Abstract

Introduction

   Sunitinib is the first-line targeted drug for the treatment of advanced
   renal cell carcinoma (RCC), but its therapeutic potential is limited by
   premature drug resistance. In an attempt to overcome this limitation, a
   sunitinib-resistant cell-derived xenograft (CDX) model of clear cell
   renal cell carcinoma (ccRCC) in vivo was constructed in this study. The
   molecular mechanism of drug resistance was analyzed using sequencing
   and bioinformatics tools.

Methods

   First, mice were injected subcutaneously with tumor cells 786-O to
   create tumors and were simultaneously treated with sunitinib. After
   three consecutive passages, a drug-resistant xenograft model was
   obtained. Then, key pathways and genes were identified via
   second-generation sequencing of the tissue and the examination of
   differentially expressed genes (DEGs) with bioinformatics tools.

Results

   Analysis of sequencing data revealed that 646 DEGs were upregulated and
   465 were downregulated in the drug-resistant tissues when compared with
   the sensitive tissues. GO showed that the DEGs were significantly
   enriched in angiogenesis, cell hypoxia response, and apoptosis. KEGG
   analysis demonstrated that the main pathways were PI3K-Akt, HIF-1,
   NF-kappa B, and MAPK. Modular analysis of the PPI network indicated
   that the GO and KEGG analyses of module 1 with the highest ranking were
   mainly related to ubiquitinase activity. Similarly, the GO and KEGG
   analyses of the top 10 hub genes were also chiefly linked to
   ubiquitinase activity. Then, comprehensive expression analysis of the
   hub genes, and finally, the genes BTRC and TRIM32 were identified,
   which were consistent in all observations.

Conclusion

   In this study, through the construction of in vitro models and
   bioinformatics analysis, the important pathways and key genes related
   to ccRCC sunitinib resistance were discovered. Among them, ubiquitinase
   may play an important role in drug resistance and may be a potential
   therapeutic target and biomarker.

   Keywords: ccRCC, sunitinib, drug resistance, xenograft model,
   ubiquitinase

Introduction

   Renal cell carcinoma (RCC) accounts for approximately 2–3% of all
   malignant tumors, and its prevalence is rising. Metastatic RCC accounts
   for 25–30% of all RCC cases, and has an exceedingly poor
   prognosis.[44]1 In 2020, among approximately 430,000 newly discovered
   cases of RCC, 179,000 died.[45]2 Clear cell renal cell carcinoma
   (ccRCC) is the most prevalent form of the disease, and accounts for
   approximately 70–80% of all RCC cases.[46]3 RCC is insensitive to
   radiotherapy and chemotherapy. Surgical resection is the main method
   for the treatment of localized RCC, but nearly 30% of patients have
   tumor recurrence and metastasis after surgery.[47]3 Therefore, a new
   treatment method, targeted therapy, was introduced.

   At present, receptor tyrosine kinase inhibitors (TKIs), such as
   sunitinib, constitute the first-line treatment for advanced RCC.[48]4
   However, sunitinib use poses challenges when significant efficacy is
   achieved. According to statistics, approximately 20% of patients will
   show innate resistance when receiving sunitinib for the first time, and
   most patients will develop secondary resistance to the drug after 6–11
   months of treatment.[49]5,[50]6 Sunitinib-resistant RCC tissues are
   difficult to obtain for ethical reasons; hence, there are few studies
   on the mechanism of sunitinib resistance. Therefore, the molecular
   mechanism of drug resistance is yet to be elucidated.

   To screen and evaluate new antitumor drugs and to study the mechanism
   of tumor resistance, researchers have established tumor xenograft
   models. The commonly used models are cell-derived xenograft (CDX)
   model[51]7 and patient-derived xenograft (PDX) model.[52]8 While the
   CDX model has the characteristics of fast modeling and high success
   rate, the PDX model has the advantages of retaining the patient’s
   molecular phenotype and genotype and can, therefore, predict the
   clinical outcome well.

