Pemphigus Vulgaris Acantholysis Ameliorated by Cholinergic Agonists FREE

Pemphigus vulgaris (PV) is an autoimmune-mediated disease of skin andmucosa leading to progressive blistering and nonhealing erosions. Therapyfor patients with PV relies on the long-term use of systemic glucocorticosteroids in relatively large doses, which, although lifesaving, may cause severe adverseeffects, including death. Active disease state is characterized by the presence of serum IgG autoantibodies binding to the keratinocyte cell membrane proteins.Although the antiacantholytic effect of glucocorticosteroids is attributed to immunosuppression, high doses of glucocorticosteroids can directly blockPV IgG–induced acantholysis in vitro^{1 - 2} and rapidly (within 48 hours) stop blistering in patients with pemphigus withoutaltering the titer of autoantibodies or blocking antibody binding to keratinocytes ("pulse therapy").^{3 - 5} Patientsdevelop autoantibodies to keratinocyte cholinergic receptors regulating cell adhesion.⁶ Activation of these receptors mimicsantiacantholytic effects of glucocorticosteroids in vitro.⁷ We report that stimulation of the keratinocyte cholinergic receptors controlsPV IgG–induced acantholysis in neonatal mice and ameliorates the natural course of disease in a patient with PV. Cholinergic agonists stimulated expressionof E-cadherin and abolished phosphorylation of E-cadherin and plakoglobinin keratinocytes caused by PV IgG. Thus, novel antiacantholytic therapiesmay be developed based on the antiacantholytic effects of cholinomimetic drugs.

Neonatal athymic nude mice were used to test direct antiacantholytic effects of cholinergic agonists. At the third day of life, athymic nude miceweigh approximately 1.5 g and can develop gross skin blisters on passive transfer of PV antibodies. This study had been approved by the University of CaliforniaDavis Review Committee on the Use of Animals in Research, Sacramento, Calif. We injected 52 mice intraperitoneally through a 30-gauge needle with PV (experiment)or normal human IgG (negative control). One subgroup of mice received PV IgG alone and the other subgroup received PV IgG together with a test drug. Thepatients' and control human IgG serum fractions were obtained using 40% ammonium sulfate followed by dialysis against phosphate-buffered saline (PBS; GibcoBRL, Gaithersburg, Md). The fractions were then lyophilized and reconstituted in PBS as detailed elsewhere.⁸ The negativecontrol IgG was isolated from normal human serum purchased from Sigma Chemical Co, St Louis, Mo.

Gross skin lesions were observed approximately 24 to 40 hours after the first injection of PV IgG. Histological evaluation of lesional and perilesionalskin in mice with clinical symptoms of induced PV revealed extensive intraepidermal split and nonspecific changes. To allow accurate analysis of the extent ofacantholysis, we euthanized the mice at the preclinical stage of the disease (ie, approximately 20 hours after the injection) . The animals were cross-sectionedat the umbilicus level and freshly frozen. The need to compare skin specimens from the same body region stemmed from the fact that acantholysis in neonatalmice may vary between different anatomical regions.⁹ Three 6-µm thick cross sections of each mouse body were placed on the glassslide and stained with hematoxylin-eosin. The images (original magnification ×100) of 5 randomly selected longitudinal skin areas in each specimenwere photographed using a computer-linked microscope and printed. We determined the extent of acantholysis in the photographs by measuring the length of theintraepidermal split (at least 4 basal cells long) and expressed the results as percentage of the total length of epidermis in the field, taken as 100%.

Direct and indirect immunofluorescence assays were performed as detailed elsewhere.⁸ For the direct immunofluorescenceassay, a tissue specimen was incubated for 1 hour at room temperature with a fluorescently labeled goat antihuman IgG antibody (Pierce, Rockford, Ill;dilution 1:200 in PBS). For the indirect immunofluorescence experiments, a tissue substrate was first treated with a primary antibody to E-cadherin orplakoglobin (both from BD Transduction Laboratories, Lexington, Ky; dilution, 1:200) and then exposed to a secondary fluorescently labeled antibody (Pierce).Semiquantitative indirect immunofluorescence assay was performed as detailedpreviously,^{10 - 11} using a computer-assistedimage analysis with a software package purchased from Scanalytics (Fairfax, Va). For each tissue specimen, a minimum of 3 different segments in at least3 different microscopic fields were analyzed and the results compared.

