Google search
 
Gout and Hyperuricemia March 2017
Gout and Hyperuricemia. 2017; 4(1): 12-20
DOI: 10.3966/GH1703040103
Abstract  |  Full Text (HTML)  |  Full Text (PDF)
Genetic variants associated with tophi occurrence by a genome wide association study of 1888 gout patients
Changgui Li1,#,*, Chung-Jen Chen2,*, Wei-Ting Liao3,*, Ya-Ting Chan4, Can Wang1, Tien-Tsai Cheng5, Lingling Cui1, Xinde Li1, Shun-Jen Chang1,4,#, Yongyong Shi1,6,7,#
1Shandong Provincial Key Laboratory of Metabolic Disease, the Affiliated Hospital of Qingdao University, Qingdao 266003, P.R. China. 2Department of Internal Medicine, Kaohsiung Medical University Hospital, Kaohsiung Medical University, Kaohsiung city, Taiwan. 3Department of Biotechnology, College of Life Science, Kaohsiung Medical University, Kaohsiung, Taiwan. 4Department of Kinesiology, Health and Leisure Studies, National University of Kaohsiung, Taiwan. 5Division of Rheumatology, Allergy and Immunology, Department of Internal Medicine, Kaohsiung Chang Gung Memorial Hospital and Chang Gung University College of Medicine, Kaohsiung, Taiwan. 6Bio-X Institutes, Key Laboratory for the Genetics of Developmental and Neuropsychiatric Disorders (Ministry of Education), Shanghai Jiao Tong University, Shanghai 200030, P.R. China. 7Institute of Social Cognitive and Behavioral Sciences, Shanghai Jiao Tong University, Shanghai 200240, P. R. China. *These authors contributed equally to this study.

*Corresponding authors:
Dr. Changgui Li
Affiliated Hospital of Qingdao University.16 Jiangsu Road, Qingdao 266003, PR China.
E-mail: lichanggui@medmail.com.cn

Dr. Shun-Jen Chang
National University of Kaohsiung, Taiwan. No. 700, Kaohsiung University Rd, Nanzih District, 811, Kaohsiung, Taiwan.
E-mail: changsj1104@gmail.com

Dr. Yongyong Shi
Bio-X Institutes, Key Laboratory for the Genetics of Developmental and Neuropsychiatric Disorders (Ministry of Education), Shanghai Jiao Tong University, Shanghai 200030, P.R. China.
E-mail: shiyongyong5@gmail.com


Submitted on Apr. 27, 2017; accepted on May. 8, 2017.
©2017, Gout and Hyperuricemia. Published by Dong Fong Health Co. LTD in Taiwan. All right reserved.
Abstract
Objective:Gouty tophi are nodular mass deposits of monosodium urate (MSU) under the skin. Although environmental factors such as age and disease duration are considered as the major risks for gouty tophi, the genetic factors remain unclear. This study intended to explore the genetic risk components related to tophus occurrence using a genome-wide association study (GWAS).
Methods:Male gout patients were recruited with the tophus status recorded as a major outcome. Each tophaceous patient was matched to four non-tophus controls by their age and gouty disease duration. Affymetrix CHB SNP arrays were used to genotype the DNA samples and analyze associations with tophus occurrence.
Results:A total of 1888 male gout patients were included. Age and disease duration were the significant risk factors related to tophus occurrence. After matching by age and disease duration, 713 patients, composed of 145 gout patients with tophi and 568 without tophus, were included for further analysis. The GWAS results showed 12 SNPs that had significant associations with tophi occurrence (p<1×10-7), which were located in or near MSX2, CXCR5, PRKCE, MARCKS, PTPRD, DIRC3, TTLL7, KRT39, POLA2, PITX2, PITX1 and ADAM20 genes. By signaling pathway analysis, four SNPs near PKC-epsilon, MARCKS, PITX2, and MSX2 showed most significantly associations with each other, and were found to be associated with the process of tophi formation.
Conclusion:We concluded PKC-epsilon, MARCKS, Pitx2, and MSX2 were strongly associated with tophi occurrence via a potential role in the phagocytosis of MSU, NETosis, osteoblast retraction, and fibroblast formation of tophus, respectively.

Key words:Tophi; GWAS; Gout; PRKCE; MARCKS; PITX2; MSX2
Abbreviation
Gene symbol Full name
MSX2 Muscle segment homeobox 2
CXCR5 Chemokine, cxc motif, receptor 5
PRKCE Protein kinase c, epsilon
MARCKS Myristoylated alanine-rich protein kinase c substrate
PTPRD Protein-tyrosine phosphatase, receptor-type, delta
DIRC3 Disrupted in renal carcinoma 3
TTLL7 Tubulin tyrosine ligase-like family member 7
KRT39 Keratin 39
POLA2 Polymerase (DNA directed), alpha 2
PITX2 Paired-like homeodomain transcription factor 2
PITX1 Paired-like homeodomain transcription factor 1
ADAM20 A disintegrin and metalloproteinase domain 20

