Data-driven discovery of targets for bipotent anticancer drugs identifies Estrogen Related Receptor Alpha

Avinash Sahu, Xiaoman Wang, Phillip Munson, Jan P.g. Klomp, Xiaoqing Wang, Shengqing Gu, Gege Qian, Phillip Nicol, Zexian Zeng, Chenfei Wang, Collin Tokheim, Wubing Zhang, Jingxin Fu, Jin Wang, Nishanth Ulhas Nair, Joost A.p. Rens, Meriem Bourajjaj, Bas Jansen, Inge Leenders, Jaap Lemmers, Mark Musters, Sanne van Zanten, Laura van Zelst, Jenny Worthington, Myles Brown, Jun S. Liu, Dejan Juric, Cliff Meyer, Arthur Oubrie, Xia Liu, David E. Fisher, Keith T. FlahertybioRxiv 2021 · DOI

TL;DR

This paper introduces BipotentR, a computational tool that identifies drug targets capable of simultaneously impacting both tumor immunity and a critical oncogenic pathway. They demonstrate its power by discovering Estrogen Related Receptor Alpha (ESRRA) as a promising target that simultaneously disrupts tumor metabolism and enhances anti-tumor immune responses, offering a novel strategy for developing more effective cancer therapies.

What Problem Does This Paper Address?

Cancer cells often develop resistance to single-target therapies by evolving resistance mechanisms or exploiting alternative pathways. The authors propose that "bipotent" drugs, which simultaneously hit two distinct oncogenic pathways, could offer broader clinical benefits, overcome resistance, and be effective in diverse patient populations. However, identifying such bipotent drug targets has been challenging due to a lack of systematic discovery approaches. This paper aims to fill that gap by developing a computational framework to predict these crucial dual-action targets.

Methods

The core innovation is the BipotentR computational approach, which integrates multiple data types to predict bipotent regulators. It consists of two modules:

  • Regulation Module: This module uses ChIP-seq data from ~700 transcription factors and chromatin regulators (TFCRs) across ~24,000 samples to estimate their binding affinity to cis-regulatory elements near genes involved in user-defined oncogenic pathways (e.g., energy metabolism, angiogenesis). This module identifies TFCRs that can regulate these specific pathways.
  • Immune Module: This module leverages bulk RNA-seq data from patient tumors to assess the immunomodulatory potential of TFCRs by correlating their expression with a proinflammatory signature. Crucially, it then uses single-cell RNA-seq and ATAC-seq data to identify TFCRs that are cancer-cell-specific, meaning they are active in cancer cells but not in CD8+ T cells, thus minimizing off-target effects on the anti-tumor immune response.

By integrating the outputs of these two modules, BipotentR identifies TFCRs that can simultaneously regulate an oncogenic pathway and modulate the immune response in a cancer-cell-specific manner.

Datasets

The study utilized a variety of datasets for training and validation:

  • ChIP-seq data: Approximately 24,000 samples were used to identify TFCR binding sites.
  • Bulk tumor transcriptomes: Data from ~5,000 human transcriptome datasets and 33 cancer types from The Cancer Genome Atlas (TCGA) were used to assess correlations between predicted regulators and gene expression, and to estimate immunomodulatory potential.
  • Single-cell RNA-seq (scRNA-seq) data: Datasets from 5 cancer cohorts and 78 single-cell transcriptome datasets from patients with 27 major cancer types were used to determine cancer-cell specificity of TFCRs.
  • Single-cell ATAC-seq (scATAC-seq) data: Used to confirm the functional activity of TFCRs in cancer cells versus CD8+ T cells.
  • CRISPR screen data: Published datasets were used to validate the predicted bipotent targets by assessing the effect of gene knockouts on T-cell mediated killing and immune response.
  • Patient tumor transcriptomes: 33,000 tumor transcriptomes from TCGA and PRECOG datasets were analyzed to correlate ESRRA activity with clinical immune phenotypes.
  • In vivo mouse models: 4T1 (breast cancer) and B16F10 (melanoma) models were used for experimental validation.

Key Insights and Ideas

  • The power of dual-action: The paper strongly advocates for the therapeutic potential of targeting two pathways simultaneously, moving beyond single-target approaches.
  • Computational integration is key: BipotentR's success lies in its ability to integrate diverse data types (epigenomics, bulk transcriptomics, single-cell transcriptomics) to identify targets with specific, dual functions.
  • Cancer-cell specificity is paramount: The immune module's focus on identifying TFCRs active in cancer cells but not CD8+ T cells is crucial for developing therapies that boost immunity without compromising it.
  • ESRRA as a prime example: The identification of ESRRA as a top bipotent immune-metabolic regulator highlights the approach's ability to uncover clinically relevant targets. Its dual role in energy metabolism and immune modulation is a compelling proof-of-concept.
  • ESRRA's role in resistance: The finding that ESRRA activity is high in immune-suppressive and immunotherapy-resistant tumors suggests it could be a key player in treatment failure, making its inhibition a promising strategy for overcoming resistance.

