The Involvement of ADAMTS-5 Genetic Polymorphisms in Predisposition and Diffusion Tensor Imaging Alterations of Lumbar Disc Degeneration Nan Wu,1 Jun Chen,1 Hao Liu,2 Luo Zhao,1 Sen Liu,1 Jiaqi Liu,1 Xinlin Su,1 Wenliang Wu,3 Jin Cong,4 Guixing Qiu,1 Zhihong Wu1 1 Department of Orthopedic Surgery, Peking Union Medical College Hospital, Peking Union Medical College and Chinese Academy of Medical Sciences, No.1 Shuaifuyuan, Beijing 100730, P.R. China, 2Biology and Biomedical Sciences, Division of Medical Sciences, Harvard Medical School, Boston, Massachusetts 02115, 3Department of Orthopedic Surgery, Shandong University Qilu Hospital, Shandong Province 250012, P.R. China, 4Department of Internal Medicine, Navy Yantai Hospital, Shandong Province 264001, P.R. China
Received 19 October 2013; accepted 17 December 2013 Published online 10 January 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jor.22582
ABSTRACT: Purpose: Low back pain is a global health problem in which more than 40% is caused by lumbar intervertebral disc degeneration (LDD). ADAMTS-5 (A disintegrin and metalloproteinase with thrombospondin motifs-5) was shown to be involved in LDD by functional analyses. To identify whether there is an association between ADAMTS-5 and LDD, and what is the contribution of ADAMTS-5 genetic polymorphisms to MD (Mean diffusivity) changes in lumbar IVD (Intervertebral disc). We firstly genotyped selected ADAMTS-5 SNPs (Single nucleotide polymorphisms) in a Chinese Han population. After the primary analyses of allelic, genotypic, and haplotypic association, we performed SNP–SNP interaction analysis. We subsequently genotyped another 50 participants and acquired the corresponding MD values from individual lumbar IVDs. The association analysis between the genotypic groups divided by the above positive SNPs and the corresponding MD values were also performed. Significant associations were identified in rs151058, rs229052, and rs162502. None of the 2-SNP haplotypic analysis survived the 10,000 permutation test. The following interaction analysis demonstrated that rs151058 was strong associated with LDD when conditioning on rs162502. Significant difference of MD values between AA and Gþ carriers was identified in rs162502. This is the first study indicating that the SNPs of ADAMTS-5 may contribute to predisposition of LDD. An interaction between rs151058 and rs229052 may exist in ADAMTS-5 with LDD. The rs162502 might be associated with altered MD values. ß 2014 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 32:686–694, 2014. Keywords: lumbar disc degeneration (LDD); a disintegrin and metalloproteinase with thrombospondin motifs-5 (ADAMTS-5); single nucleotide polymorphism (SNP); diffusion tensor imaging (DTI)
As a global health problem, low back pain (LBP) was reported to affect about 80% of population during a lifetime.1 LBP, defined as pain radiating from the back into the dermatome of the affected nerve trunk, is a common musculoskeletal disorder characterized by symptomatic lumbar disc herniation with or without sciatica.2 More than 40% of LBP is caused by lumbar intervertebral disc degeneration (LDD).3 It was reported that about 5% of the Finnish population suffered from LDD, leading to disability in the workingage population.4 Though several environmental and anthropometric risk factors have been showed to be related to LDD, the roles of genetic factors can never be exaggerated.5–10 The extracellular matrix (ECM), composed of collagens and proteoglycans, is crucial to normal intervertebral disc (IVD) functions.11 Previous studies indicated that the synthesis and degradation balance of ECM was interrupted by diverse matrix proteases during LDD.12–16 As a large family of metalloproteases, ADAMTS-5 seems to be the most active Conflicts of interest: none. Nan Wu, Jun Chen, and Hao Liu contributed equally to this study. Grant sponsor: National Natural Science Foundation of China; Grant numbers: 81071513, 81130034, 81271942; Grant sponsor: Public Service of the Health Ministry of China; Grant number: 201002018. Correspondence to: Zhihong Wu (T: þ8601069156081; F: þ8601069156081; E-mail: [email protected]) # 2014 Orthopaedic Research Society. Published by Wiley Periodicals, Inc.
686
JOURNAL OF ORTHOPAEDIC RESEARCH MAY 2014
one,15,17,18 and has been paid increasing attentions due to their abilities in the cartilage proteoglycan cleavage.19,20 Besides mRNA and protein of ADAMTS5 were detected in IVDs,21 increased levels in patients with chronic LBP and intervertebral disc herniation during LDD can be found.22–24 Although several functional analyses revealed the involvement of ADAMTS-5 in LDD22–25 and even a pedigree-based research was also conducted,26 no positive conclusion has been achieved in the relationship between genetic alterations of ADAMTS-5 and LDD. As a newly developed magnetic resonance imaging (MRI) technique, diffusion tensor imaging (DTI) can non-invasively provide the microstructural properties of soft tissue. By adding a diffusion-gradients signal to the standard MRI sequence, DTI can deliver a more detailed imaging for random water molecule motion than normal MRI. Numerous literatures have shown that changes in DTI parameters of lumbar IVDs are related to the degree of LDD.27,28 Several methods have been used to measure DTI, among which, mean diffusivity (MD) is one of the most effective parameters for the IVD degenerative status evaluation. MD can reflect the average diffusion rate in a tissue and represent changes in the water content of intervertebral disc.29 However, up to our knowledge, literatures combining genetic association study with DTI in the field of orthopedics can rarely be found till now. The present study was designed to answer the following questions: (1) Is there an association between ADAMTS-5 gene and LDD in Chinese Northern Han
LUMBAR DISC DEGENERATION
population? (2) What is the contribution of ADAMTS-5 genetic polymorphisms to DTI changes in lumbar IVDs.
MATERIALS AND METHODS Subjects A total of 489 unrelated sporadic cases with LDD (240 males and 249 females, aged 42.6 14.2 years) and 558 control subjects (271 males and 287 females, aged 41.4 11.1 years) from Chinese Northern Han population were recruited with no differences of age and gender between the two groups. All the patients and controls were diagnosed through lumbar disc MRI scanning by at least two experienced radiologists. Primary exclusion criteria included spinal and joint diseases such as trauma, spinal tumor, inflammatory, scoliosis, and osteoarthritis. Individuals with known environmental risk factors, including heavy manual labor, occupational driving, or heavy smoking, were also excluded. Written informed consent were obtained from all participants. The Ethics Committee of Peking Union Medical College Hospital, Peking Union Medical College and Chinese Academy of Medical Sciences approved this study. Genotyping of SNPs Twenty SNPs ranging from 5 kb upstream and downstream of ADAMTS-5 were selected from the NCBI SNP database (www.ncbi.nlm.nih.gov/SNP) and the HapMap database (www.hapmap.org). The SNPs included were as follows: rs229070 (SNP 1), rs151065 (SNP 2), rs229079 (SNP 3), rs226794 (SNP 4), rs2830585 (SNP 5), rs2830586 (SNP 6), rs162499 (SNP 7), rs162495 (SNP 8), rs162489 (SNP 9), rs151058 (SNP10), rs229052 (SNP 11), rs229054 (SNP 12), rs9984329 (SNP 13), rs2249350 (SNP 14), rs233896 (SNP 15), rs162509 (SNP 16), rs162506 (SNP 17), rs162502 (SNP 18), rs2132824 (SNP 19) and rs1974415 (SNP 20) (shown in Supplementary Table S1). Among them, SNP 1 and SNP 2 are in the 30 UTR region, SNP 5 in the exon region, SNP 19 and SNP 20 in the 50 near region, and others are in the introns. All the SNPs selected have a minor allele frequency (MAF) of more than 0.05. Genomic DNA was extracted from peripheral blood leukocytes using a standard phenol-chloroform method. The SNP genotyping was performed on the Sequenom MassARRAY SNP genotyping platform (Sequenom, San Diego, CA). Five percentage duplicates of randomly chosen samples were used for quality controls. MRI Data Acquisition Another 50 participants (26 female, 24 male; mean age ¼ 45.5 9.5 years; age range ¼ 26–61 years) who met the following criteria were enrolled: (1) No history of LBP or other spinal diseases (trauma, infection, inflammation and deformity, etc.); (2) no contraindications for MRI scanning; (3) no systemic disorders, such as diabetes, hypertension, hyperlipidemia, etc. Written informed consent were obtained from all participants. All subjects underwent MRI scans on a clinical 3.0-T scanner (Signa EXCITE, GE Medical Systems, Milwaukee, WI) from January to April 2013. Sagittal T2 images of five IVDs (from L1-2 to L5-S1) were acquired through T2weighted spin echo sequences. The sagittal DTI images were then acquired with a standard single-shot diffusion-weighted echo-planar imaging (DW-EPI) sequence (b ¼ 800 s/mm2) in 21 noncollinear diffusion-encoding directions. Parameters for acquisition were as follows: TR/TE ¼ 5,000/85.5 ms,
687
thickness/gap ¼ 4/0 mm, matrix ¼ 128 128, FOV ¼ 32 32 mm2, 14 slices which lasted about 6 min. Images Processing FuncTool Performance software (GE Medical Systems) was used to calculate the MD maps according to the standard formulas described by Basser et al.30. For evaluation of MD values, a region of interest (ROI) analysis was then performed. The elliptical ROIs (90–110 mm2) were manually placed in the center of the highest signal area at each level (b ¼ 0 s/mm2). They were then projected to the MD maps, and the ROI mean signal intensities were thus calculated. The individual MD value for one participant was the mean of all five IVD MD intensities. A senior spine radiologist (with 11 years of experience) and a spine surgeon (with 6 years of experience) drew the ROIs independently, and the mean values were then acquired for the following analysis. Data Analysis
Association analysis Hardy–Weinberg equilibrium (HWE), haplotype blocks, and the linkage disequilibrium (LD) patterns were estimated by using the Haploview program (version 4.2, Broad Institute of MIT and Harvard, Cambridge, MA).31 Primary analyses were performed for allelic, genotypic and haplotypic association based on the UNPHASED software (v.3.1.5 Dudbridge F, MRC Biostatistics Unit, Cambridge, UK). In our study, the sliding window size for haplotype analysis was two. To test an underlying interaction among selected SNPs within the ADAMTS-5 gene, SNP–SNP interaction analyses were then performed. Ten thousand times of permutation test, a built-in program of UNPHASED, was conducted to obtain an adjusted p-value. QUANTO software (http://hydra.usc.edu/ gxe) was used to calculate the power value, in which the prevalence of LDD was set at 32%, the false positive rate was 0.05 with a minor allele frequency of 0.10. Association of Positive SNPs and MD Values Fifty participants were further grouped according to the individual genotypes of SNPs associated with LDD in the above analysis. The Student’s t-test was then used to detect the differences of total MD values between paired groups. Differences were considered statistically significant when p < 0.05. Statistical analysis was performed by using SPSS 17 (SPSS, Inc., Chicago, IL).
