Introduction

Aplastic anemia (AA) is the paradigm of human bone marrow (BM) failure syndromes, characterizing BM hypoplasia and pancytopenia in the peripheral blood (PB). Epidemiological studies suggest the prevalence of AA is ~ 2 per million persons in Western countries, and it is 2–3 times higher in East Asia, of which the incidence is ~ 7 per million persons in China [1,2,3,4]. Although its specific pathogenesis has not been completely elucidated, immune-mediated suppression of hematopoiesis and genetic vulnerability are considered to play vital roles.

Accumulating evidence has suggested that BM hematopoietic stem cells of most AA patients are attacked by auto-reactive cytotoxic T cells, such as interferon-γ (IFN-γ)-secreting CD4+Th1 cells and interleukin-17 (IL-17)-secreting CD17+Th17 cells [5,6,7,8]. Studies suggest interleukin-12 (IL-12) could induce the activation of signal transcription and transducer factor 4 (STAT4), which promotes the secretion of IFN-γ of Th1 cells [5]. A high concentration of IFN-γ could affect hematopoietic stem cell proliferations through inducing apoptosis and decreasing self-renewal [6]. The IL-12/STAT4/IFN-γ signaling pathway, as an important signaling pathway in mediating T cells immunity, contributes to pathogenesis of AA. Besides, current studies also demonstrate the absolute number of CD4+Th17 cells and the level of IL-17 are elevated in the AA patients and elucidate the specific pathological mechanism pathway of Th17 cells affecting AA risk [7]. Interleukin-23 (IL-23), as a proinflammatory and immunoregulatory agent, activates the transcription of STAT3, and increases the secretion of IL-17 of Th17 cells. The IL-23/STAT3/IL-17 signaling pathway is reported involving in the occurrence and progress of various autoimmune diseases, including AA [8].

Immunosuppressive therapy (IST), consisting of antithymocyte globulin (ATG)/antilymphocyte globulin (ALG), and cyclosporine (CsA), is the preferred treatment modality for AA patients. About 70% of AA patients could reach hematological recovery with IST, through suppressing the over-activation of T cells, regulating T cells subsets balance, inhibiting the over-production of myelosuppressive cytokines [9]. However, some patients with IST could not reach hematological response and some responders relapse. It is not fully understood which factors could affect patients’ response and relapse.

Numerous studies have reported that single-nucleotide polymorphisms (SNPs) of some immunoregulatory cytokines may have significant association with the occurrence, development, and therapy outcome of some diseases [10,11,12]. Therefore, we wondered whether genetic polymorphisms of cytokines in the IL-12/STAT4/IFN-γ and IL-23/STAT3/IL-17 signaling pathways were associated with inclination and IST outcome of AA individuals. Herein, in the current study, we performed an analysis to detect which SNPs in IL23R (rs11209032, rs1884444); STAT3 (rs1053005, rs4796793, and rs744166); IL12B (rs3212227); and STAT4 (rs7574865, rs897200) were associated with AA susceptibility, severity, and whether the significant SNPs affected response to IST of AA patients in the Han population in southwest China.

Materials and methods

Ethics statement

The study protocols were approved by the Ethics Review Board of West China Hospital of Sichuan University. The study conformed to the principles outlined in the declaration of Helsinki. All patients provided written informed consent before participating in the study.

Study subjects

From Sep. 2013 to Dec. 2014, consecutive patients with AA were recruited from the inpatient and outpatient units of the department of hematology in West China Hospital of Sichuan University. Patients were excluded if they fulfilled the following criteria: (1) Ethnicity other than Han-Chinese, (2) patients with bone marrow transplantation, (3) patients with AA/PNH syndrome, and (4) Patients with pure red cell aplastic anemia (PRCA). The controls were sex- and age-matched healthy unrelated individuals who were selected from those coming to the hospital for regular health examinations. At recruitment, each participant donated approximate 2–4 ml of blood for genomic DNA extraction.

