Plk1-Mediated Phosphorylation of TSC1 Enhances the Efficacy of Rapamycin Free

The PI3K/AKT/mammalian target of rapamycin (mTOR) pathway is highly conserved, and its activation is tightly controlled via a multistep process (1). Upon stimulation with growth factors, PI3K is activated by receptor tyrosine kinases (RTK) to convert phosphatidylinositol 3,4-bisphosphate (PIP2) to phosphoinositide 3,4,5-trisphosphate (PIP3). Phosphoinositide-dependent kinase 1 (PDK1) and AKT bind to PIP3, allowing PDK1 to access and phosphorylate T308 in AKT and thereby activate AKT (2, 3). AKT can subsequently phosphorylate and inactive TSC2 by inducing its release from the lysosome (4–6). The lysosomal, small Ras-like GTPase, Rheb, which is regulated by the TSC complex, activates mTORC1 (7). mTORC1 substrates include the eukaryotic translation initiation factor 4E binding protein 1 (4E-BP1), and ribosomal protein S6 kinase 1 (S6K1), which, in turn, phosphorylates the ribosomal protein S6 to promote protein synthesis (8). In addition, the PI3K/AKT/mTOR pathway is important for the regulation of cell-cycle progression (9–11). Consistent with these observations, it was reported that AKT activity is fluctuated across the cell cycle (12). Furthermore, it was shown that TSC1 is threonine-phosphorylated during nocodazole-induced G₂–M arrest (13).

A significant number of studies have pointed to failure in various critical mitotic events as a cause of aneuploidy in tumors (14–16). The regulation of proper mitotic progression is predominantly controlled by several conserved serine/threonine kinases, such as Cdk1, Plk1, and aurora kinases (17). It has been documented that Plk1 is involved in almost every step of mitosis (18). Thus, it is not surprising that Plk1 is overexpressed in many cancer types (19–22). More importantly, recent studies have also linked Plk1 with other cancer-associated pathways, such as DNA damage response (23–28), p53, and the PI3K/AKT/mTOR pathway (29, 30). For example, a cross-talk between Plk1 and the p53 tumor suppressor has been described previously (31, 32). In another study, Plk1 elevation was shown to cause PTEN inactivation (33). In line with this observation, Plk1-associated activity was demonstrated to contribute to the low-dose arsenic-mediated metabolic shift via activation of the PI3K/AKT/mTOR pathway (34). Furthermore, it was reported that the phosphorylated form of TSC1 interacts with Plk1, and that the interaction between Plk1 and the TSC1/TSC2 complex regulates local mTOR activity (35).

Here, we show that the activity of mTORC1 is correlated with Plk1 activity and inversely correlated with AKT activity during cell cycle. Mechanistically, Plk1 directly phosphorylates TSC1 at S467 and S578. We show that Plk1 phosphorylation of TSC1 leads to inactivation of the TSC1 complex, thus activation of mTORC1 in mitosis, and that cells expressing the hyperphosphorylated form of TSC1 have apparent mitotic defects, but with a higher sensitivity to rapamycin. Together, these observations and others' previous findings support a new working model in which AKT activates the TSC/mTORC1 axis in response to growth factors in interphase, whereas Plk1, instead of AKT, regulates the TSC/mTORC1 pathway during mitosis.

The cell lines were obtained from ATCC. The cell lines were authenticated by ATCC and tested for absence of Mycoplasma contamination (MycoAlert, Lonza). The cells used in the experiments were within 10 passages from thawing. HeLa and HEK293T cells were cultured in DMEM supplemented with 10% FBS, 100 U/mL penicillin, and 100 U/mL streptomycin at 37°C in 8% CO₂. PC3 cells were cultured in F-12K medium supplemented with 10% FBS. After cells were transfected with plasmids with Lipofectamine (Invitrogen) for 48 hours, cells were harvested for immunoblotting or immunofluorescence. myc-TSC1 and HA-TSC2 expression plasmids were obtained from Addgene. Various TSC1 mutants were created with the QuikChange Site-Directed Mutagenesis Kit (Stratagene). The identities of all plasmids were confirmed by sequencing.

