Synergistic antitumor interaction of valproic acid and simvastatin sensitizes prostate cancer to docetaxel by targeting CSCs compartment via YAP inhibition

Cell proliferation was measured in 96-well plates in cells untreated and treated with VPA, SIM and DTX as single agent or in combination. Cell proliferation was measured using a spectrophotometric dye incorporation assay Sulforhodamine B [31]. Drugs combination studies were based on concentration-effect curves generated as a plot of the fraction of unaffected (surviving) cells versus drug concentration after 96 h of treatment. Synergism, additivity, and antagonism were quantified after an evaluation of the combination index (CI), which was calculated by the Chou-Talalay equation with CalcuSyn software (Biosoft, Cambridge, UK), as described elsewhere [32]. A CI < 0.9, CI = 0.9–1.2, and CI > 1.2 indicated a synergistic, additive or antagonistic effect, respectively. The dose reduction index (DRI) determines the magnitude of dose reduction allowed for each drug when given in combination, compared with the concentration of a single agent that is needed to achieve the same effect.

The cells (5000 cells/well) were seeded into a 96-well plate and treated for 24 h with VPA, SIM and DTX alone or in combination. The combined caspase 3/7 activity was analyzed in triplicates using the Caspase-Glo® 3/7 Assay (Promega, Madison, WI, USA) according to the manufacturer’s protocol with some modifications. Briefly, after aspirating the medium, 50 μl of Caspase-Glo reagent and the samples were incubated at room temperature for 30 min. Subsequently, the caspase activities were assessed by measuring the luminescence in a Multilabel Reader VICTOR X4 2030 (PerkinElmer, Waltham, MA, USA).

To evaluate CD133 and CD44 surface expression 5 × 10⁵ cells were labeled with PE-conjugated anti-CD133 and FITC-conjugated anti-CD44 antibodies (see Supplementary Methods for antibodies details) for 15 min at 4 °C. Labeled cells were resuspended in Phosphate Buffer Saline (PBS)/0.5% Bovine Serum Albumine (BSA) and analyzed by FACScan flow cytometer (Becton Dickinson, Franklin Lakes, NJ, USA) acquiring 10,000 events for each sample.

Analysis of apoptosis by flow cytometry nuclear DNA staining by propidium iodide (PI) was performed by a FACScan flow cytometer (Becton Dickinson, Franklin Lakes, NJ, USA) acquiring 20,000 events for each sample. The percentage of apoptotic cells was calculated in the sub-diploid region of the DNA content, registered as FL2 signals in linear scale.

Cells were plated in 24-well, flat-bottomed plates using a two-layer soft agar system, as previously described [31]. After 3 h, the cells were treated with VPA and/or SIM at the in vitro IC₂₅^96h of the drugs. The medium (with or without drugs) was replaced every 3 days. The colonies grew for 14 days and were then stained overnight with 3-(4,5-dimethylthiazolyl-2)-2,5-diphenyl tetrazolium bromide (MTT), photographed, analyzed, and counted using Image-Pro-Plus (Immagini and Computer, Bareggio, Milano, Italy). Colonies of > 100 mm were scored as positive.

22Rv1 R_39 DTX-resistant cells were obtained by stepwise selection treating 22Rv1 with increasing doses of DTX (from 0.1 nM up to 6 nM) over 10 months. The selected cells were tested for drug resistance by evaluating the resistance index (RI) = IC₅₀^96h22Rv1 R_39/ IC₅₀^96h22Rv1.

Cells grown and treated as indicated, were washed once with ice-cold PBS and centrifugated. The cell pellet was lyses by Nonidet P40 (Thermo Fisher Scientific, Waltham, MA USA) and clarified by centrifugation. Equal amount of protein, monitored by Bradford assay, was separed on 10% Sodium Dodecyl Phosphate (SDS) polyacrilamide gel electrophoresis (PAGE). Cytosol/membrane extract was obtained according to Baghirova S. et al. [33].

