Cancer Cell-Derived PDGFB Stimulates mTORC1 Activation in Renal Carcinoma

Clear cell renal cell carcinoma (ccRCC) is a hypervascular tumor that is characterized by bi-allelic inactivation of the VHL tumor suppressor gene and mTOR signalling pathway hyperactivation. The pro-angiogenic factor PDGFB, a transcriptional target of super enhancer-driven KLF6, can activate the mTORC1 signalling pathway in ccRCC. However, the detailed mechanisms of PDGFB-mediated mTORC1 activation in ccRCC have remained elusive. Here, we investigated whether ccRCC cells are able to secrete PDGFB into the extracellular milieu and stimulate mTORC1 signalling activity. We found that ccRCC cells secreted PDGFB extracellularly, and by utilizing KLF6- and PDGFB-engineered ccRCC cells, we showed that the level of PDGFB secretion was positively correlated with the expression of intracellular KLF6 and PDGFB. Moreover, the reintroduction of either KLF6 or PDGFB was able to sustain mTORC1 signalling activity in KLF6-targeted ccRCC cells. We further demonstrated that conditioned media of PDGFB-overexpressing ccRCC cells was able to re-activate mTORC1 activity in KLF6-targeted cells. In conclusion, cancer cell-derived PDGFB can mediate mTORC1 signalling pathway activation in ccRCC, further consolidating the link between the KLF6-PDGFB axis and the mTORC1 signalling pathway activity in ccRCC.


Introduction
Clear cell renal cell carcinoma (ccRCC) is the most prevalent kidney cancer subtype, accounting for~75% of all reported cases worldwide [1]. Bi-allelic inactivation of the VHL tumor suppressor gene and consequent accumulation of pro-oncogenic hypoxia-inducible factor alpha (HIFα) is the hallmark gatekeeper in ccRCC pathogenesis, contributing to 90% of sporadic ccRCC cases [2,3]. The accumulation of cytoplasmic lipids and glycogens gives rise to the ccRCC distinctive glass-like appearance, which has been attributed to ccRCC metabolic reprogramming [4]. Moreover, ccRCC is highly vascularized due to the upregulation of pro-angiogenic factors [5]. The hyperactivation of mTOR signalling pathway is also frequently observed in ccRCC patients [6]. To this end, inhibitors targeting angiogenesis and the mTOR signalling pathway have been clinically approved as either first-line or second-line treatment against ccRCC [7]. In addition, HIF2α and immune checkpoint inhibitors have also been tested and showed clinical efficacy in some ccRCC patients [8,9]. Nonetheless, these ccRCC monotherapy approaches remain unsatisfactory in terms of patients' response rate, as well as the rapid development of acquired resistance towards the administered therapy [7]. These could be due to the widespread genetic heterogeneities, inadequate target inhibition, and biological adaptation to an alternative

Human ccRCC Cells Secreted PDGFB Extracellularly
To assess whether ccRCC cells secrete PDGFB extracellularly, we first performed ELISA to quantify the level of PDGFB in the serum-starved media that were collected from 786-M1A and OS-LM1B cells at different time points. These cells were cultured in the serum-free media to replicate the condition that we had previously used to study the link between PDGFB and the mTORC1 signalling pathway in ccRCC [32]. We found that the concentration of PDGFB in the collected culture media increased in a time-dependent manner ( Figure 1a). Additionally, we also compared the level of secreted PDGFB between ccRCC cells that were cultured in either complete or serum-free media for 16 h. The level of extracellular PDGFB was found to be higher in the ccRCC cells' complete culture media as compared to serum-free culture media (Figure 1b). Nonetheless, to be consistent with our previous study [32], all of the subsequent experiments were performed in serum-free media. Collectively, these data showed that human ccRCC cells secreted PDGFB into the extracellular milieu.

