The androgen receptor (AR) is the most critical factor in prostate cancer progression. We previously demonstrated that silencing the AR using 2 unique small interfering RNAs (no. 8 and no. 31 AR siRNA) induces apoptotic cell death in AR-positive prostate cancer cells. To develop this AR siRNA technique into a therapy for prostate cancers, we generated an adeno-associated virus (AAV) vector to stably express a short hairpin-structured RNA (shRNA) against the AR gene in vivo. In addition to the no. 8 AR shRNA (ARHP8), we also screened a group of AR shRNAs with different sequences and identified a less effective AR shRNA (ARHP4) that was used as an shRNA control. An empty AAV vector (AAV-GFP) was used as a negative control. Intratumoral injection of AAV-ARHP8 viruses significantly suppressed tumor growth of xenografts derived from either androgen-responsive or castration-resistant prostate cancer cells. Most interestingly, systemic delivery of the AAV-ARHP8 but not AAV-ARHP4 or AAV-GFP viruses via tail vein injection eliminated xenografts within 10 days. Further analysis revealed that AAV-ARHP8 viruses dramatically reduced the expression of AR-regulated cellular survival genes and caused a dramatic apoptotic response. Taken together, our data strongly suggest that AAV-ARHP8 viruses induced a strong AR gene silencing in vivo and that systemic delivery of ARHP8 siRNA via an AAV vector or any other means might be considered as novel gene therapy for prostate cancers.
Clinically, nearly all prostate cancers including castration-resistant diseases retain a functional androgen receptor (AR) signaling pathway, which is the most critical factor in prostate cancer progression.1 Current evidence favors a mechanism by which aberrant activation of intracellular signal transduction pathways stimulates the AR in the presence of residue ligand after chemical or surgical castration or even in case of androgen antagonist.2–5 Thus, it is being considered that targeting the AR itself is more relevant compared to ablating its ligand that often results in clinical failure.3, 4
RNA interference (RNAi) is a mechanism of post-transcriptional gene silencing, which can be activated by introducing small double-stranded interfering RNA molecule (siRNA) corresponding to any endogenous gene of interest, resulting in mRNA degradation of the targeted gene.6 Therefore, siRNA is referred as an extremely powerful and simple method for assessing or even eliminating gene function, although the gene-silencing efficiency of a given siRNA is largely sequence-specific.7 To maintain sustained gene silencing in cells or organism, a common approach is to stably transduce cells with a short hairpin-structured RNA (shRNA) under the control of a customized promoter in cells.8–11
Recently, siRNA-based strategies have been used to target the AR gene by different groups,1, 12–18 of which a profound apoptotic response in vitro was observed from our work12 and another group.13 In contrast, other groups reported a cytostatic or neuro-endocrinal trans-differentiation effect in vitro and tumor suppression in vivo.14–18 These different results might be due to the efficiency of AR gene silencing induced by different siRNA sequences and experimental conditions.
Adeno-associated virus type 2 (AAV-2) is a nonpathogenic human parvovirus and has been developed as a gene delivery vector for treatment of numerous diseases.19 For gene delivery purpose, all wild-type AAV open reading frames (ORFs) are replaced with a customer-favored transgene expression cassette. It has been shown that recombinant AAV (rAAV) is capable of transducing a broad range of cell types without depending on active division of the host cell, and currently, rAAV has been used extensively as gene delivery vehicles to transduce a wide range of cells in vitro and in vivo.20–22 In addition, AAV DNA has been found in human semen and testis tissue,23 representing a permission of viral transduction for prostate-derived cells.
