Tumor growth depends on the abnormal and uncontrollable proliferation of cancer cells with simultaneous changes to the microenvironment. Among the stromal cells in the microenvironment surrounding the tumor, increasing evidence has reported that CAFs are targets and inducers of tumorigenic activation signals (31,32).

CAFs produce autocrine and/or paracrine cytokines that promote the biological characteristics of tumors. In addition to classical growth factors, including EGF and hepatocyte growth factor (HGF), novel CAF-secreted proteins [secreted frizzled related protein 1, and IGF like family member (IGF) 1 and 2], and membrane molecules (integrin α11 and syndecan-1) have also been identified to possess cancer cell-supporting roles (33). These factors directly or indirectly stimulate tumor growth and survival, or enhance their migratory and invasive properties.

Previous studies have demonstrated that chemokines secreted by CAFs into the microenvironment allow for the recruitment of bone marrow-derived cells (BMCs) and immune cells (34). CXCL12 (35), CXCL14 (36) and CCL5 (37) have been identified as pro-metastatic factors. In addition, MSC-derived CAFs are recruited to the stroma of the dysplastic stomach, and express interleukin (IL)-6, Wnt family member (Wnt) 5α and bone morphogenetic protein 4, all of which promote tumor growth through DNA hypomethylation (38). Furthermore, MSC-derived CAFs are recruited to the tumor through TGF-β and CXCL12 signaling (38). In oral squamous cell carcinoma (OCC), CCL2 expression in CAFs is upregulated, promoting the production of endogenous reactive oxygen species (ROS) in OC cells (OCCs) (37). Consequently, ROS induces the expression of cell cycle regulatory proteins in OCCs, and promotes OCC proliferation, migration and invasion (39). Together, these chemokines and cytokines create a suitable microenvironment allowing for the proliferation and metastasis of cancer cells.

Vascular endothelial growth factor (VEGF) was originally identified as a multifunctional cytokine in angiogenesis and lymphangiogenesis (40). The interaction between tumor and stromal cells can result in increased VEGF expression, with CAFs being the primary source of VEGF (41). Furthermore, CAF-derived PDGF has been demonstrated to be an essential factor in activating VEGF production. PDGF/PDGF receptor (R) signaling is an important regulatory pathway primarily involved in angiogenesis (41). PDGFs indirectly promote angiogenesis by recruiting stromal fibroblasts that secrete VEGF (42). Furthermore, PDGFs are able to recruit and induce BMCs to form endothelial or smooth muscle cells. Subsequently, PDGFs promote the proliferation and migration of endothelial, and smooth muscle cells (42). PDGF subunit B, which is produced by endothelial cells can induce the migration of pericytes to the vessel wall and maintain endothelial stability, thus leading to tumor angiogenesis (43).

Nagasaki et al (44) reported that cancer cells stimulate the secretion of IL-6 from fibroblasts, subsequently inducing tumor angiogenesis. IL-6R neutralization antibody inhibited IL-6 signaling and tumor angiogenesis by inhibiting the interaction between the cancer, and stroma. This finding suggests that IL-6 is a novel target for anti-angiogenesis therapy (44).

Increasing evidence suggests a metastatic support role of CAFs in tumors (45,46), whereas data regarding the presence and role of CAFs in lymph node and distant metastasis is deficient. Stromal reactions in metastatic lymph nodes, possibly comprising metastasis-associated fibroblasts, have been described as reactive and fibrotic tissue with enhanced deposition of vitronectin and fibronectin, desmoplasia, nodal fibrosis and hyaline stroma (47). Immunohistochemical characterization of CAFs was reported in one of these studies, which assessed metastatic lymph node tissue from a patient with uterine cervix adenocarcinoma who received preoperative chemotherapy (47). Certain studies have suggested that the mesenchymal-like phenotype of CAFs is involved in enhancing the metastasis of cancer cells, whereas NFs with the epithelial-like phenotype inhibit the migration of breast cancer cells (48). Similarly, normal prostate epithelial cells induce intraepithelial neoplasia in vivo when co-injected with CAFs, but not when co-injected with NFs (49).

YAP is a transcription factor that may be a signature feature of CAFs. YAP has important roles in matrix stiffening, cancer cell invasion and angiogenesis, which are induced by CAFs (50). YAP regulates the expression of specific cytoskeletal proteins, including anillin actin binding protein, diaphanous related formin 3 and myosin regulatory light polypeptide 9 (50). Additionally, CAFs secrete proinflammatory cytokines that stimulate the nuclear factor-κB (NF-κB) signaling pathway, subsequently promoting tumorigenesis (51).

