AKT upregulates B-Raf Ser445 phosphorylation and ERK1/2 activation in prostate cancer cells in response to androgen depletion

Abstract

Upregulated ERK1/2 activity is often correlated with AKT activation during prostate cancer (PCa) progression, yet their functional relation needs elucidation. Using androgen-deprived LNCaP cells, in which ERK1/2 activation occurs in strong correlation with AKT activation, we found that AKT-mediated B-Raf regulation is necessary for ERK1/2 activation. Specifically, in response to androgen deprivation, AKT upregulated B-Raf phosphorylation at Ser445 without affecting A-Raf or C-Raf-1. This effect of AKT was abolished by Arg25 to Ala mutation or truncating (∆4-129) the pleckstrin homology domain of AKT, indicating that the canonical AKT regulation is important for this signaling. Intriguingly, although a constitutively active AKT containing N-terminal myristoylation signal could sufficiently upregulate B-Raf phosphorylation at Ser445 in LNCaP cells, subsequent MEK/ERK activation still required hormone deprivation. In contrast, AKT activity was sufficient to induce not only B-Raf phosphorylation but also MEK/ERK activation in the hormone refractory LNCaP variant, C4-2. These data indicate that androgen depletion may induce MEK/ERK activation through a synergy between AKT-dependent and -independent mechanisms and that the latter may become deregulated in association with castration resistance. In support, consistent AKT-mediated B-Raf regulation was also detected in a panel of PCa lines derived from the cPten(-/-)L mice before and after castration. Our results also demonstrate that AKT regulates androgen receptor levels partly via the Raf/MEK/ERK pathway. This study reveals a novel crosstalk between ERK1/2 and AKT in PCa cells.

Figure 1. Androgen deprivation induces AKT activation and B-Raf-mediated ERK1/2 activation in LNCaP cells

(A) LNCaP cells were maintained in the medium containing c.s.FBS over 12 days. Total cell lysates harvested at indicated time points were analyzed by Western blotting for expression of phosphorylated AKT (pAKT), phosphorylated ERK1/2 (pERK1/2), ERK1/2, and AR. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was detected to validate equal protein loading. Western blot results were then analyzed by densitometry (right panel). Data (mean ± standard error) are from three independent experiments and show optical density normalized for GAPDH levels for pERK1 and pERK2, and % changes for pAKT and AR. P < 0.05 for pAKT, pERK1 and pERK2 at days 9 and 12 relative to day 0 (Student's t test). (B) LNCaP Cells, infected with the B-Raf specific lentiviral shRNA constructs (shBRaf #1, #2, and #3) or the control pGIPZ, were maintained in FBS or c.s.FBS medium for 6 days. Total cell lysates were analyzed by Western blotting for expression of the indicated proteins. pB-Raf, pA-Raf, pC-Raf, and pMEK1/2 denote phosphorylated forms of B-Raf, C-Raf, A-Raf, and MEK1/2, respectively. β-tubulin was detected to validate equal protein loading. 74 kDa indicates the expected region for the detection of C-Raf phosphorylated at Ser338. The efficacy of Raf antibodies is validated in Supplemental Fig. 2. (C) LNCaP cells were treated with the MEK1/2 inhibitor AZD6244 in c.s.FBS medium for the indicated time periods. Total cell lysates were analyzed by Western blotting for expression of the indicated proteins.

Figure 2. AKT inhibition attenuates c.s.FBS-induced phosphorylation of B-Raf

(A) LNCaP cells, infected with the lentivirus expressing the dominant-negative AKT (AKT-dn) or the control pHAGE virus, were maintained in FBS or c.s.FBS medium for 6 days. (B and C) LNCaP cells were treated with the AKT inhibitors, MK-2206 and AKTi, in c.s.FBS medium. Total cell lysates were analyzed by Western blotting for expression of the indicated proteins. pGSK3β denotes GSK3β phosphorylated at Ser9.

Figure 3. AKT promotes B-Raf/MEK/ERK activation and AR downregulation

(A) LNCaP cells, infected with the pHAGE lentivirus expressing wild type AKT (AKT-wt) or the kinase-deficient AKT mutant (AKT-K179M), were maintained in FBS or c.s.FBS medium for 48 hours. pHAGE is the control empty virus. (B) LNCaP cells, infected with lentiviral AKT-wt, were maintained in FBS or c.s.FBS medium for 48 hours. mRNA levels of AR and B-Raf were examined by qPCR. Data (mean ± standard error) are from a representative experiment performed in triplicate and are expressed as fold changes relative to the levels in pHAGE-infected cells in FBS medium. P value is < 0.005 for AKT relative to pHAGE (Student's t test). (C) LNCaP cells, infected with the pHAGE lentivirus expressing wild type AKT, were treated with increasing dose of AZD6244 in c.s.FBS medium for 48 hours. Total cell lysates from each experiment were analyzed by Western blotting for expression of the indicated proteins.

