MRC Cancer Unit

The von Hippel-Lindau tumor suppressor (VHL) is a component of a ubiquitin ligase complex that controls cellular responses to hypoxia. Endogenous VHL is also utilized by proteolysis-targeting chimera (PROTAC) protein degraders, a promising class of anti-cancer agents. VHL is broadly essential for cell proliferation, yet it is a key tumor suppressor in renal cell carcinoma. To understand the functional consequences of VHL loss, and to identify targeted approaches for the elimination of VHL null cells, we have used genome-wide CRISPR-Cas9 screening in human renal epithelial cells. We find that, upon VHL loss, the HIF1A/ARNT complex is the central inhibitor of cellular fitness, suppressing mitochondrial respiration, and that VHL null cells show HIF1A-dependent molecular vulnerabilities that can be targeted pharmacologically. Combined VHL/HIF1A inactivation in breast and esophageal cancer cells can also provide resistance to ARV-771, a VHL-based bromodomain degrader that has anti-cancer activity. HIF1A stabilization can thus provide opportunities for early intervention in neoplastic VHL clones, and the VHL-HIF1A axis may be relevant for the development of resistance to the emerging class of PROTAC-based cancer therapies.

Myeloid-derived suppressor cells (MDSCs) are a heterogeneous immune population with diverse immunosuppressive functions in solid tumors. Here, we explored the role of the tumor microenvironment in regulating MDSC differentiation and immunosuppressive properties via signal-regulatory protein alpha (SIRPα)/CD47 signaling. In a murine melanoma model, we observed progressive increases in monocytic MDSCs and monocyte-derived dendritic cells that exhibited potent Tcell-suppressive capabilities. These adaptations could be recapitulated invitro by exposing hematopoietic stem cells to tumor-derived factors. Engagement of CD47 with SIRPα on myeloid cells reduced their phagocytic capability, enhanced expression of immune checkpoints, increased reactive oxygen species production, and suppressed Tcell proliferation. Perturbation of SIRPα signaling restored phagocytosis and antigen presentation by MDSCs, which was accompanied by renewed Tcell activity and delayed tumor growth in multiple solid cancers. These data highlight that therapeutically targeting myeloid functions in combination with immune checkpoint inhibitors could enhance anti-tumor immunity.

Cancer, a collection of more than two hundred different diseases, remains a leading cause of morbidity and mortality worldwide. Usually detected at the advanced stages of disease, metastatic cancer accounts for 90% of cancer-associated deaths. Therefore, the early detection of cancer, combined with current therapies, would have a significant impact on survival and treatment of various cancer types. Epigenetic changes such as DNA methylation are some of the early events underlying carcinogenesis. Here, we report on an interpretable machine learning model that can classify 13 cancer types as well as non-cancer tissue samples using only DNA methylome data, with 98.2% accuracy. We utilize the features identified by this model to develop EMethylNET, a robust model consisting of an XGBoost model that provides information to a deep neural network that can generalize to independent data sets. We also demonstrate that the methylation-associated genomic loci detected by the classifier are associated with genes, pathways and networks involved in cancer, providing insights into the epigenomic regulation of carcinogenesis.

The shared inherited genetic contribution to risk of different cancers is not fully known. In this study, we leverage results from 12 cancer genome-wide association studies (GWAS) to quantify pairwise genome-wide genetic correlations across cancers and identify novel cancer susceptibility loci. We collected GWAS summary statistics for 12 solid cancers based on 376 759 participants with cancer and 532 864 participants without cancer of European ancestry. The included cancer types were breast, colorectal, endometrial, esophageal, glioma, head and neck, lung, melanoma, ovarian, pancreatic, prostate, and renal cancers. We conducted cross-cancer GWAS and transcriptome-wide association studies to discover novel cancer susceptibility loci. Finally, we assessed the extent of variant-specific pleiotropy among cancers at known and newly identified cancer susceptibility loci. We observed widespread but modest genome-wide genetic correlations across cancers. In cross-cancer GWAS and transcriptome-wide association studies, we identified 15 novel cancer susceptibility loci. Additionally, we identified multiple variants at 77 distinct loci with strong evidence of being associated with at least 2 cancer types by testing for pleiotropy at known cancer susceptibility loci. Overall, these results suggest that some genetic risk variants are shared among cancers, though much of cancer heritability is cancer-specific and thus tissue-specific. The increase in statistical power associated with larger sample sizes in cross-disease analysis allows for the identification of novel susceptibility regions. Future studies incorporating data on multiple cancer types are likely to identify additional regions associated with the risk of multiple cancer types.