RESULTS The genotypic distributions of the 20 selected SNPs did not deviate from Hardy–Weinberg equilibrium in the control subjects (Supplementary Table S2). Six LD blocks were observed according to Confidence intervetals’ model31 (D0 > 0.9 and r2 > 0.8; Fig. 1). Power calculation showed that our sample had 80.96% power to detect allelic association at a false rate of 0.05. The analysis of allele frequency indicated the association signals in rs151058 (SNP 10, x2 ¼ 9.069, df ¼ 1, unadjusted p ¼ 0.003), rs229052 (SNP 11, x2 ¼ 4.445, df ¼ 1, unadjusted p ¼ 0.035), and rs162502 (SNP 18, x2 ¼ 7.393, df ¼ 1, unadjusted p ¼ 0.007), which survived the 10,000 permutation correction with a global p value of 0.042 (Table 1). The genotypic analysis showed associations of SNP 10 (x2 ¼ 9.312, df ¼ 2, p ¼ 0.010) and SNP 18 (x2 ¼ 7.179, df ¼ 2, JOURNAL OF ORTHOPAEDIC RESEARCH MAY 2014
688
WU ET AL.
Figure 1. LD structures of the twenty SNPs genotyped in ADAMTS-5 gene. The LD strengths between paired SNPs are shown in color according to confidence intervals’ model.
p ¼ 0.028) with LDD, however, neither of them survived the 10000 permutation correction (global p-value ¼ 0.131; Supplementary Table S3). All haplotypes used for analysis had a frequency of >1% in either patients or the control groups. The 2-SNP haplotypic analysis showed significant associations for SNP 9-SNP 10 (x2 ¼ 9.240, df ¼ 3, p ¼ 0.010), SNP 10-SNP 11 (x2 ¼ 9.971, df ¼ 3, p ¼ 0.019), SNP 11-SNP 12 (x2 ¼ 8.286, df ¼ 3, p ¼ 0.040), SNP 17-SNP 18 (x2 ¼ 8.532, df ¼ 3, p ¼ 0.036) and SNP 18-SNP 19 (x2 ¼ 8.930, df ¼ 3, p ¼ 0.030). The global p value after 10,000 permutation test was 0.0941 (Table 2). Three positive SNPs—rs151058 (SNP 10), rs229052 (SNP 11), and rs162502 (SNP 18) were selected to perform the following SNP-SNP interaction analyses. The results demonstrated a combined effects by confirming disease association for the SNP 10—SNP 18 combination (x2 ¼ 15.462, df ¼ 3, unadjusted p ¼ 0.001, adjusted p ¼ 0.003) and SNP 11—SNP18 combination (x2 ¼ 12.593, df ¼ 3, unadjusted p ¼ 0.006, adjusted p ¼ 0.007; Supplementary Table S4). Analyses of the cis-phase interaction demonstrated that of the two SNPs tested, SNP 10 showed strong association with LDD when conditioning on SNP 18 (x2 ¼ 9.233, df ¼ 2, unadjusted p ¼ 0.010, adjusted p ¼ 0.019), while no such interaction were observed for SNP 11 when conditioning on SNP 18 (x2 ¼ 4.940, df ¼ 2, unadjusted p ¼ 0.062, adjusted p ¼ 0.085) (Table 3). JOURNAL OF ORTHOPAEDIC RESEARCH MAY 2014
According to the positive SNPs confirmed in the above analyses, the 50 participants were defined as GG carriers (37 participants) and Aþ carriers(AG þ AA, 12 participants) for SNP10, AA carriers (28 participants), and Gþ carriers(AG þ GG, 22 participants) for SNP11, and AA carriers (17 participants) and Gþ carriers(AG þ GG, 31 participants) for SNP18. The mean MD values of the five IVDs from 50 subjects were 833.19 63.30 102 m2/ms (shown in mean SD). A representative T2 image and the corresponding MD maps in one participant were shown in Figure 2. The MD values from L1/2 to L5/S1 IVDs were 181, 191, 191, 182, and 159 102 m2/ms, respectively. Significant differences of MD values between two genotypic groups were identified for SNP18 (t ¼ 2.138, df ¼ 46, p ¼ 0.038). The mean MD values for AA and Gþ carriers were 807.23 54.34 102 m2/ms and 847.30 65.88 102 m2/ms, respectively (shown in mean SD), and the scatter plot was shown in Figure 3. No such significance was observed for SNP 10 (t ¼ 0.22, df ¼ 48, p ¼ 0.829) and SNP 11 (t ¼ 0.26, df ¼ 48, p ¼ 0.854; Table 4).
DISCUSSION For the first time, our study revealed the genetic polymorphisms of ADAMTS-5 may be associated with susceptibility to LDD in Chinese Northern Han population and contribute to DTI changes in lumbar IVDs.
LUMBAR DISC DEGENERATION
689
Table 1. Allelic Associations of ADAMTS-5 Tested SNPs With LDD LDD SNP_ID SNP 1 SNP 2 SNP 3 SNP 4 SNP 5 SNP 6 SNP 7 SNP 8 SNP 9 SNP 10a SNP 11a SNP 12 SNP 13 SNP 14 SNP 15 SNP 16 SNP 17 SNP 18a SNP 19 SNP 20
CTR
Allele
N
%
N
%
OR
95% CI
x2
p-value
C G A G A G A G C T G T A G C T G T A G A G A G C T A C A C C G A G A G A G A C
752 158 486 426 717 195 492 438 846 66 67 843 690 220 587 325 260 670 84 846 724 186 167 763 548 382 337 571 281 77 385 525 847 63 508 404 254 676 579 351
82.6 17.4 53.3 46.7 78.6 21.4 52.9 47.1 92.8 7.2 7.4 92.6 75.8 24.2 64.4 35.6 28.0 72.0 9.0 91.0 79.6 20.4 18.0 82.0 58.9 41.1 37.1 62.9 78.5 21.5 42.3 57.7 93.1 6.9 55.7 44.3 27.3 72.7 62.3 37.7
848 166 546 468 807 203 503 459 917 97 97 917 775 241 674 342 262 700 129 833 769 249 183 779 548 414 367 645 603 215 411 603 948 72 503 513 269 693 591 373
83.6 16.4 53.8 46.2 80.0 20.1 52.3 47.7 90.4 9.6 9.6 90.4 76.3 23.7 66.3 33.7 27.2 72.8 13.4 86.6 75.5 24.5 19.0 81.0 57.0 43.0 36.3 63.8 73.7 26.3 40.5 59.5 92.9 7.1 49.5 50.5 28.0 72.0 61.3 38.7
1 1.07 1 1.02 1 1.08 1 0.99 1 0.74 1 1.33 1 1.03 1 1.09 1 0.97 1 1.56 1 0.79 1 1.07 1 0.92 1 0.96 1 0.77 1 0.93 1 0.98 1 0.78 1 1.03 1 0.96
1 0.85–1.36 1 0.86–1.22 1 0.87–1.35 1 0.81–1.17 1 0.53–1.02 1 0.96–1.84 1 0.83–1.264 1 0.90–1.32 1 0.79–1.18 1 1.17–2.09 1 0.64–0.98 1 0.85–1.35 1 0.77–1.11 1 0.80–1.16 1 0.57–1.03 1 0.78–1.11 1 0.69–1.39 1 0.65–0.93 1 0.84–1.26 1 0.80–1.16
0.337
0.5616
0.060
0.8068
0.480
0.4884
0.072
0.7884
3.363
0.0667
2.986
0.0840
0.055
0.8151
0.828
0.3628
0.123
0.7254
9.069
0.0026
4.445
0.0350
0.356
0.5506
0.745
0.3879
0.149
0.6997
3.042
0.0811
0.623
0.4299
0.014
0.9071
7.393
0.0066
0.100
0.7517
0.181
0.6703
LDD, lumbar disc degeneration; CTR, control; OR, odds ratio; N, number; %, percent; 95% CI, 95% confidence interval. The adjusted pvalue between the LDD and control groups for allelic association obtained from 10,000 permutations was 0.0419. aIndicates p < 0.05 before adjusted.
This may give an explanation to the phenomenon reported by Virtanen et al.26 and lead to a new way for better understanding of LDD. The ADAMTS-5 Association With LDD Several studies have suggested that polymorphism of genes encoding ECM degradation enzymes were associated with LDD. For example, Takahashi et al.32 found that the 5A5A and 5A6A genotypes of matrix metalloproteinase (MMP)-3 gene showed a significant association with degenerative IVDs in
elderly Japanese population. Sun et al.33 reported 1562C/T polymorphism of MMP-9 gene was associated with a high risk of LDD in young adult population in Northern Chinese population. The involvement of ADAMTS-5 gene in LDD has been proved in vivo and in vitro. Patel and his colleagues revealed abundant levels of ADAMTS-5 in human cadaveric IVDs.22 Le Maitre and coworkers25 have reported the relationship between expression of ADAMTS-5 and altered phenotype of IVD cells during degeneration, which was also proved by Zhao et al.24 through immunohistochemistry JOURNAL OF ORTHOPAEDIC RESEARCH MAY 2014
690
WU ET AL.
