WO2005100596A1 - Diagnostic markers for osteoarthritis - Google Patents

Abstract

A method is provided for determining the risk of an individual developing osteoarthritis, comprising detecting the presence of osteoarthritis-associated ADAM12, THSD2, CD36 and NCOR2 alleles in a nucleic acid sample of the individual, wherein the presence of said alleles indicates the individual's risk for of developing osteoarthritis.

Description

DIAGNOSTIC MARKERS FOR OSTEOARTHRITIS

Field of invention.

The present invention relates to the fields of osteoarthritis and osteoporosis, molecular biology, and drug development. More specifically, it relates to methods and reagents for detecting nucleotide sequence variability in the ADAM12, THSD2, NCOR2, and CD36 genes that is associated with osteoarthritis onset, severity and progression.

Background of invention. Osteoarthritis (OA) is the most common type of arthritis. It differs from rheumatoid arthritis, which is an inflammatory disease first and foremost, in that it is primarily a degeneration of the joint tissue that may be accompanied by an inflammatory reaction. (Martel-Pelletier J.Osteoarthritis Cartilage. 1998 Nov;6(6):374-6.). The initiation and progression of osteoarthritis involves multiple pathogenic mechanisms. An imbalance of chondrocyte-controlled anabolic and catabolic processes results in a progressive degradation of the components of the extracellular matrix of the articular cartilage, associated with secondary inflammatory factors (Goldring MB. Curr Rheumatol Rep 2000 Dec;2(6):459-65.) The primary cause of this is unknown but possibly involves a deficiency of cellular response to normal tissue demand or insufficient cellular response to supernormal demand from mechanical loading or injury. The subsequent repair response could induce elevated levels of anabolic molecules, leading to remodeling of the bone and production of osteophytes (bone outgrowths) characteristic of the disease process.

With approximately 40 million Americans affected by arthritis and other inflammatory diseases, the cost to the healthcare system is significant. Of these 40 million people, 21 million have osteoarthritis and 2.1 million have rheumatoid arthritis. Osteoarthritis is the most common chronic condition and cause of inactivity in patients older than 65. The disease occurs usually at the beginning of the fifth decade of life, with increasing prevalence and incidence with advancing age. The prevalence of arthritis is expected to increase by 57% by the year 2020. In the same time period, arthritis-causing activity limitation will increase 66% to 11.6 million people (Lawrence et al Arthritis Rheum. 1998;41(5):778-799). The primary impact of arthritis in the elderly is decreased physical functioning. This can be due to other health-related problems, such as weight gain, cardiovascular disease, GI distress related to treatment, increased psychological distress, decreased social functioning, increased work disability, and increased healthcare utilization.

There is a general consensus that radiological changes are the preferred method for epidemiological studies on the basis of cross sectional and prospective correlations between severity of X-ray changes with the presence of pain and loss of function. In osteoarthritis, the loss of cartilage produces a narrowed space between bones. The pattern of joint space narrowing can help distinguish between osteoarthritis and rheumathoid arthritis. Bone spurs (osteophytes) also help diagnose osteoarthritis. Other relevant clinical end points are pain, disability, function, joint replacement and maintenance of joint structure. Stages of disease progression are as follows:

Early stage: focal swelling of articular cartilage followed by the appearance of irregularities in the surface. Intermediate stage: progressive degradation and loss of articular cartilage. Also characterized by fibrillation (vertical splitting), detachment (horizontal splitting) and thinning of the cartilage. Late stage: Articular cartilage is almost completely destroyed. Bony outgrowths (osteophytes) occur at the joint margins resulting in residual arthritis. Characterized by pain and limitation of joint movement.

Osteoarthritis has many causes, but the end result is a structural degradation of joint cartilage and a failure of chondrocytes (cartilage forming cells, the main component of cartilage) to repair the degraded cartilage collagen matrix.

The nature of the genetic influence in osteoarthritis may involve either a structural defect (that is, collagen), alterations in cartilage or bone metabolism, or a genetic influence on a Icnown risk factor for osteoarthritis such as obesity. Twin studies have show that between 39% and 65% of osteoarthritis in the general population can be attributed to genetic factors (MacGregor and Spector, 1999). Linkage analyses (i.e., common inheritance of affected individuals in the same family) have identified a higher risk ratio for relatives of affected loci through linkage analysis using pairs of affected relatives depends on _R, the risk ratio for type R relatives compared with population prevalence (Risch 1990). Kellgren et al. (1963) compared expected and observed incidence of osteoarthritis in first-degree relatives of probands with multiple osteoarthritis. Based on their results we have estimated λj? for nodal and non-nodal osteoarthritis.

The current OA treatment, NSAIDs are responsible for the highest number of hospitalizations of any drug category and cause a significant number of internal gastrointestinal bleeding in the elderly population.

Current treatment options for osteoarthritis focus on symptom relief whereas truly disease-modifying agents or methods are lacking. Thus, the basic therapy includes common analgesics, nonsteroidal anti-inflammatory drugs, physical therapy, walking aids, and eventually in severe cases, joint replacement surgery. Perhaps because of the difficulties involved in measuring disease progression existing medications do not address the need to prevent further cartilage degradation.

To develop such drugs the following should be in place: (l)Compounds that target appropriate biochemical pathways (2) Clinical studies must be able to measure disease progression in a cost-effective and safe fashion. (3) Disease progression should be detectable within a reasonable time scale (4) The efficacy of the new drug under development should be observable (using either the imaging or biomarker method of assessment) in a sample size comparable to that of other clinical trials.

In some patients, radiographically evident OA may stabilize with only marginal changes over a period of ten years, while in other patients with similar baseline changes, the disease may rapidly progress to end-stage joint destruction within 6-12 months (Akesson, 1999). In other words, standard clinical parameters are not sufficient for predicting disease progression. Thus the need to identify subjects prone to develop severe OA over time. The lumbar spine is a common location for osteoarthritis. The axial skeleton demonstrates the same classic alterations of cartilage loss, joint instability, and osteophytosis characteristic of symptomatic disease in the appendages. Despite these similarities, questions remain regarding the lumbar spine facet joints as a source of chronic back pain. The facet joints undergo a progression of degeneration that may result in pain. The facet joints have sensory input from two spinal levels that makes localization of pain difficult. Radiographic studies describe intervertebral disc abnormalities in asymptomatic individuals that are associated with, but not synonymous for, osteoarthritis. Patients who do not have osteoarthritis of the facet joints on magnetic resonance scan do not have back pain (Bomestein 2004). Lumbar disk disease (LDD) is one of the most common musculoskeletal diseases, with a prevalence of about 5%. It has been suggested that a genetic factor or familial predisposition contributes to the development of lumbar disc herniation. Genetic polymorphisms at the aggrecan gene and the collagen type IX gene (COL9A1) have been associated with LDD (Passilta et al 2001; Kawaguchi et al 1999)

Osteoporosis is a common disease characterized by reduced bone mineral density (BMD), deterioration of bone micro-architecture and increased risk of bone damage, such as fracture. (World Health Organization 1994. Assessment of fracture risk and its application to screening for postmenopausal osteoporosis. W.H.O. Tech Rep Ser 843).

Common types of osteoporosis include postmenopausal osteoporosis; and senile osteoporosis, which generally occurs in later life, e.g., 70+ years. It is a major public health problem which affects quality of life and increases costs to health care providers. In European populations, one in three women and one in twelve men over the age of fifty is at risk. The disease effects 25 million people in the USA, where the incidence of disease is 25% higher than it is in the UK, and a further 50 million people in Japan and Europe combined (Rosen 1997). It is estimated that by the middle of the next century the number of osteoporosis sufferers will double in the West, but may increase six-fold in Asia and South-America. Fracture is the most serious endpoint of osteoporosis, particularly fracture of the hip which affects up to

1.7 million people worldwide each year. It is estimated that by the year 2050, the number of hip fractures worldwide will increase to over 6 million, as life expectancy and age of the population increase (See Spangler et al. "The Genetic Component of Osteoporosis Mini-review"; http://www.csa.com.osteointro.html..

Peak bone mass is mainly genetically determined, though dietary factors and physical activity can have positive effects. Peak bone mass is attained at the point when skeletal growth ceases, after which time bone loss starts. In contrast to the positive balance that occurs during growth, in osteoporosis, the resorbed cavity is not completely refilled by bone (Parfitt 1988)

Osteoporosis is also a major health problem in virtually all societies (Eisman 1996;

Wark 1996;). Hip fractures are by far the most dangerous and expensive manifestation of osteoporosis (Cummings SR, Black D, Rubin S. 1989. Lifetime risks for hip and vertebral fracture among postmenopausal women. Arch Intern Med. 149:2445-2448) It is estimated that 30 million Americans are at risk for osteoporosis, the most common among these diseases, and there are probably 100 million people similarly at risk worldwide (Melton, 1995). These numbers are growing as the elderly population increases. Despite recent successes with drugs that inhibit bone resorption, there is a clear need for specific anabolic agents that will considerably increase bone formation in people who have already suffered substantial bone loss. There are no such drugs currently approved.

Current treatment for osteoporosis helps stop further bone loss and fractures. Common therapeutics include HRT (hormone replacement therapy), bisphosphonates, e.g., alendronate (Fosamax), as well as, estrogen and estrogen receptor modulators, progestin, calcitonin, and vitamin D. While there may be numerous factors that determine whether any particular person will develop osteoporosis, a step towards prevention, control or treatment of osteoporosis is determining whether one is at risk for osteoporosis. Genetic factors also play an important role in the pathogenesis of osteoporosis (Ralston 1997; see also Keen et al. 1997; Eisman 1996; Rosen 1997; Cole 1998, Johnston et al. 1995; Gong et al. 1996; Wasnich 1996 inter alia). Some attribute 50-60% of total bone variation (Bone Mineral Density; BMD), depending upon the bone area, to genetic effects (Livshits et al. 1996). Studies have shown from family histories, twin studies, and racial factors, that there may be a predisposition for osteoporosis (see, e.g., Jouanny et al. 1995; Garnero et al. 1996; Cummings 1996; Lonzer et al. 1996). Several candidate genes may be involved in this, most probably multigenic, process. Osteoporosis can be considered a complex genetic trait with variants of several genes underlying the genetic determination of the variability of the phenotype. Low bone mineral density (BMD) is an important risk factor for fractures, the clinically most relevant feature of osteoporosis. Segregation analysis in families has shown that BMD is under polygenic control while, in addition, biochemical markers of bone turnover have also been shown to have strong genetic components (Livshits et al 1996)). Several candidate genes have been analysed in relation to BMD but the most widely studied gene in this respect, the vitamin D receptor (VDR) gene, explains only a small part of the genetic effect on BMD. Numerous studies, focussing on the Bsml allele of the vitamin D receptor gene have concluded that absence of the restriction site correlates with low bone mineral density.

