Type-i interferon response and exhaustion gene signature as a biomarker for cancer immunotherapy

Abstract

Provided herein is a method of treating a patient having a tumor, including determining whether one or more immune cells from the patient express an exhaustion phenotype and administering to the patient one or more therapeutic compositions, wherein: the therapeutic composition includes either one or more checkpoint inhibitors (ICIs) in an amount effective to treat the tumor when the one or more immune cells express the exhaustion phenotype or one or more compounds effective to prime the one or more immune cells to respond to one or more ICIs when the one or more cells do not express the exhaustion phenotype.

Description

TYPE-I INTERFERON RESPONSE AND EXHAUSTION GENE SIGNATURE AS A BIOMARKER FOR CANCER IMMUNOTHERAPY CROSS REFERENCE TO RELATED APPLICATION

[0001] This application claims priority to U. S. Provisional Patent Application No.63 / 717,570, filed November 7, 2024, the disclosure of which is hereby incorporated by reference in its entirety.STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH

[0002] This invention was made with government support under CA097190, CA206517, and DE031947 awarded by the National Institutes of Health. The government has certain rights in the invention.BACKGROUND OF THE INVENTIONField of the Invention

[0003] Provided herein are compositions and methods of using the same for treatment of conditions including cancer.Description of Related Art

[0004] Clinical benefits observed with immune checkpoint inhibitors (ICI) in the recurrent and metastatic (R / M) head and neck squamous cell carcinoma (HNSCC) setting have sparked increased interest in investigating immunotherapy at earlier stages of treatment. Recent discoveries in neoadjuvant treatments for HNSCC have further proven the effectiveness of this approach. However, despite all the advancements, the response rates in HNSCC patients remain limited, and the mechanisms of immune evasion remain poorly understood. Accordingly, there is a need in the art for additional treatments for cancer, including of the head and neck.SUMMARY OF THE INVENTION

[0005] Provided herein is a method of treating a patient having a tumor, including determining whether one or more immune cells from the patient express an exhaustion phenotype; and administering to the patient one or more therapeutic compositions, wherein: the therapeutic composition includes either: one or more checkpoint inhibitors (ICIs) in an amount effective to treat the tumor when the one or more immune cells express the exhaustion phenotype; or one or more compounds effective to prime theone or more immune cells to respond to one or more ICIs when the one or more immune cells do not express the exhaustion phenotype.

[0006] Also provided herein is a method of treating a patient having a tumor, including administering, to a patient having one or more immune cells expressing an exhaustion phenotype, a therapeutic composition including one or more ICIs in an amount effective to treat the tumor.

[0007] Also provided herein is a method of treating a patient having a tumor, including administering, to a patient having one or more immune cells that does not express an exhaustion phenotype, a first therapeutic composition followed by a second therapeutic composition, wherein the first therapeutic composition includes one or more compounds effective to prime the one or more immune cells to respond to one or more ICIs and the second therapeutic composition includes one or more ICIs.

[0008] Also provided herein is a method of treating a patient having a tumor, including determining whether one or more immune cells from the patient express a type-1 IFN response signature; and administering to the patient one or more therapeutic compositions, wherein: the therapeutic composition includes either: one or more checkpoint inhibitors (ICIs) in an amount effective to treat the tumor when the one or more immune cells express the type-1 IFN response signature; or one or more compounds effective to prime the one or more immune cells to respond to one or more ICIs when the one or more immune cells do not express the type-1 interferon response signature.

[0009] Also provided herein is a method of treating a patient having a tumor, including administering, to a patient having one or more immune cells expressing a type-1 interferon response signature, a therapeutic composition including one or more ICIs in an amount effective to treat the tumor.

[0010] Also provided herein is a method of treating a patient having a tumor, including administering, to a patient having one or more immune cells that does not express a type-l interferon response signature, a first therapeutic composition followed by a second therapeutic composition, wherein the first therapeutic composition includes one or more compounds effective to prime the one or more immune cells to respond to one or more ICIs and the second therapeutic composition includes one or more ICIs.

[0011] Further non-limiting embodiments are set forth in the following numbered clauses:

[0012] 1. A method of treating a patient having a tumor, comprising: determining whether one or more immune cells from the patient express an exhaustion phenotype; and administering to the patient one or more therapeutic compositions, wherein: the therapeutic composition comprises either: one or more checkpoint inhibitors (ICIs) in an amount effective to treat the tumor when the one or more immune cells express the exhaustion phenotype; or one or more compounds effective to prime the one or more immune cells to respond to one or more ICIs when the one or more immune cells do not express the exhaustion phenotype.

[0013] 2. The method of clause 1, further comprising, when the tumor cell does not express the exhaustion phenotype, administering one or more ICIs.

[0014] 3. The method of clause 1 or clause 2, wherein the one or more ICIs comprise one or more PD-1 inhibitors, PD-L1 -inhibitors, CTLA-4 inhibitors, and / or LAG-3 inhibitors.

[0015] 4. The method of any of clauses 1-3, wherein the one or more ICIs comprise one or more antibodies that bind PD-1, PD-L1, CTLA-4, and / or LAG-3.

[0016] 5. The method of any of clauses 1-4, wherein the one or more ICIs comprise one or more of nivolumab, pembrolizumab, cemiplimab, atezolizumab, durvalumab, avelumab, ipilimumab, relatlimab, fianlimab, and / or leramlimab.

[0017] 6. The method of any of clauses 1-5, wherein the one or more ICIs comprise a PD-1 inhibitor and a LAG-3 inhibitor.

[0018] 7. The method of any of clauses 1-6, wherein the one or more ICIs comprise nivolumab and relatlimab.

[0019] 8. The method of any of clauses 1-7, wherein the one or more compounds effective to prime the one or more immune cells to respond to one or more ICIs comprise one or more of a STING agonist and a TLR9 agonist.

[0020] 9. The method of any of clauses 1 -8, wherein the one or more compounds effective to prime the one or more immune cells to respond to one or more ICIs comprise one or more of cGAMP, DMXAA, MSA-2, TAK-676, a CpG oligodeoxynucleotide, IMO-2055, and / or lefitolimod.

[0021] 10. The method of any of clauses 1 -9, wherein the one or more immune cells comprise one or more CD8+ cells, optionally CD8+ TIL cells.

[0022] 11. The method of any of clauses 1-10, wherein the one or more immune cells comprise one or more NK cells, optionally tumor-infiltrating NK cells.

[0023] 12. The method of any of clauses 1-11, wherein the one or more immune cells comprise one or more CD4+ cells, optionally CD4+ conventional T cells or CD4+ regulatory T cells.

[0024] 13. The method of any of clauses 1-12, wherein the one or more immune cells are one or more CD8+ cells and the exhaustion phenotype comprises increased expression of one or more of ACP5, CCL3, CD38, CTSW, CXCL13, EIF2AK2, ENTPD1, EPSTI1, GNLY, HAVCR2, HERC5, IFI6, IFI44, IFI44L, IFIT1, ISG15, ISG20, LAG3, MX1, OASL, OAS1, OAS2, PARP14, PLSCR1, PTMS, RBPJ, RSAD2, STAT1, TOX, and / or XAF1 compared to a cell not expressing the exhaustion phenotype.

[0025] 14. The method of any of clauses 1-13, wherein the one or more immune cells are immune cell is an NK cell and the exhaustion phenotype comprises an increase in one or more of MTOR signaling, oxidative phosphorylation, interferon-[3 signature, and / or interferon-a signature compared to a cell not expressing the exhaustion phenotype.

[0026] 15. The method of any of clauses 1 -14, wherein the immune cell is a CD4+ cell and the exhaustion phenotype comprises increased expression of one or more of LAG3, CTLA4, PDCD1, ENTPD1, HAVCR2, and / or TIGIT compared to a cell not expressing the exhaustion phenotype.

[0027] 16. The method of any of clauses 1 -15, wherein the therapeutic composition is administered orally.

[0028] 17. The method of any of clauses 1 -16, wherein the therapeutic composition is administered parenterally.

[0029] 18. The method of any of clauses 1-17, wherein the tumor is a squamous cell tumor.

[0030] 19. The method of any of clauses 1-18, wherein the patient has head and neck squamous cell carcinoma.

[0031] 20. A method of treating a patient having a tumor, comprising: administering, to a patient having one or more immune cells expressing an exhaustion phenotype, a therapeutic composition comprising one or more ICIs in an amount effective to treat the tumor.

[0032] 21. The method of clause 20, wherein the one or more ICIs comprise one or more PD-1 inhibitors, PD-L1 -inhibitors, CTLA-4 inhibitors, and / or LAG-3 inhibitors.

[0033] 22. The method of clause 20 or clause 21, wherein the one or more ICIs comprise one or more antibodies that bind PD-1, PD-L1, CTLA-4, and / or LAG-3.

[0034] 23. The method of any of clauses 20-22, wherein the one or more ICIs comprise one or more of nivolumab, pembrolizumab, cemiplimab, atezolizumab, durvalumab, avelumab, ipilimumab, relatlimab, fianlimab, and / or leramlimab.

[0035] 24. The method of any of clauses 20-23, wherein the one or more ICIs comprise a PD-1 inhibitor and a LAG-3 inhibitor.

[0036] 25. The method of any of clauses 20-24, wherein the one or more ICIs comprise nivolumab and relatlimab.

[0037] 26. The method of any of clauses 20-25, wherein the one or more immune cells comprise one or more CD8+ cells, optionally CD8+ TIL cells.

[0038] 27. The method of any of clauses 20-26, wherein the one or more immune cells comprise one or more NK cells, optionally tumor-infiltrating NK cells.

[0039] 28. The method of any of clauses 20-27, wherein the one or more immune cells comprise one or more CD4+ cells, optionally CD4+ conventional T cells or CD4+ regulatory T cells.

[0040] 29. The method of any of clauses 20-28, wherein the one or more immune cells are one or more CD8+ cells and the exhaustion phenotype comprises increased expression of one or more of ACP5, CCL3, CD38, CTSW, CXCL13, EIF2AK2, ENTPD1, EPSTI1, GNLY, HAVCR2, HERC5, IFI6, IFI44, IFI44L, IFIT1, ISG15, ISG20, LAG3, MX1, OASL, OAS1, OAS2, PARP14, PLSCR1, PTMS, RBPJ, RSAD2, STAT1, TOX, and / or XAF1 compared to a cell not expressing the exhaustion phenotype.

[0041] 30. The method of any of clauses 20-29, wherein the one or more immune cells are immune cell is an NK cell and the exhaustion phenotype comprises an increase in one or more of MTOR signaling, oxidative phosphorylation, and / or interferon-a signature compared to a cell not expressing the exhaustion phenotype.

[0042] 31. The method of any of clauses 20-30, wherein the immune cell is a CD4+ cell and the exhaustion phenotype comprises increased expression of one or more of LAG3, CTLA4, PDCD1, ENTPD1, HAVCR2, and / or TIGIT compared to a cell not expressing the exhaustion phenotype.

[0043] 32. The method of any of clauses 20-31, wherein the therapeutic composition is administered orally.

[0044] 33. The method of any of clauses 20-32, wherein the therapeutic composition is administered parenterally.

[0045] 34. The method of any of clauses 20-33, wherein the tumor is a squamous cell tumor.

[0046] 35. The method of any of clauses 20-34, wherein the patient has head and neck squamous cell carcinoma.

[0047] 36. A method of treating a patient having a tumor, comprising: administering, to a patient having one or more immune cells that does not express an exhaustion phenotype, a first therapeutic composition followed by a second therapeutic composition, wherein the first therapeutic composition comprises one or more compounds effective to prime the one or more immune cells to respond to one or more ICIs and the second therapeutic composition comprises one or more ICIs.

[0048] 37. The method of clause 36, wherein the one or more compounds effective to prime the one or more immune cells comprise one or more of a STING agonist and a TLR9 agonist.

[0049] 38. The method of clause 36 or clause 37, wherein the one or more compounds effective to prime the one or more immune cells comprise one or more of cGAMP, DMXAA, MSA-2, TAK-676, a CpG oligodeoxynucleotide, IMO-2055, and / or lefitolimod.

[0050] 39. The method of any of clauses 36-38, wherein the one or more ICIs comprise one or more PD-1 inhibitors, PD-L1 -inhibitors, CTLA-4 inhibitors, and / or LAG-3 inhibitors.

[0051] 40. The method of any of clauses 36-39, wherein the one or more ICIs comprise one or more antibodies that bind PD-1, PD-L1, CTLA-4, and / or LAG-3.

[0052] 41. The method of any of clauses 36-40, wherein the one or more ICIs comprise one or more of nivolumab, pembrolizumab, cemiplimab, atezolizumab, durvalumab, avelumab, ipilimumab, relatlimab, fianlimab, and / or leramlimab.

[0053] 42. The method of any of clauses 36-41, wherein the one or more ICIs comprise a PD-1 inhibitor and a LAG-3 inhibitor.

[0054] 43. The method of any of clauses 36-42, wherein the one or more ICIs comprise nivolumab and relatlimab.

[0055] 44. The method of any of clauses 36-43, wherein the one or more immune cells comprise one or more CD8+ cells, optionally CD8+ TIL cells.

[0056] 45. The method of any of clauses 36-44, wherein the one or more immune cells comprise one or more NK cells, optionally tumor-infiltrating NK cells.

[0057] 46. The method of any of clauses 36-45, wherein the one or more immune cells comprise one or more CD4+ cells, optionally CD4+ conventional T cells or CD4+ regulatory T cells.

[0058] 47. The method of any of clauses 36-46, wherein the first and / or second therapeutic compositions are administered orally.

[0059] 48. The method of any of clauses 36-47, wherein the first and / or second therapeutic compositions are administered parenterally.

[0060] 49. The method of any of clauses 36-48, wherein the tumor is a squamous cell tumor.

[0061] 50. The method of any of clauses 36-49, wherein the patient has head and neck squamous cell carcinoma.

[0062] 51. A method of treating a patient having a tumor, comprising: determining whether one or more immune cells from the patient express a type-I IFN response signature; and administering to the patient one or more therapeutic compositions, wherein: the therapeutic composition comprises either: one or more checkpoint inhibitors (ICIs) in an amount effective to treat the tumor when the one or more immune cells express the type-I IFN response signature; or one or more compounds effective to prime the one or more immune cells to respond to one or more ICIs when the one or more immune cells do not express the type-I interferon response signature.

[0063] 52. The method of clause 51, wherein the type-I interferon response signature comprises an increased expression of one or more of IFI44, OASL, CCL3, OAS1, IFIT1, ISG15, MX1, IFI6, LAG3, and / or IFI44L compared to a cell not expressing the type-I interferon response signature.

[0064] 53. The method of clause 51 or clause 52, wherein the one or more compounds effective to prime the one or more immune cells comprise one or more of a STING agonist and a TLR9 agonist.

[0065] 54. The method of any of clauses 51-53, wherein the one or more compounds effective to prime the one or more immune cells comprise one or more of cGAMP, DMXAA, MSA-2, TAK-676, a CpG oligodeoxynucleotide, IMO-2055, and / or lefitolimod.

[0066] 55. The method of any of clauses 51-54, wherein the one or more ICIs comprise one or more PD-1 inhibitors, PD-L1 -inhibitors, CTLA-4 inhibitors, and / or LAG-3 inhibitors.

[0067] 56. The method of any of clauses 51-55, wherein the one or more ICIs comprise one or more antibodies that bind PD-1, PD-L1, CTLA-4, and / or LAG-3.

[0068] 57. The method of any of clauses 51-56, wherein the one or more ICIs comprise one or more of nivolumab, pembrolizumab, cemiplimab, atezolizumab, durvalumab, avelumab, ipilimumab, relatlimab, fianlimab, and / or leramlimab.

[0069] 58. The method of any of clauses 51-57, wherein the one or more ICIs comprise a PD-1 inhibitor and a LAG-3 inhibitor.

[0070] 59. The method of any of clauses 51-58, wherein the one or more ICIs comprise nivolumab and relatlimab.

[0071] 60. The method of any of clauses 51-59, wherein the one or more immune cells comprise one or more CD8+ cells, optionally CD8+ TIL cells.

[0072] 61. The method of any of clauses 51-60, wherein the one or more immune cells comprise one or more NK cells, optionally tumor-infiltrating NK cells.

[0073] 62. The method of any of clauses 51-61, wherein the one or more immune cells comprise one or more CD4+ cells, optionally CD4+ conventional T cells or CD4+ regulatory T cells.

[0074] 63. The method of any of clauses 51-62, wherein the therapeutic composition is administered orally.

[0075] 64. The method of any of clauses 51-63, wherein the therapeutic composition is administered parenterally.

[0076] 65. The method of any of clauses 51-64, wherein the tumor is a squamous cell tumor.

[0077] 66. The method of any of clauses 51-65, wherein the patient has head and neck squamous cell carcinoma.

[0078] 67. A method of treating a patient having a tumor, comprising: administering, to a patient having one or more immune cells expressing a type-I interferon response signature, a therapeutic composition comprising one or more ICIs in an amount effective to treat the tumor.

[0079] 68. The method of clause 67, wherein the type-I interferon response signature comprises an increased expression of one or more of IFI44, OASL, CCL3, OAS1, IFIT1, ISG15, MX1, IFI6, LAG3, and / or IFI44L compared to a cell not expressing the type-I interferon response signature.

[0080] 69. The method of clause 67 or clause 68, wherein the one or more ICIs comprise one or more PD-1 inhibitors, PD-L1 -inhibitors, CTLA-4 inhibitors, and / or LAG-3 inhibitors.

[0081] 70. The method of any of clauses 67-69, wherein the one or more ICIs comprise one or more antibodies that bind PD-1, PD-L1, CTLA-4, and / or LAG-3.

[0082] 71. The method of any of clauses 67-70, wherein the one or more ICIs comprise one or more of nivolumab, pembrolizumab, cemiplimab, atezolizumab, durvalumab, avelumab, ipilimumab, relatlimab, fianlimab, and / or leramlimab.

[0083] 72. The method of any of clauses 67-71, wherein the one or more ICIs comprise a PD-1 inhibitor and a LAG-3 inhibitor.

[0084] 73. The method of any of clauses 67-72, wherein the one or more ICIs comprise nivolumab and relatlimab.

[0085] 74. The method of any of clauses 67-73, wherein the one or more immune cells comprise one or more CD8+ cells, optionally CD8+ TIL cells.

[0086] 75. The method of any of clauses 67-74, wherein the one or more immune cells comprise one or more NK cells, optionally tumor-infiltrating NK cells.

[0087] 76. The method of any of clauses 67-75, wherein the one or more immune cells comprise one or more CD4+ cells, optionally CD4+ conventional T cells or CD4+ regulatory T cells.

[0088] 77. The method of any of clauses 67-76, wherein the therapeutic composition is administered orally.

[0089] 78. The method of any of clauses 67-77, wherein the therapeutic composition is administered parenterally.

[0090] 79. The method of any of clauses 67-78, wherein the tumor is a squamous cell tumor.

[0091] 80. The method of any of clauses 67-79, wherein the patient has head and neck squamous cell carcinoma.

[0092] 81. A method of treating a patient having a tumor, comprising: administering, to a patient having one or more immune cells that does not express a type-I interferon response signature, a first therapeutic composition followed by a second therapeutic composition, wherein the first therapeutic composition comprises one or more compounds effective to prime the one or more immune cells to respond to one or more ICIs and the second therapeutic composition comprises one or more ICIs.

[0093] 82. The method of clause 81, wherein the type-I interferon response signature comprises an increased expression of one or more of IFI44, OASL, CCL3, OAS1, IFIT1, ISG15, MX1, IFI6, LAG3, and / or IFI44L compared to a cell not expressing the type-I interferon response signature.

[0094] 83. The method of clause 81 or clause 82, wherein the one or more compounds effective to prime the one or more immune cells comprise one or more of a STING agonist and a TLR9 agonist.

[0095] 84. The method of any of clauses 81-83, wherein the one or more compounds effective to prime the one or more immune cells comprise one or more of cGAMP, DMXAA, MSA-2, TAK-676, a CpG oligodeoxynucleotide, IMO-2055, and / or lefitolimod.

[0096] 85. The method of any of clauses 81-84, wherein the one or more ICIs comprise one or more PD-1 inhibitors, PD-L1 -inhibitors, CTLA-4 inhibitors, and / or LAG-3 inhibitors.

[0097] 86. The method of any of clauses 81-85, wherein the one or more ICIs comprise one or more antibodies that bind PD-1, PD-L1, CTLA-4, and / or LAG-3.

[0098] 87. The method of any of clauses 81-86, wherein the one or more ICIs comprise one or more of nivolumab, pembrolizumab, cemiplimab, atezolizumab, durvalumab, avelumab, ipilimumab, relatlimab, fianlimab, and / or leramlimab.

[0099] 88. The method of any of clauses 81-87, wherein the one or more ICIs comprise a PD-1 inhibitor and a LAG-3 inhibitor.

[0100] 89. The method of any of clauses 81-88, wherein the one or more ICIs comprise nivolumab and relatlimab.

[0101] 90. The method of any of clauses 81-89, wherein the one or more immune cells comprise one or more CD8+ cells, optionally CD8+ TIL cells.

[0102] 91. The method of any of clauses 81-90, wherein the one or more immune cells comprise one or more NK cells, optionally tumor-infiltrating NK cells.

[0103] 92. The method of any of clauses 81-91, wherein the one or more immune cells comprise one or more CD4+ cells, optionally CD4+ conventional T cells or CD4+ regulatory T cells.

[0104] 93. The method of any of clauses 81-92, wherein the first and / or second therapeutic compositions are administered orally.

[0105] 94. The method of any of clauses 81-93, wherein the first and / or second therapeutic compositions are administered parenterally.

[0106] 95. The method of any of clauses 81-94, wherein the tumor is a squamous cell tumor.

[0107] 96. The method of any of clauses 81 -95, wherein the patient has head and neck squamous cell carcinoma.

[0108] 97. Use of a composition comprising an immune checkpoint inhibitor to treat cancer in a patient having one or more immune cells expressing an exhaustion phenotype and / or a type-l interferon response signature.

[0109] 98. Use of a composition comprising compound effective to prime an immune cell to respond to a composition comprising an immune checkpoint inhibitor and an immune checkpoint inhibitor to treat cancer in a patient having one or more immune cells that do not express an exhaustion phenotype and / or a type-l interferon response signature.BRIEF DESCRIPTION OF THE DRAWINGS

[0110] FIGS. 1A-1D show (A) Waterfall plot of pathologic response for each patient shown in an ascending order and colored by response bins and cohorts. (B) Waterfall plot depicting patient-specific volumetric response in a descending order and colored by response bins and cohorts. (C) Disease-free survival by pathological tumor responses, Kaplan-Meier curves were generated and statistical comparisons were performed via log-rank tests. Results were not statistically significant. (D) Overall survival by pathological tumor responses, Kaplan-Meier curves were generated and statistical comparisons were performed via log-rank tests. Results were not statistically significant.

[0111] FIGS. 2A-2E show (A) Classification of CD3+ TIL from HNSCC and Uniform Manifold Approximation and Projection (UMAP) of the transcriptionally defined clusters derived from the scRNA-seq dataset. (B) Proportions of baseline CD3+ TIL sub-populations including CD4+ T cells, FOXP3+ T cells, CD8+ T cells, cycling T cells, y5 T and MAIT cells collected in tumors from pTR-0 (n = 11), pTR-1 (n = 8) and pTR-2 (n = 6) patients prior to treatments. Dots represent individual patients within each response category. Boxplots show the first to the third quartile with a line in the middle that represents the median, whiskers represent minimum and maximum values. Significant comparisons (p values) of response groups within individual cell population were calculated using a one-way ANOVA test. (C) Expression of PDCD1, LAG3, and CTLA4 by CD8+ TIL at baseline by scRNA-seq. The size of dots representsthe percentage of cells expressing the three genes, while the color scheme indicates the level of gene expression. (D) Heatmaps of the top 50 upregulated differentially expressed genes (DEG) expressed by baseline CD8+ TIL from pTR-2 patients compared to pTR-0 patients in Nivo+lpi and Nivo+Rela cohorts. The color scale indicates the Z score for each gene. (E) GSEA enrichment of baseline CD8+ TIL in pTR-2 versus pTR-0 patients from Nivo+lpi and Nivo+Rela arms (STAR Methods), p values were determined by one-tailed permutation test in GSEA. NES, normalized enrichment score.

