The term TME specifically refers to the microenvironment of solid tumors, which comprises not only malignant cells but also a large population of immune and stromal cells, along with non-cellular components (Fig. 1). The main cellular components include fibroblasts, endothelial cells, adipocytes, innate immune cells and adaptive immune cells. These cells collectively modulate the local TME through dynamic interactions, exhibiting dual roles that either promote or antagonize tumor progression.

Cancer-associated fibroblasts (CAFs), the most abundant stromal cell population within the TME, play pivotal roles in cancer progression. Their identified functions include ECM remodeling to facilitate tumor invasion (7), promotion of cancer cell stemness (8), enhancement of chemoresistance to targeted therapies (9) and reprogramming of the immune environment within tumors (10,11). Studies have revealed that CAFs constitute a heterogeneous population originating from diverse precursor cells through either local differentiation or recruitment to tumor sites (2,12,13). While activation of local tissue-resident fibroblasts and stellate cells is recognized as the primary source of CAFs (12), alternative origins include adipocytes, bone marrow-derived mesenchymal stem cells, epithelial cells undergoing epithelial-mesenchymal transition (EMT) and endothelial cells undergoing endothelial-mesenchymal transition (13,14). In malignant tumors, neoplastic cells drive the transformation of normal fibroblasts into CAFs by activating inflammatory pathways through secretion of cytokines, growth factors and functional DNAs or non-coding RNAs (15,16). Furthermore, other non-malignant cells in the TME can induce CAF conversion, as evidenced by findings showing that M2-polarized macrophages promote the transformation of intra-tumoral normal fibroblasts into CAFs via paracrine signaling pathways (17). The heterogeneity of CAFs markedly influences subsequent tumor progression. Studies have identified specific CAF subtypes associated with characteristics such as tumorigenesis, chemotherapy resistance and immunosuppression (11,18,19). However, the extent of CAF heterogeneity and the potential tumor-modulating effects exerted by distinct CAF subtypes remain under investigation.

Tumor-associated macrophages (TAMs) constitute nearly half of the cellular components within solid tumors, playing pivotal roles in tumor progression (20). These cells, which are derived from peripheral blood mononuclear cells, differentiate into macrophages when stimulated by various factors secreted by tumor and stromal cells, such as chemokines, cytokines and growth factors (21). Within the TME, TAMs are markedly involved in tumor proliferation, invasion, metastasis and angiogenesis, with TAM-derived cytokines acting as key regulators in these processes (22,23). TAMs are traditionally classified into M1 and M2 subtypes, with M1-like macrophages involved in pathogen clearance, inflammatory response and anti-tumor immune functions (24), while M2-polarized macrophages exhibit anti-inflammatory properties and promote tumor cell proliferation, metastasis and immune evasion (25,26). Recent evidence suggests that TAM phenotypic diversity in vivo exceeds this binary classification due to tumor heterogeneity (27). As a highly diverse immune cell population with various phenotypes and functions, TAMs undergo differentiation influenced by multiple factors, leading to heterogeneous pro-tumorigenic capabilities. For instance, Wang et al (28) demonstrated that hepatocellular carcinoma-derived C-C motif chemokine ligand 2 (CCL2) and CCL5 attract TAMs and induce their polarization towards pro-tumorigenic M2-like macrophages. Similarly, Su et al (29) found that breast cancer cells activate macrophages to a TAM-like phenotype by secreting granulocyte-macrophage colony-stimulating factor (GM-CSF). In prostate cancer, CAFs recruit and activate monocytes through C-X-C motif chemokine ligand 12 (CXCL12) and CXCL14 chemokines to generate M2-polarized macrophages (30,31). This heterogeneity leads to complex regulatory interactions between TAMs and tumor tissues.

