Cancer-Associated Fibroblasts in Esophageal Cancer

Abstract

Cancer-associated fibroblasts (CAFs), a heterogenous population, can promote cancer cell proliferation, migration, invasion, immunosuppression, and therapeutic resistance in solid tumors. These effects are mediated through secretion of cytokines and growth factors, remodeling of the extracellular matrix, and providing metabolic support for cancer cells. The presence of CAFs in esophageal carcinoma are associated with reduced overall survival and increased resistance to chemotherapy and radiotherapy; thus, identifying therapeutic vulnerabilities of CAFs is a necessity. In esophageal cancer, the mechanisms for CAF recruitment, CAF-mediated promotion of tumorigenesis, metastatic dissemination, and therapeutic resistance have yet to be fully evaluated. Here, we provide an overview of the current understanding of CAFs in esophageal cancer, namely in esophageal squamous cell carcinoma and esophageal adenocarcinoma, as well as in the preneoplastic conditions that predispose to these cancers. Interestingly, there is a discrepancy in our knowledge of CAF biology between esophageal cancer subtypes, with very few studies in esophageal adenocarcinoma, and its precursor lesion Barrett’s esophagus, compared with esophageal squamous cell carcinoma. We propose that although great strides have been made, certain questions remain to which answers hopefully will emerge to have an impact on biomarker diagnostics and translational therapeutics.

Summary.

Review of how fibroblasts affect esophageal cancer (squamous cell carcinoma and adenocarcinoma) progression including their role in preneoplastic conditions, primary tumor growth, metastasis, and therapeutic resistance. While addressing the current state of the field, we also highlight current limitations and unresolved questions.

Esophageal cancers, represented primarily by esophageal squamous cell carcinoma (ESCC) and esophageal adenocarcinoma (EAC), are currently the sixth leading cause of cancer deaths worldwide.¹ Because the presence of esophageal cancer risk factors, such as obesity, alcohol consumption, and smoking, continue to plague the global population, the number of deaths per year is predicted to increase.¹ It is estimated that there will be 2 million cases by the year 2040.² Therefore, there is an urgent need to understand the mechanisms of esophageal cancer progression and therapeutic resistance to facilitate the development of new and durable treatment approaches.

Solid tumors are composed of approximately 20%–80% stroma,³ and recent advances have improved our understanding of how the stroma affects tumor progression. The proportion of stroma in solid tumors has been associated consistently with clinical outcomes across a range of cancer types with stroma-high tumors (>50% stroma) associated with reduced survival in colorectal cancer, non–small-cell lung cancer, breast cancer, and esophageal cancer.^4,5 This worsening outcome is likely a consequence of multiple functional roles of the stroma in tumorigenesis such as promotion of tumor cell proliferation, migration, invasion, immune escape, and increased therapeutic resistance, among other factors.⁶ The stroma comprises the noncellular extracellular matrix (ECM) and cellular components such as cancer-associated fibroblasts (CAFs), endothelial cells, pericytes, immune cells, inflammatory cells, adipocytes, and neurons.⁷ The ECM, composed of various matrix proteins (including collagens, laminins, glycoproteins, proteoglycans), is dynamically remodeled during tumor progression, and matrix stiffness has been associated with metastasis, immunosuppression, and therapeutic resistance.^8,9 We direct readers to recent reviews for an in-depth analysis of how the ECM influences cancer progression,10, 11, 12 and specifically how ECM proteins influence esophageal cancer.¹³

