The Multiple Regulations Of Rictor In Colorectal Cancer Progression

Part one:Post-transcriptional regulation of Rictor by protein kinase D in human colorectal cancer cellsBACKGROUNDmTORC2consists of mTOR, mLST8, Rictor, Sinl and PROTOR/PRR5. mTORC2phosphorylates Akt at Ser473thus resulting in Akt activation, knockdown of Rictor blocks Ser473phosphorylation. Knockdown of Rictor leads to growth inhibition and induces apoptosis in colorectal cancer cells. However, the molecular mechanism(s) involved in the regulation of Rictor expression remains unclear.Protein kinase D (PKD), a serine/threonine kinase family that includes PKD1, PKD2and PKD3, has been implicated in the regulation of cell proliferation and apoptosis. Selective PKD inhibitor has shown potential inhibitory roles in certain cancer cell proliferation, cell migration, and invasion. Recently, PKD3has been shown to increase Akt phosphorylation at Ser473. The role of PKD and its downstream effectors with regard to CRC proliferation and metastasis are not known.MATERIALS Go6976, G66983, MG132and cycloheximide (CHX) were purchased from Calbiochem (San Diego, CA). CID755673was from TOCRIS bioscience (Ellisville, Missouri). Antibodies against PKD2and PKD3were obtained from Abgent, Inc.(San Diego, CA). Antibodies against mTOR, Raptor, Rictor, PKD1, PKD2, PKD3, p-Akt, and Akt were obtained from Cell Signaling (Beverly, MA). GST-tagged PKD plasmids expressing wild-type PKD1and2were from Dr. Vivek Malhotra,(Universityof California, San Diego). Recombinant adenovirus expressing wild type (WT)-PKD3was kindly provided by Dr. Q. Jane Wang (University of Pittsburgh, Pittsburgh, PA). Tissue culture media and reagents were obtained from Invitrogen. Polyvinylidene difluoride (PVDF) membranes for Western blots were from Millipore Corp.(Bedford, MA). The enhanced chemiluminescence (ECL) system for Western immunoblot analysis was purchased from Amersham Biosciences.METHODSCell culture and transfectionThe human colon cancer cell lines, HT29and HCT116, were purchased from ATCC. HT29and HCT116cells were maintained in McCoy’s5A supplemented with10%fetal bovine serum (FBS). PKD inhibitors were initially dissolved in dimethyl sulfoxide (DMSO) and compared with cells treated with DMSO at the same final concentration. HT29and HCT116cells were transfected with the siRNA duplexes and plasmids using electroporation (Gene Pulser; Bio-Rad, Hercules, CA) and lipofectamine2000(Invitrogen, Carlsbad, CA), respectively.Stably Rictor knockdown HT29and HCT116cells were generated. Cells were infected with the control shRNA or shRNA to human Rictor lentivirus particles and stably expressing cells were selected with puromycin at a concentration of2.5μg/ml. The effective knockdown of Rictor was monitored by Western blot analysis.RNA extraction and Real-time PCRTotal RNA was extracted and DNase-treated (RQ1, Promega). Synthesis of cDNA was performed with1μg of total RNA using the reagents in the Taqman Reverse Transcription Reagents Kit from ABI (#N8080234). Quantitative real time RT-PCR analysis was performed with an Applied Biosystems Prism7000HT Sequence Detection System using TaqMan universal PCR master mix according to the manufacturer’s specifications (Applied Biosystems Inc., Foster City, CA). The TaqMan probe and primers for human Rictor was purchased from Applied Biosystems. Human GAPDH gene was used as endogenous control. Each sample was run in triplicate. Average Ct values were calculated and normalized to Ct values for GAPDH. The results were graphed with the corresponding standard deviation indicated with error bars in the figures.Protein preparation and Western blot analysisTotal protein (60μg) was resolved on a4-12%Bis-Tris gel and transferred to polyvinylidene difluoride (PVDF) membranes. Filters were incubated for1h at room temperature in blotting solution. Raptor, Rictor, mTOR, PKD1, PKD2, PKD3, p-Akt, Akt, and P-actin were detected with specific antibodies following blotting with a horseradish peroxidase-conjugated secondary antibody.ImmunoprecipitationCells were transfected with different plasmids as