Figure 4. Roles of the PH domain for AKT-mediated B-Raf/MEK/ERK activation and AR downregulation

(A) LNCaP cells, infected with the pHAGE lentivirus expressing wild type AKT or the PH domain-defective AKT mutant (AKT-R25A), were maintained in c.s.FBS medium for 48 hours. (B) LNCaP cells, infected with the pHAGE lentivirus expressing the PH domain-truncated AKT mutant (AKT-null) or the PH domain-truncated and N-terminal-myristoylation-tagged AKT mutant (AKT-myr), were maintained in FBS or c.s.FBS medium for 48 hours. (C) LNCaP cells, infected with the pHAGE lentivirus expressing wild type AKT or the constitutively active AKT mutant (AKT-myr or AKT-E40K), were maintained in c.s.FBS medium for 48 hours. Total cell lysates from each experiment were examined by Western blotting for expression of the indicated proteins. AKT-myr migrates faster than other AKT constructs because of the N-terminal truncation.

Figure 5. Involvement of PAK in AKT-mediated B-Raf/MEK/ERK1/2 regulation

(A) LNCaP cells, infected with the pHAGE lentivirus expressing AKT-dn, were maintained in c.s.FBS medium for 6 days. Cell lysates are identical as those in Fig. 1B. The GAPDH image was also used in Fig. 1B. (B to D) LNCaP cells, infected with the pHAGE lentivirus expressing wild type AKT, AKT-K79M, AKT-R25A, AKT-null or AKT-myr, were maintained in FBS or c.s.FBS medium for 48 hours. Cell lysates are identical as those in Figures 2A, 3A and 3B. The GAPDH images were also used in those figures. (E) LNCaP cells, infected with the pHAGE lentivirus expressing wild type AKT, were maintained in c.s.FBS medium containing different amount of the PAK1 inhibitor IPA-3 for 48 hours. Total cell lysates were analyzed by Western blotting for expression of the indicated proteins. pPAK1/2 denotes PAK1 and PAK2 phosphorylated at Thr423 and Thr402, respectively.

Figure 6. Effects of constitutively active AKT expression in different PCa cells

Cells of LNCaP C4-2 (A) CWR22Rv1 (B), and LAPC4 (C), infected with the pHAGE lentivirus expressing AKT-null or AKT-myr, were maintained in c.s.FBS medium for 48 hours. Total cell lysates from each experiment were analyzed by Western blotting for expression of the indicated proteins.

Figure 7. AKT regulates B-Raf/ERK1/2 in PCa cells derived from the prostate-specific Pten knockout mice

(A and B) E8 cell line (derived before castration and androgen-dependent), CE1 and CE2 cell lines (derived from the castration resistant PCa), and E2 and E4 lines (androgen refractory progenies of E8 generated in vitro) were maintained in FBS medium for 48 hours before harvest. (C and D) CE2 and E2 cells were treated with increasing dose of the AKT inhibitor AKTi or the PAK1 inhibitor IPA-3 in c.s.FBS medium for 48 hours. (E) CE2 cells were treated with increasing dose of AKT-specific siRNA or the control scrambled siRNA oligomers in c.s.FBS medium for 72 hours. Total cell lysates from each experiment were analyzed by Western blotting for expression of the indicated proteins. (F) CE2 cells were treated with increasing dose of AKTi in c.s.FBS medium for 48 hours. mRNA levels of AR were examined by qPCR. Data (mean ± standard error) are from a representative experiment performed in triplicate and are expressed as fold changes relative to the levels in untreated cells.

Figure 8. A proposed model for AKT-mediated regulation of Raf/MEK/ERK and AR in PCa

In a subset of PCa cells, androgen deprivation activates AKT to increase the levels of Ser445 phosphorylated B-Raf partly via PAK1 (i). For B-Raf/MEK/ERK to be fully activated, additional AKT-independent signals are required (ii). As a physiological effect, AKT mediates AR downregulation partly via Raf/MEK/ERK (iii) and partly via a Raf/MEK/ERK-independent pathway (iv).