Paracrine signaling in cancer-associated fibroblasts: central regulators of the tumor immune microenvironment

Abstract

Despite groundbreaking advances in cancer immunotherapy, clinical efficacy remains constrained by the immunosuppressive tumor microenvironment (TME). As key stromal components within this TME, cancer-associated fibroblasts (CAFs) emerge as pivotal regulators of drug resistance and immune evasion. Beyond establishing physical barriers that exclude cytotoxic T cells from tumor nests, that is creating an immune “desert”, CAFs dynamically reprogram the TME through multifaceted paracrine signaling, orchestrating crosstalk among tumor cells, stromal components, and immune cells. The complex paracrine signaling network jointly promotes the recruitment of immunosuppressive cells, alters the dynamics of immune cells, remodels the extracellular matrix, and ultimately establishes the immunosuppressive TME. Emerging strategies aimed at undermining the paracrine signaling Network of CAF-TME have shown potential in clinical studies to enhance the response to immunotherapy. Natural compounds such as curcumin and Baicalein and their derivatives have further expanded therapeutic approaches by regulating the paracrine phenotype of CAF due to their inherent multi-target intervention advantages. This review describes CAF and its paracrine effect as the central regulators of TME immunosuppression, emphasizing its key role in the immunotherapy response and providing new possibilities for clinical treatment strategies to restore CAFs paracrine-mediated immunosuppression and improve the efficacy of immunotherapy.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12967-025-06744-4.

Introduction

Cancer is a systemic disease that progresses through genetic mutations and dynamic interactions within the tumor microenvironment (TME) [1]. The complex pathological features limit the efficacy of the therapeutic interventions. Advances in the understanding of tumor biology and genetics have reshaped precision oncology, particularly targeted therapies, such as immunotherapy. Although promising clinical outcomes have been achieved in certain cancers (e.g., melanoma), primary or secondary resistance remains common, particularly in immunologically “cold” tumors characterized by an immunosuppressive TME [2]. A key limitation of existing therapies lies in their focus on rapidly proliferating tumor cells (the “seeds”) while neglecting the supportive role of the TME (the “soil”) [3].

The TME includes components such as tumor cells, immune cells, cancer-associated fibroblasts (CAFs), extracellular matrix (ECM). As the most abundant stromal population of TME, CAFs, upon activation, secrete fibronectin and collagen to form dense physical barriers, which not only impede effector T cell infiltration but also create an immune “desert” that suppresses antitumor immunity [4]. Furthermore, CAFs act as central hubs for cross-communication among malignant cells, immune cells, and stromal components. Through direct contact and paracrine signaling, they facilitate multicellular crosstalk and dampen local immune surveillance and antitumor responses [5, 6] (Fig. 1). Notably, CAF-derived paracrine signals exert effects beyond the local TME, broadly influencing immune regulation, therapeutic resistance, and metastatic progression [5, 7, 8],. Previous studies have demonstrated that elevated CAFs abundance correlates with reduced responses to anti-PD-1/L1 and anti-CTLA-4 therapies [9], and that CAFs mediate immunotherapy resistance by excluding CD8 + T cells from tumor nests [10]. Therefore, targeting the signals mediated by CAFs may represent a promising strategy to overcome immune exclusion and enhance the response to immunotherapy.

Fig. 1.

Schematic of heterogeneity and roles of CAFs in the TME

CAFs induce the recruitment and functional changes of immune cells by releasing and regulating intercellular communication, reshaping ECM and affecting matrix stiffness, thereby promoting tumor growth, invasion and metastasis, reshaping angiogenesis, regulating the immune microenvironment, and inducing therapeutic resistance [11, 12]. Notably, paracrine signaling encompasses not only the secretion of diverse mediators (including cytokines, chemokines, ECM components, and other signaling molecules that modulate cellular behavior and the local tissue microenvironment) but also exosome-mediated communication that facilitates long-distance interactions between CAFs and distal tumor cells [13–15]. Myofibroblastic CAFs (myCAFs) upregulate transforming growth factor-beta (TGF-β) to suppress T-cell activation, promote regulatory T (Treg) cell differentiation, and increase ECM deposition, forming a barrier that blocks CD8⁺ T-cell infiltration and activation, thereby promoting an immunosuppressive microenvironment [16–20]. Inflammatory CAFs (iCAFs) mediate the upregulation of cytokines (e.g., IL-6) and chemokines (e.g., C-X-C chemokine ligand [CXCL]12), which recruit immunosuppressive (e.g., Treg cells) to tumor sites, improving immune escape [21]. Antigen-presenting CAFs (apCAFs) are defined by the expression of antibodies such as HLA-DR and CD74. They exert antigen-presenting effects through MHC Class II molecules, induce the formation of Treg cells, and regulate tumor immunity [22, 23]. Owing to insufficient research on the heterogeneity of CAFs, a functional overlap exists among different subtypes, and crosstalk effects may exist among the paracrine signals mediated by them. For example, TGF-β regulates the expression of molecules such as IL-6 and CXCL12 to achieve multi-directional regulation [24, 25].

Clinical trials targeting CAF paracrine pathways are underway, with promising results [26–28]. For example, the TGF-β receptor I kinase inhibitor galunisertib reportedly enhances T-cell immunity and induces durable immune responses to kill tumor cells in mice, increasing response rates in patients with advanced colorectal cancer [29, 30]. Targeting CAF-derived IL-6 is beneficial for ECM remodeling and intervenes in the IL-6/JAK/STAT3 signaling pathway. The combination of IL-6 neutralizing antibodies and the JAK inhibitor tofacitinib is used to overcome matrix-induced drug resistance [31]. CAF-derived exosomes carry immunosuppressive molecules, such as PD-L1 and miRNAs, further promoting immune evasion [32]. Blocking their secretion may increase the efficacy of immune checkpoint inhibitors. Owing to their inherent multi-target intervention characteristics and immune regulatory properties, natural products, such as curcumin, simultaneously inhibit the activation of CAF and restore the function of immune cells by suppressing the secretion of TGF-β, thereby reversing immune escape [33, 34]. The importance of CAF paracrine signaling in tumor immunity is highlighted by its role in the resistance to checkpoint inhibitors, targeted therapies, CAR-T cell therapy, and cancer vaccines.

The TME comprises cancer cells, surrounding immune cells, CAFs, and various cytokines and chemokines. As central regulators of the TME, CAFs have inconsistent classification criteria and exhibit strong heterogeneity. Based on molecular markers, myCAFs, iCAFs, and apCAFs were identified. Spatially, CAFs can be divided into core and interfacial types. Glycolytic and lipogenic CAFs can be further distinguished based on their metabolic products and characteristics. Notably, CAF-TME crosstalk predominantly relies on paracrine signaling to drive pro-tumorigenic processes, and promoting angiogenesis, invasion and metastasis, immune modulation, and therapy resistance. In these pathways, cytokines (e.g., IL-1β, IL-6, TGF-β) and chemokines (e.g. CXCXL12, CXCL1) are released in large quantities, recruiting immunosuppressive cells (Treg cells and TAMs) and affecting the dysfunction of T/NK cells, collaboratively completing the re-editing of immune microenvironment. CAFs: cancer-associated fibroblasts; TME: tumor mincroenvironment; myCAFs: myofibroblast-like CAFs; iCAFs: inflammatory CAFs; apCAFs: antigen-presenting CAFs; IL-1β: Interleukin-1β; IL-6: Interleukin-6; TGF-β: transforming growth factor; CXCL12: Chemokine (C-X-C Motif) Ligand 12; CXCL1: Chemokine (C-X-C Motif) Ligand 1; PGE: Prostaglandin E; MMP: matrix metalloproteinases; Treg: regulatory T cell; TAMs: tumor-associated macrophages.

