Immune response and long-term clinical outcome in advanced melanoma patients vaccinated with tumor-mRNA-transfected dendritic cells

Abstract

The most effective anticancer immune responses are probably directed against patient-specific neoantigens. We have developed a melanoma vaccine targeting this individual mutanome based on dendritic cells (DCs) loaded with autologous tumor-mRNA. Here, we report a phase I/II trial evaluating toxicity, immune response and clinical outcome in 31 metastatic melanoma patients. The first cohort (n = 22) received the vaccine without any adjuvant; the next cohort (n = 9) received adjuvant IL2. Each subject received four weekly intranodal or intradermal injections, followed by optional monthly vaccines. Immune response was evaluated by delayed-type hypersensitivity (DTH), T cell proliferation and cytokine assays. Data were collected for 10 y after inclusion of the last patient. No serious adverse events were detected. In the intention-to-treat-cohort, we demonstrated significantly superior survival compared to matched controls from a benchmark meta-analysis (1 y survival 43% vs. 24%, 2 y 23% vs. 6.6%). A tumor-specific immune response was demonstrated in 16/31 patients. The response rate was higher after intradermal than intranodal vaccination (80% vs. 38%). Immune responders had improved survival compared to non-responders (median 14 mo vs. 6 mo; p = 0.030), and all eight patients surviving >20 mo were immune responders. In addition to the tumor-specific response, most patients developed a response against autologous DC antigens. The cytokine profile was polyfunctional and did not follow a Th1/Th2 dichotomy. We conclude that the favorable safety profile and evidence of a possible survival benefit warrant further studies of the RNA/DC vaccine. The vaccine appears insufficient as monotherapy, but there is a strong rationale for combination with checkpoint modulators.

Keywords: Cancer vaccine; T cell; clinical outcome; dendritic cell; mRNA; melanoma; survival.

Figure 1.

T-cell responses specific for antigens encoded by the transfected mRNA (cohort DCM-2). T-cell proliferation or INFγ ELISPOT after stimulation with tDCs and nDC controls. A T-cell response was considered tDC-specific if the response to tDCs was significantly (p <0.05; ANOVA/SNK) higher than in the controls (T+nDC and T cells only). For all five patients shown, a tDC-specific response was demonstrated in post-vaccination samples (week 6 onwards). In patient M108, a tDC-specific response was observed also prior to vaccination (week 0). The assays in Fig. 1 were performed on T cells pre-stimulated once in vitro with tDCs, except for the assays on follow-up samples from patient M109 (week 14–34), which were performed on freshly thawed T cells. T cell only background counts have been subtracted. Columns, mean cpm (proliferation assays) or mean spots/well (ELISPOT). Error bars, SEM.

Figure 2.

T-cell responses against antigens not encoded by the transfected mRNA (cohort DCM-2). PBMCs were obtained at study entry (baseline), week 6, month 3 and at time of later booster vaccinations. All time points measured for each patient are included in the figure. The figure shows T-cell proliferation in freshly thawed PBMCs upon stimulation with non-transfected autologous DCs (nDCs). Columns represent mean cpm from triplicates.

Figure 3.

Intranodal versus intradermal vaccination. The patients in the melanoma and prostate cancer DC-vaccination trials were allocated to intradermal (i.d.) or intranodal (i.n.) vaccine injection. The figure shows the percentage of immune responders and non-responders in the i.d. and i.n. vaccination groups, for the melanoma patients (A) and for the melanoma and prostate cancer trials taken together (B). The response rates after i.d. and i.n. vaccination were compared by use of Fishers exact test, two-sided p-values are displayed.

Figure 4.

Tumor response in patient M109. (A) Total diameter of all measurable lesions (corresponding to RECIST) at time points for CT evaluation. Three lesions were >10 mm at study entry and thus defined as target lesions. (B) Response of individual lesions at different time points (mo = months). The six measurable pre-study lesions all completely regressed. The three target lesions (>10 mm at study entry) are shown in the top half of the diagram. A new pararenal lesion (NL) appeared at month 9. (C) CT scans before and after vaccination, showing complete regression of the three target lesions defined at study entry. The displayed post-vaccination scans are from month 3 (subpleural lesion) or month 12 (lymph node I and II).

Figure 5.

(A) Survival in study patients compared to benchmark meta-analysis. Blue line: Overall survival for all enrolled study patients calculated from study entry. Dashed lines represent the 90% confidence interval (CI). Green line: Predicted survival for our patients (fixed estimate), calculated from Korn's meta-analysis. The calculation corrects for prognostic factors (performance status, site of metastases etc.) and was performed as recommended by Korn et al. The analysis suggested a difference in favor of the RNA/DC-vaccine treatment (p = 0.036). (B) Survival versus immune response. Kaplan–Meier analysis comparing overall survival between patients with/without a vaccine-specific T-cell response. The analysis suggested a difference in favor of the immune responders (log rank test; p = 0.030).