Intratumoral Electroporation of Plasmid Encoded IL12 and Membrane-Anchored Anti-CD3 Increases Systemic Tumor Immunity

This article is featured in Highlights of This Issue, p. 835

IL12 is a proinflammatory, pleiotropic cytokine that is widely recognized as a bridge between innate and adaptive immune immunity as antigen-presenting cells (APC) produce IL12, which activates both natural killer (NK) and T cells, in turn leading to IFNγ secretion and further IL12 production (refs. 1–4; see Table 1 for abbreviations). In addition to inducing upregulation of IFNγ, IL12 shifts differentiation of naïve CD4⁺ T cells toward the Th type 1 (Th1) phenotype (5), stimulates proliferation and activity of NK, CD8⁺, and CD4⁺ T cells (6), enhances antibody-dependent cellular cytotoxicity (ADCC; ref. 7), and downregulates regulatory T cells (Treg) and myeloid-derived suppressor cells (MDSC; refs. 8, 9). These mechanisms can all converge during an antitumor immune response, making IL12 a viable candidate for tumor immunotherapy. Unfortunately, systemic administration of IL12 has been associated with severe immune-related toxicities with a limited therapeutic index at doses considered safe for systemic delivery (10–19). Accordingly, development of an alternative IL12 delivery approach is needed.

Intratumoral delivery of IL12 plasmids (pIL12) via electroporation (EP) or IT-pIL12-EP has been demonstrated to yield localized IL12 protein, leading to tumor regression and prolonged overall survival in vivo (20–29). In addition to regression of treated tumors, IT-pIL12-EP induced regression of untreated distant tumors (22, 28, 30), inhibited growth of metastatic tumors (24, 25), and protected treated mice from tumor rechallenge (22, 24, 27, 28) indicating establishment of immunologic memory. Those preclinical results have translated to objective responses in clinical trials where electroporation of intratumoral plasmid IL12 (tavokinogene telseplasmid or tavo), yielded a 53% (10/19) disease control rate in phase I trial and 35.7% (10/28) overall response rate in a prospective phase II trial as a monotherapy for patients with advanced melanoma (31, 32). Of note, a total of 46% of patients in all cohorts with untreated lesions experienced regression of at least one of these lesions and 25% had a net regression of all untreated lesions (32). Importantly, intratumoral expression levels of IL12 and IFNγ were confirmed to be significantly elevated without detectable increase in paired serum samples (31), that translated to a favorable overall safety profile with no dose-limiting toxicities or grade 4 treatment-related adverse events (31, 32).

While this therapeutic strategy has shown encouraging safety and efficacy data in multiple clinical trials, longitudinal immune monitoring has underscored that on-treatment changes in the composition of the tumor microenvironment (TME) is fundamental to the regression of treated lesions as well as distant or untreated lesions and clinical responses. Specifically, a positive ratio of tumor-infiltrating T cells versus nonreactive, bystander, exhausted, or regulatory immune subsets is associated with productive systemic immunity and clinical responses (32, 33). Because tumors often lack processing and presentation of tumor-associated antigens and even MHC molecules, alternative strategies to activate T cells without the requirement of TCR engagement may increase the immunogenicity of the TME and compliment immunotherapies. In the current study, we tested whether a DNA-encodable intratumoral IL12 therapy would meaningfully benefit from broadening the scope of immune activation in the TME with the addition of membrane-anchored anti-CD3 single-chain variable fragment (scFv), a polyclonal stimulator of T cells. This concept of utilizing transfected tumor and stroma to produce nonspecific, pan-T-cell activation via a stimulatory anti-CD3-based immune synapse is an increasingly viable strategy in immune-oncology with growing evidence of clinical efficacy from bispecific T-cell engager (BiTE) and other similar antibody trials (34, 35).

In brief, we found that combined intratumoral electroporation of IL12 and membrane-anchored anti-CD3 plasmids can enhance cytokine production, T-cell cytotoxicity, and proliferation while limiting the suppressive capacity within the TME. These collective antitumor effects not only improve regression of treated tumors but also drive systemic immunity with control of nontreated contralateral tumors in vivo. This DNA-encodable polyclonal T-cell stimulator (membrane-anchored anti-CD3 plasmid) may represent a key addition to intratumoral IL12 therapies in the clinic.

Female C57BL/6J mice or BALB/cJ mice, 6 to 8 weeks of age (Jackson Laboratories) were housed under specific pathogen–free conditions in accordance with Institutional Animal Care and Use Committee guidelines. OVA TCR transgenic OT-I mice were a generous gift from Heat Biologics. B16-F10 melanoma cells (CRL-6475) and HEK-293 cells (CRL-1573) were purchased from the ATCC were maintained in DMEM, supplemented with 10% FBS and 1% penicillin-streptomycin at 5% CO₂ and 37°C. 4T1 murine breast cancer cells (CRL-2539) purchased from ATCC were maintained in RPMI1640 medium, supplemented with 10% FBS and 1% penicillin-streptomycin at 5% CO₂ and 37°C. Murine B16-OVA melanoma cell lines (a gift from Heat Biologics) were established by transfecting B16-F10 cells with full-length ovalbumin and selecting for stable transformants that were cultured in Iscove's modified Dulbecco's medium (IMDM), supplemented with 10% FBS, 2 mg/mL G418, and 1% penicillin-streptomycin at 5% CO₂ and 37°C. All cell lines were tested for Mycoplasma infection prior to banking and spot-checked (IMPACT II, IDEXX). In preparation for inoculation, the cells were trypsinized, washed with sterile DPBS (Thermo Fisher, 14190250), and counted using a Cellometer (Cellometer Auto 2000, Nexcelom). Depending on the experiment, between 0.25 and 1×10⁶ live cells in 100 µL of sterile DPBS were injected into the shaved flanks of a C57BL/6J or BALB/cJ mouse. Tumors were measured using a digital caliper. Treatments were initiated when the primary tumor volume reached approximately 80 mm³, 7 to 10 days after cell inoculation. Tumor volume (TV) was calculated using the formula TV = a² × b/2, where a = smallest diameter and b = perpendicular diameter.