   Bioinformatics is an emerging interdisciplinary subject that has been
   developed on the basis of life science, computer science, and
   mathematics. This field utilizes mathematics and computer science to
   collect, process, store, disseminate, and analyze biological data to
   understand their significance.[53]9 Nowadays, more and more researchers
   are using microarray technology combined with bioinformatics tools to
   identify new genes related to cancer progression, diagnosis, prognosis,
   and drug resistance.

   In this study, a CDX model of sunitinib-resistant ccRCC was first
   constructed, the gene expression profile of the drug-resistant tissue
   was sequenced, and then the sequencing data were analyzed using
   bioinformatics tools to identify the potential genes responsible for
   sunitinib resistance in ccRCC.

Materials and Methods

Establishment of the CDX-Resistance Model

   All animal experiments were performed following the protocol approved
   by the Animal Care and Use Institutional Committee of Sun Yat-sen
   Memorial Hospital, Sun Yat-sen University (Guangzhou, China). The human
   ccRCC cell lines 786-O was sourced from the American Type Culture
   Collection (ATCC; Manassas, VA, USA). To obtain the ccRCC tissues
   resistant to sunitinib, approximately 5.0×10^6 of 786-O cells were
   subcutaneously injected into the upper back of 5-week-old female BALB/c
   nude mice. When the volume of the xenograft reached 200 mm^3, the mice
   were treated orally with saline or sunitinib (APExBIO). The treatment
   course followed the 4/2 plan (40 mg/kg/day, 4 weeks of treatment,
   followed by 2 weeks of rest). After the treatment course, the
   xenotransplanted ccRCC tissues were separated, and a 3-mm-diameter
   tissue was transplanted into the mice again, and the administration was
   continued. At the end of the treatment, the above step was repeated to
   obtain the third generation of ccRCC tissues that were resistant to
   sunitinib; the subsequent ccRCC tissues were identified by the
   pathologist at our hospital. The tumor volume (V) was calculated using
   the following formula, every 3 days: V = (W^2 × L)/2.

Acquisition of Drug-Resistant ccRCC Cells

   In order to isolate the ccRCC cells from xenografts, the xenografts
   were first mechanically decomposed, washed with cold-PBS containing
   penicillin (500 U/mL) and streptomycin (500 μg/mL), and then
   transferred to the DMEM medium containing type II/IV collagenase (1
   mg/mL) and digested in a constant temperature oscillator at 37°C for 1
   h. Finally, the cell suspension was passed through 80-μM filter (BD
   Biosciences, USA) and centrifuged at 1000 rpm for 5 min. The
   precipitate was suspended and inoculated in a Petri dish. After passage
   3 generations, the half inhibitory concentration (IC50) of sunitinib on
   the isolated tumor cells was detected.

IC50 Detection

   The tumor cells (1 × 10^3; 100 µL/well) were seeded into a
   96-well-plate and then placed in the incubator for 24 h, sunitinib was
   added to make the final concentration of 0, 1.5, 3, 6, 12, 25, 50, and
   100 µM. There were 3 replicate wells in each group. After 48 h, the
   liquid was aspirated, media containing 20% MTS was added, and the
   solution was incubated in the dark for 2 h; the optical density of the
   resultant cells was measured at 492 nm.

Sequencing and DEGs Identification

   The xenograft tumor tissues (3 resistant tissues and 3 control tissues)
   were sent to the Lianchuan Biological Company for human transcriptome
   mRNA sequencing. The “limma” package in the R software was used to
   screen the differentially expressed genes (DEGs) between drug-resistant
   and control tissues in the sequencing results. The threshold of DEGs
   was as follows: log2|(FC)|> 2, P < 0.05, the false discovery rate (FDR)
   < 0.05.

Functional Enrichment Analyses

   In order to evaluate the function of DEGs in the drug-resistance tumor,
   “Cluster Profiler” package in the R software was used to conduct GO
   annotation and KEGG pathway enrichment analyses.[54]10 The ruling
   standard was set as P < 0.05.

