Biomarkers for use in cancer treatment and predicting responsiveness to a cancer therapy
The use of biomarkers like IFI30, FCN1, HLA-DRA, CXCL8, and SLC11A1 helps predict immunotherapy responsiveness in cancer patients, addressing the limitations of current biomarkers and improving treatment outcomes for esophageal cancer by identifying suitable candidates for immunotherapy.
Patent Information
- Application Number
- PCT/EP2024/059267
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-01
- Filing Date
- 2024-04-04
- Publication Date
- 2024-11-07
AI Technical Summary
Current biomarkers for predicting responsiveness to immunotherapy in cancer treatment, such as PD-L1 expression levels, have shown variable predictive utility, particularly in combination with chemotherapy, and there is a need for novel approaches to identify patients who would benefit most from immunotherapy or immunochemotherapy, especially in esophageal cancer where conventional markers are ineffective.
The use of specific biomarkers like IFI30, FCN1, HLA-DRA, CXCL8, SLC11A1, and other T-cell inflammation signatures to determine the likelihood of response to immunotherapy by measuring their expression levels in tumor samples, which can indicate prolonged survival and treatment effectiveness.
These biomarkers effectively predict patient responsiveness to immunotherapy, improving the selection of appropriate cancer treatments and enhancing overall survival rates by identifying patients likely to benefit from immunotherapy and immunochemotherapy.
Smart Images

Figure EP2024059267_07112024_PF_FP_ABST
Abstract
Description
[0001] BIOMARKERS FOR USE IN CANCER TREATMENT AND PREDICTING RESPONSIVENESS TO A CANCER THERAPY
[0002] CROSS-REFERENCE TO RELATED APPLICATIONS
[0003] This application claims priority under 35 U.S.C. §119(e) to U.S. Provisional Patent Application No. 63 / 499,390, filed May 1, 2023. The foregoing application is incorporated by reference herein in its entirety.
[0004] FIELD OF THE INVENTION
[0005] This invention relates generally to biomarkers for use in cancer treatment and predicting responsiveness to a cancer therapy.
[0006] BACKGROUND OF THE INVENTION
[0007] Recent advancements in tumor immunology have established immunotherapy with immune checkpoint inhibitors (ICIs) as the fourth pillar of cancer care, in addition to chemotherapy (CTX), radiotherapy, and surgery. An increasing number of cancer patients have been treated with ICI agents, either alone or in combination with concurrent CTX (immunochemotherapy, also referred to herein as “ICI+CTX”). Despite exciting improvements in patient outcomes with ICI across multiple settings, only a minority of patients attain long-term benefits from these agents. With the expanding number of regulatory approvals and thousands of ongoing clinical trials for immunotherapies, improving the ability to select patients who are most likely to respond to ICI- based therapies remains a key challenge. Biomarkers predicting good patient outcomes on ICI may not predict outcomes with CTX alone or immunochemotherapy, as ICI acts by stimulating the immune system, whereas CTX with DNA-damaging agents can impair both cancer and non- malignant immune cells. Therefore, uncoupling ICI-specific signals from the confounding effect of CTX is one of the most important impediments to characterizing biomarkers that could help select patients for the growing range of ICI+CTX indications.
[0008] Esophageal cancer is the sixth leading cause of cancer mortality worldwide. The predominant subtype in the Western world, esophageal adenocarcinoma (EAC), is among the cancer types with the highest increase in incidence over the past few decades. About 40% of esophageal cancers present with distant metastases at diagnosis. For these inoperable patients, median overall survival (OS) with conventional first-line fluoropyrimidine and platinum CTX agents is less than one year. In 2021, the U.S. Food and Drug Administration (FDA) approved first-line ICI+CTX regimens using anti-PD-1 ICI in addition to platinum / fluoropyrimidine CTX for inoperable gastro-esophageal cancers. However, conventional predictive biomarkers for ICI have shown variable predictive utility for this new ICI+CTX treatment. For example, PD-L1 expression level is not associated with improved outcomes in other anti-PD-1 ICI+CTX gastroesophageal cancer trials (Kang, etal., 2022. Lancet Oncol 23, 234-247).
[0009] Therefore, there remains a pressing need for novel approaches for identifying cancer patients who would benefit most from immunotherapy or immunochemotherapy.
[0010] SUMMARY OF THE INVENTION
[0011] This disclosure addresses the need mentioned above in a number of aspects. In one aspect, this disclosure provides a method of identifying a subject as likely responding to treatment of an immunotherapy or predicting survival of the subject undergoing the immunotherapy. In some embodiments, the method comprises: (a) determining an expression level of each of biomarkers comprising one or more of IFI30, FCN1, HLA-DRA, CXCL8, and SLC11A1 in a sample from the subject; (b) determining a change in the expression level of each of the biomarkers as compared to a respective reference expression level for each of the biomarkers; and (c) identifying the subject as likely responding to the treatment of the immunotherapy or as likely having prolonged survival after the treatment, if the expression level of each of the biomarkers is elevated as compared to the respective reference expression level for each of the biomarkers.
[0012] In some embodiments, the biomarkers comprise IFI30, FCN1, HLA-DRA, CXCL8, and SLC11A1. In some embodiments, the biomarkers consist of IFI30, FCN1, HLA-DRA, CXCL8, and SLC11A1.
[0013] In some embodiments, the biomarkers comprise one or more of FI30, FCN1, HLA-DRA, CXCL8, SLC11A1, TYROBP, LYZ, PLAUR, HLA-DRB1, and EREG. In some embodiments, the biomarkers comprise FI30, FCN1, HLA-DRA, CXCL8, SLC11 Al, TYROBP, LYZ, PLAUR, HLA-DRB1, and EREG. In some embodiments, the biomarkers consist of FI30, FCN1, HLA- DRA, CXCL8, SLC11A1, TYROBP, LYZ, PLAUR, HLA-DRB1, and EREG. In some embodiments, the sample is obtained before the treatment. In some embodiments, the sample comprises tumor tissue containing the monocytes.
[0014] In another aspect, this disclosure provides a method of identifying a subject as likely responding to treatment of an immunotherapy or predicting survival of the subject undergoing the immunotherapy. In some embodiments, the method comprises (i) determining an expression level of each of biomarkers comprising one or more of KLRD1, KLRK1, FASLG, GZMK, KLRC1, KLRC4, CD8A, PRF1, APOBEC3G, APOL3, CD244, and FCRL6 in a sample from the subject; (ii) determining a change in the expression level of each of the biomarkers as compared to a respective reference expression level for each of the biomarkers; and (iii) identifying the subject as likely responding to the treatment of the immunotherapy or as likely having prolonged survival after the treatment, if the expression level of each of the biomarkers is elevated as compared to the respective reference expression level for each of the biomarkers.
[0015] In another aspect, this disclosure provides a method of treating a cancer in a subject in need thereof. In some embodiments, the method comprises identifying a subject as likely responding to treatment of an immunotherapy or as likely having prolonged survival after the treatment according to the method as described herein, and administering to the subject a therapeutically effective amount of the immunotherapy.
[0016] In some embodiments, the expression level of each of the biomarkers is determined from the sample of the subject who has completed a treatment cycle of a cancer therapy. In some embodiments, the respective reference expression level for each of the biomarkers is determined from samples of one or more subjects before the treatment cycle of the cancer therapy. In some embodiments, the treatment cycle is about 4 weeks.
[0017] In some embodiments, the immunotherapy comprises one or more immune checkpoint inhibitors. In some embodiments, the one or more immune checkpoint inhibitors comprise a CTLA-4 inhibitor, a PD-1 inhibitor, a PD-L1 inhibitor a LAG-3 inhibitor, or a TIM-3 inhibitor.
[0018] In some embodiments, the one or more immune checkpoint inhibitors comprise an anti- CTLA-4 antibody, an anti -PD-1 antibody, an anti-PD-Ll antibody, an anti-LAG-3 antibody, an anti-TIM-3 antibody, or a combination thereof. In some embodiments, the one or more immune checkpoint inhibitors comprise an anti- PD-L1 antibody, an anti-CTLA-4 antibody, or a combination thereof. In some embodiments, the one or more immune checkpoint inhibitors comprise durvalumab. In some embodiments, the immunotherapy is administered to the subject in one or more doses or in one or more treatment cycles.
[0019] In some embodiments, the method comprises administering to the subject an additional therapeutic agent or therapy. In some embodiments, the additional therapeutic agent or therapy comprises a chemotherapy. In some embodiments, the chemotherapy comprises oxaliplatin, capecitabine, or a combination thereof.
[0020] In some embodiments, the method comprises determining a therapeutic response in the subject that comprises complete or partial response as defined by response evaluation criteria in solid tumors (RECIST), stable disease as defined by RECIST, or long-term survival despite disease progression or response as defined by immune-related response criteria (irRC).
[0021] In some embodiments, the sample comprises a biological sample that comprises a plasma sample, a blood sample, or a tissue sample. In some embodiments, the sample is obtained from a primary tumor or a metastasis. In some embodiments, the expression level comprises a mRNA or protein level.
[0022] In some embodiments, the subject has an esophageal cancer. In some embodiments, the subject has inoperable esophageal adenocarcinoma (EAC). In some embodiments, the subject has an EBV- / MSS gastric cancer.
[0023] In another aspect, this disclosure provides a method of identifying a subject as likely responding to treatment of a cancer therapy. In some embodiments, the method comprises: (a) determining a level of a tumor monocyte content (TMC) in a sample from the subject; (b) determining a change in the level of the tumor monocyte content as compared to a reference level; and (c) identifying the subject as likely responding to the treatment of the cancer therapy if the level of the tumor monocyte content is elevated as compared to the reference level, wherein the cancer therapy comprises an immunotherapy and a chemotherapy.
[0024] In another aspect, this disclosure provides a method of identifying a subject as likely responding to treatment of a cancer therapy. In some embodiments, the method comprises: (i) determining a level of a tumor monocyte content (TMC) and a level of tumor mutational burden (TMB) in a sample from the subject; (ii) determining a first change in the level of the tumor monocyte content as compared to a first reference level and a second change in the level of the tumor mutational burden as compared to a second reference level; and (iii) identifying the subject as likely responding to the treatment of the cancer therapy if the level of the tumor monocyte content is elevated as compared to the first reference level; and / or if the level of the tumor mutational burden is elevated as compared to the second reference level, wherein the cancer therapy comprises an immunotherapy and a chemotherapy.
[0025] In another aspect, this disclosure provides a method of treating a cancer with a cancer therapy in a subject in need thereof. In some embodiments, the method comprises identifying a subject who is likely responsive to the treatment of the cancer therapy comprising the immunotherapy and the chemotherapy according to the method described herein, and administering to the subject a therapeutically effective amount of the cancer therapy.
[0026] In some embodiments, the immunotherapy comprises one or more immune checkpoint inhibitors. In some embodiments, the one or more immune checkpoint inhibitors comprise a CTLA-4 inhibitor, a PD-1 inhibitor, a PD-L1 inhibitor a LAG-3 inhibitor, or a TIM-3 inhibitor.
[0027] In some embodiments, the one or more immune checkpoint inhibitors comprise an anti- CTLA-4 antibody, an anti -PD-1 antibody, an anti-PD-Ll antibody, an anti-LAG-3 antibody, an anti-TIM-3 antibody, or a combination thereof. In some embodiments, the one or more immune checkpoint inhibitors comprise an anti-PD-Ll antibody, an anti-CTLA-4 antibody, or a combination thereof. In some embodiments, the one or more immune checkpoint inhibitors comprise durvalumab.
[0028] In some embodiments, the chemotherapy comprises oxaliplatin, capecitabine, or a combination thereof.
[0029] In yet another aspect, this disclosure provides a method of identifying a subject as likely not responding to treatment of a cancer therapy. In some embodiments, the method comprises: (a) determining an expression level of each of biomarkers comprising one or more of T cell-specific interferon-stimulated genes (ISGs) in a sample from the subject; (b) determining a change in the expression level of each of the biomarkers as compared to a respective reference expression level for each of the biomarkers; and (c) identifying the subject as likely not responding to the treatment of the cancer therapy if the expression level of each of the biomarkers is elevated as compared to the respective reference expression level for each of the biomarkers.
[0030] In some embodiments, interferon-stimulated genes comprise one or more of OAS1, MX1, IFI6, and XAF1.
[0031] In some embodiments, the cancer therapy comprises an immunotherapy and optionally a chemotherapy.
[0032] The foregoing summary is not intended to define every aspect of the disclosure, and additional aspects are described in other sections, such as the following detailed description. The entire document is intended to be related as a unified disclosure, and it should be understood that all combinations of features described herein are contemplated, even if the combinations of features are not found together in the same sentence, or paragraph, or section of this document. Other features and advantages of the invention will become apparent from the following detailed description. It should be understood, however, that the detailed description and the specific examples, while indicating specific embodiments of the disclosure, are given by way of illustration only, because various changes and modifications within the spirit and scope of the disclosure will become apparent to those skilled in the art from this detailed description.
[0033] BRIEF DESCRIPTION OF THE DRAWINGS
[0034] FIGS. 1A, IB, and 1C show the design and clinical outcomes of the study LUD2015-005. FIG. 1A shows the relative chronology of treatment, sampling, and computed tomography (CT) response assessment events within LUD2015-005. On-treatment CT scans were named according to their timing relative to the ICI+CTX period (C1D8 = cycle 1, day 8). FIGS. 1B-C show Kaplan- Meier curves for (FIG. 1A) overall survival (OS) and (FIG. IB) progression-free survival (PFS) for all LUD2015-005 inoperable patients. The risk table (below) shows the number of patients at each timepoint with ongoing OS or PFS. Certain patients are highlighted to show difficulty in summarizing outcomes using response: J, = lesions grew during the ICI-4W period, but eventually attained irPR during immunochemotherapy; * = irCRbut average PFS (progressed at 9.3 months);A= PFS >12 months but no irRECIST response (two additional cases were non-assessable for target lesion size). FIG. 2 shows further detail of relative chronology of treatment, sampling, and computed tomography (CT) response assessment events within LUD2015-005 clinical trial also including operable cohorts C and D.
[0035] FIGS. 3A, 3B, 3C, and 3D show that increased TMC is an ICI+CTX-specific predictor of improved clinical outcomes. FIG. 3A shows Kaplan-Meier plots showing OS of TMC-high and - low groups, as split by the cohort-specific median, for the LUD2015-005 EAC cohort. FIG. 3B shows TMC Kaplan-Meier plots as in FIG. 3A, but for TCGA (left) and ICGC (right) cohorts of stage III and IV EAC tumors, used as reference cohorts for EAC without ICI treatment. P-value for the difference between groups was calculated by the log-rank method. FIG. 3C shows Kaplan- Meier curves comparing TMC-high (>subgroup median) and TMC-low (<subgroup median) patients for the aPD-Ll (left) and aPD-Ll+aCTLA-4 (right) subgroups. P-values were calculated by the log-rank method. FIG. 3D shows validation of TMC predictive power using an independent cohort of pooled ICI -treated GC. Deconvolution-assessed TMC values from LUD2015-005 and the pooled validation cohort were split into EBV- / MSS and EBV+ / MSI facets, and grouped by response. Response criteria for the pooled gastric cohort were RECIST (Cheong) and RECIST vl.l (Kim), differing from LUD2015-005 (irRECIST). Post-hoc review of LUD2015-005 data determined binary response calls were the same with irRECIST and RECIST vl. l criteria for all patients shown. Displayed p-values were calculated using a Mann- Whitney U-test.
[0036] FIGS. 4A, 4B, 4C, 4D, and 4E show that pre-treatment TMB and monocyte proportion are complementary predictors of ICI+CTX outcomes. FIG. 4A shows TMB for each PreTx tumor sample in the LUD2015-005 inoperable cohort, grouped by clinical benefit (CB) status. P-value between groups was calculated by Welch’s t-test. FIGS. 4B-C show Kaplan-Meier plots of TMB- high and TMB-low groups defined using the cohort median of coding mutations / Mb, for the (FIG. 4B) LUD2015-005 and (FIG. 4C) non-ICI -treated reference cohorts, split into TCGA (top) and ICGC (bottom) as in FIG. 3B. For TCGA and ICGC, only the subset of stage III and IV EAC tumors with pre-calculated mutation rates were assessed. The displayed p-value for the difference between survival curves was calculated by the log-rank method. FIG. 4D shows Kaplan-Meier plots of four groups, defined by splitting the PreTx TMC and PreTx TMB by the corresponding cohort medians for each variable. P-value is calculated by the log-rank test, testing the overall difference between the four groups. FIG. 4E shows a forest plot of multivariable Cox regression for OS (top) and PFS (bottom) with PreTx TMC (loglO-transformed) and TMB. Both values were scaled and centred prior to regression. Hazard ratio (HR) and 95% CI are shown; HR <1 indicates an association with longer survival, and >1 association with reduced survival.
[0037] FIGS. 5A, 5B, 5C, and 5D show that tumor monocyte content (TMC) at baseline was most significantly associated with outcomes from immunotherapy. FIG. 5A shows a Forest plot depicting an association of loglO-transformed PreTx tumor monocyte content (TMC) with survival in days (OS) in LUD2015-005 operable esophageal cancer (OC) patients. FIG. 5B shows a Boxplot comparing loglO-transformed PreTx TMC in LUD2015-005 operable OC patients who are alive or deceased at the end of the study (FIG. 5C) and comparing patients with and without major pathological response (mPR) at Surgery. FIG. 5D shows a Kaplan-Meier plot depicting OS of PreTx TMC-high and -low groups from LUD2015-005 operable OC patients, split by the cohort median. P-value by the log-rank test.
[0038] FIGS. 6A, 6B, and 6C show an assessment of the fit of regression models with Mallow’s Cp and the Bayesian Information Criterion. The goodness of fit of the linear models resulting from best subset regression analysis was assessed with Mallows’ Cp (FIG. 6 A) and the Bayesian Information Criterion (BIC) (FIG. 6B), displaying the number of covariates on the x axis. The subset is selected from top 30 monocyte maker genes: CD68, PPIF, KYNU, EREG, S100A12, S100A8, S100A9, BCL2A1, IL1B, CXCL8, PLAUR, G0S2, FCN1, C5AR1, SLC11A1, LYZ, SERPINA1, VCAN, TIMP1, HLA-DRB5, HLA-DRB1, HLA-DRA, HLA-DPA1, LST1, CD14, SPI1, IFI30, FCER1G, TYROBP, and AFF1. FIG. 6C shows intercept and covariates for the linear model generated from the 10-variable subset from best subset regression analysis.
[0039] FIGS. 7A, 7B, 7C, and 7D show that the monocyte marker signature recreates the monocyte signature generated from bulk transcriptomic deconvolution. FIG. 7A shows Box plots comparing PreTx 10-gene monocyte gene expression score between LUD2015-005 operable OC patients alive and deceased at the end of the study (left) and patients with and without mPR at surgery (right). Expression score is calculated by multiplying vst-stabilizedC expression with linear model coefficients. FIG. 7B shows a Kaplan-Meier plot showing OS of PreTx 10-gene expressions core high and low groups from LUD2015-005 operable OC patients, split by cohort median. P value by log rank test. FIG. 7C shows Box plots comparing PreTx 5-gene monocyte gene expression scores between LUD2015-005 operable OC patients alive and deceased at the end of the study (left) and patients with and without mPR at surgery (right). Expression score is calculated from the sum of z-scaled vst-stabilized expression. FIG. 7D shows a Kaplan-Meier plot showing OS of PreTx 5-gene expressions score high and low groups from LUD2015-005 operable OC patients, split by cohort median. P-value by the log-rank test.
[0040] FIGS. 8A, 8B, 8C, and 8D show the monocyte marker signature for predicting clinical benefits in the metastatic cohorts. FIG. 8A shows Box plots comparing PreTx 10-gene monocyte gene expression scores between LUD2015-005 metastatic OC patients alive and deceased at the end of the study (left) and patients with and without clinical benefit (PFS > 1 year) (right). Expression score was calculated by multiplying vst-stabilized expression with linear model coefficients. FIG. 8B shows a Kaplan-Meier plot showing OS of PreTx 10-gene expressions score high and low groups from LUD2015-005 metastatic OC patients, split by cohort median. P-value by the log-rank test. FIG. 8C shows Box plots comparing PreTx 5-gene monocyte gene expression score between LUD2015-005 metastatic OC patients alive and deceased at the end of the study (left) and patients with and without clinical benefit (right). Expression score is calculated from the sum of z-scaled vst-stabilized expression. FIG. 8D shows a Kaplan-Meier plot showing OS of PreTx 5-gene expressions core high and low groups from LUD2015-005 metastatic OC patients, split by cohort median. P-value by the log-rank test.
[0041] FIGS. 9A, 9B, 9C, and 9D show that the monocyte marker signature outperforms the discriminatory potential of TMC by deconvolution between patients with and without clinical benefit. FIG. 9 A shows a Forest plot depicting association of logio-transformed PreTx tumor monocyte content (TMC) and tumor mutational burden (TMB, coding mutations / Mb) with survival in days (OS) in LUD2015-005 metastatic OC patients. FIG. 9B shows a Box plot comparing logio-transformed PreTx TMC in LUD2015-005 metastatic OC patients who are alive or deceased at the end of the study (FIG. 9C) and comparing patients with and without clinical benefit (PFS > 1 year). FIG. 9D shows a Kaplan-Meier plot showing OS of PreTx TMC-high and TMC-low groups from LUD2015-005 metastatic OC patients, split by the cohort median. P-value by the log-rank test.
