Elucidating the link between collagen and pancreatic cancer: what’s next?

KEYWORDS:

• Epithelial-mesenchymal transition
• extracellular matrix
• integrin
• Kindlin
• pancreatic stellate cells

1. Introduction

A dense fibrotic stroma, so-called desmoplasia, is a characteristic feature of pancreatic cancer tissue [1]. The amount of stroma sometimes surpasses that of the cancer cells, resulting in massive fibrosis surrounding the cancer cells. Pancreatic stellate cells (PSCs), which play critical roles in pancreatic fibrosis, largely contribute to the development of the desmoplasia [1]. Pancreatic cancer is notorious as one of the most malignant neoplasms with an extremely poor prognosis. In addition to the vicious nature of cancer cells, the above-mentioned tissue structure also contributes to the intractability of pancreatic cancer. Desmoplasia itself is a result of perpetuated inflammation and cell-to-cell interactions during pancreatic carcinogenesis [1]. The interplay between cancer cells and stromal cells via cytokine production, hypoxia, oxidative stress, and so on leads to the establishment of a specific microenvironment that protects the cancer cells. Platelet-derived growth factor (PDGF) and transforming growth factor-β (TGF-β) are typical mediators of this interaction [2]. This tissue structure, especially three-dimensional collagen, has a direct effect on the chemoresistance of cancer cells. Pancreatic cancer cells cultured in three-dimensional collagen show increased extracellular signal-regulated kinase (ERK) signaling, leading to decreased sensitivity to gemcitabine [3]. Furthermore, stimuli from the stroma also promote the malignant behavior of cancer cells, such as increased motility or apoptosis resistance [1]. These stimuli from stroma include inflammatory cytokines, growth factors and the specific microenvironment formed by the extracellular matrix (ECM) deposition such as hypoxia [4].

Multiple molecules and cell types are involved in this complex process, and the discovery of novel mechanisms of interaction is still ongoing. Among these issues, this editorial article focuses on the role of collagen in pancreatic cancer. Collagens are proteins abundant in connective tissue and have numerous family members [5]. In a normal state, the tissue-specific, organized expression of collagens is essential for epithelial cell attachment, polarity and maintenance of the normal organ function. For example, collagen type IV is a major component of basement membrane, required for epithelial structure maintenance [5]. In the case of fibrosis, excess deposition of ECM proteins including collagens is a characteristic change. From this point of view, collagens have indispensable roles during pancreatic carcinogenesis, which involves destruction of the normal tissue structure, survival in an inflamed microenvironment and invasion/metastasis. A previous study identified pathological collagen deposition in pancreatic cancer tissue [6] and, furthermore, the cancer-promoting roles of collagens. The aberrant activation of collagen-related signaling molecules such as particularly interesting new cysteine-histidine-rich protein (PINCH) or CD151 has also been reported, and some of them correlate with the patient prognosis [7,8].

2. Pancreatic cancer cell function and collagen

The impact of excess deposition of collagen in pancreatic cancer has been studied [9]. Cell culture on collagen type I- or type III-coated dishes cause E-cadherin reduction in pancreatic cancer cell lines, leading to less cell-to-cell contact [10]. This E-cadherin repression is attenuated by treatment with Src kinase pyrazolo pyrimidine-type inhibitor (PP1) and herbimycin A, suggesting contact with collagens affects the intracellular signals in cancer cells. Another study identified that attachment to ECM proteins of the cancer cells causes growth promotion and chemoresistance [11]. Collagen type I increased the cell proliferation of the pancreatic cancer cell line AsPC-1 and attenuated apoptosis induced by 5-fluorouracil [12]. Pancreatic cancer cells also produce collagen for proliferation and migration. Cancer cell-derived collagen type IV stimulates cell growth and migration, which was confirmed by RNA interference-mediated knockdown [13]. These in vitro experiments suggest a protective role of ECM proteins including collagen for pancreatic cancer cells.

Mediators of the cancer-protective effects from ECM have since been identified. For example, collagen type I-mediated E-cadherin downregulation turned out to be regulated by Smad-interacting protein 1 (SIP1) [14]. SIP is a direct transcriptional repressor of E-cadherin promoter that specifically binds to enhancer box (E-box), a regulatory sequence. Pancreatic cancer cells increase Snail expression, which is another inducer of cancer invasiveness, in response to contact with collagen type I [15]. This induction of Snail, a master regulator of epithelial-to-mesenchymal transition (EMT), is mediated by the TGF-β/Smad3 and Smad4 axis. These intracellular signals are also activated in PSCs by inflammatory stimuli or various types of extracellular stress. As well as the feedforward loop of PSC activation, cell-to-cell interaction between cancer cells and PSCs is an alternative driver of the continuing inflammation in pancreatic cancer tissue. Together with inflammation in the microenvironment, ECM proteins including collagens play pivotal roles during pancreatic carcinogenesis.

3. Collagen-related signaling pathway

Pancreatic cancer cells express a wide variety of cell-surface molecules to interact with ECM proteins. Among them, integrins are transmembrane glycoproteins with numerous family members. Heterodimers of α- and β-integrins recognize ECM proteins, including collagens [16]. The expression of these integrins in pancreatic cancer cells was described in the 1990s [17,18]. Antibody-based inhibition of integrins attenuated tumor-cell adhesion to collagen [18]. The activation of intracellular signals, such as Ras-induced Raf/MEK/ERK, alters the pattern of integrin expression, leading to modified cell adhesion and migration [19]. The inflammatory cytokine interleukin-1 upregulates α6β1 integrin, leading to ERK activation and increased proliferation and migration [20]. Interaction between specific integrin and collagen also has cancer-promoting roles. For example, the α2β1 integrin is highly expressed in human pancreatic cancer cell lines and mediates collagen type I-induced cellular proliferation and migration [21].

