Desmoglein-driven dynamic signaling in pemphigus vulgaris: a systematic review of pathogenic pathways

Abstract

Epithelial tissue integrity is maintained through specialized intercellular junctions known to coordinate homeostatic processes. In this context, outside-in signaling and mechanotransduction through desmosomal cadherins, the building blocks of desmosomes and main stress bearers in epithelial tissue, are only starting to emerge. To better understand the dual function of desmosomal cadherins in structural integrity and cellular signaling, we here performed a systematic, unbiased review on pathogenic signaling effectors identified in models and patients with pemphigus vulgaris (PV). PV is an autoimmune blistering disorder characterized by disruption of desmosomal transadhesion through autoantibodies mainly targeting the desmosomal cadherins desmoglein (Dsg) 3 or Dsg1 and Dsg3. The survey of functionally validated pathogenic pathways published since inception in 1977 up to mid-2024 identifies 128 studies and 128 signaling molecules, highlighting a coherent network of biomechanical, bioelectrical, and biochemical signaling events. This in-depth analysis will stimulate future research as well as development of potential therapeutic applications beyond PV.

Introduction

Intercellular adhesion is essential to maintain tissue integrity and intercellular communication in epithelial homeostasis^1,2. Structurally, junctional cell-cell adhesion is mediated by adherens junctions and desmosomes. The adhesive building blocks of adherens junctions are classical cadherins that communicate through their plaque proteins to the actin cytoskeleton. Desmosomes are robust, tightly packed, knob-like structures that confer strongest adhesion and are built by desmosomal cadherins which act as anchoring sites for intermediate filaments. It is well established that classical cadherins detect external mechanical forces through plasma membrane deformation and molecular conformational changes, transmitted intracellularly in a process known as mechanotransduction, ultimately leading to changes in cell behavior during homeostasis and regenerative processes^3–6. These outside-in mechanotransduction mechanisms are less well understood for desmosomal cadherins.

In the epidermis, the outermost layer of the skin, desmosomal cadherins, particularly Dsg3 and Dsg1 have been studied most extensively for their role as the predominant targets in the autoimmune blistering disease pemphigus vulgaris (PV). Dsg3 is mainly expressed in the basal and suprabasal layer of stratified epithelia, from where the highly complex homeostatic processes are governed⁷, whereas Dsg1 expression is low in basal layers and gradually increases up to the granular layer beneath the stratum corneum⁸. Antibody binding to these desmosomal cadherins in PV results in potentially life-threatening, basal-suprabasal loss of intercellular adhesion (acantholysis) and skin blistering^8,9. It is noteworthy that anti-Dsg3 antibodies are sufficient to induce blisters in tissues lacking Dsg1 expression, suggesting that Dsg1 has a compensatory function in this disease. Whether this function pertains to interfering signaling with Dsg3 or to structurally more resilient desmosomes, or both, will be an interesting topic for future investigations. Complementing these findings, a milder and cutaneous-limited variant of pemphigus named pemphigus foliaceus (PF), which is characterized by autoantibodies targeting Dsg1 without Dsg3, results in superficial epidermal blistering within the granular and upper spinous layers where Dsg1 is the main desmoglein isoform¹⁰. The distinct expression patterns and clinical phenotypes of PV and PF underscore functional redundancies between these adhesion molecules while simultaneously highlighting Dsg-specific functions required for the homeostatic balance within different layers of the epidermis^11–14. Studies on functional disruption of Dsg3 transadhesion by PV antibodies¹⁵ have highlighted that desmosomal cadherins not only provide mechanical resilience, but also participate in outside-in signaling that is critical in the homeostasis of stratified epithelia in terms of stem cell maintenance, control of proliferation and differentiation as well as tissue regeneration^16,17. This underlines the significance of Dsg3 and potentially Dsg1 beyond mere adhesion functions.

Seminal work on the turnover of desmosomal cadherins has indicated that upon plasma membrane delivery through the Golgi apparatus, desmosomal cadherins are stabilized by transadhesion prior to lateral incorporation into tightly packed desmosomes^18–20. This extra-desmosomal cadherin pool, which can be fractionated biochemically (with 0.5–1% Triton-X 100), is characterized by absence of desmoplakin and presence of associated junctional plakoglobin (JUP), plakophilins and actin. It is believed to serve as an intermediate regulatory adhesive signaling structure during the transport to and from desmosomes^8,20–25. Early research has further demonstrated that PV autoantibodies predominantly target Dsg3 expressed in the more accessible extra-desmosomal compartment at the cell surface^{8,9,18,24–28}, while antibody binding to desmosomes occurs at later stages when these structures become functionally defective^18,29,30. The implication of cell signaling in PV started to be studied three decades ago by the group of Dr Kitajima. Seconds after antibody binding, cytosolic calcium influx was recorded as the first pathogenic signaling event³¹. Within the first hour, the turn-over of plasma membrane exposed, antibody-targeted extra-desmosomal Dsg3 as well as JUP, a plaque protein of Dsg1 and Dsg3 and transcription factor in the Wnt pathway^25,32, increase significantly, as shown by metabolic labeling, and drive pathogenic PV signaling²⁵. Consistently, JUP knockout keratinocytes failed to respond to PV IgG by means of disruption of the desmosomal plaque and loss of intercellular adhesion, despite defective desmosomes and normal PV IgG binding^24,33. Furthermore, the actin cytoskeleton machinery, including RhoA, a critical mechanotransductive regulatory molecule mediating the crosstalk between adherens junctions and desmosomes to support homeostatic processes³⁴, was suggested to contribute to PV pathogenesis and regeneration^35,36. Although these results highlight outside-in signal transduction by antibody-mediated disruption of Dsg3 transadhesion in conjunction with Dsg1³⁷, to date there has been no comprehensive analysis of the mechanotransducive effector network in PV. The latter is important to identify existing gaps in our understanding of both the effects of the pathogenic anti-Dsg3/Dsg1 autoantibodies and Dsg3/Dsg1 functions as well as to identify potentially druggable pathways for the first line treatment of PV patients.

Here we present the first comprehensive, pathogenic effector signaling network in PV based on the systematic review of literature published since inception in 1977 up to April 2024. The effector framework categorizes in three fundamental signaling modalities mediated by Dsg3 or Dsg3/Dsg1 which also typify classical cadherin and integrin mechanotransduction: biomechanical, bioelectrical, and biochemical signaling^3–6. These networks highlight a dynamic and complex interplay of a variety of intricate pathogenic signaling pathways that initiate blister formation in PV and promote tissue regeneration. This proposes a model whereby effector pathways such as calcium influx are activated within seconds to minutes followed by PLC/PKC, EGFR and p38MAPK activation as well as JUP dissociation from its Dsg3 anchor. This results in fate conversion including cortical actin cytoskeletal depolymerization, mitosis, migration, compromised differentiation and junctional reductions, which leads to blistering. This is followed by regenerative processes reminiscent of mechanotransduction of the desmosomal system to adherens junctions to establish force equilibrium and epithelial homeostasis.

The systematic review will shed new light on PV and desmosomal cadherin signaling while the established framework allows for identification of unexplored hierarchy and synergies between signaling modalities, with the overarching goal to develop efficient treatment options for PV and potentially other acantholytic disorders while allowing for a more effective translation of findings across disciplinary boundaries.

Results

Study characteristics

A comprehensive search was performed with the focus question of “identifying signaling molecules in keratinocytes that are essential for loss of intercellular adhesion and blister formation in PV based on pharmacological or genetic manipulation (inhibition/activation; knockout/knockdown)” (see methods; flow diagram, Supplementary Fig. S1 and Table S1 and PRISMA checklists, Supplementary Information and Supplementary Data 1). The initial search, covering all in vitro and in vivo models for PV³⁸, yielded 11,426 articles, from which 128 articles were eligible for final assessment. The PRISMA flow diagram for identification and screening of the studies is shown in Supplementary Fig. S1, Supplementary Information. The literature review identified a total of 332 interventions with 234 unique approaches, which identified 128 signaling molecules. These signaling molecules were categorized according to three main signaling modalities: biomechanical, bioelectrical and biochemical (Figs. 1–3). The most employed models in these studies were human or mouse 2D keratinocyte cultures (83 studies), the passive transfer neonatal mouse model (PTNM, 38 studies), and ex vivo human skin organ culture (30 studies). Additional models included the passive transfer adult mouse model (PTAM, 4 studies), 3D culture (2 studies), ex vivo human oral mucosa culture (1 study), and patient skin samples (1 study). A wide range of pathogenic patients and experimental IgG antibodies or antibody fractions were used to investigate the biological targets. The most commonly used PV induction methods were IgG purified from PV patients’ sera (PV IgG) (88 studies), the experimental anti-mouse Dsg3 IgG antibody AK23 (32 studies), and sera from PV patients (23 studies). Others included enzymatically digested IgG subfragments (Fab or F(ab´)₂ IgG fragments) of PV IgG (4 studies), Px43 scFv or other scFv (single chain variable fragment anti-Dsg1/3 cloned from PV patient; 4 studies), humanized AK23 (1 study), 4B3 (an anti-Dsg1/3 antibody isolated from patient) (1 study), the monoclonal anti-Dsg3 IgG antibody 2G4 (from humanized Dsg3 mice³⁹ against the amino terminal region EC5 of human Dsg3; 1 study), a rabbit polyclonal anti-Dsg3 antibody (1 study), and canine PV IgG (1 study). The details of the eligible studies are given in Supplementary Data 1. Effectors and key effectors (the latter indicated in bold were confirmed multiple times by several PV models, approaches and laboratories) are depicted in Figs. 1–3 and Fig. 4, respectively.

Fig. 1. Schematic representation of the signaling network of biomechanical effectors identified upon PV antibody binding and loss of transadhesion between extra-desmosomal Dsg3 or Dsg1/3 receptors.

The diagram illustrates key molecular interactions and pathways which have been involved in loss of intercellular adhesion in PV, with arrows indicating the direction of interaction between molecules. Effector molecules identified and tested across various experimental PV models and antibody types are highlighted in bold, underscoring their roles in downstream signaling events leading to shrinkage of desmosomes (to the right) and blister formation. This figure provides an overview of three main biomechanical signaling cascades initiated upon antibody engagement with Dsg receptors, capturing the critical points of intervention, molecular crosstalk and consequences thereof (boxes at the bottom). This figure was created in BioRender. Hariton, W. (2025) https://BioRender.com/2iinwlr.

Fig. 3. Overview of the signaling network of biochemical effectors activated by PV antibody binding to Dsg3 or Dsg1/3.

This figure illustrates key molecular interactions and highlights the four major signaling pathway families cooperating with Dsg1/3 to induce loss of intercellular adhesion and blistering, with arrows indicating interactions between specific molecules within the signaling cascade. Effector molecules tested across various experimental PV models and antibody types are highlighted in bold, emphasizing their roles in the biomechanical signaling process. The schematic captures the intricate web of biomechanical signaling events and consequences thereof (boxes at the bottom) following antibody engagement with Dsg receptors, identifying critical points for potential intervention and inter-molecular crosstalk. This figure was created in BioRender. Hariton, W. (2025) https://BioRender.com/opaf5k5.

Fig. 4. Summary scheme of major effectors identified in biomechanical, bioelectrical and biochemical signaling leading to desmosome shrinkage and loss of intercellular adhesion in PV.

The model proposes a sequence of events ranging from PV antibody binding to Dsg3 or Dsg1/3 receptors inducing loss of transadhesion, to influx of calcium from extracellular and intracellular stores, activation of cell surface receptors and their effectors through to altered transcription in the nucleus. This figure was created in BioRender. Hariton, W. (2025) https://BioRender.com/h2vc44h.

Biomechanical signaling

Biomechanical signal transducers from the plasma membrane to the nucleus have been identified in many biological systems^5,6, of which some can be assigned to PV pathogenicity. These are the extra-desmosomal cell-cell adhesion receptor Dsg3 acting as upstream regulator of Rho GTPases and actin^19,35,40, Dsg3 associated signaling molecules JUP^24,25, plakophilin 1–3^41,42, growth factor receptors such as EGFR⁴³ and the pleotropic, stress responsive p38MAPK⁴⁴. Characteristics of the 66 eligible studies are compiled in Table 1 and pathogenic effectors in PV are depicted in Fig. 1.

Table 1.

