ORIGINAL ARTICLESee related Editorial on page iv

Actin Reorganization Is Abnormal and Cellular ATP Is Decreasedin Hailey-Hailey Keratinocytes

Ida Aronchik,� Martin J. Behne,�w Laura Leypoldt,w Debbie Crumrine,� Ervin Epstein,�z Shigaku Ikeda,yMasayuki Mizoguchi,y Graham Bench,yyTullio Pozzan,z and Theodora Mauro�w�Department of Dermatology, University of California, San Francisco, California, USA; wDermatology Service,VA Medical Center, San Francisco,California, USA; zDepartment of Dermatology, San Francisco General Hospital, San Francisco, California, USA; zDepartment of Biomedical Sciences andConsiglio Nazionale delle Ricerche (CNR) Center for the Study of Biomembranes, University of Padova, Padova, Italy; yDepartment of Dermatology,Juntendo University School of Medicine,Tokyo, Japan; yyCenter for Accelerator Mass Spectrometry, Lawrence Livermore National Laboratory, Livermore,California, USA

Actin reorganization and the formation of adherensjunctions are necessary for normal cell-to-cell adhesionin keratinocytes. Hailey-Hailey disease (HHD) is blis-tering skin disease, resulting from mutations in theCa2þ ATPase ATP2C1, which controls Ca2þ concentra-tions in the cytoplasm and Golgi of human keratino-cytes. Because actin reorganization is among the ¢rstresponses to raised cytoplasmic Ca2þ, we examinedCa2þ -induced actin reorganization in normal andHHD keratinocytes. Even though HHD keratinocytesdisplay raised baseline cytoplasmic Ca2þ , we found thatactin reorganization in response to Ca2þ was impairedin HHD keratinocytes. Defects in actin reorganizationwere linked to a marked decrease in cellular ATP inHHD keratinocytes, which persists, in vivo, in HHDepidermis. Defective actin reorganization was repro-

duced in normal keratinocytes in which the intra-cellular ATP concentration had been lowered pharma-cologically. ATP concentrations in undi¡erentiatedkeratinocytes markedly declined after extracellularCa2þ was increased, but then recovered to a new base-line that was approximately 150% of the previousbaseline. In contrast, ATP concentrations in HHDkeratinocytes did not change in response to increasedextracellular Ca2þ. This report provides new insightsinto how the ATP2C1-controlled ATP metabolismmediates Ca2þ -induced cell-to-cell adhesion in normalkeratinocytes. In addition, these ¢ndings implicate in-adequate ATP stores as an additional cause in the patho-genesis of HHD and suggest novel therapeutic options.Keyword: calcium/adhesion. J Invest Dermatol 121:681 ^687,2003

Hailey-Hailey disease (HHD, OMIM 16960) is a blis-tering dermatosis inherited in an autosomal domi-nant fashion. Also known as familial benignchronic pemphigus, it was ¢rst described in 1939(Hailey and Hailey, 1939). The disorder, which

usually presents in adulthood, is characterized by vesicles andcrusted erosions, as a result of impaired cell-to-cell adhesion(acantholysis), causing an intercellular split among suprabasal ker-atinocytes in the epidermis. HHD is caused by mutations inATP2C1, which encodes an intracellular Ca2þ pump (Hu et al,2000; Sudbrak et al, 2000). Recent studies have shown that thisCa2þ ATPase is localized to the Golgi in keratinocytes (Behneet al, 2002), consistent with the location of its homolog, PMR1,in yeast (Antebi and Fink, 1992). Studies in Caenorhabditis elegansdemonstrate that this class of Ca2þ ATPases is capable of trans-porting Ca2þ from the cytosol into intracellular stores (Van Bae-len et al, 2001). Because the Golgi is known to function as aninositol 1,4,5-trisphosphate-sensitive Ca2þ store (Pinton et al,

1998), it is not surprising that regulation of cytosolic Ca2þ wasimpaired in keratinocytes cultured from clinically uninvolvedskin of HHD patients (Hu et al, 2000). These discoveries suggestthat intracellular Ca2þ stores play a major role in regulation ofepidermal cell-to-cell adhesion.The formation of adherens junctions depends both on adhe-

rens junction proteins, such as E-cadherin, and on remodelingof the actin cytoskeleton (Vasioukhin et al, 2000).When intracel-lular Ca2þ increases, the actin cytoskeleton rapidly remodels,forming ¢lopodia that intercalate with ¢lopodia from adjoiningcells to form ‘‘adhesion zippers,’’ the sca¡old on which stable ad-herens junctions form (Vasioukhin et al, 2000). Because actinpolymerization is stimulated by high concentrations of Ca2þ ,and we previously have shown that cytoplasmic Ca2þ is abnor-mally increased in HHD keratinocytes (Hu et al, 2000), we exam-ined whether actin reorganization was upregulated in HHDkeratinocytes. Surprisingly, we found that actin reorganizationwas instead downregulated in HHD keratinocytes and that thisimpairment was linked to decreased concentrations of ATP.Whereas attention has focused on defects in adherens junctionproteins in HHD, defective actin processing also has been re-ported. Skin specimens from patients with HHD reveal de¢cientactin reorganization with abnormal stress ¢bers (Metze et al, 1996)and abnormal localization of actin ¢laments (Inohara et al,1990).We report that these de¢ciencies persist in vitro and that de-creased ATP levels are found both in HHD keratinocytes and in

