Keratins Regulate p38MAPK-Dependent Desmoglein Binding Properties in Pemphigus

Abstract

Keratins are crucial for the anchorage of desmosomes. Severe alterations of keratin organization and detachment of filaments from the desmosomal plaque occur in the autoimmune dermatoses pemphigus vulgaris and pemphigus foliaceus (PF), which are mainly caused by autoantibodies against desmoglein (Dsg) 1 and 3. Keratin alterations are a structural hallmark in pemphigus pathogenesis and correlate with loss of intercellular adhesion. However, the significance for autoantibody-induced loss of intercellular adhesion is largely unknown. In wild-type (wt) murine keratinocytes, pemphigus autoantibodies induced keratin filament retraction. Under the same conditions, we used murine keratinocytes lacking all keratin filaments (KtyII k.o.) as a model system to dissect the role of keratins in pemphigus. KtyII k.o. cells show compromised intercellular adhesion without antibody (Ab) treatment, which was not impaired further by pathogenic pemphigus autoantibodies. Nevertheless, direct activation of p38MAPK via anisomycin further decreased intercellular adhesion indicating that cell cohesion was not completely abrogated in the absence of keratins. Direct inhibition of Dsg3, but not of Dsg1, interaction via pathogenic autoantibodies as revealed by atomic force microscopy was detectable in both cell lines demonstrating that keratins are not required for this phenomenon. However, PF-IgG shifted Dsg1-binding events from cell borders toward the free cell surface in wt cells. This led to a distribution pattern of Dsg1-binding events similar to KtyII k.o. cells under resting conditions. In keratin-deficient keratinocytes, PF-IgG impaired Dsg1-binding strength, which was not different from wt cells under resting conditions. In addition, pathogenic autoantibodies were capable of activating p38MAPK in both KtyII wt and k.o. cells, the latter of which already displayed robust p38MAPK activation under resting conditions. Since inhibition of p38MAPK blocked autoantibody-induced loss of intercellular adhesion in wt cells and restored baseline cell cohesion in keratin-deficient cells, we conclude that p38MAPK signaling is (i) critical for regulation of cell adhesion, (ii) regulated by keratins, and (iii) targets both keratin-dependent and -independent mechanisms.

Keywords: atomic force microscopy; desmoglein; desmosome; keratin; p38MAPK.

Figure 1

Keratin retraction in murine keratinocytes and regulation of desmoglein binding properties through keratins. (A) Immunostaining of wild-type murine keratinocytes (KtyII wt) using keratin 14 monoclonal antibody and Alexa488-phalloidin for actin staining. Control-IgG treated keratinocytes show a dense keratin network throughout the whole cell whereas keratin filament retraction (arrows) and formation of thick filament bundles were detectable after treatment with pemphigus vulgaris (PV)-IgG, pemphigus foliaceus (PF)-IgG, and AK23. Pictures are representatives of n > 4. (B) Quantification of keratin retraction using an intensity coefficient confirms occurrence of keratin retraction in all autoantibody treated conditions. (C) Dissociation assay of KtyII wt and k.o. keratinocytes reveals that keratin-deficient keratinocytes show impaired intercellular adhesion. Treatment with pathogenic autoantibodies (PF-IgG, PV1- and PV2-IgG, and AK23) reduced intercellular adhesion in wild-type (wt) but not in KtyII k.o. cells. n = 5, *p > 0.05 vs. wt control and wt control-IgG. (D,E) Atomic force microscopy adhesion measurements using Dsg3-Fc- or Dsg1-Fc-functionalized cantilevers on the same scanning area. In adhesion panels, each blue or green pixel represents a specific Dsg3- or Dsg1-binding event, respectively. (F) Merged panels of adhesion measurements reveal distinct clustering of Dsg1- and 3-binding events. (G,H) Analysis of binding frequency and unbinding forces of Dsg1 and 3 interactions. n = 6 from ≥3 independent coating procedures, 1,200 force–distance curves/adhesion maps (*p < 0.05).