Description of related art.

The ADAM12 (meltrin alpha) gene encodes a metalloprotease and has been shown to be important in mediating cellular interactions and responses [Tian et al 2002]. It appears to regulate the formation of macrophage-derived giant cells possibly by mediating the effects of 1,25 hydroxy vitamin D3 on cell-cell fusion (which itself has been implicated in OA). Moreover, the addition of antisense ADAM12 mRNA to mononuclear osteoclast precursor cells results in 50% decrease in giant cell formation (Abe et al 1999). Its role in osteoclast formation might explain the association of this gene with osteophyte presence and progression.

NCOR2 is a silencing mediator for retinoid and thyroid hormone receptors and also as a modulator of both basal and ligand-activated transcription of genes controlled by RAR/RXR heterodimers in a dose-dependent manner [Dong and Tweardy 2002; Ghosh 2002 ] The information available on ADAM12, THSD2, CD36 and NCOR2 on the OMEVI and LocusLink public websites is presented beneath:

ADAM12: a disintegrin and metalloproteinase domain 12 (meltrin alpha) http://www.ncbi.nlm.nih.gov/LocusLink/LocRpt.cgi?l=8038

OMIM link: http://www.ncbi.nlm.nih.gov/entrez/dispomim.cgi?id=602714 LocusID: 8038

This gene encodes a disintegrin and metalloprotease (ADAM) domain 12, which is a member of the ADAM protein family. Members of this family are membrane- anchored proteins structurally related to snake venom disintegrins, and have been implicated in a variety of biologic processes involving cell-cell and cell-matrix interactions, including fertilization, muscle development, and neurogenesis. This gene has 2 alternative splicing transcript forms: a shorter secreted form and a longer membrane-bound form, that diverge at their 3' ends. The shorter form is found to stimulate myogenesis. Locus Type: gene with protein product, function known or inferred Product: a disintegrin and metalloprotease domain 12 isoform 1 preproprotein a disintegrin and metalloprotease domain 12 isoform 2 preproprotein

Alternate Symbols: MCMP, MLTN, MLTNA, MCMPMltna

Alias: meltrin alpha

A disintegrin and metalloproteinase domain 12 (Meltrin-alpha, mouse, homolog of)

NCBI Reference Sequences (RefSeq):

Category: REVIEWED mRNA: NM_003474

Protein: NPJ303465 a disintegrin and metalloprotease domain 12 isoform 1 preproprotein BL

Domains: ADAM Cysteine-Rich Domain score: 422

Homologues of snake disintegrins score: 241 Reprolysin (M12B) family zinc metalloprotease. The members of this family are enzymes that cleave peptides. These proteases require zinc for catalysis. Members of this family are also known as adamalysins. Most members of this family are snake venom endo score: 721

Reprolysin family propeptide. This region is the propeptide for members of peptidase family M12B. The propeptide contains a sequence motif similar to the "cysteine switch" of the matrixins. This motif is found at the C terminus of the alignment but is no score: 314

Transcript Variant: This variant (1) lacks an alternative exon, as compared to variant 2. As a result, isoform 1 and isoform 2 have different C-terminal amino acid sequences.

GenBank Source: AF023476 mRNA: NM_021641

Protein: NP_067673 a disintegrin and metalloprotease domain 12 isoform 2 preproprotein BL

Domains: ADAM Cysteine-Rich Domain score: 414

Homologues of snake disintegrins score: 235

Reprolysin (M12B) family zinc metalloprotease. The members of this family are enzymes that cleave peptides. These proteases require zinc for catalysis. Members of this family are also known as adamalysins. Most members of this family are snake venom endo score: 711

Reprolysin family propeptide. This region is the propeptide for members of peptidase family M12B. The propeptide contains a sequence motif similar to the "cysteine switch" of the matrixins. This motif is found at the C terminus of the alignment but is no score: 313

Transcript Variant: This variant (2) contains an alternative exon, which is absent in variant 1. This exon encodes an isoform 2-specific carboxyl terminus, 3' UTR and polyA site.

GenBank Source: AF023477

*602714

A Disintegrin And Metalloproteinase Domain 12; ADAM12 Alternative titles; symbols

MELTRIN- ALPHA, MOUSE, HOMOLOG OF; MLTN

Gene map locus 10q26.3

To isolate genes related to fertilin expressed in muscle, Yagami-Hiromasa et al.

(1995) amplified cDNAs prepared from a mouse myogenic cell line by PCR using degenerative primers for conserved amino acids between fertilin-alpha and -beta (601533). They identified 3 novel mouse sequences, which they called meltrins. Similarly to myogenin, a marker of early muscle differentiation, mouse meltrin-alpha is expressed in neonatal muscle and bone, and its expression increases dramatically in response to the induction of differentiation. Immunocytochemical localization and functional expression studies suggested that meltrin-alpha may be involved in myotube formation. Gilpin et al. (1998) screened a yeast 2-hybrid cDNA library with laminin beta-2 (150325) cDNA. They isolated a human gene, which they termed ADAM12, that is the human homolog of mouse meltrin-alpha. They found 2 differentially spliced isoforms, a shorter secreted form (ADAM12S) and a larger membrane-bound form (ADAM12L), that diverge at their 3-prime ends. ADAM12L more closely resembles meltrin-alpha; the mature 881-amino acid protein contains a transmembrane domain at its C terminus. The mature ADAM12S 718-amino acid protein has no transmembrane domain. Both forms contain a metalloprotease domain, a disintegrin domain, and a cysteine-rich domain, components characteristic of the ADAM family.

Galliano et al. (2000) found by RT-PCR and immunoblot analyses that expression of mouse Adaml2 increases during muscle regeneration, while the levels of other

ADAMs remain constant. Immunofluorescence analysis revealed staining of small, newly formed muscle fibers in regenerating but not normal adult muscle cells. Using a yeast 2-hybrid screen of a human skeletal muscle cDNA library with the cytoplasmic tail of human AD AMI 2 as bait, Galliano et al. (2000) determined that the membrane proximal portion of the C-terminal half of myristoylated ADAM12 interacts with muscle-specific alpha-actinin-2 (ACTN2; 102573). Galliano et al. (2000) determined that overexpression of cytosolic ADAM12 containing the ACTN2-binding site inhibits mouse myoblast fusion.

Gilpin et al. (1998) used fluorescence in situ hybridization to map the ADAM12 gene to human chromosome 10q26.3. Radiation hybrid mapping placed the gene between markers D10S216 and D10S575. A sequence tagged site (WI-17472) is identical to part of the 3-prime untranslated region of ADAM12. By FISH, Kurisaki et al. (2003) mapped the mouse Adam 12 gene to the F3 distal-F4 band of chromosome 7.

By gene targeting, Kurisaki et al. (2003) developed Adaml2-deficient mice. Mutant embryos developed at the expected mendelian ratios; however, 30% of the null pups died before weaning. Viable homozygous mutants appeared normal and were fertile. Most muscles of the Adaml2 null mice appeared normal, and regeneration in experimentally damaged skeletal muscle was unimpeded. In some Adaml2 null pups, the interscapular brown adipose tissue was reduced, although the penetrance of this phenotype appeared low. Impaired formation of the neck and interscapular muscles was also seen in some homozygotes. Kurisaki et al. (2003) hypothesized that Adaml2 may be involved in regulating adipogenesis and myogenesis through a linked developmental pathway. They also found that phorbol ester treatment of cultured fibroblasts of null embryos showed reduced ectodomain shedding of heparin-binding epidermal growth factor-like growth factor (HBEGF; 126150), suggesting that this growth factor is a substrate for Adaml2.

CD36: CD36 antigen (collagen type I receptor, thrombospondin receptor) http ://w ww .ncbi .nlm.nih . gov/LocusLink/LocRpt.c gi ?1=948 http://www.ncbi.nlm.nih.gov/entrez/dispomim.cgi ?id=173510 LocusID: 948

CD36 is the receptor for thrombospondin and collagen in platelets. It plays important functions in cell adhesion. Locus Type: gene with protein product, function known or inferred. Product: CD36 antigen (collagen type I receptor, thrombospondin receptor) Alternate Symbols: FAT, GP4, GP3B, GPIV, SCARB3

Alias: cluster determinant 36, fatty acid translocase, CD36 antigen (collagen type I) scavenger receptor class B, member 3

Function Submit GeneRIF (All Pubs): Phenotype: Macrothrombocytopenia Malaria, cerebral, reduced risk of Malaria, cerebral, susceptibility to Platelet glycoprotein IV deficiency

GeneRIF: Gene References into Function: 12031598 Genomic heterogeneity of type II CD36 deficiency 11834946 CD36 deficiency induced by antiretro viral therapy. 12664607 Expression in melanoma cells is regulated by NSAIDs. 11668637 novel variants in individuals from malaria-endemic Ghana 12606036 mediates endocytic uptake of advanced glycation end products 11867619 promoter structure and trans activation by pparalpha and ppargamma ligands 12576469 polySia of milk CD36 is significant for neonatal development in terms of protection and nutrition 12598312 Macrophage recognition and phagocytosis of apoptotic fibroblasts is critically dependent on this protein. 12618277 Expression of the monocyte-macrophage LDL scavenger receptor CD36 is increased in people with diabetes mellitus, type 2. 12496189 mediates nonopsonic phagocytosis of erythrocytes infected with stage I and IIA gametocytes of Plasmodium falciparum in monocytes and macrophages 11872368 Advanced glycation endproducts generated in situ are recognized by

CD36, which might contribute to the pathogenesis of diabetic macrovascular complications. 11718687 CD36 Pro90Ser mutation is not necessarily related to the insulin resistance syndrome, but is associated with high free fatty acid concentrations in

Japanese 12023894 identification of terminal six amino acids of carboxy cytoplasmic tail as a functional domain implicated in binding and capture of oxidized low-density lipoprotein 12716760 data suggest that platelet CD36 has a key role in VLDL-induced collagen- mediated platelet aggregation, possibly contributing to atherothrombosis associated with increased VLDL levels 12105195 Results offer structural insights into the molecular patterns recognized by the scavenger receptor CD36 and provide a platform for the development of potential therapeutic inhibitory agents. 12479587 Data show that expression of CD36 on platelet membranes was increased in myeloproliferative, but was not significantly different from normal controls in non- insulin-dependent diabetes mellitus. 12516552 Adhesion of dendritic cells derived from CD34+ progenitors to resting human dermal microvascular endothehal cells is down-regulated upon maturation and partially depends on CDlla-CD18, CDllb-CD18 and CD36. 12224819 Heterozygotes, CD36+/-, have a significant reduction of long-chain fatty acids uptake in the heart independent of heart diseajje, suggesting genotype dependency and that CD36 might be a fundamental determinant of myocardial LCFA uptake. 11686358 The metabolic role of CD36 was tested by glucose loading in type I or II CD36 deficiency pts. Changes in triglyceride and glucose metabolism were seen in CD36 deficiency due to impaired fast FA clearance by muscle and increased liver FA uptake. 11714819 A redox-dependent conformational fraction of CD36, localized in the

CD36 C-terminal cysteine-rich region, mediates cytoadherence of Plasmodium falciparum-infected erythrocytes to CD36-expressing cells and is inducible by low toxicity reducing agents. Relationships : Mouse Homology Maps: NCBI vs. MGD 5 2.00 cM Cd36 Hs Mm Map Information : Chromosome: 7 mv Cytogenetic: 7qll.2 HUGO Markers: Chr. 7 SHGC-13954 mv Chr. 7 CD36 CD36 mv Chr. - GDB:3755118 Chr. - D7S2791 D7S2791 Chr. 7 RH69590 mv NCBI Reference Sequences (RefSeq) Category: REVIEWED mRNA: NM_000072 Protein: NP_000063 CD36 antigen (collagen type I receptor, thrombospondin receptor) BL Domains: CD36 family. The CD36 family is thought to be a novel class of scavenger receptors. There is also evidence suggesting a possible role in signal transduction.

