Pertinence of glioma and single nucleotide polymorphism of TERT, CCDC26, CDKN2A/B and RTEL1 genes in glioma: a meta-analysis

Background Previous genetic-epidemiological studies considered TERT (rs2736100), CCDC26 (rs4295627), CDKN2A/B (rs4977756) and RTEL1 (rs6010620) gene polymorphisms as the risk factors specific to glioma. However, the data samples of previous genetic-epidemiological studies are modest to determine whether they have definite association with glioma. Method The study paid attention to systematically searching databases of PubMed, Embase, Web of Science (WoS), Scopus, Cochrane Library and Google Scholars. Meta-analysis under 5 genetic models, namely recessive model (RM), over-dominant model (O-DM), allele model (AM), co-dominant model (C-DM) and dominant model (DM) was conducted for generating odds ratios (ORs) and 95% confidence intervals (CIs). That was accompanied by subgroup analyses according to various racial groups. The software STATA 17.0 MP was implemented in the study. Result 21 articles were collected. According to data analysis results, in four genetic models (AM, RM, DM and C-DM) TERT gene rs2736100 polymorphism, CCDC26 gene rs4295627 polymorphism, CDKN2A/B gene rs4977756 polymorphism and RTEL1 gene rs6010620 polymorphisms increased the risk of glioma in Caucasians to different degrees. In Asian populations, the CCDC26 gene rs4295627 polymorphism and CDKN2A/B gene rs4977756 polymorphism did not exhibit a relevance to the risk of glioma. It is suggested to cautiously explain these results as the sample size is small. Conclusion The current meta-analysis suggested that the SNP of TERT (rs2736100), CCDC26 (rs4295627), CDKN2A/B (rs4977756) and RTEL1 (rs6010620) genes in glioma might increase risk of glioma, but there are ethnic differences. Further studies evaluating these polymorphisms and glioma risk are warranted.


Introduction
Gliomas, which occupy more than 80% primary malignant tumors of central nervous system, are presumably caused by the aberrant growth of glial cells, mostly in the brain and spinal cord (1).Based on the cell type from which they originate, gliomas may be divided into several categories such astrocytomas, oligodendrogliomas, and ependymomas.The World Health Organization first used molecular pathology in the diagnostic classification of gliomas in 2016 (2).Adult diffuse glioma currently has 3 molecular subtypes, according to the most recent 2021 version : g lioblas toma multiforme (GBM), isocitrate dehydrogenase-wild-type (IDH-W-T); astrocytoma, IDH-mutant; and oligodendroglioma, IDH-mutant and 1p/19q-codeleted (3).Major current treatment of glioma is the combination of surgery with radiotherapy and chemotherapy, and for certain patients, immunotherapy and targeted therapy were implemented additionally (4)(5)(6)(7)(8).Nevertheless, the median survival time for gliomas is approximate 12 to 15 months merely, and both mortality and morbidity are tremendous (9).
Glioma occurs due to both environmental factors and genetic factors, and its precise pathogenesis remains opaque (10,11).Risk factors for glioma may include smoking, nitrosamines, race, ionizing radiation, and brain injuries (12).Single nucleotide polymorphisms (SNPs) in genetic variables partially increase glioma risk to some extent (10).Genome-related studies have materialized some advancements in the identification of the risk of glioma with the development of genomic research, which reveal the link between numerous SNPs and glioma risk, including 5p15.33 (rs2736100, TERT), 8q24.21 (rs4295627, CCDC26), 9p21.3 (rs4977756, CDKN2A/B), and 20q13.33(rs6010620, RTEL1) (10, 13,14).Telomerase reverse transcriptase (TERT) is crucial for telomerase activity in preserving telomeres and cellular immortality (15).In contrast, coiled-coil domain containing 26 (CCDC26) downregulates telomerase activity and increases apoptosis.Cyclin-dependent kinase inhibitor 2A/B (CDKN2A/B) can partially control the cell cycle.Regulator of telomere elongation helicase 1 (RTEL1), as a DNA helicase, can inhibit homologous recombination for directly maintaining the genomic stability (16).These genes all have varied degrees of influence on glioma pathogenesis.Due to limited sample sizes or relatively homogeneous ethnicities, current literature investigating the relationship between glioma and SNPs susceptibility frequently resulted in inconsistent findings.
In this meta-analysis, we expanded sample size by enrolling up 21 articles and used five genetic models to examine the relationship between glioma and SNPs of TERT (RS2736100), CCDC26 (RS4295627), CDKN2A/B (RS497756), and RTEL1 (rs6010620).

Materials and methods
The meta-analysis in the study was performed as per the Preferred Reporting Items for Systematic reviews and Meta-Analysis version.

