The relationship between interferon-gamma (INF-γ) single nucleotide polymorphism +874(T/A) and occurrence risk of aplastic anemia: a meta-analysis

ABSTRACT

Objective: This meta-analysis was designed to investigate the association between interferon-gamma (IFN-γ) polymorphisms and occurrence risk of aplastic anemia.

Methods: Literature search was conducted in PubMed, Embase and Cochrane Library up to April 2018. The pooled odds ratios (ORs) and 95% confidence interval (CI) were calculated by R 3.12.

Results: Total five studies with 304 aplastic anemia patients and 588 controls were included. The statistically significant results were found in the following models: allele genetic model (T vs A: OR = 2.1749, 95% CI = 1.6825–2.8114, P < 0.01), additive genetic model (TA vs AA: OR = 2.1071, 95% CI = 1.3962–3.1799, P < 0.01; TT vs AA: OR = 4.5788, 95% CI = 2.6606–7.8797, P < 0.01), recessive genetic model (TT vs AA + TA: OR = 2.5579, 95% = 1.6680–3.9226, P < 0.01), dominant genetic model (TT + TA vs AA: OR = 2.5599, 95% = 1.7424–3.7611, P < 0.01), the results suggested that the increased occurrence risk of aplastic anemia is significantly associated with the IFN-γ polymorphism.

Conclusions: Patients with IFN-γ genotype carrying allele T have higher occurrence risk of aplastic anemia.

KEYWORDS:

• IFN-γ
• aplastic anemia
• polymorphisms
• meta-analysis

Introduction

Aplastic anemia is a rare hematological disease with abnormal number and functions of T lymphocytes as a main pathogenic factor and is characterized by pancytopenia and bone marrow hypoplasia which are mostly caused by immune-mediated destruction of hematopoietic stem cells [1]. Currently, although there is technical development recently, the mortality rate caused by aplastic anemia is still high [1–3]. The unclear pathogenesis is one of the main factor limiting the development of prevention and treatment strategy of aplastic anemia.

Previous studies have found the high levels of interferon gamma (IFN-γ) protein in peripheral blood mononuclear cells in patients with aplastic anemia [4–6]. It was reported that the IFN-γ protein may be associated with the pathogenesis of aplastic anemia via T-bet/IFN-γ Signal Transduction Pathway, inflammatory factors (such as Tumor necrosis factor and interleukin) or cytokines (CD4 and CD8) [6–9]. Considering the protein levels generally regulated by its gene, it was speculated that the IFN-γ gene may be also associated with the occurrence or development of aplastic anemia by regulating the expression of IFN-γ protein. Moreover, some studies have proved that the IFN-γ gene polymorphism +874(A/T) is associated with the occurrence risk of aplastic anemia [10–13]. However, we could not definitely conclude the relationship between IFN-γ gene polymorphisms and occurrence risk of aplastic anemia due to small sample size and regionalism of participants in each study. Thus, we performed this meta-analysis to synthetically assess the relationship between IFN-γ gene polymorphisms and occurrence risk of aplastic anemia based on the data of recent studies.

Materials and methods

Search strategy

We conducted a literature search on databases including Embase, PubMed and Cochrane Library with the deadline of April 2018. The key words included ‘anemia, aplastic’ OR ‘Aplastic Anemia’ OR ‘aplastic’ OR ‘pure red cell aplasia’ OR ‘aplastic anemia’ AND IFN-γ AND polymorphism AND IFN-γ (interferon OR IFN) AND (polymorphi*). We also manually scanned the reference lists of some relevant reviews for obtaining additional studies.

Inclusion and exclusion criteria

The inclusion criteria of the included studies in this meta-analysis were (1) Studies were case–control studies; (2) participants were patients with aplastic anemia and controls (healthy individuals or non-aplastic anemia patients); (3) the IFN-γ polymorphism was detected; (4) the frequencies of alleles (T, A) and genotypes (TT, TA, AA) was reported or can be obtained by calculation.