In attempt to develop novel therapeutic approaches for prostate cancers, we sought to use recombinant AAV approach to express the no. 8 AR shRNA, as described in our previous publications,12, 24 in prostate cancer cells. We have shown that AAV-ARHP8 viruses induced AR silencing and apoptotic cell death in AR-positive prostate cancer cells.24 In the present study, we tested the efficiency of AAV-ARHP8 on tumor growth in a preclinical mouse xenograft model of prostate cancer. AAV-ARHP8, along with other control AAV viruses, was delivered into prostate cancer xenografts in nude mice locally or systemically. Surprisingly, when the most effective AAV-ARHP8 viruses were delivered systemically in nude mice through tail vein injection, xenograft tumors disappeared within 10 days, while a less effective AAV-ARHP4 virus only partially reduced tumor growth compared to a negative virus (AAV-GFP).
AAV: adeno-associated virus; AR: androgen receptor; ARHP: androgen receptor hairpin-structure RNA; BrdU: 5-bromo-2′-deoxyuridine; FBS: fetal bovine serum; mRNA: messenger RNA; PBS: phosphate-buffered saline; PCR: polymerase chain reaction; rAAV: recombinant AAV; RIPA: radio-immunoprecipitation assay; RNAi: RNA interference; RT: reverse transcription; s.c.: subcutaneous; SEM: standard error of mean; shRNA: short-hairpin RNA; siRNA: small interfering RNA; SGK-1: serum/glucocorticoid-induced protein kinase 1; TBS-T: Tris-buffered solution plus Tween 20; TEM: transmission electron microscope; TUNEL: terminal deocynucleotide transferase dUTP nick end labeling
Material and methods
Cell culture and reagents
Human prostate cancer cell lines LAPC-4, LNCaP, C4-2, 22Rv1, PC-3 and PC-3/AR were grown in the conditions as described previously in our publication.12 Antibodies for green fluorescent protein (GFP), serum/glucocorticoid-induced kinase 1 (SGK-1), antiapoptotic protein Bcl-xL and cytoskeleton protein Actin were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-AR antibodies (Clone PG21) were obtained from Upstate (Charlottesville, VA). Charcoal-stripped fetal bovine serum (cFBS, steroid-depleted) was obtained from Atlanta Biologicals (Norcross, GA).
Plasmid construction and AAV production
To silence AR gene expression in vivo, a hairpin-structured expression vector was generated using pSilencer-U6 1.0 plasmid vector (Ambion, Austin, TX) under the guidance of the manufacturer's manual. The sense sequences for the ARHP constructs are listed in Table 1. The sequences for no. 8 and no. 31 were reported in our previously publication.12 Oligonucleotides assembling individual hairpin-structure cassettes were synthesized by IDT (Coralville, IA). The AAV Helper-free system (Stratagene, La Jolla, CA) was utilized to produce the AAV particles, as described in our recent publication24 for AAV-ARHP8 and AAV-ARHP31. Briefly, annealed DNA fragments were subcloned into ApaI/EcoRI sites of the pSilencer-U6 vector to generate constructs of pSilencer-U6-ARHP#(1,2,4,5,7,34). To produce an AAV vector for ARHP expression, the ARHP expression cassette was released from pSilencer-U6-ARHP by XbaI digestion and subcloned into pCMV-MCS from the AAV Helper-free system. Then, this fragment was moved into pAAV-IRES-hGFP vector at the EcoRI/XhoI sites to generate AAV-ARHP viral-producing vector (i.e., pAAV-ARHP4-IRES-hGFP). The pAAV-IRES-hGFP construct was used to produce an empty AAV virus (AAV-GFP) that expresses the GFP protein only. The AAV particles were produced, packaged and purified by Applied Viromics LLC (Fremont, CA) at a titer up to ∼2 × 1012 viral particles/ml.