Additional regulators of integrin signals also affect cellular functions. Integrins regulate cytokinesis, a final step in cell division, and the activation of p90 ribosomal S6 kinase signal activation in addition to the ERK activation by integrin signal is essential for this system [22]. The migration-promoting role of αvβ6 integrin in pancreatic cancer cells depends on the existence of epidermal growth factor receptor pathway substrate 8, which is involved in remodeling of the cytoskeleton [23]. Kindlin-2, an intracellular regulator of integrin activation, is reported to be a TGF-β target gene in pancreatic cancer cells. This induction promotes cell proliferation and migration, and higher expression of Kindlin-2 in cancer cells correlates with a poor prognosis [24]. The following study performed knockdown of Kindlin-2 expression in PSCs, which attenuated the cancer-supporting roles in a subcutaneous implantation model using immune-deficient mice [25]. Stromal expression of Kindlin-2 is also associated with shorter recurrence-free survival, indicating a cancer stem-cell supporting mechanism mediated by the integrin/Kindlin-2 axis. The intracellular signals in PSCs that regulate these supportive roles should be clarified thoroughly.

Compared to integrin signal, the role of other receptors for collagens have not yet been fully reported. Discoidin domain receptor 1 (DDR1), which is a receptor tyrosine kinase activated by collagen binding, promotes tumorigenesis through the activation of protein tyrosine kinase 2 and pseudopodium-enriched atypical kinase1 in pancreatic cancer [26]. Pharmacological inhibition of this signaling pathway repressed cancer promotion and chemoresistance, which could represent a novel therapeutic strategy. The ATP-competitive DDR1 kinase inhibitor 7rh enabled the specific inhibition. This orally available small-molecule inhibitor efficiently repressed tumor growth and sensitized cancer cells to chemotherapy in an orthotopic implantation model, suggesting its possibility for clinical application [26]. Little is known about other collagen receptors in pancreatic cancer, and further study is necessary to clarify the roles of complex collagen-related signals. A comprehensive approach might be necessary to address this issue.

4. Expert commentary

The cancer promoting roles of PSCs have been studied since their discovery. The main ECM protein, collagen, from PSCs also enhances the malignant phenotype of cancer cells via the activation of multiple signaling pathways [1]. Activations of PSCs by cancer cells and vice versa, are mediated by soluble factors, mechanistic and environmental changes that are perpetuated from the primary fibrosis [1]. Inflammatory cytokines and growth factors are typical examples of soluble factors such as PDGF and TGF-β [2]. Hypoxia and increased interstitial pressure are examples of mechanistic and environmental changes [4,27]. However, simple removal of PSCs paradoxically promoted pancreatic cancer progression in mouse models, suggesting cancer-inhibiting roles of PSCs [28,29]. These studies indicated the importance of accurate targeting on cancer cell–stellate cell interactions. This point is a key weakness for stroma-targeting therapy.

Dissection of truly cancer-promoting mechanisms in the cancer cell–stromal cell interaction is still on its way, but a possible strategy has been found. Administration of a vitamin D receptor ligand, calcipotriol, improved gemcitabine delivery into the tumor, leading to prolonged survival of a pancreatic cancer model mouse [30]. This finding indicates reprogramming, rather than removal, of activated PSCs could be a therapeutic approach. Downstream targets of collagen-related signals might be responsible for these effects, which have a certain influence on the pancreatic cancer cell function and clinical outcomes. From this point of view, integrin signal or its intracellular mediator, Kindlin-2 could be attractive targets for novel therapies. However, global knockout of β1-integrin causes embryonic lethality [31]. Since β1-integrin also has indispensable roles in the maintenance of normal exocrine pancreatic functions [32], simple inhibition could result in unexpected side effect. Therefore, careful dissection of these mechanisms is necessary for clinical application. These lines of evidence also highlight the importance of targeting other collagen-related signals such as DDR1. The availability of specific inhibitors, such as rh7 [26], would provide an additional advantage.

Despite its multifaceted roles, the tumor stroma still remains an attractive target in pancreatic cancer. A recent report identified another role of collagen as a reservoir of proline, rescuing energy depletion [33]. Understanding such novel mechanisms could enable the identification of alternative therapeutic targets. The formation of pancreatic cancer is the result of continuous interactions between cancer cells and host cells, and a simple solution will not be obtained easily. The combination of accumulating knowledge and incorporation of new delivery technologies will help to establish radical therapies for pancreatic cancer, one of the last standing intractable cancers.

5. Five-year view

The delivery of a therapeutic intervention is still a problem in pancreatic cancer. The application of new delivery technologies such as nanoparticles, host-derived cell modification and mimics of cell-to-cell communication mechanisms will need to be achieved. As therapeutic targets, a combination of truly cancer-promoting functions of the stroma will be selected. Animal models recapitulating human pancreatic cancer will be used to validate novel therapies and avoid unexpected cancer-promoting results. The influence of perpetual interactions between cancer cells and stromal cells/ECM might be used for the early detection.

Declaration of interest

The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was supported in part by JSPS KAKENHI (26293171, 17K09452, 15H04804), the Mitsui Life Social Welfare Foundation to Masamune A, and the Smoking Research Foundation to Masamune A.