Predicting outcome from features of Lymph Nodes
Lymph nodes are the initial site of breast cancer metastasis, however as a secondary lymphoid organ form an essential part of the adaptive immune response. We have previously found that morphological structures within the lymph nodes are correlated with breast cancer patients’ survival (Grigoriadis, 2018; Liu, 2021). Based on this and the development of computational biology, we established a multiscale deep-learning framework to detect, quantify and measure these morphological features in the lymph nodes of breast cancer patients (Verghese, 2023). We are subsequently combining RNA-sequencing, spatial transcriptomics and these deep-learning technologies to further understand these changes and elucidate their prognostic potential.
Evolution of the premetastatic lymph node niche
The presence of lymph node (LN) metastasis is one of the most important prognostic factors in breast and many other cancers, and the overall survival decreases as the number of cancerous (involved) LNs increases. Despite, a subset of these patients respond well to treatment and achieve long-term survival. Most commonly, regional LNs are excised, histopathologically processed and examined by a pathologist to determine if LNs harbour cancerous cells. Although tumour cell metastasis is often preceded by alterations in the microenvironment of the metastatic organ in preparation for the arrival of and an effective colonisation by malignant cells, little attention is given to cancer-free (uninvolved) LNs and the progression from an uninvolved LN to an involved LN. In collaboration with Professors Tony Ng, Professor Sarah Pinder and Professor Ton Coolen, we were the first to report on morphological changes in the uninvolved LNs being risk predictive of developing distant metastasis and are now elucidating the underlying biological and translational relevance of the pre-metastatic LNs.
We are currently using multiple methods and sample types to understand the function of the LN in the anti-tumour immune response. In collaboration with Dr. Dinis Calado at the Francis Crick Institute, we have established mouse models of breast cancer to investigate these histopathological features in more depth. We are further interrogating archival and fresh patient material to investigate the intricacies of the nodal microenvironment. The techniques we utilise to do so include gene expression analyses, immunofluorescence, imaging mass cytometry, flow cytometry, in vitro chemotaxis assays and spatial transcriptomics.
A biomarker of BRCAness
‘BRCAness’ is a term used to describe sporadic breast cancers that share clinical, pathological, and molecular features with breast cancers that occur in women with a germline BRCA1 mutation. One such defining phenotype is the sensitivity of tumour cells to DNA damaging drugs, including platinum agents. In our previous work, we have shown that the expression of HORMAD1, a gene that is normally expressed only during meiosis, is associated with genomic instability in triple-negative breast cancers (TNBCs). To further our understanding of the role of HORMAD1 in breast cancer and its association with BRCAness, we are investigating the clinical, pathological, and molecular data from 4,663 breast cancers. Although follow-up studies are warranted, preliminary findings suggest that this gene has the potential to be an accessible biomarker of BRCAness that may be suitable to inform treatment in women with breast cancer.
Engagement of the immune system
In this project we seek to understand the relationship between HORMAD1 as a driver of genomic instability and as a mediator of the innate immune response in triple-negative breast cancer by analysing transcriptomic data from primary TNBCs. While the genomic landscape provides valuable information of on-going and historical mutagenic processes, the transcriptome can give insights into compensatory pathways required to maintain cell viability and to interrogate the tumour immune microenvironment. Using bulk RNA-sequencing, we are exploring how genomic instability in HORMAD1-positive triple-negative breast cancers trigger the activation of cell stress ligands and subsequently evokes an immune response. With single-cell RNA sequencing and spatial transcriptomics, we can further explore interactions between cancer and immune cells on a single-cell level. Insights into the immunobiology and DNA damage response may identify new subgroups of patients benefitting from immunotherapy, chemotherapy, or potentially novel targeted therapies. Based on biological features, these patient subgroups can be investigated retrospectively in clinical trials with regards to appropriate treatment response.
HORMAD1 and immunological visibility
As a cancer/testis antigen, HORMAD1 is a meiotic cancer/testis antigen whose expression is usually restricted to germline cells, but we have shown that it is aberrantly expressed in around 60% of triple negative breast cancers and is related to genomic instability. We want to establish the immunogenicity of HORMAD1 by demonstrating that as a cancer/testis antigen it can engage an adaptive T-cell based immune response. T cell receptors (TCRs) determine a T cells ability to recognize a specific antigen and we aim to identify HORMAD1 TCRs. We’ve collected blood from breast cancer patients to stimulate in-vitro T cells with a HORMAD1 peptide library. By leveraging state of the art TCR sequencing technologies we can characterize the clonality of the T cell repertoires to find HORMAD1 specific T cell receptors. If HORMAD1 can induce a tumour-rejecting response, the HORMAD1 TCR candidates could be used to develop targeted T cell-based immunotherapies or vaccines for HORMAD1 positive breast cancer patients.
Image-based risk assessment of BRCA1/2 mutation carriers
We are dedicated to provide image-based individual risk assessment for germline BRCA1/2 mutation carriers. An image repository named “OASIS” has been created to continously collect whole slide images (WSIs) of normal breast tissue derived from multiple cohorts and patients from different cancer risk groups (https://github.com/cancerbioinformatics/OASIS). We will use image analysis techniques to recognise histological patterns of lobules and stroma and extract features to build up models to predict cancer risk for these at-risk women.
Image by brgfx on Freepik