Characteristics of studies included in biomechanical signaling axis

Author	Year	Target molecule	Intervention	Treatment Modality	Model	Antibody	Preventing Pathology	Measured Outcome
Cell adhesion molecules
Hunziker and Morgenthaler95	1987	Fibronectin	Fibronectin (20–500 µg/mL)	Pretreatment	Ex vivo	PV serum	+	Brightfield microscopy examination of blister formation
Caldelari et al.24	2001	JUP	JUP KO	-	In vitro	PV IgG	+	KDA and keratin retraction measured by immunofluorescence microscopy
Berkowitz et al.44	2005	p38MAPK	SB202190 (100 µM)	Pretreatment for 1 h	In vitro	PV IgG	+	Keratin retraction measured by confocal immunofluorescence microscopy
Berkowitz et al.65	2006	p38MAPK	SB202190 (6.25 µg/injection), SB203580 (12.5 µg/ injection)	Pretreatment for 2 h and cotreatment	PTNM	PV IgG	+	Gross and microscopic evaluation of acantholysis, Nikolsky’s sign, and blister formation
de Bruin et al.58	2007	JUP	JUP KO	-	In vitro	PV IgG/PV Fab	+	Keratin retraction measured by immunofluorescence microscopy
Cirillo et al.175	2007	Desmoglein 1/3	IgG depletion from PV serum	-	In vitro	PV serum	-	Cell-cell detachment and disruption of intercellular contacts were seen after 24 and 48 hours, respectively.
Chernyavsky et al.84	2007	p38MAPK, Src	PD169316 (10 µM), PP2 (10 µM)	Pretreatment	In vitro	PV IgG	PD169316 (-), PP2 (+)	Cell volume and keratin aggregation measurements
Lee et al.66	2009	p38MAPK	SB202190 (6.25 μg)	Pretreatment for 2 h or treatment after 3 h	PTNM	PV IgG	Pretreatment (+), Treatment (-)	Keratin retraction measured by confocal immunofluorescence microscopy
Cirillo et al.42	2009	Plakophilin 3	siRNA PKP3 KD	-	In vitro	PV IgG	+	KDA, morphometric analysis of cell-cell detachment and acantholysis
Heupel b et al.37	2009	Desmoglein 3	Tandem peptide (20 µM)	Cotreatment	In vitro	PV IgG	+	KDA and keratin retraction measured by immunofluorescence microscopy
Heupel a et al.92	2009	Src	PP2 (10 µM)	Pretreatment for 2 h	In vitro	PV IgG 1–5	-	KDA and keratin or F-actin retraction measured by immunofluorescence microscopy
Nguyen et al.49	2009	Perp	Prep KO	-	In vitro	PV IgG	-	KDA
Pretel et al.86	2009	Src	PP1 (1 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Clinical activity and microscopic evaluation of acantholysis
Jolly et al.67	2010	p38MAPK	SB202190 (100 µM)	Pretreatment for 2 h	In vitro	PV IgG	+	Dsg3 internalization was measured by immunofluorescence microscopy
Mao et al.80	2011	p38MAPK	SB202190 (2 µM), p38α MAPK epidermal KO, p38MAPK siRNA	Pretreatment for 1 h	In vitro, PTNM	PX43	In vitro (+), PTNM (-)	In vitro: Dsg3 internalization was measured by immunofluorescence microscopy. PTNM: Gross and microscopic evaluation of blister formation.
Jennings et al.47	2011	Desmoglein 3	DSG3 overexpression	-	In vitro	PV IgG	+	KDA
Saito et al.68	2012	p38MAPK	SB202190 (In vitro: 20 µM, ex vivo: 5 µg)	In vitro: Pretreatment for 1 h. Ex vivo: Pretreatment for 2 h	In vitro, ex vivo	PV IgG, AK23	PV IgG (+), AK23 (-)	In vitro: KDA. Ex vivo: Gross and microscopic evaluation of blister formation
Gil et al.87	2012	FAK, Src	PF-573228 (5 µg/g), PP1 (1 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Gross and microscopic evaluation of tissue
Espana et al.88	2013	FAK, Src	PF-573228 (5 µg/g), PP1 (1 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Gross and microscopic evaluation of tissue
Spindler et al.48	2013	Desmoglein 3	Tandem peptide (in vitro: 20 µM, PTNM: 20 μM/L in 50 μl PBS for injection, 2 nmol mixed in 100 mg of linimentum aquosum for topical administration	In vitro: Cotreatment and posttreatment at 6 and 24 h. PTNM: Pretreatment for 2 h or posttreatment after 6 h	PTNM/ In vitro	PV IgG/ AK23	+	In vitro: KDA and Dsg3 immunofluorescence microscopy. PTNM: Gross and microscopic evaluation and scoring of blisters
Bektas et al.69	2013	p38MAPK	SB202190 (10 µM)	Pretreatment for 1 h	In vitro	PV IgG	+	Keratin retraction and Dsg3 internalization was visualized with immunofluorescence microscopy. Dsg3 depletion was measured by western blotting
Dehner et al.136	2014	Desmoplakin	Desmoplakin point mutation (S2849G)		In vitro	PV IgG/ AK23	+	KDA and keratin retraction measured by immunofluorescence microscopy
Spindler et al.59	2014	JUP, Desmoplakin	JUP siRNA KD, Desmoplakin siRNA KD	-	In vitro	AK23	x	KDA
Tucker et al.41	2014	Plakophilin 1, JUP	Plakophilin 1 overexpression, JUP overexpression	-	In vitro	PV IgG, AK23	Plakophilin 1 (-), JUP (x)	KDA and desmosome length was measured by electron microscopy
Hartlieb et al.70	2014	p38MAPK	SB202190 (30 µM)	Pretreatment for 1 h	In vitro	AK23	+	KDA
Mao et al.71	2014	p38MAPK	SB202190 (In vitro: 2 µM, PTNM: 5 µg)	Pretreatment for 2 h	In vitro, PTNM	PX43	+	In vitro: Dsg3 internalization was measured by immunofluorescence microscopy. PTNM: Gross and microscopic evaluation of blister formation.
Spindler et al.59	2014	p38MAPK	SB202190 (30 µM)	Cotreatment	In vitro	AK23	+	KDA
Rötzer et al.72	2014	p38MAPK	SB202190 (30 µM)	Cotreatment	In vitro	AK23	+	KDA
Vielmuth et al.73	2015	p38MAPK	SB202190 (NA)	Pretreatment for 1 h	In vitro	PV IgG, AK23	+	KDA. Keratin retraction and intercellular widening were visualized with immunofluorescence, and atomic force microscopy, respectively.
Cirillo et al.89	2014	Src	Src inhibitor-1 (1 µM)	Pretreatment	In vitro	PV IgG	+	Morphometric analysis of cell-cell detachment and KDA
Rötzer et al.21	2015	E-cadherin	E-cadherin overexpression	-	In vitro	PV IgG	+	KDA
Egu et al.74	2017	p38MAPK	SB202190 (50 µL of 30 µM)	Pretreatment for 1 h	Ex vivo	mPV IgG, mcPV IgG, AK23	+	Gross and light microscopic evaluation of blister formation or electron microscopy for blister size quantification and desmosome size.
Walter et al.75	2017	p38MAPK, Src	SB202190 (NA), PP2 (NA)	Pretreatment for 1 h	In vitro	AK23, mcPV IgG, mPV IgG, atPV IgG, PF IgG	+	KDA
Vielmuth et al.76	2018	p38MAPK	SB202190 (30 µM)	Cotreatment	In vitro	PV IgG, PF IgG	+	KDA
Hofrichter et al.203	2018	Desmoglein 1/3	Anti-Dsg1/3 antibody depletion	-	In vitro/ PTNM	PV IgG	+	In vitro: KDA and Dsg3 degradation assay. PTNM: Gross and microscopic histology before and after application of mechanical stress
Walter et al.46	2019	Desmoglein 2/3	Desmoglein 2 KO, Desmoglein 3 KO	-	In vitro	mPV IgG, mcPV IgG, PF IgG	Desmoglein 2 (x), Desmoglein 3 (+)	KDA
Kugelmann et al.90	2019	Src	PP2 (In vitro: 10 µM, ex vivo and PTNM: 50 µL of 10 µM)	In vitro: Pretreatment for 2 h, ex vivo and PTNM: Cotreatment	PTNM, In vitro, ex vivo	PV1-4 IgG, AK23	+	In vitro: KDA. Ex vivo and PTNM: Blister score was counted in H&E stained samples
Ivars et al.91	2020	Src	PP1 (1 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Gross and microscopic evaluation of blister formation with histological scoring
Burmester et al.77	2020	p38MAPK	BIRB 796 (0.1, 1, and 10 µM)	In vitro: Pretreatment for 2 h then discarding supernatant. Ex vivo: Cotreatment	In vitro, ex vivo	In vitro: Immunophresis material from PV patient. ex vivo: PX43 scFv	+	In vitro: KDA. Ex vivo: Gross and light microscopic evaluation of blister formation.
Egu et al.204	2020	p38MAPK	SB202190 (50 µL of 30, 60, 120 µM), SB203580 (30 µM)	Pretreatment for 1 h	Ex vivo human oral mucosa culture	AK23, mPV IgG	-	Gross and light microscopic evaluation of blister formation.
Sigmund et al.60	2023	JUP	Phosphodeficient mutant of JUP S665	-	In vitro	AK23	-	KDA
Schmitt et al.78	2023	p38MAPK	SB202190 (60 µM), EO1428 (40 µM)	Pretreatment for 1 h	In vitro	AK23, 2G4	+	KDA. Dsg3 internalization, keratin retraction, and desmosome numbers were visualized using STED microscopy.
Hariton et al.16	2023	Desmoglein 3	Desmoglein 3 KO	-	In vitro	AK23	+	KDA
Schmitt et al.78	2023	Src	PP2 (10 µM)	Pretreatment for 1 h	In vitro	AK23, 2G4	+	KDA. Dsg3 internalization, keratin retraction, and desmosome numbers were visualized using STED microscopy.
Cytoskeleton
Patel et al.27	1984	Actin, Tubulin	Cytochalasin D (0.5 µg/mL), Colchicine (40 µg/mL), Combination	Pretreatment for 30 min	In vitro	PV sera	-	Antibody internalization by electron microscopy
Sánchez-Carpintero et al.99	2004	Calmodulin	W-7 hydrochloride (200 µg/g)	Pretreatment for 3 h	PTNM	PV IgG	+	Gross and microscopic evaluation of tissue
Waschke et al.35	2006	Rho A, Rac 1, and Cdc42	CNF-1 (1200 ng/mL), CNFy (HSOC: 1200 ng/mL, 2D Cell culture: 600 ng/mL)	Cotreatment	Ex vivo/ in vitro	PV IgG	+	Ex vivo: Gross and microscopic evaluation of tissue. In vitro: F-actin immunofluorescence microscopy
Heupel a et al.92	2009	Rho A, Rac 1, and Cdc42	CNF-1 (300 ng/mL), CNFy (900 ng/mL)	Pretreatment for 2 h	In vitro	PV IgG	+	KDA and keratin retraction measured by immunofluorescence microscopy
Gliem et al.40	2010	Actin, Rho A, Rac 1, and Cdc42	Cytochalasin D (100 nM), Jasplakinolide (100 nM), Actin-RFP overexpression, CNF-1 (300 ng/mL)	Pretreatment for 3, 1, and 0.5 h, respectively	In vitro	PV IgG	Cytochalasin D and Actin-RFP overexpression (-), Jasplakinolide and CNF-1 (+)	KDA
Jennings et al.47	2011	Actin	Lantraculin A (250 nM)	Pretreatment for 1 h	In vitro	PV IgG	-	Dsg3 internalization measured by immunofluorescence microscopy
Rötzer et al.72	2014	α-adductin	siRNA KD	-	In vitro	AK23	x	KDA
Sumigray et al.50	2014	Myh9	Myosin IIA KO	-	In vitro	AK23	+	KDA
Vielmuth et al.76	2018	Keratin cluster II	Keratin cluster II knockout (KtyII)	-	In vitro	PV IgG, PF IgG, AK23	-	KDA
Kugelmann et al.90	2019	Cortactin (CTTN)	CTTN KO	-	PTNM/In vitro	AK23	-	In vitro: KDA. PTNM: Gross and microscopic evaluation of tissue
Jin et al.36	2021	RhoA	Cyclic mechanical stretch (10% 1 Hz cyclic stretch for 4 h), CN01 (1 unitl/mL)	Cotreatment	In vitro	AK23	+	KDA
Plasma membrane and Endocytosis
Hunziker and Vassalli et al.116	1987	Plasminogen	Plasminogen (300 µg/mL)	-	Ex vivo	PV sera	x	H&E staining light microscopy examination of blister formation and acantholysis
Morioka et al.117	1987	Urokinase-type plasminogen activator (uPA)	Anti-Urokinase IgG (NA)	-	Ex vivo	PV IgG	+	H&E staining light microscopy examination of blister formation and acantholysis
Hashimoto et al.118	1989	Urokinase-type plasminogen activator (uPA)	Plasminogen Activator Inhibitor 2 (100 µg/mL)	-	Ex vivo	PV IgG	+	H&E staining light microscopy examination of blister formation and acantholysis
Wilkinson et al.119	1989	Urokinase-type plasminogen activator (uPA)	Human placental urokinase inhibitor (50 U/mL)	Cotreatment	Canine oral keratinocytes grown at the air-liquid interface on synthetic membranes	Human and canine PV IgG	+	H&E staining light microscopy examination of blister formation and acantholysis
Xue et al.120	1998	Urokinase-type plasminogen activator (uPA)	Anti-Urokinase-type plasminogen activator receptor (100 µg/mL)	Cotreatment	Ex vivo	PV IgG	+	H&E staining light microscopy examination of blister formation and acantholysis
Mahoney et al.123	1999	Tissue type Plasminogen Activator (tPA), Urokinase-type plasminogen activator (uPA)	Tissue type Plasminogen Activator (tPA) KO, Urokinase-type plasminogen activator (uPA) uPA + tPA knockout	-	PTNM	PV IgG	-	Gross and microscopic evaluation of tissue
Asano et al.121	2001	Urokinase-type plasminogen activator receptor, Plasmin and plasminogen	Phosphatidylinositol-specific phospholipase C (PI-PLC) (10 µg/mL), Tranexamic acid (15 mM)	Pretreatment for 30 minu	In vitro	PV IgG	+	keratin retraction and desmoplakin were measured by immunofluorescence microscopy
Feliciani et al.122	2003	Urokinase-type plasminogen activator (uPA)	Anti-Urokinase IgG (2 µg/mL)	Pretreatment for 30 min and cotreatment	In vitro	PV sera	+	Acantholysis was measured by cell detachment assay
Delva et al.113	2008	Dynamin II, Caveolin-1	Dynamin II (Dyn K44A) mutant, Caveolin-1 (Y14F) mutant	-	In vitro	AK23	-	Dsg3 internalization assay
Vielmuth et al.73	2015	Cholesterol	β-Methylcyclodextrin (NA)	Pretreatment for 1 h	In vitro	AK23	+	KDA
Schlögl et al.81	2018	Cholesterol	β-Methylcyclodextrin (1 mM)	Pretreatment for 1 h	In vitro	AK23	+	KDA, keratin retraction was measured by immunofluorescence microscopy. Dsg3 internalization was measured by immunofluorescence microscopy and western blotting.
Choi et al.114	2019	RPGRIP1L	RPGRIP1L siRNA KD	Cells were transfected prior to calcium switch	In vitro	PV IgG	-	Dsg3 internalization was measured by immunofluorescence microscopy

Treatment modalities: Pretreatment, Intervention added before antibody application. Cotreatment, Intervention added simultaneously with the antibody. Posttreatment, Intervention added after antibody application. Model: Ex vivo, human skin organ culture. In vitro, 2D cultured keratinocyte cell lines or primary cells. PTNM, Passive transfer neonatal mouse model. PTAM, Passive transfer adult mouse model. Antibodies: mPV IgG, mucosal PV IgG. mcPV IgG, mucocutaneous PV IgG. atPV IgG, Atypical PV IgG. PF IgG, pemphigus foliaceus IgG. Preventing Pathology: ( + ) effective in preventing antibody-induced pathology, (-) Ineffective in preventing antibody-induced pathology, (x) worsened antibody-induced pathology. KDA Keratinocyte dissociation assay, NA Not Available.

Cell adhesion molecules and associated proteins

Dsg3 and Dsg1 have been identified as major targets in PV. In support of this clinical finding, the pathogenic impact of anti-Dsg3 and anti-Dsg1/3 antibodies can be completely abolished by removal of Dsg3 or Dsg1 antibodies from patients serum⁴⁵. In line, cultured Dsg3 gene knockout keratinocytes or Dsg3-deficient mice mimic the PV phenotype in terms of blister sites and altered signaling^16,46. Further supporting the unique functional role of Dsg3 and Dsg1, PV antibodies and to some extent also PF antibodies remained pathogenic after Dsg2 knockout. In agreement with these findings as well as the fact that Dsg3 is endocytosed upon PV antibody binding, Dsg3 overexpression exerted an inhibitory effect on cell sheet fragmentation in vitro⁴⁷. Two other studies have demonstrated that the stabilization of Dsg adhesion, achieved by cross-linking the binding pockets of two trans-interacting Dsg3 or Dsg1 molecules, can partially mitigate the PV phenotype of intercellular gap formation, altered Dsg3 localization, loss of intercellular adhesion, and blister formation in both in vitro and in vivo models^37,48. Building upon the finding on predominant targeting of extra-desmosomal Dsg3 by PV antibodies^{8,9,18,24–28}, Nguyen et al.⁴⁹ have proposed that the gene knockout of Perp is associated with a decrease in Dsg3 levels in desmosomes and concurrent increase of Dsg3 in the extra-desmosomal compartment. Consistently, Perp knockout was found to exacerbate the PV antibody-induced weakening of intercellular adhesion, an effect that was attributed to elevated levels in the extra-desmosomal Dsg3 pool (conceivably increasing the number of antibody targets). These observations also suggested that Perp plays an integral role in modulating the distribution of Dsg3 between the desmosomal and non-desmosomal compartments.

Other than desmosomal cadherins, the role of E-cadherin has also been investigated in PV. Multiple laboratories reported that the plasma membrane-bound expression and localization as well as related Wnt/β-catenin signaling were unaffected during blister initiation^16,21,25,50, while a considerable increase in E-cadherin, β-catenin and JUP expression at the plasma membrane marked regenerative processes of morphological repair in a mouse model for hair follicle blistering¹⁶, suggesting mechanosensitive feedback mechanisms. Rötzer et al.²¹ further reported that overexpression of E-cadherin was able to mitigate the PV IgG-induced fragmentation in HaCaT keratinocytes, by so far unknown mechanisms, while E-cadherin knockdown decreased desmosome formation and intercellular adhesive strength²¹, supporting a mutual crosstalk and direct association of these adhesion receptors^34,51–53.