Immunoblotting was performed as detailed by us previously.¹² Briefly, monolayers of normal human keratinocytes were either untreated (control) orincubated overnight with 0.25mM carbachol or 0.25mM pyridostigmine bromide at 37°C. They were then washed and lysed, and the lysates were resolvedby a 7.5% sodium dodecyl sulfate–polyacrylamide gel electrophoresis, followed by electroblotting. The membranes were blocked and stained with aprimary antibody to either E-cadherin (dilution, 1:2000) or plakoglobin (dilution, 1:1000) or with antibody to β-actin (Sigma Chemical Co; dilution, 1:2000)to standardize the measurements. Binding of the primary antibody was visualized with horseradish peroxidase–conjugated secondary antibodies and developedwith the chemoluminescence method of the ECL + Plus system (Amersham Pharmacia Biotech, Inc, Piscataway, NJ). The results were expressed as actual densitometryvalues of each protein band representing an adhesion molecule.

Quantitative phosphorylation assay was designed based on the established protocols, as detailed elsewhere.¹³ Briefly,cultured DJM-1 cells were grown to approximately 75% confluence in 75-cm² flasks (Corning Costar, Cambridge, Mass) in serum-free keratinocyte growth medium (Clonetics Corp, San Diego, Calif) containing 0.09mM Ca²⁺ at 37°C in a humid 5% carbon dioxide incubator. The cells werethen treated for 14 hours with phosphate-free minimal essential medium containing 1.8mM Ca²⁺ and 5% dialyzed fetal calf serum (Gibco BRL) and thentreated with the same medium containing 200 µCi/mL of ³²P orthophosphate (Amersham Pharmacia Biotech, Inc). The cells were labeled for8 hours and then exposed for 1 hour to 1 mg/mL of pooled normal human IgG(Sigma Chemical Co) or PV IgG pooled from 5 PV sera samples in the absenceor presence of 0.5mM carbachol or pyridostigmine bromide. After incubation, the cells were lysed and the lysates were cleared by centrifugation at 14 000g for 10 minutes. The adhesion molecules under considerationwere immunoprecipitated, resolved by 7.5% sodium dodecyl sulfate–polyacrylamidegel electrophoresis, electroblotted, and analyzed with the PhosphorImagerfeature of the Storm system (Molecular Dynamics, Mountain View, Calif). To determine protein concentration of each precipitated adhesion molecule, thesamples were analyzed by Western blotting, as described above. The results were expressed as ratios of the radioactivity value of each adhesion molecule to its protein quantity in each sample, compared with the ratios obtained in control cultures (taken as 1.00).

The results of quantitative experiments were expressed as mean ± SD. Significance was determined using the t test.The results were deemed significant if the P value was less than .05.

A recently reported case of PV that had improved by cigarette smoking¹⁴ and a study showing successful use of nicotinamide as a steroid-sparing agent in pemphigus,¹⁵ suggestedthat pharmacological regulation of keratinocyte acetylcholine (ACh) axis may be a novel antiacantholytic therapy for pemphigus because (1) cigarette smokecontains the cholinomimetic agent nicotine and (2) nicotinamide exhibits cholinomimetic effects¹⁶ owing to the stimulation of ACh release¹⁷ and inhibition of acetylcholinesterase (AChE).¹⁸ To determine whether a pharmacologic stimulationof keratinocyte cholinergic receptors can be used as a steroid-sparing regimen in the treatment of pemphigus, we administered pyridostigmine bromide (Mestinon;ICS Pharmaceuticals, Costa Mesa, Calif; 60-mg tablets) to a patient with activePV at the dose of 360 mg/d. The use of Mestinon in a patient with PV had beenapproved by the University of California Davis Human Subjects Review Committee. This patient, an 82-year-old white man, had been treated for almost 8 yearswith a mid-dose of prednisone, ranging from 15 to 30 mg/d, and occasional intralesional corticosteroid injections. He had a recalcitrant erosion onhis nose, which would never completely heal and would become active (ie, turn red and/or get painful, enlarge in size, and produce exudate). No other lesionswere seen. The lesion began to improve starting from the third week of treatment with pyridostigmine bromide. After 2 months of treatment with pyridostigminebromide, the patient's condition dramatically improved (Figure 1). The dose of pyridostigmine bromide was then decreased to 300 mg/d. The patient was treated with pyridostigmine bromide for an additional3 months. Occasional redness and/or itching, burning, or tingling sensations of the skin lesion could be alleviated by increasing the pyridostigmine bromidedose from 300 to 360 mg/d without changing the dose of prednisone. While receiving pyridostigmine bromide treatment, the daily dose of prednisone was taperedto 10 mg. Further decrease of prednisone dose was associated with a flare of his skin lesion. No other therapy, except for prednisone tapering, wasused. The titer of intercellular autoantibodies did not change. During the course of pyridostigmine bromide treatment, he infrequently developed adverseeffects of pyridostigmine bromide, such as nausea, abdominal cramps, and increased peristalsis. Other adverse effects of pyridostigmine bromide may include skinflushes, sweating, vomiting, diarrhea, increased salivation, increased bronchial secretions, miosis and diaphoresis, fasciculation, and weakness.