To arrest cells at mitosis, cells growing in 100-mm dishes were treated with 100 ng/mL nocodazole for 24 hours. After floating cells were collected into 50 mL tubes containing 10 mL of precold PBS, additional mitotic cells were collected by shaking off dishes for 10 minutes on ice. The procedure was repeated one more time. Cells were spun down at 2,000 rpm for 2 minutes, resuspended in precold 20 mL of PBS, and kept on ice for 30 minutes. The procedure was repeated 2 more times to completely remove nocodazole. After cells were checked microscopically to ensure they were in a good condition, cells were reseeded at 80% confluence to ensure they were ready for experiments 24 hours later. To arrest cells at G₁–S boundary, cells were treated with 2 mmol/L of thymidine for 16 hours, released for 8 hours, and treated with 2 mmol/L of thymidine for 16 hours again. After washing with PBS for three times, cells were released into fresh medium for different times and harvested.

The phospho-specific antibodies against TSC1-S578 were generated by Proteintech. In brief, a peptide containing phospho-S578 was synthesized and used to immunize rabbits. After the antibodies were affinity purified, a series of control experiments were performed to confirm its specificity. The antibodies against mTOR (2983), p-mTOR-S2448 (2971), TSC1 (6935), TSC2 (4308), p-TSC2-T1462 (3611), AKT (9272), p-AKT-S473 (4060), p-AKT-T308 (4056), p-S6K-T389 (9205), S6 ribosomal protein (2217), p-S6-S240/244 (2215), p-S6-S235/236 (4858), p-4E-BP-1-S65 (9451), phospho-Histone H3 (9701), histone H3 (9717), and PTEN (9188) used in this study were purchased from Cell Signaling, whereas the antibodies against Plk1 (05-844) were from Millipore. We obtained antibodies against LAMP2 (sc-18822) and ubiquitin (sc-8017) from Santa Cruz Biotechnology. Antibodies against Rheb (MAB3426) and Myc (M5546) were purchased from Novus Biologicals and Sigma, respectively.

All the animal experiments described in this study were approved by the Purdue University Animal Care and Use Committee. HeLa cells were transfected with different myc-TSC1 constructs [wild type (WT), 2A, or 2E] and treated with 2 mg/mL G418 for 1 week to select transfection-positive cells. Then, 1 × 10⁶ selected cells were mixed with an equal volume of Matrigel (Corning) and inoculated into the right flank of a nude mouse (Envigo, 5 mice/group, 4 groups, 20 mice total). One week later, mice were treated with 10 mg/kg rapamycin once a week, and tumor volumes were measured twice a week using the following formula: V = L × W²/2 (V, mm³; L, mm; W, mm).

All data are presented as means ± SD. Statistical calculations were performed with Microsoft Excel analysis tools. Although a two-tailed, unpaired Student t test was used to assess the difference between the effects of treatment in cell lines, one-way ANOVA was used to determine statistically significant differences from the means in the animal study. P values of <0.05 were considered statistically significant. *, P < 0.05; **, P < 0.01.

To elucidate the potential role of the PI3K/AKT/mTOR pathway during mitosis, we treated HeLa cells with nocodazole to induce cell-cycle arrest at prometaphase. Although phosphorylation of AKT was minimal at prometaphase, it gradually increases as cells exit from mitosis and enter the interphase, confirming the work of others (Fig. 1A; ref. 12). Unexpectedly, phosphorylation of S6 (at both S235/S236 and S240/S244) was detected to be very high at prometaphase, but quickly decreased upon mitotic exit (Fig. 1A; Supplementary Fig. S1A). Of note, the finding is not cell type specific, as the similar phenomena was observed in HEK 293T cells as well (Supplementary Fig. S1B). Therefore, our finding suggested an alternative activation of the PI3K/AKT/mTOR pathway in response to growth factors, in which the levels of p-S6 and p-AKT are strongly correlated (Fig. 1B). To further validate our finding, we also synchronized HeLa cells with the double thymidine block protocol to arrest cells at the G₁–S boundary. Consistent with the results based on the release from prometaphase, the p-AKT signal was high in G₁–S phase, but gradually decreased as cells entered G₂–M phase. In contrast, the p-S6 level was minimal in G₁–S phase, but was significantly increased as cells entered G₂ and peaked in mitosis, matching the Plk1 level during cell-cycle progression (Fig. 1C; Supplementary Fig. S1C). To confirm these novel findings, we conducted immunofluorescence studies to assess phosphorylation of S6 at various stages of mitosis in HeLa cells. Consistent with the results of the immunoblot analysis, the level of p-S6 becomes most abundant at mitosis (Fig. 1D and E).

