CDK4/6 Inhibitor Priming Enhances PD-1 Blockade via Sellhi Neutrophil-Induced Stat5a+ Progenitor Exhausted CD8+ T Cell

Abstract

Cell cycle pathway, especially via cyclin D1-CDK4/6 signaling, is enriched in immunotherapy-resistant and immune-excluded tumors. CDK4/6 inhibitor (CDK4/6i) induces antitumor immune phenotypes by targeting both tumor and immune cells, enhancing immune checkpoint blockade (ICB), but optimal combination modalities and the corresponding cellular mechanisms remain unclear. Here, it is shown that activation of tumor cell-intrinsic cyclin D1-CDK4/6 signaling is associated with low tumor-infiltrating lymphocyte populations and immunotherapy resistance in head and neck squamous cell carcinoma (HNSCC). Comparison of sequential versus combinatorial regimens in subcutaneous or orthotopic HNSCC mice revealed that CDK4/6i priming before anti-PD-1 enhances response durability by promoting CD8⁺ and CD4⁺ T cell infiltration and decreasing overall neutrophil abundance. Mechanistically, IL15-secreted Sell(hi) neutrophils induced Stat5a+ progenitor exhausted CD8⁺ T cells contributed to the antitumor effect of CDK4/6i priming modalities. Together with corroborating evidence from a clinically relevant patient-derived-organoid-TIL (PDO-TIL) co-culture model, these findings support further clinical testing of brief CDK4/6i dosing before anti-PD-1 to improve ICB efficacy.

Keywords: CD8+Tpex; CDK4/6i priming; HNSCC; Sell(hi) neutrophils; Stat5a.

Figure 1

Cyclin D1‐CDK4/6 signaling is associated with T cell exclusion and immunotherapy resistance. a) Heatmap comparing cell cycle signature, H‐score of Rb phosphorylation, Cyclin D1, counts of CD3⁺, CD4⁺, CD8⁺ T cells, p16 positive percentage and the combined positive score of PD‐L1 (CPS), with genomic aberrations annotated. Cell cycle signature and genomic aberrations of 30 PDC models were annotated using RNA‐sequencing and Whole exome sequencing respectively. b) Representative images of IHC staining of Cyclin D1, Phospho‐Rb, PD‐L1, CD3, CD4, CD8 in cell cycle signature high and low HNSCC patients. Scale bar, 100, 20 µm. c) Representative images of mIF showing cell cycle signature high HNSCC patient coupling with high expression of Cyclin D1, Phospho‐Rb and low PD‐L1 expression and CD3, CD8 infiltration, and cell cycle signature low HNSCC patient coupling with low expression of Cyclin D1, Phospho‐Rb and high PD‐L1 expression and CD3, CD8 infiltration. Scale bar, 100, 20 µm. d) Mean tumor growth curves showing tumor volume in MOC1 and MOC2 bearing‐mice treated with IgG or anti‐PD‐1 antibodies. Data are mean ± standard error of mean (s.e.m) (n = 5–6 mice per group). Two‐way ANOVA analysis. e) Flow cytometry of Phospho‐Rb, Ki67, PD‐L1 expressed in Epcam⁺ tumor cell in MOC1 or MOC2 tumors (n = 5). MFI, mean fluorescence intensity. Data are presented as mean ± standard deviation (s.d). Statistical significance was determined by two‐sided unpaired t‐test.

