There are multiple mechanisms through which CAFs elicit a range of pro-tumorigenic processes. One mechanism through which CAFs are able to facilitate tumorigenesis involves epithelial-mesenchymal transition (EMT). EMT is a physiological process that is implicated in embryonic development, wound healing and stem cell behavior (16). However, during tumor development, EMT also fosters fibrosis and promotes cancer progression (17). CAFs have also been demonstrated to produce multiple biologically active molecules, including cytokines and growth factors, such as transforming growth factor-β (TGF-β), platelet-derived growth factor (PDGF), vascular endothelial growth factor (VEGF), matrix metalloproteinase 9 (MMP9) and multiple interleukins (ILs) (8,18). All these molecules contribute to CAF-promoted tumorigenesis and its associated processes, such as angiogenesis (19,20).

Under physiological conditions, normal fibroblasts are at rest, and have little to no metabolic activity prior to their activation (21). However, this 'normality' is disrupted in the TME, where CAFs become permanently stimulated and undergo metabolic reprogramming. The reprogrammed fibroblasts subsequently initiate a support for tumor growth and cancer cell invasion (22). This support is materialized through the secretion of ECM proteins and the supply of energy and nutrients to the TME (6), the latter of which is mediated by catabolic processes in the activated CAFs, leading to the release of nutrients and metabolites into the TME, including lactate, ketone bodies, fatty acids and glutamine (23). The catabolic processes in CAFs are driven by autophagy and this process is initiated and driven by cancer cells, which undergo stress in the TME. This stress leads to the activation of multiple signaling pathways, resulting in the generation of reactive oxygen species, which, in turn, induce catabolic processes in CAFs (24). Specifically, the lactate that is released from CAFs causes acidification of the TME; this, in turn, activates MMP9, thereby contributing to the development of chemoresistance (25).

Further enhancement of the CAF-mediated support of cancer cells and tumor development and growth is mediated via pro-inflammatory cytokines. Cancer cells and CAFs both secrete pro-inflammatory cytokines such as IL-6 and IL-1, which promote further epigenetic modifications of CAFs (26). This process subsequently provides CAFs with the capacity to promote tumor development through enhancing actomyosin contractility and ECM remodeling, which thereby facilitates the migration of cancer cells (27). A previous study of Biffi et al (28) suggested that the secretion of multiple cytokines may even provide the main driving force that shapes the heterogeneity of CAFs in the TME.

An essential property of CAFs within the TME lies in their capacity to induce immunosuppression. This is achieved through the formation of fibrotic barriers, which have the effect of impeding the infiltration of immune cells, notably cytotoxic CD8⁺ T cells, into the TME (29,30). CAF- or cancer cell-produced collagen (31,32), which is the main composite of the fibrotic barriers (33), has also been shown to inhibit the effector (cytotoxic) functions of the tumor-infiltrating cytotoxic CD8⁺ T cells through the immune checkpoint molecule, leukocyte-associated immunoglobulin-like receptor 1 (34,35). Another important immunosuppressive feature is the capacity of CAFs to recruit or induce polarization of myeloid-derived suppressor cells (MDSCs), T cell-suppressing neutrophils and regulatory T cells (Tregs), subsequently inhibiting CD8⁺ T cell effector functions (10,36). In addition to this, as antigen presenters, CAFs have also been shown to prompt the transformation of naive CD4⁺ T cells into antigen-specific Tregs (12), thereby promoting the suppression of antigen-specific anti-tumor immunity.

Overall, the role of CAFs in cancer treatment has been considered to be mainly pro-tumorigenic. However, studies have indicated that, under certain conditions, CAFs are also able to exert anti-tumorigenic effects. For example, an inflammatory subpopulation of CAFs has been found to be associated with improved overall survival rates in patients with colorectal cancer (37), and their depletion was found to accelerate pancreatic cancer, thereby reducing survival (38). The mechanisms underlying their anti-tumorigenic activities may be that CAFs either limit tumor growth (39) or inhibit tumor cell proliferation (40).

The dynamic phenotypes of CAFs may also be interchanged (37). For instance, CAFs were shown to be modulated by the vitamin D analog calcipotriol into an anti-tumorigenic phenotype in pancreatic cancer (41). In the same study, CAFs under the influence of calcipotriol were shown to have a decreased production of several pro-tumorigenic factors, including prostaglandin E2, IL-6 and leukemia inhibitory factor (41). These findings provide further evidence of the little-known anti-tumorigenic activities of CAF populations, which may, in the future, force us to modify our views on their role in cancer and how they may be therapeutically targeted (15). This is especially important in the context of sarcomas, where their pro- or anti-tumorigenic activities are, at present, largely unexplored.

Breast cancer is the most frequently diagnosed type of cancer among women globally and is one of the leading causes of death in females (44). Over the past decade, notable advances in its treatment have been made, and these are attributable to the introduction of novel therapeutic agents, including anti-HER2 monoclonal antibodies and cell cycle inhibitors (45,46). Despite these improvements, however, advanced and metastatic breast cancer remains incurable and continues to be a major contributor to cancer-associated deaths.

CAFs exhibit different phenotypes in breast cancer (47), and these phenotypes are currently categorized according to the expression of surface markers into the following subtypes: CAF-S1, CAF-S2, CAF-S3 and CAF-S4 (48). Whereas CAF-S2 and CAF-S3 are also found in healthy tissues, CAF-S1 and CAF-S4 are only present in tumors and metastatic lymph nodes (48), and have been shown to have higher expression levels of α-SMA, a marker often associated with poorer prognosis (49). The CAF-S1 type is known for its utilization of C-X-C motif chemokine ligand 12 (CXCL12)/TGF-β-dependent signaling pathways, whereas CAF-S4 exerts its contractile abilities in promoting cancer cell motility through Notch-driven signaling pathways (50). Therefore, the presence of CAF-S1 and CAF-S4 in breast cancer has been assumed to be associated with a worse prognosis and the development of metastases (50). Other contributors to the CAF-mediated promotion of aggressive tumor behavior are the NOD-like receptor (NLR) family pyrin domain containing 3 (NLRP3)/IL-1β signaling pathway (51) and exosomal microRNAs released from CAFs, both of which promote cancer cell stemness and EMT (52-55).

