In contrast to the scarce literature on tumor-suppressing functions of CAFs, there are quite some studies which have related various tumor-promoting functions to specific markers expressed by certain CAFs found in human CRCs (Table 1). However, as most of these CAFs have not been extensively characterized, it often remains unclear how they relate to the two major CAF clusters and whether they represent phenotypically distinct subsets or overlap with other tumor-promoting or -suppressing CAFs. In this entry, we provided an overview on (the markers of) these possible CAF subsets, stratified per tumor-promoting function which they have been linked to.

Table 1.

It is well known that CAFs can also facilitate cancer progression by promoting tumor cell resistance to cytotoxic therapies ^[61][62][120,121]. Several studies have demonstrated that certain CAFs in human CRCs are also able to do so. For example, radiation-induced apoptosis in CRC cells can be reduced by CAFs which express microRNA-31 ^[45][122] or microRNA-93–5p ^[46][123]. Moreover, CAFs which express long non-coding RNA (lncRNA) H19 ^[48] [124] and colorectal cancer-associated lncRNA (CCAL) ^[47][125], can promote CRC cell stemness and chemoresistance via exosomal transfer of these lncRNAs to tumor cells. TGF-β2 secreting CAFs have also been shown to enhance chemoresistance in a paracrine manner by upregulating glioma-associated oncogene family zinc finger 2 (GLI2) expression in CRC stem cells ^[49][126]. These CAFs may serve as an important target for prognostic applications, since high expression levels of TGF-β2 were strongly associated with an increased risk of relapse in chemotherapy-treated CRC patients ^[49][126]. In addition, therapeutic targeting of these CAFs may also become feasible in the (near) future, with several TGF-β inhibitors already being tested in clinical trials ^{[63][64][65][66]}[127,128,129,130]. Next to supporting therapy resistance, certain CAFs in advanced CRCs can actively contribute to immune evasion of tumor cells (e.g., via aberrant TGF-β signaling ^[67][131]) and inhibition of anti-tumor immune responses ^{[68][69][70][71]}[132,133,134,135]. Li et al. reported that CAFs expressing C-X-C motif chemokine 5 (CXCL5) can promote expression of programmed death-ligand 1 (PD-L1) ^[50][136], an important suppressor of T-cell activity ^[72][137]. The fact that CAFs are able to produce a defense against T-cells could have important implications for immunotherapies against CRC ^{[68][69][70][71]}[132,133,134,135]. Moreover, CD70-expressing CAFs were shown to increase the survival of naturally occurring regulatory T-cells ^[51][138], which are key mediators of immunosuppression ^[73][139]. CD73-expressing CAFs were also able to enhance immune suppression via adenosine receptor 2A signaling ^[52][140]. Notably, the abundance of CD70- and CD73-expressing CAFs was significantly associated with a worse prognosis in CRC patients ^[51][52][74][138,140,141], thereby emphasizing the involvement of immunomodulatory CAFs in CRC progression.

2.3. Metastasis

Currently, metastatic disease remains the major cause of death in CRC patients ^[75][153]. CAFs can play an important role in mediating CRC metastasis, as summarized by Tommelein et al. ^[76][154]. Recent work from our group identified several metastasis-promoting CAFs present at the invasive front of primary CRCs, which were also significantly correlated with poor metastasis-free patient survival. These include CAFs which express endoglin ^[53] [150] or BMP2 ^[54][151]. Interestingly, BMP2 expression in CAFs seemed to be regulated by tumor necrosis factor-related apoptosis-inducing ligand (TRAIL), a cytokine which is overexpressed in CRC cells with a deficiency for mothers against decapentaplegic homolog 4 (SMAD4) ^[54][151]. These findings provide further support for the idea that the interaction between CRC cells (with a certain mutational status) and CAFs could eventually progress to a tumor-promoting reciprocal loop ^{[77][78][79][80]}[155,156,157,158]. Other examples of metastasis-promoting CAF populations include IL-11 secreting CAFs, which could activate apoptosis-suppressing programs in metastatic tumor cells ^[55][63], and stanniocalcin-1 (STC1) expressing CAFs ^[56][152]. Lastly, next to CAFs in the primary tumor, several studies have suggested that CAFs at the metastatic site, which could originate from resident fibroblasts in remote organs ^[81][82][83] [159,160,161] or CAFs co-travelling with tumor cells ^[84][85][162,163], can also contribute to CRC metastasis. For instance, resident liver and lung fibroblasts can induce formation of the pre-metastatic niche via (primary) tumor cell-driven upregulation of pro-inflammatory cytokines such as IL-6 and IL-8  ^[81][82][83][159,160,161]. Studies on circulating CAFs originating from primary CRCs have not been conducted yet, but evidence from other cancer types suggests that circulating CAFs may be importantly involved in tumor metastasis and, thus, may serve as useful prognosticators ^[86][164]. For breast and prostate cancer, it has been shown that circulating CAFs are predominantly found in patients with metastatic disease, and rarely in patients with localized cancer ^[87][88][165,166].

3. Conclusions

In brief, various tumor-promoting functions (e.g., enhancing tumor invasion, therapy resistance, or metastasis) have been linked to markers expressed by certain CAF populations in advanced CRC. Some of them also seem to serve as promising targets for prognostic or therapeutic applications. However, clinical translation of these findings should be cautiously considered, since it is yet unknown whether these CAF markers demarcate functionally distinct subsets or show overlap with other tumor-promoting or -suppressing CAFs. To enable more precise identification and targeting of “unfavorable” CAFs in advanced CRC, attention needs to be paid to comprehensive characterization and reporting of the subsets being investigated in CAF studies.