View metadata, citation and similar papers at core.ac.uk brought to you by CORE To cite this article: JID 122:1070–1077, 2004 provided by Elsevier - Publisher Connector

Published by the

ology Progress in Dermatology

Editor: Alan N. Moshell, M.D.

Staphylococcus aureus exfoliative : How they cause disease. Lisa R.W. Plano, M.D., Ph.D. Departments of Pediatrics and Microbiology & Immunology University of Miami School of Medicine, Miami, Florida

Abbreviations: cell surface molecules associated with adhesion and BI- bullous impetigo multiple antibiotic resistances including methicillin and ET- exfoliative toxins vancomycin resistance (Centers for Disease Control and EDIN- epidermal cell differentiation inhibitor Prevention, 1997; 2000a; 2000b), all contributing to the ETA- exfoliative A (epidermolysisn A, exfoliatin A) pathogenicity of these organisms. A minimum of 34 ETB- exfoliative toxin B (epidermolysisn B, exfoliatin B) different extracellular proteins are produced by S. ETD- exfoliative toxin D (epidermolysisn D, exfoliatin D) aureus, and many of these have defined roles in the PF- pemphigus foliaceus pathogenesis of their associated diseases (Iandolo, SSSS- Staphylococcal scalded skin syndrome, (pemphi- 1989). Infectious conditions caused by these organisms gus neonatorum, dermatitis exfoliativa neonatorum, can be divided into three major categories; (i) superficial Ritter’s disease) skin infections, skin abcesses and wound infections TEN- toxic epidermal necrolysis including bullous impetigo (BI) and furuncles, (ii) systemic or infections of deep seeded tissues including osteomyelitis, endocarditis, pneumonia and sepsis, and Introduction (iii) conditions caused by intoxication with one of the General Microbiology: Staphylococci are hardy excreted toxins. Among the conditions caused by intoxi- Gram-positive cocci found as bacterial pathogens or cation with an are toxic shock syndrome caused commensal organisms in both humans and animals. by toxic shock syndrome toxin (TSST-1) (Bloster-Hauta- These organisms are resistant to harsh conditions and maa et al, 1986), staphylococcal food poisoning caused can be recovered from non-physiologic environments up by the staphylococcal (SE) (Bergdoll, 1972; to months after inoculation. They grow easily under Bergdoll, 1983; Bergdoll et al, 1981) and staphylococcal numerous conditions and are classified based on coagu- scalded skin syndrome (SSSS) mediated by the exfolia- lase activity. Coagulase positive strains are classified as tive toxins (ETs), primarily ETA and ETB (Arbuthnott et . Approximately 35% of the al, 1972). This discussion will focus on the exfoliative general population are commensal nasal carriers and toxins of S. aureus and the mechanism by which they most newborns will be colonized within the first week of cause disease in humans. life (Dancer and Noble, 1991). Of the staphylococci, S. aureus is the most significant pathogen causing both General characteristics of the exfoliative toxins. human and animal diseases. S. aureus causes a wide There are two major biologically and serologically variety of infections acquired in both the community and distinct S. aureus ET isoforms that are primarily respon- hospital settings. It is the most common cause of sible for the skin manifestations of SSSS and BI in pyogenic infections of the skin. A variety of predisposing humans (Wiley and Rogolsky, 1977) ETA and ETB. Five factors may lead to more serious infections such as percent of clinical S. aureus isolates produce either ETA, conjunctivitis, pneumonia, septicemia, osteomyelitis, ETB or both toxins (Piemont et al, 1984) and there are septic arthritis, empyema, meningitis, pericarditis or reported geographic differences in the prevalence of the endocarditis. Diseases caused by S. aureus can be the particular toxins that are expressed. In North America, result of direct tissue invasion or due to the action of a Europe and Africa ETA is the predominant ET (Adesiyun variety of released by the bacteria (Iandolo, et al, 1991; Ladhani, 2001), in contrast ETB has been 1989; Marrack and Kappler, 1990). S. aureus strains reported to be more common in Japan (Kondo et al, capable of causing disease express a wide variety of 1975). These toxins exhibit exquisite species specificity, virulence factors including exported toxins (exotoxins), being active in humans, mice, monkeys and hamsters

The production of this issue of Progress in Dermatology has been underwritten by Galderma Laboratories, L.P. ©2003 Dermatology Foundation, 1560 Sherman Avenue, Evanston, Illinois 60201

1070 122 : 5 MAY 2004 STAPHYLOCOCCUS AUREUS EXFOLIATIVE TOXINS 1071

but not in rats, rabbits, dogs, hedgehogs, voles, guinea exfoliative toxins made by S. hyicus, (ShETA, ShETB pigs, chickens or frogs (Bailey et al, 1995; Elias et al, and ShETC) have also been identified and partially 1975; Ladhani et al, 1999). characterized (Anderson, 1998). These toxins have been associated with comparable skin diseases in different Both ETA and ETB have been cloned and their resul- animal populations but are likewise not associated with tant protein products characterized (Lee et al, 1987; disease in humans. O’Toole and Foster, 1987). The primary human active ETs, ETA and ETB, share significant amino acid identity The ETs, like most staphylococcal proteins associat- with each other and have similar biophysical properties ed with virulence, are encoded on mobile genetic (Bailey et al, 1980; Lee et al, 1987; O’Toole and Foster, elements (Novick, 2003) allowing for horizontal transfer 1987). A comparison of these toxins is shown in Table 1. of these virulence factors. The gene encoding ETA is ETA is a very stable protein that is resistant to extreme located on the bacterial chromosome within a 43 kb heat while ETB is heat sensitive. Both ETA and ETB temperate phage, ␾ETA (Yamaguchi et al, 2000), that share amino acid identity with staphylococcal V8 was recently shown to be capable of transducing a toxin protease (glutamyl endopeptidase I), a well character- negative S. aureus strain to positive. ETB is encoded on ized general serine protease of S. aureus (26% with ETA approximately 38.2-38.5 kb plasmids (Lee et al, 1987; and 31% with ETB). Most significantly, this identity Yamaguchi et al, 2001). ETD was identified within a includes the key amino acid residues of the