Introduction

    Gout is characterized by the deposition of monosodium urate (MSU) crystals in the joints or soft tissue, and has four disease phases including asymptomatic hyperuricemia, acute gouty arthritis, intercritical gout and chronic tophaceous gout. The acute phase involves the activation of vascular endothelial cells, increased blood flow, increased permeability to plasma proteins, recruitment of leucocytes into the tissue [1], and the secretion of cytokines to promote secondary neutrophils capture by endothelial cells under physiologic flow in response to MSU crystal uptake [2-6]. Meanwhile, after the monocytes differentiate to macrophages, the differentiated macrophages play an anti-inflammatory role in terminating an acute attack by ingesting MSU crystals without proinflammatory cytokine secretion or endothelial cell activation [2,7], promoting the production of cytokines or chemokines to inhibit the responsiveness of the endothelial cells to IL-1β and TNF-α, and suppressing proinflammatory cytokine production [1,8,9].
    Even in the absence of clinical intervention, acute gouty arthritis usually resolves within a few days, meaning that an anti-inflammatory reaction will be invoked after the acute phase. However, MSU crystals can persist in joints following acute inflammation and have been identified within mononuclear cells in synovial fluid [10-12]. Gouty tophi are the landmarks of chronic tophaceous gout. They are composed of mononuclear and multinucleated macrophages surrounding a core of MSU crystals and encapsulated by connective tissue [13]. In 2000, Schweyer et al. found that perivascular localized mononuclear cells are freshly migrated monocytes and/or macrophages. They furthermore revealed that macrophages co-express tumor necrosis factor (TNF)-alpha and matrix metalloproteinases (MMPs) 2 and 9 in gouty tophi [14]. In 2010, Dalbeth et al. also revealed that numerous CD68+ mononuclear and multinucleated cells were present within the corona zone of tophi [15]. However, the development of the tophus is a dynamic process with a low-level continuous recruitment, proinflammatory activation, maturation and turnover of monocyte-macrophages. Therefore, if a down-regulation of inflammatory signals was not terminated completely by the differentiated macrophages, the inflammation process and neutrophil infiltration will continue. Thus, the tophus, i.e. macrophages, debris, and MSU crystals surrounded by dense connective tissue, may be expected to be observed in the chronic tophaceous gout phases.
    The prevalence of tophus in gout patients is varied, from 3.9-19.4% in US adults [16,17], 12.7-21.1% in Chinese [18], and 30.5% in Spanish patients [19]. The mean follow-up time was 47 months for tophus occurrence, and more importantly, the presence of tophi was a significant mortality risk after adjustment for baseline serum urate levels by applied survival analysis methodology [19].
    The factors causing gout tophus are relatively complex and not as well understood compared to acute gouty arthritis. A previous study revealed transforming growth factor β1 (TGF-β1) gene was associated with tophi occurrence [20]. TGF-β1 has been well known to be related to higher serum uric acid level and disease duration, but is not the only factor responsible for the development of tophi. Therefore, other factors that contribute to the development of tophi are worthy of further investigation. In this study, we use a genome-wide association study to discover risk genetic variants contributing to the development of tophi by matching the patients with tophus to those without tophus with same age and disease duration.

Methods

Participant selection
    A total of 1888 male gout patients were included in the study from the Affiliated Hospital of Qingdao University, Qingdao, China. We matched each gout patient who had tophus to four non-tophus gout controls with the same age and disease duration of gout, which are two of the most important risk factors for tophus occurrence. A total of 713 patients were included, which composed of 145 gout patients with tophi greater than one granule and 568 gout patients without any tophus.
    All gout patients analyzed in the study were interviewed by an endocrinologist, diagnosed according to the American College of Rheumatology criteria for gout, and were recruited using the same diagnostic criteria. The number of tophus was estimated from the legs and hands. Collection of samples and clinical information from subjects were undertaken with informed consent and relevant ethical review board approval in accordance to the tenets of the Declaration of Helsinki.

Quality control of GWAS data set
    Using an Affymetrix genome-wide human SNP array to identify all participants’ genetic information (Axiom GW-Hu-CHB-SNP), a total of 25 samples selected randomly from the same population were designed for duplication to test the quality of GWAS. A dish quality control (DQC) value greater than 0.82 was set as the primary quality-control step. The member of the duplicate pair with the lower quality was excluded from the analysis. Genotype data were generated with Axiom Genotyping Algorithm v1 (Axiom GT1). For sample filtering, arrays with generated genotypes for <95% of the loci were excluded (n=40). Heterozygosity rates were calculated, deviations of more than 6 standard deviation (SD) from the mean were excluded and no samples were excluded. PLINK’s identity by descent analysis was used to detect hidden relatedness. When pairs of individuals had PI_HAT >0.25, the member of the pair with the lower call rate was excluded from the analysis (n=92). A total of 1888 gout patients were retained for further analysis. For SNP filtering (after sample filtering), SNPs with call rates <95%, minor allele frequency (MAF) <3% (n=12,267) or SNPs that deviated significantly (p<1×10−5) from Hardy-Weinberg equilibrium in non-tophus patients (n=15,707) were removed. A total of 603,697 SNPs passed the quality criteria and were used in subsequent analyses.

Statistical analysis
    Chi-squared tests were applied to analyze the associations between SNPs and tophus occurrence. Two-sample t-tests estimated the mean difference of demographic data between gout patients with and without tophus. An ANOVA test was applied to test the significant mean difference between genotypes and continuous data such as age, uric acid, creatinine and body mass index (BMI). A Kruskal-Wallis test was applied to estimate the significance between three genotypes and tophus number. An odds ratio (OR) test was used to test the association between single SNP and tophaceous occurrence by using SAS statistical program. A Manhattan plot of log10 (1/P values) was generated by SNPEVG program [21]. A p-value less than 1×10-7 was considered as a significant association in GWAS. In SNP-associated analysis the p-values were corrected by 0.05 times the number of multiple comparisons for Bonferroni method.

