Cancer Associated Fibroblast: Mediators of Tumorigenesis

Abstract

It is well accepted that the tumor microenvironment plays a pivotal role in cancer onset, development, and progression. The majority of clinical interventions are designed to target either cancer or stroma cells. These emphases have been directed by one of two prevailing theories in the field, the Somatic Mutation Theory and the Tissue Organization Field Theory, which represent two seemingly opposing concepts. This review proposes that the two theories are mutually inclusive and should be concurrently considered for cancer treatments. Specifically, this review discusses the dynamic and reciprocal processes between stromal cells and extracellular matrices, using pancreatic cancer as an example, to demonstrate the inclusivity of the theories. Furthermore, this review highlights the functions of cancer associated fibroblast, which represent the major stromal cell, as important mediators of the known cancer hallmarks that the two theories attempt to explain.

Introduction

Technological advances, such as whole genome and single cell sequencing, have deepened our understanding of cancer pathophysiology. As a result, these advances have facilitated breakthroughs in drug development, which have significantly prolonged survival and improved the quality of life for numerous patients. However, the lack of a uniform patient response to therapeutics, along with the national and global rise in cancer incidence and mortality [1, 2]; reveal the limitations in our understanding of tumor biology. To address these limitations, a framework by which the functional complexities of cancer could be studied was established and collectively referred to as the Hallmarks of Cancer [3, 4].

Cancer hallmarks describe normal physiological processes that are inappropriately instigated and/or terminated due to embellished/altered regulation of the tissues’ homeostatic equilibrium. Healthy tissue homeostasis is typically maintained via a coordinated balance between intrinsic (i.e., intracellular) programs and extrinsic (i.e., extracellular matrix -ECM-, growth factor availability, and more) factors. When the balance between the cell and its collective microenvironment is perturbed, the system initiates discrete programs that, in the case of cancer, enable cells to acquire growth and survival advantages such as sustained proliferation, restrained cell death and the ability to metastasize. For most cancers, this is a common theme that is often exploited by anti-cancer therapies, which are designed to interfere with cancer hallmarks. While there has been much success in some cancers with this treatment approach, the development of drug resistance and cancer recurrence remain among the major concerns [5]. The revised hallmarks of cancer [3], together with these challenges, highlight the fact that cancer involves more than just mutational cell transformation and cell-intrinsic programs. In fact, the hallmarks suggest that tumorigenesis consists of a conglomerate of rather complex, dynamic, and reciprocal systems amid cancer cells and their microenvironment [6], which continuously evolve in response to homeostatic perturbations. Taken together, this suggests that cancer hallmarks are both drivers and consequences of tumorigenesis to which the role of the tumor microenvironment is central to both disease initiation and progression. To support this notion, we highlight the two prevailing theories that attempt to explain the process of tumorigenesis: Somatic Mutation Theory (SMT) and the Tissue Organization Field Theory (TOFT). We then unify the two theories, to emphasize these are mutually inclusive. Finally, we depict the functional reciprocity between one of the major stromal cells, cancer associated fibroblasts (CAFs), and the ECM they produce, to highlight the role of CAFs in tumorigenesis accounting for the Cancer Hallmarks explained by SMT and TOFT.

Prevailing Theories of Tumorigenesis: SMT vs TOFT

Somatic Mutation Theory (SMT)

The premise of the Somatic Mutation Theory (SMT) was first established by Theodor Boveri, who described cancer as a disease solely driven by specific permanent changes in chromatin complexes that prompt mature cells to continuously divide [7, 8]. As additional prominent genetic discoveries were made, Boveri’s theory was later refined to state cancer ensued from driver mutations in DNA that cause gain (oncogenes) or loss of function (tumor suppressor genes), which altogether promote accelerated cell growth [9–12]. The idea that cancer is instigated by “driver mutations” has led researchers to query the required mutational load for cancer to initiate, whereby the Knudson hypothesis proposed that a minimum of two genetic alterations are needed for cancer to develop [13]. Boveri’s SMT was further strengthened by the demonstration that increased frequency of mutational burden strongly correlates with cancer aggressiveness [9–12, 14, 15]. SMT states that cancer is a cellular based disease in which explicit genetic mutations give rise to tumors due to intrinsic loss of cell regulatory programs that induce cell proliferation and/or alter cell differentiation [4, 16].

