Clinicopathologic Significance of VHL Gene Alteration in Clear-Cell Renal Cell Carcinoma: An Updated Meta-Analysis and Review

The von Hippel-Lindau (VHL) gene is inactivated frequently in sporadic clear-cell renal cell carcinomas (ccRCCs) by genetic alteration (mutation, loss of heterozygosity, or promoter hypermethylation). However, the pathological or prognostic significance of VHL gene alteration has not been well defined. We conducted this meta-analysis to evaluate the association between VHL alteration and clinopathologic findings in ccRCCs. We performed a systematic computerized search of online databases, including PubMed, EMBASE, Web of Science, and Google Scholar (up to July 2018). From ten studies, 1,082 patients were included in the pooled analyses of odds ratios (ORs) with 95% confidence intervals (CIs) for pathological features (nuclear grade and disease stage) or hazard ratios (HRs) with 95% CIs for overall survival (OS). VHL alteration was not significantly associated with nuclear grade (OR = 0.79, 95% CI: 0.59–1.06, p = 0.12) or disease stage (OR = 1.07, 95% CI: 0.79–1.46, p = 0.65). There was also no significant correlation between VHL alteration and OS (HR = 0.75, 95% CI: 0.43–1.29, p = 0.30). When we pooled HRs for OS according to the VHL alteration types, the combined HRs were 0.72 (95% CI: 0.47–1.11, p = 0.14) for VHL mutations and 1.32 (95% CI: 0.70–2.47, p = 0.39) for methylation. In conclusion, this meta-analysis indicates that VHL gene alteration is not significantly associated with the pathological features and survival in patients with ccRCC.


Introduction
Renal cell carcinoma (RCC) is the second common malignancy after bladder cancer in the urinary system, accounting for approximately 63,000 and 5000 new cases diagnosed each year in the United States and South Korea, respectively [1,2]. RCCs comprise a group of heterogeneous tumors with distinct molecular alterations. Clear-cell RCC (ccRCC) is the most common histotype representing 75% of cases, followed by papillary RCCs (12%), chromophobic RCC (4%), oncocytomas (4%), and rare subtypes [3]. Complete surgical resection is considered to be a curative treatment option for RCCs at the early or locally advanced stage. However, about 30% of patients who received curative resection experience disease recurrence [4]. Despite the recent introduction of new molecular targeted agents, the prognosis of recurrent or metastatic RCC is very poor, with 5-year survival rate of about 10% [5][6][7][8].
The von Hippel-Lindau (VHL) gene is a tumor suppressor gene that was identified in patients with VHL syndrome [9]. VHL gene has an important role in regulation of the hypoxia pathway via the hypoxia inducible factors (HIFs) in sporadic RCCs [10,11]. Functional loss of VHL protein (pVHL), which is induced from VHL gene alteration, allows HIFs to act as a transcription factor of various pro-tumorigenic genes including vascular endothelial growth factor (VEGF) and subsequently leads to renal tumorigenesis and progression by inducing angiogenesis [12]. The VHL gene alteration events include VHL gene mutation, promoter hypermethylation, and loss of heterozygosity (LOH) by allele deletion with concomitant alteration of the contralateral gene. These genetic or epigenetic alterations play a major role in the deregulated expression of VHL gene. Somatic VHL mutations are identified in 30-60% of ccRCCs, accounting for the vast majority of sporadic ccRCCs [13]. The majority of ccRCCs containing somatic VHL mutations also exhibits contralateral allele loss (LOH), consistent with a two-hit hypothesis of tumorigenesis [13,14]. Together with 3p LOH, VHL mutation is considered a rate-limiting event in the development of RCC [13,15]. Hypermethylation of DNA in the promoter region of VHL gene results in transcriptional gene silencing. VHL promoter methylation has been observed in 10-20% of sporadic ccRCCs [16,17].
However, most individual studies had a small sample size and the statistical power was limited. Therefore, we conducted this meta-analysis to gain a better insight into the clinicopathologic significance of VHL gene alteration in patients with ccRCC. Figure 1 shows the flow diagram of the search process. A total of 722 potentially relevant articles were initially retrieved, but 697 of them were excluded after careful screening of the titles and abstracts. Of the remaining 25 potentially eligible studies, 15 were further excluded by the inclusion criteria. Eventually, the remaining 10 studies with 1082 ccRCC patients were included in the meta-analysis [22][23][24][25][26][27][28][29][30][31].  Table 1 summarizes the main characteristics and clinicopathologic data of the 11 included studies. Most studies were performed retrospectively and published between 2001 and 2017. DNAsingle-strand conformation polymorphism (DNA-SSCP) or polymerase chain reaction (PCR) followed by direct sequencing was used to detect VHL mutations, and methylation-specific PCR was used to assess VHL promoter methylation status. The incidence of VHL gene alteration including mutations, methylation, or LOH was various from 28.6% [24] to 88.8% [29]: mutations ranged from 26% [30] to 72.2% [28] and methylation ranged from 16% [23] to 31.3% [28].

