Persistent Osteopenia in Adult Cystic Fibrosis
Transmembrane Conductance Regulator–deﬁcient Mice
Christina K. Haston1, Wei Li2, Ailian Li2, Melanie Laﬂeur1, and Janet E. Henderson2
Meakins-Christie Laboratories and 2J.T.N. Wong Laboratories for Mineralised Tissue Research, Department of Medicine, McGill University,
Montreal, Quebec, Canada
Rationale: A loss of function mutation in the cystic ﬁbrosis trans-
membrane conductance regulator gene is believed to be an in- AT A GLANCE COMMENTARY
dependent risk factor for bone disease in patients with cystic ﬁbrosis.
Objectives: The objective of this work was to use congenic mice as Scientiﬁc Knowledge on the Subject
a preclinical model to examine the bone phenotype of Cftr2/2 mice Bone disease in patients with cystic ﬁbrosis is multifacto-
and control littermates at 8, 12, and 28 weeks of age. rial, but the contribution of the cystic ﬁbrosis transmem-
Methods: The bone phenotype of control and Cftr2/2 mice was brane conductance regulator (CFTR) gene to bone disease
evaluated by quantitative imaging, histologic and histomorphomet- remains undeﬁned.
ric analyses, and serum levels of bone biomarkers.
Measurements and Main Results: At 12 weeks of age, Cftr2/2 mice were
smaller, had lower bone mineral density, cortical bone thinning, and What This Study Adds to the Field
altered trabecular architecture compared with Cftr1/1 or Cftr1/2
control mice. In skeletally mature 28-week-old mice, there were We show that adult Cftr-deﬁcient mice have low femoral
persistent deﬁcits in cortical and trabecular bone structure in Cftr2/2 bone mineral density and compromised bone architecture
mice despite signiﬁcant, quantiﬁable improvements. Cftr2/2 mice in the absence of other overt disease symptoms. The results
also had lower serum insulin-like growth factor-I levels at 12 weeks suggest that deﬁciency of Cftr contributes to the osteopenic
of age than did control mice, whereas parathyroid hormone and phenotype of Cftr2/2 mice.
25-hydroxyvitamin D levels were not signiﬁcantly different.
Conclusions: Persistent osteopenia and structural abnormalities in
adult Cftr2/2 mice, in the absence of overt respiratory and gastro-
intestinal disease, suggest that loss of Cftr function has a direct
impact on bone metabolism in Cftr2/2 mice that is not sex speciﬁc or physical inactivity, and steroid therapy. Studies of serum bio-
subject to haplotype insufﬁciency. markers for bone turnover in patients with CF suggest that bone
resorption exceeds formation (2). It has also been suggested
Keywords: bone disease; lung disease; preclinical model; genetically that low BMD in adults with CF may arise from a combination
modiﬁed mouse of insufﬁcient bone accrual during puberty (7) and subsequent
bone loss in young adulthood (2). King and coworkers (8) have
Cystic ﬁbrosis (CF) is a disease in which mutations in the cystic
shown that the common DF508 mutation in the chloride ion
ﬁbrosis transmembrane conductance regulator gene (CFTR) in
channel encoded by CFTR is an independent risk factor for
humans lead to abnormalities affecting principally the lung,
bone disease in the CF population, in addition to malnutrition
intestine, and pancreas (1). Patients with CF may also suffer
and lung disease and Sermet-Gaudelus and coworkers (9) report
from bone disease, characterized by low bone mineral density
low BMD in children with CF who are younger than 6 years of
(BMD) and increased fracture rates (2). In general, suscepti-
age. The mechanisms through which the primary CF defect inﬂu-
bility to fracture is determined by a combination of factors
ences bone metabolism, however, remain undeﬁned.
including BMD, geometry, microarchitecture, and bone cell
Mouse models that exhibit the altered electrophysiology
activity (3, 4). The diagnosis and response to treatment of
expected to arise from mutations in Cftr have been extensively
osteopenia and osteoporosis, therefore, are currently monitored
characterized (10), but there have been relatively few studies
with tests that determine BMD and bone architecture (using
investigating their bone phenotype (11, 12). Dif and coworkers
peripheral quantitative micro–computed tomography) and bone
(11) evaluated 3-week-old female UNC Cftr-deﬁcient mice,
cell activity (using serum biomarker assays).
which carry a mutation that results in the absence of the Cftr
In cystic ﬁbrosis, according to the consensus statement on
protein in a mixed genetic background (13). They showed the
bone health (2), bone disease arises from a variety of factors
weanling mice to have lower BMD, reduced cortical bone width,
including decreased vitamin absorption, pancreatic insufﬁ-
and thinner trabeculae compared with their wild-type littermates.
ciency, altered sex hormone production, chronic lung infection
This study suggested that bone disease in weanling Cftr-deﬁcient
resulting in increased levels of bone-active cytokines (5, 6),
mice was due primarily to the absence of Cftr, as the mice had
not yet developed pancreatic insufﬁciency or lung disease and
had not been subjected to therapeutic doses of steroids.
Under physiological conditions, skeletal maturity in mice has
(Received in original form May 3, 2007; accepted in ﬁnal form November 9, 2007) been determined to occur at about 16 weeks of age. Comparable
Supported by the CCFF and Canadian Institutes of Health Research, the Basic to the situation seen in humans, skeletal maturity in the mouse
Research and Therapy (BREATHE) Program, Valorisation Recherche Quebec, and marks the cessation of longitudinal and appositional bone growth.
Fonds de la Recherche en Sante Quebec.
