- Email: [email protected]
1 What is the evidence that osteoarthritis is genetically determined?
1 What is the evidence that osteoarthritis is genetically determined? F l a v i a M . C i c u t t i n i MSc, MBBS, FRACP, PhD Senior Lecturer Department of Epidemiology and Preventive Medicine, Alfred Hospital, Prahran, Victoria, 3181, Australia
Tim D. Spector MSc, MD, FRCP Consultant Rheumatologist
The Twin Research Unit, St Thomas" Hospital, London SE1 7EH, UK
Although for many years it was speculated that osteoarthritis was genetically determined, little data were available to support this contention. A major problem with early work was a lack of consistency in the definition of osteoarthritis. Based on a radiographical definition of osteoarthritis, which is currently the optimal method for epidemiological and genetic studies, data from a recent twin study have provided an estimate of the hereditable component of osteoarthritis to be in the order of 50 to 65%. In addition, sophisticated molecular biology techniques are being increasingly used to explore potential genetic abnormalities in cartilage and matrix components in osteoarthritis. These exciting new data are examined as we address the role of genetic factors in osteoarthritis. Key words: osteoarthritis; genetics: inheritance.
Osteoarthritis (OA) is the most frequent cause of musculoskeletal disability in developed countries. The multifactorial nature of OA is well recognized, with a number of environmental risk factors such as obesity (Felson et al, 1988; Hart and Spencer, 1993), previous injury (Felson et al, 1988) and menisectomy (Cooper et al, 1994) being strongly associated with its development. The role of genetic factors in the development of OA has received considerable attention, aided by the rapidly expanding knowledge of molecular biology. The genetic contribution to a disease can be determined by epidemiological studies of family history, extended family correlations, mother daughter studies, and twins. These methodologies have been used successfully in determining the contribution of genetic factors to another common disease of old age, osteoporosis, the trait of which, bone density, has an estimated heritability of 70 to 90% (Pocock et al, 1987). However, a major difference between osteoporosis and OA is the superior disease definition in Bailliere's Clinical Rheumatolegy-Vol. 11, No. 4, November 1997 ISBN 0-7020-2368-X 0950-3579/97/040657 + 13 $12.00/00
657 Copyright © 1997, by Bailliere Tindall All rights of reproduction in any form reserved
658
F.M. Cicuttini and T. D. Spector
osteoporosis compared with OA, that is, the use of bone densitometry. Although the definition of disease in OA epidemiological research remains a problem, there is a general consensus that radiological changes in populations are superior to clinical signs and symptoms (Spector and Cooper, 1993). The advantages of using X-rays include better reproducibility, reasonable correlation with symptoms and disability, and an objective permanent record which can be assessed blind to clinical details. However, not all studies have used radiology as the definition of OA. This has resulted in some difficulties in the interpretation of results, particularly in determining the family history of OA in studies which rely on patient reporting.
Family studies Heberden's nodes For over 50 years a strong genetic component to certain forms of OA was thought to be present. The familial occurrence of Heberden's nodes of the fingers was first documented by Stecher (1941) who demonstrated that the nodes were three times more common in the sisters of 64 affected subjects as in the general population. In 1944, Stecher et al concluded that these lesions were inherited as a single autosomal dominant gene with a strong female predominance.
Primary generalized OA Primary generalized OA is the most common form of inherited OA. It was first described as a clinical entity by Kellgren and Moore (1952). Primary generalized OA is described by the presence of Heberden's and Bouchard's nodes and premature degeneration of articular cartilage, often in a concentric pattern in many joints. In the early 1960s, family studies of 20 male and 32 female probands with generalized OA, based on the presence of 5 radiologically-affected joint groups, suggested that first degree relatives were twice as likely to also have radiographical generalized disease (Kellgren et al, 1963). The familial occurrence of OA was further supported by a study of 391 cases of OA in which 120 patients, largely middle-aged women, were identified to have a polyarthritis characterized by signs of inflammation and an acute onset of symptoms. A family history of similar joint disease was given by 20% of these patients (Kellgren and Moore, 1952). This, and other studies, suggested a polygenic form of inheritance of primary generalized OA rather than a single gene defect (Kellgren et al, 1963). Further support for a genetic predisposition is suggested by the association with the HLA-AIB8 (Pattrick et al, 1989) and HLA-B8 (Brodsky et al, 1979) haplotypes and with 0d-anti-trypsin isoform patterns
Evidence for genetically-determined OA 659
(Pattrick et al, 1989), although not all studies confirmed this association (Benavides et al, 1985). Familial disorders associated with OA
A number of rare diseases that have a genetic basis are associated with severe, early onset OA. Familial calcium pyrophosphate-deposition disease
Familial calcium pyrophosphate-deposition disease (CPDD), is characterized by the deposition of calcium-containing crystals in joint tissue leading to arthritis-like symptoms. Early-onset CPDD has been described in several large families in which the disease progresses to severe degenerative OA (Ryan and McCarty, 1993). In these families, an autosomal dominant mode of inheritance is observed, with an age at onset between the second and fifth decades of life (Ryan and McCarty, 1993). Hydroxyapatite arthritis
Hydroxyapatite arthritis is another form of inherited crystal deposition disease and is due to the deposition of hydroxyapatite crystals in articular cartilage (Marcos et al, 1981). The clinical features and pathological manifestations of this disease are identical to CPDD, except.for the nature of the crystals deposited. The published pedigrees indicate that this disease also has an autosomal dominant inherited pattern with full penetrance (Hajiroussou and Webley, 1986). Stickler syndrome
The Stickler syndrome, also knOwn as the hereditary arthro-ophthalmopathy, is a familial form of OA with prominent ocular involvement (Stickler et al, 1965). This syndrome is a relatively common autosomal dominant disease (1 in 10 000) characterized by vitreous degeneration, retinal detachment, and premature degenerative joint disease. Several forms of familial OA associated with chondrodysplasia also have an autosomal dominant pattern of inheritance (Zitnan and Sitaj, 1963). Chondrodysplasias
The chondrodysplasias are a group of heterogeneous disorders that are characterized by abnormal cartilage growth (Spranger, 1976). The most common of these conditions is achondroplasia which has a frequency of 1 in 50 000 births and represents the classic form of 'short-limb' dwarfism. Although this often occurs sporadically, it may also be inherited as an autosomal dominant trait, and mutations have recently been found in the fibroblast growth factor gene (Shiang et al, 1994). Although linkage to the
660
F.M. Cicuttini and To D. Spector
COL2AI gene has been demonstrated in some cases, this is not consistent (Anderson et al, 1990).
Epiphyseal dysplasias The epiphyseal dysplasias are a group of disorders characterized by growth of the epiphyses which result in joint deformity and may lead to premature degenerative joint disease (Spranger, 1976). This group of disorders includes autosomal recessive, autosomal dominant, and X-linked patterns of inheritance. The expressions of disease range from death in the perinatal period to less severe forms characterized by short stature. The epiphyseal abnormalities often result in crippling OA in both weightbearing and nonweightbearing joints at a very early age.
Multiple epiphyseal dysplasias Among the epiphyseal dysplasias is a heterogeneous group of conditions known as multiple epiphyseal dysplasia (MED) (Fairbanks, 1947). MED is inherited in an autosomal dominant manner with a high degree of penetrance, and several large families with many affected members have been described. In contrast to other epiphyseal dysplasias, patients with MED do not have ocular or retinal abnormalities. In MED, the predominant clinical manifestations occur in the extremities rather than the spine.
