Initially, myCAFs were discerned from pancreatic ductal adenocarcinoma, delineated as contractile matrix-producing fibroblasts (FAP+ αSMAhigh CAFs), showcasing heightened expression of ACTA2 and TGFβ response genes including CTGF and COL1A1 (Öhlund et al., 2017). The advent of scRNA-seq has uncovered a similar fibroblast phenotype across diverse solid human tumors. However, transcriptional profiles attributed to myCAFs exhibit variations across studies. Broadly, myCAFs characterization via scRNA-seq can be grouped into two categories.

In the first category, myCAFs are distinguished as a distinct cell population typified by elevated expression of the blood vessel wall gene RGS5, contractility genes (MYH11, ACTA2, TAGLN, MYL9), and other pericyte markers (NDUFA4L2, ADIRF) (Elyada et al., 2019; Chen et al., 2020; Galbo et al., 2021; Pu et al., 2021; Li t al., 2022; Chen et al., 2022; Hu et al., 2022; Luo et al., 2022; Obradovic et al., 2022; Yang et al., 2022; Zhao et al., 2022; Liu et al., 2023; Qin et al., 2023). Functionally, these cells are implicated in processes such as angiogenesis, smooth muscle contraction, and vascular wound healing. Apart from being referred to as myCAFs, they are also labeled as contractile CAFs (Che et al., 2021; Giguelay et al., 2022), stellate-like CAFs (Wang et al., 2021), MCAM+ CAFs (Wang et al., 2023), RGS5+ CAFs (Olbrecht et al., 2021; Zhang et al., 2023) and MYH11+ CAFs (Olbrecht et al., 2021). However, evidence suggests that a similar transcriptional profile is characteristic of pericytes and other mural cells (van Splunder et al., 2024). Consequently, several scRNA-seq investigations propose categorizing this population as pericytes rather than fibroblasts (Lin et al., 2020; Zhang et al., 2021; Bischoff et al., 2022; Li et al., 2022; Cords et al., 2023; Werba et al., 2023). Furthermore, scRNA-seq analysis of normal muscular organs discerned a 90-gene signature capable of distinguishing fibroblasts from mural cells (Muhl et al., 2020). Consistent with this, heightened expression of genes including MYL9, MYH11, MCAM, TAGLN, and ACTA2 suggests affiliation with mural cells. This raises the question: does a distinct population of RGS5 + MYH11+ CAFs truly exist or they should be considered as pericytes or other mural cells? Recent studies suggest that the analysis of scRNA-seq data allows to separate these two similar types of cells within tumors: pericytes and vascular CAFs (vCAFs), likely originating from vascular cells (Cords et al., 2023). Notably, vCAFs, akin to pericytes, exhibit elevated expression of angiogenesis-related genes (MYH11, ACTA2, MCAM, and ADIRF) (Cords et al., 2023). However, RGS5 gene expression is much lower in vCAFs compared to pericytes. Similar to pericytes, vCAFs localize around endothelial cells within vascular structures. Nevertheless, accurate classification of all RGS5 + MYH11+ cell populations is imperative for further exploration of vascular tumor cells and CAFs.