The SMT is highly revered in the cancer field and has rendered enormous leaps in our understanding of malignant diseases as well as in the development of targeted therapies. However, this theory does not fully explain all aspects of cancer. For instance, SMT does not explicate why neoplasms of inheritable cancers only develop in specific organs when the “germline” mutation exists in every cell [17–20]. More importantly, the SMT model suggests that cancer is irreversible in that mutated cells remain committed to malignant transformation (despite external stimuli). However, tumor regression (spontaneous and intervention based) is evident, in spite of the presence of tumor promoting mutations [11, 21–25]. Considering two of the most broadly used cancer interventions (radiotherapy and chemotherapy) are considerably effective in reducing tumor growth, often indiscriminate of driver mutations, dampens the validity of this SMT claim. [21, 26–28].

SMT seems insufficient for explaining functional variations imparted solely by altering cell culturing conditions [28–30]. For example, a groundbreaking study examined morphological and proliferative rates of a tumorigenic cell-line and its non-tumorigenic breast epithelial precursor. In this study, cells were initially cultured using classic culturing conditions (i.e., in 2D), where little differences in morphology and proliferation were apparent. However, when the same cells were cultured using a three-dimensional (3D) culturing system that mimics the in vivo environment, differences became apparent between non-tumorigenic and tumorigenic cells. This study, and others, ultimately revealed the importance of including microenvironmental appropriate conditions for the study of normal vs. malignant cells [31]. Collectively, these findings bring to light that in addition to genetic alterations, there are extrinsic factors that also influence tumorigenic programs. The tissue organization field theory (TOFT) takes these observations into consideration.

Tissue Organization Field Theory (TOFT)

The premise of TOFT began with careful histological observations of tumor tissue. Pathologists observed that tissues, and not just cancer cells, lacked the “normal” defined physiological architecture. This observation established the basis for pathology-informed diagnosis and led German scientists to define cancer as a tissue-based disease prior to Boveri’s cell-based speculation [32]. TOFT was thereafter framed to state that cancer is a systemic disease and suggests that proliferation is the default state of cells regulated by the local microenvironment [33, 34]. Hence, TOFT considers the dynamic and reciprocal relationship between cancer cells and their microenvironment, also known as the stroma, to be essential in restraining or promoting tumorigenesis. Of note, renewed interest in considering TOFT arose from the recent excitement generated by anti-cancer immune therapies.

Additionally, TOFT is further supported by studies where stromal cells of the primary tumor, which lack oncogenic or tumor suppressive mutations, were found to enhance the metastatic potential of the cancer cells [28, 35]. Likewise, stromal influences became evident in experiments that observed local tissue aggravation suffices to enable tumor development. In these studies, mammary fat pads were pre-cleared from epithelial components (i.e., leaving only the stroma) and subjected to ionizing radiation. After radiation, treated fat-pads were repopulated with un-irradiated non-tumorigenic epithelial cells, which consequently developed tumors. Results revealed that radiation-activated stroma promotes tumor formation in otherwise non-tumorigenic cells [36]. Additionally, Maffini et al., conducted an experiment involving isolated stroma and epithelial cells from rat mammary glands treated with a known carcinogen. The cells and/or stroma were separately treated with carcinogen/irritant, and then recombined to determine which of the treated cellular components would induce tumorigenesis. From this study, researchers found that neoplasms only developed in animals whose stroma was exposed to the treatment. Maffini et al., observed that animals whose stroma was not exposed to treatment (i.e., stroma exposed to vehicle only) did not develop neoplasms, regardless of whether the epithelial cells were exposed to the carcinogen [37]. These and similar studies went on to suggest that neoplastic development and progression is restricted by normal/naive stroma, while it is tolerated and often enabled by an altered microenvironment (i.e., tumor permissive) that includes both activated stromal cells and a remodeled ECM, akin to chronic wounds [37–44].