At this stage, net bone acquisition is replaced by remodeling to
Correspondence and requests for reprints should be addressed to Christina maintain bone mass. Given the changes in bone physiology with
Haston, Ph.D., Meakins-Christie Laboratories, 3626 rue St. Urbain, Montreal, PQ,
H2X 2P2 Canada. E-mail: email@example.com
age, the current study was undertaken to determine how loss of
function of the gene encoding the cystic ﬁbrosis transmembrane
Am J Respir Crit Care Med Vol 177. pp 309–315, 2008
Originally Published in Press as DOI: 10.1164/rccm.200705-659OC on November 15, 2007 conductance regulator, Cftr, inﬂuences the bone phenotype in
Internet address: www.atsjournals.org skeletally mature, juvenile, and adult CF mice.
310 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 177 2008
METHODS Cftr genotype, using a single section from three different mice per
group. Osteoclast numbers were determined by counting the number of
Mice TRAP-positive cells present in ﬁve individual ﬁelds of bone at a ﬁnal
Animal procedures were performed in accordance with McGill Uni- magniﬁcation of 3400, and normalizing this count to the bone volume.
versity (Montreal, PQ, Canada) guidelines set by the Canadian Council Osteoblast cell numbers were similarly determined by examination of
on Animal Care (Ottawa, ON, Canada). Congenic BALB Cftr UNC1/2 von Kossa–stained sections.
mice were originally obtained from J. Hu of the Hospital for Sick
Children (Toronto, ON, Canada) and maintained in a breeding colony Blood Analysis
at the Meakins-Christie Laboratories of McGill University. The Cftr1/2 Blood was collected by cardiac puncture of anesthetized animals and
mice were intercrossed to produce Cftr1/1, Cftr1/2, and Cftr2/2 mice, centrifuged, and the supernatant serum was stored at –858C until
which were identiﬁed by genotyping as described previously (14). We biochemical analyses for bone biomarkers were undertaken with the
used mice that are congenic in the BALB background to control for pos- following commercial assays: radioimmunoassay for 25-hydroxyvitamin
sible effects of background strain variation on the CF phenotype (14–16). D (ImmunoDiagnostic Systems, Tyne and Wear, UK), ELISA for
To circumvent premature death of Cftr2/2 mice as a result of intestinal insulin-like growth factor-I (Quantikine mouse IGF-I; R&D Systems,
disease, control and Cftr2/2 mice were fed standard chow and received Minneapolis, MN), and ELISAs for parathyroid hormone and osteo-
PegLyte (polyethylene glycol [17.8 mmol/L] and electrolytes; Phar- calcin (Immunotopics, Inc., San Clemente, CA).
mascience, Montreal, PQ, Canada) in their drinking water as described
previously (17). PegLyte is used clinically to cleanse the intestinal tract Statistical Analysis
while preserving electrolyte balance. To determine the effect of PegLyte Results are expressed as means 6 SD and differences between groups of
on bone growth in normal mice, one group of female Cftr1/1 mice was mice were determined by the Student t test and the Mann-Whitney test.
fed regular drinking water. Mixed sex groups were used for experiments
conducted at 12 weeks (Cftr2/2: 12 wk, ﬁve males and three females
[5 M/3 F]; control: 12 wk, 8 M/6 F) and 28 weeks (Cftr2/2: 28 wk, 1 M/7 F;
control: 28 wk, 1 M/5 F). Effect of PegLyte, Sex, and Haplotype on 12-Week-Old
Phenotyping procedures were performed essentially as described pre- Intervention with PegLyte to minimize lethality due to the gas-
viously (18, 19). At the indicated ages mice were weighed and killed with trointestinal phenotype enabled the survival of all the Cftr2/2
a lethal dose of anesthetic before capturing high-resolution X-rays of the mice to adulthood. To determine whether this treatment altered
femora and vertebrae with a Faxitron general science radiography body weight or skeletal development of control mice we assessed
system (model MX-20/DC2; Faxitron X-ray Corporation, Wheeling, the morphology, mineral density, and structure of femoral bone
IL), equipped with an FPX-2 imaging system (MedOptics/DALSA Life in 12-week-old mice receiving PegLyte in their drinking water,
Sciences, Waterloo, ON, Canada). X-rays were used to qualitatively and compared the results with those of age- and sex-matched
assess bone morphology and mineral density and to measure the lengths control mice receiving plain water. As shown in Table 1, a small
of femora and vertebrae. The right femur was measured from the plateau
but signiﬁcant decrease in bone volume and altered trabecular
of the knee to the tip of the femoral head and the vertebral measurement
represents the length from the proximal end of lumbar vertebra 3 (L3) to architecture was seen in the control mice receiving PegLyte
the distal end of L5 (L3–L5). A Lunar PIXImus 1.46 instrument (GE compared with untreated control mice.
Healthcare Lunar, Madison, WI) was then used to measure BMD of the To eliminate the possibility that haplotype insufﬁciency
intact animal before dissecting bones free of soft tissue. Bones were ﬁxed might contribute to a skeletal phenotype we analyzed the bones
overnight in 4% paraformaldehyde, rinsed three times with phosphate- of each of four 3-month-old female Cftr1/1 and Cftr1/2 mice.
buffered saline, and scanned with a SkyScan 1072 micro–computed No signiﬁcant differences in morphology, BMD, or architecture
tomography (micro-CT) instrument (SkyScan, Antwerp, Belgium) to were observed between Cftr1/1 and Cftr1/2 mice (data not
assess bone morphometry. Image acquisition was performed at 45 kV, shown). These data are consistent with a previous report for
222 mA for a 2.24-second exposure at 350 for femur and 330 for lumbar
weanling 3-week-old littermates (11). Cftr1/1 and Cftr1/2 mice
vertebrae with a 0.98 rotation between frames. These two-dimensional
images were used to reconstruct three-dimensional images for quantita-
were subsequently used interchangeably as control animals.
tive analysis, using 3D Creator software supplied with the instrument.
