Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Exploring Breast Cancer Associated Fibroblast Subpopulations By Using Single-Cell RNA Sequencing

Boge, Joel (2018) BINP50 20181
Degree Projects in Bioinformatics
Abstract
In the tumor microenvironment, Cancer-Associated Fibroblasts (CAFs) are the most represented cell type. Research has shown increasingly the key role of CAFs in cancer growth and evolution, creating the need to identify more clearly and further characterize the CAF population. In this study, looking at single cell RNA sequencing of 768 CAF transcriptomes, we define transcriptionally distinct subpopulations and isolate surface marker candidates for each of them. Using publicly available studies, we identify a gene set exclusive and specific to CAFs, improving their taxonomy and allowing better focus on this cell type in future experiments. We use bulk RNA sequencing data of the MMTV-PyMT mouse model of breast tumor samples, featuring tumors... (More)
In the tumor microenvironment, Cancer-Associated Fibroblasts (CAFs) are the most represented cell type. Research has shown increasingly the key role of CAFs in cancer growth and evolution, creating the need to identify more clearly and further characterize the CAF population. In this study, looking at single cell RNA sequencing of 768 CAF transcriptomes, we define transcriptionally distinct subpopulations and isolate surface marker candidates for each of them. Using publicly available studies, we identify a gene set exclusive and specific to CAFs, improving their taxonomy and allowing better focus on this cell type in future experiments. We use bulk RNA sequencing data of the MMTV-PyMT mouse model of breast tumor samples, featuring tumors between 8-16 weeks of age, to investigate the temporal evolution of CAF signatures. While the CAF signatures evolution could not be detected, our results shines light on new markers defining the CAF population more precisely, as well as identify each CAF subpopulation’s gene profile and specific marker genes. (Less)
Popular Abstract
Characterization of Breast Cancer Associated Fibroblasts (CAFs)

The tumor stroma has been shown to influence initiation and evolution of cancer. Among this microenvironment, CAFs play an important role as main component of this stroma in breast cancer. Unfortunately, no specific markers are available to identify CAFs or differentiate the existing populations. In this project we used single-cell RNA sequencing, analyzing the expression profile of isolated mice breast CAFs to further characterize the CAF populations. We aimed to reach our goal in three steps: identifying specific surface markers for CAF populations, compare populations’ gene signature to bulk RNA sequence data and compare our CAF signature to publicly available datasets.
... (More)
Characterization of Breast Cancer Associated Fibroblasts (CAFs)

The tumor stroma has been shown to influence initiation and evolution of cancer. Among this microenvironment, CAFs play an important role as main component of this stroma in breast cancer. Unfortunately, no specific markers are available to identify CAFs or differentiate the existing populations. In this project we used single-cell RNA sequencing, analyzing the expression profile of isolated mice breast CAFs to further characterize the CAF populations. We aimed to reach our goal in three steps: identifying specific surface markers for CAF populations, compare populations’ gene signature to bulk RNA sequence data and compare our CAF signature to publicly available datasets.

In order to characterize the CAF populations, we used dimensionality-reduction techniques such as Principal Component Analysis (PCA) and t-distributed stochastic neighbor embedding (t-SNE) allowing conversion of our data into 2 or 3D data while preserving its structure. A clustering algorithm, DBSCAN (Density-Based Spatial Clustering of Applications with Noise), identified 4 distinct CAF populations. The reproducibility optimized test statistic (ROTS) method was then performed on the clusters to detect the differentially expressed genes between the populations. For each CAF population, we isolated potential surface markers candidates and analysed their expression

Following the identification of CAF populations’ signature, we used these signatures in a temporal evolution analysis between our 12weeks old tumor samples and bulk RNA sequenced data of 8-16weeks old breast tumors to detect changes in gene expression and subpopulations over time. Bowtie, TopHat and Cufflinks software were used to align and map the bulk data to the reference genome. The expression variability of CAF signatures over time revealed no pattern in population 1-3, but an increase in gene expression as the tumor progress for population 4.

In the last part of our project, we used two publicly available single-cell RNA sequenced datasets respectively of human colo-rectal tumor and human primary breast tumor to refine the overall CAF population signature. After identifying convertible genes from human to mouse, we integrated the datasets together in order to further analyse and compare the gene expressions. ROTS was then performed on the integrated data to isolate potential surface pan-CAF markers overexpressed in CAFs. Using the Human protein Atlas, we identified candidates with high expression in CAFs and presenting a strong stromal signature in breast tumor samples.

Our work did not only further characterize the different CAF populations identifying suitable surface marker candidates, but also uncovered potential global surface pan-CAF markers specific and exclusive to fibroblasts. Our analysis of the signature evolution of CAF populations following cancer progression didn’t bring the expected result but could have revealed a possibility of discerning tumor in earlier stage and following its evolution. Our results open possibilities of novel clinically relevant biomarkers and more precise drug targeting in breast tumor treatment.


Master’s Degree Project in Bioinformatics 30 credits 2018
Department of Biology, Lund University

Advisor: Kristian Pietras, Michael Bartoschek
Translational Cancer Research, Faculty of Medicine, Lund University (Less)
Please use this url to cite or link to this publication:
author
Boge, Joel
supervisor
organization
course
BINP50 20181
year
type
H2 - Master's Degree (Two Years)
subject
language
English
id
8958752
date added to LUP
2018-09-14 10:12:57
date last changed
2018-09-14 10:12:57
@misc{8958752,
  abstract     = {{In the tumor microenvironment, Cancer-Associated Fibroblasts (CAFs) are the most represented cell type. Research has shown increasingly the key role of CAFs in cancer growth and evolution, creating the need to identify more clearly and further characterize the CAF population. In this study, looking at single cell RNA sequencing of 768 CAF transcriptomes, we define transcriptionally distinct subpopulations and isolate surface marker candidates for each of them. Using publicly available studies, we identify a gene set exclusive and specific to CAFs, improving their taxonomy and allowing better focus on this cell type in future experiments. We use bulk RNA sequencing data of the MMTV-PyMT mouse model of breast tumor samples, featuring tumors between 8-16 weeks of age, to investigate the temporal evolution of CAF signatures. While the CAF signatures evolution could not be detected, our results shines light on new markers defining the CAF population more precisely, as well as identify each CAF subpopulation’s gene profile and specific marker genes.}},
  author       = {{Boge, Joel}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{Exploring Breast Cancer Associated Fibroblast Subpopulations By Using Single-Cell RNA Sequencing}},
  year         = {{2018}},
}