CAFs are the most abundant cell type in the stroma and can support tumor progression through matrix remodeling, secretion of growth factors and cytokines, and providing metabolic support to cancer cells through metabolite shuttling.¹⁴ Although most CAF functions are associated with tumor promotion, studies in pancreas cancer have identified that CAFs also can show tumor-restraining functions under certain circumstances.¹⁵ The recent explosion in single-cell RNA-sequencing (scRNA-seq) data have been fundamental to our understanding of the heterogenous nature of CAFs, with numerous CAF subpopulations identified.¹⁶ The myofibroblastic CAF (myCAF) and inflammatory CAF (iCAF) subpopulations, which can be characterized by an ECM and inflammatory signature, respectively, have been identified consistently across a range of cancers.¹⁶ Although myCAF and iCAF subpopulations are studied most extensively, other CAF subpopulations have been identified in a more tissue-specific manner, such as antigen-presenting CAFs in pancreas cancer.¹⁶ This heterogeneity is likely a consequence of multiple factors including the cell of origin (resident fibroblast, endothelial cell, bone marrow–derived mesenchymal stem cell, adipocyte, and pericyte) and the diverse signals and stressors within a tumor. CAFs rarely exhibit genetic or genomic alterations, instead there is a premise in the field that CAFs are regulated at the epigenetic level.¹⁷ The recruitment and reprogramming of CAFs can be modulated by many factors, including cancer cell secretions, cell contact signals, mechanical stress, and physiological stress.¹⁴ Among these signals, the fibroblast growth factors (FGFs) and their receptors (FGFRs), which are upstream of 4 key signaling cascades: phosphoinositide 3-kinase/protein kinase B (PI3K/Akt), rat sarcoma (RAS)/mitogen-activated protein kinase, Janus kinase/signal transducer and activator of transcription (JAK/STAT), and phospholipase C gamma (PLCγ), and have been shown to be important factors in the activation of CAFs.¹⁸ Interestingly, FGFs derived from CAFs can drive tumorigenesis in multiple cancers including lung,¹⁹ pancreatic,²⁰ ovarian,²¹ breast,²² and colorectal.²³ Despite the nomenclature, FGF–FGFR signaling has implications beyond fibroblasts in the tumor microenvironment (TME); specifically, they are implicated in the promotion of immunosuppression and immune cell escape, which has been reviewed in depth.¹⁸ As such, the development of FGFR inhibitors offers potential therapeutic options to target both CAF and immune cell–mediated effects in the TME.¹⁸

The need to understand how CAFs are recruited, the functional role of each CAF subpopulation, and, ultimately, the therapeutic vulnerabilities of CAFs have been highlighted by recent reviews detailing how CAFs promote radiotherapy resistance,²⁴ chemotherapy resistance,²⁵ and metastatic spread.²⁶ This review details our current understanding of how CAFs affect esophageal cancer progression and therapeutic resistance and highlights key unresolved questions that still need to be addressed in the field (Figure 1).

Figure 1.

The known roles of fibroblasts in esophageal cancer progression and remaining unresolved questions. GERD, gastroesophageal reflux disease; PI3K, phosphoinositide 3-kinase.

Image created with BioRender.com.

The Role of Fibroblasts in Preneoplastic Esophageal Disease

The presence of squamous dysplasia and intestinal metaplasia, the latter referred to as Barrett’s esophagus (BE), are common precursor lesions in ESCC and EAC, respectively.²⁷ In a long-term study of 682 patients, Wang et al²⁸ showed that the presence of squamous dysplasia increases the probability of ESCC development, with the highest relative risk in those patients presenting with high-grade dysplasia. In oral squamous cell carcinoma and laryngeal carcinoma, a high number of infiltrating α-smooth muscle actin–positive fibroblasts was observed in invasive squamous cell carcinomas, but not in the corresponding dysplastic tissues.^29,30 Studies investigating the role of fibroblasts in esophageal squamous dysplasia and ESCC initiation have been very limited, but some reports have suggested that fibroblasts may inhibit dysplasia. Loomans et al³¹ illustrated that stromal-derived Activin A could inhibit invasion of dysplastic esophageal squamous cells. Further studies are required to identify the functional role of dysplasia-associated fibroblasts in esophageal squamous dysplasia progression and ESCC development.

