Endothelium dependent vasodilatation is resistant to insulin in

Document Sample
Endothelium dependent vasodilatation is resistant to insulin in Powered By Docstoc
					Effect of insulin therapy on
endothelial function and
insulin-stimulated endothelial function
in patients with type 2 diabetes
and ischemic heart disease.

PhD thesis
Christian Rask Madsen
University of Copenhagen, February 2001.
                                                                 Christian Rask Madsen: PhD thesis. Page i.


AGE        advanced glycation end-products

ADMA       assymmetrical dimethylarginine
BH4        tetrahydrobiopterin
BMI        body mass index
DAG        diacylglycerol
DCCT       The Diabetes Control and Complications Trial
DDAH       dimethylarginine dimethylaminohydrolase
DIGAMI     Diabetes Mellitus, Insulin Glucose Infusion in Acute Myocardial Infarction
EDHF       endothelium-derived hyperpolarising factor
eNOS       endothelial nitric oxide synthase
ET         endothelin
FFA        free fatty acids
HDL        high density lipoprotein
IR         insulin receptor
IRS        insulin receptor substrate
LDL        low density lipoprotein
L-NAME     NG-nitro-L-arginine methyl ester
L-NMMA     NG-monomethyl-L-arginine
L-NNA      NG-nitro-L-arginine
MAPK       mitogen activated protein kinase
NADH       nicotinamide adenine dinucleotide (in its reduced form)
NADPH      nicotinamide adenine dinucleotide phosphate (in its reduced form)
NO         nitric oxide
PI3K       phosphatidyl inositol-3 kinase
PKC        protein kinase C
UKPDS      United Kingdom Prospective Diabetes Study
VSMC       vascular smooth muscle cell
                                                                 Christian Rask Madsen: PhD thesis. Page ii.

The Finnish researcher, Professor Hannele Yki-       now that it was established at the department.
Järvinen, often begins her lectures by               We thus planned to measure insulin-stimulated
proposing that “Diabetes is a vascular disease       glucose uptake and endothelial function in
diagnosed by measuring blood glucose”. When          patients at the same examination day. Lars
Dr. Christian Torp-Pedersen asked me, in             said that I had not described the aim clearly
1997, to start the work which has resulted in        enough in the protocol, and I realized it was
the thesis you are about to read, the only           because I had not expressed a well-defined
aspect of the project which seemed decided           hypothesis of whether insulin resistance
on from the start was that it should somehow         causes endothelial dysfunction, or the other
embrace type 2 diabetes and cardiovascular           way around. I suggested that we could study
disease. Early plans for the study was               an intervention in our patients that would be
discussed with Drs. Per Hildebrandt, Thomas          expected to improve endothelial function but
Melchior, and Lars Køber and with Christian’s        did not by other known or proposed
wife, Dr. Marie Seibæk. Before acknowledging         mechanisms       influence   insulin-stimulated
those that contributed to the study, I would         glucose uptake: antioxidants, for example.
like to describe the preparations that went          Such a study has since been made in patients
ahead of the work presented in the thesis            with essential hypertension, a condition
proper.                                              associated with insulin resistance, and it did
                                                     not support that endothelial dysfunction
Per had then accepted a position in the
                                                     causes insulin resistance [2]. In the end,
steering comittee of the DIGAMI-2 trial and
                                                     measuring insulin-stimulated glucose uptake
thought we should attempt to explain a
                                                     was abandoned at a meeting in Christian and
mechanism for the improved survival seen
                                                     Marie's house, when Marie, momentarily
with insulin therapy of patients with diabetes
                                                     recovering from absent-mindedness during the
and acute myocardial infarction. We had clever
                                                     children’s bed-time manoeuvres, said (with a
ideas that we were not perseverant enough to
                                                     very concerned look): “You want to make 60
pursue. To identify abnormalities of insulin
                                                     clamps? That’s a lot of work”. We restricted
effects in the myocardium which could be a
                                                     ourselves to measuring endothelial function.
target for insulin therapy, Thomas wrote a
protocol draft which described experiments           Christian said that there would be plenty of
aimed at measuring insulin-stimulated glucose        patients to recruit. I first screened patients
uptake during coronary catheterisation and           who had been examined by coronary
thus to measure directly whether the human           arteriography in our department. It is true that
heart is insulin resistant in patients with type 2   at least 10% of the more than 2000 patients
diabetes. This idea was abandoned because            that receives this procedure each year at our
we agreed that it would not be possible to           institution are expected to have type 2
measure the supposedly quite small coronary          diabetes. But the inclusion criterium of
arterio-venous glucose differences. But it           glycemic dysregulation and the exclusion
seems that given a sound hypothesis one              criterium of left ventricular dysfunction
should not readily accept that a study cannot        severely decimated my regimen of potential
be done for technical reasons. Measuring the         research subjects. Not being able to find the
arterio-venous glucose difference in the             patients was the most frustrating part of my
human coronary circulation is indeed                 PhD work.
feasible [1].
                                                     My first thanks go to Christian. He has a large
Marie, in her PhD work, had measured whole-          capacity for learning new things and an
body insulin sensitivity in patients with            admirable ability to organize, in his mind,
ischemic heart disease at Gentofte University        complicated and contradictory facts, and make
Hospital. Insulin resistance, like endothelial       arbitrary decisions when it is necessary. I
dysfunction, may be one of the earliest              beleive that when he respects a collegue it is
changes leading to vascular disease, and             regardless of age and accreditation. And he is
Christian was eager to preserve the                  very good at making us remember, in his own
euglycemic hyperinsulinemic clamp techique           belief, that we work because it is enjoyable.
                                                                   Christian Rask Madsen: PhD thesis. Page iii.

My supervisor and I like the idea that this            Drs. Thure Krarup and Erik Christiansen made
research will ultimatively benefit the patients,       important contributions to the protocol and to
but his and my primary motivation are selfish.         the article manuscript and were thorough and
Realizing this makes me even more amazed of            succesful at initiating and maintaining insulin
the attitude of the patients I have adressed.          therapy in the treatment group of patients.
Very few of them have denied to participate,           Anne Marie Råby-Magle kindly helped with
and     those    that    did   went     through        dietary advice to patients. I was fortunate to
uncomfortable study days and accepted                  be able to consult Dr. Philip Hougaard about
thorough changes to their life-styles during the       statistical analysis.
intervention period without second thoughts. I
                                                       Ulla Jørgensen, Anne-Mette Budde, Kirstine
admire and respect their motivation and I
                                                       Serup-Hansen, and Berit Thomsen all worked
thank both those that participated and those
                                                       hard as study nurses in clinical trials that
that wanted to, but could not.
                                                       partly financed the project. Dr. Jens Otto Lund
Two years ago, Dr. Nikolaj Ihlemann started to         provided us with office and laboratory space.
expand the work in this PhD thesis by                  Trine Schnor, pharmacist at the Central
establishing a model of local forearm                  Hospital Pharmacy of Copenhagen County,
hyperglycemia and studying the effect of               helped with procedures for producing drugs
hyperglycemia on the acetylcholine vasodilator         for infusion. Niels Jensen, engineer at Gentofte
response       and     the    insulin-stimulated       Hospital, was of assistance many times when
acetylcholine response in the human forearm.           there was technical problems with the
It has been wonderful to have someone with             laboratory equipment.
whom I could always discuss methods and
                                                       Per Hilderbrandt, Thomas Melchior, Marie
data and who could help to digest the
                                                       Seibæk, and Professor Niels-Henrik Holstein-
literature. Nikolaj has a wonderful ironic sense
                                                       Rathlou gave important intellectual input at
of humour, and at some point he will not be
                                                       various stages of the study.
able to not make fun of this preface. He is the
best collegue I have had.                              Finally, I would like to thank University of
                                                       Copenhagen for my salary, the Danish Heart
I thank Lars for saying “I’m not sure that I
                                                       Foundation      and    Emil    Torp-Pedersen's
understand this” at the right times, thus
                                                       Foundation for financial contributions, and
pointing out weaknesses in design or data
                                                       Bayer Denmark for donating blood glucose
interpretation. I am grateful to Lene Eskildsen,
                                                       analyzers for home-monitoring of blood
who screened hundreds of patient records and
                                                       glucose. The Danish Medical Research Council
was my first technical assistance, and to
                                                       and the Danish Diabetes Foundation have
Dorthe Baunbjerg Nielsen, who has been a
                                                       financially supported the projects described in
very skilful technician and data organizer.
                                                       the        section       on       perspectives.

1.      McNulty PH, Pfau S, and Deckelbaum LI. Effect of plasma insulin level on myocardial blood
        flow and its mechanism of action. Am.J Cardiol. 2000;85(2):161-5.
2.      Natali A, Sironi AM, Toschi E, Camastra S, Sanna G, Perissinotto A, Taddei S, and Ferrannini
        E. Effect of vitamin C on forearm blood flow and glucose metabolism in essential
        hypertension. Arterioscler.Thromb.Vasc.Biol. 2000;20(11):2401-6.
                                                              Christian Rask Madsen: PhD thesis. Page iv.

Background. Decreased endothelial function        affected by insulin infusion in any group. In
and blunted insulin-stimulated endothelial        healthy controls, insulin infusion increased the
function may be mechanisms responsible for        ACh response by ∼ 120%, but had no effect in
development of atherothrombotic disease in        patients. Surprisingly, both the ACh and SNP
type 2 diabetes, but it is unknown whether        response had decreased after 3 days of insulin
hypoglycemic drug therapy can modulate            therapy in the treatment group of patients.
these abnormalities. The aim of this study was    After 2 months, hemoglobin A1c in the
to examine whether long-term intensive insulin    treatment group had decreased by 2.5
therapy can improve endothelium-dependent         ± 0.4%. In this group, the ACh and SNP
vasodilation as well as the acute effect of       responses had not changed significantly, but
insulin to stimulate stimulate endothelium-       insulin stimulation now increased the ACh
dependent vasodilation.                           response by ∼ 80%. In the time control group,
                                                  vascular responses were unchanged. The
Patients and methods. Patients with type 2
                                                  inhibitory effect of L-NMMA on the insulin-
diabetes and stable ischemic heart disease and
                                                  stimulated ACh response was similar in
lean, healthy control subjects were studied.
                                                  patients and healthy controls and had not
Forearm blood flow was measured by venous
                                                  changed after 2 months in the treatment
occlusion plethysmography during dose-
                                                  group of patients.
response studies of acetylcholine (ACh) and
sodium nitroprusside (SNP) infused into the       Conclusions. The patients with type 2
brachial artery. ACh was repeated during          diabetes and ischemic heart disease in this
intrabrachial infusion of insulin, resulting in   study had decreased endothelium-dependent
high physiological serum insulin concentrations   and endothelium-independent vasodilation.
in the forearm with minor elevation of            Their vascular dysfunction may be improved
systemic serum insulin. Co-infusion of insulin    by BH4, suggesting a relative deficiency of this
an ACh was repeated during infusion of L-         cofactor in the vasculature, but further studies
NMMA, an inhibitor of nitric oxide (NO)           are needed to confirm this preliminary finding.
production. Finally, ACh infusion was repeated    In lean, healthy subjects, insulin increased
during co-infusion of tetrahydrobiopterin         endothelium-dependent vasodilation, likely by
(BH4), a cofactor or NO synthase. Patients        a local effect on the vasculature. This effect
were re-studied after 3 days and after 2          was absent in patients with poor glycemic
months of therapy with 4 daily subcutaneous       control, but present after 2 month’s insulin
insulin injections (treatment group, n = 19) or   therapy. The mediator of this effect was not
after 2 months without hypoglycemic drug          established. Surprisingly, after 3 days of
therapy (time control group, n = 9). Certain      insulin therapy, endothelium-dependent and -
dose-response studies were only performed in      independent vasodilation was paradoxically
subgroups of patients. Thus, studies with         decreased. Planned research include studies to
insulin and L-NMMA was not performed in all       determine whether tumor necrosis factor-α
patients and studies with SNP were not            has a role in mediating endothelial insulin
performed in all healthy controls. The BH4        resistance. Future research may show whether
studies were only performed in a subgroup of      therapy directed against endothelial insulin
patients.                                         resistance can prevent atherothrombotic
                                                  complications in type 2 diabetes.
Results. The ACh and SNP responses were
lower in patients compared to healthy
controls. BH4 infusion increased the ACh
response in patients. Basal blood flow was not
                                                               Christian Rask Madsen: PhD thesis. Page v.

Baggrund.         Nedsat     insulin-stimuleret    virkning        hos       patienter.      Basal
endotelfunktion er muligvis en årsag til           blodgennemstrømning blev ikke påvirket af
udvikling af hjertekarsygdom ved type 2            insulin-infusion i nogen gruppe. Hos raske
diabetes, men det vides ikke om blodsukker-        kontrolpersoner øgede insulin-infusion AK-
nedsættende behandling kan bedre denne             virkning med ∼ 120%, men påvirkede ikke AK-
tilstand. Formålet med det nærværende studie       virkning hos patienterne. AK- og NNP-virkning
var     at    undersøge    hvorvidt     intensiv   var overraskende nedsat efter 3 dages insulin-
insulinbehandling kan bedre endotel-afhængig       behandling i behandlingsgruppen. Efter 2
vasodilatation samt den akutte virkning af         måneders var hemoglobin A1c faldet 2.5
insulin på endothel-afhængig vasodilatation.       ± 0.4% i behandlingsgruppen. I denne gruppe
                                                   havde AK- og NNP-virkninger ikke ændret sig,
Patienter og metoder. I studiet undersøgtes
                                                   men insulin-stimulation øgede nu AK-virkning
patienter med type 2 diabetes og stabil
                                                   med ∼ 80%. I tidskontrolgruppen var alle
iskæmisk hjertesygdom og slanke, raske
                                                   vaskulære       effekter    uændrede.       Den
forsøgspersoner. Underarms-gennemblødning
                                                   hæmmende effekt af L-NMMA på AK-virkning
blev målt med venøs okklusions-pletysmografi
                                                   var lige stor hos patienter og raske
under       dosis-virknings-undersøgelser     af
                                                   kontrolpersoner og ændrede sig ikke efter 2
acetylkolin (AK) og natrium nitroprussid (NNP),
                                                   måneder           i        behandlingsgruppen.
infunderet i a. brachialis. AK blev gentaget
                                                   Konklusioner. Patienterne med type 2
under intra-arteriel infusion af insulin, som
                                                   diabetes og iskæmisk hjertesygdom i dette
øgede serum-koncentrationer af insulin i
                                                   studie havde både nedsat endothel-afhængig
underarmen til høje fysiologiske værdier med
                                                   og endothel-uafhængig vasodilatation. Deres
en kun lille stigning i systemisk serum-insulin.
                                                   vaskulære dysfunktion blev muligvis bedret af
Samtidig infusion af insulin og AK blev
                                                   BH4, hvilket antyder en relativ mangel på
gentaget med infusion af L-NMMA, en
                                                   denne kofaktor i vaskulært væv, men
hæmmer af nitrogenoxid (NO) produktion.
                                                   yderligere studier er nødvendige for at
Desuden blev den samtidige infusion of insulin
                                                   bekræfte dette præliminære fund. Hos slanke,
og     AK     gentaget     med     infusion   af
                                                   raske forsøgspersoner øgede insulin endothel-
tetrahydrobiopterin (BH4), en kofaktor til NO
                                                   afhængig vasodilatation, formentlig via en
syntase. Patienterne blev genundersøgt efter 3
                                                   lokal effekt på karrene. Denne effekt
dage og efter 2 måneder i behandling med
                                                   fraværende hos patienter med dårlig
insulin, givet som 4 daglige subkutane
                                                   glykæmisk kontrol, men var til stede efter 2
injektioner (behandlingsgruppe, n = 19) eller
                                                   måneders insulinbehandling. Den medierende
efter 2 måneder uden antidiabetisk behandling
                                                   faktor af denne effekt blev ikke påvist. Efter 3
(tidskontrolgruppe, n = 9). Visse dosis-
                                                   dages insulinbehandling var både endothel-
virknings-undersøgelser blev kun udført i
                                                   afhængig og -uafhængig vasodilatation
undergrupper af deltagere. Således blev
                                                   overraskende nedsat. Egne planlagte projekter
undersøgelser med insulin og L-NMMA ikke
                                                   vil søge at vise om TNF-α har betydning ved
udført hos alle patienter og studier med NNP
                                                   endotelial insulinresistens. Fremtidige studier
ikke udført hos alle raske kontrolpersoner.
                                                   vil måske kunne vise om behandling rettet
Resultater. AK- og NNP-virkninger var mindre
                                                   mod endotelial insulinresistens kan forebygge
hos patienter sammelignet med raske
                                                   hjertekarsygdom ved type 2 diabetes.
kontrolpersoner. BH4-infusion øgede AK-
                                                                         Christian Rask Madsen: PhD thesis, table of contents. Page vi.

Table of contents
Abbreviations                                                                                                                                i.
Preface                                                                                                                                     ii.
Abstract                                                                                                                                   iv.
Resumé (abstract in Danish)                                                                                                                 v.
  Introduction...................................................................................................................... 1
  Background....................................................................................................................... 1
          Endothelial function and dysfunction ................................................................................. 1
          Endothelium-derived vasorelaxing or vasoconstricting factors .............................................. 2
          Insulin sensitivity and insulin resistance............................................................................. 2
          Hyperglycemia, insulin resistance, and cardiovascular disease. ............................................ 3
  Hypotheses ....................................................................................................................... 3
  Methods ............................................................................................................................ 4
  Applied methods ..................................................................................................................... 4
  Subjects ................................................................................................................................. 4
          Patients .......................................................................................................................... 4
          Healthy controls .............................................................................................................. 4
  Experimental protocol.............................................................................................................. 4
          Preparations and measurement of blood flow..................................................................... 4
          Acetylcholine and sodium nitroprusside response ............................................................... 5
          Insulin-stimulated acetylcholine response .......................................................................... 5
          L-NMMA-inhibited         insulin-stimulated acetylcholine response................................................. 5
          Acute effect of BH4 on the acetylcholine response.............................................................. 5
  Intervention design ................................................................................................................. 5
          Patient allocation ............................................................................................................. 5
          Early and late repeat examination ..................................................................................... 6
          Acetylcholine reproducibility study..................................................................................... 6
          Biochemical analyses. ...................................................................................................... 6
  Calculations and statistical analyses .......................................................................................... 6
  Ethical considerations .............................................................................................................. 7
  Choice of methods: alternatives and criticism ............................................................................ 8
          Endothelium-dependent stimulus and vascular bed............................................................. 8
          Measurement of blood flow and determinants of variability related to experimental design .... 9
          Acetylcholine................................................................................................................. 10
          Sodium nitroprusside ..................................................................................................... 11
          L-NMMA ........................................................................................................................ 11

          Insulin-stimulated acetylcholine response ........................................................................ 11
                                                                   Christian Rask Madsen: PhD thesis, table of contents. Page vii.

       Experiments to control for the insulin-stimulated acetylcholine response ............................ 11
       Local insulin stimulation ................................................................................................. 12
       Serial pharmacological studies on each experimental day .................................................. 12
       Design of clinical intervention study ................................................................................ 13
       Patient selection ............................................................................................................ 13
       Choice of healthy control group ...................................................................................... 13
       Concommitant medication and competing risk factors for endothelial dysfunction................ 14
       Statistical analyses......................................................................................................... 14
Results ............................................................................................................................ 15
       Insulin stimulation ......................................................................................................... 15
       Metabolic control ........................................................................................................... 16
       Blood flow data ............................................................................................................. 16
       Acetylcholine response................................................................................................... 16
       Sodium nitroprusside response ....................................................................................... 16
       Insulin-stimulated acetylcholine response ........................................................................ 16
       L-NMMA      inhibition of insulin-stimulated acetylcholine response .......................................... 16
       Effect of BH4 on the acetylcholine response..................................................................... 17
       Effect of short-term insulin therapy on acetylcholine and sodium nitroprusside responses .... 17
       Effect of long-term insulin therapy on vascular responses ................................................. 17
       Acetylcholine and sodium nitroprusside responses and insulin-stimulated responses in the time
       control group ................................................................................................................ 17
       Acetylcholine reproducibility ........................................................................................... 17
       Systemic circulatory responses........................................................................................ 17
Discussion....................................................................................................................... 17
Interpretation of own results.................................................................................................. 17
       Acetylcholine responses before treatment ........................................................................ 17
       Local insulin stimulation ................................................................................................. 18
       Metabolic effects of insulin treatment .............................................................................. 18
       Insulin-stimulated acetylcholine response ........................................................................ 18
       Effect of BH4 on the acetylcholine response..................................................................... 19
       Effect of short-term insulin therapy on acetylcholine and sodium nitroprusside responses .... 19
       Effect of long-term insulin therapy on acetylcholine and sodium nitroprusside responses ..... 19
       Effect of long-term insulin therapy on insulin-stimulated acetylcholine response.................. 19
       L-NMMA      inhibition of the insulin-stimulated acetylcholine response .................................... 20
Discussion of related previous literature .................................................................................. 20
       High glucose concentrations and endothelial function ....................................................... 20
       Effects of insulin on vasodilation in vivo in animals and healthy humans ............................. 21
       Effects of insulin on vasodilation in insulin resistant subjects ............................................. 22
       Mediators of insulin-stimulated vasodilation ..................................................................... 22
                                                                    Christian Rask Madsen: PhD thesis, table of contents. Page viii.

       Insulin-stimulated PI3K-Akt-eNOS activation .................................................................... 23
       Effects of insulin on eNOS expression .............................................................................. 24
       Mechanisms of endothelial insulin resistance .................................................................... 24
       Effects of insulin on vascular smooth muscle cell function in vitro ...................................... 24
       Effects of insulin on endothelium-independent vasorelaxation ex vivo or vasodilation in vivo 26
       Hypoglycemic or insulin sentisizing interventions and endothelial function in vivo ................ 26
       Hypothetical mechanism of improved insulin-stimulated acetylcholine response .................. 26
       Endothelial dysfunction as a determinant of peripheral insulin resistance ............................ 27
Conclusions..................................................................................................................... 27
Implications.................................................................................................................... 27
       Implications for improved endothelial function and clinical outcomes.................................. 27
       Implications for the concept of endothelial insulin resistance ............................................. 28
Perspectives.................................................................................................................... 29
       BH4.............................................................................................................................. 29
       TNF-α........................................................................................................................... 29
       Birth weight .................................................................................................................. 29
       FFAs and PKC................................................................................................................ 30
References ...................................................................................................................... 31
Table of contents
Tables 1-5
Figures 1-18
Article manuscript
                                                           Christian Rask Madsen: PhD thesis, Page 1 of 40.

The world-wide prevalence of type 2 diabetes
                                                   Endothelial function and dysfunction
is increasing because sedentary lifestyles and     Functions of the vascular endothelium are
obesity are becoming more frequent, and it is      numerous. In short, these functions include
estimated that 221 million peple will have type    vasomotor control, regulation of thrombosis
2 diabetes in 2010 (1). The most important         and fibrinolysis, influence on inflammatory
complication    of   type    2    diabetes    is   responses, and regulation of vascular
cardiovascular disease. At my institution,         remodelling. For example, the endothelium
almost half of patients referred for coronary      synthesizes vasorelaxing factors acting locally
arteriography are suffering from type 2            in an autocrine or paracrine fashion and
diabetes or impaired glucose tolerance (2).        controls mononuclear cell adhesion to its own
Endothelial dysfunction is regarded as the         surface by regulating the expression of
                                                   adhesion molecules. Endothelium-dependent
earliest detectable causal factor in the
                                                   vasorelaxation and mononuclear cell adhesion
development of atherothrombotic disease (3).
This abnormality is found in patients with type    have been the endothelial functions most
2 diabetes and in subjects with insulin            commonly examined in vitro or in vivo.
resistance (4), a central factor in the            The most thoroughly studied endothelium-
pathogenesis of type 2 diabetes. Furthermore,      derived vasodilating factor is NO, which has
the degree of endothelial dysfunction is           been shown to be a mediator of all the
associated with the severity of insulin            endothelial functions mentioned above. In
resistance (5-7),   and    hyperglycemia      is   vascular tissue, it is synthesized by the isoform
associated with endothelial dysfunction (see       of NO synthase called endothelial NO synthase
page 20). As insulin resistance and                (eNOS). Apart from the synthesis of NO, it has
hyperglycemia, the two most important extra-       been suggested that, during physiological
pancreatic abnormalities in type 2 diabetes,       conditions, the enzyme catalyzes the
are both independent risk factors for              production of superoxide, which may regulate
cardiovascular disease, they may determine         gene expression (14) or be a precursor for a
the development of atherothrombosis through        putative hyperpolarizing and vasodilating
detrimental    effects   on     the    vascular    factor, hydrogen peroxide (15).
                                                   There are several lines of evidence to suggest
Insulin stimulates endothelial function, but       that endothelial dysfuntion, especially NO-
individuals with insulin resistance associated     mediated endothelial function, has a causal
with obesity (8) or type 2 diabetes (9) are        role in atherothrombotic disease. As an
resistant to effects of insulin on the             example, a single reference to a review or
endothelium. Accordingly, insulin resistance       original article is mentioned with each item.
may exist at the level of the endothelium,
which is supported by studies in vitro (10,        1. NO inhibits cellular mechanisms thought to
11).                                                  be early determinants of atherosclerosis
Insulin therapy improves both glycemic control        and thrombosis (16).
and insulin sensitivity in patients with type 2    2. Endothelial dysfunction is temporally
diabetes (12). In addition, insulin therapy of        related to the development and regression
patients with type 2 diabetes, started during         of atheroma in animal models of
admission for acute myocardial infarction,            atherosclerosis (17).
improve their long-term mortality (13). The
                                                   3. Inhibition of vascular NO production
aim of the present study was to examine
                                                      accelerates atheroma formation in the
whether insulin therapy improved endothelial
                                                      abscence of increased blood pressure in
function and insulin-stimulated endothelial
                                                      animal models of atherosclerosis (18).
function in patients with type 2 diabetes and
ischemic heart disease.                            4. Gene transfer of NO synthase inhibits
                                                      atheroma formation in animal models of
                                                      atherosclerosis (19).
                                                   5. Endothelium-dependent vasodilatation of
                                                      coronary and limb conduit and resistance
                                                          Christian Rask Madsen: PhD thesis, Page 2 of 40.

   vessels is decreased in subjects with risk     experimental conditions (30, 31). Among
   factors for cardiovascular disease (20).       endothelium-derived      vasoconstrictors are
                                                  prostaglandins (e.g. thromboxane A2) and
6. Endothelial dysfunction is associated with
                                                  endothelins (ETs) (e.g. ET-1).
   future risk for cardiovascular events in
   patients with established coronary artery      Insulin sensitivity and insulin resistance
   disease in small selected populations (21-
   23).                                           Insulin has effects on carbohydrate, lipid, and
                                                  amino acid metabolism. In glucose metabolism
7. In subjects with risk factors for vascular     alone, several intermediary reactions are
   disease, therapy known to prevent              regulated by insulin. However, insulin
   cardiovascular     events  also   improve      sensitivity is frequently synonymous with the
   endothelial function (24).                     ability of insulin to stimulate glucose uptake.
8. Certain human eNOS polymorphisms are           Insulin-stimulated glucose uptake can be
   associated with impaired endothelium-          measured in vivo by the hyperinsulinemic
   dependent vasodilatation as well as            euglycemic clamp (32), which was used in
   cardiovascular disease in case-control         many of the studies referred to in this thesis.
   studies. Such polymorphisms include a          The principle of this test is to infuse insulin
   G894T mutation in exon 7 (with the             intravenously at a constant dose, often
   corresponding change in amino acid             sufficient to cause complete inhibition of
   sequence Glu298Asp) (25); a mutation in        hepatic glucose output. Hyperinsulinemia
   intron 4 (4a/b); and a T786C mutation in       induced this way would quickly lead to
   the 5’-end of the eNOS promotor region.        profound hypoglycemia, but during the test,
                                                  glucose is infused at a variable rate to
9. Treatment that likely has an effect            maintain plasma glucose at a constant
   specifically on endothelial dysfunction        concentration in the normal range. Whole-body
   improves clinical end-points of ischemic       insulin sensitivity can thus be expressed as the
   disease. There is very sparce information      rate of glucose infusion necessary to maintain
   to support this point, but an example is a     euglycemia (32).
   study where oral supplements of L-
   arginine, the substrate for NO synthase,       Insulin resistance is the result of a decreased
   improved walking distance in patients with     effect of insulin to stimulate glucose uptake, or
   intermittent claudication and peripheral       the result of requirements for increased insulin
   arterial occlusive disease, supposedly by      concentrations to stimulate the same glucose
   restoring vascular NO formation and            uptake. Skeletal muscle is quantitatively the
   vasodilator function (26).                     most important tissue responsible for insulin
                                                  resistance (33).
Endothelium-derived        vasorelaxing     or
                                                  Insulin resistance is probably the earliest
vasoconstricting factors
                                                  abnormality in the pathogenesis of type 2
Several factors synthesized in and released       diabetes (34). Transition from the insulin-
from the vascular endothelium has effects on      resistant nondiabetic phenotype to the state of
local vasomotor function. NO, which is a lipid-   diabetes, however, also requires development
soluble gas, diffuses to VSMC and is              of a relative β-cell insufficiency (35). The
responsible for relaxetion in response to cGMP    sequence of development of insulin resistance
produced after activation of soluble guanylate    and deficient insulin production, respectively,
cyclase when this enzyme binds NO. The            in the natural history of type 2 diabetes is
vasorelaxing pathway distal to cGMP formation     controversial (36). Insulin resistance may
in VSMC is incompletely understood (27). Apart    result from genetic causes as well as acquired
from NO availability, NO-cGMP-mediated            factors, primarily obesity (37). Circulating
vasodilation is regulated by guanylate cyclase    molecules secreted by adipocytes represent a
activity (28).                                    major focus to understand the association
                                                  between obesity and insulin resistance.
Other endothelium-dependent vasodilators are
                                                  Furthermore, such substances are particularly
prostacyclin      and      endothelium-derived
                                                  important for the discussion in the present
hyperpolarizing factor (EDHF). EDHF is
                                                  study because they may explain the
probably representative of several factors not
                                                  simultaneous occurence of derangements in
yet completety characterized. Endothelium-
                                                  insulin-dependent glucose transport and
derived vasodilating factors may substitute for
                                                  metabolism and insulin-dependent endothelial
each other during pathological (29) or
                                                              Christian Rask Madsen: PhD thesis, Page 3 of 40.

function (4). Such adipocyte-derived signals          (DIGAMI) study, which included patients with
include free fatty acids (FFA) (see page 24),         diabetes (> 80% type 2) admitted for acute
TNF-α (see page 29) and leptin (37), but in the       myocardial infarction, showed a relative risk
near future, newly discovered factors (38) may        reduction of 28% for total mortality (13)
prove to be important.                                extended to over 3 years after the
                                                      infarction (46). One purpose for the on-going
The molecular causes of insulin resistance are
                                                      DIGAMI-2 trial is to determine whether this
incompletely understood. Since the rate-
                                                      effect is due to insulin-glucose-potassium
limiting step in insulin-stimulated glucose
                                                      infusion during the first few days of the acute
metabolism is thought to be glucose
                                                      phase of myocardial infarction or due to long-
uptake (39), much effort is directed at
                                                      term effects.
identifying the molecular causes of alterations
in     the    insulin    signaling    transduction    Insulin resistance has been shown in several
pathway (40) and in the translocation of the          studies to be associated with an increased risk
glucose transproter protein 4 (GLUT4) from            of cardiovascular disease (47), and a
intracellular     vesicles       to    the     cell   prospective study of the role of insulin
membrane (41). The insulin signaling pathway,         resistance is on-going (48). The results of the
which will be discussed in several sections of        Diabetes Control and Complications Trial
this thesis (see page 23) is activated by the         (DCCT) (44) and the UKPDS (45) have refuted
binding of insulin to the insulin receptor (IR).      previous concerns for the safety of insulin
This stimulates tyrosine autophosporylation of        therapy, based on observations in certain
the IR and tyrosine phosporylation and                animal studies that insulin had pro-atherogenic
activation of insulin receptor substate (IRS).        properties. Although fasting circulating insulin
IRS acts as a docking protein for phosphatidyl        concentratrations have been shown to be an
inositol-3 kinase (PI3K), which it activates. The     independent risk factor for cardiovascular
primary        substrate       for     PI3K      is   disease (49), the current prevailing view is that
phosphatidylinositol       (4,5)-phosphate.     Its   elevated basal insulin levels in patients with
product phosphatidylinositol (3,4,5)-phosphate        type 2 diabetes is a marker of insulin
interacts with Akt (protein kinase B) (42) which      resistance, resulting from a compensatory
is recruited to the cell membrane and activated       increase of insulin secretion to maintain
by serine/threonine phosphorylation by a              insulin-stimulated glucose uptake (47).
kinase not yet identified (41). The final
                                                      Several mechanisms linking hyperglycemia and
coupling between Akt and GLUT4 translocation
                                                      insulin resistance to endothelial dysfunction
has not been identified. There may be parallel
                                                      have been described (see page 20).
signaling pathways than the one outlined here,
including activation of protein kinase C (PKC) ζ
and PKC λ (41). Decreased expression or               Hypotheses
impaired activity of the IR, IRS, and Akt may
                                                      The study aimed at testing the following
each play a role in insulin resistance (41).
                                                      hypotheses, pertaining to patients with type 2
Hyperglycemia,     insulin     resistance,     and    diabetes in poor glycemic control and
cardiovascular disease.                               complicated with ischemic heart disease,
                                                      versus age-matched, lean, healthy control
Several prospective observational studies have        subjects:
shown hyperglycemia to be an independent
risk factor for cardiovascular disease (43).          1. In patients compared to healthy controls,
Disappointingly, insulin therapy to patients             endothelium-dependent vasodilation is
with type 1 diabetes (44), or and intensive              depressed      whereas        endothelium-
treatment policy with insulin or oral                    independent vasodilation is unaffected.
sulphonylureas to patients with type 2                2. In healthy controls, local forearm
diabetes (45) did not improve cardiovascular             hyperinsulinemia   acutely      increases
events or total mortality, although the The              endothelium-dependent vasodilation.
United Kingdom Prospective Diabetes Study
(UKPDS) study (45) did show a non-significant         3. In patients, the effect of local forarm
16% reduction of the relative risk of                    hyperinsulinemia     on       endothelium-
myocardial infarction in patients managed with           dependent vasodilation is decreased
an intensive hypoglycemic treatment policy.              compared to healthy controls.
However, the Diabetes Mellitus, Insulin
Glucose Infusion in Acute Myocardial Infarction
                                                           Christian Rask Madsen: PhD thesis, Page 4 of 40.

