Targeting lysosome function causes selective cytotoxicity in VHL-inactivated renal cell carcinomas

Abstract

The inactivation of the tumor suppressor gene, von Hippel-Lindau (VHL), has been identified as the earliest event in renal cell carcinoma (RCC) development. The loss of heterogeneity by chromosome 3p deletion followed by inactivating mutations on the second VHL copy are events present in close to 90% of patients. Our study illustrates a lysosomal vulnerability in VHL-inactivated RCC in vitro. By investigating the mechanism of action of the previously identified STF-62247, a small bioactive compound known for its selective cytotoxic properties towards VHL-defective models, we present the promising approach of targeting truncal-driven VHL inactivation through lysosome disruption. Furthermore, by analyzing the open platform for exploring cancer genomic data (cbioportal), we uncover the high alteration frequency of essential lysosomal and autophagic genes in sequenced biopsies from clear cell RCC patient primary tumors. By investigating lysosome physiology, we also identify VHL-inactivated cells’ inability to maintain their lysosomes at the perinuclear localization in response to STF-62247-induced stress and accumulate cytoplasmic inclusion bodies in response to an inefficient lysosomal degradative capacity. Finally, by testing other known lysosomal-disrupting agents (LDAs), we show that these are selectively cytotoxic to cells lacking VHL functions. Our study builds a strong platform that could specifically link genetic clonal ccRCC evolution to lysosomal and trafficking vulnerabilities.

Inactivation of the von Hippel-Lindau (VHL) gene is the earliest event leading to full blown renal cell carcinoma (RCC). Our study highlights the presence of a lysosome vulnerability specific to VHL-inactivated RCC that could be further exploited.

Introduction

Renal cell carcinoma (RCC) is the most frequent type of kidney cancer with an increasing incidence rate in developed countries (1). Approximately one third of all patients are diagnosed with metastatic disease for which available therapies have a limited efficacy. The clear cell histologic subtype (ccRCC) is the most commonly diagnosed and is characterized by the combined loss of chromosome 3p with inactivating mutations on the von Hippel-Lindau (VHL) tumor suppressor gene (2). VHL loss makes ccRCC highly vascularized and aggressive because of the consequent stabilization of hypoxia-inducible factors-alpha (HIF-α) (3). Multi-region exome sequencing studies have highlighted high intratumoral heterogeneity in ccRCC and have shown that the presence of geographically localized subclonal driver events in these tumors is likely responsible for the acquisition of resistance to available therapies in metastatic disease (4). VHL inactivation, however, has been shown to be present on the trunk of tumor phylogenetic trees characterizing it as the earliest event in ccRCC evolution (5). Thus, targeting truncal drivers such as VHL reveals a promising approach to overcome tumor resistance and ameliorate patient survival.

The possibility of achieving this type of strategy was demonstrated with the identification of a 4-pyridyl-2-anilinothiazole (PAT), STF-62247, in a library screen of 64,000 compounds (6). This molecule showed therapeutically relevant properties by triggering selective cell death in VHL-inactivated models and reducing tumor growth in vivo (7). Our most recent study has re-classified STF-62247 as a potent blocker of late stages of autophagy and has linked STF-62247’s selectivity to lysosomal disruption, thus, suggesting a possible role for VHL in maintaining lysosome fitness (8). Specifically, we have shown that cells lacking VHL functions were unable to surmount a block in autophagy and could not restore their lysosome numbers during prolonged STF treatment time. This type of targeted treatment could be beneficial in combinatory treatments in hopes of sensitizing resistant tumors to already available anti-cancer strategies (9). Autophagy confers stress tolerance and promotes cell survival in established tumors, in this way, targeting late stages of this pathway has also shown promise (10,11). During malignant transformation, lysosomes up-regulate their cellular movement and accordingly increase the peripheral lysosome population (12). In fact, to fulfill the excessive needs of cancer cells, lysosomes undergo a series of molecular and functional changes such as increased motility, biogenesis and hydrolase activity (13,14).

This study sought to investigate lysosomes from VHL-defective ccRCC models. Following our previous findings of an inability of VHL-inactivated cells to restore lysosomal numbers in response to long-term STF-62247 treatments, we hypothesized that lysosomes in VHL-null ccRCCs were more vulnerable and targeting them could present a promising strategy to overcome tumor resistance (8). By analyzing sequenced biopsies originating from primary ccRCC patient tumors, we uncover the presence of altered lysosomal and autophagy genes in ~25% of all cases. We show that STF-62247 selectively affects the degradative function and mobility of lysosomes from VHL-defective cells and renders them unable to maintain their perinuclear localization. We demonstrate that other widely used lysosomal-disrupting agents (LDAs) selectively decrease VHL-inactivated cell survival without affecting cells with a functioning wild-type gene. Interestingly, a newly generated VHL-inactive STF-62247-resistant cell model was also resistant to LDAs. Our work succeeds in identifying a lysosomal vulnerability in VHL-defective models that could be further exploited for the targeting of truncal-driven VHL-inactivated ccRCC.

Methods/Materials

Cell lines and treatments

Parental ccRCC cell lines (RCC4, RCC10), their subclone counterparts expressing VHL (RCC4 VHL and RCC10 VHL) were a gift from Dr Amato Giaccia (Stanford University). The VHL-deficient A498 were a gift from Dr Réjean Lapointe (CRCHUM, Montreal). Authentication of all parental ccRCC cell lines was performed by short tandem repeat DNA profile at Genetica DNA Laboratories (Burlington, NC) in June 2016. HeLa, A549, HCT-116, PANC10.05 and BXPC3 models were purchased from ATCC (#CCL-2, CCL-185, CCL-247, CRL-2547 and CRL-1687). All cell lines tested negative for mycoplasma contamination. To determine the cell volume, cells were trypsinized and concentration (proliferation) and volume were monitored on an Orflo Moxi Z automated cell counter (Ketchum, ID).

