An Oncolytic Virus Expressing a T-Cell Engager Simultaneously Targets Cancer and Immunosuppressive Stromal Cells Joshua D

Published OnlineFirst November 18, 2018; DOI: 10.1158/0008-5472.CAN-18-1750

Cancer Translational Science Research

An Oncolytic Virus Expressing a T-cell Engager Simultaneously Targets Cancer and Immunosuppressive Stromal Cells Joshua D. Freedman1, Margaret R. Duffy1, Janet Lei-Rossmann1, Alice Muntzer2, Eleanor M. Scott1, Joachim Hagel1, Leticia Campo1, Richard J. Bryant3, Clare Verrill3,4, Adam Lambert3, Paul Miller5, Brian R. Champion2, Leonard W. Seymour1, and Kerry D. Fisher1

Abstract

Effective immunotherapy of stromal-rich tumors requires of antigen presentation, T-cell function, and trafficking. simultaneous targeting of cancer cells and immunosuppres- M2-like ascites macrophages exhibited a proinflammatory sive elements of the microenvironment. Here, we modified repolarization, indicating spectrum-wide alteration of the the oncolytic group B adenovirus enadenotucirev to express tumor microenvironment. With this approach, we have a stroma-targeted bispecific T-cell engager (BiTE). This BiTE actively killed both cancer cells and tumor fibroblasts, bound fibroblast activation protein on cancer-associated reversing CAF-mediated immunosuppression and yielding fibroblasts (CAF) and CD3e on T cells, leading to potent a potent single-agent therapeutic that is ready for clinical T-cell activation and fibroblast death. Treatment of fresh assessment. clinical biopsies, including malignant ascites and solid prostate cancer tissue, with FAP-BiTE–encoding virus Significance: An engineered oncolytic adenovirus that þ induced activation of tumor-infiltrating PD1 T cells to kill encodes a bispecific antibody combines direct virolysis with CAFs. In ascites, this led to depletion of CAF-associated endogenous T-cell activation to attack stromal fibroblasts, immunosuppressive factors, upregulation of proinflamma- providing a multimodal treatment strategy within a single tory cytokines, and increased gene expression of markers therapeutic agent. Cancer Res; 78(24); 6852–65. 2018 AACR.

þ Introduction CAF-produced CXCL12 can block entry of CD8 cells into the tumor and attract regulatory T cells, inhibiting effector T-cell Cancer-associated fibroblasts (CAF) facilitate invasion (1), proliferation (11, 12). coordinate angiogenesis (2), and maintain an immunosuppres- CAFs are pivotal to tumor immunology, making it difficult to sive microenvironment in solid carcinomas (3). Their immuno- envisage cancer immunotherapy achieving its full potential with- modulatory functions include production of indoleamine out addressing their deleterious effects. CAF depletion can reverse 2,3-dioxygenase (IDO) and regulatory cytokines such as VEGF, local immune suppression and improve tumor immunotherapy. FGF, IL10, and TGFb (4–8). Notably secreted TGFb can accumu- Genetic-based CAF depletion in an autochthonous pancreatic late in the stromal matrix, exerting a powerful immunosuppres- cancer model uncovered the ability of anti-PD-L1 to inhibit tumor sive effect on newly infiltrating na€ve immune cells (9, 10), while growth and improve survival (13). While such an approach has a strong therapeutic rationale, implementation can be difficult due to the lack of unique target antigens on the CAF surface, with most fi 1Department of Oncology, University of Oxford, Oxford, United Kingdom. of their known surface markers also present on normal broblasts. 2PsiOxus Therapeutics Ltd., Abingdon, United Kingdom. 3Nuffield Department One promising target antigen is fibroblast activation protein of Surgical Sciences, University of Oxford, Oxford, United Kingdom. 4Oxford (FAP), which is upregulated on CAFs across a broad range of solid NIHR Biomedical Research Centre, University of Oxford, Oxford, United malignancies (14) but also found on normal fibroblasts in con- 5 Kingdom. Churchill Hospital, Oxford University Hospital NHS Trust, Oxford, nective tissue in the muscle, gall bladder, bladder, and bone United Kingdom. marrow stromal cells (BMSC; ref. 15). Elimination of FAP-positive Note: Supplementary data for this article are available at Cancer Research cells with mAbs or FAP-targeted CAR-T cells demonstrated the Online (http://cancerres.aacrjournals.org/). potential to reverse tumor-associated immune suppression, par- Current address for J. Hagel: Experimental Medicine Division, Nuffield Depart- ticularly when combined with immunotherapeutic strategies such ment of Medicine, University of Oxford, Oxford, United Kingdom. as cancer vaccines (16, 17). However, FAP expression on extra- Corresponding Author: Leonard W. Seymour, University of Oxford, Old Road tumoral cells is concerning, with previous FAP-targeting preclin- Campus Research Building, Old Road Campus, Headington, Oxford, OX3 7DQ, ical studies showing extensive bone marrow toxicity and cachexia United Kingdom. Phone: 4418-6561-7040; Fax: 4418-6561-7028; that would caution against clinical development of systemic FAP- E-mail: [email protected] targeted treatments (15, 18). doi: 10.1158/0008-5472.CAN-18-1750 Bispecific T-cell engagers (BiTE) show powerful targeted killing 2018 American Association for Cancer Research. of cancer cells, but can also be deployed against stromal targets

6852 Cancer Res; 78(24) December 15, 2018

Oncolytic Adenovirus Expressing FAP BiTE

such as CAFs. BiTEs crosslink T cells (via CD3e) to antigen-positive promoter–driven expression vector (pSF-CMV-Amp; Oxford target cells, independent of HLA presentation, and can activate Genetics) by standard cloning techniques. Recombinant BiTE any T-cell to engage with and destroy adjacent target cells (19). protein was produced by transfecting HEK293A cells with poly- Moreover, BiTE-mediated T-cell activation can overcome ele- ethylenimine [PEI, linear, MW 25000, Polysciences; DNA:PEI ments of tumor-associated immunosuppression that limit phys- ratio of 1:2 (w/w)]. Cells maintained in serum-free DMEM. iologic immune responses, leading to reactivation and prolifer- Supernatants were harvested, concentrated 50-fold using ation of exhausted tumor-specific T cells (20–22). BiTEs targeted 10,000 MWCO Amicon Ultra-15 Filter Units (Millipore), and to a CAF marker such as FAP could be a potent strategy to activate stored at 80C. BiTE protein concentration was determined by intratumoral T cells to attack and deplete CAFs. However, systemic dot blot using decahistidine-tagged cathepsin D (BioLegend) as a delivery of the BiTE would likely mediate significant toxicity by standard. Specific binding of the FAP BiTE to recombinant FAP activating circulating T cells to attack normal fibroblasts and protein was confirmed by ELISA (data not shown). BMSCs. Accordingly, this potentially powerful approach is frus- trated by challenges of site-specific delivery. Generation of BiTE-expressing enadenotucirev With their ability to encode and specifically express biologics in Modified EnAd were produced by direct insertion of the BiTE disseminated tumors, oncolytic viruses (OV) are an ideal solu- cassette into the parental EnAd cloning plasmid pEnAd2.4 tion. One promising candidate is enadenotucirev (EnAd), which using Gibson assembly (30, 31). Additional viruses with FAP has demonstrated good blood stability and systemic bioavail- BiTE expression linked to red fluorescent protein (RFP) via a ability in several early-phase clinical trials (23–25). An encoded P2A site were also generated. Plasmid DNA was linearized by FAP-specific BiTE would be produced and secreted only upon restriction digest with AscI (New England Biolabs) and trans- virus infection of tumor cells, allowing it to access tumor-infil- fected into HEK293A cells for virus production in DMEM (2% trating lymphocytes (TIL). This approach has been validated FBS). Upon extensive plaque formation, cells were harvested, using BiTEs to target T-cell cytotoxicity to tumor cell antigens and virus released by three freeze–thaw cycles. Single clones (21, 26–28). However, a virus-encoded BiTE that activates T cells were selected by serial dilution and amplified by serial infec- to kill tumor stromal fibroblasts would provide a "multimodal" tion, followed by double CsCl banding to produce concentrat- therapeutic agent that simultaneously targets two distinct cell ed virus stocks. Stocks were titered by the Quant-iT Picogreen types within the tumor. Alongside direct OV-mediated cytolysis of dsDNA assay (Thermo Fisher Scientific) and infectious dose tumor cells, which is often proinflammatory (29), secretion of determined by serial titration on A549 cells. FAP-specific BiTEs should activate TILs to attack and deplete CAFs, acting to reverse CAF-induced immune suppression. This Processing and culture of human PBMCs and clinical biopsy approach combines direct cytotoxicity, immune stimulation, and samples reversal of local immunosuppression, thereby transforming an PBMCs were isolated from leukocyte cones (NHS Blood and þ immunologically inactive "cold" tumor into one that is "hot," that Transplant, UK) by density-gradient centrifugation. CD3 cells is, with greater immune infiltration, yielding an integrated and were extracted by depleting non-CD3 cells using the Pan T-cell þ þ more effective immunotherapeutic response. Isolation Kit (Miltenyi Biotec). For CD4 and CD8 cells, þ CD4 Microbeads were used (Miltenyi Biotec). Primary human malignant ascites samples and human prostate tissue samples Materials and Methods were obtained from the Churchill Hospital (Oxford University Cell lines Hospitals NHS Foundation Trust) following written informed DLD, SKOV3, A549, HEK293A (ATCC), and normal human patient consent and approval by the institutional review board dermal fibroblast (NHDF; Lonza) cells were cultured in DMEM and research ethics committee of the Oxford Centre for (Sigma-Aldrich). Chinese Hamster Ovary (CHO, ATCC), normal Histopathology Research (Reference 09/H0606/5þ5) in accor- human bronchial epithelial (NHBE) cells (Lonza) were cultured dance to the UK Human Tissue Act 2004 and the Declaration of in RPMI1640 (Sigma-Aldrich). All cells were authenticated by Helsinki. For ascites, samples were immediately processed with short tandem repeat profiling (CRUK Cambridge Institute, United cells and fluid separated by centrifugation (300 g), with the Kingdom) and routinely tested each month for Mycoplasma cellular fractions treated with red blood cell lysis buffer (MycoAlert Mycoplasma Detection Kit, Lonza). Cell lines were (Qiagen). For ex vivo T-cell activation and cytotoxicity, cells passaged no more than ten passages after thawing before use in were used immediately, or adherent cells were expanded by experiments. Growth medium was supplemented with 10% (v/v) serial passage. For human prostate tissue specimens, tissue was FBS (Thermo Fisher Scientific). Cells were incubated at 37 C and transported in RPMI and stored on ice until slicing within two 5% CO2. A FAP-expressing stable CHO cell line was generated hours of surgery. Tissue cores were embedded in UltraPure low using the FAP gene sequence (ID: 1149, NCBI) as described melting-point agarose (4% w/v, Thermo Fisher Scientific), and previously (21). 300-mm tissue slices were prepared using a vibratome (Leica VT 1200S, Leica Microsystems). Each ex vivo tissue slice was trans- BiTE engineering and production ferredtoa0.6cm2 PTFE insert (Millipore) in 24-well plates A FAP-targeted BiTE was produced by joining the DNA containing 1 mL of cultivation media for prostate tissue (Sup- encoding two single-chain antibody fragments (scFvs) recog- plementary Material). After overnight culture, the media were nizing human FAP and CD3e with a sequence encoding a replaced, and tissue slices were treated with BiTE or recombi- flexible glycine-serine (GS) linker. An N-terminal immuno- nant virus. On day 0, 4, and 7 postinfection, 30% of the globulin signal sequence for mammalian secretion and supernatant was collected, frozen, and replaced. On day 7, C-terminal decahistidine tag for detection were added. slices were fixed in paraformaldehyde (4%) and embedded in DNA sequences were synthesized and inserted into a CMV paraffinforIHC.