These types of observations are addressed in TOFT, which suggests that the healthy stroma imposes constraints on cells and that loss of these environmental constraints (i.e., a result of chronic stromal alterations, remodeling or damage), enable cancer onset and progression [30]. Evidence for this was reported in 1910 by Rous who suggested that the tumor problem is a tissue predicament [45]. When considering normal embryonic development, TOFT aligns with the current view that the role of the mesenchymal stroma, represented by fibroblastic cells, is to impart a tightly controlled environment upon epithelial cells, and that the stroma becomes altered during (and/or to initiate) tumorigenesis. In other words, the “normal” stroma balances tissue growth/regeneration by, for example, permitting cell proliferation solely during temporary assaults (i.e., acute wounds), and restoring restraint of cell growth following wound resolution [46–49]. However, instances such as chronic inflammation and cancer, can impede restoration of an “innate” homeostatic state and thus perpetuate regenerative cues of unrestraint cell growth [44, 50]. When this occurs, stromal components collectively evolve to establish a “new” equilibrium that consequently maintains a constitutively, pathological tissue regeneration condition that sustains (or enables) tumorigenesis [43, 51–53]. Another principle of TOFT, supported in Rous’ early observations [45], states that by “normalizing” stromal components (e.g. restoring the initial homeostatic tissue balance) cancer can be restrained and even reversed [33]. Several studies have demonstrated the involvement of epithelial cancer cells and normal stromal cells (i.e., local mesenchymal/fibroblastic cells) in influencing cancer development and progression. For example, Mintz and colleagues demonstrated that teratoma/carcinoma cells can be reprogrammed to give rise to normal adult tissue when injected into intact blastocysts. Follow-up studies from this group also demonstrated that the initial tumor cells were able to contribute to different cell lineages with normal functioning progeny, despite their oncogenic driver mutations and their ability to grow tumors when inoculated into in vivo permissive environments [25, 54]. These findings were later supported by McCullough et al., who observed tumorigenic liver cells differentiate into normal hepatocytes when injected into the liver (i.e., stroma) of healthy rats [55]. Similar results were achieved when a carcinogenic transformed epithelial cell-line was injected into the stroma of mammary glands and was observed to participate in normal development of the mammary gland. This study again highlighted the fact that despite the tumorigenic properties of the epithelial cells, normal stroma can guide tumor cells to not only participate in normal development, but also to produce progeny that contributes to typical/physiological tissue growth [56]. These studies suggested that the plasticity of malignant cells enables “reversal to normalcy” such that cells can potentially differentiate in a non-pathological manner and incorporate into healthy, functional tissues/organs [54]. These studies demonstrated that unaltered natural stromal is restrictive of tumor growth while alterations in stromal cells, such as activation of local fibroblastic cells (i.e., generating CAFs), and chronic concomitant remodeling of CAF generated interstitial ECM [57, 58], alter the microenvironment to endorse or to at least tolerate tumor development and progression. These studies suggest that the normal tissue architecture, often regulated by local fibroblastic cells, can reinstate regulatory controls in epithelial/solid cancer cells [59–64].

SMT and TOFT: Mutually inclusive tumorigenic theories

At first glance, TOFT appears to address centuries of experimental and clinical observations most accurately regarding tumorigenesis in comparison to SMT. However, we interpret the underlying concepts of both theories to be mutually inclusive in that both “shed light” on diverse aspects of tumorigenic processes (Figure 1). Together, both theories provide strong rationale that explains the complexities of tumor biology in that mutations in epithelial cancer cells and the stromal microenvironment are actively involved in driving all aspects of epithelial tumorigenesis. For example, early experimental studies demonstrated that tissue formation during development, such as in mammary gland development, is guided by the mesenchyme [65]. Likewise, the rate of tumor foci formation is heavily dependent on mesenchymal stromal cells (i.e. naïve fibroblasts vs. CAFs) and their concurrent ECM [66, 67]. Reciprocally, the stroma is also regulated by changes in cellular programs, which can be instigated by accompanying mutated/malignant epithelial cells and/or injury [43, 65]. Uncontrolled cell proliferation is often accompanied by a loss of normal tissue architecture [68, 69]. In fact, both cancer cells and stroma cells secrete matrix modulating proteins, such as matrix metalloproteinases (MMPs) and tissue serine proteases, which are known to degrade the basement membrane and remodel the interstitial (i.e., fibroblastic) ECM, thereby altering ECM topography [70, 71]. Dysregulation of ECM turnover is believed to actively promote cancer onset, which can both stimulate mutagenesis (evidence for SMT) and/or provide growth advantage to cells with mutations (in support of TOFT). This was demonstrated in studies in which matrix degrading enzyme, stromelysin-1 induction was shown to activate the stroma such that the reactive stroma induced tumorigenesis [72–75]. Importantly, while cancer cells produce some ECM, the majority of ECM detected in the tumor milieu is generated by CAFs [76]. Thus, a comprehensive (i.e., mechanistic) understanding of CAF function and regulation, including de novo ECM production and remodeling, is necessary to achieve clinically relevant interventions aimed at restoring the anti-tumor regulation of the stroma [77]. Notably, concepts presented in SMT and TOFT regarding tumorigenesis should be strongly considered to aid our understanding of cancer dynamics as well as in the design of effective therapies. As such, this review focuses on the role of the stromal cells; particularly on the reciprocity between activated and wound healing-like fibroblasts (i.e., CAFs) and their self-generated ECMs in promoting and/or sustaining tumor supportive programs (Figure 2).

