Analysis of p13k independent survival pathways in the prostate cancer cell line LN cap 2

Inhibition of EGFR kinase activity with AG1478 did not have any effect on the ability of serum to rescue cells from LY-induced DNA fragmentation Figure 29a and caspase-3 activation Figur

**

3.1.8 Role of ErbB receptors in EGF- and serum-mediated survival of LNCaP

cells

There is accumulating evidence supporting the role of ErbB family of receptor tyrosine kinases in the progression of many types of human cancer including prostate cancer (Salomon et al 1995; Nicholson et al 2001; Hynes and Lane 2005) They consist of four members up to date, the epidermal growth factor receptor – EGFR (ErbB-1), ErbB2, ErbB3 and ErbB4, and play a major role in the proliferation, differentiation, migration and survival of mammalian cells (Schlessinger 2000; Jorissen et al 2003) Tumors with alterations in EGFR and ErbB2 receptors most often signal a more aggressive disease and is associated with poor prognosis for the patient (Allred et al 1992; Sjogren et al 1998; Nicholson et al 2001) Increased EGFR expression has been correlated to the progression of prostate cancer to the androgen-independent stage (Di Lorenzo et al 2002; Mimeault et al 2003; Hernes et

al 2004) The role of ErbB2 in PCa is more controversial due to conflicting reports

of the degree of gene amplification and protein expression in PCa (Ross et al 1997; Signoretti et al 2000; Osman et al 2001; Calvo et al 2003; Hernes et al 2004) However, it has been proposed that ErbB2 role in PCa progression is more significant at the late stages of hormone refractory PCa (Torring et al 2003) It is likely that the importance of ErbB receptors in prostate cancer maintenance is not only attributed by the level of expression but other events as well that contribute to aberrant receptor signaling such as establishment of autocrine signaling (Kambhampati et al 2005), receptor mutation (Olapade-Olaopa et al 2004) and their interaction with the androgen receptor

Serum contains a complex mixture of active components that have been shown to protect against apoptosis in various cell types including fibroblasts (Evan et

al 1992), Chinese hamster ovary cells (Zanghi et al 1999), neuronal cells (Uto et al 1994; Lucca et al 1997), pancreatic cancer cells (Levitt and Pollak 2002) and human amnion-like WISH (Wistar Institute Susan Hayflick) epithelial cells (Berry et al 2004) Other than growth factors like EGF, IGF and PDGF that promote survival and growth, small peptides in serum like cytokines, endothelin-1, thrombin, bombesin and lysophosphotidic acid have been shown to be able to transactivate the ErbB receptors when they bind to their cognate receptors Cytokine receptors can transactivate the ErbB receptor tyrosine kinase by directly phosphorylating the receptor itself via Janus tyrosine kinases (Jak) (Yamauchi et al 1997; Yamauchi et al 2000), whereas G-protein coupled receptors (GCPR) transactivate the ErbB receptors indirectly by activating metalloproteinases which in turn cleave and release ErbB ligand precursors (e.g: pro-HB-EGF) which then bind and activate ErbB kinase activity (Prenzel et al 1999)

As ErbB receptors play such an important role in prostate cancer maintenance,

we wanted to investigate if serum-mediated cell survival required activation of the ErbB receptors whether directly or indirectly via transactivation of the receptor In order to assess this, we utilized two widely used ErbB kinase inhibitors – AG1478 and AG879 They both belong to the tryphostin group of tyrosine kinase inhibitors that function by blocking the ATP acceptor site (lysine 721) located within the receptor’s intracellular kinase domain (Levitzki and Gazit 1995) The EGFR kinase inhibitor, AG1478, has a selectivity for EGFR (IC50, 3nM) at least 4 orders of magnitude higher than for ErbB2 (IC50, > 100µM), whereas the ErbB2 kinase inhibitor, AG879, has approximately 500-fold higher selectivity for ErbB2 (IC50, 1µM) than EGFR (IC50, > 500µM) in cell-free systems (Levitzki and Gazit 1995)

The dose response of LNCaP cells to both AG1478 and AG879 and a combination of the two inhibitors at various ratios were tested to determine the appropriate concentration that inhibits the kinase activity of the respective ErbB receptors without inducing excessive cell death Serum-starved LNCaP cells were pretreated for 1 hour with various doses of the inhibitors in serum-free media Then a final concentration of 10% serum was added to the media to trigger activation of the receptors After 24 hours incubation, crystal violet assay was performed to determined cell viability Cells were also harvested at 20 hours for determination of caspase-3 activity For AG1478, there was no significant cell death observed with all three doses used (0.5, 1 and 5µM) (Figure 27a) Accordingly, caspase-3 activity measured was no higher than basal level for even the highest dose of AG1478 used (Figure 27d) For the ErbB2 inhibitor, AG879, there was significant decrease in cell viability after 24 hours treatment with 10µM of the inhibitor (Figure 27b), and this was coincides with a noticeable increase in caspase-3 activity at 20 hours compared

to basal levels Therefore we proceeded to use a concentration of 5µM for both the AG1478 and AG879, and a combination of both inhibitors at a concentration ratio of 5µM:5µM (AG1478:AG879) (Figure 27c) as there was no significant cell death or caspase-3 activation at these concentrations

At the same time, the effects of the ErbB kinase inhibitors on the phosphorylation status of EGFR and ErbB2 receptors were analyzed Serum starved LNCaP cells were pretreated with AG1478, AG879 or a combination of both inhibitors in serum-free media for 1 hour before EGF (100ng/ml) or 10% serum was added Whole cell lysates were analyzed for EGFR and ErbB2 phosphorylation status using antibodies specific for phosphotyrosine site Tyr1173 of EGFR or Tyr1248 of ErbB2 (Figure 28) EGFR Tyr1173 is a major autophosphorylation site activated by

d)

Figure 27: Dose-response of LNCaP cell viability and caspase-3

activation to ErbB receptor kinase inhibitors

LNCaP cells were serum starved for 20 hours before 1 hour

preincubation with varying concentration of AG1478, AG879 or a

combination of both in fresh serum-free media Then 10% (v/v) serum

was added to the media (a), (b) and (c) After 24 hours incubation,

crystal violet assay was performed to determined cell viability (d)