[0042] Figs. 10A, 10B, 10C, and 10D show TMC assessment in an independent pooled cohort of ICI-treated GC. FIG. 10A Box plots comparing PreTx 10-gene monocyte gene expression score between ICI-treated gastric cancer patients with and without response according to RECIST / RECIST vl .1. Expression score was calculated by multiplying vst-stabilized expression with linear model coefficients. All patients are shown. FIG. 10B shows that patients were faceted by EBV and microsatellite status (MSI = microsatellite instable, MSS = microsatellite stable). FIG. IOC shows Box plots comparing PreTx 5 -gene monocyte gene expression scores between all gastric cancer patients with and without response and faceted by EBV and microsatellite status (FIG. 10D)
[0043] DETAILED DESCRIPTION OF THE INVENTION
[0044] This disclosure describes novel biomarkers for use in cancer treatment and predicting responsiveness of a patient to treatment of a cancer therapy (such as immunochemotherapy (ICI+CTX) or immunotherapy (ICI)). The biomarkers include T-cell inflammation signature (INCITE) whose upregulation unexpectedly correlates with ICI-induced tumor shrinkage. Additional biomarkers include monocyte marker signature and high tumor monocyte content (TMC) which can be used as an effective predictor of greater overall survival in cancer patients. Furthermore, tumor mutational burden is an additional independent and additive predictor of patients’ overall survival. The disclosed predictors can improve patient selection for cancer treatment with emerging cancer therapies.
[0045] Methods of Predicting Responsiveness to a Cancer Therapy
[0046] Accordingly, in one aspect, this disclosure provides a method of identifying a subject as likely responding to treatment of an immunotherapy or predicting survival of the subject undergoing the immunotherapy. In some embodiments, the method comprises: (a) determining an expression level of each of biomarkers comprising one or more of IFI30, FCN1, HLA-DRA, CXCL8, and SLC11A1 in a sample from the subject; (b) determining a change in the expression level of each of the biomarkers as compared to a respective reference expression level for each of the biomarkers; and (c) identifying the subject as likely responding to the treatment of the immunotherapy or as likely having prolonged survival after the treatment, if the expression level of each of the biomarkers is elevated as compared to the respective reference expression level for each of the biomarkers.
[0047] In some embodiments, the biomarkers comprise IFI30, FCN1, HLA-DRA, CXCL8, and SLC11A1. In some embodiments, the biomarkers consist of IFI30, FCN1, HLA-DRA, CXCL8, and SLC11A1. In some embodiments, the biomarkers comprise one or more of FI30, FCN1, HLA-DRA, CXCL8, SLC11A1, TYROBP, LYZ, PLAUR, HLA-DRB1, and EREG. In some embodiments, the biomarkers comprise FI30, FCN1, HLA-DRA, CXCL8, SLC11 Al, TYROBP, LYZ, PLAUR, HLA-DRB1, and EREG. In some embodiments, the biomarkers consist of FI30, FCN1, HLA- DRA, CXCL8, SLC11A1, TYROBP, LYZ, PLAUR, HLA-DRB1, and EREG.
[0048] In some embodiments, the sample is obtained before the treatment. In some embodiments, the sample comprises tumor tissue containing the monocytes.
[0049] In one aspect, this disclosure provides a method of identifying a subject as likely responding to treatment of an immunotherapy or predicting survival of the subject undergoing the immunotherapy. In some embodiments, the method comprises (i) determining an expression level of each of biomarkers comprising one or more of KLRD1, KLRK1, FASLG, GZMK, KLRC1, KLRC4, CD8A, PRF1, APOBEC3G, APOL3, CD244, and FCRL6 in a sample from the subject; (ii) determining a change in the expression level of each of the biomarkers as compared to a respective reference expression level for each of the biomarkers; and (iii) identifying the subject as likely responding to the treatment of the immunotherapy or as likely having prolonged survival after the treatment, if the expression level of each of the biomarkers is elevated as compared to the respective reference expression level for each of the biomarkers.
[0050] In some embodiments, the subject has been administered a cancer therapy (e.g., immunotherapy, chemotherapy, or a combination thereof). In some embodiments, the cancer therapy comprises an immunotherapy. In some embodiments, the immunotherapy comprises one or more immune checkpoint inhibitors. In some embodiments, the one or more immune checkpoint inhibitors comprise a CTLA-4 inhibitor, aPD-1 inhibitor, a PD-L1 inhibitor a LAG-3 inhibitor, or a TIM-3 inhibitor. In some embodiments, the one or more immune checkpoint inhibitors comprise an anti-CTLA-4 antibody, an anti-PD-1 antibody, an anti-PD-Ll antibody, an anti-LAG-3 antibody, an anti-HM-3 antibody, or a combination thereof.
[0051] In some embodiments, the one or more immune checkpoint inhibitors comprise an anti- PD-Ll antibody, an anti-CTLA-4 antibody, or a combination thereof. In some embodiments, the one or more immune checkpoint inhibitors comprise durvalumab.
[0052] In some embodiments, the expression level of each of the biomarkers is determined from the sample of the subject who has completed at least one treatment cycle of a cancer therapy. In some embodiments, the respective reference expression level is determined from samples of one or more subjects who have not been administered a cancer therapy.
[0053] In some embodiments, the treatment cycle is from about 1 week to 16 weeks (e.g., 1 week, 2 weeks, 3 weeks, 4 weeks, 5 weeks, 6 weeks, 7 weeks, 8 weeks, 9 weeks, 10 weeks, 11 weeks, 12 weeks, 13 weeks, 14 weeks, 15 weeks, 16 weeks). In some embodiments, the treatment cycle is about 4 weeks.
[0054] The term “biomarker,” as used herein, refers to an indicator, e.g., predictive, diagnostic, and / or prognostic, which can be detected in a sample. The biomarker may serve as an indicator of a particular subtype of a disease or disorder (e.g., cancer) characterized by certain molecular, pathological, histological, and / or clinical features, and / or may serve as an indicator of a particular cell type or state (e.g., epithelial, mesenchymal, etc.) and / or response to therapy. Biomarkers include, but are not limited to, polynucleotides (e.g., DNA and / or RNA), polynucleotide copy number alterations (e.g., DNA copy numbers), polypeptides, polypeptide and polynucleotide modifications (e.g., posttranslational modifications), carbohydrates, and / or glycolipid-based molecular markers. A biomarker may be present in a sample obtained from a subject before the onset of a physiological or pathophysiological state (e.g., primary cancer, metastatic cancer, etc.), including a symptom thereof (e.g., response to therapy). Thus, the presence of the biomarker in a sample obtained from the subject can be indicative of an increased risk that the subject will develop the physiological or pathophysiological state or symptom thereof. Alternatively and / or additionally, the biomarker may be normally expressed in an individual, but its expression may change (i.e., it is increased (upregulated; over-expressed) or decreased (downregulated; underexpressed)) before the onset of a physiological or pathophysiological state, including a symptom thereof. Thus, a change in the level of the biomarker may be indicative of an increased risk that the subject will develop the physiological or pathophysiological state or symptom thereof. Alternatively, or in addition, a change in the level of a biomarker may reflect a change in a particular physiological or pathophysiological state, or symptom thereof, in a subject, thereby allowing the nature (e.g., severity) of the physiological or pathophysiological state, or symptom thereof, to be tracked over a period of time.
[0055] In some embodiments, a level of a biomarker may include the concentration of the biomarker, the expression level of the biomarker, or the activity of the biomarker. The terms “expression level” and “level of expression” are used interchangeably and generally refer to the amount of a biomarker in a sample. “Expression” refers to the process by which information (e.g., gene-encoded and / or epigenetic) is converted into the structures present and operating in the cell. Therefore, as used herein, “expression” may refer to transcription into a polynucleotide, translation into a polypeptide, or even polynucleotide and / or polypeptide modifications (e.g., posttranslational modification of a polypeptide). Fragments of a transcribed polynucleotide, a translated polypeptide, or polynucleotide and / or polypeptide modifications (e.g., post-translational modification of a polypeptide) shall also be regarded as expressed whether they originate from a transcript generated by alternative splicing or a degraded transcript, or from a post-translational processing of the polypeptide, e.g, by proteolysis. The term “expression” with respect to a gene sequence refers to transcription of the gene to produce a RNA transcript (e.g, mRNA, antisense RNA, siRNA, shRNA, miRNA, etc.) and, in some embodiments, translation of a resulting mRNA transcript to a protein. Thus, expression of a coding sequence may result from transcription and translation of the coding sequence. Conversely, expression of a non-coding sequence results from the transcription of the non-coding sequence.
[0056] The terms “biomarker signature,” “signature,” “biomarker expression signature,” or “expression signature” are used interchangeably herein and refer to one or a combination of biomarkers whose expression is an indicator, e.g, predictive, diagnostic, and / or prognostic. The biomarker signature may serve as an indicator of a particular subtype of a disease or disorder (e.g, primary cancer, metastatic cancer, etc.) or symptom thereof (e.g, response to therapy, drug resistance, and / or disease burden) characterized by certain molecular, pathological, histological, and / or clinical features. In some embodiments, the biomarker signature is a “gene signature.” The term “gene signature” is used interchangeably with “gene expression signature” and refers to one or a combination of polynucleotides whose expression is an indicator, e.g, predictive, diagnostic, and / or prognostic. In some embodiments, the biomarker signature is a “protein signature.” The term “protein signature” is used interchangeably with “protein expression signature” and refers to one or a combination of polypeptides whose expression is an indicator, e.g, predictive, diagnostic, and / or prognostic.
[0057] In some embodiments, a biomarker signature may include 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, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, or 100 or more biomarkers. In some embodiments, the change in the expression level of each of the set of biomarkers comprises an increase or decrease in the expression level. In some embodiments, the change can include a distribution of changes or a pattern of changes of expression levels of biomarkers.
[0058] The term “distribution of changes” or “pattern of changes,” as used herein with reference to biomarkers, refers to an aggregated analysis of changes in biomarker levels (e.g., expression level, concentration, amount, activity) of each biomarker in a set of selected biomarkers relative to a corresponding reference biomarker level. The set of biomarkers may include 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, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, or 100 or more biomarkers. As will be understood by a person skilled in the art, a reference level or reference value for each biomarker may be the same or different. Such a distribution of changes or a pattern of changes in biomarker levels may include increased levels (e.g., expression levels) for a first group of biomarkers and / or decreased levels (e.g., expression levels) for a second group of biomarkers. The changes in biomarker levels for the first and second groups of biomarkers collectively constitute a distribution of changes or a pattern of changes. In some embodiments, a distribution of changes or a pattern of changes in biomarker levels may include ratios of levels between two or more biomarkers.
[0059] As used herein, the terms “increase,” “elevate,” “elevated,” “upregulate,” “enhance,” and “activate” all generally refer to an increase by a statically significant amount as compared to a reference level (e.g., a reference expression level). For the avoidance of any doubt, these terms mean an increase of at least 5% (e.g., 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, 100%) as compared to a reference level, for example, an increase of at least about 10%, at least about 20%, at least about 30%, at least about 40%, at least about 50%, at least about 60%, at least about 70%, at least about 80%, at least about 90% or at least about 100%, as compared to a reference level. In some embodiments, these terms may refer to an increase of 10-20%, 10-30%, 10-40%, 10-50%, 10-60%, 10-70%, 10-80%, 10- 90%, 10-100%, 10-110%, 10-120%, 10-130%, 10-140%, 10-150%, 10-160%, 10-170%, 10-180%, 10-190%, 10-200%, 10-210%, 10-220%, 10-230%, 10-240%, 10-250%, 10-260%, 10-270%, 10- 280%, 10-290%, or 10-300%, as compared to a reference level. In some embodiments, these terms may refer to an increase of 10-300%, 20-300%, 30-300%, 40-300%, 50-300%, 60-300%, 70-300%, 80-300%, 90-300%, 100-300%, 110-300%, 120-300%, 130-300%, 140-300%, 150-300%, 160- 300%, 170-300%, 180-300%, 190-300%, 200-300%, 210-300%, 220-300%, 230-300%, 240- 300%, 250-300%, 260-300%, 270-300%, 280-300%, or 290-300% as compared to a reference level. In some embodiments, these terms may refer to an increase of at least 2-fold, at least 3 -fold, at least 4-fold, at least 5-fold, at least 10-fold or greater, as compared to a reference level.
[0060] An increased level (e.g., expression level) of a biomarker as compared to a predetermined reference value can be an increase of at least 5% (e.g., 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, 100%) as compared to a predetermined reference value, for example, an increase of at least about 10%, at least about 20%, at least about 30%, at least about 40%, at least about 50%, at least about 60%, at least about 70%, at least about 80%, at least about 90%, at least about 100%, or any increase from 10% to 100%, as compared to a predetermined reference value; or at least a 2-fold, at least a 3 -fold, at least a 4-fold, at least a 5-fold or at least a 10-fold increase, or any increase from 2-fold to 10-fold or greater, as compared to a predetermined reference value.
[0061] In some embodiments, an increased level of a biomarker as compared to a predetermined reference value can be an increase of 10-20%, 10-30%, 10-40%, 10-50%, 10-60%, 10-70%, 10- 80%, 10-90%, 10-100%, 10-110%, 10-120%, 10-130%, 10-140%, 10-150%, 10-160%, 10-170%, 10-180%, 10-190%, 10-200%, 10-210%, 10-220%, 10-230%, 10-240%, 10-250%, 10-260%, 10- 270%, 10-280%, 10-290%, or 10-300% as compared to a predetermined reference value. In some embodiments, an increased level of a biomarker as compared to a predetermined reference value can be an increase of 10-300%, 20-300%, 30-300%, 40-300%, 50-300%, 60-300%, 70-300%, 80- 300%, 90-300%, 100-300%, 110-300%, 120-300%, 130-300%, 140-300%, 150-300%, 160-300%, 170-300%, 180-300%, 190-300%, 200-300%, 210-300%, 220-300%, 230-300%, 240-300%, 250- 300%, 260-300%, 270-300%, 280-300%, or 290-300% as compared to a predetermined reference value. In some embodiments, an increased level of a biomarker as compared to a predetermined reference value can be an increase of 10-20%, 20-30%, 30-40%, 40-50%, 50-60%, 60-70%, 70- 80%, 80-90%, 90-100%, 100-110%, 110-120%, 120-130%, 130-140%, 140-150%, 150-160%, 160-170%, 170-180%, 180-190%, 190-200%, 200-210%, 210-220%, 220-230%, 230-240%, 240- 250%, 250-260%, 260-270%, 270-280%, 280-290%, or 290-300% as compared to a predetermined reference value.
[0062] As used herein, the terms “decrease,” “reduce,” “downregulate,” and “inhibit” all generally refer to a decrease by a statistically significant amount. However, for avoidance of doubt, the term “reduced,” “decrease,” “reduce,” or “inhibit” means a decrease by at least 5% (e.g., 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, 100%) as compared to a reference level, for example, a decrease by at least about 10%, a decrease by at least about 20%, or at least about 30%, or at least about 40%, or at least about 50%, or at least about 60%, or at least about 70%, or at least about 80%, or at least about 90% or up to and including a 100% decrease (e.g., absent level as compared to a reference sample), or any decrease of 10- 100% as compared to a reference level. In some embodiments, these terms refer to a decrease of 10-20%, 10-30%, 10-40%, 10-50%, 10-60%, 10-70%, 10-80%, 10-90%, 10-100%, 10-110%, 10- 120%, 10-130%, 10-140%, 10-150%, 10-160%, 10-170%, 10-180%, 10-190%, 10-200%, 10- 210%, 10-220%, 10-230%, 10-240%, 10-250%, 10-260%, 10-270%, 10-280%, 10-290%, or 10- 300%, as compared to a reference level. In some embodiments, these terms refer to a decrease of 10-300%, 20-300%, 30-300%, 40-300%, 50-300%, 60-300%, 70-300%, 80-300%, 90-300%, 100- 300%, 110-300%, 120-300%, 130-300%, 140-300%, 150-300%, 160-300%, 170-300%, 180- 300%, 190-300%, 200-300%, 210-300%, 220-300%, 230-300%, 240-300%, 250-300%, 260- 300%, 270-300%, 280-300%, or 290-300%, as compared to a reference level. In some embodiments, these terms refer to a decrease of 10-20%, 20-30%, 30-40%, 40-50%, 50-60%, 60- 70%, 70-80%, 80-90%, 90-100%, 100-110%, 110-120%, 120-130%, 130-140%, 140-150%, 150- 160%, 160-170%, 170-180%, 180-190%, 190-200%, 200-210%, 210-220%, 220-230%, 230- 240%, 240-250%, 250-260%, 260-270%, 270-280%, 280-290%, or 290-300%, as compared to a reference level.
[0063] As used herein, the term “higher” with reference to a biomarker measurement refers to a statistically significant and measurable difference in the level of a biomarker measurement compared to the level of another biomarker or to a control level where the biomarker measurement is greater than the level of the other biomarker or the control level. The difference may be of at least about 10%, at least about 20%, at least about 30%, at least about 40%, or at least about 50%.
[0064] As used herein, the term “lower” with reference to a biomarker measurement refers to a statistically significant and measurable difference in the level of a biomarker measurement compared to the level of another biomarker or to a control level where the biomarker measurement is less than the level of the other biomarker or the control level. The difference may be of at least about 10%, at least about 20%, at least about 30%, at least about 40%, or at least about 50%. The terms “reference level,” “reference value, “control level,” “control value,” “predetermined value,” and “predetermined level” are used interchangeably herein. The terms “reference sample,” “reference cell,” “reference tissue,” “control sample,” “control cell,” and “control tissue” are used interchangeably herein.
[0065] In some embodiments, a reference level or a control level of biomarkers may be determined from a reference sample, reference cell, reference tissue, control sample, control cell, or control tissue that is obtained from a healthy and / or non-diseased part of the body (e.g., tissue or cells) of the same subject or individual, but at different time-points, e.g., before and after therapy. In some embodiments, a reference level or a control level of biomarkers may be determined from a reference sample, reference cell, reference tissue, control sample, control cell, or control tissue that is obtained from a healthy individual who is not the subject or individual being assessed. In some examples, a reference sample, reference cell, reference tissue, control sample, control cell, or control tissue is or may include a functional T-cell, a dysfunctional T-cell (e.g., an exhausted T- cell), T-cells from a subject that is responsive or sensitive to therapy or T-cells from a subject that is non-responsive or resistant to therapy. In some embodiments, T-cells may include CD8+ T-cells. In some embodiments, T-cells may include CD 8+ T-cells from a subject that is non-responsive or resistant to a cancer therapy.
[0066] In some embodiments, the expression level may include a mRNA or protein level.
[0067] As used herein, the “amount” or “level” of a biomarker is a detectable level or amount in a sample. These can be measured by methods known to one skilled in the art. These terms encompass a quantitative amount or level (e.g., weight or moles), a semi-quantitative amount or level, a relative amount or level (e.g., weight % or mole % within class), a concentration, and the like. Thus, these terms encompass absolute or relative amounts or levels or concentrations of a biomarker in a sample. The expression level or amount of biomarker assessed can be used to determine the response to treatment.
[0068] In some embodiments, the mRNA level is determined by at least one technique selected from reverse transcription polymerase chain reaction (RT-PCR), competitive RT-PCR, real-time RT-PCR, ribonucleic acid sequencing (RNA-seq), immunohistochemistry (IHC), immunofluorescence, RNase protection assay (RPA), northern blotting, and DNA chip. In some embodiments, the protein level is determined by at least one technique selected from western blot, enzyme-linked immunosorbent assay (ELISA), radioimmunoassay (RIA), radioimmunodiffusion, Ouchterlony immunodiffusion, rocket immunoelectrophoresis, immunohistochemical staining, immunoprecipitation assay, complement fixation assay, fluorescence-activated cell sorter (FACS), and protein chip.
[0069] The terms “sample” or “biological sample,” as used herein, include any biological specimen obtained (isolated, removed) from a subject. Samples may include, without limitation, organ tissue (e.g., primary or metastatic tumor tissue), whole blood, plasma, serum, whole blood cells, red blood cells, white blood cells (e.g., peripheral blood mononuclear cells), saliva, urine, stool (feces), tears, sweat, sebum, nipple aspirate, ductal lavage, tumor exudates, synovial fluid, cerebrospinal fluid, lymph, fine needle aspirate, amniotic fluid, any other bodily fluid, exudate or secretory fluid, cell lysates, cellular secretion products, inflammation fluid, semen, and vaginal secretions. In some embodiments, a sample may be readily obtainable by non-invasive or minimally invasive methods, such as blood collection (“liquid biopsy”), urine collection, feces collection, tissue (e.g., tumor tissue) biopsy, or fine-needle aspiration, allowing the provision / removal / isolation of the sample from a subject. The term “tissue,” as used herein, encompasses all types of cells of the body, including cells of organs but also including blood and other body fluids recited above. The tissue may be healthy or affected by pathological alterations, e.g, tumor tissue. The tissue may be from a living subject or may be cadaveric tissue. In some embodiments, useful samples are those known to comprise, expected, or predicted to comprise, known to potentially comprise, or expected or predicted to potentially comprise tumor cells.
[0070] The biological sample may be any sample in which the methylation level of the relevant gene(s) can be determined. In some embodiments, the biological sample is a neoplastic tissue sample, such as a tumor sample, e.g. , a primary or metastatic tumor sample. The biological sample may also be derived from a biological fluid or body fluid, for example, whole blood, blood, urine, lymph fluid, serum, plasma, nipple aspirate, ductal fluid, and tumor exudate. It has been shown in the literature that cancer or tumor cells often release genomic DNA in circulating or other bodily fluids. Since said genomic DNA has the same methylation profile as the DNA inside the tumor or cancer cell, said methylation profile can be detected in the circulating or other bodily fluid sample. This has, for example, been reviewed by Qureshi et al., 2010 (Int. J. Surgery 2010, 8: 194-198), hereby incorporated by reference in its entirety. In some embodiments, the sample is a body fluid comprising neoplastic cells.
