Gene signature for prediction of treatment response in cancers with microsatellite instability or mismatch repair deficiency

Abstract

A gene signature for evaluating likely sensitivity of a cancer cell to treatment with inhibitors of Werner syndrome RecQ helicase (WRN) is provided. For cancer cells that have microsatellite instability (MSI) and / or mismatch repair deficiency (dMMR), sensitivity to treatment with a WRN inhibitor is likely if the cancer cell has a loss of function mutation in at least two of KMT2D, ARID1A or RAD50, or if the cancer cell has a loss of function mutation in ARID1A only, without loss of function mutations in KMT2D and RAD50. Cancer cells that are MSI and / or dMMR in which both ARID1A and at least one of KMT2D and RAD50 are wild type or have only missense mutations are unlikely to be sensitive to treatment with WRN inhibitors.

Description

GENE SIGNATURE FOR PREDICTION OF TREATMENT RESPONSE IN CANCERS WITH MICROSATELLITE INSTABILITY OR MISMATCH REPAIR DEFICIENCYCROSS-REFRENCE TO RELATED APPLICATIONS

[0001] This application claims priority to, and the benefit of, United States provisional patent application No. 63 / 679964 filed 6 August 2024. The foregoing application is incorporated by reference herein in its entirety for all purposes.TECHNICAL FIELD

[0002] Some embodiments relate to the detection of mutations in the genes K.MT2D (histone-lysine N-methyltransferase 2D), ARID1A (AT-rich interaction domain 1A) and RAD50 for use as diagnostic biomarkers to predict the clinical efficacy of agents, including but not limited to small molecule inhibitors, that inhibit the protein activity and / or function of the gene Werner syndrome RecQ helicase (1 / R / V) in the context of cancers having microsatellite instability (MSI) or mismatchrepair deficiency (dMMR).BACKGROUND

[0003] Werner syndrome RecQ helicase (1 / R / V) is an enzyme important for DNA repair and genome stability. WRN encodes a DNA helicase and nuclease that functions in DNA replication, repair, and transcription [1-3] and has recently been identified as a synthetic lethal (SL) vulnerability of mismatch repair-deficient (MMR- deficient) (dMMR) MSI cancers, including some colorectal, endometrial, gastric, and ovarian cancers [4-9]. There are currently several drug development programs underway to target WRN [10,11] in MSI / dM MR cancer, and WRN has recently been identified as a potential synthetic lethal target for cancers having microsatellite instability (MSI) or mismatch-repair deficient (dMMR) cancers. Many patients with MSI solid tumours do not respond to immune checkpoint inhibitors (ICI) or develop resistance, e.g. Yap et al., 2025

[0012] . Inhibitors of WRN offer a potential alternative therapy for such patients. However, early reports by Ferretti et al.

[0013] and Baltgalvis et al.

[0014] relating to WRN inhibitors, HRO761 and VVD-133214, respectively, which are in phase I clinical trials (NCT05838768 and NCT06004245)recruiting patients with advanced metastatic solid tumours that are MSI and / or MMR gene-deficient (dMMR), reported MSI COAD / READ cell lines that were insensitive to WRN inhibitors. This suggests that MSI and / or dMMR alone is not a sufficient predictor of the sensitivity of cancers to treatment using WRN inhibitors and that additional genetic context may be relevant to appropriately identify patients who are likely to respond to therapy with WRN inhibitors.

[0004] WRN inhibitors may also have potential utility in combination therapy with immune checkpoint inhibitors (ICI). Clinical trial NCT06004245 is a Phase I study which may examine combination therapy using WRN inhibitor VVD-133214 in combination with the immune checkpoint inhibitor pembrolizumab, and early results suggest that WRN inhibitors seem to be well tolerated and safe when used after treatment with immune checkpoint inhibitors or in combination with ICI treatment. Further, it is known based on an I CI -refractory mouse model that treatment with WRN inhibitors is still effective at reducing tumour volume

[0014] .

[0005] There remains an unmet need in the field for new approaches / methods to identify drug or ICI (immune checkpoint inhibitor) targets and other cellular vulnerabilities in cells, tissues or patients with loss-of-function mutations in tumour suppressor genes. In particular there remains a need for novel methods to identify / stratify cancer patients who have been diagnosed with MSI (microsatellite instability)-positive and / or mismatch repair deficient (dMMR) cancers that will or will not respond to inhibitors of the gene / protein WRN to assist in selecting and administering an appropriate therapy to patients.SUMMARY

[0006] In some aspects, a method of determining whether a cancer of a mammalian subject is likely to be sensitive to inhibitors of Werner syndrome RecQ helicase (WRN) is provided. In some aspects, cells of the cancer are evaluated for microsatellite instability (MSI) or mismatch repair deficiency (dMMR). In some aspects, the cells of the cancer are evaluated for inactivating alterations in two or all of KMT2D, RAD50 and ARID1A and, if the cells of the cancer are MSI or dMMR and either somatic inactivating alterations are present in two or all of KMT2D, RAD50 and ARID1A or somatic inactivating alterations are present in ARID1A,determining that the cancer is likely to be sensitive to inhibitors of WRN. If it is determined that the cancer is MSS, or if it is determined that the cells of the cancer are MSI or dMMR and either ARID1A is wild type or has only missense mutations or one of KMT2D and RAD50 are wild type or have only missense mutations, then it is determined that the cancer is likely to be insensitive to inhibitors of WRN.

[0007] In some aspects, all of KMT2D, RAD50 and ARID1A are evaluated for the presence of inactivating alterations. In some aspects, only KMT2D, RAD50 and ARID1A are evaluated for the presence of inactivating alterations.

[0008] In some aspects, sensitivity of a mammalian subject to combination therapy using inhibitors of WRN and immune checkpoint inhibitors is evaluated by determining if the cancer is likely to be sensitive to immune checkpoint inhibitors and determining if the cancer is likely to be sensitive to inhibitors of WRN as described above.

[0009] In some aspects, an appropriate therapy for a patient including a WRN inhibitor is selected based on the determination that a cancer of the patient is likely to be sensitive to inhibitors of WRN, and in some aspects, the appropriate therapy including a WRN inhibitor is administered to the patient based on such determination. In some aspects, a method of identifying a patient as a candidate for therapy using an inhibitor of WRN is provided based on a determination that a cancer of the patient is likely to be sensitive to inhibitors of WRN, and in some aspects, the appropriate therapy including a WRN inhibitor is administered to the patient based on such determination.

[0010] In some aspects, a kit for predicting sensitivity of a cancer patient to treatment with inhibitors of Werner syndrome RecQ helicase (WRN) is provided. In some aspects, the kit includes or consists of reagents for sequencing two or all of KMT2D, RAD50, and ARID1A. In some aspects, the kit includes or consists of primers specific for two or all of KMT2D, RAD50, and ARID1A. In some aspects, the kit includes instructions for using the reagents to sequence two or all of KMT2D, RAD50, and ARID1A in a sample of a cancer and in a sample of normal tissue from the patient to evaluate the presence of somatic inactivating alterations in two or all of KMT2D, RAD50, and ARID1A. In some aspects, the instructions direct a user to determine if somatic inactivating alterations are present in two or all of KMT2D,RAD50 and ARID1A in the cells of the cancer; and, if it is determined that somatic inactivating alterations are present in two or all of KMT2D, RAD50 and ARID1A in the cells of the cancer, or if it is determined that somatic inactivating alterations are present in ARID1A in the cells of the cancer, the instructions direct the user to conclude that the cancer patient is likely to be sensitive to treatment with inhibitors of WRN. In some aspects, the instructions direct a user to conclude that the cancer patient is unlikely to be sensitive to treatment with inhibitors of WRN if it is determined that ARID1A is wild type or has only a missense mutation in the cells of the cancer; and one or both of KMT2D and RAD50 are wild type or have only a missense mutation in the cells of the cancer.

[0011] Further aspects and embodiments will become apparent with reference to the following drawings and detailed description, which is illustrative and not limiting in nature.BRIEF DESCRIPTION OF THE DRAWINGS

[0012] FIG. 1A shows a lollipop plot of single nucleotide variants (SNVs) and small insertions and deletions in KMT2D identified by the inventors in DepMap cell lines, by cancer type. FIG. IB and FIG. 10 show equivalent lollipop plots for SNVs and small insertions and deletions identified by the inventors in ARID1A and RAD50, respectively.

[0013] FIG. 2 shows an example embodiment of a method for determining if a cancer is likely to be sensitive or insensitive to inhibitors of WRN.

[0014] FIG. 3 shows an example embodiment of a method for determining if a cancer is likely to be sensitive to combination therapy with inhibitors of WRN and immune checkpoint inhibitors.

[0015] FIGS. 4A-4B show the identification of pan-cancer and cancer typespecific KMT2D genetic interaction networks. FIG. 4A on the left shows the percentage of KMT2D small nucleotide mutations (SNVs) detected across 35 cancer types and a pan-cancer dataset. FIG. 4A in the middle shows the percentage of SNVs that are K.MT2D loss-of-function (LOF) mutations. FIG.4A on the right shows the number of DepMap cancer cell lines for each category. Dotted lines show the threshold for each panel. Underlined labels indicate the ten datasets that passed allthree thresholds and were selected for in silico screening. See Methods for cancer type abbreviations.

[0016] FIG. 4B shows ranked genetic interaction (GI) score summary for each of the ten in silico screens. (Bottom right corners) The total numbers of candidates. Class A, B, and the candidate with the highest and lowest GI scores are labelled.

[0017] FIGS. 5A-5E show data supporting that K T2DL0FTCGA-C0AD / READ cases may be vulnerable to ICI treatment. FIG. 5A shows Kaplan-Meier curves (top) and risk table (bottom) of overall survival of MSI TCGA-COAD / READ KMT2D'mand KMT2DL0Fcases. FIGs. 5B-5E show a comparison of ICI response markers TMB (FIG. 5B), neoantigen scores (FIG. 5C), of single base pair substitution (SBS) mutational signature (FIG. 5D), and PredictIO ICI response scores (FIG. 5E) between TCGA-COAD / READ KMT2D'mand KMT2DL0FMSI / MSS cases.

[0018] FIGS. 6A and 6B show data demonstrating that ICI response markers are elevated in KMT2DL0FPOG-MSI solid cancer cohort cases. A comparison of ICI response markers between KMT2D'mand KMT2DL0FMSI / MSS cases in POG- COAD / READ and solid cancer cohorts. ICI response markers included: (FIG. 6A) TMB, and (FIG. 6B) PredictIO ICI response score. Welch's t-test p-value + < 0.1, * < 0.05, ** < 0.01, *** < 0.001, and NS > 0.1. No statistical analysis was performed for the MSI POG-COAD / READ case, due to small sample size (N < 3). FIG. 6C shows TMB of MSK-IMPACT cohort. A comparison of TMB between KMT2Dmand KMT2DL0FMSI / MSS cases in MSK-IMPACT solid cancer cohorts. Welch's t-test p- value * < 0.05 and *** < 0.001.

[0019] FIGs. 7A-7E show data supporting that KMT2D mutational status stratifies WRN inhibitor-sensitive MSI COAD / READ cell lines. FIGs. 7A-7B show respectively WRN (A) and KMT2D (B) KO lethality probabilities in KMT2D (A) and WRN (B) WT and LOF cancer cell lines. FIG. 7C shows [AT]nmicrosatellite repeat expansion size estimated by ExpansionHunter Denovo in six DepMap KMT2DL0Fcell lines from panel A that had WGS data and three KMT2D'm+lines. WT+ denotes both wildtype alleles and missense mutations. Dashed lines indicate the average [AT]nmicrosatellite repeat expansion size across all events. FIG. 7D shows K.MT2D and MMR gene mutational status annotated using DepMap data for MSI COAD / READ cell lines treated with the WRN inhibitors HRO761 and VVD-133214, described inFerretti et al. (2024) and Baltgalvis et al. (2024), respectively [13,14]. Cancer cell line names in red font (KM12L4 AND LS174T) indicate cell lines that did not have DepMap mutation information to verify mutational status. FIG. 7E shows a contingency table comparing the frequency of KMT2DL0Falterations and WRN inhibitor-treatment sensitivity in MSI COAD / READ cell lines in panel FIG. 7D. WT+Mis. denotes both wildtype allele and missense mutations. KM12L4 and LS174T cell lines were excluded from this analysis due to a lack of KMT2D status information.

[0020] FIGs. 8A-8D show data supporting that MSI cancer cells that require WRN for survival frequently harbour KMT2DL0Fmutations. FIG. 8A shows KMT2D and MMR gene mutational status annotated using DepMap data for COAD / READ, STAD, UCEC, and OV cancer cell lines, including MSI and MSS lines, in which WRN perturbation was performed, as described in Behan et al. (2019) [4], Chan et al. (2019) [6], Kategaya et al. (2019) [5], Lieb et al. (2019) [7]. FIG. 8B, 8C, and 8D show contingency tables comparing the frequency of mutations and WRN perturbation sensitivity of cell lines in FIG. 8A. Specifically, FIG. 8B compares the frequency of KMT2DL0Falterations and WRN perturbation sensitivity, FIG. 8C compares the frequency of MMR genes LOAalterations and WRN perturbation sensitivity, and FIG. 8D compares the frequency of MMR genes LOAalterations and KMT2DL0Falterations. WT+ denotes both wildtype alleles and missense mutations.

[0021] FIGS. 9A and 9B show KMT2D protein expression levels in MSI COAD / READ cell lines. FIG. 9A shows a representative western blot image of KMT2D, WRN, and VINC (house keeping control) expression from one replicate. FIG. 9B shows expression of KMT2D relative to Ponceau S staining, total protein loaded. Each point represents one technical replicate. Error bars shows standard deviation.

[0022] FIG. 10 shows the top 10 most frequently altered WRN SL partners in pan-cancer cell lines with validated WRN KO effects.DESCRIPTION

[0023] Throughout the following description, specific details are set forth in order to provide a more thorough understanding of the invention. However, the inventionmay be practiced without these particulars. In other instances, well known elements have not been shown or described in detail to avoid unnecessarily obscuring the invention. Accordingly, the specification and drawings are to be regarded in an illustrative, rather than a restrictive sense.

