Effect of polymorphism on IL1A to cancer susceptibility: Evidence based on 34,016 subjects

Abstract

IL-1α is closely related to the development and metastasis of cancer, and its polymorphisms have been reported affecting the susceptibility of malignancy tumors, yet the conclusions are controversial. Present an overall meta-analysis was performed to reach more general findings. Relevant literature was searched from Google Scholar, Web of Science, PubMed, EMBASE, CNKI database and Wanfang database, and the association among polymorphism and cancer risk was appraised by counting ORs and 95% CIs of overall and stratification analysis. The date from 15,586 cases and 18,430 controls in 40 publications were enrolled. There is significantly upregulated risk leads by rs3783553 in all genetic models, while the same tendency was found in all cancer types. The results also suggest a high risk of cancer susceptibility in patients carried rs1800587 polymorphism, of which draw out form allelic and homozygote models in overall studies, especially for cervical cancer. However, there are no significant results in analysis of rs17561. In a word, IL1A SNPs play an essential role in upregulating cancer susceptibility, and current analysis provides detail date for further study in the future.

Keywords:

• IL1A
• cancer risk
• meta-analysis
• gene polymorphism
• rs1800587
• rs3783553

Introduction

Worldwide, the incidence and mortality of cancer continue to rise year by year. Due to many intricate reasons, of which are related to the growth of population, aging, and environmental implication, cancer became the primary reason for death and resulted in more than 9.6 million people lost their lives in 2018. Across all cancer ranges, the incidence of malignancy tumor is about 20% higher in male than in female. Lung cancer is associated with the first incidence and mortality in male, and prostate cancer, the most widespread malignancy tumor among countries, is not far behind. It is estimated that in 2019, 174,650 people will newly suffer from prostate cancer in the United States, and there are about 31,620 confirmed patients meet the finally death. For female, the incidence of breast cancer is head and shoulder above any other cancers, followed by colorectal and cervical cancer, nonetheless, lung cancer is still the primary strength which contributes to the most of cancer-death^[1,2].

Interleukin-1(IL-1), cognized firstly in all Interleukins, has been studied for a long history. It is well known to us all, as indispensable members of cytokines in our body, Interleukin-1 contains IL-1 alpha(IL-1α), IL-1 beta(IL-1β) and IL-1 receptor antagonist(IL-1Ra), which are edited by different genes and interact with each other to influence innate immunity and acquired inflammation^[3]. In the past few years, a lot of researches pay attention to the function of IL-1α. IL-1α is mainly produced by monocytes and macrophages, which play the role of proproteins in the inflammation, and may contribute to tumor growth, invasion and distant metastasis in different cancers ^[4]. Previous studies have demonstrated that IL-1α is significantly related to the risk of prostate cancer, lung cancer, colon cancer, cervical cancer, and breast cancer, etc. ^[4–13]. These come to conclusions that IL-1α may be responsible for cancer development and treatment.

IL1A, which encodes IL-1α protein, locates at chromosome 2q and consists of 7 exons and 6 introns. rs17561, rs1800587, and rs3783553 are three SNPs on IL1A gene, which are believed as functional roles. rs17561 is a mutation at +4845 of 5^th exon of IL1A, which leads to the G-to-T variation, as well as the change of amino acid^[14]. Located at -889 of IL1A promoter region, rs1800587(C-to-T) was reported might affect the expression of IL-1α^[15,16]. In addition, SNP occurred within the IL1A 3’-untranslated regions (UTR) may affects the expression of IL-1α by influencing the binding of miR-122. That’s an insertion/deletion variation, what we called rs3783553.

Recently, relationship between the polymorphisms of IL1A and different types of cancer has been reported by many researchers, but the conclusions are still not consistent. Hence, an update mete-analysis among all cancers has been carried out, to further discover the effects of IL1A polymorphism on tumors.

Methods

Retrieval of literature

As of May 20, 2019, comprehensive searches of relevant literatures were carried out from several online databases: Web of Science, Google Scholar, PubMed, EMBASE, CNKI database, and Wanfang database. For purpose of hitting more accurate studies mentioning the IL1A polymorphisms and cancer risks, we used “IL1A OR (IL-1 Alpha) OR (Interleukin 1 Alpha) OR (Interleukin-1 Alpha) OR Pro-Interleukin-1-Alpha OR (Preinterleukin 1 Alpha) OR Hematopoietin-1” AND “tumor OR cancer OR carcinoma OR neoplasm” AND “SNP OR variant OR mutation OR polymorphism” as the search items. Moreover, references from selected articles were also searched and viewed for adding any other relevant studies. All the processes were completed along with the PRISMA 2009 statement^[17].

Criteria for identification

Inclusions should meet the following criteria after checking the full-text: (1) English or Chinese articles, (2) case-control studies concerning the link between cancer susceptibility and IL1A polymorphism, (3) the studies performed in patients diagnosed with cancer, (4) the data of genotype frequency from cases and controls are sufficient reported or (5) can be calculated from other information in the article.

Otherwise, if the above conditions were not met, or had the following characteristics: (1) repeated studies using overlapping data, (2) the studies aim at animals or cell lines, (3) genotype frequency cannot be obtained, the article would be excluded.

Quality assessment

Researchers assessed the quality of study by putting into force the Newcastle–Ottawa Scale (NOS)^[18]. It can be explained by the following standard rules: (1) “high” quality choices were identified with a “star”, (2) each numbered item within the Selection and Exposure categories will acquire up to 1 star maximumly, (3) each study of case-control comparability could be awarded a maximum of 2 stars, (4) “*” means “Yes”, while “NA” means “not applicable”. Two authors give a to unified dissents after comprehensive discussion.

Date extraction

When the articles passed the selection criteria, two authors extracted the date we need from these articles, to ensure the accuracy of the recorded data. The extracted items include: the year of publication, last name of the first author, numbers and genotype frequency of controls and cases, cancer type, genotyping method, ethnicity and source of control. If the date was inconsistent or controversial, the two authors would check the articles, and determine the final data after discussion.

Statistical methods

Firstly, we generated 5 analysis models based on the allele frequency in case group and control group, including allele contrast model, homozygote comparison model, heterozygote comparison model, dominant comparison model, and recessive comparison model. They were (G vs. T), (G/G vs. T/T), (G/T vs. T/T), (G/G + G/T vs. T/T), (G/G vs. G/T + T/T) for rs17561; (C vs. T), (C/C vs. T/T), (C/T vs. T/T), (C/C + C/T vs. T/T), (C/C vs. C/T + T/T) for rs1800587 and (Ins vs. Del), (Ins/Ins vs. Del/Del), (Ins/Del vs. Del/Del), (Ins/Ins + Ins/Del vs. Del/Del), (Ins/Ins vs. Ins/Del + Del/Del) for rs3783553. Then, STATA software (Stata 12.0, College Station, Texas) was run to execute statistical calculations, and the entire inspection process were all two-tailed. We also performed χ2 test to checkout Hardy–Weinberg equilibrium (HWE) of control group, which can compare actual frequencies of genotype with expected value. We used odds ratios (ORs) with 95% confidence intervals (CIs) to demonstrate the computation result of each genic model in overall and stratified groups to appraise the correlation between IL1A polymorphism and cancer susceptibility. The subgroups were stratified by cancer type, source of control, ethnicity and whether comfort to HWE. Meanwhile, the chi-squared (χ²)-based Q test was employed to assess heterogeneity, which was regarded as significant when P < 0.1^[19]. For each group stratified date, if existed heterogeneity (P < 0.1)^[20], we used random-effects model to pool the date, otherwise fixed-effects model would be applied^[21]. Furthermore, p values from Egger’s regression test and symmetrical feature from Begg’s funnel plot were obtained to evaluate whether there is a significant bias. Above all analysis, if p values ≤ 0.05, the results were considered as statistically significant.