4. In patients, insulin therapy administered       last few years were screened and potential
   with the goal of optimizing glycemic control    participants contacted by telephone. A decision
   improve             endothelium-dependent       to include or exclude each patient was made
   vasodilation; and this effect is dependent      after an out-patient visit. All patients who were
   on long-term therapy (2 months) rather          found eligible and who accepted participation
   than short-term therapy (3 days).               were included.
In patients, insulin-stimulated endothelium-       Details of clinical characteristics of patients are
dependent vasodilation is improved by long-        listed in tables 1 and 2. All patients except two
term insulin therapy; and this improvement is      had a history of hypertension, and ACEI
due to increased nitric oxide (NO)-mediated        therapy was originally initiated because of
vasodilation.                                      hypertension in all cases.

                                                   Healthy controls
                                                   Healthy volunteers were recruited after
Applied methods                                    newspaper advertisements. They were non-
SUBJECTS                                           smoking and did not take high-dose vitamin
                                                   supplements. All were lean (BMI < 25 kg m-2)
Patients                                           and all had normal fasting blood glucose (< 5.6
                                                   mmol l-1) on two seperate occasions, normal
Patients with type 2 diabetes and ischemic
                                                   plasma cholesterol (< 6.0 mmol l-1) and a
heart disease were included. Ischemic heart
                                                   normal electrocardiogram.
disease was documented by prior myocardial
infarction    or    coronary    atherosclerosis    Details of clinical characteristics of healthy
demonstrated during angiography. Thus, this        controls are listed in table 1.
diagnosis was confirmed by prior acute
                                                   EXPERIMENTAL PROTOCOL
myocardial infaction in 10 patients, and
because of examinations made in relation to        Preparations and measurement of blood flow
coronary artery by-pass grafting in 15 patients
and percutaneous transluminal coronary plasty      In patients, oral hypoglycemic drugs were
in 3. To ensure inclusion of patients with a       withdrawn 2 weeks before the study. All
typical type 2 diabetes phenotype and a high       medicine was interrupted for 24 hours
probability of remaining clinically stable         preceeding any of the examinations, and the 4
without their usual hypoglycemic medicine,         patients who were smokers abstained from
inclusion was restricted to obese patients         smoking for 8 hours before each study. Both
(body mass index (BMI) > 27 kg m−1) with           patients and healthy controls arrived at the
diabetes onset at age 50 years or older and        laboratory at 8 AM after 8 hours of fasting. An
with a diabetes duration of > 2 years without      arterial cannula with an external diameter of 1
periods of ketoacidosis. The mean diabetes         mm (Ohmeda, Swindon, UK) was inserted into
duration was 7.6 ± 1.2 years. In addition, the     the brachial artery of the non-dominant arm.
patients were required to have a hemoglobin        During the whole experiment, isotonic saline
A1c (HbA1c) ≥ 8.0%.                                was infused through the cannula with
                                                   intermittent co-infusion of solutions of
Patients were only included if they had been       vasoactive drugs, keeping the total infused
receiving a statin (atorvastatin 10-20 mg or       volume constant at a rate of 1 ml min−1. A
simvastatin 10-20 mg) for > 3 weeks if             venous cannula was placed in the medial
clinically indicated (total cholesterol without    cubital vein of both arms for aspiration of
cholesterol-lowering therapy ≥ 5.0 mM).            blood samples. In the perfused arm, this
Patients with systolic dysfunction (left           cannula was inserted retrogradely into a deep
ventricular ejection fraction > 50%) or            vein where the tip could not be palpated, to
decreased renal function (plasma creatinine >      allow aspiration of blood from metabolically
130 mmol/l) were excluded.                         active muscle tissue, rather than from
Recruitment of patients took place at a            skin (50).
department of cardiology and at two                The study subjects were examined in the
departments of endocrinology at hospitals in       supine position with both forearms at
Copenhagen, Denmark (the teaching hospitals        horizontal level with the right atrium. Forearm
of Gentofte, Frederiksberg, and Hvidovre).         blood flow was measured by an electrically
Patient records from hospitalizations within the   calibrated strain gauge venous occlusion
                                                            Christian Rask Madsen: PhD thesis, Page 5 of 40.

plethysmograph (D.E. Hokanson, Bellevue,            Insulin-stimulated acetylcholine response
WA, USA) (51). Upper arm cuffs were inflated
                                                    Following the sodium nitroprusside dose-
to 40 mm Hg for ~ 15 seconds with ~ 15
                                                    response study, the acetylcholine dose-
second intervals. Measurements were made
                                                    response study was repeated during co-
simultaneously in both arms immediately
                                                    infusion of insulin (Actrapid, Novo Nordisk
before the start of any drug infusion and after
                                                    Scandinavia, Malmö, Sweden) 0.05 mU kg−1
completion of the periods described for each
                                                    min−1 started 20 minutes before and
drug infusion below. Any infusion period was
                                                    maintained during the dose-response study.
extended by the duration of blood flow
                                                    This study was performed in all patients and in
measurement, which lasted approximately 2
                                                    the same 10 consecutively included healthy
minutes. Generally, four recordings were made
                                                    controls mentioned above (see figure 2a).
with the plethysmograph. The recordings are
sensitive to mechanical noise, i.e. artifacts due   L-NMMA-inhibited                  insulin-stimulated
to voluntary or involuntary muscle contractions     acetylcholine response
in the forearm, movement of the arm because
of breathing, and untowards contact between         Following the insulin-stimulated dose-response
the strain gauge and, for examle, the bed           study, the combined insulin and acetylcholine
cover or infusion lines. Furthermore, large         study was repeated during co-infusion of NG-
variations in blood flow could sometimes            monomethyl-L-arginine      acetate   (L-NMMA)
readily be indentified during recordings. When      (Clinalfa) 1.6 and 3.3 mg min-1 for 5 minutes
such artifacts or large variations were             at each dose preceeding acetylcholine infusion
recognized during recordings, an extra              and maintained at 3.3 mg min-1 during
recording or extra recordings were made until       acetylcholine infusion. This study was carried
at least three stable recordings had been           out in the last 17 consecutively included
obtained. After the examination, the tangent to     patients and in the same 10 healthy controls
each curve was marked with pencil on the            as mentioned above (see figure 2c).
printed recording using a ruler, and the
                                                    Acute effect of BH4 on the acetylcholine
inclination of the tangent measured on a
digitizing board connected to a personal
computer with custom-made software. The             As a pilot study for a planned study (see the
blood flow was calculated as the mean of all        section on perspectives), the effect of
stable recordings.                                  tetrahydrobiopterin (BH4) (Clinalfa) on the
                                                    acetylcholine response was examined as the
The protocol for the pharmacological dose-
                                                    last of 4 dose-reponse studies in that days
response studies are shown in figures 2a-2c.
                                                    infusion protocol (see figure 2a). This study
Acetylcholine    and    sodium     nitroprusside    was carried out in the first 7 consecutively
response                                            included patients and was thus preceeded by
                                                    dose-response studies of acetylcholine, sodium
Accumulated dose-response studies were              nitroprusside, and acetylcholine during insulin
made with acetylcholine chloride (Clinalfa,         stimulation. To avoid oxidation, BH4 was
Läufelfingen, Switzerland) 7.5, 15, 30, and 60      brought into solution immediately before use
µg min−1 for 5 minutes at each dose (total          in saline bubbled with nitrogen. BH4 500 µg
infusion time at each dose appoximately 7           min−1 was infused for 5 minutes preceeding a
minutes including duration of blood flow            repeated acetylcholine dose-response study.
measurement). Subsequently, a dose-response         As for sodium nitroprusside, BH4 was
study with sodium nitroprusside dihydrate           administered in opaque syringes and infusion
(Roche, Basel, Switzerland) 1, 3, and 10 µg         lines.
min−1 were made with the same duration of
each dose as for acetylcholine. Because of its      INTERVENTION DESIGN
sensitivity to light, sodium nitroprusside was
                                                    Patient allocation
administered in opaque syringes and infusion
lines.                                              A flow diagram of the intervention design is
                                                    shown in figure 1. The first 18 patients
The acetylcholine response was examined in
                                                    consecutively included were randomised 1:1 to
all patients and all healthy controls, and the
                                                    a treatment group and time control group. To
sodium nitroprusside response was examined
                                                    obtain a more precise estimate of the changes
in all patients and in 10 consecutively included
                                                    in insulin-stimulated endothelial function, 10
healthy controls.
                                                    additional patients were allocated to the
                                                            Christian Rask Madsen: PhD thesis, Page 6 of 40.

treatment group. Clinical characteristics of        patient at the late repeat examination day due
patients, according to treatment group, are         to technical reasons in both cases. Therefore,
listed in table 2 and 3.                            the L-NMMA studies in the treatment group
                                                    only included 11 paired observations.
Patients in the treatment group started insulin
therapy in the evening of the initial               Acetylcholine reproducibility study
examination, following a similar administration
of insulin as the one used for long-term            An acetylcholine reproducibility study was
treatment in the DIGAMI trial (13). Thus, they      performed at the early repeat examination in 9
injected themselves subcutaneously with fast-       patients and at the late repeat examination in
acting insulin (Actrapid Pen, Novo Nordisk) 3       4 patients by a series of 3 dose-response
times daily at meals and intermediate-acting        studies      with    acetylcholine,    sodium
insulin (Insulatard Pen, Novo Nordisk) at bed-      nitroprusside, and acetylcholine, respectively
time. Patients were given instructions              (see figure 2b).
regarding insulin dose and diet several times
                                                    Biochemical analyses.
daily in the beginning of the treatment period
and at least once a week in the later stages of     HbA1c was measured by latex enhanced
the treatment period.                               turbidimetric immunoassay and expressed
                                                    relative to hemoglobin measured by the
The patients in the time control group
                                                    colorimetric cyanide free alkaline hematin
continued without hypoglycemic drugs while
                                                    method (COBAS MIRA, F. Hoffmann-La Roche,
monitoring and reporting their own blood
                                                    Basel, Swizterland).
glucose as a safety measure.
                                                    Blood samples were drawn immediately before
Early and late repeat examination                   measurements of blood flow. Blood glucose
In the first 9 consecutively included patients in   was analysed in full blood by enzymatic
the treatment group, acetylcholine and sodium       colometry (COBAS MIRA) and insulin was
nitroprusside responses were re-examined            analysed in serum by double antibody
after 3 days (“early repeat examination”)           radioimmunoassay      (Insulin     RIA     100,
according to the protocol illustrated in figure     Pharmacia & Upjohn Diagnostics, Uppsala,
2b. In all patients, the acetylcholine and          Sweden). Blood pressure and heart rate were
sodium nitroprusside responses were re-             read from a recording of the intra-arterial
examined after 2 months (“late repeat               pressure    and      an      electrocardiogram
examination”). At the early and late repeat         immediately after each measurement of blood
examination, patients in the treatment group        flow.
took no insulin, the last insulin dose being the    CALCULATIONS AND STATISTICAL ANALYSES
intermediate-acting insulin at bed-time the
night before the study. Medical therapy             Analyses of blood flow during acetylcholine or
additional to hypoglycemic drugs was                sodium nitroprusside does-response studies
unchanged throughout the study, except in           are described below. Comparisons of
one patient in the time control group who           catagorical variables of patients characteristics
stopped atorvastatin due to side effects. One       were made with tests of proportions as well as
patient in the time control group was taken out     exact binomial testing (Fisher’s exact test). All
of the study before the late repeat examination     other comparisons were analyzed with two
because of symptomatic hyperglycemia, but           sample or paired t       tests as appropriate.
data from the initial examination were included     Assumptions of Normality was tested by visual
in analyses.                                        inspection of Normal plots and assumptions of
                                                    equal variances in two sample t tests were
The insulin-stimulated acetylcholine response       tested with the F test (variance ratio test)
was re-examined at the late repeat                  (Microsoft Excel 97 software).
examination in the last 17 consecutively
included patients (13 in the treatment group, 4     Blood flow responses during dose-response
in the time control group) (see figure 2c). The     studies were compaired by mixed models using
L-NMMA         inhibited       insulin-stimulated   the mixed procedure in the Statistical Analysis
acetylcholine response was re-examined at the       Software, version 8 (SAS Institute, Cary, NC,
late repeat examination in the same 17              USA). Blood flow was log transformed because
patients. However, the L-NMMA study had not         the residual variation increased for increasing
been accomplished in one patient at the initial     values of predicted blood flow (the
examination day and was not completed in one        untransformed plot of residuals as a function
                                                            Christian Rask Madsen: PhD thesis, Page 7 of 40.

of predicted blood flow was “<” shaped),            Ethical considerations
whereas the residual variation based on
                                                    The study was carried out according to the
transformed data was independent of
                                                    Declaration of Helsinki II. It was approved by
predicted blood flow (the residual plot had an
                                                    the local ethics comittee and by the Danish
even scatter).
                                                    Medicinal Agency. All subjects gave written
Pharmacological study (e.g., the acetylcholine      informed consent.
dose-response study or insulin-stimulated
                                                    The primary ethical considerations of the study
acetylcholine dose-response study) entered the
                                                    pertained to the risk of arterial cannulation. In
model as a fixed effect, thus testing for a
                                                    a review of 8208 catheterisations (the majority
systematic difference of overall effect between
                                                    being left-sided heart catheterisations, most
two dose-response studies. The interaction
                                                    frequently with cannulation of the femoral
between pharmacological study and dose of
                                                    artery), the incidence of complications
vasodilator     (i.e.,   the    dose-dependent
                                                    requiring operation was 0.28% (52). These
progression of either the acetylcholine or
                                                    complications        include        hemorrhage,
sodium nitroprusside response) also entered
                                                    pseudoaneurism, or thrombosis. The arterial
the model as a fixed effect, thus testing for a
                                                    cannulas used in the experiments in the
systematic difference of dose-dependent
                                                    present study have a narrower outside
progression of the vasodilator response in each
                                                    diameter (1.0 mm), are shorter (4.5 cm) and
dose-response study. With a linear dose-
                                                    have less contact with the arterial wall than
response curve, this would translate into a
                                                    catheters used for heart catheterisation;
different inclination of the acetylcholine dose-
                                                    therefore, the risk for complications may be
response curve and the insulin-stimulated
                                                    smaller than that referred to above.
acetylcholine dose-esponse curve, independent
                                                    Accordingly, an British research group has
of any difference in overall effect). Other fixed
                                                    carried out cannulation of the brachial artery
effects were group allocation (i.e., patients
                                                    to perform experiments similar to those in the
versus healthy controls and treatment group
                                                    present study (51). Furthermore, a group at
versus time control group); and experimental
                                                    the University of Pisa, Italy, has performed
day. Study subject and the interaction between
                                                    approximately 2000 similar experiments
study subject and dose of vasodilator (i.e., the
                                                    without observing serious complications
dose-dependent progression of the vasodilator
                                                    (Stefano Taddei, personal communication).
response in each study subject) entered the
model as random effects, thus taking into           Another concern was the withdrawel of oral
account the supposedly random variation of          hypoglycemic drug therapy from the patients
both the overall individual response and the        in the time control group. Generally, there is
dose-dependent progression of the vasodilator       no reason to beleive that the long-term
response, both with a mean of zero. P-values        prognosis for patients such as those in this
from tests of the interactions mentioned will       study is affected by withdrawing such therapy
not be quoted. The statistical models used are      for a period of 8 weeks. In fact, at the time of
further described on page 14.                       initiation of the study there was no evidence
                                                    from clinical trials indicating that hypoglycemic
Statistical significance was defined as a two-
                                                    drug therapy improved the prognosis in
sided p < 0.05. Values (including geometric
                                                    patients with type 2 diabetes. Thus,
symbols with error bars in the figures) are
                                                    withdrawel of oral hypoglycemic drug therapy
presented as mean and standard error of the
                                                    according to protocol was found to be safe,
mean, unless otherwise specified.
                                                    given sufficient monitoring of metabolic
Differences between values for plasma lipids,       control. Symptomatic hyperglycemia was a
HbA1c, insulin, and glucose were calculated as      predefined secondary exclusion criteria.
the mean of individual differences and may
                                                    Finally, the risk of hypoglycemia in the patients
therefore not be identical to differences
                                                    in the treatment group had to be considered.
between mean values of groups (as read in the
                                                    In patients with type 2 diabetes, the risk of
                                                    severe hypoglycemia is relatively small (45).
Please note that in the cases where graphs
                                                    Attempts to prevent complications as the ones
show paired data, observations were included
                                                    discussed were made by adequate training of
only for patients with complete data. However,
                                                    the person that performed the arterial
statistical analyses considered all observations.
                                                    cannulations (the author of this thesis); by
                                                    cannulating the non-dominant arm with the
                                                           Christian Rask Madsen: PhD thesis, Page 8 of 40.

purpose of limiting short- or long-term            endothelium-dependent agonists. As in the
functional loss if complications to cannulation    coronary circulation, acetylcholine is usually
should occur; by instructions to the subjects to   administered as intra-arterial infusion to avoid
avoid physical exercise for 24 hours after         systemic effects, i.e. hypotension and
arterial cannulation; by teaching self-            secondary reflex vasoconstriction (51).
monitoring of blood glucose to patients in both
                                                   Increased     forearm    blood    flow    during
treatment groups; by frequent telephone
                                                   acetylcholine infusion in the human forearm
consultations by the author to monitor
                                                   model is determined by vasodilation of
metabolic control of the patients in both
                                                   resistance vessels, i.e. arterioles and small
treatment     groups;    and     by    thorough
                                                   arteries.    In     contrast,     flow-mediated
explanation of the symptoms of hypo- and
                                                   vasodilation (61, 62) has been used as a
hyperglycemia, and the actions to be taken if
                                                   measure         of      endothelium-dependent
such symptoms were to occur.
                                                   vasodilation of a conduit arteries (usually the
Patients with diabetes and ischemic heart          brachial artery) (63). Here, the stimulus is
disease have an adverse prognosis (13). As         shear stress, i.e. the tangential force on the
the study was expected to yield information        vessel wall exerted by the flowing blood. Flow-
with potential importance for the future           mediated vasodilation is non-invasive, both by
improvement of prevention and treatment of         virtue of the experimental stimulus, post-
cardiovascular disease in such patients, the       ischemic hyperemia, as well as the fact that
risks described above were found to be in          vessel diameter can be measured by
proportion to the possible benefits of study.      ultrasound; it has been shown to be
                                                   endothelium-dependent (in animals (64)) and
Choice of methods: alternatives and
                                                   mediated by NO (in humans (65)); it is
                                                   probably the most important physiological
Endothelium-dependent stimulus and vascular        stimulus for endothelium-derived NO (66); and
bed                                                it correlates with coronary endothelium-
                                                   dependent vasodilation (67). In contrast,
Although a plethora of important vascular          agonist-stimulated vasodilator response in the
endothelial functions exists, only a few           human forearm model has never been
different strategies have been used to measure     compared       with   coronary      endothelium-
it in humans in vivo. Investigators have           dependent vasodilation or brachial artery flow-
studied plasma concentrations of markers of        mediated vasodilation. The relative importance
endothelial activation, such as von Willebrandt    of the different mediators of the acetylcholine
factor and circulating adhesion molecules (53);    vasodilator response of forearm resistance
concentrations of nitrate (NO3−) and nitrite       vessels are not identical with those in conduit
(NO2−), the stable catabolites of NO, in plasma    vessels where atherosclerosis occur (68)). The
or urine (54); or whole-body 15N-arginine-15N-     primary reason that the acetylcholine response
citrulline kinetics (55). However, most or the     is perceived as a measure of endothelial
literature in this field describe endothelium-     function with implications for the pathogenesis
dependent vasodilation. It is directly involved    of atherothrombotic disease is the well-
in normal vasomotor control (56, 57) and is        established fact that the classical risk factors
probably responsible for symptoms because of       for atherosclerosis - hypercholesterolemia,
insufficient circulation (58). In addition, as     hypertension, and diabetes - are associated
several endothelial functions are commonly         with      abnormal      endothelium-dependent
affected       simultaneously,     endothelium-    vasodilation in both the coronary arteries (69-
dependent vasodilation is frequently viewed as     71), the brachial artery (63) (72) (73), and
a measure of general endothelial function (20).    forearm resistance vessels (74-76) (references
Endothelium-dependent vasodilation was first       only made to the first published studies in
studied in man by visualizing coronary artery      each catagory).
diameter by arteriography during intra-            Helmuth Drexler and co-workers have
coronary infusion of acetlycholine (59). Since     combined advantages of the forearm model
the    introduction   of   measurement       of    and flow-mediated vasodilation by using one-
endothelium-dependent vasodilation in the          dimensional ultrasound (resembling A-mode
forearm (in the following referred to as “the      echocardiography, no longer in clinical use) to
human forearm model”) (60), this method has        image flow-mediated vasodilation of the radial
become the most popular in studying                artery during pharmacological manipulation in
vasodilator responses to acetylcholine or other    the forearm by drug infusion into the brachial
                                                           Christian Rask Madsen: PhD thesis, Page 9 of 40.

artery. Endothelium-dependent vasodilation         transducer or flow probe, for example, may be
has also been measured as the response of          applied or placed), but excluding blood supply
forearm skin perfusion to acetylcholine applied    that does not result in distension of the
by iontophoresis (77, 78), or as maximum           forearm during venous return, for example
hyperaemic response to local heating of the        part of the blood supply to bone. Thus,
foot skin (79).                                    different methods may estimate blood flow to
                                                   various tissues (muscle, skin, bone) differently,
We chose to use the technique of intra-
                                                   and furthermore, hyperemic stimuli (exercise,
brachial infusion of acetylcholine because it
                                                   heating, drugs) may differentially increase
can be combined with local insulin stimulation
                                                   blood flow in certain tissues.
(see below), does not require extensive
training to operate adequately (as does one-       In 8 patients with valvular or ischemic heart
dimensional ultrasound, Helmuth Drexler,           disease, measurements with plethysmography
personal communication), and does not have         of blood flow at rest or stimulated by isotonic
limitations in obese subjects (as with two-        hand grip correlated well with measurements
dimensional ultrasound of the brachial artery      using an electromagnetic flow-meter (85). The
because of impaired image quality, Keld            flow-meter was applied to the surgically
Sørensen, personal communication).                 isolated brachial artery and calibrated against
                                                   controlled blood flow in canine femoral
In the present study, acetylcholine (and
                                                   artery (85). The correlation was best at blood
sodium nitroprusside, see page 10) was
                                                   flows below 15-20 ml (100 ml)−1 min−1 (the
examined in a cumulative dose-response
                                                   flow range relevant to the present study), and
study. The method makes it possible to detect
                                                   plethysmography overestimated blood flow
differences in effects without prior knowledge
                                                   relative to that measured by the flow-meter at
of whether they exist at low or high doses. In
                                                   rest and during hyperemia (85). In contrast,
addition, the comparison of dose-response
                                                   plethysmography was shown to underestimate
studes, when a difference is present at several
                                                   leg blood flow at rest compared to the dye-
doses, has a statistical power which exceeds
                                                   dilution method (86).
comparison at a single dose (see page 14).
                                                   In small populations, the mean acetylcholine
Measurement of blood flow and determinants
                                                   response has been shown to be reproducible
of variability related to experimental design
                                                   on the same day, and the sodium nitroprusside
Forearm blood flow was measured by mercury-        response on the same day as well as on
in-silastic strain gauge venous occlusion          different days (see the sections below for each
plethysmography (51, 80, 81). Very few             drug). Such presentation of reproducibility can
studies have employed methods other than           only suggest the presence or the absence of a
plethysmography        for     measurement   of    systematic error, for instance due to
endothelium-dependent forearm vasodilation,        tachyphylaxia. Reports on intra-individual
but methods, probably more accurate, exist         variability of basal or stimulated blood flow are
wich involve invasive manipulation (e.g.           rare in the literature. The intra-individual
measurement with doppler flow wire in the          coefficient of variation for basal flow was 13 ±
brachial artery (82-84) or demand great            2% (87) or 15.5 ± 2.3% (88) on the same day
ressources      (e.g.     magnetic    resonance    but 31% on different days (89). For the
imaging (83)). Venous blood flow as a              vasodilator response to unilateral forearm
substitute for limb blood flow cannot be           exercise it was 17%. The interquartile range of
measured by the thermodilution method as in        the     difference    between     vasoconstrictor
the leg (8), because a vein collecting the         responses to intra-brachial infusion of
majority of venous return, as the fermoral vein    angiotensin II or noradrenaline from one day
in the leg, is not accessible.                     to the other was ≤ 27% (89).
Forearm or leg strain gauge plethysmography        Because blood flow through arterio-venous
have been validated against other methods of       anastomoses in the hand is more variable than
blood flow measurement. Discrepancies              flow to forearm skeletal muscle and
between flow measurements with different           determined, among other factors, by
techniques may be caused by differences in         temperature      regulation,   the     standard
what     they     measure:     plethysmography     procedure is to occlude the hand circulation by
measures total limb blood flow, including that     a wrist cuff inflated to suprasystolic pressure
through skin collaterals branching off before      during measurement of forearm blood
the distal brachial artery (where a doppler flow   flow (51). However, a wrist cuff may not be
                                                            Christian Rask Madsen: PhD thesis, Page 10 of 40.

important when studying         pharmacological      vasodilating factors accounting for the
vasodilator responses (90).                          vasodilating effect of acetylcholine have not
                                                     been identified as only approximately 40%
A resting period seperating the time of arterial
                                                     (dependent upon species and vascular bed) of
cannulation from the first blood flow
                                                     vasodilation     can     be     blocked    by
measurements, as in the present study, is
                                                     pharmacological inhibition of NO and
standard in the literature, although arterial or
                                                     prostacyclin production.
venous cannulation have been shown to not
alter basal blood flow (86).                         Acetylcholine has a half-life in plasma of a few
                                                     seconds because it is rapidly catabolised by
Blood flow has been shown not to be
                                                     plasma cholinesterase. This may an explain
dependent on forearm circumference (90), but
                                                     why vasodilatation to acetylcholine may (93)
a large part of the interindividual variability of
                                                     (or may not (90)) depend on forearm length.
insulin-stimulated vasodilation in healthy
                                                     Plasma cholinesterase activity is variable, in
subjects seems to be explained by relative limb
                                                     part atributtable to genetic variation of its
muscle content (87) and density of muscle
                                                     encoding gene (94). To avoid variability of the
capillaries (91).
                                                     acetylcholine response due to rapid and
Comparison of vasodilator (to acetylcholine or       unpredictable catabolism of acetylcholine,
sodium nitroprusside) or vasoconstrictor (to L-      some authors have used the synthetic
NMMA) responses between groups have                  cholinergic agonist metacholine (see table 5c).
limited value because the pharmacological            However, the use of metacholine has been
response is dependent on factors independent         critisized, because although it activates the
of the pharmacodynamics properties, namely           same receptor as acetylcholine (95), the
vessel wall tension (due to blood pressure and       response has not been found to be inhibited by
preconstriction) and vessel wall geometry            L-NMMA (96).
(wall-lumen ratio) (51). For the same reasons,
                                                     Acetylcholine is considered a physiological
and because absolute and relative increments
                                                     stimulus for the endothelium, as endothelial
of blood flow during drug infusion are
                                                     cells may be stimulated by cholinergic
dependent on basal blood flow (51), it is
                                                     perivascular axons or by acetylcholine
impossible to determine if comparisons
                                                     synthesized and released by endothelium in an
between groups of absolute or blood flow
                                                     autocrine or paracrine fashion (97). However,
relative to base-line are most valid; this is true
                                                     it is unknown whether intra-brachial infusion,
for comparisons in the same group during
                                                     as     in  the   present    study,    produces
dose-response studies on the same day or
                                                     physiologically relevant concentrations of
different days.
                                                     acetylcholine in plasma or at the luminal or
Acetylcholine                                        abluminal surface of the endothelium.