Mice with tumors ranging from 75 to 100 mm³ (right flank) and 20 to 50 mm³ (left flank) were randomized into different treatment groups. This protocol was used as a standard model to test simultaneously for the effect on the electroporated lesion (EL) and nontreated contralateral lesion (CL). Tumor volumes were measured three times weekly. Mice were euthanized when either of the tumors reached 1,000 mm³ or the total tumor burden reached 2,000 mm³.

Patients with unresectable, stage III/IV melanomas were treated with IT-tavo-EP based on injected tavo dosage of 0.5 mg/mL followed by electroporation (six pulses, 1,500 V/cm) on days 1, 5, and 8 of 6-week cycles (NCT01502293). Tumor biopsies were collected from 13 patients at baseline prior to treatment and again at 4 weeks posttreatment induction (32). Tissue was fixed in formalin and embedded in paraffin. Eight 10-µm tissue curls were cut and deparaffinized for RNA extraction and NanoString analysis.

Plasmids encoding human and mouse pIL12P2A (IL12p35-P2A-IL12p40 in pUMVC3) have been previously described (30, 36). The sequences of variable heavy and light chains of 145-2C11 have been previously described (37). The sequence of the human anti-CD3 antibody OKT3 clone is publicly available (38, 39). Additional details are available in Supplementary Materials and Methods. All constructs were validated using restriction enzyme digests and DNA sequencing. Large-scale endotoxin-free preps were purchased from Genewiz (South Plainfield) or Aldevron (Fargo) and resuspended in sterile 0.9% saline.

Mice were anesthetized with 2.5% isoflurane for treatment. Tumors were electroporated with pUMVC3 (empty vector, EV) or pIL12P2A or p-anti-CD3-IL12P2A or p-anti-CD3 when tumor size reached approximately 80–100 mm³. In all, circular plasmid DNA, diluted to the desired concentration in sterile 0.9% saline, was injected centrally into tumors using a 0.3 mL syringe with a 28-gauge needle. Immediately afterwards, electroporation was applied using a low voltage generator optimized for gene electrotransfer (36), that delivers eight unidirectional pulses (400 V/cm, 10 ms per pulse) with two-needle electrode array. HEK293 cells or B16-F10 were transfected with indicated plasmid by using TransIT-LT1 transfection reagent (Mirus, MIR2300) according to the manufacturer's instructions.

Tumors harvested from mice were flash-frozen in liquid nitrogen. Tissues were homogenized using GentleMACS Octo Dissociator and a Mouse Tumor Dissociation Kit (Miltenyi Biotec). Total RNA was isolated using a TRIzol-chloroform extraction with RNase-free DNase I to remove the genomic DNA (Thermo Fisher Scientific). One-hundred nanograms of tissue RNA was hybridized with the mouse PanCancer IO360 Panel (digital profiling of 770 genes; NanoString Technologies, Inc.) at 65°C overnight for 16 hours. The probe set-target RNA complexes from each reaction were immobilized and processed on nCounter Cartridges using an nCounter MAX Prep Station, and transcripts were quantified on the Digital Analyzer (GEN 2). The nCounter Advanced Analysis software was used for detailed analysis.

For clinical samples, total RNA was isolated from deparaffinized tissue curls using the RecoverAll Total Nucleic Acid Kit (Thermo Fisher Scientific). 100 ng of tissue RNA was hybridized with the Human Immunology Panel (NanoString Technologies, Inc.) at 65°C overnight for 20 hours. Hybridized samples were then digitally analyzed using the nCounter SPRINT profiler. The nCounter Advanced Analysis software was used for detailed analysis.

Mice were implanted with 1×10⁶ B16-F10 cells on the right flank (primary treated lesion) and 2.5 × 10⁵ cells on the left flank (contralateral untreated lesion). Eight days postimplant, mice were randomized and primary tumors were electroporated with 50 μg pUMVC3 or 50 μg pIL12P2A plasmids. Untreated mice were used as a control. Forty-eight hours or 7 days after electroporation, both primary and contralateral tumors were collected, and single cells were dissociated by mechanical mincing and enzymatic digestion (Mouse Tumor Dissociation Kit, Miltenyi Biotec) on a GentleMACS Dissociator (Miltenyi Biotec) at 37°C. CD45⁺ cells were sorted from tumors disassociated for single-cell RNA sequencing (scRNA-seq) by using 10× Genomics. The scRNA-seq libraries were generated using Chromium Single Cell 3′ Library & Gel Bead Kit v3 (10× Genomics) according to the manufacturer's protocol. Briefly, live CD45⁺ cells were sorted by flow cytometry and loaded onto the 10× Genomics single-cell chip G. After droplet generation, samples were reverse transcribed, and the barcoded cDNA was amplified for 12 cycles. cDNA quality and concentration were assessed by Fragment Analysis (Agilent). Following fragmentation, end repair, and A-tailing, sample indexes were added during index PCR. Library quality and concentration were assessed by Fragment Analysis (Agilent) and qPCR using the KAPA Library Quantification Kit (Roche).

