Increased AKT Activity Contributes to Prostate Cancer Progression by Dramatically Accelerating Prostate Tumor Growth and Diminishing p27Kip1 Expression*

The PTEN tumor suppressor gene is frequently inactivated in human prostate cancers, particularly in more advanced cancers, suggesting that the AKT/protein kinase B (PKB) kinase, which is negatively regulated by PTEN, may be involved in human prostate cancer progression. We now show that AKT activation and activity are markedly increased in androgen-independent, prostate-specific antigen-positive prostate cancer cells (LNAI cells) established from xenograft tumors of the androgen-dependent LNCaP cell line. These LNAI cells show increased expression of integrin-linked kinase, which is putatively responsible for AKT activation/Ser-473 phosphorylation, as well as for increased phosphorylation of the AKT target protein, BAD. Furthermore, expression of the p27Kip1 cell cycle regulator was diminished in LNAI cells, consistent with the notion that AKT directly inhibits AFX/Forkhead-mediated transcription of p27Kip1. To assess directly the impact of increased AKT activity on prostate cancer progression, an activated hAKT1 mutant was overexpressed in LNCaP cells, resulting in a 6-fold increase in xenograft tumor growth. Like LNAI cells, these transfectants showed dramatically reduced p27Kip1 expression. Together, these data implicate increased AKT activity in prostate tumor progression and androgen independence and suggest that diminished p27Kip1expression, which has been repeatedly associated with prostate cancer progression, may be a consequence of increased AKT activity.

The molecular alterations that facilitate human prostate cancer (hPCa) 1 progression and the emergence of androgenindependent tumor cells are unclear but may involve a progressive decrease in the normal apoptotic response (1-3) as well as a release from the cell cycle arrest that follows androgen withdrawal (4). Indeed, hPCA progression has been associated repeatedly with decreased expression of the cell cycle regulator p27 Kip1 (4 -9).
The PTEN/MMAC tumor suppressor gene is frequently inactivated in primary human prostate cancers, particularly in the more advanced cancers (10), and in human prostate cancer xenografts and cell lines including PC-3, Du145, and LNCaP (11)(12)(13). These studies suggest that components of the phosphatidylinositol 3-kinase pathway that are negatively regulated by PTEN, such as the key cell survival kinase AKT (14 -16), may be increasingly activated with prostate tumor progression. Indeed, activated AKT regulates a number of intracellular targets implicated in prostate tumor progression and androgen independence. For instance, AKT-dependent inactivation of pro-apoptotic proteins such as BAD and caspase-9 (17,18) may suppress the normal apoptotic response. Additionally, AKT may enhance cell cycle progression by suppressing AFX/Forkhead transcription factor activity (19 -22), which would result in diminished expression of AFX target genes such as the cell cycle inhibitor p27 Kip1 (23). Furthermore, AKT can elicit enhanced translation of key growth-regulatory proteins, like cyclin D1 (24), by stimulating FRAP/mTOR kinase activity and de-repressing translation initiation (Refs. 25 and 26; depicted in Fig. 1).
Because AKT is a central regulator of many intracellular processes implicated in prostate tumor progression and because PTEN, the negative regulator of AKT, is functionally inactivated in a significant proportion of advanced hPCa, we explored whether increased AKT activity may be involved in prostate tumor progression and androgen independence. To better model androgen-independent hPCa, we established a novel series of AR ϩ , PSA ϩ , androgen-independent hPCa cells (LNAI cells) derived from xenograft tumors of the androgensensitive/dependent LNCaP prostate cancer cell line. Our analyses of these cells reveals that AKT activity is increased with androgen-independent progression corresponding to increased expression of integrin-linked kinase, the kinase putatively responsible for the activation of AKT by phosphorylation at Ser-473 (Refs. 27 and 28; Fig. 1), increased phosphorylation of the pro-apoptotic protein BAD and markedly reduced expression of the cell cycle regulator p27 Kip1 . In addition, overexpression of an activated AKT-1 cDNA in LNCaP cells dramatically accelerates xenograft tumor growth and suppresses p27 Kip1 expression. These data implicate increased AKT activity in human prostate cancer progression and androgen independence and * 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  suggest that the reduction in p27 Kip1 expression (4 -9), which has been routinely associated with prostate tumor progression, may be a consequence of increased AKT activity.