www.aacrjournals.org Cancer Res; 78(24) December 15, 2018 6853

Freedman et al.

In vitro and ex vivo coculture experiments each case. All antibodies were acquired from BioLegend unless For in vitro coculture studies, PBMCs were seeded with the stated otherwise. Analysis was performed on a FACSCalibur flow appropriate target cells (E:T ratio, 5:1) in flat-bottom 96-well cytometer (BD Biosciences) and data processed with FlowJo plates in 100-mL medium. Target cell lines were prepared with cell v10.0.7r2 software (TreeStar Inc.). dissociation buffer to preserve cell surface antigens. For ex vivo experiments, unpurified total cells from bone marrow or ascites IHC were seeded in culture medium or fluid from the same exudate Automated staining was carried out with the Leica BOND- sample, respectively. To assess T-cell activation by virus-infected MAX autostainer (Leica Microsystems). Antigen retrieval was cells or BiTE-containing supernatants (300 ng/mL), cocultures performed at 100C using Epitope Retrieval Solution 2 (Leica were treated with 100-mL supernatant or infected with 100 vp/cell Biosystems), followed by incubation with antibodies for CD8 in 100-mL medium. Where appropriate, CD3/CD28 Dynabeads (Agilent Technologies), CD25 (Atlas Antibodies), EpCAM (Thermo Fisher Scientific) were included as positive controls for (BioLegend), FAP (R&D Systems) or adenoviral hexon (Milli- T-cell activation. T cells were harvested by pooling the culture pore). Detection was performed using the BOND Polymer media and a subsequent PBS wash. If adherent cells are also Refine Detection System (Leica Biosystems). Alternatively, required, cell dissociation buffer was used to detach from plate for the FAP primary antibody only, anti-sheep HRP-DAB surface and cells were pooled with nonadherent cells. Staining Kit (R&D Systems) was used. Sections were incubated with hematoxylin and imaged (Aperio CS2 slice scanner, Characterization of human T-cell activation Leica Microsystems). T-cell activation was measured by staining for surface expression of activation markers (CD69, CD25) and analyzed by flow Quantitative PCR cytometry. To study T-cell proliferation, T cells were labeled with Adenovirus genomes were measured by qPCR using primers 5 mmol/L carboxyfluorescein succinimidyl ester (CFSE) dye and probe against hexon (primers: 50-TACATGCACATCG- (Thermo Fisher Scientific) prior to culturing with target cells. CCGGA-30/50-C GGGCGAACTGCACCA-30, probe: 50-FAM- After five days, T cells were harvested and analyzed by CCGGACTCAGGTACTCCGAAGCATCCT-TAMRA-30). At the spe- flow cytometry. As a surrogate for proliferation in mixed cell cific timepoint, total cell and supernatants were harvested, and populations (e.g., whole ascites samples), total T-cell number per DNA extracted (PureLink Genomic DNA Mini Kit; Thermo Fisher well was determined using precision counting beads (BioLegend). Scientific). In brief, primers and probe were mixed with To measure T-cell degranulation, the externalization of CD107a DNA samples and added to QPCRBIO Probe Mix Hi-Rox (PCR was assessed by adding a CD107a antibody directly to the well at Biosystems) Master Mix. To measure levels of FAP mRNA in the start of the experiment. After 1-hour incubation, GolgiStop ascites, reverse transcription qPCR was performed. The total cell (6 mg/mL, BD Biosciences) was added, followed by flow cytometry fraction was harvested after 72 hours of treatment, RNA was analysis after an additional five hours. IL2 and IFNg quantities extracted (RNAqueous-Micro Total RNA Isolation Kit; Thermo were measured using the Human IL-2 ELISA MAX kit (BioLegend) Fisher Scientific), and cDNA prepared (Superscript III First-Strand or Human IFNg ELISA MAX Kit (BioLegend). A flow cytometric Synthesis SuperMix; Thermo Fisher Scientific). FAP expression multiplex bead immunoassay was performed using LEGENDplex was quantified using FAP-specific primers (50-TCAGTGTGAG- Th Kit (BioLegend). TG CTCTCATTGTAT-30/50-GCTGTGCTTGCCTTATTGGT-30) and 2xqPCRBIO SyGreen Blue Mix Hi-ROX Master Mix (PCR Bio- Target cell cytotoxicity assay systems). Expression of the 18S gene was also measured as a To assess target cell cytotoxicity by free BiTE or virus, release of normalization control (50-GCCCGAAGCGTTTACTTTGA-30/50- LDH into the supernatant (CytoTox 96 Non-Radioactive Cyto- TCCATTAT TCCTAGCTGCGGTATC-30). All qPCR was run on ABI toxicity Assay; Promega) or MTS viability assay (CellTiter 96 Cell PRISM 7000 (Applied Biosystems). Proliferation Assay, Promega) were used. To determine viability of specific cell types, total cells were harvested by cell dissociation Gene expression analysis buffer, and residual number of viable target cells measured by Gene expression analysis was performed using the nCounter flow cytometry using an amine-reactive fluorescence live–dead PanCancer Immune Profiling Panel (NanoString Technologies). stain. For observation of cell viability in real-time, xCELLigence The nSolver Advanced Analysis module was used for data analysis, technology (Acea Biosciences) was used. TGFb and VEGF quan- in accordance with NanoString guidelines. Background thresh- tities were measured using TGF beta-1 Human/Mouse ELISA Kit olding was performed, followed by normalization of the data via (Thermo Fisher Scientific) and LEGENDplex Growth Factor Kit the mean of the internal NanoString positive controls, and dif- (BioLegend), respectively. ferential expression determined, with reference to uninfected cells. A gene set's directed global significance score for a covariate Flow cytometry measures the cumulative evidence for the up- or downregulation To classify different cellular populations, antibodies specific for of genes in a pathway and is calculated as the square root of the CD11b (ICRF44), EpCAM (9C4), FAP (427819, R&D Systems, mean squared t-statistic of genes, with t-statistics generated from USA), CD3 (HIT3a), CD4 (OKT4), CD8a (HIT8) were used. To the linear regression algorithm within the nSolver Advanced analyze T-cell populations, the following antigens were used: Analysis module. CD69 (FN50), CD25 (BC96), IFNg (4S.B3), CD107a (H4A3), PD1 (H4A3). To analyze macrophage populations, cells were Microscopy treated with Fc receptor block (Miltenyi Biotec) and stained with Brightfield and fluorescence images were obtained on a Zeiss CD163 (GHI/61), CD206 (15-2), CD64 (10.1), and CD86 Axiovert 25 microscope and captured with a Nikon DS5M camera. (IT2.2). The appropriate isotype control antibody was used in For time-lapse sequences, images were obtained on a Nikon TE

