`inhibition of FRAPymTOR
`
`Mehran S. Neshat*†, Ingo K. Mellinghoff*, Chris Tran*, Bangyan Stiles‡, George Thomas§, Roseann Petersen¶,
`Philip Frost¶, James J. Gibbons¶, Hong Wu‡i, and Charles L. Sawyers*†**
`Departments of *Medicine, ‡Medical and Molecular Pharmacology, §Pathology, †Molecular Biology Institute, and iHoward Hughes Medical Institute,
`University of California School of Medicine, Los Angeles, CA 90095; and ¶Wyeth Ayerst Research, Pearl River, NY 10965
`
`Edited by Richard D. Klausner, National Institutes of Health, Bethesda, MD, and approved June 26, 2001 (received for review February 15, 2001)
`collective evidence supports a role for mTOR in PI3-kinaseyAkt
`Recent evidence places the FRAPymTOR kinase downstream of the
`phosphatidyl inositol 3-kinaseyAkt-signaling pathway, which is
`function, but the relationship is more complex than that of a
`up-regulated in multiple cancers because of loss of the PTEN tumor
`linear signaling pathway (1).
`suppressor gene. We performed biological and biochemical studies
`Genomic amplification of either PI3-kinase or Akt has been
`to determine whether PTEN-deficient cancer cells are sensitive to
`reported in cervical, ovarian, and pancreatic cancers (23–25). In
`pharmacologic inhibition of FRAPymTOR by using the rapamycin
`addition, mutations in the tumor suppressor phosphatase gene
`PTEN, which regulates signaling through the PI3-kinaseyAkt-
`derivative CCI-779. In vitro and in vivo studies of isogenic PTEN1y1
`and PTEN2y2 mouse cells as well as human cancer cells with defined
`signaling pathway, occur commonly in prostate, glioblastoma,
`PTEN status showed that the growth of PTEN null cells was blocked
`and endometrial tumors (26–30). Because mTOR may function
`preferentially by pharmacologic FRAPymTOR inhibition. Enhanced
`in the PI3-kinaseyAkt pathway, we examined the potential
`tumor growth caused by constitutive activation of Akt in PTEN1y1
`antitumor properties of mTOR inhibitors in PTEN null tumors.
`PTEN2y2 mouse cells and human tumor lines lacking PTEN
`cells also was reversed by CCI-779 treatment,
`indicating that
`FRAPymTOR functions downstream of Akt in tumorigenesis. Loss
`were more sensitive than wild-type PTEN cells to the growth-
`of PTEN correlated with increased S6 kinase activity and phosphor-
`inhibitory effects of CCI-779, an ester of rapamycin. Transfor-
`ylation of ribosomal S6 protein, providing evidence for activation
`mation mediated by expression of activated Akt in cells with an
`of the FRAPymTOR pathway in these cells. Differential sensitivity
`intact PTEN gene also was reversed by CCI-779 treatment.
`to CCI-779 was not explained by differences in biochemical block-
`Biochemical analysis of the mTOR target S6 kinase showed
`ade of the FRAPymTOR pathway, because S6 phosphorylation was
`CCI-779-dependent constitutive activation in PTEN null cells,
`inhibited in sensitive and resistant cell lines. These results provide
`indicating up-regulation of the mTOR pathway. These results
`rationale for testing FRAPymTOR inhibitors in PTEN null human
`suggest that drugs that target mTOR may have clinical utility in
`cancers.
`human cancers lacking PTEN.
`
`Rapamycin is a natural product macrolide that induces G1
`
`growth arrest in yeast, Drosophila, and mammalian cells (1).
`Genetic and biochemical studies have established that the target
`of rapamycin is the protein kinase FRAPymTOR (hereafter
`called mTOR) (1–5), an evolutionarily conserved member of the
`phosphoinositide kinase-related kinase family that includes
`DNA-PK, ATM, and ATR (6). Rapamycin forms a complex with
`the immunophilin prolyl isomerase FKBP12, which binds spe-
`cifically to mTOR and inhibits its ability to phosphorylate
`substrates such as S6 kinase and 4E-BP1. Clinically, rapamycin
`is an approved immunosuppressive agent, based on its ability to
`block T cell activation (7, 8). Because rapamycin also can induce
`growth arrest and apoptosis in certain tumor cells (9), it is under
`investigation as a potential anticancer drug.
`Studies in yeast and mammalian cells suggest that mTOR
`functions as part of a nutrient-sensing mechanism, regulating the
`cellular response to starvation conditions such as amino acid
`deprivation (1, 10). This function is consistent with its biochem-
`ical activity in regulating S6 kinase and 4E-BP1, two mTOR
`targets that play fundamental roles in ribosome biogenesis and
`cap-dependent translation, respectively (11, 12). S6 kinase and
`4E-BP1 are also regulated, in part, through the phosphatidyl
`inositol 3-kinase (PI3-kinase)yAkt-signaling pathway (13, 14).