Reprint requests to: Martin Behne, MD, Dermatology Service (190),Veterans A¡airs Medical Center, 4150 Clement Street, San Francisco,CA 94121. Email: behnemj@itsa.ucsf.eduAbbreviations: HHD, Hailey-Hailey disease; P, phosphorous; PIXE,

proton-induced X-ray emission.

Manuscript received December 12, 2002; revised January 29, 2003;accepted for publication March 5, 2003

0022-202X/03/$15.00 . Copyright r 2003 by The Society for Investigative Dermatology, Inc.

681

epidermis. These studies describe a new pathophysiologicmechanism underlying the acantholysis that characterizes HHD.

MATERIALS AND METHODS

HHD patient selection Three patients had clinical ¢ndings (blisteringand erosions in characteristic locations) and histological pathology(acantholysis of suprabasal cells without apoptosis) characteristic of HHD.This study was approved by the University of California at San FranciscoCommittee on Human Research. Normal human keratinocytes, matchedto age, location of skin biopsy (trunk), and pass number were used ascontrols. All keratinocytes used were pass 3^7.

Cell culture Cultured human keratinocytes from adult normal skin(surgical skin margins) or clinically normal HHD skin (punch biopsies)were grown on 60-mm dishes, six-well plates, or chamber slides in0.06 mM Ca2þ Epilife medium or 0.07 mM Ca2þ 154 medium (CascadeBiologics, Eugene, OR) until 60%^70% con£uent; all cells were thentreated with 0.07 mM Ca2þ 154 medium until approximately 80%^90%con£uence (or 0.5 million cells/well) for most experiments or con£uencefor ATP measurements (see below). All cells then were treated with 154medium containing 0.07 or 1.2 mM Ca2þ for the intervals detailed ineach experiment. Antimycin A (Cat. No. 10792 Sigma-Fluka, St. Louis,MO) was dissolved in ethanol (50 mg/mL) and then added at a ¢nalconcentration of 0.1 mM immediately before experiments.Smooth muscle cells were isolated from adult healthy rats by collecting

tissue from thoracic artery. The tissue was cut into small pieces (severalmillimeters in diameter), washed several times with PBS, and suspendedin a solution of collagenase and dispase. After a 1-week-long incubation,colonies of smooth muscle cells were isolated and seeded on 100-mmplastic plates (08-757-12, Fisher, Pittsburgh, PA) in FBS-enriched DMEM.The cells were passed at con£uency by applying 3 mL of trypsin per100-mm plate, incubating for 5 min at 371C, and collecting into 7 mLof media. The cells were centrifuged, and the pellet was resuspended andseeded at a 1:20 ratio for colony propagation. The cells were previouslyfound to be stable for 22 passes (J. Karliner, personal communication); thecells used in the ATP measurements were of pass 15.

ATP £uorescence assay Normal and HHD keratinocytes were culturedas described above in noncoated six-well plates (Falcon, Becton-Dickinson,Franklin Lakes, NJ). All cells then were treated with 154 mediumcontaining 0.07 or 1.2 mM Ca2þ for the intervals detailed in eachexperiment. After treatment, medium from the wells was aspirated andcells were washed three times with 2 mL of ice-cold sterile PBS. Theplates were incubated for 5 min at 371C after each wash. The cells thenwere bathed in 1 mL of PBS and agitated from the surface of the plateswith plastic cell scrapers. The suspension was homogenized by pipettingrepeatedly up and down, and the entire volume of each well was collectedinto 1.5-mL centrifuge tubes. The protocol for measurement of ATPconcentrations was adapted from an ATP luminescence kit (FL-ASC,Sigma, St. Louis, MO). Standard dilutions were prepared from a knownATP standard. The range of applicable standards was established between1:10 and 1:2000 of the original solution. The reaction for each standardand each sample was performed by using two vials, one of whichcontained 0.1 mL of ATP reaction mix and the second containing 0.1 mLof cell-lysing reagent and 0.05 mL of ddH2O.When 0.05 mL of sample(or standard) was placed into the vial containing cell-lysing reagent, thereaction was allowed to proceed for 30 s, after which 0.1 mL of themixture was transferred into the glass vial containing reaction mix andthe vial was placed into a luminometer. The readings from theluminometer were averaged over 10 s for each sample. Each data point(bars in Figs 3, 4, 6) represents the average of four to six wells, where allwells received identical treatment. Data are presented7SD. Statisticalsigni¢cance was calculated using the two-tailed Student’s t test.