Table 1. Sense sequences of AR hairpin siRNA
RNA extraction, RT-PCR and Western blot assay
Total RNA was prepared using TriZol™ reagent (Invitrogen, Carlsbad, CA) from cultured cells, xenograft tissues and mouse organs. To assess mRNA expression, RT-PCR was carried out using a RETROscript™ kit (Ambion, Austin, TX). The primers for the AR gene, AR-regulated genes PSA, Bcl-xL and SGK-1, as well as 28S ribozyme RNA, were described in our recent publications.12, 24 Primers for GFP gene are 5′-ACG GCA AGC TGA CCC TGA AG-3′ and 5′-GGG TGC TCA GGT AGT GGT TG-3′. The oligonucleotide primers were synthesized by IDT (Coralville, IA). The amplification profile was as follows: 95°C for 30 sec, 56°C for 30 sec, and 72°C for 1 min running in 25 cycles. After 25 amplification cycles, the expected PCR products were size-fractionated onto a 2% agarose gel and stained with ethidium bromide. For semiquantitative RT-PCR assay, serial dilution of the total RNA samples were utilized for conducting the first-strand cDNA synthesis, as described in our previous report.25
For protein analysis, xenograft tumors were snap-frozen and stored at −80°C before processing. Proteins were extracted from cultured cells and xenograft tumors as described in our previous study.26 Equal amount of proteins was subjected to SDS-PAGE and then Western blot for assessing AR and GFP expression. Actin blot served as protein-loading control.
Mouse xenografts, castration and AAV virus injection
Athymic male mice (Charles River, Wilmington, MA) were maintained in accordance with the Institutional Animal Care and Use Committee (IACUC) procedures and guidelines. Xenograft tumors were generated as described in our recent publication.26 Briefly, exponentially grown prostate cancer cells were trypsinized and resuspended in PBS. A total of 2.0 × 106 cells was resuspended in RPMI-1640/10%FBS with a 4:1 v/v ratio of Matrigel™ (BD Bioscience, Bedford, MA), and was injected subcutaneously (s.c.) into the flanks of 6-week-old mice using a 27-gauge needle and 1-ml disposable syringe. The total volume was 0.1 ml/site at 2 sites per mouse. When tumors were palpable (∼30–50 mm3 in 4–6 weeks), animals were castrated or sham-operated through scrotal incision if necessary. As described in figure legends, AAV viruses were injected either locally into the tumor mass or systemically via tail vein. Tumor growth was monitored by measuring the length (L) and the width (W), and the volume was calculated by the formula of V = (L × W2)/2. Wet weights of dissected xenograft tumors were recorded at the end of experiment.
BrdU incorporation assay, TUNEL assay and immunohistochemistry
To label proliferating cells in vivo, on the last day of experiment, 1 hr before sacrifice, animals were injected intraperitoneally with 0.5 ml of a 10 mM BrdU solution that was obtained from an in situ proliferation assay kit (Roche Diagnostics, Indianapolis, IN), as recommended by the manufacturer and described previously in our publication.26 Immunohistochemistry for integrated BrdU in the nuclear compartment was conducted using the Roche-supplied BrdU kit. Apoptotic cell death was determined by in situ TUNEL analysis with the ApoAlert® DNA fragmentation assay kit (Clontech, Mountain View, CA), as described in our publication.26
For evaluating GFP expression, xenograft specimens were snap-frozen in liquid nitrogen and embedded in OCT frozen-embedding media. Frozen sections were prepared and viewed under a fluorescent microscope.
For immunostaining analysis of gene expression, specimens were fixed in 4% paraformadehyde and paraffin-embedded. Tissue sections (4 μm) were stained with hematoxylin and eosin (H&E) to evaluate the tumor structure. Primary antibodies for AR, SGK-1 and Bcl-xL were purchased from Santa Cruz Biotech (Santa Cruz, CA). Immunostaining was carried out as described in our publications.26, 27
Transmission electron microscope of the xenografts
After dissection, xenograft tumor samples were chopped into small pieces (3 × 3 mm2) and fixed in modified Karnovsky's buffer (2.5% gluteraldehyde/2% paraformaldehyde in 0.1 M Sodium Cacodylate) at 4°C till further processing. Postfixation was carried out in 0.2 M osmium tetroxide. Ultrathin sections were cut with a LKB Nova Ultramicrotome. Transmission electron microscope (TEM) images were taken using JEOL 100CX II Transmission Microscope. The magnification is indicated in the figure legend.
