Biomarkers for immunotherapy responsiveness in colorectal cancer

Abstract

Provided herein are methods for predicting or assessing the response of a subject to treatment with an immunotherapeutic agent, wherein the subject has mismatch repair proficient (pMMR) colorectal cancer. Also provided herein are novel biomarkers for identifying subjects having mismatch repair proficient (pMMR) colorectal cancer who will respond to immunotherapy.

Description

[0001] BIOMARKERS FOR IMMUNOTHERAPY RESPONSIVENESS IN COLORECTAL CANCER

[0002] FIELD OF THE INVENTION

[0003] The present invention relates to methods for selecting subjects with colorectal cancer for treatment with an immunotherapeutic agent. In particular, the invention provides new biomarkers to identify colorectal cancer patients who will respond to immunotherapy.

[0004] BACKGROUND

[0005] Colorectal cancer is the most common adult cancer and the fourth leading cause of cancer mortality. Colorectal cancers can be classified by mismatch repair status into two subtypes: proficient (pMMR, -85% of cases) and deficient (dMMR, -15% of cases), which show low and high tumour mutational burdens (TMB), respectively (The Cancer Genome Atlas Network. Comprehensive molecular characterization of human colon and rectal cancer. Nature 487, 330-337 (2012); Front. Immunol., 12 August 2020 Sec. Cancer Immunity and Immunotherapy Volume 11 . 2020).

[0006] Anti-PD1 immunotherapy can be highly effective in treating dMMR / high TMB colorectal cancers, but is rarely used as a treatment option given the low occurrence of dMMR / high TMB cancers in the general population. Even within dMMR / high TMB colorectal cancers, about half of patients treated with anti- PD1 immunotherapy show no response (Ganesh K, Stadler ZK, Cercek A, et al. Immunotherapy in colorectal cancer: rationale, challenges and potential. Nat Rev Gastroenterol Hepatol 2019;16:361- 375). This means potential non-responders are exposed to anti-PD1 or anti-PDL1 treatment regimens unnecessarily. pMMR / low TMB colorectal cancers are much more common relative to dMMR / high TMB colorectal cancers, but are considered non-responsive to PD-1 / PD-L1 immunotherapy (Kalyan et al. Updates on immunotherapy for colorectal cancer. J Gastrointest Oncol. 2018. 9:160-169). Treatment options for a large portion of the colorectal cancer population are therefore limited.

[0007] In other cancers, such as lung cancer, the expression of PD-1 or PD-L1 itself can be used as a biomarker for responsiveness to treatment, with higher expression levels indicating a greater level of sensitivity to treatment. However, PD-1 and PD-L1 expression levels are not a useful biomarker in colorectal cancer and do not predict responsiveness to anti-PD1 immunotherapy.

[0008] Thus, there is a need to identify new biomarkers for identifying those colorectal cancer patients who will respond to PD-1 / PD-L1 immunotherapy. SUMMARY OF THE INVENTION

[0009] In a first aspect, provided herein is a method for predicting or assessing the response of a subject to treatment with an immunotherapeutic agent, wherein the subject has mismatch repair proficient (pMMR) colorectal cancer, the method comprising: determining a level of CD74 expression in a sample obtained from the subject and comparing it to a reference value, wherein a CD74 expression level higher than the reference value is indicative of an immunotherapy responder, and wherein a CD74 expression level lower than the reference value is indicative of an immunotherapy non-responder.

[0010] In a second aspect, provided herein is a method of selecting a subject for treatment with an immunotherapeutic agent, wherein the subject has mismatch repair proficient (pMMR) colorectal cancer, the method comprising: determining a level of CD74 expression in a sample obtained from the subject; and comparing the level of CD74 expression in the sample to a reference value; wherein the subject is selected for treatment with the immunotherapeutic agent if the level of CD74 in the sample is higher than the reference value.

[0011] In a third aspect, provided herein is a method of selecting a subject for treatment with anti-PD-1 or anti-PD-L1 immunotherapy, wherein the subject has mismatch repair proficient (pMMR) colorectal cancer, the method comprising: a. providing a tumour sample obtained from the subject, b. determining a percentage of tumour stromal cells in the sample that express CD74, and c. selecting the subject for treatment with anti-PD-1 or anti-PD-L1 immunotherapy if greater than 19% of tumour stromal cells in the sample express CD74.

[0012] In a fourth aspect, provided herein is a method of treating colorectal cancer in a subject, wherein the subject has been identified as having mismatch repair proficient (pMMR) colorectal cancer, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; b. comparing the level of CD74 expression in the sample to a reference value; and administering a therapeutically effective amount of an immunotherapeutic agent to the subject if the level of CD74 expression in the sample is higher than the reference value.

[0013] In a fifth aspect, provided herein is a method of selecting a treatment regimen for a subject having mismatch repair proficient (pMMR) colorectal cancer, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; b. comparing the level of CD74 expression in the sample to a reference value; and c. selecting a treatment regimen for the subject if the level of CD74 expression in the sample is higher than the reference value

[0014] In a sixth aspect, provided herein is a method of stratifying mismatch repair proficient (pMMR) colorectal cancer patients by response to immunotherapy, the method comprising: a. determining a level of CD74 expression in samples obtained from the patients; and b. comparing the level of CD74 expression in the samples to a reference value; wherein patients with a CD74 expression level higher than the reference value are classified as responders, and patients with a CD74 expression level lower than the reference value are classified as non-responders.

[0015] In a seventh aspect, provided herein is a method, comprising: a. providing a sample obtained from a subject having mismatch repair proficient (pMMR) colorectal cancer; b. detecting CD74 expression in the sample; and analysing the CD74 expression in the sample, wherein said analysing comprises calculating a CD74 stroma positive score (SPS).

[0016] In an eighth aspect, provided herein is a PD-1 inhibitor or PD-L1 inhibitor for use in a method of treating colorectal cancer in a subject, wherein the subject has been identified as having mismatch repair proficient (pMMR) colorectal cancer.

[0017] In a ninth aspect, provided herein is a PD-1 inhibitor or PD-L1 inhibitor for use in a method of treating mismatch repair proficient (pMMR) colorectal cancer in a subject, the method comprising determining a level of CD74 expression in a sample obtained from the subject, and administering a therapeutically effective amount of the PD-1 inhibitor or PD-L1 inhibitor to the subject if the level of CD74 expression in the sample is higher than a reference value.

[0018] In a tenth aspect, provided herein is a method for predicting or assessing the response of a subject to treatment with an immunotherapeutic agent, wherein the subject has colorectal cancer with a low tumour mutational burden, the method comprising: determining a level of CD74 expression in a sample obtained from the subject and comparing it to a reference value, wherein a CD74 expression level higher than the reference value is indicative of an immunotherapy responder, and wherein a CD74 expression level lower than the reference value is indicative of an immunotherapy non-responder. In an eleventh aspect, provided herein is a method of selecting a subject for treatment with an immunotherapeutic agent, wherein the subject has colorectal cancer with a low tumour mutational burden, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; and b. comparing the level of CD74 expression in the sample to a reference value; wherein the subject is selected for treatment with the immunotherapeutic agent if the level of CD74 in the sample is higher than the reference value.

[0019] In a twelfth aspect, provided herein is a method of selecting a subject for treatment with an immunotherapeutic agent, wherein the subject has mismatch repair deficient (dMMR) colorectal cancer, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; and b. comparing the level of CD74 expression in the sample to a reference value; wherein the subject is selected for treatment with the immunotherapeutic agent if the level of CD74 in the sample is higher than the reference value.

[0020] BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 : (A) Experimental design. Regions at the core (n = 57) and invasive margin (n = 59) of 59 formalin-fixed paraffin-embedded (FFPE) CRC blocks from 58 colorectal cancer (CRC) patients were identified by a clinical pathologist from haematoxylin and eosin (H&E)-stained slides. Regions were subsequently laser capture micro-dissected (LCM) to obtain tumour epithelium (Te), and associated tumour stroma (Ts) compartments. Normal epithelium (Ne) was separately obtained by LCM from tumour-adjacent areas for six MMR balance samples. (B) Twenty-nine dMMR and 30 pMMR CRCs were included in this study. Twenty-four dMMR CRCs received anti-PD1 immunotherapy in the metastatic setting. Response to therapy was assessed with Response Evaluation Criteria in Solid Tumours v1 .1 (Eisenhauer EA, Therasse P, Bogaerts J, Schwartz LH, Sargent D, Ford R, Dancey J, Arbuck S, Gwyther S, Mooney M, et al: New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer 2009, 45:228-247). (C) Comparison of tumour mutational burden (TMB) between dMMR and pMMR CRCs. The two dotted lines correspond to the TMB threshold of immunotherapy eligibility (Marabelle A, Le DT, Ascierto PA, Di Giacomo AM, De Jesus-Acosta A, Delord JP, Geva R, Gottfried M, Penel N, Hansen AR, et al: Efficacy of Pembrolizumab in Patients With Noncolorectal High Microsatellite Instability / Mismatch Repair- Deficient Cancer: Results From the Phase II KEYNOTE-158 Study. J Clin Oncol 2020, 38:1-10; Marcus L, Fashoyin-Aje LA, Donoghue M, Yuan M, Rodriguez L, Gallagher PS, Philip R, Ghosh S, Theoret MR, Beaver JA, et al: FDA Approval Summary: Pembrolizumab for the Treatment of Tumor Mutational Burden-High Solid Tumors. Clin Cancer Res 2021 , 27:4685-4689) and TCGA subtype (Cancer Genome Atlas N: Comprehensive molecular characterization of human colon and rectal cancer. Nature 2012, 487:330-337). Distributions were compared using two-sided Wilcoxon’s rank sum test. (D) Significantly altered pathways between dMMR and pMMR CRCs. Proportions were compared using corrected two-sided Fisher’s test. Percentage of mutated samples are shown for each subtype and pathway. (E) Percentages of CRCs wild-type or mutant for both KRAS and hypoxia signalling pathways versus different pathway statuses. Numbers per bar represent n tumours and p- value reported calculated by two-sided Fisher’s exact test of the two status groups indicated. (F) Number of covered genes across samples groups defined by MMR status and tissue compartment. Within compartment distributions are compared used two-sides Wilcoxon’s rank sum test. (G) Whole transcriptome dimensionality reduction by multidimensional scaling (MDS) plot of expression data from all samples used in this study. Shaded clustering overlay represents the extent of Ts and Te clusters. (H) MDS plot of Ts samples shaded by MMR subtype where shaded clustering overlay represents the extend of region clustering. (I) As H, for Te. (J) Hierarchical clustering of LCM-derived tumourepithelium, Te, samples. (K) Hierarchical clustering of LCM-derived tumour associated stroma, Ts, samples. (L) Summary of Te samples categorised by adjacency of cognate samples from the tumour invasive margin or core being nearest neighbours in the hierarchical clustering of J. (M) Summary of Ts samples categorised by adjacency of cognate samples from the tumour invasive margin or core being nearest neighbours in the hierarchical clustering of K. (N) Density plots of all pairwise Manhattan distances within dMMR or pMMR Te. Distribution compared used two-sided Wilcoxon’s rank sum test. In all cases, significance is indicated when statistic <0.05.

[0022] (O) Oncoplot showing the ten most frequently mutated genes, TMB distribution and baseline clinical characteristics of CRCs analyzed in this study. Dashed lines correspond to TMB thresholds for immunotherapy eligibility (10 muts / Mbp) and CRC subtyping (12 muts / Mbp).

[0023] Figure 2: (A) Intersection of altered genes across genomic profiling approaches. 17,406 genes were altered in at least one sample sequenced by whole exome sequencing (WES). The Foundation Medicine, FoundationOne CDx (FM1) test is a clinical targeted exome sequencing approach that profiles 324 cancer driver genes. 767 genes from the CARIS Life Sciences Molecular Intelligence (CARIS) test were altered in at least one patient in our cohort. The gene intersection between the clinically available profiling approaches was 324 genes, which we used for in-house targeted exome sequencing where other profiling were not available. (B) Alteration status of 20 pathways between dMMR and pMMR CRCs. Twenty cancer-related pathways covering 324 genes (Methods: Curation of mutated biological pathways) were tested for more frequent mutation in hypermutated or nonhypermutated CRC. Circularised bar plots show the proportion of samples by hypermutation status carrying at least one altered gene within the respective pathway with the number representing the percentage within the group. Statistics results represent Fisher test p-values shown are corrected for multiple testing using Benjamini-Hochberg FDR correction. (C) Clinical characteristics of dMMR and pMMR CRCs. Age distribution was compared using two-sided Wilcoxon’s rank sum test. Boxplots show first and third quartiles, whiskers extend to the lowest and highest value within the 1 .5X interquartile range and the line indicates the median. Proportions of categorical features were compared using two-sided Fisher’s test.

[0024] Figure 3: (A) RNA-sequencing experimental design. Two independent libraries were prepared using 2 ng of total RNA from each of the 235 samples derived from the laser captured micro-dissected normal epithelium (Ne, n=6), tumour epithelium (Te, n=114), and associated tumour stroma (Ts, n=115) compartments. Libraries were sequenced in 7 batches approximately balanced for the number of Te, Ts and Ne samples coming from pMMR and dMMR tumours. (B) Number of covered genes from libraries generated from 2, 5, 10 or 40ng input cDNA by tissue compartment. The 2ng libraries were independently produced in quadruplicate and all combination duplex libraries from these obtained from concatenation of fastq read counts. (C) Number of covered genes per sample in pairs of tissue compartments from tumour regions. A total of 59 CRC tumours underwent laser capture microdissection of tumour stroma (Ts, upper bars) and tumour epithelium (Te, lower bars) at the core and invasive margin regions of the tumour. (D) Dimensionality reduction of whole transcriptome microdissected samples. The 235 micro-dissected samples (229 CRC, six adjacent normal epithelium) were visually inspected for sequencing batch effects by reduction of the whole transcriptomes to a two- dimensional space representing the highest proportion of variance explained (multi-dimensional scaling (MDS), within the limma R package). Each dot represents an independent sample shaded by one of seven sequencing batches. (E) Whole transcriptome dimensionality reduction by multidimensional scaling (MDS) plot of expression data of all dMMR samples used in this study. Shaded clustering overlay represents the extent of Ts and Te clusters. (F) Whole transcriptome dimensionality reduction by multidimensional scaling (MDS) plot of expression data of all pMMR samples used in this study. Shaded clustering overlay represents the extent of Ts and Te clusters.

[0025] Figure 4: (A) Comparison normalised enrichment scores (NES) of IFNa signature from tumour epithelium (Te) from core versus invasive margin regions of the tumour separated by MMR status. Adjacent normal epithelium (Adj.N) shown for comparison. (B) Increased (NES Te > Ne) and decreased (NES Te < Ne) intrinsic signatures in dMMR and pMMR Te compared to normal epithelium (Ne) shown by significance. (C) Comparison of NES of intrinsic signatures between dMMR and pMMR Te obtained by micro-dissection. dMMR versus pMMR comparisons shown above with MMR-specific comparisons to adjacent normal shown below. Distribution compared used corrected two-sided Wilcoxon’s rank sum test. In all cases, significance is indicated when statistic <0.05.

[0026] Figure 5: (A) Internal validation of the LM22-derived gene signatures. Heatmap of the normalised enrichment scores (NES) of CIBERSORT-derived immune population signatures. The LM22 signatures from CIBERSORT {Newman, 2015 #31} were used to curate gene lists for the 22 cell types described assigning each of the 547 genes given to a cell in which it is highest expressed. The resultant signatures were back-validated onto the pure cell type expression profiles available from CIBERSORT using single sample geneset enrichment analysis (ssGSEA {Barbie, 2009 #38}. The 22 cell types are given in a least duplicate where column shading annotations represent each of the 22 populations. (B) External validation of the immune population signatures. Boxplots of normalised enrichment scores of immune signatures tested in human CRC-obtained pseudobulk expression profiles. The filtered and / or re-curated signatures representing final 15 immune cell populations were externally validated against immune populations in a human CRC dataset {Pelka, 2021 #32}. Expression data from this dataset were made into pseudobulk data by summing counts for cell sharing the same annotation in the ‘ClusterMidway’ annotation setting, respecting TMB subtype status, and enrichment score for each of the 15 signatures calculated. Panels show mismatch repair-deficient (dMMR) and mismatch repair-proficient (pMMR) boxplots for each signature tested across all the pseudobulk profiles with the cognate positive control profile shown in blue. (C) Comparison on ESTIMATE-derived tumour purity of micro-dissected tumour epithelium between hypermutated (HM-T, N = 54) and non-hypermutated (nHM-T, N = 58), and between core and margin within the same TMB status (HM-Tc versus HM-Tm (A / = 27, N = 28, respectively), and nHM-Tc versus nHM-Tm (A / = 29, A / = 30, respectively)). Wilcoxon statistic result reported. (D) Tumour epithelium samples passing ESTIMATE purity thresholds by MMR subtype. The number of samples is reported on the y-axis while data point annotations report the percentage per subtype. Wilcoxon statistic result reported for the comparison of surviving samples’ Cytotoxic-23 NES of dMMR versus pMMR Te.

[0027] Figure 6: (A) Comparison of normalised enrichment scores (NES) of extrinsic immune signatures between dMMR and pMMR tumour stroma (Ts) obtained by micro-dissection. Distribution compared used corrected two-sided Wilcoxon’s rank sum test. In all cases, significance is indicated when statistic <0.05. (B) As in A, of extrinsic stromal signatures. (C) Chord plot representing the global correlation of signature scores between connected intrinsic, extrinsic immune from Ts and extrinsic stromal from Ts filtered by Pearson r >|0.4| and p-value <0.05. (D) As in C for only dMMR correlations. (E) As in C for only pMMR correlations. (F) Chord plot representing the global correlation of signature scores within extrinsic immune from Ts filtered by Pearson’s coefficient r >|0.4| and p-value <0.05. (G) Comparison of normalised enrichment scores (NES) of significant extrinsic immune signatures between dMMR and pMMR tumour epithelium (Te) obtained by micro-dissection. Distribution compared used corrected two-sided Wilcoxon’s rank sum test. In all cases, significance is indicated when statistic <0.05. (H) Enrichment plots of Cytotoxic-23, M1-16 and NK-33 signatures using genelevel fold-change as the pre-rank between dMMR (left) and pMMR (right) Te. (I) Correlation plots between IHC quantification of CD3 / CD68 / CD57 in Te and Ts against Cytotoxic-23 / M1-16 / NK-33 NES for a subset of representative samples across the distribution of NES values. (J) Representative IHC images of CRCs showing stromal and intra-epithelial infiltration (blue arrows) of CD68-positive or CD57-positive cells. (K) Chord plot representing the global correlation of signature scores within extrinsic immune from Te filtered by Pearson’s coefficient r >|0.4| and p-value <0.05. (L) Chord plot representing the global (left), dMMR (middle) and pMMR (right) correlation of signature scores between connected extrinsic immune from Ts paired Te, filtered by Pearson’s coefficient r >|0.4| and p-value <0.05. (M) Representative positive cells for CD57, GZMB, and CD68 immunostains. Arrows indicate IE positive cells. Scale bars = 100 pm.

[0028] Figure 7: (A) A. Hierarchical clustering of dMMR CRC by MMR-defining immune enrichment (Cyt-23, NK-33, M1-16, IFNag / STING) showing heatmap of these enrichments and mutation status of genes found differentially mutated between defined sub-groups (hot versus cold; hot / IFN-response versus cold / hot-stroma; hot versus hot-stroma; hot versus IFN-response; hot-stroma versus IFN-response; hot / hot-stroma versus IFN- response). (B) Hierarchical clustering of pMMR CRC by MMR-defining immune enrichment (Cyt-23, NK-33, M1-16, IFNag / STING) showing heatmap of these enrichments and mutation status of genes found differentially mutated between defined dMMR sub-groups. (C) Tumour mutational burden score of dMMR CRCs grouped by cold versus rest as defined in (A). P- value reported by two-sided Wilcoxon test with significance at p<0.05. (D) Hierarchical clustering and heatmap of MMR-defining enrichment signatures in a combined KRAS validation dataset of hypermutated samples from TCGA and CPTAC showing also KRAS-28 NES and KRAS gene status. (E) Enrichment scores of interferon-related processes across sub-groups defined within the KRAS validation dataset. (F) Enrichment scores of MMR-defining immune signatures across sub-groups defined within the KRAS validation dataset. (G) Enrichment scores of KRAS-27 mutational signature across sub-groups defined within the KRAS validation dataset. (E-G) FDR p-value reported by Wilcoxon test with significance at p <0.05. (H) Proportion of dMMR CRCs grouped cold versus rest as defined in (A) and response to anti-PD1 immunotherapy; all samples (left) and excluding samples having received neoadjuvant therapy of any kind (right). P-value reported by two-sided Fisher’s exact test with significance at p <0.05.

[0029] Figure 8: (A) Hierarchical clustering and heatmap of dMMR and pMMR CRCs by MMR-defining immune enrichment (Cyt-23, NK-33, M1-16, IFNag / STING). (B) Proportion plot of dMMR CRCs grouped by infiltration status from (A) and response to anti-PD1 immunotherapy. (C) Proportion plot of dMMR and pMMR CRCs grouped by infiltration status from (A) and response to anti-PD1 immunotherapy. (D) Enrichment scores of predictive ICI response signatures across CRCs grouped by infiltration status from (A) in the tumour stroma and tumour compartments of dMMR (top) and pMMR (bottom). P-value reported by two-sided Wilcoxon test with significance at p <0.05. (E) CD74 mRNA expression across CRCs grouped by infiltration status from (A) in the tumour stroma and tumour compartments of dMMR (top) and pMMR (bottom). P-value reported by two-sided Wilcoxon test with significance at p <0.05. (F) Correlations of MMR-defining immune enrichment scores (Cyt-23, top; NK-33, middle, and M1-16, bottom) against CD74 mRNA expression in tumour-adjacent normal epithelium (left), tumour stroma (middle) and tumour (right). Pearson’s correlation and significance are reported globally and per MMR subtype. Figure 9: (A) Experimental setup of mouse experiments. Mice were implanted with either MC38 (n=16) or CT26 (n=16) cell lines and pulsed with murine anti-PD1 (n=5 per cell line) or IgG (n=11 per cell line) isotype control on the days indicated. Tumours were harvested at approximately 21 days for investigation. (B) Representative tumour growth curve of MC38 tumours treated with murine anti-PD1 (red) or isotype control (black). P-value reported by two-sided Wilcoxon test of end-point distributions with significance at p <0.05. (C) As in B for representative CT26 tumours correlations. (D) Proportion of cells expressing CD74 by immunohistochemistry (IHC) protein quantification from MC38 or CT26 murine tumours. P-value reported by two-sided Wilcoxon test of end-point distributions with significance at p <0.05. (E) Representative flow cytometry acquisition from MC38- (left) and CT26- derived (right) live cells stained and plot for CD45 and CD74 fluorescence. (F) CD74 scores from IHC protein quantification of 24 human dMMR CRCs separated by immunotherapy response. CD74 was quantified in from both stromal (S) and tumour (T) cells to calculate: combined positive score (CPS: S- positive plus T-positive over all detected cells); stroma positive against tumour score (SPT: S-positive over all detected tumour cells); tumour positive score (TPS: T-positive over all detected tumour cells); stroma positive score (SPS: S-positive over all detected stromal cells). P-value reported by two-sided Wilcoxon test with significance at p <0.05. (G) Stroma positive score (SPS) of all 59 human CRCs included in this study separated by MMR status and grouped by infiltration status from (Figure 8A). A stromal positivity-proportion of 19% appropriately separates clinical (dMMR, dot shade) and putative immunotherapy response (boxplot shade). P-value reported by two-sided Wilcoxon test of end-point distributions with significance at p <0.05.

[0030] Figure 10: Enrichment signatures representative of tumour intrinsic (Intrinsic), immune (Extrinsic immune) and stromal (Extrinsic stroma) were curated for application using single-sample geneset enrichment analysis (ssGSEA). Where signatures were derived from more than one source, genes appearing in more than one list were retained, and genes appearing in more than one signature within the same signature category were removed also to minimise feature overlap. Intrinsic signatures were experimentally validated and tested for pathway enrichment using EnrichR online tool as described. Extrinsic immune signatures were validated by testing against pseudobulk samples derived from single-cell expression study of human colorectal cancer.

[0031] Figure 11 : CRC clustering based on extrinsic and intrinsic properties

[0032] (A) Hierarchical clustering of dMMR and pMMR CRCs based on NK cell, cytotoxic lymphocyte, CXCL10+ TAM, IFN production and response, and STING activation signature NES values.