Treatments were as follows; STF-62247 (Cayman Chemical, #13084), 1.25 µM in RCC4, RCC10, RCC4VHL, RCC10VHL, RCC4VHL cr.VHL, A498, PANC10.05, and BXPC3, 3 µM in A549 and HCT-116 and 5 µM in HeLa cells. Bafilomycin A1 (Cayman Chemical, #11038), 1 and 2 nM in RCC4 and RCC4VHL. Chloroquine (Cayman Chemical, #14194), 2.5 and 5 µM in RCC4 and RCC4VHL. Concanamycin A (Cayman Chemical, #11050), 0.25 and 0.50 nM in RCC4 and RCC4VHL. Ciprofloxacin (Cayman Chemical, #14286), 100 and 125 µg/ml in RCC4 and RCC4VHL.

cBIOportal Cancer Genomics analysis

This open-access tool provides web resource for exploring, visualizing, and analyzing freely-available cancer genomics data and contains cancer genomes data sets from 69 cancer studies with a total number of 17,177 samples (15,16). A set of nine genes (ATG5, ATG12, CTSD, RAB5A, RAB7A, RAB11A, SQSTM1, LAMP-1 and LAMP2) were chosen. A minimum of 5% alteration summary and a minimum of 50 samples per study were required to be included in the analysis.

CRISPR/Cas9 models

Crispr/Cas9 sgRNA sequences for VHL are the following; 5′-CACCGTGACTAGGCTCCGGACAAC-3′ (forward); 5′-AAACGGTTGT CCGGAGCCTAGT CAC-3′ (reverse) and 5′-CACCGCGCTCTTTC AGAGTATACAC-3′ (forward); 5′-AAACG-TGTATACTCTGAAAGAG CGC-3′ (reverse). These sequences were inserted in lentiCRISPRv2 plasmid, a gift from Feng Zhang (Addgene; #52961) as previously described (17,18).

Establishment of STF-62247-resistant VHL-defective cell line

VHL-inactivated RCC4 cells were exposed to increasing concentrations of STF-62247 and kept in culture for 10 weeks. Resistant RCC4 cells were kept in Dulbecco's modified Eagle's medium (DMEM) high-glucose supplemented with 10% fetal bovine serum (FBS) and 6 µM of STF-62247. Experiments were performed after culturing RCC4-resistant cells for at least 2 weeks in the absence of STF-62247.

XTT viability and clonogenic survival assays

XTT viability assays were performed as previously described (19). Briefly, 5000 (RCC4, RCC4VHL, A498, PANC10.05 and BXPC3) or 2500 (A549 and HCT-116) cells were seeded per well in 96-well plates. Cells were treated with serial dilutions of STF-62247 after which plates were incubated at 37°C under 5% CO₂ for 96 h. XTT solution (comprising 0.3 mg/ml of XTT powder (Sigma–Aldrich, #X4626), DMEM high glucose without phenol red (Wisent Bio, #319-051-CL), 20% FBS and 2.65g/ml phenazine methosulfate (Sigma–Aldrich, #P9625) was added to each well and plates were incubated at 37°C for 1 h. Absorbance was read at 450 nm on a Spectramax Plus spectrophotometer (Molecular Devices, Sunnyvale, CA). For clonogenic survival assays, 500 cells were plated in triplicate in 60 mm plates and treated 4 h later with STF-62247, Bafilomycin A1, Chloroquine, Concanamycin A or Ciprofloxacin. Plates were left at 37°C for 8 days after which they were stained with a solution of crystal violet. Colony formation was quantified by calculating the plating efficiency (PE) [(#formed colonies/# cells plated) × 100%] and then by calculating the surviving factor (SF) (#formed colonies after treatment/# plated cells × PE).

Immunoblot assays

Cells were lysed in M-PER buffer and quantified using Pierce BCA protein assay Kit (Thermo Scientific, #23225). Protein samples were separated on sodium dodecyl sulfate polyacrylamide gels (SDS-PAGE) and transferred onto polyvinylidene difluoride membranes. Chemiluminescence detection was performed on a Chemi-Doc XRS+ imager (BIO-RAD Inc, Mississauga, ON, Canada). Primary antibodies used are LC3B (#3868), LC3A (#4599), LC3C (#14736), Beclin-1 (#3495), ubiquitin (#3936), Lamp-1 (#3243) ATG9A (#13509), Rab5 (#2143), Rab7 (#9367) and EEA1 (#3288) from Cell Signaling Technologies, phospho-p62 (Serine403) from MBL (#D343-3), HIF-1alpha from BD biosciences (#610958), β-Actin (sc-47778), ubiquitin (sc-8017) p62/SQSTM1 (sc-28359), Lamp-1 (sc-20011), Lamp-2 (sc-18822) and VHL (sc-55506) from Santa Cruz Biotechnologies, and GABARAP from Abgent (#ap1821a).

Immunofluorescence

Cells were grown on coverslips at 30% confluency and were fixed in 3.7% formaldehyde (Sigma–Aldrich, #F8775) and then permeabilized with 0.25% Triton x-100 in phosphate-buffered saline. FBS was used for blocking, primary and secondary antibodies. All images were taken on an Olympus Fluoview FV1000 confocal microscope (Olympus, Center Valley, PA) with a 60× oil-immersion lens.