Notably, CAFs in the stroma of triple-negative breast cancer samples have been demonstrated to select for bone metastatic cells (52). CAFs produce CXCL12 and IGF1, which are prognostic markers for bone relapse and activators of the phosphatidylinositol 3-kinase (PI3K)/AKT serine/threonine kinase (AKT) signaling pathway (52). Cancer cells are primed for metastasis in the CXCL12-rich microenvironment of the bone marrow, thus suggesting an important role of CAFs in tumor metastasis. Another study demonstrated that a reduction in miR-148a expression in CAFs results in increased Wnt activity through the upregulation of its target gene WNT10B. Consequently, increased Wnt activity results in increased migration of endometrial cancer cells (53).

A study reported that the downregulation of miR-26b in CAFs stimulates the migration of fibroblasts, which is a dominant characteristic of the CAF phenotype. Furthermore, CAFs with reduced expression of miR-26b promote the migration and invasion of human breast cancer cells (54). Additionally, the PTEN/miR-320/ETS2 axis secretes proteins, such as Emilin2, that distinguish between normal and malignant stroma, and is associated with a higher rate of relapse in patients with breast cancer (55). This demonstrates that miR-320 is an essential regulator of the signaling pathway in fibroblasts involved in the regulation of the tumor microenvironment. Similar to in breast cancer, in prostate cancer, the downregulation of miR-15 and −16 in CAFs is mediated through activation of the AKT, and extracellular signal-regulated kinase signaling pathways, promoting prostate cancer migration, and angiogenesis (56).

Compared with cancer cells, CAFs are relatively genetically stable with a reduced probability of developing drug-resistance, thus representing as a potential therapeutic target with lower chances for the development of chemoresistance (57,58). However, an increasing amount of data has suggested that fibroblasts have a protective role that allows cancer cells to evade therapy, as described below.

The interstitial fluid pressure (IFP) in the center of solid tumors is increased compared with that in the surrounding tumor tissue (59). Higher IFP reduces the efficiency of drug penetration into the tumor tissue, thus reducing the concentration of the drug reaching the tumor cells and increasing tumor cell viability (58). Strategies on improving chemotherapy have focused on reducing tumor IFP in order to increase the efficiency of drug transport and penetration into tumors (60).

PDGF and other associated tyrosine kinase receptors are expressed in various types of cancer. STI571, a receptor tyrosine kinase inhibitor (TKI), reduces tumor IFP and increases Taxol uptake in subcutaneously injected undifferentiated anaplastic thyroid carcinoma KAT-4 cell line-induced transplantable tumors in severe combined immune deficient mice (61,62).

HGF has been identified as an essential factor in of CAF-mediated resistance to B-Raf proto-oncogene serine/threonine kinase (BRAF) inhibitor therapy in melanoma with BRAF^V600E mutation, as well as lapatinib resistance in HER2⁺ breast cancer (63,64).

TKIs exhibit strong inhibitory effects against NSCLC with epidermal growth factor receptor (EGFR)-activating mutations (65). However, the possibility of intrinsic or developing acquired resistance is an important consideration in the management of patients with cancer. The overexpression of HGF in CAF, a ligand of HGF receptor (MET), has been reported to contribute to resistance to EGFR-TKIs (66).

EGFR and HGF are coexpressed in colorectal cancer (CRC) cell lines, and the activation of both receptors synergistically induces the proliferation of cancer cells (67). Cetuximab suppresses cell growth through dephosphorylation of EGFR, mitogen-activated protein kinase (MAPK), and/or the AKT signaling pathway (68). It was demonstrated that CAF-derived HGF phosphorylates MET, but not EGFR or receptor tyrosine-protein kinase erbB-3 in cetuximab-treated cells. Subsequently, this was revealed to restore cell proliferation and rescue cells from G₁ phase arrest, and apoptosis through restimulation of the MAPK and AKT signaling pathways (68). Notably, this effect is inhibited by suppressing MET activation with PHA-665752, a highly specific MET kinase inhibitor, or by knocking down MET expression using RNA interference (69).

Together, these data demonstrate that the presence of fibroblasts secreting HGF confers resistance to therapy. In addition, HGF can activate MET, which is expressed on cancer-initiating cells (CICs) in colon cancer, through paracrine signaling (70). This can sustain typical CIC properties, including long-term self-renewal, ultimately leading to resistance to anti-EGFR therapy (70).

Increasing evidence supports the presence of stromal cytokines that are important in the development of tumor chemoresistance.

