PD-1 defines a distinct, functional, tissue-adapted state in Vδ1+ T cells with implications for cancer immunotherapy

Abstract

Checkpoint inhibition (CPI), particularly that targeting the inhibitory coreceptor programmed cell death protein 1 (PD-1), has transformed oncology. Although CPI can derepress cancer (neo)antigen-specific αβ T cells that ordinarily show PD-1-dependent exhaustion, it can also be efficacious against cancers evading αβ T cell recognition. In such settings, γδ T cells have been implicated, but the functional relevance of PD-1 expression by these cells is unclear. Here we demonstrate that intratumoral TRDV1 transcripts (encoding the TCRδ chain of Vδ1⁺ γδ T cells) predict anti-PD-1 CPI response in patients with melanoma, particularly those harboring below average neoantigens. Moreover, using a protocol yielding substantial numbers of tissue-derived Vδ1⁺ cells, we show that PD-1⁺Vδ1⁺ cells display a transcriptomic program similar to, but distinct from, the canonical exhaustion program of colocated PD-1⁺CD8⁺ αβ T cells. In particular, PD-1⁺Vδ1⁺ cells retained effector responses to TCR signaling that were inhibitable by PD-1 engagement and derepressed by CPI.

Fig. 1. Transcriptomic presence of Vδ1⁺ cells within melanoma predicts the response to anti-PD-1 and anti-PD-L1 CPI therapy.

a, Expression (transcripts per million (TPM)) of TRDV1, TRBC2, CD4 and CD8B plotted according to the objective response to anti-PD-1 and anti-PD-L1 CPI therapy. Mann–Whitney U-test. b, Left, expression of TRDV1 in tumors with below-median neoantigen loads from Liu et al. plotted according to the objective response to anti-PD-1 CPI therapy. Right, progression-free survival (PFS) of patients with below-median neoantigen loads from Liu et al. split on the median expression of TRDV1. log-rank test. c, Left, expression of TRDV1 in tumors with above-median neoantigen loads from Liu et al. plotted according to the objective response to anti-PD-1 CPI therapy. Mann–Whitney U-test. Right, PFS of patients with above-median neoantigen loads from Liu et al. split on the median expression of TRDV1. log-rank test. All P values presented are two-sided where relevant. For the box plots, the boxes denote the medians and interquartile ranges (IQRs); the whiskers denote the minimum and maximum values. NR, nonresponder; NS, not significant; R, responder. Source data

Fig. 2. Expression of PD-1 on Vδ1⁺ cells is associated with distinct biology compared with its expression on CD8⁺ T cells.

a, Schematic of nonenzymatic extraction of lymphocytes from primary human tissue and subsequent expansion protocol. b, Summary flow cytometry data of Vδ1⁺ cell enrichment in skin-expanded (Exp.) lymphocytes compared with skin lymphocytes obtained by direct enzymatic digestion (Enz.) and grid isolation (Iso.). Repeated measures one-way analysis of variance (ANOVA) followed by Holm–Šídák’s multiple comparisons test. Adjusted P values as indicated. c, Representative flow cytometry plot of PD-1 expression on skin-expanded CD8⁺ T cells (left) and Vδ1⁺ cells (right). PD-1 gates were set on paired unstained samples. Representative of n = 5 independent donors. d, Summary flow cytometry data of PD-1 expression on Vδ1⁺ and CD8⁺ T cells in skin-expanded lymphocytes from n = 5 independent donors. Paired t-test. The boxes denote the medians and IQRs and the whiskers denote the minimum and maximum values. e, Summary flow cytometry data of PD-1 expression on skin-expanded Vδ1⁺ cells from n = 4 independent donors after 48-h culture in vitro in the presence of plate-bound IgG (control) or anti-TCR-Vδ1 antibody. Results were plotted as the percentage of Vδ1⁺ cells positive for PD-1 (left) and mean fluorescence intensity (MFI) of the PD-1⁺ Vδ1 population (right). Data points are paired according to donor. Paired t-test. f, PCA of cell populations sorted from skin-expanded lymphocytes from healthy skin of independent donors based on normalized counts of all 757 genes in the nCounter Immune Exhaustion Panel (excluding γδ and CD8⁺ T cell lineage genes and PDCD1; Methods) highlighted according to cell type (Vδ1⁺ cells, gray circles; CD8⁺ T cells, white squares). g, PCA of cell populations sorted from skin-expanded lymphocytes highlighted according to PD-1 status. Left, PD-1⁻Vδ1⁺ (white circles, n = 8 independent donors) and PD-1⁺Vδ1⁺ (red circles, n = 8 independent donors) cells are highlighted. Right, PD-1⁻CD8⁺ T (white squares, n = 9 independent donors) and PD-1⁺CD8⁺ T (red squares, n = 6 independent donors) cells are highlighted. All P values presented are two-sided. Source data

