Cancer and Metastasis Reviews

, Volume 31, Issue 1, pp 195–208

Cancer-associated-fibroblasts and tumour cells: a diabolic liaison driving cancer progression


  • Paolo Cirri
    • Department of Biochemical ScienceUniversity of Florence
    • Department of Biochemical ScienceUniversity of Florence

DOI: 10.1007/s10555-011-9340-x

Cite this article as:
Cirri, P. & Chiarugi, P. Cancer Metastasis Rev (2012) 31: 195. doi:10.1007/s10555-011-9340-x


Several recent papers have now provided compelling experimental evidence that the progression of tumours towards a malignant phenotype does not depend exclusively on the cell-autonomous properties of cancer cells themselves but is also deeply influenced by tumour stroma reactivity, thereby undergoing a strict environmental control. Tumour microenvironmental elements include structural components such as the extracellular matrix or hypoxia as well as stromal cells, either resident cells or recruited from circulating precursors, as macrophages and other inflammatory cells, endothelial cells and cancer-associated fibroblasts (CAFs). All these elements synergistically play a specific role in cancer progression. This review summarizes our current knowledge on the role of CAFs in tumour progression, with a particular focus on the biunivocal interplay between CAFs and cancer cells leading to the activation of the epithelial–mesenchymal transition programme and the achievement of stem cell traits, as well as to the metabolic reprogramming of both stromal and cancer cells. Recent advances on the role of CAFs in the preparation of metastatic niche, as well as the controversial origin of CAFs, are discussed in light of the new emerging therapeutic implications of targeting CAFs.


Cancer-associated fibroblastsTumour microenvironmentTumour metabolismEpithelial–mesenchymal transitionECM remodellingHormonal dependence

1 Stromal reactivity in cancer

Fibroblasts are the most abundant cell type in connective tissues in which they form the structural framework by secreting extracellular matrix (ECM) components [1]. Quiescent fibroblasts undergo activation during tissue remodelling and become myofibroblasts (MFs), as originally described by Giulio Gabbiani in 1971 [1]. During wound healing or fibrosis, these MFs gain contractile stress fibres by expressing α-smooth muscle actin (α-SMA) and form cell–cell contacts through gap junctions [2]. Upon completion of the wound-healing process, activated fibroblasts undergo a particular type of programmed cell death, called nemosis, and are removed by the granulation tissue [3, 4]. Fibrosis is characterized by the deregulation of this process and persistence of activated fibroblasts in the tissue, with the consequent accumulation of redundant ECM.

As tumours have been described as “wounds that do not heal” [5], it is reasonable that CAFs share some similarities with MFs, including expression of SMA. Conversely, CAFs are persistently present in tumour stroma and are not removed by apoptosis. In addition, their activation is not reversible, and, like in fibrosis, this leads to tumour desmopasia through excessive ECM deposition. CAFs are extremely abundant in the stroma of many tumours, including breast, prostate and pancreatic carcinomas [6, 7]. Of note is that CAFs are heterogeneous populations and their relative composition greatly different among different tumours. These subpopulations collectively share the “activation state”, although their expression of acknowledged activation markers may differ. The main activation markers are α-SMA and fibroblast-specific protein, although the overexpression of platelet-derived growth factor (PDGF) receptors-β and fibroblast activation protein (FAP) have been frequently observed in the stromal fibroblasts of solid tumours [6, 8, 9]. Other proteins expressed by stromal fibroblasts have been shown to have a prognostic value for solid tumours. In particular, a poor prognosis has been associated with expression in CAFs of p53 tumour suppressor in ductal carcinoma [10], or the hypoxia marker carbonic anidrase IX in human lung adenocarcinoma [11], or periostin in cholangiocarcinoma [12]. On the contrary, the expression of caveolin-1, PTEN or podoplanin in tumour stroma correlates with a favourable prognosis for several cancers [1315]. In this line, p53 inactivation or the genetic inactivation of PTEN in the stromal fibroblasts has been reported to exert a tumour promoter effect in breast carcinoma, thereby accelerating both cancer onset and progression [1517]. This large heterogeneity in marker expression for CAF subpopulations or in CAFs originating from different tumours may be explained by their possible miscellaneous origin. Indeed, CAFs are variously reported to stem from resident fibroblasts, bone marrow-derived progenitor cells, or from epigenetic transitions from endothelial or cancer cells through endothelial–mesenchymal transition or epithelial–mesenchymal transition (EMT) [1821].

