If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Centre for Cell Biology and Cutaneous Research, Blizard Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK
Centre for Cell Biology and Cutaneous Research, Blizard Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK
Centre for Cell Biology and Cutaneous Research, Blizard Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK
Centre for Cell Biology and Cutaneous Research, Blizard Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK
Centre for Cell Biology and Cutaneous Research, Blizard Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK
Centre for Cell Biology and Cutaneous Research, Blizard Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK
Centre for Cell Biology and Cutaneous Research, Blizard Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK
Centre for Cell Biology and Cutaneous Research, Blizard Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK
Centre for Cell Biology and Cutaneous Research, Blizard Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK
5 These senior authors contributed equally to this work.
Ryan F.L. O'Shaughnessy
Footnotes
5 These senior authors contributed equally to this work.
Affiliations
Livingstone Skin Research Centre for Children, UCL Institute of Child Health, London, UKDepartment of Immunobiology, UCL Institute of Child Health, London, UK
5 These senior authors contributed equally to this work.
Daniele Bergamaschi
Correspondence
Correspondence: Daniele Bergamaschi, Centre for Cell Biology and Cutaneous Research, Blizard Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London E1 2AT, UK.
5 These senior authors contributed equally to this work.
Affiliations
Centre for Cell Biology and Cutaneous Research, Blizard Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK
Epidermal keratinocytes migrate through the epidermis up to the granular layer where, on terminal differentiation, they progressively lose organelles and convert into anucleate cells or corneocytes. Our report explores the role of autophagy in ensuring epidermal function providing the first comprehensive profile of autophagy marker expression in developing epidermis. We show that autophagy is constitutively active in the epidermal granular layer where by electron microscopy we identified double-membrane autophagosomes. We demonstrate that differentiating keratinocytes undergo a selective form of nucleophagy characterized by accumulation of microtubule-associated protein light chain 3/lysosomal-associated membrane protein 2/p62 positive autolysosomes. These perinuclear vesicles displayed positivity for histone interacting protein, heterochromatin protein 1α, and localize in proximity with Lamin A and B1 accumulation, whereas in newborn mice and adult human skin, we report LC3 puncta coincident with misshaped nuclei within the granular layer. This process relies on autophagy integrity as confirmed by lack of nucleophagy in differentiating keratinocytes depleted from WD repeat domain phosphoinositide interacting 1 or Unc-51 like autophagy activating kinase 1. Final validation into a skin disease model showed that impaired autophagy contributes to the pathogenesis of psoriasis. Lack of LC3 expression in psoriatic skin lesions correlates with parakeratosis and deregulated expression or location of most of the autophagic markers. Our findings may have implications and improve treatment options for patients with epidermal barrier defects.
The epidermis is a multilayered structure continuously renewed by keratinocytes of the basal layer that divide and differentiate to form cells of the spinous, granular, and cornified layers. The proliferating basal layer is a heterogeneous population comprising epidermal stem cells and transit-amplifying cells that have limited self-renewal capacity and undergo differentiation after a few cycles (
). In the granular layer, keratinocytes begin to lose their organelles, and express structural protein characteristic for epidermal terminal differentiation, leading to the flattening, collapse, and the eventual death of the cells. These flattened cells, corneocytes, which form the cornified layer, are rich in proteins and are embedded in a lipid matrix, giving the epidermis its water-retaining, chemical, and mechanical properties, ensuring effective epidermal barrier function (
). Macroautophagy (commonly referred to as autophagy) is a conserved catabolic process characterized by formation of intracellular double-membrane structures that degrade and recycle cytosolic proteins and organelles (
). However, most published work has been performed in monolayer keratinocyte culture that may not fully represent the stratified in vivo situation. A key regulator of epidermal development and differentiation is the acutely transforming retrovirus AKT8 in rodent T-cell lymphoma (AKT)/mechanistic target of rapamycin pathway. Downstream of AKT is mTORC1 that regulates anabolic and catabolic processes including autophagy. Hyperactivation of mTORC1 signaling has been associated with a defective epidermal barrier in psoriasis (
). Here we investigate the role of epidermal autophagy and its link with epidermal terminal differentiation.
