DNA methylation change of HIPK3 in Chinese rheumatoid arthritis and its effect on inflammation

Abstract

Introduction: Homeodomain-interacting protein kinase 3 (HIPK3) plays an important role in cell proliferation, apoptosis, and inflammation. Over-expression of HIPK3 in immune cells in rheumatoid arthritis (RA) has been reported. In this study, we investigated blood methylation levels and clinical characteristics of RA in a Chinese population.

Methods: A total of 235 patients with RA, 30 with osteoarthritis (OA), and 30 matched healthy controls were recruited. The methylation status of seven CpGs in the differentially methylated region of HIPK3 (cg05501357) was measured using targeted methylation-sequencing technology. The association between methylation haplotypes and the overall methylation status of HIPK3 with clinical characteristics was assessed using generalized linear regression.

Results: All seven CpGs showed hypomethylation status in RA blood compared with OA and normal individuals (overall p= 1.143×10^-8 and FDR= 2.799×10^-7), which is consistent with the previously reported high expression of HIPK3 in RA immune cells. Among all seven CpGs, 33286785 showed the highest predictive power with an area under the curve (AUC) of 0.829; we received a higher AUC=0.864 when we combined HIPK3 with anti-citrullinated protein antibodies (ACPA -) and rheumatoid factor (RF +) in the prediction model, indicating that when a patient's ACPA is negative, HIPK3 can assist RF as a new clinical index for the diagnosis of RA. We also found that HIPK3 methylation levels were negatively correlated with C-reactive protein (CRP; r= -0.16, p= 0.01). Methylation haplotypes were analyzed, and the full methylation haplotype (FMH; r= 0.16, p= 0.01) and full non-methylation haplotype (FNH; r= 0.18, p= 0.0061) were negatively correlated with CRP.

Conclusion: Circulating blood methylation levels in the protein region of HIPK3 can be utilized as a supportive diagnostic biomarker and CRP level indicator for RA.

Keywords: DNA methylation; HIPK3; inflammation; methylation haplotypes; rheumatoid arthritis.

Figure 1

Differences in methylation level and correlation of HIPK3. (A), Methylation level of HIPK3 between RA, OA patients and healthy controls; The result found that the methylation level of HIPK3 was down-regulated in patients with RA (P= 1.143×10^-8). (B), Methylation level of seven CpGs in different patients; Further analysis showed that HIPK3 has seven CpGs, including 33286847, 33286832, 33286800, 33286785, 33286752, 33286747, and 33286724; The difference analysis showed that the methylation levels of these CpGs in patients with RA were down-regulated, and the statistical difference of 33286785 is the most significant. (C), Correlation analysis in different CpGs; The association analysis among seven CpGs showed that 6847 and 6832, 6800 and 6752 have strong correlation (|r|>0.8), and 6847 and 6785, 6832 and 6785, 6800 and 6785, 6724 and 6785, 6752 and 6724 have weak correlation. In general, there is a relatively close relationship between these CpGs. (D), Correlation between HIPK3 and CRP in patients with 2.6 ≤ DAS28-CRP ≤ 3.2; In patients with 2.6 ≤ DAS28-CRP ≤ 3.2, there was a stronger correlation between HIPK3 and CRP (r= -0.49, p= 0.00049), indicating that HIPK3 has a higher clinical value in predicting CRP levels in these patients. (E) Multi-factor joint ROC analysis of HIPK3 combined with RF and ACPA; In multi-factor joint analysis, the results showed that compare with RF/ACPA, HIPK3 has more valuable clinical diagnostic significance in patients with RF (-)/ACPA (-) (AUC=0.742 vs 0.023). In addition, in patients with RF (+)/ACPA (+), adding the detection of HIPK3 methylation level was helpful to improve the diagnosis of RA (AUC=0.988), and HIPK3 could assist RF/ACPA as a new clinical index for the diagnosis of RA. *: P<0.05.

Figure 2

Differences in methylation level and correlation of haplotype sites. (A) Methylation level of different haplotype sites between RA, OA patients and healthy controls; Further analysis showed that HIPK3 has eight haplothpe sites, including TTTTTTT, TTTTCTT, TTTTCCT, TTTTTCT, TTCTCTT, CCCCCCT, CCCCTCC, CCCCCCC; The difference analysis showed that the methylation levels of these haplotype sites in patients with RA were different regulated, and among them, the methylation level of CCCCCCC in RA showed a downward trend, and the statistical difference was the most significant. (B) Correlation of haplotype sites with CRP were different compared with TTTTTTT; The correlation between different haplotype sites and CRP was analyzed, and the results were compared with TTTTTTT; Taking TTTTTTT as the standard, the trend of correlation between different haplotype sites and CRP was shown. (C) Correlation of haplotype sites with ESR was different compared with TTTTTTT; The correlation between different haplotype sites and ESR was analyzed, and the results were compared with TTTTTTT; Taking TTTTTTT as the standard, the trend of correlation between different haplotype sites and ESR was shown. (D) In patients with RF (-)/ACPA (-), CCCCCCC was negatively correlated with CRP; Further analysis of the correlation between haplotype sites and clinical indexes in different subgroups; The results showed that in patients with RF (-)/ACPA (-), correlation between CCCCCCC and CRP was the highest and the P value was statistically significant; CCCCCCC was negatively correlated with CRP level (r= -0.73, p= 0.01). (E) In patients with 2.6 ≤ DAS28-CRP ≤ 3.2, CCCCCCC were negatively correlated with CRP; Similarly, subgroup analysis also found that In patients with 2.6≤DAS28-CRP ≤ 3.2, there was also a statistically significant negative correlation trend between CCCCCCC and CRP (r= -0.44, p= 0.0019), and which was helpful to judge the change trend of clinical CRP level.