The role of CAFs in tumour progression is multifaceted as they can either inhibit or promote malignant growth. Similar to immune cells, CAFs repress the early stages of tumour progression, mainly facilitating the formation of gap junctions between activated fibroblasts and thereby exerting a contact inhibition on cancer cells [19, 20]. Conversely, due to the conspiracy of several tumour-recruited cells, as well as to tumour cells themselves, local CAFs become activated and circulating cells are recruited to transdifferentiate into activated CAFs. To this end, multiple cells are required, including macrophages, lymphocytes, neutrophils, endothelial cells, bone marrow-derived cells, etc. (Fig. 1). Upon recruitment of fibroblasts, their activation to CAFs is indeed achieved by tumour-secreted factors and promote both tumour growth and progression. Two intimately interactive pathways are established in the cross talk linking cancer and stromal cells: (a) in the “efferent” pathway, cancer cells prompt a reactive response in the stroma; (b) in the “afferent” pathway, the activated stromal cells influence cancer cell malignancy [22, 23] (Fig. 2).
Fig. 1

The stromal context in which CAFs evolve. Degradation of the basement membrane allows tumour cells to interact with resting resident fibroblast, thereby altering their phenotype towards CAFs. Besides resident fibroblasts, CAFs may also originate from bone marrow-derived mesenchymal stem cells, which differentiate into CAFs in response to tumour-derived stimuli. Invading tumour cells secrete growth factors that stimulate angiogenesis and inflammation through the recruitment and activation of several other cells, including endothelial precursor cells or tumour-associated neuthrophils or macrophages. These activated stromal cells, together with tumour cells themselves, secrete several ECM-degrading enzymes, including MMPs, uPA and cathepsins, whose collective activity further fuels stromal CAFs proliferation and tumour cell invasion
Fig. 2

Interplay between CAFs and tumour cells: efferent and afferent pathways. Tumour progression needs a reciprocal feed-forward loop between CAFs and cancer cells. Cancer cells induce and maintain the fibroblast-activated phenotype which, in turn, produces a series of growth factors and cytokines that sustain tumour progression by promoting ECM remodelling, cell proliferation, angiogenesis and EMT. a Efferent pathway: in the box we provide a list of factors secreted by cancer cells involved in the mesenchymal–mesenchymal transition (MMT) of fibroblasts towards their activation. b Afferent pathway: in the box we provide a list of factors secreted by CAFs to affect cancer cell behaviour/aggressiveness. Phenotypes affected by CAFs, listed in the arrow, include survival, proliferation, metabolism reprogramming, angiogenic shift, ECM remodelling, inflammatory cell recruitment, EMT activation and stem cell trait achievement

The transdifferentiation of CAFs, commonly identified as mesenchymal–mesenchymal transition (MMT) [6], is at present inadequately understood. Tumour growth factor-1 (TGF-β1) is one of the major pro-fibrotic tumour cell-derived factors affecting CAF activation [24]. Nevertheless, PDGF-α/β [25, 26], basic fibroblast growth factor (bFGF) [27] or interleukin (IL)-6 [23] can also be secreted by cancer cells to activate CAFs.

Compelling data indicate that CAF activation is a redox-regulated process. TGF-β1 leads to the generation of reactive oxygen species (ROS) in CAFs, which are responsible messengers for their achievement of MF phenotype, for the downregulation of gap junctions between CAFs, as well as for their tumour-promoting activity in skin tumours [28, 29]. The role of ROS in the MMT of skin cancer CAFs is further strengthened by the ability of antioxidants to prevent CAF activation, as well as their tumour-promoting function [28]. In prostate carcinoma, the activation of CAFs by tumour-secreted IL-6 is once more redox-dependent [30], and the oxidative stress due to JunD genetic inactivation promotes MMT and tumour spreading in breast adenocarcinoma [31]. In addition, antioxidant treatments abolish the secretion by CAFs of matrix metalloproteases (MMPs) or stromal-derived factor (SDF)-1, thereby disturbing the CAF “efferent” pathway.

2 Role of CAFs in cancer progression

2.1 ECM remodelling

The tumour microenvironment is composed of both cellular (fibroblasts, endothelial and immune cells) and non-cellular components (proteins, proteases, cytokines, etc.): the extracellular matrix (ECM). It has been recently shown that the architecture of ECM impacts on many important cell functions such as proliferation, differentiation and cancer [32, 33]. In fact, ECM structure affects cell morphology, integrin signalling and the cellular cytoskeleton, thereby participating in the control of the cell cycle [34].

The ECM is remodelled physiologically and also pathologically in diseases such as cancer and fibrosis. In the tumour microenvironment, the ECM is subjected to both qualitative and quantitative changes with respect to normal tissue that have been reported to influence the proliferation, survival and migration of cancer cells [35, 36]. CAFs retain a major role in ECM remodelling since they are mainly responsible for the production of ECM proteins (i.e. collagens, fibronectin) as well as proteases and other enzymes involved in the posttranscriptional modification of ECM proteins themselves. In solid tumours, an increased matrix deposition and a concurrent, progressive stiffening of ECM are commonly observed. Indeed, many cancers are associated with desmoplasia, a fibrotic state characterized by an accumulation of type I and III collagens and by a degradation of type IV collagen [37, 38]. Tumour desmoplasia has been associated with the poor prognosis of cancers [39], and it has been observed also in metastatic sites [40].

