Mechanotransduction through hemidesmosomes during aging and longevity

The interface between structural cells and their surrounding extracellular matrix (ECM) is highly dynamic and heterogeneous. This heterogeneity stems not only from the fact that every cell type produces its own unique ECM but also that this ECM is dynamically modulated depending on the cellular state or demand of the cell (Ewald, 2020; Sacher et al., 2021; Statzer and Ewald, 2020). The cell–ECM interface functions as a border between the intra- and extra-cellular environment, but it also delineates the different time scales of the proteome – the intracellular proteome with protein half-lives of typically in the range of minutes to hours and the much longer half-lives of ECM proteins of weeks to years (Table 1) (Jariwala et al., 2022). The strength and continuity of interactions at these interfaces, which reciprocally transmit biochemical and mechanical information between the cell and ECM, heavily determine the functionality of tissues under homeostatic conditions and their ability to heal after injury (Has and Nyström, 2015). Tissues that interact with each other and need to withstand or transmit mechanical challenges are equipped with specialized structures enforcing cell–ECM cohesion and uninterrupted reciprocal information exchange. At the cell surface, integrins bind the adjacent ECM and are essential to transduce and translate mechanical cues into biochemical signaling (i.e. mechanotransduction or mechanosignaling) (Humphrey et al., 2014). Intracellularly, these integrin ECM receptors associate with the actomyosin cytoskeleton, forming focal adhesions, which are generally regarded as major sites of physical force transfer between the intra- and extra-cellular environment (Has and Nyström, 2015).

Intermediate filaments are less recognized in cellular force transfer because of their more elastic and tension-bearing nature than actin filaments (Wagner et al., 2007). Nevertheless, the role of intermediate filaments in mechanosignaling is emerging. In epithelia, intermediate filaments are connected to the ECM via hemidesmosomes (HDs) (Fig. 1). HDs are large multiprotein anchoring and signaling platforms that connect intermediate filaments to basement membrane. Broadly speaking, they come in two flavors, the type I HDs present in stratified and pseudostratified epithelia, and type II HDs present in simple epithelia (Litjens et al., 2006). Type I and type II HDs share integrin α6β4 and plectin as core HD components (Fig. 1) (Walko et al., 2015). In addition to this core, type I HDs are equipped with the plakin family protein dystonin (also known as bullous pemphigoid antigen 1 isoform e, BPAG1e, or BP230), the transmembrane collagen XVII (also known as bullous pemphigoid antigen 2, BPAG2, or BP180) and CD151, which is a tetraspanin (Fig. 2) (Has and Nyström, 2015; Nishie et al., 2011; Sterk et al., 2000). The integration of these proteins in type I HDs is believed to increase their mechanical stability as a consequence of the tissues facing higher mechanical challenges than sites where type II HDs are present (Walko et al., 2015). However, type I HDs are also found in the pancreatic ductal epithelium, an epithelium that is not expected to experience high mechanical or frictional challenges (Laval et al., 2014). Thus, this might suggest that the function of a mature HD and its components is broader than providing secure cell–ECM adhesion.

Although recently HDs have emerged as important for tissue regeneration, the role of HDs in aging is less established. In this Review, using examples from C. elegans, mice and humans, we will discuss the non-structural functions of hemidesmosomes and hemidesmosome-like structures, as well as their importance for healthy aging. The conservation of these structures is discussed in another review (Zhang and Labouesse, 2010). We will specifically address the tissue-regenerative and mechanobiological functions of hemidesmosomes and their potential as therapeutic targets to combat age-related tissue and organism degeneration. Based on conserved functional outcomes of hemidesmosome preservation or obliteration from C. elegans to humans, we argue that hemidesmosomes are central in mechanoregulation and tissue maintenance and, thus, for healthy aging.