[0071] A sample can be obtained from a subject in any way typically used in clinical settings for obtaining a sample comprising the required cells or nucleic acid, including RNA, genomic DNA, mitochondrial DNA, and protein-associated nucleic acids. For example, the sample can be obtained from fresh, frozen, or paraffin-embedded surgical samples or biopsies of an organ or tissue comprising the suitable cells or nucleic acid to be tested. If desired, the sample can be mixed with a fluid purified, amplified, or otherwise treated. For example, samples may be treated in one or more purification steps to increase the purity of the desired cells or nucleic acid in the sample, or they may be examined without any purification steps. Any nucleic acid specimen in purified or non-purified form obtained from such sample can be utilized in the methods as taught herein.
[0072] In some embodiments, the sample is obtained from neoplasia tissue, tumor microenvironment, or tumor-infiltrating immune cells. In some embodiments, the sample may include a biological sample that may include a plasma sample, a blood sample, or a tissue sample. In some embodiments, the sample is obtained from a primary tumor or a metastasis. In some embodiments, the sample may include immune cells. In some embodiments, the immune cells are selected from T cells, macrophages, dendritic cells, fibroblasts, NK cells, NKT cells, and NK-DC cells. In some embodiments, the sample may include protein, DNA, or RNA.
[0073] The terms “predicting,” “prediction,” or “predictive,” as used herein, refer to an advance declaration, indication, or foretelling of a response or reaction to a therapy (e.g., immunotherapy, immunochemotherapy) in a subject not (yet) having been treated with the therapy. For example, a prediction of responsiveness (or sensitivity or susceptibility) to a cancer therapy in a subject may indicate that the subject will respond or react to the cancer therapy, for example, within a certain time period, e.g., so that the subject will have a clinical benefit from the cancer therapy. A prediction of unresponsiveness (or insensitivity or insusceptibility) to a cancer therapy in a subject may indicate that the subject will minimally or not respond or react to the cancer therapy, for example, within a certain time period, e.g., so that the subject will have no clinical benefit from the cancer therapy.
[0074] The terms “responsiveness,” “sensitivity,” or “susceptibility” may be used interchangeably herein and refer to the quality that predisposes a subject having a neoplastic disease to be responsive or reactive to a cancer therapy (e.g. , immunotherapy, immunochemotherapy). A subject is “responsive,” “sensitive,” or “susceptible” (which terms are used interchangeably) to immunotherapy (i.e., treatment with an immunotherapy, immunochemotherapy); in particular, a subject “responds positively,” if the subject will have a clinical benefit from the treatment. A neoplastic tissue, including a tumor, is “responsive,” “sensitive,” or “susceptible” to a cancer therapy if the proliferation rate of the neoplastic tissue is inhibited as a result of contact with the cancer therapy, compared to the proliferation rate of the neoplastic tissue in the absence of contact with the cancer therapy, e.g., treatment with an immunotherapy, immunochemotherapy.
[0075] The terms “unresponsiveness,” “insensitivity,” “insusceptibility,” or “resistance” may be used interchangeably herein and refer to the quality that predisposes a subject having a neoplastic disease (e.g., cancer) to a minimal (e.g., insignificant) or no response to a cancer therapy (e.g., immunotherapy, immunochemotherapy). A subject is “unresponsive,” “insensitive,” “unsusceptible,” or “resistant” (which terms are used interchangeably) to a cancer therapy (i.e., treatment with a cancer therapy), in particular, a subject “responds negatively,” if the subject will have no clinical benefit from the treatment. A neoplastic tissue, including a tumor, is “unresponsive,” “insensitive,” “unsusceptible,” or “resistant” to a cancer therapy if the proliferation rate of the neoplastic tissue is not inhibited or inhibited to a very low (e.g., therapeutically insignificant) degree, as a result of treatment of the cancer therapy, compared to the proliferation rate of the neoplastic tissue in the absence of treatment the cancer therapy. The methods as disclosed herein may allow making a prediction that a subject having a neoplastic disease will be responsive to a cancer therapy or will be unresponsive to the cancer therapy. This may, in some embodiments, include predicting that a subject having a neoplastic disease will have a comparatively low probability (e.g., less than 50%, less than 40%, less than 30%, less than 20% or less than 10%>) of being responsive to a cancer therapy; or that a subject having a neoplastic disease will have a comparatively high probability (e.g., at least 50%, at least 60%, at least 70%, at least 80%) or at least 90%) of being responsive to the cancer therapy.
[0076] The terms “determining responsiveness,” “predicting responsiveness,” and “assessing a likelihood of a therapeutic response” may be used interchangeably herein.
[0077] The term “prognosis,” as used herein, refers to anticipation of progression of a disease (e.g., cancer) or condition and prospect (e.g., the probability, duration, and / or extent) of recovery. A good prognosis of the diseases or conditions may generally encompass anticipation of a satisfactory partial or complete recovery from the diseases or conditions, such as within an acceptable time period. A good prognosis of such may more commonly encompass anticipation of not further worsening or aggravating within a given time period. A poor prognosis of the diseases or conditions as taught herein may generally encompass anticipation of a substandard recovery and / or unsatisfactorily slow recovery, or to substantially no recovery or even further worsening of such.
[0078] In some embodiments, a therapeutic response may include an anti-tumor response when referring to a cancer patient treated with a cancer therapy. For example, an anti-tumor response may include at least one positive therapeutic effect, such as a reduced number of cancer cells, reduced tumor size, reduced rate of cancer cell infiltration into peripheral organs, reduced rate of tumor metastasis or tumor growth, or progression-free survival. Positive therapeutic effects in cancer can be measured in a number of ways (see, e.g., W. A. Weber, J. Null. Med. 5O: 1S-1OS (2009); Eisenhauer etal., 2009 European Journal of Cancer, 45: 228-247). In some embodiments, an anti-tumor response to a cancer therapy is assessed using RECIST 1.1 criteria, bidimensional irRC, or unidimensional irRC. In some embodiments, an antitumor response is any of stable disease (SD), partial response (PR), complete response (CR), progression-free survival (PFS), and disease-free survival (DFS). In some embodiments, one or more biomarkers of this disclosure predict whether a subject with a solid tumor is likely to achieve a complete response or a partial response.
[0079] The disclosed method may be used to predict or determine the likelihood of a complete response or partial response, or whether a response is likely to be a complete response or a partial response. As used herein, a “complete response” or “CR” to a therapy refers to disappearance of all detectable signs of cancer in response to treatment. As used herein, a “partial response” to a therapy refers to a decrease in tumor load in an individual, for example, in terms of tumor number, size, and growth rate, and or an increase in the time of disease progression.
[0080] In some embodiments, the likelihood of the therapeutic response in the subject may include complete or partial response as defined by response evaluation criteria in solid tumors (RECIST 1 .0 criteria (Therasse P. et al., 2000 J. Natl Cancer Inst 92:2015-16)), stable disease (SD) as defined by RECIST, or long-term survival despite disease progression or response as defined by immune-related response criteria (irRC).
[0081] In another aspect, this disclosure also provides a method of identifying a subject as likely responding to treatment of a cancer therapy. In some embodiments, the method comprises: (a) determining a level of a tumor monocyte content (TMC) in a sample from the subject; (b) determining a change in the level of the tumor monocyte content as compared to a reference level; and (c) identifying the subject as likely responding to the treatment of the cancer therapy if the level of the tumor monocyte content is elevated as compared to the reference level, wherein the cancer therapy comprises an immunotherapy and a chemotherapy.
[0082] In another aspect, this disclosure further provides a method of identifying a subject as likely responding to treatment of a cancer therapy. In some embodiments, the method comprises: (i) determining a level of a tumor monocyte content (TMC) and a level of tumor mutational burden (TMB) in a sample from the subject; (ii) determining a first change in the level of the tumor monocyte content as compared to a first reference level and a second change in the level of the tumor mutational burden as compared to a second reference level; and (iii) identifying the subject as likely responding to the treatment of the cancer therapy if the level of the tumor monocyte content is elevated as compared to the first reference level; and / or if the level of the tumor mutational burden is elevated as compared to the second reference level, wherein the cancer therapy comprises an immunotherapy and a chemotherapy.
[0083] In yet another aspect, this disclosure further provides a method of identifying a subject as likely not responding to treatment of a cancer therapy. In some embodiments, the method comprises: (a) determining an expression level of each of biomarkers comprising one or more of T cell-specific interferon-stimulated genes (ISGs) in a sample from the subject; (b) determining a change in the expression level of each of the biomarkers as compared to a respective reference expression level for each of the biomarkers; and (c) identifying the subject as likely not responding to the treatment of the cancer therapy if the expression level of each of the biomarkers is elevated as compared to the respective reference expression level for each of the biomarkers.
[0084] In some embodiments, interferon-stimulated genes comprise one or more of OAS1, MX1, IFI6, and XAFl. In some embodiments, the cancer therapy comprises an immunotherapy and, optionally a chemotherapy.
[0085] The terms “patient,” “subject,” “host,” and “individual” are used interchangeably herein and refer to any subject, particularly a vertebrate subject, and even more particularly a mammalian subject, for whom therapy or prophylaxis is desired. Suitable vertebrate animals that fall within the scope of the invention include, but are not restricted to, any member of the subphylum Chordata including primates (e.g., humans, monkeys and apes, and includes species of monkeys such from the genus Macaca (e.g., cynomologus monkeys such as Macaca fascicularis, and / or rhesus monkeys ( Macaca mulatta )) and baboon ( Papio ursinus), as well as marmosets (species from the genus Callithrix), squirrel monkeys (species from the genus Saimiri ) and tamarins (species from the genus Saguinus), as well as species of apes such as chimpanzees ( Pan troglodytes)), rodents (e.g., mice rats, guinea pigs), lagomorphs (e.g., rabbits, hares), bovines (e.g., cattle), ovines (e.g., sheep), caprines (e.g., goats), porcines (e.g., pigs), equines (e.g., horses), canines (e.g., dogs), felines (e.g., cats), avians (e.g., chickens, turkeys, ducks, geese, companion birds such as canaries, budgerigars etc.), marine mammals (e.g., dolphins, whales), reptiles (snakes, frogs, lizards etc.), and fish. In some embodiments, a subject is a human with cancer.
[0086] In some embodiments, the subject has a cancer. In some embodiments, the cancer is a carcinoma, a sarcoma, a lymphoma, a melanoma, a pediatric tumor, or a leukemia.
[0087] In some embodiments, the cancer is selected from adrenal gland tumors, biliary cancer, bladder cancer, brain cancer, breast cancer, carcinoma, central or peripheral nervous system tissue cancer, cervical cancer, colon cancer, endocrine or neuroendocrine cancer, or hematopoietic cancer, esophageal cancer, fibroma, gastrointestinal cancer, glioma, head and neck cancer, Li-Fraumeni tumors, liver cancer, lung cancer, lymphoma, melanoma, meningioma, multiple neuroendocrine type I and type II tumors, nasopharyngeal cancer, oral cancer, oropharyngeal cancer, osteogenic sarcoma tumors, ovarian cancer, pancreatic cancer, pancreatic islet cell cancer, parathyroid cancer, pheochromocytoma, pituitary tumors, prostate cancer, rectal cancer, renal cancer, respiratory cancer, sarcoma, skin cancer, stomach cancer, testicular cancer, thyroid cancer, tracheal cancer, urogenital cancer, and uterine cancer.
[0088] In some embodiments, the cancer comprises inoperable esophageal adenocarcinoma (EAC). In some embodiments, the cancer comprises a BV- / MSS gastric cancer. Methods of Treatment
[0089] In another aspect, this disclosure provides a method of treating a cancer in a subject in need thereof with a cancer therapy. In some embodiments, the method comprises identifying a subject as likely responding to treatment of an immunotherapy or as likely having prolonged survival after the treatment according to the method as described herein, and administering to the subject a therapeutically effective amount of a cancer therapy. In some embodiments, the cancer therapy comprises an immunotherapy or an immunochemotherapy.
[0090] In some embodiments, the cancer therapy, such as immunotherapy or immunochemotherapy, is administered to the subject in one or more doses or in one or more treatment cycles.
[0091] In some embodiments, the immunotherapy comprises one or more immune checkpoint inhibitors. In some embodiments, one or more immune checkpoint inhibitors comprise a CTLA-4 inhibitor, a PD-1 inhibitor, a PD-L1 inhibitor, a LAG-3 inhibitor, or a TIM-3 inhibitor. In some embodiments, one or more immune checkpoint inhibitors comprise an anti-CTLA-4 antibody, an anti-PD-1 antibody, an anti-PD-Ll antibody, an anti-LAG-3 antibody, an anti-TIM-3 antibody, or a combination thereof.
[0092] In some embodiments, one or more immune checkpoint inhibitors comprise an anti-PD-Ll antibody, an anti-CTLA-4 antibody, or a combination thereof. In some embodiments, one or more immune checkpoint inhibitors comprise durvalumab.
[0093] In some embodiments, the chemotherapy comprises oxaliplatin, capecitabine, or a combination thereof.
[0094] Also described herein, the cancer therapy can be used in combination with additional chemotherapy, radiation, immunosuppressive agents, such as cyclosporin, azathioprine, methotrexate, mycophenolate, and FK506, antibodies, or other immunoablating agents, such as CAMPATH, anti-cancer antibodies, CD3 or other antibody therapies, cytoxine, fludarabine, cyclosporine, FK506, rapamycin, mycophenolic acid, steroids, FR901228, cytokines, and irradiation.
[0095] In some embodiments, the disclosed methods lead to increased efficacy and duration of anti-tumor response. Methods according to this aspect of the disclosure may include selecting a subject with cancer and administering to the subject a therapeutically effective amount of a cancer therapy (e.g., immunotherapy, immunochemotherapy). In some embodiments, the methods provide for increased tumor inhibition, e.g., by about 20%, more than 20%, more than 30%, more than 40%, more than 50%, more than 60%, more than 70%, or more than 80% as compared to an untreated subject.
[0096] In some embodiments, the methods provide for increased duration of the anti-tumor response, e.g., by about 20%, more than 20%, more than 30%, more than 40%, more than 50%, more than 60%, more than 70% or more than 80% as compared to an untreated subject. In some embodiments, administration of a cancer therapy (e.g., immunotherapy, immunochemotherapy) increases response and duration of response in a subject, e.g, by more than 2%, more than 3%, more than 4%, more than 5%, more than 6%, more than 7%, more than 8%, more than 9%, more than 10%, more than 20%, more than 30%, more than 40% or more than 50% more than an untreated subject.
[0097] In some embodiments, the disclosed methods lead to a delay in tumor growth and development, e.g, tumor growth may be delayed by about 3 days, more than 3 days, about 7 days, more than 7 days, more than 15 days, more than 1 month, more than 3 months, more than 6 months, more than 1 year, more than 2 years, or more than 3 years as compared to an untreated subject.
[0098] In some embodiments, administration of any of the combinations disclosed herein prevents tumor recurrence and / or increases duration of survival of the subject, e.g, increases duration of survival by 1-5 days, by 5 days, by 10 days, by 15 days, more than 15 days, more than 1 month, more than 3 months, more than 6 months, more than 12 months, more than 18 months, more than 24 months, more than 36 months, or more than 48 months more than the survival of an untreated subject.
[0099] In some embodiments, administration of a cancer therapy to a subject with a cancer leads to complete disappearance of all evidence of tumor cells (“complete response”). In some embodiments, administration of a cancer therapy (e.g, immunotherapy, immunochemotherapy) to a subject with a cancer leads to at least 30% or more decrease in tumor cells or tumor size (“partial response”). In some embodiments, administration of a cancer therapy (e.g, immunotherapy, immunochemotherapy) to a subject with a cancer leads to complete or partial disappearance of tumor cells / lesions, including new measurable lesions. Tumor reduction can be measured by any methods known in the art, e.g., X-rays, positron emission tomography (PET), computed tomography (CT), magnetic resonance imaging (MRI), cytology, histology, or molecular genetic analyses.
[0100] In some embodiments, a therapeutic response in the subject comprises complete or partial response as defined by response evaluation criteria in solid tumors (RECIST), stable disease as defined by RECIST, or long-term survival despite disease progression or response as defined by immune-related response criteria (irRC).
[0101] In some embodiments, administration of a cancer therapy to a subject with a cancer leads to an improved overall response rate, as compared to an untreated subject.
[0102] In some embodiments, administration of a cancer therapy to a subject with a cancer leads to increased overall survival (OS) or progression-free survival (PFS) of the subject as compared to a untreated subject. In some embodiments, the PFS is increased by at least one month, at least 2 months, at least 3 months, at least 4 months, at least 5 months, at least 6 months, at least 7 months, at least 8 months, at least 9 months, at least 10 months, at least 11 months, at least 1 year, at least
[0103] 2 years, or at least 3 years as compared to a untreated subject.
[0104] In some embodiments, the OS is increased by at least one month, at least 2 months, at least
[0105] 3 months, at least 4 months, at least 5 months, at least 6 months, at least 7 months, at least 8 months, at least 9 months, at least 10 months, at least 11 months, at least 1 year, at least 2 years, or at least 3 years as compared to a untreated subject.
[0106] In some embodiments, the cancer is a carcinoma, a sarcoma, a lymphoma, a melanoma, a pediatric tumor, or a leukemia.
[0107] As used herein, the term “treatment” refers to a clinical intervention designed to alter the natural course of the individual or cell being treated during clinical pathology. Desirable effects of treatment include decreasing the rate of disease progression, ameliorating, or palliating the disease state, and remission or improved prognosis. For example, an individual is successfully “treated” if one or more symptoms associated with a cancer are mitigated or eliminated, including, but are not limited to, reducing the proliferation of (or destroying) cancerous cells, reducing pathogen infection, decreasing symptoms resulting from the disease, increasing the quality of life of those suffering from the disease, decreasing the dose of other medications required to treat the disease, and / or prolonging survival of individuals. The phrase “treatment with a therapy,” “treating with a therapy,” “treatment with an agent,” “treating with an agent,” and the like refers to the administration of an effective amount of a therapy or agent, including a cancer therapy and optionally an agent, (e.g., a cytotoxic agent or an immunotherapeutic agent) to a patient, or the concurrent administration of two or more therapies or agents, including cancer therapies or agents, (e.g., two or more agents selected from cytotoxic agents and immunotherapeutic agents) in effective amounts to a patient.
[0108] As used herein, “cancer,” “tumor,” and “malignancy” all relate equivalently to hyperplasia of a tissue or organ. If the tissue is a part of the lymphatic or immune system, malignant cells may include non-solid tumors of circulating cells. Malignancies of other tissues or organs may produce solid tumors. The methods described herein can be used in the treatment of lymphatic cells, circulating immune cells, and solid tumors.
[0109] As used herein, the term “cancer” refers to a malignant neoplasm characterized by deregulated or unregulated cell growth. The term “cancer” includes primary malignant cells or tumors (e.g., those whose cells have not migrated to sites in the subject’s body other than the site of the original malignancy or tumor) and secondary malignant cells or tumors (e.g., those arising from metastasis, the migration of malignant cells or tumor cells to secondary sites that are different from the site of the original tumor). The term “metastatic” or “metastasis” generally refers to the spread of a cancer from one organ or tissue to another non-adjacent organ or tissue. The occurrence of the neoplastic disease in the other non-adjacent organ or tissue is referred to as metastasis.
[0110] Examples of cancer include but are not limited to carcinoma, lymphoma, blastoma, sarcoma, and leukemia or lymphoid malignancies. More particular examples of such cancers include without limitation: squamous cell cancer (e.g., epithelial squamous cell cancer), lung cancer including small-cell lung cancer, non-small cell lung cancer, adenocarcinoma of the lung, squamous carcinoma of the lung and large cell carcinoma of the lung, cancer of the peritoneum, hepatocellular cancer, gastric or stomach cancer including gastrointestinal cancer, pancreatic cancer, glioma, glioblastoma, cervical cancer, ovarian cancer, liver cancer, bladder cancer, hepatoma, breast cancer, colon cancer, rectal cancer, colorectal cancer, endometrial cancer or uterine carcinoma, salivary gland carcinoma, kidney or renal cancer, prostate cancer, vulvar cancer, thyroid cancer, hepatic carcinoma, anal carcinoma, penile carcinoma, as well as CNS cancer, melanoma, head and neck cancer, bone cancer, bone marrow cancer, duodenum cancer, esophageal cancer, thyroid cancer, or hematological cancer.
[0111] In some embodiments, the subject has an esophageal cancer. In some embodiments, the subject has inoperable esophageal adenocarcinoma (EAC). In some embodiments, the subject has an EBV- / MSS gastric cancer.
[0112] Additional Definitions
[0113] To aid in understanding the detailed description of the compositions and methods according to the disclosure, a few express definitions are provided to facilitate an unambiguous disclosure of the various aspects of the disclosure. Unless otherwise defined, 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.
[0114] Unless defined otherwise, all technical and scientific terms used herein have the meaning commonly understood by a person skilled in the art to which this invention belongs. The following references provide one of skill with a general definition of many of the terms used in this invention: Singleton et al., Dictionary of Microbiology and Molecular Biology (2nd ed. 1994); The Cambridge Dictionary of Science and Technology (Walker ed., 1988); The Glossary of Genetics, 5th Ed., R. Rieger et al. (eds.), Springer Verlag (1991); and Hale & Marham, The Harper Collins Dictionary of Biology (1991). As used herein, the following terms have the meanings ascribed to them below, unless specified otherwise.
[0115] As used herein, the phrases “nucleic acid,” “polynucleotide,” “oligonucleotide,” and “nucleic acid molecule” are used interchangeably to refer to a polymer of DNA and / or RNA, which can be single-stranded, double-stranded, or multi-stranded, synthesized or obtained (e.g., isolated and / or purified) from natural sources, which can contain natural, non-natural, and / or altered nucleotides, and which can contain natural, non-natural, and / or altered internucleotide linkages including, but not limited to phosphoroamidate linkages and / or phosphorothioate linkages instead of the phosphodiester found between the nucleotides of an unmodified oligonucleotide.