Characteristics of the Included Studies
While four studies reported VHL alteration as a favorable pathological or prognostic factor [22,[28][29][30], one observed its association with worse survival [31]. The remaining five studies failed to demonstrate the significant pathological or prognostic value of VHL alteration in ccRCCs [23][24][25][26][27].  Table 1 summarizes the main characteristics and clinicopathologic data of the 11 included studies. Most studies were performed retrospectively and published between 2001 and 2017. DNA-single-strand conformation polymorphism (DNA-SSCP) or polymerase chain reaction (PCR) followed by direct sequencing was used to detect VHL mutations, and methylation-specific PCR was used to assess VHL promoter methylation status. The incidence of VHL gene alteration including mutations, methylation, or LOH was various from 28.6% [24] to 88.8% [29]: mutations ranged from 26% [30] to 72.2% [28] and methylation ranged from 16% [23] to 31.3% [28].

Publication Bias
We did not perform publication bias tests for OS because a limited number of studies were included. Visual inspection of the funnel plots for NG and stage showed symmetry, suggesting there was no substantial publication bias ( Figure 4A,B). Egger's tests also indicated the absence of significant publication biases (p = 0.357 for NG and p = 0.658 for disease stage).

Publication Bias
We did not perform publication bias tests for OS because a limited number of studies were included. Visual inspection of the funnel plots for NG and stage showed symmetry, suggesting there was no substantial publication bias ( Figure 4A,B). Egger's tests also indicated the absence of significant publication biases (p = 0.357 for NG and p = 0.658 for disease stage).

Publication Bias
We did not perform publication bias tests for OS because a limited number of studies were included. Visual inspection of the funnel plots for NG and stage showed symmetry, suggesting there was no substantial publication bias ( Figure 4A,B). Egger's tests also indicated the absence of significant publication biases (p = 0.357 for NG and p = 0.658 for disease stage).