Exclusion criteria were (1) non-original articles (such as letters, comments or reviews) (2) no available data, (3) duplicated publications (only the study with complete data can be included in this meta-analysis).

Quality assessment and data extraction

The quality of included studies was assessed independently by two investigators. Different evaluation results were resolved by discussion to ensure consistency of evaluation. The extracted data should be recorded using a predesigned form including first author name, publication year, study location, age, gender, sample size, method of detecting IFN-γ polymorphisms, and outcomes.

The quality of these included case–control studies was evaluated by the Newcastle-Ottawa-Scale (NOS) [14], which comprised of three key domains and eight questions (Selection: total 4 questions, 1 score assigned to each question; Comparison: total 1 question, 2 scores assigned to this question; Explore: total 3 questions, 1 score assigned to each question). Thus, total of 9 scores can be obtained. The study with less than 5 scores was regarded as a low-quality study by NOS scale in this meta-analysis.

Statistical analysis

All analyses were performed using R 3.12. Hardy–Weinberg equilibrium was tested in controls by a χ² test [15]. The Chi-square and I² tests were used to evaluate the heterogeneity across included studies. The presence of significantly statistical heterogeneity of outcomes across included studies was assessed when P < 0.05 and I² > 50%. If there is no significant heterogeneity (P-value for heterogeneity ≥ 0.05 or I² ≤ 50%), a fixed-effects model was chosen to pool the effect size; otherwise, a random-effects model was used. The pooled odds ratios (ORs) as well as their 95% confidence interval (CI) were used to assess the relationship between IFN-γ polymorphisms and occurrence risk of aplastic anemia. In addition, sensitivity analysis was performed by omitting one study at a time. Publication bias was assessed by Egger’s Test.

Results

Characteristics of included studies

Literature search and study selection process are shown in Figure 1. The initial search by literature search on databases identified 67 articles (PubMed: n =  25, from Embase: n = 41, Cochrane Library: n =  1). After excluding duplicates, 54 potentially relevant articles were remained. Afterward, 34 irrelevant articles were excluded by scanning titles. Among the 20 remaining articles, 8 studies (2 letter/editorial; 2 case series/report; 4 animal/cell study) were removed by scanning abstracts, 7 articles (2 review; 3 cannot extract data; 2 duplicated populations) were excluded. No additional studies were found by manual search. Finally, five articles were included in this meta-analysis [10–13,16].

Figure 1. Flow diagram of the study selection process.

The characteristics of the five included studies are listed in Table 1. The publication year ranged from 2004 to 2015. In the five included case–control studies, a total of 304 aplastic anemia patients and 588 controls were reanalyzed in this meta-analysis. All these studies detected the polymorphisms of IFN-γ by PCR-RFLP (polymerase chain reaction-restriction fragment length polymorphism), PCR-SSP (polymerase chain reaction-sequence-specific primer) or ARMS-PCR (amplification refractory mutation system-polymerase chain reaction), except one study did not describe the detection method [13]. The polymorphisms of IFN-γ of the controls in each included study met the Hardy–Weinberg equilibrium, except El Mahgoub IR 2014 [11] and Zayed RA 2015 [16]. In addition, based on the NOS scores, no low-quality studies were included in this meta-analysis (Table 1).

Table 1. Characteristics of included studies.

Author	Public Year	Location	Study Year	NOS	Detection	Groupe	N	Male/Female	age, years	AA	AT	TT	HWE
													Χ^2*	P
Chang H	2010	China	2006.2–2007.12	6	ARMS-PCR	AA	54	28/26	37.74 ± 17.37	6	25	23	0.460	0.4976
						Co	51	27/24	26.9 ± 11.2	20	22	9
El Mahgoub IR	2014	Egypt	2010.9–2011.5	7	ARMS-PCR	AA	50	24/26	11.2 ± 7.135	16	21	13	4.340	0.0372
						Co	50	27/23	11.51 ± 7.25	31	13	6
Fermo E	2004	Italy	NA	5	PCR-SSP	AA	80	NA	NA	6	19	15	0.997	0.3180
						Co	363	NA	NA	116	170	77
Lee YG	2010	Korea	2000.1–2008.12	6	NA	AA	80	41/39	38(26–49)	61	15	3	1.080	0.2986
						Co	84	44/40	37(26–49)	66	17	0
Zayed RA	2015	Egypt	NA	7	PCR-PFLP	AA	40	20/20	11(2–50)	9	17	14	5.181	0.0228
						Co	40	21/19	12(2–50)	19	12	9

