Research ArticleCancer

Differential Requirement of mTOR in Postmitotic Tissues and Tumorigenesis

See allHide authors and affiliations

Science Signaling  27 Jan 2009:
Vol. 2, Issue 55, pp. ra2
DOI: 10.1126/scisignal.2000189

Abstract

The mammalian target of rapamycin (mTOR) is a crucial effector in a complex signaling network commonly disrupted in cancer. mTOR exerts its multiple functions in the context of two different multiprotein complexes: mTOR complex 1 (mTORC1) and mTOR complex 2 (mTORC2). Loss of the tumor suppressor PTEN (phosphatase and tensin homolog deleted from chromosome 10) can hyperactivate mTOR through AKT and represents one of the most frequent events in human prostate cancer. We show here that conditional inactivation of mTor in the adult mouse prostate is seemingly inconsequential for this postmitotic tissue. Conversely, inactivation of mTor leads to a marked suppression of Pten loss–induced tumor initiation and progression in the prostate. This suppression is more pronounced than that elicited by the sole pharmacological abrogation of mTORC1. Acute inactivation of mTor in vitro also highlights the differential requirement of mTor function in proliferating and transformed cells. Collectively, our data constitute a strong rationale for developing specific mTOR inhibitors targeting both mTORC1 and mTORC2 for the treatment of tumors triggered by PTEN deficiency and aberrant mTOR signaling.

Introduction

The mammalian target of rapamycin (mTOR) is a critical serine/threonine kinase that integrates diverse inputs, including signals from growth factors, nutrients, energy, and stress, to regulate protein synthesis, cell growth, and proliferation (1).

mTOR was discovered in the early 1990s as an inhibitory target of the bacterial macrolide rapamycin (2). Rapamycin has been a valuable tool for uncovering many physiological functions of mTOR in health and disease. Over the years it has become clear that mTOR represents a critical node in a complex signaling network that is often deregulated in human cancer (3, 4).

In particular, an intricate relationship links mTOR to the AKT signaling cascade (5), a pathway that is reported to be hyperactive in many human cancers, including prostate cancer (CaP) (4). In human cancer, the AKT pathway is commonly activated through the inactivation of at least one allele of the tumor suppressor PTEN, an event that occurs in 30 to 70% of primary CaP (6, 7).

In mammalian cells, mTOR affects AKT signaling at multiple levels as a component of two distinct protein complexes, mTOR complex 1 (mTORC1) and mTOR complex 2 (mTORC2). mTORC1, which is activated downstream of AKT, is sensitive to rapamycin-mediated inhibition. mTORC1 promotes cell growth largely through the phosphorylation and subsequent inactivation of eIF4E-binding proteins (4E-BPs) and by the phosphorylation and activation of ribosomal S6 kinase (S6K1) (1). mTORC2 is rapamycin insensitive and has recently been shown to phosphorylate AKT at Ser473 (8). This modification, in conjunction with AKT phosphorylation on Thr308 by the phosphoinositide-dependent kinase–1 (PDK1), triggers full activation of AKT in response to mitogenic stimuli (8).

Downstream of AKT, mTORC1 has been described as the most essential effector in driving cell proliferation and susceptibility to oncogenic transformation (9). This led to the targeting of mTORC1 as a therapeutic strategy in cancer, with rapamycin and its analogs (rapalogs) emerging as the first generation of mTOR inhibitors to enter cancer clinical trials. However, the rapalogs were only modestly successful, with different subsets of cancer showing a highly variable response to rapalog treatment (3, 10). This unexpected outcome might be related to the inhibition of critical negative feedback loops arising from mTORC1 (11, 12). The recent finding that mTORC2 phosphorylates and activates AKT (8) suggests that the use of rapalogs may be limited by their preferential inhibition of mTORC1. This raises the question of whether inactivation of mTORC1 and mTORC2 may represent a more effective therapeutic strategy.

Here, we report that conditional genetic inactivation of mTor in the mouse prostate, which results in the abrogation of mTORC1 and mTORC2 functions, does not have major consequences on postpubescent prostate gland physiology. Conversely, mTor inactivation represses prostate tumorigenesis driven by Pten loss, resulting in a reduction of intraepithelial neoplasia at early stages and the abrogation of invasive cancer at later time points. This reduction was markedly more efficient than that obtained by treatment with the mTORC1 preferential inhibitor RAD001. In line with these data, mTor deletion markedly affects proliferating cells in vitro. Hence, this study represents a proof of principle rationale for the therapeutic use of mTOR kinase inhibitors in the treatment of human cancers.

Results

mTor deletion has a negligible effect on mouse prostate morphology and function

mTOR has been reported to play a critical role in cell growth and proliferation in early mouse embryos (13, 14), consistent with results in Saccharomyces cerevisiae (15) and Drosophila melanogaster (16, 17). Indeed, complete inactivation of mTor in mice results in embryonic lethality at embryonic day 5.5 (E5.5), confirming the importance of mTOR in early development (13, 14). To analyze the consequences of mTor inactivation in a fully developed adult mouse tissue, we conditionally deleted mTor in the prostate, a tissue that is exquisitely sensitive to the Akt-mTOR signaling cascade (4).

