The epidermal integrin-mediated secretome regulates the skin microenvironment during tumorigenesis and repair

Abstract

Integrins are cellular transmembrane receptors that physically connect the cytoskeleton with the extracellular matrix. As such, they are positioned to mediate cellular responses to microenvironmental cues. Importantly, integrins also regulate their own microenvironment through secreted factors, also known as the integrin-mediated secretome. Epidermal integrins, or integrins expressed by keratinocytes of the skin epidermis, regulate the cutaneous microenvironment through the contribution of matrix components, via proteolytic matrix remodeling, or by mediating factors like cytokines and growth factors that can promote support for nearby but distinct cells of the stroma, such as immune cells, endothelial cells, and fibroblasts. This role for integrins is enhanced during both pathological and repair tissue remodeling processes, such as tumor growth and progression and wound healing. This review will discuss examples of how the epithelial integrin-mediated secretome can regulate the tissue microenvironment. Although different epithelial integrins in various contexts will be explored, emphasis will be given to epidermal integrins that regulate the secretome during wound healing and cutaneous tumor progression. Epidermal integrin α3β1 is of particular focus as well, since this integrin has been revealed as a key regulator of the keratinocyte secretome.

1.0. INTEGRINS: A FAMILY OF CELL ADHESION RECEPTORS

Integrins are a major family of cellular receptors for extracellular matrix (ECM) adhesion [1]. Integrins are obligate heterodimers comprised of an α and a β subunit, each having an extracellular domain, a single-pass transmembrane domain, and a cytoplasmic domain. Monomers consist of 18 α subunits and 8 β subunits that come together in limited combinations to generate 24 distinct integrins with varying ligand-binding specificities (Fig. 1). The extracellular domain of integrins bind ligand, while the cytoplasmic domain simultaneously interacts with cytoskeletal proteins [1]. In this way, a physical linkage between the ECM and the cytoskeleton is forged, which is critical for regulating cell shape and numerous other cellular processes such as adhesion, polarization, and motility [1–5]. Additionally, integrins can interact with various signaling effectors either directly or indirectly, enabling integrins to function as bidirectional signal transducers [1, 2, 6, 7]. Such integrin-guided signal transduction controls critical cellular functions including proliferation, migration, ECM remodeling, survival, and gene expression [1, 3, 6] that are important for both normal and pathological processes.

Figure 1. The integrin family of adhesion receptors and their ligand-binding specificities.

There are 18 integrin α subunits and eight integrin β subunits that can dimerize in limited combinations to form 24 integrins with different ligand-binding specificities. The original classification of integrins is shown [1]. Note that keratinocyte integrin α9β1 binds to several ligands of the skin ECM including tenascin, fibronectin, and osteopontin, although it is classically categorized as a leukocyte receptor due to its expression by leukocytes [8]. This figure was adapted from earlier work [1] and created using BioRender.

Integrins are expressed on both epithelial and stromal cells within tissues, where they are known to control many autonomous cellular functions. This review, however, will explore roles for epithelial integrins in the regulation of factors that are secreted into the tissue microenvironment (i.e., components of the secretome). Additional emphasis will be given to integrin α3β1 which has been revealed as a critical regulator of the secretome in keratinocytes. In general, the model organ for this review will be skin, wherein the epidermis is the epithelial compartment mainly comprised of keratinocytes, and the dermis is the connective tissue, or stromal, compartment. The focus of this review will be on the epidermal integrin-dependent secretome that influences matrix assembly/remodeling and supports distinct cells within the skin stroma, principally in the contexts of cutaneous tumor progression and wound healing.