Cells were also harvested at 20 hours for measurement of caspase-3

activity Results shown are a mean of 3 experiments performed in

1 10

5 5

1 5

Figure 28: Effects of AG1478 and AG879 on EGFR and ErbB2 phosphorylation

by EGF and serum

LNCaP cells were serum starved for 20 hours before 1 hour preincubation with AG1478 (5µM), AG879 (5µM) or DMSO control in fresh serum-free media Then EGF (100ng/ml) or 10% (v/v) serum was added to the media Whole cells were harvested after 15 minutes for SDS-PAGE analysis and immunoblotted for phosphorylated EGFR (P-EGFR) and ErbB2 (P-ErbB2) using anti-phospho-EGFR Tyr1173 and anti-phospho-ErbB2 Tyr1248 antibodies respectively Membrane were stripped and reprobed for total EGFR and ErbB2 β-actin was used as loading control Results shown in are from one representative experiment of at least three

receptor dimerization (Downward et al 1984) whereas Tyr1248 was previously reported to be an ErbB2 autophosphorylation site (Hazan et al 1990) EGF induced strong phosphorylation of EGFR at Tyr1173 that was inhibited significantly by AG1478 However the EGFR kinase inhibitor seemed to inhibit phosphorylation of the ErbB2 at Tyr1248 as well even to a greater extent than the ErbB2 kinase inhibitor This suggests that this site maybe subjected to transphosphorylation by EGFR kinase region and to a much lesser extent autophosphorylation by ErbB2 kinase in the presence of EGF It is known that EGF binding to EGFR can lead to the formation of both homo- and heterodimers ErbB2 is an orphan receptor without a physiologic ligand and is the preferred partner for heterodimerization for all ErbB proteins

P-EGFR (Tyr1173)

EGFR β-actin

P-ErbB2 (Tyr1248)

ErbB2 β-actin

AG1478 (5µM) AG879 (5µM)

-

-+ -

+

-+ +

EGF (100ng/mL)

-

-+ -

+

-+ + 10%FBS

P-EGFR (Tyr1173)

EGFR β-actin

P-ErbB2 (Tyr1248)

ErbB2 β-actin

AG1478 (5µM) AG879 (5µM)

-

-+ -

+

-+ +

EGF (100ng/mL)

-

-+ -

+

-+ + 10%FBS

binding would be the EGFR/ErbB2 dimers The use of a combination of EGFR and ErbB2 kinase inhibitors did not have much synergistic effect on both their respective tyrosine phosphorylation sites as AG1478 alone inhibited almost completely EGF-induced phosphorylation on both EGFR and ErbB2 These results support that the phosphorylation of ErbB2 at Tyr1248 is mainly due to transactivation by EGFR kinase domain following EGF-induced heterodimerization In contrast, serum was not able to induce detectable phosphorylation of EGFR at Tyr1173 site while weak phosphorylation of ErbB2 Tyr1248 was detected This was totally inhibited by a combination of AG1478 and AG879 but only partially when either inhibitor was used, suggesting that both EGFR and ErbB2 kinase activity contribute to serum-induced ErbB2 phosphorylation at Tyr1248 These results show that while EGF binding to its receptor induces robust phosphorylation of the EGFR homodimers and EGFR/ErbB2 heterodimers as expected, serum is only able to weakly phosphorylate the ErbB2 receptor via a mechanism that is dependent on both EGFR and ErbB2 kinase activity

Coinciding with EGF’s robust phosphorylation of EGFR, we observed a significant decrease in the level of detectable EGFR It has been observed that EGF binding to the EGFR activates the receptor and induces rapid internalization and degradation of the receptor by the ubiquitin system (Ceresa and Schmid 2000) (Waterman and Yarden 2001) Interestingly, incubation with EGFR kinase inhibitor prevented the internalization and degradation of the receptor It is possible that the kinase activity of EGFR of may play a role in the degradation of the receptor as phosphorylation of a single tyrosine residue on EGFR cytoplasmic region (Tyr1045) has been shown to be essential for Cbl-mediated EGFR degradation The Cbl ubiquitin ligase, when recruited to the internalized ligand-receptor complex in early

endosomes targets the receptor for lysosomal degradation by promoting receptor ubiquitination (Levkowitz et al 1998) This may explain why in the case of serum, EGFR degradation is not observed, possibly due to its inability to strongly activate the EGFR kinase activity (as seen by the absence of EGFR phosphorylation)

We proceeded to examine the relevance of the EGFR and ErbB2 receptors in serum and EGF-mediated survival in LY-treated LNCaP cells Serum-starved LNCaP cells were pretreated with AG1478, AG879 or a combination of both inhibitors as before, followed by addition of LY with either EGF or serum Cells were harvested for PI staining of DNA and caspase-3 activity at various time points Inhibition of EGFR kinase activity with AG1478 did not have any effect on the ability of serum to rescue cells from LY-induced DNA fragmentation (Figure 29a) and caspase-3 activation (Figure 28b) On the contrary, AG1478 completely blocked the effect of EGF on LY-induced cell death The ErbB2 inhibitor, AG879, was less efficient in inhibiting the effect of EGF-mediated survival (Figure 29a and c) Moreover the combination of both inhibitors seems to have the same effect on survival as the presence of EGFR inhibitor alone Taken together with the receptor phosphorylation experiments, these results demonstrate that EGF binding to its receptor, EGFR, induces its activation and ErbB2 transactivation, and subsequent downstream signaling The survival signal of EGF however, seems to be transmitted

by the EGFR kinase activity and not ErbB2’s On the other hand, serum-mediated survival signal does not appear to be conveyed via the EGFR or ErbB2 activation as inhibition of their kinase function did block serum’s protection on LY-mediated cell death

c)

* : p < 0.01 (vs LY)

Figure 29: Inhibition of EGFR and ErbB2 activity do not prevent

serum-mediated inhibition of LY-induced death in LNCaP cells

LNCaP cells serum starved for 20 hours were preincubated for 1 hour

without (control) or with AG1478 (5µM) or AG879 (5µM) before being

treated with 10% (v/v) serum (serum), LY (25µM) in serum-free media

(LY), LY in media with 10% serum (LY+serum) and LY with EGF

(100ng/ml) in serum-free media (LY+EGF) (a) 12 hours later % of

apoptotic cells was determined by propidium iodide staining (b) and (c)