[0024] In one embodiment, an "inactivating alteration", also known as a loss-of- function (LOF) alteration or loss-of-fu notion mutation, in a gene is a nonsense mutation (i.e. a mutation leading to insertion of a premature stop codon), a frameshift insertion, or a frameshift deletion. In one embodiment, an "inactivating alteration" further includes deep copy number loss or a methylation status of a gene and / or its promoter which suppresses expression of that gene.

[0025] In some embodiments as described further herein, the inactivating alteration in ARID1A, RAD50 or KMT2D is present at any location within the gene body. For example, with reference to FIG. 1A, IB and 1C, the inventors have confirmed through the examples described herein that inactivating alterations can be found at a plurality of different locations throughout each of ARID1A, RAD50 and KMT2D.

[0026] As used herein, the term WT+ refers to gene sequences which are wild type or which include missense mutations (i.e. gene sequences which do not incorporate inactivating alterations). A missense mutation is a type of point mutation where a change in the DNA sequence results in a different amino acid being incorporated into a protein during translation.

[0027] As used herein, a reference to a patient or subject being sensitive to therapy with a particular agent, e.g. inhibitors of WRN or immune checkpoint inhibitors, means that a favourable or beneficial clinical result is likely to occur due to administration of the agent. E.g. a cancer of the patient is likely to respond to therapy with that agent, i.e. that administration of that agent to the patient is likely to have a therapeutic or prophylactic effect on the patient.

[0028] The inventors have now determined that sensitivity of cancer patients with MSI-positive or dMMR cancers to treatment with WRN (Werner syndrome RecQ helicase) inhibitors can be predicted in advance by determining the mutational status of the tumour suppressor gene KMT2D. In some aspects, the sensitivity of cancer patients with MSI-positive cancers to treatment with WRN inhibitors can bemore precisely predicted in advance by determining the mutational status of ARID1A. In some aspects, the sensitivity of cancer patients with MSI-positive cancers to treatment with WRN inhibitors can be more precisely predicted in advance by determining the mutational status of RAD50. In some aspects, the sensitivity of cancer patients with MSI-positive cancers to treatment with WRN inhibitors can be more precisely predicted in advance by determining the mutational status of one, two or all of KMT2D, ARID1A and RAD50.

[0029] In one embodiment, novel methods for the prediction of cancer patient (with microsatellite instability positive (MSI) and / or mismatch repair deficiency (dMMR) cancer) sensitivity to inhibitors of the WRN (Werner syndrome RecQ helicase) protein by determining the mutational status of the tumour suppressor gene KMT2D are provided. In one embodiment, cancer patients with MSI and / or dMMR cancers harbouring loss-of-function mutations (LOF) in the tumour suppressor gene KMT2D are predicted to be therapeutically sensitive to treatment with inhibitors of the WRN protein. In one embodiment, cancer patients with MSI and / or dMMR cancers harbouring loss-of-function mutations (LOF) in two of KMT2D, ARID1A and RAD50 are predicted to be therapeutically sensitive to treatment with inhibitors of the WRN protein. In one embodiment, cancer patients with MSI and / or dMMR cancers harbouring loss-of-function mutations (LOF) in ARID1A are predicted to be therapeutically sensitive to treatment with inhibitors of the WRN protein. In one embodiment, cancer patients with MSI and / or dMMR having a genotype such that the tumour suppressor genes ARID1A and RAD50 are both wild-type or have a missense mutation (normal protein function), or such that ARID1A and KMT2D are both wild type or have a missense mutation (normal protein function), or ARI DI A, RAD50 and KMT2D are all wild-type or have only missense mutations, are predicted to be therapeutically insensitive to treatment with inhibitors of the WRN protein.

[0030] In one embodiment, new processes for the diagnostic identification / stratification of MSI-positive and / or dMMR cancer patients who will respond to WRN inhibitor treatment is provided. In one embodiment, whole-genome or exome sequencing of tumour cells can be used to detect / screen for / identify KMT2D, RAD50 and ARID1A LOF mutations in MSI-positive and / or dMMR cancer patients. In one embodiment, novel methods to identify ICI (immune checkpoint inhibitor) resistantcancer patients who may be amenable / sensitive to treatment with WRN inhibitors through determination of KMT2D, RAD50 and ARID1A mutational status (e.g. positive for LOF mutations in KMT2D, RAD50, and ARID1A according to the selection criteria described herein) is provided. In one embodiment, detection of KMT2D, RAD50 and ARID1A mutational status comprises a novel method for identifying cancer patients who are amenable to combination therapy with ICI and WRN inhibitors. For example, using FDA approved methods for selecting patients for ICI treatment, solid cancers with high TMB (greater than or equal to 10 mutations / megabase

[0016] ) and who further meet the selection criteria disclosed in this specification for sensitivity to WRN inhibitors, patients amenable to combination therapy with ICI and WRN inhibitors may be identified.

[0031] In particular, the inventors have determined that sensitivity of a cancer to inhibition of WRN can be predicted by examining a gene signature based on the presence of loss-of-function alterations in KMT2D, RAD50 and ARID1A according to the matrix set forth in Table 1, and that evaluation of this three-gene signature is both necessary and sufficient to predict sensitivity of the cancer to inhibitors of WRN, at least in the genetic contexts examined. In Table 1, LOF denotes a loss-of- function mutation (i.e. an inactivating alteration in the gene), while WT+ denotes a wild type or missense mutation that does not abrogate protein function.Table 1. Gene Signature to Evaluate Sensitivity to Inhibitors of WRN in MSI and / or dMMR Cancers.

[0032] For example, with reference to FIG. 2, an example embodiment of a method 100 for determining whether a cancer such as a solid tumour of a patient is likely to be sensitive to therapy using inhibitors of WRN is illustrated. At 102, cells of the solid tumour are evaluated to determine if the cancer is microsatellite stable (MSS) or microsatellite instable (MSI) and / or if the cancer has mismatch repair deficiency (dMMR). At 104, cells of the solid tumour are evaluated to determine if somatic loss-of-fu notion alterations are present in two or all of KMT2D, ARID1A and RAD50. At 106, an evaluation of the MSI and / or dMMR status and mutational status of two or all of KMT2D, ARID1A and RAD50 is made. If it is determined at 108 that the cancer is MSI or dMMR and that there are loss-of-function alterations in two or all of KMT2D, ARID1A and RAD50, or if it is determined at 110 that the cancer is MSI or dMMR and there is a loss-of-function alteration in ARID1A (even if KMT2D and RAD50 are wild type or have only missense mutations), then at 112 it is concluded that the cancer is likely to be sensitive to inhibitors of WRN. In such a case, a therapeutic agent that is an inhibitor of WRN now known or developed in future, such as VVD-133214 (also referred to as RO7589831), HRO761, GSK4418959 (also referred to as IDE275 or GSK959), or NDI-219216 (also referred to as NTX-452), which are currently under clinical investigation, can be selected for administration to the patient and if desired administered to the patient in any suitable manner at 114. Other inhibitors of WRN are also known, such as NSC617145, and could be used if desired in other embodiments, although these compounds have not been selected for further clinical development.

[0033] The structures of HRO761, VVD-133214, and GSK4418959 are as follows:HRO761 VVD-133214 GSK4418959

[0034] If it is determined at 116 that the cancer is MSS, or if it is determined at 116 that the cancer is MSI or dMMR and both ARID1A and RAD50 are wild type or have only missense mutations (i.e. are WT+), or if it is determined at 116 that the cancer is MSI or dMMR. and both ARID1A and KMT2D are wild type or have only missense mutations (i.e. are WT+), then at 118 it is concluded that the cancer is likely to be insensitive to inhibitors of WRN. In such a case, alternative therapeutic agents other than inhibitors of WRN, for example immune checkpoint inhibitors (ICI), may be selected for administration to the patient and may be administered to the patient in any suitable manner at 120.

[0035] In some embodiments, the evaluation of whether the cancer has microsatellite instability (MSI) and / or mismatch repair deficiency (dMMR) at 102 is made in any suitable manner, for example by using polymerase chain reaction (PCR) to evaluate the length of microsatellite repeats in tumour tissue versus normal tissue, or by using immunohistochemistry (IHC) to assess mismatch repair (MMR) protein expression levels, as is known in the art.

[0036] In some embodiments, the evaluation of KMT2D, ARID1A or RAD50 for the presence of somatic loss-of-function mutations at 104 is carried out in any suitable manner now known or later developed. For example, a sample of the cancer such as cells obtained via a biopsy of a solid tumour may be subjected to genetic sequencing and / or DNA methylation sequencing in any suitable manner such as short-read or long-read whole genome sequencing; short-read or long-read DNA methylation sequencing; short-read targeted genome sequencing of one, two or all of KMT2D, ARID1A and RAD50; nanopore adaptive sampling-based long-read sequencing of one, two or all of KMT2D, ARID1A and RAD50; tiling PCR followed by Sanger sequencing of one, two or all of KMT2D, ARID1A and RAD50, or the like. In one example embodiment, nanopore adaptive sampling-based long read sequencing of one, two or all of KMT2D, ARID1A and RAD50 is carried out to evaluate both DNA sequence and DNA methylation sequence using a single sequencing technique.

[0037] In some embodiments, to evaluate whether a loss-of-function mutation in one or more of KMT2D, ARID1A and RAD50 is a somatic mutation, the sequence determined from the cancer cells is compared with the corresponding sequenceobtained from a sample of normal tissue such as a blood sample. In some embodiments, the normal cells may be subjected to sequencing in any suitable manner, including by genetic sequencing and / or DNA methylation sequencing in any suitable manner such as short-read or long-read whole genome sequencing; shortread or long-read DNA methylation sequencing; short-read targeted genome sequencing of one, two or all of KMT2D, ARID1A and RAD50; nanopore adaptive sampling-based long-read sequencing of one, two or all of KMT2D, ARID1A and RAD50; tiling PCR followed by Sanger sequencing of one, two or all of KMT2D, ARID1A and RAD50, or the like.

[0038] In some embodiments, method 100 is carried out to select an inhibitor of WRN as a therapy for a patient if it is determined that a cancer of the patient is likely to be sensitive to therapy with inhibitors of WRN. In some embodiments, method 100 is carried out to identify a patient as a candidate for therapy using an inhibitor of WRN if it is determined that a cancer of the patient is likely to be sensitive to therapy with inhibitors of WRN.

[0039] In some embodiments, a kit for sequencing two or more of KMT2D, ARID1A and RAD50 is provided. In some embodiments, the kit can be used to sequence only KMT2D, ARID1A and RAD50, and cannot be used to sequence any other genes. In some embodiments, the kit contains reagents for conducting nanopore adaptive sampling-based long-read sequencing of two or more of KMT2D, ARID1A and RAD50. In some such embodiments, the kit contains only reagents for conducting nanopore adaptive sampling-based long-read sequencing of two or more of KMT2D, ARID1A and RAD50, and cannot be used to sequence any other genes. In some embodiments, the kit contains primers specific for two or all of KMT2D, ARID1A and RAD50.

[0040] In some embodiments, the kit contains instructions for using the reagents to sequence two or all of KMT2D, RAD50, and ARID1A in a sample of a cancer and in a sample of normal tissue from the patient to evaluate the presence of somatic inactivating alterations in two or all of KMT2D, RAD50, and ARID1A. In some embodiments, the instructions direct a user to determine if somatic inactivating alterations are present in two or all of KMT2D, RAD50 and ARID1A in the cells of the cancer, and, if it is determined that somatic inactivating alterationsare present in two or all of KMT2D, RAD50 and ARID1A in the cells of the cancer, or if it is determined that somatic inactivating alterations are present in ARID1A in the cells of the cancer, directing the user to conclude that the cancer patient is likely to be sensitive to treatment with inhibitors of WRN. In some embodiments, the kit further comprises instructions to conclude that the cancer patient is unlikely to be sensitive to treatment with inhibitors of WRN if it is determined that ARID1A is wild type or has only a missense mutation in the cells of the cancer, and that one or both of KMT2D and RAD50 are wild type or have only a missense mutation in the cells of the cancer.

[0041] In some embodiments, the MSI and / or dMMR cancer is a solid tumour. Examples of solid tumours in which MSI or dMMR regularly occurs include colorectal cancer, endometrial cancer, ovarian cancer, gastric cancer, or esophageal cancer. In some embodiments, the patient with the MSI and / or dMMR cancer is a mammalian subject including a human. In some embodiments, the patient is resistant to treatment with immune checkpoint inhibitors (ICI).

[0042] In some embodiments, the inhibitors of WRN are administered together with an immune checkpoint inhibitor as a combination therapy. For example, there are observations based on an I CI- refractory patient-derived mouse model that treatment with WRN inhibitors is still effective at reducing tumour volume

[0014] , that in another patient-derived mouse model WRN inhibitor GSK4418959 had combinatorial effects with ICIs and showed promise as a treatment option for MSI- H cancer patients in a clinical setting

[0017] , and that in human trials early phase I data shows that in a cohort with 89% of cases that had received ICIs as a prior treatment, administration of WRN inhibitors post-ICI treatment is well tolerated with a promising disease control rate. Thus, there may be options to use WRN inhibitors as monotherapy in appropriate cases, or in combination with existing therapies such as immune checkpoint inhibitors for enhanced clinical benefit.

[0043] Examples of immune checkpoint inhibitors that can be used in various embodiments include inhibitors of CTLA4, PD-1 or PD-L1. In some embodiments, the immune checkpoint inhibitor is pembrolizumab, ipilimumab, tremelimumab, nivolumab, atezolizumab, avelumab, durvalumab, cemiplimab, dostarlimab,relatlimab, or any other immune checkpoint inhibitor now known or developed in future.

[0044] With reference to FIG. 3, an example embodiment of a method 200 for selecting patients to receive combination therapy using immune checkpoint inhibitors and inhibitors of WRN is provided. At step 202, it is determined if the solid tumour of the patient is likely to be sensitive to therapy using immune checkpoint inhibitors using any method now known or developed in future, for example by using FDA approved methods for selecting patients for ICI treatment such as determining that the solid cancer has a high tumour mutation burden (TMB) greater than or equal to 10 mutations / megabase

[0016] . At step 204, it is determined if a solid tumour of a patient is likely to be sensitive to therapy using inhibitors of WRN, for example by carrying out method 100.