Western blot, immunostaining and RT-PCR results were presented from a representative experiment. The mean and standard error of the mean from animal experiments are shown. The significance of the differences between treatment and control was analyzed using the SPSS software (SPSS, Chicago, IL).
AAV-ARHP8 viruses induce AR gene silencing in vivo
Previously, we reported that a siRNA duplex against human AR gene (no. 8 AR siRNA) caused a massive apoptotic cell death in AR-positive prostate cancer cells.12 Later on, based on this no. 8 AR siRNA sequence, we generated a recombinant AAV vector that expresses a hairpin-structured RNA (AAV-ARHP8) to silence the AR gene.24 In this study, we used AAV-ARHP8 virus to test its in vivo effect of AR silencing on tumor growth of prostate cancer xenografts.
First, a pilot experiment was carried out to determine a proper viral dose for achieving a successful AR gene silencing in mouse xenograft model of prostate cancer. Xenograft tumors were generated in nude mice using prostate cancer cells PC-3/AR, which was established by stably expressing human AR gene in PC-3 cells and which did not respond to AR gene silencing, as described in our previous publication.12 When tumors were palpable (around 50 mm3 in size), 7 different doses (log-dilution, 5 × 103–5 × 109 viral particles in 20 μl total volume) of the AAV-ARHP8 viruses were injected locally into xenograft tumors. In addition, 2 other animals received an empty control virus AAV-GFP (2 × 109 viral particles in 20 μl), or 20 μl PBS as the negative controls. One week later, xenograft tumors were harvested for gene expression analysis. Western blot results showed gradually increasing levels of GFP protein along with the increasing doses of AAV viruses (Fig. 1b), reflecting a successful dose-dependent expression of AAV-mediated gene expression.
Next, we checked the efficiency of AR gene silencing at both mRNA and protein levels. As shown in Figure 1b, AAV-ARHP8 virus injection dramatically reduced AR expression in a dosage-dependent fashion. The peak effect was seen at a dose equal to or higher than 2 × 106 particles per 50 mm3 of tumor volume. However, AAV-GFP viruses did not cause any notable changes in AR mRNA (Fig. 1a) and protein levels (Fig. 1b) compared to the PBS control. These data confirmed that AAV-ARHP8 viruses successfully induced AR gene silencing in vivo.
Locally injected AAV-ARHP8 viruses completely block tumor growth in vivo
We then tested if AAV-ARHP8 virus suppresses tumor growth in vivo. Prostate cancer LAPC-4 cell line, which harbors a wild-type AR gene and is androgen responsive, was used in the initial experiment. Once xenografts were palpable (30–50 mm3), animals were randomly divided into 3 groups to receive intratumoral injection of PBS, AAV-ARHP8 viruses or AAV-GFP viruses. Xenograft tumors were monitored for 4 weeks. As shown in Figure 2a, tumor growth was significantly suppressed in AAV-ARHP8 group compared to that in AAV-GFP or PBS group. Tumor wet weights were also significantly lower in AAV-ARHP8 viruses-injected tumors than that in AAV-GFP group. There was no significant difference in animal body weight at the end of experiments between AAV-ARHP8 and AAV-GFP groups (Fig. 2b). No visible side effect from the mice was noticed during the experiments. These data suggest that AAV-ARHP8 viruses suppressed tumor growth in vivo.
We next expanded our observation to multiple xenograft models derived from different prostate cancer cells. We also compared tumor growth under castrated or gonad-intact conditions. First, androgen-responsive LNCaP cell line, which displays a castration-resistant phenotype after a short period of castration,28 was used. Once LNCaP xenografts were palpable, animals were randomly divided into 2 groups to receive bilateral castration or sham operation. One day after operation, animals in each group received an intratumoral injection of either AAV-ARHP8 or AAV-GFP viruses. Tumor growth was monitored for another 8 weeks. As shown in Figure 3a, after castration, AAV-GFP virus-injected xenografts was arrested for 3–4 weeks, but their growth resumed later on day 42. Conversely, tumor growth of AAV-ARHP8 virus-injected xenografts was totally suppressed in both castrated and sham-operated animals. These data supported the findings observed in LAPC-4 xenografts (Fig. 2a) and also suggest that AAV-ARHP8 viruses blocked castration-resistant progression in LNCaP xenograft tumors.