Figure 1: SMT and TOFT are mutually inclusive.

The Somatic Mutation Theory (SMT) premise is that mutations in tumor suppressor and/or oncogenes in cancer cells are the drivers of tumorigenic programs (cancer hallmarks), such as uncontrolled proliferation and cell invasion. The Tissue Organization Field Theory (TOFT) posits that the stroma dictates tumorigenesis regardless of cell mutations. TOFT explains that breakdowns or changes in stromal constraints, whether due to damage and/or pro-tumor stroma remodeling, enable tumorigenic programs to develop. However, many studies have demonstrated that cancer hallmarks can be dynamically and reciprocally driven by principles underlying both SMT and TOFT. The example provided intends to depict an early low grade pancreatic intraepithelial neoplasm (PanIN) lesion that indicates potential for cancer development. It “sheds light” on altered/mutated ductal epithelial cells with typical elongated cell bodies (cyan) and sustained nuclear polarity (yellow) highlighting SMT, as well as on activated fibroblasts (CAFs) with aligned CAF-generated ECM, depicted in the stroma (light magenta), to highlight TOFT. Illustration by P. Ovseiovich, Mexico, DF.

Figure 2: Mechanisms and functions by which CAFs and desmoplastic ECM regulate tumorigenesis.

This illustration highlights the fact that stromal dynamic reciprocity, between CAFs and ECM, controls CAF functions through mechanisms that alter numerous aspects of tumorigenesis such as tumor immunity, systemic inflammation, angiogenesis, metastasis, tumor growth and local neoneurogenesis (i.e., de novo microenvironmental innervation). The illustration includes a partial list of CAF functions as well as some of the mechanism by which these functions are achieved. The illustration highlights the fact that ECM regulated CAF function results in: secretion of distinct key factors, alteration of the metabolic state of the tumor, and the generation of a remodeled interstitial ECM, which together are active participants of all aspects of tumorigenesis. Note that this illustration was inspired by a figure displayed in Sahai et. al., Nat Rev Cancer 2020 [77]. Illustration by P. Ovseiovich, Mexico, DF.

Stromal Dynamic Reciprocity of Cancer

Maintenance of a “normal” tissue homeostatic equilibrium relies on the dynamic and reciprocal influences of stromal elements that collectively dictate tissue structure and physiology. Stromal dynamic reciprocity (SDR) of cancer highlights the biophysical and biochemical interactome amid stromal cells and extrinsic components; primarily between CAFs and the local ECM [38–40, 78]. These interactions are not restricted solely to tumor biology, albeit this dynamic reciprocity persists during tumorigenesis [78]. In normal development, fibroblasts are known to regulate tumor restrictive programs by reinstating balance between intrinsic and extrinsic occurrences in response to growth demands or temporary assaults [43, 44]. Thus, in instances of chronic inflammation like in cancer, in which a new homeostatic balance is achieved, it is not surprising to observe that activated fibroblast, such as CAFs, and their altered ECM, regulate disease programs (Figure 2) [77]. However, in cancer, the roles of CAFs is quite paradoxical in that CAFs present with multiple functions that both support and restrict tumor development and progression [77, 79]. Pancreatic ductal adenocarcinoma (PDAC) is one of the best examples to discuss this conundrum due to the complexity and notably vast PDAC stroma (Figure 3) [80–84].

Figure 3: Human PDAC; “gating” on desmoplasia.

