Prediction of Type 1 Diabetes (T1DM) & related Autoimmune Diseases (AD)
Marco Songini, MD Diabetes Unit
Azienda Ospedaliera Brotzu Cagliari (Italy)
Type 1 diabetes develops from the interaction between susceptibility genes and enviromental determinants. The major genetic susceptibility to type 1 diabetes is conferred by markers from HLA locus, but other genes are involved. The non genetic contribution to the disease (i.e. nutritional factors and infective agents) is even less wll-defined. This may imply aetiological heterogeneity in patients so that particular combinations of genetic susceptibility factors require exposure to specific non-genetic factors in order to initiate the disease developing process in type 1 diabetes. It is well known that immune markers (ICA, GADA, IA2, IAA) appear many years before clinical onset of type 1 diabetes. These “windows” offers the chance to pinpoint subjects at risk eventually suitable to preventive therapies. At present, intervention trials are recommended in the small subset of the population at high risk identified by genetic and immune markers.
Complementary strategies in the prediction of T1DM Strategy 1 Strategy 2
AIM: AIM:
TEST INTERVENTION STRATEGIES High specificity/ low sensitivity
families immune markers high risk subgroup
REDUCE INCIDENCE OF IDDM Low specificity/ high sensitivity general population genetic + immune markers
moderate risk subgroup
Bingley, E. Bonifacio & E. Gale;Diabetes, vol. 42, feb. 1993
Preventive strategies for T1DM (1)
Selective immunosuppression, using depleting or nondepleting monoclonal antibodies to lymphocyte cell surface molecules such as CD3, CD4, CD8, T cell receptor and major histocompatibility complex (MHC) antigens, or blocking peptides to T cell receptors Immunostimulation by viruses, cytokines, calcitriol, concanavalin A, bacille Calmette-Guèrin (BCG), Freund’s adjuvant or tranfusion of deficient lymphocyte subsets B-Cell rest by suppressive therapy with insulin
E. Bosi & G.F. Bottazzo; Clin. Immunother. 3 (2) 1995
Preventive strategies for T1DM (2)
Protection from oxygen radical-mediated and nitric oxidemediated damage by nicotinamide, (desferrioxamine) and aminoguanidine deferoxamine
Environmental
by manipulation temperature, diet (gluten free) and hormonal milieu
intervention
of
Induction of tolerance to B-cells by bone marrow
transplantation, lymphocyte transfusion, intrathymic islet transplantation, neonatal B-cell stimulation and administration (intravenous, intrathymic, intraperitoneal or oral) of putative B-cell autoantigens such as insulin or glutamic acid decarboxylase
E. Bosi & G.F. Bottazzo; Clin. Immunother. 3 (2) 1995
Tests to predict T1 DM & AD
Autoantibodies: ICA, GADA, IA2-A, IAA, AD-Abs HLA-phenotype: DR3/DR4 (DQ2/DQ8), AD phenos HLA-genotype: Eterodimers 57Non Asp/53Arg DQ beta/DQ Alfa, AD genos ? Cell mediated markers: Alteration of lymphocyte subsets CD4/CD8, etc.