Total RNA was isolated from cells, using Trizol® total RNA isolation reagent (Gibco, Gaitherburg, MD, USA), according to the manufacture’s recommendations. cDNA for qRT-PCR analyses was synthesized with the QuantiTect Reverse Transcription Kit (Qiagen, Valencia, CA, USA). mRNA expression levels were quantified by the fluorescent dye SYBR-green method (Qiagen, Valencia, CA, USA). Gene expression modulation was meseaured by the 2^−ΔΔCT method and normalized to β-actin levels as endogenous control.

Spheroids were cultured as described before [34] in Sphere Medium (DMEM/F12 supplemented with BSA, glucose, heparin, FGF, EGF, neuronal cell culure B27, insulin). The cells (40,000 cells/ml) were plated in low-attachment multiwell plates and treated with indicated drugs. Times and doses of treatments are described in results section. Spheroids were scored with CellTiter-Glo® 3D Cell Viability Assay (Promega, Madison, WI, USA).

Adherent 22Rv1 and 22Rv1 R_39 cells were transfected with YAP wild-type and YAP5SA plasmids as previously described by Noto A. et al. [35] using Lipofectamine 2000 Reagents (Invitrogen, Carlsbad, CA, USA), according to the manufacturer’s recommendation. After 48 h from transfection, cells were collected and western blotting, real-time PCR and immunofluorescent experiments were performed as described before.

Cells, plated on slides in 24-wells plate at 50000 cell/well, were treated with drugs as indicated in figure legends. Then cells were fixed in 4% paraformaldehyde (20 min at RT), blocked by 0.2% PBS/BSA solution (5 min at RT) and incubated with primary anti-YAP antibody for 1 h at 37 °C. After washes, cells were incubated with anti-rabbit Alexa Fluor 488 for 30 min at 37 °C and mounted on slide holder using mountant medium with 4′,6-diamidin-2-fenilindolo (DAPI) (Life technologies, Gaitherburg, MD, USA). Images were taken at 63X magnification by fluorescent microscope (AxioScope A1, Zeiss, Oberkochen, Germany).

The cell pellets (2 × 10⁶ cells) and the tissues (100 mg) were subjected to a chemical extraction using methanol:water:chloroform (700 μL:520 μL:700 μL) as previously reported [36]. The apolar phases were collected, evaporated by SpeedVac system and was re-suspended in 700 μL of deuterated chloroform and trimethylsilylpropanoic acid. A 600-MHz Bruker Avance spectrometer equipped with a cryoprobe was used to acquire ¹H spectra at 300 K for 256 scans. The spectral 0.50–6 ppm regions were integrated by the AMIX package in buckets normalized to the total spectrum area using Pareto scaling and Metaboanalyst tool [37]. We use as reference the proton signal of the cholesterol at 0.66 ppm because it was not overlapped with proton signals of other lipids. Significant differences between the proton signals of the cholesterol were evaluated by T-test and p-values < 0.05.

22Rv1 spheroid cultures were dissociated and live cells were FACS deposited using FACSaria (BD BiosciencesFranklin Lakes, NJ, USA) in a limiting dilution manner at 1, 2, 4, 8, 16, 32, 64 cells per well in ultra-low 96-well plates (Corning, NY, USA) in sphere medium. Stem cell frequency was evaluated after 3 weeks with the Extreme Limiting Dilution Analysis ‘limdil’ function as described by Colak S. et al. [38].

