Background
Renal cell carcinoma (RCC) is one of the most common human malignancies, with an estimated 63,990 new cases and 14,400 deaths occurring annually in the United States [
1]. In China, RCC is not reported among the top 10 cancer incidences and mortalities [
2]. Among the different histological subtypes of RCC, clear cell renal cell carcinoma (ccRCC) is the most common type, followed by papillary renal cell carcinoma (PRCC) and chromophobe renal cell carcinoma (ChRCC). The molecular profiles of those three common subtypes of RCC have been studied using next generation sequencing (NGS) in a multitude of research projects such as The Cancer Genome Atlas (TCGA) and other projects from Japan, the European Union and France.
In ccRCC,
VHL is the gene most frequently altered by germline and somatic mutations. According to TCGA analysis,
VHL, PBRM1, BAP1 and
SETD2 are the four most frequently somatically mutated genes in human ccRCC, all of which are typically mutated in combination with the loss of chromosome 3p, followed by
KDM5C, PTEN, MTOR and
TP53 [
3]. PRCC consists of two subtypes, type 1 and 2, based on distinct histological and genetic characteristics. In the TCGA database, several significantly mutated genes have been identified, including
MET, SETD2, NF2, KDM6A, SMARCB1, FAT1, BAP1, PBRM1, STAG2, NFE2L2 and
TP53. Notably, somatic mutations in
MET are mainly found in type 1 PRCC, whereas type 2 PRCC is primarily associated with somatic mutations in
SETD2, BAP1 and
PBRM1, all of which are also frequently mutated in human ccRCC. Furthermore,
TFE3 and
TFEB gene fusion and loss of
CNKD2A have been shown to be dominant in type 2 PRCC [
4]. In contrast to ccRCC and PRCC, ChRCC mainly manifests copy number variations of chromosomes, while relatively few somatic mutations are shown.
TP53 is the most recurrently somatically mutated gene in the TCGA dataset, followed by
PTEN [
5].
Up to now, the overwhelming majority of genomic datas of RCC have originated from the USA and European countries. As a consequence, most specimens have been collected from Caucasian and black patients, while very few Asian patients have been included. In the cBioPortal for Cancer Genomics (
http://www.cbioportal.org), only 98 ccRCC samples from Japanese patients have been investigated. According to the International Cancer Genome Consortium (ICGC) Data Portal (
https://dcc.icgc.org), only 10 Chinese donors are available in kidney cancer projects. The discrepancy between the somatic mutation profiles of RCC from Chinese patients and the published data still requires elucidation.
As a biomarker of response to the immune checkpoint inhibitor, PD-L1 expression in tumor cells was shown to correlate with the efficacy of immunotherapy involving programmed death 1 (PD-1)/PD-L1 inhibitors in many cancers. A recent study indicated that a longer progression-free survival was achieved with nivolumab plus ipilimumab than with sunitinib among advanced RCC patients with ≥ 1% PD-L1 expression but not among those with < 1% PD-L1 expression. Furthermore, PD-L1 was shown to serve as a predictive factor in terms of response and overall survival benefit from the nivolumab plus ipilimumab combination or nivolumab monotherapy as second-line treatment [
6]. However, the association between PD-L1 expression and somatic mutations in RCC has not been widely investigated.
In this study, we aimed to uncover the somatic alterations in RCC from Chinese patients diagnosed with primary RCC including ccRCC, PRCC and ChRCC by using WES, as well as tried to find some correlations between somatic mutations and PD-L1 expression.
Discussion
In recent years, the landscape of genomic somatic alterations in RCC has been depicted by several research projects including the TCGA, supported by multiple countries, which could be helpful for studying oncogenesis and new treatment strategies. However, racial differences might also contribute to the diversity of genomic somatic aberrations in tumors. For instance, in a study focusing on racial differences in the sequencing results of hereditary malignancies, Caswell and colleagues reported that a higher proportion of whites than nonwhites carried deleterious
CHEK2 mutations [
15]. It’s well known that the vast majority of somatic mutation profiles associated with ccRCC, PRCC and ChRCC were unmasked by foreign researchers. The subjects were mainly Caucasian and black patients. Only 10 Chinese patients with ccRCC have been evaluated using WES before [
16]. These data can be acquired via the ICGC data portal. The discrepancies and similarities of the somatic mutation landscapes in ccRCC, PRCC and ChRCC between Chinese patients and sufferers from other countries and regions still need to be elucidated with a larger sample size. In this study, we performed WES on paired fresh-frozen tissues obtained from 26 RCC cases.