1.1. Epidermal Integrins

As mentioned previously, integrins bind various ECM proteins. Often times, they are classified by their ability to bind the tripeptide sequence Arg, Gly, Asp, or RGD, that is present in some ligands. For example, fibronectin, tenascin, and vitronectin have accessible RGD motifs but collagen and laminin do not [1]. The most constitutively expressed and highly expressed integrins in skin keratinocytes are laminin receptors α6β4 and α3β1, and collagen receptor α2β1 (Fig. 2); all non-RGD-binding integrins [8]. Most remaining epidermal integrins are RGD-binding integrins, including α5β1, αvβ5, and αvβ6 (Fig. 2) [8]. Integrin α9β1 is also expressed by skin keratinocytes (Fig. 2), although it is categorized classically as a leukocyte receptor (Fig. 1). Integrin α9β1 has several known ligands including tenascin, fibronectin, and osteopontin [8]. The repertoire of integrins on keratinocytes are necessary to mediate adhesion or movement on different substrates, to facilitate distinct signal transduction, and/or to provide structural support. The main structural integrin in skin is α6β4, a key protein found in hemidesmosomes which are structures that connect basal keratinocytes to the underlying dermis, essential for maintaining epidermal-dermal adhesion [4, 9]. The primary role of α6β4 is structural, while it has minimal involvement in signal transduction [10]. In general, the expression of epidermal integrins is spatially limited to basal keratinocytes that are proximal to matrix proteins of the epidermal basement membrane (BM), a thin, specialized ECM structure that underlies the epidermis and divides the epidermis from the dermis. It is important to note that some ligands are revealed during tissue remodeling processes wherein the BM is lost or degraded, such as during wound healing or tumorigenesis. An example of this is integrin α2β1 binding to dermally located fibrillar type I collagen (Fig. 2) [11]. Indeed, several epidermal integrins including α2β1, α3β1, α9β1, αVβ5, and αVβ6 are upregulated during wound healing [10, 12–14], indicating critical roles for these integrins in this tissue remodeling process [8].

Figure 2. Epidermal integrins and their ECM ligands.

Main epidermal integrins include laminin receptors, hemidesmosomal integrin α6β4 (inset box) and α3β1, collagen receptor α2β1, RGD-binding receptors α5β1, αvβ5, and αvβ6, as well as integrin α9β1. Ligands are depicted as being attached to their respective integrin receptors and are labeled in italics. The basement membrane zone (BMZ) is roughly demarcated by the dotted line. This simplified schematic illustrates some known interactions between keratinocyte integrins and ECM ligands. These interactions are not thought to occur simultaneously as depicted. Certain ligands become available during remodeling processes whereby the BM is absent or degraded (e.g., dermal type I collagen). See text for supporting literature and expanded discussion. This model was created using BioRender.com.

2.0. THE EPIDERMAL INTEGRIN-MEDIATED SECRETOME

Although first classified as adhesion receptors, it is now appreciated that integrins have a variety of cellular functions. Such additional functions are not limited to cell-autonomous roles, but also include paracrine functions that are facilitated by integrin-regulated secreted factors. The so-called epidermal integrin-mediated secretome can influence skin matrix and epidermal BM composition and remodeling (see sections 3.1 and 3.2) and can mediate crosstalk to different cell types within the skin stroma, thereby impacting numerous distinct cellular processes in several different contexts, which will be discussed in more detail below (see section 3.3). First, a brief overview of general mechanisms that underlie secretome regulation by integrins will be reviewed, followed by a discussion on integrin α3β1 as a regulator of the secretome in keratinocytes. There is a focus on integrin α3β1 throughout this review, since the role for this integrin in regulating the secretome is well-characterized. Certainly, there are examples whereby other integrins regulate certain secreted factors, although the full extent of their ability to mediate a cellular secretome is currently not as well-understood.

2.1. General mechanisms of secretome regulation by integrins

The regulation by epithelial integrins of keratinocyte-generated secreted factors can happen at the level of gene expression or at the level of secretion (e.g., via exosomes; Fig. 3). A major mode of such secretome regulation is thought to occur at the level of transcription, whereby the integrin mediates an intracellular signal transduction cascade, as classically reviewed [15], that may result in the up- or downregulation of a gene that encodes a secreted protein. Alternatively, integrin regulation of the secretome can occur more directly at the level of protein secretion. Recent work indicates extensive exosomal regulation of the secretome by epithelial integrins, as reviewed [16], through mediating cargo within exosomes or by guiding the exocytic process. For example, the regulation of exocytic machinery was shown to be mediated through integrin-dependent signaling in neurons during the process of neurite sprouting [17]. In general, it is not currently understood whether the ability of an integrin to regulate a secreted factor (either transcriptionally or by regulation of exosomes) is mediated by ligand binding or by lateral interactions, or if a particular mode of integrin activation promotes either of these pathways preferentially.

Figure 3. The keratinocyte integrin-mediated secretome regulates the skin microenvironment.

