Pemphigus Vulgaris Acantholysis Ameliorated by Cholinergic Agonists

Home , Acantholysis, Pemphigoid, Pemphigus vulgaris

OBSERVATION Pemphigus Vulgaris Acantholysis Ameliorated by Cholinergic Agonists

Vu Thuong Nguyen, PhD; Juan Arredondo, PhD; Alexander I. Chernyavsky, PhD; Mark R. Pittelkow, MD; Yasuo Kitajima, MD, PhD; Sergei A. Grando, MD, PhD, DSc

Background: Pemphigus vulgaris (PV) is an autoim- phorylation level of E-cadherin and plakoglobin was in- mune, IgG autoantibody–mediated disease of skin and mu- creased by PV IgG, whereas this effect of PV IgG was at- cosa leading to progressive blistering and nonhealing ero- tenuated in the presence of 0.5mM carbachol. sions. Patients develop autoantibodies to adhesion Pyridostigmine bromide, an acetylcholinesterase inhibi- molecules mediating intercellular adhesion and to kerat- tor, produced effects similar to those of carbachol, which inocyte cholinergic receptors regulating cell adhesion. helps explain its clinical efficacy in a patient with active PV that was resistant to treatment with systemic gluco- Observations: To determine whether a cholinergic ago- corticosteroids. Treatment with pyridostigmine bro- nist can abolish PV IgG–induced acantholysis, litter mates mide (360 mg/d) in a patient with PV allowed to keep of neonatal athymic nude mice were injected with PV IgG his disease under control at a lower dose of prednisone together with carbachol (0.04 µg/g body weight). None than that used before starting pyridostigmine bromide of these mice developed skin lesions. Through in vitro treatment. experiments, we measured the expression of adhesion molecules in monolayers of normal human keratino- Conclusion: Elucidation of the cholinergic control of ke- cytes incubated overnight in the presence of 0.25mM car- ratinocyte adhesion merits further consideration be- bachol using semiquantitative Western blot and immu- cause of a potential for the development of novel anti- nofluorescence. Carbachol caused an elevation of the acantholytic therapies using cholinergic drugs. relative amount of E-cadherin in keratinocytes (PϽ.05) without changing that of plakoglobin (PϾ.05). The phos- Arch Dermatol. 2004;140:327-334

EMPHIGUS VULGARIS (PV) IS antholytic effects of glucocorticosteroids an autoimmune-mediated in vitro.7 We report that stimulation of the disease of skin and mucosa keratinocyte cholinergic receptors con- leading to progressive blis- trols PV IgG–induced acantholysis in neo- tering and nonhealing ero- natal mice and ameliorates the natural sions.P Therapy for patients with PV relies course of disease in a patient with PV. Cho- on the long-term use of systemic gluco- linergic agonists stimulated expression of corticosteroids in relatively large doses, E-cadherin and abolished phosphoryla- which, although lifesaving, may cause se- tion of E-cadherin and plakoglobin in ke- vere adverse effects, including death. Ac- ratinocytes caused by PV IgG. Thus, novel tive disease state is characterized by the antiacantholytic therapies may be devel- presence of serum IgG autoantibodies oped based on the antiacantholytic ef- binding to the keratinocyte cell mem- fects of cholinomimetic drugs. brane proteins. Although the antiacantholytic effect of glucocorticosteroids is at- METHODS tributed to immunosuppression, high doses PASSIVE TRANSFER OF PV From the Departments of of glucocorticosteroids can directly block Dermatology, University of PV IgG–induced acantholysis in vitro1,2 and Neonatal athymic nude mice were used to test California Davis, Sacramento rapidly (within 48 hours) stop blistering direct antiacantholytic effects of cholinergic ago- (Drs Nguyen, Arredondo, in patients with pemphigus without alter- nists. At the third day of life, athymic nude mice Chernyavsky, and Grando), ing the titer of autoantibodies or block- weigh approximately 1.5 g and can develop gross Mayo Clinic, Rochester, Minn skin blisters on passive transfer of PV antibod- ing antibody binding to keratinocytes (Dr Pittelkow), and Gifu 3-5 ies. This study had been approved by the Uni- University, Gifu City, Japan (“pulse therapy”). Patients develop au- versity of California Davis Review Committee (Dr Kitajima). The authors toantibodies to keratinocyte cholinergic on the Use of Animals in Research, Sacra- have no relevant financial receptors regulating cell adhesion.6 Acti- mento, Calif. We injected 52 mice intraperito- interest in this article. vation of these receptors mimics antiac- neally through a 30-gauge needle with PV (ex-

