Association of DYNC1H1 gene SNP/CNV with disease susceptibility, GCs efficacy, HRQOL, anxiety, and depression in Chinese SLE patients

Abstract Background Systemic lupus erythematosus is a heterogeneous autoimmune disease characterized by multi‐system injuries and overproduction of autoantibodies. There are many genetic studies on SLE, but no report has considered the relationship between cytoplasmic dynein and SLE susceptibility. Objectives Our study intends to investigate whether DYNC1H1 gene SNP/CNV is related to SLE susceptibility, GCs efficacy, HRQOL, anxiety, and depression in Chinese SLE patients. Methods A total of 502 cases and 544 healthy controls were recruited into the case‐control study, and 472 subjects from the case group were followed up for 12 weeks to evaluate GCs efficacy, HRQOL, anxiety, and depression. Multiplex SNaPshot technique was applied to genotype the seven SNPs of DYNC1H1, and AccuCopyTM method was conducted to quantify the copy number of DYNC1H1. Anxiety and depression were evaluated using HAMA and HAMD‐24 scales, respectively. The SF‐36 scale was used to assess HRQOL. Results The significant association between SNP rs1190606 and SLE susceptibility was displayed in the dominant model (PBH = 0.004) as well as its allele model (PBH = 0.004). We also found that SNP rs2273440 was related to photosensitization symptom in SLE patients (PBH = 0.032). In the follow‐up study, SNP rs11160668 was connected with the improvement of BP in male patients (PBH = 0.011). However, no association of DYNC1H1 gene with GCs efficacy, anxiety, and depression was found. No CNV in DYNC1H1 was detected. Conclusions The study suggests that DYNC1H1 gene polymorphisms may have an effect on SLE susceptibility and BP improvement of HRQOL in Chinese SLE patients.


| INTRODUC TI ON
Systemic lupus erythematosus (SLE) is an intricate autoimmune disease characterized by a broad spectrum of organ damage manifestations and flare-remission pattern. 1 Although the pathogenesis of SLE has not been clearly elucidated, the fact that it is a multifactorial disease involving genetic, epigenetic, and environmental factors has been identified. 2 The report pointed out that SLE mainly harms women, whose disease risk is more than nine times that of men. 3 In the past decade, more than 100 genetic loci related to SLE have been identified based on genome-wide association studies (GWAS). 4 The study on SLE genetics has attracted more and more attention from researchers.
As the most frequently used drug in SLE treatment, glucocorticoids (GCs) mainly exert their biological functions through the glucocorticoid receptor (GR). GCs are divided into endogenous and exogenous, both of which play an essential part in the pathology of SLE. 5 Generally speaking, SLE patients can be treated effectively with GCs, whereas some unpredictable patients respond poorly to GCs therapy and even have undesirable outcomes. 6 Therefore, it is worth noting how to identify and deal with individuals who are ineffective in GCs treatment.
Cytoplasmic dynein is a necessary molecular motor belonging to the AAA family of ATPases. 7 Dynein is mainly responsible for facilitating the retrograde movement of numerous intracellular cargos along microtubules to ensure accurate and efficient cell activities. 8 Dynein 1 heavy chain 1 (DYNC1H1) encodes an essential subunit, which is the core structure of dynein complex. 9 DYNC1H1 is located on chromosome 14q32.31, which has a miRNA cluster that regulates 48 lupus susceptibility genes. 10 Researchers have discovered that dynein can regulate the development of B cells closely related to autoimmune diseases with distinctive genetic mechanisms. 11 The above findings seem to suggest that DYNC1H1 may be associated with SLE susceptibility. After GCs bind to GR, the formed complex must rely on dynein to drive the movement toward the nucleus. 12 We assumed that the structural changes of dynein might have an impact on the transport of the complex, leading to glucocorticoid resistance. Previous research has also revealed that GR gene polymorphisms affect the GCs efficacy in SLE patients. 13 Thus, we hypothesized that DYNC1H1 gene polymorphisms might make a difference to GCs efficacy in SLE. Additionally, the relationship between copy number variation (CNV) of DYNC1H1 and SLE susceptibility was also explored in this paper.
SLE is closely related to central nervous system damage, resulting in a wide range of neurologic sequels, of which anxiety and depression are the most common disorders. 14 Possible causes of the high incidence of mental disorders in SLE patients include long-term GCs use, psychological pressure, and recurrent symptoms. 15 Dynein plays a vital role in the development of neurons, such as neuronal migration and neuronal regeneration. 16 Mutations in DYNC1H1 induce dysfunction of dynein-dependent signaling pathways, resulting in various neurological diseases. 17 No one has yet considered linking the variations in DYNC1H1 with anxiety and depression.
Meanwhile, frequent disease activity tends to worsen physical and mental health, resulting in a deterioration in SLE patients' healthrelated quality of life (HRQOL). 18 Given previous research showing the impact of GR gene polymorphisms on HRQOL, we assumed that DYNC1H1 (the driving force for GR transport) might similarly affect HRQOL. 19 Therefore, we sought to probe whether DYNC1H1 gene polymorphisms have an effect on mental disorders (anxiety and depression) and HRQOL in SLE patients.