Signaling through extra-desmosomal Dsg3, described and investigated in PV by means of biochemical and electron microscopical approaches since more than three decades^{8,9,18,24–28}, presumably occur in main conjunction with associated plaque proteins JUP and plakophilin 1–3, as well as p38MAPK, and actin^24,54. Once integrated into desmosomes, the desmosomal cadherins and their plaque proteins anchor to the keratin filament cytoskeleton through association with desmoplakin⁵⁵. The plaque proteins are also known transducers of mechanical signals as further discussed below^34,56,57. Studies have demonstrated that PV IgG or Fab-induced keratin retraction, disruption of the desmosomal plaque and loss of cell-cell adhesion were eliminated in JUP knockout keratinocytes, supporting the implication of JUP in the process, while pathogenic effects were rescued upon moderate ectopic expression of JUP^24,58. However, conflicting findings have been reported, indicating that both knockdown or overexpression of JUP exacerbate antibody-induced fragmentation^41,59. Given the functional implication of JUP at the plasma membrane, in the cytoplasm and in the nucleus in PV^24,25, alterations in its distribution resulting from knockdown or overexpression may contribute to the observed discrepancies which prompt further investigations. Notably, a recent study by Sigmund et al.⁶⁰ unveiled that intracellular adhesion is impaired in keratinocytes with S665 phospho-deficient JUP, and the exposure to PV antibodies do not result in further alteration. This highlights the importance of JUP phosphorylation in PV pathogenesis and loss of desmosomal integrity. In addition to JUP, the role of other desmosomal plaque proteins has also been investigated in PV pathogenesis. Specifically, studies have shown that the knockout of plakophilin 3⁴² or the overexpression of plakophilin 1⁴¹ attenuated the antibody-induced loss of intercellular adhesion. Conceivably, the knockdown of desmoplakin was found to facilitate and thus enhance the AK23 antibody-induced fragmentation of keratinocytes⁵⁹. These findings highlight the complex and variable roles that different desmosomal cadherin components in transmitting the pathological effects of PV autoantibodies to cell-cell adhesion, which is still largely unresolved

In mouse and human keratinocytes under homeostatic conditions, co-immunoprecipitation revealed that extra-desmosomal Dsg3 and Dsg1 also associate with p38MAPK^35,59,61. Upon PV antibody binding, pathogenic activation/phosphorylation of p38MAPK has been consistently emphasized and was shown to causally link loss of Dsg3 transadhesion to proteins anchoring Dsg1/3 to the cytoskeletal, such as actin or keratin intermediate filaments, followed by loss of intercellular adhesion^48,61. The four p38MAP kinase isoforms (p38α, p38β, p38γ, p38δ (MAPK14, 11, 12, 13)) as well as multiple p38 splice variants of this ubiquitously expressed MAP kinase superfamily are pleotropic responders to environmental, intracellular and pathological stresses to ensure cell survival and they exhibit distinct but also overlapping and compensatory functions⁶². Of interest in the context of PV is that p38MAP kinases are typically activated by mechanical insult, to phosphorylate, in case of the most studied family member p38α, over 100 protein targets throughout all cellular compartments, including the nucleus. P38MAPK targets involve transcription factors, DNA/DNA binding proteins as well as kinases and structural proteins including actin and keratins. Hence, p38MAPK regulate a wide range of processes in response to mechanical stress from transcription, chromatin remodeling, mRNA stability, cell cycle regulation and translation to protein degradation and localization, endocytosis, metabolism, and cytoskeleton dynamics including migration^21,63, of which only a few processes have been assigned to PV so far. In the epidermis and other epithelial tissues, the three mainly expressed p38MAPK isoforms are p38α, p38γ and p38δ. In this context, it is noteworthy that current p38MAPK inhibitors, of interest in therapeutic approaches in PV, only effectively inhibit the p38α and p38β but not or to a lesser extent the p38γ and p38δ isoforms⁶². Hence, it is currently not known to which extent these isoforms operate and compensate for each other in PV, which might be the reason for some discrepant results described below.

PV IgG-induced p38MAPK phosphorylation occurs rapidly, with peaks observed as early as after 15 min⁶⁴ which further propagate in waves up to the final blister formation⁶⁵. The studies utilizing p38α and p38β or p38α inhibitors SB202190 and EO1428, and the broad range BIRB 796 inhibitor showed significant inhibition of intercellular adhesion loss in different models and with several antibodies^{44,59,65–78}. Interestingly, Mao et al. revealed the involvement of p38MAPK in Dsg3 internalization triggered by an experimental Dsg3/1 monoclonal antibody cloned from a PV patient⁷⁹. However, conditional deletion of p38α in mouse epidermis was not sufficient to protect against this antibody-induced blistering⁸⁰. A later study by Mao et al. showed that while some knockout animals developed subcorneal blisters, suprabasal blister formation was significantly decreased in mice lacking epidermal p38α⁷¹. This said, it is worth mentioning that the study by Mao et al. is the only one which addressed a specific p38MAPK isoform in PV. The result of this seminal work evokes the possibility that not only p38α but also the p38γ or p38δ isoforms contribute to PV blistering, or, alternatively, that p38γ or p38δ compensate in case of p38α deletion⁶². Involvement or compensation of other isoforms might also be the reason that p38α and p38β inhibitors like PD169316 and SB203580 were in some cases less effective or ineffective in preventing intercellular adhesion loss. Some of these studies also concluded on p38MAPK activation down- instead of upstream of PV blistering, which however might pertain to the finding that keratin retraction/p38MAPK activation and Dsg3 endocytosis are complementary processes⁸¹. Similarly, Saito et al. discovered that SB202190 inhibited PV IgG-induced blistering and fragmentation but not the one of AK23 in vitro and ex vivo⁶⁸; however, subsequent studies demonstrated its efficacy in inhibiting AK23-induced pathogenic effects^{70,73–75,78}.

Cell-Matrix adhesion molecules constitute an important mechanosensory apparatus of epithelial cells that cross-talk to cell-cell adhesion molecules as might also be the case in PV^17,82. Recruitment and activation/phosphorylation of Src to cell-cell contacts and receptor tyrosine kinases such as EGFR family members is considered a mechano-switch to tune tissue-specific functions. This includes functional deregulation of E-cadherin and integrin adhesion implicating focal adhesion kinase (FAK) downstream of RhoA activation⁸³. Chernyavsky et al. were the first to demonstrate that Src is activated by PV IgG treatment; however, knockdown of Dsg1 or Dsg3 did not reduce Src activation⁸⁴. Recently, Egu et al. showed that PV IgG, but not PF IgG, activate several Src family kinases, including Blk, Yes, Fyn, Frk, Lck, and Src⁸⁵. Although numerous studies across various models have reported the beneficial effects of Src inhibition downstream of EGFR^{75,78,84,86–91}, its efficacy was found to be antibody- and time-dependent^78,90,92. However, decoupling Src from actin remodeling via the E-cadherin-based cortactin seemed not critical for epidermal blistering in an ex-vivo model in presence of anti-Dsg3 and Dsg1 antibodies as described by Kugelmann et al.⁹⁰. This may suggest that the anti-Dsg1 antibodies in PV can activate rescue signaling pathways to prevent blistering. Gil et al. revealed that FAK is significantly phosphorylated at tyrosine 397/925 by PV IgG through epidermal growth factor receptor (EGFR) activation, and FAK inhibition abolished PV IgG induced blister formation in the neonatal passive transfer mouse model⁸⁷. Later, they further elucidated the signaling pathways involved and showed that activation of EGFR induces phosphorylation of FAK through the neuronal nitric oxide synthase (nNOS) pathway in PV⁸⁸. FAK phosphorylation at tyrosine residue 397 is considered an autophosphorylation site, while tyrosine residue 925 is Src or EGFR dependent. As EGFR and contextual Src activation play an important role in PV pathogenesis, mechanosensing through FAK activation can also be anticipated in PV^93,94. Interestingly, addition of fibronectin, a binding partner of integrins and an activator of FAK to an ex vivo model of PV was found to reduce acantholysis and intraepidermal binding of the pemphigus antibodies⁹⁵.

Cytoskeleton

Keratins, the key components of intermediate filaments in the epidermis, play a pivotal role in the anchoring and stabilization of desmosomes and are essential for maintaining proper desmosomal adhesion and tissue integrity^96,97. Furthermore, they are long known to contribute to mechanosensing for example via Keratin8/18-mediated modulation of the small GTPase RhoA⁸² which can be controlled by p38MAPK⁶³. The typical morphological characteristic of keratin filament retraction from cell borders is a common feature observed across all clinical phenotypes of pemphigus and can be detected in in vitro and ex vivo pemphigus models as well as in patients’ non-lesional and lesional epithelia, thus preceding bona fide blister formation. Furthermore, keratins account for the mechanical properties of keratinocytes, underscoring their critical importance in the maintenance of tissue integrity. Vielmuth et al.⁷⁶ found that the knockout of keratin cluster II significantly impaired intercellular adhesion compared to wild-type cells while the subsequent treatment with AK23, PV-IgG or PF-IgG did conceivably not further compromise intercellular adhesion. Although the mechanism by which the desmoplakin/keratin anchorage is compromised in PV is still largely unknown, it can be anticipated to occur through inside-out signaling under control of the many competent effectors downstream of PV antibody-mediated Dsg3/Dsg1 receptor signaling (Figs. 1–3).

Studies on PV attest to the critical role of the actin cytoskeleton in regulating the crosstalk between E-cadherin- and Dsg3 to assemble and disassemble desmosomes^34,98. Patel et al.²⁷ were first to address the direct causal contribution of changes in actin and tubulin cytoskeletal in PV. Using mouse keratinocytes, they demonstrated that the preincubation with inhibitors of actin (Cytochalasin D) or tubulin (Colchicine) polymerization, both individually and in combination, failed to diminish the internalization and binding of PV IgG upon stimulation with PV sera. Conversely, Gliem et al.⁴⁰ found that pretreatment with a 10-fold lower dose of Cytochalasin D, enhanced internalization and aggravated PV IgG induced loss of intercellular adhesion in HaCaT keratinocytes. Consistently, the incubation of cells with the actin-stabilizing agent jasplakinolide enhanced junction stability, reduced the pathogenicity in terms of PV IgG-induced sheet fragmentation and rescued Dsg3 localization at cell borders. Jennings et al.⁴⁷ tested another inhibitor of actin polymerization (Lantraculin D) on human keratinocytes with similar results.

A change in mechanical forces requires actin cytoskeleton re-organization, more precisely contraction of actin through the actomyosin motor proteins. The actomyosin activity can be modulated by the calcium-binding protein calmodulin, which triggers the phosphorylation of myosin light chains and the subsequent contraction of the actomyosin complex. Sánchez-Carpintero et al.⁹⁹ utilized the calmodulin inhibitor W-7 hydrochloride to prevent PV IgG induced blister formation in neonatal mice. Similarly, Sumigray et al.⁵⁰ revealed that treating mouse keratinocytes with AK23 leads to increased contractility, and knocking out Myh9 (Myosin IIA) inhibits AK23-induced cell sheet fragmentation, suggesting a role for myosin IIA and actomyosin contraction in PV loss of adhesion. Intriguingly, short-term RhoA-mediated actin contraction and PKCα activation of desmoplakin/keratin interaction was found to result in desmosome assembly.

Modulators of actin dynamics are the Rho family of small GTPases as well as cortactin, actomyosin and α-adducin, which also stabilize the junctional cortex^100–102. Specifically, the Rho family of GTPases are mechano-transducers which function as highly complex molecular switches between fates by regulating actin polymerization, cell polarity, vesicular trafficking, cell contraction, and mitosis¹⁰³. Accordingly, transient RhoA/Rac activation is involved in the stabilization of transadhering E-cadherin-based contacts and maintenance of Myosin II association to uphold junctional contractile force regulated, for example, by cortactin and α-adducin phosphorylation to drive the desmoplakin-keratin interaction for desmosome assembly and keratinocyte differentiation^{90,98,104,105}. However, under mechanical tension, as introduced above, sustained RhoA activation can implicate K8/K18 and JUP recruitment to sites of cellular tension to reinforce F-actin stress fibers⁸² while compromising desmosome plaque maturation⁹⁸, a process likely implicating p38MAPK⁶³. The subsequent loss of tension and RhoA/Rac-1 inhibition would then lead to rapid disruption of junctional integrity^34,105,106.

Indicative of inhibition of Rho family members in desmosome disassembly in PV, the non-selective and selective RhoA activators CNF-1 (RhoA, Rac1, and Cdc42 activators) and CNFy (RhoA) significantly reduced weakening of desmosomes in ex vivo and in vitro models in response to PV IgG^35,40,72,92. Waschke et al.³⁵ also showed that CNFy increased stress fiber formation along junctions and abolished cell sheet fragmentation and desmoplakin re-organization. In a follow-up study the group further confirmed that inhibition of actin polymerization (using Lantraculin B), the target of RhoA, abolished the protective effect of this RhoA activator⁴⁰. Although the above-mentioned studies used patient-derived PV IgG, Jin et al. and Rötzer et al. reported similar results with PV IgG and the pathogenic monoclonal anti-Dsg3 antibody AK23^36,72,107. Additionally, Rötzer et al. discovered that the protective effect of pharmacological RhoA activation is diminished by silencing α-adducin, increasing fragmentation, compatible with the understanding that the actin-binding protein α-adducin is regulated by RhoA and PKC¹⁰⁸ to promote E-cadherin/Dsg3 interaction and stabilization at the plasma membrane preceding desmosome formation¹⁰⁹. Intriguingly, Rötzer et al. further demonstrated that α-adducin is serine-phosphorylated, hence activated, in response to AK23 or PV IgG, and found this to occur via calcium influx and PKC⁷². The authors thus discussed that α-adducin activation through calcium/PKC-mediated phosphorylation represents a rescue mechanism in PV. Collectively, these studies suggested that Dsg3 antibody-mediated inhibition of Rho GTPases alters the cytoskeletal architecture in keratinocytes preceding acantholysis, potentially paralleled by a rescue mechanism involving RhoA activation and tyrosine phosphorylation of α-adducin on Ser726. Beyond pharmacological modulation of mechanical structures and mechanosensing pathways, Jin et al.³⁶ applied cyclic mechanical stretching of cell sheets. They conceivably found this to have the opposite effect than the suggested release of tension/transadhesion on Dsg3 by PV antibodies described above, namely, to reduce p38MAPK phosphorylation and enhance RhoA recruitment to cell boarders which, in line with the findings above, mitigated AK23-induced loss of intercellular adhesion. This is compatible with the concept of protective mechanisms discussed by Rötzer et al.¹⁰⁸ and potentially with regenerative mechanisms following shrinkage of desmosomes and blister formation reported in hair follicle stem cells of the adult passive transfer mouse model for PV¹⁶. However, as discussed by Jin et al.³⁶, it can currently not be distinguished whether pharmacological Rho GTPase activation prior blistering prevents an initial Rho GTPase inhibition in PV (Table 1) or strengthens the E-cadherin/actin/actomyosin complex to bear increased mechanical force and balance loss of tension at desmosomal complexes. Notwithstanding, both possibilities are indicative of inside-out signaling weakening desmosomes, urging the application of advanced longitudinal techniques to respond to these questions.

Plasma membrane and Endocytosis

Endocytosis is a mechanosensitive process that involves precise orchestration of tightly regulated mechanical cues such as dynamic remodeling of the plasma membrane that occurs at the sub-micrometer level. Mechanistically, a range of conditions influence the membrane remodeling process including plasma membrane tension¹¹⁰. Signaling-wise, actin-based endocytosis can potentially implicate some of the here described key effectors in PV such as p38MAPK, RhoA, cdc42 and PKCα^111,112. Delva et al.¹¹³ evaluated the role of different endocytic pathways in PV IgG-induced-Dsg3 internalization. They found that caveolin-1 or dynamin II knockout did not affect AK23-induced Dsg3 internalization in cells. However, the alteration of lipid rafts via filipin or nystatin treatment drastically prevented the PV-IgG or AK23-induced internalization of Dsg3, suggesting a clathrin- and dynamin-independent but cholesterol-dependent mechanism for Dsg3 endocytosis in PV. Consistent with these findings, two other studies using another lipid raft inhibitor, methyl-β-cyclodextrin, reported similar results^73,81. Interestingly, the inhibition of Dsg3 endocytosis was found to prevent the antibody-induced loss of intercellular adhesion, but did not affect keratin retraction or p38MAPK phosphorylation, indicating that these pathological phenomena may be independent processes⁸¹. Most intriguingly, Pigmentosa GTPase regulator interacting protein 1-like (RPGRIP1L) was found to specifically control the endocytosis of desmoglein but not desmocollin cadherins, and the knockout of RPGRIP1L in mice led to shrinkage of desmosomes and spontaneous development of suprabasal blisters reminiscent of PV. Furthermore, in HaCaT keratinocytes, RPGRIP1L knockdown resulted in the internalization of Dsg3 and a loss of intercellular adhesion. Notably, the treatment of RPGRIP1L knockdown cells with PV antibodies further reduced cell-cell adhesion, potentially suggesting complementary pathways¹¹⁴, such as described above with regards to keratin retraction/p38MAPK activation and Dsg3 endocytosis⁸¹.