[0045] At step 206, an evaluation of the likely sensitivity of the solid cancer to both inhibitors of WRN and immune checkpoint inhibitors is made. If at 208 it has been determined that the solid cancer is both likely to be sensitive to inhibitors of WRN and likely to be sensitive to immune checkpoint inhibitors, then at 210 the patient can be selected to receive combination therapy using inhibitors of WRN and immune checkpoint inhibitors and both therapies can be administered to the patient. If at 212 it is determined that the solid cancer is likely to be sensitive to inhibitors of WRN and likely to be insensitive to immune checkpoint inhibitors, then at 214 the patient can be selected to receive only inhibitors of WRN and such therapy can be administered to the patient. If at 216 it is determined that the solid cancer is likely to be sensitive to immune checkpoint inhibitors and likely to be insensitive to inhibitors of WRN, then at 218 the patient can be selected to receive only immune checkpoint inhibitors and such therapy can be administered to the patient. If at 206 it is determined that the solid cancer is likely to be insensitive to both inhibitors of WRN and immune checkpoint inhibitors, then neither therapy should be administered to the patient and the patient can be considered for alternative therapies.

[0046] The therapeutic agents described herein can be administered to patients in any suitable manner now known or later developed for delivering therapeutic agents, including by being formulated into a suitable pharmaceutical compositionfor administration in any desired manner such as orally, intravenously, intramuscularly, intraperitoneally, or so on. The compound can be administered to the mammalian patient including a human using any suitable dosage regime or dosage regimen. The co-administration of any therapeutic agents described herein, e.g. inhibitors of WRN and immune checkpoint inhibitors, can be accomplished in any suitable manner, including administering the combination of active agents together, whether at the same time and / or formulated together into a single pharmaceutical composition, or at different times, e.g. through separate modes of administration or through sequential administration (e.g. via the administration of a course of one of the active agents followed in time by the administration of a course of the second one of the active agents).

[0047] Further aspects of the invention will become apparent from consideration of the ensuing description of preferred embodiments of the invention. A person skilled in the art will realise that other embodiments of the invention are possible and that the details of the invention can be modified in a number of respects, all without departing from the inventive concept. Thus, the following drawings, descriptions and examples are to be regarded as illustrative in nature and not restrictive.Examples

[0048] Certain embodiments are further described with reference to the following examples, which are intended to be illustrative and not limiting in nature.Example 1.0 - Methods

[0049] Identifying KMT2D mutations across cancer types. Mutation annotation format (MAF) file from TCGA pan-cancer dataset, representing 10,217 tumour samples across 33 cancer types

[0018] , were downloaded on June 15, 2022 (mc3.vO.2.8.PUBLIC.maf). Additionally, the inventors downloaded mutation annotations from supplementary files from B-cell non-Hodgkin lymphoma (B-NHL) datasets (117 samples) [19-21], medulloblastomas (MED) dataset (53 samples)

[0022] , and a small cell lung cancers (SCLC) dataset (110 samples)

[0023] , and annotated LOF mutations (nonsense mutations, frameshift insertions and deletions,and nonstop mutations). The inventors used TCGA study abbreviations, except for B-NHL, which includes diffuse large B-cell lymphoma, follicular lymphoma, mantle cell lymphoma, and nodal marginal zone lymphomas, and COAD / READ, which includes colon and rectal adenocarcinomas. The following are the cancer type abbreviations used: adrenocortical carcinoma (ACC), bladder urothelial carcinoma (BLCA), B-cell non-Hodkin lymphoma (B-NHL), breast invasive carcinoma (BRCA), breast fibroepithelial (BRFE), cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC), cholangiocarcinoma (CHOL), colon and rectal adenocarcinoma (COAD / READ), esophageal carcinoma (ESCA), glioblastoma multiforme (GBM), head and neck squamous cell carcinoma (HNSC), kidney chromophobe (KICH), kidney renal clear cell carcinoma (KIRC), kidney renal papillary cell carcinoma (KIRP), acute myeloid leukemia (LAML), brain lower grade glioma (LGG), liver hepatocellular carcinoma (LIHC), lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LUSC), medulloblastoma (MED), mesothelioma (MESO), ovarian serous cystadenocarcinoma (OV), pancreatic adenocarcinoma (PAAD), pheochromocytoma and paraganglioma (PCPG), prostate adenocarcinoma (PRAD), sarcoma (SARC), small cell lung cancer (SCLC), skin cutaneous melanoma (SKCM), stomach adenocarcinoma (STAD), testicular germ cell tumours (TGCT), thyroid carcinoma (THCA), thymoma (THYM), uterine corpus endometrial carcinoma (UCEC), uterine carcinosarcoma (UCS), and uveal melanoma (UVM).

[0050] In silico genetic network mapping.

[0051] Identifying DepMap cell lines. Using GRETTA (0.99.2)

[0024] the inventors queried KMT2D and WRN mutations from a total of 739 cancer cell lines in DepMap. These lines have WGS or whole-exome sequencing (WES) and RNA sequencing data, in addition to, genome-wide CRISPR-Cas9 KO screens [15,25]. Whole proteome quantification and global histone quantification were only available for a subset of 375 and 897 cancer cell lines, respectively [15,26]. The microsatellite status of DepMap cell lines was identified from supplementary files provided by Ghandi et al.

[0015] . As performed previously [24,27], the inventors used GRETTA to query the genomic data (MAF and copy number data files) to identify KMT2DL0Fand WRNL0Fmutant cancer cell lines (homozygous, trans-heterozygous, and heterozygous LOF mutants) and KMT2DWTand WRNWTcontrol cell lines. Onlytrans-heterozygous lines (lines with more than one LOF mutation or a combination of LOF mutation and copy number loss) and heterozygous lines were available for the KMT2D pan-cancer group. Since trans-heterozygous LOF mutant groups are more likely to be KMT2D deficient than heterozygous lines

[0028] , the inventors only used these lines in the KMT2 DL0Fce\\ line group (16 lines). All LOF mutants (trans- heterozygous and heterozygous KMT2DL0Flines) lines were considered in the KMT2D context-specific screens. For the pan-cancer WRN genetic screen, only heterozygous LOF mutants (176 lines) were available. To match DepMap cell lines to the appropriate cancer types, the inventors used the "disease" and "disease subtype" columns provided in the cell line annotation file (samplejnfo.csv

[0015] ) according to Supplemental Table S9 of Takemon et al.

[0029] . The normalised KMT2D TPM gene expression and protein expression values were extracted using GRETTA.

[0052] Differential lethality analysis. As described previously [24,27], the inventors used GRETTA to perform pairwise Mann-Whitney U tests comparing lethality probabilities for all 18,333 genes targeted in DepMap's CRISPR-Cas9 KO screen between LOF mutant lines and WT lines to obtain p-values. P-values were adjusted for multiple testing using a permutation approach by randomly resampling 10,000 of the lethality probability scores. The inventors used a threshold of adjusted p-value < 0.05 to identify candidate GIs of KMT2D in each screen, which were then prioritised (described in detail below). For visualisation, the inventors calculated the genetic interaction score:Genetic interaction score median LOF group lethality probability— log10[adjusted p value] x log2[ ] median WT group lethality probability

[0053] Prioritising candidate GIs. Candidate KMT2D interactors were prioritised according to their GI significance and their drug tractability, similar to the approaches used by Behan et al. [4]. Firstly, candidates were grouped into three tiers (tiers I, II, and III) based on their GI significance. Tier I candidates were the most significant group with adjusted Mann Whitney-U test p-values < 0.01, an absolute Iog2 fold change > 2 between median lethality probabilities (LOF group / WT group), and a minimum lethality probability of 0.5 in at least one group (WT orLOF). The following GI tiers II and III used progressively less stringent thresholds of Mann Whitney-U test p-values < 0.05, an absolute . Iog2fold change > 2 between median lethality probabilities (LOF group / WT group), and a minimum lethality probability of 0.5 (tier II) or none (tier III) in at least one group (WT or LOF).

[0054] Next, candidates were grouped based on drug tractability (group I, II, and III). The inventors used the Open Target Platform

[0030] , a database that contains a collection of tractability assessments based on sources, such as UniProt, HPA, PDBe, DrugEBIlity, ChEMBL, Pfam, InterPro, Complex Portal, DrugBank, Gene Ontology, and BioModels (accessed on February 15th, 2022), which is a strategy developed by Brown et al.

[0031] and Schneider et al.

[0032] ). Using this Open Target, the inventors assessed whether the candidate GI encodes a protein for which there is already a known drug or, based on its protein structure, the likelihood of a drug being developed. Drug tractability group I contains GI candidates that are most tractable, and encodes proteins that are currently targetable using approved or phase I, II and II small molecule inhibitors or antibodies (buckets 1-3). Group II contains targets that have high quality protein structure, ligand, and pocket annotations and cellular location annotation (buckets 4-6). Group III contains targets with proteins that have medium to low quality pocket annotation, cellular location annotation, or are known members of the druggable protein family (buckets 7-9).

[0055] Finally, GI candidates were prioritised by combining the GI significance and drug tractability assessments (class A, B, C, and D). With progressively less likelihood of targetable GIs, class A contained GIs with the highest significance with targetable proteins, and class D containing GI with the least significance with no evidence of tractability.

[0056] TCGA cohort analyses. Identifying KMT2D LOF mutations. The inventors used cBioPortal [33,34] to download clinical (including MSIsensor, MANTIS, and TMB scores) and genomic data from TCGA Pan-Cancer Atlas Studies (33 cancer types). A TCGA case was considered to have MSI if the MSIsensor and MANTIS scores were >= 10 and >= 0.4, respectively, as previously determined [35,36]. The inventors annotated cases with KMT2D LOF alterations as havingnonsense mutation, frameshift insertions, frameshift deletions, or deep copy number loss. Cases with missense or silent mutations were excluded.

[0057] Mutational signature analysis. A matrix of 96 trinucleotide substitutions were downloaded from Alexandrov et al.

[0037] . The matrix was analysed in Python (v3.9) using the SigProfilerAssignment

[0038] library to score single base pair substitution (SBS) signature activity scores with COSMIC (v.3.3), exome normalisation, and default settings.

[0058] Neoantigen scores and RNA-seq data. Neoantigen scores were downloaded from Thorsson et al.

[0039] and RNA-seq count data were downloaded from UCSC Xena Hub

[0040] . Additional immune marker analyses are described below.

[0059] PPG cohort analyses. Ethics approval, consent to participate, and enrollment criteria. This work, involving POG program patients, was approved by the University of British Columbia-BC Cancer Research Ethics Board (H12-00137, H14-00681), and the POG program is registered under clinical trial number NCT02155621. All patients in this study gave informed written consent and were enrolled into POG as described previously, between July 2012 and May 2022 [41- 43]. The main objectives for our POG cohort analyses were to retrospectively validate TCGA cohort findings using whole-genome and whole-transcriptome analysis.

[0060] Tissue collection. Tissue collection and sequencing were performed as in Pleasance et al.

[0043] for 820 patient samples. Tumour specimens were collected using needle core biopsies or tissue resection, snap frozen generally in optimal cutting temperature compound, reviewed by pathologists, and DNA and RNA libraries constructed and sequenced. Genome data were used to detect somatic alterations, including SNVs, copy number variants, loss of heterozygosity, and structural variants (such as gene fusions).

[0061] Whole-genome and transcriptome libraries and sequencing. PCR- free genomic DNA libraries and poly-A selected RNA libraries were constructed and sequenced, and reads were aligned to the GRCh38 genome reference. Tumour genomes were sequenced to a target depth of 80X coverage and normal peripheral blood samples to a target depth of 40X coverage on Illumina HiSeq instruments.Details regarding library preparation, sequencing, and quality control can be found in Pleasance et al. (2022)

[0043] . In cases with RIMA integrity numbers < 7.0, ribosomal RNA depletion RNA sequencing was employed as described in Pleasance et al.

[0043] . Recent samples (since POG1046) were processed using an updated ribosomal RNA depletion RNA sequencing protocol, which is described below.

[0062] To remove cytoplasmic rRNA and mitochondrial ribosomal RNA (rRNA) species from total RNA, NEBNext rRNA Depletion Kit for Human / Mouse / Rat was used (NEB, E6310X). Enzymatic reactions were set-up in a 96-well plate (Thermo Fisher Scientific) on a Microlab NIMBUS liquid handler (Hamilton Robotics, USA). 150 ng of DNase treated total RNA in 6 pL was hybridized to rRNA probes in a 8.5 pL reaction. Heat-sealed plates were incubated at 95°C for 2 minutes followed by incremental reduction in temperature by 0.1°C per second to 22°C (730 cycles). The rRNA in DNA hybrids were digested using RNase H in a 11 pL reaction incubated in a thermocycler at 50°C for 30 minutes. To remove excess rRNA probes (DNA) and residual genomic DNA contamination, DNase I was added in a total reaction volume of 26 pL and incubated at 37°C for 30 minutes. RNA was purified using RNA MagClean DX beads (Aline Biosciences, USA) with 15 minutes of binding time, 7 minutes clearing on a magnet followed by two 70% ethanol washes, 5 minutes to air dry the RNA pellet and elution in 18pL DEPC water. The plate containing RNA was stored at -80°C prior to cDNA synthesis.

[0063] First-strand cDNA was synthesized from the purified RNA (minus rRNA) using the Maxima H Minus First Strand cDNA Synthesis kit (Thermo-Fisher, USA) and random hexamer primers at a concentration of 8 ng / pL along with a final concentration of 0.04 pg / pL Actinomycin D, followed by PCR Clean DX bead purification on a Microlab NIMBUS robot (Hamilton Robotics, USA). The second strand cDNA was synthesized following the NEBNext Ultra Directional Second Strand cDNA Synthesis protocol (NEB) that incorporates dUTP in the dNTP mix, allowing the second strand to be digested using USERTM enzyme (NEB) in the post-adapter ligation PCR and thus achieving strand specificity.