Pancreatic cancer remains one of the most lethal malignancies, with a 5-year survival rate of ~10% (56). The majority of patients are diagnosed at an advanced stage due to the absence of early symptoms, or the presence of non-specific symptoms during the localized stage of the disease (56). Surgical resection remains the cornerstone of curative treatment for patients with small, localized pancreatic tumors, providing significant survival benefits. However, for patients with unresectable, metastatic or recurrent pancreatic cancer, surgical resection can only offer limited benefit. Management of metastatic or recurrent disease primarily involves chemotherapy, with or without targeted therapy. Given the limitations of current treatments in terms of achieving a cure, it is recommended to identify patients who may benefit from other potentially effective therapies, such as immunotherapy with immune checkpoint inhibitors (ICIs), or those who may be suitable candidates for clinical trials (57).

In pancreatic cancer, CAFs have also been shown to play a crucial role in shaping the TME and disease progression. Pancreatic ductal adenocarcinoma (PDAC) is the most common type of pancreatic cancer, and also one of the leading causes of cancer-associated mortality globally (58). Its common occurrence is mainly attributed to the delayed detection of tumors and limited therapy options. Pancreatic tumors are characterized by a dense stroma that is rich in CAFs (59). CAFs in pancreatic cancer can also be divided into subpopulations on the basis of their origin. CAFs infiltrating PDAC mostly originate from resident fibroblasts, pancreatic stellate cells, tumor-infiltrating mesenchymal stem cells (MSCs) or mesothelial cells. In the majority of cases, three subsets of CAFs in pancreatic cancer are recognized: iCAFs, myCAFs and apCAFs (60). MyCAFs express α-SMA molecules, whereas iCAFs are known rather for the secretion of pro-inflammatory cytokines, including IL-6. ApCAFs are characterized by their high expression levels of MHC II (61). The three identified subsets of CAFs within pancreatic tumors exhibit distinct functional profiles that are tailored to support various aspects of pancreatic cancer progression distributed across different spatial regions within the TME. MyCAFs are identified close to the tumor margin, whereas IL-6-positive iCAFs are predominantly observed in regions that were more distant from the tumor (62). IL-6 has been demonstrated to be secreted by iCAFs through paracrine signaling from the malignant epithelium, which is particularly relevant since elevated levels of IL-6 are generally indicative of a higher risk of developing extensive hepatic metastases in patients with PDAC (63,64). The balance between the myCAF and iCAF phenotypes is relatively dynamic and this balance is mainly regulated by differential engagement of the IL-1/JAK/STAT and TGF-β signaling pathways (28). The presence of iCAFs in PDAC has been shown to be associated with the activation of NF-κB signaling, as evidenced by reduced iCAF gene expression upon specific inhibition of NF-κB signaling. Activated NF-κB signaling in iCAFs has been shown to induce gene expression associated with tumor progression (65). Regarding apCAFs, these cells are highly potent antigen presenters, whose interaction with naive CD4⁺ T cells leads to the induction of antigen-specific Tregs (11). In PDAC, there is also evidence to suggest that CAF-induced regulation is mediated via extracellular vesicles (EVs) (66). Moreover, chemotherapy-exposed CAFs have also been shown to regulate the survival and proliferation of cancer cells, as they respond to the therapy through releasing increased numbers of EVs, with a concomitant increase in the level of the chemoresistance-inducing factor Snail, thereby promoting cancer cell survival and chemotherapy drug resistance (67).

Intraductal papillary mucinous neoplasms (IPMNs) are pancreatic cystic tumors that are derived from cells lining the pancreatic ducts (68). The association between IPMNs and PDAC is both unclear and complex; however, it has been demonstrated that IPMN lesions may increase the risk of developing PDAC (69). It has also been shown that iCAFs are abundant in PDAC, and myCAFs are abundant in high-grade IPMNs. This may suggest that iCAFs could contribute to the transformation of IPMNs into PDAC (70).

Primary liver cancer is the sixth most prevalent cancer worldwide, and the third leading cause of cancer-associated deaths. Hepatocellular carcinoma is the predominant type of primary liver cancer, comprising 75-86% of all cases of liver cancer. The incidence of hepatocellular carcinoma is two to three times higher in men compared with women, and these higher rates of incidence and mortality for men have been shown for most countries. At present, the treatment landscape for hepatocellular carcinoma is undergoing a major transformation, with combinations of different immunotherapies emerging as key components alongside targeted therapies (71-73).

Hepatocellular carcinoma typically arises from fibrosis and cirrhosis, and is characterized by an abundance of stromal myofibroblasts (74). Similarly to other types of tumor, CAFs are associated with poor prognosis in cases of hepatocellular carcinoma, and they have been shown to promote the migration and invasion of hepatocellular carcinoma cancer cells. Moreover, Liu et al (75) described an associated activation of the Hedgehog/TGF-β signaling pathway in hepatocellular carcinoma in response to several CAF-secreted chemokines. CAFs in hepatocellular carcinoma have been shown to promote cancer cell proliferation by secreting hepatocyte growth factor, thereby preventing cancer cells from undergoing necrosis (76). CAFs in hepatocellular carcinoma are also reported to enhance the stemness of cancer cells through producing hepatocyte growth factor and IL-6 (77,78). There is also emerging evidence to suggest that CAFs may also provide an immunosuppressive TME in hepatocellular carcinoma, as they attract MDSCs (79) and neutrophils that are subsequently primed by CAFs to impair T cell function via programmed cell death protein 1 (PD-1)-mediated inhibition (36).