6854 Cancer Res; 78(24) December 15, 2018 Cancer Research

Oncolytic Adenovirus Expressing FAP BiTE

2000-E Eclipse inverted microscope and captured with a Hama- Generation of BiTE-armed EnAd matsu Orca-ER C4742-95 using MetaMorph imaging software. EnAd is a conditionally replicating chimeric group B adenovi- Images were collected at 15-minute intervals with videos rus generated by bioselection (Fig. 2A; ref. 32). The FAP-BiTE and (12 frames/second) generated using ImageJ software (NIH, control-BiTE sequences were inserted downstream of the fiber Bethesda, MD). Where appropriate, cells were stained with Cell- gene under transcriptional control of either an exogenous CMV Tracker Orange CMTMR Dye (Thermo Fisher Scientific) and promoter or a splice acceptor (SA) site for the adenoviral major CellTrace Violet Cell Proliferation Kit (Thermo Fisher Scientific). late promoter (MLP). The former drives immediate transgene Apoptosis was visualized using CellEvent Caspase 3/7 Detection expression upon successful cell infection, whereas MLP-driven Reagent (Thermo Fisher Scientific). expression occurs only in cells permissive to virus replication, such as human tumor cells. Viruses were rescued and purified Statistical analysis from HEK293A cells (Supplementary Table S1). Where experiments produced two datasets, significance was Colorectal adenocarcinoma (DLD) cells were infected at evaluated using a Student two-tailed t test. In all cases of more 100 vp/cell of the parental or recombinant viruses to assess than two experimental conditions, statistical analysis was replication kinetics. Viral genome copies reached between 3–6 performed using a one-way ANOVA test with Tukey post hoc 1012 genomes/mL (Fig. 2B), indicating that BiTE expression did analysis or two-way ANOVA test using Bonferroni post hoc not impair replication relative to the parental virus. Cytotoxicity analysis. All data are presented as mean SD. Significance of all recombinant viruses was also comparable with parental levels used were P ¼ 0.01–0.05 (), 0.001–0.01 (), 0.0001– EnAd (Fig. 2C). Therefore, modification of the viral genome to 0.001 (). Experiments were performed in biological incorporate the BiTE transgene had little effect on viral replication triplicate, unless stated otherwise. or oncolytic activity. FAP-BiTE secretion by virus-infected HEK293A cells was demonstrated by immunoblotting of the supernatant (Supplementary Results Fig. S1B). Functionality of these secreted virus-encoded BiTEs was Fibroblast-targeted BiTE engineering assessed by adding supernatants to cocultures of PBMC-derived þ A FAP-targeted BiTE was engineered recognizing human FAP CD3 T cells and either CHO or CHO-FAP cells. T cells cocultured and CD3e.ABiTEspecificforCD3e and an irrelevant antigen, with CHO-FAP cells showed strong CD25 induction and target filamentous hemagglutinin adhesin (FHA) of Bordetella pertus- cell lysis when incubated with supernatants from EnAd-CMV- sis, was used to control for unspecific binding. An N-terminal FAP-BiTE– or EnAd-SA-FAP-BiTE–infected cells (Fig. 2D and E). immunoglobulin signal sequence and C-terminal decahisti- No activation or cytotoxicity was observed with supernatants dine tag were added for mammalian secretion and detection. from cells infected with unmodified or control-BiTE–expressing BiTE protein production was assessed by transfection of EnAd, in the presence of parental CHO cells, or in the absence of T- HEK293A cells. cells. FAP-BiTE yield from DLD cells infected with modified To confirm specificity of the FAP-BiTE for surface FAP, we viruses was measured by comparing T-cell–mediated NHDF cyto- established a human FAP-positive stable cell line using FAP- toxicity induced by 72-hour–infected DLD supernatants to a negative CHO cells. Peripheral blood mononuclear cell standard curve. Following a 24-hour cytotoxicity assay, we mea- (PBMC)-derived T cells were activated 24 hours after cocultur- sured FAP-BiTE at 9.8 and 49.2 mg/106 cells after 72 hours for ing with CHO-FAP cells and FAP BiTE–containing supernatants EnAdCMV-FAP-BiTE and EnAdSA-FAP-BiTE, respectively. This from transfected HEK293A (Fig. 1A) and mediated CHO-FAP was consistent with previous reports, suggesting that while tran- cell lysis (Fig. 1B). Neither T-cell activation nor lysis were scriptional initiation is delayed, there is superior total transgene observed in cultures with parental CHO cells or control-BiTE, expression when driven by the endogenous MLP compared with indicating that surface FAP expression is required for T-cell the CMV promoter (30). The FAP-BiTE showed impressive poten- activation, presumably via surface CD3 clustering and pseu- cy, with cytotoxicity detectable in supernatants diluted 10,000- doimmunologic synapse formation. fold (Supplementary Fig. S1C). FAP-BiTE–induced T-cell activation was also evaluated in coculture with NHDFs, which express surface FAP when cultured in EnAd-FAP-BiTE–mediated oncolysis induces T-cell–mediated high serum (10% FBS, Supplementary Fig. S1A). Incubation of fibroblasts killing NHDF- and PBMC-derived T cells from six donors with FAP-BiTE- EnAd kills carcinoma cell lines by direct oncolysis (33), but containing supernatants for 24 hours induced significant T-cell does not effectively replicate in, or directly kill, fibroblasts or other activation and NHDF lysis, with the control-BiTE having no effect nonepithelial stromal cells (23). However, FAP-targeted BiTE (Fig. 1C; EC50, 2.5 ng/mL). CFSE-labeled PBMC T cells cocultured from infected tumor cells should allow T-cell activation and with NHDF and FAP-BiTE underwent multiple rounds of T-cell mediate targeted killing of FAP-expressing stromal fibroblasts. proliferation (Fig. 1D) and showed at least 10-fold increase in Cocultures of fibroblasts, moderately permissive SKOV3 ovarian IFNg, IL2, TNFa, IL17F, IL22, and IL10 (Fig. 1E), with IFNg carcinoma cells killed by EnAd 5–7 days postinfection (acting production 10-fold higher than that induced by physiologic as BiTE producers), and PBMC-derived T cells were measured in anti-CD3/CD28 activation beads (Fig. 1F). Interestingly, FAP- real-time by cell index, a unitless measure of cell viability (Fig. 2F). BiTE induced activation and degranulation of CD4 and CD8 T In the absence of T cells, tumor cells and fibroblasts cells persisted cells, directing both subsets to kill NHDF cells with similar for 100–120 hours, independent of virus infection. In the potency (Fig. 1G–I). Importantly, no induction of activation presence of T cells, FAP-BiTE expression from infected SKOV3 markers, proliferation and cytokines was observed with con- cells led to complete NHDF cytotoxicity, with lysis observed trol-BiTE or in the absence of NHDF target cells, confirming that within 22 hours postinfection (hpi) by EnAd-CMV-FAP-BiTE and CD3 clustering is essential for T-cell activation. 42 hpi for EnAd-SA-FAP-BiTE. Crucially, no cytotoxicity was

www.aacrjournals.org Cancer Res; 78(24) December 15, 2018 6855

Freedman et al.