Literature retrieval
Literature search was carried out comprehensively in eight databases, namely PubMed, Embase, WoS, Scopus, Cochrane Library, and Google Scholars.The MeSH term were as follows: " ((Nucleotide Polymorphism, Single) OR (SNP) OR (Polymorphisms, Single Nucleotide) OR (SNP) OR (SNP) OR (SNPs) OR (variation) OR (mutation)) AND ((Gliomas) OR (Glioma) OR (Glial Cell Tumors) OR (Glial Cell Tumor) OR (Tumor, Glial Cell) OR (Tumors, Glial Cell) OR (Mixed Glioma) OR (Glioma, Mixed) OR (Malignant Glioma) OR (Glioma, Malignant) OR (Astrocytoma) OR (Glioblastoma) OR (Diffuse Intrinsic Pontine Glioma) OR (Ependymoma) OR (Glioma, Subependymal) OR (Ganglioglioma) OR (Medulloblastoma) OR (Oligodendroglioma) OR (Optic Nerve Glioma)) AND (rs xxxxxxx)".In addition, we screened relevant reviews and references from previous meta-analyses to increase the number of other eligible studies.The retrieval process was performed by two researchers independently.Any discrepancy was adjudicated by a senior investigator.

Inclusion criteria
(1) case-control or cohort study and genome-related study regarding the association of rs2736100 polymorphism of TERT gene, rs4295627 polymorphism of CCDC26 gene, rs4977756 polymorphism of CDKN2A/B gene and rs6010620 polymorphism of RTEL gene with glioma susceptibility; (2) studies containing allelic and genotypic data or ORs and 95% CIs for susceptibility; (3) appropriate statistical methods and reliable data with clear and unambiguous expression of results, allowing calculation of the ORs and 95% CIs; (4) no overlapped data, and studies involving the same data or overlapped by the same authors, only the one that had the largest sample size or the most recently published.

Data extraction and literature quality evaluation
Each study covered the data of: the first author's name, publication year, ethnicity (Caucasian, Asian, Multiple), number of cases and controls with relevant genotypes, ORs and 95% CIs.In addition, the genetic model used in the article was checked.We applied the Newcastle-Ottawa Scale for evaluating the literature quality.Two researchers performed independent data extraction and evaluated the literature quality.A senior investigator took charge of discrepancy arbitration.

Statistical analysis
STATA 17.0 MP was implemented for all statistical analyses in this study.First, we performed the Hardy-Weinberg Equilibrium (HWE) test (p<0.05reported statistical significance).Second, we measured the association degree of TERT (rs2736100), CCDC26 (rs4295627), TERT (rs2736100) and RTEL1 (rs6010620) with glioma risk under five genetic models using the ORs and 95% CIs, which were the RM, O-DM, AM, C-DM and DM.
We merged the ORs and 95% CIs that satisfied the criteria under the same models.In addition, in all genetic models, subgroup analyses were carried out according to ethnic differences and the I 2 values were adopted for evaluating study heterogeneity.When I 2 ≤ 50%, a fixed effects model served for data analysis; when I 2 = 50%-75%, a random effects model served for data analysis; and when I 2 >75%, the Galbraith plot method was used to attenuate the heterogeneity of the study.In order to statistically analyze publication bias, we applied the Egger's Test (p<0.05:statistical significance).

Literature retrieval and quality evaluation
We initially searched 640 literature and eventually enrolled 21 eligible articles covering Caucasian, Asian, and others (10, 17-36).Twenty-one articles covered Caucasian, Asian, and multiracial studies.15 out of the 21 studies were able to extract the complete genotype frequencies regarding the two groups, and we performed the HWE test on the control group (Table 1).Six studies were able to extract only ORs and 95% CIs under the correlation genetics model, as shown in Table 2. Figure 1 displays the entire search process.The included articles have a moderate quality (Table 3).

rs2736100
In the TERT gene rs2736100 polymorphism, allele G is the risk gene (Figure 2).According to the meta-analysis analysis, the TERT gene rs2736100 polymorphism led to increased glioma risk in 5 genetic models (AM (G vs T): OR=1.29, 95% CI: 1. 25  When stratifying the analysis according to ethnicity, we obtained similar results in Caucasians, although in the O-DM results only showed a trend of association with glioma (AM (G vs T): OR=1.28, 95% CI: 1.23 to 1.33; RM (GG vs TT+TG): OR=1.29, 95%CI: 1.13 to 1.47; DM (TG+GG vs TT): OR=1.51, 95% CI: 1.29 to 1.76; C-DM (GG vs TT): OR=1.64, 95% CI: 1.50 to 1.80; C-DM (GT vs TT): OR=1.37, 95% CI: 1.27 to 1.49; O-DM (TT+GG vs TG): OR=0.93, 95% CI: 0.87 to 1.00).Among Asians, we included only one study, and the results were similar to those of Caucasians.In the study of a multi-ethnic population, we included the study of Wrensch and Safaeian which revealed the relevance of the TERT gene rs2736100 polymorphism to increased glioma risk.