In conclusion, the current targeted treatment methods for eliminating CAFs have limited therapeutic efficiency due to insufficient subtype identification and other factors. The extensive role of paracrine signals provides the possibility for further targeting the multi-pathway regulation of the TME by CAFs. Although the mechanism through which CAFs reshape the TME through paracrine signals has been widely reported, most studies have focused on single subtypes [35] or isolated pathways (CXCL12/ C-X-C chemokine receptor [CXCR]4) [36]. The dynamic paracrine signaling network of CAFs and their interactions with immune cells have not been systematically analyzed and summarized. Therefore, we classified and sorted the executors of the paracrine pathways of CAFs, such as cytokines, chemokines and matrix components, comprehensively summarized the role of CAF paracrine signals in regulating the tumor immune microenvironment, and summarized the latest progress in targeting this pathway (including small-molecule inhibitors, antibody drugs, and natural products). The clinical transformation potential of this combination with immunotherapies (such as PD-1 inhibitors, CAR-T, etc.) has been extensively explored. This transforms the “cancer-promoting signals” of CAFs into “therapeutic mediators,” providing a novel perspective for clarifying CAF-mediated immunosuppression, with the expectation of developing effective combination therapy protocols and overcoming the barrier of tumor immunotherapy.

CAFs reshape the tumor immune microenvironment through paracrine signaling

CAFs interact with various components of the TME via multiple mechanisms [37], including direct ligand receptor interactions, autocrine signaling, paracrine signaling, and ECM remodeling. Generally, the paracrine pathway refers to the secretion of substances by cells into the extracellular space, which then act on neighboring cells. In addition to cytokine and chemokine secretion, the release of suppressive factors, exosomes and matrix components is considered a paracrine mechanism [38, 39]. To effectively understand the treatment dilemma caused by immunosuppression, we summarized and discussed the mechanism of action. The mechanism plot is shown in Fig. 2.

Fig. 2.

Mechanistic diagram of the regulatory role of CAFs in the TME via paracrine signaling

By cytokine- (e.g., TGF-β, IL-6) and chemokine- (e.g., CXCL12, CCL2) driven paracrine signaling, CAFs recruit immunosuppressive cells (e.g., M2-macrophages, N2-neutrophils, MDSCs, Treg cells), suppress the DC maturation and antigen presentation, induce immunosuppressive differentiation of monocytes into M2-TAMs and T cells into Treg cells, and downregulate effector functions of cytotoxic immune cells (e.g., CD8 + T cells). CAFs further mediate intercellular communication and metabolic reprogramming through exosomes release of circRNAs, miRNAs, and metabolites. CAFs promote extracellular matrix remodeling and increase stromal stiffness by activating integrin signaling pathways and secreting matrix metalloproteinases, thereby facilitating the formation of physical barriers that limit effector T-cell infiltration. Hypoxic conditions induced by CAF enrichment stimulate the secretion of inhibitory molecules (e.g., VEGF), which disrupt DCs and T-cell differentiation and upregulate immune checkpoint molecules (e.g., PD-L1 and CTLA-4), which drive immune evasion and resistance. CAFs: cancer-associated fibroblasts; TME: tumor microenvironment; TGF-β: transforming growth factor; IL-6: Interleukin-6; CXCL12: chemokine (C-X-C Motif) Ligand 12; CCL2: chemokine (C-C motif) ligand 2; MDSCs: myeloid-derived suppressor cells; DC: dendritic cell; M2-TAMs: M2-like tumor-associated macrophages; Treg cells: regulatory T cells; VEGF: vascular endothelial growth factor; PD-L1: programmed death-ligand 1; CTLA-4: cytotoxic T lymphocyte-associated antigen 4.

Secretion of cytokines regulates surrounding cell functions

As tumors progress rapidly, TGF-β is abundantly secreted by CAFs into the TME, where it exerts various pro-cancer effects through paracrine signaling [40]. For example, it promotes tumor metastasis via epithelial-mesenchymal transition (EMT), enhances angiogenesis, and assists tumor cells in immune evasion. TGF-β can also convert normal fibroblasts into CAFs. Therefore, elucidating the intrinsic mechanisms behind the “betrayal” of TGF-β has become a key objective in reversing cancer treatment dilemmas.

As a critical regulator of both innate and adaptive immunity, TGF-β controls almost every stage of the immune response triggered by tumors. In innate immunity, the activation and cytotoxic activity of NK cells are significantly inhibited by TGF-β [41, 42]. TGF-β secreted by CAFs can downregulate the expression of MHC class II molecules and the costimulatory molecules CD40, CD80, and CD86 on dendritic cells (DCs), thereby reducing their ability to activate cytotoxic T-cell responses and their antigen-presenting functions [43]. TGF-β can also promote the conversion of monocytes into M2 phenotype tumor-associated macrophages (TAMs) [44, 45]. Similar to TAMs, neutrophils can be induced by TGF-β to transform into a tumor-promoting phenotype, accelerating CD8⁺ T-cell exhaustion and forming an immunosuppressive microenvironment. Research has revealed that blocking TGF-β enhances the recruitment and activation of antitumor neutrophils [46–49].

In adaptive immunity, CAF-derived TGF-β can induce naive T cells to differentiate into Tregs cells [50]. CAFs reduce the recruitment of CD8⁺ T cells by releasing IL-6 and TGF-β while inhibiting their cytotoxic activity against tumor cells [51, 52]. Coculturing CAFs with tranilast (a CAF inhibitor) resulted in reduced levels of TGF-β and inhibited the induction of Treg cells and myeloid-derived suppressor cells (MDSCs) [53]. TGF-β pathway inhibitors, such as galunisertib, can modulate antitumor immune responses and enhance the efficacy of anti-PD-L1 monotherapy [30]. In summary, TGF-β plays a significant role in the interaction between CAFs and immune cells, and inhibiting its activity is considered an effective antitumor strategy.

There is a close association between inflammation and tumor development. The interleukin family plays an important role in tumor progression and affects the immune response. Furthermore, IL-32 derived from CAFs binds to the cell surface integrin β3, activating the p38 MAPK pathway and further promoting the invasion and metastasis of breast cancer cells [54]. IL-33 expression increases in metastasis-associated fibroblasts, reshaping the immune microenvironment into an inflammatory niche suitable for metastasis [55]. Moreover, IL-6 can induce changes in the TME, promote tumor stemness, and influence tumor development and immune responses [56]. In esophageal cancer, IL-6 secreted by CAFs has been confirmed as a crucial mediator in the TME that interacts with tumor cells through paracrine mechanisms [57]. Recent studies have revealed that IL-6 secreted by CAFs may help maintain the paracrine loop, connect communication between CAFs and tumor cells, enhance EMT, and increase immune evasion [58].

CAFs activate the STAT3 pathway through IL-6 paracrine signaling, inducing PD-L1⁺ neutrophils, and thereby promoting immune suppression in hepatocellular carcinoma [59]. Similarly, in hepatocellular carcinoma, IL-6 secreted by CAFs can mediate STAT3 pathway activation, causing recruited DCs to express Indoleamine-2,3-Dioxygenase and impair T-cell responses [60]. CAFs recruit more monocytes through IL-6 paracrine signaling and convert them into M-MDSCs [45], leading to the widespread suppression of T-cell proliferation and function [61, 62]. Monocytes can also be transformed into M2-like TAMs under the synergistic induction by IL-6 and GM-CSF [63]. In addition to inducing and recruiting immunosuppressive cells, CAFs can directly reduce CD8⁺ T-cell recruitment through IL-6 release [63].

Secretion of chemokines affects immune cell migration and infiltration

CXCL12, a CXC chemokine, is an important factor in various pathological and physiological processes. CXCL12 is secreted primarily by CAFs and can recruit specific immune cell subsets by binding to CXCR4 or CXCR7, leading to an immunosuppressive state in the TME and promoting tumor progression [64, 65].

In breast cancer, CXCL12 influences innate immunity by promoting the recruitment of monocytes to the tumor site [66]. A similar role has been observed in oral squamous cell carcinoma in, which CAF-derived CXCL12 attracts monocytes with an M2-like phenotype and induces their conversion into M2 macrophages [67]. Houthuijzen et al.. recently revealed that CAFs in breast cancer recruit myeloid cells in a CXCL12-dependent manner [68]. The CXCR4/CXCL12 pathway mediates numerous interactions between myeloid cells and CAFs, including regulating breast cancer metastasis [69, 70], tumor cell migration, and EMT [71], and aids in the recruitment of Treg cells within the tumor [72]. Another study reported that the estrogen-induced binding of CAF-derived CXCL12 to CXCR4 participates in the recruitment of mast cells [73].