Libraries were pooled in equimolar amounts and sequenced using the Illumina Novaseq 6000 system according to 10× Genomics recommendations. Sequencing depth was targeted for 50,000 reads per cell. We performed demultiplexing, alignment to GRCm38 reference sequence, filtering, barcode, and UMI counting using CellRanger v2.2.0 (10× Genomics). Cells were classified into different subsets using Louvain clustering: The graph-based method is done by the igraph package (40) with a flexible number of nearest neighbors. This number is no larger than 20 and estimated by the elbow method of k-means clustering on the PCA results. The cell-type identity (i.e., CD4, CD8) of the resulting clusters was defined manually, based on both RNA and protein (cite-seq) expression. The definition of cluster identity (i.e., CD4, CD8), was performed using both protein (cite-seq) and RNA expression. Uniform Manifold Approximation and Projection (UMAP) plots with gene or protein expression of marker genes identifying different clusters are shown in Supplementary Fig. S1A. Data were analyzed and visualized using BBrowser 2 (41)

OT-I (GFP) and naïve lymphocytes were isolated and labeled with BlueTrace Tracker. 1:1 OT-I/naïve mice lymphocyte mix was adoptively transferred (intravenously) to B16-OVA tumor-bearing mice. Tumors were then electroporated with either p-anti-CD3-IL12P2A or pUMVC3 (day 0). Spleen and lymph nodes were harvested 5 days postelectroporation to analyze the adoptive transferred lymphocytes’ proliferation rates by flow cytometry.

Splenocytes isolated from naïve C57BL/6J mice were pulsed with SIINFEKL (2 µg/mL), then labeled with high-dose 1 µmol/L CFSE, nonpulsed cells were labeled with low-dose 0.1 µmol/L CFSE as control cells. Then, CFSE^hi (target) and CFSE^lo (control) cells were mixed at a 1:1 ratio. 10⁷ mixed cells were injected intravenously into B16-OVA tumor-bearing mice. One day before intravenous injection, tumors were electroporated with pUMVC3 or p-anti-CD3. Eighteen hours later, spleen and draining lymph nodes were collected and cytotoxicity was analyzed by flow cytometry. r = (% control/% target); r is the ratio of mixed cells without injection; % cytotoxicity = [1− (r0/r)] × 100.

Murine splenocytes were isolated from C57BL/6J mice. Naïve CD4⁺ T cells were purified by negative selection using a mouse naïve CD4⁺ T Cell Isolation Kit (Stemcell Technologies, 19765). CD3⁺CD25⁻ conventional T cells (Tconv) were isolated by negative and positive selection by using the Mouse T Cell Isolation Kit (Stemcell Technologies, 19851) and mouse CD25⁺ Regulatory T Cell Positive Selection Kit (Stemcell Technologies, 18782), respectively. Naïve CD4⁺ T cells were differentiated into Treg cells by incubating in activation cocktail [RP10 (RPMI1640, 10% FBS, 100 mmol/L NEAA, 2 mmol/L glutamine, 5 mmol/L HEPES, 1% penicillin-streptomycin, 55 µmol/L β-mercaptoethanol], 10 ng/mL IL2, 10 ng/mL TGFβ, CD3/28 activation beads (Thermo Fisher, 11452D) for 4 days. CD3/28 activation beads were removed by EasySep magnet before downstream use. CD3⁺CD25⁻ Tconv cells were labeled with CFSE (Thermo Fisher, C34554) according to manufacturer's instructions. CFSE-labeled CD3⁺CD25⁻ Tconv cells were cocultured with CD4⁺CD25^hi Tregs in RP10 medium with CD3/28 activation beads for 4 days. After 4 days, cells were harvested and washed with DPBS. CD3, CD4, CD8, CD25, FOXP3, IFNγ, and Granzyme B were detected by flow cytometry. In a separate experiment, induced differentiated Tregs were cocultured with transfected B16-F10 for 24 hours, followed by washing with DPBS then coculture with Tconv cells for 4 days as described previously.

Different types of T cells (murine naïve T cells, Treg cells, and human TILs) were labeled with CFSE and cocultured with transfected tumor cell lines (B16-F10 or HEK293) with or without 100 ng/mL recombined mouse IL12 or human IL12. B16-F10 cells were transfected with pUMVC3 or p-anti-CD3. HEK293 cells were transfected with pUMVC3 or pOKT3. Conditioned medium was collected at 24 and 72 hours for ELISA. Proliferation and intracellular markers were determined by flow cytometry. Naïve T cells were enriched CD3⁺ T cells (Pan T Cells Isolation Kit II from Miltenyi Biotec, 130-095-130) isolated from 8-week-old naïve C57BL/6J spleens and lymph nodes. Treg cells were induced as described above.

A punch biopsy from a patient with melanoma who had previously progressed on checkpoint therapy and had not received any intervening therapies was collected in X-VIVO 15 media and shipped overnight at 4°C. Tissue dissociation was performed by mechanical mincing and enzymatic digestion (Human Tumor Dissociation Kit, Miltenyi Biotec) on a GentleMACS Dissociator (Miltenyi Biotec) at 37°C. Cells isolated from the biopsy were cultured in X-VIVO 15 media supplemented with 10% heat-inactivated human AB serum (Sigma-Aldrich), penicillin G, and streptomycin, β-mercaptoethanol (Gibco), and 6,000 IU/mL recombinant human IL2 (Novartis). Cells were expanded for 21 days, frozen down in CryoStor 10 freezing media (Sigma-Aldrich), and stored in liquid nitrogen until use.