Barrett’s esophagus is characterized by the replacement of esophageal squamous epithelium with a columnar epithelium and is a consequence of chronic bile acid exposure and inflammation caused by gastroesophageal reflux disease.²⁷ The infiltration of activated fibroblasts is observed in dysplastic BE.³² This finding was confirmed by scRNA-seq, which identified an increase in ECM-producing and contractile fibroblasts in BE compared with adjacent normal epithelium (Table 1).³³ Treatment of fibroblasts with nitric oxide or acid exposure to mimic gastroesophageal reflux disease inhibits the wound healing ability of fibroblasts by inhibition of Rho-Rho-associated kinase(ROCK) signaling and activates nuclear factor-κB (NF-κB) signaling, respectively.^34,35 Interesting studies in gastric metaplasia have identified specific fibroblast subpopulations associated with this condition³⁶ and scRNA-seq indicates that similar populations also are present in BE.³³ Intriguingly, co-culture of gastric metaplastic organoids with fibroblasts derived from either gastric metaplasia or gastric cancer can induce a dysplastic transition in the organoids.³⁶ It is unknown if BE-associated fibroblasts can induce dysplasia, but it is a question that could be addressed.³⁷ Certain outstanding questions remain. What role do these BE-associated fibroblasts play in BE pathogenesis and EAC tumorigenesis? How does chronic exposure to bile acids and the resulting inflammation affect fibroblast identity and plasticity? The use of coculture models, specifically 3-dimensional (3D) organotypic and 3D organoid cultures,^38,39 should aid future investigations into the role of fibroblasts in esophageal cancer precursor lesions based on the development of these model systems by our group.

Table 1.

List of Fibroblast Subpopulations Identified Across BE, ESCC, and EAC by Single-Cell RNA Sequencing Projects

Esophageal histology	Tissues analyzed	CAF subpopulations	CAF subpopulation defining markers	Predicted function	Additional notes
Barrett’s esophagus33	Normal esophagus, stomach, and intestine tissues vs esophageal and gastric metaplasia tissues	S1: early immune	PDGFRAhi, PDPNhi, CD34hi	Immune cell recruitment	-7 clusters identified by scRNA-seq that were classified as 3 subtypes -Expansion of S2/S3 populations in BE
		S2: desmoplastic/ECM depositing	PDGFRAhi, PDPNhi, CD34lo	ECM remodeling and promotion of angiogenesis-
		S3: contractile	ACTA2, MYL9, ACTG2	Active contractile fibroblasts-
ESCC40	Surgical resections of ESCC and adjacent normal tissue	CAF 1 (iCAF)	CCL2, CCL11, CXCL14	Inflammatory and immune cell recruitment	-CAF 2 expressed activated pericyte-specific cytokines -CAF 3 and 4 both have up-regulation of EMT pathways
		CAF 2 (iCAF)	CXCL5, CXCL8, CSF3	Inflammatory and immune cell recruitment-
		CAF 3 (myCAF)	MMP1, MMP11, TAGLN, PDGRB	ECM remodeling-
		CAF 4 (myCAF)	MMP11, TAGLN, ACTA2, ACTG2, PDGFRB	ECM remodeling and promotion of angiogenesis-
ESCC41	Surgical resections of ESCC and adjacent normal tissue	Fibroblast 1	GSN, CFD, IGFBP6	Not predicted by authors	7 fibroblast clusters identified with F4 the only cluster increased markedly in ESCC
		Fibroblast 2	MMP1, TFPI2, THBS2, PTGDS-
		Fibroblast 3	HLA-DRA, CD74, HLA-DRB1	ECM organization-
		Fibroblast 4	POSTN, SPARC, TAGLN	Not predicted by authors-
		Fibroblast 5	FOSB, JUN, ATF3-
		Fibroblast 6	RGS1, CXCR4, IL7R-
		Fibroblast 7	CRYAB, FGL2,-
ESCC42	Surgical resections of ESCC primary tumor, adjacent normal and metastatic lymph node	CAF 1: antigen presenting	HLA-DRA, CD74, HLA-DPB1	Modulate tumor immunity	myCAFs represented approximately 50% of fibroblasts from primary tumor and metastatic lymph node
		CAF 2: iCAF	IL6, HAS1, COL14A1	Inflammatory and immune cell recruitment-
		CAF 3: myCAF	THBS2, COL12A1, THY1, POSTN, MMP1	ECM organization, support adhesion and migration of epithelial cells-
		CAF 4: myCAF	THBS2, COL12A1, THY1, IGKC	ECM organization-
EAC43	Surgical resections of EAC and adjacent normal tissue obtained from chemotherapy-treated and treatment-naive patients	CAF 1	COL1A2	Not predicted by authors	-7 clusters identified, but CAF 1 and 2 are the predominant subtypes -Chemotherapy-induced transcriptional alterations were observed in CAF 1 and 2 -Increase in aCAF and myCAFs after chemotherapy
		CAF 2	IER2
		CAF 3: complement expression CAFs	MFAP5
		CAF 4/5: aCAFs	PTGDS
		CAF 6/7: myofibroblasts (myCAFs)/smooth muscle cells	MYH11, STEAP4