The clinical results in this patient suggested a direct interrelationship between the use of pyridostigmine bromide and the ability to keep his PV undercontrol. However, while the use of pyridostigmine bromide in this patientallowed to keep his disease under control at a lower dose of prednisone thanthat used before starting treatment with pyridostigmine bromide, it did not allow corticosteroid withdrawal.

To induce experimental PV, for 2 consecutive days we injected intraperitoneally 3-day-old athymic nude mice with IgG fraction (7 mg/g of body weight per day)pooled from the sera of 5 patients with acute PV containing anti–desmoglein 1 and anti–desmoglein 3 antibodies. Gross skin blisters developed approximately40 hours after the first injection of PV IgG (Figure 2A). At the onset of skin blistering, the Nikolskiy sign representing a loss of intraepidermal cohesion¹⁹ could be elicited in mice by applying lateral tractionwith a pencil eraser to the skin. When blisters became generalized, the mice were euthanized. The abdominal skin was then examined by light microscopyand direct immunofluorescence, revealing intraepidermal clefting due to extensive acantholysis and intercellular epidermal staining consistent with bindingof PV IgGs to murine keratinocytes, respectively. None of the 4 negative control mice that were injected with pooled normal human IgG at the daily dose of10 mg/g of body weight for 2 days developed any gross or microscopic signs of pemphigus or showed any deposition of human IgGs in their skin during 40hours of observation. To standardize assessment of the extent of acantholysis, we computed the areas of intraepidermal splitting in the images of skin harvestedfrom the euthanized mice at the umbilical level 20 hours after the injection.The results showed that in mice treated with PV IgG alone, the acantholysisextended to 73.4% ± 11% of the epidermis (Table 1).

To determine whether a cholinergic agonist can abolish PV IgG–inducedacantholysis, litter mates of the athymic nude mice were injected with thesame dose of PV IgG together with carbachol (0.04 µg/g body weight). In addition to being a mixed, nicotinic, and muscarinic agonist, carbacholis also a reversible AChE inhibitor.²⁰ We chose carbachol because in the past this cholinomimetic agent has been shown toantagonize PV IgG–induced acantholysis in keratinocyte monolayers.⁷ In contrast to mice in the positive control group, none of the mice injected with PV IgG together with carbachol developed anyvisible skin lesions 40 hours after the first injection (Figure 2B). At this point, our most vigorous efforts to induce Nikolskiy sign failed in the skin of 4 of 7 mice in this subgroup. Microscopic examinationof the skin from euthanized mice, however, was performed 20 hours after theinjection because at this time visible skin lesions were also absent in thepositive control mice. Compared with an extensive intraepidermal splitting seen in the positive control litter mates (Figure 2C), skin samples from carbachol-treated mice showed onlylimited areas of epidermal acantholysis (Figure 2D). Morphometric analysis revealed that administration of carbachol resulted in a significant (P<.001) decreaseof the extent of epidermal splitting from approximately 73% to 37% (Table 1).