We collected the basic characteristics of AA patients through the hospital information system and laboratory information system, including sex, age, red blood cells counts (RBCs), hemoglobin (HGB), platelet counts (PLTs), white blood cells counts (WBCs), absolute lymphocyte counts (ALCs), absolute neutrophil counts (ANCs), absolute reticulocyte counts (ARCs), mean corpuscular volume (MCV), and red cell distribution width (RDW). We divided the recruited AA patients into severe aplastic anemia (SAA) and non-severe aplastic anemia (NSAA), according to the criteria by Camitta for diagnosis in aplastic anemia [13, 14].

We followed up the recruited AA patients and recorded the therapeutic regimens and outcomes with IST. Therapeutic regimens of IST included standard ATG/ALG, ATG/ALG + CsA, and only regular CsA administration. We followed up the IST outcomes of AA patients more than 3 months before 31 Jan. 2015. Response was confirmed by two or more blood counts at least 4 weeks apart. Specific criteria of therapeutic effect was shown in Table 1 [14].

Table 1 Criteria of response to IST in AA
Full size table

SNP sites selection and genotyping

We screened polymorphic sites in genes IL23R, STAT3, IL12B, and STAT4 by reading articles. We searched the tag SNPs in genes IL12B, STAT4, IL23R, and STAT3 through Hapmap database and Haploview software. We hunted for the functional SNPs in genes IL23R, STAT3, IL12B, and STAT4 through the online software SNPinfo (http://snpinfo.niehs.nih.gov/). As a result, we picked up eight SNPs in our current study, including IL23R (rs11209032, rs1884444); STAT3 (rs1053005, rs744166, and rs4796793); IL12B (rs3212227); and STAT4 (rs7574865, rs897200).

Genomic DNA was extracted using a QIAamp DNA Blood mini kit (QIAGEN, Germany) and diluted to 10 ng/μl with buffer, according to the manufacturer’s instructions. The target fragments containing reference alleles from all of the study subjects were amplified by PCR with corresponding specific primers (Table 2). The specific PCR amplification and corresponding genotyping of the eight SNPs were performed by the method of high-resolution melting (HRM) in the LightCycler® 480 (Roche Diagnostics) and direct sequencing.

Table 2 The primers of reference alleles studied
Full size table

Statistical analysis

Continuous variables conforming to normal distribution were described as mean ± standard deviation (SD); otherwise, by median and interquartile range (P25–P75). Categorical variables were described as frequency and percentage (%). Variables conforming to normal distribution and homogeneity between two groups were compared by t test; otherwise, by Mann-Whitney U test.

Hardy-Weinberg equilibrium was evaluated by chi-square test for each SNP. P < 0.05 was considered statistically significant.

Allele and genotype case-control association analysis was conducted using all genotype data. For each SNP, we calculated empirical significance value on the basis of 10,000 permutations. This ensures that deviation from small sample size will not cause false positives. All the statistical analysis was performed by the software PLINK version 1.07 (http://pngu.mgh.harvard.edu/~purcell/plink). Two-sided P < 0.05 was considered statistically significant.

To assess which factors affected AA IST outcome, we utilized single-sample t test to screen the significant variables, then logistic method to adjust the variables effect. The statistical analysis was conducted by SPSS 17.0. Two sided P < 0.05 was considered statistically significant.

Meta-analysis

For further assessment of the association between genetic polymorphisms in genes IL23R, STAT3, IL12B, and STAT4 and AA susceptibility, we searched MEDLINE, EMBASE, Cochrane library, and Chinese databases (CNKI, CQVIP, and Wan-fang Databases) to collect the related literatures published in English and Chinese till Jan 2015, utilizing the theme words of “aplastic anemia,” “SNP,” “rs11209032,” “rs1884444,”, “rs744166,” “rs4796793,” “rs1053005,” “rs3212227,” “rs897200,” and “rs7574865.” We extracted the relative data from each eligible article and conducted a meta-analysis by Review Manager Version 5.3 (http://www.cc-ims.net/RevMan).