Caspase-3 activity was measured 8 hours post-treatment Results are a

5000 10000 15000 20000 25000 30000

0 2000 4000 6000 8000 10000

*

*

3.1.9 Bax is required for LY-mediated apoptosis in LNCaP cells

While the BH3-only protein, Bad is clearly responsible for EGF-mediated protection in LNCaP cells treated with LY, silencing Bad only partially prevented total cell death induced by LY (Figure 19) This coincides with the observation that EGF also did not totally prevent cell death by LY especially after six hours of treatment It remains possible that incomplete silencing of Bad could be the cause of partial inhibition of cell death, nevertheless it also implicates involvement of other apoptosis-inducing factors This is in agreement with popular views that propose BH3-only proteins like Bad are “enablers” of apoptosis by conveying apoptotic signals and sensitizing cells to apoptosis by engaging pro-survival proteins while the multi-BH domain (BH1-BH3) pro-apoptotic proteins like Bax and Bak are “effectors”

of apoptosis (Letai et al 2002) Cells deficient of both Bax and Bak have been shown

to be resistant to apoptosis induced by a variety of stimuli (Wei et al 2001; Zong et al 2001) In addition, ectopic expression of BH3-only protein like Bad, Bid and Bim was unable to induce apoptosis in these double-deficient cells (Cheng et al 2001a; Wei et

al 2001; Zong et al 2001) implying an upstream function of the BH3-only proteins Bax and Bak appear to play partially redundant roles as effectors of apoptosis although most studies so far have focused on Bax

In order to assess the role of the pro-apoptotic Bax in our system, LNCaP were transfected with either siRNA specific for Bax or control siRNA before treatment with LY Silencing Bax alone completely abrogated apoptotic cell death induced by

LY as can be seen in Figure 30a, where the percentage of cells with DNA fragmentation was equal to cells transfected with control siRNA without LY treatment Similar results were obtained when caspase-3 activation was analyzed

inhibiting LY-induced apoptosis that silencing Bad (Figure 30c) Moreover silencing both proteins at the same time gave the same level of inhibition of caspase-3 activation as seen when Bax alone was silenced (Figure 30b and c) These data strongly argue in favor of an “effector “function of Bax whereas suggesting a

“enabler” role for Bad in this system

a) b)

* : p < 0.01 (vs control siRNA in LY-treated cells)

Figure 30: Bax is required for induction of apoptosis in LY-treated LNCaP cells

LNCaP were transfected with control siRNA, Bax siRNA, Bad siRNA or both Bax siRNA and Bad siRNA At 48 hours post-transfection, cells were serum starved followed by treatment either media with 10% serum (serum) or LY in serum-free medium (LY) Apoptosis in LY-treated cells at various time-points post-treatment were shown by (a) % of cells in sub-G1 phase and (c) caspase-3 activity (a) and (c) are a mean of three experiments performed in duplicates ± SE (b) Western blot of lysates collected 72 hours post-transfection with anti-Bax and anti-Bad antibodies demonstrated specific silencing of Bax and Bax β-actin was used to show equal loading Result shown in (b) is one representative experiment of at least three

3.1.10 Serum promotes cell survival in LY-treated LNCaP cells via inhibition of Bad and Bax translocation

Similar to Bad, induction of apoptosis activated Bax proteins requires its translocation from the cytosol to the mitochondria There are a lot of speculations on how various apoptotic stimuli lead to the activation of Bax, including direct phosphorylation by cellular kinases (Gardai et al 2004; Linseman et al 2004), alterations in intracellular pH (Khaled et al 1999; Tafani et al 2002; Ahmad et al 2004) or directinteraction with certain BH3-only proteins, the best documentedbeing the active truncated form of Bid, tBid (Desagher et al 1999; Eskes et al 2000; Kuwana et al 2002; Roucou et al 2002) an possibly Bim (Marani et al 2002) and other molecules like Ku70 (Sawada et al 2003), Humanin (Guo et al 2003) and 14-

3-3θ (Nomura et al 2003a) Nevertheless, upon receipt of apoptotic signals, Bax, otherwise a monomeric soluble protein in the cytosol, undergoes a conformational change and translocates to the mitochondrial membranes (Hsu et al 1997b; Wolter et

al 1997) Once inserted into the outer mitochondrial membrane, it forms oligomers (Antonsson et al 2001; Annis et al 2005) eventually leading to mitochondrial membrane permeabilization and release of pro-apoptogenic factors and caspase activation (Wei et al 2001; Adams 2003; Danial and Korsmeyer 2004; Green and Kroemer 2004)

homo-Figure 31 shows that in LNCaP cells growing in medium containing serum, a minimal amount of Bad and Bax proteins are present in the mitochondria (Figure 31a: serum) This could be contributed by a small percentage of apoptotic cells that represent less than 5% of the population (see Figure 11) Upon serum starvation of cells for twenty hours, only slightly more Bad and Bax proteins could be detected in the mitochondria compared to cells maintained in media with serum (Figure 31a:

serum, 0hr vs no serum, 0hr) This is correlates with the minimal increase in percentage apoptotic cells under this condition whereas in certain cell types serum withdrawal induces massive cell death with a concomitant significant redistribution

of Bax to the mitochondria (Colombaioni et al 2002; Trevisan et al 2004; Hou and

Hsu 2005) There is a possibility that PTEN mutation in LNCaP that bestows a

constitutive Akt activity even in the absence of serum may provide a means for retaining Bax on the cytosol (Tsuruta et al 2002; Gardai et al 2004) On the other hand, exposure of serum-starved cells to LY resulted in a strong accumulation of both Bax and Bad in the mitochondrial fraction (Figure 31a-c: LY, 90min) However addition of serum inhibited the LY-induced Bax and Bad translocation (Figure 31a–c:

LY, 90min vs LY + serum, 90min) The mechanism of serum’s inhibition of Bax translocation to the mitochondria is not due to PI3K-Akt signaling as it is inhibited in the presence of LY (see Figure 8) Also while serum could not maintain Bad phosphorylation in the presence of LY (see Figure 18), it was able to inhibit its translocation to the mitochondria It may be possible that serum is acting upon other factors to retain the dephosphorylated Bad in the cytosol Interestingly, and in accordance with the inhibition of Bad activation through its phosphorylation at Ser75, the presence of EGF at the time of incubation with LY strongly prevented Bad translocation to the mitochondria as shown before, but did not significantly inhibit Bax translocation to the mitochondria (Figure 31a–c: LY, 90min) vs LY + EGF, 90min) Of note the level of Bcl-2 in the mitochondria did not change significantly in any of the treatments as it is predominantly a mitochondrial protein The purity of the mitochondrial fraction was verified by demonstrating the presence of the voltage-dependent anion channel (VDAC), usually expressed in the outer mitochondrial