The principal ECM interaction partner of HDs is laminin-332 (comprising laminin α3, laminin β3 and laminin γ2 chain), which can bind directly to both integrin α6β4 and collagen XVII (Fig. 1) (Has and Nyström, 2015; Nishie et al., 2011). HDs, via laminin-332, also link to collagen VII; we refer to this whole interactome encompassing the core HD and its direct and indirect interactions partners as the extended HD (Figs 1 and 2) (Hartwig et al., 2007), which collectively ensures the stable attachment of various epithelial basement membranes by forming anchoring fibrils that entrap collagen fibrils in the more cell-distal interstitial ECM (Has and Nyström, 2015). These proteins also perform essential functions outside of assuring epidermal-to-dermal adhesion, such as modulating keratinocyte migration during wound healing and tissue regeneration (Frijns et al., 2010; Hartwig et al., 2007; Jacków et al., 2016a; Nyström et al., 2013; Raymond et al., 2005), a situation when HDs are dismantled. Thus, cell–matrix anchoring complexes and their member proteins promote tissue preservation both under homeostasis and after injury. Their functions during these two states are, however, diametrically opposite – under homeostatic conditions, they ensure tissue cohesion, whereas, after injury, they support movement and regeneration. This functional dualism is, in part, achieved by post-transcriptional modifications (Frijns et al., 2010; Jacków et al., 2016b; Muir et al., 2016; Nishie, 2020; Rousselle and Beck, 2013). Serine or tyrosine phosphorylation of the cytoplasmic tail of integrin β4 induces integrin α6β4 dissociation from plectin and promotes HD disassembly that might facilitate migration (Germain et al., 2009; Kashyap et al., 2011; Seltmann et al., 2015; Wilhelmsen et al., 2007). Keratin filaments have been suggested to stabilize HDs, and phosphorylation of keratins has been linked to keratin filament disruption (Sawant et al., 2018; Seltmann et al., 2015). For collagen XVII and laminin-332, extracellular proteolysis via members of the a disintegrin and metalloproteinase (ADAM) family, matrix metalloproteinases (MMPs) or astscin-like proteinases changes their cell migration-modulating properties (Muir et al., 2016; Nishie, 2020; Rousselle and Beck, 2013). When laminin-332 is associated with HDs, it is usually processed in the N-terminal short arm of its γ2 chain and the C-terminus of its α3 chain, with the subsequent release of the LG4 and LG5 domains, suggesting that the processing transforms laminin-332 from a migration-supporting to an adhesion-supporting molecule. The ectodomain of collagen XVII can undergo ADAM-mediated shedding and be deposited in the ECM (Jacków et al., 2016b; Nishie et al., 2011). Shedding can be stimulated by reactive oxygen species (ROS) (Franzke et al., 2009), but the consequences of this are not completely understood. Analysis of a mouse model deficient in ectodomain shedding indicates that removal of the ectodomain does not impact HD formation or appearance but appears to slow wound re-epithelialization (Jacków et al., 2016b).

There is a cross-regulation of HDs and focal adhesions that can influence cell migration. Intracellularly, plectin connects to keratin intermediate filaments; however, plectin is also an F-actin-binding protein (Geerts et al., 1999) that modulates actin dynamics (Andrä et al., 1998). This is one means of an interplay between the actomyosin cytoskeleton and intermediate filaments and, by extension, between focal adhesions and HDs. Such exchanges, which regulate intermediate filament and vice versa actin filament activity, occur during wound healing when immature HDs and focal adhesions ‘treadmill’ during epithelial cell migration but also during force sensing under homeostatic conditions (Pora et al., 2019, 2020; Wang et al., 2020). In part, this might be mediated through a shift of integrin α6β4 from being associated to HDs to being released from them through a mechanism, whereby a scarcity of extracellular HD ligands prevents integrin α6β4 from assembling into mature HDs (Chaudhuri et al., 2014).

Genetic deficiency of components of the extended HD causes mechanobullous skin blistering diseases; these comprise three out of the four types of genetic epidermolysis bullosa (EB): EB simplex, junctional EB and dystrophic EB (Bardhan et al., 2020; Nyström et al., 2013; Rosa et al., 2019; Shirai et al., 2019; Solis et al., 2021; Tayem et al., 2021). The diseases manifest in fragile skin or other epithelia with blistering upon minor mechanical challenges. These symptoms are linked to the tissue-stabilizing functions of HDs; however, in addition, humans with EB and mouse models with HD-affected EBs frequently display manifestations that are indicative of a defective tissue-regeneration capacity and premature aging, including altered epidermal differentiation, slow healing or chronic wounds and alopecia (Bardhan et al., 2020; Nyström et al., 2013; Rosa et al., 2019; Shirai et al., 2019; Solis et al., 2021; Tayem et al., 2021). There are obvious challenges in making conclusions regarding primary and subsequent consequences of HD deficiency from manifestations of EB, as the disease in a natural setting involves many different external aggressors of disease (friction, microbiota, etc.) (Fuentes et al., 2023; Nyström et al., 2021). However, when collectively considering preclinical data and clinical observations, we believe there is strong support for diverse functions of the extended HD during tissue homeostasis and regeneration.

