PDF hosted at the Radboud Repository of the Radboud University Nijmegen

The following full text is a publisher's version.

For additional information about this publication click this link. http://hdl.handle.net/2066/74419

Please be advised that this information was generated on 2021-09-26 and may be subject to change. T C

CYSTATIN M/E AND ITS TARGET A novel biochemical pathway that controls stratum corneum homeostasis

TSING CHENG

Tsing Cheng Cystatin M/E and its target proteases. A novel biochemical pathway that controls stratum corneum homeostasis

Thesis, Radboud University Nijmegen Medical Centre, The Netherlands With summary in Dutch – 176 p. © 2009

Cover illustration Predicted 3‐D structure of cystatin M/E (modeled by Gys J. de Jongh)

Print PrintPartners Ipskamp, Enschede

ISBN 978‐90‐9024266‐8

No part of this book may be reproduced in any form without written permission of the author. All published papers are reprinted with permission and with credit to their resource.

CYSTATIN M/E AND ITS TARGET PROTEASES A novel biochemical pathway that controls stratum corneum homeostasis

Een wetenschappelijke proeve op het gebied van de Medische Wetenschappen

Proefschrift

ter verkrijging van de graad van doctor aan de Radboud Universiteit Nijmegen op gezag van de rector magnificus prof. mr. S.C.J.J. Kortmann, volgens besluit van het College van Decanen in het openbaar te verdedigen op

woensdag 17 juni 2009 om 13.30 uur precies

door

Tsing Cheng geboren op 3 augustus 1979 te Heerlen

Promotor: Prof. dr. J. Schalkwijk

Copromotor: Dr. P.L.J.M. Zeeuwen

Manuscriptcommissie: Prof. dr. R.E. Brock (voorzitter) Dr. W. Boelens Dr. T. Reinheckel (Albert‐Ludwigs‐Universität Freiburg) Contents

Abbreviations 6

Chapter 1 General introduction 9

Chapter 2 Cystatin M/E is a high affinity inhibitor of V and 51 by a reactive site that is distinct from the legumain‐ . A novel clue for the role of cystatin M/E in epidermal cornification

Chapter 3 Colocalization of cystatin M/E and in lamellar granules 59 and corneodesmosomes suggests a functional role in epidermal differentiation

Chapter 4 A regulatory role for cystatin M/E in epidermal desquamation. 69 Identification of a two‐chain form of cathepsin V

Chapter 5 Colocalization of cystatin M/E and its target proteases suggests a 87 role in terminal differentiation of human hair follicle and nail

Chapter 6 The cystatin M/E controlled pathway of skin barrier formation: 99 expression of its key components in psoriasis and atopic dermatitis

Chapter 7 Transgenic expression of cystatin M/E in epidermis to rescue the 119 lethal ichq mouse phenotype

Chapter 8 Summary and Discussion 133

Nederlandse samenvatting 145

Dankwoord 154

Curriculum Vitae 157

List of Publications 158

Color illustrations 161

Abbreviations

AD atopic dermatitis AEP asparaginyl endopeptidase (legumain) BCC basal cell carcinoma BSA bovine serum albumin CDSN corneodesmosin CpG cytosine‐phosphate‐guanine CTSB CTSD cathepsin D CTSL cathepsin L CTSS CTSV cathepsin V DAPI 4’,6‐diamine‐2’‐phenylindole dihydrochloride DED de‐epidermized dermis DSC desmocollin DSG desmoglein E‐64 trans‐epoxysuccinyl‐L‐leucylamido‐(4‐guanidino)butane EDTA ethylenediaminetetraacetic acid eGFP enhanced green fluorescent protein ELISA ‐linked immunosorbent assay ER endoplasmic reticulum FCS fetal calf serum GEO Expression Omnibus HLH hemophagocytic syndrome (hemophagocytic lymphohistiocytosis) HRP horseradish peroxidase INV involucrin IRES internal ribosomal entry site IRS inner root sheath KGM keratinocyte growth medium KLK kallikrein LEKTI lympho‐epithelial Kazal type inhibitor LG lamellar granule LGMN legumain MALDI‐MS matrix‐assisted laser desorption ionisation – mass spectrometry MES 2‐(N‐morpholino)ethanesulfonic acid MHC major histocompatibility complex MMP matrix metalloprotease NS normal skin ORS outer root sheath

6 PBS phosphate buffered saline PCR polymerase chain reaction PDAC pancreatic ductal adenocarcenoma PS psoriasis qPCR real‐time quantitative PCR RT‐PCR reverse transcriptase – PCR sc single‐chain SC stratum corneum SCC squamous cell carcinoma SDS‐PAGE sodium dodecyl sulfate – polyacrylamide gel electrophoresis SG stratum granulosum siRNA small interference RNA SPR small proline‐rich protein SV40 simian virus 40 tc two‐chain TGase transglutaminase TGM transglutaminase THH trichohyalin Z benzyloxycarbonyl

7

8 CHAPTER 1

General introduction

Tsing Cheng

Parts of this chapter have been published in: Zeeuwen, P.L., Cheng, T., and Schalkwijk, J. (2009) The biology of cystatin M/E and its cognate proteases. J.Invest.Dermatol. In press, doi: 10.1038/jid.2009.40

9

10 General Introduction

Skin is our largest organ and plays an important protective and aesthetic role as it is the interface of the body with the environment. The skin forms a structural barrier that protects the inner tissues against physical stress and invasion of micro‐organisms and toxic agents, but it also prevents the loss of essential body fluids. The social and psychological function of skin is shown by its influence in a person’s attraction, self‐esteem, social contacts and standing in society. Human skin consists of an outer layer of stratified, keratinizing epithelium, the epidermis, which is anchored to a basement membrane, separating the epidermis from the underlying connective tissue layer, the dermis, which contains the vascular network, the hair follicles, the sweat glands, and the sebaceous glands. Correct development of the epidermis is vital for skin barrier function as abnormalities in cornification and desquamation can lead to severe skin disorders. Therefore, elucidation of the molecular processes that underlie epidermal differentiation is of great importance for understanding skin barrier formation and function, and for future therapy of skin barrier disorders.

SKIN BARRIER FORMATION AND FUNCTION

In order to protect the body against invasion of micro‐organisms, penetration of toxic agents, and loss of essential body fluids, the skin has evolved an intricate terminal differentiation process that results in a tough, water‐impermeable outer covering that is constantly renewing. The outcome of this epidermal differentiation process is an organized tissue in which morphologically distinguishable cells are arranged in discrete layers of basal, spinous, granular, and cornified cells (Figure 1)1,2. The stratum basale is the layer of cells directly contacting the basement membrane and supposedly harbours epidermal stem cells as this layer provides the germinal cells that are necessary for epidermal regeneration3.

Figure 1: Structure of human epidermis. Human skin consists of dermis (connective tissue) and epidermis (epithelium), which are separated by a basement membrane. The epidermis is a stratified structure in which keratinocytes migrate from the basal layer to the stratum corneum as they differentiate into dead corneocytes and eventually are shed from the skin surface. Reprinted from: Alonso, L., Fuchs, E. (2003) Proc.Natl.Acad.Sci.U.S.A 100 Suppl 1, 11830‐11835. © 2003 The National Academy of Sciences of the USA

11 Figure 2: Protein crosslinking by transglutaminases. Ca2+‐dependent transglutaminases (TGM) catalyse the acyl‐transfer reaction of the glutamine residue of protein 1 to the lysine residue of protein 2. The two proteins are linked via a covalent ε(γ‐glutamyl)lysyl isopeptide bond.

Newly formed cells move upwards into the stratum spinosum and the process of terminal differentiation starts as the cells are migrating towards the skin surface. In the spinous layers the intermediate filament network of the cytoskeleton is composed of large amounts of keratins4. The intermediate filaments are anchored to desmosomes, intercellular junctions, which number increases during terminal differentiation to improve tissue integrity and resistance to mechanical forces. The stratum granulosum is the last layer of viable cells and is characterized by the accumulation of keratohyalin granules. The granular cells also present “lamellar granules” containing lipids that after incorporation into the cornified envelope form a waterproof barrier to prevent the loss of body fluids. As keratinocytes terminally differentiate, rupture of the stratum granulosum cell integrity results in release of lysosomal proteases into the cytoplasm that eventually cause cell death. These cells that have lost their nuclei and cytoplasmic organelles, are now called corneocytes and form the stratum corneum. The corneocyte contains a stable and insoluble cornified envelope that was synthesized just beneath the disintegrating plasma membrane. This cornified envelope consists of several specific proteins that are cross‐linked together in an orchestrated way by Ca2+‐dependent transglutaminases (TGMs; Figure 2). The terminal stage of the differentiation process is desquamation, which involves degradation of lipids in the intercellular spaces and loss of residual intercellular desmosomal connections. Cells reaching the skin surface, which consist largely of dead proteinaceous sacs of keratin intermediate filament cables, are continuously sloughed and replaced by inner cells differentiating and moving outward5‐8.

Cornification of keratinocytes Different experimental approaches have led to a model implicating that the cornified envelope assembly is an ordered process, that proceeds in three principal stages: i) initiation of cornified envelope assembly; ii) formation of the corneocyte lipid envelope; and iii) reinforcement of the cornified envelope9‐11. As intracellular Ca2+‐concentrations rise in suprabasal cells, envoplakin, periplakin and involucrin are expressed, which then associate with keratin intermediate filaments and desmosomal junctions (Figure 3). Nearly at the same time, TGM1 is expressed, which docks into the plasma membrane where it catalyzes the intrachain head‐to‐head and head‐to‐tail cross‐linking of involucrin, as well as interchain cross‐linking between both plakins and involucrin. These events first occur at interdesmosomal sites on the cell membrane, and eventually, involucrin, the plakin proteins

12 (and perhaps other proteins) develop a monomolecular layer along the entire inner surface of the plasma membrane, including the desmosomes, thereby forming a scaffold. The second step in this model is the formation of the corneocyte lipid envelope. At a particular time point, perhaps coincident with initiation of scaffold assembly, lamellar granules are thought to fuse with the plasma membrane of the granular cell and discharge their lipid membranes into the intercellular space to form a water‐repelling envelope around the cornified envelope. Finally, other structural proteins including loricrin and small proline rich proteins (SPRs), which together comprise about 80% of the total mass of the cornified envelope, are crosslinked onto the scaffold to reinforce the initial structure. The cytosolic, activated TGM3 enzyme catalyzes the homodimerization of loricrin and the heterotrimerization of loricrin and SPRs, which are subsequently translocated to the border of the cell, where TGM1 crosslinks them onto the pre‐existing scaffold. Minor amounts of other proteins (cystatin A, SKALP/elafin, LCE proteins, repetin, and trichohyalin) also become crosslinked to the cornified envelope. The last stage of the cornified envelope assembly is characterized by rupture of the cell integrity, which results in degradation of most cell organelles and internal structures. Together, the complete cornified envelope, consisting of the protein and lipid envelope and the intercellular lipid lamellae, provide the vital mechanical and water barrier functions for the skin and entire organism.

Figure 3: Cornified envelope assembly. Initiation of cornified envelope assembly starts in the spinous layers with the attachment of envoplakin, periplakin, and involucrin to the plasma membrane by transglutaminases. This is followed by incorporation of lipids to form the lipid envelope for water barrier function. In the granular layers additional structural proteins including loricrin and SPRs are crosslinked to the scaffold for reinforcement of the cornified envelope. Eventually, other proteins are also crosslinked to the cornified envelope. Reprinted with permission from: Candi, E., Schmidt, R., Melino, G. (2005) Nat.Rev.Mol.Cell Biol. 6, 328‐340. ©2005 Nature Publishing Group

13 Desquamation of corneocytes In the final step of terminal differentiation dead corneocytes of the stratum corneum are shed from the skin surface by a process called desquamation. Desquamation requires the degradation of corneodesmosomes, which are modified desmosomes of the epidermis that provide strong cell‐cell adhesion and that are anchoring points for the intermediate filaments of the cytoskeleton. These functions are important in maintaining the integrity of the tissue and in providing the cells resistance to physical stress and pressure12,13. Desmosomes constitute proteins from three separate families: desmosomal cadherins, armadillo proteins, and plakins (Figure 4). The adhesive intercellular junction is formed by the extracellular domains of two types of transmembrane cadherins: desmogleins and desmocollins14. Four desmogleins (DSG1‐4) and three desmocollins (DSC1‐3) have been identified, all clustered on 18, each with its own epidermal expression pattern, e.g. DSG1 and DSC1 are more expressed in the upper granular layers, whereas DSG3 and DSC3 are expressed in the basal and spinous layers (Figure 5). These differential expression patterns could have a role in epidermal morphogenesis as DSG1 and DSC1 seem to promote strong cell‐cell adhesion and keratinocyte terminal differentiation, whereas expression of DSG3 and DSC3 may be related to inhibition of differentiation and stimulation of keratinocyte proliferation15,16. Intracellularly, the desmosomal cadherins interact directly with armadillo proteins such as plakoglobin and plakophilins, which also show differentiation dependent expression patterns in epidermis. The armadillo proteins basically act as a link between the desmosomal cadherins and the intermediate filament binding plakins. Several plakin‐family members including desmoplakin, envoplakin, and periplakin bind intermediate filaments and are localized to desmosomes, but only desmoplakin is indispensible for these intercellular junctions. Deletion of the desmoplakin gene in mice results in lethality during embryonic development, whereas a double knockout of both envoplakin and periplakin in mice does not affect skin barrier formation and function17,18. In corneodesmosomes the intercellular junction is enhanced with the secreted glycoprotein corneodesmosin (CDSN),

Figure 4: Structure of desmosome super‐ imposed on electron micrograph. The extracellular domains of desmosomal cadherins (DSG and DSC) form the adhesive intercellular junction shown as a dense midline (DM). The outer dense plaques (ODP) represent the armadillo proteins (PG and PKP) that form the link between desmosomal cadherins and plakins (DP). Plakins are anchoring points for intermediate filaments (IF) of the cytoskeleton. Reprinted with permission from: Green, K.J., Simpson, C.L. (2007) J.Invest.Dermatol. 127, 2499‐2515. © 2007 The Society for Investigative Dermatology

14 which has homophilic, adhesive properties19,20. Since desquamation requires degradation of intercellular junctions between corneocytes, of corneodesmosomal proteins, especially the desmosomal cadherins and CDSN, is essential in this process. Degradation of corneodesmosomal proteins is mediated by secreted skin proteases, primarily serine, but also aspartic and cysteine proteases. It has been suggested that desquamation results from pH‐dependent activation of several families as the pH of the stratum corneum changes from neutral in the lower layers to acidic in the upper layers21. Previous studies have demonstrated that the serine proteases kallikrein‐5 and 7 (KLK5, KLK7) are able to degrade DSG1, DSC1, and CDSN, whereas aspartic protease cathepsin D and cathepsin V have been suggested as desquamatory based on their activity and localization22‐25.

PROTEOLYTIC ENZYMES

Proteolytic enzymes are proteins that hydrolyse peptide bonds and are commonly termed proteases, proteinases, peptide or peptidases, of which the latter term is actually recommended by the Nomenclature Committee of the International Union of Biochemistry and Molecular Biology (NC‐IUBMB), but here the term protease is used. Proteases act as either endopeptidases, that cleave internal bonds within a protein, or exopeptidases, that cleave at or near the N‐ or C‐terminus of the polypeptide chain. Since hydrolysis is essential for enzyme activation and protein degradation, proteases are key molecules in multiple systems that are active in health and disease, including processes such

Figure 5: Differential expression patterns of desmosomal proteins. In the epidermis desmosomal cadherins (DSG1‐4, DSC1‐3), armadillo proteins (PG, PKP1‐3), and plakins (DP, PPL, EVPL) are expressed in a differentiation dependent manner. Desmosomal proteins are degraded and replaced by new ones during differentiation, while the number of desmosomes increases.

15

Figure 6: Example of a Summary page from the MEROPS database. The Summary page of CTSL is shown: active links are shown by underlined text; information on this Summary page has been edited to fit the figure.

16 as apoptosis, antigen presentation, inflammation, tissue remodelling, angiogenesis, wound healing, fertilization, neurodegeneration, tumorigenesis, and metastasis. Regulation of protease activity is essential as alterations in protease expression, structure, or regulatory mechanisms could lead to physiologic disturbance and pathology26‐28. 15 years ago Rawlings and Barrett introduced a system for classification of all proteases, that has developed since and is implemented in the MEROPS peptidase database, that contains “Summary pages” with information and links for ~3000 proteases (Figure 6; .sanger.ac.uk)29. In this system the proteolytic enzymes are categorized into six classes of proteases depending on the nature of their catalytic mechanism: aspartic proteases, glutamic proteases, metalloproteases, serine proteases, threonine proteases, and cysteine proteases; and each class is denoted with its initial letter (A, G, M, S, T, and C, respectively). Some proteases are grouped as “Unclassified” (U) due to uncertainty of catalytic mechanism. Metalloproteases incorporate a metal ion in their catalytic site that is essential for hydrolysis, while in the other protease classes certain amino acid residues in the catalytic site are essential. Remarkably, aspartic proteases, glutamic proteases and metalloproteases use an activated water molecule as nucleophile in catalysis, while the attacking nucleophile of serine, threonine and cysteine proteases is an internal oxygen or sulphur atom in the side chain of the amino acid (Figure 7). Each class of proteases is subdivided into clans and families, denoted with a second capital letter or a number after the initial letter, respectively. A clan contains one or more families that appear to have come from a unique origin, based on similarities in 3D‐structure, arrangement of the catalytic core

Figure 7: Cleavage of peptide bonds by cysteine proteases. Proteases cleave peptide bonds via hydrolysis. In serine proteases, threonine proteases, and cysteine proteases the attacking nucleophile is an internal hydroxyl (OH) or sulfhydryl (SH) group. Here, the sulfhydryl group of the cysteine residue is activated by transfer of the proton (H+) to the histidine residue (step 1). Subsequently, the peptide bond is cleaved by the attack of the sulfur atom (step 2) and the proton is used to create the new amino terminus of the cleaved substrate (step 3). A

water molecule (H2O) catalyzes release of the

protease‐substrate complex (step 4). H2O is used to restore the sulfhydryl group of the protease and to create a new carboxy terminus of the remaining substrate (step 5). Contrary, in aspartic proteases, glutamic proteases, and metalloproteases the attacking nucleophile is a bound H O molecule, which becomes activated 2 (OH‐ and H+) and performs steps 2 and 3.

17 or amino acid sequence around the catalytic amino acid residues. Families are arranged on the basis of the homology among proteases that show a statistically significant relationship in amino acid sequence. Some clans are denoted with the letter ‘P’ and these contain protease families of more than one catalytic type, e.g. both serine and cysteine proteases. Table 1 shows a simplified example of the classification of proteolytic enzymes according to Rawlings and Barrett26.

Aspartic proteases One remarkable feature of aspartic proteases is that they all act as endopeptidases as no aspartic exopeptidases have been described so far. In aspartic proteases the catalytic residues are two aspartate (Asp) residues (for names, abbreviations, and structures of amino acids see figure 8), that are centrally located in the substrate‐binding cleft, although the second Asp is replaced by another amino acid in some families30. Aspartic proteases differ from other common “amino acid type” proteases (i.e. serine and cysteine proteases) in the use of the attacking nucleophile, since aspartic proteases use a water molecule to attack the peptide bond instead of the side chain of an amino acid residue. The two Asp residues bind and activate the water molecule that is located between the two carboxyl groups of the Asp residues31. Most aspartic proteases are stable at neutral pH, but show optimal activity at acidic pH. Six clans of aspartic proteases have been assigned of which clan AA contains the large family A1 of pepsin‐like proteases. Family A1 includes gastric enzymes pepsin A, gastricsin, and chymosin, which are involved in digestion32; lysosomal proteases D and E, which have been implicated in several physiologic and pathophysiologic mechanisms including skin barrier formation and tumorigenesis33‐35; β‐secretase BACE‐1 (memapsin‐2), which is responsible for generation of amyloid plaques in Alzheimer disease36; and renin, which is part of the renin‐angiotensin system that is involved in control of body fluid osmolality and volume37. Clan AA also includes family A2 of retroviral aspartic proteases such as HIV‐1 protease, which is essential for virus particle maturation38.

Table 1: Classification of proteolytic enzymes according to Rawlings and Barrett Class Clan Family Protease Aspartic AA A1 Pepsin and others and others Cysteine CA C1 C2 and others PA(C) C3 Picornain and others CD C13 Legumain C14 and others and others Glutamic GA G1 Scytalidoglutamic peptidase Etc.

18

Figure 8: Amino acids. Names, abbreviations, and structures of the 20 common amino acids are shown. Amino acids are classified into seven categories, each with its own characteristic. Location of the proton acceptor/donor site (H+) in basic amino acids is depicted.

19 Glutamic proteases Recently, the group of glutamic proteases was added to the classification of proteolytic enzymes due to the identification of a new catalytic mechanism in fungal proteases39,40. Glutamic proteases share common features with aspartic proteases since these too are all endopeptidases and use a water molecule to attack the peptide bond. Instead of two Asp residues, glutamic proteases have glutamate (Glu) and glutamine (Gln) as catalytic residues of which Glu activates the water molecule. So far only clan GA, formed by family G1, has been assigned comprising only fungal proteases; no glutamic proteases have been described yet for other eukaryotic organisms.

Metalloproteases Like aspartic and glutamic proteases, hydrolysis by metalloproteases is mediated by a water molecule, but the water molecule is bound to a divalent metal cation, usually zinc but also cobalt, manganese, nickel or copper. Most metalloproteases require one zinc ion for catalysis, but some families, including metalloproteases that incorporate cobalt or manganese, use two metal ions that act co‐catalytically. Co‐catalytical metalloproteases act as exopeptidases, whereas metalloproteases with one metal ion may be exopeptidases or endopeptidases. In the catalytic site specific amino acid residues (histidine, glutamate, aspartate or lysine) act as ligands for positioning of the metal ion, while another one is the catalytic residue that activates the water molecule, usually Glu. The largest clan of metalloproteases is clan MA comprising zinc‐dependent metalloproteases and is characterized by the zinc‐binding motif containing two histidine (His) residues, that act as zinc‐binding ligands, and one Glu as catalytic residue: His‐Glu‐Xaa‐Xaa‐His (“HEXXH motif”)41. Clan MA can be divided into two subclans based on the third zinc ligand that is located C‐ terminally of the HEXXH motif. The gluzincins of subclan MA(E) have Glu as third ligand, whereas His or Asp is the third ligand in metzincins of subclan MA(M). The metzincins are also distinguished by the presence of a conserved β‐turn that contains a methionine residue (“Met‐turn”) and all metzincins are endopeptidases42. Family M10 of the metzincins are known as matrix metalloproteases (MMPs), including collagenases, gelatinases, stromelysins, matrilysins, and membrane‐type MMPs, that are involved in degradation of extracellular matrix components in physiologic and pathophysiologic processes such as tissue remodelling, wound healing, angiogenesis, and tumor invasion43,44.

Serine proteases In serine proteases, the nucleophile that attacks the peptide bond in hydrolysis is the hydroxyl group of the serine residue (Figure 8). The catalytic mechanism of most serine proteases (categorized in clans PA(S), SB, SC, and SK) requires two other amino acid residues: His as proton donor and Asp for orientation of the imidazolium ring of His. Other clans use lysine (Lys) as proton donor or have a catalytic dyad that does not require the use of a third amino acid residue45,46. Serine proteases are mostly active at a neutral to alkaline pH with generally little activity in acidic environments. Family S1, categorized in clan PA(S), is the

20 largest family of all proteases and compromises only endopeptidases that are involved in several pathways of differentiation and development, blood coagulation, immunity, and digestion. Within family S1 three main activity types can be distinguished: trypsin‐like, chemotrypsin‐like, and elastase‐like activity with cleavage following arginine or lysine residues, hydrophobic amino acids (e.g. leucine), and alanine at the P1 position, respectively. Family members include granzymes, which induce cell death in virus‐infected cells47; kallikreins and matriptase, which are involved in tumorigenesis and epithelial differentiation22,48,49.

Threonine proteases Threonine proteases form a small class with five families of which four families are categorized in clan PB(T) containing N‐terminal nucleophile hydrolases (Ntn‐hydrolases) and one family not assigned to a clan yet. Ntn‐hydrolases are endopeptidases in which the N‐ terminal residue of the mature protein is catalytic (serine, threonine or cysteine). In all families of clan PB, except family T1, protease activity is restricted to autocatalytic activation of the proenzyme and once the mature protein is generated it ceases to have protease activity. In family T1 of proteasome proteases the hydroxyl group of the N‐terminal threonine (Thr) is the attacking nucleophile with Asp/Glu and Lys as other catalytic residues that act in a similar way as His and Asp in serine proteases50,51. The identification of threonine proteases is linked to the resolution of the structure of proteasomes, large protein complexes that degrade proteins tagged by ubiquitination, thereby preventing accumulation and aggregation of misfolded, damaged, and potentially toxic proteins, and controlling cellular processes through signal‐mediated proteolysis of regulatory proteins. A proteasome consists of multiple subunits of which the 20S core particle acts as threonine protease in protein degradation52‐54.

Cysteine proteases The catalytic mechanism of cysteine proteases is very similar to that of serine proteases, but the attacking nucleophile in cysteine proteases is the sulfhydryl group of the cysteine (Cys) residue (Figure 7). Cysteine proteases have a catalytical dyad of Cys and His, where His serves as proton donor in all cysteine proteases. Often a third residue is required for orientation of the imidazolium ring of His, usually asparagine (Asn, clan CA) or Glu (clans PA(C), CD, and others), but other amino acid residues have also been found in certain families such as the that use proline (Pro) or Thr55,56. Contrary to serine proteases, most cysteine proteases are active at acidic pH with generally little activity in neutral environments. The class of cysteine proteases is divided into twelve clans of which clans CA, PA(C), and CD are best‐known. Clan CA includes the papain family (C1), that is formed by lysosomal cysteine proteases known as cathepsins (more about cathepsins in further sections), and the calpain family (C2) of cytosolic calcium‐dependent proteases, that are involved in muscular dystrophy and neurodegenerative diseases through their function in cellular processes such as cytoskeletal remodelling, signal transduction pathways, and

21 apoptosis57,58. Clan PA(C) comprises viral proteases including family C3 of picornains, endopeptidases from RNA viruses of the Picorna group such as polio and hepatitis A viruses. The legumain family (C13) of asparaginyl endopeptidases is categorized in clan CD (see ‘Legumain’ section), and this clan also includes family C14 of caspases, which are cytosolic endopeptidases that are strictly specific for hydrolysis of aspartyl bonds. Caspases perform critical functions in apoptosis and inflammation59.

Cathepsins Lysosomal proteases were long considered to be primarily responsible for intralysosomal protein degradation, mediating “housekeeping” functions in the cell. However, numerous investigations have shown that these proteases, the classical cathepsins, are involved in a variety of physiologic processes such as proenzyme activation, enzyme activation, antigen presentation, hormone maturation, tissue remodeling, and bone matrix resorption60. Recent studies also suggest that cathepsins act as mediators of programmed cell death, and have demonstrated apoptosis pathways to be dependent on lysosomal cathepsin activity61,62. Cathepsins are cysteine proteases of the papain family (C1 ) with the exception of aspartic proteases cathepsins D and E and the serine protease cathepsin G; so far eleven human cysteine cathepsins have been decribed (cathepsins B, C, F, H, K, L, O, S, V/L2, W, and X). The subfamily of cathepsin L‐like cathepsins (including cathepsins K, L, S, and V) can be recognized as they have larger proregions (compared to cathepsin B) and contain two conserved motifs in their amino acid sequence (“GNFD” and “ERFNIN” motifs)63,64. Most cathepsins are monomeric proteins with sizes of 20‐40 kDa except (also known as dipeptidyl‐peptidase I) that can form functional oligomers of 200 kDa65. Some cathepsins are widely expressed, such as cathepins B, C, and L, which are considered to be responsible for general “housekeeping” functions next to their specific functions, whereas others have a more restricted tissue distribution, such as cathepsins K, S, and V, suggesting that these only fulfil organ specific functions. The production, activation, and compartmentalization of cysteine cathepsins is an intricate process, that involves multiple factors and pathways, and varies among cell types and tissues. Most studies have focused on cathepsins B, L, and aspartic cathepsin D, but generally the main mechanisms that are described in this section, can be applied to all cathepsins due to their overall similarity. Cathepsins are synthesized as inactive preproenzymes at the rough endoplasmic reticulum (ER) and targeted into the lumen of the ER by the signal peptide. After cleavage of the signal peptide by ER signal peptidase and proper folding of the proteins, the procathepsins are transported to the Golgi apparatus for post‐translational glycosylation and mannose‐6‐phosphorylation. The proregion is essential for these processes since loss of the proregion results in improper folding with protein instability, ER retention, and lack of mannose‐6‐phosphorylation as secondary effects66,67. At the trans‐Golgi network procathepsins are transported into clathrin‐coated vesicles via mannose‐6‐phosphate receptor‐mediated sorting and subsequently directed to late endosomes and lysosomes, acidic compartments where further proteolytic maturation

22 occurs (Figure 9: route A)68,69. The acidic environment in late endosomes initiates processing of the inactive procathepsins into mature active forms and their cleaved propeptides, which have shown to be inhibitors of the mature enzymes70‐74. The mechanisms behind the processing and functions of the various cathepsin forms are not completely elucidated yet, but several studies have shown that autoactivation mechanisms, both intermolecularly (mature cathepsin processes procathepsin) and intramolecularly (procathepsin processes itself), as well as proteolytic processing by other lysosomal proteases including other cysteine cathepsins, legumain, and cathepsin D, are involved in cysteine cathepsin maturation. X‐ray crystallography of procathepsin B, L, and S has shown that the proregion interacts with the mature part of the protease at the substrate binding cleft, but interestingly it binds to the cleft in the reverse substrate orientation75‐81. Procathepsin is first processed into the cleaved propeptide and an active single‐chain form, that subsequently can be further cleaved into an active two‐chain form. The latter processing step presumably occurs in lysosomes after mature single‐chain cathepsins have been transported there from late endosomes. No functions specifically for the single‐chain or two‐chain forms have been found, and it is not known whether this conversion regulates cathepsin activity or substrate specificity, but the distribution and ratio of both mature forms may differ among tissues as has been shown for cathepsin L82. Besides the default endosomal‐lysosomal pathway, additional trafficking mechanisms for cathepsins exist depending on cell type or tissue. In thyrocytes, keratinocytes, and osteoclasts mature cathepsins can be selectively loaded from lysosomes into transport vesicles, directed to the plasma membrane, and subsequently secreted into the extracellular space, where they may process and activate other proteases or contribute to extracellular matrix remodelling83‐86. Another secretory pathway involves direct secretion of inactive cathepsins via the tubulovesicular structures of the trans‐Golgi network by alternative sorting signals located in the proregion, or by absence of mannose‐6‐phosphorylation receptor‐mediated sorting in cells such as macrophages and fibroblasts due to missing receptors or lack of mannose‐6‐phosphorylation. Acidic microenvironments in the extracellular space could then trigger processing and activation of procathepsins, or re‐ internalization at the plasma membrane via endocytosis could traffic the previously secreted procathepsins back to endosomes and lysosomes (Figure 9: routes B‐D)69,87,88. The finding that cleaved cathepsin propeptides are reversible, pH dependent inhibitors of the mature cathepsin forms, is interesting in the context of secretion and activity of cathepsins in body fluids. Non‐covalently bound complexes between propeptide and cathepsin can be formed at pH 5.5‐6.0, which are highly stable at neutral to slightly alkaline pH contrary to mature cathepsins. Since at pH < 5.0 dissociation of the complex and degradation of the propeptide occurs, procathepsins secreted by the trans‐Golgi network avoid the acidic lysosomal compartments and could form these complexes autocatalytically in slightly acidic environments during transport through the trans‐Golgi network. In body fluids that have a neutral pH extracellular cathepsin, bound to its propeptide, can lie dormant until regional acidification mediates its activity through dissociation of the complex71,89.

23

Figure 9: Trafficking pathways of cathepsins. Multiple trafficking pathways of cathepsins exist depending on cell type and tissue. RNA is translated for protein synthesis by ribosomes. The classical route via the endoplasmic reticulum, trans‐Golgi network, and endosomes to the lysosomes is depicted as route A. Secretory pathways of the (pro)cathepsins from the lysosomes and the trans‐ Golgi network are routes B and C, respectively. Endocytosis could traffic previously secreted inactive procathepsins back to endosomes and lysosomes (route D). Route E represents release of cathepsins from lysosomes into the cytosol through lysosomal destabilization or membrane permeabilization leading to cell death. Alternative splicing and exon skipping results in transport of cathepsins to mitochondria and the nucleus (routes F and G, respectively).

24 Lysosomal destabilization or membrane permeabilization by oxidative stress, lysosomotropic agents, and other molecules results in leakage of mature cathepsins from lysosomes into the cytosol where they perform proteolytic functions in apoptotic pathways61,62,90. Alternative splicing and exon skipping of cathepsin mRNA may occur resulting in protease variants that lack the signal peptide and are not directed to the ER. In chondrocytes, a cathepsin B splice variant lacking exon 2 and 3 is directed to the mitochondria, because N‐terminal truncation exposes a latent mitochondrial import signal91. Experiments with N‐terminally truncated cathepsins have shown that these can also enter the cell nucleus and participate in processing of transcription factors (Figure 9: routes E‐ G)92,93. Furthermore, unknown sorting and import signals, specific protein folding or co‐ import pathways could be responsible for alternative trafficking and localization outside the classical cathepsin compartments.

Cathepsin L Cathepsin L is one of the cathepsins that is ubiquitously expressed and has a cellular “housekeeping” function in lysosomal bulk proteolysis, but also can be secreted and perform tissue specific functions. The secreted form of cathepsin L was previously known as major excreted protein (MEP). The human cathepsin L gene spans ~5.1 kb on chromosome 9q21‐22 and consists of eight exons and seven introns. The promoter region misses the TATA box, but has SP1, SP3, and NF‐Y sites for transcriptional regulation. Alternative splicing results in multiple cathepsin L variants that differ at the 5’ untranslated mRNA region60,94,95. Generally, mRNA translation produces preprocathepsin L of 333 amino acids with a mass of ~37 kDa, and during intracellular transport to the lysosomes the preproenzyme is subsequently modified to procathepsin L of ~35 kDa (residues 18‐333) and processed into the mature forms of cathepsin L (residues 114‐333). Processing of procathepsin L results in a ~30 kDa single‐chain form of cathepsin L, which in turn can be further processed into the two‐chain form by cleavage of the single peptide chain after Asp291 into heavy and light chains (25 and 5 kDa; Figure 10)96. Human cathepsin L contains three internal disulfide bonds and one N‐glycosylation site (Asn221); its is formed by Cys138, His276, and Asn300. Cathepsin L mainly acts as an endopeptidase and preferentially cleaves peptide bonds with aromatic residues in the P2 position and hydrophobic residues in the P3 position97,98. The protease is catalytically active within the pH range of 3.5‐6.5, but optimal pH is 5.5, while at neutral pH cathepsin L is highly unstable and irreversibly inactivated99‐101. Inhibitors of cathepsin L include cystatins, cathepsin L‐like propeptides, leupeptin, and the synthetic inhibitor E‐6471,100,102,103. Cathepsin L is involved in many physiologic and pathophysiologic processes including lysosomal proteolysis, apoptosis, tumor invasion and metastasis, epidermal differentiation, cardiac function, antigen presentation, and metabolism. In lysosomes, cathepsin L could process and activate other proteases such as cathepsin D, but also perform degradation of proteins such as membrane receptors and molecular chaperones104‐107. Insufficient lysosomal protein degradation may lead to enlargement and accumulation of lysosomes

25 disturbing cellular homeostasis and inducing apoptosis. In addition, cytosolic cathepsin L could cleave the anti‐apoptotic Bcl‐2 family member Bid into a truncated form, which is involved in apoptotic signaling pathways. Truncated Bid translocates into the mitochondria, where it stimulates release of cytochrome c that leads to apoptosome formation and cell death61,62. In many cancers overexpression of cathepsin L is reported, which could contribute to tumor invasion and metastasis as secreted cathepsin L is able to degrade extracellular matrix proteins108‐110. Several types of cathepsin L deficient mice have been described (ctsl‐/‐, furless, and nackt)111‐113, which show similar phenotypes with minimal differences, however, most studies have used ctsl‐/‐ mice. Cathepsin L deficient mice are viable and fertile, but have several abnormalities, whereas heterozygous mice do not differ from wild type mice. The most evident characteristic of cathepsin L deficient mice is the skin and hair phenotype showing epidermal hyperplasia and hyperkeratosis with abnormal hair follicle morphogenesis and cycling113. Enhanced recycling of epidermal growth factor receptors in ctsl‐/‐ mice, probably due to longer half‐life of the receptors because of insufficient degradation in lysosomes, results in increased keratinocyte proliferation114. Ctsl‐/‐ mice also suffer from dilated cardiomyopathy as cathepsin L is essential for maintaining the structure of the endosome‐lysosome compartment in cardiomyocytes. In addition, sarcomeric alterations and impaired mitochondrial respiration were detected in cathepsin L deficient hearts contributing to the development of cardiomyopathy115‐117. Furthermore, cathepsin L

Figure 10: Processing of human cathepsin L, cathepsin V, and legumain. Cathepsin L (CTSL), cathepsin V (CTSV), and legumain (LGMN) are produced as inactive preproenzymes. Cleavage of the signal peptide by signal peptidase (A) at the ER results in the inactive proenzyme. Proenzymes are processed in multiple steps into the active mature forms via autocatalytic mechanisms (B) and activation by other lysosomal proteases (C) in acidic environments. The mechanisms behind the processing of these proteases are not fully elucidated yet. Arrowheads indicate cleavage sites.

26 is important for degradation of the class II major histocompatibility complex (MHC) invariant chain in cortical thymic epithelial cells, but not in bone‐marrow derived antigen‐presenting cells. Consequently, cathepsin L deficient mice showed reduced numbers of CD4+ T‐cells in the thymus and periphery118,119. A role in obesity and diabetes has been suggested as cathepsin L might participate in adipogenesis and glucose intolerance. Cathepsin L in vitro degrades the matrix protein fibronectin that restrains pre‐adipocytes from differentiation, thus inhibition of cathepsin L resulted in reduced lipid accumulation. Cathepsin L deficient mice that were put on a rich diet gained little body weight and stayed lean, whereas control littermates showed increased body mass and accumulation of adipose tissue. The glucose metabolism is affected since insulin receptors are degraded by cathepsin L resulting in low serum levels of glucose and insulin due to an increase of muscle insulin receptors. In addition, cathepsin L levels were increased in human obese and type 2 diabetic patients120. Other abnormalities reported in cathepsin L deficient mice include high postnatal mortality rate111,121, periodontal acanthosis122, abnormal bone development123, and increased atrophy of seminiferous tubules of the testis with incomplete differentiation of spermatocytes124.

Cathepsin V Cathepsin V, also known as , is closely related to cathepsin L sharing ~80% protein sequence identity for the mature proteins resulting in a similar structure for cathepsin V as described previously for cathepsin L. Their encoding are located on adjacent sites on chromosome 9q21‐22 with similar genomic organization, suggesting that cathepsin V and cathepsin L have evolved from an ancestral gene. Interestingly, no orthologue of cathepsin V in mouse has been discovered yet and cathepsin V is phylogenetically more related to mouse cathepsin L than to human cathepsin L. The cathepsin V gene spans ~6.4 kb, consists of eight exons and seven introns, and its promoter misses a TATA box and has a SP1 site for transcription initiation125‐128. Cathepsin V is produced as a preproenzyme of 334 amino acids with a mass of ~37 kDa and subsequently processed into procathepsin V (residues 18‐334, ~35 kDa) and mature cathepsin V (residues 114‐334). Only a single‐chain form of ~30 kDa has been described (Figure 10)125,126,129. Cathepsin V contains three internal disulfide bonds, but has two N‐glycosylation sites (Asn221 and Asn292) compared to only one in human cathepsin L; its catalytic triad is formed by Cys138, His277, and Asn301. Cathepsin V differs from cathepsin L in substrate specificity as both aromatic and non‐aromatic hydrophobic residues are preferred in the P2 position, however, cathepsin V is inhibited by the same natural and synthetic inhibitors of cathepsin L126,130,131. Another difference with cathepsin L is that cathepsin V is significantly more stable at mildly acidic and neutral pH, which is reflected in its activity range from pH 4.1 to 7.4, although the pH optimum of 5.7 is close to that of cathepsin L126. Little is known about the biological functions of cathepsin V, but the protease probably fulfills tissue specific functions as its expression is limited to specific tissues: thymus, testis, epidermis, and cornea of the eye25,125‐127. In thymus cathepsin V is highly expressed and able to degrade class II MHC invariant chain for antigen presentation by CD4+

27 T‐cells. A similar function has been described for cathepsin L in mice, but in human cortical thymic epithelial cells cathepsin V, and not cathepsin L, is the dominant cysteine protease126,127,129. Its role in testis has not been studied yet, but cathepsin V may be involved in fertilization processes as has been suggested for other lysosmal proteases127,132. Previously named as stratum corneum thiol protease (SCTP)25,133, cathepsin V was identified as a secreted cysteine protease of late epidermal differentiation that is probably associated with desquamation as its pH optimum reflects the acidic environment of the stratum corneum. Specific epidermal transgenic expression of human cathepsin V rescued the skin and hair phenotype in cathepsin L deficient mice indicating that cathepsin V can compensate for cathepsin L in epidermal differentiation134. Furthermore, cathepsin V expression correlates inversely with skin color, both on the mRNA and protein level, as expression is higher in lightly pigmented skin compared to darkly pigmented skin. However, no differential expression of cathepsin V in melanocytes could be found suggesting that differences in ethnic skin are not directly associated with melanogenesis, but expand beyond the pigmentation process135. Cathepsin V has a role in eye physiology as it is highly expressed in corneal epithelium. Recent studies have suggested that cathepsin V is involved in the control of angiogenesis in the eye, which is important for maintenance of corneal avascularity and transparency that are essential for optimal visual performance. Cathepsin V from limbocorneal epithelial cells was able to generate endostatin as proteolytic product from collagen XVIII degradation, while in vitro studies with recombinant cathepsin V showed that angiostatin‐like fragments, but not plasmin, are produced from plasminogen degradation by cathepsin V. Angiogenesis inhibition activity of the angiostatin‐like fragments on endothelial cells was confirmed by a Matrigel assay. In addition, plasminogen degradation was shown to be a cathepsin V specific function as cathepsins B, K, and L were not able to proteolyse plasminogen136,137. Cathepsin V has also been implicated in cardiovascular pathologies such as atherosclerosis since cathepsins K, S, and V contribute to the elastolytic activity of macrophages in atherosclerotic plaques and diseased blood vessels. Increased expression of cathepsins K, S, and V has already been shown in aortic stenosis and may accelerate destruction of elastin matrix138,139. A role in tumorigenesis has been suggested as cathepsin V is expressed in colorectal and breast carcinomas, but not in normal colon or breast tissues127.

Legumain Legumain, also known as aparaginyl endopeptidase (AEP), is a lysosomal protease that belongs to family C13 of cysteine proteases and is strictly specific for hydrolysis of asparaginyl bonds. However, glycosylation of the Asn residue in the substrate prevents hydrolysis by legumain. Optimal activity is seen at pH 5.8, while at pH 7.0 or higher the protease is irreversibly denatured140,141. Like other lysosomal cysteine proteases legumain is produced as inactive preproenzyme of 433 amino acids and processed into its mature active form via autoactivation and proteolytic processing by other proteases. After cleavage of the signal peptide autocatalytic processing of prolegumain (56 kDa) occurs in multiple steps at

28 acidic pH. First, at pH 5.5 part of the C‐terminus is cleaved from the proenzyme at Asn323 leaving a 47 kDa intermediate, followed by cleavage of eight residues from the N‐terminus at pH 4.5 resulting in an active legumain form of 46 kDa that cannot be further processed autocatalytically (Figure 10). Autoactivation of legumain is an intermolecular mechanism as activated legumain is able to process and activate prolegumain. However, since legumain in vivo has a mass of ~36 kDa, the active 46 kDa legumain is not the final mature form and further processing by a yet unknown lysosomal protease is required142,143. Legumain is inhibited by α1‐macroglobuline, some of the cystatins (C, F, and M/E), iodoacetate, iodoacetamide, and maleimides, but not by E‐64, that inhibits most cysteine proteases including the cathepsins140,141,144. Legumain is expressed in various tissues, but found abundantly in kidney, more specifically in the proximal tubule, where in mice legumain might have a role in extracellular matrix remodeling via intracellular degradation of fibronectin145. Studies in legumain deficient mice showed enlarged lysosomes in kidney cells and complete absence of two‐ chain forms of cathepsins B, H, and L, indicating a function for legumain in lysosomal protein degradation and proteolytic processing of other lysosomal proteases. Although legumain is essential in processing of single‐chain cathepsin L to two‐chain cathepsin L in mice, the single‐chain form alone is sufficient enough for cathepsin L function in vivo146,147. Legumain has been shown to be involved in class II MHC antigen presentation, but legumain is not essential in this process as no differences in processing of invariant chain or maturation of class II MHC products were found between legumain deficient mice and wild type mice147‐149. Furthermore, legumain deficient mice develop a phenotype resembling hemophagocytic syndrome (hemophagocytic lymphohistiocytosis, HLH), which is characterized by fever, anemia, cytopenia, hepatosplenomegaly, and hemophagocytosis (ingestion of red blood cells) by macrophages in bone marrow, suggesting that legumain might have a role in HLH pathophysiology150. Overexpression of legumain in cancer has been correlated with tumor invasion and metastasis, possibly through activation of the matrix degrading protein gelatinase A (MMP‐2) by legumain151‐153.

Proteases and their inhibitors in skin physiology The analysis of the pathophysiology of inflammatory diseases, genetic diseases, the invasive and metastatic mechanism of malignant neoplasias, and the course of infectious diseases, has revealed that fundamental knowledge of the regulatory mechanisms of proteases and their inhibitors is crucial for our understanding of these diverse problems27,154,155. Regulation of proteolytic enzyme activity is essential for cells and tissues at many levels (activation, compartmentalization, specific inhibition), because proteolysis at a wrong time and location may have disastrous consequences. Interestingly, an increase of lysosomal cysteine protease activity has been observed in the terminal differentiation process of keratinocytes156,157. Recent studies have reported that lysosomal proteases play important roles in physiologic processes not restricted to lysosomes only. For example, protease activity and regulation outside lysosomes potentially

29 contributes to propagation of apoptosis, a process that is distinct from terminal differentiation of the epidermis but nevertheless shares some molecular and cellular features158. The importance of regulated proteolysis in epithelia159 is well demonstrated by the identification of the SPINK5 serine protease inhibitor as the defective gene in Netherton syndrome (MIM 256500)160, cathepsin C mutations in Papillon‐Lefevre syndrome (MIM 245000)161, matriptase deficiency due to mutations in the ST14 gene in ichthyosis with hypotrichosis syndrome (MIM 610765)162, cathepsin L deficiency in mice111‐113, targeted ablation of cathepsin D in mice33, and the phenotype of targeted epidermal overexpression of stratum corneum chymotryptic enzyme (KLK7) in mice163.

CYSTATIN M/E

The regulation of activity of cysteine proteases for correct functioning is a fragile balance of many factors, one of the most crucial being their cognate protease inhibitors, which include members of the cystatin gene family. Cystatins are small proteins that are widely distributed in several human tissues and body fluids. They are generally tight binding inhibitors of cysteine proteases such as cathepsins B, H, K, L, and S102, and should therefore serve a protective function to regulate the activities of such endogenous proteases. It is thought that cysteine proteases and their inhibitors play an important role in pathophysiologic processes such as invasion of normal tissues by tumors or micro‐ organisms, viral infections, and inflammation164‐167. Cystatins are members of a superfamily of evolutionary related proteins and can be divided into three major families: family‐1 cystatins (A and B), family‐2 cystatins (C, D, F, G, M/E, S, SN, and SA), and the kininogens (L‐ and H‐kininogens), which belong to family‐3 cystatins. In general, cystatins are competitive reversible, tight‐binding proteins that inhibit cysteine proteases in a µM to pM range168. Family‐2 cystatins are viewed as “emergency” inhibitors, which means that they rapidly trap a protease without delay and keep it in a stable complex, preventing any additional proteolysis155,169. In normal situations such an inhibitor would have a physiologic concentration in large excess over the putative protease target(s). These inhibitors are in general physically separated from their target proteases and primarily act on escaped endogenous proteases or exogenous proteases of invading micro‐organisms.