V8 protease pathogenicity island which also encodes EDIN-B active site, serine-histidine-aspartate. This catalytic triad (Yamaguchi et al, 2002). Pathogenicity islands are is highly conserved and forms the active site of trypsin- varied sized segments of potentially mobile DNA that like serine proteases (Dancer et al, 1990). The X-ray encode virulence associated genes (Kaper and Hacker, crystal structures of both of these toxins have been 1999). These islands have been described for numerous determined (Cavarelli et al, 1997; Papageorgiou et al, types of bacteria and at least three different types of 1999; Vath et al, 1999; Vath et al, 1997). These three- pathogenicity islands have been reported in staphylo- dimensional structures show marked similarity with each coccal species (Kuroda et al, 2001). Also like most other and V8 protease demonstrating that the ETs are staphylococcal virulence factors, the ETs are regulated members of the trypsin-like serine protease family, but by the staphylococcal accessory gene regulator locus, do have some significant differences at the amino- and agr locus (Ji et al, 1995; Novick, 2003). This locus carboxyl-termini. encodes a two-component regulatory system that coordi- nates the expression of specific virulence factors not Recently a new potential human active exfoliative required for bacterial viability, including the exotoxins. toxin, ETD, was identified from an isolate of S. aureus The ETs have a strong association with a particular agr taken from a wound site. It was initially identified as an group, agr group IV, (Jarraud et al, 2000) but the clinical open reading frame which had significant homology to significance of this association remains to be estab- the known ETs (Yamaguchi et al, 2002) but has also lished. been associated with a clinical isolate from a patient with BI (Yamaguchi et al, 2002; Yamaguchi et al, 2002). ETD expressed as a recombinant protein product demon- Clincal conditions caused by the exfoliative strated specific protease activity in the neonatal mouse toxin producing S. aureus. model and exhibits cleavage of recombinant mouse or Staphylococcal scalded skin syndrome and human Dsg1 which is indistinguishable from that of ETA Bullous impetigo. Staphylococcal scalded skin by Western blot analysis (Hanakawa et al, 2002). It was syndrome (SSSS) is an exfoliative dermatitis mediated further shown to be serologically distinct from ETA and by infection with S. aureus capable of producing one or ETB. A fourth potential human active ET produced from both of the exfoliative toxins ETA or ETB. This condition S. aureus, ETC, was isolated from a horse infection and is well described and discussed in several recent was shown to have activity in the neonatal mouse model reviews (Ladhani, 2001; Ladhani and Evans, 1998; (Sato et al, 1994); however, it has not been associated Prevost et al, 2003). It is typically a condition of infants with human disease. The clinical significance of either and young children, although it may also affect adults ETD or ETC has not been established. Other similar (Cribier et al, 1994; Gemmell, 1995). SSSS is character- ized by the separation of extended areas of the upper epidermis specifically at the level Table 1. Comparison of human active exfoliative toxins of the stratum granulosum by disruption of the after specific cleavage of Toxin Size Molecular weight Accession # desmoglein 1 (Amagai et al, 2000; Amagai ETA 242 aa 26.9 kDa PO9331 et al, 2002). The condition can initiate with ETB 246 aa 27.3 kDa AAA26628 fever and erythema, followed by a positive Nikolsky’s sign and progress to the forma- ETD 244 aa* 27.2 kDa AB036767 tion of large bullae and exfoliation of exten- sive areas of the upper epidermis of the skin. In SSSS this exfoliation occurs at a site Percent Amino Acid Identity that is separate from that of the infection. ETA ETB ETD The infecting bacteria produce the exotoxin ETA 100 40 40 and release it into the blood stream where it makes it’s way to the skin and causes exfoli- ETB 100 59 ation/epidermolysis. Infections that result in ETD 100 SSSS may be as serious as pneumonia or sepsis in adults, or as trivial as otitis media *Predicted number of amino acids after removal of signal sequence. or uncomplicated conjunctivitis in children. 1072 PLANO THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

In infants and children the syndrome is usually associat- In addition to the associations with the leukotoxins ed with a trivial infective focus, 3% mortality with appro- LukE-LukD and P-V luekocidin, and the multidrug resis- priate antibiotic therapy, and the patients are rarely tance plasmid pKS41 (Gravet et al, 2001; Koning et al, bacteremic (Gemmell, 1995). Mortality in children is 2003), there are further reports linking the ETs with other usually associated with secondary infections and other virulence factors. ETD is also clonally linked to EDIN-B complications from the loss of the upper epidermis of the (C3stau) (Yamaguchi et al, 2002) and ETB has been skin or failure to appropriately treat the underlying infec- linked to EDIN-C (Yamaguchi et al, 2001). There are tion. However, hospital nursery outbreaks continue to be three EDIN isoforms produced by S. aureus. EDIN-A reported and are becoming increasingly associated with was the first discovered as an inhibitor of morphological drug resistant strains of S. aureus (MRSA) (Cribier et al, differentiation of epidermal keratinocytes in vitro and 1994; Ito et al, 2002; Mackenzie et al, 1995). thus called epidermal cell differentiation inhibitor (EDIN) (Sugai et al, 1992). It is now accepted that the EDINs In contrast, the syndrome in adults is usually associ- belong to the C3 family of bacterial Rho-specific mono- ated with a chronic underlying medical condition. It is ADP-ribosyltransferases that specifically modify eukary- most often associated with old age, an immunocompro- otic small GTP-binding proteins belonging to the Rho mised state, renal deficiencies or diabetes mellitus family (Sugai et al, 1992). Rho GTPases are central (Cribier et al, 1994) but has been described in healthy regulators of the eukaryotic actin cytoskeleton and their adults (Oyake et al, 2001). There is often bacteremia and inactivation blocks important cellular functions including a greater than 50% mortality rate even with appropriate