`That mTOR is phosphorylated by Akt (15) raises the possibility
`of a direct signaling pathway from PI3-kinaseyAkt to mTOR.
`Genetic studies in Drosophila are consistent with this hypothesis,
`as dTOR is downstream and epistatic to the PI3-kinaseyAkt
`pathway (16, 17). However, dTOR loss gives a more severe
`phenotype than PI3-kinase, Akt, or S6 kinase loss (18–21). In
`addition, the Akt phosphorylation site on mTOR is not required
`for S6 kinase activation (15). Finally, only membrane-targeted
`alleles of Akt are sufficient to activate S6 kinase, whereas
`4E-BP1 phosphorylation appears to be Akt-dependent (22). The
`
`Methods
`Protein Analysis. S6 kinase activity was measured by in vitro
`immune complex assay by using an artificial consensus peptide
`as substrate as described (31). Phosphorylated (Ser-235y236)
`and pan-S6 antibodies were provided by Cell Signaling Tech-
`nology (Beverly, MA). Akt and MAPK activation was measured
`by immunoblot assay by using phosphospecific antibodies with
`controls for total Akt and MAPK protein as described (32).
`Pan-4E-BP1 antibody was provided by N. Sonenberg (McGill,
`Montreal, Canada). The level of 4E-BP1 bound to eIF4E was
`measured by precipitation of eIF4E by using 7methyl-GTP
`Sepharose (Amersham Pharmacia), followed by 4E-BP1 immu-
`noblot as described (14). Cyclin D1 and actin antibodies were
`obtained from Santa Cruz Biotechnology. eIF4E antibody was
`obtained from Signal Transduction Laboratories (San Diego).
`
`Tissue Culture Experiments. PTEN1y1 and PTEN2y2 embryonic
`stem (ES) cells and mouse embryo fibroblasts (MEFs) were
`derived as described previously (33). 9L rat glioblastoma cells
`were provided by L. Liau (University of California at Los
`Angeles). All other cell lines were from American Type Culture
`
`This paper was submitted directly (Track II) to the PNAS office.
`Abbreviations: FRAPymTOR, mammalian target of rapamycin; PI3-kinase, phosphatidyl
`inositol 3-kinase; MEF, mouse embryo fibroblasts; ES, embryonic stem; SCID, severe com-
`bined immunodeficient.
`
`See commentary on page 10031.
`
`**To whom reprint requests should be addressed at: University of California Los Angeles,
`Hematology–Oncology, 11–934 Factor Building, 10833 Le Conte Avenue, Los Angeles, CA
`90095-1678. E-mail: csawyers@mednet.ucla.edu.
`
`The publication costs of this article were defrayed in part by page charge payment. This
`article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C.
`§1734 solely to indicate this fact.
`
`10314 –10319 u PNAS u August 28, 2001 u vol. 98 u no. 18
`
`www.pnas.orgycgiydoiy10.1073ypnas.171076798
`
`ARGENTUM
`Exhibit 1024
`
`000001
`
`
`
`MEDICALSCIENCES
`
`PTEN null cells have enhanced sensitivity to mTOR inhibition. Isogenic MEFs derived from PTEN knockout mice were cultured for 72 h in the presence
`Fig. 1.
`of vehicle alone (open bars), 1.0 nM CCI-779 (hatched bars), or 10 nM CCI-779 (solid bars). Cell growth was measured by cell counts determined by Trypan blue
`staining. The data are plotted as the number of cells relative to vehicle control for three experiments. The IC50 for growth inhibition of the cell lines in the table
`was determined by 3H-uptake studies. (*IC50 for 9L was obtained based on lack of growth inhibition by up to 10 nM of CCI-779 as measured by direct cell count.)
`Examples are shown for U87MG and MDA-435. The PTEN status of these lines has been reported previously (26, 35) with the exception of SF-268, SF-295, and
`SF-539, which were kindly provided by R. Parsons (personal communication).
`
`Collection. Cell growth was measured by [3H]thymidine uptake
`andyor cell counts, which were determined by Trypan blue
`staining. CCI-779 was obtained from Wyeth Ayerst Laboratories
`(Marietta, PA).
`
`Mouse Experiments. PTEN1y1 or PTEN2y2 ES cells were in-
`jected s.c. into nude mice. LAPC-4 and LAPC-9 prostate cancer
`xenografts were maintained by serial passage in male severe
`combined immunodeficient (SCID) mice as described and in-
`jected as single cell suspensions (34). Tumor growth was mea-
`sured daily, and mice were randomized to CCI-779 vs. vehicle
`when tumors reached the size indicated in each experiment.