Fig. 3. Expression of PD-1 marks a transcriptional program of tissue residency and survival in Vδ1⁺ cells, which is distinct from its associations in CD8⁺ T cells.

a, Differential expression of genes associated with a tissue-resident memory phenotype in human T cells in sorted skin-expanded lymphocytes, plotted as the mean log₂ fold change between PD-1⁺ and PD-1⁻Vδ1⁺ and CD8⁺ T cells (Vδ1⁺n = 7 paired independent donors; CD8⁺n = 6 paired independent donors). The color denotes directionality and the size of the circle denotes the fold change. Empty circles plotted where no samples within a cell type had detectable counts above the threshold (for example, SELL in Vδ1⁺ cells) are shown. Paired ratio t-test. The black border denotes P < 0.05. b, Gene expression of CD101 and ZNF683 in Vδ1⁺ (left) and CD8⁺ T (right) cells plotted according to PD-1 status. Data points are paired according to the independent donors. Paired ratio t-test. The dotted line signifies the detection threshold (normalized count = 20). c, Differential expression of genes encoding the surface markers of T cell exhaustion and transcription factors implicated in T cell survival, stemness and exhaustion in sorted skin-expanded lymphocytes. Plotted as the mean log₂ fold change between PD-1⁺ and PD-1⁻Vδ1⁺ and CD8⁺ T cells (Vδ1⁺, n = 7 paired independent donors; CD8⁺, n = 6 paired independent donors). The color denotes directionality and the size of the circle denotes the fold change. Paired ratio t-test. The black border denotes P < 0.05. d, Gene expression of BCL6, IRF4 and TOX in Vδ1⁺ (left) and CD8⁺ T (right) cells plotted according to PD-1 status. e, Gene expression of XCL1 and XCL2 in Vδ1⁺ (left) and CD8⁺ T (right) cells plotted according to PD-1 status. The data points are paired according to the independent donors. Paired ratio t-test. The dotted line signifies the detection threshold (normalized count = 20). All P values presented are two-sided. Source data

Fig. 4. PD-1⁺Vδ1⁺ cells are functionally competent.

a, Differential expression of genes associated with pro-tumor and anti-tumor T cell functions in sorted skin-expanded lymphocytes, plotted as the mean log₂ fold change between PD-1⁺ and PD-1⁻ Vδ1⁺ and CD8⁺ T cells (Vδ1⁺, n = 7 paired independent donors; CD8⁺, n = 6 paired independent donors). The color denotes directionality and the size of the circle denotes the fold change. Empty circles plotted where no samples within a cell type had detectable counts above threshold are shown. Paired ratio t-test. The black border denotes P < 0.05. b, Gene expression of FASLG, GZMH, GZMK, PRF1, TNF, KLRK1, NCR1 and IL17A in Vδ1⁺ (left) and CD8⁺ T (right) cells plotted according to PD-1 status. The data points are paired according to the independent donors. Paired ratio t-test. The dotted lines denote the limit of detection (normalized count = 20). c, Summary flow cytometry data of surface CD107A and intracellular TNF, IFNγ and IL-17A staining of in vitro-activated, skin-expanded PD-1⁻ (P⁻, white circle) and PD-1⁺ (P⁺, black circle) Vδ1⁺ cells. Cells were activated in vitro with PBS (negative control), plate-bound MICA, plate-bound anti-TCR-Vδ1 or plate-bound anti-CD3 as indicated. The data points are paired according to n = 5 independent donors. Paired t-test. Source data

Fig. 5. PD-1⁺Vδ1⁺ cells can be regulated by PD-1 engagement and derepressed by therapeutic CPI.