EXPERIMENTAL PROCEDURES
Cell Culture and Establishment of Cell Lines ex Vivo-Cells were cultured in phenol red-free RPMI 1640 media (Life Technologies, Inc.) supplemented with 10% fetal bovine serum (Hyclone, Logan, UT). All xenograft studies were performed by injecting 4 ϫ 10 6 cells subcutaneously in 100 l of a 1:1 mix of 1ϫ phosphate-buffered saline and Matrigel (Becton Dickinson, Bedford, MA) into castrated male nude mice (Harlan Sprague-Dawley nu/nu) or into intact male nude mice supplemented with 90-day release testosterone pellets (Innovative Research, Tampa, FL). First-generation LNAI cell lines, LNAI.T1 and LNAI.T2, were derived from a xenograft tumor of LNCaP established in an intact mouse that continued to grow following castration. Second generation LNAI cells, designated T1.8, T1.16, T2.9, T2.10, and T2.11, were established from individual xenograft tumors of LNAI.T1 and LNAI.T2, respectively, growing in castrated mice. To evaluate tumor formation and growth in the presence or absence of circulating androgens, these secondary LNAI cell lines were also then injected into intact and castrated male nude mice. Tumor volumes were calculated by the equation A 2 ϫ B ϫ 0.4, where A is the smallest tumor diameter, B is the largest tumor diameter, and 0.4 is a correction factor (29).

Establishment of AR ϩ , PSA ϩ Androgen-independent Human
Prostate Cancer Cell Lines-The majority of androgen-independent hPCa express both the androgen receptor and prostate-specific antigen (33). The most widely available and commonly used androgen-independent human prostate cancer cell lines, PC-3 and Du145, lack expression of both AR and PSA (34) and therefore may not reflect the majority of advanced, androgen-independent hPCa. To better model androgen-independent prostate cancer, we developed a panel of cell lines from the AR ϩ , PSA ϩ , androgen-dependent/sensitive hPCa cell line, LN-CaP, following in vivo selection for xenograft tumor growth persisting after castration. These cell lines, designated LNAI cells, readily formed tumors both in intact, testosterone-supplemented male nude mice and in castrated male nude mice (Table I), indicating that these cell lines are androgen-independent. Notably, the second-generation LNAI cells, T1.8, T1.16, T2.11, formed tumors more rapidly in both intact and castrated hosts than the first-generation LNAI cells, LNAI.T1 and LNAI.T2. All of these LNAI cells expressed PSA (Fig. 2) and AR (data not shown). PSA was also evident in the serum from LNAI xenograft-bearing intact and castrated mice (data not shown). Therefore, we have successfully established androgenindependent hPCa cells that, like the majority of androgenindependent hPCa, retain expression of both AR and PSA and share the same genetic background as the androgen-depend-FIG. 1. AKT activation and signaling pathways. As a downstream consequence of phosphatidylinositol 3,4,5-triphosphate (PIP 3 ) generation by phosphatidylinositol 3-kinase (PI-3 kinase), AKT is activated via phosphorylation at Thr-308 by phosphatidylinositol-dependent kinase-1 (PDK-1) and at Ser-473 by phosphatidylinositol-dependent kinase-2 (PDK-2), which may be ILK (27,28). The PTEN lipid phosphatase degrades phosphatidylinositol 3,4,5-triphosphate , thereby negatively regulating AKT activation. PTEN loss abrogates this negative regulation (15). Upon activation, AKT can suppress the pro-apoptotic functions of BAD, via Ser-136 phosphorylation, and caspase 9 (17,18). AKT can also phosphorylate and inactivate the AFX/Forkhead (AFX/FKHR) family of transcription factors (19 -22) which in turn suppresses AFX-mediated transcription of target genes such as the p27 Kip1 cell cycle regulator (23). AKT can specifically enhance translation of key genes such as cyclin D1 (24) by activating the FRAP/mTOR kinase and de-repressing mRNA translation initiation (25,26). ent/sensitive hPCa cell line, LNCaP. As such, differences between LNAI and LNCaP cells can be related to androgen independence and tumor progression without being complicated by differing genetic backgrounds.
