Phase II Trial of IL-12 Plasmid Transfection and PD-1 Blockade in Immunologically Quiescent Melanoma

Abstract

Purpose: Tumors with low frequencies of checkpoint positive tumor-infiltrating lymphocytes (cpTIL) have a low likelihood of response to PD-1 blockade. We conducted a prospective multicenter phase II trial of intratumoral plasmid IL-12 (tavokinogene telseplasmid; "tavo") electroporation combined with pembrolizumab in patients with advanced melanoma with low frequencies of checkpoint positive cytotoxic lymphocytes (cpCTL).

Patients and methods: Tavo was administered intratumorally days 1, 5, and 8 every 6 weeks while pembrolizumab (200 mg, i.v.) was administered every 3 weeks. The primary endpoint was objective response rate (ORR) by RECIST, secondary endpoints included duration of response, overall survival and progression-free survival. Toxicity was evaluated by the CTCAE v4. Extensive correlative analysis was done.

Results: The combination of tavo and pembrolizumab was well tolerated with adverse events similar to those previously reported with pembrolizumab alone. Patients had a 41% ORR (n = 22, RECIST 1.1) with 36% complete responses. Correlative analysis showed that the combination enhanced immune infiltration and sustained the IL-12/IFNγ feed-forward cycle, driving intratumoral cross-presenting dendritic cell subsets with increased TILs, emerging T cell receptor clones and, ultimately, systemic cellular immune responses.

Conclusions: The combination of tavo and pembrolizumab was associated with a higher than expected response rate in this poorly immunogenic population. No new or unexpected toxicities were observed. Correlative analysis showed T cell infiltration with enhanced immunity paralleling the clinical activity in low cpCTL tumors.

Figure 1.

A, Study consort diagram. B, Combination study schema. C, Patient demographics (n = 22) at baseline and study schema for combination therapy including prior treatment for each patient and subsequent responses to i.t.-tavo-EP with pembrolizumab on this study. D, Intratumoral CD8⁺ TILs from included patients were < 25% cpCTL (PD-1hiCTLA-4hi).

Figure 2.

A, Predicted responses and peCTL frequencies by the flow cytometry assay as well as the best observed clinical and RECIST v1.1 responses for each patient (blue, response; yellow, nonresponse). B, Waterfall plot of best response by clinical assessment. Clinical stage depicted on each bar progression-free survival (C) and overall survival (D) were determined using the Kaplan–Meier method. The median PFS was 5.6 months but the median OS was not reached at a median follow-up of 20.1 months. E, Individual patient RECIST responses over time (dashed lines represent patients with progression of nontarget lesions). F–I, A nonresponding anti-CTLA-4, anti-PD1 antibody refractory patient had a flare (F–G) of multiple unresectable in transit lesions followed by (H) an objective clinical response. I, A biopsy at 23 month shows only a siderotic scar with no residual melanoma. §, Patient did not have RECIST measurable disease at baseline.

Figure 3.

Increasing TILs and adaptive resistance in i.t.-tavo-EP–treated lesions. FFPE tumor biopsies and PBMCs collected at pre- and posttreatment time points were stained for immune cell markers by multispectral mIHC and flow cytometry. A, An increase of intratumoral CD8⁺ T cell density in FFPE biopsies of patients by mIHC analysis was observed after a single cycle of combination therapy. B and C, After a single cycle of intratumoral plasmid IL-12 with electroporation, intratumoral PD-L1 expression (B) measured by IHC was significantly increased while (C) transcriptional analysis demonstrated a significant posttreatment expression of PD-L1 in responding patients. D, mIHC images of a patient with progressive disease (left) shows increased CD163⁺ TILs and PD-L1⁺ cells whereas a patient undergoing a complete response (right) showed a brisk increase in CD8⁺ TILs over one course of combination therapy. E, mIHC spatial analysis showed a significant increase in the number of FoxP3⁺ cells < 15 μm from CD8⁺ T cells in patients after a single cycle of combination treatment. F, Posttreatment, nonresponders showed a significantly higher amount of FoxP3⁺ cells < 15 μm from CD8⁺ T cells when compared with responders. G, Patients responding to combination treatment showed a significant increase in the ratio of CD8:M2 macrophages (CD163⁺) after a single cycle of treatment. Pre, pretreatment; post, posttreatment, cycle 2 day 1; NR, nonresponder (stable disease and progressive disease); R, responder (complete response and partial response). Significance calculated by Mann–Whitney for all unpaired samples [C (n = 6 (NR), n = 7 (R)), E (n = 18 (pre), n = 10 (post), F (n = 2 (NR), n = 9 (R)) and g (n = 5 (NR), n = 7 (R))] and by Wilcoxon matched-pairs signed rank test for all paired samples [A (n = 10) and B (n = 8)]. ns, not significant; *, P < 0.05; **, P < 0.01; N.I., noninjected lesions were noted (C, E, and F only).

Figure 4.

Combination therapy enhances intratumoral expression of immune-based gene sets. Transcriptional analysis of biopsies collected at screen and post-treatment. A, Volcano plot of both nonresponding and responding patients highlight changes in Th1/antitumor-associated genes. On-treatment (after 1 cycle of treatment) increases in gene expression related to innate immunity (B), adaptive resistance and immune activation (C), immune trafficking and costimulation were noted in patients receiving combination therapy (D). Pre, pretreatment; post, posttreatment, (cycle 2 day 1). *, P = 0.05; **, P = 0.01; ***, P = 0.001 by Mann–Whitney t test (n = 13). Note, 1 untreated sample (after 1 cycle of treatment) was used in this analysis.

Figure 5.

Treatment-related increases in intratumoral CD8⁺ T cell infiltration, adaptive resistance, and peripheral T cells. A An increase in lymphocyte infiltration was observed in both injected (EP) and noninjected (non-EP) tumors of patients regardless of response while the treatment-related ratio of CD8⁺ T cells:CD163⁺ M2 macrophages was increased in the responding patient. B, A significant increase was noted in unique intratumoral TCR clones emerging on treatment. C and D, Combination therapy resulted in a significantly increased percentage of proliferating CD8⁺ effector memory RA⁺ T cells and PD-1⁺ T cells in the blood (PBMC) posttreatment. E, Transcriptional analysis demonstrated a significant increase in the treatment-related correlation between intratumoral IRF1 and PD-L1 expression. NR, nonresponder; R, responder; pre, pretreatment; post, posttreatment, cycle 2 day 1; EP lesion, injected and electroporated lesion; non-EP lesion, noninjected and nonelectroporated lesion. Significance calculated by Wilcoxon matched-pairs signed rank test n = 10 (B), n = 13 (C and D); Pearson correlation analysis with D’Agostino and Pearson normality test (E) n = 16 (pre), n = 17 (post; ns, not significant; *, P < 0.05; **, P < 0.01; N.I., noninjected lesions were noted (B and E only).