ONCR-177, an Oncolytic HSV-1 Designed to Potently Activate Systemic Antitumor Immunity

Virotherapy with oncolytic viruses (OV) has shown promise in combination with immune checkpoint blockade (ICB) therapy in patients with cancers that have low T-cell infiltration. The mechanism of action of OV therapy involves preferential replication of OVs in tumor cells, resulting in cell lysis and release of new viral particles, which allows for continued infection of tumor cells to expand oncolysis and contribute to tumor debulking. The immunogenic cell death elicited by OVs results in the release of tumor antigens and danger-associated signals that enhance the expression of the antigen-presentation machinery, recruitment, activation, and maturation of CD8⁺ and CD4⁺ T cells and antigen-presenting cells (APC) to promote a tumor-specific immune response (1, 2).

Talimogene laherparepvec (T-VEC) is a herpes simplex virus-1 (HSV-1)–based OV and is currently the only OV therapy approved for the treatment of advanced melanoma in the United States and Europe (3, 4). T-VEC includes mutations to increase tumor cell selectivity, most notably the deletion of the ICP34.5 gene that results in reduced HSV-1 pathology, particularly in neurons. T-VEC expresses GM-CSF as a single transgene (CSF2). Clinical experience has shown that intratumoral administration of T-VEC into melanoma lesions frequently results in regression of the injected tumor. However, efficacy against visceral, untreated tumors has been limited, possibly explaining its modest uptake as a single-agent therapeutic (3, 5, 6). Reasons for lack of robust efficacy are likely multifactorial (7, 8). In relation to OV design, alterations within the viral vector to prevent potential infection/toxicity to normal tissues can simultaneously compromise replicative activity in tumors. For example, the aforementioned deletion of ICP34.5 in T-VEC may compromise its lytic potential, particularly in inflammatory tumor microenvironments (TME; ref. 9). Most OVs in the clinic or in development are either unarmed or include a single transgene, which may be insufficient for the optimal activation of antitumor immune responses. Comprehensive arming strategies are likely needed to achieve robust systemic antitumor effects. To this end, we introduce ONCR-177, an oncolytic HSV-1 designed to have several modes of action. An miRNA-mediated attenuation strategy (10) that allows for selective oncolysis of tumor cells, while sparing healthy tissue, was coupled with the expression of five immune-active transgenes. The overall strategy was to recruit and activate key immune cell types known to be involved in effective antitumor responses, while countering anticipated upregulation of clinically validated immune checkpoint pathways.

The following cell lines were obtained from the ATCC: NCI-H1299 (human lung carcinoma, CRL-5803), NCI-H1975 (human NSCLC, CRL-5908), A253 (human salivary gland carcinoma, CCL-81), A375 (human melanoma, CRL-1619), SKMEL-28 (human melanoma, HTB-72), COLO 205 (human colorectal carcinoma, CCL-222), SW837 (human colorectal carcinoma, CCL-235), FaDu (human squamous cell carcinoma, HTB-43), SCC25 (human squamous cell carcinoma, CRL-1628), A20 (mouse B-cell lymphoma, TIB-208), CT26 (mouse colon carcinoma, CRL-2638), B16F10 (mouse melanoma, CRL-6475), EMT6 (mouse mammary carcinoma, CRL-2755), Vero (monkey epithelial line, CCL-81), and Vero serum-free (SF) (CCL-81.5). MC38 (mouse colon carcinoma) cell line was from Joe Glorioso (University of Pittsburgh, Pittsburgh, PA). COLO800 (human melanoma) cell line was purchased from Sigma Aldrich (Cat. No. 93051123-1VL). Cell lines were verified to be free of mycoplasma and common mouse (IMPACT II Profile) or human (hIMPACT Profile III) pathogens by PCR testing (Idexx Bioanalytics). NCI-H1299, NCI-H1975, COLO 205, SW837, COLO 800, A20, and CT26 cell lines were cultured in RPMI (Gibco Thermo Fisher Scientific; Cat. No. 11965-092) plus 10% FBS (Gibco; Cat. No. 10100 NZ). A253, A375, SKMEL-28, SCC25, and Vero cell lines were cultured in DMEM (Gibco; Cat. No. 11965-092) containing 10% FBS (Gibco; Cat. No. 10100 NZ). The FaDu cell line was cultured in EMEM (ATCC; Cat. No. 30-2003) plus 10% FBS (Gibco; Cat. No. 10100 NZ). The EMT6 cell line was cultured in Waymouth's Media (Gibco; Cat. No. 11220035) plus 10% FBS (Gibco; Cat. No. 10082147). Vero-SF cells were cultured in SF Vero Growth Media (Gibco; Cat. No. 11681-020) containing 4 mmol/L l-glutamine (Gibco; Cat. No. 25030-081). To generate a B16F10 cell line expressing the HSV receptor Nectin1 (B16F10N1), 2 × 10⁵ cells were transduced with lentivirus-pCDH_Nectin1 that was constructed in-house with standard restriction enzyme cloning with a PCR insert derived from an open reading frame clone (Origene; Cat. No. MR20826). Cells were selected under puromycin dihydrochloride (Gibco; Cat. No. A1113803, 0.75 μg/mL), and single-cell clones were isolated. B16F10N1 tumor cells were maintained in vitro as a monolayer cell culture in DMEM (Gibco; Cat. No. 11965-092) plus 10% FBS (Gibco; Cat. No. 10100 NZ) and puromycin (0.75 μg/mL). All cell cultures were incubated in a humidified atmosphere with 5% CO₂ at 37°C and subcultured once reaching 80% confluency. Low passage numbers' (<10 generations) cells were used.

All vectors were designed and cloned in house using KG:124^BAC as the backbone for development of ONCR-159 (11). The generation, oncolytic potency, and the evaluation of the effectiveness of the miRNA attenuation of the base vector ONCR-159 have been previously characterized (10). Briefly, ONCR-159, the base vector of ONCR-177 and mONCR-171, was engineered by modification of the vector ONCR-125 (Supplementary Table S1). ONCR-125 is a replication-attenuated HSV-1 derived from KG:124^BAC that differs from the HSV-1 KOS strain with the following modifications: a deletion of the joint region (ΔJoint), the introduction of a gateway recombination cassette in the UL3/UL4 intergenic region for transgene insertion, a null mutation in US12 to promote antigen presentation on MHC Class I, the mutation of amino acids D285N and A549T in fusogenic glycoprotein B (gB) to enhance viral entry, and insertion of a single miR-T-124-3p in the 3′ untranslated region (UTR) of the essential gene ICP4 as described previously (11–15). In addition to these alterations, ONCR-159 was engineered with additional miRNA target (miR-T) cassette 3′UTR sequences in ICP4 (miR-T-1-3p and miR-T-143-3p), ICP27 (miR-T-128-3p, miR-T-219a-5p, and miR-T-122-5p), UL8 (miR-T-217-5p, miR-T-137-3p, and miR-T-126-3p), and ICP34.5 (miR-T-128-3p, miR-T-204-5p, and miR-T-219a-5p; ref. 10). ONCR-159 also harbors point mutations in the R2 region of UL37, shown to eliminate retrograde transport to the central nervous system (16).

For ONCR-177 and mONCR-171, a dual bidirectional promoter expression cassette was constructed and subcloned in the ONCR-159 base vector between UL3 and UL4 loci. Transgenes driven by the same promoter were separated by a T2A furin sequence (17). For ONCR-177, transgenes driven by the CAG promoter (CMV enhancer element/first intron and exon of chicken β-actin/splice acceptor of the rabbit β-globin gene) included the anti-human CTLA-4, ipilimumab (hIgG1, IgH, and IgL) (18), the extracellular domain (ECD) of human FLT3LG, and human CCL4. Transcripts were polyadenylated with the rabbit β-globin polyA signal sequence. Expressions of IL12A and IL12B of human IL12 and anti–PD-1 single variable heavy chain domain (VHH)-Fc (hIgG4 GenBank: AAB59394.1 with S228P mutation; ref. 19) were driven by a MND promoter, a synthetic promoter that contains the U3 region of a modified Moloney murine leukemia virus long terminal repeat (MoMuLV LTR) with a myeloproliferative sarcoma virus enhancer (20). Transcripts were polyadenylated with the SV40 polyA signal sequence.