Results

    A total of 1888 male gout patients were included in this study, 18.80% (n=355) of which had been diagnosed with tophus (Table 1). The mean age of gout patients with tophus (54.05 ± 12.93 years) was significantly higher than those gout patients without tophus (50.96 ± 13.33 years, p≤0.001). Furthermore, the disease duration of gout also showed a significant difference between those with tophus (9.42 ± 7.03 years) and without tophus (4.99 ± 5.50 years) (p<0.001). Biochemical markers such as uric acid, blood urea nitrogen (BUN), Creatinine, BMI and hip line also displayed significant differences between these two populations (all p<0.05). However, both age and disease duration were the major demographic risk factors for tophus occurrence, thus we matched each gout patient who has tophus with up to four non-tophus controls with the same age and disease duration. A total of 145 gout patients with tophus and 568 patients without tophus were included. The mean ages of those with tophus and without tophus were 51.17 years (±13.26) and 52.54 years (±12.87), respectively (p=0.253). Furthermore, the mean disease duration also did not show any significant difference (45.73 ± 13.46 vs 47.39 ± 13.06 respectively, p=0.175). Besides, the biochemical markers listed in table 1 also did not show significant difference (all p-values >0.05).
Table 1. The associations between demographic data and tophus occurrence before and after matching.
Variables All participants (n=1888) Before matching (n=1888) After matching$ (n=713)
Tophus number p-values Tophus number p-values
≥1 (%) 0 (%) ≥1 0
Sample size   355 (18.80) 1533 (81.20)   145 (20.34) 568 (79.66)  
Age# (years) 51.72 ± 13.35 54.05 ± 12.93 50.96 ± 13.33 <0.001 51.17 ± 13.26 52.54 ± 12.87 0.253
Age onset# (years) 45.94 ± 13.01 45.49 ± 13.79 46.04 ± 12.82 0.483 45.73 ± 13.46 47.39 ± 13.06 0.175
Disease duration#(years) 5.82 ± 6.07 9.42 ± 7.03 4.99 ± 5.50 <0.001 5.43 ± 2.73 5.14 ± 2.77 0.259
Uric acid (µmol/L) 465.70 ± 111.70 497.60 ± 119.40 458.30 ± 108.50 <0.001 478.30 ± 119.90 464.00 ± 109.60 0.169
BUN (mmol/L) 5.63 ± 3.30 6.03 ± 2.86 5.53 ± 3.39 0.005 5.85 ± 2.86 5.70 ± 3.30 0.629
Creatinine (μmol/L) 88.31 ± 39.84 92.91 ± 38.83 87.30 ± 40.02 0.020 89.54 ± 34.90 89.02 ± 40.13 0.890
BMI (kg/m2) 27.16 ± 4.49 26.58 ± 3.28 27.29 ± 4.71 <0.001 27.03 ± 3.23 27.29 ± 3.52 0.428
Waist line (cm) 97.28 ± 9.25 96.75 ± 8.31 97.41 ± 9.46 0.223 97.36 ± 7.87 97.72 ± 10.61 0.676
Hip line (cm) 104.98 ± 7.88 104.20 ± 6.80 105.20 ± 8.12 0.039 104.70 ± 6.61 105.70 ± 8.49 0.560
Sugar (mmol/L) 5.97 ± 1.24 6.01 ± 1.20 5.97 ± 1.25 0.531 5.92 ± 1.06 5.98 ± 1.25 0.597
Triglycerides (mmol/L) 2.47 ± 3.03 2.35 ± 2.15 2.50 ± 3.20 0.284 2.69 ± 2.90 2.62 ± 4.63 0.866
Cholesterol (mmol/L) 5.22 ± 1.15 5.18 ± 1.08 5.23 ± 1.17 0.416 5.23 ± 0.96 5.32 ± 1.16 0.383
$: Each gout patient with tophus was matched with four controls by age and gout disease duration.
#: mean ± standard deviation; BUN: blood urea nitrogen; BMI: body mass index.

    A total of 22 SNPs showed significant associations with tophus occurrence and are listed in Figure 1 (all p-values <1×10-7). After removing six SNPs that did not have assigned reference SNP ID numbers and four SNPs that showed linkage disequilibrium with other markers, a total of 12 SNPs were included for further analysis. Table 2 lists the 12 SNPs’ locations, allele information and gene names (in or near genes), such as MSX2, CXCR5, PRKCE, MARCKs, PTPRD, DIRC3, TTLL7, KRT39, POLA2, PITX2, PITX1 and ADAM20.
figure 1
Figure 1. The Manhattan plot of GWAS in a comparison of tophaceous occurrence between tophus patients and non-tophus controls after matching with age and disease duration among gout patients.
The number indicates similarly to Table 2-4.
NR: without a reference SNP ID number.
Table 2. The basic data of SNPs which were significantly associated with tophus occurrence in matched GWAS.
No. SNP Chromosome Physical position Gene names Alleles p-values (*10-8) Minor-allele/frequency in non-tophus patients Location
1 rs1427963 5 173744624 MSX2 A/G 0.030 G/0.230 Downstream
2 rs12224530 11 118691473 CXCR5 A/G 0.049 G/0.280 Upstream
3 rs513386 2 45905193 PRKCE A/T 0.200 T/0.192 Intron
4 rs73545414 6 113154528 MARCKS T/G 0.384 G/0.259 Upstream
5 rs7020371 9 8923715 PTPRD T/C 2.584 C/0.259 Intron
6 rs2252488 2 217857450 DIRC3 T/C 3.260 C/0.270 Downstream
7 rs7537839 1 84219339 TTLL7 A/G 3.374 G/0.291 Downstream
8 rs17843021 17 39116728 KRT39 A/G 3.946 G/0.297 Missense
9 rs673147 11 65060427 POLA2 A/G 4.554 G/0.310 Intron
10 rs7660246 4 111533065 PITX2 A/G 4.690 G/0.334 Intron
11 rs6859913 5 134412802 PITX1 A/G 5.472 G/0.329 Intron
12 rs8009032 14 70985876 ADAM20 T/G 5.700 G/0.194 Downstream
13 rs12656362# 5 173722032 MSX2 A/G 0.162 G/0.221 Downstream
14 rs473298# 11 118679387 CXCR5 A/G 1.824 G/0.363 Upstream
15 rs13161216# 5 173793409 MSX2 C/G 8.41 G/0.147 Intron
16 rs1378689# 6 113157184 MARCKS A/G 9.17 G/0.153 Upstream
#: the SNP showed linkage disequilibrium with other polymorphisms.