Table 2. Association Analyses of 2-SNP Haplotypes in ADAMTS-5 With LDD Combination
Haplotype
LDDa
CTRa
OR
Individual p-Value
Global p-Value
SNP1-SNP2
C-A C-G G-A G-G A-A A-G G-A G-G A-A A-G G-A G-G A-C A-T G-C G-T C-G C-T T-G T-T G-A T-A T-G A-C A-T G-C G-T C-G C-T T-G T-T G-G T-A T-G A-A A-G G-A G-G A-A A-G G-A G-G A-C A-T G-C G-T C-A C-C T-A T-C A-A A-C C-A C-C A-C A-G C-C C-G
327.5 424.5 156.5 1.5 486 0 231 195 448.2 237.8 7.8 180.2 454.4 1.6 356.6 61.4 1 843 66 0 67 623 220 582.7 107.3 2.3 217.7 53.24 510.8 193.8 116.2 260 84 586 1.1 79.9 692.9 98.1 2.1 691.9 152.9 25.1 0 167 548 215 320.8 187.2 1.2 360.8 135 0 123 74 56.4 203.6 71.7 2.3
480 466 166 0 542.3 2.7 264.7 200.3 461.2 242.8 2.8 177.2 462.3 1.7 336.7 85.3 1.0 912 96 1.0 97 674 241 668.8 102.2 2.2 238.8 44.0 540 202 104 262 129 571 0 121 672 97 1.1 670.9 170.9 47.1 1.3 181.7 546.7 232.3 316.5 185.5 2.5 381.6 278.5 1.5 280.5 199.5 96.1 464.9 199.9 1.1
1 1.06 1.09 1.90eþ008 1 0 0.97 1.09 1 1.01 2.87 1.05 1 0.96 1.08 0.73 1 0.93 0.69 2.05e007 1 1.34 1.32 1 1.20 1.16 1.05 1 0.78 0.79 0.92 1 0.66 1.03 1 0 0 0 1 0.53 0.46 0.27 1 1.536eþ008 1.674eþ008 1.546eþ008 1 0.10 0.49 0.93 1 1.41e008 0.90 0.77 1 0.75 0.61 3.54
0.463 0.773 0.618 0.292 0.837 0.169 0.683 0.421 0.753 0.850 0.179 0.708 0.946 0.829 0.233 0.049 0.941 0.068 0.076 0.342 0.082 0.385 0.853 0.349 0.238 0.922 0.861 0.375 0.387 0.883 0.369 0.725 0.003 0.103 0.320 0.004 0.046 0.821 0.552 0.047 0.366 0.010 0.315 0.590 0.363 0.609 0.625 0.754 0.562 0.492 0.214 0.489 0.950 0.158 0.067 0.984 0.098 0.184
0.568
SNP2-SNP3
SNP3-SNP4
SNP4-SNP5
SNP5-SNP6
SNP6-SNP7
SNP7-SNP8
SNP8-SNP9
SNP9-SNP10b
SNP10-SNP11b
SNP11-SNP12b
SNP12-SNP13
SNP13-SNP14
SNP14-SNP15
SNP15-SNP16
JOURNAL OF ORTHOPAEDIC RESEARCH MAY 2014
0.337
0.566
0.218
0.214
0.216
0.663
0.577
0.010
0.019
0.040
0.534
0.841
0.357
0.084
691
LUMBAR DISC DEGENERATION
Table 2. ðContinuedÞ Combination SNP16-SNP17
SNP17-SNP18b
SNP18-SNP19b
SNP19-SNP20
Haplotype
LDDa
CTRa
OR
Individual p-Value
Global p-Value
C-A C-G G-A A-A A-G G-A G-G A-A A-G G-A G-G A-A A-C G-A G-C
322 63 525 487.3 359.7 18.7 44.3 8.64 473.4 234.4 157.6 2.2 251.8 576.8 99.2
341 70 603 475.3 466.7 23.7 46.3 6.39 431.6 245.6 206.4 0 269 590 103
1 0.95 0.92 1 0.75 0.77 0.93 1 0.81 0.71 0.56 1 0 0 0
0.419 0.986 0.430 0.005 0.005 0.955 0.978 0.484 0.017 0.710 0.007 0.148 0.673 0.750 0.966
0.709
0.036
0.030
0.386
LDD, lumbar disc degeneration; CTR, control; OR, odds ratio. The global p-value between the LDD and control groups for haplotypic analysis obtained from 10,000 permutations was 0.0941. aStands for the number of the parameter. bIndicates p < 0.05 before adjusted.
and in vitro cell culture. Meanwhile, Seki et al.22 showed that use of ADAMTS-5 siRNA was effective in suppressing the degeneration of the nucleus pulposus in rabbit model, which further proved the participation of ADAMTS-5 in the process of LDD. Although functional studies have investigated the contribution of ADAMTS-5 to LDD, no definite genetic results have been reported so far. In 2007, Virtanen et al.26 performed the first genome-wide linkage study on LDD in 14 Finnish families. Two candidate genes (ADAMTS-1 and ADAMTS-5) were then analyzed in the exons and boundaries. Nevertheless, no significant association was found between phenotypes and the variations, which may be due to exclusion of noncoding regions. No further study was conducted to clarify the exact mechanism of this phenomenon. In our study, several alleles (rs151058, rs229052, and rs162502) locating in the introns of ADAMTS-5 were revealed to
be significantly associated with LDD, providing the evidence of the association of ADAMTS-5 with LDD. The G allele of rs151058, A allele of rs229052 and A allele of rs162502 were pathogenic alleles. Whereas, no significant associations were found in the genotypic and haplotypic analyses, which indicating that other than additive effects of ADAMTS-5 may be involved in the pathogenicity of LDD. The SNP-SNP Interactions May Exist Among Positive SNPs in ADAMTS-5 Though the introns do not influence protein function by changing amino acid sequence, regulatory polymorphisms occurring outside exonic regions which may affect transcriptional mRNA process or others have been supposed to be important modulators of gene expression and evolutionary change.34,35 Cisphase interaction may play an important role in the
Table 3. Interaction Analyses Among Allelic Positive SNPs in ADAMTS-5
SNP_ID
Conditional SNP
SNP 11
SNP 18
SNP 10
SNP 18
Haplotype
LDDa
CTRa
x2
OR
95% CI
p-Value
A–A A–G G–A G–G A–A A–G G–A G–G
406.3 100.7 317.7 85.3 10.5 471.5 70.5 321.5
400 101 368 145 3.1 434.9 117.9 334.1
0.143 0.428 1.634 3.968 1.426 0.029 5.816 2.339
1 0.98 0.85 0.58 1 0.32 0.17 0.28
1 0.68–1.41 0.68–1.06 0.42–0.80 1 0.07–1.53 0.03–0.88 0.58–1.34
0.705 0.513 0.201 0.046 0.232 0.864 0.016 0.126
Global p-Value 0.062b
0.010c
LDD, lumbar disc degeneration; CTR, control; OR, odds ratio; 95% CI, 95% confidence interval.aStands for the number of the parameter.bx2 ¼ 4.940, df ¼ 2, the adjusted p ¼ 0.085 after 10,000 permutation.cx2 ¼ 9.233, df ¼ 2, the adjusted p ¼ 0.019 after 10,000 permutation. JOURNAL OF ORTHOPAEDIC RESEARCH MAY 2014
692
WU ET AL.
Figure 2. T2 image and DTI map of lumbar IVDs in one participant. In this subject, the MD values from L1-2 to L5-S1 were 181, 191, 191, 182, and 159 102 m2/ms.
pathogenicity of complex trait disease. For example, Zhang et al.36 tested the effects of cis-acting SNPs on several gene expressions in peripheral white blood cell, and a significant difference of cis-effects on the MYO10 gene expression was firstly found to be correlated with metabolic syndrome. A cis-phase interaction between VNTR polymorphism and functional SNPs was also identified within the MAOA gene in major depressive disorder.37 Meanwhile, the synonymous SNPs within gene could also act as cis-phase regulators. Thumma et al.38 reported a synonymous SNP in Eucalyptus nitens COBRA-like gene, which could affect allelic expression of the downstream SNP through cis-phase interaction. In our study, the SNP-SNP interaction analysis showed the strong interactions between SNP 10—SNP
Figure 3. MD value scatter plots for the comparison of AA and Gþ carriers in rs162502. The y-axis displays the individual MD values of five lumbar IVDs. Lines represent the mean MD values. JOURNAL OF ORTHOPAEDIC RESEARCH MAY 2014
18 and SNP 11—SNP 18, while cis-phase interaction was revealed in SNP 10 when conditioning on SNP18. Although SNP 10 and SNP 18 are not in the same block, our result revealed the possibility that an interaction between the two blocks may contribute to the susceptibility to LDD in Chinese Han population. However, further functional study is required to confirm this finding. The ADAMTS-5 Association With Altered MD Values Numerous studies have addressed the association between genotypes and imaging analyses. By using tractography of DTI, Carballedo et al.39 demonstrated that the met-allele of the brain-derived neurotrophic factor (BDNF) polymorphism probably rendered subjects more vulnerabilities to dysfunctions of the uncinate fasciculus. Spalletta et al.40 found that patients with the GSTA1 B risk allele had an increased MD value in bilateral thalami. However, most of these reports were concentrated on brain related disease, no such investigation combining genetics and functional MRI has ever been performed in the field of LDD. Degenerative changes of lumbar IVDs in LDD have been proved to accompany with water content reduction and shrinkage of the disc volume.41 As a valuable and precise DTI quantitative parameter, MD was shown to be worthwhile in evaluating microstructural alterations in IVD during LDD. It was reported that the more degenerated discs harbor lower MD values.29 As has been demonstrated in our study, the A allele of rs162502 was the pathogenic allele associated with LDD, it is reasonable to believe that patients with A allele of rs162502 will experience more degenerative changes in IVDs. Furthermore, our results of the second-stage analysis was also consistent with the
LUMBAR DISC DEGENERATION
693
Table 4. Individual MD Values in Different Genotypes of Positive SNPs SNP_ID
Genotype
N
M SD (102)a
t-value
df
p-value
SNP 10
GG AG þ AA AA AG þ GG AA AG þ GG
37 13 28 22 19 31
834.02 57.36 829.37 83.92 834.93 59.25 830.97 69.47 807.23 54.34 847.30 65.88
0.22
48
0.829
0.26
48
0.854
2.138
48
0.038b
SNP 11 SNP 18
N stands for the number of different genotypes; M SD stands for mean standard deviation. aThe unit is m2/ms. bIndicates p < 0.05.