In adaptive immunity, the immunosuppressive effects of CXCL12 primarily involve T cells. In TNBC, the CAF-S1 subgroup induces a high content of immunosuppressive FOXP3⁺ Treg cells and downregulates CD8⁺ T-cell levels through CXCL12 secretion [72]. The importance of the CXCL12 signaling pathway in regulating the migration of tumor-infiltrating CD8⁺ T cells induced by FAP⁺ CAFs has been confirmed in several reports [74, 75]. Dudeja et al. [76]. reported that the possible mechanism through which pancreatic cancer is an immune “cold” tumor is related to CAFs; specifically, CAF-derived CXCL12 prevents cytotoxic T-cell infiltration into tumors and induces immune evasion. By blocking the binding of CXCL12 and its receptor CXCR4, AMD3100 (a CXCR4 inhibitor) synergizes with anti-PD-L1 antibodies to rapidly promote T-cell accumulation and effectively eliminate cancer cells [77]. In a metastatic breast cancer model that is typically resistant to immunotherapy, blocking the CXCL12 receptor, CXCR4, reduces connective tissue proliferation, increases cytotoxic T lymphocyte (CTL) infiltration, and enhances the efficacy of immunotherapy [78].

CCL2 is an effective chemotactic factor that recruits immunosuppressive cells and is crucial for regulating the TME [79]. CAFs are the primary source of CCL2 and participate in the recruitment of MDSCs and TAMs through the paracrine pathway, thereby forming an immunosuppressive TME [75, 80, 81]. In colon cancer, CCL2 contributes to the accumulation of MDSCs within tumors, affecting MDSC-mediated CD4 + and CD8 + T-cell suppression [82]. In lung squamous cell carcinoma, CAFs promote the migration of peripheral CCR2⁺ monocytes through CCL2, which are then reprogrammed into M-MDSCs [83]. In intrahepatic cholangiocarcinoma, CAFs enhance the recruitment of MDSCs via the STAT3‒CCL2 pathway [75]. In breast cancer, monocytes are converted into TAMs through CCL2 induction [84]. Through the paracrine pathway of CCL2, CAFs not only influence MDSCs and TAMs but also expand the Treg cell population within tumors [85].

CCR2 is the primary functional receptor for CCL2. Targeting the CCL2/CCR2 axis can inhibit the recruitment of inflammatory monocytes, infiltration of TAMs, and M2 polarization, thereby reversing the immunosuppressive state of the TME and activating antitumor CD8⁺ T-cell responses [86]. Previous clinical trials have reported that the CCR2 inhibitor PF-04136309 combined with chemotherapy significantly reduced the number of TAMs and Treg cells while increasing the infiltration of tumor-infiltrating lymphocytes, providing a novel approach for pancreatic cancer treatment [87]. Additionally, CCL5 released by CAFs into the TME can increase the ability of Treg cells to kill CD8⁺ T cells, which has been a proven mechanism of immune evasion in colorectal cancer [88]. Further clinical studies are needed to verify its safety and efficacy as a promising therapeutic target, which could lead to the development of additional treatment options for patients.

Release of exosomes involves intercellular information transfer

Exosomes carry various types of biological information, facilitating communication and information transfer between cells. Although current studies have mostly focused on the exosomes produced by cancer cells, the behavior of exosomes derived from CAFs has received increasing attention. The interaction between the TME and cancer cells largely depends on CAFs [89]. Exosomes originating from CAFs package bioactive molecules and can reshape the TME by interacting with other cells and the ECM [37]. Owing to their diverse cellular precursors, exosomes can broadly participate in several processes related to tumor progression, including immune responses, cell metabolism, and angiogenesis [90]. Information is transmitted through signal exchange via exosomes among stromal, immune, and tumor cells, regulating signaling pathway activity and promoting tumor progression and immune-related inflammation.

Previous studies have revealed that CAF-derived exosomes containing circTBPL1 can be transferred to breast cancer cells, linking the TME with cancer cells and reshaping the microenvironment to promote tumor progression [91]. In hepatocellular carcinoma, exosomes containing circHIF1A promote EMT process, suppress CD8 + T-cell cytotoxicity and activity, upregulate PD-L1 expression, and induce immune evasion [92]. In pancreatic cancer, exosomes released by CAFs activate the NOD1 pathway through PTGS2, promoting M2 macrophage polarization [93]. A systematic review [94] indicated that CAF-derived exosomes containing miR-20a-5p in hepatocellular carcinoma may control the TME by downregulating the Wnt/β-catenin signaling pathway. In the prostate, CAF-derived exosomes regulate GSK3β/β-catenin signaling, influencing cell growth and metastasis.

Furthermore, exosomes carry various bioactive substances, including proteins, metabolites, and nucleic acids, which can be transported in circulation and internalized by recipient cells to exert their effects. Therefore, exosomes play a vital role in intercellular communication. Compared with other substances, they provide increased specificity to certain cells and resistance to circulating RNases, making them ideal carriers for bidirectional communication [95, 96]. The dysregulation of miRNAs in melanoma promotes CAF activation through EMT induction, which in turn alters the secretion phenotype of CAF exosomes, creating a feedback signal linking CAFs with tumor cells, influencing information expression, and promoting immune suppression and metastasis [97].

Altering matrix components affects intercellular interactions

Matrix metalloproteinases (MMPs) are a large family of Zn2+- and Ca2+-dependent proteolytic enzymes produced by tumor and stromal cells that participate in the degradation of various proteins in the ECM [98]. Their activities are regulated by the tissue inhibitors of metalloproteinases (TIMPs). MMPs have been detected in various human cancers, and their expression is typically associated with reduced survival rates in colorectal, lung, and breast cancers [99–101]. As research progresses, Khalil et al. reported that MMPs can promote cell differentiation, closely linking them to angiogenesis and immune responses during tumor progression [102]. MMPs form a complex paracrine network with cytokines such as TNF-α, CXCL12, and IL-2, regulating T lymphocyte interactions with tumor cells [103].

CAFs are the main source of MMPs in the TME [104–106]. They remodel ECM and regulate immune cell proliferation by secreting MMPs [107]. The ECM protein network remodeled by CAFs acts as a physical barrier to immune cells (particularly T lymphocytes), inhibiting their recruitment to cancer sites and reducing opportunities for participation in TME immune responses [108]. NK cells are important innate immune cells in the body, and MMPs secreted by CAFs have been found to interfere with the antitumor activity of that [109]. Moreover, the pan-MMP inhibitor, GM6001, restored the sensitivity of melanoma tumor cells to NK cell-mediated lysis. Furthermore, MMP-2/9 inhibitors, such as SB-3CT [110], can downregulate PD-L1 expression, inhibit MDSC and Treg cell infiltration, and increase CD8 + T-cell infiltration rates. These results suggest that targeting MMPs is important for antitumor drug development, although their large and complex compositions hinder the identification of effective targeted drugs.

Production of inhibitory molecules directly suppresses immune cell function

In addition to the major cytokines and chemokines mentioned earlier, various CAF-derived inhibitory molecules can exert immunosuppressive effects on the TME via paracrine pathways. Under hypoxic conditions, CAFs secrete vascular endothelial growth factor (VEGF), which is involved in the abnormal differentiation of DCs and impaired antigen presentation, driving immune suppression in the TME [111]. In a review by Terme et al., VEGF-A was shown to modulate immune cells (DCs, MDSCs, and TAMs) to induce the accumulation of Treg cells, while suppressing T-cell function. VEGF also promotes immune tolerance by upregulating PD-L1 on DCs and CTLA-4 on CD8 + T cells [112, 113].

In addition to inducing the recruitment and function of immune cells, CAFs upregulate the expression of immune checkpoints on the cell surface via paracrine pathways, thereby inducing immune tolerance [5]. α-SMA⁺ CAFs increase PD-L1 expression in lung adenocarcinoma cells by secreting CXCL2, thereby affecting antitumor immunity [114]. In melanoma and colorectal cancer, CAF-derived CXCL5 increases PD-L1 expression in a PI3K/AKT-dependent manner, promoting the immunosuppressive microenvironment [115].