    Table 3 lists the odds ratio of tophus occurrence related to genotypes of 12 significant SNPs in matched GWAS. All 12 SNPs showed significant associations with tophus occurrence (p<0.05 after Bonferroni correction); the ORs of both rs17843021 (KRT39) and rs7660246 (PITX2) showed recessive models in relation to tophus occurrence and the other markers were in dominant models.
Table 3. The odds ratios of tophus occurrence related to 12 significant SNPs in matched GWAS.
No. Polymorphism (genes name) Sample size. (%) Tophi number (%)   OR (95% CI)$
≥1 0
1 rs1427963 ( MSX2)        
  GG 392 (55.29) 54 (37.76) 338 (59.72)   1.0
  AG 254 (35.83) 58 (40.56) 196 (34.63)   1.85 (1.23-2.79)
  AA 63 (8.89) 31 (21.68) 32 (5.65)   6.06 (3.42-10.74)***
2 rs12224530 (CXCR5)        
  AA 350 (49.16) 48 (33.10) 302 (53.26)   1.0
  AG 267 (37.50) 55 (37.93) 212 (37.39)   1.63 (1.07-2.50)
  GG 95 (13.34) 42 (28.97) 53 (9.35)   4.99 (3.00-8.28)***
3 rs513386 (PRKCE)        
  AA 429 (60.76) 56 (39.16) 373 (66.25)   1.0
  AT 232 (32.86) 68 (47.55) 164 (29.13)   2.76 (1.85-4.11)***
  TT 45 (6.37) 19 (13.29) 26 ( 4.62)   4.87 (2.53-9.37)***
4 rs73545414 (MARCKS)        
  GG 376 (52.88) 45 (31.03) 331 (58.48)   1.0
  TG 245 (34.46) 68 (46.90) 177 (31.27)   2.83 (1.86-4.29)***
  TT 90 (12.66) 32 (22.07) 58 (10.25)   4.06 (2.38-6.91)***
5 rs7020371 (PTPRD)        
  TT 359 (52.79) 47 (33.81) 312 (57.67)   1.0
  CT 239 (35.15) 61 (43.88) 178 (32.90)   2.27 (1.49-3.47)***
  CC 82 (12.06) 31 (22.30) 51 (9.43)   4.04 (2.34-6.94)***
6 rs2252488 (DIRC3)        
  CC 366 (51.48) 43 (29.86) 323 (56.97)   1.0
  CT 265 (37.27) 83 (57.64) 182 (32.10)   3.43 (2.27-5.17)***
  TT 80 (11.25) 18 (12.50) 62 (10.93)   2.18 (1.18-4.03)
7 rs7537839 (TTLL7)        
  AA 337 (47.40) 39 (26.90) 298 (52.65)   1.0
  AG 280 (39.38) 73 (50.34) 207 (36.57)   2.69 (1.76-4.13)***
  GG 94 (13.22) 33 (22.76) 61 (10.78)   4.13 (2.41-7.09)***
8 rs17843021 (KRT39)        
  GG 331 (46.88) 52 (36.11) 279 (49.64)   1.0
  AG 286 (40.51) 54 (37.50) 232 (41.28)   1.25 (0.82-1.90)
  AA 89 (12.61) 38 (26.39) 51 (9.07)   4.00 (2.39-6.68)***
9 rs673147 (POLA2)        
  GG 328 (46.26) 37 (25.69) 291 (51.50)   1.0
  AG 272 (38.36) 74 (51.39) 198 (35.04)   2.94 (1.90-4.54)***
  AA 109 (15.37) 33 (22.92) 76 (13.45)   3.42 (2.00-5.82)***
10 rs7660246 (PITX2)        
  GG 300 (42.37) 48 (33.10) 252 (44.76)   1.0
  AG 300 (42.37) 54 (37.24) 246 (43.69)   1.15 (0.75-1.77)
  AA 108 (15.25) 43 (29.66) 65 (11.55)   3.47 (2.12-5.69)***
11 rs6859913 (PITX1)        
  GG 366 (52.59) 102 (71.83) 264 (47.65)   1.0
  AG 248 (35.63) 33 (23.24) 215 (38.81)   0.40 (0.26-0.61)***
  AA 82 (11.78) 7 (4.93) 75 (13.54)   0.24 (0.11-0.54)***
12 rs8009032 (ADAM20)        
  GG 417 (60.61) 56 (40.00) 361 (65.88)   1.0
  GT 234 (34.01) 73 (52.14) 161 (29.38)   2.92 (1.97-4.34)***
  TT 37 (5.38) 11 (7.86) 26 (4.74)   2.73 (1.28-5.83)
$: the *** indicates that p-values are less than 0.05 after correction by Bonferroni method (crude p-value times 24).

