Kidney International, Vol. 63, Supplement 85 (2003), pp. S14–S18

Influence of polymorphisms in VDR and COLIA1 genes onthe risk of osteoporotic fractures in aged men

DANIEL ALVAREZ-HERNANDEZ, MANUEL NAVES, J. BERNARDINO DIAZ-LOPEZ, CARLOS GOMEZ,INIGO SANTAMARIA, and JORGE B. CANNATA-ANDIA

Bone and Mineral Research Unit, Instituto Reina Sofıa de Investigacion, Hospital Universitario Central de Asturias,Oviedo, Spain

Influence of polymorphisms in VDR and COLIA1 genes on [2, 3]. In addition, other candidate genes, such as collagenthe risk of osteoporotic fractures in aged men. type I, have been studied [4, 5]; however, the effect of

Background. Osteoporosis in chronic renal failure is a com- these polymorphisms seems to be less strong that themon finding caused by several factors, including age. In theinitial studies reported [6, 7]. Various suggestions havelast decade, the likely effect of genetic markers related withbeen made as to why these discrepancies exist, includingthe appearance and evolution of osteoporosis has been mainly

studied in women, with no categorical results. The aim of this ethnic, racial, or environmental differences among thestudy was to assess the influence of polymorphisms of the populations studied. Other possible explanations includevitamin D receptor (VDR) and COLIA1 genes on the risk of

an inadequate population sample, therefore resulting inosteoporotic fractures in men older than 50 years.a lack of statistical data in order to detect the discrepan-Methods. The study population comprised 156 men, aged

64 � 9 (50–86), randomly selected from the population list of cies, or, in many cases, the absence of a proper popula-Oviedo, Spain. Prevalent vertebral fractures and incident non- tion selection criterion.vertebral fractures were identified, as well as several genetic Although it has always been recognized that osteopo-polymorphisms. Prevalent vertebral fractures were considered

rosis has a higher prevalence in women, in recent yearsaccording to the Genant grade 2 classifications. The analyzedmany studies have begun focusing on its effect on men.genetic polymorphisms were located on restriction sites BsmI

(B,b), ApaI (A,a), and TaqI (T,t) in the VDR and on Sp1 (S,s) The few studies which have dealt with genetic poly-in COLIA1. morphisms and fractures in men have been carried out

Results. Although none of the VDR gene polymorphismsin populations in which the number of subjects was lowseparately analyzed showed any differences between fractured[8, 9]. Only one study was performed with a suitableand non-fractured men, the utilization of haplotypes could be

employed in order to find osteoporotic fractures in men. By number of people [10]; however, it focused on the riskcontrast, the COLIA1 polymorphism was associated with os- of incident fractures.teoporotic fractures. The percentage of prevalent vertebral The aim of this study was to evaluate the effect of thefractures was significantly higher in the “ss” genotype with

VDR and collagen type I gene polymorphisms on therespect to the other genotypes. These results show that in men,prevalence and incidence of osteoporotic fractures in athe “ss” genotype of COLIA1 polymorphism could be the best

osteoporotic fracture risk genetic predictor, independent of male study group monitored over an eight-year follow-bone mass values. up period.

METHODSReduction in bone mass in chronic renal failure occursdepending upon the stage of renal failure and also on The study group consisted of 156 men aged 50 and above,osteoporosis caused by age. Many candidate genes are mean age 64 � 9 (range, 50–86), randomly selected frominvolved in the onset of osteoporosis. Since an initial the population list of Oviedo who had previously partici-report finding a relationship between the vitamin D re- pated in the European Vertebral Osteoporosis Studyceptor (VDR) gene polymorphisms and bone mass [1], (EVOS) [11]. They underwent several biochemical markera number of studies subsequently carried out using this tests and bone mineral density measurements. Each par-gene polymorphism have reached differing conclusions ticipant had two lateral radiographs of the dorsal and

lumbar spinal area to check for any vertebral fractures.Vertebral fracture was defined using the Genant grade 2Key words: genetic, bone, vertebral fracture, incident non-vertebral

fracture. method [12]. This method uses more restrictive criteriathan Genant grade 1, therefore minimizing the presence of 2003 by the International Society of Nephrology