Structure and expression In order to obtain a systematic and quantitative overview of genes expressed by human epidermal cells transcriptomes of cultured human keratinocytes and purified epidermal cells were established via Serial Analysis of Gene Expression (SAGE) and micro‐ arrays (GEO Series accession numbers GSE31 and GSE6601). These analyses revealed quantitative expression data on numerous proteases and protease inhibitors170‐172. Cystatin M/E, a protein not previously known to be expressed in human keratinocytes, was found at considerable levels in these analyses. Cystatin M was initially identified by differential display as a down‐regulated mRNA in metastatic breast tumor cells when compared to normal and

30 primary breast tumor cells173. Independently, the same molecule was found by others via expressed sequence tag sequencing in cDNA libraries derived from epithelial cells, and was designated cystatin E174. It was shown that it was really a protease inhibitor, as the recombinant protein was able to inhibit the plant protease papain. However, in vivo target proteases and the function of this protease inhibitor were unknown. Cystatin M/E is a 14 kDa secreted protein that shares only 35% homology with the human family‐2 cystatins. Cystatin M/E has an overall structure similar to other family‐2 cystatins, such as a signal peptide and two intrachain disulfide bonds, but possesses the unusual characteristic of being a glycoprotein. This protein is only distantly related to the other known family members as reflected by the genomic position of the cystatin M/E gene on chromosome 11q13175, whereas all other family‐2 cystatin genes are clustered in a narrow region on chromosome 20p11.2176. The full‐length cystatin M/E gene was cloned and sequenced, and deposited as CST6 in the GenBank database (accession no. AY145051)177. The cystatin M/E gene is organized in three exons and two introns, and from the translation start site to the translation termination site it spans 1354 bp of genomic DNA. Characterization of the genomic sequence of cystatin M/E shows some differences with the other type 2 cystatins. The size of both introns was remarkably smaller compared to other cystatin family‐2 members (e.g. cystatin S, SA, SN, C, D). These cystatins harbor introns of respectively 1.1 kb to 2.2 kb, whereas the introns of the cystatin M/E gene are only 542 bp and 365 bp long. Besides, these family‐2 cystatins contain an identifiable TATA‐box in their 5’‐flanking sequence, whereas cystatin M/E lacks this conserved sequence. Database analysis predicted three transcription start sites at respectively –75 bp, –42 bp, and –34 bp upstream to the translation initiation codon. This is in accordance with the average length of the expressed sequence tags found in the public database, in which the start positions vary between approximately –60 bp and –20 bp. The expression pattern of human cystatin M/E is unique among cystatins, as high expression levels are largely confined to cutaneous epithelia, while other cystatins are mainly ubiquitously expressed. In human skin, cystatin M/E is expressed in the differentiating and cornifying layers of the epidermis178. This is in accordance with the observation that in vitro cystatin M/E is only expressed in differentiating but not in proliferating keratinocytes178. Cystatin M/E, which has a putative signal peptide, is indeed a secreted protein and is found in vitro in culture supernatant and in vivo in human sweat and semen178,179. Furthermore, cystatin M/E is also highly expressed in the appendages of the skin like the secretory coils of the eccrine sweat glands, the sebaceous glands, and the hair follicles178. Immunohistochemical data have also shown a moderate but specific expression of cystatin M/E in the bronchi of the lung, the proximal tubuli of the kidney, and in the cornea of the eye180. Cystatin M/E protein expression was recently also reported in normal human breast tissue181, breast ductal epithelium182, and in oligodendrocyte‐like cells and astrocyte‐like cells of normal human brain183.

31 Cystatin M/E deficient mice Analysis of adult mouse tissues has revealed an expression pattern that is less restricted than observed previously in humans184. Relatively high levels of cystatin M/E mRNA expression were found in skin, ileum, stomach, eye, and cerebellum; moderate to low levels of expression were found in tongue, palatum, nasal cavity, colon, bladder, skeletal muscle, placenta, and thymus. During embryogenesis, mRNA expression of cystatin M/E was observed from day 16 onwards, which is a developmental stage where epidermal stratification is completed. Cystatin M/E protein was found to be highly expressed in ciliated tracheal epithelium and in bronchial epithelium. Similar to the expression pattern in humans, cystatin M/E was found in the infundibular epithelium of hair follicles and in the stratum granulosum of interfollicular epidermis. The mouse cystatin M/E gene, which was subsequently cloned and characterized (deposited as Cst6 in the GenBank database, accession no. AY093591), has an exon‐intron structure identical to that of all other type 2 cystatins and encodes a 149 amino acid protein184. Alignment with its presumed human orthologue revealed 69% amino acid identity and 82% amino acid conservation. The chromosomal localization of mouse cystatin M/E was determined by fluorescence in situ hybridization and mapped to the proximal end of chromosome 19, a region that is syntenic to human chromosome 11q13 where the human cystatin M/E gene resides. With the knowledge of the cystatin M/E tissue expression and the locus of the gene in mice, it was hypothesized that cystatin M/E could be a candidate gene for the ichq mouse, a naturally occurring mutant . This mouse had morphological and biochemical similarities to a human form of ichthyosis caused by an unknown gene defect185. This hypothesis was confirmed by the identification of a mutation in the cystatin M/E gene of these ichq mice184. Homozygosity for a single nucleotide deletion in exon 1 of the cystatin M/E gene was identified in these mutant mice. This 42delG deletion alters the reading frame, resulting in a premature stop codon at amino acid position 20. Immunohistochemistry confirmed the absence of cystatin M/E at the protein level in ichq mice. Mice that were homozygous for two null alleles displayed a hyperplastic, hyperkeratotic epidermis and abnormal hair follicles and died between 6 and 12 days of age. Heterozygous mice and wild type littermates displayed a normal cystatin M/E expression pattern suggesting that cystatin M/E functions as an inhibitor of hitherto unidentified proteases. The mouse ichq mutant is another example of a disturbed protease‐antiprotease balance causing faulty differentiation processes in the epidermis and hair follicle. Autopsy on neonatal ichq mice strongly suggested that these mice died of dehydration, probably due to a disturbed skin barrier function186. It was shown that the transepidermal water loss through the skin of the cystatin M/E deficient mice is increased when compared to phenotypically normal littermates. Also, a progressive weight loss was observed in the cystatin M/E deficient mice starting at the time point that transepidermal water loss was apparent. At day 11, body weights were approximately 50% compared to normal littermates, and most mice died shortly thereafter186.

32 Morphological and functional analysis has indicated a compromised stratum corneum and aberrant cornification of the hair follicles in cystatin M/E deficient mice, which could be the result of abnormal expression or processing of structural components of the cornified envelope. In two previous studies no abnormalities were found in structural proteins such as cytokeratins, involucrin, and filaggrin185,187. This was investigated again by studying the expression and distribution of loricrin, involucrin, and filaggrin, which are important stratum corneum proteins that are involved in skin barrier function. Immunohistochemical analysis showed that all three proteins are normally expressed in the skin of cystatin M/E deficient mice. However, Western blot analysis revealed that loricrin monomers and loricrin dimers are strongly diminished in skin extracts of cystatin M/E deficient mice186. Although two abundant protein bands of approximately 13 and 15 kDa were observed in the extracts of cystatin M/E deficient mice, in‐gel tryptic digestion followed by matrix‐assisted laser desorption/ionization time‐of‐flight (MALDI‐TOF) mass spectrometry revealed that these were not low molecular weight degradation products of loricrin. The 13 kDa band was identified as S100 calcium binding protein A9 (or calgranulin B), and the 15 kDa band as epidermal fatty acid binding protein (E‐FABP), which have been described as epidermal proteins that are strongly upregulated during epidermal inflammation and following skin barrier disruption, respectively188,189. The strongly diminished expression of loricrin monomers and dimers, and the observation that this phenomenon was not due to degradation, suggested the authors to consider misregulated crosslinking as an alternative explanation. As previously described (see ‘Cornification of keratinocytes’ section), it is known that the cytosolic activated TGM3 enzyme is responsible for the crosslinking of loricrin molecules hereby forming homodimers and heterodimers, which are subsequently translocated to the border of the cell, where TGM1 crosslinks them onto the pre‐existing scaffold. Proteolytic cleavage of TGM3 is required to achieve maximal specific activity of the enzyme190, and one important question was how TGM3 could be processed into its active form. TGM3 activation during keratinocyte differentiation involves cleavage of the 77 kDa zymogen by an unknown protease resulting in the release of 30 kDa and 47 kDa fragments, which then associate together and form the active enzyme191. Western blot analysis revealed an abundant presence of proteolyzed (and hence activated) TGM3 in skin extracts of cystatin M/E deficient mice at day 6, whereas no appreciable levels of proteolyzed TGM3 could be detected in skin extract of wild type littermates186. This indicated that premature, high levels of activated TGM3 are present at the time when skin lesions develop in cystatin M/E deficient mice. It was hypothesized that cystatin M/E controls the activity of an unknown protease that is responsible for TGM3 activation, which leads to crosslinking of loricrin molecules and skin barrier formation. If cystatin M/E is missing in the skin, unrestrained protease activity is followed by too much and premature activity of TGM3, probably leading to pathology as seen in cystatin M/E deficient mice.

33 Cystatin M/E and cancer Several studies have shown that cysteine proteases such as cathepsins B and L are involved in cancer through their role in proteolytic pathways that enable tumor growth, migration, invasion, and metastasis. This is supported by increased cysteine protease expression in a wide variety of cancers154. The importance of the fine balance between cysteine proteases and their inhibitors in cancer is further underlined by the abnormal expression of cystatins in many types of tumors192. Cystatin M/E was initially identified by differential display as a downregulated mRNA in metastatic breast tumor cells when compared to normal and primary breast tumor cells, suggesting that loss of expression of cystatin M/E is likely associated with the progression of a primary tumor to a metastatic phenotype173. This hypothesis was tested in vitro by transfection of the highly tumorigenic and metastatic human breast cancer cell line MDA‐ MB‐435S, which normally does not express cystatin M/E, with a cystatin M/E cDNA expression vector. The constitutive cystatin M/E expression in transfected cells resulted in a significant reduction of cell proliferation, migration, Matrigel invasion, and endothelial cell adhesion193. However, only cell migration and matrix invasion appeared to be controlled by the inhibition of cysteine protease activity as the synthetic cysteine protease inhibitor E‐64 showed no effect on cell proliferation and endothelial adhesion. This finding made the authors suggest that cystatin M/E may act as a tumor and metastasis suppressor through other mechanisms (than cathepsin inhibition), perhaps via its inhibitory effects on legumain or even through modulation of gene transcription. Cystatin M/E expression in MDA‐MB‐435S cells significantly changed the expression of several genes involved in tumor growth, invasion, and angiogenesis including downregulation of autotaxin, a signaling molecule that has been linked to breast cancer invasiveness194. Subsequently, the transfected MDA‐MB‐ 435S cells were injected into scid mice to test if cystatin M/E expression might have in vivo tumor suppressing functions195. The mice were analyzed for tumorigenesis and spontaneous metastasis in the lungs and compared with mice implanted with mock controls. During the first 45 days, scid mice bearing orthotopically injected cystatin M/E‐expressing MDA‐MB‐ 435S cells developed significantly smaller tumors than mock‐injected controls, however, no difference was found in the incidence of metastasis and number of lesions in the lungs. Cystatin M/E expression showed no effect on seeding and survival of tumor cells, but reduced the growth of established lung tumors, which suggests that cystatin M/E has anti‐ proliferative activity. One should keep in mind that there is some controversy about the origin of the MDA‐MB‐435S cancer cell line, as it was reported that these cells may be derived from melanocytes and as such cannot be considered with absolute certainty as a model for breast cancer196. Preliminary studies using laser‐capture dissection of human breast cancer cells varied in outcome. Zhang and co‐workers reported a downregulation of cystatin M/E mRNA and protein expression in invasive ductal carcinomas compared to normal breast tissue195, which is in contrast to another study that showed cystatin M/E expression in most of the breast carcinomas including primary and metastatic tumors197.

34 Reduced or absent expression of cystatin M/E was also found in basal cell carcinomas (BCC) of the skin, squamous cell carcinomas (SCC) from the lung, and in lung cancer198,199, but in oropharyngaeal SCC cystatin M/E expression was found to be increased with metastasis along with elevated cathepsin expression200. Silencing of cystatin M/E by small interference RNA (siRNA) in a metastatic oral cancer cell line increased in vitro cell proliferation, migration, and Matrigel invasion, emphasizing the tumor and metastasis suppressing abilities of cystatin M/E201. Gene expression profiling of cutaneous SCC and psoriasis, a study designed to compare malignant versus benign keratinocyte hyperproliferation, revealed that cystatin M/E and several proteases were among the genes that were significantly upregulated in SCC202. Overexpression of cystatin M/E was also confirmed in pancreatic ductal adenocarcenoma (PDAC). Silencing of cystatin M/E by siRNA reduced PDAC cell growth, while in a PDAC cell line that normally does not express cystatin M/W, introduced expression of cystatin M/E promoted cell proliferation203. Collectively these data suggest that cystatin M/E could act as a tumor suppressor through various mechanisms of action including inhibition of cysteine protease activity and modulation of gene transcription. However, it is not yet clear what the exact role of cystatin M/E is in tumorigenesis and metastasis as both down‐ and upregulation has been found depending on the type of cancer, suggesting that altered cystatin M/E expression could either have a causative effect or is, alternatively, a protective response. In a previous study we concluded that cystatin M/E expression was not a useful marker to discriminate between benign (keratoacanthoma), pre‐malignant (actinic keratosis), and malignant epidermal neoplasias (BCC, SCC). Cystatin M/E expression in these skin‐derived neoplasias appeared to be limited to the keratinized and well‐differentiated cells. SCC from lung and the head and neck region were found to be negative for cystatin M/E. However, one cutaneous metastasis of a head and neck tumor was positive for cystatin M/E suggesting that for these squamous neoplasias a cutaneous environment is necessary for cystatin M/E expression198.

Epigenetic regulation of cystatin M/E expression Epigenetics is a term that refers to mechanisms that influence gene activity without altering DNA sequence, and it explains how gene activity is regulated in response to environment. For example, in 3‐year‐old identical twins epigenetic differences are hardly detectable, but these increased markedly with age resulting in an increased number of genes that differ in activity between 50‐year‐old twins, although both siblings share the same DNA sequence204. Epigenetic instructions are vital to key processes such as cell differentiation and development, mental health, and ageing, and therefore, are thought to be involved in the pathophsyiology of multifactorial diseases and cancer. Genomic DNA is wrapped in chromatin, which can be structured as decondensed euchromatin, that is transcriptionally active, and heterochromatin, that is highly condensed and becomes inaccessible to transcription factors. Two main epigenetic modifications of DNA exist: histone modification and DNA methylation. Histones are chromatin proteins around which DNA is wrapped and these are tagged with several molecules that influence the structure of chromatin and gene

35 transcription (e.g. histone acetylation is associated with transcriptional activity). DNA methylation involves methylation of the cytosine in CpG dinucleotides, especially those located in CpG islands within gene promoters, and represses transcription by either blocking the binding of specific transcription factors to DNA or by recruiting methyl‐CpG‐binding proteins that are associated with chromatin remodeling205‐209. Recently, epigenetic studies reported DNA methylation‐dependent silencing of the cystatin M/E gene in breast cancer cell lines, primary breast tumors, and metastases. Hypermethylation, especially of the CpG island present in the CST6 proximal promoter region, was significantly associated with gene silencing and loss of cystatin M/E protein expression, while treatment with the demethylating agent 5‐aza‐2’deoxycytidine induced cystatin M/E mRNA expression in breast cancer cell lines182,210,211. Cystatin M/E promoter methylation (and loss of cystatin M/E protein expression) occurred more frequently in lymph node metastases, whereas primary tumors were found to be negative for cystatin M/E promoter methylation. This suggests that epigenetic silencing of cystatin M/E is involved in breast cancer tumorigenesis and progression into metastasis181. Furthermore, it was recently demonstrated that epigenetic silencing of cystatin M/E is frequent in gliomas183.

AIM OF THIS THESIS

At the start of this project studies on cystatin M/E deficient mice had suggested that cystatin M/E was the main inhibitor of legumain in skin since absence of cystatin M/E resulted in free cellular activity of legumain in epidermis and hair follicles. This matched the localization of tissue pathology in these mice such as excessive cornification in hair follicles and epidermis leading to plugging of the hair follicle and ichthyosis. No legumain activity could be demonstrated in the skin of wild type littermates, which suggested that the presence of cystatin M/E normally regulates legumain activity in skin in vivo. However, cystatins are generally considered to be inhibitors of the papain‐like cathepsins, but despite biochemical enzyme assays on cathepsin inhibitory activity of cystatin M/E, the cognate target cathepsin of cystatin M/E was still not identified. Although expression of cystatin M/E has been found in other organs such as kidney, lung, and eye, studies in ichq mice could not indicate if cystatin M/E deficiency has (patho)physiologic consequences in these organs as offspring dies 6‐12 days postnatally from the lethal skin phenotype before other organs are fully developed. The studies performed in this thesis were aimed on investigating the following issues:

I. Identification of target proteases of cystatin M/E, more specifically: which cysteine cathepsin(s) are inhibited by cystatin M/E?

II. Localization and compartmentalization of cystatin M/E and its target proteases in human epidermis and epidermal appendages to obtain clues on their biological relevance and possible functions.

36

III. Elucidation of the role of cystatin M/E in epidermal differentiation. Cathepsins and legumain generally act in protein degradation and processing of other proteases. However, their exact functions and targets in epidermis as well as the biological processes that are regulated and executed by these proteases remained to be solved.

IV. Role of cystatin M/E and its target proteases in skin diseases with impaired barrier formation and function.

V. Function of cystatin M/E in other organs: transgenic expression of cystatin M/E in epidermis could rescue ichq mice from the lethal skin phenotype and might provide the opportunity to study the effects of cystatin M/E deficiency in other organs.

REFERENCES

1. Eady, R. A., Leigh, E. M., Pope, F. M. (1998) Anatomy and organization of human skin. In Textbook of Dermatology (Champion, R. H., Burton, J. L., Burns, D. A., and Breathnach, S. M., eds) pp. 37‐111, Blackwell Science, Inc., Malden. 2. Archer, C. B. (1998) Functions of skin. In Textbook of Dermatology (Champion, R. H., Burton, J. L., Burns, D. A., and Breathnach, S. M., eds) pp. 113‐122, Blackwell Science, Inc., Malden. 3. Alonso, L., Fuchs, E. (2003) Stem cells of the skin epithelium. Proc.Natl.Acad.Sci.U.S.A 100 Suppl 1, 11830‐11835 4. Smack, D. P., Korge, B. P., James, W. D. (1994) Keratin and keratinization. J.Am.Acad.Dermatol. 30, 85‐102 5. Reichert, U., Michel, S., Schmidt, R. (1993) The cornified envelope: a key structure of terminally differentiating keratinocytes. In Molecular Biology of the Skin. The Keratinocyte (Darmon, M. and Blumenberg, M., eds) pp. 107‐179, Academic Press, Inc., San Diego. 6. Fuchs, E., Raghavan, S. (2002) Getting under the skin of epidermal morphogenesis. Nat.Rev.Genet. 3, 199‐209 7. Candi, E., Schmidt, R., Melino, G. (2005) The cornified envelope: a model of cell death in the skin. Nat.Rev.Mol.Cell Biol. 6, 328‐340 8. Segre, J. A. (2006) Epidermal barrier formation and recovery in skin disorders. J.Clin.Invest 116, 1150‐1158 9. Nemes, Z., Steinert, P. M. (1999) Bricks and mortar of the epidermal barrier. Exp.Mol.Med. 31, 5‐19 10. Kalinin, A., Marekov, L. N., Steinert, P. M. (2001) Assembly of the epidermal cornified cell envelope. J.Cell Sci. 114, 3069‐3070 11. Kalinin, A. E., Kajava, A. V., Steinert, P. M. (2002) Epithelial barrier function: assembly and structural features of the cornified cell envelope. Bioessays 24, 789‐800 12. Getsios, S., Huen, A. C., Green, K. J. (2004) Working out the strength and flexibility of desmosomes. Nat.Rev.Mol.Cell Biol. 5, 271‐281

37 13. Green, K. J., Simpson, C. L. (2007) Desmosomes: new perspectives on a classic. J.Invest Dermatol. 127, 2499‐2515 14. Garrod, D. R., Merritt, A. J., Nie, Z. (2002) Desmosomal cadherins. Curr.Opin.Cell Biol. 14, 537‐ 545 15. Kottke, M. D., Delva, E., Kowalczyk, A. P. (2006) The desmosome: cell science lessons from human diseases. J.Cell Sci. 119, 797‐806 16. Elias, P. M., Matsuyoshi, N., Wu, H., Lin, C., Wang, Z. H., Brown, B. E., Stanley, J. R. (2001) Desmoglein isoform distribution affects stratum corneum structure and function. J.Cell Biol. 153, 243‐249 17. Gallicano, G. I., Kouklis, P., Bauer, C., Yin, M., Vasioukhin, V., Degenstein, L., Fuchs, E. (1998) Desmoplakin is required early in development for assembly of desmosomes and cytoskeletal linkage. J.Cell Biol. 143, 2009‐2022 18. Sevilla, L. M., Nachat, R., Groot, K. R., Klement, J. F., Uitto, J., Djian, P., Maatta, A., Watt, F. M. (2007) Mice deficient in involucrin, envoplakin, and periplakin have a defective epidermal barrier. J.Cell Biol. 179, 1599‐1612 19. Simon, M., Jonca, N., Guerrin, M., Haftek, M., Bernard, D., Caubet, C., Egelrud, T., Schmidt, R., Serre, G. (2001) Refined characterization of corneodesmosin proteolysis during terminal differentiation of human epidermis and its relationship to desquamation. J.Biol.Chem. 276, 20292‐20299 20. Jonca, N., Guerrin, M., Hadjiolova, K., Caubet, C., Gallinaro, H., Simon, M., Serre, G. (2002) Corneodesmosin, a component of epidermal corneocyte desmosomes, displays homophilic adhesive properties. J.Biol.Chem. 277, 5024‐5029 21. Elias, P. M. (2005) Stratum corneum defensive functions: an integrated view. J.Invest Dermatol. 125, 183‐200 22. Caubet, C., Jonca, N., Brattsand, M., Guerrin, M., Bernard, D., Schmidt, R., Egelrud, T., Simon, M., Serre, G. (2004) Degradation of corneodesmosome proteins by two serine proteases of the kallikrein family, SCTE/KLK5/hK5 and SCCE/KLK7/hK7. J.Invest Dermatol. 122, 1235‐1244 23. Horikoshi, T., Igarashi, S., Uchiwa, H., Brysk, H., Brysk, M. M. (1999) Role of endogenous cathepsin D‐like and chymotrypsin‐like proteolysis in human epidermal desquamation. Br.J.Dermatol. 141, 453‐459 24. Igarashi, S., Takizawa, T., Takizawa, T., Yasuda, Y., Uchiwa, H., Hayashi, S., Brysk, H., Robinson, J. M., Yamamoto, K., Brysk, M. M., Horikoshi, T. (2004) Cathepsin D, but not cathepsin E, degrades desmosomes during epidermal desquamation. Br.J.Dermatol. 151, 355‐361 25. Watkinson, A. (1999) Stratum corneum thiol protease (SCTP): a novel cysteine protease of late epidermal differentiation. Arch.Dermatol.Res. 291, 260‐268 26. Rawlings, N. D., Barrett, A. J., Woessner, J. F. (2004) Handbook of Proteolytic Enzymes.Elsevier Academic Press, London. 27. Lopez‐Otin, C., Bond, J. S. (2008) Proteases: multifunctional enzymes in life and disease. J.Biol.Chem. 283, 30433‐30437 28. Hansen, K. K., Oikonomopoulou, K., Baruch, A., Ramachandran, R., Beck, P., Diamandis, E. P., Hollenberg, M. D. (2008) Proteinases as hormones: targets and mechanisms for proteolytic signaling. Biol.Chem. 389, 971‐982 29. Rawlings, N. D., Morton, F. R., Kok, C. Y., Kong, J., Barrett, A. J. (2008) MEROPS: the peptidase database. Nucleic Acids Res. 36, D320‐D325

38 30. James, M. N., Sielecki, A. R. (1985) Stereochemical analysis of peptide bond hydrolysis catalyzed by the aspartic proteinase penicillopepsin. Biochemistry 24, 3701‐3713 31. James, M. N., Sielecki, A. R. (1983) Structure and refinement of penicillopepsin at 1.8 A resolution. J.Mol.Biol. 163, 299‐361 32. Richter, C., Tanaka, T., Yada, R. Y. (1998) Mechanism of activation of the gastric aspartic proteinases: pepsinogen, progastricsin and prochymosin. Biochem.J. 335 (Pt 3), 481‐490 33. Egberts, F., Heinrich, M., Jensen, J. M., Winoto‐Morbach, S., Pfeiffer, S., Wickel, M., Schunck, M., Steude, J., Saftig, P., Proksch, E., Schutze, S. (2004) Cathepsin D is involved in the regulation of transglutaminase 1 and epidermal differentiation. J.Cell Sci. 117, 2295‐2307 34. Liaudet‐Coopman, E., Beaujouin, M., Derocq, D., Garcia, M., Glondu‐Lassis, M., Laurent‐ Matha, V., Prebois, C., Rochefort, H., Vignon, F. (2006) Cathepsin D: newly discovered functions of a long‐standing aspartic protease in cancer and apoptosis. Cancer Lett. 237, 167‐ 179 35. Zaidi, N., Hermann, C., Herrmann, T., Kalbacher, H. (2008) Emerging functional roles of cathepsin E. Biochem.Biophys.Res.Commun. 377, 327‐330 36. Cole, S. L., Vassar, R. (2008) The role of amyloid precursor protein processing by BACE1, the beta‐secretase, in Alzheimer disease pathophysiology. J.Biol.Chem. 283, 29621‐29625 37. Sarzani, R., Salvi, F., Dessi‐Fulgheri, P., Rappelli, A. (2008) Renin‐angiotensin system, natriuretic peptides, obesity, metabolic syndrome, and hypertension: an integrated view in humans. J.Hypertens. 26, 831‐843 38. Weber, I. T., Miller, M., Jaskolski, M., Leis, J., Skalka, A. M., Wlodawer, A. (1989) Molecular modeling of the HIV‐1 protease and its substrate binding site. Science 243, 928‐931 39. Fujinaga, M., Cherney, M. M., Oyama, H., Oda, K., James, M. N. (2004) The molecular structure and catalytic mechanism of a novel carboxyl peptidase from Scytalidium lignicolum. Proc.Natl.Acad.Sci.U.S.A 101, 3364‐3369 40. Yabuki, Y., Kubota, K., Kojima, M., Inoue, H., Takahashi, K. (2004) Identification of a glutamine residue essential for catalytic activity of aspergilloglutamic peptidase by site‐directed mutagenesis. FEBS Lett. 569, 161‐164 41. Jongeneel, C. V., Bouvier, J., Bairoch, A. (1989) A unique signature identifies a family of zinc‐ dependent metallopeptidases. FEBS Lett. 242, 211‐214 42. Stocker, W., Grams, F., Baumann, U., Reinemer, P., Gomis‐Ruth, F. X., McKay, D. B., Bode, W. (1995) The metzincins‐‐topological and sequential relations between the astacins, adamalysins, serralysins, and matrixins (collagenases) define a superfamily of zinc‐ peptidases. Protein Sci. 4, 823‐840 43. Nagase, H., Visse, R., Murphy, G. (2006) Structure and function of matrix metalloproteinases and TIMPs. Cardiovasc.Res. 69, 562‐573 44. Page‐McCaw, A., Ewald, A. J., Werb, Z. (2007) Matrix metalloproteinases and the regulation of tissue remodelling. Nat.Rev.Mol.Cell Biol. 8, 221‐233 45. Warshel, A., Naray‐Szabo, G., Sussman, F., Hwang, J. K. (1989) How do serine proteases really work? Biochemistry 28, 3629‐3637 46. Rawlings, N. D., Barrett, A. J. (2004) Introduction: serine peptidases and their clans. In Handbook of Proteolytic Enzymes, Second Edition (Barrett, A. J., Rawlings, N. D., and Woessner, J. F., eds) pp. 1417‐1439, Elsevier Academic Press, London. 47. Bots, M., Medema, J. P. (2006) Granzymes at a glance. J.Cell Sci. 119, 5011‐5014

39 48. Hollenberg, M. D., Oikonomopoulou, K., Hansen, K. K., Saifeddine, M., Ramachandran, R., Diamandis, E. P. (2008) Kallikreins and proteinase‐mediated signaling: proteinase‐activated receptors (PARs) and the pathophysiology of inflammatory diseases and cancer. Biol.Chem. 389, 643‐651 49. List, K., Bugge, T. H., Szabo, R. (2006) Matriptase: potent proteolysis on the cell surface. Mol.Med. 12, 1‐7 50. Bogyo, M., McMaster, J. S., Gaczynska, M., Tortorella, D., Goldberg, A. L., Ploegh, H. (1997) Covalent modification of the active site threonine of proteasomal beta subunits and the Escherichia coli homolog HslV by a new class of inhibitors. Proc.Natl.Acad.Sci.U.S.A 94, 6629‐ 6634 51. Lowe, J., Stock, D., Jap, B., Zwickl, P., Baumeister, W., Huber, R. (1995) Crystal structure of the 20S proteasome from the archaeon T. acidophilum at 3.4 A resolution. Science 268, 533‐ 539 52. Goldberg, A. L. (2003) Protein degradation and protection against misfolded or damaged proteins. Nature 426, 895‐899 53. Baumeister, W., Walz, J., Zuhl, F., Seemuller, E. (1998) The proteasome: paradigm of a self‐ compartmentalizing protease. Cell 92, 367‐380 54. Seemuller, E., Lupas, A., Stock, D., Lowe, J., Huber, R., Baumeister, W. (1995) Proteasome from Thermoplasma acidophilum: a threonine protease. Science 268, 579‐582 55. Rawlings, N. D., Barrett, A. J. (2004) Introduction: the clans and families of cysteine peptidases. In Handbook of Proteolytic Enzymes, Second Edition (Barrett, A. J., Rawlings, N. D., and Woessner, J. F., eds) pp. 1051‐1071, Elsevier Academic Press, London. 56. Polgar, L. (2004) Catalytic mechanisms of cysteine peptidases. In Handbook of Proteolytic Enzymes, Second Edition (Barrett, A. J., Rawlings, N. D., and Woessner, J. F., eds) pp. 1072‐ 1079, Elsevier Academic Press, London. 57. Fanin, M., Nascimbeni, A. C., Fulizio, L., Trevisan, C. P., Meznaric‐Petrusa, M., Angelini, C. (2003) Loss of calpain‐3 autocatalytic activity in LGMD2A patients with normal protein expression. Am.J.Pathol. 163, 1929‐1936 58. Vosler, P. S., Brennan, C. S., Chen, J. (2008) Calpain‐mediated signaling mechanisms in neuronal injury and neurodegeneration. Mol.Neurobiol. 38, 78‐100 59. Yuan, J., Horvitz, H. R. (2004) A first insight into the molecular mechanisms of apoptosis. Cell 116, S53‐6, 1 60. Dickinson, D. P. (2002) Cysteine peptidases of mammals: their biological roles and potential effects in the oral cavity and other tissues in health and disease. Crit Rev.Oral Biol.Med. 13, 238‐275 61. Stoka, V., Turk, V., Turk, B. (2007) Lysosomal cysteine cathepsins: signaling pathways in apoptosis. Biol.Chem. 388, 555‐560 62. Chwieralski, C. E., Welte, T., Buhling, F. (2006) Cathepsin‐regulated apoptosis. Apoptosis. 11, 143‐149 63. Karrer, K. M., Peiffer, S. L., DiTomas, M. E. (1993) Two distinct gene subfamilies within the family of cysteine protease genes. Proc.Natl.Acad.Sci.U.S.A 90, 3063‐3067 64. Vernet, T., Berti, P. J., de Montigny, C., Musil, R., Tessier, D. C., Menard, R., Magny, M. C., Storer, A. C., Thomas, D. Y. (1995) Processing of the papain precursor. The ionization state of a conserved amino acid motif within the Pro region participates in the regulation of intramolecular processing. J.Biol.Chem. 270, 10838‐10846

40 65. Dolenc, I., Turk, B., Pungercic, G., Ritonja, A., Turk, V. (1995) Oligomeric structure and substrate induced inhibition of human cathepsin C. J.Biol.Chem. 270, 21626‐21631 66. McIntyre, G. F., Godbold, G. D., Erickson, A. H. (1994) The pH‐dependent membrane association of procathepsin L is mediated by a 9‐residue sequence within the propeptide. J.Biol.Chem. 269, 567‐572 67. Tao, K., Stearns, N. A., Dong, J., Wu, Q. L., Sahagian, G. G. (1994) The proregion of cathepsin L is required for proper folding, stability, and ER exit. Arch.Biochem.Biophys. 311, 19‐27 68. Chapman, H. A., Riese, R. J., Shi, G. P. (1997) Emerging roles for cysteine proteases in human biology. Annu.Rev.Physiol 59, 63‐88 69. Brix, K., Dunkhorst, A., Mayer, K., Jordans, S. (2008) Cysteine cathepsins: cellular roadmap to different functions. Biochimie 90, 194‐207 70. Fox, T., de Miguel, E., Mort, J. S., Storer, A. C. (1992) Potent slow‐binding inhibition of cathepsin B by its propeptide. Biochemistry 31, 12571‐12576 71. Carmona, E., Dufour, E., Plouffe, C., Takebe, S., Mason, P., Mort, J. S., Menard, R. (1996) Potency and selectivity of the cathepsin L propeptide as an inhibitor of cysteine proteases. Biochemistry 35, 8149‐8157 72. Billington, C. J., Mason, P., Magny, M. C., Mort, J. S. (2000) The slow‐binding inhibition of by its propeptide. Biochem.Biophys.Res.Commun. 276, 924‐929 73. Guay, J., Falgueyret, J. P., Ducret, A., Percival, M. D., Mancini, J. A. (2000) Potency and selectivity of inhibition of cathepsin K, L and S by their respective propeptides. Eur.J.Biochem. 267, 6311‐6318 74. Burden, R. E., Snoddy, P., Jefferies, C. A., Walker, B., Scott, C. J. (2007) Inhibition of cathepsin L‐like proteases by cathepsin V propeptide. Biol.Chem. 388, 541‐545 75. Mach, L., Mort, J. S., Glossl, J. (1994) Maturation of human procathepsin B. Proenzyme activation and proteolytic processing of the precursor to the mature proteinase, in vitro, are primarily unimolecular processes. J.Biol.Chem. 269, 13030‐13035 76. Rowan, A. D., Mason, P., Mach, L., Mort, J. S. (1992) Rat procathepsin B. Proteolytic processing to the mature form in vitro. J.Biol.Chem. 267, 15993‐15999 77. Menard, R., Carmona, E., Takebe, S., Dufour, E., Plouffe, C., Mason, P., Mort, J. S. (1998) Autocatalytic processing of recombinant human procathepsin L. Contribution of both intermolecular and unimolecular events in the processing of procathepsin L in vitro. J.Biol.Chem. 273, 4478‐4484 78. Nishimura, Y., Kawabata, T., Furuno, K., Kato, K. (1989) Evidence that aspartic proteinase is involved in the proteolytic processing event of procathepsin L in lysosomes. Arch.Biochem.Biophys. 271, 400‐406 79. Ishidoh, K., Kominami, E. (2002) Processing and activation of lysosomal proteinases. Biol.Chem. 383, 1827‐1831 80. Coulombe, R., Grochulski, P., Sivaraman, J., Menard, R., Mort, J. S., Cygler, M. (1996) Structure of human procathepsin L reveals the molecular basis of inhibition by the prosegment. EMBO J. 15, 5492‐5503 81. Quraishi, O., Storer, A. C. (2001) Identification of internal autoproteolytic cleavage sites within the prosegments of recombinant procathepsin B and procathepsin S. Contribution of a plausible unimolecular autoproteolytic event for the processing of zymogens belonging to the papain family. J.Biol.Chem. 276, 8118‐8124

41 82. Ishidoh, K., Saido, T. C., Kawashima, S., Hirose, M., Watanabe, S., Sato, N., Kominami, E. (1998) Multiple processing of procathepsin L to cathepsin L in vivo. Biochem.Biophys.Res.Commun. 252, 202‐207 83. Linke, M., Herzog, V., Brix, K. (2002) Trafficking of lysosomal cathepsin B‐green fluorescent protein to the surface of thyroid epithelial cells involves the endosomal/lysosomal compartment. J.Cell Sci. 115, 4877‐4889 84. Buth, H., Wolters, B., Hartwig, B., Meier‐Bornheim, R., Veith, H., Hansen, M., Sommerhoff, C. P., Schaschke, N., Machleidt, W., Fusenig, N. E., Boukamp, P., Brix, K. (2004) HaCaT keratinocytes secrete lysosomal cysteine proteinases during migration. Eur.J.Cell Biol. 83, 781‐795 85. Drake, F. H., Dodds, R. A., James, I. E., Connor, J. R., Debouck, C., Richardson, S., Lee‐ Rykaczewski, E., Coleman, L., Rieman, D., Barthlow, R., Hastings, G., Gowen, M. (1996) Cathepsin K, but not cathepsins B, L, or S, is abundantly expressed in human osteoclasts. J.Biol.Chem. 271, 12511‐12516 86. Andrews, N. W. (2000) Regulated secretion of conventional lysosomes. Trends Cell Biol. 10, 316‐321 87. Von Figura, K. (1991) Molecular recognition and targeting of lysosomal proteins. Curr.Opin.Cell Biol. 3, 642‐646 88. Ludwig, T., Munier‐Lehmann, H., Bauer, U., Hollinshead, M., Ovitt, C., Lobel, P., Hoflack, B. (1994) Differential sorting of lysosomal enzymes in mannose 6‐phosphate receptor‐deficient fibroblasts. EMBO J. 13, 3430‐3437 89. Mach, L., Mort, J. S., Glossl, J. (1994) Noncovalent complexes between the lysosomal proteinase cathepsin B and its propeptide account for stable, extracellular, high molecular mass forms of the enzyme. J.Biol.Chem. 269, 13036‐13040 90. Ferri, K. F., Kroemer, G. (2001) Organelle‐specific initiation of cell death pathways. Nat.Cell Biol. 3, E255‐E263 91. Muntener, K., Zwicky, R., Csucs, G., Rohrer, J., Baici, A. (2004) Exon skipping of cathepsin B: mitochondrial targeting of a lysosomal peptidase provokes cell death. J.Biol.Chem. 279, 41012‐41017 92. Goulet, B., Baruch, A., Moon, N. S., Poirier, M., Sansregret, L. L., Erickson, A., Bogyo, M., Nepveu, A. (2004) A cathepsin L isoform that is devoid of a signal peptide localizes to the nucleus in S phase and processes the CDP/Cux transcription factor. Mol.Cell 14, 207‐219 93. Goulet, B., Sansregret, L., Leduy, L., Bogyo, M., Weber, E., Chauhan, S. S., Nepveu, A. (2007) Increased expression and activity of nuclear cathepsin L in cancer cells suggests a novel mechanism of cell transformation. Mol.Cancer Res. 5, 899‐907 94. Chauhan, S. S., Popescu, N. C., Ray, D., Fleischmann, R., Gottesman, M. M., Troen, B. R. (1993) Cloning, genomic organization, and chromosomal localization of human cathepsin L. J.Biol.Chem. 268, 1039‐1045 95. Jean, D., Guillaume, N., Frade, R. (2002) Characterization of human cathepsin L promoter and identification of binding sites for NF‐Y, Sp1 and Sp3 that are essential for its activity. Biochem.J. 361, 173‐184 96. Ritonja, A., Popovic, T., Kotnik, M., Machleidt, W., Turk, V. (1988) Amino acid sequences of the human kidney cathepsins H and L. FEBS Lett. 228, 341‐345

42 97. Bromme, D., Bonneau, P. R., Lachance, P., Storer, A. C. (1994) Engineering the S2 subsite specificity of human cathepsin S to a cathepsin L‐ and cathepsin B‐like specificity. J.Biol.Chem. 269, 30238‐30242 98. Portaro, F. C., Santos, A. B., Cezari, M. H., Juliano, M. A., Juliano, L., Carmona, E. (2000) Probing the specificity of cysteine proteinases at subsites remote from the active site: analysis of P4, P3, P2' and P3' variations in extended substrates. Biochem.J. 347 Pt 1, 123‐129 99. Barrett, A. J., Kirschke, H. (1981) Cathepsin B, , and cathepsin L. Methods Enzymol. 80 Pt C, 535‐561 100. Mason, R. W., Green, G. D., Barrett, A. J. (1985) Human liver cathepsin L. Biochem.J. 226, 233‐241 101. Turk, B., Dolenc, I., Turk, V., Bieth, J. G. (1993) Kinetics of the pH‐induced inactivation of human cathepsin L. Biochemistry 32, 375‐380 102. Abrahamson, M., Alvarez‐Fernandez, M., Nathanson, C. M. (2003) Cystatins. Biochem.Soc.Symp. 179‐199 103. Barrett, A. J., Kembhavi, A. A., Brown, M. A., Kirschke, H., Knight, C. G., Tamai, M., Hanada, K. (1982) L‐trans‐Epoxysuccinyl‐leucylamido(4‐guanidino)butane (E‐64) and its analogues as inhibitors of cysteine proteinases including cathepsins B, H and L. Biochem.J. 201, 189‐198 104. Wille, A., Gerber, A., Heimburg, A., Reisenauer, A., Peters, C., Saftig, P., Reinheckel, T., Welte, T., Buhling, F. (2004) Cathepsin L is involved in cathepsin D processing and regulation of apoptosis in A549 human lung epithelial cells. Biol.Chem. 385, 665‐670 105. Laurent‐Matha, V., Derocq, D., Prebois, C., Katunuma, N., Liaudet‐Coopman, E. (2006) Processing of human cathepsin D is independent of its catalytic function and auto‐activation: involvement of cathepsins L and B. J.Biochem. 139, 363‐371 106. Sorkin, A., von Zastrow, M. (2002) Signal transduction and endocytosis: close encounters of many kinds. Nat.Rev.Mol.Cell Biol. 3, 600‐614 107. Zwad, O., Kubler, B., Roth, W., Scharf, J. G., Saftig, P., Peters, C., Braulke, T. (2002) Decreased intracellular degradation of insulin‐like growth factor binding protein‐3 in cathepsin L‐ deficient fibroblasts. FEBS Lett. 510, 211‐215 108. Kane, S. E., Gottesman, M. M. (1990) The role of cathepsin L in malignant transformation. Semin.Cancer Biol. 1, 127‐136 109. Maciewicz, R. A., Wotton, S. F., Etherington, D. J., Duance, V. C. (1990) Susceptibility of the cartilage collagens types II, IX and XI to degradation by the cysteine proteinases, cathepsins B and L. FEBS Lett. 269, 189‐193 110. Urbich, C., Heeschen, C., Aicher, A., Sasaki, K., Bruhl, T., Farhadi, M. R., Vajkoczy, P., Hofmann, W. K., Peters, C., Pennacchio, L. A., Abolmaali, N. D., Chavakis, E., Reinheckel, T., Zeiher, A. M., Dimmeler, S. (2005) Cathepsin L is required for endothelial progenitor cell‐ induced neovascularization. Nat.Med. 11, 206‐213 111. Roth, W., Deussing, J., Botchkarev, V. A., Pauly‐Evers, M., Saftig, P., Hafner, A., Schmidt, P., Schmahl, W., Scherer, J., Anton‐Lamprecht, I., Von Figura, K., Paus, R., Peters, C. (2000) Cathepsin L deficiency as molecular defect of furless: hyperproliferation of keratinocytes and pertubation of hair follicle cycling. FASEB J. 14, 2075‐2086 112. Benavides, F., Starost, M. F., Flores, M., Gimenez‐Conti, I. B., Guenet, J. L., Conti, C. J. (2002) Impaired hair follicle morphogenesis and cycling with abnormal epidermal differentiation in nackt mice, a cathepsin L‐deficient mutation. Am.J.Pathol. 161, 693‐703

43 113. Tobin, D. J., Foitzik, K., Reinheckel, T., Mecklenburg, L., Botchkarev, V. A., Peters, C., Paus, R. (2002) The lysosomal protease cathepsin L is an important regulator of keratinocyte and melanocyte differentiation during hair follicle morphogenesis and cycling. Am.J.Pathol. 160, 1807‐1821 114. Reinheckel, T., Hagemann, S., Dollwet‐Mack, S., Martinez, E., Lohmuller, T., Zlatkovic, G., Tobin, D. J., Maas‐Szabowski, N., Peters, C. (2005) The lysosomal cysteine protease cathepsin L regulates keratinocyte proliferation by control of growth factor recycling. J.Cell Sci. 118, 3387‐3395 115. Stypmann, J., Glaser, K., Roth, W., Tobin, D. J., Petermann, I., Matthias, R., Monnig, G., Haverkamp, W., Breithardt, G., Schmahl, W., Peters, C., Reinheckel, T. (2002) Dilated cardiomyopathy in mice deficient for the lysosomal cysteine peptidase cathepsin L. Proc.Natl.Acad.Sci.U.S.A 99, 6234‐6239 116. Petermann, I., Mayer, C., Stypmann, J., Biniossek, M. L., Tobin, D. J., Engelen, M. A., Dandekar, T., Grune, T., Schild, L., Peters, C., Reinheckel, T. (2006) Lysosomal, cytoskeletal, and metabolic alterations in cardiomyopathy of cathepsin L knockout mice. FASEB J. 20, 1266‐1268 117. Spira, D., Stypmann, J., Tobin, D. J., Petermann, I., Mayer, C., Hagemann, S., Vasiljeva, O., Gunther, T., Schule, R., Peters, C., Reinheckel, T. (2007) Cell type‐specific functions of the lysosomal protease cathepsin L in the heart. J.Biol.Chem. 282, 37045‐37052 118. Nakagawa, T., Roth, W., Wong, P., Nelson, A., Farr, A., Deussing, J., Villadangos, J. A., Ploegh, H., Peters, C., Rudensky, A. Y. (1998) Cathepsin L: critical role in Ii degradation and CD4 T cell selection in the thymus. Science 280, 450‐453 119. Benavides, F., Venables, A., Poetschke, K. H., Glasscock, E., Rudensky, A., Gomez, M., Martin, P. N., Guenet, J. L., Richie, E. R., Conti, C. J. (2001) The CD4 T cell‐deficient mouse mutation nackt (nkt) involves a deletion in the cathepsin L (CtsI) gene. Immunogenetics 53, 233‐242 120. Yang, M., Zhang, Y., Pan, J., Sun, J., Liu, J., Libby, P., Sukhova, G. K., Doria, A., Katunuma, N., Peroni, O. D., Guerre‐Millo, M., Kahn, B. B., Clement, K., Shi, G. P. (2007) Cathepsin L activity controls adipogenesis and glucose tolerance. Nat.Cell Biol. 9, 970‐977 121. Benavides, F., Giordano, M., Fiette, L., Bueno Brunialti, A. L., Martin, P. N., Vanzulli, S., Baldi, P., Schmidt, R., Dosne, P. C., Guenet, J. (1999) Nackt (nkt), a new hair loss mutation of the mouse with associated CD4 deficiency. Immunogenetics 49, 413‐419 122. Nishimura, F., Naruishi, H., Naruishi, K., Yamada, T., Sasaki, J., Peters, C., Uchiyama, Y., Murayama, Y. (2002) Cathepsin‐L, a key molecule in the pathogenesis of drug‐induced and I‐ cell disease‐mediated gingival overgrowth: a study with cathepsin‐L‐deficient mice. Am.J.Pathol. 161, 2047‐2052 123. Potts, W., Bowyer, J., Jones, H., Tucker, D., Freemont, A. J., Millest, A., Martin, C., Vernon, W., Neerunjun, D., Slynn, G., Harper, F., Maciewicz, R. (2004) Cathepsin L‐deficient mice exhibit abnormal skin and bone development and show increased resistance to osteoporosis following ovariectomy. Int.J.Exp.Pathol. 85, 85‐96 124. Wright, W. W., Smith, L., Kerr, C., Charron, M. (2003) Mice that express enzymatically inactive cathepsin L exhibit abnormal spermatogenesis. Biol.Reprod. 68, 680‐687 125. Adachi, W., Kawamoto, S., Ohno, I., Nishida, K., Kinoshita, S., Matsubara, K., Okubo, K. (1998) Isolation and characterization of human cathepsin V: a major proteinase in corneal epithelium. Invest Ophthalmol.Vis.Sci. 39, 1789‐1796