Long-lived protein degradation

RCC4, RCC4 VHL, RCC10 and RCC10 VHL cells were incubated overnight with labeling medium (DMEM/10% dialyzed FBS containing [¹⁴C] valine (0.2 μCi/ml) in 65 μM unlabeled valine) The following day, the labeling medium was exchanged to a chase medium (DMEM/10% FBS/10 mM unlabeled valine) and incubated for 2 h to remove the contribution of short-lived proteins. Afterwards, DMEM/10 mM valine with 1.25 µM of STF-62247 added to the cells for 24 h. Cells and media were collected and separately subjected to trichloroacetic acid (TCA) precipitation by the addition of ice-cold TCA to a final concentration of 10%. The percentage of lysosomal degradation was quantified as (TCA-soluble counts from medium)/(TCA-soluble counts from medium) + (TCA-insoluble counts from the cells).

Human PCR autophagy array

The Human Autophagy RT2 Profiler PCR array (Qiagen, Germantown, MD) was used to monitor 84 autophagy-specific gene expression profiles in accordance with the manufacturer’s recommendations. Briefly, total RNA from RCC4 and RCC4VHL cells, treated with STF-62247 for 48 h, was extracted using Trizol (Invitrogen) and purified using the RNeasy Mini kit (Qiagen). Reverse transcription was accomplished using RT² First Strand kit (Qiagen). Following cDNA synthesis, real-time polymerase chain reaction (PCR) was performed using the RT² SYBR Green Mastermix and the RT2 profiler array format in 384-well plates. Fluorescence was analyzed with a LightCycler 480 Realtime PCR (Roche) and results were quantified by ΔΔCt method accordingly to Qiagen resources.

RNA isolation and real-time-PCR (qRT-PCR)

Total RNA was extracted by TRIzol/chloroform according to the manufacturer’s protocol (Invitrogen, ThermoFisher). The reverse transcription was performed with SuperScript III Reverse Transcriptase (Invitrogen, USA) according to the manufacturer’s instructions using 5 µg of RNA followed by 1 µL of RNase H. Primer sequences for p62 and LC3B were provided by PrimerBank (see http://pga.mgh.harvard.edu/primerbank/). Fluorescence was analyzed with a Mastercycler Realplex² (Eppendorf) and results were quantified by ΔΔCt method.

Results

Autophagy and lysosomal genes are frequently altered in ccRCC patient samples

We have previously identified STF-62247, a compound presenting cytotoxicity towards VHL-defective ccRCC models without affecting the viability of cells with a functioning gene (6,7) (Figure 1A) To fully assess the selectivity of the compound, non-renal cancerous cell lines were evaluated with treatment (Figure 1B and C). Endogenous VHL protein levels and the ability of this tumor suppressor to regulate HIF-1α were investigated by immunoblot assays in a pulmonary carcinoma (A549), a colon colorectal carcinoma (HCT-116) and in two pancreatic adenocarcinomas models (PANC10.05 and BXPC3) and compared with two VHL-inactivated and one VHL-proficient ccRCC models (RCC4, A498 and RCC4VHL, respectively). A549, HCT-116, PANC10.05 and BXPC3 expressed either VHL30 or the most predominant and functionally redundant isoform VHL19 and showed no detectable protein levels of HIF-1α, demonstrating a normal VHL function in these cell models (20). VHL-inactivated RCC4 and A498 models showed an absence of VHL proteins (30 and 19) with a concomitant HIF-1α accumulation (red arrows) (Figure 1B). A truncated HIF-1α is seen in the A498 model (21) (Figure 1B). Next, viability assays were performed to assess STF-62247’s selectivity (Figure 1C). In response to increasing concentrations of STF, only VHL-inactivated models (RCC4 and A498) showed a significant decrease in cell viability, confirming our previous findings in VHL-defective ccRCC models (6,8) (Figure 1C). Cell models containing endogenous VHL protein levels (A549, HCT-116, PANC10.05 and BXPC3) remained unaffected in response to STF-62247, even at the highest concentration tested (5 µM) (Figure 1C).

Figure 1.

Autophagy and lysosomal genes are frequently altered in ccRCC patient samples. (A) Chemical structure of STF-62247. (B) Immunoblot analysis of HIF- 1α and VHL in VHL-defective cell lines (RCC4 and A498) and cancerous VHL-proficient models (RCC4VHL, A549, HCT-116, PANC10.05 and BXPC3). (C) XTT cell viability assay testing the sensitivity of RCC4, A498, RCC4VHL, A549, HCT-116, PANC10.05 and BXPC3 models to increasing doses of STF (0–5 µM). (D) Immunoblot analysis of HIF-1α and VHL levels in control RCC4, RCC4VHL, RCC4 VHL.670–1 and RCC4VHL Cr.VHL. The added line denotes non-contiguous lanes on the membrane. (E) Assessment of endo-lysosomal swelling by inverted-light microscopy images of RCC4, RCC4VHL and RCC4VHL Cr.VHL in response to 24 h of STF treatment (1.25 µM). (F) Clonogenic survival assay testing increasing concentrations of STF (0–1.25 µM) in VHL-proficient cells (RCC4VHL and RCC4VHL.670–1) as well as VHL-inactivated cells (RCC4 and RCC4VHL Cr.VHL). (G) Mutation, Deletion, Amplification and multiple alterations analysis of cBIOPortal for cancer genomics data of a set of nine autophagy and lysosomal-related genes (ATG5, ATG12, SQSTM1, RAB11A, RAB5A, RAB7, CTSD, LAMP-1, LAMP-2) are shown. Mean and SEM were calculated from at least three independent experiment. Statistical analysis compared untreated cells with treated cells (*P < 0.05, **P < 0.01, ***P < 0.001) or the difference between cell models (#P < 0.05, ##P < 0.01, ###P < 0.001) for each concentration using Student’s T-test.