Conditioned medium from transfected cells or cocultured cells were harvested, spun down to remove any cells or debris, aliquoted, and stored at −80°C. IL12 was quantified using mouse IL12p70 DuoSet ELISA (DY419 R&D Systems). IFNγ was quantified using mouse IFNγ DuoSet ELISA (DY485 R&D Systems) or human IFNγ DuoSet ELISA (DY285B, R&D Systems).

Cells were surface labeled with the indicated antibodies for 20 minutes at 4°C. Intracellular staining was performed after fixation using a FOXP3/Transcription Factor Staining Buffer Set (eBioscience) according to the manufacturer's instructions. For HA tag staining, transiently transfected HEK293 cells described above were washed twice and harvested with DPBS. Cells were washed and resuspended in DPBS at a density of 10⁷/mL. Using an anti-HA, PE-labeled antibody (14904S Cell Signaling Technology) at 1:50 dilution, 1 × 10⁶ cells were stained for 30 minutes at room temperature. Cells were washed twice with DPBS + 2% FBS, resuspended in 400 μL DPBS. Flow cytometry was performed on BD LSRFortessa X20 platform and results were analyzed by FlowJo software version 10.4.2. Dead cells were excluded by LIVE/DEAD Fixable Far Red Dead Cell Stain Kit (Thermo Fisher Scientific). Antibodies are listed in Supplementary Materials and Methods (Supplementary Table S1).

Transiently transfected HEK293 cells described above were washed and harvested with DPBS. Cell pellets were then lysed with RIPA lysis and extraction buffer (Thermo Fisher Scientific) supplemented with protease and phosphatase inhibitor tablet (Roche). Homogenates were centrifuged at 14,000 rpm for 10 minutes at 4°C. Supernatant was collected and protein concentration was determined using DC Protein Assay Kit (Bio-Rad, 5000111). Murine tumors were intratumorally electroporated with p-anti-CD3. After 2 days, tumors were harvested and processed by using Mem-Per Plus Protein Extraction Kit (Thermo Fisher Scientific, 89842) to separate membrane and soluble fractions. Protein concentrations were determined using DC protein assay (Bio-Rad, 5000111). Equivalent amounts of protein lysate were separated with 4%–12% Bis-Tris gels (Thermo Fisher Scientific, NP0321) and transferred onto polyvinylidene difluoride membranes using the iBlot 2 Transfer System (Thermo Fisher Scientific). The following antibodies were used: anti-HA (Cell Signaling Technology, CSTC29F4) and donkey anti-rabbit secondary antibody conjugated to horseradish peroxidase (Jackson ImmunoResearch). SuperSignal West Pico Chemiluminescent Substrates (Thermo Fisher Scientific) were used to image blots in a FluorChemE instrument (Protein Simple).

Statistical analysis was performed with GraphPad Prism 9 (GraphPad Software). Statistical significance of differences between groups was calculated by one-way or two-way ANOVA or two-tailed t test based as appropriate. P < 0.05 was deemed as statistically significant.

To understand how intratumoral IL12 with electroporation (IT-pIL12-EP) can impact the immune composition of B16-F10 murine melanoma tumors, scRNA-seq was performed on TILs from IT-pIL12-EP–treated and -untreated (distant) lesions. While B cells and neutrophils were not significantly impacted by this therapy, IT-pIL12-EP drove brisk on-treatment CD8⁺ T-cell infiltration by day 7 in both treated and untreated distant lesions (treated EV:13.5% vs. treated pIL12P2A: 45.7%; untreated EV: 10.4% vs. untreated pIL12P2A: 32.6%; Fig. 1A). A notable pIL12-driven increase in the frequency of M1 macrophages was also observed at day 7 in both treated and untreated distant tumors (treated: EV: 2.5% vs. pIL12P2A: 32.4%; untreated: EV: 4.1% vs. pIL12P2A: 37.1%; Fig. 1A). This robust effect on tumor-associated macrophage subsets was seen at early timepoints as well. Data from 2 days posttreatment demonstrate that IT-pIL12-EP treatment altered the frequencies of these intratumoral macrophage subsets; specifically, M1 macrophages (EV: 10.7% vs. pIL12P2A: 47.4%) and M2 macrophages (EV 20.3% vs. pIL12P2A: 5.1%; Supplementary Fig. S1B). Gene expression profiling analysis within the CD8⁺ T-cell population (isolated from treated tumors) also demonstrated increased IFNγ and KLRG1 expression, indicative of increase in short-lived effector T cells (Supplementary Fig. S1C). Interestingly, not all immune subsets expanded on treatment of pIL12P2A, evident by relatively small frequencies of suppressive cells such as Treg cells (CD4⁺ Treg) and M2 macrophages that were minimized by day 7. While these observations collectively suggest that IT-pIL12-EP induces an immune-primed TME with increased TILs, clinical data highlight that this productive TME can quickly trigger adaptive resistance and lead to a suppressive TME (32).

