A New Pharmacologic Action of CCI-779 Involves FKBP12-Independent Inhibition of mTOR Kinase Activity and Profound Repression of Global Protein Synthesis

CCI-779 (temsirolimus) is an anticancer drug recently approved by Food and Drug Administration for the treatment of advanced renal cell carcinoma. Multiple clinical and preclinical studies have shown promising antitumor activity of CCI-779 in other tumor types, including breast cancer (1), glioma (2), endometrial cancer (3), and mantle cell lymphoma (4). The anticancer activity of CCI-779 likely involves multiple pharmacologic actions, including its antiangiogenic and antiproliferative properties (5, 6). The molecular mechanism of CCI-779 antitumor effect involves inhibition of mammalian target of rapamycin (mTOR) kinase function. mTOR is an evolutionary conserved serine/threonine kinase, the founding member of the phosphoinositide-3-kinase (PI3K)–related kinase family. In mammals, mTOR kinase acts as a key metabolic sensor that coordinates cell growth and proliferation with the availability of growth factors, nutrients, and intracellular ATP (7). Rapamycin or its analogues form tight binary complex with the small immunophilin FKBP12 (FK506 binding protein), which is highly abundant in human cells. The rapamycin-FKBP12 complex binds mTOR in a unique region called FKBP12-rapamycin binding domain (FRB; ref. 8). Accumulating evidence points to the potential regulatory role for the FRB region in ligand-mediated mTOR activity and/or mTOR signal transduction in cells. Although the exact mechanism by which rapamycin inhibits mTOR catalytic activity remains unclear, earlier findings suggest that the rapamycin-FKBP12 complex acutely destabilizes mTOR complex 1 and, in certain cell types, interferes with the assembly of the mTOR complex 2 (9–11). Rapamycin might also compete with phosphatidic acid for binding to FRB, thereby interfering with activation of mTOR by phosphatidic acid (12, 13).

In vitro, CCI-779 and rapamycin at the commonly used nanomolar dose range effectively suppress the cellular phosphorylation of the mTOR substrates S6K1 and 4E-BP1. However, the growth inhibitory effects at these same drug concentrations are generally modest and selective against certain tumor cell lines, including brain, breast, prostate, rhabdomyosarcoma cells. Recent clinical studies in patients with various solid malignancies showed that concentrations of CCI-779 in whole blood generally reach micromolar levels (1, 14–16). Although most of the in vitro studies that investigate CCI-779 and/or rapamycin action in cancer cells use nanomolar drug concentrations, they do not address the inhibitory response and underlying cellular mechanism(s) to the drug at the clinically relevant micromolar doses.

In this report, we have identified a distinct pharmacologic mode of action of the rapamycin and its analogues. We show that a broad and profound suppression of tumor cell growth can be achieved by the clinically relevant low micromolar concentrations of CCI-779, a phenomenon we term as the “high-dose effect.” This enhanced antitumor activity is mechanistically linked to a previously unreported FKBP12-independent suppression of mTOR kinase activity and mTOR signaling in cancer cells. We show that the micromolar but not the nanomolar concentrations of CCI-779 uniquely affect the function of the cellular translation machinery, leading to a marked decline in global protein synthesis, and profound inhibition of cell growth. This high-dose drug effect could be highly relevant to the treatment of human cancer patients with CCI-779 and other rapamycin analogues.

Chemicals. All general chemicals used for buffers and assays were obtained from Sigma-Aldrich unless otherwise specified. mTOR inhibitors CCI-779, rapamycin, and the rapamycin derivative WAY-179639 were provided by Wyeth Chemical and Pharmaceutical Development. FK506 was obtained from EMD Calbiochem. All inhibitors were dissolved in DMSO as concentrated stocks and diluted before use in DMSO for in vitro enzyme or binding assays or diluted in culture media for cell-based assays.