44 126. Brömme, D., Li, Z., Barnes, M., Mehler, E. (1999) Human cathepsin V functional expression, tissue distribution, electrostatic surface potential, enzymatic characterization, and chromosomal localization. Biochemistry 38, 2377‐2385 127. Santamaria, I., Velasco, G., Cazorla, M., Fueyo, A., Campo, E., Lopez‐Otin, C. (1998) Cathepsin L2, a novel human cysteine proteinase produced by breast and colorectal carcinomas. Cancer Res. 58, 1624‐1630 128. Itoh, R., Kawamoto, S., Adachi, W., Kinoshita, S., Okubo, K. (1999) Genomic organization and chromosomal localization of the human cathepsin L2 gene. DNA Res. 6, 137‐140 129. Tolosa, E., Li, W., Yasuda, Y., Wienhold, W., Denzin, L. K., Lautwein, A., Driessen, C., Schnorrer, P., Weber, E., Stevanovic, S., Kurek, R., Melms, A., Bromme, D. (2003) Cathepsin V is involved in the degradation of invariant chain in human thymus and is overexpressed in myasthenia gravis. J.Clin.Invest 112, 517‐526 130. Somoza, J. R., Zhan, H., Bowman, K. K., Yu, L., Mortara, K. D., Palmer, J. T., Clark, J. M., McGrath, M. E. (2000) Crystal structure of human cathepsin V. Biochemistry 39, 12543‐12551 131. Puzer, L., Cotrin, S. S., Alves, M. F., Egborge, T., Araujo, M. S., Juliano, M. A., Juliano, L., Bromme, D., Carmona, A. K. (2004) Comparative substrate specificity analysis of recombinant human cathepsin V and cathepsin L. Arch.Biochem.Biophys. 430, 274‐283 132. McDonald, J. K., Culbertson, J. T., Owers, N. O. (1993) Identification and localization of a novel cathepsin S‐like proteinase in guinea pig spermatozoa. Arch.Biochem.Biophys. 305, 1‐8 133. Bernard, D., Mehul, B., Thomas‐Collignon, A., Simonetti, L., Remy, V., Bernard, M. A., Schmidt, R. (2003) Analysis of proteins with caseinolytic activity in a human stratum corneum extract revealed a yet unidentified cysteine protease and identified the so‐called "stratum corneum thiol protease" as cathepsin L2. J.Invest Dermatol. 120, 592‐600 134. Hagemann, S., Gunther, T., Dennemarker, J., Lohmuller, T., Bromme, D., Schule, R., Peters, C., Reinheckel, T. (2004) The human cysteine protease cathepsin V can compensate for murine cathepsin L in mouse epidermis and hair follicles. Eur.J.Cell Biol. 83, 775‐780 135. Chen, N., Seiberg, M., Lin, C. B. (2006) Cathepsin L2 levels inversely correlate with skin color. J.Invest Dermatol. 126, 2345‐2347 136. Ma, D. H., Yao, J. Y., Kuo, M. T., See, L. C., Lin, K. Y., Chen, S. C., Chen, J. K., Chao, A. S., Wang, S. F., Lin, K. K. (2007) Generation of endostatin by matrix metalloproteinase and cathepsin from human limbocorneal epithelial cells cultivated on amniotic membrane. Invest Ophthalmol.Vis.Sci. 48, 644‐651 137. Puzer, L., Barros, N. M., Paschoalin, T., Hirata, I. Y., Tanaka, A. S., Oliveira, M. C., Bromme, D., Carmona, A. K. (2008) Cathepsin V, but not cathepsins L, B and K, may release angiostatin‐like fragments from plasminogen. Biol.Chem. 389, 195‐200 138. Yasuda, Y., Li, Z., Greenbaum, D., Bogyo, M., Weber, E., Bromme, D. (2004) Cathepsin V, a novel and potent elastolytic activity expressed in activated macrophages. J.Biol.Chem. 279, 36761‐36770 139. Helske, S., Syvaranta, S., Lindstedt, K. A., Lappalainen, J., Oorni, K., Mayranpaa, M. I., Lommi, J., Turto, H., Werkkala, K., Kupari, M., Kovanen, P. T. (2006) Increased expression of elastolytic cathepsins S, K, and V and their inhibitor cystatin C in stenotic aortic valves. Arterioscler.Thromb.Vasc.Biol. 26, 1791‐1798 140. Chen, J. M., Dando, P. M., Rawlings, N. D., Brown, M. A., Young, N. E., Stevens, R. A., Hewitt, E., Watts, C., Barrett, A. J. (1997) Cloning, isolation, and characterization of mammalian legumain, an asparaginyl endopeptidase. J.Biol.Chem. 272, 8090‐8098

45 141. Dando, P. M., Fortunato, M., Smith, L., Knight, C. G., McKendrick, J. E., Barrett, A. J. (1999) Pig kidney legumain: an asparaginyl endopeptidase with restricted specificity. Biochem.J. 339 (Pt 3), 743‐749 142. Chen, J. M., Fortunato, M., Barrett, A. J. (2000) Activation of human prolegumain by cleavage at a C‐terminal asparagine residue. Biochem.J. 352 Pt 2, 327‐334 143. Li, D. N., Matthews, S. P., Antoniou, A. N., Mazzeo, D., Watts, C. (2003) Multistep autoactivation of asparaginyl endopeptidase in vitro and in vivo. J.Biol.Chem. 278, 38980‐ 38990 144. Alvarez‐Fernandez, M., Barrett, A. J., Gerhartz, B., Dando, P. M., Ni, J., Abrahamson, M. (1999) Inhibition of mammalian legumain by some cystatins is due to a novel second reactive site. J.Biol.Chem. 274, 19195‐19203 145. Morita, Y., Araki, H., Sugimoto, T., Takeuchi, K., Yamane, T., Maeda, T., Yamamoto, Y., Nishi, K., Asano, M., Shirahama‐Noda, K., Nishimura, M., Uzu, T., Hara‐Nishimura, I., Koya, D., Kashiwagi, A., Ohkubo, I. (2007) Legumain/asparaginyl endopeptidase controls extracellular matrix remodeling through the degradation of fibronectin in mouse renal proximal tubular cells. FEBS Lett. 581, 1417‐1424 146. Shirahama‐Noda, K., Yamamoto, A., Sugihara, K., Hashimoto, N., Asano, M., Nishimura, M., Hara‐Nishimura, I. (2003) Biosynthetic processing of cathepsins and lysosomal degradation are abolished in asparaginyl endopeptidase‐deficient mice. J.Biol.Chem. 278, 33194‐33199 147. Maehr, R., Hang, H. C., Mintern, J. D., Kim, Y. M., Cuvillier, A., Nishimura, M., Yamada, K., Shirahama‐Noda, K., Hara‐Nishimura, I., Ploegh, H. L. (2005) is not essential for class II MHC antigen presentation but is required for processing of cathepsin L in mice. J.Immunol. 174, 7066‐7074 148. Manoury, B., Hewitt, E. W., Morrice, N., Dando, P. M., Barrett, A. J., Watts, C. (1998) An asparaginyl endopeptidase processes a microbial antigen for class II MHC presentation. Nature 396, 695‐699 149. Watts, C., Matthews, S. P., Mazzeo, D., Manoury, B., Moss, C. X. (2005) Asparaginyl endopeptidase: case history of a class II MHC compartment protease. Immunol.Rev. 207, 218‐228 150. Chan, C. B., Abe, M., Hashimoto, N., Hao, C., Williams, I. R., Liu, X., Nakao, S., Yamamoto, A., Li, S. Y., Hara‐Nishimura, I., Asano, M., Ye, K. (2008) Mice lacking asparaginyl endopeptidase develop disorders resembling hemophagocytic syndrome. Proc.Natl.Acad.Sci.U.S.A 151. Chen, J. M., Fortunato, M., Stevens, R. A., Barrett, A. J. (2001) Activation of progelatinase A by mammalian legumain, a recently discovered cysteine proteinase. Biol.Chem. 382, 777‐783 152. Liu, C., Sun, C., Huang, H., Janda, K., Edgington, T. (2003) Overexpression of legumain in tumors is significant for invasion/metastasis and a candidate enzymatic target for prodrug therapy. Cancer Res. 63, 2957‐2964 153. Murthy, R. V., Arbman, G., Gao, J., Roodman, G. D., Sun, X. F. (2005) Legumain expression in relation to clinicopathologic and biological variables in colorectal cancer. Clin.Cancer Res. 11, 2293‐2299 154. Mohamed, M. M., Sloane, B. F. (2006) Cysteine cathepsins: multifunctional enzymes in cancer. Nat.Rev.Cancer 6, 764‐775 155. Turk, B., Turk, D., Salvesen, G. S. (2002) Regulating cysteine protease activity: essential role of protease inhibitors as guardians and regulators. Curr.Pharm.Des 8, 1623‐1637

46 156. Kawada, A., Hara, K., Kominami, E., Hiruma, M., Noguchi, H., Ishibashi, A. (1997) Processing of cathepsins L, B and D in psoriatic epidermis. Arch.Dermatol.Res. 289, 87‐93 157. Tanabe, H., Kumagai, N., Tsukahara, T., Ishiura, S., Kominami, E., Nishina, H., Sugita, H. (1991) Changes of lysosomal proteinase activities and their expression in rat cultured keratinocytes during differentiation. Biochim.Biophys.Acta 1094, 281‐287 158. Lippens, S., Denecker, G., Ovaere, P., Vandenabeele, P., Declercq, W. (2005) Death penalty for keratinocytes: apoptosis versus cornification. Cell Death Differ. 12 Suppl 2, 1497‐1508 159. Zeeuwen, P. L. (2004) Epidermal differentiation: the role of proteases and their inhibitors. Eur.J.Cell Biol. 83, 761‐773 160. Chavanas, S., Bodemer, C., Rochat, A., Hamel‐Teillac, D., Ali, M., Irvine, A. D., Bonafe, J. L., Wilkinson, J., Taieb, A., Barrandon, Y., Harper, J. I., de Prost, Y., Hovnanian, A. (2000) Mutations in SPINK5, encoding a serine protease inhibitor, cause Netherton syndrome. Nat.Genet. 25, 141‐142 161. Toomes, C., James, J., Wood, A. J., Wu, C. L., McCormick, D., Lench, N., Hewitt, C., Moynihan, L., Roberts, E., Woods, C. G., Markham, A., Wong, M., Widmer, R., Ghaffar, K. A., Pemberton, M., Hussein, I. R., Temtamy, S. A., Davies, R., Read, A. P., Sloan, P., Dixon, M. J., Thakker, N. S. (1999) Loss‐of‐function mutations in the cathepsin C gene result in periodontal disease and palmoplantar keratosis. Nat.Genet. 23, 421‐424 162. Basel‐Vanagaite, L., Attia, R., Ishida‐Yamamoto, A., Rainshtein, L., Ben, A. D., Lurie, R., Pasmanik‐Chor, M., Indelman, M., Zvulunov, A., Saban, S., Magal, N., Sprecher, E., Shohat, M. (2007) Autosomal recessive ichthyosis with hypotrichosis caused by a mutation in ST14, encoding type II transmembrane serine protease matriptase. Am.J.Hum.Genet. 80, 467‐477 163. Hansson, L., Backman, A., Ny, A., Edlund, M., Ekholm, E., Ekstrand, H. B., Tornell, J., Wallbrandt, P., Wennbo, H., Egelrud, T. (2002) Epidermal overexpression of stratum corneum chymotryptic enzyme in mice: a model for chronic itchy dermatitis. J.Invest Dermatol. 118, 444‐449 164. Henskens, Y. M., Veerman, E. C., Nieuw Amerongen, A. V. (1996) Cystatins in health and disease. Biol.Chem.Hoppe Seyler 377, 71‐86 165. Calkins, C. C., Sloane, B. F. (1995) Mammalian cysteine protease inhibitors: biochemical properties and possible roles in tumor progression. Biol.Chem.Hoppe Seyler 376, 71‐80 166. Bjorck, L. (1990) Proteinase inhibition, immunoglobulin‐binding proteins and a novel antimicrobial principle. Mol.Microbiol. 4, 1439‐1442 167. Korant, B. D., Towatari, T., Ivanoff, L., Petteway S Jr, Brzin, J., Lenarcic, B., Turk, V. (1986) Viral therapy: prospects for protease inhibitors. J.Cell Biochem. 32, 91‐95 168. Turk, B., Turk, V., Turk, D. (1997) Structural and functional aspects of papain‐like cysteine proteinases and their protein inhibitors. Biol.Chem. 378, 141‐150 169. Bode, W., Huber, R. (2000) Structural basis of the endoproteinase‐protein inhibitor interaction. Biochim.Biophys.Acta 1477, 241‐252 170. Jansen, B. J., van Ruissen, F., de Jongh, G., Zeeuwen, P. L., Schalkwijk, J. (2001) Serial analysis of gene expression in differentiated cultures of human epidermal keratinocytes. J.Invest Dermatol. 116, 12‐22 171. van Ruissen, F., Jansen, B. J., de Jongh, G. J., Zeeuwen, P. L., Schalkwijk, J. (2002) A partial transcriptome of human epidermis. Genomics 79, 671‐678

47 172. van Ruissen, F., Jansen, B. J., de Jongh, G. J., van Vlijmen‐Willems, I. M., Schalkwijk, J. (2002) Differential gene expression in premalignant human epidermis revealed by cluster analysis of serial analysis of gene expression (SAGE) libraries. FASEB J. 16, 246‐248 173. Sotiropoulou, G., Anisowicz, A., Sager, R. (1997) Identification, cloning, and characterization of cystatin M, a novel cysteine proteinase inhibitor, down‐regulated in breast cancer. J.Biol.Chem. 272, 903‐910 174. Ni, J., Abrahamson, M., Zhang, M., Fernandez, M. A., Grubb, A., Su, J., Yu, G. L., Li, Y., Parmelee, D., Xing, L., Coleman, T. A., Gentz, S., Thotakura, R., Nguyen, N., Hesselberg, M., Gentz, R. (1997) Cystatin E is a novel human cysteine proteinase inhibitor with structural resemblance to family 2 cystatins. J.Biol.Chem. 272, 10853‐10858 175. Stenman, G., Astrom, A. K., Roijer, E., Sotiropoulou, G., Zhang, M., Sager, R. (1997) Assignment of a novel cysteine proteinase inhibitor (CST6) to 11q13 by fluorescence in situ hybridization. Cytogenet.Cell Genet. 76, 45‐46 176. Schnittger, S., Rao, V. V., Abrahamson, M., Hansmann, I. (1993) Cystatin C (CST3), the candidate gene for hereditary cystatin C amyloid angiopathy (HCCAA), and other members of the cystatin gene family are clustered on chromosome 20p11.2. Genomics 16, 50‐55 177. Zeeuwen, P. L., Dale, B. A., de Jongh, G. J., van Vlijmen‐Willems, I. M., Fleckman, P., Kimball, J. R., Stephens, K., Schalkwijk, J. (2003) The human cystatin M/E gene (CST6): exclusion as a candidate gene for harlequin ichthyosis. J.Invest Dermatol. 121, 65‐68 178. Zeeuwen, P. L., van Vlijmen‐Willems, I. M., Jansen, B. J., Sotiropoulou, G., Curfs, J. H., Meis, J. F., Janssen, J. J., van Ruissen, F., Schalkwijk, J. (2001) Cystatin M/E expression is restricted to differentiated epidermal keratinocytes and sweat glands: a new skin‐specific proteinase inhibitor that is a target for cross‐linking by transglutaminase. J.Invest Dermatol. 116, 693‐ 701 179. Tjabringa, G. S., Vos, J. B., Olthuis, D., Ninaber, D. K., Rabe, K. F., Schalkwijk, J., Hiemstra, P. S., Zeeuwen, P. L. (2005) Host defense effector molecules in mucosal secretions. FEMS Immunol.Med.Microbiol. 45, 151‐158 180. Zeeuwen, P. L. (2004) The role of cystatin M/E in epidermal differentiation. From tag to function. Thesis. Radboud University Nijmegen, The Netherlands. 181. Rivenbark, A. G., Livasy, C. A., Boyd, C. E., Keppler, D., Coleman, W. B. (2007) Methylation‐ dependent silencing of CST6 in primary human breast tumors and metastatic lesions. Exp.Mol.Pathol. 83, 188‐197 182. Ai, L., Kim, W. J., Kim, T. Y., Fields, C. R., Massoll, N. A., Robertson, K. D., Brown, K. D. (2006) Epigenetic silencing of the tumor suppressor cystatin M occurs during breast cancer progression. Cancer Res. 66, 7899‐7909 183. Qiu, J., Ai, L., Ramachandran, C., Yao, B., Gopalakrishnan, S., Fields, C. R., Delmas, A. L., Dyer, L. M., Melnick, S. J., Yachnis, A. T., Schwartz, P. H., Fine, H. A., Brown, K. D., Robertson, K. D. (2008) Invasion suppressor cystatin E/M (CST6): high‐level cell type‐specific expression in normal brain and epigenetic silencing in gliomas. Lab Invest 88, 910‐925 184. Zeeuwen, P. L., van Vlijmen‐Willems, I. M., Hendriks, W., Merkx, G. F., Schalkwijk, J. (2002) A null mutation in the cystatin M/E gene of ichq mice causes juvenile lethality and defects in epidermal cornification. Hum.Mol.Genet. 11, 2867‐2875 185. Sundberg, J. P., Boggess, D., Hogan, M. E., Sundberg, B. A., Rourk, M. H., Harris, B., Johnson, K., Dunstan, R. W., Davisson, M. T. (1997) Harlequin ichthyosis (ichq): a juvenile lethal mouse mutation with ichthyosiform dermatitis. Am.J.Pathol. 151, 293‐310

48 186. Zeeuwen, P. L., van Vlijmen‐Willems, I. M., Olthuis, D., Johansen, H. T., Hitomi, K., Hara‐ Nishimura, I., Powers, J. C., James, K. E., op den Camp, H. J., Lemmens, R., Schalkwijk, J. (2004) Evidence that unrestricted legumain activity is involved in disturbed epidermal cornification in cystatin M/E deficient mice. Hum.Mol.Genet. 13, 1069‐1079 187. Dunnwald, M., Zuberi, A. R., Stephens, K., Le, R., Sundberg, J. P., Fleckman, P., Dale, B. A. (2003) The ichq mutant mouse, a model for the human skin disorder harlequin ichthyosis: mapping, keratinocyte culture, and consideration of candidate genes involved in epidermal growth regulation. Exp.Dermatol. 12, 245‐254 188. Broome, A. M., Ryan, D., Eckert, R. L. (2003) S100 protein subcellular localization during epidermal differentiation and psoriasis. J.Histochem.Cytochem. 51, 675‐685 189. Yamaguchi, H., Yamamoto, A., Watanabe, R., Uchiyama, N., Fujii, H., Ono, T., Ito, M. (1998) High transepidermal water loss induces fatty acid synthesis and cutaneous fatty acid‐binding protein expression in rat skin. J.Dermatol.Sci. 17, 205‐213 190. Kim, H. C., Nemes, Z., Idler, W. W., Hyde, C. C., Steinert, P. M., Ahvazi, B. (2001) Crystallization and preliminary X‐ray analysis of human transglutaminase 3 from zymogen to active form. J.Struct.Biol. 135, 73‐77 191. Ahvazi, B., Kim, H. C., Kee, S. H., Nemes, Z., Steinert, P. M. (2002) Three‐dimensional structure of the human transglutaminase 3 enzyme: binding of calcium ions changes structure for activation. EMBO J. 21, 2055‐2067 192. Keppler, D. (2006) Towards novel anti‐cancer strategies based on cystatin function. Cancer Lett. 235, 159‐176 193. Shridhar, R., Zhang, J., Song, J., Booth, B. A., Kevil, C. G., Sotiropoulou, G., Sloane, B. F., Keppler, D. (2004) Cystatin M suppresses the malignant phenotype of human MDA‐MB‐435S cells. Oncogene 23, 2206‐2215 194. Song, J., Jie, C., Polk, P., Shridhar, R., Clair, T., Zhang, J., Yin, L., Keppler, D. (2006) The candidate tumor suppressor CST6 alters the gene expression profile of human breast carcinoma cells: down‐regulation of the potent mitogenic, motogenic, and angiogenic factor autotaxin. Biochem.Biophys.Res.Commun. 340, 175‐182 195. Zhang, J., Shridhar, R., Dai, Q., Song, J., Barlow, S. C., Yin, L., Sloane, B. F., Miller, F. R., Meschonat, C., Li, B. D., Abreo, F., Keppler, D. (2004) Cystatin m: a novel candidate tumor suppressor gene for breast cancer. Cancer Res. 64, 6957‐6964 196. Ross, D. T., Scherf, U., Eisen, M. B., Perou, C. M., Rees, C., Spellman, P., Iyer, V., Jeffrey, S. S., Van de Rijn, M., Waltham, M., Pergamenschikov, A., Lee, J. C., Lashkari, D., Shalon, D., Myers, T. G., Weinstein, J. N., Botstein, D., Brown, P. O. (2000) Systematic variation in gene expression patterns in human cancer cell lines. Nat.Genet. 24, 227‐235 197. Vigneswaran, N., Wu, J., Muller, S., Zacharias, W., Narendran, S., Middleton, L. (2005) Expression analysis of cystatin C and M in laser‐capture microdissectioned human breast cancer cells‐‐a preliminary study. Pathol.Res.Pract. 200, 753‐762 198. Zeeuwen, P. L., van Vlijmen‐Willems, I. M., Egami, H., Schalkwijk, J. (2002) Cystatin M/E expression in inflammatory and neoplastic skin disorders. Br.J.Dermatol. 147, 87‐94 199. Werle, B., Schanzenbacher, U., Lah, T. T., Ebert, E., Julke, B., Ebert, W., Fiehn, W., Kayser, K., Spiess, E., Abrahamson, M., Kos, J. (2006) Cystatins in non‐small cell lung cancer: tissue levels, localization and relation to prognosis. Oncol.Rep. 16, 647‐655

49 200. Vigneswaran, N., Wu, J., Zacharias, W. (2003) Upregulation of cystatin M during the progression of oropharyngeal squamous cell carcinoma from primary tumor to metastasis. Oral Oncol. 39, 559‐568 201. Vigneswaran, N., Wu, J., Nagaraj, N., James, R., Zeeuwen, P., Zacharias, W. (2006) Silencing of cystatin M in metastatic oral cancer cell line MDA‐686Ln by siRNA increases cysteine proteinases and legumain activities, cell proliferation and in vitro invasion. Life Sci. 78, 898‐ 907 202. Haider, A. S., Peters, S. B., Kaporis, H., Cardinale, I., Fei, J., Ott, J., Blumenberg, M., Bowcock, A. M., Krueger, J. G., Carucci, J. A. (2006) Genomic analysis defines a cancer‐specific gene expression signature for human squamous cell carcinoma and distinguishes malignant hyperproliferation from benign hyperplasia. J.Invest Dermatol. 126, 869‐881 203. Hosokawa, M., Kashiwaya, K., Eguchi, H., Ohigashi, H., Ishikawa, O., Furihata, M., Shinomura, Y., Imai, K., Nakamura, Y., Nakagawa, H. (2008) Over‐expression of cysteine proteinase inhibitor cystatin 6 promotes pancreatic cancer growth. Cancer Sci. 99, 1626‐1632 204. Fraga, M. F., Ballestar, E., Paz, M. F., Ropero, S., Setien, F., Ballestar, M. L., Heine‐Suner, D., Cigudosa, J. C., Urioste, M., Benitez, J., Boix‐Chornet, M., Sanchez‐Aguilera, A., Ling, C., Carlsson, E., Poulsen, P., Vaag, A., Stephan, Z., Spector, T. D., Wu, Y. Z., Plass, C., Esteller, M. (2005) Epigenetic differences arise during the lifetime of monozygotic twins. Proc.Natl.Acad.Sci.U.S.A 102, 10604‐10609 205. Feinberg, A. P. (2007) Phenotypic plasticity and the epigenetics of human disease. Nature 447, 433‐440 206. Robertson, K. D. (2005) DNA methylation and human disease. Nat.Rev.Genet. 6, 597‐610 207. Arney, K. L., Fisher, A. G. (2004) Epigenetic aspects of differentiation. J.Cell Sci. 117, 4355‐ 4363 208. Vaissiere, T., Sawan, C., Herceg, Z. (2008) Epigenetic interplay between histone modifications and DNA methylation in gene silencing. Mutat.Res. 659, 40‐48 209. Klose, R. J., Bird, A. P. (2006) Genomic DNA methylation: the mark and its mediators. Trends Biochem.Sci. 31, 89‐97 210. Rivenbark, A. G., Jones, W. D., Coleman, W. B. (2006) DNA methylation‐dependent silencing of CST6 in human breast cancer cell lines. Lab Invest 86, 1233‐1242 211. Schagdarsurengin, U., Pfeifer, G. P., Dammann, R. (2007) Frequent epigenetic inactivation of cystatin M in breast carcinoma. Oncogene 26, 3089‐3094

50 CHAPTER 2

Cystatin M/E is a high affinity inhibitor of cathepsin V and cathepsin L by a reactive site that is distinct from the legumain‐binding site A novel clue for the role of cystatin M/E in epidermal cornification

Tsing Cheng, Kiyotaka Hitomi, Ivonne M.J.J. van Vlijmen‐Willems, Gys J. de Jongh, Kanae Yamamoto, Koji Nishi, Colin Watts, Thomas Reinheckel, Joost Schalkwijk, and Patrick L.J.M. Zeeuwen

This chapter has been published in: J.Biol.Chem. (2006) 281, 15893‐15899

51 THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 281, NO. 23, pp. 15893–15899, June 9, 2006 © 2006 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A.

Cystatin M/E Is a High Affinity Inhibitor of Cathepsin V and Cathepsin L by a Reactive Site That Is Distinct from the Legumain-binding Site A NOVEL CLUE FOR THE ROLE OF CYSTATIN M/E IN EPIDERMAL CORNIFICATION* Received for publication, January 24, 2006, and in revised form, March 13, 2006 Published, JBC Papers in Press, March 24, 2006, DOI 10.1074/jbc.M600694200 Tsing Cheng‡1, Kiyotaka Hitomi§1, Ivonne M. J. J. van Vlijmen-Willems‡, Gys J. de Jongh‡, Kanae Yamamoto§, Koji Nishi§, Colin Watts¶, Thomas Reinheckelʈ, Joost Schalkwijk‡2, and Patrick L. J. M. Zeeuwen‡ From the ‡Department of Dermatology, Nijmegen Centre for Molecular Life Sciences, Radboud University Nijmegen Medical Centre, P. O. Box 9101, 6500 HB Nijmegen, The Netherlands, the §Department of Applied Molecular Biosciences, Graduate School of Bioagricultural Sciences, Nagoya University, 464-8601 Nagoya, Japan, the ¶Division of Cell Biology and Immunology, School of Life Sciences, University of Dundee, DD1 5EH Scotland, United Kingdom, and the ʈInstitute of Molecular Medicine and Cell Research, Albert-Ludwigs-University, D-79104 Freiburg, Germany

Cystatin M/E is a high affinity inhibitor of the asparaginyl endopep- cysteine proteases and have shown important regulatory and protective tidase legumain, and we have previously reported that both proteins functions in cells and tissues against proteolysis by cysteine proteases of are likely to be involved in the regulation of stratum corneum forma- host, bacterial, and viral origin (1–3). The inhibitory activity of cystatins tion in skin. Although cystatin M/E contains a predicted binding is regulated by a reversible, tight-binding interaction between the site for papain-like cysteine proteases, no high affinity binding for protease inhibitor and its target protease (4). Disturbance of the any member of this family has been demonstrated so far. We report normal balance between cysteine proteases and their inhibitors at a that human cathepsin V (CTSV) and human cathepsin L (CTSL) are wrong time and location can lead to several pathological conditions strongly inhibited by human cystatin M/E. Kinetic studies show that such as chronic inflammatory reactions (5), tumor malignancy (6),

Ki values of cystatin M/E for the interaction with CTSV and CTSL and faulty differentiation processes in the epidermis and hair follicle are 0.47 and 1.78 nM, respectively. On the basis of the analogous (7). Little is known on the specific biological functions of cystatin sites in cystatin C, we used site-directed mutagenesis to identify family members. However, mutations in the genes encoding the cys- the binding sites of these proteases in cystatin M/E. We found that tatin family members cystatin B and C cause neurological pheno- the W135A mutant was rendered inactive against CTSV and CTSL types in humans (8, 9). but retained legumain-inhibiting activity. Conversely, the N64A Cystatin M/E is a 14-kDa secreted protein that shares only 35% mutant lost legumain-inhibiting activity but remained active against homology with other human type 2 cystatins. Nevertheless, it has a the papain-like cysteine proteases. We conclude that legumain and similar overall structure including the two characteristic intrachain papain-like cysteine proteases are inhibited by two distinct non-over- disulfide bridges (10, 11). Expression of cystatin M/E is found to be lapping sites. Using immunohistochemistry on normal human skin, restricted to the epidermis, more specifically in the stratum granulo- we found that cystatin M/E co-localizes with CTSV and CTSL. In sum, sweat glands, sebaceous glands, and the hair follicles (12, 13). In addition, we show that CTSL is the elusive enzyme that processes addition to its function as a cysteine protease inhibitor, cystatin M/E and activates epidermal transglutaminase 3. The identification of also serves as a target for cross-linking by transglutaminases (12). These CTSV and CTSL as novel targets for cystatin M/E, their (co)- findings have suggested an important role for cystatin M/E in skin phys- expression in the stratum granulosum of human skin, and the iology. We have previously reported that a null mutation in the mouse activity of CTSL toward transglutaminase 3 strongly imply an cystatin M/E gene causes the murine ichq phenotype, which is charac- important role for these enzymes in the differentiation process of terized by neonatal lethality and abnormalities in cornification and des- human epidermis. quamation, demonstrating an essential role for cystatin M/E in the final stages of epidermal differentiation (14). However, cystatin M/E was excluded as the causative gene in a lethal form of ichthyosis in humans (15). Recently, we proposed the involvement of the asparaginyl endopepti- The cellular activity of a protease is the result of many regulatory dase legumain in disturbed epidermal cornification in cystatin M/E-defi- mechanisms such as the concentration and compartmentalization of cient mice (16). Legumain belongs to family C13 of cysteine proteases, substrates, the enzyme itself, and its cognate inhibitors. Cystatins are which is unrelated to the family of papain-like cysteine proteases, and is the natural and specific inhibitors of endogenous mammalian lysosomal strictly specific for hydrolysis of asparaginyl bonds (17). This lysosomal protease was shown to be involved in the processing of other lysosomal proteases such as cathepsins B, H, and L (18, 19). Furthermore, a bio- * This work was supported by a grant from the Radboud University Nijmegen Medical Centre (to T. C.) and by VENI-Grant 916.56.117 from ZONMW, the Netherlands Orga- chemical study on legumain has indicated that cystatin M/E in vitro nization for Health Research and Development (to P. L. J. M. Z.). The costs of publica- binds to this protease with high affinity (20). We have shown that tion of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Sec- cystatin M/E deficiency in mice leads to free cutaneous legumain tion 1734 solely to indicate this fact. activity, increased transepidermal water loss, and dehydration. We 1 Both authors contributed equally to this study. 2 To whom correspondence should be addressed. Tel.: 31-24-3614094; Fax: 31-24- have reported that disturbed cornification is caused by abnormalities 3541184; E-mail: [email protected]. in loricrin processing, which may be the result of abnormal activation of

JUNE 9, 2006•VOLUME 281•NUMBER 23 JOURNAL OF BIOLOGICAL CHEMISTRY

52 Novel Targets for Cystatin M/E transglutaminase 3 (TGase 3)3 by cathepsins, whose activities in their and the recombinant human cystatins A, B, C, D, F, and S (all from R&D turn could be regulated by legumain (16). During keratinocyte differen- Systems, Minneapolis, MN) against recombinant human legumain (24), tiation, TGase 3 is activated by limited proteolysis of a 77-kDa zymogen human CTSV (R&D Systems), and human CTSL (one-chain form, (21, 22), a process that is apparently under control of cystatin M/E, at recombinant protein, R&D Systems; two-chain form, purified from least during skin morphogenesis in the neonatal phase. These data sug- human liver, Sigma) was determined by measuring inhibition of the gest that cystatin M/E plays a key role in the regulation of proteolytic proteases using the fluorogenic synthetic substrates Z-Ala-Ala-Asn- events that are involved in barrier formation and maintenance. To MCA (4-methyl-coumaryl-7-amide) (Peptides International, Louisville, understand this process in more detail at the molecular and cellular KY) for legumain and Z-Leu-Arg-AMC (7-amino-4-methyl-coumarin) level, we set out to identify the relevant enzymes and inhibitors and to (R&D Systems) for CTSV and CTSL. Protease inhibition was measured study their localization in human skin. The identification of CTSV and after an incubation period of 5 min (CTSL) or 10 min (legumain, CTSV) CTSL as novel targets for cystatin M/E, their expression in human skin, at room temperature. The buffer that was used in the legumain assay and the processing activity of CTSL toward TGase 3 strongly imply an contained 0.1 M phosphate (pH 5.7), 2 mM EDTA, 1 mM dithiothreitol, important role for these enzymes in the differentiation process of 2.67 mML-cysteine, and 100 ␮g/␮l bovine serum albumin. CTSV and human epidermis. CTSL assays were performed in buffer (pH 5.5) containing 0.1 M acetate, 1mM EDTA, 2 mM dithiothreitol, and 100 ␮g/␮l bovine serum albumin. EXPERIMENTAL PROCEDURES Reactions were stopped by adding 0.1 M Na2CO3 (pH 8.5). Ki values Three-dimensional Modeling of Cystatin M/E—A three-dimensional were determined using the Easson-Stedman plot as described by Bieth model of human cystatin M/E was generated using MODELLER (23). (25) and as indicated in the legend of Fig. 2. The crystal structure of human cystatin D (Protein Data Bank identifi- Immunohistochemistry—Human skin biopsies were processed for cation number 1Roa) was used as a starting point. Human cystatin D immunohistochemistry as described previously (13, 26). Immunohisto- and cystatin M/E have a 26% sequence identity. The sequences were chemical staining was performed according to the avidin biotinylated- aligned using the internal aligner of MODELLER; this routine uses a gap enzyme-complex method (Vector, Burlingame, CA) using polyclonal function that depends on the secondary structure. rabbit-anti-human cystatin M/E antibodies (12), polyclonal sheep-anti- Production of Cystatin M/E Variants—Cystatin M/E variants N64A human legumain antibodies (24), monoclonal mouse-anti-human and W135A were obtained using the recombinant plasmid pGEX-2T/ CTSV antibodies (R&D Systems), and monoclonal rat-anti-human/ cystatin M/E (a kind gift of Dr Georgia Sotiropoulou) and QuikChange mouse CTSL antibodies (R&D Systems). site-directed mutagenesis kit (Stratagene, La Jolla, CA). Primers used Digestion of TGase 3 Zymogen with Cathepsins—Baculovirus expressed Ј for the generation of the N64A variant were 5 -GCTACAACAT- recombinant human TGase 3 zymogen (22) was processed by treatment Ј Ј GGGCAGCGCCAGCATCTACTACTTCC-3 (sense) and 5 -GGAAG- with recombinant human CTSL, CTSV, cathepsin B (CTSB), cathepsin Ј TAGTAGATGCTGGCGCTGCCCATGTTGTAGC-3 (antisense). S (CTSS), and cathepsin D (CTSD) (all from R&D Systems). To prepare Ј Primers for the W135A variant were 5 -GGTCCTTGTGGTTCCCGCG- the proteolyzed form of TGase 3, 4 ␮g of zymogen was treated with Ј CAGAACTCCTCTCAGC (sense) and 5 -GCTGAGAGGAGTTCT- cathepsins using a molar ratio of ϳ0.001: CTSL (2.8 ng), CTSV (2 ng), Ј GCGCGGGAACCACAAGGACC-3 (antisense). All primers were from CTSS (2 ng), and CTSB (1.9 ng). In the case of CTSD, the ratio is about Biolegio (Malden, The Netherlands). Sequence analysis was performed to 0.05 (80 ng). Each incubation was performed in the appropriate buffers check for the desired mutations in the cDNA. as follows: for CTSL, CTSV, CTSB, and CTSS, 100 mM MES (pH 6.0), 16 Production of Recombinant Proteins—Production and purification of mM dithiothreitol, 1.6 mM EDTA; and for CTSD, 100 mM sodium ace- recombinant wild type and mutant forms of human cystatin M/E were tate (pH 4.0), 100 mM KCl. For controls, the same reaction mixtures performed as described previously (10) with minor changes. pGEX-2T were prepared but with the addition of pepstatin (Nacarai, Kyoto, plasmids containing the wild type or mutated cystatin M/E cDNAs were Japan), which is an inhibitor of the aspartic protease CSTD, or E-64 transformed into Escherichia coli BL21 Star (DE3) bacteria (Invitrogen). (trans-epoxysuccinyl-L-leucylamido-(4-guanidino)butane, Peptide Insti- Cultures were grown from one single colony overnight at 37 °C in LB tute Inc., Osaka, Japan), a synthetic inhibitor of the cysteine proteases medium with ampicillin followed by induction with isopropyl-1-thio- CTSL, CTSV, CTSB, and CTSS. Both controls were used at a final concen- ␤-D-galactopyranoside (Roche Applied Science) at a final concentration tration of 10 ␮M. Reactions were performed at 37 °C for the indicated times of 0.5 mM for2hat37°C.Cells were lysed by four cycles of freezing in (see Fig. 4). liquid nitrogen and thawing. Glutathione-Sepharose 4B beads (Amer- Immunoblotting—Reaction mixtures with proteolyzed TGase 3 were sham Biosciences) were used for purification of glutathione S-transfer- subjected to SDS-PAGE and electroblotted onto polyvinylidene difluo- ase fusion protein from the lysate. Cleavage of the fusion protein into ride membrane as described previously (27). The membrane was incu- glutathione S- and cystatin M/E occurred overnight at 4 °C bated with anti-human TGase 3 monoclonal antibodies C2D and C9D, with thrombin (Sigma) at a thrombin to fusion protein ratio of 1:500 in which specifically recognize the zymogen form and the proteolyzed 30- a buffer containing 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 2.5 mM and 47-kDa fragments, respectively (27). Detection of the proteins was CaCl . Recombinant cystatin M/E was purified from glutathione S-trans- 2 established with peroxidase-conjugated anti-mouse IgG and IgM (Jack- ferase using glutathione-Sepharose 4B beads and dialyzed against phos- son ImmunoResearch Laboratories, West Grove, PA) and SuperSignal phate-buffered saline with an YM3 filter (Millipore, Billerica, MA). Protein West Pico chemiluminescent substrate (Pierce) according to the man- concentration was determined using a BCA protein assay (Pierce). ufacturer’s protocol. Fluorimetric Enzyme Assays—Protease inhibitory activity of recom- TGase 3 Activity—Prior to the assay, E-64 was added to the reaction binant human wild type cystatin M/E, the N64A and W135A variants, mixture to block cathepsin activity. TGase 3 activity was measured by a 96-well plate assay as described previously (28, 29). Briefly, 1% dimethylca- 3 The abbreviations used are: TGase, transglutaminase; CTSV, cathepsin V; CTSL, cathep- sein was fixed at the plate, and the uncoated sites were blocked with sin L; CTSB, cathepsin B; CTSS, cathepsin S; CTSD, cathepsin D; MES, 2-(N-morpho- lino)ethanesulfonic acid; E-64, trans-epoxysuccinyl-L-leucylamido-(4-guanidino)bu- skimmed milk. The sample was incubated with 5-(biotinamido-)pen- tane; Z, benzyloxycarbonyl. tylamine (Pierce) in a buffer containing 100 mM Tris-HCl (pH 8.0), 10 mM

15894 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281•NUMBER 23•JUNE 9, 2006 53 Novel Targets for Cystatin M/E

TABLE 1

Ki values of wild type cystatin M/E and mutated variants for legumain, cathepsin V, and cathepsin L Cystatin Ki legumain Ki cathepsin V Ki cathepsin L M/E

nM nM nM Wild type 0.25 0.47 1.78 N64A Ͼ100a 0.34 1.77 W135A 0.72 Ͼ100b 11.7 a v i/v0 Ͼ 0.60 at ͓I͔ϭ100 nM. b v i/v0 Ͼ 0.50 at ͓I͔ϭ100 nM.

TABLE 2

Ki values of human cystatins for legumain, cathepsin V, and cathepsin L

Cystatin Ki legumain Ki cathepsin V Ki cathepsin L

nM nM nM FIGURE 1. Predicted three-dimensional representation of cystatin M/E. The crystal cystatin A Ͼ100a 0.11 0.05 structure of cystatin D was used as a basis for this model. The regions of the inhibitory cystatin B Ͼ100a 0.13 0.05 sites for papain-like cysteine proteases and for legumain are marked. The locations of the cystatin C 7.28 0.02 0.08 amino acids substituted in the cystatin M/E variants are shown in white. cystatin D Ͼ100a 29.6 2.81 cystatin F Ͼ100a Ͼ100b 0.49 cystatin S Ͼ100a Ͼ100c Ͼ100d a v i/v0 Ͼ 0.99 at ͓I͔ϭ100 nM. b v i/v0 Ͼ 0.70 at ͓I͔ϭ100 nM. c v i/v0 Ͼ 0.90 at ͓I͔ϭ100 nM. d v i/v0 Ͼ 0.85 at ͓I͔ϭ100 nM.

cystatin D (30), which revealed that the overall predicted structure of cys- tatin M/E closely matched that of cystatin D (Fig. 1). To study the presumed reactive sites and the inhibitory activities of cystatin M/E, variants of the protein with an inactive site for either legumain or papain-like cysteine proteases were generated by substitution of one of the conserved amino acids. As shown in Fig. 1, the N64A variant has a substitution of the Asn64 residue of the proposed legumain inhibitory site by an Ala residue, whereas the W135A variant has its Trp135 residue of the papain-like cysteine prote- K FIGURE 2. Determination of the dissociation constant i of the protease/protease- ase inhibitory site substituted by an Ala residue. Wild type and mutant inhibitor complex. The dissociation constant Ki was determined by measuring the residual enzymatic activity of a fixed concentration of enzyme (as shown here for CTSV) cystatin M/E was tested for inhibitory activity against legumain and several incubated with increasing concentrations of inhibitor (cystatin M/E). An Easson-Stedman papain-like cysteine proteases. In previous studies, it was found that cysta- plot (inset) was used to calculate the Ki. The inset is a replot of the data according to the Ϫ ϭ ϩ 0 0 tin M/E was a poor inhibitor of most lysosomal cathepsins. Here we following equation: [I]/1 a (Ki/a) E , where I is the inhibitor concentration, E the enzyme concentration at time 0, and a is the fractional activity, i.e. the ratio between extended our studies to include CTSV and we also included the one-chain the enzyme activity in the presence (V ) and absence (V ) of the inhibitor. The plot yields a i 0 form of CTSL, as this form is the one expressed in skin. K values were straight line whose slope is K . A representative curve of three experiments is shown. i i determined using an Easson-Stedman plot as illustrated in Fig. 2 for CTSV

and wild type cystatin M/E. When the Ki could not be determined in the dithiothreitol, and 5 mM CaCl2. After various periods of incubation, the case of very low affinity, the ratio of protease activity in the presence (Vi ) reaction was blocked by adding EDTA at a final concentration of 10 mM. and absence (V0) of inhibitor was determined, and the remaining protease

TGase-catalyzed conjugation of 5-(biotinamido-)pentylamine to dimethyl- activity at a fixed high inhibitor concentration was calculated. Ki values of casein was measured by avidin-conjugated peroxidase, hydrogen peroxide, wild type human cystatin M/E and the N64A and W135A variants for o-phenylenediamine, and hydroxy peroxide. An equal volume of 2 M H2SO4 interaction with human legumain and the human cysteine proteases CTSV was added, and the absorbance was measured at 450 nm. and CTSL are shown in Table 1. Wild type cystatin M/E is a high affinity

Determination of the TGase 3 Cleavage Site—Ten micrograms of inhibitor for legumain, CTSV, and CTSL as witnessed by their respective Ki human TGase 3 was digested with 10 ng of CTSL, and the resulting values of 0.25, 0.47, and 1.78 nM. Similar Ki values were found for interaction fragments were subjected to SDS-PAGE, blotted onto polyvinylidene of the N64A variant with CTSV and CSTL but not with legumain (Vi / Ͼ ϭ difluoride membrane, and subsequently stained with Coomassie Blue. V0 0.60 at [I] 100 nM). The Ki value of the W135A variant for interac- The N-terminal amino acid sequence of the 30-kDa fragment was deter- tion with legumain was in the same range as that of wild type cystatin M/E, mined by automated Edman degradation. whereas the Ki value for interaction with CTSV could not be determined Ͼ ϭ (Vi /V0 0.50 at [I] 100 nM). The W135A mutation did not decrease the RESULTS affinity for CTSL at a similar order of magnitude as for CTSV but still

Inhibition of Protease Activity by Cystatin M/E—Based on the homology caused a 6-fold increase of the Ki value as compared with that of wild type with cystatin C, we predicted that cystatin M/E would contain two distinct cystatin M/E (11.7 nM versus 1.78 nM). Similar experiments with several binding sites for papain-like cysteine proteases and for the asparaginyl other human cystatins were performed and Ki values for the interaction endopeptidase legumain. A three-dimensional model of human cystatin with human legumain, CTSV, and CTSL are shown in Table 2. CTSV was

M/E was generated on the basis of the published crystal structure of human efficiently inhibited by cystatins A, B, and C (Ki values below 0.15 nM),

JUNE 9, 2006•VOLUME 281•NUMBER 23 JOURNAL OF BIOLOGICAL CHEMISTRY 15895 54 Novel Targets for Cystatin M/E

FIGURE 3. Immunostaining for cystatin M/E, cathepsin V, cathepsin L, and legumain in nor- mal human skin and its appendages. Note the stratum-specific expression of cystatin M/E and CTSV in the stratum granulosum of the epidermis. CTSV is only present in the inner layer of the outer root sheet of the hair follicle, whereas cystatin M/E is also expressed in many more layers of the outer root sheet of the hair follicle. The secretory coil epi- thelium of the sweat glands is positively stained for cystatin M/E (arrowheads), but the ductal part is almost negative (the asterisk is surrounded by ducts). No expression of CTSV was found in the sweat glands. CTSL and legumain are also expressed in the stratum granulosum. Furthermore, positive staining was also found in the other layers of the epi- dermis and its appendages. The sweat gland ducts as well as the secretory coils are both positively stained for CTSL and legumain. Scale bar ϭ 100 ␮m.