Integrins expressed by basal keratinocytes of the skin epidermis (e.g., α3β1) promote the generation/secretion of proteins through several modes of regulation (e.g., transcriptional, exosomal). These secreted proteins make up the integrin-mediated secretome. This secretome can modulate the skin microenvironment by (1) promoting matrix generation through the regulation of ECM/BM proteins, (2) promoting matrix remodeling through the regulation of proteases, or (3) supporting stromal cells through the regulation of growth factors and/or cytokines. This model was created using BioRender.com.

Additionally, integrins themselves can be exosome cargo, and this has been shown to mediate the docking and uptake of exosomes [18], thereby directing intercellular crosstalk. Interestingly, as reviewed elsewhere [18, 19], integrins derived from exosomes have been found to promote cancer progression. For example, integrins and their ligands within tumor cell-derived exosomes can mediate the selection of metastatic tissue targets and trigger integrin-mediated signal transduction to aid in the formation of a new metastatic niche. Moreover, the Languino group has shown that integrin αVβ3 is transferred in exosomes from tumorigenic to nontumorigenic cells, enhancing migration of the latter [20], and that integrin αVβ6 can also undergo exosome-mediated transfer [21].

2.2. Integrin α3β1 regulates the keratinocyte secretome

Mass spectrometry (MS)-based proteomic analyses on medium conditioned by cultured keratinocytes have shown that integrin α3β1 is a critical regulator of the epidermal secretome [22, 23]. The Has group performed a study profiling the keratinocyte secretome from patients with mutations in ITGA3, the gene encoding the α3 integrin subunit. These patients lack α3β1, causing a disorder identified as interstitial lung disease, nephrotic syndrome, and epidermolysis bullosa (ILNEB). Characteristics of ILNEB resemble those seen in α3β1-deficient mice with a global knockout of the Itga3 gene [24, 25]. Although the primary concerns in ILNEB are respiratory and renal issues, the skin fragility initially suggested a diagnosis [26], as the distinctive blistering is reminiscent of that observed in neonatal α3 knockout mice wherein rupture occurs at the epidermal-dermal BM [24]. Furthermore, global Itga3 gene knockout in mice led to perinatal mortality [25], which parallels the short lifespan of most ILNEB patients [26, 27], and ILNEB patients exhibit disorganized BMs and impaired barrier functions in the skin, kidneys, and lungs, mirroring phenotypes observed in global α3 knockout mice [24, 25]. Interestingly, MS studies have shown that keratinocytes from ILNEB patients regulate the microenvironment by promoting fibronectin matrix deposition and the expression of integrins that bind fibronectin [22], perhaps as a mechanism to compensate for loss of α3β1.

Using keratinocyte cell lines, early work from the DiPersio group demonstrated the α3β1-dependent regulation of genes that encode matrix proteins or proteases with recognized roles in modification of the microenvironment, such as matrix metalloprotease (MMP)-9, mitogen related protein (MRP)-3, bone morphogenetic protein (BMP)-1, interleukin (IL)-1α and fibulin-2 [28–33]. More recently, we employed MS analysis to determine a more complete α3β1-dependent secretome in cultured keratinocytes. Indeed, many proteins detected as part of the α3β1-dependent secretome are ECM-interacting/matricellular proteins, such as multimerin-2, syndecan-4, SPARC, and urokinase-type plasminogen activator (uPA) receptor [34]. Other identified proteins have established roles in modulating the microenvironment, such as proteases tPA, uPA, MMPs 3, 10, and 13, protease inhibitors TIMP1 and 2, and PAI-1 and -2, as well as growth factors and cytokines vascular endothelial growth factor (VEGF)-C, CXCL2, 3, and 5, colony stimulating factors (CSF)-1, −2, and −3, and several prolactins [34] (several factors are discussed in more detail throughout section 3.0). Examples of secreted factors regulated by keratinocyte integrin α3β1, and their known functions in mediating matrix composition, matrix remodeling or crosstalk to stromal cells, are listed in Table 1. In the future, it would be informative to evaluate the extent to which other epidermal integrins regulate the secretome using MS-based proteomics, and to confirm results with in vivo models. Other future work may identify which factors are secreted directly versus as exosome cargo. Additionally, the recent development of what is known as the “secretome mouse” will allow for the in vivo determination of the secretome directly, as this transgenic mouse uses proximity biotinylation to robustly distinguish a tissue-specific set of secreted proteins [35]. This will be important to validate secretome data obtained from 2D cell cultures, as discussed above, which may not successfully recapitulate the in vivo microenvironment.