AKT Activation and Activity Are Increased in the Androgenindependent LNAI Cells-Because AKT is a key regulator of many intracellular processes that have been implicated in prostate cancer progression and androgen independence (1, 25), we examined AKT expression and activation in androgen-independent LNAI cells relative to androgen-dependent LNCaP cells. Although AKT expression levels were unchanged (Fig.  3A), AKT activation (Ser-473 phosphorylation) was markedly increased in LNAI cells relative to LNCaP, particularly in the more aggressively growing second-generation LNAI cells, T1.8, T1.16, and T2.11, (Fig. 3A) and in xenografts (data not shown). Phosphorylation of AKT at Thr-308, which is also required for activation, was not increased in LNAI cells (data not shown). Moreover, AKT activity was also increased in LNAI cells relative to LNCaP cells (Fig. 3B). Together, these data indicate that AKT activation (Ser-473 phosphorylation) and activity are increased in androgen-independent LNAI cells, particularly the more aggressively growing, second-generation lines (T1.8, T1.16, and T2.11), relative to androgen-dependent LNCaP cells.
Androgen-independent LNAI Cells Express AKT-1 and AKT-2 but not AKT-3-Three isoforms of human AKT have been isolated, AKT-1, AKT-2, and AKT-3 (25,35). Expression of AKT-3 has been specifically associated with hormone-independent primary human breast cancers and breast cancer cell lines. Likewise, the androgen-independent human prostate cancer cell lines PC-3 and Du145 express AKT-3, whereas the androgen-dependent/sensitive LNCaP hPCa cell line expresses only AKT-1 and AKT-2, suggesting that androgen-independent hPCA progression may involve AKT-3 induction (32). We therefore examined whether the increased AKT activity in LNAI cells might reflect AKT-3 expression. RT-PCR analyses with primers designed specifically for each AKT isoform (32) revealed that, like LNCaP cells, LNAI cells express only AKT-1 and AKT-2 (Fig. 4A), suggesting that the evolution of androgen-independent hPCa cells does not necessarily involve upregulated AKT-3 expression. To evaluate further whether AKT-3 expression is induced during hPCa development or progression, we analyzed primary human prostate cancers, prostate cancer metastases to bone and soft tissues, and normal prostate tissue for expression of AKT-1 and AKT-3 (Fig. 4, B  and C). Our analyses reveal that, although not all tissues express AKT-3, the majority of both normal and tumor tissue shows AKT-3 expression, thereby indicating that AKT-3 ex-  a These data were pooled from two experiments. LNCaP tumors did not grow to 1300 mm 3 within 100 days-post injection. N/A, not applicable. pression is not induced in, or restricted to, malignancy.
LNAI Cells Overexpress Integrin-linked Kinase-Integrinlinked kinase (ILK) has been putatively identified as responsible for AKT Ser-473 phosphorylation (27,28). Western blot analyses revealed that ILK is overexpressed in LNAI cells and xenografts (Fig. 5), particularly in the second-generation LNAI cells in which AKT activation was highest (e.g. T1.16, Fig. 5), suggesting that the enhanced AKT Ser-473 phosphorylation may be a downstream consequence of up-regulated ILK expression.
Increased AKT Activity in LNAI Cells Corresponds to Increased BAD  Phosphorylation and Decreased p27 Kip1 Expression-Because AKT activity was increased in androgenindependent LNAI cells, we addressed the question of whether molecules downstream from AKT were differentially affected in LNAI cells compared with LNCaP. The pro-apoptotic protein BAD can be functionally inactivated by phosphorylation at Ser-136 by AKT (17,36,37). Xenograft tissue and cell culture extracts of LNAI showed increased phosphorylation of BAD at the AKT target site, Ser-136, relative to LNCaP cell and xenograft extracts (Fig. 6).