The plasminogen activator enzyme (Pa) system has been linked to diverse cellular effects. Notably, Pa functions as both proteolytic enzyme (to convert e.g., plasminogen to plasmin) and signal transducer and interacts with a variety of mechanosensitive receptors, including EGFR, cadherins and integrins¹¹⁵. Two seminal studies by Hunziker and Morioka evaluated the role of this system in PV. Hunziker and Vessalli¹¹⁶ demonstrated that either addition of plasminogen or plasmin to an ex vivo PV model or high doses of plasmin alone were capable of enhancing or inducing acantholysis, respectively. Furthermore, Morioka et al.¹¹⁷ revealed that an anti-urokinase plasminogen activator (uPA) antibody completely blocked PV IgG-induced acantholysis, subsequently confirmed by multiple groups using various uPA inhibitors^118–122. The discussion regarding the involvement of the Pa system came to a conclusion by a thorough analysis by Mahoney et al.¹²³, which showed uPA or tissue-type plasminogen activator (tPA) KO alone or combined have no impact on PV IgG-induced blister formation in neonatal mice. It is noteworthy, however, that tPA KO animals had a tendency for milder disease.

These collective results support the notion that biomechanical signaling through desmosomal cadherins, which is a relatively recent field of research, plays a key role in desmosome shrinkage in PV. It implicates Rho GTPases and the re-organization of the actin/actomyosin cytoskeleton while highlighting a crosstalk between E-cadherin- and Dsg3/1-based plasma membrane exposed molecules and/or junctions in balancing cellular forces. These results give guidance for further research on identifying and discriminating between signals by these respective stress-bearing cadherins including the crosstalk to integrins in epithelial cells and PV.

Electrical signaling

Only a limited number of studies, in total 22, have addressed causative electric potential/ion signaling in PV pathogenesis with main focus on the role of calcium ions and calcium-responsive molecules summarized in Table 2 and Fig. 2. Occurring within seconds after anti-Dsg1 or anti-Dsg3 antibody incubation, as first reported by Seishima et al. in the squamous cell carcinoma keratinocyte line DJM-1³¹, an increase in intracellular calcium flux is the fastest signal recorded so far after antibody incubation. It can be mediated through various mechanisms, such as activating calcium channels at the plasma membrane or releasing calcium from intracellular stores like the endoplasmic reticulum¹²⁴. Upon stimulation, opening of plasma membrane calcium channels allows extracellular calcium to enter the cytosol, while calcium release from intracellular stores can be triggered by IP3R activation caused by Inositol 1,4,5-trisphosphate (IP3) production through Phosphatidylinositol 4-Kinase alpha (PI4Kα) and Phospholipase C (PLC). Seishima et al. showed an increase in IP3 and release of calcium from intracellular stores following anti-Dsg1 or anti-Dsg3 antibody treatment in keratinocyte 2D cultures³¹.

Table 2.

Characteristics of studies included in bioelectrical signaling axis

Author	Year	Target molecule		Intervention	Treatment Modality	Model	Antibody	Preventing Pathology	Measured Outcome
Calcium, calcium channels, and calcium activated proteins
Arredondo et al.125	2005	Calpain I and II		MDL-28170 (10 µM), PD-150606 (50 µM), Calpastatin peptide (50 µM), MDL-28170 (10 µM) + PD-150606 (50 µM) + Calpastatin peptide (50 µM)	Cotreatment	In vitro, PTNM	PV IgG	+/-	Morphometric analysis of acantholysis in vitro and in vivo
Kalantari-Dehaghi et al.126	2013	Calcineurin		Cyclosporine A (0.1 μM)	Cotreatment	PTNM	PV IgG	+	Morphometric anaylsis of acantholysis
Rötzer et al.72	2014	Intracellular Calcium		BAPTA-AM (50 µM)	Cotreatment for 4 h	In vitro	PV IgG	+	KDA
Walter et al.46	2019	Intracellular Calcium		BAPTA-AM (200 µM)	Pretreatment for 4 h	In vitro	mPV IgG, mcPV IgG, PF IgG	+	KDA
Schmitt et al.61	2021	CRAC, PI4KA		BTP-2 (NA), GSK-F1 (NA)	Pretreatment for 1 h	In vitro, Ex vivo	mPV IgG, mcPV IgG, AK23, PF IgG	+	In vitro: KDA and keratin retraction measured by immunofluorescence microscopy. Ex vivo: Gross and microscopic evaluation of tissue.
Schmitt et al.78	2023	CRAC, PI4K		BTP-2 (10 µM), GSK-F1 (10 nM)	Pretreatment for 1 h	In vitro	2G4	-	KDA
Xie et al.127	2023	Calcineurin		FK506 (100 nM)	Cotreatment	In vitro	PV sera, AK23	+	KDA, Immunofluorescence microscopy, and western blotting was performed for Dsg3 depletion
PLC, IP3R and PI3K
Esaki et al.128	1995	PLC	U73122 (10 µM)		Pre and cotreatment	In vitro	PV IgG or sera	+	Keratin retraction measured by immunofluorescence microscopy
Sánchez-Carpintero et al.99	2004	PLC	U73122 (100 µg/g)		Pretreatment for 3 h	PTNM	PV IgG	+	Gross and microscopic evaluation of tissue
Burmester et al.77	2020	PI3Kα, PI3Kβ	A66 (0.1, 1, and 10 µM), TGX-221 (0.1, 1, and 10 µM)		In vitro: Pretreatment for 2 h then discarding supernatant. Ex vivo: Cotreatment	A66: In vitro, ex vivo. TGX-221: in vitro	In vitro: Immunophresis material from PV patient. ex vivo: PX43 scFv	A66 ( + ), TGX-221 (-)	In vitro: KDA. Ex vivo: Gross and light microscopic evaluation of blister formation.
Schmitt et al.61	2021	PLC, IP3R	U73122 (Cell culture: NA, HSOC: 50 µL of 4 µM), Xestospongin C (Cell culture: NA, HSOC: 50 µL of 2 µM)		Pretreatment for 1 hour	In vitro, Ex vivo	mPV IgG, mcPV IgG, AK23, PF IgG	+	In vitro: KDA and keratin retraction measured by immunofluorescence microscopy. Ex vivo: Gross and microscopic evaluation of tissue.
Egu et al.130	2022	PLC, IP3R	U73122 (50 µL of 4 µM), Xestospongin C (50 µL of 2 µM)		Pretreatment for 1 h	Ex vivo	PV IgG	+/-	length of desmosomes, percentage of split desmosomes, and keratin retraction was measured by electron microscopy.
Hiermaier et al.129	2022	PLC, IP3R	U73122 (4 µM), Xestospongin C (0.4 µM)		Pretreatment for1 hour	In vitro	PV IgG, PF IgG	+	KDA
Schmitt et al.78	2023	PLC, IP3R	U73122 (4 µM), Xestospongin C (2 µM)		Pretreatment for 1 h	In vitro	2G4	-	KDA
PKC
Kowalewski et al.133	1994	PKC	Isoquinoline sulfonamide (H7)		-	In vitro	PV sera	+	Cell detachment assay
Sánchez-Carpintero et al.99	2004	PKC	Bisindolylmaleinamide (100 µg/g)		Pretreatment for 3 h	PTNM	PV IgG	+	Gross and microscopic evaluation of tissue
Cirillo et al.134	2010	PKC	Go6976 (5 nM)		Pretreatment for 1 h	In vitro	PV sera	+	Number of single cells were counted as the index of detachment
Spindler et al.135	2011	PKC, PKCα	Go6976 (PTNM: 500 nM in 50 µL, in vitro: 500 nM, ex vivo: NA), Safingol (PTNM: 40 µM in 50 µL, in vitro: 40 µM, ex vivo: NA), PKCα siRNA		PTNM: pretreatment for 2 h and Cotreatment. Invitro and ex vivo: Cotreatment	In vitro, ex vivo, PTNM. PKC siRNA only in vitro	PV IgG	Go6976 and Safingol (+), PKC siRNA (-)	PTNM: Gross and microscopic evaluation of blister formation. In vitro and ex vivo: immunofluorescence microscopy and KDA
Dehner et al.136	2014	PKC	Bisindolylmaleimide X hydrochloride (Bim-X, 1 µM)		Cotreatment	In vitro	PV IgG	+	KDA and keratin retraction measured by immunofluorescence microscopy
Rötzer et al.72	2014	PKC	Bim-X (1 µM)		Cotreatment	In vitro	PV IgG	+	KDA
Walter et al.75	2017	PKC	Bim-X (NA)		Pretreatment for 1 h	In vitro	AK23, mcPV IgG, mPV IgG, atPV IgG, PF IgG	+	KDA
Egu et al.205	2019	PKC	Bim-X (50 µL of 1 µM)		Pretreatment for 1 h	Ex vivo	PV IgG	-	Blister score and cleft length was calculated based on H&E stained sections. Evaluation of desmosomes were performed by electron microscopy

Treatment modalities: Pretreatment, Intervention added before antibody application. Cotreatment, Intervention added simultaneously with the antibody. Posttreatment, Intervention added after antibody application. Model: Ex vivo, human skin organ culture. In vitro, 2D cultured keratinocyte cell lines or primary cells. PTNM, Passive transfer neonatal mouse model. PTAM, Passive transfer adult mouse model. Antibodies: mPV IgG, mucosal PV IgG. mcPV IgG, mucocutaneous PV IgG. atPV IgG, Atypical PV IgG. PF IgG, pemphigus foliaceus IgG. Preventing Pathology: ( + ) effective in preventing antibody-induced pathology, (-) Ineffective in preventing antibody-induced pathology, (x) worsened antibody-induced pathology. KDA Keratinocyte dissociation assay, NA Not Available.

Fig. 2. Diagram of the signaling network of bioelectrical effectors activated by PV antibody binding to Dsg3 or Dsg1/3.

The figure depicts the reported pathways and molecular interactions centering around the increase of intracellular calcium concentration and consequences thereof (boxes at the bottom). Arrows indicate the interactions between specific molecules involved in the signaling cascade. Effector molecules that have been tested across various experimental PV models and antibody types are highlighted in bold to emphasize their roles within the bioelectrical pathways. This scheme outlines bioelectrical signaling events initiated by PV antibody binding to Dsg proteins, illustrating the potential intervention points and molecular interactions. This figure was created in BioRender. Hariton, W. (2025) https://BioRender.com/3f3hvc2.

Calcium, calcium channels, and calcium activated proteins

Rötzer et al. and Walter et al. confirmed that chelation of intra- and extracellular calcium with BAPTA-AM reduces fragmentation in vitro caused by anti-Dsg1, Dsg3/1 and Dsg3 antibodies, while, conversely, only anti-Dsg1 containing PV IgG caused a measurable cytosolic calcium influx^46,72. Furthermore, Arredondo et al. demonstrated that inhibiting the calcium-activated cysteine proteases Calpain I and II using MDL-28170, PD-150606, Calpastatin peptide, alone or in combination ameliorates PV IgG-induced acantholysis in both in vitro and in vivo PV models. However, the effectiveness was only seen for one of the two tested PV IgG antibodies in the in vivo model, with no information on the anti-Dsg1/3 profile¹²⁵. Calcineurins, activated by Calcium influx and calmodulins upstream of nuclear translocation of NFAT transcription factors, were also shown to be involved in PV pathogenesis^126,127. Schmitt et al., also investigated the role of Calcium release-activated channels (CRAC), and PI4KA in two separate studies^61,78. Their research demonstrated that PI4KA and CRAC inhibitors effectively prevented antibody-induced loss of intercellular adhesion restricted to anti-Dsg1 antibodies or PV IgG containing both anti-Dsg1 and 3 antibodies but not anti-Dsg3 antibodies alone or AK23 or 2G4. Although these results contrast with initial findings³¹, it is noteworthy that anti-Dsg3 antibody containing PV IgG used in these two studies were not able to fragment normal epidermal keratinocyte sheets, suggesting experimental conditions with high expression of compensatory Dsg1 levels preventing fragmentation¹¹. This suggested that complementary pathogenic signaling pathways can be activated by the anti-Dsg1 antibodies in PV IgG. However, recent studies have challenged this notion, revealing that intracellular calcium influx is specific for the anti-Dsg3/Dsg1-targeted epitope^46,78. Notably, anti-Dsg3 antibodies targeting the extracellular domain 1 (EC1) such as AK23, a monoclonal anti-Dsg3 antibody, failed to induce calcium influx in normal human keratinocytes, whereas 2G4, another monoclonal antibody against the membrane proximal part of Dsg3, triggered calcium influx. These results need further investigations as they seemingly are not compatible with findings on IP3 inhibition which is later discussed in detail⁸¹. Furthermore, only inhibition of CRAC, but not PLC, IP3R, or PI4Kα, resulted in the suppression of 2G4-induced intracellular calcium influx, suggesting an alternative pathway for opening intracellular calcium channels besides the PLC pathway. Inhibitors of upstream signaling of calcium have also shown varying results with monoclonal anti-Dsg3 antibodies. While PLC inhibition partially abrogated AK23 induced loss of intercellular adhesion, this effect was not observed with 2G4^61,78. Altogether, these results suggest the testable hypothesis that the Dsg1 expression level of keratinocytes is crucial to study the ability of anti-Dsg3 antibody alone on triggering a calcium influx.

PLC, IP3, IP3R, Ca2 + , PKC

The PLC/IP3/calcium release axis from intracellular stores relies on PLC recruitment to cell adhesion receptors at the plasma membrane such as Dsg3 and Dsg1⁶¹; hence, hindering PLC can diminish calcium release from intracellular stores as well as through extracellular ion channels. Several studies have explored the involvement of PLC in PV pathogenesis through various models. Esaki et al., Hiermaier et al. Schmitt et al. 2021 and 2023 utilized in vitro models. Except for Schmitt et al. 2023 who used 2G4 (a monoclonal pathogenic antibody against the juxtaposed membrane domain EC5 of Dsg3), all studies showed PLC inhibition to be effective in preventing PV IgG-mediated (with and without anti-Dsg1 antibodies) loss of intercellular adhesion^{61,78,128,129}. This was further supported by the findings of Sánchez-Carpintero et al. and Schmitt et al. in 2021, using passive transfer neonatal mouse and ex vivo PV models, respectively^61,99. Through electron microscopy, Egu et al. further illustrated that inhibiting PLC notably reduced PV IgG-induced keratin retraction, though it did not affect desmosome shrinkage, numbers, or splitting¹³⁰, consistent with p38MAPK-dependent and independent pathways in PV⁸¹ (Fig. 1).

Following the activation of PLC and the subsequent production of IP3, the inositol IP3 receptor (IP3R) is activated, leading to the opening of calcium channels on the tricellular organelles such as the endoplasmic reticulum. Schmitt et al. 2021, 2023, Egu et al. and Hiermaier et al. used IP3R inhibitors to validate implication of this receptor in PV pathogenesis^{61,78,129,130}. In line with results of PLC inhibition, all studies showed that IP3R inhibition is effective in preventing antibody induced loss of intercellular adhesion, except for 2G4. Moreover, electron microscopic evaluation revealed that IP3R inhibition significantly decreased PV IgG-induced desmosome splitting and keratin retraction while it had no effect on shrinkage of desmosomes, presumably being modulated by enhanced Dsg3 turn-over, which was unexpected in light of the activities of PLC described above⁸¹.