differentiation, chemotaxis, phagocytosis and oxidative antibiotic therapy (Cribier et al, 1994). The increased burst (Aktories et al, 2000). The clinical significance of mortality rate is likely heavily contributed to by the under- the association of the EDINs with the ETs and the role of lying medical condition that might have predisposed the EDINs in the pathogenesis of S. aureus are not clear. adult to infection with S. aureus or to the affects of the Likewise the role of ETD is unclear as it has only been exfoliative toxin itself. When SSSS has been reported in rarely isolated (1 in 88 isolates) from S. aureus obtained healthy adults there is usually insignificant morbidity or from impetigo patients (Yamaguchi et al, 2002; mortality (Oyake et al, 2001). Yamaguchi et al, 2002), but warrants further investigation. In SSSS the bullous fluid is characteristically a sterile The characteristic separation of the epidermis at the transudate. There is no inflammatory cell infiltration stratum granulosum seen in SSSS is also demonstrated associated with the bullous lesions or exfoliated areas in the neonatal mouse model (Melish and Glasgow, unless or until there is a secondary infection at that 1970) Figure 1. Animal models of SSSS show identical location. This is in contrast to bullous impetigo, which is response to either of the purified ETs. This cleavage is a localized skin infection with exfoliative toxin producing associated with a disruption of the desmosomes, with S. aureus. In BI there are bullous lesions with purulent the surrounding cells remaining intact. The mechanism exudates, local signs of inflammation, inflammatory cell by which these toxins cause exfoliation is now known to infiltrates and necrosis, and S. aureus can usually be involve cleavage of desmoglein 1 (Dsg1), a desmosomal isolated from the lesions. Recently Gravet and cowork- protein member of the cadherin family of cell adhesion ers showed that ET positive strains of S. aureus isolated molecules. This is accomplished by a unique serine from patients with impetigo were associated with the protease activity of the exfoliative toxins acting at a leukotoxin, LukE-LukD, in 78% of cases studied (Gravet specific extracellular site on Dsg1, as recently described et al, 2001). These pore forming toxins may contribute to (Amagai et al, 2000). Dsg1 is the predominant cell-cell the local inflammation and necrosis seen in BI, but their adhesion molecule in the upper epidermis where the role in the condition has not been clearly established. To separation occurs. Dsg1 amino acid sequences are date, no association has been recognized with S. aureus highly conserved among mammalian species and share strains isolated from clinical cases of SSSS and the a high percentage of identity (Koch and Franke, 1994; leukotoxin, LukE-LukD. Koning and coworkers have also Nilles et al, 1991). In other studies ETA has been shown shown an association of ETB producing stains with P-V to cleave ␣- and ␤-melanocyte stimulating hormones , a toxin that is linked to necrotic pneumonias after a lengthy incubation period (Rago et al, 2000); (Gillet et al, 2002), and the multidrug resistant plasmid however, it is unclear how this might result in or pSK41 (Koning et al, 2003). They report that this combi- contribute to SSSS or BI. Also, other investigators study- nation of virulence factors contributes to the severity of ing ETB have purified a distinct 20 kDa protease from a nonbullous impetigo. However no association of these supernatant of preparations of neonatal mouse epider- factors with BI or SSSS has been demonstrated. The mis treated with ETB which was able to reproduce role of the exfoliative toxins in the pathogenesis of epidermolysis similar to that of SSSS when injected into bullous impetigo clearly appears to be the enhancement neonatal mice (Ninomiya et al, 2000). These results of local spread of the organism in a relatively protected would suggest the presence of an additional target for plane, just under the stratum corneum in the upper the ETB versus ETA, as Dsg1 has been shown to be epidermis, thus allowing for a successful infection by the organism. The presence of additional toxins such as the or other potential virulence factors may augment this process and contribute to a successful infection by thwarting the host’s immune response. In contrast, the role of these toxins to promote bacterial survival and enhance their pathogenicity in SSSS is not so clear and warrants additional research. For a more extensive review of SSSS see references (Ladhani and Evans, 1998; Ladhani et al, 1999). Figure 1. Panel A: normal newborn Balb/c mouse skin; Panel B, Newborn Balb/c mouse injected with purified ETA. 122 : 5 MAY 2004 STAPHYLOCOCCUS AUREUS EXFOLIATIVE TOXINS 1073

cleaved by both ETB and ETA (Amagai et al, 2000; given repeated injections of toxin so that the maximal Amagai et al, 2002). level of toxin was maintained for a period of time (4 hours). Exfoliation occurred in these mice confirming the Relative protection of healthy adults against the presence and accessibility of the target for ETA in this effects of the toxins. SSSS and BI are primarily population and the ability of the toxins to affect adults diseases of infants and children and SSSS only rarely under certain conditions. It was concluded that whereas develops in adults (Cribier et al, 1994; Gemmell, 1995). the adaptive immune response is not needed for protec- In the mouse model of SSSS it was originally shown that tion of adult mice from SSSS, efficient clearance of the newborn mice would exfoliate in response to injection toxin from the bloodstream is a critical factor (Plano et al, with ET producing S. aureus, but that mice greater than 2001). The acuteness with which the mice were no 7 days old would not exfoliate. A time course of suscep- longer susceptible to the affects of the toxin suggests tibility of newborn mice to these toxins was recently that in addition to the development of the system generated and showed that until day 7 of life, mice were required to clear the toxin (probably renal), other events, sensitive to the affects of a single injection of a low dose perhaps differential expression of unidentified compo- of the toxin. However, after 8 days of life there is a nents in the skin, may be contributing to the protection of dramatic