Although acetylcholine is often called a             Other endothelium-dependent agonists have
receptor-dependent agonist of eNOS, the              been used in the human forearm model.
French researcher Paul Vanhoutte has                 Serotonin has the potential advantage that its
suggested that such an agent is merely               vasodilating effect can be completely inhibited
eliciting a “response to a vasodilator that          by L-NMMA (98), but it is a disadvantage
causes endothelium-dependent relaxation in           when comparing with one’s own results that it
vitro” (92). In humans, the primary argument         has only been used in few human studies.
for acetylcholine being an endothelium-              Bradykinin (99) and substance P (99, 100)
dependent vasodilator is that its effect can be      have also been used, but have no obvious
partly inhibited by L-NMMA (see page 11).            advantages over acetylcholine.

Acetylcholine stimulates endothelial NO              The vasodilator response to acetylcholine
production after binding to endothelial              reach a plateau at 3-14 minutes after the
muscarinic G-protein coupled receptors, thus         adminstration of a new dose in a dose-
stimulating phospholipase C which mediates           response study similar to that in the present
Ca2+ influx through the cell membrane and            study (101). Thus, the infusion period of 5
activation of eNOS after increased binding of        minutes was chosen to allow measurement of
Ca2+ to eNOS-associated calmodulin. Increased        blood flow during a steady state of
intracellular Ca2+ also stimulates production of     vasodilation.
arachidonic acid metabolites, e.g. prostacyclin,     The        mean      acetylcholine-stimulated
through activation of phospholipase A2. All          vasodilatation in the leg of patients with
                                                             Christian Rask Madsen: PhD thesis, Page 11 of 40.

ischemic heart disease was reproducible when          streptozotocin-treated      rats,    a    model      of
the dose-reponse study was repeated on the            diabetes (109).
same day (102).
                                                      Increased inhibition by L-NMMA of basal blood
Doses of acetylcholine were chosen to elicit          flow or a pharmacological vasodilator response
expected vasodilator responses from one close         does not prove the existance of increased NO
to the smallest detectable to one as large as         production. Rather, it suggests increased NO
possible whilst still avoiding systemic effects,      available for signaling (because of increased
guided by previous literature.                        NO production or because of decreased NO
                                                      break-down, e.g. due to increased vascular
Sodium nitroprusside                                  superoxide production) or an increase of NO
The mechanism of sodium nitroprusside as a            signaling (e.g. because of decreased activity of
NO donor in vascular tissue is unknown. The           guanylate cyclase, the “NO receptor” in
NO generating activity has been localized to          vascular smooth muscle cells responsible for
membrane fractions of VSCMCs (103), and               production of NO’s second messenger, cGMP).
there is evidence that NO is released on the          The IC50 of L-NMMA for eNOS activity is 3-7
extracellular VSMC surface by reduction of            µmol l−1 (110). The dose of L-NMMA chosen in
sodium nitroprusside through activity of              the present study is similar to high doses in
membrane-bound NADH oxidase (104). On the             previous studies. It produces, calculated by the
other hand, it has been suggested that sodium         formular in the appendix, a plasma
nitroprusside      actually  donates      NO+         concentration of approximately 1.4 mmol l−1.
intracellularly (105).                                The intracellular L-NMMA concentration in
The only frequently used alternative NO donor         endothelium is unpredictable. There are no
in human studies is nitroglycerin. It is              reports on unspecific effects of        L-NMMA
dependent     on     adequate      intracellular      administered in similar doses.
concentrations of cysteine. It does not have
                                                      Insulin-stimulated acetylcholine response
any clear advantages or disadvantages
compared to sodium nitroprusside.                     The infusion of insulin was first planned with
                                                      the intention of studying the very early effects
The mean sodium nitroprusside response in 10
                                                      of insulin on endothelium-dependent and -
healthy controls was different by ≤ 6.8% when
                                                      independent vasodilation, as opposed to the
repeated after 2 hours or after 3 weeks (90).
                                                      effects after 3 days and after 2 months, and
Doses of sodium nitroprusside were chosen to          the insulin-stimulated acetylcholine response
make vasodilator responses match the                  was not a pre-specified end-point in the
expected acetylcholine responses, guided by           original    protocol.     This    pharmacokinetic
previous literature.                                  understanding was superseded by the concept
                                                      of     the     insulin-stimulated    acetylcholine
L-NMMA                                                response as a test for endothelial insulin
At the time of starting the present study, the        sensitivity, in the sense that is used in this
only eNOS inhibitor available for human use           thesis. Therefore, only 13 patients in the
was L-NMMA. As a methylated analogue of L-            treatment group and 4 patients in the time
arginine, L-NMMA is a competitive inhibitor of        control group had this response tested at the
NO      synthase.    Physiologically,    a     low    late repeat examination day.
endogenous        production     of       L-NMMA
                                                      Experiments to control for the                 insulin-
exists (106). L-NMMA inhibits NO synthase
                                                      stimulated acetylcholine response
activity, but not the reduction of O2, leaving
intrinsic     superoxide-generating        activity   The studies of the L-NMMA inhibited insulin-
unaffected (14, 107). On the other hand, NG-          stimulated     acetylcholine   response   was
nitro-L-arginine (L-NNA), which has recently          designed to detect whether any change of the
been used for humans by infusing its inactive         insulin-stimulated acetylcholine response was
precursor NG-nitro-L-arginine methyl ester (L-        endothelium-dependent and mediated by NO.
NAME) (108), inhibits both NO production and
                                                      An alternative to the L-NMMA control would be
O2 reduction. Preliminary evidence suggest
                                                      to measure the insulin-stimulated sodium
that such actions of L-arginine analogue
                                                      nitroprusside response. An increase in this
inhibitors may be relevant physiologically,
                                                      response after clinical intervention would
because L-NNA inhibits the upregulated
                                                      suggest that an increase of the insulin-
superoxide     production    in     aorta    from
                                                      stimulated    acetylcholine  response    was
                                                            Christian Rask Madsen: PhD thesis, Page 12 of 40.

endothelium-dependent only if the former was         coupling of local glucose concentrations and
unchanged or had increased; and vice versa.          local blood flow, because co-infusion of L-
An increase of the insulin-stimulated sodium         glucose (which is not recognised by glucose
nitroprusside response would not preclude that       transporter proteins) did not have the same
an    increase    in   the   insulin-stimulated      effect (116).
acetylcholine response was also endothelium-
                                                     A link between glucose transport and blood
dependent. Furthermore, it would not suggest
                                                     flow is further substantiated by the observation
which factor (e.g. NO, prostacyclin, EDFH) was
                                                     that during experiments where whole-body
mediating the change of the insulin-stimulated
                                                     and leg glucose uptake were constant in
acethylcholine response.
                                                     subjects with a wide variety of insulin
Local insulin stimulation                            sensetivity, namely in lean, healthy subjects,
                                                     obese subjects, and patients with type 2
The method of local insulin stimulation in the       diabetes, leg blood flow increased by the same
forearm has been employed in for metabolic           amount (from ∼ 0.2 to ∼ 0.4 l min−1) (118).
studies for decades since it was first               These conditions were achieved during
described (111). In subjects with a history of       euglycemic hyperinsulinemic clamps with step-
regional sympathectomy, vasodilation to              wise increments of systemic insulin infusion
systemic insulin infusion is delayed in the          (40-1200 mU m−2 min−1) and hyperglycemic
innervated calf and abolished in the                 hyperinsulinemia with step-wise increments of
denervated forearm by intravenous infusion of        systemic glucose infusion (resulting in blood
L-NMMA, suggesting a direct NO-mediated              glucose concentrations of 4-20 mmol
effect of insulin which is modulated by              l−1) (118).   Furthermore,     vasodilation   to
sympathetic activity (112). This finding is          systemic insulin infusion is intact in patients
explained by the well-established ability of         with type 2 diabetes if blood glucose is
insulin to stimulate central nervous system          clamped at each patient’s prevailing fasting
sympathetic activity (113). Insulin-stimulated       blood glucose (119). In contrast, insulin’s
sympathetic effects on peripheral vasomotor          vasodilatory effect does not seem to be
function, probably vasoconstrictor effects, may      dependent on carbohydrate metabolism (120).
be different in insulin sensitive and insulin        Glucose infusion, but not fructose infusion at
resistant individuals because sympathetic            the same molar dose, caused an increase in
activity in obese (114) and elderly (115)            circulating concentrations of endogenously
subjects is increased in the fasting, basal state,   produced insulin. Indirect calometry showed
but reduced during insulin stimulation               that carbohydrate oxidation was similar during
compared to lean controls.                           the two infusions, but only glucose infusion
Intra-arterial infusion of insulin without           increased calf blood flow; and co-infusion of
concommitant infusion of glucose has been            insulin and glucose further elevated circulating
critisized for being unphysiological because it      insulin concentrations as well as blood
results in a reduction of local glucose              flow (120).
concentrations, whereas physiological increase       Studying vasodilation to intra-arterial infusion
in circulating insulin levels in any vascular bed    of insulin in the forearm rather than the leg
is usually accompanied by an increase in blood       has the advantage that a larger rise of
glucose (116). Such experimental conditions          systemic insulin levels can be expected during
during intra-brachial insulin infusion have also     intra-femoral infusion, because the basal blood
been made responsible for the fact that many         flow to the leg is approximately 10 times
studies have not found an effect of                  higher than in the forearm, and because the
intrabrachial insulin infusion to increase basal     dose necessary to raise the concentration of
blood flow (116). Notably, one study found           insulin is directly proportional to forearm or
that similar insulin concentrations in the           blood flow.
forearm caused increased basal forearm blood
flow in healthy individuals during systemic          Serial pharmacological         studies     on     each
hyperinsulinemia, but not during local               experimental day
hyperinsulinemia (117). On the other hand,
                                                     It is generally acknowledged that it is a big
intra-arterial co-infusion of D-glucose to
                                                     advantage to compare pharmacological
maintain local euglycemia augmented the
                                                     vasomotor responses on the same day in a
vasodilatory effect (from 20 to 47%) of
                                                     single subject to avoid the inter-individual
insulin (116). The authors speculated that the
                                                     variability related to day-to-day variation (51).
local rate of glucose uptake may explain this
                                                           Christian Rask Madsen: PhD thesis, Page 13 of 40.

However, this design could be flawed by carry-      prior infusion of any of the other drug
over effects, tachyphylaxia or time effects.        combinations was found more important than
There are no reports in the literature of carry-    controlling for systematic effects of serial
over effects or tachyphylaxia of acetylcholine      infusions. Therefore, the infusion protocol was
or sodium nitroprusside responses (see the          planned with a fixed sequence.
sections on each drug above).
                                                    Design of clinical intervention study
There may have been a positive carry-over
effect of insulin from the first infusion           The intervention design used was an open
(together with acetylcholine) to the second         parallel group intervention. A cross-over design
(together with sodium nitroprusside or with         would have had the benefit of paired
acetylcholine and L-NMMA). There is reason to       observations between insulin treatment and
suspect this because protein expression and         withdrawal of pharmacological treatment. This
activity of eNOS is already apparant after one      would be less suitable considering the potential
hour in cell culture experiments (121).             carry-over effects of insulin treatment, depend
Accordingly, the increase in basal limb blood       to an unpredictable degree on the duration of
flow during local or systemic insulin stimulation   wash-out     period.    Furhtermore,     insulin-
in humans is time-dependent within a few            treatment is expected to result in weight gain
hours (122).                                        (as was indeed the case), which may in itself
                                                    unfavorably influence insulin sensitivity and
A circadian rhytm has been described for basal      insulin dose requirements, and possibly
and stimulated forearm blood flow with a            endothelial function.
lower basal forearm blood flow and a larger
vasodilation to the α-adrenergic antagonist         Patient selection
phentolamine in the morning than in the
                                                    The patient selection was primarily motivated
afternoon or evening (123). Incidentally, there
                                                    by the observations made in the DIGAMI trial
was no diurnal variation of vasodilation to
                                                    as described in the introductory section (see
sodium nitroprusside in this study (123).
                                                    page 1). By including patients with manifest
The pharmacological vasomotor response has          ischemic heart disease we selected patients
been found to be dependent on basal blood           individually prone to the vascular complications
flow in some (51, 93), but not all (90), studies.   of type 2 diabetes. This population can be
If wash-out periods between pharmacological         expected to have more severe insulin
studies do not allow for a complete return of       resistance than patients with type 2 diabetes
blood flow to the initial basal level, a larger     without clinical signs of vascular disease
response can be expected.                            (124).
A final consideration is that the duration of the   Choice of healthy control group
vasoconstrictor effects of L-NMMA is several
hours. Therefore, L-NMMA can only be used as        The most obvious alternatives to the control
the last of a series of infusions.                  group the healthy controls included in the
                                                    present study would be healthy volunteers
One pharmacological study on each of several        matched with patients for BMI or another
experimental days in the same individual            index of obesity; or patients with ischemic
would not be feasible because of the time           heart disease and normal glucose tolerance,
invested by the study subject, the invasive         matched with diabetic patients for BMI. It
nature of the methods, and ressources. One          would have been an attempt to answer if the
pharmacological study on a single day in            vascular function studied was dependent on
different individuals would require large groups    the hyperglycemic phenotype or obesity,
for sufficient statistical power. If untoward       perhaps obesity-related insulin-resistance; or if
effects of serial infusions has to be accepted,     it was dependent on the manifestation of
randomisation of the sequence of infusion may       ischemic heart disease. Other authors have
ensure that these effects are not systematically    suceeded, in human studies of endothelial
favoring a certain result. But sequence             function, to show that cardiovascular risk
randomisation in a small sample size may            factors are additive with respect to their
easily confound results unpredictably. The          association with endothelial dysfunction (125).
main focus was the acetylcholine response and       However, matching patient groups would
the insulin-stimulated acetylcholine response.      probably not be feasible because insulin
Obtaining an estimate of the acetylcholine          sensitivity would likely not be similar in obese
response without the possible influence of a        and type 2 diabetic patients, respectively; and
                                                            Christian Rask Madsen: PhD thesis, Page 14 of 40.

matching for additional factors that would           Statistical analyses
compete for an influence on endothelial
                                                     The analytical strength of the dose-response
function (medication and plasma lipoproteins
                                                     study is that if the response is different in two
(see     page       14),    blood    pressure,
                                                     situations, most often the difference is present
microalbuminuria, and smoking status) would
                                                     at several doses, thus increasing the
be very difficult. Therefore, the more modest
                                                     confidence of the observation made at one
ambition was chosen of simply studying
                                                     dose. To take advantage of this fact, an
whether the diabetic patients, with their
                                                     analysis comparing the difference at each
combined characteristics, had a vascular
                                                     dose, i.e. in seperate two-sample tests, is
function different from that of lean, healthy
                                                     inadequate (130). The analysis of dose-
                                                     response studies cannot be studied with very
Concommitant medication and competing risk           simple methods such as those available in
factors for endothelial dysfunction                  standard spreadsheets. The general problem is
                                                     that all such methods require that observations
The 3-hydroxy-3-methylglutaryl-CoA reductase         are independent – which statistically implies
inhibitors (statins) may improve endothelial         that the observations are not correlated
function independent of their effect on              (covariance equal to zero). We considered
cholesterol metabolism, for example by               using simple data reduction (average or slope
increasing endothelial NO synthase (eNOS)            of observations in one individual). We were
expression (126), and statin therapy of              reluctant to use such methods because they
subjects with desirable cholesterol levels           use data reduction, the consequence being
increases      their     endothelium-dependent       that more experiments are required to study a
vasodilatation (127). In a recent study,             phenomenon than without data reduction.
cholesterol levels, even in the range
considered desirable according to current            We therefore sought professional help, and Dr.
guidelines (< 5.0 mmol l−1), was shown to be         Philip Hougaard accepted to advise us. The
negatively correlated with endothelium-              choices we had were analysis of variance with
dependent vasodilatation (128). According to         repeated measures; and mixed models. The
this study, a difference in plasma cholesterol of    clear recommendation was mixed models. In
1.0 mmol l−1 would be associated with a              addition to the help of Dr. Hougaard I
difference in maximal endothelium-dependent          consulted a textbook on the subject (131).
vasodilatation of 64% (read from figure 4 in         An outline understanding of mixed models
this article (128)). Thus, plasma cholesterol        must take an outset in the more familiar
concentrations and statin therapy status were        analysis of variance. In a standard analysis of
the characteristics that were most likely to         variance a model is formed to explain the
influence endothelial function independent of        overall variance (sum of squares of all
the characteristics we wanted to modify,             differences between the overall mean and
namely hyperglycemia and insulin resistance.         each observation) of a variable (blood flow) as
With       the      purpose      of      avoiding    composed of variance caused by variables of
hypercholesterolemia and establishing a              interest (e.g. treatment) and unspecified
homogeneous patient group with regard to             variability. The effects of a variable is termed a
cholesterol levels and statin medication,            fixed effect. In the general framework of my
patients were only included if they were             study a fixed effect will imply that the dose
treated with a statin when clinically indicated.     response curve of a individual is parallel
Treatment with aspirin (102) and angiotensin         shifted according to each influential effect.
converting enzyme inhibitors (24) may have           Repeated measures (each dose response
improved        the      endothelium-dependent       curve) can be handled with analysis of
responses in the patients, and smoking may           variance with repeated measures, but there
have impaired it (129). Medication in the            are limitations. The only effects that can be
patients and differences in cholesterol levels       studied are fixed effects (see below) and data
and blood pressure between the patients and          must be complete. If one flow is missing, all
healthy controls makes it impossible to              observations for that individual are excluded
conclude which factors are responsible for any       from the analysis. This last problem can be
differences in acetylcholine responses, but          solved by various methods of interpolation, but
does not limit the interpretation of the effect of   the analysis then addresses real as well as
insulin therapy on vascular responses in the         virtual data. For the present study, the
patients.                                            repeated measures would also not be handled
                                                           Christian Rask Madsen: PhD thesis, Page 15 of 40.

completely correctly. Models can be formulated      other effects (and interactions) were specified
which take into account that each dose              as fixed effects.
response curve are repeated measures, but it
                                                    As for model assumptions we received the
is not possible to formulate models which also
                                                    advice to study in detail plots of predicted
take into account that several dose response
                                                    values versus residuals (differences between
curve are from the same individual.
                                                    the model and the original observations). The
Mixed models is a general term for a large          assumptions of the model require that
variety of analysis of continuous and               residuals are normally distributed and not
categorical data. Within this framework, only       dependent on the predicted value. Our initial
random effect models of continuous data were        models of untransformed data did indicate
used. The basic difference between mixed            normal distribution of residuals, but the
models and a standard analysis of variance is       distribution was heavily dependent on the
that the models include covariance terms            value of the predicted value. After logarithmic
(correlation) and can therefore fit correlated      transformation of blood flow data the model
data. In contrast to analysis of variance, the      assumptions appeared fulfilled.
models can handle interaction effects including
                                                    The acutal data handling was performed by
several variables. Thus, in the present study,
                                                    myself in cooperation with Christian Torp-
effects of interest (and interactions), most
                                                    Pedersen (who has extensive experience with
importantly long-term therapy, were entered
                                                    SAS programming). SAS programs were
as fixed effects. Other effects were entered as
                                                    checked by Philip Hougaard.
random effects. Random effects are a
particular strength of mixed models. As             The biologist reading this account may worry
described above, a fixed effect implies a           that the results of the experiments depended
parrallel shift of a dose response curve.           on the use of complex models, or even that
Furthermore, it implies that for the important      the complex models were used to obtain a
interaction between dose and individual, each       positive results. Fortunately the models were
individual is associated with a particular slope    robust. Omission of interactions terms or
of a dose response curve. It is more likely that    omission of transformation gave the same
the slopes of a dose response curve represent       overall results (i.e. significant results for the
random variation around a mean slope. By            same experiments, but with different p-
entering the interaction between dose of            values).
acetylcholine and individual as a random effect
                                                    With this description of mixed models I do not
is thereby assumed that the slopes vary
                                                    pretend to be able to account for the models
normally with a mean of zero. This appears
                                                    used in a mathematical sense. The biologist
reasonable, except for the fact that the slope
                                                    using mathematical methods need to
is obviously not zero. This problem is handled
                                                    understand the use, the limitations and model
by adding interaction between individual and
                                                    checking of complex methods – and to ensure
treatment as a fixed effect, thus assuming that
                                                    by professional advice that the methods are
each treatment is associated with a mean
                                                    correctly used. My understanding of mixed
slope, and that individuals receiving that
                                                    models is limited to this approach.
treatment have slopes that vary from this
average slope. In this account I have used the
term slope to make the subject intuitively          Results
understandable. It is also possible not to
assume any particular relationship between          Insulin stimulation
dose and response. This was actually the            During insulin infusion, local serum insulin was
method used because the large variation of          raised to 133 ± 14.6 mU l−1 and 105 ± 16 in
the results make strict assumption of the           the total patient group at the initial
relationship between dose and response              examination and in the healthy controls,
uncertain.                                          respectively (p = 0.3). Local serum insulin was
Model formulation and model assumptions are         not different during insulin stimulation
critical in any use of models. The final model      between any group of subjects or any
used for the data specified all variables of        examination day (p > 0.2).
interest and all interactions. Individual and the   Local serum insulin was not statistically
interaction between individual and dose             different in healthy controls (105 ± 16 mU l−1,
(without assuming a linear relationship) was        p = 0.3) or during insulin stimulation with or
the only random effect variable included; all
                                                           Christian Rask Madsen: PhD thesis, Page 16 of 40.

without L-NMMA in the different groups of            blood pressure did not change in the treatment
subjects or on different examination days in         or the time control group (table 2, p > 0.5).
the two groups of patients.
                                                     Blood flow data
Systemic insulin was raised by 1.3 ± 0.6 mU l−1
during insulin infusion in healthy controls and      All data were missing at the late repeat
by ≤ 3.7 mU l−1 in patients in either group at       examination for one patient who was
any day. Local and systemic blood glucose            withdrawn from the time control group
decreased by ≤ 0.8 mmol l−1 at any day in any        because of symptomatic hyperglycemia. Data
group (table 1).                                     from dose-response studies with L-NMMA were
                                                     missing in one patient at the initial
All local insulin data were missing, due to          examination and in another patient at the late
failure of placing a venous cannula, in 2            repeat examination for technical reasons.
patients at the initial examination, in 3 patients   Blood flow data from measurements at certain
at the late repeat examination, and in one           doses (between one and three doses in a
healthy control subject.                             single dose-response study) were missing in
Please note that the fasting blood glucose,          five patients.
measured     immediately       after     venous
                                                     Acetylcholine response
cannulation in the morning, was generally
somewhat higher than the levels measured             Basal blood flow was not different in healthy
immediately before insulin stimulation (table 1-     controls and in the total patient group at the
4).                                                  initial examination (2.1 ± 0.2 and 1.9 ± 0.2 ml
                                                     (100 ml)−1 min−1, respectively, p = 0.5) The
Metabolic control                                    acetylcholine response was lower in the total
Fasting serum insulin was more than twice as         patient group compared with healthy controls
high in patients compared with healthy               (figure 3, p = 0.03).
controls (16.2 ± 7.8 versus 6.6 ± 0.5 mU l−1)
                                                     Sodium nitroprusside response
which, considering their elevated fasting blood
glucose, indicates that they were insulin            The sodium nitroprusside response was lower
resistant (132). In the treatment group, fasting     in the total patient group at the initial
systemic serum insulin increased significantly       examination than in healthy controls (figure 4,
from the initiale examination to the late repeat     p = 0.03).
examination (p = 0.01), which was probably a
result of insulin administration. There was no       Insulin-stimulated acetylcholine response
hypoglycemic events that required other              Twenty minutes of insulin infusion had no
management           than       patient-initiated    effect on basal forearm blood flow in healthy
carbohydrate intake.                                 controls (before insulin stimulation 2.5 ± 0.5;
Plasma cholesterol was not different in the          during insulin stimulation 2.9 ± 0.6 ml (100
total patient group (4.6 ± 0.2 mmol l−1)             ml)−1 min−1; n = 10, p = 0.2) or in the total
compared with the total group of healthy             patient group at the initial examination (2.1 ±
controls (4.9 ± 0.1 mmol l−1, p = 0.1), but          0.2 ml (100 ml)−1 min−1 without change, n =
plasma triglycerides were higher (table 2, p <       28).
0.0001).                                             In healthy controls, insulin had a large
In the treatment group, fasting blood glucose        stimulatory effect on the acetylcholine
decreased after 2 months of insulin therapy to       response (149 ± 47, 110 ± 33, 100 ± 45, and
about half the value at the initial examination      106 ± 44% increase of blood flow during the 4
(table 3) and HbA1c decreased from 10.0 ± 0.4        doses of acetylcholine, figure 5, p < 0.0001).
to 7.5 ± 0.2%. In the time control group,            In contrast, such an effect was absent in the
fasting blood glucose was unchanged from the         total patient group at the initial examination
initial to the late repeat examination (table 2),    (figure 6, p = 0.3).
and although HbA1c deteriorated slightly, this
                                                     L-NMMA      inhibition of        insulin-stimulated
was not statistically significant (table 2, p =
                                                     acetylcholine response
0.07). Plasma triglycerides decreased in the
treatment group (table 2, p = 0.0004), but           In healthy controls, L-NMMA decreased blood
were unchanged in the time control group.            flow in response to co-infusion of insulin and
Total plasma cholesterol, HDL cholesterol and        acetylcholine by 46 ± 9, 43 ± 9, 30 ± 15 and
                                                     21 ± 14% during the 4 doses of acetylcholine
                                                            Christian Rask Madsen: PhD thesis, Page 17 of 40.

(figure 5, p < 0.0001). In patients at the initial   no effect of insulin on the acetylcholine
examination, the corresponding relative              response (figure 17, p = 0.7).
vasoconstriction to L-NMMA was 41 ± 6, 25 ±
15, 40 ± 5 and 33 ± 8% (figure 19, p <               Acetylcholine reproducibility
0.0001). The vasoconstriction to L-NMMA was          The acetylcholine response was reproducible
not different between these subgroups of             when repeated after the sodium nitroprusside
patients and healthy controls (p = 0.5).             infusion (2.3 ± 0.2, 3.6 ± 0.5, 6.2 ± 1.1, and
There was no difference between decrements           9.3 ± 1.7 ml (100 ml)−1 min−1 at each dose of
of insulin- and acetylcholine-stimulated blood       acetylcholine during the first dose-response
flow during L-NMMA infusion in healthy               study; 2.8 ± 0.4, 4.9 ± 0.8, and 7.6 ± 1.4 ml,
controls or in patients at the initial examination   and 10.6 ± 1.8 (100 ml)−1 min−1 at each dose
day (figure 7).                                      during the repeated study; n = 13, p = 0.2).