`Treatment was given by i.p. injection for 5 consecutive days.
`Serum prostate-specific antigen levels were measured by ELISA
`as described (34). Hematoxylinyeosin, Ki-67 antibody, and ter-
`minal deoxynucleotidyltransferase-mediated UTP end-labeling
`staining of tumor tissue was performed as described previously
`(34). Cell size (in m2) was measured by using Microcomp
`SCIMEASURE analytical software, calibrated at 3200 magnifica-
`tion. Statistical analysis was performed by using Student’s t test
`and the ANOVA statistical model (Stata Software, College
`Station, TX).
`
`Results
`Enhanced Sensitivity of PTEN Null Cells to mTOR Inhibition. To
`examine the effect of mTOR pathway inhibition on the growth
`of PTEN wild-type vs. PTEN-deficient cells, we exposed mouse
`and human cells with defined PTEN status to varying doses of
`CCI-779 in vitro. CCI-779 is an ester of rapamycin with equiv-
`alent activity and specificity for mTOR.†† First, we examined this
`question in a genetically defined background by comparing the
`growth of PTEN1y1 and PTEN2y2 MEFs that contain a tar-
`geted deletion of the PTEN phosphatase domain (33). The
`growth of PTEN2y2 MEFs was inhibited 40% and 60% by 1.0
`nM and 10 nM CCI-779, respectively, whereas PTEN1y1 MEFs
`
`††Gibbons, J. J., Discafani, C., Peterson, R., Hernandez, R., Skotnicki, J. & Fruman, D. (1999)
`Proc. Am. Assoc. Cancer Res. 40, 301 (abstr.).
`
`were unaffected (Fig. 1 Left). This result cannot be explained by
`an increased growth rate of PTEN null cells because the
`doubling times of PTEN null and PTEN wild-type cells were
`comparable in the absence of CCI-779 (data not shown). We
`expanded the analysis to a panel of 10 tumor cell lines with
`defined PTEN status. The concentration of CCI-779 required
`for 50% growth inhibition (IC50) was less than 10 nM for all of
`the PTEN null tumor cell lines (Fig. 1; see table and [3H]thy-
`midine uptake curve of U87MG vs. MDA-435). It is important
`to note that DU145 cells, which have an intact PTEN gene, are
`relatively sensitive to CCI-779 compared with MDA-435 or
`SF-268, thereby providing an exception to this correlation. One
`explanation might be the elevated Akt3 expression and consti-
`tutive mTOR phosphorylation reported in these cells (15, 36),
`suggesting that either PTEN loss or Akt activation may enhance
`sensitivity to CCI-779. Alternatively, there may be mechanisms
`independent of the PTENyAkt pathway that also affect sensi-
`tivity to CCI-779.
`
`Enhanced S6 Kinase Activation in PTEN Null Tumor Lines. Because
`PTEN-deficient cells have elevated Akt activation, we asked
`whether these cells also have enhanced mTOR activity. We
`explored this question by examining S6 kinase, a downstream
`target of mTOR. Upon activation by mTOR, S6 kinase phos-
`phorylates the 40S ribosomal protein S6, which functions in
`translation of mRNAs with a 59 terminal oligopolypyrimidine
`(TOP) sequence (37, 38). Because S6 mRNA itself contains a 59
`TOP sequence, both translation and phosphorylation of S6
`protein are measures of S6 kinase activity. Under serum star-
`vation conditions, S6 protein level and phosphorylation were
`increased in PTEN null U87MG cells compared with PTEN
`wild-type 9L cells. After serum challenge, S6 protein levels and
`phosphorylation increased in both cell lines and were mTOR-
`dependent, because treatment with CCI-779 reduced S6 levels
`and phosphorylation to undetectable levels (Fig. 2A). These data
`(and data from additional models shown in Figs. 3C and 6)
`provide evidence that loss of PTEN is accompanied by an
`increase in mTOR-dependent S6 kinase activity.
`
`Neshat et al.
`
`PNAS u August 28, 2001 u vol. 98 u no. 18 u 10315
`
`000002
`
`
`
`Enhanced inVivoSensitivity of PTEN-Deficient Tumors to CCI-779. To
`assess the role of mTOR in tumor growth mediated by PTEN
`loss, we examined the effect of CCI-779 on the growth of
`isogenic PTEN1y1 or PTEN2y2 tumors produced by s.c. injec-
`tion of murine ES cells into immunodeficient nude mice. When
`tumors reached 200 mm3 in size, mice were randomly assigned
`to treatment with CCI-779 or vehicle. The growth of PTEN1y1
`ES tumors was slowed during the 5 days of drug administration
`but resumed after the drug was stopped (Fig. 3A). In contrast,
`the growth of PTEN2y2 tumors was completely blocked for the
`duration of the experiment, providing evidence that PTEN-
`deficient tumors derived from a defined genetic background
`have enhanced dependence on mTOR for growth.