For the ONCR-177 mouse surrogate virus mONCR-171, transgenes included anti-mouse CTLA-4 scFv-mFc (mIgG2a AAB59660.1; ref. 21), the ECD of Flt3l, and Ccl4, all driven by a CAG promoter and polyadenylated with the rabbit β-globin polyA signal sequence. The anti-mouse CTLA-4 blocker was from the 9D9 mAb (22) on a Fc effector–competent mIgG2a backbone. Expressions of l12a and Il12b subunits of mouse IL12 and anti–PD-1 VHH-mFc (mIgG1 BAX57196.1) were driven by an MND promoter and polyadenylated with the SV40 polyA signal sequence.

All genetic alterations leading to ONCR-159, ONCR-177, and mONCR-171 were conducted on their respective full-length genome cloned in a bacterial artificial chromosome (BAC) originating from KOS-37 BAC (10, 11, 23). All modifications were confirmed by Sanger sequencing, and transgene expression was demonstrated by Western blotting or ELISA. The excision of the BAC vector sequence, including the ori-replication and the antibiotic selection marker, was achieved by Cre-loxP recombination and confirmed by X-gal staining of ONCR-159–, ONCR-177–, or mONCR-171–infected cells. BAC-removed viruses were plaque purified in the Vero-SF cell line prior to expansion, purification by centrifugation, determination of titer, and vialing (10, 11).

PD-1 VHH-mFc was generated by WuXi Biologics by immunizing calemids with the human and mouse PD-1 ECD and screening for specific binders and inhibition of PD-L1 binding. Association and dissociation rate constants (ka and kd), and the dissociation equilibrium binding constant (Kd) for recombinant mouse (rmu) PD-1 and rhu PD-1 binding to anti–PD-1 molecules were determined using surface plasmon resonance “Biacore” technology. Anti–PD-1 [VHH-Fc (mIgG1)] was purified by Olympic Protein Technology. Proteins included rhu PD-1/PDCD1- His1 10377-H08H (Sino Biological) and rmu PD-1/PDCD1- His1 50127-M08H (Sino Biological). Biosensor analysis was conducted at 25°C in a HBS-P buffer system (10 mmol/L HEPES, pH 7.4, 150 mmol/L NaCl, and 0.05% Surfactant P20) using a Biacore 3000 optical biosensor equipped with a CM5 sensor chip (GE; Cat. No. BR100012). The autosampler was maintained at 8°C. Goat anti-mouse IgG capture antibody (Jackson ImmunoResearch Laboratories, Inc.; Cat. No. 115-005-071) was immobilized to Fc1 and Fc2, and goat anti-mouse IgG capture antibody (Jackson ImmunoResearch Laboratories, Inc.; Cat. No. 109-005-098) was immobilized to Fc3 and Fc4 of the sensor chip using standard amine coupling chemistry to a level of 11,000 to 15,000 RU. This surface type provided a format for stable capture of Fc proteins (ligand). Kinetic rate coefficients were recovered from binding analysis experiment performed with a Biacore 3000 biosensor. Results are reported as the global fits to 1:1 binding model ± standard error. The affinity (K_D) of anti–PD-1 VHH-mFc for human and mouse PD-1 was determined by surface plasmon resonance to be 3.44 nmol/L and 28.7 nmol/L, respectively. Anti–PD-1 VHH-mFc inhibited the binding of human PD-L1 and PD-L2 to cell surface human PD-1 with IC₅₀ of 0.6 and 0.9 nmol/L, respectively.

Human cell lines A375, SK-MEL-28, SW-837, COLO 205, FaDu, and SCC25 were plated at 10,000 per well and infected with a mock (OptiMEM-reduced serum media, Gibco; Cat. No. 31985062) or serially diluted virus suspension with a multiplicity of infection (MOI) range of 0.001 to 30. Seventy-two hours after infection, cell viability was determined by addition of an equal media volume of CellTiter-Glo reagent (Promega), as a readout for total ATP consumption and cell viability. IC₅₀ was calculated with GraphPad Prism 8.0 as a mean of four replicates.

NCI-H1299 or NCI-H1975 (1 × 10⁵) human lung cancer cell lines were pretreated with or without recombinant human IFNα-2α (5,000 UI/mL, R&D Systems; Cat. No. 11100-1) per well, in compete cell culture media (RPMI + 10% FBS). On the following day, media were removed and substituted with 100 μL of OptiMEM (Gibco; Cat. No. 11058-021) containing 0.03 MOI of ONCR-177 or G207 (24). G207, an ICP34.5-deleted oncolytic HSV-1, was kindly provided to Oncorus by Michael Christini, Aettis, Inc. Conditioned media and cells were collected 48 hours after infection, and infectious viral particles were determined in a plaque assay on Vero cells. Statistical analysis was performed using GraphPad Prism 8.0 using two-way ANOVA, Sidak multiple comparison test.

A253 salivary gland tumor cells were transiently transfected with the indicated miRNA mimics and infected with ONCR-177 (MOI 0.1), and supernatants were collected after 3 days. A plaque titer assay, where supernatants were diluted 10-fold serially and added to Vero cells, was performed in biological and technical duplicates to assess the miRNA-mediated effect on ONCR-177 viral replication. The titer ratio was determined by normalizing the test conditions to the plaque counts obtained by the negative mimic control. Statistical significance was calculated using Sidak multiple comparisons test versus negative mimic control. Transgene expressions from cellular lysates were determined using ELISAs as described below.

Vero cells (CCL-81, ATCC, passage 11) were cultured at 1.0 × 10⁵ cells/well in 1 mL Vero Growth Medium (VGM, DMEM + 10% FBS) and incubated at 37°C in a CO₂ incubator. The following day, ONCR-177 (Lot 18-177-070) stock was diluted to a concentration of 2.5 × 10⁵ plaque-forming unit (PFU)/mL in DMEM. Media were removed from Vero cells' culture, and viral infection was performed with 500 PFU (200 μL) of diluted ONCR-177 per well. The virus was adsorbed for a period of 90 minutes at 37°C in a CO₂ incubator with rocking, and then acyclovir (20 mg/mL; USP 1012065, LOT: K0L516) in DMSO was added to VGM to a final concentration of 8 μg/mL. Ten serial 2-fold dilutions of acyclovir in VGM were prepared. At the end of adsorption period, 1 mL of VGM with or without acyclovir was added to wells in triplicate. Plates were then incubated for 18 hours before the addition of 1 mL VGM with the corresponding concentration of acyclovir and 4% pooled human serum. The plates were incubated for an additional 6 days, media were aspirated, and plates were stained with a 1% crystal violet solution. The plaque number was counted for each well, and the PFU/mL of each sample incubated with acyclovir was calculated and compared with the infection without acyclovir to calculate the percent plaque reduction. Percent plaque reduction was calculated according to the following formula:

IC₅₀ values were calculated using GraphPad Prism 7.05 using the nonlinear fit of (inhibitor) versus response – variable slope (four parameters). To analyze human transgene expression, Vero cell lysates were analyzed as detailed below using ELISA to detect human transgenes.

Tissue (injected and contralateral A20 mouse tumor, liver, blood) lysates processed after in vivo oncolytic HSV (oHSV) administration were analyzed by ELISA (all from R&D Systems), according to the manufacturer's instructions. Briefly, frozen (−80°C) lysates were thawed, resuspended in PE-LB lysis buffer (an organic buffer containing a nonionic detergent for the extraction of soluble proteins from tissue samples, G-Biosciences), and placed on ice. Standards were prepared following the human Quantikine ELISA kit protocol (all R&D Systems) for mouse IL12 (Cat. No. SM1270), FLT3LG (Cat. No. MFK00), or CCL4 (Cat. No. MMB00), according to the manufacturer's instructions. The A20 tumor lysates were diluted either 1:2 (25 μL per well) for IL12 and FLT3LG or 1:20 (2.5 μL per well) for CCL4. Absorbance reading at 450 nm was obtained using the end point ELISA HRP-TMB protocol, which was located on the protocol library of the SpectraMax i3x plate reader. Absorbance at 570 nm was obtained in order to correct for plate imperfections. Net absorbance was obtained by subtracting the absorbance values of the samples and standards from that of the blank controls, using Microsoft Excel software. A standard curve was generated for each experiment from the net absorbance values, and concentration of the unknown samples was calculated. Average of duplicates measured from each biological replicate was calculated. Graphs were generated using GraphPad Prism version 8.0. Each bar represents the mean and ± SD. The assay lower limits of quantification for mouse IL12, FLT3LG, or CCL4 were 7.81, 31.3, and 7.8 pg/mL, respectively.