CD36 is involved in cell adhesion score: 1156 GenBank Source: L06850 Category: NCBI Genome Annotation Genomic Contig: NTJ307933 gb sv mv ev mm Haplotype reference Annotation for this locus: Evidence: supported by alignment with both mRNA and ESTs (29) mRNA: NM_000072 Protein: NP_000063 OMLM:

CD36 Antigen; CD36 Alternative titles; symbols Leukocyte Differentiation Antigen CD36 Platelet Glycoprotein IV; GP4 Glycoprotein Illb

GP Illb; GP3B Thrombospondin Receptor Collagen Receptor, Platelet Fatty Acid Translocase; Fat Platelet Glycoprotein IV Deficiency Gene map locus 7qll.2

Platelet glycoprotein TV is immunologically related to the leukocyte differentiation antigen CD36. It is alternatively known as GP nib and is the fourth major glycoprotein of the platelet surface, the others being GP lb, the platelet for thrombin, and von Willebrand factor (231200); whereas the complex of GP lib (607759) and GP Ilia (173470) is the platelet-binding site for fibrinogen and fibronectin (134820), GP

IV is the receptor for thrombospondin (see 188060) in platelets and various cell lines. Since thrombospondins are adhesive proteins widely distributed and involved in a variety of adhesive processes, GP IV may have important functions as a cell adhesion molecule (see review of membrane glycoprotein CD36 by Greenwalt et al., 1992).

Tandon et al. (1989) isolated and characterized platelet GP IV. Tandon et al. (1989) demonstrated that GP IV is the primary receptor for adhesion of platelets to collagen. (See 120340 for another type of receptor involved in cell adhesion to collagen.)

Savill et al. (1992) found that CD36, using thrombospondin as a molecular bridge with ITGAV (193210), mediates macrophage scavenging of senescent polymorphonuclear cells undergoing apoptosis.

Armesilla and Vega (1994) demonstrated that the CD36 gene comprises 15 exons and is more than 32 kb in size. The 5-prime untranslated region is encoded by 3 exons and the 3-prime untranslated region by 2 exons. The transcription initiation site is located 289 nucleotides upstream from the translation start codon. Using a CD36 genomic probe, Fernandez-Ruiz et al. (1993) mapped the CD36 gene to 7qll.2 by fluorescence in situ hybridization.

Yufu et al. (1990) found decreased glycosylation of platelet membrane glycoprotein IV in a 45-year-old male who had been found to have macrothrombocytopenia on routine blood examination. He had no history of hemorrhagic diathesis. Four members of his family, including a son, also had macrothrombocytopenia without notable bleeding tendency. The Bernard-Soulier syndrome (231200) is another form of familial macrothrombocytopenia, which is caused by a defect in platelet glycoprotein

lb. Van Schravendijk et al. (1992) showed that normal human erythrocytes express CD36; thus, this adhesion molecule may have a biologic role in normal individuals as well as in the pathology of falciparum malaria. In a thrombocytopenic patient with refractoriness to HLA-matched platelet transfusion, Ikeda et al. (1989) demonstrated a new platelet-specific antigen, Nak(a); Tomiyama et al. (1990) demonstrated that the corresponding antibody reacts with GP IV. Yamamoto et al. (1990) demonstrated that Nak(a)-negative platelets lacked detectable GP IV. These individuals with deficiency of platelet GP IV are apparently healthy and suffer no obvious hemostatic problems, but they are at risk for developing isoantibodies after infusion of Nak(a)-positive platelets.

CD36 deficiency can be divided into 2 subgroups (Yamamoto et al., 1994). The type I phenotype is characterized by platelets and monocytes/macrophages exhibiting CD36 deficiency; indeed, probably no cells express CD36. The type II phenotype lacks the surface expression of CD36 in platelets, but expression in monocytes/macrophages is near normal.

CD36 deficiency is present in 2 to 3% of Japanese, Thais, and Africans, but in less than 0.3% of Americans of European descent (Ikeda et al., 1989; Yamamoto et al.,

1990; Kashiwagi et al, 1995; Urwijitaroon et al., 1995; Curtis and Aster, 1996).

Patients with CD36 deficiency show reduced fatty acid uptake in heart; up to 40% of

Japanese patients with hereditary hypertrophic cardiomyopathy (192600) have CD36 deficiency (Tanaka et al., 1997). Others have reported an increased frequency of CD36 deficiency in Japanese patients with coronary heart disease, and the occurrence of type II diabetes with either insulin resistance or hypertriglyceridemia, hypertension, and coronary heart disease in patients with CD36 deficiency. Lee et al. (1999) found that CD36 deficiency is frequent in sub-Saharan Africans, as it is in Asians, and that development of anti-CD36 can lead to serious complications in multiply transfused patients, such as those with sickle cell disease.

Aitman et al. (2000) found that African populations contain an exceptionally high frequency of mutations in CD36. Unexpectedly, these mutations that cause CD36 deficiency are associated with susceptibility to severe cerebral malaria, suggesting that the presence of distinct CD36 mutations in Africans and Asians is due to some selection pressure other than malaria.

CD36 is a major receptor for Plasmodium falciparum-infected erythrocytes. In 475 adult Thai patients with P. falciparum malaria, Omi et al. (2003) screened for variation in the CD36 gene and examined possible association between CD36 polymorphisms and the severity of malaria. They identified 9 CD36 polymorphisms with a frequency of more than 15% for the minor allele. Of these, the -14T-C allele in the upstream promoter region and the -53G-T allele in the downstream promoter region were significantly decreased in patients with cerebral malaria compared with those with mild malaria. Linkage disequilibrium (LD) analysis between the 9 common polymorphisms revealed 2 blocks with strong LD in the CD36 gene; the -14T-C and -

53G-T polymorphisms were within the upstream block of 35 kb from the upstream promoter to exon 8. Another polymorphism, consisting of 12 TG repeats in intron 3 (173510.0004), was strongly associated with reduction in the risk of cerebral malaria. Omi et al. (2003) demonstrated by RT-PCR amplification that this IVS3(TG)12 polymorphism is involved in the nonproduction of the variant CD36 transcript that lacks exons 4 and 5. Because exon 5 of the gene is known to encode the ligand- binding domain for P. falciparum-infected erythrocytes, IVS3(TG)12 itself or a primary variant on the haplotype with IVS3(TG)12 may be responsible for protection from cerebral malaria in Thailand. Results of this study suggested that LD mapping has potential for detecting a disease-associated variant on the basis of haplotype blocks. Kashiwagi et al. (2001) found that the 478C-T mutation (173510.0001) had a greater than 50% frequency in their series of Japanese patients; however, none of the 4 subjects who possessed isoantibodies against CD36 had the 478C-T mutation, suggesting that this mutation prevents the production of isoantibodies against CD36. Kashiwagi et al. (2001) stated that the incidence of type I and type II subjects in Japanese is 0.3% and 4.0%, respectively. Type I subjects may produce isoantibodies against CD36 during pregnancy or transfusion, leading to neonatal immune thrombocytopenia, refractoriness to HLA-matched platelet transfusion, or posttransfusion purpura.

Griffin et al. (2001) reported a glucose-mediated increase in CD36 mRNA translation efficiency that resulted in increased expression of CD36, and proposed that a link between diabetes and atherosclerosis may be indicated by the findings. Expression of CD36 was increased in endarterectomy lesions from patients with a history of hyperglycemia. Macrophages that were differentiated from human peripheral blood monocytes in the presence of high glucose concentrations showed increased expression of cell surface CD36 secondary to an increase in translational efficiency of CD36 mRNA. They obtained similar data from primary cells isolated from human vascular lesions. They concluded that increased translation of macrophage CD36 transcripts under high glucose conditions provides a mechanism for accelerated atherosclerosis in diabetics.

The human insulin-resistance syndromes—type II diabetes, obesity, combined hyperlipidemia, and essential hypertension-are complex disorders. Determining the genetic basis of such disorders is difficult; see Reaven (1988), Groop et al. (1989),

Reaven et al. (1996), and Aitman et al. (1997). The spontaneously hypertensive rat (SHR) is insulin resistant and a model of these human syndromes. Quantitative trait loci (QTLs) for SHR defects in glucose and fatty acid metabolism, hypertriglyceridemia, and hypertension map to a single locus on rat chromosome 4. Aitman et al. (1999) combined use of cDNA microarrays, congenic mapping, and radiation hybrid mapping to identify a defective SHR gene, Cd36 (also known as Fat, as it encodes fatty acid translocase), at the peak of linkage to these QTLs. They found that SHR Cd36 cDNA contains multiple sequence variants, caused by unequal genomic recombination of a duplicated ancestral gene. The encoded protein product was undetectable in SHR adipocyte plasma membrane. Transgenic mice overexpressing Cd36 had reduced blood lipids. Aitman et al. (1999) concluded that Cd36 deficiency underlies insulin resistance, defective fatty acid metabolism, and hypertriglyceridemia in SHR, and may be important in the pathogenesis of human insulin-resistance syndromes.

Pravenec et al. (1999) developed a congenic strain in which the segment of chromosome 4 with the deletion of Cd36 in SHR was replaced by a corresponding segment from the normotensive BN rat. This replacement induced significant reduction of systolic blood pressure and ameliorated fructose-induced glucose intolerance, hyperinsulinemia, and hypertriglyceridemia.

