Dissecting the phenotype and function of sub-types of breast cancer associated fibroblasts (CAFs)

Primary supervisor: John Marshall, Queen Mary University of London

Secondary supervisor: J Louise Jone, Queen Mary University of London

Project

We have tested a panel of eight Cancer Associated Fibroblasts (CAFs) isolated from human breast cancer tissues by the Breast Cancer Now Tissue bank team. We used the 3D tumour-stroma interface mimetic system of organotypic gels we optimised (1) for the ability of the fibroblasts to support the growth and invasion of MDA MB468 and MCF1-A.CA1a breast cancer cell lines. The data showed that the response of the different CAFs was highly mixed. The fibroblasts affected three distinct parameters-matrix production (the organotypic gels got thicker), tumour cell proliferation and tumour cell invasion. The ability of each CAF to modulate these three elements varied suggesting that breast cancer CAFs have highly variable abilities and phenotypes. It is now clear through genetic studies that breast cancers (2), like other cancers (3) have multiple different fibroblast phenotypes and that some may be tumour promoting or tumour suppressive. However what is less well known is at the molecular level how these different subtypes of fibroblast influence the behaviour different cancer cells.

Project Aims

1. Test the functional phenotypes in 3D organotypic assays with MDA MB468 and MCF10A.CA1a of 20 CAFs (an additional 12 CAFs to those tested) from 20 patients with breast cancer. Gels will be analysed by H&E and cytokeratin staining (tumour cell marker) for functional phenotypes, and for stiffness using Atomic Force Microscopy (in collaboration with Dr Matteo Palma, Chemistry, QMUL) and correlated with collagen fibre density and pattern (picosirius red) and immunohistology for phosphorylated myosin-light chain kinase.
2. Use RNAseq to analyse the transcriptome of these same 20 CAF strains. Analyse the transcriptional outputs of the different subtypes of fibroblasts identified.
3. Correlate the transcriptional data with the functional observations (matrix production , tumour cell proliferation, tumour cell invasion). Determine if different fibroblast sub-types correlate with functional readouts.
4. Choose up to 5-10 genes that most closely correlate with functional phenotypes and KO (CRISPR) in CAFs showing the same phenotype (n=3 different strains) and repeat in vitro 3D organotypic studies.
5. Develop organotypic gels with CAF/cancer cell pairs that show matrix, proliferation or invasion phenotypes (n=3 strains each) for 10 days before implanting under the flank skin of athymic mice (4). Monitor tumour formation (MRI) over time and local invasion and metastasis at harvest. (CAFs will be labelled with luciferase to confirm they remain local to tumour. We have shown this works). Analyse tumours with multiplex spectral flow cytometry and multiplex T-CyCIF immunochemistry to determine total, spatial and subtype of fibroblast (mouse and human), macrophage and neutrophil. Determine how in vitro phenotype correlates with in vivo behaviour.
6. Use KO CAFs from Aim 4 (best 3 genes) to assess their effects in vivo (following methods in Aim 5)
7. Compare the transcriptional data against human breast cancer tissues in silico (TCGA, METABRIC) to establish genes that correlate with poor survival. Compare these genes with genes for in vitro phenotypes. Test up to three (novel) gene targets in 3D organotypics with suitable CAFs deficient (CRISPR) in lead candidates in vivo as above.

Project goals: (1) identify which subtypes of breast CAFs more strongly promote breast cancer proliferation and invasion in tumour mimetic gels in vitro (2) identify which subtypes of breast CAFs more strongly promote breast cancer proliferation and invasion in vivo and identify CAF-derived molecular drivers of these effects.

References

1. Nyström ML, Thomas GJ, Stone M, Mackenzie IC, Hart IR and Marshall JF. Development of a quantitative method to analyse tumour cell invasion in organotypic culture. J Path.2005 205:468-475.
2. Bartoschek M et al. Spatially and functionally distinct subclasses of breast cancer-associated fibroblasts revealed by single cell RNA sequencing. Nat Commun. 2018;9:5150
3. Neuzillet C et al. Inter- and intra-tumoural heterogeneity in cancer-associated fibroblasts of human pancreatic ductal adenocarcinoma. J Pathol. 2019 May;248(1):51-65
4. Nyström ML, McCulloch D, Weinreb PH, Violette SM, Speight PM, Marshall JF, Hart IR and Thomsas GJ. Cyclo-oxygenase-2 inhibition suppresses alphav beta6 integrin-dependent oral squamous carcinoma invasion. Cancer Res 2006 66: 10833-10842.