indicated,1%Triton lysis buffer was used. After pre-incubation with protein G PLUS-Agarose beads, equal amount of protein (500μg) were incubated with the indicated antibodies (1μg) using an end-to-end rotor overnight at4℃, followed by4h incubation with20μl of protein G PLUS-Agarose beads at4℃. Indicated buffer was used to wash the beads three times. Reactions were stopped by adding15μl of2×loading buffer. Samples were denatured by boiling for7min and separated by NuPAGE4-12%Bis-Tris gels.Cell proliferation analysesEqual numbers of cells were seeded onto24-well plates at a density of1×104per well in the appropriate culture medium with supplements. Cells carried with PKD shRNA or Rictor shRNA respectively were cultured for7days. Cells were trypsinized and counted using a cell counter (Beckman-Coulter).Soft agar assayMelt1.6%Agarose (www.lonza.com Cat.#50101) in distilled water in microwave, cool to40℃in a water bath, warm culture medium to40℃in water bath as well. Allow at least30minutes for temperature to equilibrate. Mix equal volumes of1.6% of Agar and medium to give0.8%Agar in medium. Add2ml/well of60mm dish, allow cooling down for30min. For plating, mix1ml of cell in medium and1ml0.8%Agar to a tube, mix gently and add2ml to each well (usually plate out in triplicate). Cool down in hood for30-60min. Add culture medium on the top agar for keeping humidity. Incubate assay at37℃in humidified incubator for10-14days. Stain plates with0.5ml of0.005%Crystal Violet (dilute with PBS) for>1hour, wash each well with PBS until every colonies could be seen clearly, count colonies using a dissecting microscope.Immunoh istoch emical AnalysisTissue microarrays containing normal and cancer tissues, A203(IV), were purchased from ISUABXIS through Accurate Chemical&Scientific Corporation. Sections were fixed to the slide by incubation in a dry oven at58℃for30min, and then sequentially transferred to xylene (5min,2changes),100%ethanol (3min,2changes),95%ethanol (3min,2changes) and rinsed with deionized water. Slides were allowed to cool at room temperature and rinsed twice with deionized water. Endogenous peroxidase was blocked by placing slides in3%H2O2/methanol block solution for10min, washed with deionized water, and placed in phosphate-buffered saline for5min. Slides were incubated with primary antibodies against human PKD1, PKD2, PKD3or Rictor overnight at4℃. Avidin-biotin peroxidase complex amplification and detection system (LSAB2, DAKO) with diaminobenzidine as chromagen was used. Negative controls (including no primary antibody or isotype matched mouse IgG) were used in each assessment.Statistical analysisThe data of real time-PCR, proliferation and soft agar were expressed as the mean of three independent expression±S.D, evaluated with t-test. P<0.05was defined as different signicantly. All of these analyses were made using SPSS Version13(SPSS, Chicago, IL, USA).RESULTSPKD2and PKD3regulate expression of RictorIn HT29and HCT116cells, both of Go6976(PKC and PKD inhibitors) and CID755673(PKD specific inhibitors) treatment inhibited expression of Rictor, while Go6983(PKC inhibitor) treatment had no affection on it (Fig.1-2A). Knocking down of PKD1, PKD2or PKD3individually could not inhibit Rictor expression (Fig.1-2B); overexpression of PKDl and PKD2did not increase Rictor (Fig.1-2C) as well. But we found that overexpressed PKD3using adenovirus increased expression of Rictor, with downstream target of Rictor-Akt S473phosphorylation increased at the same time (Fig.1-2D). Next, we set up PKD2and PKD3double knockdown stable cells, western blot analysis showed that both Rictor and p-Akt were decreased compared with control shRNA cells (Fig.1-2E).PKD2and PKD3bind with RictorHow do PKDs regulate expression of Rictor? First, we investigated if PKD regulated the mRNA level of Rictor by using real-time PCR. As shown in Fig.1-3A, no affection of Rictor mRNA was detected in PKD2and PKD3double knockdown cells. Similar result was gained in PKD3highly expression cells (Fig.1-3B), indicating the regulation of Rictor by PKD was post-transcriptional. Then, the interaction