S-14

Alvarez-Hernandez et al: Polymorphisms and osteoporotic fractures in men S-15

fractures caused by trauma, which a radiographic analysis (New England Biolabs, Beverly, MA, USA) and electro-phoresed in 2% agarose.alone will not always accurately determine. In short, all

vertebrae between T4 and L4 were diagnosed as being For the statistical analysis of the results, the propor-tional fracture of the different polymorphisms was car-fractured when a reduction of at least 25% in the anterior,

medium, or posterior length was detected. The subject ried out using the chi-square test. Likewise, the relativerisks of both prevalent vertebral fracture and incidentalwas only diagnosed as having an osteoporotic vertebral

fracture when a wedge, biconcavity, or crush of at least nonvertebral fracture were calculated, with 95% confi-dence interval in each of the studied alleles, respectively.grade 2 in one of the 13 analyzed vertebrae was present.

All subjects were prospectively monitored by four Statistical significance was considered to be P � 0.05.postal questionnaires over an eight-year period in orderto find out if any incident osteoporotic fracture of non-

RESULTSvertebral site had occurred.No differences between the different gene polymor-A peripheral blood sample was taken from all subjects

phisms and age, body mass index, or lumbar and neckso that DNA could be later extracted and analyzed tobone mineral density (BMD) were found (Table 1).determine the different VDR and collagen type I gene

The presence of prevalent vertebral fractures was notpolymorphisms.associated with any of the three VDR gene polymor-A fragment of 1365 bp of the VDR gene (spanningphisms studied (Table 2); however, baT and BAt haplo-intron 8 and exon 9) containing BsmI, ApaI, and TaqItypes showed less vertebral fracture than the bAT haplo-target sites was then amplified by polymerase chain reac-type (P � 0.023).tion (PCR) using the following primers: forward: GCA

The “ss” genotype of the collagen type I gene polymor-GAGTGTGCAGGCGATTCG and reverse: TGAAGphism was associated with a high frequency of prevalentCTCGTGTTCCCCGCAAT in a 30 �L reaction volume.vertebral fracture when compared with the rest of thePCR conditions were 30 seconds at 95�C, 15 seconds atgenotypes (P � 0.003) (Table 3). The presence of the “ss”64�C, and 25 seconds at 72�C for 30 cycles. The PCRgenotype increased the risk of vertebral fracture by aproduct was divided into three different aliquots andfactor of 6 and 14, respectively, compared with SS anddigested with the BsmI, ApaI, and TaqI restriction en-Ss genotypes. When the prevalent vertebral fracture waszymes (Stratagene, La Jolla, CA, USA). The samplesassociated with incident nonvertebral fracture, which oc-were then electrophoresed in a 2% agarose gel. Thecurred during the follow-up period, this relative risk fac-absence of the target site for the three endonucleasestor rose to 6.6 and 21, respectively. Incident nonvertebralproduced a single band of 1365 bp (alleles B, A, and Tfractures were not associated with the collagen type Ifor BsmI, ApaI, and TaqI, respectively). The allele “b”

(BsmI) produced two bands of 113 and 1252 bp, the gene polymorphism.allele “a” (ApaI) 284 and 1081 bp, and the allele “t”(TaqI) 208 and 1157 bp. To avoid technical and analytic

DISCUSSIONproblems, the digestion of the three recognition sites wasThe genotypic distribution of the different VDR geneperformed individually and the complete digestion of

polymorphisms has proved to be similar to other Cauca-the PCR products for each of the restriction enzymessian populations [6, 13]. Although the COLIA1 geneused was confirmed.polymorphism showed a slightly higher proportion ofThe three VDR gene polymorphisms were grouped“ss” subjects compared with other populations [8, 14–16],into the most common haplotypes (more than 95% ofthe random recruitment of subjects minimizes any selec-the sample): baT, bAT, and BAt. To obtain the differenttion bias in this population.haplotypes we combined the BB and Bb genotypes for

