Reovirus mutant jin-3 exhibits lytic and immune-stimulatory effects in preclinical human prostate cancer models

Treatment of castration-resistant prostate cancer remains a challenging clinical problem. Despite the promising effects of immunotherapy in other solid cancers, prostate cancer has remained largely unresponsive. Oncolytic viruses represent a promising therapeutic avenue, as oncolytic virus treatment combines tumour cell lysis with activation of the immune system and mounting of effective anti-tumour responses. Mammalian Orthoreoviruses are non-pathogenic human viruses with a preference of lytic replication in human tumour cells. In this study, we evaluated the oncolytic efficacy of the bioselected oncolytic reovirus mutant jin-3 in multiple human prostate cancer models. The jin-3 reovirus displayed efficient infection, replication, and anti-cancer responses in 2D and 3D prostate cancer models, as well as in ex vivo cultured human tumour slices. In addition, the jin-3 reovirus markedly reduced the viability and growth of human cancer cell lines and patient-derived xenografts. The infection induced the expression of mediators of immunogenic cell death, interferon-stimulated genes, and inflammatory cytokines. Taken together, our data demonstrate that the reovirus mutant jin-3 displays tumour tropism, and induces potent oncolytic and immunomodulatory responses in human prostate cancer models. Therefore, jin-3 reovirus represents an attractive candidate for further development as oncolytic agent for treatment of patients with aggressive localised or advanced prostate cancer.


INTRODUCTION
Prostate carcinoma is the second most common cancer and the fifth leading cause of cancer-related death in men worldwide [1]. The development of therapy resistance and incurable metastatic disease represents major clinical problems.
Immunotherapy has emerged as a viable and attractive strategy for the treatment of different solid cancers [2]. Despite the success of immunotherapeutic approaches in various cancers, prostate cancer has remained largely unresponsive for single-agent immune therapies, including cancer vaccines and immune checkpoint inhibitors [3]. Accumulating evidence suggests that prostate cancer cells escape from immune surveillance by creating an immunesuppressive and immune-exclusive tumour microenvironment [4]. This immunosuppressive barrier impairs the generation and maintenance of a clinically desired anti-tumour immune response. Treatment modalities that overcome this immunosuppressive state could represent a promising option for prostate cancer.
Oncolytic viruses specifically infect, replicate, and lyse malignant tumour cells, while minimising harm to normal cells. Moreover, oncolytic viruses have the ability to promote adaptive and innate immune responses upon infection and killing of cancer cells, e.g., mediated by the release of danger-associated molecular patterns (DAMPs) like high mobility group box 1 (HMGB1) [5][6][7][8]. Previously, we have demonstrated that the oncolytic potency of mammalian Orthoreoviruses can be enhanced by natural selection and genetic modification [9,10]. Reoviruses are double-stranded RNA viruses and have not been associated with severe disease in humans [11]. Wild-type reovirus type 3 Dearing (T3D) has oncolytic properties in a variety of tumour types. It binds to a cancer cell by interaction of viral spike protein Sigma-1 to sialic acids and to junction adhesion molecule A (JAM-A) [10,12,13]. However, JAM-A expression is often reduced in solid cancers and this correlates with a poor survival and a worse prognosis [14]. Mutant reoviruses with enhanced tumour tropism, that can also infect cancer cells independently of JAM-A (i.e., via negatively charged sialic acids) represent a promising treatment modality [10]. Our group has generated spontaneous reovirus mutants (i.e., jin-1, −2, and −3) with extended tropism. These reovirus mutants are able to infect a wide range of cell lines that normally resist wild-type reovirus T3D infection [10]. In the absence of JAM-A, these reovirus mutants depend on negatively charged sialic acids on the cell surface for infection.
In this study, the direct oncolytic and indirect immunomodulatory effects of jin-3 reovirus were determined in state-of-the-art preclinical prostate cancer models, including monolayer and three-dimensional cell cultures, ex vivo cultured human prostate cancer tissue slices, and cell line-and patient-derived prostate cancer xenograft models in vivo [15].