Age-related Changes in Bone Morphology and BMD in
The area of interest selected for quantiﬁcation of trabecular bone was
located immediately below the growth plate and extended for 1 mm Control and Cftr2/2 Mice
toward the diaphysis of the femur or the midpoint of the vertebra. In a previous study it was reported that 3-week-old weanling
Reported parameters include the following: bone volume as a percentage Cftr-deﬁcient mice had reduced BMD and osteopenia charac-
of tissue volume (BV/TV); the structure model index (SMI), which is the terized by increased bone resorption relative to formation (11).
ratio of rodlike structures to platelike structures, with the ideal ratio
Table 2 conﬁrms and expands on the previous data by showing
being 1.0. A high ratio is indicative of a reduction in trabecular
connectivity; the thickness of individual trabeculae (TrTh); trabecular
that the highly signiﬁcant reduction in body weight, bone length,
separation (TrSp), which is an indirect measure of trabecular thinning; and BMD in Cftr2/2 mice compared with control mice persists
and the number of trabeculae in a given area (TrNo). at 8 and 12 weeks of age. These changes were apparent in both
male (n 5 5) and female (n 5 3) Cftr2/2 mice compared with
Histology and Histochemistry their sex-matched control mice at the age of 12 weeks and were
After micro-CT analysis the right femur of each of four mice from each not signiﬁcantly different between 12-week-old male and female
group was embedded in polymethylmethacrylate at low temperature Cftr2/2 mice (data not shown). Skeletally mature 28-week-old
and 4-mm sections were cut. Sections were stained with von Kossa Cftr2/2 mice had largely normalized with respect to skeletal
to identify mineralized bone and counterstained with toluidine blue to morphology and BMD in the vertebra, although the femoral
identify unmineralized tissue. Midsagittal sections, corresponding to BMD remained signiﬁcantly lower.
the micro-CT images, were cut and images were captured at a magni-
ﬁcation of 32.5, using a Zeiss Axioskop microscope (Carl Zeiss Age-related Changes in Trabecular Bone Architecture in
Canada Ltd, Toronto, ON, Canada) equipped with a Zeiss AxioCam Control and Cftr2/2 Mice
MRc camera. Adjacent sections were stained for alkaline phosphatase
(ALP) activity in osteoblasts or tartrate-resistant acid phosphatase To further explore the differences in the skeletal phenotypes of
(TRAP) in osteoclasts as described previously (18). Cell numbers were control and Cftr2/2 mice we assessed architectural parameters
scored by three independent investigators, blinded to mouse age and by micro-CT and bone composition by classic histology. At 12
Haston, Li, Li, et al.: Bone Phenotype of Adult Cftr2/2 Mice 311
TABLE 1. EFFECT OF SEX AND TREATMENT WITH PEGLYTE ON BONE CHARACTERISTICS OF
12-WEEK-OLD Cftr1/1 AND Cftr1/2 CONTROL MICE
1 PegLyte P Value
– PegLyte PegLyte:
Parameter Female (n 5 5) Female (n 5 6) Male (n 5 8) PegLyte: – vs. 1 Female vs. Male
Body weight, g 22.44 6 2.91 21.48 6 1.43 28.91 6 1.18 NS ,0.001
Femur length, cm 1.498 6 0.026 1.457 6 0.027 1.495 6 0.050 ,0.029 NS
Vertebral length, cm 0.918 6 0.056 0.955 6 0.024 1.003 6 0.025 NS ,0.004
Femur BMD, mg/cm2 70.06 6 1.90 69.67 6 3.88 72.48 6 6.58 NS NS
Vertebral BMD, mg/cm2 53.20 6 1.51 55.83 6 2.40 59.30 6 3.35 NS NS
BV/TV, % 16.20 6 1.67 12.36 6 2.58 15.82 6 1.34 ,0.019 ,0.007
SMI 1.37 6 0.66 1.76 6 0.20 1.56 6 0.14 NS ,0.050
TrTh, mm 59.4 6 1.14 56.67 6 2.16 51.38 6 1.51 ,0.032 ,0.001
TrSp, mm 213.4 6 7.40 240.7 6 21.30 204.9 6 33.88 ,0.024 ,0.043
TrNo, n 2.82 6 0.19 2.17 6 0.42 3.09 6 0.26 ,0.012 ,0.001
Deﬁnition of abbreviations: BMD 5 bone mineral density; BV/TV 5 bone volume as a percentage of tissue volume; NS 5 not
signiﬁcant; SMI 5 structure model index (i.e., the ratio of rodlike structures to platelike structures); TrNo 5 number of trabeculae
in a given area; TrSp 5 trabecular separation, an indirect measure of trabecular thinning; TrTh 5 thickness of individual trabeculae.
BALB Cftr1/1 and Cftr1/2 mice were maintained with or without PegLyte in their drinking water until 12 weeks of age, the time
of death. Bone lengths were measured on the basis of high-resolution X-rays (Faxitron general science radiography system model
MX-20/DC2) and bone mineral density (BMD) was determined with a Lunar PIXImus densitometer, as described in METHODS. Bone
volume; the ratio of platelike to rodlike structures; and trabecular thickness, separation, and number were quantiﬁed by micro–
computed tomography. Values represent means 6 SD.
weeks of age (Figure 1), reconstruction of micro-CT images into little difference between control and Cftr2/2 mice, except for
a three-dimensional model (Figure 1a) shows little difference a marginal increase in TRAP staining in the Cftr2/2 mice. By
between the control (Figure 1, top) and Cftr2/2 (Figure 1, 28 weeks of age (Figure 2) there appears to be an increase in
bottom) mice, except perhaps for a thinner diaphyseal shaft in trabecular bone (Figure 2b) and cortical thickness (Figure 2c) in
the Cftr2/2 mice. Two-dimensional coronal sections adjacent to the Cftr2/2 mice (Figure 2, bottom) as evidenced by micro-CT
the growth plate (Figure 1b) show the reduced trabecular bone and von Kossa staining (Figure 2e). ALP activity (Figure 2f) is
and cortical thinning in the diaphysis (Figure 1c) of Cftr2/2 mice greatly increased in Cftr2/2 mice compared with control mice
compared with control mice. These data are corroborated in the and TRAP activity (Figure 2g) also appears up-regulated.