[0064] cDNA was fragmented by Covaris LE220 sonication for 130 seconds (2 x 65seconds) at a "Duty cycle" of 30%, 450 Peak Incident Power (W) and 200 Cycles per Burst in a 96-well microTUBE Plate (P / N : 520078) to achieve 200-250 bpaverage fragment lengths. The paired-end sequencing library was prepared following Canada's Micheal Smith Genome Sciences Centre strand-specific, platebased library construction protocol on a Microlab NIMBUS robot (Hamilton Robotics, USA). Briefly, the sheared cDNA was subject to end-repair and phosphorylation in a single reaction using an enzyme premix (NEB) containing T4 DNA polymerase, Klenow DNA Polymerase and T4 polynucleotide kinase, incubated at 20oC for 30 minutes. Repaired cDNA was purified in 96-well format using PCR Clean DX beads (Aline Biosciences, USA), and 3' A-tailed (adenylation) using Klenow fragment (3' to 5' exo minus) and incubation at 37°C for 30 minutes prior to enzyme heat inactivation. Ligation using TruSeq adapters at 20°C for 15 minutes. The adapter- ligated products were purified using PCR Clean DX beads followed by indexed PCR using NEBNext Ultra II Q5 Master Mix (NEB) with USER enzyme (1 U / pL, NEB) and dual indexed primer set. PCR parameters: 37°C for 15 minutes, 98°C for 1 minute followed by 13 cycles of 98°C 15 seconds, 65°C 30 seconds and 72°C 30 seconds, and then 72°C 5 minutes. The PCR products were purified and size-selected using a 1 : 1 PCR Clean DX beads-to-sample ratio (twice), and the eluted DNA quality was assessed with Caliper LabChip GX for DNA samples using the High Sensitivity Assay (PerkinElmer, Inc. USA) and quantified using a Quant-iT dsDNA High Sensitivity Assay Kit on a Qubit fluorometer (Invitrogen) prior to library pooling and size- corrected final molar concentration calculation for Illumina sequencing with paired- end 150 base reads.

[0065] Variant calling and KMT2D mutation selection. Somatic point mutations and small insertions and deletions are identified using Strelka 2.6.2

[0044] and Mutect2 from GATK 4.0.10.0

[0045] . Variants from these tools are intersected using RTGTools

[0046] to generate the calls as previously described

[0047] . The inventors defined POG cases with KMT2D LOF alterations as those having somatic nonsense mutations, frameshift insertions, frameshift deletions, or deep copy number loss indicating likely homozygous deletions. Cases with missense or silent mutations were not considered and were excluded from the analysis entirely. POG cases with WT KMT2D alleles were defined as those with no somatic alteration and neutral copy number. Germline alterations were not considered in this study.

[0066] MSK-IMPACT cohort analyses. Targeted panel gene sequencing, survival, and TMB data from the MSK-IMPACT cohort were downloaded from cBioPortal (accessed March 15th, 2022;

[0048] ). The MSK-IMPACT assay identifies somatic exonic mutations in 468 cancer-associated genes using both tumour- derived and matched germline normal DNA

[0048] . The inventors identified cases with KMT2DL0Falterations, when a case harboured either a frameshifting INDEL or a nonsense deleterious mutation. We inferred a case was MSI when a LOF alteration(s) (ie. frameshift INDELs and nonsense mutations) in one of the MMR genes (MLH1, MLH3, PMS2, MSH2, MSH3, and MSH6') was observed. TMB was calculated as per Samstein et al.

[0048] as the total number of somatic mutations normalised to the exonic coverage of the MSK-IMPACT panel in megabases.

[0067] Survival analyses. Survival analyses were performed using R package ggsurvfit (vO.2.1). Differences in nonparametric survival functions were assessed across groups using log-rank tests with the ggsurvfit package. Groups with less than 10 cases or without enough observations to determine 50% survival probability were not analysed.

[0068] Additional immune marker analyses. Immune checkpoint response scores were calculated using the PredictIO webportal, which uses a gene signaturebased method to predict responders to ICI treatment

[0049] .

[0069] Analysis of WRN KO dependency and WRN inhibitor sensitivity validated cell lines. Cancer cell lines that were validated to be WRN KO- (in)dependent by Behan et al. [4], Chan et al. [6], Kategaya et al. [5], Lieb et al.[7] and Picco et al. [9] and validated to be (in)sensitive to WRN inhibitors (HRO761 and VVD-133214) by Ferretti et al.

[0013] and Baltgalvis et al.

[0014] , were identified and cross referenced to DepMap cell lines with WGS / WES annotation (20Q1 and 23Q4). The inventors note that the WRN inhibitor insensitive cell line HCT8, tested by Baltgalvis et al.

[0014] , is identical to HRT18. For each cell line, the inventors annotated with the reported WRN KO or WRN inhibitor dependence and microsatellite status. Using DepMap WGS / WES annotations, we identified and annotated LOF alterations, including copy number loss, nonsense mutations, frameshift insertions and deletions, and nonstop mutations, affecting KMT2D and MMR genes (namely EX01, MBD4, MLH1, MSH2, MSH3, MSH5, MSH6, PMS1, PSM2,and RFC1). Fisher's exact tests were performed to determine co-occurrence between KMT2D / MMR gene LOF mutations and WRN dependence, and between KMT2D and MMR gene LOF alterations.

[0070] Microsatellite repeat expansion analysis. The inventors identifiedMSI DepMap cancer cell lines with WGS data that carried a WT KMT2D allele or LOF alterations (WT line: SKMEL2 and LOF lines: CCK81, LOVO, SW48, LS180, MFE319, and RK0) as well as two MSS lines (NCIH747 and COLO201; SRA project ID: PRJNA523380; SRA run IDs in Supplemental Table S11A). The inventors downloaded the raw .fastq data using SRA-Toolkit

[0111] and aligned to GRCh38 using bwa (vO.7.17)

[0050] . The inventors used default ExpansionHunter Denovo (v0.9.0)

[0051] settings to detect AT-motifs and filtered for regions that contained at least two in-repeat reads (i.e. one read pair maps within and the other outside of a repeat region).

[0071] Statistical analyses. All statistical analyses were conducted using R statistical software (v4.2.2)

[0052] unless otherwise stated.

[0072] Availability of data and materials. The scripts for reproducing the analysis, figures, and tables are available at GitHub; the GRETTA R package and DepMap data version 20Q1 used in this study are publicly available on the GRETTA GitHub repository; GRETTA has been archived with a citable DOI on Zenodo and a Singularity container has been made available on Sylabs

[0029] ). ChlP-MS data have been deposited with the ProteomeXchange Consortium via the Proteomics Identification (PRIDE

[0053] ) database (accession PXD048272). Genomic and transcriptom ic sequence datasets, including metadata with library construction and sequencing approaches have been deposited at the European Genome-phenome Archive

[0054] with accession numbers as listed in Supplemental Table S8 of Takemon et al.

[0029] . All other relevant data supporting the key findings of this study are available in Supplemental Tables of Takemon et al.

[0029] .Example 2.0 - Results

[0073] In silico genetic screens reveal pan-cancer and context-specific KMT2D GIs associated with mitotic processes, DNA repair, metabolism, and immune response. To map KMT2D's genetic network, the inventors used GRETTA

[0024] , a software approach that they developed to leverage the public cancer dependency map (DepMap) data platform to map genetic networks (gene essentiality networks and genetic interaction networks). GR.ETTA uses genomic and transcriptomic data and CRISPR perturbation screening data generated by DepMap from cancer cell lines to map genetic networks in siiico

[0024] . Similar approaches have been used to map genetic networks of cancer-associated genes [4,55,56]. The inventors used GR.ETTA

[0024] to perform pan-cancer and cancer type-specific in siiico genome-wide KO screens, with the aim of identifying SL and AL interactors of KMT2D. To determine the cancer types that were most relevant to KMT2DL0Fcancers, the inventors queried the datasets from The Cancer Genome Atlas (TCGA; 10,217 tumour samples across 33 cancer types

[0018] ) and included additional datasets from cancer types that are known to frequently harbour KMT2D alterations (see Methods for details). The inventors selected 14 cancer types in which at least 10% of samples harboured KMT2D mutations and where 25% of the total KMT2D mutations were LOF (FIG. 4A; Methods). Next, the inventors selected cancer types with at least ten DepMap cancer cell lines, using the criterion of Behan et al. [4]. This resulted in the identification of nine cancer types (namely bladder urothelial carcinomas [BLCA], B-cell non-Hodgkin lymphoma [B-NHL], colorectal adenocarcinomas [COAD / READ], esophageal carcinomas [ESCA], head and neck squamous cell carcinomas [HNSC], lung squamous cell carcinomas [LUSC], small cell lung cancers [SCLC], stomach adenocarcinomas [STAD], and uterine corpus endometrial carcinoma [UCEC]), in which cancer type-specific in siiico GI screen was performed (FIG. 4A; Methods). In addition to these nine, the inventors also performed a pan-cancer screen that included all DepMap cell lines. Given that the pan-cancer screen would include cell lines consisting of various cancer types, the inventors expected this screen to contain substantial noise. Also, the inventors expected positive signals (i.e. GI candidates) arising from a pan-cancer screen would represent strong signals, and GI candidates appearing in both pan-cancer and cancer type-specific screens to be high-confidence candidates. Therefore, the inventors rationalised the inclusion of the pan-cancer dataset, which resulted in the analysis of ten datasets.

[0074] For the pan-cancer dataset and each of the nine cancer type-specific datasets, the inventors used GRETTA [24,27] to identify KMT2D-Q?and KMT2D'mDepMap cancer cell lines (FIG. 1A; Methods). Next, the inventors compared the K.MT2D mRNA and protein expression between the KMT2D'mand KMT2DL0Fcell lines within each dataset. KMT2D mRNA expression was significantly higher in KMT2DL0Flines than in KMT2DVJlines in the pan-cancer and UCEC datasets (Welch's t-test p- value < 0.05). On the other hand, KMT2D protein expression (quantified by DepMap using mass-spectrometry) in KMT2DL0Fcell lines was significantly lower in the pan-cancer dataset and trended lower than in KMT2D'mcell lines across all nine cancer type-specific datasets for which data were available, including COAD / READ, LUSC, SCLC, STAD, and UCEC. The inventors' data showed that KMT2DL0Falterations were associated with lower KMT2D protein expression so, from this, it was inferred that KMT2D activities are reduced or lost in most cancer types, as expected. KMT2D mRNA levels, on the other hand, were elevated or stable in some KMT2DL0Fcell lines, although the inventors did not attempt to re-construct the isoforms being expressed or otherwise assess the extent to which they might encode function, judging this to be beyond the scope of this study. The discordance in protein and mRNA expression is well documented for many genes [57-59] but has not yet been documented for KMT2D. The inventors also compared the relative global levels of active histone marks captured by DepMap, namely H3K4mel, H3K4me2, and H3K27ac, which are known to be reduced in KMT2D-deficient cells [60-64]. The inventors found a significant reduction of global H3K4mel and H3K4me2 levels in KMT2D^ cell lines of the pan-cancer and COAD / READ-specific datasets compared to their respective KMT2D'mcell lines (Welch's t-test p-value < 0.05). B-NHL KMT2DL0Fcell lines also had significantly reduced levels of H3K4me2 compared to KMT2D'mcell lines (Welch's t-test p-value < 0.05), and both H3K4mel and H3K4me2 levels trended lower in KMT2DL0FBLCA, SCLC, and UCEC cell lines, consistent with a reduction in KMT2D activity. Interestingly, the combined histone mark levels H3K27aclK36meO were significantly higher in B-NHL KMT2DL0Fcell lines compared to KMT2DWTcell lines (Welch's t-test p-value < 0.05), which suggests that KMT2C, EP300, or CREBBP activity may be compensating to maintain homeostasis of transcriptional activity disrupted by KMT2D loss

[0065] . Altogether,these results are consistent with the notion that KMT2D function is reduced in the selected KMT2DL0Fcell lines and that such cell lines are suitable for in silico genetic screening analyses.

[0075] Finally, using GRETTA

[0024] , the inventors performed differential lethality analyses to predict KMT2D GIs. In this analysis, a gene KO that led to significantly higher lethality probability in the KMT2DL0Fcell lines compared to KMT2Dm' cell lines indicated a potential SL interactor, whereas a KO that led to significantly lower lethality probability in KMT2DL0Fce\\ lines indicated a potential AL interactor. Using this method, the inventors predicted 4,396 GIs with 3,692 unique GIs across all screens (Methods). In context-specific screens, most predicted KMT2D GIs (~80%) were unique to the cancer type in which they were screened, which is consistent with the view that GIs are highly context-dependent

[0066] . As expected, 46% of GIs predicted in the pan-cancer screen were also found in context-sensitive screens, given that cell lines in the pan-cancer screens are made up of several cancer types. The largest overlap was seen between the pan-cancer and COAD / READ screens, which is likely due to the large proportion of KMT2DL0Fcell lines that are COAD / READ (5 / 16 cell lines).

[0076] To highlight pathways that are associated with candidate KMT2D GIs, the inventors annotated the SL and AL candidates with the biological processes in which they are involved. The SL candidates were associated with 67 biological processes (Methods). Of these, three (4%) were associated with mitotic processes and homologous recombination, five (7%) were associated with metabolic processes, and ten (15%) were associated with T-cell cytotoxicity and immune cell response. The AL candidates were associated with 301 biological processes. Five (~2%) were associated with mitotic processes, 26 (9%) with metabolism, and four (1%) with immune response. To determine whether these biological processes were more abundantly represented in SL or AL interactors, the inventors performed an equality of proportion analysis

[0067] to determine whether the proportion of mitotic, metabolic, and immune-related terms was more abundant among SL or AL interactors. The inventors found the proportion of immune-related terms was significantly higher in SL interactors than in AL interactors (equality of proportions p-value < 0.0001); however, no difference was found for mitotic or metabolic-associated terms. These results indicate that genes associated with immune response were more abundant among SL candidates. Therefore, KMT2DL0Fcells may be vulnerable to perturbations of immune response-associated pathways.

[0077] GI prediction scores and target tractability assignments prioritise therapeutically promising KMT2D GIs. Following the method introduced by Behan et al. [4] to determine high-quality KMT2D interactors and prioritise those that might be candidate drug targets, the inventors combined two classification systems, one based on GI prediction statistics and the other based on drug tractability (i.e. the likelihood of identifying a drug targeting the protein encoded by the genetic interactor; Methods). The inventors classified GI candidates into three tiers based on statistical significance (GI tiers I, II, and III), where GI tier I candidates represented the most statistically significant group. The second classification was based on drug tractability (drug groups I, II, and III), with group I containing candidates that are targets of existing drugs. Finally, the inventors combined the GI tiers and drug tractability groups to create four priority classes (A, B, C, and D), with priority class A representing the highest level of GI prediction significance and the highest evidence for drug tractability.

