Single-cell analysis of basal cell carcinoma reveals heat shock proteins promote tumor growth in response to WNT5A-mediated inflammatory signals

Christian F. Guerrero‐Juarez, Gun Ho Lee, Yingzi Liu, Shuxiong Wang, Yutong Sha, Rachel Y. Chow, Tuyen T. L. Nguyen, Sumaira Z. Aasi, Matthew Karikomi, Michael L. Drummond, Qing Nie, Kavita Y. Sarin, Scott X. AtwoodbioRxiv 2021 · DOI

TL;DR

This study uses single-cell RNA sequencing to dissect the complex cellular ecosystem of basal cell carcinoma (BCC), revealing a critical crosstalk between tumor cells and their surrounding stroma. They found that inflammatory signals from fibroblasts, mediated by WNT5A, trigger tumor cells to produce heat shock proteins (HSPs), which in turn promote tumor growth, suggesting HSP inhibitors as a potential therapeutic avenue.

What Problem Does This Paper Address?

Basal cell carcinoma (BCC), the most common human cancer, is known to interact with its tumor microenvironment, but the precise mechanisms driving this interaction and promoting tumor growth remain unclear. Understanding this complex interplay is crucial for developing more effective treatments, especially for advanced or resistant cases. This paper aims to unravel the cellular ecosystem of BCC, distinguishing between normal and malignant cells, and to identify the signaling pathways that govern tumor-stroma communication and influence tumor progression.

Methods

The researchers employed a multi-pronged approach:

  • Single-cell RNA sequencing (scRNA-seq): This was the cornerstone technique, allowing for the analysis of individual cells within BCC tumors and surrounding normal tissue. They used 3'-droplet-enabled scRNA-seq to capture a high-resolution view of cellular heterogeneity.
  • Computational analyses:
    • InferCNV: To identify putative malignant epithelial cells by detecting copy number variations.
  • Seurat, SCTransform, LIGER, Harmony, and scMC: Various integration and clustering algorithms were benchmarked to find the best approach for analyzing and integrating diverse single-cell datasets, particularly for handling the mutational burden in cancer cells. scMC proved most effective in distinguishing normal from malignant epithelial cells.
  • Pseudotime and RNA velocity analysis (scVelo): To reconstruct developmental trajectories and understand cell state transitions within the stromal compartment.
  • Cellular entropy estimation (CEE): To quantify the stability and potential for transition of different cell states.
  • Gene Regulatory Network (GRN) analysis (pySCENIC): To identify key transcription factors and regulatory modules driving specific cellular behaviors.
  • Cell-cell communication analysis (CellChat): To model and infer signaling interactions between different cell types.
  • In situ validation: Protein expression and localization of key molecules were confirmed using immunohistochemistry and confocal microscopy on human tumor samples.
  • In vitro and in vivo experiments: The functional role of heat shock proteins (HSPs) was investigated using a BCC cell line and a mouse model, employing an HSP70 inhibitor.

Datasets

The study utilized:

  • Human BCC samples: Six surgically discarded tissues were obtained, including four basal cell carcinoma samples and two peri-tumor normal skin (PTS) samples. These were classified into different BCC subtypes (superficial, nodular, infiltrative, and infiltrative with perineural invasion).
  • Murine BCC cell line (ASZ001): Used for in vitro experiments.
  • BCC mouse model (Gli1-CreERT2; Ptch1fl/fl): Used for in vivo studies.

Key Insights and Ideas

  • Stromal rewiring is key: The study highlights that the tumor microenvironment, specifically fibroblasts, undergoes a significant "rewiring" process in BCC. This isn't just a passive bystander effect; it's an active reprogramming that fuels tumor growth.
  • WNT5A as an inflammatory conductor: The non-canonical WNT5A pathway emerges as a central player, induced by inflammatory signals in fibroblasts, which then maintains a pro-inflammatory state in the stroma.
  • HSPs as tumor's stress response: BCC cells respond to this WNT5A-driven inflammatory signaling by upregulating heat shock proteins (HSPs). This suggests HSPs are not just a general stress response but a specific mechanism that tumor cells exploit to promote their own growth.
  • Therapeutic vulnerability in HSPs: The finding that inhibiting HSP70 significantly suppresses BCC growth both in vitro and in vivo points to HSPs as a promising therapeutic target.