The inclusion criteria were (1) studies evaluating the association between “rs11209032,” “rs1884444,” “rs744166,” “rs4796793,” “rs1053005,” “rs3212227,” “rs897200,” and “rs7574865” polymorphisms and AA risk. (2) Available data for calculating allelic odds ratios (ORs) with corresponding 95% confidence interval (95% CI). (3) Genotypes in controls conforming to Hardy-Weinberg equilibrium (P > 0.05). Reviews and case reports were excluded.

The following data were extracted from each eligible study: first author’s name, year of publication, study design, geographic location or ethnicity of study population, sample size, and frequency of allele in cases and controls. Heterogeneity across all the eligible studies was estimated by the Cochran’s Q statistic. Heterogeneity was considered evident at P < 0.05 for the Q statistic. Random-effect model was used when heterogeneity among studies existed; otherwise, fixed-model was utilized. The allele model with combined ORs with 95% CIs was used to assess the association between the investigated eight SNPs and AA risk. P value was two sides and P < 0.05 was considered statistically significant.

Results

Characteristics of the study population

During our study period, we consecutively recruited a total of 180 AA patients. Out of the 180 patients, 4 patients were not Han nationality, 3 patients suffered from PRCA, 3 patients were diagnosed AA/PNH syndrome, and 4 patients were because of specimen failure of DNA extraction. After excluding these ineligible patients, a total of 164 patients were included in the current study, including 88 NSAA and 76 SAA. The characteristics of the patients were summarized in Table 3.

Table 3 Basic characteristics of study population
Full size table

The association between genetic polymorphisms of IL23R (rs11209032, rs1884444); STAT3 (rs1053005, rs744166, and rs4796793); IL12B (rs3212227); and STAT4 (rs7574865, rs897200) and AA risk

The genotypic distribution did not deviate from the Hardy-Weinberg equilibrium for the eight target SNPs in the cases and controls (Table 4). The minor allele frequency of the eight SNPs were similar to the ones reported by NCBI database and some articles in Chinese population [10, 12, 15]. There was no statistically significant difference in genotype and allele frequency for the SNPs of rs11209032 (C > A), rs1884444 (G > T), rs1053005 (C > T), rs744166 (G > A), rs4796793 (G > C), rs3212227 (G > T), rs897200 (C > T) in the 164 AA cases and 211 controls. Our current study showed that T allele and TT genotype of rs7574865 variant were more frequent in the 164 cases than in the 211 controls (42.7 vs 34.6%; 18.9 vs 10.5%). In the additive model, individual carrying the rs7574865 T allele demonstrated a 37% (OR (95% CI) = 1.37 (1.02–1.85), Pper = 0.036) increased AA risk. In the recessive model, carrier with rs7574865 TT genotype had an increased AA risk with an OR of 2.08 (OR (95%CI) = 2.08 (1.14–3.70), Pper = 0.017). In the stratified analysis, our study found that rs7574865 T allele and TT genotype were mainly associated with NSAA occurrence. Specific results are shown in Tables 5, 6, 7, and 8.

Table 4 The results of HWE of the genotypic distribution of 8 SNPs (rs1884444 (G > T), rs11209032 (G > A), rs7574865 (T > G), rs897200 (C > T), rs3212227 (G > T), rs744166 (G > A), rs1053005 (C > T), and rs4796793 (G > C)) in the cases and controls
Full size table
Table 5 The association between rs1884444 (G > T), rs11209032 (G > A), rs7574865 (T > G), and rs897200 (C > T) polymorphisms and AA risk
Full size table
Table 6 The association between rs3212227 (G > T), rs744166 (G > A), rs1053005 (C > T), and rs4796793 (G > C) polymorphisms and AA risk
Full size table
Table 7 The association between rs1884444 (G > T); rs11209032 (G > A); rs7574865 (T > G), rs897200 (C > T), rs3212227 (G > T), rs744166 (G > A), rs1053005 (C > T), and rs4796793 (G > C) polymorphisms and AA risk (NSAA vs controls)
Full size table
Table 8 The association between rs1884444 (G > T), rs11209032 (G > A), rs7574865 (T > G), rs897200 (C > T), rs3212227 (G > T), rs744166 (G > A), rs1053005 (C > T), and rs4796793 (G > C) polymorphisms and AA risk (SAA vs controls)
Full size table

