Unfavorable prognosis and clinical consequences of APOBEC3B expression in breast and other cancers: A systematic review and meta-analysis

1 Introduction

Cancer is regarded as major cause of mortality and morbidity all over the world [1]. In the year 2020, 19.3 million new cancer patients and 10.0 million cancer-associated deaths were reported worldwide. According to GLOBOCAN, 28.4 million new cancer cases are predicted for the year 2040 worldwide [2]. Despite many advances in cancer diagnosis and treatment, the cancer mortality rate is still high and cancer remains a major health problem all over the world [3–5]. It is generally believed that the accumulation of diverse mutations are important driving force for the development of neoplasia [6]. Mutations may originate from exogenous sources which are categorized as environmentally induced or rooted intracellularly [7]. DNA repair systems and their enzyme components are speculated as endogenous mutation sources. Based on the analyses of data from The Cancer Genome Atlas (TCGA), it is reported that APOBEC cytidine deaminase mutagenesis patterns are evident in cancers and could be implicated in the induction of somatic mutations leading to genome instability and carcinogenesis [8, 9].

APOBEC3 or A3 proteins are DNA cytidine deaminases that have a remarkable role in mutations in many different types of cancers [10, 11]. This family is specific to mammals, and in humans comprises seven members (A, B, C, D, F, G, and H) which are highly homologous. Their genes are located in a tandem cluster on chromosome 22q13; the proteins are involved in innate immunity and anti-viral immune response [12–14]. Current evidence, mainly resulting from studies on yeasts, suggest that APOBEC3A and APOBEC3B are responsible for somatic mutations in human cancers [15, 16] but due to the conflicting outcomes from studies performed on tumors, the mutagenic nature and role of APOBEC subtypes in various human cancers are not fully elucidated yet [17, 18].

Out of the seven members of the APOBEC3 enzyme family, APOBEC mutagenesis in tumors is attributed to A3A and A3B enzymes. Positive connection between mRNA level of A3A and A3B enzymes and APOBEC-signature mutations in tumors as well as in vitro overexpression studies strongly confirmed that high activity of these enzymes is mutagenic and genotoxic [7, 19]. Conclusions regarding the role of A3A and A3B expression in mutagenetic processes were mainly drawn through RNA-seq studies despite the fact that these highly homologous APOBECs are poorly resolved and their mRNAs cannot be confidently quantified by short sequencing reads. The dominant mutagenic potentials of A3A or A3B expression in different cancers are highly debated. For example, Banday et al. [20] based on their observations reported that A3B and not A3A is expressed in bladder cancer cell lines, which show a high load of APOBEC-signature mutations, demonstrating that in some cancer cell lines APOBEC-mediated mutagenesis is dominantly caused by the A3B activity. Tsubo M et al. [21] quantified A3B mRNA expression by RT-qPCR in fresh frozen tumors and paired adjacent normal breast tissue samples and reported that A3B mRNA expression was significantly higher in tumors comparing to the adjacent normal tissues.

Recently the clinical significance of A3B is extensively investigated because of its relevance to survival endpoints including DFS, OS, RFS in cancer patients [22, 23]. A3B is a unique member of the family with permanent nuclear localization [24]. This gene has cytidine deaminase activity and contains one catalytically active C-terminal domain (CTD) that bound to ssDNA and N-terminal pseudo-catalytic domain (NTD) [16, 25].

A3B is considered to induce mutation in breast cancer and overexpression of APOBEC3B is reported in variety of cell lines [26]. Also, A3B is introduced as the main factor contributing to mutations in various human cancers [16]. The expression of A3B is linked to the accumulation of somatic mutations during tumorigenesis and progression [26–28]. Recent studies have reported that A3B is significantly up-regulated in many cancers like ovarian, cervix, breast, lung, head & neck, and bladder [29, 30]. Abnormal overexpression of A3B promotes mutations in the PIK3CA gene in several human cancers [31, 32]. Through the knockdown experiments, it is revealed that endogenous A3B correlates with higher levels of genomic uracil, elevated frequencies of pathogenic mutations, and C-to-T transitions. Also, experimentally A3B overexpression results in cell cycle deviation and cell death, DNA fragmentation and aggregation of γH2AX. Phosphorylation of H2AX in the Ser-139 residue causes the formation of γH2AX which is biomarker of DNA damage and genotoxicity [26, 34].