We recently reported data from our phase II prospective trial (NCT01502293) of electroporated plasmid IL12 in patients with advanced melanoma (32). The objective responsive rate from the main study was 35.7%, with a complete response rate of 17.9%. Analysis of melanoma tumor samples collected from patients treated with IT-tavo-EP demonstrated that clinical responses were associated with treatment-related antitumor immunity within the TME, including a significantly increased IFNγ pathway score as well as evidence of activated immune subsets in the periphery (32). Additional transcriptomic-based immune profiling of TMEs illustrated enhanced immunogenicity in on-treatment lesions with increased mRNA levels of T-cell costimulatory molecules (CD28, CD74, CD160, and ICOS) and genes associated with antigen-processing/presenting machinery (AMP), such as BATF3, CCL19, CIITA, and IFNγ (Fig. 1B). As expected, these treatment-related increases were seen only in patients with clinical benefit (both responding and stable disease) but not in patients with progressive disease (Fig. 1B). Further transcriptomic analysis of intratumoral T-cell scores revealed significant on-treatment increases not only in CD8⁺ T cells and cytotoxic CD8⁺ T cells, but also in exhausted CD8⁺ T cells across all patients (Fig. 1C). Interestingly, this exhaustion signature was not seen in patients with progressive disease but was instead limited to patients with partial responses and stable disease. While these observations suggest that IT-pIL12-EP monotherapy induces an immune-primed TME with increased T-cell infiltration, the therapy also induced adaptive resistance and/or T-cell exhaustion that may blunt deeper clinical responses. Collectively, these data provide a rationale for an additional therapeutic to reinvigorate TILs while limiting potential suppression in the TME to enhance clinical responses even in patients with demonstrated clinical benefit.

Previously, we generated a bicistronic P2A-linked murine IL12 plasmid (pIL12P2A; Supplementary Fig. S2A) and used it with optimized electroporation parameters (reduced electrical field strength combined with extended pulse durations) to significantly enhance electroporation efficiency, which yielded increased systemic antitumor responses (36). Here, we build on this platform and introduce a plasmid encoding a membrane­anchored, scFv of anti-mouse CD3 (p-anti-CD3), based on the clone 145-2C11 (Supplementary Fig. S2B; ref. 37). In addition, a multicistronic plasmid-encoding membrane-anchored murine anti-CD3 and P2A-linked murine IL12 (p-anti-CD3-IL12P2A) was also created to simultaneously express both membrane-anchored murine anti-CD3 and murine IL12p70 (Supplementary Fig. S2C). To compare the mIL12p70 expression levels after electroporation with either the pIL12P2A plasmid or the multicistronic p-anti-CD3-IL12P2A plasmid in vitro, HEK293 cells were transfected with either plasmid. As expected, ELISA quantitation reported 5.8 times more mIL12p70 in the pIL12P2A transfection conditioned media than in those from p-anti-CD3-IL12P2A–transfected cells (Supplementary Fig. S2D).

Murine membrane-anchored anti-CD3 plasmid expression cell location was also quantified in vitro via transfection of HEK293 cells. Seventy-two hours posttransfection, p-anti-CD3 expression on cell surface was confirmed with FACS (Supplementary Fig. S2E) and whole-cell lysate Western blot analysis (Supplementary Fig. S2F). In addition, in vivo expression of membrane-anchored anti-CD3 was confirmed in B16-F10 tumors models with Western blots 48 hours postelectroporation (Supplementary Fig. S2I).

To evaluate potential synergistic effects between IL12 and membrane-anchored anti-CD3 scFV, B16-F10 melanoma cells, transfected with pUMVC3 or p-anti-CD3, were cocultured with lymphocytes isolated from naïve C57BL/6 mice with or without recombinant murine IL12. CD8⁺ T cells proliferated (5.8%) when cocultured with B16-F10 cells transfected with pUMVC3 [B16-F10 (EV)] while 11.4% of CD8⁺ T cells proliferated with the addition of 100 ng/mL recombinant mIL12 [B16-F10 (EV) + IL12; Fig. 2A]. In contrast, 7.2% of the CD8⁺ T cells proliferated when cocultured with p-anti-CD3-transfected B16-F10 cells [B16-F10 (αCD3)] while a striking 52% of the CD8⁺ T cells proliferated when cocultured p-anti-CD3-transfected B16-F10 cells in the presence of 100 ng/mL recombinant mIL12 [B16-F10 (αCD3) + IL12; Fig. 2A]. Moreover, a significant increase in production of effector cytokines, defined by intracellular staining of IFNγ and Granzyme B, were also detected in CD8⁺ T cells cocultured with p-anti-CD3-transfected B16-F10 in the presence of mIL12 (Fig. 2B and C). While the percentage of IFNγ⁺CD8⁺ T cells was not significantly altered when cocultured with p-anti-CD3-transfected cells alone; 0.1% in B16-F10 (EV) versus 0.05% in B16-F10 (αCD3), the combination of membrane-anchored anti-CD3 and IL12 [B16-F10 (αCD3) + IL12] led to a 7.6-fold increase in the frequency of IFNγ⁺ CD8⁺ T cells compared with B16-F10 (EV) + IL12 (13.3% vs. 1.75%, respectively; Fig. 2B). Similarly, coculture of T cells with B16-F10 (αCD3) + IL12 produced a 4.9-fold increase in Granzyme B⁺CD8⁺ T cells compared with B16-F10 (EV) + IL12 (57.0% vs. 11.7%, respectively; Fig. 2C).

To further explore the relationship between IL12 and membrane-anchored anti-CD3, effector cells (lymphocytes) and target cells (transfected B16-F10 cells) were cocultured at variable ratios. With relatively reduced number of cells expressing membrane-anchored anti-CD3 (E:T ratio of 2 to 1), 72 hours of coculture was necessary to accumulate more than 2,000 pg/mL IFNγ in the supernatant, compared with 48 hours required to reach 2,000 pg/mL IFNγ at an E:T ratio of 1, and 24 hours needed to reach 2,000 pg/mL IFNγ at a E:T ratio of 1 to 2 (Fig. 2D). These observations demonstrate a dose-response relationship driven by membrane-anchored anti-CD3. Interestingly, at a high enough density of membrane-anchored anti-CD3 scFV (4×10⁵ p-anti-CD3 transfected B16-F10 cells), the addition of IL12 did not further enhance effector function at 48 and 72 hours.