Effect of BH4 on the acetylcholine response          Systemic circulatory responses
After 5 minutes of BH4 infusion, blood flow          Blood flow in the arm contralateral to the
was unchanged compared to basal (3.6 ± 1.2           perfused arm were unchanged throughout the
versus 3.8 ± 1.0, p = 0.9). The acetylcholine        studies. As an example, blood flow in the
response during BH4 infusion was larger than         control arm is displayed in figure 20. In the
during acetylcholine alone (figure 8, p <            subgroup of healthy controls who were studied
0.0001).                                             during insulin stimulation and in the total
                                                     patient group at any examination day, systolic
Effect of short-term insulin therapy on              blood pressure was unchanged during
acetylcholine and sodium nitroprusside               acetylcholine infusion, but decreased by 7-12
responses                                            (range at highest dose) mmHg during sodium
                                                     nitroprusside infusion and increased by 4-10
At the early repeat examination in patients in
                                                     mmHg during L-NMMA infusion (p ≤ 0.01).
the treatment group, both the acetylcholine
(figure 9) and the sodium nitroprusside (figure
10) response had decreased (p = 0.007 and            Discussion
0.009, respectively).
                                                     Interpretation of own results
Effect of long-term insulin therapy on vascular
                                                     Acetylcholine responses before treatment
                                                     Like the present study, all previous studies,
After 2 months of insulin therapy, the
                                                     except     one (133),      have    demonstrated
acetylcholine  (figure   11)   and sodium
                                                     decreased                endothelium-dependent
nitroprusside responses (figure 10) were
                                                     vasodilatation in patients with type 2 diabetes
unchanged (p = 0.09 for both).
                                                     with similar methods (6, 24, 76, 134-142).
In contrast, insulin-stimulation now had a           Indeed, several studies have also found
significant effect in patients in the treatment      impaired flow-mediated vasodilation (73, 143,
group (58 ± 25, 84 ± 66, 120 ± 93, and 69 ±          144). Most of these studies included patients
36% increase of blood flow during the 4 doses        without clinical signs of cardiovascular disease.
of acetylcholine, figure 12 and 13, p = 0.0002).
                                                     In the majority of previous studies of forearm
This result was similar when analyzed by
                                                     responses to nitrovasodilators, their effect
repeated measurements ANOVA (p = 0.02).
                                                     have been shown to be similar in patients with
The decrements of insulin- and acetylcholine-
                                                     type 2 diabetes and healthy controls, although
stimulated blood flow during L-NMMA infusion
                                                     some have found an impaired response (76,
were unchanged by insulin therapy (figure 14,
                                                     136). Given that such an impairment of
p = 0.9).
                                                     endothelium-independent      vasodilation    is
Acetylcholine and sodium nitroprusside               representative of the general pathophysiology
responses and insulin-stimulated responses in        of type 2 diabetes - or other risk factor for
the time control group                               atheroscleroses - it is probably smaller than
                                                     the     decreased       endothelium-dependent
In the time control group, acetylcholine (figure     vasodilation, and some studies may not have
15) and sodium nitroprusside responses (figure       had the statistical power to detect it. Thus,
16) did not change after 2 months (p = 0.09          decreased brachial artery vasodilation to
and 0.6, respectively). At this time, there was      sublingual nitroglycerin was associated with
                                                          Christian Rask Madsen: PhD thesis, Page 18 of 40.

risk factors for cardiovascular disease in a        and more efficient therapy. For example, in the
study with a large sample size (n = 800) (145)      DIGAMI trial HbA1c decreased 1.1 ± 0.1% in
and preliminary data from a study of 90             the intervention group and 0.4 ± 0.1% in the
patients with type 1 diabetes showed                control group after 3 months (46).
decreased forearm vasodilation to nitroglycerin
                                                    After a new steady state of glycemic control,
infusion (146). A decreased response to
                                                    HbA1c changes to a new level after at least one
exogenous NO may be explained by some of
                                                    month, so HbA1c at the initial study day
the mechanisms that have also been
                                                    probably overestimated glycemic control after
implicated in the endogenous NO production,
                                                    2 weeks without hypoglycemic medication, and
e.g. breakdown of NO by reactive oxygen
                                                    the small (and statistically insignificant)
species before reaching targets for NO
                                                    increase in HbA1c most likely did not reflect a
signaling, or dysregulation of the signal
                                                    real change in glycemic control.
pathway distal to NO, e.g. decreased
guanylate cyclase activity (see page 24).           The decreasing plasma concentrations of
                                                    triglycerides was probably associated with a
Local insulin stimulation                           decrease of FFA, which, unfortunatley, were
By design, local serum insulin during insulin       not measured. FFA may directly influence
stimulation was not different in patients and in    insulin sensitivity (149) and both endothelial
healthy controls. A stimulating effect of insulin   function and insulin-stimulated endothelial
on endothelium-dependent vasodilatation may         function (see page 24).
exist at a higher concentration of insulin. This
                                                    Insulin-stimulated acetylcholine response
consideration was taken into account in a
study       where       endothelium-dependent       In the healthy controls, insulin had no effect
vasodilation was measured after more than 3         when infused alone, but was a relative potent
hours of systemic insulin infusion and the          stimulus for the acetylcholine response. This
lacking effect of insulin was present despite       confirms previous findings in lean, healthy
insulin levels 38 times higher than levels          subjects (150, 151). An explanation of this
sufficient for an effect in lean, healthy           synergistic effect of insulin and acetylcholine
controls (6).                                       has recently been suggested at the molecular
                                                    level (see page 23).
In a study from our institution (2) of patients
with type 2 diabetes and ischemic heart             The present study is the first to examine
disease, mean systemic insulin concentrations       insulin-stimulated vasodilator responses in
2 hours after a glucose load during a standard      patients with type 2 diabetes and manifest
oral glucose tolerance test were 106 mU l−1         atherosclerotic disease.
(measured by ELISA, which may be expected
                                                    The decreased insulin-stimulated acetylcholine
to yield lower results than the RIA used in the
                                                    response in patients compared to controls may
present study). The local increase of insulin
                                                    be explained by changes in the insulin
concentrations during insulin stimulation in the
                                                    signalling pathway in endothelial cells (see
present study is thus comparable to a
                                                    page 24).
physiological increase of insulin concentrations
in a similar patients group.                        The comparisons of seperate and combined
                                                    effects of insulin and acetycholine has the
The increases in systemic serum insulin were
                                                    limitation that the experimental protocol only
small, but they are not trivial. Circulating
                                                    allows direct comparison of the acetylcholine
insulin concentrations 2-3 times the fasting
                                                    responses with or without insulin infusion. The
level stimulates central nervous system
                                                    question of whether insulin had any effect on
sympathetic activity occur with similar potency
                                                    its own would be answered correctly if the
to high physiological levels (114, 147, 148). It
                                                    protocol had included an infusion of insulin
is unknown if a relative increase of systemic
                                                    alone during a period of the same length as
serum insulin as small as 28 ± 12%, as in our
                                                    sum of the insulin (20 minutes) and
healthy controls, stimulates sympathetic
                                                    acetylcholine (∼ 28 minutes) infusions.
                                                    Previous studies of insulin’s effect on the
Metabolic effects of insulin treatment              response to an endothelium-dependent
The 2.5 ± 0.4% decrease of HbA1c was larger         vasodilator in patients with type 2 diabetes
than that seen in large clinical trials, which      have all used systemic hyperinsulinemia (see
may both be explained by a high initial value       table 5c). The present study is the first to
                                                         Christian Rask Madsen: PhD thesis, Page 19 of 40.

examine the effect of locally elevated insulin    have occured in the present study, when
concentrations on this response in such           chronic     hyperglycemia was     quickly
patients. This is important because systemic      ameliorated.
mechanisms (117)            or          glucose
                                                  Insulin-stimulated vasodilation may be coupled
administration (116) has been suggested to be
                                                  to glucose uptake (see page 12). If this is also
necessary elements in insulin-stimulated
                                                  true for NO-mediated vasodilation (to
vasodilation. Furthermore, there has previously
                                                  endogenously produced or exogenously
been concerns that impaired insulin-stimulated
                                                  provided NO), and if basal glucose uptake after
vasodilation in subjects with obesity or type 2
                                                  3 days of insulin therapy had decreased due to
diabetes are due to test conditions of
                                                  the sudden lowering of blood glucose, such a
euglycemia (119). This does not seem to be
                                                  mechanism could also explain the result from
relevant in the patient group of the present
                                                  the early examination day. Again, this
study, where the potentiation by insulin of the
                                                  explanation is not supported by direct
vasodilator response to an endothelium-
                                                  experimental evidence.
dependent agonist exist during during
spontaneous hyperglycemia.                        Effect of long-term insulin             therapy on
                                                  acetylcholine and sodium                nitroprusside
Effect of BH4 on the acetylcholine response
The effect of BH4 on the acetylcholine
                                                  A recent study showed that adding
response was large. However, it should be
                                                  intermediate-acting bed-time insulin to existing
interpreted with hesitation claimed by the fact
                                                  treatment with metformin in patients with type
that control experiments were not made to
                                                  2 diabetes, but no clinical signs of
demonstate      that    the    response    was
                                                  cardiovascular disease, increased the forearm
endothelium-dependent        and     by     the
                                                  vasodilator response to acetylcholine, whereas
circumstance that the study design was not
                                                  the sodium nitroprusside response was
optimal for comparing the acetylcholine
                                                  unchanged (153). Apart from patient selection,
response and the combined BH4 and
                                                  the discrepancy between results from that
acetylcholine response, divided as they were
                                                  study and the present study may be that
by 3 dose-response studies with infusion of
                                                  although the effect on glycemic control was
acetylcholine,     insulin,     and     sodium
                                                  less (HbA1c decreased from 9.0 ± 0.3% to 7.6
                                                  ± 0.1%), the treatment duration was longer (6
Effect of short-term insulin therapy on           months).
acetylcholine and sodium nitroprusside            The fact that the sodium nitroprusside
responses                                         response was unchanged after 2 months in the
There is no previous literature that offers an    treatment group shows that insulin therapy did
explanation for the decreased acetylcholine       not result in any apparant change of vascular
and sodium nitroprusside responses after 3        sensitivity to NO.
days of insulin therapy. However, there is
                                                  Effect of long-term insulin therapy on insulin-
evidedence that hyperglycemia at short term
                                                  stimulated acetylcholine response
may increase endothelium-dependent and -
independet vasodilation. The acetylcholine-       The present study is the first to show that the
induced vasodilation of arterioles of the         local effect of insulin on peripheral vasodilation
supraspinatus muscle in rats, studied by          to an endothelium-dependent agonist is
intravital microscopy, was unchanged during       impaired in patients with type 2 diabetes
one hour of superfusion with 200 mg (100          compared to lean, healthy controls. It is also
ml)−1 glucose but increased by superfusion of     the first to show that this effect of insulin can
300 mg (100 ml)−1 glucose (152). This increase    be improved by glycemic control.
could not be bloced by L-NAME or
                                                  In previous studies, long-term treatment with
meclofenamate (a cyclooxygenase inhibitor)
                                                  the insulin sentisizer troglitazone improved
and was accompanied by a decreased
                                                  brachial artery flow-mediated vasodilation in
production of NO, measured directly with an
                                                  patients with impaired glucose tolerance (154),
NO-sensitive microelectrode, but an increased
                                                  but did not change forearm acetylcholine
vasodilator response to sodium nitroprusside.
                                                  responses or vasodilation to systemic
Thus, hyperglycemia apparantly increase
                                                  hyperinsulinemia in patients with obesity-
vascular NO sensitivity acutely in vivo.
                                                  associated insulin resistance (155). In these
Speculatively, the opposite mechanism might
                                                          Christian Rask Madsen: PhD thesis, Page 20 of 40.

studies, glycemic control was unchanged by         High glucose concentrations and endothelial
troglitazone therapy.                              function
L-NMMA    inhibition of the insulin-stimulated     Hyperglycemia is the most conspicuous trait of
acetylcholine response                             the diabetic phenotype, and high glucose
                                                   concentrations seem to directly affect
The acetylcholine vasodilator response is only     endothelial function. In a study in vitro,
partly mediated by NO. One study,                  endothelium-dependent      vasorelaxation    of
representative of several, showed inhibition of    rabbit aorta decreased after incubation for 6
39% of vasodilation to acetylcholine by L-         hours    in    media    with    high    glucose
NMMA (at an acetylcholine dose of 30 µg            concentrations (157), and in another study ex
min−1) (156). In healthy controls, we observed     vivo, 72 hours of intravenous glucose infusion
a similar relative L-NMMA inhibition of the far    in rats decreased acetylcholine-stimulated
greater combined insulin and acetylcholine         vasorelaxation (158).         Glucose-induced
response. This is consistent with a substantial    endothelial dysfunction seems mediated by
part of the insulin potentiation of the            reactive oxygen species, which can degrade
acetylcholine response being mediated by NO.       NO, because it was prevented by treatment in
There is only one published human study in         vivo with probucol, a scavenger of reactive
which the effect of L-NMMA on insulin-             oxygen species, and by antioxidant enzymes
stimulated vasodilation to an endothelium-         ex vivo (157). Such antioxidant protection may
dependent agonist was examined (151). The L-       occur physiologically, as high glucose
NMMA experiments in this study suffer from         concentrations induced endothelial antioxidant
the same limitations of interpretation, namely     enzymes in cell culture (159). It was also
that definite proof of involvement of NO in the    achieved by in vitro transfection of the Mn2+
insulin-stimulated acetylcholine response is       SOD gene (160). In healthy humans, local
only acheived if L-NMMA causes a decrement         forearm hyperglycemia for 6 hours decreased
of     the    insulin-stimulated   acetylcholine   endothelium-dependent vasodilatation (161).
response larger than the L-NMMA-induced            NO production decreases when endothelial
decrement of the acetylcholine response alone.     cells are cultured in high levels of glucose for
In fact, these decrements were similar (based      5-7 days (162, 163). Conversely, the in vitro
on incomplete numerical data and read from         transfection of eNOS gene to aorta (164) and
figure 4) (151).                                   carotid artery (165) of alloxan-treated rabbits
The L-NMMA response was not different in           prevent     impaired    endothelium-dependent
patients and healthy controls, suggesting that     vasorelaxation. The mechanisms responsible
the NO-mediated part of insulin-stimulated         for decreased NO production are unclear, as
acetylcholine response in healthy controls         there are reports of downregulation of eNOS
(where insulin potentiated the acetylcholine       protein (163) as well as eNOS mRNA and
response) is similar to that in patients (where    protein upregulation accompanied by increased
insulin had no effect).                            superoxide production which can degrade
                                                   NO (162). A source of superoxide may be
Taken together, the L-NMMA experiments are         eNOS itself. Thus, preliminary results show
compatible with a role of NO in the insulin        that elevated superoxide production in aorta
potentiation of the acetylcholine response in      from streptozotocin-treated rats can be
healthy controls and do not exclude a role for     inhibited by L-NNA, a NO synthase inhibitor
NO in improving the insulin-stimulated             (see page 11) (109). However, the quantitively
acetylcholine response after insulin therapy.      most important source of superoxide is
However, these experiments do not prove a          thought to be an NAD(P)H oxidase which
role for NO in the insulin potentiation of the     resembles the phagocytic NADPH oxidase, but
acetylcholine response, and the most straight-     favors NADH as a substrate (166). Activity of
forward interpretation is that NO does not         such an enzyme has been demonstrated in
mediate the improvement of the insulin-            endothelial cells and VSMCs cultured in high
stimulated acetylcholine response after insulin    glucose concentrations (167) and human
therapy.                                           saphenous veins ex vivo (168), and may be
Discussion of related previous literature          caused by activation of PKC (167).

The discussion is limited to studies relevant to   Oxidative stress, i.e. an imbalance between
the    regulation    of    endothelium-derived     reactive   oxygen     species  and    cellular
vasomotor substances.                              antioxidant defense systems (169) may also be
                                                             Christian Rask Madsen: PhD thesis, Page 21 of 40.

generated through alterations of glucose              proteins, resulting in formation of advanced
metabolism. Increased cellular glucose uptake         glycation end-products (AGEs). They may
increases the flux of glucose through the             directly   catabolize    NO (182),   decrease
sorbitol pathway (also known as the pyolol            prostacyclin production (183), and reduce the
pathway), which consumes NADPH by the                 expression of eNOS (184).
aldose reductase reaction and reduces NAD+
                                                      Another mechanism for hyperglycemia-induced
by the sorbitol reductase reaction (170).
                                                      endothelial dysfunction may be through
Among the consequences of an overactive
                                                      upregulation of assymmetrical dimethyl
sorbitol pathway is depletion of cytosolic
                                                      arginine (ADMA). ADMA is an endogenously
NADPH, a cofactor for eNOS and necessary to
                                                      produced L-arginine analogue and an inhibitor
maintain the primary intracellular antioxidant,
                                                      of NO synthase (like L-NMMA, another
glutathion; and an increase of the cytosolic
                                                      methylated       L-arginine    analogue) (106).
NADH/ NAD+ ratio, which is likely to promote
                                                      Preliminary     result     showed    that     in
the production of reactive oxygen species, in
                                                      streptozotocin-treated rats, hyperglycemia was
part through stimulating superoxide production
                                                      correlated with plasma ADMA levels, and
by vascular NAD(P)H oxidase (171). These
                                                      activity of DDAH, the enzyme that catabolizes
metabolic changes may be the reason that
                                                      ADMA, was decreased in aorta although DDAH
aldose reductase antagonism was able to
                                                      expression was unchanged (185). Similarly,
prevent     impaired      endothelium-dependent
                                                      DDAH activity, but not expression, was
vasorelaxation of aorta from alloxan-treated
                                                      decreased in cultured VSMC after incubation
diabetic    rabbits (172).    By     a    different
                                                      for 2 weeks at high concentrations of glucose
mechanism, increased glucose concentrations
                                                      compared               to          physiological
may       inhibit       glucose       6-phosphate
                                                      concentrations (185).
dehydrogenase, which catalyzes the first
intermediary      reaction   in    the    pentose     Finally, eNOS function and endothelium-
phosphate pahtway, the primary source of              dependent vasodilation associated with insulin
NADPH (173,          174).      High       glucose    resistance or hyperglycemia may be impaired
concentrations may also lead to oxidative             because of a relative vascular deficiency of a
stress through glucose autoxidation or lipid          cofactor for eNOS, BH4 (see page 29).
peroxidation (169).
                                                      Effects of insulin on vasodilation in vivo in
Hyperglycemia is associated with activation of        animals and healthy humans
certain isoforms of PKC, and with increased
intracellular concentrations of diacylglycerol        A vasodilating effect of insulin has been
(DAG), a physiological activator of PKC, in           demonstrated in several vascular beds in
vascular tissues (175). Impaired endothelium-         humans in vivo. Studies with inhibition of NO
dependent relaxation during high glucose              production with L-NMMA and studies in isolated
concentrations was prevented by a PKC                 vessels in vitro suggest that the effect is direct
inhibitor (176). Furthermore, an inhibitor of         rather than dependent on circulating factors or
the PKC β isoform prevented impaired                  the autonomic nervous system. Insulin-
endothelium-dependent           vasodilation     in   stimulated vasodilation in the arm or the leg is
mesenteric or intestinal arterioles in vivo, both     for the major part dependent on vasodilation
after oral administration to streptozotocin-          of resistance vessels in skeletal muscle and
treated     diabetic    rats (177)     and    after   independent of vasodilation in skin (87, 186).
administration     during      45    minutes     of   Systemic hyperinsulinemia for ≥ 1 hour or local
superfusion         with        high       glucose    forearm or leg hyperinsulinemia for ≥ 20
concentrations (178). Specifically, the latter        minutes with insulin levels in the high
study showed that decreased NO production,            physiological or low supraphysiological range
measured directly with a NO-sensitive                 increased blood flow in many studies by ∼ 50%
electrode,      was      prevented      by     PKC    (see table 5a).
inhibition (178). The same PKC β inhibitor
increased           insulin-induced          eNOS     Systemic euglycemic hyperinsulinemia also
expression (179). Interestingly, vitamin E may        caused vasodilation of the human internal
inhibit DAG-PKC activation (180) independent          carotid artery (measured by 2-dimensional
of its antioxidant effects (181).                     ultrasound) after 15-60 minutes; arterial
                                                      diameter had regained baseline diameter 15
Moreover, chronic hyperglycemia may impair            minutes after the termination of intravenous
endothelial function by irreversibly modifying        insulin-glucose-potassium infusion (187).
                                                           Christian Rask Madsen: PhD thesis, Page 22 of 40.

All studies of local limb hyperinsulinemia,         Effects of insulin on vasodilation in insulin
except two (150, 188), did not demonstrate          resistant subjects
any effect of insulin in healthy, lean subjects,
                                                    A decreased or even abolished effect of insulin
regardless of dose and duration of intra-
                                                    in subjects with obesity or type 2 diabetes
arterial insulin infusion (see table 5b).
                                                    have been demonstrated in several studies
However, intracoronary insulin infusion in non-
                                                    using systemic hyperinsulinemia (8, 114, 155).
diabetic patients increased coronary blood flow
                                                    In one study of subjects with obesity-
by 22% after 60 minutes (189). Furthermore,
                                                    associated      insulin    resistance,    only
local insulin stimulation caused vasodilation of
                                                    pharmacological doses of insulin had a
hand veins with (190) or without (191)
                                                    vasodilator effect (8).
precontraction       during      infusion     of
norepinephrine.                                     Vasodilation to local limb hyperinsulinemia has
                                                    not been studied in obese or diabetic subjects.
The present study examined the effect of
                                                    However, the vasodilator response in hand
insulin on the acetylcholine vasodilator
                                                    veins of subjects with low-grade obesity and
response.     Previous     studies    from   two
                                                    borderline hypertension was blunted compared
laboratories showed that insulin potentiated
                                                    to lean, healthy controls (191).
vasodilation      to     endothelium-dependent
agonists: local forearm hyperinsulinemia            Mediators of insulin-stimulated vasodilation
increased     forearm     blood    flow    during
intrabrachial acetylcholine infusion by 41% at      It is beleived that NO is an important mediator
the highest dose (151); and systemic                of     insulin-stimulated    vasodilation.    The
hyperinsulinemia increased leg blood flow to        vasodilating effect of systemic (150, 196) and
intra-femoral metacholine at the highest dose       local (151) hyperinsulinemia could be inhibited
by 20% (150).                                       by L-NNMA. The results in the former
                                                    studies (150, 196) actually suggest that the
In keeping with human studies, in vitro studies     majority of insulin’s effect is mediated by NO.
have shown a vasodilating effect of insulin on      Furthermore, sodium nitroprusside vasodilator
isolated skeletal muscle arterioles (192,           responses       were    not     potentiated    by
193)and conduit arteries (194) from animals.        systemic (150)       or    local (151)     insulin
Interestingly, in rat gastrocnemius arterioles      stimulation (see page 21).
the vasodilating effect was present during
insulin concentrations of 10 mU l−1, i.e. the       Apart from stimulation of endothelium-
order of magnitude of basal insulin                 dependent     vasodilators   (e.g.    NO      or
concentrations (192).                               prostacyclin)   or    vasoconstrictors     (e.g.
                                                    thromboxane A2 or ET-1), insulin may exert
In humans studies, a dependency on both time        effects on vascular tone by modulating
and dose has been described (122), although         sympathetic nervous system activity (see page
it is not clear in this presentation (122)          12) or modulating the effects of circulating or
whether data of time and dose are based on          endothelium-independent       vasoconstrictors.
sequential and accumulative doses of insulin. A     This hampers the interpretation of in vivo
direct comparison of the time-dependency of         responses to insulin.
insulin-stimulated vasodilation and insulin-
potentiated vasodilation to an endothelium-         In canine conduit coronary arteries, insulin-
dependent agonist cannot be made because            stimulated relaxation was completely inhibited
infusion of acetylcholine or metacholine only       by combined pharmacological inhibition of
lasted < 30 minutes. However, it is interesting     eNOS (with L-NNA) and cyclooxygenase (with
that insulin-stimulated vasodilation is minimal     indomethacin) or removal of the endothelium;
within the first 30 minutes (195) and shows a       but there was no effect of pharmcalological
gradual increase within the first 3 hours (8),      inhibition in coronary microvessels (0.1-100
whereas insulin's effect on vasodilation to         ng/ml) (197).      Unfortunately,   results   of
acetylcholine or metacholine (during local          inhibition with either L-NNA or indomethacin
stimulation) is potent within 20 minutes (in the    were not reported (197). In isolated arterioles
present study and a previous study (151)).          from         rat       cremaster (192)        or
Therefore, it is possible that the preferential     gastrocnemius (193) muscle, a vasorelaxing
mechanisms in these two situations are in fact      effect of insulin (10-104 mU l−1) was completely
very different, e.g. by induction of eNOS and       absent after exposure to L-NNA or removal of
stimulation of immediate eNOS activation,           endothelium.
respectively (see page 23).
                                                           Christian Rask Madsen: PhD thesis, Page 23 of 40.

The vasodilator effects of insulin has been         Certain studies suggest that insulin-stimulated
suspected to modify vascular responses to           vasodilation is dependent of glucose uptake or
autonomic nervous system stimulation or             metabolism (see pages 12 and 24). Perhaps
hormones. Thus, insulin inhibited the               the most interesting mechanism of insulin-
vasoconstrictor response of norepinephrine          stimulated vasodilation, discussed in the next
and angiotension II in the rabbit femoral           section, is the recently discovered insulin signal
artery (198) and in the human forearm (188).        transduction pathway in endothelium, which is
Adrenergic innervation of VSMC is primarily         similar to that in skeletal muscle and other
relayed through α1 (contraction) and β2             tissues where the insulin signal is primarily
(relaxation) receptors. However, insulin’s          affecting glucose metabolism.
vasodilating effect does not seem to be
mediated by adrenergic or cholinergic               Insulin-stimulated PI3K-Akt-eNOS activation
mechanisms in humans, as the increase of calf       A      newly     described      Ca2+-independent
blood flow during systemic hyperinsulinemia         mechanism        of     eNOS      activation   is
could not be blocked by propranolol (a non-         phosporylation of eNOS on serine 1177 by Akt,
selective β-adrenoceptor blocker) (at a high        which is itself activated by phosporylation
dose, 1 mg kg−1 min−1), prazosin (an α1-            following PI3K activation (207). The PI3K-Akt
selective      adrenoceptor      blocker),   or     pathway of eNOS activation is discussed in the
atropin (199). This study, however, is a            manuscript article (208). Until know, stimuli for
contradiction of an earlier report (200), where     this pathway has been shown to include VEGF,
propranolol blocked the vasodilating effect of      insulin, fluid shear stress, and estrogen (207).
intra-brachial insulin infusion, even at a dose     Akt-dependent activation of endothelial NO
of insulin (0.1 mU kg−1 min−1) which did not        production exists in vivo (209). Interestingly,
affect systemic blood glucose; phentolamine (a      the acetylcholine response, a Ca2+-dependent
non-selective α-adrenoceptor blocker) did not       stimulus, was shown to be dependent on Akt:
affect insulin’s effect (200).                      in vivo transfection of dominant-negative Akt
In a study of healthy, lean volunteers, insulin     abolished acetylcholine-induced vasodilation of
was infused intrabrachially at a dose of 0.1 mU     rabbit femoral artery in vivo, and in vitro
kg−1 min−1 for 2-4 hours with intravenous           transfection of dominant-negative Akt inhibited
infusion of glucose to maintain local glucose       acetylcholine-induced vasorelaxation of mouse
levels. Neither infusion of insulin alone nor       aorta (209).     The     transduction    of  the
intrabrachial infusion of ET-1 antagonists alone    mechanical stimulus of shear stress to the
changed forearm blood flow. However, ET-1           PI3K-Akt-eNOS pathway is not known, but
antagonists increased blood flow during insulin     preliminary results in cultured endothelial cells
infusion, and in the presence of ET-1               suggest that PI3K activation induced by fluid
antagonism, vasoconstriction to L-NMMA was          shear stress may involve phosporylation of
larger with insulin infusion than without (201).    platelet-endothelial cell adhesion molecule-1, a
Thus, the study did not identify a vasodilating     constituent of endothelial cell junctions (210),
effect of insulin, but suggested an insulin-        or IRS-1 (211).
induced change in the balance of endothelium-       A suggestion of how a Ca2+-dependent
derived vasodilating and vasoconstricting           stimulus (e.g. acetylcholine (212)) and a Ca2+-
factors.                                            independent stimulus (Akt phosporylation) may
Insulin stimulates NO production from cultured      interact was recently given by a study of cell-
rat aortic endothelial cells (202) and human        free function of bovine eNOS with a serine-
umbilical vein endothelial cells (10, 203) within   aspartate mutation at the serine residue that is
a few minutes (10) and with an ED50 of 500-         phosphorylated by Akt (213). This mutation,
700 nM (∼ 72 U l−1, an order of magnitude 103       through the negative charge of aspartate, was
larger than physiological concentrations).          designed to mimick phosphorylation. The
Insulin also stimulates the production of ET-1      mutated enzyme exhibited an increase of
from cultured endothelial cells (204, 205) and      catalytic activity at any concentration of Ca2+,
VSMC (206). There are no studies of the effect      perhaps by increasing the electron flux
of insulin-stimulation on the release of other      through the reductase domain of the enzyme
endothelium-derived      vasodilating     or    -   and by inhibiting the Ca2+-binding cofactor
constricting factors from isolated endothelial      calmodulin from eNOS (213).
                                                           Christian Rask Madsen: PhD thesis, Page 24 of 40.

Effects of insulin on eNOS expression               concentration-dependently       inhibited    PI3K
                                                    activity, Akt phosphorylation and eNOS
Insulin may not only affect eNOS expression
                                                    activity (221). The molecular mechanism by
through a glucose-lowering effect. Insulin
                                                    which high glucose inhibits insulin signaling in
increased the expression of eNOS mRNA in rat
                                                    endothelial cells is not known, but in a cultured
epididymal microvessels (121) and eNOS
                                                    fibroblast cell line, high glucose activates PKC
protein in endothelial cells cultured from
                                                    and       inhibits     IR-dependent       tyrosine
bovine      aorta (121),    human      umbilical
                                                    phosphorylation as well as dephosphorylation
vein (214), human aorta (214), and human
                                                    of focal adhesion kinase, whereas TNF-α
coronary arteries (163) with a maximum after
                                                    inhibits insulin signaling through activation of
12 hours of insulin stimulation (121). This
                                                    phosphotyrosine phosphatase and inhibition of
effect may be mediated by PI3K (121), which
                                                    tyrosine phosphorylation, but not through
may thus be a mediator of both immidiate (10,
                                                    inhibition of dephosporylation of focal adhesion
11, 203) (see the previous section) and late
                                                    kinase (222).
effects of insulin on NO production. Incubation
with a high glucose concentration inhibited         Some studies have found elevated circulating
insulin induction of eNOS protein after 24          concentrations of ET-1 in patients with type 2
hours (163).                                        diabetes, and ET receptor expression and ET-1
                                                    effects are altered during insulin resistance
Mechanisms of endothelial insulin resistance        and hyperglycemia (223). ET-1 has compound
Insulin resistance of glucose uptake in obesity     effects on glucose transport. Thus, in
may be mediated by elevated FFA (149).              adipocytes      ET-1       stimulated      GLUT4
Endothelial NO producition may also be              translocation and glucose uptake through
influenced by FFA, which inhibit eNOS activity      activation of the ETA receptor (224, 225), but
in vitro (215). Moreover, in lean healthy           inhibited       insulin-stimulated        glucose
subjects, leg metacholine vasodilator response      uptake (226). In rats, ET-1 impaired indirect
(but not sodium nitroprusside response) was         measures of insulin sensitivity (227), and in
impaired during intravenous infusion of FFA or      healthy humans, ET-1 inhibited leg glucose
during elevated FFA due to inhibition of            uptake during hyperinsulinemic euglycemic
endogenous insulin production by somatostatin       clamp without altering leg blood flow (and
infusion (216). Furthermore, leg vasodilation       inhibited whole-body glucose uptake, although
induced by systemic hyperinsulinemia was            not examined during steady-state) (228). In
inhibited by 4 hours of intravenous infusion of     cultured VSMCs ET-1 inhibits insulin-stimulated
FFA (217).      Leg     blood     flow     during   association of PI3K with IRS-2 as well as PI3K
hyperinsulinemia was similar with and without       activity,     apparently       through        PKC
FFA infusion, so the decreased insulin              activation (229). Although speculative, it is
vasodilator response was only present as a          thus possible that ET-1 can inhibit insulin-
smaller increase relative to baseline blood flow    stimulated    PI3K-Akt-eNOS       activation    in
(which was increased during FFA infusion            endothelium.
alone, before the hyperinsulinemic clamp was
                                                    Effects of insulin on vascular smooth muscle
started) (217). Apart from being small, this
                                                    cell function in vitro
effect was delayed, as infusion for 2 hours did
not affect insulin-stimulated vasodilation (218).   Insulin may have vasodilatory effects by a
The      impairment      of    insulin-stimulated   direct effect on VSMCs. Two groups in
vasodilation was apparantly due to a decrease       particular have pursued studies of the effects
of      NO-mediated         vasodilation,      as   of insulin on VSMCs, either primary cultures of
vasoconstriction to L-NMMA was lower with           vascular smooth muscle cells (on rat tail
FFA infusion than without (217).                    collagen gels, a situation where they do not
                                                    proliferate) from canine femoral artery grown
Both TNF-α (see page 29) and high glucose
                                                    for 1-2 weeks (230-235), or cultures, after
concentrations may inhibit the PI3K-Akt-eNOS
                                                    several passages, of vascular smooth muscle
signaling pathway. TNF-α inhibits Akt
                                                    cells from human abdominal tissue (236-238).
phosphorylation      in      endothelial     cell
culture (219) and preliminary results show that     One study showed that incubation with insulin
TNF-α       inhibit     IRS-1     by      serine    (10-40 mU l−1 for 20 minutes) inhibited VSMC
phosphorylation (220).     Other     preliminary    contraction as well as the increase in
results showed that in endothelial cell culture,    intracellular  Ca2+    concentration     after
incubation with glucose for 12 hours                stimulation with serotonin or angiotensin
                                                           Christian Rask Madsen: PhD thesis, Page 25 of 40.

