Highly active microbial phosphoantigen induces rapid yet sustained MEK/Erk- and PI-3K/Akt-mediated signal transduction in anti-tumor human gammadelta T-cells

Abstract

Background: The unique responsiveness of Vgamma9Vdelta2 T-cells, the major gammadelta subset of human peripheral blood, to non-peptidic prenyl pyrophosphate antigens constitutes the basis of current gammadelta T-cell-based cancer immunotherapy strategies. However, the molecular mechanisms responsible for phosphoantigen-mediated activation of human gammadelta T-cells remain unclear. In particular, previous reports have described a very slow kinetics of activation of T-cell receptor (TCR)-associated signal transduction pathways by isopentenyl pyrophosphate and bromohydrin pyrophosphate, seemingly incompatible with direct binding of these antigens to the Vgamma9Vdelta2 TCR. Here we have studied the most potent natural phosphoantigen yet identified, (E)-4-hydroxy-3-methyl-but-2-enyl pyrophosphate (HMB-PP), produced by Eubacteria and Protozoa, and examined its gammadelta T-cell activation and anti-tumor properties.

Methodology/principal findings: We have performed a comparative study between HMB-PP and the anti-CD3epsilon monoclonal antibody OKT3, used as a reference inducer of bona fide TCR signaling, and followed multiple cellular and molecular gammadelta T-cell activation events. We show that HMB-PP activates MEK/Erk and PI-3K/Akt pathways as rapidly as OKT3, and induces an almost identical transcriptional profile in Vgamma9(+) T-cells. Moreover, MEK/Erk and PI-3K/Akt activities are indispensable for the cellular effects of HMB-PP, including gammadelta T-cell activation, proliferation and anti-tumor cytotoxicity, which are also abolished upon antibody blockade of the Vgamma9(+) TCR Surprisingly, HMB-PP treatment does not induce down-modulation of surface TCR levels, and thereby sustains gammadelta T-cell activation upon re-stimulation. This ultimately translates in potent human gammadelta T-cell anti-tumor function both in vitro and in vivo upon transplantation of human leukemia cells into lymphopenic mice,

Conclusions/significance: The development of efficient cancer immunotherapy strategies critically depends on our capacity to maximize anti-tumor effector T-cell responses. By characterizing the intracellular mechanisms of HMB-PP-mediated activation of the highly cytotoxic Vgamma9(+) T-cell subset, our data strongly support the usage of this microbial antigen in novel cancer clinical trials.

Figure 1. Nanomolar HMB-PP replicates saturating TCR/CD3 ligation for Vγ9⁺ T-cell activation.

(A) Flow cytometry analysis for the expression of the activation marker CD69 in MACS-sorted (97–98% purity) γδ PBL, stimulated for 48 hours with the indicated amounts of HMB-PP or anti-CD3 mAb (OKT3). Shaded are non-stimulated Vγ9⁺ T-cells. Percentages refer to cells above the threshold bar. (B) Time-course of the experiment described in (A) for 1 nM HMB-PP and 1 µg/ml OKT3; cells were also stained with Annexin V to assess their viability (Annexin V⁻). (C) Cytokine bead array analysis of supernatants of MACS-sorted γδ PBL (of which 80–90% Vγ9⁺) cultures after 24 hours of stimulation with HMB-PP or OKT3. RPMI refers to cells kept in media not supplemented with activating compounds. (D) CFSE dilution assays to monitor T-cell proliferation in total PBMC cultures supplemented with HMB-PP (1 nM) or OKT3 (1 µg/ml), with or without 100 U/mL rhIL-2. Cells (gated on Vγ9⁺ or CD4⁺) were analyzed by flow cytometry after 4 days in culture; shaded are non-divided cells. Percentages indicate cells that have undergone more than 5 rounds of division. (E) CFSE dilution in gated Vγ9⁺ T-cells within 6-day cultures of total PBMC activated with 1 nM HMB-PP in the presence or absence of blocking anti-TcRVγ9 antibody. Dashed is a control incubated in 10% RPMI without HMB-PP. Results shown in this figure are representative of 3 independent experiments.

Figure 2. HMB-PP stimulation kinetically mimics Vγ9⁺ TCR/CD3 signal transduction.

(A) Phosphoimmunoblotting for kinases implicated in TCR signaling. MACS-sorted γδ PBL (of which 80–95% Vγ9⁺) were incubated with OKT3 (1 µg/ml) or HMB-PP (1 nM), in the absence (left panel) or presence (right panel) of 100 U/mL rhIL-2, for the times indicated, or kept in control media (NS, non-stimulated). Results shown in this figure are representative of 4 independent experiments. (B) TNFα and IFNγ levels were measured by CBA in the culture supernatants of MACS-sorted γδ PBL. Results were compared with the total amounts present in parallel cultures stimulated with 1 nM HMB-PP, and were expressed as percentages (IPP/HMB-PP).