Table 1.

Examples of secreted factors regulated by keratinocyte integrin α3β1, and their known functions in regulating matrix composition, matrix remodeling or crosstalk to stromal cells.

Class	Secreted Factor	Known Function
Matricellular	Fibulin-2	ECM component, BM stabilizing
	Multimerin-2	ECM component
	SPARC	ECM component
Proteases	BMP-1	ECM remodeling, Laminin processing
	MMP-3	ECM remodeling
	MMP-9	ECM remodeling, Pro-angiogenic
	MMP-10	ECM remodeling
	MMP-13	ECM remodeling
	TPA	ECM remodeling
	UPA	ECM remodeling
Protease Inhibitors	PAI-1	ECM remodeling
	PAI-2	ECM remodeling
	TIMP-1	ECM remodeing
	TIMP-1	ECM remodeling
Growth Factors	CCN-2	Enhanced colony formation, Growth
	MRP-3	Pro-angiogenic
	VEGF-C	Pro-angiogenic
Cytokines	CXCL-2	Immune cell stimulating
	CXCL-3	Immune cell stimulating
	CXCL-5	Immune cell stimulating
	CSF-1	Immune cell stimulating; Macrophages
	CSF-2	Immune cell stimulating
	CSF-3	Immune cell stimulating
	IL1-α	Immune cell homing, Role in fibroblast diff’n

See text for abbreviations, supporting literature, and expanded discussion.

3.0. REGULATION OF THE SKIN MICROENVIRONMENT BY THE EPIDERMAL INTEGRIN-MEDIATED SECRETOME

During periods of active tissue remodeling, keratinocytes regulate the microenvironment by directing the secretion of growth factors, cytokines, proteases and ECM proteins. The contribution of ECM proteins and proteases can modify the nearby microenvironment more directly by generating and remodeling matrix (e.g., the epidermal BM), while the secretion of cytokines and growth factors can, in a paracrine fashion, regulate stromal cells to support various processes such as angiogenesis, fibroblast contraction, and the inflammatory response (Fig. 3) [60–62]. In this section, we will first discuss roles of integrins in the general regulation of growth factors, cytokines, proteases and matrix proteins, followed by the α3β1-dependent regulation of the epidermal BM. Lastly, in this section we will discuss paracrine crosstalk to stromal cells via the epidermal integrin-mediated secretome in the contexts of tumorigenesis and repair.

3.1. Regulation of matrix proteins, proteases, cytokines and growth factors

Certainly, it is understood that changes in matrix composition can alter the repertoire of integrin activation via ligation. However, integrins can also regulate the ECM by mediating the expression and assembly of matrix components. For example, fibronectin-binding integrins are known to support fibronectin expression and fibrillogenesis [36]. Also, it has been shown that integrin α2β1 promotes expression of type 1 collagen genes [37], whereas α1β1 inhibits dermal collagen synthesis [38].

Extracellular proteases influence the ECM by regulating the degradation and remodeling of matrix [39]. For example, MMPs are implicated in all stages of wound resolution and cancer progression [40, 41]. In addition, proteases permit growth factors release from the cell surface or from matrix reservoirs [40, 42]. ECM remodeling by integrin-regulated proteases is reviewed elsewhere [43, 44]. Some examples of protease regulation by epithelial integrins include the regulation of MMP-9 and MMP-3 by α5β1 [45], and the induction of MMP-9, BMP-1, or uPA by α3β1 in keratinocytes [29, 32, 46]. Moreover, in the cancer setting, αvβ6 has been demonstrated to mediate the induction of several proteases that contribute to invasive behavior, including uPA in epithelial ovarian cancer [47] and MMPs-9 [48] and −3 [49] in squamous cell carcinoma.

Integrins can also regulate the generation or activation of growth factors. For instance, αvβ6 has been shown to activate latent transforming growth factor β (TGFβ) that is bound by matrix [50], and α6β4 can promote the translation of VEGF in breast cancer cells [51]. Keratinocyte α3β1 promotes several growth factors including connective tissue growth factor (CCN2) [52] and MRP-3 [28], as discussed below in sections 3.3.1.2 and 3.3.2, respectively. In general, the differential regulation of matrix components, proteases, and growth factors by integrins is likely mediated by changes in the expression of those integrins, or by alteration in the availability or expression of ligand.