AA: aplastic anemia group; Co: control group; NOS: Newcastle-Ottawa Scale; N: The total number of including; ARMS-PCR: amplification refractory mutation system-polymerase chain reaction; PCR-SSP: Polymerase chain reaction-sequence-specific primer; PCR-PFLP: polymerase chain reaction-restriction fragment length polymorphism; *: likelihood-ratio Χ²; HWE: Hardy-Weinberg equilibrium tests of control group.

Meta-analysis for relationship between polymorphisms of IFN-γ and aplastic anemia

As shown in Figure 2 and Table 2, no significant heterogeneity (I² ≤ 50%, P ≥ 0.05) among studies was observed in all the analysis models, so the fixed effects model was used.

Figure 2. Overall meta-analyses for assessing relationship of IFN-γ polymorphisms with aplastic anemia risk.

Table 2. Results of meta-analysis.

Gene model	Sample size			Test of association			Model	Test of heterogeneity^a,b			Publication bias^c
	K	AA	Co	OR(95%CI)	Z	P		Q	P	I^2 (%)	t	P
T vs A	5	526	1174	2.1749 [1.6825; 2.8114]	5.93	<0.01	F	3.74	0.44	0	0.5748	0.6057
TA vs AA	5	195	486	2.1071 [1.3962; 3.1799]	3.55	<0.01	F	6.25	0.18	36.0	2.4697	0.0901
TT vs AA	5	166	353	4.5788 [2.6606; 7.8797]	5.49	<0.01	F	1.64	0.80	0	0.8015	0.4814
TT vs AA + TA	5	263	587	2.5579 [1.6680; 3.9226]	4.31	<0.01	F	1.52	0.82	0	1.2684	0.2941
TT + TA vs AA	5	263	587	2.5599 [1.7424; 3.7611]	4.79	<0.01	F	6.94	0.14	42.4	1.3346	0.2743

^aRandom-effects model was used when the P for heterogeneity test < 0.05, otherwise the fixed-effect model was used; ^bP <  0.05 is considered statistically significant for Q statis­tics; ^cEgger‘s test to evaluate publication bias, P <  0.05 is consid­ered statistically significant; OR: Odds ratio; CI: confidence interval; K: The number of included studies; AA: aplastic anemia group; Co: control group; F: Fixed-effect model.

In allele genetic model, the results showed that aplastic anemia was significantly associated with the allele T (T vs A: OR = 2.1749, 95% CI = 1.6825–2.8114, P < 0.01). In additive genetic model, the pooled analyses showed that the patients with genotype TA or TT had higher occurrence risk of suffering from aplastic anemia than patients with genotype AA (TA vs AA: OR = 2.1071, 95% CI = 1.3962–3.1799, P < 0.01; TT vs AA: OR = 4.5788, 95% CI = 2.6606–7.8797, P < 0.01). In recessive genetic model, the results showed higher occurrence risk of aplastic anemia in patients with genotype TT than patients with genotype AA or TA (TT vs AA + TA: OR =  2.5579, 95% =  1.6680–3.9226, P <  0.01). Moreover, in the dominant genetic model, the analyses showed lower occurrence risk of aplastic anemia in patients with genotype AA than patients with genotype TT or TA (TT + TA vs AA: OR =  2.5599, 95% =  1.7424–3.7611, P <  0.01). Thus, we conclude that the higher occurrence risk of aplastic anemia is associated with the allele T or genotypes carrying allele T compared with the allele A or genotypes carrying allele A.