Cell culture, growth assays, and transfections. Human lung tumor lines (A549, H157, H460, H446), colon tumor lines (HCT116, HT29, SW480, DLD1, Caco2), prostate tumor lines (LNCap, DU145), breast lines (MDA468, MDA231), and the human embryonic kidney cell line HEK293 were obtained from American Type Culture Collection. PC3-MM2 were obtained from Dr. C. Discafani of Wyeth Oncology. Cells were cultured in a 37°C incubator with 5% CO₂ using standard cell culture methods. All culture media, supplements, and transfection reagents were purchased from Invitrogen. For growth assays, cells were plated in 96-well culture plates at 1 × to 3 × 10³ cells per well. After 24 h, cells were treated with various doses of inhibitors. Three days after drug treatment, viable cell densities were determined by MTS dye conversion using CellTiter AQ assay kit (Promega). To create an mTOR-SI cell line, the empty vector pIRESneo3 (Clontech) or pIRESneo3-mTOR-SI (S²⁰³⁵I, created by site-directed mutagenesis) were transfected into HEK293 cells using the Lipofectamine 2000 reagent and selected by G418. The effect of each treatment was calculated as percentage of control growth relative to the vehicle DMSO-treated cells grown in the same culture plate. For small interfering RNA (siRNA) knockdown experiments, HCT116 cells were seeded at 30% confluence on 24-well culture plates and transfected with 60 nmol/L siRNA using Lipofectamine RNAiMax. siRNA transfections were repeated 48 h later, and the cells were then incubated for additional 24 h, followed by the drug treatment. The eEF2K siRNA pool and nontargeting siRNA pool were obtained from Dharmacon.

Protein lysates and immunoblotting. For all experiments, cells were plated in six-well plates in complete growth media for overnight. Various doses of CCI-779 or rapamycin analogues were added alone or with FK506 for indicated periods of time. Total cellular lysates were prepared using NuPAGE-LDS sample buffer (Invitrogen). Equal amounts of proteins were subjected to immunoblotting analysis using NuPAGE electrophoresis system (Invitrogen). Antibodies for phosphorylated S6K1 (Thr³⁸⁹), total S6K1, phosphorylated 4E-BP1 (Thr⁷⁰ and Thr⁴⁶), phosphorylated AKT (Ser⁴⁷³), total AKT, phosphorylated GSK-3 (Ser^9/21), phosphorylated mTOR (Ser²⁴⁸¹), phosphorylated eEF2 (Thr⁵⁶), total eEF2, phosphorylated eIF2α (Ser⁵¹), phosphorylated eEF2K (Ser³⁶⁶), and cyclin D3 were obtained from Cell Signaling Technology. Antibodies for total GSK-3 and activating transcription factor 4 (ATF-4; H-290) were obtained from Santa Cruz Biotechnology. The p21/Cip1 antibody was from BD Transduction Laboratories and the β-actin antibody was from Chemicon-Millipore. Immunoblots were probed with appropriate primary and secondary antibodies following manufacturer's instructions and detected using enhanced chemiluminescence (GE Healthcare). β-Actin immunoblotting and/or reversible Ponceau-S staining of nitrocellulose membranes after the transfer (17, 18) were used to visualize the total protein loading.

Production and binding assays of FKBP12 and FRB. The human FKBP12 cDNA was obtained by PCR amplification from human placenta Quick Clone cDNA (Clontech) and was inserted into the XhoI/HindIII sites of the Escherichia coli vector ptrcHis (Invitrogen). The FRB domain of human mTOR (amino acids 2015–2114) was generated by PCR from the mTOR cDNA (19) and inserted into the BamHI/EcoRI sites of the E. coli vector pGEX2T (GE Healthcare). His6-FKBP12 and glutathione S-transferase–FRB were produced and purified according to the respective vendor's instructions. FKBP12 binding assay was done in the 96-well format nickel chelate flashplate with [³H]FK506 (NEN-PerkinElmer; Supplementary Materials and Methods). FRB binding by the rapamycin analogues was assessed in an FKBP12-rapamycin-FRB ternary complex formation assay in the 96-well MaxiSorp Fluoro plates (Nunc; Supplementary Materials and Methods).