`We explored the potential clinical utility of this observation by
`using human tumors grown in immunodeficient SCID mice.
`LAPC-4 (PTEN wild type) and LAPC-9 (PTEN null) (39) are
`androgen-dependent human prostate cancer xenografts that
`grow in male SCID mice with defined kinetics and secrete the
`tumor marker prostate-specific antigen into the serum (34).
`Because phase I clinical studies of CCI-779 suggest that doses in
`the 3-mgykg range are well tolerated in humans and give serum
`levels well above that required to inhibit mTOR,‡‡ we examined
`a range of CCI-779 doses. Male mice were injected s.c. with
`LAPC-4 or LAPC-9 cells and randomly assigned to treatment
`with three different doses of CCI-779 (0.1 mgykg, 4.0 mgykg, or
`40 mgykg) or vehicle when tumors reached ’200 mm3 in size.
`CCI-779 impaired the growth of wild-type PTEN LAPC-4
`tumors in a dose-dependent fashion, with minimal activity at the
`0.1-mgykg dose, partial activity at 4.0 mgykg, and nearly com-
`plete growth suppression at 40 mgykg. In contrast, all three doses
`were equally effective at completely blocking PTEN null
`LAPC-9 tumor growth in three independent experiments (Fig.
`3B). These results from isogenic mouse tumors and human
`tumor lines establish an increase in sensitivity of PTEN null
`tumors to mTOR inhibition.
`We examined the effect of CCI-779 on S6 kinase inhibition by
`measuring S6 levels and phosphorylation in tumors harvested on
`day 5 of treatment with the 0.1-mgykg dose (Fig. 3C). In
`vehicle-treated mice, both the level and phosphorylation state of
`S6 protein were increased in PTEN null LAPC-9 tumors com-
`pared with PTEN wild-type LAPC-4 tumors, consistent with the
`notion that PTEN loss leads to up-regulation of S6 kinase
`activity. In the CCI-779-treated animals, S6 phosphorylation was
`almost completely blocked in LAPC-9 tumors and the level of S6
`protein was reduced, indicating effective inhibition of mTOR. It
`is difficult to make similar conclusions for the LAPC-4 tumors
`because basal S6 phosphorylation is low. The possibility of
`differential inhibition of mTOR in PTEN wild type vs. PTEN
`null cells is addressed further in Fig. 6.
`To determine the effect of CCI-779 treatment on growth and
`apoptosis, we compared proliferation (as determined by anti-
`Ki-67 staining) and apoptosis (as determined by apoptotic bodies
`and terminal deoxynucleotidyltransferase-mediated UTP end-
`labeling staining) in LAPC-4 and LAPC-9 tumors by examining
`histologic sections obtained after 5 days of treatment with
`vehicle or 0.1 mgykg CCI-779. Measurements were based on
`examination of 10 slides per treatment group and 12–18 high-
`power fields per slide. The number of Ki-67-positive nuclei fell
`1.4-fold in LAPC-4 cells vs. 2.6-fold in LAPC-9 cells, and the
`number of apoptotic cells increased 3.4-fold in LAPC-4 cells and
`1.9-fold in LAPC-9 cells (Fig. 4). Statistical analysis indicated
`that the decrease in proliferation was significantly greater in the
`
`‡‡Raymond, E., Alexandre, J., Depenbrock, H., Mekhaldi, S., Angevin, E., Hanauske, A.,
`Baudin, E., Escudier, B., Frisch, J., Boni, J., et al. Am. Soc. Clin. Oncol. 36th Annual Meeting,
`May 19 –23, 2000, New Orleans, LA, 728 (abstr.).
`
`Enhanced S6 kinase activation in PTEN null tumor lines. (A) Phos-
`Fig. 2.