FIGURE 4. Proteolysis of human TGase 3 by ca- thepsins. Four micrograms of recombinant human TGase 3 was treated with several cathepsins at 37 °C for the indicated times. The reaction mixtures were subjected to SDS-PAGE followed by immunoblot- ting analysis using monoclonal antibodies C2D and C9D. Reactions were also performed in the pres- ence of specific inhibitors (E-64 or pepstatin) at a final concentration of 10 ␮M. The 77-kDa TGase 3 zymogen is rapidly proteolyzed into its 30- and 47-kDa fragments by CTSL (A) but not by CTSV (B), CTSS (C), CTSB (D), and CTSD (E). The open and closed arrowheads indicate zymogen and proteo- lyzed TGase 3, respectively.

whereas the affinity of CTSV for cystatin D was weaker (Ki of 29.6 nM) and of CTSV, whereas the secretory coil epithelium but not the ductal part not detectable with cystatins F and S. CTSL was efficiently inhibited by of the sweat glands was positive for cystatin M/E. CTSV and cystatin cystatins A, B, C, and F (Ki values below 0.5 nM), whereas the affinity for M/E both show a positive staining in the inner, mature cells of the cystatin D was somewhat weaker (Ki of 2.81 nM). Cystatin S showed no sebaceous glands. In contrast to cystatin M/E and CTSV, the expression inhibition with any of the tested cysteine proteases, whereas the ubiqui- of legumain and CTSL was not limited to a specific cell layer but was tously expressed cystatin C demonstrated strong inhibitory activity for all of found throughout all layers of the epidermis and its appendages (Fig. 3). them, although its affinity for legumain is much weaker than that of cystatin Activation of TGase 3 by CTSL—In a previous study, we have found M/E (7.28 nM versus 0.25 nM). that lysosomal cysteine proteases are probably involved in the process- Co-expression of Cystatin M/E and Its Biological Targets in the Stra- ing of human epidermal TGase 3 in vitro, suggesting that this could be a tum Granulosum—To establish a functional link between cystatin M/E mechanism for activation. So far, the responsible proteases for this acti- and its putative target proteases, we investigated their expression at the vation remained unknown. To address this issue more in depth at the tissue level in human skin. Immunohistochemistry was performed on functional level, recombinant human TGase 3 was incubated with a skin biopsies from healthy volunteers. In normal human skin, cystatin panel of recombinant human lysosomal cysteine proteases to investi- M/E and CTSV are strongly and specifically expressed in the stratum gate their proteolytic activity toward the zymogen form of TGase 3. We granulosum of the epidermis (Fig. 3). In addition, both proteins are also found that CTSL is able to proteolyze TGase 3 into 30- and 47-kDa expressed in the hair follicle. CTSV is only present in a small layer of the fragments (Fig. 4A) in a concentration- and time-dependent manner. hair follicle root sheet, whereas cystatin M/E is expressed throughout When the cathepsin inhibitor E-64 was added to the reaction mixture, the inner half of the root sheet. The sweat glands showed no expression no TGase 3 proteolysis could be detected. Other cathepsins were not

15896 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281•NUMBER 23•JUNE 9, 2006 55 Novel Targets for Cystatin M/E

FIGURE 6. Processing of TGase 3. The 77-kDa zymogen form of TGase 3 is proteolyzed into the activated form consisting of 30- and 47-kDa fragments, which then associate non-covalently to form the active enzyme (21). The N-terminal sequence of the 30-kDa fragment of human TGase 3 was determined by automated Edman degradation. For comparison, we show the cleavage site of TGase 3 after digestion by dispase, a neutral bacterial protease that is commonly used to activate TGase 3 in vitro. The open and closed triangles represent the cleavage sites of dispase and CTSL digestion, respectively. FIGURE 5. Activity of TGase 3 proteolyzed by cathepsin L. The enzymatic activity of proteolyzed TGase 3 was measured after digestion with CTSL for 3 min. The enzymatic reaction products are shown as the value of A . The open and closed circles indicate 450 nm group of Alvarez-Fernandez. Cystatins are known as inhibitors of lyso- the reactions using 100 ng of zymogen and the proteolyzed forms, respectively. somal cysteine proteases of the papain family (C1). Although cystatin M/E contains a predicted binding site for papain-like cysteine proteases, able to cleave the TGase 3 zymogen in this experimental setup (Fig. 4, no high affinity binding for any member of this family had been dem- B–E), except in the case of longer incubation times with cathepsin S (Fig. onstrated so far (11, 12, 16). The present study shows that cystatin M/E 4C). In the presence of relatively high amounts of cathepsins B, S, and V is a high affinity inhibitor of CTSV and CTSL. For this study, we used in the reaction mixture (molar ratio of 1:10), TGase 3 was limitedly recombinant CTSL that was obtained as a one-chain form, which is the proteolyzed (data not shown). Next, we measured the enzymatic activity relevant form for skin (31). In general, CTSL is translated as prepro- of the proteolyzed products of TGase 3 zymogen that was digested by CTSL, and during intracellular transport to the lysosomes, the preproen- CTSL. Although the TGase 3 zymogen showed no enzymatic activity, zyme is subsequently modified to pro-CTSL and processed into the TGase proteolyzed by the action of CTSL demonstrated an apparent mature one-chain and two-chain forms of CTSL. Processing of pro- enzymatic activity (Fig. 5). N-terminal sequencing of the 30-kDa TGase CTSL results in a ϳ30-kDa one-chain form of CTSL, which in turn 3 fragment revealed that the CTSL cleavage site in TGase 3 is before can be further processed into the two-chain form (5 and 24 kDa) Ala467, two amino acids downstream as compared with the cleavage site (32). It is not known whether this conversion regulates CTSL activity generated by dispase digestion (Fig. 6). or substrate specificity, but recently, it has been reported that both forms are proteolytically active and can execute the essential functions DISCUSSION of CTSL (19, 32). As we had previously found that CTSL purified from In this study, we show that the protease inhibitor cystatin M/E inhib- human liver (the two-chain form) was only weakly inhibited by cystatin its legumain and papain-like cysteine proteases due to two distinct non- M/E, the strong inhibition by recombinant one-chain form CTSL was overlapping sites. The papain-like cysteine proteases CTSV and CTSL unexpected. Comparison of human liver CTSL from different suppliers are strongly inhibited by cystatin M/E, suggesting that these proteases revealed that we have previously underestimated the inhibition of cys- are new targets for cystatin M/E. Immunostaining of cystatin M/E, tatin M/E for reasons we cannot explain. Based on the CTSL prepara- CTSV, CTSL, and legumain in normal human skin shows co-expres- tions used in this study both the one-chain and two-chain form of CTSL sion, especially in the stratum granulosum and the hair follicles, suggest- are inhibited equally well by cystatin M/E (not shown). CTSV is a ing that these enzymes have an important role in human epidermal recently discovered protease, also known as cathepsin L2 or stratum differentiation and hair follicle morphogenesis. Our data indicate that corneum thiol protease. CTSV is closely related to CTSL sharing ϳ80% CTSL is the elusive protease that can process and activate TGase 3, an protein sequence identity (33–36). Their encoding genes are located on enzyme that is involved in cornified envelope formation. adjacent sites on chromosome 9q22.2, suggesting that CTSV and CTSL Based on the homology with cystatin D, we generated a three-dimen- have evolved by duplication from an ancestral gene (35). However, sional model of cystatin M/E. This model shows that a loop located on although CTSL is ubiquitously expressed, CTSV expression is more the opposite side to the cysteine protease-binding surface, between the restricted and predominantly found in thymus, testis, cornea of the eye, ␣-helix and the first strand of the main ␤-sheet of the cystatin M/E and epidermis (33, 34, 36). Besides the general function of lysosomal structure, contains the asparaginyl residue that is probably involved in cathepsins, i.e. bulk proteolysis, studies using a knock-out mouse model legumain binding. Site-directed mutagenesis was used to identify the have suggested that CTSL has specific functions in keratinocyte prolif- presumed binding sites of papain-like cysteine proteases and that of the eration and hair follicle morphogenesis and cycling (37–39). asparaginyl endopeptidase legumain in cystatin M/E. Substitution of a The role of CTSV and CTSL in terminal differentiation of keratino- conserved amino acid by alanine in the two reactive sites resulted in loss cytes and hair follicle morphogenesis is unknown, but these proteins of inhibitory activity. We could confirm that the conserved Asn64 resi- probably process and activate other enzymes in pathways that lead to due (cystatin M/E numbering) in the legumain inhibitory site is impor- skin barrier formation. Our biochemical assays show that CTSL is the tant for cystatin-legumain interaction. Our Ki values for this interaction first identified mammalian protease that is able to cleave and activate are somewhat higher than the Ki values found by Alvarez-Fernandez TGase 3 in vitro (Figs. 4 and 5). TGase 3 is an epidermis-specific enzyme et al. (20), but this could be explained by our use of recombinant human responsible for cross-linking loricrin and small proline-rich proteins in legumain, whereas pig legumain extracted from kidney was used by the the stratum granulosum (27, 40). TGase 3 is also the most efficient

JUNE 9, 2006•VOLUME 281•NUMBER 23 JOURNAL OF BIOLOGICAL CHEMISTRY 15897 56 Novel Targets for Cystatin M/E cross-linking enzyme of soluble trichohyalin, a major structural protein (47, 48). No orthologue of CTSV in mice has been discovered yet, but it in hair follicle differentiation (41). Remarkably, a reduced expression of is plausible that murine CTSL is actually the orthologue of human trichohyalin was found in CTSL-deficient mice (38). Our results provide CTSV, in view of the fact that human CTSV is phylogenetically more evidence that CTSL initiates the cross-linking activity of TGase 3 in the closely related to murine CTSL than to human CTSL (47). Our immu- process that leads to cornification, by cleavage of the 77-kDa TGase 3 nohistochemical data provide a novel clue in support of this theory since zymogen and thereby activation of the enzyme. This mechanism could in epidermis, CTSV expression in the upper epidermal layers is similar possibly be regulated by the inhibitory activity of cystatin M/E against to that of murine CTSL. Another possibility is that murine CTSL con- CTSL. Legumain inhibition by cystatin M/E could have a regulatory role trols the specific functional enzymatic activities of both human CTSL in CTSL activity since legumain is involved in CTSL processing (18, 19). and CTSV. A closer inspection of hair follicle morphogenesis in CTSL- Cystatin M/E-deficient mice show epidermal and follicular hyperkera- deficient mice showed that defects in the differentiation, keratinization, tosis resulting in a scaly skin, abnormal hair follicles, and an impaired and desquamation of the inner root sheet form the basis of the abnormal barrier function (14). This could be explained by elevated CTSL and/or hair phenotype (38). Therefore, a more in-depth study of cellular and legumain activities due to cystatin M/E deficiency that subsequently subcellular localization of CTSL and TGase 3 in the entire human hair lead to an early or excessive activation of TGase 3, resulting in a prema- follicle is required to elucidate the potential role of these molecules in ture or enhanced cross-linking of loricrin molecules and small proline- hair follicle morphogenesis. rich proteins, incomplete desquamation, thickening of the stratum cor- We conclude that cystatin M/E has an important regulatory function neum, and abnormal hair morphogenesis. in human epidermal differentiation. Cystatin M/E could regulate cross- N-terminal sequencing of the 30-kDa fragment revealed that CTSL linking of structural proteins by TGase 3 in cornified envelope forma- cleaves TGase 3 between amino acids Ala466 and Ala467 (Fig. 6). TGase 3, tion through inhibitory activities against CTSL and legumain. In addi- as well as other TGases, consists of four domains: a ␤-sandwich, a cat- tion, cystatin M/E could also have a role in the desquamation process as alytic core, ␤-barrel-1, and ␤-barrel-2. In TGase 3, the unique hinge CTSV activity is involved in the degradation of corneodesmosomes. region (Pro461–Glu479) separates the catalytic core and ␤-barrel-1 Further studies are necessary to elucidate the regulatory role of cystatin domain (21). In our previous studies, the cleavage site of human TGase M/E in pathways that are involved in desquamation. CTSV and CTSL 3 by dispase-digestion was Phe465, which is located in this hinge region may activate other transglutaminases such as TGase 1 and TGase 5, or (42). Since treatment of TGase 3 with thrombin or trypsin produces 30- other proteases, such as cathepsin D, that are involved in terminal dif- and 47-kDa fragments, this hinge region appears to be a domain that is ferentiation. Recently, it has been demonstrated that cathepsin D-defi- sensitive to proteolysis (43). As CTSL preferentially cleaves peptide cient mice show a reduced TGase 1 activity and impaired stratum cor- bonds with aromatic residues in P2 and hydrophobic residues in the P3 neum morphology, similar to the skin of TGase 1 knock-out mice and position (44), we believe that we have identified the exact cleavage site of the human skin disease lamellar ichthyosis (49). Collectively, this find- human TGase 3 by CTLS. This conclusion is based on the fact that in the ing and the data from our studies strongly support a functional link P2 and P3 position of the TGase 3 cleavage site, an aromatic residue between protease activity and activation of TGases in the epidermis. (Phe465) and a hydrophobic residue (Pro464) are found, respectively. REFERENCES Future studies should clarify whether this is the physiological cleavage 1. Abrahamson, M., Alvarez-Fernandez, M., and Nathanson, C. M. (2003) Biochem. Soc. site in vivo as well. Symp. 179–199 It is likely that CTSV has a role in skin that is distinct from CTSL 2. Bobek, L. A., and Levine, M. J. (1992) Crit. Rev. Oral Biol. Med. 3, 307–332 3. Turk, B., Turk, V., and Turk, D. (1997) Biol. Chem. 378, 141–150 because its expression is specifically restricted to the upper epidermal 4. Turk, B., Turk, D., and Salvesen, G. S. (2002) Curr. Pharm. Des. 8, 1623–1637 layers. The secretion of both cystatin M/E and CTSV by differentiating 5. Henskens, Y. M., Veerman, E. C., and Nieuw Amerongen, A. V. (1996) Biol. Chem. keratinocytes (12, 36) makes them both plausible candidates to be Hoppe-Seyler 377, 71–86 involved in desquamation. A detailed description of the cellular and 6. Calkins, C. C., and Sloane, B. F. (1995) Biol. Chem. Hoppe-Seyler 376, 71–80 7. Zeeuwen, P. L. (2004) Eur. J. Cell Biol. 83, 761–773 subcellular localization of cystatin M/E and its target proteases in epi- 8. Pennacchio, L. A., Lehesjoki, A. E., Stone, N. E., Willour, V. L., Virtaneva, K., Miao, J., dermis and hair follicles was beyond the scope of this report. Our immu- D’Amato, E., Ramirez, L., Faham, M., Koskiniemi, M., Warrington, J. A., Norio, R., de nohistochemical data comprise only a small part of the human hair la, C. A., Cox, D. R., and Myers, R. M. (1996) Science 271, 1731–1734 follicle, and the transverse sections at the level of the infundibulum 9. Palsdottir, A., Abrahamson, M., Thorsteinsson, L., Arnason, A., Olafsson, I., Grubb, A., and Jensson, O. (1988) Lancet 2, 603–604 (Fig. 3) preclude a detailed analysis of the inner and outer root sheet at 10. Sotiropoulou, G., Anisowicz, A., and Sager, R. (1997) J. Biol. Chem. 272, 903–910 the deeper levels. 11. Ni, J., Abrahamson, M., Zhang, M., Fernandez, M. A., Grubb, A., Su, J., Yu, G. L., Li, Y., Biochemical enzyme assays with cystatins other than cystatin M/E Parmelee, D., Xing, L., Coleman, T. A., Gentz, S., Thotakura, R., Nguyen, N., Hessel- berg, M., and Gentz, R. (1997) J. Biol. Chem. 272, 10853–10858 show that cystatins A, B, and C are strong inhibitors of CTSV and CTSL 12. Zeeuwen, P. L., van Vlijmen-Willems, I. M., Jansen, B. J., Sotiropoulou, G., Curfs, J. H., as well. Cystatin A is known to be expressed in skin and participates in Meis, J. F., Janssen, J. J., van, R. F., and Schalkwijk, J. (2001) J. Investig. Dermatol. 116, epidermal barrier formation and function (45, 46). Cystatin A belongs to 693–701 the type 1 cystatins, which are primarily intracellular proteins. Cystatin 13. Zeeuwen, P. L., van Vlijmen-Willems, I. M., Egami, H., and Schalkwijk, J. (2002) Br. J. Dermatol. 147, 87–94 M/E on the other hand can be secreted, probably into the extracellular 14. Zeeuwen, P. L., van Vlijmen-Willems, I. M., Hendriks, W., Merkx, G. F., and Schalk- space between corneocytes. Keeping this in mind, it is possible that wijk, J. (2002) Hum. Mol. Genet. 11, 2867–2875 CTSV and/or CTSL activity in the upper epidermal layers of the skin is 15. Zeeuwen, P. L., Dale, B. A., de Jongh, G. J., van Vlijmen-Willems, I. M., Fleckman, P., intracellularly regulated by cystatin A and extracellularly regulated by Kimball, J. R., Stephens, K., and Schalkwijk, J. (2003) J. Investig. Dermatol. 121, 65–68 16. Zeeuwen, P. L., van Vlijmen-Willems, I. M., Olthuis, D., Johansen, H. T., Hitomi, K., cystatin M/E. Compartmentalization studies at the ultrastructural level Hara-Nishimura, I., Powers, J. C., James, K. E., op den Camp, H. J., Lemmens, R., and could shed further light on the exact role of cystatin M/E and its prote- Schalkwijk, J. (2004) Hum. Mol. Genet. 13, 1069–1079 ase targets. 17. Chen, J. M., Dando, P. M., Rawlings, N. D., Brown, M. A., Young, N. E., Stevens, R. A., Recently, the phenotype of CTSL knock-out mice was rescued by Hewitt, E., Watts, C., and Barrett, A. J. (1997) J. Biol. Chem. 272, 8090–8098 18. Shirahama-Noda, K., Yamamoto, A., Sugihara, K., Hashimoto, N., Asano, M., Nishimura, transgenic epidermal re-expression of either murine CTSL or human M., and Hara-Nishimura, I. (2003) J. Biol. Chem. 278, 33194–33199 CTSV, indicating that human CTSV can compensate for murine CTSL 19. Maehr, R., Hang, H. C., Mintern, J. D., Kim, Y. M., Cuvillier, A., Nishimura, M.,

15898 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281•NUMBER 23•JUNE 9, 2006 57 Novel Targets for Cystatin M/E

Yamada, K., Shirahama-Noda, K., Hara-Nishimura, I., and Ploegh, H. L. (2005) J. Im- 36. Watkinson, A. (1999) Arch. Dermatol. Res. 291, 260–268 munol. 174, 7066–7074 37. Roth, W., Deussing, J., Botchkarev, V. A., Pauly-Evers, M., Saftig, P., Hafner, A., 20. Alvarez-Fernandez, M., Barrett, A. J., Gerhartz, B., Dando, P. M., Ni, J., and Abra- Schmidt, P., Schmahl, W., Scherer, J., Anton-Lamprecht, I., Von Figura, K., Paus, R., hamson, M. (1999) J. Biol. Chem. 274, 19195–19203 and Peters, C. (2000) FASEB J. 14, 2075–2086 21. Ahvazi, B., Kim, H. C., Kee, S. H., Nemes, Z., and Steinert, P. M. (2002) EMBO J. 21, 38. Tobin, D. J., Foitzik, K., Reinheckel, T., Mecklenburg, L., Botchkarev, V. A., Peters, C., 2055–2067 and Paus, R. (2002) Am. J. Pathol. 160, 1807–1821 22. Hitomi, K., Kanehiro, S., Ikura, K., and Maki, M. (1999) J. Biochem. (Tokyo) 125, 39. Benavides, F., Starost, M. F., Flores, M., Gimenez-Conti, I. B., Guenet, J. L., and Conti, 1048–1054 C. J. (2002) Am. J. Pathol. 161, 693–703 23. Sali, A., and Blundell, T. L. (1993) J. Mol. Biol. 234, 779–815 40. Candi, E., Tarcsa, E., Idler, W. W., Kartasova, T., Marekov, L. N., and Steinert, P. M. 24. Li, D. N., Matthews, S. P., Antoniou, A. N., Mazzeo, D., and Watts, C. (2003) J. Biol. (1999) J. Biol. Chem. 274, 7226–7237 Chem. 278, 38980–38990 41. Tarcsa, E., Marekov, L. N., Andreoli, J., Idler, W. W., Candi, E., Chung, S. I., and 25. Bieth, J. G. (1980) Bull. Eur. Physiopathol. Respir. 16, (suppl.) 183–197 Steinert, P. M. (1997) J. Biol. Chem. 272, 27893–27901 26. El Ghalbzouri, A., Lamme, E., and Ponec, M. (2002) Cell Tissue Res. 310, 189–199 42. Hitomi, K., Ikeda, N., and Maki, M. (2003) Biosci. Biotechnol. Biochem. 67, 2492–2494 27. Hitomi, K., Presland, R. B., Nakayama, T., Fleckman, P., Dale, B. A., and Maki, M. 43. Kim, H. C., Lewis, M. S., Gorman, J. J., Park, S. C., Girard, J. E., Folk, J. E., and Chung, (2003) J. Dermatol. Sci. 32, 95–103 S. I. (1990) J. Biol. Chem. 265, 21971–21978 28. Slaughter, T. F., Achyuthan, K. E., Lai, T. S., and Greenberg, C. S. (1992) Anal. Bio- 44. Barrett, A. J., Rawlings, N. D., and Woessner, J. F. (2004) Handbook of Proteolytic chem. 205, 166–171 Enzymes, Second Edition, pp. 1097–1102, Elsevier Academic Press, London 29. Hitomi, K., Yamagiwa, Y., Ikura, K., Yamanishi, K., and Maki, M. (2000) Biosci. Bio- technol. Biochem. 64, 2128–2137 45. Takahashi, M., Tezuka, T., and Katunuma, N. (1996) Arch. Biochem. Biophys. 329, 30. Alvarez-Fernandez, M., Liang, Y. H., Abrahamson, M., and Su, X. D. (2005) J. Biol. 123–126 Chem. 280, 18221–18228 46. Takahashi, M., Tezuka, T., Kakegawa, H., and Katunuma, N. (1994) FEBS Lett. 340, 31. Ishidoh, K., Saido, T. C., Kawashima, S., Hirose, M., Watanabe, S., Sato, N., and 173–176 Kominami, E. (1998) Biochem. Biophys. Res. Commun. 252, 202–207 47. Hagemann, S., Gunther, T., Dennemarker, J., Lohmuller, T., Bromme, D., Schule, R., 32. Ishidoh, K., and Kominami, E. (2002) Biol. Chem. 383, 1827–1831 Peters, C., and Reinheckel, T. (2004) Eur. J. Cell Biol. 83, 775–780 33. Adachi, W., Kawamoto, S., Ohno, I., Nishida, K., Kinoshita, S., Matsubara, K., and 48. Reinheckel, T., Hagemann, S., Dollwet-Mack, S., Martinez, E., Lohmuller, T., Zlatkovic, Okubo, K. (1998) Investig. Ophthalmol. Vis. Sci. 39, 1789–1796 G., Tobin, D. J., Maas-Szabowski, N., and Peters, C. (2005) J. Cell Sci. 118, 3387–3395 34. Santamaria, I., Velasco, G., Cazorla, M., Fueyo, A., Campo, E., and Lopez-Otin, C. 49. Egberts, F., Heinrich, M., Jensen, J. M., Winoto-Morbach, S., Pfeiffer, S., Wickel, M., (1998) Cancer Res. 58, 1624–1630 Schunck, M., Steude, J., Saftig, P., Proksch, E., and Schutze, S. (2004) J. Cell Sci. 117, 35. Bro¨mme, D., Li, Z., Barnes, M., and Mehler, E. (1999) Biochemistry 38, 2377–2385 2295–2307

JUNE 9, 2006•VOLUME 281•NUMBER 23 JOURNAL OF BIOLOGICAL CHEMISTRY 15899 58 CHAPTER 3

Colocalization of cystatin M/E and cathepsin V in lamellar granules and corneodesmosomes suggests a functional role in epidermal differentiation

Patrick L.J.M. Zeeuwen, Akemi Ishida‐Yamamoto, Ivonne M.J.J. van Vlijmen‐Willems, Tsing Cheng, Mieke Bergers, Hajime Iizuka, and Joost Schalkwijk

This chapter has been published in: J.Invest.Dermatol. (2007) 127, 120‐128

59 ORIGINAL ARTICLE

Colocalization of Cystatin M/E and Cathepsin V in Lamellar Granules and Corneodesmosomes Suggests a Functional Role in Epidermal Differentiation Patrick L.J.M. Zeeuwen1,3, Akemi Ishida-Yamamoto2,3, Ivonne M.J.J. van Vlijmen-Willems1, Tsing Cheng1, Mieke Bergers1, Hajime Iizuka2 and Joost Schalkwijk1

Cystatin M/E is a cysteine protease inhibitor with two distinct binding sites for papain-like cysteine proteases (family C1) and the asparaginyl endopeptidase (AEP) legumain of family C13. We have previously demonstrated that deficiency of cystatin M/E in mice causes ichthyosiform skin changes and barrier disruption, which could be caused by unrestrained AEP activity. Recently, we provided biochemical evidence that human cathepsin V (CTSV) and cathepsin L (CTSL) are additional biological targets for human cystatin M/E. To address the possible role of these three proteases and their inhibitor in epidermal differentiation, we investigated the localization of these proteins in normal human skin. Whereas CTSL and AEP were broadly expressed in epithelial cells of the skin, we found a specific colocalization of cystatin M/E and CTSV in the stratum granulosum and in the root sheets of the hair follicle, using immunofluorescence microscopy. Immunoelectron microscopy revealed that cystatin M/E and CTSV are separately transported within the lamellar granules. Cystatin M/E was also found in the extracellular space in the stratum corneum associated with corneodesmosomes, where it was closely associated with CTSV. Based on the striking stratum-specific colocalization of cystatin M/E and CTSV, we propose that these molecules could have an important role in epidermal differentiation and desquamation. Journal of Investigative Dermatology (2007) 127, 120–128. doi:10.1038/sj.jid.5700480; published online 27 July 2006

INTRODUCTION absence of cystatin M/E and appearance of asparaginyl Cystatin M/E is a cysteine protease inhibitor whose expression is endopeptidase (AEP) activity colocalizes with the observed largely confined to cutaneous epithelia. In human skin, it is pathology in ichq mice, that is, excessive cornification in hair expressed in the sweat glands, hair follicles, and stratum follicles and epidermis. Furthermore, these investigations have granulosum (SG) of the epidermis. In addition to its function as a revealed that AEP could indirectly mediate the activation of cysteine protease inhibitor, cystatin M/E could also serve as a TGase 3 zymogen, via lysosomal cysteine proteases in vitro target for crosslinking by transglutaminases (TGase) (Zeeuwen (Zeeuwen et al., 2004). A biochemical study from another et al., 2001). We have previously reported that a null mutation group has indicated that cystatin M/E in vitro binds to AEP with in the mouse cystatin M/E gene (Cst6) causes the murine ichq high affinity (Alvarez-Fernandez et al., 1999), which was an phenotype, which is characterized by neonatal lethality and unexpected finding, as the cystatins were primarily regarded as abnormalities in cornification and desquamation, demonstrating inhibitors of the papain-like cysteine proteases in the unrelated an essential role for cystatin M/E in the final stages of epidermal family C1 (Abrahamson et al., 2003). Although cystatin M/E differentiation (Zeeuwen et al., 2002). We have shown that contains a predicted binding site for papain-like cysteine proteases, no high-affinity binding for any member of this family has been published so far. Recently, we have shown that 1Department of Dermatology, Nijmegen Centre for Molecular Life Sciences, cystatin M/E is a high-affinity inhibitor of cathepsin V (CTSV) Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands 2 and cathepsin L (CTSL) by a reactive site that is distinct from the and Department of Dermatology, Asahikawa Medical College, Asahikawa, AEP-binding site (Cheng et al., 2006). On the basis of the Japan analogous sites in cystatin C (Alvarez-Fernandez et al., 1999), 3These authors contributed equally to this work. we used site-directed mutagenesis to identify the binding sites of Correspondence: Dr Patrick L.J.M. Zeeuwen, Department of Dermatology, Nijmegen Centre for Molecular Life Sciences, Radboud University Nijmegen these proteases in human cystatin M/E, and we found that AEP Medical Centre, PO Box 9101, 6500 HB Nijmegen, The Netherlands. and papain-like cysteine proteases are inhibited by two distinct E-mail: [email protected] non-overlapping sites. In addition, we showed that human CTSL Abbreviations: AEP, asparaginyl endopeptidase; CTSL, cathepsin L; is the elusive enzyme that is able to process and activate human CTSV, cathepsin V; KLK, kallikrein; LG, lamellar granule; SC, stratum TGase 3 (Cheng et al., 2006). Altogether, these data strongly corneum; SG, stratum granulosum; TGase, transglutaminase imply an important role for cystatin M/E and its protease targets Received 10 March 2006; revised 25 May 2006; accepted 26 May 2006; published online 27 July 2006 in the differentiation process of human epidermis.

Journal of Investigative Dermatology (2007), Volume 127 & 2006 The Society for Investigative Dermatology 60 PLJM Zeeuwen et al. Cystatin M/E in Lamellar Granules

Desquamation is the final event in terminal differentiation binding site. For a comprehensive biochemical and kinetic of the epidermis, which occurs by the action of proteolytic analysis of the interaction between cystatin M/E and its target enzymes. During desquamation, breakdown of lipids in the proteases, see Cheng et al. (2006). intercellular spaces and loss of residual intercellular desmo- In order to examine the localization of cystatin M/E and its somal connections leads to cell shedding of the dead physiological target proteases CTSV, CTSL, and AEP in corneocytes from the stratum corneum (SC) surface (Elias, normal human skin, immunofluorescence microscopy was 2005). A number of studies have shown the involvement of performed (Figures 1–3). A strong and consistent staining of different epidermal proteases in the progressive degradation cystatin M/E was found in the SG (Figure 1a) and in the inner of the desmosomal proteins desmocollin-1, desmoglein-1, half of the root sheet of the hair follicle (Figure 1d). Staining of plakoglobin, and corneodesmosin (Lundstrom and Egelrud, the SC was present but often discontinuous (Figure 1a). 1990; Suzuki et al., 1996; Simon et al., 2001; Caubet et al., Furthermore, cystatin M/E was found in the secretory coil 2004). A convincing role in desquamation has been demons- epithelium of the eccrine sweat glands (Figure 1g), but not in trated for at least two epidermis-specific serine proteases, the the ductal portion. In the sebaceous glands, only the inner SC chemotryptic and SC tryptic enzymes (currently known as mature cells were positive (Figure 1j). Cystatin M/E specifi- kallikreins KLK7 and KLK5), which are able to degrade cally colocalizes with CTSV in the most superficial granular corneodesmosome proteins (Hansson et al., 1994; Brattsand layer, whereas colocalization in the SC shows a patchy and Egelrud, 1999; Ekholm et al., 2000; Caubet et al., 2004). distribution (Figure 1c and m). Double immunostaining KLK7 and KLK5 are highly expressed in the SG, and are revealed that cystatin M/E was expressed earlier during present in the intercellular spaces of the SC (Sondell et al., epidermal differentiation than CTSV (Figure 1c and m). At 1995). Both these enzymes are most active at neutral to higher magnifications, immunofluorescent staining for cysta- alkaline pH, but also showed some activity at a more acidic tin M/E and CTSV shows a typical granular pattern (Figure 1n pH. In addition to these serine proteases, both cysteine and and o). Cystatin M/E and CTSV also colocalize in the most aspartic protease activities with acidic pH optima have been inner layer of the root sheet of the hair follicle (Figure 1f), and found in the SC (Horikoshi et al., 1999; Watkinson, 1999). in a few cells of the sebaceous glands (Figure 1l). CTSL and We have previously suggested that a new cysteine protease of AEP are both broadly expressed throughout the epidermis and late epidermal differentiation, stratum corneum thiol pro- its appendages (Figures 2 and 3). At higher magnifications, tease, could be a possible target for cystatin M/E (Zeeuwen immunofluorescent staining for CTSL and AEP shows a diffuse et al., 2001). Stratum corneum thiol protease appears to be a staining pattern (not shown). Colocalization of cystatin M/E specific product of differentiated keratinocytes, and further- with CTSL was found at the cellular level in the SG, in the more it is apparently not stored within these cells but is inner half of the hair follicles root sheet, in the secretory coil actively secreted (Watkinson, 1999). Because stratum cor- epithelium of the eccrine sweat glands, and in the inner neum thiol protease is able to degrade desmocollin-1 in in mature cells of the sebaceous glands (Figure 2c, f, i, and l). vitro degradation experiments, an extracellular proteolytic We could not detect CTSL expression in the SC (Figure 2c). function for this protease in mediating desquamation has Furthermore, staining for cystatin M/E overlapped with that been proposed. It was shown by others that ‘‘stratum for AEP in the SG and the SC, in the inner half of the root corneum thiol protease’’ activity was also able to hydrolyze sheet of the hair follicle, in the secretory coil epithelium of corneodesmosin, and finally it appears that this so-called the eccrine sweat glands, and in the inner mature cells of the stratum corneum thiol protease was identical to CTSL2 sebaceous glands (Figure 3c, f, i, and l). (Bernard et al., 2003), also known as CTSV (Adachi et al., As shown above, using immunofluorescence microscopy, 1998). Our recent data on CTSV and CTSL as additional we could demonstrate co-expression of cystatin M/E and its biological targets for cystatin M/E suggest that these enzymes presumed targets at the cellular level, in the most superficial could possibly play an important regulatory role in epidermal granular cells of the epidermis. In order to define the barrier formation and desquamation. expression of cystatin M/E and its target proteases at the In this study, we sought to address the possible role of subcellular level, immunoelectron microscopy was per- cystatin M/E and its physiological target proteases in formed. We used two different methods that we have epidermal differentiation and desquamation. We have recently described: post-embedding and cryo-ultramicrotomy examined the localization and compartmentalization of electron microscopy (Ishida-Yamamoto et al., 2004). In the cystatin M/E, CTSV, CTSL, and AEP in normal human skin first method, LG appeared as isolated granules as in by immunofluorescence microscopy and immunoelectron conventional transmission electron microscopy. In contrast, microscopy. We found that cystatin M/E and CTSV are the cryo-ultramicrotomy method showed different images of separately transported within lamellar granules (LGs). Both LG, namely branched tubules with a partial internal lamellar proteins are secreted in the extracellular space where they structure. Using post-embedding immunoelectron micro- associate with corneodesmosomes, suggesting a regulatory scopy, we could show that cystatin M/E and CTSV gold role for cystatin M/E in the desquamation process. labels are found in LG within the cells of the SG (Figures 4a and 5b), and in the extracellular spaces of the SC where both RESULTS molecules are associated with desmosomes (Figures 4b–d and We have recently shown that cystatin M/E is an inhibitor of 5c and d). We used cryo-ultramicrotomy electron microscopy CTSV and CTSL by a reactive site that is distinct from the AEP- to study the coexpression of cystatin M/E and CTSV. It was

www.jidonline.org 61 PLJM Zeeuwen et al. Cystatin M/E in Lamellar Granules

a Cystatin M/E b CTSV c Merge

d e f

g hi

jkl

m n o

Figure 1. Colocalization of cystatin M/E and CTSV in human normal skin. Immunofluorescence staining of (green color; a, d, g, j) cystatin M/E and (red color; b, e, h, k) CTSV, and (yellow-orange merge color; c, f, i, l) double staining for both proteins in (a–c) the epidermis, (d–f) the hair follicle, (g–i) the sweat glands, and (j–l) the sebaceous glands. The rectangular area marked in (c) is shown enlarged in (m). Note that the SC is detached from the SG. The rectangular areas marked in (d and e) are shown enlarged in (n and o). DNA staining by 40,6-diamidine-20-phenylindole dihydrochloride is colored in blue. Bar= (a–I)50mm; (j–l) 100 mm; and (m–o) 12.5 mm.

found that both proteins are aggregated but not mixed with DISCUSSION each other in the LG of the cells of the SG (Figure 6a and b). In this study, we analyzed the tissue localization and In the extracellular spaces of the SC, cystatin M/E was subcellular compartmentalization of cystatin M/E and its colocalized with CTSV, and both proteins are found to be physiological target proteases in normal human skin. We closely associated with desmosomes (Figure 6c and d). showed for the first time that cystatin M/E and CTSV are Furthermore, post-embedding immunoelectron microscopy separately transported within LG. Moreover, both proteins are revealed that AEP and CTSL are not LG proteins. Labeled secreted in the extracellular space where they associate with proteins were found in the cytoplasm of SG cells (Figures 7 corneodesmosomes, suggesting a regulatory role for cystatin and 8), where some of the labeling was associated with M/E and CTSV in epidermal desquamation. keratin filaments. No expression of AEP and CTSL was found Here, we showed by immunofluorescence microscopy in the extracellular spaces of the SC. that cystatin M/E appeared to be present together with AEP,

Journal of Investigative Dermatology (2007), Volume 127 62 PLJM Zeeuwen et al. Cystatin M/E in Lamellar Granules

a Cystatin M/E b CTSL c Merge

d e f

g h i

j k l

Figure 2. Colocalization of cystatin M/E and CTSL in human normal skin. Immunofluorescence staining of (green color; a, d, g, j) cystatin M/E and (red color; b, e, h, k) CTSL, and (yellow-orange merge color; c, f, i, l) double staining for both proteins in (a–c) the epidermis, (d–f) the hair follicle, (g–i) the sweat glands, and (j–l) the sebaceous glands. DNA staining by 40,6-diamidine-20-phenylindole dihydrochloride is colored in blue. Bar= (a–I)50mm and (j–l) 100 mm.

CTSL, and CTSV in some parts (mainly the most superficial vesicular population, but are consistent with sections through cells) of the epidermis (Figures 1–3). The granular staining a tubular network (Elias et al., 1998), which suggest that pattern in the cells that we found for cystatin M/E and CTSV keratinocyte LG are trans-Golgi network structures and not suggested that both proteins are associated with LG. To vesicles. Norle´n has therefore proposed a new model for the evaluate this hypothesis, post-embedding immunoelectron delivery of the contents of LG to the plasma membrane microscopy was performed, and we indeed could confirm the (Norlen, 2001). This so-called ‘‘membrane folding model’’ localization of cystatin M/E and CTSV in LG, which appear as implies that the trans-Golgi network and LG of the uppermost isolated oval-shaped granules (Figures 4 and 5). LG, or SG cells as well as the multilamellar lipid matrix of the lamellar bodies, are secretory granules of keratinizing intercellular space at the border zone between SG and SC squamous epithelia and are thought to be essential in barrier could be representations of one and the same continuous formation and desquamation. It is widely accepted that LG membrane structure. Indeed, direct visualization of the skin originate from the Golgi apparatus, and that the assembly of barrier formation process using cryotransmission electron these secretory granules begins at the trans most cisternae of microscopy confirmed the presence of folded multilamellar the Golgi complex (Madison, 2003). Recent work on the continuous structures adjacent to apparent active sites of skin trans-Golgi network, which is the highly tubulated sorting barrier formation (Norlen et al., 2003). and delivery part of the Golgi apparatus, has revealed that Recently, it was shown that epidermal LG transport transport to the plasma membrane is mediated by pleio- different cargoes (e.g., cathepsin D, glucosylceramides, morphic tubulovesicular structures, rather than by vesicles corneodesmosin, and KLK7) as distinct aggregates, which (Mironov et al., 2001). It was shown by high magnification were delivered to the apical region of granular keratinocytes electron microscopy that LG do not represent a uniform (Ishida-Yamamoto et al., 2004). Furthermore, it was shown

www.jidonline.org 63 PLJM Zeeuwen et al. Cystatin M/E in Lamellar Granules

a Cystatin M/E b AEP c Merge

d e f

g h i

j k l

Figure 3. Colocalization of cystatin M/E and AEP in human normal skin. Immunofluorescence staining of (green color; a, d, g, j) cystatin M/E and (red color; b, e, h, k) AEP, and (yellow-orange merge color; c, f, i, l) double staining for both proteins in (a–c) the epidermis, (d–f) the hair follicle, (g–i) the sweat glands, and (j–l) the sebaceous glands. DNA staining by 40,6-diamidine-20-phenylindole dihydrochloride is colored in blue. Bar= (a–I)50mm and (j–l) 100 mm.

that the serine protease inhibitor LEKT1 (lympho-epithelial 4–6), suggesting that these molecules could have an kazal type inhibitor) is localized in LG, separated from its important role in desquamation. Because an acidic pH epidermis-specific target proteases KLK7 and KLK5, and is dominates in the outer SC layers (Elias, 2004), it is secreted in the extracellular spaces of the superficial SG conceivable that CTSV activity in these layers fulfills a (Ishida-Yamamoto et al., 2005). LEKT1 is encoded by the significant function in the desquamation process (e.g., SPINK5 gene, which is the causative gene of the severe degradation of desmosomal constituents), as opposed to ichthyotic skin disorder called Netherton syndrome (Chava- serine proteases which act in the neutral pH environment of nas et al., 2000). Recent studies have reported that defective the lower SC layers. One could speculate that separate inhibitory regulation of epidermal serine protease activity transport of cystatin M/E and CTSV by LG may prevent owing to LEKT1 deficiency results in the early degradation of interaction within the cells. Altogether, we propose a desmosomal proteins, possibly causing the skin pathology in regulatory role for cystatin M/E in desquamation by control- Netherton syndrome (Caubet et al., 2004; Ishida-Yamamoto ling the CTSV activity in the SC (Figure 9). et al., 2005). We have observed a similar distribution for In contrast with cystatin M/E and CTSV, the staining cystatin M/E and CTSV in LG using the cryo-ultramicrotomy pattern for AEP and CTSL that we observed using fluoresence method. Cystatin M/E and CTSV are both localized in the LG, microscopy was diffuse and was not associated with granular however, double staining revealed that these proteins are structures. Immunoelectron microscopy showed that AEP and localized separately within the presumed trans-Golgi LG CTSL are not located within LG, but are found in the network (Figure 6). Moreover, cystatin M/E and CTSV are cytoplasm of the SG cell (Figures 7 and 8). As both AEP and detected in the extracellular spaces of the upper SC layers CTSL are considered to be endosomal/lysosomal proteins, the where both proteins are associated with desmosomes (Figures cytosolic localization of both molecules in SG keratinocytes

Journal of Investigative Dermatology (2007), Volume 127 64 PLJM Zeeuwen et al. Cystatin M/E in Lamellar Granules

a a b

d

d 200 nm b

0.5 m 200 nm d c d 100 nm C5 c C4

d d C3 d 200 nm d C2 C4 d C1

d G 200 nm 1 m 200 nm

Figure 4. Cystatin M/E is an LG protein and is associated with desmosomes Figure 5. CTSV is an LG protein and is associated with desmosomes in the after secretion. Post-embedding immunoelectron microscopy using Lowicryl SC. Post-embedding immunoelectron microscopy using Lowicryl K11M resin. HM20 resin. (a) Cystatin M/E labels (5 nm gold, black arrows) are within the (a, b) SG of the epidermis where CTSV labels (5 nm gold, black arrows) are LGs. Note oval-shaped LGs with partial lamellar internal structures (white associated with LGs within the cells (white arrowhead) and those released arrowheads). (b) In the SC, cystatin M/E labels are associated with into the extracellular spaces (upper right) between desmosomes (d). (c, d)SG desmosomes (d). (c, d) Double labeling of cystatin M/E (5 nm gold, smaller (G) and the 1st (C1) to the 5th (C5) layer of the SC. CTSV labels are associated arrows) and desmoglein-1 (10 nm gold, larger arrows). with desmosomes. The rectangular areas marked in (a) and (c) are shown at higher magnification in (b) and (d), respectively. is at least remarkable. However, other studies have also specific enzyme responsible for the crosslinking of loricrin shown that precursors of CTSL are found in the cytoplasm of and small proline-rich proteins during the cornification epidermal keratinocytes (Kawada et al., 1997), and in the process. The presence of TGase 3 in the cytoplasm of the cytoplasm of cortical thymic epithelial cells (Arudchelvan cell is in accordance with the cytosolic localization of CTSL. et al., 2002). It has been coined that AEP, like CTSL, could be At this point, we think that in normal healthy skin, post- secreted from cells under some conditions, and may be active translational proteolytic processing of TGase 3 is completed in the pericellular and intracellular environment (Barrett at the final stage of the granular cell layer when cell integrity et al., 2004). Furthermore, recent studies have reported is lost and proteases are released leading to progressively increasing evidence that lysosomal proteases play important more TGase 3 enzyme activity. Despite the apparent loss of roles in physiological processes not solely restricted to cell integrity, organization, and compartmentalization in the lysosomes (Turk et al., 2002). Protease activity and regulation terminally differentiating keratinocyte, this must be a highly outside lysosomes potentially contributes to the propagation ordered and well-orchestrated process. We suppose that of apoptosis, a process that is distinct from terminal processing and activation of TGase 3 in terminally differen- differentiation of the epidermis, but nevertheless shares some tiating keratinocytes could be regulated by the inhibitory molecular and cellular features (Lippens et al., 2005). The activity of cytoplasmic cystatin M/E against CTSL (Figure 9). signals for the release and activation of lysosomal proteases AEP inhibition by cystatin M/E could have a regulatory role in during cornification are unknown, and also the mechanism CTSL activity as AEP is involved in CTSL processing (Maehr by which they degrade the cellular organelles. et al., 2005). In addition, TGase 3 is also expressed in hair In the epidermis, there is a balanced regulation of protease follicles, where it crosslinks soluble trichohyalin, a major activity that, when disturbed, could lead to faulty cornifica- structural protein in hair follicle differentiation (Tarcsa et al., tion processes in the epidermis and upper part of the hair 1997). In this study, we show that cystatin M/E and its target follicle (Zeeuwen, 2004). We have recently shown that proteases are all present in the root sheet of the hair follicle, human CTSL is the elusive enzyme that is able to process and which suggests that these molecules could play an important activate human TGase 3 (Cheng et al., 2006), an epidermis- role in hair follicle morphogenesis. Moreover, both cystatin

www.jidonline.org 65 PLJM Zeeuwen et al. Cystatin M/E in Lamellar Granules

a b a b

C C1

G G

500 nm 200 nm

Figure 8. CTSL is not an LG protein. Post-embedding immunoelectron microscopy using Lowicryl HM20 resin. (a) A lower magnification view of the SC and SG. The rectangular area is shown at higher magnification in (b), which shows CTSL labels (5 nm gold, black arrows) are not associated with LGs (white arrowheads).

1 m 100 nm c d Cystatin M/E

C5 C5 C4 Control Control C3 −− C2 Pro-cathepsins Pro-AEP Active-cathepsins Active C1 pH AEP pH CTSV CTSL G intra-and intra-and 1 m 100 nm extracellular extracellular

Figure 6. Cystatin M/E and CTSV are colocalized at the desmosomes after Degradation of TGase 3 Crosslinking of secretion. Immunoelectron microscopy using the cryo-ultramicrotomy Desquamation desmosomal proteins activation stratum corneum proteins method. (a) Granular layer (G) and the cornified layer (C1). The marked rectangular area is shown at higher magnification in (b). Cystatin M/E labels Figure 9. Regulation of epidermal protease activity by cystatin M/E. (10 nm gold, lager arrows) are aggregated and not mixed with CTSV labels A simplified scheme of the presumed regulatory role of cystatin M/E in (5 nm gold, smaller arrows) in the LGs of the granular layer. (c) Granular layer processes that control epidermal cornification and desquamation. Cystatin M/ and the 1st to 5th cornified layer. (d) Higher magnification of the area marked E is an inhibitor of AEP and the cysteine proteases CTSL and CTSV. Inhibition in (c), where cystatin M/E and CTSV labels are colocalized at desmosomes. of CTSV activity could possibly regulate desquamation, as CTSV is able to degrade (corneo)-desmosomal proteins. Inhibition of CTSL activity by cystatin M/E is presumably important in the cornification process, as CTSL is the elusive processing and activating enzyme for TGase 3. Inhibition of AEP might a C b regulate the processing of (pro)-cathepsins, a process that is also under control d d G of intra- and extracellular pH changes.

that human CTSV can compensate for murine CTSL (Hagemann et al., 2004; Reinheckel et al., 2005). We conclude that cystatin M/E has probably two important 500 nm 200 nm regulatory functions in epidermal differentiation and hair Figure 7. AEP is not an LG protein. Post-embedding immunoelectron follicle morphogenesis. First, cystatin M/E could have a role microscopy using Lowicryl HM20 resin. (a) A lower magnification view of the in desquamation by the regulation of CTSV protease activity, SC and SG. The rectangular area is shown at higher magnification in (b), which is involved in the degradation of corneodesmosomal which shows AEP labels (10 nm gold, black arrows) are not associated with LGs (white arrowheads) but with keratin filaments. d, desmosomes. components. Second, cystatin M/E could regulate cross- linking of structural proteins by TGase 3 in the cornification process of the epidermis and the hair follicle by controlling CTSL and AEP activities. We have recently developed a M/E-deficient mice (as described in the Introduction) as well human reconstructed skin model that recapitulates the CTSL-deficient mice show a skin/hair phenotype (Roth et al., expression pattern of several of the proposed players in this 2000; Benavides et al., 2002). Recently, the phenotype of process (unpublished results), and as such will be a valuable CTSL knockout mice was rescued by transgenic epidermal re- tool in further delineation of the role of cystatin M/E and its expression of either murine CTSL or human CTSV, indicating target proteases in the final steps of epidermal differentiation.