Simultaneous multiplex immunofluorescence of human PDAC surgical sample. The large left circle includes 3, of the 7 colors used to indicate: epithelial/tumor areas in cyan, stromal/desmoplasia areas in magenta, and cellular nuclei in yellow. The desmoplastic area, generated by the SMIA-CUKIE 2.1.0 algorithm, is shown in the white middle circle and corresponds to the magenta area. This area was used by the software to “gate” on desmoplastic/stromal positive pixels while excluding all other areas. Using desmoplastic “gating” the last 4 circles show 4 biomarkers that together depict a pro-tumoral stroma signature that is regulated by PDAC associated CAFs and ECM (figure relates to study by Franco-Barraza et. al. 2017 [84]). The 4 stromal biomarkers correspond to: the activated conformation of α5β1-integrin (green), 3D-adhesions (red) enriched in activated focal adhesion kinase (i.e., pFAK in orange), and activated SMAD (blue; indicative of canonical TGFβ1 signaling. This stromal signature was correlated with short overall survival in PDAC as well as in renal cell carcinoma patients.

Pancreatic Ductal Adenocarcinoma and its Associated Fibroblasts

PDAC will soon be the 2^nd leading cause of cancer related deaths in the United States [2]. There are several defining characteristics of PDAC. These include evidence of progressive precancerous lesions with oncogenic mutations [85] that are accompanied by significant local stromal alterations [86, 87] (Figures 1 and 3). PDAC is also recognized by its early cancer cell dissemination [88], substantial metastatic events despite locally collapsed nonfunctional blood vessels [89, 90], immune evasion [91, 92], reprogrammed cellular metabolism [93], and a notably expansion of a fibrotic-like stroma reaction, known as desmoplasia [94, 95]. Desmoplasia constitutes the bulk of the tumor mass (Figure 3) and is primarily driven by the SDR events between CAFs and CAF-generated ECM [76, 96, 97].

Numerous studies have demonstrated a symbiotic relationship between PDAC cells and its associated stroma (i.e., PDAC’s desmoplasia). For example, conditioned media from pancreatic CAFs were shown to desensitize PDAC cells to radiation and chemotherapy while also stimulating cancer cell growth, proliferation, migration, and invasion [98]. Moreover, co-injection of cancer cells and CAFs in immunocompromised mice not only increased tumor growth rates, but phenocopied some key human PDAC pathologies [99]. While normal stroma is clearly tumor restrictive, these studies highlight the significance of PDAC associated desmoplastic stroma in promoting tumorigenesis.

To this end, the level of pro-tumor stromal activation was proposed, by more than one study, to serve as a prognostic marker for PDAC disease recurrence [84, 100]. Figure 3 highlights a recent example whereby four CAF biomarkers, regulated by the CAF-generated ECM, constitute a stromal signature indicative of patient outcomes [84]. Another study defined the levels of stroma activity as an activated “stromal index” in which patients were stratified based on the ratio of alpha smooth muscle (αSMA) positive fibroblast (i.e., myofibroblast) to the amount of fibrillar collagen deposition. Accordingly, a high myofibroblastic ratio was found to correlate with decreased patient survival whereas high collagen deposition, in the absence of αSMA, significantly correlated with enhanced overall patient survival [100]. In addition to the above-listed studies, some investigators have highlighted the role of fibroblast activation protein (FAP) positive CAFs in pro-tumoral function [101], while others demonstrated that CAF elimination enhances PDAC progression [102, 103]. Together, these types of studies provide the possible notion of desmoplasia playing a dual role in PDAC progression. In other words, these studies emphasize that the role of the tumor microenvironment can enable both tumor-permissive and tumor-restrictive functions. Thus, untangling the mechanisms regulating pro and anti-tumor programs amid the desmoplastic stroma, may provide insight into harnessing the anti-tumor functions for more effective therapeutic intervention. Ultimately, this positions the SDR between CAFs and the ECM, which is needed to sustain CAF functions [84, 100], as an important constituent of the biology of human PDAC.