Tumour ECM stiffening is correlated to an increased number of covalent cross-link between collagen molecules [41]. Collagen cross linking is predominantly catalyzed by lysyl oxidase (LOX), expressed in fibroblasts during the early stages of breast carcinogenesis, whilst in a later stage LOX is induced also in hypoxic carcinoma cells, promoting aggressive growth [42]. Hence, ECM remodelling promoted by LOX activity positively affects tumour cell migration and invasion. Indeed, in a mouse model of breast carcinoma, treatment with LOX inhibitors led to a decrease of ECM cross links, preventing ECM stiffening and delaying tumour progression [41]. In addition, ECM rigidity in tumours contributes, together with impaired vascular and lymphatic mesh, to increase interstitial fluid pressure, leading to a reduced chemotherapeutics delivery inside tumours. In a mouse model of pancreatic adenocarcinoma, the reduction of tumour-associated CAFs leads to an increased efficiency of chemotherapy, thereby suggesting a key role for CAFs also in intra-tumour drug delivery [43]. Together with ECM stiffness, interstitial flow is an important mechanical stress in the tumour stroma. Recently, Shieh et al. [44] have shown that interstitial flow drives TGF-β1- and MMP-dependent fibroblast tissue invasion which, in turn, enhances tumour cell invasion by ECM remodelling. These findings confirm the notion that CAFs may serve as guidance structures that direct the migration of epithelial cancer cells by inducing protease-mediated ECM remodelling. As a matter of fact, imaging of the collective migration of squamous cell carcinoma cells and CAF demonstrated that CAFs behave as leading cells, degrading ECM and creating the path for cancer cells moving within these tracks [45].

Besides collagen, two other ECM components, mostly produced by CAFs—fibronectin and hyaluronan—have been shown to play a role in tumour microenvironment. Fibronectin can be a ligand for a dozen members of the integrin receptor family [46], including the α5β1 receptor, and regulates collagen fibril structure [47], playing important roles in cell adhesion, migration and growth [48]. Tumour-activated fibroblasts show an enhanced expression of fibronectin, together with de novo expression of its variant ED-A [6]. Stromal fibronectin is positively associated with human tumour metastatic potential and MMP secretion [40], and its upregulation, together with transglutaminase, facilitates the metastatic spreading of A431 tumour cells [49]. In addition, fibronectin regulates ovarian cancer metastatic potential by promoting a ligand-independent activation of the c-Met proto-oncogene through binding to α5β1 receptor [50]. CAFs mediated the overexpression of hyaluronan within the tumour microenvironment, and it has been shown to have a role in the recruitment of tumour-associated macrophages, which are key regulatory cells involved in tumour neovascularization through endothelial cell recruitment [51].

2.2 Secretion of soluble factors

The contribution of CAFs to tumour cell proliferation and motility include several factors which drive the so-called afferent pathway (Fig. 2). Hepatocyte growth factor (HGF), epidermal growth factor (EGF), bFGF, as well as cytokines such as SDF-1 and IL-6, are all hugely expressed by CAFs upon contact with different tumour histotypes. For instance, CAFs extracted from lung cancers produce a large amount of HGF, with the consequent activation of the c-Met pathway in neighbouring cancer cells. CAF-secreted HGF gives rise to an increased resistance of tumour cells to conventional tyrosine kinase inhibitors against EGF receptor. Of note is that blocking antibodies against HGF circumvent this acquired resistance [52]. In addition, in breast cancer cells, HGF secreted by CAFs enhanced the activation of uPA/uPAR protease, and treatment with the Met inhibitor SU11274 decreases the activation of this proteolytic system, as well as the invasion of cancer cells, thereby validating the anti-HGF approach for antimetastatic purposes [53]. To complete the cross talk, soluble factors secreted by cancer cells force CAFs to produce HGF [54]. Indeed, cancer cells affect HGF secretion by CAFs through IL-6 [23], prostaglandins [55], or PDGF-β, whilst TGF-β1 suppresses HGF expression in CAFs [56].

Pro-inflammatory cytokines, such as interleukins, interferons and members of the tumour necrosis factor family, are produced both by stromal and cancer cells and exert tumour-modulating effects [4]. First of all, CAF-secreted cytokines and chemokines lead to the infiltration of immune cells, which in turn promotes de novo angiogenesis and helps metastatic spread [57]. In pancreatic cancer cells, SDF-1 and IL-8 cooperate in the promotion of a complete angiogenic response in recruited endothelial cells [58]. SDF-1 secreted by CAFs has a double role in tumour progression as it is involved in the mobilization of endothelial precursor cells from the bone marrow, thereby inducing de novo vasculogenesis, but also in the CXCR4-triggered tumour cell growth [59]. In addition, prostate cancer stromal fibroblasts largely secrete CXCL14 chemokine, active in tumour growth, angiogenesis and macrophage infiltration [60]. Of note is that CAFs associated with developing neoplasia exhibit a pro-inflammatory signature, mainly driven by nuclear factor-κB (NF-κB), expressing SDF-1, IL-6, and IL-1β and recruiting pro-angiogenic macrophages and promoting tumour growth [61]. Furthermore, in breast adenocarcinoma, SDF-1 secreted by CAFs has been found driven by hypoxia-inducible factor-1 (HIF-1), which is in turn activated by hypoxia-mediated oxidative stress [31]. Interestingly, we have recently reported that in prostate carcinoma, contact with CAFs leads cancer cells to activate the same pro-inflammatory gene signature engaging NF-κB and HIF-1. Again, this effect is mediated by oxidative stress and leads cancer cells to achieve a motile phenotype through EMT, thereby confirming that stromal and tumour cells share common pathways during tumour progression [31].