ONCR-177 oncolysates of A253 cells, Vero cells, or tissue lysates processed after in vivo oHSV IT administration (FaDu xenograft tumor and liver) were analyzed by ELISA (all from R&D Systems), according to the manufacturer's instructions. Briefly, frozen (−80°C) lysates were thawed, resuspended in PE-LB lysis buffer, and placed on ice. Standards were prepared following the human Quantikine ELISA kit protocol (all R&D Systems) for IL12 (Cat. No. D1200), FLT3LG (Cat. No. DFK00), or CCL4 (Cat. No. DMB00). To detect the anti–PD-1 or anti–CTLA-4 transgenes, a human IgG4 (Thermo Fisher Scientific; Cat. No. BMS2095) or human IgG1 ELISA (Abcam; Cat. No. ab100548) was used according to the manufacturer's instructions. The diluent used to prepare the standards was included as a blank control. The samples were diluted 1:20 in the same diluent as that used to prepare the standards. The ELISA was performed in technical duplicates following the kit protocol. Note that 2N sulfuric acid was used to stop the substrate reaction, and the absorbance reading at 450 nm was obtained using the end point ELISA HRP-TMB protocol on the SpectraMax i3x plate reader. Absorbance at 570 nm was obtained in order to correct for plate imperfections. Further analyses were as described for mouse ELISAs.

Human induced pluripotent stem cell (iPSC)–derived neurons, hepatocytes, and cardiomyocytes were provided by Dr. Gustavo Mostoslavsky (Boston University Center for Regenerative Medicine, Boston, MA). Cells were differentiated and validated as described previously (25, 26) in a 12-well format. SK-MEL-28 cells (1 × 10⁵) or iPSC-derived cells (1 × 10⁶) were infected with 0.2 MOI of ONCR-125 or ONCR-177. Supernatants were collected 24 hours after infection, and virus infectious particles were determined in a plaque assay in Vero cells. Statistical comparisons were conducted by Tukey multiple comparisons test.

To generate lysates, MART1-expressing COLO800 cells at 80% confluence were infected with either mONCR-171 or ONCR-159 at an MOI of 1 and incubated for 1 hour at 37°C with 5% CO₂. After an hour, complete media [RPMI 1640 (Gibco) + 2 mmol/L Glutamine (Gibco) + 10% FBS] were added. On the following day when cells were more than 90% detached, the cells were collected. Control lysate was generated by 5 cycles of freezing and thawing (freezing at −196°C followed by thawing at 37°C). mONCR-171, ONCR-159, and the control cell lysates were centrifuged at 97 × g at 18°C to 20°C for 10 minutes to remove large cellular debris. The supernatant was then filtered with a 0.22-μm sterile low protein-binding filter unit (Millex). Proteins were concentrated with a Vivaspin 20–3K tube (Sartorius), as recommended by the manufacturer.

Dendritic cells (DC) and MART-1⁺ CD8⁺ T cells were generated or expanded from peripheral blood mononuclear cell (PBMC) donors. Whole blood was collected from healthy donors as described in the ethical protocol/amendment IXP-001_V3 (Belgium; Reg. Nr. B6702014215858), protocol IXP-003_V1 (Belgium; Reg. Nr. B707201627607), or protocol IXP-004_V1 (The Netherlands; Reg. Nr. NL57912.075.16). For DC generation, monocytes were isolated by positive magnetic separation (Miltenyi; Cat. No. 130- 050-201) with >90% purity, following the manufacturer's instructions, and cultured for 5 days in SF CellGenix medium (CellGenix GmbH) supplemented with human IL4 (Miltenyi; Cat. No. 130-093-922) at 50 ng/mL and human GM-CSF (Miltenyi; Cat. No. 130-093-864) at 50 ng/mL. On day 5, the monocytes were differentiated into immature (i)DCs. IDCs were then loaded with a known maturation marker, LPS (1 μg/mL; from Escherichia coli 0111:B4, γ-irradiated, BioXtra, suitable for cell culture; Sigma-Aldrich; Cat. No. L4391-1MG), or lysate [200 μg lysate/1 × 10⁶ cells/mL (1)]. Twenty-four hours after loading, the iDCs are matured into antigen-presenting mature (m)DCs. To expand MART-1⁺ CD8⁺ T cells, after 5 days of culture, the iDCs were matured for 24 hours using MEGA CD40L maturation agent (Enzo Life Sciences; Cat. No. ALX-522-110-C010) at 0.1 μg/mL. Upon maturation, the cells were loaded with MART-1 peptide (50 μg/mL), gamma-irradiated, washed before coculturing with 1 × 10⁶ autologous CD8⁺ T cells, and then isolated using positive selection (Miltenyi; Cat. No. 130-096-495) in 48-well plates at a 1:8 ratio in RPMI 1640 (Thermo Fisher Scientific; Cat. No. 31870074) with 10% human serum AB Serum (heat-inactivated; Biowest; Cat. No. S419H-100), IL7 (5 ng/mL; Bio-Techne R&D Systems; Cat. No. 207-IL-005), 2 mmol/L l-glutamine (Life Technologies; Cat. No. 25030024), gentamycin (50 μg/mL; Westburg; Cat. No. LO 17-518L), and 100 μmol/L 2-mercaptoethanol. IL2 (10 IU/mL) was added to the culture medium every 2 days, and culture medium was refreshed once per priming/stimulation round. After the 7 days of initial priming, the cells were further enriched and restimulated by addition of MART-1 peptide–loaded autologous monocytes at a 1:8 ratio for another 7 days.

For IFNγ ELISpots, MART-1–specific CD8⁺ T cells were diluted to a concentration of 5 × 10⁵ cells/50 μL in RPMI and rested for 1 hour at 37°C. After resting, CD8⁺ T cells were seeded on IFNγ precoated ELISpot plates (Mabtech; Cat. No. 3420-2APW-10). After 20 hours of incubation at 37°C, IFNγ⁺ spots were visualized, and data were acquired with a Mabtech IRIS reader and analyzed using the Mabtech Apex TM software. ELISpot data were presented as delta of spot-forming units (ΔSFU), which is the number of spots in the test conditions subtracted by the number of spots in the reference condition. To obtain the ΔSFU IFNγ (per million), the SFU were first multiplied by 1 million divided by the number of cells used in the assay to obtain SFU per million. The average of all replicates per condition was then subtracted by the average of the CD8⁺ T-cell only condition.

In vivo experiments with mouse subcutaneous tumor models were conducted with 8- to 12-week-old, 18 to 22 g BALB/c, C57BL/6, or athymic nude mice female mice, all sourced from Charles River Laboratories. Animals had ad libitum access to sterile pelleted feed and reverse osmosis–purified water and were maintained on a 12:12-hour light:dark cycle with access to environmental enrichment. All animal protocols were approved by the Oncorus Institutional Animal Care and Use Committee (IACUC) and were performed in accordance with IACUC regulations.

For the A20 syngeneic tumor model, BALB/c female mice were inoculated s.c. with 5 × 10⁶ A20 cells in 0.1 mL on the upper right flank, and, for bilateral tumor studies, 5 × 10⁶ A20 cells also in the upper left flank. For the bilateral MC38 and B16F10N1 syngeneic tumor models, C57BL/6 mice were inoculated s.c. with 5 × 10⁵ tumor cells on the upper right flank and upper left flank. For the bilateral CT26 syngeneic tumor model, BALB/c mice were inoculated s.c. with 1 × 10⁶ CT26 cells per implant on both the upper right and upper left flanks. For the FaDu xenograft tumor model, athymic nude mice were inoculated s.c. with 2 × 10⁶ tumor cells on the upper right flank. For all in vivo studies, inoculated cells were >90% viable and implanted in 0.1 mL SF DPBS (Gibco).