decrease in response to these exotoxins such adults. that these mice failed to respond to even 10 times the dose that caused exfoliation in the population 24 hours prior (Plano et al, 2001). The mechanism of the relative Mechanisms of action: protection in healthy adults who are exposed to the ETs Exfoliative toxins as potential . is not completely understood but likely is a multifactorial There was an ongoing debate as to whether the exfolia- process. It is known that in mice the adaptive immune tive toxins function as superantigens that contribute to response matures within the same time frame as they the pathogenesis of SSSS, and there are several publi- developed resistance to ETA and ETB (Adkins, 1999). It cations addressing this question (Bailey et al, 1995; has been shown that over time healthy humans will Monday et al, 1999; Plano et al, 2000; Vath et al, 1997). produce antibodies that recognize the exfoliative toxins. Bacterial superantigens are a family of proteins able to In one study, more than 50% of persons over the age of bind simultaneously to the major histocompatibility 10 years had antibodies that recognize ETA (Melish et complex class II and to the T-cell receptor (TCR), result- al, 1981), however it was undetermined whether these ing in stimulation of a large number of T-cells expressing were specific antibodies raised against ETA or against a specific V␤ subsets of the T-cell repertoire (Marrack and cross reacting material. The fact that only about 5% of all Kappler, 1990). This stimulation then leads to the clinical S. aureus isolates are capable of producing ETA release of numerous cytokines and subsequent systemic argues against these being specific antibodies. In a symptoms. Classic mediated diseases, recent study conducted in our labs we investigated the such as toxic shock syndromes, are associated with role of the adaptive immune response in the protection of erythematous rash, hypotension and multiorgan system adult mice and concluded that it does not play a primary failure. Although SSSS can include erythema, fever and role in defense of 10 day old mice against the effects of purulent infective foci, these are not always present. The the ETs (Plano et al, 2001). The role of the adaptive controversy as to whether the ETs were superantigens immune response was investigated using two immune stemmed from the fact that although for over 30 years it compromised mouse populations: (i) mice deficient in T has been known that the ETs are responsible for exfoli- cells (BALB/c mice thymectomized as newborns and ation/epidermolysis (Melish, 1971) in SSSS, the precise allowed to grow to adulthood) and (ii) mice deficient in mechanism by which this occurs was not known until both T and B cell (CB-17 SCID mice) and therefore recently (Amagai et al, 2000). Although stimulation of unable to produce antibodies. Neither of these immun- specific V␤ populations of T-cells has been demonstrat- odeficient populations of mice was affected by exposure ed (Rago et al, 2000), the SSSS generally lacks the to the exotoxin. Furthermore, attempts to protect severe systemic symptoms associated with a superanti- newborn mice from the effects of the ETs with intra- gen mediated syndrome. It is unlikely that this activity venous injections of normal adult mouse spleen cells plays a major role in the pathogenesis of SSSS but may either prior to or at the time of exposure to the ETs failed. contribute to other disease processes. It is now general- From these data it was concluded that although clearly ly accepted that in SSSS and BI the exfoliative toxins the adaptive immune response can play a role in protect- function as unique serine proteases that specifically ing against the exotoxins once specific antibodies are cleave desmoglein 1. made, it does not play a primary role in the protection of Exfoliative toxins are unique serine proteases. adult mice from the effects of these toxins. For many years the mechanism of action of the exfolia- Rapid clearance of the ETs from the serum was tive toxins was unclear, however there was mounting however shown to play a key role in protection of adult indirect evidence that ETA and ETB were serine proteas- mice (Plano et al, 2001). Serum clearance from 1 day es (Dancer et al, 1990). Both ETA and ETB share amino and 10 day old mice were compared after a single injec- acid identity with staphylococcal V8 protease (glutamyl tion with ETA. While the maximal level of toxin found in endopeptidase I), (26% with ETA and 31% with ETB) the serum was similar for both ages, levels peaked earli- and most significantly, this identity includes residues of er and were cleared earlier in the 10 day old mice, which the V8 protease serine-histidine-aspartate catalytic triad. showed no signs of response to the toxin, as compared This catalytic triad is a signature sequence common to to the one day old mice which began to exfoliate after 2 trypsin-like serine proteases. Mutant forms of ETA that hours of exposure to the toxin. Therefore, ETA was not contain amino acid substitutions at any one of the cleared as efficiently in the one-day old mice as in the 10 residues of the catalytic triad (S195, H72, D120, ETA day old mice. To address whether a sustained high level numbering) have no activity as exfoliants in the animal of ETA would result in exfoliation, 10 day old mice were model of SSSS (Prevost et al, 1992; Prevost et al, 1991). 