In our study of males with a mean age similar to that“B” allele, AA and Aa for “A” allele, and Tt and tt forof our dialysis population, the presence of a specific geno-“t” allele.type of collagen type I gene polymorphism was relatedA guanine-to-thymidine polymorphism in the Sp1with a high frequency of fractures. However, this rela-binding site in the COLIA1 gene was determined bytionship was not observed in any of the different VDRPCR using a mismatched primer to introduce a diallelicgene polymorphisms analyzed. Langdahl et al [8] notedMsc1 restriction site, as previously described [4]. Underthat the COLIA1 Sp1 polymorphism was significantlythese conditions, the assay discriminates two alleles, Gassociated with osteoporotic vertebral fractures in men,and T, which correspond to the presence of guanineand the effect on fracture risk appears to be independent,and thymidine, respectively, as the first base in the Sp1-in part, of BMD.binding site in the first intron of the COLIA1 gene. The

Some studies in women using a population of limitedPCR conditions were 15 seconds at 95�C, 15 seconds atsize failed to detect a significant difference between68�C, and 20 seconds at 72�C for 35 cycles. The PCR

product was digested with the Van91 I restriction enzyme COLIA1 alleles and bone mass or osteoporotic fractures,

Alvarez-Hernandez et al: Polymorphisms and osteoporotic fractures in menS-16

Table 2. Distribution of BsmI, ApaI, and TaqI genotypes and VDRhaplotypes in fractured and non-fractured men

Fractured Non-fractured

Genotypes BsmIBB 3 (12.0%) 22 (88.0%)Bb 9 (13.2%) 68 (96.8%)bb 8 (15.4%) 44 (84.6%) �2 �0.818

Genotypes ApaIAA 4 (10.8%) 33 (89.2%)Aa 12 (16.7%) 60 (83.3%)aa 1 (4.0%) 24 (96.0%) �2 �0.240

Genotypes TaqITT 7 (14.9%) 40 (85.1%)Tt 7 (10.4%) 60 (89.6%)tt 7 (15.0%) 17 (85.0%) �2 �0.738

VDR haplotypesbaT 1 (4.2%) 23 (95.8%)bAT 6 (31.6%) 13 (68.4%)BAt 10 (11.6%) 76 (88.4%) �2 �0.023

Table 3. Distribution of COLIA1 genotypes in fracturedand non-fractured men

Fractured Non-fractured

GenotypesSS 10 (11.8%) 75 (88.2%)Ss 2 (5.4%) 35 (94.6%)ss 5 (45.5%) 6 (54.5%)

�2 (genotypes) �2 �11.42Significance P�0.003Prevalent vertebral fracture

ss vs. SSa 6.3 (1.6–24.3)Prevalent vertebral fracture

ss vs. Ssa 14.6 (2.3–92.6)Incident osteoporotic fractures

ss vs. SSa 5.2 (0.4–66.7)Incident osteoporotic fractures

ss vs. Ssa 5.9 (0.3–121.9)Prevalent vertebral fracture �

incident osteoporotic fracturess vs. SSa 6.6 (1.8–25.0)

Prevalent vertebral fracture �incident osteoporotic fracture

ss vs. Ssa 21.0 (3.3–133.3)a Relative risk and 95% confidence interval

however [17–20]. It is currently unclear whether the dif-ferences between different studies represent true hetero-geneity between populations in terms of the contributionthat COLIA1 alleles make to regulation of bone massand fracture risk, or to other factors, such as small samplesize, population admixture, or patient selection.