Immunocytochemistry Keratinocytes were grown in glass two-chamber slides (45091, USP 37, Lab-Tek, Fisher Scienti¢c) in 0.07 mMCa2þ , as described previously. Upon reaching con£uency the cells wereswitched to high-Ca2þ medium for various times, ¢xed in 4% neutralbu¡ered formaldehyde, and stored until labeling. After repeated washingwith PBS, F-actin was labeled with tagged phalloidin (A-12380, AlexaFluor 568 phalloidin, Molecular Probes, Eugene, OR). Slides weremounted and visualized on either the Leica TCS-SP confocal microscopeor on a Zeiss Axioscope, equipped for epi£uorescence.

Electron microscopy Actin was examined by electron microscopy(Taunton et al, 2000). Brie£y, cultured normal or HHD keratinocytes werepermeabilized for 5 min in a phalloidin/bu¡er combination thatpreferentially stains actin and lyses noncytoskeletal components,consisting of (in mM): 10 MES, pH 6.1; 138 KCl; 3 Mg2Cl; 2 EGTA; 0.1%Triton X-100, and 1 mg/mL phalloidin. Cells were ¢xed for 10 min in alysine ¢xative solution consisting of 50 mM lysine, 3% gluteraldehyde,and 0.05 M cacodylate, pH 7.0, and then subjected to an additional¢xation for 20 min in a gluteraldehyde ¢xative solution containing 3%glutaraldehyde and 0.1 M cacodylate bu¡er. After being rinse three timesin cacodylate bu¡er, cells were treated at 41C in the dark for 15 min with1% OsO4, 0.8% potassium ferricyanide in 0.1 M cacodylate bu¡er. Cellswere stained for 2 h with 2% aqueous uranyl acetate, again at 41C in thedark. Samples then were dehydrated in ethanol and embedded in epoxyresin. Thin sections were examined in a Zeiss 10-A electron microscopeoperated at 60 kV.

Proton-induced X-ray emission Proton-induced X-ray emission(PIXE) analysis was performed using a modi¢cation of Antolak andBench (1994). Clinically normal trunk skin biopsies from normal andHHD patients were snap-frozen in liquid propane/liquid nitrogen.Thirty-micrometer sections were transferred to a metal-free nylon foiland freeze-dried for 12 h at ^801C. Data were obtained using 3-MeVproton beams, a 0.8-mA current, and a 5-mM spot diameter, with a scansize of 300�30 mm, and binned into 10-mm segments for analysis. X-rayswere detected with a Si(Li) detector, located at an angle of 1351 withrespect to the incident beam, which subtended a solid angle of b100 msr.After PIXE analysis, the samples were counterstained with hematoxylinand eosin, and epidermal thickness was measured using a lens micrometer.Data were reduced o¡-line so that X-ray spectra from subregions could beextracted from each irradiated region. X-ray spectra were analyzed withthe PIXE spectrum-¢tting code (Antolak and Bench, 1994). Thin ¢lmcalibration standards containing Ca2þ were used to measure the e⁄ciencyof the X-ray. Each sample was measured in three separate areas. Data arepresented as the mean7SD.

RESULTS

Actin reorganization is incomplete in HHD keratino-cytes We previously have shown that baseline cytoplasmicCa2þ concentrations in HHD keratinocytes were abnormallyincreased, consistent with an impaired intracellular Ca2þ

sequestration. In fact, cytoplasmic Ca2þ in undi¡erentiatedHHD keratinocytes was as high as cytoplasmic Ca2þ con-centrations in normal di¡erentiated keratinocytes (Hu et al,2000). Because actin reorganization is stimulated by highconcentrations of cytoplasmic Ca2þ (Vasioukhin et al, 2000), wecompared actin reorganization in normal and HHD keratinocytesto examine whether the abnormally high Ca2þ concentrations inundi¡erentiated HHD keratinocytes prematurely stimulated actinreorganization. Both normal and HHD keratinocytes displayedstress ¢bers without peripheral actin organization when grownin low-Ca2þ medium (Fig 1A,C), conditions that produceundi¡erentiated keratinocytes. Normal keratinocytes reorganizedtheir actin cytoskeleton after extracellular Ca2þ was raised from0.07 to 1.2 mM (Fig 1B), forming regular honeycomb patternsat the periphery of each cell. Cytoskeletal reorganization wasincomplete, however, in HHD keratinocytes exposed to 1.2 mMCa2þ for up to 24 h (Fig 1D). Moreover, electron micrographs,comparing normal and HHD keratinocytes, revealed that normalextension of cytoskeletal elements to the plasma membrane thatoccurs in normal keratinocytes at 1.2 mM Ca2þ was stunted inHHD keratinocytes (Fig 2A,B). These studies demonstrate thatactin reorganization is impaired in HHD keratinocytes.