As well as macrophages, other immune cells also play critical roles in tumor progression or anti-tumor surveillance. Cancer patients exhibit diminished immune surveillance capacity and immune dysregulation, including imbalances in CD4⁺ T cells, CD8⁺ T cells and associated cytokines (32,33). Thymus-derived naive CD4⁺ T cells differentiate into distinct subsets upon antigenic stimulation in the periphery: T-helper 1 cells (Th1) are characterized by secretion of interferon-γ (IFN-γ), while Th2 is defined by interleukin-4 (IL-4) production (34). Another subset comprises regulatory T cells (Tregs), which play a pivotal role in attenuating anti-tumor immune responses (35). Tregs express the transcription factor Foxp3 and surface markers CD25/CD127, suppressing the function of effector T cells and antigen-presenting cells through direct cell-cell contact and secretion of inhibitory cytokines like transforming growth factor-β (TGF-β) and IL-10 (35,36). Myeloid-derived suppressor cells (MDSCs), originating from aberrant myeloid differentiation of hematopoietic stem cells, exhibit immunosuppressive properties (37,38). Accumulating evidence highlights MDSCs as central components of the malignant TME, critically driving tumor progression and chemoresistance through secretion of inflammatory factors and chemokines such as IL-6 and CXCL family members (39,40). Neutrophils, the most abundant leukocytes in human blood, represent a significant proportion of the TME in multiple malignancies (41,42). These cells demonstrate functional plasticity and are recruited to tumor tissues via specific cytokines and chemokines during solid tumor progression, where they adopt context-dependent roles (42,43). Neutrophils can be classified into anti-tumor N1 and pro-tumor N2 subtypes (44). A subset of neutrophils may exert anti-cancer effects through reactive oxygen species (ROS) and neutrophil elastase release (45,46); however, the majority are polarized into pro-tumorigenic phenotypes by factors within the TME (47). N2 neutrophils promote tumor cell proliferation, angiogenesis and metastasis via secretion of pro-tumorigenic molecules, while also fostering an immunosuppressive TME through interactions with macrophages, NK cells and T cells (43,48,49).

Cytokines are a class of secreted proteins that mediate intercellular communication. Broadly defined, cytokines encompass a diverse group of secreted molecules including chemokines, growth factors, angiogenic factors, soluble receptors and extracellular proteases (50). These proteins not only exhibit cancer-suppressive functions but also participate in regulating physiological processes closely associated with tumor initiation, progression and metastasis, such as inflammation, apoptosis, migration and angiogenesis (2,4).

Multiple factors directly regulate tumor progression by modulating tumor cell proliferation and intra-tumoral vascular formation. For instance, vascular endothelial growth factors (VEGF) and its downstream signaling pathways are overexpressed in most malignancies, demonstrating dual functions in promoting angiogenesis and enhancing vascular permeability through specific induction of endothelial cell division, proliferation and migration (51,52). Similarly, insulin-like growth factor-1 (IGF-1) binds to its receptor IGF-1R to activate PI3K/AKT and MEK/ERK signaling pathways, thereby regulating tumor cell proliferation, invasion and metastasis (53). Notably, IGF-1R has been found widely expressed across various cell types in the TME, including epithelial cancer cells, CAFs and myeloid cells (54). TGF-β, a pleiotropic growth factor, exhibits dual functionality by suppressing tumor cell growth while promoting cancer progression through EMT induction under specific conditions, conferring stem-like properties to cancer cells (55). Additionally, aberrant expression of factors such as fibroblast growth factor (FGF), epidermal growth factor (EGF) and hepatocyte growth factor (HGF) markedly promotes tumor and stromal cell proliferation along with angiogenesis (56,57).

Furthermore, cellular components within the TME secrete inflammatory cytokines and chemokines that exert complex effects through autocrine and paracrine mechanisms. IL-6 demonstrates pleiotropic pro-inflammatory functions, promoting B and T cell differentiation while directly stimulating tumor cell proliferation and chemoresistance (58,59). The pro-inflammatory cytokine IL-8 enhances tumor cell proliferation, survival and exhibits potent pro-angiogenic activity, while also participating in the recruitment of lymphocytes, monocytes and neutrophils (59). Conversely, the anti-inflammatory cytokine IL-10 protects cancer cells from immune attack through potent immunosuppressive mechanisms (40,60).

Extensive research has focused on soluble factor-mediated molecular interactions between tumor cells and their microenvironment, collectively revealing the tip of the iceberg in intra-tumoral regulatory networks (3,61,62). Although the current understanding of the involvement of cytokine networks in TME communication remains incomplete, cytokines represent valuable therapeutic targets and biomarkers. This necessitates a deeper exploration of their signal transduction and regulatory functions in the TME.

The TME is a dynamic ecosystem where stromal and immune cells interact with cancer cells through a complex network of secreted proteins, such as inflammatory mediators, chemokines and growth factors, to drive the development of tumorigenic phenotypes. Studies have made breakthroughs in exploring the molecular mechanisms underlying this process.