Protein Kinase C (PKC) which is activated downstream of deacyl glycerol (DAG) or Ca2+ release from intracellular stores (Figs. 2 and 3), plays a crucial role in signal transduction, regulating the structure, composition, and function of desmosomes¹³¹. Osada et al. were the first to demonstrate that antibody stimulation activates Ca2 + - and diacylglycerol (DAG)-dependent PKCα, as well as DAG-dependent PKCs δ and η, and atypical PKC (ζ), with recruitment to the plasma membrane, occurring as early as 30 s post-stimulation¹³². This activation aligns with increased Ca2+ influx and phospholipase C (PLC) activation described above. Various studies employing in vitro and patient-derived PV models have indicated that PKC inhibition can prevent PV-related pathologies^{72,75,99,133–136}. However, findings from ex vivo models remain inconclusive; while Spindler et al. reported that PKCα inhibition prevents Dsg3 depletion, Egu et al. recently found that PKC inhibitors are ineffective in preventing blister formation, desmosome shrinkage, or splitting^74,135, results requiring further investigations.

In summary, the electrical signaling axis implicated in PV pathogenesis has been primarily associated with intracellular calcium influx and its related signaling effectors. While it was previously believed that calcium influx depends on Dsg1 and not Dsg3, this finding will be revisited by carefully monitoring Dsg1 expression levels. Furthermore, other ion channels will need to be addressed in PV pathogenesis.

Biochemical signaling

Activation or inhibition of biochemical effectors downstream of Dsg3/Dsg1 has been most widely studied in PV and integrates a variety of components belonging to known signaling pathways. These include specific plasma membrane receptors such as growth factor receptors, canonical Wnt or G-protein-coupled receptors as well as receptors of the TNF superfamily. A total of 91 studies addressed the modulation and pathogenicity of these effectors functionally by pharmacological modulators or gene knockout. All 91 studies of the biochemical signaling axis are summarized in Table 3 and effector signals in Fig. 3.

Table 3.

Characteristics of studies included in biochemical signaling axis

Author	Year	Target molecule	Intervention	Treatment Modality	Model	Antibody	Preventing Pathology	Measured Outcome
Receptor tyrosine kinase growth factor receptors and signaling
Hunziker et al.139	1986	PDGF	Stimulated or sonicated platelets (100 µL/mL)	Cotreatment	Ex vivo	PV sera	-	H&E staining light microscopy examination of blister formation and acantholysis
Frušić-Zlotkin et al.43	2006	EGFR	AG1478 (10 µM)	Cotreatment	In vitro	PV IgG	+	Qualitative visualization of acanthoylsis by phase contrast microscopy
Heupel a et al.92	2009	EGFR	GW2974 (10 µM)	Pretreatment for 2 h	In vitro	PV IgG	-	KDA and keratin or F-actin retraction measured by immunofluorescence microscopy
Pretel et al.86	2009	EGFR	Erlotinib (100 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Clinical activity and microscopic evaluation of acantholysis
Gil et al.87	2012	EGFR	CL-387785 (10 µg/g)	Pretreatment for 2 hours	PTNM	PV IgG	+	Gross and microscopic evaluation of blister
Bektas et al.69	2013	EGFR	CL-387785 (10 µM), PD15330 (10 µM), BPIQ-II (10 µM), Gefitinib (5–10 µM), Erlotinib (5–10 µM), AG1478 (in vitro: 0.1–10 µM, PTNM: 4 µg)	AG1478: Pretreatment for 2 h. Other inhibitors: overnight incubation or 2 h	In vitro, PTNM	PV IgG, AK23	+	In vitro: keratin retraction and Dsg3 internalization was visualized with immunofluorescence microscopy. Dsg3 depletion was measured by western blot. Loss of intercellular adhesion was measured by KDA. PTNM: Gross and microscopic evaluation of blister formation
Espana et al.88	2013	EGFR	CL-387785 (10 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Gross and microscopic evaluation of blister formation
Sayar et al.137	2014	EGFR	Erlotinib (In vitro: 0.4–4 µg/mL, PTNM: 4–100 µg/g), Lapatinib (4–45 µg/g), EGFR KO	In vitro: Pretreatment for 6 h. PTNM: Pretreatment for 4 h	In vitro, PTNM	AK23 with suboptimal dose of patient derived anti-dsg1 antibody, PV IgG	+	In vitro: KDA. PTNM: Gross and microscopic histology of skin and palate. Length of blister was quantified relative percentage to total length of section.
Walter et al.46	2019	EGFR	Erlotinib (2.5 µM)	Pretreatment for 1 h	In vitro	mPV IgG, mcPV IgG, PF IgG	+	KDA
Ivars et al.91	2020	EGFR	CL-387785 (10 µg/g), Cetuximab (100 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Gross and microscopic evaluation of blister formation with histological scoring
Burmester et al.77	2020	VEGFR2	Vandetanib (0.1, 1, and 10 µM)	In vitro: Pretreatment for 2 h then discarding supernatant. Ex vivo: Cotreatment	In vitro, ex vivo	In vitro: Immunophresis material from PV patient. ex vivo: PX43 scFv	+	In vitro: KDA. Ex vivo: Gross and light microscopic evaluation of blister formation.
Hariton et al.16	2023	EGFR	Lapatinib (9 µg/g)	Pretreatment for 4 h and post-treatment after 4 days	PTAM	AK23	+	Percentage of hair follicle blistering was evaluated by H&E staining
Egu et al.85	2024	EGFR (ErbB1 and ErbB1/2), TrkA	Lapatinib (44 nM), Erlotinib (2.5 µM), siTRKA (siRNA)	Lapatinib and Erlotinib: Pretreatment for 1 h. siTRKA: Pretreatment for 24 h	Lapatinib and Erlotinib: ex vivo and in vitro. siTRKA: in vitro	Lapatinib and Erlotinib: PV IgG. siTRKA: AK23, PV IgG	+	in vitro: KDA and STED microscopy. Ex vivo: Blister scoring based on H&E staining and EM.
Src
Chernyavsky et al.84	2007	Src	PP2 (10 µM)	Pretreatment	In vitro	PV IgG	+	Cell volume and keratin aggregation were measured
Heupel a et al.92	2009	Src	PP2 (10 µM)	Pretreatment for 2 h	In vitro	PV IgG 1–5	-	KDA and keratin or F-actin retraction measured by immunofluorescence microscopy
Pretel et al.86	2009	Src	PP1 (1 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Clinical activity and microscopic evaluation of acantholysis
Gil et al.87	2012	Src	PP1 (1 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Gross and microscopic evaluation of blister
Espana et al.88	2013	Src	PP1 (1 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Gross and microscopic evaluation of blister formation
Cirillo et al.89	2014	Src	Src inhibitor-1 (1 µM)	Pretreatment	In vitro	PV IgG	+	Morphometric analysis of cell-cell detachment and KDA
Walter et al.75	2017	Src	PP2 (NA)	Pretreatment for 1 h	In vitro	AK23, mcPV IgG, mPV IgG, atPV IgG, PF IgG	+	KDA
Kugelmann et al.90	2019	Src	PP2 (In vitro: 10 µM, ex vivo and PTNM: 50 µL of 10 µM)	In vitro: Pretreatment for 2 h, ex vivo and PTNM: Cotreatment	PTNM, In vitro, ex vivo	PV1-4 IgG, AK23	+	In vitro: KDA. Ex vivo and PTNM: Blister score was counted in H&E stained samples
Ivars et al.91	2020	Src	PP1 (1 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Gross and microscopic evaluation of blister formation with histological scoring
Schmitt et al.78	2023	Src	PP2 (10 µM)	Pretreatment for 1 h	In vitro	AK23, 2G4	+	KDA. Dsg3 internalization, keratin retraction, and desmosome numbers were visualized using STED microscopy.
Caspases
Wang et al.160	2004	Caspase 1	YVAD-CHO (100 nM)	Cotreatment	Ex vivo	PV IgG	+	Gross and microscopic evaluation of acantholysis
Arredondo et al.125	2005	Caspases	Z-DCB-MK, DEVD-CHO, Z-DEVD-FMK, Z-DCB-MK + DEVD-CHO + Z-DEVD-FMK (all 10 µM)	Cotreatment	In vitro	PV IgG-1a/PV IgG-2b	+	Morphometric analysis of acantholysis
Pretel et al.86	2009	Caspases	cpm-VAD-CHO (1.6 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	clinical activity and microscopic evaluation of acantholysis
Schmidt et al.149	2009	Caspases	VAD-fmk (20–300 µM)	Pretreatment for 1 h	In vitro	PV IgG	-	KDA and Dsg3 Immunofluorescence staining to detect keratin retraction, and intercellular gaps
Pacheco-Tovar et al.157	2011	Caspases	Ac-DEVD-CMK (20 mM in 50 µL)	Pretreatment for 2 h	PTNM	PV IgG	+	Gross and microscopic evaluation of blister formation
Gil et al.87	2012	Caspases	cpm-VAD-CHO (1.6 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Gross and microscopic evaluation of blister
Espana et al.88	2013	Caspases	cpm-VAD-CHO (1.6 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Gross and microscopic evaluation of blister formation
Luyet et al.158	2015	Caspase 2, Caspase 3, Caspase 8, Caspase 9, Caspase 12, Caspases	Z-DEVD-FMK (40 µM), Ac-DEVD-CMK (In vitro: 40 µM, PTNM and PTAM: 6 µg/g), Z-IETD-FMK (100 µM), Z-LEHD-FMK (50 µM), Z-ATAD-FMK (10 µM), Z-VAD-FMK (40 µM)	Pretreatment	All In vitro, Ac-DEVD-CMK: PTNM and PTAM	Ac-DEVD-CMK and Z-DEVD-FMK: PV IgG1/2, AK23, AK23/PF. Z-IETD-FMK, Z-LEHD-FMK, and Z-ATAD-FMK: PV IgG1/2, AK23. Z-VAD-FMK: AK23	Ac-DEVD-CMK, Z-DEVD-FMK, and Z-VAD-FMK ( + ), Z-IETD-FMK, Z-LEHD-FMK, and Z-ATAD-FMK (-)	In vitro: KDA. PTNM and PTAM: Quantified hair follicle blister and length of blister
Hariton et al.159	2017	Caspase 3	K14 promoter-driven keratinocyte-specific caspase-3-deficient mice (casp3EKO)	-	PTAM	AK23	+	Histological quantification of telogen hair follicle blisters
Eichkorn et al.161	2022	Caspases + UVA	Z-VAD-FMK + UVA (NA)	Pretreatment	In vitro	AK23	+	KDA
Wnt Pathway
Williamson et al.25	2006	c-Myc, GSK3β	c-Myc inhibitor 3 (5404711, 6 µM), MI-1 (5521700), SB216763 (2 µg/g), Lithium Chloride (200 µg/g)	Pretreatment for 2 h	PTNM	c-Myc inhibitor 3: PV IgG, AK23. SB216763: AK23	+	Gross and microscopic evaluation of blister formation
Schlögl et al.81	2018	Casein Kinase 1 (CK1)	D4476 (100 µM)	Pretreatment for 1 h	In vitro	PV IgG, AK23	x	KDA and immunofluorescence microscopy
Hariton et al.16	2023	BIO	GSK-3 (10 µg/g daily for 2 days)	Injection after 4 days of AK23 treatment	PTAM	AK23	+	Percentage of hair follicle blistering was evaluated by H&E staining
Glucocorticoid and G protein-coupled receptors
Schiltz et al.164	1979	Glucocorticoid receptor	Hydrocortisone, Triamcinolone acetide (10 µM)	Pretreatment	Ex vivo	PV IgG	-	dIF and H&E staining light microscopy examination of acantholysis
Anhalt et al.165	1986	Glucocorticoid receptor	Dexamethansone (10 and 20 mg/kg/day in 48 h period)	24 h pretreatment and simultaneous with PV IgG injection	PTNM	PV IgG	-	Clinical extent scoring, and H&E staining light microscopy evaluation of blister formation and acantholysis
Naito et al.166	1989	Glucocorticoid receptor	Dexamethansone (20 µg/g)	Pretreatment	PTNM	PV IgG	-	Visual examination for Nikolsky sign as well as histological examination of acantholysis
Grando et al.168	1993	Acetylcholine receptor	Acetylcholine, Bethanechol, Carbachol or Methacholine (All 1 nM to 100 µM)	1 h after addition of PV IgG	In vitro	PV IgG	+	Qualitative visualization of acanthoylsis by phase contrast microscopy. Measuring permeability of [‘H]thymidine by transwell experiment
Asano et al.121	2001	Glucocorticoid receptor	Dexamethansone (0.1, 1 mM)	Pretreatment for 24 h	In vitro	PV IgG	-	Immunofluorescence examination of keratin retraction and desmoplakin distribution
Nguyen a et al.154	2004	Glucocorticoid receptor	Methylprednisolone (15 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Gross and microscopic evaluation of acantholysis and blister formation
Nguyen b et al.155	2004	Acetylcholine receptor, acetylcholine esterase	Carbacol (0.04 µg/g), Pyridostigmine bromide (0.1 µg/g)	Cotreatment	PTNM	PV IgG	+	Gross and microscopic evaluation of acantholysis and blister formation
Chernyavsky et al.169	2008	α7 nicotinic acetylcholine receptor, Muscarinic acetylcholine receptor M1	AR-R17779 (100 µM), Pilocarpine (50 µM)	Cotreatment	In vitro	PV IgG	+	Measuring permeability monolayer 1 hour after treatment with [3H]thymidine by transwell experiment
Spindler et al.170	2010	ß-Adrenergic receptor, Adenylyl cyclase (cAMP)/ Phosphodiesterase-4 (cAMP), Protein Kinase A (PKA)	Isoproterenol (in vitro: 100 µM, PTNM: 10 µM), Propranolol (10 µM), Forskolin/ Rolipram (5 µM/10 µM, H89 (10 µM)	PTNM: Pretreatment for 2 h or simultaneous. in vitro: simultaneous, 24 h after PV IgG treatment with media exchange for 2 h	Isoproterenol and propranolol: In vitro and PTNM. Forskolin/Rolipram, H89: In vitro	PV IgG	Isoproterenol (+), Propranolol (-), Forskolin/ Rolipram (+), H89 (-)	PTNM: Gross and microscopic evaluation of blister formation. In vitro: immunofluorescence microscopy and KDA
Mao et al.167	2017	Glucocorticoid receptor, STAT3	Hydrocortisone (In vitro: 100 µg/mL, PTNM: topical cream with 5% hydrocortisone), STAT3 Inhibitor XVIII (BP-1-102, In vitro: 25 nM, PTNM: topical cream with 12.5 nM rapamycin), constitutively activated Stat3 (Stat3C) overexpression	In vitro: Pretreatment. In vivo: Topical application for 3 days before antibody treatment.	In vitro, PTNM	Hydrocortisone and STAT3C: PV IgG. STAT3 Inhibitor XVIII: PV IgG, PX43	Hydrocortisone and STAT3 Inhibitor XVIII (+), Stat3C (-)	In vitro: KDA. PTNM: Gross and histological evaluation of blister formation
Sigmund et al.60	2023	Phosphodiesterase 4 (cAMP)	Apremilast (In vitro: 1–100 µM, ex vivo: 50 µL of 1 µM, PTNM: 2 × 30 mg/day)	In vitro and ex vivo: Pretreatment for 1 h, PTNM: Pretreatment for 2 h	In vitro, ex vivo, PTNM	PV IgG, AK23	+	In vitro: KDA and immunofluorescence imaging. Ex vivo: Blister length in H&E histology. PTNM: Blister length was quantified in H&E histology.
Xie et al.127	2023	Glucocorticoid receptor, Glucocorticoid receptor + Calcineurin	Clobetasol propionate (1 µM), Clobetasol + FK506 (1 µM + 100 nM)	Cotreatment	In vitro	PV sera, AK23	+	KDA, Immunofluorescence microscopy, and western blotting was performed for Dsg3 depletion
TNF-α and inflammatory mediator signaling
Feliciani et al.140	1999	IL1α, TNF-α	Anti-IL1α antibody (2 µg/mL), Anti-TNF-α antibody (2 µg/mL)	Pretreatment for 30 min and cotreatment	In vitro	PV sera	+	Acantholysis measured by cell detachment assay
Feliciani et al.141	2000	IL1α, TNF-α, IL1α + TNF-α, IL1, IL1β, IL2	Anti-IL1α antibody (2 µg/mL), anti-TNF-α antibody (2 µg/mL), anti-IL1α + anti-TNF-α antibody (2 µg/mL), Anti-IL2 antibody (2 µg/mL), ICE knockout, IL1β knockout, TNFR1R2 knockout	Pretreatment for 30 min and cotreatment	In vitro: Anti-IL1α, Anti-TNF-α, Anti-IL1α + Anti-TNF-α, Anti-IL2. PTNM: ICE knockout, IL1β knockout, TNFR1R2 knockout	PV IgG	All except anti-IL2 antibody (+), Anti-IL2 antibody (-)	In vitro: Acantholysis measured by cell detachment assay. PTNM: Gross and microscopic evaluation of acantholysis and blister formation
Toto et al.143	2000	CD28, IL10	CD28 knockout, IL10 knockout, IL10 (50 ng)	Pretreatment for 15–60 min	In vitro: CD28 knockout. PTNM: CD28 knockout, IL10 knockout, IL10 injection	PV sera	CD28 knockout (PTNM X, in vitro -), IL10 knockout (X), IL10 injection (+)	In vitro: Acantholysis measured by cell detachment assay. PTNM: Gross and microscopic evaluation of acantholysis and blister formation
Feliciani et al.122	2003	IL1α, TNF-α, IL1α + TNF-α, IL2	Anti-IL1α antibody (2 µg/mL), anti-TNF-α antibody (2 µg/mL), anti-IL1α + anti-TNF-α antibody (2 µg/mL), Anti-IL2 antibody (2 µg/mL)	Pretreatment for 30 min and cotreatment	In vitro	PV sera	All except anti-IL2 antibody (+), Anti-IL2 antibody (-)	Acantholysis measured by cell detachment assay
Orlov et al.146	2006	Fas, TNF-α, Fas + TNF-α	Fas-L (25 ng/ml), TNF-α (25 ng/ml), Fas-L + TNF-α	Cotreatment	In vitro, EpiDermFT (3D Culture)	PV IgG1/2	x	Quantitative analysis of morphologic changes such as cell volume, cell detachment, and acantholysis in monolayers and EpiDermFT
Marquina et al.145	2008	iNOS, eNOS, nNOS, NOS, NF-κB	1400 W (3 µg/g), L-N5-(1-Iminoethyl) ornitine dihydrochloride (L-NIO, 30 µg/g), S-methyl-L-thiocitruline dihdrochloride (SMTC, 50 ug/g), [N(G)-monomethyl-L-arginine (L-NMMA, 0.5 µg/g), Parthenolide (6 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	1400 W (-), L-NIO (-), SMTC ( + ), L-NMMA (+), Parthenolide (-)	Gross and microscopic evaluation of acantholysis and blister formation
Schmidt et al.149	2009	FLIP	FLIP-L (overexpression), FLIP-S (overexpression)	Pretreatment for 1 h	In vitro	PV IgG	-	KDA and Dsg3 Immunofluorescence staining to detect keratin retraction, and intercellular gaps
Schulze et al.153	2012	Rag2 knockout	Rag2	-	PTAM	AK23	-	Histologcal examination for hair follicle and suprabasal blisters of skin and palate
Espana et al.88	2013	nNOS	SMTC (10 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Gross and microscopic evaluation of acantholysis and blister formation
Lotti et al.147	2018	FasL, Soluble FasL, Membrane bound FasL	Anti-FasL Antibody (CD178, in vitro: 1 or 15 µg/mL, PTNM: 40 µg), Soluble FasL knockout, Membrane bound FasL knockout	In vitro: Pretreatment for 1 h, PTNM: 1, 2, or 3 h after PV IgG	In vitro: Anti-FasL Antibody. PTNM: Anti-FasL antibody, Soluble FasL knockout, Membrane bound FasL knockout	PV IgG	Anti-FasL Antibody (+), Soluble FasL knockout (+), Membrane bound FasL knockout (-)	In vitro: KDA. PTNM: Gross and microscopic evaluation of acantholysis and blister formation
Radeva et al.150	2019	ST18	ST18 overexpression	-	In vitro	PV IgG1-2, AK23	x	KDA
Assaf et al.142	2021	ST18, TNF-α	ST18 overexpression, Anti-TNF-α antibody (20 µg/mL)	Cotreatment	In vitro	PV sera	TNF-α (+), ST18 overexpression (-)	KDA
Assaf et al.151	2022	ST18	ST18 overexpression	-	In vitro	AK23	x	Immunofluorescence staining of membrane Dsg3
Lin et al.156	2022	CD4 + SOCS3	CD4 + T-Cell SOCS3 knockdown	-	In vitro	PV IgG	x	KDA
Lotti et al.148	2023	Soluble FasL	PC111 (Anti-solube FasL Antibody, in vitro: 0.001-10 µg/mL, ex vivo: 1, 10, 50, and 100 μg)	Treatment was applied 2 h after PV antibody treatment	In vitro, ex vivo	PV IgG, PX43 scFv	+	In vitro: KDA. Ex vivo: Microscopic evaluation of acantholysis and blister formation
Liang et al.144	2023	IL-37, IL-37 + STAT3	Recombinent human IL-37 (100 ng/mL), Colivelin + IL-37 (0.5 µM + 100 ng/mL)	-	In vitro wildtype and Caveolin-1 shRNA cells	Rabbit polyclonal anti-Dsg3 antibody	In vitro (+), in vitro Caveolin-1 shRNA cells (-),Colivelin + IL-37 (-)	KDA
Luan et al.152	2023	Inflammation	Thalidomide (10–50 mg/mL)	Pretreatment for 1 h	PTNM	PV IgG	+	Gross and microscopic evaluation of acantholysis and blister formation
Fc Receptors
Schiltz et al.171	1980	Serum, Fc portion of antibody	IgG depletion of conditioned medium by affinity chromatography, PV IgG digestion by paparin	-	Ex vivo	From PV IgG	-	H&E staining light microscopy examination of blister formation
Hunziker et al.206	1985	Fc receptor	IVIG (Sandoglobin, 0.3–30 mg/mL)	Pretreatment for 6 and 24 h or 1 h with PV sera	Ex vivo	PV sera	+	dIF and H&E staining light microscopy examination of blister formation and acantholysis
Kawana et al.172	1985	Fc receptor	PV F(ab´)2 fragments	-	In vitro	From PV IgG	+	In vitro epidermal cell detachment visualization using phase microscopy and coulter counter
Anhalt et al.173	1986	Fc receptor	PV F(ab´)2, Fab´ fragments	-	PTNM	From PV IgG	PV F(ab´)2 (-), Fab´ ( + )	Clinical extent scoring, dIF and H&E staining light microscopy evaluation of blister formation and acantholysis
Mascaro Jr et al.174	1998	Fc receptor	PV F(ab´)2, Fab´ fragments	-	PTNM	From PV IgG	-	dIF and H&E staining light microscopy examination of blister formation and acantholysis
Arredondo et al.125	2005	Fc receptor	IVIG	Pretreatment for 1 h	In vitro	PV IgG-1a/PV IgG-2b	+	Morphometric analysis of acantholysis in
Mimouni et al.176.	2005	Fc Receptor	IVIG (5 mg/mouse), normal F(ab´)2 (6 mg/mouse), normal Fc portion (7 mg/mouse), recombinant Dsg3 (2 mg, premixed with PV IgG)	Cotreatment	PTNM	PV IgG	+	Gross and microscopic evaluation of acantholysis and blister formation
Cirillo et al.175	2007	Serum	PV IgG-free serum (0.5 mg)	-	In vitro	From PV sera	-	KDA and immunofluorescence imaging
de Bruin et al.58	2007	Fc Receptor	PV Fab	-	In vitro	From PV IgG	-	Keratin retraction was measured by immunofluorescence microscopy
Mimouni et al.177	2010	Fc Receptor	PV specific anti-idiotypic antibody (30 µg/mouse), IVIG (30 and 2000 µg/mouse)	-	PTNM	Anti-Dsg 1 and 3 scFv	+	Gross and microscopic evaluation of acantholysis and blister formation
Zakrzewicz et al.207	2022	Neonatal Fc receptor (FcRn)	Efgartigimode (25 µg/mL)	Pretreatment for 30 min or treatment after 30 min of antibody	In vitro wildtype Fc-IHH knockout cells	hAK23, mAK23, 4B3, PV IgG	Wildtype cells (+), Fc-IHH knockout cells (-), mAK23 (-)	KDA and Dsg3 depletion by western blotting
Others
Lanza et al.208	2008	Cyclin-dependent kinase 2	Cdk2 siRNA, Roscovitine (100 µg/g)	Pretreatment for 2 h	In vitro: Cdk2 siRNA. PTNM: Roscovitine	PV sera	+	Morphometric analysis of cell-cell detachment
Pretel et al.86	2009	mTOR	Rapamycin (5 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Clinical activity and microscopic evaluation of acantholysis
Lanza et al.209	2011	Perk	Perk siRNA	-	In vitro	PV sera and immunoglobulin-free sera	+	KDA
Gil et al.87	2012	mTOR	Rapamycin (5 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Gross and microscopic evaluation of tissue
Espana et al.88	2013	mTOR	Rapamycin (5 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Gross and microscopic evaluation of tissue
Mao et al.71	2014	MK2	MK2 inhibitor III (In vitro: 2.5 µg/mL, PTNM: 6.25 µg), MK2 shRNA, MK2 knockout	Pretreatment for 2 hours	In vitro: MK2 inhibitor III, MK2 shRNA. PTNM: MK2 inhibitor III, MK2 knockout	PX43	+	In vitro: Dsg3 internalization was measured by immunofluorescence microscopy and western blotting. PTNM: Gross and microscopic evaluation of blister formation.
Walter et al.75	2017	MEK 1/2	U0126 (NA)	Pretreatment for 1 h	In vitro	AK23, mcPV IgG, mPV IgG, atPV IgG, PF IgG	mcPV IgG, PF IgG (+), AK23, mPV IgG, atPV IgG (-)	KDA
Egu et al.205	2019	MEK 1/2	U0126 (50 µL of 5 µM)	Pretreatment for 1 h	Ex vivo	PV IgG	+	Blister score and cleft length was calculated based on H&E stained sections. Evaluation of desmosomes were performed by electron microscopy
Radeva et al.150	2019	MEK 1/2	U0126 (5 µM)	Pretreatment for 1 hour	In vitro ST18 overexpression	PV IgG, AK23	+	KDA
Burmester et al.77	2020	PDK1, PLK1, MEK 1	BX-795, Rigosertib, Selumetinib (0.1, 1, and 10 µM)	In vitro: Pretreatment for 2 h then discarding supernatant. Ex vivo: Cotreatment	In vitro, ex vivo	In vitro: Immunophresis material from PV patient. ex vivo: PX43 scFv	BX-795, Selumetinib (+), Rigosertib (in vitro: +, ex vivo: -)	In vitro: KDA. Ex vivo: Gross and light microscopic evaluation of blister formation.
Ivars et al.91	2020	ADAM10	GI254023X (150 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Gross and microscopic evaluation of blister formation with histological scoring
Kugelmann et al.210	2022	ADAM10, ADAM17	GI254023X (in vitro: 20 µM, ex vivo: 50 µL of 20 µM), Tapi-1 (10 µM)	Cotreatment	In vitro, ex vivo	mPV IgG 2/3, mcPV IgG1/4, AK23	In vitro for GI254023X: mcPV IgG1, mPV IgG 2/3, and Ak23 (+), mcPV IgG4 (-). Ex vivo for GI254023X: was tested with mcPV IgG4 only (-). Tapi-1 (-)	In vitro: KDA. Ex vivo: Blister scoring based on H&E staining was evaluated
Hiermaier et al.129	2022	MEK 1/2	U0126 (10 µM)	Pretreatment for 1 h	In vitro	PV IgG, PF IgG	+	KDA
Hariton et al.16	2023	Smoothened receptor	SAG (5 µg/g daily for 2 days)	Injection after 4 days of AK23 treatment	PTAM	AK23	x	Percentage of hair follicle blistering was evaluated by H&E staining
Valentino et al.211	2023	miR-148a	miR-148a mimic	-	In vitro	mPV IgG	+	KDA