[0112] FIGS. 3A-3E show (A) Quantification and comparison of post-treatment CD3+CD8+ TIL proportions in pTR-0 (n = 9), pTR-1 (n = 11 ) and pTR-2 (n = 6) patients based on scRNA-seq data. Dots represent individual patient proportions. Boxplots show the first to the third quartile with a line in the middle that represents the median, whiskers represent minimum and maximum values. Statistically significant comparisons of post-treatment CD3+CD8+ TIL are determined by a one-way ANOVA test. (B) Comparisons of paired baseline and post-treatment CD3+CD8+ TIL frequencies in Nivo, Nivo+lpi and Nivo+Rela patients based on scRNA-seq data. Left panel: pTR-0 (n = 6) in Nivo cohort; middle panel: pTR-0 (n = 3), pTR-1 (n = 1) and pTR-2 (n = 3) patients in the Nivo+lpi cohort; right panel: pTR-0 (n = 1), pTR-1 (n = 4) and pTR-2 (n = 3) in the Nivo+Rela cohort. Statistically significant comparisons are determined by a two-sided paired t-test. (C) Representative multiplex IF images of patient-matched baseline (top) and post-treatment (bottom) tumors in pTR-0 (left), pTR-1 (middle) and pTR-2 (right) patients. Epithelial cells were stained with pan-CK, CD8+ T cells were detected using CD3, CD8, and DAPI was used to identify nucleated cells (STAR Methods). Scale bar (bottom left, white bar), 100pm. (D) Comparisons of the changes in CD8+ TIL density across the whole (total) tumor region. CD8+ TIL densities within the whole tumor region are presented. Paired data collected from patients treated with Nivo (n = 9), Nivo+lpi (n = 9) and Nivo+Rela (n = 10) are shown. Dots represent individual patients. Colors represent the treatment received. Boxplots show the first to the third quartile with a line in the middle that represents the median, whiskers represent minimum and maximum values. Statistically significant comparisons (baseline vs. post) were determined by a two-sided paired t-test, and nominal p values are reported. (E) Correlation of baseline-to-post-treatment changes of CD3+CD8+ TIL densities in the whole tumor region to pathologic response in Nivo (n = 9), Nivo+lpi (n = 9), and Nivo+Rela (n = 10) arms are shown. Correlationcoefficients and p values were calculated using Spearman’s correlation between the change in CD3+CD8+ TIL density from baseline to post-treatment and pathologic responses. Nominal p values are reported.

[0113] FIGS. 4A-4G show (A) UMAP clustering and classification of CD8+ TIL from HNSCC of the transcriptionally defined subsets from the scRNA-seq data. Subsets were identified based on the expression of canonical markers and top DEG defining each cluster. (B) Bar graphs depict the proportion of CD8+ TIL subsets detected in tumors before and after neoadjuvant Nivo+lpi, and Nivo+Rela therapies. Each bar represents the cumulative number of patients within the same group at baseline or post-therapy. Only patients with matching baseline and post-treatment samples were included in the analysis. Colors identify CD8+ TIL subsets as defined in Figure 4A. (C) The comparison of the fraction of C04_TEM-ISGlow, C05_TEM-ISGhigh, C07_TRM-ICRhigh, and C08_TEX-ISGhigh CD8+ TIL subsets analyzed in pTR-2 patients’ tumors before and after Nivo+lpi (n = 3) and Nivo+Rela (n = 3) therapies. Dots represent individual patients, with lines connecting baseline-to-post-treatment-matched evaluations. Dot and line colors correspond to the treatment received. Statistically significant comparisons (baseline vs. post-treatment) are determined by a two-sided paired t-test. (D) Volcano plot showing the DEGs among C04_TEM-ISGlowcellsdetected in post-treatment tumors, comparing pTR-0 (n = 2) and pTR-2 (n = 3) patientstreated with Nivo+Ipi. Red dots denote genes with adjusted p-values < 0.001 (two-sided Wilcoxon rank-sum test) and an average LogFC >0.5 (upregulated in pTR-2) or an average LogFC < - 0.5 (downregulated in pTR-2). Representative top genes are labeled, with their gene names highlighted in black. (E) Volcano plot illustrating the DEGs among C04_TEM-ISGlowcells detected in post-treatment tumors, comparing pTR-0 (n = 2) and pTR-2 (n = 3) patients treated with Nivo+Rela. Red dots denote genes with adjusted p-values of <0.001 (two-sided Wilcoxon rank-sum test) and an average LogFC >0.5 (upregulated in pTR-2) or an average LogFC < -0.5 (downregulated in pTR-2). Representative top genes are labeled, with their gene names highlighted in black. (F) Unique and shared upregulated and downregulated DEGs among C04_TEM-ISGlow cells in pTR-2 patients compared to pTR-0 after Nivo+Ipi andNivo+Rela treatments. The Venn diagram is generated by counting the number of up-and down-regulated DEGs that are uniquely associated with or shared by pTR-2 patients from Nivo+lpi and Nivo+Rela cohorts. (G) List of all shared genes that are upregulated by pTR-2 versus pTR-0 patients in post-treatment C04_TEM-ISGlow cellsbetween Nivo+lpi (left) and Nivo+Rela (right) arms. The color bar represents the Z score for each gene.

[0114] FIGS. 5A-5E show (A) CD8+ TIL TCR diversity estimation before and after Nivo+lpi (pTR-0 = 3, pTR-2 = 3) and Nivo+Rela (pTR-0 = 1, pTR-1 = 4, pTR-2 = 3) using Inverse Simpson index. Boxplots show the first to the third quartile with a line in the middle that represents the median, whiskers represent minimum and maximum values. Statistically significant comparisons between baseline vs. post were determined by two-sided paired t-test. (B) TCR sharing between baseline and posttreatment CD8+ TIL subsets among pTR-2 patients treated with Nivo+lpi or Nivo+Rela, as calculated using the STARTRAC transition index (pTrans index). (C) Mapping of pre-existing TCR clonotypes detected in pTR-2 from Nivo+lpi (n = 3) and Nivo+Rela (n = 3) cohorts on the CD8+ TIL UMAP. Dense areas are highlighted by arrows and annotated with associated CD8+ TIL phenotypes present in baseline (left) and post-treatment (right) samples. (D) Slingshot analysis of CD8+ T cells derived from baseline and post-treatment tumor biopsies, revealing 3 trajectories: TEX-ischigh(solid), TRM-ICR|0W(long dashed), and TEX-ISG|0W(dashed). (E) Expression of IFN-I response, T cell exhaustion and T cell-mediated immune response gene programs across cells within TEM-ISG|0Wand TEX-ischighpopulations in pTR-2 from Nivo+lpi (n = 3) and Nivo+Rela (n = 3) cohorts. Boxplots in violin plots show the first to the third quartile with a line in the middle that represents the median, whiskers represent minimum and maximum values. Statistical comparisons were performed using a linear mixed effect model (LMM) and two-tailed Mann-Whitney U test.

[0115] FIGS. 6A-6F show (A) (Left panel) Proportion of TEM-ISG|0Wcells identified from an external cohort before and after neoadjuvant anti-PD-1+anti-CTLA-4 therapy. Color of boxes represent the pathological response groups. Two-sided paired t-test was performed when compare baseline-post paired samples and two-sided Wilcoxon rank-sum test was performed for comparing RE (n = 18) vs. NR (n = 12). (Right panel) Expression of TEM-ISG|0Wgene signatures in CD8+ TILs collected before and after anti-PD-1+anti-CTLA-4 from HNSCC cohort 1. Statistical comparisons were performed using LMM and two-tailed Mann-Whitney U test. (B) TCR sharing between baseline and post-treatment CD8+ TIL subsets in the NR and RE patients as calculated using the STARTRAC transition index (pTrans index). (C) Proportion of post-treatment CD8+ TIL subsets that share TCR clonotypes with baseline TRM-ICR|0Wcells. Statistical comparisons were performed using one-way ANOVA. (D) Proportion of posttreatmentCD8+ TIL subsets share TCR clonotypes with baseline TEM-isGhighcells. Statistical comparisons were performed using one-way ANOVA. (E) Proportion of posttreatment TEM -ISG|0Wand TRM cells identified from a cohort2 after neoadjuvant anti-PD-1 mono / or in combination with anti-CTLA-4 therapy(combo). Color of boxes represent the pathological response groups. Two-sided Wilcoxon rank-sum test was performed for comparing High vs. Low. (F) Expression of TRM and IFN-I response gene signatures in posttreatment CD8+ TILs collected from HNSCC cohort 2. Statistical comparisons were performed using LMM and two-tailed Mann-Whitney U test. Boxplots throughout the figure show the first to the third quartile with a line in the middle that represents the median, whiskers represent minimum and maximum values.

[0116] FIGS. 7A-7D show (A) Correlation between IFN-I gene signature score and LAG3 expression in the Nivo+Rela cohort at baseline and post-treatment. Dot colors represent response groups. Correlation coefficients and p values are calculated based on Spearman’s correlation between averaged IFN-I gene signature scores and averaged LAG3 expression. (B) Representative image for the m IF panel that evaluates CD3, CD8, CD39 and LAG-3 expression. Scale bar (bottom left), 100pm. (C) Quantification of CD3+CD8+CD39+LAG-3+ cell density in the CD3+CD8+ cell compartment from whole tumor regions of FFPE tissues collected from Nivo+lpi (left, 3 pTR-0 and 4 pTR-2) and Nivo+Rela (right, 4 pTR-0 and 3 pTR-2) cohorts. Statistical significances between baseline and post-treatment tumors were determined using two-sided paired t-test. (D) Quantification of CD3+CD8+CD39-LAG-3- cell densities in patients treated with Nivo+lpi (left, 3 pTR-0 and 4 pTR-2) and Nivo+Rela (right, 3 pTR-0 and3 pTR-2). Statistical significance was determined using the two-sided paired t-test. In (C) and (D), boxplots show the first to the third quartile with a line in the middle that represents the median, whiskers represent minimum and maximum values.

[0117] FIGS. 8A-8E show (A) GSEA of each CD8+ TIL subset to identify signatures of previously reported tumor-reactive CD8+ T cells. The color scale indicates the Z score for enrichment and dot sizes represent the log-transformed p value, p values were determined by a one-tailed permutation test. (B) Illustration demonstrating the definition and selection of putative tumor reactive CD8+ TIL clones. (C) Heatmap reporting average gene expression in putative tumor reactive CD8+ TIL from baseline or post-therapy tumors collected from pTR-2 and pTR-0 patients treated with Nivo+lpi or Nivo+Rela. Top upregulated genes in putative tumor reactive CD8+TIL from pTR-2 patients treated with Nivo+Rela are reported, along with representative T cell-related genes categorized by their known functions. Color scale represents the Z score. (D-E) Bidimensional plot depicting quantification of gene signature scores of exhaustion (y axis) and IFN-I response gene programs (x axis) in top expanding putative tumor-reactive TCR clones in tumors collected at baseline or post-treatment from patients treated with Nivo+lpi (pTR-0, n = 3; pTR-2, n = 3) and Nivo+Rela (pTR-0, n = 1; pTR-2, n = 3). Dots represent single putative tumor-reactive CD8+ TIL cells, and colors represent identical TCR clones. The size of dots represents the frequency of each TCR clone.

[0118] FIG. 9 shows upregulation of TIM-3 (HAVCR2) in non-responsive versus responsive HNSCC patients, primarily in an anti-PD-1 monotherapy-treated group.

[0119] FIG. 10 shows CTLA4, LAG3 and PDCD1 expression by baseline CD4+ conventional T cells (Tconv) and correlation to pathologic response.

[0120] FIG. 11 shows CTLA4, LAG3 and PDCD1 expression by baseline regulatory T cells (Treg) and correlation to pathologic response.

[0121] FIGS. 12A-12J show (A) Overall survival by groups based on baseline PD-L1 and LAG-3 IHC staining. Staining was evaluated in a binary format with negative and positive staining. Kaplan-Meier curves were generated, and statistical comparisons were performed via log-rank tests (n=38 patients). (B-D) Correlation of average mRNA expression of PDCD1, LAG3 and CTLA4 at baseline by Tconv with pathological response in each regimen. Dots represent baseline gene expression values averaged for each patient. Correlation coefficients and p-values were calculated based on Spearman’s rank correlation (Spearman’s rho (p)) in Nivo+lpi (n=8) and Nivo+Rela (n=12) arms. (E-G) Correlation of average mRNA expression of PDCD1, LAG3 and CTLA4 at baseline by Treg with pathological response in each regimen. Dots represent baseline gene expression values averaged for each patient. Correlation coefficients and p-values were calculated based on Spearman’s rank correlation (Spearman’s rho (p)) in Nivo+lpi (n=8) and Nivo+Rela (n=12) arms. (H-J) Overall survival by CTLA4 and LAG3 mRNA expression level. Patients were grouped into high vs low based on median baseline gene expression in Nivo+lpi treated Tconv and Treg (n=29 patients). Kaplan-Meier curves were generated and statistical comparisons were performed via log-rank tests.

[0122] FIGS. 13A-13D show (A) Heatmap of representative differentially expressed genes (DEG) between pTR-2 patients (n=3 Nivo+lpi, n=4 Nivo+Rela)compared to pTR-0 patients (n=3 Nivo+lpi, n=3 Nivo+Rela), grouped by differentially enriched Hallmark and previously published gene signatures at baseline in Tconv. The color scale indicates scaled gene expression values. (B) GSEA enrichment of baseline Tconv in pTR-2 versus pTR-0 patients from Nivo+lpi and Nivo+Rela arms, p-values were determined by one-tailed permutation test using the fgsea package. Dot size represents adjusted p-values and the color bar represents the normalized enrichment score (NES). (C) Heatmap of representative differentially expressed genes (DEG) between pTR-2 patients (n=3 Nivo+lpi, n=4 Nivo+Rela) compared to pTR-0 patients (n=3 Nivo+lpi, n=3 Nivo+Rela), grouped by differentially enriched Hallmark and previously published gene signatures at baseline in Treg. The color scale indicates scaled gene expression values. (D) GSEA enrichment of baseline Treg in pTR-2 versus pTR-0 patients from Nivo+lpi and Nivo+Rela arms, p-values were determined by one-tailed permutation test using the fgsea package. Dot size represents adjusted p-values and the color bar represents the normalized enrichment score (NES).

[0123] FIGS. 14A-14D show (A) Correlation of proportion change of TFH subsets (C08_TCF7hi _early_TFH and C10_TOXhi987 _TFH) from baseline and posttreatment samples to pathological response. Correlation coefficients and p-values were calculated based on Spearman’s rank correlation coefficients (Spearman’s rho (p)) in Nivo+lpi (n = 6) and Nivo+Rela (n = 9) arms. (B) GSEA enrichment posttreatment vs. baseline in tumor-infiltrating CD4+ Tconv of Nivo-lpi and Nivo+Rela pTR-2. p-values were determined by one-tailed permutation test using the fgsea package. Dot size represents adjusted p-values and the color bar represents the normalized enrichment score (NES). (C) Volcano plot comparing DEGs detected in Tconv posttreatment between Nivo-lpi pTR-0 and pTR-2. Red dots highlight representative significant DEGs with gene names labeled in black, p-values and logFC are calculated using the wilcox-limma test implemented in the Seurat package. (D) Volcano plot comparing DEGs detected in Tconv post-treatment between Nivo-Rela pTR-0 and pTR-2. Red dots highlight representative significant DEGs with gene names labeled in black, p-values and logFC are calculated using the wilcox-limma test implemented in the Seurat package.

[0124] FIGS. 15A-15D show (A) GSEA enrichment post-treatment vs. baseline in tumor-infiltrating CD4+ Treg of Nivo+lpi and Nivo+Rela pTR-2. p-values were determined by one-tailed permutation test using the fgsea package. Dot size represents adjusted p-values and the color bar represents the normalized enrichmentscore (NES). (B) Fraction of IFNGhi Treg analyzed in paired Nivo+lpi (n=3) and Nivo+Rela (n=3) pTR-2 and pTR-0 at baseline and post-treatment. Dots represent individual patients, with lines connecting baseline to post-treatment matched measurements. Box colors correspond to the treatment timepoint. Statistically significant comparisons (baseline vs. post treatment) are determined by a two-sided paired t-test. (C) Volcano plot comparing DEGs detected in Treg post-treatment between Nivo-lpi pTR-0 and pTR-2. Red dots highlight representative significant DEGs with gene names labeled in black, p-values and logFC are calculated using the wilcox-limma test implemented in the Seurat package. (D) Volcano plot comparing DEGs detected in Treg post-treatment between Nivo-Rela pTR-0 and pTR-2. Red dots highlight representative significant DEGs with gene names labeled in black, p-values and logFC are calculated using the wilcox-limma test implemented in the Seurat package.

[0125] FIGS. 16A-16C show (A) Density plot showing the distribution of tumorinfiltrating Tconv along trajectory 4 (ICR TH1) (lineage 4ICR-TH1, as calculated by Slingshot) at baseline and post-treatment in Nivo+lpi pTR-2 (left) Nivo+Rela pTR-2 (right). The color of the density plots represents the timepoint. Heatmaps (bottom panels) depict gene expression of selected genes along each trajectory. Black arrows indicate differences between baseline and post-treatment. (B) Barplot showing distribution of clones post-treatment by categories: newly detected clones (n=1, n>2) and pre-existing clones (n=1, n>2) across different Tconv subsets post treatment in Nivo+lpi (n=3) and Nivo+Rela pTR-2 (n=3) patients. (C) Heatmap highlighting representative genes that were differentially expressed between expanded Tconv clones and other Tconv for Nivo+lpi and Nivo+Rela pTR-2 patients. The color shows the scaled gene expression values.

[0126] FIGS. 17A-17C show (A) Density plot showing the distribution of tumorinfiltrating Treg along trajectory 1NFKBI active (lineage 1 NFKBI-active, as calculated by Slingshot) post-treatment in Nivo+Rela pTR-2. The color of the density plots represents the timepoint. Heatmap (bottom panel) depict gene expression of selected genes along each trajectory. Black arrows indicate differences between baseline and post-treatment. Density plot showing the distribution of tumor-infiltrating Treg along trajectory 3GIMAP (lineage 3GIMAP, as calculated by Slingshot) post-treatment in Nivo+Rela pTR-2. The color of the density plots represents the timepoint. Heatmap (bottom panel) shows gene expression of selected genes along each trajectory. Blackarrows indicate differences between baseline and post-treatment. (B) Barplot showing distribution of clones post-treatment by categories: newly detected clones (n=1, n>2) and pre-existing clones (n=1, n>2) across different Treg subsets post treatment in Nivo+lpi (n=3) and Nivo+Rela (n=3) pTR-2 patients. (C) Heatmap highlighting representative genes that were differentially expressed between expanded Treg clones and other Treg for Nivo+lpi and Nivo+Rela pTR-2 patients. The color shows the scaled gene expression values, p-values and logFC are calculated using the wilcox-limma test implemented in the Seurat package.DESCRIPTION OF THE INVENTION

[0127] The Other than in the operating examples, or where otherwise indicated, all numbers expressing quantities of ingredients, reaction conditions, and so forth used in the specification and claims are to be understood as being modified in all instances by the term "about". Accordingly, unless indicated to the contrary, the numerical parameters set forth in the following specification and attached claims are approximations that may vary depending upon the desired properties sought to be obtained by the present invention. At the very least, and not as an attempt to limit the application of the doctrine of equivalents to the scope of the claims, each numerical parameter should at least be construed in light of the number of reported significant digits and by applying ordinary rounding techniques.

[0128] Notwithstanding that the numerical ranges and parameters setting forth the broad scope of the invention are approximations, the numerical values set forth in the specific examples are reported as precisely as possible. Any numerical values, however, inherently contain certain errors necessarily resulting from the standard deviation found in their respective testing measurements. Furthermore, when numerical ranges of varying scope are set forth herein, it is contemplated that any combination of these values inclusive of the recited values may be used.

[0129] As used herein, the terms “comprising,” “comprise” or “comprised,” and variations thereof, in reference to elements of an item, composition, apparatus, method, process, system, claim etc. are intended to be open-ended, meaning that the item, composition, apparatus, method, process, system, claim etc. includes those elements and other elements can be included and still fall within the scope / definition of the described item, composition, apparatus, method, process, system, claim etc.As used herein, "a" or "an" means one or more. As used herein "another" may mean at least a second or more.

[0130] As used herein, the terms "patient" or "subject" refer to members of the animal kingdom, including, but not limited to human beings.

[0131] Unless otherwise explained, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs. Although methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present disclosure, suitable methods and materials are described below.

[0132] The terms “treat,” “treating”, “treatment,” “ameliorate” or “ameliorating” and other grammatical equivalents as used herein, can include alleviating, or abating a disease or condition symptoms, inhibiting a disease or condition, e.g., arresting the development of a disease or condition, relieving a disease or condition, causing regression of a disease or condition, relieving a condition caused by the disease or condition, or stopping symptoms of a disease or condition.

[0133] The term “preventing” can mean preventing additional symptoms, ameliorating or preventing the underlying metabolic causes of symptoms, and can include prophylaxis.

[0134] In some instances, “treat,” “treating”, “treatment,” “ameliorate” or “ameliorating” and other grammatical equivalents can include prophylaxis. “Treat,” “treating”, “treatment,” “ameliorate” or “ameliorating” and other grammatical equivalents can further include achieving a therapeutic benefit and / or a prophylactic benefit. Therapeutic benefit can mean eradication of the underlying disease being treated. Also, a therapeutic benefit can be achieved with the eradication of one or more of the physiological symptoms associated with the underlying disease such that an improvement can be observed in a subject notwithstanding that, in some embodiments, the subject can still be afflicted with the underlying disease.

[0135] The terms “effective amount”, “therapeutically effective amount” or “pharmaceutically effective amount” as used herein, can refer to a sufficient amount of a compound being administered which will at least partially ameliorate a symptom of a disease or condition being treated.

[0136] Therapeutic compositions, including those containing the compositions disclosed herein, may comprise a pharmaceutically acceptable carrier, or excipient. An excipient is an inactive substance used as a carrier for the active ingredients of amedication. Although "inactive," excipients may facilitate and aid in increasing the delivery or bioavailability of an active ingredient in a drug product. Non-limiting examples of useful excipients include: adjuvants, antiadherents, binders, rheology modifiers, carriers, coatings, disintegrants, emulsifiers, oils, buffers, salts, acids, bases, fillers, diluents, solvents, flavors, colorants, glidants, lubricants, preservatives, antioxidants, sorbents, vitamins, sweeteners, etc., as are available in the pharmaceutical / compounding arts.

[0137] Useful dosage forms for the compositions disclosed herein include, for example and without limitation: parenteral, intravenous, intramuscular, intraocular, or intraperitoneal solutions, oral tablets or liquids, topical drops, ointments, or creams, and transdermal devices (e.g., patches). The compound may be a sterile solution comprising the active ingredient, and a solvent, such as water, saline, lactated Ringer's solution, or phosphate-buffered saline (PBS). Additional excipients, such as polyethylene glycol, emulsifiers, salts, and buffers may be included in the solution. Suitable dosage forms may include single-dose, or multiple-dose vials or other containers, such as medical syringes or droppers, e.g., eye droppers. Pharmaceutical formulations adapted for administration include aqueous and non-aqueous sterile solutions which may contain, in addition to the active pharmaceutical ingredient or drug, for example and without limitation, adjuvants, anti-oxidants, buffers, bacteriostats, lipids, liposomes, lipid nanoparticles, emulsifiers, suspending agents, and rheology modifiers. The formulations may be presented in unit-dose or multi-dose containers, for example, sealed ampoules and vials, and may be stored in a freeze-dried (lyophilized) condition requiring only the addition of the sterile liquid carrier, for example, water for injections, immediately prior to use. Extemporaneous solutions and suspensions may be prepared.

[0138] Therapeutic / pharmaceutical compositions as described herein may be prepared in accordance with acceptable pharmaceutical procedures, such as described in Remington: The Science and Practice of Pharmacy, 21st edition, ed. Paul Beringer et al., Lippincott, Williams & Wilkins, Baltimore, MD Easton, Pa. (2005) (see, e.g., Chapters 39, 41, and 42 for examples of liquid, parenteral, and intravenous formulations and methods of making such formulations).

[0139] Therapeutic compositions typically must be sterile and stable under the conditions of manufacture and storage. For example, sterile injectable solutions can be prepared by incorporating the active agent in the required amount in an appropriatesolvent with one or a combination of ingredients enumerated herein, as required. Generally, dispersions are prepared by incorporating the active compound into a sterile vehicle that contains a basic dispersion medium and the required other ingredients from those enumerated above. In the case of sterile powders for the preparation of sterile injectable solutions, typical methods of preparation are vacuum drying and freeze-drying that yields a powder of the active ingredient plus any additional desired ingredient from a previously sterile-filtered solution thereof. The proper fluidity of a solution can be maintained, for example, by the use of a rheology modifier. Prolonged absorption of injectable compositions can be brought about by including in the composition an agent that delays absorption, for example, monostearate salts, gelatin or a hydrogel.

[0140] The phrase "pharmaceutically-acceptable carrier" as used herein means a pharmaceutically-acceptable material, composition or vehicle, such as a liquid or solid filler, diluent, excipient, manufacturing aid (e.g., lubricant, talc magnesium, calcium or zinc stearate, or steric acid), solvent, or encapsulating material, involved in carrying or transporting an active agent from one organ, or portion of the body, to another organ, or portion of the body. Each carrier must be "acceptable" in the sense of being compatible with the other ingredients of the formulation and not injurious to the subject being treated. Some examples of materials which can serve as pharmaceutically-acceptable carriers include: (1) sugars, such as lactose, glucose and sucrose; (2) starches, such as corn starch and potato starch; (3) cellulose, and its derivatives, such as sodium carboxymethyl cellulose, ethyl cellulose and cellulose acetate; (4) powdered tragacanth; (5) malt; (6) gelatin; (7) lubricating agents, such as magnesium state, sodium lauryl sulfate and talc; (8) excipients, such as cocoa butter and suppository waxes; (9) oils, such as peanut oil, cottonseed oil, safflower oil, sesame oil, olive oil, corn oil and soybean oil; (10) glycols, such as propylene glycol; (11) polyols, such as glycerin, sorbitol, mannitol and polyethylene glycol; (12) esters, such as ethyl oleate and ethyl laurate; (13) agar; (14) buffering agents, such as magnesium hydroxide and aluminum hydroxide; (15) alginic acid; (16) pyrogen-free water; (17) isotonic saline; (18) Ringer's solution; (19) ethyl alcohol; (20) pH buffered solutions; (21) polyesters, polycarbonates and / or polyanhydrides; (22) bulking agents, such as polypeptides and amino acids (23) serum component, such as serum albumin, HDL and LDL; and (22) other non-toxic compatible substances employed in pharmaceutical formulations. Remington: The Science and Practice of Pharmacy, The University ofthe Sciences in Philadelphia, Editor, Lippincott, Williams, & Wilkins, Philadelphia, Pa., 21st Edition (2005) (see above), describes compositions and formulations suitable for pharmaceutical delivery of one or more therapeutic compositions and additional pharmaceutical agents.