NOTE. All projects were completed using the 10x genomics scRNA-sequencing platform.

aCAF, adipogenic CAF; ACTA2, alpha smooth muscle actin; ACTG2, gamma smooth muscle actin; ATF3, activating transcription factor 3; CCL, chemokine (C-C motif) ligand; CFD, complement factor D; COL, collagen type; CRYAB, crystallin alpha B; EMT, epithelial-to-mesenchymal transition; FGL2, fibrinogen like 2; FOSB, FosB proto-oncogene, AP-1 transcription factor subunit; GSN, gelsolin; HAS1, hyaluronan synthase 1; HLA-DRA, major histocompatibility complex, class 2, DR alpha; HLA-DRB1, major histocompatibility complex, class 2, DR beta 1; HLA-DPB1, major histocompatibility complex, class 2, DP beta 1; IER2, immediate early response 2; IGFBP6, insulin like growth factor binding protein 6; IGKC, immunoglobulin kappa constant; JUN, Jun proto-oncogene, AP-1 transcription factor subunit; MFAP5, microfibril associated protein 5; MMP, matrix metallopeptidase; MYH11, myosin heavy chain 11; MYL9, myosin light chain 9; PDGFRA, platelt derived growth factor receptor alpha; PDGFRB, platelet derived growth factor receptor beta; PDPN, podoplanin; POSTN, periostin; PTGDS, prostaglandin D2 synthase; RGS1, regulator of G protein signaling 1; SPARC, secreted protein acidic and rich in cysteine; STEAP4, six-transmembrane epithelial antigen of prostate 4; TAGLN, transgelin; TFPI2, tissue factor pathway inhibitor 2; THBS2, thrombospondin-2; ; THY1, Thy-1 cell surface antigen.

The Role of Fibroblasts in Esophageal Squamous Cell Carcinomas

What Is the Origin of ESCC CAFs?

Elegant transplantation models have shown that most CAFs are recruited from the local microenvironment,⁴⁴ although the precise identity of these precursor cells in esophageal cancer is not fully understood. In ESCC, tumor cells can reprogram normal fibroblasts, mesenchymal stem cells, and FGFR2+ fibrocytes into CAFs through release of cytokines (chemokine (C-X-C motif) ligand [CXCL]1, interleukin [IL]6, tumor necrosis factor-α [TNF-α], CXCL12), growth factors (transforming growth factor-β [TGF-β]), other secreted proteins (plasminogen activator, urokinase [PLAU]), and noncoding RNAs (long noncoding RNA POU class 3 homeobox 3 [POU3F3], microRNA-21).^45,46 Projection analysis of scRNA-seq data indicate a pericyte to myofibroblast transition also can occur in ESCC.⁴⁰ A recent publication showed that loss of annexin A1 from epithelial cells, which occurs during ESCC progression owing to transcription repression, promotes CAF activation.⁴⁷ This variety in recruitment may explain the different CAF populations that have been identified in esophageal cancer (Table 1). Comprehensive analysis identified 7 fibroblast clusters across ESCC and normal esophageal epithelium, but only 1 cluster that expressed common myCAF markers (periostin [POSTN], transgelin [TAGLN], secreted protein acidic and rich in cysteine [SPARC]) was enriched in ESCC.⁴¹ Zhang et al⁴⁰ identified 4 CAF subpopulations in ESCC. Two were categorized as iCAF and 2 as myCAF, which raises the question: do all CAF subpopulations exist in a continuum between iCAF and myCAF? Intriguingly, there is evidence from pancreatic cancer that CAFs are very dynamic and can transition bidirectionally between iCAFs and myCAFs depending on the tumor microenvironmental cues.⁴⁸ This lineage plasticity in CAFs makes investigations, particularly in vitro, technically challenging because fibroblasts will adjust to environmental conditions and thus may not represent the heterogenous in vivo CAF population.¹⁴ The use of 3D coculture models, including culturing of CAFs with organoids and in organotypic cultures,³⁶ improves biological relevance, but there is still a lack of in vitro models that allow investigators to re-create and analyze the function of each CAF subpopulation. The current gold standard to identify CAF subpopulations remains scRNA-seq of dissociated tissue followed by either multiplex immunofluorescence or spatial transcriptomics to determine the localization of CAFs within the tumor. Indeed, the spatial relationship between CAF populations and other components of the tumor (cancer cell, immune cell, blood vessels) is key to understanding the role of these populations.⁴⁹