[0033] (B and C) Comparison of stromal IFNG expression levels across the four CRC clusters (B) and between IFN-high and IFN-low IP CRCs (C). Cluster B was classified as IFN-high IP since the median IFNG expression was comparable to cluster A and significantly higher than cluster C.

[0034] (D) TMB comparisons between IFN-high and IFN-low IP CRCs. (E and F) Proportions of patients responding or not to ICI between IFN-high and IFN-low IP dMMR (E) and all (F) CRCs. All pMMR CRC patients were considered as non-responders.

[0035] (G) Proportions of patients with liver metastasis in IFN-high and IFN-low IP CRCs. Only 55 patients for which this information was available were compared. Distributions in (B)-(D) were compared using two-sided Wilcoxon’s rank-sum test. Boxplots show first and third quartiles, whiskers extend to the lowest and highest value within the 1 .5x interquartile range and the line indicates the median. Proportions in (E)-(G) were compared using two-sided Fisher’s exact test. Numbers below each plot represent CRCs used in the analysis, except in (B) and (C) where IFNG expression levels in the core and invasive margin regions of each CRC were used. CRC, colorectal cancer; dMMR, mismatch repair deficient; ICI, immune checkpoint inhibition; IE, intra-epithelial; IFN, interferon; IP, immunophenotype; NES, normalized enrichment score; pMMR, mismatch repair proficient; TAM, tumour-associated macrophage; Te, tumour epithelium; TMB, tumour mutational burden; Ts, tumour stroma.

[0036] Figure 12: T cell induced CD74 overexpression in TAMs and CRC cells

[0037] (A) Representation of TAMs defined by proximity to T cells. Cells of interest could be identified in the CosMx slides using their coordinates and cell neighbourhood (STAR Methods).

[0038] (B) Genes up- and down-regulated in TAMs adjacent to T cells. Numbers of cells used for comparison across all four samples are indicated.

[0039] (C) Representation of tumour epithelial and T cells defined by local neighbourhood (STAR Methods). (D and E) Genes up- and down-regulated in IE T cells (D) and in epithelial cells next to T cells (E). Numbers of cells used for comparison across all four samples are indicated.

[0040] (F) CD74 median gene expression levels in Ts and Te of IFN-high and IFN-low IP dMMR and pMMR CRCs.

[0041] (G and H) Experimental setup (G) and CD74 relative expression (H) in CRC cell lines cultured with or without activated CD8+ T cells and IFNy inhibitor. Numbers represent biological replicates used in the analysis.

[0042] (I) Genes up- and down-regulated in HCT116 cells cultured with activated CD8+ T cells. Differentially expressed genes in (B), (D), and (E) have FC > 1 .1 or < 0.9 and FDR < 0.01 , while in (I) have FC > 2 or < 0.5 and FDR < 0.05. Representative APP and IFN genes (B, E, and I) and cytotoxicity markers

[0043] (D) are shown. Distributions in (F) were compared using two-sided Wilcoxon’s rank-sum test; distributions in (H) were compared using Student’s t test on the mean of means of the technical replicates across biological batches. Shown are the means ± SD. APP, antigen processing and presentation; CRC, colorectal cancer; dMMR, mismatch repair deficient; Epi, epithelial; FC, fold change; FDR, false discovery rate; IE, intraepithelial; IFN, interferon; pMMR, mismatch repair proficient; SD, standard deviation; TAM, tumour-associated macrophage; Te, tumour epithelium; Ts, tumour stroma. Figure 13: CD74 expression in human and murine CRCs treated with anti-PD-1 ICI

[0044] (A) Schematics of CD74 scores per tissue compartment.

[0045] (B) ROC curves to assess sensitivity and specificity of the four CD74 scores in the UCLH cohort.

[0046] (C) Proportions of dMMR CRC responders and non-responders in the UCLH cohort by 18% SPS threshold.

[0047] (D) CD74 SPS distributions in responder and non-responder dMMR CRCs from the GONO cohort.

[0048] (E) Proportions of dMMR CRC responders and non-responders in the GONO cohort by 18% SPS threshold.

[0049] (F) CD74 SPS distributions in responder and non-responder dMMR CRCs treated with a combination of anti-PD-1 and anti-CTLA-4 agents.

[0050] (G) Proportions of dMMR CRC responders and non-responders treated with a combination of anti-PD- 1 and anti-CTLA-4 agents by 18% SPS threshold.

[0051] (H) CD74 SPS distribution in patients from the UCLH and GONO cohorts with information about the presence of liver metastasis.

[0052] (I) Proportions of responders and non-responders in patients from the UCLH and GONO cohorts with information about the presence of liver metastasis.

[0053] (J and K) TMB (J) and MSI score (K) of MC38 and CT26 cells. TMB was calculated as non- synonymous mutations per mega base-pairs. MSI score was measured using MSI-sensor-pro28 as compared to the corresponding background genome (C57BL / 6 for MC38 and BALBc for CT26, respectively).

[0054] (L) Relative Cd74 expression in MC38 or CT26 tumours. Bars represent the standard deviation of RQ values across tumours. Samples from one experimental batch were analyzed.

[0055] Distributions in (D), (F), (H), (L) and (K) were compared using two-sided Wilcoxon’s rank-sum test. Boxplots show first and third quartiles, whiskers extend to the lowest and highest value within the 1 .5x interquartile range and the line indicates the median. Proportions in (C), (E), (G), and (I) were compared using two-sided Fisher’s exact test. Numbers below each plot represent CRCs used in the analysis, except for (L) and (Figure 9D) where they represent individual mice and tumours. Shown are the means ± SD. aCTLA-4, anti-CTLA-4; aPD-1 , anti-PD-1 ; AUC, area under the curve; CRC, colorectal cancer; CPS, combined positive score; dMMR, mismatch repair deficient; FDR, false discovery rate; ICI, immune checkpoint inhibition; IFN, interferon; IP, immunophenotype; Mbp, mega base pair; MSI, microsatellite instability; muts, mutations; NR, non-responder; pMMR, mismatch repair proficient; R, responder; ROC, receiver operating characteristic; RQ, relative quantification; SD, standard deviation; SPS, stroma proportion score; SPT, stroma proportion over tumour cells; Te, tumour epithelium; TMB, tumour mutational burden; TPS, tumour proportion score Ts, tumour stroma.

[0056] Figure 14: CD74 quantification in human pMMR CRCs

[0057] (A) CD74 SPS distributions in IFN-high and IFN-low IP pMMR CRCs. (B) CD74 staining in representative dMMR (UH17 and UH22) and pMMR (CR10 and CR35) CRCs with 18%> SPS >18%. Scale bar = 100pm.

[0058] (C) Schematics of the samples from the CAMILLA trial used in this study.

[0059] (D) CD74 gene expression levels in stroma and tumour regions of responders and non-responders from the CAMILLA trial.

[0060] (E) Schematics of samples from the AtezoTRIBE trial used in this study.

[0061] (F) CD74 SPS distributions in IS-IC high and IS-IC low pMMR CRCs.

[0062] (G-J) Kaplan-Meier estimates of PFS of patients who received (G) or not (H) surgery of the primary tumour and patients in the experimental (I) and control (J) arm of the AtezoTRIBE trial. Dashed black lines indicate median PFS values.

[0063] Distributions in (A), (D), and (F) were compared using two-sided Wilcoxon’s rank-sum test. Boxplots show first and third quartiles, whiskers extend to the lowest and highest value within the 1 .5x interquartile range and the line indicates the median. Progression-free survival probabilities in (G)-(J) were compared using the Log rank statistic. HR and p values of the interaction between CD74 SPS and treatment (G and H) as well as CD74 SPS and surgery of the primary (I and J) were calculated using Cox PH models. Numbers under each plot represent CRCs used in the analysis. Cl, confidence interval; CRC, colorectal cancer; HR, hazard ratio; IFN, interferon; IP, immunophenotype; IS-IC, Immunoscore-Immune-Checkpoint; NR, non-responder; pMMR, mismatch repair proficient; R, responder; ROIs, regions of interest; SPS, stroma proportion score.

[0064] Figure 15: (A) Comparisons of intrinsic NES distributions between Te and Ne in dMMR and pMMR CRCs. Distributions were compared using two-sided Wilcoxon’s rank sum test and corrected for FDR with the Benjamini-Hochberg method. Dashed lines correspond to FDR = 0.05. (B) Significant pairwise correlations (Pearson coefficient r >|0.4| and FDR <0.1) between NES values of intrinsic (blue letters) and immune (red numbers) signatures in dMMR (left) and pMMR (right).

[0065] (C) Significant pairwise correlations (Pearson coefficient r >|0.4| and FDR <0.1) between NES values of NK cell, cytotoxic lymphocyte, and CXCL10+ TAM signatures in Te, and all immune signatures in the Ts of the whole CRC cohort. CRC, colorectal cancer; dMMR, mismatch repair deficient; FDR, false discovery rate; Ne, normal epithelium; NES normalized enrichment score; NK, natural killer cell; pMMR, mismatch repair proficient; ssGSEA, single sample gene set enrichment analysis; TAM, tumor associated macrophage; Te, tumor epithelium; Ts, tumor stroma.

[0066] Figure 16: Validation of IFN-high and IFN-low IPs in CRC, related to Figure 11 .

[0067] (A, B) Hierarchical clustering of CRCs from CPTAC (A) and TCGA (B) based on NK cell, cytotoxic lymphocyte, CXCL10+ TAM, IFN production and response, and STING activation signatures.

[0068] (C) IFNG expression levels between IFN-high and IFN-low CRCs in CPTAC and TCGA.

[0069] (D) Proportions of iCMS3_MSI and MSS_F CRCs in IFN-high and IFN-low CRCs. CRCs were classified into iCMS subtypes, as described Figure 12A, using nearest template prediction within the CMScaller v2.0.1 package on centered and scaled expression data from Te samples. Samples with a prediction FDR <0.1 from 100 permutations were considered indeterminate.

[0070] (E) CD274 expression levels between Ts of IFN-high and IFN-low in dMMR and pMMR CRCs. A normalized expression value of 4.97 represents no expression.

[0071] (F) Comparison of stromal and epithelial SOX17 gene expression levels between IFN-high and IFN- low CRCs. A normalized expression value of 4.97 represents no expression.

[0072] (G) Hierarchical clustering of dMMR and pMMR CRCs based on all 21 signatures significantly different between dMMR and pMMR CRCs (Figure 4C, 6A and 6H) . CRC subtype, immunophenotype (as defined in Figure 11 A) and response to ICI are shown.

[0073] (H) Stromal IFNG expression levels across CRC clusters defined in Figure 16G.

[0074] (I) Proportions of dMMR CRCs responding or not to ICI in the three CRC clusters defined in Figure 16G. CRCs untreated / NA were excluded from the analysis. Distributions in panels C, E, F and H were compared using two-sided Wilcoxon’s rank sum test and panel E was corrected for FDR with the Benjamini-Hochberg method. Boxplots show first and third quartiles, whiskers extend to the lowest and highest value within the 1 .5X interquartile range and the line indicates the median. Proportions in D and I were compared using two-sided Fisher’s exact test. Numbers in each plot represent samples used in the analysis. CRC, colorectal cancer; dMMR, mismatch repair deficient; FDR, false discovery rate; ICI, immune checkpoint inhibition; iCMS, intrinsic-consensus molecular subtypes; IFN, interferon; IP, immunophenotype; MSI, microsatellite instable; MSS, microsatellite stable; NES, normalized enrichment score; NK cell, natural killer cell; pMMR, mismatch repair proficient; TAM, tumor associated macrophage; Te, tumor epithelium; Ts, tumor stroma.

[0075] Figure 17: Single-cell spatial transcriptomic clustering and phenotyping, related to Figure 12.

[0076] (A) UMAP plots of single cells in CosMx samples colored by cell type. The number of cells per type and total per sample are reported in brackets.

[0077] (B) Proportions of myeloid (left) and T / NK cell (right) subpopulations annotated using single-cell CRC data.

[0078] (C) Significantly over-represented (FDR <0.1) Reactome pathways in the 94 upregulated genes in TAMs adjacent to T cells compared to other TAMs. Gene ratios refer to the proportion of upregulated genes over the total genes in the Reactome term.

[0079] (D) Significantly upregulated proteins in TAMs and epithelial cells adjacent to T cells from CosMx single-cell proteomic CRC data. Main cell types were annotated from raw data using InSituType. Epithelial cells within a centroid distance of 15pm from >1 CD8 T were considered as adjacent to T cells and compared to epithelial cells with >4 epithelial cells within 15pm centroid distance. TAMs within a centroid distance of 15pm from >1 CD8+ T were defined as TAMs adjacent to T cells and compared to TAMs with no CD8+ T cells within 15pm centroid distance. (E, F) Comparisons of gene expression values of genes upregulated in TAMs and epithelial cells next to T cells (E) and of anti-PD-1 response-associated TAM markers (F) between IFN-high and IFN-low CRCs across subtypes (dMMR and pMMR) and compartments (Ts and Te).

[0080] Distributions in panel D, E and F were compared using two-sided Wilcoxon’s rank sum test and corrected for FDR with the Benjamini-Hochberg method. Dashed lines in D, E and F correspond to FDR = 0.05. CRC, colorectal cancer; dMMR, mismatch repair deficient; Epi, epithelial cell; FDR, false discovery rate; IFN, interferon; NK cell, natural killer cell; pMMR, mismatch repair proficient; TAM, tumor associated macrophage; Te, tumor epithelium; Ts, tumor stroma; pm, micrometer; UMAP, uniform manifold approximation and projection.

[0081] Figure 18: CD74 expression in CRC subtypes, tissue compartments and cell cultures, related to Figure 12.

[0082] (A) Stromal (left) and epithelial (right) CD74 expression in IFN-high and IFN-low dMMR and pMMR CRCs.

[0083] (B-D) CD74 normalized expression distributions in epithelial cells and TAMs next to T cells from CosMx (B), LCM (C) and single-cell (D) CRC data.

[0084] (E) Correlation between CD74 normalized expression in Ts and Te from dMMR and pMMR CRCs. Pearson correlation coefficient (r) and associated p value are reported.

[0085] (F-G) Experimental setup (F) and CD74 relative expression (G) in CRC cell lines cultured with or without IFNy and anti-IFNy antibody. Numbers represent biological replicates. Shown are the means ± SD.

[0086] (H) Representative gating strategy for FACS identification of CD8+ T cells and epithelial (EpCam+) populations in co-cultures.

[0087] Distributions in A-D were compared using two-sided Wilcoxon’s rank sum test, and panel A was corrected for FDR with the Benjamini-Hochberg method. Boxplots show first and third quartiles, whiskers extend to the lowest and highest value within the 1 .5X interquartile range and the line indicates the median. Distributions in G were compared using Student’s t-test on the mean across biological replicates. Numbers in each plot represent samples used in the analysis. CRC, colorectal cancer; dMMR, mismatch repair deficient; Epi, epithelial cell; FDR, false discovery rate; IFN, Interferon; IP, immunophenotype; LCM, laser-captured microdissection; pMMR, mismatch repair proficient; SD, standard deviation; TAM, tumor associated macrophage; Te, tumor epithelium; Ts, tumor stroma.

[0088] Figure 19: CD74 SPS and response to ICI in pMMR CRCs in IFN-high and IFN-low IPs, related to Figure 14.

[0089] (A) Proportions of responders and non-responders in the experimental arm of the AtezoTRIBE cohort by IS-IC category. (B) Comparison of CD74 SPS distributions in responders and non-responders in the experimental arm of the AtezoTRIBE cohort.

[0090] (C) Proportions of responders and non-responders in the experimental arm of the AtezoTRIBE cohort by CD74 SPS category.

[0091] (D) Kaplan-Meier estimates of PFS in the AtezoTRIBE cohort in patients who underwent or not surgery on their primary tumor.

[0092] (E) Proportions of patients with liver metastases in patients who underwent surgery on their primary tumor and received ICI in the AtezoTRIBE cohort by CD74 SPS category. Distributions in panel (B) were compared using two-sided Wilcoxon’s rank sum test. Boxplots show first and third quartiles, whiskers extend to the lowest and highest value within the 1 .5X interquartile range and the line indicates the median. Proportions in (A), (C) and (E) were compared using two-sided Fisher’s exact test. PFS probabilities in (D) were compared using the log-rank statistic. Numbers in each plot represent samples used in the analysis. CRC, colorectal cancer; dMMR, mismatch repair deficient; ICI, immune checkpoint inhibition; IFN, interferon; IP, immunophenotype; IS-IC, Immunoscore-IC; NR, non-responder; pMMR, mismatch repair proficient; PFS, progression-free survival; R, responder; SPS, stroma proportion score; Te, tumor epithelium; Ts, tumor stroma.

[0093] DETAILED DESCRIPTION

[0094] The present invention identifies CD74 as a novel biomarker to identify mismatch repair proficient colorectal cancer patients who will respond to immunotherapy, in particular to anti-PD1 / PD-L1 immunotherapy. pMMR colorectal cancers, and colorectal cancers with low tumour mutational burden, have typically been considered non-responders to immunotherapy, so the surprising identification of CD74 as a biomarker for immunotherapy responders in this patient group will enable access to new treatment options for these patients.

[0095] Definitions

[0096] Below are provided certain definitions of terms, technical means, and embodiments used herein.

[0097] DNA mismatch repair status

[0098] In general, cancers including colorectal cancers can be characterised as DNA mismatch repair proficient (pMMR) or DNA mismatch repair deficient (dMMR). dMMR status occurs when the normal mismatch repair (MMR) mechanism fails to correct DNA replication errors (for example, due to lack of MMR genes in tumour cells, or other defects in replication repair), resulting in the accumulation of mutations and microsatellites: segments of 1-6 nucleotides typically repeated multiple times at a particular genomic location (https: / / www.genome.gov / genetics-glossary / Microsatellite). dMMR cancers can also be known as high microsatellite instability (MSI-H) or with high tumour mutational burden (TMB). Conversely, pMMR cancers are considered to have normal DNA mismatch repair (also known as having low microsatellite instability (MSI-L) or being microsatellite stable (MSS), or with low TMB)). While anti-PD-1 immunotherapy has been shown to be effective in dMMR colorectal cancer, they have provided little to no clinical benefit in pMMR colorectal cancer (Kalyan et al. Updates on immunotherapy for colorectal cancer. J Gastrointest Oncol. 2018. 9:160-169). pMMR colorectal cancers are therefore generally considered non-responders to immunotherapy, and have even been described as “PD-L1 resistant” (Lin et al. PD-1 and PD-L1 inhibitors in cold colorectal cancer: challenges and strategies. Cancer Immunol Immunother. 2023. 72:3875-3893).

[0099] Surprisingly, despite pMMR colorectal cancers generally being considered non-responders to immunotherapy, the inventors have discovered that CD74 expression can be used to identify a subset of pMMR colorectal cancer subjects who will respond to immunotherapy. Accordingly, described herein are methods for predicting or assessing the response of a subject to treatment with an immunotherapeutic agent, wherein the subject has pMMR colorectal cancer, the method comprising comparing a level of CD74 expression in a sample from the subject to a reference value. Also described herein are methods of selecting a subject for treatment with an immunotherapeutic agent, wherein the subject has pMMR colorectal cancer, in which the subject is selected for treatment with the immunotherapeutic agent if a level of CD74 expression in a sample obtained from the subject is higher than a reference value. dMMR / pMMR status is routinely assessed for colorectal cancers. dMMR / pMMR status can be determined by testing for loss of MMR proteins (such as MLH1 , MSH2, MSH6, PMS2), or for the presence of microsatellite-high markers in the genome, according to standard methods in the art (Kawakami et al. MSI testing and its role in the management of colorectal cancer. Curr. Treat. Options Oncol. 2016. 16:30). For example, microsatellite-high testing can be performed on tumour tissue using PCR-based assays to establish if microsatellite markers show instability (Buhard O, Cattaneo F, Wong YF, et al. Multipopulation analysis of polymorphisms in five mononucleotide repeats used to determine the microsatellite instability status of human tumors. J Clin Oncol. 2006;24:241-51). MMR protein expression can be assessed using immunohistochemistry, either in addition to or instead of microsatellite testing. In general, a subject is identified as having pMMR if tumour tissue slides show a positive result for MMR protein stain (Luchini C, ESMO recommendations on microsatellite instability testing for immunotherapy in cancer, and its relationship with PD-1 / PD-L1 expression and tumour mutational burden: a systematic review-based approach. Ann Oncol. 2019; Lindor NM, Burgart LJ, Leontovich O, Goldberg RM, Cunningham JM, Sargent DJ, Walsh-Vockley C, Petersen GM, Walsh MD, Leggett BA, Young JP, Barker MA, Jass JR, Hopper J, Gallinger S, Bapat B, Redston M, Thibodeau SN. Immunohistochemistry versus microsatellite instability testing in phenotyping colorectal tumors. J Clin Oncol. 2002 Feb 15;20(4):1043-8. doi: 10.1200 / JC0.2002.20.4.1043. PMID: 11844828).

[0100] In colorectal cancers in general, regardless of MMR status, the invention establishes a level of CD74 expression which can stratify subjects according to response to immunotherapy. Subjects with CD74 expression higher than a reference value can be classified as immunotherapy responders, and subjects with CD74 expression lower than a reference value can be classified as immunotherapy nonresponders.

[0101] Tumour mutational burden

[0102] Tumour mutational burden refers to the number of mutations per megabase (muts / Mb) harboured by tumour cells in a given neoplasm. A tumour mutational burden of 12 or more muts / Mb is generally considered a high tumour mutational burden in colorectal cancer (Cancer Genome Atlas N.

[0103] Comprehensive molecular characterization of human colon and rectal cancer. Nature 2012;487:330- 337). A threshold of 10 muts / Mb has been identified as the threshold for eligibility to immunotherapy, regardless of cancer type (Marabelle et al Association of tumour mutational burden with outcomes in patients with advanced solid tumours treated with pembrolizumab: prospective biomarker analysis of the multicohort, open-label, phase 2 KEYNOTE-158 study Lancet 2020 DOI: https : / / doi .org / 10.1016 / S1470-2045(20)30445-9). When described herein, it will be appreciated that the term “low tumour mutational burden” generally refers to subjects having colorectal cancers which have a tumour mutational burden below the threshold for receiving immunotherapy, i.e. less than 10, as established in Marabelle et al. As discussed elsewhere herein, such subjects would typically be considered non-responders to immunotherapy. Low tumour mutational burden as used herein generally refers to fewer than 10 muts / Mb. Subjects with low tumour mutational burden may have fewer than 10, fewer than 9, fewer than 8, fewer than 7, fewer than 6, fewer than 5, fewer than 4, fewer than 3, fewer than 2, or fewer than 1 muts / Mb. Low tumour mutational burden may variously be referred to herein as low TMB, TMB-L, pMMR, or microsatellite low. High tumour mutational burden may variously be referred to herein as high TMB, TMB-H, dMMR, microsatellite high, or hypermutated.

[0104] Subjects described herein may have a low TMB. Generally, this means the subject has been identified as having a tumour with a low tumour mutational burden; has not been identified as having a tumour with a high tumour mutational burden; and / or has been identified as not having a tumour with a high tumour mutational burden. TMB can be assessed by any suitable method known in the art, including but not limited to next generation sequencing, for example as described in Cancer Genome Atlas N. Comprehensive molecular characterization of human colon and rectal cancer. Nature 2012;487:330-337, or Lawlor RT et al. (2021) Tumor Mutational Burden as a Potential Biomarker for Immunotherapy in Pancreatic Cancer: Systematic Review and Still-Open Questions, Cancers (Basel). 13: 3119.

[0105] Colorectal cancer

[0106] Described herein are methods for selecting a subject for treatment with an immunotherapeutic agent, wherein the subject has colorectal cancer. By “has colorectal cancer”, it is generally meant that the subject has been diagnosed with colorectal cancer. The subject may have been identified as having colorectal cancer. The subject may be currently suffering from colorectal cancer. The subject may have previously been diagnosed with colorectal cancer but since entered remission. The subject may be at risk of developing colorectal cancer. The subject may be at risk of having a recurrence of colorectal cancer. The colorectal cancer may be at any clinical stage (Stage 0, Stage 1 , Stage 2, Stage 3, Stage 4, etc.), for example the subject may have metastatic colorectal cancer.