Journal of Investigative Dermatology (2007), Volume 127 66 PLJM Zeeuwen et al. Cystatin M/E in Lamellar Granules

Together with the use of knockout mouse models that we are cryosections was carried out according to the Tokuyasu method currently generating (cystatin M/E – AEP and cystatin M/E – (Tokuyasu, 1989), with slight modifications as described previously CTSL double-knockout mice), this work extends our know- (Ishida-Yamamoto et al., 2004). ledge on the components that are involved in epidermal cornification and desquamation. We expect that these data CONFLICT OF INTEREST will lead to further understanding of human disorders with The authors state no conflict of interest. disturbed skin barrier function. ACKNOWLEDGMENTS MATERIALS AND METHODS P.Z. was financially supported by VENI-Grant 916.56.117 from ZONMW, The Netherlands Organization for Health Research and Development. A.I-Y. Antibodies was supported by grants from the Ministry of Education, Culture, Sports, Primary antibodies used in this study were as follows: affinity- Science, and Technology, and the Ministry of Health, Labor, and Welfare of purified polyclonal rabbit anti-human cystatin M/E antibodies Japan. We thank Dr Colin Watts (Division of Cell Biology and Immunology, University of Dundee, UK) for providing polyclonal sheep anti-human AEP (Zeeuwen et al., 2001), affinity-purified polyclonal sheep anti- antibodies. human AEP antibodies (Li et al., 2003), monoclonal mouse anti- human CTSV, monoclonal rat anti-human/mouse CTSL, monoclonal mouse anti-human cystatin M/E (all from R&D Systems, Minneapo- REFERENCES lis, MN), and mouse anti-DSG1-P23 (Progen, Heidelberg, Germany). Abrahamson M, Alvarez-Fernandez M, Nathanson CM (2003) Cystatins. For immunofluorescence analysis, the following secondary Biochem Soc Symp 70:179–99 reagents were used: Alexa-Fluor 488 goat anti-rabbit IgG highly Adachi W, Kawamoto S, Ohno I, Nishida K, Kinoshita S, Matsubara K et al. cross-absorbed, and Alexa-Fluor 594 goat anti-rat, goat anti-mouse, (1998) Isolation and characterization of human cathepsin V: a major and donkey anti-sheep IgG highly cross-absorbed (Molecular proteinase in corneal epithelium. Invest Ophthalmol Vis Sci 39:1789–96 Probes, Eugene, OR). Secondary antibodies used for immunoelec- Alvarez-Fernandez M, Barrett AJ, Gerhartz B, Dando PM, Ni J, Abrahamson tron microscopy were 5 or 10 nm gold-conjugated goat anti-mouse M (1999) Inhibition of mammalian legumain by some cystatins is due to a novel second reactive site. J Biol Chem 274:19195–203 IgG, gold-conjugated rabbit anti-goat IgG, and gold-conjugated goat anti-rabbit IgG (all from Amersham Bioscience, Buckinghamshire, Arudchelvan Y, Nishimura Y, Tokuda N, Sawada T, Shinozaki F, Fukumoto T (2002) Differential expression of MHC class II antigens and cathepsin L by UK), and 10 nm gold-conjugated donkey anti-sheep IgG (BBInterna- subtypes of cortical epithelial cells in the rat thymus: an immunoelectron tional, Cardiff, UK). microscopic study. J Electron Microsc (Tokyo) 51:173–81 Barrett AJ, Rawlings ND, Woessner JF (2004) Handbook of proteolytic Immunofluorescence analysis enzymes. 2nd ed. London: Elsevier Academic Press, 1–2140 Biopsies of normal skin from healthy volunteers were rinsed in Benavides F, Starost MF, Flores M, Gimenez-Conti IB, Guenet JL, Conti CJ phosphate-buffered saline, fixed for 4 hours in buffered 4% formalin, (2002) Impaired hair follicle morphogenesis and cycling with abnormal epidermal differentiation in nackt mice, a cathepsin L-deficient muta- and embedded in paraffin wax (skin biopsies were collected with tion. Am J Pathol 161:693–703 informed consent using protocols approved by the local medical Bernard D, Mehul B, Thomas-Collignon A, Simonetti L, Remy V, Bernard MA ethics committee that complies with the Declaration of Helsinki et al. (2003) Analysis of proteins with caseinolytic activity in a human Principles). All material was cut in 7-mm sections, mounted on stratum corneum extract revealed a yet unidentified cysteine protease and identified the so-called ‘‘stratum corneum thiol protease’’ as SuperFrost slides (Menzel, Braunschweig, Germany), deparaffinized, cathepsin l2. J Invest Dermatol 120:592–600 and rehydrated. The sections were subsequently incubated with Brattsand M, Egelrud T (1999) Purification, molecular cloning, and expression primary antibodies diluted in 1% BSA/phosphate-buffered saline for of a human stratum corneum trypsin-like serine protease with possible 1 hour at room temperature. After washing in phosphate-buffered function in desquamation. J Biol Chem 274:30033–40 saline, fluorescent secondary antibodies were applied for 30 minutes Caubet C, Jonca N, Brattsand M, Guerrin M, Bernard D, Schmidt R et al. at room temperature. For double labeling with antibodies raised in (2004) Degradation of corneodesmosome proteins by two serine proteases of the kallikrein family, SCTE/KLK5/hK5 and SCCE/KLK7/hK7. different animals, a mixture of primary antibodies was applied and J Invest Dermatol 122:1235–44 this was followed by incubation with a mixture of secondary Chavanas S, Bodemer C, Rochat A, Hamel-Teillac D, Ali M, Irvine AD et al. antibodies conjugated with different fluorescent dyes. Nuclei were (2000) Mutations in SPINK5, encoding a serine protease inhibitor, cause 0 0 stained with 4 ,6-diamidine-2 -phenylindole dihydrochloride (Dako- Netherton syndrome. Nat Genet 25:141–2 Cytomation, Copenhagen, Denmark), and the slides were mounted Cheng T, Hitomi K, Vlijmen-Willems IM, de Jongh GJ, Yamamoto K, Nishi K using Prolong Gold Antifade reagent (Molecular Probes). The slides et al. (2006) Cystatin M/E is a high affinity inhibitor of cathepsin V and cathepsin L by a reactive site that is distinct from the legumain-binding were subsequently examined using an immunofluorescence micro- site. A novel clue for the role of cystatin M/E in epidermal cornification. scope (Axioskop 2 MOT, Zeiss, Sliedrecht, The Netherlands), and J Biol Chem Papers 281:15893–9 images were photographed using a digital camera (Sony DXC-390P Ekholm IE, Brattsand M, Egelrud T (2000) Stratum corneum tryptic enzyme in 3 CCD, Scanalytics, Fairfax, VA) and edited using Axiovision normal epidermis: a missing link in the desquamation process? J Invest Software (Zeiss). Dermatol 114:56–63 Elias PM (2004) The epidermal permeability barrier: from the early days at Immunoelectron microscopy Harvard to emerging concepts. J Invest Dermatol 122:vi–ix Post-embedding immunoelectron microscopy using Lowicryl K11M Elias PM (2005) Stratum corneum defensive functions: an integrated view. J Invest Dermatol 125:183–200 resin and Lowicryl HM20 resin (Chemische Werke Lowi, Waldkrai- Elias PM, Cullander C, Mauro T, Rassner U, Komuves L, Brown BE et al. burg, Germany) was performed as described previously (Ishida- (1998) The secretory granular cell: the outermost granular cell as a Yamamoto et al., 2004). Immunoelectron microscopy on ultrathin specialized secretory cell. J Investig Dermatol Symp Proc 3:87–100

www.jidonline.org 67 PLJM Zeeuwen et al. Cystatin M/E in Lamellar Granules

Hagemann S, Gunther T, Dennemarker J, Lohmuller T, Bromme D, Schule R Reinheckel T, Hagemann S, Dollwet-Mack S, Martinez E, Lohmuller T, et al. (2004) The human cysteine protease cathepsin V can compensate Zlatkovic G et al. (2005) The lysosomal cysteine protease cathepsin L for murine cathepsin L in mouse epidermis and hair follicles. Eur J Cell regulates keratinocyte proliferation by control of growth factor recycling. Biol 83:775–80 J Cell Sci 118:3387–95 Hansson L, Stromqvist M, Backman A, Wallbrandt P, Carlstein A, Egelrud T Roth W, Deussing J, Botchkarev VA, Pauly-Evers M, Saftig P, Hafner A et al. (1994) Cloning, expression, and characterization of stratum corneum (2000) Cathepsin L deficiency as molecular defect of furless: hyperpro- chymotryptic enzyme. A skin-specific human serine proteinase. J Biol liferation of keratinocytes and pertubation of hair follicle cycling. FASEB Chem 269:19420–6 J 14:2075–86 Horikoshi T, Igarashi S, Uchiwa H, Brysk H, Brysk MM (1999) Role of Simon M, Jonca N, Guerrin M, Haftek M, Bernard D, Caubet C et al. (2001) endogenous cathepsin D-like and chymotrypsin-like proteolysis in Refined characterization of corneodesmosin proteolysis during terminal human epidermal desquamation. Br J Dermatol 141:453–9 differentiation of human epidermis and its relationship to desquamation. Ishida-Yamamoto A, Deraison C, Bonnart C, Bitoun E, Robinson R, O’Brien TJ J Biol Chem 276:20292–9 et al. (2005) LEKTI is localized in lamellar granules, separated from KLK5 Sondell B, Thornell LE, Egelrud T (1995) Evidence that stratum corneum and KLK7, and is secreted in the extracellular spaces of the superficial chymotryptic enzyme is transported to the stratum corneum extracellular stratum granulosum. J Invest Dermatol 124:360–6 space via lamellar bodies. J Invest Dermatol 104:819–23 Ishida-Yamamoto A, Simon M, Kishibe M, Miyauchi Y, Takahashi H, Yoshida Suzuki Y, Koyama J, Moro O, Horii I, Kikuchi K, Tanida M et al. (1996) The S et al. (2004) Epidermal lamellar granules transport different cargoes as role of two endogenous proteases of the stratum corneum in degradation distinct aggregates. J Invest Dermatol 122:1137–44 of desmoglein-1 and their reduced activity in the skin of ichthyotic Kawada A, Hara K, Kominami E, Hiruma M, Noguchi H, Ishibashi A (1997) patients. Br J Dermatol 134:460–4 Processing of cathepsins L, B and D in psoriatic epidermis. Arch Tarcsa E, Marekov LN, Andreoli J, Idler WW, Candi E, Chung SI et al. (1997) Dermatol Res 289:87–93 The fate of trichohyalin. Sequential post-translational modifications by Li DN, Matthews SP, Antoniou AN, Mazzeo D, Watts C (2003) Multistep peptidyl-arginine deiminase and transglutaminases. J Biol Chem autoactivation of asparaginyl endopeptidase in vitro and in vivo. J Biol 272:27893–901 Chem 278:38980–90 Tokuyasu KT (1989) Use of poly(vinylpyrrolidone) and poly(vinyl alcohol) for Lippens S, Denecker G, Ovaere P, Vandenabeele P, Declercq W (2005) cryoultramicrotomy. Histochem J 21:163–71 Death penalty for keratinocytes: apoptosis versus cornification. Cell Turk B, Stoka V, Rozman-Pungercar J, Cirman T, Droga-Mazovec G, Oreic K Death Differ 12(Suppl 2):1497–508 et al. (2002) Apoptotic pathways: involvement of lysosomal proteases. Lundstrom A, Egelrud T (1990) Evidence that cell shedding from plantar Biol Chem 383:1035–44 stratum corneum in vitro involves endogenous proteolysis of the Watkinson A (1999) Stratum corneum thiol protease (SCTP): a novel cysteine desmosomal protein desmoglein I. J Invest Dermatol 94:216–20 protease of late epidermal differentiation. Arch Dermatol Res 291:260–8 Madison KC (2003) Barrier function of the skin: ‘‘la raison d’etre’’ of the Zeeuwen PL (2004) Epidermal differentiation: the role of proteases and their epidermis. J Invest Dermatol 121:231–41 inhibitors. Eur J Cell Biol 83:761–73 Maehr R, Hang HC, Mintern JD, Kim YM, Cuvillier A, Nishimura M et al. Zeeuwen PL, Vlijmen-Willems IM, Hendriks W, Merkx GF, Schalkwijk J (2005) Asparagine endopeptidase is not essential for class II MHC (2002) A null mutation in the cystatin M/E gene of ichq mice causes antigen presentation but is required for processing of cathepsin L in mice. juvenile lethality and defects in epidermal cornification. Hum Mol Genet J Immunol 174:7066–74 11:2867–75 Mironov AA, Beznoussenko GV, Nicoziani P, Martella O, Trucco A, Kweon Zeeuwen PL, Vlijmen-Willems IM, Jansen BJ, Sotiropoulou G, Curfs JH, Meis HS et al. (2001) Small cargo proteins and large aggregates can traverse JF et al. (2001) Cystatin M/E expression is restricted to differentiated the Golgi by a common mechanism without leaving the lumen of epidermal keratinocytes and sweat glands: a new skin-specific protei- cisternae. J Cell Biol 155:1225–38 nase inhibitor that is a target for cross-linking by transglutaminase. Norlen L (2001) Skin barrier formation: the membrane folding model. J Invest J Invest Dermatol 116:693–701 Dermatol 117:823–9 Zeeuwen PL, Vlijmen-Willems IM, Olthuis D, Johansen HT, Hitomi K, Hara- Norlen L, Al Amoudi A, Dubochet J (2003) A cryotransmission electron Nishimura I et al. (2004) Evidence that unrestricted legumain activity is microscopy study of skin barrier formation. J Invest Dermatol involved in disturbed epidermal cornification in cystatin M/E deficient 120:555–60 mice. Hum Mol Genet 13:1069–79

Journal of Investigative Dermatology (2007), Volume 127 68 CHAPTER 4

A regulatory role for cystatin M/E in epidermal desquamation Identification of a two‐chain form of cathepsin V

Tsing Cheng, Christiaan Dolk, Cécile Caubet, Michel Simon, Colin Watts, Marijke Kamsteeg, Guy Serre, Joost Schalkwijk, and Patrick L.J.M. Zeeuwen

Submitted

69 A Regulatory Role for Cystatin M/E in Epidermal Desquamation IDENTIFICATION OF A TWO‐CHAIN FORM OF CATHEPSIN V

Tsing Cheng1, Christiaan Dolk1, Cécile Caubet2, Michel Simon2, Colin Watts3, Marijke Kamsteeg1, Guy Serre2, Joost Schalkwijk1, and Patrick L.J.M. Zeeuwen1 1Department of Dermatology, Nijmegen Centre for Molecular Life Sciences, Radboud University Nijmegen Medical Centre, PO Box 9101, 6500 HB Nijmegen, The Netherlands, 2University of Toulouse III‐CNRS UMR5165, Institut Fédératif de Recherche Claude de Préval, IFR30 (INSERM–CNRS–Université Paul Sabatier–Centre Hospitalier Universitaire de Toulouse), 31059 Toulouse, France, and 3Division of Cell Biology & Immunology, School of Life Sciences, University of Dundee, DD1 5EH Dundee, Scotland, United Kingdom

Desquamation is the physiological process of shedding dead corneocytes from the stratum corneum of the epidermis. This process requires degradation of corneodesmosomes, intercellular junctions that provide strong cell‐cell adhesion. Recently, we reported that cathepsin V and its natural inhibitor cystatin M/E co‐localize in the stratum granulosum and the stratum corneum of human epidermis. In the extracellular space of the upper stratum corneum layers both proteins are closely associated with corneodesmosomes, indicating a potential role for these molecules in epidermal desquamation. In the present study we analyzed the ability of cathepsin V to cleave recombinant and natural forms of corneodesmosomal components. Cathepsin V proteolyses desmoglein‐1, desmoglein‐3, and corneodesmosin at pH 5.5, which resembles the acidic environment of the upper stratum corneum layers. At a neutral pH of 7.0, which is the pH environment of the lower stratum corneum layers, little proteolysis was observed and the corneodesmosomal proteins remained largely unaffected. Furthermore, we found that legumain, another physiological target protease of cystatin M/E, is able to process active single‐chain cathepsin V in vitro into a previously unknown active two‐chain form. Subsequently, we determined the cleavage site by N‐terminal sequencing and MALDI‐MS analysis. Immunoblotting of cathepsin V in human epidermal extracts shows that the two‐ chain form of cathepsin V exists in vivo. Altogether, we propose a regulatory role for cystatin M/E in desquamation by controlling cathepsin V protease activity in the stratum corneum.

INTRODUCTION

During terminal differentiation of the epidermis, keratinocytes from the basal layer gradually differentiate into the dead corneocytes of the stratum corneum (SC), and are eventually shed from the skin surface by the process of desquamation. Desquamation

70 requires degradation of corneodesmosomes, which are modified desmosomes of the epidermis that provide strong cell‐cell adhesion. These functions are important in maintaining the integrity of the tissue and in providing the cells resistance to physical stress and pressure1,2. Desmosomes are composed of proteins from multiple families (cadherins, armadillo proteins, and plakins), but the adhesive intercellular junction is formed by the extracellular domains of two types of transmembrane cadherins: desmogleins and desmocollins3. Four desmogleins (DSG1‐4) and three desmocollins (DSC1‐3) have been identified, all clustered on chromosome 18, but each with its own epidermal expression pattern, e.g. DSG1 and DSC1 are more expressed in the upper granular layers, whereas DSG3 and DSC3 are expressed in the basal and spinous layers. These differential expression patterns could have a role in epidermal morphogenesis as DSG1 and DSC1 seem to promote strong cell‐cell adhesion and keratinocyte terminal differentiation, whereas expression of DSG3 and DSC3 may be related to inhibition of differentiation and stimulation of keratinocyte proliferation4,5. In corneodesmosomes the intercellular junction is enhanced by the secreted glycoprotein corneodesmosin (CDSN), which has homophilic, adhesive properties6,7. Various human diseases and mouse models have shown the importance of these adhesive components in epidermal barrier formation and function2,4,8,9. Proteolysis of corneodesmosomal proteins is essential for desquamation, and this process is mediated by secreted skin proteases, primarily serine, but also aspartic and cysteine proteases10‐12. Previous studies have mainly focused on the potential role of serine proteases in desquamation, especially the kallikreins (KLKs). KLK5 and KLK7 are able to degrade DSG1, DSC1, and CDSN6,13, while other KLKs such as KLK6 and KLK14 also have been suggested as desquamatory enzymes14. Mutations in the SPINK5 gene, encoding the KLK inhibitor lympho‐epithelial Kazal type inhibitor (LEKTI), have been identified as the cause of Netherton syndrome, which features ichthyosis and hair shaft defects15. Patients with Netherton syndrome show reduced levels of DSG1 and DSC1 while KLK5 and KLK7 activities were increased due to the absence of their inhibitor16. Since KLKs show optimal activity at neutral pH, proteolysis of the corneodesmosomal proteins by KLKs occurs faster at a neutral pH than at an acidic pH14. The gradual change in pH of the SC from neutral in the lower layers to acidic in the upper layers has implications for KLK activity and favors the activity of other proteases such as the acid cathepsins. Cystatin M/E is a cysteine protease inhibitor that is highly expressed in the stratum granulosum and SC of human epidermis, hair follicles, sebaceous glands, and sweat glands17. We have previously demonstrated the importance of cystatin M/E in epidermal homeostasis in a mouse model. The cystatin M/E deficient ichq mouse, which is characterized by abnormal SC and hair follicle formation, displays a disturbed skin barrier function and neonatal death18‐20. The mouse model taught us that cystatin M/E was indispensible for correct barrier formation, and provided evidence for a role of cystatin M/E target proteases in transglutaminase‐mediated crosslinking of structural proteins. In a next step we decided to study the role of human cystatin M/E and its target proteases more closely, using in vitro biochemical and cell biological approaches. As a result, we have revealed a novel

71 biochemical pathway that controls skin barrier formation by regulation of both cornification and desquamation of the SC in humans. Cystatin M/E is a central molecule in this pathway as it controls the activity of the papain‐like cysteine proteases cathepsins L and V (CTSL; CTSV) as well as the asparaginyl endopeptidase legumain (LGMN) via two distinct binding sites21. These and other findings revealed a discrepancy between the mouse and human pathway, because mice lack the CTSV orthologue and mouse LGMN is not inhibited by mouse cystatin M/E (unpublished data). We found that cystatin M/E regulates crosslinking of structural proteins by transglutaminase 3 (TGM3) in the cornification process, by controlling CTSL activity, as human CTSL is the elusive enzyme that processes and activates human TGM321. Furthermore, another study provided morphological evidence that cystatin M/E may have a role in desquamation by regulating CTSV activity. We have shown that CTSV and cystatin M/E co‐localize in the stratum granulosum and the SC of the epidermis, that both proteins are separately transported within lamellar granules, and that after secretion in the extracellular space of the SC both proteins are closely associated with corneodesmosomes, indicating a potential role for these molecules in epidermal desquamation22. As cystatin M/E is also a potent inhibitor of human LGMN and it has been reported that LGMN is involved in the processing of lysosomal proteases such as cathepsins B, H, and L in mice23,24, LGMN could have a regulatory role in processing of human cathepsins L and V. These cathepsins are produced as inactive prepro‐enzymes and are subsequently processed in multiple steps into mature active forms via autoactivation and/or cleavage by other, hitherto unknown, proteases. The aim of the present study was to further investigate a possible role of cystatin M/E and CTSV in epidermal desquamation. As we had already obtained evidence based on morphological data, we now complement these findings by a biochemical approach. Since CTSV has its optimal activity at acidic pH, we hypothesized that CTSV could be responsible for degradation of corneodesmosomes in the upper SC layers. Therefore, we analyzed the ability of CTSV to degrade corneodesmosomal proteins at an acidic pH as well as at a neutral pH. In addition, we investigated the ability of LGMN to process human CTSL and CTSV and vice versa as these proteases may be involved in their mutual activation. Here we show that CTSV degrades DSG1, DSG3, and CDSN at pH 5.5. Furthermore, we demonstrate that CTSV is processed in vitro by LGMN into a previously unknown active two‐chain form, which could also be demonstrated in human epidermis in vivo. These findings suggest a regulatory role for cystatin M/E in desquamation by controlling CTSV protease activity in the outer layers of the SC.

EXPERIMENTAL PROCEDURES

Protein extraction from human epidermis 4 mm biopsies were taken from lesional psoriatic skin and from normal abdominal skin of patients undergoing plastic surgery. The study was approved by the local medical ethical committee (Commissie Mensgebonden Onderzoek Arnhem‐Nijmegen) and

72 conducted according to the Declaration of Helsinki principles. Written informed consent was obtained from each volunteer. Biopsies were taken under local anesthesia with 4 mm punches. Subsequently, the epidermis was separated from the dermis by incubation for 5 min in PBS at 56°C. Epidermal proteins were extracted by repeated freeze/thawing of epidermis in extraction buffer (40 mM Tris‐HCl, 10 mM EDTA, 0.5% Nonidet‐P40, pH 7.0), followed by centrifugation for 10 min at 15000 × g. The supernatants that harbor the soluble epidermal proteins were stored at ‐80°C. For determination of CTSV in human epidermis, biopsies from normal skin and psoriasis skin were subjected to dermo/epidermal separation by incubation for 4 h in PBS (with Ca2+/Mg2+) containing 12 mg/ml dispase (Roche Diagnostics, Basel, Switzerland) at 4˚C. Epidermal proteins were extracted in PBS containing protease inhibitor cocktail Complete (Roche Diagnostics) by repeated freeze/thawing as described above. 10 µg of protein extract was subjected to SDS‐PAGE followed by immunoblotting for determination of CTSV in human epidermis.

In vitro degradation of corneodesmosomal proteins Human recombinant DSG1/Fc chimera, DSG3/Fc chimera (150 ng, both from R&D Systems, Minneapolis, MN), and CDSN (100 ng)13 were incubated with/without human recombinant CTSV (R&D Systems, activated according to manufacturers protocol for 2 h at 37°C) and human recombinant cystatin M/E21 for various time intervals at 37°C in CTSV assay buffer (0.1 M acetate, 1 mM EDTA, and 2 mM dithiothreitol) a pH of 5.5 or 7.0. Reactions contained 2 ng CTSV (with DSG1/Fc, DSG3/Fc) or 4 ng CTSV (with CDSN), while 12.5 ng cystatin M/E per ng CTSV was added to reaction samples with CTSV and its inhibitor. Human epidermal protein extracts (2.5 μg) were incubated with 100 ng CTSV at 37ºC. For visualization by Coomassie blue or silver staining the assay buffer also contained 0.05% Tween20, while 5 ng/μl bovine serum albumin was added when the reactions were followed by immunoblotting. Reactions were rapidly stopped by freezing in liquid nitrogen.

Processing of CTSV by legumain GST‐proCTSV fusion protein, GST‐proCTSL fusion protein (both from Abnova, Taipei, Taiwan), CTSV and CTSL (recombinant scCTSL protein, R&D systems; tcCTSL: purified from human liver, Sigma) were incubated with activated LGMN25 (12.5 ng LGMN per ng CTSV/CTSL) for various time intervals at 37°C in LGMN assay buffer (0.1 M phosphate pH 5.7, 2 mM EDTA, 1 mM dithiothreitol, 2.67 mM L‐cysteine, and 5 ng/µl bovine serum albumin). Reactions were rapidly stopped by freezing in liquid nitrogen.

Visualization of proteins Proteins were separated by SDS‐PAGE using the NuPAGE electrophoresis system and pre‐cast Bis‐Tris polyacrylamide gels under reducing conditions (Invitrogen). Subsequently, proteins were visualized by Coomassie blue staining (Fluka Chemie, Buchs, Switzerland), silver staining or immunoblotting. Silver staining of gels was performed using the SilverXpress kit (Invitrogen) according to manufacturer’s protocol with a minor change in

73 order to minimize background: amounts of pre‐diluted solutions were reduced by 50% in the total volume. For immunoblotting the proteins were electroblotted onto polyvinylidene difluoride membrane (Invitrogen). The membrane was incubated with monoclonal mouse anti‐human DSG1 antibodies (R&D Systems), monoclonal mouse anti‐human CDSN antibodies26, monoclonal mouse anti‐human CTSV antibodies (R&D Systems), or monoclonal rat anti‐human/mouse CTSL antibodies (R&D Systems). Detection of the proteins was established using the Phototope‐HRP detection kit (Cell Signaling, Beverly, MA) according to manufacturer’s protocol. CTSL detection required the use of biotinylated goat anti‐rat IgG antibody (Vector, Burlingame, CA) followed by incubation with anti‐biotin‐HRP antibody (Cell Signaling).

CTSV activity assay Activity of scCTSV and tcCTSV was determined with fluorometric assays using the CTSV‐specific fluorogenic substrate Z‐Leu‐Arg‐AMC (7‐amino‐4‐methyl‐coumarin, R&D Systems) in CTSV assay buffer with 5 ng/μl bovine serum albumin21. The amount of fluorescent signal released from substrate hydrolysis by CTSV during a period of time is measured as CTSV activity with a fluorescence spectrometer (Perkin Elmer, Waltham, MA).

Determination of the CTSV cleavage site CTSV was incubated with LGMN as described above followed by visualization of the proteins by Coomassie blue staining and isolation of the ~23 kDa fragment. The N‐terminal amino acid sequence of the fragment was determined by automated Edman degradation. Subsequently, the fragment was digested with trypsin and the sequence of the resulting peptides were identified by Matrix Assisted Laser Desorption Ionisation‐Mass Spectrometry (MALDI‐MS) analysis at the Nijmegen Proteomics Facility (Nijmegen, The Netherlands).

RESULTS

Degradation of corneodesmosomal proteins by CTSV To study the presumed role of CTSV in the desquamation process of the skin we incubated recombinant corneodesmosomal proteins as well as human epidermal protein extracts with recombinant human CTSV. Reactions were performed using two buffers that only differed in pH from each other (pH 5.5 and pH 7.0) as CTSV has its optimal activity at an acidic pH, while at a neutral pH its activity is greatly reduced. Recombinant DSG1/Fc is rapidly degraded by CTSV at pH 5.5; within 30 min the DSG1/Fc protein is degraded completely (Fig. 1A). At a neutral pH proteolysis of DSG1/Fc by CTSV is limited and after 240 min of incubation most of the protein is still intact. To verify that CTSV is responsible for the degradation of DSG1/Fc, we incubated DSG1/Fc with CTSV at pH 5.5 in the presence of recombinant human cystatin M/E, a natural and strong inhibitor of CTSV. This reaction showed that after 2 h of incubation the DSG1/Fc protein is still intact due to inhibition of CTSV activity by cystatin M/E, whereas in the sample without cystatin M/E all DSG1/Fc is

74 degraded (Fig. 1B). As the previous experiments were performed with recombinant fusion protein, we subsequently studied the ability of CTSV to degrade native epidermal DSG1. Human epidermal protein extracts were incubated with and without an excess of recombinant CTSV for 2 and 4 h at pH 5.5. Immunoblotting against DSG1 showed that the epidermal DSG1 band of ~110 kDa disappears in the reactions with recombinant CTSV, indicating that CTSV also degrades the natural form of DSG1 (Fig. 1C). Experiments performed with human recombinant DSG3/Fc and recombinant CDSN showed similar results (Fig. 2 and 3). Degradation of DSG3/Fc by CTSV occurred rapidly at an acidic pH, as all DSG3/Fc protein is completely degraded within 1 h (Fig. 2, left panel). A small proportion of the recombinant DSG3/Fc protein is degraded at a neutral pH, as shown by the 50 to 65 kDa

Figure 1: Degradation of recombinant DSG1/Fc and epidermal DSG1 by CTSV. A) recombinant DSG1/Fc (lane 1) was incubated with CTSV for 1, 5, 15, 30, 60, 120, and 240 min (lanes 2‐8) at 37°C. Reactions were performed in CTSV assay buffers, both pH 5.5 and pH 7.0, and visualized by silver staining. Note that at pH 5.5 full length DSC1/Fc is degraded into smaller peptides within 30 min, and that these are completely degraded thereafter. No significant degradation is visible at pH 7.0. B) recombinant DSG1/Fc (lane 1) was incubated with CTSV alone (lane 2) or with CTSV in the presence of its inhibitor cystatin M/E (lane 3) at pH 5.5 for 120 min at 37°C. Proteins were visualized by Coomassie blue staining. Note that in lane 2 DSG1/Fc is completely degraded, and this is effectively inhibited by cystatin M/E (lane 3). C) human epidermal protein extract (lane 1) was incubated with CTSV (lanes 2 and 3) at pH 5.5 for 2 h (lanes 2 and 4) and 4 h (lanes 3 and 5) at 37°C. The proteins were visualized by immunoblotting using an anti‐DSG1 antibody. A band of 110 kDa, corresponding to endogenous DSG1 is detected. In addition, three other low‐molecular weight bands, presumably breakdown products extracted from the differentiated epidermis and SC, are visible. Arrowheads indicate the intact DGS1/Fc and DSG1 proteins.

75 bands that appear in lane 2 to 6 (Fig. 2, right panel). These intermediate degradation products of DSG3/Fc are only observed within the first 30 min incubation at an acidic pH, after longer incubation periods these intermediate bands are fully degraded. Cystatin M/E prevents degradation of DSG3/Fc at pH 5.5 by inhibition of CTSV activity (lane 8). Note that the degradation of DSG3/Fc at neutral pH is fully inhibited by cystatin M/E. Furthermore, CTSV is able to degrade recombinant CDSN (52‐56 kDa) (Fig. 3A) as well as native CDSN in human epidermal protein extracts (Fig. 3B). Using epidermal protein extracts, some degradation of CDSN was also seen in the control sample (Fig.3B, lane 5), although at a lower level as in the samples with added recombinant CTSV. Apparently, prolonged incubation (8 h) of the epidermal extract allows endogenous proteases to degrade CDSN.

LGMN processes CTSV in vitro into a two‐chain form Cathepsins are produced as prepro‐enzymes that need to be processed into their mature forms via autoactivation and/or cleavage by other proteases. CTSL has two mature active forms, single‐chain (sc) CTSL and two‐chain (tc) CTSL, while CTSV is until now only known as a mature single‐chain form. Since LGMN is involved in the processing of scCTSL into tcCTSL in mice, we investigated whether human LGMN could process human CTSL and CTSV in vitro. Therefore, we performed incubations of each (pro)cathepsin with LGMN and visualized the reaction by immunoblotting. Both GST‐tagged proCTSL and proCTSV were not processed into mature proteins by LGMN (data not shown). Interestingly, incubations of mature (single‐chain) CTSV protein of ~30 kDa with LGMN in a series of time points showed the disappearance of the scCTSV, while a smaller protein band of ~23 kDa appeared, which is

Figure 2: Degradation of recombinant DSG3/Fc by CTSV. Recombinant DSG3/Fc (lane 1) was incubated with CTSV for 15, 30, 60, 120, and 240 min (lanes 2‐6); DSG3/Fc was incubated in buffer without CTSV for 240 min (lane 7); DSG3/Fc was incubated with CTSV and cystatin M/E for 240 min (lane 8). Reactions were performed in CTSV assay buffers pH 5.5 and pH 7.0 at 37°C and visualized by silver staining. As found for DSG1/Fc, only at pH 5.5 we observed breakdown of the target protein by CTSV. Arrowheads indicate the intact DGS3/Fc protein.

76 Figure 3: Degradation of recombinant and epidermal CDSN by CTSV. A) recombinant CDSN (lane 1) was incubated with CTSV for 30, 60, 120, 240, and 360 min (lanes 2‐6); CDSN was incubated alone for 360 min (lane 7); CDSN was incubated with CTSV and cystatin M/E for 360 min (lane 8). Reactions were performed in CTSV assay buffer pH 5.5 at 37°C, and visualized by immunoblotting using an anti‐CDSN antibody. After 360 min, CDSN is completely degraded by CTSV (lane 6), whereas after 360 min in buffer (lane 7) or with CTSV in the presence of cystatin M/E (lane 8) the CDSN band is still present. B) human epidermal protein extract (lane 1) was incubated with CTSV for 2, 4, and 8 h (lanes 2‐ 4); human epidermal extract was incubated alone for 8 h (lane 5). Reactions were performed in CSTV assay buffer pH 5.5 at 37°C, and visualized by immunoblotting using an anti‐ CDSN antibody. After 8 h a significant proportion of the endogenous CDSN was found to be degraded by added recombinant CTSV (lane 4), whereas 8 h of incubation in buffer

alone caused significantly less degradation. Arrowheads indicate the intact CDSN protein. probably tcCTSV, based on its size and the homology between CTSV and CTSL. Within 1 h small amounts of scCTSV are processed into tcCTSV and after 8 h only tcCTSV is detected (Fig. 4A). When scCTSV was incubated only in LGMN assay buffer for similar periods of time, the single‐chain form remained intact and no two‐chain form could be detected, suggesting that LGMN activity is responsible for processing of scCTSV into tcCTSV and that no autoprocessing occurred (data not shown). Subsequently, we examined the potential protease activity of the tcCTSV form for its identification as a truly mature form of CTSV. Samples that were partly and fully processed into tcCTSV were subjected to fluorometric enzyme assays using a CTSV‐specific substrate showing that after processing into tcCTSV the protease remains active with a similar level of activity as scCTSV (Fig. 4B). Experiments with human CTSL and LGMN did not show processing of scCTSL into tcCTSL by LGMN as the single‐chain form was still visible, while no two‐chain form appeared after incubation of scCTSL with LGMN for 8 h (Fig. 4C). Since cathepsins are capable of processing and activating proteins we investigated whether CTSL and CTSV are involved in the activation of LGMN as proLGMN also has to be processed into its active form. Various experiments in which we incubate proLGMN and mature LGMN with scCTSL, tcCTSL, and scCTSV did not show any processing of LGMN by the cathepsins (data not shown).

77

Figure 4. Processing of CTSV by LGMN. A, recombinant scCTSV (lane 1) was incubated with LGMN for 1, 2, 4, 6, and 8 h (lanes 2‐6) in LGMN assay buffer at 37°C. Reactions were visualized by immunoblotting using an anti‐CTSV antibody. A time‐dependent processing of scCTSV was found. B, CTSV activity was measured by hydrolysis of the fluorogenic substrate Z‐Leu‐Arg‐AMC for samples shown in fig. 4A: CTSV (lane 1), CTSV incubated with LGMN for 4 h (lane 4), and CTSV incubated with LGMN for 8 h (lane 6). Only a minor decrease in hydrolytic activity towards the synthetic substrate was noted upon processing of CSTV by LGMN. C, Recombinant scCTSL (lane 1) was incubated with LGMN for 8 h (lane 2) in LGMN assay buffer at 37°C. tcCTSL is shown as a control (lane 3). Reactions were visualized by immunoblotting using anti‐CTSL antibody. In contrast to CTSV (figure 4A), no processing of CTSL to its two‐chain form by LGMN was observed. The closed and open arrowheads indicate the single‐chain and two‐chain forms of the cathepsin proteins, respectively.

Determination of the CTSV cleavage site Since human CTSV is very homologous to human CTSL and mouse CTSL, protein sequence alignment of these proteins could indicate the expected cleavage site in CTSV. In human CTSL and mouse CTSL the cleavage site between the heavy and light chains is at residue 291 (Asp and Asn, respectively), which corresponds to Asn292 in CTSV (Fig. 5). N‐ terminal sequencing was performed on the tcCTSV fragment of ~23 kDa (Fig. 4A), identifying the N‐terminus as LPKS (residues 114‐117, according to the amino acid numbering of CTSV), which is the same N‐terminal sequence of the single‐chain form, confirming its identity as a mature form of CTSV and indicating that processing by LGMN occurs at the C‐terminal end.

78 Subsequently, the tcCTSV fragment was subjected to trypsin digestion followed by MALDI‐ MS analysis in order to elucidate the sequence of the resulting tryptic and semi‐tryptic peptides. A semi‐tryptic peptide is predicted to contain the end of the heavy chain and, could therefore indicate the cleavage site in CTSV. The MALDI‐MS analysis identifies a semi‐ tryptic peptide ending at Asn290 suggesting that this is indeed the cleavage site (Fig. 5). tcCTSV is present in human epidermis Our experiments have revealed an active two‐chain form of CTSV in vitro, that has not been reported in tissues before. To address if the two‐chain form exists in vivo we used immunoblotting to detect CTSV in human epidermal protein extracts from normal and psoriatic skin. Before extraction protease inhibitor cocktail was added to prevent processing of CTSV by endogenous proteases in the extracts. In the epidermal protein extracts of normal skin all forms of CTSV are detected: the inactive prepro‐ and proCTSV (~35‐37 kDa), mature scCTSV (~30 kDa), and tcCTSV (~23 kDa). Interestingly, very little CTSV could be detected in the epidermal protein extracts of psoriatic skin (Fig. 6).

DISCUSSION

Previous morphological data suggested to us that cystatin M/E and CTSV could be involved in epidermal desquamation. Here we provide complementary biochemical evidence that this might indeed be the case. In this study we analyzed the ability of CTSV to degrade the corneodesmosomal proteins DSG1, DSG3, and CDSN at an acidic and neutral pH. Degradation of the corneodesmosomal components was efficient at pH 5.5, while little proteolysis was observed at pH 7.0. During terminal differentiation dead corneocytes of the SC are eventually shed from the skin surface by the process of desquamation. This process is mediated by secreted skin proteases that proteolyze corneodesmosomal components and thereby breaking

Figure 5: Determination of the CTSV cleavage site. Protein sequence alignment and cleavage sites of human CTSL, mouse CTSL and human CTSV. The cleavage site in CTSV was determined by MALDI‐MS analysis and found to be at N290 (open arrowhead), not at N292 (asterisk). The closed arrowheads indicate the cleavage sites in human and mouse CTSL at D291 and N291, respectively. P1 amino acids are underlined. For comparison, the cleavage sites of human CTSL and mouse CTSL are given as reported previously by others44,45.

79 intercellular adhesive junctions. Several proteases, including serine, aspartic, and cysteine proteases, have been suggested as desquamatory proteins, although specific proteolysis of corneodesmosomal proteins has only been shown for the KLK family of serine proteases so far6,13,14. The present study shows that the cysteine protease CTSV is also able to proteolyze corneodesmosomal proteins, both recombinant and natural forms, and this degradation was prevented in the presence of its natural inhibitor cystatin M/E (Fig. 1‐3). Although DSG1 is the main desmoglein component in corneodesmosomes, we have also included DSG3 in this study. DSG3 is mainly expressed in the basal and spinous layers of the epidermis, while in the granular layers its expression is gradually reduced. CTSV could be responsible for degradation of DSG3 remnants in the stratum granulosum or the SC. Our initial studies in cystatin M/E deficient mice did not suggest a role for cystatin M/E target proteases in desquamation. Retention of thick scales and plugging of hair follicles was found from day 5 to neonatal death between day 10‐12. Apparently, the excessive CTSL activity (CTSV is absent in mice) caused massive crosslinking and retention of corneocytes rather than promoting desquamation and loss of scales. Obviously, the phenotype of a mouse completely deficient for cystatin M/E does not necessarily reveal the subtleties of the normal physiological functions of a protein, nor is it necessarily relevant for human interfollicular epidermis which is quite different from mouse epidermis. For this reason we have studied the cystatin M/E pathway in human systems such as skin biopsies from normal and diseased skin21,22,27 and in submerged and organotypic keratinocyte cultures27. Based on our findings with immunoelectron microscopy on human skin22 and biochemical data from the present study we conclude that cystatin M/E has a regulatory role in epidermal desquamation by controlling CTSV protease activity.

Figure 6: Natural forms of CTSV in extracts of human epidermis. Protein extracts from human epidermis of normal skin (NS, n=3) and psoriasis skin (PS, n=3) were subjected to immunoblotting using an anti‐CTSV antibody. Arrowheads show the multiple forms of CTSV, including the newly discovered two‐chain form of CTSV. Expression of CTSV appears to be strongly decreased in PS compared to NS.

80 Interestingly, it has been suggested that desquamation results from pH‐dependent activation of the several protease families as the pH of the SC changes from neutral in the lower layers to acidic in the upper layers. In the lower SC layers desquamation could be initiated by the serine proteases (KLKs), while the process is continued by cysteine proteases (CTSV) and aspartate proteases (cathepsin D) in the upper SC layers10. We showed optimal degradation of corneodesmosomal proteins by CTSV at pH 5.5 whereas it was limited at pH 7.0, which can be explained by the acidic pH optima (4.5‐6.5) of CTSV. This is opposite to what was found for the proteolytic abilities of KLKs that have neutral to alkaline pH optima, and showed decreased proteolytic activity at pH 5.4 compared to pH 7.614. Our findings suggest that desquamation is regulated by protease inhibitors that control protease activity in the SC and that gradual change in pH of the SC indicates which proteases are mainly active in each of the different layers. Recently, it was shown that the interaction between the serine protease inhibitor LEKTI and KLK5 is pH‐regulated. LEKTI had a strong affinity for KLK5 at pH 7.5, but acidification of the environment decreased the strength of the binding resulting in an increased dissociation of the LEKTI‐KLK5 complex and thereby releasing KLK528. Despite reduced KLK5 activity at lower pH, the pH‐regulated interaction between LEKTI and KLK5 might be necessary to prevent premature desquamation at the stratum granulosum‐SC interface. Although CTSV activity is limited at neutral pH, the interaction between cystatin M/E and CTSV could be influenced by pH gradient in a similar way. CTSV and CTSL belong to the family C1A of papain‐like cysteine proteases that are generally lysosomal proteases responsible for bulk proteolysis, but also can be secreted and perform tissue specific functions29,30. CTSV and CTSL are closely related to each other sharing ~80% protein sequence identity and their encoding genes are located on adjacent sites on chromosome 9q22.2, suggesting that CTSV and CTSL have evolved from an ancestral gene31‐ 33. Legumain belongs to family C13 of cysteine proteases, which is unrelated to the family of papain‐like cysteine proteases, and is strictly specific for hydrolysis of asparaginyl bonds34. All three proteases are produced as inactive enzymes that are processed into their mature active forms by autoactivation and/or processing by other proteases. However, the mechanisms behind the processing and functions of the various forms are not completely elucidated yet. After cleavage of the signal peptide from preproCTSL, processing of proCTSL into mature scCTSL is initiated by an acidic environment and happens both intermolecularly (mature CTSL is able to process proCTSL) and intramolecularly (proCTSL processes itself)35. In addition, an aspartate protease, presumably cathepsin D, is involved in proCTSL processing36. Processing of CTSV has not been studied extensively, but proCSTV is probably processed in a similar way. The cleaved propeptides of both CTSV and CTSL have shown to be inhibitors of the mature cathepsin forms37,38. LGMN is involved in the processing of scCTSL into tcCTSL, since complete absence of tcCTSL was detected in LGMN deficient mice, although the single‐ chain form alone was sufficient enough for CTSL function in vivo. LGMN deficiency did not affect processing of proCTSL to scCTSL indicating that LGMN is not involved in this processing step23,24. However, our results show that human LGMN does not process human CTSL, we do however find that LGMN processes human CTSV into a hitherto unknown active two‐chain

81 form (Fig. 4). The existence of a two‐chain form of CTSV is not surprising since two‐chain forms of both human and mouse CTSL are known23,39. No functions specifically for the single‐ chain or two‐chain forms of CTSL have been found, and it is not known whether this conversion regulates CTSL activity or substrate specificity, but the distribution and ratio of both mature forms may differ among tissues40. Similar matters regarding the single‐chain and two‐chain forms of CTSV remain to be solved. Autocatalytic processing of proLGMN produces an active intermediate form of LGMN that requires further processing by another protease to yield the final mature LGMN25. We found that human CTSV and CTSL were not able to process proLGMN or the active intermediate form of LGMN, and these results are consistent with a previous report in which the processing of the intermediate form of LGMN to the final mature form in vivo is not blocked by the cysteine protease inhibitors E64 or leupeptin25. At first sight it is unexpected that human LGMN could process human CTSV and mouse CTSL into their two‐chain forms, but not human CTSL. However, comparison of the cleavage sites of these proteases shows that cleavage between the heavy and light chains in human CTSV and mouse CTSL occurs after Asn‐residues (N290 and N291, respectively) while human CTSL has a Asp‐residue (D291) in the cleavage site. Since LGMN is specific for hydrolysis of asparaginyl bonds, the Asp‐residue in human CTSL could explain why LGMN is not able to process human CTSL into the two‐chain form. Based on the homology of CTSV with human and mouse CTSL the cleavage site in CTSV was expected to be at Asn292. However, our analysis showed that the actual cleavage site is at Asn290 (Fig. 5), probably because hydrolysis by LGMN is prevented by N‐linked glycolysation as Asn292 is potentially glycolysated due to the position of Ser29441. Our finding that both human CTSV and mouse CTSL are processed by human LGMN supports the suggestion that mouse CTSL is actually the orthologue of human CTSV, since no orthologue of CTSV in mice has been discovered yet. In addition, human CTSV can compensate for mouse CTSL and is phylogenetically more related to mouse CTSL than to human CTSL42. Another possibility is that mouse CTSL controls the specific functional activities of both human CTSL and CTSV. In epidermal protein extracts of normal skin all forms of CTSV were clearly visible, but very little CTSV could be detected in the extracts of psoriatic skin (Fig. 6). Previous experiments have shown that CTSV expression is decreased in psoriatic skin, both on the mRNA and the protein level27. Decreased CTSV activity could lead to delayed desquamation and retention of the corneocytes and thereby induce thickening and visible scaling of psoriatic skin. This report shows for the first time detection of tcCTSV protein in human tissue. CTSV protein expression in tissues has only been studied in two previous studies so far, that performed Western blot analysis in extracts of SC and thymus. In SC multiple forms of CTSV of 33‐39 kDa were detected, that were assumed to be proCTSV, scCTSV and degradation products, but the two‐chain form of CTSV was not seen12. However, the CTSV antibody used in that study was raised against the end‐part of the propeptide of CTSV, and the detected bands probably represent proCTSV and intermediate forms of scCTSV with extended N‐termini, whereas the fully matured single‐chain form and the two‐chain form

82 are not recognized by the antibody. In thymus only one form of ~30 kDa was detected, probably indicating scCTSV43. The antibody used for detection in thymus extracts is reported to interact specifically with CTSV, but production and characterization of this antibody has not been published. Another possibility is that tcCTSV is not produced in thymus since distribution of the mature forms may vary among tissues, which has been shown for CTSL40. Altogether, we have shown that CTSV degrades corneodesmosomal proteins and that CTSV is processed by LGMN into a previously unknown active two‐chain form that exists in human epidermis. Since the activities of both CTSV and LGMN can be inhibited by cystatin M/E, we propose a regulatory role for cystatin M/E in epidermal desquamation by controlling CTSV protease activity in the SC.