matched, midsagittal micro-CT (Figure 1d) and von Kossa– and Quantitative micro-CT analysis of bone at 8 weeks (skele-
toluidine blue–stained histologic sections (Figure 1e). Adjacent tally immature), 12 weeks (juvenile), and 28 weeks (skeletally
sections stained for ALP (Figure 1f) and TRAP (Figure 1g) to mature adult) of age is shown in Table 3. All parameters except
localize osteoblasts and osteoclasts, respectively, show there is for trabecular thickness changed signiﬁcantly between 8 and
12 weeks in the control mice and there was minimal signiﬁcant
TABLE 2. AGE-RELATED CHANGES IN MORPHOLOGY AND change after that time. In contrast, bone architecture in the
MINERAL DENSITY OF BONE TISSUE OF CONTROL AND Cftr2/2 mice did not change signiﬁcantly between 8 and 28 weeks
Cftr2/2 MICE of age, despite the signiﬁcant increase in ALP- and TRAP-positive
cells. Thus, at 8 weeks of age, trabecular bone volume in Cftr2/2
Age (wk) Parameter Control Cftr2/2 P Value
femora is similar to that in control littermates, despite the signif-
8* Number 6 6 icant difference in BMD, but it was signiﬁcantly reduced at 12 and
Body weight, g 18.30 6 2.79 14.55 6 1.65 ,0.018 28 weeks of age, primarily because of reduced numbers of tra-
Femur length, cm — — — beculae. As shown in Figure 3A, there was a signiﬁcant increase
Vertebral length, cm — — —
Femur BMD, mg/cm2 55.60 6 5.86 43.00 6 1.00 ,0.003
in TRAP-positive cells in Cftr2/2 mice at both 12 and 28 weeks,
Vertebral BMD, mg/cm2 47.00 6 5.05 42.60 6 3.13 NS whereas the increase in ALP-positive cells did not reach signiﬁ-
12 Number 14 8 cance until 28 weeks. The alteration in bone cell activity in Cftr2/2
Body weight, g 25.73 6 4.01† 16.89 6 2.21 ,0.001 mice was reﬂected in a decrease in cortical width of the bones of
Femur length, cm 1.479 6 0.045 1.381 6 0.100 ,0.001 Cftr2/2 mice compared with that seen in control mice. The dis-
Vertebral length, cm 0.982 6 0.034 0.884 6 0.053 ,0.001 crepancy in cortical bone width between control and Cftr2/2 mice
Femur BMD, mg/cm2 71.27 6 5.58† 47.24 6 7.23 ,0.001
Vertebral BMD, mg/cm2 57.81 6 3.38† 43.04 6 4.86 ,0.001
appeared to diminish over time, as shown in Figure 3B.
28 Number 8 6 To determine whether alterations in the bone phenotype of
Body weight, g 27.42 6 2.69 21.85 6 2.05‡ ,0.004 Cftr2/2 mice were due in part to changes in the circulating levels
Femur length, cm 1.529 6 0.074 1.458 6 0.100 NS of bone-active factors, we examined serum biomarkers in
Vertebral length, cm 1.006 6 0.039 0.978 6 0.062‡ NS representative mice at 12 and 28 weeks of age. As shown in
mg/cm2 73.20 6 6.48 59.00 6 5.40‡ ,0.001 Table 4, Cftr2/2 mice had lower levels of IGF-I at 12 weeks and
mg/cm2 64.76 6 9.69‡ 59.17 6 7.20‡ NS
osteocalcin at 28 weeks compared with age-matched control
Deﬁnition of abbreviations: BMD 5 bone mineral density; NS 5 not signiﬁcant. mice. No signiﬁcant differences were seen in the calciotropic
BALB control and Cftr2/2 mice were maintained on PegLyte-treated drinking hormones parathyroid hormone and 25-hydroxyvitamin D.
water to 12 or 28 weeks of age, the time of death. Bone lengths were measured on
the basis of high-resolution X-rays (Faxitron MX-20/DC2) and BMD was de-
termined with a Lunar PIXImus densitometer, as described in METHODS. Values DISCUSSION
represent means 6 SD.
* Eight-week-old mice were all female and received normal drinking water. This study was undertaken to determine how loss of function of
P , 0.05; signiﬁcantly different from 8-week-old mice. the Cftr gene inﬂuences bone development and metabolism. We
P , 0.05; signiﬁcantly different from 12-week-old mice. have shown that at 12 weeks of age, Cftr2/2 male and female
312 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 177 2008
Figure 1. Architecture and composition
of distal femur of 12-week-old control
and Cftr2/2 mice. Bones from control
(top) and Cftr2/2 (bottom) mice were
dissected free of soft tissue, ﬁxed, and
scanned with a SkyScan 1072 micro–
computed tomography (micro-CT) sys-
tem equipped with 3D-Creator analytical
software. Three-dimensional reconstruc-
tion (a) and two-dimensional cross-
sectional scans (b–d) demonstrated fewer
trabeculae and cortical thinning of Cftr2/2
bones. Plastic-embedded sections of the
same bones, stained with von Kossa and
counterstained with toluidine blue (e) and
photographed at original magniﬁcations
of 32.5 and 320 (inset), conﬁrmed the
micro-CT data. Adjacent sections stained
for alkaline phosphatase (f, brown) and
tartrate-resistant acid phosphatase (g,
red) showed little difference between con-
trol and Cftr2/2 mice in osteoblast (brown)
or osteoclast (red) activity along the edges
of bony trabeculae. CF 5 cystic ﬁbrosis
mice are smaller, have low BMD, cortical bone thinning, and absence of overt, life-threatening lung and pancreatic disease
compromised trabecular architecture compared with wild-type or steroid hormone treatment and in the presence of signiﬁcant
or heterozygous control mice. At 28 weeks of age, when the alterations in bone cell numbers and activity. The data thus
mice are skeletally mature, there are persistent deﬁcits in support the conjecture that loss of Cftr function has a direct
cortical and trabecular bone structure despite signiﬁcant, quan- impact on skeletal metabolism that is not sex speciﬁc or subject
tiﬁable improvements. These differences were seen in the to haplotype insufﬁciency in a mouse model.