1074 PLANO THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Both ETA and ETB have esterase activity versus a express cloned mouse Dsg1 or Dsg3. They treated the synthetic substrate, a common property of serine transfected cells with ETA and showed by Western blot proteases (Bailey and Redpath, 1992). In addition, as analysis, that only Dsg1 was cleaved. To confirm these stated above, after prolonged incubation times, ETA and findings in vivo they took extracts of skin from neonatal ETB are able to cleave ␣ and ␤-melanocyte-stimulating mice that developed blisters after injection with ETA and hormone (Rago et al, 2000). Until the recent report that showed by Western blot analysis using antibodies ETA cleaves desmoglein 1, the strongest evidence that against Dsg1, Dsg3 and epithelial cadherin that only ETA and ETB are proteases came from the X-ray crystal Dsg1 was cleaved. These data confirmed that treatment structures of these proteins. (Cavarelli et al, 1997; with ETA resulted in Dsg1 cleavage but did not demon- Papageorgiou et al, 1999; Vath et al, 1999; Vath et al, strate direct cleavage by the ET. To confirm that ETA 1997). Comparisons of the X-ray crystal structures of was directly responsible for Dsg1 cleavage they isolated both of these toxins as well as an inactive ETA with a the extracellular domains of both mouse and human mutation at the active site serine (Ser195Ala variant) Dsg1, and Dsg3 as a control, using a eukaryotic expres- with that of glutamyl endopeptidase I clearly demonstrate sion system to produce these protein products. Using that the ETs are members of the trypsin-like serine these purified proteins they successfully demonstrated protease family. The structures also show that they have that ETA directly cleaved the extracellular domain of significant differences at the amino- and carboxy- termi- both mouse and human Dsg1 but not Dsg3 in a dose ni and loop regions when compared to other like serine dependent manner. In similar experiments these investi- proteases (Papageorgiou et al, 1999). The catalytic site gators have now demonstrated that ETB (Amagai et al, of both ETs also are very similar to those of glutamate- 2002) and ETD (Hanakawa et al, 2002) also specifically specific serine proteases (V8) suggesting a similar cleave Dsg1 but not Dsg3 of mouse and humans. They catalytic mechanism. However, differences in the struc- have also identified a single cleavage site within the tures show that a part of the catalytic site is in an inactive extracellular segment of Dsg1 at glutamate 381 (E381) conformation in crystalized ETA, but in an active confor- between extracellular cadherin domains 3 and 4. Previ- mation in ETB. This suggests that some action is neces- ous studies have demonstrated that the first three extra- sary to put ETA into an active confirmation. It has been cellular cadherin domains of desmoglein are needed for postulated that this could be accomplished by the calcium dependent heterophilic adhesion to the desmo- binding of ETA to a specific target possibly involving the collins (Chitaev and Troyanovsky, 1997). Cleavage at unique amino-terminus (Papageorgiou et al, 1999). this site, E381, would release the portion of the Dsg1 Analysis of the structure of the inactive Ser195Ala involved in this heterophilic adhesion disrupting the variant showed no significant perturbation at the active . In additional experiments using immunoflu- site, suggesting that its loss of biological activity as an orescence colocalization and coimmunoprecipitation, exfoliant is attributed solely to disruption of the catalytic these investigators have also demonstrated that an serine residue (Papageorgiou et al, 1999). exfoliation inactive mutated ET specifically binds to Dsg1 (Hanakawa et al, 2002). These data together clearly Exfoliative toxins specifically cleave the cadherin demonstrate that the mechanism of action of these protein Dsg1 but not the closely related Dsg3. The toxins is specific recognition and proteolytic cleavage of autoimmune condition pemphigus foleacious (PF) is the structurally critical adhesion molecule, Dsg1. characterized by blistering in the superficial epidermis, essentially identical to SSSS (Amagai, 1999; Mahoney et Differential diagnosis of the conditions caused by al, 1999; Stanley, 1993). This blistering is caused by the the ETs. Distinguishing between SSSS and toxic epider- disruption of desmosomes after loss of function of the mal necrolysis (TEN) is critical as misdiagnosis of one for cadherin protein Dsg1 because of the interaction with a the other could lead to harmful therapies. Like SSSS, specific autoantibody to Dsg1. The desmosomal TEN is a generalized blistering condition that can involve cadherin proteins, desmogleins and desmocollins, are erythema and skin tenderness. The blisters in TEN are the major cell adhesion molecules and Dsg1 is the deeper in the epidermis, primarily in the basal cell layer, predominant desmosomal cadherin protein in the upper and are brought on by reactions, mainly to drugs. Antibi- epidermis of the skin while Dsg3 is found in the lower otic treatment indicated for the S. aureus infection epidermis (Wu et al, 2000). With the evidence that the associated with SSSS could worsen the prognosis in ETs appeared to be potential serine proteases, and the TEN. A hallmark difference in these conditions is the lack knowledge that PF antibodies injected into neonatal mice of mucous membrane involvement in SSSS. This differ- cause blistering in the superficial epidermis, like that ence along with the clinical history are usually sufficient caused by the ETs, Amagi and coworkers reasoned that to make the diagnosis, although definitive diagnosis can a logical target for ETs would be Dsg1. They initially be made by histology from a full thickness skin biopsy. examined skin from neonatal mice injected with ETA by The mechanism of action of the ETs, specific cleavage of immunofluorescence using antibodies against Dsg1 and Dsg1, along with the distribution of this precise target Dsg3 as a control. They focused on the deeper layers of explains the clinical differences seen in SSSS versus the epidermis, basal and immediate suprabasal areas TEN. Desmoglein 1 is expressed in the epidermis; there- that have Dsg1 and Dsg3 but do not blister. Treatment fore, it is now clear that there is no mucous membrane with ETA caused significant changes in Dsg1 staining involvement in SSSS because there is no target avail- consistent with cleavage and subsequent internalization able for the toxins to act upon. of the cleavage products (Amagai et al, 2000). Dsg3 Pemphigus foliaceus and vulgaris are blistering condi- staining was not affected by ETA, supporting its role in tions caused by reaction of auto-antibodies to Dsg1 and maintaining adhesion at this level of the epidermis after Dsg3 respectively. In either of these conditions antibodies exposure to an ET, (desmoglein compensation) to both desmogleins can be present and the depth of the (Mahoney et al, 1999). To demonstrate actual cleavage lesions will be affected based on the predominant of Dsg1 they used a human keratinocyte cell line to 122 : 5 MAY 2004 STAPHYLOCOCCUS AUREUS EXFOLIATIVE TOXINS 1075