The second category of myCAFs, also termed as such in studies, is typified by expression of collagen-encoding genes (COL1A1, COL10A1, COL11A1, CTHRC1, etc.), matrix metalloproteinases (MMP11), non-collagenous ECM genes (POSTN, THBS2, and VCAN, FN1), and signal transduction genes (SULF1, INHBA) (Kieffer et al., 2020; Lin et al., 2020; Wang et al., 2021; Bischoff et al., 2022; Loret et al., 2022; Liu et al., 2023; Werba et al., 2023). These fibroblasts, also referred to as matrix CAFs (mCAFs) (Zhang et al., 2020; Cords et al., 2023; Ma et al., 2023), ECM-remodeling CAFs (Che et al., 2021; Giguelay et al., 2022), desmoplastic CAFs (dCAFs) (Galbo et al., 2021), TGF-β-activated CAFs (TGFB CAFs) (Hornburg et al., 2021), classical CAFs (cCAFs) (Chen et al., 2021), and others (Olbrecht et al., 2021; Deng et al., 2022; Pan et al., 2023), are intricately associated with ECM organization and remodeling, collagen production, TGF-β response and pathways governing the epithelial to mesenchymal transition (EMT). Studies suggest that these fibroblasts arise from resident fibroblasts and are localized within collagen-rich regions of the tumor (Bartoschek et al., 2018; Zhang et al., 2020; Cords et al., 2023). The upregulation of genes linked to TGF-β-induced reactive stroma implies a pivotal role for TGF-β in activating resident fibroblasts and shaping the myCAF phenotype (Bates et al., 2015; Biffi et al., 2019; Kieffer et al., 2020; Hornburg et al., 2021; Giguelay et al., 2022). Importantly, TGF-β1-activated fibroblasts accelerate cancer cell motility and invasion (Yeung et al., 2013; Huang et al., 2021; Yoon et al., 2021). The localization of matrix-associated myCAFs at the invasive tumor front further underscores their role in tumor invasion (Zhang et al., 2020).

The variability in characteristics among myCAFs contributes to ambiguity when comparing results across different studies. Hence, caution is warranted in analyzing the RGS5+MYH11+ cell population within tumors, emphasizing the necessity to confirm whether this population represents pericytes. Given that the term myCAFs was initially coined for fibroblasts capable of producing a collagen matrix, we propose categorizing the second group of cells as myCAFs.

The expression of markers characteristic of myCAFs serves as a distinguishing feature among various CAF populations. However, the specificity of this gene signature to myCAFs amidst all tumor cell populations remains a subject of inquiry. scRNA-seq studies indicate that the expression of most myCAFs marker genes is primarily confined to fibroblasts within the tumor milieu (Lambrechts et al., 2018; Elyada et al., 2019; Chen et al., 2020; Wang et al., 2021; Zhang et al., 2021; Che et al., 2021; Olbrecht et al., 2021; Pu et al., 2021; Luo et al., 2022; Zhao et al., 2022; Qin et al., 2023; Zhang et al., 2023; Ma et al., 2023; Wang et al., 2023; Werba et al., 2023). Furthermore, analysis of immunohistochemistry staining (IHC) images from the Human Protein Atlas database (Uhlen et al., 2017) reveals predominant localization of proteins such as COL3A1, POSTN, VCAN, FN1, COL12A1 in the stroma of solid tumor types (Supplementary Figure S1). These findings suggest that the myCAF gene signature is not characteristic of any other cell population within the tumor and its microenvironment. Consequently, alterations in the expression of this gene signature in tumor tissues likely reflect changes in the abundance of the myCAF population.

According to bulk RNAseq analyses, myCAF markers are predominantly upregulated at transcriptional level in tumor tissues compared to normal tissues across various cancer types (Figure 2; Supplementary Table S1). These observations are further substantiated at the protein level through comparisons of IHC staining patterns in tumor and normal tissues (Figure 2; Supplementary Table S1). Notably, the myCAF gene signature exhibits a significant correlation with poor survival outcomes among patients diagnosed with stomach adenocarcinoma, kidney carcinoma, mesothelioma, low-grade glioma, cervical squamous cell carcinoma, bladder urothelial carcinoma, skin cutaneous melanoma, lung carcinoid tumors and prostate cancer (Galbo et al., 2021; Bischoff et al., 2022; Liu et al., 2023; Pan et al., 2023). Furthermore, elevated expression of most myCAFs marker genes correlates with poor overall survival across various cancer types (Figure 2; Supplementary Table S2). Thus, the heightened presence of the myCAF population emerges as a characteristic feature of tumor tissues and is associated with an unfavorable prognosis.