Recently, we showed that STF-62247’s cytotoxicity was due to uncontrolled swelling of endo-lysosomal structures that could only be resolved by VHL-functioning cells (8). To evaluate the ability of VHL in maintaining the swelling phenotype, VHL was knocked-out in the parental VHL-proficient RCC4VHL model (RCC4VHL Cr.VHL). To confirm the functionality of this model, VHL and its ability to regulated HIF-1α was assessed (Figure 1D). A complete VHL knockout was observed in the RCC4VHL cr.VHL model with a concomitant accumulation of HIF-1α band (Figure 1D). Next, light microscopy images were taken and VHL-inactivated RCC4 cells showed a striking vacuolization phenotype due to endo-lysosomal swelling in response to STF treatment while RCC4VHL cells were able to recover from this stress (Figure 1E). Interestingly, the VHL knock-out model (RCC4VHL Cr.VHL) rendered the cells unable to control the vacuolization in response to STF and endo-lysosomal swelling developed uncontrollably, similarly to parental VHL-defective RCC4 cells, strongly suggesting a role for VHL in maintaining endo-lysosomal integrity (Figure 1E). Then, STF’s selectivity for VHL loss was tested by clonogenic survival assay (Figure 1F). VHL knockout in RCC4VHL (RCC4VHL Cr.VHL) significantly reduced the cells’ survival comparable to RCC4 cells, confirming STF’s selective properties for VHL inactivation and correlating with the unresolved swelling phenotype.

STF-62247’s mechanism of action was previously linked to autophagy disruption (8,19). To elucidate the impact of autophagy in ccRCC development, the alteration frequency of essential autophagy and lysosomal genes was analyzed with the open platform for exploring multidimensional cancer genomics data, cBioPortal (15,16) (Figure 1G). Autophagy-related gene (ATG) 5, ATG12, Cathepsin D (CTSD), Ras-related protein RAB5A, RAB7A, RAB11A, SQSTM1, lysosome-associated-membrane-protein (LAMP)-1 and LAMP2 were chosen as they play critical roles in autophagy and lysosomal integrity (22–24). The presence of mutations, deletions, amplifications and multiple alterations was analyzed for the gene query in different cancer studies available. Strikingly, ccRCC TCGA study had the most alterations in these autophagic and lysosomal genes, second only to pancreatic cancer, with patient primary tumor samples being altered in ~25% of all cases (Figure 1G). Altogether, these results demonstrate the selective properties of STF-62247 for VHL loss and highlight the presence of multiple alterations in autophagy and lysosomal genes in approximately 25% of ccRCC patient tumor samples.

VHL-functioning cells overcome STF-62247’s block of autophagy as well as endo-lysosomal swelling

To measure the effects of the endo-lysosomal swelling, cell volume was quantified (Figure 2A). In response to STF, the cell volume of VHL-inactivated RCC4 cells was significantly increased at each time point while RCC4VHL cells retained a cell volume similar to the untreated control, confirming their ability to surmount the phenotype (Figure 2A). Cell counts showed that STF did not affect the growth of RCC4VHL cells contrastingly to VHL-defective cells which were unable to proliferate (Figure 2B). To correlate cell viability with endo-lysosomal swelling, multiple cancerous cell models were monitored in response to a time-dependent STF treatment. Initially, endo-lysosomal swelling was observed in all cell lines in response to 24 h of STF (Figure 2C). However, prolonged exposure to STF (48 and 72 h) only caused an unregulated vacuolization in VHL-inactivated RCC4 and A498 cell models which correlated with the compound’s selective cytotoxicity (Figure 2C). Contrastingly, A549, HCT-116 and PANC10.05 models started resolving the swelling phenotype at the 48 h time point with a complete recovery observed at 72 h. BXPC3 only showed modest endo-lysosomal swelling at the 24 h time point (Figure 2C). To evaluate STF’s effects on autophagy in different cell models, autophagic effectors LC3B and p62 were measured by immuoblot assays (Figure 2D and E). During autophagy, LC3B is lipidated (LC3B-II) and inserted in the autophagosome membrane on which p62 can bind to assure degradation of specific ubiquitinated cargo. Autophagy induction is thus monitored by the lysosomal degradation of p62 combined with an increased in newly formed autophagosomes, marked by lipidated LC3B (LC3B-II) (25). Accordingly, a block in autophagy can be assessed by the simultaneous accumulation of LC3B-II and p62 protein levels. Confirming our previous findings, STF acts by initially blocking autophagy in all human cancer cell models, independently of VHL status as assessed by the increased protein levels of LC3B-II and p62 (Figure 2D and E). In accordance with the resolved endo-lysosomal swelling in A549, HCT-116, PANC10.05 and BXPC3, levels of p62 and LC3B greatly decreased after 24 h and returned to levels comparable to the control untreated samples, demonstrating the ability of these models to resolve autophagy and endo-lysosomal degradation capacity (Figure 2E). These results succeed in widening STF-62247’s mechanism of action on other human cancerous cell models and show the ability of VHL-functioning cells to surmount STF-induced endo-lysosomal swelling and a block in autophagy.

Figure 2.