CAF-derived cardiotrophin-like cytokine 1 (CLCF1) promotes the infiltration and polarization of tumor-associated neutrophils (TANs) through paracrine mechanisms [116]. Similarly, CAF-secreted CXCR2 enhances the migration of TANs [117]. CAFs induce M2 polarization of macrophages and suppress NK cell function via IL-8 secretion. Other molecules with similar functions include monocyte chemotactic protein-1, macrophage colony-stimulating factor (M-CSF), and chitinase 3-like 1 [118, 119]. According to research by Goehrig et al. [120]., CAFs directly inhibit the function of CD8⁺ T cells through the secretion of βig-h3 (an ECM protein also known as TGF-βi), affecting their proliferation, activation, and cytotoxic activity.

Interactions of paracrine signals reshape the immune microenvironment

The immunosuppressive microenvironment mediated by CAFs is not driven by isolated signaling pathways, but by the dynamic crosstalk synergy among paracrine factors. Firstly, TGF-β has pleomorphic effects on the biological behavior of CAFs. Paracrine TGF-β increases the contractility of CAFs and regulates ECM remodeling owing to its activation of non-classical TGF-β /RhoA/ROCK axis signaling [121]. In addition, TGF-β can stimulate the expression of factors such as IL-6, CXCL12 and VEGFA; recruit immune cells; induce angiogenesis and promote the process of tumor metastasis [24]. Among them, regulating CXCL12/CXCR4 recruits Treg cells and MDSCs to enhance immune rejection [25], while simultaneously triggering the G-protein coupled/PI3K/AKT/NF-κB axis and the Ras-MEK1/2-Erk1/2 axis, leading to angiogenic responses [122]. Notably, bidirectional regulation exists between CXCL12 and TGF-β, which cooperatively promotes cell invasion [123] and activates the Wnt/β-catenin signaling pathway to drive epithelial-mesenchymal transition (EMT) and drug resistance [124]. Additionally, TGF-β upregulates the expression of IL-6, promotes the survival of TAMs, and activates the STAT3 signaling pathway, creating a tumor-promoting and immune resistance environment [125, 126]. IL-6 further contributes to systemic immunosuppression through metabolic reprogramming [127]. In the hepatocellular carcinoma (HCC) TME, CAF-derived CLCF1 increases CXCL6 and TGF-β secretion, promoting TAN infiltration and polarization in a paracrine manner. Concurrently, HCC cell-secreted CXCL6 and TGF-β activate ERK1/2 phosphorylation, forming a positive feedback loop that accelerates tumor progression [128].

In addition, exosome-mediated communication is central to establishing paracrine signaling networks. In bladder cancer, normal fibroblasts internalize tumor-derived exosomes, promoting CAF proliferation, activation, and marker expression. Exosome-mediated TGF-β/SMAD signaling sustains the malignant CAF phenotype [129]. Exosomal miRNAs and lncRNAs extensively modulate pathways including TGF-β/SMADs, JAK/STAT, and MAPK, reinforcing feedback loops during CAF transformation and regulating CAF-supported angiogenesis, glucose metabolism, and stromal remodeling [94, 130]. Exosome-packaged microRNA-21 synergizes with IL-6 to activate downstream STAT3 signaling, driving M-MDSC generation [62]. Exosome-mediated signaling in the crosstalk among M2 macrophages, CAFs, and cancer stem cells represents a pivotal event in tumorigenesis, promoting a self-sustaining immunosuppressive niche [131]. This multilayered crosstalk underscores the necessity to target nodal points (e.g., TGF-β, IL-6, CXCL12 axis, or exosome trafficking) rather than single factors to overcome CAF-mediated immune evasion.

Targeting therapeutic strategies

With a deeper understanding of the immunosuppressive microenvironment mediated by CAF-derived paracrine signaling, targeting key effector molecules and pathways has emerged as a viable approach for reversing immunotherapy resistance. Paracrine signaling has garnered widespread attention as the most critical intervention target. Therapeutic strategies targeting CAF paracrine pathways focus on suppressing CAF activation, blocking secretion of immunosuppressive factor, and modulating tumor cell–CAF–immune cell crosstalk.

Advances in targeted pharmacological agents

CAFs promote pro-tumorigenic environments through aberrant paracrine signaling and matrix remodeling, and their bidirectional interactions with cancer cells accelerate tumor progression. We have summarized and reviewed the current research progress on CAF paracrine signaling drugs and their therapeutic targets in Table 1. Therapeutic interventions targeting critical CAF components and effectors have yielded promising results in both preclinical models and clinical trials (Supplementary Table S1).

Table 1.

Names, classifications, and mechanisms of action of small-molecule inhibitors that target CAFs

Class	Name	Mechanism	Cancer type	Reference
TGF-β receptor inhibitor	LY2157299	Upregulation of chemokine CXCL9, CXCL10, and CXCL11 and chemokine receptor CXCR3 of CAR-T, enhancing CAR-T accumulation at lymphoma sites	Lymphoma	[151]
	LY3200882	Downregulation of ECM expression in CAFs Promote the infiltration of effector T cells Induction of dendritic cell maturation	Breast cancer	[133]
	SB525334	Breaking CAF barrier helps to normalize the microvasculature Inhibition of CAFs activity marked by α-SMA and FAPα Improves the function of cytotoxic T lymphocytes Restore the immunosuppressive TME activity	Oral squamous cell carcinoma, Pancreatic cancer	[8, 152]
	SB431542	Inhibition of CAF activation and tumor cell activity Inhibition of epithelial–mesenchymal transition	Breast cancer, Non-small cell lung cancer	[153, 154]
	Tranilast	Increases TME perfusion and oxygenation Improves immune stimulation Inhibition of CAF-induced carcinoma cell survival Promotes normalization of TME	Breast cancer, Nasopharyngeal carcinoma	[155, 156]
IL-6 inhibitor	Tocilizumab	Inhibition of IL-6Rα and downstream signaling pathways and altering dynamic crosstalk between tumor cells and CAF in the TME Inhibition of CAF activity Inhibition of CAF-promoted epithelial–mesenchymal transition Inhibiting the secretion of pro-angiogenic factors by CAFs, inhibiting angiogenesis	Breast cancer, Cholangiocarcinoma, Hepatocellular carcinoma, Head and neck squamous cell carcinoma	[139, 157–159]
CXCL12/CXCR4 antagonist	Plerixafor (AMD3100)	Inhibiting the differentiation of CAFs Alter the recruitment of Treg cells Promote T-cell activation Reduce the growth stimulating activity of CAFs under hypoxia	Breast cancer, Cholangiocarcinoma, Gastric cancer, Human parathyroid neoplasia Pancreatic cancer	[77, 146, 160–163]
	Olaptesed pegol (NOX-A12)	Enhanced infiltration of T and NK cells in a dose-dependent manner, improving immunotherapy	Colorectal cancer, Pancreatic cancer	[144, 145]
COX-2 inhibitor	Celecoxib	Enhanced Th1 phenotype and decreased production of phenotypic markers of Treg cells Reduced PGE2 production and α-SMA expression in CAFs	Gastrointestinal stromal tumor	[164]
Exosomal inhibitor	miRNA-320a	Inhibiting CAF-derived exosomes Inhibiting the M2 polarization of macrophages Inhibiting the malignant behavior of cancer cells	Pancreatic cancer	[150]
	GW4869	Inhibiting the survival of epithelial cells cocultured with CAFs and the expression of chemoresistance-inducing factor in recipient epithelial cells	Pancreatic cancer	[165]
IGF inhibitor	Linsitinib	Enhanced response to immune checkpoint blockade Promote T-cell infiltration	Breast cancer, Colorectal cancer	[149]

Abbreviations: α-SMA, alpha-smooth muscle actin; CAF, cancer-associated fibroblast; COX-2, cyclooxygenase-2; CXCR, C-X-C chemokine receptor; CXCL, C-X-C chemokine ligand; FAPα, fibroblast activation protein-alpha; IGF, insulin-like growth factor; IL-6, interleukin 6; TGF-β, transforming growth factor beta; TME, tumor microenvironment

Among the paracrine pathways, TGF-β and related cytokines remain the most extensively studied therapeutic targets. The TGF-β and IL-1 signaling pathways are pivotal for CAF activation, making their inhibition a potential strategy to enhance antitumor immunity [132]. Accumulating evidence shows that TGF-β receptor 1 inhibitors (e.g., galunisertib) or TGF-β-neutralizing antibodies synergize with immune checkpoint blockade (ICB) to induce robust and sustained T-cell responses [18]. In a previous study, TGF-β blockade generated a CD73⁺IFN- licensed CAF subset that secreted CXCL9/10/11 chemokines, promoted T-cell recruitment, suppressed tumor growth, and synergized with ICB [4]. Another study demonstrated that SB525334 effectively restored CTL function and normalized the immunosuppressive TME [8]. Zhang et al. [133]. developed a site-specific nanodelivery system in which the TGF-β receptor inhibitor LY3200882 and PD-L1 siRNA were co-administered to improve the immune response. This system significantly downregulates CAF-derived ECM expression, enhances effector T-cell infiltration and drug penetration, triggeres immunogenic tumor cell death, and induces DC maturation.