PPI Network Construction and Module Function Analyses

   In order to evaluate the interaction between DEGs, the online search
   tool STRING database was employed to construct a protein-protein
   interaction (PPI) network, which can provide a comprehensive
   interaction between proteins and genes.[55]11 The PPI network was
   visualized using the Cytoscape.[56]12 The mcode plug-in of Cytoscape
   was used to extract important modules from the PPI network for
   analyzing the interaction relationship of DEGs. The default threshold
   included a degree cut-off: 2, node score cut-off: 0.2, k-core: 2, and
   max depth: 100.[57]13,[58]14

Identification of Hub Genes

   Hub gene is a gene with high connectivity. The Maximum Group Centrality
   (MCC) algorithm in cytohubba plug-in was used to identify hub genes,
   and the top 10 hub genes with the highest MCC score were subsequently
   analyzed.[59]15

Verify the mRNA and Protein Expression Levels of Hub

   We downloaded the ccRCC cohort data from the TCGA database in July
   2021, which included the patient gene expression data and clinical
   information. We then used the “limma” and “beeswarm” packages in the R
   software to extract the mRNA expression of the Hub genes and compared
   the results between cancerous tissues and the adjacent tissues. In
   addition, the HPA database was referred to conduct an online
   immunohistochemical analysis to determine the distribution of the hub
   genes and the protein levels.[60]16

Prognostic Analysis

   The prognosis of hub genes was analyzed by the “survival” package in
   the R software. A total of 500 ccRCC patients followed up for >90 days
   in the TCGA database were selected. The survival curve was drawn by
   Kaplan-Meier method. P < 0.05 was considered to indicate statistical
   significance.

Results

Construction and Identification of Sunitinib-Resistant ccRCC CDX Model

   In this study, a CDX model of ccRCC was established by injecting 786-O
   subcutaneously into mice and treating them with sunitinib to obtain
   sunitinib-resistant tumors in the third generation. The specific
   operation process is depicted in [61]Figure 1A. The growth curves of
   the tumor in the CDX model of the control group, the first generation
   experimental group, and the third generation experimental group are
   shown in [62]Figure 1B. The results indicated that the third-generation
   tumor had developed resistance to sunitinib. The tumor tissues were
   then dispersed to obtain isolated tumor cells, and the IC50 test of
   sunitinib was performed, which suggested that the IC50 of the third
   generation tumor cells was 12.5μM. This value was significantly higher
   than that of the control group (4.0μM), which indicated that the
   isolated ccRCC cells had also developed resistance to sunitinib in
   vitro ([63]Figure 1C). In summary, our established CDX model of
   sunitinib-resistant ccRCC was successful and reliable.

Figure 1.

   [64]Figure 1
   [65]Open in a new tab

   Construction and identification of the CDX model of ccRCC sunitinib
   resistance. (A) Flow chart of ccRCC sunitinib-resistance model
   construction. (B) The curve of tumor volume with time in the control
   group, the CDX1-Su group, and the CDX3-Su group. (C) The half
   inhibitory concentration (IC50) of sunitinib on tumor cells in the
   control and CDX3-Su groups.

Identification of DEGs

   The R package was used to analyze the sequencing data of
   sunitinib-resistant and sensitive tumors in ccRCC. The volcano map of
   differential gene expression is presented in [66]Figure 2A (all
   differential genes are shown in [67]Supplementary Excel S1: Sheet1).
   Based on the thresholds of P < 0.05 and log2|FC| > 2, there were 646
   upregulated and 465 downregulated genes in the resistant tissues
   ([68]Figure 2B).

Figure 2.

   [69]Figure 2
   [70]Open in a new tab

   Identification of genes that were differentially expressed in
   drug-resistant tissues and the relatively sensitive tissues. (A) The
   volcano map reflecting the differentially expressed genes. Red color
   represents upregulated genes, while the blue color represents
   downregulated genes. The cutoff threshold was |logFC|>2.0 and adj. P <
   0.05. (B) The histogram depicts that, among the genes with differential
   expression, there were 646 upregulated genes and 465 downregulated
   genes.