All studies have been performed in compliance with institutional guidelines and regulations (Directive 2010/63/EU; Italian Legislative Decree DLGS 26/2014) and after approval from the appropriate institutional review board (N.865/2015-PR). Five weeks old female NOD/SCID athymic mice (Charles River,Wilmington, MA, USA) were used for 22Rv1, 22Rv1 R_39 xenograft models and 4 weeks old female CD1 nude mice (Charles River,Wilmington, MA, USA) were used for DU145R80 xenograft model. Mice were acclimatized in the Animal Care Facility of Laboratory of Mercogliano (AV) Istituto Nazionale Tumori -“Fondazione G. Pascale” – IRCCS. After 1 week, cells (5 × 10⁶) diluted in 200 μl [PBS/Matrigel GF (Becton Dickinson) 1/1] were injected subcutaneously (s.c) in the flank regions of the mice. Based on pilot studies (data not shown), the mice were treated intraperitoneally (i.p.) with VPA (melted in water and diluted in a physiological solution) and SIM (melted in DMSO and diluited in physiological solution), plus DTX (melted in DMSO and diluited in physiological solution) once a week at the indicated concentrations. Mice in the control groups were treated with both physiological solution and/or DMSO plus physiological solution 1:1. Tumor volume (TV) (mm³), Tumor growth delay (TGD) and the percent change in the experimental groups was compared with that of the vehicle control groups as described before [29]. Tumor incidence curves to analyze tumor engraftment (first appearance of a palpable mass) was performed taking advantage of Kaplan-Meier approach.

We investigated the antitumor effect of VPA in combination with SIM in a panel of PCa cell lines (PC3, DU145, LNCaP, 22Rv1) with different molecular features. All cell lines resulted sensitive to the antiproliferative effects of both agents in monotherapy (Supplementary Table S1). DU145R80 cells, selected for resistance to the inhibitor of the prenylation arm of MVP zoledronic acid (ZOL) [39], were cross-resitant to SIM (RI of DU145R80 vs parental DU145 cells: 12.77) and sensitive to VPA. Then, we combined the two drugs, exploring different cytotoxic ratios, either equipotent doses (50:50 ratio) or one of the two drugs in excess (75:25 and 25:75 ratio) (Supplementary Table S2), and different treatment schedules, either simultaneously or sequentially (24 h delay between the two agents) (Supplementary Table S3).

We obtained consistent antitumor synergistic effects with low CIs, calculated at 50% (CI₅₀) of cell lethality, independently from the ratio of the two drugs used or the schedule tested, in all cell lines, except the LNCaP cells where an additive/antagonistic effect was observed (Fig. 1a; Supplementary Table S2). Interestingly, we also demonstrated that VPA treatment completely reverts SIM-resistance in DU145R80 cells, suggesting an impact of HDACi on MVP (Fig. 1b).

The synergistic antiproliferative effect induced by VPA/SIM correlated with a significant induction of apoptosis measured as caspase 3/7 activity after 24 h at IC₅₀^96h, with the exception of DU145R80 where only a slight pro-apoptotic effect was observed (Fig. 1c). Notably, in normal epithelial EPN cells we did not observe any pro-apoptotic effect of either agent or the combination, suggesting a selective action on tumor cells (Fig. 1c).

We also confirmed the synergistic antitumor effect of VPA/SIM combination in anchorage-independent condition on 22Rv1 (colony formation inhibition: VPA ∼ 58%; SIM ∼ 43%; VPA + SIM ∼ 86%) using low doses (IC₂₅^96h) (Fig. 1d), and similar data were obtained in DU145 and DU145R80 cells (Supplementary Fig. S1A and S1B).

Finally, to better recapitulate tumor growth complexity, we tested VPA/SIM combination also on PCa 3D-self-assembled spheroids. For these experiments we focused on 22Rv1 spheroids since this cell line resulted the most suitable for the growth in low attach condition using sphere medium compared with the other PCa cell lines (Fig. 1e and Supplementary Fig. S2A). We used different approaches to highlight different effects: (a) by evaluating treatments on 1st generation sphere formation (cells plated in low-attached plate in sphere medium and concomitantly treated), we investigated the capacity of treatment to prevent/reduce tumor formation (spheres A); (b) by treating 2nd generation sphere formation (cells were grown for 72 h, then disaggregated and plated again in the presence of drugs), we evaluated the impact of treatment to prevent/reduce more aggressive tumors (spheres B); (c) by treating formed-spheres (spheres allowed to grow for 72 h and then treated), we evaluated the capacity of treatment to induce tumor regression (spheres C). Our results showed that VPA/SIM combination, compared to single agents, strongly inhibits spheroid formation (spheres inhibition vs control: ∼76% in spheres A, ∼81% in spheres B), and induced ∼56% formed-sphere regression in spheres C vs control (Fig. 1e). Notably, compared to cell adhesion condition, 1st and 2nd generation spheres are normally described as enriched in CSC compartment [40‐42] with self-renewal capacity. Indeed in both these 22Rv1 3D-models we showed the increased expression levels of CSC markers such as NANOg and OCT4 (Supplementary Fig. S2B-C) as well as CD44⁺ and CD133⁺ surface expression, compared to adherent cells (Supplementary Fig. S2D).

