Desmoglein-3 (Dsg3) was first identified as a pemphigus vulgaris antigen (PVA) from the human keratinocyte expression libraries in 1991 by Amagai et al. in an attempt to search for the targets of pemphigus vulgaris autoantibodies (Amagai et al. 1991). Due to its significant homology with the cadherin family of cell adhesion molecules and most remarkably to Dsg1, Dsg3 is characterized as a member of cadherin superfamily and specifically the one in desmoglein subfamily which contains Dsg1. Pemphigus vulgaris (PV) is a potentially lethal autoimmune blistering disease that affects oral mucosa and skin with the manifestation of acantholysis characterized as the loss of adhesion between epithelial cells and structural components maintaining cell cohesion in the tissues caused by the action of autoantibodies against Dsg3 as well as Dsg1. Dsg1 is also a target antigen in another autoantibody-mediated blistering disease of the skin, pemphigus foliaceus. There are two subfamilies of desmosomal cadherins, desmoglein (Dsg) and desmocollin (Dsc), both of which are members of the cadherin superfamily of calcium-dependent adhesive proteins in desmosomes, one type of intercellular junctions between epithelial, myocardial, and other cell types that confer strong cell–cell adhesion (Thomason et al. 2010). There are four human desmoglein isoforms (Dsg1–4) and three desmocollin isoforms (Dsc1–3), which exhibit different tissue and cell type–specific expression patterns. For instance, Dsg3/Dsc3 and Dsg1/Dsc1 are more restricted to complex epithelial tissues, whereas Dsg2/Dsc2 are widely expressed in desmosomes of the basal layer of stratified epithelia, simple epithelia, and nonepithelial cells such as the cardiomyocytes and lymph node follicles. Dsg3, in particular, is found confined in the basal and immediate suprabasal layers of epidermis but uniformly distributed across the entire stratified squamous epithelia in the oral mucosa.
All the genes that encode Dsgs and Dscs are located within a tightly linked cluster in chromosome 18q12.1 (Frank et al. 2001). Although mutations in the human Dsg3 gene have not been described to date, mutations in DSG3 underlie the balding mice (Davisson et al. 1994; Pulkkinen et al. 2002) and overt squeaky phenotype with a spectrum of severe pathology such as cyclic hair loss, obstructed airways, and severe immunodeficiency subsequent to the development of oral lesions and malnutrition (Kountikov et al. 2015). The ablation of DSG3 gene in mice leads to fragility of the skin and oral mucous membrane, analogous to those found in PV patients, along with runting and hair loss (Koch et al. 1997, 1998). However, a recent report has revealed expression of a 31-kDa truncated protein of Dsg3 containing 282 amino acids, which corresponds to the N-terminal truncated intracellular domain of Dsg3, in differentiating keratinocytes of human epidermis (Lee et al. 2009b). This truncated Dsg3 was also found to be upregulated in psoriatic epidermis and skin tumors, including Bowen’s disease and squamous cell carcinoma (SCC). In addition, recently, upregulation of the Dsg3 gene and protein (wild type) has also been reported in cancer, especially in SCC with increased levels of expression correlating with the clinical stage of malignancy, implicating its procancerous role and potentiality to serve as a diagnostic and prognostic marker (Brown and Wan 2015).
The mature form of Dsg3 (50–999 amino acids) is cleaved from a precursor protein and contains the unglycosylated peptide of 950 amino acids. The first 23 amino acids are the signal peptide. Like other cadherins, Dsg3 is a single-pass transmembrane glycoprotein consisting of the extracellular domain with five motifs of about equal size, EC1 through EC5, which, except for EC5, have homology with each other, in particular among EC1 to EC3, the most amino-terminal domains. All five extracellular motifs that contain 566 amino acids (50–615 amino acids), also show significant homology with the corresponding region in classical cadherin, such as P-cadherin (Amagai et al. 1991). Both Dsg1 and Dsg3 have a RAL site (128–130 amino acids) in EC1 that corresponds to the conserved cell adhesion recognition sequence HAV in an equivalent position in classical cadherins that is responsible for homophilic binding. Other conserved sequences in the extracellular domain are several putative Ca2+-binding sites with all cadherins. The transmembrane region of Dsg3 encompasses 25 amino acids (616–640 amino acids) and the cytoplasmic domain contains 359 amino acids (641–999 amino acids) which is substantially longer than that of typical cadherins (approximately 150 residues) but shorter than that of Dsg1 (480 residues). The cytoplasmic tail consists of several domains, such as the intracellular anchoring domain (IA), intracellular cadherin-specific domain (ICS), proline-rich linker domain (IPL), repeating unit domain (RUD), and lastly the desmoglein-specific terminal domain (DTD). The proteins that have been identified to directly bind to the cytoplasmic domain of Dsg3 includes p120 that binds to IA, plakoglobin (Pg) and caveolin-1 (Cav-1) that bind to ICS, and actin with the specific domain not yet defined (Andl and Stanley 2001; Brown et al. 2014; Wan et al. 2016; Kanno et al. 2008). The evidence for the interaction with actin proteins came from the mass spectrometry analysis of the Halo immunoprecipitates from A431 cells transfected with plasmid containing the entire Dsg3 cytoplasmic tail tagged with Halo at the N-terminus (Brown et al. 2014). Dsg3 has also been found to form complexes with many other proteins, in particular the signal molecules that are described below. Although Dsg3 belongs to the desmoglein subfamily, it has been found to be uniformly distributed on keratinocyte cell surface of stratified squamous epithelia, suggesting that Dsg3 does not solely function in desmosome adhesion but rather plays an extrajunctional role associating with other cellular processes than cell–cell adhesion.