3.2. Integrin α3β1-dependent regulation of the epidermal basement membrane

The cutaneous BM is a laminin-332 (LN332)-rich matrix that physically separates the epidermis from the dermis. Additionally, the epidermal BM provides supportive cues to keratinocytes that aid in their polarization, differentiation, survival, and migration. Murine studies have uncovered critical roles for β1 integrins in the stability and organization of BMs [53, 54]. Integrin α3β1, in particular, has been revealed as a critical regulator of BM integrity and assembly [55–57]. Disorganization of BMs in many tissues and organs (e.g., skin, lung, kidney) has been reported in ILNEB patients with ITGA3 mutations and in α3 knockout mice. Importantly, lack of BM integrity is a likely culprit for the observed morphological defects in these organs [24–26, 58].

The underlying mechanisms of α3β1-dependent BM integrity and organization have mostly been investigated in the skin. Deletion of α3β1 from murine skin [24, 58] or from skin of ILNEB patients [26] results in disorganization of the epidermal BM, which causes BM rupture, generating small blisters that are characterized by the distribution of BM proteins, such as LN332, to both epidermal and dermal sides of said blisters. Skin blistering in murine α3-null neonates was linked to reduced levels of a matricellular protein called fibulin-2, and deletion of fibulin-2 alone sufficed to cause similar skin blisters in neonates [31]. This important role for α3β1-dependent fibulin-2 expression during skin development is reprised in adult mice during wound healing, as adult mice with epidermis-specific deletion of α3β1 also display blistering within the BM of the neo-epidermis in cutaneous wounds that have re-epithelialized [31].

It has also been demonstrated that keratinocyte α3β1 regulates the expression, organization and proteolysis of LN332 both in vivo and in vitro [58–60]. As mentioned previously, α3β1 promotes expression of the extracellular protease BMP-1 [32] which is known to proteolyze the γ2 chain of LN332 [61], and BMP-1 expression has been associated with the α3β1-dependent processing of the laminin γ2 chain in keratinocytes [32], which is thought to regulate keratinocyte motility, and is thought to be a key step in BM maturation [62, 63]. Overall, these findings indicate a role for α3β1 in the regulation of proteins that influence the assembly, stability and maturation of the epidermal BM during skin development and wound healing. As LN332 is the major ligand for α3β1, it follows that this regulatory role may provide an opportunity for this integrin to alter its own cues from the ECM and generate feedback to control its own functions within epidermis.

3.3. Secretome support for stromal cells in context

The following sections will discuss how secreted factors regulated by epidermal integrins support distinct stromal cell populations in the contexts of tumorigenesis and repair.

3.3.1. Tumorigenesis

It is now well-appreciated that epithelial cell transformation is not adequate for carcinoma development, but that a supportive tumor microenvironment (TME) is also necessary, as reviewed elsewhere [64–66]. Secreted factors from tumor cells are a main pathway through which transformed epithelial cells communicate with the required stromal cells to co-opt their support for tumor growth and progression. Important roles are emerging for the tumor cell integrin-mediated secretome in the stimulation of distinct cell populations within the stroma that promote cancer progression. The following subsections will discuss this topic as it relates to vascular endothelial cells (ECs; section 3.3.1.1), as well as tumor-associated fibroblasts (TAFs; section 3.3.1.2) and macrophages (TAMs; section 3.3.1.3).

3.3.1.1. Vascular endothelial cells

In studies performed in models of breast cancer, tumor cell integrins have been shown to crosstalk to ECs through the generation of pro-angiogenic signals. As mentioned above, integrin α6β4 on breast cancer cells enhances VEGF expression, which was shown to stimulate angiogenesis and survival of tumor cells [51]. Moreover, integrin α3β1 on breast cancer cells can promote the expression of pro-angiogenic factors, MMP-9 [67] and cyclooxygenase-2 (Cox-2) [68]. Paradoxically, however, expression of the same integrins (α6 or α3) on ECs was linked to the inhibition of angiogenesis in pathological settings [69, 70]. This demonstrates the critical concept that roles for particular integrins are context dependent and may depend on which type of cell is expressing the integrin. When it comes to exploiting integrins as therapeutic targets (see section 4.1), this is an important consideration that may necessitate advanced cellular targeting techniques rather than a blanket approach.