To investigate whether the synergistic interaction between VPA and SIM occurred via MVP (schematically summarized in Fig. 2a) we evaluated the antitumor effect of the single agents or the combination, in the presence or absence of mevalonic acid (Mev), that overcomes the inhibition of HMGCR activity. Notably, the addition of Mev antagonized both the synergistic antiproliferative (Fig. 2b) and pro-apoptotic effect (Fig. 2c-d) induced by VPA/SIM combination on 22Rv1 cells grown in adherent condition or as spheres A (Fig. 2e).

To further evaluate the impact of treatment on putative CSCs, we analyzed the effect of the combination using an additional spheres growth system [34] (spheres D - Fig. 2f). In detail, 22Rv1 cells grown as spheroids were treated in 1st generation with VPA and SIM as single agents or in combination with or without Mev for 72 h; survived spheroids, were then disaggregated and plated again to form 2nd generation spheroids without additional treatment. Remarkably, a single VPA/SIM combination treatment in 1st generation, is able to affect 2nd generation spheroids formation (∼57% of inhibition vs control) and this effect was completely reverted by the addition of Mev (Fig. 2f).

Finally, as a readout of MVP inhibition we investigated the cholesterol content of 22Rv1 cell line in the different treatment setting, taking advantage of ¹H-NMR metabolomic analysis of the cellular lipophilic (apolar) phase. As shown in Fig. 2g we observed a clear reduction of cholesterol content upon SIM treatment or in the combination setting and a slight reduction upon VPA treatment while all these effects were reverted by Mev.

Overall these data suggested that the synergistic interaction between VPA and SIM in PCa models could occur by targeting CSCs compartment via concurrent inhibition of the MVP.

To further disclose the molecular mechanism behind the synergistic antitumor interaction of VPA/SIM combination we performed an ingenuity pathway analysis (IPA) on “mevalonate pathway enzymes” and “HDAC inhibitors” combined search. As shown in Fig. 3a we revealed a network with direct and indirect relationships connecting HDAC1 and MVP enzymes (i.e. HMGCR, HMG-CoA synthase), as well as the transcription factors SREBF1 and 2 regulating MVP genes expression, all together confirming a functional relationship between the targets of our treatment combination. Indeed, we demonstrated the reciprocal ability of both VPA and SIM to target histone acetylation within 24 h (Supplementary Fig. S3A) and HMGCR mRNA expression within 2 h and up to 8 h of treatment at the IC₅₀^96h (Supplementary Fig. S3B), in 22Rv1 cells. Moreover, at similar early time point (4 h), both VPA and SIM were able to reduce specifically HDAC1 and HDAC2 mRNA expression (Supplementary Fig. S3C), but not HDAC3 and HDAC6 (data not shown). Furthermore, upon VPA or SIM treatment we also showed the increase of RhoA cytoplasmatic and inactive form, that was reverted by either Mev or GGOH (Supplementary Fig. S3D), confirming the ability of both drugs to affect the MVP prenylation arm (see Fig. 2a).