Results
Autophagy is involved in epidermal granular layer formation
During fetal development, expression studies are possible because of the temporal separation of epidermal terminal differentiation and skin barrier formation. We examined expression patterns of ULK1, beclin 1, WIPI1, autophagy related 5-autophagy related 12 complex, and LC3 as key markers of sequential stages of autophagy in the mouse embryo (Figure 1a). Epidermal barrier formation correlates with development of the cornified layer that occurs between E15.5 and E18.5 in mice. In E16.5 mouse fetal skin, filaggrin expression confirmed activation of terminal differentiation and granular layer formation. Filaggrin is also constitutively expressed in the granular layer of adult human skin (Figure 1b and Supplementary Figure S1b online), and is further increased in fetal granular layers at E17.5 and E18.5, when AKT1 is also expressed, indicating the presence of an intact granular layer. At E15.5, before granular layer development, the epidermis consists of proliferating basal and spinous layers (Supplementary Figure S1a). At this time point, LC3 as well as other autophagy proteins are present at low levels (Figure 1a). Initiation of granular layer formation at E16.5 is accompanied by upregulation of LC3 in the uppermost layers that correspond to the early granular layers. Levels of ULK1, WIPI1, and granular ATG5-ATG12 are also further increased at E16.5, accompanied by switching-on of BECN1 expression in the basal and upper epidermal layers. This autophagy marker expression pattern is maintained until birth and persists in adult mouse and human skin (Figure 1a). In summary, expression of autophagy markers is highly upregulated at E16.5 indicating that autophagy occurs during epidermal development and differentiation, and may contribute to granular layer formation.
Figure 1Induction of epidermal terminal differentiation during fetal development is accompanied by activation of autophagy marker expression. (a) Expression of acutely transforming retrovirus AKT8 in rodent T-cell lymphoma 1 (AKT1), epidermal terminal differentiation marker, filaggrin, and autophagy proteins LC3, ULK1, WIPI1, BECN1, and ATG5-ATG12 in mouse fetal skin development. Epidermal expression levels of LC3, ULK1, and WIPI1 were quantified in n = 3 fetal mouse samples from E15.5 to E18.5. ANOVA for average intensities from n = 3 samples of epidermal LC3 (P < 0.00005), ULK1 (P < 0.0005), WIPI1 (p < 0.005), respectively. (b). Expression of epidermal terminal differentiation and autophagy markers in both adult mouse and adult human epidermis. Bar = 20 μm. Dotted line = basement membrane. ANOVA, analysis of variance; ATG5-ATG12, autophagy related 5-autophagy related 12; BECN1, beclin 1; LC3, microtubule-associated protein light chain 3; ULK1, Unc-51 like autophagy activating kinase 1; WIPI1, WD repeat domain phosphoinositide interacting 1.
Detection of epidermal autophagic vesicles in newborn mouse epidermis
We then aimed to establish the presence and localization of autophagic vesicles in normal epidermis. For this purpose we used 3-day-old mouse epidermis because this time point allows a short recovery period after birth to exclude artifacts due to neonatal starvation-induced autophagy (
). We first characterized by immunofluorescence the expression of the main autophagic markers that reveal the same patterns described during mouse epidermal development as well as adult mouse and human skin (Figure 2a). By using transmission electron microscopy we investigated the ultrastructure of the epidermal granular layer (Figure 2b) and we identified autophagic vesicles in both transitional cells and the cornified envelope (Figure 2c and d). Autophagic vesicles within the epidermal layers varied in shape (from rounded to oval-shaped) with cross-sections within the range of 300–500 nm. These data confirmed previous reports describing epidermal autophagosomes in cultured keratinocytes, and also reveal that the autophagy process occurs continuously, even in newborn mouse skin, and not only during epidermal development.
Figure 2Epidermal autophagic vesicles in epidermal layers of newborn mice skin. (a) Immunofluorescent analysis of d3 mouse back skin for the key autophagic markers LC3, ULK1, ATG5-ATG12, and BECN1. (b) Transmission electron microscopy analysis (TEM) of the granular layer of the epidermis shows degrading nuclei (N). The cornified layer in this image is to the right. (c) Double-membrane autophagic vesicles are present in the granular layer keratinocytes. (d) Autophagic vesicles are present in the cornified layer of the epidermis. (e) Another image of a double membrane putative autophagic vesicle in the basal layer of the epidermis. Dotted line (A) indicates the dermal-epidermal boundary. Bars = 50 μm (a), 2 μm (b), 500 nm (c), 200 nm (d and e). ATG5-ATG12, autophagy related 5-autophagy related 12; BECN1, beclin 1; LC3, microtubule-associated protein light chain 3; ULK1, Unc-51 like autophagy activating kinase 1.