STF-62247 blocks autophagy in different cancer models but only causes cell death in VHL-mutated cell models. 1.25 µM in RCC4, RCC10, RCC4VHL, RCC10VHL, RCC4VHL cr.VHL, A498, PANC10.05, and BXPC3, 3 µM in A549 and HCT-116 and 5 µM in HeLa cells. (A) Quantification of cell volume (pL) in RCC4 and RCC4VHL in response to 24, 48 and 72 h of STF treatment (1.25 µM) by a MOXI Z mini automated cell counter. (B) Cell number quantification in untreated and STF-treated RCC4 and RCC4VHL (1.25 µM) every day for 6 days. (C) Inverted-light microscopy images of RCC4, A498, A549, HCT-116, PANC10.05 and BXPC3 in response to 24, 48 and 72 h STF-62247. (D) Immunoblot analysis of autophagic effectors, LC3B-II and p62 in RCC models RCC4 and A498, and in (E) various non-renal cancer models A549, HCT-116, PANC10.05 and BXPC3. RCC4, A498, PANC10.05 and BXPC3 cell models were treated at 1.25 µM and A549 with 3 µM of STF. Mean and SEM were calculated from at least three independent experiments. Statistical analysis compared untreated cells with treated cells (*P < 0.05, **P < 0.01, ***P < 0.001) at each concentration using Student’s T-test.

STF-62247 deregulates lysosomal transcriptome assessed by autophagy PCR array

To understand the ability of VHL-functioning cells to restore a functional autophagic flux, a human autophagy PCR array investigating the expression of 84 key genes involved in autophagy and lysosome integrity was performed. mRNA expression was quantified in VHL-mutated RCC4 cells as well as its isogenic clonal counterpart RCC4VHL in response to STF (Figure 3A). Among the 84 genes tested, the most up-regulated mRNAs were genes with roles in lysosome integrity and/or lysosome maturation and fusion; CTSD, NPC1, CTSB, CTSS, GAA, GABARAP, UVRAG and APP (Figure 3A) (26,27). We have previously shown that CTSD trafficking was severely altered in response to STF treatment with an accumulation of the enzyme’s inactive form and a concomitant decrease in its active form which correlates here with an increase in its mRNA levels (8). Down-regulated mRNAs were genes responsible for lysosomal stability such as the chaperones HSP90AA1 and HSPA8, involved in chaperone-mediated-autophagy, as well as TMEM74, a lysosomal and autophagosomal protein (28,29). To further validate individual genes, the relative mRNA expression of GABARAP, MAP1LC3B and SQSTM1 was quantified in response to 48 h of STF (Figure 3B). In VHL-inactivated cells, mRNA levels of all three effectors were up-regulated in response to STF (RCC4), validating the expression of these transcripts in the human autophagy array (Figure 3B). Interestingly, only GABARAP mRNA levels were highly upregulated in VHL-proficient RCC4VHL cells while MAP1LC3B levels were only moderately elevated with STF treatment (Figure 3B). To correlate mRNA data at the protein level, immunoblot analyses of ATG9A and Beclin-1 (BECN1) were performed in the parental RCC4 and RCC4VHL models (Figure 3C). In response to STF, the autophagy array revealed downregulated ATG9A mRNA in RCC4VHL cells while upregulation was observed in RCC4 cells (Figure 3A). ATG9A protein levels correlated in the RCC4VHL cells with a marked decrease in ATG9A protein levels. However, ATG9A levels also decreased in RCC4 cells (Figure 3C). Beclin-1 levels did not fluctuate and also did not correlate with transcriptomic data (Figure 3C). While the levels of autophagic transcripts can be highly informative, it has been shown that they do not alone represent a general readout for autophagic activity and thus, do not always correlate at the protein level (25). Members of the ATG8 family of proteins and p62 were also detected by immunoblot in RCC4 and RCC4VHL cells (Figure 3D). Following STF treatment, all MAP1LC3s (LC3A, LC3B and LC3C), GABARAP and p62 protein levels accumulated due to a block in autophagy. Interestingly, these results suggest that all ATG8 members are associated with the autophagosomes in these cell models (Figure 3D). Finally, LAMP-1 expression levels were assessed in RCC4 and RCC4VHL models (Figure 3E). Glycosylation of lysosome-associated proteins is crucial as it protects the lysosomal membrane from lytic enzymes found inside the lysosome (30). In response to treatment, levels of glycosylated LAMP-1 greatly increased in both cell models, indicating the presence of important STF-induced lysosomal changes (Figure 3E). Collectively, these results suggest that STF-62247 mainly deregulates genes associated with autophagy and lysosomal integrity as assessed by the human autophagy array and lamp-1 glycosylation states.

Figure 3.

STF-62247 deregulates lysosomal transcriptome assessed by autophagy PCR array. (A) Heat-map comparing transcriptomic signatures of autophagic and lysosomal genes in STF treated RCC4 and RCC4VHL models. (B) Relative mRNA expression of GABARAP, MAP1LC3B and SQSTM1 in response to 48 h of STF treatment in RCC4 and RCC4VHL cells. (C) Immunoblot analysis of ATG9A, beclin-1, D) members of the ATG8 family of proteins, LC3A, LC3B, LC3C, GABARAP and p62 in control and STF-treated (4 h) RCC4 and RCC4VHL cells. (E) Immunoblot analysis of the levels of glycosylated LAMP-1 in RCC4 and RCC4VHL in response to 4 h of STF treatment. RCC4 and RCC4VHL cell models were treated at 1.25 µM of STF. Mean and SEM were calculated from at least three independent experiment. Statistical analysis compared untreated cells with treated cells (*P < 0.05, **P < 0.01, ***P < 0.001) using Student’s T-test.