In a series of studies on cervical squamous cell carcinoma, Wei et al (71) investigated the tumor-promoting mechanisms of CAFs through two distinct signaling axes. First, CAF-derived plasminogen activator inhibitor-1 (PAI-1) was identified as promoting lymphatic metastasis by triggering EndoMT in lymphatic endothelial cells (LECs). This process was mediated through LDL receptor-related protein 1 (LRP1)-dependent activation of the AKT/ERK signaling pathway, ultimately facilitating tumor dissemination. This identifies the PAI-1/LRP1 axis as a potential therapeutic target for inhibiting metastasis. Parallel to this discovery, the team further identified periostin⁺ CAFs as another critical mediator of lymphatic metastasis. Periostin secreted by these CAFs activated the integrin-FAK/Src signaling cascade in LECs, leading to phosphorylation and subsequent degradation of VE-cadherin. This disrupted LEC barrier integrity, facilitating tumor cell intravasation and dissemination (72). In addition to fibroblasts, Sun et al (73) demonstrated how omental adipocytes facilitate tumor peritoneal metastasis in ovarian cancer in a murine model. Adipocyte-secreted CCL2 bound to CCR2 on cancer cells, activating the PI3K/AKT/mTOR pathway, which upregulated hypoxia inducible factor-1α (HIF-1α) and VEGF-A secretion. This stimulated intra-tumor angiogenesis and tumor metastasis to the omentum. Blocking the CCL2/CCR2 axis represents a potential therapeutic strategy to prevent peritoneal metastasis in ovarian cancer. A complex bi-directional crosstalk between geminin-overexpressing (GemOE) triple-negative breast cancer (TNBC) cells and stromal cells was investigated as a driver of metastasis. GemOE-cells secreted acetylated HMGB1, which activated receptor for advanced glycation end-products (RAGE) and CXCR4 expression on mesenchymal stem cells (MSCs) and recruited them into tumors. These MSCs differentiated into CAFs and secreted S100A4, which reciprocally stimulated GemOE-cells to release CCL2, attracting and polarizing macrophages into M2-TAMs. Gas6 secreted by the TAMs combined with AXL on GemOE cells and synergized with RAGE signaling to amplify cancer stemness, EMT and CD151/α3β1-integrin-mediated interactions, enhancing invasiveness (74). While CAFs typically promote tumor growth, Han et al (75) revealed that hypoxic fibroblasts demonstrated opposing effects. Conditioned medium (CM) from hypoxic dermal fibroblasts (H-CM) suppressed cervical cancer cell (HeLa) viability by enhancing apoptosis via caspase-3/7 activation, mitochondrial dysfunction and G₀/G₁ arrest. Proteomics analysis identified lymphotoxin-beta receptor (LTBR), a member of the TNF receptor family, as a key factor upregulated in H-CM, which suppressed HeLa cell proliferation. These findings suggest microenvironment-dependent fibroblast plasticity in regulating malignant progression (75).