antibodies. Therefore, PF with only antibody to Dsg1, will or mechanism separate from those that determine target look indistinguishable from SSSS by histology. Clinical specificity (Plano unpublished data). The crystal struc- history is usually sufficient to make the diagnosis versus ture of the ETs revealed unique areas in both the amino SSSS. One would need to establish a history of erythe- and carboxy-terminal regions that could potentially play ma, fever and try to identify the infectious focus with S. a role in specifically targeting these toxins to the upper aureus to establish the diagnosis of SSSS. epidermis. It was recently reported that there is differential expression of Dsg3 in human neonatal versus adult skin Acknowledgements (Wu et al, 2000). These investigators reported that in the Part of the work presented was supported by grant neonate Dsg3 is expressed throughout the epidermis #K08 AI01466 from the National Institutes of Health. I while in the adult it is limited to the basal and suprabasal would like to thank Michelle Perez for help in preparation layers. The knowledge of the differential expression of of the manuscript and Dr. George Munson for critical Dsg1 and Dsg3 in neonatal versus adult skin can explain review. the clinical finding that newborns of mothers with PF appear to be protected against blistering from maternal References anti-Dsg1 antibodies. It has been postulated that the Adesiyun AA, Lenz W, Schaal KP: Exfoliative toxin produc- presence of Dsg3 in the upper epidermis compensates tion by Staphylococcus aureus strains isolated from animals for antibody induced loss of Dsg1 function thereby and human beings in Nigeria. Microbiologica 14:357-362, inhibiting blister formation. These facts could be confus- 1991 ing when trying to explain the lack of desmoglein Adkins B: T cell function in newborn mice and humans. compensation in SSSS in this same population. Immunol. Today 20:330-335, 1999 Likewise, if the primary desmoglein in the adult upper epidermis is Dsg1 why are adults not more sensitive to Aktories K, Schmidt G, Just I: Rho GTPases as targets of the affects of these toxins? Recent studies showed that bacterial protein toxins. Biol Chem 381:421-426, 2000 sustained high levels of ETs were sufficient to overcome Amagai M: Autoimmunity against desmosomal cadherins in any desmoglein compensation that protected 10 day old pemphigus. J Dermatol Sci 20:92-102, 1999 mice from the affects of the ETs (Plano et al, 2001). Amagai M, Matsuyoshi N, Wang ZH, Andl C, Stanley JR: The protection of adults is directly related to serum Toxin in bullous impetigo and staphylococcal scalded-skin clearance of the toxins most probably due to a maturing syndrome targets desmoglein 1. Nature Medicine 6:1275- renal system, but the timing of protection demonstrated 1277, 2000 in the mouse time course (Plano et al, 2001) was Amagai M, Yamaguchi T, Hanakawa Y, Nishifuji K, Sugai M, extremely acute and might not be completely explained Stanley JR: Staphylococcal exfoliative toxin B specifically by a gradual maturation of the system required for clear- cleaves desmoglein 1. J Invest Dermatol 118:845-850, 2002 ance. This is also supported by the occurrence of SSSS in healthy adults without any renal deficiencies (Oyake et Anderson LO: Differentiation and distribution of three types al, 2001). This would suggest that maturation and or of exfoliative toxin produced by Staphylococcus aureus strain isolated from a horse phlegmon. FEMS Immunol Med expression of as yet unidentified or uncharacterized skin Microbiol 20:301-310, 1998 components might also be contributing to the relative protection of adults from the affects of the toxins. Arbuthnott JP, Kent J, Lyell A, Gemmell CG: Studies of Recently newly identified desmogleins were reported in staphylococcal toxins in relation to toxic epidermal necroly- mice (Whittock, 2003). The newly identified mouse Dsg5 sis (the scalded skin syndrome). Br. J. Derm. 86:35-39, has significant nucleotide homology, 96%, with Dsg1 and 1972 has identical amino acid sequence at the cleavage site of Bailey CJ, Deazavedo J, Arbuthnott JP: A comparitive study Dsg1. No information is available at this time about of two serotypes of epidermolytic toxins from Staphylococ- levels of expression in the adults or humans, however cus aureus. Biochim. Biophys. Acta 624:111-120, 1980 given the highly conserved nature of this family of Bailey CJ, Lockhart BP, Redpath MB, Smith TP: The epider- proteins one would expect that similar proteins could be molytic (exfoliative) toxins of Staphylococcus aureus. Med. found in human skin and might be contributing to protec- Microbiol. Immunol. 184:53-61, 1995 tion in the adult. Whittock and Bower have also reported genetic evidence for an additional desmoglein isoform Bailey CJ, Redpath MB: The esterolytic activity of epider- (Dsg4) in the human genome (Whittock and Bower, molytic toxins. Biochem. J. 284:177-180, 1992 2003). These data considered together supports the Bergdoll MS (1972) in The Staphylococci, ed. Cohen, J. O. need for further studying this area. (Wiley, New York), pp. 187-248. Future areas of research. One of the most signifi- Bergdoll MS (1983) in Staphylococci and Staphylococcal cant remaining questions and an area of potential Infections, eds. Easmon, C. S. F. & Adlam, C. (Academic ongoing research is what causes these toxins to be Press, London), Vol. 2, pp. 559-598. targeted to the skin. In 1976, Fritsch and colleagues, Bergdoll MS, Crass BA, Reiser RF, Robbins RN, Davis JP: using radioactive iodine labeled ET, showed that the A new staphylococcal , enterotoxin F, associated toxins preferentially concentrate in the skin in newborn with toxic shock syndrome Staphylococcus aureus isolates. mice (Firtsch et al, 1976). In recent studies in our labora- Lancet 1:1017-1021, 1981 tory we have shown the presence of ETs in the skin of Bloster-Hautamaa DA, Kreiswirth BN, R.P. N, Schlievert both newborn susceptible species and non-susceptible PM: The nucleotide and partial amino acid sequence of toxic species (newborn Balb/c mice and newborn rats respec- shock syndrome toxin-1. J. Bio. Chem. 261:15783-15786, tively) after injection with purified toxin. This suggests 1986 that targeting to the skin may be governed by a process 1076 PLANO THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Cavarelli J, Prevost G, Bourguet W, et al: The structure of Ji G, Beavis