However, the IPA network reported in Fig. 3a also highlighted additional hubs such as AMP-activate protein kinase (AMPK) or p53, both known regulators of the MVP. AMPK, a sensor of cellular energy status, is a known regulator of HMGCR activity [43, 44] and can be activated by both HDACi and statins [45]. We confirmed that the activating AMPK Tyr172 phosphorylation is induced by either VPA or SIM within 40 min (Supplementary Fig. S4A). Notably, this effect is paralleled by the induction of HMGCR inhibitory phosphorylation occurring within 2 h of VPA or SIM treatment (Supplementary Fig. S4A). Mutant p53 has been shown to trigger in cancer cells the MVP, leading to the aberrantly activation of YAP, an essential oncogene for cancer initiation/growth of most solid tumors, including PCa, an effect that can be reverted by statins [46]. Interestingly, recent reports also highlighted the ability of AMPK to regulate the Hippo pathway, directly inducing YAP inhibitory phosphorylation [47‐49]. On these bases we evaluated YAP expression and activity in both p53 mutant 22Rv1 and p53 null PC3 cells upon VPA and/or SIM treatment. One of the most critical findings of our study was the clear synergistic induction of YAP inhibitory Ser127-phosphorylation induced by VPA/SIM combination within 4 h of treatment (Fig. 3b) and up to 24 h, (Supplementary Fig. S4B) paralleled by the synergistic activation of AMPK, its well-known downstream substrate acetyl-CoA carboxylase (ACC) and the inhibitory phosphorylation of HMGCR, in both 22Rv1 and PC3 cells (Fig. 3b, Supplementary Fig. S4C-D).

As a consequence of increased Ser127-phosphorylation induced by VPA/SIM combined treatment, YAP protein was retained in the cytoplasm (Supplementary Fig. S5A) and cannot translocate into the nucleus (Supplementary Fig. S5B). Indeed a clear reduction of YAP direct and indirect transcriptional targets CTGF, Cyr61, BIRC5 and NANOg [50‐52] was observed upon VPA/SIM combined treatment (Fig. 3c). Notably, the inhibition of YAP activation was completely reverted by Mev which bypasses HMGCR inhibition, or GGOH, which bypasses prenylation arm inhibition (Supplementary Fig. S5A and S5B), thus confirming that YAP inhibition is dependent on VPA/SIM synergistic inhibition of MVP, at least in part via AMPK activation. Indeed, pharmacological inactivation of AMPK, with the specific inhibitor compound C, partially reverts the antiproliferative and apoptotic effect induced by VPA/SIM combination, both in PCa adherent cells and in 3D spheroids (Supplementary Fig. S6). Consistently with all the data reported above, the additional IPA network (Supplementary Fig. S5C), obtained by combining HMGCR and AMPK search, highlighted as hubs HDAC2 as well as CYR61, BIRC5 and CTGF YAP-transcriptional targets, further corroborating our results.

Interestingly compared to cell adhesion condition, 22Rv1 1st generation spheres showed an increased expression of YAP (Supplementary Fig.S2C) and one of its target gene CTGF (Supplementary Fig.S2E). We confirmed increased YAP inhibitory Ser127-phosphorylation induced by VPA, SIM and VPA/SIM combination in 22Rv1 3D spheroids (Fig. 3d) paralleled by a significant reduction of one of its target gene BIRC5 (Fig. 3e) and, most importantly, by the impairment of CSC CD44⁺/CD133⁺ surface markers (Fig. 3f).

Anyhow, the observed inhibition of YAP in our cell models is most likely p53-independent, because has been demonstrated in both 22Rv1 mut- and PC3 null-p53. Interestingly, LNCaP castration-sensitive wt-p53 cells, where we did not observe a VPA/SIM synergistic antitumor effect (Fig. 1a), expressed lower basal levels of AMPK, HMGCR and YAP (Supplementary Fig. S5D). Moreover, in this cell line we did not observe neither AMPK activation upon single agents or VPA/SIM combination, nor YAP increased inhibitory phosphorylation by combination treatment (Supplementary Fig. S5E). Furthermore, in this cell line combined VPA and SIM treatment significantly downregulated p53 levels (Supplementary Fig. S5E). Hence, the mechanism of the lack of VPA/SIM synergism in LNCaP cells require further investigation. Noteworthy, the three PCa cell lines where we reported a clear VPA/SIM synergistic antiproliferative and pro-apoptotic effect (PC3, 22Rv1, DU145), expressed high baseline protein levels of AMPK, HMGCR and YAP compared to EPN and LNCaP cells (Supplementary Fig. S5D), again suggesting that the mechanism by which VPA and SIM synergized requires the coordinated activation/addiction of/to these pathways. SIM-resistant DU145R80 cells, although expressed similar high levels of AMPK, showed reduced HMGCR and YAP basal protein levels, potentially related to the mechanism of SIM-resistance (Supplementary Fig. S5D).