(REPRINTED) ARCH DERMATOL / VOL 140, MAR 2004 WWW.ARCHDERMATOL.COM 327

©2004 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021 periment) or normal human IgG (negative control). One subgroup PHOSPHORYLATION ASSAY of mice received PV IgG alone and the other subgroup received Quantitative phosphorylation assay was designed based on the PV IgG together with a test drug. The patients’ and control hu- 13 man IgG serum fractions were obtained using 40% ammonium established protocols, as detailed elsewhere. Briefly, cultured DJM-1 cells were grown to approximately 75% conflu- sulfate followed by dialysis against phosphate-buffered saline (PBS; 2 Gibco BRL, Gaithersburg, Md). The fractions were then lyophi- ence in 75-cm flasks (Corning Costar, Cambridge, Mass) in 8 serum-free keratinocyte growth medium (Clonetics Corp, San lized and reconstituted in PBS as detailed elsewhere. The nega- 2+ tive control IgG was isolated from normal human serum pur- Diego, Calif) containing 0.09mM Ca at 37°C in a humid 5% chased from Sigma Chemical Co, St Louis, Mo. carbon dioxide incubator. The cells were then treated for 14 hours with phosphate-free minimal essential medium contain- 2+ ASSESSMENT OF ACANTHOLYSIS ing 1.8mM Ca and 5% dialyzed fetal calf serum (Gibco BRL) and then treated with the same medium containing 200 µCi/mL Gross skin lesions were observed approximately 24 to 40 hours of 32P orthophosphate (Amersham Pharmacia Biotech, Inc). The after the first injection of PV IgG. Histological evaluation of le- cells were labeled for 8 hours and then exposed for 1 hour to 1 sional and perilesional skin in mice with clinical symptoms of mg/mL of pooled normal human IgG (Sigma Chemical Co) or induced PV revealed extensive intraepidermal split and nonspe- PV IgG pooled from 5 PV sera samples in the absence or pres- cific changes. To allow accurate analysis of the extent of acan- ence of 0.5mM carbachol or pyridostigmine bromide. After in- tholysis, we euthanized the mice at the preclinical stage of the cubation, the cells were lysed and the lysates were cleared by disease (ie, approximately 20 hours after the injection). The ani- centrifugation at 14000g for 10 minutes. The adhesion mol- mals were cross-sectioned at the umbilicus level and freshly fro- ecules under consideration were immunoprecipitated, re- zen. The need to compare skin specimens from the same body solved by 7.5% sodium dodecyl sulfate–polyacrylamide gel region stemmed from the fact that acantholysis in neonatal mice electrophoresis, electroblotted, and analyzed with the Phos- may vary between different anatomical regions.9 Three 6-µm thick phorImager feature of the Storm system (Molecular Dynam- cross sections of each mouse body were placed on the glass slide ics, Mountain View, Calif). To determine protein concentra- and stained with hematoxylin-eosin. The images (original mag- tion of each precipitated adhesion molecule, the samples were nification ϫ100) of 5 randomly selected longitudinal skin areas analyzed by Western blotting, as described above. The results in each specimen were photographed using a computer-linked were expressed as ratios of the radioactivity value of each ad- microscope and printed. We determined the extent of acan- hesion molecule to its protein quantity in each sample, com- tholysis in the photographs by measuring the length of the in- pared with the ratios obtained in control cultures (taken as 1.00). traepidermal split (at least 4 basal cells long) and expressed the results as percentage of the total length of epidermis in the field, STATISTICS taken as 100%. The results of quantitative experiments were expressed as IMMUNOHISTOCHEMISTRY mean±SD. Significance was determined using the t test. The results were deemed significant if the P value was less than .05. Direct and indirect immunofluorescence assays were performed as detailed elsewhere.8 For the direct immunofluorescence assay, a tissue specimen was incubated for 1 hour at room REPORT OF A CASE temperature with a fluorescently labeled goat antihuman IgG an- A recently reported case of PV that had improved by ciga- tibody (Pierce, Rockford, Ill; dilution 1:200 in PBS). For the in- 14 direct immunofluorescence experiments, a tissue substrate was rette smoking and a study showing successful use of 15 first treated with a primary antibody to E-cadherin or plakoglo- nicotinamide as a steroid-sparing agent in pemphigus, bin (both from BD Transduction Laboratories, Lexington, Ky; suggested that pharmacological regulation of keratino- dilution, 1:200) and then exposed to a secondary fluorescently cyte acetylcholine (ACh) axis may be a novel antiacan- labeled antibody (Pierce). Semiquantitative indirect immuno- tholytic therapy for pemphigus because (1) cigarette fluorescence assay was performed as detailed previously,10,11 us- smoke contains the cholinomimetic agent nicotine and ing a computer-assisted image analysis with a software package (2) nicotinamide exhibits cholinomimetic effects16 ow- purchased from Scanalytics (Fairfax, Va). For each tissue speci- ing to the stimulation of ACh release17 and inhibition of men, a minimum of 3 different segments in at least 3 different acetylcholinesterase (AChE).18 To determine whether a microscopic fields were analyzed and the results compared. pharmacologic stimulation of keratinocyte cholinergic re- WESTERN BLOT ceptors can be used as a steroid-sparing regimen in the treatment of pemphigus, we administered pyridostig- Immunoblotting was performed as detailed by us previously.12 Briefly, monolayers of normal human keratinocytes were mine bromide (Mestinon; ICS Pharmaceuticals, Costa either untreated (control) or incubated overnight with 0.25mM Mesa, Calif; 60-mg tablets) to a patient with active PV at carbachol or 0.25mM pyridostigmine bromide at 37°C. They the dose of 360 mg/d. The use of Mestinon in a patient were then washed and lysed, and the lysates were resolved by with PV had been approved by the University of Cali- a 7.5% sodium dodecyl sulfate–polyacrylamide gel electropho- fornia Davis Human Subjects Review Committee. This resis, followed by electroblotting. The membranes were blocked patient, an 82-year-old white man, had been treated for and stained with a primary antibody to either E-cadherin (di- almost 8 years with a mid-dose of prednisone, ranging lution, 1:2000) or plakoglobin (dilution, 1:1000) or with an- ␤ from 15 to 30 mg/d, and occasional intralesional corti- tibody to -actin (Sigma Chemical Co; dilution, 1:2000) to stan- costeroid injections. He had a recalcitrant erosion on his dardize the measurements. Binding of the primary antibody was nose, which would never completely heal and would be- visualized with horseradish peroxidase–conjugated secondary antibodies and developed with the chemoluminescence come active (ie, turn red and/or get painful, enlarge in method of the ECL+Plus system (Amersham Pharmacia Bio- size, and produce exudate). No other lesions were seen. tech, Inc, Piscataway, NJ). The results were expressed as ac- The lesion began to improve starting from the third week tual densitometry values of each protein band representing an of treatment with pyridostigmine bromide. After 2 months adhesion molecule. of treatment with pyridostigmine bromide, the patient’s

(REPRINTED) ARCH DERMATOL / VOL 140, MAR 2004 WWW.ARCHDERMATOL.COM 328

Figure 1. Clinical results in a patient with pemphigus vulgaris treated with pyridostigmine bromide (Mestinon; ICS Pharmaceuticals, Costa Mesa, Calif). A, Before pyridostigmine bromide treatment; B, 2 months later. Prior to starting treatment with pyridostigmine bromide, this patient took prednisone at the dose of 20 mg/d. Pyridostigmine bromide was administered at the daily dose of 360 g. While the patient was receiving pyridostigmine bromide treatment, the dose of prednisone was decreased to 10 mg/d.