RC, Novick RP: Cell density control of staphy- Staphylococcus aureus epidermolytic toxin A, an atypic lococcal virulence mediated by an octapeptide pheromone. serine protease, at 1.7 A resolution. Structure 5:813-824, Proc Natl Acad Sci U S A 92:12055-12059, 1995 1997 Kaper JB, Hacker J: Pathogenicity islands and other mobile Centers for Disease Control and Prevention: Staphylococ- virulence elements. (1999) Pathogenicity islands and other cus aureus with reduced susceptibility to vancomycin-- mobile virulence elements (ASM Press, Washington, DC). United States, 1997. Morb. Mortal. Wkly. Rep. 46:765-766, 1997 Koch PJ, Franke W: Desmosomal cadherins: another growing multigene family of adhesion molecules. Current Centers for Disease Control and Prevention: Staphylococ- Biology 6:682-687., 1994 cus aureus with reduced susceptibility to vancomycin-- Illinois, 1999. JAMA 283:597-598, 2000a Kondo I, Sakurai S, Sarai Y, Futaki S: Two serotypes of exfoliatin and their distribution in staphylococcal strains Centers for Disease Control and Prevention: Staphylococ- isolated from patients with scalded skin syndrome. J Clin cus aureus with reduced susceptibility to vancomycin-- Microbiol 1:397-400, 1975 Illinois, 1999. Morb. Mortal. Wkly. Rep. 48:1165-1167, 2000b Koning S, van Belkum A, Snijders S, et al: Severity of nonbullous Staphylococcus aureus impetigo in children is Chitaev NA, Troyanovsky SM: Direct Ca2+-dependent associated with strains harboring genetic markers for heterophilic interaction between desmosomal cadherins, exfoliative toxin B, Panton-Valentine leukocidin, and the desmoglein and desmocollin, contributes to cell-cell multidrug resistance plasmid pSK41. J Clin Microbiol adhesion. J Cell Biol 138:193-201, 1997 41:3017-3021, 2003 Cribier B, Piemont Y, Grosshans E: Staphylococcal scalded Kuroda M, Ohta T, Uchiyama I, et al: Whole genome skin syndrome in adults. J. Amer. Acad. Derm. 30:319-324, sequencing of meticillin-resistant Staphylococcus aureus. 1994 Lancet 357:1225-1240, 2001 Dancer SJ, Garrett R, Saldanha J, Jhoti H, Evans R: The Ladhani S: Recent developments in staphylococcal scalded epidermolytic toxins are serine proteases. FEBS Lett. skin syndrome. Clin Microbiol Infect 7:301-307, 2001 268:129-132, 1990 Ladhani S, Evans RW: Staphylococcal scalded skin Dancer SJ, Noble WC: Nasal, axillary, and perineal carriage syndrome. Archives of Disease in Childhood 78:85-88, 1998 of Staphylococcus aureus among women: identification of strains producing epidermolytic toxin. J Clin Pathol 44:681- Ladhani S, Joannou CL, Lochrie DP, Evans RW, Poston 684., 1991 SM: Clinical, microbial, and biochemical aspects of the exfoliative toxins causing staphylococcal scalded skin Elias PM, Fritsch P, Dahl MV, Wolff K: Staphylococcal toxic syndrome. Clin. Microbiol. Rev. 12:224-242, 1999 epidermal necrolysis: pathogenesis and studies on the subcellular site of action of exfoliatin. J Invest Dermatol Lee CY, Schmidt JJ, Johnson-Winegar AD, Spero L, Iando- 65:501-512, 1975 lo JJ: Sequence determination and comparison of the exfoliative toxin A and toxin B genes from Staphylococcus Firtsch P, Elias P, Varga J: The fate of Staphylococcal exfoli- aureus. J. Bacteriol. 169:3904-3909, 1987 atin in newborn and adult mice. British J. Derm. 95:275-284, 1976 Mackenzie A, Johnson W, B. H, Norrish B, Jamieson F: A prolonged outbreak of exfoliative toxin A-producing Staphy- Gemmell CG: Staphylococcal scalded skin syndrome. J. lococcus aureus in a Newborn Nursery. Diagn. Microbiol. Med. Microbiol. 43:318-327, 1995 Infect. Dis. 21:69-75, 1995 Gillet Y, Issartel B, Vanhems P, et al: Association between Mahoney MG, Wang Z, Rothenberger K, Koch PJ, Amagai Staphylococcus aureus strains carrying gene for Panton- M, Stanley JR: Explanations for the clinical and microscopic Valentine leukocidin and highly lethal necrotising pneumonia localization of lesions in pemphigus foliaceus and vulgaris. J in young immunocompetent patients. Lancet 359:753-759, Clin Invest 103:461-468, 1999 2002 Marrack P, Kappler J: The staphylococcal enterotoxins and Gravet A, Couppie P, Meunier O, et al: Staphylococcus their relatives. Science 248:705-711, 1990 aureus isolated in cases of impetigo produces both epider- molysin A or B and LukE-LukD in 78% of 131 retrospective Melish ME: Staphylococcal scalded skin syndrome: the and prospective cases. J Clin Microbiol 39:4349-4356, 2001 expanded clinical syndrome. J. Pediatrics 78:958-967, 1971 Hanakawa Y, Schechter NM, Lin C, et al: Molecular mecha- Melish ME, Chen FS, Sprouse S, Stuckey M, Murata MS: nisms of blister formation in bullous impetigo and staphylo- Epidermolytic toxin in staphylococcal infection: toxin levels coccal scalded skin syndrome. J Clin Invest 110:53-60, and host response. Zentralbl. Bakteriol. Suppl. 10:287-298, 2002 1981 Iandolo JJ: Genetic analysis of extracellular toxins of Melish ME, Glasgow LA: The staphylococcal scalded-skin Staphylococcal aureus. Annu. Rev. Microbiol. 43:375-402, syndrome. N. Engl. J. Med. 282:1114-1119, 1970 1989 Monday SR, Vath GM, Ferens WA, et al: Unique superanti- Ito Y, Funabashi Yoh M, Toda K, Shimazaki M, Nakamura T, gen activity of staphylococcal exfoliative toxins. J. Immunol. Morita E: Staphylococcal scalded-skin syndrome in an adult 162:4550-4559, 1999 due to methicillin-resistant Staphylococcus aureus. J Infect Nilles LA, Parry DA, Powers EE, Angst BD, Wagner RM, Chemother 8:256-261, 2002 Green KJ: Structural analysis and expression of human Jarraud S, Lyon GJ, Figueiredo AM, et al: Exfoliatin-produc- desmoglein: a cadherin-like component of the desmosome. ing strains define a fourth agr specificity group in Staphylo- Journal of Cell Science 99:809-821., 1991 coccus aureus. J Bacteriol 182:6517-6522, 2000 Ninomiya J, Ito Y, Iwao T: Purification of protease from a mixture of exfoliative toxin and newborn-mouse epidermis. Infection and Immunity 68:5044-5049, 2000 122 : 5 MAY 2004 STAPHYLOCOCCUS AUREUS EXFOLIATIVE TOXINS 1077