[0107] By “colorectal cancer”, it is generally meant any type of cancer that affects the bowel. Colorectal cancer includes slow-growing or low-grade cancers, or fast-growing or high-grade cancers. The colorectal cancer may be small bowel cancer, large bowel cancer, colon cancer, rectal cancer, anal cancer, or any combination thereof. The colorectal cancer may be an adenocarcinoma, a mucinous tumour, a signet ring tumour, a squamous cell tumour, a carcinoid tumour, a sarcoma, a lymphoma, a melanoma, or any combination thereof.

[0108] The subject may be undergoing chemotherapy. The subject may have previously undergone chemotherapy. The subject may be chemotherapy-naive. The subject may have not previously been treated with an immunotherapeutic agent.

[0109] Other cancers

[0110] CD74 expression may also be used to distinguish immunotherapy responders from non-responders in other cancer types, aside from colorectal cancer. In particular, the methods described herein may also be used to predict the response of a subject to treatment with an immunotherapeutic agent, wherein the subject has endometrial cancer, stomach cancer, or oesophageal cancer. Like colorectal cancer, these cancer types can be divided into dMMR and pMMR tumours and therefore may benefit from immunotherapy if the expression of CD74 exceeds a certain value. In this way, CD74 expression may be used as a cancer-agnostic test to identify those patients who will respond to treatment with an immunotherapeutic agent. CD74

[0111] CD74 (also known as invariant chain protein or li) is a transmembrane protein expressed by cells including macrophages, dendritic cells, B cells, epithelial cells and endothelial cells. CD74 is known for its function in antigen presentation and formation of MHC II and MHC I complexes. As described herein, determining a level of CD74 expression in a sample (in particular a tumour sample) obtained from a subject with colorectal cancer can be used to establish whether that subject will be responsive to immunotherapy, such as anti-PD-1 immunotherapy. Surprisingly, within a patient group typically considered as non-responders to immunotherapy (low TMB / pMMR colorectal cancer), CD74 expression can be used to identify subjects who will respond to immunotherapy.

[0112] As described herein, CD74 expression is assessed in a sample obtained from a subject. Generally, the level of CD74 expression is then compared to reference value (or cut-off or threshold value). If the level of CD74 expression is higher than the reference value, the subject is predicted or assessed to be responsive to immunotherapy (an “immunotherapy responder”). If the level of CD74 expression is lower than the reference value, the subject is predicted or assessed to be non-responsive to immunotherapy (an “immunotherapy non-responder”). In general, a reference value as described herein may be 10%, 11%, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21 %, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31%, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41%, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%, in particular 15%, 16%, 17%, 18%, 19%, 20%, 21%, 22%, or 23%, in particular 15%, 17%, 19%, 21%, or 23%, in particular 19%.

[0113] References herein to “higher than” also generally include “higher than or equal to”. For example, if the reference value is 19%, a level of CD74 expression of 19% or higher would indicate an immunotherapy responder, and a level of CD74 expression of 18% would indicate an immunotherapy non-responder.

[0114] The level of CD74 expression may be determined by determining a percentage of cells in the sample that express CD74. The reference value may therefore correspond to a percentage (%) of cells in the sample that express CD74. The reference value may be 10%, 11%, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21%, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31%, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41%, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21%, or greater than 23% of the cells in the sample express CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%. The level of CD74 expression may be determined by determining a percentage of the sample that is positive for CD74. The reference value may therefore correspond to a percentage (%) of the sample that is positive for CD74. The reference value may be 10%, 11%, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21%, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31%, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41%, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21%, or greater than 23% of the sample is positive for CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%.

[0115] In many cases, the sample will be a tumour sample, such as a tumour biopsy. Tumours (and corresponding tumour samples) can generally be considered to comprise multiple compartments, such as a tumour epithelial compartment and a tumour stromal compartment. Determining the level of CD74 expression may comprise determining CD74 expression in one, two, or more tumour compartments. The tumour epithelial compartment is generally considered to comprise the neoplastic tumour cells. The tumour-associated stroma (“tumour stroma”) is a critical component of the tumour microenvironment, and can substantially influence tumour growth, metastasis and therapeutic resistance. The tumour stroma is typically considered to comprise all non-neoplastic cells of the bulk lesion, such as immune cells (such as macrophages), fibroblasts, mesenchymal stromal cells, osteoblasts, chondrocytes, the extracellular matrix, endothelial cells, pericytes, and adipocytes. CD74 expression can be assessed across the whole tumour stroma. Alternatively, CD74 expression by a specific cell type within the tumour stroma can be assessed, such as CD74 expression by macrophages. CD74 expression can also be assessed in the tumour microenvironment (TME).

[0116] The level of CD74 expression may be determined by determining a percentage of stromal cells in the sample that express CD74. The reference value may be a percentage of stromal cells in the sample that express CD74. The reference value may be 10%, 11%, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21%, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31%, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41%, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21%, or greater than 23% of the stromal cells in the sample express CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%.

[0117] The level of CD74 expression may be determined by determining a percentage of the stroma that is positive for CD74. The reference value may be a percentage of the stroma that is positive for CD74. The reference value may be 10%, 11%, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21%, 39%, 40%, 41 %, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21 %, or greater than 23% of the stroma is positive for CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%.

[0118] The level of CD74 expression may be determined by determining a percentage of tumour stromal cells in the sample that express CD74. The reference value may be a percentage of tumour stromal cells in the sample that express CD74. The reference value may be 10%, 11 %, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21 %, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31 %, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41 %, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21 %, or greater than 23% of the tumour stromal cells in the sample express CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%.

[0119] The level of CD74 expression may be determined by determining a percentage of the tumour stroma that is positive for CD74. The reference value may be a percentage of the tumour stroma that is positive for CD74. The reference value may be 10%, 11 %, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21 %, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31 %, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41 %, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21 %, or greater than 23% of the tumour stroma is positive for CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%.

[0120] The level of CD74 expression may be determined by determining a percentage of tumour epithelial cells in the sample that express CD74. The reference value may be a percentage of tumour epithelial cells in the sample that express CD74. The reference value may be 10%, 1 1 %, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21 %, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31 %, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41 %, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21 %, or greater than 23% of the tumour epithelial cells in the sample express CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%.

[0121] The level of CD74 expression may be determined by determining a percentage of the tumour epithelium that is positive for CD74. The reference value may be a percentage of the tumour epithelium that is positive for CD74. The reference value may be 10%, 11 %, 12%, 13%, 14%, 15%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41%, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21%, or greater than 23% of the tumour epithelium is positive for CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%.

[0122] The level of CD74 expression may be determined by determining a percentage of tumour cells in the sample that express CD74. The reference value may be a percentage of tumour cells in the sample that express CD74. The reference value may be 10%, 11%, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21%, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31%, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41%, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21%, or greater than 23% of the tumour cells in the sample express CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%.

[0123] The level of CD74 expression may be determined by determining a percentage of the tumour that is positive for CD74. The reference value may be a percentage of the tumour that is positive for CD74. The reference value may be 10%, 11%, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21%, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31%, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41%, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21%, or greater than 23% of the tumour is positive for CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%.

[0124] The level of CD74 expression may be determined by determining a percentage of cells in the tumour microenvironment of the sample that express CD74. The reference value may be a percentage of cells in the tumour microenvironment of the sample that express CD74. The reference value may be 10%, 11%, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21%, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31%, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41%, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21%, or greater than 23% of the cells in the tumour microenvironment of the sample express CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%.

[0125] The level of CD74 expression may be determined by determining a percentage of the tumour microenvironment that is positive for CD74. The reference value may be a percentage of the tumour microenvironment that is positive for CD74. The reference value may be 10%, 11%, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21 %, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31 %, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41 %, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21 %, or greater than 23% of the tumour microenvironment is positive for CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%.

[0126] Determining the level of CD74 expression may comprise determining CD74 expression in a particular tumour compartment, normalised to total compartment, total tumour, or total sample.

[0127] The level of CD74 expression may be calculated as a CD74 score, with the reference value also as a CD74 score. For example, CD74 expression may be quantified in tumour stromal cells, tumour epithelial cells, and / or total tumour cells, and normalised to other values to generate a CD74 score.

[0128] In some cases, the CD74 score is a combined positive score (CPS). The CPS is calculated by combining the CD74+ve tumour stromal cells and CD74+ve tumour epithelial cells, and dividing by all the detected cells in the tumour sample. The reference value may be a CPS of 10%, 11 %, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21 %, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31%, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41 %, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if the CPS is greater than 15%, greater than 17%, greater than 19%, greater than 21 %, or greater than 23%. In some cases, the reference value is a CPS of 17%. In some cases, the reference value is a CPS of 19%.

[0129] In some cases, the CD74 score is a stroma positive against tumour (SPT) score. The SPT score is calculated as the CD74+ve stromal cells divided by all detected tumour epithelial cells. The reference value may be an SPT of 10% , 11 % , 12% , 13% , 14% , 15% , 16% , 17% , 18% , 19% , 20% , 21 % , 22% , 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31 %, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41 %, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if the SPT is greater than 15%, greater than 17%, greater than 19%, greater than 21 %, or greater than 23%. In some cases, the reference value is an SPT of 17%. In some cases, the reference value is an SPT of 19%.

[0130] In some cases, the CD74 score is a tumour positive score (TPS). The TPS is calculated as the CD74+ve tumour epithelial cells divided by all the detected tumour epithelial cells. The reference value may be a TPS of 10%, 11 %, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21 %, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31 %, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41 %, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if the TPS is greater than 15%, greater than 17%, greater than 19%, greater than 21 %, or greater than 23%. In some cases, the reference value is a TPS of 17%. In some cases, the reference value is a TPS of 19%.

[0131] In some cases, the CD74 score is a stroma positive score (SPS). The SPS is calculated as the CD74+ve stromal cells divided by all the stromal cells. The reference value may be an SPS of 10%, 11%, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21 %, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31 %, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41 %, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if the SPS is greater than 15%, greater than 17%, greater than 19%, greater than 21 %, or greater than 23%. In some cases, the reference value is an SPS of 17%. In some cases, the reference value is an SPS of 19%.

[0132] The level of CD74 expression may be determined by determining a percentage of macrophages in the sample that are positive for CD74. The reference value may be a percentage of macrophages in the sample that are positive for CD74. The reference value may be 10%, 1 1 %, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21 %, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31 %, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41 %, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21 %, or greater than 23% of macrophages in the sample are positive for CD74. In many cases, the reference value is 19%.

[0133] The level of CD74 expression may be determined by determining a percentage of macrophages in the sample that express CD74. The reference value may be a percentage of macrophages in the sample that express CD74. The reference value may be 10%, 11 %, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21 %, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31 %, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41 %, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21 %, or greater than 23% of macrophages in the sample express CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%.

[0134] The level of CD74 expression may be determined by determining a percentage of macrophages in the tumour stroma that express CD74. The reference value may be a percentage of macrophages in the tumour stroma that express CD74. The reference value may be 10%, 11 %, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21 %, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31 %, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41 %, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21 %, or greater than 23% of macrophages in the tumour stroma express CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%.

[0135] The level of CD74 expression may be determined by determining a percentage of the tumour margin that is positive for CD74. The reference value may be a percentage of the tumour margin that is positive for CD74. The reference value may be 10%, 11 %, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21 %, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31 %, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41 %, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21 %, or greater than 23% of the tumour margin that is positive for CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%.

[0136] The level of CD74 expression may be determined by determining a percentage of cells in the tumour margin that are positive for CD74. The reference value may be a percentage of cells in the tumour margin that are positive for CD74. The reference value may be 10%, 11 %, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21 %, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31 %, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41 %, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21 %, or greater than 23% of cells in the tumour margin are positive for CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%.

[0137] The level of CD74 expression may be determined by determining a percentage of cells in the tumour margin that express CD74. The reference value may be a percentage of cells in the tumour margin that express CD74. The reference value may be 10%, 11 %, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21 %, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31 %, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41 %, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21 %, or greater than 23% of cells in the tumour margin express CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%.

[0138] The level of CD74 expression may be determined by determining a percentage of the tumour core that is positive for CD74. The reference value may be a percentage of the tumour core that is positive for CD74. The reference value may be 10%, 11 %, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 38%, 39%, 40%, 41%, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21%, or greater than 23% of the tumour core that is positive for CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%.

[0139] The level of CD74 expression may be determined by determining a percentage of cells in the tumour core that are positive for CD74. The reference value may be a percentage of cells in the tumour core that are positive for CD74. The reference value may be 10%, 11%, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21%, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31%, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41%, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21%, or greater than 23% of cells in the tumour core are positive for CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%.

[0140] The level of CD74 expression may be determined by determining a percentage of cells in the tumour core that express CD74. The reference value may be a percentage of cells in the tumour core that express CD74. The reference value may be 10%, 11%, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, 20%, 21%, 22%, 23%, 24%, 25%, 26%, 27%, 28%, 29%, 30%, 31%, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41%, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%. Accordingly, a subject may be selected for treatment with the immunotherapeutic agent if greater than 15%, greater than 17%, greater than 19%, greater than 21%, or greater than 23% of cells in the tumour core express CD74. In some cases, the reference value is 17%. In some cases, the reference value is 19%.

[0141] In the methods described herein, determining the level of CD74 expression may comprise determining a level of CD74 expression in the tumour stroma. Determining the level of CD74 expression may comprise determining a level of CD74 expression in the tumour epithelium. Determining the level of CD74 expression may comprise determining a level of CD74 expression in the tumour. Determining the level of CD74 expression may comprise determining a level of CD74 expression by stromal cells. Determining the level of CD74 expression may comprise determining a level of CD74 expression by tumour stromal cells. Determining the level of CD74 expression may comprise determining a level of CD74 expression by tumour epithelial cells. Determining the level of CD74 expression may comprise determining a level of CD74 expression by tumour cells. Determining the level of CD74 expression may comprise determining a level of CD74 expression by tumour stromal cells and tumour epithelial cells. Determining the level of CD74 expression may comprise determining a level of CD74 expression by tumour stromal cells and tumour cells. Determining the level of CD74 expression may comprise determining a level of CD74 expression by tumour epithelial cells and tumour cells. Determining the level of CD74 expression may comprise determining a level of CD74 expression by tumour stromal cells, tumour epithelial cells, and tumour cells. Determining the level of CD74 expression may comprise determining a level of CD74 expression by macrophages in the sample.

[0142] Determining the level of CD74 expression may comprise determining a level of CD74 expression by macrophages in the sample. Determining the level of CD74 expression may comprise determining a level of CD74 expression by macrophages in the tumour stroma. Determining the level of CD74 expression may comprise determining a level of CD74 expression by macrophages in the tumour.

[0143] Generally, the level of CD74 expression may comprise a level of CD74 protein, for example where CD74 is assessed by immunohistochemistry, ELISA or similar. The level of CD74 expression may comprise a level of cell surface expression of CD74.

[0144] In some cases, the level of CD74 expression may comprise a level of CD74 DNA, CD74 RNA, or a combination thereof.

[0145] The method of determining the level of CD74 expression may comprise immunohistochemistry, gene expression analysis, flow cytometry, proteomics, transcriptomics, in situ hybridisation, Western blot, in- situ immunofluorescence, imaging mass cytometry, ELISA, or any combination thereof. For example, in the case of immunohistochemistry, tissue sections from a tumour sample may be stained or labelled with a commercially available anti-CD74 antibody in order to detect CD74 expression in the sample. The level of CD74 expression in the sample can then be detected and quantified using conventional methods in the art, such as microscopy or fluorescence microscopy.

[0146] Also described herein is the use of an anti-CD74 antibody in a method of predicting or assessing the response of a subject to treatment with an immunotherapeutic agent, wherein the subject has mismatch repair proficient (pMMR) colorectal cancer, the method comprising: a. providing a sample obtained from the subject, b. detecting CD74 expression in the sample, wherein said detecting comprises treating the sample with the anti-CD74 antibody, and c. determining a level of CD74 expression in the sample and comparing it to a reference value; wherein a CD74 expression level higher than the reference value is indicative of an immunotherapy responder, and wherein a CD74 expression level lower than the reference value is indicative of an immunotherapy non-responder.

[0147] Predicting the response to treatment with an immunotherapeutic agent Described herein are methods for predicting the response of a subject to treatment with an immunotherapeutic agent. Also described herein are methods for assessing the response of a subject to treatment with an immunotherapeutic agent. In general, the subject has pMMR colorectal cancer. The subject may have colorectal cancer with a low tumour mutational burden. The methods described herein can be used to identify “responders” and “non-responders” to immunotherapy, such as anti- PD1 immunotherapy. A responder to immunotherapy or treatment with an immunotherapeutic agent may show complete or partial response to treatment with the immunotherapeutic agent, such as a reduction in tumour size, or complete remission. A non-responder to immunotherapy or treatment with an immunotherapeutic agent may show stable or progressive disease following treatment with the immunotherapeutic agent, such as tumour growth, metastases, or no change in tumour size. “Stable” and “progressive” disease are defined clinically from the RECIST criteria (Eisenhauer EA, Therasse P, Bogaerts J, Schwartz LH, Sargent D, Ford R, Dancey J, Arbuck S, Gwyther S, Mooney M, et al: New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer 2009, 45:228-247). The method may comprise determining a level of CD74 expression in a sample obtained from the subject and comparing it to a reference value, wherein a CD74 expression level higher than the reference value is indicative of an immunotherapy responder, and wherein a CD74 expression level lower than the reference value is indicative of an immunotherapy non-responder.

[0148] Selecting a subject for treatment with an immunotherapeutic agent

[0149] The methods described herein use expression of CD74 as a biomarker to select subjects with colorectal cancer for treatment with an immunotherapeutic agent. By identifying responders and non- responders to immunotherapy based on the expression of CD74, the methods described herein can stratify patients and select only those predicted responders for subseguent treatment with immunotherapy, such as anti-PD-1 / PD-L1 immunotherapy.

[0150] Described herein are methods for selecting a subject for treatment with an immunotherapeutic agent, wherein the subject has pMMR colorectal cancer, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; and b. comparing the level of CD74 expression in the sample to a reference value; wherein the subject is selected for treatment with the immunotherapeutic agent if the level of CD74 in the sample is higher than the reference value.

[0151] Described herein are methods for selecting a subject for treatment with an immunotherapeutic agent, wherein the subject has colorectal cancer with a low tumour mutational burden, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; and b. comparing the level of CD74 expression in the sample to a reference value; wherein the subject is selected for treatment with the immunotherapeutic agent if the level of CD74 in the sample is higher than the reference value.

[0152] Described herein is a method of selecting a subject for treatment with anti-PD-1 or anti-PD-L1 immunotherapy, wherein the subject has mismatch repair proficient (pMMR) colorectal cancer, the method comprising: a. determining a percentage of cells in a tumour sample obtained from the subject that express CD74, and b. selecting the subject for treatment with anti-PD-1 or anti-PD-L1 immunotherapy if greater than 19% of cells in the tumour sample express CD74.

[0153] Methods for selecting a subject for treatment with an immunotherapeutic agent may optionally further comprise a step of administering the immunotherapeutic agent to the subject to treat the colorectal cancer if the level of CD74 in the sample is higher than the reference value.

[0154] An immunotherapeutic agent as described herein may comprise an immune checkpoint inhibitor. The immunotherapeutic agent may comprises an agent that acts on the PD-1 / PD-L1 axis. The immunotherapeutic agent may comprise a PD-1 inhibitor, a PD-L1 inhibitor, or a combination thereof. The immunotherapeutic agent may comprise an anti-PD-1 antibody, an anti-PD-L1 antibody, or a combination thereof. The immunotherapeutic agent may comprise an antagonistic anti-PD-1 antibody, an antagonistic anti-PD-L1 antibody, or a combination thereof. The immunotherapeutic agent may comprise pembrolizumab, nivolumab, atezolizumab, or a combination thereof.

[0155] Although generally the immunotherapeutic agents described herein refer to agents that act on the PD- 1 axis or PD-1 / PD-L1 axis, it will be appreciated that other immunotherapeutic agents, such as CTLA- 4 inhibitors (in particular anti-CTLA-4 antibodies) may also be useful in the approaches described herein. Accordingly, the immunotherapeutic agent may comprise a CTLA-4 inhibitor, such as an anti- CTLA-4 antibody. The immunotherapeutic agent may comprise ipilimumab, tremelimumab, or a combination thereof. In some cases, the immunotherapeutic agent may comprise a combination of PD-1 inhibitor, PD-L1 inhibitor and / or CTLA-4 inhibitor.

[0156] Methods of treatment

[0157] The methods described herein may further comprise a step of administering an immunotherapeutic agent to the subject. If the subject is identified as having CD74 expression higher than the reference value, the method may further comprise administering a therapeutically effective amount of the immunotherapeutic agent to the subject. For example, the methods described herein may further comprise administering a therapeutically effective amount of the immunotherapeutic agent to the subject if greater than (e.g.) 19% of cells in a tumour sample obtained from the subject express CD74. The method may comprise administering a therapeutically effective amount of the immunotherapeutic agent to the subject such that the colorectal cancer is treated.

[0158] Described herein are methods of selecting a treatment regimen for a subject having mismatch repair proficient (pMMR) colorectal cancer, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; b. comparing the level of CD74 expression in the sample to a reference value; and c. selecting a treatment regimen for the subject if the level of CD74 expression in the sample is higher than the reference value.

[0159] The treatment regimen may comprise treatment with an immunotherapeutic agent, such as an immune checkpoint inhibitor.

[0160] Described herein are methods of selecting a treatment regimen for a subject having low TMB colorectal cancer, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; b. comparing the level of CD74 expression in the sample to a reference value; and c. selecting a treatment regimen for the subject if the level of CD74 expression in the sample is higher than the reference value.

[0161] The treatment regimen may comprise treatment with an immunotherapeutic agent, such as an immune checkpoint inhibitor.

[0162] Described herein are methods of stratifying mismatch repair proficient (pMMR) colorectal cancer patients by response to immunotherapy, the method comprising: a. determining a level of CD74 expression in samples obtained from the patients; and b. comparing the level of CD74 expression in the samples to a reference value; wherein patients with a CD74 expression level higher than the reference value are classified as responders, and patients with a CD74 expression level lower than the reference value are classified as non-responders.

[0163] Described herein are methods of stratifying low TMB colorectal cancer patients by response to immunotherapy, the method comprising: a. determining a level of CD74 expression in samples obtained from the patients; and b. comparing the level of CD74 expression in the samples to a reference value; wherein patients with a CD74 expression level higher than the reference value are classified as responders, and patients with a CD74 expression level lower than the reference value are classified as non-responders.

[0164] Described herein are inhibitors of PD-1 and / or PD-L1 for use in methods of treating colorectal cancer in a subject. Generally, the subject has been identified as having pMMR colorectal cancer. The subject may have a low tumour mutational burden. The subject may have been identified as having CD74 expression levels that are higher than a reference value, for example the subject may have been identified as having a CD74 score greater than 19%, or having greater than 19% of a tumour sample express CD74. Accordingly, described herein is a PD-1 inhibitor for use in a method of treating pMMR colorectal cancer in a subject, wherein the subject has been identified as having a level of CD74 expression that is higher than a reference value. Also disclosed herein is a PD-L1 inhibitor for use in a method of treating pMMR colorectal cancer in a subject, wherein the subject has been identified as having a level of CD74 expression that is higher than a reference value. Also described herein is a PD-1 inhibitor for use in a method of treating pMMR colorectal cancer, the method comprising determining a level of CD74 expression in a sample obtained from the subject and administering a therapeutically effective amount of the PD-1 inhibitor to the subject if the level of CD74 expression in the sample is higher than a reference value. Also described herein is a PD-L1 inhibitor for use in a method of treating pMMR colorectal cancer, the method comprising determining a level of CD74 expression in a sample obtained from the subject and administering a therapeutically effective amount of the PD-L1 inhibitor to the subject if the level of CD74 expression in the sample is higher than a reference value.

[0165] The methods described herein may comprise administering a therapeutically effective amount of the immunotherapeutic agent to the subject in combination with a further therapeutic agent. The further therapeutic agent may comprise a further immunotherapeutic agent, a chemotherapeutic agent, an anti-angiogenic agent, a growth inhibitor, or a combination thereof.

[0166] Methods of analysing a sample

[0167] Provided herein is a method, comprising: a. providing a sample obtained from a subject having mismatch repair proficient (pMMR) colorectal cancer; b. detecting CD74 expression in the sample; and c. analysing the CD74 expression in the sample, wherein said analysing comprises calculating a CD74 stroma positive score (SPS).

[0168] Also provided herein is a method comprising: a. providing a sample obtained from a subject having low TMB colorectal cancer; b. detecting CD74 expression in the sample; and c. analysing the CD74 expression in the sample, wherein said analysing comprises calculating a CD74 stroma positive score (SPS).

