Local delivery of CpG-B and GM-CSF induces concerted activation of effector and regulatory T cells in the human melanoma sentinel lymph node

Abstract

Impaired immune effector functions in the melanoma sentinel lymph node (SLN) may allow for early metastatic events. In an effort to determine the optimal way to strengthen immune defenses, 28 clinical stage I-II melanoma patients were randomized in a 3-arm Phase II study to receive, prior to excision and sampling of the SLN, i.d. injections of saline or low-dose CpG-B (CpG), alone or combined with GM-CSF (GM), around the melanoma excision site. We previously described the combined administration of these DC-targeting agents to result in activation and recruitment of potentially cross-presenting BDCA3(+) DCs to the SLN. In this report we describe the effects on effector and regulatory T and NK cell subsets. Local low-dose CpG administration resulted in lower CD4/CD8 ratios, Th1 skewing, increased frequencies of melanoma-specific CD8(+) T cells and possible recruitment of effector NK cells, irrespective of GM co-administration. These immune-potentiating effects were counterbalanced by increased IL-10 production by T cells and significantly higher levels of FoxP3 and CTLA4 in regulatory T cells (Tregs) with correspondingly higher suppressive activity in the SLN. Notably, CpG ± GM-administered patients showed significantly lower numbers of SLN metastases (saline: 4/9, CpG + GM: 1/9, CpG: 0/10, p = 0.04). These findings indicate that i.d. delivery of low-dose CpG ± GM potentially arms the SLN of early-stage melanoma patients against metastatic spread, but that antitumor efficacy may be further boosted by counteracting the collateral activation of Tregs.

Keywords: CpG; GM-CSF; Immunotherapy; Melanoma; Regulatory T cells; Sentinel lymph node.

Fig. 1

T cell cytokine profiles reveal CpG-associated Th1 skewing. a Cytokine secretion profiles and Th skewing (represented as the average of all IFN-γ/IL-4 ratios) of freshly harvested SLN T cells after in vitro stimulation. Means (in bar graphs with SEM) are shown. b Intracellular cytokine levels in expanded SLN CD4⁺ (top row) and CD8⁺ (bottom row) T cells after in vitro stimulation. Shown are percentages of cells positive for the indicated intracellular Th1 and Th2 cytokines and Th skewing. *p < 0.05; **p < 0.01

Fig. 2

CpG/GM effects on NK cells. a Absolute changes in percentages of blood NK cells between t = −7 (pre-treatment) and t = 0 (post-treatment) are shown. b Changes in peripheral blood NK cell frequencies for CpG ± GM-administered patients are significantly and reversely correlated to corresponding SLN NK cell frequencies. c Shift from predominant regulatory CD56^bright to more effector CD56^dim NK cells in the melanoma SLN. Average CD56^dim/CD56^bright ratios are shown for each group by box and whisker plot. d Surface TRAIL expression on SLN NK cells. *p < 0.05; **p < 0.01

Fig. 3

Increased Treg activation and suppressive activity in SLN, but not in peripheral blood. a CTLA-4 and FoxP3 expression levels [by mean fluorescence intensity (MFI)] in freshly isolated SLN Tregs and IL-10 secretion after stimulation of freshly harvested SLN T cells. *p < 0.05; **p < 0.01. b Representative Treg staining of CD25-depleted and CD25-enriched expanded SLN T cells and of T cells from the peripheral blood,  % Tregs are listed. c Representative suppression assays from a saline- and a CpG-administered patient. Gray bars show the percentage of proliferated CD4⁺CD25⁻ effector cells in the presence of CD4⁺CD25⁺-enriched fractions of expanded SLN T cells at different ratios. d Closed bars indicate the suppressive activity of CD4/CD25-enriched, expanded SLN T cells for all three groups. Open bars indicate the suppressive activity of CD4/CD25-enriched T cells from the peripheral blood at the same time point as SLN harvest (t = 0). Numbers of patients tested in the SLN and peripheral blood are, respectively: saline: 5/6, CpG + GM: 7/6, CpG: 6/5. e Representative CD25/LAP and FoxP3/LAP staining after pre-gating on CD3⁺CD4⁺ cells. f LAP expression of CD3⁺CD4⁺CD25⁺ cells from expanded SLN T cells. Average percentages of LAP with SEM are shown for each group. N = 4 in each group

Fig. 4

MAA-specific CD8⁺ T cells in the SLN. a MAA-specific tetramer⁺CD8⁺ T cell rates in the SLN of HLA-A2⁺ saline-administered patients are shown for tumor-negative and tumor-positive SLNs. Each dot represents one melanoma-specific tetramer-binding CD8⁺ population. The cutoff threshold set for positive tetramer responses is shown as a dashed line. b MAA-specific tetramer⁺CD8⁺ T cell rates in tumor-negative SLNs for all three groups. Below both graphs response numbers relative to evaluated numbers of patients and epitopes are shown. *p < 0.05