Rethinking cancer therapies in the frame of cellular dormancy
Lead Research Organisation:
University College London
Department Name: Genetics Evolution and Environment
Abstract
Cellular plasticity, the ability to undergo molecular and physical changes, is a key adaptive mechanism that enables organisms to develop, regenerate and respond to stress. In cancer, however, it is one of the major underlying causes for resistance to therapy. This process allows cells to transition to distinct states and thereby gain the ability to evade immune responses, to accumulate new mutations which could enable resistance to antiproliferative drugs, colonise other organs or help metastatic tumour cells to adapt to new environments. All of these can lead to tumour progression and/or recurrence. Plasticity manifests in many forms, but one of particular clinical relevance is cellular dormancy. Dormant cells enter a reversible cell cycle arrest named quiescence which enables them to resist to high levels of stress such as those encountered during tumour development or treatment, e.g. chemotherapy targeting cycling cells. There is increasing evidence indicating that tumours contain subpopulations of slow-cycling or entirely quiescent cells which can evade therapy.
Dormancy results from a complex equilibrium between DNA damage and cell growth processes, involving master regulators like the p53 and Rb proteins, to enable the cells to cope with increasing genomic instability and levels of stress. These processes are crucial during early tumour development and manifest as a consequence of various mutagenic insults accumulated in the genome. Yet, the specific mutational processes inducing cells into a dormant phenotype are completely unexplored. Moreover, the extent to which dormancy levels vary within a single tumour or between multiple affected individuals is unknown. Addressing these knowledge gaps could lead to a paradigm shift into cancer treatment management.
A cancer with particular relevance for the phenomenon of dormancy is oesophageal adenocarcinoma, an aggressive disease with consistently poor outcomes from both chemotherapy and immunotherapy and a 5-year survival rate of only 15%. We have shown that chemotherapy does not change the mutational make-up of this cancer significantly, suggesting that something beyond the genome must be driving resistance to this therapy. Literature evidence and our own analyses indicate that dormancy could play a role. I propose to investigate the mechanisms leading to dormancy in this cancer and their potential therapeutic relevance. Beyond oesophageal cancer, various other tumour types show evidence of this phenomenon. Brain tumours, particularly gliomas, show consistently high rates of dormancy which confers stem-like characteristics and resistance to standard therapeutic strategies. Sarcomas are highly metastatic, often due to latent micrometastases enabled by dormant cells. Understanding the emergence and clinical impact of dormancy in various cancer tissues could help rationalise tailored therapies for cancer patients.
Statistical methods such as matrix decomposition have recently enabled us to identify the distinctive genomic footprints of various mutagens (such as smoking or UV light) from DNA sequencing data. These inform us on how specific cancers developed and can be linked to gene expression programmes active in dormancy and their manifestation in the tissue. In this regard, I will employ and develop cutting edge statistics, data integration and artificial intelligence methods on bulk, single cell and imaging datasets from oesophageal, brain and bone cancers to elucidate the genomic drivers, spatial heterogeneity and therapeutic relevance of dormancy in tumours.
The aim of this proposal is to provide an integrated view of how dormancy arises as a result of various mutational processes in cancer and how this impacts the broader cellular architecture of the tumour. I will employ this knowledge to redesign how cancer treatment is allocated by predicting the risk of resistance to therapies where dormancy may play a role.
Dormancy results from a complex equilibrium between DNA damage and cell growth processes, involving master regulators like the p53 and Rb proteins, to enable the cells to cope with increasing genomic instability and levels of stress. These processes are crucial during early tumour development and manifest as a consequence of various mutagenic insults accumulated in the genome. Yet, the specific mutational processes inducing cells into a dormant phenotype are completely unexplored. Moreover, the extent to which dormancy levels vary within a single tumour or between multiple affected individuals is unknown. Addressing these knowledge gaps could lead to a paradigm shift into cancer treatment management.
A cancer with particular relevance for the phenomenon of dormancy is oesophageal adenocarcinoma, an aggressive disease with consistently poor outcomes from both chemotherapy and immunotherapy and a 5-year survival rate of only 15%. We have shown that chemotherapy does not change the mutational make-up of this cancer significantly, suggesting that something beyond the genome must be driving resistance to this therapy. Literature evidence and our own analyses indicate that dormancy could play a role. I propose to investigate the mechanisms leading to dormancy in this cancer and their potential therapeutic relevance. Beyond oesophageal cancer, various other tumour types show evidence of this phenomenon. Brain tumours, particularly gliomas, show consistently high rates of dormancy which confers stem-like characteristics and resistance to standard therapeutic strategies. Sarcomas are highly metastatic, often due to latent micrometastases enabled by dormant cells. Understanding the emergence and clinical impact of dormancy in various cancer tissues could help rationalise tailored therapies for cancer patients.
Statistical methods such as matrix decomposition have recently enabled us to identify the distinctive genomic footprints of various mutagens (such as smoking or UV light) from DNA sequencing data. These inform us on how specific cancers developed and can be linked to gene expression programmes active in dormancy and their manifestation in the tissue. In this regard, I will employ and develop cutting edge statistics, data integration and artificial intelligence methods on bulk, single cell and imaging datasets from oesophageal, brain and bone cancers to elucidate the genomic drivers, spatial heterogeneity and therapeutic relevance of dormancy in tumours.
The aim of this proposal is to provide an integrated view of how dormancy arises as a result of various mutational processes in cancer and how this impacts the broader cellular architecture of the tumour. I will employ this knowledge to redesign how cancer treatment is allocated by predicting the risk of resistance to therapies where dormancy may play a role.
Planned Impact
The proposed research programme will enable scientific and clinical advancement for the management of both early stage cancers as well as progressive disease. Understanding response to treatment and predicting resistance to chemotherapy is one of the key unmet needs in the cancer field which this proposal aims to tackle through a tumour-dormancy angle. This strategy will lead to improvements in cancer diagnostic and treatment strategies with long-term impact on healthcare and the general public. The specific impacts are detailed below.
1. Academic benefit:
This research will generate new knowledge around the biology of dormancy in cancer, specifically:
(a) The mutational drivers of tumour dormancy and the extent of variation in dormancy signals among individuals or different types of cancer (see Objective 1 in Case for Support).
(b) The role of the immune system in tumour dormancy (see Objective 2 in Case for Support).
(c) The link between tumour dormancy and relapse/resistance to chemotherapy (see Objective 3 in Case for Support).
Beyond generating new knowledge into the basic biology of dormancy, this research will also generate novel tools for the research community: bioinformaticians, biostatisticians and data/AI scientists will benefit from open access to methods and code developed within the project. Specifically, this refers to artificial intelligence (AI) methods for image analysis and methods for data integration of bulk and single cell genomics/transcriptomics/proteomics. These methods are of wide interest in the biology field and could be utilised widely to study biological systems in a disease and non-disease context.
Furthermore, AI is an expanding field that is of high priority for the UK research portfolio, and the methods I will be developing will help progress the application of AI in healthcare and in particular in the cancer field.
Timescale: as soon as the first research outputs emerge, estimated within 1-2 years of the commencement of the project.
2. Healthcare benefit:
This research will develop an AI-based predictor of response to chemotherapy from digital pathology images of cancer tissue which are routinely collected in the clinic. This can be translated into a clinical tool that provides AI-assisted support for oncologists in deciding patient treatment. The creation of this technology will benefit cancer patients by tailoring their treatment to avoid the unnecessary side effects of chemotherapy for those individuals who are unlikely to benefit from this approach. As a consequence, the NHS will benefit long-term through improved healthcare strategies and cost reductions of cancer treatments. Moreover, policy makers and the national government could benefit from evidence provided by this research to argue for a wider integration of data-driven technologies in healthcare.
Timescale: from year 4.