[0169] In the methods described herein, said detecting may comprise treating the sample with an agent that labels CD74, such as an anti-CD74 antibody. Said detecting may be performed using immunohistochemistry.

[0170] The approaches described herein may involve determining a level of CD74 expression in a sample obtained from a subject. In general, the sample is a tumour sample, such as a tumour biopsy. However, the sample may also be whole blood, serum, plasma, blood cells, circulating tumor cells, urine, or any combination thereof. Generally, the sample has been previously obtained from a subject.

[0171] The term “subject” generally refers to an animal. The subject may be a mammalian subject, such as a human.

[0172] The terms “treatment” and “treating,” as used herein, are intended to refer to all processes wherein there may be a slowing, interrupting, arresting or stopping of the progression of a disorder, or amelioration of one or more symptoms thereof, but does not necessarily indicate a total elimination of all symptoms.

[0173] The term “therapeutically effective amount” as used herein generally means an amount of active compound or pharmaceutical agent that elicits the biological or medicinal response in a tissue system, animal or human that is being sought by a researcher, veterinarian, medicinal doctor or other clinician, which includes alleviation or reversal of the symptoms of the disease or disorder being treated.

[0174] MMR-aqnostic methods

[0175] In general, the methods described herein relate to pMMR or low TMB colorectal cancer. However, it will be appreciated that determining a level of CD74 expression can surprisingly be used to classify any colorectal cancer patient as an immunotherapy responder or non-responder.

[0176] As such, the present disclosure also encompasses a method for predicting or assessing the response of a subject to treatment with an immunotherapeutic agent, wherein the subject has colorectal cancer, the method comprising: determining a level of CD74 expression in a sample obtained from the subject and comparing it to a reference value, wherein a CD74 expression level higher than the reference value is indicative of an immunotherapy responder, and wherein a CD74 expression level lower than the reference value is indicative of an immunotherapy non-responder.

[0177] Corresponding methods of selecting a subject for treatment with an immunotherapeutic agent, methods of treating a colorectal cancer, and methods of stratifying colorectal cancer patients, wherein the subject has either pMMR or dMMR, or the MMR status is not known, are also within the scope of this disclosure.

[0178] Aspects and embodiments described herein with the term “comprising” may include other features or steps within the scope. It is also understood that aspects and embodiments described as “comprising” also describes aspect and embodiments wherein the term “comprising” is replaced by the term “consisting essentially of’ or “consisting of’.

[0179] The phrase "selected from the group comprising" may be substituted with the phrase "selected from the group consisting of and vice versa, wherever they occur herein.

[0180] It is also understood that the application discloses all combinations of any of the above aspects and embodiments described above with each other, unless the context demands otherwise. Similarly, the application discloses all combinations of the preferred and / or optional features either singly or together with any of the other aspects, unless the context demands otherwise.

[0181] The invention will now be further described by way of the following Examples, which are meant to serve to assist one of ordinary skill in the art in carrying out the invention and are not intended in any way to limit the scope of the invention, with reference to the Figures.

[0182] EXAMPLES

[0183] Materials & Methods

[0184] Patient and sample cohort

[0185] FFPE blocks of 59 primary tumour resections were obtained from 58 CRC patients consented at the UCL Cancer Institute Pathology Biobank - REC reference 15 / YH / 0311 . All patients were chemo-naTve at the time of resection, except ten who received neoadjuvant treatment. Twenty-three patients (corresponding to twenty-four tumours) received anti-PD1 immunotherapy in the metastatic setting with outcomes assessed according to the RECIST1.1 criteria (Eisenhauer EA, Therasse P, Bogaerts J, Schwartz LH, Sargent D, Ford R, Dancey J, Arbuck S, Gwyther S, Mooney M, et al: New response evaluation criteria in solid tumours: revised RECIST guideline (version 1 .1). Eur J Cancer 2009, 45:228-247). Patients were considered responders if post-treatment lesions showed complete or partial responses, and as non-responders if they showed stable or progressive disease.

[0186] For validation of CD74 SPS in dMMR CRCs, unstained dMMR CRC sections from 43 patients who received anti-PD-1 ICI in the metastatic setting alone or in combination with anti-CTLA-4 ICI were provided by the GONO foundation (REC reference EM 186_2017 granted by Comitato Etico Istituto Oncologico Veneto, Padua, Italy). For the AtezoTRIBE cohort, FFPE unstained slides from 124 unresectable metastatic pMMR CRC patients treated with FOLFOXIRI (chemotherapy) and bevacizumab (anti-VEGF antibody) alone or in combination with atezolizumab (anti-PD-L1 ICI) were received from the GONO foundation (REC ref. 13582_2019, granted by Comitato Etico Area Vasta Nord Ovest, Pisa, Italy). Informed consent was obtained from all patients participating in this study.

[0187] Targeted gene re-sequencing and mapping of mutated pathways

[0188] Mutation and TMB data for 35 and five tumours were obtained from the Sarah Cannon Research Institute UK and (Bortolomeazzi M, Keddar MR, Montorsi L, Acha-Sagredo A, Benedetti L, Temelkovski D, Choi S, Petrov N, Todd K, Wai P, et al: Immunogenomics of Colorectal Cancer Response to Checkpoint Blockade: Analysis of the KEYNOTE 177 Trial and Validation Cohorts. Gastroenterology 2021 , 161 :1179-1193), respectively, while 324 consensus genes (Figure 2) were sequenced in the remaining 19 tumours using a custom panel (Twist Biosciences). For this, three 10pm sections per tumour were macro-dissected with a needle under a stereo microscope. Genomic DNA was extracted using QIAamp DNA FFPE Advanced UNG kit (Qiagen) and libraries were prepared using 0.5-1 pg of genomic DNA with the Twist Bioscience Human Exon kit (NEB Library Prep) and sequenced using a 100bp paired-end protocol on Illumina NovaSeq 6000 at 300x coverage per sample. Because matched normal was not available, somatic nucleotide variants (SNVs) and indels were identified with two independent approaches. In both approaches, reads were aligned using GRCh38. SNVs, and indels were then identified using Strelka v2.9.2 (Kim S, Scheffler K, Halpern AL, Bekritsky MA, Noh E, Kallberg M, Chen X, Kim Y, Beyter D, Krusche P, Saunders CT: Strelka2: fast and accurate calling of germline and somatic variants. Nat Methods 2018, 15:591-594) and annotated using ANNOVAR v16.04.2018 (Wang K, Li M, Hakonarson H: ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res 2010, 38:e164), against a single unmatched normal. Only damaging mutations (truncating stop-gain, stop-loss and frameshift alterations, indels and missense SNVs) identified by both approaches were retained and damaged genes were defined as those containing >1 damaging mutation. TMB was calculated as the number of damaging variants over sequenced target region length according to standard methods in the art.

[0189] All 324 consensus genes were mapped to pathways from levels 1-3 of KEGG v.94.1 (Kanehisa M, Furumichi M, Tanabe M, Sato Y, Morishima K: KEGG: new perspectives on genomes, pathways, diseases and drugs. Nucleic Acids Res 2017, 45:D353-D361) (274 genes), levels 1-3 of Reactome v.72 (Fabregat A, Jupe S, Matthews L, Sidiropoulos K, Gillespie M, Garapati P, Haw R, Jassal B, Korninger F, May B, et al: The Reactome Pathway Knowledgebase. Nucleic Acids Res 2018, 46:D649-D655) (289 genes), and MSigDB Hallmark databases (Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES, Mesirov JP: Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 2005, 102:15545-15550) (186 / 324 genes), for a total of 257 mapped genes (79%). To reduce redundancy, pathways from each database were manually integrated using common keywords and the original mapping genes intersected and merged to obtain 20 pathways containing the 257 / 324 genes.

[0190] Haematoxylin and eosin (H&E) and immunohistochemistry staining.

[0191] H&E and CD3 (clone LN10, Leica Biosystems, 1 :100 dilution), CD57 (clone NK1 , Leica Biosystems ready to use), CD74 (clone LN2, BioLegend, 1 :300 dilution) and Granzyme B (clone 11 F1 , 1 :100) stains in human samples from UCLH, GONO, and AtezoTribe cohorts were performed on 4pm thick slides using a BOND autostainer (Leica Biosystems) as per manufacturer instruction. CD68 staining (clone KP1 , Biolegend, 1 :2000 dilution) was performed manually upon slide dewaxing and heat- induced epitope retrieval (HIER) using Antigen Retrieval Reagent-Basic (R&D Systems). Tissues were blocked by incubation and subsequently with horseradish peroxidase conjugated anti-mouse antibody (Abeam). Slides were then stained with 3,3' diaminobenzidine (DAB) substrate (Abeam) and haematoxylin. Digital acquisition of stained slides was performed using Axioscan Z1 (Zeiss) at 20x resolution. A board-certified surgical pathologist (M.R.J.) used H&E and CD3 stains to assess tumour content and select regions at the core and invasive margin with good T cell infiltration. Meanwhile, CD74 staining on mouse tumour samples was performed manually on 7pm frozen slides. After initial incubation with PBS at room temperature, tissues were blocked with Goat F (ab) Anti-Mouse IgG H&L (1 ; 1000), incubated with anti-CD74 (SC-6262, 1 :1600) followed by horseradish peroxidase conjugated anti-rabbit / anti-mouse secondary antibody (Abeam) and finally stained with 3,3’ diaminobenzidine (DAB) substrate (Abeam) and haematoxylin.

[0192] To detect and quantify CD3, CD68, CD57, and CD74 positive cells, whole slide images were loaded into QuPath v0.5.0 (Bankhead P, Loughrey MB, Fernandez JA, Dombrowski Y, McArt DG, Dunne PD, McQuaid S, Gray RT, Murray LJ, Coleman HG, et al: QuPath: Open source software for digital pathology image analysis. Sci Rep 2017, 7:16878) and the “Estimate Stain Vector” function was run as pre-processing step to increase the contrast between DAB and haematoxylin. The outlines of the core and margin regions delimited by the pathologist in the H&E slide were projected in all other slides. A classifier was trained on at least three tumour stroma (Ts) and tumour epithelium (Te) samples from representative CRC blocks. Where the classifier was inadequate, additional Te and Ts samples were used. For mouse samples, necrotic regions were identified in an adjacent H&E slide and removed from the cell detection areas. QuPath “positive cell detection” was used on core and invasive margin regions (CD3, CD68, CD57, GZMB) or whole slides (CD74), or non-necrotic areas (mouse CD74).

[0193] Laser-capture microdissection and RNA sequencing

[0194] Ten-micron FFPE sections were mounted and air-dried overnight onto PEN-Membrane slides (Leica Microsystems). Slides were dewaxed using xylene, rehydrated through graded alcohols, and stained with haematoxylin, blueing agent and eosin. Graded alcohol and xylene hydration was reversed to dehydrate the stained slides, which were then airdried and stored at -80 °C until use. Cutting outlines were drawn to separate 2mm2areas of Te and Ts within the core and invasive margin regions using a H&E slide as a reference. A total of 235 laser captured areas were collected on the cap of Axygen 0.2 ml tubes (Corning Life Sciences) using the LMD 7000 (Leica Microsystems).

[0195] Total RNA was extracted from the 235 LCM samples using the High Pure FFPEt DNA isolation kit (Roche). cDNA libraries were prepared in duplicate from 2ng of total RNA using the NuGEN Ovation V2 system (Tecan) and sequenced using a 100bp paired-end protocol on Illumina NovaSeq 6000 according to manufacturer’s instructions to obtain >20 million reads per sample. Demultiplexed FASTQ files from independent library duplicates were concatenated and used for QC, read pseudo-alignment and count matrix generation with the nf-core rnaseq v3.2 pipeline (Ewels PA, Peltzer A, Fillinger S, Patel H, Alneberg J, Wilm A, Garcia MU, Di Tommaso P, Nahnsen S: The nf-core framework for community-curated bioinformatics pipelines. Nat Biotechnol 2020, 38:276-278). Reads were aligned to the GRCh38 genome using the “star_salmon” aligner with default options. Salmon raw counts were mapped to 19756 human genes (Dressier L, Bortolomeazzi M, Keddar MR, Misetic H, Sartini G, Acha- Sagredo A, Montorsi L, Wijewardhane N, Repana D, Nulsen J, et al: Comparative assessment of genes driving cancer and somatic evolution in non-cancer tissues: an update of the Network of Cancer Genes (NCG) resource. Genome Biol 2022, 23:35) and RNA sequencing data sparsity was controlled for by applying SMIXnorm (RSeqNorm R package vO.0.0.9000) (Yin S, Zhan X, Yao B, Xiao G, Wang X, Xie Y: SMIXnorm: Fast and Accurate RNA-Seq Data Normalization for Formalin-Fixed Paraffin- Embedded Samples. Front Genet 2021 , 12:650795), with probability threshold of 0.99. Expression data were normalised using estimateSizeFactors and varianceStabilizingTransformation functions of DESeq2 v1 .36.0 (Love Ml, Huber W, Anders S: Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 2014, 15:550) with default parameters.

[0196] Curation of tumour intrinsic and extrinsic signatures

[0197] Fourteen tumour intrinsic gene signatures representing dMMR and pMMR CRC biology and their interaction with the TME were assembled from the literature. In case more than one literature source was available, only genes found in at least two sources were retained and genes shared across signatures were removed. The resulting 14 signatures were validated using independent CRC datasets and tested for significant enrichment in the corresponding KEGG v.94.1 (Kanehisa M, Furumichi M, Tanabe M, Sato Y, Morishima K: KEGG: new perspectives on genomes, pathways, diseases and drugs. Nucleic Acids Res 2017, 45:D353-D361) and MSigDB Hallmark (Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES, Mesirov JP: Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 2005, 102:15545-15550) pathways (Figure 10).

[0198] Extrinsic immune signatures were derived assigning each of the 547 LM22 CIBERSORT genes (Newman AM, Liu CL, Green MR, Gentles AJ, Feng W, Xu Y, Hoang CD, Diehn M, Alizadeh AA: Robust enumeration of cell subsets from tissue expression profiles. Nat Methods 2015, 12:453-457) to one of the 22 immune cell types by highest expression. Genes in each list were then removed if they were more expressed in unrelated than the designated cell types within a single cell human CRC dataset (Pelka K, Hofree M, Chen JH, Sarkizova S, Pirl JD, Jorgji V, Bejnood A, Dionne D, Ge WH, Xu KH, et al: Spatially organized multicellular immune hubs in human colorectal cancer. Cell 2021 , 184:4734-4752 e4720) using the Single Cell Portal and ‘ClusterMidway’ annotation setting (https: / / singlecell.broadinstitute.org / single_cell / study / SCP1162 / human-colon-cancer-atlas-c295). Accuracy of the resulting signatures was further tested on the CIBERSORT LM22 expression data by single sample gene set enrichment analysis (ssGSEA) using the GSVA R package v1 .44.5 (Barbie DA, Tamayo P, Boehm JS, Kim SY, Moody SE, Dunn IF, Schinzel AC, Sandy P, Meylan E, Scholl C, et al: Systematic RNA interference reveals that oncogenic KRAS-driven cancers require TBK1 . Nature 2009, 462:108-112) (Figure 5A). ssGSEA results were also used to merge very similar cell state signatures for B cells, dendritic cells, mast cells, NK cells and CD4 T cells (Figure 5A). Because of its relevance in anti-tumour immunity, a specific cytotoxicity cell signature was assembled to contain CD3D / E / G and CD8A / B genes, the top five genes (CCL3, GNLY, GZMB, HAVCR2, PRF1) representing the CRC-associated T cell cytotoxicity programme (Pelka K, Hofree M, Chen JH, Sarkizova S, Pirl JD, Jorgji V, Bejnood A, Dionne D, Ge WH, Xu KH, et al: Spatially organized multicellular immune hubs in human colorectal cancer. Cell 2021 , 184:4734-4752 e4720 ) and the cytotoxic cells signature from (Tamborero D, Rubio-Perez C, Muinos F, Sabarinathan R, Piulats JM, Muntasell A, Dienstmann R, Lopez-Bigas N, Gonzalez-Perez A: A Pan-cancer Landscape of Interactions between Solid Tumors and Infiltrating Immune Cell Populations. Clin Cancer Res 2018, 24:3717-3728) excluding APOL3, ZBTB16 and WHAMMP3 because they are not expressed CRC associated T cells (Pelka K, Hofree M, Chen JH, Sarkizova S, Pirl JD, Jorgji V, Bejnood A, Dionne D, Ge WH, Xu KH, et al: Spatially organized multicellular immune hubs in human colorectal cancer. Cell 2021 , 184:4734-4752 e4720). This resulted in 23 genes fully contained the LM22 CD8 signature

[0199] (Figure 5B), which therefore was dropped. Similarly, the LM22 signatures for follicular helper, gamma delta and regulatory T cells were not included because signatures resulted in enrichment scores highly over-lapping with closely related cell types. A general T cell signature was derived from the intersection of genes uniquely mapping to KEGG v.94.1 (Kanehisa M, Furumichi M, Tanabe M, Sato Y, Morishima K: KEGG: new perspectives on genomes, pathways, diseases and drugs. Nucleic Acids Res 2017, 45:D353-D36T) T cell receptor signalling pathway and Gene Ontology (accessed via ToppGene v2021-Mar-29) (Chen J, Bardes EE, Aronow BJ, Jegga AG 2009. ToppGene Suite for gene list enrichment analysis and candidate gene prioritization. Nucleic Acids Research doi: 10.1093 / nar / gkp427) T cell pathway (GO: 0050852) that were expressed in CRC-associated T cells (Pelka K, Hofree M, Chen JH, Sarkizova S, Pirl JD, Jorgji V, Bejnood A, Dionne D, Ge WH, Xu KH, et al: Spatially organized multicellular immune hubs in human colorectal cancer. Cell 2021 , 184:4734- 4752 e4720). Finally, an IFN production signature was derived from (Nicolet BP, Guislain A, van Alphen FPJ, Gomez-Eerland R, Schumacher TNM, van den Biggelaar M, Wolkers MC: CD29 identifies IFN-gamma-producing human CD8(+) T cells with an increased cytotoxic potential. Proc Natl Acad Sci U S A 2020, 117:6686-6696). Genes shared by any two signatures were removed. The resulting 15 immune signatures were tested against the corresponding pseudo-bulk profiles from the single cell human CRC dataset (Pelka K, Hofree M, Chen JH, Sarkizova S, Pirl JD, Jorgji V, Bejnood A, Dionne D, Ge WH, Xu KH, et al: Spatially organized multicellular immune hubs in human colorectal cancer. Cell 2021 , 184:4734-4752 e4720) to confirm that the median of their ssGSEA- derived NES distribution was significantly higher than that of the rest of cell types.

[0200] Seven extrinsic stromal signatures composed on at least five genes were taken from (Kieffer Y, Hocine HR, Gentric G, Pelon F, Bernard C, Bourachot B, Lameiras S, Albergante L, Bonneau C, Guyard A, et al: Single-Cell Analysis Reveals Fibroblast Clusters Linked to Immunotherapy Resistance in Cancer. Cancer Discov 2020,10:1330-1351). Genes shared by more than one signature were removed and the resulting eight signatures were tested against the corresponding pseudo-bulk profiles from the single cell human CRC dataset (Pelka K, Hofree M, Chen JH, Sarkizova S, Pirl JD, Jorgji V, Bejnood A, Dionne D, Ge WH, Xu KH, et al: Spatially organized multicellular immune hubs in human colorectal cancer. Cell 2021 , 184:4734-4752 e4720) to confirm that the median of their ssGSEA-derived NES distribution was significantly higher than that of the rest of cell types.

[0201] Sample Clustering

[0202] Visualisation of transcriptomic data was performed using multi-dimensional scaling (plotMDS, limma R package) and plotting the top two dimensions containing the highest variance. Whole transcriptome clustering was achieved through calculation of a distance matrix using a Manhattan distance measure and complete hierarchical clustering (stats R package v4.2.1). Clustering of samples by enrichment signatures was performed within pheatmap (R package v1 .0.12) using default parameters after scaling of signature row scores between zero (minimum) and one (maximum). Human cell lines

[0203] Human colorectal adenocarcinoma cell lines HCT116 (dMMR) and CaCo2 (pMMR) were obtained from Cell Services (Francis Crick Institute) and cultured in low glucose Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) (Gibco). All cell lines were authenticated by short tandem repeat genotyping, tested for mycoplasma and cultured at 37 °C in a humidified 5% carbon dioxide atmosphere.

[0204] Mice and murine cell lines

[0205] Female C57BL / 6 and BALB / c (6- to 8-week-old) mice were purchased from Charles River and housed under specific pathogen-free conditions in individually ventilated cages at a temperature of 21 °C with water and food ad libitum. Prior studies in the laboratory dictated cohort sizes. Animal work was conducted under the Home Office project licence number 135562374 and personal licence number PP3601022. The Ethical Review Committee at King’s College London and Home Office permitted this work.

[0206] CT26 cells (BALB / c background) were sourced from Cell Services at the Francis Crick Institute. MC38 cells (C57BL / 6 background) were kindly gifted by the Sahai Lab (Francis Crick Institute). Cell lines were cultured in Roswell Park Memorial Institute (RPMI) medium (Gibco) supplemented with 10% foetal bovine serum (FBS, Gibco) and incubated at 37°C with 5% carbon dioxide. Cell lines were confirmed to be mycoplasma free using the MycoAlert Mycoplasma Detection Kit (Lonza).

[0207] CosMx sample preparation, data acquisition and quality control

[0208] Four FFPE samples were selected based on their CRC subtype and immunophenotype (IP). CosMx cyclic RNA readout was performed as previously described. In brief, following deparaffinization and rehydration slides underwent target retrieval for 15 min, protease digestion for 30 min, and postfixation. After blocking, tissue incubation with the probe set hybridization was performed overnight and unbound probes removed by slide flushing. Cell populations were identified using a DAPI nuclear stain and antibodies against B2M / CD298 (CosMx Human Universal Cell Segmentation Kit), PanCK (epithelial cells), CD45 (immune cells) (CosMx Human IO PanCK / CD45 Supplemental Segmentation Kit) and CD68 (myeloid cells) (CosMxHuman CD68 A La Carte Marker, Ch5). Nanostring flow cell- assembled slides were loaded onto the SMI instrument. Fields of view (FoVs) were selected (CR21 , UH20: 100 FoVs each; CR36, CR48: 93 FoVs each) to maximize the coverage of areas with high CD68 and CD45 staining. RNA readout and image capture were preformed according to manufacturer’s instructions using the CosMx Universal Cell Characterization Panel. Raw data, Seurat object expression data and slide-specific polygon files were obtained from the AtoMx Spatial Informatics Platform (Nanostring). Cells with a total number of counts > 20, total number of expressed genes > 15, and a NegProbes / total count ratio 0.05 were kept for downstream analyses.

[0209] CosMx cell annotation, local neighbourhood, and gene expression analysis

[0210] Cell populations were identified in each samples using the R package InSitutype (vl .0.0). Negative probe counts per cell were used to normalize raw counts. To refine singe cell annotation, the mean staining intensity for the five immunofluorescence markers were used and the insitutypefl function then implemented (n_starts=7, max_iters=20, n_c / usfs=10:14). Cluster phenotypes were defined using Seurat, first normalizing raw expression counts with NormalizeDataQ followed by FindAIIMarkersO function to define overexpressed genes for each cell cluster versus the rest. Cluster marker genes were defined as those with a fold change >1 .5 and a Bonferroni-corrected P-value <0.01 . Clusters with highly overlapping markers were merged using the InSituType refineClustersO function. In the case of T / NK cells, to filter out ambiguous cells, an additional clustering step was performed using refineClustersO with parameter subcluster = 2:4. T / NK cell subcluster markers have been identified using Seurat FindAIIMarkersO function for each subcluster versus the rest of the dataset. Samplespecific cell type clusters were visualized by Uniform Manifold Approximation and Projection (UMAP) using the top 30 principal components obtained through Seurat RunPCAO function followed by UMAP embedding taking the first 25 principal components to run the function RunUMAP() with default parameters.

