mTORC2 mediate FLCN-induced HIF2α nuclear import and proliferation of clear cell renal cell carcinoma

Clear cell renal cell carcinoma (ccRCC), as the most important type of renal carcinoma, has a high incidence and easy metastasis. Folliculin (FLCN) was identified as a tumor suppressor gene. Its deletions and mutations are associated with a potential risk of kidney cancer. At present, the specific molecular mechanism of FLCN-induced proliferation, invasion and migration in clear cell renal cell carcinoma remains elusive. In this study, we demonstrated that FLCN controled cell proliferation, invasion and migration through PI3K/mTORC2 pathway. FLCN combined with HIF2α in various normal and cancerous renal cells, and mTORC2 mediate FLCN effectively alleviated the deterioration of renal cancer cells by degrading HIF2α. Silencing of FLCN showed promotion of HIF2α protein expression, which in turn led to an increase in downstream target genes Cyclin D1 and MMP9. Moreover, when interfering with siFLCN, HIF2α degradation rate was delayed, and the time of entry into the nucleus was advanced. Taken together, our study illustrated that mTORC2 promoted the specific molecular mechanism of HIF2α by down-regulated FLCN, and might be a new therapeutic target against renal cancer progression.

For Western blotting，equal amounts of protein were run on SDS polyacrylamide gels and transferred to nitrocellulose membrane. The resulting blots were blocked with 5% non-fat milk (in TBST) and incubated with primary antibodies overnight at 4 . Then protein bands were detected by incubating with secondary antibodies for 1-2h at room-temperature and visualized with ECL reagent (Millipore, Billerica, MA, USA) by Chemi Doc XRS+gel imaging system (Bio-Rad, USA). Densitometry analysis was performed using Quantity One software, and band intensities were normalized to those of β -Actin and GAPDH. IP was performed as previously described 19 . Briefly, the cell lysates were centrifuged to remove the cell debris and then were incubated with beads (Abmart) for 1-2 h. Endogenous FLCN and HIF2α was immunoprecipitated using an anti-FLCN or anti-HIF2α polyclonal antibody. The beads were boiled after extensive washing; resolved via SDS-PAGE gel electrophoreses, and analyzed via immunoblotting. The protein concentration was detected using the Odyssey system.

Cytoplasmic and nuclear protein extraction
Cytoplasmic and nuclear proteins were obtained using the Nuclear and Cytoplasmic Protein Extraction Kit (Beyotime) according to the manufacturer's instructions. Cells were harvested by centrifugation and resuspended in cytoplasmic extraction agent A. The solution was vigorously vortexed and then incubated on ice for 10 minutes. Then, cytoplasmic extraction agent B was added to the cell pellet. The pellet was vortexed and incubated on ice for 1 minute. The pellet was vortexed again and centrifuged for 5 minutes at 5000 rpm. The Supernatant was collected as cytoplasmic extract. The insoluble (pellet) fraction was suspended by nuclear extraction agent. After vortexed several times, the mixture was centrifuged for 10 minutes at 12000 rpm. The supernatant was collected as nuclear extract.

Real time quantitative PCR
Total RNA was extracted using Trizol reagent (Invitrogen) and reversely transcribed with HiScript ® Q RT SuperMix for qPCR (Vazyme, Nanjing, China) according to the protocol. Real time PCR analyses were performed with AceQ ® qPCR SYBR ® Green Master Mix (High ROX Premixed) (Vazyme) on ABI StepOne™ Real Time PCR System (Applied Biosystems, Foster City, CA) at the recommended thermal cycling settings: one initial cycle at 95°C for 10 minutes followed by 40 cycles of 15 seconds at 95°C and 60 seconds at 60°C. The gene expression levels were calculated with Rt (2 −ΔΔCT ) values by StepOne Software v2.1 (Applied Biosystems). Primer sequences used in qRT PCR were listed in TABLE 2.

Immunofluorescence microscopy
For immunofluorescent staining, cells were seeded on coverslips. The cells were transfected with or without siFLCN and FLCN plasmids for the indicated time, fixed in ice-cold 4% paraformaldehyde for 20 min at room temperature, rinsed with 3×phosphate-buffered saline (PBS) for 5 min and permeabilized with 0.1% Triton X-100 before blocking in 1% BSA for 1 h at room temperature. The cells were incubated with primary antibodies at 4 °C overnight, washed, and then incubated with Alexa or FITC-coupled secondary antibodies for 1.5 h at room temperature in a moist chamber. After washing with PBS, the samples were mounted with DAPI Fluoromount G (Southern Biotech, Birmingham, AL). Images were acquired using an Olympus BX51 microscope coupled with an Olympus DP70 digital camera and Carl ZEISS MicroImaging GmbH (Jena,Germany,700). Images were representatives of three independent experiments.