Prior to study initiation, animals were identified by ear-tag, and the tumor implantation sites were shaved. Prior to tumor cell implantation, animals were anesthetized with 5% isoflurane to ensure subcutaneous delivery of cells. Randomization, enrollment, and treatment were initiated when an adequate number of mice with properly sized tumors was identified. For most studies, median right and left starting tumor volumes (TV) of approximately 100 ± 25 and 100 ± 50 mm³, respectively, were used. For the B16F10N1 model, due to its aggressive growth rate, a median right and left flank starting TV of approximately 80 ± 30 mm³ was followed. Animals were pair-matched based on right TV size and randomly placed into treatment groups.

Doses were formulated according to the following formula:

where V_s is volume of virus stock solution (μL), V is volume of virus preparation being prepared for dosing (μL), D is dose in PFU, C_S is the concentration of the virus stock (PFU/μL), and V_I is volume of injection (μL).

All viruses were formulated immediately before dosing and kept on ice until administration. A prespecified volume of virus aliquot (V_S) was diluted in a volume of sterile DPBS (V_DPBS) and gently mixed. Intratumoral administration of virus (25 μL) into the tumor tissue was done using a sterile 31G insulin needle (bevel up) attached to a sterile syringe. For most efficacy studies, mice received 3 intratumoral injections on days 1, 4, and 7, with day 1 being considered the first day of dosing. Dose (PFU) of mONCR-171, which varied depending upon the tumor model and study, is indicated in the figure legends. For studies testing mONCR-171 in combination with PD-1 antibody, mice in the appropriate treatment groups received three i.t. injections on days 1, 4, and 7 of either PBS or mONCR-171 (3 × 10⁶ PFU/dose, 25 μL dosing volume) or three i.p. injections on days 1, 4, and 7 of either mouse IgG2a isotype control or anti-mouse PD-1 (200 μg/dose, 100 μL dosing volume, Bio X Cell). Study duration depended upon the tumor growth rate and response to therapy. Some studies used a “survival” endpoint, which in the case of subcutaneous tumors represented a predetermined total (summarized right and left) TV, typically >2,000 mm³. Takedowns and tissue (right and left tumor) harvests for mode of action studies typically occurred on day 8, which was 1 day after completion of a dosing regimen of 3 doses given every 3 days. Throughout the course of the in vivo study, animals were monitored daily for adverse events, and body weight was recorded twice per week. TVs, measured at least 2 times per week, were calculated using the following formula: TV = a × b²/c, where “a” is tumor length and “b” is tumor width. Mice were humanely euthanized once the combined tumor burden reached the endpoint value of >2,000 mm³. Longitudinal TV and body weight data were captured in Microsoft Excel format and then transferred to GraphPad Prism 8.0 software for plot generation and statistical analyses. Tumor growth curves were fitted with a linear mixed model with the statistical software SAS JMP version 13 (SAS Institute Inc.), after log transformation of TVs. Data were plotted as mean TV ± SEM or as tumor growth curves for individual mice in each treatment group. Survival curves were plotted in 1GraphPad Prism 8.0 software and analyzed with a log-rank test.

Depleting antibodies were administered to BALB/c female mice by intraperitoneal injection on a twice-per-week schedule for 2 weeks preceding A20 tumor cell implantation and continued throughout the course of the efficacy study. Anti-mouse CD8α (Clone 2.43, Bio X Cell) and its isotype control rat IgG2b (Clone LTF-2, Bio X Cell) were dosed at 400 μg for the first dose and then 200 μg for subsequent doses. Polyclonal anti-asialo GM1 (Wake Pure Chemicals Industries) and its control normal rabbit serum (ImmunoReagents) were dosed at 50 μL for the first dose and then 25 μL for the subsequent doses. Selective immune cell subset depletion was confirmed 1 week after initiation of the depleting treatments by analysis of CD45⁺ blood leukocyte populations by flow cytometry. For the efficacy portion of the study, approximately 100 mm³ starting volume tumors were dosed i.t. with PBS or mONCR-171 (at 3 × 10⁶ PFU) on days 1, 4, and 7, with day 1 being the first day of dosing. Animals were monitored until a total TV of approximately 2,000 mm³ was reached.

For IHC and flow cytometric analysis, A20 or MC38 tumor–bearing mice, respectively, were generated in an identical manner as for efficacy studies. Animals with 100 mm³ right and left starting TVs were treated i.t. on days 1, 4, and 7 with DPBS, ONCR-159 (5 × 10⁵ PFU/dose), or mONCR-171 (5 × 10⁵ PFU/dose). On day 8, which was 1 day after the third and final dose, right and left tumors (n = 5 mice/treatment group each for IHC and flow cytometry) were harvested, weighed, and processed. For IHC, tumors were fixed in 10% formalin for 24 hours, transferred to 70% ethanol, followed by processing to paraffin-embedded blocks. Slides from these blocks were stained with a 1:100 dilution of mouse CD3ϵ antibody (clone SP7, Thermo Fisher Scientific). Morphometric enumeration of murine CD3⁺ T cells was determined after whole slide scanning followed by evaluation using QuPath image analysis software (27). Regions of nonviable tissue were excluded from the analysis. Data were graphed using GraphPad Prism software.

For flow cytometry, tumors were disaggregated to single-cell suspensions prepared using a Miltenyi GentleMACs Octo-dissociator with heaters, following the manufacturer's instructions. LIVE/DEAD Fixable Red Dead Cell Stain Kit, for 488 nm excitation (Invitrogen; Cat. No. L32102), was used to assess the viability. Surface cell staining was performed using BD Stain Buffer (Becton Dickinson; Cat No. 554656) with the antibodies for mouse CD45 (30-F11), CD3ϵ (17A2), CD8α (53-6.7), CD4 (RM4-5), CD25 (PC61), CD69 (REA937), CD49b (DX5), MHCII (M5/114.15.2), CD11c (HL3), and CD103 (M290). True-Nuclear Transcription Factor Buffer Set (BioLegend; Cat. No. 424401) was used for FOXP3 (150D) intracellular staining. All antibodies were purchased from BioLegend, except CD69 (Miltenyi), and CD49b, CD11c, and CD103 (BD Biosciences). Data were acquired on a BD LSRFortessa using BD FACSDiva software and analyzed using FlowJo software. For the T-cell and natural killer (NK) panel, cells were first gated for lymphocytes (SSC-A vs. FSC-A) and singlets (FSC-H vs. FSC-A). The lymphocyte gate was further analyzed for their uptake of the live/dead stain to determine live versus dead cells and CD45 expression. Cells were then gated on CD3 versus SSC-A to select T cells or CD49b versus SSC-A to select the NK cells. For the T cells, the population was gated for CD4 versus CD8, and the CD4⁺ T cells were further gated for CD25⁺ and FOXP3⁺ to analyze regulatory T cells (Treg). NK cells were gated for CD49b⁺ and CD3⁻. For the myeloid panel, cells were first gated for lymphocytes (SSC-A vs. FSC-A) and singlets (FSC-H vs. FSCA). The lymphocyte gate was further analyzed for their uptake of the live/dead stain to determine live versus dead cells and CD45 expression. Then DCs were gated using CD11c⁺ IA-IE⁺ and subsequently on CD8⁺ and CD103⁺ DCs.

C57BL/6 animals initially bearing bilateral MC38 tumors (mean, 125 mm³; range, 100–165 mm³) were administered two i.t. doses, on days 1 and 4, of PBS or 3 × 10⁶ PFU of mONCR-171. Tumors were harvested and disaggregated on day 6 (as described above), followed by in vitro stimulation with the tumor-specific peptides Adpgk and Reps1 (GenScript; 1 μg/mL; ref. 28) for 5 hours in a 37°C humidified incubator in the presence of Golgi plug before surface and intracellular staining for IFNγ: CD45-APC-Fire (30-F11), CD3-PerCP-Cy5.5 (17A2), CD8-Alexa Fluor 700 (53-6.7), IFNγ-APC (XMG1.2) using a similar staining protocol as noted above. All antibodies were purchased from BioLegend. Data were acquired on a BD LSRFortessa using BD FACSDiva software and analyzed using FlowJo software. Statistics were calculated with a two-way ANOVA in GraphPad Prism.