Human ccRCC Cells Secreted PDGFB Extracellularly
To assess whether ccRCC cells secrete PDGFB extracellularly, we first performe ELISA to quantify the level of PDGFB in the serum-starved media that were collected from 786-M1A and OS-LM1B cells at different time points. These cells were cultured in the se rum-free media to replicate the condition that we had previously used to study the lin between PDGFB and the mTORC1 signalling pathway in ccRCC [32]. We found that th concentration of PDGFB in the collected culture media increased in a time-dependen manner ( Figure 1a). Additionally, we also compared the level of secreted PDGFB betwee ccRCC cells that were cultured in either complete or serum-free media for 16 h. The leve of extracellular PDGFB was found to be higher in the ccRCC cells' complete culture medi as compared to serum-free culture media (Figure 1b). Nonetheless, to be consistent wit our previous study [32], all of the subsequent experiments were performed in serum-fre media. Collectively, these data showed that human ccRCC cells secreted PDGFB into th extracellular milieu.

PDGFB Secretion Level Positively Correlated with Intracellular PDGFB Expression
We next assessed whether the modulation of intracellular PDGFB would affect th level of PDGFB secreted by ccRCC cells. To achieve this, we either overexpressed exoge nous PDGFB or targeted PDGFB using CRISPR/Cas9 in the 786-M1A cells. We confirme PDGFB overexpression in the 786-M1A cells that stably expressed the exogenous PDGF (Figure 2a,b). On the other hand, the PDGFB targeting was performed in the pooled 786 M1A cells using two independent sgRNAs, referred to as sgPDGFB_1 and sgPDGFB_ We observed a reduction in the expression of intracellular PDGFB in the 786-M1A cel that were transduced with the sgPDGFB constructs as compared to the control cells (tran duced with non-targeting sgRNA construct) (Figure 2b). We then collected the respectiv serum-starved media from these PDGFB-engineered and control 786-M1A cells and sub jected the media to PDGFB ELISA. Relative to the control cells, we found that the PDGFB targeted 786-M1A cells had the lowest level of secreted PDGFB, whereas the PDGFB ove expressing cells secreted more PDGFB into the extracellular environment ( Figure 2c).
To consolidate these observations, we also modulated the expression of intracellula PDGFB in the OS-LM1B cells similar to those established for 786-M1A cells. We confirme that they were increased and decreased in intracellular PDGFB expression in the PDGFB overexpressing and PDGFB-targeted OS-LM1B cells, respectively (Figure 2d). We foun

PDGFB Secretion Level Positively Correlated with Intracellular PDGFB Expression
We next assessed whether the modulation of intracellular PDGFB would affect the level of PDGFB secreted by ccRCC cells. To achieve this, we either overexpressed exogenous PDGFB or targeted PDGFB using CRISPR/Cas9 in the 786-M1A cells. We confirmed PDGFB overexpression in the 786-M1A cells that stably expressed the exogenous PDGFB (Figure 2a,b). On the other hand, the PDGFB targeting was performed in the pooled 786-M1A cells using two independent sgRNAs, referred to as sgPDGFB_1 and sgPDGFB_2. We observed a reduction in the expression of intracellular PDGFB in the 786-M1A cells that were transduced with the sgPDGFB constructs as compared to the control cells (transduced with non-targeting sgRNA construct) (Figure 2b). We then collected the respective serumstarved media from these PDGFB-engineered and control 786-M1A cells and subjected the media to PDGFB ELISA. Relative to the control cells, we found that the PDGFB-targeted 786-M1A cells had the lowest level of secreted PDGFB, whereas the PDGFB overexpressing cells secreted more PDGFB into the extracellular environment ( Figure 2c). that the OS-LM1B cells that had the PDGFB targeted using CRISPR/Cas9 secreted less PDGFB extracellularly compared to the OS-LM1B control cells (Figure 2e). On the other hand, a high level of extracellular PDGFB was detected in the PDGFB-overexpressing OS-LM1B cells' culture media ( Figure 2f). Overall, it can be inferred from these findings that ccRCC cells indeed secreted PDGFB into the extracellular milieu, and the secretion level positively correlated with the intracellular PDGFB expression.

