Osteoarthritis (OA) is a common musculoskeletal disease among the elderly, characterized by the degradation of articular cartilage and formation of abnormal bone (osteophyte). It is a multifactorial disorder in which aging, genetic, hormonal and mechanical factors are all major contributors to its onset and progression [1]. Results of family-based and candidate gene studies have demonstrated a clear genetic component, particularly for early-onset OA [2–4]. The identification of candidate genes for OA susceptibility has mainly focused on genes encoding collagens (particularly for type II collagen), for other structural proteins of the extracellular cartilage matrix(ECM), the vitamin D and estrogen receptor genes, and for bone and cartilage growth factors [5]. And it is expected that some genes which regulate the formation, degradation, and repair of articular cartilage and subchondral bone metabolism may determine the occurrence of osteoarthritis. However, the specific underlying genetic factors and mechanisms in the development of osteoarthritis need to be further researched.

Calmodulin 1 gene (CALM1) encodes calmodulin (CaM), a ubiquitous eukaryotic Ca²⁺-binding protein which regulates many Ca²⁺-dependent cellular events by interacting with a heterogeneous group of target proteins [6]. Several studies have indicated that calmodulin probably plays a crucial role in articular cartilage on at least three aspects: regulating articular chondrogenesis [7], maintaining the cartilage phenotype in response to mechanical stimuli in mature chondrocytes [8, 9] and modulating adhesion of chondrocytes to extracellular matrix proteins during the cartilage repairing process [10]. Consequently, a deficiency of Ca²⁺-calmodulin signal may be involved in the pathogenesis of OA.

Through a genome-wide association study, Mototani et al [7] previously reported a functional single nucleotide polymorphism (SNP) (-16C/T rs12885713) in the core promoter region of the CALM1 gene associated with a markedly increased prevalence of hip OA in the Japanese population. Subsequent functional analyses revealed that the susceptibility -16T allele of this polymorphism decreased CALM1 transcription both in vitro and in vivo. In addition, higher levels of CALM1 expression in hip and knee OA cartilage than in normal were also observed [7].

Together, these findings implicated the candidacy of CALM1 as a susceptibility gene for OA. However, subsequent association studies [11, 12] failed to provide consistent results in OA patients of differently selected populations. To clarify its global relevance, the association needs to be confirmed by independent studies in different ethnic groups. The objective of this study is to assess the genetic association of the -16C/T polymorphism in CALM1 with knee OA in a Chinese Han population.

Through a large-scale association study with subsequent functional researches, the Japanese group have demonstrated a convincing association between the CALM1 core promoter polymorphism and Japanese hip OA. The well clarified role Ca²⁺-calmodulin signal played in chondrogenesis and normal cartilage phenotype maintaining, together with the results of the microarray analysis [7] showing higher levels of CALM1 expression both in hip and knee OA, have interested us in assessing whether this genetic polymorphism is also associated with OA in a Chinese Han population. However, we failed to find any positive result. Before the present study, this potential association had also been studied in UK Caucasian populations [11, 12], yielding no positive findings either in hip or knee OA patients. Together, these studies highlight the complexity and diversity of genetic contributions to such a multifactorial disease as OA.

The "T" allele and "TT" genotype frequencies of controls in present study were very similar to those in the original study of the Japan group but different to those in the UK study [12]. The "T" allele even seemed to be the major allele in UK Caucasians instead of the allele "C" which was dominant in both the present and the Japanese studies, indicating a heterogenous nature of this polymorphism between Asians and UK Caucasians. The different ethnic background and the potential linkage disequilibrium (LD) of the -16C/T locus with some other more relevant allele or alleles within the CALM1 gene that differ in different populations, may be one of the most likely explanations for the failure to replicate the association of -16C/T with OA susceptibility in UK Caucasians.

Recently, several population-based association studies for other OA susceptibility genes have shown consistent results in the Japanese and Chinese Han populations [13, 16]. Populations from neighboring regions may have greater similarities in genome for the fact that they share more recent common ancestors. Therefore, their allele frequencies are more highly correlated, a pattern that is commonly manifested as a cline of allele frequencies [17]. All these together had encouraged us to expect a positive association of -16C/T polymorphism in our population-based study, except for the fact that there is an obvious difference in clinical phenotype of OA patients between the present and the original Japanese studies -the different joint site affected, which may account for the inconsistent results of the two studies. It is becoming more apparent that the nature of OA genetic susceptibility is likely to vary between different joint sites [18]. In the RHOB and TXNDC3 studies, different associations were detected for the different affected joint sites [19–21]. A relatively small population size might have prevented from seeing the association. The proportion of genetic contribution of certain polymorphic locus to OA susceptibility may be influenced by other local environmental factors such as anatomical and biomechanical effects, and by some joint-specific genetic factors most of which were postulated to be involved in cell signalling and signal transduction [18, 22]. Interestingly, CALM1 is only one member of the CALM family by which mammalian CaM is encoded. The CALM family includes three nonallelic genes CALM1, CALM2 and CALM3 located on chromosomes 14q24-q31, 2p21.1-p21.3 and 19q13.2-q13.3, respectively, but all these three genes encode an identical 148 amino acid protein CaM [23]. However, the transcriptional activity and the proportion of each gene contributing to the local mRNA pools seem to differ during cell differentiation processes or in different biochemical microenvironments [24, 25]. Thus, the other two CALM genes that are more functionally active in the knee joint might have decreased the extent to which the CALM1 OA risk is manifest. To make our study more complete and comparable, it is necessary to collect enough samples of hip OA despite of the very much lower prevalence [26], and assess the association between this polymorphism and hip OA in the same Chinese population.

Moreover, Hoaglund [27] has recently made a comment on the ascertainment criteria used for the patient enrollment in the Japanese association study and has suggested that the CALM1 core promoter polymorphism may be associated with Japanese congenital hip disease causing secondary hip OA rather than primary hip OA. Colleagues in our department are now trying to identify this hypothesis by assessing the association in patients of developmental dysplasia of the hip (DDH) which may represent a severe phenotype of congenital hip disease.

The present study did not find any significant association between the CALM1 core promoter polymorphism (-16C/T) and susceptibility to knee OA in the Chinese Han population. Further studies are necessary to identify the association of this polymorphism with hip OA in our population and with OA in other different populations on different joint sites to give a global view of this polymorphism in pathogenesis of OA.