Figure 31: LY-induced initial Bax translocation to the mitochondria is inhibited

by serum but not EGF

LNCaP cells were plated in 100mm for 48 hours before either media with 10% serum

(serum) or serum-free media (no serum) was added for 20 hours Cells were harvested

at this point (0hr) (lane 1 and 2) for subcellular fractionation Additionally,

serum-starved cells were further treated with media with 10% serum (serum), LY (25µM) in

serum-free media (LY), LY in media with 10% serum (LY+serum) or LY with EGF

(100ng/ml) in serum-free media (LY+EGF) for 90 minutes before harvesting for

subcellular fractionation (lane 3-6) Fractionation was performed as in Materials and

methods (a) Equal amount of protein for mitochondria fraction was subjected to

SDS-PAGE and western blot analysis Purity of mitochondria fraction was determined by

blotting with mitochondria- and cytosol-specific protein, VDAC and Cu/Zn SOD

respectively (b) and (c) Densitometric analysis of amount of Bad and Bax in the

mitochondria fraction Results are shown as the ratio of mitochondria Bad or Bax over

VDAC converted to % of the ratio obtained in cells incubated with serum (% of

relative intensity) Results shown are one representative experiment of at least three

se m

No se

rum se m

LY+s

LY+EGF

membrane, and the absence of Cu/Zn superoxide dismutase (Cu/Zn SOD) which is a cytosolic protein

These data were confirmed by immunofluorescence staining of activated Bax using an antibody that recognizes the N-terminus of the Bax protein (anti-Bax6A7) It

is well documented that activation of Bax involves a conformation change that

unmasks its N-terminus (Hsu and Youle 1997; Goping et al 1998) prior to its translocation to the mitochondrial membrane Thus detection of the N-terminus of Bax has been considered as synonymous to detection of the activated state of Bax LNCaP cells treated with the protein kinase inhibitor, staurosporin (STS) was used as

a positive control for Bax activation as it has been demonstrated previously that treatment with STS induces translocation of Bax to the mitochondria leading to massive apoptosis as early as 3 to 6 hours in LNCaP cells (Marcelli et al 1999; Marcelli et al 2000; Li et al 2001) Immunofluorescence staining of Bax with anti-Bax6A7 antibody clearly illustrated that LNCaP cells treated with STS for 6 hours featured punctate staining of activated Bax that co-localizes with mitochondria stained with MitoTracker (Figure 32a: bottom panel) Activated Bax staining was not detected in LNCaP cells in control media as expected (Figure 32a: top panel)

Concurring with mitochondria translocation results, activated Bax was not detectable in serum-starved LNCaP cells, same as in cells grown in media with serum (Figure 32b: 0hr – serum and no serum) Serum-starved cells were then treated with

LY, EGF and serum as before Following 2 hours of treatment, activated Bax that localized with mitochondria was detected in LNCaP cells treated with LY in serum-free media, and the amount present did not change significantly when EGF was added together with LY In contrast, treatment with LY in presence of serum, activated Bax was not detected Together with the results from Bax translocation experiments, we

co-demonstrated that LY induced Bax activation and translocation to the mitochondria

that was inhibited by serum but not EGF

a)

Figure 32: LY-induced initial Bax activation is inhibited by serum but EGF

LNCaP cells were grown on coverslips in 12 well-plates and pretreated with 100nM

of MitoTracker CMXRos before treatment with (a) DMSO (control) or STS (4µM)

for 6 hours as positive control for Bax activation (b) LNCaP cells were also either

kept in media with serum (serum, 0hr) or serum-starved for 20 hours (no serum, 0hr)

followed by addition of fresh media with 10% serum (serum, 2 hr), LY in serum-free

media (LY), LY in media with 10% serum (LY+serum) or LY with EGF in

serum-free media After 2 hours, cells were stained for Bax activation using anti-Bax 6A7

antibody as in Materials and methods Images were captured using a fluorescence

microscope at fixed settings Results shown are one representative experiment of at

3.2 REACTIVE OXYGEN SPECIES REGULATION OF CELL SURVIVAL

3.2.1 Intracellular superoxide anions modulate serum’s inhibition of

LY-induced cell death

Reactive oxygen species (ROS) have been traditionally linked to their damaging effects on cells induced mostly by oxidative stress However in the past decade, ROS are increasingly recognized as important mediators in cell signaling that affect cell growth, differentiation and survival (Thannickal and Fanburg 2000) (Sauer

et al 2001b) (Pervaiz and Clement 2002; Pervaiz and Clement 2004) While oxidative stress involves high concentrations of ROS and mainly irreversible oxidation of lipids, proteins and DNA, redox signaling is usually associated with low, physiological amounts of ROS that reversibly oxidize signaling molecules leading to modifications

of their function (Forman et al 2004) Not surprising then that cells are armed with a comprehensive anti-oxidant system that tightly regulates the intracellular redox status Previous results from our lab have demonstrated that by varying the intracellular levels of ROS, in particular superoxide (O2·−), in tumor cells, we were able to regulate sensitivity of the cells to apoptosis induced by both the death receptor engagement or chemotherapeutic drugs (Clement and Stamenkovic 1996; Pervaiz et al 1999; Pervaiz

et al 2001; Clement et al 2003) Likewise we wanted to investigate if O2·− was involved in serum’s protection against LY-induced apoptosis Although mitochondria are the main site of ROS production via the process of oxidative phosphorylation, superoxide is also produced by a family of flavoprotein-dependent enzymes known as NADPH oxidases found on cellular membranes Several works recently have highlighted the role of these NADPH oxidases in enhancing the growth, tumorigenicity and survival of prostate cancer cells (Sauer et al 2001a; Arbiser et al