Figure 2. Architecture and composition of
distal femur of 28-week-old control and
Cftr2/2 mice. Bones from control (top) and
Cftr2/2 (bottom) mice were processed as
described in Figure 1. Trabecular bone (b–
d) and cortical width (c and d) were increased
in Cftr2/2 mice as compared with 12-week-
old Cftr2/2 mice, although the values still had
not normalized relative to control mice. A
signiﬁcant increase in alkaline phosphatase
activity (f, brown) and tartrate-resistant acid
phosphatase activity (g, red) was seen along
the edges of bony trabeculae in Cftr2/2 mice
compared with control littermates. CF 5
cystic ﬁbrosis (Cftr2/2 mice).
Haston, Li, Li, et al.: Bone Phenotype of Adult Cftr2/2 Mice 313
TABLE 3. AGE-RELATED CHANGES IN FEMUR BONE TABLE 4. SERUM LEVELS OF BONE BIOMARKERS IN 12- AND
ARCHITECTURE OF CONTROL AND Cftr2/2 MICE 28-WEEK-OLD CONTROL AND Cftr2/2 MICE
Age (wk) Parameter Control Cftr2/2 P Value Age (wk) Parameter Control Cftr2/2 P Value
8* Number 6 6 12 Number 4 5
BV/TV, % 9.02 6 2.56 7.75 6 2.28 NS 25 VitD, ng/ml 47 6 5 49 6 10 NS
SMI 2.18 6 0.14 2.13 6 0.21 NS IGF-I, ng/ml 411 6 36 208 6 97 ,0.05
TrTh, mm 57.20 6 3.42 49.20 6 2.78 ,0.003 PTH, pg/ml 23 6 4 27 6 15 NS
TrSp, mm 297.2 6 77.10 334.2 6 35.3 NS Osteocalcin, ng/ml 50 6 5 67 6 24 NS
TrNo 1.74 6 0.43 1.57 6 0.41 NS 28 Number 5 4
12 Number 14 8 25 VitD, ng/ml 49 6 5 41 6 20 NS
BV/TV, % 14.34 6 2.58† 5.83 6 2.71 ,0.001 IGF-I, ng/ml 459 6 18 389 6 102 NS
SMI 1.65 6 0.19† 2.16 6 0.24 ,0.001 PTH, pg/ml 24 6 6 20 6 6 NS
TrTh, mm 53.64 6 3.23 46.75 6 3.45 ,0.001 Osteocalcin, ng/ml 67 6 15 49 6 14 ,0.05
Tr Sp, mm 220.2 6 33.6† 307.2 6 65.3 ,0.001
TrNo 2.70 6 0.57† 1.23 6 0.51 ,0.001 Deﬁnition of abbreviations: 25 VitD 5 25-hydroxyvitamin D; IGF-I 5 insulin-like
28 Number 8 6 growth factor-I; NS 5 not signiﬁcant; PTH 5 parathyroid hormone.
BV/TV, % 13.17 6 3.22 7.11 6 1.84 ,0.001 BALB control and Cftr2/2 mice were maintained on PegLyte-treated drinking
SMI 1.52 6 0.31 1.97 6 0.12 ,0.005 water to 12 or 28 weeks of age, the time of death. Blood was drawn by cardiac
TrTh, mm 59.38 6 2.77‡ 47.67 6 2.42 ,0.001 puncture from anesthetized animals and serum biomarkers were assessed with
TrSp, mm 280.0 6 54.4‡ 266.0 6 20.1 NS a commercial radioimmunoassay assay for 25-hydroxyvitamin D and commercial
TrNo 2.22 6 0.55 1.48 6 0.33 ,0.013 ELISAs for IGF-I, parathyroid hormone, and osteocalcin. Values represent medians
6 95% conﬁdence interval.
Deﬁnition of abbreviations: BV/TV 5 bone volume as a percentage of tissue
volume; NS 5 not signiﬁcant; SMI 5 structure model index (i.e., the ratio of
rodlike structures to platelike structures); TrNo 5 number of trabeculae in a given complications with associated altered levels of inﬂammatory
area; TrSp 5 trabecular separation, an indirect measure of trabecular thinning; cytokines (6) and did not receive corticosteroids, which could
TrTh 5 thickness of individual trabeculae.