3.3.1.2. Tumor-associated fibroblasts

Both epithelial and stromal cell integrins can regulate TAF differentiation and function, as reviewed [71]. The expression of integrin α9β1 on breast cancer cells was shown to foster TAF recruitment as well as the upregulation of osteopontin, a matrix protein that contributes to primary tumor growth and metastasis [72]. In an interesting example of direct tumor cell to fibroblast crosstalk, integrin α6β1 on pancreatic cancer cells was shown to interact with the fibroblast uPA receptor to stimulate a MMP-2-driven proteolytic cascade from fibroblasts that supports tumor progression [73]. Additionally, in a study from the Sonnenberg group, stem cells of the hair follicle bulge from mice treated with the two-step chemical tumorigenesis protocol were analyzed by RNA sequencing, comparing control mice to mice with epidermal deletion of integrin α3β1 [52]. Expression profiling determined 15 differentially expressed genes, four of which encode secreted proteins. This study also showed that integrin α3β1 on keratinocytes supports the expression of a growth factor called CCN2 that stimulated 3D growth and colony formation of transformed keratinocytes in vitro [52], consistent with a requirement for integrin α3β1 expression in the formation of skin papillomas [74]. Since CCN2 is well-known to induce the transdifferentiation of myofibroblasts [75], it will be interesting in the future to examine whether integrin α3β1-dependent production of CCN2 from tumor cells promotes tumor progression at least in part by impacting the TAF population.

3.3.1.3. Tumor-associated macrophages

The accumulation of TAMs has been linked with tumor cell expression of integrin αvβ3 in several epithelial tumors from humans and mice, as reviewed [76]. Consistently, our own recent work has shown that integrin α3β1 on epidermal tumor cells supports the TAM population in cutaneous papillomas [23]. As mentioned previously, MS analysis interrogating the keratinocyte secretome showed that several α3β1-dependent secreted factors are colony stimulating factors that are known to support TAMs. The mRNA expression of one such factor, called CSF-1, was reduced from epidermal tumor cells following integrin α3 subunit deletion, with an accompanying decrease in the number of TAMs in the tumor stroma [23].

3.3.1.4. Paradoxical roles for α3β1 in tumor progression

Pro-tumorigenic roles for integrin α3β1 have been established in several cancer types, including breast cancer, squamous cell carcinoma, glioma, gastric cancer, and melanoma, [56, 57, 77, 78] and increased expression of α3β1 and/or its laminin ligands are markers of poor prognosis in certain breast cancers and squamous cell carcinomas [57, 79, 80]. Paradoxically, however, there are also clear instances wherein α3β1 inhibits tumor growth or suppresses metastatic potential [57]. A suppressive role for α3β1 has been shown in HER2-driven breast cancer in both a murine model and in human breast carcinoma cells, wherein α3β1 downregulation promoted tumor progression and invasiveness [81]. This suppressive role was not observed in triple negative breast cancer cells, indicating that the role of α3β1 in breast cancer can be subtype-specific [81].

Interestingly, α3β1 can exert paradoxical functions within the same type of cancer, switching from a tumor-promoting integrin at early stages during tumor development to a cancer-suppressive integrin at later stages during cancer progression. This was demonstrated by the findings from two different groups that used epidermis-specific α3 knockout mouse models combined with models of chemically-induced epidermal skin tumorigenesis. Studies in these models have revealed a requirement for α3β1 in papilloma tumor formation [74] and in the maintenance of tumor growth [34]. However, when an aggressive chemically-induced carcinogenesis protocol was applied to this same genetic model, again fewer tumors formed in epidermis lacking α3β1, but they showed a significantly increased rate of progression to undifferentiated and invasive carcinoma [74]. These findings indicate that roles for integrin α3β1, and likely other integrin receptors, can be cancer type-, cancer subtype-, and even cancer stage-specific, and it is likely that changes in the integrin-dependent secretome follow suit as being either pro-tumorigenic or anti-tumorigenic, although this remains to be confirmed.