The collagen type I is the most abundant protein inbone that takes part in the skeleton, and therefore itscoding gene has been considered a possible regulatorof bone mass. Any variations in its composition couldcondition more or less stability, resulting in the appear-ance of fractures. In fact, functional studies have foundthat the unfavorable genotype in the collagen type I genepolymorphism is associated with osteoporosis due to: (1)

Tab

le1.

Cha

ract

eris

tics

ofpo

pula

tion

byV

DR

and

CO

LIA

1ge

nepo

lym

orph

ism

s

BB

(N�

25)

Bb

(N�

77)

bb(N

�52

)A

A(N

�37

)A

a(N

�72

)aa

(N�

25)

TT

(N�

47)

Tt

(N�

67)

tt(N

�20

)SS

(N�

85)

Ss(N

�37

)ss

(N�

11)

Age

year

s63

�9

65�

864

�9

66�

964

�8

65�

964

�9

66�

962

�8

65�

964

�8

61�

8B

MI

kg/m

227

.9�

3.1

27.4

�3.

327

.3�

3.5

27.6

�2.

727

.5�

3.6

27.3

�3.

527

.3�

3.6

27.3

�3.

328

.7�

2.8

27.5

�3.

527

.6�

3.0

26.9

�3.

6B

MD

lum

bar

g/cm

21.

072

�0.

192

1.01

3�

0.15

11.

010

�0.

156

1.04

4�

0.19

01.

017

�0.

162

1.03

0�

0.12

61.

020

�0.

157

1.01

9�

0.17

71.

065

�0.

127

1.07

16�

0.16

71.

060

�0.

131

0.99

3�

0.20

0B

MD

neck

g/cm

20.

854

�0.

146

0.80

3�

0.11

80.

804

�0.

123

0.82

5�

0.14

90.

816

�0.

119

0.80

5�

0.13

10.

800

�0.

135

0.81

8�

0.11

20.

850

�0.

159

0.80

1�

0.12

70.

849

�0.

140

0.82

0�

0.11

5

Abb

revi

atio

nsar

e:B

MI,

body

mas

sin

dex;

BM

D,b

one

min

eral

dens

ity;

VD

R,v

itam

inD

rece

ptor

.

evidence of allele-specific differences in binding of the

Alvarez-Hernandez et al: Polymorphisms and osteoporotic fractures in men S-17

igaciones Sanitarias (FIS 94/1901-E), and by Fundacion Renal InigoSp1 protein to the polymorphic recognition site; (2) dif-Alvarez de Toledo (Spain).

ferences in allele-specific transcription; (3) differences in Daniel Alvarez-Hernandez was supported by the Fundacion RenalInigo Alvarez de Toledo, Instituto Reina Sofıa de Investigacion, and bycollagen protein production with an increase in collagenthe FICYT (Fundacion para el Fomento en Asturias de la Investigacion(type I) chain �1 compared with the collagen chain �2;Cientıfica Aplicada y la Tecnologıa). Dr. Inigo Santamarıa is a recipient

and (4) differences in bone strength in samples derived of a Research Contract from Fondo Investigaciones Sanitarias (FIS00/3161).from patients of different genotype [7, 21].

The authors wish to thank Dr. J.R. Jimenez Blanco (Department ofIn keeping with previous studies [10], we found a smallRadiology) for technical support and language consultants Covadonga

effect of the “s” allele on new fractures. However, the Dıaz Dıaz and Francesca Pieraccini.low number of fractures that occurred and also the power

Reprint requests to Dr. Jorge Cannata Andıa, Bone and Mineralto detect any effect was limited.Research Unit, Instituto Reina Sofıa de Investigacion, Hospital Central