AKT has recently been shown to phosphorylate and inactivate the Forkhead/AFX family of transcription factors (19 -22). As AFX activates transcription of the cell cycle regulator p27 Kip1 (23), inhibition of AFX-mediated transcription by AKT may lead to decreased expression of p27 Kip1 . In LNAI cells, in which AKT is hyperactivated, the protein expression of p27 Kip1 was markedly reduced relative to that in LNCaP (Fig. 7). This decrease was most pronounced in the most aggressively growing, second-generation LNAI cells with the highest AKT activity (e.g. T1.16). These data are consistent with the notion that AKT negatively regulates AFX-mediated transcription of p27 Kip1 but do not exclude the possibility that other mechanisms, such as decreased protein or RNA stability, may contribute to p27 Kip1 loss. Together, these data indicate that the increased AKT activity in androgen-independent LNAI cells profoundly effects key downstream mediators involved in cell cycle arrest and apoptotic signaling.
Stable Overexpression of Constitutively Active AKT Dramatically Enhances LNCaP Xenograft Tumor Growth in Intact Male Nude Mice-Our data indicate that AKT activity is mark-edly enhanced in androgen-independent LNAI cells compared with androgen-dependent/sensitive LNCaP and suggest that increased AKT activity may be intimately involved in the progression of hPCa and the emergence of androgen-independent hPCa cells. To determine directly how AKT may impact upon hPCa progression, we established stable overexpressors of the constitutively active AKT-1 T308D/S473D mutant in LNCaP cells (LNCaP:AKT DD cells) (Fig. 8, inset). Subcutaneous injection of these LNCaP:AKT DD cells into intact, testosterone-supplemented male nude mice yielded rapidly growing tumors with 100% incidence (4/4) in less than 30 days post-injection, whereas vector control cells formed tumors in only three of five mice that were not evident until day 37 post-injection (Fig. 8). Moreover, the LNCaP:AKT DD cells showed a dramatic increase in tumor growth (Fig. 8), with a mean tumor volume more than 6-fold that of the vector control transfectants at all time points (by Dunnett's test on ranked tumor volumes, p Ͻ 0.05 for all time points). Like the aggressively growing LNAI cells, LNCaP: AKT DD tumor cells showed a substantial decrease in p27 Kip1 expression (Fig. 8, inset). Tumors did not form in castrated mice following injection with either LNCaP:AKT DD or vector control LNCaP cells. A second, independent experiment confirmed that LNCaP:AKT DD cells formed more rapidly growing tumors in intact male nude mice when compared with vector control LNCaP cells but were unable to form tumors in castrated male nude mice (data not shown). Therefore, although insufficient to elicit growth in castrated male nude mice, overexpression of the constitutively active AKT DD dramatically enhanced LNCaP xenograft tumor growth in intact male nude mice and resulted in markedly reduced expression of p27 Kip1 . DISCUSSION We have developed a series of androgen-independent hPCa cell lines (LNAI cells) from the androgen-dependent/sensitive hPCA cell line, LNCaP. Like the majority of androgen-independent hPCa (34), these LNAI cells retain expression of AR and PSA. We have now shown that AKT activation (Ser-473 phos- phorylation) and activity are increased in these LNAI cells when compared with androgen-dependent LNCaP cells and may be a consequence of increased expression of the integrinlinked kinase, the kinase putatively responsible for AKT Ser-473 phosphorylation (27,28). Like parental LNCaP cells, LNAI cells express only AKT-1 and AKT-2, indicating that AKT-3 expression is not necessary for androgen independence, as has been suggested in studies with the androgen-independent PC-3 and Du145 hPCa cells (32). Moreover, our data reveal that AKT-3 is expressed not only in prostate cancer tissues but also in normal prostate tissue, indicating that AKT-3 expression is not induced in, or restricted to, prostate malignancy. We have also shown that the increased AKT activity in these LNAI cells corresponds to increased phosphorylation of the pro-apoptotic protein, BAD, and markedly decreased expression of the p27 Kip1 cell cycle regulator, consistent with the notion that AKT inhibits AFX-mediated transcriptional activation of p27 Kip1 (23,38). Taken together, these data imply increased AKT activity in hPCa progression and the emergence of androgen-independent human prostate cancer cells.