[0211] A supervised clustering analysis with insitutypeMLO function using a single-cell human CRC dataset as a reference was ran to verify the relative abundance of each myeloid and lymphoid cell subpopulation within the myeloid and T / NK clusters. Pixel coordinates of cell boundaries were used to build cell polygons with st_polygon() function from the R package sf (v1 .0-14). Direct cell-cell contact was defined using the sf stJntersectsO function on the polygon objects. Cell types for downstream analysis were defined as follows: IE T cell, T / NK cells in direct contact with >3 epithelial cells; stromal T cell, T / NK cells in direct contact with 0 epithelial cells; Epi next to T cell, epithelial cells in direct contact with >1 IE T cells; Epi-only, epithelial cells in direct contact only with >4 epithelial cells; TAM next to T cell, myeloid cells in direct contact with >1 T / NK cells; TAM other, myeloid cells in direct contact with 0 T / NK cells.

[0212] Differential gene expression analysis was performed by applying the MAST hurdle model through Seurat FindMarkersO function on the CosMx log-normalized count data. Genes were defined as differentially expressed with a fold change >1.1 or <0.9, a Bonferroni-corrected P-value <0.01. Sample ID was indicated as a latent variable via the latent.vars parameter in FindMarkersO function. In the case of IE T and Stromal T cells comparison, to avoid spurious results caused by misassignment of gene transcripts due to poor cell segmentation, 185 genes previously identified as expression markers for Epi or FibroEndoMuscle clusters (Method ‘CosMx cell type annotation and local neighbourhood analysis’) in >3 samples were removed from the single-cell expression matrix before performing count normalization and differential gene expression.

[0213] In-vitro co-culture of CRC cell lines and CD8+ T cells and interferon-y treatment

[0214] Peripheral blood mononuclear cells (PBMCs) were isolated from the peripheral blood obtained from healthy donors by Ficoll-Plaque density gradient separation and cryopreserved until further used. One day before co-culture, HCT116 and CaCo2 cells were plated in 96-well U-bottomed plates at a density of 5 x 103cells / well and cultured overnight at 37°C. On the day of the co-culture, CD8+ T cells were isolated from PBMCs by magnetic cell depletion of NK cells using anti-CD56 microbeads (Miltenyi Biotec) followed by positive selection of CD8+ cells using anti-CD8 microbeads (Miltenyi Biotec) with MS and LS columns (Miltenyi Biotec). Freshly isolated CD8+ T cells were activated using anti- CD3 / CD28 beads, added to each well (2.5 x 104cells / well) and cultured for 48hrs in low glucose DMEM media, with or without anti-IFNy antibody (NIB42, eBioscience, 10 pg / mL). CRC and CD8+ T cells were then dissociated using TrypLE express, incubated with human Fc receptor block (Biolegend) and stained in FACS buffer (1 % FBS, 1 % BSA, 0.02% sodium azide) using anti-CD8- AF700 (clone SK1) and anti-EpCAM-APC (clone 9C4). Dead cells were excluded using the Zombie Aqua fixable viability dye (Biolegend). HCT116 and CaCo2 cells were sorted as Zombie EpCAM+CD8- cells while CD8+T cells were sorted as Zombie EpCAM CD8+cells using a FACSAria Fusion cell sorter. Overall, three independent experiments were carried out per CRC cell line and three wells were seeded in at least triplicate per condition per experiment (CRC- CD8+T cell co-cultures, CRC- CD8+T cell co-cultures plus anti-IFNy antibody, and CRC cell lines monocultures).

[0215] For the IFNy treatment assay, HCT116 and CaCo2 cells were plated at a density of 104cells / well in 48-well plates and cultured in low glucose DMEM media overnight at 37°C. The next day, the media was replaced with fresh media with or without 200 ng / mL human recombinant IFNy (Peprotech) and with or without anti-IFNy antibody (NIB42, eBioscience, 10 pg / mL as per manufacturer’s instructions). After 24hrs of treatment, cells were harvested and CD74 expression was measured as described below. Four independent experiments were performed for each experimental condition.

[0216] CD74 expression profiling and RNA-sequencing of human cell lines

[0217] Total RNA was extracted from CRC cell lines in the co-culture and IFNy treatment assays using the Direct-zol RNA microprep kit (ZymoResearch) and reverse transcribed using the High-capacity cDNA reverse transcription kit (Thermo Fisher Scientific) after measuring its concentration on a NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific). CD74 expression profiling was performed using predesigned Taqman gene expression assays for CD74 (Hs00269961_m1) and GAPDH (Hs02758991_g1) purchased from Life Technologies. Quantitative real-time PCR (qPCR) was performed in triplicate using QuantiTect probe PCR mastermix (Qiagen) and CD74 relative expression was calculated using the 2-AACtmethod and GAPDH as endogenous control.

[0218] RNA-sequencing was carried out in samples from one of the HCT116 co-culture experiments using 10 ng of total RNA to prepare cDNA libraries with the Watchmaker RNA Ribo / Globin depletion kit (Watchmaker Genomics). Libraries were sequenced using a 100bp paired-end protocol on Illumina NovaSeq S2 according to manufacturer’s instructions to obtain >20 million reads per sample. QC, read pseudo-alignment and count matrix generation was performed using demultiplexed FASTQ files as described before for LCM samples. After retaining only Salmon raw counts mapped to 19756 human genes for downstream analyses, the DESeq2 function DESeqO was run to perform gene expression normalization, gene-wise estimation of dispersion, and test differential gene expression between HCT116 cells cultured alone or co-cultured with activated CD8+ T cells. Differentially expressed genes from DESeq2 results were defined as those having a fold change >2 or <0.5, with a Benjamini- Hochberg corrected p-value <0.05.

[0219] Murine tumour studies

[0220] A total of 2.5 x 105cells in 100 pL serum-free RPMI MC38 and CT26 cells were implanted subcutaneously into the mammary fat pad of syngeneic female C57BL / 6 and BALB / c mice respectively. When tumours became palpable, dimensions were measured using callipers and volume was calculated using the following formula, where Length denotes the longest and Width the shortest tumour dimensions: Volume = (Length*Width2) / 2. Once the tumour volume across the cohorts reached an average of 63 mm3the mice were treated intraperitoneally with in vivo-ready anti-mouse CD279 PD1 Clone RMP1-14 (Biolegend Cat 114111) at 12 mg / kg or purified rat lgG2a K Isotype control (Biolegend Cat 400563) at 2.1 mg / kg three times per week (every 2 days) for approximately three weeks. Tumour tissue for flow cytometry analyses was enzyme-digested to release single cells as previously described (Anstee JE, Feehan KT, Opzoomer JW, Dean I, Muller HP, Bahri M, Cheung TS, Liakath-Ali K, Liu Z, Choy D, et al: LYVE-1 (+) macrophages form a collaborative CCR5-dependent perivascular niche that influences chemotherapy responses in murine breast cancer. Dev Cell 2023, 58:1548-1561 e1510). Tumours were extracted and cells were liberated by first cutting the tissues into small pieces (~1 mm thick) with a scalpel, followed by enzymatic digestion in RPMI supplemented with 1 mg / mLCollagenase I (Sigma-Aldrich) and 0.1 mg / mL DNAse I (AppliChem). Tissue was digested in a bacterial shaker for 1 hour at 37°C at 900 rpm prior to cell straining through a 70 pM nylon filter. Samples were centrifuged at 500 x g for 5 minutes at 4°C prior to cell counting using trypan blue (Thermo Fisher Scientific) exclusion prior to downstream analyses. Cd74 expression was quantified by RT-qPCR after tumour RNA was extracted using the RNeasy Mini Kit (Qiagen) and reverse transcribed using TaqMan One-Step RT-PCR Master Mix Reagents Kit (Thermo Fisher Scientific). Predesigned Taqman gene expression assays for Cd74 (Mm00658576_m1) and Actb (Mm02619580_g1) were purchased from Life Technologies and qPCRs were performed in triplicate. Cd74 relative expression was calculated using the 2AACtmethod and Actb as endogenous control. For immunohistochemistry analysis, tumours were fixed in 10% neutral buffered formalin solution (Sigma-Aldrich) for 24 h, dehydrated in 30% sucrose (Sigma-Aldrich) for 24 h, embedded in Optimal Cutting Temperature compound (VWR Chemicals) and stored at -80°C until further used. All mouse experiments were performed in at least one biological replicate.

[0221] DNA libraries were prepared using the Twist Bioscience Mouse Exome Panel (NEB Library Prep) on 100ng of genomic DNA extracted from MC38 and CT26 cells with the DNeasy Blood & Tissue kit (Qiagen). Libraries were sequenced using a 100bp paired-end protocol on Illumina NovaSeq S4 at 500x coverage per sample. Raw reads were aligned to GRCm38 mouse genome and the nf-core Sarek pipeline v3.1 .1 was used for tumour-only variant calling (Strelka) and annotation (snpEff) at default settings. Germline variants for C57BL / 6 and BALB / c mouse strains were filtered out from MC38 and CT26 variants, respectively. Only mutations passing the following filters were retained: Strelka QC passed, variant allele frequency <97%, altered allele supported by <10 reads, MODERATE or HIGH phenotypic impact according to snpEff annotation.

[0222] Analysis of pMMR CRC human cohorts

[0223] For the CAMILLA trial, raw GeoMx data from pretreatment metastatic biopsies of 20 pMMR CRC patients were downloaded from Gene Expression Omnibus (GSE254054). Data quality control was performed following the original publication, resulting in 105 regions of interest, 52 from the tumour and 53 from the stroma. Gene expression data were normalized with DESeq2 function varianceStabilizingTransformationO-

[0224] For the AtezoTRIBE cohort, FFPE tissue slides from 124 pMMR CRC patients were stained for CD74 and SPS was derived as described above. SPS values were dichotomized into high or low using receiver operating characteristic (ROC) curves to identify the SPS cut-point that best separated progression-free survival (PFS) above and below the median value within the experimental arm of the trial. This SPS threshold was then applied to all samples. Cox Proportional Hazard model using the coxphO function of the R package survminer was applied to test interaction between CD74 SPS and treatment. Kaplan-Meier PFS curves for patient subgroups were generated using survfitO and ggsurvpIotQ functions.

[0225] Flow cytometry Flow cytometry was performed as previously described (Cancer Genome Atlas N: Comprehensive molecular characterization of human colon and rectal cancer. Nature 2012, 487:330-337). Fc receptors present on any of the stromal populations were blocked with 5 pg / mL Fc Block / Fc shield (anti-CD 16 / 32) (BD Biosciences) at room temperature away from light for 30 minutes. The following antibodies against the indicated antigen were purchased from Thermo Fisher Scientific and were used at 1 pg / mL unless stated otherwise: Ter119-APC780 (TER-119, Biolegend), CD45-BV785 (30-F11 , Biolegend), F4 / 80-PECy7 (BM8, Biolegend), CD74-APC (PIN.1 , StressMarq Biosciences). Positive stains were compared to fluorescence minus one (FMO) controls. Intracellular stains were performed as previously described (Anstee JE, Feehan KT, Opzoomer JW, Dean I, Muller HP, Bahri M, Cheung TS, Liakath-Ali K, Liu Z, Choy D, et al: LYVE-1 (+) macrophages form a collaborative CCR5-dependent perivascular niche that influences chemotherapy responses in murine breast cancer. Dev Cell 2023, 58:1548-1561 e1510). Dead cells and red blood cells were excluded using 1 pL / mL Fixable Viability Dye eFluor® 780. Samples for intracellular staining were incubated on ice with fixation / permeabilization buffer (eBioScience) in the dark for 20 minutes. Cells were then washed with permeabilization buffer and stained with antibody mix for 20 minutes on ice in the dark prior to washing. Samples were run on the CytoFLEX V2-B5-R3 Flow Cytometer (Beckman Coulter). Results were analysed using FlowJo (BD Biosciences).

[0226] Haematoxylin and eosin (H&E) and immunohistochemistry staining

[0227] CD74 scores from IHC protein quantification of 24 human dMMR CRCs were separated by immunotherapy response. CD74 was quantified in from both stromal (S) and tumour (T) cells to calculate: combined positive score (CPS: S-positive plus T-positive over all detected cells); stroma positive against tumour score (SPT: S-positive over all detected tumour cells); tumour positive score (TPS: T-positive over all detected tumour cells); stroma positive score (SPS: S-positive over all detected stromal cells).

[0228] Regional micro-dissection to separate epithelium from stroma across bulk tumour regions indicates greater intertumour than intra-regional heterogeneity

[0229] We laser-capture micro-dissected (LCM) 235 samples at the core and invasive margins of 59 formalin- fixed paraffin-embedded (FFPE) tumour blocks from 58 colorectal patients for separate analysis of tumour stroma (Ts), tumour epithelium (Te) and normal epithelium (Ne) compartments (Figure 1A). Tissue blocks came from mismatch repair (MMR) proficient (pMMR, 30) and deficient (dMMR, 29) colorectal cancer (CRC) patients. Twenty-four of the latter were treated with anti-PD1 immunotherapy and further categorised into responders and non-responders (Figure 1 B). To quantify the tumour mutational burden (TMB) and genetic alterations, we sequenced a consensus of 324 cancer-related genes from three gene panels (Figure 2A). As expected, the TMB of dMMR tumours was significantly higher than that of pMMR tumours (Figure 1 C) and consistent with the TCGA threshold (>12 muts / Mbp) (Cancer Genome Atlas N: Comprehensive molecular characterization of human colon and rectal cancer. Nature 2012, 487:330-337) and immunotherapy eligibility cut-off (>10 muts / Mbp) (Marabelle A, Le DT, Ascierto PA, Di Giacomo AM, De Jesus-Acosta A, Delord JP, Geva R, Gottfried M, Penel N, Hansen AR, et al: Efficacy of Pembrolizumab in Patients With Noncolorectal High Microsatellite Instability / Mismatch Repair-Deficient Cancer: Results From the Phase II KEYNOTE-158 Study. J Clin Oncol 2020, 38:1-10; Marcus L, Fashoyin-Aje LA, Donoghue M, Yuan M, Rodriguez L, Gallagher PS, Philip R, Ghosh S, Theoret MR, Beaver JA, et al: FDA Approval Summary: Pembrolizumab for the Treatment of Tumor Mutational Burden-High Solid Tumors. Clin Cancer Res 2021 , 27:4685-4689). Gene alteration frequencies in the two CRC subtypes were concordant with previous reports (Cancer Genome Atlas N: Comprehensive molecular characterization of human colon and rectal cancer. Nature 2012, 487:330-337). APC, TP53 and KRAS mutations were more frequent in pMMR CRCs while TGFBR2, BRAF, MMR gene mutations were more frequent in dMMR CRCs (FDR <0.1 , corrected two-sided Fisher’s exact test.

[0230] We mapped the 324 genes to 20 pathways combining MSigDB (Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES, Mesirov JP: Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 2005, 102:15545-15550), Reactome (Fabregat A, Jupe S, Matthews L, Sidiropoulos K, Gillespie M, Garapati P, Haw R, Jassal B, Korninger F, May B, et al: The Reactome Pathway Knowledgebase. Nucleic Acids Res 2018, 46:D649-D655) and KEGG (Kanehisa M, Furumichi M, Tanabe M, Sato Y, Morishima K: KEGG: new perspectives on genomes, pathways, diseases and drugs. Nucleic Acids Res 2017, 45:D353-D361) and compared pathway alteration frequencies between dMMR and pMMR CRCs (Figures 2B). KRAS, chromatin, interferon, and hypoxia-related pathways accumulated significantly more alterations in dMMR than pMMR CRCs (FDR <0.1 , corrected two-sided Fisher’s exact test, Figures 1 D). Notably, a significantly high proportion CRCs (68%) were either wild-type or mutant in both KRAS and hypoxia pathways irrespective of the subtype (p = 0.03, two-sided Fisher’s exact test Figure 1 E). This was consistent with the functional association between hypoxic CRCs and KRAS signalling activation (Qi L, Chen J, Yang Y, Hu W: Hypoxia Correlates With Poor Survival and M2 Macrophage Infiltration in Colorectal Cancer. Front Oncol 2020, 10:566430). Chromatin-associated genes were near universally altered in dMMR CRCs but bisected pMMR CRCs (Figures 1 D), paralleling recent observations of clonal mutations in chromatin modifiers occurring pervasively in dMMR but only in a fraction of pMMR CRCs (Heide T, Househam J, Cresswell GD, Spiteri I, Lynn C, Mossner M, Kimberley C, Fernandez-Mateos J, Chen B, Zapata L, et al: The co-evolution of the genome and epigenome in colorectal cancer. Nature 2022, 611 :733-743). Next, we performed whole transcriptomics in each LCM sample from Te and Ts compartments. To minimise the batch effect and maximise the number of detected genes from low-input RNA (<2ng), we optimised the RNA-sequencing protocol by concatenating the read counts from two independently sequenced cDNA libraries (Figure 3A). For both Te and Ts, the number of genes detected with this approach was higher than single 2ng libraries and comparable to 10ng libraries (Figure 3B). Applied to the 235 LCM samples, our optimised protocol yielded a median of 11523 genes per sample for a total of 15652 genes overall (Figure 3C). Although there was no difference in number of genes between Te and Ts, we detected consistently higher number of genes in pMMR Te and Ts (Figure 1 F). All 15652 genes were expressed in at least one of dMMR / pMMR Ts and Te indicating the lack of type or tissuespecific genes per se, only that pMMR samples possessed more complex transcriptomes.

[0231] We performed dimensionality reduction on global transcriptomes and confirmed no need for further batch correction (Figure 3D) reflecting the optimised experimental design. Tissue compartments, rather than CRC subtypes, underscored the main separation between dMMR and pMMR CRCs (Figure 1G). This held true when the two subtypes were analysed separately (Figure 3E,F), indicating that inter-compartment transcriptional heterogeneity prevailed over inter-subtype differences. We observed no separation between the whole transcriptomes of core and invasive margin in Ts (Figure 1 H) and Te, (Figure 4), suggesting that intratumour transcriptional heterogeneity is not apparent when observed from intertumour perspective. This was surprising especially for Ts due to the likely higher stromal infiltration at the invasive margins of the tumour compared to the core. To further investigate this, we performed hierarchical clustering of samples from regions at the core and invasive margin in Ts and Te independently. In both compartments, samples tended to cluster by block ID rather than region (Figure 1J,K). However, while this occurred for most Te samples (81%, Figure 1 L), paired samples were only 49% in Ts (Figure 1M), indicating that there is more transcriptional diversity within the tumour stroma than within the tumour epithelium.

[0232] Finally, given the dispersion in the Te compartment dimensionality reduction plot (Figure 4) we sought to assess intertumour epithelial heterogeneity. To quantify this, we measured the distance between a Te sample and all its MMR subtype neighbours and compared the resulting density distributions between pMMR and dMMR CRCs. We confirmed that dMMR Te were significantly more transcriptionally heterogeneous than pMMR Te (two-sided Wilcoxon’s rank sum test, Figure 1 N).

[0233] Tumour intrinsic transcriptomic features provide greater insight into CRC biology than gene or pathway mutation status alone.

[0234] Although highly enriched in epithelial cells, LCM Te samples were still constituted cellular admixtures. We therefore applied a gene signature enrichment approach based on single-sample enrichmentbased analysis (ssGSEA) (Barbie DA, Tamayo P, Boehm JS, Kim SY, Moody SE, Dunn IF, Schinzel AC, Sandy P, Meylan E, Scholl C, et al: Systematic RNA interference reveals that oncogenic KRAS- driven cancers require TBK1 . Nature 2009, 462:108-112) to infer the activity of 14 tumour-intrinsic transcriptional programmes in each LCM Te sample. These intrinsic signatures were selected to represent the broad biology and the most common mutational phenotypes of colorectal cancer and we represented them through 14 non-overlapping gene signatures extracted from the literature and validated in CRC datasets and through MSigDB (Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES, Mesirov JP: Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 2005, 102:15545-15550) and KEGG (Kanehisa M, Furumichi M, Tanabe M, Sato Y, Morishima K: KEGG: new perspectives on genomes, pathways, diseases and drugs. Nucleic Acids Res 2017, 45:D353-D361) (Figure 10).

[0235] We compared the distributions of ssGSEA normalised enrichment scores (NES) of each intrinsic signature between core and invasive margin regions within Te of the two MMR subtypes separately. In line with the clustering results (Figure 1 J,L), we detected no significant differences in the intrinsic signatures between core and invasive margin, confirming substantial transcriptional homogeneity within the tumour epithelium. The only exception was a higher IFNa response at the invasive margins particularly of pMMR CRCs (two-sided Wilcoxon’s rank sum test, Figure 4A) likely reflecting a higher engagement with the immune system at the periphery of the tumour.

[0236] Next, we compared the 14 intrinsic signatures between Te compartment of either CRC subtype and the six Ne samples, after verifying that there were no transcriptional differences between Ne adjacent to dMMR or pMMR tumours. Except for the sternness signature, all other signatures were significantly different between Ne and the Te from at least one of the two CRC subtypes (FDR <0.05, corrected two-sided Wilcoxon’s rank sum test, Figure 4B. This was in line with the clear separation between the global transcriptomes of Ne and Te, regardless of CRC subtypes, that we observed by dimensionality reduction (Figure 11). It also confirmed that the 14 signatures captured tumour-induced deregulations. The majority of programmes were significantly more active in the tumour than in normal epithelium, with the exception of those associated with oxidative phosphorylation and, most visibly, wild type TP53 (Figure 4B).

[0237] Finally, we compared the activation status of the 14 tumour-intrinsic transcriptional programmes between dMMR and pMMR Te compartments, finding that ten of them were significantly different (FDR <0.05, corrected two-sided Wilcoxon-rank test Figure 4C). IFNa and IFNg responses and, to a lower extent, hypoxia were higher in dMMR tumours consistent with their higher mutational frequency (Figure 1 D) and indicating that the dMMR elevated immune infiltration (Pelka K, Hofree M, Chen JH, Sarkizova S, Pirl JD, Jorgji V, Bejnood A, Dionne D, Ge WH, Xu KH, et al: Spatially organized multicellular immune hubs in human colorectal cancer. Cell 2021 , 184:4734-4752 e4720) has transcriptional effect on the cancer epithelium. Consistent with increased type I IFN responses, STING-associated signature was higher in dMMR, where the accumulation of DNA damage is known to chronically activate IFN signalling via cGAS-STING (Mestrallet G, Brown M, Bozkus CC, Bhardwaj N: Immune escape and resistance to immunotherapy in mismatch repair deficient tumors. Front Immunol 2023, 14:1210164).

[0238] Other signatures reflected the different mutation frequencies in the two subtypes, indicating that the effects of somatic alterations in specific genes are measurable from the expression profile of wider gene programmes. This is the case of BRAF signature (higher in dMMR CRCs) and WNT and mutant TP53 transcriptional programmes (more active in pMMR CRCs, Figure 4C). However, the KRAS and TGFI3 signatures were not enriched in pMMR nor dMMR, respectively, as might be expected from the significantly different mutation frequencies of these genes in the two subtypes. This alluded to a potential uncoupling between KRAS and TGFI3 mutational status and the activation of the downstream pathways. Interestingly both pathways were significantly less active in Ne (Figure 4B). Similarly, the activation status of oxidative phosphorylation and glycolysis was not different between the Te of the two CRC subtypes, despite glycolysis being only marginally more active than in normal epithelium (Figure 4B). Finally, the sternness signature was most associated with pMMR Te compared to dMMR or Ne, consistent with the stem-like phenotype of pMMR CRCs (Guinney J, Dienstmann R, Wang X, de Reynies A, Schlicker A, Soneson C, Marisa L, Roepman P, Nyamundanda G, Angelino P, et al: The consensus molecular subtypes of colorectal cancer. Nat Med 2015, 21 :1350-1356; Nguyen MN, Choi TG, Nguyen DT, Kim JH, Jo YH, Shahid M, Akter S, Aryal SN, Yoo JY, Ahn YJ, et al: CRC- 113 gene expression signature for predicting prognosis in patients with colorectal cancer. Oncotarget 2015, 6:31674-31692), while metaplasia is associated with dMMR CRC (Figure 4C).

[0239] Cytotoxic and macrophage enrichment within micro-dissected CRCs shows intra-epithelial infiltration

[0240] We next undertook to measure the enrichment of a broad panel of immune populations within the Ts. Fifteen signatures representing immune cell signatures were derived from a combination of literature and re-analysis of CIBERSORT LM22 barcode genes (Newman AM, Liu CL, Green MR, Gentles AJ, Feng W, Xu Y, Hoang CD, Diehn M, Alizadeh AA: Robust enumeration of cell subsets from tissue expression profiles. Nat Methods 2015, 12:453-457), and were validated on derivative (Figure 5A) and colorectal-specific (Pelka K, Hofree M, Chen JH, Sarkizova S, Pirl JD, Jorgji V, Bejnood A, Dionne D, Ge WH, Xu KH, et al: Spatially organized multicellular immune hubs in human colorectal cancer. Cell 2021 , 184:4734-4752 e4720) (Figure 5B) expression data.