2002; Brar et al 2003; Lim et al 2005) To address the possibility that the presence of serum may be augmenting O2·− levels by activating the superoxide-producing enzyme, fresh media with 10% FBS was added to serum-starved LNCaP cells that were pre-incubated with or without diphenylene iodonium (DPI), an inhibitor of NADPH oxidase (Cross and Jones 1986; Hancock and Jones 1987) Presence of serum did not increase the level of intracellular O2·− compared to cells maintained in serum-free media as detected by lucigenin chemiluminescence assay (Figure 33a) However, pre-incubation with DPI in the presence of serum decreased intracellular superoxide level

of LNCaP cells to approximately 40% of those kept in serum alone, indirectly implying that production O2·−by NADPH oxidase contributes to a significant portion

of intracellular O2·−present under normal growth conditions Also treatment with LY did not affect the level O2·−ruling out the role of PI3K in the production of O2·−, and it also did not further decrease the level of O2·−when DPI was present As for the effect

of O2·− on cell death, lowering O2·− alone with DPI without LY did not induced significant amount of apoptosis, similar to those of cells treated with LY in the presence of serum (Figure 33b and c)

a)

0.0 0.2 0.4 0.6 0.8 1.0 1.2

se m

serum+DP I

b)

* = p < 0.01 (vs DPI+serum and LY+serum)

c)

* = p < 0.01 (vs DPI+serum and LY+serum)

Figure 33: DPI decrease O 2 · − level and abrogates serum’s protection in induced apoptosis

LY-LNCaP cells were serum-starved for twenty hours before the pre-incubation with DPI (6µM) for 1 hour Then 10% serum and LY (25µM) is added to the media and cells are harvested for a) measurement of intracellular O2·− after 1 hour using lucigenin assay as described in Materials and Methods Apoptosis was determined by measurement of b) caspase-3 activation at 24 hours and c) % cells in sub-G1 phase at

6, 12 and 24 hours post-treatment Results represent the mean of three experiments performed in duplicates ± SE

0 5 10 15 20 25 30 35

*

*

*

0 50 100 150 200 250 300

se m

DPI+s

erum

LY+serum

LY+serum+

However cells pre-incubated with DPI then treated with LY in the presence of serum showed a dramatic increase in caspase-3 activation and DNA fragmentation

By lowering the level of intracellular O2·−, the presence of serum was no longer as efficient in inhibiting the process of apoptosis in LY-induced cell death Several reports have shown the ability of serum and other growth factors to activate the NADPH oxidase leading to increased O2·− production within minutes of receptor stimulation, although the mechanism is still not well understood (Finkel 1998; Thannickal and Fanburg 2000; Vaquero et al 2004) However it appears that in LNCaP cells serum was not able to directly promote an increase in intracellular O2·−

concentration Tumor cells generally exhibit a pro-oxidant state due to increased production or decreased elimination of ROS compared to their normal counterparts, which confers a survival advantage to tumor cells (Cerutti 1985) It is possible that there is a “constitutive” generation of O2·−in prostate cancer cells that does not require growth factors or cytokines as compared to normal prostate cells Therefore, although

O2·−, does play a role in promoting serum-mediated survival as shown by results with DPI, it is unlikely due to a direct affect of serum increasing O2·−levels

Notwithstanding that DPI is widely used to inhibit NAD(P)H oxidase, it can also inhibit other flavoenzymes such as nitric oxidesynthase (Stuehr et al 1991).To confirm the pro-apoptotic effect of DPI was due to its inhibitory effect on O2·−

production by NADPH oxidase, LNCaP cells were treated with diethydithiocarbamate (DDC) which inhibits superoxide dismutase (SOD) (Iqbal and Whitney 1991; Hiraishi

et al 1992), the enzyme that catalyzes the conversion of O2·− to H2O2 to maintain a low steady-state concentration of O2·− (Fridovich 1995) Incubation with DDC increased the level of O2·−almost six-fold within 1 hour (Figure 34) Moreover DDC increased O2·− level in the presence of DPI, which is to be expected as NADPH

oxidase is only one of several sources of O2·−in the cells To note incubation with DDC alone or DDC with DPI did not induce significant amount of apoptosis or caspase-3 activation (Figure 35a and b) Importantly, DDC-mediated increase of intracellular O2·− was able to inhibit cell death promoted by DPI, supporting the essential role of O2·−in regulation of LNCaP cells sensitivity to apoptosis by LY This

is in line with other recent findings that have demonstrated that an increase in intracellular ROS can inhibit apoptotic signaling in tumor cells (Clement and Stamenkovic 1996; Pervaiz et al 1999; Pervaiz et al 2001; Brar et al 2003; Vaquero

et al 2004), which may contribute to their resistance to chemotherapeutic drugs (Pervaiz and Clement 2004)

Figure 34: DDC increase intracellular O 2 · − concentration in LNCaP cells

LNCaP cells were serum-starved for twenty hours before the incubation with DPI (6µM) and DDC (1mM) in fresh media with 10% serum for 1 hour Measurement of intracellular O2·−was done using lucigenin assay Results represent the mean of three experiments performed in duplicates ± SE

0.0 2.0 4.0 6.0

0 50 100 150 200 250 300 350 400

*

*

3.2.2 Intracellular superoxide and activation of MEK-ERK-RSK pathway

With increasing evidence supporting the inhibitory role of intracellular O2·−in apoptosis elicited by various triggers such as chemotherapeutic drugs and death receptors (Hockenbery et al 1993; Clement and Stamenkovic 1996; Hampton et al 1998; Pervaiz et al 1999; Pervaiz et al 2001; Brar et al 2003; Clement et al 2003; Vaquero et al 2004; Edderkaoui et al 2005; Mochizuki et al 2006), much interest has been geared towards deciphering the mechanisms and molecular targets of O2·−in that context Most of the focus has been on the ROS-induced modifications of kinases (Knapp and Klann 2000), phosphatases (Barrett et al 1999) and transcription factors (Lo and Cruz 1995; Teshima et al 2000) that contain redox-sensitive cysteine residues These oxidative modifications lead to either enhanced signaling of survival pathways or up-regulation of a number of anti-apoptotic proteins However in most cases, even though ROS has been shown to regulate the activation of a signaling cascade, the exact molecular target(s) remains to be elucidated In addition to oxidative modification of proteins involved in cellular signaling, the intracellular pro-oxidant milieu created by ROS has been shown to block the protease activity of the main apoptosis executioner, caspase-3, by oxidative modification of cysteine thiol group within the proteolytic site (Hampton and Orrenius 1997; Thornberry 1998; Pervaiz and Clement 2004)