TAMs are pivotal drivers of tumor progression, primarily through their secretion of cytokines and chemokines that promote immunosuppressive microenvironments and metastatic cascades. Huang et al (76) demonstrated that TAMs expressing high levels of triggering receptor expressed on myeloid cells 1 (TREM1) are key contributors to EMT and metastasis in hepatocellular carcinoma (HCC). The authors identified CCL7 as a key downstream effector secreted by TREM1⁺ TAMs that drives these effects. These findings suggest TREM1 or CCL7 as potential targets to disrupt EMT and metastasis in HCC. Furthermore, TREM1 expression positively associated with elevated levels of programmed death ligand 1 (PD-L1) and cytotoxic T lymphocyte associated protein 4 (CTLA-4), linking CCL7-driven EMT to immunosuppressive TME remodeling (76). In breast cancer models, Zheng et al (77) demonstrated that chronic unpredictable mild psychological stress activates glucocorticoid receptor (GR) signaling in TAMs, leading to increased secretion of CXCL1. This chemokine facilitated MDSC recruitment via CXCR2, enhancing their immunosuppressive capacity to inhibit CD8⁺ T cell cytotoxicity and promote pre-metastatic niche formation. Therefore, targeting GR signaling in TAMs or the CXCL1-CXCR2 axis represents a potential therapeutic strategy to counteract stress-induced metastasis and immunosuppression in breast cancer. In osteosarcoma, a group of Iba1⁺/CD163⁺ TAMs were found to enhance tumor progression and lung metastasis. Using antibody arrays, the authors confirmed the CM of tumor cells and TAMs co-cultures enriched with IL-8, which markedly enhanced osteosarcoma cell proliferation, migration and invasion through the FAK pathway (78). Macrophage polarization and cytokine release are influenced by bi-directional crosstalk with tumor-derived factors, which further enhance tumor progression. Kim et al (79) highlighted that colorectal cancer (CRC)-derived CD133⁺ microvesicles (MVs) differentiate macrophages into an M2-like phenotype within the TME. CD133⁺ MVs triggered IL-6 secretion from TAMs, which subsequently activated the STAT3 pathway in CRC cells, thereby enhancing their EMT and invasion. IL-6/STAT3 activation further established a feedback loop, reinforcing CD133 expression in cancer cells and establishing an immunosuppressive TME conducive to CRC progression. Targeting TAMs is a promising strategy to enhance anti-tumor effects. As demonstrated by Licarete et al (80), prednisolone phosphate (PLP) enhances doxorubicin (DOX) efficacy against B16.F10 melanoma cells by targeting TAMs. PLP suppresses TAM-mediated angiogenesis by downregulating pro-angiogenic factors (FGF-2, VEGF, G-CSF, IL-1β and TNF-α), thereby disrupting the TME supporting melanoma growth. Combined PLP with DOX synergistically inhibits melanoma proliferation and induces apoptosis, linked to reduced oxidative stress and potentiated anti-angiogenic effects. However, the M2 immunosuppressive phenotype and IL-10 and arginase-1 (Arg-1) secretion remains unaffected, highlighting TAM angiogenic pathways as the primary mechanism.

Other immune cells within the TME also play a role in regulating tumor progression through distinct secretory programs. In TNBC, tumor-infiltrating neutrophils are engaged in a self-reinforcing loop with cancer cells via tissue inhibitor of metalloproteinase-1 (TIMP-1) and CD90 interactions (81). Antibody array screening demonstrated that these neutrophils secrete elevated TIMP-1, which drove twist-mediated EMT in tumor cells. Reciprocally, EMT-transformed CD90⁺ tumor cells enhanced TIMP-1 production in neutrophils through CD90-Mac-1 interactions, amplifying metastatic potential of the tumor (81). Lee et al (82) demonstrated that eosinophils drove metastasis through cytokine-mediated immunosuppression and vascular remodeling in head and neck squamous cell carcinoma (HNSCC). Tumor-associated tissue eosinophilia (TATE) was found to be associated with aggressive tumor features such as angiogenesis and lymph node metastasis in HNSCC, with TATE-rich tumors exhibiting increased CD4⁺Foxp3⁺ Tregs, exhausted CD8⁺PD1⁺ T cells and reduced cytotoxic lymphocytes, collectively fostering an immunosuppressive niche that accelerates tumor progression. CCL2 was identified as a key eosinophil-derived factor promoting tumor cell migration and EMT (82). While typically associated with immunosuppression, Benzing et al (83) reported that undifferentiated monocyte-like suppresses pancreatic ductal adenocarcinoma (PDAC) invasiveness. Co-culturing PDAC cells with monocyte-like THP-1 markedly suppressed invadopodia formation and matrix degradation. Proteomic analysis of THP-1 CM identified high levels of TIMP-2, which selectively inhibits MT1-MMP (MMP-14), an enzyme critical for invadopodia function.