Since epidermal keratinocytes synthesize and release ACh,^{21 - 22} which serves as an endogenous agonist of both the nicotinic and muscarinic classesof cholinergic receptors expressed in keratinocytes,^{23 - 24} we sought to determine whether increasing the level of free ACh in the epidermiscan ameliorate the signs of experimental pemphigus. Toward this goal, we injected a group of mice with PV IgG together with the AChE inhibitor pyridostigminebromide²⁵ (0.1 µg/g body weight). The protective effect of pyridostigmine bromide on PV IgG–induced acantholysiswas found to be similar to that of carbachol. The gross skin lesions did not appear, Nikolskiy sign was negative, and the extent of acantholysis in theskin of mice treated with pyridostigmine bromide decreased significantly (P<.001) compared with the positive control litter mates(Table 1).

Since patients with PV develop antibodies to keratinocyte ACh receptors, along with autoantibodies to desmosomal cadherins (reviewed by Grando⁶ ), one of the hypothetical mechanisms that could explainthe antiacantholytic effect of carbachol and pyridostigmine bromide (which, similar to carbachol, can act directly on ACh receptors in addition to reversiblyinhibiting AChE)²⁶ was the direct competition of these drugs with PV IgG for binding to keratinocytes. To test this hypothesis,we measured the intensity of fluorescent staining of the epidermis of mice injected with PV IgG alone (positive control) or together with carbachol orpyridostigmine bromide (experiment), using fluorescein isothiocyanate–labeledgoat antihuman IgG antibody (Figure 2Eand F). Quantitative analysis of the intensity of specific staining showed an increased amount of PV IgG in the epidermis of cholinergic agonist–treatedmice compared with that determined in positive control mice (Table 1), indicating that steric hindrance could not account for the antiacantholytic effect of carbachol and pyridostigmine bromide. An unexpectedincrease of the intensity of fluorescent staining of the epidermis in experimental mice could be explained through a hypothesis that cholinergic agonists up-regulatedexpression of keratinocyte adhesion molecules targeted by PV IgG.

To investigate molecular mechanism(s) of antiacantholytic action of cholinergic agonists, we measured the expression of adhesion molecules inmonolayers of normal human keratinocytes incubated overnight in the presenceof 0.25mM carbachol or pyridostigmine bromide. Since acantholysis in pemphigusstarts at nondesmosomal areas of the keratinocyte cell membrane,^{27 - 30} we studied E-cadherin, a protein involved in assembly of the adherence junctions,and plakoglobin (ie, γ-catenin), which contributes to both the adherence and the desmosomal junction complexes and reportedly plays an important rolein mediating the pathophysiological effects of PV IgG.³¹ Formation of the adherence junctions is a prerequisite for the formation of desmosomes.³² Although plakoglobin may not be the binding siteof pemphigus autoantibodies, it is present in the autoantigen complex precipitated by PV IgG³³ and its expression is altered inPV.³⁴ The semiquantitative indirect immunofluorescenceassay demonstrated that treatment with cholinergic agonists significantly(P<.05) increased the relative amount of E-cadherinin keratinocytes (Figure 3 and Figure 4). The relative amount of plakoglobindid not increase. These indirect immunofluorescence findings were confirmed by the results of Western blot, which showed that the protein level of E-cadherinwas increased in cultures treated with cholinergic agonists compared with nontreated control monolayers (Figure 5).Both carbachol and pyridostigmine bromide caused elevation of the relativeamount of E-cadherin (P<.05) without changingthat of plakoglobin (P>.05).

Since phosphorylation has been recently recognized as a major factor in regulation of intercellular adhesion,³⁵ particularlyin disassembly of keratinocyte intercellular junctions caused by PV IgG,^{36 - 37} we sought to determine the effect of cholinergic agonists on PV IgG–induced phosphorylation of keratinocyteadhesion molecules. Toward this end, we measured the effects of carbachol and pyridostigmine bromide on PV IgG–induced phosphorylation of E-cadherinand plakoglobin using the DJM-1 cutaneous squamous cell carcinoma cell line that features high levels of phosphorylation of adhesion molecules and, therefore,is customarily used to study phosphorylation of keratinocyte adhesion molecules.^{37 - 41} Incubation of keratinocyte monolayers for 1 hour with 1 mg/mL of pooled PV IgG resultedin 1.8- and 8.6-fold increase of the phosphorylation level of E-cadherin andplakoglobin, respectively (Figure 6).In the presence of 0.5mM carbachol or pyridostigmine bromide, this effect of PV IgG was attenuated, indicating that antiacantholytic activity of cholinomimeticdrugs observed in vitro⁷ and in vivo (Figure 2) may stem from pharmacological regulation of both the expression of E-cadherin and their ability to attenuate PV IgG–inducedphosphorylation of E-cadherin and plakoglobin.