Results that appeared to support an etiologic role of Cd36 in the spontaneously hypertensive rat (SHR) had been reported. However, Gotoda et al. (1999) showed that the Cd36 mutation is absent in the original SHR strains, maintained since their development in Japan, and questioned the etiologic relevance of the Cd36 mutation to insulin resistance in SHR. They emphasized the genetic and phenotypic heterogeneity of SHR that must be considered when investigating this important animal model.

Pravenec et al. (2001) showed that transgenic expression of Cd36 in spontaneously hypertensive rats (SHR) ameliorates insulin resistance and lowers serum fatty acids. The results provided direct evidence that Cd36 deficiency can promote defective insulin action and disordered fatty-acid metabolism in spontaneous hypertension.

Coburn et al. (2000) found that Cd36 null mice had reduced uptake of 2 iodinated fatty acid analogs in heart, skeletal muscle, and adipose tissue compared with wildtype mice. Reduced uptake was associated with decreased incorporation of palmitate into triglycerides and a higher accumulation of palmitate in diglycerides, which could not be explained by changes in the specific activities of long-chain acyl- CoA synthetase (see 604443) and diacylglycerol acyltransferase (DGAT; 604900). These activities were similar in wildtype and Cd36 null mice. Cobum et al. (2000) concluded that CD36 facilitates a large fraction of fatty acid uptake by heart, skeletal muscle, and adipose tissues. Importantly, CD36 deficiency, rather than some other defect, explains the defective myocardial fatty acid uptake observed in humans.

CD36 deficiency is divided into 2 subgroups: in type I deficiency, neither platelets nor monocytes express CD36, whereas in type II deficiency, monocytes express CD36 in spite of the lack of platelet CD36. In a Japanese patient with platelet glycoprotein IN deficiency, Kashiwagi et al. (1993) demonstrated a C-to-T transition at nucleotide 478 resulting in a change at codon 90 from CCT (pro) to TCT (ser). The P90S mutation was found in 4 of 5 patients with GP IV deficiency type II.

Kashiwagi et al. (1995) demonstrated that monocyte CD36 cDΝA from 2 type π deficiency patients was heterozygous for C478 and T478 forms of CD36, while platelet CD36 cDΝA of these subjects consisted of only the T478 form. In a type I deficient subject, both platelet and monocyte CD36 cDΝA showed only the T478 form. Expression assay using the C478 or T478 form of CD36 cDΝA in transfected cells revealed that there was an 81-kD precursor form of CD36, and that the maturation of the 81-kD precursor form to the 88-kD mature form of CD36 was markedly impaired by the substitution. The mutated precursor form of CD36 was subsequently degraded in the cytoplasm. These results indicated that the C-to-T substitution at nucleotide 478 of the cDΝA (which corresponds to nucleotide 12293 in exon 4 of the genomic sequence) directly leads to CD36 deficiency via defects in posttranslational modification and that this substitution is the major defect underlying CD36 deficiency. Thus, type I individuals are presumably homozygous for P90S, whereas type II individuals are heterozygous.

Aitman et al. (2000) identified a common T-to-G transversion at nucleotide 1264 of the CD36 gene that results in a premature termination codon in exon 10. This allele accounted for 80% of mutant alleles detected and was present in 9.9% of control UK

Afro-Caribbeans. This allele was overrepresented in patients with severe cerebral malaria (see 248310). Aitman et al. (2000) identified a compound mutation in exon 12 of the CD36 gene: a G-to-C transversion in nucleotide 1439 resulting in a ala-to-pro substitution, and a frameshift deletion at nucleotide 1444. This mutation was found in the carrier state in 3.7% of control Gambians and 0.3% of healthy UK Afro-Caribbeans. This allele was overrepresented in patients with severe cerebral malaria (see 248310).

Omi et al. (2003) found that a (TG)12 repeat in intron 3 of the CD36 gene was strongly associated with reduction in the risk of cerebral malaria (see 248310) in Thailand. They showed that this variant is involved in the nonproduction of the variant CD36 transcript that lacks exons 4 and 5. Exon 5 of the CD36 gene is known to encode the ligand-binding domain of P. falciparum-infected erythrocytes.

NCOR2: nuclear receptor co-repressor 2 http://www.ncbi.nlm.nih.gov/LocusLink/LocRpt.cgi 71=9612 http://www.ncbi.nlm.nih.gov/entrez/dispomim.cgi?id=600848

LocusID: 9612

Locus Type: gene with protein product, function known or inferred. Product: nuclear receptor co-repressor 2

Alternate Symbols: SMRT, CTG26, SMRTE, TRAC1, TNRC14, TRAC-1

Alias: silencing mediator for retinoid and thyroid hormone receptors

Function Submit GeneRIF (All Pubs) : GeneRIF: Gene References into Function: 12840002 SANT motif interprets the histone code and promotes histone deacetylation 12441355 a significant role of SMRT in modulating androgen receptor transcriptional activity 11972046 Interactions that determine the assembly of a retinoid X receptor/corepressor complex 12388540 SMRT has a role as a coactivator for thyroid hormone receptor T3Ralpha from a negative hormone response element 11929749 Sstable binding of the Stat5-RARalpha fusion protein to corepressor SMRT is accompanied by an impaired response to differentiation signals in hematopoietic cells 12139968 The silencing mediator of retinoic acid and thyroid hormone receptors can interact with the aryl hydrocarbon (Ah) receptor but fails to repress Ah receptor- dependent gene expression. 11929748 the effect of STAT5b-RARalpha on the activity of myeloid transcription factors including STAT3, and STAT5 as well as its molecular interactions with the nuclear receptor corepressor, SMRT, and nuclear receptor coactivator, TRAM-1.

Relationships : Mouse Homology Maps: NCBI vs. MGD 5 cM Ncor2 Hs Mm

Map Information : Chromosome: 12 mv Cytogenetic: 12q24 OMJ-M Markers: Chr. - D12S1611 D12S1611 Chr. 12 STS-D60472 mv Chr. 12 RH80134 mv Chr. 12 RH80812 mv Chr. 12 RH79918 mv NCBI Reference Sequences (RefSeq) : Category: PROVISIONAL mRNA: NM_006312 Protein: NP_006303 nuclear receptor co-repressor 2 BL Domains: Chromosome segregation ATPases [Cell division and chromosome partitioning] score: 99 Myb-like DNA-binding domain. This family contains the DNA binding domains from Myb proteins, as well as the SANT domain family score: 141 GenBank Source: AF113003 Category: NCBI Genome Annotation Genomic Contig: NT_009755 gb sv mv ev mm Haplotype reference Annotation for this locus: Evidence: supported by alignment with mRNA mRNA: NM_006312 Protein: NP_006303 NUCLEAR RECEPTOR COREPRESSOR 2; NCOR2 Alternative titles; symbols SILENCING MEDIATOR FOR RETINOID AND THYROID HORMONE RECEPTORS; SMRT Gene map locus 12q24

Transcriptional silencing mediated by nuclear receptors is important in development, differentiation, and oncogenesis. The mechanism underlying this effect is one key to understanding the molecular basis of hormone action. Chen and Evans (1995) identified a receptor-interacting factor, which they designated SMRT, as a silencing mediator (corepressor) for retinoid and thyroid-hormone receptors. The authors cloned the SMRT gene from a HeLa cell cDNA library and showed that the association of SMRT with receptors, both in solution and bound to DNA-response elements, is destabilized by ligand. The interaction with mutant receptors correlates with their transcriptional silencing activities. The repressor function of thyroid hormone receptor and retinoic acid receptor appears to involve direct interaction with SMRT, possibly in a way that stabilizes or promotes their interaction with the transcription factor TFIIB (189963). Chen and Evans (1995) showed that ligand causes the dissociation of SMRT from the receptor, triggering the activation process. SMRT appears to belong to a family of corepressors that Horlein et al. (1995) referred to as TRAC (thyroid hormone- and retinoic acid receptor-associated corepressor). (See NCOR1; 600849.) Ordentlich et al. (1999) described splicing variants of the human and mouse SMRT genes. The forms of human and mouse SMRT that included a 1,000-amino acid extension revealed striking homology to the amino terminus of nuclear receptor corepressor-1 (NCOR1). Binding of ligand to nuclear hormone receptors induces a conformation that attracts coactivator proteins containing a Leu-x-x-Leu-Leu motif, the so-called NR box. Hu and Lazar (1999) showed that NCOR1 and SMRT contain sequences that are similar to the NR box and are repeated in each of 2 nuclear hormone receptor interaction domains. Hu and Lazar (1999) called this box (IJI-x-x-I/V-I) the CoRNR (corepressor/ nuclear receptor; 'comer') box. The corner box is required for nuclear hormone receptor interaction and the comer box peptides specifically block corepressor interaction in vitro and repression in vivo. Sequences flanking the comer box determine nuclear hormone receptor specificity. Thus, Hu and Lazar (1999) concluded that the key feature of hormone action, differential recognition of unliganded and liganded nuclear hormone receptors by coactivators and corepressors, is due to very subtle differences between CoRNR and NR boxes.

Fischle et al. (2002) showed that the catalytic domain of HDAC4 (605314) interacts with HDAC3 (605166) via the transcriptional corepressor NCOR2. All experimental conditions leading to the suppression of HDAC4 binding to NCOR2 and to HDAC3 resulted in loss of enzymatic activity associated with HDAC4. These observations indicated that class II HDACs regulate transcription by bridging the enzymatically active NCOR2-HDAC3 complex and select transcription factors. Jiang et al. (2001) determined that the NCOR2 gene contains 45 exons. By FISH, Ordentlich et al. (1999) mapped the NCOR2 gene to chromosome 12q24.

THSD2; thrombospondin, type L domain 2 LocusID: 84870

This gene encodes a protein similar to thrombospondin type 1 domain-containing proteins. In addition, the protein contains a furin-like cysteine-rich region. Furin-like repeat domains have been found in a variety of eukaryotic proteins involved in the mechanism of signal transduction by receptor tyrosine kinases. The function of this protein has not been determined. Locus Type: gene with protein product, function known or inferred. Product: thrombospondin Alternate Symbols: PWTSR, FLJ14440 Alias: thrombospondin-like gene

Function Submit GeneRIF (All Pubs) GeneRIF: Gene References into Function: 12463421 the gene is a novel member of thrombospondin type I repeat supergene family (hPWTSR)

12643280 a polymorphism in thrombospondin subtly but significantly sensitizes the calcium-binding repeats to removal of Ca2+ and thermal denaturation

Gene OntologyTM: Term Evidence Source Pub electron transporter activity IE A GO A electron transport JJEA GOA

Relationships Mouse Homology Maps: NCBI vs. MGD 2810459H04Rik Hs

Map Information Chromosome: 6 mv Cytogenetic: 6q22.33 RefSeq Markers: Chr. 6 SHGC-57345 mv Chr. 6 SHGC-57647 mv Chr. 6 RH94253 mv

NCBI Reference Sequences (RefSeq) Category: REVIEWED mRNA: NM 032784 Protein: NP_116173 thrombospondin BL Domains: KOG3525: Subtilisin-like proprotein convertase [Posttranslational modification, protein turnover, chaperones] score: 293 GenBank Source: AF251057

Category: NCBI Genome Annotation Genomic Contig: NT_025741 gb sv mv ev mm Haplotype reference Annotation for this locus: Evidence: supported by alignment with mRNA mRNA: NM_032784 Protein: NP_116173 BL

Related Sequences Nucleotide Type Protein AF086298 m AF251057 m AAK34947 BL BC022367 m AAH22367 BL

Additional Links UniGene: Hs.135254

Summary of Invention.