Migration and Matrigel invasion assays
For wound healing assay, 786-O or ACHN cells were grown as described above and were plated in 6 well plates on glass cover slips. Approximately 24 h later, when the confluence reached 95 ~ 100 %, the cells were incubated overnight in DMEM and the wounding was performed by scraping through the cell monolayer with a 10 μ L pipette tip. The Medium and the nonadherent cells were removed, and the cells were washed twice with PBS. Images were collected of the 0 h time point using an inverted microscope (Carl Zeiss Meditec, Jena, Germany). The Cells were permitted to migrate into the area of clearing for 6h, 12h and 24h in incubator. And then, the cells were removed from the incubator and imaged using the same inverted microscope. Care was taken to align the scratch along the y axis of the camera to aid subsequent image quantification.
For Matrigel invasion assay, the cell invasion assay was performed using Transwells (8 μ m pore size, millipore) with inserts coated with Matrigel (50 mg/ mL, BD Biosciences). 786-O and ACHN cells (1.0 × 10 5 cells/well) were seeded in the upper chambers with 0.1 mL matrigel and allowed to invade through matrigel for 12 h and 24 h. The cells remained on the membranes were fixed with 4% paraformaldehyde and stained with 0.5% crystal violet.
The cell pictures were taken by Nikon TS100 (Tokyo, Japan) and counted by Image J software. All assays were performed at least three times.

CCK-8 assay and Flow cytometry analysis
The Cells transfected with plasmids or siRNA were seeded at a density of 3.5 × 10 3 cells per well into 96-well plate. After culture, the cells were washed, add 10 µl Cell Counting Kit-8 (CCK-8) per 100 µl culture medium and the plate was incubated in the dark for 2 h, followed by measurement of absorbance value at 450 nm using a microplate absorbance reader (Bio-Tek, Elx800, USA). The fold growth was calculated as the absorbance of drug treated sample/control sample absorbance ×100%.
Cell cycle analysis was performed by flow cytometry. Briefly, cells were harvested and fixed in 80% ice cold ethanol overnight. Then the cells were incubated with RNase A and propidium iodide staining solution at 37°C for 30 minutes in darkness. Subsequently, the stained cells were analysed using flow cytometry.

Immunohistochemistry
Renal cancer tissue microarrays were purchased from Outdo biotech (Shanghai, China). 75 cases of Renal clear cell carcinoma samples and their corresponding paracancerous tissue samples were used for immunohistological staining in our study. Briefly, after microwave antigen retrieval, microarray tissues were incubated FLCN (Proteintech), or HIF2α antibody (Sigma) overnight at 4°C, followed by 1 hour incubation with secondary antibody. The sections were developed in DAB solution under microscopic observation and counter stained with haematoxylin. Immunohistochemical staining results were taken by using Axioskop 2 plus microscope (Carl Zeiss). FLCN and HIF2α immunostaining was scored by immune reactive score (IRS) as described previously 20 .

Statistical analysis
Student's t-test and repeated-measures were used to analyze differences between groups by using the SPSS statistical software program (Version 19.0; SPSS, Chicago, IL, USA). Data were presented as mean ± S.E.M. Values of P < 0.05 were considered statistically significant. All experiments were repeated at least three times.

FLCN is involved in proliferation of ccRCC cells
Firstly, we tested whether FLCN plays a crucial role in the HIF2α expression in ccRCC cells and verified the definite mechanisms involved. We first detected the protein levels of FLCN in normal renal tubular epithelial cells and clear cell renal cell carcinoma by Western blotting. The results showed that FLCN was poorly expressed in the 786-O and ACHN ccRCC cells compared to renal tubular epithelial cells HK-2. However, HIF2α protein expression level was opposite to FLCN trends (Fig. 1A). The expression of the FLCN and HIF2α had been examined by Realtime-PCR analysis in three cell lines (Fig. 1B). Western blotting confirmed siRNA mediated specific knockdown of FLCN in 786-O and ACHN cells (Fig. 1C), and the results of CCK-8 assay showed that the knockdown efficiently promoted the cell viability (Fig. 1D).