We asked whether mitosis-associated S6 phosphorylation was under control of the PI3K/AKT pathway. Accordingly, asynchronously growing and mitosis-arrested HeLa cells were serum-starved for 12 hours and then subjected to stimulation with insulin for 10 minutes. Preincubation of serum-starved cells with LY294002 for 6 hours completely prevented insulin-induced mTORC1 kinase activity, as measured by reduced phosphorylation of S6K and S6 (Supplementary Fig. S1D). However, neither treatment of cells with starvation nor treatment of cells with insulin had any effect on S6K, S6, or 4E-BP1 phosphorylation in mitotic cells (Fig. 2A; Supplementary Fig. S1E). More interestingly, treatment of mitotic cells with LY294002 failed to block S6K, S6, or 4E-BP1 phosphorylation, indicating that S6K, S6, or 4E-BP1 phosphorylation is independent of PI3K/AKT pathway during mitosis (Fig. 2A; Supplementary Fig. S1E).

To avoid the possibility of side effect of nocodazole, we conducted immunofluorescence studies in randomly growing HeLa cells. LY294002 pretreatment markedly reduced p-S6 signal in interphase cells (Fig. 2B) but not in mitotic cells (Fig. 2B). Interestingly, mitosis-specific phosphorylation of S6 was inhibited by both mTORC1 and p70S6K inhibitors, indicating that mTORC1/p70S6K pathway is responsible for mitotic-specific S6 phosphorylation (Fig. 2B). To substantiate this further, HeLa cells were synchronized at the G₂–M boundary with RO-3306, and then allowed to go through mitosis upon release, with most cells exiting mitosis at 4 hours postrelease. Rapamycin, but not LY294002, pretreatment markedly reduced p-S6 signal (Fig. 2C and D). Finally, we also asked whether amino acid starvation affects the observation we made. As indicated, amino acid–starved cells still responded to nocodazole treatment by increased phosphorylation levels of S6 at S235/6. In addition, mobility shift of 4E-BP1, an indicator of its phosphorylation, was also clearly detected upon nocodazole treatment (Supplementary Fig. S1F).

To further characterize the role of the PI3K/AKT/mTOR signaling during mitosis, we reexamined changes in the activities of AKT and its downstream signaling proteins. Consistent with the data in Fig. 2, hyperphosphorylation of S6 and S6K was observed during mitosis in an insulin-stimulation independent manner (Fig. 3A). Interestingly, under normal growth condition, electrophoretic migration of TSC1 was predominantly detected as a single band (Fig. 3A), whereas a second more slowly migrating band was observed in mitotic cell extracts (Fig. 3A). This observation is consistent with a previous study (13). To test whether TSC1 is indeed hyperphosphorylated in mitosis, we arrested cells at G₁, S, G₂, and M phase. As indicated in Fig. 3B, the mitosis-enriched cells expressed a TSC1 form that migrated apparently slower than TSC1 from interphase cells. Because the change in electrophoretic mobility was completely reversed by treatment of extracts with λ-phosphatase (Fig. 3C), we concluded that TSC1 is hyperphosphorylated during mitosis. To identify kinases that are responsible for TSC1 phosphorylation in mitosis, we compared the TSC1 mobility upon inhibition of several mitotic kinases. Addition of RO-3306 impaired the mobility shift of TSC1, suggesting that TSC1 might be phosphorylated by Cdk1 in mitosis (Fig. 3D). In contrast, treatment with BI2536 (a specific inhibitor of Plk1), VX-680 (a pan-Aurora inhibitor), or SB202190 (an inhibitor of p38 MAP kinase) did not change the mobility of mitotic TSC1 significantly (Fig. 3E; Supplementary Fig. S2A). Consistent with the previous study (13), our data suggest that the majority of TSC1 phosphorylation during mitosis likely depends on Cdk1. Considering the fact that Cdk1-mediated substrate phosphorylation tends to create a docking site for the C-terminal polo-box domain of Plk1, we tested the combination effect of RO-3306 plus BI2536. As indicted, Plk1 inhibition alone slightly reduced mobility shifts of TSC1 and 4E-BP1, but a combination of RO-3306 and BI2536 completely blocked these events (Fig. 3F and G). In support, Plk1 overexpression led to increased levels of p-S6K, p-4E-BP1, p-S6, but not p-AKT and p-TSC2 (Fig. 3H; Supplementary Fig. S2B). Finally, hyperphosphorylation of TSC1 during mitosis is independent of AKT kinase activity (Supplementary Fig. S2C).