KLF6 Regulated PDGFB Expression and Secretion in ccRCC Cells
We have previously reported that PDGFB expression in ccRCC was directly transactivated by transcription factor KLF6 [32]. Therefore, we were prompted to measure the level of extracellular PDGFB in the previously generated 786-M1A iKLF6_2 cells, which have KLF6 stably repressed using CRISPRi approach [32]. We confirmed that there were strong KLF6 repression and consequent PDGFB downregulation in these cells (Figure 3a). We found that the 786-M1A iKLF6_2 cells secreted a lower amount of PDGFB compared to the 786-M1A control cells (Figure 3b), which correlated with the reduced expression of intracellular PDGFB in these KLF6-repressed cells. We next measured the level of secreted To consolidate these observations, we also modulated the expression of intracellular PDGFB in the OS-LM1B cells similar to those established for 786-M1A cells. We confirmed that they were increased and decreased in intracellular PDGFB expression in the PDGFBoverexpressing and PDGFB-targeted OS-LM1B cells, respectively ( Figure 2d). We found that the OS-LM1B cells that had the PDGFB targeted using CRISPR/Cas9 secreted less PDGFB extracellularly compared to the OS-LM1B control cells (Figure 2e). On the other hand, a high level of extracellular PDGFB was detected in the PDGFB-overexpressing OS-LM1B cells' culture media ( Figure 2f). Overall, it can be inferred from these findings that ccRCC cells indeed secreted PDGFB into the extracellular milieu, and the secretion level positively correlated with the intracellular PDGFB expression.

KLF6 Regulated PDGFB Expression and Secretion in ccRCC Cells
We have previously reported that PDGFB expression in ccRCC was directly transactivated by transcription factor KLF6 [32]. Therefore, we were prompted to measure the level of extracellular PDGFB in the previously generated 786-M1A iKLF6_2 cells, which have KLF6 stably repressed using CRISPRi approach [32]. We confirmed that there were strong KLF6 repression and consequent PDGFB downregulation in these cells (Figure 3a). We found that the 786-M1A iKLF6_2 cells secreted a lower amount of PDGFB compared to the 786-M1A control cells (Figure 3b), which correlated with the reduced expression of intracellular PDGFB in these KLF6-repressed cells. We next measured the level of secreted PDGFB in the 786-M1A iKLF6_2 cells that expressed either exogenous KLF6 CDS or PDGFB CDS. KLF6 overexpression and subsequent PDGFB upregulation were confirmed in the 786-M1A iKLF6_2 cells that stably expressed the exogenous KLF6 CDS (Supplementary Figure S1a  We further validated the findings in Figure 3 by measuring the level of extracellular PDGFB in an additional set of KLF6-modulated ccRCC cells, which were the UOK101 cells. CRISPRi-mediated KLF6 repression in UOK101 cells reduced PDGFB expression, whereby the reintroduction of exogenous KLF6 CDS in these cells resulted in PDGFB up- We further validated the findings in Figure 3 by measuring the level of extracellular PDGFB in an additional set of KLF6-modulated ccRCC cells, which were the UOK101 cells. CRISPRi-mediated KLF6 repression in UOK101 cells reduced PDGFB expression, whereby the reintroduction of exogenous KLF6 CDS in these cells resulted in PDGFB upregulation (Figure 4a,b and Supplementary Figure S2a). We found that the PDGFB secretion level in these UOK101 cells also positively correlated with the expression of intracellular PDGFB that was regulated by the transcription factor KLF6 ( Figure 4c). As expected, the KLF6repressed UOK101 cells that stably expressed exogenous PDGFB CDS secreted a high level of PDGFB extracellularly (Figure 4d and Supplementary Figure S2b). Collectively, these data clearly demonstrated that ccRCC cells secreted PDGFB into the extracellular environment and that the secretion level was positively correlated with the expression of intracellular PDGFB and its upstream regulator KLF6. Importantly, we have corroborated the functional link between one of the strongest super enhancers and the regulation of PDGFB expression and secretion in ccRCC. intracellular PDGFB that was regulated by the transcription factor KLF6 ( Figure 4c). As expected, the KLF6-repressed UOK101 cells that stably expressed exogenous PDGFB CDS secreted a high level of PDGFB extracellularly (Figure 4d and Supplementary Figure S2b).
Collectively, these data clearly demonstrated that ccRCC cells secreted PDGFB into the extracellular environment and that the secretion level was positively correlated with the expression of intracellular PDGFB and its upstream regulator KLF6. Importantly, we have corroborated the functional link between one of the strongest super enhancers and the regulation of PDGFB expression and secretion in ccRCC.