Immunological markers for T1DM
ICA
Islet Cell Abs
Indirect immunofluorescence on human pancreatic cryosections
Risk at 10 yrs FH+ >10 JDFU 41% >80 JDFU 80%
IAA
Insulin AutoAbs
In Children R.I.A. first antibodies Risk at 10 yrs to appear ICA + IAA 81%
Risk at 5 yrs FH+ 44%
Immunological markers for T1DM
IA2-A
Protein Tyrosin Phosphatase AutoAbs
More R.I.A. common among children
High specificity low sensitivity
GADA
Glutamic Acid Decarboxilase AutoAbs
More R.I.A. common among adults
High sensitivity low specificity
Immunological markers for T1DM Combined markers in FH+
Positivity for 3 or 4 antibodies yelds a risk of 88-100% to become diabetic in 10 years The best association of autoantibodies is: GADA + IA2-A GADA + IA2-A + IAA in young children Pastore MR et al Diabetes Care 1998, 9; 1445-50 We are able to assay GADA + IA2-A on blood spots E. Bosi, E. Bonifacio et al . Diabetes Care - March 1999
Background
The preclinical stage of type 1 diabetes and related AD can last even many years These “windows” offers the chance to pinpoint subjects at risk eventually suitable to preventive therapies
Genetic markers for T1DM
HLA typing
predisposing:
HLA DR3-DQ2, DR4-DQ8
protective:
HLA DR2-DQ6
Lernmark A Diabetes Metabolism Rev 1998, 14,3-29
Genetic markers for T1DM HLA
Molecular biology of DQ chains of class second
DQ A301, DQ B302, DQ B501 Alleles: 99% of diabetic patients 50% of normal people
Lernmark A Diabetes Metabolism Rev 1998, 14,3-29
DQ B602 is fully protective for T1DM Gianani R et al. J Autoimmunity 1996, 9; 423-425
Genetic markers for T1DM (1)
Locus
ls
6p21 11p21 15q 11q13 6q25 18q 2q31 6q27
IDDM1 2,6 IDDM2 1,29 IDDM3 IDDM4 1,07 IDDM5 1,16
% References 35 Davies (1994) 9,4 Davies (1994), Bennet (1995)
Field (1994)
IDDM6 1,1 IDDM7 1,13
2,5 Hashimoto (1994), Davies (1994) 5,5 Davies (1994) 3,5 Meriman (unpub.), Davies (1994)
Gabbay 4,5 Owerbach andCopeman(1995) Davies (1994), (1994),
IDDM8 1,42 12,9 Luo (1995), Davies (1996)
Genetic markers for T1DM (2)
Locus 3q21-q25
IDDM9
ls
%
References
1,26 8,5 Reed and Todd (unpubl.)
Gough and Todd (unpubl.) 10p11.2-q11.2 IDDM10 1,45 13,7 Davies, Hashimoto (1994) TOT. 14 95.5 GCK 7p 14q24.3-q31 IDDM11 2q33 IDDM12 (CTLA-4) 2q34 IDDM13 6q21 IDDM15 Rowe (1995) Field (1996) Nistico (1996) Morahan (1996) Delepine (1997)
Natural History of T1DM
Popul islet-related Abs+ Follow- Risk up
Identical Twins 1st degree relatives (FH+) 5 yrs
Polyendocrinopathy (FH-) 10 yrs
References
10 yrs 100%Tun RY, BMJ 1994 70% ICARUS Group Study
25% Bosi E, Diabetes 1991
Polyendocrinopathy (FH+) 10 yrs High risk newborns (FH+) 2 yrs High risk newborns (gene+) 2 yrs
70% Bosi E, Diabetes 1991 50% BABYDIAB (Germany) 50% DIPP Project (Finland)
Sardinian school children (gen)7 yrs 24% SSI Study (Sardinia)
Natural history of T1DM
Triggers ? Triggers ?
Auto Abs + 75%
tt
FPIR OGTT + Triggers ?