There are different perspectives on the link between A3B and the cell cycle. Lackey et al. reported that the expression pattern of A3B is similar during the mitosis phase and S phase [35]. But, Hirabayashi et al. found that A3B mRNA is the only member of the APOBEC3 family that is differentially expressed at the G2/M-phase in myeloma and bone marrow mononuclear cells [36]. Ma et al. hypothesized that A3B expression induces human hepatocellular carcinoma (HCC) tumorigenesis through deaminase-independent activity and increases cycle progression in HCC cells [37]. McCann et al. stated that A3B can bind to CDK4 and inhibit CDK4-dependent nuclear translocation of cyclin D1, thus delaying cell cycle progression. Overexpression of A3B leads to G1/S phase stalling and increases A3B access to its substrate (ssDNA). Thus, it may enhance the likelihood of mutation and genomic DNA deamination [38]. In addition, pan-cancer analysis showed that the expression of A3B is related to cell cycle-regulated genes [39]. Although these evidences proposed that cell cycle-dependent activities are linked to the regulation of expression of A3B, little is known about the mechanism of cell cycle-dependent regulation of A3B expression and many questions remain unresolved.

Recently, Du et al. argued that high expression of APOBEC3B was linked to worse OS and DSF in patients affected with ovarian cancer [40]. Besides, A3B upregulation can induce the expression of some immunotherapy response genes like; T-cell infiltration and PD-L1 in non-small cell lung cancer (NSCLC) patients, so it can be used as a predictive biomarker [41].

Although deregulation of APOBEC3B has been reported in many cancer types, the conclusions about the prognostic value of A3B for patients remain controversial. A large number of studies reported that high A3B expression is linked with adverse outcomes or poor prognosis in cancer patients [40, 42–44]. However, several studies reported that A3B expression improved OS in patients with cancer or have no correlations with disease-free or overall survival [32, 46]. To date, no meta-analysis has been performed to comprehensively assess the prognostic significance of APOBEC3B expression in cancer patients. Therefore, we performed a systematic review and meta-analysis based on previously published data to assess the link between APOBEC3B expression and the clinicopathological factors and survival in patients with diverse types of cancer.

2 Materials and methods

3 Results

3.1 Characteristics of studies

A total of 846 articles were obtained using the primary search according to PRISMA guidelines. Based on our defined exclusion criteria, 827 papers were excluded. Detailed excluding criteria and the number of papers failed to attain eligibility criteria are as follow: irrelevant studies (n = 321), duplicates (n = 243), reviews (n = 17), congress and scientific meeting abstracts (n = 21), letters (n = 4), noncancerous studies (n = 99), animal studies (n = 98), cell line studies (n = 4), and studies lacking sufficient data (n = 20). Finally, 19 studies were included for this meta-analysis. The flow diagram of article selection is shown in Fig. 1. Among the 19 studies, 4 were from Japan, nine from China, three from the USA, the other three from Germany, the Netherlands and South Korea. The included studies contained seven types of cancers, namely breast cancer (n = 5), ovarian cancer (n = 4), non-small cell lung cancer (n = 2), nasopharyngeal carcinoma (n = 1), renal cell carcinoma (n = 1), gastric cancer (n = 2), squamous cell carcinomas (n = 2), neuroendocrine tumors (n = 1), bladder cancer (n = 1). In 11 studies the A3B expression level was detected by immunohistochemistry (IHC) and eight studies performed real time quantitative PCR (qRT-PCR) to detect A3B expression. The cut-off values for A3B expression were different among studies due to various definitions.

OPEN IN VIEWER

Fig. 1

Flow diagram of the study selection process in meta-analysis.

3.3 Relationship of A3B overexpression and prognosis outcome

A total of 13 studies with 2502 patients had reported the relationship between OS and A3B expression level. The random-effects model was used due to heterogeneity (I² = 93%, P < 0.01).

Poor prognosis was significantly associated with higher A3B expression (pooled HR = 1.30; 95% CI:1.09–1.55) (Fig. 2). Subgroup analysis showed that A3B expression was associated with OS of patients with lung cancer (HR = 1.85, 95% CI:1.40–2.45), and breast cancer (HR = 1.38, 95% CI:1.00–1.89). For the rest of cancers, no statistically significant results were obtained (Table 2).

OPEN IN VIEWER

Fig. 2

Association between A3B expression and survival outcomes.

OPEN IN VIEWER

Table 2

Results of subgroup analysis of pooled HRs with regard to overall survival

		n. study	HR	95% CI	I2 %	p Heterogeneity
Overall		13	1.30	(1.10–1.55)	93.4	0.004
Region	East Asia	8	1.50	(1.08–2.08)	76.6	< 0.001
	USA	3	0.98	(0.87–1.09)	81.9	0.004
	Europe	2	1.09	(0.76–1.57)	60.9	0.110
Detection Method
	IHC	8	1.41	(1.02–1.97)	95.1	< 0.001
	RT-qPCR	5	1.10	(0.83–1.47)	89.5	< 0.001
Cancer Type
	Ovarian	4	0.97	(0.87–1.08)	75.0	0.007
	Breast	3	1.38	(1.00–1.89)	90.4	< 0.001
	Gastrointestinal	2	1.24	(0.41–3.74)	73.1	0.054
	Lung	2	1.85	(1.40–2.45)	0.0	0.896
	Others	2	1.12	(0.08–15.66)	93.6	< 0.001

CI = confidence interval, HR = hazard ratio, calculated by random effects meta-analysis.