| Subject and inclusion criteria
The research program has been approved by the ethics committee of Anhui Medical University before implementation. All selected subjects offered written informed consent. This study mainly consisted of a case-control study aimed at probing the association between DYNC1H1 gene and SLE and a follow-up study aimed at exploring whether DYNC1H1 gene affects the GCs efficacy, HRQOL, anxiety, and depression in SLE patients.
In the case-control study, 502 diagnosed SLE cases were collected from the First and Second Affiliated Hospitals of Anhui Medical University, and 544 healthy controls were recruited from the healthy population identified by the physical examination departments of hospitals. Each patient was diagnosed by rheumatologists referring to the SLE diagnostic criteria published by the American College of Rheumatology in 1997. Cases with other autoimmune diseases and critical illness were not included in the case group. Moreover, the included healthy controls did not have SLE or other immune system diseases and denied a family history of SLE.
A total of 472 patients with SLE were recruited in the follow-up study. All enrolled patients achieved a total point of 5 or higher on the SLE disease activity index (SLEDAI) at baseline. Meanwhile, participants did not accept GCs treatment within the first three months of enrollment or only obtained the minimum dose of GCs for maintenance treatment. Besides, subjects who met any of the following criteria would be eliminated from the study:  Overall, a GCs (prednisone) treatment dose of 10mg/day -0.5mg/ kg/day would be available to patients with total scores below 10, and patients whose total scores on ≥10 would receive a treatment dose of 0.5 -1.0mg/kg/day. In the meantime, the adjustment to GCs dose required consultation with rheumatologists. Furthermore, combination therapy of hydroxychloroquine, which exerted beneficial effects in improving clinical symptoms and retarding the onset of damage, was provided to patients. 20 Rheumatologists performed SLEDAI score assessment on all patients at baseline, 4, 8, and 12 weeks to understand their disease status. Additionally, the clinical symptoms of patients were recorded in detail. The change in SLEDAI score after 12 weeks reflected the efficacy of GCs. Patients were assigned to different groups according to the reduction of SLEDAI score and the improvement of clinical symptoms. Patients whose SLEDAI score reduced to below 5 points at 12 weeks or the diminished score (baseline -12 weeks) reached at least 5 points were divided into the GC-sensitive groups; otherwise, patients would be classified as the GC-insensitive group. Moreover, somebody treated with other immunosuppressors owing to poor efficacy would also be assigned to the GC-insensitive group. were calculated through these eight dimensions. The score range of each dimension is 0 -100, with a higher score demonstrating higherlevel HRQOL. We measured the HRQOL of each subject separately at baseline and 12 weeks. The judgment of baseline HQORL depended on the comparison of SF-36 total score and median: a total score greater than the median meant the good quality of life, while others meant the poor quality of life. The change in the score of each dimension after 12 weeks was viewed as an indicator to reflect the improvement of HRQOL in SLE patients.