The balance between tumor-promoting and tumor-suppressing immune responses is critical for tumor survival and involves multiple signaling pathways modulated by secreted factors derived from tumor cells, immune cells and non-neoplastic stromal cells within the TME. However, due to the complexity of this dynamic process, the underlying mechanisms remain incompletely understood. Li et al (84) identified the critical role of uridine phosphorylase 1 (UPP1), highly expressed in lung adenocarcinoma (LUAD) cells, in shaping an immunosuppressive TME. UPP1 upregulation in tumor cells elevated TGF-β1 secretion, which promoted Treg differentiation, CD8⁺ T cell exhaustion and macrophage polarization toward an M2 phenotype via paracrine signaling. Additionally, UPP1 activated the PI3K/AKT/mTOR pathway to enhance PD-L1 expression on LUAD cells, further impairing T cell cytotoxic function. Similarly, TGF-β3 overexpression in HCC cells was shown to activate the SMAD2/3-Sp1 axis, thereby upregulating decoy receptor 3 (DcR3), which inhibits pro-inflammatory signaling by binding to the LIGHT ligand on activated CD4⁺ T cells. DcR3 further promoted the differentiation of CD4⁺ T cells into Th2 and Treg cells while suppressing Th1 polarization, thereby impairing anti-tumor immunity in HCC (85). Xie et al (86) demonstrated that hypoxic tumor cells in mouse breast and colon cancer models produce angiotensin II (AngII) via a hypoxia-lactate-chymase-dependent axis, fostering an immunosuppressive TME characterized by increased infiltration of Tregs, TAMs, CAFs and MDSCs, alongside reduced recruitment of CD8⁺ effector T cells. Blocking AngII signaling reversed this immunosuppressive TME and markedly upregulated immune-activating cytokines (such as IL-7, IL-20 and CXCL11) while downregulating immunosuppressive cytokines (such as IL-10, GM-CSF and CCL28) in 4T1 breast cancer cells. Since the immunosuppressive microenvironment is essential for tumor persistence, its modulation represents a promising therapeutic strategy. Jiang et al (87) revealed c-Myc targeting as a strategy to overcome immune suppression in osteosarcoma. c-Myc inhibition via JQ-1 upregulated T cell-recruiting chemokines (CCL5, CXCL9 and CXCL10), thereby enhancing cytotoxic T cell infiltration. Simultaneously, c-Myc inhibition upregulated CD40 on dendritic cells (DCs) and activated CD40/CD40L costimulatory signaling to promote DC-T cell interaction and cytotoxic T cell activation. These immune-enhancing effects, rather than direct anti-proliferative actions, drove tumor regression and prolonged overall survival in murine models. Wang et al (88) found that STK3 expression is epigenetically suppressed via promoter hypermethylation in ovarian cancer, associated with poor patient prognosis. Overexpression of STK3 in ovarian cancer cells activated the NF-κB signaling pathway, leading to upregulated CXCL16 and CX3CL1. These chemokines facilitated the recruitment of CD8⁺ T cells and markedly inhibited tumor cell proliferation, migration and invasion while promoting apoptosis. These findings highlighted the role of STK3 in counteracting immune suppression.

High-stromal tumors, characterized by elevated fibroblast and MSC content, exhibit a distinct immune landscape marked by reduced infiltration of effector cells. In high-grade serous ovarian carcinoma, CXCL12 overexpression was found in epithelial-like tumor cells and cancer-associated MSCs and was shown to drive immune exclusion by binding to CXCR4 receptors on cytotoxic CD8⁺ T cells and NK cells, sequestering these effector cells in stromal compartments (89). This CXCL12-CXCR4 axis impaired anti-tumor immunity by suppressing granzyme B and IFN-γ production in CD8⁺ T cells while upregulating immunosuppressive cytokines such as CXCL1, CXCL5 and CXCL13 in myeloid cells (89). Similarly, Sheng et al (90) identified TAK1⁺ CAFs as playing a critical role in shaping the immunosuppressive microenvironment of PDAC. TAK1⁺ CAFs predominantly exhibit an inflammatory phenotype and serve as major sources of CXCL12 and IL-6, which recruit MDSCs and polarize macrophages toward M2 phenotypes, while also excluding CD8⁺ T cells from tumor nests. Notably, TAK1 inhibition shifted CAFs toward a tumor-suppressive myo-fibroblastic phenotype, which suppressed EMT and tumor cell invasion via downregulation of MAPK and NF-κB pathways. Furthermore, angiotensin-stimulated human and mouse fibroblasts were found to secrete CCL5 as a key immunosuppressive mediator in melanoma, which simultaneously reduced intra-tumoral CD8⁺ T cell infiltration and promoted Treg recruitment via CCR5 (91).