). Under nutrient-rich conditions, mTORC1 activates anabolic processes such as protein synthesis and inhibits catabolic processes such as autophagy. To establish whether epidermal autophagy is also regulated by mTORC1 signaling, the effects of mTOR inhibitors rapamycin and torin1 were analyzed using a well-established assay for in vitro skin differentiation (
). Mouse skin explants were harvested from E15.5 foetuses when the granular layer and stratum corneum are still absent (Figure 1a and Supplementary Figure S1a). After 72 hours of drug treatment (corresponding to E18.5, the time point just before birth), inhibition of mTORC1 does not have a striking effect on filaggrin expression patterns or on epidermal granular layer formation (Figure 3a and Supplementary Figure S1c). S6 phosphorylation (S240/244), a downstream target of mTORC1, is reduced in drug-treated explant epidermis (Figure 3a) confirming effective mTORC1 inhibition. The autophagosome marker LC3 is mainly expressed in the granular layer of adult epidermis (Figure 1a) as well as in vehicle-treated explants. mTORC1 inhibition strongly increases epidermal LC3 levels (Figure 3a) and induces conversion of LC3I to LC3II (Figure 3b), a measure of autophagic activity. The expression of WIPI1, present as puncta in the upper layers of vehicle-treated fetal explants, is upregulated (Figure 3a) as is ULK1 expression. Basal-lower spinous layer expression levels of ATG5-ATG12 and BECN1 were also increased (Supplementary Figure S1d). Therefore, mTORC1 inhibition increases expression of essential autophagy proteins and upregulates LC3 processing in the granular layer, consistent with induction of autophagy in the upper epidermis. Treatment in the last 4 hours of mTORC1 inhibition with bafilomycin A1—which prevents fusion of lysosomes with autophagosomes (
)—confirmed the integrity of the autophagic flux in the granular layers of these mouse skin explants (Figure 3c).
Figure 3mTOR inhibition upregulates constitutive autophagy in terminally differentiating keratinocytes. (a) Expression of epidermal S6 phosphorylation, terminal differentiation, and autophagy markers in mouse fetal explants isolated at E15.5 and treated with vehicle, 2.5 μM Torin1 or 5 μM rapamycin for 72 hours. *Two-tailed paired t-test (P < 0.05) quantified for four different fields of view. (b) Western blots analysis of the above (a) treated explant cultures. (c) Western blots analysis showing autophagic flux of Torin1 (2.5 μM for 72 hours) treated epidermal explant w/o 100 nM bafilomycin A1 (BafA1) (in the last 4 hours). (d) Immunofluorescence LC3 staining of differentiated primary keratinocytes treated with 10 nM rapamycin w/o 100 nM BafA1. (e) Measurement of autophagic flux by western blot analysis of rapamycin-treated human primary keratinocytes w/o 100 nM BafA1. Actin was used as a loading control. *Two-tailed paired t-test (P < 0.05) for LC3II/LC3I ratios in differentiated keratinocytes. (f) Measurement of autophagic flux by western blot analysis of 10 nM rapamycin-treated rat epidermal keratinocyte (REK) w/o 200 μM chloroquine. Bar = 20 μm. Dotted line = basement membrane. LC3, microtubule-associated protein light chain 3; mTOR, mechanistic target of rapamycin.
To further explore whether mTORC1 inhibition effectively upregulates autophagy in human epidermis, primary monolayer keratinocyte cultures were treated with rapamycin in the presence and absence of bafilomycin A1. Treatment with bafilomycin A1 leads to LC3II accumulation mainly in rapamycin-treated cells (Figure 3d and e), confirming inhibition of autophagic vesicle degradation. Increased LC3 expression on rapamycin combined with bafilomycin A1 (Figure 3d), as well as upregulation of LC3 turnover (Figure 3e), was mainly evident in differentiated keratinocytes. Rapamycin induces a significantly higher turnover of LC3I to LC3II in differentiated compared with undifferentiated keratinocytes. Rapamycin also strongly downregulates S6 phosphorylation confirming effective mTORC1 inhibition. These data are consistent with our observations in mouse fetal explant cultures as well as in organotypic models using rat epidermal keratinocytes (Figure 3f and Supplementary Figure S2a online) and reveal the integrity of the “epidermal” autophagic flux, but also how sensitivity of autophagy in keratinocytes could be influenced by differentiation status.