REFERENCES

1. Getsios, S., Huen, A. C., Green, K. J. (2004) Working out the strength and flexibility of desmosomes. Nat.Rev.Mol.Cell Biol. 5, 271‐281 2. Green, K. J., Simpson, C. L. (2007) Desmosomes: new perspectives on a classic. J.Invest Dermatol. 127, 2499‐2515 3. Garrod, D. R., Merritt, A. J., Nie, Z. (2002) Desmosomal cadherins. Curr.Opin.Cell Biol. 14, 537‐ 545 4. Kottke, M. D., Delva, E., Kowalczyk, A. P. (2006) The desmosome: cell science lessons from human diseases. J.Cell Sci. 119, 797‐806 5. Elias, P. M., Matsuyoshi, N., Wu, H., Lin, C., Wang, Z. H., Brown, B. E., Stanley, J. R. (2001) Desmoglein isoform distribution affects stratum corneum structure and function. J.Cell Biol. 153, 243‐249 6. Simon, M., Jonca, N., Guerrin, M., Haftek, M., Bernard, D., Caubet, C., Egelrud, T., Schmidt, R., Serre, G. (2001) Refined characterization of corneodesmosin proteolysis during terminal differentiation of human epidermis and its relationship to desquamation. J.Biol.Chem. 276, 20292‐20299 7. Jonca, N., Guerrin, M., Hadjiolova, K., Caubet, C., Gallinaro, H., Simon, M., Serre, G. (2002) Corneodesmosin, a component of epidermal corneocyte desmosomes, displays homophilic adhesive properties. J.Biol.Chem. 277, 5024‐5029 8. Lai‐Cheong, J. E., Arita, K., McGrath, J. A. (2007) Genetic diseases of junctions. J.Invest Dermatol. 127, 2713‐2725 9. Matsumoto, M., Zhou, Y., Matsuo, S., Nakanishi, H., Hirose, K., Oura, H., Arase, S., Ishida‐ Yamamoto, A., Bando, Y., Izumi, K., Kiyonari, H., Oshima, N., Nakayama, R., Matsushima, A., Hirota, F., Mouri, Y., Kuroda, N., Sano, S., Chaplin, D. D. (2008) Targeted deletion of the murine corneodesmosin gene delineates its essential role in skin and hair physiology. Proc.Natl.Acad.Sci.U.S.A 105, 6720‐6724 10. Elias, P. M. (2005) Stratum corneum defensive functions: an integrated view. J.Invest Dermatol. 125, 183‐200 11. Igarashi, S., Takizawa, T., Takizawa, T., Yasuda, Y., Uchiwa, H., Hayashi, S., Brysk, H., Robinson, J. M., Yamamoto, K., Brysk, M. M., Horikoshi, T. (2004) Cathepsin D, but not cathepsin E, degrades desmosomes during epidermal desquamation. Br.J.Dermatol. 151, 355‐361

83 12. Bernard, D., Mehul, B., Thomas‐Collignon, A., Simonetti, L., Remy, V., Bernard, M. A., Schmidt, R. (2003) Analysis of proteins with caseinolytic activity in a human stratum corneum extract revealed a yet unidentified cysteine protease and identified the so‐called "stratum corneum thiol protease" as cathepsin L2. J.Invest Dermatol. 120, 592‐600 13. Caubet, C., Jonca, N., Brattsand, M., Guerrin, M., Bernard, D., Schmidt, R., Egelrud, T., Simon, M., Serre, G. (2004) Degradation of corneodesmosome proteins by two serine proteases of the kallikrein family, SCTE/KLK5/hK5 and SCCE/KLK7/hK7. J.Invest Dermatol. 122, 1235‐1244 14. Borgono, C. A., Michael, I. P., Komatsu, N., Jayakumar, A., Kapadia, R., Clayman, G. L., Sotiropoulou, G., Diamandis, E. P. (2007) A potential role for multiple tissue kallikrein serine proteases in epidermal desquamation. J.Biol.Chem. 282, 3640‐3652 15. Chavanas, S., Bodemer, C., Rochat, A., Hamel‐Teillac, D., Ali, M., Irvine, A. D., Bonafe, J. L., Wilkinson, J., Taieb, A., Barrandon, Y., Harper, J. I., de Prost, Y., Hovnanian, A. (2000) Mutations in SPINK5, encoding a serine protease inhibitor, cause Netherton syndrome. Nat.Genet. 25, 141‐142 16. Descargues, P., Deraison, C., Prost, C., Fraitag, S., Mazereeuw‐Hautier, J., D'Alessio, M., Ishida‐ Yamamoto, A., Bodemer, C., Zambruno, G., Hovnanian, A. (2006) Corneodesmosomal cadherins are preferential targets of stratum corneum trypsin‐ and chymotrypsin‐like hyperactivity in Netherton syndrome. J.Invest Dermatol. 126, 1622‐1632 17. Zeeuwen, P. L., van Vlijmen‐Willems, I. M., Jansen, B. J., Sotiropoulou, G., Curfs, J. H., Meis, J. F., Janssen, J. J., van Ruissen, F., Schalkwijk, J. (2001) Cystatin M/E expression is restricted to differentiated epidermal keratinocytes and sweat glands: a new skin‐specific proteinase inhibitor that is a target for cross‐linking by transglutaminase. J.Invest Dermatol. 116, 693‐701 18. Zeeuwen, P. L., van Vlijmen‐Willems, I. M., Olthuis, D., Johansen, H. T., Hitomi, K., Hara‐ Nishimura, I., Powers, J. C., James, K. E., op den Camp, H. J., Lemmens, R., Schalkwijk, J. (2004) Evidence that unrestricted legumain activity is involved in disturbed epidermal cornification in cystatin M/E deficient mice. Hum.Mol.Genet. 13, 1069‐1079 19. Zeeuwen, P. L., van Vlijmen‐Willems, I. M., Hendriks, W., Merkx, G. F., Schalkwijk, J. (2002) A null mutation in the cystatin M/E gene of ichq mice causes juvenile lethality and defects in epidermal cornification. Hum.Mol.Genet. 11, 2867‐2875 20. Sundberg, J. P., Boggess, D., Hogan, M. E., Sundberg, B. A., Rourk, M. H., Harris, B., Johnson, K., Dunstan, R. W., Davisson, M. T. (1997) Harlequin ichthyosis (ichq): a juvenile lethal mouse mutation with ichthyosiform dermatitis. Am.J.Pathol. 151, 293‐310 21. Cheng, T., Hitomi, K., van Vlijmen‐Willems, I. M., de Jongh, G. J., Yamamoto, K., Nishi, K., Watts, C., Reinheckel, T., Schalkwijk, J., Zeeuwen, P. L. (2006) Cystatin M/E is a high affinity inhibitor of cathepsin V and cathepsin L by a reactive site that is distinct from the legumain‐binding site. A novel clue for the role of cystatin M/E in epidermal cornification. J.Biol.Chem. 281, 15893‐ 15899 22. Zeeuwen, P. L., Ishida‐Yamamoto, A., van Vlijmen‐Willems, I. M., Cheng, T., Bergers, M., Iizuka, H., Schalkwijk, J. (2007) Colocalization of cystatin M/E and cathepsin V in lamellar granules and corneodesmosomes suggests a functional role in epidermal differentiation. J.Invest Dermatol. 127, 120‐128 23. Shirahama‐Noda, K., Yamamoto, A., Sugihara, K., Hashimoto, N., Asano, M., Nishimura, M., Hara‐Nishimura, I. (2003) Biosynthetic processing of cathepsins and lysosomal degradation are abolished in asparaginyl endopeptidase‐deficient mice. J.Biol.Chem. 278, 33194‐33199

84 24. Maehr, R., Hang, H. C., Mintern, J. D., Kim, Y. M., Cuvillier, A., Nishimura, M., Yamada, K., Shirahama‐Noda, K., Hara‐Nishimura, I., Ploegh, H. L. (2005) Asparagine endopeptidase is not essential for class II MHC antigen presentation but is required for processing of cathepsin L in mice. J.Immunol. 174, 7066‐7074 25. Li, D. N., Matthews, S. P., Antoniou, A. N., Mazzeo, D., Watts, C. (2003) Multistep autoactivation of asparaginyl endopeptidase in vitro and in vivo. J.Biol.Chem. 278, 38980‐ 38990 26. Simon, M., Montezin, M., Guerrin, M., Durieux, J. J., Serre, G. (1997) Characterization and purification of human corneodesmosin, an epidermal basic glycoprotein associated with corneocyte‐specific modified desmosomes. J.Biol.Chem. 272, 31770‐31776 27. Cheng, T., Tjabringa, G. S., van Vlijmen‐Willems, I. M., Hitomi, K., van Erp, P. E., Schalkwijk, J., Zeeuwen, P. L. (2009) The cystatin M/E controlled pathway of skin barrier formation: expression of its key components in psoriasis and atopic dermatitis. Br.J.Dermatol. In press 28. Deraison, C., Bonnart, C., Lopez, F., Besson, C., Robinson, R., Jayakumar, A., Wagberg, F., Brattsand, M., Hachem, J. P., Leonardsson, G., Hovnanian, A. (2007) LEKTI fragments specifically inhibit KLK5, KLK7, and KLK14 and control desquamation through a pH‐dependent interaction. Mol.Biol.Cell 18, 3607‐3619 29. Rawlings, N. D., Barrett, A. J. (2004) Introduction: the clans and families of cysteine peptidases. In Handbook of Proteolytic Enzymes, Second Edition (Barrett, A. J., Rawlings, N. D., and Woessner, J. F., eds) pp. 1051‐1071, Elsevier Academic Press, London. 30. Brix, K., Dunkhorst, A., Mayer, K., Jordans, S. (2008) Cysteine cathepsins: cellular roadmap to different functions. Biochimie 90, 194‐207 31. Adachi, W., Kawamoto, S., Ohno, I., Nishida, K., Kinoshita, S., Matsubara, K., Okubo, K. (1998) Isolation and characterization of human cathepsin V: a major proteinase in corneal epithelium. Invest Ophthalmol.Vis.Sci. 39, 1789‐1796 32. Brömme, D., Li, Z., Barnes, M., Mehler, E. (1999) Human cathepsin V functional expression, tissue distribution, electrostatic surface potential, enzymatic characterization, and chromosomal localization. Biochemistry 38, 2377‐2385 33. Santamaria, I., Velasco, G., Cazorla, M., Fueyo, A., Campo, E., Lopez‐Otin, C. (1998) Cathepsin L2, a novel human cysteine proteinase produced by breast and colorectal carcinomas. Cancer Res. 58, 1624‐1630 34. Chen, J. M., Dando, P. M., Rawlings, N. D., Brown, M. A., Young, N. E., Stevens, R. A., Hewitt, E., Watts, C., Barrett, A. J. (1997) Cloning, isolation, and characterization of mammalian legumain, an asparaginyl endopeptidase. J.Biol.Chem. 272, 8090‐8098 35. Menard, R., Carmona, E., Takebe, S., Dufour, E., Plouffe, C., Mason, P., Mort, J. S. (1998) Autocatalytic processing of recombinant human procathepsin L. Contribution of both intermolecular and unimolecular events in the processing of procathepsin L in vitro. J.Biol.Chem. 273, 4478‐4484 36. Nishimura, Y., Kawabata, T., Furuno, K., Kato, K. (1989) Evidence that aspartic proteinase is involved in the proteolytic processing event of procathepsin L in lysosomes. Arch.Biochem.Biophys. 271, 400‐406 37. Burden, R. E., Snoddy, P., Jefferies, C. A., Walker, B., Scott, C. J. (2007) Inhibition of cathepsin L‐ like proteases by cathepsin V propeptide. Biol.Chem. 388, 541‐545

85 38. Carmona, E., Dufour, E., Plouffe, C., Takebe, S., Mason, P., Mort, J. S., Menard, R. (1996) Potency and selectivity of the cathepsin L propeptide as an inhibitor of cysteine proteases. Biochemistry 35, 8149‐8157 39. Ishidoh, K., Kominami, E. (2002) Processing and activation of lysosomal proteinases. Biol.Chem. 383, 1827‐1831 40. Ishidoh, K., Saido, T. C., Kawashima, S., Hirose, M., Watanabe, S., Sato, N., Kominami, E. (1998) Multiple processing of procathepsin L to cathepsin L in vivo. Biochem.Biophys.Res.Commun. 252, 202‐207 41. Dando, P. M., Fortunato, M., Smith, L., Knight, C. G., McKendrick, J. E., Barrett, A. J. (1999) Pig kidney legumain: an asparaginyl endopeptidase with restricted specificity. Biochem.J. 339 (Pt 3), 743‐749 42. Hagemann, S., Gunther, T., Dennemarker, J., Lohmuller, T., Bromme, D., Schule, R., Peters, C., Reinheckel, T. (2004) The human cysteine protease cathepsin V can compensate for murine cathepsin L in mouse epidermis and hair follicles. Eur.J.Cell Biol. 83, 775‐780 43. Tolosa, E., Li, W., Yasuda, Y., Wienhold, W., Denzin, L. K., Lautwein, A., Driessen, C., Schnorrer, P., Weber, E., Stevanovic, S., Kurek, R., Melms, A., Bromme, D. (2003) Cathepsin V is involved in the degradation of invariant chain in human thymus and is overexpressed in myasthenia gravis. J.Clin.Invest 112, 517‐526 44. Ritonja, A., Popovic, T., Kotnik, M., Machleidt, W., Turk, V. (1988) Amino acid sequences of the human kidney cathepsins H and L. FEBS Lett. 228, 341‐345 45. Troen, B. R., Gal, S., Gottesman, M. M. (1987) Sequence and expression of the cDNA for MEP (major excreted protein), a transformation‐regulated secreted cathepsin. Biochem.J. 246, 731‐ 735

86 CHAPTER 5

Colocalization of cystatin M/E and its target proteases suggests a role in terminal differentiation of human hair follicle and nail

Tsing Cheng, Ivonne M.J.J. van Vlijmen‐Willems, Kiyotaka Hitomi, Marcel C. Pasch, Piet E.J. van Erp, Joost Schalkwijk, and Patrick L.J.M. Zeeuwen

This chapter has been published in: J.Invest.Dermatol. (2009) 129, 1232‐1242

87 ORIGINAL ARTICLE

Colocalization of Cystatin M/E and its Target Proteases Suggests a Role in Terminal Differentiation of Human Hair Follicle and Nail Tsing Cheng1, Ivonne M.J.J. van Vlijmen-Willems1, Kiyotaka Hitomi2, Marcel C. Pasch1, Piet E.J. van Erp1, Joost Schalkwijk1 and Patrick L.J.M. Zeeuwen1

The cysteine protease inhibitor cystatin M/E is a key regulator of a biochemical pathway that leads to epidermal terminal differentiation by inhibition of its target proteases cathepsin L, cathepsin V, and legumain. Inhibition of cathepsin L is important in the cornification process of the skin, as we have recently demonstrated that cathepsin L is the elusive processing and activating protease for transglutaminase 3, an enzyme that is responsible for crosslinking of structural proteins in cornified envelope formation. Here, we study the localization of all players of this pathway in the human hair follicle and nail unit in order to elucidate their possible role in the biology of these epidermal appendages. We found that cathepsin L and transglutaminase 3 specifically colocalize in the hair bulb and the nail matrix, the regions that provide cells that terminally differentiate to the hair fiber and the nail plate, respectively. Furthermore, transglutaminase 3 also colocalizes with the structural proteins loricrin and involucrin, which are established transglutaminase substrates. These findings suggest that cathepsin L and transglutaminase 3 could be involved in the pathway that leads to terminal differentiation, not only in the epidermis but also in the human hair follicle and nail unit. Journal of Investigative Dermatology advance online publication, 13 November 2008; doi:10.1038/jid.2008.353

INTRODUCTION and small proline-rich proteins (SPRs) in the cornification Recent work of our group has revealed a previously process of the skin (Cheng et al., 2006). During keratinocyte unreported biochemical pathway that controls skin barrier differentiation, TGase 3 is activated by limited proteolysis formation by regulation of both crosslinking and desquama- (Hitomi et al., 1999; Ahvazi et al., 2002), a process that is tion of the stratum corneum. The cysteine protease inhibitor apparently under control of cystatin M/E, at least during skin cystatin M/E, which is a key molecule in this pathway, has morphogenesis in the neonatal phase (Zeeuwen et al., 2004). two important regulatory functions in epidermal differentia- We suppose that processing and activation of TGase 3 in tion and hair follicle morphogenesis. First, cystatin M/E terminally differentiating keratinocytes is regulated by the regulates crosslinking of structural proteins by transglutami- inhibitory activity of cytoplasmic cystatin M/E against CTSL. nase 3 (TGase 3) in the cornification process of the epidermis Legumain inhibition by cystatin M/E could have a regulatory and the hair follicle by controlling cathepsin L (CTSL) and role in CTSL activity as legumain is involved in CTSL legumain activities (Zeeuwen et al., 2004; Cheng et al., processing (Maehr et al., 2005). As a second regulatory 2006). We provide evidence for this assumption when we function, cystatin M/E could have a function in desquamation demonstrated that human CTSL is the elusive enzyme that is by the regulation of cathepsin V (CTSV) protease activity, able to process and activate human TGase 3, an epidermis- which is involved in the degradation of corneodesmosomal specific enzyme responsible for the crosslinking of loricrin components. We recently showed that cystatin M/E and CTSV are separately transported within lamellar granules and that both proteins are secreted in the extracellular space of 1 Department of Dermatology, Nijmegen Centre for Molecular Life Sciences, the stratum corneum where they associate with corneodes- Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands and 2Department of Applied Molecular Biosciences, Graduate School of mosomes as determined at the ultrastructural level (Zeeuwen Bioagricultural Sciences, Nagoya University, Nagoya, Japan et al., 2007). Correspondence: Professor Joost Schalkwijk or Dr Patrick L.J.M. Zeeuwen, Misregulation of this pathway by unrestrained protease Department of Dermatology, Nijmegen Centre for Molecular Life Sciences, activity, as seen in cystatin M/E-deficient mice, leads to Radboud University Nijmegen Medical Centre, P.O. Box 9101, 6500 HB, abnormal stratum corneum and hair follicle formation, Nijmegen, The Netherlands. E-mails: [email protected] or [email protected] disturbance of skin barrier function, and neonatal death Abbreviations: CTSL, cathepsin L; CTSV, cathepsin V; IRS, inner root sheath; (Sundberg et al., 1997; Zeeuwen et al., 2002, 2004). The ORS, outer root sheath; TGase, transglutaminase; THH, trichohyalin importance of regulated proteolysis in epithelia is well Received 29 May 2008; revised 4 August 2008; accepted 11 September 2008 demonstrated by the recent identification of the SPINK5

& 2008 The Society for Investigative Dermatology www.jidonline.org 88 T Cheng et al. Cystatin M/E in Human Nail and Hair Follicle

serine protease inhibitor as the defective gene in the known. Nevertheless, it is known that lack of TGase 1 in Netherton syndrome, cathepsin C mutations in the Papillon- lamellar ichthyosis patients could be reflected in structural Lefevre syndrome, ichthyosis because of matriptase defi- nail alterations (Rice et al., 2003). ciency, and targeted ablation of CTSL and cathepsin D in In the present study we sought to address the possible role mice (for a comprehensive review see Zeeuwen, 2004). of cystatin M/E and its physiological target proteases in the The hair follicle and nail unit are epidermal appendages biology of the human nail unit and hair follicle. We have that also show processes of terminal differentiation. Hair examined the localization of cystatin M/E, CTSV, CTSL, follicles undergo cycles of growth, regression, and rest, which TGase 3, and legumain in (longitudinal and/or transversal) are, respectively, named anagen, catagen, and telogen phase sections of the human hair follicle and nail unit by (Paus and Foitzik, 2004; Alonso and Fuchs, 2006). During the immunofluoresence microscopy. We also examined the anagen phase the entire hair from tip to root is produced localization of known TGase substrates like the structural when matrix cells in the hair bulb proliferate and differentiate proteins involucrin, loricrin, and trichohyalin (THH). We into the several layers of the hair follicle (Figure 1): the found colocalization of CTSL, TGase 3, involucrin, and medulla, cortex, and cuticle layer of the hair shaft (the actual loricrin in the hair bulb and the nail matrix, suggesting that hair fiber); the Huxley’s and Henle’s layers, and the cuticle of these proteins are key players in the pathway that initiate the inner root sheath (IRS), and the outer root sheath (ORS) biochemical changes in the terminal differentiation of these (Niemann and Watt, 2002; Fuchs, 2007). Keratinization of epidermal appendages. hair follicles occurs in specific compartments, that is the hair shaft, cuticle, and IRS, through crosslinking of structural RESULTS precursor proteins by TGases (Commo and Bernard, 1997; Hair follicle Thibaut et al., 2005). The degeneration of the IRS in the Immunofluorescence microscopy was performed on both infundibulum of the hair follicle near the opening of the longitudinal and transversal sections of human hair follicle in sebaceous duct requires desquamation, a process in which pro- order to examine the localization of cystatin M/E and its teolytic enzymes are involved (Ekholm and Egelrud, 1998). physiological target proteases in the different layers of the Contrary to hairs, nails show continuous growth of the hair follicle. For a comprehensive biochemical and kinetic hard nail plate out of the nail matrix, sliding over the analysis of the interaction between cystatin M/E and its target underlying nail bed (Dawber et al., 2001; Baran et al., 2005). proteases (see Cheng et al., 2006). Double staining was Keratinization of the nail starts in the nail matrix, but the performed for cystatin M/E with its target proteases, for TGase relative contribution of TGases to the formation of nail 3 with its activating protease CTSL, and for TGase 3 with its corneocytes which end up in the nail plate is not exactly presumed substrates (Figures 2–4).

Epidermis Inner root sheath (IRS)

Outer root sheath (ORS) Henle’s layer Huxley’s layer

Sebaceous Cuticle glands

Outer root sheath Inner root sheath

Medulla Cortex Cuticle

Fibrous sheath Medulla Hair bulb Cortex

Longitudinal section Transversal section

Figure 1. Schematic overview of longitudinal and transversal sections demonstrating the various components of the hair follicle.

Journal of Investigative Dermatology 89 T Cheng et al. Cystatin M/E in Human Nail and Hair Follicle

abcCystatin M/E Cathepsin V Merge

defCystatin M/E Cathepsin V Merge

ghiCystatin M/E Cathepsin L Merge

jklCystatin M/E Cathepsin L Merge

mnCystatin M/E Legumain oMerge

p Cystatin M/E q Legumain r Merge

Figure 2. Colocalization of cystatin M/E and its target proteases CTSV, CTSL, and legumain in human hair follicle. Immunofluorescence staining of (a, d, g, j, m, and p) cystatin M/E (green color), and (b, e, h, k, n, and q) CTSV, CTSL, or legumain (red color), and (c, f, i, l, o, and r) double staining (yellow–orange merge color) for both proteins on (c, i, and o) longitudinal and (f, l, and r) transversal sections of the hair follicle. Nuclei are blue (DNA staining by 40, 6-diamine-20-phenylindole dihydrochloride (DAPI)). Bar ¼ 100 mm.

Cystatin M/E is expressed in the IRS, the inner half of the infundibulum of the hair follicle, but is absent in the proximal ORS, and in the sebaceous glands of the hair follicle. part, the hair bulb (Figures 2a, g, m, and 3g). Furthermore, However, cystatin M/E is only expressed in the distal part and staining of the IRS is only strong in the Henle’s layer;

www.jidonline.org 90 T Cheng et al. Cystatin M/E in Human Nail and Hair Follicle

abcTGase 3 Cathepsin L Merge

defTGase 3 Cathepsin L Merge

ghiCystatin M/E Legumain Cathepsin V

Figure 3. Colocalization of TGase 3 and CTSL in human epidermis and hair follicle. Immunofluorescence staining of (a and d) TGase 3 (green color), and (b and e) CTSL (red color), and (c and f) double staining (yellow–orange merge color) for both proteins on epidermis (a–c), and both longitudinal (d–f) and transversal (insets d–f) sections of the hair follicle. The arrows (f) show the position of the transversal section (inset). Immunofluorescence staining of legumain was found in the (h) hair bulb, whereas no positive staining could be detected for (g) cystatin M/E or (i) CTSV in this part of the follicle. Nuclei are blue by DAPI staining for DNA. Bar ¼ 100 mm.

expression of cystatin M/E is weak or absent in the cuticle and granulosum (Figure 3c). Note that the yellow merge staining Huxley’s layer (Figure 2d, j, and p). Colocalization of cystatin forms a transition zone between the red CTSL staining in the M/E and CTSV can be found specifically in the Henle’s layer spinous and granular layers and the green TGase 3 staining in of the IRS, as CTSV expression is restricted to this layer of the upper granular layers and the stratum corneum, which cells of the hair follicle (Figure 2e–f). Interestingly, colocali- suggests that first the activating protease is expressed zation of cystatin M/E and CTSV is only seen in the followed by its substrate. Double staining of TGase 3 and infundibulum of the hair follicle near the opening of the CTSL in the hair follicle shows a similar overlapping sebaceous ducts into the hair canal (Figure 2c). CTSL shows a colocalization in the hair bulb (Figure 3f). First, CTSL is less restricted expression pattern and is expressed in IRS (but expressed at the bottom of the hair bulb followed by the not in the Huxley’s layer), ORS, and sebaceous glands (Figure expression of TGase 3 in the upper part of the hair bulb. 2h and k). Coexpression of cystatin M/E and CTSL can be TGase 3 expression is restricted to the medulla and cortex of seen in the distal part and infundibulum of the hair follicle the hair shaft, whereas CTSL is also expressed in the root and is located in the Henle’s layer of the IRS, the inner half of sheaths (see insets Figure 3d–f). Subsequently, CTSL expres- the ORS, and the sebaceous glands (Figure 2i and l). CTSL is sion disappears from the hair shaft and only TGase 3 also expressed in the medulla and cortex of the hair shaft, but expression remains (Figure 3f). However, staining of TGase only in the lower hair bulb (inset Figure 3e). Legumain 3 in the hair shaft was only seen in the proximal part of the expression can be found throughout the entire hair follicle hair follicle; no expression could be detected in the distal part resulting in a colocalization with cystatin M/E in the IRS, the and infundibulum of the hair follicle (data not shown). inner layers of the ORS, and the sebaceous glands (Figure 2o Remarkably, cystatin M/E is not present in the hair bulb, as is and r), but again only in the distal part and infundibulum of CTSV, whereas legumain staining was positive for this section the hair follicle. (Figure 3g–i). During epidermal cornification TGase 3 is TGases are enzymes involved in protein crosslinking of responsible for crosslinking of loricrin and small proline-rich structural proteins in cornified envelope formation. In human proteins into the cornified envelope. Loricrin is abundantly epidermis TGase 3 is expressed in the upper layers of the expressed in the hair shaft; a weak signal can be detected in stratum granulosum and the stratum corneum, where the the Henle’s layer of the IRS, but not in Huxley’s layer (Figure keratinocytes undergo cornification (Figure 3a). Previously, 4b). This expression pattern results in a colocalization with we have demonstrated that CTSL is the elusive enzyme that TGase 3 in the medulla, cortex, and cuticle of the hair shaft processes and activates TGase 3. Here, we show that in (Figure 4c). We also found colocalization of TGase 3 with human epidermis TGase 3 and CTSL colocalize in the stratum involucrin, another structural protein of cornified envelopes,

Journal of Investigative Dermatology 91 T Cheng et al. Cystatin M/E in Human Nail and Hair Follicle

abTGase 3 Ioricrin cMerge

deTGase 3 Involucrin fMerge

ghTGase 3 Trichohyalin iMerge

jklTGase 3 Trichohyalin Merge

mnTGase 3 TGase 1 oMerge

Figure 4. TGase 3 and its presumed substrates in human hair follicle. Immunofluorescence staining of (a, d, g, j, and m) TGase 3 (green color), and (b, e, h, k, and n) loricrin, involucrin, THH, or TGase 1 (red color), and (c, f, i, l, and o) double staining (yellow–orange merge color) for these proteins on transversal (a–i) and longitudinal (j–o) sections of the hair follicle. Nuclei are blue by DAPI staining for DNA. Bar ¼ 100 mm. as involucrin is expressed in the hair shaft and the IRS (Figure Nail unit 4e). However, colocalization of these proteins is only Five specific regions of the nail unit were studied in close-up, detected in the medulla and cortex, but not the cuticle, namely the epidermis (at the proximal side of the nail), the as involucrin seems to be absent in this layer as indicated proximal nail fold, the nail matrix, the nail bed, and the by green staining surrounding the medulla and cortex (Figure hyponychium (Figure 5). We performed immunofluorescence 4f). Immunofluorescence detection of THH, the major microscopy for double staining of cystatin M/E and its target structural component of hair follicles and a presumed protease CTSL, and for TGase 3 with its activating protease substrate for TGase 3, shows positive staining of granules in CTSL (Figures 6 and 7). the IRS (Figure 4h and k). However, TGase 3 and THH Cystatin M/E is expressed in the upper granular layers of show no colocalization in this part of the hair follicle epidermis and proximal nail fold (Figure 6a and d), but (Figure 4i and l). Finally, we studied the distribution of TGase expression diminishes towards the nail matrix where staining 1 in the hair follicle as TGase 3 expression is restricted to of cystatin M/E is faint and located in the cell layers bordering the hair shaft. Interestingly, TGase 1 is specifically expressed the nail plate (Figure 6g). Cystatin M/E is not present in the in the IRS at the same location where THH is expressed nail bed, but reappears at the hyponychium (Figure 6j (Figure 4k and n). Double staining of TGase 1 and TGase 3 and m). Positive staining of CTSL can be seen throughout shows no overlap between both enzymes in the hair follicle the entire nail unit and is suprabasal with the exception (Figure 4o). of the stratum corneum of the epidermis and hyponychium

www.jidonline.org 92 T Cheng et al. Cystatin M/E in Human Nail and Hair Follicle

Epidermis (Figure 6b, e, h, k, and n), resulting in a colocalization with cystatin M/E in all regions of the nail unit that express cystatin

Proximal M/E (Figure 6c, f, i, l, and o). The double staining of cystatin nail fold M/E and legumain shows a similar picture as legumain is also Matrix expressed in the entire nail unit (data not shown). No colocalization of cystatin M/E with CTSV can be found as Nail bed CTSV seems to be absent in the nail unit, except for the Hyponychium epidermis and the hyponychium (data not shown). TGase 3 is Bone expressed in all regions of the nail unit except the nail bed, which is completely negative for TGase 3 (Figure 7a, d, g, j, and m). The expression of TGase 3 in the epidermis, proximal nail fold, and hyponychium is suprabasal, but most intense in the upper granular layers, although the distal part of the Figure 5. Cross-section through the finger demonstrating the various proximal nail fold shows a weaker expression of TGase 3. components of the nail unit. Different parts of the nail unit (epidermis, However, TGase 3 staining is very strong in the nail matrix, proximal nail fold, nail matrix, nail bed, and hyponychium) that were but restricted to only the upper cell layers bordering the nail examined for immunohistochemistry in this study are marked in a plate (Figure 7g). Colocalization of TGase 3 and CTSL is seen colored box. in the granular layers of the nail unit, that is epidermis,

abcCystatin M/E Cathepsin L Merge

Epidermis

de f

Proximal nail fold

gh i

Matrix

jkl

Nail bed

mn o

Hyponychium

Figure 6. Colocalization of cystatin M/E and CTSL in the human nail unit. Immunofluorescence staining of (a, d, g, j, and m) cystatin M/E (green color), and (b, e, h, k, and n) CTSL (red color), and (c, f, i, l, and o) double staining (yellow–orange merge color) for both proteins in epidermis (a–c), proximal nail fold (d–f), nail matrix (g–i), nail bed (j–l), and hyponychium (m–o). Nuclei are blue by DAPI staining for DNA. Bar ¼ 100 mm.

Journal of Investigative Dermatology 93 T Cheng et al. Cystatin M/E in Human Nail and Hair Follicle

abcTGase 3 Cathepsin L Merge

Epidermis

def

Proximal nail fold

ghi

Matrix

jkl

Nail bed

mno

Hyponychium

Figure 7. Colocalization of TGase 3 and CTSL in the human nail unit. Immunofluorescence staining of (a, d, g, j, and m) TGase 3 (green color), and (b, e, h, k, and n) CTSL (red color), and (c, f, i, l, o) double staining (yellow–orange merge color) for both proteins in epidermis (a–c), proximal nail fold (d–f), nail matrix (g–i), nail bed (j–l), and hyponychium (m–o). Nuclei are blue by DAPI staining for DNA. Bar ¼ 100 mm. proximal nail fold, and hyponychium (Figure 7c, f, and o), but follicle. CTSL expression appears in the lower (suprabasal) is most prominent in the nail matrix (Figure 7i). layers of the epidermis where terminal differentiation of the cells starts, followed by the expression of TGase 3 in the DISCUSSION upper granular layer where processing and activation of the Cystatin M/E is a key regulator of a previously unreported crosslinking enzyme takes place (Figure 3). Likewise, CTSL is biochemical pathway that leads to epidermal terminal expressed in the bottom of the hair bulb, which contains differentiation through its inhibitory activities against its matrix cells that produce progeny cells that terminally target proteases CSTL, CTSV, and legumain. Inhibition of differentiate to form the growing hair shaft. Just above this CTSL is important in the cornification process as CTSL area, TGase 3 appears in the hair bulb at a location where processes and activates TGase 3. In this study we analyzed hair shaft precursors are normally expressed (Rogers, 2004). the localization of the key molecules of this pathway in the The colocalization of CTSL and TGase 3 in the hair bulb fits human hair follicle and nail unit. CTSL and TGase 3 well with our previous discovery that CTSL is the elusive specifically colocalize in the hair bulb and the nail matrix, protease that processes and activates TGase 3 (Cheng et al., the regions that provide cells that terminally differentiate into 2006). The nail matrix does also show a gradual transition the hair fiber and the nail plate, respectively. Moreover, the from CTSL to TGase 3 expression (Figure 7i). Terminally structural proteins loricrin and involucrin are also expressed differentiated cells comprising the nail bed originate from the at these sites suggesting an involvement in the pathway that nail matrix, which suggests that activation of TGase 3 takes leads to terminal differentiation, not only in the epidermis but place at a location that is analogous to the human hair bulb, also in the human hair follicle and nail unit. the location where the matrix cells of the hair reside. Our results show that CTSL and TGase 3 are coordinately Remarkably, cystatin M/E is absent in the hair bulb and expressed in the epidermis and the developing human hair almost not detectable in the nail matrix, which could mean

www.jidonline.org 94 T Cheng et al. Cystatin M/E in Human Nail and Hair Follicle

that full, uninhibited activity of CTSL is necessary for medulla (Rothnagel and Rogers, 1986; Hamilton et al., 1991; activating sufficient amounts of TGase 3. It is also possible Tarcsa et al., 1997; Steinert et al., 2003). However, none of that another cysteine protease inhibitor regulates CTSL these studies did use human material, and it is clear that THH activity in these regions. Cystatin A is able to inhibit CTSL expression and also the expression of TGase 3 is different and has been detected in hair shaft and nail extracts between mice and men. Furthermore, in these studies a (Tsushima et al., 1992; Tsushima, 1993). Expression of polyclonal antibody was used to detect TGase 3, which could cystatin M/E starts above the hair bulb and is further possibly crossreact with other TGases resulting in the positive continuously present in the distal part and the infundibulum, staining of the IRS. As TGase 3 does not show staining in the up to the epidermis where the hair exits the skin. One might IRS in our study, other TGases could be responsible for THH speculate that in these parts of the hair follicle CTSL has other crosslinking at this location. We found that TGase 1 is functions than processing TGase 3 (which is not expressed at specifically expressed in the IRS (Figure 4n) suggesting that these sites), and that at these locations CTSL activity has to be not TGase 3 but TGase 1 is the crosslinking enzyme of THH regulated by cystatin M/E. Nonetheless, the colocalization of in human hair follicle. This colocalization of THH and TGase CTSL and TGase 3 supports the proposed role of CTSL in the 1 in the IRS of the human hair follicle was previously also processing and activation of TGase 3 and the ensuing found by another group (Commo and Bernard, 1997). In our keratinization of the hair shaft and nail matrix. study THH could not be detected in the nail unit, whereas a In the epidermis TGases are responsible for crosslinking of previous study showed staining of the nail matrix for THH. loricrin, involucrin, and other structural barrier proteins into Interestingly, the matrix cells showed an apparent absence of the cornified envelope of keratinocytes (Candi et al., 2005). THH granules and only cytoplasmic THH was detected To our knowledge we showed for the first time colocalization (O’Keefe et al., 1993). This could explain our results as the of TGase 3 and loricrin in the hair follicle suggesting that used THH antibody of our choice, AE15, only stains THH in TGase 3 is involved in cornified envelope formation of this granules and not when THH is associated with keratin epidermal appendage. We also found expression of loricrin intermediate filament (O’Guin et al., 1992; Takahashi et al., and involucrin in the nail matrix at the same location were 2007). TGase 3 is expressed (not shown). The colocalization of We showed colocalization of cystatin M/E and CTSV in the TGase 3 and involucrin in the hair follicle was surprising as a IRS of the hair follicle, especially in Henle’s layer (Figure 2a–f). previous study, also based on immunofluorescence micro- This colocalization is restricted to the infundibulum of the hair scopy, reported that these proteins did not colocalize in the follicle around the opening of the sebaceous duct into the hair hair follicle. It was found that involucrin was abundantly canal. At this point the IRS is degenerated allowing the expressed only in the IRS and the internal layer of the ORS of secretion of sebum into the hair canal and movement of the the hair follicle, whereas TGase 3 expression was restricted to hair fiber (Stenn and Paus, 2001; Alonso and Fuchs, 2003). The the hair shaft and the hair cuticle (Thibaut et al., 2005). We expression pattern of cystatin M/E and CTSV suggests that both could confirm the involucrin expression in the IRS, but in proteins are involved in the degeneration of the IRS, possibly contrast to the study of Thibaut et al., we also found high through desquamation of the IRS cells. Epidermal desquama- expression of involucrin in the medulla and the cortex of the tion requires the proteolysis of corneodesmosomes, hair shaft at the level of the hair bulb. We suppose that the specialized junctions that keep epidermal cells together. In transversal sections of the hair follicle in our study (Figure 4) normal human skin the kallikreins KLK5 and KLK7 (two originate from a more proximal part of the hair follicle, epidermis-specific serine proteases) are able to degrade whereas the transversal sections in the study of Thibaut et al. corneodesmosomal proteins desmoglein-1, desmocollin-1, were taken at a higher level of the follicle. Nevertheless, as and corneodesmosin (Caubet et al., 2004) and it is presumed TGase 3 is not expressed in the IRS other TGases such as that cysteine proteases such as CTSV could have a similar TGase 1 and TGase 5, which are both expressed in the IRS function (Elias, 2005; Zeeuwen et al., 2007). Desquamation in and the internal layer of the ORS, could be responsible for the hair follicle could result from proteolytic activity of both involucrin crosslinking in the more distal parts of the hair serine and cysteine proteases. Desmoglein-1 and corneodes- follicle. mosin are expressed in the IRS as well as KLK7 (Ekholm and Surprisingly, our results showed no colocalization of Egelrud, 1998; Nachat et al., 2005), which both show an TGase 3 and THH in the hair follicle. THH could be a identical expression pattern as observed for CTSV in this study. substrate of TGase 3 in the hair shaft as a biochemical in vitro Altogether these findings provide a plausible explanation for study has revealed that TGase 3 is the most efficient the degeneration of the IRS at the opening of the sebaceous crosslinking enzyme of THH and also showed colocalization duct. Furthermore, we found no expression of CTSV in the of TGase 3 and THH in mouse skin (Tarcsa et al., 1997). THH ventral part of the proximal nail fold (bordering the nail plate), is a major structural hair follicle protein that is crosslinked to the nail matrix, and the nail bed. As desquamation does not cornified envelope proteins and keratin intermediate fila- occur in the regions bordering the nail plate, CTSV is not ments in cells of the hair shaft and the IRS providing expected to be expressed at this location. mechanical strength to the cells (Lee et al., 1993; Steinert It is likely that regulation of protease activity by cystatin et al., 2003). In our study THH is specifically expressed in the M/E is important in terminal differentiation of the hair Huxley’s layer of the IRS, whereas previous studies generally follicle and nail unit, and that disturbance could lead to reported localization of THH in the IRS and also in the abnormal development of these skin appendages. Cystatin

Journal of Investigative Dermatology 95 T Cheng et al. Cystatin M/E in Human Nail and Hair Follicle

M/E-deficient mice show a severe skin and hair phenotype Nail specimens (n ¼ 2) were obtained from toes of patients characterized by defects in cornification processes, although after distal limb amputations. All samples were collected with the nails on digits of both front and rear paws were normal written, informed consent using protocols approved by the local (Sundberg et al., 1997). It must be noted however, that these ethics committee, Commissie Mensgebonden Onderzoek Arnhem- mice die shortly after birth, and nail abnormalities cannot be Nijmegen, that complies with the Declaration of Helsinki principles. studied beyond day 10. CTSL-deficient mice show abnormal hair follicle cycling and morphogenesis including faulty Antibodies cornification and incomplete desquamation of the IRS (Roth Primary antibodies that were used in this study: affinity-purified et al., 2000; Benavides et al., 2002; Tobin et al., 2002). The polyclonal rabbit anti-human cystatin M/E antibody (Zeeuwen et al., latter observation is interesting as murine CTSL is likely to 2001), affinity-purified polyclonal sheep anti-human legumain harbour the specific enzymatic activities of both human CTSL antibody (Li et al., 2003), monoclonal mouse anti-human CTSV and CTSV as no murine ortholog of CTSV appears to exist antibody, monoclonal rat anti-human/mouse CTSL antibody (both (Bro¨mme et al., 1999; Hagemann et al., 2004). Moreover, from R&D Systems, Minneapolis, MN), monoclonal mouse anti- other proteases and inhibitors, as well as TGases, may also be human TGase 3 antibodies C2D and C9D (Hitomi et al., 2003), involved in terminal differentiation of the hair follicle and monoclonal mouse anti-human involucrin antibody MON150 nail unit. Mutations in the human cathepsin C gene have (Sanbio, Uden, The Netherlands), polyclonal rabbit anti-human/ been identified as the cause of Haim-Munk syndrome mouse loricrin antibody (BAbCO, Richmond, CA), and mouse anti- (MIM245010), a rare disorder resembling Papillon-Lefevre human THH antibody AE15 (Santa Cruz Biotechnology, Santa Cruz, syndrome but with the addition of nail abnormalities (Hart CA). For immunofluorescence analysis, the following secondary et al., 2000). Recently, a human autosomal recessive reagents were used: Alexa-Fluor 488 goat anti-rabbit IgG highly ichthyosis with hypotrichosis syndrome (MIM610765) was cross-absorbed, Alexa-Fluor 594 goat anti-rat, goat anti-mouse, and linked to a specific mutation in the ST14 gene, which donkey anti-sheep IgG highly cross-absorbed (Molecular Probes, encodes the serine protease matriptase (Basel-Vanagaite Eugene, OR). et al., 2007; List et al., 2007). Furthermore, the Netherton syndrome (MIM256500) features ichthyosis and hair shaft Immunofluorescence analysis defects caused by mutations in the SPINK5 gene resulting in a Tissues were rinsed in phosphate-buffered saline, fixed for 4 hours in reduced activity of the serine protease inhibitor lympho- buffered 4% formalin, and embedded in paraffin wax. All material epithelial Kazal type inhibitor (LEKTI) (Chavanas et al., 2000). was cut in 7-mm sections, mounted on SuperFrost slides (Menzel, Concerning TGases, it is known that lack of TGase 1 in Braunschweig, Germany), deparaffinized and rehydrated. The lamellar ichthyosis (MIM242300) patients could cause nail sections were subsequently incubated with primary antibodies and hair shaft abnormalities (Rice et al., 2003, 2005), diluted in 1% BSA/phosphate-buffered saline for 1 hour at room whereas patients with acral peeling skin syndrome temperature. After washing in phosphate-buffered saline, fluorescent (MIM609796), due to a missense mutation in the TGase 5 secondary antibodies were applied for 30 minutes at room tempera- gene, do not show nail nor hair defects (Cassidy et al., 2005). ture. For double labeling with antibodies raised in different animals, So far no patients are known that are deficient for TGase 3. a mixture of primary antibodies was applied and this was followed We conclude that the previously unreported biochemical by incubation with a mixture of secondary antibodies conjugated pathway regulated by cystatin M/E, could also be involved in with different fluorescent dyes. Nuclei were stained with 40,6- terminal differentiation of the hair follicle and nail unit. The diamine-20-phenylindole dihydrochloride (DakoCytomation, Copen- localization of all key molecules, from protease and protease hagen, Denmark), and the slides were mounted using Prolong Gold inhibitor to structural components, suggests a role for these Antifade reagent (Molecular Probes). The slides were subsequently proteins in both cornification and desquamation. Terminal examined using an immunofluorescence microscope (Axioskop 2 differentiation could start in the hair bulb and the nail matrix MOT; Zeiss, Sliedrecht, The Netherlands) and images were where CTSL is able to process and activate TGase 3, followed photographed using a digital camera (Sony DXC-390P 3 CCD; by crosslinking of loricrin, involucrin, and other structural Scanalytics, Fairfax, VA) and edited using Axiovision Software proteins. One should keep in mind that other TGases such as (Zeiss). TGase 1 and TGase 5 contribute to this process as well. In the hair follicle, CTSV could be involved in the proteolysis of Immunohistochemistry corneodesmosomal proteins, which is required for desqua- Immunohistochemical staining was performed using the Vectastain mation of the IRS cells. Altogether, this study provides more ABC kit (Vector Laboratories, Burlingame, CA) as previously insight in the molecular processes that underlie terminal described (Latijnhouwers et al., 1996). Sections stained for differentiation of the hair follicle and nail unit, and could lead involucrin received a pretreatment with citrate buffer and needed to further understanding of human skin disorders with antigen retrieval in the microwave oven. impaired barrier function and hair and nail abnormalities. CONFLICT OF INTEREST The authors state no conflict of interest. MATERIALS AND METHODS Tissue samples ACKNOWLEDGMENTS ¼ Human hair follicles (n 3) were obtained from surgically removed This work was supported by a grant from the Radboud University Nijmegen scalp tissue from patients with basal-cell carcinoma on the head. Medical Centre (to T.C.) and by VENI-grant 916.56.117 from ZONMW,

www.jidonline.org 96 T Cheng et al. Cystatin M/E in Human Nail and Hair Follicle

The Netherlands Organization for Health Research and Development (to Hart TC, Hart PS, Michalec MD, Zhang Y, Firatli E, Van Dyke TE et al. (2000) P.L.J.M.Z.). We thank Dr Colin Watts (Division of Cell Biology & Immunology, Haim-Munk syndrome and Papillon-Lefevre syndrome are allelic University of Dundee, UK) for providing polyclonal sheep anti-human mutations in cathepsin C. J Med Genet 37:88–94 legumain antibodies. We thank Professor Thomas Franz (Anatomisches Institut, Universita¨t Bonn, Germany) for critical reading of the paper. Hitomi K, Kanehiro S, Ikura K, Maki M (1999) Characterization of recombinant mouse epidermal-type transglutaminase (TGase 3): regula- tion of its activity by proteolysis and guanine nucleotides. J Biochem (Tokyo) 125:1048–54 REFERENCES Hitomi K, Presland RB, Nakayama T, Fleckman P, Dale BA, Maki M (2003) Ahvazi B, Kim HC, Kee SH, Nemes Z, Steinert PM (2002) Three-dimensional Analysis of epidermal-type transglutaminase (transglutaminase 3) in structure of the human transglutaminase 3 enzyme: binding of calcium human stratified epithelia and cultured keratinocytes using monoclonal ions changes structure for activation. EMBO J 21:2055–67 antibodies. J Dermatol Sci 32:95–103 Alonso L, Fuchs E (2003) Stem cells of the skin epithelium. Proc Natl Acad Sci Latijnhouwers MA, Bergers M, Van Bergen BH, Spruijt KI, Andriessen MP, USA 100(Suppl 1):11830–5 Schalkwijk J (1996) Tenascin expression during wound healing in human skin. J Pathol 178:30–5 Alonso L, Fuchs E (2006) The hair cycle. J Cell Sci 119:391–3 Lee SC, Kim IG, Marekov LN, O’Keefe EJ, Parry DA, Steinert PM (1993) The Baran R, Dawber RP, Haneke E (2005) Hair and nail relationship. Skinmed structure of human trichohyalin. Potential multiple roles as a functional 4:18–23 EF-hand-like calcium-binding protein, a cornified cell envelope pre- Basel-Vanagaite L, Attia R, Ishida-Yamamoto A, Rainshtein L, Ben AD, Lurie R cursor, and an intermediate filament-associated (cross-linking) protein. et al. (2007) Autosomal recessive ichthyosis with hypotrichosis caused J Biol Chem 268:12164–76 by a mutation in ST14, encoding type II transmembrane serine protease Li DN, Matthews SP, Antoniou AN, Mazzeo D, Watts C (2003) Multistep matriptase. Am J Hum Genet 80:467–77 autoactivation of asparaginyl endopeptidase in vitro and in vivo. J Biol Benavides F, Starost MF, Flores M, Gimenez-Conti IB, Guenet JL, Conti CJ Chem 278:38980–90 (2002) Impaired hair follicle morphogenesis and cycling with abnormal List K, Currie B, Scharschmidt TC, Szabo R, Shireman J, Molinolo A et al. epidermal differentiation in nackt mice, a cathepsin L-deficient muta- (2007) Autosomal ichthyosis with hypotrichosis syndrome displays low tion. Am J Pathol 161:693–703 matriptase proteolytic activity and is phenocopied in ST14 hypomorphic Bro¨mme D, Li Z, Barnes M, Mehler E (1999) Human cathepsin V functional mice. J Biol Chem 282:36714–23 expression, tissue distribution, electrostatic surface potential, enzymatic Maehr R, Hang HC, Mintern JD, Kim YM, Cuvillier A, Nishimura M et al. characterization, and chromosomal localization. Biochemistry 38:2377–85 (2005) Asparagine endopeptidase is not essential for class II MHC antigen presentation but is required for processing of cathepsin L in mice. Candi E, Schmidt R, Melino G (2005) The cornified envelope: a model of cell J Immunol 174:7066–74 death in the skin. Nat Rev Mol Cell Biol 6:328–40 Nachat R, Mechin MC, Charveron M, Serre G, Constans J, Simon M (2005) Cassidy AJ, van Steensel MA, Steijlen PM, van GM, van d V, Morley SM et al. Peptidylarginine deiminase isoforms are differentially expressed in the (2005) A homozygous missense mutation in TGM5 abolishes epidermal anagen hair follicles and other human skin appendages. J Invest transglutaminase 5 activity and causes acral peeling skin syndrome. Am J Dermatol 125:34–41 Hum Genet 77:909–17 Niemann C, Watt FM (2002) Designer skin: lineage commitment in postnatal Caubet C, Jonca N, Brattsand M, Guerrin M, Bernard D, Schmidt R et al. epidermis. Trends Cell Biol 12:185–92 (2004) Degradation of corneodesmosome proteins by two serine proteases of the kallikrein family, SCTE/KLK5/hK5 and SCCE/KLK7/hK7. O’Guin WM, Sun TT, Manabe M (1992) Interaction of trichohyalin with J Invest Dermatol 122:1235–44 intermediate filaments: three immunologically defined stages of tricho- hyalin maturation. J Invest Dermatol 98:24–32 Chavanas S, Bodemer C, Rochat A, Hamel-Teillac D, Ali M, Irvine AD et al. (2000) Mutations in SPINK5, encoding a serine protease inhibitor, cause O’Keefe EJ, Hamilton EH, Lee SC, Steinert P (1993) Trichohyalin: a structural Netherton syndrome. Nat Genet 25:141–2 protein of hair, tongue, nail, and epidermis. J Invest Dermatol 101:65S–71S Cheng T, Hitomi K, van Vlijmen-Willems IM, de Jongh GJ, Yamamoto K, Nishi K et al. (2006) Cystatin M/E is a high affinity inhibitor of cathepsin Paus R, Foitzik K (2004) In search of the ‘‘hair cycle clock’’: a guided tour. V and cathepsin L by a reactive site that is distinct from the legumain- Differentiation 72:489–511 binding site. A novel clue for the role of cystatin M/E in epidermal Rice RH, Crumrine D, Hohl D, Munro CS, Elias PM (2003) Cross-linked cornification. J Biol Chem 281:15893–9 envelopes in nail plate in lamellar ichthyosis. Br J Dermatol 149: Commo S, Bernard BA (1997) Immunohistochemical analysis of tissue 1050–4 remodelling during the anagen–catagen transition of the human hair Rice RH, Crumrine D, Uchida Y, Gruber R, Elias PM (2005) Structural follicle. Br J Dermatol 137:31–8 changes in epidermal scale and appendages as indicators of defective Dawber RPR, de Berker DAR, Baran R (2001) Science of the nail apparatus. TGM1 activity. Arch Dermatol Res 297:127–33 In: Baran and Dawber’s Diseases of the Nails and their Management Rogers GE (2004) Hair follicle differentiation and regulation. Int J Dev Biol (Baran R, Dawber RPR, Haneke E, Tosti A eds). Blackwell Science, Inc., 48:163–70 Malden, 1–47. Roth W, Deussing J, Botchkarev VA, Pauly-Evers M, Saftig P, Hafner A et al. Ekholm E, Egelrud T (1998) The expression of stratum corneum chymotryptic (2000) Cathepsin L deficiency as molecular defect of furless: hyperpro- enzyme in human anagen hair follicles: further evidence for its liferation of keratinocytes and pertubation of hair follicle cycling. FASEB involvement in desquamation-like processes. Br J Dermatol 139:585–90 J 14:2075–86 Elias PM (2005) Stratum corneum defensive functions: an integrated view. Rothnagel JA, Rogers GE (1986) Trichohyalin, an intermediate filament- J Invest Dermatol 125:183–200 associated protein of the hair follicle. J Cell Biol 102:1419–29 Fuchs E (2007) Scratching the surface of skin development. Nature Steinert PM, Parry DA, Marekov LN (2003) Trichohyalin mechanically 445:834–42 strengthens the hair follicle: multiple cross-bridging roles in the inner Hagemann S, Gunther T, Dennemarker J, Lohmuller T, Bromme D, Schule R root shealth. J Biol Chem 278:41409–19 et al. (2004) The human cysteine protease cathepsin V can compensate Stenn KS, Paus R (2001) Controls of hair follicle cycling. Physiol Rev for murine cathepsin L in mouse epidermis and hair follicles. Eur J Cell 81:449–94 Biol 83:775–80 Sundberg JP, Boggess D, Hogan ME, Sundberg BA, Rourk MH, Harris B Hamilton EH, Payne RE Jr, O’Keefe EJ (1991) Trichohyalin: presence in the et al. (1997) Harlequin ichthyosis (ichq): a juvenile lethal granular layer and stratum corneum of normal human epidermis. J Invest mouse mutation with ichthyosiform dermatitis. Am J Pathol 151: Dermatol 96:666–72 293–310