[0141] In general, the nature of the carrier will depend on the particular mode of administration being employed. For instance, parenteral formulations usually comprise injectable fluids that include pharmaceutically and physiologically acceptable fluids such as water, physiological saline, balanced salt solutions, aqueous dextrose, glycerol or the like as a vehicle. For solid compositions (for example, powder, pill, tablet, or capsule forms), conventional non-toxic solid carriers can include, for example, pharmaceutical grades of mannitol, lactose, starch, or magnesium stearate. In addition to biologically-neutral carriers, pharmaceutical compositions to be administered can contain minor amounts of non-toxic auxiliary substances, such as wetting or emulsifying agents, preservatives, and pH buffering agents and the like, for example sodium acetate or sorbitan monolaurate.

[0142] Compositions disclosed herein may be combined, and may be administered in any amount effective to treat the condition (e.g., cancer). Administration of a composition, pharmaceutically acceptable, or formulation can be performed for a treatment duration of at least about at least about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100 days consecutive or nonconsecutive days. In some embodiments, a treatment duration can be from about 1 to about 30 days, from about 2 to about 30 days, from about 3 to about 30 days, from about 4 to about 30 days, from about 5 to about 30 days, from about 6 to about 30 days, from about 7 to about 30 days, from about 8 to about 30 days, from about 9 to about 30 days, from about 10 to about 30 days, from about 11 to about 30 days, from about 12 to about 30 days, from about 13 to about 30 days, from about 14 to about 30 days, from about 15 to about 30 days, from about 16 to about 30 days, from about 17 to about 30 days, from about 18 to about 30 days, from about 19 to about 30 days, from about 20 to about 30 days, from about 21 to about 30 days, from about 22 to about 30 days, from about 23 to about 30 days, from about 24 to about 30 days, from about 25 to about 30 days, from about 26 to about 30days, from about 27 to about 30 days, from about 28 to about 30 days, or from about 29 to about 30 days. Treatment duration can be interrupted by one or more periods where treatment is not delivered.

[0143] Administration of a composition, pharmaceutically acceptable salt, or formulation can be performed at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, or 24 times a day. In some embodiments, administration of a composition, pharmaceutically acceptable salt, or formulation can be performed at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or 21 times a week. In some embodiments, administration of a composition, pharmaceutically acceptable salt, or formulation can be performed at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, or 90 times a month.

[0144] In some embodiments, administration of the pharmaceutical formulation comprising a composition or pharmaceutically acceptable salt occurs over a time period of from at least about 0.5 min to at least about 1 min, from at least about 1 min to at least about 2 min, from at least about 2 min to at least about 3 min, from at least about 3 min to at least about 4 min, from at least about 4 min to at least about 5 min, from at least about 5 min to at least about 6 min, from at least about 6 min to at least about 7 min, from at least about 7 min to at least about 8 min, from at least about 8 min to at least about 9 min, from at least about 9 min to at least about 10 min, from at least about 10 min to at least about 11 min, from at least about 11 min to at least about 12 min, from at least about 12 min to at least about 13 min, from at least about 13 min to at least about 14 min, from at least about 14 min to at least about 15 min, from at least about 15 min to at least about 16 min, from at least about 16 min to at least about 17 min, from at least about 17 min to at least about 18 min, from at least about 18 min to at least about 19 min, from at least about 19 min to at least about 20 min, from at least about 21 min to at least about 22 min, from at least about 22 min to at least about 23 min, from at least about 23 min to at least about 24 min, from at least about 24 min to at least about 25 min, from at least about 25 min to at least about 26 min, from at least about 26 min to at least about 27 min, from at least about 27 min to at least about 28 min, from at least about 28 min to at least about 29 min, or from at least about 29 min to at least about 30 min.

[0145] In some embodiments, a composition, pharmaceutically acceptable salt thereof, or pharmaceutical formulation comprising a composition or salt thereof described herein can be administered at a dose of from about 1 milligram (mg) to about 1000 mg, from about 5 mg to about 1000 mg, from about 10 mg to about 1000 mg, from about 15 mg to about 1000 mg, from about 20 mg to about 1000 mg, from about 25 mg to about 1000 mg, from about 30 mg to about 1000 mg, from about 35 mg to about 1000 mg, from about 40 mg to about 1000 mg, from about 45 mg to about 1000 mg, from about 50 mg to about 1000 mg, from about 55 mg to about 1000 mg, from about 60 mg to about 1000 mg, from about 65 mg to about 1000 mg, from about 70 mg to about 1000 mg, from about 75 mg to about 1000 mg, from about 80 mg to about 1000 mg, from about 85 mg to about 1000 mg, from about 90 mg to about 1000 mg, from about 95 mg to about 1000 mg, from about 100 mg to about 1000 mg, from about 150 mg to about 1000 mg, from about 200 mg to about 1000 mg, from about 250 mg to about 1000 mg, from about 300 mg to about 1000 mg, from about 350 mg to about 1000 mg, from about 400 mg to about 1000 mg, from about 450 mg to about 1000 mg, from about 500 mg to about 1000 mg, from about 550 mg to about 1000 mg, from about 600 mg to about 1000 mg, from about 650 mg to about 1000 mg, from about 700 mg to about 1000 mg, from about 750 mg to about 1000 mg, from about 800 mg to about 1000 mg, from about 850 mg to about 1000 mg, from about 900 mg to about 1000 mg, or from about 950 mg to about 1000 mg. In some embodiments, a composition, pharmaceutically acceptable salt thereof, or pharmaceutical formulation comprising a composition or salt thereof described herein can be administered at a dose of about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179 180, 181, 182, 183, 184, 184, 186, 187, 188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 210, 220, 230, 240, 250, 260, 270, 280, 290, 300, 310, 320, 330, 340, 350, 360, 370, 380, 390, 400, 410, 420, 430, 440, 450, 460, 470, 480, 490,500, 510, 520, 530, 540, 550, 560, 570, 580, 590, 600, 610, 620, 630, 640, 650, 660, 670, 680, 690, 700, 710, 720, 730, 740, 750, 760, 770, 780, 790, 800, 810, 820, 830, 840, 850, 860, 870, 880, 890, 900, 910, 920, 930, 940, 950, 960, 970, 980, 990, or 1000 mg.

[0146] In some embodiments, a pharmaceutical formulation comprising a composition or pharmaceutically acceptable salt is present at a concentration from at least about 0.01 micrograms per milliliter (µg / mL) to at least about 100 milligrams per milliliter (mg / mL). In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration from at least about at least about 0.1 mg / mL to at least about 5 mg / mL. In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration from at least about at least about 0.5 mg / mL to at least about 1 mg / mL. In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration about 1 mg / mL. In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration about 2 mg / mL. In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration about 3 mg / mL. In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration about 4 mg / mL. In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration about 5 mg / mL. In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration about 6 mg / mL. In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration about 7 mg / mL. In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration about 8 mg / mL. In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration about 9 mg / mL. In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration about 10 mg / mL. In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration about 20 mg / mL. In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration about 30 mg / mL. In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration about 40 mg / mL. In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration about 50 mg / mL. In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration about 60 mg / mL. Insome embodiments, the composition or pharmaceutically acceptable salt is present at a concentration about 70 mg / mL. In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration about 80 mg / mL. In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration about 90 mg / mL. In some embodiments, the composition or pharmaceutically acceptable salt is present at a concentration about 100 mg / mL.

[0147] In some embodiments, a pharmaceutical formulation comprising a composition or pharmaceutically acceptable salt can exhibit activity at a concentration from at least about 0.01 µg / mL to at least about 0.02 µg / mL, from at least about 0.02 µg / mL to at least about 0.03 µg / mL, from at least about 0.03 µg / mL to at least about 0.04 µg / mL, from at least about 0.04 µg / mL to at least about 0.05 µg / mL, from at least about 0.05 µg / mL to at least about 0.06 µg / mL, from at least about 0.06 µg / mL to at least about 0.07 µg / mL, from at least about 0.07 µg / mL to at least about 0.08 µg / mL, from at least about 0.08 µg / mL to at least about 0.09 µg / mL, from at least about 0.09 µg / mL to at least about 0.1 µg / mL, from at least about 0.1 µg / mL to at least about 0.2 µg / mL, from at least about 0.2 µg / mL to at least about 0.3 µg / mL, from at least about 0.3 µg / mL to at least about 0.4 µg / mL, from at least about 0.4 µg / mL to at least about 0.5 µg / mL, from at least about 0.5 µg / mL to at least about 0.6 µg / mL, from at least about 0.6 µg / mL to at least about 0.7 µg / mL, from at least about 0.7 µg / mL to at least about 0.8 µg / mL, from at least about 0.8 µg / mL to at least about 0.9 µg / mL, from at least about 0.9 µg / mL to at least about 1 µg / mL, from at least about 1 µg / mL to at least about 2 µg / mL, from at least about 2 µg / mL to at least about 3 µg / mL, from at least about 3 µg / mL to at least about 4 µg / mL, from at least about 4 µg / mL to at least about 5 µg / mL, from at least about 5 µg / mL to at least about 6 µg / mL, from at least about 6 µg / mL to at least about 7 µg / mL, from at least about 7 µg / mL to at least about 8 µg / mL, from at least about 8 µg / mL to at least about 9 µg / mL, from at least about 9 µg / mL to at least about 10 µg / mL, from at least about 10 µg / mL to at least about 20 µg / mL, from at least about 20 µg / mL to at least about 30 µg / mL, from at least about 30 µg / mL to at least about 40 µg / mL, from at least about 40 µg / mL to at least about 50 µg / mL, from at least about 50 µg / mL to at least about 60 µg / mL, from at least about 60 µg / mL to at least about 70 µg / mL, from at least about 70 µg / mL to at least about 80 µg / mL, from at least about 80 µg / mL to at least about 90 µg / mL, from at least about 90 µg / mL to at least about 0.1 mg / mL, from at least about 0.1 mg / mL to at least about 0.2 mg / mL, from at least about 0.2 mg / mL to at least about 0.3 mg / mL, from at least about 0.3mg / mL to at least about 0.4 mg / mL, from at least about 0.4 mg / mL to at least about 0.5 mg / mL, from at least about 0.5 mg / mL to at least about 0.6 mg / mL, from at least about 0.6 mg / mL to at least about 0.7 mg / mL, from at least about 0.7 mg / mL to at least about 0.8 mg / mL, from at least about 0.8 mg / mL to at least about 0.9 mg / mL, from at least about 0.9 mg / mL to at least about 1 mg / mL, from at least about 1 mg / mL to at least about 2 mg / mL, from at least about 2 mg / mL to at least about 3 mg / mL, from at least about 3 mg / mL to at least about 4 mg / mL, from at least about 4 mg / mL to at least about 5 mg / mL, from at least about 5 mg / mL to at least about 6 mg / mL, from at least about 6 mg / mL to at least about 7 mg / mL, from at least about 7 mg / mL to at least about 8 mg / mL, from at least about 8 mg / mL to at least about 9 mg / mL, from at least about 9 mg / mL to at least about 10 mg / mL, from at least about 10 mg / mL to at least about 20 mg / mL, from at least about 20 mg / mL to at least about 30 mg / mL, from at least about 30 mg / mL to at least about 40 mg / mL, from at least about 40 mg / mL to at least about 50 mg / mL, from at least about 50 mg / mL to at least about 60 mg / mL, from at least about 60 mg / mL to at least about 70 mg / mL, from at least about 70 mg / mL to at least about 80 mg / mL, from at least about 80 mg / mL to at least about 90 mg / mL, or from at least about 90 mg / mL to at least about 100 mg / mL.

[0148] In some embodiments, effective amounts of a composition or pharmaceutically acceptable salt can be a concentration from at least about 0.01 µg / mL to at least about 100 mg / mL. In some embodiments, effective amounts of a composition or pharmaceutically acceptable salt is at a concentration from at least about at least about 0.1 mg / mL to at least about 5 mg / mL. In some embodiments, effective amounts of a composition or pharmaceutically acceptable salt is at a concentration from at least about at least about 0.5 mg / mL to at least about 1 mg / mL.

[0149] In some embodiments, effective amounts of a composition or pharmaceutically acceptable salt is at a concentration about 1 mg / mL. In some embodiments, the effective amount of a composition or pharmaceutically acceptable salt is at a concentration of about 2 mg / mL. In some embodiments, the effective amount of a composition or pharmaceutically acceptable salt is at a concentration of about 3 mg / mL. In some embodiments, the effective amount of a composition or pharmaceutically acceptable salt is at a concentration of about 4 mg / mL. In some embodiments, the effective amount of a composition or pharmaceutically acceptable salt is at a concentration of about 5 mg / mL. In some embodiments, the effective amount of a composition or pharmaceutically acceptable salt is at a concentration ofabout 6 mg / mL. In some embodiments, the effective amount of a composition or pharmaceutically acceptable salt is at a concentration of about 7 mg / mL. In some embodiments, the effective amount of a composition or pharmaceutically acceptable salt is at a concentration of about 8 mg / mL. In some embodiments, the effective amount of a composition or pharmaceutically acceptable salt is at a concentration of about 9 mg / mL. In some embodiments, the effective amount of a composition or pharmaceutically acceptable salt is at a concentration of about 10 mg / mL.

[0150] In some embodiments, an effective amount of a composition or pharmaceutically acceptable salt may be a concentration from at least about 0.01 µg / mL to at least about 0.02 µg / mL, from at least about 0.02 µg / mL to at least about 0.03 µg / mL, from at least about 0.03 µg / mL to at least about 0.04 µg / mL, from at least about 0.04 µg / mL to at least about 0.05 µg / mL, from at least about 0.05 µg / mL to at least about 0.06 µg / mL, from at least about 0.06 µg / mL to at least about 0.07 µg / mL, from at least about 0.07 µg / mL to at least about 0.08 µg / mL, from at least about 0.08 µg / mL to at least about 0.09 µg / mL, from at least about 0.09 µg / mL to at least about 0.1 µg / mL, from at least about 0.1 µg / mL to at least about 0.2 µg / mL, from at least about 0.2 µg / mL to at least about 0.3 µg / mL, from at least about 0.3 µg / mL to at least about 0.4 µg / mL, from at least about 0.4 µg / mL to at least about 0.5 µg / mL, from at least about 0.5 µg / mL to at least about 0.6 µg / mL, from at least about 0.6 µg / mL to at least about 0.7 µg / mL, from at least about 0.7 µg / mL to at least about 0.8 µg / mL, from at least about 0.8 µg / mL to at least about 0.9 µg / mL, from at least about 0.9 µg / mL to at least about 1 µg / mL, from at least about 1 µg / mL to at least about 2 µg / mL, from at least about 2 µg / mL to at least about 3 µg / mL, from at least about 3 µg / mL to at least about 4 µg / mL, from at least about 4 µg / mL to at least about 5 µg / mL, from at least about 5 µg / mL to at least about 6 µg / mL, from at least about 6 µg / mL to at least about 7 µg / mL, from at least about 7 µg / mL to at least about 8 µg / mL, from at least about 8 µg / mL to at least about 9 µg / mL, from at least about 9 µg / mL to at least about 10 µg / mL, from at least about 10 µg / mL to at least about 20 µg / mL, from at least about 20 µg / mL to at least about 30 µg / mL, from at least about 30 µg / mL to at least about 40 µg / mL, from at least about 40 µg / mL to at least about 50 µg / mL, from at least about 50 µg / mL to at least about 60 µg / mL, from at least about 60 µg / mL to at least about 70 µg / mL, from at least about 70 µg / mL to at least about 80 µg / mL, from at least about 80 µg / mL to at least about 90 µg / mL, from at least about 90 µg / mL to at least about 0.1 mg / mL, from at least about 0.1 mg / mL to at least about 0.2 mg / mL, from at leastabout 0.2 mg / mL to at least about 0.3 mg / mL, from at least about 0.3 mg / mL to at least about 0.4 mg / mL, from at least about 0.4 mg / mL to at least about 0.5 mg / mL, from at least about 0.5 mg / mL to at least about 0.6 mg / mL, from at least about 0.6 mg / mL to at least about 0.7 mg / mL, from at least about 0.7 mg / mL to at least about 0.8 mg / mL, from at least about 0.8 mg / mL to at least about 0.9 mg / mL, from at least about 0.9 mg / mL to at least about 1 mg / mL, from at least about 1 mg / mL to at least about 2 mg / mL, from at least about 2 mg / mL to at least about 3 mg / mL, from at least about 3 mg / mL to at least about 4 mg / mL, from at least about 4 mg / mL to at least about 5 mg / mL, from at least about 5 mg / mL to at least about 6 mg / mL, from at least about 6 mg / mL to at least about 7 mg / mL, from at least about 7 mg / mL to at least about 8 mg / mL, from at least about 8 mg / mL to at least about 9 mg / mL, from at least about 9 mg / mL to at least about 10 mg / mL, from at least about 10 mg / mL to at least about 20 mg / mL, from at least about 20 mg / mL to at least about 30 mg / mL, from at least about 30 mg / mL to at least about 40 mg / mL, from at least about 40 mg / mL to at least about 50 mg / mL, from at least about 50 mg / mL to at least about 60 mg / mL, from at least about 60 mg / mL to at least about 70 mg / mL, from at least about 70 mg / mL to at least about 80 mg / mL, from at least about 80 mg / mL to at least about 90 mg / mL, or from at least about 90 mg / mL to at least about 100 mg / mL.

[0151] In some embodiments, an effective amount of a composition or pharmaceutically acceptable salt may be from at least about 1 microliter (µL) to at least about 2 µL, from at least about 2 µL to at least about 3 µL, from at least about 3 µL to at least about 4 µL, from at least about 4 µL to at least about 5 µL, from at least about 5 µL to at least about 6 µL, from at least about 6 µL to at least about 7 µL, from at least about 7 µL to at least about 8 µL, from at least about 8 µL to at least about 9 µL, from at least about 9 µL to at least about 10 µL, from at least about 10 µL to at least about 20 µL, from at least about 20 µL to at least about 30 µL, from at least about 30 µL to at least about 40 µL, from at least about 40 µL to at least about 50 µL, from at least about 50 µL to at least about 60 µL, from at least about 60 µL to at least about 70 µL, from at least about 70 µL to at least about 80 µL, from at least about 80 µL to at least about 90 µL, from at least about 90 µL to at least about 100 µL, from at least about 100 µL to at least about 200 µL, from at least about 200 µL to at least about 300 µL, from at least about 300 µL to at least about 400 µL, from at least about 400 µL to at least about 500 µL, from at least about 500 µL to at least about 600 µL, from at least about 600 µL to at least about 700 µL, from at least about 700 µL to at least about 800pL, from at least about 800 pL to at least about 900 pL, from at least about 900 pL to at least about 1 milliliter (mL), from at least about 1 mL to at least about 2 mL, from at least about 2 mL to at least about 3 mL, from at least about 3 mL to at least about 4 mL, from at least about 4 mL to at least about 5 mL, from at least about 5 mL to at least about 6 mL, from at least about 6 mL to at least about 7 mL, from at least about 7 mL to at least about 8 mL, from at least about 8 mL to at least about 9 mL, from at least about 9 mL to at least about 10 mL, from at least about 10 mL to at least about 20 mL, from at least about 20 mL to at least about 30 mL, from at least about 30 mL to at least about 40 mL, from at least about 40 mL to at least about 50 mL, from at least about 50 mL to at least about 60 mL, from at least about 60 mL to at least about 70 mL, from at least about 70 mL to at least about 80 mL, from at least about 80 mL to at least about 90 mL, from at least about 90 mL to at least about 100 mL, from at least about 100 mL to at least about 200 mL, from at least about 200 mL to at least about 300 mL, from at least about 300 mL to at least about 400 mL, from at least about 400 mL to at least about 500 mL, from at least about 500 mL to at least about 600 mL, from at least about 600 mL to at least about 700 mL, from at least about 700 mL to at least about 800 mL, from at least about 800 mL to at least about 900 mL, from at least about 900 mL to at least about 1 liter (L), from at least about 1 L to at least about 2 L, from at least about 2 L to at least about 3 L, from at least about 3 L to at least about 4 L, from at least about 4 L to at least about 5 L, from at least about 5 L to at least about 6 L, from at least about 6 L to at least about 7 L, from at least about 7 L to at least about 8 L, from at least about 8 L to at least about 9 L, from at least about 9 L to at least about 10 L, from at least about 10 L to at least about 20 L, from at least about 20 L to at least about 30 L, from at least about 30 L to at least about 40 L, from at least about 40 L to at least about 50 L, from at least about 50 L to at least about 60 L, from at least about 60 L to at least about 70 L, from at least about 70 L to at least about 80 L, from at least about 80 L to at least about 90 L, from at least about 90 L to at least about 100 L, from at least about 100 L to at least about 200 L, from at least about 200 L to at least about 300 L, from at least about 300 L to at least about 400 L, from at least about 400 L to at least about 500 L, from at least about 500 L to at least about 600 L, from at least about 600 L to at least about 700 L, from at least about 700 L to at least about 800 L, from at least about 800 L to at least about 900 L, from at least about 900 L to at least about 1 kiloliter (kL), from at least about 1 kL to at least about 2 kL, from at least about 2 kL to at least about 3 kL, from at least about 3 kL to at leastabout 4 kL, from at least about 4 kL to at least about 5 kL, from at least about 5 kL to at least about 6 kL, from at least about 6 kL to at least about 7 kL, from at least about 7 kL to at least about 8 kL, from at least about 8 kL to at least about 9 kL, or from at least about 9 kL to at least about 10 kL.

[0152] Those of skill in the art will appreciate that any dosing regimen (including frequency and dose) described herein includes all subranges and individual values between the disclosed ranges.

[0153] Provided herein are methods of treating a condition, for example a cancer, by identifying, based on a gene expression signature, that a treatment is or is not likely to be effective. Methods disclosed herein leverage gene expression patterns in immune cells associated with a tumor (e.g., tumor-infiltrating lymphocytes (TIL)) as an indicator of the class of drug treatment that should be utilized to provide the most efficacious outcome.

[0154] Accordingly, in non-limiting embodiments, a method of treating a patient having cancer (e.g., a tumor) may include determining whether one or more cells, for example one or more immune cells (such as, without limitation, one or more immune cells associated with the tumor) from the patient express an exhaustion phenotype and / or a type-l interferon response signature. In non-limiting embodiments, the one or more cells may be one or more tumor-infiltrating lymphocytes (TILs), for example, and without limitation, one or more cytotoxic T lymphocytes (CTLs), helper T lymphocytes (Th cells), Natural Killer (NK) cells, B cells, CD8+ cells, and / or CD4+ cells. In non-limiting embodiments, one or more cells may be a plurality of cells (e.g., two or more). In non-limiting embodiments, the one or more cells may be one or more cells within the tumor. In non-limiting embodiments, the cancer may be a squamous cell cancer. In non-limiting embodiments, the cancer may be a head and neck cancer. In non-limiting embodiments, the cancer may be a head and neck squamous cell carcinoma. In non-limiting embodiments, the cancer may be a melanoma.

[0155] In non-limiting embodiments, the one or more immune cells may be one or more CD4+ cells, one or more CD8+ cells, and / or one or more natural killer (NK) cells. The one or more immune cells may be obtained from a patient and isolated using known techniques, for example fluorescence-activated cell sorting (FACS), fluorescent dyes, immunomagnetic sorting, microfluidics, and the like, for example as described in Li et al., “Distinct CD8+ T cell dynamics associate with response to neoadjuvant cancer immunotherapies,” Cancer Cell 2025, 43(4): 757-775,incorporated herein by reference in its entirety. In non-limiting embodiments, the cells are obtained from a sample taken from the patient, for example from a tumor or other tissue. In non-limiting embodiments, the one or more immune cells are one or more of a CD8+ cell, a CD8+ TIL, an NK cell, an NK TIL, a CD4+ cell, a CD4+ TIL, a CD4+ conventional T cell, and / or a CD4+ regulatory T cell.

[0156] In non-limiting embodiments, the one or more immune cells are genotypes and / or phenotyped, for example using techniques known to those of skill in the art and include nucleic acid sequencing (including next generation sequencing), polymerase chain reaction (PCR) techniques (including quantitative PCR and allele-specific PCR), blotting (e.g,, Southern Blot, Northern Blot, Western Blot), restriction fragment length polymorphism (RFLP) techniques, enzyme-linked immunosorbent assays (ELISA), microarrays, flow cytometry, single-cell multiomics, RNA sequencing and / or protein expression (e.g, by cellular indexing of transcriptomes and epitopes by sequencing or ATAC-seq), immunohistochemistry, imaging techniques (including high-content imaging), tomographic techniques (including soft x-ray tomography), microscopy techniques (including correlative microscopy), spatialomics (RNA or protein measures by digital spatial profiling (DSP), CosMX, Visium, Visium hd), ChlP-seq (histone marks), WGBS (DNA methylation), ATAC-seq (chromatin accessibility), and the like, for example as described in Li et al., “Distinct CD8+ T cell dynamics associate with response to neoadjuvant cancer immunotherapies,” Cancer Cell 2025, 43(4): 757-775, incorporated herein by reference in its entirety. In non-limiting embodiments, the cells are genotyped and / or phenotyped to determine whether the one or more immune cells exhibit an exhaustion phenotype and / or a type-l interferon response signature.