condition dramatically improved (Figure 1). The dose desmoglein 1 and anti–desmoglein 3 antibodies. Gross skin of pyridostigmine bromide was then decreased to 300 blisters developed approximately 40 hours after the first mg/d. The patient was treated with pyridostigmine bro- injection of PV IgG (Figure 2A). At the onset of skin blis- mide for an additional 3 months. Occasional redness tering, the Nikolskiy sign representing a loss of intraepi- and/or itching, burning, or tingling sensations of the skin dermal cohesion19 could be elicited in mice by applying lesion could be alleviated by increasing the pyridostig- lateral traction with a pencil eraser to the skin. When blis- mine bromide dose from 300 to 360 mg/d without chang- ters became generalized, the mice were euthanized. The ing the dose of prednisone. While receiving pyridostig- abdominal skin was then examined by light microscopy mine bromide treatment, the daily dose of prednisone was and direct immunofluorescence, revealing intraepider- tapered to 10 mg. Further decrease of prednisone dose was mal clefting due to extensive acantholysis and intercellu- associated with a flare of his skin lesion. No other therapy, lar epidermal staining consistent with binding of PV IgGs except for prednisone tapering, was used. The titer of in- to murine keratinocytes, respectively. None of the 4 nega- tercellular autoantibodies did not change. During the course tive control mice that were injected with pooled normal of pyridostigmine bromide treatment, he infrequently de- human IgG at the daily dose of 10 mg/g of body weight veloped adverse effects of pyridostigmine bromide, such for 2 days developed any gross or microscopic signs of pem- as nausea, abdominal cramps, and increased peristalsis. phigus or showed any deposition of human IgGs in their Other adverse effects of pyridostigmine bromide may in- skin during 40 hours of observation. To standardize as- clude skin flushes, sweating, vomiting, diarrhea, in- sessment of the extent of acantholysis, we computed the creased salivation, increased bronchial secretions, miosis areas of intraepidermal splitting in the images of skin har- and diaphoresis, fasciculation, and weakness. vested from the euthanized mice at the umbilical level 20 The clinical results in this patient suggested a di- hours after the injection. The results showed that in mice rect interrelationship between the use of pyridostig- treated with PV IgG alone, the acantholysis extended to mine bromide and the ability to keep his PV under con- 73.4%±11% of the epidermis (Table). trol. However, while the use of pyridostigmine bromide To determine whether a cholinergic agonist can abol- in this patient allowed to keep his disease under control ish PV IgG–induced acantholysis, litter mates of the athy- at a lower dose of prednisone than that used before start- mic nude mice were injected with the same dose of PV IgG ing treatment with pyridostigmine bromide, it did not together with carbachol (0.04 µg/g body weight). In ad- allow corticosteroid withdrawal. dition to being a mixed, nicotinic, and muscarinic agonist, carbachol is also a reversible AChE inhibitor.20 We RESULTS chose carbachol because in the past this cholinomimetic agent has been shown to antagonize PV IgG–induced ac- EXPERIMENTAL AMELIORATION antholysis in keratinocyte monolayers.7 In contrast to mice OF PV IgG–INDUCED ACANTHOLYSIS in the positive control group, none of the mice injected To induce experimental PV, for 2 consecutive days we in- with PV IgG together with carbachol developed any vis- jected intraperitoneally 3-day-old athymic nude mice with ible skin lesions 40 hours after the first injection (Figure IgG fraction (7 mg/g of body weight per day) pooled from 2B). At this point, our most vigorous efforts to induce Ni- the sera of 5 patients with acute PV containing anti– kolskiy sign failed in the skin of 4 of 7 mice in this sub-

(REPRINTED) ARCH DERMATOL / VOL 140, MAR 2004 WWW.ARCHDERMATOL.COM 329

C D

E F

Figure 2. Carbachol inhibits pemphigus vulgaris (PV) IgG–induced skin blistering of 3-day-old athymic nude mice. A, Positive control: a representative 3-day-old homozygous athymic nude mouse 40 hours after injection of 7 mg/g of body weight per day of PV IgG showing gross skin blistering. B, Experiment: a representative mouse of the same progeny after injection of 7 mg/g body weight per day of PV IgG together with 0.04 µg/g of body weight per day of carbachol. Note lack of gross skin blistering. C, Positive control: extensive acantholysis in murine epidermis 20 hours after injection of PV IgG alone (hematoxylin-eosin, scale bar=100 µm). D, Experiment: limited acantholysis in murine epidermis 20 hours after injection of PV IgG together with 0.04 µg/g body weight per day of carbachol (hemotoxylin-eosin, scale bar=100 µm). E and F, Direct immunofluorescence staining of the epidermis in positive control (E) and experimental (F) mice 20 hours after injection of 7 mg/g body weight per day of PV IgG together with 0.04 µg/g body weight per day of carbachol, using fluorescein isothiocyanate–labeled goat antihuman IgG antibody (scale bar=50 µm). Note that the pemphiguslike epidermal staining is present in both cases.

Results of Microscopic Analysis of Skin in Control and Experimental Mice*

Morphometric Assay of the Semiquantitative Assay of Epidermal Treatment (Control/Experimental) No. of Mice Extent of Acantholysis, % IgG Binding (Relative Values) PV IgG/PV IgG + cholinomimetic carbachol 4/7 73.4 ± 11/36.8 ± 2.4 67.0 ± 9.3/106.87 ± 7.9 PV IgG/PV IgG + pyridostigmine bromide 4/7 78.2 ± 1.9/31.3 ± 7.2 116.8 ± 7.6/188.9 ± 7.3

*Data are mean ± SD value in control/experimental mice. The extent of acantholysis and the intensity of epidermal staining for human IgG were estimated 20 hours after a single injection of pemphigus vulgaris (PV) IgG with or without test drug. PϽ.001 for all results.