Results

Literature searches and characteristics

Figure 1 exhibits the course of screening process from searching to including. 317 relevant articles were downloaded from all database, as well as 23 from their references were also enrolled. We initially obtained 323 studies from getting rid of the duplicated, and then, through screening and checking the titles and abstracts, we continued to remove 260 articles which were significantly different from our objective. Finally, we included 40 eligibilities by applying the selection criteria after browsing and reading the full text of the rest 63 articles.

Figure 1. Flowchart of the enrolled studies selection procedure.

The characteristics of the enrolled researches on IL1A polymorphism and cancers were showed in Table 1. There were 5 studies about rs17561, 22 about rs1800587, and 18 about rs1783553(one article can contain more than one study about IL1A polymorphisms). Among all studies, the results of heterogeneity test were showed in Table 1. For studies on rs17561, they respectively investigated the ovarian cancer, glioma, cervical cancer, nasopharyngeal cancer, and lung cancer ^[22–26]. As for rs1800587, there are 2 studies for each type of cancer on ovarian cancer, cervical cancer, breast cancer, odontogenic tumor, nasopharyngeal cancer, lymphoma, and lung cancer. The rest studies put research centers on kaposi sarcoma, vulvar carcinoma, multiple myeloma, gastric cancer, prostate cancer, esophageal cancer, and glioma^[22,25–44]. In term of rs3783553, it is a contention of a hundred schools of thought. It includes researches on multiple systems of cancer diseases, such as hepatocellular carcinoma, nasopharyngeal cancer, glioma, prostate cancer, gastric cancer, thyroid carcinoma, oral squamous cell carcinoma, cervical cancer, ovarian cancer, colorectal cancer, colorectal cancer, and endometrial cancer^[37,45–61]. Analyzed by ethnicity, most of studies focus on Asians, yet researches direct to Caucasians are also quite a lot. Additionally, Bushley et al, Foster et al, and Rothman et al did the study from mixed races^[25,33,35]. On the whole, all the cases add up to 15,586 for patients (case group), as well as 18,430 for healthy subjects (control group). The whole sample size is large, and we could make our study get results more accurately and convincingly.

Table 1. Characteristics of the enrolled studies on IL-1A polymorphism and cancer.

SNP	First author	Year	Ethnicity	Genotyping Method	Source of Control	Cancer Type	Cases			Controls
							1/1	1/2	2/2	1/1	1/2	2/2	HWE
rs17561	Bushley et al.	2004	Mixed	PCR	HB	Ovarian Cancer	146	30	6	151	61	7	Y
rs17561	Ai et al.	2014	Asian	PCR-RFLP	PB	Glioma	118	34	2	169	25	1	Y
rs17561	Zidi et al.	2015	Africa	PCR-RFLP	HB	Cervical Cancer	12	46	30	50	124	86	Y
rs17561	Sousa et al.	2016	Caucasian	TaqMan	HB	Nasopharyngeal Cancer	73	39	10	369	279	55	Y
rs17561	Eaton et al.	2018	Caucasian	Taqman	PB	Lung cancer	303	272	41	316	249	51	Y
rs1800587	Foster et al.	2000	Mixed	PCR	HB	Kaposi Sarcoma	52	47	12	67	45	11	Y
rs1800587	Grimm et al.	2004	Caucasian	PCR	HB	Vulvar Carcinoma	27	31	10	101	114	13	N
rs1800587	Bushley et al.	2004	Mixed	PCR	HB	Ovarian Cancer	140	32	6	150	59	7	Y
rs1800587	Wang et al.	2005	Asian	PCR	HB	Cervical Cancer	18	22	10	27	20	3	Y
rs1800587	Wang et al.	2005	Asian	PCR	PB	Cervical Cancer	27	15	8	23	24	3	Y
rs1800587	Hefler et al.	2005	Caucasian	PCR	PB	Breast Cancer	123	124	22	102	112	13	N
rs1800587	Snoussi et al.	2005	Caucasian	PCR-RFLP	PB	Breast Cancer	116	132	57	83	82	35	Y
rs1800587	Rothman et al.	2006	Mixed	PCR	PB	Lymphoma	1494	1306	253	1740	1436	281	Y
rs1800587	Abazis-Stamboulieh et al.	2007	Caucasian	PCR-RFLP	PB	Multiple Myeloma	27	38	9	116	20	24	N
rs1800587	Braicu et al.	2007	Caucasian	PCR	HB	Ovarian Cancer	65	68	14	65	55	9	Y
rs1800587	Hou et al.	2007	Caucasian	PCR	PB	Gastric Cancer	143	119	43	184	174	59	Y
rs1800587	Ennas et al.	2008	Caucasian	TaqMan	PB	Chronic Lymphocytic Leukaemia	21	15	4	61	42	8	Y
rs1800587	Sáenz-López et al.	2008	Caucasian	PCR	PB	Prostate Cancer	133	135	28	152	122	29	Y
rs1800587	Hoeft et al.	2008	Caucasian	PCR-RFLP	PB	Lymphoma	330	274	46	335	274	50	Y
rs1800587	Sengüven et al.	2011	Caucasian	PCR-RFLP	HB	Odontogenic Tumor	44	20	10	16	10	4	Y
rs1800587	Yang et al.	2011	Asian	PCR-RFLP	PB	Nasopharyngeal Cancer	121	101	26	157	121	18	Y
rs1800587	Eshghyar et al.	2012	Asian	PCR-SSP	PB	Odontogenic Tumor	6	21	3	38	85	22	N
rs1800587	Bai et al.	2013	Asian	TaqMan	HB	Lung cancer	4	26	163	4	33	173	Y
rs1800587	Zheng et al.	2013	Asian	PCR	HB	Esophageal cancer	310	54	2	298	67	5	Y
rs1800587	Ai et al. et al.	2014	Asian	PCR-RFLP	PB	Glioma	114	39	1	151	44	0	Y
rs1800587	Qu et al.	2014	Asian	TaqMan	HB	Nasopharyngeal Cancer	130	29	24	182	46	3	Y
rs1800587	Eaton et al.	2018	Caucasian	Taqman	PB	Lung cancer	299	269	44	312	249	51	Y
rs3783553	Gao et al.	2009	Asian	PCR	PB	HCC	149	676	652	231	815	627	Y
rs3783553	He et al.	2010	Asian	PCR	PB	HCC	22	190	191	58	218	158	Y
rs3783553	Yang et al.	2012	Asian	PCR-RFLP	PB	Nasopharyngeal Cancer	28	106	114	56	132	108	Y
rs3783553	Si et al.	2013	Asian	PCR	PB	Glioma	12	72	77	29	99	78	Y
rs3783553	Liao et al.	2014	Asian	PCR-PAGE	PB	Prostate Cancer	8	53	70	30	118	81	Y
rs3783553	Zeng et al.	2014	Asian	PCR	PB	Gastric Cancer	15	102	90	51	183	147	Y
rs3783553	Gao et al.	2014	Asian	PCR	PB	Thyroid Carcinoma	29	132	112	56	242	211	Y
rs3783553	Zhang et al.	2014	Asian	PCR	PB	Ovarian Cancer	29	126	146	36	110	94	Y
rs3783553	Pu et al.	2014	Asian	PCR	PB	Cervical Cancer	26	140	153	60	201	163	Y
rs3783553	Du et al.	2014	Asian	PCR	PB	HCC	113	451	434	111	463	431	Y
rs3783553	Huang et al.	2015	Asian	PCR	PB	Cervical Cancer	13	134	88	33	147	134	Y
rs3783553	Zhang et al.	2015	Caucasion	PCR	PB	Oral Squamous Cell Carcinoma	35	144	146	60	163	112	Y
rs3783553	Yan et al.	2015	Asian	PCR	PB	Colorectal Cancer	37	160	142	46	143	124	Y
rs3783553	Huang et al.	2016	Asian	PCR	PB	Breast Cancer	24	113	91	45	115	81	Y
rs3783553	Yu et al.	2016	Asian	PCR-PAGE	PB	Endometrial Cancer	26	127	119	45	132	117	Y
rs3783553	Zhang et al.	2016	Asian	PCR	PB	Gastric Cancer	46	244	229	80	258	198	Y
rs3783553	Hashemi et al.	2017	Caucasion	PCR-RFLP	HB	Prostate Cancer	63	71	16	83	65	7	Y
rs3783553	Ma et al.	2018	Asian	PCR	PB	Colorectal Cancer	28	176	178	70	200	163	Y

