PBRM1 deficiency oncogenic addiction is associated with activated AKT–mTOR signalling and aerobic glycolysis in clear cell renal cell carcinoma cells

Abstract The PBRM1 (PB1) gene which encodes the specific subunit BAF180 of the PBAF SWI/SNF complex, is highly mutated (~ 40%) in clear cell renal cell carcinoma (ccRCC). However, its functions and impact on cell signalling are still not fully understood. Aerobic glycolysis, also known as the ‘Warburg Effect’, is a hallmark of cancer, whether PB1 is involved in this metabolic shift in clear cell renal cell carcinoma remains unclear. Here, with established stable knockdown PB1 cell lines, we performed functional assays to access the effects on 786‐O and SN12C cells. Based on the RNA‐seq data, we selected some genes encoding key glycolytic enzymes, including PFKP, ENO1, PKM and LDHA, and examined the expression levels. The AKT–mTOR signalling pathway activity and expression of HIF1α were also analysed. Our data demonstrate that PB1 deficiency promotes the proliferation, migration, Xenograft growth of 786‐O and SN12C cells. Notably, knockdown of PB1 activates AKT–mTOR signalling and increases the expression of key glycolytic enzymes at both mRNA and protein levels. Furthermore, we provide evidence that deficient PB1 and hypoxic conditions exert a synergistic effect on HIF 1α expression and lactate production. Thus, our study provides novel insights into the roles of tumour suppressor PB1 and suggests that the AKT–mTOR signalling pathway, as well as glycolysis, is a potential drug target for ccRCC patients with deficient PB1.

in cell cycle, cell fate, cell death, metabolism, DNA repair, transcriptional regulation and tumorigenesis, among many others. [4][5][6] In mammals, there are three distinct ATP-dependent SWI /SNF complexes and each complex has an ATPase (BRM or BRG1) and 10-14 other common or specific subunits. The classic BAF (BRG1 or BRM-associated factor) contains either ARID1A or ARID1B as a specific unit, while the PBAF (Polybromo-associated BAF) contains PB1, ARID2, BRD7 and PHF10 as specific subunits, and the newly identified GBAF (GLTSCR1 or GLTSCR1L-associated BAF) contains GLTSCR1 or GLTSCR1L and BRD9. 7,8 The genomic studies showed that genes encoding subunits of the SWI/SNF complexes are mutated in ~25% of human cancers, 9 which implies the critical roles of SWI/SNF complexes in carcinogenesis and cancer development.
PBRM1 (PB1) is the second most commonly mutated gene in ccRCC after Vhl, with a mutation rate that can be as high as 41%. 10 Mouse models showed that the loss of Vhl was not sufficient to cause RCC tumour, while the double deletion of Vhl and PB1 resulted in bilateral, multifocal, transplantable ccRCC tumours. 11,12 Moreover, the concurrent loss of PB1 and BAP1 drove ccRCC development from low grade to high grade. [13][14][15] These data imply the critical tumour suppressor role of PB1 or PBAF complex in carcinogenesis and development of ccRCC.

Metabolic reprogramming of tumour cells is a hallmark of can-
cer, and the metabolic shift to aerobic glycolysis (Warburg effect) is a feature for ccRCC. [16][17][18] ccRCC is characterized by over-deposition of glycogen and lipids in the cytoplasm, 19 and thus, kidney cancer has long been considered as a metabolic disease. [20][21][22] Furthermore, a correlation between an increase in metabolic activity and disease progression was observed in major renal cancer histological types. 18,23 There are lines of evidence showing that other mutated genes in ccRCC, such as VHL, MET, FLCN, TSC1, TSC2, FH and SDH, are involved in pathways that respond to metabolic stress or nutrient stimulation. 20 Whether PB1 plays a role in metabolic reprogramming such as aerobic glycolysis and how it functions remains unknown. 22 Here, we carried out functional studies with stably knockdown cell lines of 786-O and SN12C, in an attempt to understand the role of PB1 in ccRCC cells and its molecular mechanisms in metabolic reprogramming. Our data provide new insights into the tumoursuppressive role of PB1 and also links with glycolysis, mTOR and HIF1α in ccRCC.

| Online databases
Gene Expression Profiling Interactive Analysis (GEPIA2) database and Lianchuan online platform were used for transcriptomic analysis. The server cBioPortal was used for co-expression analysis between PB1 and several well-known genes encoding classic glycolytic enzymes. Kaplan-Meier Plotter was used to analyse the overall survival rate of ccRCC patients.