To address whether AKT alone can drive prostate tumor progression and/or the emergence of androgen-independent cells, we established LNCaP cells with stable overexpression of a constitutively active AKT-1 cDNA. Though alone insufficient to elicit xenograft tumor growth in castrated mice, overexpression of activated AKT dramatically increased tumor growth in intact mice relative to vector control transfectants. Like the most aggressively growing LNAI cells (e.g. T1.16), these LN-CaP:AKT DD cells also showed a remarkable reduction in p27 Kip1 expression, suggesting that p27 Kip1 loss, which has been repeatedly linked to hPCa progression (4 -9), may be a consequence of increased AKT activity. Taken together with the analyses of LNAI cells, these data reveal that increased AKT activity contributes to hPCa progression by dramatically accelerating prostate tumor growth, perhaps in relation to decreasing p27 Kip1 expression, and is specifically associated with the more advanced, aggressively growing androgen-independent hPCa cells (e.g. T1.16).
Increased AKT activity may play a profound role in the progression of human prostate cancers. AKT regulates many of the processes associated with metastatic progression and the emergence of androgen-independent hPCa cells, such as a diminished apoptotic response (1, 2, 39) as well as a release from the cell cycle control that follows androgen ablation (4). AKT can dampen the normal apoptotic response by suppressing the activity of numerous pro-apoptotic proteins including BAD, caspase 9, and the Forkhead family of transcription factors (Refs. 17-22; summarized in Fig. 1). AKT can also facilitate the release of cells from cell cycle control by inhibiting AFX/Forkhead-mediated transcription of the key cell cycle regulator, p27 Kip1 (23). Further, by activating FRAP/mTOR and de-repressing translation initiation, AKT may specifically enhance the translation of cyclin D1 (24) as well as other translationally controlled growth factors and growth regulatory proteins implicated in human prostate cancer progression (Refs. 40 -43; see Fig. 1).
The data in this report clearly imply increased AKT activity in hPCa progression. The most aggressively growing, androgen-independent LNAI cells had the greatest increase in AKT activity. Moreover, overexpression of an activated AKT-1 mutant dramatically accelerated LNCaP xenograft tumor growth. However, our data also reveal that overexpression of an activated AKT-1 was alone insufficient to drive xenograft tumor growth in castrated mice, suggesting that factors in addition to increased AKT activity must contribute to the ability of LNAI cells to form and grow tumors in the absence of circulating androgens. For instance, the increased expression of ILK in LNAI cells may influence alternate pathways in addition to the AKT pathway, such as the Wnt pathway (27), that may contribute to the androgen-independent phenotype. Moreover, other intracellular pathways may be activated in LNAI cells independent of, or upstream from, AKT that contribute to the androgen-independent phenotype of LNAI cells. Indeed, activation of the mitogen-activated protein kinase pathways, perhaps as a downstream consequence of HER2 activation, has been implicated in hPCa progression and androgen independence (44 -46).
In summary, the data presented in this report indicate that increased AKT activity, perhaps as a consequence of increased ILK expression, is involved in hPCa progression and suggest that the reduction in p27 Kip1 expression, which has been repeatedly associated with hPCa progression and androgen independence (4 -9), may be a consequence of enhanced AKT activity. Together with previously published reports that PTEN is inactivated in a significant proportion of advanced primary human prostate cancers (10) and that AKT regulates many of the intracellular processes associated with hPCa progression (1,25), the data in this report implicate AKT as a key mediator of prostate tumor progression and suggest that AKT may be a prime target for prostate cancer therapy.