    We further estimated the associations in mean age, BMI, uric acid, creatinine and tophus number distributed in different genotypes of 12 significant SNPs (Table 4). The results showed that only the mean tophus number showed significant association in different genotypes of all SNPs, while the mean age, BMI, uric acid and creatinine did not show any significant association between these two groups in the listed SNPs after Bonferroni correction (crude p-values times 12). After identifying these 12 SNPs significantly associated with gouty tophus occurrence, we further analyzed the possible mechanistic linkage among these genes. By using MetaCore bioinformatics tool (Build No.61585), we identified PKC-based mechanistic network as the pathway most relational to each other. In this pathway, evidence points to PKC (PKC-epsilon) directly regulating MARCKS, MSX2 and PITX2, and indirectly regulating CXCR5, PTPRD, TTLL7, PITX1, and DNA polymerase alpha (Figure 2).
Table 4. The means of age, BMI, uric acid, creatinine as well as tophus number distributed in different genotypes of significant SNPs in matched GWAS.
No. Polymorphism (Gene names) Sample size. (%) Age$ (years) Uric acid$ (μmol/L) Tophi number$, # Creatinine$ (μmol/L) BMI$ (kg/m2)
1 rs1427963 (MSX2)            
  GG 392 (55.29) 52.18 ± 12.95 464.90 ± 110.81 1.14 ± 0.35 85.89 ± 31.62 27.02 ± 3.43
  AG 254 (35.83) 52.33 ± 13.00 468.38 ± 111.23 1.22 ± 0.42 94.28 ± 47.65 27.45 ± 3.48
  AA 63 (8.89) 52.83 ± 13.20 478.45 ± 119.57 1.49 ± 0.50 90.40 ± 44.84 27.79 ± 3.53
      P=0.933 P=0.667 P<0.001 P=0.035* P=0.123
2 rs12224530 (CXCR5)            
  AA 350 (49.16) 51.81 ± 13.07 462.02 ± 112.08 1.14 ± 0.34 89.30 ± 44.63 27.06 ± 3.33
  AG 267 (37.50) 52.30 ± 12.44 470.29 ± 109.72 1.21 ± 0.41 89.01 ± 34.59 27.50 ± 3.68
  GG 95 (13.34) 53.84 ± 13.98 474.67 ± 117.33 1.44 ± 0.50 88.44 ± 26.78 27.19 ± 3.28
      P=0.400 P=0.507 P<0.001 P=0.983 P=0.297
3 rs513386 (PRKCE)            
  AA 429 (60.76) 52.58 ± 13.10 461.72 ± 112.41 1.13 ± 0.34 88.90 ± 42.27 27.27 ± 3.39
  AT 232 (32.86) 51.54 ± 12.54 476.50 ± 107.57 1.29 ± 0.46 90.23 ± 35.68 27.26 ± 3.53
  TT 45 (6.37) 52.84 ± 14.35 456.80 ± 128.86 1.42 ± 0.50 85.07 ± 23.80 26.76 ± 3.87
      P=0.586 P=0.230 P<0.001 P=0.730 P=0.640
4 rs73545414 (MARCKS)            
  GG 376 (52.88) 52.91 ± 13.69 456.79 ± 114.21 1.11 ± 0.33 90.39 ± 46.55 26.87 ± 3.43
  TG 245 (34.46) 51.07 ± 12.03 478.44 ± 103.86 1.28 ± 0.45 86.93 ± 30.44 27.76 ± 3.57
  TT 90 (12.66) 53.17 ± 11.88 475.87 ± 119.26 1.36 ± 0.48 89.19 ± 21.45 27.44 ± 3.09
      P=0.178 P=0.045* P<0.001 P=0.580 P=0.007*
5 rs7020371 (PTPRD)            
  TT 359 (52.79) 51.55 ± 12.71 463.72 ± 106.94 1.13 ± 0.34 88.28 ± 43.40 27.34 ± 3.32
  CT 239 (35.15) 52.41 ± 13.40 472.15 ± 116.05 1.26 ± 0.44 88.09 ± 28.51 27.32 ± 3.77
  CC 82 (12.06) 54.02 ± 12.92 479.52 ± 120.22 1.38 ± 0.49 94.35 ± 45.24 26.78 ± 3.27
      P=0.279 P=0.430 P<0.001 P=0.453 P=0.405
6 rs2252488 (DIRC3)            
  CC 366 (51.48) 52.45 ± 13.18 458.10 ± 114.89 1.12 ± 0.32 89.32 ± 45.85 27.24 ± 3.48
  CT 265 (37.27) 52.19 ± 12.66 474.96 ± 106.90 1.31 ± 0.46 88.41 ± 27.90 27.28 ± 3.48
  TT 80 (11.25) 52.14 ± 12.79 479.47 ± 112.50 1.23 ± 0.42 90.94 ± 37.05 27.10 ± 3.37
      P=0.961 P=0.100 P<0.001 P=0.882 P=0.925
7 rs7537839 (TTLL7)            
  AA 337 (47.40) 52.08 ± 12.52 465.62 ± 112.08 1.12 ± 0.32 89.43 ± 39.15 27.18 ± 3.41
  AG 280 (39.38) 52.00 ± 13.35 460.75 ± 107.68 1.26 ± 0.44 85.31 ± 21.17 27.28 ± 3.63
  GG 94 (13.22) 53.65 ± 13.38 490.24 ± 121.82 1.35 ± 0.48 93.45 ± 43.02 27.31 ± 3.16
      P=0.533 P=0.085 P<0.001 P=0.113 P=0.917
8 rs17843021 (KRT39)            
  GG 331 (46.88) 52.00 ± 13.27 462.95 ± 117.04 1.16 ± 0.36 87.31 ± 39.07 27.24 ± 3.48
  AG 286 (40.51) 52.40 ± 12.63 471.58 ± 105.34 1.19 ± 0.39 90.48 ± 38.88 27.19 ± 3.53
  AA 89 (12.61) 52.92 ± 12.99 467.31 ± 114.06 1.43 ± 0.50 92.11 ± 41.22 27.49 ± 3.25
      P=0.821 P=0.636 P<0.001 P=0.473 P=0.777
9 rs673147 (POLA2)            
  GG 328 (46.26) 51.97 ± 13.19 463.56 ± 108.58 1.11 ± 0.32 88.37 ± 44.43 27.06 ± 3.08
  AG 272 (38.36) 52.25 ± 12.34 469.37 ± 115.37 1.27 ± 0.46 89.15 ± 30.21 27.53 ± 3.98
  AA 109 (15.37) 52.81 ± 13.62 473.45 ± 113.26 1.30 ± 0.46 92.03 ± 41.24 27.07 ± 3.15
      P=0.840 P=0.678 P<0.001 P=0.718 P=0.214
10 rs7660246 (PITX2)            
  GG 300 (42.37) 52.43 ± 12.80 462.93 ± 111.36 1.16 ± 0.37 89.07 ± 44.54 27.06 ± 3.00
  AG 300 (42.37) 51.85 ± 13.08 465.64 ± 110.01 1.18 ± 0.38 87.33 ± 35.44 27.50 ± 3.91
  AA 108 (15.25) 53.37 ± 13.17 482.29 ± 117.72 1.40 ± 0.49 94.45 ± 32.21 27.10 ± 3.38
      P=0.573 P=0.296 P<0.001 P=0.297 P=0.267
11 rs6859913 (PITX1)            
  GG 366 (52.59) 52.69 ± 12.56 471.70 ± 112.95 1.28 ± 0.45 87.50 ± 25.94 27.32 ± 3.51
  AG 248 (35.63) 51.31 ± 13.24 467.10 ± 113.95 1.13 ± 0.34 92.86 ± 55.82 27.23 ± 3.05
  AA 82 (11.78) 52.46 ± 14.29 439.86 ± 99.11 1.09 ± 0.28 83.84 ± 23.49 27.11 ± 4.43
      P=0.428 P=0.067 P<0.001 P=0.122 P=0.865
12 rs8009032 (ADAM20)            
  GG 417 (60.61) 52.07 ± 12.97 468.03 ± 109.33 1.13 ± 0.34 88.26 ± 39.60 27.23 ± 3.44
  GT 234 (34.01) 52.65 ± 13.06 462.54 ± 120.04 1.31 ± 0.46 87.97 ± 35.26 27.18 ± 3.57
  TT 37 (5.38) 51.14 ± 11.96 480.14 ± 96.47 1.30 ± 0.46 99.80 ± 54.84 27.73 ± 3.55
      P=0.751 P=0.637 P<0.001 P=0.224 P=0.674
#: The p-values were estimated by Kruskal-Wallis test.
*: The crude p-values were significant (p<0.05), but there were no longer significant after p-values were corrected by Bonferroni method (crude p-value times 24).
$: The values express as mean ± standard deviation.
figure 2
Figure 2. The interactions between the 10 SNPs analyzed by MetaCore database, which revealed PKC, MSX2, MARCKS and PITX2 genes to directly interact. DIRC3 and KRT39 were not included in MetaCore.