The myCAF population plays a pivotal role within the tumor stromal microenvironment, actively participating in the remodeling of the ECM, a process intricately linked to the development of chemoresistance. One such modification characteristic of tumor tissues is ECM stiffening, primarily caused by the increased deposition of type I collagen and fibronectin (Cox and Erler, 2011). These proteins are coded by genes COL1A1, COL1A2, and FN1, primarily expressed in myCAFs. This stiffening of the ECM poses significant challenges to effective capillary transport of therapeutic agents into tumor tissues, owing to the resultant elevation in interstitial fluid pressure (Figure 3) (Stylianopoulos et al., 2018). Furthermore, myCAF-mediated ECM stiffness serves as a driver of angiogenesis, further fueling tumor progression (Figure 3). Inhibition of fibroblast activity using renin-angiotensin system inhibitors results in decreased ECM stiffness and enhances the anti-angiogenic effects of bevacizumab in patients with metastatic colorectal cancer (Shen et al., 2020). Notably, in various cancers such as pancreatic and breast cancer, the heightened ECM stiffness contributes to resistance against chemotherapeutic agents like paclitaxel and doxorubicin, respectively, through the induction of EMT processes (Figure 3) (Rice et al., 2017; Joyce et al., 2018). Moreover, the rigidity of the ECM may be linked to the regulation of DNA double-strand break repair processes, thereby influencing cellular sensitivity to genotoxic agents, including etoposide and cisplatin (Deng et al., 2020). In breast cancer cells growing on low stiffness ECM, the activation of MAP4K4/6/7 kinases and subsequent ubiquitin phosphorylation occurs. This leads to the impairment of RNF8-mediated ubiquitin signaling and double-stranded DNA repair deficiency (Figure 3) (Deng et al., 2020).

Another critical mechanism underlying drug resistance, facilitated by myCAFs, involves the integrin-mediated adhesion of tumor cells to ECM components, reinforcing linkages between integrins and the cytoskeleton, consequently leading to elevated formation of focal adhesions (Valdembri and Serini, 2021). The serine/threonine kinase Akt signaling pathway in cancer cells emerges as a key player in this integrin-mediated resistance mechanism. ECM proteins produced by myCAFs, such as collagens, POSTN and FN1, interact with integrins on the surface of tumor cells, thereby activating Akt signaling (Figure 3) (Elango et al., 2022; Wang et al., 2022; Sun et al., 2023). For instance, co-cultivation of ovarian cancer cells with CAFs or on the plates pre-coated with COL11A1 attenuates cisplatin-induced apoptosis (Rada et al., 2018; Nallanthighal et al., 2020). This effect is achieved by activating the Akt signaling pathway through the interaction of COL11A1 with DDR2 receptor tyrosine kinase and α1β1 integrin on the cell surface of ovarian cancer cells. In turn, Akt activates downstream NF-kB signaling, leading to the induction of the expression of three antiapoptotic proteins, including XIAP, BIRC2, and BIRC3, and decreasing levels of cleaved caspase-3 (Figure 3) (Rada et al., 2018). Moreover, Akt signaling leads to the phosphorylation of 5′ AMP-activated protein kinase and upregulates mitochondrial fatty acid oxidation, which is necessary for maintaining ovarian cancer cell survival (Figure 3) (Nallanthighal et al., 2020).

Another downstream mechanism activated by Akt leads to the upregulation of the antiapoptotic protein BCL-2 and the downregulation of proapoptotic protein BAX. Altered balance between BCL-2 and BAX suppressed the apoptotic program by reduction in cleaved caspase-3/9 expression (Figure 3). This mechanism mediates the resistance of pancreatic cancer cells cultivated on COL11A1-coated plates to gemcitabine (Wang et al., 2021) and the resistance of ovarian cancer cells exposed to recombinant POSTN or co-cultured with POSTN-overexpressing fibroblasts to cisplatin (Chu et al., 2020). Furthermore, FN1/PI3K/Akt signaling contributes to breast cancer cell resistance against docetaxel by inhibiting the ASK1/p38 apoptotic pathway (Figure 3) (Xing et al., 2008). Activation of PI3K/Akt/αvβ5 signaling axis by unfolded type III domain of FN1 protects non-small cell lung cancer cells from apoptosis induced by TNF-related apoptosis inducing ligand (Cho et al., 2016).