While IL12 has been shown to limit Treg differentiation and function in vivo, T-cell receptor signaling can promote immune function as well as triggering a compensatory suppressive response, which could blunt the therapeutic impact (42). To explore whether the combination of IL12 and membrane-anchored anti-CD3 could modulate Treg suppression, ex vivo–induced suppressive CD4⁺ Foxp3⁺ Treg cells were cocultured with transfected B16-F10 cells (pUMVC3 or p-anti-CD3) with or without mouse recombinant IL12. Treg cell function was confirmed with suppression of both CD8⁺ T cells and CD4⁺ T cells (Supplementary Fig. S3A and S3B). Coculture of these Tregs with IL12 alone had no effect on Treg proliferation but the addition of p-anti-CD3-transfected B16-F10 led to significant proliferation; additional proliferation was observed when combined with IL12 (Fig. 2E). An unexpected effector profile developed as the restimulated Tregs demonstrated upregulated intracellular expression of IFNγ and Granzyme B when exposed to both mIL12 and membrane-anchored anti-CD3 (Fig. 2F and G) accompanied by rapid and sustained secretion of IFNγ at 24 and 72 hours post coculture (Fig. 2H). Furthermore, these restimulated Tregs also lost functional immunosuppression as initial exposure to mIL12 and membrane-anchored anti-CD3 significantly reduced their ability to inhibit CD8⁺ T cells and CD4⁺ T cells (Fig. 2I).

Collectively, those observations suggest the IL12 and membrane-anchored anti-CD3 combination enhances T-cell proliferation and IFNγ secretion in different types of T cells, even altering fundamental phenotypes of Tregs and exhausted T cells which may reshape the composition and immunogenicity of the tumor microenvironment.

The impact of intratumoral membrane-anchored anti-CD3 on T-cell responses was measured using an in vivo murine adoptive T-cell transfer model to interrogate polyclonal T-cell proliferation and antigen-specific cytotoxic T-cell activity. Briefly, mice were inoculated with B16-OVA cells followed by electroporation of tumors with pUMVC3 or p-anti-CD3 before adoptively transferring GFP-labeled OT-I lymphocytes and blue trace–labeled naïve lymphocytes at a 1:1 mixture (Fig. 3A). Analysis of the draining lymph nodes demonstrated that all of the OT-I lymphocytes proliferated regardless of treatment, while a significant increase in proliferation of naïve polyclonal T cells was seen with membrane-anchored anti-CD3 (69.6% vs. 50.8%; Fig. 3A). To further elucidate the specific mechanisms of membrane-anchored anti-CD3 in vivo, TILs from B16-F10-OVA tumors treated with pUMVC3 or p-anti-CD3 were analyzed 5 days postelectroporation. Membrane-anchored anti-CD3 significantly increased the percentage of CD8⁺ T cells within the TIL compartment while having no effect on CD4⁺ T cells, discernibly increasing the percentage of OVA antigen-specific CD8⁺ T cells compared with pUMVC3 treatment control (Fig. 3B).

To assess antigen-specific cytotoxic T-cell activity in vivo, cytotoxic T-cell killing assays were conducted (Fig. 3C). Target cells–lymphocytes isolated and pooled from spleen and lymph node from naïve OT-I mice–were pulsed with OVA peptide (257-264 aa, SIINFEKL) and labeled with 1 µmol/L of CFSE (high), or left unpulsed and labeled with 0.1 µmol/L CFSE (low). One day after electroporation of established B16-OVA tumors with either pUMVC3 control or p-anti-CD3 plasmids, the two differentially CFSE-labeled groups of target cells were adoptively transferred as a 1:1 mixture into these tumor-bearing mice. Electroporation of p-anti-CD3 significantly enhanced antigen-specific killing, indicated by preferential loss of SIINFEKL⁺ CFSE^high target cells. This effect was seen both in the draining lymph node, and systemically as measured in the spleen (Fig. 3D).

To evaluate the contribution of the membrane-anchored anti-CD3 to the systemic antitumor immunity driven by IT-pIL12-EP, we first had to normalize the intratumoral mIL12p70 expression in vivo to account for the bias in mIL12p70 expression from the pIL12P2A plasmid over the expression from the multicistronic p-anti-CD3-IL12P2A plasmid as observed in Supplementary Fig. S2D. Electroporation was conducted with 10 or 100 μg pIL12P2A, or 100 μg p-anti-CD3-IL12P2A in B16-F10 tumors. Forty-eight hours postelectroporation, mIL12p70 in whole tumor lysates were quantified by ELISA, demonstrating that electroporation of 10 μg pIL12P2A produced an equivalent level of mIL12p70 as electroporation of 100 μg p-anti-CD3-IL12P2A (Fig. 4A).

A B16-F10 melanoma contralateral tumor model was used to measure tumor regression and associated immune signatures. Specifically, this bilateral model uses two tumors (one per flank) to measure regression of both the treated primary tumor and a distant nontreated contralateral tumor. While both pIL12P2A and p-anti-CD3-IL12P2A demonstrated potent antitumor effects on the treated tumors, the addition of anti-CD3 to the IL12 (p-anti-CD3-IL12P2A) demonstrated an additional systemic antitumor effect on the nontreated lesion (Fig. 4B). Similar results were seen in this model using sequential administration of plasmids encoding IL12, followed by anti-CD3 (Supplementary Fig. S4).