MATERIAL AND METHODS Virus production
Wild-type T3D reovirus strain R124 was plaque purified from the wild-type reovirus T3D (ATCC, Manassas, VA, United States) on HER911 cells [10,16]. Reovirus mutant jin-3 was isolated from JAM-A-deficient U118MG cells after passaging of the wildtype T3D strain R124 [10]. Both R124 and jin-3 reoviruses were propagated, purified, and titrated on human HER911 cells as described [10]. Cell lines were propagated for no >6 months or 30 passages after resuscitation from stocks. All cell lines were frequently tested for Mycoplasma infection, using a Mycoplasma-specific polymerase chain reaction (PCR).
Two-and three-dimensional prostate cancer cultures Human prostate cancer cell lines PC-3M-Pro4luc2, DU145, and 22Rv1 were cultured in monolayers (Table S1). Threedimensional cultures were generated from a previously established three-dimensional prostate cancer model from bone metastasis material (MSK-PCa1), or generated from newly established patient-derived xenograft (PDX) models from prostate cancer bone and liver metastases biopsies (NM78 and NM72) [17,18].
Three-dimensional cultures of prostate cancer bone and liver metastases were maintained, as previously described [17,18].

Viability assays
For cell lines, 1500 cells were seeded per well in a 96-well plate. After 24 h, cells were exposed to oncolytic reovirus at a multiplicity of infection (MOI) of 0.01-0.1-1-10 and 100 plaque forming units (p.f.u.)/cell. After 24 h, the medium was refreshed. After 6 days, the viability of the cells was assessed by performing MTS assays [19]. Three-dimensional prostate cancer cultures were treated with oncolytic reovirus for 3, 7, and 10 days. Changes in viability were assessed by the Cell Titre Glo assay, according to the manufacturer's protocol (Promega, Madison, WI, United States).

FACS analyses
Flow cytometry was performed with LSRII (BD Biosciences, Franklin Lakes, NJ, United States) and analysed with FCS express software after staining the cells with JAM-A antibody. (Table S2) Sigma-3 immunocytochemistry A total of 20,000 prostate cancer cells were seeded in eight-well chamber slides (ThermoFisher Scientific, Waltham, MA, United States). After 24 h, the cells were exposed to oncolytic reovirus After 1, 2, and 3 days of post exposure, Sigma-3 was visualised by immunofluorescence (Table S2) and confocal microscopy.
Generation of prostate cancer patient-derived xenograft models and ex vivo tumour tissue slice culture Prostate cancer tissue was obtained via either transurethral resection of the prostate (prostatectomy) or needle biopsies after informed consent (Pronet p05.85 and RBUT-ID-PROSTAAT-151; Table S3).
In order to establish new prostate cancer PDX models, tumour pieces were implanted subcutaneously in adult male immunodeficient mice. All animal experiments were performed after approval by the Animal Welfare Committee of the Leiden University Medical Center in accordance with the Dutch Act on Animal experimentation and EU Directive 2010/63/EU (project licences from Central Authority for Scientific Procedures on Animals (CCD): AVD1160020173725 and AVD1160020187004). Established and well-characterised prostate cancer PDX models were propagated as described (project licence AVD101002017867) [20] (Table  S4). Tumour growth was monitored by calliper measurements. All mice were housed under sterile conditions in accordance with Dutch guidelines.
Prostate cancer tissue was sliced and cultured, as previously described [19]. Slices were exposed to 10 8 p.f.u./ml jin-3 reovirus. Three days post exposure, the tissues were fixed with 4% PFA and processed for histology.

Real-time quantitative polymerase chain reaction (RT-qPCR)
Cells were seeded in six-well plates and exposed to oncolytic reovirus for 6, 24, and 48 h. Total RNA was isolated according to the manufacturer's protocol (Nucleospin RNA kit Macherey-Nagel, Düren, Germany). cDNA was generated by using random primers (Promega, Madison, WI, United States) and RT-qPCR was performed with GoTaq Mastermix (Promega, Madison, WI, United States), according to the manufacturer's protocol in technical duplicates and biological triplicates (Promega, Madison, WI, United States). Gene expression was normalised to GAPDH expression. The sequences of the PCR primers used for the quantitation of cellular transcripts, viral RNA, and the detection of mycoplasma contamination can be found in Table S5.
High mobility group box 1 release Cells were exposed to oncolytic reovirus with a MOI of 10. At 48 h post exposure, conditioned medium was collected. HMBG1 release was measured by performing an ELISA, according to the manufacturer's protocol (IBL International, Hamburg, Germany).