3.3.2. Repair

Consistent with aspects of the repair process mirroring tumorigenesis [82], the integrin-mediated keratinocyte secretome is also known to support stromal cells during wound healing. For example, as discussed above, integrin α3β1 on breast cancer cells offers pro-angiogenic support to the tumor [68], and this is recapitulated during wound healing. In wound epidermis, integrin α3β1 stimulates paracrine-mediated wound angiogenesis via keratinocyte secretion of pro-angiogenic MRP-3 [28]. Furthermore, work from the DiPersio group showed that epidermal integrin α9β1 can inhibit the paracrine wound angiogenesis mediated by α3β1 during wound resolution to aid in the regression of overexuberant wound vasculature [83]. Additionally, wound vasculature was shown to be enhanced in α2-null mice, indicating a suppressive role in angiogenesis for this integrin [84]. Altogether, these studies indicate important roles for integrins in coordinating wound angiogenesis through keratinocyte to EC crosstalk.

In a study from the Van De Water group, epidermal integrin α3β1 promotes crosstalk to fibroblasts in the dermis to regulate the myofibroblast phenotype in wounds [33]. Briefly, keratinocyte α3β1-mediated induction of IL-1α stimulates Cox-2/prostaglandin E₂ (PGE₂) signaling to control TGFβ-induced differentiation of wound fibroblasts [33]. A different study consistently found that β1 integrin transgene expression by epidermis enhanced secretion of IL-1α [85]. Similarly, tumor-cell associated IL-1α was shown to stimulate TAFs in a pancreatic cancer model [86], consistent with the notion that TAFs parallel wound myofibroblasts [87].

Roles for epidermal integrins in mediating paracrine crosstalk to immune cells during wound healing are just beginning to be explored. The treatment of epithelial cells with an anti-integrin α3 antibody was found to inhibit immune cell-homing interleukins and macrophage chemoattractant protein 1 (MCP-1) [88]. Consistently, our recent study has shown that keratinocyte integrin α3β1 induces CSF-1 expression through a YAP/TEAD-dependent mechanism to support the macrophage population of cutaneous wounds in vivo [89]. Again, this parallels our findings from a skin papilloma model, discussed above in section 3.3.1.3 [23], further demonstrating similarities between roles for the integrin-mediated secretome in the wound and tumor settings.

4.0. CONCLUSION

The vastness of integrin function is becoming more apparent as studies continue to uncover the wide-ranging roles that epidermal integrins have in modeling the skin microenvironment. It is likely that these important integrin-guided functions are gained or enhanced during both normal and pathological tissue remodeling processes, like wound healing and tumor progression, respectively. Additionally, it must be acknowledged that stromal cells express their own repertoire of integrins that are likely to contribute to a complex network of crosstalk within the stroma and reciprocally to epithelial cells [71, 90]. Future work should include identifying more complete integrin-mediated secretomes, as well as the entirety of distinct stromal cell populations that they support, and how these functions differ between normal and pathological processes. In future and past work alike, it is important to consider potential caveats associated with model systems utilized. For example, deriving secretome data from cells in 2D culture alone may insufficiently represent the in vivo setting. Furthermore, murine models wherein a knockout is embryonic, as is the case in widely used constitutive models, may result in compensatory mechanisms which could obscure some integrin functions. For instance, a recent study from our group identified an important role for integrin α3β1 in promoting wound re-epithelialization [91]. This role was elucidated using an inducible system whereby the integrin knockout was induced in epidermis of adult mice just prior to wounding. Interestingly, this role was not observed in constitutive models with embryonic deletion of epidermal α3β1 [28, 92]. A better understanding of accurate and complete integrin-mediated secretomes will be beneficial towards the consideration of integrins as potential therapeutic targets (see section 4.1), since even direct targeting of an integrin within a single cellular compartment is likely to have far-reaching pleiotropic effects. Indeed, topical application of the secretome from human amniotic epithelial cells was shown to attenuate psoriatic lesions in a murine model [93], demonstrating the powerful anti-inflammatory properties that can be derived from cells of epithelial origin.

4.1. Future potential for integrin-targeted therapies

The possibility of therapeutically targeting epidermal integrins in the treatment of chronic wounds or skin cancer is intriguing since the effect would not be limited to the epidermal compartment but would extend to the secretome and potentially impact the microenvironment in a powerful way. However, the development of integrin-targeting therapeutics has been challenging given the complexity of the integrin family, the vastness of integrin functions, and the ubiquitous expression of integrins across cell types. Adding to the complexity, integrin function depends on a balance between active and inactive states mediated by several mechanisms, including ligand binding, lateral protein interactions, conformational changes and trafficking [94]. Thus far, clinical targeting of integrins has been limited to integrin inhibition, as discussed more below. However, the concept of integrin targeting should not be confined to inhibition. For example, since the activity of most integrins is thought to promote wound healing, promoting the function of certain integrins is likely to be beneficial in the treatment of chronic or hard-to heal wounds. In the wound clinic, topical application of ECM-inspired biomaterials could promote healing, as reviewed [95], and this concept could be used as a means to deliver exogenous ligand for the enhancement of integrin function. Similarly, as discussed in section 3.3.1.4 above, integrin function in certain cancer contexts is anti-tumorigenic, therefore, promoting the function/activity of certain integrins may even be helpful in some cancer settings.