CAFs have also been shown to make an important contribution to therapy resistance in hepatocellular carcinoma. For example, CAF-secreted phosphoprotein 1 was identified as the most important factor responsible for the resistance of hepatocellular carcinoma cells to treatment with sorafenib and lenvatinib, both of which are offered as first-line systemic therapy for advanced hepatocellular carcinoma (80,81). Moreover, in an animal model of hepatocellular carcinoma, a higher abundance of CAFs was also demonstrated to be associated with low sensitivity of tumors to anti-PD-1 immunotherapy (82).

Colorectal cancer ranks as the third most frequently diagnosed cancer and a trend of increasing incidence has been reported among young adults in the 21st century (83). Colorectal cancer ranks among the top five leading causes of cancer-associated mortality in both men and women globally (84,85). Of patients with colorectal cancer, ~50% develop liver metastases, which markedly contribute to the lethality of the disease. For metastatic colorectal cancer, doublet or triplet chemotherapy should be offered to patients with initially unresectable metastatic colorectal cancer, with specific therapies based on tumor characteristics such as microsatellite instability and the RAS mutation status. Cytoreductive surgery in combination with chemotherapy is often considered for colorectal peritoneal metastases. For patients with potentially curable liver metastases, however, perioperative chemotherapy or surgery alone is recommended. In all cases, multidisciplinary team management and shared decision-making are crucial (86,87).

CAFs are abundantly present in colorectal cancer, and various CAF markers (for example, a-SMA, PDGFR-B, FAP and ferroptosis suppressor protein-1) are often found to be expressed in different stages of the disease (88). A study by Wikberg et al (89) demonstrates the importance of the localization of the expression of these markers: Patients with colorectal cancer exhibiting high FAP expression in the tumor center, but not in the tumor periphery, were shown to have poorer prognosis, as FAP was expressed more highly in tumors with positive microsatellite instability, suggesting FAP to be a potential negative prognostic factor.

CAF-derived EVs also have a role in colorectal cancer. These EVs are capable of increasing cancer cell stemness and EMT and, therefore, of promoting metastasis and resistance to therapy (90). A study of Aizawa et al (91) described the secretion of Wnt family member 2 (Wnt2) by CAFs as an important cancer-promoting mechanism. The high expression of Wnt2 in CAFs was also found to correlate with the depth of the tumor, lymph node metastasis, TNM stage, venous invasion and disease recurrence.

As one of the most commonly diagnosed cancers, lung cancer is also the major cause of cancer-associated deaths globally (92). Of all lung cancer cases, ~85% are of the non-small-cell lung cancer (NSCLC) type. Due to the lack of effective screening programs and the late appearance of symptoms, however, the majority of patients are diagnosed at an advanced stage, which results in a very poor prognosis (93). Over the last 20 years, major progress has been achieved in the treatment of NSCLC, which has, at the same time, deepened our understanding of its biology and tumor progression. Small-molecule tyrosine kinase inhibitors and immunotherapy have provided significant survival benefits for specific patients. Nevertheless, the overall cure and survival rates for NSCLC, especially for the cancer in its metastatic stages, remain low (92,94).

During the disease progression, CAFs fulfill a crucial role in the constitution of a supportive niche for cancer stemness, EMT and cancer cell growth (95,96), suggesting that CAFs act as feeder cells in lung cancer (97). EMT induced by CAFs has been shown to enhance the capability of cancer cells to migrate and form the secondary tumor foci via several molecular mechanisms. Secretion of IL-6 by CAFs has been described as driving the EMT via the subsequent activation of JAK2/STAT3 signaling (98,99). CAFs, as described previously, have been shown to regulate cancer cells via exosomal transfer. In lung cancer, CAFs have been shown to promote cell migration and invasion by triggering the PTEN/PI3K/Akt pathway with EVs via miR-210 (100). CAFs were also found to perform a crucial role in TME remodeling (101). CAFs were reported to promote the recruitment of CCR2⁺ monocytes via the production of chemokine (C-C motif) ligand 2. The recruited monocytes were subsequently polarized by CAFs into cells with the MDSC phenotype, causing the robust suppression of adaptive anti-tumor immunity, specifically CD8⁺ T-cell proliferation and interferon γ (IFN-γ) production.

In recent years, the use of ICIs has elicited significant results in several cancer diagnoses. In lung cancer, however, only a small portion of patients were found to respond to this type of therapy (102). This low responsiveness has been attributed to CAFs, as they have been shown to prevent immune cell infiltration, specifically of the cytotoxic CD8⁺ T cells, resulting in remodeling of the ECM, thereby promoting immune escape (103,104). Therefore, several recent studies have suggested targeting CAFs as a potential path to overcome ICI therapy resistance in lung cancer (105-107).

Prostate cancer is one of the most common malignancies in men, predominantly affecting those over the age of 65 (92,108). The majority of prostate cancers are adenocarcinomas, which typically originate in the peripheral part of the gland, and often progress slowly (109). For clinically localized cases, standard treatments include radical prostatectomy and radiation therapy. Androgen ablation, which involves either bilateral orchiectomy or the use of luteinizing hormone-releasing hormone agonists or anti-androgens, has been demonstrated to reduce tumor burden, thereby extending median progression-free survival (110). Although the treatment is initially effective, the majority of patients on hormonal therapy eventually develop castration-resistant prostate cancer, with ~70% of patients progressing to metastatic disease, primarily in the bones (80-90% of advanced and metastatic cases) (111,112).

Prostate CAFs have a crucial role in driving prostate carcinoma progression, influencing the supportive microenvironment through age-associated stromal changes and cancer-specific alterations (113,114). Multiple sources contribute to the heterogeneity of CAFs in prostate cancer, including resident fibroblasts undergoing EMT (115,116), bone marrow-derived MSCs recruited through chemotaxis (117), senescent stromal cells that promote tumor progression (118) and pericytes that potentially resemble CAFs (116). CAFs exhibit diverse functions within the prostate cancer TME, categorized into myCAFs, apCAFs and immune-regulatory CAFs (119). In particular, myCAFs occupy a central role in ECM remodeling, promoting collagen deposition and cancer cell growth, upon which they induce ECM stiffness and morphological alterations (120-122).