[0116] The term “gene” is well-known in the art and refers to a locatable region of a genomic sequence corresponding to a unit of inheritance, which is associated with regulatory regions, transcribed regions and / or other functional sequence regions. Genes typically comprise a coding sequence encoding a gene product, such as an RNA molecule or a polypeptide.
[0117] As used herein, the term “polypeptide” refers to any polymer preferably consisting essentially of any of the 20 natural amino acids regardless of its size. Although the term “protein” is often used in reference to relatively large proteins, and “peptide” is often used in reference to small polypeptides, use of these terms in the field often overlaps. The term “polypeptide” refers generally to proteins, polypeptides, and peptides unless otherwise noted. Peptides useful in accordance with the present disclosure are generally between about 0.1 to 100 KD or greater up to about 1000 KD, preferably between about 0.1, 0.2, 0.5, 1, 2, 5, 10, 20, 30, and 50 KD as judged by standard molecule sizing techniques such as centrifugation or SDS-polyacrylamide gel electrophoresis.
[0118] As used herein, the term “immunotherapy” refers to any treatment that modulates a subject’s immune system. In particular, the term comprises any treatment that modulates an immune response, such as a humoral immune response, a cell-mediated immune response, or both. An immune response may typically involve a response by a cell of the immune system, such as a B cell, cytotoxic T cell (CTL), T helper (Th) cell, regulatory T (Treg) cell, antigen-presenting cell (APC), dendritic cell, monocyte, macrophage, natural killer T (NKT) cell, natural killer (NK) cell, basophil, eosinophil, or neutrophil, to a stimulus. In the context of anti-cancer treatments, immunotherapy may elicit, induce or enhance an immune response, such as in particular an immune response specifically against tumor tissues or cells, such as to achieve tumor cell death. Immunotherapy may modulate, such as increase or enhance, the abundance, function, and / or activity of any component of the immune system, such as any immune cell, such as without limitation T cells (e.g., CTLs or Th cells), dendritic cells, and / or NK cells. Immunotherapies can be categorized as active, passive or a combination thereof. Anti-cancer immunotherapy is based on the fact that cancer cells typically have molecules on their surface, known as tumor antigens, that can be detected by the immune system. Active immunotherapy directs the immune system to attack tumor cells by targeting tumor antigens. Passive immunotherapies enhance existing antitumor responses, typically through binding and modifying the intracellular signaling of surface receptors. Immunotherapy comprises cell-based immunotherapy in which immune cells, such as T cells and / or dendritic cells, are transferred into the patient. The term also comprises an administration of substances or compositions, such as chemical compounds and / or biomolecules (e.g., antibodies, antigens, interleukins, cytokines, or combinations thereof), that modulate a subject’s immune system. Examples of cancer immunotherapy include, without limitation, treatments employing monoclonal antibodies, for example, immune checkpoint inhibitors, Fc- engineered monoclonal antibodies against proteins expressed by tumor cells, prophylactic or therapeutic cancer vaccines, immunotherapy, immunochemotherapy, and combinations thereof.
[0119] The terms “prevent,” “preventing,” “prevention,” “prophylactic treatment,” and the like refer to reducing the probability of developing a disorder or condition in a subject, who does not have, but is at risk of or susceptible to developing a disorder or condition. The term includes prevention of spread of infection in a subject exposed to the virus or at risk of having cancer.
[0120] The term “immune response,” as used herein, refers to any type of immune response, including, but not limited to, innate immune responses (e.g., activation of Toll receptor signaling cascade), cell-mediated immune responses (e.g., responses mediated by T cells (e.g., antigenspecific T cells) and non-specific cells of the immune system) and humoral immune responses (e.g., responses mediated by B cells (e.g., via generation and secretion of antibodies into the plasma, lymph, and / or tissue fluids). The term “immune response” is meant to encompass all aspects of the capability of a subject’s immune system to respond to antigens and / or immunogens (e.g., both the initial response to an immunogen (e.g., a pathogen) as well as acquired (e.g., memory) responses that are a result of an adaptive immune response).
[0121] The term “disease” as used herein is intended to be generally synonymous and is used interchangeably with the terms “disorder” and “condition” (as in medical condition), in that all reflect an abnormal condition of the human or animal body or of one of its parts that impairs normal functioning, is typically manifested by distinguishing signs and symptoms, and causes the human or animal to have a reduced duration or quality of life.
[0122] The term “effective amount,” “effective dose,” or “effective dosage” is defined as an amount sufficient to achieve or at least partially achieve a desired effect. A “therapeutically effective amount” or “therapeutically effective dosage” of a drug or therapeutic agent is any amount of the drug that, when used alone or in combination with another therapeutic agent, promotes disease regression evidenced by a decrease in severity of disease symptoms, an increase in frequency and duration of disease symptom-free periods, or a prevention of impairment or disability due to the disease affliction. A “prophylactically effective amount” or a “prophylactically effective dosage” of a drug is an amount of the drug that, when administered alone or in combination with another therapeutic agent to a subject at risk of developing a disease or of suffering a recurrence of disease, inhibits the development or recurrence of the disease. The ability of a therapeutic or prophylactic agent to promote disease regression or inhibit the development or recurrence of the disease can be evaluated using a variety of methods known to the skilled practitioner, such as in human subjects during clinical trials, in animal model systems predictive of efficacy in humans, or by assaying the activity of the agent in in vitro assays.
[0123] The term “agent” is used herein to denote a chemical compound, a mixture of chemical compounds, a biological macromolecule (such as a nucleic acid, an antibody, a protein or portion thereof, e.g., a peptide), or an extract made from biological materials such as bacteria, plants, fungi, or animal (particularly mammalian) cells or tissues. The activity of such agents may render it suitable as a “therapeutic agent,” which is a biologically, physiologically, or pharmacologically active substance (or substances) that acts locally or systemically in a subject.
[0124] The terms “therapeutic agent,” “therapeutic capable agent,” or “treatment agent” are used interchangeably and refer to a molecule or compound that confers some beneficial effect upon administration to a subject. The beneficial effect includes enablement of diagnostic determinations; amelioration of a disease, symptom, disorder or pathological condition; reducing or preventing the onset of a disease, symptom, disorder or condition; and generally counteracting a disease, symptom, disorder or pathological condition.
[0125] “Combination” therapy, as used herein, unless otherwise clear from the context, is meant to encompass administration of two or more therapeutic agents in a coordinated fashion and includes, but is not limited to, concurrent dosing. Specifically, combination therapy encompasses both co-administration (e.g., administration of a co-formulation or simultaneous administration of separate therapeutic compositions) and serial or sequential administration, provided that administration of one therapeutic agent is conditioned in some way on administration of another therapeutic agent. For example, one therapeutic agent may be administered only after a different therapeutic agent has been administered and allowed to act for a prescribed period of time. See, e.g., Kohrt etal. (2011) Blood 117:2423. Doses are often expressed in relation to body weight. Thus, a dose which is expressed as [g, mg, or other unit] / kg (or g, mg etc.) usually refers to [g, mg, or other unit] “per kg (or g, mg etc.) body weight,” even if the term “body weight” is not explicitly mentioned.
[0126] As used herein, “in vitro” refers to events that occur in an artificial environment, e.g., in a test tube or reaction vessel, in cell culture, etc., rather than within a multi-cellular organism.
[0127] As used herein, “in vivo” refers to events that occur within a multi-cellular organism, such as a non-human animal.
[0128] It is noted here that, as used in this specification and the appended claims, the singular forms “a,” “an,” and “the” include plural reference unless the context clearly dictates otherwise.
[0129] As used herein, “including,” “comprising,” “containing,” or “having” and variations thereof are meant to encompass the items listed thereafter and equivalents thereof as well as additional subject matter unless otherwise noted.
[0130] As used herein, the phrases “in one embodiment,” “in various embodiments,” “in some embodiments,” and the like do not necessarily refer to the same embodiment, but may unless the context dictates otherwise.
[0131] As used herein, the terms “and / or” or “ / ” means any one of the items, any combination of the items, or all of the items with which this term is associated.
[0132] As used herein, the term “substantially” does not exclude “completely,” e.g., a composition which is “substantially free” from Y may be completely free from Y. Where necessary, the word “substantially” may be omitted from the definition of the present disclosure.
[0133] As used herein, the term “approximately” or “about,” as applied to one or more values of interest, refers to a value that is similar to a stated reference value. In some embodiments, the term “approximately” or “about” refers to a range of values that fall within 25%, 20%, 19%, 18%, 17%, 16%, 15%, 14%, 13%, 12%, 11%, 10%, 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2%, 1%, or less in either direction (greater than or less than) of the stated reference value unless otherwise stated or otherwise evident from the context (except where such number would exceed 100% of a possible value). Unless indicated otherwise herein, the term “about” is intended to include values, e.g., weight percents, proximate to the recited range that are equivalent in terms of the functionality of the individual ingredient, the composition, or the embodiment. As used herein, the term “each,” when used in reference to a collection of items, is intended to identify an individual item in the collection but does not necessarily refer to every item in the collection. Exceptions can occur if explicit disclosure or context clearly dictates otherwise.
[0134] As disclosed herein, a number of ranges of values are provided. It is understood that each intervening value, to the tenth of the unit of the lower limit, unless the context clearly dictates otherwise, between the upper and lower limits of that range is also specifically disclosed. Each smaller range between any stated value or intervening value in a stated range and any other stated or intervening value in that stated range is encompassed within the present disclosure. The upper and lower limits of these smaller ranges may independently be included or excluded in the range, and each range where either, neither, or both limits are included in the smaller ranges is also encompassed within the present disclosure, subject to any specifically excluded limit in the stated range. Where the stated range includes one or both of the limits, ranges excluding either or both of those included limits are also included in the present disclosure.
[0135] The use of any and all examples, or exemplary language (e.g., “such as”) provided herein, is intended merely to better illuminate the invention and does not pose a limitation on the scope of the present disclosure unless otherwise claimed. No language in the specification should be construed as indicating any non-claimed element as essential to the practice of the present disclosure.
[0136] All methods described herein are performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context. In regard to any of the methods provided, the steps of the method may occur simultaneously or sequentially. When the steps of the method occur sequentially, the steps may occur in any order, unless noted otherwise. In cases in which a method comprises a combination of steps, each and every combination or sub-combination of the steps is encompassed within the scope of the disclosure, unless otherwise noted herein.
[0137] Each publication, patent application, patent, and other reference cited herein is incorporated by reference in its entirety to the extent that it is not inconsistent with the present disclosure. Publications disclosed herein are provided solely for their disclosure prior to the filing date of the present disclosure. Nothing herein is to be construed as an admission that the present disclosure is not entitled to antedate such publication by virtue of prior disclosure. Further, the dates of publication provided may be different from the actual publication dates, which may need to be independently confirmed.
[0138] It is understood that the examples and embodiments described herein are for illustrative purposes only and that various modifications or changes in light thereof will be suggested to persons skilled in the art and are to be included within the spirit and purview of this application and scope of the appended claims.
[0139] Examples
[0140] EXAMPLE 1
[0141] This example describes materials and methods for the subsequent example (Example 2).
[0142] LUD2015-005 inoperable cohort
[0143] The LUD2015-005 clinical trial is a phase I / II study of durvalumab as first-line therapy for patients with esophageal or gastro-esophageal cancer. Adult patients with a histologically confirmed diagnosis of locally advanced or metastatic esophageal or gastro-esophageal cancer were eligible for enrolment into the inoperable arms of the LUD2015-005 trial. Patients with previous systemic anti-cancer therapy for this same advanced disease were excluded from this study, as were patients who had been treated with ICI in any previous setting. Inclusion criteria included an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1, and an anticipated lifespan of greater than 4 months.
[0144] Eligible patients were enrolled into one of three treatment arms for inoperable patients based on a dose escalation strategy, which proceeded to the next stage following establishment of acceptable safety profile at each step. Patients in the first stage (n = 12) received 750 mg of intravenous durvalumab administered biweekly (Q2W) as the ICI agent, those in the second (n = 5) received the same regimen of durvalumab plus a single 37.5 mg priming dose of tremelimumab, and those in the third (n = 21) received durvalumab plus a single 75 mg priming dose of tremelimumab. Expansion at the recommended dose (third stage) to at least 20 patients was driven by the need to establish the safety profile of combination therapy as well as a preliminary assessment of efficacy. After 4 weeks of treatment with these ICI agents alone, patients in all cohorts received a maximum of six cycles of chemotherapy with oxaliplatin and capecitabine, in addition to continued Q2W durvalumab. In the absence of a reason to discontinue treatment earlier, durvalumab infusions continued until the end of the final chemotherapy cycle. The demographic details of the 38 inoperable patients who received treatment on the LUD2015-005 trial can be found in Table 1.
[0145] Supplemental patient-derived material
[0146] For scRNA-sequencing, additional biopsies were also taken from three patients from the operable arms of the LUD2015-005 trial, which enrolled patients deemed suitable for surgery with curative intent. These operable EACs were treated in the neoadjuvant setting with a similar regimen of durvalumab, to which chemotherapy or chemoradiotherapy was added following an ICI-only window. Clinical and demographic details of these patients can be found in Table 13. Additional biopsies were also collected from patients undergoing routine monitoring of known BE by the Translational Gastroenterology Unit (TGU) Biobank (John Radcliffe Hospital; REC reference 1 l / YH / 0020).
[0147] Assessment of patient outcomes
[0148] The primary outcomes of this phase I / II study involved safety: the number of subjects reporting adverse events (up to one year after first dose of study medication), number of subjects experiencing a dose-limiting toxicity (up to 10 weeks after first dose of study medication), and any changes from treatment onset in laboratory evaluations (screening through three months after last dose of study medication). Secondary outcome measures were OS, PFS, and tumor response. Response was assessed using Immune-related Response Evaluation Criteria in Solid Tumors (irRECIST) criteria from computed tomography (CT) scans conducted during the screening period and every 6 weeks thereafter during treatment. irRECIST outcomes were classified as complete responses (irCR), partial responses (irPR), progressive disease (irPD), or stable disease (irSD). One patient with an overall response of irNN, which occurs when there is no measurable target lesion at baseline and an absence of on-treatment irPD or irCR for non- target lesions, was considered to be irSD for the purposes of this work. Three patients who did not have an on- treatment CT scan for assessment of irRECIST criteria but were deemed to have clear on-treatment clinical progression were classified as irPD. irCR and irPR were classified as responders, and all other response categories were classified as non-responders. Tumor shrinkage during the ICI-4W window was calculated as the difference in sum of diameters of target lesions between pretreatment and C1D8 CT scans. One patient (EAC-ECJA) did not have an available CT scan at the C1D8 timepoint, but did have an unscheduled CT scan just before (30 days after treatment onset), which was classified with C1D8 scans for the calculation of ICI-4W tumor shrinkage in this study.
[0149] Overall survival (OS) and progression free survival (PFS) were calculated including data from the post-study follow-up period (data cut-off 16 October 2020). A post-hoc outcome metric for long-term disease control was also defined, which was termed “clinical benefit.” A patient was deemed to have attained clinical benefit if a centralized review process determined they had 12 months of survival from the initiation of treatment without confirmed disease progression. One patient who received an alternative therapy within 12 months of treatment onset was excluded from clinical benefit analyses.
[0150] Research study procedures and sample collection
[0151] Endoscopic biopsies were collected for research analysis prior to the onset of treatment, during the fourth week of treatment with ICI alone (prior to the onset of chemotherapy), and at the end of the study treatment protocol (approximately 5-6 months after treatment onset). At each endoscopy, up to five 2 mm biopsy pairs were taken from the site of the tumor, while 1-2 pairs of normal control tissue were also taken from the descending duodenum (D2), gastric cardia (at least 2 cm distal to the gastro-esophageal junction (GEJ) or distal extent of lesion if crossing the GEJ), and normal esophagus (at least 2 cm proximal to the GEJ or proximal extent of lesion, whichever is more proximal). Each biopsy pair was split into two aliquots, one of which was immediately snap-frozen using dry ice or liquid nitrogen, while the other was slow frozen with 1 mL of fetal bovine serum (FBS) or heat-inactivated human serum with 10% DMSO and placed at -80°C in a controlled-rate freezing container. Snap and slow-frozen biopsies collected at local enrolment sites were stored at -80°C, and were shipped on dry ice to the central site for long- term storage at - 80°C and subsequent analysis.
[0152] For supplemental operable EAC and Barrett’s patients profiled for scRNA-seq, a similar protocol was followed as for the LUD2015-005 inoperable cohort. For operable EAC patients, 2 mm endoscopic biopsies were obtained from tumor and normal esophagus prior to treatment and after 4 weeks of durvalumab alone, and were processed for snap and slow-freezing as above.
[0153] At the surgery marking the end of the study protocol for these operable EAC patients, core biopsies of the tumor site and normal esophagus were also obtained from resected material. For Barrett’s patients, 2 mm endoscopic biopsies of Barrett’s Esophagus (BE) and paired normal esophagus (at least 2 cm proximal to extent of BE) were collected at a single timepoint. Core biopsies from LUD2015-005 surgical specimens and the supplemental biopsies from BE patients were collected fresh in a 1.5 mL microcentrifuge tube containing DMEM / F-12 with 10% HS, and were kept on ice until processing. Supplemental biopsies were either slow-frozen in HS with 10% DMSO in a controlled-rate freezing jar as above or kept unfrozen on ice for dissociation prior to single-cell RNA-sequencing (see below).
[0154] Single-cell RNA-sequencing
[0155] Tissue dissociation
[0156] Prior to dissociation, slow-frozen biopsies were thawed by agitation of the cryovial in a 37°C water bath. For both fresh and frozen material, biopsies were washed in PBS, then transferred into either fresh or frozen tissue dissociation medium (Table 14) and dissociated under constant rotation in an incubator at 37°C and 5% CO2 until fragments reduced considerably in size. The dissociated solution was then filtered through a 100 pM CellTrics filter and the resulting filtrate was centrifuged to pellet cells. For fresh material, red blood cell lysis was performed using red blood cell lysis solution, and dead cells were depleted using a magnetic- bead-based approach. The resultant cell pellets were then resuspended in pre-chilled cell resuspension buffer (0.04% Bovine Serum Albumin in PBS), and filtered through a 70 pM Flowmi tip strainer. Filtered cell suspensions were then counted using an automated cell counter (Bio-Rad, TC20), diluted to a concentration of approximately 1 million total cells / mL in cell resuspension buffer, and kept on ice until encapsulation.
[0157] Encapsulation, library preparation, and sequencing scRNA-seq was conducted using a 5’ scRNA-seq gene expression workflow (Chromium Single Cell Immune Profiling, Solution vl.l, lOx Genomics). Encapsulation of cells was performed using the Chromium Controller, and then GEM-RT, cDNA amplification, and construction of final libraries was conducted following manufacturer’s instructions. Size profiles of amplified cDNA and final sequencing-ready libraries were verified by on-chip electrophoresis (Agilent 2100 Bioanalyzer system) using the High Sensitivity DNA Kit, and concentration of final libraries was assessed by Qubit. Libraries were sequenced on an Illumina NovaSeq 6000 or NextSeq 500 (26 cycles read 1, 8 cycles i7 index, 98 cycles read 2), targeting a minimum of 20,000 reads per cell. Bulk tissue RNA-sequencing
[0158] Total RNA was extracted from whole endoscopic biopsies using the mirVana miRNA Isolation Kit. Briefly, snap-frozen biopsies were transferred directly into pre-chilled lysis / binding buffer in a 1.5 mL microcentrifuge tube, and the tissue was homogenized on ice using a disposable RNAse-free pestle. Acid-Phenol: Chloroform RNA extraction was then performed following manufacturer’s instructions. Genomic DNA was digested from the eluted RNA using a 30-minute incubation with TURBO DNase. RNA concentration was then calculated using the Qubit RNA BR Assay and the RNA integrity number (RIN) was calculated using the Agilent 2100 Bioanalyzer system (Total RNA Assay). RNA was then plated and stored at -80°C. Prior to library preparation, thawed RNA was purified and concentrated using AMPure XP at a 2.8x ratio. Bead-based rRNA depletion and total RNA-sequencing library preparation with dual sample indexing was then conducted using the TruSeq Stranded Total RNA Library Prep Human / Mouse / Rat Kit. Libraries were sequenced on the Illumina HiSeq (75 cycles read 1, 8 cycles i7 index, 8 cycles i5 index, 75 cycles read 2) to a targeted depth of 50 million reads per library.
[0159] For the monocyte spike-in experiment, RNA extraction was conducted on additional esophageal cancer biopsies as above (3 EAC, 1 ESCC). RNA extraction from purified peripheral blood monocyte populations (see Cell sorting) was conducted using the same protocol, with homogenization by vortexing rather than by pestle. For each of four tumor biopsies, subaliquots were generated with monocyte RNA spiked in at 0%, 0.5%, 1%, 2%, and 4% of the total RNA mass. For two patients with excess RNA, additional subaliquots with 0% and 8% monocyte RNA were generated. Library preparation was conducted on these RNA mixtures using the TruSeq Stranded Total RNA Library Prep Human / Mouse / Rat as above. All libraries were sequenced on the same NextSeq 2000 flow cell, targeting a depth of 50 million reads per sample.
[0160] Whole genome sequencing
[0161] Genomic DNA was extracted from whole endoscopic biopsies using GeneJET Genomic DNA Purification Kit according to manufacturer's protocol. DNA quality and quantity was measured using Nanodrop and Qubit dsDNA HS assay kit. Extracted DNA samples were then sent out for library preparation using a PCR-free protocol and sequencing using the Illumina NovaSeq6000.