CAFs are also able to secrete several members of the MMP family as well as plasminogen activators. These enzymes may be useful essentially for three reasons: (1) direct degradation of ECM, obviously associated with the generation of space due to tumour expansion, invasion or de novo angiogenesis; (2) cleavage of membrane-bound growth factors or cytokines as well as their receptors; or (3) cleavage of cell adhesion molecules like cadherins, leading to an increased motility and EMT [62, 63]. The expression of tumour (MMP-1, MMP-2 and MMP-14) and stromal (MMP-9, MMP-13 and MMP-14) matrix metalloproteinases is mandatory for squamous cell carcinoma progression [64]. MMP-13 secreted by CAFs promotes tumour angiogenesis by releasing vascular endothelial growth factor (VEGF) from the ECM, thereby leading to the increased invasion of squamous cell carcinoma or in melanoma [65]. We have recently reported that CAFs extracted from aggressive prostate carcinoma secrete a large amount of MMP-9, which in turn induces a clear EMT in prostate carcinoma cells, likely through E-cadherin negative regulation [23]. CAFs are activated by cancer cell-secreted IL-6, a cytokine that has been associated with the senescence of prostate stroma, a known prognostic factor for the aggressiveness of this cancer [66]. We recently observed that MMP-9 secretion by prostate carcinoma CAFs is driven by CAIX acidification of the extracellular milieu. Indeed, activated CAFs strongly expressing CAIX lead to the hyperacidification of co-cultures with prostate carcinoma cells, and CAIX selective inhibitors block the secretion of MMP-9 and EMT of carcinoma cells (unpublished data).

Alongside MMPs, CAFs from colon and breast carcinoma also express uPA and its receptor uPAR [67]. Cancer cells engage with their stromal fibroblasts a specific feed-forward loop mediated by paracrine bFGF and EGF, which induce uPA transcription in the fibroblasts. In turn, the serine protease system helps in activating these paracrine factors from the tumour cell surface/ECM [68].

Accumulating evidence shows that senescence might affect tissue microenvironment reactivity and CAF-mediated responses [6971]. The most significant of effects exerted by senescence onto fibroblasts is the acquisition of a senescence-associated secretory phenotype (SASP), efficiently turning senescent fibroblasts into pro-inflammatory cells, allowing them to increase their pro-inflammatory and pro-angiogenic cytokine secretion. This SASP has been correlated with the ability of fibroblasts to promote tumour progression, at least in part by inducing an EMT in nearby epithelial cells [71].

2.3 Hormonal dependence

Androgen-dependent growth exemplifies the paradigm of epithelial–stromal cross talk in normal and transformed tissues. For example, during development, epithelial prostate growth is sustained by an androgen receptor (AR)-dependent signalling mediated by prostate stromal cell. Fully differentiated prostate epithelial cells require the expression of functional AR in both the stromal and epithelial compartments of the adult prostate [72, 73], but in this condition, stromal cells are required to maintain epithelial cell differentiation in an androgen-dependent manner. In prostate cancer cells, alteration of the androgen-regulated pathways in both cellular compartments may affect the tumour development and progression, shifting from androgen-dependent to castration-resistant tumour [74]. In particular, in the later stage of tumour progression, concurrent overexpression of non-functional AR in the malignant epithelia and loss of AR immunoreactivity in the surrounding stroma were associated with a higher PSA level, and earlier and higher relapse rates after radical prostate radical prostatectomy [75, 76]. These results suggest the need to further investigate the mechanistic basis for the loss of AR expression in the malignant stroma and its potential role in the progression of prostate carcinoma (PCa).

Breast cancer is another kind of hormone-dependent tumour where a microenvironmental interaction between the stroma and cancer cells is mandatory. Oestrogen receptor (ER) signalling through genomic and non-genomic pathways promote the proliferation of breast cancer cells. The incidence of breast cancer is high even in postmenopausal women since oestrogen is produced within the breast cancer tissue consequently to a cross talk between tumour and stromal cells in the tumour microenvironment mediated by cytokines such as TNF-α [77], prostaglandin E2, and interleukin-6 and interleukin-11 [78]. As a consequence, in postmenopausal breast cancers, tumour cells activate stromal fibroblasts to express aromatase [79], an enzyme involved in oestrogen biosynthesis, resulting in intratumoral oestrogen production. Many clinical trials have reported that aromatase inhibitors are effective in endocrine treatment for ER-positive breast cancer [80].