II (230). These effects were similar after          VSMCs has been shown to constitutively
preincubation with insulin 40 mU l−1 for 1          express NO synthase, but there is no
week (230). These effects were apparantly           agreement of which isoforms are present,
mediated by voltage-operated Ca2+-channels,         which may be dependent on species and cell
because it could be completely inhibited by         culture    procedures.     Primary     cultured
verapamil. The authors speculated that the          nonproliferated canine VSMC expressed iNOS,
mechanism could be inhibition of Ca2+-influx by     but no eNOS, by Western blot (233). In
changing the sensitivity of voltage-operated        contrast, cultured human VSMC expressed
Ca2+-channels to more depolarized values            eNOS mRNA, detected by Northern blot,
through stimulation of the Na+-K+ pump,             whereas      iNOS    expression      was    not
because the effects of insulin were only minor      examined (238). Neither study examined nNOS
during high extracellular K+ concentrations and     expression. These findings seem irreconcilable,
completely inhibited by oubain (an inhibitor of     regardless of the fact that iNOS expression
the the Na+-K+ pump) (230).                         could be an artifact of the cell isolation
                                                    procedure which takes several hours and
Removing glucose from the medium or                 although there may be a discrepancy of
inhibiting glucose uptake with phloridzin           protein and mRNA expression. It is probably
completely inhibited the effects of insulin (143    not a cell culture artifact, since it was also
mU l−1 for 30 minutes) on Ca2+ influx and cell      detected in freshly dispersed VSMC (234).
contraction (231). A later study showed that        Conversely, nNOS mRNA and protein was
pharmacological inbibition of glycolysis or         demonstrated in rat carotid artery (241).
abscence of glucose in the medium prevented
insulin-stimulated       elevation    of      the   Regardless     of     the     source,    cultured
lactate/pyruvate ratio (an index of increased       nonproliferated canine VSMC had a basal
cytosolic NADH/NAD+ ratio) (170) and cGMP           production of NO, which was inhibited by L-
production (234).       Incubation    with     β-   NMMA (232, 234). Serotonin, insulin, or their
hydroxybutyrate, which reduces NAD+ in the β-       combination      did     not      increase    NO
hydroxybutyrate      dehydrogenase      reaction,   production (234).     Insulin    increased    the
elevated the lactate/pyruvate ratio and             production of cGMP (232, 236) as opposed to
mimicked          insulin-stimulated       cGMP     inhibit its catabolism (236)) and such
production (234). It follows that these effects     stimulation could be blocked by genistein (a
of insulin on VSMC are dependent on                 tyrosine kinase inhibitor, and thereby an IR
glucolysis, possibly by a change to a more          kinase inhbibitor) (236). Furthermore, L-NMMA
electropositive standard reduction potential.        (232, 236) and methylene blue (an inhibitor of
Although speculative, this mechanism could          guanylate cyclase) (236) inhibited basal cGMP
involve the redox balance of heme of                production, and L-NMMA inhibited insulin-
guanylate cyclase, which is inactivated in its      stimulated cGMP production after serotonin
oxidized, ferric (Fe3+) form (239). Keeping the     pretreatment (232, 234). L-NMMA also blocked
heme of guanylate cyclase in its active ferrous     the inhibiting effect of insulin (143 mU l−1 for
(Fe2+) form may be controlled by a flavoprotein     30      minutes)       on      serotonin-induced
oxidoreductase (171, 239) that uses NADPH as        contraction (232). Insulin and the NO donor
a cofactor (generated, for example, by the          SNAP had additive effects on cGMP production,
pentose phosphate pathway (171), see page           and insulin-stimulated cGMP production was
20). On the other hand, superoxide production       not inhibited by L-NMMA if pretreatment with
from vascular NAD(P)H oxidase (which favors         SNAP had restored cGMP to control
NADH as a cofactor (166)) may oxidize               values (234). Thus, insulin potentiated NO-
ferreous heme. This can be counteracted by          mediated cGMP production, but this effect was
promoting lactate production in the lactate         not mediated through NO formation.
dehydrogenase reaction, thereby decreasing          A later study showed that insulin had effects
the NADH/NAD+ ratio (240) (Pyruvate + NADH          on VSMC independent of Ca2+ transients (233).
+ H+ ↔ L-lactate + NAD+). Thus, glycolysis and      Intracellular Ca2+ was clamped by adding Ca2+
the lactate/pyruvate ratio may influence            ionophores to a medium with very high
guanylate cyclase sensitivity to NO depending       calcium concentration (1 µM-1.8 mM). This
on whether cellular enzymatic activity promote      had a constrictor effect, but precluded
oxidation or reduction of guanylate cyclase         intracellular Ca2+-dependent signaling. Insulin
heme from NADPH or NADH.                            inhibited ionophore-induced constriction, and
                                                    this effect was abolished by a guanylate
                                                             Christian Rask Madsen: PhD thesis, Page 26 of 40.

cyclase inhibitor or L-NMMA (233). Inhibition of      insulin for 45 minutes did not have any
Ca2+ influx as well as cGMP production and            effect (242).
conctraction during “clamped” intracellular Ca2+
                                                      Only one published study has examined the
concentrations     may     be   mediated     by
                                                      effect of a blood glucose-lowering intervention
stimulating PKC activity (235).
                                                      on endothelium-dependent vasodilation. It
Insulin also stimulated cAMP production in            showed that the forearm vasodilator response
VSMC (236), and NO-cGMP production in                 to     intrabrachially    infused      acetylcholine
VSMC potentiated the catecholamine-induced            increased after adding bed-time intermediate-
increase in VSMC cAMP concentrations (237).           acting insulin to established therapy with
                                                      metformin in patients with type 2 diabetes
Effects of insulin on endothelium-independent         without manifest cardiovascular disease (153).
vasorelaxation ex vivo or vasodilation in vivo        In previous reports of studies with long-term
Taken together, cell culture studies show that        treatment      with     the    insulin    sentisizer
insulin potentiate responses to NO formed in          troglitazone,      glycemic        control       was
or supplied to VSMC. This could account for an        unchanged (154, 155). They showed that such
unknown part of the insulin-stimulated                treatment improved brachial artery flow-
acetylcholine response in vivo or increase the        mediated vasodilation in patients with impaired
apparant vascular smooth muscle sensitivity           glucose tolerance (154), but did not change
for NO. This is not supported by studies in           forearm       acetylcholine       responses        or
vivo. As described above (page 22), insulin           vasodilation to systemic hyperinsulinemia in
had no vasodilating effect on rat cremaster           patients     with     obesity-associated      insulin
muscle arterioles when endothelium was                resistance (155).
removed (192). In isolated rat gastrocnemius
                                                      Hypothetical mechanism of improved insulin-
muscle arterioles, removal of endothelium or L-
                                                      stimulated acetylcholine response
NNA changed the vasodilator response to a
vasoconstrictor response (193). On the other          Insulin therapy in the present study may have
hand, insulin has been shown to inhibit               improved endothelial function by increasing
phenylephrine-induced vascular contraction in         the expression of eNOS (121, 163) or by
rat aorta with endothelium removed (194). In          increasing NO production through upregulation
lean,      healthy        humans,         systemic    of endothelial BH4 concentrations (243). Such
hyperinsulinemia increased blood flow by 32%          mechanisms may be relevant even though an
(lowest dose) and 19% (two highest doses) in          effect insulin therapy could not be
lean, healthy volunteers, a difference which          demonstrated on the acetylcholine response
was not statistically significant (150). In a later   itself. It could also have increased NO
study from the same laboratory, the sodium            production by limiting any of the effects
nitroprusside response in lean, healthy               mentioned above of hyperglycemia or elevated
volunteers and patients with type 2 diabetes          FFA on endothelial function.
was neither larger nor statistically different
                                                      Finally, insulin therapy could have improved
when        repeated         during       systemic
                                                      endothelial signaling through the endothelial
hyperinsulinemia (6), but the sample size for
                                                      PI3K-Akt-eNOS pathway (see page 28). The
the insulin-stimulated sodium nitroprusside
                                                      last putative mechanism is certain to earn
study in patients with type 2 diabetes was very
                                                      inspiration from the intense research interest
small (n = 3). The sodium nitroprusside
                                                      in the similar signaling abnormalities in skeletal
response in a control group of lean, healthy
                                                      muscle and fat (244). This signaling pathway is
volunteers (n = 6) in another study (151) was
                                                      inhibited by the pro-inflammatory cytokine
neither larger nor statistically different during
                                                      TNF-α (see page 29). In patients with type 2
intra-brachial insulin infusion at the same dose
                                                      diabetes, cytokine levels in plasma (245) and
as in the present study.
                                                      cytokine secretion from monocytes (246)
Hypoglycemic       or    insulin     sentisizing      decreases with increased glycemic control.
interventions and endothelial function in vivo        Similarly, thiazolidinediones, a new class of
                                                      insulin sentisizing drugs, may improve insulin
Insulin treatment from 48 hours following             signaling through modulation of cytokine
streptozotozin treatment prevented impaired           activation, as they prevent the inhibitory
endothelium-dependent relaxation of aorta             effects of TNF-α on IR kinase, IRS-1, and PI3K
from diabetic rats, whereas incubation with           activity in fat cells (247), and decrease
                                                          Christian Rask Madsen: PhD thesis, Page 27 of 40.

cytokine levels in plasma in patients with type    5.     In patients, a potentiation of the
2 diabetes (245).                                  vasodilator response by insulin was present
                                                   after 2 months of insulin treatment in a
Endothelial dysfunction as a determinant of        condition of improved glycemic control. The
peripheral insulin resistance                      mediator of this vasodilator effect of insulin
The ability of insulin to increase NO-dependent    was not identified.
basal blood flow and a blunted vasodilator         There are tree limitations, in particular, to the
response to insulin in insulin reistance and       study. One is that the treatment allocation was
type 2 diabetes has been taken as evidence         only partly randomized. Another is that
that endothelium-dependent vasodilatation          pharmacological control experiments were not
may determine skeletal muscle glucose uptake       able to identify whether the insulin-stimulated
by regulating glucose supply to peripheral         acetylcholine response, and its changes, were
tissues (248). This issue has been the subject     NO mediated; and, indeed, whether changes
of a particularly unrelenting debate (249, 250)    were endothelium-dependent. The third is that
and is outside the scope of this discussion.       follow-up      of     the      insulin-stimulated
                                                   acetylcholine response in the time control
Conclusions                                        group after 2 months was only performed for
                                                   the last 4 patients, resulting in little statistical
The present study examined peripheral              power to detect any changes unrelated to
vasodilation of resistance vessels in patients     insulin therapy.
with type 2 diabetes and ischemic heart
disease as well as in lean, healthy volunteers.
The following conclusions are listed in the
sequence of the hypothesis (see page 3).
                                                   Implications for improved endothelial function
1.     The previous well-established findings of   and clinical outcomes
decreased endothelium-dependent vasodilation
                                                   Restoration of the normal response with
in patients with type 2 diabetes was confirmed
                                                   respect        to         endothelium-dependent
in the present group of patients with type 2
                                                   vasodilatation has been demonstrated with
diabetes and ischemic heart disease, compared
                                                   numerous clinical interventions known to
to age-matched, lean, healthy controls.
                                                   improve the prognosis of people with risk
Endothelium-independent           NO-mediated
                                                   factors from cardiovascular disease. Thus,
vasodilation was also found to be impaired.
                                                   decreased                 endothelium-dependent
This was also the result of certain previous
                                                   vasodilatation has been improved by exercise
studies of patients with diabetes or other
                                                   training in heart failure, by ACEI treatment in
conditions     associated   with    endothelial
                                                   heart failure, type 1, and type 2 diabetes, by
dysfunction, but is in disagreement with the
                                                   statin     therapy       in        patients     with
majority of studies of patients with
                                                   hypercholesterolemia or by antihypertensive
uncomplicated type 2 diabetes.
                                                   therapy in arterial hypertension. Improvement
2.     The results confirmed previous findings     of abnormal endothelial function has also been
in healthy non-obese people regarding the          linked to therapy for which the evidence for a
ability of locally elevated insulin levels to      beneficial clinical outcome is still missing or
stimulate vasodilation to an endothelium-          controversial, i.e. hypoglycemic therapy in
dependent agonist.                                 patients with type 2 diabetes, orally
                                                   administered L-arginine for coronary or
3.    In patients compared to healthy
                                                   peripheral    vascular      disease,     or    orally
controls, insulin-stimulated potentiation of the
                                                   administered vitamin C for patients with
vasodilator response to an endothelium-
                                                   ischemic heart disease. Given the evidence for
dependent agonist was absent.
                                                   the role of endothelial dysfunction in the
4.    Despite of a considerable improvement        pathogenesis of vascular disease outlined in
of glycemic control, insulin therapy did not       the background section, it is likely that
change either endothelium-dependent or -           restoration of insulin-stimulated endothelial
independent vasodilation after 2 months.           function towards the normal response is
Surprisingly, both responses were depressed        beneficial. It may be very difficult to obtain
after 3 days of insulin treatment, where a         evidence for this without means to selectively
considerable improvement of fasting blood          improve      or      inhibit      insulin-stimulated
glucose was already acheived.                      endothelial      function      in      models     of
                                                          Christian Rask Madsen: PhD thesis, Page 28 of 40.

atherosclerosis, which will more likely be         improved glucose tolerance and normalized
acheived by molecular biology techniques than      skeletal muscle Akt activity and glucose
pharmacological approaches.                        uptake (253). Furthermore, exercise training of
                                                   normal rats increase skeletal muscle PI3K
There are important theoretical reservations
                                                   activity and Akt phosphorylation (254) and
towards insulin therapy in this context. Insulin
                                                   pancreatic islet cell transplantation in
signaling pathways may be differentially
                                                   streptozotocin-treated diabetic rats improved
affected in type 2 diabetes and insulin
                                                   IR kinase, IRS, and PI3K activity (255) and Akt
resistance. The insulin signal transmitted by
                                                   activity (256) in skeletal muscle. The present
IRS may be viewed as diverging into the PI3K
                                                   study supports the suggestion of a common
pathway described above (see page 23) and
                                                   pathway of insulin signaling in muscle, fat, and
into the mitogen activated protein kinase
                                                   endothelium. Furthermore, it supports a
(MAPK) pathway, which relays insulin’s effects
                                                   shared defect of insulin-stimulated glucose
as an anabolic and mitogenic hormone. In
                                                   uptake and insulin-stimulated NO production in
insulin resistance, the PI3K pathway of insulin
                                                   insulin resistance associated with obesity and
signaling seems selectively impaired, whereas
                                                   type 2 diabetes, because insulin treatment and
the MAPK pathway has been shown to be
                                                   improved glycemic control improved a
unaffected, both in endothelium (11) and in
                                                   functional measure of insulin-stimulated
skeletal        muscle (251).      Accordingly,
                                                   endothelial NO production.
administration of insulin may beneficially
upregulate insulin’s effects on glucose            Insulin resistance secondary to hyperglycemia
transport and metabolism and endothelial NO        has been interpreted as an adaptive response
production, but adversely upregulate insulin’s     to maintain a normal cellular glucose uptake
effects on vascular proliferation, perhaps         during a chronic state of hyperglycemia (257).
leading to diabetic complications other than       Thus, even though patients with type 2
atherosclerosis, for example retinopathy.          diabetes experience hyperglycemia after a
Hyperinsulinemia may also alter vascular           meal, primarily due to decreased inhibition of
homeostasis in a pro-atherogenic fashion by        hepatic glucose output, postprandial glucose
signaling through other than the classical         uptake is normal (258). It has been suggested
insulin-receptor dependent pathways. For           that the hexosamine pathway is a metabolic
example, insulin stimulates NF-κB in vascular      sensor of glucose levels that mediates insulin
smooth muscle cells (252). This nuclear factor     resistance in response to hyperglycemia (259).
is thought to be important for several early
                                                   This notion that insulin resistance may protect
events in atherogenesis, for example by
                                                   insulin       sensitive       tissues        from
activation of the genes for TNF-α and VCAM-1.
                                                   complications (257)      has     recently    been
Obviously then, the overall vascular effects of    challenged (175). Thus, vascular disease is a
insulin administration may be complex and          major complication in type 2 diabetes even
paradoxical. Understanding vascular insulin        though vascular tissues are insulin sensitive by
signaling will be important because in the         virtue of eNOS activation through the insulin-
future, pharmacological agents with selective      PI3K-Akt pathway. Glucose uptake in
effects on insulin signaling will become           endothelium is primarily dependent on insulin-
available,     one    example     being     the    insensitive GLUT1. In contrast to vascular
thiazolidinediones (see page 26).                  smooth muscle cells, GLUT1 is not down-
                                                   regulated     in    endothelial     cells   during
Implications for the concept of endothelial        hyperglycemia (260).       In     this    context,
insulin resistance                                 endothelial insulin resistance may be viewed as
It is well-established that improved glycemic      an untoward result of an adaptive response
control in patients with type 2 diabetes           that does not protect endothelium from
increases insulin sensitivity, measued as          increased cellular glucose uptake but adversely
whole-body        insulin-stimulated  glucose      affects NO production, which is dependent on
uptake (12). It is unknown whether clinical        the same signal transduction pathway as
insulin therapy improves expression or activity    glucose uptake. Such a concept is, of course,
of the IR, IRS, PI3K, or Akt in any tissue.        entirely speculative.
However, treatment of Goko-Kakizaki rats (a
non-obese models of type 2 diabetes) with
phloridzin (which improve glucose levels by
inhibition of renal glucose reabsorption)
                                                            Christian Rask Madsen: PhD thesis, Page 29 of 40.

Perspectives                                        endothelium-independent vasodilatation, i.e.
                                                    the sodium nitroprusside response, will be
The study presented in this thesis has              made in each subject on a seperate study day.
prompted new questions regarding the natural
history of endothelial dysfunction of insulin       TNF-α
resistance and type 2 diabetes and the              Acute subclinical inflammation in healthy
mechanisms of the decreased effect of insulin       humans leads to a substantial temporary
to stimulate endothelial function. Therefore,       reduction of endothelial function (265), which
work has been planned in co-operation with          may      explain   the    increased     risk    for
several collegues, with central efforts delivered   cardiovascular events shortly after infectious
by Drs. Christian Torp-Pedersen, Nikolaj            disease (266).      On     the    other      hand,
Ihlemann, Helena Domínguez, and Thomas              atherogenesis has been described as a chronic
Hermann, all at Gentofte University Hospital.       inflammatory      condition (267).   The      pro-
The project on FFAs described below is the          inflammatory cytokine TNF-α probably plays a
focus of my current work at the Joslin Diabetes     role    in    the    development     of     insulin
Center, Boston, MA, USA, supervised by Dr.          resistance (268) which is an important risk
George L. King.                                     factor for cardiovascular disease (47). Thus,
BH4                                                 abnormal expression of TNF-α may be an
                                                    explanation for the association between
One planned study will continue the work with       vascular inflammation, metabolic insulin
BH4. Paradoxically, conditions as different as      resistance, endothelial insulin signaling, and
hypercholesterolemia, hypertension, diabetes,       atherogenesis.
and smoking may all lead to the same disease,
atherosclerosis. It is an exciting prospect that    A planned study will examine the acute effect
they     may      all   cause      a    common      of TNF-α on either the acetylcholine response
pathophysiological condition which lead to          or    the    insulin-stimulated  acetylcholine
endothelial dysfunction and ultimatively to         response in young, healthy male volunteers.
atherosclerosis (261). Such a condition is
                                                    Birth weight
beleived to be represented by increased
vascular production of superoxide anion or          In the first descriptions of the insulin
other reactive oxygen species (261). Oxidative      resistance syndrome (269), one of the central
stress may inhibit NO-mediated endothelial          observations was that the development of type
function by degrading NO (262). Alternatively,      2 diabetes and cardiovascular disease often
it may lead to decreased intracellular              occurs simultaneously in individuals with
concentrations of the reduced, active form of       insulin resistance. One of the most interesting
BH4, a cofactor for eNOS. Suboptimal BH4            theories that attempts to explain this “common
concentrations lead to “uncoupling” of the          soil” theory is that intrauterine growth
oxidase and reductase domain of NO                  retardation may predispose to insulin
synthase (263) which results in oxidation of        resistance, type 2 diabetes and cardiovascular
molecular oxygen instead of L-arginine and          disease decades later in adult life (270).
synthesis of superoxide instead of NO. Thus,        Genetic factors cannot adequately explain a
BH4 deficiency may be both an effect and            higher frequence of a history of low
cause of vascular oxidative stress (264).           birthweight in patients with type 2
                                                    diabetes (271). However, growth retardation
The forearm vasodilator response to co-
                                                    due to malnutrition or placental insufficiency
infusion of BH4 and acetylcholine will be
                                                    may adversely affect several organ systems
compared to the acetylcholine response in
                                                    permanently. Such changes are termed “fetal
patients with insulin resistance. BH4 is an
                                                    programming” (270). As metabolic insulin
antioxidant, although much less potent than
                                                    resistance is one of the earliest changes
for example vitamin C. Experiments in the
                                                    detectable in individuals at risk for type 2
planned study controlling for an anti-oxidant
                                                    diabetes (272), insulin-stimulated endothelial
effect will be made on seperate study days by
                                                    function may be an early vascular change in
infusing a stereoisomer (6S-BH4) without
                                                    individuals prone to the development of type 2
cofactor capacities. In experiments in vitro we
have shown that 6R-BH4 (the active
stereoisomer) and 6S-BH4 are equipotent anti-       The planned study will examine endothelial
oxidants (figure 18). Additional experiments        function and insulin-stimulated endothelial
will control for the effect of 6R-BH4 on            function in young men born in 1980 and
                                                     Christian Rask Madsen: PhD thesis, Page 30 of 40.

randomly selected from the Danish Medical
Birth register and Central Personal Register
into two groups of birth weight in the lowest
10%-percentile or the highest 75%-percentile
of their year, respectively. All had normal
gestational age. These subjects were recently
characterized with regard to a host of
metabolic characteristics, including data from
an euglycemic hyperinsulinemic clamp.

FFAs and PKC
The metabolic dysregulation in type 2 diabetes
includes increases in plasma concentrations of
both glucose and FFAs. Circulating FFAs are
thought to inhibit insulin-stimulated glucose
uptake, and this effect may be caused by a
direct interaction with insulin signaling (273)
rather than a previously proposed mechanism
of allosteric inhibition of glycolytic enzymes. As
FFAs activate PKC in non-vascular cells (274),
and since vascular PKC is activated in animal
models of diabetes, we plan to study the effect
of FFAs on endothelial PKC activation.
Furthermore, as insulin signaling through the
pathway involving PI3K and Akt is decreased in
the vasculature of rat models of type 2
diabetes, whereas the pathway involving
mitogen-aktivated protein kinase (MAPK) is
intact (11), we hypothesize that FFA-induced
PKC activation may inhibit the PI3K/Akt
pathway and simultaneously activate the MAPK
pathway, leading to decreased nitric oxide
(NO) expression and increased endothelin-1
(ET-1) expression, respectively. The planned
studies include experiments in endothelial cell
culture and experiments with fat infusion in
rats and mice with null mutations in PKC genes
or overexpression of selected PKC isoforms.
                                                                         Christian Rask Madsen: PhD thesis, Page 31 of 40.

                                                                hyperpolarizing factor in mice [In Process Citation]. J Clin
References                                                      Invest 2000. 106(12): 1521-30.
                                                                16.      Price DT, Loscalzo J. Cellular adhesion molecules
1.        Amos AF, McCarty DJ, Zimmet P. The rising
                                                                and atherogenesis. Am J Med 1999. 107(1): 85-97.
global burden of diabetes and its complications: estimates
and projections to the year 2010. Diabet Med 1997.              17.       Harrison DG, Armstrong ML, Freiman PC,
14(Suppl 5): S1-85.                                             Heistad DD. Restoration of endothelium-dependent
                                                                relaxation by dietary treatment of atherosclerosis. J Clin
2.        Seibaek M, Sloth C, Vallebo L, Hansen T,
                                                                Invest 1987. 80(6): 1808-11.
Urhammer SA, Burchardt H, Torp-Pedersen C, Pedersen O,
Hildebrandt P. Glucose tolerance status and severity of         18.       Cayatte AJ, Palacino JJ, Horten K, Cohen RA.
coronary artery disease in men referred to coronary             Chronic inhibition of nitric oxide production accelerates
arteriography. Am Heart J 1997. 133(6): 622-9.                  neointima formation and impairs endothelial function in
                                                                hypercholesterolemic rabbits. ArteriosclerThromb 1994.
3.       Biegelsen ES, Loscalzo J. Endothelial function
                                                                14(5): 753-9.
and atherosclerosis. Coron Artery Dis 1999. 10(4): 241-56.
                                                                19.       Channon KM, Qian H, George SE. Nitric oxide
4.        Tooke     JE.    Possible    pathophysiological
                                                                synthase in atherosclerosis and vascular injury : insights
mechanisms for diabetic angiopathy in type 2 diabetes. J
                                                                from experimental gene therapy. Arterioscler Thromb Vasc
Diabetes Complications 2000. 14(4): 197-200.
                                                                Biol 2000. 20(8): 1873-81.
5.          Petrie JR, Ueda S, Webb DJ, Elliott HL, Connell
                                                                20.        Drexler H. Endothelial dysfunction: clinical
JM. Endothelial nitric oxide production and insulin
                                                                implications. ProgCardiovascDis 1997. 39(4): 287-324.
sensitivity. A physiological link with implications for
pathogenesis of cardiovascular disease. Circulation 1996.       21.       Suwaidi JA, Hamasaki S, Higano ST, Nishimura
93(7): 1331-3.                                                  RA, Holmes DR, Jr., Lerman A. Long-term follow-up of
                                                                patients with mild coronary artery disease and endothelial
6.        Steinberg HO, Chaker H, Leaming R, Johnson A,
                                                                dysfunction. Circulation 2000. 101(9): 948-54.
Brechtel G, Baron AD. Obesity/insulin resistance is
associated with endothelial dysfunction. Implications for       22.        Schachinger V, Britten MB, Zeiher AM.
the syndrome of insulin resistance. J Clin Invest 1996.         Prognostic impact of coronary vasodilator dysfunction on
97(11): 2601-10.                                                adverse long- term outcome of coronary heart disease.
                                                                Circulation 2000. 101(16): 1899-906.
7.        Cleland SJ, Petrie JR, Small M, Elliott HL, Connell
JM. Insulin action is associated with endothelial function in   23.       Neunteufl T, Heher S, Katzenschlager R, Wolfl G,
hypertension and type 2 diabetes. Hypertension 2000.            Kostner K, Maurer G, Weidinger F. Late prognostic value of
35(1 Pt 2): 507-11.                                             flow-mediated dilation in the brachial artery of patients
                                                                with chest pain. Am J Cardiol 2000. 86(2): 207-10.
8.        Laakso M, Edelman SV, Brechtel G, Baron AD.
Decreased effect of insulin to stimulate skeletal muscle        24.       O'Driscoll G, Green D, Maiorana A, Stanton K,
blood flow in obese man. A novel mechanism for insulin          Colreavy F, Taylor R. Improvement in endothelial function
resistance. J Clin Invest 1990. 85(6): 1844-52.                 by angiotensin-converting enzyme inhibition in non-insulin-
                                                                dependent diabetes mellitus. J AmCollCardiol 1999. 33(6):
9.        Laakso M, Edelman SV, Brechtel G, Baron AD.
Impaired insulin-mediated skeletal muscle blood flow in
patients with NIDDM. Diabetes 1992. 41(9): 1076-83.             25.        Hingorani AD, Liang CF, Fatibene J, Lyon A,
                                                                Monteith S, Parsons A, Haydock S, Hopper RV, Stephens
10.      Zeng G, Quon MJ. Insulin-stimulated production
                                                                NG, O'Shaughnessy KM, Brown MJ. A common variant of
of nitric oxide is inhibited by wortmannin. Direct
                                                                the endothelial nitric oxide synthase (Glu298>Asp) is a
measurement in vascular endothelial cells. J Clin Invest
                                                                major risk factor for coronary artery disease in the UK.
1996. 98(4): 894-8.
                                                                Circulation 1999. 100(14): 1515-20.
11.       Jiang ZY, Lin YW, Clemont A, Feener EP, Hein
                                                                26.       Boger RH, Bode-Boger SM, Thiele W, Creutzig A,
KD, Igarashi M, Yamauchi T, White MF, King GL.
                                                                Alexander K, Frolich JC. Restoring vascular nitric oxide
Characterization of selective resistance to insulin signaling
                                                                formation by L-arginine improves the symptoms of
in the vasculature of obese Zucker (fa/fa) rats. J Clin
                                                                intermittent claudication in patients with peripheral arterial
Invest 1999. 104(4): 447-57.
                                                                occlusive disease. J Am Coll Cardiol 1998. 32(5): 1336-44.
12.       Reaven GM. Therapeutic approaches to reducing
                                                                27.       Sausbier M, Schubert R, Voigt V, Hirneiss C,
insulin resistance in patients with noninsulin-dependent
                                                                Pfeifer A, Korth M, Kleppisch T, Ruth P, Hofmann F.
diabetes mellitus. Am J Med 1983. 74(1A): 109-12.
                                                                Mechanisms of NO/cGMP-dependent vasorelaxation. Circ
13.        Malmberg K, Ryden L, Efendic S, Herlitz J, Nicol     Res 2000. 87(9): 825-30.
P, Waldenstrom A, Wedel H, Welin L. Randomized trial of
                                                                28.     Carvajal JA, Germain AM, Huidobro-Toro JP,
insulin-glucose infusion followed by subcutaneous insulin
                                                                Weiner CP. Molecular mechanism of cGMP-mediated
treatment in diabetic patients with acute myocardial
                                                                smooth muscle relaxation. J Cell Physiol 2000. 184(3):
infarction (DIGAMI study): effects on mortality at 1 year.
JAmCollCardiol 1995. 26(1): 57-65.
                                                                29.       Koltai MZ, Hadhazy P, Posa I, Kocsis E, Winkler
14.       Wang W, Wang S, Yan L, Madara P, Del Pilar
                                                                G, Rosen P, Pogatsa G. Characteristics of coronary
Cintron A, Wesley RA, Danner RL. Superoxide production
                                                                endothelial dysfunction in experimental diabetes.
and reactive oxygen species signaling by endothelial nitric-
                                                                Cardiovasc Res 1997. 34(1): 157-63.
oxide synthase. J Biol Chem 2000. 275(22): 16899-903.
                                                                30.      Osanai T, Fujita N, Fujiwara N, Nakano T,
15.      Matoba T, Shimokawa H, Nakashima M,
                                                                Takahashi K, Guan W, Okumura K. Cross talk of shear-
Hirakawa Y, Mukai Y, Hirano K, Kanaide H, Takeshita A.
                                                                induced production of prostacyclin and nitric oxide in
Hydrogen    peroxide    is   an    endothelium-derived
                                                                         Christian Rask Madsen: PhD thesis, Page 32 of 40.