[0162] Cell sorting PBMCs were isolated from healthy donors’ leukapheresis cones by density gradient separation (NHS Blood and Transplant, UK). Cells were resuspended in PBS containing 0.5% BSA (Sigma) and 2mMEDTA (Gibco), and stained with Zombie NIR, CD14-FITC M5E2, CD 16- APC clone 3G8, CD33-PE clone WM53, and CD56-BV510 clone 5.1H11 (all from Biolegend). Non-classical and classical monocytes were sorted according to the gating strategy in Figure 13B using a BD FACSAria Fusion. To form the monocyte population for spike-in, sorted classical and non-classical monocytes were mixed in 10: 1 ratio before RNA extraction, mirroring the proportions of intratumoral monocyte subpopulations in the LUD2015-005 scRNA-seq atlas.
[0163] Bioinformatics
[0164] Repository data access
[0165] For deconvolution, public bulk RNA-sequencing data for normal esophagus, BE, and EAC were retrieved from ENA accession PRJEB11797 (Maag, etal., 2017. Mol Cancer Res 15, 1558— 1569), and for ICI -treated GC from ENA accession PRJEB25780 (Kim, et al., 2018. Nature Medicine 24, 1449-1458) and EGA accession EGAD00001008091 (Cheong, et al., 2022. Nat Commun 13, 774). Pre-processed RNA-seq counts and clinical outcome information for EAC patients in the TCGA were downloaded from the Genomic Data Commons using the TCGAbiolinks package (Colaprico, A., etal., 2016. Nucleic Acids Res 44, e71.) in R. ICGC EAC raw sequencing and outcome data were downloaded from EGAD00001004423 (Frankell, et al., 2019. Nature Genetics 51, 506-516).
[0166] Whole genome sequencing data analyses
[0167] Whole genome sequencing data were processed for alignment to the GRCh38 human genome, mutation calling, and copy number analysis in line with current best practices using a set of fully reproducible Nextflow (Di Tommaso, P., et al., 2017. Nat Biotechnol 35, 316-319) pipelines. All pipeline definitions and associated information can be found at https: / / bitbucket.org / licroxford / carroll_etal_2022.
[0168] Briefly, mutations were called using a consensus of two of three callers (Strelka2, Mutect2 and Octopus). Following previous practice (Zehir, A., etal., 2017. Nat Med 23, 703-713), tumor mutational burden was quantified as the rate of non-synonymous mutations in the coding space, reported as mutations per megabase. The size of the coding space used in this study was 35.6 Mb, calculated as the sum of the protein-coding exons from ENSEMBL100 gene annotation. Copy number variations were called using Battenberg. Ambiguities in calls of tumor ploidy were resolved by manual inspection and agreement of two authors (JK and IP). Using the majority clone number from Battenberg, Amplification was defined as >2 copy number-adjusted ploidy while deletions were defined as change of more than half of the baseline gene copies. For the TCGA cohort, tumor mutational burden metrics were downloaded from the source publication using TCGAbiolinks (Colaprico, A., etal., 2016. Nucleic Acids Res 44, e71.; The Cancer Genome Atlas Research Network, 2017). For the ICGC cohort, mutation calls were downloaded from the ICGC data portal (Zhang, J., etal., 2019. Nat Biotechnol 37, 367-369), and the rate of non-synonymous coding variants per megabase was calculated as for the LUD2015- 005 cohort.
[0169] Genomic signatures of MSI were detected using MSIsensor (Jia, P., et al., 2020. Genomics Proteomics Bioinformatics 18, 65-71). Tumors were determined to have MSI if the MSIsensor score was at least 3.5, a threshold which has previously been used to classify tumors as MSI using this tool (Niu et al., 2014 Bioinformatics 30, 1015- 1016; Johansen etal., 2019. BMC Cancer 19, 971).
[0170] Single-cell RNA-sequencing data analyses
[0171] Processing of scRNA-seq data was conducted using the Cell Ranger pipeline from lOx Genomics. Briefly, raw BCL files were converted to demultiplexed FASTQ files using cellranger mkfastq with -use-bases-mask=Y26n*,I8,Y98n*. A custom reference sequence was then prepared using the concatenation of the GRCh38 human genome (ENSEMBL100) and supplementary contigs from the Genomic Data Commons alignment reference, namely the decoy contig hs38.dl and a collection of viral genomes commonly found in human cancer samples. The full GENCODE v34 GTF was used as a gene annotation file, filtering out readthrough transcripts and annotations to PAR regions in chromosome Y as per Cell Ranger recommendations. A custom Cell Ranger reference set was then prepared from these sequence and annotation files using cellranger mkref with default arguments. Finally, the identification of true cells and construction of a UMI counts table for each sample was performed using cellranger count, and these per-sample counts were combined into a single matrix for downstream analysis using cellranger aggr without normalization. Downstream analysis was conducted using Seurat (Stuart et al., 2019. Data. Cell 177, 1888-19O2.e21). To begin, low-quality cells with less than 200 genes detected or more than 25% mitochondrial reads were removed. For initial clustering of the full dataset, it was normalized using SCTransform (Hafemeister and Satija, 2019. Genome Biol 20, 296), performed UMAP dimensionality reduction, and identified clusters with the default Louvain algorithm. Broad cell type partitions were then identified and subsetted into separate Seurat objects. To characterize smaller cell subsets, the above workflow was repeated on each of these partitioned objects. To minimize batch effect between fresh and frozen samples for these smaller objects, SCTransform normalization was conducted on fresh and frozen samples separately, and then integrated these two datasets using FastMNN (Haghverdi et al., 2018. Nat Biotechnol 36, 421-427), a mutual nearest neighbor batch correction method, before proceeding to dimensionality reduction. Marker genes for each identified cluster were calculated using the FindAllMarkers function, and these markers were used to manually assign cell type and cell subtype labels.
[0172] Clusters representing low quality cells, detected based on abnormally high mitochondrial percentage, increased levels of a gene signature representative of dissociation-induced stress, and / or an abnormally low number of features within a given cell type, and clusters representing doublets, detected based on marker expression with an abnormally high number of features as supporting evidence, were identified and removed from the final scRNA-seq object. For detection of cell type-specific markers associated with patient outcome, a pseudobulk differential expression method was designed to minimize patient-specific genes and prioritize genes associated with outcome in multiple patients. Briefly, SCTransform-normalized counts from all cells in a given biopsy were summed, resulting in a pseudobulk sample. Differential expression was then conducted on these pseudobulk samples using DESeq2 as discussed below, with clinical benefit as the design formula. Figures visualizing the scRNA-seq findings were generated using a combination of Seurat and ggplot2.
[0173] Bulk RNA-sequencing analysis
[0174] Raw RNA-sequencing data were processed following best practices using a reproducible Nextflow pipeline. This workflow was based off of STAR (Dobin et al., 2013. Bioinformatics 29, 15-21) alignment to the same reference used for scRNA-seq alignment, and generation of final counts matrices using featureCounts (Liao et al., 2014. Bioinformatics 30, 923-930). The reproducible pipeline and associated information can be found at https: / / bitbucket.org / licroxford / carroll_etal_2022.
[0175] Differential expression was conducted using DESeq2 using a significance threshold of FDR < 0.1 to identify differentially expressed genes. Default arguments were used, except for the use of the “local” fit type to model dispersion estimates. Moderated gene-level log-fold changes were calculated within DESeq2 using the “ashr” method (Love etal., 2014. Genome Biol 15, 550), and were used as the test statistic for gene set enrichment analysis, which was conducted using FGSEA on a database of pathways from Reactome and MSigDB (Hallmark and Gene Ontology). Log-normalized counts for heatmap visualization were generated using the varianceStabilizingTransformation function. To assess changes between timepoints while correcting for patient-specific effects, a design formula of -patient+timepoint was used.
[0176] Deconvolution analysis
[0177] 45 of the 46 major cell types identified from scRNA-seq were used for bulk RNA-seq deconvolution, with one cell type (Schwann cells) excluded due to their rarity. For all deconvolution analyses, normalized counts in a linear scale were used for deconvolution, which were obtained for scRNA-seq data using the vst function from the sctransform package and for bulk RNA-seq data using the median of ratios method embedded in the estimates izeFactors function from DESeq2. The combination of these normalization methods has been shown to provide accurate results across a variety of deconvolution approaches (Avila Cobos et al. , 2020. Nat Commun 11, 5650).
[0178] For benchmarking, open-source tools Bisque (Jew et al., 2020, Nat Commun 11, 1971), BSEQ-sc (Baron etal., 2016, Cell Syst 3, 346-360.e4), MuSiC (Wang et al., 2019, Nat Commun 10, 380), and BayesPrism (Chu et al., 2022) were installed from their GitHub repositories and ran on the same HPC cluster. CIBERSORTx (Newman et al., 2019, Nat Biotechnol 37, 773-782) was run from a docker container on an AWS instance. For all conditions tested, potentially confounding gene types (mitochondrial, ribosomal, sex-specific, and TCR / BCR variable region genes) were excluded from the input reference matrix prior to each run. Two input reference matrices were used, one including all genes, and another subsetted to just the 3000 most variable genes (as calculated by FindVariableFeatures from Seurat). For methods that allowed the sample or subject ID for each cell in the input reference matrix to be used during deconvolution (Bisque and MuSiC), the patient ID was passed to the deconvolution call. BayesPrism can additionally incorporate cell subtype information and malignant cell labels from the input reference matrix. Cell type and subtype labels were therefore used with all BayesPrism runs, with EAC given as the malignant cell type label for deconvolution of all gastro-esophageal adenocarcinoma samples, and ESCC for the deconvolution of squamous cell carcinoma samples (1 ESCC in monocyte spike-in experiment). For BSEQ-sc, markers of each cell type were pre-selected using Seurat’s FindAllMarkers and a downstream filtering strategy (pct.l > 0.25, pct.2 < 0.75, p.adj < 0.01). CIBERSORTx (cibersortx / fractions docker container) was run with 100 permutations for significance calculation and recommended settings for droplet-based scRNA-seq (—fraction 0 - rmbatchSmode TRUE).
[0179] Benchmarking was conducted on 80 pseudobulk samples, created by adding together scRNA- seq counts from known numbers of cells in the LUD2015-005 atlas. These pseudobulk samples were generated to represent a range of possible outcomes in this setting, including varying degrees of tumor purity, immune infiltration, and contamination with stroma and squamous epithelium. The Spearman correlation coefficient and normalized root mean square error (Normalized RMSE, defined as RMSE divided by the range of predicted values) were calculated, both against known RNA proportions (the fraction of counts assigned to each cell type) and known cell proportions (the cell numbers of each cell type used to generate the pseudobulk sample divided by the total number of cells). The same output of all algorithms was used for both RNA and cell comparisons, except for BayesPrism; for this tool, the output of the main algorithm was used to predict RNA fraction, and the optional helper function estimate sf was used to predict the cell fraction.
[0180] Final deconvolution analysis on bulkRNA-seq samples were conducted using BayesPrism. As restricting the input reference matrix to the subset of the 3000 most variable genes significantly improved the run time of these algorithms, while returning similar or even slightly improved accuracy (Table 7), the 3000-gene matrix was used for BayesPrism deconvolution of these datasets. Potentially confounding genes were removed prior to the run, and EAC cells were labelled as malignant, as above. Normalized and scaled cell type proportion estimate for individual cell types were used to perform Cox regressions against overall survival, and estimates for cell types within broad cellular compartments were used to perform hierarchical clustering of samples (using the hclust method with “ward.D” linkage). Survival analysis
[0181] The link between variables and survival outcomes were assessed using the survivalAnalysis package in R. analyze_survival was used to conduct univariate Cox regressions and to generate Kaplan-Meier plots, while analyze multivariate was used for multivariate Cox regressions and forest plots. Unless otherwise noted, continuous variables were scaled and centered prior to regression. For propensity matching analysis, patient treatment, demographic, and clinical outcome data were extracted from the Surveillance, Epidemiology, and End Results Program (SEER), a US-based registry of disease outcomes, using SEER*STAT. For comparison with inoperable LUD2015-005 patients, only patients with stage III or IV esophageal cancers treated with chemotherapy were included. Optimal matching between LUD2015-005 and this SEER subset was performed using the Matchit package (Ho et al., 2011, Journal of Statistical Software 42, 1-28) with optimal matching and a 1:4 ratio. Tumor stage, histological subtype, sex, age, and primary ethnicity were the covariates accounted for during propensity matching.
[0182] Quantification and statistical analysis
[0183] All statistical analyses were done using R. Boxplot visualizations all have a vertical line denoting the minimum to maximum range (excluding outliers), the box marking Q1-Q3, and a horizontal line denoting median. Details for the statistical tests employed can be found in the relevant Method Details section and figure legends. In the case of multiple testing, p-value adjustment was performed using the Benjamini-Hochberg procedure (Benjamini and Hochberg, 1995) to control for the false discovery rate (FDR). A threshold of 0.05 was used to determine significance throughout, with the exception of DESeq2 differential expression testing, where the package default threshold of FDR < 0.10 was used.
[0184] EXAMPLE 2
[0185] Treating esophageal cancer with ICI-4W prior to ICI+CTX in the LUD2015-005 trial
[0186] The experimental phase I / II LUD2015-005 trial (NCT02735239, EudraCT 2015-005298- 19) was designed to understand both ICI and ICI+CTX responses in previously untreated inoperable esophageal cancers. Treatment in the LUD2015-005 trial began with a four-week ICI- only window (ICI-4W), followed by six cycles of ICI+CTX. During the ICI-4W phase of treatment, either aPD-Ll (n=12) or aPD-Ll with a single priming dose of aCTLA-4 (n=26) was administered; both groups then received aPD-Ll ICI alongside oxaliplatin and capecitabine during the ICI+CTX phase. Serial endoscopic biopsies and CT scans were collected before and throughout treatment as indicated (Figures 1 A and 2).
[0187] The demographic and clinical characteristics of the 38 patients (35 EAC; 3 esophageal squamous cell carcinoma (ESCC)) in the intention-to-treat population (ITT) for the LUD2015-005 inoperable cohorts are provided in Table 1. Primary outcomes were related to safety (see STAR Methods). All ITT patients reported at least one treatment emergent adverse event (TEAE), with 29 (76.3%) reporting grade 3 or higher TEAEs (Table 2 and Table 3). No dose-limiting toxicities were encountered during the dose escalation phase of this study (Table 2).
[0188] Secondary outcomes were OS, progression-free survival (PFS), and tumor response as measured by irRECIST. Median OS and PFS in the ITT population were 13.4 and 9.1 months, respectively (Figures 1B-C, Table 4). Comparing the OS of LUD2015-005 patients with that of a propensity-matched cohort of CTX-treated patients from SEER, a US -based registry of cancer cases, suggested a potential trend towards increased OS in the LUD2015-005 cohort, but this trend did not reach statistical significance, likely due to small cohort size of the LUD2015-005 trial.
[0189] In the ITT population, 17 / 38 patients (44.7%) showed an irRECIST response, defined as a 30% reduction in tumor burden from pre-treatment measurements (Table 4). Many patients already displayed tumor shrinkage during ICI-4W; indeed, three patients met the definition for response immediately following the ICI-4W window, indicating four weeks of ICI alone is sufficient to induce tumor shrinkage in a subset of patients. Notably, some patients showed differing responses between the different phases of therapy, with a few patients with unconfirmed progression (at least a 20% increase in tumor burden) during ICI-4W going on to attain stable disease or responses during the ICI+CTX phase. These could indicate initial progression on ICI with subsequent tumor shrinkage in response to CTX addition, or instances of pseudoprogression, where ICI-induced immune influx preceding an eventual anti-tumor response results in a transient increase in measured tumor size.
[0190] Four irRECIST non-responders still experienced prolonged PFS (>12 months). As the strength of association between response and long-term survival outcomes may vary, an additional “clinical benefit” outcome metric was established to denote long-term disease control on this treatment protocol. Briefly, patients attaining 12 months of PFS were deemed to have attained clinical benefit (CB), while the rest were classified as no clinical benefit (NCB). Based on this metric, 15 of 37 (40.5%) assessable patients attained CB (Table 4). Due to its definition, CB classification was more closely linked with long-term survival outcomes than irRECIST response. In summary, four weeks of first-line ICI-only was sufficient to induce tumor shrinkage in some patients with inoperable esophageal cancers, while around 40% achieved CB on subsequent aPD- L1 ICI+CTX.
[0191] ICI-4W treatment induces INCITE signature, an on-treatment marker of ICI efficacy
[0192] As 35 of 38 patients in the LUD2015-005 study had EAC, the molecular analyses were limited to EAC patients with available tissue samples. The design and sample collection protocol of the LUD2015-005 trial provided a unique opportunity to elucidate early response and resistance mechanisms to ICI-4W treatment in EAC. A multi-timepoint bulk RNA-seq dataset was generated using tumor samples from 33 EAC patients and paired normal GI tissues from a subset of patients. Paired tumor transcriptomes from both before (PreTx) and after (ICI-4W) the initial ICI-4W window were available for 28 EAC patients. Assessing ICI-responsive transcriptional changes between PreTx and ICI-4W across these 28 patients, it was found that the top differentially expressed genes (DEGs) upregulated during the ICI-4W window were dominated by markers of cytotoxic inflammation, including T cell chemokines (CXCL9, CXCL10, and CXCL17), T / NK cell markers (CD2, CD3D / E, CD8A / B, Killer cell lectin-like receptor (KLR) family, NKG7, and TRGC2), cytotoxic effector molecules (GZMA, GZMH, GZMK, PRF1, and FASLG), and markers of CD8+ tissue-resident memory T cells (7TGZE / CD103, ZAF6S3 / Hobit) (Table 5). As the top 70 upregulated DEGs appeared particularly enriched for these cytotoxic markers, these top 70 DEGs were used to define an ICI-responsive gene signature, termed INCITE (ImmuNe Checkpoint Inhibitor-induced T / NK-cell Enrichment).
[0193] The predominance of these T-cell inflammation markers among top DEGs demonstrated that ICI-only treatment for four weeks was sufficient to induce appreciable immune responses against EAC. Upregulation of INCITE genes was observed in nearly all patients at ICI-4W, although the extent varied considerably. Many INCITE genes were significantly upregulated both in patients receiving aPD-Ll and those receiving aPD-Ll+aCTLA-4 during the ICI-4W window, showing that both strategies could induce detectable immune responses at four weeks. The extent of INCITE upregulation was significantly correlated with shrinkage in tumor burden during the ICI-4W window, indicating the induction of this transcriptional signature was linked to overall ICI efficacy. Ordering patients by their mean upregulation of INCITE genes, patients were subdivided into INCITE-reactive (greater ICI-induced INCITE upregulation) and INCITE-quiescent (less upregulation) groups. All three patients with unconfirmed irRECIST progression during ICI-4W were INCITE-quiescent, indicating these cases did not represent pseudoprogression due to intratumoral immune cell influx, while the INCITE-reactive group included all three patients exhibiting irRECIST responses following ICI-4W. Notably, the patient with the highest INCITE upregulation (EAC-JCNP) also showed the most tumor shrinkage during ICI-4W. These findings indicate that intratumoral INCITE upregulation is a strong correlate of early responses to treatment with ICI-only.
[0194] Gene set analysis revealed that upregulation of various inflammatory and interferon gamma-related signatures were strongly linked with tumor shrinkage. Upregulation of INCITE outperformed other inflammatory gene sets as a correlate for ICI-4W response (FDR = 1.3e-21). Conversely, upregulation of E2F targets (e.g., PRDX4, MCM2,' FDR = 2.6e-24) and extracellular matrix genes (e.g., COCH, MATN3,' FDR = 8.0e-18) were strongly associated with tumor growth and resistance to ICI-4W treatment. These results show that four weeks of ICI is sufficient to induce appreciable T-cell inflammation in this setting, best captured by the INCITE signature, and that on-treatment increase in INCITE is a key marker of anti-tumor immune responses during early ICI-only treatment.
[0195] Next, whether a subset of INCITE genes could reproduce the significant correlation between the full signature and ICI-induced tumor shrinkage was assessed, as a smaller gene panel would be more amenable to development towards clinical applications. After testing all subset sizes, the top 12 INCITE genes (INCITE-12) were selected as a compact signature showing a strong correlation with tumor shrinkage while preserving robustness against randomly generated noise (Table 5). The performance of INCITE-12 was compared against five other inflammatory signatures used as predictive biomarkers in clinical ICI research. Of all signatures tested, only upregulation of INCITE-12 showed a significant correlation with ICI-induced tumor shrinkage in LUD2015-005 EAC patients (Table 6).
[0196] While INCITE-12 upregulation was greater in early responders to ICI-only in the LUD2015-005 protocol, it was not associated with long-term CB following ICI+CTX, perhaps due to some ICI-resisting tumors responding to the CTX component of therapy. To further assess the utility of INCITE as a biomarker for ICI-only protocols, upregulation of INCITE-12 was assessed alongside the five other inflammatory signatures in an independent cohort of ICI-naive melanoma patients receiving aPD-1 ICI, where sampling was conducted before and after a similar four-week ICI-only window. In these patients, upregulation of INCITE-12 during the first four weeks was significantly associated with overall ICI response and benefit, showing the strongest association with these variables of any gene set tested (Table 6). In another cohort of patients from the same report who had previously progressed on aCTLA-4 ICI, no signature showed a significant association with outcomes, indicating INCITE-12 was most suitable for the ICI-naive setting (Table 6). Together, upregulation of INCITE genes, especially INCITE-12, is a promising early on-270 treatment biomarker associated with ICI-only outcomes.