The evidence that suggests Dsg3 acting as a signaling molecule comes from plethora studies in pemphigus research over the past two decades that demonstrates that PV-IgG binding to Dsg3 on keratinocyte surface triggers a cascade of intracellular events, such as Dsg3 phosphorylation, activation of phospholipase C (PLC) signal pathway including production of inositol 1,4,5-trisphosphate (IP3) and induction of a Ca2+/PKC, signaling through p38MAPK/heat shock protein 27 (HSP27), Src, epidermal growth factor receptor (EGFR), apoptosis, c-Myc, adaptor protein plakoglobin/plakophinlin-3, as well as Rho GTPases and rearrangement of actin cytoskeleton, etc. All these observations have implicated a key role of Dsg3 in mediating outside-in signaling that leads to desmosome remodeling, cell proliferation and differentiation, or apoptosis. Deregulation of these signal pathways causes depletion of Dsg3 from the desmosomes and keratinocyte surface, resulting in cell dissociation and blister formation as observed in pemphigus disease. For convenience, the discussion below regarding the PV-IgG induced Dsg3 signaling is referred to as “PV-IgG signaling” in this review, unless otherwise specified.
Phospholipase C (PLC) and Protein Kinase C (PKC) Signaling
p38 MAPK Signaling
EGF Receptor and Apoptosis Pathways
It has been shown that p38 MAPK activation not only stimulates but also enhances cell apoptosis in PV-IgG-treated keratinocytes, in PV patients, and also in passive transfer model, suggesting that apoptosis could be a major causal factor of the acantholytic phenomenon (Lee et al. 2009a; Pelacho et al. 2004; Wang et al. 2004; Grando et al. 2009). The observations of apoptosis in PV includes (1) secretion of soluble FasL; (2) elevated cellular amounts of FasR, FasL, Bax, and p53 proteins; (3) reduction in the levels of cellular Bcl-2; (4) enrichment in caspase 8 and activation of caspases 1 and 3; and (5) coaggregation of FasL and FasR with caspase 8 in death-inducing signaling complex (Wang et al. 2004). Furthermore, in skin organ cultures and in in vitro cultured keratinocytes, PV-IgG is shown to be able to induce the caspase activation and DNA fragmentation, and the caspase inhibitors can prevent the formation of PV-IgG-induced epidermal lesions. In support, another independent report showed caspase-3 activation, Bcl-2 depletion, and Bax expression, and the p38 MAPK inhibitors were able to block the activation of caspase-3, the proapoptotic proteinase, suggesting that initiation of apoptosis is downstream to, and a consequence of, p38MAPK activation (Lee et al. 2009a). Hence, the Fas-mediated cell death seems to be involved in PV-IgG signaling and ultimately the pemphigus acantholysis.