[0157] By “exhaustion phenotype” it is meant that the cells exhibit an increase (e.g., a statistically significant increase) in expression (e.g., an increase in transcript and / or protein levels) of one or more genes associated with functional impairment and / or dysfunction in an immune cell, for example a T cell, compared to a cell that does not exhibit an exhaustion phenotype, a cell that is obtained from a patient that does not have cancer (e.g., a tumor), and / or an immune cell from the patient, where the immune cell is not associated with a tumor. In non-limiting embodiments the exhaustion phenotype includes an increase in one or more of ACP5, CCL3, CD38, CTSW, CXCL13, EIF2AK2, ENTPD1, EPSTI1, GNLY, HAVCR2, HERC5, IFI6, IFI44, IFI44L, IFIT1, ISG15, ISG20, LAG3, MX1, OASL, OAS1, OAS2, PARP14, PLSCR1, PTMS, RBPJ, RSAD2, STAT1, TOX, XAF1, mTOR (e.g., increased mTOR signaling),oxidative phosphorylation, interferon-β signature (e.g., increased expression of IFN-β), interferon-α signature (e.g, increased expression of IFN- α), CTLA4, PDCD1, and / or TIGIT. In non-limiting embodiments, the phenotype may include an increase in a plurality of any of the foregoing, in any combination.

[0158] By “type-l interferon response signature” it is meant that the cells exhibit an increase (e.g., a statistically significant increase) in expression (e.g., an increase in transcript and / or protein levels) of one or more of IFI44, OASL, CCL3, OAS1, IFIT1, ISG15, MX1, IFI6, LAG3, and / or IFI44L (In non-limiting embodiments, the signature may include an increase in a plurality of any of the foregoing, in any combination) compared to a cell that does not exhibit a type-l interferon response signature, a cell that is obtained from a patient that does not have cancer (e.g., a tumor), and / or an immune cell from the patient, where the immune cell is not associated with a tumor. In non-limiting embodiments, a type-l interferon response signature is an increase in expression of IFI44, OASL, CCL3, OAS1, IFIT1, ISG15, MX1, IFI6, LAG3, and IFI44L.

[0159] In non-limiting embodiments, the one or more immune cells is a CD8+ cell and the exhaustion phenotype and / or type-l interferon response signature is an increase in one or more of ACP5, CCL3, CD38, CTSW, CXCL13, EIF2AK2, ENTPD1, EPSTI1, GNLY, HAVCR2, HERC5, IFI6, IFI44, IFI44L, IFIT1, ISG15, ISG20, LAG3, MX1, OASL, OAS1, OAS2, PARP14, PLSCR1, PTMS, RBPJ, RSAD2, STAT1, TOX, and / or XAF1. In non-limiting embodiments, the phenotype may include an increase in a plurality of any of the foregoing, in any combination. In non-limiting embodiments, the one or more cells are one or more CD4+ cells and the exhaustion phenotype and / or type-l interferon response signature is an increase in expression of each of IFI44, OASL, CCL3, IFIT1, ISG15, MX1, IFI6, LAG3, and IFI4L. In non-limiting embodiments, the one or more immune cells are one or more CD8+ cells and the exhaustion phenotype includes an increase in expression of each of ACP5, CCL3, CD38, CTSW, CXCL13, EIF2AK2, ENTPD1, EPSTI1, GNLY, HAVCR2, HERC5, IFI6, IFI44, IFI44L, IFIT1, ISG15, ISG20, LAG3, MX1, OASL, OAS1, OAS2, PARP14, PLSCR1, PTMS, RBPJ, RSAD2, STAT1, TOX, and XAF1.

[0160] In non-limiting embodiments, the one or more immune cells are one or more NK cells and the exhaustion phenotype includes an increase in one or more of mTOR (e.g., increased mTOR signaling), increased oxidative phosphorylation, increased interferon-[3 signature (e.g., increased IFN-|3 expression), and / or interferon-a signature (e.g., increased IFN-a expression). In non-limiting embodiments, the oneor more immune cells are one or more NK cells and the exhaustion phenotype includes an increase in one or more of IFN-β expression and IFN-α expression. In non-limiting embodiments, the one or more immune cells are one or more NK cells and the cells do not exhibit an increase in expression of HAVCR2.

[0161] In non-limiting embodiments, the one or more immune cells are one or more CD4+ cells, and the exhaustion phenotype includes an increase in one or more of LAG3, CTLA4, PDCD1, ENTPD1, HAVCR2, and / or TIGIT (in non-limiting embodiments, the phenotype may include an increase in a plurality of any of the foregoing, in any combination). In non-limiting embodiments, the one or more immune cells are one or more CD4+ cells and the exhaustion phenotype includes an increase in each of LAG3, CTLA4, PDCD1, ENTPD1, HAVCR2, and TIGIT.

[0162] In non-limiting embodiments, a sample obtained from a patient may include one or more of any of CD4+ cells, CD8+ cells, and / or NK cells, and the exhaustion phenotype and / or type-l interferon response signature may be determined by genotyping and / or phenotyping any one or more of the CD4+, CD8+, and / or NK cells, alone or in combination.

[0163] With continuing reference to a method of treating a patient, a method may include administering to the patient one or more therapeutic compositions. In nonlimiting embodiments, for exam pie where one or more immune cells of the patient have been determined to exhibit an exhaustion phenotype and / or a type-l interferon response signature, the therapeutic composition may include one or more checkpoint inhibitors (ICIs) in an amount effective to treat the cancer (e.g., tumor). In non-limiting embodiments, the one or more ICIs may be one or more PD-1 inhibitors, PD-L1-inhibitors, CTLA-4 inhibitors, LAG-3 inhibitors, TIM-3 inhibitors, and / or TIGIT inhibitors. In non-limiting embodiments the ICI may include one or more antibodies, for example, and without limitation, antibodies against 4-1 BB and / or 0X40. In nonlimiting embodiments, the one or more ICIs may one or more antibodies that bind PD-1, PD-L1, CTLA-4, and / or LAG-3. In non-limiting embodiments, the one or more ICIs may be one or more of nivolumab, pembrolizumab, cemiplimab, atezolizumab, durvalumab, avelumab, ipilimumab, relatlimab, fianlimab, and / or leramlimab. In nonlimiting embodiments the one or more ICIs may include a PD-1 inhibitor and a LAG-3 inhibitor, for example nivolumab and relatlimab.

[0164] In non-limiting embodiments, for example where one or more immune cells of the patient have been determined to not exhibit an exhaustion phenotype and / or tonot exhibit a type-1 interferon response signature, the therapeutic composition may include one or more compounds effective to prime one or more cells to respond to an ICI (e.g., induce a type-1 interferon response signature) in the patient (e.g., in the tumor). In non-limiting embodiments, the one or more cells may be determined to exhibit an exhaustion phenotype, but not a type-1 interferon response signature, and the therapeutic composition may include one or more compounds effective to prime one or more cells to respond to an ICI (e.g., induce a type-1 interferon response signature) in the patient (e.g., in the tumor). In non-limiting embodiments, the one or more compounds effective to prime a cell (e.g., induce a type-1 interferon response signature) may be a STING agonist, a TLR7 agonist, a TLR8 agonist, and / or a TLR9 agonist. In non-limiting embodiments, the one or more compounds may be one or more of cGAMP, DMXAA, MSA-2, TAK-676, a CpG oligodeoxynucleotide, IMO-2055, lefitolimod, imiquimod, and / or mRNA vaccines (for example as described in Grippin et al., “SARS-CoV-2 mRNA vaccines sensitize tumours to immune checkpoint blockade,” Nature, Oct. 22, 2025).

[0165] In non-limiting embodiments, for example where one or more immune cells of the patient have been determined to not exhibit an exhaustion phenotype and / or not exhibit a type-l interferon response signature, following administration of the one or more compounds effective to prime the cells to respond (e.g., induce a type-l interferon response signature), e.g., about 5 hours to about 14 days, optionally about 6 hours to about 14 day, optionally about 7 hours to about 14 days afterwards, all values and subranges therebetween inclusive, the method may include administering to the patient an ICI in an amount effective to treat the cancer (e.g., tumor).

[0166] In non-limiting embodiments, the one or more ICIs and / or the one or more compounds effective to prime one or more cells to respond (e.g., induce a type-l interferon response signature) may be administered to the patient orally or parenterally, though as noted herein, any route of delivery may be suitable. Further, exemplary dosages for compounds are disclosed herein and are known.

[0167] In non-limiting embodiments, following treatment with the one or more compounds effective to prime one or more cells to response (e.g., induce a type-l interferon response signature) and / or the one or more ICIs, one or more immune cells (e.g., one or more CD4+ cells, NK cells, and / or CD8+ cells as described herein) may be obtained and genotyped and / or phenotyped. Based on such analyses, additional determinations as to treatments of the cancer (e.g., tumor) may be made. That is, innon-limiting embodiments, post-treatment genotyping and / or phenotyping may be performed, and, depending on that genotyping and / or phenotyping, additional decisions, for example as to surgical interventions, continued treatment, and / or the like, may be made.

[0168] Also provided herein is a method of treating a patient having a cancer (e.g, a tumor), where the method may include administering, to a patient having one or more immune cells expressing an exhaustion phenotype and / or a type-l interferon response signature as described herein, a therapeutic composition including one or more ICIs as described herein in an amount effective to treat the cancer (e.g., tumor).

[0169] Also provided herein is a method of treating a patient having a cancer (e.g, a tumor), where the method may include administering, to a patient having one or more immune cells that do not express an exhaustion phenotype and / or a type-l interferon response signature as described herein, a first therapeutic composition followed by a second therapeutic composition, wherein the first therapeutic composition may be one or more compounds effective to prime one or more cells (e.g., in the tumor) to respond (e.g., induce a type-l interferon response signature) as described herein and the second therapeutic composition may include one or more ICIs as described herein.Example 1Materials and Methods

[0170] Study design, accrual and clinical characteristics

[0171] Hillman Cancer Center (HCC) 18-139 (NCT04080804) is a phase II trial that evaluated neoadjuvant immunotherapy prior to curative intent surgical resection. The trial randomized patients 1:1:1 to either: Nivo alone, Nivo+Rela, or Nivo+lpi. Randomization was stratified by p16 status (oropharynx only), as well as LAG-3 (>1% vs <1%) and PD-L1 (>1% vs. <1%). Forty-two patients were randomized, with one patient assigned to the Nivo arm withdrawing consent during neoadjuvant treatment. Therefore, 41 patients were evaluable, with 14 randomized to Nivo, 15 to Nivo+Rela, and 12 to Nivo+lpi. The median age (range) was 62 (32-81); 65.9% of patients were male, 5% of patients were black and the remainder were white. The majority were Eastern Cooperative Group Performance status (ECOG PS) of 0 (63.4%). Primary sites included 73.2% oral cavity, 9.8% larynx, and 17% oropharynx. There were 3 patients with HPV positive oropharyngeal SCC (7.3%) in the study, and all were on the Nivo arm. Most patients were clinical T stage 3 (34.1%) orT4 (46.3%), and 68.3%had nodal disease (N1 / N2), with 58.5% of patients being overall stage IVa (58.5%). There were no differences in baseline characteristics or clinical staging by arms. After the 41 patients were accrued, the trial was closed and data analyzed with the definitive clinical data / outcomes reported in the current manuscript. Subsequently a new trial was designed under the same trial notation, eliminating the Nivo monotherapy arm, increasing the Rela dose from 160mg to 480mg, and randomizing 2:1 in favor of Nivo+Rela. This separate trial will be reported when accrual is complete.

[0172] Treatment compliance

[0173] All patients received one cycle of the prescribed treatment with one patient in the Nivo arm and one patient in the Nivo+Rela arm receiving the optional second dose before surgery. Two patients in the Nivo arm and one patient in the Nivo+lpi arm experienced disease progression and did not undergo surgery on trial and therefore were not included in the pathologic response analysis. Two of these patients underwent surgery off trial, one after further chemotherapy. One HPV positive oropharyngeal SCC patient was no longer a candidate for transoral oral robotic surgery (TORS) and underwent definitive chemoradiation.

[0174] Experimental design and patient enrollment

[0175] The trial randomized patients 1:1:1 to either: Nivolumab alone (Nivo), Nivo+Rela, or Nivo+lpi. Randomization was stratified by p16 status (oropharynx only), as well as LAG-3 (>1% vs <1%) and PD-L1 (>1% vs. <1%). The study was approved by the University of Pittsburgh IRB (NCT04080804).

[0176] Patients were included if they had overall stage III or IVA HNSCC (AJCC 8th edition) that was previously untreated and surgically resectable. The primary tumor had to be in the oral cavity, oropharynx, larynx or hypopharynx. Recurrent or second primary patients were excluded unless a patient had a HNSCC treated greater than 5 years ago with surgery alone. For HPV positive oropharyngeal squamous cell carcinoma (SCC), patients must have had a T3 / T4 primary and / or one ipsilateral lymph node greater than 3 cm, multiple ipsilateral lymph nodes, or a contralateral lymph node.

[0177] Other eligibility criteria were left ventricular ejection fraction (EF) >50% by transthoracic echo or multigated acquisition scan (MUGA), accessible tissue for baseline biopsy, age >18 years, and ECOG performance status (PS) 0-1. Patients were excluded if they had a history of: prior RT, CRT, immunotherapy, a severe reaction to mAb antibody, a troponin T (TnT) or I (Tnl) >2 X institutional upper limit ofnormal during screening, a prior history of myocarditis of any etiology, known history of hepatitis B or C, active autoimmune disease (defined as any condition requiring current therapy), concurrent second primary HNSCC, or distant disease.

[0178] During screening, patients underwent a baseline research biopsy, CT neck and chest with intravenous contrast, Troponin, CBC, liver function tests, basic metabolic panel, TSH with reflex T4, PT / PTT / INR, 12 lead EKG, and ejection fraction evaluation. The research biopsy required a sample estimated to be equivalent to at least two 4 mm punch biopsies or four, 14-18 gauge core needle biopsies. A research blood draw was also done which included: one lavender top tube, five green top (heparin) plastic 10ml tubes, one red top plastic 10ml tube. Oral rinse and stool were also collected.

[0179] Eligible patients were randomized to: Nivo monotherapy at 480mg IV every 28 days, Nivo+Rela at 480mg IV and 160mg IV respectively together on day 1, every 28 days, or Nivo+lpi with Ipi 1 mg / m2 IV on day 1, and Nivo at 240mg IV day 1 and 14, every 28 days. There was the option to give a second cycle prior to surgery based on provider preference and surgical timing. During treatment, patients were seen by the treating medical oncologist on day 1 and 14 and by the surgeon on day 14. Surgical resection was done 21-35 days from the start of the immunotherapy treatment and a repeat CT neck / chest was obtained within 72 hours prior to surgery. Patients were seen 1, 3, and 6 months after surgery by the study team. Afterwards, survival followup was conducted via phone. Research blood draws, oral rinse, and stool were collected on day 14, day 28 (if another cycle was given) and then at 1,3, and 6 months.

[0180] Endpoint

[0181] The primary objectives were to collect preliminary data to assess safety, efficacy, and effect on the tumor immune microenvironment of Nivo, Nivo+Rela, and Nivo+lpi. Treatment-related adverse events (TRAEs) and toxicity data were compiled by grade by CTCAE v 5.0, and attribution to treatment was given. Toxicity was characterized as treatment-related if it was possible, probable, or definite from treatment. The trial included a pre-specified pause rule if one patient had an unexpected grade 4 or 5 adverse event and / or if there is a delay of over 4 weeks in the expected time of surgery that is determined to be related to study treatment. If the pause rule was triggered, further accrual would be put on hold, and data and safety monitoring review would occur. Clinical information collected for efficacy endpoints included best overall response by Response Evaluation Criteria in Solid Tumours(RECIST) 1.1, percentage change radiographically, and follow up for any recurrence or death for calculation of disease free and overall survival.

[0182] Sample collection

[0183] Fresh baseline and post-surgical resected tumor tissues were collected and each was split into two sections. One section was placed into RPMI 1640 (Sigma-Aldrich, Cat# 21115) media supplemented with 10% fetal bovine serum (FBS, R& D Systems, Cat#S11150H), 1% amphotericin B (ThermoFisher Cat# 15290026), 1% penicillin-streptomycin (pen-strep; ThermoFisher Cat# 15140122), 1% non-essential amino acid (NEAA, ThermoFisher Cat# 11140050), and 5nM L-glutamine (ThermoFisher Cat# 35050061) and used for scRNA-seq studies. The other was immediately fixed in formalin. Formalin-fixed paraffin-embedded (FFPE) tissue sections were utilized for diagnostic and pathological response evaluation.

[0184] Method details

[0185] Tumor dissociation

[0186] Fresh tumors were transferred to a 5nm dish with 3-5 ml of complete media (1X DMEM supplemented with 10% FBS, 1% NEAA, 1% L-glutamine and 1% pen-strep) and minced into <2mm pieces with sterile scalpels. The tissues were then mechanically dissociated on a 70pm Falcon cell strainer (Coming Inc.; Coming, NY) using a sterile syringe plunger flange. Cells that passed through the strainer were collected into a 50 ml conical tube and centrifuged at 1500 rpm for 10 min to form a pellet. The remaining tissue pieces underwent enzymatic digestion in a 50 ml conical tube containing the dissociation cocktail (1x HBSS supplemented with 50 IU / ml collagenase I, 25 IU / ml collagenase II, 50 IU / ml collagenase IV 0.025 mg / ml DNase I (STEMCELL Technologies; Vancouver, Canada) and 3 mM calcium chloride (Sigma-Aldrich; St. Louis, MO). Tissue digestion was performed under gentle rotation at 37°C in a MACSmix™ Tube Rotator (Miltenyi Biotec, cat# 130-090-753) for 20 min. The resulting tumor suspension was centrifuged at 1500 rpm for 10 min and pooled with cells collected using mechanical dissociation. Lastly, cells were washed with 10-20 ml complete media (RPMI, 10% FBS, 1x L-glutamine, 1x non-essential amino acid, Sodium pyruvate and 10x% amphotericin B) and counted before proceeding to the scRNA-seq protocol.

[0187] Isolation of CD45+CD3+ TIL for scRNA-seq

[0188] All patient specimens, except those collected from patient #108, were freshly used following tumor biopsy. CD45+CD3+ TIL collected from all patients wereisolated through fluorescence-activated cell sorting (FACS). Strained tumor cell suspensions were washed with PBS (Gibco cat# 10010023) and checked for viability using Trypan Blue (Gibco cat# 15250061) and AO / PI dye (Nexcelom; Lawrence, MA). If necessary, a red blood cell (RBC) lysis step was performed to remove extra red blood cells. Briefly, 3-5 ml of 1x RBC lysis buffer (FisherScientific) was added to tumor cell suspension and mixed by inverting a tube for 1-2 min. Subsequently, complete media was added to a total of 50 ml volume to prevent damage to isolated tumor and immune cells. The cell suspension was centrifuged at 1500 rpm for 5 min. Pelleted cells were then counted and checked for viability using Trypan Blue exclusion. Up to 106 cells were resuspended in 100 pl FACS buffer (1x PBS, 1mM EDTA, 0.04% BSA) with the addition of TruStain Fx (Biolegend) to block Fc receptors and incubated on ice for 10min. Fluorochrome-conjugated antibodies against CD45 (Biolegend) and CD3 (Biolegend), as well as DNA-barcoded antibodies against CD4, CD8, CD39, CD103, TIM-3 and PD-1 (Biolegend) were subsequently added to the cells and incubated on ice for 20 min. Afterwards, the cells were washed twice with 5 ml FACS buffer, strained through a 40pm Flowmi cell strainer (Thermofisher) and resuspended at 5x105 cells per 1 ml in FACS buffer. For most specimens, viable CD45+CD3+ cells were sorted into a 5 ml Falcon tube containing the 10x single cell buffer (1X PBS, 0.04% BSA) using the SONY900 cell sorter (Sony Biotechnology; San Jose, CA) and immediately transferred to ice. Due to low cell numbers, only CD45+ cells were sorted and processed for the 10x Genomics workflow for baseline samples from patients #114, 127, and 134, as well as post-treatment samples from patients #119,121,127, and 133.

[0189] Gel bead-in-emulsion (GEM) generation

[0190] 10x Genomics 5’ V1 assay and protocols were performed on specimens isolated from patients #101-109, while 10x Genomics 5’ V2 assay was implemented for patients #110-142. FACS-sorted CD45+CD3+ cells were centrifuged and washed using the 10x single cell buffer. Cells were then recounted and resuspended at 800-1000 cells per 1 pl of buffer. Up to 15,000 cells were loaded to single cell Chip A (5’V1 ) or Chip K (5’V2) channel, along with reverse transcriptase reagent mixtures and 5’ gel beads per manufacturer’s protocol. Chips were loaded onto the 10x Genomics Chromium Controller iX and emulsions were immediately recovered and transferred into an 8-tube strip upon completion of single-cell partitioning.

[0191] 10x 5’ single cell library Preparation and sequencing

[0192] Emulsions were immediately incubated in the deep-well block thermocycler (Invitrogen) to initiate the reverse transcription step (cDNA generation) according to the 10x Genomics protocol (80°C: 5 min; 56°C: 45 min; 4°C: hold). cDNA was used to generate gene expression (GEX), TCR (VDJ), and antibody capture (ADT) libraries per manufacturer’s protocols. The quality of cDNA and resulting libraries was determined by D5000 high sensitivity kits on the 2200 TapeStation system (Agilent; Santa Clara, CA).

[0193] GEX, TCR and ADT libraries from baseline and post-treatment CD45+CD3+ or CD45+ TIL were pooled together from the same patient for sequencing. The resulting pooled libraries were diluted to 2 pM, denatured, and loaded on a NovaSeq 500. For sequencing, NextSeq 500 / 550 High Output v2 kits (150 cycles) was used with the following parameters: Read 1: 26 cycles; i7 Index 8 cycles; Read 2: 98 cycles, as specified by the 10x Genomics guidelines. Average 50,000 reads / cell for gene expression, 5000 reads / cell for TCR and ADT libraries.

[0194] Processing and quality control of TIL scRNA-seq data

[0195] Raw scRNA-seq sequencing reads from each sample were demultiplexed using sample index, and FASTQ files were created via Cell Ranger mkfastq (v7.0.0) and Illumina’s bcl2fastq (illumine software). FASTQ files were aligned to a hybrid reference made of human reference genome (GRCH38) and the HPV16 genome using the cellranger multi command to generate gene expression counts by barcode matrix with “-library” flag “GEX” for each sample.

[0196] The filtered counts matrices of each sample were used as input for Seurat (v4.0, R package) clustering analysis (described below). We applied several steps to filter out low quality cells and potential multiplets. 1) In Seurat, we removed cells with less than 500 unique molecular identifiers (UM I), 250 genes, and greater than 20% of UMI mapped to mitochondrial genes. 2) DoubletFinder70 (v2.0.2, R package) was used to calculate the doublet scores for each cell within individual patient samples. Cells with a doublet score >80 were not considered as single cells and removed from the matrix for further downstream analysis. 3) T cell receptor sequencing provided additional evidence on whether the cells could be doublets or multiplets. Among T cells, we removed cells containing two TCR-a or TCR-[3 chain sequences, except if they were detected multiple times and were clonally expanded populations.

[0197] Batch correction and clustering analysis

[0198] To assess the presence of batch effects, we first processed the data by following the standard Seurat (v4.0, R package) pipeline. The filtered counts matrices of each patient sample were merged into one combined cell count matrix and processed using Seurat (v4.0, R package). The “Normalization” function with default parameters (normalization method = “LogNormalization,” scale.factor = 10,000) in Seurat was applied to normalize the gene expression levels in each single cell to control for different sequencing depths in individual cells. The “FindVariableFeatures” function with “vst” method was performed on the combined cell count matrix to identify 2,000 highly variable genes. Next, “ScaleData” function was used to scale and center gene expression matrices for each cell after regressing out mitochondrial contaminations. Clusters were determined using the Leiden algorithm and calculated using the “FindNeighbors” function with the selection of the first 20 principal components to construct the shared nearest neighbor (SNN) graph. Cluster resolutions was applied when top differentially genes were uniquely associated with each cell cluster. We observed batch effects between patient samples processed with the 5’ v1 assay and the 5’ v2 assay of the 10x Genomics chemistry.

[0199] To correct for batch effects, we applied the Integration method implemented in Seurat. Cells from the combined cell counts matrix were annotated based on their assays (either 5’ v1 or or 5’ v2), and the “Splitobject” function was used to divide the data into a list of Seurat objects. Normalization was performed with “FUN” set to “SCTransform”. “SelectlntegrationFeatures” with default parameters (nfeatures = 3000) was performed to select repeated variables across datasets for integration. We then ran “PrepSCTIntegration” (anchor.features=output from “SelectlntegrationFeatures”) to ensure that the SCTransform residuals for the features were present in each Seurat objects. Next, we found common anchors for integrating the list of individual Seurat objects using the function “FindlntegrationAnchors,” with the normalization method set as “SCTransform” and dimension reduction set as “rpca”. Subsequently, the list of Seurat objects was integrated through the “IntegrateData” function with the normalization method set as “SCTransform.” The integrated dataset was further processed with the standard Seurat pipeline, with the assay set to “Integrated.” To mitigate the dominant effects of T cell receptor variable genes on clustering, all TRAV, TRBV, TRDV and TRGV genes were removed from the list of highly variable genes generated via the “VariableFeatures” function. Cell clustering was then performed by executing the “FindNeighbors” and “FindClusters” functions. Aclustering resolution of 0.3 was applied to the “FindClusters” function to identify clusters of CD3+ T cells, while a resolution of 0.5 was used to determine clusters of CD8+ T cells.