CAFs are sensitive to androgen levels and contribute to resistance to anti-androgen therapies, thereby promoting the development of castration-resistant prostate cancer (123-125). CAFs have been shown to mediate resistance to anti-androgen therapies through modulating androgen receptor signaling and enhancing cytokine secretion (126,127). Moreover, CAFs were found to decrease sensitivity to anti-androgens and chemotherapies through various signaling pathways, including the PI3K/Akt pathway (128,129).

STS accounts for ~1% of all malignancies in adults (36). STS can be formed anywhere in the body, although the extremities are the most common sites (7). The most prevalent subtypes of STS are LPS, LMS and UPS. Among other highly prevalent subtypes of STS are myxofibrosarcoma, synovial sarcoma and RMS in children (7). The TME varies extensively within each histological subtype, and also according to its anatomic location and the tumor stage (132); the same holds true for fibroblasts. Since fibroblasts can acquire a wide variety of phenotypes, and given their specific functions at distinct locations of the body (133), they should be evaluated individually within each histological subset of STS.

LPS arises from adipose tissues and represents one of the most common types of STSs in adults, representing ~20% of all cases of adult STS (134). LPS predominantly manifests in individuals aged 50-70, with an equal incidence in both sexes. The anatomical distribution of LPS is wide, with each subtype characterized by distinct morphological features, distinctive natural history, and specific genetic alterations crucial for diagnosis (134). According to the latest World Health Organization classification, LPS is divided into five subtypes: Well-differentiated LPS, dedifferentiated LPS (DDLPS), myxoid LPS (MLPS), pleomorphic LPS (PLPS) and myxoid pleomorphic LPS (MPLPS) (134).

The primary approach for treating all LPS subtypes is surgical resection. The degree of resection required depends on the location of the tumor and its histopathological characteristics. Achieving a complete gross resection is paramount for all sarcomas. The conventional first-line therapeutic approach for unresectable or metastatic disease involves systemic chemotherapy that is based on anthracyclines. Any decision to use adjuvant therapy for patients with high-risk LPS should be made on an individual basis; nevertheless, the chemotherapeutic sensitivity of LPS remains grossly low (134).

Fibroblasts are an important component of adipose tissue. Adipocytes regulate their size through expansion or shrinkage, and therefore there is an urgent requirement for continual tissue remodeling. ECM produced by fibroblasts provides structural support and has the function of storing growth factors and cytokines and also has the function of protecting adipocytes from mechanical stress (135). Moreover, adipocytes appear to be able to generate CAFs (42,136). Specifically, it has been shown that adipocytes, in the presence of cancer cells, are able to differentiate into cancer-associated adipocytes (CAAs) that express matrix remodeling proteins, leading to the overproduction of ECM. These CAAs subsequently give rise to adipocyte-derived fibroblasts, which fuel migratory properties in various types of cancer (137,138).

Our knowledge regarding the role of CAFs in LPS is relatively poor. Harati et al (139) isolated CAFs from one surgically removed high-grade PLPS and one intermediate MLPS. An in vitro co-culture with the LPS SW872 cell line induced cell proliferation and an increased viability of cancer cells compared with co-culture with dermal fibroblasts (139). Moreover, the co-cultivation of CAFs originating from high-grade PLPS led to a further enhanced proliferative capability of the cancer cells. In agreement with these results, Xu et al (140) observed increased proliferative and migratory capabilities of various LPS cell lines when co-cultured with CAFs isolated from retroperitoneal DDLPS. Notably, the secretion of thrombospondin-2 by LPS cells was shown to induce the transformation of fibroblasts into CAFs in vitro. Therefore, thrombospondin-2 may be an attractive therapeutic target for LPS (140). CAFs were also found to be able to influence clinicopathological features of retroperitoneal LPS. It was shown that the expression of α-SMA was correlated with tumor size, grade and subtype; however, it had no effect on overall survival, disease-free survival or recurrence-free survival (140). CAFs were also shown to influence the efficacy of chemotherapeutics (141) and it appears that their presence could decrease the sensitivity of LPS to the chemotherapeutic drug doxorubicin (139).

RMS is the most common STS in children and adolescents, classified as a high-grade neoplasm of skeletal myoblast-like cells (142). RMS comprises two primary subtypes, alveolar and embryonal, initially identified by their light-microscope characteristics, each of which are driven by different molecular mechanisms and presenting unique clinical challenges. Effective treatment requires control of the primary tumor, which can develop in various anatomical locations, as well as managing disseminated disease, assumed to be present in all cases.

RMS are typically skeletal muscle tumors. However, RMS may develop on sites that normally lack skeletal muscles as well. This raises a question about RMS origins; one of the proposed hypotheses points to mesenchymal progenitor cells. These cells may circulate between organs, which could explain RMS occurrence at sites without skeletal muscles. Moreover, these cells may give rise to fibroblasts as well (143).

Skeletal muscle fibroblasts are mainly located in the space between muscle fibers. Although they represent a small fraction of cells within the muscle, they have an irreplaceable role in terms of supporting force transduction, maintaining muscle structure and responding to muscle injuries. Recently, fibroblasts were also detected at the border of the perimysium and endomysium (133,144).

It has been shown that there is a limited crosstalk of CAFs and RMS cells. Cancer cells typically remodel the ECM on their own, and express functional and structural components, including metalloproteinases, fibronectin and laminin. By contrast, RMS cells secrete macrophage migration inhibitory factor (MIF), which induces activation of MAPK, enhances cell adhesion and increases tumor vascularization, but decreases the recruitment of CAFs. Moreover, MIF was shown to impede the migration of cells toward the IL-8 gradient (145). RMS has also been found to secrete IL-8 under hypoxic conditions (146). Since its receptor C-X-C motif chemokine receptor 2 (CXCR2) is expressed by fibroblasts (147,148), this MIF-IL-8 axis could explain the decreased presence of CAFs within the TME of RMS. These findings diverge from the typical behavior observed in a number of types of cancer, where tumor progression and metastasis are often facilitated by the presence and activity of CAFs (141).