previous hypothesis, indicating that AA carriers in rs162502 showed a lower MD values (807.23 54.34 102 m2/ms) compared with Gþ carriers (847.30 65.88 102 m2/ms). On the other hand, the dose effect of A allele and the relative imbalance of the sample numbers may provide a bias to this result. Limitations There are several potential limitations in our study. First, as a pilot study, functions of the positive SNPs were not investigated. Second, the sample size of imaging analysis is rather small so that the type II errors cannot be overcomed. Third, given that our Chinese sample cannot be considered as all population representative, studies in other ethnics will be greatly encouraged. Summary For the first time, we reported the genetic polymorphisms of ADAMTS-5 (rs151058, rs229052, and rs162502) may be associated with susceptibility to LDD. A cis-phase interaction between rs151058 and rs229052 may exist in ADAMTS-5 with LDD. Furthermore, we firstly combined the genetic association study with diffusion tensor imaging technique in the field of orthopedics indicating that the rs162502 might be associated with altered MD values.
REFERENCES 1. Andersson GB. 1999. Epidemiological features of chronic low-back pain. Lancet 354:581–585. 2. Frymoyer JW. 1988. Back pain and sciatica. N Engl J Med 318:291–300. 3. Luoma K, Riihimaki H, Luukkonen R, et al. 2000. Low back pain in relation to lumbar disc degeneration. Spine (Phila Pa 1976) 25:487–492. 4. Heliovaara M, Impivaara O, Sievers K, et al. 1987. Lumbar disc syndrome in Finland. J Epidemiol Community Health 41:251–258. 5. Kalichman L, Hunter DJ. 2008. The genetics of intervertebral disc degeneration. Familial predisposition and heritability estimation. Joint Bone Spine 75:383–387. 6. Varlotta GP, Brown MD, Kelsey JL, et al. 1991. Familial predisposition for herniation of a lumbar disc in patients who are less than twenty-one years old. J Bone Joint Surg Am 73:124–128. 7. Matsui H, Terahata N, Tsuji H, et al. 1992. Familial predisposition and clustering for juvenile lumbar disc herniation. Spine (Phila Pa 1976) 17:1323–1328.
8. Ala-Kokko L. 2002. Genetic risk factors for lumbar disc disease. Ann Med 34:42–47. 9. Sambrook PN, MacGregor AJ, Spector TD. 1999. Genetic influences on cervical and lumbar disc degeneration: a magnetic resonance imaging study in twins. Arthritis Rheum 42:366–372. 10. Williams FM, Popham M, Hart DJ, et al. 2011. GDF5 singlenucleotide polymorphism rs143383 is associated with lumbar disc degeneration in Northern European women. Arthritis Rheum 63:708–712. 11. Le Maitre CL, Pockert A, Buttle DJ, et al. 2007. Matrix synthesis and degradation in human intervertebral disc degeneration. Biochem Soc Trans 35:652–655. 12. Kalb S, Martirosyan NL, Kalani MY, et al. 2012. Genetics of the degenerated intervertebral disc. World Neurosurg 77:491–501. 13. Roberts S, Caterson B, Menage J, et al. 2000. Matrix metalloproteinases and aggrecanase: their role in disorders of the human intervertebral disc. Spine (Phila Pa 1976) 25:3005–3013. 14. Le Maitre CL, Freemont AJ, Hoyland JA. 2004. Localization of degradative enzymes and their inhibitors in the degenerate human intervertebral disc. J Pathol 204:47–54. 15. Pockert AJ, Richardson SM, Le Maitre CL, et al. 2009. Modified expression of the ADAMTS enzymes and tissue inhibitor of metalloproteinases 3 during human intervertebral disc degeneration. Arthritis Rheumat 60:482–491. 16. Bachmeier BE, Nerlich A, Mittermaier N, et al. 2009. Matrix metalloproteinase expression levels suggest distinct enzyme roles during lumbar disc herniation and degeneration. Eur Spine J 18:1573–1586. 17. Glasson SS, Askew R, Sheppard B, et al. 2005. Deletion of active ADAMTS5 prevents cartilage degradation in a murine model of osteoarthritis. Nature 434:644–648. 18. Stanton H, Rogerson FM, East CJ, et al. 2005. ADAMTS5 is the major aggrecanase in mouse cartilage in vivo and in vitro. Nature 434:648–652. 19. Tortorella MD, Burn TC, Pratta MA, et al. 1999. Purification and cloning of aggrecanase-1: a member of the ADAMTS family of proteins. Science 284:1664–1666. 20. Somerville RP, Longpre JM, Jungers KA, et al. 2003. Characterization of ADAMTS-9 and ADAMTS-20 as a distinct ADAMTS subfamily related to Caenorhabditis elegans GON-1. J Biol Chem 278:9503–9513. 21. Liang QQ, Zhou Q, Zhang M, et al. 2008. Prolonged upright posture induces degenerative changes in intervertebral discs in rat lumbar spine. Spine (Phila Pa 1976) 33:2052–2058. 22. Patel KP, Sandy JD, Akeda K, et al. 2007. Aggrecanases and aggrecanase-generated fragments in the human intervertebral disc at early and advanced stages of disc degeneration. Spine (Phila Pa 1976) 32:2596–2603. 23. Le Maitre CL, Freemont AJ, Hoyland JA. 2007. Accelerated cellular senescence in degenerate intervertebral discs: a
JOURNAL OF ORTHOPAEDIC RESEARCH MAY 2014
694
24.
25.
26.
27.
28. 29.
30.
31.
32.
33.
WU ET AL. possible role in the pathogenesis of intervertebral disc degeneration. Arthritis Res Ther 9:R45. Zhao CQ, Zhang YH, Jiang SD, et al. 2011. ADAMTS-5 and Intervertebral disc degeneration: the results of tissue immunohistochemistry and in vitro cell culture. J Orthop Res 29: 718–725. Le Maitre CL, Freemont AJ, Hoyland JA. 2007. Accelerated cellular senescence in degenerate intervertebral discs: a possible role in the pathogenesis of intervertebral disc degeneration. Arthritis Res Therapy 9:R45. Virtanen IM, Noponen N, Barral S, et al. 2007. Putative susceptibility locus on chromosome 21q for lumbar disc disease (LDD) in the Finnish population. J Bone Miner Res 22:701–707. Carballido-Gamio J, Xu D, Newitt D, et al. 2007. Single-shot fast spin-echo diffusion tensor imaging of the lumbar spine at 1.5 and 3 T. Magn Reson Imaging 25:665–670. Haughton V. 2006. Imaging intervertebral disc degeneration. J Bone Joint Surg Am 88:15–20. Zhang Z, Chan Q, Anthony MP, et al. 2012. Age-related diffusion patterns in human lumbar intervertebral discs: a pilot study in asymptomatic subjects. Magn Reson Imaging 30:181–188. Basser PJ, Pierpaoli C. 1996. Microstructural and physiological features of tissues elucidated by quantitative-diffusiontensor MRI. J Magn Reson B 111:209–219. Gabriel SB, Schaffner SF, Nguyen H, et al. 2002. The structure of haplotype blocks in the human genome. Science 296:2225–2229. Takahashi M, Haro H, Wakabayashi Y, et al. 2001. The association of degeneration of the intervertebral disc with 5a/6a polymorphism in the promoter of the human matrix metalloproteinase-3 gene. J Bone Joint Surg Br 83:491–495. Sun ZM, Miao L, Zhang YG, et al. 2009. Association between the 1562 C/T polymorphism of matrix metalloproteinase-9 gene and lumbar disc disease in the young adult population in North China. Connect Tissue Res 50:181–185.
JOURNAL OF ORTHOPAEDIC RESEARCH MAY 2014
34. Nackley AG, Shabalina SA, Tchivileva IE, et al. 2006. Human catechol-O-methyltransferase haplotypes modulate protein expression by altering mRNA secondary structure. Science 314:1930–1933. 35. Cordell HJ. 2009. Detecting gene-gene interactions that underlie human diseases. Nat Rev Genet 10:392–404. 36. Zhang Y, Kent JJ, Olivier M, et al. 2013. A comprehensive analysis of adiponectin QTLs using SNP association, SNP cis-effects on peripheral blood gene expression and gene expression correlation identified novel metabolic syndrome (MetS) genes with potential role in carcinogenesis and systemic inflammation. BMC Med Genomics 6:14. 37. Zhang J, Chen Y, Zhang K, et al. 2010. A cis-phase interaction study of genetic variants within the MAOA gene in major depressive disorder. Biol Psychiatry 68:795–800. 38. Thumma BR, Matheson BA, Zhang D, et al. 2009. Identification of a Cis-acting regulatory polymorphism in a Eucalypt COBRA-like gene affecting cellulose content. Genetics 183: 1153–1164. 39. Carballedo A, Amico F, Ugwu I, et al. 2012. Reduced fractional anisotropy in the uncinate fasciculus in patients with major depression carrying the met-allele of the Val66Met brain-derived neurotrophic factor genotype. Am J Med Genet B Neuropsychiatr Genet 159B:537–548. 40. Spalletta G, Piras F, Gravina P, et al. 2012. Glutathione Stransferase alpha 1 risk polymorphism and increased bilateral thalamus mean diffusivity in schizophrenia. Psychiatry Res 203:180–183. 41. Hughes SP, Freemont AJ, Hukins DW, et al. 2012. The pathogenesis of degeneration of the intervertebral disc and emerging therapies in the management of back pain. J Bone Joint Surg Br 94:1298–1304.