| RNA extraction and quantitative realtime PCR
Total RNA was extracted with Trizol (Solarbio), cDNA was synthesized by reverse transcription according to the manufacturer's instructions (Takara Holdings Inc.), quantitative real-time PCR (qRT-PCR) was performed by using SensiFast SYBR Kit (Bioline) on a CFX 96 instrument (Bio-Rad) with the following parameter: 95°C for 5 min, 40 cycles of 95°C for 5 s and 58°C for 50 s. Finally, dissociation curves were run (65°C for 5 s and 95°C for 5 s) in order to identify specific products.

| Western blot analysis
The cells were collected and lysed with RIPA buffer (Solarbio), in the presence of protease inhibitor (Roche #11836170001) and phenylmethylsulfonyl fluoride (PMSF) (Solarbio,#P0100). The protein concentration was quantified with the BCA Protein Assay Kit (Solarbio) following the manufacturer's instructions. After five volumes of proteins were mixed with one volume of 5× loading buffer and denatured at 100°C for 5 min, proteins were resolved on 10% SDS-Page gels. PVDF membrane (Epizyme Biotech, China) was activated in 100% methanol and then blocked with 5% nonfat milk in tris buffered saline with Tween 20 (TBST) for 2 h. The PVDF membrane was incubated with diluted primary antibody at 4°C overnight, washed with TBST three times and incubated with the diluted second antibody for 2 h at room temperature. The membranes were then washed with TBST, treated with ECL reagents and washed again with TBST three times. Protein bands were imaged by a Gel imaging system (Gel Doc XR, Bio-Rad) and analysed with Image J software.
The following primary antibodies were used: BAF180/PB1

| siRNAs transfection
Small interfering RNAs (siControl, siPB1-1, and siPB1-2) were designed and synthesized by Genepharma (Shanghai). Cells were digested with trypsin-EDTA and 6 × 10 5 cells were seeded in wells of a 6-well plate. The following day cells were transfected using the RNA-Mate (Genepharma) following the manufacturer's instructions.
The cells were harvested 72 h later and used for Western blot analysis or RNA extraction.

| CCK-8 cell proliferation experiment
After the cells were digested with 0.05% trypsin-EDTA, they were suspended in 1640 medium supplemented with 10% FBS and 1× penicillin/streptomycin solution. 2 × 10 3 cells per well for 786-O and 3 × 10 3 cells per well for SN12C were reseeded in 96-well plates (BKMAN), respectively. The cells were kept at 37°C in a humidified 5% CO 2 incubator. At indicated time points (Days 0, 1, 2, 3 and 4), 10 μl of CCK8 reagent (Beyotime) was added to each well of a 96well plate, samples were incubated at 37°C for 2 h followed by measuring the absorbance at 450 nm using a multifunctional microplate reader (SpectraMax M2e, Molecular Devices).

| Cell colony formation assay
Cell colony formation assay was performed basically as Franken et al described. 24 In brief, after the cells were digested and resuspended as mentioned above, 2 × 10 3 cells per well for 786-O and 4 × 10 3 cells per well for SN12C were reseeded in 6-well plates. The cells were grown for 12 days, and the medium was changed every 3 days. Then, the medium was removed, the clones were washed with phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde for 20 minutes. After the clones were stained with 1% crystal violet for 40 min and washed with PBS, the images were taken for statistical analysis.

| Wound healing assay
The cells were cultured in a 6-well plate and the wound healing assay was carried out when the cell density reached 90%. The cells were scratched evenly with a 200 μl pipette tip to simulate a wound. The floating cells were washed off and 1640 medium supplemented with 2% FBS was added to each well. The images were acquired using an inverted Leica DMIRB microscope at 0, 24 and 48 h time points.

| Transwell migration experiment
The transwell migration experiment was carried out in 24-well transwell chambers (Corning). After the cells were digested with 0.05% After the cells were stained with 1% crystal violet solution for 1 h at room temperature and gently rinsed with distilled water, five fields were randomly selected and counted under an inverted microscope.
The experiment was performed in triplicate.

| Subcutaneous tumour formation in nude mice
The tumour formation assay was conducted in compliance with animal protocols approved by the Laboratory Animal Ethics Committee at Zunyi Medical University. In total, 15 male BALB/c-Nu mice (~6 weeks old) were purchased from the Animal Center of Zunyi Medical University. Each mouse was subcutaneously inoculated with 1 × 10 7 cells (either shControl or shPB1-1 or shPB1-2) suspended in 100 μl PBS using a 1 ml syringe. The body weight and tumour size were monitored every 5 days. Mice were sacrificed after 4 weeks of inoculation. The tumours were weighed and analysed.

| Measurement of lactate concentration
The concentration of lactic acid was measured using a LA assay kit (Solarbio, #BC2230) according to the manufacturers' instructions. In brief, after the cells were scraped off and resuspended, 1 × 10 6 cells were homogenized at 4°C with an ultrasonic sonicator (CIENTZ).
Following centrifugation at 10,000 g at 4°C for 10 min, the individual supernatant and reaction reagent was mixed and kept at 37°C for 15 min before transferring to a 96-well plate. Finally, the absorbance at 450 nm was measured using a multifunctional microplate reader (SpectraMax M2e, Molecular Devices).