Journal of Investigative Dermatology 97 T Cheng et al. Cystatin M/E in Human Nail and Hair Follicle

Takahashi M, Horiuchi Y, Yoshida M, Isogai R, Kawada A, Tezuka T (2007) Zeeuwen PL (2004) Epidermal differentiation: the role of proteases and their Properties and histochemical application of a novel antibody against inhibitors. Eur J Cell Biol 83:761–73 trichohyalin granules. Arch Dermatol Res 299:33–9 Zeeuwen PL, Ishida-Yamamoto A, van Vlijmen-Willems IM, Cheng T, Bergers Tarcsa E, Marekov LN, Andreoli J, Idler WW, Candi E, Chung SI et al. (1997) M, Iizuka H et al. (2007) Colocalization of cystatin M/E and cathepsin V The fate of trichohyalin. Sequential post-translational modifications by in lamellar granules and corneodesmosomes suggests a functional role in peptidyl-arginine deiminase and transglutaminases. J Biol Chem epidermal differentiation. J Invest Dermatol 127:120–8 272:27893–901 Zeeuwen PL, van Vlijmen-Willems IM, Hendriks W, Merkx GF, Schalkwijk J Thibaut S, Candi E, Pietroni V, Melino G, Schmidt R, Bernard BA (2005) (2002) A null mutation in the cystatin M/E gene of ichq mice causes Transglutaminase 5 expression in human hair follicle. J Invest Dermatol juvenile lethality and defects in epidermal cornification. Hum Mol Genet 125:581–5 11:2867–75 Tobin DJ, Foitzik K, Reinheckel T, Mecklenburg L, Botchkarev VA, Peters C Zeeuwen PL, van Vlijmen-Willems IM, Jansen BJ, Sotiropoulou G, Curfs JH, et al. (2002) The lysosomal protease cathepsin L is an important regulator Meis JF et al. (2001) Cystatin M/E expression is restricted to differentiated of keratinocyte and melanocyte differentiation during hair follicle epidermal keratinocytes and sweat glands: a new skin-specific protei- morphogenesis and cycling. Am J Pathol 160:1807–21 nase inhibitor that is a target for cross-linking by transglutaminase. J Tsushima H (1993) Isolation of cysteine proteinase inhibitor, cystatin A, from Invest Dermatol 116:693–701 human nails. Arch Dermatol Res 285:418–22 Zeeuwen PL, van Vlijmen-Willems IM, Olthuis D, Johansen HT, Hitomi K, Tsushima H, Ueki A, Mine H, Nakajima N, Sumi H, Hopsu-Havu VK (1992) Hara-Nishimura I et al. (2004) Evidence that unrestricted legumain Purification and characterization of a cystatin-type cysteine proteinase activity is involved in disturbed epidermal cornification in cystatin M/E inhibitor in the human hair shaft. Arch Dermatol Res 284:380–5 deficient mice. Hum Mol Genet 13:1069–79

www.jidonline.org 98 CHAPTER 6

The cystatin M/E controlled pathway of skin barrier formation: expression of its key components in psoriasis and atopic dermatitis

Tsing Cheng, Geuranne S. Tjabringa, Ivonne M.J.J. van Vlijmen‐ Willems, Kiyotaka Hitomi, Piet E.J. van Erp, Joost Schalkwijk, and Patrick L.J.M. Zeeuwen

This chapter will be published in: Br.J.Dermatol. (2009) In press

99 The Cystatin M/E Controlled Pathway of Skin Barrier Formation: Expression of its Key Components in Psoriasis and Atopic Dermatitis

Tsing Cheng1, Geuranne S. Tjabringa1, Ivonne M.J.J. van Vlijmen‐Willems1, Kiyotaka Hitomi2, Piet E.J. van Erp1, Joost Schalkwijk1, and Patrick L.J.M. Zeeuwen1 1Department of Dermatology, Nijmegen Centre for Molecular Life Sciences, Radboud University Nijmegen Medical Centre, PO Box 9101, 6500 HB Nijmegen, The Netherlands, and 2Department of Applied Molecular Biosciences, Graduate School of Bioagricultural Sciences, Nagoya University, Nagoya 464‐8601, Japan

SUMMARY

Background The antiprotease activity of cystatin M/E regulates skin barrier formation, as it inhibits the activity of cathepsin V, cathepsin L and legumain, thereby controlling the processing of transglutaminase‐3. Misregulation of this pathway by unrestrained protease activity, as seen in cystatin M/E deficient mice, leads to abnormal stratum corneum and hair follicle formation, and severe disturbance of skin barrier function. Objectives Our major aim was to make a quantitative analysis of the expression of all players of this pathway in the epidermis of patients with inflammatory skin diseases. A second aim was to determine if reconstructed human skin could be used as an in vitro model system to investigate this pathway. Methods Autopsy material from normal human tissues, biopsies from normal skin of healthy volunteers, and lesional skin from atopic dermatitis and psoriatic patients were used to study the expression of the above mentioned molecules at the mRNA level by qPCR. Localization of the protein was performed by immunofluorescence microscopy, and expression was quantitated by image analysis. Results In skin, cystatin M/E is expressed at relatively higher levels than its target proteases, when compared to other tissues, which emphasizes its prominent role in cutaneous biology. We found decreased expression of cystatin M/E and cathepsin V in lesional atopic dermatitis and psoriasis epidermis at the mRNA level as well as the protein level. cathepsin L and transglutaminase‐3 are increased at the transcriptional level, however, this was not reflected by higher protein levels. Interestingly, the expression of all these molecules in reconstructed skin was qualitatively and quantitatively similar to the in vivo situation. Conclusions Disturbance of the cystatin M/E – cathepsin pathway could contribute to dysregulated skin barrier function observed in inflammatory dermatoses. Human reconstructed skin appears to be a valuable model to study this novel biochemical pathway in vitro.

100 INTRODUCTION

Cystatin M/E is a cysteine protease inhibitor encoded by the CST6 gene that is highly expressed in the upper layers of the human epidermis, hair follicles, sebaceous glands, and in the sweat glands1. Recently we unraveled a novel biochemical pathway that controls skin barrier formation by regulation of both cornification and desquamation of the stratum corneum (SC) (Fig. 1). Cystatin M/E, which targets two different classes of proteases, is a central molecule in this pathway as it controls the activity of cathepsin V (CTSV), cathepsin L (CTSL) and legumain (LGMN). Cystatin M/E regulates crosslinking of structural proteins by transglutaminase‐3 (TGM3) in the cornification process by controlling CTSL and LGMN activities, as human CTSL is the elusive enzyme that processes and activates human TGM32,3. During keratinocyte differentiation, TGM3 is activated by limited proteolysis4,5, a process that is apparently under control of cystatin M/E, at least during skin morphogenesis in the neonatal phase2. We suppose that processing and activation of TGM3 in terminally differentiating keratinocytes is regulated by the inhibitory activity of cystatin M/E against CTSL. LGMN inhibition by cystatin M/E could have a regulatory role in CTSL activity since LGMN is involved in CTSL processing6. Furthermore, cystatin M/E could have a role in

Figure 1: Regulation of epidermal protease activity by cystatin M/E. A simplified scheme of the presumed regulatory role of cystatin M/E in processes that control epidermal cornification and desquamation. Cystatin M/E is an inhibitor of the asparaginyl endopeptidase LGMN and the cysteine proteases CTSL and CTSV. Inhibition of LGMN might regulate the processing of (pro)cathepsins, a process that is also under control of intra‐ and extracellular pH changes. Inhibition of CTSV putatively regulates desquamation, as CTSV could degrade (corneo)‐desmosomal proteins. Inhibition of CTSL activity by cystatin M/E is thought to be important in the cornification process, as CTSL is the elusive processing and activating enzyme for TGM3.

101 desquamation by regulation of CTSV activity, which is involved in degradation of corneodesmosomal components. We showed that cystatin M/E and CTSV are separately transported within lamellar granules and that both proteins are secreted in the extracellular space of the SC where they associate with corneodesmosomes7. Recently, we have reported that the biochemical pathway controlled by cystatin M/E is also involved in terminal differentiation of the hair follicle and nail unit8. Misregulation of this pathway by unrestrained protease activity leads to abnormal SC and hair follicle formation, disturbance of skin barrier function and neonatal death as seen in cystatin M/E deficient ichq mice. These mice show a severe skin and hair phenotype and are characterized by defects in cornification processes2,9,10. CTSL deficient mice show abnormal hair follicle cycling and morphogenesis including faulty cornification and incomplete desquamation of the inner root sheath11‐13. Murine CTSL possibly controls the specific enzymatic activities of both human CTSL and CTSV since no murine orthologue of CTSV has been discovered14,15. Disturbed biosynthetic processing of cathepsins and accumulation of macromolecules in the lysosomes in the proximal tubule cells of the kidney was seen in LGMN deficient mice, however, no apparent skin or hair phenotype was observed16. These mouse models show that the balance between proteases and their cognate inhibitors is important for correct skin barrier formation and that imbalance may lead to disease17. Although no human skin disease has been linked to cystatin M/E deficiency yet, the importance of this balance is emphasized in view of other proteases and inhibitors that are involved in skin barrier formation. Mutations in the human cathepsin C gene have been identified as the cause of Papillon‐Lefevre syndrome (MIM 245000), which is characterized by hyperkeratosis on palms and soles and severe periodontitis18. Recently, the novel human autosomal recessive ichthyosis with hypotrichosis syndrome (ARIH, MIM 610765) was linked to a specific mutation in the ST14 gene, which encodes the serine protease matriptase19,20. Furthermore, the Netherton syndrome (MIM 256500) features ichthyosis and hair shaft defects caused by mutations in the SPINK5 gene resulting in a reduced activity of the serine protease inhibitor LEKTI21. Atopic dermatitis (AD) and psoriasis (PS) are two common chronic inflammatory skin diseases in which the expression of genes and the formation of the epidermal barrier is altered. Although both diseases are generally regarded as immune‐mediated conditions, recent genetic studies have indicated the importance of abnormalities in epithelium‐ expressed genes as a primary cause. Loss of function alleles of the skin barrier protein filaggrin were found to be a major predisposing factor for AD22, and we have recently demonstrated that a copy number polymorphism of a beta‐defensin gene cluster is associated with increased risk for PS23. In view of the fact that the cystatin M/E – cathepsin pathway is important for correct epidermal differentiation, we investigated the expression of all players of this biochemical pathway in major inflammatory skin diseases with a disturbed skin barrier function. In the present study we show the expression and localization of cystatin M/E, CTSV, CTSL, LGMN and TGM3 in AD and PS skin compared to normal skin (NS) from healthy volunteers.

102 Expression of some of these molecules was significantly different between diseased and normal skin, both on the mRNA and the protein level. Furthermore, we show that human reconstructed skin is a valuable tool for future research as it correctly recapitulates the expression pattern of all players in the pathway.

MATERIAL AND METHODS

Autopsic material and skin biopsies Patients and healthy volunteers were recruited by the Dermatology clinic of the Radboud University Nijmegen Medical Centre. All patients were diagnosed by a dermatologist. The study was approved by the local medical ethical committee (Commissie Mensgebonden Onderzoek Arnhem‐Nijmegen) and conducted according to the Declaration of Helsinki principles. Written informed consent was obtained from each volunteer. Patients who suffered from chronic AD, diagnosed according to the Hanifin criteria, were in an active phase of disease for which they were hospitalized and biopsies were taken from subacute lesions. Care was taken to exclude vesicles or scratched skin. All psoriasis patients had moderate to severe plaque type psoriasis. Biopsies of normal skin, psoriatic lesions and atopic dermatitis lesions were taken under local anesthesia with 4 mm punches. Archival autopsy material was obtained from the Department of Pathology, University of Nijmegen, The Netherlands.

Antibodies Primary antibodies that were used in this study: affinity‐purified polyclonal rabbit anti‐ human cystatin M/E antibody1, affinity‐purified polyclonal sheep anti‐human LGMN antibody24, monoclonal mouse anti‐human CTSV antibody, monoclonal rat anti‐ human/mouse CTSL antibody (both from R&D Systems, Minneapolis, MN), monoclonal mouse anti‐human TGM3 antibodies C2D and C9D25, monoclonal mouse anti‐human involucrin antibody MON150 (Sanbio, Uden, The Netherlands), polyclonal rabbit anti‐ human/mouse loricrin antibody (BAbCO, Richmond, CA), and rabbit‐anti‐human filaggrin antibody BT576 (Biomedical Technologies, Stoughton, MA). For immunofluorescence analysis, the following secondary reagents were used: Alexa‐Fluor 488 goat anti‐rabbit IgG highly cross‐absorbed, Alexa‐Fluor 594 goat anti‐rat, goat anti‐mouse and donkey anti‐sheep IgG highly cross‐absorbed (Molecular Probes, Eugene, OR).

Histology Immunohistochemical stainings were performed using the Vectastain ABC kit (Vector Laboratories, Burlingame, CA) as previously described26. Sections stained for involucrin received a pretreatment with citrate buffer and microwave for antigen retrieval. Immunofluorescence stainings were performed as described previously7.

103 Preparation of RNA from epidermal sheets In order to examine gene expression patterns exclusively in the epidermis we have previously developed a protocol that allows rapid separation of the epidermis from the underlying dermis under conditions that would minimize de novo expression or degradation of mRNA transcripts during the isolation procedure27. Total RNA for real‐time qPCR was purified from these epidermal sheets as previously described28.

Quantitative real‐time PCR First strand cDNA was synthesized using an input of 1 µg of RNA with the iScript cDNA synthesis kit (Bio‐Rad, Hercules, CA) according to the manufacturer’s recommendation. The reverse transcriptase reaction products were used for quantitative real‐time PCR, which was performed with the MyiQ Single‐Colour Real‐Time Detection System for quantification with Sybr Green and melting curve analysis (Bio‐Rad) as previously described29. Primers were designed using Primer Express 1.0 Software (Applied Biosystems) and produced by Biolegio (Nijmegen, The Netherlands). Primer validation, qPCR reactions, and determination of relative mRNA expression were performed as previously described28. Expression of target genes was normalized to that of human ribosomal phosphoprotein P0 (RPLP0). Relative quantity of gene expression was measured using a method described by Livak et al.30

ImageJ software The images were analyzed with public domain image processing software (ImageJ; http://rbs.info.nih.gov/) in a standardized manner. For expression of the epidermal proteins a region of interest (ROI) was drawn in a frame. This ROI contained the epidermis with the stratum corneum. In this ROI the intensity and percentage of positive signal in the epidermis was measured.

Table 1. List of genes (approved gene symbols, protein names) and primers used for quantitative real‐time PCR

HUGO Ensemble Description Gene Forward primer Reverse primer E* Transcript no. Gene/Protein Symbol CST6 ENST00000312134 cystatin M/E tccgagacacgcacatcatc ccatctccatcgtcaggaagtac 1.96 CTSL2 ENST00000259470 cathepsin L2/V, CTSV gctatggatgcaggccattc acaccatgatccaggtttttgc 1.97 CTSL ENST00000257498 cathepsin L gttgctattgatgcaggtcatga actgctacagtctggctcaaaataaa 1.86 LGMN ENST00000334869 legumain, AEP, PRSC1 tgcagtatggaaacaaaacaatctc gggtgaggtcaaggtgtgtga 1.85 RPLP0 ENST00000228306 ribosomal protein P0, hARP caccattgaaatcctgagtgatgt tgaccagcccaaaggagaag 2.00 TGM3 ENST00000381458 transglutaminase 3, TGE ggaaggactctgccacaatgtc tgtctgacttcaggtacttctcatactg 1.90

*E is efficiency as fold increase in fluorescence per PCR cycle

104 Generation of human skin equivalents Human skin equivalents were prepared as described previously31. Briefly, de‐epidermized dermis (DED) was punched with a 8‐mm biopter. DED was generated using abdominal skin from donors who underwent surgery for abdominal wall correction. Punches were placed in a 24 well transwell plate (Costar, NY, USA), and seeded with 105 keratinoyctes. After culturing the constructs 3 days submerged, the medium level was lowered, and the constructs were cultured for indicated time periods at the air‐liquid interface.

Statistical analysis All data were analyzed with the Statistica software package version 7.0 (StatSoft Inc). For the qPCR experiments (Fig. 3), statistical analysis was performed on the ΔCt values, which is the difference between the Ct of the target gene and the reference gene. For semi‐quantitative protein values (Fig. 5), statistical analysis was performed on the intensity (fluorescence units) of the protein staining in the whole epidermis. Regarding the expression of all genes/proteins in the epidermis of healthy controls and in AD and PS skin, the one‐way ANOVA test was performed. Post‐hoc testing for significance was done by Bonferroni testing.

RESULTS

We have previously shown by immunohistochemistry and RT‐PCR that cystatin M/E expression is largely confined to human skin and its appendages1,3,7. Immunohistochemical data have also shown a moderate but specific expression of cystatin M/E in the bronchi of the lung, the proximal tubuli of the kidney, and in the cornea of the eye (unpublished data). Some other studies have reported expression of cystatin M/E at the mRNA level in some other human tissues32,33 that are contradictory with our data. Therefore we analyzed the expression of cystatin M/E, CTSV, CTSL, LGMN, and TGM3 in a large panel of human tissues at the mRNA level by qPCR, which is currently the most accurate method to detect quantitative gene expression levels (Fig. 2). These experiments revealed that cystatin M/E, CTSV, and TGM3 show a more restricted tissue expression pattern, whereas CTSL and LGMN are ubiquitously expressed. mRNA expression of all five genes was only seen in skin, esophagus and oropharyngeal tissues. Co‐expression of cystatin M/E and its target proteases was also found in bronchus and kidney, however the expression levels were moderate. Aorta tissue showed a relatively high expression of cystatin M/E mRNA, while CTSV mRNA levels are extremely high in testis. In skin, cystatin M/E is expressed at relatively higher levels than its target proteases, when compared to other tissues, which emphasizes its prominent role in cutaneous biology. To investigate the cystatin M/E – cathepsin pathway in AD and PS we studied the expression of the above mentioned molecules in skin biopsies from these patients compared to skin from healthy controls. Figure 3 shows the relative mRNA expression levels of these molecules in purified epidermal cells. We found significant changes in expression for each

105 gene. Cystatin M/E and CTSV mRNA levels are decreased in inflamed skin compared to NS (although not significantly for CTSV in NS vs AD), while CTSL and TGM3 mRNA expression is increased in inflamed skin. We found a remarkable difference in CTSL expression, which was much higher in AD compared to PS. We also noticed significant lower levels of cystatin M/E mRNA in PS compared to AD. To study the expression at the protein level immunofluorescence microscopy was performed followed by quantitative analysis using ImageJ software. The localization of the proteins in inflamed skin did not differ from NS. Cystatin M/E, CTSV, and TGM3 expression is mainly restricted to the stratum granulosum (SG) and the SC, while CTSL and LGMN are expressed in all suprabasal layers except the SC (Fig. 4). Quantitative analysis of protein expression showed a significant decrease of cystatin M/E and CTSV in AD and PS, which corresponds to the mRNA expression levels found with qPCR analysis (Fig. 5). Western blotting on epidermal protein extracts from AD, PS, and normal skin showed similar results for CTSV (data not shown). However, the protein expression of CTSL, LGMN, and TGM3 was somewhat different compared to mRNA expression as CTSL protein expression was significantly decreased in inflamed skin compared to NS (as opposed to increased CTSL mRNA expression), while no differences are seen in at the protein level for LGMN and TGM3 between the three skin conditions. Recently, our group has adapted a model to generate normal human reconstructed skin, in which skin constructs differentiate from 1‐2 basal layers of keratinocytes at day 0 into a completely stratified epidermis on day 12 (Fig. 6). Immunohistochemical staining of differentiation markers shows that involucrin is already present in the upper layer of the

Figure 2: Relative mRNA expression levels of cystatin M/E, CTSV, CTSL, LGMN and TGM3 in a panel of human tissues. Equal amounts from each tissue were analyzed by qPCR. CTSV expression in tongue was set to the value 1. Note that the relative quantity is presented in a log10 scale. All ΔCt values were corrected for the efficiency of each primer set in relation to the housekeeping gene primers (see table 1). CysM = cystatin M/E.

106 construct on day 0 and that during differentiation of the epidermis the other structural proteins are being expressed: loricrin from day 3 and filaggrin from day 6. Subsequently, we studied the expression of all molecules of the cystatin M/E – cathepsin pathway in the human reconstructed skin model, both on the mRNA and protein level. The relative quantity of CTSL and LGMN mRNA expression levels did not change during the development of the stratified epidermis, but cystatin M/E, CTSV and TGM3 mRNA expression increased as soon as the keratinocyte monolayer started to differentiate (Fig. 7). These results were confirmed for each molecule on the protein level by immunofluorescence microscopy (Fig. 8). Cystatin M/E is not expressed in the skin construct at day 0, but expression gradually emerges in the SG during differentiation and later also in the SC. The same expression pattern can be seen for TGM3. CTSV expression is weak on day 0, but clear from day 3 with specific localization in the upper layers of the SG and the SC, while CTSL and LGMN are expressed in all cell layers during the entire differentiation process. The human reconstructed skin model shows the same expression and localization of the proteins compared to normal human skin (Fig. 4 and 8).

Figure 3: Relative mRNA expression levels of cystatin M/E, CTSV, CTSL, LGMN and TGM3 in inflamed skin. qPCR was performed on RNA from purified epidermal cells of NS (n=7 individuals), lesional AD skin (n=9 patients) and lesional PS skin (n=9 patients). LGMN expression in NS was set to the value 1. All ΔCt values were corrected for the efficiency of each primer set in relation to the housekeeping gene primers (see table 1). qPCR data are presented in mean ± SEM. Statistical significance was P < 0.001 (***), P < 0.01 (**), and P < 0.05 (*). NS=normal skin, AD=atopic dermatitis, PS=psoriasis, CysM = cystatin M/E.

107

Figure 4: Protein expression of cystatin M/E and its target proteases in inflamed skin. Representative immunofluorescence stainings (each group n=6) of Cystatin M/E, CTSV, CTSL, LGMN, and TGM3 protein expression in normal human skin (left column), and in skin of patients with AD (middle column) and PS (right column). Nuclei are blue (DNA staining by 4’,6‐diamine‐2’‐phenylindole dihydrochloride (DAPI)). Bar = 100 µm, CysM = cystatin M/E.

108 Figure 5: Quantitative measurement of protein expression. Immunofluorescence staining of Cystatin M/E, CTSV, CTSL, LGMN, and TGM3 protein expression in NS, and in skin of patients with AD and PS (each group n=6), was quantitatively analyzed using ImageJ software. Data represent intensity of proteins in epidermis (Fluoresence Units) and are given as mean ± SEM. Statistical significance was P < 0.001 (***). NS=normal skin, AD=atopic dermatitis, PS=psoriasis, CysM = cystatin M/E.

DISCUSSION

In this study we report the expression of cystatin M/E, CTSV, CTSL, LGMN, and TGM3 in a large panel of human tissues, and in healthy and inflamed skin at the mRNA level as well as at the protein level. In addition, we found that human reconstructed skin recapitulates the expression pattern of all players in the pathway. Although mRNA expression of cystatin M/E, CTSV, and CTSL in various human tissues has been studied before using other methods, qPCR analysis is currently the most accurate method for detection of quantitative gene expression levels. Here we show cystatin M/E mRNA expression in human skin, bronchus, lung, and kidney, confirming results from previous studies1,32,33. In addition, cystatin M/E expression was observed in oropharyngeal tissues, esophagus, and aorta, however cystatin M/E expression in most other tissues could not be detected. qPCR not only allows the comparison of relative expression levels between genes, but is also more accurate in detecting low mRNA levels that could not be detected using Northern blot and RT‐PCR analysis. A similar observation is made when comparing our

109 CTSV mRNA expression data with previous studies14,34,35. Using qPCR analysis we showed that CTSV mRNA expression is not only restricted to thymus, testis, skin, and the cornea of the eye, as previously reported, but the distribution is more diverse including esophagus and several gastrointestinal and oropharyngeal tissues. CTSL is ubiquitously expressed, which is in accordance with previous findings and confirmed by protein expression34,35. In this study, we also present the tissue distribution of human LGMN and TGM3 on the transcriptional level. Like CTSL, LGMN is also ubiquitously expressed at the mRNA level, while LGMN protein expression has been detected by immunohistochemistry only in human skin, kidney and liver so far3,36. Extensive expression analysis of TGM3 mRNA has only been performed in mice showing distribution in skin, stomach, and testis37. Our study suggests that the tissue distribution of human TGM3 is restricted to skin, esophagus, and oropharyngeal tissues. Remarkably, mRNA expression of all five proteins was only seen in skin, esophagus, and oropharyngeal tissues, all of which generate stratified epithelia that are directly exposed to micro‐organisms and physical stress, which is another clue that the cystatin M/E – cathepsin pathway may have a role in the differentiation process and epithelial barrier formation of these tissues.

Figure 6: Expression of epidermal differentiation markers in human reconstructed skin. Development of normal human skin equivalents. Adult human keratinocytes were seeded on DED, and cultured for 3 days submerged, followed by 0, 3, 6, 9 and 12 days culturing at the air‐liquid interface. Morphology of the skin constructs was studied by H&E staining. Differentiation of the skin constructs was studied by immunohistochemical stainings for involucrin, loricrin and filaggrin. Bar = 100 µm.

110 Recent studies have shown that the expression of genes and the formation of the epidermal barrier is altered in inflammatory skin diseases, which has encouraged us to analyze the expression pattern of all players of the novel biochemical pathway in AD and PS. qPCR analysis showed that mRNA expression levels of some genes are changed in diseased skin compared to NS. Generally, mRNA expression levels of cystatin M/E and CTSV are decreased in inflamed skin, while the levels of CTSL and TGM3 are increased. These findings are in accordance with micro‐array analyses in patients with AD or PS (GEO profiles GDS2381 and GDS2518, accessible at http://www.ncbi.nlm.nih.gov). GDS2381 contains mRNA expression profiles of lesional and non‐lesional skin from AD patients and NS from healthy volunteers. Differences in mRNA expression of the studied molecules of the cystatin M/E – cathepsin pathway are observed between AD patients and healthy volunteers, but no major differences were found within the AD patients group when lesional skin is compared to non‐ lesional skin. In contrast, GDS2518 showed a difference in mRNA expression between lesional and non‐lesional skin from PS patients, suggesting that the cystatin M/E – cathepsin pathway is disturbed during the development of the psoriatic plaques.

Figure 7: Relative mRNA expression levels of Cystatin M/E, CTSV, CTSL, LGMN and TGM3 in reconstructed skin. Adult human keratinocytes were seeded on DED, and cultured for 3 days submerged, followed by 0, 3, 6, 9 and 12 days culturing at the air‐liquid interface. qPCR was performed on RNA from epidermal cells derived from the cultured human skin equivalents. CTSV expression at day 0 was set to the value 1. All ΔCt values were corrected for the efficiency of each primer set in relation to the housekeeping gene primers (see table 1). qPCR data are presented in mean ± SEM of duplicate experiments. CysM = cystatin M/E.

111 Previously, we have reported that cystatin M/E expression in AD and PS skin was extended to some spinous layers and therefore, seemed to be increased38. However, only the number of positive cell layers was increased, but not necessarily the relative expression as staining intensity was less in AD and PS compared to normal skin. The present study confirms that relative cystatin M/E expression is actually decreased. Although our immunofluorescence stainings show that cystatin M/E is mainly expressed in the stratum granulosum and stratum corneum of AD and PS skin, we did detect cystatin M/E in the upper spinous layers, but this is difficult to see in figure 4. Differences in protocol, strength of signal and visual appearance between immunohistochemical staining and immunofluorescence could attribute to the apparent difference in expression pattern, e.g. immunohistochemical overstaining vs. adjusted fluorescence intensity.

Figure 8: Expression of cystatin M/E and its target proteases in reconstructed human skin. Adult human keratinocytes were seeded on DED, and cultured for 3 days submerged, followed by 0, 3, 6, 9 and 12 days culturing at the air‐liquid interface. Representative immuno‐fluorescence stainings (in duplo) showed the expression of Cystatin M/E, CTSV, CTSL, LGMN, and TGM3 in these reconstructed skin cultures. Nuclei are blue (DNA staining by 4’,6‐diamine‐2’‐phenylindole dihydrochloride (DAPI). Bar = 100 µm, CysM = cystatin M/E.

112 In both AD and PS decreased protein expression of cystatin M/E and CTSV was confirmed by quantification of immunofluorescent stainings, but the elevated transcriptional levels of CTSL and TGM3 were not matched by increased protein expression. Surprisingly, CTSL protein expression is decreased in inflamed skin, which not only contradicts our qPCR data, but also a previous study that reported increased expression of CTSL in PS39. However, Bylaite et al. were not able to detect CTSL expression in NS which questions their report since CTSL is expressed quite abundantly in NS3,7,40. The decrease of CTSL protein expression in inflamed skin could be explained by altered processing of CTSL. Multiple forms of the active CTSL enzyme exist and in NS only inactive proCTSL and the mature active single‐chain CTSL form are detected40. Expression of CTSL in NS appears as a diffuse staining of the entire epidermis except the SC. In PS enhanced processing of CTSL results in a mixture of active single‐chain and two‐chain CTSL forms, which shows a different, more granular and condensed staining pattern of CTSL in psoriatic skin40. It could be that our CTSL antibody cannot detect the two‐chain form as well as it does the single‐chain form, or perhaps the antibody is not suited for staining CTSL in condensed granules resulting in decreased protein detection. Our data suggest that the cystatin M/E ‐ cathepsin pathway is disturbed in AD and PS, which might contribute to an abnormal skin barrier formation and function, as an unbalanced protease/protease inhibitor interaction would have consequences for downstream processes in the pathway. Since the cystatin M/E – cathepsin pathway is involved in both cornification and desquamation, these processes are most likely to be affected3,7. Insufficient processing and activation of TGM3 by CTSL could result in aberrant cornification. Although TGM3 protein expression did not change in inflamed skin, our TGM3 antibodies cannot distinguish between inactive and active forms, thus leaving the possibility that in diseased skin most of the TGM3 protein could be inactive, resulting in impaired barrier formation by insufficient cross‐linking of structural proteins into the cornified envelope. Since CTSV could have an important role in desquamation, decreased CTSV activity could lead to delayed desquamation of corneocytes and thereby induce thickening and scaling of the skin in AD and PS. Other cystatins and cathepsins have been implicated in skin barrier formation and function and as such could have a role in the development of AD and/or PS lesions. Cystatin A is expressed in skin and participates in epidermal barrier formation and function41. It has been suggested that cystatin A inhibits exogenous cysteine proteases preventing microbial infection and inflammatory stimulation of the keratinocytes by mite allergens in AD42,43. Cystatin A expression is decreased in AD, whereas in psoriatic skin cystatin A is overexpressed44,45. Interestingly, the CTSA gene is located in the PS and AD loci on chromosome 3q2146,47. Cathepsin D (CTSD) is an aspartic protease that is involved in the processing and activation of TGM1. CTSD deficient mice show reduced TGM1 activity and impaired SC morphology, similar to the skin of TGM1 knockout mice and the human skin disease lamellar ichthyosis48. In addition, CTSD is able to degrade corneodesmosomal proteins and could have a role in the desquamation process49,50. CTSD expression is shown to be altered in PS; expression shifts from the upper granular layers and the SC in normal

113 human skin to all suprabasal layers except the SC in psoriatic skin51. Similar to CTSL, enhanced processing of CTSD was found in PS40. The use of knockout or transgenic mouse models is a cornerstone of modern biomedical research and has provided mechanistic insights in biology and pathology. We are currently using such mouse models to investigate the role cystatin M/E and its target proteases in skin biology and development. We reasoned that if a generally applicable strategy could be developed for gene knock‐down or overexpression in a model that closely resembles normal human skin, this would lead to a significant reduction of animal use and discomfort. In this study, we have used a human reconstructed skin model showing the same expression and localization of several differentiation markers and the key molecules of the cystatin M/E – cathepsin pathway compared to normal human skin. Therefore, this reconstructed skin model will be used as a tool for future research in which we aim to use lentiviral vectors for gene knockdown (by delivery of siRNA) or overexpression of certain genes.

ACKNOWLEDGMENTS

We thank Dr. Colin Watts (Division of Cell Biology & Immunology, University of Dundee, UK) for providing polyclonal sheep anti‐human LGMN antibodies.

REFERENCES

1. Zeeuwen, P. L., van Vlijmen‐Willems, I. M., Jansen, B. J., Sotiropoulou, G., Curfs, J. H., Meis, J. F., Janssen, J. J., van Ruissen, F., Schalkwijk, J. (2001) Cystatin M/E expression is restricted to differentiated epidermal keratinocytes and sweat glands: a new skin‐specific proteinase inhibitor that is a target for cross‐linking by transglutaminase. J.Invest Dermatol. 116, 693‐701 2. Zeeuwen, P. L., van Vlijmen‐Willems, I. M., Olthuis, D., Johansen, H. T., Hitomi, K., Hara‐ Nishimura, I., Powers, J. C., James, K. E., op den Camp, H. J., Lemmens, R., Schalkwijk, J. (2004) Evidence that unrestricted legumain activity is involved in disturbed epidermal cornification in cystatin M/E deficient mice. Hum.Mol.Genet. 13, 1069‐1079 3. Cheng, T., Hitomi, K., van Vlijmen‐Willems, I. M., de Jongh, G. J., Yamamoto, K., Nishi, K., Watts, C., Reinheckel, T., Schalkwijk, J., Zeeuwen, P. L. (2006) Cystatin M/E is a high affinity inhibitor of cathepsin V and cathepsin L by a reactive site that is distinct from the legumain‐binding site. A novel clue for the role of cystatin M/E in epidermal cornification. J.Biol.Chem. 281, 15893‐ 15899 4. Ahvazi, B., Kim, H. C., Kee, S. H., Nemes, Z., Steinert, P. M. (2002) Three‐dimensional structure of the human transglutaminase 3 enzyme: binding of calcium ions changes structure for activation. EMBO J. 21, 2055‐2067 5. Hitomi, K., Kanehiro, S., Ikura, K., Maki, M. (1999) Characterization of recombinant mouse epidermal‐type transglutaminase (TGase 3): regulation of its activity by proteolysis and guanine nucleotides. J.Biochem.(Tokyo) 125, 1048‐1054

114 6. Maehr, R., Hang, H. C., Mintern, J. D., Kim, Y. M., Cuvillier, A., Nishimura, M., Yamada, K., Shirahama‐Noda, K., Hara‐Nishimura, I., Ploegh, H. L. (2005) Asparagine endopeptidase is not essential for class II MHC antigen presentation but is required for processing of cathepsin L in mice. J.Immunol. 174, 7066‐7074 7. Zeeuwen, P. L., Ishida‐Yamamoto, A., van Vlijmen‐Willems, I. M., Cheng, T., Bergers, M., Iizuka, H., Schalkwijk, J. (2007) Colocalization of cystatin M/E and cathepsin V in lamellar granules and corneodesmosomes suggests a functional role in epidermal differentiation. J.Invest Dermatol. 127, 120‐128 8. Cheng, T., van Vlijmen‐Willems, I. M., Hitomi, K., Pasch, M. C., van Erp, P. E., Schalkwijk, J., Zeeuwen, P. L. (2009) Colocalization of Cystatin M/E and its Target Proteases Suggests a Role in Terminal Differentiation of Human Hair Follicle and Nail. J.Invest Dermatol. 129, 1232‐1242 9. Zeeuwen, P. L., van Vlijmen‐Willems, I. M., Hendriks, W., Merkx, G. F., Schalkwijk, J. (2002) A null mutation in the cystatin M/E gene of ichq mice causes juvenile lethality and defects in epidermal cornification. Hum.Mol.Genet. 11, 2867‐2875 10. Sundberg, J. P., Boggess, D., Hogan, M. E., Sundberg, B. A., Rourk, M. H., Harris, B., Johnson, K., Dunstan, R. W., Davisson, M. T. (1997) Harlequin ichthyosis (ichq): a juvenile lethal mouse mutation with ichthyosiform dermatitis. Am.J.Pathol. 151, 293‐310 11. Benavides, F., Starost, M. F., Flores, M., Gimenez‐Conti, I. B., Guenet, J. L., Conti, C. J. (2002) Impaired hair follicle morphogenesis and cycling with abnormal epidermal differentiation in nackt mice, a cathepsin L‐deficient mutation. Am.J.Pathol. 161, 693‐703 12. Roth, W., Deussing, J., Botchkarev, V. A., Pauly‐Evers, M., Saftig, P., Hafner, A., Schmidt, P., Schmahl, W., Scherer, J., Anton‐Lamprecht, I., Von Figura, K., Paus, R., Peters, C. (2000) Cathepsin L deficiency as molecular defect of furless: hyperproliferation of keratinocytes and pertubation of hair follicle cycling. FASEB J. 14, 2075‐2086 13. Tobin, D. J., Foitzik, K., Reinheckel, T., Mecklenburg, L., Botchkarev, V. A., Peters, C., Paus, R. (2002) The lysosomal protease cathepsin L is an important regulator of keratinocyte and melanocyte differentiation during hair follicle morphogenesis and cycling. Am.J.Pathol. 160, 1807‐1821 14. Brömme, D., Li, Z., Barnes, M., Mehler, E. (1999) Human cathepsin V functional expression, tissue distribution, electrostatic surface potential, enzymatic characterization, and chromosomal localization. Biochemistry 38, 2377‐2385 15. Hagemann, S., Gunther, T., Dennemarker, J., Lohmuller, T., Bromme, D., Schule, R., Peters, C., Reinheckel, T. (2004) The human cysteine protease cathepsin V can compensate for murine cathepsin L in mouse epidermis and hair follicles. Eur.J.Cell Biol. 83, 775‐780 16. Shirahama‐Noda, K., Yamamoto, A., Sugihara, K., Hashimoto, N., Asano, M., Nishimura, M., Hara‐Nishimura, I. (2003) Biosynthetic processing of cathepsins and lysosomal degradation are abolished in asparaginyl endopeptidase‐deficient mice. J.Biol.Chem. 278, 33194‐33199 17. Zeeuwen, P. L. (2004) Epidermal differentiation: the role of proteases and their inhibitors. Eur.J.Cell Biol. 83, 761‐773 18. Toomes, C., James, J., Wood, A. J., Wu, C. L., McCormick, D., Lench, N., Hewitt, C., Moynihan, L., Roberts, E., Woods, C. G., Markham, A., Wong, M., Widmer, R., Ghaffar, K. A., Pemberton, M., Hussein, I. R., Temtamy, S. A., Davies, R., Read, A. P., Sloan, P., Dixon, M. J., Thakker, N. S. (1999) Loss‐of‐function mutations in the cathepsin C gene result in periodontal disease and palmoplantar keratosis. Nat.Genet. 23, 421‐424

115 19. Basel‐Vanagaite, L., Attia, R., Ishida‐Yamamoto, A., Rainshtein, L., Ben, A. D., Lurie, R., Pasmanik‐Chor, M., Indelman, M., Zvulunov, A., Saban, S., Magal, N., Sprecher, E., Shohat, M. (2007) Autosomal recessive ichthyosis with hypotrichosis caused by a mutation in ST14, encoding type II transmembrane serine protease matriptase. Am.J.Hum.Genet. 80, 467‐477 20. List, K., Currie, B., Scharschmidt, T. C., Szabo, R., Shireman, J., Molinolo, A., Cravatt, B. F., Segre, J., Bugge, T. H. (2007) Autosomal ichthyosis with hypotrichosis syndrome displays low matriptase proteolytic activity and is phenocopied in ST14 hypomorphic mice. J.Biol.Chem. 282, 36714‐36723 21. Chavanas, S., Bodemer, C., Rochat, A., Hamel‐Teillac, D., Ali, M., Irvine, A. D., Bonafe, J. L., Wilkinson, J., Taieb, A., Barrandon, Y., Harper, J. I., de Prost, Y., Hovnanian, A. (2000) Mutations in SPINK5, encoding a serine protease inhibitor, cause Netherton syndrome. Nat.Genet. 25, 141‐142 22. Palmer, C. N., Irvine, A. D., Terron‐Kwiatkowski, A., Zhao, Y., Liao, H., Lee, S. P., Goudie, D. R., Sandilands, A., Campbell, L. E., Smith, F. J., O'Regan, G. M., Watson, R. M., Cecil, J. E., Bale, S. J., Compton, J. G., Digiovanna, J. J., Fleckman, P., Lewis‐Jones, S., Arseculeratne, G., Sergeant, A., Munro, C. S., El, H. B., McElreavey, K., Halkjaer, L. B., Bisgaard, H., Mukhopadhyay, S., McLean, W. H. (2006) Common loss‐of‐function variants of the epidermal barrier protein filaggrin are a major predisposing factor for atopic dermatitis. Nat.Genet. 38, 441‐446 23. Hollox, E. J., Huffmeier, U., Zeeuwen, P. L., Palla, R., Lascorz, J., Rodijk‐Olthuis, D., van de Kerkhof, P. C., Traupe, H., de Jongh, G., den Heijer, M., Reis, A., Armour, J. A., Schalkwijk, J. (2008) Psoriasis is associated with increased beta‐defensin genomic copy number. Nat.Genet. 40, 23‐25 24. Li, D. N., Matthews, S. P., Antoniou, A. N., Mazzeo, D., Watts, C. (2003) Multistep autoactivation of asparaginyl endopeptidase in vitro and in vivo. J.Biol.Chem. 278, 38980‐38990 25. Hitomi, K., Presland, R. B., Nakayama, T., Fleckman, P., Dale, B. A., Maki, M. (2003) Analysis of epidermal‐type transglutaminase (transglutaminase 3) in human stratified epithelia and cultured keratinocytes using monoclonal antibodies. J.Dermatol.Sci. 32, 95‐103 26. Latijnhouwers, M. A., Bergers, M., van Bergen, B. H., Spruijt, K. I., Andriessen, M. P., Schalkwijk, J. (1996) Tenascin expression during wound healing in human skin. J.Pathol. 178, 30‐35 27. van Ruissen, F., Jansen, B. J., de Jongh, G. J., Zeeuwen, P. L., Schalkwijk, J. (2002) A partial transcriptome of human epidermis. Genomics 79, 671‐678 28. de Jongh, G. J., Zeeuwen, P. L., Kucharekova, M., Pfundt, R., Valk, P. G., Blokx, W., Dogan, A., Hiemstra, P. S., Kerkhof, P. C., Schalkwijk, J. (2005) High expression levels of keratinocyte antimicrobial proteins in psoriasis compared with atopic dermatitis. J.Invest Dermatol. 125, 1163‐1173 29. Franssen, M. E., Zeeuwen, P. L., Vierwinden, G., van de Kerkhof, P. C., Schalkwijk, J., van Erp, P. E. (2005) Phenotypical and functional differences in germinative subpopulations derived from normal and psoriatic epidermis. J.Invest Dermatol. 124, 373‐383 30. Livak, K. J., Schmittgen, T. D. (2001) Analysis of relative gene expression data using real‐time quantitative PCR and the 2(‐Delta Delta C(T)) Method. Methods 25, 402‐408 31. Tjabringa, G., Bergers, M., van Rens, D., de Boer, R., Lamme, E., Schalkwijk, J. (2008) Development and validation of human psoriatic skin equivalents. Am.J.Pathol. 173, 815‐823 32. Sotiropoulou, G., Anisowicz, A., Sager, R. (1997) Identification, cloning, and characterization of cystatin M, a novel cysteine proteinase inhibitor, down‐regulated in breast cancer. J.Biol.Chem. 272, 903‐910

