Associations between twelve common gene polymorphisms and susceptibility to hepatocellular carcinoma: evidence from a meta-analysis

Background Associations between polymorphisms in vitamin D receptor (VDR)/vascular endothelial growth factor (VEGF)/interleukin-18 (IL-18)/mannose-binding lectin (MBL) and susceptibility to hepatocellular carcinoma (HCC) were already explored by many studies, yet the results of these studies were inconsistent. The aim of this meta-analysis was to better clarify associations between polymorphisms in VDR/VEGF/IL-18/MBL and HCC by combing the results of all relevant studies. Methods Eligible publications were searched from PubMed, Embase, WOS, and CNKI. We used Review Manager to combine the results of individual studies. Results Thirty studies were included in this study. Combined results revealed that VDR rs7975232, VDR rs2228570, VEGF rs699947, VEGF rs3025039, IL-18 rs1946518, and MBL rs7096206 polymorphisms were all significantly associated with HCC in the overall pooled population. We also obtained similar significant associations for VDR rs7975232, VDR rs2228570, IL-18 rs1946518, and MBL rs7096206 polymorphisms in Asians. Conclusions Collectively, this meta-analysis proved that VDR rs7975232, VDR rs2228570, VEGF rs699947, VEGF rs3025039, IL-18 rs1946518, and MBL rs7096206 polymorphisms may confer susceptibility to HCC in certain populations.


Background
Hepatocellular carcinoma (HCC) is one of the leading causes of death all over the world [1,2]. Although we still did not reveal the exact mechanism of its pathogenesis, it was evident that genetic components were essential in the development of HCC. Firstly, the incidences of HCC in different populations were quite different [3,4], and genetic background was probably one of the reasons behind differences in disease prevalence across different populations. Secondly, numerous susceptible genetic loci of HCC were also identified and validated by existing genetic association studies [5,6].
Mannose-binding lectin (MBL) and interleukin-18 (IL-18) are crucial modulators of immunological reactions, whereas vitamin D receptor (VDR) and vascular endothelial growth factor (VEGF) are vital for both immune-regulation and angiogenesis [7][8][9][10]. So, if a genetic polymorphism could alter the transcription activity of VDR/VEGF/IL-18/MBL or the protein structure of VDR/VEGF/IL-18/MBL, there is a possibility that this polymorphism may lead to the development of chronic inflammatory cellular injuries and also confer susceptibility to many types of malignancy including HCC.
In the past 20 years, many studies explored associations between polymorphisms in VDR/VEGF/IL-18/MBL and HCC, yet the conclusions of these studies were somehow inconsistent . To better clarify associations between polymorphisms in VDR/VEGF/IL-18/MBL and HCC, we designed this study to get a more credible conclusion by combing the results of all relevant studies.

Methods
We wrote this meta-analysis in accordance with the requirements of the PRISMA guideline [41].

Literature search and inclusion criteria
To retrieve eligible articles, we searched PubMed, WOS, Embase, and CNKI with keywords listed below: ("vitamin D receptor" or "VDR" or "vascular endothelial growth factor" or "VEGF" or "interleukin 18" or "IL 18" or "mannose-binding lectin" or "Mannose-binding protein" or "MBL" or "MBP") and ("polymorphism" or "variant" or "variation" or "mutation" or "SNP" or "genome-wide association study" or "genetic association study" or "genotype" or "allele") and ("hepatocellular carcinoma" or "HCC"). The references of retrieved articles were also screened by us to identify other potentially relevant articles.
To be included in this meta-analysis, some criteria must be met: (I) about associations between polymorphisms in VDR/VEGF/IL-18/MBL and HCC in humans; (II) Offer genotypic distribution of VDR/VEGF/IL-18/MBL polymorphisms in patients with HCC and controls; (III) full manuscript in English or Chinese is retrievable. Publications were deemed to be ineligible for inclusion if (I) not about polymorphisms in VDR/VEGF/IL-18/MBL and HCC; (II) narrative reviews, systematic reviews, or comments; (III) studies only involved HCC patients. We only included the most up to date study for analyses if duplicate publications were found during the literature search.

Data extraction and quality assessment
Two authors extracted the following essential information from eligible studies: (I) name of the leading author; (II) published year; (III) country of the leading author; (IV) ethnicity of involved participants; (V) number of patients with HCC and controls in each study; (VI) genotype distributions of polymorphisms in VDR/VEGF/ IL-18/MBL among patients with HCC and controls. P values of Hardy-Weinberg equilibrium (HWE) were also calculated.
The authors used the Newcastle-Ottawa scale (NOS) to assess the quality of eligible publications [42]. The score range of NOS is between 0 and 9, when a study got a score of 7 or more, we considered that the methodology quality of this study was good Two authors extracted data and assessed the quality of eligible studies. The authors wrote to the leadings authors for additional information if essential information was found to be incomplete.

Statistical analyses
We used Review Manager to combine the results of individual studies. Z test was employed to assess associations between polymorphisms in VDR/VEGF/IL-18/MBL and susceptibility to HCC. The statistical significance threshold of P value was set at 0.05. We used I 2 statistics to assess between-study heterogeneities. We used Randomeffect models (DerSimonian-Laird method) to combine the results if I 2 is larger than 50%. Otherwise, fixedeffect models (Mantel-Haenszel method) were used to combine the results [43,44]. We further carried out subgroup analyses by ethnicity to get ethnic-specific results. We examined the stability of combined results by deleting one study each time and combining the results of the remaining studies. We used funnel plots to estimate whether our combined results may be influenced by publication biases.