group. Microscopic examination of the skin from eutha- nificant (PϽ.001) decrease of the extent of epidermal split- nized mice, however, was performed 20 hours after the ting from approximately 73% to 37% (Table). injection because at this time visible skin lesions were also Since epidermal keratinocytes synthesize and release absent in the positive control mice. Compared with an ex- ACh,21,22 which serves as an endogenous agonist of both tensive intraepidermal splitting seen in the positive con- the nicotinic and muscarinic classes of cholinergic recep- trol litter mates (Figure 2C), skin samples from carbachol- tors expressed in keratinocytes,23,24 we sought to deter- treated mice showed only limited areas of epidermal mine whether increasing the level of free ACh in the epi- acantholysis (Figure 2D). Morphometric analysis re- dermis can ameliorate the signs of experimental pemphigus. vealed that administration of carbachol resulted in a sig- Toward this goal, we injected a group of mice with PV IgG

(REPRINTED) ARCH DERMATOL / VOL 140, MAR 2004 WWW.ARCHDERMATOL.COM 330

Figure 3. Fluorescence images of E-cadherin (scale bar=50 µm). Results of semiquantitative indirect immunofluorescence analysis of the relative amounts of E-cadherin in cultured human foreskin keratinocytes incubated overnight without (control) (A) or with 0.25mM carbachol (B) or pyridostigmine bromide (C). After incubation, keratinocyte monolayers were washed and immunostained. The images were analyzed using software for semiquantitative image analysis.

together with the AChE inhibitor pyridostigmine bro- mide25 (0.1 µg/g body weight). The protective effect of pyri- Control 125 Cholinomimetic Carbachol dostigmine bromide on PV IgG–induced acantholysis was Pyridostigmine Bromide found to be similar to that of carbachol. The gross skin lesions did not appear, Nikolskiy sign was negative, and the 100 extent of acantholysis in the skin of mice treated with pyri- Ͻ dostigmine bromide decreased significantly (P .001) com- 75 pared with the positive control litter mates (Table). Since patients with PV develop antibodies to keratinocyte ACh receptors, along with autoantibodies to des- 50 mosomal cadherins (reviewed by Grando6), one of the 25

hypothetical mechanisms that could explain the antiacan- Relative Intensity of Fluorescence tholytic effect of carbachol and pyridostigmine bromide

(which, similar to carbachol, can act directly on ACh re- 0 ceptors in addition to reversibly inhibiting AChE)26 was the Plakoglobin E-cadherin direct competition of these drugs with PV IgG for binding to keratinocytes. To test this hypothesis, we measured the Figure 4. Cholinergic effects on the expression of adhesion molecules by keratinocytes. The results are expressed as the relative amounts of intensity of fluorescent staining of the epidermis of mice fluorescence intensity expressed by experimental vs nontreated cells injected with PV IgG alone (positive control) or together (control). Error bars indicate SD. with carbachol or pyridostigmine bromide (experiment), using fluorescein isothiocyanate–labeled goat antihuman IgG antibody (Figure 2E and F). Quantitative analysis of physiological effects of PV IgG.31 Formation of the adher- the intensity of specific staining showed an increased ence junctions is a prerequisite for the formation of des- amount of PV IgG in the epidermis of cholinergic agonist– mosomes.32 Although plakoglobin may not be the binding treated mice compared with that determined in positive con- site of pemphigus autoantibodies, it is present in the au- trol mice (Table), indicating that steric hindrance could not toantigen complex precipitated by PV IgG33 and its expres- account for the antiacantholytic effect of carbachol and pyri- sion is altered in PV.34 The semiquantitative indirect im- dostigmine bromide. An unexpected increase of the inten- munofluorescence assay demonstrated that treatment with sity of fluorescent staining of the epidermis in experimen- cholinergic agonists significantly (PϽ.05) increased the rela- tal mice could be explained through a hypothesis that tive amount of E-cadherin in keratinocytes (Figure 3 and cholinergic agonists up-regulated expression of keratino- Figure 4). The relative amount of plakoglobin did not in- cyte adhesion molecules targeted by PV IgG. crease. These indirect immunofluorescence findings were confirmed by the results of Western blot, which showed UP-REGULATION OF E-CADHERIN EXPRESSION that the protein level of E-cadherin was increased in cul- To investigate molecular mechanism(s) of antiacantho- tures treated with cholinergic agonists compared with non- lytic action of cholinergic agonists, we measured the ex- treated control monolayers (Figure 5). Both carbachol and pression of adhesion molecules in monolayers of normal pyridostigmine bromide caused elevation of the relative amount of E-cadherin (PϽ.05) without changing that of human keratinocytes incubated overnight in the presence Ͼ of 0.25mM carbachol or pyridostigmine bromide. Since acan- plakoglobin (P .05). tholysis in pemphigus starts at nondesmosomal areas of the 27-30 INHIBITION OF PV IgG–INDUCED keratinocyte cell membrane, we studied E-cadherin, a PHOSPHORYLATION protein involved in assembly of the adherence junctions, and plakoglobin (ie, ␥-catenin), which contributes to both Since phosphorylation has been recently recognized as the adherence and the desmosomal junction complexes and a major factor in regulation of intercellular adhesion,35 reportedly plays an important role in mediating the patho- particularly in disassembly of keratinocyte intercellular

(REPRINTED) ARCH DERMATOL / VOL 140, MAR 2004 WWW.ARCHDERMATOL.COM 331

©2004 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021 Pyridostigmine N IgG PV IgG PV IgG/ PV IgG Control Carbachol Control Bromide Carbachol Pyridostigmine Bromide