Kniest dysplasia Kniest dysplasia, which has an autosomal dominant pattern of inheritance, is an extremely rare disorder characterized by shortening of the trunk and limbs, flattening of the face and bridge of the nose, protuberance of the eye globes, and severe joint abnormalitie~ (Kniest and Leiber, 1977). The joints are usually very large at birth and continue to enlarge during childhood and early adolescence. The majority of affected individuals develop severe, premature degenerative joint disease, particularly involving the knees and hips. The articular cartilage is soft and has decreased resilience. Histologically, the articular cartilage has large cystic lesions giving the appearance of Swiss cheese. It has been suggested that this disease may result from abnormalities in the processing of type II procollagen, as large inclusions containing the carboxyl-propeptide of type II procollagen have been found in the dilated rough endoplasmic reticulum (Poole et al, 1988).
Twin studies Classical twin studies are a very useful design for determining the extent to which genetic factors may influence the variation in a disease, since twins are uniquely matched for age and many known or unknown environmental
Evidence for genetically-determined OA
661
factors. As monozygotic twins are completely concordant for genetic factors, any intrapair variation is assumed to be due to environmental factors, while dizygotic twins share on average half of their genes so any intrapair variations are assumed to be due to both environmental and genetic factors. Thus a comparison of the similarities of monozygotic and dizygotic twins allows us to estimate the extent to which genetic factors determine variation in the measures of OA. Lawrence commented on an unfinished twin study of 10 pairs which suggested increased concordance in identical pairs at a number of joint sites (Lawrence, 1977). A full scale classical twin study was recently performed in the UK to determine whether radiographical OA of the hand and knee are predominantly genetic and to examine the extent to which genetic factors determine variation in the radiographical measures of OA. A clear genetic influence was demonstrated using 500 unselected female twins aged 45-70 years, who were screened radiologically for OA of the hand and knee (Spector et al, 1996). The correlations of OA disease status were consistently twofold higher in 130 pairs of identical compared with nonidentical twins. The influence of genetic factors was estimated to be between 39 and 65%, independent of known environmental or demographic confounders. This genetic effect was consistent whichever radiographical diagnosis of OA was used (i.e. osteophyte or narrowing) and whether it was defined clinically as knee pain or Heberden's nodes. Both the tibiofemoral and patellofemoral joints of the knee were influenced genetically. Only the proximal interphalangeal joints failed to show a genetic effect. A similar effect of inheritance was seen when the analyses were limited to osteophytes or narrowing scores separately or to the knee joint alone. These results demonstrated a clear genetic effect for primary OA of the hand and knee in women, with a heritability of up to 65% independent of environmental or demographic confounders. However, there was no clear evidence to support generalized OA as a separate genetic entity as the genetic effect for Heberden's nodes, a postulated marker for generalized OA, was similar to that for tibiofemoral and patellofemoral OA. Similar twin studies for hip OA have yet to be performed. However, one study from Scandinavia showed double the rate of hip OA in 289 siblings of 184 hip replacement patients compared with 289 controls (Linberg, 1986).
Possible genetic influence in OA The nature of the genetic influence in OA is speculative and may involve either a structural defect (i.e. collagen) or alterations in cartilage or bone metabolism. Proposed candidate genes in the inheritance of OA include the collagen genes which encode the most abundant proteins in the cartilage, with type II collagen being the commonest. There are several reasons that suggest the failure of the collagenous component of articular cartilage may be responsible for the degeneration of
662
E M. Cicuttini and T. D. Spector
joint tissue in OA. First, collagen is the main component of the organic matrix of articular cartilage and plays an important role in the maintenance of the biomechanical properties of cartilage (Kempson et al, 1970). Second, the normal assembly of cartilage collagen serves as a mechanical constraint to prevent the expansion of the highly-hydrated proteoglycans into the large hydrodynamic domains that are characteristic of proteoglycans free in solution (Urban et al, 1979). A failure of this would result in increased hydration of the tissue, softening of the matrix, and cartilage degradation. Finally, it has become apparent in recent years that articular cartilage contains a remarkable degree of molecular complexity with regard to the number of biochemical species of collagen (Mayne, 1989).
A role for type II procollagen in OA There are increasing numbers of mutations described in the COL2A1 gene for type II procollagen, which encodes the main collagenous component of articular cartilage. Genetic linkage analysis of large kindred populations were performed to examine the involvement of candidate genes in inheritable OA. Two reports on three unrelated families demonstrated coinheritance with primary generalized OA with specific alleles of the COL2A1 gene for type II procollagen on chromosome 12 (Knowlton et al, 1990). This allele has now been cloned and found to be normal except for a single base mutation (arginine to cysteine) at position 519 of the ix1 (II) chain (Alla-Kokko et al, 1990); this was found in all the affected members of a family but in none of the unaffected or unrelated individuals. A further study of generalized familial disease found that two out of seven families had the mutation, both of which had evidence of an associated chondrodysplasia (Pun et al, 1994). Linkage between COL2A1 and the development of OA has now been demonstrated in several.'families (Vikkula et al, 1993). In general, COL2A1 mutations have been shown to be associated with milder forms of chondrodysplasia, which may present with precocious generalized OA. The arginine 19 to cysteine and the arginine 75 to cysteine mutations are two such sites on COL2A1 and have now been reported in multiple, unrelated families with early-onset, generalized OA and chondrodysplasia. The observation of multiple sites where recurrent mutations occur suggests that certain areas of COL2A1 are more prone to mutational events. Recently, familial spondyloepiphyseal dysplasia tarda, brachydactyly, and precocious OA were shown to be associated with an arginine 75 to cysteine mutation in the COL2A1 gene in a kindred of Chiloe Islanders (Reginato et al, 1994). However, the relationship of this type of spondyloepiphyseal dysplasia tarda to familial CPDD and idiopathic hip dysplasia, both endemic in Chiloe Islanders, needs to be further investigated. Linkage analysis of several Stickler syndrome kindred has also demonstrated that the disease is linked to COL2A1 in approximately a quarter to
Evidence for genetically-determined OA
663
a half of families (Knowlton et al, 1989). No clear linkage has been demonstrated in CPDD. However, in a recent study of a large New England family with early-onset CPDD and severe degenerative OA (Baldwin et al, 1995), a genetic linkage was found between the disease in this family and chromosome 8q, suggesting that a defective gene at this location may be responsible. Although the available studies provide conclusive evidence that mutations in COL2A1 are present in affected individuals from some families displaying the phenotype of primary generalized OA and mild spondyloepiphyseal dysplasia, other studies have shown that COL2A1 is not the disease locus in other families with OA (Reginato et al, 1994). Primary generalized OA may be a heterogeneous disease at the genetic level, and mutations in genes other than COL2A1 are likely to be responsible for this phenotype. For example, the role of mutations in the genes encoding the minor collagen types (e.g. IX, X, XI) is yet to be explored. However, transgenic mouse models with a central deletion in the alpha-1 chain of the type IX collagen gene suggest that this gene may be important, as heterozygotes develop OA with no signs of chondrodysplasia (Nakata et al, 1993). Mice bomozygous for the transgene, display a mild chondrodysplastic syndrome with dwarfism, spinal involvement, and ophthalmopathy. In contrast, heterozygotes develop OA with no signs of chondrodysplasia (Nakata et al, 1993). Mutations in the type X collagen gene, which encodes a major collagen of growth plate cartilage, have been identified in the Schmid type of osteochondrodysplasias (Williams and Jimenez, 1995). A recent study has also suggested that mutations in collagen XI genes may have a role in OA (Vikkula et al, 1995).