CCL2 is an inflammatory chemokine, which is recruited by immune cells into the tumor microenvironment and has been demonstrated to confer resistance to paclitaxel, and docetaxel in prostate cancer (71). A previous study demonstrated that CCL2 expression is higher in three different paclitaxel-resistant ovarian cancer cell lines ES-2/TP, MES-OV/TP and OVCAR-3/TP compared with parental cells (72). Furthermore, treatment with a CCL2 inhibitor enhances the antitumor efficacy of paclitaxel and carboplatin combination therapy in ovarian cancer (72). CAFs can induce CCL2 production through signal transducer and activator of transcription 3 (STAT3) phosphorylation, and in turn, CAF-derived CCL2 promotes cancer progression by regulating cancer stem cells through activation of the Notch signaling pathway (73).

The chemokine CXCL12 is the sole ligand of CXCR4. CAFs are an important source of CXCL12 in the tumor stroma. Previous studies have indicated that CXCL12/CXCR4 signaling contributes to chemoresistance by inducing the activation of focal adhesion kinase, ERK and AKT signaling pathways, enhancing the transcriptional activities of β-catenin, and NF-κB, and the expression of survival proteins (74,75). Disruption of the CXCR4/CXCL12 signaling pathway has been demonstrated to sensitize prostate cancer cells to docetaxel (76). Similar results have been observed in colon (77) and lung (78) cancer. Therefore, these studies suggest that chemokines, including CXCL12, may act as promising targets for cancer therapy, alone and/or in combination with other cytotoxic drugs.

Emerging evidence suggests that the dynamic crosstalk between tumor cells and stromal fibroblasts underlie drug resistance. In CRC, IL-17A, which is overexpressed by CAFs in response to chemotherapy, bind to the IL-17A receptor expressed on CICs (79). Consequently, this results in the maintenance and development of therapeutic resistance of CICs through the upregulation of NF-κB (79). In ER-negative and triple-negative breast cancer, IL-17A protects from docetaxel-induced cell death through activation of ERK1, and 2, thus participating in therapy-resistance development (80).

IL-6, an inflammatory cytokine, is primarily secreted by CAFs. IL-6 promotes the growth and invasion of cancer cells through activation of STAT3 (81). NSCLC cells expressing persistently activated mutant EGFR are also associated with the IL-6 signaling pathway, which promotes the proliferation and survival of cells, leading to erlotinib resistance (82,83). IL-6 secreted by CAFs induces tamoxifen resistance through activation of the Janus kinase (JAK)/STAT3 and PI3K/AKT signaling pathways in breast cancer cells (84). Inhibition of proteasome activity, IL-6 activity or the JAK/STAT3, or PI3K/AKT signaling pathways markedly reduced CAF-induced tamoxifen resistance (84). These results demonstrate that IL-6 creates a ‘protective niche’ that maintains the survival of residual tumor cells, consequently inducing tumor relapse.

WNT16B is an important fibroblast-derived protein and treatment-induced factor that confers chemotherapy resistance. The chemotherapy resistance effects of fibroblast-derived WNT16B have been detected in vivo and in vitro, indicating that WNT16B reduces apoptosis induced by chemotherapy drugs in prostatic carcinoma (85). This study guides novel directions for combination therapies, including targeting fibroblast-derived WNT16B, which may reverse chemoresistance in breast and prostate cancer (85). Fibroblast-secreted high mobility group protein B1 is released into the tumor microenvironment and performs paracrine signaling on neighboring cancer cells, which has been suggested to induce chemoresistance in breast cancer (86).

Therapy directed at specific fibroblast markers or the antigens presented on CAFs make CAFs particularly sensitive to cancer treatment. FAP is a membrane protein that is exclusively overexpressed on CAFs (93). FAP has been shown to support tumor growth and proliferation, making it a potential target for novel anticancer therapies (94). FAP-specific molecules selectively target fibroblasts and finally inhibit the growth of surrounding cancer cells (94,95).

FAPα-specific monoclonal antibodies have demonstrated therapeutic potential in cancer treatment. FAP5-DM1, a monoclonal maytansinoid-conjugated antibody, was demonstrated to inhibit and cause the complete regression of tumor growth in xenograft models of lung, pancreatic, and head and neck cancer in vivo (96).

Inhibition of FAPα enzyme activity using specific inhibitors has also been considered a promising approach to targeting fibroblasts. Using the peptidase inhibitor, PT-100 (talabostat) was revealed to reduce the tumor growth rate in numerous types of tumor animal models (97). Knocking down FAPα expression resulted in distinct tumor growth regression in an LSL-K-ras^G12D genetic mouse model of lung cancer and in a colon cancer model, suggesting a tumor-supporting role of endogenous FAPα (98). Furthermore, treatment with PT-630 was able to inhibit tumor growth in the lung and colon cancer models (98).