We used the Cre/loxP approach (18) and specifically inactivated mTor in the prostates of mTorloxP/loxP mice by crossing them with Probasin-Cre4 (PB-Cre4) transgenic mice. PB-Cre4 transgenic mice specifically express Cre recombinase under the control of the ARR2 Probasin promoter in the prostate epithelium after puberty (19). For simplicity, we refer to the resulting mTorloxP/loxP;PB-Cre4 offspring as mTorpc−/−. Polymerase chain reaction (PCR) analysis revealed the presence of the mTor recombined allele specifically in the three lobes of the prostate, namely, anterior prostate (AP), ventral prostate (VP), and dorsolateral prostate (DLP), and a low amount of recombination in the seminal vesicles (Fig. 1A), consistent with what has been previously reported for Probasin-driven Cre expression (20). To determine the efficiency of mTor recombination in the prostates of these mice, first we compared mTor messenger RNA (mRNA) expression in the whole prostate of the mTorpc−/− mice with that of their wild-type (WT) littermates by quantitative reverse transcription PCR (RT-PCR) (Fig. 1B). Although we observed a significant reduction in the expression of mTor, the extent of recombination was underestimated because of nonepithelial WT contaminating tissue, preventing an accurate measure of mTor recombination in the prostatic epithelium, the compartment that expresses Probasin-driven Cre recombinase. Therefore, we used laser capture microdissection (LCM) to specifically extract prostatic epithelial cells from prostate cryosections of WT, mTorpc+/−, and mTorpc−/− mice (inset in Fig. 1C). DNA was extracted from these cells, and genomic quantitative real-time PCR analysis revealed that mTor is efficiently deleted in the prostates of the mTorpc−/− mice (Fig. 1C).

Fig. 1

Effect of mTor conditional inactivation in the mouse prostate. (A) PCR analysis to detect recombination of mTorΔ allele on DNA from several mouse tissues. AP, VP, and DLP stand for anterior prostate, ventral prostate, and dorsolateral prostate, respectively. (B) mTor quantitative RT-PCR analysis of the prostate mRNA from WT and mTorpc−/− mice. Error bars show SDs from three independent experiments. *P < 0.05. (C) mTor genomic quantitative real-time PCR on DNA extracted from laser capture microdissected (LCM) epithelial cells of WT, mTorpc+/−, and mTorpc−/− mice prostates. The inset shows the genotyping by PCR of the DNA extracted from the prostatic epithelial cells obtained by LCM and subsequently analyzed by genomic quantitative real-time PCR. (D) Representative images of prostates from mTorpc−/− and WT 3-month-old mice. (E) Weights of prostate from mTorpc−/− and WT mice. Error bars show SDs in the weights of five mice from each genotype. (F) H&E staining of AP, VP, and DLP sections from the same mice shown in (D). (G) Upper panel: cell suspensions from mTorpc−/− and WT mouse prostates stained with lineage markers (CD31, CD45, Ter119) and for α6 integrin (CD49f). Gates indicate the lineage-negative, α6 integrin–positive, basal epithelial cell population. Lower panel: representative overlay histogram of mTorpc−/− and WT basal epithelial cell FSC, an indicator of cell size. *P < 0.05. (H) Size of the litters from WT and mTorpc−/− males at 3 months of age. There is no significant difference between the litters from WT and from mTorpc−/− males.

Mice from mTorpc−/− and WT cohorts were killed at different time points (from 3 to 12 months of age) and their prostates were extracted and characterized. We did not observe any macroscopic difference in the prostates of the mTorpc−/− mutants when compared with those of the WT age-matched controls at all time points (Fig. 1D). Additionally, no significant differences in the weights of the prostate lobes were observed (Fig. 1E). Moreover, histopathological analysis revealed that the prostates of the mTorpc−/− mice were similar to those of age-matched WT controls and did not show any overt sign of involution or atrophy (Fig. 1F).

Functionally, the activity of the mTOR pathway is positively linked to cell size (21). We therefore sought to determine whether loss of mTor expression in the prostate of the mTorpc−/− mice resulted in a decrease of prostatic epithelial cell size. To this end, we attempted to define a discrete epithelial cell population with the use of specific markers expressed on the cell surface. In this mouse model, the Cre recombinase needed for mTor excision is driven from the ARR2 Probasin promoter, which drives Cre expression in both the basal and the luminal epithelial cell compartments of the prostate (22). The basal cell compartment of mouse prostate, which is thought to contain the epithelial progenitor cell population, expresses unique sets of both intracellular (cytokeratin5, p63) and cell surface (Sca1, α6 integrin) markers (2224). We confirmed that most basal cells of the AP did indeed coexpress both cytokeratin5 (CK5) and α6 integrin (fig. S1). Using flow cytometry to separate out this discrete population of cells, we identified basal cells from single-cell suspensions of prostate by staining for α6 integrin and nonepithelial “lineage” markers to label contaminating cells (CD31 for endothelial cells, CD45 and Ter119 for hematopoietic cells). Cell size for each population was determined from the forward scatter (FSC) parameter. When we compared the FSC profiles of the basal epithelial cell populations (α6 integrin positive, lineage negative) from mTorpc−/− with that of WT controls, we found that there was a minor but reproducible reduction in the cell size of the cells from mTorpc−/− mutants (average reduction in mean FSC-Height (FSC-H) of 7.5%) (Fig. 1G).

The reproductive capacity of mTorpc−/− was comparable with that of WT animals, as assessed by the size of the litters fathered by males of 3 months of age (Fig. 1H).