BALB control and Cftr2/2 mice were maintained on PegLyte-treated drinking
lead to severe osteopenia when combined with structural abnor-
water to 12 or 28 weeks of age, the time of death. Bone architectural parameters malities conferred by mutations in the CF gene. In contrast to the
were quantiﬁed by micro–computed tomography, as described in METHODS. BMD data, quantitative micro-CT shows that body size is not the
Values represent means 6 SD. major determinant of the observed differences in the skeletal
* Eight-week-old mice were all female and received normal drinking water. phenotypes of control and Cftr2/2 mice. At 8, 12, and 28 weeks
P , 0.05; signiﬁcantly different from 8-week-old mice. the Cftr2/2 mice weighed 20, 44, and 19% less than control mice,
P , 0.05; signiﬁcantly different from 12-week-old mice.
respectively. However, most micro-CT parameters, which reveal
changes in trabecular bone quantity and architecture that are
Impact of PegLyte on Survival and Bone Phenotype of
indicative of metabolic bone disease, were signiﬁcantly different
at 28 weeks, but not at 8 weeks, when the differences in body
In this study, we used PegLyte in the drinking water to extend the weight were approximately equivalent.
life span of Cftr2/2 mice as described by Gawenis and coworkers Using quantitative techniques such as neutron activation and
(12). The intervention had little direct effect on bone develop- bone ash analysis, Gawenis and coworkers did not detect differ-
ment, as evidenced by attainment of nearly normal bone lengths ences in the chemical composition of bones from Cftr2/2 and
by 28 weeks of age in Cftr2/2 mice, but did allow for examination control mice and concluded that Cftr does not play a direct role
of adult mice. It was reported by Gawenis and coworkers that, in bone mineralization (12). Our own studies, as well as those of
after correcting for body weight, BMD did not differ between Dif and coworkers (11), support this conjecture, as there was no
UNC Cftr2/2 mice and control littermate mice. Our own evidence of increased osteoid in either study. We also reported
observations support their data by showing no signiﬁcant de- normal circulating levels of parathyroid hormone and 25-hydrox-
crease in femoral or vertebral BMD after correcting for differ- yvitamin D, suggesting that mineral ion homeostasis and absorp-
ences in body weight in the Cftr2/2 mice. These observations are tion of vitamin D via the gut were normal in Cftr2/2 mice during
somewhat surprising given the severe vertebral osteopenia pre- and postnatal bone development. Malabsorption would have
observed in young adult patients with CF. This discrepancy resulted in low vitamin D and an excess of osteoid, as seen in
might be explained by the fact that the Cftr2/2 mice receiving vitamin D–deﬁcient rickets (20). PegLyte treatment therefore
PegLyte did not display overt gastrointestinal or respiratory had an additional indirect beneﬁcial effect on bone by allowing
Figure 3. Trabecular bone cell numbers and
cortical bone width of femur of 12- and 28-
week-old control and Cftr2/2 mice. Osteo-
blasts were counted on representative von
Kossa– and toluidine blue–stained sections,
and osteoclasts on representative tartrate-re-
sistant acid phosphatase–stained sections, of
plastic-embedded femoral bone from three or
four female mice per group (A). Cortical width
was measured at four points at the midpoint
of the femoral diaphysis of three female mice
per group (B). Values are expressed as means
6 SD. *P , 0.05, signiﬁcantly different from
control data. CF 5 cystic ﬁbrosis (Cftr2/2
mice); OB 5 osteoblasts; OC 5 osteoclasts.
314 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 177 2008
for improved absorption of calcium and vitamin D, which may be data, the 28-week-old Cftr2/2 mice demonstrated nearly normal
compromised in patients with CF (21). vertebral BMD. This discrepancy is most likely due to the
absence of steroidal treatment in the mice and its presence as
Improvement of Osteopenia over Time in BALB Cftr2/2 Mice a confounding variable in the interpretation of the clinical
In the current study we show that the bone phenotype of 12- ﬁndings. The increase in osteoclast numbers and the absence
week-old (juvenile) BALB Cftr2/2 mice resembles that of the of excessive osteoid (osteomalacia) are common features of CF
3-week-old (weanling) UNC Cftr2/2 mice studied by Dif and and adult Cftr2/2 mouse phenotypes.
coworkers. The defects in Cftr2/2 mice included low BMD; At 12 weeks of age, there was a signiﬁcant reduction in
reduced cortical bone width; and fewer, thinner trabeculae circulating IGF-I in Cftr2/2 mice compared with control mice.
compared with control littermates. The osteopenic phenotype This deﬁciency was coincident with the largest discrepancy in
of 3-week-old Cftr2/2 mice was attributed to a decrease in bone- body weight and bone mass between control and Cftr2/2 mice.
forming surfaces and in osteoclast numbers in trabecular bone. IGF-I is a major regulator of somatic cell and bone growth and
However, the mixed genetic background (14–16) and the has been identiﬁed as an independent predictor of low BMD in
immature bone age may have been confounding factors in the patients with CF (23) and compromised growth in Cftr2/2 mice
interpretation of data in the Dif study. In 28-week-old Cftr2/2 (24). Work has used genome-wide scanning of F2 mice derived
mice, osteopenia was seen in the presence of increased numbers from C57BL and C3H strains, with low and high BMD,
of trabecular osteoblasts and osteoclasts, suggesting a high respectively, to identify quantitative trait loci linked to both
turnover defect. Alternatively, the relative increase in osteo- serum IGF-I and bone acquisition (25). Adult female congenic
blasts and osteoclasts in the 28-week-old mice could reﬂect mice carrying a chromosome 10 locus from C3H on a C57BL
a ‘‘catch-up’’ phenomenon, resulting from normalization of background had signiﬁcantly higher circulating levels of IGF-I
IGF-I levels, as discussed below. Studies examining dynamic in association with higher femoral BMD and trabecular num-
labeling of mineralization fronts and including in-depth histo- ber. In BALB Cftr2/2 mice, there was a signiﬁcant increase in
morphometric analyses will be required to resolve this issue. serum IGF-I between 12 and 28 weeks and a concomitant
Cortical thinning in 3-week-old Cftr2/2 mice (11) was increase in BMD and trabecular number, suggesting that the
attributed to a relative increase in cortical osteoclasts, although late-onset ‘‘growth spurt’’ in Cftr2/2 mice could have been
the cortical osteoblast surfaces were not documented. In 12- mediated by circulating IGF-I. Of additional interest in this
week-old Cftr2/2 mice, despite the signiﬁcant reduction in respect are the reports of reduced peroxisome proliferator–
cortical width, we could not detect any signiﬁcant differences activated receptor-g in Cftr2/2 mice (26) and the negative
in the number of osteoclasts or osteoblasts in cortical bone regulation of IGF-I by activation of this receptor (27). It is
compared with the control mice. The cortical width of Cftr2/2 conceivable that the rise in IGF-I at 28 weeks was related to
bones in the study by Dif and coworkers was approximately an age-dependent downregulation of peroxisome proliferator–
50% that of control mice at 3 weeks of age, compared with 75% activated receptor-g in Cftr2/2 mice.