To differentiate between RT-qPCR and IHC techniques used to determine A3B expression, subgroup analysis was performed. Only high expression of A3B detected by IHC method (8 studies) was associated with overall survival (HR = 1.41, 95% CI:1.02–1.97). Subgroup analysis based on geographic regions showed correlation in patients from the East Asia region with 8 studies (HR = 1.50, 95% CI:11.08–2.08) (Table 2). Sensitivity analysis demonstrated that our result was stable based on estimates of the HRs (Fig. S1a). There was no evidence of publication bias by the Egger test (P = 0.123) and the Begg test (P = 0.329) (Fig. S1b).

A total of 7 studies with 1296 patients had reported the relationship between poor DFS and A3B expression level. The random-effects model was used due to heterogeneity (I² = 77%, P < 0.01). Poorer outcome in terms of DFS was significantly associated with higher A3B expression (pooled HR = 1.66; 95% CI:1.17–2.35) (Fig. 2). Sensitivity analysis displayed that the result was stable based on estimates of the HRs (Fig. S2a). There was no evidence of publication bias by the Egger test (P = 0.201) and the Begg test (P = 0.881) (Fig. S2b).

A poor prognosis in terms of RFS also was associated with higher expression of A3B (HR = 1.51, 95% CI:1.11–2.04, no. of studies = 4). A total of 4 studies reported the relationship between poor PFS and A3B expression level. The random-effects model was used due to heterogeneity (I² = 79%, P < 0.01). Poorer outcome in terms of PFS was not associated with higher A3B expression (pooled HR = 0.90; 95% CI:0.79–1.02) (Fig. 2).

3.4 Association of A3B expression with age and gender groups

Subgroup analysis on the basis of age and gender was performed in relation to A3B expression but the results were not statistically significant (OR = 0.87, 95% CI:0.59–1.26, OR = 1.28, 95% CI:0.79–2.09) respectively (Fig. 3).

OPEN IN VIEWER

Fig. 3

Forest plot regarding the association of A3B expression with age and gender.

3.5 Association of A3B expression and clinical stage

OR values and corresponding CIs were pooled to evaluate the relationship between A3B expression and clinicopathological data. From the data of 10 studies, A3B expression was not associated with cancer stage (OR = 1.54, 95% CI: 0.95–2.50, I² = 70%) (Fig. 4). In a subgroup analysis, higher expression of A3B detected by IHC method was associated with advanced tumor stage in all cancer types (OR = 1.89, 95% CI: 1.17–3.07, I² = 50.6%, no. of studies = 6). There was no evidence of publication bias by Begg (P = 0.788) and Egger’s test (P = 0.830) (Fig. S3)

OPEN IN VIEWER

Fig. 4

Forest plot for the association between A3B expression and lymph node metastasis /cancer stage.

3.7 Association of A3B expression with grade and tumor size

Eleven studies covering 2145 patients were analyzed for assessing the relationship of A3B expression and the clinical grade of tumors. A random-effect model was applied because of the inter-study heterogeneity (I² = 79.1%, P < 0. 01). Although half of the studies implied a direct association of clinical grade with higher A3B expression, overall effect size did not show a significant association (OR = 1.62, 95% CI: 0.94–2.77) (Fig. 5).

The determined OR values remained the same by excluding any single article (Fig. S5). Begg funnel plot was symmetric (P = 0.484) and Egger test did not show a significant publication bias (P = 0.822) (Fig. S5). A3B expression was not significantly correlated with tumor size (OR = 1.20, 95% CI = 0.45–2.24) (Fig. 5).

OPEN IN VIEWER

Fig. 5

Forest plot regarding the association of A3B expression and grade / tumor size.

Cancer is the most important challenge for life expectancy in both developing and developed countries of the world [1, 51]. Despite advances in clinical and molecular research to find effective biomarkers with prognostic significance in cancers, but steady rise in cancer incidence is indicative of the urgent need for reliable and effective cancer screening/prognosis biomarkers [52]. APOBEC3B is one of the main sources of mutagenesis in several human cancers including lung, bladder, head and neck, cervix, breast, etc [30, 54]. Recently, many studies reported an abnormal expression of A3B in multiple cancer types which may contribute to the poor prognosis and therapeutic resistance [41, 56]. Although the significance of A3B as a mutation source is becoming well-acknowledged, the prognostic value of A3B expression in cancers remains controversial. In this meta-analysis, we investigated the prognostic value and clinicopathological features of A3B expression in cancer patients. We evaluated survival data of 3776 patients in 19 studies. Our meta-analysis revealed that increased A3B expression was closely associated with poor OS, poor DFS and poor RFS in some of cancers.