SUPPORTING INFORMATION Additional supporting information may be found in the online version of this article at the publisher’s web-site.
Received 19 October 2013; accepted 17 December 2013 Published online 10 January 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jor.22582
ABSTRACT: Purpose: Low back pain is a global health problem in which more than 40% is caused by lumbar intervertebral disc degeneration (LDD). ADAMTS-5 (A disintegrin and metalloproteinase with thrombospondin motifs-5) was shown to be involved in LDD by functional analyses. To identify whether there is an association between ADAMTS-5 and LDD, and what is the contribution of ADAMTS-5 genetic polymorphisms to MD (Mean diffusivity) changes in lumbar IVD (Intervertebral disc). We firstly genotyped selected ADAMTS-5 SNPs (Single nucleotide polymorphisms) in a Chinese Han population. After the primary analyses of allelic, genotypic, and haplotypic association, we performed SNP–SNP interaction analysis. We subsequently genotyped another 50 participants and acquired the corresponding MD values from individual lumbar IVDs. The association analysis between the genotypic groups divided by the above positive SNPs and the corresponding MD values were also performed. Significant associations were identified in rs151058, rs229052, and rs162502. None of the 2-SNP haplotypic analysis survived the 10,000 permutation test. The following interaction analysis demonstrated that rs151058 was strong associated with LDD when conditioning on rs162502. Significant difference of MD values between AA and Gþ carriers was identified in rs162502. This is the first study indicating that the SNPs of ADAMTS-5 may contribute to predisposition of LDD. An interaction between rs151058 and rs229052 may exist in ADAMTS-5 with LDD. The rs162502 might be associated with altered MD values. ß 2014 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 32:686–694, 2014. Keywords: lumbar disc degeneration (LDD); a disintegrin and metalloproteinase with thrombospondin motifs-5 (ADAMTS-5); single nucleotide polymorphism (SNP); diffusion tensor imaging (DTI)
As a global health problem, low back pain (LBP) was reported to affect about 80% of population during a lifetime.1 LBP, defined as pain radiating from the back into the dermatome of the affected nerve trunk, is a common musculoskeletal disorder characterized by symptomatic lumbar disc herniation with or without sciatica.2 More than 40% of LBP is caused by lumbar intervertebral disc degeneration (LDD).3 It was reported that about 5% of the Finnish population suffered from LDD, leading to disability in the workingage population.4 Though several environmental and anthropometric risk factors have been showed to be related to LDD, the roles of genetic factors can never be exaggerated.5–10 The extracellular matrix (ECM), composed of collagens and proteoglycans, is crucial to normal intervertebral disc (IVD) functions.11 Previous studies indicated that the synthesis and degradation balance of ECM was interrupted by diverse matrix proteases during LDD.12–16 As a large family of metalloproteases, ADAMTS-5 seems to be the most active Conflicts of interest: none. Nan Wu, Jun Chen, and Hao Liu contributed equally to this study. Grant sponsor: National Natural Science Foundation of China; Grant numbers: 81071513, 81130034, 81271942; Grant sponsor: Public Service of the Health Ministry of China; Grant number: 201002018. Correspondence to: Zhihong Wu (T: þ8601069156081; F: þ8601069156081; E-mail: [email protected]) # 2014 Orthopaedic Research Society. Published by Wiley Periodicals, Inc.
686
JOURNAL OF ORTHOPAEDIC RESEARCH MAY 2014
one,15,17,18 and has been paid increasing attentions due to their abilities in the cartilage proteoglycan cleavage.19,20 Besides mRNA and protein of ADAMTS5 were detected in IVDs,21 increased levels in patients with chronic LBP and intervertebral disc herniation during LDD can be found.22–24 Although several functional analyses revealed the involvement of ADAMTS-5 in LDD22–25 and even a pedigree-based research was also conducted,26 no positive conclusion has been achieved in the relationship between genetic alterations of ADAMTS-5 and LDD. As a newly developed magnetic resonance imaging (MRI) technique, diffusion tensor imaging (DTI) can non-invasively provide the microstructural properties of soft tissue. By adding a diffusion-gradients signal to the standard MRI sequence, DTI can deliver a more detailed imaging for random water molecule motion than normal MRI. Numerous literatures have shown that changes in DTI parameters of lumbar IVDs are related to the degree of LDD.27,28 Several methods have been used to measure DTI, among which, mean diffusivity (MD) is one of the most effective parameters for the IVD degenerative status evaluation. MD can reflect the average diffusion rate in a tissue and represent changes in the water content of intervertebral disc.29 However, up to our knowledge, literatures combining genetic association study with DTI in the field of orthopedics can rarely be found till now. The present study was designed to answer the following questions: (1) Is there an association between ADAMTS-5 gene and LDD in Chinese Northern Han
LUMBAR DISC DEGENERATION
population? (2) What is the contribution of ADAMTS-5 genetic polymorphisms to DTI changes in lumbar IVDs.
MATERIALS AND METHODS Subjects A total of 489 unrelated sporadic cases with LDD (240 males and 249 females, aged 42.6 14.2 years) and 558 control subjects (271 males and 287 females, aged 41.4 11.1 years) from Chinese Northern Han population were recruited with no differences of age and gender between the two groups. All the patients and controls were diagnosed through lumbar disc MRI scanning by at least two experienced radiologists. Primary exclusion criteria included spinal and joint diseases such as trauma, spinal tumor, inflammatory, scoliosis, and osteoarthritis. Individuals with known environmental risk factors, including heavy manual labor, occupational driving, or heavy smoking, were also excluded. Written informed consent were obtained from all participants. The Ethics Committee of Peking Union Medical College Hospital, Peking Union Medical College and Chinese Academy of Medical Sciences approved this study. Genotyping of SNPs Twenty SNPs ranging from 5 kb upstream and downstream of ADAMTS-5 were selected from the NCBI SNP database (www.ncbi.nlm.nih.gov/SNP) and the HapMap database (www.hapmap.org). The SNPs included were as follows: rs229070 (SNP 1), rs151065 (SNP 2), rs229079 (SNP 3), rs226794 (SNP 4), rs2830585 (SNP 5), rs2830586 (SNP 6), rs162499 (SNP 7), rs162495 (SNP 8), rs162489 (SNP 9), rs151058 (SNP10), rs229052 (SNP 11), rs229054 (SNP 12), rs9984329 (SNP 13), rs2249350 (SNP 14), rs233896 (SNP 15), rs162509 (SNP 16), rs162506 (SNP 17), rs162502 (SNP 18), rs2132824 (SNP 19) and rs1974415 (SNP 20) (shown in Supplementary Table S1). Among them, SNP 1 and SNP 2 are in the 30 UTR region, SNP 5 in the exon region, SNP 19 and SNP 20 in the 50 near region, and others are in the introns. All the SNPs selected have a minor allele frequency (MAF) of more than 0.05. Genomic DNA was extracted from peripheral blood leukocytes using a standard phenol-chloroform method. The SNP genotyping was performed on the Sequenom MassARRAY SNP genotyping platform (Sequenom, San Diego, CA). Five percentage duplicates of randomly chosen samples were used for quality controls. MRI Data Acquisition Another 50 participants (26 female, 24 male; mean age ¼ 45.5 9.5 years; age range ¼ 26–61 years) who met the following criteria were enrolled: (1) No history of LBP or other spinal diseases (trauma, infection, inflammation and deformity, etc.); (2) no contraindications for MRI scanning; (3) no systemic disorders, such as diabetes, hypertension, hyperlipidemia, etc. Written informed consent were obtained from all participants. All subjects underwent MRI scans on a clinical 3.0-T scanner (Signa EXCITE, GE Medical Systems, Milwaukee, WI) from January to April 2013. Sagittal T2 images of five IVDs (from L1-2 to L5-S1) were acquired through T2weighted spin echo sequences. The sagittal DTI images were then acquired with a standard single-shot diffusion-weighted echo-planar imaging (DW-EPI) sequence (b ¼ 800 s/mm2) in 21 noncollinear diffusion-encoding directions. Parameters for acquisition were as follows: TR/TE ¼ 5,000/85.5 ms,
687
thickness/gap ¼ 4/0 mm, matrix ¼ 128 128, FOV ¼ 32 32 mm2, 14 slices which lasted about 6 min. Images Processing FuncTool Performance software (GE Medical Systems) was used to calculate the MD maps according to the standard formulas described by Basser et al.30. For evaluation of MD values, a region of interest (ROI) analysis was then performed. The elliptical ROIs (90–110 mm2) were manually placed in the center of the highest signal area at each level (b ¼ 0 s/mm2). They were then projected to the MD maps, and the ROI mean signal intensities were thus calculated. The individual MD value for one participant was the mean of all five IVD MD intensities. A senior spine radiologist (with 11 years of experience) and a spine surgeon (with 6 years of experience) drew the ROIs independently, and the mean values were then acquired for the following analysis. Data Analysis
Association analysis Hardy–Weinberg equilibrium (HWE), haplotype blocks, and the linkage disequilibrium (LD) patterns were estimated by using the Haploview program (version 4.2, Broad Institute of MIT and Harvard, Cambridge, MA).31 Primary analyses were performed for allelic, genotypic and haplotypic association based on the UNPHASED software (v.3.1.5 Dudbridge F, MRC Biostatistics Unit, Cambridge, UK). In our study, the sliding window size for haplotype analysis was two. To test an underlying interaction among selected SNPs within the ADAMTS-5 gene, SNP–SNP interaction analyses were then performed. Ten thousand times of permutation test, a built-in program of UNPHASED, was conducted to obtain an adjusted p-value. QUANTO software (http://hydra.usc.edu/ gxe) was used to calculate the power value, in which the prevalence of LDD was set at 32%, the false positive rate was 0.05 with a minor allele frequency of 0.10. Association of Positive SNPs and MD Values Fifty participants were further grouped according to the individual genotypes of SNPs associated with LDD in the above analysis. The Student’s t-test was then used to detect the differences of total MD values between paired groups. Differences were considered statistically significant when p < 0.05. Statistical analysis was performed by using SPSS 17 (SPSS, Inc., Chicago, IL).