The present invention relates to a newly discovered association between sequence variants within ADAM12, THSD2, CD36 and NCOR2 genes and osteoarthritis and related clinical manifestations of osteoarthritis namely, continued pain in the joint, joint space narrowing and osteophytes and increase in joint space narrowing and osteophytes over time as measured by X-rays. Identification of the allelic sequence variants present provides information that assists in characterizing individuals according to their risk of osteoarthritis. In the methods of the invention, the genotype of the AD AMI 2, THSD2, CD36 and NCOR2 genes is determined in order to provide information useful for assessing an individual's risk for particular joint diseases, in particular, osteoarthritis. Individuals who have at least one allele statistically associated with osteoarthritis possess a factor contributing to the risk of osteoarthritis. The statistical association of ADAM12,

THSD2, CD36 and NCOR2 genes alleles (sequence variants) with osteoarthritis is shown in the examples.

As each of these genes (ADAM12, THSD2, CD36 and NCOR2) is but one component of the complex system of genes involved in the development of osteoarthritis, the effect of the each locus is expected to be relatively small. Still, genotyping for the ADAM12, THSD2, CD36 and NCOR2 genes will contribute information that is, nevertheless, useful for a characterization of an individuals predisposition toward osteoarthritis, or, given that an individual has already been diagnosed with radiographic OA, toward increased joint narrowing, the presence of osteophytes and of pain in the joints which are clinical manifestation of more severe disease.

In a preferred embodiment of the invention, genotyping is carried out using oligonucleotide probes specific to variant sequences. Preferably, a region of the

ADAM12, THSD2, CD36 and NCOR2 genes which encompasses the probe hybridization region is amplified prior to, or concurrent with, the probe hybridization. Probe-based assays for the detection of sequence variants are well known in the art.

In the methods of the invention, the genotype of the ADAM12, CD36, NCOR2 and

THSD2, genes is determined in order to provide information useful for assessing an individual's risk for particular joint diseases, in particular, osteoarthritis. Individuals who have at least one allele statistically associated with osteoarthritis possess a factor contributing to the risk of osteoarthritis. The statistical association of ADAM12, CD36, NCOR2 and THSD2 genes alleles (sequence variants) with osteoarthritis and osteoarthritis related phenotypes is shown in the examples. The presence of one or two arginines at position 48 of the ADAM12 gene (Gly48Arg) is associated with larger change in osteophyte grade in the knee over a 10 year period (p<0.004) and with an increased risk of presence of osteophytes in the knee (O.R. = 1.92 p<0.002). It is also associated with increased risk of radiographic knee OA defined as a Kellgren-Lawrence grade of 2 or more (O.R. 1.84 p<0.004). The Arg-

Arg genotype at ADAM12 position 48 is also associated with 6% higher bone mineral density at the neck of femur than the Gly-Gly genotype and 3.6 % higher than the Gly-Arg genotype (p<0.038) and with 3.5% higher bone mineral density of the spine than the combined Gly-Gly and Gly-Arg genotypes (p<0.047).

Carrying an A instead of a C at position -120 (5') of the CD36 gene (promoter) is associated with decreased risk of presenting osteophytes in the knee ( O.R.= 0.76 p<0.015). The same allele is also associated with decreased risk of radiographic knee OA ( O.R. = 0.77 p<0.02). The AA genotype at this site is also associated 33% greater change in osteophyte grade at the spine over time than the CC and AC genotypes

(p<0.04) and with 41% larger change in disc space narrowing at the spine over time (p<0.015).

Carrying two threonines at NCOR2 position 1699 (Thrl699Ala) is associated with increased risk for the presence of osteophytes in the knee (O.R.=1.60 p<0.011), of radiographic OA of the knee (O.R.= 1.57 p<0.015) and of joint space narrowing of the knee (O.R. =1.54 p<0.04) relative to the Thr-Ala and Ala-Ala genotypes. The Thr-Thr genotype at this position is also associated with 60% higher increase in disc space narrowing of the spine over a 9 year period compared to the Ala-Ala genotype and 23% higher increase compared to the Ala-Thr genotype (p<0.06).

Carrying the CC genotype at the 73^rd nucleotide of intron 3 (position 33906 from start of CDS) of the THSD2 gene was associated with 44% higher increase in osteophyte grade of the knee over time (p<0.041) whereas the TT genotype was associated with lower risk of presenting joint space narrowing at the knee (O.R= 0.83 p<0.075). The TT genotype was also associated with 34% less change in the Kellgren-Lawerence grade in the spine over time (p<0.003).

As each of these genes (ADAM12, CD36, NCOR2 and THSD2) is but one component of the complex system of genes involved in the development of osteoarthritis, the effect of the each locus is expected to be relatively small. Still, genotyping for the ADAM12, CD36, NCOR2 and THSD2 genes will contribute information that is, nevertheless, useful for a characterization of an individuals predisposition toward osteoarthritis, or, given that an individual has already been diagnosed with radiographic OA, toward increased joint narrowing, the presence of osteophytes and of pain in the joints which are clinical manifestation of more severe disease.

Accordingly, a method of the present invention may comprise the step of deciding that an individual is at risk of developing osteoarthritis if the results of the sample analysis show one or more of any one of the following: (a) there are two arginine residues at AD AMI 2 Gly48Arg; (b) there are two threonine residues at NCOR2 Thrl699Ala; (c) there is an adenosine (A) instead of a cytosine (C) at position -120 of the CD36 gene (promoter); or (d) there are one or two cytosines (C) instead of a tymidine (T) at position 33906 of the THSD2 gene (73rd nucleotide of intron 3).

Detailed description of the invention. To aid in understanding the invention, several terms are defined below:

The terms "Kellgren-Lawrence scale", "KL scale" , and "KL grade" are used interchangeably to refer to a validated radiographic index used by clinical investigators to determine the prevalence and severity of osteoarthritis of the hand, hip, knee, apophyseal joints of the cervical and lumbar spine, and cervical disc degeneration as described by Lane and Kremer (Rheum Dis Clin North Am 1995 May;21(2):379-94). The term "JSN" denotes joint space narrowing which results from degradation and loss of articular cartilage at a joint such as the knee, the hip, the hand, the cervical disc or the apophyseal joints of the cervical and lumbar spine.

The term "ADAM12 gene" refers to the genomic nucleic acid sequence that encodes the disintegrin and metalloproteinase domain 12 (meltrin alpha) protein.

The term "CD36 gene" refers to the genomic nucleic acid sequence that encodes the collagen type I receptor protein also known as thrombospondin receptor.

The term "NCOR2 gene" refers to the genomic nucleic acid sequence that encodes the nuclear receptor co-repressor 2 protein also known as silencing mediator for retinoid and thyroid hormone receptors.

The term "THSD2 gene" refers to the genomic nucleic acid sequence that encodes the thrombospondin, type I, domain 2 protein.

The gene sequence of a Human mRNA for the two alternative isoforms of ADAM12 are provided at GenBank accession numbers mRNA NM_003474 (SEQ ID NO: 1) and NM_021641 (SEQ ID NO: 2). The protein sequences of the two alternative isoforms are provided at GenBank accession numbers NP_003465 (SEQ ID NO:3) and NP_067673 (SEQ ID NO: 4).

The gene sequence of a Human mRNA for CD36 is provided at GenBank accession number mRNA NM_000072 (SEQ ID NO: 5). The protein sequence is provided at GenBank accession number NP_000063 (SEQ ID NO:6).

The gene sequence of a Human mRNA for NCOR2 is provided at GenBank accession number mRNA NM_006312. (SEQ ID NO: 7). The protein sequence is provided at GenBank accession number NPJ306303 (SEQ ID NO:8). The gene sequence of a Human RNA for THSD2 is provided at GenBank accession number mRNA NMJ332784 . (SEQ ID NO: 9). The protein sequence is provided at GenBank accession number NP_116173 (SEQ ID NO: 10).

The above sequences are fully described in the attached drawings as Figures 1 to 10.

The term "allele" as used herein refers to a sequence variant of the gene. Alleles are identified with respect to one or more polymorphic positions, with the rest of the gene sequence unspecified. For example a the ADAM12, THSD2, CD36, and NCOR2 allele may be defined by the nucleotide present at a single SNP or by the nucleotides present at a plurality of SNPs. Examples of the ADAM12, THSD2, CD36, and NCOR2 genes can be found in table 1.

The term "predisposing allele" refers to an allele that is positively associated with the presence of the disease of the joints such as osteoarthritis, or with clinical manifestations that correlate with a more severe or faster progressing form of the disease, such as radiographic measurements of joint narrowing and their change over time. The presence of a predisposing allele in an individual could be indicative that the individual has an increased risk for the disease relative to an individual without the disease.

ADAM12, THSD2. CD36. and NCOR2 SNPs. In certain embodiments, the genotype of a single SNP at any of the AD AMI 2, THSD2, CD36, and NCOR2 genes can be used to determine an individuals risk for osteoarthritis. In other embodiments, the genotypes of a plurality of SNPs at the ADAM12, THSD2, CD36, and NCOR2 genes can be used. Table 1 shows examples of ADAM12, THSD2, CD36, and NCOR2 SNPs

Table 1

Genotyping Methods In the methods of the present invention the alleles present in a sample are detected by identifying the nucleotide present at one or more of the polymorphic sites. Any type of tissue containing the ADAM12, THSD2, CD36, and NCOR2 nucleic acid may be used for determining ADAM12, THSD2, CD36, and NCOR2 genotypes of an individual. A number of methods are known in the art for identifying the nucleotide present at a single nucleotide polymorphism. The particular method used to identify the genotype is not a critical aspect of the invention. Although considerations of cost, performance, and convenience will make particular methods to be preferred over others, any method that can identify the nucleotide present will provide the information required to identify the genotype.