ECM protein-integrin binding can activate other signaling cascades besides Akt in tumor cells. For instance, FN1 binding to β1 integrin on colorectal cancer cells activates FAK/CDC42/YAP signaling, leading to upregulation of the transcription factor SOX2 and consequent resistance to 5-fluorouracil and cisplatin (Figure 3) (Ye et al., 2020). Additionally, stromal FN1 promotes both FAK/RAS/MEK/ERK and FAK/PI3K/Akt pathways in breast cancer cells, leading to estrogen receptor-α phosphorylation and subsequent tamoxifen sensitivity (Figure 3) (Pontiggia et al., 2012). Furthermore, the interaction of ECM proteins with integrins significantly impacts the activity of ATP binding cassette (ABC) transporters, which recognize cytostatic drugs and export them from the cytosol. Cultivating breast cancer cells on the microplates coated with collagen I and fibronectin leads to enhanced resistance to doxorubicin, cisplatin, and mitoxantrone, due to increased activity of ABC efflux transporters (Figure 3) (Baltes et al., 2020).

myCAFs also contribute to anti-cancer therapy resistance through mechanisms beyond ECM remodeling. For instance, myCAFs enhance proliferation, migration, and resistance to carboplatin and cisplatin in cervical squamous cell carcinoma via the NRG1/ERBB3 pathway activation (Figure 3) (Liu et al., 2023). Furthermore, the expression of the myCAF-specific gene signature correlates with poor response to anti-PD-1 immunotherapy in patients with urothelial carcinoma, melanoma, and non-small cell lung cancer, potentially mediated by myCAF-induced modulation of the immune microenvironment (Kieffer et al., 2020; Pan et al., 2023). The impact on the effectiveness of anti–PD-1 immunotherapy may be explained by the ability of myCAFs in vitro to increase the proportion of CD4⁺ CD25⁺ FOXP3⁺ Tregs among CD4⁺CD25⁺ T lymphocytes and elevate the levels of PD-1 and CTLA4 proteins on the surface of CD4⁺ CD25⁺ FOXP3⁺ Tregs (Figure 3) (Kieffer et al., 2020).

In summary, the myCAF population serves not only as a prognostic indicator but also as a promising therapeutic target. Other myCAF markers, including INHBA, SULF1, CTHRC1, VCAN, and COL5A1, have also been implicated in influencing resistance to anticancer drugs (Table 1). A significant limitation of these studies is the use of tumor cells for investigation of myCAFs markers in the development of resistance to various anti-tumor medications. Given that this gene signature is characteristic of CAFs, not tumor cells, conducting similar studies using CAFs could provide greater clarity on acquired chemoresistance in tumors.

It is widely recognized that CAFs play a crucial role in shaping the unique tumor immune microenvironment, facilitating tumor immune evasion (Mhaidly and Mechta-Grigoriou, 2021). Among the diverse CAF subtypes, iCAFs stand out for their distinct immunomodulatory secretory profile (Figure 1). iCAFs are characterized by their expression of various genes, including members of the chemokine ligand family (CXCL12 and CXCL14), the insulin-like growth factor family (IGF1 and IGFBP6), complement-regulatory genes (C3, C7, and CFD), and some other genes (APOD, FBLN1, PTGDS, DPT, and GSN) (Elyada et al., 2019; Zhang et al., 2020; Chen et al., 2020; Wang et al., 2021; Galbo et al., 2021; Pu et al., 2021; Li et al., 2022; Chen et al., 2022; Li et al., 2022; Hu et al., 2022; Loret et al., 2022; Obradovic et al., 2022; Yang et al., 2022; Liu et al., 2023; Cords et al., 2023; Ma et al., 2023; Pan et al., 2023; Werba et al., 2023).