HHD cellular ATP is abnormal The inability of HHDkeratinocytes to reorganize cellular actin in response to raisedextracellular Ca2þ would appear to be paradoxical, becauseactin polymerization is stimulated, not impaired, by increasedcytoplasmic Ca2þ (Vasioukhin et al, 2000). Because actin reorga-nization also requires ATP consumption, we tested whether thedefective actin reorganization seen in HHD keratinocytes wasdue to altered ATP concentrations, rather than being regulated

682 ARONCHIK ETAL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

simply by increases in cytoplasmic Ca2þ . As a control, we ¢rstfound that normal proliferative keratinocytes, grown in 0.07mM Ca2þ in a precon£uent state, contain ATP concentrationsthat are approximately one-third to one-fourth of those seen ina metabolically robust type of cell, adult rat thoracic arterysmooth muscle cells (Fig 3). We compared ATP concentrationsin normal and HHD cells cultured in six-well plates. Whereasnormal and HHD keratinocyte absolute cell numbers per wellwere constant (data not shown), indicating normal HHDattachment to plastic, ATP concentrations in HHD keratinocyteswere much lower than those seen in normal keratinocytes,both at baseline cultured in 0.07 mM Ca2þ and in responseto raising extracellular Ca2þ (Fig 4A). ATP concentrationtransiently decreased after exposure to 1.2 mM Ca2þ in normalkeratinocytes, whereas HHD keratinocyte ATP concentrationsdisplayed a minimal response to raised extracellular Ca2þ

(Fig 4A). The normal keratinocyte ATP concentration decreasedrapidly after raising extracellular Ca2þ and rebounded to levelsgreater than those seen in proliferative normal keratinocytes(Fig 4A) after 48 h. In normal keratinocytes, ATPlevels decreased with even slight increases in extracellular Ca2þ

(Fig 4B), corresponding to the known sensitivity of thekeratinocyte plasma membrane Ca2þ receptor (Oda et al, 1998)and suggesting that extracellular concentrations of Ca2þ thatactivate Ca2þ signaling through this Ca2þ receptor also cause acorresponding decrease in cellular ATP. These experimentsdemonstrate that normal keratinocytes contain much greaterconcentrations of ATP than do HHD keratinocytes and suggestthat ATP may be required for normal Ca2þ -induced actinreorganization, because actin reorganization in normal keratino-cytes was accompanied by a transient decrease in ATP.

ATP concentrations are abnormal in HHD epidermis Toexamine whether lower ATP concentrations persist in HHDepidermis, we compared the phosphorous (P) content in normalversus HHD epidermis and dermis, using PIXE.Whereas P levels

in the dermis were comparable, P levels throughout the epidermiswere markedly lower in HHD epidermis (Fig 5). Thus, actinpolymerization may be impaired in HHD skin owing toinadequate intracellular ATP concentrations. These studiesdemonstrate that ATP concentrations are abnormally low inHHD epidermis and a¡ord a possible mechanism by whichactin polymerization is abnormal in HHD.

Pharmacologically decreasing cellular ATP reproducesabnormal actin reorganization Finally, we tested whetherpharmacologically lowering the ATP concentration in normalkeratinocytes would reproduce the abnormalities in actinreorganization seen in HHD keratinocytes. For these studies, wetreated normal keratinocytes with antimycin, a substance thatdecreases ATP generation from oxidative phosphorylation.Using a dose that decreased ATP levels by approximately 70%in normal keratinocytes, we found that antimycin pre-treatmentfor 90 min signi¢cantly suppressed ATP synthesis for 60 minafter antimycin was removed (Fig 6A). Cells treated with thisprotocol showed no signs of toxicity, and ATP concentrationsnormalized slowly after antimycin was removed (data notshown), demonstrating that cell viability had not beencompromised by this short depression of ATP concentration.Next, we treated normal keratinocytes with antimycin, raisedextracellular Ca2þ to 1.2 mM, and monitored actin reorga^nization. Because antimycin e¡ects were transient, and ATP levelsdecreased rapidly after raised extracellular Ca2þ , we examinedactin reorganization at the earlier time points of 15 min and 1 h.Previously, actin has been shown to reorganize in response toCa2þ within this time period (Zamansky et al, 1991).We found rapid actin reorganization in vehicle-treated, normal

keratinocytes, with disappearance of stress ¢bers and emergence ofregular honeycomb structures within 15 min after raisingextracellular Ca2þ , and formation of complete bilayeredadhesion zippers at 1 h (Fig 6B). Depletion of cellular ATP innormal keratinocytes reproduced the defects seen in untreated

Figure1. HHD actin reorganization is ab-normal. Normal and HHD keratinocytes werecultured in 0.07 mM Ca2þ (A,C) and thenswitched to 1.2 mM Ca2þ (B,D) for 24 h. Kera-tinocytes were stained for actin using labeledphalloidin (see Materials and Methods). Bothnormal and HHD keratinocytes displayed stress¢bers without peripheral actin organizationwhen grown in low-Ca2þ medium (A,C). Ac-tin in normal keratinocytes formed regularhoneycomb patterns at the periphery of eachcell after Ca2þ was raised (B). Cytoskeletal reor-ganization, however, was incomplete in HHDkeratinocytes exposed to 1.2 mM Ca2þ (D).Bar, 20 mm.