PPI Network Construction, GO Annotation, and KEGG Pathway Enrichment Analysis

   To comprehend the interaction between the DEGs, the STRING database was
   used to determine the PPI pairs among the 1011 DEGs. As shown in
   [71]Figure 3A, the constructed PPI network included 613 nodes (genes)
   and 4056 edges (interactions) (threshold: interaction score > 0.4). To
   obtain a deeper understanding of the biological effects of these DEGs,
   the “Cluster Profiler” package in R was used to perform GO annotation
   and KEGG pathway enrichment analysis (for detailed results, see
   [72]Supplementary Excel S1: Sheet2 and Excel S1: Sheet3). The top 20
   results are shown in the figure. GO analysis revealed that these DEGs
   were significantly enriched in the collagen-containing extracellular
   matrix, angiogenesis, cellular response to hypoxia, and positive
   regulation of apoptotic process ([73]Figure 3B). In addition, the KEGG
   pathways were significantly enriched in the PI3K-Akt signaling pathway,
   HIF-1 signaling pathway, NF-kappa B signaling pathway, and MAPK
   signaling pathway ([74]Figure 3C). Our results are consistent with the
   current understanding of sunitinib resistance, and bypass activation
   may be involved in sunitinib resistance.

Figure 3.

   [75]Figure 3
   [76]Open in a new tab

   PPI network diagram of differentially expressed genes, as well as GO
   and KEGG analysis. (A) The PPI network diagram is composed of 613 genes
   and is visualized by Cytoscape. The interaction score > 0.4. (B) GO
   enrichment analysis results of genes in the PPI network. (C) KEGG
   analysis results of genes in the PPI network.

PPI Network Module Analysis

   Since the PPI network constructed by us was relatively large, the MCODE
   plugin was used to calculate the modules with an MCODE score of >5 and
   identify the most important modules. Finally, a total of 6 modules (for
   detailed results, see [77]Supplementary Excel S1: Sheet4 and Figure S1)
   was identified, out of which, two (modules 1 and 3) were rich in
   biological processes and pathways related to tumor progression. These
   modules were considered to be functionally important. There were 23
   genes in module 1 (12 upregulated genes and 11 downregulated genes), as
   shown in [78]Figure 4A. There existed a total of 17 genes in module 3
   (13 upregulated genes and 4 downregulated genes), as shown in
   [79]Figure 4B. GO analysis of module 1 signified that the main
   enrichment was in ubiquitin–protein transferase activity. KEGG analysis
   showed that the main enrichment pathway was ubiquitin-mediated
   proteolysis ([80]Figure 4C). The GO analysis of module 3 implied that
   it was mainly enriched in cell adhesion molecule binding, insulin-like
   growth factor I binding, and growth factor binding. KEGG analysis
   alluded that the main enriched pathway was the p53 signaling pathway,
   and PI3K-Akt signaling pathway ([81]Figure 4D). According to module
   analysis, it appears that ubiquitinase may play an important role in
   sunitinib resistance.

Figure 4.

   [82]Figure 4
   [83]Open in a new tab

   Modular analysis of PPI network (A) Module 1 contains 23 genes. (B)
   Module 3 contains 17 genes. (C) GO and KEGG analysis of genes in Module
   1. (D) GO and KEGG analysis of genes in Module 3.

Hub Gene Identification

   To identify the hub genes in the interaction network, the CytoHubba
   plugin was used to determine the top 10 hub genes based on the MCC
   score. The results included UBC, TRIP12, ZNRF1, UBA52, BTRC, NEDD4L,
   ITCH, TRIM32, ANAPC5 and ANAPC7 ([84]Figure 5A) (upregulated genes:
   UBC, ZNRF1, UBA52, ITCH, and ANAPC5; downregulated genes: TRIP12, BTRC,
   NEDD4L, TRIM32, and ANAPC7). Interestingly, the hub genes determined
   based on the MCC score were consistent with the highest score module 1
   from MCODE analysis. The GO analysis revealed that the hub genes were
   highly enriched in ubiquitin–protein transferase activity, protein
   phosphatase binding, and CXCR chemokine receptor binding ([85]Figure
   5B). The KEGG pathway enrichment included ubiquitin-mediated
   proteolysis and Cell cycle ([86]Figure 5C). Based on the above
   analyses, the hub gene function identified from the PPI network was
   mainly found to be related to the ubiquitinating enzyme activity. This
   result suggests that the ubiquitinating enzyme system may play an
   important role in sunitinib resistance.

