An evolving cancer instigates clonally unrelated neighboring cells to form distant metastases

Jie Dong, Cédric Darini, Farhia Kabeer, Sarah Dendievel, Junxia Min, Dolores Hambardzumyan, Louis Gaboury, Tom Lenaerts, Guillaume Darrasse-Jèze, Yi Li, Katrina PodsypaninabioRxiv 2016 · DOI

TL;DR

This paper challenges the traditional clonal evolution model of cancer by demonstrating that seemingly normal, "neighboring" mammary cells can be instigated by an evolving tumor to disseminate and even form distant metastases, independent of the primary tumor's initiating oncogene. This suggests a non-linear, collaborative model of cancer spread with significant implications for prevention and treatment.

What Problem Does This Paper Address?

The prevailing model of cancer progression, the clonal evolution theory, posits that cancer arises from a single cell acquiring mutations that grant it a growth advantage. Subsequent mutations lead to the emergence of subclones, and all metastatic growths are ultimately descendants of this initial mutated cell, sharing the very first oncogenic mutation. However, this model doesn't fully explain the observed heterogeneity in metastatic tumors, where some metastases may lack mutations present in the primary tumor. This paper investigates whether cells that haven't acquired the initial cancer-driving mutation can still contribute to cancer progression and metastasis, a phenomenon not well-explained by strict clonal evolution.

Methods

The researchers employed several sophisticated mouse models and techniques to track and differentiate normal and cancerous mammary cells:

  • Chimeric Mammary Gland Models: They created chimeric mammary glands by mixing genetically distinct mammary cells (e.g., RFP-tagged normal cells with YFP-tagged oncogene-expressing cells) and implanting them into mice. This allowed for direct visualization and tracking of cell interactions and dissemination.
  • Inducible Oncogene Expression: They used doxycycline-inducible oncogenes (like PyMT and ErbB2) to precisely control when cancer-driving mutations were activated in specific cell populations.
  • Fluorescent Labeling: Cells were labeled with different fluorescent proteins (RFP, YFP, GFP) to distinguish between normal and oncogene-expressing cells, and to track their movement and fate.
  • Intraductal Injections: They injected cells directly into mammary ducts to mimic natural tumor formation.
  • Viral Vectors (RCAS): For some experiments, they used RCAS viral vectors to deliver oncogenes into specific mammary cells, allowing for tracking of infected vs. uninfected cells.
  • Immunohistochemistry and Immunofluorescence: These techniques were used to visualize and quantify specific proteins (like HA tag for oncogene expression, E-cadherin for epithelial identity, and γH2A.X for DNA damage) in tissue sections.
  • Flow Cytometry: This was crucial for quantifying the proportion of different cell populations in dissociated tissues and blood, and for identifying cell surface markers.
  • Mathematical Modeling: A birth-death process model was developed to describe the dynamics of disseminated cells in distant organs.

Datasets

Not discussed in the paper. The study primarily relies on experimental data generated from mouse models and analysis of human tissue samples.

Key Insights and Ideas

  • "Collusion" of Normal Cells: The most striking finding is that evolving tumor cells can actively "instigate" or "collude" with their clonally unrelated, normal neighboring mammary cells, prompting them to disseminate.
  • De Novo Tumorigenesis in Neighbors: These "colluded" normal cells can not only spread but also undergo their own tumorigenic transformation and form independent distant metastases, even without carrying the primary tumor's initiating oncogene.
  • Non-Linear Cancer Spread: This challenges the linear clonal evolution model, suggesting that cancer progression can be a more complex, collaborative, and non-linear process.
  • DNA Damage as a Driver: In some experimental contexts (particularly with ErbB2 oncogene), the presence of an oncogene in neighboring cells can induce DNA damage in adjacent normal cells, potentially leading to new oncogenic mutations.
  • "Adjacent Normal" Tissue as a Reservoir: The study highlights that seemingly normal tissue adjacent to a tumor is not passive but can be a source of cancer progression and spread.

Key Results

  • In mouse models, genetically marked normal mammary cells were found to disseminate to the lungs when adjacent to growing tumors, even without expressing an oncogene themselves.
  • These disseminated normal cells could form metastatic foci in the lungs that lacked the initiating oncogene of the primary tumor.
  • Experiments using RCAS viral vectors to introduce oncogenes (ErbB2, Wnt1) showed that some lung metastases were initiated by mammary cells that were not infected by the virus, suggesting they were descendants of uninfected neighbors.
  • In some cases, these uninfected neighboring cells showed evidence of DNA damage, a potential precursor to acquiring new mutations.
  • A mathematical model supported the idea of sustained non-linear dissemination from primary tumors.

Limitations and Future Work

  • Experimental Context Dependence: The study notes that the induction of DNA damage in neighboring cells was dependent on the experimental context and the type of primary tumor. This suggests that the mechanisms by which normal cells are instigated to transform might vary.
  • Human Relevance: While human lobular carcinoma samples were analyzed for E-cadherin expression patterns, the primary experimental work was in mouse models. Further validation in human studies is implied.
  • Therapeutic Implications: The authors suggest that the "adjacent normal" tissue should be considered a reservoir for cancer progression and spread, implying that it might need to be targeted in cancer prevention and treatment strategies. Future work could explore how to specifically target these "colluded" cells or prevent their transformation and dissemination.
  • Mechanisms of Instigation: While the study demonstrates the phenomenon, the precise molecular mechanisms by which tumor cells instigate their neighbors (e.g., paracrine signaling, extracellular vesicles) warrant further investigation.

Summary updated

Abstract

Abstract Based on the clonal evolution theory of cancer formation, a single cell within a tissue gains a cancer-driving mutation and thus a growth advantage. From this expanded cellular mass, another cell gains a new mutation allowing this newly mutated cell to gain new competitive advantage and to expand in number (thus clonal expansion). Another clone then emerges. Eventually all required mutations are gained, and a cancer forms. Consequently, while a primary lesion may harbor divergent subclones, all the subclones within the primary cancer as well as all metastatic growths in secondary organs share at least the very first oncogenic mutation that initiates the primary cancer. However, by tracking genetically marked mammary epithelial cells that suffered the initiating oncogenic mutation—and their neighboring mammary cells that did not-in several mouse models of human breast cancer, we found that genetically unrelated mammary epithelial cells can be colluded by neighboring mutated cells to disseminate, and that they can even undergo de novo tumorigenic transformation and form distant metastases. Therefore, clonally unrelated epithelial cells may contribute to cancer progression and to the heterogeneity of the systemic disease. The non-linear cancer spread has important implications in cancer prevention, treatment, and therapeutic resistance.
BibTeX
@misc{dong2016evolving,
  author = {Jie Dong and Cédric Darini and Farhia Kabeer and Sarah Dendievel and Junxia Min and Dolores Hambardzumyan and Louis Gaboury and Tom Lenaerts and Guillaume Darrasse-Jèze and Yi Li and Katrina Podsypanina},
  title = {{An evolving cancer instigates clonally unrelated neighboring cells to form distant metastases}},
  year = {2016},
  doi = {10.1101/085423},
  howpublished = {bioRxiv},
}
Figure