Characteristics of included studies
We found 168 articles during literature searching. Fortyfive articles were assessed for eligibility after excluding unrelated or duplicate articles. We further excluded eight reviews and six case series, and another one publication was excluded because of missing crucial data. Totally, 30 articles were ultimately found to be eligible for inclusion (Fig. 1). Extracted data of eligible articles were summarized in Table 1.

Meta-analyses results for polymorphisms in VEGF and HCC
Nine studies were eligible for the estimation of associations between polymorphisms in VEGF and HCC. VEGF rs699947 (recessive comparison OR = 0.63, 95% CI 0.41-0.98) and rs3025039 (allele comparison OR = 1.21, 95% CI 1.00-1.46) polymorphisms were found to be significantly associated with HCC in overall combined analyses. Nevertheless, we did not observe any positive associations in subgroup analyses (see Table 2 and Additional file 1: Supplementary Figure S1).

Meta-analyses results for polymorphisms in IL-18 and HCC
Ten studies were eligible for the estimation of associations between polymorphisms in IL-18 and HCC. IL-18 rs1946518 (dominant comparison OR = 0.79, 95% CI 0.68-0.92; recessive comparison OR = 1.26, 95 % CI 1.08-1.48; allele comparison OR = 0.78, 95% CI 0.67-0.91) polymorphism was found to be significantly associated with HCC in overall combined analyses. Subgroup analyses showed similar positive findings for rs1946518 polymorphism in East Asians (allele comparison), South Asians (dominant, recessive, and allele comparisons), and those with hepatitis B virus (HBV) infection (dominant and allele comparisons) (see Table 2 and Additional file 1: Supplementary Figure S1).

Sensitivity analyses
We examined the stability of combined results by deleting one study each time and combining the results of the remaining studies. The trends of associations remained consistent in sensitivity analyses, which indicated that the combined results were statistically stable.

Publication biases
Funnels plots were employed to estimate whether our combined results may be influenced by publication biases. Funnel plots of every comparison were symmetrical, which indicated that the combined results were unlikely to be seriously impacted by overt publication biases.

Discussion
The combined results of this meta-analysis revealed that VDR rs7975232, VDR rs2228570, VEGF rs699947, VEGF rs3025039, IL-18 rs1946518, and MBL rs7096206 polymorphisms were significantly associated with susceptibility to HCC in certain populations. The trends of associations remained consistent in sensitivity analyses, which indicated that the combined results were statistically stable.
To better understand the combined results of this meta-analysis, some points should be considered. First, past basic studies revealed that all investigated polymorphisms were either correlated with altered transcription activity or protein structure [45][46][47][48]. So, these variations may influence the biological function of VDR/VEGF/IL-18/MBL, result in immune dysfunction, cause chronic inflammatory hepatocellular injury, and ultimately confer susceptibility to HCC. Thus, our meta-analysis may be statistically insufficient to observe the real underlying associations between polymorphisms in VDR/VEGF/IL-18/MBL and HCC in certain subgroups. Therefore, future studies still need to confirm our findings. Second, we noticed that most eligible studies were from Asian countries, whereas studies in other countries were highly scarce, so scholars from European and African countries should also try to examine associations between polymorphisms in VDR/VEGF/IL-18/MBL and HCC. Besides, considering the functional importance of VDR/ VEGF/IL-18/MBL in regulating inflammatory reactions and angiogenesis, future studies also need to test the relationship between polymorphisms in VDR/VEGF/IL-18/ MBL and other types of malignancies. Third, the etiology of HCC is very complicated, so we highly recommend further genetic association studies to explore the effects of haplotypes and gene-gene interactions on disease susceptibility [49]. Fourth, we aimed to investigate associations between all polymorphisms in VDR/VEGF/ IL-18/MBL and HCC in the very beginning. However, we did not find any study on other VDR/VEGF/IL-18/ MBL polymorphisms, so we only focused on 12 polymorphisms in this meta-analysis. Fifth, it is worth noting that Zhu et al. [50] also performed a meta-analysis about IL-18 polymorphisms and HCC in 2016. Based on combined analyses of eight eligible studies with 3572 subjects, they did not find any positive results regarding IL-18 polymorphisms and HCC in general or subgroup analyses. Since our pooled analyses about IL-18 polymorphisms were based on more eligible studies and larger sample sizes, our results should be more statistically robust. Nevertheless, studies with larger sample sizes are still warranted to test the genetic associations between IL-18 polymorphisms and HCC in the future.
Some limitations of this meta-analysis should also be mentioned. Firstly, the results regarding associations between polymorphisms in VDR/VEGF/IL-18/MBL and HCC were based on combining unadjusted findings of eligible studies due to the lack of raw data [51]. Secondly, the relationship between polymorphisms in VDR/VEGF/IL-18/MBL and HCC may also be affected by environmental factors. Unfortunately, the majority of eligible studies only focused on associations between polymorphisms in VDR/VEGF/IL-18/MBL and HCC, so we could not explore genetic-environmental interactions in this meta-analysis [52]. Thirdly, grey literatures were not searched. So although funnel plots of every comparison were symmetrical, it is still possible that the combined results may be affected by publication biases [53].

Conclusion
In summary, this meta-analysis proved that VDR rs7975232, VDR rs2228570, VEGF rs699947, VEGF rs3025039, IL-18 rs1946518, and MBL rs7096206 polymorphisms may confer susceptibility to HCC in certain populations. These results also indicated that VDR, VEGF, IL-18, and MBL may involve in the development of HCC. However, the combined results of this metaanalysis should still be verified by studies with larger sample sizes.
Additional file 1: Figure S1. Forest plots of investigated polymorphisms.