Sensitivity analyses and publication bias

In this study, no inconsistent results were found in sensitivity analyses.

Egger’s Test showed there was no significant publication bias (P >  0.05, Table 2)

Discussion

In this study, we found the significant relationship between IFN-γ polymorphisms +874(T/A) and occurrence risk of aplastic anemia. Moreover, genotypes carrying allele T was associated with the high occurrence risk of aplastic anemia.

As known, the level of IFN-γ protein is higher in peripheral blood mononuclear cells and bone marrow cells of aplastic anemia patients than healthy individuals [17,18]. High level of IFN-γ protein may be associated with the occurrence of aplastic anemia by disrupting lineage differentiation and altering hematopoietic stem or progenitor cell composition [4]. Tumor necrosis factor alpha (TNF-α) is another higher expression protein in peripheral blood mononuclear cells and bone marrow cells from aplastic anemia patients [17,18]. Previous study reported that both IFN-γ protein and TNF-α protein could suppress early and late stages of hematopoiesis and induce programed cell death [19]. Moreover, the gene polymorphisms of TNF-α also positively related to aplastic anemia [20]. Thus, the IFN-γ/TNF-α pathway-induced suppression of human hematopoiesis may be the main pathogenesis of aplastic anemia. As shown in the previous study, the high expression of IFN-γ is associated with the genotype carrying allele T, rather than AA [21]. Abnormally high expression of these genes caused by polymorphisms may be the main factor resulting in the startup of IFN-γ/TNF-α pathway, then inducing the occurrence of aplastic anemia.

Furthermore, a previous meta-analysis also analyzed the relationship between IFN-γ polymorphisms +874(T/A) and occurrence risk of aplastic anemia [22]. Differently, this previous study did not find a significant relationship between occurrence risk of aplastic anemia with IFN-γ polymorphisms+874(T/A). The inconsistent result may be caused by different studies included for analysis. In this study, we included the study of Chang H 2010 [10] and El Mahgoub IR 2014 [11], which were not included in that previous meta-analysis. However, the study of Roderick 2013 [23] was not included in this meta-analysis because the frequencies of alleles (T, A) and genotypes (TT, TA, AA) was not reported or cannot be obtained by calculation. Given the controversies and limitations in these meta-analyses, more studies should be performed to confirm the relationship between IFN-γ polymorphisms +874(T/A) and occurrence risk of aplastic anemia.

Limitations of the study

Although it was found in this meta-analysis that IFN-γ polymorphisms +874(T/A) was associated with the occurrence risk of aplastic anemia, there were still some limitations. Firstly, the covariate adjustment and meta-regression were not conducted due to the incomplete data of some studies, especially the demographic data, which might affect the results of the meta-analysis as potential confounders. Besides, the subgroup analysis was not performed due to the small total sample size and number of included studies. Secondly, it was failed to study the interactions between genes and the environment for lack of relevant data. Also, there might exist language bias because only eligible studies with English language were included. Thirdly, although no significant heterogeneity among included studies was found in this study, the impact factors of the relationship between IFN-γ polymorphisms +874(T/A) and occurrence risk of aplastic anemia must be investigated in further studies to confirm the stability of this relationship. In addition, the other gene polymorphisms (such as TNF-α) may be also associated with occurrence risk of aplastic anemia and there may be interactions between these genes [24,25]. Thus, the effect of these gene polymorphisms on occurrence risk of aplastic anemia should be investigated and updated based on more high-quality studies with a large sample size in further studies.

Conclusion

In conclusion, IFN-γ polymorphisms +874(T/A) is associated with the occurrence risk of aplastic anemia. More studies are needed to further confirm the results of this meta-analysis.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

National Natural Science Foundation of China [grant number 81903959]; Natural Science Foundation of ZheJiang Province of China [grant number LQ18H080003]; Medical and Health Science and Technology Project of ZheJiang Province [grant number 2016KYA151].