Evidence for genetic collagen abnormalities in common forms of OA
There is currently little evidence that the common forms of OA are due to collagen mutations. The families identified are extremely rare and associated with mild chondrodysplasias, and do not allow clear conclusions to be drawn on the overall contribution of genetics to disease in the population. A recent study used gene-specific highly polymorphic markers and affected sibling pair analyses to investigate genetic linkage between OA and three cartilage matrix genes: COL2A1; CRTL1 which encodes the cartilage link protein, and CRTM which encodes the cartilage matrix protein (Loughlin et al, 1994). The analyses showed no linkage between generalized OA and the three genes in the 38 sibling pairs examined. These results suggest that COL2A1, CRTL1 and CRTM are not major susceptibility loci for generalized OA, although the power of the study was low. A recent study explored possible mutations in the COL2A1 gene in patients with cartilage diseases ranging from early-onset familial OA to lethal chondrodysplasias (Ritvaniemi et al, 1995). Mutations were detected in up to 2% of patients with early-onset familial OA. Although these
664
F.M. Cicuttini and T. D. Spector
percentages were probably only minimal estimates (as all possible mutations in the gene could not be detected with this procedure) these data still suggest that only a low proportion of OA can be explained by this genetic abnormality.
Implications for aetiological research: other candidate genes The role of the other extracellular matrix proteins such as aggrecan, decorin, and the link protein will need to be investigated (Rimoin et al, 1994). However, using the affected sibling pair method of analysis, genomic DNA from 66 sibling pairs with nodal OA was analysed for association with highly polymorphic microsatellite marker loci, and a significant association was identified between nodal OA and two loci on the short arm of chromosome 2 (2q 23-35) (Wright et al, 1996). Candidate genes for OA in this region include: fibronectin, a glycoprotein present in the extracellular matrix of normal cartilage; the alpha-2 chain of collagen type V, a major constituent of bone; and the interleukin-8 receptor, important in the regulation of neutrophil activation and chemotaxis. This chromosomal region requires further detailed study. Confirmation of these findings in large independent data sets, and further analysis of candidate genes in this region, may be important in unravelling the molecular basis of OA. Recently, new data have suggested an association between polymorphisms of the vitamin D receptor, which has recently been associated with osteoporosis, and early knee OA in a population study of 600 women (Keen et al, 1997). As some evidence suggests an inverse association between these two common conditions (Cooper et al, 1991), it may be that some polymorphisms of this gene may offer protection against OA. This will need to be tested in other populations.
Implications for novel treatment modalities The available evidence suggests that environmental risk factors, such as obesity and occupation, act differently on the risk of OA at different sites, pointing towards marked heterogeneity in environmental and genetic interactions. The fact that up to 70% of the disease variance may be genetic does not make OA 'inevitable'. There are likely to be important geneticenvironmental interactions such tl~at changes in lifestyle in predisposed individuals may have a marked effect. Identifying the extent of these interactions in different subgroups with OA will allow identification of individuals at risk and in whom efforts should be made to reduce the environmental risk factor (e.g. mechanical factors or obesity). This preventive approach aimed at individuals identified as being 'high risk' based on genetic factors is likely to be more effective than an approach that is unfocused.
Evidence for genetically-determined OA
665
Do genetic factors fulfill the Bradford-Hill Criteria for causation in osteoarthritis? Evidence relating genetic factors to OA comes from clinical and laboratory studies in humans. In a paper published in 1965, Bradford-Hill developed a set of criteria for assessing whether an association is likely to be causal (Bradford-Hill, 1965). The relevant criteria are considered in relation to this question.
Temporal relationship For a possible risk factor to be the cause of a disease it has to come before the disease. This essential criterion is clearly fulfilled by genetic factors.
Plausibility and consistency We have seen that evidence for a genetic effect in OA comes from family studies (Stecher, 1941; Kellgren and Moore, 1952; Kellgren et al, 1963; Pun et al, 1994) and twin studies (Spector et al, 1996) as well as animal models (Nakata et al, 1993). A molecular mechanism for some rare forms of OA has been identified. Thus, the association of genetics with OA is consistently seen in data obtained from different sources.
Strength of an association The influence of genetic factors on OA has been estimated to be between 39 and 65%, independent of known environmental or demographic confounders (Spector et al, 1996), hence a strong effect. In other studies, Heberden's nodes of the fingers were three times as common in the sisters of cases (Stecher et al, 1944), and first degree relatives were twice as likely to also have radiographic generalized disease (Kellgren et al, 1963).
Dose-response relationship The data from the twin studies suggest a dose-response relationship between genetic factors and OA. Identical twins, who share genes, are more concordant in terms of OA, than non-identical twins (Spector et al,
1996). Coherence Genetic factors are consistent with the natural history and biology of OA.
666
F.M. Cicuttini and T. D. Spector
Analogy Although the nature of the genetic influence in OA is speculative and may involve either a structural defect (i.e. collagen) or alterations in cartilage or bone metabolism, certain similarities exist with another common disease of old age, osteoporosis, the trait of which, bone density, has an estimated heritability of 70 to 90%. Direct comparison is difficult, however, given the vastly superior methodology of bone densitometry compared with radiography. Osteoporosis has also been associated with defects in collagen in rare families and, more recently, using twin studies, has been linked to polymorphisms of the vitamin D receptor gene (Pocock et al, 1987) which is likely to act via regulation of bone turnover. Twin studies of discordant OA pairs could be used to look for collagen gene polymorphisms in the same way. It is therefore possible that the turnover or repair of cartilage is under genetic control with different thresholds and responses in genetically distinct individuals.
Summary The available evidence suggests that genetic factors have a major role in OA. For over 50 years, a strong genetic component to certain forms of OA has been believed to be present. This genetic influence has now been estimated to be up to 65% in a large scale twin study. The nature of the genetic influence in OA is speculative and may involve either a structural defect (i.e. collagen), alterations in cartilage or bone metabolism or, alternatively, a genetic influence on a known risk factor for OA, such as obesity. Exciting work has identified mutations in type 2 collagen to be important in some rare, familial forms of OA but the genetic basis of the common forms of OA is currently unclear. Finding the genes involved is likely to lead to important therapeutic and diagnos6c advances in the most common disabling rheumatic disease.
Research agenda •
tile r~)le of genes controlling the cxtr:~cellul~lr matrix proteins. aggr~:can, decorin, and tile link protein, i,~ OA will need ro be clarilied
!
•
thc possible ass()cialion betwcc,~ polyrm~rl~hisms (~f the vih~min I) receptor and other candidates for osteopor~sis and ()A ~lced to be exph)red
,
•
any genetic-enxironme,n:fl interactions will need to be explored and their public health impact examined
.