BackgroundCoffee contains many bioactive chemicals and associations with cancer have been reported in observational studies. In this Mendelian randomisation (MR) study we investigated the causal associations of coffee consumption with a broad range of cancers. Materials and methodsTwelve independent genetic variants proxied coffee consumption. Genetically-predicted risk of any cancer (59,647 cases) and 22 site-specific cancers was estimated in European-descent individuals in UK Biobank. Univariable and multivariable MR analyses were conducted. ResultsGenetically-predicted coffee consumption was not associated with risk of any cancer in the main analysis (OR 1.05, 95% CI 0.98–1.14, p = 0.183) but was associated with an increased risk of digestive system cancer (OR 1.28, 95% CI 1.09–1.51, p = 0.003), driven by a strong association with oesophageal cancer (OR 2.79, 95% CI 1.73–4.50, p = 2.5×10−5). This association was consistent after adjustment for genetically-predicted body mass index, smoking and alcohol consumption. There was no strong evidence supporting a causal relationship between genetically-predicted coffee consumption and the majority of cancers studied. However, genetically-predicted coffee consumption was associated with increased risk of multiple myeloma (OR 2.25, 95% CI 1.30–3.89, p = 0.004) and reduced ovarian cancer risk (OR 0.63, 95% CI 0.43–0.93, p = 0.020). ConclusionsThis MR study provides strong support for a causal association of coffee consumption with oesophageal cancer, but not for the majority of cancer types, and the underlying mechanisms require investigation.

BackgroundSomatic copy number alterations (SCNAs) are an important class of genomic alteration in cancer. They are frequently observed in cancer samples, with studies showing that, on average, SCNAs affect 34% of a cancer cell’s genome. Furthermore, SCNAs have been shown to be major drivers of tumour development and have been associated with response to therapy and prognosis. Large-scale cancer genome studies suggest that tumours are driven by somatic copy number alterations (SCNAs) or single-nucleotide variants (SNVs). Despite the frequency of SCNAs and their clinical relevance, the use of genomics assays in the clinic is biased towards targeted gene panels, which identify SNVs but provide limited scope to detect SCNAs throughout the genome. There is a need for a comparably low-cost and simple method for high-resolution SCNA profiling.ResultsWe present conliga, a fully probabilistic method that infers SCNA profiles from a low-cost, simple, and clinically-relevant assay (FAST-SeqS). When applied to 11 high-purity oesophageal adenocarcinoma samples, we obtain good agreement (Spearman’s rank correlation coefficient, rs=0.94) between conliga’s inferred SCNA profiles using FAST-SeqS data (approximately £14 per sample) and those inferred by ASCAT using high-coverage WGS (gold-standard). We find that conliga outperforms CNVkit (rs=0.89), also applied to FAST-SeqS data, and is comparable to QDNAseq (rs=0.96) applied to low-coverage WGS, which is approximately four-fold more expensive, more laborious and less clinically-relevant. By performing an in silico dilution series experiment, we find that conliga is particularly suited to detecting SCNAs in low tumour purity samples. At two million reads per sample, conliga is able to detect SCNAs in all nine samples at 3% tumour purity and as low as 0.5% purity in one sample. Crucially, we show that conliga’s hidden state information can be used to decide when a sample is abnormal or normal, whereas CNVkit and QDNAseq cannot provide this critical information.ConclusionsWe show that conliga provides high-resolution SCNA profiles using a convenient, low-cost assay. We believe conliga makes FAST-SeqS a more clinically valuable assay as well as a useful research tool, enabling inexpensive and fast copy number profiling of pre-malignant and cancer samples.