endothelial cells. Am J Physiol Heart Circ Physiol 2000.         47.       Haffner SM. Epidemiology of insulin resistance
278(1): H233-8.                                                  and its relation to coronary artery disease. Am J Cardiol
                                                                 1999. 84(1A): 11J-4J.
31.       Huang A, Sun D, Smith CJ, Connetta JA, Shesely
EG, Koller A, Kaley G. In eNOS knockout mice skeletal            48.       Wagenknecht LE, Mayer EJ, Rewers M, Haffner
muscle arteriolar dilation to acetylcholine is mediated by       S, Selby J, Borok GM, Henkin L, Howard G, Savage PJ,
EDHF. Am J Physiol Heart Circ Physiol 2000. 278(3): H762-        Saad MF. The insulin resistance atherosclerosis study
8.                                                               (IRAS) objectives, design, and recruitment results. Ann
                                                                 Epidemiol 1995. 5(6): 464-72.
32.       DeFronzo RA, Tobin JD, Andres R. Glucose
clamp technique: a method for quantifying insulin                49.      Despres JP, Lamarche B, Mauriege P, Cantin B,
secretion and resistance. Am J Physiol 1979. 237(3): E214-       Dagenais GR, Moorjani S, Lupien PJ. Hyperinsulinemia as
E23.                                                             an independent risk factor for ischemic heart disease.
                                                                 NEnglJMed 1996. 334(15): 952-7.
33.        DeFronzo RA, Gunnarsson R, Bjorkman O,
Olsson M, Wahren J. Effects of insulin on peripheral and         50.      Coles DR, Cooper KE, Mottram RF, Occleshaw
splanchnic glucose metabolism in noninsulin-dependent            OV. The source of blood samples withdrawn from deep
(type II) diabetes mellitus. J Clin Invest 1985. 76(1): 149-     forearm veins via catherters passed upstream from the
55.                                                              medium cubital vein. J Physiol (Lond) 1958. 142: 258-67.
34.        Kruszynska YT, Olefsky JM. Cellular and               51.        Benjamin N, Calver A, Collier J, Robinson B,
molecular mechanisms of non-insulin dependent diabetes           Vallance P, Webb D. Measuring forearm blood flow and
mellitus. J Investig Med 1996. 44(8): 413-28.                    interpreting the responses to drugs and mediators.
                                                                 Hypertension 1995. 25(5): 918-23.
35.        Weyer C, Bogardus C, Mott DM, Pratley RE. The
natural history of insulin secretory dysfunction and insulin     52.       Groome J, Vohra R, Cuschieri RJ, Gilmour DG.
resistance in the pathogenesis of type 2 diabetes mellitus.      Vascular injury after arterial catheterization. PostgradMedJ
J Clin Invest 1999. 104(6): 787-94.                              1989. 65(760): 86-8.
36.        Gerich JE. Insulin resistance is not necessarily an   53.        Knöbl P, Schernthaner G, Schnack C,
essential component of type 2 diabetes. J Clin Endocrinol        Pietschmann P, Proidl S, Prager R, Vukovich T.
Metab 2000. 85(6): 2113-5.                                       Haemostatic abnormalities persist despite glycaemic
                                                                 improvement by insulin therapy in lean type 2 diabetic
37.        Kahn BB, Flier JS. Obesity and insulin resistance.    patients. ThrombHaemost 1994. 71(6): 692-7.
J Clin Invest 2000. 106(4): 473-81.
                                                                 54.       Huszka M, Kaplar M, Rejto L, Tornai I, Palatka K,
38.      Steppan CM, Bailey ST, Bhat S, Brown EJ,                Laszlo P, Udvardy M. The association of reduced
Banerjee RR, Wright CM, Patel HR, Ahima RS, Lazar MA.            endothelium derived relaxing factor-NO production with
The hormone resistin links obesity to diabetes. Nature           endothelial damage and increased in vivo platelet
2001. 109: 307-12.                                               activation in patients with diabetes mellitus. ThrombRes
39.       Shulman GI. Cellular mechanisms of insulin             1997. 86(2): 173-80.
resistance. J Clin Invest 2000. 106(2): 171-6.                   55.       Castillo L, Beaumier L, Ajami AM, Young VR.
40.        Virkamaki A, Ueki K, Kahn CR. Protein-protein         Whole body nitric oxide synthesis in healthy men
interaction in insulin signaling and the molecular               determined from [15N] arginine-to-[15N]citrulline labeling.
mechanisms of insulin resistance. J Clin Invest 1999.            Proc Natl Acad Sci U S A 1996. 93(21): 11460-5.
103(7): 931-43.                                                  56.       Radegran G, Saltin B. Nitric oxide in the
41.      Taha C, Klip A. The insulin signaling pathway. J        regulation of vasomotor tone in human skeletal muscle.
Membr Biol 1999. 169(1): 1-12.                                   Am J Physiol 1999. 276(6 Pt 2): H1951-60.

42.       Coffer PJ, Jin J, Woodgett JR. Protein kinase B        57.        Duffy SJ, Castle SF, Harper RW, Meredith IT.
(c-Akt): a multifunctional mediator of phosphatidylinositol      Contribution of vasodilator prostanoids and nitric oxide to
3-kinase activation. Biochem J 1998. 335(Pt 1): 1-13.            resting flow, metabolic vasodilation, and flow-mediated
                                                                 dilation in human coronary circulation. Circulation 1999.
43.       Laakso M. Hyperglycemia and cardiovascular             100(19): 1951-7.
disease in type 2 diabetes. Diabetes 1999. 48(5): 937-42.
                                                                 58.        Quyyumi AA, Dakak N, Andrews NP, Gilligan DM,
44.       The effect of intensive treatment of diabetes on       Panza JA, Cannon RO, III. Contribution of nitric oxide to
the development and progression of long-term                     metabolic coronary vasodilation in the human heart.
complications in insulin-dependent diabetes mellitus. The        Circulation 1995. 92(3): 320-6.
Diabetes Control and Complications Trial Research Group.
NEnglJMed 1993. 329(14): 977-86.                                 59.       Ludmer PL, Selwyn AP, Shook TL, Wayne RR,
                                                                 Mudge GH, Alexander RW, Ganz P. Paradoxical
45.       Intensive     blood-glucose      control     with      vasoconstriction induced by acetylcholine in atherosclerotic
sulphonylureas or insulin compared with conventional             coronary arteries. NEnglJ Med 1986. 315(17): 1046-51.
treatment and risk of complications in patients with type 2
diabetes (UKPDS 33). UK Prospective Diabetes Study               60.       Vallance P, Collier J, Moncada S. Effects of
(UKPDS) Group. Lancet 1998. 352(9131): 837-53.                   endothelium-derived nitric oxide on peripheral arteriolar
                                                                 tone in man. Lancet 1989. 2: 997-1000.
46.        Malmberg K. Prospective randomised study of
intensive insulin treatment on long term survival after          61.       Sinoway LI, Hendrickson C, Davidson WR, Jr.,
acute myocardial infarction in patients with diabetes            Prophet S, Zelis R. Characteristics of flow-mediated
mellitus. DIGAMI (Diabetes Mellitus, Insulin Glucose             brachial artery vasodilation in human subjects. Circ Res
Infusion in Acute Myocardial Infarction) Study Group. BMJ        1989. 64(1): 32-42.
1997. 314(7093): 1512-5.
                                                                        Christian Rask Madsen: PhD thesis, Page 33 of 40.

62.        Anderson EA, Mark AL. Flow-mediated and reflex       Hayes     JR.  Impaired    endothelium-dependent     and
changes in large peripheral artery tone in humans.              independent vasodilation in patients with type 2 (non-
Circulation 1989. 79(1): 93-100.                                insulin-dependent) diabetes mellitus. Diabetologia 1992.
                                                                35(8): 771-6.
63.       Celermajer DS, Sorensen KE, Gooch VM,
Spiegelhalter DJ, Miller OI, Sullivan ID, Lloyd JK, Deanfield   77.       Morris SJ, Shore AC, Tooke JE. Responses of the
JE. Non-invasive detection of endothelial dysfunction in        skin microcirculation to acetylcholine and sodium
children and adults at risk of atherosclerosis. Lancet 1992.    nitroprusside in patients with NIDDM. Diabetologia 1995.
340(8828): 1111-5.                                              38(11): 1337-44.
64.      Rubanyi GM, Romero JC, Vanhoutte PM. Flow-             78.       Martin H, Hu J, Gennser G, Norman M. Impaired
induced release of endothelium-derived relaxing factor.         endothelial function and increased carotid stiffness in 9-
Am J Physiol 1986. 250(6 Pt 2): H1145-9.                        year-old children with low birthweight. Circulation 2000.
                                                                102(22): 2739-44.
65.       Joannides R, Haefeli WE, Linder L, Richard V,
Bakkali EH, Thuillez C, Luscher TF. Nitric oxide is             79.      Jaap AJ, Pym CA, Seamark C, Shore AC, Tooke
responsible for flow-dependent dilatation of human              JE. Microvascular function in type 2 (non-insulin-
peripheral conduit arteries in vivo. Circulation 1995. 91(5):   dependent) diabetes: improved vasodilation after one year
1314-9.                                                         of good glycaemic control. Diabet Med 1995. 12(12):
66.       Busse R, Fleming I. Pulsatile stretch and shear
stress: physical stimuli determining the production of          80.       Whitney RJ. The measurement of volume
endothelium-derived relaxing factors. J Vasc Res 1998.          changes in human limbs. J PhysiolLond 1953. 121: 1-27.
35(2): 73-84.
                                                                81.       Greenfield ADM. Br Med Bull 1963. 19: 101.
67.       Anderson TJ, Uehata A, Gerhard MD, Meredith
IT, Knab S, Delagrange D, Lieberman EH, Ganz P, Creager         82.       Prasad A, Tupas-Habib T, Schenke WH,
MA, Yeung AC, et a. Close relation of endothelial function      Mincemoyer R, Panza JA, Waclawin MA, Ellahham S,
in the human coronary and peripheral circulations.              Quyyumi AA. Acute and chronic angiotensin-1 receptor
JAmCollCardiol 1995. 26(5): 1235-41.                            antagonism      reverses     endothelial   dysfunction in
                                                                atherosclerosis. Circulation 2000. 101(20): 2349-54.
68.      Boegehold MA. Heterogeneity of endothelial
function within the circulation. Curr Opin Nephrol              83.       Sorensen M, Ong P, Hayward C, Webb C,
Hypertens 1998. 7(1): 71-8.                                     Gatehouse P, Collins P, Pennell D. Gradient echo magnetic
                                                                resonance imaging of the brachial artery: a novel method
69.        Drexler H, Zeiher AM, Meinzer K, Just H.             for assessing vascular reactivity. Circulation 1999. 100(18
Correction of endothelial dysfunction in coronary               Suppl): I-49 (abstract).
microcirculation of hypercholesterolaemic patients by L-
arginine. Lancet 1991. 338(8782-8783): 1546-50.                 84.       Bank AJ, Wang H, Holte JE, Mullen K, Shammas
                                                                R, Kubo SH. Contribution of collagen, elastin, and smooth
70.        Treasure CB, Manoukian SV, Klein JL, Vita JA,        muscle to in vivo human brachial artery wall stress and
Nabel EG, Renwick GH, Selwyn AP, Alexander RW, Ganz P.          elastic modulus. Circulation 1996. 94(12): 3263-70.
Epicardial coronary artery responses to acetylcholine are
impaired in hypertensive patients. Circ Res 1992. 71(4):        85.        Longhurst J, Capone RJ, Mason DT, Zelis R.
776-81.                                                         Comparison of blood flow measured by plethysmograph
                                                                and flowmeter during steady state forearm exercise.
71.        Nitenberg A, Valensi P, Sachs R, Dali M, Aptecar     Circulation 1974. 49(3): 535-40.
E, Attali JR. Impairment of coronary vascular reserve and
ACh-induced coronary vasodilation in diabetic patients with     86.        Jorfeldt L, Rutberg H. Comparison of dye-
angiographically normal coronary arteries and normal left       dilution and plethysmographic blood flow measurements:
ventricular systolic function. Diabetes 1993. 42(7): 1017-      an evaluation of the influence of invasive techniques on
25.                                                             blood flow and on arterial and femoral venous substrate
                                                                variables in man. Clin Sci (Cholc) 1990. 79(1): 81-7.
72.       Li J, Zhao SP, Li XP, Zhuo QC, Gao M, Lu SK.
Non-invasive detection of endothelial dysfunction in            87.        Utriainen T, Malmstrom R, Makimattila S, Yki-
patients with essential hypertension [see comments]. Int J      Jarvinen H. Methodological aspects, dose-response
Cardiol 1997. 61(2): 165-9.                                     characteristics and causes of interindividual variation in
                                                                insulin stimulation of limb blood flow in normal subjects.
73.       Goodfellow J, Ramsey MW, Luddington LA,               Diabetologia 1995. 38(5): 555-64.
Jones CJ, Coates PA, Dunstan F, Lewis MJ, Owens DR,
Henderson AH. Endothelium and inelastic arteries: an early      88.       Pedrinelli R, Taddei S, Graziadei L, Salvetti A.
marker of vascular dysfunction in non-insulin dependent         Vascular responses to ouabain and norepinephrine in low
diabetes. BMJ 1996. 312(7033): 744-5.                           and normal renin hypertension. Hypertension 1986. 8(9):
74.      Chowienczyk PJ, Watts GF, Cockcroft JR, Ritter
JM. Impaired endothelium-dependent vasodilation of              89.      Petrie JR, Ueda S, Morris AD, Murray LS, Elliott
forearm resistance vessels in hypercholesterolaemia.            HL, Connell JM. How reproducible is bilateral forearm
Lancet 1992. 340(8833): 1430-2.                                 plethysmography? BrJClinPharmacol 1998. 45(2): 131-9.

75.       Linder L, Kiowski W, Buhler FR, Luscher TF.           90.       Lind L, Sarabi M, Millgard J. Methodological
Indirect evidence for release of endothelium-derived            aspects of the evaluation of endothelium-dependent
relaxing factor in human forearm circulation in vivo.           vasodilatation in the human forearm. Clin Physiol 1998.
Blunted response in essential hypertension. Circulation         18(2): 81-7.
1990. 81(6): 1762-7.                                            91.        Utriainen T, Holmang A, Bjorntorp P, Makimattila
76.     McVeigh GE, Brennan GM, Johnston GD,                    S, Sovijarvi A, Lindholm H, Yki-Jarvinen H. Physical fitness,
McDermott BJ, McGrath LT, Henry WR, Andrews JW,                 muscle morphology, and insulin-stimulated limb blood flow
                                                                       Christian Rask Madsen: PhD thesis, Page 34 of 40.

in normal subjects. Am J Physiol 1996. 270(5 Pt 1): E905-     107.      List BM, Klosch B, Volker C, Gorren AC, Sessa
E11.                                                          WC, Werner ER, Kukovetz WR, Schmidt K, Mayer B.
                                                              Characterization of bovine endothelial nitric oxide synthase
92.        Vanhoutte PM. How to assess endothelial            as a homodimer with down-regulated uncoupled NADPH
function in human blood vessels. J Hypertens 1999. 17(8):     oxidase activity: tetrahydrobiopterin binding kinetics and
1047-58.                                                      role of haem in dimerization. Biochem J 1997. 323(Pt 1):
93.       Chowienczyk PJ, Cockcroft JR, Ritter JM. Blood      159-65.
flow responses to intra-arterial acetylcholine in man:        108.        Sander M, Chavoshan B, Victor RG. A large
effects of basal flow and conduit vessel length. Clin Sci     blood pressure-raising effect of nitric oxide synthase
(Cholc) 1994. 87(1): 45-51.                                   inhibition in humans. Hypertension 1999. 33(4): 937-42.
94.        Jensen FS, Skovgaard LT, Viby-Mogensen J.          109.       Hink U, Li H, Mollnau H, Matheis E, Hartmann M,
Identification of human plasma cholinesterase variants in     Skatchkov M, Thaiss F, Munzel T. Effects of diabetes
6,688 individuals using biochemical analysis. Acta            mellitus on nitric oxide synthase (NOS III) expression,
Anaesthesiol Scand 1995. 39(2): 157-62.                       NOS III mediated superoxide production and vascular
95.        Bruning TA, Hendriks MG, Chang PC, Kuypers         nitric oxide bioavailability. Circulation 1999. 100(18 Suppl):
EA, van Zwieten PA. In vivo characterization of               I-235 (abstract).
vasodilating muscarinic-receptor subtypes in humans. Circ     110.     Faraci WS, Nagel AA, Verdries KA, Vincent LA,
Res 1994. 74(5): 912-9.                                       Xu H, Nichols LE, Labasi JM, Salter ED, Pettipher ER. 2-
96.        Chowienczyk PJ, Cockcroft JR, Ritter JM.           Amino-4-methylpyridine as a potent inhibitor of inducible
Differential inhibition by NG-monomethyl-L-arginine of        NO synthase activity in vitro and in vivo. Br J Pharmacol
vasodilator effects of acetylcholine and methacholine in      1996. 119(6): 1101-8.
human forearm vasculature. BrJPharmacol 1993. 110(2):         111.      Andres R, Cader G, Zierler KL. The quantitatively
736-8.                                                        minor role of carbohydrate in oxidative metabolism by
97.        Kummer W, Haberberger R. Extrinsic and             skeletal muscle in intact man in the basal state:
intrinsic cholinergic systems of the vascular wall. EurJ      measurements of oxygen and glucose uptake and carbon
Morphol 1999. 37(4-5): 223-6.                                 dioxide and lactate production in the forearm. J Clin Invest
                                                              1956. 35: 671-82.
98.       Stroes ES, Koomans HA, de BT, Rabelink TJ.
Vascular function in the forearm of hypercholesterolaemic     112.      Sartori C, Trueb L, Nicod P, Scherrer U. Effects
patients off and on lipid-lowering medication. Lancet 1995.   of sympathectomy and nitric oxide synthase inhibition on
346(8973): 467-71.                                            vascular actions of insulin in humans. Hypertension 1999.
                                                              34(4): 586-9.
99.        Cockcroft JR, Chowienczyk PJ, Brett SE, Ritter
JM. Effect of NG-monomethyl-L-arginine on kinin-induced       113.      Scherrer U, Sartori C. Insulin as a vascular and
vasodilation in the human forearm. BrJ Clin Pharmacol         sympathoexcitatory hormone: implications for blood
1994. 38(4): 307-10.                                          pressure regulation, insulin sensitivity, and cardiovascular
                                                              morbidity. Circulation 1997. 96(11): 4104-13.
100.      Panza JA, Casino PR, Kilcoyne CM, Quyyumi AA.
Impaired endothelium-dependent vasodilation in patients       114.      Vollenweider P, Randin D, Tappy L, Jequier E,
with essential hypertension: evidence that the abnormality    Nicod P, Scherrer U. Impaired insulin-induced sympathetic
is not at the muscarinic receptor level. J Am Coll Cardiol    neural activation and vasodilation in skeletal muscle in
1994. 23(7): 1610-6.                                          obese humans. J Clin Invest 1994. 93(6): 2365-71.

101.       Chowienczyk PJ, Cockcroft JR, Ritter JM.           115.     Hausberg M, Hoffman RP, Somers VK, Sinkey
Inhibition of acetylcholinesterase selectively potentiates    CA, Mark AL, Anderson EA. Contrasting autonomic and
NG-      monomethyl-L-arginine-resistant     actions    of    hemodynamic effects of insulin in healthy elderly versus
acetylcholine in human forearm vasculature. Clin Sci          young subjects. Hypertension 1997. 29(3): 700-5.
(Cholc) 1995. 88(1): 111-7.                                   116.      Ueda S, Petrie JR, Cleland SJ, Elliott HL, Connell
102.      Husain S, Andrews NP, Mulcahy D, Panza JA,          JM. The vasodilating effect of insulin is dependent on local
Quyyumi AA. Aspirin improves endothelial dysfunction in       glucose uptake: a double blind, placebo-controlled study. J
atherosclerosis. Circulation 1998. 97(8): 716-20.             Clin Endocrinol Metab 1998. 83(6): 2126-31.

103.      Kowaluk EA, Seth P, Fung HL. Metabolic              117.      Cardillo C, Kilcoyne CM, Nambi SS, Cannon RO,
activation of sodium nitroprusside to nitric oxide in         III, Quon MJ, Panza JA. Vasodilator response to systemic
vascular smooth muscle. JPharmacolExpTher 1992.               but not to local hyperinsulinemia in the human forearm.
262(3): 916-22.                                               Hypertension 1998. 32(4): 740-5.

104.       Mohazzab HK, Kaminski PM, Agarwal R, Wolin         118.      Mather K, Laakso M, Edelman S, Hook G, Baron
MS. Potential role of a membrane-bound NADH                   A. Evidence for physiological coupling of insulin-mediated
oxidoreductase in nitric oxide release and arterial           glucose metabolism and limb blood flow. Am J Physiol
relaxation to nitroprusside. Circ Res 1999. 84(2): 220-8.     Endocrinol Metab 2000. 279(6): E1264-70.

105.      Oszajca J, Stochel G, Wasielewska E, Stasicka Z,    119.      Dela F, Larsen JJ, Mikines KJ, Galbo H. Normal
Gryglewski     RJ,     Jakubowski      A,   Cieslik     K.    effect of insulin to stimulate leg blood flow in NIDDM.
Cyanonitrosylmetallates as potential NO-donors. J             Diabetes 1995. 44(2): 221-6.
InorgBiochem 1998. 69(1-2): 121-7.                            120.      Vollenweider P, Tappy L, Randin D, Schneiter P,
106.      Leiper J, Vallance P. Biological significance of    Jequier E, Nicod P, Scherrer U. Differential effects of
endogenous methylarginines that inhibit nitric oxide          hyperinsulinemia and carbohydrate metabolism on
synthases. Cardiovasc Res 1999. 43(3): 542-8.                 sympathetic nerve activity and muscle blood flow in
                                                              humans. J Clin Invest 1993. 92(1): 147-54.
                                                                         Christian Rask Madsen: PhD thesis, Page 35 of 40.

121.      Kuboki K, Jiang ZY, Takahara N, Ha SW,                 dependent diabetes mellitus. J Clin Invest 1996. 97(1): 22-
Igarashi M, Yamauchi T, Feener EP, Herbert TP, Rhodes            8.
CJ, King GL. Regulation of endothelial constitutive nitric
oxide synthase gene expression in endothelial cells and in       136.       Watts GF, O'Brien SF, Silvester W, Millar JA.
vivo: a specific vascular action of insulin. Circulation 2000.   Impaired     endothelium-dependent       and    independent
101(6): 676-81.                                                  dilatation of forearm resistance arteries in men with diet-
                                                                 treated     non-insulin-dependent     diabetes:    role  of
122.      Yki-Jarvinen H, Utriainen T. Insulin-induced           dyslipidaemia. Clin Sci (Cholc) 1996. 91(5): 567-73.
vasodilatation: physiology or pharmacology? Diabetologia
1998. 41(4): 369-79.                                             137.      Hogikyan RV, Galecki AT, Pitt B, Halter JB,
                                                                 Greene DA, Supiano MA. Specific impairment of
123.      Panza JA, Epstein SE, Quyyumi AA. Circadian            endothelium-dependent vasodilation in subjects with type
variation in vascular tone and its relation to alpha-            2 diabetes independent of obesity. J Clin Endocrinol Metab
sympathetic vasoconstrictor activity. NEnglJ Med 1991.           1998. 83(6): 1946-52.
325(14): 986-90.
                                                                 138.      Makimattila S, Liu ML, Vakkilainen J, Schlenzka
124.      Inchiostro S, Bertoli G, Zanette G, Donadon V.         A, Lahdenpera S, Syvanne M, Mantysaari M, Summanen P,
Evidence of higher insulin resistance in NIDDM patients          Bergholm R, Taskinen MR, Yki-Jarvinen H. Impaired
with ischaemic heart disease. Diabetologia 1994. 37(6):          endothelium-dependent vasodilation in type 2 diabetes.
597-603.                                                         Relation to LDL size, oxidized LDL, and antioxidants.
                                                                 Diabetes Care 1999. 22(6): 973-81.
125.      Heitzer T, Yla HS, Luoma J, Kurz S, Munzel T,
Just H, Olschewski M, Drexler H. Cigarette smoking               139.      Chowienczyk PJ, Brett SE, Gopaul NK, Meeking
potentiates endothelial dysfunction of forearm resistance        D, Marchetti M, Russell-Jones DL, Anggard EE, Ritter JM.
vessels in patients with hypercholesterolemia. Role of           Oral treatment with an antioxidant (raxofelast) reduces
oxidized LDL. Circulation 1996. 93(7): 1346-53.                  oxidative stress and improves endothelial function in men
                                                                 with type II diabetes. Diabetologia 2000. 43(8): 974-7.
126.        Laufs U, La F, V, Plutzky J, Liao JK. Upregulation
of endothelial nitric oxide synthase by HMG CoA reductase        140.       Heitzer T, Krohn K, Albers S, Meinertz T.
inhibitors. Circulation 1998. 97(12): 1129-35.                   Tetrahydrobiopterin improves endothelium-dependent
                                                                 vasodilation by increasing nitric oxide activity in patients
127.      Vogel RA, Corretti MC, Plotnick GD. Changes in         with Type II diabetes mellitus. Diabetologia 2000. 43(11):
flow-mediated brachial artery vasoactivity with lowering of      1435-8.
desirable cholesterol levels in healthy middle-aged men.
AmJCardiol 1996. 77(1): 37-40.                                   141.       McNeill KL, Fontana L, Russell-Jones DL, Rajman
                                                                 I, Ritter JM, Chowienczyk PJ. Inhibitory effects of low-
128.      Steinberg HO, Bayazeed B, Hook G, Johnson A,           density lipoproteins from men with type II diabetes on
Cronin J, Baron AD. Endothelial dysfunction is associated        endothelium-dependent relaxation. J AmCollCardiol 2000.
with cholesterol levels in the high normal range in              35(6): 1622-7.
humans. Circulation 1997. 96(10): 3287-93.
                                                                 142.      Butler R, Morris AD, Belch JJ, Hill A, Struthers
129.       Heitzer T, Brockhoff C, Mayer B, Warnholtz A,         AD. Allopurinol normalizes endothelial dysfunction in type
Mollnau H, Henne S, Meinertz T, Munzel T.                        2 diabetics with mild hypertension. Hypertension 2000.
Tetrahydrobiopterin improves endothelium-dependent               35(3): 746-51.
vasodilation in chronic smokers : evidence for a
dysfunctional nitric oxide synthase. CircRes 2000. 86(2):        143.      Enderle MD, Benda N, Schmuelling RM, Haering
E36-E41.                                                         HU, Pfohl M. Preserved endothelial function in IDDM
                                                                 patients, but not in NIDDM patients, compared with
130.      Matthews JN, Altman DG, Campbell MJ, Royston           healthy subjects. Diabetes Care 1998. 21(2): 271-7.
P. Analysis of serial measurements in medical research.
BMJ 1990. 300(6719): 230-5.                                      144.      Balletshofer BM, Rittig K, Enderle MD, Volk A,
                                                                 Maerker E, Jacob S, Matthaei S, Rett K, Haring HU.
131.      Brow H, Prescott R, Applied Mixed Models in            Endothelial    dysfunction   is   detectable    in   young
Medicine. Statistics in Practice series, ed. V. Barnett. 1992,   normotensive first-degree relatives of subjects with type 2
New York: Wiley.                                                 diabetes in association with insulin resistance. Circulation
132.       Matthews DR, Hosker JP, Rudenski AS, Naylor           2000. 101(15): 1780-4.
BA, Treacher DF, Turner RC. Homeostasis model                    145.      Adams MR, Robinson J, McCredie R, Seale JP,
assessment: insulin resistance and beta-cell function from       Sorensen KE, Deanfield JE, Celermajer DS. Smooth muscle
fasting plasma glucose and insulin concentrations in man.        dysfunction     occurs     independently     of    impaired
Diabetologia 1985. 28(7): 412-9.                                 endothelium-dependent dilation in adults at risk of
133.      Avogaro A, Piarulli F, Valerio A, Miola M, Calveri     atherosclerosis. J Am CollCardiol 1998. 32(1): 123-7.
M, Pavan P, Vicini P, Cobelli C, Tiengo A, Calo L, Del PS.       146.       Chan NN, Vallance P, MacAllister RJ, Hingorani
Forearm nitric oxide balance, vascular relaxation, and           AD, Colhoun HM. Impaired nitric oxide response in
glucose metabolism in NIDDM patients. Diabetes 1997.             resistance vessels in human type 1 diabetes in vivo.
46(6): 1040-6.                                                   Circulation 2000. 102(18): II-326-.
134.       Williams SB, Cusco JA,      Roddy MA, Johnstone       147.     Scherrer U, Vollenweider P, Randin D, Jequier E,
MT, Creager MA. Impaired               nitric oxide-mediated     Nicod P, Tappy L. Suppression of insulin-induced
vasodilation in patients with           non-insulin-dependent    sympathetic activation and vasodilation by dexamethasone
diabetes mellitus. J Am Coll Cardiol   1996. 27(3): 567-74.      in humans. Circulation 1993. 88(2): 388-94.
135.    Ting HH, Timimi FK, Boles KS, Creager SJ, Ganz           148.       Hausberg M, Mark AL, Hoffman RP, Sinkey CA,
P, Creager MA. Vitamin C improves endothelium-                   Anderson EA. Dissociation of sympathoexcitatory and
dependent vasodilation in patients with non-insulin-             vasodilator actions of modestly elevated plasma insulin
                                                                 levels. J Hypertens 1995. 13(9): 1015-21.
                                                                        Christian Rask Madsen: PhD thesis, Page 36 of 40.