Publication bias
Our study adopted Egger's Test for evaluating the literature publication bias and the results are shown in Table 4. P > 0.05 indicates that publication bias did not exist.
In rs2376100, there was publication bias in AM (p=0.469).In

Discussion
Our study is one of the first ones with conducting the first metaanalysis with the largest sample size as far as we concern and we performed detailed analyses of multiple associated SNPs under five genetic models.Besides, stratified analyses were carried out considering ethnicity.This indicates reliable meta-analysis results.
From the meta-analysis, allele G in the SNP at the rs2736100 locus differentially elevated glioma risk, which conforms to studies by Di Stefano (24).In particular, in the C-DM (GG vs TT), it elevated the risk by 69% (95% CI: 1.56 to 1.82).However, in the O-DM (TT+GG vs TG), the risk of glioma was reduced by 7% (95% CI: 0.88 to 0.98).We obtained basically identical results in Caucasians after stratification by ethnicity.Nevertheless, in the Asian O-DM, the risk of glioma was not associated with the SNP of the gene.Nevertheless, there is only one study on Asian populations and it is necessary to cautiously interpret the obtained results.Caucasians and Asians showed significant differences in the SNPs at the rs4295627 locus versus the rs4977756 locus.In the allele, recessive, dominant and C-DMs, all increased the risk of glioma.But in Asian populations, SNPs at both loci did not present relevance to glioma risk.This suggests the existence of significant racial differences.However, the amount of research literature on Asian populations is too small and further studies are needed to verify this conclusion.Among the SNPs at the rs6010620 locus, according to analysis of all five models, allele G elevated the risk, with the C-DM (GG vs AA) increasing the risk by 92% (95% CI: 1.71 to 2.15).However, in Chen's study, the risk of glioma was reduced by 19% (95% CI: 0.68 to 0.96) in the O-DM, which was possibly because the sample size in Chen's study was small (19).
Telomere refers to a small fragment of DNA-protein complex located at the end of eukaryotic chromosome, consisting of a sixbase repeat sequence -TTAGGG-and binding proteins, which have important roles in localizing, replicating and protecting chromosome and controlling cell growth, and exert a close relevance to apoptosis, transformation and immortalization of cells.Telomere shortening may also increase susceptibility to Flow chart of the study selection and exclusion criteria from this meta-analysis.cancer (37).TERT is the most important component in regulating telomerase activity.A few scholars found the relevance of TERT expression to glioma grade and patients' prognosis (38).Located in intron 2 in TERT, the SNP can impact the telomerase activity or clearly relates to a functional variant in TERT (39).In the study by Codd, people who possessed TT genotype (hTERT rs2736100 T > G) presented shorter telomeres relative to people who possessed TG genotype (40).Oppositely, a larger number of cell divisions may be supported at long telomeres, which makes it easier to achieve abnormalities and thereby contributes to the development of cancers (41).However, in our study, in the SNP of TERT rs2736100, allele G increases the risk of developing glioma.This implicates that telomere length possibly plays a double-edged role in cancer development.In 2019, Yuan conducted a detailed review of TERT mutation-related cancers, with the highest frequency of TERT mutation reaching 80% in glioma, and more than a dozen tumors such as hepatocellular carcinoma, thyroid cancer, and skin melanoma are also closely related to this gene mutation (42).Hence, it is necessary to conduct deep investigations on relevant function mechanisms in order to better treat clinical patients.CCDC26 can be found on chromosome 8q24.21,and the rs4295627 single nucleotide polymorphism can only be found in the intron 3 region regarding the CCDC26 gene, where a G-T base mutation can increase glioma disease risk (31).This is consistent with the meta-analysis in the study.The explanation can be attributed to that retinoic acid phosphorylates cAMP response element binding proteins for the induction of caspase 8 transcription and downregulates the telomerase activity for enhancing the apoptosis in neuroblastoma and glioblastoma cells under death stimuli (43).In addition, clinical studies employ CCDC26 to predict glioma patients' diagnosis and prognosis (22).Jenkins found that CCDC26 variants could remarkably elevate the risk of low-grade gliomas that underwent IDH1 or IDH2 mutations, suggesting the interaction between CCDC26 and IDH mutations and/or the downstream effects regarding above mutations for promoting glioma development (44).Wang studies showed that lncRNA CCDC26 silencing inhibits glioma cell growth and migration by targeting miR-203.It revealed the regulatory mechanism of CCDC26/miR-203 pathway in glioma pathogenesis and provided a new target for glioma treatment (45).In addition to gliomas, Homer-Bouthiette found that deletion of the gene reduced breast cancer incidence in a mouse model of breast cancer (46).This chromosomal region has also been studied in prostate, bladder and rectal cancers (47)(48)(49)(50).