This study demonstrates that activation of the keratinocyte ACh axis can ameliorate pemphigus acantholysis and up-regulate the expression of adhesionmolecules and protect them from PV IgG–induced phosphorylation. The knowledge on the pathophysiology of acantholysis converges with that on thephysiology of keratinocyte adhesion. We hypothesized that a nonsteroidal treatment of pemphigus can be achieved by pharmacologically interceding at the siteof intracellular biochemical events that mediate the acantholytic effects of pemphigus autoantibodies, and studied the immunopharmacology of pemphigusIgG action on keratinocytes. In the past, we reported that PV IgG–inducedphosphorylation of keratinocyte adhesion molecules could be abolished in thepresence of the corticosteroid methylprednisolone.⁴² Increased phosphorylation of desmoglein 3 in pemphigus may lead to the formation ofdesmoglein 3–depleted desmosomes and altered adhesion.^{36 - 37} In addition to the phosphorylation of desmoglein,⁴³ desmocollin,⁴⁴ and desmoplakin,⁴⁵ assemblyand disassembly of desmosomal junctions also involves phosphorylation of the keratin- and vimentin-intermediate filaments (reviewed by Eriksson et al⁴⁶ ). For instance, while phosphorylation of classiccadherins on tyrosine disables the adherence-type junctions, leading to cell-cell detachment,^{47 - 48} experimentallyinhibiting tyrosine-specific phosphatases results in a major changes in cell morphology, as manifested by a rapid rounding up of the cells, followed byreorganization of the cell monolayer.⁴⁸

A list of known targets for pemphigus antibodies includes both adhesion molecules (eg, desmogleins 1, 2, and 3, desmocollins, and plakoglobin) andthe receptor molecules (FcϵRIα, α3 and α9 nicotinic ACh receptor subunit, pemphaxin, and other annexins)(reviewed by Grando⁶ ). Anti-ACh receptor antibodies are found in patientswith PV or pemphigus foliaceus.⁸ ACh signaling in keratinocytes linked to the regulation of expression and function of adhesionmolecules and cholinergic drugs affects cell shape, adhesion, and cytoplasm motility (reviewed by Grando⁴⁹ ). Because ofthe lack of a strong correlation between the clinical phenotype of PV and the presence of anti–desmoglein 1 and 3 antibodies⁵⁰ andsince PV-like lesions can be induced in neonatal mice in the absence of these antibodies,⁹ the immunopathogenesis of PV canbe explained through the "multiple hit" hypothesis. We propose that acantholysis in PV results from synergistic and cumulative effects of autoantibodies targetingkeratinocyte cell membrane antigens of different kinds, including (1) molecules that regulate cell shape and adhesion (eg, ACh receptors) and (2) moleculesthat mediate cell-to-cell adhesion (eg, desmosomal cadherins). Severity of the disease and exact clinical picture depend on the ratio of different kindsof autoantibodies in each particular patient. Antibodies to ACh receptors can weaken desmosomal junctions by inducing phosphorylation of adhesion molecules,cause desmosome shedding owing to the apoptosis-related cleavage of desmosomal cadherins, and prevent desmosomal reassembly owing to the activation of theproteolytic cascade. In turn, the binding of pemphigus IgG to desmosomal cadherinsmay prevent formation of new desmosomes because it blocks the extracellulardomains of desmogleins mediating homophilic adhesion.^{51 - 52}

We have previously reported that PV IgG–induced acantholysis can be treated in culture with cholinergic agonists.⁷ Whileglucocorticosteroids or protease inhibitors can only block, but not reverse, acantholysis,^{1 ,53} cholinomimeticsare the only drugs capable of reversing PV IgG–induced acantholysis.The cholinergic effects on cell adhesion observed in cell monolayers^{7 ,23} have been corroborated by resultsshowing enlargement of the intercellular space between keratinocytes in the epidermis treated with the nicotinic antagonist tubocurarine⁵⁴ orthe muscarinic antagonist atropine.⁵⁵ The fact that patients with PV produce autoantibodies to ACh receptors expressed bykeratinocytes, along with the fact that these ACh receptors regulate intercellular adhesion of these cells (the function that is altered in pemphigus), promptedus to use cholinomimetic drugs to prevent skin blistering in mice with experimentally induced pemphigus.