On the other hand, EVs derived from fusion-positive or -negative RMS have been shown to stimulate fibroblast proliferation and invasiveness (149). Detailed analysis has revealed that the presence of transmembrane glycoprotein CD147 on RMS-derived EVs was critical for the induction of these properties in BJ fibroblasts (150). Moreover, CD147 was identified to be one of the key molecules required for EV uptake by BJ fibroblasts (150).

LMS is a common type of STS, and ~50% of these tumors are formed in the retroperitoneum or intra-abdominal regions (151). Managing retroperitoneal LMS is challenging due to the tendency of this sarcoma subtype to metastasize rather than to recur locally. The lungs and liver are primary sites for metastasis, and once metastasis occurs, survival rates decrease markedly, with a median survival rate of only 13 months (152).

Surgery is the main treatment for LMS, and complete removal of the tumor is essential for improving survival. However, retroperitoneal tumors frequently recur, even after a number of years post-diagnosis (153). Tumor grade and size are crucial factors affecting overall survival (154). Therefore, addressing the risk of distant metastasis is critical in treating retroperitoneal LMS. Moreover, the prognosis of patients with retroperitoneal LMS remains poor, mainly since these tumors are typically already large at the time of initial detection (155).

LMSs are classically divided into extra-uterine and uterine subtypes (156). Myofibroblasts represent smooth muscle fibroblasts, and these are morphologically enlarged, interconnected through gap junctions and facilitate contractile force and tissue repair (157). To date, only a limited number of studies have investigated the role of CAFs within LMS. It was shown by Nagao et al (158) that uterine LMSs secrete specific miRNA in EVs, which has the effect of inducing the transformation of fibroblasts into CAFs. Further studies, however, are warranted to clarify the role of CAFs within the LMS microenvironment.

UPS is characterized as a high-grade and aggressive STS, and is one of the three most common histotypes of STSs that are associated with the poorest prognosis (159). This malignancy may be manifested in various locations, including soft tissues, bones and the retroperitoneum, and it has the potential to metastasize to multiple organs (160). The primary treatment strategy for UPS involves surgical excision, with the goal of achieving microscopically negative margins. In certain cases, the use of neoadjuvant or adjuvant radiotherapy and chemotherapy may be considered to enhance treatment outcomes (160).

The origin of UPS has been controversial for a long time. Formerly classified as malignant fibrous histiocytoma, the cells were thought to originate from histiocytoma (161,162). However, later studies have challenged this hypothesis (163,164), showing that the cells have uncertain differentiation due to the absence of differentiation markers, and that they originate from often highly transformed cells at early stages of their development from MSCs (165). The cells of UPS are extremely pleiomorphic, acquiring numerous shapes and sizes, thereby making it difficult to visually distinguish them from other cell types, including fibroblasts or CAFs (32,166-168). Based on this perspective, the identification of CAFs in UPS represents an even greater challenge, and at present, hardly any studies have been published that were directly dedicated to exploring the role of CAFs in UPS.

Bone sarcoma is a group of rare tumors with distinct incidence, which most commonly appear in children and adolescents (133). It consists of osteosarcoma, EWS and chondrosarcoma (133). Fibroblasts are directly involved in cartilage degradation through matrix-degrading enzymes (133,169,170). They affect bone erosion via cathepsin K, and regulate the osteoclast/osteoblast axis; moreover, they are able to give rise to bone-forming osteoblasts (133,169,170).

EWS is an aggressive sarcoma standing at the 'crossroads' between bone and soft tissue cancer. The treatment for EWS typically involves starting with neoadjuvant chemotherapy, followed by surgical resection, which is subsequently followed by adjuvant chemotherapy. For patients with metastatic disease, the control of metastatic sites is usually postponed until after the adjuvant chemotherapy has been administered. However, although there are clear treatment plans for newly diagnosed cases, the approach to managing recurrent EWS requires individualized strategies (171).

As aforementioned, EWS usually occurs in children and young adults, with peak incidence in adolescents (172). Differences in gene pathways with prognostic effect have been determined only in EWS with stromal content (173). This underlines the need to study the tumor stroma and associated underlying mechanisms that lead to tumor progression. Notably, not only stromal cells may have the capability to produce the ECM; CD73⁺ CAF-like tumor cells have been described in EWS. These cells deposit pro-tumorigenic ECM proteins, thereby contributing to tissue remodeling. These are termed EWS/FLI1-low cells, which are predominantly present along tumor borders and invasive fronts (174).

Osteosarcoma is a bone cancer, with a peak occurrence during the second decade of life. The current treatment for osteosarcoma primarily involves surgery in conjunction with adjuvant chemotherapy. However, due to the rapid progression of the disease and the development of chemotherapy resistance, the effectiveness of chemotherapy is often limited, particularly in recurrent and certain primary cases (175,176).

Osteosarcoma mostly occurs in the extremities (177). These tumors are surrounded by ECM, which creates a physical barrier for infiltrating cells (178). However, there is an ongoing direct and indirect crosstalk between osteosarcoma cells and CAFs. The miRNA 1228 has been shown to be upregulated in CAFs from patients with osteosarcoma. It downregulates the protein, suppressor of cancer cell invasion, in osteosarcoma cells, which thereby promotes their proliferation and migration (179). On the other hand, osteosarcoma-derived EVs have been found to drive the differentiation of lung fibroblasts into CAFs, which subsequently exhibit both an increased expression of α-SMA and metalloproteinase MMP2 and MMP9, and an increased production of fibronectin (180). This may serve as a possible mechanism of metastatic niche formation. Moreover, TGF-β1 was also identified as a crucial molecule for fibroblast transdifferentiation (180). Osteosarcoma cells can secrete EVs with collagen type VI alpha 1 (COL6A1), which transforms normal fibroblasts into CAFs that secrete IL-6 and IL-8 (181). In addition, COL6A1 has been shown to induce TFG-β expression in CAFs, which subsequently triggers the invasiveness of osteosarcoma cells (181). The process of cell-projection pumping, the exchange of cellular and cytoplasmic molecules, has been reported to occur between osteosarcoma and human gingival fibroblasts. This interaction surprisingly led to a decreased production of IL-6, granulocyte-colony stimulating factor and granulocyte-macrophage colony-stimulating factor (182).