A growing number of studies have reported that neutrophils can be recruited to tumor tissues by specific cytokines and chemokines, undergo functional polarization and exert dual regulatory effects in TME immunomodulation. SenGupta et al (92) demonstrated that TNBC recruited neutrophils via tumor-derived factors. Antibody array screening identified a distinct secreted profile in aggressive TNBC cells compared with ER⁺ cells, characterized by elevated production of TGF-β and CXCR2 ligands (e.g., CXCL1, CXCL2 and CXCL3). These factors synergistically promoted neutrophil chemotaxis and polarization via TGF-β/SMAD3 signaling and the CXCR2-Gi pathway. Complementing this, Ogawa et al (93) revealed that loss of SMAD4, a tumor suppressor, in CRC cells triggered massive production of CXCL1 and CXCL8 via IκB kinase 2 and GSK-3β pathways, leading to accumulation of CXCR2⁺ neutrophils in tumors. These neutrophils further amplified CXCL1 and CXCL8 production in the TME, while polarizing neutrophils toward pro-angiogenic and immunosuppressive phenotypes marked by MMP-2 and 9, Arg-1 and indoleamine 2,3-dioxygenase expression. Germann et al (94) further identified that tumor-infiltrating neutrophils play a pivotal role in forming an immunosuppressive microenvironment in CRC. These neutrophils release MMP-9 to proteolytically convert latent TGF-β into its bio-active form. Activated TGF-β subsequently exerts dual immunosuppressive effects by suppressing anti-tumor effector T cells and promoting Treg expansion. Neutrophil depletion or MMP inhibition in mice reduces TGF-β signaling, enhances CD8⁺ T cell infiltration and diminishes the tumor burden. Conversely, Chan et al (95) reported that neutrophils acquire an anti-tumor phenotype in pancreatic adenocarcinoma (PAAD) with melatonin supplementation. Melatonin stimulates CXCL2 secretion from tumor cells, promoting neutrophil infiltration and polarization toward an N1-like phenotype in PAAD, characterized by CD11b⁺Ly6G⁺. These neutrophils exhibit elevated ROS and form neutrophil extracellular traps, inducing tumor cell death through cell-contact-dependent NETosis.

Therapeutic resistance constitutes a fundamental issue limiting the efficacy of chemotherapy in cancer patients. Besides neoplastic cells, the TME mediates therapy resistance through direct intercellular contact within TME components or dynamic alterations in local secreted factors. CAFs exhibit functional heterogeneity in shaping chemoresistance through distinct secreted profiles that sustain cancer stemness and activate survival pathways. Su et al (96) identified a pivotal CD10⁺GPR77⁺ CAF subset in breast cancers that establishes a chemoresistance niche by driving persistent NF-κB activation through GPR77-mediated complement signaling. This signaling axis promotes IL-6 and IL-8 secretion in CAFs, which enhances cancer stemness and upregulates ABCG2-mediated drug efflux in cancer cells. Moreover, the recruitment and activation of this CAF subset were found to be dynamically regulated by crosstalk with macrophages (97). M2-polarized TAMs secrete CCL18, which binds to the PITPNM3 receptor on normal breast fibroblasts, inducing their differentiation into CD10⁺GPR77⁺ CAFs. In clinical cohorts, a strong association between CCL18⁺ TAM infiltration and CD10⁺GPR77⁺ CAF density across breast cancer subtypes was confirmed, highlighting the role of immune-stromal cell crosstalk in amplifying cancer chemoresistance (97). Further demonstrating the role of CAF secretory heterogeneity in chemoresistance, Hu et al (11) established a living biobank of patient-derived CAFs in non-small cell lung cancer and used high-throughput antibody arrays to functionally classify CAFs into three subtypes based on their secreted profiles and resistance mechanisms. Subtype I CAFs secreted high levels of HGF and FGF-7, activating MET/FGFR bypass signaling to confer robust resistance to EGFR/ALK inhibitors. Subtype II CAFs predominantly secreted FGF-7, mediating moderate FGFR-dependent resistance, while subtype III CAFs showed minimal protective capacity. This classification enabled subtype-specific therapeutic strategies: Dual MET/FGFR inhibition overcame subtype I-mediated resistance, whereas subtype II required only FGFR blockade. In melanoma, Papaccio et al (98) revealed a pivotal role of CAFs in suppressing chemotherapy-induced cytotoxicity via secretion of IL-6 and IL-8. CAF-derived supernatants enriched in pro-tumorigenic growth factors (HGF, FGF-2 and VEGF) accelerated melanoma cell migration and induced sustained activation of FAK signaling in treated tumor cells. In ovarian cancer, cisplatin-treated CAFs secrete elevated CCL5, which activates STAT3 and PI3K/AKT pathways in tumor cells, attenuating cisplatin cytotoxicity both in vitro and in vivo by enhancing Bcl-2 expression and suppressing pro-apoptotic markers (99). Additionally, Che et al (100) identified cisplatin-induced PAI-1 secretion by esophageal CAFs as a key mediator of AKT and ERK1/2 activation in esophageal squamous cell carcinoma (ESCC). PAI-1 suppresses caspase-3 activity and ROS accumulation, attenuating apoptosis in a paracrine manner, with clinical data linking high stromal PAI-1 expression to poor progression-free survival in ESCC patients. Chemoresistance development also involves remodeling of ECM. Chrisochoidou et al (101) showed that mesothelioma-activated fibroblasts deposit collagen- and tenascin-enriched ECM, which sequesters TGF-β and activates PI3K/mTOR/SRC pathways to promote resistance to cisplatin and pemetrexed. This fibrotic niche recruits naïve fibroblasts through chemo-attractants such as CHI3L1 and angiopoietin, establishing a self-sustaining resistance loop.