An additional tumor model was used to understand whether systemic immunity driven from this local therapy could lead to regression of distant untreated visceral tumors. This robust in vivo model uses transplantation of 4T1 breast cancer cells in the lower flank of mice that reproducibly metastasizes to the lung. Similar to the bilateral B16-F10 melanoma model above, electroporation was performed on the subcutaneous tumors with pIL12P2A or p-anti-CD3-IL12P2A and mice were sacrificed and metastatic tumor nodules on lungs were counted when the subcutaneous tumor volumes reached 1,000 mm³. The combinatorial effect of p-anti-CD3-IL12P2A demonstrated significantly augmented antitumor effect on the primary, treated tumor while also leading to significantly reduced number of metastatic nodules in the lung (Fig. 4C).

Beyond measuring the direct systemic antitumor effect via regression of untreated tumors in the B16-F10 and 4T1 models, blood from these treated mice was collected on days 5 or 6 to measure treatment-related changes in the frequency of peripheral CD8⁺ T-cell subsets. The addition of membrane-anchored anti-CD3 drove a significant increase in the number of CD3⁺ T cells, CD8⁺ T cells, and activated CXCR3⁺CD8⁺ T cells in the periphery of mice bearing B16-F10 and 4T1 tumors (Fig. 4D and E). Previously, we demonstrated short lived effector cells (SLEC), defined as KLRG1^hi CD8⁺ effector T cells, to be upregulated with IT-pIL12-EP (30). Here, the addition of membrane-anchored anti-CD3 promoted significant increase in the number of SLECs while also increasing the number of effector T cells in peripheral blood when compared with IL12 alone in both tumor models (Fig. 4D and E).

We have demonstrated that IT-pIL12-EP can rapidly increase patient TILs with only one cycle of treatment while preclinical studies highlighted how rapidly IT-pIL12-EP could drive proimmune changes in gene expression 7 days after EP (30, 32). To identify the early contribution of membrane-anchored anti-CD3 on the immune-related transcriptomic pathways of TILs, tumors were electroporated with normalized doses of pIL12P2A or p-anti-CD3­IL12P2A (see Fig. 4A), collected at 48 hours postelectroporation and assessed with NanoString's PanCancer IO360 panel. p-anti-CD3-IL12P2A-treated tumors demonstrated a significant enhancement of not only IL12­related T-cell phenotypes, such as cytotoxic or Th1-based cell types, but also expression pathways associated with productive antitumor responses such as MHC-I/II antigen processing/presentation, TCR signaling, and costimulation (Fig. 5A and B). No significant changes were detected in Th2, Th17, Tfh, and Treg transcriptomic signatures (Supplementary Fig. S5).

While human studies in a clinical setting will ultimately be required to prove the utility of this combination therapy, a plasmid encoding human anti-CD3 (clone: OKT3) scFv (pOKT3) was used to interrogate the clinical utility of membrane-anchored human anti-CD3 (Fig. 6A). TILs isolated from a patient with melanoma actively progressing on immune checkpoint inhibitor therapy provided an opportunity to explore whether coculture of cells expressing IL12 and membrane-anchored anti-CD3 could be used to restore immune function of these likely tolerized lymphocytes. Patient T cells were cocultured ex vivo with HEK293 cells, which were transfected with either pUMVC3 or membrane-anchored human anti-CD3 (pOKT3) and cultured with or without human recombinant IL12. pOKT3 was sufficient to restimulate the patient TILs as compared with IL12, which had no discernible impact on proliferation (Fig. 6B). Similarly, pOKT3, but not IL12, stimulated secretion of IFNγ from patient TILs. However, the combination of pOKT3 and IL12 demonstrated a significant increase in IFNγ secretion compared with pOKT3 alone (Fig. 6C). Additional signs of reactivation of these tolerized TILs could be seen with PD-1 expression within the CD8⁺ T-cell population when exposed to membrane-anchored human anti-CD3 and especially with the combination of membrane-anchored human anti-CD3 and recombinant hIL12 (Fig. 6D). Collectively, these observations suggest that combination treatment of membrane-anchored anti-CD3 and IL12 may have the potential to restore or further enhance TIL function in patients.

The final assessment of our prospective phase II clinical trial evaluating intratumoral electroporation of plasmid IL12 in advanced, unresectable melanoma (NCT 01502293) recently provided encouraging evidence of clinical efficacy and safety of IT-pIL12-EP with 35.7% overall response rate without any grade 4 adverse events (32). In addition, IT-pIL12-EP triggered abscopal responses with regression of at least one untreated lesion in 46% of the patients, demonstrating that this intratumoral monotherapy can yield systemic antitumor immunity (32). While induction of adaptive resistance, particularly the PD-1/PD-L1 axis, can be an immunological byproduct of this IL12-driven therapy, combination of anti-PD-1 checkpoint blockade can help mitigate its effect (32). The addition of a therapeutic capable of broadening the immunologic activity while limiting the suppression within the TME has provided an opportunity to enhance this intratumoral IL12 therapy primary to immune checkpoint blockade. IT-pIL12-EP has previously been shown to widely inflame treated tumors regardless of response (31, 32). Relatedly, accumulation of TILs were noted in stable disease patients (Fig. 1C), providing a clinical rationale to introduce membrane-anchored anti-CD3 in the TME to increase the immunogenicity via broad T-cell activation to yield systemic immunity, durable responses, and ultimately better clinical outcomes. The DNA-encodable IL12 therapeutic was modified to include a membrane-anchored polyclonal T-cell stimulator, anti-CD3, and evaluated as an intratumoral therapeutic strategy.

