**3. Results**

### *3.1. E*ff*ects of ADT on Tumor Growth and Serum PSA Kinetics of Xenograft Derived from Co-Inoculation of E9, F10, and AIDL Cells with pcPrF-M5 Cells In Vivo*

Regardless of the presence of pcPrF-M5 cells, tumor volumes in mice inoculated with E9 cells rapidly decreased following ADT (Figure 1A) and were not altered between days 14 and 21 after ADT, i.e., the e ffects of ADT on E9 tumors were maintained until at least day 21 after ADT. In contrast, in the presence or absence of pcPrF-M5 cells, tumor volumes in mice inoculated with F10 cells temporally decreased following ADT, but increased between days 14 and 21 after ADT (Figure 1A). Moreover, in the presence or absence of pcPrF-M5 cells, tumor volumes in mice inoculated with AIDL cells gradually increased following ADT, similar to the results in the sham group (Figure 1A).

Ki-67 labeling indices in mice inoculated with E9 cells, with or without pcPrF-M5 cells, rapidly decreased following ADT (Figure 1B) and were not significantly increased between days 14 and 21 after ADT; that is, the e ffects of ADT on E9 tumors were maintained until at least day 21 after ADT. In contrast, Ki-67 labeling indices in mice inoculated with F10 or AIDL cells, with or without pcPrF-M5 cells, were not altered following ADT as compared with that in the sham group (Figure 1B).

Serum PSA titers in mice inoculated with E9 cells alone rapidly decreased after ADT (Figure 1C). Importantly, those in mice inoculated with E9 cells plus pcPrF-M5 cells rapidly decreased following ADT and became detectable on day 21 after ADT. Serum PSA titers in mice inoculated with F10 cells, with or without pcPrF-M5 cells, increased gradually (Figure 1C), whereas those in mice inoculated with AIDL cells, with or without pcPrF-M5 cells, were not detected (Figure 1C).

**Figure 1.** Effects of androgen deprivation therapy (ADT) on tumor growth and serum PSA kinetics of xenografts derived from co-inoculation of LNCaP sublines with pcPrF-M5 cells in vivo. Changes in tumor volume (**A**), Ki67 index (**B**), and serum PSA (**C**) of xenografts were compared in untreated (sham-operated) or ADT-treated mice inoculated with PCa cells with or without M5 cells on days 0, 14, and 21 after ADT. \*\* *P* < 0.01 versus sham-operated control. ADT, androgen deprivation therapy; PSA, prostate-specific antigen; M5, pcPrF-M5.

E9-derived tumors, with or without pcPrF-M5 cells, grown in mice treated with ADT showed reduced tumorigenesis as compared with those in untreated (sham-operated) mice (Figure S1). In contrast, both F10- and AIDL-derived tumors, with or without pcPrF-M5 cells, grown in mice treated with ADT showed no changes in tumorigenesis as compared with those in untreated mice (Figures S2 and S3). Additionally, AR and PSA proteins were expressed in both E9- and F10-derived tumors, with or without pcPrF-M5 cells, on day 21 after ADT, although serum PSA levels were very low in mice inoculated with E9 cells alone (Figures S1 and S2). Moreover, AR protein was expressed in AIDL-derived tumors, with or without pcPrF-M5 cells, on day 21 after ADT, whereas PSA protein was not detected because of mutated AR in AIDL cells (Figure S3). NSE staining was diffuse among ADT-treated and untreated hosts on day 21 after ADT (Figures S1–S3). Microvessel density (MVD) in all tumors, regardless of the presence of pcPrF-M5 cells, was not altered among ADT-treated and untreated hosts on days 14 and 21 after ADT (Tables S1–S3).