i
Evidence for genetically-determined OA
667
Practice points •
heritability of primary OA of the hands and knees in women is up to 65%
•
mutations in COL2A1 have been identified in rare forms of familial OA
References "~Alla-Kokko L, Baldwin CT, Moskovitz RW & Prockop DJ (1990) A single base mutation in the type II procollagen gene (COL2A1) as cause of primary osteoarthritis associated with a mild chondrodysplasia. Proceedingsof the National Academy of Science 87: 6565-6568. Anderson IJ, Goldberg RB, Marion RW et al (1990) Spondyloepiphyseal dysplasia congenita: Genetic linkage to type II collagen (COL2A1). American Journal of Human Genetics 46:
896-901. *Baldwin CT, Farrer LA, Adair R et al (1995) Linkage of early-onset osteoarthritis and chondrocalcinosis to human chromosome 8q. 56: 692-697. Benavides G, Cerantes A, Silva Bet at (1985) HLA and Herbeden's nodes in Mexican mestizos. Clinical Rheumatology 4: 97-98. Bradford-Hill AB (1965) The environment and disease; association or causation? Proceedings of the Royal Society of Medicine 58: 295-300. Brodsky T, Appelboam A & Govaerts Farney JP (1979) Antigens HLA et nodosites d'Herbeden. Acta Rheumatologica 3: 95-103. Cooper C, Cook PL, Osmond C et al (1991) Osteoarthritis of the hip and osteoporosis of the proximal femur. Annals of Rheumatic Disease 50: 540-542. Cooper C, McAlindon T, Snow S e t al (1994) Mechanical and constitutional factors for symptomatic knee osteoarthritis: differences between medial tibiofemoral and patellofemoral disease. Journal of Rheumatology 21: 307-313. Fairbanks T (1947) Dysplasia epiphysialis multiplex. British Journal of Surgery 34: 224-232. Felson DT, Anderson JJ, Naimark A et al (1988) Obesity and knee osteoarthritis. The Framingham Study. Annals of Internal Medicine 109: 18-24. Hajiroussou VJ & Webley M (1986) Familial calcific periarthritis. Annals of Rheumatic Disease 42: 469-470. Hart DJ & Spector TD (1993) The relationship of obesity, fat distribution and osteoarthritis in women in the general population: the Chingford Study. Journal of Rheumatology 20: 331-335. *Keen RW, Hart DJ, Griffith GO et al (1997) Potymorphisms of the vitamin D receptor gene are associated with osteoarthritis of the knee. Arthritis and Rheumatism 1997 (in press). Kellgren JH & Moore R (1952) Generalized osteoarthritis and Heberden's nodes. British Journal of Medicine 1: 181-187. *Kellgren JH, Lawrence JS & Bier F (1963) Genetic factors in generalised osteoarthritis. Annals of Rheumatic Disease 22: 237-255. Kempson GE, Mir H & Swanson SAV (1970) Correlations between stiffness and the chemical constituents of cartilage on the human femoral head. Biochemica Biophysica Acta 215: 70-77. Kniest W & Leiber B (1977) Kniest syndrome. Monatsschrift Kinderheitkunde 125: 970-973. Knowlton RG, Weaver EJ, Struyk AF et al (1989) Genetic linkage analysis of hereditary arthro-ophthalmopathy (Sticker syndrome) and the type II procollagen gene. American Journal of Human Genetics 45: 681-688. *Knowlton RG, Katzenstein PL, Moscovitz RW et al (1990) Genetic linkage of a polymorphism in the type II procollagen gene (CoI2A1) to primary osteoarthritis associated with mild chondrodysplasia. New England Journal of Medicine 322: 526-530.
668
E M. Cicuttini and T. D. Spector
Lawrence JS (1977) Rheumatism in populations, pp 144-155 London: Heinemann. *Linberg H (1986) Hip OA in relations. Clinical Orthopaedics and Related Research 203: 273-275. *Loughlin J, Irven C, Fergusson C & Sykes B (1994) Sibling pair analysis shows no linkage of generalized osteoarthritis to the loci encoding type II collagen, cartilage link protein or cartilage matrix protein. British Journal of Rheumatology 33: 1103-1106. Marcos JC, DeBenyacar MA, Garcia-Morteo O et al (1981) Idiopathic familial chondrocalcinosis due to apatite crystal deposition. American Journal of Medicine 71: 557-564. Mayne R (1989) Cartilage collagens: What is their function, and are they involved in articular disease? Arthritis and Rheumatism 32: 241-246. Nakata K, Ono K, Miyazaki J et al (1993) Osteoarthritis associated with mild chondrodysplasia in transgenic mice expressing al(IX) collagen chains with a central deletion. Proceedings of the National Academy of Sciences of the USA 90: 2870-2874. Pattrick M, Manhire A, Ward M & Doherty M (1989) HLA-A, B antigen, and cd-antitrypsin phenotypes in nodal and generalized osteoarthritis and erosive osteoarthritis. Annals of Rheumatic Disease 48: 470-475. Pocock N, Eisman J, Hopper J e t al (1987) Genetic determinants of bone mass in adults. Journal of Clinical Investigation 80: 706-710. Poole AR, Pidoux I, Reiner A et al (1988) Kniest dysplasia is characterized by an apparent abnormal processing of the C-propeptide of type II cartilage collagen resulting in imperfect fibril assembly. Journal of Clinical Investigation 81: 579-589. "'Pun YL, Moskowitz RW, Lie Set al (1994) Clinical correlations of osteoarthritis associated with a single-base mutation (arginine 519 to cysteine) in type II procollagen gene. Arthritis and Rheumatism 37: 264-269. *Reginato AJ, Passano GM, Neumann G et al (1994) Familial spondyloepiphyseal dysplasia tarda, brachydactyly, and precocious osteoarthritis associated with an arginine 75--~cysteine mutation in the procollagen type II gene in a kindred of Chiloe Islanders. I. Clinical, radiographic, and pathologic findings. Arthritis and Rheumatism 37: 1078-1086. Rimoin DL, Rasmussen IM & Briggs MD (1994) A large family with features of pseudoachondroplasia and multiple epiphyseal dysplasia: Exclusion of seven candidate gene loci that encode proteins of the cartilage extracellular matrix. Human Genetics 93: 236-242. *Ritvaniemi P, Korkko J, Bonaventure J e t al (1995). Identification of COL2A1 gene mutations in patients with chondrodysplasias and familial osteoarthritis. Arthritis and Rheumatism 38: 999-1004. Ryan LM & McCarty DJ (1993) Calcium pyrophosphate crystal deposition disease: Pseudogout; articular chondrocalcinosis'. In McCarty DJ, ,Koopman WJ (eds) Arthritis and Allied Conditions, 12th edition, pp 1835-1855. Philadelphia: Lea & Febiger. Shiang R, Thompson LM, Zhu Y-Z et al (1994) Mutations in the transmembrane domain of FGFR3 cause the most common genetic form of dwarfism, achondroplasia. Cell 78: 335-342. Spector TD & Cooper C (1993) Whither Kellgren and Lawrence?: The radiographic definition of osteoarthritis. Osteoarthritis and Cartilage 1: 203-206. *Spector TD, Cicuttini F, Baker J & Hart DJ (1996) Genetic influences on osteoarthritis in women: a twin study. British Medical Journal 312: 940-944. Spranger J (1976) The epiphyseal dysplasias. Clinical Orthopaedics and Related Research 114: 45-59. Stecher RM (1941) Heberden's nodes. Heredity in hypertrophic arthritis of the finger joints. American Journal of Medical Science 201:801. *Stecher RM, Hersh AH & Hauser H (1944) Heberden's nodes; The mechanisms of inheritance in hypertrophic arthritis of the fingers. Journal of Clinical Investigation 23: 699-704. Stickler GB, Belau PG & Farrell FJ (1965) Hereditary progressive arthro-ophthalmopathy. Mayo Clinic Proceedings 40: 433-435. Urban J, Maroudas A & Bayliss MT (1979) Swelling pressure of proteoglycans at the concentrations found in cartilaginous tissues. Biorheology 16: 447-464. *Vikkula M, Palotic A, Ritvaniemi Pet al (1993) Early onset osteoarthritis linked to the type II procollagen gene; detailed clinical phenotype and further analysis of the gene. Arthritis and Rheumatism 36: 40!-409.
Evidence for genetically-determined OA
669
Vikkula M, Mariman EC, Lui VC et al (1995) Autosomal dominant and recessive osteochondrodysplasias associated with the COL11A2 locus. Cell 80: 431-437. Williams CJ & Jimenez SA (1995) Heritable diseases of cartilage caused by mutations in collagen genes. Journal of Rheumatology 43 (supplement): 28-33. ~Wright GD, Hughes AE, Regan M & Doherty M (1996) Association of two loci on chromosome 2q with nodal osteoarthritis. Annals of Rheumatic Disease 55: 317-319. Zitnan D & 8itaj S (1963) Chondroarticularis. I. Clinical and radiologic study. Annals of Rheumatic Disease 22: 142-169.