Because of their relatively long residence time in tissue, HD and ECM proteins accumulate multiple modifications or localized damages, for instance, from spontaneous glycosylation (i.e. advanced glycation end products, AGEs) or ROS (Jariwala et al., 2022). AGE deposits occur in both photo- and chrono-aged skin, although the pattern differs, with dermal collagen fibrils, particularly in the deeper dermis, accumulating more AGEs in chrono-aged skin and elastin being affected in photo-aged skin (Jeanmaire et al., 2001). Moreover, basement membrane collagens and laminins also accumulate AGEs (Roig-Rosello and Rousselle, 2020). Although careful investigations of non-enzymatic glycation of the epidermal basement membrane with aging are lacking, proteomics has revealed a decreased solubility of epidermal basement membrane proteins, including collagen IV, collagen VII and laminin γ1-chain-containing laminins (McCabe et al., 2020), which could be a reflection of accumulation of AGEs. Such accumulation might compromise protein–protein interactions and, consequently, the stability of skin tissue junction structures, including HDs (Charonis et al., 1990; Goldin et al., 2006; Paul and Bailey, 1999). Collectively, age-related changes dysregulate the proteolytic balance and the biomechanical properties of ECM, leading to generally stiffer but more brittle collagen networks (Stammers et al., 2020). Increased stiffness of the basement membrane and adjacent papillary dermal extracellular matrix during aging is further linked to changes in the abundance of components that modulate stiffness, such as a reduction in the level of collagen XIV and an increase of that in tenascin-c (Koester et al., 2021; Miroshnikova et al., 2016). From a tissue-regenerative perspective, age-related stiffening of the basement membrane and the adjacent papillary dermal extracellular matrix might reduce its renewal potential by lowering the fitness of skin stem cells (Koester et al., 2021). This occurs through mechanical stress, which reduces chromatin accessibility, silences bivalent promoters and represses transcription of genes, including Lef1 and multiple Bmps, Wnts and Tcfs, all of which are essential for the self-renewal of stem cells (Koester et al., 2021).

The HD component collagen XVII is essential for the fitness of skin stem cells (Liu et al., 2019; Matsumura et al., 2016). During aging, the abundance of collagen XVII, which is a transmembrane collagen, has been shown to be increasingly lost through proteolysis, including owing to the activity of neutrophil elastase, ADAMs and MMPs (Liu et al., 2019; Matsumura et al., 2016) (Fig. 3). Phenotypically in the skin, collagen XVII-positive epidermal basal cells tend to divide parallel to the basement membrane, whereas cells low or deficient in collagen XVII undergo perpendicular cell division, generating one basal cell and one suprabasal differentiating cell. Thus, collagen XVII presence is linked to the maintenance of the conservation of the epidermal stem cell pool (Liu et al., 2019; Matsumura et al., 2016). With aging, for instance, through microenvironmental challenges such as ROS and other factors that increase proteolysis, collagen XVII is lost, and, consequently, the regenerative potential of stem cells is reduced (Liu et al., 2019; Matsumura et al., 2016). In human skin, loss of HD-associated proteins, which are either completely or partially present in the extracellular space, including laminin-332, collagen VII and XVII, occurs during both chrono- and photo-aging (Langton et al., 2016, 2021) (Fig. 3). Multiple mechanisms could contribute to this, including enhanced proteolysis, which could disturb the balance between degradation and synthesis, leading to depletion of specific proteins, for instance, laminin-332; this, in turn, might modulate basement membrane stiffness, potentially leading to a stiffer environment with reduced laminin-332 levels that would then prevent integrin α6β4 assembly into HDs (Chaudhuri et al., 2014; Koester et al., 2021; Liu et al., 2019; Matsumura et al., 2016). In addition, the epidermal basement membrane and its individual components are also targets of AGEs, which could impact their macromolecular arrangement and interactions, including giving rise to increased stiffness and brittleness as observed in the interstitial ECM (Roig-Rosello and Rousselle, 2020). Furthermore, age-related changes not only affect the HDs but also their interaction partners. For instance, collagen VII, which forms the hanging structures to bind elastin at the level just below the epidermal basement membrane, becomes fragmented, and its colocalization with elastin is lost during old age (Tohgasaki et al., 2022) (Fig. 3). Similarly, even collagen IV, which provides the basement membrane with mechanical stability (Pöschl et al., 2004), has been reported to be reduced in the epidermal basement membrane during aging (Vázquez et al., 1996). This could affect its mechanical properties and, consequently, mechanosignaling through the already degenerating HD.

