Dysregulation of the EGFR family is well known to associate with bladder cancer (14
). AR signals have also been implicated in bladder carcinogenesis and tumor progression (3
). Nonetheless, crosstalk between the AR and EGFR pathways remains unclear in bladder cancer, although it has been widely studied in prostate cancer (20
). We have recently shown that AR signals increase EGFR and ERBB2 expression and activity, suggesting androgen-mediated bladder cancer progression via the regulation of the EGFR/ERBB2 pathways (26
). In the present study, we provided evidence suggesting that EGF could regulate cell proliferation by activating AR signals in bladder cancer.
In prostate cancer, accumulating evidence has indicated that EGFR/ERBB2 signals induce AR transactivation in an androgen-dependent and -independent manner (20
). In bladder cancer cells, we here showed that EGF could activate AR transcription and PD168393, a specific inhibitor of EGFR, restored this EGF effect. These data suggest that EGF androgen-independently induces EGFR-mediated ARE reporter activity in bladder cancer. However, it was shown that the effect of EGF on AR transcription might be almost negligible compared to the induction by androgens in prostate cancer (20
). Similarly, in bladder cancer lines 5637-AR and J82-AR where a wild-type AR was stably overexpressed, the effect of EGF was less significant than that of DHT. On the other hand, in UMUC3 cells that possess endogenous AR, EGF effect (20% increase) is similar to the relatively insignificant effect of DHT (25% increase). In addition, PD168393 displayed agonist effects [1.5-fold (vs. 2.0-fold by EGF or 7.4-fold by DHT)] on AR transcription in J82-AR via unknown mechanisms. It was described in prostate cancer cells that PD168393 upregulated AR target gene expression in the presence of androgen, possibly via blocking basal activity of EGFR or ERBB2 (30
). Importantly, as shown in prostate cancer (21
), a combination of EGF and androgen further induced AR transcriptional activity in all the three bladder cancer lines tested and the AR antagonist HF completely abolished AR transactivation induced by EGF, androgen, or both at least in UMUC3. We could not evaluate antagonistic effects of HF on EGF-induced AR transcription due to the considerable agonist activity of HF which was even higher than that of EGF in 5637-AR. Thus, our results support the possibility that EGF mediates AR transcriptional activity through the EGFR and AR pathways in bladder cancer cells.
Consistent with previous findings shown by others and us (5
) androgens promoted AR-positive bladder cancer cell proliferation that was blocked by antiandrogens. These effects of androgens were suggested to be at least partially mediated through the EGFR pathway (26
). In the present study, as expected, EGF increased the growth of AR-positive cells and, less significantly, that of AR-knockdown/negative cells. In AR-positive lines, combined treatment with EGF and androgen further induced cell proliferation. Of note were inhibitory effects of the AR antagonist on EGF- and EGF+androgen-increased cell growth. Specifically, on the growth of 5637-derived lines, EGF and/or DHT showed only marginal effects (5637-V) and HF almost completely abolished EGF-mediated effects (5637-AR). These findings indicate that EGF-induced cell proliferation involves the AR pathway in bladder cancer. Nonetheless, in J82-derived lines, EGF retained its effect on cell growth without AR (J82-V) and HF failed to completely inhibit EGF-increased cell proliferation (J82-AR), suggesting the involvement of those other than the AR pathway.
It has been reported that EGF is capable of inducing AR transcription and protein expression in androgen-independent prostate cancer cells (21
). Others also described negative regulation of AR expression and activity by EGFR signaling in prostate cancer (30
). In bladder cancer cells, we previously showed increases in the expression of endogenous AR by androgen treatment (26
), which was inconsistent with the results demonstrated by Boorjian et al
). We also showed no significant increases in exogenously overexpressed AR (5637-AR) by DHT or in endogenous and exogenous ARs by EGF (26
). We confirmed our previous findings in the three lines tested and further showed EGF-enhanced AR overexpression in the presence of androgen in UMUC3 and J82-AR, but not in 5637-AR. The mechanism underlying this discrepancy in the response to the treatment of EGF+DHT between levels of exogenous AR expression in 5637 versus J82 remains uncertain. Repeatedly, the AR expression increased by androgen with or without EGF in bladder cancer cells was abolished by an AR antagonist.
EGF has been shown to enhance the expression or phosphorylation of TIF2, one of the p160 nuclear receptor coactivators, leading to an increase in AR transactivation in prostate cancer cells (21
). Indeed, the expression of major AR coactivators, including TIF2, was detected in bladder cancer cell lines as well as in AR-positive and even AR-negative bladder tumor specimens and TIF2 knockdown resulted in a decrease in androgen-mediated cell proliferation (9
). We here found that TIF2 was considerably (e.g., ≥1.5-fold) augmented in the presence of EGF and DHT in bladder cancer cells, while EGF or DHT alone could lead to marginal/only slight increases in TIF2 expression. Interestingly, like our results in AR expression/activity and cell proliferation, EGF-induced TIF2 upregulation was abolished by the antiandrogen. Although detailed mechanisms need to be clarified, these results may imply that elevated levels of TIF2 contribute to EGF/androgen-enhanced AR trans-activation in bladder cancer cells.
In hormone-responsive cells expressing both AR and ER (α and/or β), such as prostate and breast cancers, AR/ER/Src association plays a crucial role in activation of Src signals triggered by EGF and/or sex hormones (25
). It was noteworthy that either AR or ER antagonist sufficiently inhibited this EGF-mediated association and subsequent stimulatory effects (28
). It has also been shown that Src mediates EGF-induced AR tyrosine phosphorylation in prostate cancer cells, which leads to an increase in AR transcriptional activity (29
). Indeed, in many bladder cancer tissue specimens, AR and ER(s) were found to be co-expressed (3
). In this study, we showed associations of AR with ERβ and Src in UMUC3 which were enhanced by EGF treatment. These findings suggest that EGF activates Src via assembling the AR/ER/Src complex, resulting in AR transactivation and cell proliferation in bladder cancer. This may also justify the drastic inhibition of EGF-induced effects accomplished by antiandrogen treatment.
In conclusion, EGF could increase AR transcriptional activity and cell proliferation in bladder cancer. These EGF effects were likely mediated through the AR pathway involving upregulation of TIF2 expression as well as activation of Src signals due to forming an AR/ER/Src complex. These results, together with our previous findings, not only shed light on crosstalk between the AR and EGFR pathways in bladder cancer but also enhance the feasibility of androgen deprivation interfering with this crosstalk as a potential therapeutic approach.