Administration of jin-3 reovirus in vivo
Subcutaneous PC-3M-Pro4luc2 tumours were generated in male NSG mice (AVD1160020173725 and AVD1160020187004) [19]. When tumours reached a volume of 0.11 cm 3 , intra-tumoural administration with 10 8 p.f.u. jin-3 reovirus in 10 μl PBS was initiated twice a week (n = 6 per group). For sample size calculations and in/exclusion criteria, see Supplementary Information. Tumour growth was monitored by bioluminescence imaging [21]. Tumour pieces of PDX model PCa-15.01 were implanted subcutaneously in the flank of male NSG mice (AVD1160020173725 and AVD1160020187004). When tumours reached a volume of 0.11 cm 3 , mice were randomised based on body weight, tumour burden, and/or tumour size. Subsequently, tumours were treated by intra-tumoural injections of 10 8 p.f.u. jin-3 reovirus (n = 10 per group). Tumour growth was assessed by calliper measurements.
All in vitro experiments were repeated at least twice. Data are represented as mean ± standard error of the mean (SEM). Twoway ANOVA and two-sided t tests were performed on data from in vivo experiments. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Reovirus infection and replication in ex vivo cultured prostate cancer tissue slices
Prostate cancer tissue slices were generated from explanted PC-3M-Pro4luc2 tumours and infected with jin-3 reovirus for 2, 3, 4, and 7 days. Exposure to jin-3 reovirus resulted in a timedependent increase in Sigma-3 score indicating viral infection and replication (Fig. S1 and Fig. 3A, B).
Next, tumour tissue slices were generated from novel and previously established PDX models [20,22,23] followed by exposure to jin-3 reovirus for 3 days (Fig. 4A). For the duration of the ex vivo tissue culture experiment, Sigma-3 staining was observed in 91% (10/11) of the PDX models in the jin-3 reovirustreated group (Fig. 4A and Fig. S2). JAM-A protein expression was observed in all PDX tumours, but the subcellular localisation of JAM-A protein varies per patient-derived tumour (Fig. S3A). Reovirus infection was observed in both AR-positive and ARnegative PDX models (Fig. S3B). In ex vivo cultured tissue slices from patient biopsy material, Sigma-3 staining was observed in all prostate cancer biopsies after jin-3 reovirus exposure for 3 days (Fig. 4B and Fig. S4).