FLCN negatively regulates ccRCC cell cycle and invasion in vitro
Since FLCN knockdown increased the viability of ccRCC cells, the second set of analyses examined the role of FLCN in cell motility. The results showed that FLCN knockdown affected the cell cycle in 786-O cells, and the number of cells in S phase was significantly increased ( Fig. 2A left). In contrast, after transfection with the FLCN plasmid, the proportion of cells in S phase was reduced ( Fig.2A right). The above experiment was also performed with ACHN cells, and similar results were obtained (Fig.2B). The results of wound healing assays showed that the rate of migration of cells transfected with siFLCN #1 and #2 was increased when it was compared to the control group in 786-O cells (Fig. 2C). FLCN overexpression gave the opposite results. The same conclusion was also drawn in ACHN cells (Fig. 2D). These results suggested a positive role for FLCN in regulating ccRCC cells migration, and compared with long-term migration rate (24h), the short-term (12h) rate was more obvious. We checked cell invasion by transwell assay, and found that FLCN knockdown increased the invasion as compared with control in 786-O and ACHN cells (Fig.2E).

Cells lacking FLCN have elevated levels of HIF2α expression
To explore the mechanism of FLCN regulating cell proliferation and cell cycle, we focused on its putative interacting protein HIF2α. The silencing FLCN expression in 786-O cells with siRNA targeting FLCN, or the upregulating with Flag-FLCN plasmid was performed. The knockdown or overexpression efficiency was determined by Western blotting. The results showed that HIF2α and FLCN protein expression are negatively correlated (Fig. 3A). Conclusions in the ACHN cells are consistent with the above (Fig. 3B). In order to confirm HIF2α regulation by FLCN, the cells were treated with siFLCN or FLCN-overexpression plasmids, and then analysed for FLCN and HIF2α mRNA level by real-time PCR. The results showed that the expression of FLCN gene was obviously changed with the treatment, but none of this program caused a significant effect on HIF2α gene as compared with control ( Fig. 3C and 3D). Then the expression of the genes of cyclin family (FLCN/HIF2α downstream cell-viability related genes) in HK-2 (normal), 786-O and ACHN (ccRCC) cells was detected. Consistent with expectations, the expression level of cyclin D1 was significantly higher than others (Fig. 3E). According to this, the follow-up experiment used cyclin D1 as a marker, to detect if HIF2α was directly downstream of FLCN. We co-transfected siFLCN/siHIF2α in 786-O cells, and detected cyclin D1 gene expression levels by real-time PCR (Fig. 3F left). Similarly, overexpression of FLCN reduced cyclin D1 gene expression levels, and under these conditions, co-transfections with Epas1 plasmid rescued the cyclin D1 mRNA level (Fig. 3F right). We also observed the same phenomenon in ACHN cells (Fig. 3G).

FLCN knockdown restricts HIF2α degradation of ccRCC cells
In order to confirm whether FLCN bound to HIF2α directly, we further identified the endogenous protein interaction between renal cancer cells and normal renal tubular epithelial cells by immunoprecipitation (Fig. 4A). Whereas 786-O/ACHN cells underwent markable reduction or increase in FLCN, the abundance of HIF2α mRNA was not altered greatly ( Figure. 3C and D). Thus, we concluded that instead of transcription-dependent mechanism, FLCN might modulate HIF2α expression by promoting its degradation process. We found that FLCN knockdown inhibited the HIF2α protein level with CHX stimulation in 786-O ( Fig. 4B and C top) and ACHN cells ( Fig. 4B and C bottom). To explore the causes for the above phenomenon, the cells were treated with KU0063794 (mTOR inhibitor) for different times, and then with or without ubiquitinated proteasome inhibitors to verify whether FLCN degraded HIF2α through the ubiquitination pathway. After treatment with KU0063794, inhibition of mTOR was also shown to retard HIF2α degradation in 786-O ( Fig. 4D and E top) and ACHN cells ( Fig. 4D and E bottom).

FLCN knockdown induces HIF2α transportation to nucleus
In view of the above findings, we further explored the influence of FLCN on the function of HIF2α in cell proliferation. Firstly, we examined the exact location of FLCN and HIF2α in the cells, and found that these two molecules were co-localized ( Fig. 5A). Then immunofluorescence results showed that knockdown of FLCN promoted HIF2α to enter the nucleus (Fig. 5B). After 46-48 hours (approximately the time of cell division generation) of FLCN knockdown, HIF2α turned into massive in the 786-O cell nucleus at this time point (Fig. 5D) compared to the control group ( Fig  5C). Therefore, we verified the validity of this conclusion again by Western blotting in 786-O cells. The results showed visible nucleus HIF2α, which was further increased by FLCN knockdown 46-48h ( Fig. 5E bottom), and a slight decrease in expression of HIF2α in the cytoplasma (Fig. 5E top). The same experiment was also repeated in ACHN cells (Fig. 5F), and similar results were obtained.