Novick RP: Autoinduction and signal transduction in the characterization of mouse DSG4, DSG5, and DSG6. J regulation of staphylococcal virulence. Mol Microbiol Invest Dermatol 120:970-980, 2003 48:1429-1449, 2003 Whittock NV, Bower C: Genetic evidence for a novel human Novick RP: Mobile genetic elements and bacterial toxinoses: desmosomal cadherin, desmoglein 4. J Invest Dermatol the superantigen-encoding pathogenicity islands of Staphy- 120:523-530, 2003 lococcus aureus. Plasmid 49:93-105, 2003 Wiley BB, Rogolsky M: Molecular and serological differenti- O’Toole P, Foster TJ: Nucleotide sequence of the epider- ation of staphylococcal exfoliative toxin synthesized under molytic toxin A gene of Staphylococcus aureus. J. Bacteriol. chromosomal and plasmid control. Infect. Immun. 18:487- 169:3910-3915, 1987 494, 1977 Oyake S, Oh-i T, Koga M: Staphylococcal scalded skin syndrome in a healthy adult. J Dermatol 28:145-148, 2001 Wu H, Wang ZH, Yan A, et al: Protection against pemphigus foliaceus by desmoglein 3 in neonates. N Engl J Med Papageorgiou AC, Plano LRW, Collins CM, Acharya KR: 343:31-35, 2000 Structural differences in Staphylococcus aureus exfoliative toxins A and B as revealed from their crystal structures. Yamaguchi T, Hayashi T, Takami H, et al: Phage conversion Submitted 1999 of exfoliative toxin A production in Staphylococcus aureus. Mol Microbiol 38:694-705, 2000 Piemont Y, Rasoamananjara D, Fouace JM, Bruce T: Epidemiological investigation of exfoliative toxin-producing Yamaguchi T, Hayashi T, Takami H, et al: Complete Staphylococcus aureus strains in hospitalized patients. J. nucleotide sequence of a Staphylococcus aureus exfoliative Clin. Microbiol. 19:417-420, 1984 toxin B plasmid and identification of a novel ADP-ribosyl- transferase, EDIN-C. Infect Immun 69:7760-7771, 2001 Plano LRW, Adkins B, Woisnchnik M, Ewing R, Collins CM: Toxin levels in serum correlate with the development of Yamaguchi T, Nishifuji K, Sasaki M, et al: Identification of Staphylococcal scalded skin syndrome in a murine model. the Staphylococcus aureus etd pathogenicity island which Infection and Immunity 69:5193-5197, 2001 encodes a novel exfoliative toxin, ETD, and EDIN-B. Infect Plano LRW, Gutman DM, Woischnik M, Collins CM: Recom- Immun 70:5835-5845, 2002 binant Staphylococcus aureus exfoliative toxins are not Yamaguchi T, Yokota Y, Terajima J, et al: Clonal association bacterial superantigens. Infection and Immunity. 68.:3048- of Staphylococcus aureus causing bullous impetigo and the 3052, 2000 emergence of new methicillin-resistant clonal groups in Prevost G, Couppie P, Monteil H: Staphylococcal epider- Kansai district in Japan. J Infect Dis 185:1511-1516, 2002 molysins. Curr Opin Infect Dis 16:71-76, 2003 Prevost G, Rifai S, Chaix ML, Meyer S, Piemont Y: Is the Staphylococcus aureus exfoliative toxins: His72, Asp120, Ser195 triad constituative of the catalytic How they case disease. site of staphylococcal exfoliative toxin A. (1992) Is the His72, Lisa R W Plano Asp120, Ser195 triad constituative of the catalytic site of Update of recent studies staphylococcal exfoliative toxin A. (Gustav Fischer, New York). In recent studies published after submission of this Prevost G, Rifai S, Chaix ML, Piemont Y: Functional manuscript, Hanakawa, et al have further characterized the evidence that the Ser-195 residue of staphylococcal exfolia- mechanism of action of the exfoliative toxins and aspects of tive toxin A is essential for biologic activity. Infect. Immun. the interaction of the toxins and their targets desmoglein 1, 59:3337-3339, 1991 that contribute to their specificity. They have shown that Rago JV, Vath GM, Bohach GA, Ohlendorf DH, Schlievert not only do the toxins recognize a specific amino acid PM: Mutational analysis of the superantigen staphylococcal sequence, they also require a calcium dependent exfoliative toxin A (ETA). J Immunol 164:2207-2213, 2000 conformation of desmoglein 1 for cleavage. They showed Rago JV, Vath GM, Tripp TJ, Bohach GA, Ohlendorf DH: that exfoliative toxins could not cleave Dsg1 after o Staphylococcal exfoliative toxins cleave alpha- and beta- pretreatment at 56 C or at high or low pH, suggesting that a melanocyte stimulating hormones. Infection and Immunity proper conformation was needed for cleavage. They further 68:2366-2368, 2000 confirmed that calcium depletion of Dsg1 caused a Sato H, Matsumori Y, Tanabe T, Saito H, Shimizu A, conformational change that resulted in the inability of a Kawano J: A new type of staphylococcal exfoliative toxin previously active toxin to cleave its target (Hanakawaet al , from a Staphylococcus aureus strain isolated from a horse 2003). In a subsequent study they report initial kinetics with with phlegmon. Infect Immun 62:3780-3785, 1994 kcat/Km values that suggest very efficient proteolysis by these toxins. Using truncated mutant human Dsg1 and chimeric Stanley JR: Cell adhesion molecules as targets of autoanti- Dsg’s generated with human Dsg1 and either human Dsg3 bodies in pemphigus and pemphigoid, bullous diseases due or canine Dsg1 they identified regions of Dsg1 upstream to defective epidermal cell adhesion. Adv Immunol 53:291- from the cleavage site important for both binding and 325, 1993 specific cleavage by these toxins (Hanakawa et al , 2004). Sugai M, Hashimoto K, Kikuchi A, et al: Epidermal cell differ- These data support the complexity of the toxin target entiation inhibitor ADP-ribosylates small GTP-binding interactions that govern the exquisite species specificity and proteins and induces hyperplasia of epidermis. J Biol Chem probably contribute to the age specificity of the syndromes 267:2600-2604, 1992 caused by these toxins and validate the importance of Vath GM, Earhart CA, Monie DD, Iandola JJ, Schlievert PM, continued investigation of these unique toxins. Ohlendorf DH: The crystal structure of exfoliative toxin B: a Hanakawa Y, Selwood T, Woo D, Lin C, Schechter NM, superantigen with enzymatic activity. Biochem. 38:10239- Stanley JR: Calcium-dependent conformation of desmoglein 10246, 1999 1 is required for its cleavage by exfoliative toxin. J Invest Vath GM, Earhart CA, Rago JV, et al: The structure of the Dermatol 121:383-389, 2003 superantigen exfoliative toxin A suggests a novel regulation Hanakawa Y, Schechter NM, Lin C, Nishifuji K, Amagai M, as a serine protease. Biochem 36:1559-1566, 1997 Stanley JR: Enzymatic and molecular characteristics of the Whittock NV: Genomic sequence analysis of the mouse efficiency and specificity of exfoliative toxin cleavage of desmoglein cluster reveals evidence for six distinct genes: desmoglein 1. J Biol Chem 279:5268-5277, 2004

PRINTED ON RECYCLED PAPER

The Dermatology Foundation: Enhancing patient care through research in dermatology