3. Economic benefit:
This research is very likely to result in arising IP in the form of a novel chemotherapy predictor and in this regard, this project will have the potential to benefit the UK economy.
Timescale: from year 4.
1. Academic benefit:
This research will generate new knowledge around the biology of dormancy in cancer, specifically:
(a) The mutational drivers of tumour dormancy and the extent of variation in dormancy signals among individuals or different types of cancer (see Objective 1 in Case for Support).
(b) The role of the immune system in tumour dormancy (see Objective 2 in Case for Support).
(c) The link between tumour dormancy and relapse/resistance to chemotherapy (see Objective 3 in Case for Support).
Beyond generating new knowledge into the basic biology of dormancy, this research will also generate novel tools for the research community: bioinformaticians, biostatisticians and data/AI scientists will benefit from open access to methods and code developed within the project. Specifically, this refers to artificial intelligence (AI) methods for image analysis and methods for data integration of bulk and single cell genomics/transcriptomics/proteomics. These methods are of wide interest in the biology field and could be utilised widely to study biological systems in a disease and non-disease context.
Furthermore, AI is an expanding field that is of high priority for the UK research portfolio, and the methods I will be developing will help progress the application of AI in healthcare and in particular in the cancer field.
Timescale: as soon as the first research outputs emerge, estimated within 1-2 years of the commencement of the project.
2. Healthcare benefit:
This research will develop an AI-based predictor of response to chemotherapy from digital pathology images of cancer tissue which are routinely collected in the clinic. This can be translated into a clinical tool that provides AI-assisted support for oncologists in deciding patient treatment. The creation of this technology will benefit cancer patients by tailoring their treatment to avoid the unnecessary side effects of chemotherapy for those individuals who are unlikely to benefit from this approach. As a consequence, the NHS will benefit long-term through improved healthcare strategies and cost reductions of cancer treatments. Moreover, policy makers and the national government could benefit from evidence provided by this research to argue for a wider integration of data-driven technologies in healthcare.
Timescale: from year 4.
3. Economic benefit:
This research is very likely to result in arising IP in the form of a novel chemotherapy predictor and in this regard, this project will have the potential to benefit the UK economy.
Timescale: from year 4.
Organisations
- University College London (Fellow, Lead Research Organisation)
- University College London (Collaboration)
- QUEEN'S UNIVERSITY BELFAST (Collaboration)
- IMPERIAL COLLEGE LONDON (Collaboration)
- Memorial Sloan Kettering Cancer Center (Collaboration)
- University of Cologne (Collaboration)
- Cardiff University (Collaboration)
- UNIVERSITY OF GLASGOW (Collaboration)
- Institute of Cancer Research UK (Collaboration)
- UNIVERSITY OF BIRMINGHAM (Collaboration)
- Indian Institute of Science Bangalore (Collaboration)
- Genomics England (Collaboration)
- KING'S COLLEGE LONDON (Collaboration)
Publications
Abbas S
(2023)
Mutational signature dynamics shaping the evolution of oesophageal adenocarcinoma.
in Nature communications
Jacobson DH
(2023)
Multi-scale characterisation of homologous recombination deficiency in breast cancer.
in Genome medicine
Pan S
(2023)
HistoMIL: A Python package for training multiple instance learning models on histopathology slides
in iScience
Wiecek AJ
(2023)
Genomic hallmarks and therapeutic implications of G0 cell cycle arrest in cancer.
in Genome biology
Malagoli Tagliazucchi G
(2023)
Genomic and microenvironmental heterogeneity shaping epithelial-to-mesenchymal trajectories in cancer.
in Nature communications
Description | This award has enabled the development of new methods to study how cancer cells adapt to various forms of stress they are subjected to throughout their evolution. In such conditions, tumour cells will temporarily stop dividing, which often provides them with an advantage, as they are able to escape many of the standard cancer therapies targeting proliferating cells. We have developed a new signature that utilises transcriptional signals to identify these difficult to capture cells that are arrested in the G0 state of the cell cycle and that have been shown to enable resistance to a variety of therapies, leading to tumour persistence and cancer relapse. Our signature captures rapid and short-lived adaptation to drugs targeting cell cycle, tyrosine kinase signalling and epigenetic modulators, and could be used to track developing therapeutic resistance in the clinic in the longer-term. Importantly, our signature can be used as a robust baseline for studying various forms of cell cycle arrest that are clinically relevant including dormancy, quiescence and senescence. We have used this signature and machine learning approaches to identify genomic triggers and constraints of proliferation/G0 arrest decisions during cancer evolution and identify new potential targets that could be manipulated to keep tumour cells indefinitely arrested in the cell cycle or revert this process so they can be more easily killed by chemotherapeutics. Among these, we experimentally validated the centrosomal protein CEP89 as a novel modulator of proliferation/quiescence decisions. These findings have been published in Genome Biology (Wiecek et al, 2023): https://genomebiology.biomedcentral.com/articles/10.1186/s13059-023-02963-4 Objective 1: Characterising mutational processes driving cancer dormancy We have demonstrated that G0 arrest preferentially emerges in the context of more stable, less mutated genomes that maintain TP53 integrity, lack the hallmarks of DNA damage repair deficiency and present increased APOBEC mutagenesis (Wiecek et al, Genome Biol 2023). In the context of oesophageal adenocarcinoma (OAC), we find that APOBEC mutagenesis, also linked with dormancy, is a key process delineating the transition from Barrett Oesophagus to OAC. Our joint study with the Fitzgerald lab, published in Nature Communications (Abbas et al 2023, https://www.nature.com/articles/s41467-023-39957-6) suggests that the balance between APOBEC mutagenesis and SBS17, the most dominant mutational process in OAC, might help the cancer cells shift between a highly proliferative and a more adaptive slower growing (dormant) and therapy resistant state. Objective 2: Evaluating cell-to-cell heterogeneity and spatial architecture of dormant tumours To explore G0 arrest spatially within the tumour tissue, we have developed a robust and flexible AI framework applicable to digital pathology (HistoMIL; Pan and Secrier, iScience 2023). We use it to show that the expression of E2F targets, the main readout of cell proliferation/arrest decisions, is the cancer hallmark best captured in tissue slides stained with haematoxylin & eosin (H&E). We have also developed AI classifiers of G0 arrest demonstrating remarkable prediction accuracies between 83-86% in breast and lung cancers (Pan and Secrier, in prep). Analysis of the cell-cell interactions established within the dormant areas of the tumour shows that stromal barriers are increased around dormant cells, suggesting these cells are more successful in evading immune recognition (Pan et al, in prep). Objective 3: Developing a dormancy-based predictive model of resistance to chemotherapy Our G0 arrest signature can be used to detect rapid and short-lived adaptation to chemotherapy in a p53 wild type background, but also to kinase signalling and epigenetic modulators. Our studies suggest that resistance to treatment is likely to be an acquired phenomenon rather than a pre-existing one. Thus, predicting resistance is less straightforward than previously anticipated, especially when employing single sample bulk sequencing. However, our G0 arrest signature offers the opportunity to study this rapid and fleeting state of arrest linked with therapeutic resistance in a dynamic manner. Open questions remain as to the nature of this state (quiescent, senescent or dormant) as well as the TME mechanisms enabling it, which we plan to explore in the future. Some initial insights come from cancer-specific studies conducted with our collaborators. Firstly, together with Prof Adrienne Flanagan and Dr Lucia Cottone, we have shown that our G0 arrest signature is increased in osteosarcomas that are refractory to chemotherapy in the TARGET cohort (p=0.025). Using a cell lineage tracing system, Dr Cottone has demonstrated that cells surviving chemotherapy treatment enter a reversible quiescence state like the one described in our studies. We have just been awarded a CRUK City of London Development Fund (£25,000) to employ single cell sequencing to understand the signalling pathways leading to G0-enabled chemotherapy resistance in osteosarcoma in a multidisciplinary collaboration between the Secrier, Cottone and Efremova labs. We anticipate a first publication from this work within the next couple of years. Secondly, in collaboration with Prof