ACTIN IS ABNORMAL IN HAILEY-HAILEY DISEASE 683VOL. 121, NO. 4 OCTOBER 2003

HHD keratinocytes, as the antimycin-treated keratinocytes didnot reorganize their actin cytoskeleton normally at 15 min or 1 hafter raising extracellular Ca2þ . Cells treated with antimycincontinuously for 24 h displayed signs of toxicity (detachmentfrom the plate and retraction of ¢lopodia). Thus, we could notassay actin reorganization in antimycin-treated keratinocytes atthis time point.Together, these experiments suggest that ATP con-centrations, greater than those that occur in HHD keratino-cytes,are required for normal Ca2þ -induced actin polymerization.

DISCUSSION

In normal keratinocytes, calcium stimulates formation of adhe-rens junctions, causing both cytoskeletal reorganization andcomplexing of junctional proteins such as E-cadherin and a-cate-nin. Actin polymerization and formation of ¢lopodia are amongthe ¢rst responses to raised extracellular Ca2þ (Vasioukhin et al,2000). Actin reorganization also requires su⁄cient quantities ofATP, because experimental or pathologic ATP depletion, such asseen in tissue ischemia, results in disruption of adherens junctions

Figure 3. Normal keratinocyte ATP concentrations are approxi-mately one-fourth of those in muscle cells. Normal keratinocytes andmuscle cells (RSMC) were cultured as detailed under Materials and Meth-ods in 0.07 mM Ca2þ until 80% con£uence. Cells were collected and ATPconcentrations were measured using a ATP luminescence kit (Sigma; seeMaterials and Methods). Because keratinocytes and muscle cells do notcontain the same amount of protein per cell (data not shown), ATP con-centrations are presented as milligrams of ATP per cell, as opposed tomilligrams of ATP per sample as seen in subsequent ¢gures. Data are pre-sented as the mean7SD, n¼ 6 for each data point. The normal keratino-cyte ATP concentration was found to be approximately one-third toone-fourth of that found in metabolically robust muscle cells.

Figure 4. ATP concentrations in normal and HHD keratinocytes.(A) ATP decreases transiently in response to Ca2þ in normal keratinocytes,while remaining abnormally decreased in HHD keratinocytes. Normal(black columns) and HHD keratinocytes (gray columns) were cultured in 0.07mM Ca2þ and then exposed to 1.2 mM extracellular Ca2þ for the timeperiods indicated. ATP was measured using the methods detailed (see Ma-terials and Methods). n¼ 4�6 for each data point; �p o 0.0001, as assessedby two-tailed Student’s t test. Data are presented as the mean7SD. (B)ATP concentrations decrease in response to small increases in extracellularCa2þ . Normal human keratinocytes were cultured in 0.07 mM Ca2þ andthen exposed to varying extracellular Ca2þ concentrations for 24 h. ATPconcentrations were measured as detailed under Materials and Methods.n¼ 3 for each data point; �po0.001, as assessed by an ANOVA test. Dataare presented as the mean7SD.

Figure 2. Cell-to-cell contact is defective in HHD keratinocytes.Normal and HHD keratinocytes were cultured in 0.07 mMCa2þ and thenswitched to 1.2 mM Ca2þ . After 24 h, cells were processed for electronmicroscopy using a protocol that isolates cytoskeletal elements (see Materi-als and Methods). Cytoskeletal elements including actin (arrows) are foundin the entire area of cell-to-cell junctions in normal keratinocytes (A). Incontrast, cytoskeletal elements do not extend to the cell border in HHDkeratinocytes, leaving empty spaces between cells (B, open arrow). Double ar-rows, keratins; N, nuclei. Bar, 5 mm.

684 ARONCHIK ETAL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

(Bush et al, 2000). Prior studies of HHD pathogenesis have con-centrated on defects in Ca2þ -sensitive proteins that form the ad-herens junctions (Harada et al, 1994; Hashimoto et al, 1995; Cooleyet al, 1996; Tada and Hashimoto, 1998; Hakuno et al, 2000).Whereas adherens junction protein assembly is impaired inHHD, defective actin polymerization with abnormal stress ¢bers(Metze et al, 1996) and abnormal localization of actin ¢laments(Inohara et al, 1990) also have been reported. In this report,we show that Ca2þ -induced actin reorganization, a necessarycomponent of adherens junction formation, is impaired in HHDkeratinocytes, owing to insu⁄cient cellular ATP stores. Thus, wehave identi¢ed an additional pathogenic mechanism by whichadherens junction formation is impaired in HHD.ATP concentrations reach their lowest levels in normal kerati-

