Enhanced local and systemic anti-melanoma CD8+ T cell responses after memory T cell-based adoptive immunotherapy in mice

Abstract

Adoptive cell transfer (ACT) melanoma immunotherapy typically employs acutely activated effector CD8+ T cells for their ability to rapidly recognize and clear antigen. We have previously observed that effector CD8+ T cells are highly susceptible to melanoma-induced suppression, whereas memory CD8+ T cells are not. Although memory T cells have been presumed to be potentially advantageous for ACT, the kinetics of local and systemic T cell responses after effector and memory ACT have not been compared. B16F10 melanoma cells stably transfected to express very low levels of the lymphocytic choriomeningitis virus (LCMV) peptide GP33 (B16GP33) were inoculated into syngeneic C57BL/6 mice. Equal numbers of bona fide naïve, effector, or memory phenotype GP33-specific CD8+ T cells were adoptively transferred into mice 1 day after B16GP33 inoculation. The efficacy of ACT immunotherapy was kinetically assessed using serial tumor measurements and flow cytometric analyses of local and systemic CD8+ T cell responses. Control of B16GP33 tumor growth, persistence of adoptively transferred CD8+ cells, intratumoral infiltration of CD8+ T cells, and systemic CD8+ T cell responsiveness to GP33 were strongest after ACT of memory CD8+ T cells. Following surgical tumor resection and melanoma tumor challenge, only mice receiving memory T cell-based ACT immunotherapy exhibited durable tumor-specific immunity. These findings demonstrate how the use of non-expanded memory CD8+ T cells may enhance ACT immunotherapeutic efficacy.

Keywords: Adoptive transfer; Effector; Immunotherapy; Melanoma; Memory; T cell.

Fig. 1

ACT of memory CD8+ T cells results in optimal control of melanoma tumor growth. Naïve, effector, and memory Ly5.1+ CD8+ GP33-specific T cells are harvested and characterized by flow cytometry for: a TCR specificity for GP33 by staining with MHC tetramer loaded with GP33 peptide (gated on CD8+ lymphocytes) b CD62L expression, and c CD127 and KLRG1 expression (gated on CD8+ GP33-specific cells). d Serial B16GP33 tumor measurements in mice treated with no immunotherapy or ACT of 10⁵ CD8+ GP33-specific naïve, effector, or memory T cells one day after tumor inoculation demonstrated maximal inhibition of tumor growth after memory T cell ACT. e Memory T cells remained more effective than effector T cells when ACT is performed 7 days after tumor inoculation. These experiments are performed 3–4 times with similar results, with four mice per group (*p < 0.05 vs. no immunotherapy; ^† p < 0.05 vs. naïve T cell immunotherapy; ^‖ p < 0.05 vs. effector T cell immunotherapy)

Fig. 2

Memory T cell subsets are comparable in their ability to inhibit melanoma tumor growth. Memory CD8+ GP33-specific T cells are separated into central memory (CD62L+) and effector memory (CD62L−) subsets by magnetic bead column separation and then characterized by FACS analysis based on: a TCR specificity for GP33 (gated on CD8+ lymphocytes) and b CD62L expression (gated on CD8+ GP33-specific cells). c Serial B16GP33 tumor measurements in mice treated with ACT of 10⁵ CD8+ GP33-specific unsorted, central, or effector memory T cells demonstrated no statistically significant differences in tumor control. This experiment is performed three times with similar results, with four mice per group (*p < 0.05 vs. unsorted memory T cell immunotherapy)

Fig. 3

Effector and memory CD8+ T cells are similar in their ability to inhibit melanoma proliferation in vitro. 2 × 10⁵ B16F10 or B16GP33 melanoma cells are co-cultured with CD8+ GP33-specific effector or memory T cells at various effector/target ratios for 24 h. MTT assays measuring the viability of adherent melanoma cells demonstrated comparable dose-dependent inhibition of B16GP33 melanoma cell viability; non-specific inhibition of parental B16F10 melanoma cells (which do not express GP33) is not observed. This experiment is performed four times with similar results, with four wells per group

Fig. 4

Memory T cell ACT results in a stronger local T cell response. As before, no immunotherapy or ACT of 10⁵ Ly5.1+/CD8 + GP33-specific effector or memory T cells is performed on day one after B16GP33 tumor inoculation in Ly5.2+/C57BL/6 mice. a FACS analysis after tumor inoculation showed higher CD8+ T cell infiltration on day 4 after tumor inoculation in mice receiving ACT, and higher CD8+ T cell infiltration on day 14 in mice receiving memory T cell ACT. b Representative hematoxylin and eosin-stained sections of tumors resected on day 18 also demonstrated higher numbers of infiltrating lymphocytes after memory T cell ACT. c In contrast to effector T cell ACT, exogenously derived Ly5.1+/CD8+ T cell infiltration increased over time after memory T cell ACT. d No significant differences in endogenously derived Ly5.2+/CD8+ T cell infiltration is observed between effector and memory T cell ACT. These experiments are performed three times with similar results with four mice per group (*p < 0.05 vs. no immunotherapy; ^† p < 0.05 vs. effector T cell immunotherapy)

Fig. 5

Memory T cell ACT results in a stronger systemic T cell response. In the same experiments described in Fig. 4, splenocytes are also harvested and analyzed by flow cytometry. a On days 4 and 14, significantly higher numbers of exogenously derived Ly5.1+/CD8+ T cells are measured in the spleens of mice treated with memory T cell ACT (*p < 0.05 vs. no immunotherapy; ^† p < 0.05 vs. effector T cell immunotherapy). b Splenocyte reactivity to GP33 on day 14 (as indicated by GP33 peptide-stimulated intracellular levels of IFNγ) is only observed in mice treated with memory T cell ACT. This experiment is performed three times with similar results with three mice per group (*p < 0.05 vs. no peptide)

Fig. 6

Neoadjuvant memory T cell ACT promotes durable tumor antigen-specific immunity. Ly5.2+/C57BL/6 mice are inoculated with 10⁶ B16GP33 cells and then treated with ACT of 5 × 10⁵ Ly5.1+/CD8+ GP33-specific effector or memory T cells on day 7. B16GP33 tumors are then surgically resected on day 15. a Effector and memory T cell ACT both resulted in inhibition of established B16GP33 tumors. b Whereas contralateral challenge B16GP33 tumors inoculated on day 25 grew exponentially in mice that had received effector T cell ACT, no growth of challenge B16GP33 tumors is seen in mice that had received memory T cells. c In contrast, parental B16F10 melanoma tumors that did not express GP33 grew exponentially in B16GP33-immune mice. These experiments are performed two times with similar results with four mice per group (*p < 0.05 between groups)