[0200] Cluster annotation and differential expression analysis

[0201] To identify T cell populations, we evaluated the expression of CD3D, CD3E and CD3G transcripts, as well as the reconstruction of TCR-a and TCR-[3 chains. To further improve our classification, we performed multimodal reference mapping, which unbiasedly mapped our data to a published and annotated CITE-seq dataset reference in Seurat. To map our dataset, we first normalized our data with the “SCTransform” function to match the format of the reference dataset. We then found anchors between the reference and query datasets via the “FindTransferAnchors” function with precomputed supervised PCA (spca) transformation. Next, we used the “MapQuery” function to obtain predicted scores for each cell regarding their possible identities, including “CD4 T,” “CD8 T,” “Treg,” “other T,” “B,” “Myeloid,” and “NK” cell populations. Cell identities were annotated based on prediction scores > 0.6. Any cells that did not have a score > 0.6 for “CD4,” “CD8,” “Other T,” “Treg,” and cells that were predicted to be “NK” were removed. To improve our confidence in identifying CD8+ and CD4+ T cell subpopulations, we evaluated the expression of CD4 and CD8A transcripts via scRNA-seq, as well as the expression of CD4 and CD8 at the protein levels via CITE-seq sequencing.

[0202] Processing of CITE-seq data

[0203] Antibody capture FATSQ files were used to run cellranger multi with the library flag” set to “ADT” and aligned to reference sequences of each barcoded antibody (Biolegend) used in this study. This process generated a filtered counts matrix that measured the number of UMI for each antibody within each barcoded cell. The filtered matrices were then used as input for Seurat (v4.0, R package) analysis. A CLR transformation was applied to normalize the ADT data within each cell. Subsequently, PCA was utilized to reduce the dimensionality of both scRNA-seq and CITE-seq datasets, using 7 CITE-seq and 30 scRNA-seq dimensions to construct the WNN graph. When re-clustering CD8+ TIL populations, we repeated all preprocessing steps and performed the same procedure on cells identified as CD8+ TIL.

[0204] CD8+ T Cell Identification based on CITE-seq

[0205] To accurately identify CD8+ T cell populations, we utilized our CITE-seq data in addition to the multimodal mapping method. Initially, we measured CD8A andCD4 expression of CD3+ T cells identified based on multimodal reference mapping. Cells with CD4 expression > 0.5, CD8A < 1, and no TCR chains were subsequently removed. We then measured CD8 protein expression, and cells with CD8 protein expression > 1 and CD8A expression > 0.5 were identified as CD8+ T cells. In the end, we obtained 137,133 CD8+ TIL.

[0206] Processing of single cell T Cell receptor (scTCR) data from CD3+ TIL

[0207] To reconstruct scTCR sequences and annotate each VDJ library, we aligned raw scTCR FASTQ reads to the human GRCh38 V(D)J reference (v7.0.0, from 10x Genomics) using cellranger multi with the “-library” flag as “vdj-T.” The filtered contig annotations, containing amino sequences of TCR-a and TCR-[3 chains, as well as cellular barcodes, were further processed in scRepertoire74 (v1.7.2, R package). Processed baseline and post filtered contig files from same patient were then combined and integrated into metadata of CD3+ TIL according to their cellular barcodes.

[0208] Identification of expanded clonotypes and TCR diversity estimation

[0209] To identify clonally expanded TCR clonotypes, we calculated the frequencies of unique TCR sequences at the baseline and post-treatment timepoints. TCR clonotypes present at baseline and persisting through therapy were considered and identified as clonally expanded based on the following criteria: 1) post-treatment frequency / baseline frequency > 1; 2) post-treatment count > baseline count; 3) to avoid potential dropout events, TCR count > 5. TCR clonotypes that met these criteria were labeled as clonally expanded.

[0210] For TCR diversity estimation, we assigned a “ClonelD” to each unique TCR sequence detected in individual patient samples (baseline and post-treatment combined). Next, we calculated the inverse Simpson Index for each sample using the “invSimpson” function implemented in the vegan R package (vO.4.3).

[0211] TCR tracking and sharing across CD8+ TIL clusters

[0212] To examine the relationship between CD8+ TIL subsets, we calculated STARTRAC transition index (pTrans index) to analyze shared TCR sequences using STARTRAC55 (v.0.1.0, R package). Baseline and post-treatment TCR sequences from patient samples in each cluster were used as input for STARTRAC calculation. We constructed a 12x12 contingency table for each cluster pair in baseline and posttreatment samples. The pTrans index value reflected the indices calculated based on combined patient samples for a given group. To statistically evaluate TCR sharingacross timepoints and different clusters, pTrans index values between baseline and post-treatment CD8+ TIL clusters of individual patients were extracted from the STARTRAC output object. A one-way ANOVA test was used to determine whether significant TCR sharing existed between baseline and post-treatment clusters of interest.

[0213] Identification of expanded putative tumor-reactive clonotypes

[0214] To identify putative tumor reactive clones, we calculated the protein expression of CD39 and CD103 via CITE-seq. CD39 and CD103 protein expression were averaged to individual CD8+ TIL clones detected in baseline and posttreatment tumors. Putative tumor reactive clones were then identified based on four criteria 1). Clonally expanded; 2). Average CD39 and CD103 protein expression > 1; 3). CXCL13 RNA expression > 1; 4). TCR clones with dominant phenotypes TRM-ICRIow or TEX-ISGhigh.

[0215] Differential gene expression analysis

[0216] To calculate differentially expressed genes (DEG) between groups, we used the “FindAIIMarkers” function and the Wilcoxon Rank Sum test in Seurat. In general, a LogFC > 0.5 and adjusted p-value < 0.001 were applied to determine significantly upregulated DEG, while a LogFC < -0.5 and an adjusted p-value < 0.001 were used to identify significantly downregulated DEGs.

[0217] Gene set enrichment analysis (GSEA)

[0218] We used the R package SingleSetGset (vO.1.2) for GSEA. Briefly, GSEA was performed by calculating the mean gene expression per cluster, and assessing the log fold-change in gene expression between a given cluster and the mean expression of the same gene in all cells outside the given cluster as the test statistic. Enrichment scores were calculated for each CD8+ TIL cluster using gene sets obtained from published studies of human TIL, including naive / central memory, effector, effector memory, tissue resident memory, exhaustion, and IFN-I response. To calculate the enrichment scores of tumor reactivity for CD8+ TIL clusters, we utilized published gene sets of tumor-specific or neoantigen-specific CD8+ TIL.37,38,39,40 R package fgsea (v1.33.2)72 was used to calculate enriched pathways.

[0219] Gene signature score calculation

[0220] The enrichment of gene sets and signatures of cells was calculated using AUCell (R package, v1.26.0). By scoring these gene sets, the potential bias due highlyexpressed gene within such gene sets was minimized. TEM-ISGlow and TRM-ICRlow gene signatures in FIGS. 6A-6F were derived from top 30 genes ranked by log transformed the ratio of the normalized gene expression in each cluster. IFN-I response and T cell mediated immune response gene sets were derived from MSigDB (v1.12.0) and T cell exhaustion gene set was obtained from previous scRNA-seq studies. Gene set enrichment score was calculated on a per cell basis and the statistically significant comparisons (eg., baseline vs. post, Nivo+lpi vs. Nivo+Rela) and p values were determined using a linear mixed effects model and Wilcox sum ranked test (Mann-Whitney U test).

[0221] Trajectory inference analysis

[0222] The R package Slingshot56 (v0.2.0) was used to explore and build pseudotime ordering in CD8+ TILs. TN / M cells were considered as the roots when calculating the pseduotime trajectories. RNA velocity57 analysis was performed by scVelo (vO.2.4) and used to validate the results from Slingshot. Loom files were obtained via running raw scRNA-seq counts through Velocyto75 (vO.17.16). To investigate the potential differences across the trajectories (TEX-ISGhigh, TRM-ICRIow and TEX-ISGIow trajectories) between Nivo+lpi and Nivo+Rela, density plots were created from ordered pseudotime (scaled from 0 to 100). To minimize the effect of NA-values within the trajectory, we only included cells known to be part of the trajectory.

[0223] Multiplex immunofluorescent staining

[0224] Four pm FFPE sections from baseline and post-treatment tissue biopsies were mounted to slides. Automated staining of tissues was performed on the Leica Bond RX using Akoya Bioscience’s Opal 6-Plex Detection Kit (cat# NEL871001 KT), following the manufacturer’s instructions. Whole-slide scanning was carried out on Akoya Bioscience’s PhenoimagerHT platform. Briefly, tissues were baked at 65°C for 1 h, deparaffinized using xylene (3x10 min) and rehydrated using ethanol through a serial dilution (100% 1x10min; 95% 1x10min; and rinse 70%). The slides were then subjected to microwave heat-induced epitope retrieval using ER1 (citrate with a pH range of 5.9-6.1) or ER2 (EDTA with a pH range of 8.9-9.1) buffers, followed by a 10 min blocking step before incubation with primary antibodies for 10 min at room temperature (RT). The slides were then incubated in Opal Polymer HRP Ms+Rb for 10 min at RT, followed by a 10 min incubation with Opal Working Solution to generate the Opal Signal. Epitope retrieval, blocking, and Opal Polymer HRP introduction stepswere repeated for all 5 targets until Opal Polaris 780 labeling was achieved. Secondary antibodies were added, followed by Opal Polymer HRP and Opal TSA-DIG introduction. Next, the slides were incubated in Opal Polaris 780 Working Solution at RT for 10 min. Finally, cells were counterstained with DAPI for 5 min at RT and wash with Tris-buffered saline containing 0.1% Tween 20. The slides were mounted with Prolong Diamond Antifade Mountant (ThermoFisher) and stored at 4°C in the dark until imaging.

[0225] Multiplex immunofluorescent data analysis

[0226] Cell segmentation and classification were performed on whole-slide images of tissue sections annotated by a pathologist and scanned at 10x magnification following multiplexed imaging of the multispectral immunofluorescence panel. All procedures were performed in QuPath (vO.3.2). A pixel classifier was trained on pan-CK stain (Opal 781) from 10 randomly chosen samples to segment pan-CK+ tumor epithelium within pathologist-annotated regions. PanCK- regions were treated as stroma. A + / - 35pm margin was added to the pan-CK+ boundary to define the inner and outer tumor invasive margins. Tumor outside of the inner margin was described as “central tumor,” and stroma outside outer margin was described as “remaining stroma.” Next, cell detection was performed on the DAPI channel (threshold: 25) using the watershed method and 2pm expansion to define the cellular compartment. The mean signal within the cellular compartment was used to identify the following markers (fluorophore, threshold): CD8 (Opal 480, 38.922), CD3 (Opal 650, 17.814), pan-CK (Opal 780, 22.754). Using these markers, a CD8+ T cell population was defined as CD8+CD3+. For each patient sample, cell density was calculated as (# target cells) / (# DAPI+ cells)x100% in each tissue compartment. For our validation panel (CD3, CD8, CD39, LAG-3), cell density was calculated as (# target cells) / (# DAPI+CD3+CD8+ cells)x100%. Paired differences pre- vs. post-therapy were examined within each arm on the log-transformed cell densities of each patient using paired t-test and correlated with pathology response using Spearman’s rank correlation (R v4.1.3, ggpubr v0.4.0).

[0227] Quantification and statistical analysis

[0228] For clinical outcomes, treatment and response-based analyses were conducted with Fisher’s exact tests for categorical outcomes, and Mann-Whitney or Kruskal-Wallis tests for continuous outcomes. For DFS and OS, Kaplan-Meier curves were generated and compared via log rank tests. The statistical analysis for correlative data including quantification of immune populations, employed one-way ANOVA(Kruskal-Wallis test), Wilcoxon rank-sum test, and two-sided paired t-test, as described in the figure legends. Correlations between the expression of T cell makers and pathologic responses were estimated using Spearman’s rank correlation. The analysis of DEGs was calculated by the Wilcoxon rank-sum test and p values were corrected for multiple comparisons using a False Discovery Rate (FDR) of 5% or the Bonferroni correction method.Results

[0229] Clinical outcomes

[0230] Thirty-eight patients were evaluable for pTR and categorized into three groups following previous studies: pTR-0 (91–100% viable residual tumor), pTR-1 (51%–90% viable residual tumor), and pTR-2 (≤50% viable residual tumor; FIG. 1A).While there was no significant difference in pathologic response by arm, both combination arms had numerically higher pathologic response rates (73.3% Nivo+Rela, 63.6% Nivo+lpi, 41.6% Nivo), as well as the percentage of patients that had pTR-2 (26.7% Nivo+Rela, 36.4% Nivo+lpi) compared to Nivo monotherapy (8.4%;FIG. 1A). Less than 10% of viable residual tumor was observed in 13.3% Nivo+Rela, 18.2% Nivo+lpi, and 8.4% Nivo patients. The only complete pathologic response (no viable tumor) was in the Nivo+Rela arm (6.7%) in a patient with cT4N0 HNSCC of the oral cavity. Only one patient (Nivo+Rela arm) achieved a partial response by RECIST, with stable disease being observed in 75%, 78.6%, and 50% in the Nivo, Nivo+Rela, and Nivo+lpi arms, respectively (p = 0.345).

[0231] Both combination arms had a numerically higher percentage of patients that had any decrease in target lesions radiographically (42.9% Nivo+Rela, 41.7% Nivo+lpi) compared to Nivo (23%; FIG. 1B). While RECIST response did not correlate with pathologic response, any decrease in the size of target lesions on imaging significantly correlated with pathologic response (p = 0.005).

[0232] A total of 37.5% of patients had clinical to pathologic downstaging after neoadjuvant therapy. This was more common in combination arms (46.7% (7 / 15) Nivo+Rela, 41.7% (5 / 12) Nivo+lpi, 23% (3 / 13) Nivo). We then compared the extent of the planned surgery (based on baseline exams and imaging) to the performed surgery. The performed surgery matched the planned surgery for patients with a radiographic decrease in disease.

[0233] We evaluated whether LAG-3 and / or PD-L1 expression was associated with pathologic response (>50% vs. <50%), using a cutoff point of >1% to definepositivity for each biomarker. There was no significant correlation between pathologic response and expression of LAG-3, PD-L1, or both for all patients. By arm, for Nivo+lpi, a significantly higher pathologic response was observed in those that did not express PD-L1 (p = 0.003) or did not express both PD-L1 and LAG-3 (p = 0.045). In contrast, combined PD-L1 and LAG-3 expression was associated with a higher pathologic response in patients treated with Nivo+Rela (p = 0.05), whereas individual PD-L1 or LAG-3 expression was not significantly associated with pathologic response. There was no significant correlation between PD-L1 and / or LAG-3 expression and pathologic response in the Nivo arm.

[0234] The median follow-up time was 32.2 months. There was no statistically significant difference between arms. However, patients that had a pathologic response >50% tended to have a better disease free and overall survival compared to those with <50% pathologic response (3-year DFS 100% vs. 83.7%, 3-year OS 100% vs. 87.1%, respectively), although this did not reach statistical significance (FIGS. 1C-1D).

[0235] Toxicity

[0236] Overall, 60.9% (25 / 41) of patients had a TRAE of any grade. By arm, 57.1%, 46.7%, and 83.3% of patients in the Nivo, Nivo+Rela, and Nivo+lpi group had a TRAE of any grade, with grade >3 TRAEs occurring in 4 patients in the Nivo arm (G3 leukocytosis, weight loss, pain, and SAE hepatitis), 1 patient in the Nivo+Rela arm (G3 syncope) and 1 patient in Nivo+lpi arm (SAE G4 hepatitis). No patients discontinued therapy due to toxicity.

[0237] CD8+ TIL from responding patients exhibit distinct baseline transcriptional states prior to neoadjuvant combination ICI treatments

[0238] To study the association between transcriptional dynamics of T cells and pathologic responses to neoadjuvant ICI treatments, we isolated and profiled 372,914 CD45+CD3+ TIL (FIG. 2A) from 35 patients (29 baseline and 31 post-surgical fresh tumor specimens; 20 patients having matched pre- and post-treatment biopsies), using the 10x Genomics 5' scRNA-seq workflow. These cells were clustered into five distinct populations annotated as CD8+ T, CD4+ T, FOXP3+ T, y5 T / MAIT and cycling T cell clusters, and visualized by Uniform Manifold Approximation and Projection (UMAP) analysis. Clusters were annotated based on the expression of reported canonical T cell marker gene and CITE-seq expression levels (FIG. 2A).

[0239] dsdsd To evaluate if distinct CD3+ TIL subsets are associated with pathological response, we quantified the proportion of baseline CD3+ cells in eachsubset. Baseline CD8+ TIL showed a trend to the more frequent in responding patients (p = 0.153; FIG. 2B). We assessed baseline expression of PDCD1, CTLA4, and LAG3 by CD8+ TIL, observing the highest expression for PDCD1 and LAG3 (FIG. 2C). The expression of PDCD1 and LAG3 by baseline CD8+ TIL positively correlated with the degree of pathologic response only in Nivo+Rela-treated patients (LAG3: Spearman’s p = 0.64, p = 0.034; PDCD1: p = 0.69, p = 0.02). Similar analyses did not show a significant correlation between LAG3, PDCD1 and CTLA4 expression levels in baseline CD8+ TIL and the degree pathologic response in the Nivo and Nivo+lpi cohorts.

[0240] To identify baseline CD8+ TIL transcriptional signatures correlating with pathologic response, we generated a list of differentially expressed genes (DEG) between pTR-2 and pTR-0 tumors for each regimen. When analyzing within the Nivo+lpi cohort, we found that, among the top 50 DEG, genes upregulated in pTR-2 patients were associated with TCR signaling (LAT), inflammation (IL32, CXCR3), effector function (GZMM) or memory-associated phenotype (LTB) characterized by expression of classical naive / memory marker genes (e.g., CCR7, SELL, LEF1 and IL7R) and tissue residency (ZNF683, HOPX)14,21 (FIG. 2D). Gene set enrichment analysis (GSEA) indicated that pTR-2 patients in the Nivo+lpi cohort had an enrichment of effector CD8+ T cell (TEFF) gene programs (FIG. 2E). CD8+ TIL from Nivo+Rela pTR-2 showed an enrichment of genes associated with type I interferon (IFN-I) response / stimulation (e.g., MX1, IFI6, ISG15, IFI44L, STAT1),31 inhibitory receptors (HAVCR2, ENTPD1, LAG3), transcription factors often associated with T cell exhaustion (ID2, TOX), human leukocyte antigen (HLA) class II molecules (e.g., HLA-DRA, HLA-DRB1), and an TEFF state (e.g., PRF1, GNLY) as well as a chemokine associated with TEFF state and immune cell homing (CCL3) (FIG. 2D).Further GSEA analysis showed that pTR-2 patients in the Nivo+Rela cohort had an enrichment of the IFN-I response gene program, whereas those in the Nivo+lpi arm had an enrichment of TEFF gene programs (FIG. 2E). The baseline sample from the Nivo pTR-2 patient was not available. Consequently, we compared the DEG between pTR-1 and pTR-0, revealing an upregulation of IFN-I response and HLA class II genes in pTR-1 patients. Additionally, we profiled the transcriptome of baseline CD8+ TIL from two patients with progressive disease (PD) who did not have a pathologic response evaluation and were subsequently removed from the Nivo arm. Baseline CD8+ TILs from PD patients were less activated and had elevated expression ofnaive / central memory T cell-associated markers (SELL, IL7R, TCF7) compared to pTR-1 patients. GSEA further showed that baseline CD8+ TIL from pTR-1 patients were enriched in T cell receptor signaling and T cell mediated immune response gene programs. These analyses implied potentially differential response-associated activation pathways for CD8+ T cells by each ICI combination.

[0241] Nivo+Rela increases CD8+ TIL abundance in post-treatment tumors

[0242] Based on scRNA-seq data, post-treatment pTR-2 tumors contained higher CD8+ TIL frequencies irrespective of ICI regimen (p = 0.00025, FIG. 3A). Only Nivo+Rela-treated pTR-2 patients showed a statistically significant increase in CD8+ TIL abundance (p = 0.032, n = 3), while Nivo+lpi-treated patients exhibited a positive trend (p = 0.06; n = 3) in CD8+ TIL abundance from baseline to post-treatment tumors (FIG. 3B). To further evaluate CD8+ T cell infiltration into the TME post-treatment, we assessed their density using multispectral immunofluorescence (mIF) staining of patient-matched pre- and post-treatment formalin-fixed parafilm-embedded tissues (n = 28). Tumors were segregated into four compartments: central tumor, inner and outer tumor-invasive margins (as defined by the regions 35 pm inside and outside the tumor edge, respectively) and remaining stroma (FIG. 3C). The overall CD8+ T cell density was increased in the whole tumor region (p = 0.028; FIG. 3D), as well as outer tumor-invasive margins (p = 0.047) of primarily patients who received the Nivo+Rela regimen. Increased CD8+ TIL density post-therapy in the outer tumor-invasive margin correlated with tumor rejection for Nivo (p = 0.034, p = 0.79) and Nivo+Rela (p = 0.022, p = 0.74), as well as in the remaining stroma, but only for the Nivo+Rela arm (p = 0.022, p = 0.71). When evaluating the total tumor area (tumor bed and stroma), increased CD8+ TIL density post-therapy only correlated with pathologic response in the Nivo+Rela arm (p = 0.0054, p = 0.8; FIG. 3E). No significant correlations were observed in the inner tumor-invasive margin and the central tumor regions.

[0243] Phenotypic and transcriptional dynamics of intratumoral CD8+ T cells following neoadjuvant ICI treatment

[0244] To characterize the transcriptomes of CD8+ TIL and assess cellular subsets associated with response to each neoadjuvant ICI therapy, we extracted CD3+ CD8+ TIL using CD8A expression by scRNA-seq, and CD8 protein expression by CITE-seq. A total of 137,133 CD8+ TIL were identified, which were further divided into 12 transcriptionally defined subclusters (FIG. 4A). These clusters were defined by cross-linking the top DEG of each cluster with previously reported marker genes ofhuman CD8+ TIL subsets. To better characterize these CD8+ TIL subsets, we performed GSEA using published naive or central memory, effector, effector memory, resident memory, exhaustion, and IFN-I response gene signatures for human CD8+ TIL. Two effector memory populations were distinguished by low (C04_TEM-ISG|OW) and high (C05_TEM-ISGhigh) expression of gene sets associated with response to IFN-I stimulation (ISG). Cells in clusters 8 (C08_TEX-ISGhigh) and 9 (C09_TEX-ISG|OW) exhibited enrichment in the expression of T cell exhaustion-related genes (e.g., LAG3, HAVCR2, TIGIT, ENTPD1, and CTLA4), while co-expressing. Additionally, cells in cluster 7 (C07_TRM-ICR|OW) displayed a similar enrichment in the expression of tissue resident memory gene sets as C08_TEX-ISGhigh and C09_TEX-ISGIow cells but displayed low immune checkpoint receptor (ICR) expression. Moreover, cluster 1 (C01_TN / M) cells were enriched in the expression of naive and central memory genes (e.g., IL7R, SELL, LEF1, and TCF7). Effector memory cells in cluster 2 (C02_TEM-cytotoxic) expressed T cell cytotoxicity-related genes (e.g., CCL4L2, FOS, and TNF), while cluster 3 (C03_TEM-activation) expressed genes that regulate T cell activation (e.g., AHNAK1 and MACF1). Cluster s (C06_TRM-resting) expressed tissue resident memory-associated genes (e.g., HOPX and ZNF683), as well as genes associated with naive / central memory T cells. Cluster 10 (C10_THSP) expressed heat shock protein genes (e.g., HSPA1 A and HSPA1 B), cluster 11 (C11_TEMRA) cells expressed genes associated with TEMRA cells (e.g., FGFBP2, FCGR3A and CX3CR1), and cluster 12 (C12_Tcycling) expressed genes associated with T cell cycling and proliferation (e.g., MIK67).

[0245] The dynamic changes of CD8+ TIL activation states were assessed using paired baseline and post-treatment tissues. Notably, when assessing changes in the proportions of CD8+ TIL subsets from baseline to post-treatment, pTR-2 patients treated with Nivo+Rela showed a significant increase in C04_TEM-ISGIow cells (p = 0.026, n = 3). This was accompanied by a concomitant decrease of C08_TEX-ISGhigh cells (p = 0.034, n = 3; FIGS. 4B-4C). Among other baseline vs. post-therapy comparisons, only the pTR-1 cohort receiving Nivo+Rela demonstrated a significant reduction in C04_TEM-ISG|OW(p = 0.034). Evaluation of Nivo pTR-0 patients suggested that these individuals did not have any significant modulations of their CD8+ TIL activation states post-therapy.

[0246] Since C04_TEM-ISGIow was the most frequent CD8+ TIL phenotype in post-treatment pTR-2 patients within each combination treatment arm, we evaluatedthe transcriptional differences among these cells between pTR-2 and pTR-0 patients. Nivo+lpi-treated pTR-2 CD8+ TIL displayed differential enrichment of HLA class II (e.g., HLA-DRA and HLA-DQA2) and effector function-related genes (GZMK, ITGB2, CCL5), reflecting an enhanced effector phenotype compared to pTR-0. Genes that were more highly expressed in pTR-0 were associated with response to IFN-I stimulation (ISG20, IFITM3, and IFI44L) and metabolic stress (HIF1 A and LDHA) (FIG.4D). Nivo+Rela-treated pTR-2 CD8+ TIL showed an upregulation of genes associated with an TEFF cell state, including GZMK, CCL5, CD74, and GZMM (FIG. 4E), as well as enhanced expression of CXCL13, a marker associated with tumor-specific reactivity. For pTR-0 from the Nivo+Rela arm, genes associated with metabolic stress (LDHA, SLC7A5) TEFF (GZMB) and naive / or early TEFF states (CD7) and adhesion (VIM, LGALS1) were identified (FIG. 4E).