at 12 weeks and 85% at 28 weeks of age in our own work. These The precise molecular mechanisms by which mutations in
observations, together with those documenting a gradual in- CFTR can alter bone cell function remain to be elucidated,
crease in trabecular and cortical BMD and improvement in although insight may be gained from studies of other chloride
bone architectural parameters between 12 and 28 weeks of age, channels implicated in defective bone metabolism. For exam-
suggest that attainment of peak bone mass may have been ple, loss of function of the chloride channel-7 (ClC-7) results in
delayed in Cftr-deﬁcient mice. However, it is also important to a defect in chloride conductance required for efﬁcient proton
note that trabecular bone volume remained almost 50% lower pumping by the H1-ATPase of the osteoclasts (28). The
in Cftr2/2 mice at 28 weeks, with reductions in the number and catabolic activity of osteoclasts depends on the active transport
thickness of individual trabeculae and an increase in the ratio of of hydrogen ions into the extracellular space adjacent to bone to
platelike to rodlike structures measured. These features, along solubilize bone mineral. Loss of chloride channel-7 activity
with the continued elevation in osteoblasts and osteoclasts, therefore leads to accumulation of bone that would otherwise
support the conjecture that bone turnover is increased in Cftr2/2 be resorbed, or to osteopetrosis (28). Similarly, Kajiya and
mice compared with age-matched control mice. coworkers (29) reported that the K1/Cl2 cotransporter-1 ex-
The trajectory of weight gain and bone accrual in control pressed in mouse osteoclasts also participates in H1 extrusion
mice was greatest between 8 and 12 weeks of age, which would during bone resorption, thus providing a second example of
roughly correspond with childhood and adolescence in humans, a chloride channel involved in the catabolic activity of osteo-
and then slowed thereafter. In contrast, Cftr2/2 mice gained clasts. In our current work, bone acquisition continued in mice
little weight or bone mass between 8 and 12 weeks whereas up to 28 weeks of age in the presence of a signiﬁcant increase in
signiﬁcant increases in both were observed between 12 and 28 osteoclast and osteoblast activity. It is therefore unlikely that
weeks. This failure to keep pace with age-matched control mice there was a signiﬁcant defect in either cell type, although CFTR
recapitulates the clinical situation, in which low bone mass in protein has been shown to be present in both (30), but rather
adult patients with CF has been attributed to a failure to thrive that a change in the bone microenvironment inﬂuenced their
and accrue bone during childhood and adolescence (7). Sup- activity. Additional in vitro and molecular studies are required
porting this, through histomorphometric analyses, Elkin and to resolve this issue.
coworkers (5) revealed bone harvested from adult patients with
CF with low BMD to have reduced trabecular bone volume Conﬂict of Interest Statement: None of the authors has a ﬁnancial relationship
compared with age- and sex-matched healthy control subjects, with a commercial entity that has an interest in the subject of this manuscript.
which was ascribed to low bone formation. Furthermore,
a second study, in which histomorphometric analysis of bone
specimens taken at autopsy from posttransplantation patients References
treated with corticosteroids was completed, revealed the exis- 1. Davis PB, Drumm M, Konstan MW. Cystic ﬁbrosis. Am J Respir Crit
tence of signiﬁcant cortical and trabecular osteopenia in the CF Care Med 1996;154:1229–1256.
specimens, in association with decreased osteoblastic and in- 2. Aris RM, Merkel PA, Bachrach LK, Borowitz DS, Boyle MP, Elkin SL,
creased osteoclastic activity (22). In contrast to these clinical Guise TA, Hardin DS, Haworth CS, Holick MF, et al. Guide to bone
Haston, Li, Li, et al.: Bone Phenotype of Adult Cftr2/2 Mice 315
health and disease in cystic ﬁbrosis. J Clin Endocrinol Metab alization and osteopenia in young adult FGFR32/2 mice. Hum Mol
2005;90:1888–1896. Genet 2004;13:271–284.
3. McDonnell P, McHugh PE, O’Mahoney D. Vertebral osteoporosis and 19. Richard S, Torabi N, Franco GV, Tremblay GA, Chen T, Vogel G,
trabecular bone quality. Ann Biomed Eng 2007;35:170–189. Morel M, Cleroux P, Forget-Richard A, Komarova S, et al. Ablation
4. Wehren LE, Siris ES. Beyond bone mineral density: can existing clinical of the Sam68 RNA binding protein protects mice from age-related
risk assessment instruments identify women at increased risk of bone loss. PLoS Genet 2005;1:e74.
osteoporosis? J Intern Med 2004;256:375–380. 20. Holick MF. Resurrection of vitamin D deﬁciency and rickets. J Clin
5. Elkin SL, Vedi S, Bord S, Garrahan NJ, Hodson ME, Compston JE. Invest 2006;116:2062–2072.
Histomorphometric analysis of bone biopsies from the iliac crest of adults 21. Haworth CS, Jones AM, Adams JE, Selby PL, Webb AK. Randomised
with cystic ﬁbrosis. Am J Respir Crit Care Med 2002;166:1470–1474. double blind placebo controlled trial investigating the effect of
6. Shead EF, Haworth CS, Gunn E, Bilton D, Scott MA, Compston JE. calcium and vitamin D supplementation on bone mineral density
Osteoclastogenesis during infective exacerbations in patients with and bone metabolism in adult patients with cystic ﬁbrosis. J Cyst
cystic ﬁbrosis. Am J Respir Crit Care Med 2006;174:306–311. Fibros 2004;3:233–236.