Henning Walczak and Prof Scott Lowe, we have used a bespoke mouse model of lower grade glioma developed in the Walczak group that faithfully reproduces human cancers, and show that faster disease progression in this type of brain malignancy is enabled by intrinsic and TME cell regulation (Shroff et al, in prep). We will continue dissecting the role of dormancy in relation to immune inflammation in the following years. We plan to combine our recent work exploring cancer cell plasticity from an EMT perspective (Malagoli Tagliazucchi et al, Nat Commun 2023) with our G0 signatures to further characterise dormant cells in a variety of scenarios in oesophageal cancer, sarcoma and glioma. To date, this award has enabled the development of new computational methods for cell plasticity/fate evaluation as well as of integrative frameworks combining AI (multiple instance learning) for RNA-seq as well as H&E-stained slides, cell segmentation and graph approaches to explore cancer tissue structure and linked phenotypes of interest. My FLF funded research to date has produced five papers where I am senior corresponding author and has attracted additional funding of £13,000 as PI, £641,434 as Co-I and £90,000 as collaborator. Furthermore, this award has created 15 new collaborations to date (with UCL, Imperial College, Cardiff University, Queen Mary University of Belfast, Karolinska Institute, i3S Porto, Indian Institute of Science etc). |
Exploitation Route | The G0 arrest signature developed in this project could be employed to further study and track cancer cells as they rapidly adapt to and avoid various cancer therapies. Developing resistance to such therapies could be monitored in the clinic through liquid biopsies in the longer term, provided these persister cell signals can be captured with enough sensitivity from tumour DNA/RNA circulating in the blood. This could be employed both in a healthcare setting to monitor patients' responses, as well as in clinical trials led by pharmaceutical companies to understand resistance developed to their tested drugs. This signature can also serve to further study fundamental mechanisms of stress adaptation in cancer, including responses to various drugs, and several groups are already using our signature in their research to study dormancy and therapeutic resistance. All the methods and tools developed in this research are useful resources for the scientific community and are made freely available through our GitHub repositories: https://github.com/secrierlab/CancerG0Arrest, https://github.com/secrierlab/EMT, https://github.com/secrierlab/TumourHistologyDL, https://github.com/secrierlab/tumourMassDormancy, https://github.com/secrierlab/Mutational-Signatures-OAC, https://github.com/secrierlab/HistoMIL, https://github.com/secrierlab/SpottedPy. |
Sectors | Digital/Communication/Information Technologies (including Software) Healthcare Pharmaceuticals and Medical Biotechnology |
URL | https://genomebiology.biomedcentral.com/articles/10.1186/s13059-023-02963-4;https://www.sciencedirect.com/science/article/pii/S2589004223021508;https://www.nature.com/articles/s41467-023-39957-6 |
Description | To date, our findings and developed G0 arrest signature has attracted interest from leading groups in the field who are currently applying it in their own research in order to investigate the role of G0 arrest in resistence to chemotherapies and immunotherapies. These results, in addition to our other studies of cell plasticity and EMT are enabling the expansion of a broader area of studying cancer cell fate/plasticity with novel in silico tools and methods. |
First Year Of Impact | 2022 |
Sector | Other |
Description | Emerging Research Leader on the council of the Academy of Medical Sciences |
Geographic Reach | National |
Policy Influence Type | Participation in a guidance/advisory committee |
URL | https://acmedsci.ac.uk/ |
Description | Postgraduate training on cutting edge methodology to study tumour mutagenesis |
Geographic Reach | Europe |
Policy Influence Type | Influenced training of practitioners or researchers |
Description | CRUK Clinical Trial Award: "PROTIEUS - Proton Beam Therapy And Adjuvant Immunotherapy In Oesophageal Cancer" |
Amount | £616,434 (GBP) |
Organisation | University College London |
Sector | Academic/University |
Country | United Kingdom |
Start | 01/2023 |
End | 02/2028 |
Description | CRUK CoL Centre Development Fund ("Defining the metabolic landscape of immune cells in the tumor microenvironment"), Co-I with Patricia Barral (Francis Crick Institute) |
Amount | £25,000 (GBP) |
Organisation | Cancer Research UK |
Sector | Charity/Non Profit |
Country | United Kingdom |
Start | 05/2024 |
End | 05/2025 |
Description | CRUK CoL Centre Development Fund (Understanding how osteosarcoma cells become resistant to chemotherapy through 'quiescence'), Co-I with Lucia Cottone and Mirjana Efremova (QMUL) |
Amount | £25,000 (GBP) |
Organisation | Cancer Research UK |
Sector | Charity/Non Profit |
Country | United Kingdom |
Start | 01/2024 |
End | 12/2024 |
Description | EMBO workshop "The spatial biology of cancer", Co-PI with Francesca Ciccarelli and Matteo Cereda |
Amount | € 45,000 (EUR) |
Funding ID | w24/86 |
Organisation | European Molecular Biology Organisation |
Sector | Charity/Non Profit |
Country | Germany |
Start | 08/2024 |
End | 09/2024 |
Description | UCL Global Engagement Fund: "A global view of cancer cell plasticity during metastasis" |
Amount | £5,000 (GBP) |
Organisation | Indian Institute of Science Bangalore |
Sector | Academic/University |
Country | India |
Start | 12/2022 |
End | 07/2023 |
Description | UCL-IISc Joint Seed Fund: "Spatial and tumour microenvironment modelling of hybrid states during the epithelial-to-mesenchymal transition in cancer" |
Amount | £8,000 (GBP) |
Organisation | Indian Institute of Science Bangalore |
Sector | Academic/University |
Country | India |
Start | 12/2022 |
End | 07/2023 |
Title | A large language model to capture plastic cell states in cancer |
Description | We have developed a new large language AI model that allows us to capture plastic cell states in cancer. The method and application are currently being written up in a manuscript. |
Type Of Material | Computer model/algorithm |
Year Produced | 2023 |
Provided To Others? | No |
Impact | Method under development, we expect to publish it on GitHub and in a preprint within the next couple of months. |
Title | A robust transcriptomic signature of G0 arrest in cancer |
Description | We have developed novel methodology that allows us to characterise G0 arrest states such as quiescence, dormancy or senescence in a variety of cancer settings. We propose a new signature and computational approach to quantify G0 arrest in bulk, single cell and spatial transcriptomics data. We have extensively validated our method and signature in single cell datasets and cancer cell lines, and have demonstrated that it can reliably and robustly capture signals of G0 arrest both in bulk tissue as well as in single cells. This allowed us for the first time to comprehensively profile the landscape of G0 arrest across different cancer types and identify key genomic hallmarks and associated driver events. We highlight multiple genomic dependencies of this process that could be used to selectively target and kill G0 arrested cells, or to induce exit from G0 to overcome resistance to therapies that target cycling cells. Moreover, we demonstrate that our signature of dormancy can be employed in single cell data to track emerging resistance to a variety of therapies and highlight the main modalities that present increased G0 arrest-induced resistance, including cell cycle, kinase signalling and epigenetic inhibitors. Finally, we further optimise our signature to make it tractable to a variety of single cell or clinical applications. This signature and the corresponding methodology are available at: https://github.com/secrierlab/CancerCellQuiescence A preprint describing this methodology is available here: https://www.biorxiv.org/content/10.1101/2021.11.12.468410v3 |
Type Of Material | Data analysis technique |
Year Produced | 2021 |
Provided To Others? | Yes |
Impact | We provide a robust signature of G0 arrest that can be flexibly employed to study this phenomenon in bulk and single cell datasets, and make it freely available for the scientific community. We have also further optimised it to make it more feasible for measurements in the clinic. We intend to further develop this signature so that it could be used to track emerging therapy resistance through liquid biopsies (long-term goal that requires further funding). Using this methodology, we have already been able to identify several modulators of G0 arrest/proliferation decisions in cancer (see https://www.biorxiv.org/content/10.1101/2021.11.12.468410v3), including CEP89, a centrosomal gene whose suppression leads to an increase in quiescence according to our experimental validation. Such vulnerabilities could be targeted in the future in order to either maintain G0 arrest in tumours for longer periods, or to induce a cycling behaviour that is more effectively targeted by antiproliferative therapies. |
URL | https://github.com/secrierlab/CancerCellQuiescence |
Title | Framework for cell plasticity quantification in spatial transcriptomics datasets |