Tim D. Spector MSc, MD, FRCP Consultant Rheumatologist
The Twin Research Unit, St Thomas" Hospital, London SE1 7EH, UK
Although for many years it was speculated that osteoarthritis was genetically determined, little data were available to support this contention. A major problem with early work was a lack of consistency in the definition of osteoarthritis. Based on a radiographical definition of osteoarthritis, which is currently the optimal method for epidemiological and genetic studies, data from a recent twin study have provided an estimate of the hereditable component of osteoarthritis to be in the order of 50 to 65%. In addition, sophisticated molecular biology techniques are being increasingly used to explore potential genetic abnormalities in cartilage and matrix components in osteoarthritis. These exciting new data are examined as we address the role of genetic factors in osteoarthritis. Key words: osteoarthritis; genetics: inheritance.
Osteoarthritis (OA) is the most frequent cause of musculoskeletal disability in developed countries. The multifactorial nature of OA is well recognized, with a number of environmental risk factors such as obesity (Felson et al, 1988; Hart and Spencer, 1993), previous injury (Felson et al, 1988) and menisectomy (Cooper et al, 1994) being strongly associated with its development. The role of genetic factors in the development of OA has received considerable attention, aided by the rapidly expanding knowledge of molecular biology. The genetic contribution to a disease can be determined by epidemiological studies of family history, extended family correlations, mother daughter studies, and twins. These methodologies have been used successfully in determining the contribution of genetic factors to another common disease of old age, osteoporosis, the trait of which, bone density, has an estimated heritability of 70 to 90% (Pocock et al, 1987). However, a major difference between osteoporosis and OA is the superior disease definition in Bailliere's Clinical Rheumatolegy-Vol. 11, No. 4, November 1997 ISBN 0-7020-2368-X 0950-3579/97/040657 + 13 $12.00/00
657 Copyright © 1997, by Bailliere Tindall All rights of reproduction in any form reserved
658
F.M. Cicuttini and T. D. Spector
osteoporosis compared with OA, that is, the use of bone densitometry. Although the definition of disease in OA epidemiological research remains a problem, there is a general consensus that radiological changes in populations are superior to clinical signs and symptoms (Spector and Cooper, 1993). The advantages of using X-rays include better reproducibility, reasonable correlation with symptoms and disability, and an objective permanent record which can be assessed blind to clinical details. However, not all studies have used radiology as the definition of OA. This has resulted in some difficulties in the interpretation of results, particularly in determining the family history of OA in studies which rely on patient reporting.
Family studies Heberden's nodes For over 50 years a strong genetic component to certain forms of OA was thought to be present. The familial occurrence of Heberden's nodes of the fingers was first documented by Stecher (1941) who demonstrated that the nodes were three times more common in the sisters of 64 affected subjects as in the general population. In 1944, Stecher et al concluded that these lesions were inherited as a single autosomal dominant gene with a strong female predominance.
Primary generalized OA Primary generalized OA is the most common form of inherited OA. It was first described as a clinical entity by Kellgren and Moore (1952). Primary generalized OA is described by the presence of Heberden's and Bouchard's nodes and premature degeneration of articular cartilage, often in a concentric pattern in many joints. In the early 1960s, family studies of 20 male and 32 female probands with generalized OA, based on the presence of 5 radiologically-affected joint groups, suggested that first degree relatives were twice as likely to also have radiographical generalized disease (Kellgren et al, 1963). The familial occurrence of OA was further supported by a study of 391 cases of OA in which 120 patients, largely middle-aged women, were identified to have a polyarthritis characterized by signs of inflammation and an acute onset of symptoms. A family history of similar joint disease was given by 20% of these patients (Kellgren and Moore, 1952). This, and other studies, suggested a polygenic form of inheritance of primary generalized OA rather than a single gene defect (Kellgren et al, 1963). Further support for a genetic predisposition is suggested by the association with the HLA-AIB8 (Pattrick et al, 1989) and HLA-B8 (Brodsky et al, 1979) haplotypes and with 0d-anti-trypsin isoform patterns
Evidence for genetically-determined OA 659
(Pattrick et al, 1989), although not all studies confirmed this association (Benavides et al, 1985). Familial disorders associated with OA
A number of rare diseases that have a genetic basis are associated with severe, early onset OA. Familial calcium pyrophosphate-deposition disease
Familial calcium pyrophosphate-deposition disease (CPDD), is characterized by the deposition of calcium-containing crystals in joint tissue leading to arthritis-like symptoms. Early-onset CPDD has been described in several large families in which the disease progresses to severe degenerative OA (Ryan and McCarty, 1993). In these families, an autosomal dominant mode of inheritance is observed, with an age at onset between the second and fifth decades of life (Ryan and McCarty, 1993). Hydroxyapatite arthritis
Hydroxyapatite arthritis is another form of inherited crystal deposition disease and is due to the deposition of hydroxyapatite crystals in articular cartilage (Marcos et al, 1981). The clinical features and pathological manifestations of this disease are identical to CPDD, except.for the nature of the crystals deposited. The published pedigrees indicate that this disease also has an autosomal dominant inherited pattern with full penetrance (Hajiroussou and Webley, 1986). Stickler syndrome
The Stickler syndrome, also knOwn as the hereditary arthro-ophthalmopathy, is a familial form of OA with prominent ocular involvement (Stickler et al, 1965). This syndrome is a relatively common autosomal dominant disease (1 in 10 000) characterized by vitreous degeneration, retinal detachment, and premature degenerative joint disease. Several forms of familial OA associated with chondrodysplasia also have an autosomal dominant pattern of inheritance (Zitnan and Sitaj, 1963). Chondrodysplasias
The chondrodysplasias are a group of heterogeneous disorders that are characterized by abnormal cartilage growth (Spranger, 1976). The most common of these conditions is achondroplasia which has a frequency of 1 in 50 000 births and represents the classic form of 'short-limb' dwarfism. Although this often occurs sporadically, it may also be inherited as an autosomal dominant trait, and mutations have recently been found in the fibroblast growth factor gene (Shiang et al, 1994). Although linkage to the
660
F.M. Cicuttini and To D. Spector
COL2AI gene has been demonstrated in some cases, this is not consistent (Anderson et al, 1990).
Epiphyseal dysplasias The epiphyseal dysplasias are a group of disorders characterized by growth of the epiphyses which result in joint deformity and may lead to premature degenerative joint disease (Spranger, 1976). This group of disorders includes autosomal recessive, autosomal dominant, and X-linked patterns of inheritance. The expressions of disease range from death in the perinatal period to less severe forms characterized by short stature. The epiphyseal abnormalities often result in crippling OA in both weightbearing and nonweightbearing joints at a very early age.
Multiple epiphyseal dysplasias Among the epiphyseal dysplasias is a heterogeneous group of conditions known as multiple epiphyseal dysplasia (MED) (Fairbanks, 1947). MED is inherited in an autosomal dominant manner with a high degree of penetrance, and several large families with many affected members have been described. In contrast to other epiphyseal dysplasias, patients with MED do not have ocular or retinal abnormalities. In MED, the predominant clinical manifestations occur in the extremities rather than the spine.