Epidermal nucleophagy during keratinocyte terminal differentiation
Having established that keratinocytes from differentiated granular layers are capable of higher levels of autophagy compared with basal proliferating cells, the role of autophagy during epidermal terminal differentiation was further investigated. Immunofluorescence analysis of autophagy proteins showed perinuclear localization of LC3, p62, and ULK1 in differentiated keratinocytes (Supplementary Figure S2b). Detailed examination of cell morphology revealed irregular or misshaped nuclei in a significant proportion of differentiated keratinocytes, whereas these were rare in undifferentiated cells (Figure 4a). Regions of degraded nuclear material in differentiated keratinocytes are replaced by LC3 aggregates that are also positive for LAMP2, a lysosomal membrane protein (Figure 4b and Supplementary Figure S3a and b online) that labels lysosomes and autolysosomes just before autophagic degradation. 4',6-diamidino-2-phenylindole staining is reduced in intensity but is present in these LAMP2/LC3 aggregates, suggestive of active degradation of DNA in these bodies (Supplementary Figure S3b). The regions of missing nuclear material may therefore be sites of high autophagic activity, a process defined as nucleophagy. p62 acts as an adaptor, recognizing and binding both ubiquitinated autophagic cargo and LC3 within autophagosomal membranes, and is commonly used as a read-out for targeted autophagic degradation of ubiquitinated cargo (
). In keratinocyte cultures, nucleophagic regions in the differentiated population are p62 positive (Figure 4c). However, most LC3 vesicles in undifferentiated keratinocytes are negative for p62 indicating that basal autophagy in proliferating keratinocytes is not targeted. These data indicate that nucleophagy is a mechanism of targeted autophagic degradation of cargo by which terminally differentiating keratinocytes might degrade their nuclei.
Figure 4Terminal differentiation in monolayer keratinocyte cultures is accompanied by targeted autophagic degradation of nuclear material—nucleophagy. (a) DAPI staining analysis reveals presence of misshaped nuclei (red arrows) within differentiated keratinocytes. Statistical significance (*P < 0.05) was measured by a two-tailed paired t-test. Immunofluorescence staining reveals (b) coexpression of LC3/LAMP2; (c) coexpression of LC3/p62; (d) coexpression heterochromatin protein 1α (HP1α)/LAMP2 specifically within regions of missing nuclear material of primary differentiated keratinocyte cultures. Red arrows indicate HP1α aggregates. (e) Lamin A (LMNA) labeling reveals that the nuclear membrane is still present between the nucleus and nucleophagic regions. (f) Detailed investigation of lamin B1 (LMNB1) and LAMP2 expression in calcium-switched human keratinocytes. Pre indicates early events whereby a small amount of nuclear material is exposed. Early indicates a time-point where coexpression of LAMP2, LMNB1, and DAPI occurs. Mid indicates a time-point where DAPI and LMNB1 expression is lost within LAMP2 positive bodies. Late indicates misshaped nuclei in the absence of LAMP2 positive bodies (like in a). (g) Western blot analysis of WIPI1 small interfering RNA (siRNA) silencing in keratinocytes. *Two-tailed paired t-test (P < 0.05). Bar = 10 μm. Small red arrows = regions of nuclear deformity (a) and (e) or regions of perinuclear protein accumulation (d). DAPI, 4',6-diamidino-2-phenylindole; LAMP, lysosomal-associated membrane protein 2; LC3, microtubule-associated protein light chain 3; WIPI1, WD repeat domain phosphoinositide interacting 1.
To confirm whether these autophagic vesicles contain nuclear material, nucleophagic keratinocytes were analyzed for expression of a histone interacting protein, heterochromatin protein 1α (HP1α), a marker of genetically inactive, tightly packed DNA found at the periphery of the nucleus (
). HP1α is mainly expressed in nuclei of undifferentiated keratinocytes (Figure 4d). However, in differentiated keratinocytes, HP1α is present in both the nucleus and cytoplasm where it forms perinuclear aggregates in regions of missing nuclear material and overlaps with LAMP2. This observation suggests that nucleophagic vesicles contain nuclear material.