2.4 Regulation of motility and stemness

Clinical and experimental data sustain the hypothesis that CAFs regulate cell motility and the metastatic spread towards secondary organs, a key feature of cancer progression towards a malignant state [81]. Efferent signals from cancer cells trigger a response in reactive stromal cells, thereby engaging a cycle of paracrine afferent signals, sensed by cancer cells which activate a pleiotropic response toward malignancy (Fig. 2). Efferent signals impacting on fibroblast differentiation, proliferation and motility have been identified as transient heterotypic cell–cell contacts or secreted growth factors, chemokines or lipid products such as PDGF-α/β, TGFβ1, bFGF, IL-6, LPA and eicosanoids [22, 82]. Resident stromal cells are thereby activated to enhance their contractility and ECM production. In addition, other non-resident cells, including pericytes or circulating mesenchymal stem cells, may be attracted to the primary tumour site. Whatever their origin, these cells differentiate into CAFs and prompt cancer cells to exit from dormancy, affecting their motility and aggressiveness. Studies on the proteome/secretome of activated CAFs revealed that they can induce invasive growth by either cell–cell contacts or by paracrine diffusible signals, including TGFβ, HGF, VEGF, FGF, SDF-1, as well as matrix metalloproteases, cathepsins and plasminogen activators (Fig. 2) [81, 82].

The proinvasive activity of human CAFs in vitro was shown by De Wever et al. [83] in human colon cancer cells using stromal fibroblasts isolated from surgical colon cancer fragments. In keeping with this, CAFs are able to stimulate the invasive growth of breast and colon cancer cells in co-implantation tumour xenograft mouse models [59, 83]. A similar motogenic effect was also shown for CAFs isolated from surgical prostate carcinoma specimens [23]. In this model, CAFs exert a very powerful metastatic spur as they induce spontaneous lung metastases upon co-injection of cancer and stromal cells.

CAFs mainly contribute to the invasive process by inducing the EMT of tumour cells, a known epigenetic programme leading cancer cells to engage a mesenchymal, motile and proteolytic phenotype [84, 85]. In addition to the pro-migratory spur, EMT has also been correlated with the induction of a cancer stem cell phenotype. Indeed, in breast cancer cells, EMT, due to the overexpression of Snail or Twist transcription factors, is correlated with increased CD44/CD24 ratio as well as with the generation of tumour-initiating cells which spread metastases [86, 87]. In keeping with this, Giannoni [23] reported that CAFs isolated by prostate carcinoma specimens, by means of activating the EMT epigenetic programme, promote/select the generation of cancer stem cells. Indeed, CAFs affect the self-renewal ability of carcinoma cells and enhance their expression of cancer stem cell markers (CD133+ and high CD44/CD24 ratio) and the formation of non-adherent prostaspheres, a property associated with prostate stem cells [23, 8789]. Analysis of tumour-forming ability and spontaneous lung metastasis formation after contact with CAFs revealed that the activated stroma contributes to generate a population of prostate cancer stem cells with defined ability to form primary tumours and distant metastases [23]. In keeping with this, in a conditional Pten deletion mouse model of prostate adenocarcinoma, CAFs confirm their key role in the regulation of stemness properties as they enhance spheroid formation and prostate glandular structures with lesions, high proliferative index and tumour-like histopathology [90].

Whilst embryonic EMT is normally a self-resolving process, EMT serving tissue repair or metastatic dissemination is often correlated with concomitant inflammation and continues until the provoking cause is eliminated [84]. In this framework, CAFs are active in sustaining a chronic pro-inflammatory spur, eliciting multiple effects concurring to induce/select cancer cell phenotypes resistant to the hostile tumour microenvironment. In incipient tumours, CAFs have been shown to orchestrate macrophage recruitment and neovascularization in strict dependence on nuclear factor-κB (NF-κB) activation [61, 91]. Furthermore, CAFs exert their propelling role for EMT by eliciting similar pro-inflammatory pathways in metastatic cells, exploiting oxidative stress and involving the activation of COX-2, NF-kB and HIF-1 [30] (Fig. 3). The fact that cancer cells undergoing EMT in response to CAF contact share with inflammatory cells the same signals is in keeping with the idea that metastasis is a phenomenon reminiscent of the migratory/invasive behaviour of inflammatory cells.
Fig. 3

Molecular pathways engaged by CAFs to elicit EMT of cancer cells. CAFs can affect the EMT of cancer cells by secreting either cytokines or MMPs. In particular, the secretion of TNF-α induces protein stabilisation of Snail and β-catenin by inhibiting GSK-3β-mediated phosphorylation through NF-κB and Akt signalling pathways. In parallel, TNF-α leads to CSN2 expression again through a NF-κB-dependent pathway. Jointly, GSK-3β and CSN-2-dependent signalling events start EMT in tumour cells through the activation of Snail transcription factor. In parallel, CAF-delivered MMPs can activate the small GTPase Rac1 and, consequently, the production of reactive oxygen species (ROS) from cycloxygenase-2. The result is a redox-dependent activation of a transcriptional response driven by hypoxia-induced factor-1 (HIF-1). In turn, the activation of HIF-1 leads to Twist-mediated EMT activation and to the expression of carbonic anydrase-9 (CA9), which, through acidification of the tumour microenvironment, leads to a further activation of MMPs and strengthening EMT