[0197] Immune responses associated with early ICI outcomes are detectable with singletimepoint ICI-4W sampling
[0198] As on-treatment INCITE upregulation was a strong correlate of tumor shrinkage during ICI-4W, whether PreTx expression levels of INCITE or other genes could predict these responses was next examined. For this single-timepoint analysis, DESeq2 was used to calculate DEGs associated with changes in tumor burden during the ICI-4W window as a continuous variable. DEGs and GSEA revealed that PreTx levels of T-cell inflammation signatures such as INCITE (FDR = 0.66) were not associated with tumor shrinkage following ICI-4W treatment. PreTx PD- L1 expression was not significantly associated with ICI-4W tumor shrinkage (FDR = 0.93; Table 5). PreTx genes and gene sets associated with tumor shrinkage during ICI-4W include pathways related to neural and muscle cell development, while tumor growth during ICI-4W was associated with higher PreTx expression of pathways related to CYP450, a superfamily of genes involved in xenobiotic metabolism. These pathways in the PreTx analysis had relatively weaker associations with ICI-4W outcomes than did the dynamic analysis. These findings indicate that the strength of association between PreTx expression and ICI-4W outcomes may be limited, and that commonly assessed PreTx ICI biomarkers, including PD-L1 and T cell inflammation signatures, may not be strongly predictive of ICI-only outcomes in inoperable EAC.
[0199] In contrast, when the same differential expression approach was applied to stand-alone analysis of on-treatment ICI-4W biopsies, a number of highly significant associations were detected with changes in tumor size during the ICI-4W window. GSEA showed that levels of several inflammatory signatures in the ICI-4W tumor biopsy were strongly linked with tumor shrinkage, largely recapitulating the dynamic analysis; of these, INCITE again showed the most significant association. Markers of y8 T cells TRDC, TRGC1, and TRGC2) appeared to be key contributors to the enrichment of inflammatory gene signatures in this analysis. ICI-4W levels of genes related to the extracellular matrix, including trypsinogens (PRSS1, PRSS2) and matrix metalloproteases, were associated with ICI resistance and tumor growth during the ICI- 4W window. A number of genes encoding cancer antigens, such as CTAG2, SAGE1, MAGEA1 / A4 / A10, POTEE, and PRAME, were also significantly associated with tumor growth (Table 5). Since cancer antigens can be immunogenic and are mainly expressed in cancer cells, this finding indicates that ICI-4W treatment failed to generate sufficient anti-tumor immune responses in progressing patients, resulting in higher tumor growth associated with preserved higher expression of these genes following ICI-4W.
[0200] In summary, compared with PreTx sampling, sampling after four weeks of ICI allowed for a much clearer molecular distinction between ICI-sensitive patients, who show high INCITE T- cell inflammation, and ICI-insensitive patients, who are enriched for cancer-germline genes and certain extracellular matrix genes. Therefore, single-timepoint sampling after four weeks of ICI provides an opportunity to evaluate treatment-emergent molecular signatures of ICI response and resistance that are comparatively absent from pre-treatment tumors, and sampling at this timepoint could help supplement radiological response in determining expected long-term ICI-only outcomes soon after treatment onset.
[0201] IPRES signatures at ICI-4W mark INCITE-quiescent tumors resisting ICI treatment
[0202] Given the association between extracellular matrix genes and ICI resistance, both in the standalone analyses at ICI-4W and in assessing changes between PreTx and ICI-4W, Innate PD-1 Resistance (IPRES) signatures were evaluated, which were a group of gene pathways previously described to predict ICI resistance in melanoma, which include stromal modules for angiogenesis and wound healing. ICI-4W biopsies from INCITE-quiescent patients showed much higher levels of IPRES compared to INCITE-reactive patients. This finding was not just specific to the stromal modules of IPRES: a similar trend could be seen for the hypoxia, EMT / Metastasis, and TGF0 modules included in IPRES. Unlike in the original melanoma study, in this EAC cohort, only IPRES scores at ICI-4W were associated with ICI resistance; IPRES scores at PreTx showed no significant association with INCITE reactivity.
[0203] Of note, a subset of INCITE-quiescent patients with high levels of IPRES at ICI-4W went on to achieve CB during the ICI+CTX window. This signifies that ICI+CTX can overcome ICI- only resistance in some patients, but could also be due to delayed responses to ICI-only in some patients. These findings indicate that IPRES and INCITE signatures can mark early ICI response and resistance.
[0204] Innate immune signatures predict clinical benefits of immunochemotherapy
[0205] As a de novo approach to identify biomarkers that could better predict ICI+CTX outcomes in EAC, differential expression analysis of PreTx tumor biopsies between CB and NCB patients were conducted. Surprisingly, DEGs and gene sets significantly associated with CB were related to the innate immune system, particularly the myeloid compartment (TREM1, ACOD1, TNFSF14, and CSF3R), rather than the markers of T cell inflammation and checkpoint molecules which have been commonly described as predictive of positive outcomes in ICI-based regimens (Table 5). To understand whether the importance of innate immune signatures might be due to specific features of the microenvironment in EAC, the expression of various immune markers was compared between PreTx EAC (n = 33) and paired normal GI tissue biopsies (esophagus, gastric, and duodenum) from a subset of patients (n = 5-9). Notably, compared to paired normal GI tissues, myeloid markers (including CD14, FCGR3A / CD I 6A, ITGAM / CD Q, CD163) exhibited significantly higher expression in EAC. In contrasting, dendritic cell, T cell, and B cell markers were expressed at similar levels to paired normal GI tissues. Cytokines and chemokines known to promote tissue infiltration of myeloid cells or modulate their function (CXCL3, CCL2, CSF 7 / M- CSF) were also highly enriched in EAC biopsies, and increased expression of these genes was strongly correlated with increased expression of myeloid cell markers. EAC transcriptomes from public databases (TCGA / GTEX) also showed significant enrichment for myeloid markers and cytokines and chemokines when compared with normal GI biopsies. Increased expression of monocyte, macrophage, and neutrophil markers appeared to be a defining feature of the EAC microenvironment, consistent with enhanced expression of relevant cytokines and chemokines. These findings indicated that a skew in composition or phenotype of the characteristic myeloid enrichment present in EAC likely played an important role in determining ICI+CTX outcomes. Generating a 65,000-cell upper GI cell atlas and establishment of a deconvolution workflow
[0206] To identify which myeloid cell types may influence ICI+CTX outcomes, evidence from both single-cell transcriptomics and bulk RNA-seq were integrated. The LUD2015-005 upper GI cell atlas was first generated, which is a 65,000-cell single-cell RNA-sequencing (scRNA-seq) dataset derived from both diseased and normal gastro-esophageal tissues collected from 18 patients, including inoperable EAC and ESCC, operable EAC, and the pre-malignant lesion Barrett’s esophagus. 46 major cell types were identified in the LUD2015-005 atlas, with good representation of epithelial, stromal, and lymphocytic and phagocytic immune cell compartments.
[0207] Analysis of the LUD2015-005 atlas verified that INCITE genes were specifically expressed in T and NK cells, and confirmed the enrichment of monocytes, macrophages, and neutrophils in tumor samples seen in bulk data. It also showed that myeloid-targeted cytokines and chemokines were expressed by a variety of cell types: particularly by the myeloid cells themselves, but also a subset of EAC cells. Analysis of scRNA-seq expression profiles revealed that the innate immunity DEGs enriched in CB patients in the bulk RNA-seq analysis were mainly expressed in monocytes, neutrophils, and Ml -like macrophages, indicating that the skew of the myeloid infiltrate characteristic of EAC towards one or more of these cell types was predictive of better outcomes on ICI+CTX.
[0208] To further define which myeloid cell type was primarily responsible for the association with CB, a deconvolution workflow was implemented based on the LUD2015-005 scRNA-seq atlas to compute cell type composition in bulk RNA-seq datasets. To assess the accuracy of various deconvolution tools on the data, scRNA-seq data generated in this study were first pooled to create 80 pseudobulk RNA-seq samples. As the precise cell composition used to generate each pseudobulk sample was known, this dataset enabled the assessment of which deconvolution algorithm generated the most accurate results using the LUD2015-005 atlas. Across all pseudobulk samples, deconvolution estimates generated using BayesPrism (Chu et al., 2022, Nat Cancer 3, 505-517) showed the highest median correlation and smallest median error of all algorithms tested (see STAR Methods; Table 7), and BayesPrism was therefore selected for deconvolution using the LUD2015-005 atlas. Using a previously published RNA-seq dataset (Maag et al., 2017, Mol Cancer Res 15, 1558-1569), this deconvolution workflow was able to resolve key differences in the epithelial and microenvironment composition of EAC from that of other esophageal tissue types, including the closely related Barrett’s esophagus (Table 8). Following this in silico benchmarking, the BayesPrism-based deconvolution workflow was taken forward to estimate the cellular composition of bulk transcriptomes from the LUD2015- 005 trial.
[0209] Tumor monocyte content (TMC) identifies gastro-esophageal cancer patients likely to benefit from addition of ICI
[0210] The deconvolution workflow was applied to estimate cell composition of PreTx EAC biopsies from the LUD2015-005 trial. Focusing on phagocytic immune cells, a cluster of EAC biopsies was identified with high tumor monocyte content (TMC) who obtained a higher rate of CB (8 / 12 [66%]) than patients not in the monocyte-high cluster (5 / 20 [25%]). In contrast, no significant association with ICI+CTX outcomes was seen with cell types from epithelial, stromal, or lymphocytic immune compartments (Table 9). Cox regressions were conducted for testing the association between levels of each cell type and OS. PreTx TMC was strongly associated with improved OS, which remained significant following p-value correction (Benjamini-Hochberg procedure) for multiple testing (HR: 0.38 (95% CI: 0.22-0.67), p = 0.0008, FDR = 0.037; Table 9). Other cell types, including neutrophils, Ml-like macrophages, and Tregs, showed associations trending towards improved OS but not reaching significance, indicating TMC was the primary driver of improved outcomes. Using the cohort median of TMC to define monocyte-high or -low groups effectively stratified patient outcomes, with monocyte-high patients showing greatly improved OS compared to monocyte-low patients (median OS 24.3 vs. 8.6 months; Figure 3A). To verify that this predictive biomarker accurately reflected intratumoral monocyte RNA content and not technical artifacts from deconvolution, RNA from peripheral blood monocytes isolated by fluorescence-activated cell sorting (FACS) was spiked into RNA extracted from esophageal cancer biopsies in known quantities (0-8% of total RNA), and RNA-sequencing libraries were then prepared from these samples. Increasing levels of monocyte spike-in RNA were indeed significantly associated with increasing deconvolution-assessed TMC.
[0211] To assess whether TMC was ICI-specific or a general prognostic biomarker, advanced EACs from TCGA and ICGC databases were analyzed: most are treated with platinum / fluoropyrimidine with no ICI reported (Table 10). The same deconvolution workflow was applied to these transcriptomes as for the LUD2015-005 dataset and found no significant link between TMC and OS for either TCGA or ICGC EAC patients (Figure 3B), indicating that the association between TMC and survival was specific to ICI-containing regimens such as the LUD2015-005 ICI+CTX protocol. Within the LUD2015-005 study, there was a significant association between TMC and OS in both the aPD-Ll- and aPD-Ll+aCTLA-4 treatment subgroups, showing this finding was not restricted to either ICI strategy (Figure 3C). Deconvolution-assessed TMC in PreTx EAC biopsies therefore shows powerful potential utility to help identify EAC patients most likely to benefit from the addition of ICI to CTX in first-line therapy.
[0212] While EBV and MSI are associated with improved outcome on ICI-based therapies in GC, there are not yet robust predictive biomarkers for the EBV- / microsatellite stable (MSS) subtypes, which comprise the majority of all GC. Given the molecular similarity between EAC and the most prevalent form of EBV- / MSS GC (The Cancer Genome Atlas Research Network, 2017), whether TMC could also predict outcomes in an independent pooled cohort of 77 ICI-treated GCs with available RNA-seq and MSI / EBV subtyping (Cheong et al., 2022, Nat Commun 13, 774; Kim et al., 2018, Nature Medicine 24, 1449-1458) was assessed. In EBV- / MSS GC, PreTx TMC was significantly higher in responders than non-responders (p = 0.041), showing similar predictive utility for response as in EBV- / MSS EACs from the LUD2015- 005 study (p = 0.022). TMC was not associated with response in EBV+ / MSI gastresophageal cancers. These findings show TMC predicts improved outcomes on ICI-containing regimens across independent gastro-esophageal cancer cohorts, and could be particularly useful for EBV- / MSS subtypes of gastro-esophageal cancer.
[0213] TMC and TMB are independent but complementary predictors of immunochemotherapy
[0214] Whole genome sequencing (WGS) of LUD2015-005 PreTx EAC biopsies (n=33) was therefore conducted to assess TMB and other genomic alterations. Mutation and copy-number profiling showed that the most frequently altered genes in EACs from this cohort were TP 53 (70%) and CDKN2A (27%). A median TMB of 3.3 non-synonymous coding mutations per megabase (muts / Mb) were detected across all EAC patients. Patients attaining CB had a significantly higher TMB than those who did not (median TMB: 4.3 vs 2.1 muts / Mb, Figure 4A). Of the 33 EACs assessed, only one (from EAC-JCNP) was identified as having MSI; this tumor had the highest TMB (13.7 muts / Mb) in the cohort (Table 11). This patient attained CB and showed the greatest extent of tumor shrinkage and INCITE upregulation during the ICI-4W window.
[0215] Stratifying patients by their median TMB showed that patients with higher TMB had significantly improved OS (Figure 4B). A Cox regression on the TMB values themselves also showed the significant association between TMB and improved OS (hazard ratio [95% CI]: 0.50 [0.28-0.90], p=0.021), which was preserved even after excluding the MSI patient (hazard ratio [95% CI]: 0.62 [0.40-0.95], p=0.029). Similar to the findings for TMC, analysis of advanced EAC patients from TCGA and ICGC did not show any predictive power for TMB (Figure 4C), indicating this biomarker was also more specific to ICI+CTX than CTX alone. Moreover, TMB and TMC were each significantly correlated with ICI-induced tumor shrinkage during the ICI-4W window, and their combination showed an even stronger association with ICI-4W tumor shrinkage, supporting the specificity of TMB and TMC for ICI-containing protocols.
[0216] Next, whether TMB and TMC were independently associated with OS on ICI+CTX was investigated. There was no significant correlation between PreTx TMB and TMC in this cohort. EAC patients having both TMB and TMC values above their respective medians showed significantly higher OS than other patients in this cohort, indicating additive predictive power (Figure 4D). In a multivariable Cox regression including TMB and TMC, both variables were significantly associated with longer OS and PFS (Figure 4E) illustrating that TMB and TMC are both independent predictors of ICI+CTX outcomes, and that their combination achieves the best predictive power.
[0217] While TMC and T / NK influx (INCITE) both predicted tumor shrinkage during ICI-4W, these variables showed no significant correlation, and regression analysis indicated both were independent contributors to ICI-induced tumor shrinkage. Notably, TMC strongly declined on- therapy during ICI- 4W in many patients attaining CB, indicating that persistence of monocytes throughout treatment was not required to attain prolonged disease control on ICI+CTX. Given this decrease during ICI-4W and the knowledge that monocytes can differentiate into mature macrophage and monocyte-derived conventional dendritic cell (eDC) populations (Coillard and Segura, 2019), it was hypothesized that intratumoral monocytes could be stimulated by ICI to differentiate into mature inflammatory myeloid effectors that could enhance immune responses to ICI+CTX. Indeed, trajectory analysis of myeloid cells from the LUD2015-005 scRNA-seq dataset indicated that monocytes represented an intermediate population that could transition into several macrophage and dendritic cell populations, with particular proximity to eDCs (especially cDC2), differentiating Ml -like macrophages, and other tumor-associated macrophage (TAM) subsets.
[0218] Whether ICI generated differing myeloid phenotypes in TMC-high and TMC-low patients was assessed. Intratumoral myeloid cells at the PreTx timepoint in both TMC-high and TMC-low subgroups primarily consisted of a general TAM phenotype. After ICI-4W, the majority of myeloid cells from TMC-low tumors still exhibited TAM and M2-like macrophage phenotypes; contrastingly, in TMC-high patients, ICI strongly shifted the phenotype of myeloid cells, with the majority of myeloid cells showing Ml -like macrophage, cDCl, cDC2, and LAMP3-high mature eDC phenotypes at ICI-4W. Since scRNA-seq was only available from a subset of LUD2015-005 patients, deconvolution results were examined to assess whether this result was across the entire cohort. Indeed, deconvolution results confirmed that ICI-4W biopsies from TMC-high patients showed significantly increased levels of inflammatory myeloid effectors (eDCs and Ml -like macrophages) and decreased levels of TAM / M2-like macrophages than in TMC-low patients, while the same comparison in PreTx biopsies showed no significant differences in these cell types (Table 12). These data show that PreTx TMC-high patients generate a significantly more pro- inflammatory myeloid phenotype following four weeks of ICI-only. In combination with the trajectory analysis, this finding indicates that monocytes likely serve as an intratumoral progenitor pool that can be polarized by ICI towards inflammatory myeloid effectors, thereby driving the association of TMC with ICI-induced tumor shrinkage and improved long-term outcomes on ICI+CTX.
[0219] Single-cell transcriptomic analysis reveals cell type-specific differential expression markers associated with CB
[0220] Interactions between T cells and malignant cells are central to determining response and resistance to ICI-containing regimens. The detailed information present in scRNA-seq from PreTx tumor biopsies of eight inoperable EAC patients from LUD2015-005 (4 CB, 3 NCB, one excluded from CB determination) was used to investigate whether cell-type specific expression of any gene was associated with outcomes using a pseudobulk differential expression approach to prioritize cell-type specific DEGs shared between multiple patients with CB or NCB and identified 43 DEGs (FDR < 0.1). The most significant DEG enriched in EAC cancer cells from CB patients was IGFBP2, which was highly expressed in EAC cells from all patients with CB, but only rarely detectable in EAC cells from all three patients with NCB. The patient- specific expression pattern for IGFBP2 was markedly different in cells from the stromal compartment than for epithelium, demonstrating the power of scRNA-seq in identifying cell type-specific DEGs which otherwise could be masked by traditional bulk transcriptomic approaches.
[0221] Analysis of T cells from PreTx EAC found 15 DEGs (FDR < 0.1) between T cells from CB and NCB patients, most of which were associated with NCB. NCB-associated DEGs were largely interferon-stimulated genes (ISGs), including OAS1,MX1, IFI6, an&XAFl. Notably, these ISG-high NCB patients showed minimal INCITE- reactivity during the ICI-4W window, with all three having among the four lowest INCITE upregulation scores. T-cell specific ISG expression could therefore be a marker of an intratumoral immune state that is not primed to generate effective anti-tumor immune responses to ICI+CTX. Similar to IGFBP2, some ISGs showed different expression patterns in lymphocytes than in other compartments, which could prevent the detection of this phenomenon in bulk transcriptomic approaches. In this cohort, integrative scRNA-seq approaches, including both cell composition estimation and cell-type specific expression, showed significant power to reveal additional classes of biomarkers with high predictive power.
[0222] Discussion
[0223] This example reports a comprehensive biomarker discovery study based on the unique design of the phase I / II LUD2015-005 study, which treated 38 inoperable esophageal cancer patients with ICI alone for four weeks prior to combined ICI+CTX. Results showed that monitoring on-treatment changes in a T-cell inflammation signature (INCITE) during this four- week ICI-only window yields key information for assessment of ICI sensitivity, and that tumor mutational burden and tumor monocyte content (TMB+TMC) are strong predictive biomarkers for long-term ICI+CTX outcomes.
[0224] The ICI-4W window in this trial served as proof-of-concept that four weeks of first-line ICI treatment is sufficient to induce intratumoral T cell responses in this setting, and afforded a unique opportunity to assess ICI immune response without confounding effects of CTX. Early on- treatment changes associated with ICI response can be measured by a T cell inflammation signature (INCITE) identified in this study. While no differences were seen with this inflammatory signature in pre-treatment biopsies, both upregulation of INCITE genes during the first four weeks of ICI and INCITE expression at the ICI-4W timepoint were significantly associated with greater ICI-induced tumor shrinkage during the ICI-only phase of therapy. In an independent melanoma cohort (Riaz et al. , 2017, Cell 171, 934-949. el5), INCITE upregulation during the first four weeks of aPD-1 ICI was also associated with overall outcomes. These findings indicate that early on- treatment sampling is useful in establishing the likely overall ICI efficacy for a given patient, and that applicability of this strategy is not limited to EAC.
[0225] Many CTX agents are known to cause neutropenia and other forms of myelodepletion, but they can also deplete and impair the functionality of B and T cells, possibly limiting the full potential of adaptive immune responses to ICI. Hence, an initial ICI-only window can provide an opportunity to induce antitumoral immune responses in ICI-sensitive patients using the full complement of the immune system, potentially generating stronger antitumoral responses which could better withstand the negative effects of CTX. The results from this study indicate that a short window of ICI-only treatment is sufficient and potentially desirable to prime antitumoral immune responses, and is a unique opportunity to assess whether a patient’s immune system is fit-for- purpose to respond to ICI agents. These findings are particularly important given the expanding interest in ICI-only as a treatment option for selected gastro-esophageal cancer patients.