STF-62247 causes lysosomal impairment in VHL-inactivated cells and causes the accumulation of ubiquitinated protein aggregates

We have previously reported STF’s cellular localization to lysosomes as well as the presence of a divergent mechanism of STF’s cellular processing between VHL-proficient and VHL-defective models specifically at prolonged treatment times (8). To specifically investigate the extent of STF’s effects on lysosomal function, a metabolic pulse/chase experiment was used to measure the degradation of long-lived proteins in two VHL-inactivated cell models, RCC4 and RCC10, as well as the parental stable isogenic counterparts where wild-type VHL has been reintroduced, RCC4VHL and RCC10VHL (Figure 4A). In response to prolonged STF treatment, only VHL-proficient cell models (RCC4VHL and RCC10VHL) were able to increase their lysosomal protein degradation, concurrent with their ability to surmount a block in late stages of autophagy (Figure 4A). Contrastingly, no differences in long-lived protein degradation was observed in VHL-inactivated cells in response to STF. These results suggest a VHL-dependent mechanism in maintaining lysosomal integrity in response to STF-62247.

Figure 4.

STF-62247 causes lysosomal impairment in VHL-inactivated cells and causes the accumulation of ubiquitinated protein aggregates. (A) Long-lived proteins degradation by lysosomes in response to 48 h of STF in VHL-inactivated cells (RCC4 and RCC10) and VHL-functioning models (RCC4VHL and RCC10VHL). (B) Immunoblot of polyubiquitin levels in response to 4 h of STF treatment in RCC4, RCC4VHL and HeLa cell models. (C) Immunofluorescence of endogenous levels of p62 and ubiquitin and their co-localization in response to 48 h of STF treatment in RCC4 and RCC4VHL cell models. (D) Immunoblot analysis of total protein levels of p62 and serine 403 phosphorylated p62 in response to 8 h or 48 h of STF treatment in RCC4, RCC4VHL and HeLa cells. (E) Quantification of D. RCC4, RCC4VHL were treated at 1.25 µM and HeLa cells at 5 µM of STF. Mean and SEM were calculated from at least three independent experiment. Statistical analysis compared untreated cells with treated cells (*P < 0.05, **P < 0.01, ***P < 0.001) and the difference between cell models (#P < 0.05, ##P < 0.01, ###P < 0.001) using Student’s T-test.

Following autophagy and/or lysosome impairment, the constitutive clearance of aggregate-prone proteins is often altered (31). p62/SQSTM1 is an important regulatory molecule linking these ubiquitin-positive inclusions to autophagy and lysosome degradation (32). Thus, ubiquitin levels were assessed by immunoblot analysis in RCC4 and RCC4VHL models and in HeLa cells which contain a functional endogenous VHL gene (Figure 4B). Ubiquitin protein levels were greatly increased in response to STF treatment in VHL-inactivated RCC4 cells differentiating them from RCC4VHL and HeLa models for which ubiquitin levels decreased after treatment (Figure 4B). Indicative of the formation of protein aggregates, endogenous levels of p62 and ubiquitin were assessed by immunofluorescence (Figure 4C). STF caused the formation of cytoplasmic ubiquitinated protein aggregates in VHL-inactivated RCC4 cells as assessed by the co-localization of p62 and ubiquitin (Figure 4C). RCC4VHL cells showed no co-localization of the two markers, concomitant with their ability to increase their level of lysosomal degradation, resolve the endo-lysosomal swelling, and to remain insensitive to the compound. To link these protein aggregates to p62-dependent selective autophagy, levels of phosphorylated p62 on serine residue 403 were assessed (33) (Figure 4D and E). In accordance with immunofluorescence results, protein levels of phospho-p62 (S403) increased in a time-dependent manner in RCC4 cells in response to STF while being barely detectable in VHL-functioning cells (RCC4VHL and Hela) (Figure 4D and E). Altogether, the results show that STF-62247 selectively inhibits lysosome function in VHL-inactivated cells and leads to the accumulation of cytoplasmic inclusion bodies, as assessed by immunofluorescence and p-p62 (S403) immunoblot analyses.

Lysosomes from VHL-inactivated cells are unable to maintain their juxtanuclear localization when treated with STF-62247

To investigate lysosomes of VHL-inactivated cells, lysosome localization was assessed by immunofluorescence of endogenous Lamp-1 and F-actin (phalloidin) in STF-treated RCC4 and RCC4VHL models (Figure 5A). Lamp-1-positive structures in RCC4 gradually lost their perinuclear localization in a time-dependent manner (Figure 5A). In fact, lamp-1 staining became largely disorganized and occupied most of the cytoplasm at the prolonged time point (72 h). Contrastingly, lamp-1-positive staining in VHL-functioning cells gradually adopted a perinuclear localization with time (Figure 5A). These results confirm the decreased lysosomal function of VHL-inactivated cells as assessed by the measurement of long-lived protein degradation and their inability to surmount a block in autophagy. According to macroautophagic regulation, lysosomes will adopt a perinuclear localization in order to assure proper fusion events with autophagosomes and induce lysosomal activity (34). Oppositely, when macroautophagy is restrained, lysosomes are found to localize closer to the plasma membrane to assure mTORC1’s proximity to signaling receptors (35). With this in mind and to compare basal levels of autophagy according to VHL status, we visualized endogenous levels of Lamp-2 and assessed p62 and lipidated LC3B protein levels by immunoblot analyses in untreated RCC4 and RCC4VHL models (Figure 5B and C). Untreated RCC4 cells showed an extensive perinuclear localization of Lamp-2 compared to the RCC4VHL model. Concomitantly, RCC4 cells showed lower protein levels of p62 combined to higher levels of lipidated LC3 while the opposite was observed in RCC4VHL cells (Figure 5B and C). These results show higher basal levels of autophagy in untreated VHL-inactivated RCC4 cells and a striking difference in lysosomal positioning compared to the VHL-functioning RCC4VHL model.

Figure 5.