1/1: wild homozygote; 1/2: heterozygote; 2/2: mutant homozygote.

Results of NOS and HWE checking

We performed Newcastle–Ottawa Scale in evaluating the overall quality of enrolled studies. Table S1 showed that all included studies scored more than 5 stars, and more than 70% of them were higher than 7 stars. The maximum score was 8 stars, with 5 cases. Therefore, the overall quality of the study is at a high level. For three IL1A polymorphisms, showed in Table 1, apart from 4 studies on rs1800587 from Grimm et al, Hefler et al, Abazis-Stamboulieh et al, and Eshghyar et al^{[32,34,40,42]}, all studies gone through the test of HWE. Mentionable, the source of control in Grimm et al ’s study was based on hospital, the other three were patient based. These four studies all received high stars in NOS, which got 7 stars, 7 stars, 8 stars and 6 stars respectively. We consider that although they did not meet the HWE, their quality may still be high to ensure their reliability and significance.

Association between IL1A +4845- G > T (rs17561) and cancer susceptibility

As showed in Figure S1 and Table 2, the overall p values and ORs demonstrate no significant rising or declining link between IL1A rs17561 - G > T SNP and cancer risk, whatever the genotype (G vs. T: P = 0.995 OR= 1, 95% CI= 0.77– 1.3; G/G vs. T/T: P = 0.855, OR= 0.97, 95% CI= 0.71– 1.33; G/T vs. T/T: P = 0.955, OR= 1.01, 95% CI= 0.66– 1.54; G/G + G/T vs. T/T: P = 0.923, OR= 1.02, 95% CI= 0.69– 1.5; G/G vs. G/T + T/T: P = 0.647, OR= 0.94, 95% CI= 0.7– 1.24). And further stratification by cancer type, ethnicity, and control of source showed meaningless results, too.

Table 2. Results of pooled analysis for IL-1A polymorphism and cancer susceptibility.