To evaluate the effect of AAV-ARHP8-mediated AR silencing in vivo, xenograft tumors were harvested for examining virus distribution (GFP expression), AR protein expression (anti-AR immunostaining), cell proliferation (BrdU labeling) and apoptotic cell death (TUNEL assay). As shown in Figure 3b, a focal-to-diffuse distribution pattern of GFP expression was observed in either AAV-ARHP8 or AAV-GFP viruses-injected xenografts, indicating an uneven distribution of AAV viruses after local injection. Anti-AR immunostaining showed that AR expression decreased dramatically in AAV-ARHP8 virus-treated tumors compared to AAV-GFP virus-treated tumors. BrdU incorporation assay showed that AAV-ARHP8 virus caused a significant decrease in BrdU-positive cells compared to AAV-GFP virus, indicating that cell proliferation was severely disrupted after AR silencing induced by AAV-ARHP8 virus. In accordance with tumor growth pattern, TUNEL-positive cells were significantly higher in AAV-ARHP8 virus-treated xenografts than that in AAV-GFP virus-treated xenografts (Fig. 3c). These results are consistent to our previous findings that AR silencing leads to apoptotic cell death in AR-naive prostate cancer cells.12
We also tested the effect of AAV-ARHP8 virus on tumor growth of C4-2-derived xenografts, which are castration-resistant.29 C4-2 cells were inoculated into castrated nude mice to establish xenografts. Once xenografts were palpable (∼50 mm3), animals were randomly divided into 2 groups to receive either AAV-ARHP8 or AAV-GFP viruses via local injection. Tumor growth was monitored for another 3 weeks. As shown in Supporting Information Figure 1a, treatment with AAV-ARHP8 viruses significantly slowed down tumor growth compared to AAV-GFP viruses. Similar to that seen in LNCaP xenografts (as shown in Fig. 3), AR expression and BrdU labeling decreased dramatically in AAV-ARHP8 virus-injected xenografts compared to that in AAV-GFP virus-injected xenografts. Also, apoptotic index (TUNEL assay) was significantly higher in AAV-ARHP8 virus-injected xenografts than that in AAV-GFP virus-injected xenografts (Supporting Information Figures 1b and 1c). Taken together, these data demonstrate that AAV-ARHP8 viruses induced apoptotic cell death and resulted in tumor suppression from both androgen-responsive and castration-resistant xenograft tumors. These studies extended our previous in vitro findings12 into an in vivo setting, indicating that the no. 8 AR siRNA is of interest to be developed as a therapeutic agent for prostate cancers.
Systemically delivered AAV-ARHP8 viruses eradicate xenograft tumors in vivo
For a therapeutic method, tumor elimination is desirable once it is used in clinic. As discussed above, although tumor growth was strongly suppressed after local injection of AAV-ARHP8 viruses, tumor elimination was not achieved. A plausible explanation might be due to uneven distribution of the locally injected viruses. To achieve a better virus distribution, we went on to inject the viruses via tail vein for systemic delivery. A castration-resistant prostate cancer CWR22Rv1 cell line was used to generate xenograft tumors. 22Rv1 tumor grows very fast even in castrated animal.30 Once 22Rv1 xenograft tumors were palpable (∼30 mm3), AAV-ARHP8 or AAV-GFP viruses were injected via tail vein. Tumor growth and animal condition were monitored for another month. All animals did not show any obvious complication and no visible side effect was noticed due to intravenous virus injection. As shown in Figure 4a, AAV-GFP virus-injected tumors displayed a rapid growing pattern. In contrast, in a group of 5 animals that received AAV-ARHP8 viruses, 3 tumors disappeared within 10 days and another tumor disappeared within 2 weeks. The last tumor stayed as a steady tiny node during the experiment period. After dissection, its major part exhibited as a fibrous scar with few cancer cells (Fig. 4b). GFP expression was observed throughout the cross-section of the xenograft (Fig. 4c). These data clearly indicate that systemic delivery of AAV-ARHP8 virus achieved an even distribution of the viruses and a strong AR gene silencing, which resulted in elimination of xenograft tumors.