To determine whether nucleophagic vesicles are in contact or coincident with the nucleus, the expression pattern of the nuclear membrane proteins lamin A (LMNA) and lamin B1 (LMNB1) was examined. Lamins are inner nuclear membrane proteins required for the maintenance of nuclear integrity (
). In both undifferentiated and differentiated keratinocytes, LMNA and LMNB1 are expressed at the nuclear membrane. Differentiating keratinocytes display an accumulation of LMNA and LMNB1 at the border between the existing nucleus and the autophagic regions (Figure 4e and f). This shows that LAMP2 positive autophagic vesicles at nucleophagic sites are cytosolic (Figure 4f) and also confirmed the involvement of lamins in epidermal autophagic nuclear degradation (
However in differentiated keratinocytes, acetylated histone H3 (Lys14) is only expressed within the intact nuclei and absent in nucleophagic regions (Supplementary Figure S3c). Thus, nucleophagy in keratinocytes is a targeted process whereby nuclear material within the nucleophagic regions differs from that within the intact nucleus where gene transcription may still occur. Nucleophagy is not a DNA damage response, as shown by lack of phosphorylated H2AX within nucleophagic regions (Supplementary Figure S3d). Moreover, neither caspase-3 cleavage (Supplementary Figure S4a online) nor Tunel staining (unshown data) was detected within nucleophagic regions confirming that no apoptotic response is triggered during nucleophagy.
To establish whether autophagy is essential for degradation of nuclei during keratinocyte differentiation, a knockdown of key autophagy proteins WIPI1 and ULK1 was performed. WIPI1 interacts with autophagic structures under conditions of starvation or mTORC1 inhibition and physiologically colocalizes with LC3 positive puncta; its expression is associated with terminal differentiation, both in culture and in the epidermis (Supplementary Figure S4b). WIPI1 knockdown significantly reduced nucleophagic cells in differentiating keratinocytes (approximately 2%) compared with controls (approximately 15%), whereas undifferentiated keratinocytes were unaffected (Figure 4g). Likewise, knockdown of ULK1, another protein essential for autophagy, also reduced the number of nucleophagic differentiated keratinocytes, although to a lesser extent than with WIPI1 silencing (Supplementary Figure S4c), reinforcing the importance of the integrity of the autophagy pathway for nucleophagy in differentiating keratinocytes. Finally, we analyzed newborn mouse and adult human skin to validate the above observation in keratinocytes and confirm whether the nucleophagy process also occurs in epidermal models. By performing confocal microscopy analysis at higher magnification of LC3 expression within the granular layer of mice newborn skin, we clearly observed that some of the LC3 puncta in the granular layer were coincident with areas of loss of nuclear material, strongly suggesting an active nucleophagy process ongoing within the granular layer (Figure 5a). Likewise, in upper layers of adult human skin, LC3 puncta were localized to the same regions of misshapen nuclei (Figure 5b). By further analyzing the granular layer ultrastructure of newborn mice skin with electron microscopy, we also identified double-membrane vesicles containing heterochromatin material corresponding to perinuclear regions where nuclear material was lost (Figure 5c and d). These data provide evidence of an epidermal autophagy process targeting nuclear integrity within the terminally differentiated layers of the skin.
Figure 5Nucleophagy in terminal differentiated layers of mouse and human epidermis. (a) Confocal microscopy of LC3 expression in d3 postnatal back skin. Left panel: low magnification of multiple LC3 positive nuclei. Right panel: optical section demonstrates LC3 bodies through nuclear invaginations (magnification 0.5 μm). Bar graph shows percentage of nuclei colocalizing in the granular layer and the rest of the epidermis. (b) LC3 accumulation within the granular layer of adult human trunk skin. Insets show 0.5 μm optical sections through one of the nuclei undergoing nucleophagy, with LC3 puncta in an invagination of the nucleus. Bar graph shows percentage of nuclei colocalizing in the granular layer and the rest of the epidermis. (c, d) Transmission electron microscopy (TEM) images of nuclei in the granular layer of d3 postnatal back skin. Arrowheads point to putative double-membrane nucleophagic bodies associated with degrading nuclei. M, mitochondrion; N, nucleus. **P < 0.001, Fishers exact test (n = 81 mouse, n = 419 human). Bars = 50 μm (a and b, low mag) 10 μm (a and b, high mag), 200nm (c), and 500nm (d). Dotted line (a) indicates the dermal-epidermal boundary. LC3, microtubule-associated protein light chain 3.