Figure 2

Keratin-deficiency accelerates depletion of Dsg3. (A,B) Dsg3 adhesion measurements on KtyII wt and k.o. cells using AK23. Treatment of KtyII wt and k.o. cells with AK23 for 1 h reduced Dsg3 binding to comparable extent. n = 6 from ≥3 independent coating procedures, 1,200 force–distance curves/adhesion maps (*p < 0.05). (C,D) Cell surface biotinylation delineates depletion of Dsg3 in keratin-deficient keratinocytes after 1 h of pemphigus vulgaris (PV) 1-IgG treatment but not in wild-type keratinocytes. Representative of n = 5, *p < 0.05 vs. respective control. (E) Kymographs of fluorescence recovery after photobleaching experiments using Dsg3-eGFP-transfected KtyII wt and k.o. keratinocytes. (F) Immobile fractions of Dsg3-eGFP in KtyII wt and k.o. cells under basal conditions and after incubation with AK23 incubation for 1 h; n = 7, 5 cell borders/experiment; *p < 0.05.

Figure 3

aDsg1 antibodies (Abs) causes a keratin-dependent redistribution of Dsg1. (A) Atomic force microscopy (AFM) adhesion measurements using Dsg1-Fc-functionalized tips at small areas along cell borders and on the cell surface above the nucleus. Pemphigus foliaceus (PF)-IgG for 1 h and 24 h induced redistribution of Dsg1 from cell junctions in wild type. (B) Dsg1-binding frequencies under control condition and after 1 h of PF-IgG treatment show no direct inhibition of binding events. (C) 24 h of PF-IgG treatment slightly reduced Dsg1-binding frequency in KtyII wt and k.o. cells. (D,E) Distribution coefficient of Dsg1-binding events under control conditions and after 1 and 24 h of PF-IgG incubation. n = 4 from 3 independent coating procedures, 1,200 force–distance curves/adhesion map (*p < 0.05). (F,G) Unbinding forces of Dsg1-binding events under control conditions and after 1 or 24 h of PF-IgG incubation. n = 4 from 3 independent coating procedures, 1,200 force–distance curves/adhesion map (*p < 0.05). (H) AFM adhesion measurements using Dsg1-Fc functionalized tips in KtyII wt and k.o. cells under control conditions or after incubation with pemphigus vulgaris (PV) 1-IgG containing aDsg1 Abs for 1 h. (I) Dsg1-binding frequency did not change after 1 h of PV-IgG treatment in both cell lines. (J) Distribution of Dsg1-binding events was altered in KtyII wt but not in KtyII k.o. cells after PV1-IgG treatment. n = 4 from 3 independent coating procedures, 1,200 force–distance curves/adhesion map (*p < 0.05).

Figure 3

aDsg1 antibodies (Abs) causes a keratin-dependent redistribution of Dsg1. (A) Atomic force microscopy (AFM) adhesion measurements using Dsg1-Fc-functionalized tips at small areas along cell borders and on the cell surface above the nucleus. Pemphigus foliaceus (PF)-IgG for 1 h and 24 h induced redistribution of Dsg1 from cell junctions in wild type. (B) Dsg1-binding frequencies under control condition and after 1 h of PF-IgG treatment show no direct inhibition of binding events. (C) 24 h of PF-IgG treatment slightly reduced Dsg1-binding frequency in KtyII wt and k.o. cells. (D,E) Distribution coefficient of Dsg1-binding events under control conditions and after 1 and 24 h of PF-IgG incubation. n = 4 from 3 independent coating procedures, 1,200 force–distance curves/adhesion map (*p < 0.05). (F,G) Unbinding forces of Dsg1-binding events under control conditions and after 1 or 24 h of PF-IgG incubation. n = 4 from 3 independent coating procedures, 1,200 force–distance curves/adhesion map (*p < 0.05). (H) AFM adhesion measurements using Dsg1-Fc functionalized tips in KtyII wt and k.o. cells under control conditions or after incubation with pemphigus vulgaris (PV) 1-IgG containing aDsg1 Abs for 1 h. (I) Dsg1-binding frequency did not change after 1 h of PV-IgG treatment in both cell lines. (J) Distribution of Dsg1-binding events was altered in KtyII wt but not in KtyII k.o. cells after PV1-IgG treatment. n = 4 from 3 independent coating procedures, 1,200 force–distance curves/adhesion map (*p < 0.05).