SNP	Comparison	Subgroup	N	PH	PZ	Random	Fixed
rs17561	G vs. T	Overall	5	0.013	0.995	0.999 (0.771–1.295)	0.987 (0.869–1.121)
rs17561	G/G vs. T/T	Overall	5	0.673	0.855	0.967 (0.703–1.33)	0.971 (0.708–1.331)
rs17561	G/T vs. T/T	Overall	5	0.001	0.955	1.012 (0.664–1.542)	1.013 (0.854–1.201)
rs17561	G/G + G/T vs. T/T	Overall	5	0.002	0.923	1.019 (0.693–1.499)	1 (0.849–1.177)
rs17561	G/G vs. G/T + T/T	Overall	5	0.81	0.647	0.935 (0.703–1.243)	0.935 (0.703–1.244)
rs17561	G vs. T	Caucasian	2	0.292	0.655	0.962 (0.814–1.137)	0.966 (0.829–1.125)
rs17561	G/G vs. T/T	Caucasian	2	0.831	0.432	0.86 (0.591–1.251)	0.86 (0.59–1.253)
rs17561	G/T vs. T/T	Caucasian	2	0.051	0.75	0.927 (0.583–1.474)	1.015 (0.829–1.242)
rs17561	G/G + T/T vs. T/T	Caucasian	2	0.095	0.701	0.93 (0.643–1.346)	0.989 (0.815–1.2)
rs17561	G/G vs. G/T + T/T	Caucasian	2	0.495	0.395	0.853 (0.592–1.229)	0.853 (0.59–1.231)
rs17561	G vs. T	PB	2	0.025	0.373	1.317 (0.719–2.413)	1.08 (0.915–1.274)
rs17561	G/G vs. T/T	PB	2	0.326	0.546	0.872 (0.566–1.345)	0.876 (0.569–1.348)
rs17561	G/T vs. T/T	PB	2	0.087	0.2	1.396 (0.838–2.325)	1.232 (0.993–1.527)
rs17561	G/G + G/T vs. T/T	PB	2	0.05	0.266	1.388 (0.779–2.474)	1.183 (0.962–1.455)
rs17561	G/G vs. G/T + T/T	PB	2	0.347	0.358	0.819 (0.538–1.247)	0.822 (0.54–1.249)
rs17561	G vs. T	HB	3	0.088	0.338	0.855 (0.621–1.177)	0.865 (0.707–1.057)
rs17561	G/G vs. T/T	HB	3	0.635	0.701	1.092 (0.681–1.75)	1.095 (0.689–1.739)
rs17561	G/T vs. T/T	HB	3	0.043	0.357	0.775 (0.451–1.333)	0.722 (0.542–0.96)
rs17561	G/G + G/T vs. T/T	HB	3	0.055	0.373	0.8 (0.49–1.306)	0.754 (0.576-0.986)
rs17561	G/G vs. G/T + T/T	HB	3	1	0.818	1.046 (0.71–1.541)	1.046 (0.71–1.541)
rs1800587	C vs. T	Overall	22	0.002	0.018	1.119 (1.019–1.228)	1.069 (1.016–1.124)
rs1800587	C/C vs. T/T	Overall	22	0.05	0.021	1.259 (1.035–1.53)	1.16 (1.028–1.309)
rs1800587	C/T vs. T/T	Overall	22	<0.001	0.294	1.077 (0.937–1.238)	1.061 (0.991–1.136)
rs1800587	C/C + C/T vs. T/T	Overall	22	0.002	0.075	1.117 (0.989–1.261)	1.078 (1.01–1.15)
rs1800587	C/C vs. C/T + T/T	Overall	22	0.032	0.06	1.195 (0.992–1.439)	1.116 (0.996–1.25)
rs1800587	C vs. T	Asians	8	0.008	0.104	1.234 (0.958–1.59)	1.221 (1.057–1.411)
rs1800587	C/C vs. T/T	Asians	8	0.027	0.054	2.04 (0.989–4.209)	2.293 (1.546–3.401)
rs1800587	C/T vs. T/T	Asians	8	0.459	0.805	0.975 (0.803–1.184)	0.976 (0.805–1.184)
rs1800587	C/C + C/T vs. T/T	Asians	8	0.19	0.231	1.135 (0.894–1.44)	1.118 (0.931–1.343)
rs1800587	C/C vs. C/T + T/T	Asians	8	0.007	0.056	1.898 (0.985–3.658)	1.8 (1.319–2.457)
rs1800587	C vs. T	Caucasian	11	0.08	0.084	1.103 (0.987–1.233)	1.072 (0.99–1.161)
rs1800587	C/C vs. T/T	Caucasian	11	0.626	0.325	1.101 (0.916–1.324)	1.097 (0.912–1.318)
rs1800587	C/T vs. T/T	Caucasian	11	<0.001	0.143	1.193 (0.942–1.512)	1.117 (1.001–1.246)
rs1800587	C/C + C/T vs. T/T	Caucasian	11	0.002	0.102	1.172 (0.969–1.419)	1.11 (1.001–1.232)
rs1800587	C/C vs. C/T + T/T	Caucasian	11	0.587	0.66	1.045 (0.876–1.245)	1.04 (0.873–1.239)
rs1800587	C vs. T	PB	13	0.178	0.049	1.075 (0.995–1.162)	1.056 (1–1.115)
rs1800587	C/C vs. T/T	PB	13	0.806	0.246	1.078 (0.947–1.228)	1.079 (0.949–1.228)
rs1800587	C/T vs. T/T	PB	13	<0.001	0.12	1.149 (0.965–1.369)	1.084 (1.008–1.167)
rs1800587	C/C + C/T vs. T/T	PB	13	0.005	0.076	1.135 (0.987–1.305)	1.083 (1.01–1.162)
rs1800587	C/C vs. C/T + T/T	PB	13	0.672	0.58	1.034 (0.913–1.171)	1.036 (0.915–1.172)
rs1800587	C vs. T	HB	9	0.001	0.227	1.176 (0.904–1.529)	1.16 (1.009–1.334)
rs1800587	C/C vs. T/T	HB	9	0.016	0.065	1.748 (0.966–3.162)	1.957 (1.375–2.785)
rs1800587	C/T vs. T/T	HB	9	0.294	0.388	0.931 (0.752–1.153)	0.92 (0.762–1.111)
rs1800587	C/C + C/T vs. T/T	HB	9	0.044	0.578	1.078 (0.827–1.407)	1.045 (0.876–1.247)
rs1800587	C/C vs. C/T + T/T	HB	9	0.013	0.046	1.69 (1.01–2.827)	1.687 (1.258–2.263)
rs1800587	C vs. T	Breast Cancer	2	0.758	0.465	1.072 (0.889–1.293)	1.072 (0.889–1.293)
rs1800587	C/C vs. T/T	Breast Cancer	2	0.683	0.315	1.237 (0.815–1.877)	1.238 (0.816–1.878)
rs1800587	C/T vs. T/T	Breast Cancer	2	0.409	0.886	1.02 (0.78–1.334)	1.02 (0.78–1.333)
rs1800587	C/C + C/T vs. T/T	Breast Cancer	2	0.496	0.677	1.056 (0.819–1.361)	1.056 (0.819–1.361)
rs1800587	C/C vs. C/T + T/T	Breast Cancer	2	0.485	0.384	1.186 (0.804–1.75)	1.188 (0.806–1.752)
rs1800587	C vs. T	Cervical Cancer	2	0.118	0.07	1.471 (0.759–2.854)	1.475 (0.968–2.247)
rs1800587	C/C vs. T/T	Cervical Cancer	2	0.445	0.017	3.386 (1.232–9.308)	3.396 (1.242–9.285)
rs1800587	C/T vs. T/T	Cervical Cancer	2	0.065	0.909	0.938 (0.309–2.841)	0.937 (0.52–1.69)
rs1800587	C/C + C/T vs. T/T	Cervical Cancer	2	0.065	0.699	1.227 (0.436–3.455)	1.219 (0.702–2.116)
rs1800587	C/C vs. C/T + T/T	Cervical Cancer	2	0.784	0.013	3.43 (1.299–9.06)	3.439 (1.304-9.072)
rs1800587	C vs. T	Lung cancer	2	0.73	0.681	1.035 (0.878–1.22)	1.035 (0.878–1.22)
rs1800587	C/C vs. T/T	Lung cancer	2	0.952	0.632	0.904 (0.597–1.367)	0.904 (0.597–1.367)
rs1800587	C/T vs. T/T	Lung cancer	2	0.639	0.347	1.117 (0.887–1.408)	1.117 (0.887–1.408)
rs1800587	C/C + C/T vs. T/T	Lung cancer	2	0.813	0.475	1.084 (0.869–1.353)	1.084 (0.869–1.353)
rs1800587	C/C vs. C/T + T/T	Lung cancer	2	0.367	0.813	0.961 (0.692–1.335)	0.961 (0.693–1.334)
rs1800587	C vs. T	Lymphoma	2	0.609	0.402	1.03 (0.961–1.104)	1.03 (0.961–1.104)
rs1800587	C/C vs. T/T	Lymphoma	2	0.626	0.732	1.03 (0.87–1.219)	1.03 (0.87–1.219)
rs1800587	C/T vs. T/T	Lymphoma	2	0.737	0.288	1.052 (0.958–1.154)	1.052 (0.958–1.154)
rs1800587	C/C + C/T vs. T/T	Lymphoma	2	0.66	0.3	1.048 (0.959–1.145)	1.048 (0.959–1.145)
rs1800587	C/C vs. C/T + T/T	Lymphoma	2	0.677	0.94	1.006 (0.855–1.185)	1.006 (0.855–1.185)
rs1800587	C vs. T	Odontogenic Tumor	2	0.705	0.825	0.952 (0.622–1.459)	0.953 (0.621–1.461)
rs1800587	C/C vs. T/T	Odontogenic Tumor	2	0.959	0.812	0.889 (0.336–2.356)	0.889 (0.336–2.351)
rs1800587	C/T vs. T/T	Odontogenic Tumor	2	0.272	0.842	1.055 (0.498–2.236)	1.071 (0.547–2.096)