Next, in order to further confirm our observations, we evaluated the ability of VPA/SIM to affect YAP nuclear localization and activation, in 22Rv1 transiently transfected with either wild-type YAP (wt-YAP) or with the constitutively active mutated form YAP5SA [35] (Fig. 4a). As expected, we found that VPA/SIM were not capable to induce increased YAP-ser127 inhibitory phosphorylation as well as pro-apoptotic effect, as measured by increased Caspase 3 expression and PARP cleavage, in YAP5SA-transfected compared with non-trasfected or wt-YAP transfected 22Rv1 cells (Fig. 4b). Consequently, YAP nuclear translocation was not inhibited by the combined VPA/SIM treatment in YAP5SA-transfected compared with YAP wt-transfected cells, as shown by immunofluorescence experiments (Fig. 4c). Significantly a clear increased mRNA expression of CTGF and Cyr61, not signicantly changed upon treatments, was observed in YAP5SA-transfected cells compared with control non-transfected or wt-YAP transfected cells (Fig. 4d). Finally, to further investigate the effect of VPA/SIM combination on CSCs compartment in 22Rv1 control or YAP5SA-transfected cells we performed a limiting dilution assay, demonstrating a drammatic reduction in stem cell frequency induced by VPA/SIM combination (∼80% reduction), which was clearly lost in YAP5SA- transfected cells (∼43% of reduction), considering that the latter cell line also express endogenous wt-YAP (Fig. 4e).

All together, these data suggested that VPA/SIM combination induced antitumor effect in PCa models targeting CSCs compartment, via MVP-driven inhibition of YAP activation.

To explore the clinical relevance of YAP targeting in PCa we generated a signature of four genes induced by YAP (CTGF, CYR61, BIRC5 and ANRDK1) and interrogated the prostate adenocarcinoma Cancer Genome Atlas (TCGA). Notably, the four genes directly modulated by YAP were all highly enriched in the tumor tissues of relapsed PCa patients after curative resection compared with tumor free patients, suggesting a correlation of YAP activation with PCa prognosis and thus as potential drug target (Supplementary Fig. S7).

Several evidences suggest that the CSCs compartment critically contributes to chemoresistance. Thus, we next explore the potential of VPA/SIM combination to sensitize PCa cells to DTX. We investigated the triple combination of VPA, SIM and DTX at equitoxic concentrations either simultaneously or sequentially (with a 24 h delay between concomitant VPA/SIM and DTX or vice versa) (Table 1). We obtained consistent synergistic anti-proliferative effects of the triple combination VPA/SIM/DTX with the lowest CI₅₀ values in all the four PCa cell lines tested, compared with dual combinations (VPA/DTX or SIM/DTX), and independently of the schedule used. The clear potentiation of DTX cytotoxic effect by VPA/SIM combination was also confirmed by the dose reduction indexes (DRIs), the order of magnitude (fold) of dose reduction obtained for the IC₅₀ (DRI₅₀) in combination vs single drug treatments, which ranged for DTX, among the cell lines tested, from 4 up to 20-fold, in triple combination treatment (Table 1). Furthemore, we also showed a synergistic induction of apoptosis in 22Rv1 (Fig. 5a-b), DU145 and PC3 cells (Supplementary Fig. S8A-B), and of DNA damage assessed as H2AX phosphorylation (γH2AX) in 22Rv1 cells (Fig. 5c), by the triple VPA/SIM/DTX combination vs single agents or dual combinations. Notably, in EPN normal epithelial cells, we did not potentiate the pro-apototic effect of DTX in either dual or triple combination treatments (Supplementary Fig. S8A-B).