Treatment modalities: Pretreatment, Intervention added before antibody application. Cotreatment, Intervention added simultaneously with the antibody. Posttreatment, Intervention added after antibody application. Model: Ex vivo, human skin organ culture. In vitro, 2D cultured keratinocyte cell lines or primary cells. PTNM, Passive transfer neonatal mouse model. PTAM, Passive transfer adult mouse model. Antibodies: mPV IgG, mucosal PV IgG. mcPV IgG, mucocutaneous PV IgG. atPV IgG, Atypical PV IgG. PF IgG, pemphigus foliaceus IgG. Preventing Pathology: ( + ) effective in preventing antibody-induced pathology, (-) Ineffective in preventing antibody-induced pathology, (x) worsened antibody-induced pathology. KDA Keratinocyte dissociation assay, NA Not Available.

Tyrosine kinase growth factor receptors and signaling

The EGFR growth factor receptor family stands out as one of the most extensively researched biochemical signaling pathways in all models of PV, compatible with a crosstalk between these receptors and Dsg3 and Dsg1. Frušić-Zlotkin et al., were the first to demonstrate in HaCat cells that PV IgG induces EGFR phosphorylation within 30 min after PV IgG exposure and inhibiting it can halt PV IgG-induced acantholysis⁴³. With the exception of the study by Heupel et al., various research works involving different antibodies and in vitro and in vivo PV disease models have consistently shown the protective impact of EGFR inhibition in PV pathogenesis^{16,46,69,85–88,91,92,137}. Notably, Walter et al. found that only PV IgGs containing anti-Dsg3 antibodies and not Dsg3/Dsg1 could induce EGFR phosphorylation, and EGFR inhibitors were ineffective in preventing PF IgG-induced fragmentation⁴⁶. Furthermore, Ivars et al. reported that EGFR inhibition failed to prevent epidermal blistering in a neonatal mouse model with PV IgG containing anti-desmocollin (Dsc) 3 antibodies, while, conversely, Hudemann et al. have shown that anti-Dsc3 IgG induces the activation of EGFR and p38MAPK, but not Src, in HaCaT cells^91,138. Significantly, Sayar et al. underscored the importance of maintaining basal EGFR activity when inhibiting PV blistering in mouse epidermis¹³⁷. They highlighted that the EGFR inhibition needs to be within a specific range to preserve its basal epidermal activity, to prevent blister formation. Furthermore, complete EGFR inhibition through pharmaceutical agents or gene knockout could be harmful to keratinocyte survival and exacerbate blister formation triggered by autoantibodies in skin but not mucosa, aligning with the dermatological adverse events of EGFR inhibitors. The most recent study by Hariton et al. unveiled a dualistic nature of EGFR activity. EGFR inhibition before antibody injection and blistering in adult mice was found to reduce induction of new hair follicle blisters, whereas EGFR inhibitor treatment in presence of blisters aggravated lesions by hindering regeneration and healing¹⁶. The latter study discriminated for the first time between a blister-inducing and blister-healing response phase marked by initial activation followed by inhibition of EGFR, and conversely, blister-inducing Wnt inhibition followed by blister-healing Wnt activation. While the two phases may explain the discrepant findings concerning p38MAK and actin cytoskeleton/RhoA inhibition and activation, they need to be taken into consideration for therapeutic applications.