A cohort of BALB/c female mice bearing A20 single-flank subcutaneous tumors were administered either PBS or mONCR-171 (3 × 10⁶ PFU) by i.t. injection on days 1, 4, and 7. Long-term (>70 days) follow-up verified that most (93%) of the animals treated with mONCR-171 were cured of their tumors. In contrast, tumors from all PBS control–treated animals rapidly progressed, necessitating early removal from the study. Three to 5 months after confirmation of complete tumor regression, naïve BALB/c female mice, or animals cured of A20 tumors, were challenged or rechallenged with 5 × 10⁶ A20 cells or 1 × 10⁶ EMT-6 cells and monitored for 55 days. Animals were humanely euthanized once the study endpoint was reached, which was a tumor burden >1,000 mm³.

For all biodistribution studies, mice with approximately 100 mm³ subcutaneous right and left flank tumors were administered a single intratumoral dose of ONCR-159 or mONCR-171 to the right (“injected”) flank tumor. For dose-escalation studies, animals (n = 5 mice per group) were administered PBS, ONCR-159 (3 × 10⁵ PFU), or log doses of mONCR-171 that ranged from 3 × 10³ PFU to 3 × 10⁷ PFU. Tissues were collected 24 hours after dosing. For kinetic evaluation, a single 3 × 10⁶ PFU i.t. dose of ONCR-159 or mONCR-171 was administered, and tissues were collected at 4, 24, 72, and 168 hours after dosing. Tumor (injected and contralateral), liver, blood (obtained by cardiac puncture), and plasma were flash frozen in liquid nitrogen and stored at −80°C until processing for downstream assays. For tissue homogenization, tumors were placed in sterile 5 mL tubes containing 9.5 mm grinding balls. The samples were pulverized using the 1600 miniG tissue pulverizer. A small amount of tumor from the dose–response samples was collected and lysed with a tissue lysis buffer comprised of buffer ATL (an SDS-containing buffer) and 20 μL proteinase K from the DNeasy blood and tissue kit (Qiagen; Cat. No. 69506) per 25 mg of tissue. A small amount of tumor from the kinetic study samples was weighed and lysed with buffer RLT plus (a guanidinium thiocyanate–containing buffer). Large tissues, such as liver, were pulverized on the cryoPREP Pulverizer (Covaris). Small tissues were collected and lysed with 180 μL ATL buffer and 20 μL proteinase K/25 mg of tissue.

Genomic DNA from the dose–response samples was isolated using the DNEasy blood and tissue kit protocol on the Qiacube (Qiagen). Genomic DNA and total RNA from the kinetic study samples were isolated using the All prep DNA/RNA mini kit protocol (Qiagen; Cat. No. 80204) on the Qiacube. For protein purification, tumors were weighed, and based on their weight, samples were lysed using tissue PE-LB. Halt protease and phosphatase inhibitor was added to the lysis buffer, as described for ELISA. Samples were centrifuged, and the supernatant frozen at −80°C.

Genomic DNA from tumors and livers was quantified using a nanodrop and diluted to 25 ng/μL in nuclease-free water. Blood samples were diluted to 10 ng/μL. Five μL/well of either the DNA sample (corresponding to 125 ng of DNA from tumors or 50 ng of DNA from blood) or the US6 plasmid standard which was prepared in-house was used in a qPCR reaction and performed in technical triplicates to detect HSV-1 genomes on a Quantstudio5 qPCR instrument with the following thermal cycler conditions—Hold Stage: 95°C for 1 second, PCR stage: 95°C for 1 second, followed by 60°C for 30 seconds. The PCR stage was performed for 45 cycles. The plasmid standards used in the assay were used to generate a standard curve with known amount of copies corresponding to a set of Ct values. Based on this standard curve, the amount of HSV genome copies in the samples was calculated. The copy-number calculations were done on the Quantstudio software (Thermo Fisher Scientific). Because 125 ng/well of tumor DNA or 50 ng/well of blood DNA was loaded, a conversion factor of 8 (for solid tissues) or 20 (for blood) was applied to the copy number in order to plot them as copies/μg of DNA. Technical duplicates were averaged, and the results of individual animals within each group were plotted using GraphPad software. Conversion of copy number to copies/μg and calculations of averages were done on Microsoft Excel. HSV-1 US6 PCR Primer/probe: Fwd: 5′-CCCGCTGGAACTACTATGACA-3′; Rev: 5′-GCATCAGGAACCCCAGGTT-3′; Probe: 5′-TTCAGCGCCGTCAGCGAGGA-3′ with FAM as the reporter dye and MGB-NFQ as the quencher.

RNA from the injected tumors of the kinetic study samples was quantified on the nanodrop. RNA (1 μg) was subjected to 1X DNase I (Life Technologies; Cat. No. 18068-015) treatment to eliminate genomic DNA followed by a reverse transcription reaction using Superscript IV Vilo master (Thermo Fisher Scientific; Cat. No. 11756050) mix to synthesize complementary DNA (cDNA). A no reverse transcriptase control was included in the assay. The cDNA was diluted 1:5 with nuclease-free water and used for qPCR and performed in duplicate to detect either VP5 or GAPDH as the house-keeping gene. The qPCR instrument and thermal cycler conditions are identical as outlined above. The Ct values of the house-keeping gene (GAPDH) were subtracted from those of VP5 for expression normalization. The formula 2^−Ct was then applied to the Ct difference between VP5 and GAPDH. Because the resultant number was small, it was multiplied by 10⁷ to convert into a whole number. This value was then reported as the normalized VP5 expression. An average of technical duplicates was calculated, and results from individual animals were plotted on GraphPad Prism 8.0 software. All the calculations were done on Microsoft Excel. VP5 primers/probe: Fwd: 5′-TCGTATTGCAACACCCTGTC-3′; Rev: 5′-TGGTGGACCTCGAACTGTA-3′; Probe: 5′-TGTGTACCAAGTTTCCGGAGCTGG-3′ with FAM as the reporter dye and TAMRA as the quencher. The mouse Gapdh primers/probe set was obtained from Thermo Fisher Scientific (Assay ID: Mm99999915_g1).

The Mouse Cytokine/Chemokine Magnetic Bead Panel (MilliporeSigma; Cat. No. MCYTOMAG-70K) was used to assess mouse cytokines/chemokines in tumor tissue or blood. Tumor lysates resuspended in PE-LB lysis buffer were thawed and placed on ice. Standards were prepared following the instructions in the Mouse Cytokine/Chemokine magnetic bead kit protocol. The magnetic beads representing each of the analytes were vortexed, and the beads were pooled and resuspended in the assay buffer. The tumor samples (25 μL) were tested undiluted and in technical duplicates. The plasma samples were diluted 1:2 in assay buffer and tested in technical duplicates. The Luminex assay was performed following the kit protocol with an overnight incubation of samples with magnetic beads at 4°C instead of 2 hours at room temperature. Unbound beads were washed off the next day, and the assay proceeded as per the protocol. After the last wash, samples were resuspended in drive fluid (Thermo Fisher Scientific; Cat No. MPXDF4PK) and read on the Magpix instrument (Luminex). All data analyses were performed using the MAGPIX software. Based on the standard curve, the concentrations of the different analytes in the samples were calculated. The final concentrations were obtained by applying the dilution factor. Values were plotted on GraphPad Prism 8.0 software.

Flash-frozen tumors (n = 5 mice per treatment group) were pulverized, and RNA was extracted using the Rneasy Mini Qiacube extraction kit according to the manufacturer's protocol. Total isolated RNA (60 ng) was mixed in a final volume of 25 μL with 3′ biotinylated capture probe and 5′-reporter probes from Mouse PanCancer IO 360 Gene Expression Panel (NanoString Technologies). Hybridization was conducted at 65°C for 16 hours. Hybridized samples were isolated on the NanoString nCounter preparation station, where excess probe was removed, and hybridized complexes of probes and target RNA sequences were immobilized on the cartridge (Cat. No. SPRINT-CAR-1.0). The cartridge-bound samples were then analyzed using nCounter digital analyzer. Data collection was carried out on the following manufacturer's instructions. For each assay, a high-density scan (600 fields of view) was performed. Results were analyzed using the nSolver Analysis Software 4.0 (NanoString Technologies). The raw data were subjected to nSolver Advanced analysis that performed data normalization using the geometric mean of house-keeping genes, pathway scoring, and cell type profiling with a P value threshold set to 0.05.

