A Single-Cell Window into Pancreas Cancer Fibroblast Heterogeneity

The clinical impact that tumor-associated fibrosis can have on tumor aggressiveness and resistance to therapy has become increasingly clear in recent years. In addition, the role of cancer-associated fibroblasts (CAF) as master regulators of fibrotic responses within the tumor microenvironment (TME) is well established. CAFs have been implicated in various tumor-supportive processes, including tumor growth and metastasis, angiogenesis, immune suppression, and fibrosis (1). Classically, we have also appreciated that activated fibroblasts are essential to isolating pathologically injured or infected tissues to protect the host. Nonetheless, in the context of cancer, the exact functions of CAFs continue to be debated, and this debate is most highly relevant in pancreatic ductal adenocarcinoma (PDAC), a disease characterized by dramatic tumor-associated desmoplasia. In some PDAC models, depletion of specific CAF subsets slows PDAC progression and improves antitumor immunity (2), whereas in other models, depletion of CAFs accelerates PDAC progression (3). These seemingly conflicting findings may be reconciled by the previously unappreciated heterogeneity of mesenchymal subsets present in PDAC tumors (4). In this issue of Cancer Discovery, work by Elyada and colleagues using comparative single-cell transcriptomics on human and mouse PDAC brings CAF heterogeneity to the fore (5).

Previous work from this group identified two major CAF populations in mouse models of PDAC. One is an inflammatory (iCAF) subset characterized by Ly6C expression, JAK/STAT signaling, and a secretory phenotype; and the second is a myofibroblastic (myCAF) subset characterized by SMA expression, TGFβ signaling, and extracellular matrix (ECM) production (6). These subsets also have unique spatial organization, whereby myCAFs proximally surround tumor ducts and interact with tumor cells through juxtacrine mechanisms, whereas iCAFs reside at greater distances within the stroma and interact with tumor cells, myCAFs, and likely other stromal cells, through inflammatory cytokines. This study by Elyada and colleagues demonstrates that both CAF populations can be observed in human PDAC and in the KPC genetic model of PDAC. Furthermore, these CAF subsets have been independently corroborated by a more limited single-cell analysis of human PDAC and precursor neoplastic lesions, reported by Bernard and colleagues (7). The exact role of these subsets in human disease, their heterogeneity across patients and disease stages, and the impact of therapeutically targeting them remain to be explored.

The Elyada study also goes beyond previous findings to identify a novel CAF subset that the authors named antigen-presenting (apCAF) based on expression of MHC class II and Cd74. The presence of this subset was confirmed by flow cytometric and IHC evaluation of mouse and human PDAC. apCAFs were characterized by activation of STAT1, MTORC1, MYC, and antigen presentation pathways. On a functional level, ex vivo studies demonstrated a capacity for apCAFs to activate CD4⁺ T cells, albeit at much lower levels than professional immune compartment antigen-presenting cells (APC). Moreover, because apCAFs do not express MHCII costimulatory molecules at levels similar to professional APCs, they could not stimulate proliferation of CD4⁺ T cells ex vivo. Nevertheless, the existence of a functional MHCII⁺ CAF subset is intriguing, especially considering the characteristically immunosuppressive and immunotherapy-resistant TME in PDAC.

Fibroblasts are classically known for their roles as mediators of fibrosis and ECM remodeling; however, the identification of putative immunomodulatory CAF subsets in PDAC and other tumor types implies a novel role for CAFs in regulating antitumor immunity. Although this is the first report of MHCII expression on CAFs, MHCII-expressing fibroblasts have been implicated in regulating the development and priming of T cells in the thymus and lymph nodes, and in fibrotic and inflammatory disease–associated fibroblasts in tissues such as the skin and joints. Therefore, it is plausible that apCAFs might serve a unique antigen-presenting role in regulating T-cell priming and activity within the PDAC TME, potentially by acting as decoy APCs or through interactions with professional APCs such as macrophages. apCAFs might also represent an intermediate phenotype between CAFs and macrophages as the populations have shared properties, including expression of ECM remodeling regulators and antigen-presentation mediators.

The extent to which fibroblast heterogeneity is replicated across tissues and disease states remains unknown. To that end, fibroblast heterogeneity has been investigated in other tumor types and in fibrotic diseases. Single-cell transcriptomic analyses of colon cancer patient samples demonstrated the existence of unique CAF subsets, with one demonstrating myofibroblastic gene-expression signatures, similar to myCAFs in PDAC (8). Another study identified a CD10⁺GPR77⁺ CAF subset in human breast cancer and lung adenocarcinoma that drives cancer-initiating phenotypes, chemoresistance, and tumor growth (9). Also in breast cancer, two studies identified four CAF subsets in patient samples and mouse models. The human study utilized flow cytometric resolution of CAF heterogeneity and functional analyses, which demonstrated immunosuppressive properties of several CAF subsets (10). The mouse study demonstrated origin-related diversity in CAF subsets and suggested potential immunomodulatory roles, although the extent of cross-species concordance was not clear (11). One major potential contributor to CAF diversity that remains to be fully explored is their multiple distinct origins. For example, in pancreatic cancer, CAFs can derive from tissue-resident stellate cells, epithelial cells, adipose tissue, and the bone marrow. Therefore, it will be important to evaluate the specific origins of CAFs in PDAC and carcinomas of other tissues, and determine whether these origins confer unique phenotypes and functions.

The existence of CAF phenotypic heterogeneity in PDAC has now been established (Fig. 1), yet several outstanding questions remain. It will be essential to evaluate how cancer therapies shift these subsets and the relative influences of each subset on therapeutic resistance and/or response. Specifically, for the apCAFs identified in the Elyada study, the impact of this subset on tumor immunity will be important to understand whether apCAFs are functionally or quantitatively modulated in response to immunotherapies. CAF subset profiling of patient samples, such as the data provided here, will be critical in defining the extent of interpatient and intratumoral heterogeneity. As our view of fibroblast heterogeneity begins to take shape, questions like these will become critical to understanding the impact of CAF diversity on the pathobiology of tumors.

No potential conflicts of interest were disclosed.