Figure 1. FAP-BiTE–containing supernatants activate primary human T cells and target cytotoxicity to FAPþ cells. A, Activation (CD69þCD25þ) of primary human CD3 cells cultured for 24 hours with BiTE-containing supernatants and CHO or CHO-FAP cells. B, LDH release by target cells in A. C, CD25 expression on T cells (black) and NHDF lysis (LDH release, red) was assessed after 24 hours of coculture with BiTE- and PBMC-derived T cells from six healthy donors. D, Representative proliferation of CFSE-stained T cells after 5-day coculture with BiTE and NHDFs. E, Cytokine production was evaluated by a multiple human Th cytokine panel after coculturing T cells with BiTE and NHDFs for 48 hours. F, Secreted IFNg from cocultures of T cells with NHDF and BiTE-containing supernatants. Induction of CD25 (G) or degranulation (H) of CD4 and CD8 T cells following 24 hours coculture with BiTE and NHDFs. I, LDH release by NHDF cells after 24 hours in cocultures with BiTE and either CD4- or CD8-purified T cells. Data show mean SD of biological triplicates (A-C and E-I). Significance was assessed using one-way ANOVA with Tukey post hoc analysis compared with "untreated" (A, C, E,andF) or "control-BiTE" (B) or using an unpaired two-tailed t test (G–I; , P < 0.05; , P < 0.01; , P < 0.001; ns, nonsignificant).

6856 Cancer Res; 78(24) December 15, 2018 Cancer Research

Oncolytic Adenovirus Expressing FAP BiTE

Figure 2. EnAd-expressing FAP-BiTE induces T-cell–dependent cytotoxicity of stromal fibroblasts. A, Genome of the BiTE-armed oncolytic adenovirus, EnAd. ITR, inverted terminal repeat; P, promoter; pA, polyadenylation site. B, Genome replication of parental EnAd- or BiTE-expressing viruses in DLD cells infected with 100 vp/cell. C, Viability of DLD cells infected with EnAd or recombinant virus (MTS; 5 days postinfection). D and E, T-cell activation (D; CD25þ) and target cell cytotoxicity (E; LDH release) in cocultures of T cells, target cells, and infected HEK293A supernatants 24 hpi. F, Viability of NHDF and SKOV3 (4:1) cells monitored by xCELLigence in the absence or presence of CD3-purified PBMCs (5:1 effector:target). Mean (solid line) SD (dotted line) of biological triplicates. G, T-cell activation (CD25þ) in virus-infected cocultures of NHDF and SKOV3 cells. H, Representative images showing coculture of NHDF (red; stained with CellTracker Orange CMTMR), DLD cells (unstained), and T cells (blue; stained with CellTrace Violet), followed by infection with EnAd or BiTE-armed virus. Apoptosis was visualized using CellEvent Caspase 3/7 Detection Reagent (green). White arrow, dead fibroblasts. Scale bar, 100 mm. Data show mean SD of biological triplicates (D-G). Significance assessed using one-way ANOVA with Tukey post hoc analysis compared with "EnAd" (B and C) or "uninfected" (D–G; , P < 0.05; , P < 0.001).

www.aacrjournals.org Cancer Res; 78(24) December 15, 2018 6857

Freedman et al.

observed in cultures infected with EnAd or EnAd expressing the showed complete elimination of both cell types upon treatment control-BiTE. with EnAd-CMV-FAP-BiTE or EnAd-SA-FAP-BiTE 72 hpi (Supple- NHDF lysis was confirmed by lactate dehydrogenase (LDH) mentary Fig. S1G). release in similar coculture experiments (Supplementary Fig. S1D). The kinetics of T-cell activation paralleled that of NHDF EnAd-SA-FAP-BiTE–mediated T-cell activation and target cell cytotoxicity (Fig. 2G; Supplementary Fig. S1E). Importantly, lysis is tumor selective FAP-BiTE–encodedvirusesfailedtoinduceCD25expression Conventional FAP-targeted therapeutics given intravenously þ in the absence of NHDF, further demonstrating the require- arereportedtoinduceFAP cell toxicity within the bone þ ment of FAP cells for BiTE-mediated T-cell activation (Sup- marrow compartment (18). Coupling BiTE expression to virus plementary Fig. S1F). replication via the viral MLP restricts expression to the tumor þ BiTE-induced cytotoxicity of stromal fibroblasts by T cells was compartment, minimizing unwanted toxicity to FAP fibro- observed by time-lapse microscopy using cocultures of T cells, blasts in normal physiologic sites. To compare selectivity of fibroblasts, and DLD cells, which are more susceptible than virally encoded CMV- and MLP-driven BiTE expression, NHDFs SKOV3 cells to EnAd-mediated lysis (Fig. 2H; Supplementary were incubated with EnAd, EnAd-CMV-FAP-BiTE or EnAd-SA- Movie S1–S3). While infection with EnAd induced dramatic FAP-BiTE in the presence of primary T cells only. At 72 hpi with DLD killing within 48 hours, NHDFs remained viable through- EnAd-CMV-FAP-BiTE, we observed cytotoxicity in 80% of out. In contrast, EnAd-CMV-FAP-BiTE infection induced both NHDF cells (Fig. 3A). No lysis was observed in EnAd-SA- direct DLD killing and T-cell–mediated fibroblast cytotoxicity. FAP-BiTE–infected cells, consistent with the inability of EnAd Quantification of DLD and NHDF cells in parallel cultures to complete its life cycle in nonepithelial tumor cells (23).

Figure 3. EnAd-SA-FAP-BiTE does not induce T-cell activation or FAPþ cell lysis in the absence of tumor cells. A, Cytotoxicity in T-cell and NHDF cocultures was assessed by LDH release 72 hpi with BiTE-expressing viruses. B and C, T-cell activation in coculture with NHDF and NHBE or SKOV3 seeded at 5:1 and infected with BiTE-armed viruses. T cells were added 2 hpi. Activation (B, CD69þ; C, CD25þ) was assessed 72 hpi. D, LDH release in cocultures from B and C. T-cell activation (CD25þ; E) and LDH release (F) in cocultures of fresh bone marrow samples 5 days postinfection with FAP-BiTE–expressing EnAd. Prior to virus infection, bone marrow cells were plated with NHDF with or without SKOV3 (50:5:1). LDH release was calibrated against lysis of corresponding amounts of NHDF and SKOV cells without bone marrow cells, corrected for the amount released by healthy bone marrow cells. Data show mean SD of biological triplicates. Significance between more than two conditions was assessed using one-way ANOVA with Tukey post hoc analysis compared with "uninfected" (A). Significance between two conditions was assessed using an unpaired two-tailed t test (B–F; , P < 0.05; , P < 0.001; ns, nonsignificant).