[0226] PD-L1 expression and TMB are commonly assessed predictive biomarkers for aPD-1 ICI. In this study, PD-L1 expression was assessed by RNA-seq, which did not show a significant association with OS. The performance of histological PD-L1 CPS assays used in clinical practice for aPD-1 ICI was not assessed in LUD2015-005. While RNA- seq and histological methods for PD-L1 quantification are highly correlated in other settings, the absence of CPS quantification limits the ability to directly compare this result from LUD2015-005 with CPS results from aPD-1 ICI in this setting. TMB quantification is FDA-approved as a companion diagnostic marker for ICI-based therapy, however, questions remain regarding its clinical application, including whether TMB is an equally useful biomarker across cancer types. In gastro-esophageal cancers, TMB alone shows varying utility for ICI-containing regimens. It therefore remains a priority to clarify the utility of this well-studied biomarker in recently approved ICI+CTX indications for gastroesophageal cancer, and to discover additional biomarkers that could complement its predictive power. High TMB was associated with better OS in LUD2015-005, and its predictive power was significantly enhanced when combined with TMC, an independent predictor identified in this study through deconvolution of bulk transcriptomes using the LUD2015-005 cell atlas. While TMC is an NGS-derived estimate of monocyte content, validation experiments confirmed a significant association between TMC and ground- truth monocyte-derived RNA levels in EAC RNA-seq libraries. Notably, TMC-high patients included all those attaining OS >18 months in this study, indicating TMC could be a particularly useful marker of durable benefit on ICI+CTX. TMB assessment by targeted sequencing panels rather than whole-exome sequencing has improved access to this otherwise expensive biomarker.
[0227] The finding that TMB and TMC are specific biomarkers for ICI-containing regimens but not predictive of OS in CTX-treated advanced EAC cohorts from the TCGA and ICGC makes this biomarker combination an ideal tool to select EAC patients most likely to benefit from the addition of ICI to conventional CTX. While CTX largely targets cancer cells directly, ICI+CTX targets both tumor and immune cells in the tumor milieu to generate a response. Therefore, the utility of TMB and TMC in predicting outcomes on ICI+CTX combinations can derive from the ability of this biomarker pair to simultaneously evaluate features of both cancer cells and their associated microenvironment that are likely to respond well to treatment.
[0228] The finding that TMC is a predictor of improved outcomes with ICI-based therapies in both EAC and EBV- / MSS GC shows utility for both cancer types, but also highlights the importance of molecular subtyping in biomarker discovery, particularly for GC. EAC tumors have high molecular similarity to the chromosomal instability (CIN) subtype, the most prevalent form of GC, but are distinct from EBV-associated (EBV) and microsatellite instability (MSI) subtypes, which are rare or absent in EAC (The Cancer Genome Atlas Research Network, 2017). Previous work using a cohort of GC patients receiving ICI as second- or third-line therapy, which included EBV and MSI tumors, identified CXCL11 and PD-L1 as strongly associated with ICI response (Kim et al., 2018, Nature Medicine 24, 1449-1458; Noh et al., 2021, Experimental & Molecular Medicine 53, 223-234). This contrasts with the present work in EAC, where neither PD-L1 nor T cell inflammation markers were predictive of ICI-4W or ICI+CTX outcomes, and the predominant transcriptional biomarker was TMC. As noted by Kim et al., ICI responses in their gastric cohort were much higher in EBV and MSI GCs (among those with available exome sequencing, 100% and 100%) compared with EBV- / MSS subtypes (GS: 12%, CIN: 5%). This indicates that, in unselected GC cohorts, biomarkers predictive of ICI response are likely to be driven largely by EBV and MSI responders, potentially explaining why these biomarkers would not be shared by EAC cohorts. The higher response rates in EBV / MSI GC than other subtypes, the high transcriptional similarity between CIN GC and EAC, and the discovery that the major transcriptional predictive biomarker (TMC) in the EAC cohort also had significant utility in EBV- / MSS (but not EBV / MSI) GC indicate that EBV- / MSS GC subtypes behave more analogously to EAC than to EBV / MSI GC subtypes with respect to ICI.
[0229] While some myeloid subsets have immunosuppressive functions, mature myeloid-derived effector cell types are also key mediators of phagocytosis, antigen presentation, and antibodydependent cellular cytotoxicity in tumors. Inflammatory intratumoral macrophages (high CD68+ / CD163+ ratio) and other inflammatory tumor-associated myeloid cells have previously been associated with improved ICI response. A transcriptional signature of tissue-resident macrophages was also recently shown to predict response to aPD-1 ICI across melanoma, lung, and breast cancers. Nevertheless, monocyte-associated signatures, rather than those of fully differentiated inflammatory macrophages, can be a strong ICI+CTX-specific pre-treatment biomarker.
[0230] Although PreTx TMC was significantly associated with CB in this cohort, the decrease in median monocyte content at ICI-4W in CB patients indicated that monocytes were unlikely to generate improved outcomes through directly carrying out sustained anti-tumor activity; rather, ICI may drive differentiation of intratumoral monocytes into pro-inflammatory myeloid effectors responsible for this association with improved outcomes. While some differentiated myeloid cells present in solid tissues can represent long-lived tissue-resident populations derived from alternative progenitors, in the presence of cancer-associated inflammation, circulating monocytes are recruited into tumors where they can differentiate into various intratumoral macrophage and DC subsets in response to different stimuli. Monocyte-derived TAMs can exist across a continuum of Ml -like and M2-like states. While the predominant monocyte-derived intratumoral DC population is an inflammatory population resembling resident cDC2s, other DC states, including a mature LAMP3-high population arising in response to inflammatory stimuli, can also be derived from monocytes. Monocytes could therefore serve as an intratumoral progenitor pool for a variety of differentiated macrophage and DC subsets, agreeing with the myeloid trajectory analysis on LUD2015-005 scRNA-seq data, which showed monocytes existing along a continuum connecting eDCs (particularly cDC2-like cells), Ml -like macrophages, and other TAM phenotypes. Single-cell trajectory analysis and bulk RNA-seq deconvolution revealed that, while TMC- low tumors retained a higher TAM / M2-like myeloid phenotype at ICI-4W, TMC-high tumors developed a more pro-inflammatory myeloid phenotype in response to ICI, with greater levels of inflammatory myeloid effectors at ICI-4W. The pro-inflammatory myeloid skew generated at ICI- 4W in TMC-high patients consisted of higher levels of Ml -like macrophages, cDCls, cDC2s, and mature LAMP3-high DCs. These differentiated effectors have a number of features that contribute to improved anti -tumor responses on ICI+CTX. Ml -polarized macrophages have an improved ability to lyse internalized cells following phagocytosis, contributing to improved tumor control. Ml -like macrophages in the LUD2015-005 dataset also showed increased expression of secreted proteins, which can amplify anti-tumoral inflammatory responses, including IL1A / IL1B and TNF. The listed DC subsets play a key role in activating T cells by presenting antigens and providing costimulatory signals. XCR1+ cDCls are known to play a key role in cross-presenting tumor antigens on MHC I to activate antitumoral cytotoxic T cells, and mature LAMP3-high DCs in the dataset had the highest expression of CD80 / CD86 costimulatory signals needed to activate T cell responses fully. cDC2 cells are more phenotypically and functionally heterogeneous, but also play an important role in directing CD4 responses in tumors.
[0231] Therefore, the ICI-induced shift in myeloid phenotype in tumors with high PreTx monocyte content amplifies anti -tumoral responses to therapy through both direct and T-cell mediated mechanisms. The inflammatory myeloid effectors generated by ICI in TMC-high patients act together with ICI-induced lymphocyte infiltration to drive tumor shrinkage during ICI-only, and the improved ICI+CTX outcomes are observed here.
[0232] Finally, the power of scRNA-seq to identify cell type-specific predictive biomarkers was demonstrated. Although patient numbers are limited (n=8), the specificity of scRNA-seq enabled the discovery that DEGs would be masked in bulk transcriptomics due to varying expression patterns across cellular compartments. In PreTx EAC cells, IGFBP2 was most significantly associated with CB. Tumor-specific expression o IGFBP2 in melanoma has already been reported to be associated with an ICI-favorable immune environment, providing evidence that this method can successfully uncover meaningful biological signals in cell-type specific contexts. In T cells, high expression of ISGs at PreTx was found in NCB patients, showing a lack of inflammatory responses to ICI (INCITE-quiescence). While ISGs can mediate important immune responses, their aberrant overexpression in T cells is linked with T cell dysfunction and death during severe viral infection, including HIV and SARS-COV-2, mediated through ISGs such as the pro-apoptotic XAF1. Therefore, T cell-specific ISG expression can serve as a marker of dysfunctional or apoptosis-prone T cells unlikely to respond to ICI, agreeing with the low INCITE reactivity seen in these patients. These findings show how scRNA-seq can reveal additional predictive biomarkers and biological insight.
[0233] EXAMPLE 3
[0234] Bulk RNA sequencing
[0235] Paired endoscopic biopsies were collected, and of each pair, one biopsy was slow-frozen, and one snap-frozen. Snap-frozen biopsies were cut in half on dry ice, RNA was extracted from one half, and one half was embedded in OCT for histology. After frozen tissue was dissolved with a pestle, followed by addition of Mirvana lysis / binding buffer, Acid-Phenol: Chloroform RNA extraction was performed using the Mirvana RNA isolation kit (ThermoFisher, AMI 560) as per manufacturer’s instructions. DNAse was removed using the Turbo DNA-free kit (ThermoFisher, AMI 907), and DNA quantity and quality were assessed with Nanodrop (ThermoFisher) and 2100 BioAnalyzer (Agilent) using the RNA Pico Kit (Agilent, 5067-1513), respectively.
[0236] Single cell RNA sequencing
[0237] Single cell sequencing was performed on 48 samples from 16 patients, representing 13 patients with OAC, and 3 patients with OSCC. These 16 patients were selected, as a sufficient number of slow-frozen biopsies was available for all timepoints.
[0238] Tissue dissociation
[0239] Endoscopic biopsies were slow-frozen on-site in human serum with 10% DMSO or, in a few instances, in FBS with 10% DMSO where human serum was unavailable. The tissue was stored at -80°C until further use.
[0240] Slow-frozen vials were thawed by gentle agitation of the cryovial in a 37°C water bath. Tissue dissociation was performed as described above. Briefly, two biopsies per patient and timepoint were selected and transferred into a microcentrifuge containing PBS with sterile forceps to wash off any excess mucus and DMSO remnants. To increase the surface area, biopsies were snipped vigorously with sterile scissors and dissociated in 5ml dissociation media (see Table 15) at 37°C with constant rotation for 45-60 minutes. After vigorous pipette mixing, the dissociation solution was passed through a pre-wetted 100pm CellTrics filter (Sysmex, 04-004-2328) into a 15mL Falcon tube. Any remaining tissue fragments were washed with 6mL quench buffer to a final volume of ImL. Cells were counted using a hemocytometer, and viability was assessed using an automated cell counter (BioRad, TC20). Cells were pelleted at 300g in a pre-chilled centrifuge for 5 minutes to remove the dissociation media and quench buffer. Cell pellets were resuspended in pre-chilled resuspension buffer to the final concentration. Encapsulation was performed as soon as possible to maintain cell viability, and the single cell solution was kept on ice.
[0241] During optimization, enrichment for immune cells with MACS CD3 and CD 19 MicroBeads (Miltenyi, 130-050-100 and 130-050-300) was tested, but significant loss of viable epithelial cells compared to non-separated samples was observed upon quantification with flow cytometry, presumably because epithelial cells adhered to the MACS filter. Viability was also adversely affected.
[0242] Table 15. Composition of dissociation reagents. Reagents were prepared fresh.
[0243] Encapsulation, library preparation, and sequencing
[0244] Single cell sequencing was performed using the 10X chromium platform and the 10X chromium controller. The single cell emulsion was prepared using the 5’ paired-end Next GEM Single Cell 5’ v2 kit (1000286). BCR and TCR sequences were amplified using the Chromium Single Cell Human TCR (100252) and BCR (1000253) Amplification kits, and library prep was performed with the Chromium Library Construction kit (1000190). Single cell sequencing was performed according to manufacturer’s instruction, with the exception of the concentration of the single cell suspension. Due to poor viability, more cells than recommended (median 26400, range 16500-33000 compared to recommended 16000 cells) were loaded to increase cell recovery. Fragment size of prepared libraries was quantified with on-chip electrophoresis (Agilent 2100 Bioanalyzer System or Agilent TapeStation) using the High Sensitivity DNA Kit (Agilent, 5067- 4626) or High Sensitivity D5000 kit (Agilent, 5067-5592 / 3) respectively. Library concentration was determined by Qubit dsDNA HS Assay kit (Invitrogen, Q32851). Libraries were sequenced on an Illumina NextSeq2000 machine, with 26 cycles for Read 1, 10 cycles each for the i5 and i7 indices, and 90 cycles for Read 2, with a final library molarity of 0.65 nM and targeting a minimum of 20,000 reads per cell. 1% PhiX (Illumina, FC-110-3001) was spiked for QC considerations. Cell numbers were predicted from the final library DNA concentration, and verified by the CellRanger cell calling algorithm, and libraries were re-sequenced if sufficient read depth had not been achieved. For some samples, poor library quality, for example, due to contamination with ambient DNA, precluded reaching 20,000 reads per cell, and additional sequencing did not improve read depth.
[0245] Bioinformatics
[0246] The majority of bioinformatic analysis was conducted on the BMRC computation cluster (supported by the Wellcome Trust Core Award Grant Number 203141 / Z / 16 / Z with funding from the NIHR Oxford BRC).
[0247] Single cell RNA sequencing
[0248] Single cell data was processed using the Cell Ranger pipeline (v7.1.0) from lOx Genomics. Fastq files generated previously on samples from the metastatic cohort of the LUD2015-005 trial were re-analyzed as they had been processed with an older version of Cell Ranger that is not compatible with the current version. Briefly, raw BCL files were demultiplexed and converted into FASTQ files using cellranger mkfastq. A custom reference sequences was prepared as described above, from the CRCh38 human genome (ENSEMBL100) and supplementary contigs from the Genomic Data Commons alignment reference, containing the decoy contig hs38.dl and 200 viral genomes commonly found in human cancer samples. This reference sequence was generated with Cell Ranger v3.1.0 to ensure compatibility with previous analysis conducted on samples of the LUD2015-005 trial, using the cellranger mkref command and the gencode v36 annotation file supplied by Genomic Data Commons. Alignment was performed with cellranger count, specifying an expected cell recovery of 8000 cells, which yields a count matrix per sample. BCR and TCR reads were processed using the cellranger vdj command. Samples from the operable cohorts were integrated with samples from the metastatic cohorts. As these were encapsulated using different lOx chemistries (5’ v.1.1 and v.2), UMI count matrices were integrated with cellranger aggr without library normalization, but with chemistry correction. The limit of cell numbers precluded integrating all samples together, so samples were randomly ordered and processed in four batches. The presence of batch effects between chemistries with excluded following analysis with kBET (vO.99.6).
[0249] Downstream analysis was conducted using Seurat (v4.1.1) as described above. Low- quality cells with less than 200 genes and over 25% mitochondrial reads were removed. The full dataset was normalized using SCTransform (v0.3.3), following UMAP dimensionality reduction (umap-learn v0.5.3). Clusters were identified using the default Louvain algorithm at a resolution of 1.2. Broad cell types were separated into separate Seurat objects. Some biopsies from the metastatic cohort were processed fresh, and the only systematic batch effect was observed between fresh and frozen samples. SCTransform was therefore conducted separately on fresh and frozen samples. The batch effect was accounted for using the LastMNN batch correction method (SeuratWrappers v0.3.1), before re-integration and dimensionality reduction. After clustering, markers were identified using the LindAllMarkers function to manually assign cell type and cell subtype labels.
[0250] Ligures visualizing the scRNA-seq findings were generated with Seurat, ggplot2 (v.3.4.3), and pheatmap (vl.0.12). Trajectory analysis on sub-compartments was performed using monocle (v2.26.0), monocle 3 (vl.3.4), and slingshot (v2.6.0). The detection of intronic reads with the newer version of Cell Ranger was capitalized to perform RNA velocity analysis, which detects gene expression dynamics using the ratio of spliced to unspliced reads of individual genes. As intronic reads are discarded by cellranger count, alignment was additionally performed with velocyto (vO.17.17), aligning to the same reference genome, to generate loom files containing both spliced and unspliced reads. These were aggregated into a SingleCellExperiment (vl.20.1). The previously generated Seurat object of the sub-compartment of interest was also converted into a SingleCellExperiment. The SingleCellExperiment newly generated from loom files was a subset for high-quality cells belonging to the sub-compartment of interest, and spliced and unspliced reads were added as additional assays to the myeloid SingleCellExperiment and RNA velocity analysis was conducted with velociraptor (vl.8), an R wrapper for scvelo.
[0251] Immune enrichment analysis was performed using the immune dictionary (Cui, A., et al. Nature 625, 377-384 (2024)).
[0252] Bulk RNA-sequencing
[0253] Fastq files were downloaded from the sequencing provider. Initially, adapters were removed with cutadapt(v2.9) and TrimGalore (vO.6.5), and reads were aligned to the same custom reference genome used in scRNA-seq processing using STAR (v2.7.10b, (index generated with STAR v2.7.5)). BAM files were sorted with samtools (vl.11.0) Count matrices were prepared for each sample using featureCounts (subread v2.0.6). Sequencing quality metrics were generated using FastQC (v0.11.8), MultiQC (vl.8), and Picard tools (GATK v4.1.7.0) CollectRnaSeqMetrics and CollectMultipleMetrics modules, including assessment of rRNA depletion efficacy by quantifying reads mapped to rRNA regions (both nuclear and mitochondrial rRNA, as well as rRNA pseudogenes).
[0254] Differential expression was performed with DESeq2 (vl.38.3) and a significance threshold of FDR < 0.1. Default arguments were used, with the exception of the use of the “local” fit type to model dispersion estimates. Unless otherwise specified, Wald tests were specified. Moderated gene-level log-fold changes were calculated using the “ashr” method. These values were the test statistic used for gene set enrichment analysis and immune response enrichment analysis (Cui, A., et al. Nature 625, 377-384 (2024)). When comparing the expression of individual genes between groups, logio-transformed (off-set 0.1) transcripts per million (tpm), normalized for gene lengths, were calculated. When comparing the expression of sets of genes, for example, for heatmap visualization, or gene score calculations, variance stabilized counts were generated with the varianceStabilizingTransformation function of DESeq2, which were then Z-scaled. Unless otherwise specified, gene scores were calculated from the means of Z-scaled, vst-transformed counts.
[0255] Deconvolution
[0256] Deconvolution was performed on bulk transcriptomic data to estimate the proportional celltype composition of biopsies. A benchmarking study of pseudobulk samples generated from single data of the metastatic cohorts of the LUD2015-005 trial revealed that BayesPrism best recreated true cell composition so BayesPrism was used to deconvolute samples from operable LUD2105- 005 cohorts. Deconvolution was performed as described above, but using the single cell dataset comprising all four cohorts of the LUD2015-005 study as input reference. At the time that deconvolution was performed for the metastatic cohorts, BayesPrism was available as a pre-print, and the software was distributed as the package TED. Since then, BayesPrism has been published (Chu, T., et al. Nat Cancer 3, 505-517 (2022), distributed and maintained as the package BayesPrism (v.2.0). Deconvolution was performed with both packages to assess consistency. Briefly, the 3000 most variable genes were extracted for all high-quality single cells and stored as a reference matrix. Count matrices were generated separately for all OSCC and OAC bulk transcriptomes, which were normalized and scaled using SCTransform. Deconvolution was performed using default settings, specifying either squamous or non-squamous tumor cells as tumor reference, respectively.
[0257] Survival analysis
[0258] The survivalAnalysis (v0.3.0) package in R was used to assess the link between variables and survival outcomes. Kaplan-Meier plots were generated from univariate Cox regressions calculated with analyze_survival, while multivariate Cox regressions and forests plots were generated with analyze_multivariate. Unless otherwise specified, continuous variables were scaled and centred prior to regression.
[0259] Results
[0260] Tumor monocyte content predicts outcomes from immune checkpoint inhibitor immunotherapy in operable esophageal cancer
[0261] To investigate whether the tumor microenvironment composition affects the response of operable esophageal cancers to immunotherapy, bulk transcriptomic data generated from the operable samples of the LUD2015-005 samples was deconvoluted with the BayesPrism package, using annotated single cell RNAseq (scRNAseq) data as a reference. Of all cell types, tumor monocyte content (TMC) at baseline was most significantly associated with the outcome from immunotherapy, significantly predicting survival by different statistical metrics, with a trend of higher TMC in patients with the major pathological response (mPR) at Surgery (Figure 5). 5- and 10-monocyte gene signatures predict outcomes from immunotherapy
[0262] A monocyte gene signature was created to simplify the complex bioinformatic process of bulk transcriptomic deconvolution. As the scRNAseq reference matrix was subset to the top 3000 most variable genes for deconvolution, the scRNAseq dataset was also subset for these 3000 genes to extract monocyte markers. After subsetting, monocyte markers were calculated using the FindMarkers function, comparing cells annotated as monocytes to all other cells (including stroma and epithelial cells). The top 30 most differentially expressed markers were taken forward (Figure 6A).
[0263] Using the caret (version 6.0.94) and leaps (version 3.1) packages and defining TMC by deconvolution as the output metric, best subset regression analysis was conducted on these 30 genes to determine the optimal number of genes contributing to a gene signature representing TMC. Assessment of goodness of fit of regression models with Mallow’s Cp and the Bayesian Information Criterion suggests that the optimal number of variables is 10 (Figures 6B-C). A linear model was generated for these 10 variables (Figure 6D)
[0264] The 10-gene monocyte marker signature can equally split the operable cohort of the LUD2015-005 trial into patients likely to respond and less likely to respond (Figure 7B). The expression score is significantly higher in patients alive at the end of the study (Figure 7A), and patients with mPR at surgery also show a trend of a higher expression score (Figure 7A), suggesting that the 10-gene monocyte marker signature recreates the monocyte signature generated from bulk transcriptomic deconvolution.