[0241] Nine of the fifteen signatures we tested were all enriched in the Ts of dMMR samples compared to pMMR, none were enriched in the opposite direction (Figure 6A). High general T-cell infiltration within dMMR versus pMMR Ts was attributable to cytotoxic effector T cell enrichment (FDR P = 0.0008) due to lack of significant difference in CD4-(T cell)-17 signature enrichment (Figure 6A). We also found T cell cytotoxicity paralleled by high enrichment of NK cells (NK-33, FDR P = 0.0003) (Figure 6A). For non-immune stromal populations, we observed that TGFJ3-producing cancer-associated fibroblasts (CAF-S1-myCAF-TGFJ3-7, Figure 6B) were enriched in pMMR Ts (FDR P = 0.0025). Population of TGF-producing CAFs have been previously described to be immunosuppressive, partially explaining the immune excluded phenotype of pMMR CRCs (Koppensteiner L, Mathieson L, O'Connor RA, Akram AR: Cancer Associated Fibroblasts - An Impediment to Effective Anti-Cancer T Cell Immunity. Front Immunol 2022, 13:887380; Mariathasan S, Turley SJ, Nickles D, Castiglioni A, Yuen K, Wang Y, Kadel EE, III, Koeppen H, Astarita JL, Cubas R, et al: TGFbeta attenuates tumour response to PD-L1 blockade by contributing to exclusion of T cells. Nature 2018, 554:544-548).

[0242] We next explored whether correlations between signature types might reveal interplaying TME features. Globally, intrinsic features related to interferon, including STING, were highly correlated with Ts immune enrichment, particularly interferon-producing NK cells and M1 macrophages (Figure 6C). We found cytotoxicity and interferon production were anti-correlated with mutant TP53 enrichment, consistent with these features being characteristic of opposite MMR subtypes. Further, TGF-producing CAF enrichment positively correlated with mutant TP53 enrichment alluding to the potential of the tumour epithelium to shape the stromal compartment. For dMMR CRCs, we found a broad negative association of immune exclusive tumour features (sternness, TGF-production and Wnt pathway activation) with immune enrichment of the Ts, which had opposing correlations with normal fibroblast enrichment (Figure 6D). Interestingly, despite the presumed immune cold nature of pMMR CRCs, we found positive correlations between tumour intrinsic interferon response and related immune enrichment indicative of infiltration-dependent heterogeneity (Figure 6E). Wnt pathway enrichment was also a major feature of signature correlations, as expected of this subtype, showing positive correlation with M2 macrophages consistent with the pro-tumour activity of these cells (Min AKT, Mimura K, Nakajima S, Okayama H, Saito K, Sakamoto W, Fujita S, Endo H, Saito M, Saze Z, et al: Therapeutic potential of anti-VEGF receptor 2 therapy targeting for M2-tumor-associated macrophages in colorectal cancer. Cancer Immunol Immunother 2021 , 70:289-298) and the activation of canonical Wnt signalling by CD206+ macrophages in damaged mucosa (Cosin-Roger J, Ortiz- Masia D, Calatayud S, Hernandez C, Alvarez A, Hinojosa J, Esplugues JV, Barrachina MD: M2 macrophages activate WNT signaling pathway in epithelial cells: relevance in ulcerative colitis. PLoS One 2013, 8:e78128). Despite prominence of certain immune populations across inter-subtype comparisons, within Ts, regardless of subtype, we found significant co-infiltration across all immune subsets (Figure 6F).

[0243] Next, we tested whether our enrichment-based approach could resolve immune infiltrates from microdissected tumour epithelium. We calculated the purity of our tumour epithelium LCM samples using ESTIMATE (Yoshihara K, Shahmoradgoli M, Martinez E, Vegesna R, Kim H, Torres-Garcia W, Trevino V, Shen H, Laird PW, Levine DA, et al: Inferring tumour purity and stromal and immune cell admixture from expression data. Nat Commun 2013, 4:2612), finding no difference between the tumour regions of each subtype but significantly reduced purity of dMMR tumours (Figure 5C). Our group has previously reported the increased abundance of proliferating and effector CD8+ T cells in hypermutated CRC (Bortolomeazzi M, Keddar MR, Montorsi L, Acha-Sagredo A, Benedetti L, Temelkovski D, Choi S, Petrov N, Todd K, Wai P, et al: Immunogenomics of Colorectal Cancer Response to Checkpoint Blockade: Analysis of the KEYNOTE 177 Trial and Validation Cohorts. Gastroenterology 2021 , 161 :1179-1193). Mlencnik et al. (Mlecnik B, Bindea G, Angell HK, Maby P, Angelova M, Tougeron D, Church SE, Lafontaine L, Fischer M, Fredriksen T, et al: Integrative Analyses of Colorectal Cancer Show Immunoscore Is a Stronger Predictor of Patient Survival Than Microsatellite Instability. Immunity 2016, 44:698-711) similarly reported using tissue microarrays increased densities of both stromal and intra-epithelial cytotoxic T cells in MSI versus MSS CRC cases. To detect enrichment of lowly expressed immune genes from background noise in Te, we selected the minimum purity fraction (98%) where our cytotoxicity-23 signature was significantly enriched in dMMR versus pMMR Te with >50% of samples in both subtypes (Figure 5D).

[0244] We tested our panel of 15 immune-based signatures for enrichment within the expression data of the micro-dissected Te samples. Seven signatures were more enriched in Te micro-dissected from dMMR samples than pMMR counterparts (Figure 6G). However, unlike Ts, we also observed enrichment of CD4 (T cells)-17, dendritic cells (DC)-19 and mast (cells)-25 increased in dMMR compared to pMMR Te. To confirm these results, we used ranked GSEA on gene-level fold-changes between dMMR and pMMR micro-dissected tumours, showing agreement in enrichment direction and significance of Cytotoxic-23, M1 macrophages (M1-16), and NK (cells)-33 (P <0.05, NES >0) (Figure 6H). Increased intra-epithelial populations of T-lymphocytes in dMMR CRCs are well described (Gonzalez IA, Bauer PS, Liu J, Chatterjee D: Intraepithelial tumour infiltrating lymphocytes are associated with absence of tumour budding and immature / myxoid desmoplastic reaction, and with better recurrence-free survival in stages l-lll colorectal cancer. Histopathology 2021 , 78:252-264), and reports of the polarisation of M1 macrophages with increased tumour-infiltrating lymphocytes and an MSI-high high phenotype (Vayrynen JP, Haruki K, Lau MC, Vayrynen SA, Zhong R, Dias Costa A, Borowsky J, Zhao M, Fujiyoshi K, Arima K, et al: The Prognostic Role of Macrophage Polarization in the Colorectal Cancer Microenvironment. Cancer Immunol Res 2021 , 9:8-19) parallel well with our observations. High abundance of NK cells within CRCs has been reported but data reporting intra-epithelial subsets has been sparse and often linked IC1 -like populations (Huang Q, Cao W, Mielke LA, Seillet C, Belz GT, Jacquelot N: Innate Lymphoid Cells in Colorectal Cancers: A Double-Edged Sword. Front Immunol 2019, 10:3080; de Vries NL, van Unen V, Ijsselsteijn ME, Abdelaal T, van der Breggen R, Farina Sarasqueta A, Mahfouz A, Peeters K, Hollt T, Lelieveldt BPF, et al: High-dimensional cytometric analysis of colorectal cancer reveals novel mediators of antitumour immunity. Gut 2020, 69:691-703). We used CD68 and CD57 staining by immunohistochemistry (IHC) as a proxy for T cells, macrophages, and NK cells, respectively. Representative samples across the distribution of enrichment scores for the Cytotoxic-23, M1 macrophages (M1-16), and NK (cells)-33 signatures were selected for IHC validation whereby we found moderate-to-strong correlations between stained protein quantification and transcriptomic enrichment scores (Pearson r 0.47-0.79, P < 0.05) (Figure 6I). Further, by IHC we were able to in-situ visualise the intra-epithelial subset of these infiltrating immune cells (Figure 6J).

[0245] Co-infiltration of the Te by all immune populations paralleled that of the Ts (Figure 6F,K). Moreover, we tracked enrichment scores correlations across tissue compartments to reveal Ts:Te co-infitration. Globally, the signatures with the highest frequency of significant enrichment correlations were those that were most enriched in the dMMR versus pMMR comparison: Cytotoxic-23, M1 (macrophages)-16, and NK (cells)-33, indicating the importance of these populations in underscoring the immune-hot microenvironment of CRC and likely indicative of recruitment rather than intra-epithelial residency (Figure 6L). Further, we show that intra-epithelial Cytotoxic-23, M1 (macrophages)-16, and NK (cells)- 33 enrichments have the strongest correlations (Pearson’s r >0.8) with a broad array of immune populations in the Ts, possibly reflecting that a well-populated stromal microenvironment is necessary for infiltration of the tumour nest. We found these observations paralleled in both dMMR CRCs and pMMR CRCs demonstrating shared infiltration phenotypes despite distinct MMR status (Figure 6L). Although in pMMR CRCs we did observe broad co-infiltration, we also observed loss in strength of cross-compartment correlation of Cytotoxic-23 and NK (cells)-33 enrichments (r <0.6). Interestingly, intra-epithelial enrichment of M1 macrophages in pMMR CRCs was strongly correlated (r >0.8) with Ts enrichment of a broad set of adaptive and myeloid populations, including CD4, cytotoxicity, NK cells, dendritic cells, and M0 and M1 macrophages (Figure 6L), indicating that epithelial infiltration of M1 -like macrophages is a common feature of CRC regardless of MMR status.

[0246] Parallel immunophenotypes between hypermutated and non-hypermutated CRCs

[0247] Given that we observed strong correlation in addition to considerable score distribution among the MMR-defining immune enrichment signatures across both tissue compartments (Figure 6A,G), we looked to explore whether these could identify subset of tumours within their MMR subtype. By hierarchical clustering we assessed the relationship between samples using their respective Ts and Te macrophage (M1)-16, NK (cells)-33 and cytotoxicity-23, and, because of their relevance to these immune processes, interferon-related enrichments (Figure 7A, B). Within each resulting clustering we assessed for differentially mutated genes, of which only dMMR subclusters showed differences in tumours carrying gene alterations (Fisher Test, P < 0.1) (Figure 7A). These ten, more often, mutated in colder CRCs were not found mutated in pMMR cold tumours (Figure 7B), nor was the TMB score of cold dMMR tumours any different to the rest (Figure 7C). KRAS mutations are associated with pMMR and cold dMMR samples (Figure 7A). We also found a subset of mildly infiltrated dMMR tumours (“dMMR IFN response” group) that have significantly increased KRAS-27 signature compared to hot counterparts (p = 0.015, FDR = 0.12) and alike cold counterparts (p=0.25, FDR = 0.35). Interestingly, we find the high KRAS-27 enrichment of this dMMR IFN response group incongruous to its wild-type KRAS status, as such we wondered whether a subset of lowly infiltrated, KRAS-wildtype dMMR CRCs was more generally describable. To test this we used the TCGA and CPTAC dMMR cohorts to test for enrichment of macrophages (M1)-16, NK (cells)-33 and cytotoxicity-23 and STING / IFN responses alongside KRAS(-27) signature enrichment and gene status (Figure 7D). Considering only KRAS wild-type samples, we found a proportion with raised IFN responses (Figure 7E) but elevated infiltration compared to cold counterparts (Figure 7F), which we termed the “intermediate-C” group. We show that this subgroup phenotype has high enrichment of KRAS-27 like cold samples (regardless of KRAS status) and is significantly increased compared to archetypal dMMR samples with a hot microenvironment (FDR = 0.021) (Figure 7G).

[0248] Because our clustering was based on immune infiltrates known to be increased in dMMR CRCs with durable benefit to anti-PD1 blockade (Bortolomeazzi M, Keddar MR, Montorsi L, Acha-Sagredo A, Benedetti L, Temelkovski D, Choi S, Petrov N, Todd K, Wai P, et al: Immunogenomics of Colorectal Cancer Response to Checkpoint Blockade: Analysis of the KEYNOTE 177 Trial and Validation Cohorts. Gastroenterology 2021 , 161 :1179-1193), we tested cluster composition for overrepresentation of favourable immunotherapy response. We considered any level of cytotoxic and M1 macrophage infiltration within dMMR tumours as beneficial for facilitating anti-PD1 immunotherapy response, for which thirteen of our dMMR samples show and 62% of these have favourable anti-PD1 responses (Fisher’s exact test, p = 0.123). When accounting for samples that have not received neoadjuvant therapy that may alter the tumour microenvironment vis-a-vis treatment-naive samples, clusters of tumours with high cytotoxic, M1 macrophage and interferon enrichments are significantly over-represented by favourable anti-PD1 response (Fisher’s exact test, p = 0.02) (Figure 7H).

[0249] TIME immunophenotypes stratify CRCs for response to anti-PD-1 ICI

[0250] TIME comparison between CRC subtypes consistently showed differences in the abundance of stromal and IE NK cell, cytotoxic lymphocyte, and CXCL10+ TAM infiltration triggering IFN responses and STING activation in the tumour. Hierarchical clustering using the NES values of these seven signatures separated the whole cohort into four distinct clusters (Figure 11 A). Cluster A, mostly composed of dMMR CRCs, had the highest immune infiltration and associated tumour response while cluster B had high stromal but lower IE infiltration and tumour response and was a mixture of both subtypes. Clusters C and D showed lower levels of immune infiltration and tumour response.

[0251] Consistent with this, clusters A and B had significantly higher IFNG (encoding interferon y) expression than clusters C and D (Figure 11 B). We therefore classified samples into IFN-high and IFN-low immunophenotypes (IP) CRCs (Figures 11A and 11C). We tested the robustness of this classification in two ways. First, we used the seven signatures to cluster two external CRC RNA-seq datasets. Again, we observed two IPs (Figures 16A and 16B) significantly differing for IFNG expression levels (Figure 16C). IFN-high IP comprised around 20% of pMMR CRCs, in line with our results (Figure 11A). Second, we applied the iCMS classification 10 to our cohort. IFN-high CRCs were significantly enriched in iCMS3_MSI and iCMS3_MSS_F classes (Figure S3D), which have the highest IFN response and immune cytotoxicity level. Therefore, our signatures could consistently identify similar IPs in independent CRC cohorts and these IPs significantly overlapped with external classifications associated with IFN response.

[0252] In addition to the seven signatures used for clustering, IFN-high IP tumours were enriched in almost all other immune populations, supporting widespread immune co-infiltration. TMB was not different between IFN-high and IFN-low IP CRCs (Figure 11 D), confirming that immune infiltration is not triggered by the levels of tumour immunogenicity. Despite this, IFN-high IP tumours were significantly over-represented in dMMR CRC patients responsive to anti-PD-1 ICI (Figure 11 E) and the enrichment remained significant even considering all pMMR CRCs as potential non-responders (Figure 11 F). Interestingly, IFN-high IP CRCs were depleted in patients with liver metastasis (Figure 11G), which is an emerging marker of resistance to ICI. As previously reported, CD274 (encoding PD-L1) was lowly expressed and showed significant overexpression in IFN-high IP dMMR but not pMMR CRCs (Figure 16E). Recently, SOX17 has been reported to suppress the ability of early CRC cells to respond to IFNy enabling immune evasion. The S0X17 gene is barely expressed in the Te, and we detected no significant difference in S0X17 expression between IFN-high and IFN-low IP CRCs (Figure 16F). This may suggest different roles of SOX17 in early and advanced CRC.

[0253] Finally, hierarchical clustering using all 21 different signatures between CRC subtypes did not divide samples according to IP (Figure 16G), IFNG gene expression (Figure 16H), or response to ICI (Figure 161). Therefore, the separation into IPs and association with response to immunotherapy reflect specific TIME properties and associated tumour response rather than global differences between CRC subtypes.

[0254] T cell proximity induces CD74 upregulation in TAMs and tumour cells

[0255] To better understand the local interplay between tumour cells and the three immune populations (NK cells, cytotoxic lymphocytes, and CXCL10+ TAMs) defining IFN-high and IFN-low IPs, we performed single-cell spatial transcriptomics25 of two ICI-responsive dMMR (UH20 and CR36) and two IFN-high pMMR (CR21 and CR48) CRCs. To aid cell phenotyping and verify transcriptome-based cell clustering, we used immunofluorescence staining of epithelial cells (pan-keratin), TAMs (CD68), and lymphocytes (CD45) in addition to nuclear (DAPI) and membrane markers (B2M and CD298). We identified between 149645 and 253161 cells per sample that were grouped into five or six clusters (Figure 17A). Using supervised annotation based on single-cell CRC data, we confirmed that the myeloid and T / NK cell clusters were mainly composed of TAMs and T cells, respectively (Figure 17B).

[0256] After separating TAMs adjacent to T cells from the rest of TAMs based on their spatial coordinates and local cell neighbourhood (Figure 12A, Star Methods), we compared their gene expression profiles. TAMs adjacent to T cells showed significant upregulation of genes involved in antigen processing and presentation (APP) and IFN response (Figures 12B and 17C). Applying a similar approach (Figure 12C), we observed significant upregulation of genes associated with a cytotoxic and exhausted phenotype in IE T cells compared to stromal T cells (Figure 12D). Finally, we tested how T cell proximity modified the gene expression profile of tumour cells compared to those far from immune infiltrates (Figure 12C). Like TAMs, tumour cells adjacent to T cells significantly overexpressed genes involved in APP and IFN response (Figure 17E). To independently support these findings, we analysed an external single-cell proteomic CRC dataset. Again, HLA-DRA (the only APP marker in the panel) and two IFN-stimulated proteins, ICAM1 and STING, were more abundant in TAMs and tumour cells proximal to T cells (Figure 17D), confirming our findings. Therefore, a TIME locally enriched in cytotoxic T cells induces the overexpression of APP genes in professional and non-professional antigen-presenting cells.

[0257] Next, we investigated whether a single marker could identify this local TIME indicative of IFN-high IP and response to I Cl . Of the seven overexpressed genes in both TAMs and tumour cells adjacent to T cells, four (CD74, CXCL9, CCL5, and CXCL10) were also consistently overexpressed in IFN-high IP CRCs across tissue compartments and CRC subtypes (Figure 17E), Interestingly, the MHC class II invariant chain CD74 was among the 15 protein markers of TAMs significantly associated with anti- PD-1 ICI response. Fourteen of these 15 genes were significantly overexpressed in IFN-high IP CRC in at least one compartment or subtype (Figure 17F), indicating that this TAM population likely corresponded to the CXCL10+ TAMs enriched in the IFN-high IP. In particular, CD74 was overexpressed in IFN-high IP CRCs of both subtypes and compartments (Figures 12F and 18A), supporting stromal and epithelial expression. Compared to TAMs, tumour cells expressed significantly lower levels of CD74 in the spatial transcriptomic (Figure 18B), LCM (Figure 18C), and external singlecell (Figure 18D) data. CD74 tumour and stroma expression was positively correlated (Figure 18E), suggesting that epithelial CD74 expression depends on the abundance of immune cell infiltration.

[0258] To test whether T cell proximity induces CD74 in tumour cells, thus supporting the spatial transcriptomic data, we compared its expression in dMMR (HCT116) and pMMR (CaCo2) CRC cell lines cultured with or without human primary activated CD8+ T cells (Figure 12G). CD74 expression was induced (HCT116 cells) or significantly increased (CaCo2 cells) upon co-culture with T cells (Figure 12H). Mirroring tumour samples (Figure 12E), HCT116 cells exposed to activated CD8+ T cells showed global upregulation of APP and IFN response (Figure 121; Table S5). We further tested whether CD74 expression was IFNy dependent by blocking IFNy with a neutralizing antibody (Figure 12G). This reduced CD74 expression in HCT116 but not CaCo2 cells (Figure 12H). To further investigate this, we cultured both cell lines in IFNy-supplemented media with or without anti-IFNy antibody (Figure 18F). Again, we observed CD74 induction in IFNy-treated HCT116 cells, while the effect was only modest in CaCo2 cells (Figure 18G). These results demonstrated that CD74 expression, and likely that of APP globally, can be directly induced in tumour cells by the proximity of CD8+ T cells through IFNy and possibly other factors.

[0259] High CD74 expression is associated with response to anti-PD-1 ICI in dMMR CRC

[0260] All our results on ICI response rates and differences in CD74 expression between IFN-high and IFN- low IP CRCs were uniquely based on gene expression comparisons. We further tested whether CD74 protein quantification via immunostaining of tumour tissue slides was alone able to separate responders from non-responders. To this aim, we compared the proportion of CD74+ cells via four scores analogous to those currently used in clinical pathology to quantify PD-L1+ cells. Namely we quantified CD74 combined positive score (CPS), stroma proportion over tumour score (SPT), and tumour proportion score (TPS, Figure 13A). Since CD74 was highly expressed in TAMs, we additionally measured the proportion of CD74+ stroma cells over all stroma cells (stroma proportional score, SPS, Figure 13A).

[0261] All CD74 scores were significantly higher in dMMR CRC responders with SPS showing the best separation between the two groups (Figure 9F). We confirmed these results using receiver operating characteristic (ROC) curves and identified 18% SPS as the threshold with the best sensitivity and specificity (Figure 13B). This threshold separated dMMR CRC responders from non-responders (Figure 13C) even better than Ips (Figure 11 E). To seek independent support, we quantified CD74+ cells in stained tissue sections from 43 additional dMMR CRCs of patients treated with ICI in advanced or metastatic setting (GONO cohort). Like in the UCLH cohort, responders from the GONO cohort showed significantly higher CD74 SPS values (Figure 13D). Moreover, 18% CD74 SPS threshold could again significantly separate responders from non-responders (Figure 13E). Interestingly, these differences held true even in a subset of patients treated with a combination of anti-PD-1 and anti-CTLA-4 ICI (Figures 13F and 13G), suggesting that an IFN-high TIME may be required for response to immunotherapy, independently of the agent. CD74 SPS values were not different between patients with or without liver metastases (Figure 13J), mirroring the lack of association between presence of liver metastasis and objective response in our (Figure 131) and external dMMR CRC cohorts. Although the low number of cases did not allow proper statistical comparison, CD74 SPS values were above 18% in all responders for which metastatic rather than primary lesions were analysed). This may suggest that the TIME features favouring response are site-independent and likely triggered by the properties of the tumour of origin. To test this hypothesis, we injected two murine CRC cell lines (CT26 and MC38) subcutaneously into immunocompetent syngeneic mice, which were then treated with anti-PD-1 or IgG (control) antibodies. Despite similar TMB (Figure 13J) and MSI score (Figure 13K), upon treatment we found significant growth inhibition of MC38 (Figure 9B) but not CT26 (Figure 9C) tumours. Moreover, MC38 IgG-treated tumours showed significantly higher Cd74 gene expression (Figure 13L) as well as significantly higher proportion of CD74+ cells (Figure 9D) than CT26 IgG- treated tumours. Although a comparison of multiple sites within the same donors is missing, these data support the initial hypothesis that CD74 is more expressed in CRCs that respond to anti-PD-1 ICI , independently of the tumour site.

[0262] Immune correlates of favourable anti-PD1 response identify CD74 as a predictive biomarker for dMMR and pMMR CRCs

[0263] We show that MMR-defining immune enrichments define parallel phenotype clustering between dMMR and pMMR CRCs, hence we explored how these subgroups exist when considered across all CRCs (Figure 8A). The extremes of enrichment scores are reflective of the well described biology of CRC: highly infiltrated, hot dMMR tumours and cold pMMR tumours. The intermediately infiltrated CRCs show discrete clustering of hot-stroma samples when defined either by MMR subtype or globally indicative of co-infiltration independent of tumour intrinsic features. Surprisingly, we found that CRCs defined as cold relative to their MMR subtype were independent clusters when considered globally (Figure 8A). This indicated that so-called cold dMMR CRCs retain immune enrichments lacking in truly cold pMMR CRCs, and this immune enrichment was not limited to cytotoxicity and M1 macrophages but more general immune co-infiltration and pronounced IFNag response enrichment.

[0264] We again reasoned that cold CRCs of both subtypes would be resistant to immunotherapy responses. Clustering in this way (Figure 8A), we found a significant over-representation of samples from anti- PD1 responsive dMMR tumours (Fisher’s exact test, P = 0.02) (Figure 8B). When we considered pMMR samples within this comparison (pMMR as non-responders), the cluster comparison gained an order of magnitude significance (Fisher’s exact test, P = 0.001) (Figure 8C). This is highly suggestive that pMMR CRCs that co-cluster with moderately / highly infiltrated dMMR counterparts may benefit from immunotherapy intervention.