Description | We have developed a new framework for spatial transcriptomics data processing and analysis that allows us to identify cell states such as G0 arrest, dormancy or EMT in these slides and explore their interactions with other cells in the tumour microenvironment. This framework is currently being employed for several studies in the group and is due to be released by the end of 2023. |
Type Of Material | Data analysis technique |
Year Produced | 2022 |
Provided To Others? | No |
Impact | We have used this framework for the spatial transcriptomics analysis presented in Malagoli Tagliazucchi et al, Nat Commun 2023. |
Title | HistoMIL: fully automated deep learning framework for cancer histopathology slides |
Description | We have developed a fully automated pipeline for histopathology analysis in cancer, which consists of pre-processing and data augmentation of cancer whole-slide images, followed by training using a variety of deep learning models and testing in separate datasets. This has been finalised in a Python package entitled HistoMIL, which we have published in 2023. |
Type Of Material | Computer model/algorithm |
Year Produced | 2023 |
Provided To Others? | Yes |
Impact | We introduce HistoMIL, a Python package designed to streamline the implementation, training and inference process of MIL-based algorithms for computational pathologists and biomedical researchers. It integrates a self-supervised learning module for feature encoding, and a full pipeline encompassing TL and three MIL algorithms: ABMIL, DSMIL, and TransMIL. The PyTorch Lightning framework enables effortless customization and algorithm implementation. We illustrate HistoMIL's capabilities by building predictive models for 2,487 cancer hallmark genes on breast cancer histology slides, achieving AUROC performances of up to 85% and capturing cell cycle related processes with highest accuracy. These results have been published at iScience (Pan and Secrier, 2023): https://www.sciencedirect.com/science/article/pii/S2589004223021508 The package is freely available for use by the scientific community at: https://github.com/secrierlab/HistoMIL We hope this will serve as an accelerator of the research in the cancer digital pathology space. |
URL | https://github.com/secrierlab/HistoMIL |
Title | Neo4J graph database for cellular interactions inferred from histopathology data |
Description | We introduce a novel framework for surveying detailed cellular interaction landscapes within digital pathology slides by combining nuclear segmentation, classification and graph assembly, with efficient queries handled via a bespoke Neo4J graph database. We have stored cell-cell interactions inferred from histopathology slides in colon cancer that can be interactive explored and queried via our database, which is available for download here: https://github.com/secrierlab/TumourHistologyDL This work is presented in the following preprint: https://www.biorxiv.org/content/10.1101/2022.07.06.498984v1 |
Type Of Material | Database/Collection of data |
Year Produced | 2022 |
Provided To Others? | Yes |
Impact | We make use of the capabilities of the Neo4J graph database model to efficiently quantify and explore tissue landscapes and cellular interactions using histopathology data. This framework enables a faster, more extensive and more interpretable exploration of tumour cells in the context of their microenvironment, and could be easily adapted to answer a variety of biological questions in cancer as well as healthy tissue settings. Our investigation of the cellular organisation of the tumour tissue to date has highlighted niche structures, such as stromal barriers separating cancer cells from lymphocytes, which appear more frequently in highly quiescent/dormant tumours. This may have implications for immunotherapy treatment which we intend to investigate further using histopathology and spatial transcriptomics approaches. |
URL | https://github.com/secrierlab/TumourHistologyDL |
Title | Prognostic deep learning classifier of cellular dormancy in histopathology slides of cancer tissue |
Description | We employed matched RNA-seq and digital pathology data from the Cancer Genome Atlas (TCGA) to develop a prognostic tumour dormancy classifier in H&E-stained tissues. We employ the Inception v3 architecture for this purpose. We demonstrate that we can detect tumour dormancy in histopathological tissue slides with accuracies ranging between 80-82% in lung and colorectal cancers and up to 85% in breast cancer, which serves as a good working model for our intended work in oesophageal, brain and bone cancers. We show that the cellular contexture of dormancy is informative for response to chemotherapy. |
Type Of Material | Computer model/algorithm |
Year Produced | 2021 |
Provided To Others? | No |
Impact | We are currently preparing a manuscript on this work, and the code for the classifier has been released at our GitHub repository: https://github.com/secrierlab/TumourHistologyDL This framework will enable future studies into the impact of cellular quiescence in cancer and is freely available for use by the scientific community. Once applied to oesophageal cancer, we will further develop this classifier towards clinical use. |
URL | https://github.com/secrierlab/TumourHistologyDL |
Title | Quantification of epithelial-to-mesenchymal (EMT) state of tumour samples |
Description | This study has created new methodology and tools for the study of EMT in cancer. We have developed a computational framework for inferring EMT trajectories in bulk and single cell datasets and applied this methodology pan-cancer to identify three major EMT states: epithelial, hybrid E/M and mesenchymal. Briefly, we map cancer transcriptomics data from individual samples on single cell trajectories of EMT transformation using PCA and k-nearest neighbour approaches. This enables us to evaluate the metastatic potential of individual tumours for which bulk (mixed cell), single cell or spatial transcriptomic sequencing is available. Using this method, we were able to characterise a continuum of EMT states underlying metastatic transformation and link them with genetic changes. We describe a hybrid EMT state with poorer outcomes, which we believe could help enable dormancy and long term clinical relapse and wish to further study in relation to dormancy. The framework we have developed for assessing EMT trajectories has been made freely available to the research community here: https://github.com/secrierlab/EMT |
Type Of Material | Data analysis technique |
Year Produced | 2021 |
Provided To Others? | Yes |
Impact | Through the methodology developed within this project, we are for the first time able to capture a continuum of EMT transformation states within bulk primary tumours, single cells and entire tissues with transcriptomic data (spatial transcriptomics). This allows us to understand how cells move from an early, non-transformed state to a later, malignant state of cancer. This approach also allows us to link these states with genomic changes that may enable EMT and metastasis. In fact, in our study we have been able to uncover several new potential drivers of EMT and metastasis, some of which have been validated through siRNA screens. This study has been published in Nature Communications (Malagoli Tagliazucchi et al, 2023) and the code has been made available to the research community via our GitHub repository: https://github.com/secrierlab/EMT We are also working on a stand-alone R package implementing this methodology. Furthermore, this framework has opened up a new PhD project funded by HDRUK where we explore how EMT shapes the tissue architecture of tumours and cancer cell-microenvironment interactions in histopathology slides using AI (deep learning) and graph interaction methodologies developed in parallel in the lab. |
URL | https://github.com/secrierlab/EMT |
Title | SpottedPy: identification of tumour hotspots and TME relationships in spatial transcriptomics slides |
Description | We introduce a new tool for spatial transcriptomics data analyses called SpottedPy, which allows the identification of tumour hotspots with enriched signatures of interest and the exploration of their proximity to immune and stromal-rich areas within the slide. The implementation is based on the Getis Ord Gi* geospatial statistic. |
Type Of Material | Data analysis technique |
Year Produced | 2023 |
Provided To Others? | Yes |
Impact | We have used SpottedPy to explore EMT transformation and G0 arrest phenotypes in spatial transcriptomics cancer slides. The tool is publicly available at: https://github.com/secrierlab/SpottedPy, with the accompanying manuscript currently under review and released in preprint form: https://www.biorxiv.org/content/10.1101/2023.12.20.572627v1 We are preparing a 2nd manuscript utilising this method at the moment, and have also applied it to analyse data in one of our collaborations. |
URL | https://github.com/secrierlab/SpottedPy |
Description | Collaboration on deep learning approaches for cancer digital pathology data |