In vitro assay of mTOR catalytic activity. The Flag-tagged wild-type human mTOR (Flag-mTOR) and Flag-mTOR-SI (S²⁰³⁵I, created by site-directed mutagenesis) DNA constructs were transiently transfected into HEK293 cells. Protein extraction and purification of Flag-mTOR and Flag-mTOR-SI were carried out 48 h later as described previously (19). In vitro kinase assays of purified Flag-mTOR and Flag-mTOR-SI without or with FKBP12 were done and detected by dissociation-enhanced lanthanide fluorescent immunoassay (DELFIA) as described (19).

Analysis of global translation. Global protein synthesis was measured by determining incorporation of [³H]methionine into TCA-insoluble cellular components (Supplementary Materials and Methods). Analysis of polysome profiles was done as described previously (20) with minor modifications (Supplementary Materials and Methods).

CCI-779 at low micromolar concentrations broadly suppresses tumor cell proliferation via an FKBP12-independent mechanism. During our studies on tumor cell proliferation in culture, we noted that CCI-779 displayed a biphasic dose response in a panel of human tumor cell lines (Fig. 1A). In the commonly used dose range of 10 nmol/L to <5 μmol/L, the drug exhibited a generally flat and cell-selective inhibition of cell proliferation. The most sensitive cell lines that were inhibited by ≥50% included LNCap, PC3MM2, MDA468, H446, and Caco2, whereas those with ≤25% inhibition were SW480, HT29, HCT116, and H460. Interestingly, at higher dose range of 5 to 15 μmol/L, the drug manifested a more profound antiproliferative activity in nearly all cell lines (Fig. 1A). A similar biphasic inhibition pattern was observed for rapamycin (data not shown). These observations raised the possibility that at the low micromolar doses, CCI-779 may exert a previously uncharacterized inhibition mechanism. Because rapamycin in complex with FKBP12 inhibits mTOR with high affinity (∼3 nmol/L; ref. 21), we first performed FK506 competition experiments to determine whether the second phase inhibition (called micromolar inhibition) similarly requires FKBP12. CCI-779 was titrated in LNCap and MDA468 cell proliferation assay, with or without 0.3, 3, and 10 μmol/L FK506, to completely inhibit the binding of CCI-779 to FKBP12 (Fig. 1B). In both cell lines, antiproliferative effect of CCI-779 at concentrations of ≤10 μmol/L was largely reversed by FK506, indicating that the interaction with FKBP12 was essential for low-dose inhibition of cell growth by CCI-779. However, the high-dose growth inhibitory effect of the drug was completely resistant to cotreatment with 10 μmol/L FK506 in both LNCap and MDA468 cells. These data indicate that the biphasic inhibition of cell proliferation elicited by CCI-779 reflects both FKBP12-dependent and FKBP12-independent actions of this drug at low versus high concentrations respectively.

The growth inhibitory action of rapamycin at nanomolar concentrations is well known to be largely cytostatic. To determine whether the micromolar CCI-779 also elicits cytostasis, we measured viability of several representative cell lines after 24 and 48 h of exposure to 20 nmol/L or 20 μmol/L CCI-779. Consistent with the previous finding, no cell death was detected in 20 nmol/L CCI-779–treated cells. In cells treated with 20 μmol/L CCI-779, cell death for HCT116, HT29, A549, and HEK293 was minimal at both time points. Interestingly, both PTEN-negative cell lines LNCap and MDA468 exhibited a small but significant reduction in viability as early as 24 h (Supplementary Table S1). To further examine whether these cells undergo apoptosis, we performed immunoblotting for PARP cleavage as well as caspase activity assays. CCI-779 (20 μmol/L) caused PARP cleavage in LNCap and MDA468 cells but not in HT29 and A549 cells (Supplementary Fig. S1A). The caspase-3/caspase-7 activity was also increased in LNCap (4.9-fold) and MDA468 (6.6-fold) cells (Supplementary Fig. S1B). Thus, micromolar CCI-779 seems to inhibit proliferation via a largely cytostatic mechanism in most of the cells tested, whereas it induces apoptosis in the PTEN-negative tumor cells.