E-cadherin 120 kDa

WB 1345 1945 8329 15 750 ± 200 ±14∗ ± 673 ±641∗

1.00 1.73 1.27 0.60

Plakoglobin 82 kDa E-cadherin 32P

1773 2198 2305 2848 ± 330 ±101 ± 162 ±146 1.00 3.18 0.65 0.73

32P/WB 1.00 1.83 0.52 1.22

β -Actin 48 kDa WB

Figure 5. Cholinergic effects on the synthesis of adhesion molecules in 1.00 1.36 1.09 1.22 keratinocytes. Triplicate monolayers of normal human foreskin keratinocytes were treated overnight with 0.25mM carbachol or pyridostigmine bromide at 37°C, after which the total protein was isolated and measured, as detailed in Plakoglobin 32P the “Methods” section. Identical amounts of proteins from treated and nontreated control cultures were separated by 7% sodium dodecyl sulfate–polyacrylamide gel electrophoresis and electroblotted. Each 1.00 11.7 1.77 0.85 membrane was stained with a primary antibody to E-cadherin, plakoglobin, 32 or ␤-actin (to standardize the measurements). The numbers represent P/WB 1.00 8.62 1.62 0.70 mean±SD densitometry values of each protein band. Asterisks indicate significant (PϽ.05) differences from control. Figure 6. Cholinergic effects on pemphigus vulgaris (PV) IgG–induced phosphorylation of keratinocyte adhesion molecules. The levels of phosphorylation of E-cadherin and plakoglobin in cultured DJM-1 cells junctions caused by PV IgG,36,37 we sought to determine exposed for 1 hour to 1 mg/mL of normal human IgG (N IgG) or 1 mg/mL of PV IgG in the absence or presence of 0.5mM carbachol or pyridostigmine the effect of cholinergic agonists on PV IgG–induced phos- bromide. Each adhesion molecule was immunoprecipitated by the specific phorylation of keratinocyte adhesion molecules. To- monoclonal antibody, resolved by sodium dodecyl sulfate–polyacrylamide gel ward this end, we measured the effects of carbachol and electrophoresis and electroblotted. The amount of phosphorus 32 incorporated by each adhesion molecule was assayed by autoradiography. Parallel samples pyridostigmine bromide on PV IgG–induced phosphory- were analyzed by Western blot (WB), as described in the “Methods” section. lation of E-cadherin and plakoglobin using the DJM-1 The protein bands were visualized to determine the protein amount of each cutaneous squamous cell carcinoma cell line that fea- adhesion molecule in the sample. The results are expressed as ratios of the radioactivity value of each adhesion molecule to its protein quantity in each tures high levels of phosphorylation of adhesion mol- lane, compared with the ratios obtained in control cultures (taken as 1.00). ecules and, therefore, is customarily used to study phosphorylation of keratinocyte adhesion molecules.37-41 Incubation of keratinocyte monolayers for 1 hour with studied the immunopharmacology of pemphigus IgG ac- 1 mg/mL of pooled PV IgG resulted in 1.8- and 8.6-fold tion on keratinocytes. In the past, we reported that PV IgG– increase of the phosphorylation level of E-cadherin and induced phosphorylation of keratinocyte adhesion mol- plakoglobin, respectively (Figure 6). In the presence ecules could be abolished in the presence of the of 0.5mM carbachol or pyridostigmine bromide, this effect corticosteroid methylprednisolone.42 Increased phosphory- of PV IgG was attenuated, indicating that antiacantho- lation of desmoglein 3 in pemphigus may lead to the for- lytic activity of cholinomimetic drugs observed in vitro7 mation of desmoglein 3–depleted desmosomes and al- and in vivo (Figure 2) may stem from pharmacological tered adhesion.36,37 In addition to the phosphorylation of regulation of both the expression of E-cadherin and their desmoglein,43 desmocollin,44 and desmoplakin,45 assem- ability to attenuate PV IgG–induced phosphorylation of bly and disassembly of desmosomal junctions also in- E-cadherin and plakoglobin. volves phosphorylation of the keratin- and vimentin- intermediate filaments (reviewed by Eriksson et al46). For COMMENT instance, while phosphorylation of classic cadherins on tyrosine disables the adherence-type junctions, leading to cell- This study demonstrates that activation of the keratino- cell detachment,47,48 experimentally inhibiting tyrosine- cyte ACh axis can ameliorate pemphigus acantholysis and specific phosphatases results in a major changes in cell up-regulate the expression of adhesion molecules and pro- morphology, as manifested by a rapid rounding up of the tect them from PV IgG–induced phosphorylation. The cells, followed by reorganization of the cell monolayer.48 knowledge on the pathophysiology of acantholysis con- A list of known targets for pemphigus antibodies in- verges with that on the physiology of keratinocyte adhe- cludes both adhesion molecules (eg, desmogleins 1, 2, and sion. We hypothesized that a nonsteroidal treatment of pem- 3, desmocollins, and plakoglobin) and the receptor mol- phigus can be achieved by pharmacologically interceding ecules (Fc⑀RI␣, ␣3 and ␣9 nicotinic ACh receptor sub- at the site of intracellular biochemical events that mediate unit, pemphaxin, and other annexins)(reviewed by the acantholytic effects of pemphigus autoantibodies, and Grando6). Anti-ACh receptor antibodies are found in pa-