| Knockdown of PB1 promotes the migration, invasion and xenograft growth of ccRCC cells
ShRNA lentiviruses (shControl, shPB1-1 and shPB1-2) were used to infect 786-O and SN12C cell lines. The knockdown efficiency was examined by Western blotting which showed a remarkable drop at the protein level for both shRNAs (shPB1-1 & shPB1-2) ( Figure 1A). Then, the two ccRCC cell lines were used for further studies.
We performed the CCK8 assay to examine the cells' proliferation capability upon stably knockdown of PB1. There was no significant difference between shControl group and shPB1 group until Day 4 for both cell lines ( Figure 1B).

| The deficient PB1 activates AKT/mTOR signalling pathway
The AKT/ mTOR pathway is one of the most activated signalling pathways in ccRCC. 22,25 Our analysis showed that the key AKT- mutations. This is also true for other genes of the AKT-mTOR signalling pathway ( Figure 3B,C). Since these mutations seem to be mutually exclusive, we speculated that deficient PB1 may activate the AKT-mTOR signalling pathway. To test this hypothesis, we examined the protein level of Akt, p-AKT, mTOR and p-mTOR. The Western blot results demonstrated that deficient PB1 did increase the levels of Akt and mTOR in both 786-O and SN12C cells, as well as their phosphorylated forms ( Figure 3C).

| Deficient PB1 increase the protein levels of glycolytic enzymes
Previously, we generated RNA-seq data in 786-O cells with siRNA targeting PB1, in order to understand what processes or pathways were regulated by PB1 at the transcriptional level. Our data indicate a number of KEGG terms were enriched, including cell growth and death, cell motility, replication and cell cycle control ( Figure 4A).
Interestingly, we noticed that these dysregulated genes were also in-

| Deficient PB1 activates the glycolysis signalling independent of HIF1α
Hypoxia-inducible factor HIF1α is a key player in ccRCC which can be accumulated due to Vhl loss. To investigate if deficient PB1 also increases the expression of HIF1α, we examined the protein level of

| PB1 level is negatively correlated with the expression of key glycolytic enzymes in clinic samples
Next, we examined the expression of key enzymes catalysing glycolysis such as PFKP, ENO1, PKM and LDHA in ccRCC tumours and the matched controls. The analysis showed that these genes were all highly expressed in tumour samples compared with adjacent normal controls ( Figure 7A). We further analysed the relationship of PB1    Figure 7B). The overall survival analysis demonstrates that deficient or low-level PB1 was significantly correlated with worse outcomes, while high-level PB1 could prolong the patient's life expectancy ( Figure 7C).
Based on these data, we proposed that given HIF1α expression is induced under hypoxic conditions, deficient PB1 not only activates AKT-mTOR and the glycolysis signalling pathways but also exerts a synergetic effect on HI1a expression, which collectively promote ccRCC tumour growth ( Figure 7D).

| DISCUSS ION
In order to understand the functional role of PB1 and the mechanisms, we firstly performed the CCK8 assay, colony formation assay and then, migration assay using the stable cell lines with reduced PB1 expression. These functional assays demonstrate that PB1 does function as a tumour suppressor and promotes the proliferation and migration of ccRCC cells. Moreover, the nude mice assay showed that deficient PB1 promotes xenograft tumour growth in vivo.
Our immunofluorescence assay clearly showed that PB1 is localized in the nuclei, which is consistent to its role functioning in chromatin remodelling process and gene transcriptional control Both Vhl and PB1 are tumour suppressor genes highly mutated in ccRCC. Based on mouse model studies, loss of Vhl alone cannot cause kidney tumour, and only when both genes are inactivated the tumour may form. 11,12 It is well known that the HIF1α level can be elevated by loss of Vhl, our finding that HIF1α level can also be elevated by deficient PB1 provides a new explanation for why angiogenesis is featured in ccRCC samples with wild-type Vhl. It is also consistent to the observation that expression of full-length HIF1α and PB1protein seems to be mutually exclusive in most ccRCC cell lines and in some primary tumours. 26

ACK N OWLED G EM ENT
We thank Professor Bin Tean Teh for his valuable help during this study.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

DATA A N A LYS I S
We used Photoshop 6.0 to crop the images and used Graphpad Prism 6 and SPSS V 16.0 to plot the data. For statistical analysis, the results were all expressed as x ± SD. The two-tailed Student's t-test was used to examine the significance level.

DATA AVA I L A B I L I T Y S TAT E M E N T
The expression profiles were deposited to NCBI under BioProject accession No.PRJNA771094.