rs1800587	C/C + C/T vs. T/T	Odontogenic Tumor	2	0.362	0.94	1.014 (0.534–1.923)	1.025 (0.546–1.923)
rs1800587	C/C vs. C/T + T/T	Odontogenic Tumor	2	0.588	0.588	0.799 (0.328–1.949)	0.787 (0.331–1.873)
rs1800587	C vs. T	Ovarian Cancer	2	0.039	0.798	0.929 (0.531–1.628)	0.948 (0.725–1.239)
rs1800587	C/C vs. T/T	Ovarian Cancer	2	0.472	0.511	1.262 (0.625–2.547)	1.264 (0.629–2.537)
rs1800587	C/T vs. T/T	Ovarian Cancer	2	0.033	0.659	0.847 (0.404–1.774)	0.84 (0.596–1.183)
rs1800587	C/C + C/T vs. T/T	Ovarian Cancer	2	0.03	0.743	0.887 (0.433–1.816)	0.877 (0.632–1.216)
rs1800587	C/C vs. C/T + T/T	Ovarian Cancer	2	0.679	0.516	1.252 (0.631–2.486)	1.254 (0.633–2.482)
rs1800587	C vs. T	Y	18	0.02	0.091	1.082 (0.988–1.186)	1.055 (1.001–1.111)
rs1800587	C/C vs. T/T	Y	18	0.05	0.084	1.206 (0.975–1.49)	1.131 (0.998–1.282)
rs1800587	C/T vs. T/T	Y	18	0.453	0.279	1.04 (0.969–1.117)	1.04 (0.969–1.116)
rs1800587	C/C + C/T vs. T/T	Y	18	0.279	0.107	1.059 (0.97–1.155)	1.057 (0.988–1.131)
rs1800587	C/C vs. C/T + T/T	Y	18	0.047	0.109	1.173 (0.965–1.426)	1.105 (0.982–1.244)
rs1800587	C vs. T	N	4	0.021	0.113	1.338 (0.934–1.918)	1.27 (1.054–1.531)
rs1800587	C/C vs. T/T	N	4	0.511	0.037	1.657 (1.049–2.617)	1.628 (1.031–2.571)
rs1800587	C/T vs. T/T	N	4	<0.001	0.23	1.827 (0.683–4.887)	1.4 (1.077–1.821)
rs1800587	C/C + C/T vs. T/T	N	4	<0.001	0.183	1.647 (0.791–3.432)	1.413 (1.099–1.817)
rs1800587	C/C vs. C/T + T/T	N	4	0.112	0.275	1.274 (0.681–2.386)	1.263 (0.83–1.92)
rs3783553	Ins vs. Del	Overall	18	0.003	<0.001	1.3 (1.205–1.402)	1.269 (1.209–1.331)
rs3783553	Ins/Ins vs. Del/Del	Overall	18	0.006	<0.001	1.879 (1.582–2.23)	1.752 (1.57–1.954)
rs3783553	Ins/Del vs. Del/Del	Overall	18	0.235	<0.001	1.49 (1.317–1.687)	1.457 (1.31–1.621)
rs3783553	Ins/Ins + Ins/Del vs. Del/Del	Overall	18	0.046	<0.001	1.659 (1.443–1.907)	1.585 (1.432–1.754)
rs3783553	Ins/Ins vs. Ins/Del + Del/Del	Overall	18	0.014	<0.001	1.308 (1.186–1.443)	1.285 (1.202–1.372)
rs3783553	Ins vs. Del	Asian	16	0.003	<0.001	1.281 (1.183–1.387)	1.254 (1.193–1.318)
rs3783553	Ins/Ins vs. Del/Del	Asian	16	0.005	<0.001	1.835 (1.529–2.202)	1.715 (1.531–1.92)
rs3783553	Ins/Del vs. Del/Del	Asian	16	0.144	<0.001	1.512 (1.312–1.742)	1.455 (1.301–1.628)
rs3783553	Ins/Ins + Ins/Del vs. Del/Del	Asian	16	0.025	<0.001	1.666 (1.423–1.949)	1.573 (1.413–1.75)
rs3783553	Ins/Ins vs. Ins/Del + Del/Del	Asian	16	0.022	<0.001	1.281 (1.16–1.414)	1.266 (1.183–1.354)
rs3783553	Ins vs. Del	Caucasian	2	0.903	<0.001	1.501 (1.243–1.813)	1.501 (1.243–1.813)
rs3783553	Ins/Ins vs. Del/Del	Caucasian	2	0.582	<0.001	2.377 (1.545–3.658)	2.382 (1.549–3.664)
rs3783553	Ins/Del vs. Del/Del	Caucasian	2	0.881	0.022	1.476 (1.058–2.06)	1.476 (1.058–2.06)
rs3783553	Ins/Ins + Ins/Del vs. Del/Del	Caucasian	2	0.696	0.001	1.697 (1.234–2.334)	1.699 (1.236–2.335)
rs3783553	Ins/Ins vs. Ins/Del + Del/Del	Caucasian	2	0.373	<0.001	1.701 (1.263–2.292)	1.707 (1.268–2.298)
rs3783553	Ins vs. Del	PB	17	0.003	<0.001	1.293 (1.196–1.397)	1.264 (1.204–1.327)
rs3783553	Ins/Ins vs. Del/Del	PB	17	0.005	<0.001	1.855 (1.559–2.208)	1.738 (1.557–1.941)
rs3783553	Ins/Del vs. Del/Del	PB	17	0.186	<0.001	1.503 (1.317–1.716)	1.458 (1.307–1.626)
rs3783553	Ins/Ins + Ins/Del vs. Del/Del	PB	17	0.032	<0.001	1.67 (1.44–1.936)	1.585 (1.428–1.758)
rs3783553	Ins/Ins vs. Ins/Del + Del/Del	PB	17	0.017	<0.001	1.298 (1.178–1.43)	1.28 (1.198–1.367)
rs3783553	Ins vs. Del	Cervical Cancer	2	0.039	0.316	1.193 (0.845–1.685)	1.217 (1.032–1.435)
rs3783553	Ins/Ins vs. Del/Del	Cervical Cancer	2	0.552	0.001	1.977 (1.31–2.983)	1.976 (1.309–2.982)
rs3783553	Ins/Del vs. Del/Del	Cervical Cancer	2	0.402	0.003	1.83 (1.217–2.752)	1.836 (1.222–2.758)
rs3783553	Ins/Ins + Ins/Del vs. Del/Del	Cervical Cancer	2	0.855	0.001	1.908 (1.289–2.823)	1.908 (1.289–2.824)
rs3783553	Ins/Ins vs. Ins/Del + Del/Del	Cervical Cancer	2	0.009	0.76	1.097 (0.605–1.989)	1.142 (0.914–1.428)
rs3783553	Ins vs. Del	Colorectal Cancer	2	0.093	0.042	1.306 (1.01–1.688)	1.317 (1.131–1.533)
rs3783553	Ins/Ins vs. Del/Del	Colorectal Cancer	2	0.066	0.037	1.975 (1.043–3.739)	2.001 (1.419–2.821)
rs3783553	Ins/Del vs. Del/Del	Colorectal Cancer	2	0.19	0.001	1.752 (1.118–2.746)	1.766 (1.256–2.483)
rs3783553	Ins/Ins + Ins/Del vs. Del/Del	Colorectal Cancer	2	0.099	0.025	1.852 (1.08–3.177)	1.872 (1.354–2.588)
rs3783553	Ins/Ins vs. Ins/Del + Del/Del	Colorectal Cancer	2	0.2	0.02	1.272 (0.973–1.663)	1.279 (1.039–1.576)
rs3783553	Ins vs. Del	Gastric Cancer	2	0.776	<0.001	1.315 (1.137–1.522)	1.315 (1.137–1.522)
rs3783553	Ins/Ins vs. Del/Del	Gastric Cancer	2	0.929	<0.001	2.032 (1.441–2.866)	2.033 (1.441–2.867)
rs3783553	Ins/Del vs. Del/Del	Gastric Cancer	2	0.709	0.002	1.715 (1.222–2.405)	1.718 (1.225–2.409)
rs3783553	Ins/Ins + Ins/Del vs. Del/Del	Gastric Cancer	2	0.8	<0.001	1.853 (1.34–2.563)	1.856 (1.342–2.566)
rs3783553	Ins/Ins vs. Ins/Del + Del/Del	Gastric Cancer	2	0.656	0.009	1.305 (1.068–1.594)	1.305 (1.068–1.594)
rs3783553	Ins vs. Del	HCC	3	0.002	0.049	1.231 (1.001–1.514)	1.198 (1.111–1.292)
rs3783553	Ins/Ins vs. Del/Del	HCC	3	<0.001	0.067	1.645 (0.966–2.801)	1.474 (1.242–1.749)
rs3783553	Ins/Del vs. Del/Del	HCC	3	0.015	0.145	1.333 (0.906–1.963)	1.243 (1.049–1.473)
rs3783553	Ins/Ins + Ins/Del vs. Del/Del	HCC	3	0.002	0.091	1.474 (0.94–2.312)	1.345 (1.144–1.58)
rs3783553	Ins/Ins vs. Ins/Del + Del/Del	HCC	3	0.018	0.038	1.265 (1.014–1.578)	1.24 (1.119–1.375)
rs3783553	Ins vs. Del	Prostate Cancer	2	0.539	<0.001	1.655 (1.3–2.106)	1.657 (1.302–2.108)
rs3783553	Ins/Ins vs. Del/Del	Prostate Cancer	2	0.91	<0.001	3.137 (1.672–5.888)	3.142 (1.674–5.896)
rs3783553	Ins/Del vs. Del/Del	Prostate Cancer	2	0.749	0.053	1.494 (0.991–2.251)	1.497 (0.994–2.255)
rs3783553	Ins/Ins + Ins/Del vs. Del/Del	Prostate Cancer	2	0.427	0.005	1.74 (1.172–2.584)	1.758 (1.187–2.604)
rs3783553	Ins/Ins vs. Ins/Del + Del/Del	Prostate Cancer	2	0.72	<0.001	2.17 (1.462–3.22)	2.175 (1.466–3.226)