Secreted PDGFB Stimulates mTORC1 Signalling Pathway Activation
We have previously demonstrated the role of the KLF6-PDGFB axis in modulating mTORC1 activity in ccRCC. Repressing either KLF6 or PDGFB impaired mTORC1 activity, whereas supplementing the KLF6-repressed cells with recombinant human PDGFB

Secreted PDGFB Stimulates mTORC1 Signalling Pathway Activation
We have previously demonstrated the role of the KLF6-PDGFB axis in modulating mTORC1 activity in ccRCC. Repressing either KLF6 or PDGFB impaired mTORC1 activity, whereas supplementing the KLF6-repressed cells with recombinant human PDGFB reactivated the mTORC1 signalling pathway in ccRCC [32]. However, whether cancer cell-derived PDGFB is also able to activate mTORC1 activity in KLF6-depleted cells has remained unclear. Since the reintroduction of exogenous KLF6 and PDGFB increased the PDGFB secretion level of the KLF6-repressed cells (Figure 3c,d and Figure 4c,d), we postulated that these cells would have sustained mTORC1 activity. To assess this, we blotted for phosphorylated S6 protein, a read-out for mTORC1 activity, in the KLF6-and PDGFB-expressing 786-M1A iKLF6_2 cells that underwent serum starvation overnight. In line with our hypothesis, we found that the KLF6-and PDGFB-expressing cells had more mTORC1 signalling activity compared to the 786-M1A iKLF6_2 control cells (Figure 5a). mTORC1 signalling activity compared to the 786-M1A iKLF6_2 control cells (Figure 5a). We next tested whether the secreted PDGFB was able to stimulate mTORC1 activity in a paracrine manner. The mTORC1-impaired 786-M1A iKLF6_2 cells were serumstarved overnight and cultured for an hour on the following day with (i) serum-free media, (ii) 786-M1A parental cells', or (iii) PDGFB-overexpressing 786-M1A cells' conditioned media. Prior to collecting the conditioned media, these parental and PDGFB-overexpressing 786-M1A cells were cultured in serum-free media overnight. Supplementing the KLF6-repressed cells with the conditioned media from PDGFB-overexpressing cells resulted in the reactivation of the mTORC1 activity (Figure 5b). On the other hand, the conditioned media from the 786-M1A parental cells was also able to stimulate the mTORC1 activity, although the magnitude of induction was much lower than the conditioned media of PDGFB-overexpressing cells (Figure 5b). To complement these results, we replicated these experiments using the UOK101 iKLF6_2 cells. Consistent with our previous and present findings, we observed a reduced mTORC1 activity in the KLF6-repressed UOK101 cells. Furthermore, the reintroduction of exogenous PDGFB CDS or culturing these UOK101 iKLF6_2 cells with conditioned media from PDGFB-overexpressing cells were able to sustain and reactivate the mTORC1 signalling pathway, respectively (Supplementary Figure S3).  We next tested whether the secreted PDGFB was able to stimulate mTORC1 activity in a paracrine manner. The mTORC1-impaired 786-M1A iKLF6_2 cells were serum-starved overnight and cultured for an hour on the following day with (i) serum-free media, (ii) 786-M1A parental cells', or (iii) PDGFB-overexpressing 786-M1A cells' conditioned media. Prior to collecting the conditioned media, these parental and PDGFB-overexpressing 786-M1A cells were cultured in serum-free media overnight. Supplementing the KLF6-repressed cells with the conditioned media from PDGFB-overexpressing cells resulted in the reactivation of the mTORC1 activity (Figure 5b). On the other hand, the conditioned media from the 786-M1A parental cells was also able to stimulate the mTORC1 activity, although the magnitude of induction was much lower than the conditioned media of PDGFB-overexpressing cells (Figure 5b). To complement these results, we replicated these experiments using the UOK101 iKLF6_2 cells. Consistent with our previous and present findings, we ob-served a reduced mTORC1 activity in the KLF6-repressed UOK101 cells. Furthermore, the reintroduction of exogenous PDGFB CDS or culturing these UOK101 iKLF6_2 cells with conditioned media from PDGFB-overexpressing cells were able to sustain and reactivate the mTORC1 signalling pathway, respectively (Supplementary Figure S3).