`phorylated (Ser-235y236) and total S6 protein and actin were measured by
`immunoblot in 9L and U87MG cells treated with vehicle or doses of 0.1, 1.0,
`and 10 nM CCI-779 for 7 h. Serum challenge was for 15 min. Equal amounts of
`protein were loaded per lane as determined by Bio-Rad DC protein assay. (B)
`S6 kinase activity was measured by in vitro kinase assay (31) in MDCK cells
`stably transfected with vector or wild-type PTEN (Left) and in LAPC-4 prostate
`cancer cells stably infected with retrovirus expressing myr-Akt or vector con-
`trol (Right). MDCK cells were serum-starved overnight, pretreated with vehi-
`cle (open bars) or 10 nM of CCI-779 (solid bars), and then challenged with
`serum for 15 min. LAPC-4 cells were treated identically except that no serum
`challenge was given. Results from two experiments are plotted relative to
`serum-starved, untreated cells. Expression of PTEN and Akt in the transfec-
`tants was confirmed by immunoblot (data not shown). (C) Levels of total and
`activated Akt and MAPK were measured by immunoblot in lysates of PTEN
`wild-type (DU145) and PTEN null (PC3) cells by using specified antibodies after
`2 h of pretreatment with vehicle or 10 nM CCI-779.
`
`To directly test the role of the PTENyAkt pathway in regu-
`lating S6 kinase, we engineered PTEN or Akt overexpression in
`cells with wild-type endogenous PTEN. PTEN overexpression
`blunted serum-induced S6 kinase activity, as measured by using
`an immunoprecipitation kinase assay (Fig. 2B Left). Conversely,
`expression of a constitutively active, membrane-bound allele of
`Akt (myr-Akt) was sufficient to activate S6 kinase in the absence
`of serum, and this activation was mTOR-dependent (Fig. 2b
`Right). The effects of CCI-779 appear to be specific to the mTOR
`pathway, because no inhibition of Akt or MAPK activation was
`observed (Fig. 2C). Taken together, these data support the
`hypothesis that S6 kinase is activated in PTEN null cells.
`
`10316 u www.pnas.orgycgiydoiy10.1073ypnas.171076798
`
`Neshat et al.
`
`000003
`
`
`
`PTEN null cells have enhanced sensitivity to mTOR inhibition in vivo. (A) PTEN1y1 or PTEN2y2 ES cells were injected s.c. into nude mice at a dose of 5 3
`Fig. 3.
`105 cells per mouse (n 5 20). When tumor volume reached 200 mm3, mice were randomized to treatment with vehicle or 40 mgykg CCI-779. The fold change
`in tumor volume in response to treatment was plotted. (B) Single-cell suspensions of LAPC-4 or LAPC-9 prostate cancer xenografts were injected s.c. into male
`SCID mice (n 5 80) at a dose of 106 cells per mouse. When tumors became palpable, mice were randomized (arrow) to treatment with vehicle, 0.1 mgykg, 4 mgykg,
`or 40 mgykg CCI-779. The fold change in tumor volume from two independent experiments is plotted. (C) Tumors were harvested from mice after 5 days of
`treatment with vehicle or 0.1 mgykg CCI-779 and lysed by boiling in 2% SDS buffer. Immunoblots were performed by using antibodies for phosphorylated S6
`(Ser-235y236), total S6, and actin.
`
`PTEN null LAPC-9 tumor line (P 5 0.03), whereas the increase
`in apoptotic bodies was comparable in both xenografts (P 5 0.24).
`In addition to growth and apoptosis, a third parameter that
`can affect tumor volume is cell size. Genetic analysis has
`established that PI3-kinase, dPTEN, dAkt, and S6 kinase reg-
`ulate cell size in Drosophila (18–21). Similarly, S6 kinase-
`deficient mice have insulin insufficiency because of small islet
`
`cells in the pancreas (40). To determine whether mTOR inhi-
`bition affects cell size in tumors, we performed morphometric
`analysis of 20 tissue sections, in which 1,400 cells were analyzed.
`CCI-779 treatment led to a 1.4-fold reduction in cell size in
`PTEN null LAPC-9 tumors but had no effect in PTEN wild-type
`LAPC-4 tumors (P 5 0.01). These findings indicate that CCI-779
`can affect at least three parameters that influence tumor volume:
`growth, apoptosis, and cell size. Although further work is needed
`to fully account for its antitumor activity, the major difference
`in the effect of CCI-779 on PTEN null and PTEN wild-type cells
`appears to be growth suppression and reduction in cell size.
`
`Akt-Mediated Tumor Growth Is mTOR-Dependent. Because Akt is a
`major effector of transformation caused by PTEN loss, we asked
`whether introduction of constitutively active Akt into wild-type
`PTEN cells would affect CCI-779 sensitivity. To test this hy-
`pothesis, we introduced a constitutively active allele of Akt
`(myr-Akt) in wild-type PTEN LAPC-4 cells, which activated S6
`kinase in an mTOR-dependent fashion (refer to Fig. 2B). Male
`SCID mice were injected with LAPC-4ypuro or LAPC-4ymyr-
`Akt cells, tumors were allowed to reach 200 mm3 in size, and
`mice were randomly assigned to treatment with 0.1 mgykg
`
`MEDICALSCIENCES
`
`Akt-mediated growth in PTEN wild-type tumor cells is mTOR-
`Fig. 5.