Besides epigenetic mechanisms influencing cross-signalling between CAFs and cancer cells, malignancy may also be regulated by stromal mutations affecting the motility and aggressiveness of cancer cells. Mutations in stromal cells have been attributed to several reasons. First, EMT induced by CAFs in cancer cells may be directly responsible for the accumulation of mutations by tumour or stromal cells. Indeed, stromal-derived MMP-3, which is frequently upregulated in breast cancers, induces genomic instability through the upregulation of ROS [92]. Alternatively, mutated stromal cells might directly derive from cancer cells that have undergone EMT and achieve the characteristics of CAFs [93]. Finally, in mouse prostate carcinoma cells, a paracrine mechanism determines a selective pressure for the expansion of a p53-lacking subpopulation of CAFs [16]. These CAFs lacking p53 contribute to cancer invasion and to the eventual loss of p53 in the epithelium. Furthermore, p53 mutations in CAFs are also significantly associated with lymph node metastases in several sporadic breast cancers [94]. Whatever their origin, these mutated CAFs show a profoundly altered behaviour and concur to tumour malignancy.

2.5 Tumour metabolism remodelling

Glucose metabolism commonly differs between cancer and normal cells. Definitely, cancer cells mainly use glucose by glycolysis, producing lactate even in the presence of oxygen (the so-called Warburg effect or aerobic glycolysis), whilst normal cells fully exploit glucose by oxidative phosphorylation [95]. Aerobic glycolysis, coupled with increased glucose uptake due to incomplete glucose oxidation, assists in proliferating cells, an efficient anabolism from glycolytic intermediates needed to increase cancer biomass [96]. The M1 or M2 splice isoforms of pyruvate kinase (PK), a mandatory regulatory glycolytic enzyme, shift glucose metabolism towards aerobic glycolysis (PKM2) or oxidative phosphorylation (PKM1) [97]. All cancer cells studied to date exclusively express PKM2, an enzyme with lower catalytic activity with respect to PKM1, whereas cells in many normal differentiated tissues express PKM1 [97]. Recently, Vander Heiden et al. [98] clarified the molecular basis of the Warburg effect, demonstrating that cancer cells use PKM2 to short-circuit ATP production and avoid the inhibition of glycolysis. The expression of PMK2 in cancers can also be explained by its functional association with HIF-1, which enhances the transcriptional activity of the master regulator of hypoxia response, a common feature of many cancers [99].

CAFs have been shown to participate in the complex metabolism of tumours, engaging a bidirectional liaison with tumour cells, prompting them to overcome energy depletion due to the Warburg effect. Fibroblasts in contact with epithelial cancer cells undergo differentiation and produce lactate and pyruvate through aerobic glycolysis. Histopathological analysis reveals that PKM2 and lactate dehydrogenase are highly expressed in the stroma of breast cancer lacking caveolin-1 expression [100]. Fibroblasts undergoing activation due to deletion of caveolin-1, or in response to the downregulation of caveolin-1 upon oxidative stress induced by contact with cancer cells, show a stabilisation of HIF-1. The metabolic consequence in these CAFs is the HIF-1-mediated shift towards aerobic glycolysis and the elimination of mitochondrial activity through mitophagy [101]. Although it is likely that cancer cells may then gain a further benefit from their simultaneous expression of HIF-1 and PKM2 for transcriptional purposes, this hypothesis remains to be investigated.

Although these data are intriguing and pose the molecular basis for the role of CAFs in controlling the metabolism of tumour cells, for the moment, they remain limited to CAFs activated by the deletion/downregulation of caveolin-1. In colorectal carcinoma, histological analyses suggested an opposite behaviour of the tumoral stroma. Indeed, the stroma infiltrating these tumours expresses aerobic glycolysis enzymes, and the authors propose a role of these stromal cells to buffer and recycle products of anaerobic metabolism of cancer cells in order to sustain invasive cancer growth [102]. Due to these controversial findings, further confirmations in other experimental settings, as in ex vivo CAFs, are thus needed. However, in keeping with this role of CAFs as metabolic synergistic bystanders, we recently observed that the contact between prostate CAFs and their carcinoma cells gives rise to an increase in their sensitivity to stresses such as pH and hypoxia and sustains their proliferation. This “corrupted” stroma produces energy-rich metabolites like lactate, which are uploaded by cancer cells. Indeed, whilst CAFs express the monocarboxylate trasporter-4 (MCT-4), driving efflux of lactate, cancer cells in the presence of CAFs express MCT-1, which allow them to upload this lactate. Cancer cells can then use the energy-rich lactate either for TCA cycle and ATP production or in anabolic pathways allowing biomass increase and proliferation (Fig. 4, unpublished results). The advantages gained by cancer cells through CAF contact could be twofold: the establishment of a Cori cycle serving the upload of lactate, i.e. energy-rich metabolites to fuel their TCA cycle, and the protection from apoptosis induced by the hostile tumour microenvironment lacking nutrients.
Fig. 4