RESULTS The genotypic distributions of the 20 selected SNPs did not deviate from Hardy–Weinberg equilibrium in the control subjects (Supplementary Table S2). Six LD blocks were observed according to Confidence intervetals’ model31 (D0 > 0.9 and r2 > 0.8; Fig. 1). Power calculation showed that our sample had 80.96% power to detect allelic association at a false rate of 0.05. The analysis of allele frequency indicated the association signals in rs151058 (SNP 10, x2 ¼ 9.069, df ¼ 1, unadjusted p ¼ 0.003), rs229052 (SNP 11, x2 ¼ 4.445, df ¼ 1, unadjusted p ¼ 0.035), and rs162502 (SNP 18, x2 ¼ 7.393, df ¼ 1, unadjusted p ¼ 0.007), which survived the 10,000 permutation correction with a global p value of 0.042 (Table 1). The genotypic analysis showed associations of SNP 10 (x2 ¼ 9.312, df ¼ 2, p ¼ 0.010) and SNP 18 (x2 ¼ 7.179, df ¼ 2, JOURNAL OF ORTHOPAEDIC RESEARCH MAY 2014
688
WU ET AL.
Figure 1. LD structures of the twenty SNPs genotyped in ADAMTS-5 gene. The LD strengths between paired SNPs are shown in color according to confidence intervals’ model.
p ¼ 0.028) with LDD, however, neither of them survived the 10000 permutation correction (global p-value ¼ 0.131; Supplementary Table S3). All haplotypes used for analysis had a frequency of >1% in either patients or the control groups. The 2-SNP haplotypic analysis showed significant associations for SNP 9-SNP 10 (x2 ¼ 9.240, df ¼ 3, p ¼ 0.010), SNP 10-SNP 11 (x2 ¼ 9.971, df ¼ 3, p ¼ 0.019), SNP 11-SNP 12 (x2 ¼ 8.286, df ¼ 3, p ¼ 0.040), SNP 17-SNP 18 (x2 ¼ 8.532, df ¼ 3, p ¼ 0.036) and SNP 18-SNP 19 (x2 ¼ 8.930, df ¼ 3, p ¼ 0.030). The global p value after 10,000 permutation test was 0.0941 (Table 2). Three positive SNPs—rs151058 (SNP 10), rs229052 (SNP 11), and rs162502 (SNP 18) were selected to perform the following SNP-SNP interaction analyses. The results demonstrated a combined effects by confirming disease association for the SNP 10—SNP 18 combination (x2 ¼ 15.462, df ¼ 3, unadjusted p ¼ 0.001, adjusted p ¼ 0.003) and SNP 11—SNP18 combination (x2 ¼ 12.593, df ¼ 3, unadjusted p ¼ 0.006, adjusted p ¼ 0.007; Supplementary Table S4). Analyses of the cis-phase interaction demonstrated that of the two SNPs tested, SNP 10 showed strong association with LDD when conditioning on SNP 18 (x2 ¼ 9.233, df ¼ 2, unadjusted p ¼ 0.010, adjusted p ¼ 0.019), while no such interaction were observed for SNP 11 when conditioning on SNP 18 (x2 ¼ 4.940, df ¼ 2, unadjusted p ¼ 0.062, adjusted p ¼ 0.085) (Table 3). JOURNAL OF ORTHOPAEDIC RESEARCH MAY 2014
According to the positive SNPs confirmed in the above analyses, the 50 participants were defined as GG carriers (37 participants) and Aþ carriers(AG þ AA, 12 participants) for SNP10, AA carriers (28 participants), and Gþ carriers(AG þ GG, 22 participants) for SNP11, and AA carriers (17 participants) and Gþ carriers(AG þ GG, 31 participants) for SNP18. The mean MD values of the five IVDs from 50 subjects were 833.19 63.30 102 m2/ms (shown in mean SD). A representative T2 image and the corresponding MD maps in one participant were shown in Figure 2. The MD values from L1/2 to L5/S1 IVDs were 181, 191, 191, 182, and 159 102 m2/ms, respectively. Significant differences of MD values between two genotypic groups were identified for SNP18 (t ¼ 2.138, df ¼ 46, p ¼ 0.038). The mean MD values for AA and Gþ carriers were 807.23 54.34 102 m2/ms and 847.30 65.88 102 m2/ms, respectively (shown in mean SD), and the scatter plot was shown in Figure 3. No such significance was observed for SNP 10 (t ¼ 0.22, df ¼ 48, p ¼ 0.829) and SNP 11 (t ¼ 0.26, df ¼ 48, p ¼ 0.854; Table 4).
DISCUSSION For the first time, our study revealed the genetic polymorphisms of ADAMTS-5 may be associated with susceptibility to LDD in Chinese Northern Han population and contribute to DTI changes in lumbar IVDs.
LUMBAR DISC DEGENERATION
689
Table 1. Allelic Associations of ADAMTS-5 Tested SNPs With LDD LDD SNP_ID SNP 1 SNP 2 SNP 3 SNP 4 SNP 5 SNP 6 SNP 7 SNP 8 SNP 9 SNP 10a SNP 11a SNP 12 SNP 13 SNP 14 SNP 15 SNP 16 SNP 17 SNP 18a SNP 19 SNP 20
CTR
Allele
N
%
N
%
OR
95% CI
x2
p-value
C G A G A G A G C T G T A G C T G T A G A G A G C T A C A C C G A G A G A G A C
752 158 486 426 717 195 492 438 846 66 67 843 690 220 587 325 260 670 84 846 724 186 167 763 548 382 337 571 281 77 385 525 847 63 508 404 254 676 579 351
82.6 17.4 53.3 46.7 78.6 21.4 52.9 47.1 92.8 7.2 7.4 92.6 75.8 24.2 64.4 35.6 28.0 72.0 9.0 91.0 79.6 20.4 18.0 82.0 58.9 41.1 37.1 62.9 78.5 21.5 42.3 57.7 93.1 6.9 55.7 44.3 27.3 72.7 62.3 37.7
848 166 546 468 807 203 503 459 917 97 97 917 775 241 674 342 262 700 129 833 769 249 183 779 548 414 367 645 603 215 411 603 948 72 503 513 269 693 591 373
83.6 16.4 53.8 46.2 80.0 20.1 52.3 47.7 90.4 9.6 9.6 90.4 76.3 23.7 66.3 33.7 27.2 72.8 13.4 86.6 75.5 24.5 19.0 81.0 57.0 43.0 36.3 63.8 73.7 26.3 40.5 59.5 92.9 7.1 49.5 50.5 28.0 72.0 61.3 38.7
1 1.07 1 1.02 1 1.08 1 0.99 1 0.74 1 1.33 1 1.03 1 1.09 1 0.97 1 1.56 1 0.79 1 1.07 1 0.92 1 0.96 1 0.77 1 0.93 1 0.98 1 0.78 1 1.03 1 0.96
1 0.85–1.36 1 0.86–1.22 1 0.87–1.35 1 0.81–1.17 1 0.53–1.02 1 0.96–1.84 1 0.83–1.264 1 0.90–1.32 1 0.79–1.18 1 1.17–2.09 1 0.64–0.98 1 0.85–1.35 1 0.77–1.11 1 0.80–1.16 1 0.57–1.03 1 0.78–1.11 1 0.69–1.39 1 0.65–0.93 1 0.84–1.26 1 0.80–1.16
0.337
0.5616
0.060
0.8068
0.480
0.4884
0.072
0.7884
3.363
0.0667
2.986
0.0840
0.055
0.8151
0.828
0.3628
0.123
0.7254
9.069
0.0026
4.445
0.0350
0.356
0.5506
0.745
0.3879
0.149
0.6997
3.042
0.0811
0.623
0.4299
0.014
0.9071
7.393
0.0066
0.100
0.7517
0.181
0.6703
LDD, lumbar disc degeneration; CTR, control; OR, odds ratio; N, number; %, percent; 95% CI, 95% confidence interval. The adjusted pvalue between the LDD and control groups for allelic association obtained from 10,000 permutations was 0.0419. aIndicates p < 0.05 before adjusted.
This may give an explanation to the phenomenon reported by Virtanen et al.26 and lead to a new way for better understanding of LDD. The ADAMTS-5 Association With LDD Several studies have suggested that polymorphism of genes encoding ECM degradation enzymes were associated with LDD. For example, Takahashi et al.32 found that the 5A5A and 5A6A genotypes of matrix metalloproteinase (MMP)-3 gene showed a significant association with degenerative IVDs in
elderly Japanese population. Sun et al.33 reported 1562C/T polymorphism of MMP-9 gene was associated with a high risk of LDD in young adult population in Northern Chinese population. The involvement of ADAMTS-5 gene in LDD has been proved in vivo and in vitro. Patel and his colleagues revealed abundant levels of ADAMTS-5 in human cadaveric IVDs.22 Le Maitre and coworkers25 have reported the relationship between expression of ADAMTS-5 and altered phenotype of IVD cells during degeneration, which was also proved by Zhao et al.24 through immunohistochemistry JOURNAL OF ORTHOPAEDIC RESEARCH MAY 2014
690
WU ET AL.