[0265] To further simplify the gene expression signature, coefficients from the linear regression were removed, and only variables with positive coefficients were taken forward, yielding a 5-gene monocyte signature composed of IFI30, FCN1, HLA-DRA, CXCL8, and SLC11A1. The expression score generated from the sum of z-scaled, vst-stabilized expression, outperformed both the 10 gene signature and monocyte deconvolution results in terms of differential expression between patients alive and dead at the end of the study and between patients with and without mPR at surgery (Figure 7C). Splitting the cohort by the median expression of the 5-gene expression signature can also split the cohorts into patients significantly more likely to achieve long survival (Figure 7D). The same analysis was repeated for the metastatic cohorts of the LUD2015-005 trial. While expression of the 10-gene signature was similar in patients who were dead or alive at the end of the study, and between patients with and without clinical benefit (progression-free survival > one year, Figure 8 A), splitting the cohort by expression median separated patients into groups likely and less likely to attain long OS (Figure 8B). Regarding the 5 gene expression scores, expression trended higher in patients with clinical benefit, approaching significance (Figure 8C), and classification by expression median also trended towards significance (Figure 8D). Importantly, this outperforms the discriminatory potential of TMC by deconvolution between patients with and without clinical benefit (Figure 9C) and cumulative survival (Figure 9D). Notably, the difference in predictive value in PreTx TMC in the LUD2015-005 metastatic cohorts here, compared to what the results above can be attributed to the transcriptional similarity of monocytes and neutrophils, which are more profoundly similar in the scRNAseq dataset comprising samples from all four cohorts of the LUD2015-005 cohorts, compared to the scRNAseq dataset generated from the metastatic cohorts only. While PreTx neutrophil content association with overall survival is significant in a forest analysis (HR = 0.63 (CI: 0.40-0.99), p = 0.045) in metastatic OC patients, the association is much weaker than in the operable OC cohort (HR = 0.27 (CI: 0.08-0.89), p = 0.031), although TMB for the operable OC cohorts is currently unavailable, so this is the result from univariate analysis.
[0266] External validation in an independent pooled cohort of gastric cancer patients treated with ICI + chemotherapy showed no difference in expression of the 10-gene monocyte signature (Figures 10A-B), but a profound and statistically significant difference in expression of the 5 gene signature (Figures 10C-D). As observed with TMC above, PreTx TMC was only higher in responders in EBV negative and microsatellite stable (MSS) gastric cancer patients, but not EBV positive or microsatellite instable (MSI) patients (Figure 10D).
[0267] In conclusion, this data shows that PreTx TMC predicts outcomes of operable OC patients in the LUD2015-005 trial, with the survival advantage being perhaps even more profound than previously observed in the metastatic cohort. Further, expression of 5 genes, IFI30, FCN1, HLA- DRA, CXCL8, and SLC11A1, can recreate the signature observed through bulk transcriptomic deconvolution.
[0268] Expression of monocyte signature genes in myeloid compartment All genes contributing to the 5-gene and 10-gene monocyte signatures are derived from monocyte marker genes, i.e., highly enriched in monocytes compared to any other cell type according to scRNAseq. Gene expression was plotted in this compartment to assess how specific monocyte signature genes are within the myeloid compartment, both on a cell subtype level and per cell across the whole compartment. Some genes are relatively widely expressed, and have maximum expression in a non-monocyte population. Still, these are the genes with negative coefficients in the 10-gene linear model (e.g., EREG with maximum expression in MDSC, PLAUR with maximum expression in CXCLlhi Ml -like macrophages, or HLA-DRB1 with maximum expression in cDC2 dendritic cells). Neutrophils and monocytes are transcriptionally similar, which can also be observed for the 10-gene signature genes. Still, the expression is more distinct for the 5-gene signature genes, providing an explanation for the superior ability to predict the outcomes.
[0269] The present disclosure is not to be limited in scope by the specific embodiments described herein. Indeed, various modifications of the invention, in addition to those described herein, will become apparent to those skilled in the art from the foregoing description and the accompanying figures. Such modifications are intended to fall within the scope of the appended claims.
[0270] Table 1. Clinical characteristics of the inoperable LUD2015-005 cohorts
[0271] The characteristics at enrolment of the 38 patients with inoperable esophageal cancer who received treatment on the LUD2015-005 study are presented and categorized by the ICI agent used. ECOG refers to the patient performance status: 0 is fully active, while 1 is ambulatory but restricted in physically strenuous activity.
[0272] Table 2. Treatment-emergent adverse events during the LUD2015-005 study.
[0273] Top: A summary of all treatment-emergent adverse events (TEAEs) reported for the intent- to-treat (ITT) population of the inoperable LUD2015-005 trial cohorts. All TEAEs are shown, regardless of the degree of investigator-assessed likelihood that the TEAE was related to study drugs. Durvalumab-related TEAEs are defined as any degree of investigator-assessed likelihood of relationship to durvalumab. Severity of TEAEs was graded using Common Terminology Criteria for Adverse Events (CTCAE; v4.03). Bottom: Reporting of dose-limiting toxi cities (DLTs) in the dose escalation phase of the LUD2015-005 study. The pre-specified measurement window for this primary outcome was less than 10 weeks following the first dose of medication, *1 of 14 patients in the extension phase of the aPD-Ll+75mg aCTLA-4 cohort experienced an adverse event meeting DLT criteria.
[0274] (a) Treatment-emergent adverse events in ITT population:
[0275] DLTs encountered during dose escalation phase:
[0276] Table 3. Treatment-related adverse events reported in 10% or more of patients.
[0277] TEAEs were categorized using CTCAE v4.03, as in Table 2. The frequency of TEAEs meeting a 10% cut-off across all patients in the ITT population is reported by their preferred term, which is categorized by system organ classes (bold).
[0278] Table 4. Clinical outcomes of the inoperable LUD2015-005 cohorts. irRECIST BOR and CB outcomes and survival summary statistics are presented for the full ITT population (top, n = 38) and the 35 ITT adenocarcinomas (bottom). Each table shows outcomes for the full population and each ICI treatment subgroup. For survival outcomes, NA means that the upper end of the confidence interval was not yet reached due to ongoing patient survival. See Methods for full details of BOR and CB classifications.
[0279] Full ITT Population (n = 38):
[0280] EAC only (n = 35):
[0281] Table 5. Key differential expression results from bulk RNA-sequencing.
[0282] DESeq2 outputs are provided following log fold change moderation (‘ashr’ method). Results are listed as follows: Categorical comparison of ICI-4W vs. PreTx EAC samples (Sheet 1, INCITE genes are labeled, positive log2FoldChange enriched in ICI-4W); all 5 PreTx EAC samples, using the change in tumor size during ICI-4W as a continuous variable in the design formula (Sheet 2, negative log2FoldChange linked with tumor shrinkage); all ICI-4W EAC samples, using the change in tumor size during ICI-4W as a continuous variable in the design formula (Sheet 3, negative log2FoldChange linked with tumor shrinkage); categorical comparison of CB vs. NCB from PreTx EAC samples (Sheet 4, positive log2FoldChange enriched in CB). Significant DEGs with adjusted p-value < 0.1 are bolded. For sheet 5, PD-L1 (CD274) loglO-transformed TPMs (0.1 offset) were Z-score normalized and subjected to Cox regression with OS io and PFS, and summary statistics are provided. For sheet 6, Pearson and Spearman correlation coefficients between the change in VST- normalized gene expression during the ICI-4W window (ICI-4W - PreTx values) and tumor shrinkage during the ICI-4W window (negative coefficients linked with tumor shrinkage). Associated p and FDR values for these correlation tests are also provided.
[0283] Pearson coefficients were used as the GSEA statistic for Figure 2C.
[0284]
[0285]
[0286]
[0287]
[0288]
[0289]
[0290]
[0291]
[0292]
[0293]
[0294]
[0295]
[0296]
[0297]
[0298]
[0299]
[0300]
[0301]
[0302]
[0303] Table 6. Performance of INCITE-12 against other inflammatory gene signatures.
[0304] Top: Correlation test p-values (Pearson and Spearman) are provided for the association between upregulation of listed gene sets during ICI-4W (calculated as the sum of scaled VST- normalized counts) and target lesion shrinkage during ICI-4W. Middle: Significance of association between upregulation of listed gene sets and long-term ICI-only outcomes in a cohort of ipilimumab-naive melanoma patients treated with aPD-1 ICI (Riaz et al., 2017). P-values are calculated by non- parametric tests (Kruskal -Wallis for 3 -way response; Mann- Whitney U test for 2-way benefit comparison). Bottom: as above, but for a cohort of melanoma patients treated with aPD-1 ICI who had previously progressed on ipilimumab from the same publication (Riaz et al., 2017, Cell 171, 934-949.el5).
[0305] Table 7. Extended results of deconvolution algorithm assay benchmarking.
[0306] Cell type-specific results for the benchmarking pseudobulk experiment are provided for each of the five deconvolution algorithms in separate tabs. Results are shown for two input reference matrices: The full scRNA-seq reference matrix, and a reference matrix subset to just the 3000 most variable genes (using Seurat’s FindVariableFeatures). For each of these conditions, results of each 5 algorithm were compared to both the known RNA proportion (i.e., the fraction of counts assigned to each cell type) and the known cell proportion (i.e., the number of cells from each cell type divided by the total number of cells) of each pseudobulk sample. Spearman correlation and normalized RMSE (RMSE / range of predicted values) are shown. The run time of each algorithm using each input reference matrix is also displayed.
[0307]
[0308]
[0309]
[0310]
[0311]
[0312]
[0313]
[0314]
[0315]
[0316]
[0317]
[0318]
[0319]
[0320]
[0321]
[0322]
[0323]
[0324]
[0325]
[0326]
[0327]
[0328]
[0329]
[0330]
[0331]
[0332]
[0333]
[0334]
[0335]
[0336]
[0337] Table 8. BayesPrism deconvolution correctly identifies epithelial type in closely related esophageal pathologies.
[0338] The mean proportion of reads assigned to each of the indicated cell type signature (rows) for each of the tissue types (columns) in the Maag el al. dataset are indicated. BE: Barrett’s esophagus; LGD: low-grade dysplasia.
[0339] Table 9. Single-timepoint cell composition as a biomarker for OS
[0340] Sheet 1 : The proportion of each of the 45 major cell types estimated by BayesPrism were used in Cox regressions with overall survival (OS). Results of each cell-type regression are provided from the PreTx timepoint. Sheet 2: Full BayesPrism deconvolution output for all LUD2015-005 inoperable EAC biopsies (3000 gene reference matrix). Deconvolution results are reported as the proportion
[0341] 5 of RNA (out of 1) contributed by each cell type.
[0342]
[0343] Table 10. Summary of reported pharmacological agents in TCGA-ESCA and ESAD-UK (related to Figure 4 and Figure 5) Sheet 1 : Pharmaceutical treatment information, available only for 43 of 185 cases in the TCGA- ESCA cohort, was downloaded from GDC. Drug names were cleaned and standardized, and are reported for all patients with available information. For each patient, a one represents that the drug was reported to be taken at least once, while a zero represents that the drug was not reported for that patient.
[0344] 5 The Total Count column shows drug frequency across the TCGA- ESCA cohort. Sheet 2: Four fields reporting treatment agents are collated for each patient from specimen- and donor-level entries from the ICGC Data Portal (J. Zhang el al., 2019). A merged summary field is provided below to summarize the treatment type(s) reported for each patient.
[0345]
[0346]
[0347]
[0348] Table 11. Summary of patient metadata and key biomarker results
[0349] Patient demographics, treatment information, and clinical outcomes are displayed for each patient identifier, alongside selected biomarker results from this study. For pre-treatment values, tumor mutational burden is presented in coding mutations per Megabase,
[0350] 5 MSIsensor score reports the percentage of MSI somatic sites as reported by the MSIsensor algorithm, and tumor monocyte content is the raw output of BayesPrism (3000 gene reference matrix) showing the fraction of tumor RNA predicted to be derived from monocytes. For metrics derived from changes during ICI-4W, INCITE change is the normalized average of change in all INCITE genes between PreTx and ICI-4W (scaled without centering, positive values represent upregulation at ICI-4W), while tumor shrinkage shows the change in target lesion burden between PreTx and C1D8 scans as a percentage of the PreTx value.
[0351]
[0352]
[0353]
[0354]
[0355]
[0356]
[0357] Table 12. ICI-4W cell composition of patients with TMC-high and TMC-low PreTx phenotypes
[0358] Patients were split into tumor monocyte content (TMC)-high and TMC-low based on the PreTx median of the subgroup of EAC patients with available PreTx and ICI-4W biopsies (n = 28).
[0359] The value at ICI-4W for each cell type were compared between PreTx monocyte-high and PreTx monocyte-low patients, p-values were calculated using a Mann- Whitney U-test. No significant results were found following FDR adjustment (FDR < 0.1).
[0360] Table 13. Details of supplemental operable EAC patients in the LUD2015-005 scRNA- seq dataset.
[0361] Treatment and outcome details of the three patients from the operable arms of the LUD- 2015-005 trial (not reported in this paper) whose samples were used for scRNA-seq. FLOT: 5- fluorouracil, leucovorin, oxaliplatin, and docetaxel.
[0362] Table 14. Composition of reagents used for single cell sequencing.
[0363] These analyses of LUD2015-005 clinical and molecular profiles would not have been possible without the support of the full LUD2015-005 Project Team. This listing provides a more complete accounting of contributors to the LUD2015-005 trial.
Claims
CLAIMSWhat is claimed is:
1. A method of identifying a subject as likely responding to treatment of an immunotherapy or predicting survival of the subject undergoing the immunotherapy, comprising: determining an expression level of each of biomarkers comprising one or more of IFI30, FCN1, HLA-DRA, CXCL8, and SLC11 Al in a sample from the subject; determining a change in the expression level of each of the biomarkers as compared to a respective reference expression level for each of the biomarkers; and identifying the subject as likely responding to the treatment of the immunotherapy or as likely having prolonged survival after the treatment, if the expression level of each of the biomarkers is elevated as compared to the respective reference expression level for each of the biomarkers.
2. The method of claim 1, wherein the biomarkers comprise IFI30, FCN1, HLA-DRA, CXCL8, and SLC11A1.
3. The method of claim 1, wherein the biomarkers consist of IFI30, FCN1, HLA-DRA, CXCL8, and SLC11A1.
4. The method of any one of the preceding claims, wherein the biomarkers comprise one or more of FI30, FCN1, HLA-DRA, CXCL8, SLC11A1, TYROBP, LYZ, PLAUR, HLA-DRB1, and EREG.
5. The method of claim 4, wherein the biomarkers comprise FI30, FCN1, HLA-DRA, CXCL8, SLC11A1, TYROBP, LYZ, PLAUR, HLA-DRB1, and EREG.
6. The method of claim 4, wherein the biomarkers consist of FI30, FCN1, HLA-DRA, CXCL8, SLC11A1, TYROBP, LYZ, PLAUR, HLA-DRB1, and EREG.
7. The method of any one of the preceding claims, wherein the sample is obtained before the treatment.
8. The method of any one of the preceding claims, wherein the sample comprises tumor tissue containing the monocytes.
9. A method of identifying a subject as likely responding to treatment of an immunotherapy or predicting survival of the subject undergoing the immunotherapy, comprising: determining an expression level of each of biomarkers comprising one or more of KLRD1, KLRK1, FASLG, GZMK, KLRC1, KLRC4, CD8A, PRF1, AP0BEC3G, AP0L3, CD244, and FCRL6 in a sample from the subject; determining a change in the expression level of each of the biomarkers as compared to a respective reference expression level for each of the biomarkers; and identifying the subject as likely responding to the treatment of the immunotherapy or as likely having prolonged survival after the treatment, if the expression level of each of the biomarkers is elevated as compared to the respective reference expression level for each of the biomarkers.
10. A method of treating a cancer in a subject in need thereof, comprising: identifying a subject as likely responding to treatment of an immunotherapy or as likely having prolonged survival after the treatment according to the method of any one of the preceding claims, and administering to the subject a therapeutically effective amount of the immunotherapy.
11. The method of any one of the preceding claims, wherein the expression level of each of the biomarkers is determined from the sample of the subject who has completed a treatment cycle of a cancer therapy.
12. The method of claim 11 , wherein the respective reference expression level for each of the biomarkers is determined from samples of one or more subjects before the treatment cycle of the cancer therapy.
13. The method of claim 11, wherein the treatment cycle is about 4 weeks.
14. The method of any one of the preceding claims, wherein the immunotherapy comprises one or more immune checkpoint inhibitors.
15. The method of claim 14, wherein the one or more immune checkpoint inhibitors comprise a CTLA-4 inhibitor, a PD-1 inhibitor, a PD-L1 inhibitor a LAG-3 inhibitor, or a TIM-3 inhibitor.
16. The method of claim 14, wherein the one or more immune checkpoint inhibitors comprise an anti-CTLA-4 antibody, an anti -PD-1 antibody, an anti-PD-Ll antibody, an anti-LAG-3 antibody, an anti-HM-3 antibody, or a combination thereof.
17. The method of claim 14, wherein the one or more immune checkpoint inhibitors comprise an anti-PD-Ll antibody, an anti-CTLA-4 antibody, or a combination thereof.
18. The method of claim 14, wherein the one or more immune checkpoint inhibitors comprise durvalumab.
19. The method of any one of claims 10-18, wherein the immunotherapy is administered to the subject in one or more doses or in one or more treatment cycles.
20. The method of any one of claims 10-19, comprising administering to the subject an additional therapeutic agent or therapy.
21. The method of claim 20, wherein the additional therapeutic agent or therapy comprises a chemotherapy.
22. The method of claim 21, wherein the chemotherapy comprises oxaliplatin, capecitabine, or a combination thereof.
23. The method of any one of the preceding claims, comprising determining a therapeutic response in the subject that comprises complete or partial response as defined by response evaluation criteria in solid tumors (RECIST), stable disease as defined by RECIST, or long-term survival despite disease progression or response as defined by immune-related response criteria (irRC).
24. The method of any one of the preceding claims, wherein the sample comprises a biological sample that comprises a plasma sample, a blood sample, or a tissue sample.
25. The method of any one of the preceding claims, wherein the sample is obtained from a primary tumor or a metastasis.
26. The method of any one of the preceding claims, wherein the expression level comprises a mRNA or protein level.
27. The method of any one of the preceding claims, wherein the subject has an esophageal cancer.
28. The method of any one of the preceding claims, wherein the subject has inoperable esophageal adenocarcinoma (EAC).
29. The method of any one of claims 1-26, wherein the subject has an EBV- / MSS gastric cancer.
30. A method of identifying a subject as likely responding to treatment of a cancer therapy, comprising: determining a level of a tumor monocyte content (TMC) in a sample from the subject; determining a change in the level of the tumor monocyte content as compared to a reference level; andidentifying the subject as likely responding to the treatment of the cancer therapy if the level of the tumor monocyte content is elevated as compared to the reference level, wherein the cancer therapy comprises an immunotherapy and a chemotherapy.
31. A method of identifying a subject as likely responding to treatment of a cancer therapy, comprising: determining a level of a tumor monocyte content (TMC) and a level of tumor mutational burden (TMB) in a sample from the subject; determining a first change in the level of the tumor monocyte content as compared to a first reference level and a second change in the level of the tumor mutational burden as compared to a second reference level; and identifying the subject as likely responding to the treatment of the cancer therapy if the level of the tumor monocyte content is elevated as compared to the first reference level; and / or if the level of the tumor mutational burden is elevated as compared to the second reference level, wherein the cancer therapy comprises an immunotherapy and a chemotherapy.
32. A method of treating a cancer with a cancer therapy in a subject in need thereof, comprising: identifying a subject who is likely responsive to the treatment of the cancer therapy comprising the immunotherapy and the chemotherapy according to the method of claim 30 or 31, and administering to the subject a therapeutically effective amount of the cancer therapy.
33. The method of any one of claims 30-32, wherein the immunotherapy comprises one or more immune checkpoint inhibitors.
34. The method of claim 33, wherein the one or more immune checkpoint inhibitors comprise a CTLA-4 inhibitor, a PD-1 inhibitor, a PD-L1 inhibitor a LAG-3 inhibitor, or a TIM-3 inhibitor.
35. The method of claim 33, wherein the one or more immune checkpoint inhibitors comprise an anti-CTLA-4 antibody, an anti-PD-1 antibody, an anti-PD-Ll antibody, an anti-LAG-3 antibody, an anti-HM-3 antibody, or a combination thereof.
36. The method of claim 33, wherein the one or more immune checkpoint inhibitors comprise an anti-PD-Ll antibody, an anti-CTLA-4 antibody, or a combination thereof.
37. The method of claim 33, wherein the one or more immune checkpoint inhibitors comprise durvalumab.
38. The method of any one of claims 30-37, wherein the chemotherapy comprises oxaliplatin, capecitabine, or a combination thereof.
39. A method of identifying a subject as likely not responding to treatment of a cancer therapy, comprising: determining an expression level of each of biomarkers comprising one or more of T cellspecific interfer on-stimulated genes (ISGs) in a sample from the subject; determining a change in the expression level of each of the biomarkers as compared to a respective reference expression level for each of the biomarkers; and identifying the subject as likely not responding to the treatment of the cancer therapy if the expression level of each of the biomarkers is elevated as compared to the respective reference expression level for each of the biomarkers.
40. The method of claim 39, wherein interferon-stimulated genes comprise one or more of OAS1, MX1, IFI6, and XAFl.
41. The method of claim 39, wherein the cancer therapy comprises an immunotherapy and optionally a chemotherapy.
Citation Information
Patent Citations
Biomarkers for cancer immunotherapy outcomes
US20210164054A1
PD-l2 biomarkers predictive of PD-1 pathway inhibitor responses in esophagogastric cancers
WO2016144673A1
Diagnostic and therapeutic methods for cancer
WO2019090263A1
Molecular gene signatures and methods of using same
WO2019226514A2
Nucleotide combination and use thereof
WO2022012420A1