The ADAM12, THSD2, CD36, and NCOR2 alleles can be identified by DNA sequencing methods, such as the chain termination method described (Sanger et al. 1977 Proc Natl Acad Sci 74:5463-5467) which is well known in the art.

The ADAM12, THSD2, CD36, and NCOR2 alleles can be identified using amplification-based genotyping methods. A preferred method is the polymerase chain reaction (PCR) which is well known in the art and described in US Patent Nos. 4,683,195; 4,683,202; 4,965,1888. Other suitable amplification methods include the ligase chain reaction (Wu & Wallace 1988 Genomics 4:560-569); the strand displacement assay (Walker et al. 1992, Proc Natl Acad Sci USA 89:392-396) and several transcription-based amplification systems, including the methods described in

US Patent Nos 5,437,990 and 5,409,818.

Genotyping can also be carried out by detecting the ADAM12, THSD2, CD36, and NCOR2 mRNAs. Amplification of RNA can be carried out by first reverse- transcribing the target RNA using, for example, a viral reverse transcriptase, and then amplifying the resulting cDNA, or using a combined high temperature reverse- transcription -polymerase chain reaction (RT-PCR) as described for example in US Patent No.5,310,652.

The ADAM12, THSD2, CD36, and NCOR2 alleles can be identified using primer extension or allele-specific amplification methods, which are based on the inhibition by a terminal primer mismatch on the ability of a DNA polymerase to extend the primer. Allele specific amplification- or extension-based methods are described for example in US Patent No. 4,8451, 331.

An alternative probe-less method, referred herein as kinetic-PCR method in which the generation of amplified nucleic acid is detected by monitoring the increase in the total amount of double-stranded DNA in the reaction

Whatever the method for determining which oligonucleotides of the invention selectively hybridize to the ADAM12, THSD2, CD36, and NCOR2 allelic sequences in a sample, the central feature of the typing method involves the identification of the

ADAM12, THSD2, CD36, and NCOR2 alleles present in the sample by detecting the variant sequences present.

The present invention also relates to kits, container units comprising useful components for practicing the present method. A useful kit can contain oligonucleotide probes specific for the ADAM12, THSD2, CD36, and NCOR2 alleles. Such kits may also suitably comprise instructions for use in accordance with a method of the present invention.

Preferred features for the second and subsequent aspects of the invention are as for the first aspect mutatis mutandis.

The examples of the present invention presented below are provided only for illustrative purposes and not to limit the scope of the invention. Numerous embodiments of the invention within the scope of the claims that follow the example will be apparent to those of ordinary skill in the art from reading the foregoing text and following examples.

Reference is made herein and in the Examples to a number of Tables in which:

Table 1 provides examples of ADAM12, THSD2, CD36, and NCOR2 SNPs useful in the methods of the invention Table 2 provides examples of amplification primers for the ADAM12, THSD2, CD36 and NCOR2 genes (SEQ ID NO 9 -14) useful in the methods of the invention

Table 3 provides detection probes shown in the 5' to 3' orientation used for detecting

ADAM12, THSD2, CD36 and NCOR2 alleles (SEQ ID NO 15 -20) useful in the methods of the invention

Table 4 provides examples of genetic association between ADAM12, THSD2, CD36 and NCOR2 polymorphisms and the risk of radiographic osteoarthritis of the knee

Table 5 provides examples of genetic association between ADAM12, THSD2, CD36 and NCOR2 polymorphisms and the risk of presenting osteophytes in the knee

Table 6 provides examples of genetic association between an NCOR2 polymorphism and the risk of presenting joint space narrowing in the knee

Table 7 provides examples of genetic association between an ADAM12 polymorphism and the progression of KL and osteophyte grades over a 10 year period.

Table 8 provides examples of association of NCOR2, CD36 and THSD2 genotypes with change in disc space narrowing, osteophyte grade and KL grade over time, (a) NCOR2 genotypes associated with change in DSN, F(l, 685) = 3.56 ρ<0.059; (b) association of CD36 genotype with change in disc space narrowing over time, F(l,

709)= 5.91 p<0,015; (c) association of CD36 genotype with change in spine osteophyte radiographic grade over time, F(l, 709)= 4.21 p<0.040; (d) association of THSD2 with change in KL grade in the spine, F(l, 678)= 8.63 ρ<0.003

Table 9 provides examples of association of an ADAM12 polymorphism with hip and spine BMD. (a) F(l,204) = 4.37 p<0.038 (adjusted for age and BMI of subjects); (b) F (1,208)= 3.99 p<0.047 (adjusted for age and BMI of subjects) Table 10 provides examples of association of an ADAM12 polymorphisms with knee OA among men. (a) Chi-squared (1 df) = 2.12 p<0.15; (b) Chi-squared (1 df) = 8.60 p<0.003

Table 11 provides examples of association of an ADAM12 haplotypes with knee OA in men and women, the p- value as well as the odds ratios with the corresponding 95% confidence intervals are shown.

Reference is also made to a number of drawings in which:

Figure 1 shows SEQ ID NO: 1 (NM_003474) mRNA for ADAM12 isoform 1

Figure 2 shows SEQ ID NO: 2 (NM_021641) mRNA sequence for ADAM12 isoform 2

Figure 3 shows SEQ ID NO: 3 (NP_003465) protein sequence for ADAM12 isoform 1

Figure 4 shows SEQ ID NO: 4 (NP_067673) protein sequence for ADAM12 isoform

2

Figure 5 shows SEQ ID NO: 5 mRNA sequence for CD36 (NM_000072)

Figure 6 shows SEQ ID NO: 6 protein sequence for CD36 (NP_000063)

Figure 7 shows SEQ ID NO: 7 mRNA sequence for NCOR2 (NM_006312)

Figure 8 shows SEQ ID NO: 8 protein sequence for NCOR2 (NP_006303)

Figure 9 shows SEQ ID NO: 9 mRNA sequence for THSD2 (NM_032784) Figure 10 shows SEQ ID NO: 10 protein sequence for THSD2 (NP_116173)

Assays Useful for Determining the Association of a Polymorphism with osteoporosis Preventive treatment for osteoporosis is most effective at the time when bone loss is increasing and before the bones have become fragile and prone to fracturing.

Established diagnostic techniques use x-^ray and ultrasonography to measure skeletal parameters of bone size, volume and mineral density to predict fracture risk and to assess response to therapy. Such measurements give a static value which can be compared to normal values to aid diagnosis of low bone mass and fracture risk (Schott, Cormier et al. 1998). The World Health Organization defines osteoporosis as present when the bone mineral density levels are more than 2.5 standard deviations below the young normal mean. The various techniques used to measure bone mineral density are:

Dual energy X-ray absorptiometry (DXA) - used to measure bone mass at the lumbar spine ■ and hip, but it can also be applied to measuring total skeletal bone mass, soft- tissue composition and other regional bone measurements. Considered the gold standard for BMD measurement.

High-resolution quantitative computed tomography (QCT) - highly sensitive, accurate and specific spinal measurements. This technique is more costly and involves higher radiation doses than other techniques and is not widely available.

Single energy x-ray absorptiometry (SXA) - provides accurate radius BMD measurements.

■ Quantitative ultrasound (QUS) - new and promising technique which may have applicationsin both BMD measurement and assessment or architectural deterioration of bone tissue. Recent studies suggest QUS of calcaneus bone predicts hip fracture as well as DXA (Hans, Dargent-Molina et al. 1996). ^■ An alternative method to predict fracture independently of bone mass is to measure bone turnover. High turnover (bone resorption and formation) is associated with rapid bone loss and is likely to contribute to micro- architectural deterioration (Ross and Knowlton 1998). This is a dynamic measurement which is assessed with biochemical markers in urine or serum and can be used very effectively in therapy monitoring in preference to BMD measurements which alter more slowly (results of PEPI trial and Merck Research Laboratories). When used in combination with bone mass assessment, biomarkers can provide more accurate fracture predictions over bone mass measurement alone. Several markers for bone resorption (deoxypyridinoline crosslinks), and bone formation (bone alkaline phosphatase, osteocalcin) have been developed for use in diagnostic kits.

Example 1; Genotyping Protocol.

Fluorescently labeled primers were synthesized and PCR was performed on 47 DNAs from a Coriel derived Human Diversity Panel. The PCR products were electrophoresed on an ABI 377 machine and 8% nondenatureing, 12 CM SSCP gels were used. The resulting traces were aligned in ABI Genotyper software and where variant traces (indicating underlying polymorphism) were found, examples of each variant type were sequenced.

Genomic DNA was set on 384 well plates. Genotyping was carried out using Assays- on-Demand™ SNP Genotyping products from Applied Biosystems (Applera Corporation, Foster City, CA), where the primers are labeled with a reporter dye at the 5' end of each probe: the VIC dye linked to the 5' end of one of the allele' s probe and the FAM dye was linked to the 5' end of the other allele' s probe.

The PCR amplification was carried out in a total reaction volume of 5/_L. Reactions were carried out using xlO Universal PCR master mix x40 ppMix (Containing both

Probes and Primers) (Applied Biosystems, Applera Corporation, Foster City, CA).

Amplification was carried out in a Kbiosystems Super Duncan thermal cycler, using the specific temperature cycling profile shown below. Pre-reaction incubations: 95°C for 10 minutes (AmpliTaq GOLD activation, template denaturation) 40 cycles; denature: 95°C for 15 seconds anneal / extend 60°C for 60 seconds. Amplification of one regions of the gene was carried out using the primer pair shown bellow in the 5' to 3' orientation.

Table 2 shows amplification primers for the ADAM12, THSD2, CD36 and NCOR2 genes

Table 2

Preferred probes used to identify the nucleotides present at the SNPs in the amplified ADAM12, THSD2, CD36 AND NCOR2 nucleic acids are described in Table 3. The probes are shown in the 5' to 3' orientation.

Table 3 shows detection probes shown in the 5' to 3' orientation used for detecting AD AMI 2, THSD2, CD36 and NCOR2 alleles

Table 3

Cleavage by the AmpliTaq Gold® DNA polymerase separates the reporter dye from the quencher dye, which results in increased fluorescence by the reporter. The increase in fluorescence signal occurs only if the amplified target sequence is complementary to the probe. Thus, the fluorescence signal generated by PCR amplification indicates which alleles are present in the sample. A substantial increase in VIC dye fluorescence indicates homozygosity for Allele 1. A substantial increase in FAM dye fluorescence indicates homozygosity for Allele 2. A substantial increase in both fluorescent signals Indicates Allele 1 -Allele 2 heterozygosity. Each 384 well plate was read using SDS instrumentation and alleles were called using the SDS software package (all of the above manufactured by Applied Biosystems part of Applera Corporation, Foster City, CA).