a, Summary flow cytometry data of surface CD107A, and intracellular TNF and IFNγ staining, in PD-1⁺Vδ1⁺ cells activated in vitro with plate-bound anti-TCR-Vδ1 antibody in the presence of control plate-bound mouse IgG (−), rhPD-L1 (P) or rhPD-L1 and atezolizumab (P + A). b, Summary flow cytometry data of surface CD107A, and intracellular TNF and IFNγ staining in PD-1⁺Vδ1⁺ cells activated in vitro with plate-bound anti-CD3 antibody in the presence of control plate-bound mouse IgG (−), rhPD-L1 (P) or rhPD-L1 and atezolizumab (P + A). c, Summary flow cytometry data of surface CD107A, and intracellular TNF and IFNγ staining, in PD-1⁺Vδ1⁺ cells activated in vitro with plate-bound MICA in the presence of control plate-bound mouse IgG (−), rhPD-L1 (P) or rhPD-L1 and atezolizumab (P + A). A repeated measures one-way ANOVA followed by a Holm–Šídák’s multiple comparisons test was used. The data points are paired according to n = 5 independent donors. Adjusted P values are shown. All P values presented are two-sided. Source data

Fig. 6. Primary tumor-derived PD-1⁺Vδ1⁺ cells are variably regulated by PD-1 engagement and derepressed by therapeutic CPI.

a, Summary flow cytometry data of surface CD107A, and intracellular TNF and IFNγ staining, in NSCLC-expanded PD-1⁺Vδ1⁺ cells activated in vitro with plate-bound anti-CD3 antibody in the presence of control plate-bound mouse IgG (−), rhPD-L1 (P) or rhPD-L1 and atezolizumab (P + A). b, Summary flow cytometry data of surface CD107A, and intracellular TNF and IFNγ staining, in NSCLC-expanded PD-1⁺Vδ1⁺ cells activated in vitro with plate-bound MICA in the presence of control plate-bound mouse IgG (−), rhPD-L1 (P) or rhPD-L1 and atezolizumab (P + A). A repeated measures one-way ANOVA followed by a Holm–Šídák’s multiple comparisons test was used. The data points are paired according to n = 5 independent donors. Adjusted P values are shown. All P values presented are two-sided. Source data

Extended Data Fig. 1. RNAseq TCR V gene alignment.

A, Heatmap of TRAV and TRBV genes detected across all five melanoma public datasets analysed for Fig. 1 segregated by TRDV1 status (detected versus not detected) demonstrating a coordinated global lack of TCR V gene alignment. Unfilled cells represent undetected genes. Colour scale denotes Log2(TPM + 1) gene expression. B, Objective response rates of entire cohort of baseline cases (n = 216), baseline cases where TRDV1 was not detected (n = 89) and cases where TRDV1 was either below (TRDV1 low, n = 63) or above (TRDV1 high, n = 64) median (see Methods). Source data

Extended Data Fig. 2. Grid isolation and expansion enriches for skin-derived Vδ1+ cells and improves yield.

A, Representative flow cytometry dot plots demonstrating Vd1+ cell enrichment in skin-expanded lymphocytes (Exp) compared with direct enzymatic digestion (Enz) and grid isolation (Iso). Gated on live, single, CD45+ cells. Representative of n = 5 independent donors. B, Summary flow cytometry data of intracellular Ki-67 staining in grid-isolated Vδ1+ cells (Iso) and after 14 days of expansion (Exp). Ki-67 MFI index calculated as Ki-67 MFI for the whole Vδ1+ population divided by the MFI of the isotype control for the same population. Paired t test. C, Summary data of absolute numbers of Vδ1+ cells obtained from a single grid at the end of 3 weeks of grid isolation (Iso) and at the end of 3 weeks of expansion (Exp). Ratio paired t test. Datapoints represent individual skin donors. All P values presented are two-sided. For box plots, boxes denote medians and interquartile ranges, and whiskers denote minimum and maximum values. Source data

Extended Data Fig. 3. NanoString transcriptomic analysis of skin-derived grid-expanded Vδ1+ and CD8+ T cells.