On the other hand, vitamin D plays a central role in de Asturias, C/Julian Claverıa s/n, 33006 Oviedo, Spain.E-mail: metoseo@hca.escalcium homoeostasis by regulating calcium absorption,

bone resorption, bone cell differentiation, and parathy-REFERENCESroid hormone secretion, so VDR could be a possible can-

didate gene for the regulation of bone mass. Although 1. Morrison NA, Qi JC, Tokita A, et al: Prediction of bone densityfrom vitamin D receptor alleles. Nature 367:284–287, 1994the contribution of VDR to genetic regulation of bone

2. Houston LA, Grant SF, Reid DM, Ralston SH: Vitamin Dmass has been disputed [22], the association between receptor polymorphism, bone mineral density, and osteoporoticVDR and bone mass may be masked in certain popula- vertebral fracture: Studies in a UK population. Bone 18:249–252,

1996tions, caused by an interaction with environmental fac-3. Hustmyer FG, Peacock M, Hui S, et al: Bone mineral density intors such as calcium intake or vitamin D status [23, 24]. relation to polymorphism at the vitamin D receptor gene locus. J

Meanwhile, a previous study performed in the same Clin Invest 94:2130–2134, 19944. Grant SF, Reid DM, Blake G, et al: Reduced bone density andcohort showed a relationship between VDR gene poly-

osteoporosis associated with a polymorphic Sp1 binding site in themorphism and bone mass in women [25], but no effect collagen type I alpha 1 gene. Nat Genet 14:203–205, 1996was observed in men. No association between fractures 5. Uitterlinden AG, Burger H, Huang Q, et al: Relation of alleles

of the collagen type Ialpha1 gene to bone density and the risk ofand the three VDR gene polymorphisms separatelyosteoporotic fractures in postmenopausal women. N Engl J Medanalyzed was found [9]. However, when the polymor- 338:1016–1021, 1998

phisms were grouped into haplotypes, the bAT haplo- 6. Cooper GS, Umbach DM: Are vitamin D receptor polymorphismsassociated with bone mineral density? A meta-analysis. J Bonetype was associated with vertebral fractures. Other au-Miner Res 11:1841–1849, 1996thors using bone mass rather than fractures [26] found, 7. Mann V, Hobson EE, Li B, et al: A COL1A1 Sp1 binding site

similar to us, that the bAT haplotype was associated with polymorphism predisposes to osteoporotic fracture by affectingbone density and quality. J Clin Invest 107:899–907, 2001a reduced BMD.

8. Langdahl BL, Ralston SH, Grant SF, Eriksen EF: An Sp1The small frequency of people in the normal popula- binding site polymorphism in the COLIA1 gene predicts osteopo-

rotic fractures in both men and women. J Bone Miner Res 13:1384–tion with the unfavorable “ss” genotype could be a limita-1389, 1998tion in this study. However, the wide proportion between

9. Langdahl BL, Gravholt CH, Brixen K, Eriksen EF: Polymor-fractured patients with this genotype compared with the phisms in the vitamin D receptor gene and bone mass, bone turn-

over and osteoporotic fractures. Eur J Clin Invest 30:608–617, 2000other genotypes and the fact that this association was10. McGuigan FE, Armbrecht G, Smith R, et al: Prediction of osteo-also found in a group of women that participated in the

porotic fractures by bone densitometry and COLIA1 genotyping:same study, would exclude a casual finding. A prospective, population-based study in men and women. Osteo-

poros Int 12:91–96, 2001In summary, although none of the VDR gene polymor-11. O’Neill TW, Felsenberg D, Varlow J, et al: The prevalence ofphisms separately analyzed (ApaI, BsmI, and TaqI)

vertebral deformity in european men and women: The Europeanshowed any relationship with fractures, the utilization Vertebral Osteoporosis Study. J Bone Miner Res 11:1010–1018, 1996

12. Genant HK, Wu CY, van Kuijk C, Nevitt MC: Vertebral fractureof haplotypes could be employed in order to find osteo-assessment using a semiquantitative technique. J Bone Miner Resporotic fractures in men. In males, the “ss” genotype of8:1137–1148, 1993