nocytes precisely when actin reorganizes, most likely becauseATP hydrolysis is coupled to actin reorganization, imposinga signi¢cant energy burden on cells. In addition, protein synth-esis (in this case coupled to raised extracellular Ca2þ ) alsorequires signi¢cant energy consumption. For example, actomyosin-ATPase comprises 18% of ATP consumption in vascular endothe-lial cells, whereas protein synthesis consumes an additional 23%of total ATP consumption (Culic et al, 1997). Because raised extra-cellular Ca2þ stimulates a cascade of energy-requiring down-stream events in keratinocytes including gene transcription,synthesis of di¡erentiation-speci¢c proteins, migration of junc-tional proteins to the periphery, and cytoskeletal reorganization,it is not surprising that cellular ATP decreases transiently in re-sponse to raised extracellular Ca2þ .

Figure 5. P levels are decreased in HHD epidermis. Clinically nor-mal control and HHD epidermis samples were biopsied from truncal skinand measured for P concentrations using PIXE analysis (see Materials andMethods). Measurement indicates micrometers from the surface of theskin. Normal skin is denoted by ¢lled circles (�), whereas HHD skin isdenoted by open squares (&). Dermal P concentrations (deeper than ap-proximately 125^150 mm) did not di¡er between normal and HHD skin,but epidermal P concentrations (25^125 mm) were markedly lower inHHD skin. Analysis conditions: 3-MeV protons, 0.8-nA current, 5-mmspot diameter. Scan size B300�30 mm. Q¼ 4.0 mC. STIM performedafter PIXE data collection with median value of 19 ions per pixel selectedas the energy thickness of each section. Data were binned into 10-mm seg-ments. n¼ 3 for each data point.

Figure 6. Decreased ATP causes defective actin polymeriza-tion in normal keratinocytes. (A) Antimycin treatment of nor-mal keratinocytes decreases ATP concentrations to those seen inHHD keratinocytes. Normal keratinocytes were grown in 0.07mM Ca2þ and then treated with 0.1 mM antimycin. After 90 min,antimycin was removed and extracellular Ca2þ was raised to 1.2mM. ATP concentrations remained depressed for up to 60 min afterantimycin was removed and extracellular Ca2þ was raised. Eachdata point represents six experiments. po0.01, as assessed by an AN-OVA test. Data are presented as the mean7SD. (B) Pre-treatmentwith antimycin inhibits Ca2 -induced actin reorganization. Normalkeratinocytes were cultured at 0.07 mM Ca2þ (A), pretreated with0.1 mM antimycin (right, C,E) or vehicle (left, B,D) for 90 min, andthen treated with 1.2 mMCa2þ (no antimycin). After 15 min of ex-posure to 1.2 mM Ca2þ , normal actin reorganization began in nor-mal control keratinocytes (B), whereas Ca2 -stimulated actinreorganization was delayed in antimycin-treated normal keratino-cytes, as stress ¢bers (closed arrows) persisted and decreased reorgani-zation at the periphery (open arrows) was noted (C). At 60 min, well-formed actin structures were noted in normal keratinocytes (D),whereas stress ¢bers and incompletely formed peripheral actin reor-ganization still persisted in the antimycin-treated keratinocytes (E).Bar, 20 mm.

ACTIN IS ABNORMAL IN HAILEY-HAILEY DISEASE 685VOL. 121, NO. 4 OCTOBER 2003

Because we know that HHD keratinocytes su¡er from abnor-mally high cytoplasmic Ca2þ concentrations (Hu et al, 2000), wehypothesize that defective ATP synthesis might result from mito-chondrial Ca2þ overload. Although small, transient increases inmitochondrial Ca2þ concentration enhance ATP synthesis, excessmitochondrial Ca2þ accumulation causes Ca2þ overload, whichresults in uncoupling of oxidative phosphorylation and decreasedATP synthesis (reviewed in Duchen, 2000). Aging results in in-creased mitochondrial sensitivity to Ca2þ overload (Jahangir etal, 2001), consistent with the onset of HHD in adulthood. ATPdepletion also may result from increased consumption of ATPby other cellular ATPases attempting to clear excess cytoplasmicCa2þ or Naþ (Chinopoulos et al, 2000). Finally, decreased ATPmay further exacerbate cellular Ca2þ overload by inhibitingATP-requiring Ca2þ ATPases, such as the ATP2A2 and plasmamembrane Ca2þ ATPase (Wieser and Krumschnabel, 2001),which normally would compensate for decreased ATP2C1 func-tion (Hu et al, 2000). If mitochondrial Ca2þ overload underliesdecreased ATP production and defective actin reorganizationin HHD, agents that ameliorate mitochondrial Ca2þ overloadmight improve the skin lesions seen in this disease. Thus, eventhough there is no treatment that speci¢cally increases ATP inkeratinocytes, these experiments may suggest novel therapeuticapproaches to the treatment of HHD.Pharmacologic inhibition of ATP synthesis using antimycin