Cancer Associated Fibroblast: Major Stromal Architects

The term “CAF” is generally used to describe the activated (i.e., no longer quiescent) fibroblastic cell population that accompanies solid epithelial tumors [77]. CAFs are stromal fibroblastic cells that undergo phenotypic and functional changes and to regulate a plethora of tumorigenic processes (Figure 2) [77]. Fibroblastic activation, akin to wound healing, can occur in response to a disturbance that temporarily alters the resting homeostatic equilibrium [43]. In this case, the main function of an activated fibroblast is to build and remodel the ECM and thus heal the temporary wound. Yet, if the assault prevails, like in the case of cancer, fibroblastic activation (i.e., in the form of CAFs) is sustained along with the ECM remodeling process [43]. An interesting occurrence in desmoplasia, is that naïve/quiescent fibroblastic cells can be activated in response to the ECM deposited by CAFs [104]. In fact, differences amid ECM components have been shown to provide CAFs with distinct phenotypes [105]. Of interests, alterations to ECM that result in fibroblastic activation, can also include physical changes to the ECM, like in the case of increased collagen crosslinking, which in turn could control metastatic events [57]. Further, fibroblastic cell activation resulting in altered ECM deposition, can also be induced by secreted factors [44, 77, 106]. These processes are often controlled by cytoskeletal rearrangements, mediated by cytoskeletal regulatory proteins, which modify morphology and function of fibroblastic cells [69]. One such case is evident when studying the nuclear protein myocardin-related transcription factor A (MRTF-A), which is retained in the cytoplasm by binding to globular actin (G-actin). Upon mechanical stretching of the cell, MRTF-A dissociates from G-actin and translocates to the nucleus [107]. Under low G-actin or during biomechanical stretching conditions, the accumulation of nuclear MRTF-A triggers the expression of cytoskeletal-organizing proteins, like α-SMA, providing the cell with the ability to contract [108]. The newly acquired contractile forces, if sustained like in the case of CAFs, tend to enable the production of an altered/remodeled ECM [109]. The discrete ECM production is accompanied by the deposition of large amounts of collagen I and particular fibronectin splice variants (i.e., EDA fibronectin), amid others [44, 47, 94, 110, 111].

Mutations in SMAD4 are common occurrences in PDAC and result in the desensitization of cancer cells to secreted TGFβ, thereby directing this factor’s effects to stromal cells. Again, this example highlights that both SMT and TOFT are needed to elucidate tumor stroma interactions. Of note, TGFβ1 is secreted in a latent form and it is stored, inactive, within the local ECM. TGFβ1 can be released from its latent complex by fibroblastic (myofibroblastic) cell contraction [112], integrin-dependent mechanisms [106], and more [113]. Once processed (i.e., released), TGFβ1 functions locally and engages its receptor, expressed in stromal fibroblastic cells (i.e., CAFs), to continuously promote fibroblastic activation. The resultant CAFs increase ECM production and together with other cell local factors remodel the ECM, via proteases (i.e., MMPs), matrix cross-linkers such as lysyl oxidase, and/or other proteins such as osteonectin (also known as SPARC), periostin, and hyaluronic acid; all of which are essential for the production of the interstitial ECM that accompanies tumors. Ultimately, these processes establish a vicious cycle of desmoplastic expansion believed to physically and/or biochemically hinder therapeutic intervention [44, 47, 77, 78, 114].

Emphasis on how CAF Function is Mediated by Cytoskeletal Rearrangements

As hinted above, ECM regulated cytoskeletal (and other) changes provide more than just structural support to cells [115, 116]. The cytoskeleton is dynamically regulated to facilitate cell polarity/organization, migration, and division [115, 117, 118]. Together with the main ECM receptors, integrins, the cytoskeleton serves as a mechano-transductor and signaling scaffold that bi-directionally transmits intracellular and extracellular cues [69, 118, 119]. As such, abnormalities in cytoskeletal organization and regulation are associated with diseases including cancer [120–122]. For example, regulated by cytoskeletal changes, alterations in nuclear architecture, a known pathological characteristic of cancer, are attributed to changes in lamin A/C isoform ratios [123–126].