50% GENES (susc)
0
25% TYPE 1 DIABETES
Time
Screening for pre-T1DM and related AD
•France • Spain Schoolchildren • Holland • SSI • Finland • Sweden • Oxford • Estonia • USA • Germany
Newborn
• DAISY (USA) • BABYDIAB (Germany, Australia) • SNI (Sardinia) • DIPP (Finland) • DIABFIN (Italy)
Cost of predicting T1DM
Cost of insulin therapies (per year) Cost of Screening (for each enrolled case) •Conventional Therapy (CT) $1450 •Intensive Therapy (ICT) •CSII •DPT-1 •DIPP (follow up=10 yrs) (newborns) $245
Birth 100% Genetic+Abs screening 13% Counselling
Abs follow up
$ 2 x CT $ 3 x CT $1751 $733
Hahl et al. Diabetologia (1998) 41:79-85
Cost of DM (?AD) •$ 92 billions
Birth 100% Abs screening 100% Abs follow up
Prevention of T1DM and other related AD
T1DM&AD are theoretically preventable
• Because there are environmental causes
• Because we are beginning to understand the genetic and immune basis
• Because they develops very slowly • Because we have good predictive tests • Because we can stop them in animals • Because we can run clinical trials
T1DM & AD are suitable diseases for preclinical screening and intervention
• Serious consequences (in USA 50 deaths yearly from
DKA)
• Treatment following diagnosis expensive, demanding, limited effect on complications • Identifiable preclinical phase also for AD • Identifiable subjects “not at risk” also for AD • ...but as yet no preventive therapy of proved efficacy (no penicillin for prediabetes!)
Assigning risk
• Primary prevention: must be based on family
history or high risk HLA - and will miss a lot of cases!
• Secondary prevention: immune-markers
relatively stable after age 5; almost inevitable progression with multiple antibodies; excellent screening efficiency (islet imaging)
Setting up an intervention: in whom?
• Primary: Neonates with family history or high
risk HLA
• Secondary:
– Infants: HLA DR3/4 with antibodies – Children/young adults with multiple Abs (T1DM&AD) – Older adults with LADA
Setting up an intervention: with what?
Should work:
– In animal models – In newly diagnosed type 1 – In pilot trials (assessed how?)
Must have:
– An acceptable safety profile – Ease of administration
Setting up an intervention: conclusions
• At present trials must be large, structured, costly and long term • Will depend on international collaboration • We need a disciplined consensus process for evaluating and prioritizing new therapies • Role of pharmaceutical industry? clinicians should have a say
T1DM prevention trials
Primary
Cow’s milk avoidance: TRIGR Gluten free diet: PREVFIN
Secondary
Nicotinamide: DENIS, ENDIT, New Zealand Insulin: DPT-1, EPLL SCIT; Schwabing, Brunetti 1999
Tertiary
Cyclosporin: GETREM, French and Canadian studies Linomide : Franco-Swedish trial
Intervention trials: assumptions
Trial ENDIT DENIS DPT (high) DPT (inter) N 530 130 340 490 Yr Diabetes % 5 40:26 3 30:6 4 84:55 4 24:12 RRR % 35 80
35 50
Mahon and Dupre,1997
Cyclosporin before onset of T1DM
• 6 relatives vs 9 historical controls • All controls developed diabetes in 12 months • 4/6 cyclosporin treated patients developed diabetes within 4 years (5, 24, 24 and 47 months)
Carel et al., 1996
Intervention in early infancy?
• • • • • Level of risk? Safety of intervention? Long term data? Acceptability/compliance? Efficacy demonstrated in other AD? Can the intervention be tested effectively in this category of patient?
Emergence of at least one auto-Ab by the age of 2 years (n=173)
FOLLOW-UP, MONTHS 3 6 9 12 18 24 TOTAL
CASEIN 0/83 HYDROLYSATE CM-BASED FORMULA 0/87
0/75
1/72
3/71
1/67
1/62
1/84
6/79
7/78
6/77
7/76
3/84 (3.6%) p=0.06 10/89 (11.2%)
The Second TRIGR Pilot Study
EURODIAB Sardinia (1989-98) birth seasonality
20 18 16 14 12 10 8 6 4 2 0 Jan-March Jan-March
No. of births/month
P<0.001
Apr-June Jul-Sept Apr-June Jul-Sept Seasons of birth
Oct-Dec Oct-Dec
N=1928, 0-29yr
Future Directions?
• Surrogate end-points • Safety and acceptability need to be balanced against efficacy • Early “one-off” therapy would be ideal • Explicit standards for performance of trials • Fewer, better quality studies based on international consensus • Lessons from other human autoimmune disease?