To rule out the possibility that virus injection itself caused the tumor elimination, we compared the effect of AAV-ARHP8 virus on 2 different xenografts side by side at the same animal. Xenografts were generated with 22Rv1 or PC-3 cells on each side of the flanks at individual animals. PC-3 cells are AR-null and do not response to AR silencing, as shown in our previous publication.12 As shown in Figure 4d, in animals receiving AAV-GFP treatment, either 22Rv1 or PC-3 xenograft tumors displayed a fast growing path as expected. In animals that received AAV-ARHP8 viruses, PC-3 xenografts displayed a similar growing pattern as seen in AAV-GFP treatment. In contrast, 22Rv1 xenografts showed a recessional pattern compared to AAV-GFP treatment. These data strongly suggest that AAV-ARHP8-induced tumor elimination is not likely due to any viral infectious effect but due to AR silencing. Notably, tumor recession in 22Rv1 xenografts in this experiment (Fig. 4d) was slower compared to the complete elimination within 10 days in the experiment shown in Figure 4a. The possible reason for this different effect was due to the different initial tumor burden of the xenografts; since 22Rv1 xenografts were growing faster than PC-3 tumors in nude mice, when PC-3 xenografts reached at least 30 mm3 in size for intravenous virus injection, 22Rv1 tumors were already about 70–100 mm3. Thus, tumor burden of the xenografts might be a factor affecting the experimental results, as discussed in a recent study using AR siRNA for prostate cancer treatment from other group.17
To further determine if AAV-ARHP8 virus-induced tumor elimination was an shRNA-induced nonspecific effect, we compared the effect of AAV-ARHP8 with a weaker AAV-ARHP virus as a control on tumor growth. To identify a less effective AR shRNA, multiple AR shRNA constructs were generated with different sequences (Table 1) and transfected into PC-3/AR cells for 3 days. As shown in Figure 5a, the ARHP4 sequence only slightly reduced AR protein expression compared to others. Therefore, ARHP4-expressing cassette was used to produce AAV-ARHP4 virus particles and served as a weaker shRNA control. As shown in Figure 5b, similarly as shown earlier, AAV-ARHP8 treatment eliminated xenograft tumors within 10 days after systemic delivery. However, AAV-ARHP4 viruses only induced a moderate tumor suppression compared to the tumor growth pattern in AAV-GFP group. These data clearly indicate that AAV-ARHP8 virus-induced tumor elimination was a sequence-specific AR gene silencing effect but not a viral infectious effect or a shRNA-mediated nonspecific phenomenon.
AAV-ARHP8 viruses reduce the expression of AR-regulated genes in vivo
Recently, we demonstrated that AR regulates the expression of 2 survival genes, SGK-1 and Bcl-xL.24, 31 Therefore, in this study, we examined the expression of these 2 genes, SGK-1 and Bcl-xL, in tumors after AAV-ARHP8 virus treatment. Paraffin sections from 22Rv1 xenograft tumors (obtained from the experiment shown in Fig. 4d) were prepared for immunostaining. Compared to the control of AAV-GFP treatment, expression of SGK-1 and Bcl-xL decreased dramatically in AAV-ARHP8 virus-treated xenografts (Fig. 6a). Consistently, the mRNA levels of these genes also dramatically decreased in xenografts receiving AAV-ARHP8 treatment compared to the AAV-GFP control, as assessed by semiquantitative RT-PCR assay (Fig. 6b). As expected, a significant decrease of the AR gene at both protein and mRNA levels (Figs. 6a and 6b) was achieved. Similar result was also seen for the well-known prostate-specific AR-regulated gene PSA (Fig. 6b), which served as a positive control. These results confirmed our previous in vitro data24, 31 that the AR modulates cellular survival via upregulation of 2 cellular survival genes, SGK-1 and Bcl-xL.