Overall, these data show that mTor deletion does not have a major effect on the morphology and function of the prostate, indicating the dispensable nature of mTor in a postmitotic and nonproliferative context.

mTor deletion suppresses Pten-null–driven prostate tumorigenesis more efficiently than does treatment with RAD001

We next aimed to elucidate the effect of mTor inactivation in a mitotically active context. As noted, activation of mTOR is a central aspect of the pathway downstream of the phosphatidylinositol 3-kinase (PI3K)/AKT signaling cascade, which is opposed by PTEN (25, 26). Conditional inactivation of Pten in the mouse prostate triggers prostatic intraepithelial neoplasia (PIN) that later evolves to a nonlethal invasive CaP with full penetrance at 6 months of age (27). We crossed the mTorpc−/− mice with PtenloxP/loxP;PB-Cre4 mice (Ptenpc−/−) and examined the incidence of prostate-specific lesions and invasive CaP in the resulting Ptenpc−/−;mTorpc−/− mice compared to that in age-matched Ptenpc−/− controls (n = 24 mice per cohort, 4 mice per time point). Three-month-old Ptenpc−/−;mTorpc−/− mice showed a reduction in prostate enlargement compared with Ptenpc−/− mice (Fig. 2A). Histopathological analysis of these mice revealed a marked deceleration in tumor initiation in Ptenpc−/−;mTorpc−/− mice as assessed by a decrease in the number of glands affected by PIN (14% of prostate glands analyzed were affected by PIN in Ptenpc−/−;mTorpc−/− mice versus 100% in the Ptenpc−/− mice, four mice per genotype) (Fig. 2B, upper and lower panels).

Fig. 2

Genetic inactivation of mTor suppresses Pten-null–driven prostate tumorigenesis more efficiently than does treatment with RAD001. (A) Representative images of prostates from 3-month-old WT, Ptenpc−/−, and Ptenpc−/−;mTorpc−/− mice. (B) Upper panel: H&E staining of AP sections from 3-month-old Ptenpc−/− and Ptenpc−/−;mTorpc−/− mice. An example of residual PIN lesion in Ptenpc−/−;mTorpc−/− mouse prostate is indicated by the arrow and shown at higher magnification. Lower panel: percentage of glands affected by PIN in the prostate of 3-month-old Ptenpc−/− and Ptenpc−/−;mTorpc−/− mice. Error bars show SDs in the number of glands affected by PIN in four mice from each genotype. **P < 0.01. (C) IHC staining of PIN lesions present in the prostates of Ptenpc−/− and Ptenpc−/−;mTorpc−/− mice shown in (B) (upper panel) with anti–phospho-S6 and anti–phospho-S473 Akt antibodies. Quantification (±SD) of phospho-S6– and phospho-Akt–positive cells in the PIN lesions of Ptenpc−/− and Ptenpc−/−;mTorpc−/− mice is indicated. (D) Upper panel: H&E staining of AP sections from 8-week-old Ptenpc−/− mice treated with vehicle or RAD001 for 4 weeks. One of the residual PIN lesions in RAD001-treated Ptenpc−/− mouse prostate is indicated by the arrow and shown at higher magnification. Lower panel: percentage of glands affected by PIN in the prostate of 8-week-old Ptenpc−/− mice treated with vehicle or RAD001 for 4 weeks. Error bars show SDs in the number of glands affected by PIN in four vehicle-treated and four RAD001-treated Ptenpc−/− mice. *P < 0.05. (E) IHC staining of PIN lesions present in the prostate of the Ptenpc−/− mice treated with vehicle or RAD001 shown in (D) (upper panel) with anti–phospho-S6 and anti–phosphor-S473 Akt.

We performed an immunohistochemical (IHC) analysis of mTORC1 and mTORC2 activity in PIN lesions in the prostates of Ptenpc−/−;mTorpc−/− mice, with the use of the phosphorylated form of the ribosomal protein S6 (phospho-S6), as a readout of mTORC1 activation, and Akt phosphorylated on Ser473, as a readout of mTORC2 activity. The percentage of phospho-Akt– and phospho-S6–positive cells was indistinguishable between the PIN-affected glands of the Ptenpc−/− and the residual PIN lesions of the Ptenpc−/−;mTorpc−/− mice (Fig. 2C, upper and lower panels, respectively). Furthermore, IHC analysis for Pten showed that Pten recombination is similar in the residual PIN lesions of the Ptenpc−/−;mTorpc−/− mice compared with that of Ptenpc−/− mice (fig. S2A), whereas genomic quantitative real-time PCR for mTor on DNA extracted from LCM PIN lesions from the same mice revealed that mTor is not deleted in these lesions (fig. S2B). Taken together, these results suggest that the residual PIN-affected glands in the prostates of the Ptenpc−/−;mTorpc−/− mice arise from Pten-recombined cells that have escaped mTor recombination.

To compare the antitumor activity of preferential inhibition of mTORC1 relative to that of complete mTor inactivation, we treated a cohort of 4-week-old Ptenpc−/− mice with RAD001 (cohorts of four mice per treatment). There was a much less pronounced reduction in the number of PIN-positive glands in the prostates of Ptenpc−/− mice treated for 4 weeks as compared with Ptenpc−/−;mTorpc−/− mice (56% of prostate glands affected by PIN in the RAD001-treated Ptenpc−/− mice versus full penetrance in the vehicle-treated Ptenpc−/− mice; Fig. 2D). Thus, the preferential targeting of mTORC1 by RAD001 is less effective at reducing PIN lesions than is genetic inactivation of mTor, in which the functions of both mTORC1 and mTORC2 are eliminated.