Table 2. Association Analyses of 2-SNP Haplotypes in ADAMTS-5 With LDD Combination
Haplotype
LDDa
CTRa
OR
Individual p-Value
Global p-Value
SNP1-SNP2
C-A C-G G-A G-G A-A A-G G-A G-G A-A A-G G-A G-G A-C A-T G-C G-T C-G C-T T-G T-T G-A T-A T-G A-C A-T G-C G-T C-G C-T T-G T-T G-G T-A T-G A-A A-G G-A G-G A-A A-G G-A G-G A-C A-T G-C G-T C-A C-C T-A T-C A-A A-C C-A C-C A-C A-G C-C C-G
327.5 424.5 156.5 1.5 486 0 231 195 448.2 237.8 7.8 180.2 454.4 1.6 356.6 61.4 1 843 66 0 67 623 220 582.7 107.3 2.3 217.7 53.24 510.8 193.8 116.2 260 84 586 1.1 79.9 692.9 98.1 2.1 691.9 152.9 25.1 0 167 548 215 320.8 187.2 1.2 360.8 135 0 123 74 56.4 203.6 71.7 2.3
480 466 166 0 542.3 2.7 264.7 200.3 461.2 242.8 2.8 177.2 462.3 1.7 336.7 85.3 1.0 912 96 1.0 97 674 241 668.8 102.2 2.2 238.8 44.0 540 202 104 262 129 571 0 121 672 97 1.1 670.9 170.9 47.1 1.3 181.7 546.7 232.3 316.5 185.5 2.5 381.6 278.5 1.5 280.5 199.5 96.1 464.9 199.9 1.1
1 1.06 1.09 1.90eþ008 1 0 0.97 1.09 1 1.01 2.87 1.05 1 0.96 1.08 0.73 1 0.93 0.69 2.05e007 1 1.34 1.32 1 1.20 1.16 1.05 1 0.78 0.79 0.92 1 0.66 1.03 1 0 0 0 1 0.53 0.46 0.27 1 1.536eþ008 1.674eþ008 1.546eþ008 1 0.10 0.49 0.93 1 1.41e008 0.90 0.77 1 0.75 0.61 3.54
0.463 0.773 0.618 0.292 0.837 0.169 0.683 0.421 0.753 0.850 0.179 0.708 0.946 0.829 0.233 0.049 0.941 0.068 0.076 0.342 0.082 0.385 0.853 0.349 0.238 0.922 0.861 0.375 0.387 0.883 0.369 0.725 0.003 0.103 0.320 0.004 0.046 0.821 0.552 0.047 0.366 0.010 0.315 0.590 0.363 0.609 0.625 0.754 0.562 0.492 0.214 0.489 0.950 0.158 0.067 0.984 0.098 0.184
0.568
SNP2-SNP3
SNP3-SNP4
SNP4-SNP5
SNP5-SNP6
SNP6-SNP7
SNP7-SNP8
SNP8-SNP9
SNP9-SNP10b
SNP10-SNP11b
SNP11-SNP12b
SNP12-SNP13
SNP13-SNP14
SNP14-SNP15
SNP15-SNP16
JOURNAL OF ORTHOPAEDIC RESEARCH MAY 2014
0.337
0.566
0.218
0.214
0.216
0.663
0.577
0.010
0.019
0.040
0.534
0.841
0.357
0.084
691
LUMBAR DISC DEGENERATION
Table 2. ðContinuedÞ Combination SNP16-SNP17
SNP17-SNP18b
SNP18-SNP19b
SNP19-SNP20
Haplotype
LDDa
CTRa
OR
Individual p-Value
Global p-Value
C-A C-G G-A A-A A-G G-A G-G A-A A-G G-A G-G A-A A-C G-A G-C
322 63 525 487.3 359.7 18.7 44.3 8.64 473.4 234.4 157.6 2.2 251.8 576.8 99.2
341 70 603 475.3 466.7 23.7 46.3 6.39 431.6 245.6 206.4 0 269 590 103
1 0.95 0.92 1 0.75 0.77 0.93 1 0.81 0.71 0.56 1 0 0 0
0.419 0.986 0.430 0.005 0.005 0.955 0.978 0.484 0.017 0.710 0.007 0.148 0.673 0.750 0.966
0.709
0.036
0.030
0.386
LDD, lumbar disc degeneration; CTR, control; OR, odds ratio. The global p-value between the LDD and control groups for haplotypic analysis obtained from 10,000 permutations was 0.0941. aStands for the number of the parameter. bIndicates p < 0.05 before adjusted.
and in vitro cell culture. Meanwhile, Seki et al.22 showed that use of ADAMTS-5 siRNA was effective in suppressing the degeneration of the nucleus pulposus in rabbit model, which further proved the participation of ADAMTS-5 in the process of LDD. Although functional studies have investigated the contribution of ADAMTS-5 to LDD, no definite genetic results have been reported so far. In 2007, Virtanen et al.26 performed the first genome-wide linkage study on LDD in 14 Finnish families. Two candidate genes (ADAMTS-1 and ADAMTS-5) were then analyzed in the exons and boundaries. Nevertheless, no significant association was found between phenotypes and the variations, which may be due to exclusion of noncoding regions. No further study was conducted to clarify the exact mechanism of this phenomenon. In our study, several alleles (rs151058, rs229052, and rs162502) locating in the introns of ADAMTS-5 were revealed to
be significantly associated with LDD, providing the evidence of the association of ADAMTS-5 with LDD. The G allele of rs151058, A allele of rs229052 and A allele of rs162502 were pathogenic alleles. Whereas, no significant associations were found in the genotypic and haplotypic analyses, which indicating that other than additive effects of ADAMTS-5 may be involved in the pathogenicity of LDD. The SNP-SNP Interactions May Exist Among Positive SNPs in ADAMTS-5 Though the introns do not influence protein function by changing amino acid sequence, regulatory polymorphisms occurring outside exonic regions which may affect transcriptional mRNA process or others have been supposed to be important modulators of gene expression and evolutionary change.34,35 Cisphase interaction may play an important role in the
Table 3. Interaction Analyses Among Allelic Positive SNPs in ADAMTS-5
SNP_ID
Conditional SNP
SNP 11
SNP 18
SNP 10
SNP 18
Haplotype
LDDa
CTRa
x2
OR
95% CI
p-Value
A–A A–G G–A G–G A–A A–G G–A G–G
406.3 100.7 317.7 85.3 10.5 471.5 70.5 321.5
400 101 368 145 3.1 434.9 117.9 334.1
0.143 0.428 1.634 3.968 1.426 0.029 5.816 2.339
1 0.98 0.85 0.58 1 0.32 0.17 0.28
1 0.68–1.41 0.68–1.06 0.42–0.80 1 0.07–1.53 0.03–0.88 0.58–1.34
0.705 0.513 0.201 0.046 0.232 0.864 0.016 0.126
Global p-Value 0.062b
0.010c
LDD, lumbar disc degeneration; CTR, control; OR, odds ratio; 95% CI, 95% confidence interval.aStands for the number of the parameter.bx2 ¼ 4.940, df ¼ 2, the adjusted p ¼ 0.085 after 10,000 permutation.cx2 ¼ 9.233, df ¼ 2, the adjusted p ¼ 0.019 after 10,000 permutation. JOURNAL OF ORTHOPAEDIC RESEARCH MAY 2014
692
WU ET AL.
Figure 2. T2 image and DTI map of lumbar IVDs in one participant. In this subject, the MD values from L1-2 to L5-S1 were 181, 191, 191, 182, and 159 102 m2/ms.
pathogenicity of complex trait disease. For example, Zhang et al.36 tested the effects of cis-acting SNPs on several gene expressions in peripheral white blood cell, and a significant difference of cis-effects on the MYO10 gene expression was firstly found to be correlated with metabolic syndrome. A cis-phase interaction between VNTR polymorphism and functional SNPs was also identified within the MAOA gene in major depressive disorder.37 Meanwhile, the synonymous SNPs within gene could also act as cis-phase regulators. Thumma et al.38 reported a synonymous SNP in Eucalyptus nitens COBRA-like gene, which could affect allelic expression of the downstream SNP through cis-phase interaction. In our study, the SNP-SNP interaction analysis showed the strong interactions between SNP 10—SNP
Figure 3. MD value scatter plots for the comparison of AA and Gþ carriers in rs162502. The y-axis displays the individual MD values of five lumbar IVDs. Lines represent the mean MD values. JOURNAL OF ORTHOPAEDIC RESEARCH MAY 2014
18 and SNP 11—SNP 18, while cis-phase interaction was revealed in SNP 10 when conditioning on SNP18. Although SNP 10 and SNP 18 are not in the same block, our result revealed the possibility that an interaction between the two blocks may contribute to the susceptibility to LDD in Chinese Han population. However, further functional study is required to confirm this finding. The ADAMTS-5 Association With Altered MD Values Numerous studies have addressed the association between genotypes and imaging analyses. By using tractography of DTI, Carballedo et al.39 demonstrated that the met-allele of the brain-derived neurotrophic factor (BDNF) polymorphism probably rendered subjects more vulnerabilities to dysfunctions of the uncinate fasciculus. Spalletta et al.40 found that patients with the GSTA1 B risk allele had an increased MD value in bilateral thalami. However, most of these reports were concentrated on brain related disease, no such investigation combining genetics and functional MRI has ever been performed in the field of LDD. Degenerative changes of lumbar IVDs in LDD have been proved to accompany with water content reduction and shrinkage of the disc volume.41 As a valuable and precise DTI quantitative parameter, MD was shown to be worthwhile in evaluating microstructural alterations in IVD during LDD. It was reported that the more degenerated discs harbor lower MD values.29 As has been demonstrated in our study, the A allele of rs162502 was the pathogenic allele associated with LDD, it is reasonable to believe that patients with A allele of rs162502 will experience more degenerative changes in IVDs. Furthermore, our results of the second-stage analysis was also consistent with the
LUMBAR DISC DEGENERATION
693
Table 4. Individual MD Values in Different Genotypes of Positive SNPs SNP_ID
Genotype
N
M SD (102)a
t-value
df
p-value
SNP 10
GG AG þ AA AA AG þ GG AA AG þ GG
37 13 28 22 19 31
834.02 57.36 829.37 83.92 834.93 59.25 830.97 69.47 807.23 54.34 847.30 65.88
0.22
48
0.829
0.26
48
0.854
2.138
48
0.038b
SNP 11 SNP 18
N stands for the number of different genotypes; M SD stands for mean standard deviation. aThe unit is m2/ms. bIndicates p < 0.05.
previous hypothesis, indicating that AA carriers in rs162502 showed a lower MD values (807.23 54.34 102 m2/ms) compared with Gþ carriers (847.30 65.88 102 m2/ms). On the other hand, the dose effect of A allele and the relative imbalance of the sample numbers may provide a bias to this result. Limitations There are several potential limitations in our study. First, as a pilot study, functions of the positive SNPs were not investigated. Second, the sample size of imaging analysis is rather small so that the type II errors cannot be overcomed. Third, given that our Chinese sample cannot be considered as all population representative, studies in other ethnics will be greatly encouraged. Summary For the first time, we reported the genetic polymorphisms of ADAMTS-5 (rs151058, rs229052, and rs162502) may be associated with susceptibility to LDD. A cis-phase interaction between rs151058 and rs229052 may exist in ADAMTS-5 with LDD. Furthermore, we firstly combined the genetic association study with diffusion tensor imaging technique in the field of orthopedics indicating that the rs162502 might be associated with altered MD values.