The construction and characterization of ONCR-159, the base vector for ONCR-177, have been described (10). Briefly, ONCR-159 harbors tissue-specific miR-T sequences within three essential virus genes (UL8, ICP27, and ICP4) and the ICP34.5 gene (Fig. 1A). This miRNA attenuation strategy allows for replication to occur in tumor cells but not in normal tissues that are vulnerable to HSV-1 infection (10). Retention of ICP34.5 in ONCR-177 may allow for efficient viral replication in the presence of type I IFN, commonly found in the TME. Consistent with this hypothesis, replication of G207, an oncolytic ICP34.5-deleted HSV-1 (13, 24), was either abolished (H1299 line) or reduced 135-fold (H1975 line) in the presence of IFNα, whereas ONCR-177 replication was modestly suppressed 4.5-fold (H2199 line) to 6-fold (H1975 line) in the presence of IFNα (Fig. 1B). miRNA attenuation mechanisms were operational as demonstrated by potent suppression of ONCR-177 replication and transgene expression after the transient transfection of siRNAs acting as miRNA mimics (Fig. 1C and D). Compared with a transformed cancer cell line, viral replication was suppressed in nontransformed cells such as neurons, hepatocytes, and cardiomyocytes derived from iPSCs (Fig. 1E).

Mutation of the tegument protein UL37 provides an additional safety mechanism by preventing axonal retrograde transport of HSV-1 capsid and therefore the establishment of latent infection (16). As a third safety mechanism, a functional thymidine kinase gene maintained sensitivity, at the levels of viral replication and transgene expression, to clinically achievable levels seen with acyclovir (Supplementary Fig. S1). Despite extensive genetic alterations, ONCR-159 and ONCR-177 retained potent oncolytic activity across a panel of human cancer cell lines (Supplementary Table S2). Finally, DCs loaded with ONCR-177 lysates had significantly increased ability to activate melanoma antigen MART-1⁺ autologous CD8⁺ T cells compared with negative control or ONCR-159 lysates (P ≤ 0.0001 across PBMC donors; Fig. 1F). These results suggest that ONCR-177–encoded transgenes contributed to the ability of DCs to efficiently activate tumor antigen–specific CD8⁺ T cells.

Biodistribution kinetics of ONCR-177 were assessed in the sensitive (Supplementary Table S2) FaDu xenograft tumor model. After an i.t. injection of 3 × 10⁶ PFU of ONCR-177, 10⁵ to 10⁷ HSV genome copies/μg of viral DNA were detected throughout the 4- to 168-hour timeframe (Fig. 2A). In contrast, viral DNA was not detectable in liver until 72 hours after dose, at >1,000-fold less than that observed in injected tumors (Fig. 2B). Because the qPCR assay cannot distinguish between live virus and nonviable viral particles, RNA transcripts of the UL19 (VP5) viral gene, which should only be expressed by live virus, were assayed by RT-PCR. Analysis of tumor VP5 gene expression suggested live virus persisted for extended periods of time within injected tumors (Fig. 2A). In contrast, viral DNA detected at low levels at later time points in the liver was not derived from live virus (Fig. 2B). Intratumoral transgene [(hIL12, hFLT3LG (ECD), hCCL4, hIgG4 (anti–PD-1), and hIgG1 (anti–CTLA-4)] concentrations reached peak mean concentrations of 2,184, 2,277, 257, 2,100, and 1,000 pg/mL, respectively (Fig. 2C). ONCR-177 transgene expression kinetics in tumors largely mirrored that of virus DNA, with expression peaking 24 to 72 hours after dose and decreasing by the 168-hour time point.

Because human IL12 and ipilimumab are not active in mice, a mouse surrogate virus to ONCR-177, mONCR-171, was generated and employed in nonclinical pharmacology studies to assess biodistribution, efficacy, and mode of action. mONCR-171 utilized the same ONCR-159 base vector as ONCR-177 (Fig. 1A), but expressed the mouse homologs for IL12, CCL4, and FLT3LG (ECD). The same anti–PD-1 VHH nanobody for ONCR-177 (on an mIgG1 backbone) was included in mONCR-171 because it is cross-reactive to mouse with an 8-fold weaker K_D. Finally, the anti-mouse CTLA-4 blocker was from the 9D9 monoclonal antibody (22) on an Fc effector–competent mIgG2a backbone. To account for immune-mediated effects on virus biodistribution, the biodistribution of mONCR-171 was assessed in the oHSV-sensitive A20 immune-competent syngeneic tumor model (Fig. 3A). A mONCR-171 dose–dependent increase in viral DNA was noted in injected tumors, whereas little to no viral DNA was detected in contralateral tumors or livers. In blood, viral DNA was detectable for some animals at the highest doses of mONCR-171, albeit at levels approximately 1,000-fold below that was observed for the corresponding dose in injected tumors. Similarly, concentrations of mONCR-171 transgenes IL12, FLT3LG (ECD), and CCL4 increased in a mONCR-171 dose–dependent manner in injected tumors. In contrast, for contralateral tumors, transgene expression either was not detectable or was detected at minute concentrations (Fig. 3B). Assays to detect the anti–PD-1 and anti-mouse CTLA-4 in mouse tissues were not developed. IFNγ and CXCL10, which are downstream of IL12 and IFNγ, respectively, and indicative of an inflammatory antitumor T_H1 response (29), were elevated in injected tumors (Fig. 3C). Other inflammatory cytokines, such as TNFα and IL6, increased in a dose-dependent manner in injected, but not contralateral, tumors. In blood, IFNγ and CXCL10 were detectable in a mONCR-171 dose–dependent manner (Fig. 3D). TNFα and IL6 are two key cytokines implicated in cytokine release syndrome, an exaggerated systemic immune response that without clinical intervention can be lethal (30). These cytokines were undetectable in the plasma after intratumoral administration of mONCR-171. These results suggest that mONCR-171 IT administration primarily resulted in a localized inflammatory TME.

Kinetic analysis after a single i.t. dose of mONCR-171 (3 × 10⁶ PFU) demonstrated that peak viral DNA copy numbers were achieved in injected tumors 24 hours after dose (Fig. 4). Viral DNA declined, but did not return to baseline levels, by 168 hours after dose. In contrast, viral DNA was low to undetectable in contralateral tumors, livers, and blood at all the time points tested (Fig. 4A). VP5 transcripts in the injected tumor were elevated at early time points after dosing and dropped to near baseline by 168 hours, inferring that most of the viral DNA signal detected at later time points was derived from nonviable virus (Fig. 4B). Similar to viral DNA, IL12 and FLT3LG (ECD) transgene expression within injected tumors peaked by 24 hours after dose (Fig. 4C). Endogenous CCL4 was elicited at early time points by oHSV treatment, as shown by comparison of ONCR-159 and mONCR-171 treatments, with apparent contributions from vector sources by 24 hours after dose. Little to no IL12, FLT3LG (ECD), or CCL4 was detected in contralateral tumors (Fig. 4D). In conclusion, intratumoral administration of mONCR-171 resulted in a dose-dependent increase in viral DNA, transgene expression, and inflammatory cytokines that was restricted to injected tumors.

Efficacy evaluation was carried out in immune-competent animals bearing bilateral subcutaneous syngeneic tumors to allow for the determination of both in situ and systemic (abscopal) antitumor effects. A summary of mONCR-171 single-agent efficacy studies is listed in Supplementary Table S3. A dose-escalation efficacy study in the oHSV-sensitive A20 model demonstrated a positive dose–efficacy relationship for both injected and contralateral tumors, with efficacy plateauing at a 3 × 10⁶ PFU dose. No adverse effects on animal body weight or condition were noted for all tested doses (Fig. 5A). A similar positive dose–efficacy relationship and tolerability profile was noted for the oHSV-resistant (relative to A20) MC38 tumor model, albeit higher viral doses were required to deliver equivalent antitumor effects (Supplementary Table S3). Next, ONCR-159 and mONCR-171 were tested in a panel of syngeneic tumor models that represent various degrees of sensitivity to oHSV and baseline T-cell immune infiltration (1, 31–33). As shown in Fig. 5B, and noted by others (1), the A20 tumor model was sensitive to oHSV therapy, with ONCR-159 administration resulting in a 70% and 10% response rate (RR, defined as the summation of partial and complete tumor regressions), respectively, for injected and contralateral tumors. However, antitumor efficacy of mONCR-171 was significantly superior to that of ONCR-159, resulting in 100% (P = 0.0006) and 80% (P ≤ 0.0001) RR, respectively, for injected and contralateral tumors (Fig. 5B). Potent antitumor efficacy, including a significant proportion of complete tumor regressions, was also observed in other syngeneic tumor models, including the oHSV-resistant CT26 tumors and the cold “immune desert” B16F10N1 tumor (Fig. 5B; Supplementary Table S3). Long-term assessment of animals achieving bilateral complete tumor regressions suggested that they were, in effect, cured of their tumor (Fig. 5C). Cured animals were resistant to subsequent rechallenge with the same tumor cell line, but not from challenge with an antigenically unrelated cell line (Fig. 5D; Supplementary Fig. S2). These results showed that mONCR-171 could elicit durable antitumor responses and tumor antigen–specific protective memory responses.