Inhibition of other growth factor receptors, such as VEGFR2 and TrkA, have also been demonstrated to prevent the loss of intercellular adhesion induced by PV IgG, PX43 or AK23 in both in vitro and ex vivo models^77,85. Egu et al. further revealed that TrkA operates downstream of EGFR, where PV IgG led to increased expression and phosphorylation of TrkA at Y490⁸⁵. On the other hand, Hunziker et al. found that simultaneous injection of PV serum with stimulated or sonicated platelets, which are enriched with PDGF, into an ex vivo model, enhanced blister formation¹³⁹.

Other downstream signaling effectors

Important effectors downstream of EGFR and receptors such as of the TNF superfamily are PI3K/ERK/Akt. They have only been investigated to a limited extent. Burmester et al. revealed that PI3Kα but not PI3Kβ prevents PV pathologies in both in vitro and ex vivo models⁷⁷. Additional signaling mediators related to receptor tyrosine kinase growth factors are listed in Table 3.

The TNF superfamily and inflammatory mediator signaling

Several studies have evaluated the role of inflammatory cytokines in PV pathogenesis, including protective or disease-promoting effects. Neutralizing or knocking out TNF-α, IL-1α, IL-1β and IL6 but not IL2, has been demonstrated to alleviate the loss of intercellular adhesion, suggesting the implication of inflammatory mediators in blister formation^{122,140–142}. Conversely, a protective effect counteracting blistering was attributed to IL10 and IL37^143,144. Inhibiting neuronal nitric oxide synthase (nNOS) upregulation and NF-κB nuclear translocation, which can be triggered downstream of TNF/FasR signaling, also prevented skin blistering in mice¹⁴⁵. In line with these findings, soluble Fas ligand was shown to be causative in various models and interventions in PV pathogenesis, and its action could be successfully blocked with the human monoclonal antibody PC111 in an ex-vivo model^146–148. Notwithstanding, the administration of FLIP_L and FLIP_S, which inhibit NF-κB downstream of FasR, did not ameliorate antibody-induced acantholysis, keratin retraction, and fragmentation¹⁴⁹. Other mediators of inflammation were also shown to be involved in PV pathogenesis, notably p38MAPK, discussed in the section on biomechanical signaling^88,145. TNF-α signaling via pro-inflammatory cytokines and p38MAPK was further highlighted by a constitutive mutation (“PV risk variant” rs17315309) in the transcription factor Suppression of tumorigenicity 18 (ST18) gene or its overexpression in human keratinocytes in vitro, which trigger a self-amplifying cycle for Dsg3 downregulation in presence of PV IgG^142,150,151. Finally, Luan et al. revealed that PV IgG increased the intracellular proinflammatory adapter protein MyD88 along with other PV phenotypic changes such as NF-κB and blistering, which were ameliorated in HaCaT cells and mice by thalidomide, an immunomodulatory and anti-inflammatory glutamic acid derivative¹⁵². While inflammatory responses by keratinocytes have been demonstrated to be pathogenic in PV, as outlined above, a role of innate or adaptive immune cells in blister formation sensu stricto could be excluded by passive transfer of PV IgG into mice. Several studies utilizing immunocompromised mice exhibited a similar blistering phenotype than immunocompetent mice^153–155. However, Toto et al. demonstrated a higher incidence of PV blisters with low dose PV IgG in CD28 KO mice compared to wild-type mice although in vitro evaluation showed similar acantholysis between KO and wild-type cells¹⁴³. Similarly, co-incubation of keratinocytes with SOCS3 knocked-down CD4 + T cells, exhibiting a shift towards the Th1/Th17 phenotype, also aggravated PV IgG induced fragmentation¹⁵⁶.

Caspases

Involvement of caspases in PV acantholysis has been reported in 10 independent studies^{86–88,125,149,157–161}. Except for Schmidt et al. all studies found caspase 1/2/3 inhibition but not inhibition of caspase 8/9/12 to prevent PV pathology¹⁴⁹. The role of caspase activation in PV pathogenesis remains a critical and unresolved issue, particularly regarding its contribution to apoptosis-dependent versus newly recognized, non-apoptotic processes¹⁶². Some evidence indicates that caspases are activated leading to apoptosis in PV^{43,86,125,157,160}. However, recent studies suggest that while apoptosis may occur as a fail-safe mechanism, also in PV, the initial pathogenic event leading to acantholysis is the activation of caspases, rather than apoptosis itself^158,159. This is supported by the current understanding that both proteolytic and non-proteolytic activities of caspases are involved in cytoskeletal rearrangements, cell cycle regulation, cytoplasmic reduction, cell migration, and fate determination, thereby providing a rational for their roles beyond apoptosis¹⁶².

Canonical and non-canonical Wnt signaling

The Wnt pathway was implicated in PV pathogenesis in various models from three different groups. JUP, a crucial mediator of cell-cell adhesion and Wnt signaling, is implicated in both cell adhesion and fate determination and was already addressed in the biomechanical signaling section. Studies have shown that Wnt activation utilizing various GSK3 inhibitors prevents blister formation in neonatal and adult mouse models^16,25. Williamson et al. further confirmed the involvement of Myc in the pathogenesis of PV²⁵, a pro-proliferative transcription factor in the Wnt pathway. Additionally, Hariton et al. recently revealed that Wnt activation using BIO not only attenuated hair follicle blister formation but also facilitated blister regeneration and healing in mice. This finding may have significant implications for clinical applications. In line with these findings, Casein kinase-1, an enzyme that serves dual roles in the Wnt pathway, has also been shown to associate with keratin intermediate filaments via FAM83H¹⁶³. Schögel et al. demonstrated that inhibiting this molecule exacerbates AK23-induced fragmentation in vitro⁸¹. Together, these studies highlight that canonical Wnt and possibly also non-canonical Wnt (upstream of PLC/PKC) inhibition play a critical role in regulating cell adhesion as well as in influencing fate conversion and regeneration.

Glucocorticoid and other G protein-coupled receptors

Corticosteroids, which primarily exert their effects through Corticosteroid G-protein-coupled receptors, are the cornerstone of therapeutic treatment of PV patients and have also been explored in preclinical studies. Among six research studies, at keratinocyte level, only half were able to demonstrate the efficacy of corticosteroids on preventing loss of adhesion in the context of PV pathogenesis^{127,154,164–167}. Nguyen et al. found that pretreatment, and not co-treatment, with methylprednisolone reduced blistering in both neonatal Balb/c and athymic nude mice¹⁵⁴, suggesting a therapeutic effect reducing initiation of new blisters. Similarly, Mao et al. reported comparable results in an in vitro model, as well as with the topical application of hydrocortisone in a neonatal C57BL/6 J mouse model which likely acts through the prevention of STAT3 activation¹⁶⁷.

Adrenergic and acetylcholine receptors are another family of G protein-coupled receptors that have been studied in PV pathogenesis and might, in rare cases, even be directly targeted by antibodies in PV IgG other than anti-Dsg3/Dsg1. Initially, Grando et al. demonstrated that acetylcholine agonists can prevent PV IgG-induced acantholysis, provided that exposure to PV IgG is limited to less than 8 hours¹⁶⁸. These experiments were further validated using a mouse model and with additional acetylcholine receptor agonists¹⁵⁵. In a recent study, Chernyavsky et al. found that the effects of PV IgG on keratinocytes involve the phosphorylation of both p120- and β-catenin, with PKC mediating the serine phosphorylation of β-catenin and Src mediating the tyrosine phosphorylation of p120-catenin¹⁶⁹. Furthermore, downstream signaling of keratinocyte M1 muscarinic receptors interfered with the PV IgG-dependent catenin phosphorylation through the activation of both protein phosphatase 2 A (PP2A) and protein-tyrosine phosphatase (PTP). In contrast, the mechanism of action of α7 nicotinic receptors involved both the activation of PTP and the inhibition of Src. Each inhibitor only partially reversed the effect of PV IgG. Apart from cholinergic receptors, Spindler et al. investigated the role of adrenergic receptors and confirmed previous studies, that the activation of β-Adrenergic receptors prevents loss of intercellular adhesion in vitro and blister formation in an vivo model in response to PV antibodies¹⁷⁰. Also, β-adrenergic receptor agonists facilitated the recovery of cell-cell adhesion following the removal of autoantibodies¹⁷⁰. Increasing cAMP levels using phosphodiesterase inhibitors protected from the loss of intercellular adhesion through blockage of p38MAPK, while, conversely, the inhibition of protein kinase A (PKA) exacerbated the effects of the antibody^60,170. The status of phosphodiesterase activity in PV is currently unknown, however the increase in cAMP monitored after PV IgG stimulation was suggested to pertain to the repair/regenerative phase^60,170, revealed and amply described by Hariton et al.¹⁶.

Fc Receptors

In 1979, a seminal study by Schiltz et al. demonstrated that Fab fragments of PV IgG, obtained by enzymatic digestion using papain (Fab), were as effective than entire IgG, suggesting that antibody-mediated crosslinking of target receptors is not required for PV pathogenicity. Moreover, IgG depleted fraction of patient serum failed to cause acantholysis¹⁷¹. Later studies yielded inconsistent results regarding the efficacy of Fab or F(ab´)₂ in acantholysis. While Kawana et al. reported that PV F(ab´)₂ was ineffective in an in vitro model, Anhalt and Till found that all mice receiving PV F(ab´)₂ developed cutaneous blisters and erosions, whereas none of the mice receiving PV Fab fragments developed lesions^172,173. Conversely, Mascaro Jr et al. revealed that both F(ab´)₂ and Fab fragments induced acantholysis in mice¹⁷⁴. Consistently, de Bruin et al. found that the Fab portion was as effective as PV IgG in causing loss of intercellular adhesion in cultured keratinocytes⁵⁸. The role of serum-born factors in inducing acantholysis was further evaluated by Cirilo et al. who showed, on interest for intravenous IgG, that PV serum lacking IgG can partially reduce the intercellular adhesion, but this effect was not statistically significant¹⁷⁵. Mimouni et al. conducted two studies revealing that the beneficial effects of normal IgG are not solely dependent on the Fc portion but also involve the F(ab´)₂ fragments^176,177.

Fc receptor activation can also induce Src activation, which may be mistakenly interpreted as Dsg signaling. However, studies have demonstrated that Src inhibition using mouse monoclonal antibodies with low affinity for human Fc receptors is effective^90,178. Therefore, Src activation cannot be attributed to Fc receptor signaling.

Other factors

Other factors such as protease inhibitors, complement system, mitochondrial potential, among others are briefly outlined in Table 4.

Table 4.

Characteristics of other studies

Author	Year	Target molecule	Intervention	Treatment Modality	Model	Antibody	Preventing Pathology	Measured Outcome
Michel et al.212	1977	Complement (C3)	Complement inactivation by heating the serum at 56 °C for 30 min or dialysis	-	Ex vivo	PV sera	-	dIF and H&E staining light microscopy examination of blister formation and acantholysis
Farb et al.213	1978	Serine proteases, Broad spectrum proteinase	Soybean trypsin inhibitor (100 µg/mL), α2-macroglobulin (0.5 µg/mL)	Pretreatment for 24 h	In vitro	PV sera (Heat inactivated)	+	Cell detachment assay. Binding of antibodies was measured by dIF. Cellular viability was measured by incorporation of radioactive amino acid into TCA precipitable protein
Schiltz et al.164	1979	Serine proteases, unknown	Soybean trypsin inhibitor (100 µg/mL), Thioglucose salt of gold (10 µM)	Pretreatment	Ex vivo	PV IgG	-	dIF and H&E staining light microscopy examination of acantholysis
Schiltz et al.171	1980	C3, Protein synthesis	Complement inactivation by heating the serum at 56 °C for 1 h, Puromycin (0.1–100 µM)	Pretreatment	Ex vivo	PV IgG	C3 (-), Puromycin (+)	H&E staining light microscopy examination of acantholysis
Singer et al.214	1980	Serine proteases, Broad spectrum proteinase	Soybean trypsin inhibitor (100 µg/mL), α2-macroglobulin (0.5 mg/mL)	Pretreatment	In vitro: Soybean trypsin inhibitor, alpha2-macroglobulin. PTNM: Soybean trypsin inhibitor	PV sera	+	Cell detachment assay
Morioka et al.215	1981	Serine proteases, Aspartyl proteases, Cysteine peptidase	Soybean trypsin inhibitor (75–5000 µg/mL), Pepstatin A (2–40 µg/mL), N-ethylmaleimide (NEM, 0.1–1 mM)	Pretreatment	Ex vivo	PV IgG	+	dIF and H&E staining light microscopy examination of blister formation and acantholysis
Hashimoto et al.216	1983	Plasmin, Trypsin, Serine proteases	Plasminogen (200 µg/mL), Lima bean trypsin inhibitor (NA), Aprotinin (60 TIU)	Cotreatment	Ex vivo	PV IgG, PV IgG + Plasminogen	+	H&E staining light microscopy examination of blister formation and acantholysis
Woo et al.217	1983	Serine proteases, Aspartyl proteases, Trypsin, Protease	Soybean trypsin inhibitor (100 µg/mL), Aprotinin (800 units/mL), Pepstatin (50 µg/mL), Lima bean trypsin inhibitor (400 µg/mL), α2-macroglobulin (540 µg/mL), Chymostatin (600 µg/mL), Leupeptin (500 µg/mL), Antipain (100 µg/mL)	Cotreatment	In vitro	PV IgG	Soybean trypsin inhibitor, Aprotinin, α2-macroglobulin, Pepstatin, Lima bean trypsin inhibitor (+), Chymostatin, Leupeptin, Antipain (-)	Cell detachment assay
Kawana et al.172	1985	Complement, C1q, Plasmin	Complement addition (10% normal human serum), Complement depletion, C1q depleted fraction, complement inactivation by heating the serum at 56 °C for 30 min, Plasminogen depleted serum	-	In vitro	PV IgG	Complement addition, Plasminogen depleted serum (-), Complement depletion, C1q depleted fraction, complement inactivation by heating the serum at 56 °C for 30 min (+)	Cell detachment assay
Anhalt et al.173	1986	C5, C3	C5 knockout, Cobra Venom Factor (CoVF) (NA)	Pretreatment	PTNM	PV IgG	C5 knockout (-), CoVF (+)	The clinical extent scoring and H&E staining light microscopy evaluation of blister formation and acantholysis
Naito et al.166	1989	Proteases, Serine Proteases	FOY (Gabexate mesylate, ex vivo: 1–10 mg/mL, PTNM: 133 µg/g), FOY-305 (Camostat mesylate, Ex vivo: 0.5–5 mg/mL, PTNM: 100 µg/g), FUT-175 (Nafamostat Mesilate, ex vivo: 0.1–5 mg/mL, PTNM: 13 µg/g), α1-proteinase inhibitor (Ex vivo: 2.5, 25 mg/mL, PTNM: 1.67 µg/g)	Cotreatment	Ex vivo, PTNM	PV IgG	Ex vivo: all (+). PTNM: FOY-305 and α1-proteinase inhibitor (+), FUT-175 and FOY (-)	Ex vivo: H&E staining and microscopic evaluation for acantholysis. PTNM: Visual examination for Nikolsky sign as well as histological examination of acantholysis
Kawana et al.218	1990	Ultraviolet light	UVB (0.06 to 0.18 J/cm2)	-	In vivo PV patient skin	In vivo PV patient skin	-	H&E staining light microscopy and immunofluorescence examination
Dobrev et al.219	1996	Proteases	p-aminomethylbenzoic acid (PAMBA, 1 mg/mL), Aprotinin (10 ATrE/ml)	Cotreatment	Ex vivo	PV sera	+	H&E staining light microscopy examination of blister formation and acantholysis
Feliciani et al.140	1999	C3	Anti-C3 antibody (10 µg/mL)	Pretreatment for 30 min and Cotreatment	In vitro	PV sera	+	Acantholysis was measured at cell detachment assay
Asano et al.121	2001	Proteases	PAMBA (10, 50 µg/mL), Aprotinin (0.1, 1 TIU)	Pretreatment for 30 min	In vitro	PV IgG	-	Immunofluorescence examination of keratin retraction and desmoplakin
Sánchez-Carpintero et al.99	2004	Tyrosine kinases	Genistein (500 µg/g), Herbimycin (15 µg/g)	Pretreatment for 3 h	PTNM	PV IgG	+	Gross and microscopic evaluation of acantholysis and blister formation
Feliciani et al.220	2007	Unknown	Tannic acid (0.1 nM)	Cotreatment	In vitro	PV IgG	x	Acantholysis was measured by cell detachment assay
Cirillo et al.221	2008	Kinases	Staurosporine (100 nM)	Pretreatment for 1 h	In vitro	PV sera	+	Acantholysis measured by immunofluorescence microscopy
Delva et al.113	2008	Tyrosine kinases	Genistein (40 µM)	Pretreatment for 1 h	In vitro	PV IgG	+	KDA, immunofluorescence imaging, and Dsg3 internalization assay
Marquina et al.145	2008	Tyrosine kinases	Genistein (500 µg/g)	Pretreatment for 2 h	PTNM	PV IgG	+	Clinical activitiy and microscopic evaluation of acantholysis
Saito et al.68	2012	Tyrosine kinases	Genistein (Ex vivo: 6.25 µg, in vitro: 50 µM)	Ex vivo: Pretreatment for 2 h. In vitro: Pretreatment for 1 h	Ex vivo, in vitro	PV IgG, AK23	+	In vitro: KDA. Ex vivo: Gross and microscopic evaluation of blister formation.
Kalantari-Dehaghi et al.126	2013	Mitochondria, NAD	Minocycline (0.05 mM), Nicotinamide (10 mM)	Cotreatment	PTNM	PV IgG	Minocycline (+), Nicotinamide (-)	Morphometric analysis of acantholysis
Liang et al.222	2017	Unknown	Naringenin (50, 100, 200 µM)	24 h after PV sera treatment	In vitro	PV sera	+	KDA
Wei et al.223	2021	Thioredoxin-2 (Trx2)	Trx2 Overexpression	-	PTNM	PV sera	+	Gross and microscopic evaluation of acantholysis and blister formation

