Hybrid Stem Cell States: Insights Into the Relationship Between Mammary Development and Breast Cancer Using Single-Cell Transcriptomics

Tasha Thong, Yutong Wang, Michael Brooks, Christopher T. Lee, Clayton Scott, Laura Balzano, Max S. Wicha, Justin A. ColacinoFrontiers in Cell and Developmental Biology 2020 · DOI

TL;DR

This study uses single-cell transcriptomics to reveal that normal human mammary cells can exist in hybrid epithelial/mesenchymal (E/M) states, which are more developmentally immature and resemble embryonic stem cells. These hybrid states are enriched during pregnancy and are more prevalent in aggressive breast cancer subtypes, suggesting a link between developmental immaturity, hybrid cell plasticity, and cancer progression.

What Problem Does This Paper Address?

Cancer is often viewed as a "disease of dysregulated development," with similarities between cancer cells and stem cells, particularly their plasticity. Mammary stem cells (MaSCs) are implicated in breast cancer, and recent evidence suggests they can exist in distinct phenotypic states: epithelial, mesenchymal, and hybrid epithelial/mesenchymal (E/M). These hybrid states, especially E/M, are linked to metastasis and poor prognosis. However, hybrid cells are rare in normal mammary tissue, making them difficult to study. This paper aims to characterize these hybrid cell states in normal mammary tissue across different developmental stages and in response to an in vitro method (conditional reprogramming, CR) that expands stem-like populations. By integrating human and mouse single-cell data with bulk tumor data, the authors seek to understand the role of these hybrid states in normal mammary development and their connection to breast cancer.

Methods

The researchers employed single-cell RNA sequencing (scRNA-seq) to analyze normal human mammary (NM) cells before and after conditional reprogramming (CR). CR is an in vitro technique that expands stem-like cells from patient samples. They used machine learning-based dataset alignment (specifically, the CORGI algorithm) to integrate their human data with existing mouse mammary gland scRNA-seq datasets spanning various developmental stages. Pseudotime analysis was used to infer developmental trajectories. Differential gene expression analysis was performed to identify genes and pathways associated with different cell states and between NM and CR cells. They also analyzed bulk RNA-seq data from The Cancer Genome Atlas (TCGA) to compare breast cancer subtypes.

Datasets

  • Human Mammary Tissue: Normal mammary (NM) cells from three individuals undergoing voluntary reduction mammoplasty.
  • Conditionally Reprogrammed (CR) Cells: NM cells cultured in vitro using the CR method.
  • Mouse Mammary Gland Datasets:
    • Giraddi et al. (2018): Single-cell RNA-seq data spanning multiple developmental timepoints (embryonic day 16, embryonic day 18, postnatal day 0, postnatal day 4, and adult). This dataset served as the reference for developmental trajectory alignment.
  • Bach et al. (2017): Adult mouse mammary gland single-cell RNA-seq data from nulliparous, mid-gestation, lactation, and post-involution stages.
  • The Cancer Genome Atlas (TCGA): Bulk RNA-seq data from breast tumors across different subtypes.

Key Insights and Ideas

  • Hybrid States in Normal Mammary Tissue: The study demonstrates that normal mammary cells can exist in hybrid epithelial/mesenchymal (E/M) and luminal/basal (L/B) states, not just in cancer.
  • CR Induces Stem-like, Immature Phenotype: Conditional reprogramming (CR) expands a stem-like population in normal mammary cells, characterized by increased expression of embryonic stem cell (ESC) genes and a more "developmentally immature" transcriptional profile.
  • E/M Hybrids are Developmentally Immature: Hybrid E/M cells, identified by EPCAM/VIM co-expression, are the most developmentally immature among the hybrid populations, aligning with early embryonic stages in the mouse developmental trajectory.
  • Pregnancy Enriches Hybrid Cells: During gestation and lactation in mice, there's a significant enrichment of hybrid E/M and L/B cells, suggesting their crucial role in mammary gland morphogenesis during these dynamic periods.
  • Basal Breast Cancers are Most Immature: TCGA data reveals that basal breast cancer subtypes exhibit the most "developmentally immature" transcriptional signatures, aligning with the earliest developmental stages and suggesting a link between developmental immaturity and aggressive cancer.