Kniest dysplasia Kniest dysplasia, which has an autosomal dominant pattern of inheritance, is an extremely rare disorder characterized by shortening of the trunk and limbs, flattening of the face and bridge of the nose, protuberance of the eye globes, and severe joint abnormalitie~ (Kniest and Leiber, 1977). The joints are usually very large at birth and continue to enlarge during childhood and early adolescence. The majority of affected individuals develop severe, premature degenerative joint disease, particularly involving the knees and hips. The articular cartilage is soft and has decreased resilience. Histologically, the articular cartilage has large cystic lesions giving the appearance of Swiss cheese. It has been suggested that this disease may result from abnormalities in the processing of type II procollagen, as large inclusions containing the carboxyl-propeptide of type II procollagen have been found in the dilated rough endoplasmic reticulum (Poole et al, 1988).
Twin studies Classical twin studies are a very useful design for determining the extent to which genetic factors may influence the variation in a disease, since twins are uniquely matched for age and many known or unknown environmental
Evidence for genetically-determined OA
661
factors. As monozygotic twins are completely concordant for genetic factors, any intrapair variation is assumed to be due to environmental factors, while dizygotic twins share on average half of their genes so any intrapair variations are assumed to be due to both environmental and genetic factors. Thus a comparison of the similarities of monozygotic and dizygotic twins allows us to estimate the extent to which genetic factors determine variation in the measures of OA. Lawrence commented on an unfinished twin study of 10 pairs which suggested increased concordance in identical pairs at a number of joint sites (Lawrence, 1977). A full scale classical twin study was recently performed in the UK to determine whether radiographical OA of the hand and knee are predominantly genetic and to examine the extent to which genetic factors determine variation in the radiographical measures of OA. A clear genetic influence was demonstrated using 500 unselected female twins aged 45-70 years, who were screened radiologically for OA of the hand and knee (Spector et al, 1996). The correlations of OA disease status were consistently twofold higher in 130 pairs of identical compared with nonidentical twins. The influence of genetic factors was estimated to be between 39 and 65%, independent of known environmental or demographic confounders. This genetic effect was consistent whichever radiographical diagnosis of OA was used (i.e. osteophyte or narrowing) and whether it was defined clinically as knee pain or Heberden's nodes. Both the tibiofemoral and patellofemoral joints of the knee were influenced genetically. Only the proximal interphalangeal joints failed to show a genetic effect. A similar effect of inheritance was seen when the analyses were limited to osteophytes or narrowing scores separately or to the knee joint alone. These results demonstrated a clear genetic effect for primary OA of the hand and knee in women, with a heritability of up to 65% independent of environmental or demographic confounders. However, there was no clear evidence to support generalized OA as a separate genetic entity as the genetic effect for Heberden's nodes, a postulated marker for generalized OA, was similar to that for tibiofemoral and patellofemoral OA. Similar twin studies for hip OA have yet to be performed. However, one study from Scandinavia showed double the rate of hip OA in 289 siblings of 184 hip replacement patients compared with 289 controls (Linberg, 1986).
Possible genetic influence in OA The nature of the genetic influence in OA is speculative and may involve either a structural defect (i.e. collagen) or alterations in cartilage or bone metabolism. Proposed candidate genes in the inheritance of OA include the collagen genes which encode the most abundant proteins in the cartilage, with type II collagen being the commonest. There are several reasons that suggest the failure of the collagenous component of articular cartilage may be responsible for the degeneration of
662
E M. Cicuttini and T. D. Spector
joint tissue in OA. First, collagen is the main component of the organic matrix of articular cartilage and plays an important role in the maintenance of the biomechanical properties of cartilage (Kempson et al, 1970). Second, the normal assembly of cartilage collagen serves as a mechanical constraint to prevent the expansion of the highly-hydrated proteoglycans into the large hydrodynamic domains that are characteristic of proteoglycans free in solution (Urban et al, 1979). A failure of this would result in increased hydration of the tissue, softening of the matrix, and cartilage degradation. Finally, it has become apparent in recent years that articular cartilage contains a remarkable degree of molecular complexity with regard to the number of biochemical species of collagen (Mayne, 1989).
A role for type II procollagen in OA There are increasing numbers of mutations described in the COL2A1 gene for type II procollagen, which encodes the main collagenous component of articular cartilage. Genetic linkage analysis of large kindred populations were performed to examine the involvement of candidate genes in inheritable OA. Two reports on three unrelated families demonstrated coinheritance with primary generalized OA with specific alleles of the COL2A1 gene for type II procollagen on chromosome 12 (Knowlton et al, 1990). This allele has now been cloned and found to be normal except for a single base mutation (arginine to cysteine) at position 519 of the ix1 (II) chain (Alla-Kokko et al, 1990); this was found in all the affected members of a family but in none of the unaffected or unrelated individuals. A further study of generalized familial disease found that two out of seven families had the mutation, both of which had evidence of an associated chondrodysplasia (Pun et al, 1994). Linkage between COL2A1 and the development of OA has now been demonstrated in several.'families (Vikkula et al, 1993). In general, COL2A1 mutations have been shown to be associated with milder forms of chondrodysplasia, which may present with precocious generalized OA. The arginine 19 to cysteine and the arginine 75 to cysteine mutations are two such sites on COL2A1 and have now been reported in multiple, unrelated families with early-onset, generalized OA and chondrodysplasia. The observation of multiple sites where recurrent mutations occur suggests that certain areas of COL2A1 are more prone to mutational events. Recently, familial spondyloepiphyseal dysplasia tarda, brachydactyly, and precocious OA were shown to be associated with an arginine 75 to cysteine mutation in the COL2A1 gene in a kindred of Chiloe Islanders (Reginato et al, 1994). However, the relationship of this type of spondyloepiphyseal dysplasia tarda to familial CPDD and idiopathic hip dysplasia, both endemic in Chiloe Islanders, needs to be further investigated. Linkage analysis of several Stickler syndrome kindred has also demonstrated that the disease is linked to COL2A1 in approximately a quarter to
Evidence for genetically-determined OA
663
a half of families (Knowlton et al, 1989). No clear linkage has been demonstrated in CPDD. However, in a recent study of a large New England family with early-onset CPDD and severe degenerative OA (Baldwin et al, 1995), a genetic linkage was found between the disease in this family and chromosome 8q, suggesting that a defective gene at this location may be responsible. Although the available studies provide conclusive evidence that mutations in COL2A1 are present in affected individuals from some families displaying the phenotype of primary generalized OA and mild spondyloepiphyseal dysplasia, other studies have shown that COL2A1 is not the disease locus in other families with OA (Reginato et al, 1994). Primary generalized OA may be a heterogeneous disease at the genetic level, and mutations in genes other than COL2A1 are likely to be responsible for this phenotype. For example, the role of mutations in the genes encoding the minor collagen types (e.g. IX, X, XI) is yet to be explored. However, transgenic mouse models with a central deletion in the alpha-1 chain of the type IX collagen gene suggest that this gene may be important, as heterozygotes develop OA with no signs of chondrodysplasia (Nakata et al, 1993). Mice bomozygous for the transgene, display a mild chondrodysplastic syndrome with dwarfism, spinal involvement, and ophthalmopathy. In contrast, heterozygotes develop OA with no signs of chondrodysplasia (Nakata et al, 1993). Mutations in the type X collagen gene, which encodes a major collagen of growth plate cartilage, have been identified in the Schmid type of osteochondrodysplasias (Williams and Jimenez, 1995). A recent study has also suggested that mutations in collagen XI genes may have a role in OA (Vikkula et al, 1995).