Metabolic interplay between CAFs and cancer cells. Tumour cells are characterized by an high proliferation rate and consequently undergo a strong activation of the anabolic pathways that allow rapid growth. The model reported in this figure deals with PCa cells and their CAFs or healthy non-activated fibroblasts (human prostate fibroblasts, HPFs). Both cancer cells and CAFs undergo a regulated metabolic reprogramming in the tumour microenvironment. In particular, CAFs undergo a clear Warburg effect, experiencing normoxic expression of HIF-1 transcription factor, an increase of glucose uptake, expression of PKM2, lactate dehydrogenase-5 isoenzyme and the efflux lactate transporter MCT-4. In parallel, cancer cells, again through an HIF-1 dependent mechanism, exploit the energy-rich lactate (uploaded through the expression of the influx lactate transporter MCT-1) to fuel either the Krebs cycle or anabolism/growth. Both cancer cells and CAFs also express HIF-1 target carbonic anydrase IX, responsible for the hyperacidification of the extracellular milieu, further forcing EMT through MMP activation. CAIX carbonic anydrase IX, CM conditioned medium, HIF-1 hypoxia-inducible factor, MCT monocarboxylate transporter, LDH lactate dehydrogenase, PCa prostate carcinoma cells, PKM2 pyruvate kinase M2

2.6 Role of miRNAs

MicroRNAs (miRNAs) are short, natural (20–25 nucleotides) non-coding RNA molecules able to pair with the 3′ untranslated region of their target mRNAs. The effect of this interaction is either to impair mRNA translation or to deliver the mRNA transcript to degradation. Considering that the human genome contains ~1,200 miRNAs and that aberrant patterns of some miRNA expression are implicated in cancer progression [103, 104], it is not surprising that they are also involved in the regulation of the interplay between tumours and their stromal counterpart.

Very recent studies indicate miRNAs as key players of the tumour-supportive capacity of CAFs. Indeed, it has been reported that both miR-15 and miR-16 are downregulated in fibroblast adjacent prostate carcinoma, thereby concurring in promoting tumour progression, mainly acting through FGF-2 signalling. The downregulation of miR-15 and miR-16 leads to a reduced posttranscriptional repression of FGF-2 and its receptor, which enhances cancer cell survival, proliferation and migration. Moreover, the reconstitution of miR-15 and miR-16 validates their role in CAF differentiation in vivo [105]. In addition to miR-15 and miR-16, also miR-21 has been observed in fibroblast-like cells located in the stromal compartment of the tumours [106]. Noteworthy is that the expression of miR-21 was increased in activated fibroblasts after treatment with TGF-β1 or by treatment with a conditioned medium from cancer cells, acting in MFs through protection from apoptosis [107]. Eleven microRNAs have been found to be differentially expressed in CAFs isolated from endometrial cancers with respect to normal tissues, with miR-31 being the most downregulated. The overexpression of miR-31, specifically targeting the homeobox gene SATB2 responsible for chromatin remodelling and the regulation of gene expression, significantly impaired the ability of CAFs to stimulate tumour cell migration and invasion without affecting tumour cell proliferation [108].

Interestingly, miRNAs can be exchanged between stromal and cancer cells and therefore may play additional roles in the two cell types. The mechanisms of exchange are essentially twofold: gap junctions and exsomes/vescicles. In breast carcinoma, a transmission of miRNAs from bone marrow-derived stromal cells has been observed [109]. An analysis of miRNA transmitted by stromal cells via gap junctions identified numerous miRNAs implicated in cell proliferation including miR-127, miR-197, miR-222 and miR-223 targeting SDF-1. In addition, tumour-derived exosomes act as a vehicle for the exchange of genetic information via miRNAs among tumour and stromal cells [110]. In human renal cell carcinoma, the release of these exosomes containing a set of pro-angiogenic mRNAs and microRNAs triggers angiogenesis and promotes the formation of a pre-metastatic niche, contributing to prompting the angiogenic switch and coordinating metastatic diffusion.

Although studies on the role of miRNAs in tumour/stroma interplay are still at their infancy, we can speculate that this will be a fertile topic in future cancer research. For example, as one of the master effects elicited by CAFs in tumours is the activation of the EMT process, the involvement of miRNAs acknowledged to regulate EMT (as the miR200 and miR205 families) is highly expected in tumours sensible to stromal effects.