IHC analysis revealed a reduction in phospho-S6 staining of the residual PIN lesions of the RAD001-treated Ptenpc−/− mice compared with that of the PIN-affected glands of the vehicle-treated Ptenpc−/− mice (Fig. 2E, upper panel), whereas the PIN-affected glands of the RAD001-treated Ptenpc−/− mice showed strong Akt phosphorylation on Ser473 (Fig. 2E, lower panel).

Overall, these results imply that mTORC1 inhibition is only partially effective in opposing cancer initiation driven by Pten loss. We found that PIN lesions can develop in the absence of mTORC1 activity in the prostates of RAD001-treated Ptenpc−/− mice, as shown by the lack of S6 phosphorylation in these lesions. Notably, phosphorylation of Akt on Ser473 is maintained in these lesions, indicating that mTORC2 remains active. However, in the prostates of the Ptenpc−/−;mTorpc−/− mice, PIN lesions were never found to display concomitant inactivation of mTORC1 and mTORC2, indicative of the essential nature of both mTOR multiprotein complexes for prostate tumorigenesis in the Ptenpc−/− mice.

mTor inactivation opposes Pten loss–induced tumor progression from in situ to invasive cancer lesions

To further analyze the effect of mTor inactivation on Pten loss–driven CaP, we followed cohorts of Ptenpc−/− and Ptenpc−/−;mTorpc−/− mice up to 12 months of age with monthly magnetic resonance imaging (MRI) analyses. MRI detected the presence of tumors in the prostates of 6-month-old Ptenpc−/− mice [Fig. 3A, left panel (27)]. These tumors were markedly smaller in the Ptenpc−/−;mTorpc−/− cohort than in the Ptenpc−/− mice (Fig. 3A, right panel). Moreover, histopathological analysis revealed that, as previously reported (27), at 6 months of age the Ptenpc−/− mice developed focal signs of invasive CaP at full penetrance. In contrast, prostates from mice with combined inactivation of Pten and mTor showed no signs of invasion (Fig. 3B, four mice per genotype). Additionally, invasive CaPs in Ptenpc−/− mutants exhibited a strong inflammatory response (Fig. 3B, lower left panel) with a marked reaction of the stroma surrounding the epithelial glands, features that were completely absent from the Ptenpc−/−;mTorpc−/− prostates (Fig. 3B). Furthermore, the Ptenpc−/−;mTorpc−/− mice only develop invasive CaP from 10 months of age and at low penetrance (20%) (four mice per genotype were analyzed, P < 0.01).

Fig. 3

mTor inactivation opposes Pten loss–induced tumor progression from in situ to invasive cancer lesions. (A) MRI analysis of prostatic tumors (dashed yellow circles) in Ptenpc−/− and Ptenpc−/−;mTorpc−/− 6-month-old mice. (B) H&E staining of AP sections from 6-month-old Ptenpc−/− and Ptenpc−/−;mTorpc−/− mice. Focal invasion in Ptenpc−/− mouse prostate is indicated by an arrow. Inflammatory infiltration in Ptenpc−/− mouse prostate is shown at higher magnification in the lower left panel. Lower right panel: percentage of Ptenpc−/− and Ptenpc−/−;mTorpc−/− 6-month-old mice with signs of invasion in the prostate. Four mice from each genotyping were analyzed. **P < 0.01.

Together, these findings show that inactivation of mTORC1 and mTORC2 opposes Pten loss–induced tumor initiation and the subsequent progression from in situ to invasive cancer lesions.

mTor inactivation results in growth arrest in a mitotically active context

Because the residual cancer lesions present in the Ptenpc−/−;mTorpc−/− prostates showed signs of having escaped mTor recombination, we hypothesized that Ptenpc−/−;mTorpc−/− double-null cells lack the ability to support neoplastic transformation. We evaluated a number of different parameters in vivo to understand the fate of these cells. IHC analysis of the proliferation marker Ki-67 in prostates from Ptenpc−/−;mTorpc−/− mice showed a reduction in the overall proliferative rate compared with that in Ptenpc−/− mice (Fig. 4A). This proliferation was mainly restricted to the residual PIN lesions (note that the proliferation indices for PIN lesions in prostates from Ptenpc−/− and from Ptenpc−/−;mTorpc−/− mice are similar, as shown by the Ki-67 quantification in the inset of Fig. 4A), indicating that mTor deletion profoundly affects the ability of Pten−/− cells to proliferate and initiate tumorigenesis.

We next determined whether the Ptenpc−/−;mTorpc−/− cells that fail to proliferate engage an apoptotic response. Apoptosis was evaluated in cells in the prostates of WT, mTorpc−/−, Ptenpc−/−, and Ptenpc−/−;mTorpc−/− mice by a standard TUNEL [terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate (dUTP) nick end labeling] assay. No substantial apoptotic response was observed in the normal glands of the prostates of any of these genotypes (Fig. 4B), although a slight incidence of apoptosis (~1%) was observed in PIN lesions, consistent with a higher cellular turnover in the context of tumorigenesis (Fig. 4B).