CDKN2A/B is a representative altered genes in human cancers, and expression deletion facilitates malignant behavior by dysregulating cell cycle and promoting cell proliferation (51).CDKN2A/B is localized at 9p21, encoding p14, p16 (CDKN2A) and p15 (CDKN2B) oncoproteins which play multiple roles in cellular stress recognition, senescence regulation, differentiation as well as apoptosis in the developmental and proliferative phases of cells (51,52).As reported, an unknown antisense lncRNA is encoded at human CDKN2A/B locus at 9p21.3, i.e.ANRIL, which crucially impacts disease development.The rs4977756's 59 kb telomere is also mapped to CDKN2B in the 122 kb region regarding LD 9p21.3.Based on the meta-analysis, the G allele elevates the glioma risk.It is suspected that this SNP alters the expression level of ANRIL and thus affects cell proliferation, apoptosis and metastasis to promote tumorigenesis (53).Interestingly, type 2 diabetes and cardiovascular disease also appear to be linked to this gene (54, 55).RTEL1 is involved in deoxyribonucleic acid repair, regulation of replication and transcription, and maintenance of telomere length.RTEL1 can both promote and suppress tumorigenesis.RTEL1 SNPs associated with gliomas are mainly found in the non-coding (intron) regions of genes.Intron SNPs possibly facilitate aberrant exon inclusion or deletion by altering mRNA splicing and the formation and addition regarding nonsense transcripts (56,57).The RTEL1 gene rs6010620 on 20q13.33 is a vital candidate genetic variant that has been widely reported (58).As reported by previous studies, mutations in the rs6010620 gene may cause gliomas, but these findings have not been confirmed.According to meta-analysis, in the five genetic models, rs6010620 increased glioma susceptibility, most significantly in the C-DM (GG vs AA), which increased the susceptibility by 92% (OR: 1.92; CI (1.71, 2.15).Also, miR-4530 underwent downregulation in glioma tissues as well as cell lines, and overexpression inhibited malignant biological behaviors like glioma cell migration, proliferation, invasion and colony formation.RTEL1 was a direct target of miR-4530.The abnormal expression of RTEL1 could lead to obvious reversion of the miR-4530 overexpression in glioma cell lines.The miR-4530/ RTEL1 axis acts as a latent treating target specific to gliomas (59).Not only glioma, but also mutated RTEL1 has been linked to the development of astrocytoma and Hoyeraal-Hreidarsson syndrome (60).Although recent in-depth studies have confirmed that RTEL1 can maintain genomic stability by studying its ability to maintain  telomere homeostasis and promote DNA replication, there are still a lot of questions that shall be highlighted.Nowadays, researches have not confirmed the regulation of RTEL1 and its recruitment to impact the replication forks and telomeres (60).
Studying the association between the risk of glioma and SNPs is definitive, because it could help clinicians use such polymorphisms as pragmatic molecular biomarkers of glioma risk, assist drug developers innovate relevant targeted therapeutic agents for the benefit of patients, and allow for basic researchers to explore the molecular mechanisms of glioma pathogenesis and accurately regulate glioma progression.According to meta-analysis results, racial differences exist in the risk of glioma, but there is a paucity of studies on Asian ethnicity, which requires researchers to conduct relevant studies.Exploration of SNPs at other loci and risk of glioma is also necessary.

Limitations
Some limitations should be noted when interpreting the current results.The study also presents some limitations, which shall be noted during the interpretation of current studies.
First, there were scarce studies on Asian populations.Studies on results in Asian populations should be interpreted with caution.
In addition, the absence of more specific individual information about interaction within genes and that between gene and environment prohibited us from conducting a more precise analysis.
Finally, some unavoidable publication bias in the results of meta-analysis should also be taken into account.

Conclusion
In summary, our study suggested that the TERT gene rs2736100 polymorphism, CCDC26 gene rs4295627 polymorphism, CDKN2A/ B gene rs4977756 polymorphism and RTEL1 gene rs6010620 polymorphism might all increase the risk of glioma development, but there are ethnic differences.It is necessary to well develop casecontrol studies with large sample sizes and a focus on more races or glioma types to overcome the limitations described earlier to make the conclusions more accurate.

TABLE 2
Characteristics of the literature providing ORs and 95% CIs under the correlated genetic model.

TABLE 3
Results of literature quality evaluation.

TABLE 4
Publication bias results of glioma SNPs under five genetic models.