Micromolar inhibition of proliferation correlates with FKBP12-independent inhibition of cellular mTOR signaling. To define the mechanism responsible for high-dose growth inhibition by CCI-779, we initially examined the effect of CCI-779 on cell proliferation and mTOR signaling at nanomolar versus micromolar drug concentrations. These assays were done in the absence or presence of FK506, which competes with CCI-779 for binding to intracelullular FKBP12. At low concentrations (e.g., 20 nmol/L), CCI-779 exhibited an FK506-reversible inhibition of growth (Fig. 2A), which correlated with the reversible dephosphorylation of mTOR substrates S6K1 (Thr³⁸⁹) and 4E-BP1 (Thr⁷⁰; Fig. 2B). In contrast, the inhibitory effects of high-dose CCI-779 (10–20 μmol/L) on growth and mTOR substrate phosphorylation were not reversed by FK506, suggesting that high-dose CCI-779 suppresses both mTOR signaling and cell growth through an FKBP12-independent mechanism. The mTOR complex 2 is insensitive to low concentrations of rapamycin and its analogues (22, 23). However, the effects of higher drug concentrations on this second mTOR complex are unclear. Consequently, we examined the phosphorylation state of the known mTORC2 substrate AKT and its downstream target GSK-3 in CCI-779–treated cells. At concentrations of 10 and 20 μmol/L, CCI-779 dramatically inhibited phosphorylation of Ser⁴⁷³ of AKT and Ser^21/9 of GSK-3 with maximal suppression of both responses observed after 24 h of drug exposure (Fig. 2C). Partial inhibition of AKT and GSK-3 phosphorylation was also observed after 24 h of cellular treatment with a low concentration (20 nmol/L) of CCI-779, and this CCI-779 effect was fully reversed by cotreatment with FK506. The decreases in AKT and GSK-3 phosphorylation after long-term CCI-779 exposure were consistent with previous findings, and the blockade of the CCI-779 effects by FK506 indicates that the low-dose effect of CCI-779 was fully dependent on FKBP12. In contrast, FK506 cotreatment failed to reverse the inhibitory effects of 10 or 20 μmol/L CCI-779. In fact, concomitant FK506 exposure noticeably augmented the acute (6 h) effects of CCI-779 on AKT and GSK-3 phosphorylation. Although the actual mechanism underlying this enhancing effect is unknown, we speculate that FK506 may saturate irrelevant binding sites (including those exposed by FKBP family members), thereby increasing the pool of free intracellular CCI-779.

CCI-779/rapamycin micromolar action requires binding to the FRB domain of mTOR. Structural information regarding the interactions between rapamycin, FKBP12, and the FRB domain of mTOR has supported rational approaches to the generation of rapamycin derivatives with varying affinities for FKBP12 or FRB. We used select rapamycin analogues to further investigate the requirements for FKBP12 and/or mTOR binding in the dose-dependent effects of CCI-779 on cell proliferation and signaling functions. WAY-179639, a triene-perturbed derivative, retained potent FKBP12 binding (EC₅₀, 8 nmol/L) but was inactive in an FKBP12-rapamycin-FRB ternary complex formation assay (EC₅₀, >5,000 nmol/L; Fig. 3A). Interestingly, this compound was not growth inhibitory against LNCap, HEK293 (Fig. 3B), and other cells (data not shown) at the doses up to 30 μmol/L. The observed lack of activity in proliferation assays is presumably due to the lack of inhibition of cellular mTOR signaling as evidenced by the normal levels of S6K1 phosphorylation (Fig. 3C). To ascertain whether WAY-179639 actually penetrated cells, we used it as a competitor against CCI-779 in a growth inhibition assay. Cotreatment with WAY-179639 overcomes the growth inhibitory effect of CCI-779 at concentrations of ≤10 μmol/L, consistent with the ability of WAY-179639 to penetrate cells and to compete with CCI-779 for binding to FKBP12 (Fig. 3D). These results indicate that the high-dose inhibition of cell growth by CCI-779 correlates with its ability to bind the mTOR FRB domain.