HIF2α mediates FLCN induced cell invasion via PI3K/mTORC2 signaling pathways
We examined MMP9 protein and mRNA levels by Western blotting and RT-PCR respectively in 786-O cells transfected with siFLCN or FLCN plasmids ( Fig.6A and 6B). We then explored the mechanism of how PI3K/mTORC2 signaling regulated the invasion of FLCN. Recent studies have showed that mTORC is activated in kidney tumors that lack FLCN 10 . Similarly, Tiffiney R. Hartman et al. reported that downregulation of BHD reduced the TORC1 activity 21 , which suggested a close relationship between FLCN and mTORC. It was also reported that mTORC promoted HIF expression, mTORC2 focused on HIF2α , and HIF1α was sensitive to rapamycin but HIF2α was tolerant 22 . Based on the above observations, our study hypothesized that FLCN relied on the PI3K/mTORC signaling pathway to achieve regulation of HIF2α. Firstly, we detected the expression of FLCN and found that silencing or overexpression of this gene had no effect on p-AKT S473, an indicator of TORC1 activity (Fig. 6C). Then, we treated 786-0 cells with LY294002, an inhibitor of PI3K, and found that that LY294002 treatment significantly increased HIF2α expression (Fig. 6D). In order to verify the clear upstream and downstream of FLCN, we treated the cells with KU0063794 (mTORC inhibitor), which promoted FLCN expression (Fig. 6E), but could not find significant change in FLCN after adding AKT inhibitor MK2206 (Fig. 6F). These results indicated that mTORC was upstream of FLCN/HIF2α. Finally, we found that rapamycin stimulation did not affect FLCN expression (Fig.6G), which further validated FLCN regulation of HIF2 via mTORC2, but not mTORC1.

FLCN is marginally expressed in human renal cancer samples and correlated with HIF2α expression
In order to investigate whether the in vitro experimental findings were consistent with the pathogenesis and progression of renal cancer in humans, we examined FLCN and HIF2α expression patterns in renal cancers (75 paired cases) by a tissue microarray (Fig. 7A). Immunohistochemistry results indicated that FLCN was marginally expressed in tumor tissues compared with matched para-cancerous tissues ( Figure 7B) but HIF2α was reversed. The immunoreactive score (IRS) was calculated as the intensity of the staining reaction multiplied by the percentage of positive cells. Based on the analysis of 75 groups of renal cancer tissues and para-cancerous tissues, we found that the FLCN expression level was low in tumor tissues, while HIF2α expression level was in a reversed pattern ( Figure 7C and 7D). Overall, the clinical data supported our in vitro results and revealed a negative link between FLCN and HIF2α protein expression in renal cancer.

