Introduction
Osteosarcoma (OS) is the most common primary malignant bone tumor in children and young adults with a peak incidence in adolescence and a second smaller peak after age 50, affecting approximately four people per million annually in the USA [
1,
2]. The metaphyseal regions of the long bones, including the distal femur, proximal tibia and proximal humerus, which are capable of rapid growth, are most likely to be affected [
3]. Primary OS arises from primitive mesenchymal cells producing osteoid and later extends to the pulmonary region forming secondary tumors in 10–20 % of patients [
4,
5]. Although neoadjuvant chemotherapy and surgical techniques have improved the 5-year survival of OS patients from 20–30 to 70 %, the level of OS treatment has not obtained a significant improvement in recent years and 30 % of newly diagnosed patients eventually died of lung metastasis [
6‐
9]. There have been extensive studies of OS genetics, biology, pathology and clinical aspects in recent years but there has been little progress in our understanding of its pathogenesis and therapeutic effect [
10]. Therefore, it is important to study OS-related risk factors and to identify the underlying mechanism in order to be of great benefit to the reduction of morbidity and mortality of the disease.
No predisposing factor is known to be related to OS. The combined effect of genetic and environmental factors has an important role in the pathogenesis of OS. Genome-wide linkage analysis has been extremely limited in studies of polygenic diseases such as OS; so, using single-nucleotide polymorphism (SNP) as a molecular genetic marker is expected to predict disease susceptibility to guide individual treatment. Earlier studies identified many putative genetic susceptibility variants for increasing OS risk in diverse biologically plausible pathways [
11‐
22], including growth and DNA repair; however, the statistical power of these studies has been limited by small sample sizes [
23]. Data regarding the role of common genetic variation in OS risk are also sparse.
Methods for the analysis of genome-wide association studies (GWAS) have been promoted greatly by the HapMap Project data [
24]. GWAS are an examination of genome-wide common genetic variants in different individuals to find disease or quantitative trait-related variants. Recently, GWAS in large samples of Caucasians recruited by Savage et al. [
25] identified two novel susceptibility loci for OS. In the combined samples, the two novel SNPs (rs1906953 and rs7591996) showed unequivocal evidence of association with OS, with a genome-wide significance threshold of
P = 5 × 10
−8. Moreover, the other SNPs (rs10208273 and rs17206779) nearly met the genome-wide significance level for association with OS (
P = 2.9 × 10
−7 and 5.1 × 10
−7). The above four SNPs are associated with OS in Caucasians, but whether they are associated with the incidence of OS in a Chinese Han population has not been estimated. We investigated the association between rs1906953, rs7591996, rs10208273 and rs17206779 polymorphisms and OS susceptibility in a Chinese Han population.
Discussion
Genetic and environmental factors both contribute to the cause of OS, and genetic background has the most important role. The finding of more and more susceptibility loci and genes related to OS in recent years will provide important insight into the diagnosis of the disease and aid the design of new preventive and therapeutic strategies for OS patients. Importantly, we confirmed the earlier finding that rs1906953 in the glutamate receptor metabotropic 4 (GRM4) gene is associated with OS in the Chinese Han population through the association analysis, which once more provides strong evidence for this locus.
We showed for the first time that there is an association between rs1906953, located in intron 7 of
GRM4 at 6p21.3, and OS in a Chinese Han population. Our results showed the allele T frequency of the OS case group (51.19 %) was greater compared to the control group (40.05 %) and those carrying the TT genotype had a higher risk of OS.
GRM4 is a family of G protein-coupled receptors that encode the group III metabotropic glutamate receptor 4 (mGluR4), and its role in intracellular signaling and inhibition of the cyclic AMP (cAMP)-signaling cascade has been reported [
26,
27].
The important role of glutamate in intercellular communication in the central nervous system (CNS) has been extended in recent years to non-neural systems [
28‐
30]. The glutamate-signaling system has been shown to be essential in bone remodeling and homeostasis [
29,
31,
32]. The disrupted normal glutamate signaling aggravates several bone diseases. Numerous cancers, including breast, prostate and lung, that can metastasize to bone have been shown to express several glutamate receptors and transporters [
33,
34]. Human OS cell lines expressing several glutamate receptors are known to be associated with proper tumor cell function because inhibited glutamate receptors lead to limited cell growth and pharmacologic prevention of glutamate release that results in blocked differentiation and increased levels of apoptosis [
35]. The important role of the cAMP pathway in OS has been demonstrated in mice, in which a cAMP-dependent protein kinase (Prkar1a) has been shown to suppress OS tumor growth [
36,
37]. Together, these results suggest
GRM4 is implicated in OS.
The reference SNP (rs1906953) is located in intron 7 of
GRM4, which is expressed abundantly in nerve tissue (including brain, spinal cord and retina) and relatively little in other tissues. To further test whether the OS-associated SNPs regulate the expression of
GRM4, we inspected three
cis-gene expression quantitative trait loci (
cis-eQTL) databases of European Caucasian populations [
38‐
40] and found rs1906953 was not correlated with the expression of
GRM4 or other genes. The index SNP, rs1906953, maps to a DNase I hypersensitive region based on the Encyclopedia of DNA Elements (ENCODE) data set, raising the distinct likelihood that the variant is within a region accessible to open chromatin and could manifest active regulatory elements [
41].
Several studies have suggested OS has a high incidence in young children and males of all ages and is often located in long tubular bones [
42]. The results of some studies suggested being taller than average and with a greater birth weight at diagnosis are associated with increased OS risk [
43‐
45]. Therefore, we analyzed the association of rs1906953 with some clinical parameters, including gender, age, location and distant metastasis. The genotype of rs1906953 is associated with the susceptibility of OS in our study, but not specifically to cases with different gender, age and location of the tumor. This indicates
GRM4 is involved in signaling pathways that stimulate the formation of OS independent of gender, age and location.
Distant metastasis association analysis showed the genotype of rs1906953_TT was correlated significantly with metastasis of OS. Survival analysis showed for rs1906953 that the median survival time of OS patients with the TT genotype was significantly shorter compared to the TC and CC genotypes. This result is consistent with several earlier findings that
GRM4 expressed in human OS cells is associated with its proper function and its expression in pediatric CNS tumors, rhabdomyosarcoma, multiple myeloma and colorectal cancer is predicted with poor prognosis [
35,
46‐
49]. Therefore, molecular biological studies are needed to confirm these two findings. On the basis of these results, we suggest
GRM4-related signaling is involved in the metastasis of OS and affects the survival of OS patients; however, more studies of the underlying molecular and biological mechanisms are needed to further confirm these results.
In summary, we showed the GRM4 gene polymorphism rs1906953 is associated with the incidence of OS and with the metastasis and survival of OS patients in a Chinese Han population. There are some limitations in our case–control study. Because all patient cases in this study were recruited from one hospital, it is difficult to avoid the inherent selection bias. In addition, because of the low incidence of OS, the number of OS patients in the study is small, which makes the statistical power of our results low. Therefore, studies with greater numbers of patients are needed to be confident that the GRM4 gene polymorphism rs1906953 is a genetic marker in predicting the incidence and prognosis of OS in a Chinese Han population.