Direct inhibition of mTOR enzymatic activity by CCI-779 and rapamycin in the absence of FKBP12. To obtain biochemical evidence of inhibition of mTOR directly by CCI-779 and rapamycin, we performed in vitro kinase assays with purified, recombinant Flag-mTOR in the presence of various concentrations of CCI-779 without FKBP12. Assay products were first analyzed by immunoblot analysis with phosphorylation-specific antibodies. Phosphorylation of both mTOR substrates, His6-S6K1 (Thr³⁸⁹) and His6-4E-BP1 (Thr⁴⁶), tested in these assays was inhibited by CCI-779 at concentrations of ≥800 nmol/L (Fig. 4A). Direct inhibition of mTOR kinase activity was quantified with a DELFIA-based mTOR kinase assay using His6-S6K1 as the substrate (19). Pooled data from three experiments yielded mean IC₅₀ values of 1.76 ± 0.15 and 1.74 ± 0.34 μmol/L for CCI-779 and rapamycin, respectively (Fig. 4B). Our results show that both CCI-779 and rapamycin directly inhibit mTOR catalytic activity in the absence of FKBP12. Furthermore, the IC₅₀ values for rapamycin obtained in our DELFIA assay were well correlated with those reported in a recent study in which mTOR activity was measured in an immune complex kinase assay (24).

FRB mutant mTOR-SI confers resistance to nanomolar and micromolar CCI-779 inhibitions in vitro and in vivo. Rapamycin-resistant mTOR alleles can be created by substitution of a conserved residue Ser²⁰³⁵ in the FRB domain with a more bulky amino acid, such as isoleucine (21, 25). To assess whether CCI-779 shows a similar dependence on Ser²⁰³⁵ for binding to the FRB domain in the absence of FKBP12, we examined the sensitivity of an mTOR (Ser²⁰³⁵→Ile) mutant, termed mTOR-SI, to CCI-779 in the presence or absence of FKBP12. Wild-type Flag-mTOR was efficiently inhibited by the FKBP12-rapamycin complex (IC₅₀, 11 ± 1 nmol/L); however, mTOR-SI was not inhibited by the immunophilin-drug complex at concentrations of up to 160 nmol/L (Fig. 5A,, top). In the parallel assays without FKBP12, inhibition of mTOR-SI was only observed at drug concentrations considerably higher than those required for inhibition of wild-type mTOR. The mean IC₅₀ value of 20 ± 3.2 μmol/L (Fig. 5A,, bottom) for CCI-779 was ∼11-fold higher than that measured with the wild-type mTOR enzyme (Fig. 4B). Thus, the Ser²⁰³⁵→Ile mutation in the FRB domain also confers a substantial resistance to CCI-779 in the absence of FKBP12.

To determine whether these results could be extended to intact cells, we generated HEK293 cell–derived stable cell lines expressing either the control vector or mTOR-SI allele. Phosphorylation of S6K1 at Thr³⁸⁹ and AKT at Ser⁴⁷³ and autophosphorylation of mTOR itself at Ser²⁴⁸¹ was restored in mTOR-SI cells acutely (6 h) treated with 0.02 μmol/L CCI-779, consistent with previously published studies (Fig. 5B; refs. 26, 27). Importantly, compared with vector control, mTOR-SI cells were also protected against the inhibition of above markers by 10 μmol/L and, to a lesser extent, by 20 μmol/L CCI-779 (Fig. 5B). We further investigated whether such restoration of mTOR signaling is sufficient to rescue long-term (3 days) growth suppression induced by various concentrations of CCI-779. As expected, although the vector control cells showed 55% reduction in growth in response to 0.05 μmol/L CCI-779, growth of mTOR-SI cells was inhibited only minimally (Fig. 5C). Interestingly, although 15 and 20 μmol/L drug equally inhibited the growth of both vector and mTOR-SI cells, mTOR-SI cells remained significantly more resistant to 1 and 10 μmol/L CCI-779 (Fig. 5C). The failure of mTOR-SI cells to exhibit significant protection against higher doses of CCI-779 is likely due, at least in part, to the fact that the mTORC1 complexes in these cells contain both wild-type mTOR (rapamycin-sensitive) and mTOR-SI mutant (rapamycin-resistant) proteins. Collectively, our data support the hypothesis that the micromolar effects of CCI-779/rapamycin in the cells are specifically mediated through mTOR and require the binding of the drug to FRB.