IntroductionColorectal cancer (CRC) occurring after a negative colonoscopy is called a post-colonoscopy colorectal cancer (PCCRC). Until recently, it has been difficult to compare global performance due to different methods used to define PCCRC rates. In 2018, the World Endoscopy Organization (WEO) standardised the methodology to calculate unadjusted PCCRC-3yr rates (PCCRC-3yr). The PCCRC-3yr is a “false-negative” rate of CRC diagnosed by colonoscopy.A recent systematic review investigated PCCRC-3yr rates but included studies outside the WEO’s methodology. Therefore, the true global prevalence of PCCRC as per the WEO methodology remains unknown.MethodsWe searched six databases for studies calculating the PCCRC-3yr rate using the WEO methodology. We performed a systematic review and meta-analysis to calculate the global PCCRC prevalence, change in prevalence over time and associated characteristics.ResultsOur search identified 4 studies reporting 13,791 cases of PCCRC. The pooled PCCRC prevalence is 7.5% (95% CI=6.6-8.4%) with high heterogeneity observed (I2 = 98%, p<0.001)(Figure 1).Compared to the combined baseline year category of 2011-12, the odds of PCCRC occurrence were significantly higher in 2008 – 2010 (OR 1.12 (95% CI =1.01–1.24), p=0.03) and 2004 – 2007 (OR 1.19, 95% CI =1.06-1.33, p=0.03) for three studies combined with high heterogeneity.Patients with inflammatory bowel diseases (IBD) had a pooled PCCRC-3yr rate of 29.3% (95% CI =21.3 – 38.1%), 6-times the odds than those without (OR 6.17, 95% CI =1.06-1.33) based on three studies combined with high heterogeneity. The pooled PCCRC-3yr rate in the right colon (two studies) was 8.6% (95% CI =8.3 – 8.8%), 50% higher than those with disease in the left colon (OR 1.51, 95% CI =1.41-1.61).ConclusionsThe global pooled PCCRC prevalence was 7.5%. Encouragingly, the results show a reduction in the PCCRC-3yr rate over time. We also identified that IBD patients have over 6 times higher odds than those without for developing a PCCRC. P169 Figure 1Forest plot of overall PCCRC-3yr rates for colonoscopies performed between 2004-2012

Background: Oesophageal adenocarcinoma (OAC) is a lethal cancer with an overall 5-year survival of <20%. Given the presence of a pre-invasive disease stage, also known as Barrett’s oesophagus (BO), and the availability of minimally invasive treatments for BO-related neoplasia, it is thought that early detection is the best strategy to improve patient outcomes. Clinical guidelines recommend endoscopic screening in patients with symptoms of acid reflux and additional risk factors. This strategy is flawed by the cost and invasiveness of endoscopy as well as by the fact that a significant proportion of OAC patients deny a history of reflux symptoms. Summary: New research on the use of epidemiologic and clinical data has allowed the creation of risk-prediction algorithms to identify the population at risk. In addition, newer less-invasive devices such as transnasal endoscopy, Cytosponge, volumetric laser endomicroscopy, and volatile organic compounds are emerging as promising options to allow screening in the primary care setting. Finally, there is an opportunity to intervene at the pre-invasive stage with pharmacological strategies to reduce the risk burden. Key Messages: In this review, we provide a critical appraisal of the different screening approaches and chemopreventive strategies and a guide to readers on how to implement research evidence in clinical practice.

With the rapid growth of the number of sequenced ancient genomes, there has been increasing interest in using this new information to study past and present adaptation. Such an additional temporal component has the promise of providing improved power for the estimation of natural selection. Over the last decade, statistical approaches for the detection and quantification of natural selection from ancient DNA (aDNA) data have been developed. However, most of the existing methods do not allow us to estimate the timing of natural selection along with its strength, which is key to understanding the evolution and persistence of organismal diversity. Additionally, most methods ignore the fact that natural populations are almost always structured, which can result in an overestimation of the effect of natural selection. To address these issues, we introduce a novel Bayesian framework for the inference of natural selection and gene migration from aDNA data with Markov chain Monte Carlo techniques, co-estimating both timing and strength of natural selection and gene migration. Such an advance enables us to infer drivers of natural selection and gene migration by correlating genetic evolution with potential causes such as the changes in the ecological context in which an organism has evolved. The performance of our procedure is evaluated through extensive simulations, with its utility shown with an application to ancient chicken samples.