Support of positive association in family-based genetic analysis between COL27A1 and Tourette syndrome

Recently, a genome-wide association study has indicated associations between single nucleotide polymorphisms in the Collagen Type XXVII Alpha 1 gene (COL27A1) and Tourette syndrome in several ethnic populations. To clarify the global relevance of the previously identified SNPs in the development of Tourette syndrome, the associations between polymorphisms in COL27A1and Tourette syndrome were assessed in Chinese trios. PCR-directed sequencing was used to evaluate the genetic contributions of three SNPs in COL27A1(rs4979356, rs4979357 and rs7868992) using haplotype relative risk (HRR) and transmission disequilibrium tests (TDT) with a total of 260 Tourette syndrome trios. The family-based association was significant between Tourette syndrome and rs4979356 (TDT: χ2 = 4.804, P = 0.033; HRR = 1.75, P = 0.002; HHRR = 1.32, P = 0.027), and transmission disequilibrium was suspected for rs4979357 (TDT: χ2 = 3.969, P = 0.053; HRR = 1.84, P = 0.001; HHRR = 1.29, P = 0.044). No statistically significant allele transfer was found for rs7868992 (TDT: χ2 = 2.177, P = 0.158). Although the TDT results did not remain significant after applying the conservative Bonferroni correction (p = 0.005), the significant positive HRR analysis confirmed the possibility of showing transmission disequilibrium, which provides evidence for an involvement of COL27A1in the development of TS. However, these results need to be verified with larger datasets from different populations.

for the existence of a genetic background associated with a multifactorial mode of inheritance 5,6 , and numerous studies aiming to explore genetic susceptibility of TS have been published [7][8][9][10] . However, no causative candidate genes have been identified and the genetics of TS is complex and remains unclear.
In recent years, the genome-wide association study (GWAS) approach has led to the identification of many genetic associations for common complex traits 11 . This model-free approach to gene discovery has led to a greater pathophysiological understanding of many disorders and can improve pharmaco-therapeutic strategies 12 . Recently, Scharf and colleagues performed the first GWAS of TS in 1285 cases and 4964 ancestry-matched controls of European ancestry, including two European-derived population isolates, Ashkenazi Jews from North America and Israel and French Canadians from Quebec, Canada. They found the strongest association was with rs7868992, which is within the Collagen Type XXVII Alpha 1 gene (COL27A1), (P = 1.85 × 10 −6 ) in European-derived samples 13 .
COL27A1 maps to chromosome 9q32-33, is approximately 156 kb long and has 61 exons 14 , which encode a long triple helical domain, a carboxyl-terminal propeptide (C-propeptide), and a large globular amino-terminal propeptide (N-propeptide). COL27A1 is strongly expressed in developing cartilage and weakly expressed in many other tissue types such as skin, stomach, gonad and brain 15 . Several studies have unexpectedly identified novel roles of collagens and collagen-like molecules in the developing vertebrate nervous system 16 . However, the function of COL27A1 in neural development, and specifically in neural circuitry is poorly understood.
To clarify the global relevance of the SNPs that have been associated with TS, the associations need to be confirmed by independent studies in different ethnic groups. The possible associations between SNPs and TS development should also be validated. The objective of this study was to assess the genetic association of three SNPs in COL27A1 (rs4979356, rs4979357 and rs7868992) with TS in a Chinese Han population.

Results
The results showed that the allele and genotype frequencies of rs4979356, rs4979357 and rs7868992 in the parents group were not significantly different from those expected according to Hardy-Weinberg equilibrium (for rs4979356, χ2 = 0.19 and p = 0.66; for rs4979357, χ 2 = 0.24 and p = 0.62; for rs7868992, χ 2 = 1.05 and p = 0.30). For all three polymorphisms, p > 0.05, suggesting that the population was genetically balanced and that the samples were from the same Mendelian population.