Association between IL1A -889- C > T (rs1800587) and cancer susceptibility

When it comes to rs1800587 polymorphism, we found that there is a statistically significant uplifted overall risk between it and cancer susceptibility only in allelic contrast model (C vs. T: P = 0.018, OR= 1.12, 95% CI= 1.02– 1.23) and homozygote comparison model (CC vs. TT: P = 0.021, OR= 1.26, 95% CI= 1.04– 1.53) (the others: CT vs. TT: P = 0.294, OR= 1.08, 95% CI= 0.94– 1.24; C/C + C/T vs. T/T: P = 0.075, OR= 1.12, 95% CI= 0.99– 1.26; C/C vs. C/T + T/T: P = 0.06, OR = 1.19, 95% CI= 0.99– 1.44; Table 2, Figure 2).

Figure 2. Meta-analysis of the association between IL1A rs1800587 polymorphism and cancer risk.

What is more, the studies stratified by cancer type showed statistical significances in homozygote and recessive two type models (CC vs. TT: P = 0.017, OR= 3.4, 95% CI= 1.24– 9.29; CC vs. CT + TT: P = 0.013, OR= 3.44, 95% CI = 1.3– 9.07) in Cervical Cancer, but it suggested few link with the SNP to breast cancer, lung cancer, lymphoma, odontogenic and ovarian cancer. Another finding showed a high cancer risk of rs1800587 SNP in allelic comparison model of population-based studies (C vs. T: P = 0.049, OR= 1.06, 95% CI = 1– 1.12), while the same trend risk of same mutation was found in recessive comparison model in hospital-based studies (CC vs. CT + TT: P = 0.046, OR= 1.69, 95% CI= 1.01– 2.83).

We have also expounded there are four studies did not pass the Hardy–Weinberg equilibrium in rs1800587 polymorphism. According to whether they were accord with the HWE, we divided them into two parts. Notable, climbing risks were discovered in homozygote comparison model of those four who did not conform to Hardy–Weinberg equilibrium (C/C vs. T/T: P = 0.037, OR = 1.63, 95% CI= 1.03–2.57). (Table 2.)

Association between IL1A 3’UTR- ins > del (rs3783553) and cancer susceptibility

As for rs3783553, the IL1A 3’UTR-ins/del SNP, studies showed strongly significantly links with positive risk to overall cancers in all genetic models: allelic contrast model (Ins vs. Del: P < 0.001, OR = 1.3, 95% CI = 1.21– 1.4), homozygote comparison model (Ins/Ins vs. Del/Del: P < 0.001, OR= 1.88, 95% CI= 1.58– 2.23), heterozygote comparison model (Ins/Del vs. Del/Del: P = P < 0.001, OR  = 1.46, 95% CI = 1.31– 1.62), dominant comparison model (Ins/Ins + Ins/Del vs. Del/Del: P < 0.001, OR = 1.66, 95% CI = 1.44– 1.91), and recessive comparison model (Ins/Ins vs. Ins/Del + Del/Del: P < 0.001, OR = 1.31, 95% CI = 1.19– 1.44) (Table 2, Figure 3).

Figure 3. Meta-analysis of the association between IL1A rs3783553 polymorphism and cancer risk.