Discussion
By utilizing KLF6-and PDGFB-engineered ccRCC cells, we confirmed that ccRCC cells secrete PDGFB into the extracellular space. We found that PDGFB targeting reduced the level of extracellular PDGFB, whereas the overexpression of PDGFB increased the PDGFB secretion level. This was in line with the fact that the level of most, if not all, secreted proteins would depend on the expression of their intracellular counterparts. Altering the expression of these intracellular proteins, either by genetic or chemical means, would affect the secretion level. Intriguingly, we found that the PDGFB secretion level was also correlated with the expression of its upstream transcriptional activator KLF6. We have previously reported that KLF6 was highly expressed in ccRCC, and its expression was driven by one of the strongest and robust super enhancers in ccRCC [32]. Based on these findings, we postulated that the high expression of KLF6 enhances PDGFB transcriptional activation that could in turn play an important role in supporting ccRCC pathogenesis. Whilst targeting KLF6 impaired ccRCC cell growth and lung colonization capabilities [32], it is worthwhile to test the effect of direct PDGFB perturbation on ccRCC cell phenotypes in a future study. Our present findings corroborated the functional link between the KLF6 super enhancer locus and the transcriptional regulation of the angiogenesis-promoting PDGFB in ccRCC. This is in line with other reported roles of super enhancers as drivers of the expression of genes that regulate and maintain cancer phenotypes [33,34].
One of ccRCC hallmark features observed in patients is the frequent hyperactivation of the mTOR signalling pathway [6,35]. Following the report that PDGFB is one of the mTOR signalling pathway agonists [36], we previously showed that supplementing the KLF6-targeted ccRCC cells with recombinant human PDGFB re-activated the impaired mTORC1 activity in these cells [32]. The secretion of PDGFB by ccRCC cells and the ability of conditioned ccRCC media to activate mTORC1 signalling in KLF6 depleted cells, as reported herein, consolidates our previous findings. It is noteworthy to highlight that our results unraveled a molecular connection between pro-angiogenic factors and mTOR pathway activation, which are the two approved therapeutic targets in ccRCC [7]. Concordantly, a combinatorial treatment of lenvatinib (a multi RTKs inhibitor including PDGFR) and mTOR inhibitor everolimus on metastatic RCC patients who have progressed after one previous VEGF-targeted therapy increased the progression-free survival over either drug alone [37]. In a separate trial on untreated advanced RCC patients, a combination of lenvatinib with either pembrolizumab (PD-L1 monoclonal antibody) or everolimus prolonged the progression-free survival over sunitinib treatment alone [38]. The secretion of PDGFB by ccRCC cells reported herein could prompt the idea of the potential development and use of monoclonal antibodies against secreted PDGFB in treating ccRCC. There have been reports examining the correlation between intracellular PDGFB/PDGFRβ expression and RCC stages or prognosis [39,40]. However, to our knowledge, there is no study directly evaluating the serum PDGFB level in RCC in comparison to a healthy control and across different tumor stages. This knowledge is crucial to determine the possibility of not only targeting the circulating PDGFB, but also for the development of diagnostic and/or prognostic biomarkers based on the circulating PDGFB level.
In conclusion, secretion of PDGFB by ccRCC cells is able to induce mTORC1 activity in the neighboring ccRCC cells in a paracrine manner. In this regard, the secreted PDGFB could also possibly act on the same cells in an autocrine manner ( Figure 6). Our findings further establish the role of one of the strongest super enhancers in ccRCC as a modulator of mTORC1 activation via the KLF6-PDGFB transcriptional axis. Thus, the common hyperactivation of the mTORC1 signalling pathway in ccRCC could, at least in part, be

Cell Lines and Reagents
The human ccRCC cell lines used in this study were 786-M1A, OS-LM1 A498, and 769-P. The 786-M1A and OS-LM1 cell lines were obtained from J. (Memorial Sloan Kettering Cancer Center, New York, NY, USA). The 786-M1 LM1 are the metastatic derivative of 786-O and OS-RC2 cells, respectively, w established and described previously [41]. The UOK101 cell line was obtained f ton Linehan (National Cancer Institute, Bethesda, MD, USA). The A498 and 769were obtained from the American Type Culture Collection (ATCC, Manassas, These cell lines were confirmed to be mycoplasma negative using the e-Myco plasma PCR Detection Kit (Intron, Kirkland, WA, USA). All of these human c lines were maintained in RPMI-1640 medium (Nacalai Tesque, Kyoto, Japan mented with 10% FBS (Tico Europe, Amstelveen, Netherlands) and 1% penicill mycin solution (Nacalai Tesque). The HEK293T cells used for lentivirus produ cultured in DMEM (Nacalai Tesque), supplemented with 10% FBS and 1% streptomycin solution. Puromycin and Hygromycin B solution were purchased vogen and Nacalai Tesque, respectively.