6858 Cancer Res; 78(24) December 15, 2018 Cancer Research

Oncolytic Adenovirus Expressing FAP BiTE

To better simulate the multiple cell types present in a healthy cell number during treatment (Fig. 4G). Free FAP-BiTE and EnAd- þ tissue, NHDFs were cultured with exogenous PBMC-derived T CMV-FAP-BiTE induced significant depletion of FAP fibroblasts cells and either NHBE cells or SKOV3 and subsequently infected in all samples, typically to levels below 1% of those in untreated or with EnAd-SA-FAP-BiTE or EnAd-SA-control-BiTE. While EnAd- control samples, consistent with marked falls in FAP mRNA, VEGF SA-FAP-BiTE infection of SKOV3 allowed T-cell activation and secretion, and elimination of cells with a fibroblast-like morphol- target cell lysis (Fig. 3B–D), NHBE cells did not. ogy (Supplementary Fig. S3A–S3C). A similar trend was observed Finally, for maximum clinical relevance, EnAd-SA-FAP-BiTE upon infection with EnAd-SA-FAP-BiTE, although one patient activity was evaluated in three fresh human whole bone marrow (patient biopsy 1) demonstrated neither T-cell activation nor samples from healthy donors. Despite literature reports of bone fibroblast depletion (Fig. 4E and G). Infection of this sample þ marrow toxicity of FAP-targeted antibodies, no FAP cells were with EnAd-SA-GFP also showed no GFP-positive cells (Supple- þ þ detected in any samples. Accordingly, FAP NHDF 'target' cells mentary Fig. S3D; Supplementary Table S2) and no EpCAM were added prior to infection to determine whether our armed tumor cells at the outset (Supplementary Fig. S3E). The sample þ viruses triggered uncontrolled toxicity against FAP cells. Neither likely had insufficient tumor cells to support virus replication, EnAd-SA-FAP-BiTE nor EnAd-SA-control-BiTE induced endoge- demonstrating the strict necessity of tumor cells for virus repli- nous T-cell activation (Fig. 3E) or targeted cytotoxicity (Fig. 3F) in cation and MLP-drive BiTE expression, suggesting one predictor the absence of tumor cells. The addition of SKOV3 cells led to T- for potency. A cytokine array of patient biopsy 1 demonstrated cell activation and cytotoxicity following EnAd-SA-FAP-BiTE that EnAd-CMV-FAP-BiTE infection induced at least 10-fold infection, with the latter thought to reflect predominantly BiTE- increases in IL17A, IL17F, IL22, IFNg, and IL10 expression (Fig. þ mediated T-cell lysis of FAP NHDFs. These data confirm that 4H). Parallel experiments using expanded mixed cultures of MLP-driven FAP-BiTE production is restricted to tumor cells and ascites-derived fibroblasts and tumor cells showed that fibroblast þ suggest there should be no systemic toxicity against FAP cells depletion led to 50%–70% lower TGFb levels in supernatants þ within normal bone marrow. (Fig. 4I), suggesting FAP cells to be a major source of immunosuppressive TGFb within tumor ascites. Altogether, these data EnAd expressing FAP-BiTE activates tumor-associated T cells to show that treatment of malignant ascites with free or virally kill endogenous fibroblasts within patient-derived malignant encoded FAP-BiTE is able to polyclonally activate anergic T-cells, ascites leading to targeted depletion of autologous tumor-associated Malignant peritoneal ascites are frequent in several advanced fibroblasts. carcinoma types, including ovarian, pancreatic, breast, and lung cancers (34), and often associated with a poor prognosis (34, 35). Global changes in immune pathway gene expression were The fluid is routinely drained from some patients as a palliative observed in EnAd-SA-FAP-BiTE–treated samples treatment, providing a convenient and informative liquid biopsy. To assess the impact of CAF depletion, T-cell activation and There is mounting evidence that malignant ascites are sites of virolysis of cancer cells on the tumor microenvironment, we used substantial immunosuppression (36). We therefore assessed the NanoString to determine changes in the expression of 730 cancer effects of cell-free ascites fluid on PBMC-derived T-cell activation and immune pathway genes in three primary ascites samples, using anti-CD3/CD28 beads and FAP-BiTE. Bead-mediated T-cell selected to represent the spectrum of clinical possibilities. Biopsy þ þ activation was significantly inhibited by 3 of 5 ascites samples 4 had a high ratio of FAP cells relative to EpCAM cells (likely (Fig. 4A). In contrast, FAP-BiTE-mediated T-cell activation was not epithelial cancer cells; Supplementary Table S3), while biopsy 5 þ þ suppressed by any ascites fluids compared with levels observed in had a similar proportion of EpCAM and FAP cells, and biopsy þ þ normal serum (Fig. 4B). 6, a relatively low ratio of FAP to EpCAM cells. All samples had þ þ Human ascites biopsy samples typically contain tumor cells, similar levels of CD3 T cells and CD11b myeloid cells. fibroblasts, lymphocytes, and macrophages, representing a Significant T-cell activation was observed in all samples three unique tumor-like model system to assess endogenous tumor- days postinfection with EnAd-SA-FAP-BiTE, but not EnAd-SA- associated T-cell activation. Figure 4C shows the cellular compo- control-BiTE (Supplementary Fig. S4A). Approximately 40% of þ þ sition of a representative sample containing CD3 , EpCAM , all genes showing significant changes in mRNA levels of at least þ þ CD11b , and FAP cells (see also Supplementary Fig. S2A and two-fold (Supplementary Fig. S4B); only mRNA basally expressed þ S2B; Supplementary Table S2). Ascites-associated CD3 T cells above a minimum threshold level were included in the analysis. were, on average, 63% (up to 92.5%) positive for the exhaustion Considerably more genes showed changes following exposure to marker PD1 compared with only 10%–20% of PBMC-derived T EnAd-SA-FAP-BiTE than EnAd-SA-control-BiTE, except biopsy 6, þ cells (Fig. 4D). We assessed the ability of BiTE-encoding EnAd to where the high number of EpCAM cancer cells may have resulted infect ascites cancer cells and secrete sufficient amounts of BiTE, in extensive EnAd-specific gene changes due to BiTE-independent leading to endogenous T-cell activation and killing of autologous direct virolysis. Changes were grouped by immune response cancer-associated fibroblasts within an ascites sample. Total asci- category as an average of three samples (Fig. 5A) or individual tes cells from four patient biopsies were incubated in 50% ascites samples (Supplementary Fig. S4C). Although individual biopsies fluid with free FAP-BiTE or EnAd encoding FAP-BiTE. After five showed some variation, all EnAd-SA-FAP-BiTE–infected samples days, endogenous T cells were strongly activated in all ascites demonstrated increased gene expression in numerous immune þ biopsy samples (30%–80% of total CD3 cells; Fig. 4E), com- groupings including cytotoxicity, pathogen defense, and T-cell, B- þ bined with CD3 T-cell proliferation (Fig. 4F). Parental and cell, and NK-cell function. While T-cell- and NK-cell–attractant control-BiTE viruses did not induce T-cell activation or chemokines (CXCL9, CXCL10, CXCL11) were also upregulated in proliferation. all biopsies, strong decreases in fibroblast-associated CXCL6, Ascites T-cell activation and cytotoxicity toward endogenous CXCL12, and CXCL14 induced a downregulation in overall che- þ þ FAP fibroblasts was assessed by measuring the change in FAP mokine expression for biopsy 4 (Supplementary Fig. S4D).

www.aacrjournals.org Cancer Res; 78(24) December 15, 2018 6859

Freedman et al.

Figure 5B shows the dataset for biopsy 4, which had the highest decreases (up to 1,000-fold) were in ﬁbroblast-associated genes, þ number of FAP cells. The most highly upregulated genes (up to such as COL3A1, FN1, THY1, CXCL12, and IL13RA2. Similar 100-fold) following treatment with EnAd-SA-FAP-BiTE included trends were seen across all three samples, with the most modest T-cell markers (GZMB, IFNG, IL2RA, PRF1, TNF). The greatest changes in ﬁbroblast markers observed in biopsy 6, which had the

Figure 4. EnAd-expressing FAP-BiTE activates endogenous T cells and induces FAPþ cell lysis in malignant exudate samples. A and B, Activation (CD69/CD25þ)of PBMC-derived T cells cocultured with anti-CD3/CD28 Dynabeads (A) or NHDF and BiTEs (B) for 24 hours in normal serum or ascites fluid (50%). C, Representative flow cytometry plot demonstrating the cellular composition for a typical ascites biopsy sample. D, PD-1 expression by T cells as a percentage of total CD3þ cells. E, CD25 expression on endogenous T cells five days posttreatment with free BiTEs or BiTE-expressing EnAd. Total unpurified cells were treated in 50% exudate fluid from the same biopsy sample. F and G, Relative quantity of CD3þ cells (F) and residual FAPþ cells (G). Effector:target ratios were 79.4 (sample 1, black), 2.27 (sample 2, blue), 31.6 (sample 3, red), and 2.44 (sample 4, gray). H, Cytokine production was evaluated by a multiple human Th cytokine panel. I, TGFb in supernatants harvested from ascites cells incubated with PBMC T cells and BiTE-expressing virus. Data show mean SD of biological triplicates (A, B, and E-I). Significance was assessed using one-way ANOVA with Tukey post hoc analysis compared with "normal serum" (A and B) or "untreated" (E–I). D, Data show mean SD. Significance assessed using an unpaired two-tailed t test (, P < 0.05; , P < 0.01; , P < 0.001).

6860 Cancer Res; 78(24) December 15, 2018 Cancer Research

Oncolytic Adenovirus Expressing FAP BiTE

Figure 5. Global response of malignant ascites to infection with EnAd-SA-FAP-BiTE. A, Heatmap showing changes in mRNA counts (as global significance score, average of three samples) within defined gene sets compared with untreated. Left, EnAd-SA-control-BiTE; right, EnAd-CMV-FAP-BiTE. B, Fold change plot showing gene-specific differences in mRNA counts following infection of biopsy 4 with EnAd-SA-FAP-BiTE. Five T-cell activation (blue) and fibroblast (gold) markers are shown. COL3A1, alpha1 (III) collagen; THY1, Thy-1; IL13RA2, IL13Ra2; FN1, fibronectin; GZMB, granzyme B; PRF1, perforin; IL2RA, CD25; IFNG, IFNg. C–E, Fold change in mRNA counts (left, EnAd-SA-control-BiTE; right, EnAd-SA-FAP-BiTE; compared with untreated) in three ascites samples for fibroblast-specific genes (C) or genes involved in T-cell function (D) or antigen presentation and T-cell trafficking (E). Circle, biopsy 4; triangle, 5; square, 6. F and G, Expression levels of CD163, CD206 (F), and CD64 and CD86 (G) on CD11bþCD64þ cells. gMFI, geometric mean fluorescence intensity. C–E, Data show mean of biological duplicates. Significant changes were assessed using a multivariate linear regression algorithm with three patient biopsies. Significance of changes in gene expression induced by each virus versus uninfected is displayed adjacent the x-axis, and between EnAd-SA-control-BiTE or EnAd-SA-FAP-BiTE displayed above the plot. F and G, Data show mean SD of biological triplicates. Significance was assessed using one-way ANOVA with Tukey post hoc analysis compared with "untreated" (, P < 0.05; , P < 0.01; , P < 0.001).

www.aacrjournals.org Cancer Res; 78(24) December 15, 2018 6861

Freedman et al.