Intratumoral administration of mONCR-171 resulted in a significant increase in the median T-cell density for injected tumors (2,112 vs. 635 T cells/mm², P = 0.008) and contralateral tumors (1,036 vs. 552 T cells/mm², P = 0.03) compared with PBS-treated controls, and trended higher compared with that observed for ONCR-159 treatment (Fig. 6A and B). Transcriptional analysis of injected and contralateral A20 tumors revealed several significant changes in the expression of gene sets associated with CD8⁺ and CD4⁺ T cells, NK cells, and DCs with mONCR-171 treatment compared with PBS and, in some cases, with ONCR-159 (Fig. 6C). Gene sets associated with specific immune processes, such as cytotoxic activity, IFN signaling, and antigen presentation, were also upregulated with mONCR-171 treatment (Fig. 6D). Immune cell phenotyping in MC38 bilateral tumors demonstrated significantly increased counts of CD4⁺ and CD8⁺ T cells with mONCR-171 treatment, consistent with the transcriptional profiling data in A20 (Fig. 6E). mONCR-171 stimulated the functional (IFNγ-secreting) expansion of MC38 tumor antigen–specific CD8⁺ T cells from both injected and contralateral tumors (Supplementary Fig. S3). With mONCR-171 treatment, the proportion of Tregs significantly decreased, resulting in elevated CD8/Treg ratios (Fig. 6E). Similar changes in immune cell content were noted in contralateral tumors. Whereas ONCR-159 elicited immune cell–related changes compared with PBS control, mONCR-171 augmented these changes, including recruitment and activation (indicated by CD69 staining) of NK cells, CD8⁺ T cells, and CD4⁺ Th1 cells, and recruitment of antigen cross-presenting conventional (c)DC subsets. Immune subset depletion in the A20 tumor model confirmed the importance of CD8⁺ T cells and NK cells in the antitumor efficacy of mONCR-171 (Fig. 6F–H). Analysis of longitudinal TV curves suggested that the in situ antitumor efficacy of mONCR-171 was dependent upon both virus-mediated oncolysis and the presence of CD8⁺ T cells (Fig. 6F), with NK cells also playing a role (Fig. 6G). For the contralateral tumor, the abscopal activity of mONCR-171 was completely abrogated upon depletion of either CD8⁺ or NK immune cell subsets.

The antitumor efficacy of intratumorally administered mONCR-171 was tested in combination with a systemically administered anti-mouse PD-1–blocking antibody in the MC38 tumor model (Fig. 7A; Supplementary Table S4). mONCR-171 treatment resulted in a 70% RR in injected tumors and a 20% RR for contralateral tumors, both significantly different from the PBS control (P < 0.0001). As may be expected for systemic therapy, anti–PD-1 treatment resulted in similar efficacy for the right and left flank tumors, with a 20% RR for both, which was significantly (P < 0.0001) different from isotype controls. For combination therapy, efficacy in the injected tumor was primarily due to mONCR-171 (70% vs. 80% RR for mONCR-171 single-agent vs. combination treatment). In contrast, combination treatment resulted in a significantly enhanced 40% RR for contralateral tumors compared with 20% RR for either mONCR-171 (P = 0.002) or anti–PD-1 (P = 0.003) single-agent treatment. Transcriptional analyses demonstrated significantly elevated gene signature scores indicative of T-cell infiltration and CD8⁺ T-cell/NK-cell cytotoxic activity in both injected and contralateral tumors with mONCR-171 single-agent therapy, compared with PBS or isotype controls or the anti–PD-1 single-agent–treated group (Supplementary Fig. S4). Together, these results suggest that whereas mONCR-171 IT treatment was sufficient to drive efficacy for injected tumors, the abscopal efficacy of mONCR-171 was augmented by the addition of systemic anti–PD-1 treatment. These effects correlated with significant survival benefits for combination treatment over that of the single agents (Fig. 7B).

ONCR-177 is a modified recombinant oHSV designed to be a safe and efficacious therapy for the treatment of solid tumors. miRNA-dependent degradation of viral transcripts required for HSV-1 replication (ICP4, ICP27, and UL8) and neurovirulence (ICP34.5) was introduced to inhibit replication and neuropathic activity, while preserving oncolytic ability in tumor cells (10). The miRNA attenuation strategy involved identification and selection of 10 miRNAs based on high expression in healthy tissues that could be sensitive to HSV-1 infection, most notably the nervous system, but low or absent expression in malignancies. In vitro analyses confirmed the miRNA attenuation mechanisms were operational in ONCR-177 and that viral replication and transgene expression were potently suppressed by miRNA mimics and in nontransformed cell cultures compared with cancer cells.

miRNA attenuation allows ONCR-177 to retain a functional copy of the ICP34.5 gene. The physiologic function of ICP34.5 is to enable HSV-1 replication in the presence of host antiviral responses by counteracting PKR-mediated inhibition of translation, type I IFN signaling, and autophagy (9). Most oHSVs to date have deleted the ICP34.5 in order to hinder virus replication in the presence of type I IFN in healthy cells, particularly in the nervous system (8, 9). oHSVs that have retained ICP34.5 have sought to mitigate against its neuropathic effects, for example, by replacing its endogenous promoter with a tumor-specific promoter (34–36), or by attenuation of virulence due to defects in other viral genes (37). Defective IFN signaling in cancer cells is the basis for the selective replication of ICP34.5-deleted HSV-1, such as T-VEC. However, in tumors with a residual antiviral response or in the presence of type I IFN secreted by myeloid cells within the tumor stroma, ICP34.5 deletion may decrease viral replication and oncolytic activity against cancer cells. Consistent with this hypothesis, ONCR-177 retained the ability to kill cancer cells in the presence of IFNα compared with an ICP34.5-deleted oHSV.

Among 14 transgene candidates systematically screened for in vivo efficacy, 5 were selected for inclusion within ONCR-177 and mONCR-171. The individual and collective mechanisms of action of these five transgenes were anticipated to efficiently recruit key immune cell types to the tumor, including effector (CD8⁺ and CD4⁺) T cells, NK cells, and cDCs, with the intent to elicit a potent in situ immune response that would in turn bolster systemic abscopal antitumor effects.

The cytokine IL12 induces the activation, maturation, and proliferation of NK cells and effector T cells. Our study found that the intertumoral concentration of IL12 was increased following an intratumoral injection of ONCR-177 in a dose-dependent manner. CD8⁺ and CD4⁺ T cells, NK cells, and DCs were recruited to tumors injected with mONCR-171, and the activation of these immune cells stimulated the production of IFNγ (29, 38), which in turn mediated a downstream transcriptional program involved in the activation of antigen processing machinery, such as TAP-1 and MHC Class I expression, and the production of T-cell chemotactic and antiangiogenic factors, such as the chemokine CXCL10 (39). Classical DC subsets have been identified as potent APCs capable of shuttling between tumor and lymph nodes to mediate antigen cross-presentation, which is necessary to efficiently prime and activate tumor-specific antigen cytotoxic T cells (40). Consistent with previous studies, our findings suggest that gene sets associated with immune processes, such as cytotoxic activity, IFN signaling, and antigen presentation, were upregulated in tumors injected with mONCR-171. Expansion of cDCs has been shown to correlate with efficacy to ICB therapy (29). Intratumoral NK cells have been identified as the primary source of FLT3LG, which is the formative cytokine for cDCs (41). The NK–cDC axis within the TME correlates with responsiveness to anti–PD-1 therapy (38, 42). CCL4 has been shown to be important for the recruitment of CD8⁺ and CD4⁺ T cells for antiviral and antitumor responses (43–45).