[0247] We next evaluated the number of overlapping up- and downregulated DEG in C04_TEM-ISGIow CD8+ TIL from pTR-2 patients treated with Nivo+lpi or Nivo+Rela (Fig. 4F). Among the upregulated DEG in pTR-2 patients’ C04_TEM-ISGIow cells, 33 genes were shared between Nivo+lpi and Nivo+Rela, including those associated with effector function and the HLA class II pathway (FIGS. 4F-4G). Additionally, 37 genes were shared among the DEG in the pTR-0 and pTR-2 comparisons in Nivo+lpi and Nivo+Rela arms, including metabolic stress-related genes.

[0248] A similar analysis was performed on post-treatment C07_TRM-ICRIow cells, which had a modest increase in frequency in pTR-2 patients following the Nivo+Rela treatment. We observed that 8 genes were upregulated by pTR-2 and shared by Nivo+lpi and Nivo+Rela patients, including CCL5, CD52, CXCR6, and IL16, that are commonly expressed by TEFF cells. These data suggest that tumors responding to both combinational ICIs share a common TEFF cell profile posttreatment, despite varying baseline CD8+ TIL states.

[0249] Clonal dynamics during neoadjuvant ICI

[0250] We next investigated whether the increase of C04_TEM-ISGIow and C07_TRM-ICR|OWCD8+ TIL could be attributed to the expansion of pre-existing C04_TEM-ISG|OWand C07_TRM-ICRIow cells, differentiation from other cell clusters, or recruitment of novel T cells from the circulation. To address these questions, clonal dynamics and associated CD8+ TIL phenotypes were evaluated using scTCR-seq. Out of 137,133 CD8+ TIL captured, we successfully reconstructed paired TCR-a and TCR-[3 chains for 87,686 cells (64%), which resulted in 22,594 unique TCR clonotypesthat were further analyzed. We identified TCR clonotypes that were present in baseline tumors and persisted throughout the treatment, classifying them as “pre-existing” clones. Additionally, we identified novel clones that were exclusively detected in posttreatment tumors, categorizing them as “Newly detected (New)” clones. Clonal diversity estimation assessed using the inverse Simpson index, indicated an increase in TCR diversity in post-treatment pTR-2 only from the Nivo+Rela cohort (p = 0.006; FIG. 5A). No other significant baseline-to-post changes in TCR diversity were observed (FIG. 5A). We next assessed the origins of CD8+ TIL and found that cells with pre-existing clonotypes are the primary source of post-treatment cells regardless of patients’ responses and therapeutic regimens. The newly detected clones also showed the TEM-ISG|OWphenotype in pTR-2 patients after the Nivo+Rela treatment.

[0251] Effective combination regimens target CD8+TIL with different activation states at baseline and drive their differentiation toward a common effector or tissueresident memory profile

[0252] To assess whether the dynamic changes between baseline and posttreatment CD8+subsets could be attributed to the expansion of pre-existing C04_TEM-ISG! OWcells or rather a transition between cell states, we investigated the origin of posttreatment C04__TEM! SGiowTCR clonotypes at baseline. We calculated and tracked TCR sharing between baseline and post-treatment CD8* TIL clusters using STARTRAC pTrans indices. In Nivo+lpi pTR-2 patients, TCR sharing was observed between baseline C04_TEM-isGtowcells and post-treatment C04 TEM-isGiow(pTrans index - 0.53) and C02__TEM-cytotoxic cells (pTrans index = 0.56; F G. SB). By contrast, we noted greater TCR sharing in Nivo+Rela pTR-2 patients between baseline and posttreatment CD8+TIL clusters. Among the shared TCR clusters, post-treatment C0 __TEIWISG! OWcells shared TCR with baseline CO8_TEx-: SGfli0hcells (pTrans index = 0.33) and C05_TEM-isGhigcells (pTrans index - 0.32) (FIG. SB). Notably, posttreatment C07_TRM-: CR’OWcells also showed increased TCR sharing with baseline COS TEX-ISG^ cells (pTrans index = 0.58; FIG. 5B). Subsequently, we traced the transcriptional phenotypes of pre-existing CD8+TIL to examine the major cell types before and after treatment in pTR-2 patients. A majority of pre-existing CD8+TIL already displayed effector memory phenotypes (35% C04__T M-isGiowand 20% C02_TEM-cytofoxic) prior to the Nivo+lpi treatment, whereas only 8% of pre-existing cells exhibited a TEM-ISG! OWphenotype before the Nivo+Rela treatment. The major fraction of pre-existing cells at baseline in pTR-2 consisted of C08__TEx-iSGh'gand COS^TEM-ISG^1,accounting for 33% and 18.7% of CDS TIL at baseline, respectively (FIG. 5C). These transitions were not captured in the pTR-0 patients from either combination arm. Upon further analysis of pTR-0 within each treatment arm, we observed that pre-existing CD8+TIL clones of pTR-0 patients from both the Nivo and Nivo+lpi arms exhibited an enrichment of C10__THSP phenotype after treatment. Interestingly, pre-existing CDS'" TIL clones of Nivo+Rela pTR~1 patients predominantly displayed the C08_TEX-iSGfr h phenotype prior to and after treatment.

[0253] Subsequently we performed pseudotime ordering analysis using Slingshot to interrogate the transcriptional relatedness between CD8* TIL subsets and infer trajectories from pseudotime ordered scRNA-seq data. We reconstructed two TEX trajectories, starting from C01__TN / M, connected by C04__T M-; SG! OW, and ending in C08_TEx-! SGf”shand COQ TEX-ISG^T respectively (F G. 5D). We also reconstructed a TRM trajectory, starting from C01__TN / M, connected by TEM-ISG10*, and ending in CO7__TRM-icR3Was reported (FIG. 5D). These trajectories were supported by marker genes expressed along their pseudotime and confirmed using RNA velocity, an independent method. In line with our previous observations (FIGS. 4B-4C and 5C), pre-existing cells from post-treatment tumors of pTR-2 patients in both Nivo-Hpi and Nivo+Rela cohorts were enriched in C04_TEM-ISG! OW(high expression of GZMK, KLRG1, EOMES, and GZMH), which are part of the TEx-isGhighand TRM-icRit)Wtrajectories. Cells from baseline biopsies of pTR-2 in the Nivo+Rela cohort were enriched in both C08_TEx-: SGhigh(high expression of LAG3, ENTPD1, HAVCR2, and PDCD1) and C05_T^M sG!iigh(high expression of ISG15, MX1, F / 6), which are part of the TEx-isGhightrajectory.

[0254] To confirm that CD8* TIL activation in pTR-2 patients varied between the two combination arms, we calculated the density of cells along the TEx-isGhigh, TRM and TEX SG! OWtrajectories. In post-treatment pTR-2 tumors, pre-existing cells shifted away from the end of TEx-! SGhigband toward TEM-ISG! OWparts of the trajectory after the Nivo+Rela, but not the Nivo lpi treatment. In the Nivo+lpi cohort, cells from baseline pTR-2 tumors were enriched in naive / memory (high expression of! L7R, SELL, and LEF1) and TEM parts of the TEX-ISG^ and TRM-ICR! OWtrajectories, and they shifted toward T and TRM (high expression of ZNF633 and HOPX) phenotypes after therapy. Though CDS' TIL in TEX-ISG! OWtrajectory undergoes similar transcriptional changes as the TE -isGigtrajectory, the lack of TCR sharing between baseline C09__TEx-iSGiowand post-treatment C04__TEM-ISG! OWcells FIG. 58) argues against apossible transition between the two cell pools. Of note, following Nivo+Rela but not Nivo+lpi, C04__TEM-iSGiowand C08__TEx-iSGhi9hcells from pTR-2 patients showed decreased expression of IFN-I response (p < 0.001) and T cell exhaustion gene programs (p < 0.001), and a concurrent increase in the T cell mediated immune response gene program (p = 0.002; F G. 5E). In the Nivo+lpi cohort, baseline TEM-ISG! OWcells from pTR~2 showed higher expression of the T cell mediated immune response gene program compared to pTR-2 in the Nivo+Rela arm (p = 0.003), and they maintained the same profile after therapy. These changes were not observed in pTR-0 and pTR-1 patients from either mono- or combination therapies. Overall, these data suggest that post-treatment activation of CDS' TIL involves the reprogramming of pre-existing IFN-l-responsive cells by Nivo+Rela and the expansion of pre-existing effector / effector memory cells by Nivo+lpi.

[0255] To validate our findings, we utilized scRNA-seq and scTCR-seq data from two recent HNSCC neoadjuvant studies. In the first cohort, HNSCC patients received neoadjuvant Nivo or Nivo+lpi, however the data were derived only from patients who received combination ICI. Nivo+lpi responders showed an increase in the abundance of TEM-ISG! OWcells {p - 0.028), and an enrichment of the TEM-ISG! OWgene signature (p < 0.001) post-therapy (FIG. 6A). Responding patients also demonstrated an increase in the frequency of TRM-ICRSOWcells after therapy (p = 0.026) and an enhancement of TRM gene signatures compared to non-responding patients at baseline (p < 0.001). No differences in the expression of the IFN-I response gene program were observed in CDS* TIL from either responding or non-responding patients. TCR analyses further showed that the expanded TEM-ISGSOWcells in responding patients were already present in the TEM-ISG! OWcompartment at baseline (p < 0.001) and were not a result of therapy-induced differentiation from other cell states, confirming the findings in our Nivo+lpi cohort (FIGS. 6B-6D). For the second external validation cohort, baseline-to-post-therapy pairwise comparisons were limited by the lack of pre-treatment scRNA-seq data. Consequently, we specifically analyzed post-treatment CD8+TIL. We found that both TRM and IFN-I gene signatures were enriched in patients with a high response (>50% pathologic response) in the Nivo-Hpi arm (RGS. 6E-6F).

[0256] Given the lack of publicly available scRNA-seq datasets of TIL isolated from Nivo+Rela-treated cancer patients in HNSCC, we next performed mIF to validate our findings regarding the Nivo+Rela treatment. LAG-3 protein and gene expression byCDS T cells were reported to be highly inducible by IFN-I stimulation. We observed that LAG3 expression by CDS* TIL positively correlates with the IFN-I response gene signature at baseline (p = 0.71, p ~ 0.019) and post-therapy (p = 0.84, p ~ 0.0012) in the Nivo+Rela cohort, as well as in all trial patients based on scRNA-seq (FIG. 7A) Additionally, both CD39 gene (ENTPD1) and protein, as well as LAG3 were highly-expressed by C08_TEx-iFNh’9hcells. Therefore, we concluded that modulation of LAG-3 and CD39 protein expression levels could serve as surrogate biomarkers to reflect the expression of the IFN-I response gene program as well as to quantitively assess C08__TEx-iFNfl j!1cells. Patient-matched baseline and post-treatment FFPE tissues for 4 pTR-0 and 3 pTR-2 patients from the Nivo+Rela, as well as 3 pTR-0 and 4 pTR-2 patients from the Nivo-Hpi arm, were assessed by mIF for LAG-3 and CD39 modulations. Consistent with our scRNA-seq observations, only i o+Rela-treated pTR-2 patients showed a decreased density of CD3*CD8*CD39*LAG-3* cells after treatment (p = 0.005), accompanied by an increased density of CD3*CD8*CD39"LAG-3' cells (FIGS. 7B-7D). In contrast, pTR-0 patients showed an increased density of CD3*CD8*CD39*LAG-3* cells after treatment (p = 0.007) (FIG. 7C). This change was not observed in pTR-2 patients treated with Nivo-Hpi (FIG. 7D). Cumulatively, our analyses, supported by external scRNA-seq studies and m! F results, suggest that Nivo-Hpi regimen does not promote cell state transitions, but rather expands cells with enhanced TEM and TRM gene programs that are present prior to treatment. In contrast, Nivo^Rela reprograms and revives exhausted CD8* TIL with a high IFN-I response gene program into less exhausted phenotypes, mainly TEM, in pTR-2 patients.

[0257] Neoadjuvant Nivo+Rela reprograms putative tumor-reactive CD8* TIL that harbor an IFN-I response gene program

[0258] To decipher potential mechanisms governing the expansion and reprogramming of tumor-specific CDS* TIL by the three regimens, we identified putative tumor-reactive clones and evaluated alterations in their gene programs during therapy. GSEA results showed significant enrichment of COS TEX-ISG^ and C07_TRM-icRit)Win the expression of established tumor-reactive gene sets (F G. 8A). To further validate this observation, CITE-seq was employed to examine the co-expression of CD39 and CD103, markers commonly found on tumor-reactive CDS* TIL. In addition to being PDCD1 and HA VCR2 double-positive based on scRNA-seq, CD39 and CD103 evaluated by CITE-seq were also highly expressed on C08_TEx^sGh®hand C07„TRM-ICR! OWcells. Given that CXCL13 has recently been identified as a strongmarker for tumor-reactive C-D8 ' TIL, we validated and confirmed its expression across individual CDS* TIL clusters including COS TTX-ISG51^ and C07__T M-icRfowcells. Therefore, the C08_TE -isGh'gand C07_TR[ U R! OWclusters were designated as putative tumor-reactive clusters.

[0259] To study transcriptional alterations in clonally expanded putative tumor-reactive CD8+TIL, we selected clones at baseline based on three primary criteria: 1) T cell clones with a TEx-isGhi0hand TRM-! CR! OWphenotype, 2) co-expression of CD39, CD103 (CITE-seq) and CXCL13 (scRNA-seq), and 3) expansion of TOR clones after treatment (FIG. 8B). interrogation of baseline vs. post-transcriptional changes of expanded putative tumor-reactive CD8* TIL via DEG analysis revealed that pTR-2 patients in the Nivo+Rela cohort exhibited a significant reduction in the expression of genes associated with T cell exhaustion and IFN-I response compared to pTR-0 within the same trial arm (FIG. 8C). Evaluation of the top ten expanded putative turner-reactive clones in pTR-0 and pTR-2 patients from the Nivo+lpi cohort, as well as pTR-0 and pTR-1 patients in the Nivo+Rela cohort, did not show a decrease in the expression of IFN-I response and exhaustion gene programs. Instead, these clones gained expression of IFN-I and exhaustion gene programs (FIGS. 8C-8E). in contrast, pTR-2 patients in the Nivo+Rela arm demonstrated a reduction in IFN-I response and exhaustion gene signatures at the clone level (FIGS. 8D-8E). Additionally, pseudotime analysis showed that, in post pTR-2 tumors, top expanded putative pre-existing tumor reactive clones shifted away from the end of TEx-isGh’8hand toward TEM-ISG! OWparts of the trajectory only after Nivo+Rela treatment but not Nivo+lpi. Collectively, these findings suggest that Nivo+Rela specifically targets and reprograms exhausted CDS4' TIL with IFN-l-stimulated gene signatures, thereby promoting their transition to an effector profile.Example 2Materials and Methods

[0260] Patients with treatment naive, Stage H-iVa locally advanced resectable head and neck squamous cell carcinoma (HNSCC) were assessed prior to and following treatment with Nivolumab alone (Nivo), Nivolumab + Ipillimumab (Novo+lpil), or Nivolumab + Relatilmab (Novo+Rela). Assessments included gene expression and T-cell receptor (TCR) reconstruction.Results

[0261] scRNAseq analyses reveal distinct NK cell phenotypes in HNSCC and distinct proportional changes in NK cell phenotypes in HNSCC at baseline and posttreatment. In addition, an increase in NK cell cytotoxic and Type-1 interferon signature at baseline predicted HNCSS patient response. Specifically, heatmaps showed that gene signatures associated with inflamed type-1 IFN and CD56dim NK cells are associated with responsiveness to ICI treatments. Gene set enrichment analysis (GSEA) of selected Hallmark genes showed that MTOR signaling, oxidative phosphorylation, and interferon-asignatures are enriched in responding patients at baseline. In addition, proportional changes of inflamed-type, CD56 DIM and ZNF683+ subset NK cells per patients were observed based on response at baseline and post-tx.

[0262] Lastly, scRNA-seq analyses revealed an upregulation of TIM-3 (HAVCR2) in non-responsive versus responsive HNSCC patients, primarily in the anti-PD-1 monotherapy-treated group (FIG. 9).

[0263] Accordingly, in view of the data, it could be concluded that tumor-infiltrating NK cells may contribute to ICI-driven anti-tumor immunity. In addition, increased expression of cytotoxicity, MTOR signaling, oxidative phosphorylation and type-1 interferon response gene signatures at baseline in tumor-infiltrating NK cells isolated from patients that responded to ICI. Lastly, HAVCR2 is upregulated in tumorinfiltrating NK cells from patients that did not respond favorably to the three ICI regimens, particularly nivolumab monotherapy.Example 3Materials and Methods

[0264] This study investigated the T cell transcriptional and spatial dynamics of T cells in HNSCC patients treated with neoadjuvant nivolumab (anti-PD-1, Nivo) ± ipilimumab (anti-CTLA-4, Ipi) or relatlimab (anti-LAG-3, Rela) to better elucidate specific mechanisms of ICI therapy on T cells associated with favorable clinical responses.

[0265] HNSCC patient-matched pre- and post-treatment tumor samples from HNSCC patients (n=42) were processed for single-cell RNA-sequencing, TCR sequencing and cellular indexing of transcriptomes and epitopes. Thirty-nine patients completed sequencing and were analyzed using Seurat (v.5.0.1). Specific T cell marker genes and signatures were used to characterize cellular subpopulations with specific transcriptional and clonal dynamics. Tissue sections were used to performmultispectral imaging to evaluate changes in T cell composition within the tumor. Pathologic response was scored using standard criteria.Results

[0266] Two Nivo patients of the Nivo monotherapy arm were excluded due to clinical progressive disease and inability to complete the neoadjuvant treatment. Remaining patients were categorized based on their pathologic response to the neoadjuvant treatment. Non-responders (pTR-0) had a pathologic response <10%, minor responders (pTR-1) exhibit a pathologic response between 10-49%, major responders (pTR-2) had a pathologic response between 50-100%. We observed four pTR-2 pathologic major responders (pMR) in each combination arm (Nivo+lpi: 4 / 12 (30%), Nivo+Rela: 4 / 15 (27%)).

[0267] LAG-3 (p=0.099) and PDCD1 (p=0.029) expression by CD4+ Tconv cells at baseline is positively correlated with path (FIG. 10). response to Nivo / Rela treatment. CTLA4 expression by CD4+ Tconv at baseline is negatively correlated to pathologic response (p=0.047).

[0268] Expression of IFN-I response, ICR and Tfh genes in CD4+ Tconv at baseline positively correlates to Nivo / Rela pTR-2.

[0269] Nivo / lpi and Nivo / Rela treatments leads to increased frequencies of C08 early-Tfh and a decreased frequency of C11 IFN-I induced Tconv.

[0270] Nivo / lpi and Nivo / Rela treatments leads to increased frequencies of C08 early-Tfh and a decreased frequency of C11 IFN-I induced Tconv.

[0271] LAG-3 (p=0.052) and PDCD1 (p=0.04) expression by Treg cells at baseline is positively correlated with path response to the Nivo / Rela treatment (FIG. 11). CTLA4 expression by Treg at baseline is negatively correlated to pathologic response (p=0.039).

[0272] Higher expression of IFN-I response genes and ICR in Treg at baseline associates with Nivo / Rela pTR-2.

[0273] Seven different clusters of Treg could be identified and characterized according to their top differential expressed genes (DEG).

[0274] Nivo / Rela treatment leads to a decrease in C05 IFN-I induced Treg and C07 cycling Treg frequencies.

[0275] Nivo / lpi pTR-2 indicate targeted reprogramming of naive Tconv to Tfh (B, arrow). Nivo / Rela pTR-2 exhibit a ICR high population at baseline that is not present at baseline in pTR-0.

[0276] Nivo / Rela treatment causes reprogramming of IFN-I induced and activated Treg to a more naive or memory phenotype in pTR-2.Conclusions

[0277] Nivo / lpi and Nivo / Rela may target distinct CD4+ Tconv and Treg activation states. Lower CTLA4 expression by Tconv associates with better pathological response to Nivo / lpi. High ICR expression by Tconv und Treg at baseline correlates with better pathological response to Nivo / Rela. Nivo / Rela causes reprogramming of Treg with an activated phenotype to a more naive or memory state in pTR-2.Example 4Materials and Methods

[0278] Tumor dissociation, sample preparation and scRNASeq workflow

[0279] The complete workflow from tumor dissociation followed by isolation of CD45+ CD3+ TIL for scRNASeq, library preparation and use of 10x genomics 5’ kits was previously described. In addition to the previously described CITESeq antibodies, the TotalSeqTM-A Human Universal Cocktail, V1.0 was used for patients 152 and 155. The gene expression (GEX), CITESeq (ADT), and TCRseq libraries were sequenced and used for downstream bioinformatics analyses.

[0280] BIOINFORMATICS METHODS

[0281] Processing and quality control of TIL scRNAseq data

[0282] Raw scRNAseq sequencing data were pre-processed to generate filtered count matrices for each sample using the 10x Genomics Cellranger pipeline (v.7.0) as previously described previously. The Seurat package (v5.0.1) was used for downstream quality control and analyses. Low quality cells were removed using filters such as less than 500 unique molecular identifiers (UMI), 250 genes, and greater than 20% of UMI mapped to mitochondrial genes. Next, the DoubletFinder package (v2.0.4) was used to remove doublets by setting the cut-off as score >76.

[0283] Batch correction and clustering analysis

[0284] Since batch effects were previously identified in this dataset as described previously, we used the Seurat Integration pipeline to correct batch effects. Cells from the combined cell counts matrix containing GEX and ADT counts were annotated based on their 5’ chemistry version (v1 or v2), and the “split” function was used to divide the data into a list of Seurat objects. Normalization was performed followed by finding 2500 highly variable genes. To eliminate the bias in clustering and downstream analysis, of T cell receptor variable genes on clustering, all TRAV, TRBV, TRDV andTRGV genes were removed from the list of highly variable genes generated via the “FindVariableFeatures” function. The standard analysis workflow steps followed next included scaling the data using “ScaleData”, running PCA using “RunPCA”, finding nearest neighbors using “FindNeighbors” with 30 PCs, running the clustering algorithm using “RunClusters” with a resolution of 0.2. The results of the clustering were visualized after making a LIMAP using “RunllMAP”. Next, the “IntegrateLayers” function was used to perform integration using the “RPCAIntegration” method. This was followed by running, “JoinLayers”, “FindNeighbors”, “FindClusters” and “RunllMAP” on the integrated Seurat object. The ADT assay was normalized using the “CLR” normalization method for further downstream analysis.

[0285] CD4+ T cell identification and separation of Tconv and Treg cells

[0286] From all the CD3+ cells, we separated the CD4+ and CD8+ cells with the help of marker genes and confirmed this selection using ADTs. Clusters with high expression of CD4 and the CD4 ADT were further selected for downstream analyses. To remove double positive cells, a filter was used to remove CD8A+ cells (CD8A < 0.5). This purified population of CD4+ cells was used to calculate a Treg score to further assist with selection of Tregs from the CD4+ cells. The genes used for Treg selection include FOXP3 and IL2RA. Clusters with the highest score for these genes were selected and an additional filter of FOXP3 > 0.1 was applied to confirm the identity of these cells as Tregs. All the remaining cells were considered Tconv. The Seurat integration workflow as described above was followed for both Tconv and Treg cells using patient ID as the batch correction variable.

[0287] Processing of single cell T cell receptor (scTCR) data from CD3+ TIL

[0288] VDJ libraries were generated using Cellranger multi (v7.0.0) by adding the “VDJ-T” flag in the library argument. The filtered contig files generated from Cellranger were used as input for scRepertoire (v2.0.0) which helped with formatting these data and merge them with the existing Seurat object by barcode. Cells with multiple TCR-β chains or more than 2 TCR-α chains were removed from downstream analyses.

[0289] Categorization of TCRs and identification of expanded clones

[0290] TCR clones were classified into two main categories: “Pre-existing” clones were found both at baseline and post treatment time points whereas “Newly detected” clones were observed only in the post treatment samples. Clones from both these categories were further divided based on count (count > 2 and count =1). To identify expanded clones, we calculated the proportion of the TCR repertoire occupied by eachclone at baseline and post treatment. If the ratio of post / pre proportion >1 and post count >1, we considered the clone to be expanded. Cells containing the amino acid sequence CTaa of expanded clones were labeled as expanded and were used for downstream analyses.

[0291] TCR tracking and sharing across CD4+ Tconv and Treg clusters

[0292] To study the sharing of TCR clones across clusters pre and post-treatment, the STARTRAC package (v0.1.0) was used. The pTrans Index was calculated for each response group by splitting patients by timepoint to quantify the TCR sharing between clusters. TCR sharing was also calculated across the Treg and Tconv cell types to evaluate any TCR sharing between these two major CD4+ populations.