Reovirus mutant jin-3 induces oncolysis in prostate cancer xenografts in vivo
Intra-tumoural administration of reovirus mutant jin-3 reovirus in subcutaneously growing tumours from PC-3M-Pro4luc2 significantly decreased tumour burden (p = 0.0367) and diminished tumour volume (p = 0.06; Fig. 5A-D). Viral capsid protein RNA expression in tumour cells was significantly upregulated in jin-3 reovirus-treated prostate tumours (S4 p < 0.0001; Fig. 5E). In line with the tumour regression data, histological analyses revealed that jin-3 reovirus-treated tumours were depleted of proliferating cancer cells, their tissue architecture was lost and histological tumour markers significantly decreased. Conversely, high levels of viral Sigma-3 protein were detected in jin-3 reovirus-treated tumours (Fig. 5F). Similar data were obtained with our novel prostate cancer PDX model PCa-15.01, in which intra-tumoural injection of jin-3 reovirus significantly reduced tumour volume (p = 0.0072) and tumour weight (p = 0.0373; Fig. 5G, H). Likewise, jin-3 reovirus-treated tumours displayed a strong induction of viral RNA S4 transcripts (p < 0.0001; Fig. 5I). Histological evaluation  A Effect of jin-3 reovirus administration on total tumour burden was measured by whole-body bioluminescent reporter imaging (BLI) of firefly-luciferase2 expressing PC-3M-Pro4luc2 cells (n = 6 per group). B Whole-body bioluminescent optical imaging (BLI) at the start and end of the experiment. C Change of tumour burden (BLI) was significantly reduced in jin-3-treated tumours (p < 0.05). D Tumour volume (calliper measurements) was reduced after jin-3 administration. E S4 RNA expression, indicative of viral Sigma-3 gene expression, was observed in PC-3M-Pro4luc2 tumours treated with jin-3 reovirus at day 21 (p < 0.0001). F Histological evaluation depicted a strong oncolytic response, the presence of viral proteins (Sigma-3), a reduction in tumour cell proliferation (PCNA), and a loss of cytokeratins in PC-3M-Pro4luc2 tumours treated with jin-3 reovirus. G Treatment with jin-3 reovirus significantly reduced tumour volume in the PCa-15.01 PDX model (n = 10 per group; p < 0.01). H Significant tumour shrinkage upon intra-tumoural Jin 3 administration (tumour weight; p < 0.05). I In tumours treated with jin-3 reovirus, viral S4 RNA was detected (p < 0.0001). J Histological evaluation of jin-3 reovirus-mediated anti-tumour effects indicated a loss of tissue architecture, the presence of viral proteins (Sigma-3), an induction of apoptosis (cleaved caspase-3), a reduction in proliferation (PCNA), and a loss of tumour-associated cytokeratins (PANKRT). Magnification is 63×, scale bar = 25 μm. Error bars indicated ± SEM, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, two-way ANOVA and t tests.  6 Immune modulatory response of R124 parental and jin-3 mutant reovirus in prostate cancer cultures. A The role of the STING pathway and IFN signalling in recognising viral RNA and induction of anti-viral and anti-tumour immune responses. B Induction of IFNβ mRNA in prostate cancer cells after treatment with R124 or jin-3 reovirus for 48 h. C Induction of interferon-stimulated genes (ISGs) in human prostate cancer cells after treatment with MOI10 of R124 or jin-3 reovirus for 48 h (log-transformed). White boxes indicated that the gene was not expressed. D Induction of inflammatory cytokines gene expression of CXCL10, TNFα, and IL-1β and cytosolic RNA sensor RIG-I after treatment of human prostate cancer cells with reovirus. E HMGB1 protein release (danger-associated molecular pattern) by human prostate cancer cells after 48 h treatment with oncolytic reovirus. *p < 0.05, **p < 0.01, ***p < 0.001,****p < 0.0001, $ p < 0.05, $$ p < 0.01, $$$ p < 0.001, $$ p < 0.0001. Mean ± standard error of the mean (SEM), N = 2. Two-way ANOVA.
indicated a complete loss of tumour tissue architecture after jin-3 reovirus treatment (Fig. 5J).

jin-3 reovirus induces determinants of immune modulation in human prostate cancer cells
Type I interferons and interferon-stimulated genes (ISGs) are of crucial importance in oncolytic virotherapy, as expression of ISGs associated with the sensitivity to oncolytic virotherapy [24,25] (Fig. 6A). Exposure to jin-3 reovirus resulted in the significant upregulation of IFNβ gene expression in human prostate cancer cells after 48 h (Fig. 6B). Compared to R124, exposure to jin-3 reovirus resulted in a significant stronger induction in IFNβ gene expression ( $ p < 0.05; Fig. 6B). Moreover, exposure of the cancer cells to jin-3 reovirus resulted in a stronger induction of multiple ISGs compared to R124 (Fig. 6C). Furthermore, treatment with jin-3 reovirus induced a significant, dose-dependent induction of the inflammatory cytokines CXCL10, TNFα, and IL-1β. In addition, exposure to jin-3 induced a significant stronger expression of CXCL10 and IL-1β compared to R124 ( $$$ p < 0.001 and $$$$ p < 0.0001). Expression of DDX58 that encodes for the cytosolic RNA sensor RIG-I, was also found to be significantly upregulated upon jin-3 reovirus administration (Fig. 6D). In addition, shedding of HMGB1 protein, a well-established marker for immunogenic cell death, was observed after exposure of prostate cancer cells to jin-3 reovirus [26] (Fig. 6E). Moreover, additional upstream targets of the anti-viral response were significantly affected by jin-3 reovirus exposure, including IFNAR1 and IFNAR2 (Fig. S5).