The functions attributed to iCAFs encompass a broad spectrum, ranging from involvement in inflammatory responses and complement cascades to mediating cytokine-driven signaling and fostering cancer cell proliferation. Unlike myCAFs, iCAFs are located more distant from cancer cells (Öhlund et al., 2017; Wu et al., 2020). It has also been shown that iCAFs encircle vCAFs and the endothelial cells lining the tumor blood vessels (Cords et al., 2023). Notably, the induction of the iCAF phenotype predominantly occurs through the IL1-IL1R1 axis and can occur independently of direct contact with cancer cells (Biffi et al., 2019), leading researchers to frequently refer to them as IL1+ CAFs in scientific investigations (Dominguez et al., 2020; Hornburg et al., 2021).

Based on scRNA-seq data, iCAFs, as well as myCAFs, exhibit a distinct gene signature, enabling their differentiation from other tumor cell populations (Lambrechts et al., 2018; Chen et al., 2020; Wang et al., 2021; Luo et al., 2022; Liu et al., 2023; Qin et al., 2023; Wang et al., 2023). Consequently, alterations in the expression of this gene signature within tumor tissues are indicative of shifts in the abundance of iCAFs. Unlike myCAFs markers, markers associated with iCAFs tend to exhibit significant downregulation at the transcriptional level in tumor tissue compared to normal tissue across various cancer types (Figure 4; Supplementary Table S3). However, an exception to this trend is observed in pancreatic cancer, where the expression of most iCAF markers notably increases in tumor tissue (Figure 4; Supplementary Table S3). Additionally, the expression of certain genes correlates with overall survival across different tumor types (Figure 4; Supplementary Table S4).

Similar to myCAFs, iCAFs are also mentioned in the context of disease prognosis; however, these findings are contradictory. Higher iCAF abundance has been associated with poor clinical outcomes and diminished response to immunotherapy in patients with bladder cancer, low-grade glioma, and head and neck squamous cell carcinoma (Galbo et al., 2021; Chen et al., 2022; Obradovic et al., 2022; Mou et al., 2023). In colorectal cancer, increased abundance of only the IL1R1+ iCAF subgroup, but not all other iCAFs, correlates with a lower overall survival rate (Koncina et al., 2023). Conversely, in pancreatic ductal adenocarcinoma, a higher iCAF abundance is linked to better prognosis and improved response to immunotherapy (Hu et al., 2022). These contradictory observations suggest that the presence of iCAFs in tumors and their impact on tumor behavior vary depending on the specific tumor type.

Due to increased expression of inflammatory cytokines, iCAFs induce the differentiation of immune cells into pro-tumor populations and reduce the activity of immunostimulating cells, collectively contributing to the establishment of immunosuppressive TME. Across various tumor types, scRNA-seq studies consistently demonstrate that iCAFs contribute to the differentiation of macrophages towards an alternatively activated M2-like phenotype (Chen et al., 2020; Chen et al., 2022; Ma et al., 2023; Werba et al., 2023). This phenomenon is supported by the fact that cytokines typically produced by iCAFs play a pivotal role in macrophage polarization. For instance, in prostate, non-small cell lung, and colorectal cancers, CAF-derived CXCL12 and CXCL14 facilitate the M2 polarization of macrophages (Figure 5) (Augsten et al., 2014; Comito et al., 2014; Wu et al., 2022). M2 macrophages, distinct from classically activated (M1) ones, promote tumor proliferation, invasion, and immune suppression. Furthermore, M2-like macrophages can confer therapeutic tolerance to various anticancer drugs, including cisplatin, gemcitabine, tamoxifen, and doxorubicin (Wang et al., 2024). For example, CAF-derived CXCL12 accelerates colorectal cancer progression and induces cisplatin resistance in vivo by promoting the M2 polarization of macrophages (Figure 5) (Jiang et al., 2023).

In scRNA-seq studies, direct interactions between iCAFs and T cells have also been elucidated. iCAFs are presumed to interact with various tumor-infiltrating T cell populations through chemokine ligand-receptor interactions (Wu et al., 2020; Pu et al., 2021; Li et al., 2022). In triple-negative breast cancer, CXCL12+ iCAFs induce CD8⁺ T cell dysfunction (Figure 5) (Wu et al., 2020). Notably, the iCAFs signature exhibits a significant positive correlation with the exhausted CD8⁺ T cell signature but a negative correlation with cytotoxic CD8⁺ T cells in head and neck squamous cell carcinoma (Figure 5) (Mou et al., 2023). Cytotoxic CD8⁺ T cells are recognized as the most potent effectors in the anticancer immune response, forming the basis of current successful cancer immunotherapies (Raskov et al., 2021). Conversely, dysfunction and exhaustion of CD8⁺ T cells diminish the potential for sustained responses to checkpoint blockade (McGoverne et al., 2020).