[0293] Differential gene expression analysis

[0294] Differentially expressed genes (DEGs) were calculated using the “FindMarkers” and “FindAIIMarkers” functions using the “wilcoxjimma” test from the Seurat package. Significant genes were selected based on adjusted p-value <= 0.05 and the percentage difference (pct1-pct2) was used to rank the genes. The logFC values are inflated for infrequently expressed genes and hence were not solely considered for the ranking. Volcano plots and heatmaps were used to visualize the DEGs.

[0295] Gene Set Enrichment Analysis (GSEA)

[0296] Gene signature scores were assigned to each cell using the “ssGSEA” method implemented in the escape R package (v2.2.4). The scores were averaged per sample and the patients were divided into groups based on the median value of the scores for survival analysis. For gene set enrichment between 2 groups, the fgsea package (v1.26.0) was used with the significant DEGs between the two groups as an input. Gene signatures used were obtained from MSigDB using the msigdbr package (v7.5.1) and from other published studies.

[0297] Trajectory inference analysis

[0298] The R package Slingshot (v2.8.0) was used to explore and build pseudotime ordering for both Tconv and Treg. Naive cells were considered as the roots when calculating the pseudotime trajectories for both cell types.

[0299] Gene expression vs. response correlation plots

[0300] Average gene expression for each patient was calculated using “FetchData” function from Seurat, followed by averaging the score per patient at bothtimepoints and path response was used for correlation analysis. Spearman’s correlation was used, and the p-value was reported.

[0301] Identification of IFN-y and T-bet expressing Tregs

[0302] IFN-y and T-bet expressing Tregs were identified based on mean IFNG and TBX21 expression and divided into 2 groups (high vs low) based on the overall mean IFNG and TBX21 expression for downstream analyses.

[0303] Multiplex immunofluorescent staining and data analysis

[0304] The mIHC staining workflow and data processing was previously described in detail. Briefly, cell segmentation and classification was performed on whole slide scans followed identification of tumor and stroma regions using a pixel classifier. Following cell detection using the DAPI channel, CD3+ CD8- cells were considered as CD4+ and CD3+ CD8- FOXP3+ cells were considered as Treg for all downstream analyses. For each patient, the cell density for these populations was calculated in each tissue compartment. Change in cell density was correlated with pathologic response using the Spearman’s rank correlation method (package). PD-L1 and LAG-3 staining was evaluated in a binary coded way with no visible staining equals negative and visible staining equals positive staining.

[0305] Analysis of ligand / receptor interactions across cohorts

[0306] We analyzed putative ligand / receptor interactions between T cell subsets using the R package Celltaker; v1.0.0.3; https: / / github.com / CilloLaboratory / celltalker / ). Briefly, Celltalker works by scoring each ligand / receptor interaction based on the joint mean expression level of the ligand and receptor between cell types of interest and then comparing this joint mean to a mean in which the ligand and receptor are scrambled. P-values are derived by comparing the distribution of the true joint mean score to the scrambled joint mean score. Statistically significant interactions are those with a family-wise type I error of less than 5%. Celltalker was run on all timepoints, treatment groups, and response groups individually and results were compared to (i) identify ligand / receptor interactions that were unique to one group versus all others and (ii) quantify changes in the magnitude of ligand / receptor interactions between groups. In all instances in which Celltalker was run, CD4+ Tconv, CD4+ Treg and CD8+ T cells were input as the metadata grouping with a ligand or receptor requiring expression of at least 100 counts and less than 20000 counts. Input ligand / receptor gene pairs were obtained from the NicheNet database. We randomly scrambled ligand and receptor pairs 10 times to create a background distribution in all instances. For R-L normalization, we repeated subsampling steps (sampled 4000 cells from each treatment condition 10 times) and then took unique ligand / receptor interactions that appeared at least two of the random subsampling instances. Circos plots were used to visualize unique interactions in specific groups. The Iog2 was taken of inferred strength of the interaction between ligand / receptor pairs for the comparisons of interactions strength between groups.

[0307] Survival analysis

[0308] Survival analysis was performed using the Survminer R package (vO.4.9). Variables of interest were divided into 2 groups and the surv_fit() and ggsurvplot() functions were used to make the K-M curves, p-values were calculated using the log rank test.Results

[0309] Baseline PDCD1, LAG3 and CTLA4 expression by Tconv and Treg correlates with pathologic response to neoadjuvant ICI doublets

[0310] Of 451,984 CD45+CD3+ 94 TIL profiled from 39 patients (31 baseline, 35 post-surgical HNSCC specimens), we identified 202,514 CD4+ TIL. These were clustered into Tconv, Treg and cycling CD4+ cells and visualized by Uniform Manifold Approximation and Projection (UMAP). Cells were annotated using canonical CD4+ T cell transcriptional markers, with CITEseq protein expression and gene set enrichment confirming assignments. We first assessed whether baseline Tconv and Treg frequencies among CD3+ TIL correlated with pathological response and found no significant differences across groups.

[0311] Next, we compared baseline expression of PDCD1, LAG3 and CTLA4, the immune checkpoint receptors (ICR) targeted in our trial. PDCD1 expression was low in all CD4+ subsets, while CTLA4 and LAG3 were more highly expressed in cycling CD4+ 104 T cells and Treg. Meanwhile, ICR expression on Tconv and Treg at baseline was equal across the study arms. Notably, positive immunohistochemical staining of PD-L1 but not LAG-3 at baseline revealed an association with longer OS in our cohort (PD-L1: p=0.0024; LAG-3: p=0.17; FIG. 12A).

[0312] In Tconv cells, baseline PDCD1 and LAG3 expression positively correlated with pathologic response to Nivo+Rela (PDCD1 p=0.0024; LAG3: p=0.0062) but not Nivo+lpi (FIGS. 12B-12C). In contrast, baseline CTLA4 expression negatively correlated with response to Nivo+lpi (p=0.0021), but not Nivo+Rela (p=0.23; FIGS.12D)

[0313] In Treg, baseline PDCD1 expression correlated with response to Nivo+Rela (p=0.0037), 114 while baseline LAG3 showed borderline significance (p=0.062; FIGS. 12E-12F). CTLA4 expression negatively correlated with pathologic response to Nivo+lpi (p=0.00092), while trending towards a positive correlation in the Nivo+Rela arm (p=0.066; FIG. 12G). Neither LAG3 nor PDCD1 expression in the Nivo+lpi arm showed significant correlation with pathologic response in Treg (FIG.12G)

[0314] Using survival analysis that included all trial patients and treatment regimens, a higher level of CTLA4 expression by Tconv and Treg and a higher level of LAG3 expression only by Treg correlated with an improved OS (Tconv CTLA4: p=0.017; Treg CTLA4 p=0.025; Treg LAG3 p=0.035; FIGS. 12H-12J).

[0315] Baseline Tconv transcriptional profiles associate with neoadjuvant regimen-specific pathologic response

[0316] To identify baseline transcriptional signatures in tumor-infiltrating Tconv associated with pathologic response, we calculated DEGs between pTR-2 and pTR-0 patients within each treatment arm. In Nivo+Rela pTR-2 patients, baseline Tconv had an enrichment of genes associated with ICRs (LAG3, PDCD1, HAVCR2, CTLA4, TIGIT), chemokine-chemokine receptors (CXCL13, CXCR6, CCL3, CCL4, CCR5) and cytotoxicity (GZMB, GZMA, CSTB, GNLY, TNF) (FIG. 13A). While an enrichment of interferon (IFN)-I response genes (JFI44L, OAS1, ISG15, IFI6, ISG20, MX1) was present in deep responders to both combination arms (FIG. 13A), several gene families presented discordant correlations to response. Conversely, CTLA4, TIGIT and CXCL13 were enriched in Tconv from Nivo+lpi pTR-0 patients, indicating potential predictors of therapy resistance (FIG. 13A). Moreover, genes associated with oxidative phosphorylation pathway (OXPHOS; MT-ATP6, MT-ND4, MT-ND2, MT-CYB, MT-ND3) were enriched in Nivo+lpi pTR-2, but also in Nivo+Rela pTR-0 patient Tconv, indicating a treatment-specific role in response prediction (FIG. 13A). Lastly, genes associated with a naive T cell phenotype (KLF2, SELL, TCF7, LEF1, IL7R, KLF3) were enriched in pTR-0 patients in both doublet ICI arms, reinforcing the role of baseline activation in effective Tconv responses (FIG. 13A).

[0317] Gene set enrichment analysis (GSEA) revealed broader baseline pathway enrichment in Tconv from Nivo+Rela pTR-2 patients, highlighting hallmark signatures related to TNFO-NFKB, IFN-y, IFN-a, IL6-JAK-STAT3 and IL2-STAT5 144 signaling. Notably, OXPHOS enrichment was exclusive to Nivo+lpi pTR-2 at baseline (FIG.13B). In contrast, a naive gene signature was significantly enriched only in Tconv from Nivo+Rela pTR-0 patients at baseline.

[0318] Altogether, baseline expression of ICR, chemokine-chemokine receptors and OXPHOS genes by Tconv displayed a discordant role to deep response in both combination arms indicating potential regimen-specific predictors of therapy resistance.

[0319] Therapy specific baseline Treg transcriptomes associate with tumor rejection

[0320] IFN-I response genes (IFI44L, OAS1, ISG15, IFI6, ISG20, MX1) were upregulated by Treg at baseline in pTR-2 in both combination arms mirroring results presented in Tconv (FIG. 13C). Strikingly, most gene families shown in Treg presented discordant expression patterns to response. In Nivo+Rela pTR-2 patients at baseline, Treg were enriched with ICR (LAG3, PDCD1, HAVCR2, CTLA4, TIGIT) and costimulatory molecules (TNFRSF9, TNFRSF18, TNFRSF1B, TNFRSF4, ICOS) that are linked to Treg activity (FIG. 13C). In contrast, PDCD1, HAVCR2, CTLA4 and costimulatory molecules were enriched in Treg from Nivo+lpi pTR-0 patients (FIG.13C), highlighting regimen-specific baseline profiles that may mark responsiveness or resistance. NaTve-memory-associated genes (KLF2, SELL, TCF7, LEF1, IL7R, KLF3) were enriched in Treg at baseline from Nivo+Rela pTR-0 patients only, suggesting a potential baseline Treg state linked to resistance (FIG. 13C).

[0321] In Nivo+lpi pTR-2, baseline enrichment of OXPHOS-related genes (MT-ATP8, MT-ND4, MT-ND2, MT-CYB, MT-ND3) was correlated with pTR-2 while an increased OXPHOS expression was found in Nivo+Rela pTR-0 patients (FIG. 13C) mirroring discordant results observed in Tconv. Immunosuppression-associated genes (IL10RA, TGFB1, EBI3, NT5E) in baseline Treg lacked consistent regimen-associated expression (FIG. 13C).

[0322] GSEA confirmed the association of IFN-y and IFN-a response pathway enrichment to pTR-2 patients in both treatment regimens, previously associated with Treg destabilization (FIG. 13D). Additionally, only in Nivo+Rela pTR-2 responding patients, significant enrichment of TNFO-NFKB and IL6-JAK-STAT3 signaling pathways was observed (FIG. 13D). Similarly, naive-markers and OXPHOS genes exhibited a divergent association to response in both combination arms (FIG. 13D).Nivo+Rela pTR-2 patients showed an enrichment of the cytotoxic gene set at baseline.A less differentiated baseline Treg state may be involved in Nivo+Rela treatment resistance.

[0323] Post-treatment change in density of Treg in stroma is discordantly associated to neoadjuvant regimen-specific pathologic response

[0324] To assess CD4+ T cell density within the tumor microenvironment (TME) and to validate scRNAseq results, we performed multispectral immunofluorescence (mIF) staining on patient matched pre- and post-treatment tumor tissues. Pan-CK staining was used to define tumor and stromal regions, while Tconv and Treg were defined as CD3+CD8-FOXP3- and CD3+CD8-FOXP3+, respectively. While the trend was similar in both combination arms, a higher Tconv influx in both tumor (p=0.061) and stroma (p=0.0033) correlated with greater pathologic response only in the Nivo+lpi cohort. Treg density change in the tumor did not significantly correlate with pathologic response across treatment arms. Remarkably, stromal density change of Treg posttreatment did show an inverse association with pathologic response to both combination arms. Here, a higher Treg infiltration was positively correlated with response to Nivo+lpi (p=0.0015) and negatively associated with response to Nivo+Rela (p=0.021). Survival analysis including all trial patients displayed a longer OS and disease-free survival (DFS) in patients with a higher Tconv density in the stroma (OS: p=0.049; DFS: p=0.072). Baseline frequency of Treg in the stroma and Tconv in the tumor demonstrated a longer DFS while infiltration during treatment did not associate with improved DFS (Treg stroma: p=0.045; Tconv tumor: p=0.083). Based on the discordant density pattern in both combination treatment arms, we then sought to determine the transcriptional cell states of these induced Tconv and Treg infiltrating deep responding tumors.

[0325] Transcriptional characteristics of tumor-infiltrating Tconv following neoadjuvant ICI treatment

[0326] To analyze CD4+ 201 TIL transcriptomes and identify subsets associated with response to neoadjuvant ICI treatments, we first segregated CD4+ 202 T cells as described above. Within 127,034 tumor-infiltrating Tconv, twelve transcriptionally distinct subclusters were defined. Using subcluster-specific DEG analysis and canonical gene sets, we annotated each cluster to transcriptionally characterize their phenotypes.

[0327] Clusters C01 and C02 exhibited enrichment of genes associated with a naive-memory phenotype (SELL, IL7R, KLF2, TCF7, CCR7), with C02 showing higherKLF2 expression (C02_KLF2hi_Tnaive) and C01 being enriched for ANXA1 (C01_ANXA1hi_Tnaive). MHC-II genes (HLA-DRB5, HLA-DQB1, HLA-DPB1) marked cluster C03 (C03_MHCIIhi210 _Tconv). C04 expressed IL17A and RORC, consistent with T helper 17 cells (TH 17; C04_IL17Ahi_TH17). TH1 genes (TBX21 / T-bet, IFNG, GZMB, CXCL13) were enriched in C05 and C12, with C05 co-expressing ICRs (LAG3, PDCD1, TIGIT, CTLA4; C05_ICRhi_TH1), while C12 was enriched with TNF (C12_TNFhi_TH1). IKZF2 / Helios characterized C06 (C06_IKZF2hi_Tconv). Central-memory markers (IL7R, LEF1, CD69) defined C07 (C07_IL7Rhi_TCM). T follicular helper cell (TFH)-associated genes (CXCL13, T0X2, TOX) were enriched in C08 and C10, with C08 co-expressing TCF7, a gene associated with a memory-naive phenotype (C08_TCF7hi_early_TFH), and C10 expressing TOX (C10_TOXhi_TFH). C09 displayed proliferation-associated genes (MKI67, MCM7; C09_MKI67hi_Tcycling), and C11 was enriched for IFN-I response genes (ISG15, MX1, IFIT1; C11_ISGhi_Tconv).

[0328] Distinct changes in tumor-infiltrating Tconv subcluster proportions were observed posttreatment. C08_TCF7hi_early_TFH expanded in pTR-2 patients posttreatment in both doublet ICI arms, reaching significance in the Nivo+Rela arm (Nivo+lpi: p=0.11; Nivo+Rela: p=0.043). C11_ISGhi_Tconv showed a post-treatment decrease, most pronounced in the Nivo+Rela arm (p=0.084). In contrast, C03_MHCIIhi_Tconv increased exclusively in Nivo+Rela pTR-2 patients (p=0.087), while C09_MKI67hi_Tcycling decreased specifically in this patient cohort posttreatment (p=0.081). A post-treatment increase in the combined proportion of C08_TCF7hi_early_TFH and C10_TOXhi_TFH clusters was significantly correlated with favorable pathologic response in the Nivo+lpi arm (p=0.0051), but not in Nivo+Rela (p=0.5; FIG. 14A).

[0329] GSEA of published CD4+ T cell gene signatures revealed an enrichment of exhaustion and cytotoxicity signatures as well as IL2-STAT5 signaling pathway in Tconv post-treatment versus baseline in Nivo+lpi pTR-2 patients (FIG. 14B). In contrast, naive, IFN-a response, and IFN-y response were more prominent at baseline (FIG. 14B). For Nivo+Rela pTR-2 patients, only the OXPHOS pathway was significantly enriched post treatment, whereas IFN-a response, IFN-y response, IL6-JAK-STAT3 and IL2-STAT5 pathways were significantly enriched in Tconv at baseline (FIG. 14B)

[0330] Next, we investigated transcriptional changes in tumor-infiltrating Tconv post-treatment, comparing pTR-2 and pTR-0 patients within each treatment. In Nivo+lpi pTR-2 patients, Tconv exhibited enrichment of cytotoxicity-associated (GZMM, CST7, CSTB) and ICR genes (PDCD1, TIG IT), consistent with an effector phenotype (FIG. 14C). Conversely, Tconv from Nivo+lpi pTR-0 patients displayed enrichment of naive-memory (KLF2, IL7R, TCF7), chemokine / chemokine receptor (CXCR4, CCR4) and TCR signaling genes (FOSB, JUND, PRDM1, NFKBIA, NFKBIZ) (FIG. 14C). In Nivo+Rela-treated pTR-2 patients, there was an enrichment of genes associated with OXPHOS (MT-ATP8) and cytotoxic function (GZMM), while pTR-0 patients were enriched for genes associated with TCR signaling (FOS, JUN, FOSB, DUSP4, NFKBIA), and MHC-II expression (HLA-DRA, HLA-DQAT, FIG. 14D)

[0331] Post-treatment in pTR-2, upregulated and downregulated genes were identified compared to pTR-0 patients in both arms. A gene associated with chemotaxis (IL16) was upregulated in pTR-2 patients post-treatment, while genes related to immune-modulation (NFKBIA), regulation of TNF signaling (TNFAIP3) and cell cycle control (CDKN1A) were downregulated in pTR-2 patients following doublet ICI (FIGS. 14C-D).

[0332] Transcriptional features of tumor-infiltrating Treg after neoadjuvant ICI therapy

[0333] To investigate the transcriptomic profiles of tumor-infiltrating Treg, we extracted them from CD4+ T cells as described above. Based on transcriptional markers, we identified seven distinct Treg subclusters within the 75,480 Treg. Using subcluster specific DEGs and canonical human Treg markers, we annotated each subset to define its phenotype. Clusters C01 and C06 were enriched for TNFR-super-family genes (TNRFSF18 / GITR, TNFRSF4 / OX40, TNFRSF9 / 4-1BB), which are associated with activated / suppressive Treg. C06 also exhibited an enrichment of TNF signaling genes (NFKBID and NFKBIA', C06_NFKBIhi_TNFRhi_Treg), implicating it harbors a more activated state than C01. C01 expressed TNRFSF18 / GITR, but lacked 77VFRSF4 / OX40, suggesting a lesser activated profile (C01_TNFRhi_Treg). Naive-mem ory-associated genes (SELL, KLF2, TCF7) defined cluster C02 (C02_KLF2hi_naive / memory_Treg). Cluster C03 showed an enrichment of IKZF2 / HELIOS and RUNX1 / AML1 (C03_HELIOShi_Treg), genes implicated in Treg function and stability. C04 was marked by GIMAP family genes (GIMAP4, GIMAP7) and LAG3 (C04_GIMAPhi_Treg). IFN-I response genes (ISG15, MX1, IFIT1) wereenriched in C05 (C05_ISGhi_Treg), while proliferation markers (MKI67, MCM7) were found in C07 (C07_MKI67hi_cycling_Treg).

[0334] Nivo+lpi possibly increased C01_TNFRhi_Treg frequencies post-treatment in pTR-2 patients (p=0.136), while C02_KLF2hi_naive / memory_Treg increased selectively in Nivo+Rela pTR-2 patients (p=0.12). C05_ISGhi_Treg proportions decreased in both doublet arms, with borderline significance in Nivo+Rela (p=0.073). Additionally, Nivo+Rela reduced C07_MKI67hi_cycling_Treg frequencies posttreatment in all patients (p=0.106), contrasting with a significant increase in Nivo+lpi-treated pTR-2 patients (p=0.042), suggesting treatment-specific effects on proliferating Treg.

[0335] GSEA revealed a downregulation of a naive-signature, IFN-a and IFN-y response signatures in Treg of post-treatment Nivo+lpi pTR-2 patients (FIG. 15A).Interestingly, cytotoxicity-related signatures were enriched post-treatment. In contrast, Nivo+Rela pTR-2 patients showed enrichment of OXPHOS and naive pathways posttreatment, with TNFA, IL6-JAK-STAT3 and IL2-STAT5 signaling enriched at baseline (FIG. 15A)

[0336] Given the prominent overexpression of IFN-response pathways in pTR-2 at baseline in both combination treatment arms, we further analyzed the role of IFNG expressing Treg previously reported in the context of fragile / unstable Treg in immunotherapy response. IFNGhi Treg were mainly located within C04_GIMAPhi_Treg. Notably, the proportion of IFNGhi Treg was significantly increased post-treatment in Nivo+lpi pTR-0 and pTR-2 (pTR-0: p=0.039; pTR-2: p=0.021) but not in Nivo+Rela indicating a treatment specific effect (FIG. 15B). In contrast, the proportion of IFNGhi 292 Treg was decreased only in Nivo+Rela pTR-2 with no change observed in pTR-0 (p=0.0007, FIG. 15B). A previous report described an increased population of fragile Treg in peripheral blood of metastatic melanoma patients responding to Nivo+Rela. Here, a phenotype of unstable Treg was characterized by a loss of F0XP3 and beside other markers, an increased T-bet (TBX21) and LAG3 expression. In our study, T-bethi Treg were mainly located in the suppressive C06_NFKBIhi_TNFRhi_Treg population and did not overlay with IFNGhi Treg suggesting that fragile and unstable Treg might represent two distinct subpopulations but ultimately supporting a theory of a lineage-directed Treg phenotype plasticity where Treg cells adopt key phenotypic / transcriptional features of the effector population they’re suppressing.

[0337] Transcriptional profiling of tumor-infiltrating Treg post-treatment revealed distinct patterns between pTR-0 and pTR-2 patients (FIGS. 15C-15D). In Nivo+lpi-treated pTR-2 patients, Treg had an enrichment of genes involved in development and survival (GIMAP4, GIMAP7) and chemotaxis (IL16). In contrast, pTR-0 had an enrichment of activation and TCR signaling (JUND, FOSB, DUSP4, PRDM1), and cytokine / cytokine receptor genes (IL1R1 and IL1R2) (FIG. 15C). In the Nivo+Rela arm, pTR-2 showed an enrichment of genes associated with chemotaxis / adhesion (IL16, ICAM3, CXCR3 commonly found on TH1 -like Treg), and IL 10RA (FIG. 15D). Treg from pTR-0 patients had an enrichment of TCR signaling (FOSB, DUSP4, NFKBIA), MHC-II expression (HLA-DRA, HLA-DQA1, HLA-DQA2, HLA-DRB5), naive-memory phenotype (KLF2) and IFN-I response genes (IFI6 and ISG15) (FIG. 15D). Shared features across doublet arms included upregulation of IL16 and FCMR in pTR-2, while DUSP4, TNFAIP3, and PRD / W7 / Blimp-1 (linked to Treg suppression) were elevated in pTR-0 patients (FIGS. 15C-15D). LAIR2 displayed a treatment specific expression pattern, with enrichment being observed in Nivo+Rela pTR-0 and Nivo+lpi pTR-2 patients post-treatment (FIGS. 15C-15D).

[0338] Combination treatments transform specific Tconv phenotypes to distinct effector states in pTR-2 patients

[0339] To investigate therapy-induced phenotypic shifts within the Tconv compartment, we performed pseudotime analysis using Slingshot, originating from C02_KLF2hi_Tnaive cluster. Four trajectories were identified. In trajectory 4, only Nivo+Rela pTR-2 showed a high density of cells toward the end of the trajectory (ending in C05_ICRhi_TH1) at baseline, confirming the association of baseline ICR expression on Tconv in Nivo+Rela pTR-2 patients. Meanwhile, Nivo+lpi treated patients had a higher density of cells in the first half, more naive part of the trajectory and less cells in the C05_ICRhi_TH1 cluster (FIG. 16A). Here, a post-treatment shift in Tconv density towards the C05_ICRhi_TH1 area of the trajectory was observed only in Nivo+lpi treated patients regardless of response (Fig. 16A). Heatmaps with genes representing the displayed phenotypes in trajectory 4 post-treatment confirm the shift toward ICRhi Tconv exclusively in Nivo+lpi pTR-2 (FIG. 16A).

[0340] To determine whether the observed changes in Tconv phenotypes reflect expansion or reprogramming of existing clones or detection of new clones posttreatment, we performed clonal tracking using scTCRseq. Clonotypes detected at both baseline and post-treatment were classified as “pre-existing,” while those found onlypost-treatment were defined as “newly detected”. We separated clones based on cell number with n=1 from clones with n>2 numbers. In both doublet regimens and response groups (pTR-0 and pTR-2), newly detected clones dominated the posttreatment repertoire regardless of response, with lesser pre-existing clones present (FIG. 16B). Here, most newly detected clones mapped to Tconv clusters C01-C05 while ICRhi Tconv (C05) were the dominant phenotype in pre-existing clones baseline and post-treatment.

[0341] In Nivo+lpi treated patients, STARTRAC-based clonal sharing analysis revealed a distinct transition of baseline C05_ICRhi_TH1 and C09_MKI76hi_Tcycling with post-treatment C10_TOXhi_TFH. Additional TCR sharing between baseline C09_MKI76hi343 _Tcycling and post-treatment C12_TNFhi_TH1 cells was observed only in Nivo+lpi pTR-2 suggesting the recruitment of Tconv with a lesser effector state at baseline towards a more activated phenotype post-treatment. In contrast, Nivo+Rela pTR-2 patients exhibited TCR sharing between baseline C02_KLF2hi_Tnaive and post-treatment C03_MHCIIhi347 _Tconv, with lesser sharing observed between baseline C11_ISGhi_Tconv and post-treatment C03_MHCIIhi_Tconv.