between PKD and Rictor was detected by Immunoprecipitation. Both of PKD2and PKD3bound with Rictor respectively, while no binding was investigated between PKD1and Rictor (Fig.1-3C and Fig.1-3D). Based on the data above, we hypothesized that PKD2and PKD3interacted with Rictor directly to regulate its protein expression.Rictor is degradation by proteasomal pathway and overexpression of PKD3inhibits degradation of RictorRictor is rapidly degraded by the ubiquitin-proteasome pathway. HT29cells were treated with MG132for the indicated times and subjected to Western blot analyses (Fig.1-4A). HT29were treated with CHX for the indicated times and Rictor levels were detected by Western analyses (Fig.1-4B).Mosty importantly, we investigated the role of PKD3during Rictor degradation by CHX chase assay. As shown in Fig.1-4C, we found that the degradation of Rictor was decreased with highly expressed PKD3, demonstrating that overexpression of PKD3inhibited Rictor degradation and increased stability of Rictor protein.PKD and Rictor do not affect CRC cells proliferation The previous data already showed that PKD affected Akt activities by regulating Rictor (Fig.1-2D and Fig.1-2E). So we was wandering that whether PKD and Rictor were involved in cell proliferation or not. After7days culture, cell numbers were counted by Beckman-Coulter, no difference was detected between control shRNA cells and PKD shRNA/Rictor shRNA cells individually (Fig1-5).Inhibition of PKD and Rictor reduces tumorigenesis of CRC cellsNext, soft agar assay was used to investigate tumorigenesis ability in those cells carried with different shRNA. As shown in Fig.1-6, inhibition of either Rictor or PKD2&3decreased tumorigenesis ability of colorectal cancer cells.Rictor as well as PKD2and PKD3are up-regulated in CRCs and liver metastasisIn Fig.1-7, we showed Immunohistochemical analysis of Rictor, PKD2and PKD3in CRC and liver metastasis sections. Rictor as well as PKD2and PKD3were up-regulated in primary CRCs and liver metastases compared with normal mucosa.CONCLUSIONS1. Inhibition of PKD2and PKD3decreases Rictor protein expression in CRC cells; consistently, overexpression of PKD3increases Rictor as well.2. PKD2and PKD3bind with Rictor individually; Rictor is a target of26S proteasomal degradation in CRC cells and PKD3stabilzes Rictor protein3. No affection on proliferation is detected while knocking down of both Rictor and PKD, while Rictor and PKD up-regulates tumorigenesis of CRC cells.4. Rictor as well as PKD2and PKD3are up-regulated in primary CRCs and liver metastases compared with normal mucosa, implying a potential role of PKDs upstream of the Rictor signaling pathway in the pathogenesis of CRC. Part two:Rictor regulates FBXW7-dependent c-Myc and cyclin E degradation in colorectal cancer cellsBACKGROUND:Rictor (Rapamycin-insensitive companion of mTOR) forms a complex with mTOR and phosphorylates and activates Akt. Activation of Akt induces expression of c-Myc and cyclin E, which are overexpressed in colorectal cancer and play an important role in colorectal cancer cell proliferation. Here, we show that Rictor associates with FBXW7to form an E3complex participating in the regulation of c-Myc and cyclin E degradation. The Rictor/FBXW7complex is biochemically distinct from the previously reported mTORC2and can be immunoprecipitated independently of mTORC2. Moreover, knocking down of Rictor in serum-deprived colorectal cancer cells results in the decreased ubiquitination and increased protein levels of e-Myc and cyclin E while overexpression of Rictor induces the degradation of c-Myc and cyclin E proteins. Genetic knockout of FBXW7blunts the effects of Rictor, suggesting that Rictor regulation of c-Myc and cyclin E requires FBXW7.MATERIALSMG132and cycloheximide (CHX) were purchased from Calbiochem (San Diego, CA). Rabbit monoclonal anti-Rictor antibody, used for both immunoblotting and immunoprecipitation, was purchased from Bethyl Laboratories (Montgomery, TX). Rabbit anti-mTOR, rabbit anti-phospho-Akt (Ser473), mouse anti-Ubiquitin, rabbit anti-USP28and rabbit anti-myc-tag antibodies