In addition to modulating the tumor immune microenvironment, iCAFs exert direct effects on cancer cells, promoting resistance to anti-cancer therapy. Co-cultivation experiments involving ovarian, colorectal, and pancreatic cancer cells with CAFs or CAF supernatants have demonstrated a reduction in cisplatin- and gemcitabine-induced apoptosis, attributed to CAF-derived CXCL12 (Zhang et al., 2015; Morimoto et al., 2016; Wei et al., 2018; Zhang et al., 2020; Jiang et al., 2023). CXCL12-induced resistance primarily occurs via activation of the CXCL12/CXCR4 axis, initiating downstream signaling cascades in tumor cells, including Akt, ERK, and Wnt/β-catenin pathways (Figure 5) (Singh et al., 2010; Zhang et al., 2015; Zhang et al., 2020). These pathways regulate the expression of mitochondrial apoptotic proteins, such as BCL-2 and BAX. Reduced BAX/BCL-2 ratio in tumor cells leads to the inhibition of downstream caspase 3, which drives the terminal events of apoptosis. In colorectal cancer, CXCL12/CXCR4-mediated upregulation of miR-125b initiates a similar cascade to evade the effects of 5-fluorouracil (Yu et al., 2017). It is worth noting that a comparable mechanism of resistance to the cytotoxic effects of chemotherapeutic agents was observed in the interaction between myCAF-specific proteins and tumor cells. Suppressing effector caspases (caspase-3 and 9) may also occur upon activation of the CXCL12/CXCR4/survivin axis, resulting in decreased apoptosis and enhanced radiotherapy resistance in colon cancer cells (Figure 5) (Wang et al., 2017). Furthermore, the IL-6 autocrine loop, facilitated by Akt and ERK signaling pathway activation, contributes to CXCL12-induced gemcitabine chemoresistance in pancreatic cancer cells (Zhang et al., 2015). Additionally, CXCL12/CXCR4 signaling antagonizes docetaxel-induced G2/M phase cell cycle arrest by activating PAK4/LIMK1, enhancing microtubule dynamics (Figure 5). Consequently, this impedes docetaxel-mediated microtubule stabilization, suppresses its anti-mitotic effect, and promotes tumor cell survival (Bhardwaj et al., 2014). Collectively, these findings underscore the role of the CXCL12 axis in mediating cancer chemoresistance.

Less explored in the context of anticancer drug resistance, other iCAF markers warrant further investigation (Table 2). It is known that CAF-derived CXCL14 activates p38/STAT1 signaling in breast cancer cells, reducing paclitaxel-induced apoptosis (Figure 5) (Liu et al., 2017). Insulin-like growth factor IGF1 and insulin-like growth factor binding protein IGFBP6 exert contrasting effects on tumor cells, with pro-tumor and anti-tumor effects, respectively (Figure 5). Depending on the balance between secreted IGF and IGFBP proteins, CAFs can induce both osimertinib resistance and osimertinib sensitivity in lung cancer cells (Remsing Rix et al., 2022). Regarding complement system genes C3 and C7, there exists only indirect evidence suggesting their potential impact on anticancer therapy resistance. Elevated C7 expression predicts an unfavorable prognosis in patients treated with taxane-anthracycline chemotherapy (Zhang et al., 2021), while C3 expression is associated with resistance to 5-fluorouracil- and oxaliplatin-based chemotherapy in colorectal cancer (He et al., 2021). Consequently, the precise role of the iCAFs specific signature in tumor cell resistance to therapeutic agents remains incompletely characterized. A comprehensive investigation of iCAF markers could elucidate the contribution of this population to tumor resistance development.