Ptenpc−/− tumors display a robust p53-dependent senescence response, which opposes tumor progression (20). Abrogation of senescence, as shown in the prostates of Ptenpc−/− mice that also have inactivation of Trp53, allows Pten-null tumors to achieve their full proliferative potential and drives the rapid progression of the PIN lesions to invasive CaP (20). Thus, we investigated whether cellular senescence could account for the slowed progression from in situ neoplastic lesions to invasive cancer in the prostates of the Ptenpc−/−;mTorpc−/− mice. To this end, we analyzed senescence-associated β-galactosidase (SA-β-Gal) activity (28) together with induction of p53 in the Ptenpc−/−;mTorpc−/− mice. At 3 months of age, SA-β-Gal staining in the PIN lesions of Ptenpc−/−;mTorpc−/− prostates was indistinguishable from that in the prostates of the Ptenpc−/− mice (Fig. 4C, upper left and right panels). This result is consistent with the notion that the PIN glands observed in the Ptenpc−/−;mTorpc−/− prostates retain the features of the Pten-null PIN lesions because they have escaped mTor inactivation (Fig. 2C and fig. S2). The similar degree of senescence in both genotypes was confirmed by IHC analysis for p53 (Fig. 4C, lower panel).

In summary, these data suggest that most cells that experience combined Pten;mTor loss in the prostate epithelium of Ptenpc−/−;mTorpc−/− mice undergo a block in proliferation. This markedly decreases the rate of PIN initiation. The cells that lose Pten, while retaining mTor, have the propensity to proliferate more or to stochastically engage a cellular senescence response (20). This in turn opposes the progression of these PIN lesions to invasive CaP (20). The reduction in the overall number of PIN lesions in prostates from the Ptenpc−/−;mTorpc−/− mice explains why invasive lesions are virtually absent at 6 months of age. These PIN lesions arise from Pten-null cells, which retain mTor function through escape from Cre-mediated mTor recombination. As a result, it is likely that these “escaper” PIN lesions maintain the potential to bypass senescence through loss of the p53 response. Indeed, consistent with this interpretation, a few invasive CaP lesions are also observed in Ptenpc−/−;mTorpc−/− mice at later time points (10 months).

Because our in vivo data suggest that mTor deletion has a strong effect in a mitotically active context (Pten loss–driven neoplastic transformation), we sought to characterize the effects of acute mTor inactivation in greater detail in an in vitro model. We infected mTorloxP/loxP, PtenloxP/loxP, and PtenloxP/loxP;mTorloxP/loxP mouse embryonic fibroblasts (MEFs) with a Cre-PURO-IRES-GFP (Cre–puromycin-resistant–internal ribosome entry site–green fluorescent protein)–encoding retrovirus (experimental timeline shown in fig. S3A). The efficiency of Cre-mediated loxP recombination, assessed by Western blot analysis (Fig. 5A) and genomic PCR (fig. S3B), showed elimination of Pten and mTor. It should be noted that the abundance of both phospho-S6 and phospho-Akt is decreased upon mTor deletion in both the Pten-WT and Pten-recombined context, consistent with the functional loss of mTor (Fig. 5A). The remaining phospho-S6 and phospho-Akt may be due to residual mTOR protein remaining after recombination (Fig. 5A).

Fig. 4

Biological outcome of mTor deletion in the WT and Ptenpc−/− prostate. (A) Left panel: Ki-67 staining on prostate sections from 3-month-old mice of the indicated genotypes. Right panel: quantification of Ki-67. Pos. stands for positive. Three different areas of one section were counted to determine an average and a representative value for each slide. Slides from three independent mice were counted in this way to determine a standard deviation for the population. *P < 0.05. The inset shows the Ki-67 quantification exclusively in the PIN lesions of the Ptenpc−/− and Ptenpc−/−;mTorpc−/− mice prostates. (B) TUNEL assay for apoptosis on prostate sections from 3-month-old mice of the indicated genotypes. The arrows show examples of TUNEL-positive cells. The inset in the Ptenpc−/−;mTorpc−/− prostate section shows the TUNEL staining in a residual PIN lesion. (C) Upper left panel: senescence-associated β-galactosidase staining (SA-β-Gal) on prostate sections from 3-month-old mice of the indicated genotypes. Upper right panel: quantification of the SA-β-Gal staining in the glands affected by PIN of the prostate sections from 3-month-old mice of the indicated genotypes. Quantifications were done on three representative sections from three mice. There is no significant difference between the PIN lesions of Ptenpc−/− and Ptenpc−/−;mTorpc−/− mice prostates. Lower panel: IHC for p53 on the same mice shown in the upper panel.