### *3.2. E*ff*ects of Indirect Coculture with Fibroblasts on E9, F10, and AIDL Cells In Vitro*

Cell proliferation of E9 and AIDL cells was significantly increased when cells were cocultured with PrSCs or pcPrFs, whereas that of F10 cells was not affected (Figure 2A). Expression levels of PSA proteins were increased in E9 cells cocultured with pcPrFs but not PrSC and were not affected in F10 cells (Figure 2B). In contrast, expression levels of PSA proteins were decreased in F10 cells cocultured with PrSC but not pcPrFs. PSA protein expression was not detected in AIDL cells cocultured with PrSC or pcPrFs (Figure 2B). NSE protein expression and STAT3 phosphorylation were increased in E9 cells cocultured with PrSC or pcPrFs (Figure 2B). NSE proteins were not changed in F10 cells cocultured with PrSC or pcPrFs, whereas STAT3 phosphorylation was increased in F10 cells cocultured with pcPrF-M6 cells (Figure 2B). NSE protein expression was decreased in AIDL cells cocultured with pcPrFs but not PrSC, whereas phosphorylation of STAT3 was increased in AIDL cells cocultured with pcPrFs (Figure 2B).

### *3.3. E*ff*ects of Growth Factors and Cytokines on E9, F10, and AIDL Cells In Vitro*

Cell proliferation of E9 and AIDL cells was significantly increased by treatment with growth factors and cytokines, such as EGF and IL-6, whereas that of F10 cells was significantly decreased by treatment with FGF-10, HGF, IGF-1, and TGFβ1 (Figure 3A). Phosphorylation of Akt and p44/42 MAPK in E9 and AIDL cells was strongly increased by treatment with EGF, whereas that in F10 cells was not affected (Figure 3B). Phosphorylation of Akt and p44/42 MAPK in E9, F10, and AIDL cells was not affected by treatment with HGF (data not shown). EGFR protein was detectable in all PCa cells, including LNCaP, E9, F10, AIDL cells, whereas expression of EGFR protein in AIDL cells was considerably higher than that in E9 and F10 cells (Figure S4). Detection of EGFR protein in BPH-1 cells was used as a positive control for anti-EGFR antibodies. In addition, PSA secretion from E9 cells was significantly increased by treatment with EGF but not HGF, whereas that from F10 cells was significantly decreased by treatment with both EGF and HGF (Figure S5). Notably, PSA secretion from AIDL cells was not detected.

Full-length AR protein was detectable in all PCa cells, including LNCaP, E9, F10, AIDL, and 22Rv1 cells, whereas AR-V7 protein was detectable only in F10 cells (Figure 4). Detection of AR-V7 protein in 22Rv1 cells was used as a positive control for anti-AR-V7 antibodies.

**Figure 2.** Effects of indirect coculture with fibroblasts on cell proliferation and PSA expression in LNCaP sublines in vitro. (**A** and **B**) LNCaP sublines were co-cultured with fibroblasts using cell culture inserts for 4 days in phenol red (−) RPMI-1640 with 1% CS-FBS containing DHT (0.1 nM). (**A**) Cell proliferation. \* *P* < 0.05, \*\* *P* < 0.01 versus LNCaP sublines alone. (**B**) Cell lysates from co-cultures were subjected to western blotting and probed with antibodies against each protein. Protein levels were compared using actin expression as a loading control. AR, androgen receptor; DHT, dihydrotestosterone; NSE, neuron-specific enolase; PSA, prostate-specific antigen; M5, pcPrF-M5; M6, pcPrF-M6; M7, pcPrF-M7.

**Figure 3.** Effects of growth factors and cytokines on cell proliferation and cellular signaling in LNCaP sublines in vitro. (**A** and **B**) LNCaP sublines were treated with 10 ng/mL of each growth factor and cytokine for 4 days in phenol red (−) RPMI-1640 with 1% CS-FBS containing DHT (0.1 nM). (**A**) Cell proliferation. \*\* *P* < 0.01 versus untreated control. (**B**) Cell lysates from cultures of LNCaP sublines were subjected to western blotting and probed with antibodies against each target protein. Protein levels were compared with actin expression as a loading control. DHT, dihydrotestosterone.

**Figure 4.** Expression of AR-V7 protein in human PCa cell lines. Cell lysates from growing cultures of parental LNCaP cells, LNCaP sublines (E9, F10, and AIDL cells), and 22Rv1 cells were subjected to western blotting and probed with antibodies against each protein. Protein levels were compared using actin as a loading control. 22Rv1 cells were used as a positive control for detection of AR-V7 protein. AR, androgen receptor; AR-V7, androgen receptor splice variant 7.