Given that HDs are sensitive to environmental changes (i.e. AGEs, ROS and UV) and proteases, including neutrophil elastase and MMPs released from inflammatory cells and stressed structural cells (Liu et al., 2019, 2022), the integrity of HDs is an important readout for tissue homeostasis and maintenance of cellular repair, not only in the skin but in all organs where they or similar structures are present. Given that HDs are present at the cellular interface linking tissues, it makes sense that they are at the nexus regulating cellular homeostasis across tissues. There is tantalizing evidence in model organisms such as Caenorhabditis elegans suggesting that the mechanotransduction function of HDs is central to tissue homeostasis and healthy longevity (Teuscher et al., 2022 preprint). To provide some insights into the role of HDs in organismal aging, we discuss below their functions in C. elegans and the possible relationship to mammalian aging.

The nematode C. elegans is a pioneering model organism to study ECM protein homeostasis during aging and longevity (Ewald, 2020; Statzer et al., 2021, 2022). C. elegans ages similarly to mammals, but within 3 weeks. Furthermore, the transparency of C. elegans also allows for non-invasive, in vivo monitoring of tagged ECM proteins incorporated into matrices during aging. C. elegans has two main ECMs, basement membranes, sheet-like structures surrounding organs, and the cuticle that forms the exoskeleton (Kramer, 2005; Page and Johnstone, 2007). These ECMs are connected by HD-like structures (Fig. 2) (Zhang and Labouesse, 2010). Although direct experimental proofs for protein–protein interactions are missing, based on accumulating genetics, microscopy and yeast-two-hybrid data, a speculative model of the C. elegans HD-like composition has been suggested (Cox and Hardin, 2004). The cuticular ECM binds transmembrane receptors MUA-3 (fibrillin in mammals) and MUP-4 (matrilin), which connect to VAB-10 (plectin) in the apical hypodermis plasma membrane (Fig. 2). Intermediate filaments span through the hypodermis to the basal site, where again VAB-10 binds to the LET-805 (myotactin) receptor for anchoring into the basement membrane ECM (Fig. 2). The basement membrane collagen IV and UNC-52 (perlecan) bind to the heterodimer formed by PAT-2 (integrin α) and PAT-3 (integrin β) at the muscle cell surface (Fig. 2) (Zhang and Labouesse, 2010). Together, these components make up the C. elegans HD-like structure (ceHD).

For many cases where mammals have a whole protein family, C. elegans only contains a single protein ortholog, as well as, sometimes, a more ancestorial protein, such as DAF-2, which resembles the insulin receptor and IGF-1 receptor in its more ancestral form (Ewald et al., 2018). Here, we speculate, based on orthology, that the ceHD might be an ancestral structure that combines two conserved mechano-regulators: mammalian HDs and focal adhesions. The upper part of the ceHD resembles HDs found in mechanical exposed epithelial tissue (e.g. skin, esophagus, intestine), while its lower part, from the basement membrane to the muscle, resembles focal adhesions (Fig. 2). This view is supported by the fact that, as discussed in mammals, there is dynamic crosstalk between HDs and focal adhesions. For instance, during cell migration and wound healing, when HDs cluster as ordered arrays interspersed by actin-associated focal adhesions, the more stable HDs work together with, the more transient focal adhesions to generate cell motility (Hatzfeld and Magin, 2019; Molder et al., 2021). We also speculate that the tetraspanin TSP-15, which has similarity to CD151, might also associate with ceHDs, based on its expression pattern and localization to ceHDs as well as the fact that mutations in tsp-15 result in ceHD phenotypes that are reminiscent of mutations in ceHD components (Moribe et al., 2004).