[0078] The inventors identified three priority class A candidates— namely the SL interactors MDM2, TUBA1B (COAD / READ), and NDUFB5 (STAD) and the AL interactor CDK6 (in the pan-cancer dataset)— that encode targets of approved anticancer drugs or that are the focus of drug development efforts (FIG. 4B

[0030] .

[0079] Priority class B candidates (targets without drugs in clinical development but with high-confidence evidence supporting target tractability, such as having high-quality ligand structures for pharmaceutical development

[0030] ) include seven genes, specifically the SL interactors WRN (pan-cancer and COAD / READ), MRPS17 (ESCA), and MTG1 (STAD) and the AL interactors ZNF217 (COAD / READ), RAB7A (ESCA), and ACTR5 (HNSC; priority class B; FIG. 4B). WRN encodes a DNA helicase and nuclease that functions in DNA replication, repair, and transcription [1-3] and has recently been identified as a SL vulnerability of MMR-deficient MSI cancers, including some colorectal, endometrial, gastric, and ovarian cancers [4-9]. Although WRN was categorised as a drug target group II candidate by the Open Target Platform

[0030] (Methods), there are currently several drug development programsunderway to target WRN [10,11] in MSI cancer; therefore, the inventors reclassified it as a drug target group I and priority class A gene.

[0080] Priority class C candidates consisted of interactors without drugs in clinical development and medium to low-confidence evidence supporting tractability. They included 19 genes.

[0081] Given that SL candidates NDUFB5 in the STAD screen; MDM2 and TUBA1B in the COAD / READ screen; and WRN in the COAD / READ and pan-cancer screens were identified as the most promising therapeutic targets (priority class A) for KMT2DL0Fcancers, the inventors sought to understand the relationship between KMT2D and these genes further.

[0082] KMT2DLOF MSI TCGA-COAD / READ cases have significantly elevated ICI response indicators. MSI is an approved tissue / site-agnostic biomarker for ICI treatments [69,70]. A recently completed analysis of KEYNOTE- 177 (in April of 2022), a randomised, open-label, phase-three study comparing MSI / MMR-deficient metastatic colorectal cancer cases treated with pembrolizumab versus chemotherapy, showed that although pembrolizumab-treated patients had fewer treatment-related adverse effects and a progression-free survival improvement was noted, no significant difference in overall survival between the two treatment groups was observed

[0071] . Therefore, there is a need to improve ICI treatment stratification. Given that the inventors identified several SL candidates with roles in T-cell cytotoxicity and the immune response, identified TUBAIB as a priority class A target in the COAD / READ screen, and showed that MSI can elicit specific vulnerabilities in KMT2DL0Fcancer cell lines; the inventors sought to understand the impact of KMT2DL0Fon immune markers in MSS and MSI COAD / READ cases and whether KMT2D mutational status might be relevant in the context of ICI treatment. To this end, the inventors profiled the clinical, mutational, and transcriptomic profiles of untreated COAD / READ MSI cases from TCGA. The inventors identified 43 MSI COAD / READ cases (27 KMT2D'mMSI and 16 KMT2DL0FMSI cases; Methods) and observed a trend towards lower overall survival probability in KMT2DL0FMSI cases compared to KMT2D'mMSI cases (log-rank test p-value = 0.10; FIG. 5A; Methods). A survival analysis was not performed for MSS cases due to the small sample size of KMT2DL0Fcases (417 KMT2D'mMSS and twoKMT2DL0FMSS cases). This supports previous observations showing that KMT2D deficiency leads to inferior survival outcomes compared to KMT2D-proficient patients with lymphoma, lung, and breast cancers [72-75].

[0083] Next, in the TCGA-COAD / READ cohort, the inventors characterised and compared the expression of TUBA1B, as an elevated expression has been associated with superior response to ICI therapy

[0104] , and markers of immune response, including TMB [76,77], neoantigen scores

[0078] , CD274 / PD-L1 and CTI_A4 expression

[0079] , cytolytic activity scores [80,81], M1 / M2 macrophage scores [82,83], interferon-gamma (IFNy) signalling scores

[0084] , and homologous recombination deficiency (HRD) scores

[0085] . Due to a small KMT2DL0FMSS sample size, statistical analyses of MSS cases were not performed for any of the immune markers. However, in MSI cases, the inventors found a trend towards higher TUBA1B expression in KMT2DL0FMSI cases compared to KMT2D'mMSI cases (Welch's t-test p-value = 0.057; Methods). The inventors also observed a significantly higher TMB in KMT2DL0FMSI compared to KMT2DmMSI cases (Welch's t-test p-value < 0.001; FIG. 5B; Methods), which supports previous observations showing that KMT2D deficiency led to increased genomic instability in mouse embryonic fibroblast and human colorectal cell line models

[0086] . A comparison of mutational signatures using the Catalogue Of Somatic Mutations In Cancer (COSMIC) indicated that the single base pair substitution (SBS) signatures associated with MMR deficiency were not significantly different between KMT2DL0Fand KMT2DWTMSI cases (BH-corrected Welch's p-value > 0.05; Methods). However, several other mutational signatures, namely SBS1, SBS5, SBS22, and SBS54, were elevated in KMT2DL0FMSI cases compared to KMT2DWTMSI cases (uncorrected Welch's p-value < 0.05 but > 0.05 when using BH correction). SBS54 is a possible sequencing artefact, whereas SBS22 is associated with aristolochic acid exposure [37,87]. Interestingly, increased SBS1 and SBS5 mutation rates are associated with increased mutational burden and have been linked to clock-like signatures [37,87]. However, a comparison of chronological age did not show a significant difference between KMT2DF0Fand KMT2D'irTcases (Welch's t-test p-value > 0.05). SBS1 arises from an endogenous mutational process that generates guanine:thymine mismatches in double-stranded DNA due to the deamination of 5-methylcytosine tothymine [37,87,88]. Therefore, the elevated SBS1 signature in KMT2DL0Fcases indicates that LOF alterations in KMT2D may induce endogenous mismatches in addition to those acquired due to MSI (FIG. 5D). The inventors also observed significantly higher neoantigen scores in KMT2DF0FMSI cases than in KMT2D'mMSI cases in COAD / READ (Welch's t-test p-value < 0.05; FIG. 50; Methods). Next, the inventors compared the expression of CD274 (encoding PD-L1) and CTLA4.Although not statistically significant, the inventors observed a trend towards higher CD274 expression in KMT2DF0FMSI cases (4.38 mean FPKM) compared to KMT2D'mMSI cases (2.67 mean FPKM; Welch's t-test p-value = 0.15). No such trend was observed in CTLA4 expression (Welch's t-test p-value = 0.45; Methods). Furthermore, the inventors also compared the difference in immune cytolytic activity scores, a measure for anti-tumour immune cell activity [80,81], and found that KMT2DL0FMSI cases had significantly higher immune cytolytic activity scores compared to KMT2D'mMSI cases (Welch's t-test p-value < 0.05; Methods).Together with the trending increase in CD274 expression, without being bound by theory these results may indicate that KMT2DF0FMSI COAD / READ cancers have increased T-cell infiltration and anti-tumour immune cell activity compared to KMT2DmMSI cases. The inventors next compared M1 / M2 macrophage scores. While there was not a statistically significant difference, M1 / M2 scores did show elevated trends in KMT2DL0FMSI COAD / READ cases compared to KMT2D'mMSI COAD / READ cases (Methods). Next, the inventors calculated the IFNy signalling scores

[0089] and found that KMT2DWFMSI cases had significantly higher scores than KMT2D'mMSI (Welch's t-test p-value < 0.05; Methods). The inventors also compared HRD scores and did not observe a significant difference in HRD scores between KMT2DL0FMSI and KMT2D'mMSI cases (Methods).

[0084] Finally, to predict whether KMT2DF0FMSI COAD / READ cases might respond more favourably to ICI treatment than KMT2D'mMSI COAD / READ cases, the inventors calculated their ICI response scores using PredictIO

[0049] . Briefly, PredictIO uses a signature based on 100 genes that best predict response to ICIs across various cancer types

[0049] . While the inventors did not find a statistically significant difference, they did observe a higher trend in ICI response scores in KMT2DF0FMSI cases (Welch's t-test p-value < 0.1; FIG. 5E; Methods), indicatingthat KMT2DL0FMSI COAD / READ cases may exhibit a more favourable response when treated with ICIs than KMT2D'mMSI cases. Overall, despite the small number of MSI cases challenging the statistical analyses, the inventors were still able to show, for the first time, that KMT2DL0Falterations in MSI cases appear to be associated with significantly higher TMB neoantigen scores and SBS1 activity scores than MSI alone (Welch's t-test p-value < 0.001; FIG. 5B). Furthermore, the inventors showed that KMT2DL0FMSI cases had significantly higher expression of neoantigens, immune-activation signatures, cytotoxic T-cell proportions, and ICI- response scores (FIG. 5C), all consistent with the possibility of more favourable ICI responses compared to KMT2D'mMSI COAD / READ cases.

[0085] KMT2DLOf:MSI alterations may be associated with elevated trends in immune response indicators in the POG and MSK-IMPACT cohorts. The inventors next analysed two cohorts, the advanced and metastatic cancer patients from the Personalized OncoGenomics (POG) Project at BC Cancer (NCT02155621 [41,42]) and advanced cancer patients treated with ICI from MSK-IMPACT (NCT01775072

[0048] ). POG and MSK-IMPACT cases differ from untreated TCGA cases, as they are at an advanced stage, are incurable, and have been heavily treated. They closely represent the patient populations typically receiving seconder third-line immunotherapy. The inventors first identified COAD / READ cases harbouring KMT2DL0Fmutations and their MS status. However, the number of MSS and MSI cases harbouring KMT2DL0Falterations was small (six and one case, respectively). Therefore, the inventors also included all solid cancers, which included COAD / READ and 26 other cancer types, and identified 124 KMT2DL0FMSS cases and five KMT2DL0FMSI cases. Similar to the trends seen in TCGA-COAD / READ MSI cohort, the inventors computed statistically significant lower overall survival probability in POG KMT2DF0FMSS solid cancer cases than in KMT2D'mMSS solid cancer cases (Log-rank test p-value = 0.01). However, the inventors did not conduct survival analyses in COAD / READ MSS and MSI solid cancer cases due to limited case numbers. MSK-IMPACT data included few KMT2DL0FMSS and MSI cases, and no significant difference in survival was observed. The analysis thus showed that, at least in the MSS heavily treated POG patients, KMT2DF0Fcases appeared to have reduced overall survival probabilities compared to KMT2DvrFcases, although this did not reach significance, as was also seen in untreated TCGA cases.

[0086] Next, the inventors analysed ICI response markers in POG and MSK- IMPACT cases (FIGs. 6A and 6B; Methods), as done with TCGA cases. For the POG cohort, no comparisons were made in COAD / READ MSI cases for any ICI response markers since there was only one case harbouring KMT2DL0Falterations, as mentioned above. In the MSK-IMPACT cohort, only TMB was estimated since this study sequenced a targeted panel of genes and lacked transcriptome data

[0048] . In POG and in MSK-IMPACT data for COAD / READ and solid cancers, the inventors saw significantly higher TMB in KMT2DL0FMSS cases compared to KMT2D'mMSS cases (Welch's t-test p-value < 0.05; FIG. 6A). Furthermore, in the MSK-IMPACT cohort, compared to KMT2DVJMSI solid cancer cases, TMB was also significantly elevated in KMT2DL0FMSI solid cancer cases (Welch's t-test p-value < 0.05; FIG. 6C). In the POG solid cancer MSI cases and MSK-IMPACT COAD / READ cases, KMT2DL0Fcases trended towards elevated TMB compared to KMT2DWTcases (FIGs. 6A, 6C). The inventors also saw trends in the same direction for TUBA1B, CD274 and CTLA4 expression; M1 / M2 scores; IFNy scores; CD8+ T-cell scores; and ICI response scores (FIG. 6B) in KMT2DL0FMSI solid cancer compared to KMT2DWTMSI solid cancer cases. The substantially smaller sample sizes (a total of one KMT2DL0Fcase in the COAD / READ MSI cases and five in the solid cancer MSI cases) resulted in insufficient power to detect significant differences in the POG cohort. Even so, these findings are consistent with a similar effect of KMT2DL0Fin the POG and MSK- IMPACT cohorts as observed in TCGA. These results reinforce the notion that MSI cases with KMT2DL0Falterations may respond more favourably to ICI treatment than KMT2D'mMSI cases and that consideration of KMT2D mutational status may improve patient stratification for ICI treatment.

[0087] KMT2D mutational status significantly stratifies MSI COAD / READ cell lines sensitive to WRN inhibitors. Given that WRN is a known essential gene in MSI cancer cell lines [4,6-9], the inventors investigated whether microsatellite status played a role in the SL interaction predicted between WRN and K.MT2D. To this end, the inventors used MSI status to stratify the WRN KO lethality probability scores of KMT2DL0Fcell lines and KMT2Dmcell lines from the pan-cancerdataset. The inventors show that lethality probabilities in KMT2DL0FMSI cell lines, resulting from WRN KO, were statistically significant and higher than KMT2Dmcell lines, including the KMT2D'mMSI group (ANOVA followed by Tukey's HSD tests p- value < 0.001; FIG. 7A; Methods). The inventors also found that KMT2DF0FMSI lines had significantly higher WRN KO lethality probabilities than KMT2DFQFmicrosatellite stable (MSS) cell lines (ANOVA followed by Tukey's HSD test p-value < 0.001). However, no difference in WRN KO lethality probabilities was observed between KMT2D'mMSI and KMT2D'mMSS cell lines. This result indicated that WRN KO was specifically lethal in MSI cell lines harbouring KMT2DL0Falterations. This analysis was not performed in a COAD / READ-specific manner due to a lack of COAD / READ cell lines in DepMap. Next, the inventors used GRETTA to assess whether cancer cell lines with WRN LOF (1 / R / VLOF) alterations might require K.MT2D for survival. The inventors found that 1 / R / VLOFMSS lines were significantly more susceptible to KMT2D perturbation than WRN WT (1 / R / VWT) MSS lines (ANOVA followed by Tukey's HSD test p-value < 0.01; FIG. 7B; Methods). In FIGs. 7A and 7B, ell lines were separated by microsatellite status as determined in Ghandi et al. (2019)

[0015] . ANOVA followed by Tukey's HSD p-value * < 0.05, ** < 0.01, *** < 0.001, NS > 0.05. NA indicates cancer cell lines without MSI / MSS status annotations. These results are consistent with the notion that WRN is a SL interactor of KMT2D and that MSI alone may not confer a dependency on WRN for survival. In other words, the inventors' analysis indicated that WRN may be required for MSI cancer cell survival only when KMT2DL0Fmutations are present and that MSI alone may not result in WRN dependence, in contrast to initial descriptions [4,6-9].