Plakoglobin (Pg), a binding partner of Dsg3 and with dual functions in both desmosomes and nucleus, plays a role in suppression of transcription factor c-Myc and is described as the principle effecter of PV-IgG-mediated signals downstream of c-Myc (Williamson et al. 2006; Williamson et al. 2007; de Bruin et al. 2007). When keratinocytes treated by PV-IgG, the nonjunctional pool of Dsg3 along with Pg undergoes a process of enhanced protein turnover leading to depletion of nuclear Pg and attenuation of its suppression of c-Myc (Fig. 3). As a consequence, a marked increase in c-Myc accumulation and proliferating Ki67-positive cells was presented in PV of both human and animal patients. In line with this finding, Pg-knockout mouse exhibited high levels of c-Myc in epidermal keratinocytes, further confirming a suppressive role of Pg in c-Myc activity. What is more, another desmosomal adaptor protein plakophilins 3 (Pkp3) is implicated as an effecter of Src signaling, and upon PV-IgG binding to keratinocytes and activation of Src, tyrosine phosphorylation of Pkp3 was induced leading to the detachment of Pkp3 from Dsg3 and cytoplasmic accumulation accompanied with cell dissociation (Cirillo et al. 2014). Taken together, all these data support the notion that Dsg3 cross talks with EGFR and regulates its signaling pathway, including the downstream effecters Src, ERK, p38 MAPK, and c-Jun as well as the desmosomal adaptor proteins Pg and Pkp3 that collectively result in gross effects on cell proliferation, apoptosis, as well as cell–cell detachment.
Although strong evidence suggests activation of EGFR is necessary for PV-IgG-induced apoptosis and cell–cell detachment, there is also reports, however, arguing against this theory with the findings implying that PV-IgG-mediated signaling does not require EGFR and its internalization, especially the phosphorylation at Tyrosine (Tyr)-1173 (the canonical activation of EGFR) and the c-Src-dependent site Tyr-845, nor apoptosis as absence of apoptosis was observed in the early lesions of PV acantholysis in vivo and also in the keratinocyte cultures treated with PV-IgG in vitro (Heupel et al. 2009; Schmidt and Waschke 2009; Schmidt et al. 2009). Even if both PV-IgG and EGF elicit cell discohesion and cytokeratin retraction, only the effects of EGF can be blocked by inhibition of EGFR and c-Src. In support, laser tweezer experiments showed that impaired bead binding of Dsg3 and Dsg1 in response to PV-IgG was not affected by inhibition of either EGFR or c-Src. These findings implicate that neither EGFR nor Src attributes to the loss of Dsg-mediated adhesion, cytokeratin retraction, and keratinocyte dissociation and suggest that apoptosis detected in PV may occur secondary to acantholysis event.
Signaling to E-Cadherin/Src and the Wnt/Beta-Catenin Pathways
Dsg3 is also implicated to play a part in influencing Wnt/beta-catenin signal pathway (Fig. 4). In addition to a cell adhesive role in adherens junctions, beta-catenin is known to act as a transcriptional coregulator in Wnt-signal pathway (Novak and Dedhar 1999; Kikuchi et al. 2006). The conserved Wnt/beta-catenin signaling plays an important role in regulating development, cell proliferation, migration, and cell-fate decision. Activation of this signal pathway stabilizes beta-catenin in the cytoplasm, leading its translocation to the nucleus via Rac1 and other factors, where it binds to LEF/TCF transcription factors and increases the expression of Wnt target genes, such as Myc, cyclin D1, TCF-1, PPAR-δ, MMP-7, Axin-2, and CD44, etc., in order to facilitate cell migration. In the absence of Wnt-signal, beta-catenin is targeted to ubiquitination and proteasomal degradation. In vitro study showed that overexpression of Dsg3 in A431 cancer cell line activates beta-catenin and its transcriptional activity to some degree (unpublished data). In line with this, a recent report suggested that Dsg3 regulates Wnt/β-catenin signaling indirectly, in a Pg-dependent manor, by a mechanism of sequestering Pg and preventing its nuclear translocation and suppression of LEF/TCF transcriptional activity (Chen et al. 2013). Dsg3 deletion caused an inverse effect with increased nucleus translocation where Pg interacts with and inhibits TCF/LEF transcription activity, resulting in suppression of tumor growth and invasion. Correspondingly, the increased levels of Dsg3 were shown to be correlated with reduced nuclear Pg accompanied with the elevated expression of the LEF/TCF transcriptional targets, cyclin D1, c-Myc, and MMP7, in both the head and neck cancer tissues and the cultured oral cancer cell lines (Chen et al. 2013). Thus, upregulation of Dsg3 in squamous cancer could potentially tip the balance in favor of the β-catenin-LEF/TCF interaction and activation via the suppression of the Pg nuclear translocation.