It has been documented that Cdk1 tends to function as a priming kinase for subsequent phosphorylation by Plk1 for mitotic proteins (36). Cdk1 phosphorylates TSC1 at three possible sites T417, S584, and T1047 (13). Furthermore, TSC1 interacts with Plk1 in a T310 phosphorylation-dependent manner (35). Therefore, we hypothesized that TSC1 might be a novel PIk1 substrate. Previously, we also identified TSC1 as a potential Plk1 substrate (37). Coimmunoprecipitation assays showed that endogenous TSC1 coimmunoprecipitated with endogenous Plk1 in nocodazole-treated cells (Fig. 4A). Next, we asked whether Plk1 regulates mTORC1 kinase activity. As indicated, depletion of Plk1 reduced the phosphorylation levels of S6K and S6 (Fig. 4B; Supplementary Fig. S2D). These observations confirmed the previously published results (35)_. To directly test whether TSC1 is a Plk1 substrate, in vitro kinase assays were conducted and showed that TSC1 is a direct Plk1 substrate (Fig. 4C). To determine whether TSC1 phosphorylation by Plk1 was modulated by Cdk1, we performed the sequential kinase assay. Preincubation with Cdk1 clearly enhanced the subsequent Plk1-mediated phosphorylation of TSC1 (Fig. 4D). To directly map potential phosphorylation sites of TSC1, various recombinant glutathione S-transferase (GST)-fused TSC1 regions were generated. In vitro kinase assays showed that two regions of TSC1, amino acids (aa) 301-474 and aa 475-660, contain the Plk1 phosphorylation sites (Fig. 4E). Then, virtually every single serine and threonine residues of aa 301-474 and aa 475-660 was mutated into alanine to identify S467 and S578 as two Plk1 phosphorylation sites (Fig. 4F and G). To further characterize the phosphorylation sites we mapped, we also generated a polyclonal antibody that specifically recognizes the phosphorylated form of TSC1 at S578 (Fig. 4H). In vivo, the pS578 antibody detected signals from cells expressing WT TSC1, but not from cells expressing the S578A variant (Fig. 4I). More importantly, the pS578 antibody was able to detect phosphorylation signals on TSC1 from HeLa cells, but the epitope was significantly reduced in cells depleted of Plk1 (Fig. 4J). To determine whether Cdk1 serves as a priming kinase for Plk1 in cells, we depleted Cdk1 using RNAi and analyzed the TSC1-S578 phosphorylation. As indicated in Fig. 4K, the pS578 level was significantly reduced in Cdk1-depleted cells. Finally, overexpression of Plk1-WT form, but not the kinase-deficient mutant (K82M), led to elevation of the pS578-TSC1 epitope (Fig. 4L). Alignment of TSC1 sequences indicates that the two phosphorylation sites are highly conserved across different species (Fig. 4M).