mimicked the defective actin polymerization seen in HHD kera-tinocytes. However, additional factors required for actin reorga-nization, such as activation of Rho or Rac molecules, may bedefective in HHD keratinocytes (Braga et al, 1997; Malliri et al,1998). Future studies will be necessary to determine the molecularbasis of defective actin reorganization in HHD. Defective actinreorganization may be responsible for the decreased cell-to-celladhesion found in HHD, or conversely, decreased cell-to-cell ad-hesion may be responsible for defective actin reorganization. It isclear that decreased cellular ATP causes defective actin reorganiza-tion, because impaired actin reorganization in response to a phy-siologic stimulus, Ca2þ , is seen in Hailey-Hailey keratinocytes,correlating with decreased ATP concentrations. More impor-tantly, impaired actin reorganization can be reproduced by phar-macologically decreasing keratinocyte ATP. Whether decreasedcell-to-cell adhesion is a consequence of impaired actin reorgani-zation or an intermediate step leading to impaired actin reorgani-zation cannot be deduced from our data.Although PIXE is exquisitely sensitive to changes in elemental

ion concentration and very precise in its ability to localize ionconcentrations within tissue, it is not able to distinguish amongthe varieties of P species in tissue, such as inorganic P, ATP, ADP,phosphocreatine, and P linked to other molecules. Thus, our ex-periment correlating decreased ATP seen in vitrowith decreased Plevels seen in vivo must be interpreted with some caution. Pre-vious studies, however, have shown that decreased ATP concen-trations correlate with decreases in total P (Park et al, 1988;Masson et al, 1993), suggesting that total P may be used as a surro-gate marker for ATP. Moreover, conditions speci¢cally associatedwith ATP depletion demonstrate decreased total P concentrationswhen measured by PIXE or related methods (Rehwald et al,2002; Kisters et al, 2001), again demonstrating that decreasedATP causes a corresponding decrease in total P. Because the PIXEdata are used speci¢cally to localize changes in P in normal versusHHD epidermis, we believe that the decrease in total P measuredin HHD epidermis is an appropriate surrogate marker fordecreased ATP.

The authors thank DrJ. Karliner for assistance with cell culture of smooth muscle cells.This work was supported by NIH AR44341 (T.M.) and the San FranciscoVeteran’sA¡airs Hospital and was partially performed under the auspices of the U.S. Depart-ment of Energy at the University of California Lawrence Livermore National Labora-tory, under ContractW-7405-Eng-48.

REFERENCES

Antebi A, Fink GR: The yeast Ca(2þ )-ATPase homologue, PMR1, is required fornormal Golgi function and localizes in a novel Golgi-like distribution. MolBiol Cell 3:633^654, 1992

Antolak A, Bench G: PIXEF: The Livermore PIXE spectrum analysis package. NuclInstr Methods B90:596^601, 1994

Behne MJ,Tu Chia-Ling, Aronchik I: Human keratinocyte ATP2C1 localizes to theGolgi in keratinocytes and controls Golgi Ca2+ stores. J Invest Dermatol (inpress)

Braga VM, Machesky LM, Hall A, Hotchin NA: The small GTPases Rho and Racare required for the establishment of cadherin-dependent cell-cell contacts.J Cell Biol 137:1421^1431, 1997

Bush KT,TsukamotoT, Nigam SK: Selective degradation of E-cadherin and dissolu-tion of E-cadherin-catenin complexes in epithelial ischemia. AmJ Physiol RenalPhysiol 278:F847^F852, 2000

Chinopoulos C,Tretter L, Rozsa A, Adam-Vizi V: Exacerbated responses to oxidativestress by an Na(þ ) load in isolated nerve terminals. the role of ATP depletionand rise of [Ca(2þ )](i). J Neurosci 20:2094^2103, 2000

Cooley JE, Briggaman RA, Cronce DJ, Banes AJ, O’Keefe EJ: Hailey-Hailey diseasekeratinocytes: Normal assembly of cell-cell junctions in vitro. J Invest Dermatol107:877^881, 1996

Culic O, Gruwel ML, Schrader J: Energy turnover of vascular endothelial cells. AmJ Physiol 273:C205^C213, 1997

Duchen MR: Mitochondria and Ca2þ in cell physiology and pathophysiology. CellCalcium 28:339^348, 2000

Hailey H, Hailey H: Familial benign chronic pemphigus. Arch Dermatol Syphilol39:679^685, 1939

Hakuno M, Shimizu H, Akiyama M, Amagai M,Wahl JK,Wheelock MJ, NishikawaT: Dissociation of intra- and extracellular domains of desmosomal cadherinsand E-cadherin in Hailey-Hailey disease and Darier’s disease. Br J Dermatol142:702^711, 2000

Harada M, Hashimoto K, Fujiwara K: Immunohistochemical distribution of CD44and desmoplakin I & II in Hailey-Hailey’s disease and Darier’s disease. J Derma-tol 21:389^393, 1994