Blocking the checkpoint molecules PD-1 and CTLA-4 has been clinically validated to promote antitumor immunity, primarily by efficient reactivation of CD8⁺ effector T cells. However, systemic administration as a monotherapy or in combination is associated with immune-mediated adverse events that can be severe (46, 47). Intratumoral administration of mONCR-171 was not associated with adverse events, including animal body weight or condition, suggesting a tolerable safety profile. Immune activation by oHSV expressing IL12, FLT3LG (ECD), and CCL4 is anticipated to result in compensatory upregulation of immune checkpoint pathways such as PD-1 and CTLA-4. For example, intratumoral administration of T-VEC increases the secretion of inflammatory cytokines, including IFNγ, which in turn upregulates the expression of PD-L1 (2). Activation of the PD-1/L1 pathway can be circumvented by blockade, either systemically or encoded within the virus. Consistent with this hypothesis, the enhanced antitumor effect that was observed with mONCR-171 combination therapy with anti–PD-1 treatment may be in part due to efficient release of T-cell suppression.

Numerous lines of experimental evidence demonstrate the contribution of the transgenes to the efficacy and to the proposed immune mechanisms of action of ONCR-177/mONCR-171. In vitro DC and T-cell cocultures consistently demonstrated that DCs loaded with ONCR-177 lysates were superior to those loaded with ONCR-159 lysates in their ability to activate tumor antigen–specific CD8⁺ T cells. Across in vivo efficacy studies representing a diverse set of syngeneic tumor models (1, 32, 33), mONCR-171 typically exhibited significantly greater in situ and, in particular, abscopal efficacy compared with the same dose and schedule of ONCR-159. Enhanced efficacy resulted in long-term bilateral cures that translated to robust and significantly longer survival benefits for mONCR-171–treated compared with ONCR-159–treated animals. This included the CT26 model, which has been reported to be resistant to OncoVex^mGM-CSF, an oHSV expressing mouse GM-CSF (1). These results suggest that oHSV1 therapeutic efficacy may be improved upon by expression of additional and/or alternative (other than GM-CSF) immune payloads in the TME. Efficacy of mONCR-171 correlated with local and distant intratumoral infiltration of immune cell types and processes predicted to have antitumor function. Together, these results are consistent with the proposed biological functions of the transgenes and further show that the activity of the expressed transgenes contributed to the systemic antitumor efficacy.

Immune depletion confirmed the importance of CD8⁺ T cells and NK cells in the antitumor efficacy of mONCR-171. In situ efficacy of mONCR-171 was dependent upon both virus-mediated oncolysis, the presence of CD8⁺ T cells, and to a lesser extent, NK cells. The requirement of NK cells for the abscopal efficacy of mONCR-171 was unexpected but may be related to the critical role of this innate immune cell type in the expansion and activation of motile DCs within the TME (38). For contralateral tumors, in which biodistribution studies showed that live virus was not present, abscopal activity was primarily immune-mediated. It is possible that depletion of other cell types that express CD8 or asialo GM1 (48) may in part contribute to observed response profiles. It is likely that immune cell types in addition to CD8⁺ T cells and NK cells, such as effector CD4⁺ T cells and cDCs, contribute to the in situ and/or abscopal efficacy of mONCR-171, as has been shown for other immune therapies (29, 49, 50) as well as in this report.

ONCR-177 is being developed as an intratumoral therapy. Compared with systemic dosing, intratumoral administration permits the administration of large amounts of active agent within the target (tumor) tissue. This direct dosing strategy increases the opportunity for locally efficient viral replication, oncolysis, and transgene expression, while potentially reducing systemic toxic effects. Efficient replication and transgene expression may nucleate a potent local and systemic antitumor immune response, although it could also accelerate immune-mediated viral clearance. Biodistribution studies presented here suggest that viral DNA and transgene expression were relegated primarily to the injected tumor, consistent with other published oHSV biodistribution studies (1, 51). In the FaDu xenograft (immune-deficient) model, live virus persisted in the ONCR-177–injected tumor up to the 168-hour time point. In contrast, for the A20 tumor model, a similar dose of mONCR-171 resulted in rapid loss of live virus to nearly undetectable levels by 72 to 168 hours after intratumoral injection. Robust oHSV replication in human (relative to mouse) cancer cells coupled with impaired immune-mediated clearance may explain the extended persistence of live virus and transgene expression in the injected tumor of the xenograft compared with syngeneic tumor model.

Efficient transgene expression in the tumor is considered therapeutically beneficial to optimally recruit and activate targeted immune cell types, particularly if transgenes remain largely in the injected tumor to avoid excessive immune activation that could lead to toxic effects. mONCR-171 transgene biodistribution patterns largely mirrored that of viral DNA, being primarily restricted to the injected tumor and peaking 24 to 72 hours after dose. However, IFNγ and CXCL10, which are downstream of IL12 and IFNγ, respectively, were elevated in both injected tumors and blood, consistent with the observed dose-dependent in situ and abscopal antitumor efficacy and proposed mechanism of action of mONCR-171. mONCR-171 did not elicit systemic expression of IL6 and TNFα, two cytokines implicated in cytokine release syndrome (30). Although examination of cytokine release in mice may not fully reflect or predict that which occurs in humans, collectively, these data suggest that intratumoral administration of mONCR-171 elicited a local (TME) rather than systemic inflammatory response.

In summary, ONCR-177 and its murine surrogate mONCR-171 are designed for enhanced oncolysis, tolerability, and antitumor activity. Efficacy was dependent on the activation of systemic immunity elicited by the expressed transgenes and could be further enhanced by cotreatment with an immune checkpoint inhibitor.

B.B. Haines reports other from FBL ClinWriters, LLC (editorial review and assistance with manuscript submission), ImmunXperts (conducted fee-for-service in vitro immunogenicity assays), and Sofie Denies (statistical analysis) during the conduct of the study, as well as other from Oncorus, Inc. (employee; holds company stock) outside the submitted work. P. Grzesik reports other from Oncorus, Inc. (employee) outside the submitted work. L. Kong reports other from Oncorus, Inc. (employee) outside the submitted work. E.M. Kennedy reports a patent for US 2020/0206285 A1 pending. L. Lerner reports other from Oncorus, Inc. (employee) outside the submitted work, as well as a patent for US 2020/0206285 A1 pending. C. Quéva reports personal fees and other from Oncorus, Inc. (officer, employee, and shareholder) outside the submitted work, as well as a patent for PCT/US2018/043938 pending to Oncorus, Inc. No disclosures were reported by the other authors.

B.B. Haines: Conceptualization, formal analysis, supervision, methodology, writing–original draft, project administration, writing–review and editing. A. Denslow: Formal analysis, methodology. P. Grzesik: Methodology. J.S. Lee: Methodology. T. Farkaly: Formal analysis, validation. J. Hewett: Methodology. D. Wambua: Software, formal analysis, methodology. L. Kong: Methodology. P. Behera: Methodology. J. Jacques: Formal analysis. C. Goshert: Methodology. M. Ball: Validation. A. Colthart: Validation. M.H. Finer: Conceptualization. M.W. Hayes: Methodology. S. Feau: Methodology. E.M. Kennedy: Conceptualization, software, formal analysis, validation, methodology, writing–original draft, project administration, writing–review and editing. L. Lerner: Conceptualization, formal analysis, validation, methodology, writing–original draft, project administration, writing–review and editing. C. Quéva: Conceptualization, formal analysis, supervision, writing–original draft, project administration, writing–review and editing.

This study was sponsored by Oncorus, Inc. The authors acknowledge and thank Sofie Denies for statistical analysis, Sofie Pattijn and Jana Schockaert (ImmunXperts) for in vitro immunogenicity assays, James B. Rottman (Athenaeum Pathology Consulting) for morphometric analysis of T-cell infiltration, and Michael Cristini and Jim Markert for providing G207. The authors also thank Chastity Bradley of FBL ClinWriters, LLC for editorial review of the article and article submission.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.