There are several genetic observations during development that point to a reciprocal role of ceHDs in mechanosensation and ECM remodeling. The ceHDs are required for contractile force transmission for normal locomotion of C. elegans (Hresko et al., 1999). Muscle contractions are essential for embryonic elongation by mechanical coupling and for proper assembly of these filamentous ceHD-containing structures from the muscle basement membrane through the hypodermis to the outer cuticle that acts as a soft exoskeleton (Zhang et al., 2011). Muscular contraction induces mechanical force transmission that promotes the reorganization of actin bundles in the hypodermis (Lardennois et al., 2019), indicating there is a mechanotransductive remodeling that adapts the inner and outer tensions and forces to these scaffolding structures.

Besides having a role in force transmission and cytoskeleton adaptation, ECM proteins in the vicinity of ceHDs also influence ECM remodeling. For instance, the zona pellucida domain-containing ECM proteins NOAH-1 and NOAH-2 are important for maintaining mechanoreceptor potentials and cuticular ECM remodeling (Frand et al., 2005; Vuong-Brender et al., 2017). During C. elegans development, cuticular ECM remodeling occurs during molting. Intriguingly, loss of unc-52, pat-3 or unc-95 (paxillin in mammals) results in ECM-associated molting defects (Frand et al., 2005; Zaidel-Bar et al., 2010). By contrast, mutations in the muscle myosin unc-54, which is important for muscle contraction, or mutations in unc-13, important for neurotransmitter release, did not induce cuticular collagen expression (Broday et al., 2007). The surface of the C. elegans cuticle consists of annuli (circumferential ridges) and, at the lower points of these annuli (i.e. the furrows), the cuticle attaches to ceHDs (McMahon et al., 2003). Scanning electron microscopy has revealed that there are abnormal, branched or flat cuticular annuli in unc-52 mutants but not in unc-13 mutants, which show neurotransmitter-release defects (Broday et al., 2007), suggesting that defects in unc-52 (perlecan), which is important for ceHD function, can impair proper cuticle ECM morphology. Furthermore, pxn-2 (peroxidasin) is involved in promoting sulfilimine crosslinks of the basement membrane collagen IV, thereby regulating its mechanical properties. Defects in pxn-2 can be bypassed by mutations in ceHD components, such as let-805, vab-10 and unc-52 (Gotenstein et al., 2018), suggesting that changes in either the cuticular or basement membrane ECM are both detected and mediated by ceHDs.

In alignment with the idea that ceHDs coordinate an adaptation of ECM remodeling through the mechanical induction of gene expression, what are the physiological consequences of mechanical manipulation on the worm? Mechanical compression of ceHDs by placing C. elegans into hypergravity impairs the migration of motor neurons over the muscle, and mutations in vab-10, unc-52 and other ceHD components rescue neuronal migration (Kalichamy et al., 2020). Vice versa, stretching ceHDs unmasks the SH3 domain of VAB-10, enabling mechanosensitive signaling essential for embryonic elongation (Suman et al., 2019). Interestingly, hydrostatic pressure increases the production of cuticular collagen col-107 mRNA, as well as increasing lifespan (Watanabe et al., 2020). These findings suggest that mechanical compression affects a wide range of physiological implications from development to aging. This might not be specific to the worm, as similarly, hypogravity owing to a three-month space flight induces collagen turnover in murine skin (Neutelings et al., 2015), suggesting ECM turnover. Future studies are needed to mechanistically link such ECM turnover to specific cell physiological processes and perhaps improve age-related changes.