þ lowest levels of FAP cells. Changes in expression of key genes untreated slices nor those infected with EnAd-SA-control-BiTE- compared with untreated samples are shown by individual biopsy RFP showed detectable T-cell activation, although some samples in Fig. 5C–F. For example, expression of the fibroblast marker demonstrated modest increases in IFNg and IL2 following EnAd- collagen type III (COL3A1) is dramatically reduced upon infec- SA-control-BiTE-RFP infection, likely a direct result of virolysis. In tion with EnAd-FAP-BiTE compared with EnAd-control-BiTE in benign prostate tissue, there was little increase in IFNg or IL2 with all three samples, while THY1 and IL13RA2 (also used as fibro- any treatment (Supplementary Fig. S5D and S5E). blast markers; Fig. 5C) showed FAP-BiTE–dependent decreases in BiTE-mediated activation of T cells is expected to lead to two biopsies. Basal expression of these genes in biopsy 6 did not fibroblast killing. Cells were visualized undergoing apoptosis in pass the minimum threshold for analysis. T-cell activation mar- real-time within ex vivo prostate tissues (Fig. 6F). EnAd-SA-FAP- kers (GZMB, PRF1, and IL2RA; Fig. 5D), checkpoint markers BiTE-RFP–infected cells strongly associated with apoptotic nuclei, (PDCD1, CTLA4, and LAG3; Supplementary Fig. S4E), T-cell– suggesting BiTE-mediated induction of proximal cytotoxicity of recruiting chemokine CXCL9, and DC maturation/antigen pre- surrounding cells. FAP-BiTE–mediated cytotoxicity was observed sentation markers LAMP3 and TAP1 (Fig. 5E) all increased in a in all patient biopsy samples, with intrinsic EnAd activity also FAP-BiTE–dependent manner. These latter findings are particu- inducing small increases, potentially due to a greater number of larly encouraging, raising the possibility of immunosuppression cancer cells in some samples (Fig. 6G). Indeed, regions of high reversal in the tumor microenvironment following EnAd-SA-FAP- T-cell activation showed an absence or degradation of surround- BiTE infection. ing tissue or stroma, with tissue architecture remaining intact in uninfected samples (Fig. 6H). Crucially, FAPlow benign prostate Treatment of ascites samples with EnAd-SA-FAP-BiTE induces tissue showed negligible increases in cytotoxicity within the repolarization of resident tumor-associated macrophages duration of the study (Supplementary Fig. S5F). Macrophages are known to show plasticity between proinflammatory "M1" and wound-healing "M2" phenotypes, with tumor- Discussion associated macrophages (TAM) usually skewed toward "M2." To investigate the influence of FAP-BiTE on macrophage polariza- Here, we have developed an armed oncolytic adenovirus com- tion, patient-derived malignant ascites samples were treated with bining three distinct therapeutic strategies: direct virus-mediated free or EnAd-encoded FAP-BiTE to determine activation of endog- cytotoxicity toward cancer cells, creation of a proinflammatory þ þ enous T cells and CD11b CD64 cells. Treatment with free FAP- immune environment, and removal of a key stromal cell mediator BiTE or EnAd-SA-FAP-BiTE induced strong T-cell activation and of tumor immunosuppression. Encoding BiTEs within OVs IFNg secretion (Supplementary Fig. S4F). We observed simulta- exploits the strengths of both virotherapy and immunotherapy neous induction of an activated M1-like macrophage phenotype, while overcoming limitations of each agent alone. When manifested by strong decreases in CD206 and CD163 (Fig. 5F) expressed locally, the short plasma half-lives of BiTEs will become and increased CD64 expression (Fig. 5G). Infection with EnAd- advantageous, minimizing systemic exposure and avoiding "on- SA-FAP-BiTE induced a large increase in CD86 expression, while target, off-tumor" toxicities (38). Conversely, arming an OV with a free FAP-BiTE, EnAd-SA-control-BiTE, or IFNg alone had no effect BiTE provides an additional mechanism of cell killing and broad- (Fig. 5G). ens the range of target cells to include OV-resistant stromal cells. Combining OVs and BiTEs also has a potentially synergistic þ EnAd-SA-FAP-BiTE activates TILs and induces BiTE-mediated effect on infiltrating T-cells. Increased CD8 T-cell infiltration into cytotoxicity in solid prostate tumor biopsies the tumor bed has been observed in several OV clinical trials, We obtained seven fresh paired punch biopsies of malignant including studies using EnAd and FDA-approved Imlygic (24, 39), prostate tissue from patients undergoing radical prostatectomy, cut likely providing more effector cells for BiTE-mediated cytotoxic- into thin sections for ex vivo cultures. Prostate tissue slices showed a ity. Simultaneously, BiTE-mediated redirection of TILs (poten- characteristically complex architecture, with glandular structures of tially virus-specific) toward chosen targets may delay viral clear- þ (malignant) EpCAM epithelial cells interspersed with large ance and increase intratumoral spread. regions of intervening stroma containing scattered CD8 T cells For maximal translational relevance, our current and future (Supplementary Fig. S5A). FAP expression was generally weak in studies will focus on primary human tumor biopsies maintained benign prostate tissue and high in malignant prostate tissue (Fig. ex vivo, rather than compromise with imperfect murine models 6A). Fibroblasts showing the strongest FAP expression were often that may not provide the desired tumor heterogeneity or realistic adjacent to malignant epithelial cells (Supplementary Fig. S5A). levels of immune suppression (23, 40). Clinical samples retain To facilitate assessment of virus activity, we developed reporter the heterogeneous and multifaceted cellular interactions of viruses linking FAP-BiTE and RFP expression (Supplementary Fig. advanced human cancer, and, in the case of organotypic prostate S5B). Following infection with EnAd-SA-FAP-BiTE-RFP, malig- tumor slices, the stromal architecture and extracellular matrix of a nant tissue slices showed RFP expression, demonstrating success- solid tumor. Treatment of both solid and liquid tumor biopsies ful viral infection, replication and BiTE expression (Supplemen- with EnAd-FAP-BiTE led to tumor-associated T-cell activation and þ tary Fig. S5C). Positive staining for viral hexon confirmed EnAd destruction of endogenous FAP fibroblasts, alongside secretion replication, apparently limited to malignant epithelial cells (Fig. of large quantities of proinflammatory cytokines and chemo- 6B). Malignant prostate tissue infected with EnAd-SA-FAP-BiTE kines, including IFNg, IL2, TNFa, IL17, and CXCL9. Crucially, showed an increase in activated endogenous TILs seven days this demonstrated that the patient's own tumor-associated T cells postinfection (Fig. 6C). Slices from all patients showed significant can be used for therapeutic purposes in the realistic environment induction of IFNg production, with IL2 levels also increasing in of an advanced human tumor. It was particularly encouraging þ four samples (Fig. 6D and E); both cytokines are associated with that T cells within all tested patient biopsies, shown to be PD1 activated CD4 Th1 and CD8 cytotoxic T cells (37). Neither and likely anergic, were readily activated and rendered functional

6862 Cancer Res; 78(24) December 15, 2018 Cancer Research

Oncolytic Adenovirus Expressing FAP BiTE

Figure 6. EnAd-expressing FAP-BiTE activates tumor-infiltrating T cells and mediates cytotoxicity in malignant prostate slice cultures. A, FAP expression patterns in benign and malignant prostate tissue slices. B, Expression of viral hexon protein in prostate tissue slice after EnAd infection. C, Representative image showing tumor-infiltrating T-cell activation (CD25þ) in prostate tissue slices 7 days postinfection. IFNg (D) and IL2 (E) levels in malignant prostate tissue slices infected with BiTE-expressing EnAd. F, Active caspase-3/7 (CellEvent Caspase 3/7 Green Detection Reagent, green) in malignant prostate tissue infected with EnAd-SA-FAP-BiTE-RFP (red). White arrows, dual-positive cells. Scale bar, 100 mm. G, LDH release of malignant prostate tissue slices from 5 patients infected with recombinant EnAd. H, Histologic architecture and T-cell activation (CD25þ) in EnAd-infected prostate slices. Data show mean SD of technical triplicates (D, E, and G). For D and E, n ¼ 7; for G, n ¼ 5. Significance was assessed using one-way ANOVA with Tukey post hoc analysis compared with "uninfected" (, P < 0.05; , P < 0.01; , P < 0.001).

www.aacrjournals.org Cancer Res; 78(24) December 15, 2018 6863

Freedman et al.