Passive transfer of PV IgG to neonatal Balb/c mice caused gross and microscopic changes consistent with experimental pemphigus in vivo.⁵⁶ In pilot studies,⁵⁷ wefound that 3- to 5-day-old Balb/c mice may respond to antiacantholytic treatments with corticosteroids or cholinomimetics, but untreated positive controls donot always produce PV lesions because rapidly developing hair follicles may reinforce their epidermal integrity. Therefore, we sought to develop a morereliable animal model and tested athymic nude mice. The 1- to 2-day-old mice died despite any antiacantholytic treatments, whereas older mice respondedto treatments and survived subsequent injections of PV IgG with a test drug.Therefore, we selected 3-day-old athymic nude mice as a model for testingantiacantholytic drugs. The cholinergic drugs used in the in vivo experiments were first tested in intact mice to assure that the doses at which they areused in adult humans are not lethal for neonatal mice. Once a safe dose of each drug was established, it was used for experimental treatment of pemphigus.We found that the cholinergic agonists carbachol and pyridostigmine bromidepartially inhibit PV IgG–induced acantholysis by passive transfer ofautoantibodies without altering the binding of IgG to the keratinocyte cellmembrane.

The cholinergic drugs exerted their antiacantholytic action through yet poorly understood intracellular signaling pathways. Both drugs can reversiblyinhibit AChE and ligate ACh receptors. While carbachol is a well-known mixed muscarinic and nicotinic agonist,²⁰ pyridostigminebromide interacts with the ACh-ionic channel complex, blocking it in open conformation.⁵⁸ Thus, an important role ofphosphorylation of adhesion proteins for normal reorganization of cadherin-cytoskeletal interactions, along with the fact that ligation of ACh receptor types expressedin keratinocytes (eg, α9 ACh–gated ion channels) has been reported to induce phosphorylation of the cell membrane–associated proteins,⁵⁹ suggests that cholinomimetics ameliorated acantholysisin keratinocytes exposed to PV IgG by inhibiting phosphorylation-mediated alterations in the assembly and disassembly of intercellular attachment units.Additionally and/or alternatively, activation of the keratinocyte ACh axiswith cholinomimetic drugs could up-regulate expression of adhesion moleculestargeted by pemphigus antibodies, as suggested by increased binding of PV IgG in the epidermis of mice treated with cholinergic agonists. Indeed, wehave recently found that ACh receptors expressed by keratinocytes couple a signaling pathway leading to activation of the adhesion molecules that mediateintercellular attachment of these cells.^{13 ,60} The expression of desmogleins 1 and 3 was increased in keratinocytes treated withcarbachol or pyridostigmine bromide.¹³ Therefore, we speculate that pemphigus acantholysis in the skin of patients with PV couldbe treated by pharmacologically stimulating keratinocyte ACh axis. Elucidation of the cholinergic control of keratinocyte adhesion has a potential for thedevelopment of treatment regimens using safer drugs to control blistering in a variety of other skin diseases. First results of the clinical trial ofpyridostigmine bromide in patients with pemphigus have been recently published.⁶¹

Corresponding author and reprints: Sergei A. Grando, MD, PhD, DSc, Department of Dermatology, University of California Davis Medical Center,4860 Y St, Room 3400, Sacramento, CA 95817 (e-mail: sagrando@ucdavis.edu).

Accepted for publication July 2, 2003.

This work was supported by a research grant from the Robert Leet & Clara Guthrie Patterson Trust, Hartford, Conn (Dr Grando), and by the InternationalPemphigus Research Fund, Sacramento, Calif.

This study was presented in part at the Fourth Joint Meeting of the European Society for Dermatological Research, Japanese Society for InvestigativeDermatology, and Society for Investigative Dermatology, May 4, 2002, Miami Fla, and was published in the form of an abstract (Grando SA, Arredondo J,Chernyavsky A, Pittelkow MR, Kitajima Y, Nguyen VT. Cholinergic stimulation inhibits pemphigus IgG-induced acantholysis and ameliorates clinical diseasein a patient with pemphigus vulgaris. J Invest Dermatol. 2003;121[1]:abstract 42).