Example 2; Association with Prevalence of Osteoarthritis of the Knee Subjects. Genotyping of ADAM12, THSD2, CD36 and NCOR2 SNPs was carried out using the genotyping method described in Example 1 on 749 English women of Caucasian race. X-rays of the knee with an antero-posterior. At baseline, all subjects underwent anteroposterior weight-bearing radiography of the knee, taken in full extension Radiographs of the knee were repeated 10 years later. At both baseline and 10 years later, all subjects completed a standardized questionnaire on medical history. Blood and urine samples were collected at baseline and annually thereafter and stored at - 45°C in aliquots, without freeze-thawing until the time of analysis. 749 women gave their informed consent to extract DNA from the blood samples donated and to use the DNA for genetic analyses.

Definition of OA. Knee radiographs were taken with subjects in the weight-bearing, fully extended position. Films were read by a trained examiner for the presence of knee osteophytes and JSN, using a 0-3 scale of severity. A Kellgren and Lawrence (KL) grade >= 2 was used as diagnosis of radiographic osteoarthritis. Knee OA was defined as KL grade >=2. Baseline groups were compared using a composite radiography scale, which was calculated by adding all possible individual scores for JSN and osteophytes. Radiographic joint space narrowing (JSN) is the hallmark of progression in osteoarthritis. Osteophytes are bony outgrowths that occur at the joint margins strongly correlating with the severity of the disease. The knee with the highest KL grade at year 10 was chosen for each volunteer.

Statistical Methods:

Odds ratio: the odds ratio is the ratio of the odds of an individual selected from the study being a case (e.g. having OA) given that the subject carries the genetic risk factor (genotype) to the odds that he/she is affected given that the individual does not carry the genetic risk factor. The odds ratio gives a reasonable estimate of the relative risk when (1) the proportion of subjects classified with the disease is small (2) the cases and controls are random samples from the same relevant population group.

Another approach commonly used consists in comparing the mean of a clinical trait of particular relevance to the disease (e.g. JSN grade, Kellgren-Lawrence, etc) between genotypes. This enables the investigator to determine what proportion of the variance in a quantitative trait (e.g. KL grade) is determined by the presence of a genetic risk factor (genotype carried by subject). The statistical technique used in this case is either an analysis of variance (ANOVA) or of covariance (ANCOVA) or a t-test.

In the preferred embodiment of the invention in order to examine the role of the SNP genotypes as a predictors of OA, joint space narrowing and logistic regression models were fitted. These models included the SNP genotypes, age and BMI. Analyses of variance (ANOVA) were used to compare the mean radiographic grade progression (change over ten years) between SNP genotypes. Age and BMI were used as covariates in the ANOVA. The outcome variables used were the change over 10 years in KL grade, osteophyte grade and joint space narrowing (JSN) grade. For the binary traits the odds ratios derived by logistic regression with 95% confidence intervals are presented. For the quantitative traits (change over ten years in radiographic grade) the adjusted means of the radiographic grade of each trait with standard error are shown.

The variants at polymorphisms at NCOR2 and ADAM12 studied were found to significantly increase the risk of OA whereas the variant at CD36 was found to decrease the risk of radiographic O A (Table 4)

Example 3: Association with the presence of osteophytes and the presence of joint space narrowing of the knee

As the osteoarthritis progresses, pieces of the articular cartilage are lost into the joint resulting in tufts and irregular areas of the surface. Fragments in the joint, some microscopic, some grossly visible, can be found that are released into the joint. But the thinning of the articular cartilage does not occur because of the loss of these pieces of tissue. It occurs because of the release of enzymes that break down the matrix and cause the actual loss of most of the articular cartilage. Grossly, the main features observed are loss of the joint space, formation of osteophytes, and thickening of the synovial membrane. At the end stage of the disease, there is full-thickness loss of articular cartilage, thickening of the subchondral bone, cysts, obvious osteophytes, and stiffness of the joint. This is the end stage of the disease which is recognized in most of the patients that are treated with total joint replacement. Radiographic joint space narrowing (JSN) is thus the hallmark of progression in osteoarthritis and osteophytes are bony outgrowths that occur at the joint margins strongly correlating with the severity of the disease.

The knee with the highest KL grade at year 10 was chosen for each volunteer. We tested by logistic regression adjusting for body mass index (BMI) and age whether the ADAM12, THSD2, CD36 and NCOR2 polymorphisms were associated with the presence of joint space narrowing and of osteophytes at year 10 of the study in the 749 participating women.

The variants at polymorphisms at NCOR2 and ADAM12 studied were found to significantly increase the risk of presenting osteophytes whereas the variant at CD36 was found to decrease the risk of presenting osteophytes (Table 5). The variant genotype at NCOR2 was also found to significantly increase the risk of presenting joint space narrowing (Table 6) whereas the variant genotype at THSD2 was associated with a lower risk of presenting joint space narrowing.

Example 4: Association with increase in the osteophyte and KL radiographic grades over time. The changes in, KL grade and osteophyte grade were computed by substracting the grade at baseline from the grade at year 10. Analysis of covariance was used to determine if there existed a statistically significant association between radiographic progression of osteoarthritis, as measured by JSN grade, osteophyte grade or KL- grade, and genotype at the ADAM12, THSD2, CD36 and NCOR2 genes. Carrying one or two copies of the C allele at the polymorphism in ADAM 12 was found to be associated with a statistically significant larger increase, and thus higher progression, in both KL grade and osteophyte grade (Table 7).

Example 5: Association with radiographic features of lumbar spine osteoarthritis.

Subjects were 710 post-menopausal women participating the Chingford study. The Chingford Study population, established in 1988, is a well-described prospective longitudinal cohort of 1,003 women seen annually and described in detail previously (Hart et al 1993).The original response rate of the sample was 78%. After 9-year followup, 821 women remained for examination ,and 796 paired radiographs were available. Genotypes for 710 of those 796 were carried out. Women from this practice are similar to women in the UK general population in terms of weight, height, and smoking characteristics.- Each woman was asked to undergo a radiographic examination of hands, knees, hips, and thoraco-lumbar spine. Local ethics committee approval was obtained for both the original study and the 9-year followup.

Radiographic assessment. Lateral lumbar spine radiographs at years 1 and 9 were taken centered on the L3 vertebrae with the subjects in the left lateral recumbent position by the same radiographer at both time points. A single trained observer blinded to patient identity and chronologic order read all radiographs. Each lateral lumbar spine radiograph was graded 0 -3 for the individual features of DSN and osteophyte formation, using the semiquantitative method reported by Lane et al, summarized as grade 0 = normal, grade 1 = mild, grade 2 = moderate, and grade 3 = severe (Lane et al 1993). Within-observer variation was assessed by test-retest analysis of 40 randomly selected radiographs from the study. Good within-observer reproducibility (kappa = 0.78 -0.89) was found. Definitions of progression were based on radiographic assessments. The association with genetic polymorphisms was determined by analysis of covariance including age and body mass index as covariates. Change in disc space narrowing radiographic grade was defined as the grade at year 9 - the grade at year 1 (baseline) divided by the grade at baseline plus one. Change in osteophyte radiographic grade was defined as the average grade for both anterior and posterior osteophytes at year 9 minus the grade at year 1 (baseline) divided by the grade at baseline plus one.

Individuals carrying two threonines at codon position 1699 of the NCOR2 gene had a greater change in disc space narrowing than those carrying only one threonine which in turn had a worse outcome than those carrying two alanines (Table 8). Also, individuals with the genotype "AA" at the -120 5' position of the gene CD36 had a statistically significantly larger (worse) change in disc space narrowing and in spine osteophyte radiographic grade than those carrying one or two "C" alleles at this position (Table 8). Further, individuals carrying genotype TT at the THSD2_int3 had a significantly lower change in KL grade at the spine than those carrying the CC or

CT genotypes (Table 8).

References :

Hart DJ,Spector TD. The relationship of obesity, fat distribution and osteoarthritis in women in the general population: the Chingford Study. J Rheumatol 1993;20:331 -5.

Lane N, Nevitt MC, Genant HK, Hochberg MC. Reliability of new indices of radiographic osteoarthritis of the hand and hip and lumbar disc degeneration. J Rheumatol 1993 ;20: 1911 -8.

Example 6; Association with bone mineral density of the spine and the neck of femur

Genetic risk factors for a disease or trait can be identified through direct analysis of candidate genes through association studies. The approach often used for such studies is the case-control design, in which a difference in allele frequency is sought between affected individuals and unrelated unaffected controls. A case control study starts with the identification of persons with the disease or other outcome variable of interest and a suitable reference or control group. The relationship of a risk factor, in this case a genetic risk factor, to the disease is examined by comparing the affected, in this case, with low bone mineral density, and unaffected (high bone mineral density) with regards to how frequency the risk factor is present.

Subjects: 210 post-menoμausal women who had not taken hormone replacement therapy (HRT) in their lifetime were assesses by DXA scan for bone mineral density of the hip (neck of femur) and spine. Subjects carrying two arginines at codon 48 of

ADAM12 had on average, after adjustment for age and body mass index , 4.7% higher bone mineral density at the neck of femur (hip) and 4.1% higher BMD at the spine sites measured (Table 9). These differences are statistically significant as tested by analysis of covariance.

Example 7; Association of genotype at ADAM12 polymorphisms with knee osteoarthritis in men

DNA samples from of 296 male knee OA cases and 297 age, ethnicity and gender matched controls were obtained. These were English men of Caucasian race from the Nottinghamshire, London and Oxfordshire areas. Osteoarthritis was assessed both clinically and radiographically. Each of the cases had standardised extended anteroposterior radiographs of their knees with weight bearing obtained. Matched controls aged 50-80 have also been recruited without signs or symptoms of osteoarthritis

Two polymorphisms in the AD AMI 2 gene, ADAM_48 as described in example 2, and an intronic polymorphism rsl871054, were genotyped in both cases and controls.

An odds ratio was computed and a statistical test to compare the genotype frequencies between knee OA cases and controls was carried out. An increased risk due to the presence of the GG genotype was found, which is consistent the influence of this genotype on increased progression of OA in women. The odds ratio was 1.29 p<0.15. As illustrated in Table 10 the presence of genotype AA at the intronic SNP rsl871054 resulted in an increased risk of knee OA (odds ratio= 1.78 p<0.003) in men. These data indicate that polymorphisms at the ADAM12 can also predict increased risk of osteoarthritis in men.

Example 8: Association of genotype at ADAM12 haplotypes with knee osteoarthritis in both men and women

DNA samples from of 603 knee OA cases (298 men and 305 women) and 596 age- matched controls (299 women and 297 men) from Nottingham, Oxford and London were studied. All individuals were of Caucasian race. Osteoarthritis was assessed both clinically and radiographically. Each of the cases had standardised extended anteroposterior radiographs of their knees with weight bearing obtained. In addition 596 age and ethnicity matched controls aged 50-80 were recruited without signs or symptoms of osteoarthritis from the same areas in the UK.