149.      Bergman RN, Ader M. Free fatty acids and              164.       Zanetti M, Sato J, Katusic ZS, O'Brien T. Gene
pathogenesis of type 2 diabetes mellitus. Trends                transfer of endothelial nitric oxide synthase alters
Endocrinol Metab 2000. 11(9): 351-6.                            endothelium-dependent relaxations in aortas from diabetic
                                                                rabbits. Diabetologia 2000. 43(3): 340-7.
150.       Steinberg HO, Brechtel G, Johnson A, Fineberg
N, Baron AD. Insulin-mediated skeletal muscle vasodilation      165.       Lund DD, Faraci FM, Miller FJ, Jr., Heistad DD.
is nitric oxide dependent. A novel action of insulin to         Gene transfer of endothelial nitric oxide synthase improves
increase nitric oxide release. J Clin Invest 1994. 94(3):       relaxation of carotid arteries from diabetic rabbits.
1172-9.                                                         Circulation 2000. 101(9): 1027-33.
151.       Taddei S, Virdis A, Mattei P, Natali A, Ferrannini   166.      Griendling KK, Sorescu D, Ushio-Fukai M.
E, Salvetti A. Effect of insulin on acetylcholine-induced       NAD(P)H oxidase: role in cardiovascular biology and
vasodilation in normotensive subjects and patients with         disease. Circ Res 2000. 86(5): 494-501.
essential hypertension. Circulation 1995. 92(10): 2911-8.
                                                                167.       Inoguchi T, Li P, Umeda F, Yu HY, Kakimoto M,
152.      Lash JM, Nase GP, Bohlen HG. Acute                    Imamura M, Aoki T, Etoh T, Hashimoto T, Naruse M, Sano
hyperglycemia depresses arteriolar NO formation in              H, Utsumi H, Nawata H. High glucose level and free fatty
skeletal muscle. Am J Physiol 1999. 277(4 Pt 2): H1513-         acid stimulate reactive oxygen species production through
H20.                                                            protein kinase C-dependent activation of NAD(P)H oxidase
                                                                in cultured vascular cells. Diabetes 2000. 49(11): 1939-45.
153.       Vehkavaara S, Makimattila S, Schlenzka A,
Vakkilainen J, Westerbacka J, Yki-Jarvinen H. Insulin           168.      Guzik TJ, West NE, Black E, McDonald D,
therapy improves endothelial function in type 2 diabetes.       Ratnatunga C, Pillai R, Channon KM. Vascular superoxide
Arterioscler Thromb Vasc Biol 2000. 20(2): 545-50.              production by NAD(P)H oxidase: association with
                                                                endothelial dysfunction and clinical risk factors. CircRes
154.      Avena R, Mitchell ME, Nylen ES, Curry KM,             2000. 86(9): E85-E90.
Sidawy AN. Insulin action enhancement normalizes
brachial artery vasoactivity in patients with peripheral        169.      Baynes JW. Role of oxidative stress in
vascular disease and occult diabetes. J Vasc Surg 1998.         development of complications in diabetes. Diabetes 1991.
28(6): 1024-31.                                                 40(4): 405-12.
155.       Tack CJ, Ong MK, Lutterman JA, Smits P.              170.      Williamson JR, Chang K, Frangos M, Hasan KS,
Insulin-induced vasodilatation and endothelial function in      Ido Y, Kawamura T, Nyengaard JR, van den Enden M, Kilo
obesity/insulin resistance. Effects of troglitazone.            C, Tilton RG. Hyperglycemic pseudohypoxia and diabetic
Diabetologia 1998. 41(5): 569-76.                               complications. Diabetes 1993. 42(6): 801-13.
156.      Meredith IT, Currie KE, Anderson TJ, Roddy MA,        171.      Gupte SA, Rupawalla T, Phillibert D, Jr., Wolin
Ganz P, Creager MA. Postischemic vasodilation in human          MS. NADPH and heme redox modulate pulmonary artery
forearm is dependent on endothelium-derived nitric oxide.       relaxation and guanylate cyclase activation by NO. Am J
Am J Physiol 1996. 270(4 Pt 2): H1435-H40.                      Physiol 1999. 277(6 Pt 1): L1124-32.
157.      Tesfamariam B, Cohen RA. Free radicals mediate        172.      Tesfamariam B, Palacino JJ, Weisbrod RM,
endothelial cell dysfunction caused by elevated glucose.        Cohen RA. Aldose reductase inhibition restores endothelial
Am J Physiol 1992. 263(2 Pt 2): H321-H6.                        cell function in diabetic rabbit aorta. J Cardiovasc
                                                                Pharmacol 1993. 21(2): 205-11.
158.      Pieper GM, Meier DA, Hager SR. Endothelial
dysfunction in a model of hyperglycemia and                     173.      Zhang Z, Apse K, Pang J, Stanton RC. High
hyperinsulinemia. Am J Physiol 1995. 269(3 Pt 2): H845-         glucose inhibits glucose-6-phosphate dehydrogenase via
H50.                                                            cAMP in aortic endothelial cells. J Biol Chem 2000.
                                                                275(51): 40042-7.
159.      Ceriello A, dello Russo P, Amstad P, Cerutti P.
High glucose induces antioxidant enzymes in human               174.       Asahina T, Kashiwagi A, Nishio Y, Ikebuchi M,
endothelial cells in culture. Evidence linking hyperglycemia    Harada N, Tanaka Y, Takagi Y, Saeki Y, Kikkawa R,
and oxidative stress. Diabetes 1996. 45(4): 471-7.              Shigeta Y. Impaired activation of glucose oxidation and
                                                                NADPH supply in human endothelial cells exposed to H2O2
160.       Zanetti M, Sato J, Katusic ZS, O'Brien T. Gene       in high-glucose medium. Diabetes 1995. 44(5): 520-6.
transfer of endothelial nitric oxide synthase and
manganese superoxide dismutase improves vascular                175.      King GL, Wakasaki H. Theoretical mechanisms
reactivity of aortas from diabetic rabbits. Circulation 1999.   by which hyperglycemia and insulin resistance could cause
100(18 Suppl): I-624 (abstract).                                cardiovascular diseases in diabetes. Diabetes Care 1999.
                                                                22 Suppl 3: C31-C7.
161.       Williams SB, Goldfine AB, Timimi FK, Ting HH,
Roddy MA, Simonson DC, Creager MA. Acute                        176.      Tesfamariam B, Brown ML, Cohen RA. Elevated
hyperglycemia       attenuates     endothelium-dependent        glucose impairs endothelium-dependent relaxation by
vasodilation in humans in vivo. Circulation 1998. 97(17):       activating protein kinase C. J Clin Invest 1991. 87(5):
1695-701.                                                       1643-8.
162.      Cosentino F, Hishikawa K, Katusic ZS, Luscher         177.      Jack A, Cameron NE, Cotter MA. Treatment with
TF. High glucose increases nitric oxide synthase expression     the protein kinase Cβ inhibitor, LY333531, attenuates the
and superoxide anion generation in human aortic                 development      of   impaired     endothelium-dependent
endothelial cells. Circulation 1997. 96(1): 25-8.               vasodilatation in the mesenteric vasculature of diabetic
                                                                rats. Diabetes 1999: Abstract no. 558.
163.     Ding Y, Vaziri ND, Coulson R, Kamanna VS, Roh
DD. Effects of simulated hyperglycemia, insulin, and            178.      Bohlen HG, Nase GP. Arteriolar nitric oxide
glucagon on endothelial nitric oxide synthase expression.       concentration is decreased during hyperglycemia-induced
Am J Physiol Endocrinol Metab 2000. 279(1): E11-E7.             betaII PKC activation. Am J Physiol Heart Circ Physiol
                                                                2001. 280(2): H621-7.
                                                                         Christian Rask Madsen: PhD thesis, Page 37 of 40.

179.       Ha S, Park J, Kuboki K, Yamauchi T, King G.          193.       Schroeder CA, Jr., Chen YL, Messina EJ.
Inhibition of protein kinase C (PKC) normalized endothelial     Inhibition of NO synthesis or endothelium removal reveals
nitric oxide synthase (eNOS) expression in vascular tissues     a vasoconstrictor effect of insulin on isolated arterioles. Am
of obese zucker rats. Diabetes 1999. 48(Suppl. 1): A270.        J Physiol 1999. 276(3 Pt 2): H815-H20.
180.      Kunisaki M, Bursell SE, Umeda F, Nawata H,            194.      Zemel MB, Reddy S, Sowers JR. Insulin
King GL. Normalization of diacylglycerol-protein kinase C       attenuation of vasoconstrictor responses to phenylephrine
activation by vitamin E in aorta of diabetic rats and           in Zucker lean and obese rats. Am J Hypertens 1991. 4(6):
cultured rat smooth muscle cells exposed to elevated            537-9.
glucose levels. Diabetes 1994. 43(11): 1372-7.
                                                                195.     Baron AD, Brechtel-Hook G, Johnson A, Cronin J,
181.      Tasinato A, Boscoboinik D, Bartoli GM, Maroni P,      Leaming R, Steinberg HO. Effect of perfusion rate on the
Azzi A. d-alpha-tocopherol inhibition of vascular smooth        time course of insulin-mediated skeletal muscle glucose
muscle cell proliferation occurs at physiological               uptake. Am J Physiol 1996. 271(6 Pt 1): E1067-E72.
concentrations, correlates with protein kinase C inhibition,
and is independent of its antioxidant properties. Proc Natl     196.       Scherrer U, Randin D, Vollenweider P,
Acad Sci U S A 1995. 92(26): 12190-4.                           Vollenweider L, Nicod P. Nitric oxide release accounts for
                                                                insulin's vascular effects in humans. J Clin Invest 1994.
182.       Bucala R, Tracey KJ, Cerami A. Advanced              94(6): 2511-5.
glycosylation products quench nitric oxide and mediate
defective    endothelium-dependent      vasodilatation   in     197.       Oltman CL, Kane NL, Gutterman DD, Bar RS,
experimental diabetes. J Clin Invest 1991. 87(2): 432-8.        Dellsperger KC. Mechanism of coronary vasodilation to
                                                                insulin and insulin-like growth factor I is dependent on
183.       Yamagishi S, Fujimori H, Yonekura H, Yamamoto        vessel size. Am J Physiol Endocrinol Metab 2000. 279(1):
Y, Yamamoto H. Advanced glycation endproducts inhibit           E176-81.
prostacyclin production and induce plasminogen activator
inhibitor-1 in human microvascular endothelial cells.           198.      Yagi S, Takata S, Kiyokawa H, Yamamoto M,
Diabetologia 1998. 41(12): 1435-41.                             Noto Y, Ikeda T, Hattori N. Effects of insulin on
                                                                vasoconstrictive responses to norepinephrine and
184.       Rojas A, Romay S, Gonzalez D, Herrera B,             angiotensin II in rabbit femoral artery and vein. Diabetes
Delgado R, Otero K. Regulation of endothelial nitric oxide      1988. 37(8): 1064-7.
synthase expression by albumin- derived advanced
glycosylation end products. CircRes 2000. 86(3): e50-4.         199.      Randin D, Vollenweider P, Tappy L, Jequier E,
                                                                Nicod P, Scherrer U. Effects of adrenergic and cholinergic
185.      Ito A, Asagami T, Tsao PS, Adimoolam S, Kimoto        blockade on insulin-induced stimulation of calf blood flow
M, Tsuji H, Reaven GM, Cooke JP. Dysregulation of               in humans. Am J Physiol 1994. 266(3 Pt 2): R809-R16.
dimethylarginine dimethylaminohydrolase. A mechanism of
endothelial dysfunction in diabetes mellitus. Circulation       200.      Creager MA, Liang CS, Coffman JD. Beta
1999. 100(18 Suppl): I-473 (abstract).                          adrenergic-mediated vasodilator response to insulin in the
                                                                human forearm. J Pharmacol Exp Ther 1985. 235(3): 709-
186.       Raitakari M, Nuutila P, Ruotsalainen U, Laine H,     14.
Teras M, Iida H, Makimattila S, Utriainen T, Oikonen V,
Sipila H, Haaparanta M, Solin O, Wegelius U, Knuuti J, Yki-     201.       Cardillo C, Nambi SS, Kilcoyne CM, Choucair WK,
Jarvinen H. Evidence for dissociation of insulin stimulation    Katz A, Quon MJ, Panza JA. Insulin stimulates both
of blood flow and glucose uptake in human skeletal              endothelin and nitric oxide activity in the human forearm.
muscle: studies using [15O]H2O, [18F]fluoro-2-deoxy-D-          Circulation 1999. 100(8): 820-5.
glucose, and positron emission tomography. Diabetes             202.      Guo JP, Murohara T, Buerke M, Scalia R, Lefer
1996. 45(11): 1471-7.                                           AM. Direct measurement of nitric oxide release from
187.      Chaudhuri A, Kanjwal Y, Mohanty P, Rao S,             vascular endothelial cells. J Appl Physiol 1996. 81(2): 774-
Sung BH, Wilson MF, Dandona P. Insulin-induced                  9.
vasodilatation of internal carotid artery. Metabolism 1999.     203.        Zeng G, Nystrom FH, Ravichandran LV, Cong LN,
48(11): 1470-3.                                                 Kirby M, Mostowski H, Quon MJ. Roles for insulin receptor,
188.     Jern S. Effects of insulin on vascular responses       PI3-kinase, and Akt in insulin-signaling pathways related to
to mental stress and norepinephrine in human forearm.           production of nitric oxide in human vascular endothelial
Hypertension 1994. 24(6): 686-94.                               cells. Circulation 2000. 101(13): 1539-45.

189.     McNulty PH, Pfau S, Deckelbaum LI. Effect of           204.        Oliver FJ, de la Rubia G, Feener EP, Lee ME,
plasma insulin level on myocardial blood flow and its           Loeken MR, Shiba T, Quertermous T, King GL. Stimulation
mechanism of action. Am J Cardiol 2000. 85(2): 161-5.           of endothelin-1 gene expression by insulin in endothelial
                                                                cells. J Biol Chem 1991. 266(34): 23251-6.
190.       Grover A, Padginton C, Wilson MF, Sung BH,
Izzo JL, Jr., Dandona P. Insulin attenuates norepinephrine-     205.      Hattori Y, Kasai K, Nakamura T, Emoto T,
induced venoconstriction. An ultrasonographic study.            Shimoda S. Effect of glucose and insulin on
Hypertension 1995. 25(4 Pt 2): 779-84.                          immunoreactive endothelin-1 release from cultured porcine
                                                                aortic endothelial cells. Metabolism 1991. 40(2): 165-9.
191.       Feldman RD, Bierbrier GS. Insulin-mediated
vasodilation: impairment with increased blood pressure          206.     Anfossi G, Cavalot F, Massucco P, Mattiello L,
and body mass. Lancet 1993. 342(8873): 707-9.                   Mularoni E, Hahn A, Trovati M. Insulin influences
                                                                immunoreactive endothelin release by human vascular
192.       Chen YL, Messina EJ. Dilation of isolated skeletal   smooth muscle cells. Metabolism 1993. 42(9): 1081-3.
muscle arterioles by insulin is endothelium dependent and
nitric oxide mediated. Am J Physiol 1996. 270(6 Pt 2):          207.      Govers R, Rabelink T. Cellular regulation of
H2120-H4.                                                       endothelial nitric oxide synthase. Am J Physiol 2001. 280:
                                                                        Christian Rask Madsen: PhD thesis, Page 38 of 40.

208.      Rask-Madsen C, Ihlemann N, Krarup T,                  different mechanisms and different effects on insulin
Christiansen E, Kober L, Kistorp CN, Torp-Pedersen C.           signaling. J Clin Invest 1996. 97(6): 1471-7.
Insulin therapy improves insulin-stimulated endothelial
function in patients with type 2 diabetes and ischemic          223.      Hopfner RL, Gopalakrishnan V. Endothelin:
heart disease. Diabetes 2001;50(11):2611-8.                     emerging role in diabetic vascular complications.
                                                                Diabetologia 1999. 42(12): 1383-94.
209.      Luo Z, Fujio Y, Kureishi Y, Rudic RD, Daumerie
G, Fulton D, Sessa WC, Walsh K. Acute modulation of             224.       Wu-Wong JR, Berg CE, Wang J, Chiou WJ, Fissel
endothelial Akt/PKB activity alters nitric oxide-dependent      B. Endothelin stimulates glucose uptake and GLUT4
vasomotor activity in vivo. J Clin Invest 2000. 106(4): 493-    translocation via activation of endothelin ETA receptor in
9.                                                              3T3-L1 adipocytes. J Biol Chem 1999. 274(12): 8103-10.

210.        Fleming I, Fisslthaler B, Busse R. Shear stress     225.      Imamura T, Ishibashi K, Dalle S, Ugi S, Olefsky
elicits the association of phosphatidylinositol 3-kinase with   JM. Endothelin-1-induced GLUT4 translocation is mediated
PECAM-1 in cultured human endothelial cells. Circulation        via Galpha(q/11) protein and phosphatidylinositol 3-kinase
1999. 100(18 Suppl): I-338.                                     in 3T3-L1 adipocytes. J Biol Chem 1999. 274(47): 33691-
211.      Kim F, Gallis BM, Corson MA. IRS-1
phosphorylation and endothelial NO production: reciprocal       226.      Chou YC, Perng JC, Juan CC, Jang SY, Kwok CF,
regulation by fluid shear stress and TNF-α. Circulation         Chen WL, Fong JC, Ho LT. Endothelin-1 inhibits insulin-
1999. 100(18 Suppl): I-338 (abstract).                          stimulated glucose uptake in isolated rat adipocytes.
                                                                Biochem Biophys Res Commun 1994. 202(2): 688-93.
212.       Fleming I, Busse R. Signal transduction of eNOS
activation. Cardiovasc Res 1999. 43(3): 532-41.                 227.       Juan CC, Fang VS, Huang YJ, Kwok CF, Hsu YP,
                                                                Ho LT. Endothelin-1 induces insulin resistance in conscious
213.       McCabe TJ, Fulton D, Roman LJ, Sessa WC.             rats. Biochem Biophys Res Commun 1996. 227(3): 694-9.
Enhanced electron flux and reduced calmodulin
dissociation may explain "calcium-independent" eNOS             228.       Ottosson-Seeberger A, Lundberg JM, Alvestrand
activation by phosphorylation. J Biol Chem 2000. 275(9):        A, Ahlborg G. Exogenous endothelin-1 causes peripheral
6123-8.                                                         insulin resistance in healthy humans. Acta Physiol Scand
                                                                1997. 161(2): 211-20.
214.      Aljada A, Dandona P. Effect of insulin on human
aortic endothelial nitric oxide synthase. Metabolism 2000.      229.      Jiang ZY, Zhou QL, Chatterjee A, Feener EP,
49(2): 147-50.                                                  Myers MG, Jr., White MF, King GL. Endothelin-1 modulates
                                                                insulin signaling through phosphatidylinositol 3-kinase
215.       Davda RK, Stepniakowski KT, Lu G, Ullian ME,         pathway in vascular smooth muscle cells. Diabetes 1999.
Goodfriend TL, Egan BM. Oleic acid inhibits endothelial         48(5): 1120-30.
nitric oxide synthase by a protein kinase C-independent
mechanism. Hypertension 1995. 26(5): 764-70.                    230.     Kahn AM, Seidel CL, Allen JC, O'Neil RG, Shelat
                                                                H, Song T. Insulin reduces contraction and intracellular
216.       Steinberg HO, Tarshoby M, Monestel R, Hook G,        calcium concentration in vascular smooth muscle.
Cronin J, Johnson A, Bayazeed B, Baron AD. Elevated             Hypertension 1993. 22(5): 735-42.
circulating free fatty acid levels impair endothelium-
dependent vasodilation. J Clin Invest 1997. 100(5): 1230-       231.       Kahn AM, Lichtenberg RA, Allen JC, Seidel CL,
9.                                                              Song T. Insulin-stimulated glucose transport inhibits Ca2+
                                                                influx and contraction in vascular smooth muscle.
217.      Steinberg HO, Paradisi G, Hook G, Crowder K,          Circulation 1995. 92(6): 1597-603.
Cronin J, Baron AD. Free fatty acid elevation impairs
insulin-mediated vasodilation and nitric oxide production.      232.      Kahn AM, Husid A, Allen JC, Seidel CL, Song T.
Diabetes 2000. 49(7): 1231-8.                                   Insulin acutely inhibits cultured vascular smooth muscle
                                                                cell contraction by a nitric oxide synthase-dependent
218.       Vollenweider L, Tappy L, Owlya R, Jequier E,         pathway. Hypertension 1997. 30(4): 928-33.
Nicod P, Scherrer U. Insulin-induced sympathetic
activation and vasodilation in skeletal muscle. Effects of      233.      Kahn AM, Husid A, Odebunmi T, Allen JC, Seidel
insulin resistance in lean subjects. Diabetes 1995. 44(6):      CL, Song T. Insulin inhibits vascular smooth muscle
641-5.                                                          contraction at a site distal to intracellular Ca2+
                                                                concentration. Am J Physiol 1998. 274(5 Pt 1): E885-92.
219.       Hermann C, Assmus B, Urbich C, Zeiher AM,
Dimmeler S. Insulin-mediated stimulation of protein kinase      234.      Kahn AM, Allen JC, Seidel CL, Lichtenberg DS,
Akt: A potent survival signaling cascade for endothelial        Song T, Zhang S. Insulin increases NO-stimulated
cells. Arterioscler Thromb Vasc Biol 2000. 20(2): 402-9.        guanylate cyclase activity in cultured VSMC while raising
                                                                redox potential. Am J Physiol EndocrinolMetab 2000.
220.       Rui L, Aguirre V, Kim JK, Shulman GI, Lee A,         278(4): E627-E33.
Corbould A, Dunaif A, White MF. Insulin/IGF-1 and TNF-α
stimulate phosphorylation of IRS-1 at inhibitory Ser 307        235.      Kahn AM, Allen JC, Seidel CL, Song T. Protein
via distinct pathways. J Clin Invest 2001. 107: 181-9.          kinase C mediates insulin-inhibited Ca2+ transport and
                                                                contraction of vascular smooth muscle. Am J Hypertens
221.      Bernardini    B,   Liao  JK.   Activation     of      2000. 13(4 Pt 1): 383-8.
phosphatidylinositol 3-kinase and endothelial nitric oxide
synthase by insulin is inhibited by moderate                    236.      Trovati M, Massucco P, Mattiello L, Cavalot F,
hyperglycemia. Circulation 2000. 102(18): Abstract              Mularoni E, Hahn A, Anfossi G. Insulin increases cyclic
publishing ID no. 67.                                           nucleotide content in human vascular smooth muscle cells:
                                                                a mechanism potentially involved in insulin-induced
222.      Kroder G, Bossenmaier B, Kellerer M, Capp E,          modulation of vascular tone. Diabetologia 1995. 38(8):
Stoyanov B, Muhlhofer A, Berti L, Horikoshi H, Ullrich A,       936-41.
Haring    H.    Tumor    necrosis   factor-alpha-     and
hyperglycemia-induced insulin resistance. Evidence for          237.     Trovati M, Massucco P, Mattiello L, Cavalot F,
                                                                Mularoni EM, Hahn AW, Anfossi G. Studies on the
                                                                           Christian Rask Madsen: PhD thesis, Page 39 of 40.

influence of insulin on cyclic adenosine monophosphate in         Mandarino LJ. Insulin resistance differentially affects the PI
human vascular smooth muscle cells: dependence on                 3-kinase- and MAP kinase-mediated signaling in human
cyclic guanosine monophosphate and modulation of                  muscle. J Clin Invest 2000. 105(3): 311-20.
catecholamine effects. Diabetologia 1996. 39(10): 1156-
64.                                                               252.       Golovchenko I, Goalstone ML, Watson P,
                                                                  Brownlee M, Draznin B. Hyperinsulinemia enhances
238.       Trovati M, Massucco P, Mattiello L, Costamagna         transcriptional activity of nuclear factor- kappaB induced
C, Aldieri E, Cavalot F, Anfossi G, Bosia A, Ghigo D. Human       by angiotensin II, hyperglycemia, and advanced
vascular smooth muscle cells express a constitutive nitric        glycosylation end products in vascular smooth muscle cells
oxide synthase that insulin rapidly activates, thus               [In Process Citation]. Circ Res 2000. 87(9): 746-52.
increasing guanosine 3':5'-cyclic monophosphate and
adenosine 3':5'-cyclic monophosphate concentrations.              253.      Krook A, Kawano Y, Song XM, Efendic S, Roth
Diabetologia 1999. 42(7): 831-9.                                  RA, Wallberg-Henriksson H, Zierath JR. Improved glucose
                                                                  tolerance restores insulin-stimulated Akt kinase activity
239.      Iesaki T, Gupte SA, Wolin MS. A flavoprotein            and glucose transport in skeletal muscle from diabetic
mechanism appears to prevent an oxygen-dependent                  Goto-Kakizaki rats. Diabetes 1997. 46(12): 2110-4.
inhibition of cGMP-associated nitric oxide-elicited
relaxation of bovine coronary arteries. CircRes 1999.             254.       Chibalin AV, Yu M, Ryder JW, Song XM, Galuska
85(11): 1027-31.                                                  D, Krook A, Wallberg-Henriksson H, Zierath JR. Exercise-
                                                                  induced changes in expression and activity of proteins
240.     Gupte SA, Rupawalla T, Mohazzab HK, Wolin                involved in insulin signal transduction in skeletal muscle:
MS. Regulation of NO-elicited pulmonary artery relaxation         differential effects on insulin-receptor substrates 1 and 2.
and guanylate cyclase activation by NADH oxidase and              Proc Natl Acad Sci U S A 2000. 97(1): 38-43.
SOD. Am J Physiol 1999. 276(5 Pt 2): H1535-42.
                                                                  255.      Giorgino F, Logoluso F, Davalli AM, Napoli R,
241.      Boulanger CM, Heymes C, Benessiano J, Geske             Laviola L, Hirshman MF, Horton ES, Weir GC, Smith RJ.
RS, Levy BI, Vanhoutte PM. Neuronal nitric oxide synthase         Islet transplantation restores normal levels of insulin
is expressed in rat vascular smooth muscle cells: activation      receptor and substrate tyrosine phosphorylation and
by angiotensin II in hypertension. Circ Res 1998. 83(12):         phosphatidylinositol 3-kinase activity in skeletal muscle
1271-8.                                                           and myocardium of streptozocin-induced diabetic rats.
                                                                  Diabetes 1999. 48(4): 801-12.
242.       Pieper GM. Divergent actions of chronic insulin
treatment in vivo versus acute treatment ex vivo on               256.      Markuns JF, Napoli R, Hirshman MF, Davalli AM,
diabetic-induced endothelial dysfunction. Life Sci 1997.          Cheatham B, Goodyear LJ. Effects of streptozocin-induced
60(25): L371-L6.                                                  diabetes and islet cell transplantation on insulin signaling
                                                                  in rat skeletal muscle. Endocrinology 1999. 140(1): 106-
243.      Shinozaki K, Kashiwagi A, Nishio Y, Okamura T,          11.
Yoshida Y, Masada M, Toda N, Kikkawa R. Abnormal
biopterin metabolism is a major cause of impaired                 257.       Yki-Jarvinen H, Makimattila S. Insulin resistance
endothelium-dependent relaxation through nitric oxide/O2-         due to hyperglycaemia: an adaptation protecting insulin-
imbalance in insulin-resistant rat aorta. Diabetes 1999.          sensitive tissues. Diabetologia 1997. 40 Suppl 2: S141-S4.
48(12): 2437-45.
                                                                  258.      Mitrakou A, Kelley D, Veneman T, Jenssen T,
244.       Pessin JE, Saltiel AR. Signaling pathways in           Pangburn T, Reilly J, Gerich J. Contribution of abnormal
insulin action: molecular targets of insulin resistance. J Clin   muscle and liver glucose metabolism to postprandial
Invest 2000. 106(2): 165-9.                                       hyperglycemia in NIDDM. Diabetes 1990. 39(11): 1381-90.
245.        Desfaits AC, Serri O, Renier G. Normalization of      259.      Marshall S, Bacote V, Traxinger RR. Discovery of
plasma lipid peroxides, monocyte adhesion, and tumor              a    metabolic     pathway   mediating    glucose-induced
necrosis factor-alpha production in NIDDM patients after          desensitization of the glucose transport system. Role of
gliclazide treatment. Diabetes Care 1998. 21(4): 487-93.          hexosamine biosynthesis in the induction of insulin
                                                                  resistance. J Biol Chem 1991. 266(8): 4706-12.
246.       Morohoshi M, Fujisawa K, Uchimura I, Numano
F. Glucose-dependent interleukin 6 and tumor necrosis             260.      Kaiser N, Sasson S, Feener EP, Boukobza-Vardi
factor production by human peripheral blood monocytes in          N, Higashi S, Moller DE, Davidheiser S, Przybylski RJ, King
vitro. Diabetes 1996. 45(7): 954-9.                               GL. Differential regulation of glucose transport and
                                                                  transporters by glucose in vascular endothelial and smooth
247.       Peraldi   P,     Xu    M,     Spiegelman      BM.      muscle cells. Diabetes 1993. 42(1): 80-9.
Thiazolidinediones block tumor necrosis factor-alpha-
induced inhibition of insulin signaling. J Clin Invest 1997.      261.      Patterson C, Madamanchi NR, Runge MS. The
100(7): 1863-9.                                                   oxidative paradox: another piece in the puzzle. CircRes
                                                                  2000. 87(12): 1074-6.
248.     Baron AD. The coupling of glucose metabolism
and perfusion in human skeletal muscle. The potential role        262.      Ohara Y, Peterson TE, Harrison DG.
of endothelium-derived nitric oxide. Diabetes 1996. 45            Hypercholesterolemia increases endothelial superoxide
Suppl 1.                                                          anion production. J Clin Invest 1993. 91(6): 2546-51.
249.       Steinberg HO, Baron AD. Insulin-mediated               263.      Mayer B, John M, Bohme E. Purification of a
vasodilation: why one's physiology could be the other's           Ca2+/calmodulin-dependent nitric oxide synthase from
pharmacology. Diabetologia 1999. 42(4): 493-4.                    porcine cerebellum. Cofactor-role of tetrahydrobiopterin.
                                                                  FEBS Lett 1990. 277(1-2): 215-9.
250.      Yki-Jarvinen H, Utriainen T. Insulin-mediated
vasodilatation: the authors' reply. Diabetologia 1999.            264.      Wever R, Luscher TF, Cosentino F, Rabelink T.
42(4): 494-5.                                                     Atherosclerosis and the two faces of endothelial nitric
                                                                  oxide synthase. Circulation 1998. 97: 108-12.
251.     Cusi K, Maezono K, Osman A, Pendergrass M,
Patti ME, Pratipanawatr T, DeFronzo RA, Kahn CR,
                                                               Christian Rask Madsen: PhD thesis, Page 40 of 40.

265.      Hingorani AD, Cross J, Kharbanda RK, Mullen
MJ, Bhagat K, Taylor M, Donald AE, Palacios M, Griffin GE,
Deanfield JE, MacAllister RJ, Vallance P. Acute systemic
inflammation impairs endothelium-dependent dilatation in
humans. Circulation 2000. 102(9): 994-9.
266.     Meier CR, Jick SS, Derby LE, Vasilakis C, Jick H.
Acute respiratory-tract infections and risk of first-time
acute myocardial infarction. Lancet 1998. 351(9114):
267.      Ross R. Atherosclerosis - an inflammatory
disease. NEnglJ Med 1999. 340(2): 115-26.
268.      Hotamisligil GS. The role of TNFα and TNF
receptors in obesity and insulin resistance. J InternMed
1999. 245(6): 621-5.
269.      Reaven GM. Banting lecture 1988. Role of insulin
resistance in human disease. Diabetes 1988. 37(12):
270.      Godfrey KM, Barker DJ. Fetal nutrition and adult
disease. AmJ Clin Nutr 2000. 71(5 Suppl): 1344S-52S.
271.      Poulsen P, Vaag AA, Kyvik KO, Moller JD, Beck-
Nielsen H. Low birth weight is associated with NIDDM in
discordant monozygotic and dizygotic twin pairs.
Diabetologia 1997. 40(4): 439-46.
272.      Martin BC, Warram JH, Krolewski AS, Bergman
RN, Soeldner JS, Kahn CR. Role of glucose and insulin
resistance in development of type 2 diabetes mellitus:
results of a 25-year follow-up study. Lancet 1992.
340(8825): 925-9.
273.       Griffin ME, Marcucci MJ, Cline GW, Bell K,
Barucci N, Lee D, Goodyear LJ, Kraegen EW, White MF,
Shulman GI. Free fatty acid-induced insulin resistance is
associated with activation of protein kinase C theta and
alterations in the insulin signaling cascade. Diabetes 1999.
48(6): 1270-4.
274.       Khan WA, Blobe GC, Hannun YA. Arachidonic
acid and free fatty acids as second messengers and the
role of protein kinase C. Cell Signal 1995. 7(3): 171-84.