Discussion

    The GWAS study revealed 22 SNPs which were significantly associated with susceptibility to tophi formation in male gout patients. Furthermore, using MetaCore analysis, four SNPs near or in the genes of PRKCE, MARCKS [22-28], MSX2 [29] and PITX2 [30] were found to be linked and function together. This GWAS study provides robust results by with matching age and disease duration between tophaceous and control cohorts, and with the restriction of only including male subjects; moreover, this study also used a large sample size. However, this study did not match the uric acid variable even uric acid is the major contributor to develop gout, but the result showed uric acid did not have a significant association between gout patients and controls after matching. All aspects of the study design can ensure a complete exploration of the genetic risk components attributing to tophus occurrence while avoiding confounding effects resulting from varying age, disease duration, gender and insufficient sample size.
    The development of tophus is a dynamic process with hyperuricemia as well as a low-level continuous recruitment of leukocytes, proinflammatory activation, maturation and turnover of monocyte-macrophages. Chronic tophaceous gout usually develops after years of acute intermittent gout. However, asymptomatic hyperuricemia usually precedes the onset of acute gouty attack. There are two mechanisms that drive the phagocytosis of MSU that leads to hyperuricemia. The first route, in which MSU binds IgG and induces phagocytosis in cells (such as macrophages) with Fc-gamma receptors is regarded as the key player in gouty inflammation. The second mechanism involves direct contact of MSU crystal with lipid membrane and proteins in cells. Larsen et al. demonstrated that protein kinase C (PKC)-epsilon in phagosomes can enhance the rate of IgG-dependent phagocytosis [31], which is the major route of MSU phagocytosis. We believe this is the major route for MSU phagocytosis.
    In addition, another study revealed that intracellular interaction of PKC-epsilon with IL-32α and STAT3 promotes STAT3 binding to the IL-6 promoter, resulting in increased production of IL-6, a prominent pro-inflammatory cytokine [32]. Moreover, PKC-epsilon was found to be a critical component of TLR-4 signaling pathway and thereby to play a key role in macrophage and dendritic cell activation in response to LPS. However, controlling the kinase activity of PKC-epsilon might represent an efficient strategy to prevent or treat certain inflammatory disorders (of microbial origin) [33]. The aforementioned findings support our GWAS finding that PKC-epsilon variant is potentially an independent and key risk factor for the development of tophus in relation to phagocytosis.
    The tophi structure is composed of three zones: a microscopically-central core, corona zone and fibrovascular zone [14,34,35]. Tophi appear as chronic granulomatous lesions comprised of collections of mononucleated and multinucleated macrophages (cellular corona zone) surrounding a core of MSU crystals (central crystalline core) and encased by dense connective tissue (fibrovascular zone). Moreover, mast cells are present throughout both the corona and fibrovascular zones at similar densities. In contrast, neutrophils are often observed in the corona zone [15].
    The role of MARCKS encoded by the MARCKS gene acts as a substrate for protein kinase C, which has never been reported in association with either gout or tophus. Green et al. investigated the role of MARCKS protein in directing the migration of macrophages toward a chemoattractant in mice [36], and concluded that MARCKS is involved in directed migration of macrophages via a process involving its phosphorylation, cytoplasmic translocation, and interaction with actin. [36] Eckert et al. reported that treatment of isolated human neutrophils with myristoylated aminoterminus significantly inhibited both their migration and beta2 integrin-dependent adhesion [37] and suggested that MARCKS is a key regulator of neutrophil migration and adhesion. Recently, it has been proposed that the ejection of chromatin from neutrophils into the extracelluar space to immobilize and kill pathogens has been described as neutrophil extracellular traps (NETs). Schuer et al. reported that gouty tophi are composed of extended areas of extracellular DNA co-localizing with materials from neutrophil granules [38], which is typical of neutrophils that have undergone NETosis. This finding implies that neutrophils are the key cells in the core area of tophi and neutrophils undergoing NETosis is the key mechanism for tophi development. However, under high neutrophil densities, these NETs aggregate and degrade both cytokines and chemokines to promote the resolution of neutrophilic inflammation. Furthermore, Schauer et al. have revealed that tophi share characteristics with aggregated NETs [38], so we proposed that MARCKS might play a key role in aggregating NETs and contributing to the development of tophi in relation to neutrophilic aggregation.
    Besides, the development of tophi is usually accompanied with erosive bone destruction. Dallben et al. revealed a strong relationship between bone erosion and the presence of intraosseous tophus [34], and concluded that tophus infiltration into bone is the dominant mechanism for the development of bone erosion and joint damage in gout. However, the underlying basis or molecular component of bone erosion in tophaceous gout remains unclear. In 2011, MSU crystals were revealed to have profound inhibitory effects on osteoblast viability and differentiation [39]; bone erosion in tophaceous gout occurs with both excessive osteoclast formation and osteoblast retraction induced by adherent neutrophils promoting osteoclast bone resorption [40]. Furthermore, Hayashi et al. found that that Pitx2 suppresses osteogenic signals induced by bone morphogenetic proteins (BMPs) in myoblasts to prevent their osteoblastic conversion [41]. Here, we revealed that variant PITX2 is associated with tophi formation and propose that variant PITX2 might inhibit the differentiation of osteoblasts, contributing to the development of bone erosion, a characteristic manifestation in cases with tophi.
    Most tophi are located in subcutaneous areas. Our finding that MSX2 variant is associated with tophi formation may provide some insight. The rationale is supported by findings from Yeh et al, who revealed that keratinocytes of MSX2 in null mice exhibited increased cell migration and the fibroblasts showed stronger collagen gel contraction [42]. The above findings suggest that MSX2 variant may have aberrant presentation of both fibroblasts and keratinocytes to contribute towards the development of cutaneous tophi.
    In conclusion, our study here presents the first matched genome-wide association analyses for molecular association related to tophus occurrence. Using signaling pathway analysis, four loci in or near PKC-epsilon, MARCKS, PITX2, and MSX2 were found to be closely associated with each other in the process of tophi formation. Based on previous findings, we propose that PKC-epsilon might be related to macrophage mediated phagocytosis of MSU, MARCKS might be related to NETosis in the tophus core, PITX2 might contribute to osteoblast retraction, and MSX2 might be associated with the formation of cutaneous tophus [42]; all of the functions can explain the mechanism of tophus formation. However, whether these risk variants of genes can prospectively predict the development of tophi warrants further study.