were obtained from Cell Signaling (Beverly, MA). Mouse monoclonal anti-Flag-tag and anti-β-actin antibodies, and non-targeting control shRNA and Rictor shRNA lentiviral particles were from Sigma (St. Louis, MO). Protein G PLUS-Agarose beads, rabbit anti-cyclin E and mouse anti-Aktl antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA). Rabbit anti-c-Myc antibody was from Epitomics (Burlingame, CA). The plasmids encoding c-Myc, myc-tagged Rictor and myc-tagged mTOR were from Addgene (Cambridge, MA). The plasmid encoding Flag-tagged FBXW7a was kindly provided by Dr. Bruce E. Clurman (Fred Hutchinson Cancer Research Center, Seattle, WA). Human Rictor and non-targeting control siRNA METHODSCell culture and transfectionThe human CRC cell lines SW620, HT29and HCT116were from ATCC (Manassas, VA). SW620cells were maintained in DMEM supplemented with10%of FBS. HT29and HCT116cells were cultured in McCoy’s5A supplemented with10%of FBS. Wild-type and FBXW7-/-HCT116cells, kindly provided by Dr. Bert Vogelstein (The Johns Hopkins University School of Medicine, Baltimore, MD), were maintained in McCoy’s5A with10%FBS. SW620and HT29cells were transfected with the siRNA duplexes and plasmids using electroporation (Gene Pulser; Bio-Rad, Hercules, CA) and lipofectamine2000(Invitrogen, Carlsbad, CA), respectively.Stable Rictor knockdown SW620, HT29and HCT116cells were generated. Cells were infected with the control shRNA or shRNA to human Rictor lentivirus particles and stably expressing cells were selected with puromycin at a concentration of2.5μg/ml. The effective knockdown of Rictor was monitored by Western blot.RNA extraction and RT-PCRTotal RNA was extracted and DNase-treated (RQ1, Promega). Synthesis of cDNA was performed with1μg of total RNA using the reagents in the Taqman Reverse Transcription Reagents Kit from ABI (#N8080234). Reverse transcriptional PCR was performed by Hotstart DNA polymerase (Qiagen, Hilden, Germany), forward and reverse primers, and deoxynucleoside triphosphates in a final volume of25μl. The amplification product was of the expected sizes.Protein preparation and Western blot analysisCells were collected at48h after transfection or at the indicated time points after treatment. Equal amounts of cell lysates were resolved on a4-12%Bis-Tris gel and transferred to polyvinylidene fluoride membranes. Membranes were blocked by5%non-fat milk for1h at room temperature. Rictor, mTOR, phospho-Akt (Ser473), Akt1, c-Myc, cyclin E, Myc-tag, Flag-tag and β-actin were detected with specific antibodies following blotting with a horseradish peroxidase-conjugated secondary antibody and visualized using a chemiluminescence detection system. Immunoprecipitation and in vivo ubiquitination analysisCells were transfected with different plasmids as indicated,1%Triton lysis buffer or CHAPS lysis buffer was used. After pre-incubation with protein G PLUS-Agarose beads, equal amount of protein (500μg) were incubated with the indicated antibodies (1μg) using an end-to-end rotor overnight at4℃, followed by4h incubation with20μl of protein G PLUS-Agarose beads at4℃. Indicated buffer was used to wash the beads three times. Reactions were stopped by adding15μl of2×loading buffer. Samples were denatured by boiling for7min and separated by NuPAGE4-12%Bis-Tris gels. For in vivo ubiquitination analysis, His-ubiquitin plasmid was co-transfected with the indicated plasmids. Transfected cells were incubated with MG132(20μM) in serum free medium for4h before harvesting.Statistical analysisThere is no statistical analysis in this section.RESULTSRictor regulates protein expression of c-Myc and cyclin EActivation of Akt increases c-Myc and cyclin E expression. Since inhibition of mTORC2by knockdown of Rictor inhibits Akt activation, we were interested to know whether Rictor regulates c-Myc and cyclin E expression. Human colorectal cancer cells, SW620and HT29, were transfected with shRNA targeting Rictor and stable cell lines were established. Knockdown of Rictor did not obviously affect the expression of c-Myc and cyclin E; however, with serum starvation, knockdown of