How Do CAFs Promote ESCC Progression?

The functional consequence of CAF-tumor interactions in the primary tumor includes increased cancer cell proliferation, migration, and invasion.¹⁴ Cytokine arrays have identified numerous cytokines and growth factors that are increased upon in vitro coculture of CAFs with ESCC cells such as IL6, chemokine (C-C motif) ligand 5 (CCL5), IL8, CXCL1, TGF-β, and many more.^45,50,51 This diversity in signals results in activation of multiple signaling pathways including Akt, STAT3, extracellular signal-regulated kinase 1/2 (ERK1/2), and nuclear factor-κB, thereby resulting in protumorigenic behaviors such as increased cancer cell proliferation and migration.^45,50,51 The abundance of CAF-derived cytokines hints at the ability of CAFs to modulate the local immune environment. Indeed, ESCC CAFs can promote the polarization and migration of macrophages via cytokine (CCL2, IL6, CXCL8) secretion, and the presence of CAFs was shown to correlate with a reduction in tumor-infiltrating lymphocytes.^14,45 Separately, high FGF-2 expression in the stroma has been correlated with increased macrophage (CD163+ and CD204+) infiltration and increased local invasion in ESCC, however, the underlying mechanisms have not been elucidated.⁵² In addition, CAFs can support tumorigenesis through remodeling of the ECM¹³ and metabolite reprogramming.¹⁴ Exchange of metabolites and amino acids between tumor cells and CAFs can help support the metabolic demands of rapidly growing tumors. A recent scRNA-seq analysis of ESCC described that different CAF subpopulations had distinct metabolic signatures,⁵³ which may reflect the local environmental conditions that CAFs were exposed to such as nutrient deprivation or hypoxia. Cytokine arrays have identified many proteins involved in CAF-tumor communication, but the complexity of the CAF and ESCC tumor cell secretomes, as identified by mass spectrometry,⁵⁴ suggests there is more to learn. CAFs and tumor cells also can communicate by secretion of noncoding RNAs, DNA, and lipids.⁵⁵ Intriguingly, the mutational burden of the tumor cell, especially the presence of mutant tumor protein p53 (p53), has been shown to alter the tumor cell secretome.⁵⁶ In non–small-cell lung cancer, the type of driver mutation (p53 or epidermal growth factor receptor [EGFR]) correlated with the abundance of specific CAF subpopulations.⁵⁷