[0088] Defects in MMR functions as a result of LOF mutations or epigenetic silencing of MMR-associated genes (including MLH1, MLH3, MSH2, MSH3, MSH6, and PMS2") are causes of MSI, resulting in numerous unrepaired mutations across the genome, mainly in repetitive sequences

[0090] . Although several studies have shown that WRN is a SL target of MSI cancer cells, restoration of MMR functions through ectopic expression of MMR-associated genes in MSI cell lines only showed partial rescue of cell viability [6] upon WRN KO, and generating MSI conditions through MMR gene KOs in MSS cell lines failed to recapitulate the lethal effects ofWRN KO [8]. Therefore, the inventors wanted to see whether KMT2DL0Falterations in MSI cell lines would lead to a dependency on WRN for survival. To this end, the inventors identified cancer cell lines, in which WRN KO was performed and the cell's viability was validated (i.e. viability assays were performed on cell lines using a single sgRNA targeting WRN instead of a pooled sgRNA KO screen) and analysed whether MSI cell lines that harboured KMT2DL0Falterations showed more consistent lethal effects upon WRN KO compared to KMT2D'mMSI cell lines. The inventors identified 32 cancer cell lines consisting of COAD / READ, STAD, UCEC, and ovarian (OV) cancers that had genomic sequencing data from DepMap and (in)dependence on WRN for survival validated by studies conducted by either Behan et al. [4], Chan et al. [6], Kategaya et al. [5], Lieb et al. [7], or Picco et al. [9]. For these cell lines, the inventors annotated the presence of LOF alterations in KMT2D and MMR genes, namely EX01, MBD4, MLH1, MSH2, MSH3, MSH5, MSH6, PMS1, PSM2, and RFC1 [9], reported MSI status, and the outcomes of WRN KO viability assays (Methods). Among the 14 cell lines that showed WRN was an essential gene (i.e. that required WRN for survival), the inventors identified LOF alterations of MMR genes in 85% (12 / 14) and concurrent KMT2DF0Falterations in 64% (9 / 14). Among the 18 cell lines where WRN was reported as a non-essential gene, the inventors found only 33% with MMR or KMT2D alterations (4 / 18 with MMR gene LOF and 2 / 18 with KMT2D-Q?. These results indicate that most MMR-deficient lines harbour KMT2DF0Falterations and that not all MMR-deficient cell lines require WRN for survival. Furthermore, both KMT2DL0Falterations and LOF alteration in MMR genes were significantly enriched in cell lines where WRN was essential (Fisher's exact test p- value = 0.0007 with odds ratio [OR] = 20 and p-value = 0.0009 with OR = 21, respectively; FIG. 7B-7D). However, LOF alterations in KMT2D and MMR also cooccurred (Fisher's exact test p-values = 0.0091 and OR = 9; FIG. 7D). Given the limited sample size and mixture of cancer types, the inventors were not able to distinguish the independent effects of KMT2DWFalterations and MMR deficiency in specific cancer types.

[0089] Next, given that MSI cells require WRN to resolve expanded microsatellite regions containing [AT]nrepeat elements for survival and that these expansions could not be recapitulated by MSS cells upon MMR gene KO [8], the inventorssought to determine whether the loss of KMT2D affected the expansion of [AT]nmicrosatellite regions in MSI cell lines. To this end, the inventors used ExpansionHunter Denovo

[0051] (Methods) to profile [AT]n microsatellite regions in WGS data from DepMap cancer cell lines (SRA project: PRJNA523380). Briefly, ExpansionHunter Denovo is a catalogue-free method for genome-wide repeat expansion detection and has been used previously to estimate [AT]nrepeat element expansion [8]. The inventors characterised microsatellite regions of 14 cell lines, including nine KMT2DL0FMSI, two KMT2D'mMSI, and two KMT2D'mMSS lines, and subsequently filtered for lines that contained more than one anchored in-repeat read (i.e. one read pair maps within and the other outside of a repeat region) in [AT]nregions. This resulted in profiles of nine cell lines, specifically CCK81, LOVO, SW48, LS180, MFE319, RKO, SKMEL2, NCIH747, and COLO201 lines. Interestingly, between KMT2DL0FMSI, KMT2D'mMSI, and KMT2D'mMSS lines, [AT]nsites were estimated to be the most expanded in KMT2DL0FMSI cell lines, namely LOVO, SW48, LS180 (COAD / READ), and MFE319 (UCEC; FIG. 7C). The LOVO [5] and SW48 [4] lines were previously shown to depend on WRN for survival. Two KMT2DL0?MSI lines, CCK81 and RKO (a line that also depends on WRN for survival) [7], showed below average [AT]nexpansion size across events (FIG. 7C; mean [AT]nexpansion size across samples = 193); however, the number of [AT]nregions detected was among the highest of all samples in both lines. Two KMT2D'mlines, SKMEL2 (MSI) and COLO201 (MSS), also showed below-average [AT]nexpansion size. The KMT2D'mMSS cell line NCIH747 showed above-average [AT]nexpansion size; however, the [AT]nsize was >50% below that of the expanded KMT2D'-0FMSI lines. Although the inventors are not able to perform statistical comparisons of [AT]nexpansion between KMT2D'mMSI and KMT2D-Q?MSI cell lines due to lack of available data for MSI cell lines, these results showed a pattern consistent with the notion that KMT2DL0Falteration is associated with increased size and number of expanded [AT]nmicrosatellite repeat regions.

[0090] Ferretti et al.

[0013] and Baltgalvis et al.

[0014] independently reported WRN inhibitors, HRO761 and VVD-133214, respectively, which are in phase I clinical trials (NCT05838768 and NCT06004245). Both phase I trials are recruiting patients with advanced metastatic solid tumours that are MSI or MMR gene-deficient.Interestingly, both papers reported MSI COAD / READ cell lines that were insensitive to WRN inhibitors [13,14]. The inventors sought to determine whether cell line sensitivity to these WRN inhibitors could be explained by KMT2D mutational status. Using DepMap data, the inventors annotated the KMT2D and MMR gene mutational statuses of the MSI COAD / READ cell lines treated with HRO761 or VVD-133214 (Methods), and compared these annotations to the inhibitor sensitivity data reported for the cell lines in

[0013] and

[0014] . With the exception of the KM12L4 and LS174T lines, which were WRN inhibitor-sensitive but did not have DepMap mutational data available for this analysis, the inventors were able to show that KMT2DL0Falterations were present in all drug-sensitive cell lines (Fisher's exact test p-value = 0.0035; FIGs. 7D-7E). Interestingly, KM12L4 and LS174T originate from the same primary COAD / READ patient biopsy sample as the KMT2DWFMSI cell lines KM12

[0092] and LS180 [93,94], respectively (shown in FIGs. 7D and 8A). The KM12L4 cell line was established from spontaneous metastatic cells that were produced from the KM 12 cell line being injected into the mouse spleen

[0092] . The LS174T and LS180 cell lines were derived from the same primary culture, where LS174T and LS180 lines were established from a trypsinized and a scrapped non- trypsinized passage, respectively

[0093] . Without being bound, it is thus possible that the KM12L4 and LS174T cell lines harbour the same KMT2DL0?alterations as KM12 and LS180. However, even without mutational data for the KM12L4 and LS174T cell lines, these results significantly associate sensitivity to WRN inhibitors of MSI COAD / READ cell lines with KMT2DL0Fmutations as the inventors predicted using in silico GI mapping. FIG. 8B, 8C, and 8D show contingency tables comparing the frequency of mutations and WRN perturbation sensitivity of cell lines in FIG. 8A. Specifically, FIG. 8B compares the frequency of KMT2DL0Falterations and WRN perturbation sensitivity, FIG. 8C compares the frequency of MMR genes LOF alterations and WRN perturbation sensitivity, and FIG. 8D compares the frequency of MMR genes LOAalterations and KMT2DL0Falterations. WT+ denotes both wildtype alleles and missense mutations. Thus, using in silico GI mapping predictions and in vitro experimental data, the inventors have now shown that KMT2DL0Fmutations are able to identify MSI COAD / READ cell lines that are sensitive to WRN perturbation.Example 2.0 - Discussion

[0091] The inventors computationally mapped genetic networks of KMT2D, a tumour suppressor gene frequently mutated in several cancer types. Using KMT2D loss-of-function {KMT2DL0^ mutations as a model, the inventors illustrate the utility of in siiico genetic networks in uncovering novel functional associations and vulnerabilities in cancer cells with LOF alterations affecting tumour suppressor genes. The inventors revealed genetic interactors with functions in histone modification, metabolism, and immune response and synthetic lethal (SL) candidates, including some encoding existing therapeutic targets. Notably, the inventors predicted WRN as a novel SL interactor and, using recently available WRN inhibitor (HRO761 and VVD-133214) treatment response data, the inventors observed that KMT2D mutational status significantly distinguishes treatmentsensitive MSI cell lines from treatment-insensitive MSI cell lines.

[0092] In this study, the inventors demonstrated a framework for revealing tumour suppressor gene functions and vulnerabilities of cancer cells harbouring LOF alterations in them. Using the tumour suppressor gene KMT2D as an example for this framework, the inventors performed in siiico genetic network analyses to gain further insights into its functions and identify its cancer type-specific genetic interactors. The inventors identified genes in KMT2D's essentiality network and proteins in its proteomic network that were known interactors and also new candidate interactors that have roles associated with the regulation of cell cycle, cell division, telomere maintenance, chromosome segregation, DNA replication and metabolism, thus expanding the extent of KMT2D's role in these processes. The inventors identified SL and AL candidates that have roles in mitotic processes, metabolic processes, and immune response. Furthermore, results from the in siiico GI screens identified several SL candidates— namely NDUFB5 (in the STAD screen), MDM2 and TUBA1B (in the COAD / READ screen), and WRN (in the pan-cancer and COAD / READ screens)— which encode proteins that can be inhibited using existing and in-development drugs. Interestingly, in addition to showing genes associated with immune response as K.MT2D SL interactors, which included TUBAIB, the inventors also showed that ICI response makers were elevated in KMT2DL0Fcases.Using untreated TCGA-COAD / READ cases, the inventors showed that KMT2DL0FMSI cases have a trending increase in TUBA1B, CD274, and CTLA4 mRNA expression, CD8+T-cell infiltration scores, and favourable ICI treatment response scores and have significantly higher TMB, neoantigen scores, cytolytic activity, IFNy signalling scores, and yd T-cell infiltration scores compared to KMT2DVJcases. In a retrospective analysis of heavily treated POG cohorts, the inventors found similar trends as in TCGA cohorts. Finally, the inventors also showed that KMT2D may mediate MSI cell line dependency on WRN for survival, possibly through further expansion of [TA]nmicrosatellite regions. The inventors also show that KMT2D mutational status is significantly associated with WRN inhibitor sensitivity in MSI cell lines. These results expand KMT2D's roles in genomic stability, metabolism, and immune response; presents promising SL interactors that can potentially be targeted using drugs, including two in clinical trials; and proposes ICIs and / or WRN inhibitors as potential therapeutic avenues for KMT2DWFMSI cases.

[0093] Although KMT2D is best characterised for its role as a histone methyltransferase, loss of KMT2D has been associated with increased genomic instability, dysregulation of DNA repair / replication [86,95], and metabolic dysregulation [96,97]. However, the extent of KMT2D's involvement in regulating genomic stability and metabolism is still unclear. In addition to supporting KMT2D's known role as a histone modifier, this analysis of KMT2D's co-essential network and chromatin-specific proteome interaction network reveal an enrichment of genes / proteins related to cell cycle processes, mitotic segregation, and DNA replication / repair that have not previously been linked to KMT2D. These novel candidate SL interactors were also involved in mitotic regulation and homologous recombination. These processes are known to affect genomic stability

[0098] and thus further implicate KMT2D in this role.

[0094] For the first time, the inventors predicted KMT2D candidate SL interactors, namely NDUFB5, MDM2, TUBA1B, and WRN, which encode targets of existing drugs or those that are in development.

[0095] The inventors also contribute results that further implicate KMT2D with a role in immune response regulation. MSI is a marker for ICI treatment; however, only ~53% of MSI cases have an objective response to ICI treatment

[0099] ,suggesting additional biomarkers are necessary to stratify MSI cases. Several biomarkers have been proposed for ICI treatment, including HRD

[0100] , ARID1ALOF

[0101] , and SETD2L0F

[0102] . Wang et al.

[0103] also demonstrated that mice transplanted with KMT2D-deficient tumours were responsive to ICI treatment and that TMB was elevated in TCGA cases with KMT2DL0F. The elevated TMB in the MSS KMT2DL0Fcases compared to the MSS KMT2D'min the POG and MSK-IMPACT COAD / READ cohorts is consistent with previous studies showing loss of KMT2D leading to increased genomic instability [86,95]. In this study, the inventors show for the first time that candidate KMT2D SL interactors were involved in T-cell cytotoxicity and immune response regulation. Notably, the inventors identified TUBAIB as a highly attractive SL candidate, where its overexpression has recently been implicated in favourable ICI treatment outcomes

[0104] . The inventors not only show that TUBAIB was overexpressed in KMT2DL0FMSI cancers compared to KMT2DmMSI cancers in two cohorts, TCGA and POG, but the inventors also show that several established ICI response markers are elevated in these cases. Although the TCGA, POG, and MSI-IMPACT cohort analyses had a small number of KMT2DL0FMSI cases, these data are compatible with the notion that KMT2DL0FMSI cases might be more sensitive to ICI treatment than KMT2DmMSI cases. This is consistent with previous studies, which also used small cohorts, to show that a larger proportion of KMT2DL0Fcases had a durable response to ICI treatment compared to KMT2Dmcases [103,105]. However, given that large ICI-treated cohort studies, such as KEYNOTE-177 [NCT02563002], do not collect genomic data, there is an unmet future need for cohorts with sufficient sample sizes of KMT2DWTMSI cases and KMT2DL0FMSI cases to perform more robust statistical analyses. Altogether, these results suggest that KMT2DL0FMSI cancers may respond to ICI treatments, and thus identify KMT2D as a biomarker to further stratify MSI cases for ICI treatment.