Key Results

  • Cellular heterogeneity: scRNA-seq revealed significant heterogeneity within BCC tumors, both at the epithelial and stromal levels.
  • Fibroblast subtypes: Four distinct fibroblast (FIB) populations and two FIB-like populations were identified, with specific subtypes showing altered expression patterns in BCC compared to normal tissue.
  • ECM remodeling: BCC stroma showed extensive remodeling, with upregulation of collagen and metalloproteinase genes.
  • WNT5A-driven inflammation: Cell-cell communication analysis identified ncWNT signaling, primarily driven by WNT5A, as highly active in BCC and linked to inflammatory pathways like TNF, IL6, and IFN-I.
  • HSP signature in BCC: A significant heat shock protein (HSP) signature was observed in BCC epithelial cells.
  • HSP70 inhibition efficacy: Treatment with an HSP70 inhibitor (Ver155008) significantly reduced ASZ001 cell proliferation and Gli1 expression in vitro, and suppressed micro-tumor growth in vivo in a dose-dependent manner.

Limitations and Future Work

  • Integration challenges: While scMC showed promise, integrating single-cell data from highly heterogeneous and mutated tumors like BCC remains a technical challenge. The authors note that some algorithms struggled to distinguish normal from malignant cells due to high mutational burden.
  • Functional validation of specific pathways: While WNT5A and HSPs were highlighted, further functional studies could explore the precise roles of other identified signaling pathways and transcription factors in BCC progression.
  • Combinatorial therapies: The authors suggest that HSP inhibition might be more effective in combination with other treatments, a direction for future clinical trials.
  • Translational potential: The identification of HSP70 as a key regulator opens avenues for developing novel therapeutic strategies for BCC, but further preclinical and clinical validation would be necessary.

Summary updated

Abstract

Abstract How basal cell carcinoma (BCC) interacts with its tumor microenvironment to promote growth is unclear. Here we use singe-cell RNA sequencing to define the human BCC ecosystem and discriminate between normal and malignant epithelial cells. We identify spatial biomarkers of both tumors and their surrounding stroma that reinforce the heterogeneity of each tissue type. Combining pseudotime, RNA velocity, cellular entropy, and regulon analysis in stromal cells reveal a cancer-specific rewiring of fibroblasts where STAT1, TGF-β, and inflammatory signals induce a non-canonical WNT5A program that maintains the stromal inflammatory state. Cell-cell communication modeling suggests that tumors respond to the sudden burst of fibroblast-specific inflammatory signaling pathways by producing heat shock proteins, which we validated in situ . Finally, dose-dependent treatment with an HSP70 inhibitor suppresses in vitro BCC cell growth and Hedgehog signaling and in vivo tumor growth in a BCC mouse model, validating HSP70’s essential role in tumor growth and reinforcing the critical nature of tumor microenvironment crosstalk in BCC progression.
BibTeX
@misc{guerrerojuarez2021single,
  author = {Christian F. Guerrero‐Juarez and Gun Ho Lee and Yingzi Liu and Shuxiong Wang and Yutong Sha and Rachel Y. Chow and Tuyen T. L. Nguyen and Sumaira Z. Aasi and Matthew Karikomi and Michael L. Drummond and Qing Nie and Kavita Y. Sarin and Scott X. Atwood},
  title = {{Single-cell analysis of basal cell carcinoma reveals heat shock proteins promote tumor growth in response to WNT5A-mediated inflammatory signals}},
  year = {2021},
  doi = {10.1101/2021.10.07.463571},
  howpublished = {bioRxiv},
}
Figure