p-anti-CD3-IL12P2A demonstrated a variety of promising antitumor effects in the present preclinical study. Notably, synergistic upregulation of T-cell proliferation and markers of cytotoxic activity were observed compared with IL12 or membrane-anchored anti-CD3 alone (Fig. 2A–C). In addition, membrane-anchored anti-CD3 demonstrated a dose-response relationship with IFNγ secretion (Fig. 2D), which is foundational to the IL12/IFNγ feed-forward loop and critical to efficacy of this therapy (32, 43). Of note, Treg proliferation was also upregulated by combination treatment of IL12 and membrane-anchored anti-CD3 in vitro, which was an anticipated concern of this approach as anti-CD3 treatment has been previously noted by others to promote proliferation of Tregs while not altering their capacity to suppress effector T cells in vivo (44, 45). Surprisingly, Tregs cocultured with tumors expressing membrane-anchored anti-CD3 and recombinant IL12 demonstrated not only increased IFNγ secretion in vitro but also a reduced ability to suppress the proliferation of CD4⁺ or CD8⁺ T effector cells (Fig. 2F–I), suggesting diminished Treg function when sensitized with membrane-anchored anti-CD3. While enhanced cytolytic function is associated with this therapy, inclusion of additional atypical antitumor T-cell subsets such as suppressive T cells may be critical in producing a deeper and more durable clinical response as TME becomes reshaped via engagement with the membrane-anchored anti-CD3.

In preclinical in vivo models, tumor electroporation of p-anti-CD3-IL12P2A treatment reflected in vitro findings as the combined therapeutic approach induced polyclonal T-cell expansion, productive innate and adaptive transcriptomic immune signatures, and enhanced antigen-specific tumor cell cytotoxicity leading to augmented antitumor effects in B16-F10 and 4T1 models compared with IL12 alone (Figs. 3 and 4). Interestingly, tumor electroporation of p-anti-CD3-IL12P2A therapy demonstrated significant antitumor effects on subcutaneous primary (electroporated) 4T1 tumors with a notable control of lung metastasis compared with IL12 alone (Fig. 4C). While this study was not designed to determine whether p-anti-CD3-IL12P2A can enhance the clinical efficacy relative to the current intratumoral IL12 platform, ex vivo coculture experiments with dysfunctional TILs grown from a patient with metastatic melanoma suggest that the benefits of IL12 were greatly amplified with the addition of membrane-anchored anti-CD3 (Fig. 6). The complex immune contexture of the TME currently cannot be recapitulated with an ex vivo coculture; however, these results suggest that this combination could drive additional lymphocytic function from a tolerized and potentially immunosuppressed TME.

Collectively, observations from this study suggest that the combination of IL12 and membrane-anchored anti-CD3 activates more than naïve or effector T cells by mobilizing a diverse group of TILs toward an effective antitumor response, capable of broadly reshaping the TME to yield systemic immunity.

M. Han reports personal fees and other support from OncoSec Medical Inc outside the submitted work; in addition, M. Han has a patent for application no. PCT/US2019/063590 pending to OncoSec Medical Inc. J.Y. Lee reports other support from OncoSec Medical Inc outside the submitted work. E. Browning reports other support from Oncosec outside the submitted work; in addition, E. Browning has a patent for PCT/US2019/063590 pending to Oncosec. R. Hermiz reports personal fees from OncoSec during the conduct of the study; personal fees from OncoSec outside the submitted work. A.S. Rolig reports grants from OncoSec during the conduct of the study. W.L. Redmond reports grants from OncoSec during the conduct of the study; grants from Bristol Myers Squibb, Nektar Therapeutics, GlaxoSmithKline, Inhibrx, Galectin Therapeutics, MiNA Therapeutics, Veana Therapeutics, Turn Bio, Galecto, Shimadzu, and grants from Calibr outside the submitted work; in addition, W.L. Redmond has a patent for US20180236084A1 issued and licensed to Galectin Therapeutics; and is a scientific advisory board member at Vesselon and Medicenna. A.P. Algazi reports personal fees and other support from OncoSec Medical, Inc. during the conduct of the study; personal fees, non-financial support, and other support from OncoSec Medical, Inc., Sensei Biosciences; and other support from Valitor Biosciences outside the submitted work. A.I. Daud reports grants from Oncosec, Merck, BMS, Pfizer, Incyte, Novartis, and grants from Roche during the conduct of the study; in addition, A.I. Daud has a patent for Oncosec issued. D.A. Canton reports other support from Oncosec Medical Incorporated outside the submitted work; in addition, D.A. Canton has a patent for PCT/US2019/063590 pending. C.G. Twitty reports nonfinancial support from OncoSec outside the submitted work; in addition, C.G. Twitty has a patent for PCT/US2019/063590 issued. No disclosures were reported by the other authors.

M. Han: Conceptualization, data curation, supervision, writing–original draft. B. Nguyen: Investigation, methodology. J.Y. Lee: Investigation, methodology, writing–review and editing. E. Browning: Investigation, writing–review and editing. J. Zhang: Investigation, writing–review and editing. A. Mukhopadhyay: Investigation, writing–review and editing. R. Gujar: Investigation, writing–review and editing. J. Salazar: Investigation. R. Hermiz: Investigation. L. Svenson: Investigation. A.S. Rolig: Investigation, methodology, writing–review and editing. W.L. Redmond: Supervision, writing–review and editing. A.P. Algazi: Supervision, writing–review and editing. A.I. Daud: Supervision, writing–review and editing. D.A. Canton: Conceptualization, supervision, writing–review and editing. C.G. Twitty: Conceptualization, supervision, writing–original draft, writing–review and editing.

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.