Due to the pro-survival role of the MEK-ERK signaling cascade in LNCaP cells, we wanted to investigate if O2·−was mediating its anti-apoptotic effects through stimulation of this pathway Both exogenous and endogenously produced ROS have been shown to activate the MEK-ERK pathway (Stevenson et al 1994; Guyton et al 1996; Muller et al 1997; Wilmer et al 1997; Abe et al 2000; Greene et al 2000; Mukhin et al 2000; Sekharam et al 2000), however the mechanism and the precise

molecular target(s) of ROS is still unclear When LNCaP cells were treated with DPI

prior to incubation with 10% serum, there was a marked decrease in the ability of serum to phosphorylate ERK1/2 (Figure 36: serum vs DPI+serum) In contrast, pre-incubation with DDC enhanced the phosphorylation of ERK1/2 by serum (serum vs DDC+serum), and when added together with DPI, managed to revert the effects of DPI on ERK1/2 phosphorylation (DPI+serum vs DPI+DDC+serum) These results support the role of intracellular O2·−in regulating the activation of ERK1/2.In addition, phosphorylation of ERK1/2 increased in a dose-dependent manner with DDC concentration (Figure 37), supporting the direct effect of DDC on ERK phosphosrylation status Moreover the direct downstream target of ERK, RSK also showed a dose dependent increase with DDC Activation of RSK by ROS in prostate

cancer cells has been demonstrated by Sauer et al (Sauer et al 2001a) and various cell types by Abe et al (Abe et al 2000) However they demonstrated that while both

ERK and RSK activation are regulated by ROS, activation of RSK by ROS is contributed by other factors and not solely dependent on the activation status of its upstream kinase ERK1/2, suggesting that ROS differentially regulates the various

members of the MEK-ERK-RSK signaling cascade

As mentioned earlier, phosphorylation of ERK1/2 by serum appears to be partially dependent on PI3K activity (see Figure10), implying a cross-talk between the PI3K-AKT and MEK-ERK pathways To check if O2·−was mediating its effects by influencing the PI3K signaling, DDC’s ability to induce ERK1/2 phosphorylation was determined in the presence of LY While LY did not affect DDC-mediated ERK1/2 phosphorylation at one hour post-treatment, LY significantly decreased phosphorylation of ERK1/2 after 5 hours (DDC+serum vs LY+DDC+serum) These results seem to imply that regulation of prolonged ERK activation by O2·−may be

sensitive to PI3K inhibition However it does not rule out that O2·−effects on ERK phosphorylation is due to event(s) parallel to PI3K signaling that ‘feeds’ to MEK-ERK pathway One likelihood is O2·− inactivation of phosphatases that normally function to antagonize activation of MEK-ERK pathway This is supported by the observation that adding LY in the presence of DDC only partially dephosphorylate ERK1/2 compared to LY alone (LY+DDC+serum vs LY+serum) indicating that increasing intracellular O2·−still leads to enhanced ERK activation in spite of PI3K inhibition

Figure 36: Activation of MEK-ERK pathway is regulated by intracellular superoxide

LNCaP cells were serum-starved for 20 hours before treatment with DPI (6µM), DDC (1mM), LY (25µM) and 10% FBS as in Figure 35 Cells were harvested at 1 hour and 5 hours post-treatment for SDS-PAGE and immunoblotted with antibody against phosphorylated ERK1/2 Blots were stripped and reprobed with antibody against total ERK1/2 and β-actin as loading controls Blots shown are from one representative experiment of at least three

DDC+

se m

DPI+

DDC+seru

LY+D

DC +serum Serum

DPI+

erum

LY+DP

I +DD

C+serum

LY+D

DC +seru

DDC+

se m

DPI+

DDC+seru

LY+D

DC +serum

Seru

m

DPI+

se m

DDC+

se m

DPI+

DDC+seru

LY+D

DC +serum Serum

DPI+

erum

LY+DP

I +DD

C+serum

LY+D

DC +seru

Serum

DPI+

erum

LY+DP

I +DD

C+serum

LY+DDC +seru

LNCaP cells were serum-starved for 20 hours before treatment with DPI (6µM),

LY (25µM), 10% FBS and increasing concentration of DDC (1-20mM) as in

Figure 35 (a) Cells were harvested after 1 hour for SDS-PAGE and

immunoblotted with antibody against phosphorylated ERK1/2 and RSK Blots

were stripped and reprobed with antibody against total ERK1/2, RSK1 and

β-actin as loading controls (b) and (c) Densitometry analysis of results from (a)

Results are shown as the ratio of phosphorylated RSK or ERK1/2 respectively

over total protein and β-actin converted to % of the ratio obtained in cells

incubated with serum (% of relative intensity) Results shown in are from one

representative of at least three

+ +

1

+ +

10

+ +

+ - -

+ + -

+ - 1

+ + 1

+ + 10

+ + 20

0 100 200 300 400 500 600

+ - -

+ + -

+ - 1

+ + 1

+ + 10

+ + 20

3.2.3 Bad phosphorylation is regulated by intracellular O 2 · level

We have previously shown the BH3-only protein, Bad as an important downstream target of the MEK-ERK-RSK pathway that regulates survival in LNCaP cells Therefore we wanted to determine if Bad is regulated by O2·− in a similar fashion as MEK-ERK-RSK, and therefore play a role in O2·− regulation of cell survival As shown in Figure 38, treatment with DPI for one hour decreased phosphorylation of Bad approximately 40%, compared to the more dramatic decrease

in ERK phosphorylation This is in line with our earlier observation that phosphorylation of Bad on Ser75 by serum does not seem to be regulated by the MEK-ERK signaling; and in this case, O2·−may be affecting different targets in the regulation of ERK and Bad phosphorylation In addition, treatment with LY together with DPI caused a further decrease in phosphorylation status of Bad which correlates with increased apoptosis and caspase-3 activation (see Figure 33) compared to cells incubated in LY alone To support that dephosphorylated Bad plays a role in inducing cell death when LNCaP cells are treated with DPI and LY, cells were transfected with either control or Bad siRNA before treatment with the apoptotic triggers As shown in Figure 39, silencing Bad significantly reduced caspase-3 activation in cells treated with DPI and LY In line with this, addition of DDC which increased intracellular O2·−

concentration was able to re-phosphorylate Bad in the presence of DPI (Figure 40: DPI+serum vs DPI+DDC+serum) Furthermore DDC was able to re-phosphorylate Bad in the presence of LY (LY+DPI+serum vs LY+DPI+DDC+serum), which correlates with a decrease in percentage of cells in the sub-G1 population and caspase-3 activity

a)

b)