The effect of the triple combination was further investigated in two different 3D 22Rv1 cell colture systems such as hanging-drop microtissues (Supplementary Fig. S8C-D) and self-assembled spheroids (Supplementary Fig. S8E). We confirmed a strong inhibitory effect of VPA/SIM combination on both microtissues or spheres formation and regression but we also observed a potentiation of DTX effect, which, as single agent, was poorly effective. More importantly, in sphere D system, the pre-treatment with VPA/SIM combination of 22Rv1 cells during the 1st generation of spheroids formation, strongly potentiated the efficacy of DTX given alone to surviving cells during 2nd generation of spheroid formation, sustaining again the hypothesis that VPA/SIM combination, is able to sensitize PCa tumorsphere to DTX treatment by targeting the CSCs compartment (Fig. 5d).

Consistently with the latest results, taking advantage of the DTX-resistant 22Rv1cells (22Rv1_ R39), generated in our lab by stepwise incresing concentrations of DTX (DTX resistance index vs parental 22Rv1 cells: 39.5), we also demonstrated that VPA/SIM combination was able to revert DTX resistance (Fig. 6a). Indeed we confirmed a clear synergism in triple combination (VPA/SIM/DTX), with a CI₅₀ smaller then 0.9 and DRI₅₀ for DTX of up to 180 fold in DTX-resistant 22Rv1_R39 cells (Fig. 6b). Notably, compared to 22Rv1 parental cells 22Rv1 R_39 showed an increased basal level expression of NANOg mRNA both in adhesion and 1st generation spheres conditions (Fig. 6c) and reduced basal YAP-ser127 inhibitory phosphorylation (Fig. 6e). Moreover VPA and/or SIM treatment clearly reduced CD44⁺/CD133⁺ subpopulation in 22Rv1 R_39 grown as CSC enriched 1st generation spheroids (Fig. 6d, Supplemetary Fig.S9A). Consistently, by using the sphere D system, we confirmed VPA/SIM ability to revert DTX resistance on pre-treated tumorsphere, but this effect was partially loss in 22Rv1 cells transfected with mutated YAP (< 79% vs < 43% of spheroid formation reduction in 22Rv1 cells and 22Rv1-YAP5SA, respectively) (Fig. 6e-f; Supplemetary Fig. 9B).

Altogether these evidences demonstratd that VPA/SIM combination potentiate the efficacy of the chemotherapeutic DTX and revert DTX-resistance, by targeting CSCs compartments via YAP-activity inhibition.

Finally, to assess whether the synergistic antitumor effects demonstrated in vitro could be confirmed in vivo we evaluated VPA and SIM in combination with DTX in both 22Rv1 and DU145R80 xenograft models as well as in 22Rv1 R_39 DTX resistant cells. This was accomplished through the measurement of tumor volume (Fig. 7a-b and f), the percent of tumor volume change and the tumor growth delay (TGD) (Supplemetary Fig. S10A-C). Specifically, twenty-eight mice were injected with either 22Rv1 or DU145R80 cells, and randomly assigned to four groups to receive DTX (10 mg/Kg i.p weekly for 2 weeks), VPA/SIM combination (200 mg/Kg and 2 mg/Kg, respectively, i.p. daily for 2 weeks), the triple combination, or their vehicles. Notably, the dosages of DTX, VPA and SIM correspond or were lower to those reported in the literature for in vivo experiments [32, 53, 54], and were consistent with corresponding doses used in patients [32, 54, 55].