Figure 5.

   [87]Figure 5
   [88]Open in a new tab

   Hub gene analysis of PPI network (A) The top ten genes calculated by
   the MCC algorithm, the darker the color, the higher the ranking. (B) GO
   enrichment results of hub genes. (C) KEGG analysis result of Hub genes.

Comprehensive Expression Analysis of the Hub Genes at the mRNA and Protein
Levels and Prognosis in ccRCC

   The transcriptome data of ccRCC was downloaded from the TCGA database,
   which included 539 cases of cancer tissues and 72 cases of adjacent
   tissues. As shown in [89]Figure 6, the mRNA expression of UBC, UBA52,
   ANAPC5 and ANAPC7 in the ccRCC tissues were significantly higher than
   that in the adjacent tissues (P<0.05). The mRNA expression of TRIP12,
   ZNRF1, BTRC, NEDD4L, ITCH and TRIM32 in the ccRCC tissues were
   significantly lower than the adjacent tissues (P<0.05).

Figure 6.

   [90]Figure 6
   [91]Open in a new tab

   TCGA database was used to analyze the difference mRNA expression of hub
   genes between ccRCC and adjacent tissues.

   The protein expressions of the hub genes in ccRCC were analyzed through
   the online HPA database, as shown in [92]Figure 7. The protein
   expression of UBC, TRIP12, UBA52, NEDD4L and ANAPC7 in the ccRCC
   tissues were significantly higher than that in normal tissues. On the
   other hand, the protein expression of ZNRF1, BTRC, ITCH, TRIM32 and
   ANAPC5 in the ccRCC tissues were significantly lower than that in the
   normal tissues.

Figure 7.

   [93]Figure 7
   [94]Open in a new tab

   HPA database was used to analyze the difference protein expression of
   hub genes between ccRCC tissues and normal tissues. (A) UBC protein is
   highly expressed in ccRCC. (B) TRIP12 protein is highly expressed in
   ccRCC. (C) ZNRF1 protein is low expressed in ccRCC. (D) UBA52 protein
   is highly expressed in ccRCC. (E) BTRC protein is low expressed in
   ccRCC. (F) NEDD4L protein is highly expressed in ccRCC. (G) ITCH
   protein is low expressed in ccRCC. (H) TRIM32 protein is low expressed
   in ccRCC. (I) ANAPC5 protein is highly expressed in ccRCC tissues and
   normal tissues. (J) ANAPC7 protein is highly expressed in ccRCC.

   The clinical information of patients with ccRCC was downloaded from the
   TCGA database, and the prognosis based on the hub genes was analyzed.
   The results indicated that the high expression of TRIP12, BTRC, NEDD4L,
   ITCH, and TRIM32 improved the prognosis of patients with ccRCC, while
   the high expressions of ANAPC5 and ANAPC7 suggested a poor prognosis
   ([95]Figure 8).

Figure 8.

   [96]Figure 8
   [97]Open in a new tab

   TCGA database was used to analyze the effect of Hub genes expression on
   the prognosis of ccRCC.

   The Venn diagram showed that there were no common genes in the
   intersection of hub genes with upregulated mRNA expression in the
   drug-resistant CDX model, hub genes with up-regulated mRNA and protein
   expressions in ccRCC, and hub genes related to poor prognosis
   ([98]Figure 9A). On the other hand, there were two genes (BTRC and
   TRIM32) in the intersection of hub genes with the downregulated mRNA
   expression in the drug-resistant CDX model, hub genes with
   downregulated mRNA and protein expressions in ccRCC, and hub genes
   related to good prognosis ([99]Figure 9B). The details of BTRC and
   TRIM32 genes are presented in [100]Table 1. It was found that BTRC and
   TRIM32 belonged to E3 ubiquitinase. In conclusion, we believe that BTRC
   and TRIM32, as E3 ubiquitinases, may play key roles in the sunitinib
   resistance in RCC.