Cell Lines and Reagents
The human ccRCC cell lines used in this study were 786-M1A, OS-LM1, UOK101, A498, and 769-P. The 786-M1A and OS-LM1 cell lines were obtained from J. Massagué (Memorial Sloan Kettering Cancer Center, New York, NY, USA). The 786-M1A and OS-LM1 are the metastatic derivative of 786-O and OS-RC2 cells, respectively, which were established and described previously [41]. The UOK101 cell line was obtained from Marston Linehan (National Cancer Institute, Bethesda, MD, USA). The A498 and 769-P cell lines were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). These cell lines were confirmed to be mycoplasma negative using the e-MycoTM Mycoplasma PCR Detection Kit (Intron, Kirkland, WA, USA). All of these human ccRCC cell lines were maintained in RPMI-1640 medium (Nacalai Tesque, Kyoto, Japan), supplemented with 10% FBS (Tico Europe, Amstelveen, The Netherlands) and 1% penicillin-streptomycin solution (Nacalai Tesque). The HEK293T cells used for lentivirus production were cultured in DMEM (Nacalai Tesque), supplemented with 10% FBS and 1% penicillin-streptomycin solution. Puromycin and Hygromycin B solution were purchased from Invivogen and Nacalai Tesque, respectively.

sgPDGFB Cloning and Bacteria Transformation
The complementary sgPDGFB top and bottom strands were purchased separately. These strands were designed to harbor BbsI restriction overhangs at their respective 5 and 3 ends for ligation into BbsI-digested pKLV-U6-gRNA(BbsI)-PGKHygro2AeGFP plasmid. The top and bottom strands were annealed and subjected to 5 end phosphorylation using T4 polynucleotide kinase (New England Biolabs, Ipswich, MA, USA). Ligation was performed using T4 Ligase (New England Biolabs) at 16 • C overnight and the ligated plasmid was transformed into the chemically competent DH5α E. coli strain (New England Biolabs). The presence of sgPDGFB construct within the expression plasmid was verified via Sanger sequencing.

Plasmid and Total RNA Extraction
Plasmid was extracted from the transformed bacteria culture using Monarch Plasmid Miniprep Kit (New England Biolabs) according to the manufacturer's protocols. Total RNA was extracted from the cells using Trizol TM reagent (Thermo Fisher, Waltham, MA, USA) by following the manufacturer's protocols. The yield and purity of the extracted plasmids and total RNA were determined using the NanoDrop TM 2000c Spectrophotometer (Thermo Fisher).

Lentivirus Transduction
The HEK293T cells were co-transfected with the mixture of lentivirus packaging plasmids and plasmid of interest using Attractene transfection reagent (Qiagen, Hilden, Germany) according to the manufacturer's recommendations. Media containing the lentivirus was collected 72 h post-transfection and filtered through MinisartNML 0.45 µM syringe filter (Sartorius, Göttingen, Germany). For lentiviral transduction, the filtered media containing the lentivirus was added onto the cells, which were at 60-70% confluency, in the presence of 8 µg/mL Polybrene (Millipore, Burlington, MA, USA).

cDNA Synthesis and qRT-PCR
Total RNA was converted into cDNA using the LunaScript RT SuperMix Kit (New England Biolabs) according to the manufacturer's recommendations. The qRT-PCR was performed using the Luna Universal Probe qPCR Master Mix (New England Biolabs) and 20× pre-designed TaqMan gene expression probes (Thermo Fisher) on the CFX96 Touch Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) according to the manufacturer's recommendations. The following TaqMan probes were used: KLF6 (Hs00810569_m1), PDGFB (Hs00966522_m1), and TBP (Hs00427620_m1). The Ct values of the gene of interest were normalized using the Ct value of the housekeeping control, TBP. The gene expression fold change between the samples was calculated using the 2 −∆∆Ct method.