Organisation | University of Glasgow |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | The expertise of the postdoc employed on this grant has enabled us to expand beyond exploring genomic alterations linked with cancer progression and has allowed us to integrate imaging data from cancer histopathology tissue (H&E stained slides). As a result of the methods developed to integrate genomics and imaging data, I have set up a collaboration with Dr Ke Yuan at the University of Glasgow, where we are exploring supervised (us) and unsupervised (them) approaches to detecting patterns of G0 arrest and metastasis within cancer tissue images. |
Collaborator Contribution | My collaborator has provided new unsupervised methodology and data analysis which we have integrated with our approaches to gain further insights into the cellular context and spatial distribution of transformed cells. |
Impact | Our analyses have yielded new insights on how G0 arrested, EMT transformed cells interact with their TME which can be captured in digital pathology slides. We have used these insights to test more specific hypotheses and build new grant applications. We will scale up our analyses in the upcoming year to allow us to produce a manuscript for publication. |
Start Year | 2020 |
Description | Collaboration on new mouse models of IDH mutant low grade gliomas |
Organisation | Memorial Sloan Kettering Cancer Center |
Country | United States |
Sector | Academic/University |
PI Contribution | This collaboration entails a joint effort between our lab and those of Prof Henning Walczak (UCL Cancer Institute and the University of Cologne) and Prof Scott Lowe (MSKCC). Prof Walczak's lab has developed a new mouse model for low grade glioma which he wanted to characterise using our approaches. We have contributed methods and data analysis for their sequenced tumour datasets, and have driven the research focused on understanding linked tumour plasticity and TME phenotypes in this model of glioma. |
Collaborator Contribution | The collaborators provided expression, methylation and genomic data from >60 mouse tumours developed through their innovative model. |
Impact | In this collaboration, we have helped demonstrate that the mouse model developed mimics human cancers with high fidelity and have obtain important insights into the biology and microenvironment of low grade gliomas. Importantly, we have uncovered two subtypes of LGG with distinct prognosis, one of which is driven by a G0 arrest phenotype. We are about to submit a manuscript describing these findings. This collaboration has also helped us build new ideas which have gone into my UKRI Future Leaders Fellowship. |
Start Year | 2020 |
Description | Collaboration on new mouse models of IDH mutant low grade gliomas |
Organisation | University College London |
Department | UCL Cancer Institute |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | This collaboration entails a joint effort between our lab and those of Prof Henning Walczak (UCL Cancer Institute and the University of Cologne) and Prof Scott Lowe (MSKCC). Prof Walczak's lab has developed a new mouse model for low grade glioma which he wanted to characterise using our approaches. We have contributed methods and data analysis for their sequenced tumour datasets, and have driven the research focused on understanding linked tumour plasticity and TME phenotypes in this model of glioma. |
Collaborator Contribution | The collaborators provided expression, methylation and genomic data from >60 mouse tumours developed through their innovative model. |
Impact | In this collaboration, we have helped demonstrate that the mouse model developed mimics human cancers with high fidelity and have obtain important insights into the biology and microenvironment of low grade gliomas. Importantly, we have uncovered two subtypes of LGG with distinct prognosis, one of which is driven by a G0 arrest phenotype. We are about to submit a manuscript describing these findings. This collaboration has also helped us build new ideas which have gone into my UKRI Future Leaders Fellowship. |
Start Year | 2020 |
Description | Collaboration on new mouse models of IDH mutant low grade gliomas |
Organisation | University of Cologne |
Country | Germany |
Sector | Academic/University |
PI Contribution | This collaboration entails a joint effort between our lab and those of Prof Henning Walczak (UCL Cancer Institute and the University of Cologne) and Prof Scott Lowe (MSKCC). Prof Walczak's lab has developed a new mouse model for low grade glioma which he wanted to characterise using our approaches. We have contributed methods and data analysis for their sequenced tumour datasets, and have driven the research focused on understanding linked tumour plasticity and TME phenotypes in this model of glioma. |
Collaborator Contribution | The collaborators provided expression, methylation and genomic data from >60 mouse tumours developed through their innovative model. |
Impact | In this collaboration, we have helped demonstrate that the mouse model developed mimics human cancers with high fidelity and have obtain important insights into the biology and microenvironment of low grade gliomas. Importantly, we have uncovered two subtypes of LGG with distinct prognosis, one of which is driven by a G0 arrest phenotype. We are about to submit a manuscript describing these findings. This collaboration has also helped us build new ideas which have gone into my UKRI Future Leaders Fellowship. |
Start Year | 2020 |
Description | Collaboration on positive selection in cancer |
Organisation | Institute of Cancer Research UK |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | The work funded by this fellowship has enabled me to establish a new collaboration with Dr Luis Zapata at the ICR in London. We have jointly supervised an MSc student on positive selection in the context of cancer dormancy/quiescence and are currently further collaborating on linked projects. My lab has provided expertise in evaluating cancer cell quiescence,and training in cancer genomics and data analysis for the joint student. |
Collaborator Contribution | Dr Luis Zapata has provided methodology and expertise for investigating positive and negative selection during cancer evolution, as well as useful discussions and feedback. |
Impact | This is a multidisciplinary collaboration: mathematics/statistics, bioinformatics We have identified two novel genes that are positively selected in the context of tumour dormancy/quiescence, indicating specific evolutionary forces towards enabling this phenotype. We are currently writing up these results into a publication. Update 2023: A separate analysis of immune evasion employing the same methods and showcasing our approach that we are currently refining for dormancy is already out as a preprint since June 2022 and currently under review: https://www.biorxiv.org/content/10.1101/2022.06.20.496910v1 Update 2024: Our manuscript is in revision at Genome Biology. We have in the meantime initiated another collaboration looking at mutational processes driving response/resistance to immunotherapy in oesophageal adenocarcinoma, and co-supervised a PhD rotation student on this topic last term. A manuscript is planned from this second project. |
Start Year | 2021 |
Description | Collaboration on the role of p53 mutations in cancer cell quiescence |
Organisation | University of Birmingham |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | This grant has enabled me to set up a new collaboration with Prof Gareth Bond at the University of Birmingham to investigate the role of p53 mutations in the context of cancer cell quiescence. Our contribution has been the methodology to evaluate quiescence in large scale sequencing datasets. |
Collaborator Contribution | The Bond group have contributed expertise in evaluating germline and somatic alterations to link different p53 mutations with tumour quiescence levels. |
Impact | This collaboration has uncovered that p53 functionality is a pervasive, but not obligatory dependency of cancer cell quiescence. These findings open up new directions in exploring what drives quiescence/dormancy in p53 mutant cancers, which is currently very poorly understood. The results of this collaboration have been published in Wiecek et al, Genome Biol 2023: https://pubmed.ncbi.nlm.nih.gov/37221612/ This collaboration was multidisciplinary: genetics, bioinformatics |
Start Year | 2021 |
Description | Collaboration on the role of proton therapy in modulating the tumour and liver microenviroment to facilitate response to IO |
Organisation | University College London |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | As a result on the methods developed to evaluate G0 arrest within this grant, I was asked to join a collaboration led by Prof Maria Hawkins at UCL on the role of proton therapy in combination with IO in the context of liver metastases in oesophageal cancer. My lab will contribute our methods on quantifying G0 arrest from RNA-seq and histopathology data to investigate the link between dormancy and responses to combined radiation and immunotherapy treatment. |
Collaborator Contribution | The collaborators will provide whole-exome sequencing, RNA-seq and histopathology images from tumours before and after proton and IO treatment. |