To reduce the heterogeneity of studies, the sub-group and meta-regression analysis were conducted by factors including study region, cancer type, detection method and expression measurement method either by protein or mRNA quantification. Results from stratified analysis by study region showed that the associations between high A3B expression and poor OS were significant in the East Asia region but not in non-Asian populations, suggesting that the relation between A3B expression and poor OS could be ethnic-dependent. In the subgroup analysis based on cancer type, APOBEC3B expression was associated with poor OS in lung and breast cancer but this was not true in the case of ovarian cancer. This may be indicative of that A3B might have different biological behaviors and clinical characteristics in various cancer types.

Results from stratified analysis by method showed that the associations between high A3B expression and poor OS were significant in the studies that used the IHC method but not RT-qPCR. There is a possible explanation for these results. APOBEC3 family have considerable nucleotide identity at the mRNA level so these high identities make some complication to the quantification of the mRNA expression [57–59]. Some of the included studies used RT-qPCR and their result may be affected by these high identities.

In addition, the role of A3B in cancer metastasis and progression should also be considered. For example, A3B overexpression can cause genomic instability in human cells by inducing C-to-T transitions which might lead to tumorigenesis [6]. Additionally, Yamazaki et al. found that highly expressed A3B in myeloma cells causes DNA substitutions and deletions. This shows that A3B could induce mutations in the tumor genome in a way that bypasses the DNA repair mechanism, potentially leading to genomic instability in myeloma [60]. Saito et al. also observed that irradiation can induce APOBEC3B expression and may lead to radiation-induced mutation in HepG2, HeLa and SAS cells [61].

The role of A3B expression in different malignancies suggested that common signaling pathways in various cancer types are likely to be involved in dysregulation of A3B expression. Maruyama et al. found that high expression of A3B has been induced by the PKC/IKK/NFkB signaling pathway in various cancer cells, which might lead to the accumulation of somatic mutations and genetic diversity in many common cancers [62]. Also, P53 is a serious inhibitor for A3B expression in cells and limits its mutagenicity. In cancer cells due to the loss of p53 through mutation, A3B overexpression can elevate the power of mutagenesis and tumor evolution [63].

Chemotherapy is the first-line approach to destroy cancer cells for many cancers. Recent, studies showed that A3B expression is correlated with resistance to chemotherapy. Periyasamy et al. found that chemotherapy can promote A3B expression via DNA-PKcs and ATM-mediated activation of the NF-κB pathways and therefore decrease the efficacy of chemotherapeutic treatments [64]. Serebrenik et al. suggested that in ovarian cancer, APOBEC3B is a predictive biomarker of response to platinum-based chemotherapy [45]. Furthermore, in ER-positive breast cancer, A3B expression can induce tamoxifen resistance. Such observations suggest that A3B has a role in the progression of tumors and inhibiting APOBEC3B expression might improve the efficacy of cancer treatments [65].

Mutation in A3B gene was reported to play role in malignancy development. For example, Radmanesh et al. demonstrated that truncating variant at the A3B locus (c.783delG) is related to the risk of breast cancer and early onset of the disease [66]. The results of one meta-analysis revealed that 30 kb APOBEC3B deletion can decrease A3B expression and does not link to the risk of breast and ovarian cancers in the European population [67]. However, Cescon et al. proved that in breast cancer, A3B overexpression reflects mitosis and cell cycle-related gene expression, whereas a deletion polymorphism is linked to immunological activation. The authors proposed that the A3Bdel polymorphism could be utilized to predict cancer immunotherapy [68].

In spite of the inspiring results, there are several limitations in the current meta-analysis that need to be taken into consideration. First, the cut-off value and detection method for APOBEC3B expression varied among the included studies, which may cause heterogeneity and bias. Second, all the studies included in our meta-analysis were in English, which might bring selection bias. Third, only a few included studies evaluated PFS and DFS, which might result in bias. Fourth, most of the eligible studies were conducted in Asia, this might also lead to bias and our results might be more relevant to Asian patients. Fifth, immune infiltration into the tumor microenvironment may contribute to an increase of APOBEC3B expression in qPCR measurements and may lead to bias. Finally, some HRs and 95% CIs were not mentioned directly in the included studies. Then, we extracted them from Kaplan–Meier curves which could not be precise enough.

In conclusion, to the best of our knowledge, the present study is the first comprehensive meta-analysis exploring the prognostic value of APOBEC3B overexpression in various cancers. The results indicated that high APOBEC3B expression is significantly related to poor clinical outcomes in cancer patients especially breast cancer. Therefore, APOBEC3B overexpression might be used as an adverse prognostic biomarker for cancer patients. However, further high quality studies are needed to strengthen these findings.