Protein Extraction and Western Blot
Cells were either trypsinized or scraped, followed by cell lysis on ice using 1× RIPA lysis buffer containing 1:100 protease inhibitor cocktail (Nacalai Tesque) and 1:100 phosphatase inhibitor cocktail (Thermo Fisher). The Pierce BCA Protein Assay Kit (Thermo Fisher) was used to determine the protein lysate concentration according to the manufacturer's protocols. Equal amount of protein samples were boiled in 1× Trident Laemmli SDS Sample Buffer (GeneTex, Irvine, CA, USA) containing 8% Beta-mercaptoethanol. The protein samples were separated in 10% SDS-PAGE gel and transferred onto nitrocellulose membrane (GE Healthcare Amersham, (GE Healthcare, Chicago, IL, USA; Amersham, UK). The membrane was blocked with 5% non-fat dry milk in 0.1% TBS-Tween and blotted with primary antibody overnight at 4 • C. Secondary antibody was added onto the membrane on the following day and incubated for an hour at room temperature. Signals were developed using Pierce™ ECL Western Blotting Substrate (Thermo Fisher) and visualized using the ChemiDoc XRS+ Gel Imaging System (Bio-Rad). For mTORC1 activity assessment, the cells were serum starved overnight and subjected to Western blotting as described above. Instead of non-fat dry milk, 5% BSA was used for blocking the membrane and diluting the antibodies. Primary antibodies used were PDGFB (Santa Cruz, Dallas, TX, USA, sc-365805, 1:1000), P-S6 ribosomal (Cell Signaling Technology, Danvers, MA, USA, Ser235/236, #4857, 1:3000), S6 ribosomal (Cell Signaling Technology, #2317, 1:1000), and B-actin (Santa Cruz, sc-69879, 1:5000). Secondary antibodies were polyclonal rabbit anti-mouse IgG/HRP (Dako, Santa Clara, CA, USA, P0260, 1:1000) and polyclonal swine anti-rabbit IgG/HRP conjugated (Dako, P0217, 1:1000).

Exogenous PDGFB Expression
The PDGFB CDS was amplified from the cDNA of 786-M1A cells using the Accuprime Pfx Supermix (Thermo Fisher). The amplified PDGFB CDS harbored the EcoRI and XbaI restriction sites upstream of the start codon and downstream of the stop codon, respectively, for ligation into the pLVX-Puro plasmid. The ligation and bacteria transformation were performed according to the cloning strategy described in the previous subsection. The presence of the ligated PDGFB CDS within the expression plasmid was confirmed via Sanger sequencing.

Acetone Protein Precipitation
The cells were washed twice with 1× PBS and cultured in serum free media overnight. The media was collected and spun down at 1000 RPM for 5 min to pellet the cells and debris. The media was transferred into conical tube, and 4 volume of ice-cold acetone was added and mixed well. The mixture was incubated at −20 • C overnight and spun down at 13,000× g for 10 min at 4 • C on the following day. The supernatant was removed, and the precipitated proteins were dissolved in 1× RIPA lysis buffer containing 1:100 protease inhibitor cocktails. The dissolved proteins were subjected to PDGFB Western blot.

PDGF-BB ELISA
The cells were washed twice with 1× PBS and cultured in serum free media overnight. On the next day, the media was collected and spun down at 1000 RPM for 5 min to pellet the cells and debris. The media were then subjected to PDGFB ELISA using RayBio Human PDGF-BB ELISA Kit (RayBiotech, Peachtree Corners, GA, USA) according to manufacturer's recommendations. The absorbance of standards and samples were read at 450 nm using the Varioskan LUX Multimode Microplate Reader (Thermo Fisher). The concentration of extracellular PDGFB in the media was determined using the Four Parameter Logistic Curve (https://www.myassays.com/four-parameter-logistic-curve.assay (accessed on 15 April 2022)).

Statistical Analysis
Two-tailed unpaired t-test or one-way ANOVA were used for the PDGF-BB ELISA experiments, and p-values lower than 0.05 were considered statistically significant. For the qRT-PCR, three independent experiments are shown unless stated otherwise in the figure legend. Each of the experiment is the average of three technical replicates.