Cytoskeletal organizing proteins are commonly associated with disease progression in many cancers [127]. Interestingly, during myofibroblastic activation, CAFs are known to upregulate cytoskeletal proteins such as α-SMA, palladin, and others. Palladin is a scaffolding protein that is associated with disease progression and unfavorable patient outcomes [128]; especially among epithelial cancers rich in desmoplastic stromal interactions such as PDAC [86, 129–132]. For example, palladin is upregulated in early precancerous lesions and its expression increases during malignant transformation [129, 130, 133–135]. Further, desmoplastic ECM production is facilitated by cytoskeletal rearrangements mediated in part by palladin [110, 132, 136–140]. Moreover, palladin was identified as an independent prognostic marker in PDAC and serves as a potential prognostic biomarker following chemo radiation [132]. Thus, unveiling the role of palladin in pro-tumor stroma dynamics is necessary to therapeutically harness the tumor restrictive stromal properties. Importantly, palladin overexpression is robust in PDAC stroma, in comparison to PDAC cells, and is a conserved feature of desmoplasia [131]. Hence, stromal palladin may play a role in PDAC development, as well as other cancers.

In addition to palladin, α-SMA is often found to be highly upregulated in malignant diseases [141–144]. α-SMA is one of six actin isoforms that enable cells to appropriately respond and adapt to the dynamic microenvironment [145–147]. While in some instances, α-SMA expression is a mere consequence of tumor progression, α-SMA is also upregulated during precancerous conditions, such as inflammation [148]. Although α-SMA expression is known to be regulated by TGFβ1 (i.e., in a smad3 dependent manner), the mechanism by which α-SMA expression contributes to tumorigenesis, in both tumor and stromal cells, remains elusive. In fact, specifically targeting α-SMA expressing cells in PDAC resulted in accelerated progression of this disease [103]. This suggests that α-SMA positive CAFs may also restrain tumor growth and metastasis. Nevertheless, CAF contractile roles are central in generating the earlier noted pro-tumoral anisotropic topography of the desmoplastic ECM.

In PDAC, the desmoplastic ECM includes high volumes of the known aligned “tumor associated collagen signatures 2 and 3” (TACs 2 & 3) [149–151], which positively correlate with poor patient outcomes [109, 152, 153]. ECMs enriched in TACs 2 & 3 are produced by contractile CAFs (i.e., α-SMA positive CAFs), but not by their normal fibroblastic cell counterparts [104]. Of interest, the ECM has been referred to as a fundamental mediator of cell function [116]. It is therefore not surprising that changes in ECM can affect chromatin and gene expression [154]. In fact, ECM dynamics participate in all the hallmarks of cancer [155]. Multiple studies have characterized cytoskeletal regulating Rho GTPases working synergistically through feedback loops and mutual activation/inhibition to control the (fibroblastic) cell morphodynamics [156, 157]. Then again, fibroblastic proteins that regulate the function of Rho GTPases during CAF contraction have been proposed to enable changes in clinically relevant ECM topography (i.e., TACs 2 & 3) [153, 158, 159]. In fact, Rho GTPase regulators, like GEF-H1, were shown by the late Dr. Keely to mediate Rho activation in response to matrix stiffness [160, 161]. Also, when fibroblastic cells contract within a fibrillar ECM, the motility of immune cells (i.e., macrophages) is modified [162]. Thus, CAF contractile functions that alter the ECM are of paramount importance for desmoplastic activation and desmoplastic field expansion. This idea is fully supported by the above mentioned SDR concept, suggesting a central role for actin regulated fibroblastic contraction and ECM interactions in tumor onset and progression [38, 78, 109, 153, 161]. For SDR to occur and be maintained, ECM production and remodeling, as well as the contractile function of CAFs, are essential.

Functional Roles of CAFs in Cancer

One of the best ways to describe CAFs is using an analogy to the sustained force-dependent contractile granulose reaction (i.e., contractile connective tissue) that is evident in chronic wounds [43]. Using this analogy, CAFs present with several functions (Figure 2) [77]. The best known CAF role is described by contractile myofibroblastic CAFs, known to be induced by TGFβ1 (often through cytoskeletal alteration) to produce the typically described pro-metastatic ECM [57, 77, 153, 161]. It is therefore not surprising that investigators recently found that particular posttranslational modifications, in key ECM proteins, profoundly alter integrin clustering to improve stromal cell invasion in wounds [163]. CAFs are also recognized as microenvironmental cells that provide metabolic support to cancer cells (i.e., promoting tumor cell growth under nutritional stress) [93, 164] as well as cells that regulate angiogenesis through secretion of vascular endothelial growth factor [165]. CAFs are also immuno-modulatory cells with both immunosuppressive and immunogenic functions [81, 82, 166, 167]. In fact, inflammatory CAFs, simulating the granulose-like reaction that includes production and secretion of cytokines, such as IL-8, IL-6, IL-1, CXCL12, and more, have also been described [77, 81, 167–169].