Deregulation of granular layer autophagy in psoriatic skin lesions
Psoriasis is a skin barrier defect characterized by epidermal hyperplasia, abnormal terminal differentiation of keratinocytes, infiltration of immune cells into the dermis and epidermis, and increased vascularization (
). Lesional psoriatic skin has a hyperproliferative, thickened epidermis, elongated rete ridges, and a thicker cornified layer with retained nuclei or parakeratosis (Figure 6a and b) (
), suggesting incomplete terminal differentiation. However, uninvolved skin in psoriasis has a clearly defined granular layer and corneocytes similar to healthy epidermis. To further investigate our hypothesis that autophagy is a mechanism for organelle degradation during terminal differentiation in epidermis, the expression pattern of several autophagy markers were analyzed in skin from six patients with psoriasis and compared with several samples from healthy skin. In healthy adult epidermis, LC3 is expressed in all layers of the epidermis with the strongest expression in the granular layer (Figures 1a and 6b). In psoriatic lesions, LC3 is completely absent from all layers of the epidermis (Figure 6b). In nonlesional psoriatic epidermis where cornification occurs, LC3 is present in the granular layer, but its expression is reduced. Therefore in nonlesional psoriatic skin, where keratinocyte terminal differentiation and cornification are intact, the granular layer LC3 is present but at lower levels compared with healthy epidermis. In psoriatic lesions where terminal differentiation is impaired leading to nuclear retention and improper cornification, LC3 is absent. WIPI1 is only expressed in the upper layers where cornification can occur, as in nonlesional psoriatic skin and in healthy skin. Its expression is reduced (like LC3) in parakeratotic regions of psoriatic epidermis. ULK1 is expressed in all layers of healthy epidermis with highest expression in the upper layers (Figure 6b). Its expression significantly increases in suprabasal and granular layers of nonlesional psoriatic epidermis, whereas in lesional psoriatic epidermis, ULK1 is highly expressed everywhere except in the uppermost parakeratotic layers. Likewise, BECN1, mainly expressed in the proliferating basal layer of healthy epidermis, shifts to basal and parabasal layers of nonlesional psoriatic epidermis, and suprabasal layers of lesional psoriatic skin (Figure 6b). ATG5-ATG12 is similar to BECN1 with a strong basal layer expression in healthy epidermis and no changes in nonlesional psoriatic skin, whereas a shift into suprabasal layers is observed in lesional psoriatic epidermis. These data show deregulation of autophagy markers in both lesional and nonlesional psoriatic epidermis, providing a possible reason for the defective epidermal barrier observed in this disease.
Figure 6Constitutive granular layer autophagy is deregulated in psoriasis, an epidermal barrier defect disease. (a) Hematoxylin and eosin (H&E) staining of psoriatic lesion, nonlesional psoriasis, and healthy epidermis showing thicker uppermost layer and nuclei retention in psoriatic lesion (blue vertical line), whereas the cornified layer of nonlesional psoriatic skin and healthy epidermis are completely free from nuclei (blue vertical line). In psoriatic lesions, there is no clear granular layer compared with nonlesional psoriatic skin and healthy epidermis (blue arrows). Black bar = 20 μm. (b) Immunofluorescence analysis of autophagy markers on lesional psoriatic skin compared with nonlesional psoriatic and healthy epidermis. Red vertical bars = cornified layers. White bar = 20 μm. This figure is representative of n = 6 samples of psoriatic lesions and n = 5 samples of healthy epidermis.
In this report, we demonstrate that autophagy is an active process playing a critical role in normal epidermal development and differentiation and that impaired autophagy may contribute to the pathogenesis of human disorders of epidermal differentiation such as psoriasis.