A, Representative gating strategy for sorting PD-1- and PD-1 + Vδ1+ and CD8+ T cells from bulk, skin-expanded lymphocytes. PD-1 gate set on fluorescence minus one (FMO) control (blue). PD-1- and PD-1 + Vδ1+ and CD8+ T cells (red) were sorted and lysed in RLT buffer for downstream analysis using the NanoString nCounter Immune Exhaustion Panel. Representative of n = 9 independent donors. B, Expression of anticipated canonical lineage genes in each sorted population assessed by NanoString. Dotted line = detection threshold (normalised count = 20). Datapoints represent gene expression in each cell type sorted from individual donors. Source data

Extended Data Fig. 4. PD-1 expression on Vδ1+ cells is associated with distinct biology compared to its expression on CD8+ T cells.

A, PC1 from Fig. 2F plotted by cell type. Datapoints paired by PD-1 status and donor (n = 14 pairs). Paired t test. B, PC2 from Fig. 2G plotted by cell type and PD-1 status and paired by donor (Vδ1 n = 7 pairs, CD8 n = 6 pairs). Paired t test. All P values presented are two-sided. n.s. = not significant. Source data

Extended Data Fig. 5. In vitro activation of skin-expanded Vδ1+ cells.

A, Schema of in vitro activation assay. Plates were coated with anti-TCR-Vδ1, anti-CD3 or MICA along with either mouse IgG (control) or rhPD-L1 overnight. Atezolizumab was added 30 minutes prior to addition of skin-expanded lymphocytes to relevant wells. Skin-expanded lymphocytes were seeded in the presence of brefeldin A and incubated for 5 hours at 37^oC before staining for flow cytometry and acquisition. Representative flow cytometry plots of surface CD107A, and intracellular TNFα, IFNγ and IL-17A staining after activation with B, MICA, C, anti-TCR-Vδ1 or D, anti-CD3. Gated either PD-1- or PD-1 + Vδ1+ cells. Representative of n = 5 independent donors.

Extended Data Fig. 6. PD-1+ CD8+ T cells are suppressed after PD-1 engagement with variable de-repression by therapeutic CPI.

Summary flow cytometry data of surface CD107A, and intracellular TNFα and IFNγ staining in skin-expanded PD-1+ CD8+ T cells activated in vitro with plate-bound anti-CD3 in the presence of control plate-bound mouse IgG (-), rhPD-L1 (P), or rhPD-L1 and atezolizumab (P + A). Repeated measures one-way ANOVA followed by Holm-Sidak’s multiple comparisons test. Datapoints paired by n = 5 independent donors. Adjusted P values indicated. All P values presented are two-sided. n.s. = not significant. Source data

Extended Data Fig. 7. Functional regulation of primary NSCLC-expanded Vδ1+ and CD8 T+ cells by PD-1.

A, Donor level flow cytometry staining of surface CD107A, and intracellular TNFα, IFNγ and IL-17A in NSCLC-expanded PD-1+ Vδ1+ cells activated in vitro with plate-bound anti-CD3 (top) and MICA (bottom) in the presence of control plate-bound mouse IgG (-), rhPD-L1 (P), or rhPD-L1 and atezolizumab (P + A). B, Donor level flow cytometry staining of surface CD107A, and intracellular TNFα, IFNγ and IL-17A in PD-1+ NSCLC-expanded CD8+ T cells activated in vitro with plate-bound anti-CD3 (top) and MICA (bottom) in the presence of control plate-bound mouse IgG (-), rhPD-L1 (P), or rhPD-L1 and atezolizumab (P + A). Cell values and shading denote percentage of parent gate (PD-1+ Vδ1 cells or PD-1+ CD8+ T cells) positive for stain. C, Summary flow cytometry data of surface CD107A, and intracellular TNFα and IFNγ staining in NSCLC-expanded PD-1+ CD8+ T cells activated in vitro with plate-bound anti-CD3 antibody in the presence of control plate-bound mouse IgG (-), rhPD-L1 (P), or rhPD-L1 and atezolizumab (P + A). Repeated measures one-way ANOVA followed by Holm-Sidak’s multiple comparisons test. Datapoints paired by n = 5 independent donors. Adjusted P values indicated. All P values presented are two-sided. n.s. = not significant. Source data