Discussion
Recent advances in decreasing the cost of SNP genotyping and rigorous statistical methodology for analyzing large numbers of samples have made GWAS a feasible method for the genetic study of complex disorders 17 . Scharf and colleagues reported the first GWAS of TS 13 . In a primary meta-analysis of GWAS data from European ancestry samples, the top association was found for rs7868992 within COL27A1 (P = 1.85 × 10 −6 ). A secondary analysis including an additional 211 cases and 285 controls also identified rs7868992 to have the highest association (P = 3.6 × 10 7 ). Plumb reported that COL27A1 plays an important structural role in the pericellular extracellular matrix of the growth plate and that homozygotes for an 87 amino acid deletion exhibit severe chondrodysplasia 18 . To date few studies have reported on the correlation between COL27A1 and TS (in addition to that of Scharf et al. 13 ), and no strong evidence can support the view that COL27A1 is a candidate for TS. Nevertheless, recent studies have suggested the importance of collagens in directing neurite extension and connection in the nervous system 19 . COL27A1 morphant zebrafish show dysmorphic vertebrae lacking hemal and neural spines and impaired notochord development 20 .
In this study, we evaluated the genetic contributions of rs4979356, rs4979357 and rs7868992 polymorphisms in COL27A1 using a TDT-HRR design with a total of 260 TS trios. Although the traditional case/ control study design has many advantages, results from these studies must be interpreted with caution. The most serious concern is that the results can be explained by spurious associations due to mismatches in ethnicity or geographical regions between the control and patient groups, such as population stratification. Moreover, it seems not to be an issue if the call rate of each SNP is 100% and absence of Hardy Weinberg disequilibrium. However, TDT-HRR analysis is a family-based test, and it can be used to find an association due to linkage with an etiological mutation rather than due to spurious associations. By examining the frequency with which a marker allele is transmitted from parents to affected offspring, we can look for an allele to be transmitted more often than by chance. Such family-based approaches clearly complement the case/control design when doing research on association in multiple studies (especially in different populations, ethnic groups and/or different laboratories).
The TDT test showed that rs4979356 had significant transmission disequilibrium, which indicated that rs4979356 was associated with the etiology of TS development. Similar positive results were found in the subsequent HRR and HHRR analysis for rs4979356 suggesting that allele G and genotype G(+ ) are risk factors for TS. The results also showed that there might be transmission disequilibrium of rs4979357 in TS because the P value of TDT was approximately equal to the test level of significance, 0.05. However, the TDT test results did not remain significant after applying the conservative Bonferroni correction for multiple testing. The reason may be that the TDT and Bonferroni test are conservative themselves to generate significant results. Moreover, the significant positive HRR analysis of rs4975396 confirmed the possibility of showing transmission disequilibrium which provides evidence for an involvement of COL27A1 in the development of TS. Similarly, the allele C and genotype C(+ ) of rs4979357 were also risk factors for TS. According to the primary meta-analysis of GWAS data, the highest association was found for rs7868992. It was disappointing that the TDT results for rs7868992 did not revealed statistical significance of allele transfer in any cases. It's confirmed by the results of linkage disequilibrium test that the 3 SNPs were inherited independently, and thus, differences in genetic ancestry between Chinese and European may account for the insignificant TDT result of rs7868992.
To the best of our knowledge, this is the first report on association between COL27A1 and the development of TS in a Han Chinese population, in which a new COL27A1 variant, rs4979356, shows association with TS. Our study lays the groundwork for the eventual identification of further TS candidate variants in larger cohorts. However, this family-based study performed on 260 trios with TS did not demonstrate the influence of rs7868992 on the incidence of TS, as reported by Scharf et al. in 2012 13 . All the trios were genotyped by a direct sequencing method which has a relatively high sensitivity and accuracy rate, so the reliability of the data can be effectively guaranteed. A minimum number of samples rs4979356(C > G) rs4979357(T > C) rs7868992(G > A) Transmitted allele  227  293  232  288  187  333   Non-transmitted allele  192  328  200  320  164  356 Results (  is required to ensure the testing power of TDT analysis; hence, it would be appropriate to perform further analysis with a larger sample size to quantify the link between rs7868992 polymorphism and TS. This study and others have only demonstrated associations between polymorphisms in COL27A1 and TS at the level of statistics and we look forward to results from functional analyses that will, or will not, verify the associations. These findings give hope for possible TS treatment, based on COL27A1 as a therapeutic target. Clarification of the genetic mechanism of TS will not only help us understand the etiology of TS and comorbid conditions, such as OCD and ADHD, but will also shed light on neural development and growth of the central nervous system.

Materials and Methods
Study population. Study subjects between 5 and 18 years of age with a diagnosis of TS, together with their parents, were recruited from the Affiliated Hospital of Qingdao University and Linyi People's hospital, China. The TS cases comprised 35 female and 225 male outpatients. All probands were diagnosed independently by two experienced psychiatrists according to the DSM-IV criteria and the TS Classification Study Group. They were assessed by means of neurological examination and mental status examination. Subjects were excluded if they presented with unclear diagnosis or incomplete medical record data. The study protocol was approved by the Human Ethics Committee of the Affiliated Hospital of Qingdao University and the National Research Institute for Family Planning and is compliant with the 1975 Declaration of Helsinki. Informed written consent was obtained from every participant or their legal guardians after a complete and extensive explanation of the study. All experiments were performed in accordance with the approved guidelines.
Laboratory methods. Blood samples were collected from TS patients and their parents, and stored at − 20 °C until analyzed. Genomic DNA was extracted from peripheral blood leukocytes according to standard methods. The polymorphic alleles of COL27A1 were identified by polymerase chain reaction (PCR). The DNA fragment containing the three polymorphisms (rs4979356, rs4979357, rs7868992, which are located close to each other) was amplified using the following primers: 5′ -AGACAGGCTGCCTAGTGT-3′ and 5′ -GATAGCGTCATTGAACTCC-3′ . PCR reactions were carried out in a final volume of 20 μ L, containing 2× PCR MasterMix, 0.4 mol/L of each primer and 100 ng of genomic DNA. The reaction was carried out as follows: 94 °C for 5 min, followed by 35 cycles of 95 °C for 30 s, 56 °C for 30 s, 72 °C for 1 min and 72 °C for 10 min. Amplified PCR products were purified and sequenced using the appropriate PCR primers and the BigDye Terminator Cycle Sequencing kit (Applied Biosystems, Foster City, CA, USA) and run on an automated sequencer, a ABI 3730XL (Applied Biosystems), to determine genotype. Statistical analysis. All data analyses were carried out using the Statistical Package for the Social Sciences (Windows Version 12.0; SPSS, Chicago, IL, USA). The degree of LD between the adjacent SNPs were estimated using HAPLOVIEW 4.0 (available at http://www.broadinstitute.org/haploview), in which the transmitted and untransmitted group served as case and control group respectively. For trio data from the 260 TS cases, a family-based study was performed to assess genetic association by means of haplotype relative risk (HRR) and transmission disequilibrium tests (TDT) statistics. To increase efficiency of the test, we effectively enlarged the number of cases by haplotype-based haplotype relative risk (HHRR) analysis. In order to deal with the multiple testing problem (TDT-HRR test), we performed a Bonferroni correction test. The conventional P value of 0.05 was divided by the total number of tests performed in the present study (10 tests) accounting for a new P value threshold of 0.005. The 10 tests were calculated as follows: Some studies are done with experiment-wide significance, and here 3 SNPs were examined with three tests each (albeit correlated tests). Thus formally, a Bonferroni correction of 0.05/10 = 0.005 was set for experiment-wide significance, and still a chance observation cannot be totally ruled out.