Phase II clinical trial data indicated that continuous administration of galunisertib alone for 14 days improved hematological parameters in patients with low-to-intermediate-risk myelodysplastic syndrome, resulting in a median response duration of 90 days with favorable clinical safety [134]. In clinical trials targeting solid tumors (e.g., lung, liver, and colorectal cancers), galunisertib combined with chemotherapy or targeted therapies has also demonstrated expected clinical efficacy, showing initial therapeutic benefits along with good tolerability [29, 135, 136]. In addition, phase I/II clinical trials are currently ongoing for other small-molecule TGF-β receptor inhibitors, such as TEW-7197 and LY3200882. Although formal research results have not yet been published, they are worth looking forward to.

Activated CAFs secrete high levels of pro-invasive/metastatic IL-6. IL-6 blockade therapy can effectively improve T-cell function, increase the efficacy of conventional immunotherapy, and improve patient prognosis [51, 52]. The IL-6-targeted monoclonal antibody tocilizumab reduces the number of MDSCs, stimulates CD8⁺ T-cell activation, and enhances antitumor efficacy [137, 138]. Tocilizumab can also normalize activated mammary CAFs in breast cancer by suppressing the IL-6/STAT3/AUF1 pathway and inhibiting their pro-tumorigenic paracrine effects [139]. It also overcomes CAF-mediated 5-FU resistance in gastric cancer and disrupts IL-6-dependent crosstalk between esophageal cancer cells and CAFs [31, 57]. In head and neck squamous cell carcinoma, IL-6 inhibition alters Th17-CAF interactions and reshapes immune cell infiltration [140]. Recently, Diab et al. [141]. reported surprising findings that IL-6 blockers can not only promote the infiltration of CD4⁺/CD8⁺ effector T cells but also reduce Th17, macrophages, and myeloid cells, importantly alleviating immune-related adverse events (irAEs). IL-6 plays a crucial role in the tumor immune microenvironment, potentially enhancing the efficacy of ICB and serving as a viable therapeutic intervention.

The CXCL12/CXCR4 signaling axis has been demonstrated to modulate tumor progression and reshaped microenvironmental states in diverse preclinical models. As classical inhibitors targeting this axis, olaptesed pegol and plerixafor have undergone extensive clinical investigations. Olaptesed pegol, an L-ribose oligonucleotide, binds to and neutralizes CXCL12. In patients with relapsed/refractory chronic lymphocytic leukemia, this agent exhibites favorable tolerability as a monotherapy. A combination regimen of bendamustine and rituximab achieved an overall response rate exceeding 80%, including complete response (CR) in 11% of patients [142]. Clinical trials combining pegol with radiotherapy have further confirmed its satisfactory safety profile [143]. Notably, the CXCL12 inhibitor NOX-A12 (olaptesed pegol) [144] enhances tumor-infiltrating T cells, disrupts the immunosuppressive TME, and expands the applicability of immune checkpoint inhibitors by improving antitumor immune responses. Similar efficacy was observed in colorectal and pancreatic cancers [144, 145].

Plerixafor (AMD3100), a small-molecule CXCR4 antagonist, has received Food and Drug Administration approval for use in patients with lymphoma or myeloma undergoing autologous transplantation. Preclinical studies have revealed that plerixafor synergizes with bortezomib to reverse chemoresistance in cholangiocarcinomas. This combination suppresses CAF activity while promoting T-cell activation, thereby sensitizing cholangiocarcinoma to anti-PD-1 therapy and significantly reducing the tumor burden [146]. In addition, in a study by Feig [77], FAP + CAFs were a major source of CXCL12, directing immune escape in a pancreatic ductal carcinoma model. AMD3100 treatment can significantly alter the infiltration and recruitment of T cells, and its combination with anti-PD-L1 immunotherapy can significantly prevent immune escape and improve therapeutic effects.

In addition to the abovementioned targeted inhibitors, novel approaches are being developed. CAFs can also mediate the production of the tumor inflammatory microenvironment by secreting COX-2. Celecoxib, a COX-2 inhibitor combined with neoadjuvant toripalimab, is a potential treatment option for patients with locally advanced colorectal cancer with mismatch repair deficiency or high microsatellite instability [147]. This treatment is associated with a high pathological CR rate and an acceptable safety profile for patients. In breast cancer, it can reduce PGE2 production and α-SMA expression in CAFs, enhance Th1 responses (T-bet/IFN-γ), and suppress Treg cell markers (FoxP3/IL-10/TGF-β1), thereby improving immune responses [148].

In immunologically “cold” tumors, insulin-like growth factor 2 (IGF2) secreted by CAFs contributes to immune exclusion. Genetic ablation of IGF2 or pharmacological inhibition via linsitinib (an IGF1R inhibitor) significantly potentiates the ICB response [149]. Exosomal miRNA-320a blockers counteract CAF-derived exosome-induced M2 macrophage polarization and inhibit cancer malignancy [150]. These preclinical and clinical findings highlight the therapeutic potential of targeting the CAF paracrine pathways to reverse immunosuppression and remodel the TME.

Advances in natural products and derivatives

Although small-molecule-targeted drugs are promising, challenges such as adaptive resistance and narrow therapeutic spectra limit their clinical utility [166]. Natural products exhibit inherent advantages in terms of immune regulation and anticancer potential. Their clinical efficacy has gained recognition, particularly when modified with nanotechnology, novel material encapsulation, or combination therapies. Notably, more than 75% of antitumor agents originate from natural sources [167]. These natural products typically modulate multiple signaling pathways and functions within the TME, exerting comprehensive effects on cancer progression and immune remodeling [168]. Details are presented in Table 2.

Table 2.

Names, classifications, and mechanisms of action of natural drugs and their derivatives that target CAFs