Fig. 5

Biological outcome of mTor deletion in MEFs. (A) Western blot analysis of lysates of mTorloxP/loxP, PtenloxP/loxP, and PtenloxP/loxP;mTorloxP/loxP MEFs infected with PURO-IRES-GFP (vector) or Cre-PURO-IRES-GFP (Cre) (see experimental timeline shown in fig. S3A). Quantification of the phospho-Akt/total Akt ratio is shown. Quantifications were done by densitometry analysis performed with ImageJ software. For each genotyping, the phospho-Akt/total Akt value of Cre-infected cells is normalized against the corresponding vector-infected control. (B) Flow cytometric analysis (forward scatter, FSC-H) of the mTor null (Pten+/+;mTorΔ/Δ-Cre) and Pten;mTor double-null MEFs (PtenrΔ/Δ;mTorΔ/Δ-Cre) compared to the WT (Pten+/+;mTor+/+-Cre). (C) Cell proliferation curve analysis of the same MEFs analyzed in (B) followed over a 6-day period. (D) Cell proliferation curve analysis of mTorloxP/loxP and PtenloxP/loxP;mTorloxP/loxP primary MEFs first immortalized with SV40 large T antigen and subsequently infected with PURO-IRES-GFP (vector) or Cre-PURO-IRES-GFP (Cre). (E) TUNEL assay on the same MEFs analyzed in (B). *P < 0.05. (F) Flow cytometric analysis of mTorloxP/loxP and PtenloxP/loxP;mTorloxP/loxP MEFs infected with PURO-IRES-GFP (vector) or Cre-PURO-IRES-GFP (Cre) (see experimental timeline shown in fig. S3A). To accurately evaluate cell cycle populations, data were gated to exclude the sub-G1 population. PE-A, phycoerythrin channel area. (G) Western blot analysis on the lysates of mTorloxP/loxP and PtenloxP/loxP;mTorloxP/loxP MEFs infected with PURO-IRES-GFP (vector) or Cre-PURO-IRES-GFP (Cre) (see experimental timeline shown in fig. S3A).

Flow cytometric analysis of the mTor-null (Pten+/+;mTorΔ/Δ-Cre) and mTor;Pten double-null MEFs (PtenrΔ/Δ;mTorΔ/Δ-Cre) revealed that both cell types display decreased cell size compared to that of WT (Pten+/+;mTor+/+-Cre) MEFs, as measured by the FSC parameter (Fig. 5B).

In addition, cell growth analysis revealed that mTor deletion affects the proliferative capability of MEFs both in WT (Pten+/+;mTorΔ/Δ-Cre) and in Pten-null (PtenrΔ/Δ;mTorΔ/Δ-Cre) genetic backgrounds (Fig. 5C). These results corroborate our in vivo data and highlight the fact that mTor deletion has a marked effect on mitotically active primary cells. We next immortalized primary MEFs with simian virus 40 (SV40) large T antigen and subsequently inactivated mTor alone or mTor and Pten together in these cells. Immortalization acts as a primary oncogenic event, which confers a strong proliferative advantage even in a scenario in which primary cells would normally undergo growth arrest. As shown in Fig. 5D, SV40-Pten+/+;mTorΔ/Δ-Cre–infected cells show complete inhibition of proliferation, indicating that mTor deletion opposes the strong proliferative potential bestowed on these cells by the SV40 large T antigen. Moreover, even in the context of a second oncogenic hit, the inactivation of Pten, mTor deletion still acts to block proliferation (SV40-PtenΔ/Δ;mTorΔ/Δ-Cre in Fig. 5D).

We next sought to analyze cell death in the Pten+/+;mTorΔ/Δ-Cre and PtenΔ/Δ;mTorΔ/Δ-Cre MEFs. TUNEL assays revealed a significant increase in the apoptotic index of Pten+/+;mTorΔ/Δ-Cre and PtenΔ/Δ;mTorΔ/Δ-Cre MEFs compared with that of WT MEFs (Fig. 5E), consistent with the idea that a prolonged growth arrest may result in an increase in cell death.

Cell cycle status was analyzed in the same cells by flow cytometric analysis (Fig. 5F) and by Western blot utilizing cell cycle–specific markers (Fig. 5G). Results from the flow cytometry revealed that mTor deletion in both the WT (Pten+/+;mTorΔ/Δ-Cre) and the Pten-null (PtenΔ/Δ;mTorΔ/Δ-Cre) genetic backgrounds leads to G1 arrest (Fig. 5F). This G1 arrest is associated with a down-regulation of cyclin D1, a well-characterized cell cycle marker that is translationally regulated by mTOR (Fig. 5G) (29). An analysis of cyclin-dependent kinase inhibitors revealed that although there was no detectable change in p21Cip1/Waf1 (Fig. 5G), there was an increase in p27Kip1 abundance (Fig. 5G). We also observed enhanced nuclear compartmentalization of p27Kip1 upon mTor deletion (fig. S3C), consistent with its function in regulating the cell cycle (30).

Taken together, these results reinforce the concept that the functionality of mitotically active cells depends on the activity of mTor, whose deletion impairs tumor initiation and progression in a Pten loss–driven model of CaP. Our findings indicate that acute inactivation of mTor in vitro results in G1 growth arrest and a modest increase in apoptosis in the Pten+/+;mTorΔ/Δ-Cre and PtenΔ/Δ;mTorΔ/Δ-Cre MEFs.

Discussion

Overall, our results are consistent with a model in which the consequences of mTor deletion depend on the proliferative context of the cell. It is clear from previous knockout models that mTor plays a crucial role in the developing embryo (13, 14). Here, we show the consequences for mTor deletion in the adult prostate of the mouse. In contrast with what has been previously observed in the embryo, mTor had only minor effects on the adult mouse prostate, suggesting that mTor is dispensable in the postmitotic setting.