Table 1. Clinical characteristics of patients and healthy controls.

                                                                Healthy      Healthy
                                                                controls    controls*

 n                                               28                   31       10

 Age [years]                                    63 ± 2           58 ± 2      56 ± 2

 Sex (male/female)                              26/2              25/6        10/0

 Body mass index [kg m−2]                     32.0 ± 1.1       22.9 ± 0.4   23.7 ± 0.6

 Hemoglobin A1c [%]                           9.8 ± 1.8               .         .

 Fasting blood glucose [mmol l−1]             14.2 ± 3.3       5.1 ± 0.1    5.3 ± 0.1

 Fasting serum insulin [mU l−1]               16.2 ± 7.8       5.4 ± 0.3    6.6 ± 0.5

 Systolic blood pressure [mm Hg]               142 ± 5          117 ± 4      117 ± 5

 Diastolic blood pressure [mm Hg]              71 ± 2            63 ± 1      61 ± 2

 Plasma cholesterol [mmol l−1]                4.6 ± 0.1        4.9 ± 0.1    5.3 ± 0.2

 Plasma triglycerides [mmol l−1]              2.1 ± 0.2        0.8 ± 0.1    1.0 ± 0.1

*in whom the insulin-stimulated acetylcholine response was studied.
Table 2. Clinical characteristics of patients in the two treatment groups.

                                          Treatment group
                                                               Time control    p

 n                                               19                 9          .

 Age [years]                                   63 ± 2            63 ± 3       n.s.

 Sex (male/female)                              17/2              10/0        n.s.

 Body mass index [kg m−2]                    31.4 ± 1.2         33.4 ± 2.2    n.s.

 Smokers                                      3 (11%)           1 (11%)       n.s.

 Microalbuminuria                           4/19 (21%)         4/9 (44%)      n.s.

 Oral hypoglycemic drugs                    16/19 (84%)        5/9 (56%)      n.s.

 Statin therapy                             15/19 (79%)        9/9 (100%)     n.s.

 ACEI therapy                               4/19 (21%)         1/9 (11%)      n.s.

 Long-acting nitrates                       3/19 (16%)         1/9 (11%)      n.s.

 Calcium blockers                           6/19 (32%)         2/9 (22%)      n.s.

Abbreviations: ACEI: angiotensin converting enzyme inhibitor
Table 3. Serum insulin and blood glucose during insulin stimulation in healthy controls and in patients at the initial examination.

                                    Treatment group                        Time control group                      Healthy control group

                               Before             During                Before             During                Before             During

 Local insulin               13.3 ± 1.4        121.2 ± 19.5           13.9 ± 2.3        154.2 ± 20.4           5.4 ± 0.5           105 ± 16

 Systemic insulin            13.2 ± 1.44        16.2 ± 1.4            14.0 ± 2.4         17.7 ± 2.3            5.5 ± 0.4          6.8 ± 0.5

 Local glucose               13.2 ± 0.9         12.7 ± 0.8            11.5 ± 0.8         10.7 ± 0.8            4.9 ± 0.1          4.4 ± 0.1

 Systemic glucose            12.8 ± 0.8         12.5 ± 0.8            11.4 ± 0.8         10.9 ± 0.8            5.1 ± 0.2          4.6 ± 0.2

The heading “Insulin” represents the concentration of insulin in serum, expressed in units of mU l−1. The heading “Glucose” represents the
concentration of glucose in blood, expressed in units of mmol l−1. The heading “Local” represents the arm where insulin is infused.
The heading “Before” represents concentrations just before insulin infusion. The heading “During” represents concentrations 20 minutes after
the start of insulin infusion.
Table 4. Clinical parameters for patients according to treatment group and for healthy controls.

                                                             Treatment group                    Time control group
                                                             Examination day:                    Examination day:       Healthy
                                                                   Early          Late                        Late       group
                                                 Initial                                       Initial
                                                                  repeat         repeat                      repeat

 n                                                 19               9              19            9              8          31

 Body mass [kg]                                  94 ± 4             .            98 ± 5       104 ± 8        104 ± 8     74 ± 2

 Fasting blood glucose [mmol l−1]              14.7 ± 0.9       8.3 ± 0.6       7.5 ± 0.2    13.1 ± 0.9    13.0 ± 1.0   5.1 ± 0.1

 Hemoglobin A1c [%]                            10.0 ± 0.4           .           7.5 ± 0.2    9.5 ± 0.5     10.0 ± 0.6       .

 Fasting serum insulin [mU l−1]                15.4 ± 1.4       17.1 ± 3.1      19.4 ± 2.4   16.2 ± 2.9    14.7 ± 3.3   5.5 ± 0.3

 Systolic blood pressure [mm Hg]                142 ± 5          138 ± 6         148 ± 6     143 ± 11       145 ± 10    117 ± 5

 Diastolic blood pressure [mm Hg]                71 ± 2           64 ± 3         76 ± 3        71 ± 3        75 ± 3      63 ± 1

 Plasma cholesterol [mmol l−1]                 4.6 ± 0.2            .           4.5 ± 0.2    4.6 ± 0.3      4.8 ± 0.4   4.9 ± 0.1

 Plasma HDL cholesterol [mmol l−1]             1.1 ± 0.1            .           1.3 ± 0.1    1.0 ± 0.1      1.1 ± 0.1   1.3 ± 0.1

 Plasma triglycerides [mmol l−1]               2.0 ± 0.3            .           1.4 ± 0.2    2.4 ± 0.5      1.6 ± 0.5   0.8 ± 0.1

Abbreviations: Hg: mercury; HDL: high density lipoprotein.
Table 5a. Previous studies of vasodilation during systemic hyperinsulinemia in lean, healthy subjects.

First author          Year     See       Age            BMI                    Vascular    Blood flow             Insulin level [mU l−1]   Duration [minutes]      BF increase [%]
                               table                                           bed         measurement

Laakso                1990     5a        33 ± 2         (weight 68 ± 2 kg)     leg         thermodilution          21, 49, 3492*           180-550                 −4, +37, +65a
(Baron) [1]

Laakso                1990     5a        35 ± 3         (weight 67 ± 2 kg)     leg         thermodilution          524*                    60, 120, 180            +36, +44, +72
(Baron) [1]

Kelley [2]            1990     .         54 ± 2         27.1± 1.0              leg         impedance              77                       240                     no

Anderson [3]          1991     .         20.7 ± 0.3     25.1 ± 0.1             forearm     plethysmography         72, 144                 60, 120                 +55, 65b

Bak [4]               1992               49 ± 1         27.3 ± 1.0             forearm     plethysmography         231                     180                     +4 (NS)

Bonadonna             1993     .         23 ± 1         lean                   forearm     indocyanine dye        30, 70, 498, 1587*c      130*                    (not available, NS), 16 (NS), 23,
(Ferrannini) [5]                                                                           dilution                                                                25

Vollenweider          1993     .         28± 4          21.8 ± 0.8             calf        plethysmography         63                      120                     35
(Scherrer) [6]

Utriainen (Yki-       1995     .         24 ± 1         22.2 ± 0.7             forearm     plethysmography         64-484                  120−360                 +15-115
Jarvinen) [7]

Dela [8]              1995     .         59 ± 1         25.7 ± 0.8             leg         thermodilution          416,770, 15936          120-360                 19, 26, 50d

Abbreviations: not statistically significant (NS). Blood flow increase at highest dose of agonist or NO donor (∆BF).
Certain studies have different protocols for insulin infusion that fits the catagories of more than one of the tables 5a-c. In those cases, cross-reference to additional notes have been made in the
column “See table”.
* Value selected from several similar or calculated from data presented in the article.
    Read from figure 2.
    Read from figure 5.
    These doses administered on different days.
    Read from figure 1.
Table 5b. Previous studies of vasodilation during local hyperinsulinemia in lean, healthy subjects.

First author                   Year        See table      Age            BMI                    Vascular      Insulin level   Duration         ∆BF [%]
                                                                                                bed           [mU l−1]        [min]

Creager [9]                    1985        .                                                    forearm       .a              15               + 55

Natali (Ferrannini) [10]       1990        .              23 ± 3         22.4 ± 1.5             forearm       125             100              0b (NS)

Lembo [11]                     1993        .              28 ± 3         lean                   forearm       63*             30               + 5 (NS)

Steinberg (Baron) [15]         1994        5c             36 ± 3         25.9 ± 1.4             leg           50              20               + 49

Jern [12]                      1994        .              28 (range:     22.7 (range: 19.6-     forearm       98              90               + 81 (+ 61 in
                                                          22-36)         25.5)                                                                 control arm)

Taddei (Ferrannini) [13]       1995        5c             47 ± 6         23.1 ± 1.4             forearm       48              20               + 3 (NS)

Cardillo [14]                  1998        .              52 ± 2         24.6 ± 0.2             forearm       231             120              − 8c (NS)

Present study                  2001        5c             56 ± 2         23.7 ± 0.6             forearm       133             20               + 16 (NS)

Abbreviations: not statistically significant (NS). Blood flow increase at highest dose of agonist or NO donor (∆BF).
Certain studies have different protocols for insulin infusion that fits the catagories of more than one of the tables 5a-c. In those cases, cross-reference to additional notes have been made in the
column “See table”.
* Value selected from several similar or calculated from data presented in the article.
    Dose 0.1−1 mU kg−1 min−1
    Read from figure 1.
    Read from figure 1.
Table 5c. Previous studies of vasodilation to endothelium-dependent and -independet agonists during local or systemic hyperinsulinemia.

First author           Year     See       Age            Phenotype     BMI            Vascular      Insulin          Insulin     Duration        Agonist & dose     ∆BF [%]     NO donor &         ∆BF [%]
                                table                                                 bed           administration   level       pre-agonist     [µg min−1]                     dose [µg
                                                                                                                     [mU l−1]    [min]                                          min−1]

Steinberg              1994     5b        38 ± 1a        healthy       28.2 ± 1.4a    leg           systemic         43a         180             MCh 2.5-10         + 20        SNP 1.8-7.0        + 19
(Baron) [15]                                                                                                                                                                                       (NS)

Taddei                 1995     5b        47 ± 6         healthy       23.1 ± 1.4     forearm       Intra-brachial   48          20              ACh 1.6 -165*      + 41        SNP 11-44          − 2 (NS)
(Ferrannini) [13]

Present study          2001     5b & c    56 ± 2         healthy       23.7 ± 0.6     forearm       Intra-brachial   105         20              ACh 3.8-60         + 106       .                  .

Steinberg              1996     .         34 ± 2a        obese         33.8 ± 1.6a    leg           systemic         257a        200             MCh 2.5-12.5       − 26b       SNP 1.8-7.0        − 60c
(Baron) [16]                                                                                                                                                        (NS)                           (NS)

Steinberg              1996     .         40 ± 3a        type 2        36.0 ± 3.4a    leg           systemic         1023a       200             MCh 2.5-12.5       − 33b       SNP 1.8-7.0        − 44c
(Baron) [16]                                             diabetes                                                                                                   (NS)                           (NS)

Present study          2001     5b & c    56 ± 2         type 2        32.0 ± 1.1     forearm       Intra-brachial   133         20              ACh 3.8-60         + 15        .                  .
                                                         diabetes                                                                                                   (NS)

Abbreviations: metacholine (MCh), acetylcholine (ACh), not statistically significant (NS). Blood flow increase at highest dose of agonist or NO donor (∆BF).
”Agonist” designates endothelium-depent agonist. For aid in identifying studies from the same laboratory, the name of a key researcher (often the senior researcher) is mentioned in parenthesis
after the first author.
Certain studies have different protocols for insulin infusion that fits the catagories of more than one of the tables 5a-c. In those cases, cross-reference to additional notes have been made in the
column “See table”.
    Values are from studies with metacholine, but are similar to values for sodium nitroprusside studies.
    Read from figure 1 and 2.
    Read from figure 3.
* Value selected from several similar or calculated from data presented in the article.
References to table 5a-c.

1.    Laakso M, Edelman SV, Brechtel G, and Baron AD. Decreased effect of insulin to stimulate skeletal muscle
      blood flow in obese man. A novel mechanism for insulin resistance. J.Clin.Invest. 1990;85(6):1844-52.
2.    Kelley DE and Mandarino LJ. Hyperglycemia normalizes insulin-stimulated skeletal muscle glucose oxidation
      and storage in noninsulin-dependent diabetes mellitus. J.Clin.Invest 1990;86(6):1999-2007.
3.    Anderson EA, Hoffman RP, Balon TW, Sinkey CA, and Mark AL. Hyperinsulinemia produces both sympathetic
      neural activation and vasodilation in normal humans. J.Clin.Invest. 1991;87(6):2246-52.
4.    Bak JF, Moller N, Schmitz O, Saaek A, and Pedersen O. In vivo insulin action and muscle glycogen synthase
      activity in type 2 (non-insulin-dependent) diabetes mellitus: effects of diet treatment. Diabetologia
5.    Bonadonna RC, Saccomani MP, Seely L, Zych KS, Ferrannini E, Cobelli C, and DeFronzo RA. Glucose transport
      in human skeletal muscle. The in vivo response to insulin. Diabetes 1993;42(1):191-8.
6.    Vollenweider P, Tappy L, Randin D, Schneiter P, Jequier E, Nicod P, and Scherrer U. Differential effects of
      hyperinsulinemia and carbohydrate metabolism on sympathetic nerve activity and muscle blood flow in
      humans. J.Clin.Invest. 1993;92(1):147-54.
7.    Utriainen T, Malmstrom R, Makimattila S, and Yki-Jarvinen H. Methodological aspects, dose-response
      characteristics and causes of interindividual variation in insulin stimulation of limb blood flow in normal
      subjects. Diabetologia 1995;38(5):555-64.
8.    Dela F, Larsen JJ, Mikines KJ, and Galbo H. Normal effect of insulin to stimulate leg blood flow in NIDDM.
      Diabetes 1995;44(2):221-6.
9.    Creager MA, Liang CS, and Coffman JD. Beta adrenergic-mediated vasodilator response to insulin in the
      human forearm. J Pharmacol Exp Ther 1985;235(3):709-14.
10.   Natali A, Buzzigoli G, Taddei S, Santoro D, Cerri M, Pedrinelli R, and Ferrannini E. Effects of insulin on
      hemodynamics and metabolism in human forearm. Diabetes 1990;39:490-500.
11.   Lembo G, Rendina V, Iaccarino G, Lamenza F, Volpe M, and Trimarco B. Insulin reduces reflex forearm
      sympathetic vasoconstriction in healthy humans. Hypertension 1993;21(6 Pt 2):1015-9.
12.   Jern S. Effects of insulin on vascular responses to mental stress and norepinephrine in human forearm.
      Hypertension 1994;24(6):686-94.
13.   Taddei S, Virdis A, Mattei P, Natali A, Ferrannini E, and Salvetti A. Effect of insulin on acetylcholine-induced
      vasodilation in normotensive subjects and patients with essential hypertension. Circulation 1995;92(10):2911-
14.   Cardillo C, Kilcoyne CM, Nambi SS, Cannon RO, III, Quon MJ, and Panza JA. Vasodilator response to systemic
      but not to local hyperinsulinemia in the human forearm. Hypertension 1998;32(4):740-5.
15.   Steinberg HO, Brechtel G, Johnson A, Fineberg N, and Baron AD. Insulin-mediated skeletal muscle
      vasodilation is nitric oxide dependent. A novel action of insulin to increase nitric oxide release. J.Clin.Invest.
16.   Steinberg HO, Chaker H, Leaming R, Johnson A, Brechtel G, and Baron AD. Obesity/insulin resistance is
      associated with endothelial dysfunction. Implications for the syndrome of insulin resistance. J.Clin.Invest.
                                              Patients                  Controls

                                              n = 28                    n = 31

                                   n = 19                   n=9         n = 31

                    Treatment      Insulin                   No

                   Early repeat

                   Late repeat
                                   n = 19                   n=8

                                  Treatment              Time control
                                    group                   group


Figure 1. Flow diagram for clinical interventions and examinations of vascular function. All 19
patients in the treatment group participated in the late repeat examination after 2 months; the first
9 consecutively included patients in the treatment group also participated in the early repeat
examination after 3 days. Hypoglycemic drugs were withdrawn in all patients 2 weeks prior to the
initial examination. Treatment consisted of insulin 3 times daily at meals and intermediate-acting
insulin at bed-time. Hypoglycemic therapy was witheld in patients in the time control group until
the late repeat examination day. One patient in the time control group was taken out of the study
before the late examination because of symptomatic hyperglycemia.

                                                                    Insulin                        BH4
log(ACh dose) [µg min ]

                                                                                                                   log(SNP dose) [µg min ]

                           30                                                                                 10
                          7.5                                                                                 3















                                               Time after arterial cannulation [min]

Figure 2a. Infusion protocol at the initial examination day. The narrow, black graphs
representsacetylcholine dose-response studies (doses at left ordinate), the bold, gray graph
represents sodium nitroprusside infusion (doses at right ordinate). Insulin and BH4 was given in
doses of 50 µU kg−1 min−1 and 500 µg min−1, respectively. This protocol was followed for the first
10 consecutively included patients. Only 7 of these participated in the BH4 study at the end of the
day. Wash-out periods between dose-response studies were of 20 minutes duration (as shown) or
more, until blood flow had returned to basal. Abbreviations: acetylcholine (ACh);
tetrahydrobiopterin (BH4); sodium nitroprusside (SNP).
log(ACh dose) [µg min ]

                                                                                                 log(SNP dose) [µg min ]

                           30                                                               10
                          7.5                                                               3







                                     Time after arterial cannulation [min]

Figure 2b. Infusion protocol followed for 9 patients at the early repeat examination day and for 4
patients at the late repeat examination day. Abbreviations as in figure 2a.



                                                                            Insulin                                        Insulin
log(ACh dose) [µg min ]

                                                                                                                                               log(SNP dose) [mg min ]


                           30                                                                                                             10
                          7.5                                                                                                             3
















                                                Time after arterial cannulation [min]

Figure 2c. Infusion protocol followed for 10 healthy controls and for 17 patients at the initial
examination day and for 15 patients at the late repeat examination day. Abbreviations: L-NMMA
(NG-monomethyl-L- arginine), otherwise as in figure 2a.
   [ml (100 ml)−1 min−1]

        Blood flow

                                    0       7.5   15          30                       60

                                                  Acetylcholine dose [µg min−1]

Figure 3. Forearm blood flow during intra-brachial infusion of acetylcholine in 28 patients with
type 2 diabetes and ischemic heart disease (closed circles ●) and in 31 lean, healthy control
subjects (open circles ○). The blood flow response to acetylcholine was lower in patients than in
controls (p = 0.03). Error bars represent standard error of the mean.

 [ml (100 ml)−1 min−1]

      Blood flow

                                0       1           3                                  10
                                              Sodium nitroprusside dose [µg min ]

Figure 4. Forearm blood flow during intra-brachial infusion of sodium nitroprusside in 28 patients
with type 2 diabetes and ischemic heart disease (closed circles ●) and in 10 lean, healthy control
subjects (open circles ○). The blood flow response to sodium nitroprusside was lower in patients
than in controls (p = 0.0008). Error bars represent standard error of the mean.
    [ml (100 ml)−1 min−1]

         Blood flow

                                 0   7.5   15          30                  60

                                           Acetylcholine dose [µg min−1]

Figure 5. Forearm blood flow during intra-brachial infusion of acetylcholine (closed circles ●) and
during combined infusion of acetylcholine and insulin (0.05 mU kg−1 min−1, started 20 minutes
before, and maintained during, acetylcholine infusion), with (open triangles ▽) or without (open
circles ○) infusion of NG-monomethyl-L-arginine (L-NMMA) in 10 lean, healthy control subjects.
Blood flow at dose = 0 µg min−1 was recorded as base-line before start of the insulin infusion. The
blood flow response to actylcholine was higher during insulin stimulation and L-NMMA inhibited the
combined insulin and acetylcholine response (p < 0.0001). Error bars represent standard error of
the mean.

 [ml (100 ml)−1 min−1]

      Blood flow

                                 0   7.5   15         30                   60
                                           Acetylcholine dose [µg min ]

Figure 6. Forearm blood flow during intra-brachial infusion of acetylcholine (closed circles ●) and
during combined infusion of acetylcholine and insulin (0.05 mU kg−1 min−1, started 20 minutes
before, and maintained during, acetylcholine infusion) (open circles ○) before long-term
intervention in 28 patients with type 2 diabetes and ischemic heart disease. Blood flow at dose = 0
µg min−1 was recorded as base-line before start of the insulin infusion. The blood flow response to
actylcholine was unchanged during insulin stimulation (p = 0.3). Error bars represent standard
error of the mean.
  [ml (100 ml) min ]

      ∆ blood flow



                               0   7.5   15          30                  60

                                         Acetylcholine dose [µg min−1]

Figure 7. Decrements of forearm blood flow during intrabrachial infusion of NG-monomethyl-L-
arginine (L-NMMA) as a co-infusion with acetylcholine and insulin (0.05 mU kg−1 min−1, started 20
minutes before, and maintained during, acetylcholine infusion) compared with combined infusion
of acetylcholine and insulin in 10 lean, healthy control subjects (open triangles ▽) and in 17
patients with type 2 diabetes and ischemic heart disease (closed triangles ▼). Blood flow at dose =
0 µg min−1 was recorded as base-line before start of the insulin infusion. Blood flow decrements
were not different in patients and controls. Error bars represent standard error of the mean.

  [ml (100 ml)−1 min−1]

       Blood flow

                               0   7.5   15          30                  60
                                         Acetylcholine dose [µg min ]

Figure 8. Forearm blood flow during intra-brachial infusion of acetylcholine (closed circles ●) and
during combined infusion of acetylcholine and tetrahydrobiopterin (BH4) (500 µg min−1, started 5
minutes before, and maintained during, acetylcholine infusion) (open circles ○), a cofactor for
endothelial nitric oxide synthase (eNOS), in 7 patients with type 2 diabetes and ischemic heart
disease. Blood flow at dose = 0 µg min−1 was recorded as base-line before start of the insulin
infusion. The blood flow response to acetylcholine was higher during BH4 infusion (p < 0.0001).
Error bars represent standard error of the mean.
   [ml (100 ml)−1 min−1]

        Blood flow

                                    0       7.5   15          30                       60

                                                  Acetylcholine dose [µg min−1]

Figure 9. Forearm blood flow during intra-brachial infusion of acetylcholine in 9 patients with type
2 diabetes and ischemic heart disease before (closed circles ●) and after (closed squares ■) 3
days of insulin therapy. The blood flow response to acetylcholine was lower after 3 days of insulin
therapy (p = 0.007). Error bars represent standard error of the mean.

 [ml (100 ml)−1 min−1]

      Blood flow

                                0       1           3                                  10
                                              Sodium nitroprusside dose [µg min ]

Figure 10. Forearm blood flow during intra-brachial infusion of sodium nitroprusside in 9 patients
with type 2 diabetes and ischemic heart disease before (closed circles ●) and after 3 days (closed
squares ■) and 2 months (closed triangles ▲) of insulin therapy. Compared to the situation before
insulin therapy, the blood flow response to sodium nitroprusside was lower after 3 days of insulin
therapy (p = 0.009), but unchanged after 2 months (p = 0.09). Error bars represent standard error
of the mean.
  [ml (100 ml)−1 min−1]

       Blood flow

                               0   7.5   15          30                  60

                                         Acetylcholine dose [µg min−1]

Figure 11. Forearm blood flow during intra-brachial infusion of acetylcholine before (closed
circles ●) and after (closed triangles ▲) 2 months of insulin therapy in 19 patients with type 2
diabetes and ischemic heart disease. The blood flow response to acetylcholine was unchanged by
insulin therapy (p = 0.09). Error bars represent standard error of the mean.
    [ml (100 ml)−1 min−1]

         Blood flow

                                 0   7.5   15          30                  60

                                           Acetylcholine dose [µg min−1]

Figure 12. Forearm blood flow during intra-brachial infusion of acetylcholine (closed circles ●)
and during combined infusion of acetylcholine and insulin (0.05 mU kg−1 min−1, started 20 minutes
before, and maintained during, acetylcholine infusion) (open circles ○) before insulin therapy in 13
patients with type 2 diabetes and ischemic heart disease. Blood flow at dose = 0 µg min−1 was
recorded as base-line before start of the insulin infusion. The blood flow response to actylcholine
was unchanged during insulin stimulation (p = 0.5). Error bars represent standard error of the

 [ml (100 ml)−1 min−1]

      Blood flow

                                 0   7.5   15         30                   60
                                           Acetylcholine dose [µg min ]

Figure 13. Forearm blood flow during intra-brachial infusion of acetylcholine (closed circles ●)
and during combined infusion of acetylcholine and insulin (0.05 mU kg−1 min−1, started 20 minutes
before, and maintained during, acetylcholine infusion) (open circles ○) after 2 month’s insulin
therapy in the same 13 patients as in figure 12. Blood flow at dose = 0 µg min−1 was recorded as
base-line before start of the insulin infusion. The blood flow response to actylcholine was higher
during insulin stimulation (p = 0.0002). Error bars represent standard error of the mean.
  [ml (100 ml) min ]

      ∆ blood flow



                             0   7.5   15          30                  60

                                       Acetylcholine dose [µg min−1]

Figure 14. Decrements of forearm blood flow during intrabrachial infusion of NG-monomethyl-L-
arginine (L-NMMA) as a co-infusion with acetylcholine and insulin (0.05 mU kg−1 min−1, started 20
minutes before, and maintained during, acetylcholine infusion) compared with combined infusion
of acetylcholine and insulin in 11 patients with type 2 diabetes and ischemic heart disease before
(closed triangles ▼) and after (open triangles ▽) 2 months of insulin therapy. Blood flow at dose =
0 µg min−1 was recorded as base-line before start of the insulin infusion. Blood flow decrements
were not different before and after insulin therapy (p = 0.9). Error bars represent standard error of
the mean.
  [ml (100 ml)−1 min−1]

       Blood flow

                               0   7.5    15          30                       60

                                          Acetylcholine dose [µg min−1]

Figure 15. Forearm blood flow during intra-brachial infusion of acetylcholine before (closed circles
●) and after (open circles ○) 2 months without hypoglycemic drug therapy in 8 patients with type
2 diabetes and ischemic heart disease. The blood flow response to actylcholine was unchanged
during this period (p = 0.09). Error bars represent standard error of the mean.

  [ml (100 ml)−1 min−1]

       Blood flow

                               0   1           3                               10
                                       Sodium nitroprusside dose [µg min ]

Figure 16. Forearm blood flow during intra-brachial infusion of sodium nitroprusside before
(closed circles ●) and after (open circles ○) 2 months without hypoglycemic drug therpay in 8
patients with type 2 diabetes and ischemic heart disease. The blood flow response to sodium
nitroprusside was unchanged during this period (p = 0.6).
  [ml (100 ml)−1 min−1]        16
       Blood flow

                                                              0      7.5   15             30                   60
                                                                               Acetylcholine dose [µg min ]

Figure 17. Forearm blood flow during intra-brachial infusion of acetylcholine (closed circles ●)
and during combined infusion of acetylcholine and insulin (0.05 mU kg−1 min−1, started 20 minutes
before, and maintained during, acetylcholine infusion) (open circles ○) in 8 patients with type 2
diabetes and ishcemic heart disease. Furthermore, blood flow during acetylcholine infusion (closed
triangles △) and combined insulin and acetylcholine infusion (closed triangles ) in the same
patients after 2 months without hypoglycemic therapy. Blood flow at dose = 0 µg min−1 was
recorded as base-line before start of the insulin infusion. Insulin did not have an effect on the
acetylcholine response before or after this period. (p > 0.7). Error bars represent standard error of
the mean.


                                                          8         BH4 0 µM
                          Light emission [10 counts s ]


                                                                    BH4 1 µM


                                                                  BH4 2.5 µM
                                                                    BH4 5 µM

                                                          0       BH4 100 µM
                                                                               1            2            3          4

                                                                           Time from adding xanthine oxidase [minutes]

Figure 18: anti-oxidant effect of 6R- and 6S-BH4 in vitro. Luminol-enhanced
chemiluminescence of the superoxide producing reaction between xanthine 400 nM and xanthine
oxidase 2 mU ml−1 in the presence of 6R-BH4 (closed circles ●), 6S-BH4 (open circles ○), or
superoxide dismutase (closed triangles ▼). 6R- and 6S-BH4 are equipotent anti-oxidants in the
dose range examined.
  [ml (100 ml)−1 min−1]

       Blood flow

                               0   7.5   15          30                  60

                                         Acetylcholine dose [µg min−1]

Figure 19. Forearm blood flow during intra-brachial infusion of acetylcholine (closed circles ●)
and during combined infusion of acetylcholine and insulin (0.05 mU kg−1 min−1, started 20 minutes
before, and maintained during, acetylcholine infusion), with (open triangles ▽) or without (open
circles ○) infusion of NG-monomethyl-L-arginine (L-NMMA) in 17 patients with type 2 diabetes and
ischemic heart disease before long-term intervention. Blood flow at dose = 0 µg min−1 was
recorded as base-line before start of the insulin infusion. The blood flow response to actylcholine
was higher during insulin stimulation and L-NMMA inhibited the combined insulin and acetylcholine
response (p < 0.0001). Error bars represent standard error of the mean.


  [ml (100 ml)−1 min−1]
       Blood flow



                               0   7.5   15          30                  60
                                         Acetylcholine dose [µg min ]

Figure 20. Forearm blood flow in the forearm contralateral to the perfused forearm during intra-
brachial infusion of acetylcholine (closed circles ●) and during combined infusion of acetylcholine
and insulin (0.05 mU kg−1 min−1, started 20 minutes before, and maintained during, acetylcholine
infusion) in 10 lean, healthy control subjects. Blood flow at dose = 0 µg min−1 was recorded as
base-line before start of the insulin infusion. The blood flow response to actylcholine in this arm
was unchanged during acetylcholine infusion with or without co-infusion of insulin (p = 0.8). Error
bars represent standard error of the mean.