An increasing number of publications has shown that keratinocytes are capable of autophagy (
). Most of these reports have used monolayer keratinocyte culture as a research model, although this does not accurately represent late terminal differentiation in vivo. To date, an accurate profile of autophagy marker expression in healthy epidermis has not been established. The epidermal autophagy marker expression pattern suggests that autophagy plays an important role in the terminal differentiation process of the granular layer. We have recently identified a mechanism modulating the autophagy process in keratinocytes mediated by inhibitor of apoptosis stimulating protein of P53 (
). Decreased iASPP expression in keratinocytes not only triggers autophagy by derepression of the ATG5/ATG12-ATG16L1 complex, promoting autophagosome maturation, but also induces terminal differentiation. Likewise, another elegant study has recently highlighted the importance of BNIP3 in inducing autophagy but also terminal differentiation of epidermal keratinocytes (
Here we have shown that induction of autophagy marker expression during fetal development coincides with the initiation of epidermal terminal differentiation. This suggests that autophagy is constitutively active in granular layer keratinocytes and may be important for granular layer formation. Transmission electron microscopy analysis of 3-day-old mouse epidermis shows that double-membrane vesicles containing engulfed material are present in basal and granular layers. Another report has recently shown autophagosome-like structures in the granular layer of mouse skin (
). Autophagy is regulated by different cell signaling pathways of which mTORC1 is a central player. In epidermis, the AKT/mTORC1 pathway is a key regulator of epidermal development and differentiation (
), and we have previously shown that mTORC1 inhibition upregulates activity of epidermal AKT1, the AKT isoform associated with terminal differentiation (
). We were therefore interested to determine whether epidermal autophagy is also regulated by mTORC1. We observed that mTORC1 inhibition in epidermal explant cultures significantly increased the expression levels of most of the autophagy markers. These data support observations made in monolayer cultures where mTORC1 inhibition induces a striking increase in LC3 turnover in differentiated keratinocytes. Moreover, increased expression of lipidated LC3 was also detected on block of the autophagic flux, confirming integrity of the autophagy process in epidermal keratinocytes. Our findings strongly suggest that epidermal mTORC1 not only mediates AKT1 activity but also regulates constitutive granular layer autophagy. Interestingly, not all the autophagic markers localized within the granular epidermal layers. It is currently unclear whether interdependency exists between basal (BECN1, ATG5-ATG12) epidermal layer autophagy markers and upper layer epidermal markers, or whether different streams of epidermal autophagy occur within the skin. Future experiments would elucidate the role of BECN1 and ATG5-ATG12 in the epidermal basal layer and may explore the function of basal layer epidermal autophagy.
Analysis of autophagy in monolayer cultures revealed that differentiating keratinocytes undergo nucleophagy—autophagic degradation of the nucleus. This differs from micronucleophagy in which satellite nuclei are formed because of stress or genome instability and then engulfed by autophagosomes (
). Recently, a novel mechanism of nucleophagy mediated by Atg39, a nuclear envelope receptors that induces autophagic sequestration of part of the nucleus, has been identified in Saccharomyces cerevisiae (
). Although we cannot exclude at this stage whether Atg39 could also be involved in our proposed mechanism, in differentiating keratinocytes nucleophagy is characterized by accumulation of LC3/LAMP2/p62 positive autolysosomes also containing HP1α positive cargo in the regions of missing nuclear material, suggesting selective autophagic degradation is occurring. Lack of DAPI detection within the nucleophagic regions might be due to chromatin conformational changes or partial degradation of the DNA leading to reduced DAPI incorporation into DNA. Our data suggest that nuclear material within the autophagosome may be damaged or modified and bound to HP1α, which is targeted for autophagic degradation. However, the material within the nucleophagic vesicles is not positive for the marker of DNA double-strand breaks, γ-H2AX, nor for cleaved Caspase-3, suggesting that nucleophagy degradation does not trigger an apoptotic response. A similar type of nucleophagy has recently been described in the aging intestinal epithelia of Caenorhabditis elegans, probably due to aging-related changes in LMNA, and compromised nuclear integrity (
). In our monolayer keratinocyte cultures, nucleophagic vesicles are outside the partially degraded nucleus. Another model has just been reported in human primary fibroblasts also based on nuclear LC3 directly interacting with LMNB1. LMNB1 lysosomal degradation would rely on LC3-LMNB1 interaction as a general mechanism to protect the cells from oncogene-induced senescence and tumorigenesis (
). These two models of nucleophagy resemble our data in differentiating keratinocytes where we observe LMNA and LMNB1 accumulation exactly corresponding to regions of nuclear degradation. The rarity of nucleophagic events in vitro and in vivo suggests a rapid process that is dependent on the differentiating stage of keratinocytes. Finally, we reinforced the importance of the integrity of the epidermal autophagy confirmed by lack of nucleophagy in differentiating keratinocytes depleted from key autophagy proteins WIPI1 or ULK1.