The key question is how do changes in HDs affect longevity? During C. elegans aging, a dissociation of basement membrane collagen IV from its integrin receptor has been observed (Teuscher et al., 2022 preprint), suggesting that there is an age-dependent loss of basement membrane and ceHD integrity. During the early adulthood of C. elegans, the protein homeostasis network collapses, which can be slowed by longevity interventions (Ben-Zvi et al., 2009; Labbadia and Morimoto, 2014). Similarly, longevity interventions are able to slow this age-dependent uncoupling of ECM from its receptor (Teuscher et al., 2022 preprint). We favor the idea that loss of proteostasis drives the loss of ceHD tension-coupling given that a less-stable perlecan mutant, which serves as a readout of ECM proteostasis (Ben-Zvi et al., 2009), accelerates the dissociation of collagen IV from its integrin receptor (Teuscher et al., 2022 preprint). The mutated form of perlecan is disorganized and forms aggregates, which can be removed by enhanced chaperone, autophagy or proteasomal activities (i.e. enhanced protein homeostasis) as seen in longevity mutants or with drugs (Alavez et al., 2011). Furthermore, calreticulin crt-1 (CALR), which regulates the folding and protein levels of UNC-52 (Zahreddine et al., 2010), is also essential for ceHD-mediated collagen homeostasis (Teuscher et al., 2022 preprint). Thus, longevity interventions might improve the integrity of HDs and so ensure their dynamic adaptation and signaling.

Detachment of a cell from the ECM has been shown to lead to apoptosis or loss of cellular identity (He et al., 2015), and several recent findings point toward the notion that HDs might directly regulate cellular homeostasis and tissue adaptation. For instance, the chaperone HSP-43 is constitutively expressed and stored at HDs (Fu et al., 2020). Upon heat shock, HSP-43 is released from the HDs in a fast response, analogous to the release of small heat-shock chaperones that are associated with desmosomes and focal adhesions and help to ensure resistance against heat-induced damage in mammalian cells (Fu et al., 2020). Disturbances or disintegration of ceHDs induce V-ATPase and activate lysosomes to facilitate the turnover of ECM during C. elegans development (Miao et al., 2020). Hemidesmosomal integrity is also linked to mitochondrial ATP production and muscle protein synthesis (Etheridge et al., 2015). Moreover, disruption of the cuticle or the upper part of ceHDs (mup-4) through silencing of genes essential for ceHD formation induces anti-pathogenic responses (Zhang et al., 2015). Similarly, in primary human epidermal keratinocytes, disruption of HDs by small molecules, antibodies or siRNA targeting stimulates the production of antimicrobial peptides (Zhang et al., 2015). Moreover, disruption of the cuticle or the upper part of ceHDs (mup-4) induces a pathogen response (Zhang et al., 2015). Similarly, in primary human epidermal keratinocytes, disruption of HDs induces an antimicrobial peptide pathogen response (Zhang et al., 2015). Intriguingly, in C. elegans, a ceHD feedback loop that enhances collagen expression leads to cellular adaptation of the cytoskeleton, metabolic, oxidative stress response and pathogen response proteins (Teuscher et al., 2022 preprint), all processes that are important in maintaining cellular homeostasis and thus provide means for longevity interventions.

In an extensive lifespan screen involving over 55,000 C. elegans, we identified the mechano-responsive transcriptional co-activator Yes-associated protein YAP-1 (YAP1, but commonly known as just YAP, in mammals) as being essential for longevity (Teuscher et al., 2022 preprint). YAP1 is a transcriptional regulator that, upon its activation by mechanical cues, such as increased pressure or stiffer ECM, translocates from the cytosol to the nucleus to modulate gene expression (Elosegui-Artola et al., 2017; Pelissier et al., 2014). In C. elegans, YAP-1 is found in the cytoplasm and nucleus under normal conditions (Iwasa et al., 2013). Reducing Hippo signaling drives YAP-1 to accumulate in the nucleus, a response conserved in mammals. In C. elegans, yap-1 is required for cytoskeletal responses to stress and pathogens (Iwasa et al., 2013; Lee et al., 2019a; Ma et al., 2020). Interestingly, under conditions that promote longevity, such as reduced insulin or IGF-1 signaling (Ewald et al., 2015), YAP-1 is found at ceHDs, but this is not the case under normal conditions (Teuscher et al., 2022 preprint). Furthermore, YAP-1 responds to externally applied pressure and promotes an increase in collagen expression in a process that requires intact ceHDs (Teuscher et al., 2022 preprint). This suggests that YAP-1 localization to ceHD might serve as a mechanical sensor for muscular force pulling across tissues to the cuticle to adapt ECM strength to mechanical loading (Teuscher et al., 2022 preprint). Interestingly, several of the phenotypes we observed in C. elegans are conserved in mammals. For instance, YAP is involved in heat stress response (Luo et al., 2020), cytoskeletal dynamics (Morikawa et al., 2015) and pathogen response (Wang et al., 2017).