Acknowledgements

    This work was supported by the 973 Program (2010CB534902), the National Science Foundation of China (31371272, 31471195).

References

  1. Dalbeth N, Haskard DO. Mechanisms of inflammation in gout. Rheumatology (Oxford). 2005; 44: 1090-6.
  2. Landis RC, Yagnik DR, Florey O, et al. Safe disposal of inflammatory monosodium urate monohydrate crystals by differentiated macrophages. Arthritis Rheum. 2002; 46: 3026-33.
  3. Di Giovine FS, Malawista SE, Nuki G, Duff GW. Interleukin 1 (IL 1) as a mediator of crystal arthritis. Stimulation of T cell and synovial fibroblast mitogenesis by urate crystal-induced IL 1. J Immunol. 1987; 138: 3213-8.
  4. di Giovine FS, Malawista SE, Thornton E, Duff GW. Urate crystals stimulate production of tumor necrosis factor alpha from human blood monocytes and synovial cells. Cytokine mRNA and protein kinetics, and cellular distribution. J Clin Invest. 1991; 87: 1375-81.
  5. Terkeltaub R, Zachariae C, Santoro D, Martin J, Peveri P, Matsushima K. Monocyte-derived neutrophil chemotactic factor/interleukin-8 is a potential mediator of crystal-induced inflammation. Arthritis Rheum. 1991; 34: 894-903.
  6. Chapman PT, Yarwood H, Harrison AA, et al. Endothelial activation in monosodium urate monohydrate crystal-induced inflammation: in vitro and in vivo studies on the roles of tumor necrosis factor alpha and interleukin-1. Arthritis Rheum. 1997; 40: 955-65.
  7. Yagnik DR, Hillyer P, Marshall D, et al. Noninflammatory phagocytosis of monosodium urate monohydrate crystals by mouse macrophages. Implications for the control of joint inflammation in gout. Arthritis Rheum. 2000; 43: 1779-89.
  8. Fadok VA, Bratton DL, Konowal A, Freed PW, Westcott JY, Henson PM. Macrophages that have ingested apoptotic cells in vitro inhibit proinflammatory cytokine production through autocrine/paracrine mechanisms involving TGF-beta, PGE2, and PAF. J Clin Invest. 1998; 101: 890-8.
  9. Yagnik DR, Evans BJ, Florey O, Mason JC, Landis RC, Haskard DO. Macrophage release of transforming growth factor beta1 during resolution of monosodium urate monohydrate crystal-induced inflammation. Arthritis Rheum. 2004; 50: 2273-80.
  10. Weinberger A, Schumacher HR, Agudelo CA. Urate crystals in asymptomatic metatarsophalangeal joints. Ann Intern Med. 1979; 91: 56-7.
  11. Pascual E, Jovani V. A quantitative study of the phagocytosis of urate crystals in the synovial fluid of asymptomatic joints of patients with gout. Br J Rheumatol. 1995; 34: 724-6.
  12. Pascual E, Batlle-Gualda E, Martinez A, Rosas J, Vela P. Synovial fluid analysis for diagnosis of intercritical gout. Ann Intern Med. 1999; 131: 756-9.
  13. Palmer DG, Highton J, Hessian PA. Development of the gout tophus. An hypothesis. Am J Clin Pathol. 1989; 91: 190-5.
  14. Schweyer S, Hemmerlein B, Radzun HJ, Fayyazi A. Continuous recruitment, co-expression of tumour necrosis factor-alpha and matrix metalloproteinases, and apoptosis of macrophages in gout tophi. Virchows Arch. 2000; 437: 534-9.
  15. Dalbeth N, Pool B, Gamble GD, et al. Cellular characterization of the gouty tophus: a quantitative analysis. Arthritis Rheum. 2010; 62: 1549-56.
  16. Liote F, Lancrenon S, Lanz S, et al. GOSPEL: prospective survey of gout in France. Part I: design and patient characteristics (n = 1003). Joint Bone Spine. 2012; 79: 464-70.
  17. Zhu Y, Pandya BJ, Choi HK. Prevalence of gout and hyperuricemia in the US general population: the National Health and Nutrition Examination Survey 2007-2008. Arthritis Rheum. 2011; 63: 3136-41.
  18. Yu KH, Luo SF. Younger age of onset of gout in Taiwan. Rheumatology (Oxford). 2003; 42: 166-70.
  19. Perez-Ruiz F, Martinez-Indart L, Carmona L, Herrero-Beites AM, Pijoan JI, Krishnan E. Tophaceous gout and high level of hyperuricaemia are both associated with increased risk of mortality in patients with gout. Ann Rheum Dis. 2014; 73: 177-82.
  20. Chang SJ, Chen CJ, Tsai FC, et al. Associations between gout tophus and polymorphisms 869T/C and -509C/T in transforming growth factor beta1 gene. Rheumatology (Oxford). 2008; 47: 617-21.
  21. Li J, German R, King J, et al. Recent trends in prostate cancer testing and incidence among men under age of 50. Cancer Epidemiology. 36: 122-7.