Figure 2

CDK4/6i priming before PD‐1 blockade shows superior antitumor activity than concurrent regimens and overcome immunotherapy resistance. a–d) Preclinical treatment regimen with anti‐PD‐1 day0‐7, CDK4/6i day0‐7, concurrent treatment of anti‐PD‐1 and CDK4/6i day0, palbociclib priming 7 days before anti‐PD‐1 combination, CDK4/6i priming 7 days before anti‐PD‐1 monotherapy and anti‐PD‐1 priming 7 days before CDK4/6i combination cisplatin for mice bearing orthotopic and subcutaneous MOC1 and MOC2. For the CDK4/6i priming strategy, CDK4/6i was administered from day 0 (randomization time) to day 4 for orthotopic and from day 0 (randomization time) to day 7 for subcutaneous tumor after tumor inoculation. CDK4/6i priming was initiated on day 3–4 after orthotopic tumor inoculation. For subcutaneous tumors, it was initiated when the tumor volumes reached 100–150 mm³. Dark gray circles indicate regimen and time points for CyTOF analysis in Figure 3. Tongue tumor tissues were collected from three syngeneic subcutaneous or orthotopic tumor models at time points and in treatment regimens including: 1) vehicle, 2) CDK4/6i, 3) anti‐PD‐1, 4) CDK4/6i + anti‐PD‐1 (on d4 for orthotopic and d7 for subcutaneous tumor); 5) C‐PC, 6) C‐P, 7) P‐CP (on d11 for orthotopic and d13 for subcutaneous tumor) for CyTOF analysis. Mouse tumor growth curves in response to treatment strategy outlined in (a), n = 5‐6, two‐way ANOVA test. Error bands were shaded. e,f) Mouse tumor growth curves in response to treatment strategy outlined in (d), n = 5, two‐way ANOVA test. Error bands were shaded. Data in (b, c, e, f) are presented as mean ± S.E.M.

Figure 3

T cell expansion is predominantly associated with response to optimized therapy regimen. a) t‐SNE maps (left) of tumor‐infiltrating CD45⁺ cells analyzed by CyTOF in three indicated syngeneic subcutaneous or orthotopic tumor models at time points and in treatment regimens indicated in Figure 2g–j. Three pooled tumor tissues were collected from syngeneic subcutaneous or orthotopic tumor models at time points and in treatment regimens including: 1) vehicle, 2) CDK4/6i, 3) anti‐PD‐1, 4) CDK4/6i + anti‐PD‐1 (on d4 for orthotopic and d7 for subcutaneous tumor); 5) C‐PC, 6) C‐P, 7) P‐CP (on d11 for orthotopic and d13 for subcutaneous tumor). Heatmaps (right) showing the expression values of immune phenotypic protein markers normalized to the maximum mean value across subsets. b) Frequencies of immune cell types in the CD45⁺ population of three pooled tumor tissues collected from syngeneic tumor models at time points and treatment regimens indicated in Figure 2g–j. c,d) Frequencies of PD1⁺CD44⁺CD4⁺ and PD1⁺CD44⁺CD8⁺ in the CD45⁺ population of three syngeneic tumor models at time points and treatment regimens indicated in Figure 2g–j, n = 3, one‐way ANOVA test. e) NLR calculated as absolute neutrophil counts divided by absolute lymphocyte counts in three syngeneic tumor models at time points and treatment regimens indicated in Figure 2i–l. f,g) Representative images of mIF staining of CK5, Ly‐6G, CD4, CD8, PD‐1 in V, CP and C‐P of MOC2 orthotopic tumors and related histological quantification are shown. Scale bar, 1000 µm, n = 5, one‐way ANOVA test. h,i), MOC2 tumor‐bearing mice were treated with αLy6G, αCD4, αCD8, or isotype control 2 days prior to treatment with C‐P, then every 3 days until end of C‐P or vehicle treatment. Tumor volume was measured every 2 to 4 days. Data in (c, d, h, i) are presented as mean ± S.E.M. Data in g) are presented as mean ± S.D.

Figure 4

CD8⁺ Tex tended to form progenitor‐like Tex and retain proliferative and cytotoxic capacity under C‐P regimens. a) Maintype TSNE dimensionality reduction visualization of scRNA data in MOC2 syngeneic orthotopic tumor models at time points and in treatment regimens indicated in Figure S4a (Supporting Information). Three pooled tongue tumor tissues were collected from MOC2 syngeneic orthotopic tumor models at time points and in treatment regimens including: 1) vehicle, 2) CDK4/6i, 3) anti‐PD‐1, 4) CDK4/6i + anti‐PD‐1 (on d4); 5) C‐PC, 6) C‐P, 7) P‐CP (on d10). b) Maintype TSNE dimensionality reduction visualization by sample grouping. c) Visualization of TSNE dimensionality reduction for T cell subgroups. d) Visualization of T cell subgroups marker gene expression using a dot plot. e) Visualization of pseudotime analysis for CD8⁺ T cell subgroups using scVelo, including RNA velocity_embedding_stream and latent time. f) Quantification of flow cytometry of PD‐1, TIM3, SLAMF6 in CD8⁺ T cells from vehicle, CP, and C‐P groups (= 4 mice per group). g) Quantification of flow cytometry of Ki‐67, IFNγ, Granzyme B in CD8⁺ T cells from vehicle, C‐PC, and C‐P groups (n = 4 mice per group). Data are presented as mean ± S.D in (f) and (g). Statistical significance was determined by one‐way ANOVA test in (f) and (g).