Tumor-associated immune cells and neoplastic cells cooperatively establish chemoresistance through secreted protein networks that reprogram the TME. In clear cell renal cell carcinoma, Wang et al (102) identified a self-reinforcing loop where SOX17 loss activated YAP/TEAD1 signaling, driving CCL5 secretion to recruit and polarize M2-like TAMs. These TAMs, in turn, suppressed SOX17 via CCR5/STAT3 signaling, amplifying CCL5 production and fostering resistance to tyrosine kinase inhibitors. Similarly, in KRAS-mutant CRC, Liu et al (103) demonstrated that mutant KRAS stabilizes HIF-1α via ROS-mediated prolyl hydroxylase inhibition, leading to overexpression of colony GM-CSF and lactate. These factors synergistically polarized macrophages toward an immunosuppressive CD206^high/HLA-DR^low phenotype characterized by anti-inflammatory cytokine secretion, which shielded tumor cells from cetuximab-induced apoptosis. The metabolic reprogramming of CRC cells further reinforced TAM-mediated resistance, as lactate enhanced GM-CSF-driven immunosuppression. In HCC, macroH2A1 loss was demonstrated to enhance cancer cell stemness marked by reduced IL-6 and IL-8 secretion. This cytokine-depleted secretome reprogramed neighboring tumor cells to adopt chemoresistance and expanded immunosuppressive CD4⁺CD25⁺FoxP3⁺ Tregs, linking epigenetic dysregulation to paracrine-mediated immune evasion (104).

Resistance to radiotherapy also involves stromal secreted factors that alter tumor cell behavior. Guo et al (105) demonstrated that CAFs in breast cancer secrete IL-6 to promote radioresistance. CAF-derived IL-6 enhanced post-irradiation survival and proliferation of cancer cells in vitro, while murine co-injection models revealed that IL-6/STAT3 signaling diminished radiation efficacy. Chu et al (106) found that CAFs enhanced cervical cancer growth and resistance to radiation. CM from CAFs alone or from CAF-HeLa co-cultures suppressed DNA damage response genes (GADD45α, BTG2) and promoted p38 phosphorylation in HeLa cells following irradiation. Using antibody arrays, the authors identified distinct cytokine profiles in CAFs and HeLa cells: CAFs secreted high levels of IGF-2, EGF and FGF-4, whereas HeLa cells produced platelet-derived growth factor (PDGF)-AA and PDGF-BB, suggesting that this bidirectional crosstalk establishes a growth factor-enriched, radioprotective microenvironment through integrated growth factor signaling. Beyond fibroblasts, Cao et al (107) showed that sub-lethally irradiated non-parenchymal liver cells in HCC elevated MMP-8 production, which suppressed AMPK phosphorylation and activated mTOR signaling in HCC cells, driving TME remodeling and metastatic progression. Arshad et al (108) reported that simultaneous irradiation of stromal and tumor cells further modulates secretory dynamics. Irradiated wild-type fibroblasts secreted TGF-β1 to induce EMT in lung carcinoma cells, enhancing migration via Vimentin and Snail upregulation, while RhoB-deficient fibroblasts promoted pro-metastatic MMP secretion. Notably, co-irradiation of tumor and stromal cells suppressed TGF-β1 and MMP release, in contrast to stromal-only irradiation. RhoB-deficient fibroblasts additionally upregulated IL-6 and FGF-2 in tumor cells, suggesting genotype-specific shifts in paracrine signaling that may influence immune modulation.