Investigators often seek to classify fibroblastic cell populations [92] as either anti-tumor, similar to local naïve fibroblastic cells such as pancreatic stellate cells (PSCs), or pro-tumor, such as FAP expressing CAF [81, 101–103]. Yet, this binary characterization remains controversial [77, 79, 101–103, 170, 171]. Due to the dichotomy in pro- and anti-tumor functions as well as the various phenotypic and functional/mechanistic characteristics of CAFs, the notion of heterogeneous CAF populations has become an important focus for epithelial tumors that incorporate substantial stromal remodeling (i.e., desmoplasia) such as in breast, lung, and PDAC [82, 172–177]. Of note, despite being identified as distinct molecular signatures, differentially functional CAFs (such as myofibroblastic and inflammatory CAFs) describe fibroblastic cell functions (Figure 2) that are dynamically interchangeable [77, 81–83]. Then again, regardless of this functional heterogeneity, it seems that CAFs arise from anti-tumor fibroblastic (local or recruited) cell populations, for which cell lineages are only beginning to be traced [92, 169].

Hence, it is possible that the cytoskeletal dependent SDR, between fibroblastic cells and ECMs, is responsible for the local dynamic regulation of the CAF function that affect tumorigenesis (Figure 2), which encompass both the SMT and TOFT (Figure 1), and support all the recognized Hallmarks of Cancer (Figure 4).

Figure 4: Dynamic reciprocity between CAFs and ECM partakes in Cancer Hallmarks.

Illustration intended to highlight the fact that the dynamic reciprocity between CAFs and ECM results in the active regulation of all known Cancer Hallmarks. Note that this illustration was inspired by a figure displayed in Hanahan and Weinberg Cell 2011 [3]. Illustration by P. Ovseiovich, Mexico, DF.

Conclusion:

It is well understood that cancer is not uniform across patient cohorts. Instead, cancer is found to display tremendous heterogeneity, including both cancer and stromal cells, even within a single tumor. With this notion arose the Hallmarks of Cancer, to which treatments were designed to disrupt specific mechanisms that contributed to one or more trademark recognized to aid in tumor formation, growth, and progression (Figure 4). However, these therapeutic “magic bullets” have only proved to work transiently where resistance eventually evades the treatment and the disease continues to progress [178]. Thus, understanding and monitoring the long-term effects associated with targeting cancer hallmarks are crucial in effectively combating cancer progression as well as in preventing its development. Accordingly, a comprehensive understanding of SDR in general, and of the role of CAFs and ECM, may be key to establishing novel and more effective therapeutic efforts. From an evolutionarily standpoint, the body is physiologically conditioned to address and resolve insults that prompt temporary deviations from innate tissue functions through both local and systemic repair processes. These temporary occurrences are typically resolved whereby the acutely altered SDR is reverted to an innate homeostatic SDR equilibrium. It is plausible that this innate system is exploited during tumorigenesis thereby shifting the SDR balance towards a new equilibrium that supports tumorigenesis, such as desmoplastic SDR [78]. The desmoplastic SDR, between CAFs and ECM, then maintains the vicious cycle of tumorigenesis. While this pathological adaptation favors the aggressive nature of the disease state, ablation of compensatory elements in the SDR process impairs the possibility of restoring an innate tumor-restrictive stromal status. Thus, a thorough delineation of SDR mechanisms (i.e., amid CAFs and ECM), especially those that influence fibroblastic cell functions (Figure 2), may be vital to dissolving the regulatory features of cancer hallmarks (Figure 4), as well as consolidating SMT and TOFT (Figure 1), to eventually generate research that improves the development of effective and long lasting therapies.

Highlights:

• When considering the stromal dynamic reciprocity between cancer associated fibroblasts and the extracellular matrix, two seemingly opposing cancer theories are mutually inclusive.

• Stromal dynamic reciprocity is imperative to sustain cancer associated fibroblast functions during tumorigenesis.

• Functions of cancer associated fibroblasts influence all known Cancer Hallmarks.