Four polymorphisms in the ADAM12 gene, ADAM_48 as described in example 2, an intronic polymorphism rs 1871054, and two synsonymous coding SNPs mapping to codon positions 504 (rsl278279) and 825 (rsl044122) in the ADAM12 gene were genotyped in both cases and controls. Haplotype frequencies were estimated separately in cases and controls using the expectation maximisation algorithm (Excoffier L, Slatkin MW 1995 Maximum likelihood estimation of molecular haplotype frequencies in a diploid population. Mol Biol Evol 12;921-927, 1995) as implemented by the software Arlequin version 1.1

(http ://anthropolo ie.uni ge. ch/arlequin) . The resulting haplotype frequencies are shown in Table 11. For clarity the SNPs have been labeled as follows: ADAM_48= rsl044122, ADAM_int= rsl278279, ADAM_504= rsl871054, ADAM_825= rs3740199. The flanking sequences for each of these polymorphisms are freely available in the public domain from the NCBI web site: http ://ww w .ncbi ,n lm.nih. go v/SNP/index .html

The AD AM 12 haplotype frequencies were found to be significantly different between knee OA cases and controls (Table 11). In particular, chromosomes carrying a C at position 48, an A at the intronic SNP, an A and at position 504 and a T at position 825 resulted in a much increased risk of knee OA as reflected by the odds ratio of 4.55 (95% confidence interval 2.47-8.41). This haplotype was found at a frequencie of 5.2% among knee OA cases but only of 1.19% among healthy controls. These data indicate that combinations of polymorphisms at the ADAM12 can be used to predict increased risk of osteoarthritis in both men and women.

References:

Akesson K. (1999) in Dynamics of bone and cartilage metabolism, Seibel MJ, Robins SP, Bilezikian JP (eds), Academic Press S. Diego USA

Borenstein D. Curr Rheumatol Rep. 2004 Feb;6(l):14-9. Kawaguchi et al Spine. 1999 Dec l;24(23):2456-60. Lanes et al (1997) Arthritis & Rheumatism 40, 1475-1481.

Lawrence et al. Arthritis Rheum. 1998;41(5):778-799.

Petersson, IF (1996) Ann Rheum Dis 55, 659-664.

Paassilta et al JAMA. 2001 Apr ll;285(14):1843-9. Risch N, Am J Hum Genet 1990 Feb ;46 (2 : 229-41

Yelin E, & Callahan LF (1995). Arthritis Rheum. Oct;38(10): 1351-62.

Saito et al (1999) Biol Pharm Bull ;22(7):727-30

Hart et al (1999) Arthritis Rheum;42(l): 17-24

Zhang et al (1998).Arthritis Rheum. 41(10): 1867-73. Reddi, A. H. (1981) Collagen Rel. Res., 1:209-226;

Melton, L. J. III. (1995) Bone Miner. Res., Vol. 10, page 175.

Consensus Development Conference: Prophylaxis and Treatment of Osteoporosis:

Am. J. Med. 94:646-650, 1993;

Parfitt A M, Calcif Tissue Int 28: 1-5, 1979 Parfitt A M, Res Staff Physician, December:60-72, 1981

Parfitt A M in Osteoporosis: Etiology, Diagnosis, and Management (Riggs B L,

Melton L J, eds.). Raven Press, New York, 1988, pp 74-93

Garnero et al. J Clin Endocrinol Metab. 81(1): 140-46, 1996.

Johnston et al.: Pathogenesis of osteoporosis. Bone 17(2 Suppl): 19S-22S, 1995. Rosen, C J. Calcif Tissue Int 60(3):225-28, 1997.

Keen et al Drugs Aging (New Zealand) ll(5):333-7, 1997.

Jouanny et al. Arthritis Rheum 38(l):61-67, 1995.

Gong et al Am J Hum Genet 59(1):146-51, 1996.

Cummings Bone 18(3 Suppl): 165S-167S, 1996. Wasnich. Bone 18(3 Suppl): 179S-183S, 1996.

Wark Maturitas (Ireland) 23(2): 193-207, 1996.

Livshits et al. Hum Biol 68:540-54 1996.

Lonzer et al Clin Pediatr 35(4): 185-9, 1996.

Tian et al Mol Pathol. 2002 Dec;55(6):394-7. Abe et al Calcif Tissue Int. 1999 Jun;64(6):508-15.

Dong et al Blood. 2002 Apr 15;99(8):2637-46.

Ghosh et al.Proc Natl Acad Sci U S A. 2002 Apr 30;99(9):5842 Galliano et al J. Biol. Chem. 275: 13933-13939, 2000.

Gilpin et al J. Biol. Chem. 273: 157-166, 1998.

Kurisaki et al Molec. Cell. Biol. 23: 55-61, 2003.

Yagami-Hiromasa et al Nature 377: 652-656, 1995. Chen, J. D.& Evans, R. M. Nature 377: 454-457, 1995.

Fischle et al Molec. Cell 9: 45-57, 2002.

Horlein et al Nature 377: 397-403, 1995.

Hu, X & Lazar, M. A Nature 402: 93-96, 1999.

Jiang et al Cytogenet. Cell Genet. 92: 217-220, 2001. Ordentiich et al Proc. Nat. Acad. Sci. 96: 2639-2644, 1999.

Aitman et al Nature 405: 1015-1016, 2000.

Aitman et al Nature Genet. 21: 76-83, 1999.

Aitman et al Arterioscler. Thromb. Vase. Biol. 17: 748-754, 1997.

Armesilla, A. L & Vega, M. A. J. Biol. Chem. 269: 18985-18991, 1994. Cobum et al. J. Biol. Chem. 275: 32523-32529, 2000.

Curtis, B. R & Aster, R. H. Transfusion 36: 331-334, 1996.

Fernandez-Ruiz et al Genomics 17: 759-761, 1993.

Gotoda et al Nature Genet. 22: 226-228, 1999.

Greenwalt et al Blood 80: 1105-1115, 1992. Griffin et al Nature Med. 7: 840-846, 2001.

Groop et al J. Clin. Invest. 84: 205-213, 1989.

Ikeda et al Vox Sang. 57: 213-217, 1989.

Kashiwagi et al Thromb. Haemost. 69: 481-484, 1993.

Kashiwagi et al J. Clin. Invest. 95: 1040-1046, 1995. Kashiwagi et al, Hum. Genet. 108: 459-466, 2001.

Lee et al Transfusion 39: 873-879, 1999.

Omi et al Am. J. Hum. Genpt. 72: 364-374, 2003.

Pravenec et al Nature Genet. 27: 156-158, 2001.

Pravenec et al J. Clin. Invest. 103: 1651-1657, 1999. Reaven, G. M. Diabetes 37: 1595-1607, 1988.

Reaven et al New Eng. J. Med. 334: 374-381, 1996.

Savill et al J. Clin. Invest. 90: 1513-1522, 1992. Tanaka et al J. Molec. Cell. Cardiol. 29: 121-127, 1997. Tandon et al J. Biol. Chem. 264: 7576-7583, 1989. Tandon et al J. Biol. Chem. 264: 7570-7575, 1989. Tomiyama et al Blood 75: 684-687, 1990. Urwijitaroon et al Transfusion 35: 868-870, 1995. van Schravendijk et al Blood 80: 2105-2114, 1992. Yamamoto et al Blood 83: 392-397, 1994. Yamamoto et al Blood 76: 1698-1703, 1990. Yufu et al Am. J. Hemat. 33: 271-273, 1990.

Table 4

Table 5 SNP alias Genotype trait Test statistic Presence of Odds 95% CI Walds P< Osteophytes: Ratio Chi-sq ADAM .48 GG=0 26.7% 1.92 1.26, 2.91 9.35 0.002 GC+CC=1 40.3% CD36_5p AA=0 42.1% 0.76 0.60, 0.95 5.96 0.015 AOl 39.2% CC=2 32.1% NCOR_1699 GG+GA=0 28.8% 1.60 1.11, 2.31 6.47 0.011 AA=1 40.5%

Table 6 SNP alias Genotype trait Test statistic Presence of Joint odds 95% CI Walds P< Space Narrowing ratio Chi-sq NCOR_1699 GG+GA=0 17.3% 1.54 1.022.32 4.23 0.040 AA=1 24.9% THSD2_int3 CC+CT=0 25.4% 0.83 0.68 1.02 3.12 0.075 TT=1 18.9% Table 7 SNP alias Genotype trait Test statistic Change in KL Std. Err F-value (df) P< grade ADAM_48 GG 0.448 0.070 F(l,737)= GC+CC 0.590 0.035 3.29 0.070

Change in Std. Err P< Osteophyte grade ADAM_48 GG 0.229 0.060 F(l,737)= GC+CC 0.424 0.030 8.31 0.004 Change in Std. Err Osteophyte grade THSD2_int3 CC 0.501 0.066 F=(l,608) CT+TT 0.348 0.034 4.16 P<0.041

Table 8

Table 9

(a)

(b)

Table 10 (a)

(b)

Claims

1. A method for determining the risk of an individual developing osteoarthritis, comprising detecting the presence of osteoarthritis-associated ADAM12, THSD2, CD36 and NCOR2 alleles in a nucleic acid sample of the individual, wherein the presence of said alleles indicates the individual's risk for of developing osteoarthritis.

2. The method of claim 1, wherein the nucleic acid sample comprises DNA

3. The method of claim 1, wherein the nucleic acid sample comprises RNA

4. The method of claim 1, wherein the allele is detected by amplification.

5. The method of claim 4, wherein the allele or alleles are detected by polymerase chain reaction.

6. A kit for determining the risk of an individual developing osteoarthritis, comprising: (a) one or more sequence-specific oligonucleotides each individually comprising a sequence that is fully complementary to a sequence in and osteoarthritis-associated ADAM12, THSD2, CD36 and NCOR2 alleles wherein said sequence comprises one or more polymorphisms associated with said alleles (b) instructions to use the kit to determine the risk of an individual developing osteoarthritis.

7. The kit of claim 6 which contains additional sequencing primers.

8. A method for determining the risk of an individual developing osteoporosis, comprising detecting the presence of osteoporosis-associated ADAM12 alleles in a nucleic acid sample of the individual, wherein the presence of said alleles indicates the individual's risk for of developing osteoporosis.

9. A method for determining the risk of an individual developing spine osteoarthritis, comprising detecting the presence of spine osteoarthritis-associated THSD2, CD36 and NCOR2 alleles in a nucleic acid sample of the individual, wherein the presence of said alleles indicates the individual's risk for of developing spine osteoarthritis.