[0342] To further elaborate on our TCR analysis, we profiled the transcriptomes of expanded TCR clones. Expanded clones were defined as clones with an increase in proportion in the TCR repertoire of a sample post therapy, as compared to baseline. Nivo+lpi showed a higher amount of Tconv originated from expanded clones in pTR-2 patients compared to Nivo+Rela (Nivo+lpi: 23.41%; Nivo+Rela: 5.64%). In contrast, the proportion of Tconv from expanded clones in pTR-0 was higher in the Nivo+Rela arm (Nivo+lpi: 11.24%; Nivo+Rela: 19.78%). Most of the cells from expanded clones presented an ICR enriched phenotype (C05_ICRhi357 _TH1) in all responders and treatment arms. To explore the transcriptome of the expanded Tconv clones, we compared DEGs between expanding and non-expanding clones in pTR-2 patients. Expanding clones revealed an upregulation of genes associated with cytotoxicity (CST7, GZMM, GZMA, GZMB), ICRs (PDCD1, LAG3, HAVR2), activation (PRDM1, TOX, HLA-DPB1, TNFRSF18, TNFRS4, ZEB2, KLRB1), chemokine-chemokine receptors (CXCL13, CXR6, CCL4, CCL5) (FIG. 16C).

[0343] In conclusion, Nivo+lpi seems to drive activation of pre-existing effectorbased and proliferative clones, while Nivo+Rela promotes differentiation of naive-like clones toward an MHC-ll-enriched phenotype in pTR-2 patients.

[0344] Nivo+Rela reprograms TNFRhi 367 Treg to a more memory phenotype in major responders

[0345] We utilized Slingshot pseudotime analysis to gain a comprehensive understanding of therapy-induced phenotypic shifts within the Treg compartment. Three trajectories were identified. Nivo+lpi had a minimal effect on Treg trajectories in both pTR-0 and pTR-2 patients. Conversely, Nivo+Rela pTR-0 patients showed a baseline predominance of C02_KLF2hi_naive / memory_Treg, which shifted posttreatment toward activated C01_TNFRhi_Treg cluster (trajectory 1 and 2) or C04_GIMAPhi_Treg states (trajectory 3). In Nivo+Rela pTR-2 patients, the baseline Treg had activated C01_TNFRhi_Treg and C05_ISGhi_Treg phenotypes, particularly along trajectories 1 and 3 (FIG. 17A). Strikingly, post-treatment, the Treg transitioned to a more naive-memory state, with increased density in the C02_KLF2hi_naive / memory_Treg cluster (FIG. 17A). This phenotype shift was accompanied by downregulation of activation (TNFRSF18, TNFRSF4, TNFRSF9, ICOS, IKZF2) and IFN-I response genes (ISG15, IFI6), with concurrent enrichment of naive-memory genes (IL7R, CCR7, SELL, KLF2, LEF1, TCF7) (FIG. 17A). These data highlight Nivo+Rela-specific phenotype changes of TNFRhi activated and IFN-I-responsive Treg toward a less differentiated, naive-memory phenotype exclusively in pTR-2 patients.

[0346] To evaluate therapy-induced changes in Treg clonal dynamics, we analyzed TCR-defined pre-existing and newly detected clones in all pTR-2 patients. Similar to Tconv, newly detected clones dominated the post-treatment repertoire, with lesser pre-existing clones present in both doublet regimens and response groups (FIG.17B). Most pre-existing and newly detected clones presented a C01_TNFRhi_Treg and C02_KLF2hi_naive / memory_Treg phenotype. Following Nivo+Rela, C02_KLF2hi_naive / memory_Treg proportion increased non significantly (p=0.134), while C05_ISGhi389 _Treg decreased across both arms.

[0347] To further investigate the therapy-induced phenotype changes, we performed TCR sharing analysis using STARTRAC. In Nivo+lpi pTR-2 patients, TCR sharing was noted between baseline C06_NFKBIhi_TNFRhi_Treg and post-treatment C03_HELIOShi_Treg. Only in Nivo+Rela p-TR-2 patients, TCR sharing was observed between baseline C01_TNFRhi_Treg and C05_ISGhi_Treg clones with posttreatment C02_KLF2hi_naive / memory_Treg. These results underline the Nivo+Rela-induced reprogramming of activated Treg toward a naive-memory phenotype already described above.

[0348] The proportion of Treg originally from expanded clones was lower in pTR-2 compared to pTR-0 in both combination arms (Nivo+lpi pTR-2: 4.21%; Nivo+Rela pTR-2: 9.88%; Nivo+lpi pTR-0: 13.57%; Nivo+Rela pTR-0: 20.19%). The majority of Treg from expanded clones in pTR-2 had a C01_TNFRhi_Treg phenotype. In contrast to Nivo+Rela pTR-2, no naive-memory Treg (C02_KLF2hi_naive / memory_Treg) were found in expanded Treg in pTR-0. Interestingly, expanding Treg clones presented an upregulation of the chemokine receptor CXCR6 and suppressive surface lysosomal protease CTSC while non-expanding Treg were enriched with KLF2, a naive-memory marker gene (FIG. 17C). These findings highlight the role of expanding Treg with specific trafficking and activation capabilities in therapy resistance.

[0349] Treg-to-Tconv conversion selectively occurs in Nivo+Rela-treated patients

[0350] To explore potential Treg-to-Tconv plasticity, previously reported in contexts involving naive Tconv or TH17 cells transitioning to Treg and vice versa, we integrated TCR tracking with CD4+ T cell phenotyping. Treg-to-Tconv plasticity was only observed in the Nivo+Rela arm among clones with >2 cells. TCR sharing was detected between baseline C01_TNFRhi_Treg with post-treatment C03_MHCIIhi_Tconv and C06_IKZF2hi_Tconv in both pTR-0 and pTR-2 patients. Additionally, in pTR-0 patients, shared TCR were found between baseline C03_MHCIIhi_Tconv and post-treatment C07_MKI67hi_cycling_Treg. However, no clonal overlap was observed between baseline Treg subsets and post-treatment C04_IL17A_TH17 cells, or vice versa, indicating that TH17-Treg transitions were not evident in this setting.

[0351] Receptor-Ligand analysis revealed treatment-specific T cell interactions at baseline and post-treatment in pTR-2 patients

[0352] To identify T cell receptor-ligand (R-L) interactions between CD3+ T cell subsets that are associated with response and resistance to immunotherapy and to expand the repertoire of potential interactions, we combined our CD4+ T cell dataset with the published CD8+ T cell dataset. At baseline and post-treatment, the number of unique (i.e. interactions that were only present in a given normalized condition) R-L interactions in Nivo+lpi pTR-2 was low compared to Nivo+Rela pTR-2 (9 versus 25 for baseline and 15 versus 37 for post-treatment). Interestingly, there were more uniqueR-L interactions detected post-treatment in both treatment arms than baseline (Nivo+lpi:+60%; Nivo+Rela: +48%). Notably, the majority of those interactions are outgoing from CD8+ T cells post-treatment in Nivo+Rela pTR-2. Analyzing the differential regulation of intercellular interaction pairs (i.e. those interactions that are present in multiple conditions but have a differential inferred interaction strength) between baseline and post-treatment in Nivo+lpi pTR-2 (267 pairs) and Nivo+Rela pTR-2 (342 pairs) revealed a strong TNFSF9-TNFRSF9 interaction between Tconv and Treg subpopulations, especially involving C06_Treg_NFKBIhi_TNFRhi in Nivo+Rela pTR-2. Interestingly, there were no interactions with higher strength involving Treg in Nivo+lpi pTR-2 post-treatment compared to baseline. Notably, Nivo+lpi pTR-0 exhibits a larger number of unique R-L interactions post-treatment compared to Nivo+lpi pTR-2, whereas the number of interactions was comparable in Nivo+Rela responders.

[0353] Comparing the differential regulation of intercellular interaction pairs between pTR-2 and pTR-0 post-treatment exposed a much higher number of detected pairs in Nivo+lpi compared to Nivo+Rela (Nivo+lpi: 637 pairs; Nivo+Rela: 120 pairs). Furthermore, it confirmed the Nivo+Rela pTR-2 specific involvement of TNFSF9-TNFRSF9 interaction of C06_Treg_NFKBIhi_TNFRhi, especially with C04_TH17_IL17Ahi that was not prominent in Nivo+Rela pTR-0 and much lower in the Nivo+lpi arm. Strikingly, in Nivo+lpi patients, there was a high number of differentially expressed intercellular interaction pairs posttreatment especially in pTR-0. Here, an increased association of CD8+ T cells and Tconv involving ITGA1 and TNF was detected in pTR-2 while pTR-0 exhibit interactions involving CD200R1, IL17A, SEMA7A and SLAMF7 indicating a targetable mechanism of Nivo+lpi resistance.

[0354] The present invention has been described with reference to certain exemplary embodiments, dispersible compositions and uses thereof. However, it will be recognized by those of ordinary skill in the art that various substitutions, modifications or combinations of any of the exemplary embodiments may be made without departing from the spirit and scope of the invention. Thus, the invention is not limited by the description of the exemplary embodiments, but rather by the appended claims as originally filed.

Claims

THE INVENTION CLAIMED IS1. A method of treating a patient having a tumor, comprising: determining whether one or more immune cells from the patient express an exhaustion phenotype; andadministering to the patient one or more therapeutic compositions, wherein:the therapeutic composition comprises either:one or more checkpoint inhibitors (ICIs) in an amount effective to treat the tumor when the one or more immune cells express the exhaustion phenotype; orone or more compounds effective to prime the one or more immune cells to respond to one or more ICIs when the one or more immune cells do not express the exhaustion phenotype.

2. The method of claim 1, further comprising, when the tumor cell does not express the exhaustion phenotype, administering one or more ICIs.

3. The method of claim 1, wherein the one or more ICIs comprise one or more PD-1 inhibitors, PD-L1 -inhibitors, CTLA-4 inhibitors, and / or LAG-3 inhibitors.

4. The method of claim 1, wherein the one or more ICIs comprise one or more antibodies that bind PD-1, PD-L1, CTLA-4, and / or LAG-3.

5. The method of claim 1, wherein the one or more ICIs comprise one or more of nivolumab, pembrolizumab, cemiplimab, atezolizumab, durvalumab, avelumab, ipilimumab, relatlimab, fianlimab, and / or leramlimab.

6. The method of claim 1, wherein the one or more ICIs comprise a PD-1 inhibitor and a LAG-3 inhibitor.

7. The method of any of claims 1-6, wherein the one or more ICIs comprise nivolumab and relatlimab.

8. The method of claim 1, wherein the one or more compounds effective to prime the one or more immune cells to respond to one or more ICIs comprise one or more of a STING agonist and a TLR9 agonist.

9. The method of claim 1, wherein the one or more compounds effective to prime the one or more immune cells to respond to one or more ICIs comprise one or more of cGAMP, DMXAA, MSA-2, TAK-676, a CpG oligodeoxynucleotide, IMO-2055, and / or lefitolimod.

10. The method of claim 1, wherein the one or more immune cells comprise one or more CD8+ cells, optionally CD8+ TIL cells.

11. The method of claim 1, wherein the one or more immune cells comprise one or more NK cells, optionally tumor-infiltrating NK cells.

12. The method of claim 1, wherein the one or more immune cells comprise one or more CD4+ cells, optionally CD4+ conventional T cells or CD4+ regulatory T cells.

13. The method of claim 1, wherein the one or more immune cells are one or more CD8+ cells and the exhaustion phenotype comprises increased expression of one or more of ACP5, CCL3, CD38, CTSW, CXCL13, EIF2AK2, ENTPD1, EPSTI1, GNLY, HAVCR2, HERC5, IFI6, IFI44, IFI44L, IFIT1, ISG15, ISG20, LAG3, MX1, OASL, OAS1, OAS2, PARP14, PLSCR1, PTMS, RBPJ, RSAD2, STAT1, TOX, and / or XAF1 compared to a cell not expressing the exhaustion phenotype.

14. The method of claim 1, wherein the one or more immune cells are immune cell is an NK cell and the exhaustion phenotype comprises an increase in one or more of MTOR signaling, oxidative phosphorylation, interferon-[3 signature, and / or interferon-a signature compared to a cell not expressing the exhaustion phenotype.

15. The method of claim 1, wherein the immune cell is a CD4+ cell and the exhaustion phenotype comprises increased expression of one or more ofLAG3, CTLA4, PDCD1, ENTPD1, HAVCR2, and / or TIGIT compared to a cell not expressing the exhaustion phenotype.

16. The method of claim 1, wherein the therapeutic composition is administered orally.

17. The method of claim 1, wherein the therapeutic composition is administered parenterally.

18. The method of claim 1, wherein the tumor is a squamous cell tumor.

19. The method of claim 1, wherein the patient has head and neck squamous cell carcinoma.

20. A method of treating a patient having a tumor, comprising: administering, to a patient having one or more immune cells expressing an exhaustion phenotype, a therapeutic composition comprising one or more ICIs in an amount effective to treat the tumor.

21. The method of claim 20, wherein the one or more ICIs comprise one or more PD-1 inhibitors, PD-L1 -inhibitors, CTLA-4 inhibitors, and / or LAG-3 inhibitors.

22. The method of claim 20, wherein the one or more ICIs comprise one or more antibodies that bind PD-1, PD-L1, CTLA-4, and / or LAG-3.

23. The method of claim 20, wherein the one or more ICIs comprise one or more of nivolumab, pembrolizumab, cemiplimab, atezolizumab, durvalumab, avelumab, ipilimumab, relatlimab, fianlimab, and / or leramlimab.

24. The method of claim 20, wherein the one or more ICIs comprise a PD-1 inhibitor and a LAG-3 inhibitor.

25. The method of claim 20, wherein the one or more ICIs comprise nivolumab and relatlimab.

26. The method of claim 20, wherein the one or more immune cells comprise one or more CD8+ cells, optionally CD8+ TIL cells.

27. The method of claim 20, wherein the one or more immune cells comprise one or more NK cells, optionally tumor-infiltrating NK cells.

28. The method of claim 20, wherein the one or more immune cells comprise one or more CD4+ cells, optionally CD4+ conventional T cells or CD4+ regulatory T cells.

29. The method of claim 20, wherein the one or more immune cells are one or more CD8+ cells and the exhaustion phenotype comprises increased expression of one or more of ACP5, CCL3, CD38, CTSW, CXCL13, EIF2AK2, ENTPD1, EPSTI1, GNLY, HAVCR2, HERC5, IFI6, IFI44, IFI44L, IFIT1, ISG15, ISG20, LAG3, MX1, OASL, OAS1, OAS2, PARP14, PLSCR1, PTMS, RBPJ, RSAD2, STAT1, TOX, and / or XAF1 compared to a cell not expressing the exhaustion phenotype.

30. The method of claim 20, wherein the one or more immune cells are immune cell is an NK cell and the exhaustion phenotype comprises an increase in one or more of MTOR signaling, oxidative phosphorylation, and / or interferon-a signature compared to a cell not expressing the exhaustion phenotype.

31. The method of claim 20, wherein the immune cell is a CD4+ cell and the exhaustion phenotype comprises increased expression of one or more of LAG3, CTLA4, PDCD1, ENTPD1, HAVCR2, and / or TIGIT compared to a cell not expressing the exhaustion phenotype.

32. The method of claim 20, wherein the therapeutic composition is administered orally.

33. The method of claim 20, wherein the therapeutic composition is administered parenterally.

34. The method of claim 20, wherein the tumor is a squamous cell tumor.

35. The method of claim 20, wherein the patient has head and neck squamous cell carcinoma.

36. A method of treating a patient having a tumor, comprising: administering, to a patient having one or more immune cells that does not express an exhaustion phenotype, a first therapeutic composition followed by a second therapeutic composition, wherein the first therapeutic composition comprises one or more compounds effective to prime the one or more immune cells to respond to one or more ICIs and the second therapeutic composition comprises one or more ICIs.

37. The method of claim 36, wherein the one or more compounds effective to prime the one or more immune cells comprise one or more of a STING agonist and a TLR9 agonist.

38. The method of claim 36, wherein the one or more compounds effective to prime the one or more immune cells comprise one or more of cGAMP, DMXAA, MSA-2, TAK-676, a CpG oligodeoxynucleotide, IMO-2055, and / or lefitolimod.

39. The method of claim 36, wherein the one or more ICIs comprise one or more PD-1 inhibitors, PD-L1 -inhibitors, CTLA-4 inhibitors, and / or LAG-3 inhibitors.

40. The method of claim 36, wherein the one or more ICIs comprise one or more antibodies that bind PD-1, PD-L1, CTLA-4, and / or LAG-3.

41. The method of claim 36, wherein the one or more ICIs comprise one or more of nivolumab, pembrolizumab, cemiplimab, atezolizumab, durvalumab, avelumab, ipilimumab, relatlimab, fianlimab, and / or leramlimab.

42. The method of claim 36, wherein the one or more ICIs comprise a PD-1 inhibitor and a LAG-3 inhibitor.

43. The method of claim 36, wherein the one or more ICIs comprise nivolumab and relatlimab.

44. The method of claim 36, wherein the one or more immune cells comprise one or more CD8+ cells, optionally CD8+ TIL cells.

45. The method of claim 36, wherein the one or more immune cells comprise one or more NK cells, optionally tumor-infiltrating NK cells.

46. The method of claim 36, wherein the one or more immune cells comprise one or more CD4+ cells, optionally CD4+ conventional T cells or CD4+ regulatory T cells.

47. The method of claim 36, wherein the first and / or second therapeutic compositions are administered orally.

48. The method of claim 36, wherein the first and / or second therapeutic compositions are administered parenterally.

49. The method of claim 36, wherein the tumor is a squamous cell tumor.

50. The method of claim 36, wherein the patient has head and neck squamous cell carcinoma.

51. A method of treating a patient having a tumor, comprising:determining whether one or more immune cells from the patient express a type-1 IFN response signature; andadministering to the patient one or more therapeutic compositions, wherein:the therapeutic composition comprises either:one or more checkpoint inhibitors (ICIs) in an amount effective to treat the tumor when the one or more immune cells express the type-1 IFN response signature; orone or more compounds effective to prime the one or more immune cells to respond to one or more ICIs when the one or more immune cells do not express the type-1 interferon response signature.

52. The method of claim 51, wherein the type-I interferon response signature comprises an increased expression of one or more of IFI44, OASL, CCL3, OAS1, IFIT1, ISG15, MX1, IFI6, LAG3, and / or IFI44L compared to a cell not expressing the type-I interferon response signature.

53. The method of claim 51, wherein the one or more compounds effective to prime the one or more immune cells comprise one or more of a STING agonist and a TLR9 agonist.

54. The method of claim 51, wherein the one or more compounds effective to prime the one or more immune cells comprise one or more of cGAMP, DMXAA, MSA-2, TAK-676, a CpG oligodeoxynucleotide, IMO-2055, and / or lefitolimod.

55. The method of claim 51, wherein the one or more ICIs comprise one or more PD-1 inhibitors, PD-L1 -inhibitors, CTLA-4 inhibitors, and / or LAG-3 inhibitors.

56. The method of claim 51, wherein the one or more ICIs comprise one or more antibodies that bind PD-1, PD-L1, CTLA-4, and / or LAG-3.

57. The method of claim 51, wherein the one or more ICIs comprise one or more of nivolumab, pembrolizumab, cemiplimab, atezolizumab, durvalumab, avelumab, ipilimumab, relatlimab, fianlimab, and / or leramlimab.

58. The method of claim 51, wherein the one or more ICIs comprise a PD-1 inhibitor and a LAG-3 inhibitor.

59. The method of claim 51, wherein the one or more ICIs comprise nivolumab and relatlimab.

60. The method of claim 51, wherein the one or more immune cells comprise one or more CD8+ cells, optionally CD8+ TIL cells.

61. The method of claim 51, wherein the one or more immune cells comprise one or more NK cells, optionally tumor-infiltrating NK cells.

62. The method of claim 51, wherein the one or more immune cells comprise one or more CD4+ cells, optionally CD4+ conventional T cells or CD4+ regulatory T cells.

63. The method of claim 51, wherein the therapeutic composition is administered orally.

64. The method of claim 51, wherein the therapeutic composition is administered parenterally.

65. The method of claim 51, wherein the tumor is a squamous cell tumor.

66. The method of claim 51, wherein the patient has head and neck squamous cell carcinoma.

67. A method of treating a patient having a tumor, comprising:administering, to a patient having one or more immune cells expressing a type-1 interferon response signature, a therapeutic composition comprising one or more ICIs in an amount effective to treat the tumor.

68. The method of claim 67, wherein the type-1 interferon response signature comprises an increased expression of one or more of IFI44, OASL, CCL3, OAS1, IFIT1, ISG15, MX1, IFI6, LAG3, and / or IFI44L compared to a cell not expressing the type-l interferon response signature.

69. The method of claim 67, wherein the one or more ICIs comprise one or more PD-1 inhibitors, PD-L1 -inhibitors, CTLA-4 inhibitors, and / or LAG-3 inhibitors.

70. The method of claim 67, wherein the one or more ICIs comprise one or more antibodies that bind PD-1, PD-L1, CTLA-4, and / or LAG-3.

71. The method of claim 67, wherein the one or more ICIs comprise one or more of nivolumab, pembrolizumab, cemiplimab, atezolizumab, durvalumab, avelumab, ipilimumab, relatlimab, fianlimab, and / or leramlimab.

72. The method of claim 67, wherein the one or more ICIs comprise a PD-1 inhibitor and a LAG-3 inhibitor.

73. The method of claim 67, wherein the one or more ICIs comprise nivolumab and relatlimab.

74. The method of claim 67, wherein the one or more immune cells comprise one or more CD8+ cells, optionally CD8+ TIL cells.

75. The method of claim 67, wherein the one or more immune cells comprise one or more NK cells, optionally tumor-infiltrating NK cells.

76. The method of claim 67, wherein the one or more immune cells comprise one or more CD4+ cells, optionally CD4+ conventional T cells or CD4+ regulatory T cells.

77. The method of claim 67, wherein the therapeutic composition is administered orally.

78. The method of claim 67, wherein the therapeutic composition is administered parenterally.

79. The method of claim 67, wherein the tumor is a squamous cell tumor.

80. The method of claim 67, wherein the patient has head and neck squamous cell carcinoma.

81. A method of treating a patient having a tumor, comprising: administering, to a patient having one or more immune cells that does not express a type-I interferon response signature, a first therapeutic composition followed by a second therapeutic composition, wherein the first therapeutic composition comprises one or more compounds effective to prime the one or more immune cells to respond to one or more ICIs and the second therapeutic composition comprises one or more ICIs.

82. The method of claim 81, wherein the type-I interferon response signature comprises an increased expression of one or more of IFI44, OASL, CCL3, OAS1, IFIT1, ISG15, MX1, IFI6, LAG3, and / or IFI44L compared to a cell not expressing the type-I interferon response signature.

83. The method of claim 81, wherein the one or more compounds effective to prime the one or more immune cells comprise one or more of a STING agonist and a TLR9 agonist.

84. The method of claim 81, wherein the one or more compounds effective to prime the one or more immune cells comprise one or more of cGAMP, DMXAA, MSA-2, TAK-676, a CpG oligodeoxynucleotide, IMO-2055, and / or lefitolimod.

85. The method of claim 81, wherein the one or more ICIs comprise one or more PD-1 inhibitors, PD-L1 -inhibitors, CTLA-4 inhibitors, and / or LAG-3 inhibitors.

86. The method of claim 81, wherein the one or more ICIs comprise one or more antibodies that bind PD-1, PD-L1, CTLA-4, and / or LAG-3.

87. The method of claim 81, wherein the one or more ICIs comprise one or more of nivolumab, pembrolizumab, cemiplimab, atezolizumab, durvalumab, avelumab, ipilimumab, relatlimab, fianlimab, and / or leramlimab.

88. The method of claim 81, wherein the one or more ICIs comprise a PD-1 inhibitor and a LAG-3 inhibitor.

89. The method of claim 81, wherein the one or more ICIs comprise nivolumab and relatlimab.

90. The method of claim 81, wherein the one or more immune cells comprise one or more CD8+ cells, optionally CD8+ TIL cells.

91. The method of claim 81, wherein the one or more immune cells comprise one or more NK cells, optionally tumor-infiltrating NK cells.

92. The method of claim 81, wherein the one or more immune cells comprise one or more CD4+ cells, optionally CD4+ conventional T cells or CD4+ regulatory T cells.

93. The method of claim 81, wherein the first and / or second therapeutic compositions are administered orally.

94. The method of claim 81, wherein the first and / or second therapeutic compositions are administered parenterally.

95. The method of claim 81, wherein the tumor is a squamous cell tumor.

96. The method of claim 81, wherein the patient has head and neck squamous cell carcinoma.

97. Use of a composition comprising an immune checkpoint inhibitor to treat cancer in a patient having one or more immune cells expressing an exhaustion phenotype and / or a type-l interferon response signature.

98. Use of a composition comprising compound effective to prime an immune cell to respond to a composition comprising an immune checkpoint inhibitor and an immune checkpoint inhibitor to treat cancer in a patient having one or more immune cells that do not express an exhaustion phenotype and / or a type-l interferon response signature.

Images