After that, tumor type classification analysis revealed an interesting finding. We found risk of cancer is upregulated in homozygote, heterozygote and dominant models (Ins/Ins vs. Del/Del: P = 0.001, OR = 1.98, 95% CI = 1.31– 2.98; Ins/Del vs. Del/Del: P = 0.003, OR = 1.84, 95% CI = 1.22– 2.76; Ins/Ins + Ins/Del vs. Del/Del: P = 0.001, OR = 1.91, 95% CI = 1.29– 2.82) in cervical cancer, and there were the similar trend in cerival cancer in homozygote and recessive models of rs1800587 as well. Then, when we analyzed rs3783553 separately, except cervical cancer, there were other statistically significant rising risk between it and different type of cancer, including colorectal cancer for all comparison models(Ins vs. Del: P = 0.042, OR = 1.31, 95% CI = 1.01– 1.69; Ins/Ins vs. Del/Del: P = 0.037, OR = 1.98, 95% CI = 1.04– 3.74; Ins/Del vs. Del/Del: P = 0.001, OR  = 1.77, 95% CI = 1.26– 2.48; Ins/Ins + Ins/Del vs. Del/Del: P = 0.025, OR = 1.85, 95% CI = 1.08– 3.18; Ins/Ins vs. Ins/Del + Del/Del: P = 0.02, OR = 1.28, 95% CI = 1.04– 1.58), gastric cancer for all comparison models(Ins vs. Del: P < 0.001, OR = 1.32, 95% CI = 1.14– 1.52; Ins/Ins vs. Del/Del: P < 0.001, OR = 2.03, 95% CI = 1.44– 2.87; Ins/Del vs. Del/Del: P = 0.002, OR  = 1.72, 95% CI = 1.23– 2.41; Ins/Ins + Ins/Del vs. Del/Del: P < 0.001, OR = 1.86, 95% CI = 1.34– 2.57; Ins/Ins vs. Ins/Del + Del/Del: P = 0.009, OR = 1.31, 95% CI = 1.07– 1.59), hepatocellular carcinoma for 2 comparison models(Ins vs. Del: P = 0.049, OR = 1.23, 95% CI = 1– 1.51; Ins/Ins vs. Ins/Del + Del/Del: P = 0.038, OR = 1.27, 95% CI = 1.01– 1.58) and prostate cancer for 4 models(Ins vs. Del: P < 0.001, OR = 1.66, 95% CI = 1.3– 2.11; Ins/Ins vs. Del/Del: P < 0.001, OR = 3.14, 95% CI = 1.67– 5.9; Ins/Ins + Ins/Del vs. Del/Del: P = 0.005, OR = 1.76, 95% CI = 1.19– 2.6; Ins/Ins vs. Ins/Del + Del/Del: P < 0.001, OR = 2.18, 95% CI = 1.47– 3.23).

Another result that caught our attention is, for source of control stratification analysis, population-based subgroup analysis found a greatly increased risk in all five comparison models (C vs. T: P < 0.001, OR = 1.29, 95% CI = 1.2– 1.4; CC vs. TT: P < 0.001, OR = 1.86, 95% CI = 1.56– 2.21; CT vs. TT: P < 0.001, OR = 1.46, 95% CI = 1.31– 1.63; C/C + C/T vs. T/T: P < 0.001, OR = 1.67, 95% CI = 1.44–1.94; C/C vs. C/T + T/T: P < 0.001, OR  = 1.3, 95% CI = 1.18– 1.43). But noticeable, the hospital-based subgroup is missing in this part. (Table 2.)

Analysis using ethnicity as subgroup hinted a closely association, in both Asian and Caucasian, between ins/del mutation and high cancer susceptibility risk in all five models. For Asian, there were 16 case-controls consisting of 6493 cases and 7528 controls, whose significant genotype models were: Ins vs. Del: P < 0.001, OR = 1.28, 95% CI = 1.18– 1.39; Ins/Ins vs. Del/Del: P < 0.001, OR = 1.84, 95% CI = 1.53– 2.2; Ins/Del vs. Del/Del: P < 0.001, OR  = 1.46, 95% CI = 1.3– 1.63; Ins/Ins + Ins/Del vs. Del/Del: P < 0.001, OR = 1.67, 95% CI = 1.42– 1.95; Ins/Ins vs. Ins/Del + Del/Del: P < 0.001, OR = 1.28, 95% CI = 1.16– 1.41. In contrast, 2 studies about Caucasian included 475 cases and 490 controls, with significant models as following: Ins vs. Del: P < 0.001, OR = 1.5, 95% CI = 1.24– 1.81; Ins/Ins vs. Del/Del: P < 0.001, OR = 2.38, 95% CI = 1.55– 3.66; Ins/Del vs. Del/Del: P = 0.022, OR  = 1.48, 95% CI = 1.06– 2.06; Ins/Ins + Ins/Del vs. Del/Del: P = 0.001, OR = 1.7, 95% CI = 1.24– 2.34; Ins/Ins vs. Ins/Del + Del/Del: P < 0.001, OR = 1.71, 95% CI = 1.27– 2.3.

Sensitivity analysis and bias evaluate

Revealed from sensitivity analysis, see Table S2, Figure 4 and Figure 5, there had no significant impacts of any individual study on the pooled ORs. Besides, Egger’s test and Begg’s funnel plot showed that in each allelic contrast and genotype model, there had no publication bias all over the comparison of models (Table S3, Figure 6).

Figure 4. Sensitivity analysis for IL1A rs1800587 polymorphism and cancer risk.

Figure 5. Sensitivity analysis for IL1A rs3783553 polymorphism and cancer risk.

Figure 6. Begg’s funnel plot for IL1A polymorphisms and cancer risk. The x-axis is log (OR), and the y-axis is natural logarithm of OR. The horizontal line in the figure represents the overall estimated log (OR). The two diagonal lines indicate the pseudo 95% confidence limits of the effect estimate.

Summary and discussion

IL1A, together with other members of IL-1 family, locates on chromosome 2 and form a cluster of cytokines. There are 7 exons and 6 introns in IL1A, the 1^st exon encodes the 5’-untranslated regions, the 2^nd, 3^rd and 4^th exons encode the enzymatic hydrolysis fragment of the pro-proteins, the 5^th, 6^th, 7^th exons encode IL-1α, and the 7^th exon encode 3’-untranslated regions^[62]. The promoter of IL1A lacks of typical regulatory areas, but can be regulated by variable transcription factors through the binding sites, such as AP1, Sp1, NF- κB, etc.^[63,64].

IL-1α, the production of IL1A, can be rapidly up-regulated by multiple molecular factors, and enhance the pro-inflammatory effects, accompany with the other interleukin family molecules^[65]. Furthermore, IL-1α can reverse the transport of transcriptase form cytoplasm to nucleus and independently regulate gene expression, which may have anti-apoptotic effects on non-inflammatory cells^[4]. Researches also illustrated that the existing of IL-1α in the tumor microenvironment can enhance growth, invasion and metastasis of cancer cell, and its necrosis production could promote the vessel formation and invasion risk of tumor through boosting the expression of vascular endothelial growth factor (VEGF)^[4,66]. On the other way, there are also some findings suggested that the immune response mediated by IL1α would build up an immune memory for tumor cells, so as to continuously play an anti-tumor role by guiding the immune system to attack tumor cells^[67]. Therefore, IL1α may have both cancer-causing and anti-cancer effects, this kind of mysterious and powerful molecule is worth for a further investigating.