[0096] Finally, the SL interaction between KMT2D and WRN contributes a new layer of understanding in MSI cancer cell vulnerability and potential strategies for identifying patients that are sensitive to WRN inhibitors. The inventors show for the first time that MSI cell lines harbouring KMT2DL0Falterations are more sensitive to WRN KO than KMT2D'mMSI cell lines, contrary to the previous studies indicatingthat the MSI feature alone is sufficient for cancer cells to confer a survival vulnerability to WRN perturbation [4,6-9]. Although it is unclear how KMT2D may directly mediate survival dependency on WRN in MSI cancer cells, it is interesting that MSH6, a gene that functions in MMR

[0106] , appeared as a co-essential gene and protein interactor of KMT2D. The inventors also show that [AT]nrepeat elements were relatively larger and more abundant than KMT2DWTMSI cell lines. Further studies are needed to better understand the possible mechanisms of the KMT2D -MMR. relationship. Most notably, the inventors were able to use the KMT2DL0Fmutational status to identify MSI cancer cell lines that are sensitive to WRN inhibitors, which are in phase I clinical trials. The inventors have shown that determining the KMT2D mutational status of cancers could play a role in selecting patients who are most likely to respond to WRN inhibitors in future clinical trials, where drug efficacy is assessed. Furthermore, given that ICI-resistant MSI xenograft models have been shown to be sensitive to the VVD-133214 WRN inhibitor

[0014] , the inventors predict that I CI -refractory patients with KMT2DL0Fmutations may respond to WRN inhibitors.

[0097] The inventors' study of in silico genetic networks identified several potential roles for KMT2D and further implicated it in the regulation of genomic stability and metabolism. The inventors highlight cancer type-specific genetic interactors, namely NDUFB5, MDM2, TUBA1B, and WRN, which are promising targets of existing and in-development drugs. Furthermore, using cancer patient data, the inventors provide evidence for KMT2DL0Falterations as a potential biomarker for ICI and WRN inhibitor treatments. Altogether, this work serves as an example for identifying novel functional associations and potential targeted treatment opportunities for cancers with tumour suppressor gene LOF alterations.

[0098] Here, the inventors report the use of an in silico genetic screening approach to systematically characterise tumour suppressor gene function and vulnerabilities of cancer cells harbouring LOF alterations in a tumour suppressor gene. The inventors applied their method to map the genetic networks of K.MT2D, a frequently mutated tumour suppressor gene across cancer types

[0107] .Examination of KMT2D's essentiality network revealed novel associations with genes that play roles in histone modification, transcription, mitotic cell cycle regulation,glycolysis, and DNA replication, and those encoding proteins that interact with KMT2D on the chromatin.

[0099] Through mapping cancer type-specific in siiico KMT2D GI networks, the inventors revealed several SL candidates, namely NDUFB4, MDM2, TUBA1B, and WRN, that encode targets of existing and in-development therapeutics, making them potentially viable drug targets in cancers harbouring KMT2DL0Falterations. Using The Cancer Genome Atlas (TCGA) data, the inventors showed that dysregulated SL candidate-associated functions, such as in p53 regulation and metabolism. Using data from TCGA, Memorial Sloan Kettering Cancer Center Integrated Mutation Profiling of Actionable Cancer Targets (MSK-IMPACT; NCT01775072

[0048] ), and the Personalized OncoGenomics (POG) Program at BC Cancer (NCT021556210 [41,42]), a cohort of advanced and metastatic cancer patients that have been heavily treated; the inventors also showed that MSI cases with KMT2DL0Falterations show significantly elevated immune checkpoint response markers compared to KMT2DVJMSI cases, indicating KMT2DL0Fmutations may be a biomarker for immune checkpoint inhibitor (ICI) treatment stratification. Strikingly, compared to KMT2DVJcancer cell lines with microsatellite instability (MSI), the inventors show that KMT2DL0FMSI cell lines are more sensitive to WRN KO and thus, as predicted, to treatment with two recently published WRN inhibitors, which are in phase I clinical trials (NCT05838768

[0013] and NCT06004245

[0014] ). The inventors' work thus models a more general approach, in which in siiico genetic network maps are used to identify novel functional associations, cancer cell vulnerabilities, and novel treatment opportunities associated with tumour suppressor gene alterations.

[0100] MSI is found in ~10-30% of colorectal, endometrial, ovarian, gastric and other cancer types [108,109], and KMT2D LOF mutations are found in ~40-60% of these MSI cases. The inventors have demonstrated that, in COAD / READ, MSI cases harbouring KMT2D LOF mutations will be sensitive to WRN inhibitor treatment, but cases harbouring the wildtype KMT2D allele will be insensitive to treatment. The inventors believe that KMT2D mutational status should be considered by physicians to determine whether a patient will respond to WRN inhibitor treatments. The inventors also believe that other cancer types with high prevalence of MSI cases(eg. Endometrial, ovarian, and gastric cancers) will likely benefit from determining the KMT2D mutational status of the tumour prior to WRN inhibitor treatment. WRN inhibitors have been shown to be effective in patient xenograft models harbouring MSI cancers resistant to resistant to immune checkpoint inhibitors (ICIs

[0014] ).

[0101] Two breakthrough Nature papers were published in May 2024 [13,14] by authors from Vividion / Roche and Novartis, respectively, demonstrating the efficacy of their WRN inhibitors in inhibiting the growth of MSI cancer cell lines. Both these inhibitors are now being tested in phase I clinical trials (NCT06004245 [Vividion / Roche] and NCT05838768 [Novartis]). Using data from both Nature papers, the inventors have shown that KMT2D loss-of-function mutational status significantly distinguishes MSI cancer cells that are sensitive to these inhibitors from MSI cells that are not. Thus, the inventors believe that their data show that KMT2D mutational status has considerable potential to predict response to both WRN inhibitors. These results are consistent with and predicted by the study results described in the specification herein.

[0102] In conclusion, the inventors have discovered that loss-of-function (LOF) mutations in the gene KMT2D appeared to be required to render cancer cells with microsatellite instability (MSI), such as many colorectal cancers, sensitive to WRN inhibitors. This surprising discovery emerged from use of the inventors' software GRETTA, leading to a predicted synthetic lethal interaction between KMT2D and WRN in colorectal adenocarcinoma (COAD / READ) cell lines. This prediction showed that, surprisingly, only MSI cancer cell lines with KMT2D LOF mutations were selectively killed upon WRN knockout, contrary to current understanding that MSI is the sole biomarker needed to sensitize cells to WRN knockout [8]. The inventors were able to validate their prediction using data from two recent Nature publications [13,14], where they treated MSI cancer cell lines with their WRN inhibitors (VVD-133214 [Vividion / Roche] and HRO761 [Novartis]). Analysing data from these publications, the inventors discovered that the presence of KMT2D LOF mutations significantly distinguished MSI cells that were sensitive to WRN inhibitors from MSI cells that were insensitive. There are now two phase I clinical trials (NCT06004245 [Vividion / Roche] and NCT05838768 [Novartis]) that have been initiated, testing toxicity and dosage of the two WRN inhibitors. These WRNinhibitors are currently indicated to be effective against all MSI cases. However, the results presented herein demonstrate that WRN inhibitors will largely be ineffective against MSI cases harbouring KMT2D wild type alleles, which make up ~40-60% of MSI cases. The inventors' discovery will allow oncologists to identify patients sensitive to WRN inhibitors by sequencing for KMT2D LOF alterations in MSI cases.Example 3.0 - KMT2D Protein Expression Levels do not Distinguish WRN Inhibitor Sensitive Cancer Cells

[0103] Given that mass spectrometry- based methods of protein quantification is not an accessible technology for pathology labs, the inventors conducted additional experiments to evaluate whether an western blot-based approach to quantify expression levels of KMT2D protein could be used to predict the likely sensitivity of cancer cells to treatment with WRN inhibitors. Briefly, using standard western blot protocols, the inventors loaded 50ug of protein per sample. Ponceau S staining was used to determine total proteins loaded. KMT2D, WRN, and vinculin (VINC) protein expression was detected and quantified using anti-KMT2D (Abeam #ab213721), anti-WRN (Novus #NB100-472), and anti-VINC (Abeam #abl29002) antibodies, respectively. KMT2D expression was then normalised to Ponceau S staining to determine relative protein abundance across cell lines.

[0104] Results are shown in FIGs. 9A and 9B. Representative results of one of three replicate Western blots are shown in FIG. 9A, with expression levels of KMT2D shown relative to Ponceau S, quantifying total protein loaded, shown in FIG. 9B. Three separate MSI KMT2D'mcell lines DLD1, HCT8, and HCT15 were assessed, as well as six separate MSI KMT2DL0Fcell lines LS180, LS174T, KM12, KM12L4, HCT116 and RKO. FIG. 9B plots the relative abundance of KMT2D for each cell line relative to Ponceau S. Each point shows relative expression levels from a replicate. Error bars show standard deviation. No significant difference was found between KMT2D relative protein abundance between the MSI-positive KMT2Dmcell lines and MSI KMT2DL0Fcell lines. As the differences in relative expression of KMT2D protein are not significantly different between the KMT2D'mand KMT2DL0Fcell lines evaluated, these data suggest that assessment of KMT2D expression levels shouldnot be used to evaluate the likelihood that cancer cells will be sensitive to treatment with WRN inhibitors.Example 4.0 - Development of a Gene Signature for Predicting Sensitivity to Treatment with WRN Inhibitors

[0105] Given that KMT2D LOF alteration did not capture all cell lines that were vulnerable to WRN inhibitors, as mentioned above, the authors performed GRETTA in siiico genetic interaction screens on all ~ 18,000 genes affected by LOF alterations (in the presence of MSI) to identify other SL partners of WRN.

[0106] FIG. 10 shows 31 pan-cancer cell lines with validated WRN sensitivity data, and the mutational status of genes predicted to be WRN SL partners. Only the top 10 most frequently altered WRN SL partners are shown. ARID1A, KMT2D and RAD50 were the most commonly altered WRN SL interactor among the cell lines sensitive to WRN disruption. Based on this data, the inventors have determined that a more precise gene signature to consider when determining if a cancer cell that is MSI is likely to be sensitive to treatment with WRN inhibitors. WRN inhibitor sensitive cancer cells are MSI-positive and have loss-of-function mutations in at least two of KMT2D, ARID1A and RAD50; or are MSI-positive with LOF mutation in ARID1A only, without loss-of-function mutations in either of KMT2D and RAD50. On the other hand, the cell is unlikely to be sensitive to treatment with WRN inhibitors if the cancer cell is MSS; or the cell is MSI and both ARID1A and RAD50 are wild type, the cell is unlikely to be sensitive to treatment with WRN inhibitors.

[0107] While a number of exemplary aspects and embodiments have been discussed above, those of skill in the art will recognize certain modifications, permutations, additions and sub-combinations thereof. It is therefore intended that the following appended claims and claims hereafter introduced are interpreted to include all such modifications, permutations, additions and sub-combinations as are consistent with the broadest interpretation of the specification as a whole.REFERENCES