However, our study of mTor inactivation in Pten loss–driven prostate tumorigenesis showed the pivotal role of this protein in preserving the proliferative capability of mitotically active cells. Loss of mTor in this scenario markedly reduced the potential of Pten-null cells to initiate tumor formation. The essential requirement of mTor for tumorigenesis is also consistent with observations from mTor knockout models, given the high mitotic index and proliferative demands of the developing embryo.

In vitro analysis in MEFs confirms the concept that inactivation of mTor impairs cell proliferation so profoundly that it cannot even be rescued by immortalization of these cells. Additionally, we show that acute inactivation of mTor in these MEFs results in a combined G1 arrest and apoptosis.

The in vivo effect of inactivation of both mTORC1 and mTORC2 on prostate tumorigenesis driven by Pten loss is much more pronounced than that observed upon treatment of Ptenpc−/− mice with the mTORC1 inhibitor, RAD001.

In a mouse model of PIN generated through overexpression of a constitutively active myr-Akt transgene in the prostate (31), RAD001 treatment completely reverts the PIN phenotype (32). In contrast, RAD001 treatment of our Ptenpc−/− mice, which show a phenotype more pronounced than that of the myr-Akt transgenic mice, is only partially effective. These data also emphasize that the type of murine preclinical model used is critical for the extrapolation of the therapeutic potential of compounds studied. Furthermore, comparing results from mTORC1 inhibition by RAD001 and concomitant abrogation of mTORC1 and mTORC2 in Ptenpc−/− mice, we conclude that mTORC2 may play an essential role in prostate tumorigenesis driven by Pten loss.

The critical role of mTORC2 in tumorigenesis is also highlighted by recent findings of Masri et al., showing that elevated mTORC2 activation, through overexpression of rictor, is observed in glioma cell lines and primary tumor cells (33).

Thus, these data strongly support the rationale for the development of inhibitors that target the catalytic subunit of mTOR. It is tempting to speculate that these inhibitors, through the targeting of a wider spectrum of mTOR downstream effectors, will hold added therapeutic potential over rapalogs, the first generation of mTOR inhibitors, which are currently being tested in several clinical trials. However, the differential outcome of mTOR abrogation in proliferating and nonproliferating tissues shown here warrants further studies to establish a dosage that will minimize any potential toxicity of these compounds on rapidly proliferating cells.

Materials and Methods

mTor and Pten mice

Previously generated mTorloxP/loxP and PtenloxP/loxP mice (13, 27) were crossed with the PB-Cre4 mice (19). All work with mice was performed in accordance with our Institutional Animal Care and Use Committee–approved protocol. For genotyping, tail DNA was subjected to PCR following the protocols previously described (13, 27). To detect the presence of the recombined mTorΔ allele, PCR analysis was carried out on DNA extracted from several tissues and the previously reported primers were used (13).

Quantitative mRNA RT-PCR

Total RNA was prepared from mice prostates with the Trizol method (Invitrogen). Complementary DNA was obtained with Transcriptor (Roche). A TaqMan probe specific for the 5 to 6 exon boundary on mTor (note that exon 5 is deleted upon mTor recombination) was obtained from Applied Biosystems (Foster City, CA). Amplifications were run in a 7900 Real-Time PCR System (Applied Biosystems). Each value was adjusted with glucuronidase B levels as reference.

LCM and genomic quantitative real-time PCR

LCM was performed on 5-μm frozen sections with a Veritas Microdissection Instrument (Arcturus Bioscience, Mountain View, CA). Sections were cut from Tissue-Tek O.C.T.-embedded samples on PEN Membrane Glass Slides (Arcturus Bioscience), then fixed for 10 min in methanol, washed for 5 min with phosphate-buffered saline (PBS), counterstained with hematoxylin, and dehydrated through washes with 95% ethanol, 100% ethanol, and xylene. Sections were finally air-dried for 1 to 2 hours. LCM prostatic epithelial cells were collected on CapSure HS LCM Caps (Arcturus Bioscience). DNA was extracted with QIAamp DNA Micro Kit (Qiagen) according to the manufacturer’s instructions. Quantitative real-time PCR was performed on a Roche LightCycler using the Quantitect SYBR Green PCR kit (Qiagen) and the following amplification protocol: 15 min at 95°C, 40 cycles (15 s at 94°C, 20 s at 57°C, 20 s at 72°C), followed by determination/confirmation of amplicon melting temperature. Reactions were performed in triplicates; primer pairs were confirmed to yield a single amplicon band by 3% agarose gel electrophoresis. The following mouse primer set was used to amplify specifically a 185-bp amplicon on exon 2 of mTor, which is deleted upon recombination: forward primer, TCTGTGCACATCTTCCTTGC; reverse primer, TGCTACCAGAGGCTGTCCTT. The following primers for β-actin were used as reference standards: forward primer, GGCTGTATTCCCCTCCATCG; reverse primer, CCAGTTGGTAACAATGCCATGT.

Flow cytometry

To generate a single-cell suspension from mouse prostate, the lobes (AP, VP, DLP) were dissected, and minced, and cells were extracted essentially as described (34). After cells were passed sequentially through 70- and 40-μm cell strainers to remove debris and cell clumps, cells were resuspended in PBS supplemented with 2% fetal bovine serum. Cells were then incubated with the following antibodies: biotin-conjugated antibody to CD31 (BD Bioscience), CD45, and Ter119 (eBioscience) plus Alexa647-conjugated antibody to CD49f (α6 integrin, clone GoH3; Biolegend). Cells were washed and then incubated in phycoerythrin-conjugated streptavidin for the detection of biotin-labeled antibodies. DAPI (4′,6′-diamidino-2-phenylindole) was used for live-dead discrimination and stained cells were analyzed on an LSRII flow cytometer (BD).