SNPs have a great value in cancer susceptibility, especially for the mutations which do alter cancer risk but are difficult to cause apparent symptoms. When it happens to IL1A, it may increase or decrease the risk of cancers through affecting gene transcription rates, mRNA stability, expression or activity of IL1α, etc.^{[14–16,68,69]}. Previous studies had revealed 51 unique variations in IL1A, many of them are exact functional. In the 5^th exon region +4845 site, IL1A has a G > T transform SNP (rs17561), of which could lead to the conversion of encoding amino acid at position 144 from alanine (Ala) to serine (Ser), and such change could increase the release of IL-1α through enhancing the efficiency of specific enzymes which process IL-1α precursors, thereby enhancing the inflammatory response^[14]. Also, IL1A owns a C > T SNP at -889 site of 5’-regulatory region (rs1800587). Accumulating evidence suggested that this mutation may affect gene expression by affecting transcriptional activity^[15,16]. Besides, previous studies had shown that a 4-bp insertion/deletion (rs3783553) polymorphism happening in the gene 3’-UTR could alter the strength of miRNA binding, thereby affecting cancer risks^[68,69]. In the past decades, lots of researches stated the connection between cancer risk and IL1A polymorphisms, but the conclusions were contradictory. Some case-control studies reported a high risk of cancer susceptibility with IL1A polymorphism^[47,61]. On the contrary, others suggested a low risk in the same polymorphism and the same kind cancer^[45,60]. According to the law of statistics, limitation of sample size, difference of methods, ethnicity, control’s source, systematic and sampling errors, any of them may cause the conflicting results. In order to eliminate potential errors to obtain a more accurate conclusion, we performed an overall meta-analysis in current study.

In the current work, we enrolled 45 eligible case-control studies, containing 15,586 cancer cases and 18, 430 healthy controls. The five genetic model of three SNPs were under analyzing, and comparison distributions of alleles were pooled to get a more precise result by the larger sample size. Evidence from NOS and sensitivity analysis ensured a high quality and credibility of each enrolled study. Final conclusion drew a vivid, specific and accurate picture. It revealed a prominent relationship between IL1A rs3783553 ins/del SNP and rising risk of overall cancer susceptibility. The similar findings were found in all cancer type we enrolled, including cervical cancer, colorectal cancer, gastric cancer, HCC, and prostate cancer. This conclusion shows the importance of IL1A ins/del SNP for tumor susceptibility, which means that this mutation may be used as an early screening marker for cancer, or even help explore a touchable and useful way to treat cancer in the future. The final result also suggested IL1A rs1800587 SNP may influence the overall risk with an increased trend, particularly for cervical cancer. This finding can only be found in allele and homozygote model for overall study, while it can be found in homozygote and dominant model in cervical cancer. And there is an interesting founding that we cannot ignore, rs1800587 and rs3783553 both have pointed to a link with high risk to cervical cancer. It reminds us that different genetic variations may indeed affect gene regulation in a certain link, separately and mutually. From this, the gynecologists and scholars can set out to explore the mechanism at deeper levels. Nevertheless, IL1A rs17561 SNP was found to have no effect on risk of overall cancer susceptibility. Likewise, results of pooled analysis did not demonstrate any low risk of cancer susceptibility to IL1A SNP in neither of three type mutation.

Stratification analysis based on ethnicity was done in Asian and Caucasian, it showed that IL1A rs3783553 ins/del SNP promotes the risk of cancer in both subgroups. We can make the explanation as the allele frequency of the SNP is similar between 2 kind ethnicities. In terms of OR-values, the link of this kind SNP to high cancer risk may be a little stronger among Caucasians, but the difference is not huge. This result can reinforce the idea that the ins/del mutation could enhance cancer susceptibility. Yet, we cannot extend this result to overall ethnicity because we only included Asian and Caucasian, and the allele frequency may change in other ethnicity, such as Africans.

Admittedly, there are still some limitations such as the ethnicity problem mentioned above in our study. On the one hand, the result from the source of control stratification analysis of IL1A rs1800587 SNP suggested possible bias. On the other hand, the result from the HWE based stratification analysis of same SNP implied the same condition. Although we have conducted a quality assessment on the use of NOS for those who do not comply with HWE to ensure the credibility of the results, it need to be supported by more case-control studies. Another important thing to note is, according to the result of meta-analysis, the conclusion about how IL1A rs1800587 SNP influence cancer susceptibility remain cautious. What need to ponder is the meaningless result in allele contract model in cervical cancer. It requires to be interpreted with caution that IL1A -889 T may be a potential risk allele for cancer and it is worthy of further study. Then, all SNP considered, we only enrolled English and Chinese articles, this may exist underlying selection biases. Besides, we cannot exclude influence of some confounding factors, such as gender, age, life habits like drinking and smoking, other diseases, etc. Because of that, we have no way to take potential interactions of these factors to gene into consideration. Furthermore, the development of cancer is an extremely complex process, various members of interleukin family may be involved in. Therefore, in addition to the study of SNP, we need to establish a more comprehensive study of the interaction between different genes to verify the influence of gene SNP on cancer susceptibility.

Limitations cannot belittle virtues, there does have many merits in recent research. First, by retrieving, we found that no similar study about IL1A-SNP in overall cancer susceptibility have been published before. In present study we included a large number of case-controls, so that it can supplement powerful and accurate evidence for overall study to help understand the role of how IL1A-SNPs could impact the development of cancer. Second, we evaluated the quality of enrolled studies on using NOS scale, and results from it ensure every enrolled research is with a high-quality to fulfill the pre-set criteria. Third, sensitivity analysis was preformed to test for bias, meanwhile, stratification analysis was done by cancer type, ethnicity, source of controls. All these measures are designed to avoid potential biases and lower heterogeneities so as to get the most precise conclusion in the end.

Conclusion

Taken as a whole, our study had confirmed in detail that IL1A rs3783553 SNP play an important role in cancer development process, it can significant up-regulation risk of cancer susceptibility. Also, the mutation link with high risk of cancer both in Caucasians and Asians. By comparison, the link seemed stronger among Caucasians, but the difference was not huge. Another implication is that IL1A rs1800587 SNP may be responsible for high cancer risks, especially for cervical cancer. However, this conclusion needs to be carefully explained due to its limitations. Lastly, there is no evidence that IL1A rs17561 SNP associates with cancer risk, no matter high or low. Above, present study suggests the closely relationship of IL1A SNP and risk of cancers. We are pleased to be able to contribute to the search for mechanisms and treatments for cancers, at the same time, we look forward to further researches and more comprehensive conclusions in the future.

Author contributions

Haoran Xia and Yiding Chen performed the literature search, data extraction, and statistical analysis and wrote the manuscript. Haoran Xia and Jialin Meng supervised the literature search, data extraction, analysis, Jialin Meng and ChaozhaoLiang reviewed the manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This manuscript was supported by National Natural Science Foundation of China (81630019 to C Liang).