As shown in Fig. 7a vehicles-treated 22Rv1 xenografted tumors grew rapidly and reached the endpoint size within 3 weeks; at this time-point DTX was not effective, VPA/SIM combination produced a clear statistically significant tumor growth inhibition compared with control mice, while triple combination treatment induced a further significant inhibition of tumor growth. Notably, in SIM-resistant DU145R80 xenograft model (Fig. 7b), where the endpoint size of control tumors was reached in 4 weeks, triple combination was able to completely block the tumor growth, compared with controls or other treated groups. VPA/SIM combination slightly reduced tumor growth only at last time point (day 32), while DTX alone clearly reduced growth in this tumor model. The combined treatment was well tolerated by both 22Rv1 and DU145R80 cells xenografted mice, as shown by the maintenance of body weight (Fig. 7a-b, insets) and by the absence of other signs of acute or delayed toxicity. Moreover, in 22Rv1 model, by calculating the percent change in tumor volume from the time of initial treatment (day 0) to the end of the study (day 21), VPA/SIM and triple combination treatment reduced the tumor burden by 15 and 29%, respectively, in spite of DTX that did not reduce tumor burden (Supplementary Fig. S10A). Conversely, in DU145R80 from day 7 to the end of the study (day 32) the triple combination reduced tumor burden by almost 90% compared with 40% reduction induced by VPA/SIM and about 62% by DTX (Supplementary Fig. S10B). In 22Rv1 xenograft model the synergistic intereaction between VPA/SIM and DTX was confirmed also by the evaluation of the TGD induced by the triple combination that reached a peak of more than 100% indicating that the mean rate of tumor growth in the control were approximately 2-fold higher, while compared to VPA/SIM the rate of control tumors was approximately 1.5-fold higher (Supplementary Fig. S10A). In DU145R80 xenograft model the mean rate of tumor growth in the control compared to the triple combination was approximately 3-fold higher (Supplemetary Fig. S9B). Moreover, we demonstrated also a clear increase of PARP cleavage in all the triple combination tumor samples compared with the other groups, in line with in vitro data showing increased apoptosis, paralleled by induction of DNA damage, measured as H2AX protein phosphorylation (Fig. 7c). Induction of AcH3 and HMGCR protein expression was used as a read-out of VPA and SIM, respectively (Fig. 7c) [24, 56]. Moreover, in agreement with our in vitro findings, a significant increase of phospho-YAP (Ser127) (Fig. 7c), together with a clear reduction of YAP-target genes BRC5 (Fig. 7d, left panel) and NANOg (Fig. 7d, right panel), were shown in both VPA/SIM and VPA/SIM/DTX treated mice compared with controls or DTX-treated tumors, further supporting our hypothesis. Furthermore, as shown in Fig. 7e VPA/SIM combination reduced cholesterol content also in vivo in tumor samples, and this effect is further potentiated in triple combination setting.

Next, we confirmed our previous observation also in 22Rv1 R_39 DTX resistant cells xenograft model. In details, fiftyfour mice were injected with 22Rv1 R_39 cells, and ten days after implantation the mice were randomly assigned to six groups to receive DTX, VPA and /or SIM, and triple combination, or their vehicles, for ten days, at the dosages indicated above for parental cells xenograft model. As shown in Fig. 7f VPA/SIM and triple combination produced a clear statistically significant tumor growth inhibition compared with control and single treatments groups. This effect was confirmed through the measurement of tumor volume (Fig. 7g, left panel) and tumor weight (Fig. 7g, right panel) ex vivo. The combined treatment was well tolerated by 22Rv1 R_39 cells xenografted mice, as shown by the maintenance of body weight (Fig. 7f, insets). Moreover, by calculating the percent change in tumor volume from the time of initial treatment (day 3) to the end of the study (day 13), VPA/SIM and triple combination treatment reduced the tumor burden by 29.2 and 34.2%, respectively, in spite of the other treated groups that did not reduce tumor burden (Supplemetary Fig.S10C). In addition, the mean rate of tumor growth in the control compared to the double and triple combination was approximately 1,3-fold higher for both groups (Supplemetary Fig. S10C). Interestingly, tumor incidence curves analyzing tumor engraftment (first appearance of a palpable mass) in cohorts of 9 mice/group injected with either 22Rv1 or 22Rv1 R_39 cells, showed that 22Rv1 cells xenograft model developed tumors with a latency of 8 days compared to 22Rv1 R_39 cells xenograft model, suggesting increasing aggressivity of the resistant subline, potentially related with CSC enrichment (Supplemetary Fig.S10D).

Overall we confirmed the efficacy of VPA/SIM combination to potentiate the antitumor effect of DTX and to revert DTX resistance also in vivo in PCa tumor models by targeting the MVP/YAP axis.