The association between genetic polymorphisms of IL23R (rs11209032, rs1884444); STAT3 (rs1053005, rs744166, and rs4796793); IL12B (rs3212227) and STAT4 (rs7574865, rs897200) and AA severity

To assess the association between these eight SNPs polymorphisms and AA severity, we compared the frequency distribution of alleles and genotypes of these eight SNPs between SAA and NSAA cases. Consequently, there was no statistically significant difference in allele and genotype frequency for the SNPs of rs1884444 (G > T), rs1053005 (C > T), rs744166 (G > A), rs4796793 (G > C), rs3212227 (G > T), rs897200 (C > T) in the 76 SAA cases and 88 NSAA ones. Our current study demonstrated that G allele and GG genotype of rs11209032 variant were more frequent in the 88 NSAA cases than in the 76 SAA ones (42.1 vs 55.7%; 14.5 vs 33.0%). Specific results are shown in Table 9.

Table 9 The association between rs1884444 (G > T), rs11209032 (G > A), rs7574865 (T > G), rs897200 (C > T), rs3212227 (G > T), rs744166 (G > A), rs1053005 (C > T) and rs4796793 (G > C) polymorphisms and AA severity (NSAA vs SAA)
Full size table

The results of affected factors of IST outcome of AA patients

During our follow-up period, a total of 54 patients received IST. Out of the 54 patients, 13 patients were treated with ATG/ALG, 3 patients with ATG combined with CsA, and 38 patients with CsA administration. A total of 14 patients lost follow-up or the follow-up time was less than 3 months. Finally, we assessed the 3-month immunosuppressive treatment outcome of the 40 AA patients; 21 patients achieved hematological response and 19 ones did not. The response rate was 52.5%. Our study found G allele and GG genotype of rs11209032 variant, and GG genotype of rs744166 site were associated with the IST outcome of AA patients. Our current study did not show that these factors of patient onset age, sex, PLTs, RBCs, HGB, WBCs, ALCs, ANCs, ARCs, RDW, and MCV were associated with IST outcome of AA patients. Specific results are shown in Tables 10 and 11.

Table 10 The results of rs1884444 (G > T), rs11209032 (G > A), rs7574865 (T > G), rs897200 (C > T), rs3212227 (G > T), rs744166 (G > A), rs1053005 (C > T) and rs4796793 (G > C) polymorphisms and IST outcome of AA patients
Full size table
Table 11 The results of the basic characteristic and IST outcome of AA patients
Full size table

The results of meta-analysis

As shown in Fig. 1, a total of 187 articles were retrieved. After reading the titles, 181 unrelated studies were excluded. After further reading the abstract and full text of the remaining six papers, only one paper which focused on the association between rs7574865 polymorphism and AA risk, met the inclusion criteria [16]. Our meta-analysis included the Feng’s and our current study, consisting of 366 AA patients and 406 controls. The basic information of included studies was shown in Table 12. We found there was no heterogeneity between studies; therefore, we utilized fixed-model to assess the association between rs7574865 polymorphism and AA risk in the combined population. Our meta-analysis also showed T allele and TT genotype increased AA risk. In the additive model, individual carrying the rs7574865 T allele demonstrated a 44% (OR (95% CI) = 1.44 (1.17–1.78), P = 0.001) increased AA risk. In the recessive model, carrier of rs7574865 TT genotype had an increased AA risk with an OR (95% CI) of 1.81 (1.15–2.85, P < 0.001). Specific results are summarized in Figs. 2 and 3.