Dsg3 Regulates Rho GTPase, Ezrin, c-Jun/AP-1, and Organization of Actin Cytoskeleton
In vitro study using gain-of-function approach showed that overexpression of full-length human Dsg3 in epithelial cells causes remarkable induction of Rac1/Cdc42 and to a lesser extent, RhoA (Tsang et al. 2012a). Such an activation of Rho GTPases is accompanied with significant increase of actin turnover and pronounced membrane protrusions and dynamics in various epithelial cell lines, including MDCK and A431 epidermoid cells (Tsang et al. 2012a). As a consequence, enhanced cell migration and invasion is consistently observed in these cells when compared to the respective control cells with relatively low levels of endogenous Dsg3 expression. Knockdown of Dsg3 results in perturbation in junction formation and accompanied cell polarization that require cortical actin assembly. Using various techniques, such as immunofluorescence, biochemical methods, as well as mass spectrometry, it was shown that Dsg3 not only colocalizes with but also physically interacts with actin proteins although the actual nature of their association remains undefined (Tsang et al. 2012a; Brown et al. 2014). In line with this finding, it is not surprising that Dsg3 is found to interact with an actin binding protein Ezrin, a member of the ERM family. Partial colocalization between Dsg3 and Ezrin is displayed at the plasma membrane, especially in the membrane projections of A431 cells. Furthermore, this interaction is proved by coimmunoprecipitation analysis with the rabbit anti-Ezrin antibody in Dsg3-dosage-dependent manner (Brown et al. 2014). In addition, Dsg3 is found to be capable of regulating the Ezrin activation by enhancing phosphorylation at Thre567 residual which is known to attribute to the accelerated motility of cancer cells. The functional importance of such an interaction was further validated by colocalization analysis of Ezrin/F-actin or CD44/F-actin at the plasma membrane that suggests that Dsg3 is required for the proper function of Ezrin at the plasma membrane since Dsg3 knockdown impaired the associations between these proteins (Brown et al. 2014). The increased expression and activation of Ezrin is known to have positive correlation with cancer development, progression, and metastasis (Kong et al. 2013; Clucas and Valderrama 2014). The increased Ezrin phosphorylation in Dsg3-overexpressing cells could be abrogated substantially by inhibition of PKC, p38 MAPK, Rho kinases that are known to be involved in Ezrin activation (Brown et al. 2014) (Fig. 5). This finding indicates that Ezrin is likely a downstream effecter of Dsg3 signaling that regulates cortical actin organization as well as junction assembly. What is more, activation of c-Jun/AP-1 has also been identified in Dsg3-overexpressing cells by phospho-kinase array analysis and the luciferase activity assay in various cell lines with a marked increase in c-Jun S63 phosphorylation (Brown et al. 2014) (unpublished data) (Fig. 5). Knockdown of Dsg3 in A431 cell line showed more than twofold reduction of AP-1 activity relative to the matched control cells, assessed by the luciferase activity assay. Altogether, these additional data place Dsg3 as a key surface regulator for PKC-dependent Ezrin phosphorylation (activation) as well as the c-Jun/AP-1 activation that likely contributes to its procancerous role in Dsg3 associated cancer progression and metastasis.
Although Dsg3 was originally described as a desmosomal adhesion protein, accumulating evidence suggests that it is a versatile molecule with its role more related to cell signaling in epithelial cells. It is worth noting that Dsg3 is not only found in the core domain of desmosomes but also recognized to be distributed on the entire surface at the plasma membrane, and thus its signaling roles as described above are likely associated with its nonjunctional pool beyond the desmosome adhesion. In fact, the protein complex containing Dsg3 and other signal molecules such as Src and Ezrin is detectable in nonionic detergent soluble fraction that is disassociated with the cytoskeleton. Of interest, when expressing chimeric construct that encodes the extracellular domain of Dsg3 and the cytoplasmic domain of E-cadherin in fibroblast L-cells, they only displayed slight aggregation in contrast to that transfected with the wild type of E-cadherin (Amagai et al. 1994). In line with this finding, overexpression of wild-type Dsg3 in epithelial cell lines failed to render enhanced cell–cell adhesion but instead showed promoted cell migration and invasion accompanied with the downregulation of E-cadherin adhesion due to the activation of Src and other signaling pathways (Brown et al. 2014; Tsang et al. 2010).