TSC1 and TSC2 are two tumor suppressor genes mutated in tumor syndrome TSC (35), whereas increased Plk1 expression is detected in many types of cancer (19). We hypothesized that Plk1 phosphorylation of TSC1 during mitosis might have an inhibitory effect on the TSC1/TSC2 complex. Compared with cells expressing WT-TSC1, cells expressing the phosphomimetic TSC1-2E mutant clearly had an increased level of p-S6 upon insulin stimulation (Fig. 5A). Next, we cotransfected HeLa cells with different forms of myc-TSC1 with HA-TSC2 and treated with nocodazole. As indicated in Fig. 5B, although the p-S6 signals were decreased during mitosis in both WT- and 2A-TSC1–expressing cells, expression of TSC1-2E failed to do so. To further substantiate this, we synchronized HeLa cells in prometaphase by the mitotic shake-off protocol and then allowed cells to exit mitosis upon release. As expected, compared with cells expressing TSC1-2A, cells expressing TSC1-2E apparently showed a higher p-S6 level during mitotic exit (Fig. 5C; Supplementary Fig. S3A). We hypothesized that phosphorylation of TSC1 might also play a role in regulating its interaction with TSC2. As expected, WT TSC1 and TSC2 formed a stable complex. However, we found that compared with TSC1-WT and -2A mutant, the amount of TSC1-2E mutant bound to TSC2 was apparently reduced (Fig. 5D). To address whether phosphomimetic mutant of TSC1 is less stable than TSC1-WT or -2A mutant, we transfected HeLa cells with different forms of myc-TSC1 constructs, followed by treatment with cycloheximide. As indicated in Fig. 5E, 12-hour cycloheximide treatment did not affect the levels of TSC1-WT and -2A mutant, but significantly reduced the level of TSC1-2E mutant. To ensure our finding is not cell type specific, we repeated the experiment in PC3 cells and obtained a similar result (Supplementary Fig. S3B). A similar observation was made when HEK 293T cells were cotransfected with Myc-TSC1 (WT, 2A, or 2E mutant) with HA-TSC2 (Fig. 5F). To confirm these findings, we further showed that the level of TSC1-2E mutant was markedly increased upon treatment with the proteasome inhibitor MG132 (Fig. 5G). Finally, a much higher level of polyubiquitination was detected in TSC1-2E mutant than that of TSC1-WT and -2A mutant (Fig. 5H).

Because phosphorylation of TSC1 is cell cycle regulated, we asked whether Plk1 dependent phosphorylation of TSC1 affects cell-cycle progression. Interestingly, we found that overexpression of WT TSC1 affects mitotic progression (Fig. 6A). Next, we compared mitotic exit of cells expressing different forms of TSC1 at the Plk1 phosphorylation sites and found that cells expressing different forms of TSC1 (2A or 2E) go through mitosis with different kinetics. As indicated, the level of p-H3 was already fairly low at 2-hour postrelease in cells expressing TSC1-2A, but the level of p-H3 remained high even at 4-hour postrelease in cells expressing TSC1-2E (Fig. 6B). To rule out the possibility that the mitotic delay was caused by nocodazole, we performed time-lapse live cell imaging of HeLa cells stably expressing GFP-H2B. Consistent with the nocodazole release experiments, live cell imaging analysis revealed that cells expressing TSC1-2E showed a prolonged prometaphase associated with a significantly extended average time (63.2 ± 2.4 minutes) from nuclear envelope breakdown to the onset of anaphase compared with those of control cells (41.8 ± 1.5 minutes; Fig. 6C). Experiments were then conducted to determine whether the mitotic defect observed in TSC1-2E–expressing cells was sensitive to mTOR inhibitor rapamycin. Cells expressing TSC1-2E mutant showed slower mitotic exit (Fig. 6B); rapamycin treatment reduced pS6 level (Supplementary Fig. S3C), but did not reverse the TSC1-2E expression-induced mitotic delay (Fig. 6D).

The mTOR signaling plays a role in coordinating cell growth and proliferation signals with protein synthesis in a raptor-dependent manner during mitosis (38). To test whether the Plk1/TSC1/mTORC1 axis also mediates cell growth, we exogenously expressed different forms of myc-TSC1 constructs in HeLa cells. Under normal growth conditions, expression of different TSC1 constructs all led to a decrease in the rate of cell growth. Unexpectedly, the growth rate of TSC1-2E–expressing cells, which showed the constitutively active mTOR pathway (Supplementary Fig. S3D), was similar to those of cells expressing TSC1-WT and -2A (Fig. 6E). Interestingly, at a lower dose of rapamycin, 15 nmol/L, we observed that the overall growth rate was appreciably lower in TSC1-2E–expressing cells than cells expressing TSC1-WT and -2A (Fig. 6F and G). Finally, cells expressing different forms of TSC1 constructs were used to generate xenograft tumors. Consistent to cell culture–based data in Fig. 6, tumors expressing TSC1-2E mutant were more sensitive to rapamycin treatment than tumors derived from cells expressing TSC1-WT or -2A mutant (Fig. 7A and B; Supplementary Fig. S4A and S4B). Further IHC analysis indicates that tumors expressing the TSC1-2E mutant had the lowest proliferation rate, indicated by Ki67 staining (Supplementary Fig. S4C). Interestingly, both tumors expressing TSC-2A and -2E mutants had high levels of cleaved caspase-3 signals (Supplementary Fig. S4D). Thus, the small tumor size of group expressing TSC1-2E is due to both reduced proliferation and increased apoptosis, whereas the small tumor size of group expressing TSC1-2A is mainly due to increased cell death. We also measured the baseline of apoptosis in an in vitro experiment in the presence and absence of rapamycin in cells expressing different forms of TSC1 (WT, 2A, 2E). Consistent with the general concept that rapamycin usually does not induce apoptosis, we failed to detect apparent apoptosis in cultured cells (Supplementary Fig. S4E). Apparently, the tumors expressing TSC1-2E respond differently to rapamycin, and more experimentation is needed to clarify the issue.