Fig. 1
figure 1

Specific screening flow diagram of articles in meta-analysis

Full size image
Table 12 The results of Meta-analysis of rs7574865 polymorphism and AA risk
Full size table
Fig. 2
figure 2

The results of meta-analysis of rs7574865 polymorphisms and AA risk under additive model

Full size image
Fig. 3
figure 3

The results of meta-analysis of rs7574865 polymorphisms and AA risk under recessive mode

Full size image

Discussion

STAT4, as a critical immune regulation factor in some signal pathways, was first found in GWAS that rs7574865 (T > G)genetic polymorphism involved in rheumatic arthritis and systemic lupus erythematosus occurrence [17]. Studies have shown similar autoimmune diseases share the same susceptibility genes [18]. Since then, plenty of researchers also reported that STAT4 genetic polymorphisms were associated with some other autoimmune diseases [19,20,21]. Our current data showed T allele and TT genotype of rs7574865 increased AA risk. In the additive model, individual carrying the rs7574865 T allele demonstrated a 37% increased AA risk. In the recessive model, carrier of rs7574865 TT genotype had an increased AA risk with an OR of 2.08. In the stratified analysis, we found that T allele and TT genotype of rs7574865 were mainly associated with NSAA occurrence. Our current study results are consistent with the Feng’s [16]. To avoid false positive results caused by small sample size, we performed the Meta-analysis by combining our current study and Feng’s, finding T allele and TT genotype of rs7574865 really increased AA risk. Previous study results have demonstrated rs7574865 T allele was associated with high expression of gene of TBX21. Gene TBX21 could regulate the differentiation of transcript factor of Th1 cells [22]. However, due to our current limited study condition, we did not perform further research about how rs7574865 affect AA occurrence.

IL-23R mainly mediates the intracellular signal transduction of IL-23, which could promote various cells differentiation and development and inflammatory cytokines secretion through STAT3 signal pathways [23]. Our current data showed individuals carrying rs11209032 G allele and GG genotype could protect them from NSAA to SAA development. Maybe, due to our current small sample size, we did not find the association between rs11209032 genetic polymorphisms and AA occurrence, but AA severity. Although, many researchers have reported rs11209032 genetic polymorphisms related to some T lymphocyte cell-mediated autoimmune diseases, few studies explore the specific function mechanism of rs11209032 site. We predict rs11209032 location is the transcription factor binding site by SNPinfo software, indicating a possible role in affecting the transcription factor action and mRNA expression, even the translation of related protein, finally impacting the diseases occurrence and development. However, our current study results need to be further verified by larger sample, multiple nationalities, and multi-area studies.

In our evaluation of IST outcome of AA patients, the hematological response rate was 52.5%, different with those reported by Scheinberg [24] and Atta [25]. We found G allele and GG genotype of IL23R-rs11209032 were associated with better IST outcome of AA patients. By contrast, GG genotype of STAT3-rs744166 was associated with IST failure of AA patients. Previous researchers have reported that the onset age and ARCs, and ALCs of AA patients could be the predictive factors of IST outcome [26]. Our current study was unable to show the association. These different study findings may be caused by following reasons. Firstly, different treatment regimens (ATG/ALG, ATG/ALG + CsA, just CsA), recruited study subjects, and follow-up time maybe the main reasons of different response rates. Secondly, studies suggest that the treatment effect of horse ATG is better than rabbit ATG; ATG from different sources may have different response rates [24, 25]. Thirdly, different drug dose and response of study population may lead to different response effect. Lastly, our current study retrospectively collected the data of AA patients, some patients did not have ARCs tests in the first clinic appointment, maybe, causing the result difference of many researches.