COLIA1 polymorphism could be the best osteoporotic 13. Marco MP, Martinez I, Amoedo ML, et al: Vitamin D receptorgenotype influences parathyroid hormone and calcitriol levels infracture risk genetic predictor, independent of bone masspredialysis patients. Kidney Int 56:1349–1353, 1999values. It would be of interest to know if the same find- 14. Garnero P, Borel O, Grant SF, et al: Collagen Ialpha1 Sp1

ings could be applied to our recently published results polymorphism, bone mass, and bone turnover in healthy Frenchpremenopausal women: The OFELY study. J Bone Miner Resconcerning male dialysis patients the same age as the13:813–817, 1998general population [27]. 15. Keen RW, Woodford-Richens KL, Grant SF, et al: Associationof polymorphism at the type I collagen (COL1A1) locus withreduced bone mineral density, increased fracture risk, and in-ACKNOWLEDGMENTScreased collagen turnover. Arthritis Rheum 42:285–290, 1999

This study has been supported in part by European Vertebral Osteo- 16. Weichetova M, Stepan JJ, Michalska D, et al: COLIA1 polymor-porosis Study (EVOS); European Community (1.991-1.993), by Euro- phism contributes to bone mineral density to assess prevalent wristpean Prospective Osteoporosis Study (EPOS); European Community, fractures. Bone 26:287–290, 2000

17. Aerssens J, Dequeker J, Peeters J, et al: Polymorphisms of theBIOMED 93-95; BMHI-CT 092-0182 (1.993-1.997) by Fondo de Invest-

Alvarez-Hernandez et al: Polymorphisms and osteoporotic fractures in menS-18

VDR, ER and COLIA1 genes and osteoporotic hip fracture in supplementation on the bone mineral density of the femoral neckis associated with vitamin D receptor genotype. J Bone Miner Reselderly postmenopausal women. Osteoporos Int 11:583–591, 2000

18. Hustmyer FG, Liu G, Johnston CC, et al: Polymorphism at an 12:1241–1245, 199724. Ferrari S, Rizzoli R, Chevalley T, et al: Vitamin-D-receptor-Sp1 binding site of COL1A1 and bone mineral density in premeno-

pausal female twins and elderly fracture patients. Osteoporos Int gene polymorphisms and change in lumbar-spine bone mineraldensity. Lancet 345:423–424, 19959:346–350, 1999

19. Liden M, Wilen B, Ljunghall S, Melhus H: Polymorphism at 25. Gomez C, Naves ML, Barrios Y, et al: Vitamin D receptor genepolymorphisms, bone mass, bone loss and prevalence of vertebralthe Sp 1 binding site in the collagen type I alpha 1 gene does not

predict bone mineral density in postmenopausal women in Sweden. fracture: Differences in postmenopausal women and men. Osteo-poros Int 10:175–182, 1999Calcif Tissue Int 63:293–295, 1998

20. Lander ES, Schork NJ: Genetic dissection of complex traits. Sci- 26. Uitterlinden AG, Pols HA, Burger H, et al: A large-scale popu-lation-based study of the association of vitamin D receptor geneence 265:2037–2048, 1994

21. Qureshi AM, McGuigan FE, Seymour DG, et al: Association polymorphisms with bone mineral density. J Bone Miner Res11:1241–1248, 1996between COLIA1 Sp1 alleles and femoral neck geometry. Calcif

Tissue Int 69:67–72, 2001 27. Rodrıguez-Garcıa M, Gomez C, Dıaz Lopez JB, et al: Prevalenciade fracturas vertebrales y calcificaciones aorticas en pacientes en22. Peacock M: Vitamin D receptor gene alleles and osteoporosis: A

contrasting view. J Bone Miner Res 10:1294–1297, 1995 hemodialisis: Comparacion con una poblacion de la misma edady sexo. Nefrologia, 200323. Graafmans WC, Lips P, Ooms ME, et al: The effect of vitamin D