We then asked whether cells expressing different forms of TSC1 have different mTORC1 activities. As indicated, HeLa cells expressing TSC1-2E mutant indeed had a higher phosphorylation level of S6K than cells expressing the 2A mutant (Fig. 7C). To ensure that our finding is not HeLa cell specific, we repeated the experiment in PC3 cells and obtained similar results (Fig. 7D; Supplementary Fig. S3D). It was shown that AKT-mediated phosphorylation of TSC2 results in dissociation of the TSC complex from the lysosome in response to insulin stimulation (6). As expected, treatment of cells with AKT inhibitor MK2206 prevented the dissociation of the TSC complex from the lysosome (Fig. 7E). Interestingly, treatment of cells with Plk1 inhibitor BI2536 also reduced cytosolic localization, but increased lysosomal localization, of the TSC complex (Fig. 7E). To further dissect how Plk1 phosphorylation of TSC1 affects the mTORC1 activation at the lysosome, we followed colocalization of different members of the mTOR pathway upon expression of different forms of TSC1 (WT vs. 2A). mTOR was colocalized with its positive regulator Rheb at lysosome in cells expressing TSC1-WT, but expression of TSC1-2A mutant apparently reduced the colocalization of Rheb with mTORC1 at lysosome (Fig. 7F and G). In striking contrast, colocalization of TSC2 with LAMP2, was increased in cells expressing TSC1-2A mutant in comparison with cells expressing TSC1-WT (Fig. 7H and I). Altogether, we conclude that Plk1 phosphorylation of TSC1 leads to increased dissociation of the TSC complex from the lysosome.

PI3K, AKT, and mTOR are three hubs in the PI3K pathway to regulate cell growth and proliferation (39). Despite the fact that all the inhibitors demonstrated very promising preclinical activities, none of them is currently able to cure a single cancer patient, largely due to the development of drug resistance. Unfortunately, the underlying molecular mechanism of resistance to these inhibitors remains largely unknown (40, 41). In case of mechanisms responsible for resistance to mTOR inhibitors, one potential mechanism is due to the fact that activation of oncogenic mTOR pathway can be achieved in diverse ways (41, 42). Consistent with this idea, our data clearly showed that the PI3K/AKT pathway, the well-established upstream activating pathway of mTORC1, is downregulated during mitosis, whereas mTORC1 is clearly active. To probe the mechanism, we found that mitotic TSC1 is highly phosphorylated. Although it is well documented that AKT phosphorylation of TSC2 leads to inactivation of the TSC1/2 complex, kinase(s) response for and functional significance of TSC1 phosphorylation have not been described. Our observations confirmed the previous study that TSC1 is phosphorylated by Cdk1 and the phosphorylated TSC1 interacts with Plk1 (13, 35). Furthermore, we indicate that Plk1 cooperates with Cdk1 to generate the mitotic-specific hyperphosphorylated TSC1, providing a novel mechanism for the regulation of mTORC1 activity during mitosis. In this model, the TSC complex functions as a downstream target of Plk1. We showed that Plk1-mediated phosphorylation of TSC1 causes disruption of TSC complex formation and destabilization of TSC1 protein, eventually resulting in the activation of mTORC1 in mitosis. Interestingly, it was recently shown that Plk1 protein levels are increased in TSC1^−/− and TSC2^−/− cells (43). When combined with what we described here, these data imply that a feedback mechanism might exist between Plk1 and TSC1.