Meanwhile, we examined the apoptotic response and cell proliferation in these xenograft tumors. As revealed by TUNEL and BrdU incorporation assays (Fig. 6a), increased TUNEL-positive cells and decreased BrdU-labeled cells were observed in AAV-ARHP8-treated xenografts compared to AAV-GFP group, suggesting that systemic delivery of AAV-ARHP8 viruses induced apoptotic cell death and reduced cell proliferation in vivo.
Systemically delivered AAV-ARHP8 viruses were readily detectable in vivo
Finally, we examined the tissue distribution of AAV-ARHP8 viral particles after systemic delivery. Major organs and 22Rv1 xenograft tumors were harvested on day 3 postintravenous injection of the AAV-ARHP8 viruses. Total RNAs were extracted for RT-PCR analysis of GFP expression since the AAV vector harbors a GFP expression cassette. As shown in Supporting Information Figure 2a, different levels of GFP expression were detected from various tissues, of which xenograft tumor, liver, spleen and brain tissues showed higher expression levels than that in other tissues. Notably, prostate tissue also showed a fair amount of GFP expression, suggesting that systemic delivery of AAV-ARHP8 viral particles is feasible to target prostate tissue.
We also examined AAV virus distribution using transmission electron microscopy (TEM) approach in xenograft tumors after systemic delivery. As shown in Supporting Information Figure 2b, viral particles were readily seen inside cells, based on the morphologic features described in a previous report.32 Meanwhile, certain typical morphologic features of apoptotic response were also seen in tumors receiving AAV-ARHP8 injection, including nuclear shrinkage and condensation of nuclear compartment (Supporting Information Figure 2b, panels c and e). Conversely, a smooth looking nuclear membrane was seen in tumors receiving AAV-GFP injection (Supporting Information Figure 2b, panels a and b). These data further support the notion that AR silencing leads to apoptotic cell death in prostate cancer cells.
Currently, it is well established that the AR is the dominant factor for castration-resistant progression of prostate cancer.1, 2 Targeting the AR gene and its regulated signal pathway represents the next generation of therapy for prostate cancers.3–5 In addition to previous reports from in vitro studies,1, 12–14 recent data from mouse xenograft experiments provided further evidence that silencing the AR gene leads to tumor suppression from either androgen-responsive or castration-resistant prostate cancer cells.15–17, 33 The siRNA delivery approaches include daily intraperitoneal injection of chemically synthetic siRNA, stable transfection with an inducible shRNA system or ex vivo infection with a lentiviral AR shRNA vector. In the present report, we utilized an AAV delivery system to express the previously reported no. 8 AR shRNA, which induces apoptotic cell death.12 By using an in vivo mouse xenograft model, a surprising effect of tumor elimination was observed when AAV-ARHP8 viruses were systemically delivered in xenograft tumor-bearing mice (Fig. 4). Further analysis demonstrated that this dramatic tumor elimination effect was an AR shRNA sequence-specific effect but not due to an AAV viral infection or RNA hairpin-mediated nonspecific effect on tumor growth in nude mice (Fig. 5).
