Abstract |
Prostate cancer is a leading cause of premature death in Western countries. After
many decades, androgen-ablation remains the principal treatment paradigm for PCa.
However, relapse from hormonal therapy is almost certain and most patients with
metastasis progress to end-stage disease regardless of treatment strategy. The lack of
progress on clinical options for PCa patients reflects our poor understanding of the
molecular and cellular mechanisms that underlie disease etiology and progression.
The non-genomic action of steroids is an active field of research during the last
decade. Whereas the classical (genomic) model of steroid action involves binding to
specific intracellular steroid receptors, translocation to the nucleus, DNA binding, and
activation of specific genes, a process which requires hours to be completed, a nongenomic
effect usually occurs within minutes, is insensitive to inhibitors of
transcription and translation and it may occur in cells not expressing or expressing
non-functional classical steroid receptors.
Recently, membrane androgen receptors (mAR’s) have been detected in the human
prostate cancer cell lines LNCaP and DU145, the second of which is devoid of
functional classical intracellular receptors (iAR’s). R activation through nonpermeable
androgen analogues like testosterone-BSA induces the apoptotic regression
of both prostate cancer cell lines in vitro and in vivo. Moreover, a new, non-genomic
mechanism of androgen action was found to alter actin cytoskeleton dynamics in
LNCaP cells through a signaling cascade involving phosphorylation and activation of
FAK, PI-3K and the downstream signaling molecules Rac1 and Cdc42. Actin
cytoskeleton reorganization is one of the earliest cellular responses to many
extracellular stimuli. Recently it was reported to control apoptosis in different cellular
systems.
Based on these findings the main goal of the present study was to examine a possible
link between rapid actin cytoskeleton reorganization and the apoptotic response of
prostate cancer cells following mAR activation. DU145 cells expressing nonfunctional
intracellular androgen receptors (iAR’s) were chosen as the principal
cellular model for this study, in order to exclude interference of genomic androgen
action in the examined cellular responses.
In line with previous results in LNCaP cells, both dihydrotestosterone and the nonpermeable
analogue testosterone-BSA, induced rapid actin polymerization followed
by the formation of filopodia and stress fibers in DU145 cells, as indicated by the
19
decrease of the G/total actin ratio and morphological analysis using confocal laser
scanning microscopy. Furthermore, mAR activation led to actin microfilament
redistribution and apoptosis in PC3 cells, another prostate cancer cell line expressing
non-functional iAR’s. The mAR induced apoptotic response of all three cell lines was
inhibited by pre-incubation of the cells with actin microfilament blocking agents such
as cytochalasin B (cytB) or phallacidin, indicating a key role for actin cytoskeleton
dynamics in regulating apoptosis in mAR stimulated prostate cancer cells.
We then addressed the signaling pathway downstream of mAR activation in DU145
cells and identified a novel, non-genomic signaling cascade leading to actin
cytoskeleton reorganization, different from the one described in LNCaP cells. Indeed,
the levels of phosphorylated FAK were found to be comparatively high in nonstimulated
DU145 cells, probably due to the cell’s high invasive and metastatic
potential. Cell stimulation by testosterone-BSA did not induce further
phosphorylation of FAK or its immediate substrate PI-3K. The levels of endogenous
phospho-PI-3K instead decreased after prolonged incubation with the androgen, while
co-treatment of cells with the PI-3K inhibitor wortmannin did not affect androgen
induced actin reorganization. In addition, mAR activation had no effect on Rac1-GTP
levels in DU145 cells. These findings indicate that the FAK PI-3K Rac1
pathway is not involved in regulating actin cytoskeleton dynamics in DU145 cells.
On the contrary, three other members of the family of small RhoGTPases, namely
RhoA, RhoB and Cdc42, are rapidly activated following mAR stimulation in DU145
cells. Cells expressing dominant-negative RhoA, RhoB or Cdc42 failed to form new
actin stress fibers and filopodia in response to testosterone-BSA, whereas the
surrounding non-transfected cells showed robust actin reorganization. Furthermore,
ROCK was found to be their downstream effector, since inhibition of ROCK by the
specific inhibitor Y-27632 completely blocked actin polymerization. RhoGTPases are
known to regulate the activity of LIM kinases. We therefore performed in vitro kinase
assays and detected rapid testosterone-BSA induced LIMK2 phosphorylation. Destrin,
an isoform of cofilin, co-precipitated with LIMK2 and became phosphorylated with
kinetics similar to those of LIMK2. Since the members of the ADF/cofilin family
have an established role as actin depolymerizing factors, phosphorylation and hence
deactivation of destrin contributes to increased actin polymerization in our system. In
addition, ROCK inhibition by Y-27632 completely blocked androgen induced
20
LIMK2/destrin phosphorylation. These findings suggest that mAR activation leads to
actin cytoskeleton reorganization through a RhoA/B, Cdc42 ROCK LIMK2
destrin signaling cascade in androgen-independent DU145 cells.
Interestingly, RhoA and RhoB were activated in response to testosterone-BSA in
androgen-dependent LNCaP cells too, though their activation required longer
incubation times with the androgen. Treatment with the specific ROCK inhibitor, Y-
27632, also blocked actin polymerization in LNCaP cells. Furthermore, the androgen
induced apoptotic response of both DU145 and LNCaP cells was abolished upon
treatment with Y-27632. On the contrary cell treatment with the PI-3K inhibitor,
wortmannin, decreased apoptotic rates in LNCaP cells exclusively, probably by
blocking the signal transmission to the actin cytoskeleton upstream of Rho/ROCK.
Therefore, it seems that the Rho/ROCK pathway regulates mAR induced apoptosis by
controlling actin cytoskeleton reorganization in prostate cancer cells, bearing or not
functional intracellular androgen receptors (iAR's).
Finally, we studied the pro-apoptotic factors involved in the apoptotic response of
DU145 cells. The expression of members of the Bcl-2 protein family was not
effected, but testosterone-BSA treatment resulted in rapid activation of caspase-3 and
increased FasL expression. Both factors are controlled by Rho/ROCK and actin
cytoskeleton dynamics, as indicated by the abolishment of their activation in the
presence of the specific ROCK inhibitor, Y-27632, or actin blocking agents. In
addition, the rapid decrease of Akt phosphorylation in response to testosterone-BSA,
with kinetics similar to those of PI-3K, might suggest the deactivation of the main
survival pathway PI-3K/Akt, thereby contributing to apoptosis.
In conclusion, the present study has identified, for the first time, the biological
significance of the mAR induced rapid actin cytoskeleton reorganization in human
prostate cancer cells, demonstrating a key role for the actin cytoskeleton in regulating
the cell's apoptotic response. In addition, a new, non-genomic signaling pathway
leading to androgen induced actin polymerization in DU145 prostate cancer cells,
expressing non-functional iAR's, was identified. This pathway involves the activation
of RhoGTPases and of their downstream kinases ROCK and LIMK2, transmitting a
signal to destrin, thereby modulating its ability to shift the actin polymerization
equilibrium in a positive manner. Interestingly, Rho/ROCK were also shown to
21
control actin cytoskeleton reorganization in iAR-positive LNCaP cells. Furthermore,
the regulatory role of Rho/ROCK in the apoptotic response of both iAR- positive
LNCaP and iAR-negative DU145 cells has been highlighted. Finally, experimental
data indicating the rapid activation of pro-apoptotic factors in response to mAR
stimulation, such as the increase in caspase-3 activity and the deactivation of the PI-
3K/Akt survival pathway, provide a starting point for a mechanistical approach of the
interaction between the actin cytoskeleton and apoptosis in prostate cancer cells.
|