It is becoming clear that a variety of CAFs with diverse functions and prognostic impact are involved in osteosarcoma. In an in silico study performed by Xu et al (183), six subclusters of fibroblasts were identified. Specifically, TOP2A⁺ CAFs showed the highest association with carcinogenic pathways (p53 and cell cycle) and crosstalk with osteosarcoma cells. Only these TOP2A⁺ CAFs were found to serve as a prognostic factor (183). In a different study (184), similarly, high levels of two more CAF subsets that either resembled iCAFs or apCAFs were found to negatively influence survival probability in patients with osteosarcoma.

Fibroblasts and sarcomas have a common mesenchymal origin. Putting aside the heterogeneous origins of CAFs that have been described previously, fibroblasts and connective/muscle tissues share developmental lineage. They both arise from the true mesenchyme during embryonic development. Moreover, certain fibroblasts, similarly to other cells of the connective/muscle tissues, have been shown to arise from MSCs in the post-natal period (185,186). These complicated developmental associations are further emphasized in RMS, which can arise in human body parts lacking skeletal muscles. One of the possible hypotheses underlying this phenomenon could be its development from mesenchymal progenitor cells that failed to develop into myocytes (143). Due to these tight developmental links, numerous sarcoma cancer cells share features with fibroblasts/CAFs with respect to the shared gene expression involved in cell adhesion, motility and ECM production. The most closely related sarcoma cells are histological subtypes of sarcomas that directly arise from fibroblasts (either fibrosarcoma or myxofibrosarcoma). At present, no information is available on whether CAFs are present in these tumors, or, if they are present, what the nature of their origin is; namely, neighboring tissue-resident fibroblasts, endothelial to mesenchymal transition, MSCs/tumor-associated MSCs, EMT, transdifferentiated fibroblasts or fibrocytes from the bone marrow (4). The absence of this information is primarily due to the absence of markers that could distinguish CAFs from sarcoma cells. This limitation is probably the main reason contributing to the scarcity of studies on sarcomas and CAFs in comparison with the abundance of studies on CAFs in carcinomas. Therefore, sarcomas and CAFs currently represent largely uncharted territory in terms of both clinical and academic research.

ECM within the TME is a complex milieu, which is responsible for orchestrating numerous processes, including cell migration and immunomodulation. Although no comparative studies between sarcomas and carcinomas have been published, it could be hypothesized that sarcomas are likely to have a higher accumulation of ECM compared with epithelial cancers due to both their shared developmental processes with fibroblasts and the abundance of fibroblasts in connective tissues. However, certain sarcomas, such as RMS, which is a skeletal muscle-like tumor, may exhibit an abundance of ECM similar to that of epithelial tumors due to the lower presence of fibroblasts. In carcinomas, CAFs are able to drive EMT. However, sarcoma cells already exhibit mesenchymal features, and these may need to be induced in carcinomas to enable metastases. Sannino et al (187) proposed that sarcomas are able to undergo phenotypic changes between EMT and mesenchymal-epithelial transition and can retain an intermediate, metastable phenotype. The varying degrees of epithelial/mesenchymal-like differentiation in different types of sarcomas supports this hypothesis.

Lineage plasticity is increasingly acknowledged as the hallmark of cancer, as it is associated not only with cancer cells, but also with other cells in the TME (188). CAFs have been shown to exert an important role in tumoral plasticity and their mutual interactions with cancer cells and other cells of the TME determine tumor behavior, which is translated into cancer progression, invasiveness and therapy resistance (188). In carcinomas, the plasticity of CAFs and cancer cells has been extensively studied both in vitro and in vivo. Via either basic experimental models exploring the changes elicited after co-culture of fibroblasts/CAFs with cancer cells (189,190) or in vivo models and advanced special-based analytic techniques (43), or via the possible implementation of novel strategies and techniques for the evaluation of lineage plasticity (188), our understanding of CAF plasticity in carcinomas has advanced compared with the past.

The research trajectory of CAFs in sarcomas, however, has been different. Pioneering studies on sarcomas and fibroblasts attempted to delineate the association between sarcoma cells and fibroblasts as long ago as the 1950s (191), with a small amount of further progress being made in the 1970s (192,193). These studies already pointed to the behavioral similarities between sarcoma cells and fibroblasts in terms of their in vitro invasiveness (191,192) or the acquisition of their features (193). Since then, attempts that have been made to distinguish fibroblasts/CAFs from sarcomas either morphologically/phenotypically or functionally (194,195), or to delineate their mutual association (193), have been rather sporadic, falling far short of the research efforts dedicated to CAF research in carcinomas. However, novel strategies and technologies are currently changing the opportunity landscape enabling exploration of CAF heterogeneity within sarcomas to be made. The application of novel strategies in combination with metabolic profiling may lead to the identification of novel CAF populations in sarcomas (196); alternatively, applying single-cell sequencing techniques in combination with spatial profiling may serve to unravel sarcoma cell populations with CAF-like properties (174). These novel strategies, in combination with multidimensional high-definition techniques, may now challenge the previously unexplored identities of CAFs or CAF-like cells in sarcomas.