Figure 5

Stat5a mediates the functionally antitumor phenotype of CD8⁺Tpex cells. a) Volcano plot showing the DEG between CD8⁺ Tex and other CD8⁺ T cells in C‐P and C‐PC group. b) Bar charts showing the proportion of Stat5a⁺Tpex in the C‐P group and the vehicle group at various time points (n = 5). c) Experimental design. d) Quantification of flow cytometry analysis of frequencies of CD8⁺ Tpex, GZMB CD8⁺ cells in WT, Stat5a EV and OE groups (n = 6), two‐way ANOVA test. e) Specific cell killing of GFP+ target cells during the co‐incubation of MOC2‐OVA‐GFP and WT, Stat5a EV or Stat5a OE OT‐I CD8⁺ T‐cells for 48h under indicated E:T ratio (n = 5). two‐way ANOVA test. f) Experimental design. g,h) Absolute number of GFP⁺ target cells during the co‐incubation of MOC2‐OVA‐GFP and WT OT‐I‐Tpex, Stat5‐EV OT‐I‐CD8⁺Tpex or Stat5‐OE OT‐I‐CD8⁺Tpex at the indicated times (n = 3). The representative FACS plots for GFP⁺ target cells are shown (g). Quantification of specific killing percentage in indicated time is shown (h), (n = 3), two‐way ANOVA test. Data are presented as mean ±s.d in (b, d, e, h).

Figure 6

Sell^(hi) neutrophils induced CD8⁺Tpex cells through IL15‐Stat5a axis. a) UMAP dimensionality reduction visualization of neutrophil clusters. b) Visualization of marker gene expression for neutrophil subgroups using a pheatmap. c) UMAP visualization of gene expression density for ll15 in neutrophil subgroups. d) Experimental design. e) IL15 secretion measured by ELISA in Sell^(hi) and Sell^(lo) neutrophil supernatant. One‐way ANOVA test. f) Bar plots showing the frequencies of Tpex, Ttex, p‐Stat5⁺ Tpex, IFNγ⁺, and GZMB⁺OT‐I CD8⁺ T cell in indicated treatment groups, one‐way ANOVA test. g) Specific cell killing of GFP⁺ target cells during the co‐incubation of MOC2‐OVA‐GFP, sorted Sell^(hi) and Sell^(lo) neutrophils and CD8⁺ T‐cells from OT‐I mice under indicated treatments and at the indicated times (n=3), one‐way ANOVA test. h) Experimental design of in vivo adoptive cell transfer using OT‐I CD8⁺ cells and neutrophils. i,j), Tumor images and volumes of in vivo experiment using MOC2‐OVA tumor orthotopic model treated with indicated treatments. Sell^(hi) and Sell^(lo) neutrophil population were defined as CD45+/CD11b+/Ly6C‐Ly6G+Sell+ and CD45+/CD11b+/Ly6C‐Ly6G+Sell‐. Data in (e‐g, j) are presented as mean ± S.D.

Figure 7

Blocking IL15‐Stat5a signaling dampens antitumor activity of C‐P regimen. a) Mean tumor growth curves showing tumor volume in mice treated with IgG, C‐P regimen, C‐P regimen plus anti‐IL15 antibody and C‐P regimen plus anti‐Ly6G antibody. (n = 5 mice per group), two‐way ANOVA test. b) Frequencies of Sell^(hi) neutrophils in indicated treatment groups, n=5, one‐way ANOVA test. c) Frequencies of Tpex, Ttex and p‐Stat5⁺ Tpex in indicated treatment groups, n=5, one‐way ANOVA test. d) Experimental design. e) Representative bright fields and Calcein/PI fluorescence staining images of Organoid/TIL coculture assays derived from two HNSCC patients (left). Relative cell viability of each treatment group (right). Data in Data in (e) are presented as mean ± s.d. P+T: PDO+TIL; C‐P: CDK4/6 lead‐in before PD‐1.