Metastases are responsible for most cancer deaths and esophageal cancers tend to metastasize early. Thus, a better understanding of how CAFs contribute to metastasis and the identification of therapeutic strategies to negate their effect is needed. Kashima et al⁵⁸ showed that high numbers of fibroblast activation protein (FAP)+ CAFs within an ESCC primary tumor was associated with lymph node metastases and decreased survival. Co-implantation of ESCC cell lines and human fetal esophageal fibroblasts, which were used as a surrogate for CAFs, in orthotopic mouse models increased metastasis compared with ESCC cells alone.⁵⁸ Recent studies have shown that CAFs can promote both lymphangiogenesis⁵⁹ and angiogenesis⁶⁰ in ESCC, which can increase the ability of cancer cells to enter both the lymphatic and blood circulation. Once tumor cells enter the circulation, forming complexes with CAFs can increase cell survival and improve colonization at metastatic sites,⁶¹ although this has yet to be shown in esophageal cancers. In summary, CAFs can promote metastasis through increasing tumor cell migration and invasion, promotion of lymphangiogenesis and angiogenesis, and increasing the ability of circulating tumor cells to survive and colonize. Jia et al⁴² used scRNA-seq to compare the fibroblast populations in ESCC primary tumors and lymph node metastases. In metastatic lymph nodes there was a higher proportion of pericyte-derived CAFs and the lymph node–derived myCAFs had up-regulated multiple genes involved in monocyte differentiation and fibroblast proliferation.⁴² A direct comparison between primary and distant metastatic CAFs is more challenging because distant metastatic CAFs likely will be recruited from the resident cells of that organ, such as lung or liver fibroblasts, rather than travel from the primary tumor. Thus, are differences reflective of a different cell of origin or inherent differences between primary and metastatic tumors?

How Do CAFs Mediate Therapeutic Resistance in ESCC?

Resistance to chemotherapy and radiotherapy is a major obstacle in esophageal cancer treatment with only a small proportion of patients achieving a complete pathologic response to neoadjuvant chemoradiotherapy.⁶² Therapeutic resistance is multifactorial, with CAFs representing a key factor. CAF-secreted cytokines (TGF-β and IL6) and growth factors (FGFs) have been shown to confer chemoresistance in ESCC.^45,63 Adjuvant treatment with an IL6 inhibitor could abrogate the CAF-induced cisplatin resistance in ESCC cell lines.⁶⁴ Interestingly, CAF-derived IL6, via activation of STAT3 signaling, was reported to promote the differentiation of monocytic myeloid-derived suppressor cells, which are associated with cisplatin resistance and poor overall survival.⁶⁵ Several studies have shown that CAF-conditioned medium can inherently promote chemotherapeutic resistance in ESCC cell lines, but this resistance is enhanced significantly if CAFs have been exposed to cisplatin, suggesting chemotherapy induced resistance mechanisms.⁶⁶ After cisplatin exposure, CAFs increased secretion of plasminogen activator inhibitor-1 (PAI-1), a TGF-β target gene, which can reduce apoptosis and promote cisplatin resistance.⁶⁶ Importantly, treatment with a PAI-1inhibitor improved cisplatin sensitivity in vivo.⁶⁶ The ability of CAF-conditioned media to promote chemotherapeutic resistance in ESCC has been reported consistently, but there has yet to be a comprehensive investigation of the ESCC CAF secretome and although this work would be complex it would enhance our understanding of how CAFs promote therapeutic resistance.

The role of CAFs in radioresistance has not been fully explored but there is evidence that CAF-conditioned media confers resistance to radiotherapy in solid tumors, which may be caused by inducing cancer stem cells within the tumor.²⁴ This induction of cancer stem cells, and subsequent radioresistance, has been described as another function of CAF-secreted TGF-β.²⁴ In ESCC, CAFs can promote radiotherapy resistance by reducing the levels of superoxide dismutase 1, thus increasing reactive oxygen species and enhancing DNA damage repair.⁶⁷ This effect was modulated by CAF-derived CXCL1, which also may confer radiotherapy resistance through activation of mitogen-activated protein kinase kinase (MEK)/ERK signaling.⁶⁷ Incubation of ESCC cell lines with CAF-conditioned media increased expression of the long noncoding RNA dynamin 3 opposite strand (DNM3OS), which can increase radiotherapy resistance in ESCC cell lines both in vitro and in vivo.⁶⁸ The induction of DNM3OS was abrogated by inhibition of the platelet derived growth factor receptor beta/forkhead box O1 (PDGFRβ/FOXO1) signaling axis.⁶⁸ As expected, exposure of CAFs to ionizing radiation alters the CAF transcriptome and secretome,²⁴ but the importance of these alterations in radiotherapy resistance has not been established.