Figure 3. HMB-PP-mediated Vγ9⁺ T-cell activation requires functional MEK/Erk and PI-3K/Akt signaling pathways.

Effects of MEK/Erk inhibitor UO126 and PI-3K/Akt inhibitor LY294002 on the activation and function of MACS-sorted γδ PBL (of which 85–95% Vγ9⁺). (A) Expression of activation marker CD69, assessed by flow cytometry. (B) secretion of TNFα after 24 hours of stimulation, measured by CBA. (C) Cell proliferation, assessed by CFSE dilution after 4 days in culture (percentages indicate cells that have undergone 2 or more rounds of division). (D) Jurkat leukemia cell killing, assessed by Annexin V staining and flow cytometry analysis after 6 hrs of co-incubation with pre-activated (for 3 days) γδ PBL. Results shown in this figure are representative of 3 independent experiments. Error bars represent SD and significant differences refer to controls without addition of chemical inhibitors (n = 3, *p<0.05 and **p<0.01).

Figure 4. HMB-PP treatment reproduces the transcriptional alterations induced by TCR/CD3 ligation on γδ T-cells.

(A) Volcano plots of DNA microarray comparisons between αCD3 (OKT3) mAb-treated, HMB-PP-treated and non-stimulated (“resting”) MACS-sorted γδ PBL (of which 85–95% Vγ9⁺). After 18 hours of incubation with the stimuli, RNA was extracted and submitted to Affymetrix GeneChip analysis. Represented are Fold-changes (“biological significance”) versus statistical significance (B-values). Black dots represent genes over 4-fold differentially expressed (DE) between samples; all other probed genes are depicted in grey. Genes selected as differentially expressed had adjusted p-values lower than 0.005. Results are representative of 3 independent microarray experiments (see Figure S2). (B) Real-time PCR validation of microarray results for a selection of genes similarly induced by OKT3 and HMB-PP (from Table 1). Gene expression was quantified in independent samples of control and treated cells, also including an IL-2-treated sample. Error bars represent SD and significant differences refer to “resting” cells (n = 3, *p<0.05 and **p<0.01).

Figure 5. HMB-PP does not induce down-modulation of surface Vγ9⁺ TCR and sustains cytokine production upon re-stimulation.

(A) MACS-sorted γδ PBL (of which 80–90% Vγ9⁺) were incubated for the indicated times with HMB-PP or OKT3, and stained with anti-Vγ9 mAb for flow cytometry analysis. Bold lines represent treated cells, while shaded are non-stimulated Vγ9⁺ cells (time = 0 hrs). (B) Confocal microscopy photos of γδ T-cells cultured for 24 hrs as in (A) and then stained for Vγ9⁺ TCR. (C–D) Experimental design (C) and CBA analysis (D) of the re-stimulation response of MACS-sorted γδ PBL. After 36 hrs of stimulation, cells were re-plated for secondary activation during 24 hrs, when supernatants were collected and analyzed for Th1 cytokines by CBA. Error bars represent SD and differences refer to IL-2 controls (ns, non-significant; **p<0.01; ***p<0.001). Results shown in this figure are representative of 2–5 independent experiments.

Figure 6. Leukemia cell killing by HMB-PP-activated γδ T-cells.

(A) In vitro lysis of Molt-4 leukemia cells. MACS-sorted γδ PBL (of which 85–95% Vγ9⁺) were pre-activated for 72 hours with 1 or 10 µg/ml αCD3 mAb (OKT3), or 1 or 10 nM HMB-PP in the absence of IL-2, and also combined at the lower concentrations with IL-2 (100 U/ml). For the killing assay, DDAO-SE-labelled Molt-4 cells and pre-activated γδ PBL were co-incubated for 3 hours in media devoid of activating compounds. Samples were then stained with Annexin V to identify dying (Annevin V⁺) tumor (DDAO-SE⁺) cells by flow cytometry. (B) Data summary for killing assays (as in A) performed with three distinct leukemia cell lines. (C–D) Bioluminescent imaging of NOD/SCID mice inoculated with luciferase⁺ Molt-4 leukemic cells, with (D) or without (C) co-injection of pre-activated γδ PBL, analyzed weekly as described in Materials and Methods. (E) LivingImage quantification of photon signals (tumor load) collected at day 28 of the experiment illustrated in (C–D). Comparison of γδ-treated and control animals (n = 5, p<0.05). Data in this figure are representative of 3 (A–B) or 2 (C–E) independent experiments.