| Assessment of anxiety and depression
Both the Hamilton Depression Rating Scale (HAMD-24) and the Hamilton Anxiety Rating Scale (HAMA) are currently widely adopted psychometric instruments. 22,23 The HAMD-24 was applied to evaluate depression, and a score of 8 or above was diagnosed as depression.
The HAMA was employed to assess anxiety, and a score greater than 6 was determined as anxiety. Changes in HAMA and HAMD-24 scale scores after 12 weeks (score at baseline -score at 12 weeks) reflected the improvement of patients' anxiety and depression.

| Tag SNPs electing and genotyping
The HapMap database was adopted to search for single nucleotide polymorphisms (SNPs) in DYCN1H1 gene representing Han-Chinese.
The QIAGEN DNA extraction kit (Germany) was employed to acquire genomic DNA from peripheral blood samples. DNA samples were preserved in the medical refrigerator at −80℃ until being used.
SNPs genotyping was performed using the Multiplex SNaPshot technique based on ABI fluorescence-based assay discrimination method (Applied Biosystems, Foster City, CA). The SNaPshot information of DYNC1H1 gene was presented in Supplementary Table S1.

| Detection of DYNC1H1 copy number
The quantification of DYNC1H1 copy number was accomplished through a well-established AccuCopy TM method. AccuCopy TM method is based on competitive PCR amplification, which can accurately obtain the CNV status at multiple genomic loci in the same assay. 24 The primer information of the detected fragment was shown in Supplementary Table S2.

| Statistical analysis
For continuous variables, arithmetic mean with standard deviation (mean (SD)) was employed to represent normally distributed data, while median with inter-quartile range (median, (P 25 -P 75 )) was applied to describe abnormally distributed variables. Categorical variables were presented as counts (percentage). In order to assess the differences between groups, Student's t test was adopted in the case of normally distributed data and Mann-Whitney U test in the case of abnormally distributed data. Chi-squared or Fisher's exact test was utilized to compare differences between qualitative variables. Similarly, Hardy-Weinberg equilibrium (HWE) also was evaluated by chi-squared test. We conducted univariate logistic regression to discover potential associations and presented the results in the form of odds ratios (OR) with 95% confidence intervals (95% CIs) and P value. To control the interference of confounding factors in the case-control study (sex, age) and the follow-up study (sex, age, height, weight, marital status, cigarette smoking, alcohol drinking, tea consumption, history of GCs therapy, GCs dose, and SLEDAI scores at baseline), analysis results were corrected with multivariate logistic regression.
The above analyses were performed by SPSS version 21.0 (SPSS, Inc., Chicago, IL). This study also performed haplotype analysis using SHEsis software, with the criterion that the haplotype had a minimum frequency greater than 0.03. For false-positive error and multiple comparisons problems, the Benjamini-Hochberg (BH) method based on the false discovery rate (FDR) was conducted to adjust P value. Additionally, we have further analyzed data using gender as a stratification factor.
The threshold of significance was P < 0.05 (2-sided) for all tests.

| Characteristics and HWE analysis
The basic and clinical characteristics of cases and controls were shown in Supplementary Table S3. The case-control study consisted of 502 patients and 544 healthy controls. There were 50 males and 452 females in the case group with an average age of 35.32 (12.22) years, while the control group had 63 males and 481 females with an average age of 35.31 (9.65) years. No statistical differences in age and gender between groups were found. Genotype frequencies of DYNC1H1 gene and HWE results were summarized in Supplementary Table S4. Since rs10131545 and rs10132469 did not meet HWE, subsequent analyses would exclude these two SNPs.

| Allele of SNPs and SLE susceptibility
We analyzed the distribution of SNP alleles between SLE cases and controls. As shown in Table 1, compared with the control group, the frequency of three minor alleles in the case group was significantly higher, including rs1004903 (allele A, P = 0.039), rs1190606 (allele G, P = 5.30×10 −4 ), and rs2273440 (allele A, P = 0.038). After adjusting P value, the result showed that the higher frequency of allele G in rs1190606 might contribute to an increased risk of SLE (P BH = 0.004) ( Figure 1A). The results of gender stratification demonstrated this association was more significant in the female subgroup (P BH = 0.001) (Supplementary Tables S5 and S6).