(REPRINTED) ARCH DERMATOL / VOL 140, MAR 2004 WWW.ARCHDERMATOL.COM 332

©2004 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021 tients with PV or pemphigus foliaceus.8 ACh signaling in safe dose of each drug was established, it was used for keratinocytes linked to the regulation of expression and experimental treatment of pemphigus. We found that the function of adhesion molecules and cholinergic drugs af- cholinergic agonists carbachol and pyridostigmine bro- fects cell shape, adhesion, and cytoplasm motility (re- mide partially inhibit PV IgG–induced acantholysis by viewed by Grando49). Because of the lack of a strong cor- passive transfer of autoantibodies without altering the relation between the clinical phenotype of PV and the binding of IgG to the keratinocyte cell membrane. presence of anti–desmoglein 1 and 3 antibodies50 and since The cholinergic drugs exerted their antiacantholytic PV-like lesions can be induced in neonatal mice in the ab- action through yet poorly understood intracellular signal- sence of these antibodies,9 the immunopathogenesis of PV ing pathways. Both drugs can reversibly inhibit AChE and can be explained through the “multiple hit” hypothesis. We ligate ACh receptors. While carbachol is a well-known propose that acantholysis in PV results from synergistic and mixed muscarinic and nicotinic agonist,20 pyridostigmine cumulative effects of autoantibodies targeting keratino- bromide interacts with the ACh-ionic channel complex, cyte cell membrane antigens of different kinds, including blocking it in open conformation.58 Thus, an important role (1) molecules that regulate cell shape and adhesion (eg, of phosphorylation of adhesion proteins for normal reor- ACh receptors) and (2) molecules that mediate cell-to- ganization of cadherin-cytoskeletal interactions, along with cell adhesion (eg, desmosomal cadherins). Severity of the the fact that ligation of ACh receptor types expressed in disease and exact clinical picture depend on the ratio of keratinocytes (eg, ␣9 ACh–gated ion channels) has been different kinds of autoantibodies in each particular pa- reported to induce phosphorylation of the cell membrane– tient. Antibodies to ACh receptors can weaken desmo- associated proteins,59 suggests that cholinomimetics ame- somal junctions by inducing phosphorylation of adhesion liorated acantholysis in keratinocytes exposed to PV IgG molecules, cause desmosome shedding owing to the apop- by inhibiting phosphorylation-mediated alterations in the tosis-related cleavage of desmosomal cadherins, and pre- assembly and disassembly of intercellular attachment units. vent desmosomal reassembly owing to the activation of the Additionally and/or alternatively, activation of the kerati- proteolytic cascade. In turn, the binding of pemphigus IgG nocyte ACh axis with cholinomimetic drugs could up- to desmosomal cadherins may prevent formation of new regulate expression of adhesion molecules targeted by pem- desmosomes because it blocks the extracellular domains phigus antibodies, as suggested by increased binding of of desmogleins mediating homophilic adhesion.51,52 PV IgG in the epidermis of mice treated with cholinergic We have previously reported that PV IgG–induced agonists. Indeed, we have recently found that ACh recep- acantholysis can be treated in culture with cholinergic tors expressed by keratinocytes couple a signaling path- agonists.7 While glucocorticosteroids or protease inhibi- way leading to activation of the adhesion molecules that tors can only block, but not reverse, acantholysis,1,53 cho- mediate intercellular attachment of these cells.13,60 The ex- linomimetics are the only drugs capable of reversing pression of desmogleins 1 and 3 was increased in keratino- PV IgG–induced acantholysis. The cholinergic effects on cytes treated with carbachol or pyridostigmine bromide.13 cell adhesion observed in cell monolayers7,23 have been Therefore, we speculate that pemphigus acantholysis in the corroborated by results showing enlargement of the in- skin of patients with PV could be treated by pharmaco- tercellular space between keratinocytes in the epider- logically stimulating keratinocyte ACh axis. Elucidation of mis treated with the nicotinic antagonist tubocurarine54 the cholinergic control of keratinocyte adhesion has a po- or the muscarinic antagonist atropine.55 The fact that pa- tential for the development of treatment regimens using tients with PV produce autoantibodies to ACh receptors safer drugs to control blistering in a variety of other skin expressed by keratinocytes, along with the fact that these diseases. First results of the clinical trial of pyridostig- ACh receptors regulate intercellular adhesion of these cells mine bromide in patients with pemphigus have been re- (the function that is altered in pemphigus), prompted us cently published.61 to use cholinomimetic drugs to prevent skin blistering in mice with experimentally induced pemphigus. Accepted for publication July 2, 2003. Passive transfer of PV IgG to neonatal Balb/c mice This work was supported by a research grant from the caused gross and microscopic changes consistent with Robert Leet & Clara Guthrie Patterson Trust, Hartford, Conn experimental pemphigus in vivo.56 In pilot studies,57 we (Dr Grando), and by the International Pemphigus Re- found that 3- to 5-day-old Balb/c mice may respond to search Fund, Sacramento, Calif. antiacantholytic treatments with corticosteroids or cho- This study was presented in part at the Fourth Joint linomimetics, but untreated positive controls do not al- Meeting of the European Society for Dermatological Re- ways produce PV lesions because rapidly developing hair search, Japanese Society for Investigative Dermatology, and follicles may reinforce their epidermal integrity. There- Society for Investigative Dermatology, May 4, 2002, Mi- fore, we sought to develop a more reliable animal model ami Fla, and was published in the form of an abstract (Grando and tested athymic nude mice. The 1- to 2-day-old mice SA, Arredondo J, Chernyavsky A, Pittelkow MR, Kitajima died despite any antiacantholytic treatments, whereas Y, Nguyen VT. Cholinergic stimulation inhibits pemphigus older mice responded to treatments and survived sub- IgG-induced acantholysis and ameliorates clinical disease sequent injections of PV IgG with a test drug. There- in a patient with pemphigus vulgaris. J Invest Dermatol. fore, we selected 3-day-old athymic nude mice as a model 2003;121[1]:abstract 42). for testing antiacantholytic drugs. The cholinergic drugs Corresponding author and reprints: Sergei A. Grando, used in the in vivo experiments were first tested in in- MD, PhD, DSc, Department of Dermatology, University of tact mice to assure that the doses at which they are used California Davis Medical Center, 4860 Y St, Room 3400, in adult humans are not lethal for neonatal mice. Once a Sacramento, CA 95817 (e-mail: [email protected]).