Impact | This collaboration is multi-disciplinary, with clinicians, physicists, bioinformaticians and pathologists participating in the project. Update 2023: Our collaborative grant proposal "PROTIEUS: Proton Beam Therapy and Adjuvant Immunotherapy in Oesophageal Cancer" has been funded by a CRUK Clinical Trial Award (£616,434). The trial will commence recruitment in August 2023 and we expect to generate rich genomics, RNA-seq and imaging datasets to explore within the next 4 years. We will also be seeking further funding to support some of the analyses proposed. Update 2024: We have generated some pilot in situ data that explores the effects of Proton Beam Therapy in liver metastases at 0, 3, 6 and 24 hours from radiation using spatial transcriptomics. We are applying our newly developed tool, SpottedPy, to identify therapy resistant and susceptible areas, and intend to use this pilot data for planned grant applications this year. We also plan to publish a manuscript on these data by the end of the year. |
Start Year | 2021 |
Description | Collaboration on various aspects of G0 arrest in cancer |
Organisation | Imperial College London |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | This grant has funded research in collaboration with Dr Alexis Barr, leader of the Cell Cycle Control cell biology group at Imperial College, to explore various aspects of cancer cell quiescence. In particular, we have collaborated with Dr Barr so far to experimentally validate and optimise a transcriptomic signature of quiescence using cancer cell lines, and to validate several quiescence modulators that had arisen from our analysis using siRNA screens. Furthermore, we are also collaborating on a distinct project where we explore secreted proteins that promote quiescence. We provide the in silico validation and help with prioritisation of candidate proteins. |
Collaborator Contribution | The Barr lab has performed microscopy experiments to quantify quiescence fractions in cancer cell lines and siRNA screens to validate some of our top targets linked with quiescence. |
Impact | This collaboration has resulted in a new, highly robust transcriptomic signature of G0 arrest, which can be employed to track resistance to a variety of cancer therapies. We have also validated a new modulator of quiescence, CEP89. These findings are now published in Genome Biology (Wiecek et al, 2023): https://pubmed.ncbi.nlm.nih.gov/37221612/ We have also contributed to two manuscripts from the Barr lab looking at the role of G0 arrest in lung cancer and the secretome that may enable G0 arrest. These manuscripts are due to be submitted within the next couple of months. This collaboration is multidisciplinary: cell biology, microscopy, functional genomics, bioinformatics |
Start Year | 2019 |
Description | Cross-species programmes of cellular dormancy |
Organisation | University College London |
Department | Biosciences |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | As a result of our work on cellular dormancy and quiescence in primary tumours, we have established a new collaboration with Prof Jurg Bahler at the Institute of Healthy Ageing at UCL. We will use the RNA-seq data generated by the Bahler lab in yeast, killifish and C.elegans to study gene expression programmes driving dormancy in these organisms. To date, we have integrated these datasets and compared them to our in house signature of G0 arrest in human cancers (see preprint by Wiecek et al, bioRxiv 2021). |
Collaborator Contribution | The partner is contributing RNA-seq datasets of dormant and non-dormant individuals profiled from multiple species. |
Impact | We have refined our signature of G0 arrest following this integration exercise, helping us improve our dormant cell identification in cancer and are currently preparing a manuscript to describe the outcomes of this research. This is a multidisciplinary collaboration involving genetics and bioinformatics approaches. |
Start Year | 2022 |
Description | Defining the metabolic landscape of immune cells in the tumour microenvironment |
Organisation | King's College London |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | This is a new collaboration with Patricia Barral (King's College London) which aims to investigate the metabolic regulation of TAMs in the tumour microenvironment. We find a strong association between these cells and G0 arrested, hybrid E/M cells with increased invasive potential in our studies, and will be exploring how changes in TAMs might impact them. We will be contributing data analysis/bioinformatics expertise to this project. |
Collaborator Contribution | All wet lab experiments and single cell RNA-seq data. |
Impact | This is a multidisciplinary collaboration involving immunologists and bioinformatics/AI experts. We have been recently awarded a CRUCK CoL Development Fund (£25,000) to kick-start this collaboration and are currently waiting for the experiments to be performed. |
Start Year | 2024 |
Description | Dormancy-enabled immune evasion in pre-malignant disease |
Organisation | Cardiff University |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | Following the attendance of an MRC Cancer Sandpit in January 2023, I have formed a new collaboration with Dr Catherine Hogan (Cardiff University) aiming to investigate the role of dormancy in immune evasion of cancer precursor cells. We will do this by analysing spatial transcriptomics data in mouse healthy tissue, where we will identify PDAC precursor cells. We are in the early stages of planning the collaboration and seeking funding. My group will provide all the computational expertise and analyses for the project, including a novel method recently developed for spatial transcriptomics data (SpottedPy). |
Collaborator Contribution | The collaborators will provide spatial transcriptomics data and experimental validation of the findings. |
Impact | We are currently awaiting decisions on grant applications to the BBSRC and CRUK. This project is highly interdisciplinary, involving cell biology, clinical aspects and computational biology/bioinformatics. |
Start Year | 2023 |
Description | Dormancy-enabled immune evasion in pre-malignant disease |
Organisation | Queen's University Belfast |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | Following the attendance of an MRC Cancer Sandpit in January 2023, I have formed a new collaboration with Dr Catherine Hogan (Cardiff University) aiming to investigate the role of dormancy in immune evasion of cancer precursor cells. We will do this by analysing spatial transcriptomics data in mouse healthy tissue, where we will identify PDAC precursor cells. We are in the early stages of planning the collaboration and seeking funding. My group will provide all the computational expertise and analyses for the project, including a novel method recently developed for spatial transcriptomics data (SpottedPy). |
Collaborator Contribution | The collaborators will provide spatial transcriptomics data and experimental validation of the findings. |
Impact | We are currently awaiting decisions on grant applications to the BBSRC and CRUK. This project is highly interdisciplinary, involving cell biology, clinical aspects and computational biology/bioinformatics. |
Start Year | 2023 |
Description | Exploring immune evasion phenotypes in melanoma |
Organisation | University College London |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | Following our successful establishment of methods to explore tumour microenvironment interactions and immune evasion, we have started to collaborate with Prof Ariberto Fassati at UCL. Ariberto has a bespoke mouse model of transplantable melanoma which shows remarkable immune evasion, and we will be employing our data analysis and machine learning expertise to explore these signatures of immune evasion in bulk and single cell data from human cancers. |
Collaborator Contribution | Provision of transcriptomics signatures linked with immune evasion from the transplantable melanoma model and linked expertise. |
Impact | This is a multi-disciplinary collaboration, involving immunologists and bioinformatics/AI experts. We have just started to collaborate on this project and at the moment are exploring the clinical relevance of the immune evasion signatures in bulk human cancer datasets. |
Start Year | 2023 |
Description | Investigating the role of G0 arrest in tumour immune evasion |
Organisation | University College London |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | I have initiated a new collaboration with Dr Kevin Litchfield at the UCL Cancer Institute where we are employing our G0 arrest signature to investigate immune evasion in the context of immunotherapy. We are providing the G0 arrest signature to the collaboration. |
Collaborator Contribution | The partner provides a carefully curated database on immunotherapy clinical trial outcomes and RNA-seq data, as well as organoid models to study this phenomenon. |
Impact | We have obtained encouraging results linking our G0 arrest signature to immunotherapy outcomes, and this is currently being investigated further. We plan to use an improved signature capturing G0 arrest in tumour cells and programmes from the microenvironment to develop a better predictor of checkpoint blockade outcomes. A manuscript is in preparation from this work. |
Start Year | 2022 |
Description | Mechanisms of aberrant cell elimination to maintain tissue homeostasis |