Figure 4

p38MAPK activation induces redistribution of Dsg1- and Dsg3-binding events. (A) Immunostaining of keratin 14 and Alexa488-phalloidin staining of actin in KtyII wt after anisomycin treatment for 1 h revealed keratin filament retraction (red arrows), which was blocked by p38MAPK inhibitors SB202190 and SB203580 when preincubated for 1 h. Representative images from n = 4. (B) Dissociation assay after treatment with anisomycin alone or after preincubation with SB202190 or SB203580 in keratin-deficient keratinocytes compared with wild-type (wt) (n > 4, *p < 0.05). (C–E) Dsg3 adhesion measurements show no change in binding frequency after anisomycin treatment but reveal a redistribution of binding events away from cell borders in KtyII wt but not in KtyII k.o. cells. (F) Unbinding forces of Dsg3 adhesion measurements were not significantly altered in both cell lines. n = 4 from 3 independent coating procedures, 1,200 force–distance curves/adhesion map (*p < 0.05). (G–I) Dsg1 adhesion measurements revealed redistribution of binding events away from cell borders in KtyII wt but not in KtyII k.o. cells. (J) Unbinding forces of Dsg1 adhesion measurements were not significantly altered in wt but reduced in KtyII k.o. cells. n = 4 from 3 independent coating procedures, 1,200 force–distance curves/adhesion map (*p < 0.05).

Figure 4

p38MAPK activation induces redistribution of Dsg1- and Dsg3-binding events. (A) Immunostaining of keratin 14 and Alexa488-phalloidin staining of actin in KtyII wt after anisomycin treatment for 1 h revealed keratin filament retraction (red arrows), which was blocked by p38MAPK inhibitors SB202190 and SB203580 when preincubated for 1 h. Representative images from n = 4. (B) Dissociation assay after treatment with anisomycin alone or after preincubation with SB202190 or SB203580 in keratin-deficient keratinocytes compared with wild-type (wt) (n > 4, *p < 0.05). (C–E) Dsg3 adhesion measurements show no change in binding frequency after anisomycin treatment but reveal a redistribution of binding events away from cell borders in KtyII wt but not in KtyII k.o. cells. (F) Unbinding forces of Dsg3 adhesion measurements were not significantly altered in both cell lines. n = 4 from 3 independent coating procedures, 1,200 force–distance curves/adhesion map (*p < 0.05). (G–I) Dsg1 adhesion measurements revealed redistribution of binding events away from cell borders in KtyII wt but not in KtyII k.o. cells. (J) Unbinding forces of Dsg1 adhesion measurements were not significantly altered in wt but reduced in KtyII k.o. cells. n = 4 from 3 independent coating procedures, 1,200 force–distance curves/adhesion map (*p < 0.05).

Figure 5

p38MAPK inhibition restores loss of intercellular adhesion caused by autoantibodies and keratin deficiency. (A) Western blot showing activation of p38MAPK after treatment with AK23, pemphigus vulgaris (PV) 1-IgG and pemphigus foliaceus (PF)-IgG for 1 h. Representative of n > 4. (B) Dissociation assay of KtyII wt and k.o. cells after incubation with pathogenic autoantibodies alone or in combination with SB202190 for 24 h (n > 5). *p < 0.05 vs. respective control (control panel for SB202190-treated cells and control-IgG for conditions treated with pathogenic autoantibodies); ^#p < 0.05 vs. respective condition treated with pathogenic autoantibodies alone. (C–E) Dsg1 adhesion measurements reveal a redistribution of binding events away from cell borders in KtyII wt but not in KtyII k.o. cells after 1 h PF-IgG. SB202190 preincubation for 1 h blocked PF-IgG-induced redistribution in KtyII wt cells and alone caused redistribution of Dsg1-binding events to cell borders in k.o. cells. Binding frequencies were not altered in all conditions. n = 3 from 3 independent coating procedures, 1,200 force–distance curves/adhesion map (*p < 0.05).