(REPRINTED) ARCH DERMATOL / VOL 140, MAR 2004 WWW.ARCHDERMATOL.COM 333

©2004 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021 32. Lewis JE, Wahl JK III, Sass KM, Jensen PJ, Johnson KR, Wheelock MJ. Cross- REFERENCES talk between adherens junctions and desmosomes depends on plakoglobin. J Cell Biol. 1997;136:919-934. 1. Swanson DL, Dahl MV. Methylprednisolone inhibits pemphigus acantholysis in 33. Korman NJ, Eyre RW, Klaus-Kovtun V, Stanley JR. Demonstration of an adhering- skin cultures. J Invest Dermatol. 1983;81:258-260. junction molecule (plakoglobin) in the autoantigens of pemphigus foliaceus and 2. Jeffes EW III, Kaplan RP, Ahmed AR. Acantholysis produced in vitro with pem- pemphigus vulgaris. N Engl J Med. 1989;321:631-635. phigus serum. J Clin Immunol. 1984;4:359-363. 34. Mignogna MD, Pannone G, Lo Muzio L, Staibano S, Bucci E. Catenin dislocation 3. Fine JD, Appell ML, Green LK, Sams WM Jr. Pemphigus vulgaris: combined treat- in oral pemphigus vulgaris. J Oral Pathol Med. 2001;30:268-274. ment with intravenous corticosteroid pulse therapy, plasmapheresis, and aza- 35. Amar LS, Shabana al HM, Oboeuf M, Martin N, Forest N. Desmosomes are regu- thioprine. Arch Dermatol. 1988;124:236-239. lated by protein kinase C in primary rat epithelial cells. Cell Adhes Commun. 1998; 4. Chryssomallis F, Dimitriades A, Chaidemenos GC, Panagiotides D, Karakatsanis 5:1-12. G. Steroid-pulse therapy in pemphigus vulgaris long term follow-up. Int J Der- 36. Aoyama Y, Kitajima Y. Pemphigus vulgaris-IgG causes a rapid depletion of des- matol. 1995;34:438-442. moglein 3 (Dsg3) from the triton X-100 soluble pools, leading to the formation 5. Werth VP. Treatment of pemphigus vulgaris with brief, high-dose intravenous of Dsg3-depleted desmosomes in a human squamous carcinoma cell line, DJM-1 glucocorticoids. Arch Dermatol. 1996;132:1435-1439. cells. J Invest Dermatol. 1999;112:67-71. 6. Grando SA. Autoimmunity to keratinocyte acetylcholine receptors in pemphi- 37. Aoyama Y, Owada MK, Kitajima Y. A pathogenic autoantibody, pemphigus vulgaris- gus. Dermatology. 2000;201:290-295. IgG, induces phosphorylation of desmoglein 3, and its dissociation from plako- 7. Grando SA, Dahl MV. Activation of keratinocyte muscarinic acetylcholine recep- globin in cultured keratinocytes. Eur J Immunol. 1999;29:2233-2240. tors reverses pemphigus acantholysis. J Eur Acad Dermatol Venereol. 1993;2: 38. Wakita H, Furukawa F, Baba S, Takigawa M. Human squamous-cell-carcinoma cell 72-86. line (DJM-1) cells synthesize P-cadherin molecules via an elevation of extracellular 8. Nguyen VT, Lee TX, Ndoye A, et al. The pathophysiological significance of non- calcium. Int J Cancer. 1997;73:432-439. desmoglein targets of pemphigus autoimmunity. Arch Dermatol. 1998;134:971- 39. Kitajima Y, Owada MK, Fujisawa Y, et al. A hemidesmosomal transmembrane 980. collagenous molecule, the 180-kDa bullous pemphigoid antigen (BPA II), is phos- 9. Nguyen VT, Ndoye A, Shultz LD, Pittelkow MR, Grando SA. Antibodies against phorylated with 12-O-tetradecanoylphorbol-13-acetate in a human squamous cell keratinocyte antigens other than desmogleins 1 and 3 can induce pemphigus carcinoma cell line (DJM-1). Epithelial Cell Biol. 1995;4:70-75. vulgaris-like lesions. J Clin Invest. 2000;106:1467-1479. 40. Kitajima Y, Aoyama Y, Seishima M. Transmembrane signaling for adhesive regu- 10. Zia S, Ndoye A, Nguyen VT, Grando SA. Nicotine enhances expression of the ␣3, lation of desmosomes and hemidesmosomes, and for cell-cell detachment in- ␣4, ␣5, and ␣7 nicotinic receptors modulating calcium metabolism and regu- duced by pemphigus IgG in cultured keratinocytes: involvement of protein ki- lating adhesion and motility of respiratory epithelial cells. Res Commun Mol Pathol nase C. J Investig Dermatol Symp Proc. 1999;4:137-144. Pharmacol. 1997;97:243-262. 41. Matsuoka Y, Yamada T, Seishima M, Hirako Y, Owaribe K, Kitajima Y. Transient 11. Ndoye A, Buchli R, Greenberg B, et al. Identification and mapping of keratino- translocation of hemidesmosomal bullous pemphigoid antigen 1 from cytosol cyte muscarinic acetylcholine receptor subtypes in human epidermis. J Invest to membrane fractions by 12-O-tetradecanoylphorbol-13-acetate treatment and Dermatol. 1998;111:410-416. Ca2+-switch in a human carcinoma cell line. J Dermatol Sci. 2001;27:206-214. 12. Nguyen VT, Ndoye A, Grando SA. Novel human ␣9 acetylcholine receptor regu- 42. Nguyen VT, Arredondo J, Chernyavsky A, Kitajima Y, Pittelkow MR, Grando SA. lating keratinocyte adhesion is targeted by pemphigus vulgaris autoimmunity. Pemphigus vulgaris IgG and methylprednisolone exhibit reciprocal effects on ke- Am J Pathol. 2000;157:1377-1391. ratinocytes. J Biol Chem. 2004;279:2135-2146. 13. Nguyen VT, Arredondo J, Chernyavsky AI, Kitajima Y, Grando SA. Keratinocyte 43. Rees DA. Junctional complexes of epithelial cells. Ciba Found Symp. 1987;125: acetylcholine receptors regulate cell adhesion. Life Sci. 2003;72:2081-2085. 1-273. 14. Mehta JN, Martin AG. A case of pemphigus vulgaris improved by cigarette smok- 44. Parrish EP, Marston JE, Mattey DL, Measures HR, Venning R, Garrod DR. Size ing. Arch Dermatol. 2000;136:15-17. heterogeneity, phosphorylation and transmembrane organisation of desmo- 15. Chaffins ML, Collison D, Fivenson DP. Treatment of pemphigus and linear IgA somal glycoproteins 2 and 3 (desmocollins) in MDCK cells. J Cell Sci. 1990;96: dermatosis with nicotinamide and tetracycline: a review of 13 cases. J Am Acad 239-248. Dermatol. 1993;28:998-1000. 