Organisation | Cardiff University |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | While attending an MRC Sandpit on "Technological innovation for understanding cancers of unmet need" in January 2023, I have met Catherine Hogan who has been studying dormancy in pancreatic cancer precursors and is interested in how dormancy might shape tissue homeostasis preceding cancer. We have established a new collaboration to investigate this jointly using our newly developed methods to explore malignant transformation (dormancy/EMT) in spatial transcriptomics data (see our recent publication Malagoli Tagliazzuchi et al, Nature Commun 2023). We will characterise the transcriptomic changes in tissue homeostasis spatially in murine pancreas tissue available in the Hogan lab to investigate what happens when we manipulate mutant cell elimination capability, which we believe may be hijacked through dormancy, and couple these with specific signalling pathways. I am listed as a collaborator for a BBSRC Research Grant that Catherine Hogan and Fisun Hamaratoglu at Cardiff University have submitted at the end of January. |
Collaborator Contribution | The Hogan and Hamaratoglu labs will provide genomics and spatial transcriptomics data from murine pancreas models of tissue homeostasis for us to analyse, which will enable us to investigate the role of malignant cell elimination in tissue homeostasis. |
Impact | BBSRC Research Grant application submitted at the end of January, with me listed as collaborator. This is a multidisciplinary collaboration, with the following disciplines involved: cell biology, developmental biology, computational biology/bioinformatics. Update 2024: This application has not been successful, but a rewritten version is currently again under consideration with BBSRC. |
Start Year | 2023 |
Description | Participation in the Genomics England (100,000 Genomes) Sarcoma GeCIP |
Organisation | Genomics England |
Country | United Kingdom |
Sector | Public |
PI Contribution | I have been invited by Prof Adrienne Flanagan to be part of the Sarcoma Working Group of the 100,000 Genomes Project (Genomics England). This provides me access to whole-genome sequecing data and RNA-seq data for ~2000 sarcoma tumours as part of this consortium. I have an ongoing collaboration with Prof Flanagan on investigating aspects of tumour dormancy and metastasis in sarcoma, which has been funded by my UKRI Future Leaders Fellowship. Our contribution is methodology for scoring tumour progression and dormancy in bulk sequencing datasets based on transcriptomics signatures, and the entire analysis of the datasets provided. |
Collaborator Contribution | The collaborators provide whole-genome sequencing data for ~2000 tumours and RNA-seq for ~1000 tumours. |
Impact | This is a multidisciplinary collaboration, involving clinicians, geneticists and computational biologists. I regularly take part in the Sarcom GeCIP meetings and collaborate with several members. In particular, participation in this consortium has helped me obtain a UKRI Future Leader Fellowship to investigate tumour dormancy in sarcoma (among other cancers). Update 2024: A joint postdoc has just started to work on a collaboration between the Flanagan, Pillay and Secrier labs aiming to explore chromosomal instability in digital pathology, and AI approaches to enable this. We anticipate to obtain results within the next year. |
Start Year | 2019 |
Description | Synergistic systems approaches to investigating cellular plasticity in cancer |
Organisation | Indian Institute of Science Bangalore |
Country | India |
Sector | Academic/University |
PI Contribution | I have initiated a new collaboration with Dr Mohit Kumar Jolly (IISc, India) where we would like to explore synergies in our expertise in order to develop new integrated methods to study cellular plasticity in cancer. We have been successful in obtaining two seed awards (UCL-IISc Joint Seed Fund and UCL Global Engagement Fund) to organise a symposium entitled "Systems approaches towards cancer cell plasticity" and enable a collaborative visit between our labs. Through this award, we wish to create momentum for the development of new tools integrating cutting-edge spatial technologies with mathematical modelling, and build a community of experts contributing to this endeavour. This will foster new opportunities for collaborative grant applications. Our group contributes expertise in genomics and cell fate/plasticity quantification from bulk, single cell and spatial transcriptomics data. |
Collaborator Contribution | The Jolly group provides expertise in mathematical modelling of gene regulatory circuits. |
Impact | Two seed awards totalling £13000 to be used for the organisation of a symposium on cancer cell plasticity, a collaborative visit of Dr Jolly at UCL and the writing of a review article and joint grant proposal. The collaboration is multidisciplinary, involving mathematics, modelling, genomics, bioinformatics. Update 2024: In March 2023 we have organised a symposium entitled "Systems approaches towards cancer cell plasticity" which attracted high interest from the community (>100 participants) and fostered new collaborations. We are currently writing up a review on the topic, which we expect to submit by the end of the year, and this will help set the groundwork for joint grant applications. |
Start Year | 2022 |
Description | The role of metabolic heterogeneity in cancer cell plasticity |
Organisation | Imperial College London |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | I have established a collaboration with Dr Stefan Antonowicz at Imperial College that aims to investigate the role of metabolic heterogeneity in cancer in an adjuvant setting in oesophageal adenocarcinoma. We will employ spatial transcriptomics methods we have developed in the group to identify cellular plasticity states such as cell cycle arrest or EMT (see Malagoli Tagliazucchi, Nat Commun 2023) to explore their link with metabolic regulation in this context. |
Collaborator Contribution | The partner will provide spatial transcriptomics data of cancer tissue profiled before and after chemotherapy. |
Impact | I am a collaborator on a Rosetrees Trust Major Project Grant application led by Dr Antonowicz which has been submitted at the end of January. This is a multidisciplinary collaboration involving clinical and bioinformatics expertise. Update 2024: The grant application has been successful and we are currently waiting for data to be generated so we can proceed with the analyses. |
Start Year | 2023 |
Title | Signature of quiescence-linked therapeutic resistance |
Description | We have developed a robust transcriptomic signature of cancer cell quiescence which can be used to track resistance to multiple targeted therapies, including kinase signalling inhibitors and epigenetic regulators, as well as chemotherapy in specific settings. We plan to further test its use in tracing developing resistance, as well as predicting it. |
Type | Support Tool - For Fundamental Research |
Current Stage Of Development | Initial development |
Year Development Stage Completed | 2023 |
Development Status | Under active development/distribution |
Impact | In the long term, we hope this signature can be used to inform therapeutic decisions in the clinic. |
URL | https://genomebiology.biomedcentral.com/articles/10.1186/s13059-023-02963-4 |
Description | Invited speaker at the "Cancer Genomics ONLINE - Webinar 3: Unravelling the genetics of the tumour microenvironment" |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | Online event organised by Front Line Genomics, aimed to bring together experts using genetics approaches to understand the tumour microenvironment. I gave a talk on "Genomic and digital pathology approaches to elucidate the tumour microenvironment of plastic cells". |
Year(s) Of Engagement Activity | 2022 |
URL | https://frontlinegenomics.com/cancer-genomics-online-october-2022/ |
Description | Invited speaker at the EMBL-EBI workshop: Targeting non-oncogene addiction for cancer therapies |
Form Of Engagement Activity | Participation in an activity, workshop or similar |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Industry/Business |
Results and Impact | The purpose of this industry organised workshop in collaboration with EBI was to explore opportunities in targeting non-oncogene addictions, which could result in effective cancer therapies with acceptable therapeutic window. Invited speakers from academia and industry were joined in a forum to initiate such discussions, with focus on targeting cellular stress and the tumour microenvironment, as well as bioinformatics resources and computational tools for targeting non-oncogenic addictions. I have presented our current insights into EMT and quiescence-linked cellular vulnerabilities that could be targeted to slow down tumour growth or to render cells sensitive to chemotherapy. The anticipated workshop outputs included the following: • Sharing knowledge and experiences on successful stories and lessons learned • Novel technologies to measure / screen for non-oncogene addiction • Novel approaches for assessing whether any protein plays a role in various forms of non-oncogenic addiction • Provide opportunities for networking and collaborations between industry members and academia As a result of this workshop, I was contacted by Bristol Myers Squibb regarding a potential collaboration, which we intend to pursue through a joint application for a UCL Case PhD studentship (academia-industry studentship). |