Rictor increased protein expression of c-Myc and cyclin E (Fig.2-3A). In addition, treatment with MG132, a specific cell-permeable proteasome inhibitor, attenuated the increases of c-Myc and cyclin E protein expression resulting from Rictor knockdown (Fig.2-3B), indicating that the26S proteasome pathway was involved in this regulation. To further confirm our findings, we used Rictor siRNA containing a different sequence from the Rictor shRNA to decrease Rictor expression. Consistently, knockdown of Rictor by transient transfection with siRNA targeting Rictor resulted in a significant increase of c-Myc and cyclin E protein expression (Fig.2-3C). To further determine the role of Rictor in the regulation of c-Myc and cyclin E protein expression, SW620and HT29cells were transiently transfected with a myc-tagged Rictor plasmid and the transfected cells were serum starved for24h before harvesting. As shown in Fig.2-3D, overexpression of Rictor decreased c-Myc and cyclin E protein levels in both SW620and HT29cells, suggesting that Rictor regulates the protein levels of c-Myc and cyclin E associated with serum deprivation. Interestingly, RT-PCR assay showed that knockdown of Rictor inhibited mRNA level of c-Myc and cyclin E (Fig.2-3E). Considering the increases of c-Myc and cyclin E protein expression by Akt activation and knockdown of Rictor resulting in the dephosphorylation and inhibition of Akt, our results demonstrate that Rictor regulates c-Myc and cyclin E protein in an mTORC2/Akt pathway independent fashion.Rictor interacts with FBXW7without mTORDegradation of c-Myc and cyclin E was targeted by FBXW7as an E3component. And it has been shown that Rictor forms a complex with cullinl to degrade SGK1protein. To determined whether Rictor interacts with FBXW7. As shown in Fig2-4A, Myc-Rictor constructs were co-transfected with Flag-FBXW7a, Flag-FBXW7p and Flag-FBXW7y individually. The interactions between Rictor and three FBXW7isforms were detected. FBXW7a is expressed at a much higher level than FBXW7P and FBXW7y in most human cell lines, and usually plays a major role as an E3ligase to the downstream targets. Therefore, we transfected HCT116cells with Flag-FBXW7a together with Myc-Rictor and the interactions between FBXW7and Rictor were detected. Two different lysis buffers were used in this assay:1%Triton and CHAPS lysis buffers. CHAPS buffer was used as a mild buffer to keep mTORC2intact while1%Triton was used to dissociate mTORC2. As shown in Fig.2-4C, Flag-FBXW7a was clearly copurified when Rictor was immunoprecipitated from cell lysates extracted with both1%Triton lysis buffer and CHAPS buffer, but the interaction between mTOR and Rictor was only observed in the lysates extracted using CHAPS buffer. Although it has been reported that FBXW7interacts with mTOR to promote its ubiquitination and degradation, our results indicate that Rictor forms a complex with FBXW7independent of mTORC2. Furthermore, the binding between Rictor and FBXW7is not affected by serum-deprivation (Fig2-4D). Rictor interacts with FBXW7to regulate c-Myc and cyclinEWe next determine whether Rictor participates in FBXW7-dependent regulation of c-Myc and cyclin E degradation, we generated Rictor shRNA stable cell lines based on wild-type and FBXW7-/-HCT116cells. As shown in Fig.2-5A, knockdown of Rictor induced the expression of c-Myc and cyclin E in wild type HCT116cells but not in FBXW7-/-cells. These data suggest that Rictor regulation of c-Myc and cyclin E is FBXW7dependent. To further demonstrate the role of Rictor in the FBXW7E3ligase complex, HCT116control shRNA and Rictor shRNA stable cell lines were transfected with empty vector (control) or Flag-FBXW7a plasmid and c-Myc and cyclin E protein levels were analyzed (Fig.2-5B). Overexpression of FBXW7αdecreased the protein levels of c-Myc and cyclin E as expected. However, this degradation by FBXW7a was significantly attenuated in Rictor shRNA cells, demonstrating a role of Rictor, bound with FBXW7a as a part of E3complex, to induce the degradation of c-Myc and cyclin E.Rictor regulates stability of FBXW7Rictor didn’t