DISCUSSION
Current clinical and preclinical evidence demonstrates that antitumour immune responses can eliminate existing malignant cells, resulting in the protection against tumour recurrence [27,28]. However, the majority of prostate tumours does not adequately respond to immune therapy, which is mediated-among others-by the immunosuppressive tumour microenvironment, the low mutation burden resulting in limited expression of neoantigens, and immune evasion and exclusion processes in prostate tumours [4,8,29].
Oncolytic viruses are increasingly being exploited to counteract these immunosuppressive mechanisms by selectively killing cancer cells and, more importantly perhaps, by the initiation of anti-tumour immunity. Oncolytic viruses either have a natural preference to enter, replicate in, and/or kill cancer cells as opposed to normal cells or they are engineered to do so [30]. To achieve improved anti-tumour immunity in prostate cancer, the identification and clinical development of optimised oncolytic viruses, alone or combined with other treatment modalities, may be advantageous.
Here we report, for the first time, the use of jin-3 reovirus in human prostate cancer models. Reovirus mutant jin-3 reovirus induces potent direct anti-tumour effects in human prostate cancer. In ex vivo cultured human prostate cancer tissue slices and 2D/3D cell cultures, we found that jin-3 reovirus is able to infect and replicate in the cancer cells. Moreover, jin-3 reovirus exposure significantly induced the expression of ISGs and inflammatory cytokines. This is further substantiated by the release of immunogenic cell death marker HMGB1 after exposure of cancer cells to jin-3 reovirus.
A number of preclinical studies have reported the anti-tumour effects of wild-type mammalian Orthoreovirus [31,32]. In addition, clinical trials have revealed that wild-type reovirus has inherent oncolytic properties, is not associated with serious human disease, and has a favourable safely profile in cancer patients [11]. However, clinical trials have demonstrated that the clinical benefit of wild-type reovirus administration in prostate cancer patients is limited and so far phase II studies have revealed no survival benefit [11,12,[32][33][34][35][36].
We previously described the isolation of reoviruses with an expanded tropism [10]. These so-called jin mutant reoviruses can also infect cancer cells independently of the canonical reovirus entry receptor JAM-A [10]. The potential oncolytic and immunomodulatory properties of these mutant reoviruses were not previously established in preclinical models of human prostate cancer. Here, we show that the reovirus mutant jin-3 displays strong oncolytic and immunomodulatory properties and represents an interesting candidate oncolytic virus for the treatment of prostate cancers. We observed efficacious jin-3 infection in the majority of ex vivo cultured prostate cancer tissue slices derived from PDXs and primary prostate cancer patient material. Moreover, jin-3 induced significant tumour regression in various human prostate cancer xenografts in vivo as indicated by real-time optical imaging measurements (firefly luciferase2 expressing cancer cells) to determine the total tumour burden, changes in tumour volume, tumour weight, and reduction of pan-cytokeratin. In prolonged in vivo experiments and in a clinical setting, it will also be worthwhile to assess the effect of jin-3 Reovirus exposure on additional tumour markers in circulation, e.g., serum PSA. In this study, we observed that jin-3 reovirus administration induces several ISGs and inflammatory cytokines. The jin-3 reovirusinduced gene expression of key immunomodulators (activators of the adaptive and innate immune system) is further substantiated by the release of HMGB1 as one of the key DAMPs.
In conclusion, we describe the evaluation as an oncolytic agent of the mutant reovirus jin-3 that displays strong oncolytic and immunomodulatory properties in state-of-the-art preclinical prostate cancer models, including patient-derived tumour slice models. Our findings support the notion that jin-3 reovirus can be potentially exploited in strategic treatment combinations to with otherwise less efficacious immuno-oncological or chemotherapeutical approaches to achieve improved and durable antitumour responses.