Year(s) Of Engagement Activity | 2021 |
Description | Invited speaker at the ISCB Webinar series |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | A webinar series organised by ISCB on multiple computational biology/bioinformatics topics. I gave a talk on "Reconstructing the mutational histories of healthy and cancer genomes". |
Year(s) Of Engagement Activity | 2022 |
URL | https://www.youtube.com/watch?v=FCgrojRr0RE&ab_channel=ISCB |
Description | Invited talk at the "ELRIG Research and Innovation Conference" |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | National |
Primary Audience | Industry/Business |
Results and Impact | Topic: "INNOVATIONS FOR TOMORROW'S DRUG DISCOVERY" This is a conference organised by pharmaceutical companies and biotech. I gave a talk on "Genomic and digital approaches to elucidate cancer cell quiescence", which was well received and sparked a lot of questions and discussions from the public (~150 participants). |
Year(s) Of Engagement Activity | 2022 |
URL | https://www.elrig.org/portfolio/research-innovation-2022/ |
Description | Invited talk at the 3rd International IBSE Symposium |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Other audiences |
Results and Impact | Topic of the symposium: CLINICAL GENOMICS TO SYSTEMS MEDICINE: COMPUTATIONAL APPROACHES FOR TRANSFORMING HEALTHCARE I gave a talk on "Genomics and digital pathology approaches to elucidate cancer dormancy". |
Year(s) Of Engagement Activity | 2022 |
URL | https://ibse.iitm.ac.in/symposium-03/ |
Description | Invited talk at the AMS Springboard Alumni Event |
Form Of Engagement Activity | Participation in an activity, workshop or similar |
Part Of Official Scheme? | No |
Geographic Reach | National |
Primary Audience | Professional Practitioners |
Results and Impact | I gave a talk at the AMS Springboard Alumni event about my PI journey since the Springboard award. This generated plenty of discussion and questions, and I am grateful for the opportunity to share my experience and provide advice for the new Springboard awardees. |
Year(s) Of Engagement Activity | 2022 |
Description | Invited talk at the Tri-Omics Summit |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | This summit organised in London in a hybrid format aimed to bring together experts in multi-omics technologies. I was a speaker in the Cancer Genomics session and discussed "Genomics and digital pathology approaches to elucidate cellular quiescence in cancer". |
Year(s) Of Engagement Activity | 2022 |
URL | https://frontlinegenomics.com/the-tri-omics-summit-europe/ |
Description | Organisation of the TransMed special session at the ISMB conference |
Form Of Engagement Activity | Participation in an activity, workshop or similar |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | As co-chair of the Translational Medicine Informatics (TransMed) group within the International Society of Computational Biology, I play a pivotal role in bringing together computational and AI experts in translational medicine. I organize the annual TransMed special session at the ISMB conference (>300 attendees in 2023). |
Year(s) Of Engagement Activity | 2023 |
URL | https://transmed.github.io/team/ |
Description | Organiser of the "Systems approaches towards cancer cell plasticity" symposium |
Form Of Engagement Activity | Participation in an activity, workshop or similar |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | Together with Mohit Kumar Jolly (IISc India) we received seed funding to jointly organise a a symposium entitled "Systems approaches towards cancer cell plasticity", which has taken place on 31st March 2023 at UCL. We had a stellar list of international and UK-based speakers, including keynotes from Prof Giovanni Ciriello (University of Lausanne) and Prof Luca Magnani (Imperial College London), as well as a Rising Stars session where we showcased future leaders in the field. The symposium was highly successful, with >120 registered participants and excellent feedback, with 100% of the participants indicating they would like this to be an annual event. Indeed, we intend to raise further funds from companies in order to continue organising this symposium as a yearly or biennial event. The symposium has led to new collaborations with external groups, including Yolanda Calle Patino (University of Roehampton) for the Secrier lab (joint grant funding applications in progress) and Alejandra Bruna (ICR) for the Jolly lab. The Secrier and Jolly labs are in the process of writing a review on the topic together, and are planning collaborative grant applications. |
Year(s) Of Engagement Activity | 2023 |
Description | Panelist on session discussing "The changing landscape of scientific outputs and impacts on funding and career decisions" |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | Local |
Primary Audience | Postgraduate students |
Results and Impact | I was invited to take part in the UCL Cancer Institute Annual Conference as panelist in a session entitled "The changing landscape of scientific outputs (publications, data, algorithms etc) and their respective impacts on funding and career decisions". I was asked to discuss my perspective on this topic as a recent recipient of a UKRI Future Leaders Fellowship. This session has received an overwhelmingly positive feedback, with participants being very appreciative of the advice received. |
Year(s) Of Engagement Activity | 2021 |
Description | Poster spotlight at the "EACR The Invisible Phase of Metastasis" |
Form Of Engagement Activity | Participation in an activity, workshop or similar |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | I was awarded a poster spotlight short talk at this conference, aiming to discuss various aspects of metastases. I presented about "Genomic and microenvironmental heterogeneity shaping EMT in cancer". The talk and poster were well received and generated a lot of discussion. |
Year(s) Of Engagement Activity | 2022 |
URL | https://www.eacr.org/conference/metastasisdormancy2022virtual |
Description | Talk at the "EACR Cellular Bases for Patient Response to Conventional Cancer Therapies" |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | Conference bringing together experts in cancer therapies. I presented "A novel quiescence signature captures rapid drug tolerance development in cancer persister cells", which generated a lot of interest and opportunities for future collaboration. |
Year(s) Of Engagement Activity | 2022 |
URL | https://www.eacr.org/conference/cellularbases2022/introduction |
Description | Talk at the EACR Congress |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | Largest conference on cancer research in Europe. I presented on "Genomic hallmarks and therapeutic implications of cellular quiescence in cancer" in the Tumour Dormancy session, which resulted in requests for further information about our developed signatures and collaboration opportunities. |
Year(s) Of Engagement Activity | 2022 |
URL | https://www.eacr.org/congress |
Description | Talk at the Health Data Research UK Lunchtime Seminar Series |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | National |
Primary Audience | Postgraduate students |
Results and Impact | I was invited to discuss computational and statistical methods in cancer, and their application to study cancer evolution at the HDRUK Lunchtime Seminar Series, which is organised for PhD students enrolled in the HDRUK-Turing PhD programme. This is a diverse audience formed mainly of statisticians, mathematicians and computer scientists, so my talk was helpful for providing them with a broader view of the applications of various math/statistical methods in the field of biomedicine. |
Year(s) Of Engagement Activity | 2022 |
URL | https://www.hdruk.ac.uk/careers-in-health-data-science/further-education/phd-programme/ |
Description | Talk on "Mutational processes and cancer evolution" at the UCL Biological Society |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | Local |
Primary Audience | Undergraduate students |
Results and Impact | I was invited by the UCL Biological Society to discuss computational/statistical and bioinformatics methods, and their application to the study of cancer genomes and cancer evolution. The result was inspiring the younger generation about potential careers in this STEM area. |
Year(s) Of Engagement Activity | 2021 |
Description | UCL Digital Pathology Deep Learning Working Group |
Form Of Engagement Activity | A formal working group, expert panel or dialogue |
Part Of Official Scheme? | No |
Geographic Reach | Local |
Primary Audience | Professional Practitioners |
Results and Impact | We have organised a UCL-wide working group to exchange expertise and initiate new collaborations in the field of Deep Learning for Digital Pathology. An initial meeting was held in September 2021, and we intend to organise further meetings on a regular basis. This will be established as an official working group within UCL and we will share data and code for this purpose. The overall aim is to enhance the profile of UCL in this area of research and enable interdepartmental collaborations. |
Year(s) Of Engagement Activity | 2021 |