Micromolar CCI-779 profoundly inhibits cellular translation. Treatment of sensitive cell types with low concentrations of rapamycin provokes the accumulation of cells in G₁ phase of cell cycle, in part due to the translational suppression of a subset of proliferation-related mRNAs, including the c-Myc and cyclin D1 transcripts (28, 29). The considerably more profound inhibition of cellular proliferation induced by high-dose CCI-779 treatment prompted more detailed analysis of the effects of micromolar CCI-779 on global protein synthesis. We first measured serum-stimulated incorporation of radiolabeled methionine into newly synthesized proteins in LNCap, HCT116, and HEK293 cells after acute treatment with 20 nmol/L and 20 μmol/L CCI-779. In cell populations treated with 20 nmol/L CCI-779, no significant inhibition of protein synthesis was observed in any of the tested populations (Fig. 6A and Supplementary Fig. S2A). In contrast, treatment with 20 μmol/L CCI-779 strongly inhibited global translational activity to a level comparable with that observed with a benchmark protein synthesis inhibitor, cycloheximide. Such profound suppression of global protein synthesis by micromolar CCI-779 is unlikely due to a secondary response of the cells to the growth delay, as an unrelated control drug Taxol, a potent cell growth inhibitor, was not active in this assay. To extend these findings, we analyzed polysome profiles of HEK293 and HCT116 cells treated with 0.02 or 20 μmol/L CCI-779 (Fig. 6B and Supplementary Fig. S2B). These studies revealed that 0.02 μmol/L CCI-779 caused a significant decrease in the ratio of polysomes to monosomes, an effect consistent with previously reported for rapamycin (30, 31). However, exposure to 20 μmol/L CCI-779 caused much more profound reduction in polysomes, accompanied by a strong increase in the 80S monosomal peak. Thus, the dramatic decrease in translational activity induced by 20 μmol/L CCI-779 is likely related to the disassembly of actively translating polysomes into monosomes.

Modulation of phosporylation of multiple translation factors. To shed light on the mechanism underlying the translational arrest provoked by high-dose CCI-779, we investigated phosphorylation states of several translation factors in HEK293 cells treated with 0.02, 10, or 20 μmol/L CCI-779, with or without 10 μmol/L FK506, to competitively inhibit the binding of CCI-779 to FKBP12. Consistent with the profound suppression of translational activity (Fig. 6A and B), we observed striking increase in inhibitory phosphorylations of eukaryotic translation initiation factor eIF2α on Ser⁵¹ and eukaryotic translation elongation factor eEF2 on Thr⁵⁶ (Fig. 6C). Both of these alterations occurred within 1-hour exposure to 10 to 20 μmol/L CCI-779 and were not reversed by concomitant exposure to FK506 (Fig. 6C). Similar results were obtained in HCT116 colon carcinoma cells exposed to 10 to 20 μmol/L CCI-779 (Supplementary Fig. S2C, left). It is also noteworthy that micromolar CCI-779 induced accumulation of the transcription factor ATF-4 (Fig. 6C , left), which might be secondary to the phosphorylation of eIF2α.

The nucleotide exchange activity of eIF2 complex is negatively regulated by phosphorylation of eIF2α subunit on Ser⁵¹ in response to numerous growth inhibitory signals (32), although a role for mTOR in regulating eIF2α in mammalian cells has not been reported. The phosphorylation of key inhibitory residue Thr⁵⁶ in eEF2 is attributed to the activation of eEF2K, which itself is regulated by an inhibitory phosphorylation event on Ser³⁶⁶. The mTORC1-stimulated protein kinase S6K1 has been identified as one of the upstream effectors of eEF2K phosphorylation at Ser³⁶⁶ (33). Nevertheless, the eEF2 phosphorylation elicited by micromolar CCI-779 seemed to involve an S6K1-independent mechanism because it did not correlate with S6K1 activity or Ser³⁶⁶ phosphorylation of eEF2K (Fig. 6C , right). To further explore the role of eEF2K in eEF2 phosphorylation, we depleted this protein from HCT116 cells using specific siRNAs. Knockdown of eEF2K completely abolished the eEF2 (Thr⁵⁶) phosphorylation induced by high-dose CCI-779 (Supplementary Fig. S2C, right), indicating that eEF2K was fully responsible for the dramatic increase in eEF2 (Thr⁵⁶) phosphorylation seen in these cells. A similar result was observed in U87MG glioma cells (data not shown). The exact mechanism whereby high-dose CCI-779 activates eEF2K remains to be elucidated; regardless, the combined effects of this drug on eIF2α and eEF2 phosphorylation suggest that high-dose CCI-779 interferes with cellular protein synthesis by blocking both the initiation and elongation steps of translation.