Another important consideration for CAF-mediated therapeutic resistance is understanding how CAFs remodel the ECM to reduce therapeutic efficiency.⁶⁹ This was shown by treating solid tumors with FAP targeting chimeric antigen receptor (CAR) T-cell therapy, which after loss of FAP+ CAFs, forces ECM remodeling to allow immune cell infiltration, thus boosting the performance of immunotherapy.⁷⁰ These insights highlight the importance of CAFs as architects of the TME and emphasize the many potential mechanisms of therapeutic resistance CAFs may be orchestrating. Given their ability to remodel the ECM, support cancer cells metabolically, and generate an immunosuppressive environment, do CAFs alone promote therapeutic resistance? Or do chemotherapeutic agents and radiotherapy induce alterations in CAF transcriptome and secretome signatures that promote resistance? If the first supposition is true, then any key factor required for CAF recruitment and maintenance could be a potential therapeutic target to overcome CAF-induced resistance.

The Role of Fibroblasts in Esophageal Adenocarcinoma

To date, most esophageal CAF studies have focused on ESCC, and, hopefully, future studies will involve studies in EAC-related CAFs as well. Proteomic analysis of matched normal and EAC-derived fibroblasts have illustrated the heterogeneity of EAC CAFs and shown, as expected, pathway enrichment for ECM organization, migration, and angiogenesis.⁷¹ Croft et al⁴³ identified 2 predominant CAF populations in EAC that displayed transcriptomic alterations after chemotherapy (Table 1). Similar to ESCC, conditioned media from EAC CAFs can confer chemotherapy and radiotherapy resistance in EAC with CAF-secreted IL6 serving as a key mediator of therapeutic resistance through the induction of epithelial-to-mesenchymal transition in EAC cells.⁷² Inhibiting the myCAF subpopulation, through inhibition of phosphodiesterase type 5, which is expressed in both the esophageal epithelium and stroma, improved the sensitivity of EAC 3D models to chemotherapy.⁷³ In addition to promoting therapeutic resistance, EAC-derived CAFs can promote EAC invasion through secretion of periostin, which binds EAC cell integrins and activates PI3K-Akt signaling.⁷⁴ It is likely that EAC and ESCC CAFs have overlapping functions and therapeutic vulnerabilities, but more EAC CAF-specific investigations are required to elucidate common and divergent properties. Indeed, the pathogenesis of EAC, specifically the role of chronic bile acid exposure and the metaplastic changes observed in BE, would suggest the EAC CAFs may have unique features such as acid-induced transcriptome and functional alterations.

Future Directions

The tumor microenvironment is exceedingly complex and although great strides have been made, there are many unresolved questions. Often, the model systems have limited investigations, but the development of 3D coculture models as a more biologically representative model should aid future studies into CAF–tumor cell (and other cell types) communication and functionality. The use of in vivo models with an active immune system, such as in genetic or syngeneic mouse models, will facilitate the investigation of how CAFs modulate the local immune environment and vice versa. Recent scRNA-seq projects have delineated the CAF subpopulations and transcriptome signatures in esophageal cancer,^{33,40,41,42,43} and now we require mechanistic studies to establish functionality and nomination of targets that may be therapeutically feasible. How do these CAFs affect tumor progression, metastatic dissemination, and therapeutic resistance? How does the mutational burden of a tumor affect CAF recruitment and functionality? What is the interplay between CAFs and immune cells in esophageal cancer progression? Can we identify new therapeutic approaches to target CAFs? Our knowledge of CAFs has increased exponentially in the past decade and, as such, we are hopeful that the field will be able to address these challenges and move CAF-targeting therapeutics into the clinic.