by the BiTE to mediate cytotoxicity. This may reflect the high agent actively kills two different cell types using two distinct, yet level of activating stimuli each T-cell can receive using a BiTE, targeted, cytotoxic mechanisms. EnAd provides a particularly where in principle every CD3 can be crosslinked to the target promising virus platform to achieve targeted BiTE expression in antigen (>100,000 copies per cell) without being limited by the disseminated tumors, exploiting the blood stability and systemic relatively small number of HLA-presented TCR-cognate peptides bioavailability of the virus, which has been studied in several early available, likely to be less than 100 per cell. Indeed, the efficacy of phase clinical trials. This strategy to induce proinflammatory cell BiTE-mediated T-cell stimulation is augmented when targeting death while reversing TME-mediated immunosuppression may a high-density receptor like FAP, a result also seen with other be what is ultimately required to turn intransigent, stromal-rich antigens (21, 22). carcinomas into targets for a complete and durable immunother- NanoString analysis confirmed the extensive depletion of apeutic response. fibroblast-associated RNA in human malignant ascites samples treated with EnAd-FAP-BiTE, together with strong induction of Disclosure of Potential Conflicts of Interest genes involved in T-cell function. Despite their varying cellular B.R. Champion is a chief scientificofficer and has ownership interest compositions, similar immune-activating trends were seen in all (including stock, patents, etc.) in PsiOxus Therapeutics. L.W. Seymour reports receiving a commercial research grant, has ownership interest (including stock, samples following EnAd-FAP-BiTE infection, with stimulation of patents, etc.), and is a consultant/advisory board member for PsiOxus Thera- fi RNAs involved in leukocyte traf cking, dendritic cell maturation, peutics. K.D. Fisher is a scientific advisor at PsiOxus Therapeutics and has and antigen presentation. Surface markers on ascites TAMs ownership interest (including stock, patents, etc.) in PsiOxus Therapeutics. revealed a clear shift from an M2-like phenotype to one that is No potential conflicts of interest were disclosed by the other authors. more proinflammatory. We expect that newly infiltrating mono- cytes, recruited by OV-mediated induction of monocyte-attractant Authors' Contributions chemokines, such as CCL2, CCL7, and RANTES, will acquire an Conception and design: J. Freedman, M.R. Duffy, J. Hagel, B.R. Champion, L.W. Seymour, K.D. Fisher M1 "activated" phenotype. Significantly, surface expression of Development of methodology: J. Freedman, J. Hagel, R.J. Bryant costimulatory ligand CD86 on TAMs was only induced by the Acquisition of data (provided animals, acquired and managed patients, combination of virus and FAP-BiTE (EnAd-FAP-BiTE). We provided facilities, etc.): J. Freedman, A. Muntzer, E.M. Scott, J. Hagel, hypothesize that virus stimulation of pathogen-associated molec- L. Campo, R.J. Bryant, P. Miller ular patterns, IFN signaling, or STING and removal of CAF- Analysis and interpretation of data (e.g., statistical analysis, biostatistics, mediated suppression are required for CD86 activation. These computational analysis): J. Freedman, J. Lei-Rossmann, J. Hagel, L.W. Seymour Writing, review, and/or revision of the manuscript: J. Freedman, findings indicate that coupling CAF depletion with potent acti- J. Lei-Rossmann, A. Muntzer, R.J. Bryant, C. Verrill, B.R. Champion, vatory stimuli (T-cell activation and viral-mediated immunogenic L.W. Seymour, K.D. Fisher cell death) synergistically repolarize the tumor microenvironment Administrative, technical, or material support (i.e., reporting or organizing toward promotion of an effective anticancer immune response data, constructing databases): A. Lambert (41). Similarly, although we expect that suppressive markers will Study supervision: J. Freedman, L.W. Seymour, K.D. Fisher increase in tandem with activation markers, this potential barrier Other (sample processing): A. Muntzer Other (provided primary tumor material): C. Verrill to continued virus activity could be counteracted by combining OVs with checkpoint inhibitors. Acknowledgments Using OVs for cancer-targeted transgene expression has now The authors gratefully acknowledge support from the Medical Research been validated both preclinically and clinically (24, 42). Here, we Council (MRC-Oxford Doctoral Training Partnership, MR/K501256/1 regulated BiTE expression using the adenoviral MLP, limiting BiTE to J.D. Freedman) and Cancer Research UK (grant #C552/A17720 to J. Lei- production to cells permissive to the virus life cycle. In the absence Rossmann, K. Fisher, L. Seymour; studentship C5255/A20936 to E.M. Scott). M. of cancer cells, we observed no BiTE production or cytotoxicity R. Duffy is funded by the Kay Kendall Leukaemia Fund (grant KKL1050). J. Lei- Rossmann is supported by Linacre College, Oxford. pEnAd2.4 was kindly (Fig. 3). This is particularly important in light of the "on-target off- provided by PsiOxus Therapeutics. We are grateful to Egon Jacobus (University tumor" toxicities observed with FAP-targeted antibodies or CAR-T of Oxford, Oxford, United Kingdom) for the use of his primers. Special thanks to þ cells toward FAP bone marrow cells. Infection of primary cultures Alison Carr and her team for their helpful collection of ascites. C. Verrill's of freshly isolated human bone marrow by EnAd-SA-FAP-BiTE research time is part-funded by the Oxford NIHR Biomedical Research Centre showed no T-cell activation or bone marrow cell toxicity in the (Molecular Diagnostics Theme/Multimodal Pathology Subtheme). We þ absence of tumor cells. Endogenous FAP cells also likely occur at acknowledge the contribution to this study made by the Oxford Centre for Histopathology Research and the Oxford Radcliffe Biobank, which are sup- frequencies too low ( 0.01%) to identify in the mononuclear cell ported by the NIHR Oxford Biomedical Research Centre. fraction. Hence, our elegant targeted expression strategy is expected to avoid such toxicities while exploiting the potent effects of the The costs of publication of this article were defrayed in part by the payment of FAP-BiTE within the microenvironment of each tumor deposit. page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. We therefore believe that arming OVs to express BiTEs targeting stromal elements, such as CAFs, can provide a powerful new Received June 8, 2018; revised September 18, 2018; accepted October 16, multimodal approach to cancer therapy. In this way, a single- 2018; published first November 18, 2018.

References 1. Itoh G, Chida S, Yanagihara K, Yashiro M, Aiba N, Tanaka M. Cancer- 3. Harper J, Sainson RCAA. Regulation of the anti-tumour immune response associated fibroblasts induce cancer cell apoptosis that regulates invasion by cancer-associated fibroblasts. Semin Cancer Biol 2014;25:69–77. mode of tumours. Oncogene 2017;36:4434–44. 4. Connell JTO, Sugimoto H, Cooke VG, MacDonald BA, Mehta AI, LeBleu VS, 2. Tang D, Gao J, Wang S, Ye N, Chong Y, Huang Y, et al. Cancer-associated et al. VEGF-A and Tenascin-C produced by S100A4þ stromal cells are fibroblasts promote angiogenesis in gastric cancer through galectin-1 important for metastatic colonization. Proc Natl Acad Sci U S A 2011; expression. Tumor Biol 2016;37:1889–99. 108:16002–7.