CCI-779 blocks growth of different types of tumors, in part, by inducing G₁ arrest and inhibiting proliferation of tumor cells. Although some malignancies respond well to CCI-779 treatment, others exhibit intrinsic or acquired resistance to the drug both in preclinical or clinical settings. In this report, we showed for the first time that increasing concentrations of CCI-779 to 10 to 20 μmol/L not only led to a sharp decrease in proliferation of sensitive cells, but also overcame the resistance of many tumor cell lines to the drug, such as colon cancer cells HCT116 or HT29. Similar doses of rapamycin were used in several reports; however, this phenomenon has been neither noted nor studied previously (34–36). We showed that the profound antiproliferative effect of micromolar CCI-779 is largely attributed to its cytostatic action on most of the tested cells; however, it also elicits apoptotic response in some PTEN-negative tumor cells.

Whereas the mechanisms of intrinsic resistance of tumor cells to the nanomolar CCI-779 and rapamycin analogues in general are not fully defined, deregulation of several signaling pathways is known to contribute to the sensitivity of cancer cells to rapamycin and its analogues in vitro and in vivo. For example, breast, prostate, or renal cancer cell lines that display a loss of PTEN or VHL tumor suppressors or an activation of AKT signaling are more sensitive to CCI-779 inhibition (37–39). Response of tumor cells to the commonly used nanomolar concentrations of rapamycin may be limited by the feedback loop, whereas suppression of S6K1 signaling by rapamycin stimulates IRS-PI3K pathway to promote AKT activity. Herein, we found that micromolar CCI-779 dramatically reduced mTORC1 and mTORC2 signaling, as evidenced by the decrease in the phosphorylation of Thr³⁸⁹ of S6K1, Thr⁷⁰ of 4E-BP1, as well as phosphorylation of Ser⁴⁷³ of AKT and Ser^9/21 of GSK-3, which may in part contribute to the uniform inhibition of proliferation in our cell line panel by micromolar CCI-779. It is also interesting to speculate that micromolar CCI-779 may disrupt the function of both mTORC1 and mTORC2 complexes in cancer cells in a more efficient manner than nanomolar compound. Further studies will be needed to address this hypothesis.

We found that the profound inhibition of mTOR signaling and cell proliferation elicited by micromolar CCI-779 did not require FKBP12, suggesting that FKBP12 independent binding of CCI-779 to the FRB domain of mTOR is required for the observed phenotypes. This view is further supported by the observation that an FRB-binding defective rapamycin analogue failed to display any growth inhibitory properties, even at micromolar concentrations. These results implicate the rapamycin triene interaction with FRB as necessary for both the nanomolar and micromolar dose activity of the drug. Additional findings helped to explain the mechanism of micromolar CCI-779 action at the biochemical level. First, CCI-779 was able to inhibit mTOR kinase activity at micromolar doses directly in the absence of FKBP12, with an IC₅₀ value of 1.76 μmol/L. Second, the kinase activity of the FRB mutant mTOR-SI was 11-fold more resistant to the micromolar CCI-779 compared with the wild-type mTOR. mTOR-SI also conferred a partial protection to mTOR signaling and cell growth in response to micromolar doses of CCI-779. The lack of significant resistance of mTOR-SI cells to the growth inhibitory effect of the drug at the highest concentrations (≥15 μmol/L) can be explained by the fact that (a) endogenously expressed wild-type mTOR allele renders at least a subset of mTORC1 complexes sensitive to CCI-779, (b) mTOR-SI allele seems to be hypomorphic with respect to reconstitution of mTORC1 signaling functions (25), and (c) the mTOR-SI mutant enzyme is partially inhibited by 10 to 20 μmol/L CCI-779 in vitro (Fig. 5A). Together, our current findings are in accord with recent reports that rapamycin is able to bind FRB domain of mTOR directly, without FKBP12, and our IC₅₀ value for CCI-779 agrees well with K_d value reported in these studies (12, 24, 40). These reports, however, do not address biological effects of direct interactions between rapamycin and FRB on mTOR signaling in the cells.

