Enhanced Phenotype Definition for Precision Isolation of Precursor Exhausted Tumor-Infiltrating CD8 T Cells

Abstract

In the context of adoptive T cell transfer (ACT) for cancer treatment, it is crucial to generate in vitro large amounts of tumor-specific CD8 T cells with high potential to persist in vivo. PD-1, Tim3, and CD39 have been proposed as markers of tumor-specific tumor-infiltrating CD8 T lymphocytes (CD8 TILs). However, these molecules are highly expressed by terminally differentiated exhausted CD8 T cells (Tex) that lack proliferation potential. Therefore, optimized strategies to isolate tumor-specific TILs with high proliferative potential, such as Tcf1+ precursor exhausted T cells (Tpe) are needed to improve in vivo persistence of ACT. Here we aimed at defining cell surface markers that would unequivocally identify Types for precision cell sorting increasing the purity of tumor-specific PD-1+ Tcf1+ Tpe from total TILs. Transcriptomic analysis of Tpe vs. Tex CD8 TIL subsets from B16 tumors and primary human melanoma tumors revealed that Tpes are enriched in Slamf6 and lack Entpd1 and Havcr2 expression, which encode Slamf6, CD39, and Tim3 cell surface proteins, respectively. Indeed, we observed by flow cytometry that CD39- Tim3- Slamf6+ PD-1+ cells yielded maximum enrichment for tumor specific PD-1+ Tcf1+ OT1 TILs in B16.OVA tumors. Moreover, this population showed higher re-expansion capacity upon an acute infection recall response compared to the CD39+ counterparts or bulk PD-1+ TILs. Hence, we report an enhanced sorting strategy (CD39- Tim3- Slamf6+ PD-1+) of Tpes. In conclusion, we show that optimization of CD8 TIL cell sorting strategy is a viable approach to improve recall capacity and in vivo persistence of transferred cells in the context of ACT.

Keywords: adoptive cell transfer; melanoma; memory-like CD8 T cells; precursor exhausted T cells; tumor-infiltrating CD8 T cells.

Figure 1

Surface marker identification of PD-1+ Tcf1+ TILs. (A) Volcano plots showing differential gene expression between Pdcd1+ Tcf7+ (encoding PD-1 and Tcf1, respectively) progenitor exhausted (Tpe) and Pdcd1+ Tcf7– terminal exhausted (Tex) CD8+ T cells. Genes with p-value < 0.01 (–log₁₀[p-value] > 2) are plotted. (B) Percentage of Tcf7+ cells in Pdcd1+, Pdcd1+ Tim3– or Pdcd1+ Tim3– Slamf6+ CD8+ TILs. (C) Pie charts showing mean frequency of Entpd1– and Entpd1+ (which encodes CD39) cells in total Pdcd1+ Tim3– Slamf6+ CD8+ TILs. (D) Percentage of Tcf7+ cells in Pdcd1+ Tim3– Slamf6+ Entpd1– vs. Pdcd1+ Tim3– Slamf6+ Entpd1+ cells. (A–D) show analysis of scRNAseq data from murine B16 tumors and human melanoma patients' primary tumors. (E) Representative dotplots and quantification of the percentage of Tcf1+ cells in endogenous CD8+ PD-1+ and transferred PD-1+ OT1 cells from B16.OVA tumors before and after subset definition based on the expression of Tim3, Slamf6, and CD39. A repeated measure 1-way ANOVA was performed followed by Tukey's multiple comparison test (n = 13–16). Three mice were excluded in the left panel for having acquired <100 endo PD1+ cells. (F) Percentage of each marker based subset in total endogenous CD8+ and OT1 cells from B16.OVA tumors (n = 16) (E,F) show flow cytometry analysis of CD8+ TILs where each dot represents a mouse. Pooled data of two independent experiments.

Figure 2

Re-expansion capacity assessment of PD-1+ TIL subsets in Lm-OVA infection. (A) Experimental design of panels B-E. Briefly TILs of 6 B16.OVA tumors (21 days-post engraftment) were pooled before marker based cell sorting (Supplementary Figure 3A) and 300 isolated cells were transferred to each recipient B6 mice (n = 5–6/group). The same day recipients were infected with 2000 cfu Lm-OVA. (B) Representative dotplots and quantification of the percentage of transferred OT1 cells in total CD8+ T cells following the gating strategy of Supplementary Figure 3B. n.d: non-detected. Representative of two independent experiments. (C) Representative dotplots and quantification of Tcf1+ cells in total transferred OT1 cells. (D) Representative dotplots and quantification of the percentage of IFNγ+, TNFα+ and double IFNγ+ TNFα+ cells in total transferred OT1 cells following the gating strategy of Supplementary Figure 3C. (E) IFNγ median fluorescent intensity (MFI) and TNFα MFI in IFNγ+ and TNFα+ transferred OT1 cells respectively. (B–E) show analysis of transferred OT1 cells from spleens of secondary hosts 7 days after Lm-OVA infection. (D,E) Show ex vivo analysis of OT1 cells restimulated with SIINFEKL peptide. A 1-way ANOVA followed by Tukey's multiple comparison test in panel B and an unpaired t-test between mice receiving CD39− vs. CD39+ Tim3− Slamf6+ PD1+ OT1 cells in panels (D,E) was performed. Bars represent the mean ± SEM and each dot a mouse (n = 2–6/group).