Evidence for genetic collagen abnormalities in common forms of OA
There is currently little evidence that the common forms of OA are due to collagen mutations. The families identified are extremely rare and associated with mild chondrodysplasias, and do not allow clear conclusions to be drawn on the overall contribution of genetics to disease in the population. A recent study used gene-specific highly polymorphic markers and affected sibling pair analyses to investigate genetic linkage between OA and three cartilage matrix genes: COL2A1; CRTL1 which encodes the cartilage link protein, and CRTM which encodes the cartilage matrix protein (Loughlin et al, 1994). The analyses showed no linkage between generalized OA and the three genes in the 38 sibling pairs examined. These results suggest that COL2A1, CRTL1 and CRTM are not major susceptibility loci for generalized OA, although the power of the study was low. A recent study explored possible mutations in the COL2A1 gene in patients with cartilage diseases ranging from early-onset familial OA to lethal chondrodysplasias (Ritvaniemi et al, 1995). Mutations were detected in up to 2% of patients with early-onset familial OA. Although these
664
F.M. Cicuttini and T. D. Spector
percentages were probably only minimal estimates (as all possible mutations in the gene could not be detected with this procedure) these data still suggest that only a low proportion of OA can be explained by this genetic abnormality.
Implications for aetiological research: other candidate genes The role of the other extracellular matrix proteins such as aggrecan, decorin, and the link protein will need to be investigated (Rimoin et al, 1994). However, using the affected sibling pair method of analysis, genomic DNA from 66 sibling pairs with nodal OA was analysed for association with highly polymorphic microsatellite marker loci, and a significant association was identified between nodal OA and two loci on the short arm of chromosome 2 (2q 23-35) (Wright et al, 1996). Candidate genes for OA in this region include: fibronectin, a glycoprotein present in the extracellular matrix of normal cartilage; the alpha-2 chain of collagen type V, a major constituent of bone; and the interleukin-8 receptor, important in the regulation of neutrophil activation and chemotaxis. This chromosomal region requires further detailed study. Confirmation of these findings in large independent data sets, and further analysis of candidate genes in this region, may be important in unravelling the molecular basis of OA. Recently, new data have suggested an association between polymorphisms of the vitamin D receptor, which has recently been associated with osteoporosis, and early knee OA in a population study of 600 women (Keen et al, 1997). As some evidence suggests an inverse association between these two common conditions (Cooper et al, 1991), it may be that some polymorphisms of this gene may offer protection against OA. This will need to be tested in other populations.
Implications for novel treatment modalities The available evidence suggests that environmental risk factors, such as obesity and occupation, act differently on the risk of OA at different sites, pointing towards marked heterogeneity in environmental and genetic interactions. The fact that up to 70% of the disease variance may be genetic does not make OA 'inevitable'. There are likely to be important geneticenvironmental interactions such tl~at changes in lifestyle in predisposed individuals may have a marked effect. Identifying the extent of these interactions in different subgroups with OA will allow identification of individuals at risk and in whom efforts should be made to reduce the environmental risk factor (e.g. mechanical factors or obesity). This preventive approach aimed at individuals identified as being 'high risk' based on genetic factors is likely to be more effective than an approach that is unfocused.
Evidence for genetically-determined OA
665
Do genetic factors fulfill the Bradford-Hill Criteria for causation in osteoarthritis? Evidence relating genetic factors to OA comes from clinical and laboratory studies in humans. In a paper published in 1965, Bradford-Hill developed a set of criteria for assessing whether an association is likely to be causal (Bradford-Hill, 1965). The relevant criteria are considered in relation to this question.
Temporal relationship For a possible risk factor to be the cause of a disease it has to come before the disease. This essential criterion is clearly fulfilled by genetic factors.
Plausibility and consistency We have seen that evidence for a genetic effect in OA comes from family studies (Stecher, 1941; Kellgren and Moore, 1952; Kellgren et al, 1963; Pun et al, 1994) and twin studies (Spector et al, 1996) as well as animal models (Nakata et al, 1993). A molecular mechanism for some rare forms of OA has been identified. Thus, the association of genetics with OA is consistently seen in data obtained from different sources.
Strength of an association The influence of genetic factors on OA has been estimated to be between 39 and 65%, independent of known environmental or demographic confounders (Spector et al, 1996), hence a strong effect. In other studies, Heberden's nodes of the fingers were three times as common in the sisters of cases (Stecher et al, 1944), and first degree relatives were twice as likely to also have radiographic generalized disease (Kellgren et al, 1963).
Dose-response relationship The data from the twin studies suggest a dose-response relationship between genetic factors and OA. Identical twins, who share genes, are more concordant in terms of OA, than non-identical twins (Spector et al,
1996). Coherence Genetic factors are consistent with the natural history and biology of OA.
666
F.M. Cicuttini and T. D. Spector
Analogy Although the nature of the genetic influence in OA is speculative and may involve either a structural defect (i.e. collagen) or alterations in cartilage or bone metabolism, certain similarities exist with another common disease of old age, osteoporosis, the trait of which, bone density, has an estimated heritability of 70 to 90%. Direct comparison is difficult, however, given the vastly superior methodology of bone densitometry compared with radiography. Osteoporosis has also been associated with defects in collagen in rare families and, more recently, using twin studies, has been linked to polymorphisms of the vitamin D receptor gene (Pocock et al, 1987) which is likely to act via regulation of bone turnover. Twin studies of discordant OA pairs could be used to look for collagen gene polymorphisms in the same way. It is therefore possible that the turnover or repair of cartilage is under genetic control with different thresholds and responses in genetically distinct individuals.
Summary The available evidence suggests that genetic factors have a major role in OA. For over 50 years, a strong genetic component to certain forms of OA has been believed to be present. This genetic influence has now been estimated to be up to 65% in a large scale twin study. The nature of the genetic influence in OA is speculative and may involve either a structural defect (i.e. collagen), alterations in cartilage or bone metabolism or, alternatively, a genetic influence on a known risk factor for OA, such as obesity. Exciting work has identified mutations in type 2 collagen to be important in some rare, familial forms of OA but the genetic basis of the common forms of OA is currently unclear. Finding the genes involved is likely to lead to important therapeutic and diagnos6c advances in the most common disabling rheumatic disease.
Research agenda •
tile r~)le of genes controlling the cxtr:~cellul~lr matrix proteins. aggr~:can, decorin, and tile link protein, i,~ OA will need ro be clarilied
!
•
thc possible ass()cialion betwcc,~ polyrm~rl~hisms (~f the vih~min I) receptor and other candidates for osteopor~sis and ()A ~lced to be exph)red
,
•
any genetic-enxironme,n:fl interactions will need to be explored and their public health impact examined
.