REFERENCES 1. Andersson GB. 1999. Epidemiological features of chronic low-back pain. Lancet 354:581–585. 2. Frymoyer JW. 1988. Back pain and sciatica. N Engl J Med 318:291–300. 3. Luoma K, Riihimaki H, Luukkonen R, et al. 2000. Low back pain in relation to lumbar disc degeneration. Spine (Phila Pa 1976) 25:487–492. 4. Heliovaara M, Impivaara O, Sievers K, et al. 1987. Lumbar disc syndrome in Finland. J Epidemiol Community Health 41:251–258. 5. Kalichman L, Hunter DJ. 2008. The genetics of intervertebral disc degeneration. Familial predisposition and heritability estimation. Joint Bone Spine 75:383–387. 6. Varlotta GP, Brown MD, Kelsey JL, et al. 1991. Familial predisposition for herniation of a lumbar disc in patients who are less than twenty-one years old. J Bone Joint Surg Am 73:124–128. 7. Matsui H, Terahata N, Tsuji H, et al. 1992. Familial predisposition and clustering for juvenile lumbar disc herniation. Spine (Phila Pa 1976) 17:1323–1328.
8. Ala-Kokko L. 2002. Genetic risk factors for lumbar disc disease. Ann Med 34:42–47. 9. Sambrook PN, MacGregor AJ, Spector TD. 1999. Genetic influences on cervical and lumbar disc degeneration: a magnetic resonance imaging study in twins. Arthritis Rheum 42:366–372. 10. Williams FM, Popham M, Hart DJ, et al. 2011. GDF5 singlenucleotide polymorphism rs143383 is associated with lumbar disc degeneration in Northern European women. Arthritis Rheum 63:708–712. 11. Le Maitre CL, Pockert A, Buttle DJ, et al. 2007. Matrix synthesis and degradation in human intervertebral disc degeneration. Biochem Soc Trans 35:652–655. 12. Kalb S, Martirosyan NL, Kalani MY, et al. 2012. Genetics of the degenerated intervertebral disc. World Neurosurg 77:491–501. 13. Roberts S, Caterson B, Menage J, et al. 2000. Matrix metalloproteinases and aggrecanase: their role in disorders of the human intervertebral disc. Spine (Phila Pa 1976) 25:3005–3013. 14. Le Maitre CL, Freemont AJ, Hoyland JA. 2004. Localization of degradative enzymes and their inhibitors in the degenerate human intervertebral disc. J Pathol 204:47–54. 15. Pockert AJ, Richardson SM, Le Maitre CL, et al. 2009. Modified expression of the ADAMTS enzymes and tissue inhibitor of metalloproteinases 3 during human intervertebral disc degeneration. Arthritis Rheumat 60:482–491. 16. Bachmeier BE, Nerlich A, Mittermaier N, et al. 2009. Matrix metalloproteinase expression levels suggest distinct enzyme roles during lumbar disc herniation and degeneration. Eur Spine J 18:1573–1586. 17. Glasson SS, Askew R, Sheppard B, et al. 2005. Deletion of active ADAMTS5 prevents cartilage degradation in a murine model of osteoarthritis. Nature 434:644–648. 18. Stanton H, Rogerson FM, East CJ, et al. 2005. ADAMTS5 is the major aggrecanase in mouse cartilage in vivo and in vitro. Nature 434:648–652. 19. Tortorella MD, Burn TC, Pratta MA, et al. 1999. Purification and cloning of aggrecanase-1: a member of the ADAMTS family of proteins. Science 284:1664–1666. 20. Somerville RP, Longpre JM, Jungers KA, et al. 2003. Characterization of ADAMTS-9 and ADAMTS-20 as a distinct ADAMTS subfamily related to Caenorhabditis elegans GON-1. J Biol Chem 278:9503–9513. 21. Liang QQ, Zhou Q, Zhang M, et al. 2008. Prolonged upright posture induces degenerative changes in intervertebral discs in rat lumbar spine. Spine (Phila Pa 1976) 33:2052–2058. 22. Patel KP, Sandy JD, Akeda K, et al. 2007. Aggrecanases and aggrecanase-generated fragments in the human intervertebral disc at early and advanced stages of disc degeneration. Spine (Phila Pa 1976) 32:2596–2603. 23. Le Maitre CL, Freemont AJ, Hoyland JA. 2007. Accelerated cellular senescence in degenerate intervertebral discs: a
JOURNAL OF ORTHOPAEDIC RESEARCH MAY 2014
694
24.
25.
26.
27.
28. 29.
30.
31.
32.
33.
WU ET AL. possible role in the pathogenesis of intervertebral disc degeneration. Arthritis Res Ther 9:R45. Zhao CQ, Zhang YH, Jiang SD, et al. 2011. ADAMTS-5 and Intervertebral disc degeneration: the results of tissue immunohistochemistry and in vitro cell culture. J Orthop Res 29: 718–725. Le Maitre CL, Freemont AJ, Hoyland JA. 2007. Accelerated cellular senescence in degenerate intervertebral discs: a possible role in the pathogenesis of intervertebral disc degeneration. Arthritis Res Therapy 9:R45. Virtanen IM, Noponen N, Barral S, et al. 2007. Putative susceptibility locus on chromosome 21q for lumbar disc disease (LDD) in the Finnish population. J Bone Miner Res 22:701–707. Carballido-Gamio J, Xu D, Newitt D, et al. 2007. Single-shot fast spin-echo diffusion tensor imaging of the lumbar spine at 1.5 and 3 T. Magn Reson Imaging 25:665–670. Haughton V. 2006. Imaging intervertebral disc degeneration. J Bone Joint Surg Am 88:15–20. Zhang Z, Chan Q, Anthony MP, et al. 2012. Age-related diffusion patterns in human lumbar intervertebral discs: a pilot study in asymptomatic subjects. Magn Reson Imaging 30:181–188. Basser PJ, Pierpaoli C. 1996. Microstructural and physiological features of tissues elucidated by quantitative-diffusiontensor MRI. J Magn Reson B 111:209–219. Gabriel SB, Schaffner SF, Nguyen H, et al. 2002. The structure of haplotype blocks in the human genome. Science 296:2225–2229. Takahashi M, Haro H, Wakabayashi Y, et al. 2001. The association of degeneration of the intervertebral disc with 5a/6a polymorphism in the promoter of the human matrix metalloproteinase-3 gene. J Bone Joint Surg Br 83:491–495. Sun ZM, Miao L, Zhang YG, et al. 2009. Association between the 1562 C/T polymorphism of matrix metalloproteinase-9 gene and lumbar disc disease in the young adult population in North China. Connect Tissue Res 50:181–185.
JOURNAL OF ORTHOPAEDIC RESEARCH MAY 2014
34. Nackley AG, Shabalina SA, Tchivileva IE, et al. 2006. Human catechol-O-methyltransferase haplotypes modulate protein expression by altering mRNA secondary structure. Science 314:1930–1933. 35. Cordell HJ. 2009. Detecting gene-gene interactions that underlie human diseases. Nat Rev Genet 10:392–404. 36. Zhang Y, Kent JJ, Olivier M, et al. 2013. A comprehensive analysis of adiponectin QTLs using SNP association, SNP cis-effects on peripheral blood gene expression and gene expression correlation identified novel metabolic syndrome (MetS) genes with potential role in carcinogenesis and systemic inflammation. BMC Med Genomics 6:14. 37. Zhang J, Chen Y, Zhang K, et al. 2010. A cis-phase interaction study of genetic variants within the MAOA gene in major depressive disorder. Biol Psychiatry 68:795–800. 38. Thumma BR, Matheson BA, Zhang D, et al. 2009. Identification of a Cis-acting regulatory polymorphism in a Eucalypt COBRA-like gene affecting cellulose content. Genetics 183: 1153–1164. 39. Carballedo A, Amico F, Ugwu I, et al. 2012. Reduced fractional anisotropy in the uncinate fasciculus in patients with major depression carrying the met-allele of the Val66Met brain-derived neurotrophic factor genotype. Am J Med Genet B Neuropsychiatr Genet 159B:537–548. 40. Spalletta G, Piras F, Gravina P, et al. 2012. Glutathione Stransferase alpha 1 risk polymorphism and increased bilateral thalamus mean diffusivity in schizophrenia. Psychiatry Res 203:180–183. 41. Hughes SP, Freemont AJ, Hukins DW, et al. 2012. The pathogenesis of degeneration of the intervertebral disc and emerging therapies in the management of back pain. J Bone Joint Surg Br 94:1298–1304.
SUPPORTING INFORMATION Additional supporting information may be found in the online version of this article at the publisher’s web-site.