In this study, we first injected the AAV viruses directly into existing xenograft tumors. Similar to the reports from others, this local delivery approach only resulted in tumor suppression (Figs. 2 and 3, and Supporting Information Fig. 1). After careful evaluation of the tumor section, we realized that the locally injected AAV viruses were not distributed evenly inside xenograft tumors. To mimic the clinical situation of cancer treatment and to achieve more profound effect, we delivered the AAV viruses via intravenous injection. Surprisingly, almost all the xenograft tumors disappeared within 10 days after AAV-ARHP8 treatment, but the control empty AAV viruses (AAV-GFP) did not show such effect (Fig. 4). To rule out the possibility of any nonspecific effect from AAV preparation, we established 2 different xenografts in each individual animal with 2 different prostate cancer cell lines, AR-positive 22Rv1 or AR-null PC-3 cells. These 2 cell lines responded to AR silencing differently as described in our previous publication.12 Consistently, intravenous injection of AAV-ARHP8 viruses caused a dramatic recession of 22Rv1 xenografts but had no effect on the growth of PC-3 xenografts. Furthermore, we also used a less effective AAV-ARHP4 virus produced from a different sequence as a comparison for AAV-ARHP8 in eradicating 22Rv1 xenografts. As expected, AAV-ARHP4 viruses did not induce tumor recession but a moderate suppression of tumor growth after delivered systemically (Fig. 5). All together, these data strongly suggest that systemic delivery of AAV-ARHP8 viruses eradicated xenograft tumors by effectively inducing AR gene silencing in vivo. As the authors are aware, this is the first study to report a tumor elimination effect by AR silencing in prostate cancers.
Consistent with our previous findings12 and reports from others,13 xenograft tumors displayed an apoptotic response after AAV-ARHP8 treatment in addition to reduced cell proliferation. This apoptotic response might account for the dramatic tumor recession (Figs. 4 and 5), which is significant compared to tumor suppression reported by others.15–17 The difference between tumor recession (Figs. 4 and 5) and tumor suppression from others might be due to the unique sequence property of the no. 8 AR siRNA. In addition, we also noticed that tumor recession was somewhat dependent on the initial size of tumor burden in xenograft experiment, which was also discussed in a recent report from other group.17 Consistently, this notion is supported by a very recent study that in vivo knockdown of the AR with an inducible AR silencing system stably expressed in prostate cancer cells led to tumor recession of pre-established xenografts.33
Currently, recombinant AAV has been used extensively as gene delivery vehicles to transduce a wide range of cells in vitro and in vivo.20, 34, 35 Remarkable success has also been demonstrated recently in phase I/II clinical trials in AAV-based gene therapy studies.21, 22, 36 Nonetheless, potential side effects, such as insertional mutagenesis and immunogenesis, are the major limitations or concerns when applied to humans. To avoid these potential side effects in human, we are currently developing an alternative approach with nanoparticles that are designed to deliver the ARHP8 plasmid DNA specifically to prostate cancer cells.
In conclusion, in this study, we demonstrated that AAV-ARHP8 viruses induced a dramatic AR gene silencing, which resulted in apoptotic cell death and tumor elimination in vivo. This effect was not due to AAV viral infection or shRNA nonspecific response from the host animal but due to an shRNA sequence-specific AR silencing. Also, silencing the AR gene caused a dramatic decrease of cellular survival genes, SGK-1 and Bcl-xL, which might play a role in tumor recession. These data were supported by several reports from other groups,13–17, 33 indicating that targeting the AR is a promising approach for prostate cancer intervention, although the clinical efficiency of the ARHP8 shRNA in human prostate cancers needs to be determined in the future.
We thank Dr. Dongsheng Duan, Department of Molecular Microbiology and Immunology, University of Missouri, for his critical discussion in the use of AAV vector for gene delivery. We also thank Barbara Fegley for her excellent assistance in electronic microscopy and KUMC LAR facility for animal maintenance. This study was supported by grants from Kansas Mason's Foundation, Department of Defense PCRP and SWOG HOPE Foundation to Dr. Benyi Li. The project was also partially supported by grants from Zhejiang Provincial Natural Science Foundation of China to Dr Aijing Sun, who was an award recipient of KUMC Biomedical Research Training Program.