Discussion
Although VHL gene alteration accounts for the vast majority of sporadic ccRCCs and provides plausible therapeutic target for anti-angiogenic agents, its clinicopathologic significance is still controversial with conflicting results among studies. In the current meta-analysis, we evaluated the pathological and prognostic value of VHL alteration in patients only with ccRCC. Our results failed to demonstrate the significant association of VHL alteration with pathological features and prognosis in ccRCCs.
From an improved understanding of the molecular biology of RCCs during the last decade, new molecular targeted agents have emerged for patients with advanced RCC [5][6][7][8]17]. VHL gene encodes a multifunctional protein (pVHL) which has an important role in regulation of the hypoxia pathway via the HIFs in sporadic RCCs [11]. With functional loss of pVHL arising from the genetic or epigenetic changes of VHL (VHL alteration), HIFs can act as a transcription factor of various pro-tumorigenic genes including VEGF [12]. Recently introduced targeted agents (bevacizumab, sunitinib, sorafenib, regorafenib, or pazopanib) for RCCs have anti-angiogenic effects to modulate this VHL-HIF pathway [5][6][7][8][32][33][34]. Therefore, it is logical to assume that VHL alteration may have important implications for disease prognosis.
There are two meta-analyses in the literature that investigated the pathological or prognostic impact of VHL alteration in RCC [35,36]. Our previous study evaluated the association of VHL alteration and clinical outcomes in patients with ccRCC and found that VHL alteration had no predictive or prognostic values [35]. However, we included only 3 studies in the meta-analysis and did not evaluate the pathologic impact of VHL alteration. Yang et al. investigated pathologic significance of VHL promoter hypermethylation with 13 studies and found that VHL hypermethylation was not correlated with specific clinicopathologic characteristics [36]. However, they included studies with other histotypes of RCC which rarely carry VHL alteration and did not evaluate the prognostic significance of VHL methylation. Ultimately, it is not clear that VHL gene alteration has specific pathological or prognostic value in sporadic ccRCCs. Most studies to date have been limited mainly by the retrospective design with a small number of patients. To give more statistical power, we conducted this comprehensive meta-analysis comprising an extended number of studies. Because the vast majority of VHL alteration events usually occur in ccRCCs, we excluded studies with patients diagnosed with other histotypes of RCC [20,37]. We systematically investigated the pathological and survival data of 1,082 patients with ccRCC from the ten studies [22][23][24][25][26][27][28][29][30][31]. Our results revealed that VHL alteration was not significantly associated with NG (OR = 0.79, 95% CI: 0.59-1.06, p = 0.12) and disease stage (OR = 1.07, 95% CI: 0.79-1.46, p = 0.65). In addition, there was no significant association between VHL alteration and prognosis of ccRCCs (HR = 0.75, 95% CI: 0.43-1.29, p = 0.30). These results indicate that VHL alteration has no significant pathological or prognostic impact in patients with ccRCC.
As mentioned above, VHL gene alteration is a broad concept of genetic abnormality which includes mutations, promoter hypermethylation, and LOH. Thus, we further performed the subgroup analyses to determine the difference of the polled HR for two major VHL alteration events (mutation and hypermethylation). However, the results showed that both mutation (HR = 0.72, 95% CI: 0.47-1.11, p = 0.14) and hypermethylation (HR = 1.32, 95% CI: 0.70-2.47, p = 0.39) were not significantly correlated with prognosis of ccRCCs. Recently, Lessi et al. reported an interesting result regarding VHL gene variation and prognosis in early-stage ccRCCs [38]. Tumors carrying biallelic alterations (mutation in homozygosis or presence at the same time of LOH and mutation, LOH and methylation, or methylation and mutation) had a shorter DSS compared with those with no VHL alteration. However, as the study had too small number of cases with biallelic alterations (only 15) to reach statistical significance, large sample size studies are warranted to exploit this issue.
There are several mutation types such as silent, nonsense, missense, and frameshift mutation. Loss of function (LOF) mutations can be defined as events which alter VHL transcriptional read through such as nonsense or frameshift mutations predicted to interfere with protein stability and missense mutation which alter VHL start codon [38]. Several studies reported that LOF mutations rather than other types of VHL alteration showed a meaningful correlation with worse survival [24,39]. However, these findings are premature to accept due to the small number of cases that contains LOF mutations. There are also other studies which failed to observe the significant correlation between LOF mutations and survival [25,27]. Interestingly, LOF mutations acted as a good predictive marker for VEGF-targeted therapy in patients with RCC [34,40]. Choueiri et al. investigated the VHL gene status and response to VEGF-targeted agents in patients with metastatic ccRCC [34]. In patients with tumors containing VHL mutations or methylation, the response rate was 37% compared with 31% in the wild-type group (p = 0.34). However, patients with LOF mutations obtained a significantly higher response rate than those with VHL-wild-type tumor (52% vs. 31%, p = 0.04). In this meta-analysis, we could not evaluate the impact of LOF mutations as a prognostic marker because VHL mutations were not classified into the subgroups in most studies.
This study has some inherent limitations that need to be discussed. First, this meta-analysis included a small number of studies. Second, almost all studies were retrospectively performed and therefore might carry the biases of the retrospective design. Third, patients had several potential confounders, such as ethnicities, tumor stage, and modalities of treatment. Forth, the substantial heterogeneity observed across the studies could not be completely interpreted although the random-effects model was selected for pooling HRs for OS. Finally, only articles published in English were included, which might have led to selection bias.
In conclusion, this meta-analysis indicates that VHL gene alteration is not significantly associated with the pathological features or survival in patients with ccRCC. However, large-scale studies are needed to reveal the predictive or prognostic role according to the mutational subtypes of VHL gene in patients with ccRCC.