In human cell cultures, YAP translocates to the nucleus to induce expression of its target genes in response to the stiffness of the ECM (Elosegui-Artola et al., 2017), which typically increases, or becomes more heterogeneous, with age, leading to dysregulation of YAP-mediated mechanotransduction and induction of cellular senescence and altered stem cell differentiation observed in vitro and in vivo, respectively (He et al., 2019; Pelissier et al., 2014; Xie et al., 2013). Although activation of YAP has beneficial effects, such as promoting the regeneration of the heart, liver, intestine and skin in old mice, prolonged hyperactivation of YAP is found in human cancers and can also drive cancers in mice, as well as skeletal muscle degeneration, suggesting that YAP activity needs to be tightly regulated in mammalian cells. Consequently, there may be potential to regulate YAP activity for regenerative medicine purposes (Moya and Halder, 2019); however, care has to be taken to consider cell- or tissue-specific and time-limited stimulations to limit risks and increase efficacy.

HDs, via their associated intermediate filament system, have been shown to resist mechanosignaling and force generation exerted through focal adhesions and actomyosin (Wang et al., 2020). Genetic evidence that HDs indeed repress these events comes from naturally occurring mutations in HD components that give rise to diseases, including, as above introduced, certain types of EB. On a cellular level, such deficient keratinocytes display enhanced activity of the actomyosin cytoskeleton, as displayed by increased focal adhesions and mechanosignaling (Wang et al., 2020). One mechanism by which HDs oppose mechanotransduction is through integrin α6β4 and plectin-mediated decrease of FAK (also known as PTK2)–phosphoinositide 3-kinase (PI3K) signaling, which increases the inactivating serine phosphorylation of YAP (Lee et al., 2019b; Wang et al., 2020). Under low-stress conditions, YAP associates with integrin α6β4, whereas under high stress, YAP becomes associated with integrins that contain a β1 subunit; this prompts it becoming associated with the actomyosin cytoskeleton, which might facilitate subsequent nuclear translocation via F-actin-mediated opening of the nuclear pore (Lee et al., 2019b; Wang et al., 2020). Furthermore, YAP activity might also, in such situations, be concomitantly enhanced by activated FAK signaling (Lachowski et al., 2018; Lee et al., 2019b; Wang et al., 2020).

In mammary epithelial cells, stiffness has been shown to reduce HD assembly (Chaudhuri et al., 2014). However, there is evidence that not only mechanical but biochemical attributes, including specific HD-associated ECM proteins and their densities, are essential for protection from aging and the maintenance of healthy stem cells (Liu et al., 2019; Matsumura et al., 2016; Rosa et al., 2019). Conceptually, increasing the density of an integrin α6β4 ECM ligand (in the referenced study laminin-111, comprising α1, β1 and γ1 chains) in mammary epithelial cells was shown to overcome mechanically induced HD repression (Chaudhuri et al., 2014). Evidence directly related to pathophysiology comes from the depletion of epidermal stem cells occurring in individuals with laminin-332-deficient junctional EB (Rosa et al., 2019). Laminin-332, via integrin α6β4, mediates sustained YAP activity in epidermal stem cells (Rosa et al., 2019), and depletion of the laminin-332–integrin α6β4 axis abolishes YAP activity, resulting in epidermal stem cell loss. These findings might be somewhat at odds with reports that HDs and integrin α6β4 repress YAP signaling (Lee et al., 2019b; Wang et al., 2020). However, they suggest that the outcome depends on the context, such as exact experimental conditions and tissue state-specific outcomes. Specifically, in the above work (Rosa et al., 2019), keratinocytes were maintained in a low-Ca²⁺ medium, which does not stimulate excessive in vitro HD formation, whereas in Wang et al. (2020), cells were cultured in high-Ca²⁺ DMEM, which promotes HD assembly. Nevertheless, despite the experimental and contextual differences, the data collectively support the involvement of HDs in the regulation of YAP-mediated signaling.