Class	Name	Mechanism	Cancer type	Reference
Phenols	Curcumin	Inhibition of CAF paracrine pathways, such as TGF-β, and α-SMA and p-SMAD2/3 expression in CAFs Inactivating CAFs, promote CAFs to reverse their phenotype to peritumoral fibroblast-like cells Inhibiting the invasion and migration of cancer cells and increasing the levels of tumor	Breast cancer, Tongue squamous cell carcinoma.	[33, 169, 171]
	EGCG	Reduced the secretion function of CAFs Regulates T-cell activity and enhances immunity Targeting c-Met in tumor cells and HGF and VEGF production in interstitial fibroblasts Targeting aerobic glycolysis and reduces CAF lactate production to impede cancer cell metabolic coupling	Colorectal cancer, Prostate Cancer.	[179, 181]
	Eugenol	Inhibition of CAF paracrine, migration, and proliferation potential Inhibition of DNA methyltransferase gene expression in breast CAF cells Inhibiting the production of angiogenic effectors VEGF-A and IL-8 Inhibiting the paracrine pro-angiogenic effect of CAF cells	Breast cancer	[191, 192]
	Salvianolic acid B	Inhibiting TGF-β1 secretion to interfere with CAF activation Inhibiting the infiltration of MDSCs and Treg cells Increased recruitment of CD4 + and CD8 + T cells and expression of Th1 cytokines in the tumor region	Breast cancer	[188]
Flavonoids	Apigenin	Downregulation of TNF-α-mediated CCL2 release prevented the infiltration and migration of CAFs into the tumor environment	Breast cancer	[193]
	Baicalein	Inhibition of TGF-β signaling pathway Reduce the expression of IL-6, CCL2, and CCL5 in the tissue to avoid the activation of CAFs Promote the switch of tumor-associated macrophages from the M2 phenotype to the M1 phenotype Increase the infiltration of cytotoxic T cells and activate the tumor immune microenvironment	Breast cancer	[189]
	Isoliquiritigenin	Inhibition of CAF paracrine pathways, such as TGF-β, and α-SMA and p-SMAD2/3 expression in CAFs	Gastric cancer	[194]
	Puerarin	Reduced profibrotic cytokines including FGF-2, PDGF-B, and TNF-α Promote CAF inactivation and downregulate CCL2 and CCL5 expression Promote CD8 and CD4 T-cell infiltration, reduce the infiltration of Treg cells and MDSCs and promote the transformation of M2 macrophages to M1 phenotype	Breast cancer	[190]
	Quercetin	Downregulating Wnt16 expression in CAFs and significantly downregulating fibrosis and reducing matrix barrier	Breast cancer	[195]
	Silibinin	Reducing the expression of monocyte chemoattractant protein-1 and related biomarkers in CAFs Promote the recruitment of immune cells in the TME	Prostate cancer	[196]
Polysaccharides	MPSSS	Targeting the CAF paracrine pathway to weaken the inhibitory effect of CAFs on CD4 + and CD8 + T cells	Colorectal cancer, Prostate Cancer.	[177, 197]
	Polysaccharides from polygonatum	Stimulating autophagy in prostate CAFs and inhibition of CAF growth in prostate cancer	Prostate cancer	[178]
	Prunella vulgaris Polysaccharide	Inhibition of CAF proliferation, increase the proportion of CAFs in apoptosis, and blocking the crosstalk between CAFs and breast cancer tumor cells by inhibiting the expression of FGF-2	Breast cancer	[198]
Alkaloids	Alkaloids	Inhibition of fibroblast proliferation and ECM formation Inhibit the synthesis and secretion of paracrine factors to inhibit tumor–matrix interaction	Hepatocellular Carcinoma, Pancreatic cancer.	[173, 174]
	Berberine	Inhibition of epithelial–mesenchymal transition Promote CAF-induced apoptosis of colonic epithelial cells	Colon cancer	[199]
	Matrine	Reduce the secretion of CAF exosomal circSLC6A5	Colorectal cancer	[200]
Terpenoids	Artemisinin	Inhibition of TGF-β activation	Breast cancer	[201]
	Asiatic acid	Inhibit the levels of CAF-derived collagen and reduce the collagen deposition in the matrix region	Colon cancer	[202]
	Cannabinoids	Inhibition of TGF-β activation and α-SMA and MMP-2 expression	Prostate Cancer	[203]
	Paeoniflorin	Inhibit the secretion of CAFs by activating microRNA-149	Gastric Cancer	[172]
	Triptolide	Downregulating the expression of fibrosis-related genes in CAFs of pancreatic cancer Correcting the abnormal expression of microRNA in gastric cancer CAFs	Gastric cancer, Pancreatic cancer	[204, 205]
	Ursolic acid	Inhibition of CXCL12 paracrine in CAFs	Papillary thyroid carcinoma	[206]
Saponins	Astragaloside IV	Inhibition of CAF activation in gastric cancer by HOXA6/ZBTB12 axis	Gastric cancer	[207]

Abbreviations: α-SMA, alpha-smooth muscle actin; CAF, cancer-associated fibroblast; CCL, C-C motif ligand; c-Met, cellular-mesenchymal epithelial transition factor; CXCL12, C-X-C chemokine ligand 12; ECM, extracellular matrix; FGF-2, fibroblast growth factor-2; HGF, hepatocyte growth factor; HOXA6, homeobox A6; IL, interleukin; MDSC, myeloid-derived suppressor cell; MMP-2, matrix metalloproteinase-2; PDGF-B, platelet-derived growth factor subunit B; p-SMAD2/3, phosphorylated mothers against decapentaplegic homolog 2/3; TGF-β, transforming growth factor beta; TME, tumor microenvironment; TNF-α, tumor necrosis factor alpha; VEGF, vascular endothelial growth factor; Wnt16, wingless-type MMTV integration site family member 16; ZBTB12, Zinc Finger and BTB Domain Containing 12

TGF-β is a dominant paracrine pathway in the TME, and tumor cells can activate the synthesis of TGF-β by CAFs, thereby inducing tumor cell activation. After curcumin treatment, the expression of TGF-β and the phosphorylation of Smad2 are significantly reduced [169]. Curcumin, a natural bioactive compound derived from turmeric, exhibits anticancer and immune-stimulating activities against many types of tumors. It can be targeted and delivered to specific cells after nanomaterial encapsulation to exert its effects [170]. Ba et al. [171]. f found that curcumin downregulated α-SMA and p-SMAD2/3 expression in CAFs; inhibited the expression of the CAF paracrine signals MMP-2, stromal cell-derived factor-1 (SDF-1), and TGF-β1; reversed the phenotype of CAFs to peritumoral fibroblast-like cells; and inhibited the proliferation and tumorigenicity of tongue squamous cell carcinoma cells. In breast cancer cells, curcumin inhibits the invasion/migration, inactivates breast CAFs, and reshapes the immune microenvironment by reducing the secretion of TGF-β, IL-6, MMP-2, MMP-9, and SDF-1 [33].

The interleukin family is widely involved in regulating the activation, proliferation and differentiation of immune cells, as well as the regulation of inflammatory responses. Paeoniflorin inhibits IL-6 secretion from CAFs via microRNA-149 activation, reducing p-STAT3, MMP-2, and MMP-9 levels [172]. In pancreatic cancer, Conophylline (CnP) inhibits tumor-matrix interactions by inhibiting the synthesis and secretion of CAF paracrine factors, specifically IL-6, IL-8, CCL2, and CXCL12. The degree of connective tissue hyperplasia in tumors containing SUIT-2 ⁺ CAF was reduced. This enhanced the drug delivery rate of gemcitabine, providing a novel approach for combination therapy [173]. In HCC, a previous study [174] confirmed that CnP targets CAF paracrine pathways, such as IL-6, IL-8, CCL2, angiogenin, and osteopontin pathways; induces PD-L1⁺ neutrophils; enhances the recruitment of MDSCs; and promotes immunosuppression of HCC. In further in vivo experiments, CnP was combined with the small-molecule drug sorafenib to verify its synergistic and toxic effects.

Chemokine levels tend to be regulated by natural polysaccharides, which can suppress MMP expression, balance Th1/Th2 responses, enhance DC antigen presentation, and eliminate TAMs to stimulate antitumor immunity [175]. ZSP4 extracted from Dictyophora indusiata reversed CAF-mediated immunosuppression and restored CD4⁺/CD8⁺ T-cell proliferation [176]. Lentinus edodes polysaccharide MPSSS attenuates CAF-induced suppression of CD4⁺/CD8⁺ T cells via TLR4‒NF-κB pathway activation [177]. In addition to their effective immune regulation, the safety of polysaccharides has been widely recognized. Polygonatum polysaccharides selectively induce autophagy in prostate CAFs by upregulating Beclin-1 and LC3 and inhibiting CAF growth without affecting normal fibroblasts [178]. Selective inhibition of CAFs provides new possibilities for targeted CAF therapy.

In addition, the CAF-mediated paracrine pathways involve diverse mechanisms, including the secretion of immunosuppressive molecules, metabolites, and exosomes. In vitro, epigallocatechin gallate (EGCG) reduces hepatocyte growth factor and VEGF production in interstitial fibroblasts [179] and disrupts CAF metabolic coupling by inhibiting aerobic glycolysis, thereby increasing TNF-α-mediated T-cell bystander killing [180, 181] and reducing immune escape. Duan et al. [182]. showed that quercetin downregulates Wnt16 in CAFs, reducing fibrosis and matrix barriers to improve paclitaxel efficacy in breast cancer [182]. In gastric cancer, astragaloside IV suppresses CAF-derived M-CSF and upregulates TIMP2, thereby inhibiting MMP activity and M2 macrophage polarization [183]. Artemisinin derivatives (artesunate and dihydroartemisinin) inhibited CAF activation and reduced MMP, vimentin, and α-SMA expression [184].