Treatment modalities: Pretreatment, Intervention added before antibody application. Cotreatment, Intervention added simultaneously with the antibody. Model: Ex vivo, human skin organ culture. In vitro, 2D cultured keratinocyte cell lines or primary cells. PTNM, Passive transfer neonatal mouse model. Antibodies: PV IgG or sera, pemphigus vulgaris IgG/sera. PF IgG, pemphigus foliaceus IgG. Preventing Pathology: ( + ) effective in preventing antibody-induced pathology, (-) Ineffective in preventing antibody-induced pathology, (x) worsened antibody-induced pathology. KDA Keratinocyte dissociation assay, NA Not Available.

Discussion

This systematic review allows for the first time to present a comprehensive network of 128 functionally validated effector signals in PV published since inception in 1977 up to mid-2024. These effectors can be categorized according to three general signaling axis typifying adhesion receptor signaling^5,6: biomechanical (Fig. 1), bioelectrical (Fig. 2), and biochemical (Fig. 3). While the understanding of biomechanical signaling is only starting to emerge and bioelectrical effectors, in particular ion channels, have drawn little attention so far, the best studied effectors in loss of intercellular adhesion relate to biochemical signal transduction upon loss of Dsg transadhesion.

It is noteworthy that modulation of the great majority of the 128 effector molecules in PV have been revealed through cell biological approaches followed by functional validation through pharmacological activation, deactivation or gene deletion prior to antibody application, thus mainly addressing effectors responsible for blister initiation. Among these, we defined as key effectors in PV (indicated in bold in Figs. 1–3), those molecules which have been consistently identified and validated by multiple sources using various models, antibodies and manipulations. In all signaling axis combined (Fig. 4), these are, remarkably, the molecules which were found to be associated with extra-desmosomal Dsg3 or Dsg1; JUP, p38MAPK, PLCγ/PKC, EGFR/ErB2 and GSK-3β/JUP (Wnt signaling). The implications of these effectors are consistent with stress-induced cytoskeleton reorganization, junctional release, increased migration and proliferation. Activation of growth factor and tyrosine kinase receptors, including EGFR/ErbB2, govern three major processes: mitotic activation, cytoskeletal rearrangement and migration, encompassing endocytosis of junctional components and release of adhesive strength. The pro-migratory and pro-mitotic phenotype in PV is consistently seen throughout the biomechanical signaling axis and is supported by Wnt inhibition, closing the gateway towards keratinocyte differentiation through nuclear signaling and stabilization of junctions at the plasma membrane^25,179. Several other studies have demonstrated the influx of intracellular calcium and activation of tyrosine kinase receptor-dependent signaling pathways further supporting enhanced migration. Moreover, the suggested inhibition of small GTPases, such as RhoA, Rac-1 and Cdc42 in PV³⁵ or the early redistribution of activated RhoA away from the plasma membrane³⁶, and the reported actin depolymerization⁴⁰, would also support migration. However, to define the rebalance of forces through the Rho GTPase family of highly dynamic, complex and compartmentalized molecular switches, more sophisticated technologies like live cell imaging or experimental manipulations such as FRET¹⁸⁰, as initiated by Jin et al.¹⁸¹, will be required. Consistent with calcium influx and a migration/mitosis phenotype, calcium release-activated channels (CRAC), and calmodulins were identified as effectors in the bioelectrical axis, which were however mainly identified to be activated in presence of low Dsg1 protein expression³¹ or complementary anti-Dsg1 antibodies. The increase in intracellular calcium concentration^8,46,61,72 also favors migration in addition to the release of inflammatory cytokines^125–127. Additional key effectors defining the biochemical axis are soluble FasL and Caspases (TNF superfamily). Furthermore, the TNF-α superfamily regulates survival and apoptotic genes under stress conditions. Taken together, this signaling network is highly consistent and supports actin rearrangement⁴⁰, and contractility⁵⁰ as well as increased proliferation^{16,25,153,182,183} and migration while reducing differentiation^17,25. Other studies outside the scope of PV have also confirmed the role of Dsg3, and in some cases also Dsg1, in cell migration^184–187, proliferation^188–190, and the regulation of the actin cytoskeleton. Consistently, extra-desmosomal Dsg3 has been shown to be associated with actin fibers and ezrin, functioning as an upstream regulator of Rho GTPases and Src^22,23,186. Finally, morphological regeneration and restoration of desmosomal junctions were reported to occur via downregulation of EGFR, activation of Wnt signaling, the upregulation of RhoA/Rac-1 and the increase of adherens junction proteins E-cadherin, β-catenin and JUP. The latter presumably increases the tension load to offset desmosomal instability, consistent with α-adducin phosphorylation and p38MAPK inhibition^16,72,181.

Collectively, the binding of autoantibodies to desmosomal antigens supports an astonishingly coherent model whereby outside-in signaling mediated by the major PV target receptors Dsg3/Dsg1 leads within seconds to minutes through a rapid influx of calcium to activation of PLC and PKC, p38MAPK, and EGFR as well as Wnt inhibition (Fig. 4). These cascades are known to lead to actin remodeling, enhanced Dsg3/JUP endocytosis and desmosomal shrinkage as well as keratin retraction, contributing to the destabilization of desmosomes and weakening of cell-cell adhesion presumably along a process of inside-out signaling such as p38MAPK and PKC activation. Also, Rho GTPases, particularly RhoA, and their inactivation in PV pathogenesis may further facilitate weakening of desmosomal adhesion, while reactive oxygen species (ROS) and apoptotic pathways survey the signaling events associated with acantholysis.

While a coherent picture is provided for blister initiating effectors reported in PV, the longitudinal analyses, and in particular the study of effectors of protective or regenerative processes during and after blister formation will be essential to be considered when designing therapeutic approaches to avoid adverse effects of effector modulation on already existing blisters, as reported by Hariton et al.¹⁶. Furthermore, studying only one time point can introduce a bias, as signaling molecules are subject to dynamic changes, which has been shown in case of EGFR, Wnt and RhoA^{16,137,181,191–193}. This is also compatible with the biological understanding that the cell populations in the basal layer of stratified epithelia reflected in keratinocyte 2D cultures are heterogeneous (stem cells, transit amplifying, differentiating progenitor cells)⁷. Therefore, and once more, measuring a change in signaling molecules at a single time point may lead to conclusions that cannot be validated in clinical translation. Lastly, in order to understand Dsg3 and Dsg1 receptor signaling in PV and particularly in epithelial homeostasis in detail, it is imperative to investigate the extra-desmosomal (without desmoplakin) and desmosomal (with desmoplakin) fraction separately through biochemical fractionation or imaging.

Concerning inhibitor/activator screening studies, it must be mentioned that validation of the molecular target in the biological setting with and without the pharmacological drug is mandatory, as some molecules may not be directly involved or causative in PV pathogenesis but may instead strengthen the cytoskeleton network, desmosomes or other adhesion molecules as a compensatory mechanism, thereby seemingly preventing or attenuating blister formation^60,170. Moreover, when probing Dsg3 signaling, the status of Dsg1 expression needs to be monitored, as high Dsg1 levels may compromise signaling results, or compensate for loss of Dsg3 cell-cell contacts. Further discriminating between Dsg3 and Dsg1 receptor signaling will help to better address this gap. As another caveat, some signaling molecules respond within seconds to minutes after antibody binding, inferring that the application of inhibitors or activators occurs too late and may not effectively interrupt the signaling cascade, yielding misleading results¹⁹⁴. Another issue identified is that in many studies chemical concentrations were not titrated and only one single dose was used. This might not be effective or come with secondary, PV-unrelated effects, and was for example shown to be particularly important for fundamental biological pathways like EGFR. It was shown in one study that EGFR inhibition is ineffective in preventing the loss of intercellular adhesion, while Sayer et al. demonstrated a biphasic dose-response relationship where only doses within a hermetic zone effectively prevent blister formation^92,137. Additionally, none of the studies measured confounding contaminants, such as endotoxins, which can activate inflammatory pathways and potentially confound PV signaling^195,196.

In summary, in this systematic literature review, we identified targets that have been tested in various models using different PV antibodies (Fig. 4). The therapeutic use of these targets necessitates careful consideration of their feasibility and efficacy on both phases of blistering and regeneration. For example, while EGFR is implicated in PV pathogenesis, Hariton et al. demonstrated that lapatinib administration can hinder the healing process of blisters, and thus is not eligible for therapeutic application¹⁶. Similarly, p38MAPK plays a crucial role in numerous physiological processes, which will continue to be identified while making it a challenging drug target^62,197; several clinical trials have failed due to unacceptable toxicities^198,199. More recently developed p38MAPK inhibitors, which have shown greater tolerability in patients, were found to act primarily as EGFR inhibitors rather than direct p38MAPK inhibitors²⁰⁰. Hence, targeting downstream pathways may offer a more feasible approach with fewer side effects, although efficacy remains a concern, as some inhibitors, such as MK2, have only been shown to prevent spontaneous blistering, not Nikolsky-induced blisters^71,201. Thus, a more advanced understanding of the intertwined signaling network downstream of PV IgG binding potentially presents the possibility to design combined targets for efficient treatment.

In conclusion, the signaling network downstream of transadhesion loss by extra-desmosomal Dsg3/Dsg1 receptors and desmosomal structures, unveiled through half a century of signaling studies on PV, increases the complexity of our understanding of epithelial homeostasis and regeneration. Suggested to be strong enough to reprogram epidermal cell states^16,25, the cross-talk triggered from desmosomal cadherins to major cellular circuits and evolutionary older structures such as classical cadherins and integrins²⁰² revealed here is astonishing and will require further attention. The great challenge ahead will be to follow through from desmosomal cadherin signaling to global transcriptional reprogramming expected to represent a fine balance between a plethora of intersecting signaling cascades.

Methods

Search strategy

The present systematic review was performed according to the methodology outlined by the 2020 version of the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines. Four databases—PUBMED, EMBASE, Web of Science, and SCOPUS—were searched for eligible articles using the search terms “pemphigus” AND (“activator” OR “inhibitor” OR “Dsg3 signaling”) from inception until June 12, 2023. The syntax of search terms in each database is mentioned in Supplementary Table S1, Supplementary Information. The reference lists of all relevant articles previously identified through electronic database searches were also checked. After the completion of data extraction, another search was performed to screen for newly published investigations in this interval until April 25, 2024.

Data selection and collection

Search results were combined, and duplicates were excluded using EndNote (version 20.6, Thomson Reuters, 2020). Two reviewers independently reviewed all citations and selected eligible studies. Eligibility was determined after the inclusion criteria were met as follows: (i) studies that used PV sera, PV IgG, or monoclonal antibodies against Dsg3 or Dsg1/3; (ii) Studies that utilized one or more of the following models: human skin explants, 3D keratinocyte cultures, 2D cultured primary or cell line keratinocytes, passive antibody transfer models in neonatal or adult mice; (iii) studies that evaluated a signaling molecule through chemical or genetic manipulation (intervention) to inhibit or aggravate the PV phenotype; (iv) studies that measured at least one of the following pathological outcomes: desmosomal shrinkage, keratin retraction, cell-cell and cell-matrix dissociation, Dsg3 internalization, acantholysis, or blister formation in suprabasal layers or hair follicles.

Exclusion criteria included: (i) non-English published studies; (ii) non-original researches (e.g., editorials, reviews, book chapters, commentaries); (iii) studies in humans or in non-human animals other than mice (iv) studies conducted on active transfer PV mouse models; (v) outcome measures other than PV pathological phenotypes mentioned in inclusion criteria; (vi) studies that used non-Dsg antibodies or IgG from atypical PV cases; (vii) unavailability of the full text of articles.

Data extraction

The authors extracted data using a pilot-tested form. The study characteristics extracted from the full-text articles eligible for qualitative analysis included the following categories: (i) bibliographic details (title, authors, publication year, journal); (ii) model characteristics (type of preclinical model); (iii) intervention characteristics (type of experimental manipulation, dose, incubation duration, antibodies used); (iv) details of the examined signaling molecule (signaling molecule, signaling pathway); and (v) outcome measures (type of outcome, primary and secondary outcomes, effectiveness).

Quality assessment

All studies were included in the review; therefore, a formal quality assessment was not considered necessary.

Author contributions

All authors met the following criteria: substantial contributions to the conception and design of the study, acquisition of data, or analysis and interpretation of data; drafting the article or critically revising it for important intellectual content; final approval of the version to be submitted; and accountability for all aspects of the work, ensuring that questions related to the accuracy or integrity of any part of the work were appropriately investigated and resolved. The specific contributions of the authors are detailed as follows: Conception and design of the study: E.J.M. and S.R. Literature search and review of bibliography: S.R., F.K., and W.V.J.H. Analysis and interpretation of data: E.J.M. and S.R. Drafting the article: E.J.M. and S.R. Preparation of figures: W.V.J.H. and E.J.M. Critical revision of the article for important intellectual content: E.J.M., S.R., W.V.J.H., R.L., and L.B. Funding acquisition: E.J.M.

Data availability

All data utilized in this study are fully detailed and available within the article and its supplementary information files, ensuring transparency and facilitating further open research.

Competing interests

Ralf J. Ludwig has received honoraria for speaking or consulting or has obtained research grants from Monasterium Laboratories, Novartis, Lilly, Bayer, Dompe, Synthon, Argen-X, TriNetX, and Incyte during the last 3 years. All other authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41536-025-00426-x.