Unraveling the identity of CAF heterogeneity, however, will bring not only opportunities, but also challenges. At present, a number of CAF subsets with various functions and phenotypes have already been identified. This variability may complicate efforts to target them effectively. Furthermore, a universal therapeutic strategy is unlikely to be uncovered without having a comprehensive understanding of the intricate TME. Recent advancements in technology, such as single-cell sequencing and spatial transcriptomics, may facilitate a more detailed exploration of specific CAF subsets (197). However, truly effective targeting of these subsets will require an integrated, deeper and context-dependent understanding of their roles. This understanding will certainly need a greater stress and reliance on an understanding of the functional diversity of CAFs (198). It may be assumed that the functional diversity of CAFs is probably still largely hidden beneath the universal CAF phenotype; moreover, this as yet still-hidden and unexplored functional diversity, driven by the reprogramming forces from the tumor and tumor-infiltrating cells, will presumably complicate the future development of universal therapeutic strategies targeting CAFs or CAF-like cells in sarcomas.

Regardless of the aforementioned discussion, at present, the exact role and underlying mechanisms of CAF involvement in sarcomas have yet to be fully elucidated, which certainly further contributes to the complexity of CAF functionalities and the ensuing challenges with their identity in sarcomas. Owing to such complexity, research based on cancer cell lines and fibroblast lineages needs to be more frequently corroborated both by in vivo analyses and the investigation of CAFs isolated directly from sarcomas and advanced approaches and techniques should be implemented to enhance our understanding of CAF and/or CAF-like cell functions within these challenging tumors.

One particular therapeutic avenue that has come under consideration is to focus on depleting the immunosuppressive FAP⁺ CAF population. In order to target FAP⁺ CAFs, a study by Kakarla et al (12) used chimeric antigen receptor T cells (CAR-T), which target FAP. Another study by Wang et al (200) also reported that CAR-T cells were not only able to inhibit tumor growth, but they could also cause an increase in the responsiveness of CD8⁺ T cells to tumors. A recent study also demonstrated that these CAR-T cells could improve the anti-tumor efficacy of claudin18.2-targeting CAR-T cells via suppressing the recruitment of MDSCs in pancreatic cancer, suggesting that targeting CAFs may be beneficial in combined therapies (201). The strength of the depletion strategy stems from a considerable (antigen-mediated) specificity of the CAF targeting. However, the weakness of this strategy is that CAFs are not a uniform population, consisting of populations with occasionally opposing roles in the tumor microenvironment. Depleting all CAFs may thus inadvertently remove subtypes that have anti-tumorigenic activities, or removal of CAFs with pro-tumorigenic activities may induce compensatory mechanisms. Therefore, depleting CAFs can also lead to unintended consequences and possible weaknesses, as was shown in a preclinical animal study by Ozdemir et al (38). In the authors study involving a transgenic mouse model of PDAC, the removal of α-SMA⁺ myCAFs led to increased invasiveness of the cells and enhanced EMT. Additionally, this depletion of α-SMA⁺ myCAFs resulted in decreased survival rates among the experimental animals, and was associated with a suppression of anti-tumor immunity (38). These contrasting findings therefore raise strong awareness of the importance of not employing CAF depletion approaches without a full understanding of how CAF biology operates under specific disease state conditions.

The therapeutic impact can be attained not only through the direct targeting of CAFs, but also through targeting their products, including EVs. The EVs produced by CAFs are able to support cancer progression, create a pre-metastatic niche for certain cancer diagnoses, and facilitate resistance to therapy. In a study by Gao et al (202), CD63⁺ CAFs were found to drive the resistance to tamoxifen in breast cancer via the secretion of miR-22-rich EVs. This study further demonstrated that this resistance could be overcome by the simultaneous neutralization of CD63 and miR-22 to enhance the therapeutic effect of tamoxifen (202). Another study (67) showed that CAF-released EVs could mediate CAF-promoted survival of cancer cells as the response to chemotherapy, and inhibiting the release of EVs was shown to overcome the chemotherapy-induced resistance. However, EVs have also been shown to serve dual roles as both targets and tools in cancer therapy (38). Zhou et al (203) recently demonstrated that EVs have significant potential to reprogram CAFs, thereby enhancing therapeutic responses in pancreatic cancer. In their study, these researchers loaded EVs with miR-138-5p and the anti-fibrotic agent pirfenidone. They found that miR-138-5p could inhibit both TGF-β signaling and proline-mediated collagen synthesis. This effect was synergistically enhanced by addition of the anti-fibrotic agent pirfenidone, resulting in the reprogramming of CAFs and the subsequent suppression of their pro-tumorigenic effects. Therefore, targeting EVs represents a promising strategy for elucidating the underlying key pro-tumorigenic mechanisms associated with CAFs. The potential strength of EV targeting may stem from its enhanced specificity. This is because pro-tumorigenic EVs are selectively affected, while the functions of anti-tumorigenic CAFs and their released EVs may remain intact. However, similar to CAF depletion strategies, there are still significant challenges. The inherent heterogeneity of EVs, varying in size, composition, and origin, along with a lack of clear and EV-specific molecular targets that could identify the pro-tumorigenic nature of these EVs, still needs to be addressed.

Given the important role of TGF-β in CAF biology, and how CAFs affect ECM deposition, targeting TGF-β signaling has also been proposed as a potential strategy for cancer treatment. However, immune dysregulation, or employing an incorrect dosing of anti-TGF-β agents, may cause potential systemic cytotoxicity (204). Nevertheless, the TGF-β signaling pathway has been shown to be regulated by Wnt signaling (205,206). There is growing evidence to suggest that Wnt signaling in CAFs fulfills a significant role in tumor development. For example, cancer cells located in regions enriched with CAFs were shown to have upregulated Wnt signaling (207) and autocrine Wnt signaling in CAFs was shown to promote cancer cell migration and invasiveness and EMT in breast cancer (208). Another study by Avgustinova et al (209) also demonstrated that non-canonical Wnt7a, which is exclusively secreted by aggressive breast cancer cells, could recruit and activate fibroblasts via a TGF-β-dependent process. Wnt signaling in CAFs has also been shown to be involved in inducing tumor resistance to various types of therapies, including chemotherapy. For example, increased expression of Wnt16B by CAFs was found to attenuate the cytotoxic effects of the chemotherapeutic drugs mitoxantrone and docetaxel in patients with prostate cancer (210). Wnt signaling in CAFs has also been demonstrated to modulate the immune components of the tumor microenvironment; for example, autocrine Wnt2 signaling in fibroblasts was shown to promote: i) Colorectal cancer progression (211); ii) the recruitment and polarization of immune cells, such as tumor-associated macrophages or Tregs [as described in (212)]; and iii) Wnt2 signaling may also inhibit in vitro differentiation and stimulatory activity of dendritic cells (213). Therefore, targeting Wnt2 signaling in CAFs may afford a promising therapeutic strategy, and this has already been corroborated by a study in which targeting Wnt2-producing CAFs was shown to restore dendritic cell-mediated immunity against tumors (213).