A report recently showed that epidermal ATG7 knockout mice have no skin phenotype suggesting that autophagy in the epidermis does not require ATG7 (
). Our observations of strong basal layer expression of BECN1 and ATG5-ATG12 suggest that these proteins may also play other roles in epidermis. It is likely that the precise mechanism of autophagy in epidermis may differ depending on the stage of keratinocyte differentiation and may also vary between mice and humans. It is worth mentioning that although no cornification defects were revealed by ultrastructural analysis, an increase in corneocyte thickness was identified in the back skin where Atg7 was specifically inactivated (
). Another group has recently grafted Atg7-deficient skin onto severe combined immunodeficient mice and observed acanthosis, hyperkeratosis, abnormal hair growth, and an overall retardation of granular layer differentiation in the autophagy-deficient grafts (
Psoriasis was used as a human disease model to validate our hypothesis and observations. Psoriasis is characterized by keratinocyte hyperproliferation, deregulated terminal differentiation and parakeratosis, inflammation, and abnormal vascularization, resulting in the defective epidermal barrier present in psoriatic plaques (
). We show that LC3 puncta were still present in the granular layer of nonlesional psoriatic epidermis but at a much lower level compared with healthy adult epidermis, whereas in lesional psoriatic skin, LC3 was not expressed at all. Likewise, no expression of ULK1 or WIP1 was detected in the parakeratotic regions of psoriatic skin, whereas a shift of BECN1 and ATG5-ATG12 expression was observed in psoriatic lesions. Therefore, the overall autophagy expression marker profile does only apply to an epidermal model capable of complete terminal differentiation. Moreover, hyperactivation of mTORC1 has been shown in lesional psoriatic skin (
). High mTORC1 signaling would inhibit autophagy and may explain why LC3 aggregates formation is absent in psoriatic lesions. Previous reports have shown that activation of toll-like-receptor signaling in keratinocytes causes hyperproliferation and inflammation and induces p62 expression that further increases toll-like-receptor signaling. However, in healthy keratinocytes, autophagy keeps epidermal p62 levels in check thereby preventing epidermal inflammation (
). The absent or reduced LC3, ULK1, WIP1 expression levels in psoriatic skin suggest that autophagy may be blocked or impaired, explaining the high p62 levels and contributing to excessive inflammation (
). These data suggest that patients with epidermal barrier defect diseases such as psoriasis may benefit from treatment with mTORC1 inhibitors. Such an approach could restore constitutive epidermal granular layer autophagy that may lead to normalization of terminal differentiation and barrier formation, and possibly reduce inflammation and hyperproliferation.
In conclusion, our data reveal the extent and location of autophagy in epidermal development and differentiation and highlight its critical importance in ensuring normal epidermal function and its potential role in disease of epidermal barrier function.
Materials and Methods
Cultures of primary keratinocytes and skin explants
Human neonatal primary keratinocytes purchased from Invitrogen were expanded in medium 154 (Invitrogen, Paisley, UK) with 0.2 mM calcium chloride on collagen-coated plates (rat-tail collagen, BD Biosciences). More details of this technique are included in the Supplementary Materials and Methods online.
Immunohistochemistry
Tissue specimens were fixed in Bouin’s solution and embedded in paraffin. More details of this technique and a list of the antibodies used are included in the Supplementary Materials and Methods.
Transmission electron microscopy
The processing, embedding, cutting, and imaging of the transmission electron microscopy sections were performed by the Pathology Core Facility of Queen Mary University of London. Three-day-old dorsal mouse skin from CD1 mice was harvested and cut into strips of 1 mm width. The strips were fixed in 4% glutaraldehyde in 0.1 M phosphate buffer (3.1 g sodium phosphate monobasic monohydrate, 10.9 g sodium phosphate dibasic [pH 7.4] per L), and embedded in LR White.
We are grateful to David Kelsell and John Connelly for kindly providing antibodies. Funding support was provided by MRC, Cancer Research-UK and Barts, and The London Charity.
Keratinization of the stratum corneum involves a highly choreographed sequence of events in which granular cells lose their nuclei and become desiccated corneocytes. Akinduro et al. detail the molecular machinery underlying removal of the nucleus (nucleophagy) during the final stages of keratinization. They provide evidence that nucleophagy is induced when the keratinocytes differentiate and that failure in the initiation of nucleophagy is associated with parakeratosis.
30973-3&id=gr1.jpg){kind=link}
30973-3&id=gr2.jpg){kind=link}
30973-3&id=gr3.jpg){kind=link}
30973-3&id=gr4.jpg){kind=link}
30973-3&id=gr5.jpg){kind=link}
30973-3&id=gr6.jpg){kind=link}