45. Mueller H, Franke WW. Biochemical and immunological characterization of des- 16. Romanenko AV. The action of nicotinamide on neuromuscular transmission. Fi- moplakins I and II, the major polypeptides of the desmosomal plaque. J Mol Biol. ziol Zh. 1987;33:51-56. 1983;163:647-671. 17. Koeppen A, Klein J, Erb C, Loeffelholz K. Acetylcholine release and choline avail- 46. Eriksson JE, Opal P, Goldman RD. Intermediate filament dynamics. Curr Opin ability in rat hippocampus. J Pharmacol Exp Ther. 1997;282:1139-1145. Cell Biol. 1992;4:99-104. 18. Stoytcheva M, Zlatev R. Bioelectrocatalytical studies of the effect of some phar- 47. Behrens J, Vakaet L, Friis R, et al. Loss of epithelial differentiation and gain of maceuticals on the acetylcholinesterase activity. Electroanalysis. 1996;8:676- invasiveness correlates with tyrosine phosphorylation of the E-cadherin/␤- 679. catenin complex in cells transformed with a temperature-sensitive v-SRC gene. 19. Grando SA, Grando AA, Glukhenky BT, Doguzov V, Nguyen VT, Holubar K. His- J Cell Biol. 1993;120:757-766. tory and clinical significance of mechanical symptoms in blistering dermatoses: 48. Volberg T, Zick Y, Dror R, et al. The effect of tyrosine-specific protein phosphory- a reappraisal. J Am Acad Dermatol. 2003;48:86-92. lation on the assembly of adherens-type junctions. EMBO J. 1992;11:1733-1742. 20. DiPalma JR. Basic Pharmacology in Medicine. 4th ed. Willow Grove, Pa: Medi- 49. Grando SA. Biological functions of keratinocyte cholinergic receptors. J Inves- cal Surveillance Inc; 1994. tig Dermatol Symp Proc. 1997;2:41-48. 21. Grando SA, Kist DA, Qi M, Dahl MV. Human keratinocytes synthesize, secrete 50. Jamora MJ, Jiao D, Bystryn JC. Antibodies to desmoglein 1 and 3, and the clini- and degrade acetylcholine. J Invest Dermatol. 1993;101:32-36. cal phenotype of pemphigus vulgaris. J Am Acad Dermatol. 2003;48:976-977. 22. Wessler I, Kirkpatrick CJ, Racke K. Non-neuronal acetylcholine, a locally acting 51. Stanley JR. Defective cell-cell adhesion in the epidermis. Ciba Found Symp. 1995; molecule, widely distributed in biological systems: expression and function in 189:107-120. humans. Pharmacol Ther. 1998;77:59-79. 52. Amagai M. Autoimmunity against desmosomal cadherins in pemphigus. J Der- 23. Grando SA, Horton RM, Pereira EFR, et al. A nicotinic acetylcholine receptor regu- matol Sci. 1999;20:92-102. lating cell adhesion and motility is expressed in human keratinocytes. J Invest 53. Naito K, Morioka S, Nakajima S, Ogawa H. Proteinase inhibitors block formation Dermatol. 1995;105:774-781. of pemphigus acantholysis in experimental models of neonatal mice and skin 24. Grando SA, Zelickson BD, Kist DA, et al. Keratinocyte muscarinic acetylcholine explants. J Invest Dermatol. 1989;93:173-177. receptors. J Invest Dermatol. 1995;104:95-100. 54. Wessler I. New use of active ingredients which affect non-neuronal acetylcho- 25. Taylor P. Anticholinesterase agents. In: Gilman AG, Goodman LS, Rall TW, Mu- line functions. In: PCT Gazette. Mannheim, Germany: Boehringer Ingelheim KG; rad F, eds. Goodman and Gilman’s Pharmacological Basis of Therapeutics.7th 1998:WO 98/00119. ed. New York, NY: Macmillan Publishing Co Inc; 1985:110-127. 55. Wessler I, Kirkpatrick CJ, Racke K. The cholinergic “pitfall.” Clin Exp Pharmacol 26. Akaike A, Ikeda SR, Brookes N, Pascuzzo GJ, Rickett DL, Albuquerque EX. The na- Physiol. 1999;26:198-205. ture of the interactions of pyridostigmine with the nicotinic acetylcholine receptor- 56. Anhalt GJ, Labib RS, Voorhees JJ, Beals TF, Diaz LA. Induction of pemphigus in ionic channel complex: II, patch clamp studies. Mol Pharmacol. 1984;25:102-112. neonatal mice by passive transfer of IgG from patients with the disease. N Engl 27. Hashimoto K, Lever WF. An electron microscopic study on pemphigus vulgaris J Med. 1982;306:1189-1196. of the mouth and the skin with special reference to the intercellular cement. J Invest 57. Nguyen VT, Grando SA. Novel animal model for testing anti-acantholytic treat- Dermatol. 1967;48:540-552. ments of pemphigus [abstract 919]. J Invest Dermatol. 2001;117:543. 28. Wilgram GF, Caulfield JB, Lever WF. An electron microscopic study of acan- 58. Albuquerque EX, Akaike A, Shaw KP, Rickett DL. The interaction of anticholin- tholysis in pemphigus vulgaris. J Invest Dermatol. 1961;36:373-382. esterase agents with the acetylcholine receptor-ionic channel complex. Fundam 29. Barnett ML, Beutner EH, Chorzelski TP. Organ culture studies of pemphigus an- Appl Toxicol. 1984;4(2 pt 2):S27-S33. tibodies: II, ultrastructural comparison between acantholytic changes in vitro and 59. Szonyi M, Csermely P, Sziklai I. Acetylcholine-induced phosphorylation in iso- human pemphigus lesions. J Invest Dermatol. 1977;68:265-271. lated outer hair cells. Acta Otolaryngol. 1999;119:185-188. 30. Hu CH, Michel B, Schiltz JR. Epidermal acantholysis induced in vitro by pem- 60. Nguyen VT, Chernyavsky AI, Arredondo J, et al. Synergistic control of keratino- phigus autoantibody: an ultrastructural study. Am J Pathol. 1978;90:345-361. cyte adhesion through muscarinic and nicotinic acetylcholine receptor sub- 31. Caldelari R, de Bruin A, Baumann D, et al. A central role for the armadillo protein types. Exp Cell Res. 2004;294:534-549. plakoglobin in the autoimmune disease pemphigus vulgaris. J Cell Biol. 2001; 61. Grando SA. New approaches to the treatment of pemphigus. J Investig Derma- 153:823-834. tol Symp Proc. 2004;9:84-91.

(REPRINTED) ARCH DERMATOL / VOL 140, MAR 2004 WWW.ARCHDERMATOL.COM 334