Lysosomes from VHL-inactivated cells are unable to maintain their juxtanuclear localization in response to STF-62247. (A) Lysosomal localization assessed by immunofluorescence of Lamp-1 and phalloidin (F-actin) in response to 24, 48 and 72 h of STF-62247 in RCC4 and RCC4VHL cell models. (B) Immunofluorescence of LAMP-2 and immunoblot analysis of autophagic effectors, LC3B and p62 in untreated RCC4 and RCC4VHL cells. (C) Quantification of B. (D) Immunoblots of autophagic effectors, LC3B and p62 in STF-treated cells (24 and 120 h) in RCC4, RCC4 VHL and RCC4VHL Cr.VHL (E–F) Quantifications of D. RCC4, RCC4VHL and RCC4VHL Cr.VHL cell lines were treated at 1.25 µM of STF. Mean and SEM were calculated from at least three independent experiment. Statistical analysis compared the difference between cell models (#P < 0.05, ##P < 0.01, ###P < 0.001) using Student’s T-test.

As shown here and previously, STF causes a block of the autophagic process and leads to an accumulation of P62 and LC3B-II (8) (Figures 2 and 3). We sought to investigate VHL’s role in the acclimatization of VHL-proficient cells to STF treatment by quantifying the protein levels of P62 and lipidated LC3B with 24 h and 120 h of treatment (Figure 5D–F). Quantification of p62 and LC3B-II at each time point was accomplished by normalizing protein levels with respective untreated controls and β-actin. In response to 120 h of STF treatment, RCC4VHL cells showed significantly reduced levels of p62 and lipidated LC3B compared to VHL-defective RCC4 cells, indicating VHL’s ability to reduce autophagy, surmount the block in flux, and maintain a level of lysosomal degradation in response to prolonged STF treatment (Figure 5E and F). Knockout of VHL in RCC4VHL rendered cells incapable of surmounting the autophagic block and levels of p62 and LC3B-II increased significantly, indicating a lysosomal disruption and a VHL dependence in lysosomal processes (Figure 5E and F). These results confirm our previous findings that gave prominence to a dependence of the autophagic process in VHL-defective models and further suggest a role for VHL in regulating levels of basal autophagy. We show that STF-62247 renders VHL-inactivated cells unable to surmount a block of autophagy because of their inability to retain their juxtanuclear positioning and assure proper fusion events.

Lysosomal-disrupting agents causes selective cell death in VHL-inactivated cells

The ability of VHL-functioning cells to surmount endo-lysosomal swelling and retain their lysosomes in a juxtanuclear localization in response to STF-62247 led us to hypothesize that VHL-loss leads to a lysosomal vulnerability that could be exploited. Consequently, different LDAs targeting the vacuolar H+-ATPase (V-ATPase) (bafilomycin A1 and concanamycin A) and that accumulate and/or raise lysosomal pH (chloroquine, ciprofloxacin) were tested by clonogenic survival assay in VHL-inactivated and VHL-functioning cells (Figure 6A). RCC4 cells showed significant sensitivity to 1nM and 2nM of bafilomycin A1 (BAF) with cell survival decreasing by 58% and 92%, respectively (Figure 6A). RCC4VHL cells however, showed no sensitivity to BAF treatment. Similarly, chloroquine treatment at 2.5 and 5 µM showed selective cytotoxicity for RCC4 cells and reduced their survival from 60% and 92%, respectively (Figure 6A). Moreover, concanamycin A (ConA) was selectively toxic at concentrations as low as 0.25 nM and reduced cell survival of RCC4 cells to ~15% while RCC4VHL cells maintained a cell survival of ~95%. While treatment at 0.50 nM showed cytotoxic selectivity for RCC4 cells and reduced their survival to 3%, this concentration also affected RCC4VHL cells and reduced their cell survival to ~40%, indicating the necessity to use low dosages of this compound. These results demonstrate that lysosomes from VHL-inactivated cells are noticeably more vulnerable than cells with a functional VHL gene.

Figure 6.

Lysosomal-disrupting agents causes selective cell death in VHL-inactivated cells. (A) Clonogenic survival assay of RCC4 and RCC4VHL in response to lysosomal disrupting agents (Bafilomycin A1 (BAF), Chloroquine (CQ), Concanamycin A (ConA) and Ciprofloxacin (Cipro)). (B) Clonogenic survival assay of RCC4, RCC4VHL and a STF-resistant RCC4 model in response to increasing concentrations of STF (0–2.5 µM). (C) Immunoblot analysis of RAB5, EEA1, RAB7 and LAMP-1 in RCC4 and RCC4 Resistant models. Total protein stain is shown. The added line denotes non-contiguous lanes on the membrane. (D) Clonogenic assays of RCC4, RCC4VHL and RCC4-Res models in response to Bafilomycin A1 (0–2 nM), (E) concanamycin A1 (0–0.25 nM), and (F) Chloroquine (0–5 µM). Mean and SEM were calculated from at least three independent experiment. Statistical analysis compared untreated cells with treated cells (*P < 0.05, **P < 0.01, ***P < 0.001) and the difference between cell models (#P < 0.05, ##P < 0.01, ###P < 0.001) at each concentration using Student’s T-test.