In previous investigations,
VHL was always reported to be the most frequently mutated gene in ccRCC. As revealed in a study about WES performed on 106 ccRCC specimens from Tokyo [
14],
VHL gene was somatically mutated at a frequency of 40.57%. According to the TCGA database [
3], the mutation frequency of
VHL was 51.42%, which was much lower (20%) in the WES study performed on 10 Chinese patients with ccRCC [
16]. In our study on 15 paired tumor-normal ccRCC samples from Chinese patients, the mutation frequency of
VHL was 66.67%, which was much higher than that in the WES study performed on 10 Chinese with ccRCC previously. The big difference in the mutation frequency of
VHL between those two Chinese studies is probably due to the distinct sample amount, diverse sequencing platforms and different sample source centers. In this study, all of the somatic mutations in
VHL were located in the known domain for
VHL and determined to be deleterious to protein function. In other words, the function of VHL protein (pVHL) was altered or even lost. As a part of the ubiquitin-mediated proteolysis pathway, the pVHL plays an important role in the degradation of several cellular proteins containing hypoxia-induced factors (HIF). HIF includes two subunits, namely, HIF1α and HIF2α, which participates in the transcription of some genes regulating metabolism and angiogenesis [
17,
18]. Hence, the absence of pVHL function can result in the accumulation of HIF, which can contribute to the dysregulation of signaling pathways associated with metabolism, inflammation and angiogenesis, accelerating oncogenesis consequently [
19]. Considering these published ideas together, we can speculate that the deleterious mutations in
VHL identified in our study might play a leading role in the oncogenesis of ccRCC. However, loss of
VHL activity is unable to induce ccRCC by itself, as there are some other ingredients cooperating with that towards the oncogenesis of ccRCC. Amrita and colleagues demonstrated that the deficiencies of
Vhl and
Pbrm1 in the mouse kidney can lead to multifocal ccRCC with a tendency of metastasis [
20]. Sabine and colleagues showed that the combined deletion of
Vhl,
Trp53 and
Rb1 targeted in renal epithelial cells in mice caused ccRCC, which shared molecular markers and mRNA expression with human ccRCC [
21].
As the second most frequently mutated gene in ccRCC both in the TCGA and COSMIC databases,
PBRM1 is located at chromosome 3p21 encoding the BAF180 protein, which is a vital component of the PBAF SWI/SNF chromatin remodeling complex [
22]. In this study, only a stop-gain mutation (p.E981X) in
PBRM1 was detected in one ccRCC case, which had been reported previously. Compared with the data documented in the TCGA (30.6%) and COSMIC datasets (31%), the mutation frequency of
PBRM1 in ccRCC in this study was relatively lower (6.7%). Varela and partners disclosed truncating mutations in
PBRM1 at a frequency of 41% in 227 ccRCC cases [
23]. The discrepancy in the mutation frequency of
PBRM1 in ccRCC between our study and previous studies might result from racial differences in the subjects. Moreover, the smaller sample amount in our study might also contribute to that, which should be taken into consideration. Therefore, additional analysis with a larger sample size still needs to confirm the data reported herein. Nowadays, it has been generally accepted that
PBRM1 acts as a tumor suppressor gene in the kidney and plays a critical role in the pathogenesis and progression of ccRCC [
19]. It had been demonstrated that loss of
Vhl and
Pbrm1 in mouse kidney could generate ccRCC [
24]. As revealed in our study,
PBRM1 and
VHL were somatically mutated in the same ccRCC case. Consequently, we speculated that somatically altered
PBRM1 and
VHL genes worked cooperatively for the oncogenesis of ccRCC in our study. More recently, another study showed that depressed
PBRM1 and
VHL expression was associated with elevated tumor aggressiveness [
25]. In addition, the
PBRM1 mutation was also identified in one type 2 PRCC case in this study, which was consistent with the previous finding that mutated
PBRM1 was mainly associated with type 2 PRCC [
4].