  22. Soh JW, Lee EH, Prywes R, Weinstein IB. Novel roles of specific isoforms of protein kinase C in activation of the c-fos serum response element. Mol Cell Biol. 1999; 19: 1313-24.
  23. Larsson C. Protein kinase C and the regulation of the actin cytoskeleton. Cell Signal. 2006; 18: 276-84.
  24. Hartwig JH, Thelen M, Rosen A, Janmey PA, Nairn AC, Aderem A. MARCKS is an actin filament crosslinking protein regulated by protein kinase C and calcium-calmodulin. Nature. 1992; 356: 618-22.
  25. Akita Y, Kawasaki H, Ohno S, Suzuki K, Kawashima S. Involvement of protein kinase C epsilon in thyrotropin-releasing hormone-stimulated phosphorylation of the myristoylated alanine-rich C kinase substrate in rat pituitary clonal cells. Electrophoresis. 2000; 21: 452-9.
  26. Fujise A, Mizuno K, Ueda Y, et al. Specificity of the high affinity interaction of protein kinase C with a physiological substrate, myristoylated alanine-rich protein kinase C substrate. J Biol Chem. 1994; 269: 31642-8.
  27. Suzuki Y, Zhang H, Saito N, Kojima I, Urano T, Mogami H. Glucagon-like peptide 1 activates protein kinase C through Ca2+-dependent activation of phospholipase C in insulin-secreting cells. J Biol Chem. 2006; 281: 28499-507.
  28. Heidkamp MC, Iyengar R, Szotek EL, Cribbs LL, Samarel AM. Protein kinase Cepsilon-dependent MARCKS phosphorylation in neonatal and adult rat ventricular myocytes. J Mol Cell Cardiol. 2007; 42: 422-31.
  29. Jeong HM, Jin YH, Choi YH, et al. PKC signaling inhibits osteogenic differentiation through the regulation of Msx2 function. Biochim Biophys Acta. 2012; 1823: 1225-32.
  30. Espinoza HM, Ganga M, Vadlamudi U, et al. Protein kinase C phosphorylation modulates N- and C-terminal regulatory activities of the PITX2 homeodomain protein. Biochemistry. 2005; 44: 3942-54.
  31. Larsen EC, Ueyama T, Brannock PM, et al. A role for PKC-epsilon in Fc gammaR-mediated phagocytosis by RAW 264.7 cells. J Cell Biol. 2002; 159: 939-44.
  32. Kang JW, Park YS, Lee DH, et al. Intracellular interaction of interleukin (IL)-32alpha with protein kinase Cepsilon (PKCepsilon ) and STAT3 protein augments IL-6 production in THP-1 promonocytic cells. J Biol Chem. 2012; 287: 35556-64.
  33. Aksoy E, Goldman M, Willems F. Protein kinase C epsilon: a new target to control inflammation and immune-mediated disorders. Int J Biochem Cell Biol. 2004; 36: 183-8.
  34. Dalbeth N, Clark B, Gregory K, et al. Mechanisms of bone erosion in gout: a quantitative analysis using plain radiography and computed tomography. Ann Rheum Dis. 2009; 68: 1290-5.
  35. Palmer DG, Hogg N, Denholm I, Allen CA, Highton J, Hessian PA. Comparison of phenotype expression by mononuclear phagocytes within subcutaneous gouty tophi and rheumatoid nodules. Rheumatol Int. 1987; 7: 187-93.
  36. Green TD, Park J, Yin Q, et al. Directed migration of mouse macrophages in vitro involves myristoylated alanine-rich C-kinase substrate (MARCKS) protein. J Leukoc Biol. 2012; 92: 633-9.
  37. Eckert RE, Neuder LE, Park J, Adler KB, Jones SL. Myristoylated alanine-rich C-kinase substrate (MARCKS) protein regulation of human neutrophil migration. Am J Respir Cell Mol Biol. 2010; 42: 586-94.
  38. Schauer C, Janko C, Munoz LE, et al. Aggregated neutrophil extracellular traps limit inflammation by degrading cytokines and chemokines. Nat Med. 2014; 20: 511-7.
  39. Chhana A, Callon KE, Pool B, et al. Monosodium urate monohydrate crystals inhibit osteoblast viability and function: implications for development of bone erosion in gout. Ann Rheum Dis. 2011; 70: 1684-91.
  40. Allaeys I, Rusu D, Picard S, Pouliot M, Borgeat P, Poubelle PE. Osteoblast retraction induced by adherent neutrophils promotes osteoclast bone resorption: implication for altered bone remodeling in chronic gout. Lab Invest. 2011; 91: 905-20.
  41. Hayashi M, Maeda S, Aburatani H, et al. Pitx2 prevents osteoblastic transdifferentiation of myoblasts by bone morphogenetic proteins. J Biol Chem. 2008; 283: 565-71.
  42. Yeh J, Green LM, Jiang TX, et al. Accelerated closure of skin wounds in mice deficient in the homeobox gene Msx2. Wound Repair Regen. 2009; 17: 639-48.