`dependent. LAPC-4ypuro or LAPC-4ymyr-Akt cells were injected s.c. into male
`SCID mice (n 5 20) at a dose of 106 cells per mouse. When tumor volume
`reached 50 –200 mm3, mice were randomized to treatment with vehicle or 0.1
`mgykg CCI-779, given by i.p. injection for 5 consecutive days in weeks 1 and 4.
`The fold changes in tumor volume (Left) and in serum prostate-specific
`antigen (PSA) (Right) are plotted.
`
`Effect of CCI-779 on cell size, proliferation, and apoptosis. Photomi-
`Fig. 4.
`crographs (3200) of hematoxylinyeosin- and Ki-67-stained sections of LAPC-4
`and LAPC-9 tumors are shown after 5 days of treatment with vehicle or
`CCI-779. Cell size was measured by morphometry (1,200 total cells on 20
`sections). Proliferation and apoptosis were measured by counting the number
`of Ki-67-positive nuclei or apoptotic bodies [terminal deoxynucleotidyltrans-
`ferase-mediated UTP end labeling (TUNEL)-positive] divided by high-power
`field (hpf) (12–18 hpfytumor, 5 tumorsygroup). All values are mean 6 SEM.
`Both cell size and proliferation were reduced significantly in CCI-779-treated
`LAPC-9 tumors but not LAPC-4 tumors (LAPC-4 size: P 5 0.95; LAPC-9 size: P 5
`0.001; LAPC-4 proliferation: P 5 0.212; LAPC-9 proliferation: P 5 0.01, Stu-
`dent’s t test). Apoptosis was increased significantly in treated LAPC-4 and
`LAPC-9 tumors (P 5 0.01 and P 5 0.005, respectively, Student’s t test). The
`relative effect of CCI-779 treatment on size, proliferation, and apoptosis in
`LAPC-4 vs. LAPC-9 was compared by using an ANOVA model, with P values
`shown above the bar graphs.
`
`Neshat et al.
`
`PNAS u August 28, 2001 u vol. 98 u no. 18 u 10317
`
`000004
`
`
`
`PTEN2y2 cells (Fig. 6 Bottom). Treatment with CCI-779 caused
`a dose-dependent increase in eIF4E-bound 4E-BP1 in PTEN1y1
`cells and a dose-dependent decrease in hyperphosphorylated as
`well as total 4E-BP1 in PTEN2y2 cells, indicating that 4E-BP1
`phosphorylation is, in part, mTOR-dependent in both genetic
`backgrounds. Surprisingly, the decrease in 4E-BP1 phosphory-
`lation in PTEN2y2 cells was not accompanied by a detectable
`increase in eIF4E-bound 4E-BP1. A full explanation of this
`result and the apparent increase in total levels of 4E-BP1 in
`PTEN2y2 cells require further analysis. We also examined cyclin
`D1, a potential downstream parameter of 4E-BP1 phosphory-
`lation whose translation is dependent on eIF4E function (and is
`rapamycin-sensitive) because of complex secondary structure of
`the 59 untranslated region of the mRNA (44, 45). Cyclin D1
`protein levels were elevated in PTEN2y2 compared with
`PTEN1y1 MEFs and were reduced by CCI-779 treatment in
`both cell lines (Fig. 6). Because cyclin D1 is also regulated by
`transcriptional and posttranslationalydegradative mechanisms
`independent of mTOR (46, 47), we cannot make definitive
`conclusions about how each of these variables influences cyclin
`D1 levels in PTEN2y2 cells. Nevertheless, the collective evi-
`dence from analysis of S6, 4E-BP1, and cyclin D1 indicates that
`doses of CCI-779 (10 nM) that cause selective growth inhibition
`in PTEN2y2 cells block mTOR signaling in PTEN1y1 and
`PTEN2y2 cells.
`Discussion
`Our studies in human tumor lines and cells from PTEN
`knockout mice demonstrate that PTEN-deficient cells are
`sensitive to growth inhibition caused by pharmacologic mTOR
`blockade. We also show that a membrane-targeted Akt allele,
`which induces mTOR pathway activation in PTEN wild-type
`cells, promotes tumor growth in a fashion that is reversed by
`mTOR inhibition. Although targeted mTOR gene deletion is
`required to conclusively establish a genetic relationship be-
`tween PTEN, Akt, and mTOR in tumorigenesis, our data
`provide pharmacologic evidence that mTOR is an important
`effector of transformation mediated by perturbation of the
`PTENyAkt pathway. This conclusion is consistent with recent
`work in chicken embryo fibroblasts showing that the mTOR
`inhibitor rapamycin specifically blocks focus formation in-
`duced by oncogenic alleles of PI3-kinase or Akt, but not by Src,
`Ras, Myc, or other oncogenes (48). Because of evidence from
`other studies that PI3-kinaseyAkt- and mTOR-mediated sig-
`nals to S6 kinase and 4E-BP1 can be separated (15, 22), it
`important to stress that the relationship between these signal-
`ing pathways is not strictly linear.