Discussion
There are many reports on genes associated with kidney cancer pathogenesis 23,24 , including VHL, FH, MET,FLCN,etc. 25,26 . The FLCN gene is a tumor suppressor 27 , but the mechanism by which FLCN deficiency causes renal cancer is not completely understood. The PI3K/AKT/mTOR pathway, which drives cell proliferation, motility, and migration in a variety of malignancies 23,28 . Loss of FLCN is linked with activity of the AKT/mTORC pathway 29,30 , leading to Birt-Hogg-Dubé (BHD) autosomal dominant syndrome, increasing the risk of hybrid oncocytic RCC and pulmonary cysts 21,31 . It can be seen that these four genes are closely related to HIF and RCC. In summary, we decided to focus on FLCN and study the effects of its inactivation on ccRCC, HIF and the related signaling pathways involved.
A current literature reports that there is a controversy about the relationship between FLCN and HIFα 4 . HIFα is highly expressed in BHD tumors, and FLCN knockout leads to increased activity of HIFα. However, most of the studies just stayed in the phenomenon. The mechanism is not discussed and the specific relationship between FLCN and HIF1α/HIF2α is unclear. The main points are as follows: RS Preston et al. thought the absence of the FLCN gene did not affect HIF1α levels under normoxia/hypoxia, and it is also mentioned that HIF2α is highly expressed under hypoxic conditions 32 . In another report, IHC results indicate that FLCN deficiency increases HIF1α and its downstream expression 33 . A document in the same year claimed that HIF1α increased in the lack of FLCN and promoted tumor growth 34 . In addition, our study showed that silencing FLCN enhanced the expression of HIF2α protein, and implicated a novel molecular mechanism that FLCN interacted HIF2α and impacted ccRCC physiological activity. In the process, we also detected the mRNA level of HIF2α, and found that there was no significant change, suggesting that FLCN did not regulate HIF2α at the transcriptional level but might control the degradation of HIF2α.
In the experiment, it could be clearly seen that the RNA and protein levels of FLCN in ccRCC were relatively low compared to normal renal tubular epithelial cells. This again proved that BHD acted as a tumor suppressor gene, but its role might be more prone to maintain the normal physiological function of the cells. And BHD deletion activated the downstream signaling pathway, resulting in tumor formation, rather than saving the deterioration of existing tumor cells. Based on the above observations, we transfected siFLCN with ccRCC to demonstrate that FLCN deficiency had a detailed mechanism for inducing renal cell carcinoma. At the same time, FLCN overexpression experiments were also performed to verify whether the increasing FLCN expression levels in renal tumors caused by other factors contributed to the alleviation of tumor cell deterioration. In the experiment, we selected 786-O and ACHN two ccRCC cell lines, of which 786-O was a cell line with VHL gene deletion, and HIF2α was still highly expressed under normoxia, which was more convenient for us to observe the results. Then, we added CHX interfering protein translation to the cell line knocked down by FLCN. The results showed that compared with the control group, the content of HIF2α in the FLCN interference group could be maintained at a relatively stable level within 0-8 hours after the addition of CHX, indicating that FLCN might accelerate the degradation of HIF2α protein. We hypothesized that this mechanism might have some similarities with VHL ubiquitination and degradation of HIF1α 35-37 . To further verify the specific mechanism of this degradation, we added the mTORC inhibitor to up-regulate the FLCN, and at the same time added the proteasome inhibitor MG-132 to inhibit the ubiquitination degradation pathway. The results showed that although increased FLCN could reduce HIF2α, but after adding MG132 the expression of HIF2α was significantly increased, which indicated that FLCN was likely to achieve the degradation of HIF2α by ubiquitination, thereby inhibiting the proliferation of tumors.
In addition, during the experiment, we also unexpectedly found that the transcription factor HIF2α in the siFLCN group increased significantly in the nucleus at 46-48 hours compared with the control group. In order to study the occurrence of this phenomenon, it was caused by the FLCN knockdown which promoted the acceleration of HIF2α into the nucleus, rather than the increase of the total protein of HIF2α. We transfected siFLCN in ccRCC and observed the cell morphological phenomenon from 0 h with immunofluorescence, and then detected the distribution of HIF2α in the cells every two hours (only the distribution of 44 h-50 h was shown in the figures). The results showed that HIF2α in the siFLCN group of the 786-O cell line began to enter the nucleus in 46-48 hours, rather than gradually increasing with time. At the same time, the distribution and content of HIF2α in the nucleus of the control group did not change significantly. This indicated that FLCN could not only bound and degraded HIF2α, but also participated in the regulation of HIF2α nuclear import: silencing FLCN, HIF2α might move into the nucleus and activated downstream related tumor signaling pathways.
In conclusion, we demonstrate the specific mechanism by which FLCN regulates HIF2α-induced renal cancer cells proliferation, migration, and invasion. FLCN interacts with HIF2α and promotes HIF2α degradation in human renal cancer cells. Although the current study has contributed to the mechanistic understanding the role of FLCN regulates renal cancer cell physiological activities, the issue as to how FLCN precisely adjusts HIF2 degradation in ccRCC cells is unlikely to be settled in this paper. Our findings point out that FLCN/HIF2α expression may be a novel therapeutic target for preventing renal cancer proliferation and invasion.

6.Conflict of interest
The authors confirm that there are no conflicts of interest.  i  c  k  e  r  s  o  n  M  L  ,  W  a  r  r  e  n  M  B  ,  T  o  r  o  J  R  ,  e  t  a  l  .  M  u  t  a  t  i  o  n  s  i  n  a  n  o  v  e  l  g  e  n  e  l  e  a  d  t  o  k  i  d  n  e  y  t  u  m  o  r  s  ,   l  u  n  g  w  a  l  l  d  e  f  e  c  t  s  ,  a  n  d  b  e  n  i  g  n  t  u  m  o  r  s  o  f  t  h  e  h  a  i  r  f  o  l  l  i  c  l  e  i  n  p  a  t  i  e  n  t  s  w  i  t  h  t  h  e  B  i  r  t  -H  o  g  g  -D  u  b