[0265] To test this hypothesis, we used transcriptomic signatures characterising favourable immune checkpoint inhibition response in non-small cell lung cancer (NSCLC) and pMMR CRC (Xue W, Shi J: Identification of genes and cellular response factors related to immunotherapy response in mismatch repair-proficient colorectal cancer: a bioinformatics analysis. J Gastrointest Oncol 2022, 13:3038- 3055). In the Ts, where much of the immune infiltrate is located, both response signatures were increased in infiltrated pMMRs versus cold counterparts (NSCLC score Wilcoxon P = 1.2e-05, pMMR CRC score Wilcoxon P = 1 .8e-4) (Figure 8D).

[0266] We next considered single molecular markers that might identify the clustering we observe. We performed gene-wise Wilcoxon tests for increased expression in putative response groups and found top among these genes was CD74. Notably, CD74 expression in the Ts for both MMR subtypes shows a gradient of expression from hot to cold tumour clusters (Figure 8E). We report that for comparisons of interest, CD74 expression is highly increased in both Ts and Te compartments of dMMR cases, but only in the Ts of pMMR cases in the comparison of putative responders versus cold (Figure 8E). Confirming this, we show that CD74 expression is highly correlated with our clusterdefining immune signatures (Cytotoxic-23, NK-(cells)-33 and M1-(macrophages)-16) in both Ts and Te tissue compartments and considering MMR subtypes together or independently (Pearson’s r >0.5, P <0.001) (Figure 8F).

[0267] CD74 is a biomarker for favourable immunotherapy response

[0268] Since in general pMMR / low TMB CRCs are considered non-responders to immunotherapy, there is a lack of available expression data from anti-PD1 treated pMMR CRCs. Given this, we used murine models of CRC with divergent responses to immunotherapy. MC38 and CT26 mouse CRC cell lines were grown subcutaneously in syngeneic mice and were selected for their mismatch repair proficient status (REF: 10.1038 / nature24673) (Figure 9A). Confirming previous characterisations of MC38 and CT26 responses to anti-PD1 (REF: 10.3389 / fimmu.2022.1011943; 10.1186 / s12885-021-08974-3), we found significant inhibition of tumour growth only in MC38 models when treated with anti-PD1 versus IgG control (p <0.05, two-sided Wilcoxon test of end-point tumour volumes) (Figure 9B). CT26 were resistant to anti-PD1 immunotherapy (Figure 9C).

[0269] Given these observations, we assumed that these two models parallel our putative human pMMR groups: those that co-cluster with dMMR responders (MC38-like) and the rest of the pMMR CRCs (CT26-like) (Figure 8A). We performed CD74 staining by immunohistochemistry on 19 MC38 (n=10) and CT26 (n=9) tumours. The nature of these subcutaneously injected tumour cell lines is such that dense tumours make infiltration difficult by the host immune system, which precluded positive cell detection by tissue compartment. We quantified total CD74-positive staining normalised for all detected cells, showing a stark increase in CD74 expression in the MC38 anti-PD1 responsive model (p = 4.3x10-5, two-sided Wilcoxon test) (Figure 9D). To assess the contribution by tissue compartment we performed flow cytometry on dissociated tumours from each model. Analysis of CD45 and CD74 staining shows that MC38 tumours have an increase in both CD45-negative (tumour cells, upper left quadrant) and CD45-positive (host immune cells, upper right quadrant) expressing CD74 versus CT26 (Figure 9E). CD74 expression is therefore not limited to immune cells but also endogenously expressed by the tumour cells.

[0270] To test if CD74 can demarcate human anti-PD1 responders versus non-responders, we quantified its expression by IHC per tissue compartment using human CRC-trained classifiers. The cohort of 24 anti-PD1 -treated dMMR tumours from our human study group provided the control of this quantification. In all tissue compartments there was a trend of increased CD74 score in those that had favourable responses to anti-PD1 versus those that did not (Figure 9F). The quantification methods using compartment-specific positive detection normalised for the same compartment size (tumour positive score, TPS; stroma positive score, SPS) proved most significant between groups (p < 0.01 , two-sided Wilcoxon’s test) (Figure 9F). However, for a single marker to have clinical utility, we reasoned that the distributions of the responder and non-responder groups should minimally overlap. Therefore, we selected the SPS quantification method with most promise for clinical quantification of CD74 to determine immunotherapy response (Figure 9F).

[0271] Next, to determine a threshold by which putative responders may be identified for both MMR subtypes, we tested the CD74 generated by the SPS approach against the clustering performed in Figure 8A. For both subtypes, as predicted, our CD74 stroma positive score is significantly increased in samples from the infiltrated clusters (p < 0.01 , two-sided Wilcoxon’s test) (Figure 9G).

[0272] The infiltrated cluster of Figure 8A includes several non-responsive dMMR tumours (31 %, Figure 8B), using a CD74 SPS quantification threshold of 19%, we resolve these false positives as indicated by their low value for this score (Figure 9G). Hence, stromal CD74 quantification and separating samples by the defined cut-off is superior to hierarchical clustering in demarking dMMR responders versus non- responders (Fisher test 2.89x10-4 versus 0.02 (Figure 8B)). In this situation, only one responder and one non-responder are incorrectly predicted. For pMMR CRCs, using the same CD74 SPS threshold of 19% identifies all putative responders with a high score alongside three putative non-responders (Figure 9G). These data are highly suggestive that there exists a subset of pMMR CRCs that would benefit from anti-PD1 immunotherapy in the metastatic setting and they can be identified, along with dMMR responders, by quantification of stromal CD74-positive proportion in excess of 19%.

[0273] Hiqh CD74 SPS predicts clinical benefit in ICI-treated pMMR CRC patients

[0274] Hierarchical clustering showed a subset of pMMR CRCs with IFN-high IP (Figure 11 A). Moreover, all these tumours had CD74 SPS higher than 18% (Figure 14A) and their CD74 immunostains resembled those of dMMR CRCs (Figure 14B). This suggested that patients with pMMR CRCs and high CD74 SPS may have the local TIME required for response to ICI . Currently, anti-PD-1 ICI monotherapy is not used for the treatment of pMMR CRC because of the overall low response and inability to effectively identify upfront the few benefitting patients. To test whether CD74 abundance may help stratify pMMR CRC patients, we analyzed samples from two clinical trials testing ICI in combination with other therapies.

[0275] The first was the CAMILLA trial (NCT03539822) testing cabozantinib (tyrosine kinase inhibitor) and durvalumab (anti-PD-L1 ICI) in chemo-refractory gastrointestinal cancers. We gathered GeoMX spatial transcriptomic data from pretreatment metastatic biopsies of 20 pMMR CRC patients (Figure 14C). After data normalization and quality control (Star Methods), we compared CD74 expression levels in tumour and stroma regions of responders and non-responders. In both compartments, responders showed significantly higher CD74 expression (Figure 14D), confirming its association with response. These samples were metastatic biopsies from patients treated with an anti-PD-L1 agent, supporting our previous observations that CD74 abundance may be a marker of response to ICI irrespective of agent (Figures 13G and 13H) or tumour site (Figure 13L).

[0276] The second trial was the AtezoTRIBE trial (NCT03721653) assessing clinical benefit from addition of atezolizumab (anti-PDL1 ICI) to FOLFOXIRI (chemotherapy) and bevacizumab (anti-VEGF antibody). In this case, we stained and quantified CD74 SPS in FFPE tissue slides from 124 unresectable metastatic pMMR CRC patients who received (experimental arm) or not (control arm) atezolizumab (Figure 14E. A subset of 97 patients had been previously categorized as being high or low for immunoscore-immune-checkpoint (IS-IC), a spatial TIME quantification of PD-L1 + and CD8+ cells that showed positive association with PFS in the experimental arm of the trial. Interestingly, CD74 SPS values were significantly higher in high IS-IC pMMR CRCs (Figure 14F), suggesting that the two assays are markers of the same local TIME.

[0277] Neither IS-IC (Figure 19A) nor CD74 SPS (Figures 19B and 19C) showed significant association with objective response in this trial. We therefore tested whether SPS, like IS-IC, was predictive of PFS upon ICI treatment. To this aim, we dichotomized CD74 SPS into high and low values using the SPS cut point best separating ICI-treated patients according to the PFS median value (Star Methods).

[0278] Since patients with resected primary tumours had longer survival (Figure 19D), we analyzed this group separately from patients with unresected primary tumours. Although not statistically significant, high CD74 SPS was indicative of lower risk of progression in resected patients within the experimental but not control arm (median PFS of 26.4 and 15.0 months, respectively, Figure 14G). This was not the case for unresected patients (Figure 14H). When treating the two arms separately, we confirmed the significant prognostic value of high CD74 SPS in resected patients from the experimental (Figure 141) but not control (Figure 14J) arm. Moreover, resected patients treated with atezolizumab in the high CD74 SPS group were significantly depleted in liver metastases (Figure 19E), in line with the reported association between liver metastases and ICI resistance in pMMR CRC. Altogether, these results supported the role of CD74 as a marker of the local IFN-high TIME required to achieve clinical benefit. EMBODIMENTS

[0279] The invention may also be defined by reference to the following numbered embodiments:

[0280] 1 . A method for predicting or assessing the response of a subject to treatment with an immunotherapeutic agent, wherein the subject has mismatch repair proficient (pMMR) colorectal cancer, the method comprising: determining a level of CD74 expression in a sample obtained from the subject and comparing it to a reference value, wherein a CD74 expression level higher than the reference value is indicative of an immunotherapy responder, and wherein a CD74 expression level lower than the reference value is indicative of an immunotherapy non-responder.

[0281] 2. A method of selecting a subject for treatment with an immunotherapeutic agent, wherein the subject has mismatch repair proficient (pMMR) colorectal cancer, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; and b. comparing the level of CD74 expression in the sample to a reference value; wherein the subject is selected for treatment with the immunotherapeutic agent if the level of CD74 in the sample is higher than the reference value.

[0282] 3. The method of embodiment 1 or 2, wherein the reference value is a percentage of cells in the sample that express CD74.

[0283] 4. The method of embodiment 1 or 2, wherein the reference value is a percentage of the sample that is positive for CD74.

[0284] 5. The method of embodiment 1 or 2, wherein the reference value is a percentage of stromal cells in the sample that express CD74.

[0285] 6. The method of embodiment 1 or 2, wherein the reference value is a percentage of the stroma that is positive for CD74.

[0286] 7. The method of embodiment 1 or 2, wherein the reference value is a percentage of tumour stromal cells in the sample that express CD74.

[0287] 8. The method of embodiment 1 or 2, wherein the reference value is a percentage of the tumour stroma that is positive for CD74. 9. The method of embodiment 1 or 2, wherein the reference value is a percentage of tumour epithelial cells in the sample that express CD74.

[0288] 10. The method of embodiment 1 or 2, wherein the reference value is a percentage of the tumour epithelium that is positive for CD74.

[0289] 11. The method of embodiment 1 or 2, wherein the reference value is a percentage of tumour cells in the sample that express CD74.

[0290] 12. The method of embodiment 1 or 2, wherein the reference value is a percentage of the tumour that is positive for CD74.

[0291] 13. The method of any preceding embodiment, wherein the reference value is 15%.

[0292] 14. The method of any preceding embodiment, wherein the reference value is 16%.

[0293] 15. The method of any preceding embodiment, wherein the reference value is 17%.

[0294] 16. The method of any preceding embodiment, wherein the reference value is 18%.

[0295] 17. The method of any preceding embodiment, wherein the reference value is 19%.

[0296] 18. The method of any preceding embodiment, wherein the reference value is 20%.

[0297] 19. The method of any preceding embodiment, wherein the reference value is 21%.

[0298] 20. The method of any preceding embodiment, wherein the reference value is 22%.

[0299] 21. The method of any preceding embodiment, wherein the reference value is 23%.

[0300] 22. The method of any preceding embodiment, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 15% of the cells in the sample express CD74.

[0301] 23. The method of any preceding embodiment, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 17% of the cells in the sample express CD74.

[0302] 24. The method of any preceding embodiment, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 19% of the cells in the sample express CD74.

[0303] 25. The method of any preceding embodiment, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 21% of the cells in the sample express CD74.

[0304] 26. The method of any preceding embodiment, wherein the reference value is a CD74 score. 27. The method of any preceding embodiment, wherein the reference value is a CD74 stromapositive score (SPS).

[0305] 28. The method of any of embodiments 1 , 2 or 13-27, wherein the reference value is a CD74 stroma-positive score (SPS) of 15%.

[0306] 29. The method of any of embodiments 1 , 2, or 15-28, wherein the reference value is a CD74 stroma-positive score (SPS) of 17%.

[0307] 30. The method of any of embodiments 1 , 2, or 17-29, wherein the reference value is a CD74 stroma-positive score (SPS) of 19%.

[0308] 31. The method of any of embodiments 1 , 2, or 19-30, wherein the reference value is a CD74 stroma-positive score (SPS) of 21%.

[0309] 32. The method of any of embodiments 1 , 2, or 21-31 , wherein the reference value is a CD74 stroma-positive score (SPS) of 23%.

[0310] 33. The method of any of embodiments 1 , 2, or 13-26, wherein the reference value is a CD74 tumour-positive score (TPS).

[0311] 34. The method of any of embodiments 1 , 2, or 13-26, wherein the reference value is a CD74 combined positive score (CPS).

[0312] 35. The method of any of embodiments 1 , 2, or 13-26, wherein the reference value is a CD74 stroma positive against tumour (SPT) score.

[0313] 36. The method of any preceding embodiment, wherein determining the level of CD74 expression comprises determining a level of CD74 expression in the tumour stroma.

[0314] 37. The method of any preceding embodiment, wherein determining the level of CD74 expression comprises determining a level of CD74 expression in the tumour epithelium.

[0315] 38. The method of any preceding embodiment, wherein determining the level of CD74 expression comprises determining a level of CD74 expression by stromal cells.

[0316] 39. The method of any preceding embodiment, wherein determining the level of CD74 expression comprises determining a level of CD74 expression by tumour stromal cells. 40. The method of any preceding embodiment, wherein determining the level of CD74 expression comprises determining a level of CD74 expression by tumour epithelial cells.

[0317] 41. The method of any preceding embodiment, wherein determining the level of CD74 expression comprises determining a level of CD74 expression by tumour cells.

[0318] 42. The method of any preceding embodiment, wherein determining the level of CD74 expression comprises determining a level of CD74 expression by tumour and stromal cells.

[0319] 43. The method of any preceding embodiment, wherein determining the level of CD74 expression comprises assessing a level of CD74 expression by macrophages in the sample.

[0320] 44. The method of any preceding embodiment, wherein determining the level of CD74 expression comprises determining a level of CD74 expression by macrophages in the tumour stroma.

[0321] 45. The method of any preceding embodiment, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 15% of macrophages in the tumour stroma express CD74.

[0322] 46. The method of any preceding embodiment, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 17% of macrophages in the tumour stroma express CD74.

[0323] 47. The method of any preceding embodiment, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 19% of macrophages in the tumour stroma express CD74.

[0324] 48. The method of any preceding embodiment, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 21 % of macrophages in the tumour stroma express CD74.

[0325] 49. The method of any preceding embodiment, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 23% of macrophages in the tumour stroma express CD74.

[0326] 50. The method of any preceding embodiment, wherein determining the level of CD74 expression comprises determining a level of CD74 protein. 51. The method of any preceding embodiment, wherein determining the level of CD74 expression comprises determining a level of cell surface expression of CD74.

[0327] 52. The method of any preceding embodiment, wherein determining a level of CD74 expression comprises determining a level of CD74 DNA, CD74 RNA, or a combination thereof.

[0328] 53. The method of any preceding embodiment, wherein determining the level of CD74 expression comprises immunohistochemistry, gene expression analysis, flow cytometry, proteomics, transcriptomics, in situ hybridisation, Western blot, in-situ immunofluorescence, imaging mass cytometry, ELISA, or any combination thereof.

[0329] 54. The method of any preceding embodiment, wherein the subject has been identified as having colorectal cancer with a tumour mutational burden of fewer than 12 mutations per megabase.

[0330] 55. The method of any preceding embodiment, wherein the subject has been identified as having colorectal cancer with a tumour mutational burden of fewer than 11 mutations per megabase.

[0331] 56. The method of any preceding embodiment, wherein the subject has been identified as having colorectal cancer with a tumour mutational burden of fewer than 10 mutations per megabase.

[0332] 57. The method of any preceding embodiment, wherein the subject has been identified as having colorectal cancer with a tumour mutational burden of 9 or fewer mutations per megabase.

[0333] 58. The method of any preceding embodiment, wherein the sample is a tumour sample.

[0334] 59. The method of any preceding embodiment, wherein the sample is a tumour biopsy.

[0335] 60. The method of any preceding embodiment, wherein the immunotherapeutic agent comprises an immune checkpoint inhibitor.

[0336] 61. The method of any preceding embodiment, wherein the immunotherapeutic agent comprises an agent that acts on the PD-1 / PD-L1 axis.

[0337] 62. The method of any preceding embodiment, wherein the immunotherapeutic agent comprises a PD-1 inhibitor, a PD-L1 inhibitor, or a combination thereof.

[0338] 63. The method of any preceding embodiment, wherein the immunotherapeutic agent comprises an anti-PD-1 antibody, an anti-PD-L1 antibody, or a combination thereof. 64. The method of any preceding embodiment, wherein the immunotherapeutic agent comprises an antagonistic anti-PD-1 antibody, an antagonistic anti-PD-L1 antibody, or a combination thereof.

[0339] 65. The method of any preceding embodiment, wherein the immunotherapeutic agent comprises pembrolizumab, nivolumab, atezolizumab, or a combination thereof.

[0340] 66. The method of any preceding embodiment, wherein the immunotherapeutic agent comprises a CTLA-4 inhibitor.

[0341] 67. The method of any preceding embodiment, wherein the immunotherapeutic agent comprises an anti-CTLA-4 antibody.

[0342] 68. The method of any preceding embodiment, wherein the immunotherapeutic agent comprises ipilimumab, tremelimumab, or a combination thereof.

[0343] 69. The method of any preceding embodiment, wherein the method further comprises administering a therapeutically effective amount of the immunotherapeutic agent to the subject if the level of CD74 in the sample is higher than the reference value.

[0344] 70. A method of selecting a subject for treatment with anti-PD-1 or anti-PD-L1 immunotherapy, wherein the subject has mismatch repair proficient (pMMR) colorectal cancer, the method comprising: a. providing a tumour sample obtained from the subject, b. determining a percentage of tumour stromal cells in the sample that express CD74, and c. selecting the subject for treatment with anti-PD-1 or anti-PD-L1 immunotherapy if greater than 19% of tumour stromal cells in the sample express CD74.

[0345] 71. A method of treating colorectal cancer in a subject, wherein the subject has been identified as having mismatch repair proficient (pMMR) colorectal cancer, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; b. comparing the level of CD74 expression in the sample to a reference value; and c. administering a therapeutically effective amount of an immunotherapeutic agent to the subject if the level of CD74 expression in the sample is higher than the reference value.

[0346] 72. A method of selecting a treatment regimen for a subject having mismatch repair proficient (pMMR) colorectal cancer, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; b. comparing the level of CD74 expression in the sample to a reference value; and c. selecting a treatment regimen for the subject if the level of CD74 expression in the sample is higher than the reference value. The method of embodiment 72, wherein the treatment regimen comprises treatment with an immunotherapeutic agent, such as an immune checkpoint inhibitor. A method of stratifying mismatch repair proficient (pMMR) colorectal cancer patients by response to immunotherapy, the method comprising: a. determining a level of CD74 expression in samples obtained from the patients; and b. comparing the level of CD74 expression in the samples to a reference value; wherein patients with a CD74 expression level higher than the reference value are classified as responders, and patients with a CD74 expression level lower than the reference value are classified as non-responders. A method, comprising: a. providing a sample obtained from a subject having mismatch repair proficient (pMMR) colorectal cancer; b. detecting CD74 expression in the sample; and c. analysing the CD74 expression in the sample, wherein said analysing comprises calculating a CD74 stroma positive score (SPS). The method of embodiment 75, wherein said detecting comprises treating the sample with an agent that labels CD74, such as an anti-CD74 antibody. The method of embodiment 75 or 76, wherein said detecting is performed using immunohistochemistry. A PD-1 inhibitor or PD-L1 inhibitor for use in a method of treating colorectal cancer in a subject, wherein the subject has been identified as having mismatch repair proficient (pMMR) colorectal cancer. The PD-1 inhibitor or PD-L1 inhibitor for use of embodiment 78, wherein a tumour sample obtained from the subject has been identified as having CD74 expression that is greater than a reference value. 80. The PD-1 inhibitor or PD-L1 inhibitor for use of embodiment 78 or 79, wherein the subject has been identified as being a responder to immunotherapy by the method of any of embodiments 1-70.

[0347] 81. A PD-1 inhibitor or PD-L1 inhibitor for use in a method of treating mismatch repair proficient (pMMR) colorectal cancer in a subject, the method comprising determining a level of CD74 expression in a sample obtained from the subject, and administering a therapeutically effective amount of the PD-1 inhibitor or PD-L1 inhibitor to the subject if the level of CD74 expression in the sample is higher than a reference value.

[0348] 82. A method of selecting a subject for treatment with an immunotherapeutic agent, wherein the subject has mismatch repair deficient (dMMR) colorectal cancer, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; and b. comparing the level of CD74 expression in the sample to a reference value; wherein the subject is selected for treatment with the immunotherapeutic agent if the level of CD74 in the sample is higher than the reference value.

[0349] 83. A method for predicting or assessing the response of a subject to treatment with an immunotherapeutic agent, wherein the subject has colorectal cancer with a low tumour mutational burden, the method comprising: determining a level of CD74 expression in a sample obtained from the subject and comparing it to a reference value, wherein a CD74 expression level higher than the reference value is indicative of an immunotherapy responder, and wherein a CD74 expression level lower than the reference value is indicative of an immunotherapy non-responder.

[0350] 84. A method of selecting a subject for treatment with an immunotherapeutic agent, wherein the subject has colorectal cancer with a low tumour mutational burden, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; and b. comparing the level of CD74 expression in the sample to a reference value; wherein the subject is selected for treatment with the immunotherapeutic agent if the level of CD74 in the sample is higher than the reference value.

[0351] 85. The method of embodiment 83 or 84, wherein the reference value is a percentage of cells in the sample that express CD74.

[0352] 86. The method of embodiment 83 or 84, wherein the reference value is a percentage of the sample that is positive for CD74. 87. The method of embodiment 83 or 84, wherein the reference value is a percentage of stromal cells in the sample that express CD74.

[0353] 88. The method of embodiment 83 or 84, wherein the reference value is a percentage of the stroma that is positive for CD74.

[0354] 89. The method of embodiment 83 or 84, wherein the reference value is a percentage of tumour stromal cells in the sample that express CD74.

[0355] 90. The method of embodiment 83 or 84, wherein the reference value is a percentage of the tumour stroma that is positive for CD74.

[0356] 91. The method of embodiment 83 or 84, wherein the reference value is a percentage of tumour epithelial cells in the sample that express CD74.

[0357] 92. The method of embodiment 83 or 84, wherein the reference value is a percentage of the tumour epithelium that is positive for CD74.

[0358] 93. The method of embodiment 83 or 84, wherein the reference value is a percentage of tumour cells in the sample that express CD74.

[0359] 94. The method of embodiment 83 or 84, wherein the reference value is a percentage of the tumour that is positive for CD74.

[0360] 95. The method of any of embodiments 83-94, wherein the reference value is 15%.

[0361] 96. The method of any of embodiments 83-95, wherein the reference value is 16%.

[0362] 97. The method of any of embodiments 83-96, wherein the reference value is 17%.

[0363] 98. The method of any of embodiments 83-97, wherein the reference value is 18%.

[0364] 99. The method of any of embodiments 83-98, wherein the reference value is 19%.

[0365] 100. The method of any of embodiments 83-99, wherein the reference value is 20%.

[0366] 101 . The method of any of embodiments 83-100, wherein the reference value is 21%.

[0367] 102. The method of any of embodiments 83-101 , wherein the reference value is 22%.

[0368] 103. The method of any of embodiments 83-102, wherein the reference value is 23%. 104. The method of any of embodiments 83-103, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 15% of the cells in the sample express CD74.