Figure 9.

   [101]Figure 9
   [102]Open in a new tab

   Venn diagram. (A) The intersection of the upregulated hub genes and the
   hub genes with poor prognosis. Blue circles represent hub genes with
   upregulated mRNA expression in drug-resistant tissues; yellow circles
   represent hub genes with upregulated mRNA expression in ccRCC; green
   circles represent hub genes with upregulated protein expression in
   ccRCC; red circles represents hub genes related to poor prognosis. (B)
   The intersection of downregulated hub genes and hub genes with good
   prognosis. The blue circles represent hub genes with downregulated mRNA
   expression in drug-resistant tissues; the yellow circles represent hub
   genes with downregulated mRNA expression in ccRCC; the green circles
   represent hub genes with downregulated protein expression in ccRCC; and
   the red circles represent hub genes related to good prognosis.

Table 1.

   Summary of the Functions of 2 Hub Genes
   Gene Symbol Full Name Function
   BTRC Beta-transducin repeat containing E3 ubiquitin protein ligase This
   gene encodes a member of the F-box protein family which is
   characterized by an approximately 40 amino acid motif, the F-box. The
   F-box proteins constitute one of the four subunits of ubiquitin protein
   ligase complex called SCFs (SKP1-cullin-F-box), which function in
   phosphorylation-dependent ubiquitination.
   TRIM32 Tripartite motif containing 32 The protein encoded by this gene
   is a member of the tripartite motif (TRIM) family. The TRIM motif
   includes three zinc-binding domains, a RING, a B-box type 1 and a B-box
   type 2, and a coiled-coil region. It is considered a new type of
   ubiquitin E3 ligase.
   [103]Open in a new tab

Discussion

   RCC is a highly vascularized tumor. In the early stages of tumor
   formation, RCC acquires an ability to synthesize angiogenic factors and
   form neovascularization.[104]17 Since the advent of antiangiogenic
   therapy, the overall survival (OS) rate of patients with RCC has
   increased drastically from 13 months in the cytokine era 10 years
   ago[105]18 to 30 months in 2014.[106]19,[107]20 Sunitinib is the first
   generation of multitarget TKI-targeted drugs, and can non-covalently
   bind to several tyrosine kinase receptors and inhibit their
   phosphorylation. In a mice model of RCC tumors, sunitinib has been
   shown to reduce vascular permeability, inhibit neovascularization, and
   destroy the generated tumor vessels.[108]21

   The application of targeted drugs has improved the prognosis of
   patients with RCC. However, with the application of targeted drugs,
   drug resistance has gradually emerged and has lowered the clinical
   effects. Although combined medication can help overcome drug
   resistance, it is fraught with the disadvantages of high cost, drug
   toxicity, and side effects, which have posed several hurdles in the
   treatment of patients.[109]22

   The mechanisms of primary and secondary drug resistance of sunitinib
   often overlap each other. The drug resistance mechanisms can be divided
   into activation of angiogenesis signaling pathway, alteration of tumor
   microenvironment, increase of tumor invasion and metastasis, role of
   non-coding RNA, and activation of other signaling pathways.[110]23–25
   Some reports have been published on the mechanism of targeted drug
   resistance in RCC; however, the underlying molecular mechanism is not
   clear. Furthermore, there is a lack of effective biomarkers to predict
   targeted drug resistance. Also, there are no effective treatments to
   reverse or delay drug resistance after it occurs.