mTOR kinase coordinates protein synthesis and abundance of the various components of the translational machinery with the continuous changes in nutrient, mitogen, or energy status of the cell. Despite numerous studies implicating mTOR in control of general translational factors, the effects of nanomolar rapamycin on global protein synthesis rates are subtle and cell selective. A marked reduction in general protein synthesis in response to micromolar CCI-779, as indicated by both methionine incorporation and polysome profiles data, suggests a potential mechanism for the suppression of cell proliferation at elevated concentrations of the drug. In our studies, micromolar CCI-779 dramatically increased the phosphorylation of the translation initiation factor eIF2α on Ser⁵¹ and elongation factor eEF2 on Thr⁵⁶ in an FKBP12-independent manner. It is likely that modulation of the above factors, as well as reduction in phosphorylation of S6K1 and 4E-BP1, may collectively contribute to the observed decrease in global protein synthesis. Phosphorylation of eIF2α on Ser⁵¹ by four kinases (PERK, GCN2, PKR, HRI) inhibits eIF2α-GTP-Met-tRNA ternary complex formation and results in attenuation of global protein synthesis under diverse stress conditions (32). The kinase(s) that phosphorylates eIF2α following the exposure to micromolar CCI-779 remains to be determined. It is well established that phosphorylation of Thr⁵⁶ of eEF2 by eEF2K decreases elongation rates and blocks protein synthesis in transformed human cells (41, 42). mTOR mediates inhibitory phosphorylation of three sites on eEF2K: Ser³⁶⁶ via S6K1, Ser³⁵⁹ and Ser⁷⁸ through the unknown kinases (43). Our results indicate that stimulation of Thr⁵⁶ of eEF2 by micromolar CCI-779 is not mediated by dephosphorylation of eEF2K at Ser³⁶⁶ or modulated by S6K1. It is likely that phosphorylation changes in residue(s) other than Ser³⁶⁶ and/or other biochemical mechanism(s) may mediate eEF2K activation and eEF2 hyperphosphorylation. Whereas AMPK pathway also impinges on eEF2K in certain cellular contexts, preincubation of cancer cells with compound C, a direct and specific inhibitor of AMPK, did not eliminate the effect of micromolar CCI-779 on eEF2 phosphorylation (data not shown). Finally, the silencing of eEF2K using siRNA prevented the effects of micromolar CCI-779 on eEF2 phosphorylation, thus substantiating the role of eEF2K in regulating eEF2 activity under above conditions. Alternative mode of regulation independent on eEF2K cannot be ruled out where the exposure to micromolar dose of CCI-779 may suppress the activity of PP2A phosphatase, thereby delaying the rate of Thr⁵⁶ dephosphorylation. Similar mechanism was recently proposed for the ethanol-induced eEF2 phosphorylation in C2C12 myocytes (44).

Whereas it remains an open question whether this newly described mechanism of CCI-779 action contributes to antitumor efficacy in the clinic, a number of recent clinical studies suggest that micromolar drug levels are present in blood of CCI-779–treated patients with breast cancer (1), advanced renal cancer (15), and malignant glioma (16). Therefore, the high-dose drug concentrations that are used in our work are well within the clinically relevant drug concentrations. Interestingly, in a randomized phase II renal cancer trial, patients who received the highest weekly dose of CCI-779 of 250 mg (whole blood exposure of 16 μg/mL) had a higher median survival and higher probability of survival than those in the lower dosage groups (15). Therefore, we speculate that the newly described pharmacologic effects exerted by micromolar CCI-779/rapamycin action in cancer cells contribute to its therapeutic efficacy in vivo.

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.

We thank Dr. Kim Arndt, Dr. Tianmin Zhu, Judy Lucas, and Celine Shi for technical assistance and helpful discussions.