Other potential CAF-targeting therapies may include the stimulation of IL-2 pathways (214,215) or inhibition of IL-1 signaling through IL-1 receptor agonists (28). It has also been demonstrated that TGF-β-activated CAFs contribute to tumor resistance in breast cancer therapy, especially in the case of treatment with trastuzumab, which is often associated with diminished IL-2 activity. This suggests that stimulation of the IL-2 pathway in the stroma may restore the efficacy of trastuzumab (216). To inhibit the IL-1 signaling pathway, the inhibition of NLRP3 in CAFs has been proposed to offer a promising therapeutic strategy. This proposal was based on the finding that activation of the NLRP3/IL-1β signaling axis in CAFs has been associated with both the inflammation-mediated immunosuppressive milieu and the promotion of tumor progression (51,217).

Targeting cell signaling could be a promising approach for future anti-cancer therapies, as new and often highly specific inhibitors of these signaling pathways are being developed. However, a significant challenge with targeting the signaling pathways involved with CAFs is that these targets are essential for maintaining homeostasis. As a result, these inhibitors need to be effective only within the TME. Their full therapeutic potential is likely to be unleashed when combined with effective drug delivery strategies, such as antibody-drug conjugates or nanocarrier-based systems (218).

Although conventional 2D monolayer cell culture is an easy and often used system with high cell viability, it is not as effective as 3D models are in terms of replicating the complexity of the TME. CAFs cultured under 3D conditions secrete relatively more paracrine signaling molecules, and promote the invasive behavior of breast cancer cells compared with 2D conditions (243). Additionally, when CAFs are grown in a 3D model using a mixture of Matrigel and collagen I, the expression of important proteins for cell-ECM contacts, such as tenascin, are upregulated. This was found not to be the case for 2D cultures (244).

Various methods have been employed to enhance 3D tumor spheroid models in order to replicate more closely the complexity of the TME. One way to study the interactions between tumor cells and CAFs is through using a multicellular (heterotypic) spheroid system, where CAFs are co-cultured with tumor cells (245). This system enables an analysis of ECM production and remodeling by CAFs to be made; however, it does not incorporate an important component of the tumor stroma, which is the vasculature. Tumoroids, which consist of tumor cells embedded in a stromal environment comprising monomeric type I rat-tail collagen, patient-specific CAFs and ECM components such as laminin and human umbilical vein endothelial cells, can be utilized to study the impact of CAFs on the vasculature. Additionally, 3D-printed scaffolds may be employed to mimic the intricate architecture of the TME more closely when investigating CAF-cancer cell interactions. Synthetic materials, such as poly(ε-caprolactone) (PCL) (246) and natural biomaterials, such as sodium alginate or gelatin (247), have both been demonstrated to offer multifaceted support for cells and to facilitate direct cell interactions, consequently improving the reproducibility of the 3D tumor models.

To create a realistic 3D model, it is important to consider immune cells. Cytokines and chemokines produced by CAFs are mainly analyzed by ELISA, or via various multiplex cytokine analyses from co-culture supernatants (248) or conditioned media (249). To obviate the need for complex analyses of individual populations in co-culture models, Balachander et al (250) performed a study using only CAFs cultured on engineered fibrous matrices of PCL scaffolds. They demonstrated proliferation of the CAFs, with ECM remodeling, induced stemness and increased invasiveness of the cancer cells, and the CAFs retained a pro-inflammatory phenotype independently of the cancer cells.

The most commonly used in vivo models for studying CAFs are mouse and zebrafish xenografts with an injected mix of tumor cells and CAFs (53,251-253). Additionally, Cre-driver transgenic mice are often used (254-256). The co-injection models have the disadvantage that the injected CAFs become mixed up with the host-derived CAFs as the tumor grows. However, even transgenic models using Cre drivers are not flawless models, as a CAF-specific Cre driver line does not currently exist (8). Therefore, further research is required to develop improved in vivo models that are suited to the study of CAFs.

For an improved understanding of the association between sarcomas and CAFs, researchers often utilize the aforementioned strategies. In the case of RMS, both 3D spheroids and xenograft mice have been employed (5). Similarly, in a study by De Vita et al (257), both 3D cultures with a collagen scaffold and a xenotransplantation model were used to study CAFs in UPS, LPS and LMS. Comparatively speaking, the EWS model is quite complex, and requires the inclusion of bone architecture. To create this model, in a study by Molina et al (258), human MSCs were seeded on to and cultured on PCL scaffolds for 12 days in an osteogenic medium to enable their differentiation into osteoblastic cells. Subsequently, the constructs were decellularized and used to inoculate EWS cells (258).

Developing technologies, such as tumor-on-chip models, could further improve the models for CAF/sarcoma studies (259). However, due to the shared mesenchymal origins of CAFs and sarcoma cells, their heterogeneity and significant lineage plasticity, techniques and models that allow for lineage tracing and monitoring should also be incorporated into contemporary models (260,261). This incorporation would help towards an improved understanding of the phenotypical/functional heterogeneities and dynamics of CAFs/CAF-like cells and sarcoma cells in the TME.