6864 Cancer Res; 78(24) December 15, 2018 Cancer Research

Oncolytic Adenovirus Expressing FAP BiTE

5. Henriksson ML, Edin S, Dahlin AM, Oldenborg P-AA, Oberg€ Å, Van 23. Illingworth S, Di Y, Bauzon M, Lei J, Duffy MR, Alvis S, et al. Preclinical fi Guelpen B, et al. Colorectal cancer cells activate adjacent fibroblasts safety studies of enadenotucirev, a chimeric group B human-speci c – resulting in FGF1/FGFR3 signaling and increased invasion. Am J Pathol oncolytic adenovirus. Mol Ther Oncolytics 2017;5:62 74. 2011;178:1387–94. 24. Garcia-Carbonero R, Salazar R, Duran I, Osman-Garcia I, Paz-Ares L, 6. Kojima Y, Acar A, Eaton ENN, Mellody KT, Scheel C, Ben-Porath I, et al. Bozada JM, et al. Phase 1 study of intravenous administration of the Autocrine TGF- and stromal cell-derived factor-1 (SDF-1) signaling drives chimeric adenovirus enadenotucirev in patients undergoing primary the evolution of tumor-promoting mammary stromal myofibroblasts. Proc tumor resection. J Immunother Cancer 2017;5:71. Natl Acad Sci U S A 2010;107:20009–14. 25. Calvo E, Martín MG, Cubillo A, Machiels J, Rottey S, Mardjuadi F, et al. 7. Chen JY, Li CF, Kuo CC, Tsai KK, Hou MF, Hung WC. Cancer/stroma 1064P A phase 1 study of enadenotucirev, an oncolytic Ad11/Ad3 chimeric interplay via cyclooxygenase-2 and indoleamine 2,3-dioxygenase pro- group B adenovirus, administered intravenously - analysis of dose expan- motes breast cancer progression. Breast Cancer Res 2014;16:410. sion and repeat cycle cohorts in patients with metastatic colorectal cancer 8. Soliman H, Mediavilla-Varela M, Antonia S. Indoleamine 2,3-dioxygenase: (MCRC). Ann Oncol 2014;25:iv367. is it an immune suppressor? Cancer J 2010;16:354–9. 26. Speck T, Heidbuechel JPWW, Veinalde RR, Jaeger D, von Kalle C, Ball CR, 9. Taipale J, Saharinen J, Keski-Oja J. Extracellular matrix-associated trans- et al. Targeted bite expression by an oncolytic vector augments therapeutic fi – forming growth factor-beta: role in cancer cell growth and invasion. Adv ef cacy against solid tumors. Clin Cancer Res 2018;24:2128 37. Cancer Res 1998;75:87–134. 27. Fajardo CA, Guedan S, Rojas LA, Moreno R, Arias-Badia M, De Sostoa J, 10. Flavell RA, Sanjabi S, Wrzesinski SH, Licona-Limon P. The polarization of et al. Oncolytic adenoviral delivery of an EGFR-targeting t-cell engager fi – immune cells in the tumour environment by TGFbeta 2. Nat Rev Immunol improves antitumor ef cacy. Cancer Res 2017;77:2052 63. 2010;10:554–67. 28. Yu F, Wang X, Guo ZS, Bartlett DL, Gottschalk SM, Song X-TT. T-cell fi 11. Fearon DT. The carcinoma-associated fibroblast expressing fibroblast acti- engager-armed oncolytic vaccinia virus signi cantly enhances antitumor – vation protein and escape from immune surveillance. Cancer Immunol Res therapy. Mol Ther 2014;22:102 11. 2014;2:187–93. 29. Guo ZS, Liu Z, Bartlett DL. Oncolytic immunotherapy: dying the right way 12. Costa A, Kieffer Y, Scholer-Dahirel A, Pelon F, Bourachot B, Cardon M, et al. is a key to eliciting potent antitumor immunity. Front Oncol 2014;4:74. Fibroblast heterogeneity and immunosuppressive environment in human 30. Marino N, Illingworth S, Kodialbail P, Patel A, Calderon H, Lear R, et al. breast cancer. Cancer Cell 2018;33:463–479.e10. Development of a versatile oncolytic virus platform for local intra- 13. Feig C, Jones JO, Kraman M, Wells RJB, Deonarine A, Chan DS, et al. tumoural expression of therapeutic transgenes. PLoS One 2017;12: Targeting CXCL12 from FAP-expressing carcinoma-associated fibroblasts e0177810. synergizes with anti-PD-L1 immunotherapy in pancreatic cancer. Proc Natl 31. Gibson DG, Young L, Chuang R-YY, Venter JC, Hutchison CA, Smith HO. Acad Sci U S A 2013;110:20212–7. Enzymatic assembly of DNA molecules up to several hundred kilobases. – 14. Garin-Chesa P, Old LJ, Rettig WJ. Cell surface glycoprotein of reactive Nat Methods 2009;6:343 5. stromal fibroblasts as a potential antibody target in human epithelial 32. Kuhn I, Harden P, Bauzon M, Chartier C, Nye J, Thorne S, et al. Directed cancers. Proc Natl Acad Sci U S A 1990;87:7235–9. evolution generates a novel oncolytic virus for the treatment of colon 15. Roberts EW, Deonarine A, Jones JO, Denton AE, Feig C, Lyons SK, et al. cancer. PLoS One 2008;3:e2409. Depletion of stromal cells expressing fibroblast activation protein-a from 33. Dyer A, Di Y, Calderon H, Illingworth S, Kueberuwa G, Tedcastle A, et al. skeletal muscle and bone marrow results in cachexia and anemia. J Exp Med Oncolytic group B adenovirus enadenotucirev mediates non-apoptotic cell fl 2013;210:1137–51. death with membrane disruption and release of in ammatory mediators. – 16.WangL-CS,LoA,SchollerJ,SunJ,MajumdarRS,KapoorV,etal. Mol Ther Oncolytics 2017;4:18 30. Targeting fibroblast activation protein in tumor stroma with chime- 34. Ayantunde AA, Parsons SL. Pattern and prognostic factors in patients with – ric antigen receptor T cells can inhibit tumor growth and augment malignant ascites: a retrospective study. Ann Oncol 2007;18:945 9. host immunity without severe toxicity. Cancer Immunol Res 2014;2: 35. Zamboni MM, da Silva CT, Baretta R, Cunha ET, Cardoso GP. Important 154–66. prognostic factors for survival in patients with malignant pleural effusion. 17. Ostermann E, Garin-Chesa P, Heider KH, Kalat M, Lamche H, Puri BMC Pulm Med 2015;15:29. C, et al. Effective immunoconjugate therapy in cancer models 36. Simpson-Abelson MR, Loyall JL, Lehman HK, Barnas JL, Minderman H, fl targeting a serine protease of tumor fibroblasts. Clin Cancer Res 2008; O'Loughlin KL, et al. Human ovarian tumor ascites uids rapidly and 14:4584–92. reversibly inhibit T cell receptor-induced NF-kB and NFAT signaling in 18. Tran E, Chinnasamy D, Yu Z, Morgan RA, Lee C-CR, Restifo NP, et al. tumor-associated T cells. Cancer Immun 2013;13:14. Immune targeting of fibroblast activation protein triggers recognition of 37. Romagnani S. Lymphokine production by human T cells in disease states. – multipotent bone marrow stromal cells and cachexia. J Exp Med Annu Rev Immunol 1994;12:227 57. 2013;210:1125–35. 38. Teachey DT, Rheingold SR, Maude SL, Zugmaier G, Barrett DM, Seif AE, 19. Brischwein K, Schlereth B, Guller B, Steiger C, Wolf A, Lutterbuese R, et al. et al. Cytokine release syndrome after blinatumomab treatment related to MT110: A novel bispecific single-chain antibody construct with high abnormal macrophage activation and ameliorated with cytokine-directed – efficacy in eradicating established tumors. Mol Immunol 2006;43: therapy. Blood 2013;121:5154 7. 1129–43. 39. Ribas A, Dummer R, Puzanov I, VanderWalde A, Andtbacka RHI, Michielin fi 20. Dao T, Pankov D, Scott A, Korontsvit T, Zakhaleva V, Xu Y, et al. Therapeutic O, et al. Oncolytic virotherapy promotes intratumoral T cell in ltration and – bispecific T-cell engager antibody targeting the intracellular oncoprotein improves anti-PD-1 immunotherapy. Cell 2017;170:1109 19. WT1. Nat Biotechnol 2015;33:1079–86. 40. Lei J, Jacobus EJ, Taverner WK, Fisher KD, Hemmi S, West K, et al. 21. Freedman JD, Hagel J, Scott EM, Psallidas I, Gupta A, Spiers L, et al. Expression of human CD46 and trans-complementation by murine ade- Oncolytic adenovirus expressing bispecific antibody targets Tcell cytotox- novirus 1 fails to allow productive infection by a group B oncolytic icity in cancer biopsies. EMBO Mol Med 2017;9:1067–87. adenovirus in murine cancer cells. J Immunother Cancer 2018;6:55. 22. Deisting W, Raum T, Kufer P, Baeuerle PA, MunzM.Impactofdiverse€ 41. Weerasinghe P, Buja LM. Oncosis: an important non-apoptotic mode of – immune evasion mechanisms of cancer cells on T cells engaged by cell death. Exp Mol Pathol 2012;93:302 8. EpCAM/CD3-bispecific antibody construct AMG 110. PLoS One 42. Miller A, Russell SJ. The use of the NIS reporter gene for optimizing – 2015;10:e0141669. oncolytic virotherapy. Expert Opin Biol Ther 2016;16:15 32.

www.aacrjournals.org Cancer Res; 78(24) December 15, 2018 6865

An Oncolytic Virus Expressing a T-cell Engager Simultaneously Targets Cancer and Immunosuppressive Stromal Cells

Joshua D. Freedman, Margaret R. Duffy, Janet Lei-Rossmann, et al.

Cancer Res 2018;78:6852-6865. Published OnlineFirst November 18, 2018.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-18-1750

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2018/12/01/0008-5472.CAN-18-1750.DC1

Cited articles This article cites 42 articles, 14 of which you can access for free at: http://cancerres.aacrjournals.org/content/78/24/6852.full#ref-list-1

Citing articles This article has been cited by 7 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/78/24/6852.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/78/24/6852. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.