Western blot analysis and immunohistochemistry

Cell lysates were prepared with radioimmunoprecipitation assay buffer [1× PBS, 1% Nonidet P40, 0.5% sodium deoxycholate, 0.1% SDS, and protease inhibitor cocktail (Roche)] and cleaned by centrifugation. The following antibodies were used for Western blot analysis: mouse antibodies to β-actin (AC-72; Sigma), mTOR (Cell Signaling Technology), Pten (Cell Signaling Technology), cyclin D1 (BD Pharmingen), cyclin B1 (Cell Signaling Technology), p21 (C-19; Santa Cruz), and p27 (Upstate).

For immunohistochemistry, prostate tissues were fixed in 10% neutral-buffered formalin (Sigma) overnight, subsequently washed once with PBS, transferred into 50% ethanol, and stored in 70% ethanol. Prostate lobes were embedded in paraffin, sectioned, and stained with hematoxylin and eosin (H&E) in accordance with standard procedures. Sections were stained with the following antibodies: antibodies to phospho-S6 (S235/S236) (Cell Signaling Technology), phospho-Akt (S473) (Cell Signaling Technology), Pten (Ab-2; NeoMarkers), Ki-67 (Novacastra), (Nanotools), and p53 (FL-393; Santa Cruz).

Treatment with RAD001

RAD001 (everolimus) was obtained from Novartis and was administered by oral gavage at 10 mg/kg per day. The treatment started when the mice were 4 weeks old, before PIN onset, and was maintained for 4 weeks.

MRI

Individual mice were subjected to MRI analysis to detect the presence of prostatic tumors. The analysis was carried out as previously described (20, 27).

TUNEL assay and senescence detection in vivo

For TUNEL assay, sections were pretreated with proteinase K (Sigma Aldrich). The sections were then incubated with terminal deoxynucleotidyl transferase (TdT, Roche Diagnostics) and biotion 16-dUTP (Roche Diagnostics). Detection was performed with horseradish peroxidase–streptavidin conjugate (1:100) and developed with DAB Peroxidase Substrate kit (SK-100, Vector Laboratories). After the incubation, the slides were counterstained with hematoxylin and mounted in Permount. For detection of senescence in vivo, prostates were frozen in Tissue-Tek O.C.T. and frozen sections 5 μm thick were stained for β-galactosidase activity with the Senescence Detection Kit (Calbiochem). Quantifications were performed on three representative sections from three mice for each genotype.

MEF generation, infection, and analysis

Primary MEFs were prepared as described previously (35) and infected with a retrovirus expressing Cre-PURO-IRES-GFP or the corresponding empty control. Virus production and MEF infection were performed as previously described (20). Two days after selection, cells were plated for growth curves, flow cytometric analysis, TUNEL, and Western blot as reported (20). TUNEL analysis was performed with the in situ cell death detection Kit (Roche). To obtain cell cycle profiles by flow cytometry, cells were trypsinized, harvested, and fixed and permeabilized in ice-cold 70% ethanol. After incubation on ice for 30 min or overnight at −20°C, fixed cells were washed twice with PBS and resuspended in propidium iodide/ribonuclease solution (BD Bioscience). Cells were kept at 4°C overnight before analysis on a BD LSRII flow cytometer.

Statistical analysis

Statistical evaluations were carried out with SigmaStat 2.03 (SPSS). For all analyses, the level of statistical significance was set at P < 0.05 (indicated with one asterisk on the graphs) or P < 0.01 (indicated with two asterisks on the graphs). Unpaired Student’s t test was performed to determine the statistical significance.

Acknowledgments

We thank R. Bernardi, M. S. Song, and all the members of the Pandolfi laboratory for discussion and comments. We are grateful to the Beth Israel Deaconess Medical Center pathology core for technical expertise in antibody characterization and immunohistochemistry. We are grateful for the MRI imaging work by MR small-animal imaging core at Memorial Sloan-Kettering Cancer Center, which is partially supported by Small Animal Imaging Resource Program (R24CA83084) and an NIH Center Grant (P30 CA08748). We thank Novartis for RAD001. C.N. was supported in part by a fellowship from the American-Italian Cancer Foundation. A.C. was supported by a European Molecular Biology Laboratory long-term fellowship. This work was supported by NIH/National Cancer Institute grants (U01 CA-84292; R01 CA-82328) to P.P.P. C.N., A.C., A.A., R.M.H, and P.P.P. designed and conceived the experiments. C.N., A.C., A.A., R.M.H., and Z.C. performed the experiments. C.N., J.G.C., A.C., and P.P.P. wrote the manuscript.

Supplementary Materials

www.sciencesignaling.org/cgi/content/full/2/55/ra2/DC1

Materials and Methods

Fig. S1. α6 integrin is predominantly expressed by cytokeratin 5 (CK5)–positive basal prostatic epithelial cells.

Fig. S2. Residual PIN-affected glands in Ptenpc−/−;mTorpc−/− mice arise from Pten-recombined cells that have escaped mTor recombination.

Fig. S3. Deletion of mTor in MEFs.

References and Notes

View Abstract

Navigate This Article