It was shown that TSC1 physically and functionally interacts with Plk1 to regulate centrosome biology and mitotic progression (35). We hypothesized that Plk1 phosphorylation of TSC1 in the M phase also regulates mitotic progression. Supporting this notion, we found that the expression of phosphomimetic TSC1 mutant impairs normal mitotic progression. We then tested the hypothesis that impaired mitotic progression observed in TSC1-2E–expressing cells might be mTOR dependent. Unexpectedly, pretreatment of cells expressing TSC1-2E with rapamycin had no impact on mitotic progression. These results indicate that Plk1 phosphorylation of TSC1 affects its function through at least two mechanisms: First, phosphorylation of TSC1 impairs normal mitotic progression; second, phosphorylation of TSC1 activates the mTORC1 pathway. Indeed, cells expressing TSC1-2E showed a slower growth in the presence of rapamycin, supporting the notion that Plk1 phosphorylation of TSC1 plays a central role in the coordination of cell-cycle progression and cell growth. Plk1, a serine/threonine protein kinase that is normally expressed in mitosis, is frequently overexpressed in multiple types of human tumors regardless of the cell-cycle stage. However, the causal relationship between overexpression of Plk1 and tumorigenesis has not been fully investigated. In this study, we provide evidence that Plk1-mediated phosphorylation of TSC1 leads to aberrant mitosis and hyperactivation of mTOR pathway, both of which are hallmarks of cancer.

A previous study reported that shRNA-mediated Plk1 depletion decreased the phosphorylation of mTOR substrates (44). It was also reported that Plk1 is a physical mTORC1 interactor and is involved in autophagy pathway (45). However, the underlying mechanism how Plk1 controls the mTORC1 pathway is still unclear. Here, we report a novel signaling pathway where Plk1 regulates mTOR independently of AKT in mitosis. The data support the signaling findings in both cancer and normal cell lines used. We acknowledge that the effects of the TSC1 mutants in proliferation and tumor growth appear puzzling. We are also surprised that the phosphorylation mimic of the TSC1 phosphorylation site delays mitosis but has no effect on cell proliferation when compared with WT TSC1 or alanine mutant. This could be due to the different culture conditions. For the mitotic exit experiments, we used shake-off protocol to synchronize all cells in prometaphase. The delayed mitosis exit observed in the phosphorylation mimic mutant could be due to the defected recovery from the nocodazole arrest. The delayed mitotic exit could also be the potential mechanism of unexpected observation, rapamycin-induced apoptosis in TSC1-2E cells. Rapamycin almost never induces apoptosis, both in vitro and in vivo. It is very surprising that TSC1-2A and TSC1-2E tumors treated with rapamycin have increased apoptosis. Thus, it is possible that rapamycin-induced apoptosis under this condition is due to defects in mitosis or due to aberrant activation of prosurvival pathways. The exact mechanism deserves further experimentation.

In summary, we propose that mTOR activity is controlled by two different pathways during cell cycle. In interphase, the PI3K/AKT pathway plays a key role to activate the mTOR pathway by AKT-mediated phosphorylation of TSC2 in response to intracellular signaling. However, in mitosis, Plk1 is the major kinase to activate the mTOR pathway by targeting TSC1. Instead of activation of the mTOR pathway, Plk1 phosphorylation of TSC1 also leads to mitotic defects in an mTOR-independent manner.

No potential conflicts of interest were disclosed.

Conception and design: Z. Li, X. Hou, X. Liu

Development of methodology: Z. Li, X. Hou

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Z. Li, Y. Kong, J. Liu, X. Hou

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Z. Li, Y. Kong, Q. Luo, X. Liu

Writing, review, and/or revision of the manuscript: Z. Li, X. Liu

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Z. Li, Y. Kong, L. Song, Q. Luo, C. Shao, X. Liu

Study supervision: Z. Li, X. Liu

This work was supported by NIH grants R01 CA157429 (to X. Liu), R01 CA192894 (to X. Liu), R01 CA196835 (to X. Liu), R01 CA196634 (to X. Liu), and P30 CA023168 (Purdue Center for Cancer Research).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.