| DYNC1H1 gene polymorphisms and SLE susceptibility
The results of regression analyses were displayed in  Table S9).
The result demonstrated that there were statistical differences in the distribution of haplotypes GAGAGGG and GAGGGGG between SLE patients and controls (P = 0.017 and P = 0.024). These associations disappeared after correcting P value by FDR method.
However, the analysis of the female subgroup indicated that haplotypes GAGAGGG and GAGGGGG still had associations with SLE susceptibility (P BH , 0.015 and 0.006, respectively) ( Table 3), whereas no connection was found in males (Supplementary Table S10). These associations further supported the hypothesis that DYNC1H1 gene polymorphisms may make a difference to SLE susceptibility.

| DYNC1H1 gene polymorphisms and clinical manifestations
The relationship between genotype distribution in diverse inherited models and clinical manifestations of SLE patients was presented in Supplementary

| Allele of SNPs and GCs efficacy
Differences in allelic distribution between the GC-sensitive group and the GC-insensitive group were shown in Supplementary Table   S16. No association between SNP alleles and GCs efficacy was observed. Similarly, gender stratification results found no statistical differences (Supplementary Tables S17 and S18).

F I G U R E 1
The association between rs1190606 of DYNC1H1 gene and SLE susceptibility: A, Allele distribution of rs1190606 between case and control groups; B, Genotype frequency of rs1190606 in dominant model between case and control groups

| DYNC1H1 gene polymorphisms and GCs efficacy
No statistical association was observed in logistic regression analysis (

| CNV analysis of DYNC1H1 gene
For 547 samples, the copy number of DYNC1H1 gene fragment was tested. The original data were displayed in Figure 2. After testing, the copy number of DYNC1H1 in all samples was 2, and no copy number variation was discovered. Thus, no more analysis was performed. This is the first study to link DYNC1H1 gene polymorphisms with SLE susceptibility. We presented the following opinions to illustrate our findings.

| DISCUSS ION
The pathological similarities between autoimmune diseases and cancer, such as abnormal cell proliferation and tissue microenvironment disorders, have been recognized for decades. 25 Due to the function of dynein in maintaining the cell cycle, its role in cancer has been extensively studied. Somatic mutations of DYNC1H1 have been detected in several cancers, including intraductal papillary mucinous neoplasm, pancreatic neuroendocrine neoplasm, and glioblastoma multiforme. 26,27 DYNC1H1 has been recognized as a driver gene for colorectal cancer to progress to stage-II. 28 Huang et al.
stated that, compared with cholecystitis and normal gallbladder tissues, the expression of DNYC1H1 in human primary gallbladder cancer (PGC) was markedly down-regulated. Their results indicated that DYNC1H1 was a potential diagnostic biomarker for PGC. 29 The reported data showed that DYNC1H1 was overexpressed in gastric cancer cell lines resistant to 5-fluorouracil, paclitaxel, and cisplatin, suggesting that DYNC1H1 may be used as a biomarker for predicting chemotherapy efficacy in gastric cancer. 30 Moreover, some studies have revealed that dynein seems to play a key role in certain autoimmune diseases. It has been shown that dynein gene may be associated with rheumatoid arthritis, an auto- vation. 35 Meanwhile, it has been proved that dynein is required for the movement of B-cell receptor microclusters and antigen gathering during B-cell activation. 36 Thus, it is reasonable to speculate that mutations in DYNC1H1 gene may interfere with its biological function and affect the activation of B and T cells, thereby changing the susceptibility to SLE. The above evidence seems to support our findings regarding the relationship between DYNC1H1 gene polymorphisms and SLE susceptibility. Remarkably, we observed a gender difference in the association of DYNC1H1 gene with SLE, which may be related to the high risk of SLE in women, or to the limited male cases included in this study.
As the drug of choice for SLE treatment, GCs mainly rely on GR to exert anti-inflammatory and immunomodulatory functions.
Mediated by transport-related proteins, GR forms the GR-Hsp90-Hsp70-immunophilin complex, which then is transported retrogradely with the assistance of cytoplasmic dynein. 37