   To decipher the molecular mechanism of sunitinib resistance in ccRCC, a
   CDX mouse model of sunitinib resistance in ccRCC was established in
   this study. By sequencing the drug-resistant and sensitive tissues,
   DEGs were identified and their expression data were analyzed using
   bioinformatics tools. The results suggested the following: (1) When
   compared with the sensitive tissues, sunitinib-resistant tissues had
   646 upregulated genes and 465 downregulated genes. GO analysis showed
   that these DEGs were mainly enriched in angiogenesis, cellular response
   to hypoxia, and positive regulation of apoptotic process. KEGG analysis
   implied that phosphatidylinositol-3-kinase-(PI3K)-Akt,
   hypoxia-inducible factor (HIF)-1, nuclear factor kappa-B (NF-kappa B),
   and MAPK signaling pathways were significantly enriched. (2) Through
   the construction of PPI network and module analysis, six important
   modules were identified. The GO and KEGG analyses of module 1 showed
   that the network was mainly enriched in ubiquitin-mediated proteolysis,
   which indicates that ubiquitin may play an important role in sunitinib
   resistance. The GO and KEGG analyses of module 3 showed that the p53
   and PI3K-Akt signaling pathways were significantly enriched, which
   suggests that these pathways may have key functions in regulating
   sunitinib resistance. (3) Ten hub genes were identified in the DEGs. By
   analyzing the expressions of mRNA and protein and their impacts on the
   prognosis of patients with ccRCC, BTRC and TRIM32 were finally found to
   be consistent in expression and prognosis. It is speculated that these
   two genes may be chiefly involved in the sunitinib resistance in ccRCC.

   At present, the research on sunitinib resistance in RCC mainly includes
   cell lines, CDX models, and PDX models. However, it is difficult to
   obtain drug-resistant tissues from patients owing to ethical reasons.
   Studies on animal models of sunitinib resistance in RCC are limited.
   Zhang established a CDX model of sunitinib resistance in ccRCC and
   found that LncARSR promotes the expressions of AXL and c-MET and the
   activation of downstream STAT3, AKT, and ERK signaling pathways, which
   leads to sunitinib resistance in ccRCC.[111]26 Diaz-Montero created a
   sunitinib-resistant PDX model for ccRCC. They discovered that during
   single-agent TKI treatment, the tumor showed reduced MEK
   phosphorylation. However, after combined TKI/MEKi treatment, it
   effectively inhibited phosphorylated MEK and phosphorylated ERK, which
   can prolong the antitumor effect of sunitinib.[112]27 In this study,
   resistance-related DEGs were found to be enriched in PI3K-Akt, HIF-1,
   NF-kappa B, and MAPK signaling pathways. However, network module
   analysis revealed that the GO and KEGG analyses results of the highest
   ranked module 1 (containing 23 genes) were mainly enriched in
   ubiquitin–protein transferase activity and ubiquitin-mediated
   proteolysis. Our finding suggests that the ubiquitinase system may play
   an important role in the resistance of ccRCC to sunitinib.

   Furthermore, the hub genes in the PPI network were calculated using the
   MCC algorithm. The top 10 genes were UBC, TRIP12, ZNRF1, UBA52, BTRC,
   NEDD4L, ITCH, TRIM32, ANAPC5, and ANAPC7. It is worth noting that these
   10 genes completely overlapped with the genes in module 1. The GO and
   KEGG analyses of these hub genes were predominantly enriched in
   ubiquitin–protein transferase activity and ubiquitin-mediated
   proteolysis. Module and hub gene analyses further proved that
   ubiquitinase may play an important role in sunitinib resistance.
   Additionally, the mRNA and protein expressions of the hub genes in
   ccRCC and their influence on the prognosis of patients with ccRCC were
   analyzed. As a result of these analyses, BTRC and TRIM32 were finally
   identified to be the key genes, with consistent results in all studies.
   Hence, we speculate that these two genes may play a pertinent role in
   sunitinib resistance.