Searching Strategy
This meta-analysis was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [41]. We performed a systematic computerized search of online databases, including PubMed, EMBASE, Web of Science, and Google Scholar (up to July 2018). The following searching terms were used: "kidney" or "renal" and "carcinoma" or "cancer" or "neoplasm" or "malignancy" and "von Hippel-Lindau" or "VHL" and "alteration" or "mutation" or "methylation" or "loss of heterozygosity." All eligible studies were retrieved and their bibliographies were checked for other relevant publications. Reference lists of identified studies and reviews were also hand-searched.

Inclusion Criteria
Eligible studies should meet the following inclusion criteria: (i) prospective or retrospective cohort studies investigating the correlation of VHL alteration (mutation, LOH, or hypermethylation) with pathological features or survival in patients with histologically confirmed ccRCC; (ii) the use of tumor tissues and adequate methods to assess VHL alteration events; (iii) sufficient data for odds ratio (OR) with 95% confidence interval (CI) for pathological findings or hazard ratio (HR) with 95% CI for overall survival (OS); (iv) studies published only in peer-reviewed journals; and (v) articles written in English.

Data Extraction
Two investigators (H.S.K and H.J.J.) independently screened relevant studies and extracted the data from each eligible study. If these two researchers did not agree, the principle investigator (J.H.K.) was consulted to resolve the disagreement through discussion.
The following data were extracted from the included studies: first author, year of publication, country, number of patients, methods for VHL gene alteration, and pathological or survival data for ORs and HRs with their 95% CIs. When both univariate and multivariate analysis were performed to get the HR for OS, the data from multivariate analysis were extracted preferentially.

Statistical Analyses
Statistical values were directly extracted from the original articles or were indirectly estimated from the given data. If HRs or ORs with their 95% CIs were not presented, the Engauge Digitizer software was used to estimate them from the corresponding data and Kaplan-Meier curves, respectively. The strength of the association between VHL alteration and pathological features was shown as ORs and their 95% CIs.
The RevMan version 5.3 (Cochrane Collaboration, Copenhagen, Denmark) was used to combine the data. The plots show a summary estimate of the results from all studies combined. The size of each square represents the estimate from each study, reflecting the statistical "weight" of the study. Outcomes are presented as forest plots with diamonds representing the estimate of the pooled effect. The width of each diamond implies its precision. The line of no impact is number one for binary outcomes, which depicts statistical significance if not crossed by the diamond [42]. The heterogeneity across studies was estimated by the Cochran's Q statistics and I 2 inconsistency test. The fixed-effect model (Mantel-Haenszel method) was used for pooling homogeneous outcomes (p ≥ 0.1 or I 2 ≤ 50%), and the random-effects model (DerSimonian-Laird method) was selected in cases with significant heterogeneity (p < 0.1 and I 2 > 50%). Statistical significance of the pooled HR or OR was determined by the Z-test for overall effect. The pooled OR or HR < 1.0 implies better pathological or survival outcomes for ccRCCs with VHL alteration.
Publication bias was evaluated by funnel plot and quantified by the Egger's test to assess funnel plot asymmetry [43,44]. Statistical significance was considered for a p-value of less than 0.05. Funding: This work had no specific funding.

Conflicts of Interest:
The authors declare no conflict of interest.