Figure 38: DPI induces Bad dephosphorylation

LNCaP cells were serum-starved for twenty hours before they were pre-incubated for one hour with DPI (6µM) Then LY (25µM) and 10% FBS was added to the media and cells were harvested one hour later for SDS-PAGE analysis (a) Samples were immunoblotted with antibodies for phospho-Bad Membranes were stripped and reprobed for total Bad and β-actin as loading control (b) Densitometry analysis of results from (a) Results are shown as the ratio of phosphorylated Bad over total protein and β-actin converted to % of the ratio obtained in cells incubated with serum (% of relative intensity) Results shown in are from one representative of at least three

P-Bad (S75) t-Bad β-actin

serum DPI+s

erum

LY+s

erumLY+D

PI +serum

P-Bad (S75) t-Bad β-actin

serum DPI+s

erum

LY+s

erumLY+D

PI +serum

0 20 40 60 80 100 120

se m

DPI+s

erum

LY+serum

DPI+LY+s erum

Figure 40: DDC induces Bad phosphorylation

LNCaP cells were serum-starved for 20 hours before treatment with DPI (6µM), DDC (1mM), LY (25µM) and 10% FBS as in Figure 35 Cells were harvested at 5 hours post-treatment for SDS-PAGE and immunoblotted with antibody against phosphorylated Bad Blots were stripped and reprobed with antibody against total Bad and β-actin as loading controls Blots shown are from one representative experiment

of at least three

0 50 100 150 200 250 300 350

*

P-Bad (S75)

t-Bad β-actin

se

m

erum DDC

se

m

erum DDC

se

m

erum DDC

Phorbol esters are known to promote cellular transformation and tumorigenesis which can be inhibited by antioxidants (Cerutti 1985) They induce production of superoxide by NADPH oxidase in various cell types including neutrophils (Watson et al 1991) and tumor cell lines (Clement and Stamenkovic 1996) Incubation of LNCaP cells with 2 µM of phorbol 12-myristate 13-acetate (PMA) in the presence of LY, increased intracellular O2·− (Figure 41a) with a simultaneous increase in Bad phosphorylation (Figure 42b and c) as compared to cells

in serum alone Pre-incubation with DPI caused a marked decreased in PMA-induced

O2·−production but remained significantly higher compared to cells incubated without PMA This correlates with the level of Bad phoshorylation supporting the role of O2·−

in regulation of Bad activation However, Bad phosphorylation is higher in cells incubated with PMA in the presence of DPI and LY compared to those in serum alone (Figure 41b and c: lane 5 vs lane 1), even though O2·−concentration is relatively higher in cells incubated in serum alone This indicates that other factors may be involved in PMA regulation of Bad phosphorylation Nevertheless, incubation with PMA significantly inhibited apoptosis induced by DPI and LY (Figure 41d) These results implicates the pro-apoptotic Bad as one important component of O2·− -mediated regulation of cell survival in LNCaP cells, where a low intracellular concentration of O2·− facilitates dephosphorylation and subsequent activation of Bad, and thus sensitizing the cells to apoptotic triggers Even though the MEK-ERK-RSK signaling pathway upstream of Bad is shown to be sensitive to regulation by O2·−, the exact molecular target of O2·− is still not known

MA+serum

I+serum

LY+DPI +PM A+serum

P-Bad (S75) t-Bad β-actin

PMA+

se m

LY+P

MA+serum

I+serum

LY+DPI +PM A+serum

PMA+

se m

LY+P

MA+serum

I+serum

LY+DPI +PM A+serum

P-Bad (S75) t-Bad β-actin

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 serum

LY+PMA+serum LY+DPI+serum LY+DPI+PMA+serum

Relative Intracellular Superoxide (RLU/µg protein of treated vs control)

#

*

0 100 200 300 400 500 600

se m

PMA+

serum+PMA+serum

LY+DP

I+serum

PI+PMA +serum

of at least three (c) Densitometry analysis of results from (a) Results are shown as the ratio of phosphorylated Bad over total protein converted to % of the ratio obtained

in cells incubated with serum (% of relative intensity) (d) Apoptosis was determined

at various time-points by PI staining Results represent the mean of three experiments performed in duplicates ± SE

Time (hr)

0 5 10 15 20 25 30 35

*

*

CHAPTER 4: DISCUSSION

The PI3K-Akt signaling cascade is widely accepted as an important survival pathway in various tumors cells It features prominently in PCa as a loss of function

mutation of PTEN which antagonizes PI3K signaling is one of the most common

mutations found in PCa Not surprisingly, therapies based on inhibition of PI3K-Akt signaling have been intensely explored as it is predicted to profoundly affect the development and progression of PCa However, there has yet to emerge an effective therapy from targeting the PI3K-Akt pathway alone Perifosine, an oral alkylphospholipid was demonstrated to decrease Akt activation and induce apoptosis

in various cancer cell lines including prostate cancer (Ruiter et al 2003; Hideshima et

al 2006; Li et al 2006) However, phase II trials using perifosine as a single agent in patients with hormone refractory prostate cancer did not result in any significant therapeutic benefit (Posadas et al 2005) Similar results were obtained from clinical trials using perifosine in patients with pancreatic carcinoma (Marsh Rde et al 2007) and metastatic head and neck cancer (Argiris et al 2006) One possible explanation for failure of such monotherapy is that most cancers cells possess multiple anti-apoptotic mechanisms that work synergistically or in parallel to ensure the survival of the cells In fact, often cancer cells gather mutations that promote cell survival as the tumor progresses, resulting in resistance to therapy

In prostate cancer, growth factor signaling has been proposed to gain importance as the disease progresses as witnessed by the upregulation of various growth factors such as EGF and its receptors as well as a switch from paracrine to autocrine growth factor signaling As growth factors have been demonstrated to not only promote proliferation of cancer cells but also their survival, it is reasonable to

propose that the deregulated growth factor signaling can induce alternative or latent survival signaling pathway(s) in PCa that contribute to the resistance of prostate cancer cells to conventional therapy

4.1 EGF AND SERUM ACTIVATE PI3K-AKT-INDEPENDENT SURVIVAL PATHWAY(S) IN LNCAP CELLS

The PI3K-Akt pathway has been proposed to be the major survival pathway in LNCaP cells due to PTEN inactivation in this PCa cell line (Lin et al 1999) Our results and reports from others (Carson et al 1999) demonstrate that while PI3K-Akt survival pathway is important for LNCaP survival, there exist alternative survival pathway(s) that can come into play when the PI3K-Akt signaling is inhibited Indeed, while inhibition of PI3K-Akt signaling with LY led to apoptotic cell death, the presence of serum or growth factors such as EGF protected the LNCaP cells from LY-induced apoptosis, pointing to the existence of PI3K-Akt-independent survival pathways that can activated by serum or EGF