i
Evidence for genetically-determined OA
667
Practice points •
heritability of primary OA of the hands and knees in women is up to 65%
•
mutations in COL2A1 have been identified in rare forms of familial OA
References "~Alla-Kokko L, Baldwin CT, Moskovitz RW & Prockop DJ (1990) A single base mutation in the type II procollagen gene (COL2A1) as cause of primary osteoarthritis associated with a mild chondrodysplasia. Proceedingsof the National Academy of Science 87: 6565-6568. Anderson IJ, Goldberg RB, Marion RW et al (1990) Spondyloepiphyseal dysplasia congenita: Genetic linkage to type II collagen (COL2A1). American Journal of Human Genetics 46:
896-901. *Baldwin CT, Farrer LA, Adair R et al (1995) Linkage of early-onset osteoarthritis and chondrocalcinosis to human chromosome 8q. 56: 692-697. Benavides G, Cerantes A, Silva Bet at (1985) HLA and Herbeden's nodes in Mexican mestizos. Clinical Rheumatology 4: 97-98. Bradford-Hill AB (1965) The environment and disease; association or causation? Proceedings of the Royal Society of Medicine 58: 295-300. Brodsky T, Appelboam A & Govaerts Farney JP (1979) Antigens HLA et nodosites d'Herbeden. Acta Rheumatologica 3: 95-103. Cooper C, Cook PL, Osmond C et al (1991) Osteoarthritis of the hip and osteoporosis of the proximal femur. Annals of Rheumatic Disease 50: 540-542. Cooper C, McAlindon T, Snow S e t al (1994) Mechanical and constitutional factors for symptomatic knee osteoarthritis: differences between medial tibiofemoral and patellofemoral disease. Journal of Rheumatology 21: 307-313. Fairbanks T (1947) Dysplasia epiphysialis multiplex. British Journal of Surgery 34: 224-232. Felson DT, Anderson JJ, Naimark A et al (1988) Obesity and knee osteoarthritis. The Framingham Study. Annals of Internal Medicine 109: 18-24. Hajiroussou VJ & Webley M (1986) Familial calcific periarthritis. Annals of Rheumatic Disease 42: 469-470. Hart DJ & Spector TD (1993) The relationship of obesity, fat distribution and osteoarthritis in women in the general population: the Chingford Study. Journal of Rheumatology 20: 331-335. *Keen RW, Hart DJ, Griffith GO et al (1997) Potymorphisms of the vitamin D receptor gene are associated with osteoarthritis of the knee. Arthritis and Rheumatism 1997 (in press). Kellgren JH & Moore R (1952) Generalized osteoarthritis and Heberden's nodes. British Journal of Medicine 1: 181-187. *Kellgren JH, Lawrence JS & Bier F (1963) Genetic factors in generalised osteoarthritis. Annals of Rheumatic Disease 22: 237-255. Kempson GE, Mir H & Swanson SAV (1970) Correlations between stiffness and the chemical constituents of cartilage on the human femoral head. Biochemica Biophysica Acta 215: 70-77. Kniest W & Leiber B (1977) Kniest syndrome. Monatsschrift Kinderheitkunde 125: 970-973. Knowlton RG, Weaver EJ, Struyk AF et al (1989) Genetic linkage analysis of hereditary arthro-ophthalmopathy (Sticker syndrome) and the type II procollagen gene. American Journal of Human Genetics 45: 681-688. *Knowlton RG, Katzenstein PL, Moscovitz RW et al (1990) Genetic linkage of a polymorphism in the type II procollagen gene (CoI2A1) to primary osteoarthritis associated with mild chondrodysplasia. New England Journal of Medicine 322: 526-530.
668
E M. Cicuttini and T. D. Spector
Lawrence JS (1977) Rheumatism in populations, pp 144-155 London: Heinemann. *Linberg H (1986) Hip OA in relations. Clinical Orthopaedics and Related Research 203: 273-275. *Loughlin J, Irven C, Fergusson C & Sykes B (1994) Sibling pair analysis shows no linkage of generalized osteoarthritis to the loci encoding type II collagen, cartilage link protein or cartilage matrix protein. British Journal of Rheumatology 33: 1103-1106. Marcos JC, DeBenyacar MA, Garcia-Morteo O et al (1981) Idiopathic familial chondrocalcinosis due to apatite crystal deposition. American Journal of Medicine 71: 557-564. Mayne R (1989) Cartilage collagens: What is their function, and are they involved in articular disease? Arthritis and Rheumatism 32: 241-246. Nakata K, Ono K, Miyazaki J et al (1993) Osteoarthritis associated with mild chondrodysplasia in transgenic mice expressing al(IX) collagen chains with a central deletion. Proceedings of the National Academy of Sciences of the USA 90: 2870-2874. Pattrick M, Manhire A, Ward M & Doherty M (1989) HLA-A, B antigen, and cd-antitrypsin phenotypes in nodal and generalized osteoarthritis and erosive osteoarthritis. Annals of Rheumatic Disease 48: 470-475. Pocock N, Eisman J, Hopper J e t al (1987) Genetic determinants of bone mass in adults. Journal of Clinical Investigation 80: 706-710. Poole AR, Pidoux I, Reiner A et al (1988) Kniest dysplasia is characterized by an apparent abnormal processing of the C-propeptide of type II cartilage collagen resulting in imperfect fibril assembly. Journal of Clinical Investigation 81: 579-589. "'Pun YL, Moskowitz RW, Lie Set al (1994) Clinical correlations of osteoarthritis associated with a single-base mutation (arginine 519 to cysteine) in type II procollagen gene. Arthritis and Rheumatism 37: 264-269. *Reginato AJ, Passano GM, Neumann G et al (1994) Familial spondyloepiphyseal dysplasia tarda, brachydactyly, and precocious osteoarthritis associated with an arginine 75--~cysteine mutation in the procollagen type II gene in a kindred of Chiloe Islanders. I. Clinical, radiographic, and pathologic findings. Arthritis and Rheumatism 37: 1078-1086. Rimoin DL, Rasmussen IM & Briggs MD (1994) A large family with features of pseudoachondroplasia and multiple epiphyseal dysplasia: Exclusion of seven candidate gene loci that encode proteins of the cartilage extracellular matrix. Human Genetics 93: 236-242. *Ritvaniemi P, Korkko J, Bonaventure J e t al (1995). Identification of COL2A1 gene mutations in patients with chondrodysplasias and familial osteoarthritis. Arthritis and Rheumatism 38: 999-1004. Ryan LM & McCarty DJ (1993) Calcium pyrophosphate crystal deposition disease: Pseudogout; articular chondrocalcinosis'. In McCarty DJ, ,Koopman WJ (eds) Arthritis and Allied Conditions, 12th edition, pp 1835-1855. Philadelphia: Lea & Febiger. Shiang R, Thompson LM, Zhu Y-Z et al (1994) Mutations in the transmembrane domain of FGFR3 cause the most common genetic form of dwarfism, achondroplasia. Cell 78: 335-342. Spector TD & Cooper C (1993) Whither Kellgren and Lawrence?: The radiographic definition of osteoarthritis. Osteoarthritis and Cartilage 1: 203-206. *Spector TD, Cicuttini F, Baker J & Hart DJ (1996) Genetic influences on osteoarthritis in women: a twin study. British Medical Journal 312: 940-944. Spranger J (1976) The epiphyseal dysplasias. Clinical Orthopaedics and Related Research 114: 45-59. Stecher RM (1941) Heberden's nodes. Heredity in hypertrophic arthritis of the finger joints. American Journal of Medical Science 201:801. *Stecher RM, Hersh AH & Hauser H (1944) Heberden's nodes; The mechanisms of inheritance in hypertrophic arthritis of the fingers. Journal of Clinical Investigation 23: 699-704. Stickler GB, Belau PG & Farrell FJ (1965) Hereditary progressive arthro-ophthalmopathy. Mayo Clinic Proceedings 40: 433-435. Urban J, Maroudas A & Bayliss MT (1979) Swelling pressure of proteoglycans at the concentrations found in cartilaginous tissues. Biorheology 16: 447-464. *Vikkula M, Palotic A, Ritvaniemi Pet al (1993) Early onset osteoarthritis linked to the type II procollagen gene; detailed clinical phenotype and further analysis of the gene. Arthritis and Rheumatism 36: 40!-409.
Evidence for genetically-determined OA
669
Vikkula M, Mariman EC, Lui VC et al (1995) Autosomal dominant and recessive osteochondrodysplasias associated with the COL11A2 locus. Cell 80: 431-437. Williams CJ & Jimenez SA (1995) Heritable diseases of cartilage caused by mutations in collagen genes. Journal of Rheumatology 43 (supplement): 28-33. ~Wright GD, Hughes AE, Regan M & Doherty M (1996) Association of two loci on chromosome 2q with nodal osteoarthritis. Annals of Rheumatic Disease 55: 317-319. Zitnan D & 8itaj S (1963) Chondroarticularis. I. Clinical and radiologic study. Annals of Rheumatic Disease 22: 142-169.