affect the mRNA level of FBXW7by RT-PCR assays (Fig.2-6A), which means Rictor regulates FBXW7through post-transcriptional regulation. Next, in vivo ubiquitination assay was performed to investigate the ubiquitination of FBXW7with or without Rictor expression. As expected in Fig.2-6B, the ubiquitination of FBXW7a was diminished in Rictor knockdown cells compared with control cells. Moreover, in Rictor stably knockdown cells, the stability of FBXW7αwas decreased; futher confirmed our hypothesis (Fig.2-6C).Rictor/FBXW7promotes c-Myc and cyclin E ubiquitinationFBXW7degrades c-Myc and cyclin E through the ubiquitination of c-Myc and cyclin E proteins. To address whether Rictor promotes the ubiquitination of c-Myc and cyclin E, in vivo ubiquitination assays were performed. c-Myc or cyclin E and His-ubiquitin were expressed by transient transfection. Knockdown of Rictor reduced the ubiquitination of both c-Myc (Fig.2-7A) and cyclin E (Fig.2-7C). As expected, the ubiquitination of c-Myc and cyclin E was diminished in FBXW7-/-HCT116cells compared with wild-type and, importantly, Rictor-dependent regulation of c-Myc and cyclin E ubiquitination was also blunted in FBXW-/-HCT116cells (Fig.2-7B and Fig.2-7D). Taken together, these data indicate a role of Rictor in the regulation of c-Myc and cyclin E protein ubiquitination and degradation.Rictor reduces the stability of c-Myc and cyclin ETo determine whether Rictor regulates the stability of c-Myc and cyclin E, we next performed CHX chase assays. Time-course experiments showed that the half-lives of c-Myc (Fig.2-8A) and cyclin E (Fig.2-8C) were prolonged in HCT116cells with Rictor shRNA compared with HCT116cells with non-targeting control shRNA. The sequence of Rictor shRNA used in this study was designed to target the3’-UTR region of Rictor and would not affect the expression of transfected myc-Rictor plasmid. Therefore, Rictor protein levels can be rescued in Rictor shRNA stable cell lines with myc-Rictor overexpression. Indeed, transfection of HCTl16Rictor shRNA stable cells with myc-Rictor plasmid decreased the half-life of c-Myc (Fig.2-8B) and cyclin E (Fig.2-8D). Our results demonstrate that Rictor is required for the integrity of the FBXW7E3complex to promote the degradation of c-Myc and cyclin E.Rictor binds with USP28through FBXW7USP28is a deubquitin specific protein, which can bind with FBXW7to deubiquitinate its downstreams. c-Myc, as one of USP28specific targets, was decreased with USP28knockdown in HT29; consistently, overexpression of USP28in HCT116induced c-Myc protein level (Fig.2-9A). Rictor bound with USP28via FBXW7(Fig.2-9B), but Rictor did not have affect on the deubiquitination regulation of c-Myc by USP28(Fig.2-9C and Fig.2-9D).Rictor/FBXW7may mediate Rapamycin resistance in CRCRictor activity was also inhibited by long-time exposure of Rapamycin. Some of cancer cells were rapamycin sensitive, while some of them were rapamycin resistant and the mechanisms of rapamycin resistance is still unclear. As shown in Fig.2-10A,24hours exposure of rapamycin in rapamycin resistant cells HT29and S W620led to the decrease of Rictor, with increase of c-Myc, cyclin E and Akt phosphorylation at the same time. And no increase of c-Myc, cyclin E mRNA level was detected in the same cell lines with rapamycin treatment (Fig.2-10B), indicating rapamycin induced c-Myc and cyclin E was mTORC2/Akt pathway independent. Based on these data, we hypothesize that rapamycin resistance is mediated by increased c-Myc and cyclin E via Rictor/FBXW7regulation.CONCLUSIONS1. Our findings identify Rictor as an important component of FBXW7E3ligase complex participating in the regulation of c-Myc and cyclin E protein ubiquitination and degradation.2. Our results suggest that elevated growth factor signaling may contribute to decrease Rictor/FBXW7-mediated ubiquitination of c-Myc and cyclin E, thus leading to accumulation of cyclin E and c-Myc in colorectal cancer cells.3. Our study figures out a possible mechanism of Rapmycin resistance in colorectal cancer cell, which yields another role of Rictor during tumor progression.