The exploration of HD in mechanosensing and longevity is only in its infancy, as thus far, focal adhesions have been the main focus in investigating mechanosensing and mechanotransduction (Sarker et al., 2020). Interestingly, an earlier computer simulation study found that focal adhesions were a central network associated with human longevity (Wolfson et al., 2009). Considering the reciprocal interactions between focal adhesions and HDs, and the genetic evidence in C. elegans for ceHD function in promoting longevity, it is reasonable to assume that mammalian HDs also have a central role in aging and longevity. This is directly supported by the observation of a positive correlation between collagen XVII depletion and impaired stem cell fitness in the human skin (Liu et al., 2019; Matsumura et al., 2016). The correlation in the skin is not limited to collagen XVII but destabilized HDs and their interacting ECM on multiple levels. For instance, facial skin biopsies from humans over 60 years old show disassociation of β4 integrin from HDs and decreased collagen IV levels compared to younger individuals (Varlet et al., 1998), whereas collagen VII is not only reduced but also shows less colocalization with elastic fibers in aged skin (Tohgasaki et al., 2022). Consequently, restoring HDs and their interacting structures might promote the retention of tissue functionality during aging and ensure its longevity.

The successful development of therapies targeting genetic diseases caused by HD or associated HD-structure deficiencies, including gene-therapy approaches, gene-editing, cell therapy, small molecules and protein-replacement therapy (Chakravarti et al., 2022; Payne, 2022; Peking et al., 2018) have generated a toolbox that could also potentially be applied during natural aging. Indeed, studies have shown that systemic or topical protein replacement therapies using recombinant ECM proteins are feasible. For instance, a recent clinical trial (NCT03536143; https://www.clinicaltrials.gov/study/NCT03536143) on recessive dystrophic EB has demonstrated that it is feasible and safe to increase the production and subsequent deposition of collagen VII in wounded skin by topical viral gene therapy (Gurevich et al., 2022), suggesting that targeting the age-dependent decline of HDs might be possible as a therapeutic intervention. A current limitation of such therapies is the lack of delivery through the intact epidermis, requiring either treatment of wounded tissue or intradermal injections.

The concept of HDs as main regulators of general ECM homeostasis through the coordination of biochemical and mechanical signaling is supported by genetic diseases linked to defective HD functionality, in which progressive ECM remodeling is generally observed (Nyström et al., 2018). The consequences of loss of junction structures that support tissue integrity are not only seen in the skin but are a universal feature – fibrosis and ECM remodeling are hallmarks of genetic muscular dystrophies (Zhou and Lu, 2010). The ECM remodeling that occurs in these conditions is commonly linked to stiffer tissue arising from excessive collagen production or altered collagen organization manifested by thickened and more parallel arrangement of collagen fibrils (Bernasconi et al., 2021; Sahani et al., 2022). The mechanisms underlying these changes are multiple and include inflammation evoked by tissue damage (Bernasconi et al., 2021; Sahani et al., 2022). However, the consequences of loss of HD-regulated tissue integrity and ECM homeostasis are both dysregulated biochemical and mechanical outside-in (deficient ECM) and inside-out (integrin α6β4 and collagen XVII deficiency) feedback loops. Thus, HDs are not only structural components of tissue but actively ensure tissue homeostasis and, as such, are centrally placed in aging and longevity. Taking this knowledge to therapeutic reality will require an increased mechanistic understanding of HD in tissue homeostasis and detailed knowledge of processes driving HD destruction, together with novel, safe and efficacious methods for the topical delivery of HD rejuvenation therapeutics.

We thank Dr Ingrid Hausser for help with the TEM imaging and the Nyström and Ewald labs for critical reading.