[0369] 105. The method of any of embodiments 83-104, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 17% of the cells in the sample express CD74.

[0370] 106. The method of any of embodiments 83-105, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 19% of the cells in the sample express CD74.

[0371] 107. The method of any of embodiments 83-106, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 21% of the cells in the sample express CD74.

[0372] 108. The method of any of embodiments 83-107, wherein the reference value is a CD74 score.

[0373] 109. The method of any of embodiments 83-108, wherein the reference value is a CD74 stroma-positive score (SPS).

[0374] 110. The method of any of embodiments 83, 84, or 95-109, wherein the reference value is a CD74 stroma-positive score (SPS) of 15%.

[0375] 111. The method of any of embodiments 83, 84, or 96-110, wherein the reference value is a CD74 stroma-positive score (SPS) of 17%.

[0376] 112. The method of any of embodiments 83, 84, or 97-111 , wherein the reference value is a CD74 stroma-positive score (SPS) of 19%.

[0377] 113. The method of any of embodiments 83, 84, or 98-112, wherein the reference value is a CD74 stroma-positive score (SPS) of 21%.

[0378] 114. The method of any of embodiments 83, 84, or 99-113, wherein the reference value is a CD74 stroma-positive score (SPS) of 23%.

[0379] 115. The method of any of embodiments 83, 84, or 98-108, wherein the reference value is a CD74 tumour-positive score (TPS). 116. The method of any of embodiments 83, 84, or 98-108, wherein the reference value is a CD74 combined positive score (CPS).

[0380] 117. The method of any of embodiments 83, 84, or 98-108, wherein the reference value is a CD74 stroma positive against tumour (SPT) score.

[0381] 118. The method of any of embodiments 83-117, wherein determining the level of CD74 expression comprises determining a level of CD74 expression in the tumour stroma.

[0382] 119. The method of any of embodiments 83-118, wherein determining the level of CD74 expression comprises determining a level of CD74 expression in the tumour epithelium.

[0383] 120. The method of any of embodiments 83-119, wherein determining the level of CD74 expression comprises determining a level of CD74 expression by stromal cells.

[0384] 121. The method of any of embodiments 83-120, wherein determining the level of CD74 expression comprises determining a level of CD74 expression by tumour stromal cells.

[0385] 122. The method of any of embodiments 83-121 , wherein determining the level of CD74 expression comprises determining a level of CD74 expression by tumour epithelial cells.

[0386] 123. The method of any of embodiments 83-122, wherein determining the level of CD74 expression comprises determining a level of CD74 expression by tumour cells.

[0387] 124. The method of any of embodiments 83-123, wherein determining the level of CD74 expression comprises determining a level of CD74 expression by tumour and stromal cells.

[0388] 125. The method of any of embodiments 83-124, wherein determining the level of CD74 expression comprises assessing a level of CD74 expression by macrophages in the sample.

[0389] 126. The method of any of embodiments 83-125, wherein determining the level of CD74 expression comprises determining a level of CD74 expression by macrophages in the tumour stroma.

[0390] 127. The method of any of embodiments 83-126, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 15% of macrophages in the tumour stroma express CD74. 128. The method of any of embodiments 83-127, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 17% of macrophages in the tumour stroma express CD74.

[0391] 129. The method of any of embodiments 83-128, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 19% of macrophages in the tumour stroma express CD74.

[0392] 130. The method of any of embodiments 83-129, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 21% of macrophages in the tumour stroma express CD74.

[0393] 131 . The method of any of embodiments 83-130, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 23% of macrophages in the tumour stroma express CD74.

[0394] 132. The method of any of embodiments 83-131 , wherein determining the level of CD74 expression comprises determining a level of CD74 protein.

[0395] 133. The method of any of embodiments 83-132, wherein determining the level of CD74 expression comprises determining a level of cell surface expression of CD74.

[0396] 134. The method of any of embodiments 83-133, wherein determining a level of CD74 expression comprises determining a level of CD74 DNA, CD74 RNA, or a combination thereof.

[0397] 135. The method of any of embodiments 83-134, wherein determining the level of CD74 expression comprises immunohistochemistry, gene expression analysis, flow cytometry, proteomics, transcriptomics, in situ hybridisation, Western blot, in-situ immunofluorescence, imaging mass cytometry, ELISA, or any combination thereof.

[0398] 136. The method of any of embodiments 83-135, wherein the subject has been identified as having colorectal cancer with a tumour mutational burden of fewer than 12 mutations per megabase.

[0399] 137. The method of any of embodiments 83-136, wherein the subject has been identified as having colorectal cancer with a tumour mutational burden of fewer than 11 mutations per megabase. 138. The method of any of embodiments 83-137, wherein the subject has been identified as having colorectal cancer with a tumour mutational burden of fewer than 10 mutations per megabase.

[0400] 139. The method of any of embodiments 83-138, wherein the subject has been identified as having colorectal cancer with a tumour mutational burden of 9 or fewer mutations per megabase.

[0401] 140. The method of any of embodiments 83-139, wherein the sample is a tumour sample.

[0402] 141. The method of any of embodiments 83-140, wherein the sample is a tumour biopsy.

[0403] 142. The method of any of embodiments 83-141 , wherein the immunotherapeutic agent comprises an immune checkpoint inhibitor.

[0404] 143. The method of any of embodiments 83-142, wherein the immunotherapeutic agent comprises an agent that acts on the PD-1 / PD-L1 axis.

[0405] 144. The method of any of embodiments 83-143, wherein the immunotherapeutic agent comprises a PD-1 inhibitor, a PD-L1 inhibitor, or a combination thereof.

[0406] 145. The method of any of embodiments 83-144, wherein the immunotherapeutic agent comprises an anti-PD-1 antibody, an anti-PD-L1 antibody, or a combination thereof.

[0407] 146. The method of any of embodiments 83-145, wherein the immunotherapeutic agent comprises an antagonistic anti-PD-1 antibody, an antagonistic anti-PD-L1 antibody, or a combination thereof.

[0408] 147. The method of any of embodiments 83-146, wherein the immunotherapeutic agent comprises pembrolizumab, nivolumab, atezolizumab, or a combination thereof.

[0409] 148. The method of any of embodiments 83-147, wherein the immunotherapeutic agent comprises a CTLA-4 inhibitor.

[0410] 149. The method of any of embodiments 83-148, wherein the immunotherapeutic agent comprises an anti-CTLA-4 antibody.

[0411] 150. The method of any of embodiments 83-149, wherein the immunotherapeutic agent comprises ipilimumab, tremelimumab, or a combination thereof. 151 . The method of any of embodiments 83-150, wherein the method further comprises administering a therapeutically effective amount of the immunotherapeutic agent to the subject if the level of CD74 in the sample is higher than the reference value.

[0412] 152. A method of selecting a subject for treatment with anti-PD-1 or anti-PD-L1 immunotherapy, wherein the subject has colorectal cancer with a low tumour mutational burden, the method comprising: a. providing a tumour sample obtained from the subject, b. determining a percentage of tumour stromal cells in the sample that express CD74, and c. selecting the subject for treatment with anti-PD-1 or anti-PD-L1 immunotherapy if greater than 19% of tumour stromal cells in the sample express CD74.

[0413] 153. A method of treating colorectal cancer in a subject, wherein the subject has been identified as having colorectal cancer with a low tumour mutational burden, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; b. comparing the level of CD74 expression in the sample to a reference value; and c. administering a therapeutically effective amount of an immunotherapeutic agent to the subject if the level of CD74 expression in the sample is higher than the reference value.

[0414] 154. A method of selecting a treatment regimen for a subject having colorectal cancer with a low tumour mutational burden, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; b. comparing the level of CD74 expression in the sample to a reference value; and c. selecting a treatment regimen for the subject if the level of CD74 expression in the sample is higher than the reference value.

[0415] 155. The method of embodiment 154, wherein the treatment regimen comprises treatment with an immunotherapeutic agent, such as an immune checkpoint inhibitor.

[0416] 156. A method of stratifying low TMB colorectal cancer patients by response to immunotherapy, the method comprising: a. determining a level of CD74 expression in samples obtained from the patients; and b. comparing the level of CD74 expression in the samples to a reference value; wherein patients with a CD74 expression level higher than the reference value are classified as responders, and patients with a CD74 expression level lower than the reference value are classified as non-responders.

[0417] 157. A method, comprising: a. providing a sample obtained from a subject having colorectal cancer with a low tumour mutational burden; b. detecting CD74 expression in the sample; and c. analysing the CD74 expression in the sample, wherein said analysing comprises calculating a CD74 stroma positive score (SPS).

[0418] 158. The method of embodiment 157, wherein said detecting comprises treating the sample with an agent that labels CD74, such as an anti-CD74 antibody.

[0419] 159. The method of embodiment 157 or 158, wherein said detecting is performed using immunohistochemistry.

[0420] 160. A PD-1 inhibitor or PD-L1 inhibitor for use in a method of treating colorectal cancer in a subject, wherein the subject has been identified as having colorectal cancer with a low tumour mutational burden.

[0421] 161. The PD-1 inhibitor or PD-L1 inhibitor for use of embodiment 160, wherein a tumour sample obtained from the subject has been identified as having CD74 expression that is greater than a reference value.

[0422] 162. The PD-1 inhibitor or PD-L1 inhibitor for use of embodiment 160 or 161 , wherein the subject has been identified as being a responder to immunotherapy by the method of any of embodiments 83-152.

[0423] 163. An PD-1 inhibitor or PD-L1 inhibitor for use in a method of treating colorectal cancer in a subject, the method comprising determining a level of CD74 expression in a sample obtained from the subject, and administering a therapeutically effective amount of the PD-1 inhibitor or PD-L1 inhibitor to the subject if the level of CD74 expression in the sample is higher than a reference value.

[0424] 164. A method of selecting a subject for treatment with an immunotherapeutic agent, wherein the subject has colorectal cancer with a high tumour mutational burden, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; and b. comparing the level of CD74 expression in the sample to a reference value; wherein the subject is selected for treatment with the immunotherapeutic agent if the level of CD74 in the sample is higher than the reference value.

[0425] 165. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a percentage of cells in the sample that express CD74.

[0426] 166. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a percentage of the sample that is positive for CD74.

[0427] 167. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a percentage of stromal cells in the sample that express CD74.

[0428] 168. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a percentage of the stroma that is positive for CD74.

[0429] 169. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a percentage of tumour stromal cells in the sample that express CD74.

[0430] 170. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a percentage of the tumour stroma that is positive for CD74.

[0431] 171. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a percentage of tumour epithelial cells in the sample that express CD74.

[0432] 172. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a percentage of the tumour epithelium that is positive for CD74.

[0433] 173. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a percentage of tumour cells in the sample that express CD74. 174. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a percentage of the tumour that is positive for CD74.

[0434] 175. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is 15%.

[0435] 176. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is 16%.

[0436] 177. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is 17%.

[0437] 178. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is 18%.

[0438] 179. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is 19%.

[0439] 180. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is 20%.

[0440] 181. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is 21%.

[0441] 182. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is 22%.

[0442] 183. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is 23%.

[0443] 184. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 15% of the cells in the sample express CD74.

[0444] 185. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 17% of the cells in the sample express CD74. 186. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 19% of the cells in the sample express CD74.

[0445] 187. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the subject is selected for treatment with the immunotherapeutic agent if greater than 21% of the cells in the sample express CD74.

[0446] 188. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a CD74 score.

[0447] 189. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a CD74 stroma-positive score (SPS).

[0448] 190. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a CD74 stroma-positive score (SPS) of 15%.

[0449] 191. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a CD74 stroma-positive score (SPS) of 17%.

[0450] 192. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a CD74 stroma-positive score (SPS) of 19%.

[0451] 193. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a CD74 stroma-positive score (SPS) of 21%.

[0452] 194. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a CD74 stroma-positive score (SPS) of 23%. 195. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a CD74 tumour-positive score (TPS).

[0453] 196. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a CD74 combined positive score (CPS).

[0454] 197. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the reference value is a CD74 stroma positive against tumour (SPT) score.

[0455] 198. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein determining the level of CD74 expression comprises determining a level of CD74 expression in the tumour stroma.

[0456] 199. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein determining the level of CD74 expression comprises determining a level of CD74 expression in the tumour epithelium.

[0457] 200. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein determining the level of CD74 expression comprises determining a level of CD74 expression by stromal cells.

[0458] 201. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein determining the level of CD74 expression comprises determining a level of CD74 expression by tumour stromal cells.

[0459] 202. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein determining the level of CD74 expression comprises determining a level of CD74 expression by tumour epithelial cells.

[0460] 203. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein determining the level of CD74 expression comprises determining a level of CD74 expression by tumour cells.

[0461] 204. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein determining the level of CD74 expression comprises determining a level of CD74 expression by tumour and stromal cells. . The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein determining the level of CD74 expression comprises assessing a level of CD74 expression by macrophages in the sample. . The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein determining the level of CD74 expression comprises determining a level of CD74 expression by macrophages in the tumour stroma. . The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the patient is selected for treatment with the immunotherapeutic agent if greater than 15% of macrophages in the tumour stroma express CD74. . The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the patient is selected for treatment with the immunotherapeutic agent if greater than 17% of macrophages in the tumour stroma express CD74. . The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the patient is selected for treatment with the immunotherapeutic agent if greater than 19% of macrophages in the tumour stroma express CD74. . The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the patient is selected for treatment with the immunotherapeutic agent if greater than 21% of macrophages in the tumour stroma express CD74. . The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the patient is selected for treatment with the immunotherapeutic agent if greater than 23% of macrophages in the tumour stroma express CD74. . The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein determining the level of CD74 expression comprises determining a level of CD74 protein. . The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein determining the level of CD74 expression comprises determining a level of cell surface expression of CD74. . The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein determining a level of CD74 expression comprises determining a level of CD74 DNA, CD74 RNA, or a combination thereof. . The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein determining the level of CD74 expression comprises immunohistochemistry, gene expression analysis, flow cytometry, proteomics, transcriptomics, in situ hybridisation, Western blot, in-situ immunofluorescence, imaging mass cytometry, ELISA, or any combination thereof. . The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the subject has been identified as having colorectal cancer with a tumour mutational burden of fewer than 12 mutations per megabase. . The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the subject has been identified as having colorectal cancer with a tumour mutational burden of fewer than 11 mutations per megabase. . The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the subject has been identified as having colorectal cancer with a tumour mutational burden of fewer than 10 mutations per megabase. . The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the subject has been identified as having colorectal cancer with a tumour mutational burden of 9 or fewer mutations per megabase. . The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the sample is a tumour sample. 221. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the sample is a tumour biopsy.

[0462] 222. The method of any of embodiments 71-77, 82, 153-159 or 164, or the PD-1 or PD-L1 inhibitor for use of any of embodiments 78-81 or 160-163, wherein the immunotherapeutic agent comprises an immune checkpoint inhibitor.

[0463] 223. The method of any of embodiments 71-77, 82, 153-159 or 164, wherein the immunotherapeutic agent comprises an agent that acts on the PD-1 / PD-L1 axis.

[0464] 224. The method of any of embodiments 71-77, 82, 153-159 or 164, wherein the immunotherapeutic agent comprises a PD-1 inhibitor, a PD-L1 inhibitor, or a combination thereof.

[0465] 225. The method of any of embodiments 71-77, 82, 153-159 or 164, wherein the immunotherapeutic agent comprises an anti-PD-1 antibody, an anti-PD-L1 antibody, or a combination thereof.

[0466] 226. The method of any of embodiments 71-77, 82, 153-159 or 164, wherein the immunotherapeutic agent comprises an antagonistic anti-PD-1 antibody, an antagonistic anti- PD-L1 antibody, or a combination thereof.

[0467] 227. The method of any of embodiments 71-77, 82, 153-159 or 164, wherein the immunotherapeutic agent comprises pembrolizumab, nivolumab, atezolizumab, or a combination thereof.

[0468] 228. The method of any of embodiments 71-77, 82, 153-159 or 164, wherein the immunotherapeutic agent comprises a CTLA-4 inhibitor.

[0469] 229. The method of any of embodiments 71-77, 82, 153-159 or 164, wherein the immunotherapeutic agent comprises an anti-CTLA-4 antibody.

[0470] 230. The method of any of embodiments 71-77, 82, 153-159 or 164, wherein the immunotherapeutic agent comprises ipilimumab, tremelimumab, or a combination thereof.

[0471] 231. The method of any of embodiments 71-77, 82, 153-159 or 164, wherein the method further comprises administering a therapeutically effective amount of the immunotherapeutic agent to the subject if the level of CD74 in the sample is higher than the reference value. 232. The method of any of embodiments 75-77 or 157-159, further comprising a step of comparing the SPS to a reference value, optionally wherein the reference value is as defined in any of embodiments 3-35 or 85-103.

[0472] EQUIVALENTS AND SCOPE

[0473] Those skilled in the art will appreciate that the present invention is defined by the appended claims and not by the Examples or other description of certain embodiments included herein.

[0474] Similarly, the singular forms “a”, “an”, and “the” include plural referents unless the context clearly dictates otherwise.

[0475] Unless defined otherwise above, 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 invention belongs. Any methods and materials similar or equivalent to those described herein can also be used in the practice or testing of the present invention. Generally, nomenclatures used in connection with, and techniques of, cell and tissue culture, molecular biology, immunology, genetics and protein and nucleic acid chemistry described herein are those well-known and commonly used in the art, or according to manufacturer’s specifications.

[0476] All publications mentioned herein are incorporated herein by reference to disclose and describe the methods and / or materials in connection with which the publications are cited.

Claims

CLAIMS1 . A method for predicting or assessing the response of a subject to treatment with an immunotherapeutic agent, wherein the subject has mismatch repair proficient (pMMR) colorectal cancer, the method comprising: determining a level of CD74 expression in a sample obtained from the subject and comparing it to a reference value, wherein a CD74 expression level higher than the reference value is indicative of an immunotherapy responder, and wherein a CD74 expression level lower than the reference value is indicative of an immunotherapy non-responder.

2. A method of selecting a subject for treatment with an immunotherapeutic agent, wherein the subject has mismatch repair proficient (pMMR) colorectal cancer, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; and b. comparing the level of CD74 expression in the sample to a reference value; wherein the subject is selected for treatment with the immunotherapeutic agent if the level of CD74 in the sample is higher than the reference value.

3. The method of claim 1 or 2, wherein the reference value is a percentage of cells in the sample that express CD74.

4. The method of claim 1 or 2, wherein the reference value is a percentage of the sample that is positive for CD74.

5. The method of claim 1 or 2, wherein the reference value is a percentage of tumour stromal cells in the sample that express CD74.

6. The method of claim 1 or 2, wherein the reference value is a percentage of the tumour stroma that is positive for CD74.

7. The method of any preceding claim, wherein the reference value is 15%, 16%, 17%, 18%, 19%, 20%, 21%, 22%, or 23%.

8. The method of any preceding claim, wherein the reference value is 19%.

9. The method of any preceding claim, wherein the reference value is a CD74 score.

10. The method of any preceding claim, wherein the reference value is a CD74 stroma-positive score (SPS).

11. The method of any preceding claim, wherein the reference value is a CD74 stroma-positive score (SPS) of 15%, 16%, 17%, 18%, 19%, 20%, 21%, 22%, or 23%.

12. The method of claim 11 , wherein the reference value is a CD74 stroma-positive score (SPS) of 19%.

13. The method of any preceding claim, wherein the subject has been identified as having colorectal cancer with a tumour mutational burden of fewer than 10 mutations per megabase.

14. The method of any preceding claim, wherein the sample is a tumour sample.

15. The method of any preceding claim, wherein the immunotherapeutic agent comprises an immune checkpoint inhibitor.

16. The method of any preceding claim, wherein the immunotherapeutic agent comprises an agent that acts on the PD-1 / PD-L1 axis.

17. The method of any preceding claim, wherein the immunotherapeutic agent comprises a PD-1 inhibitor, a PD-L1 inhibitor, or a combination thereof.

18. The method of any preceding claim, wherein the immunotherapeutic agent comprises an anti- PD-1 antibody, an anti-PD-L1 antibody, or a combination thereof.

19. The method of any preceding claim, wherein the immunotherapeutic agent comprises pembrolizumab, nivolumab, atezolizumab, or a combination thereof.

20. The method of any preceding claim, wherein the method further comprises administering a therapeutically effective amount of the immunotherapeutic agent to the subject if the level of CD74 in the sample is higher than the reference value.

21. A method of selecting a subject for treatment with anti-PD-1 or anti-PD-L1 immunotherapy, wherein the subject has mismatch repair proficient (pMMR) colorectal cancer, the method comprising: a. providing a tumour sample obtained from the subject, b. determining a percentage of tumour stromal cells in the sample that express CD74, and c. selecting the subject for treatment with anti-PD-1 or anti-PD-L1 immunotherapy if greater than 19% of tumour stromal cells in the sample express CD74.

22. A method of treating colorectal cancer in a subject, wherein the subject has been identified as having mismatch repair proficient (pMMR) colorectal cancer, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; b. comparing the level of CD74 expression in the sample to a reference value; and c. administering a therapeutically effective amount of an immunotherapeutic agent to the subject if the level of CD74 expression in the sample is higher than the reference value.

23. A method of selecting a treatment regimen for a subject having mismatch repair proficient (pMMR) colorectal cancer, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; b. comparing the level of CD74 expression in the sample to a reference value; and c. selecting a treatment regimen for the subject if the level of CD74 expression in the sample is higher than the reference value.

24. The method of claim 23, wherein the treatment regimen comprises treatment with an immunotherapeutic agent, such as an immune checkpoint inhibitor.

25. A method of stratifying mismatch repair proficient (pMMR) colorectal cancer patients by response to immunotherapy, the method comprising: a. determining a level of CD74 expression in samples obtained from the patients; and b. comparing the level of CD74 expression in the samples to a reference value; wherein patients with a CD74 expression level higher than the reference value are classified as responders, and patients with a CD74 expression level lower than the reference value are classified as non-responders.

26. A method, comprising: a. providing a sample obtained from a subject having mismatch repair proficient (pMMR) colorectal cancer; b. detecting CD74 expression in the sample; and c. analysing the CD74 expression in the sample, wherein said analysing comprises calculating a CD74 stroma positive score (SPS).

27. The method of claim 26, wherein said detecting comprises treating the sample with an agent that labels CD74, such as an anti-CD74 antibody.

28. The method of claim 26 or 27, wherein said detecting is performed using immunohistochemistry.

29. A PD-1 inhibitor or PD-L1 inhibitor for use in a method of treating colorectal cancer in a subject, wherein the subject has been identified as having mismatch repair proficient (pMMR) colorectal cancer.

30. The PD-1 inhibitor or PD-L1 inhibitor for use of claim 29, wherein a tumour sample obtained from the subject has been identified as having CD74 expression that is greater than a reference value.

31. The PD-1 inhibitor or PD-L1 inhibitor for use of claim 29 or 30, wherein the subject has been identified as being a responder to immunotherapy by the method of any of claims 1-21 or 25.

32. A PD-1 inhibitor or PD-L1 inhibitor for use in a method of treating mismatch repair proficient (pMMR) colorectal cancer in a subject, the method comprising determining a level of CD74 expression in a sample obtained from the subject, and administering a therapeutically effective amount of the PD-1 inhibitor or PD-L1 inhibitor to the subject if the level of CD74 expression in the sample is higher than a reference value.

33. A method for predicting or assessing the response of a subject to treatment with an immunotherapeutic agent, wherein the subject has colorectal cancer with a low tumour mutational burden, the method comprising: determining a level of CD74 expression in a sample obtained from the subject and comparing it to a reference value, wherein a CD74 expression level higher than the reference value is indicative of an immunotherapy responder, and wherein a CD74 expression level lower than the reference value is indicative of an immunotherapy non-responder.

34. A method of selecting a subject for treatment with an immunotherapeutic agent, wherein the subject has colorectal cancer with a low tumour mutational burden, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; and b. comparing the level of CD74 expression in the sample to a reference value; wherein the subject is selected for treatment with the immunotherapeutic agent if the level of CD74 in the sample is higher than the reference value.

35. A method of selecting a subject for treatment with an immunotherapeutic agent, wherein the subject has mismatch repair deficient (dMMR) colorectal cancer, the method comprising: a. determining a level of CD74 expression in a sample obtained from the subject; and b. comparing the level of CD74 expression in the sample to a reference value;wherein the subject is selected for treatment with the immunotherapeutic agent if the level of CD74 in the sample is higher than the reference value.

Images