`Why does loss of PTEN sensitize tumors to mTOR inhibition?
`Signaling through the PI3-kinaseyAkt pathway normally is
`tightly regulated, such that PIP3 levels are maintained at low
`concentrations and rise transiently in response to specific growth
`factor signals. In contrast, PIP3 levels are constitutively elevated
`in tumors lacking PTEN, raising the possibility that sustained
`activation of signaling molecules downstream of PIP3 renders
`cells more dependent on this pathway for growth. If this is the
`case, cells with PTEN loss could be more sensitive to the
`biological effects of mTOR inhibition than PTEN wild-type
`cells, even though biochemical inhibition of the mTOR pathway
`occurs efficiently in both cell types. Our data showing efficient
`blockade of S6 phosphorylation in all cell lines tested, regardless
`of their biological sensitivity to CCI-779, provide support for this
`hypothesis. A similar conclusion was drawn in the analysis of
`rapamycin-sensitive and rapamycin-resistant rhabdomyosarco-
`mas (49). In that study, elevated c-myc expression was correlated
`with rapamycin resistance, raising the possibility of parallel
`pathways that can rescue cells from mTOR dependence. In
`preliminary studies, we were unable to demonstrate a similar
`correlation in PTEN1y1 and PTEN2y2 MEFs (M.S.N. and
`
`Fig. 6. Analysis of S6, 4E-BP1, and cyclin D1 in PTEN1y1 and PTEN2y2 MEFs.
`PTEN1y1 or PTEN2y2 MEFs were treated with the indicated concentrations of
`CCI-779 for 6 h, lysed by three freezeythaw cycles in 50 mM Tris, pH 7.5y150
`mM KCl mM EDTAy1 mM EGTAy1 mM DTTy50 mM 2-mercaptoethanol, sup-
`plemented with protease and phosphatase inhibitors, and probed with anti-
`bodies to phosphorylated S6, total S6, total 4E-BP1, or actin. The level of
`4E-BP1 bound to eIF4E was measured by precipitation of eIF4E by using
`7methyl-GTP Sepharose, followed by 4E-BP1 immunoblot. Comparable pre-
`cipitation of eIF4E was confirmed by immunoblot.
`
`CCI-779 or vehicle for 5 days. In the absence of treatment,
`LAPC-4ymyr-Akt tumors grew more quickly than LAPC-4y
`puro tumors,
`indicating that activation of the Akt pathway
`enhances tumor growth in this model. The enhanced growth of
`LAPC-4ymyr-Akt tumors was completely inhibited by CCI-779
`treatment, whereas this dose had no discernible effect on
`LAPC-4ypuro tumors (Fig. 5 Left). Similar results were obtained
`by using serum prostate-specific antigen as the treatment
`endpoint (Fig. 5 Right). These results establish that mTOR
`is required for the enhanced tumor growth caused by Akt
`activation.
`
`Analysis of S6 and 4E-BP1 in PTEN1y1 and PTEN2y2 MEFs. One
`potential explanation for the differential sensitivity of PTEN
`wild-type and PTEN null tumor cells to CCI-779 might be a
`difference in biochemical inhibition of the mTOR pathway. We
`explored this issue by comparing S6 protein levels and phos-
`phorylation in PTEN1y1 and PTEN2y2 MEFs after exposure to
`concentrations of CCI-779 ranging from 0.05 to 10 nM. As
`observed previously in tumor cell lines (Fig. 2 A) and xenografts
`(Fig. 3C), the level and phosphorylation state of S6 were
`markedly elevated in PTEN2y2 cells (Fig. 6 Top). S6 phosphor-
`ylation and protein level were inhibited in a dose-dependent
`fashion in PTEN2y2 cells. In PTEN1y1 MEFs, S6 protein level
`was also reduced, but this effect was observed only at the 10-nM
`dose (S6 phosphorylation was too low in untreated PTEN1y1
`MEFs to measure any changes).
`We examined this issue further by characterizing the status of
`4E-BP1, a protein that blocks translation of 59 cap mRNAs by
`binding the initiation factor eIF4E (41). 4E-BP1 phosphoryla-
`tion, which is modulated by PI3-kinaseyAkt and mTOR as well
`as other pathways, occurs on multiple residues and prevents
`complex formation with eIF4E (22, 42, 43). The level of hyper-
`phosphorylated 4E-BP1 (detected as a slow-mobility isoform in
`SDSyPAGE) was elevated in PTEN2y2 relative to PTEN1y1
`MEFs (Fig. 6 Middle). Conversely, the level of 4E-BP1 (unphos-
`phorylated) bound to eIF4E was increased in PTEN1y1 vs.