There are numerous reports that support the role of the pro-survival Bcl-2 and Bcl-xL in PCa cells’ resistance to apoptosis in response to various chemotherapeutic drugs as overexpression of these proteins protected the cells while downregulation with specific antisense oligonucleotides sensitized them to apoptosis (Lebedeva et al 2000; Shi et al 2001; Yamanaka et al 2006) Moreover upregulation of Bcl-xL expression by serum was proposed to mediate PI3K-independent survival in prostate cancer cell lines (Yang et al 2003) However, our findings show that incubation with 10% serum or EGF did not increase Bcl-2 and Bcl-xL in LNCaP even after 24 hours incubation, whereas the protective effect of both serum and EGF was observed very much earlier on This indicates that upregulation of the two pro-survival Bcl-2 protein

is not the main mechanism involved in inhibition of apoptosis under these conditions Nonetheless, the upregulation of pro-survival Bcl-2 members may still play a role in the long term survival of PCa cells as apoptosis induction is determined by the interactions between pro-survival and pro-apoptotic Bcl-2 members Moreover Bcl-2 and Bcl-xL have been reported to be upregulated in PCa following androgen ablation therapy, allowing PCa cells to survive under condition of low androgen (Kajiwara et

al 1999; Catz and Johnson 2003; McCarty 2004)

4.2 EGF-MEDIATED SURVIVAL IS DEPENDENT ON MEK-ERK ACTIVATION IN LNCAP CELLS

Binding of EGF to its receptor, EGFR, leads to activation of its receptors’ tyrosine kinase activity which propagates the growth factor signals to various downstream signaling cascades Among the most well studied is the Ras-Raf-MEK-ERK signaling cascade which regulate cellular functions including proliferation, differentiation and survival Indeed our findings demonstrate that EGF induced a rapid and robust activation of MEK-ERK signaling Moreover our results support that the activation of MEK-ERK signaling by EGF led to inhibition of apoptotic cell death induced by LY Inhibition of MEK-ERK signaling using a specific inhibitor of MEK, U0126, almost completely abolished EGF protective effect in LNCaP cell, implying that MEK-ERK signaling cascade is the main survival signal that is transmitted by EGF In support of this, studies by others demonstrated that inhibition of ERK1/2 signaling sensitized DU-145 prostate cancer cell line to radiation-induced apoptosis (Hagan et al 2000; Qiao et al 2002; Yacoub et al 2003) Similarly, MEK1/2 inhibitors were shown to potentiate apoptosis induced by Tryphostin AG825 and

that MAPK activity, particularly ERK1/2 activity, correlates with progression to an increasingly advanced and hormone-independent disease, where non-neoplastic human prostate tissue specimen exhibited minimal MAPK activity whereas in prostate tumors, MAPK activity was elevated with increasing Gleason score and tumor stage (Gioeli et al 1999; Gioeli et al 2001) Taken toegether, the MEK-ERK signaling is hypothesized to play an important role in prostate cancer progression and may contribute to resistance to conventional therapy

4.3 EGF-MEDIATED SURVIVAL REQUIRES EGFR’S TYROSINE KINASE ACTIVITY

There have been many studies done to investigate the expression of EGFR family member through out the progression of prostate cancer especially EGFR and ErbB2 While most demonstrate an increase in the expression of EGFR that correlates with progression of the disease to androgen-independence (Scher et al 1995; Di Lorenzo et al 2002; Hernes et al 2004), results on ErbB-2 expression have been more contradictory (Ratan et al 2003; Hernes et al 2004) These inconsistencies could stem from differences in methodology, specimen preparation, specimen heterogeneity and methods of scoring Unlike in breast cancer where ErbB-2 have been shown to be overexpressed in a subset of 20-30% of tumors due to gene amplification and is associated with poor prognosis (Slamon et al 1987; Slamon et al 1989; Berchuck et

al 1990; Naber et al 1990), several reports have demonstrated a lack of ErbB-2 gene amplification in prostate cancer (Bubendorf et al 1999; Savinainen et al 2002) This may explain why treatment with anti-ErbB-2 antibody traztuzumab (Herceptin), which is effective in breast cancer (Pegram et al 2000), was not successful in advanced prostate cancer (Morris et al 2002) However it is generally agreed upon

that overexpression of EGFR and ErbB2 when they do occur, often correlates with poor prognosis

Our findings suggest that EGF mediates the survival signal through activation

of EGFR’s receptor tyrosine kinase activity as inhibition of EGFR’s kinase activity abrogated EGF protective effect in LNCaP cells Interestingly, several natural products such as silymarin, genistein, epigallocatechin 3-gallate and curcumin demonstrated growth inhibitory effect on PCa cells which was shown to be due to inhibition of EGF/EGFR signaling (Zi et al 1998; Ye et al 1999; Dorai et al 2000; Dorai et al 2001) Particularly, curcumin, which was shown to inhibit EGFR tyrosine kinase activity, also induced apoptosis and inhibited proliferation and angiogenesis of

LNCaP cells (Dorai et al 2000; Dorai et al 2001) Additionally, although EGF also

induces phosphorylation of ErbB2 receptor, probably through transphosphorylation

by EGFR, ErbB2 tyrosine kinase activity was not important for EGF-mediated survival LNCaP cells express EGFR, ErbB2 and ErbB3 but no detectable ErbB4 (Torring et al 2003; El Sheikh et al 2004) Studies by others have also shown that EGF induced phosphorylation of both EGFR and ErbB2 receptors in LNCaP cells No detectable tyrosine phosphorylation of ErbB3 and ErbB4 was observed; therefore they

do not appear to be the significant dimerization partners of EGFR in this cell line (Torring et al 2003; El Sheikh et al 2004) Notably, while ErbB2 kinase activity did not seem to relevant for EGF-mediated survival, it may still play a role in proliferation of LNCaP cells In fact, phosphorylation of both EGFR and ErbB2 has been reported to be a reliable predictor of prostate cancer cell proliferation in response

to EGF (El Sheikh et al 2004)