Hashimoto K, Fujiwara K, Harada M, Setoyama M, Eto H: Junctional proteins ofkeratinocytes in Grover’s disease, Hailey-Hailey’s disease and Darier’s disease.J Dermatol 22:159^170, 1995

Hu Z, Bonifas JM, Beech J, et al: Mutations in ATP2C1, encoding a calcium pump,cause Hailey-Hailey disease. Nat Genet 24:61^65, 2000

Inohara S, Tatsumi Y, Tanaka Y, Sagami S: Immunohistochemical localization ofdesmosomal and cytoskeletal proteins in the epidermis of healthy individualsand patients with Hailey-Hailey’s disease. Acta Dermatol Venereol 70:239^241,1990

Jahangir A, Ozcan C, Holmuhamedov EL,Terzic A: Increased calcium vulnerabilityof senescent cardiac mitochondria: Protective role for a mitochondrial potas-sium channel opener. Mech Ageing Dev 122:1073^1086, 2001

Kisters K, Krefting RE, Dietl HK: Decreased intracellular phosphate and magne-sium concentrations in vascular smooth muscle cells from spontaneously hy-pertensive rats. Int J Angiol 10:50^52, 2001

Malliri A, Symons M, Hennigan RF, Hurlstone AF, Lamb RF,Wheeler T, OzanneBW: The transcription factor AP-1 is required for EGF-induced activation ofrho-like GTPases, cytoskeletal rearrangements, motility, and in vitro invasionof A431 cells. J Cell Biol 143:1087^1099, 1998

Masson S, Desmoulin F, Sciaky M, Cozzone PJ: Catabolism of adenine nucleo-tides and its relation with intracellular phosphorylated metabolite concentra-tion during ethanol oxidation in perfused rat liver. Biochemistry 32:1025^1031,1993

Metze D, Hamm H, Schorat A, Luger T: Involvement of the adherens junction-actin¢lament system in acantholytic dyskeratosis of Hailey-Hailey disease: A histo-logical, ultrastructural, and histochemical study of lesional and non-lesionalskin. J Cutan Pathol 23:211^222, 1996

Oda Y, Tu CL, Pillai S, Bikle DD: The calcium sensing receptor and its alterna-tively spliced form in keratinocyte di¡erentiation. J Biol Chem 273:23344^23352, 1998

Park JH, Brown RL, Park CR, Cohn M, Chance B: Energy metabolism of the un-trained muscle of elite runners as observed by 31P magnetic resonance spectro-scopy: Evidence suggesting a genetic endowment for endurance exercise. ProcNatl Acad Sci USA 85:8780^8784, 1988

Pinton P, Pozzan T, Rizzuto R: The Golgi apparatus is an inositol 1,4,5-trispho-sphate-sensitive Ca2þ store, with functional properties distinct from those ofthe endoplasmic reticulum. EMBO J 17:5298^5308, 1998

RehwaldWG, Fieno DS, Chen EL, Kim RJ, Judd RM: Myocardial magnetic reso-nance imaging contrast agent concentrations after reversible and irreversibleischemic injury. Circulation 105:224^229, 2002

Sudbrak R, Brown J, Dobson-Stone C, et al: Hailey-Hailey disease is caused by mu-tations in ATP2C1 encoding a novel Ca(2þ ) pump. Hum Mol Genet 9:1131^1140, 2000

Tada J, Hashimoto K: Ultrastructural localization of cell junctional components(desmoglein, plakoglobin, E-cadherin, and beta-catenin) in Hailey-Haileydisease, Darier’s disease, and pemphigus vulgaris. J Cutan Pathol 25:106^115,1998

686 ARONCHIK ETAL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Taunton J, Rowning BA, Coughlin ML,Wu M, Moon RT, Mitchison TJ, LarabellCA: Actin-dependent propulsion of endosomes and lysosomes by recruitmentof N-WASP. J Cell Biol 148:519^530, 2000

Van Baelen K, Vanoevelen J, Missiaen L, Raeymaekers L, Wuytack F: The GolgiPMR1 P-type ATPase of Caenorhabditis elegans: Identi¢cation of the geneand demonstration of calcium and manganese transport. J Biol Chem276:10683^10691, 2001

VasioukhinV, Bauer C,Yin M, Fuchs E: Directed actin polymerization is the drivingforce for epithelial cell-cell adhesion. Cell 100:209^219, 2000

Wieser W, Krumschnabel G: Hierarchies of ATP-consuming processes: Direct com-pared with indirect measurements, and comparative aspects. Biochem J 355:389^395, 2001

Zamansky GB, Nguyen U, Chou IN: An immuno£uorescence study of the calcium-induced coordinated reorganization of micro¢laments, keratin intermediate ¢-laments, and microtubules in cultured human epidermal keratinocytes. J InvestDermatol 97:985^994, 1991

ACTIN IS ABNORMAL IN HAILEY-HAILEY DISEASE 687VOL. 121, NO. 4 OCTOBER 2003