Apart from
VHL and
PBRM1, there are some other genes significantly mutated in ccRCC based on the TCGA and COSMIC datasets, such as
SETD2 and
BAP1, which are both located at chromosome 3p21. For
BAP1, a missense mutation (p.H193Q) and a frameshift-deletion (p.P352fs) were found in two different ccRCC cases in this study. Regarding
SETD2, we also identified two somatic mutations in two distinct ccRCC cases consisting of a missense mutation (p.W1562C) and a stop-gain mutation (p.S512X). All of those mutations in
SETD2 and
BAP1 had not been reported before and were predicted to be deleterious. Serving as tumor suppressor genes in ccRCC,
BAP1 and
SETD2 mutations were related to worse cancer-specific survival [
26]. In the TCGA database, only mutations in
BAP1 were reported to be associated with poor survival outcome [
3]. Miura and colleagues unraveled in their research that deficiency of
BAP1 protein expression at metastatic sites indicated poor progression in patients with ccRCC [
27]. Unfortunately, no prognostic information was available in our study. Thus, ccRCC patients who were confirmed to have
BAP1 and
SETD2 mutations should be followed up regularly. Further research with a larger sample size focusing on Chinese ccRCC patients, mainly concerning about the progression and prognosis of patients with altered
BAP1 and
SETD2, should be considered.
It has been widely known that
TP53 is the most frequently mutated gene in ChRCC, with a frequency of 30.77% according to the TCGA dataset, which was only somatically mutated in one ChRCC case in this study and predicted to be deleterious. While Casuscelli and partners unraveled that
TP53 was mutated at a frequency of 58% in 38 metastatic ChRCC cases, which was much higher than that unmasked by the TCGA project and our study. In addition, those researchers found that mutations in
TP53 and
PTEN and imbalanced chromosome duplication in primary ChRCC were associated with worse survival [
28]. In contrast, all the specimens in our study were harvested from patients with no metastasis. It seemed that metastasis might underlie the discrepancy in the reported TP53 mutation frequencies. Thus, we hypothesised that somatically mutated
TP53 might serve as an important factor contributing to the aggressiveness of ChRCC. However, more further studies should be performed to confirm this hypothesis.
The PI3K/AKT/mTOR signaling pathway has been demonstrated to be highly involved in a variety of cancer types by contributing to the regulation of a series of cellular mechanisms, including proliferation, angiogenesis, metastasis and survival [
29]. It was also reported that the PI3K/AKT/mTOR signaling pathway was significantly altered and activated in ccRCC [
3,
14], playing a dominant role in the tumorigenesis in distal tubules of rats and human beings [
30]. In our study, a multitude of somatically mutated genes associated with the PI3K/AKT signaling pathway were identified in all the three RCC types, while none of the mutated genes was involved in the mTOR signaling pathway. As an important therapeutical target, mTOR inhibitors, such as everolimus, have been recommended for the treatment of patients with metastatic ccRCC. To the best of our knowledge, investigations concerning about mTOR inhibitors and metastatic RCC have been launched to search for predictive factors among the components of the PI3K/AKT/mTOR signaling pathway [
31]. However, in order to better use mTOR inhibitors for the treatment of metastatic RCC, further more studies focusing on the correlation between the PI3K/AKT/mTOR signaling pathway and RCC are still required.
Currently, PD-L1 expression in tumor cells has become a predictor of the response to immunotherapy with PD-1/PD-L1 inhibitors among diverse cancers including RCC [
6]. In this study, among the 26 RCC cases, PD-L1-positive rate in tumor cells was significantly higher in specimens with 4 somatically mutated genes, including
CSPG4,
DNAH11,
INADL and
TMPRSS13, than in samples without those (P < 0.05). None of those gene mutations were reported to correlate with PD-L1 expression in RCC tumor cells previously. In other words, those altered genes could serve as predictors of the PD-L1-positive expression in RCC tumor cells. Consequently, it could be speculated that those four somatically mutated genes might become the potential targeted genes for predicting responses to immunotherapy with PD-1/PD-L1 inhibitors in RCC. Nevertheless, whether those four mutated genes can influence the expression of PD-L1 in RCC is in need of further investigation. Previous studies had revealed that PD-L1 expression had an association with poor overall survival in ccRCC [
32], while the TCGA database indicated that only mutations in
BAP1 were associated with poor survival in ccRCC [
3]. Both of somatically mutated
BAP1 and PD-L1 expression were demonstrated to correlate with the poor prognosis of ccRCC patients. As revealed in this study,
BAP1 was altered in only two ccRCC specimens, both of which exhibited PD-L1-positive in tumour cells. Those two mutations in
BAP1 were predicted to be deleterious. Therefore, we hypothesised that somatically altered
BAP1 might serve as a critical ingredient contributing to the PD-L1 expression in ccRCC tumor cells, and most likely work in concert with PD-L1 in tumor cells contributing to the aggressiveness of ccRCC. The interaction between somatic mutations in
BAP1 and PD-L1 expression in ccRCC needs to be further elucidated in additional studies.