Further studies have shown that updating the modification of materials such as nanoparticles, liposomes, and drug delivery systems provides stronger support for the use of natural products as alternative adjuvant therapies to overcome the limitations of traditional drugs. CFH peptide-decorated liposomal oxymatrine (CFH/OM-L) can inhibit CAF activity and disrupt the collagen barriers constructed by CAFs, repolarizing TAMs to the M1 phenotype via TGF-β1/IL-6 modulation and enhancing NK cell infiltration [185]. Ginsenoside Rg3-based liposomes loaded with paclitaxel (Rg3-PTX-LPs) inhibit IL-6/STAT3 signaling, repolarize M2 macrophages to the M1 phenotype, and reduce CAF and collagen deposition [186]. Similarly, Rg3-loaded liposomes (Rg3-Lp/DTX) inhibit TGF-β/Smad signaling, reverse CAF activation, convert immunologically “cold” TNBC cells into “hot” tumors enriched with CD4⁺/CD8⁺ T cells, reduce the MDSC/Treg cell ratio, and increase the M1/M2 ratio [187]. PEGylated salvianolic acid B liposomes (PEG-SAB-Lips) suppress TGF-β1 secretion in CAFs, increase T-cell recruitment and Th1 cytokine expression, and inhibit MDSC/Treg infiltration [188]. Baicalein-loaded nanoparticles (PMs-Ba) inhibit TGF-β signaling, reduce IL-4/6/10 and CCL2/5 levels, and shift TAMs from the M2 phenotype to the M1 phenotype, activating the tumor immune microenvironment [189]. Puerarin nanoemulsions downregulate fibroblast growth factor-2 (FGF-2), platelet-derived growth factor-B (PDGF-B), and TNF-α, inactivating CAFs while increasing CD8⁺/CD4⁺ T-cell infiltration and promoting M1 macrophage polarization [190].

Challenges and future perspectives

Although targeting the paracrine signaling pathway of CAFs has achieved certain effects, it still faces multiple challenges in clinical application. First, the high heterogeneity of CAFs and mixed spatial distribution may lead to functional redundancy and signal compensation [208, 209], Single-cell sequencing studies have revealed that CAFs subsets can undergo dynamic phenotypic transformation (e.g. inducing iCAF to myCAF transformation) [210, 211] under therapeutic stress, further increasing the difficulty in targeting. Secondly, existing drugs (such as TGF-β inhibitor galunisertib or CXCR4 antagonist plerixafor) are prone to cause off-target toxicity owing to sharing signaling pathways with normal tissues, such as aggravated tissue fibrosis or imbalance of immune homeostasis. In addition, CAF-driven ECM hardening forms a physical barrier that restricts drug penetration, whereas anti-fibrotic strategies can temporarily soften the matrix but may accelerate the risk of metastasis [212].

Currently, most drugs targeting key CAF regulatory factors are undergoing clinical or preclinical evaluation, while conclusive findings primarily stem from Phase II clinical trials, with only a minority advancing to Phase III [213]. Future research should pay more attention to analyzing the functional characteristics of CAFs in order to seek more precise treatment plans. On one hand, integrating technologies such as single-cell multi-omics and spatial metabolomics is crucial to develop predictive biomarkers for stratifying and selecting patient populations suitable for CAF-targeted therapies. On the other hand, multi-dimensional combination strategies, such as the sequential use of drugs targeting CAF paracrine signaling, nanoparticle delivery of subtype-specific inhibitors, combined with PD-1 blockade, may overcome the limitations of monotherapies. For instance, IL-1β inhibition of presenilin 1 promotes CD8 + T cell proliferation and infiltration in ovarian cancer models, and restoring functional CD8 + T cells within the TME enhances immunotherapy efficacy [214]. Feig et al. found that CXCL12 signaling contributes to T cell exhaustion and unresponsiveness to α-PD-L1 therapy. Blocking the CXCL12/CXCR4 signaling pathway with AMD3100 inhibited CAF-directed tumor immune evasion, increased T cell infiltration, and significantly reduced tumor volume when combined with α-PD-L1 [77].

Notably, the vivo characteristics of CAFs often change due to alterations of culture conditions and passaging [215, 216], which poses challenges for drug validation. Current strategies are reflected in 3D co-culture models incorporating tumor organoids, CAFs, and Transwell systems, which demonstrate the tumor-promoting effects of paracrine signaling. Concurrently, xenograft models are used to validate how widespread CAF enrichment promotes drug resistance [217]. Future efforts should also further optimize organoid-immune co-culture and humanized models to more directly elucidate CAF-mediated targeting effects. Additionally, dynamic monitoring technologies (e.g., ctDNA or real-time exosome analysis) hold promise for enabling real-time treatment adjustments to counteract CAF signal compensation and the evolution of resistance. Despite these significant challenges, interdisciplinary convergence (e.g., nanomedicine, bioinformatics, immunology, metabolomics) will provide novel perspectives for addressing CAF-mediated immunosuppression, thereby further reshaping the landscape of cancer therapy.

Conclusions

With advancements in diagnostic and therapeutic approaches, immunotherapy has achieved significant clinical benefits; however its efficacy is typically limited by immune escape [218]. The TME is a highly heterogeneous and dynamic ecosystem. CAFs, as enforcers of immune escape in the TME, create a supportive niche for tumor growth while altering intercellular communication networks to influence immune responses. Cytokines, chemokines, exosomes, and metabolites secreted by CAFs recruit immunosuppressive cells, impair effector immune cell function (e.g., cytokine production and cytotoxicity), and modulate antitumor immune responses, thereby facilitating immune escape [219, 220]. Notably, CAF-derived TGF-β, IL-6, and other effectors act as paracrine signals to recruit immunosuppressive cells, compromise T-cell function, remodel the ECM, and establish an immunosuppressive TME. Additionally, CAF-secreted exosomes carrying circRNAs, miRNAs, and metabolites further reprogram the TME by altering the cellular composition, intercellular communication, and metabolic activity. Despite the early setbacks in clinical trials targeting CAFs, recent advances in understanding their paracrine mechanisms have revitalized therapeutic interest.

Recent strategies have focused on blocking immunosuppressive ligands in CAF paracrine pathways, such as TGF-β, IL-6, and CXCL12. The interaction between CAFs and T lymphocytes regulates and reconnects the TME to support the malignant biological characteristics of tumors. Chakravarthy et al. discovered that the dysregulation of pan-cancer ECM is related to CAF-mediated TGF-β signaling, which is closely related to immune activity and can predict the failure of PD-1 blockade [221]. Clinical trials (NCT02734160 and NCT03436563) have demonstrated that TGF-β inhibitors (e.g., LY3200882 and galunisertib) synergize with ICBs to increase effector T-cell infiltration, DC maturation, and antitumor immunity. Some TGF-β inhibitors also directly improve CTL function, normalize the TME, and overcome myCAF-mediated barriers. The CXCL12/CXCR4 axis contributes to tumor progression and immunosuppression. CAF-derived CXCL12 recruits CXCR4⁺ endothelial progenitor cells and immunosuppressive DCs to promote angiogenesis and tumor growth. CXCR4 inhibitors (e.g., plerixafor) reduce fibrosis, increase cytotoxic T-cell infiltration, and enhance checkpoint inhibitor efficacy. CXCL12 inhibitors (e.g., tipifarnib and NOX-A12) improve immunotherapeutic outcomes in pancreatic and colorectal cancers in a dose-dependent manner [144, 145, 222]. The strategies discussed herein – targeting CAF-derived factors, disrupting CAF-immune cell crosstalk, or modulating CAF activation states – hold substantial translational potential to shift the immunosuppressive TME towards a more permissive state for immune attack.

Furthermore, the success of CAF-targeted therapies also critically depends on robust patient stratification. Wessolly M et al. found in high-grade serous ovarian cancer (HGSOC) [223] that stratification CAFs and consideration of alternative treatment options might be helpful. Multiple pathways, including MAPK, Ras, and TGF-β, have been proven to impair the treatment response. In pancreatic cancer, patients with high CAF risk sores are associated with a lower response to immunotherapy and higher tumor mutational burden (TMB) [224]. Similarity, in gastric cancer, the high-CAF-score group exhibits more active cancer-related pathways and later TNM stages, while the low-CAF-score group shows a higher abundance of follicular helper T (Tfh) cells, increased immune activation, and a greater likelihood of sensitivity to immunotherapy [225]. Thus, CAF-based stratification plays a crucial role in clarifying the status of the tumor immune microenvironment and guiding clinical treatment strategies.