[0108] The following references are of interest with respect to the subject matter described herein. Each of the following references is incorporated by reference herein in its entirety:1. Futami et al. Front Genet. 2014;5:441.2. Mukherjee et al. Int J Mol Sci. 2018;19.3. Marabitti et al. Int J Mol Sci. 2022;23.4. Behan et al. Nature. 2019;568:511-6.5. Kategaya et al. iScience. 2019;13:488-97.6. Chan et al. Nature. 2019;568:551-6.7. Lieb et al. Elife. 2019;8. doi.org / 10.7554 / eLife.433338. van Wietmarschen et al. Nature. 2020;586:292-8.9. Picco et al. Cancer Discov. 2021. doi.org / 10.1158 / 2159-8290. CD-20-150810. Mullard A. Nat Rev Drug Discov. 2022;21 :477-9.11. Morales-Juarez and Jackson, npj Precision Oncology. 2022;6: 1-6.12. Yap et al. Cancer Res. 2025;85 (8_Supplement_2): CT016.13. Ferretti et al. Nature. 2024;629:443-9.14. Baltgalvis et al. Nature. 2024;629:435-42.15. Ghandi et al. Nature. 2019;569:503-8.16. US Food & Drug Administration, "FDA approves pembrolizumab for adults and children with TMB-H solid tumours", 6 / 17 / 2020.17. Lee et al. Cancer Res. 2025;85 (8_Supplement_l): 2913.18. Ellrott et al. Cell Syst. 2018;6:271-81. e7.19. Morin et al. Nature. 2011;476:298-303.20. Spina et al. Blood. 2016;128: 1362-73.21. Bea et al. Proc Natl Acad Sci U S A. 2013;110: 18250-5.22. Parsons et al. Science. 2011;331:435-9.23. George et al. Nature. 2015;524:47-53.24. Takemon and Marra. Bioinformatics. 2023; 39(6):btad381.25. Meyers et al. Nat Genet. 2017;49: 1779-84.26. Nusinow et al. Cell. 2020;180:387-402. el6.27. Takemon et al. Cancers. 2023;15:2805.28. Ura et al. Biology. 2021; 10. doi.org / 10.3390 / biologyl004025629. Takemon et al. Genome Med. 2024;16, 136.30. Ochoa et al. Nucleic Acids Res. 2023;51: D1353-9.31. Brown et al. Medchemcomm. 2018;9:606-13.32. Schneider et al. Nat Rev Drug Discov. 2021;20:789-97.33. Cerami et al. Cancer Discov. 2012;2:401-4.34. Gao et al. Sci Signal. 2013;6: 1.35. Middha et al. JCO Precis Oncol. 2017;2017. doi.org / 10.1200 / PO.17.0008436. Kautto et al. Oncotarget. 2017;8:7452-63.37. Alexandrov et al. Nature. 2020;578:94-101.38. Diaz-Gay et al. bioRxiv. 2023 [cited 2023 Nov 1]. p. 2023.07.10.548264.39. Thorsson et al. Immunity. 2018;48:812-30. el4.40. Goldman et al. Nat Biotechnol. 2020;38:675-8.41. Laskin et al. Cold Spring Harb Mol Case Stud. 2015;l :a000570.42. Pleasance et al. Nature Cancer. 2020;1 :452-68.43. Pleasance et al. Ann Oncol. 2022;33:939-49.44. Kim et al. Nat Methods. 2018;15:591-4.45. Benjamin et al. bioRxiv. 2019 [cited 2023 Dec 14]. p. 861054. biorxiv.org / content / 10.1101 / 861054vl46. Cleary et al. bioRxiv. 2015 [cited 2023 Dec 14]. p. 023754. biorxiv.org / content / 10.1101 / 023754v247. Titmuss et al. BMC Med Genomics. 2022;15: 190.48. Samstein et al. Nat Genet. 2019;51:202-6.49. Bareche et al. Ann Oncol. 2022;33: 1304-17.50. Li and Durbin. Bioinformatics. 2009;25: 1754-60.51. Dolzhenko et al. Genome Biol. 2020;21: 102.52. R Core Team. R: A Language and Environment for Statistical Computing [Internet]. Vienna, Austria: R Foundation for Statistical Computing; 2020.53. Perez-Riverol et. al. Nucleic Acids Res. 2019;47: D442-50.54. Freeberg et al. Nucleic Acids Res. 2022;50: D980-7.55. Pan et al. Cell Syst. 2018;6:555-68. e7.56. Wainberg et al. Nat Genet. 2021;53:638-49.57. El-Brolosy et al. Nature. 2019;568: 193-7.58. Liu et al. Cell. 2016;165:535-50.59. Takemon et al. Elife. 2021;10. doi.org / 10.7554 / eLife.6258560. Hu et al. Mol Cell Biol. 2013;33:4745-54.61. Lee et al. Elife. 2013;2:e01503.62. Dhar et al. Mol Cell. 2018;70:825-41. e6.63. Ortega-Molina et al. Nat Med. 2015;21 : 1199-208.64. Zhang et al. Nat Med. 2015;21 : 1190-8.65. Froimchuk et al. Gene. 2017;627:337-42.66. Henkel et al. Curr Opin Genet Dev. 2019;54:73-82.67. Wilson .J Am Stat Assoc. 1927;22:209.68. Marabitti et al. Int J Mol Sci [Internet]. 2022;23. doi.org / 10.3390 / ijms23031547.69. Center for Drug Evaluation, Research. FDA grants accelerated approval to pembrolizumab for first tissue / site agnostic indication [Internet]. U.S. Food and Drug Administration. FDA; 2019 [cited 2023 Sep 22].70. Marcus et al. Clin Cancer Res. 2019;25:3753-8.71. Diaz et al. Lancet Oncol. 2022;23:659-70.72. Ardeshir-Larijani et al. Clin Lung Cancer. 2018;19:e489-501.73. Ferrero et al. Haematologica. 2020;105: 1604-12.74. Li et al. Biomark Res. 2020;8:27.75. Morcillo-Garcia et al. PLoS One. 2019; 14:e0209134.76. Subbiah et al. Ann Oncol. 2020;31 : 1115-8.77. Sha et al. Cancer Discov. 2020;10: 1808-25.78. Thorsson et al. Immunity. 2018;48:812-30. el4.79. Galon and Bruni. Nat Rev Drug Discov. 2019;18: 197-218.80. Rooney et al. Cell. 2015;160:48-61.81. Johnson et al. Proc Natl Acad Sci U S A. 2003;100:2657-62.82. Jayasingam et al. Front Oncol. 2019;9: 1512.83. Pender et al. Clin Cancer Res [Internet]. 2020; doi.org / 10.1158 / 1078- 0432.CCR-20-1163.84. Li et al. Br J Cancer. 2022;126: 1663-75.85. Fares et al. Biomark Med. 2022;16:771-82.86. Kantidakis et al. Genes Dev. 2016;30:408-20.87. Tate et al. Nucleic Acids Res. 2019;47: D941-7.88. Nik-Zainal et al. Cell. 2012;149:979-93.89. Ayers et al. J Clin Invest. 2017;127:2930-40.90. Velho et al. World J Gastroenterol. 2014;20: 16433-42.91. Dolzhenko et al. Genome Biol. 2020;21: 102.92. Morikawa et al. Cancer Res. 1988;48:6863-71.93. Tom et al. In Vitro. 1976;12: 180-91.94. Rutzky et al. Cancer Res. 1980;40: 1443-8.95. Ray et al. Nature. 2016;535:382-7.96. Alam et al. Cancer Cell. 2020;37:599-617. e7.97. Pacelli et al. Cells [Internet]. 2020;9. doi.org / 10.3390 / cells9071685.98. De Marco Zompit and Stucki. DNA Repair . 2021;108: 103215.99. Ribas and Wolchok. Science. 2018;359: 1350-5.100. You et al. Front Immunol. 2022;13:871756.101. Li et al. Cancer Cell Int. 2022;22: 347.102. Zheng et al. BMC Cancer. 2023;23:686.103. Wang et al. Cancer Discov. 2020;10: 1912-33.104. Hu et al. Diagnostics (Basel). 2022;12. doi.org / 10.3390 / diagnosticsl2040858.105. Liu et al. Cancer Sci. 2023;114: 1229-39.106. Pecina-Slaus et al. Front Mol Biosci. 2020;7: 122.107. Kandoth et al. Nature. 2013;502:333-9.108. Aaltonen et al., 1993. Science 260(5109):812-816.109. Bonneville et al., 2017. JCO Precision Oncology 1 : 1-15.110. Kim et al. bioRxiv 2025.05.04.652100, 9 May 2025.111. Kodama et al., 2011. Nucleic Acids Res 40(Dl): D54-D56.

Claims

WHAT IS CLAIMED IS:

1. A method of determining whether a cancer of a mammalian subject is likely to be sensitive to inhibitors of Werner syndrome RecQ helicase (WRN) comprising : evaluating cells of the cancer for microsatellite instability (MSI) or mismatch repair deficiency (dMMR); evaluating the cells of the cancer for somatic inactivating alterations in two or all of KMT2D, RAD50 and ARID1A; and if it is determined both : that the cells of the cancer are MSI or dMMR; and one of the following is true: that somatic inactivating alterations are present in two or all of KMT2D, RAD50 and ARID1A in the cells of the cancer; or that somatic inactivating alterations are present in ARID1A in the cells of the cancer; determining that the cancer is likely to be sensitive to inhibitors of WRN .

2. The method as defined in claim 1, wherein : if it is determined that the cells of the cancer are MSS; or if it is determined that: the cells of the cancer are MSI or dMMR;ARID1A is wild type or has only missense mutations in the cells of the cancer; and one of KMT2D and RAD50 are wild type or have only missense mutations in the cells of the cancer; determining that the cancer is likely to be insensitive to inhibitors of WRN .

3. A method of determining whether a cancer that has microsatellite instability (MSI) or mismatch repair deficiency (dMMR) is likely to be sensitive to inhibitors of Werner syndrome RecQ helicase (WRN) comprising : evaluating cells of the cancer for somatic inactivating alterations in two or all of KMT2D, RAD50 and ARID1A; andif it is determined either: that somatic inactivating alterations are present in two or all of KMT2D, RAD50 and ARID1A in the cells of the cancer; or that somatic inactivating alterations are present in ARID1A in the cells of the cancer; determining that the cancer is likely to be sensitive to inhibitors of WRN.

4. The method as defined in claim 3, wherein: if it is determined that:ARID1A is wild type or has only missense mutations in the cells of the cancer; and one or both of KMT2D and RAD50 are wild type or have only missense mutations in the cells of the cancer; determining that the cancer is likely to be insensitive to inhibitors of WRN.

5. The method as defined in any one of claims 1-4 wherein evaluating the cells of the cancer for somatic alterations in two or all of KMT2D, RAD50 and ARID1A comprises evaluating the cells of the cancer for somatic alterations in all of KMT2D, RAD50 and ARID1A.

6. The method as defined in any one of claims 1-4, wherein evaluating the cells of the cancer for somatic alterations in two or all of KMT2D, RAD50 and ARID1A consists of evaluating the cells of the cancer for somatic alterations in all of KMT2D, RAD50 and ARID1A.

7. A method of determining whether a cancer of a mammalian subject is likely to be sensitive to combination therapy using inhibitors of WRN and immune checkpoint inhibitors (ICI), comprising: carrying out the method as defined in any one of claims 1 to 6 to determine if the cancer is likely to be sensitive to inhibitors of WRN; determining if the cancer is likely to be sensitive to immune checkpoint inhibitors; and if it is determined that both the cancer is likely to be sensitive to inhibitors of WRN; andthe cancer is likely to be sensitive to immune checkpoint inhibitors; determining that the cancer is likely to be sensitive to combination therapy using inhibitors of WRN and immune checkpoint inhibitors.

8. The method as defined in claim 7, wherein determining if the cancer is likely to be sensitive to immune checkpoint inhibitors comprises determining that a tumour mutation burden of the cancer is greater than or equal to 10 mutations / megabase.

9. A method of selecting a therapy for a cancer of a mammalian subject comprising: carrying out the method as defined in any one of claims 1 to 8 and, if it is determined that the cancer is likely to be sensitive to inhibitors of WRN, selecting an inhibitor of WRN as the therapy for the cancer.

10. The method as defined in claim 9, wherein, if it is further determined that the cancer is likely to be sensitive to immune checkpoint inhibitors, selecting combination therapy using an inhibitor of WRN and immune checkpoint inhibitors as the therapy for the cancer.

11. A method of identifying a mammalian subject as a candidate for therapy using an inhibitor of WRN comprising: carrying out the method as defined in any one of claims 1 to 8 on a cancer of the mammalian subject and, if it is determined that the cancer is likely to be sensitive to inhibitors of WRN, identifying the mammalian subject as a candidate for therapy using an inhibitor of WRN.

12. The method as defined in claim 11 wherein, if it is further determined that the cancer is likely to be sensitive to immune checkpoint inhibitors, identifying the mammalian subject as a candidate for combination therapy using an inhibitor of WRN and immune checkpoint inhibitors.

13. The method as defined in any one of claims 1 to 12, wherein the inhibitor of WRN comprises VVD-133214, HRO761, GSK4418959, or NDI-219216.

14. The method as defined in any one of claims 1, 3 or 5 to 13, further comprising administering the inhibitor of WRN to the mammalian subject.

15. The method as defined in any one of claims 7, 8, 10 or 12, wherein the immune checkpoint inhibitor inhibits one or more of: CTLA4, PD-1 or PD-L1; or wherein the immune checkpoint inhibitor is pembrolizumab, ipilimumab, tremelimumab, nivolumab, atezolizumab, avelumab, durvalumab, cemiplimab, dostarlimab, or relatlimab.

16. The method as defined in any one of claims 7, 8, 10, 12, 14 or 15, further comprising administering the immune checkpoint inhibitor to the mammalian subject.

17. The method as defined in any one of claims 1 to 16, wherein the step of evaluating the cells of the cancer for somatic inactivating alterations in two or all of KMT2D, RAD50 and ARID1A comprises evaluating a sample of the cancer by genetic sequencing and / or by DNA methylation sequencing and comparing a resultant sequence of two or all of KMT2D, RAD50 and ARID1A in the cells of the cancer to a corresponding sequence of KMT2D, RAD50 and ARID1A obtained from a sample of normal tissue.

18. The method as defined in claim 17, wherein the sample of the cancer comprises a tumour biopsy.

19. The method as defined in either one of claims 17 or 18, wherein the sample of normal tissue comprises a blood sample.

20. The method as defined in any one of claims 1 to 19, wherein the step of evaluating the cells of the cancer for somatic inactivating alterations in two or all of KMT2D, RAD50 or ARID1A comprises sequencing two or all of KMT2D, RAD50 and ARID1A in the cells of the cancer.

21. The method as defined in claim 20, wherein the sequencing comprises: shortread or long-read whole genome sequencing; short-read or long-read DNA methylation sequencing; short-read targeted genome sequencing of two or all of KMT2D, RAD50 or ARI DI A nanopore adaptive sampling-based long-read sequencing of two or all of KMT2D, RAD50 or ARID1A; or tiling PCR followed by Sanger sequencing of two or all of KMT2D, RAD50 or ARID1A.

22. The method as defined in any one of claims 1 to 21, wherein the cancer is a solid tumour.

23. The method as defined in claim 22, wherein the cancer is colorectal cancer, endometrial cancer, ovarian cancer, gastric cancer or esophageal cancer.

24. The method as defined in any one of claims 1 to 23, wherein the mammalian subject is a human.

25. The method as defined in any one of claims 1 to 6, 9, 11, 13 or 14, wherein the mammalian subject is resistant to treatment with immune checkpoint inhibitors (ICI).

26. A kit for predicting sensitivity of a cancer patient to treatment with inhibitors of Werner syndrome R.ecQ helicase (WR.N) comprising reagents for sequencing two or all of KMT2D, RAD50, and ARID1A, optionally wherein the kit comprises reagents for conducting nanopore adaptive sampling-based long-read sequencing of two or all of KMT2D, RAD50, and ARID1A, optionally wherein the kit comprises primers specific for two or all of KMT2D, RAD50, and ARID1A.

7. The kit as defined in claim 26, comprising instructions for using the reagents to sequence two or all of KMT2D, RAD50, and ARID1A in a sample of a cancer and in a sample of normal tissue from the patient to evaluate the presence of somatic inactivating alterations in two or all of KMT2D, RAD50, and ARID1A, the instructions further containing instructions to:determine if somatic inactivating alterations are present in two or all of KMT2D, RAD50 and ARID1A in the cells of the cancer; and if it is determined that somatic inactivating alterations are present in two or all of KMT2D, RAD50 and ARID1A in the cells of the cancer; or if it is determined that somatic inactivating alterations are present in ARID1A in the cells of the cancer; conclude that the cancer patient is likely to be sensitive to treatment with inhibitors of WRN.

28. The kit as defined in either one of claims 26 or 27, further comprising instructions to conclude that the cancer patient is unlikely to be sensitive to treatment with inhibitors of WRN if it is determined that:ARID1A is wild type or has only a missense mutation in the cells of the cancer; and one or both of KMT2D and RAD50 are wild type or have only a missense mutation in the cells of the cancer.

29. The kit as defined in any one of claims 26 to 28 wherein the reagents for sequencing two or all of KMT2D, RAD50, and ARID1A consist of reagents for sequencing only KMT2D, RAD50, and ARID1A.

Images