Fig 2: correlation of cellular adhesion molecules, E-cadherin (E-cad) is positively and significantly associated with desmoglein (DSG) 3 but not with DSG1; in addition, DSG1 and DSG3 did not show any significant correlation (*p-value = 0.003 by Spearman’s correlation test)

Fig 3: Inhibition of Dsg3 internalization through cholesterol depletion did not prevent PV-IgG-driven keratin alterations. HaCaT-CK5 keratinocytes were pre-incubated with either PBS or ß-MCD as a cholesterol-depleting agent for 1 h and then exposed to PV2-IgG for 6 h. Images shown (A) are representative for four independent experiments. For analysis a bar of 15 µm was applied perpendicularly over the cell border of two adjacent cells and the fluorescence intensity was plotted for both Dsg3 and CK5 (n = 100 cells from four independent experiments; *p < 0.05). Bar represents 10 µm. (B) Streptavidin pulldown of biotinylated membrane Dsg3 was performed with HaCaT-CK5 keratinocytes (n = 3–5, *p < 0.05). (C) Phosphorylation of p38MAPK was determined by densitometry. (D) Intercellular adhesion was determined by dispase-based dissociation assays in HaCaT keratinocytes (n = 8, *p < 0.05).

Fig 4: Structural changes of human keratinocytes in response to pemphigus vulgaris (PV) antibody exposure. HaCaT keratinocytes expressing cytokeratin5 (CK5)-YPF (HaCaT-CK5) were incubated with PV1-IgG for 2, 12, or 24 h. Images shown are representatives of >3 independent experiments. (A) Desmoglein (Dsg)3 staining and CK5 expression in response to the antibody binding. Loss of keratin filaments in the cell periphery is marked by arrows and Dsg3 alterations by arrowheads. Comparison of fluorescence profiles on a 15 µm line perpendicularly to the membrane of two adjacent cells (n = 75 cells from three independent experiments, *p < 0.05 vs. control). Bar represents 10 µm. (B) Triton X-100 insoluble fraction, representing the cytoskeletal-bound fraction, of HaCaT-CK5 lysates after incubation with PV1-IgG for the indicated period of time (n = 6). Densitometric analysis of structure proteins shown as fold of control (n = 6, *p < 0.05 vs. control). (C) Dispase-based dissociation assays in HaCaT-CK5 keratinocytes after PV1-IgG treatment (n = 5, *p < 0.05 vs. control). Representative images of cell sheets after applied sheer stress stained with 10 µM MTT for better visibility.

Fig 5: Keratin changes and Dsg3 clustering occurred in parallel and preceded Dsg3 internalization. HaCaT keratinocytes expressing CK5-YFP were transfected with Dsg3-mCherry. In three-dimensional confocal time-lapse microscopy experiments, cells were incubated with either PV1-IgG or Control-IgG and imaged every 2 min for 120 min. Dashed lines indicate cell border areas (A). Ratio of CK5 fluorescence intensity and in the cell periphery vs. intensity in the cytoplasm was assessed (B) (n = 5–6 cells, each from an independent experiment, *p < 0.05 vs. Control-IgG at respective time point). Ratio of Dsg3 distribution in the membrane vs. the cytoplasm (C) (n = 5–6, *p < 0.05 vs. Control-IgG). Streptavidin pulldown of biotinylated membrane Dsg3 after 2, 12, and 24 h of PV2-IgG incubation and densitometric analysis of protein levels (n = 3–4, *p < 0.05 vs. control) (D). Dsg3 clustering was evaluated with intensity profile plotting linearly along the membrane (n = 5–6, *p < 0.05 vs. Control-IgG at respective time point) (E). Loss of cell adhesion caused by incubation with PV2-IgG was measured in parallel to structural changes at early time points by dispase-based dissociation assay (F) (n = 4–5, *p < 0.05 vs. respective Control-IgG incubation).