   Beta-transducing repeat containing E3 ubiquitin protein ligase (BTRC)
   encodes F-box protein, which contains multiple WD-40 repeat sequences.
   On the one hand, BTRC has been shown to ubiquitinate phosphorylated
   NFKBIA and target its degradation, thereby activating NF-κB, promoting
   its translocation to the nucleus, and activating the transcription of
   metastasis-related genes.[113]28,[114]29 For instance, Zhang found that
   TSPAN15 interacts with BTRC to promote phosphorylated-inhibitor of
   NF-κB alpha (p-IκBα) ubiquitination and proteasomal degradation,
   thereby triggering NF-κB nuclear translocation and the subsequent
   transcriptional activation of several metastasis-related genes. These
   events promote the metastasis of esophageal squamous cell
   carcinoma.[115]30 Similarly, in triple-negative breast cancer, Lim
   found that WBP2 induces the stability BTRC, promotes its targeting of
   IκBα for ubiquitination and degradation, and enhances cancer
   progression.[116]31 On the other hand, BTRC has also been shown to play
   a vital role in the epithelial–mesenchymal transition (EMT) process
   because it facilitates the ubiquitination of Snail in cancer. The
   inhibition of BTRC leads to the upregulation of Snail, thereby inducing
   EMT.[117]32,[118]33 For example, Zheng identified that miR-224 inhibits
   the expression of its target BTRC protein and promotes the associated
   Wnt/β-catenin pathway to cause the migration and invasion of rectal
   cancer cells.[119]34 Zhou discovered that miR-193a-3p is upregulated in
   patients with glioma and promotes the invasion, migration, and EMT of
   glioma by inhibiting the expression of BTRC protein.[120]35

   Tripartite motif containing 32 (TRIM32) is a new type of ubiquitin E3
   ligase. The protein encoded by this gene is a member of the tripartite
   motif (TRIM) family. The TRIM motif includes three zinc binding
   domains, a RING, a B-box type 1 and a B-box type 2, and a coiled-coil
   region. Studies have demonstrated that TRIM32 plays an important role
   in regulating cell death induced by oxidative stress. TRIM32 activates
   the caspase-mediated apoptosis pathway under conditions of oxidative
   stress.[121]36 A recent study has shown that TRIM32 enhances X-linked
   inhibitor-of-apoptosis protein (XIAP) turnover during TNF-mediated
   apoptosis.[122]37 A decrease in the level of TRIM32 may increase the
   level of XIAP. Furthermore, XIAP has been proven to act as an
   antioxidant by regulating the expressions of several antioxidant genes,
   including superoxide dismutase type 1 (SOD1), SOD2, NAD (P) H: quinone
   oxidoreductase 1, heme-oxygenase-1, and thioredoxin 2.[123]38

   The findings of this study showed that BTRC and TRIM32 belong to the
   same E3 ubiquitin ligase. Although some functions have been studied in
   tumors, there is no research on angiogenesis or antiangiogenesis
   resistance. Moreover, another gene of the same E3 ubiquitin ligase, von
   Hippel–Lindau (VHL), was noted to be involved in tumor angiogenesis.
   The activation of hypoxia signaling pathway in tumors is closely
   related to the increase in vascular endothelial growth factor (VEGF)
   synthesis and angiogenesis. The core of the hypoxia pathway is the
   formation of the HIF complex. The HIF-1 complex can bind to the hypoxia
   response elements (HRE) on the promoter of the target gene to regulate
   downstream gene transcription.[124]39 The HIF-1 complex consists of two
   subunits, namely HIF-1α and HIF-1β. Among them, HIF-1α undergoes
   proline hydroxylation under the action of proline hydroxylase and then
   combines with VHL E3 ligase to be degraded via the ubiquitinated
   proteasome pathway.[125]40 In ccRCC, the deletion of the VHL gene
   results in the inability to degrade HIF-1α, and its accumulation
   enhances the expression of its downstream products (VEGF, transforming
   growth factor-α, CXC motif chemokine receptor type 4, etc.) and
   promotes angiogenesis.[126]41 Whether BTRC and TRIM32 have functions
   similar to those of VHL remains to be investigated.

Conclusion

   In this study, a CDX model of ccRCC sunitinib resistance was
   successfully constructed, and a series of bioinformatics analyses were
   performed on the DEGs related to the resistance via sequencing
   analysis. Finally, some key hub genes that may be related to the
   resistance of ccRCC to sunitinib treatment were identified. Among the
   identified genes and their related pathways, ubiquitin ligase activity
   was established for the first time to play a key role in drug
   resistance. The dysregulation of ubiquitin ligase expression was
   predicted to be a potential therapeutic target or biomarker for ccRCC
   sunitinib resistance. However, further molecular and biological
   experiments are required to confirm the functions of the key genes in
   ccRCC sunitinib resistance.

Acknowledgments