Although CAF subgroups have been gradually discovered, the high heterogeneity of CAFs caused by different tumors, even different parts of the tumor, and spatiotemporal dynamic changes in the tumor remain a major problem in achieving precise targeted therapy. Natural products with antitumor and immunomodulatory properties have obvious advantages in targeting CAF paracrine pathways because of their multi-target and multi-pathway characteristics [226, 227]. Natural products modulate multiple immune pathways while maintaining systemic balance, making them valuable resources for anticancer drug development. Their polypharmacology minimizes side effects caused by single-target overactivation or suppression. Polyphenols regulate inflammatory cytokines to disrupt CAF–tumor cell crosstalk, polysaccharides influence chemokine-driven immune cell recruitment, and saponins enhance the efficacy of ICB as immunostimulants. Nanotechnology-optimized natural products (e.g., PEG-SAB-Lip, puerarin nanoparticles, ZSP4, astragaloside IV, Rg3-PTX-LPs, and CFH/OM-L) selectively remodel the TME by reducing Treg cells/MDSC infiltration, repolarizing M2 macrophages to the M1 phenotype, and protecting CD8⁺/CD4⁺ T cells.

To further overcome the limitations and resistance brought about by the single therapy, combination therapies integrating small-molecule inhibitors, natural products, chemotherapy, and immunotherapy have immense potential for enhancing treatment efficacy. Jiao et al. [228]. reported that TGF-β blockers synergize with ICB in bone metastasis models. The IL-1 receptor antagonist anakinra sensitizes pro-tumorigenic iCAFs in pancreatic cancer to therapy, which is correlated with increased CD8 + T cells [210]. CnP combined with gemcitabine or sorafenib has exhibited promising safety and efficacy in preclinical studies [173, 229]. In addition, certain achievements have also been made in the development of cell-based therapy and new drug delivery systems. Gao Y et al. Constructed dual-targeted CAR-T cells with PD-L1/CAF nanoantibodies, upregulated IFN-γ and TNF-α, inhibited the expression of IL-10, significantly enhanced the matrix penetration ability, and promoted immune activity and anti-tumor activity [230].Yuan S et al. preparing a lipid-polymer hybrid drug delivery system (PI/JGC/L-A) to upregulate IL-12 mediated by CAF, overcome the CAF barrier and activated antitumor immunity at the tumor site [231]. The development of multiple drug delivery systems based on nanoplatforms [232, 233] has further confirmed that targeting the CAF-mediated pathway can not only reshape the matrix density and permeability, but also emphasize the alteration of the tumor immune microenvironment and the improvement of anti-tumor activity.

As central regulators of the TME, CAFs undeniably possess the capacity to remodel the immune landscape. CAF-directed therapies, including combination strategies with ICIs, have generated encouraging preclinical results and are transitioning to clinical trials. However, CAF heterogeneity and dynamic paracrine signaling necessitate careful consideration of tumor type, stage, TME composition, and prior treatments in therapeutic design. Their functional plasticity also provides opportunities to modulate CAF activity for therapeutic benefit. Meanwhile, with the development of high-throughput sequencing technology, the tumor suppression microenvironment landscape mediated by CAFs has been further identified. Current genetic tools remain largely confined to subtype-specificity CAFs ablation for immune homeostasis in the tumor microenvironment. Developing genetic interventions targeting hub genes within paracrine signaling pathways may substantially enhance therapeutic efficacy, thereby unlocking novel avenues for multimodal modulation of CAFs to potentiate anti-tumor immunity. These insights offer novel perspectives for targeting CAFs to reprogram the TME and improve anticancer outcomes, warranting continued exploration.

In conclusion, CAFs, as central regulators of the TME, represent both a challenge and an opportunity. While their heterogeneity and pleiotropic signaling complicate therapeutic targeting, their plasticity enables functional modulation to restore antitumor immunity. Future strategies must consider tumor-specific CAF subtypes, crosstalk with TME composition, and prior treatments. The integration of CAF-targeted therapies with natural products and nanotechnology holds promise for overcoming stroma-driven resistance and achieving durable clinical responses

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

AKT

Protein kinase B

apCAFs

Antigen-presenting cancer-associated fibroblasts

α-SMA

Alpha-smooth muscle actin

AUF1

Adenylate-uridylate-rich binding factor 1

CAFs

Cancer-associated fibroblasts

CAR-T

Chimeric antigen receptor T cell therapies

CCL

C-C motif ligand

CCR

C-C chemokine receptor

CLCF1

CAF-derived cardiotrophin-like cytokine 1

c-Met

Cellular-mesenchymal epithelial transition factor

CnP

Conophylline

COX-2

Cyclooxygenase-2

CR

Complete response

CTLA-4

Cytotoxic T lymphocyte-associated protein 4

CTL

Cytotoxic T lymphocyte

CXCL

C-X-C chemokine ligand

CXCR

C-X-C chemokine receptor type

DCs

Dendritic cells

ECM

Extracellular matrix

EGCG

Epigallocatechin gallate

EMT

Epithelial‒mesenchymal transition

FAPα

Fibroblast activation protein-alpha

FGF-2

Fibroblast growth factor

Foxp3

Forkhead box protein P3

GM-CSF

Granulocyte-macrophage colony-stimulating factor

GSK3β

Glycogen synthase kinase-3 beta

HCC

Hepatocellular carcinoma

HGF

Hepatocyte growth factor

HLA

Human leukocyte antigen

HOXA6

Homeobox A6

iCAFs

Inflammatory cancer-associated fibroblasts

ICB

Immune checkpoint blockade

ICIs

Immune checkpoint inhibitors

IFNγ

Interferon gamma

IGF

Insulin-like growth factor

IGF1R

Insulin-like growth factor 1 receptor

IL

Interleukin

irAEs

Immune-related adverse events

MAPK

Mitogen-activated protein kinase

M-CSF

Macrophage colony-stimulating factor

MDSC

Myeloid-derived suppressor cell

MHC

Major histocompatibility

M-MDSCs

Monocytic myeloid-derived suppressor cells

MMPs

Matrix metalloproteinases

myCAFs

Myofibroblast cancer-associated fibroblasts

NF-κB

Nuclear factor kappa B

NK cells

Natural killer cells

NOD1

Nucleotide-binding oligomerization domain 1

PDGF-B

Platelet-derived growth factor subunit B

PD-1/L1

Programmed death receptor 1/ligand 1

PGE2

Prostaglandin E2

PI3K

Phosphoinositide 3-kinase

PTGS2

Prostaglandin endoperoxide synthase 2

SDF-1

Stromal cell-derived factor-1

STAT3

Signal transducer and activator of transcription 3

TAMs

Tumor-associated macrophages

TANs

Tumor-associated neutrophils

TGF-β

Transforming growth factor beta

Th17

T helper cell 17

TIMP

Tissue inhibitor of metallopeptidases

TLR4

Toll-like receptor 4

TME

Tumor microenvironment

TNBC

Triple-negative breast cancer

TNFα

Tumor necrosis factor alpha

Treg cells

Regulatory T cells

VEGF

Vascular endothelial growth factor

ZBTB12

Zinc finger and BTB domain containing 12

Author contributions

SCG, MWZ, WQB, and WJB offered direction and guidance for the manuscript. LY and WLY drafted the initial manuscript. LY illustrated the figures for the manuscript. WLY wrote the table for the manuscript. MWZ, SCG, WQB and WJB contributed to the revision of the manuscript. All the authors approved the final manuscript.

Funding

This work was supported by funds from the Project supported by the Major Program of the National Natural Science Foundation of China (Grant No. 82430123), the National Natural Science Foundation of China (Grant Number: 82174222), the Shandong Taishan Scholars Special Expert Talent Project (tstp20221166), the Science and Technology Development Fund of Macau SAR (No: 0098/2021/A2 and 0048/2023/AFJ), and the Chinese Medicine Guangdong Laboratory (HQCML-C-2024006).

Data availability

Not applicable.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Written informed consent for publication was obtained from all participants.

Competing interests

The authors have declared that no competing interest exists.