`
`10318 u www.pnas.orgycgiydoiy10.1073ypnas.171076798
`
`Neshat et al.
`
`000005
`
`
`
`C.L.S, unpublished observations). Identification of these path-
`ways in PTEN1y1 vs. PTEN2y2 cells requires further investiga-
`tion. However,
`it remains possible that loss of PTEN may
`predispose cells to a greater degree of biochemical inhibition of
`the mTOR pathway at very low doses of CCI-779, as suggested
`by our analysis of S6 protein levels in MEFs. If so, this could have
`important implications with regard to the degree of mTOR
`inhibition in clinical situations in which drug levels vary depend-
`ing on the frequency of dosing.
`Aberrant activation of the PI3-kinaseyAkt pathway occurs in
`a large number of human cancers—primarily through PTEN
`loss, but also through PI3-kinase or Akt gene amplification (23,
`50). Phase I clinical trials of mTOR inhibitors in various cancers
`currently are underway and have shown preliminary evidence of
`
`antitumor activity.‡‡ Our findings suggest that PTEN null can-
`cers are one group of human tumors in which mTOR inhibitors
`may be effective. Because factors other than the PTENyAkt
`pathway also can affect mTOR activation, it will be important
`not to exclude patients with wild-type PTEN tumors from
`clinical trials.
`
`We are grateful to Peter Houghton for sharing data before publication,
`Nahum Sonenberg for providing 4E-BP1 antibody, Fred Dorey for help
`with statistics, Ramon Parsons for providing data on PTEN status of cell
`lines, and Lisa Rose for manuscript preparation. This work was sup-
`ported by grants from the National Institutes of Health, Department of
`Defense, and CaP CURE.
`
`1. Schmelzle, T. & Hall, M. N. (2000) Cell 103, 253–262.
`2. Brown, E. J., Albers, M. W., Shin, T. B., Ichikawa, K., Keith, C. T., Lane, W. S.
`& Schreiber, S. L. (1994) Nature (London) 369, 756–758.
`3. Sabatini, D. M., Erdjument-Bromage, H., Lui, M., Tempst, P. & Snyder, S. H.
`(1994) Cell 78, 35–43.
`4. Chiu, M. I., Katz, H. & Berlin, V. (1994) Proc. Natl. Acad. Sci. USA 91,
`12574–12578.
`5. Sabers, C. J., Martin, M. M., Brunn, G. J., Williams, J. M., Dumont, F. J.,
`Wiederrecht, G. & Abraham, R. T. (1995) J. Biol. Chem. 270, 815–822.
`6. Keith, C. T. & Schreiber, S. L. (1995) Science 270, 50–51.
`7. Schreiber, S. L. (1991) Science 251, 283–287.
`8. Kuo, C. J., Chung, J., Fiorentino, D. F., Flanagan, W. M., Blenis, J. & Crabtree,
`G. R. (1992) Nature (London) 358, 70–73.
`9. Dilling, M. B., Dias, P., Shapiro, D. N., Germain, G. S., Johnson, R. K. &
`Houghton, P. J. (1994) Cancer Res. 54, 903–907.
`10. Hara, K., Yonezawa, K., Weng, Q.-P., Kozlowski, M. T., Belham, C. & Avruch,
`J. (1998) J. Biol. Chem. 273, 14484–14494.
`11. Brown, E. J., Beal, P. A., Keith, C. T., Chen, J., Shin, T. B. & Schreiber, S. L.
`(1995) Nature (London) 377, 441–446.
`12. Brown, E. J. & Schreiber, S. L. (1996) Cell 86, 517–520.
`13. Brunn, G. J., Williams, J., Sabers, C., Wiederrecht, G., Lawrence, J. C., Jr., &
`Abraham, R. T. (1996) EMBO J. 15, 5256–5267.
`14. Gingras, A. C., Kennedy, S. G., O’Leary, M. A., Sonenberg, N. & Hay, N.
`(1998) Genes Dev. 12, 502–513.
`15. Sekulic, A., Hudson, C. C., Homme, J. L., Yin, P., Otterness, D. M., Karnitz,
`L. M. & Abraham, R. T. (2000) Cancer Res. 60, 3504–3513.
`16. Zhang, H., Stallock, J. P., Ng, J. C., Reinhard, C. & Neufeld, T. P. (2000) Genes
`Dev. 14, 2712–2724.
`17. Oldham, S