The therapeutic targeting of integrins has been established clinically. For example, vedolizumab is an antibody targeting α4β7 that is utilized to inhibit lymphocyte trafficking to the gut mucosa for the management of ulcerative colitis and Chron’s disease [96, 97]. In a recent review, it was stated that about 90 kinds of integrin-targeting therapeutics were in clinical trials as integrin antagonists, including antibodies, antibody-drug conjugates, synthetic mimetic peptides, and small molecules, for the treatment of a variety of maladies, including cancers, autoimmune diseases, viral diseases, and fibrotic diseases [94]. In the cancer clinic, integrin-blocking RGD mimetic, Cilengitide, has been utilized with limited success [98, 99]. It was proposed that Cilengitide would reduce angiogenesis in tumors by inhibiting RGD-binding integrins on ECs, but this methodology is likely confounded by the expression of RGD-binding integrins on other cells within the TME. Moving forward, the consideration of how integrin-targeting agents will impact the entire milieu, either directly through integrin-decorated cells or indirectly through the integrin-regulated secretome, is necessary for their success.

Moreover, RGD mimetics only inhibit RGD-binding integrins. As discussed in this review, there is ample preclinical data to suggest the importance of laminin-binding integrins, such as keratinocyte α3β1, in regulating critical wound healing and tumorigenic functions, and the power of such targeting is likely to be amplified through this integrin’s ability to contribute a secretome that robustly supports the stroma. This hypothesis has not been clinically tested, and more work should be done preclinically to understand the intricacies of integrin function and how these functions may flux during these dynamic processes. Of note, a synthetic high-affinity peptide, called LXY30, has been shown to faithfully bind to α3β1 [100, 101]. While it is not thought to necessarily impact the activity of the integrin itself, it may be useful in a combined approach to deliver therapy, or an inhibitory microRNA for example, to tumor cells that express high levels of integrin α3β1 in contexts where this integrin has been shown to promote tumor growth and/or progression.

Highlights.

• Integrins are transmembrane receptors for extracellular matrix adhesion.

• Keratinocyte integrins regulate factors such as matrix proteins, proteases, growth factors, and cytokines that are secreted into the microenvironment (i.e., the secretome).

• Integrin α3β1 is emerging as a regulator of the secretome in keratinocytes.

• Integrin-regulated secreted factors mediate crosstalk to the microenvironment by: (1) promoting matrix generation through the regulation of ECM/BM proteins, (2) promoting matrix remodeling through the regulation of proteases, or (3) supporting stromal cells through the regulation of cytokines and/or growth factors.

• Secreted factors regulated by certain integrins can promote paracrine crosstalk to distinct cells such as immune cells, endothelial cells, and fibroblasts to control both pathological (i.e., tumor growth and progression) and repair (i.e., wound healing) processes.

ABBREVIATIONS

ECM

extracellular matrix

RGD

arginine, glycine, aspartic acid tripeptide sequence

BM

basement membrane

MS

mass spectrometry

ILNEB

interstitial lung disease, nephrotic syndrome, and epidermolysis bullosa

MMP

matrix metalloprotease

MRP-3

mitogen-regulated protein-3

BMP-1

bone morphogenetic protein-1

IL-1α

interleukin 1α

uPA

urokinase-type plasminogen activator

VEGF

vascular endothelial growth factor

CSF

colony stimulating factor

TGFβ

transforming growth factor β

CCN2

connective tissue growth factor

LN332

laminin-332

TME

tumor microenvironment

EC

endothelial cell

TAF

tumor-associated fibroblast

TAM

tumor-associated macrophage

Cox-2

cyclooxygenase-2

PGE₂

prostaglandin E₂

MCP-1

macrophage chemoattractant protein 1