7. Buntain HM, Schluter PJ, Bell SC, Greer RM, Wong JC, Batch J, 22. Haworth CS, Webb AK, Egan JJ, Selby PL, Hasleton PS, Bishop PW,
Lewindon P, Wainwright CE. Controlled longitudinal study of bone Freemont TJ. Bone histomorphometry in adult patients with cystic
mass accrual in children and adolescents with cystic ﬁbrosis. Thorax ﬁbrosis. Chest 2000;118:434–439.
2006;61:146–154. 23. Gordon CM, Binello E, LeBoff MS, Wohl ME, Rosen CJ, Colin AA.
8. King SJ, Topliss DJ, Kotsimbos T, Nyulasi IB, Bailey M, Ebeling PR, Relationship between insulin-like growth factor I, dehydroepiandros-
Wilson JW. Reduced bone density in cystic ﬁbrosis: DF508 mutation is terone sulfate and proresorptive cytokines and bone density in cystic
an independent risk factor. Eur Respir J 2005;25:54–61. ﬁbrosis. Osteoporos Int 2006;17:783–790.
9. Sermet-Gaudelus I, Souberbielle JC, Ruiz JC, Vrielynck S, Heuillon B, 24. Rosenberg LA, Schluchter MD, Parlow AF, Drumm ML. Mouse as a
Azhar I, Cazenave A, Lawson-Body E, Chedevergne F, Lenoir G. model of growth retardation in cystic ﬁbrosis. Pediatr Res 2006;59:
Low bone mineral density in young children with cystic ﬁbrosis. Am J 191–195.
Respir Crit Care Med 2007;175:951–957. 25. Delahunty KM, Shultz KL, Gronowicz GA, Koczon-Jaremko B, Adamo
10. Davidson DJ, Rolfe M. Mouse models of cystic ﬁbrosis. Trends Genet ML, Horton LG, Lorenzo J, Donahue LR, Ackert-Bicknell C, Kream
2001;17:S29–S37. BE, et al. Congenic mice provide in vivo evidence for a genetic locus
11. Dif F, Marty C, Baudoin C, de Vernejoul MC, Levi G. Severe that modulates serum insulin-like growth factor-I and bone acquisi-
osteopenia in CFTR-null mice. Bone 2004;35:595–603. tion. Endocrinology 2006;147:3915–3923.
12. Gawenis LR, Spencer P, Hillman LS, Harline MC, Morris JS, Clarke LL. 26. Ollero M, Junaidi O, Zaman MM, Tzameli I, Ferrando AA, Andersson
Mineral content of calciﬁed tissues in cystic ﬁbrosis mice. Biol Trace C, Blanco PG, Bialecki E, Freedman SD. Decreased expression of
Elem Res 2001;83:69–81. peroxisome proliferator activated receptor g in CFTR2/2 mice. J Cell
13. Snouwaert J, Brigman KK, Latour AM, Malouf NN, Boucher RC, Physiol 2004;200:235–244.
Smithies O, Koller B. An animal model for cystic ﬁbrosis made by 27. Lecka-Czernik B, Ackert-Bicknell C, Adamo ML, Marmolejos V,
gene targetting. Science 1992;257:1083–1088. Churchill GA, Shockley KR, Reid IR, Grey A, Rosen CJ. Activation
14. Haston CK, Cory S, Lafontaine L, Dorion G, Hallett MT. Strain- of peroxisome proliferator-activated receptor g (PPARg) by rosigli-
dependent pulmonary gene expression proﬁles of a cystic ﬁbrosis tazone suppresses components of the insulin-like growth factor
mouse model. Physiol Genomics 2006;25:336–345. regulatory system in vitro and in vivo. Endocrinology 2007;148:
15. Haston CK, Corey M, Tsui LC. Mapping of genetic factors inﬂuencing 903–911.
the weight of cystic ﬁbrosis knockout mice. Mamm Genome 2002;13: 28. Kornak U, Kasper D, Bosl MR, Kaiser E, Schweizer M, Schulz A,
614–618. Friedrich W, Delling G, Jentsch TJ. Loss of the ClC-7 chloride
16. Haston CK, McKerlie C, Newbigging S, Corey M, Rozmahel R, Tsui LC. channel leads to osteopetrosis in mice and man. Cell 2001;104:
Detection of modiﬁer loci inﬂuencing the lung phenotype of cystic 205–215.
ﬁbrosis knockout mice. Mamm Genome 2002;13:605–613. 29. Kajiya H, Okamoto F, Li JP, Nakao A, Okabe K. Expression of mouse
17. Clarke LL, Gawenis LR, Franklin CL, Harline MC. Increased survival of osteoclast K–Cl co-transporter-1 and its role during bone resorption.
CFTR knockout mice with an oral osmotic laxative. Lab Anim Sci J Bone Miner Res 2006;21:984–992.
1996;46:612–618. 30. Shead EF, Haworth CS, Condliffe AM, McKeon DJ, Scott MA,
18. Valverde-Franco G, Liu H, Davidson D, Chai S, Valderrama-Carvajal Compston JE. Cystic ﬁbrosis transmembrane conductance regulator
H, Goltzman D, Ornitz DM, Henderson JE. Defective bone miner- (CFTR) is expressed in human bone. Thorax 2007;62:650–651.