Key Results

  • BipotentR identified 38 candidate immune-metabolic regulators, including 30 immunosuppressive and 8 immunostimulatory TFCRs.
  • The top candidate, ESRRA, was shown to inhibit tumor growth through direct effects on energy metabolism and two immune mechanisms: cytokine induction (leading to proinflammatory macrophage polarization) and antigen-presentation stimulation (recruiting CD8+ T cells).
  • Inhibition of ESRRA (ESRRAi) in mouse models markedly reduced tumor growth and prevented tumor relapse.
  • ESRRAi treatment led to increased infiltration and activation of CD8+ T cells and suppressed regulatory T cells (Tregs).
  • ESRRAi also promoted M1 macrophage polarization, shifting them towards a more anti-tumorigenic state.
  • Analysis of patient data revealed that high ESRRA activity is associated with immunosuppressive and immunoresistant tumors, suggesting broad clinical relevance.
  • The BipotentR approach was also successfully applied to angiogenesis and growth-suppressor pathways, demonstrating its generalizability.

Limitations and Future Work

  • Translational challenges: While ESRRAi showed promise in preclinical models, further clinical investigation is required to confirm its efficacy and safety in human patients, particularly in immunotherapy-resistant settings.
  • Off-target effects: Although the immune module aims to minimize off-target effects on CD8+ T cells, the possibility of unintended consequences from inhibiting a nuclear receptor like ESRRA warrants careful consideration in drug development.
  • Broader pathway exploration: While the paper demonstrates BipotentR's applicability to other pathways, further exploration of these targets and their therapeutic potential is needed.
  • Mechanism refinement: While the paper elucidates key mechanisms of ESRRA inhibition, a deeper understanding of the intricate interplay between ESRRA, metabolism, and the immune system could lead to more refined therapeutic strategies.
  • BipotentR accessibility: The authors have made BipotentR publicly available, which is a crucial step for enabling further research and drug discovery in this area.

Summary updated

Abstract

Abstract Drugs that kill tumors through multiple mechanisms have potential for broad clinical benefits, with a reduced propensity to resistance. We developed BipotentR, a computational approach to find cancer-cell-specific regulators that simultaneously modulate tumor immunity and another oncogenic pathway. Using tumor metabolism as proof-of-principle, BipotentR identified 38 candidate immune-metabolic regulators by combining epigenomes with bulk and single-cell tumor transcriptomes from patients. Inhibition of top candidate ESRRA (Estrogen Related Receptor Alpha) killed tumors by direct effects on energy metabolism and two immune mechanisms: (i) cytokine induction, causing proinflammatory macrophage polarization (ii) antigen-presentation stimulation, recruiting CD8 + T cells into tumors. ESRRA is activated in immune-suppressive and immunotherapy-resistant tumors of many types, suggesting broad clinical relevance. We also applied BipotentR to angiogenesis and growth-suppressor pathways, demonstrating a widely applicable approach to identify drug targets that act simultaneously through multiple mechanisms. BipotentR is publicly available at http://bipotentr.dfci.harvard.edu/ . One-Sentence Summary BipotentR identifies targets for bipotent anticancer drugs, as shown by the energy and immune effects of ESRRA inhibition.
BibTeX
@misc{sahu2021data,
  author = {Avinash Sahu and Xiaoman Wang and Phillip Munson and Jan P.G. Klomp and Xiaoqing Wang and Shengqing Gu and Gege Qian and Phillip Nicol and Zexian Zeng and Chenfei Wang and Collin Tokheim and Wubing Zhang and Jingxin Fu and Jin Wang and Nishanth Ulhas Nair and Joost A.P. Rens and Meriem Bourajjaj and Bas Jansen and Inge Leenders and Jaap Lemmers and Mark Musters and Sanne van Zanten and Laura van Zelst and Jenny Worthington and Myles Brown and Jun S. Liu and Dejan Juric and Cliff Meyer and Arthur Oubrie and Xia Liu and David E. Fisher and Keith T. Flaherty},
  title = {{Data-driven discovery of targets for bipotent anticancer drugs identifies Estrogen Related Receptor Alpha}},
  year = {2021},
  doi = {10.1101/2021.10.25.465724},
  howpublished = {bioRxiv},
}

Loading related content...

Figure