2.7 Preparation of metastatic niche

Only a minority of transformed cells undertake the metastatic route, and of those, an even smaller fraction succeeds in this task [111]. Metastatic cells share some properties with somatic or embryonic stem cells, i.e. self-renewal capability, which led to the hypothesis that metastatic cells are indeed cancer stem cells [112]. CAFs within the primary tumour are involved also in the preparation of the metastatic site and have shown stem-like traits (see above) [23] themselves (Fig. 5). It has been known that metastatic cancer cells preferentially grow in secondary sites with a particular and selected microenvironment [113].
Fig. 5

Role of CAFs alongside the metastatic route. Within the primary tumour, resident or bone marrow-derived cells undergo differentiation into CAFs and cooperate to increase the aggressiveness of cancers cells, affecting their invasiveness through epithelial–mesenchymal transition (EMT), their resistance to several stresses such as hypoxia, acidity or chemotherapeutic drugs, as well as their metabolism sustaining the growth of tumour mass even in nutrient restriction. Cells that achieve the metastatic potential then bring their own CAFs from the primary site. These co-travelling CAFs first provide a pro-survival spur to cancer cells, protecting them from naturally occurring anoikis (i.e. apoptosis due to lack of proper adhesion) and facilitating extravasation. Then, within the metastatic niche, CAFs provide an early growth advantage to the accompanying metastatic cancer cells in the secondary site. In addition, CAFs also concur to the secretion of soluble cytokines/chemokines which establish the final organization of the metastatic tumour in the new tissue, regulate mesenchymal–epithelial transition (MET) and allow subsequent colony growth

It has been recently reported that metastatic cells bring CAFs originating from the primary tumours to the metastatic site [114]. In a first instance, CAFs protect metastatic cells circulating in the bloodstream from apoptosis, and once arriving in the metastatic site, the co-travelling CAFs provide an early growth advantage to cancer cells that do not enter dormancy [114]. In addition, prostate cancer cells and fibroblast residents in the metastatic niche co-evolve, thereby accelerating metastatic cell growth [115]. Recent studies suggest that mesenchymal stem cells recruited to the human cell carcinoma secrete discrete microvesicles acting as a vehicle for the exchange of genetic information among stromal and tumour cells. These vesicles contain several pro-angiogenic mRNAs and microRNAs, trigger angiogenesis and promote the formation of a pre-metastatic niche [110]. Hence, metastatic cells affect the reactivity of stromal fibroblasts, i.e. induction of expression of ECM proteins and chemokines, in a similar manner to the biunivocal interplay engaged within the primary tumour [115].

3 Concluding remarks and perspectives

CAFs have recently emerged in recent years as a promising therapeutic target to counteract cancer progression. The main reasons are: (1) CAFs are genetically stable in comparison to cancer cells, reducing the risk of drug resistance onset; (2) CAFs are mainly responsible for the structure of tumour ECM that hampers the diffusion of anticancer agents through solid tumours; (3) CAFs favours the survival, proliferation and invasive features of cancer cells.

Tumour stroma-directed therapies can interfere with growth factor/cytokine-mediated cross talk within the tumour microenvironment, promoting an anticancer effect. In a mouse model of cervical carcinogenesis, the blockade of PDGF receptor signalling in CAFs inhibits the progression of premalignant lesions [116], whilst the overexpression of trombospondin-1, an angiogenesis inhibitor, has been shown to inhibit cervical tumour growth accompanied by a decrease of two markers of CAFs, αSMA and desmin [117]. Preclinical studies using NK4, a competitive antagonist of Met, as well as anti-HGF monoclonal antibodies showed a remarkable inhibition of tumour growth and metastasis [118, 119]. In addition, inhibition of stromal cell proliferation in a pancreatic cancer model, using a specific inhibitor of the Hedgehog receptor, allows the improved delivery of gemcitabine to the tumour and increased survival [43]. PDGF-C produced by CAFs is able to induce VEGF production of tumour cells, thereby sustaining the angiogenic shift. It has been shown that anti-PDGF-C antibodies can be used in order to inhibit angiogenesis in tumour not sensible to anti-VEGF treatment [120].

The involvement of COX-2 in CAF–tumour cell cross talk has been reported in prostate carcinoma malignancy. COX-2 is a protein involved in the inflammatory response whose expression is markedly increased when fibroblasts are co-cultured with cancer cells [121]. The upregulation of COX-2 in a mouse xenograft model resulted in increased VEGF and MMP14 expression, which contributes to cancer progression and invasion [122]. Hence, COX-2 inhibitors such as celecoxib or refecoxib represent a promising therapeutic tool for targeting CAF-induced effects on tumour progression.

Another promising approach is the inhibition of CAF MMT in order to eliminate efferent ways affecting cancer aggressiveness. DNA methyl-tranferase 1 by 5-aza-2-deoxycytidine blocks in vitro hepatic stellate cell/myofibroblast differentiation [123], whilst a monoclonal Ab against FAP, a protein involved in the MMT process, is extremely promising in clinical trials [124]. In addition, the MMT of skin cancer fibroblasts in response to TGF-β is abolished by antioxidant treatment using trolox or selenite, thus inhibiting cancer progression towards most aggressive phenotypes [28].

In conclusion, the large range of therapeutic approaches that are emerging from basic and applied research are promising and encouraging. On the other hand, considering the variability of the roles played by CAFs in each tumour histotype as well as the multiple ways in which activated fibroblasts can be generated, every future therapy directed against CAFs will be more effective in the presence of additional studies both on CAF taxonomy and in a more accurate and detailed comprehension of CAF–tumour relationship.

Copyright information

© Springer Science+Business Media, LLC 2011