Dsg3 is found to be associated with two human diseases, PV and squamous cell carcinoma. As a major target of PV-IgGs, Dsg3 is able to trigger a cascade of intracellular events, including the phosphorylation of itself and its depletion from the desmosomes and cell surface and concomitant activation of various signal molecules, such as PLC, PKC, EGFR, Src, p38 MAPK, and c-Myc. As a consequence, cell–cell detachment occurs, leading to blistering and pemphigus acantholysis in Dsg3 baring tissues such as skin and oral mucosa. Disruption of cell cohesion is also accompanied with changes in cell proliferation and induction of apoptosis due to the activation of c-Myc and p38 MAPK pathways, respectively, that attribute to pemphigus acantholysis. On the other hand, upregulation of Dsg3 is found in squamous cell carcinoma in various organs and shows a positive correlation with clinical grade and poor differentiation of cancers. In support, in vitro studies have demonstrated that overexpression of Dsg3 in cancer cell lines (gain-of-function) elicits activation of signal pathways, including Src, Rac1/Cdc42 GTPases, Ezrin, as well as transcription factors c-Jun/AP-1, all of which are known to be responsible for cancer invasion and metastasis. Thus, deregulation of Dsg3 signaling may play a key part in cancer development and progression. However, our current understanding of the biological function of Dsg3 and its related cellular and molecular mechanisms remains limited, and there are still many questions unanswered. For instance, we still do not understand why the distinct expression patterns of Dsg3 exist in epidermis and oral mucous membrane, both of which are the stratified squamous epithelia, with more restricted basal distribution in the former and broad uniform expression in the latter. Is there any specific function for Dsg3 that is correlated to such a distinct tissue distribution? What is the actual ligand or environmental cue that triggers the Dsg3 signaling? Does it exert a function in response to tissue mechanics as many of the pathways regulated by Dsg3 are known to be involved in mechanotransduction? Can Dsg3 serve as a drug target in cancer therapy? Of course, it is challenging to address all these questions. Future investigation will advance our knowledge and shed light on the additional role of Dsg3 in cell biology.
- Brown L, Waseem A, Cruz IN, Szary J, Gunic E, Mannan T, Unadkat M, Yang M, Valderrama F, O’Toole EA, Wan H. Desmoglein 3 promotes cancer cell migration and invasion by regulating activator protein 1 and protein kinase C-dependent-Ezrin activation. Oncogene. 2014;33(18):2363–74.PubMedCrossRefGoogle Scholar
- Esaki C, Seishima M, Yamada T, Osada K, Kitajima Y. Pharmacologic evidence for involvement of phospholipase C in pemphigus IgG-induced inositol 1,4,5-trisphosphate generation, intracellular calcium increase, and plasminogen activator secretion in DJM-1 cells, a squamous cell carcinoma line. J Invest Dermatol. 1995;105(3):329–33.PubMedCrossRefGoogle Scholar
- Grando SA, Bystryn JC, Chernyavsky AI, Frusic-Zlotkin M, Gniadecki R, Lotti R, Milner Y, Pittelkow MR, Pincelli C. Apoptolysis: a novel mechanism of skin blistering in pemphigus vulgaris linking the apoptotic pathways to basal cell shrinkage and suprabasal acantholysis. Exp Dermatol. 2009;18(9):764–70.PubMedCrossRefGoogle Scholar
- Kitajima Y, Aoyama Y, Seishima M. Transmembrane signaling for adhesive regulation of desmosomes and hemidesmosomes, and for cell-cell datachment induced by pemphigus IgG in cultured keratinocytes: involvement of protein kinase C. J Investig Dermatol Symp Proc. 1999;4(2):137–44.PubMedCrossRefGoogle Scholar
- Koch PJ, Mahoney MG, Ishikawa H, Pulkkinen L, Uitto J, Shultz L, Murphy GF, Whitaker-Menezes D, Stanley JR. Targeted disruption of the pemphigus vulgaris antigen (desmoglein 3) gene in mice causes loss of keratinocyte cell adhesion with a phenotype similar to pemphigus vulgaris. J Cell Biol. 1997;137(5):1091–102.PubMedPubMedCentralCrossRefGoogle Scholar
- Kountikov EI, Poe JC, Maclver NJ, Rathmell JC, Tedder TF. A spontaneous deletion within the desmoglein 3 extracellular domain of mice results in hypomorphic protein expression, immunodeficiency, and a wasting disease phenotype. Am J Pathol. 2015;185(3):617–30.PubMedPubMedCentralCrossRefGoogle Scholar
- Seishima M, Esaki C, Osada K, Mori S, Hashimoto T, Kitajima Y. Pemphigus IgG, but not bullous pemphigoid IgG, causes a transient increase in intracellular calcium and inositol 1,4,5-triphosphate in DJM-1 cells, a squamous cell carcinoma line. J Invest Dermatol. 1995;104(1):33–7.PubMedCrossRefGoogle Scholar
- Tsang SM, Brown L, Lin K, Liu L, Piper K, O’Toole EA, Grose R, Hart IR, Garrod DR, Fortune F, Wan H. Non-junctional human desmoglein 3 acts as an upstream regulator of Src in E-cadherin adhesion, a pathway possibly involved in the pathogenesis of pemphigus vulgaris. J Pathol. 2012b;227(1):81–93.PubMedCrossRefGoogle Scholar