By gradually increasing STF concentration treatments over time, we have developed a new isogenic VHL-defective RCC4 cell model that is resistant to STF-62247 (RCC4-Res) as shown by clonogenic survival assay (Figure 6B). Next, immunoblots of key trafficking effectors were chosen (Rab5 and EEA1 marking early endosomes, Rab7 and Lamp1 marking late endosomes and lysosomes) to monitor differences in endogenous protein levels between RCC4 cells and the newly generated STF-resistant cells (Figure 6C). The RCC4-Res model showed decreased protein levels of Rab5, EAA1 and Rab7 while higher total levels of the lysosomal transmembrane protein Lamp-1 were observed, highlighting changes in the membranes of lysosomes (Figure 6C). Thus, to link STF’s selective properties to lysosomal defects, RCC4-Res cells were also treated with increasing concentrations of all three LDAs (Figure 6D–F). RCC4-Res model showed higher clonogenic survival capacity to each compound, comparable to the RCC4VHL model (Figure 6B). Similarly to VHL-functioning RCC4VHL cells, the RCC4-Res model demonstrated a significantly higher clonogenic survival capacity to BAF, ConA and chloroquine compared to the parental RCC4 model with 60%, 100% and 80% cell survival at the highest concentrations for each compound tested, respectively (Figure 6D–F). Altogether, these results succeed in highlighting a lysosomal vulnerability in VHL-defective cells, uncovering a role for VHL in maintaining endo-lysosomal integrity. Moreover, the resistance of the newly generated RCC4-resistant model to all LDAs tested confirms that STF-62247’s selective properties for VHL inactivation is due to lysosomal targeting.

Discussion

Our work succeeds in identifying a lysosomal vulnerability when VHL functions are lost. By investigating STF-62247, a small compound shown to block late stages of autophagy, we present the promising approach of targeting truncal-driven VHL inactivation through lysosomes. For the first time, we demonstrate the high alteration frequency of essential autophagic and lysosomal genes in ccRCC patient tissue samples derived from primary tumors (cbioportal). These results strongly suggest the presence of oncogenic transformations in lysosomes and autophagy in ccRCC which reveals potential and promising new therapeutic targets. In fact, the gene query comprised crucial endocytic, autophagic and lysosomal effectors, linking deregulation of lysosomal pathways to ccRCC development.

By investigating the selectivity of STF-62247, we showed that this compound does not affect the viability of other cancerous cell models tested but is in fact specific for ccRCC cells. As mentioned, the primary and ubiquitous event leading to the development of ccRCC is the biallelic inactivation of the tumor suppressor gene VHL. Hence, we demonstrated that STF-62247 is specifically cytotoxic to cells lacking VHL functions. Furthermore, knocking out VHL in a parental ccRCC model rendered cells noticeably more sensitive to the compound and could no longer resolve endo-lysosomal swelling, linking a role for VHL in maintaining lysosome integrity. Next, by monitoring lysosome positioning, we showed the ability of VHL-proficient cells to surmount STF-induced stress by increasing the motility of lamp-1-positive compartments towards the nucleus which correlates with the measured increase in lysosome degradation. In contrast, lysosomes from VHL-defective cells were severely altered, were unable to adopt a perinuclear localization in response to STF which explains their inability to increase long-lived protein degradation and the consequent accumulation of ubiquitin-positive aggregates. These results confirm our previous finding that fusion with autophagosomes is in fact inhibited when VHL functions are lost (8). Interestingly, the positioning of lysosomes and particularly their fusion with autophagosomes has been shown to be coordinated by a lysosome-assocaited multiportien complex named BLOC-1-related complex (BORC) (36). It is a possibility that STF-62247 blocks autophagy by inhibiting a BORC effector. As such, inhibiting fusion through the targeting of a BORC subunit is a strategy that would warrant further investigations in hopes of exploiting a VHL-dependent lysosome sensitivity.

Our results also demonstrate clear differences in levels of basal autophagy as well as lysosomal positioning between untreated VHL-defective and VHL-proficient cells. Specifically, we demonstrated that lysosomes from VHL-inactivated cells were localized near the perinuclear area which correlated with higher measured levels of basal autophagy, a characteristic that could render lysosome inhibition more effective. Exploiting this oncogenic feature by blocking late stages of autophagy through lysosomes could reveal a promising modality to overcome ccRCC tumor resistance to available targeted therapies. Furthermore, by testing several known LDAs, we demonstrated that cells lacking VHL functions were specifically sensitive to these compounds. Excitingly, the utilization of a VHL-defective STF-resistant cell model also showed similar resistance to all LDAs tested, linking STF-62247’s selectivity to lysosome-related pathways.

The results gathered in this study present strong preliminary evidence that targeting lysosomes in truncal VHL-driven ccRCC could be a promising approach. A challenge with this cancer remains the limited ccRCC cell models available for mechanistic studies and their inability to recapitulate the disease’s highly vascular features which can complicate translatability (37). Nonetheless, it has been shown that xenografts tumors derived from these subcutaneously injected cell models recapitulated some crucial documented clinical response and thus, could be used to study resistance mechanisms (38,39). Our study builds a strong platform that can lead to additional studies that could specifically link genetic clonal ccRCC evolution to lysosomal vulnerabilities. In vivo studies as well as cell death mechanisms should be the next steps in uncovering the possibility in targeting lysosomes as a promising approach in truncal-driven VHL-inactivated ccRCC.

Glossary

Abbreviations

ARG

autophagy-related gene

DMEM

Dulbecco’s modified Eagle’s medium

FBS

fetal bovine serum

HIF

hypoxia-inducible factors

LDA

lysosomal-disrupting agents

PAT

4-pyridyl-2-anilinothiazole

PCR

polymerase chain reaction

PE

plating efficiency

RCC

renal cell carcinoma

SEM

standard error of the mean

SF

surviving factor

STF

STF-62247

TCA

trichloroacetic acid

VHL

von Hippel-Lindau

Funding

This study was supported by Canadian Institutes of Health Research (CIHR) (#326436); New Brunswick Innovation Foundation (NBIF); Research Chair of the Canadian Cancer Society to S.T. The authors want to acknowledge the support of the Atlantic Cancer Research Institute.

Conflict of Interest Statement: None declared.