Key Results

  • CR of human mammary cells led to an expanded stem cell state, with increased expression of ESC-associated genes.
  • CR cells showed higher embryonic stem cell scores compared to normal mammary cells.
  • Three hybrid populations were identified post-CR: KRT18+/KRT14+ (L/B), EPCAM+/VIM+ (E/M), and a quadruple-positive population.
  • E/M hybrids were found to be the most developmentally immature, followed by quadruple hybrids, and then L/B hybrids, both in mouse and CR cells.
  • In mouse models, hybrid E/M cells were enriched during gestation and lactation.
  • TCGA analysis showed that basal breast cancer subtypes mapped most closely to the most developmentally immature stages of the mouse mammary gland, with the most immature pseudotime estimates.

Limitations and Future Work

  • Source of Mammary Tissue: The study used mammoplasty tissue, which may not be fully representative of "normal" breast tissue.
  • De-identification of Samples: Lack of demographic data for the mammoplasty samples limited detailed correlation with individual characteristics.
  • CR Methodology: The CR method relies on co-culture with J2 fibroblasts, which might influence the observed cellular states.
  • Limited Gene Coverage: Single-cell technology captures only a subset of expressed genes.
  • Potential Bias in Alignment: Using a subset of CORGI-selected genes for alignment might introduce bias.

Future Work: * Investigate the localization of hybrid states in adult mammary glands using spatial transcriptomics. * Explore the role of hybrid E/M cells as therapeutic targets for aggressive cancers. * Study the influence of environmental exposures and genetic predisposition on stem cell heterogeneity. * Examine the impact of conditional reprogramming on breast cancer samples to understand enhanced stemness and developmental immaturity in tumors. * Further investigate the role of stromal/epithelial interactions in regulating hybrid stem cell states. * Characterize hybrid populations and developmental maturity in pregnancy-associated breast cancer (PABC).

Summary updated

Abstract

Similarities between stem cells and cancer cells have implicated a role for mammary stem cells in breast carcinogenesis. Recent evidence suggests that normal breast stem cells exist in multiple phenotypic states: epithelial, mesenchymal, and hybrid epithelial/mesenchymal. The proportion of cells in these states vary between individuals, suggesting that state dynamics may be influenced by genetics or environment. Conditional reprogramming (CR), an in vitro method of expanding patient derived tissue samples, promotes the rapid induction of a stem-like state in the absence of genetic manipulation. The goal of this study was to use single-cell RNA sequencing to quantify the cell state distributions of normal human mammary (NM) cells isolated from patients undergoing voluntary reduction mammoplasty (n=3) before and after CR, investigating stem cell populations, and identifying gene and pathway drivers of stem cell phenotypes. Unbiased clustering revealed that post-CR, myoepithelial and luminal cell populations are retained, while fibroblast, endothelial, and immune cell populations are depleted. Compared to NM cells, CR cells show higher expression of an embryonic stem cell gene signature and differentially express stem cell and cancer related genes (LGALS1, SKA2, MKI67, HJURP, BIRC5, CCNB1, and BUB1). Pseudotime analysis and alignment to a mouse single-cell transcriptome atlas spanning mammary gland development revealed that NM cells align most closely to adult mouse cells and CR cells align across the trajectory with a population aligning to the embryonic mouse cells. We identified the emergence of three hybrid populations, a KRT14+/KRT18+ population (L/B), consistent with luminal progenitor cells, an EPCAM+/VIM+ (E/M) population, associated with cells undergoing the epithelial to mesenchymal transition, and a quadruple positive hybrid population, expressing all four markers. Pseudotime analysis and alignment to the mouse developmental trajectory revealed that E/M hybrids are the most developmentally immature, aligning along both luminal and basal developmental trajectories. Finally, pseudotime analysis and alignment of bulk breast tumors from the cancer genome atlas (TCGA), revealed that breast cancer subtypes express distinct developmental signatures, with basal tumors expressing the most “developmentally immature” phenotype. These results highlight phenotypic plasticity of normal mammary stem cells and provide insight into the relationship between hybrid cell populations, stemness, and cancer.
BibTeX
@inproceedings{thong2020hybrid,
  author = {Tasha Thong and Yutong Wang and Michael Brooks and Christopher T. Lee and Clayton Scott and Laura Balzano and Max S. Wicha and Justin A. Colacino},
  title = {{Hybrid Stem Cell States: Insights Into the Relationship Between Mammary Development and Breast Cancer Using Single-Cell Transcriptomics}},
  booktitle = {Frontiers in Cell and Developmental Biology},
  year = {2020},
  doi = {10.3389/fcell.2020.00288},
}
Figure