116 33. Ni, J., Abrahamson, M., Zhang, M., Fernandez, M. A., Grubb, A., Su, J., Yu, G. L., Li, Y., Parmelee, D., Xing, L., Coleman, T. A., Gentz, S., Thotakura, R., Nguyen, N., Hesselberg, M., Gentz, R. (1997) Cystatin E is a novel human cysteine proteinase inhibitor with structural resemblance to family 2 cystatins. J.Biol.Chem. 272, 10853‐10858 34. Santamaria, I., Velasco, G., Cazorla, M., Fueyo, A., Campo, E., Lopez‐Otin, C. (1998) Cathepsin L2, a novel human cysteine proteinase produced by breast and colorectal carcinomas. Cancer Res. 58, 1624‐1630 35. Adachi, W., Kawamoto, S., Ohno, I., Nishida, K., Kinoshita, S., Matsubara, K., Okubo, K. (1998) Isolation and characterization of human cathepsin V: a major proteinase in corneal epithelium. Invest Ophthalmol.Vis.Sci. 39, 1789‐1796 36. Liu, C., Sun, C., Huang, H., Janda, K., Edgington, T. (2003) Overexpression of legumain in tumors is significant for invasion/metastasis and a candidate enzymatic target for prodrug therapy. Cancer Res. 63, 2957‐2964 37. Hitomi, K., Horio, Y., Ikura, K., Yamanishi, K., Maki, M. (2001) Analysis of epidermal‐type transglutaminase (TGase 3) expression in mouse tissues and cell lines. Int.J.Biochem.Cell Biol. 33, 491‐498 38. Zeeuwen, P. L., van Vlijmen‐Willems, I. M., Egami, H., Schalkwijk, J. (2002) Cystatin M/E expression in inflammatory and neoplastic skin disorders. Br.J.Dermatol. 147, 87‐94 39. Bylaite, M., Moussali, H., Marciukaitiene, I., Ruzicka, T., Walz, M. (2006) Expression of cathepsin L and its inhibitor hurpin in inflammatory and neoplastic skin diseases. Exp.Dermatol. 15, 110‐118 40. Kawada, A., Hara, K., Kominami, E., Hiruma, M., Noguchi, H., Ishibashi, A. (1997) Processing of cathepsins L, B and D in psoriatic epidermis. Arch.Dermatol.Res. 289, 87‐93 41. Takahashi, M., Tezuka, T., Katunuma, N. (1996) Filaggrin linker segment peptide and cystatin alpha are parts of a complex of the cornified envelope of epidermis. Arch.Biochem.Biophys. 329, 123‐126 42. Takahashi, M., Tezuka, T., Korant, B., Katunuma, N. (1999) Inhibition of cysteine protease and growth of Staphylococcus aureus V8 and poliovirus by phosphorylated cystatin alpha conjugate of skin. Biofactors 10, 339‐345 43. Kato, T., Takai, T., Mitsuishi, K., Okumura, K., Ogawa, H. (2005) Cystatin A inhibits IL‐8 production by keratinocytes stimulated with Der p 1 and Der f 1: biochemical skin barrier against mite cysteine proteases. J.Allergy Clin.Immunol. 116, 169‐176 44. Seguchi, T., Cui, C. Y., Kusuda, S., Takahashi, M., Aisu, K., Tezuka, T. (1996) Decreased expression of filaggrin in atopic skin. Arch.Dermatol.Res. 288, 442‐446 45. Bowcock, A. M., Shannon, W., Du, F., Duncan, J., Cao, K., Aftergut, K., Catier, J., Fernandez‐Vina, M. A., Menter, A. (2001) Insights into psoriasis and other inflammatory diseases from large‐ scale gene expression studies. Hum.Mol.Genet. 10, 1793‐1805 46. Enlund, F., Samuelsson, L., Enerback, C., Inerot, A., Wahlstrom, J., Yhr, M., Torinsson, A., Riley, J., Swanbeck, G., Martinsson, T. (1999) Psoriasis susceptibility locus in chromosome region 3q21 identified in patients from southwest Sweden. Eur.J.Hum.Genet. 7, 783‐790 47. Lee, Y. A., Wahn, U., Kehrt, R., Tarani, L., Businco, L., Gustafsson, D., Andersson, F., Oranje, A. P., Wolkertstorfer, A., Berg, A., Hoffmann, U., Kuster, W., Wienker, T., Ruschendorf, F., Reis, A. (2000) A major susceptibility locus for atopic dermatitis maps to chromosome 3q21. Nat.Genet. 26, 470‐473

117 48. Egberts, F., Heinrich, M., Jensen, J. M., Winoto‐Morbach, S., Pfeiffer, S., Wickel, M., Schunck, M., Steude, J., Saftig, P., Proksch, E., Schutze, S. (2004) Cathepsin D is involved in the regulation of transglutaminase 1 and epidermal differentiation. J.Cell Sci. 117, 2295‐2307 49. Horikoshi, T., Igarashi, S., Uchiwa, H., Brysk, H., Brysk, M. M. (1999) Role of endogenous cathepsin D‐like and chymotrypsin‐like proteolysis in human epidermal desquamation. Br.J.Dermatol. 141, 453‐459 50. Igarashi, S., Takizawa, T., Takizawa, T., Yasuda, Y., Uchiwa, H., Hayashi, S., Brysk, H., Robinson, J. M., Yamamoto, K., Brysk, M. M., Horikoshi, T. (2004) Cathepsin D, but not cathepsin E, degrades desmosomes during epidermal desquamation. Br.J.Dermatol. 151, 355‐361 51. Chen, S. H., Arany, I., Apisarnthanarax, N., Rajaraman, S., Tyring, S. K., Horikoshi, T., Brysk, H., Brysk, M. M. (2000) Response of keratinocytes from normal and psoriatic epidermis to interferon‐gamma differs in the expression of zinc‐alpha(2)‐glycoprotein and cathepsin D. FASEB J. 14, 565‐571

118 CHAPTER 7

Transgenic expression of cystatin M/E in epidermis to rescue the lethal ichq mouse phenotype

Tsing Cheng, Wiljan J.A.J. Hendriks, Thomas Reinheckel, Piet E.J. van Erp, Joost Schalkwijk, and Patrick L.J.M. Zeeuwen

119 Transgenic Expression of Cystatin M/E in Epidermis to Rescue the Lethal Ichq Mouse Phenotype

Tsing Cheng1, Wiljan J.A.J. Hendriks2, Thomas Reinheckel3, Piet E.J. van Erp1, Joost Schalkwijk1, and Patrick L.J.M. Zeeuwen1 Departments of 1Dermatology and 2Cell Biology, Nijmegen Centre for Molecular Life Sciences, Radboud University Nijmegen Medical Centre, PO Box 9101, 6500 HB Nijmegen, The Netherlands, and 3Institute of Molecular Medicine and Cell Research, Albert‐Ludwigs‐University, D‐79104 Freiburg, Germany

Cystatin M/E is a member of a superfamily of inhibitors that provide regulatory and protective functions against uncontrolled proteolysis by cysteine proteases. Although most cystatins are ubiquitously expressed, high levels of cystatin M/E expression are mainly restricted to epithelia of the skin and a few other organs, such as lung and eye. Homozygosity for cystatin M/E null alleles in ichq mice leads to disturbed epidermal cornification, impaired barrier function, and dehydration. Unfortunately, the consequences of cystatin M/E deficiency in other tissues and organs cannot be studied in these mice as they die 6 to 12 days after birth. In order to overcome this, we devised a strategy to rescue the lethal skin phenotype of cystatin M/E deficient mice by transgenic expression of cystatin M/E in epidermis. A vector was generated in which mouse cystatin M/E (the Cst6 gene product) was placed under the control of the involucrin (INV) promoter. In addition, the Cts6 coding sequence was fused to an IRES – eGFP (enhanced green fluorescent protein) cassette, resulting in the production of both cystatin M/E and eGFP from a bicistronic mRNA in epidermis. Expression of the cystatin M/E – eGFP transgene (Tg(INV‐Cst6‐IRES‐eGFP)), was validated in vitro by transient transfection of cultured primary human keratinocytes. Transfected keratinocytes that were stimulated to differentiate indeed displayed eGFP protein expression as measured by flow cytometry and fluorescence microscopy. Also, mouse cystatin M/E encoding transcripts as well as production and secretion of mouse cystatin M/E protein were successfully detected in these cultures. As a next step the transgene was used for conventional micro‐injection into mouse zygotes and several transgenic founder mice have been obtained. These transgenic, epidermis‐specific cystatin M/E expressing mice will be cross‐bred with cystatin M/E deficient mice in order to test feasibility of rescuing the ichq neonatal lethal skin phenotype.

INTRODUCTION

Cystatin M/E is a cysteine protease inhibitor that is mainly expressed in the epidermis, more specifically in the upper layers of the stratum granulosum, sweat glands, sebaceous glands, and the hair follicles1,2, suggesting an important role for cystatin M/E in skin and hair physiology. Recently, we have elucidated a novel biochemical pathway that

120 controls skin barrier formation by regulating both cornification and desquamation of the SC. Cystatin M/E plays a key role in this pathway as it controls the activity of both the papain‐like proteases cathepsins L and V (CTSL, CTSV) and the asparaginyl endopeptidase legumain (LGMN) via two distinct binding sites3. The importance of cystatin M/E and its target proteases in terminal differentiation of the epidermis is illustrated by the severe skin phenotype of the cystatin M/E deficient ichq mouse, which is characterized by abnormal SC and hair follicle formation, disturbed skin barrier function, and neonatal lethality4‐6. Cystatin M/E might regulate crosslinking of structural proteins by transglutaminase 3 (TGM3) in the cornification process by controlling CTSL activity, as human CTSL is the elusive enzyme that processes and activates human TGM33. Furthermore, cystatin M/E could also have a role in desquamation by regulating CTSV activity, since in the extracellular space of the SC both proteins are closely associated with corneodesmosomes, and CTSV is able to degrade corneodesmosomal proteins7,8. However, no orthologue of CTSV in mouse has been discovered yet and it has been suggested that mouse CTSL could actually be the orthologue of human CTSV9. Another possibility is that mouse CTSL controls the specific functional activities of both human CTSL and CTSV. Inhibition of LGMN by cystatin M/E probably also regulates cathepsin activity as LGMN was shown to be involved in the processing of lysosomal proteases such as cathepsins B, H, and L in mice10,11. The neonatal lethal skin phenotype of cystatin M/E deficient mice makes it impossible to study the consequences of cystatin M/E deficiency in other tissues and organs in adult animals. The generation of double knockout and/or transgenic “rescued” mice will therefore be a good approach to circumvent this problem. Recently, we have generated CTSL‐cystatin M/E double knockout mice to study the role of cystatin M/E and its target protease in skin physiology. Preliminary results showed that the CTSL‐cystatin M/E double knockout mouse is viable, which suggests that the balance between the protease CTSL and its inhibitor cystatin M/E is highly important in epidermal barrier formation and function. Remarkably, the combination of cystatin M/E deficiency and heterozygosity for CTSL (Cst6‐/‐ /Ctsl+/‐) resulted in an “intermediate” skin phenotype with severe non‐lethal epidermal abnormalities and corneal inflammation (keratitis) (Zeeuwen et al., manuscript in preparation). To be able to study effects of unrestricted protease activity due to the ichq mutation in other tissues than skin (e.g. bronchi, eye, kidney) we aim to rescue cystatin M/E deficiency in skin by classical transgenesis involving a keratinocyte‐specific mouse cystatin M/E expression construct. For specific epidermal expression the involucrin (INV) promoter was chosen since this is active in the suprabasal, differentiating layers of epidermis and hair follicles12. Furthermore, an IRES (internal ribosomal entry site) – eGFP (enhanced green fluorescent protein) sequence was included to enable easy monitoring of transgene expression by virtue of the intrinsic fluorescence of the eGFP reporter protein13,14. The resulting pBS‐INV‐Cst6‐IRES‐eGFP plasmid was first functionally tested in vitro in human keratinocytes before the transgene was used for micro‐injection into mouse zygotes. Several transgenic founder mice were obtained and are currently used in breeding experiments with

121 heterozygous ichq mice. This may allow the derivation of viable mice that express cystatin M/E specifically in the epidermis, but remain cystatin M/E deficient in all other tissues.

EXPERIMENTAL PROCEDURES

RNA isolation and RT‐PCR cloning Total RNA from mouse tissues was isolated using TRIzol Reagent (Life Technologies, Gaithersburg, MD) as previously described4. Oligo‐dT primed single strand cDNA was synthesized from total RNA using Moloney murine leukemia virus RNase‐H‐ reverse transcriptase (Boehringer, Mannheim, Germany). To amplify the full mouse Cst6 coding sequence we used the following oligonucleotide primers: forward primer 5’‐ CTGAATCCGCGGCTATGGAG‐3’ and reverse primer 5’‐CCTGACTCTGTCACCCTGG‐3’ (corresponding to nucleotide positions 13‐32 and 483‐501, respectively, in the mouse Cst6 cDNA sequence; accession ID: MGI:1920970). PCR conditions (1 mM MgCl2) were: 94°C for 6 min followed by 35 cycles of 94°C for 1 min, 57°C for 1 min, and 72°C for 2 min. Resulting mouse Cst6 cDNA fragment was cloned into the pCR2.1‐TOPO cloning vector (Invitrogen, Carlsbad, CA).

Construction of the cystatin M/E expression vector The Cst6 cDNA vector, together with the human involucrin (INV) promoter‐containing plasmid pBS‐3700 (a kind gift from Dr. Joseph M. Carroll), and the pIRES2‐EGFP vector (BD Biosciences Clontech, Mountain View, CA), were used for construction of the final cystatin M/E expression vector (pBS‐INV‐Cst6‐IRES‐eGFP). First, pIRES2‐EGFP was modified by inserting a BamH I adaptor (5’‐TTAACGGGATCCCG‐3’) at the unique Afl II site. Subsequently, the internal ribosomal entry site (IRES), enhanced green fluorescent protein (eGFP), and simian virus 40 (SV40) Poly(A) signal sequences were excised with BamH I and inserted into the unique BamH I site in pBS‐3700. Finally, the Cst6 coding sequence was cloned into the intermediate vector using EcoR I resulting in the expression vector pBS‐INV‐Cst6‐IRES‐eGFP. All cloning steps were examined by DNA sequence analysis.

Cell culture and transfection of human keratinocytes Human epidermal keratinocytes, obtained from abdominal skin after surgical correction, were cultured in keratinocyte growth medium (KGM) as described previously15. Transfection of keratinocytes with pBS‐INV‐Cst6‐IRES‐eGFP or pIRES2‐eGFP was performed at 90‐100% confluency followed by switching to KGM/5% FCS medium to stimulate keratinocyte differentiation. Transfections were performed in OptiMEM I medium (Invitrogen) using Lipofectamine 2000 (LF2000, Invitrogen) according to manufacturer’s protocol with a DNA : LF2000 ratio of 1 µg : 5 µl. Finally, keratinocytes were harvested for flowcytometry or RNA isolation, while the culture supernatant was used for ELISA.

122 Detection of eGFP protein expression Cultered keratinocytes that were harvested and resuspended in PBS were sorted using an EPICS Elite flowcytometer (Coulter, Luton, UK) that detects fluorescent signals. Viable cells were separated from dead cells based on morphology and amount of propidium iodide uptake, and analyzed for eGFP expression. In addition, keratinocyte suspensions were also directly analyzed by fluorescence microscopy.

Detection of Cst6 transgenic transcripts RNA was isolated from cultured keratinocytes, treated with DNase I (Invitrogen) to prevent DNA contamination, and used for cDNA synthesis as described above. Subsequently, the cDNA templates were used in a PCR reaction for amplification of the mouse cystatin M/E part of the construct. The PCR (35 cycles) contained 2.5 mM MgCl2 and was performed at a melting temperature of 61°C using the following primers: forward primer 5’‐ ATGGAGCGTCCTCACTTCC‐3’ and reverse primer 5’‐CCTGACTCTGTCACCCTGG‐3’ (corresponding to nucleotide positions 27‐45 and 483‐501, respectively, in the mouse Cst6 cDNA sequence; accession ID: MGI:1920970). PCR products were analyzed on ethidium bromide‐stained 1.5% agarose gels.

Detection of mouse cystatin M/E by ELISA We developed an ELISA for specific detection of mouse cystatin M/E in medium from cultured human keratinocytes. Recombinant mouse cystatin M/E and human cystatin M/E (both from R&D Systems, Minneapolis, MN) were used as a standard and as a negative control, respectively, in concentrations varying from 5 ng/ml to 0.156 ng/ml. The wells of a 96‐well plate were coated overnight at 4°C with polyclonal rabbit anti‐human cystatin M/E antibody1, followed by incubation with 1% BSA (ICN Biomedicals, Aurora, OH) and 1% normal goat serum (Vector Laboratories Inc., Burlingame, CA) in PBS for 30 min. Subsequently, standards, controls, and samples (undiluted up to 32x diluted) were incubated for 1 hour, followed by incubation with monoclonal rat anti‐mouse cystatin M/E (R&D Systems) in PBS/1% normal rabbit serum/0.1% BSA/0.05% Tween‐20 for 30 min. Next, wells were incubated with goat anti‐rat biotinylated antibody (Vector Laboratories) for 30 min, followed by a final incubation with avidin‐biotinylated horseradish peroxidase complexes (Vector Laboratories) for 30 min. The above incubation steps were performed at 37°C and separated by repeated washing steps with PBS/0.05% Tween‐20. Chromogenic substrate 1‐step Ultra TMB (Thermo Scientific, Rockford, IL) was used as substrate for detection and the reaction was stopped by adding 4 M H2SO4. Each well was measured for mouse cystatin M/E at an absorbance of 450 nm with an ELISA micro‐plate reader (Bio‐Rad, Hercules, CA).

Micro‐injection of the transgene in mouse zygotes The generation of transgenic mice was done according to standard protocols16 and the guidelines of the local Committee for Animal Welfare. The transgenic expression plasmid was isolated from bacteria using an endotoxin‐free midi‐prep kit (Sigma, St. Louis, MO),

123 followed by excision of the INV‐Cst6‐IRES‐eGFP transgene (Tg(INV‐Cst6‐IRES‐eGFP)) using Sal I and Spe I restriction enzymes and subsequent purification of the linear DNA fragment. Superovulation was induced in female C57BL/6J mice and at 0.5 dpc (days post‐coitum) the mice were sacrificed and the fertilized eggs (zygotes) collected. The transgene was micro‐ injected into the male pronucleus and surviving zygotes were subsequently cultured in vitro overnight onto the two‐cell stage. Viable micro‐injected zygotes were then transferred into the oviduct of a 0.5 dpc pseudo‐pregnant female C57BL/6J mouse to resume embryonic development.

Genotyping of Cst6 alleles Tail tissue biopsies of mice born from Tg(INV‐Cst6‐IRES‐eGFP) micro‐injected zygotes were collected and samples were incubated overnight at 55°C for cell lysis in 200 µl DirectPCR solution (Viagen, Los Angeles, CA) and 100 µl proteinase K (0.5 mg/ml; Roche, Mannheim, Germany). Genomic DNA was isolated by phenol‐chloroform extraction and alcohol precipitation. The genotyping PCR (42 cycles) contained 2.5 mM MgCl2 and was performed at a melting temperature of 53°C using the following primers: forward primer (exon 2) 5’‐TACTACCTGACTTTGGACATA‐3’ and reverse primer (exon 3) 5’‐ AAAGTTGCAACGCAGTTT‐3’ (corresponding to nucleotide positions 285‐305 and 393‐410, respectively, in the mouse Cst6 cDNA sequence; accession ID: MGI:1920970). PCR products were analyzed on ethidium bromide‐stained 1.5% agarose gels.

Determination of transgene copy numbers To determine the copy number of integrated mouse Cst6 cDNA fragments we used a quantitative PCR. As an internal control a part of the genomic sequence of Tgm3 (intron

5/exon 6) was amplified. The PCR (35 cycles) contained 5 mM MgCl2, involved a melting temperature of 60°C and was performed using the following primers sets: for amplification of Cst6, forward primer 5’‐ATGGAGCGTCCTCACTTCC‐3’ with a 3’‐end tagged fluorescent FAM‐label, and reverse primer 5’‐TTTGCATCGATGACTTTGGTGT‐3’, for amplification of Tgm3, forward primer 5’‐CCCTCCCTCTAAGTGATTTGTT‐3’ with a 3’‐end tagged fluorescent FAM‐ label, and reverse primer 5’‐GGATCTCCACACTACCATTCCA‐3’. These correspond to nucleotide positions 27‐45 and 238‐259 in the mouse Cst6 cDNA sequence (accession ID: MGI:1920970) and positions 17716‐17737 and 17955‐17976 in mouse Tgm3 gene sequence (accession ID: MGI:98732), respectively. 2 µl PCR product (amplicon size is 233 and 261 bp for Cst6 and Tgm3, respectively) was added to 10 µl HiDi formamide (Applied Biosystems, Foster City, CA), and analyzed by electrophoresis on an ABI3730 analyzer using POP7 polymer (Applied Biosystems). Data were analyzed with PeakScanner software (Applied Biosystems).

124 RESULTS

Cystatin M/E transgene design To ensure proper tissue specificity and easy verification of expression we generated the plasmid pBS‐INV‐Cst6‐IRES‐eGFP that harbors five main elements relevant for correct expression and function (Figure 1). The first is a 3.7 kb fragment covering the promoter and the first exon and intron of the human involucrin (INV) gene. Involucrin is a structural protein of keratinocytes that is expressed in the suprabasal, differentiating layers of epidermis. The involucrin promoter is followed by the complete mouse cystatin M/E (Cst6) coding sequence that was amplified from mouse total RNA by reverse transcriptase PCR. Then an IRES sequence and the eGFP coding sequence are included, making up a fluorescent reporter for transgene expression, and finally the SV40 Poly(A) signal sequence ensures proper transcription termination and mRNA processing. The IRES sequence provides an additional entry site for ribosomes on the mRNA, enabling that not only the cystatin M/E open reading frame but also the one encoding the enhanced green fluorescent protein will be translated from this bicistronic mRNA. Thus, the involucrin promoter‐driven transcript should lead to the production of two separate proteins, i.e. mouse cystatin M/E and eGFP, in epidermal cells.

Expression of eGFP in transfected keratinocytes Human keratinocytes were transiently transfected with the cystatin M/E transgene‐ containing plasmid pBS‐INV‐Cst6‐IRES‐eGFP or with a control vector (pIRES2‐eGFP), and stimulated to differentiate for 72h by adding 5% FCS to the culture medium. The IRES sequence in the transgene construct will give rise to the parallel production of cystatin M/E and eGFP proteins, the latter of which is easily detectable by flowcytometry and fluorescence microscopy. For flowcytometry, viable keratinocytes were first selected based on forward scatter (indicating morphology) and propidium iodide uptake, and subsequently each selected keratinocyte was measured for eGFP expression level. The relative number of eGFP‐positive keratinocytes can be calculated based on the total number of viable

Figure 1: Map of the cystatin M/E –

eGFP bicistronic expression construct. The transgene construct for specific epidermal expression of cystatin M/E comprises five main elements: the 3.7 kb human involucrin promoter, a Cst6 cDNA fragment encoding mouse cystatin M/E, an IRES sequence, the eGFP coding sequence, and a poly(A) addition signal from the SV40 genome. The transgene can be excised from the pBS‐INV‐Cst6‐ IRES‐eGFP expression plasmid using Sal I and Spe I restriction sites.

125

Figure 2: Expression of eGFP in transfected keratinocytes. A) eGFP protein expression analysis by flowcytometry in non‐transfected and pBS‐INV‐ Cst6‐IRES‐eGFP transfected keratinoctyes. The transfected sample shows an increased number of fluorescent keratinocytes as compared to non‐ transfected cells. B) Fluorescence microscopy shows eGFP protein expression in a pBS‐INV‐Cst6‐ IRES‐eGFP transfected keratinocyte. keratinocytes that have been selected. Non‐transfected keratinocytes display almost no fluorescence (0.4%), whereas 2.8% of the keratinocytes transfected with pBS‐INV‐Cst6‐IRES‐ eGFP are eGFP‐positive. Moreover, the observed fluorescence intensity of transfected keratinocytes is 10 to 100‐fold higher than that of the non‐transfected, negative control (Figure 2A). Keratinocytes transfected with pIRES2‐eGFP displayed similar results (data not shown). Also, using fluorescence microscopy the eGFP‐positive keratinocytes could be detected in pBS‐INV‐Cst6‐IRES‐eGFP transfected keratinocyte cell suspensions (Figure 2B).

Detection of mouse cystatin M/E expression by transfected human keratinocytes Next we tested whether the eGFP‐encoding transcript in pBS‐INV‐Cst6‐IRES‐eGFP transfected keratinocytes also resulted in the production of cystatin M/E. As a first step we tested whether Cst6 sequences were indeed represented in the bicistronic messenger. Total RNA was subjected to RT‐PCR amplification of Cst6 coding sequences, which resulted in an amplicon of 475 bp (Figure 3A) using RNA derived from keratinocytes transfected with pBS‐ INV‐Cst6‐IRES‐eGFP. RNA from mock transfected and pIRES2‐eGFP transfected keratinocytes remained negative, confirming that the transgene gives rise to the bicistronic mRNA. Since cystatin M/E can be secreted by keratinocytes, culture medium was used to determine production of mouse cystatin M/E by transfected keratinocytes. We developed an ELISA for

126 specific detection of mouse cystatin M/E but not the endogenous human cystatin M/E. The polyclonal rabbit anti‐human cystatin M/E antibody, coated on the bottom of the plate, recognizes both human and mouse cystatin M/E, whereas the second antibody, monoclonal rat anti‐mouse cystatin M/E, is specific for mouse cystatin M/E only. Medium from keratinocytes transfected with pBS‐INV‐Cst6‐IRES‐eGFP contained approximately 15 ng/ml mouse cystatin M/E, whereas in media from mock transfected keratinocytes and keratinocytes transfected with the control pIRES2‐eGFP vector no mouse cystatin M/E could be detected (Figure 3B).

Generation of transgenic founders After excision from the expression vector the linear INV‐Cst6‐IRES‐eGFP transgene (Tg(INV‐Cst6‐IRES‐eGFP)), was micro‐injected into wild type C57BL/6J mouse zygotes. Resulting transgenic founder mice may contain one or multiple copies of the construct incorporated by random insertion into the host genomic DNA. Since the Cst6 transgene is intron‐less, specific genotyping of the litters for endogenous and transgenic Cst6 sequences can be performed using a PCR with primers that span exon borders. A mouse exon 2 forward and exon 3 reverse primer pair resulted in a 473 bp fragment from the endogenous Cst6 gene and a smaller fragment of 126 bp that represents the transgene (Figure 4) on DNA from transgenic founders (mouse No. 3). DNA from non‐transgenic littermates yielded only the 473 bp fragment. In this way we could identify two different transgenic founders (C57BL/6J‐ Tg(INV‐Cst6‐IRES‐eGFP)). To determine the number of incorporated transgenes we amplified an identical part in the endogenous and the transgenic Cst6 genes together with a segment

Figure 3: Expression of mouse cystatin M/E in transfected keratinocytes. A) Reverse transcriptase PCR on cDNA from non‐transfected (lane 1), pIRES2‐eGFP transfected (lane 2), and pBS‐INV‐Cst6‐ IRES‐eGFP transfected keratinocytes (lane 3) using primers specific for mouse Cst6. Only keratinocytes transfected with the cystatin M/E expression vector are positive for Cst6 mRNA as indicated by the 475 bp fragment (arrowhead). Water (lane 4) served as negative control for the PCR reaction. B) ELISA specific for the detection of mouse cystatin M/E in medium taken from non‐ transfected, pIRES2‐eGFP transfected, and pBS‐INV‐Cst6‐IRES‐eGFP transfected keratinocyte cultures. Recombinant human cystatin M/E serves as a negative control.

127 Figure 4: Genotyping of potentially Tg(INV‐ Cst6‐IRES‐eGFP) carrying mice. Amplification of Cst6 alleles on genomic DNA from a litter obtained after micro‐injection of the transgene (mice Nos. 1‐9), wild type control (wt), Cst6 cDNA control (cDNA), mixture of wild type genomic DNA and Cst6 cDNA (mix), and a control for the PCR reaction (water). The 473 bp fragment represents wild type Cst6 alleles in the genomic DNA, whereas the 126 bp fragment represents Tg(INV‐Cst6‐IRES‐eGFP) (arrowheads). Mouse No. 3 has incorporated the transgene into its genomic DNA (arrow).

Figure 5: Determination of Tg(INV‐Cst6‐IRES‐eGFP) copy numbers. Amplification of all Cst6 alleles (A1, A2) is compared to amplification of Tgm3 alleles (B1, B2) that serves as reference. The ratio between A2 and A1 is calculated by (A2/A1)*(B2/B1). A ratio of 5 : 1 indicates that besides the 2 wild‐type alleles 8 copies of Tg(INV‐Cst6‐IRES‐ eGFP) have been incorporated. from the endogenous Tgm3 gene using fluorescently tagged primers for quantification. In this assay the resulting signal coming from Cst6 loci corresponds to the two endogenous alleles as well as any incorporated transgene copy, while the signal from Tgm3 corresponds to the only two endogenous alleles (Figure 5). The data revealed that the transgenic founders have incorporated 2 and 4 copies of Tg(INV‐Cst6‐IRES‐eGFP), respectively, into their genomic DNA. Further breeding with these transgenic founders has demonstrated stable germline transmission of the transgene in the resulting offspring. Cross‐breeding with heterozygous ichq mice (Cst6+/‐) is now underway to assess whether successful rescue of the lethal skin barrier phenotype can occur.

DISCUSSION

In this study we have designed, generated and validated a transgene construct, pBS‐ INV‐Cst6‐IRES‐eGFP, for specific epidermal expression of mouse cystatin M/E. Human cultured keratinocytes that were transfected with the plasmid and were subsequently stimulated to differentiate did contain the bicistronic mRNA transcript and did successfully express the encoded proteins (i.e. mouse cystatin M/E protein and the fluorescent marker eGFP). The INV‐Cst6‐IRES‐eGFP transgene was subsequently micro‐injected into mouse zygotes and thus far the genotyping of resulting litters has revealed two transgenic founder mice that stably transmit the transgene insertion to their offspring.

128 Since cystatin M/E in skin is expressed largely in the differentiated keratinocytes of the epidermis, we aimed to generate a construct with a similar epidermal specific expression. We chose to use the human involucrin promoter that is active in the suprabasal, differentiating layers of epidermis12. The 3.7 kb involucrin promoter fragment used in our study has been optimized for maximal expression in keratinocyte cultures and covers the promoter and the first exon and intron17. Importantly, this promoter segment has been successfully used in vivo for the generation of transgenic mice that express the β‐ galactosidase reporter exclusively in the suprabasal layers of the epidermis and the hair follicles. Using the sensitive reverse transcriptase PCR assay it was shown that the reporter gene was also active in some other tissues with stratified squamous epithelia such as bladder and cervix, but X‐Gal histochemistry showed that expression was less uniform in those tissues as compared to epidermis. No transgene expression was detectable in organs with simple, non‐stratified epithelia including brain, heart, liver, lung, and kidney18. Transgenic mice with human involucrin promoter‐driven (over)expression of integrins or interferon‐γ have also been generated and show skin and hair phenotypes that resemble psoriasis and eczema, respectively, underscoring the differentiation and tissue specific regulatory potential of this involucrin promoter fragment19,20. Initiation of translation usually occurs by cap‐scanning of the RNA as ribosomal proteins bind at the cap‐structure at the 5’‐end of the RNA strand followed by linear scanning in the 5’‐3’ direction until the initiation site is reached. Originally identified in picornaviruses, that have uncapped RNA, the IRES enables the binding of ribosomes and initiation of translation at sites that are downstream of the 5’‐end of the RNA strand21,22. In this study we have used the encephalomyocarditis virus (ECMV) IRES, that has been successfully used in other studies23,24 and is commercially available in the pIRES2‐eGFP plasmid, as an internal binding site for ribosomes to allow the expression of the fluorescent eGFP reporter protein in cis with mouse cystatin M/E. Both proteins were indeed detected in human keratinocytes transfected with the bicistronic expression vector. The relative number of eGFP‐positive keratinocytes after transfection with the pBS‐INV‐Cst6‐IRES‐eGFP was only 2.4% (following background adjustment), indicating low transfection efficiency. Transfection with pIRES2‐eGFP resulted in a similar efficiency, and both are in accordance with previous findings25,26. However, for our specific purpose a low transfection efficiency is not problematic. The human keratinocyte transfection experiments indicated the suitability of the expression construct for animal studies. Transgenic founder mice were generated by micro‐ injection of the excised linear INV‐Cst6‐IRES‐eGFP transgene into wild type mouse zygotes. These transgenic mice are currently used for further breeding. Importantly, to verify proper transgene expression we do not rely on mRNA or protein isolation from epidermis, but by virtue of the IRES‐eGFP sequence we may confine ourselves to the monitoring of eGFP expression in skin using straightforward fluorescence. After confirmation of proper transgene expression, mice will be crossed with Cst6+/‐ mice in order to introduce the transgene into the cystatin M/E null background. The resulting compound transgenic mice

129 will then be intercrossed to generate the desired “rescued” mouse (Cst6‐/‐‐Tg(INV‐Cst6‐IRES‐ eGFP); Figure 6). Assuming that our conjecture regarding the cause of neonatal lethality is correct, “rescued” cystatin M/E deficient mice will not develop the lethal skin phenotype due to epidermal expression of the cystatin M/E transgene. Such a rescue would offer long‐ awaited opportunities to study the effects of cystatin M/E deficiency in other tissues and organs that normally express cystatin M/E (although at lower levels), such as lung, stomach, and brain4. Cystatin M/E has also been detected in cornea of the eye, and since keratitis was observed in Cst6‐/‐/Ctsl+/‐ mice we hypothesize that regulation of CTSL activity by cystatin M/E is important in the development of corneal epithelium as well (Zeeuwen et al., manuscript in preparation). Unfortunately, we may not be in the position to test this hypothesis in our “rescued” mouse model since a recent study demonstrated expression of involucrin in human cornea27. Extrapolating this finding to mouse tissue, the cornea of “rescued” mice would probably no longer be cystatin M/E deficient as transgenic cystatin M/E expression is controlled by the INV promoter. A previous study that used the INV promoter for transgenic expression of integrins in mice has reported abnormalities of the eyelid and cornea including inflammatory exudate and corneal scarring19, suggesting that the INV promoter is indeed active in the cornea. Irrespective, a successful “rescue” of the skin phenotype in cystatin M/E deficient mice will provide a unique opportunity to gain more insight into the function of cystatin M/E in various other tissues and organs besides the epidermis.

Figure 6: Breeding schedule for generation of “rescued” mouse. In this figure Tg refers to Tg(INV‐ Cst6‐IRES‐eGFP). First, the transgenic founder mice (Cst6+/+‐Tg ) are crossbred with heterozygous ichq mice (Cst6+/‐) to generate F1 Cst6+/‐‐Tg mice. Subsequently, the offspring of two Cst6+/‐‐Tg mice (F1) would contain “rescued” cystatin M/E deficient mice that will not develop the lethal skin phenotype due to epidermal expression of the cystatin M/E transgene (Cst6‐/‐‐Tg).

130 REFERENCES

1. Zeeuwen, P. L., van Vlijmen‐Willems, I. M., Jansen, B. J., Sotiropoulou, G., Curfs, J. H., Meis, J. F., Janssen, J. J., van Ruissen, F., Schalkwijk, J. (2001) Cystatin M/E expression is restricted to differentiated epidermal keratinocytes and sweat glands: a new skin‐specific proteinase inhibitor that is a target for cross‐linking by transglutaminase. J.Invest Dermatol. 116, 693‐701 2. Cheng, T., van Vlijmen‐Willems, I. M., Hitomi, K., Pasch, M. C., van Erp, P. E., Schalkwijk, J., Zeeuwen, P. L. (2009) Colocalization of cystatin M/E and its target proteases suggests a role in terminal differentiation of human hair follicle and nail. J.Invest Dermatol. 129, 1232‐1242 3. Cheng, T., Hitomi, K., van Vlijmen‐Willems, I. M., de Jongh, G. J., Yamamoto, K., Nishi, K., Watts, C., Reinheckel, T., Schalkwijk, J., Zeeuwen, P. L. (2006) Cystatin M/E is a high affinity inhibitor of cathepsin V and cathepsin L by a reactive site that is distinct from the legumain‐binding site. A novel clue for the role of cystatin M/E in epidermal cornification. J.Biol.Chem. 281, 15893‐ 15899 4. Zeeuwen, P. L., van Vlijmen‐Willems, I. M., Hendriks, W., Merkx, G. F., Schalkwijk, J. (2002) A null mutation in the cystatin M/E gene of ichq mice causes juvenile lethality and defects in epidermal cornification. Hum.Mol.Genet. 11, 2867‐2875 5. Zeeuwen, P. L., van Vlijmen‐Willems, I. M., Olthuis, D., Johansen, H. T., Hitomi, K., Hara‐ Nishimura, I., Powers, J. C., James, K. E., op den Camp, H. J., Lemmens, R., Schalkwijk, J. (2004) Evidence that unrestricted legumain activity is involved in disturbed epidermal cornification in cystatin M/E deficient mice. Hum.Mol.Genet. 13, 1069‐1079 6. Sundberg, J. P., Boggess, D., Hogan, M. E., Sundberg, B. A., Rourk, M. H., Harris, B., Johnson, K., Dunstan, R. W., Davisson, M. T. (1997) Harlequin ichthyosis (ichq): a juvenile lethal mouse mutation with ichthyosiform dermatitis. Am.J.Pathol. 151, 293‐310 7. Zeeuwen, P. L., Ishida‐Yamamoto, A., van Vlijmen‐Willems, I. M., Cheng, T., Bergers, M., Iizuka, H., Schalkwijk, J. (2007) Colocalization of cystatin M/E and cathepsin V in lamellar granules and corneodesmosomes suggests a functional role in epidermal differentiation. J.Invest Dermatol. 127, 120‐128 8. Cheng, T., Dolk, C., Caubet, C., Simon, M., Watts, C., Kamsteeg, M., Serre, G., Schalkwijk, J., Zeeuwen, P. L. (2009) A regulatory role for cystatin M/E in epidermal desquamation. Identification of a two‐chain form of cathepsin V. Submitted 9. Hagemann, S., Gunther, T., Dennemarker, J., Lohmuller, T., Bromme, D., Schule, R., Peters, C., Reinheckel, T. (2004) The human cysteine protease cathepsin V can compensate for murine cathepsin L in mouse epidermis and hair follicles. Eur.J.Cell Biol. 83, 775‐780 10. Shirahama‐Noda, K., Yamamoto, A., Sugihara, K., Hashimoto, N., Asano, M., Nishimura, M., Hara‐Nishimura, I. (2003) Biosynthetic processing of cathepsins and lysosomal degradation are abolished in asparaginyl endopeptidase‐deficient mice. J.Biol.Chem. 278, 33194‐33199 11. Maehr, R., Hang, H. C., Mintern, J. D., Kim, Y. M., Cuvillier, A., Nishimura, M., Yamada, K., Shirahama‐Noda, K., Hara‐Nishimura, I., Ploegh, H. L. (2005) Asparagine endopeptidase is not essential for class II MHC antigen presentation but is required for processing of cathepsin L in mice. J.Immunol. 174, 7066‐7074 12. Rice, R. H., Green, H. (1979) Presence in human epidermal cells of a soluble protein precursor of the cross‐linked envelope: activation of the cross‐linking by calcium ions. Cell 18, 681‐694

131 13. Young, W. S., III, Iacangelo, A., Luo, X. Z., King, C., Duncan, K., Ginns, E. I. (1999) Transgenic expression of green fluorescent protein in mouse oxytocin neurones. J.Neuroendocrinol. 11, 935‐939 14. Hadjantonakis, A. K., Dickinson, M. E., Fraser, S. E., Papaioannou, V. E. (2003) Technicolour transgenics: imaging tools for functional genomics in the mouse. Nat.Rev.Genet. 4, 613‐625 15. van Ruissen, F., de Jongh, G. J., Zeeuwen, P. L., van Erp, P. E., Madsen, P., Schalkwijk, J. (1996) Induction of normal and psoriatic phenotypes in submerged keratinocyte cultures. J.Cell Physiol 168, 442‐452 16. Hogan, B., Beddington, R., Constantini, F., Lacy, E. (1994) Manipulating the Mouse Embryo: A Laboratory Manual.Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 17. Carroll, J. M., Taichman, L. B. (1992) Characterization of the human involucrin promoter using a transient beta‐galactosidase assay. J.Cell Sci. 103 (Pt 4), 925‐930 18. Carroll, J. M., Albers, K. M., Garlick, J. A., Harrington, R., Taichman, L. B. (1993) Tissue‐ and stratum‐specific expression of the human involucrin promoter in transgenic mice. Proc.Natl.Acad.Sci.U.S.A 90, 10270‐10274 19. Carroll, J. M., Romero, M. R., Watt, F. M. (1995) Suprabasal integrin expression in the epidermis of transgenic mice results in developmental defects and a phenotype resembling psoriasis. Cell 83, 957‐968 20. Carroll, J. M., Crompton, T., Seery, J. P., Watt, F. M. (1997) Transgenic mice expressing IFN‐ gamma in the epidermis have eczema, hair hypopigmentation, and hair loss. J.Invest Dermatol. 108, 412‐422 21. Jackson, R. J., Howell, M. T., Kaminski, A. (1990) The novel mechanism of initiation of picornavirus RNA translation. Trends Biochem.Sci. 15, 477‐483 22. Stoneley, M., Willis, A. E. (2004) Cellular internal ribosome entry segments: structures, trans‐ acting factors and regulation of gene expression. Oncogene 23, 3200‐3207 23. Stenman, J., Toresson, H., Campbell, K. (2003) Identification of two distinct progenitor populations in the lateral ganglionic eminence: implications for striatal and olfactory bulb neurogenesis. J.Neurosci. 23, 167‐174 24. Fenske, T. S., Pengue, G., Mathews, V., Hanson, P. T., Hamm, S. E., Riaz, N., Graubert, T. A. (2004) Stem cell expression of the AML1/ETO fusion protein induces a myeloproliferative disorder in mice. Proc.Natl.Acad.Sci.U.S.A 101, 15184‐15189 25. Jiang, C. K., Connolly, D., Blumenberg, M. (1991) Comparison of methods for transfection of human epidermal keratinocytes. J.Invest Dermatol. 97, 969‐973 26. Hawley‐Nelson, P., Ciccarone, V. (1996) Transient transfection efficiency of human keratinocytes in two serum‐free media. Focus 18, 43‐44 27. Tong, L., Corrales, R. M., Chen, Z., Villarreal, A. L., De Paiva, C. S., Beuerman, R., Li, D. Q., Pflugfelder, S. C. (2006) Expression and regulation of cornified envelope proteins in human corneal epithelium. Invest Ophthalmol.Vis.Sci. 47, 1938‐1946

132 CHAPTER 8

Summary and Discussion

Tsing Cheng

Parts of this chapter have been published in: Zeeuwen, P.L., Cheng, T., and Schalkwijk, J. (2009) The biology of cystatin M/E and its cognate proteases. J.Invest.Dermatol. In press, doi:10.1038/jid.2009.40

133

134 Summary and Discussion

Our skin protects the body against invasion of micro‐organisms, penetration of toxic agents and loss of essential body fluids. Correct development of the epidermis is essential for skin barrier function as abnormalities in cornification and desquamation can lead to severe skin disorders. Elucidation of the molecular processes that underlie epidermal differentiation is of great importance for understanding skin barrier formation and function, and for future therapy of skin barrier disorders. The main subject of the studies presented in this thesis is cystatin M/E, a cysteine protease inhibitor that is highly expressed in skin, and has an important role in epidermal differentiation as indicated by the lethal skin phenotype of cystatin M/E deficient mice. However, the physiologic processes that are influenced by cystatin M/E during epidermal differentiation remained unclear. The general aim of this thesis was the elucidation of the exact function of cystatin M/E in epidermal differentiation.

SUMMARY OF RESULTS

The human skin barrier is a result of terminal differentiation of the epidermis, an intricate process of cornification and desquamation that is described in chapter 1. Epidermal differentiation involves a variety of proteins including proteases and their inhibitors. An overview on the classification, characteristics, and functions of proteases in general is presented, and in addition a more in‐depth study on lysosomal cysteine proteases is given, which describes the processing and cellular trafficking of cathepsins and provides profiles of cathepsin L, cathepsin V, and legumain. Furthermore, this chapter summarizes our current knowledge on cystatin M/E including studies on the cystatin M/E deficient mouse, the role of cystatin M/E as a tumor suppressor in cancer biology, and epigenetic silencing of cystatin M/E expression. Previous studies have shown that cystatin M/E is a potent inhibitor of the asparaginyl endopeptidase legumain and it was suggested that legumain activity in the epidermis and hair follicles is regulated by cystatin M/E. Cystatins are generally considered to be inhibitors of the papain‐like cathepsins, however, the cognate target cathepsin of cystatin M/E had not yet been identified at the start of this PhD‐project. Chapter 2 describes the identification of two novel target proteases. By using biochemical assays we showed that in vitro cystatin

M/E is a high affinity inhibitor of the cathepsins L and V with dissociation constants (Ki) of the protease inhibitor‐protease complexes in the nM range. Using recombinant cystatin M/E variants with mutations in predicted binding sites we showed that inhibition of cathepsin L and cathepsin V by cystatin M/E is mediated by a reactive site that is distinct from the legumain binding site. Following identification of the target proteases of cystatin M/E, localization and compartmentalization studies, as described in chapters 2 and 3, were performed to

135 investigate the biological relevance and function of these proteases in vivo. Chapter 2 includes immunohistochemical stainings, which showed that all molecules are expressed in normal human epidermis. Localization of cystatin M/E and cathepsin V is restricted to the stratum granulosum and the stratum corneum, whereas cathepsin L and legumain are expressed in all layers except the stratum corneum. Using immunofluorescence microscopy, specific co‐localization of cystatin M/E with cathepsin L and legumain was detected in the stratum granulosum, whereas cystatin M/E and cathepsin V co‐localize in the upper granular layers and the stratum corneum (Chapter 3). In addition, immunoelectron microscopy revealed that cystatin M/E and cathepsin V are separately transported within lamellar granules and secreted into the extracellular space of the stratum corneum, where both proteins are associated with corneodesmosomes. In contrast, cathepsin L and legumain were not found in lamellar granules, but in the cytoplasm of granular cells, which is remarkable as cathepsin L and legumain are considered to be lysosomal proteins (Chapter 3). However, lysosomal proteases may participate in physiologic processes outside lysosomes such as apoptosis, a process that is distinct from terminal differentiation of the epidermis but nevertheless shares some molecular and cellular features. During epidermal cornification cytosolic transglutaminase 3 (TGM3) is responsible for crosslinking of loricrin molecules and small proline‐rich proteins. TGM3 is produced as an inactive 77 kDa zymogen, that is activated via proteolytic cleavage into 47 kDa and 30 kDa fragments by an unknown protease. Our previous finding of premature TGM3 activity in cystatin M/E deficient mice has led to the suggestion that one of the target proteases of cystatin M/E might be responsible for TGM3 activation. In chapter 2 we demonstrated that cathepsin L is the elusive protease that can process and activate human TGM3, whereas other cathepsins (B, D, S, and V) were not able to cleave the TGM3 zymogen into the 47 kDa and 30 kDa fragments. Cleavage of the TGM3 zymogen by cathepsin L resulted in biologically active TGM3. N‐terminal sequencing of the 30 kDa fragment revealed that cathepsin L cleaves TGM3 at Ala466. Additionally, we showed co‐localization of cathepsin L and TGM3 in the stratum granulosum of normal human epidermis by immunofluorescence microscopy, which is described in chapter 5. The co‐localization of cystatin M/E and cathepsin V in the extracellular space of the stratum corneum near the corneodesmosomes suggested a role for both proteins in desquamation. Desquamation requires degradation of the corneodesmosomes, which is mediated through the proteolytic activities of secreted proteases. Therefore, the ability of cathepsin V to degrade the (corneo‐)desmosomal proteins desmoglein‐1, desmoglein‐3, and corneodesmosin was analyzed in chapter 4. Cathepsin V degrades these proteins at pH 5.5, which resembles the acidic environment of the upper layers of the stratum corneum, while at pH 7.0 little proteolysis was observed. Interestingly, it has been suggested that desquamation is a pH‐dependent process in which the gradual change in pH of the stratum corneum, from neutral in the lower layers to acidic in the upper layers, determines which proteases are mainly active in each of the different layers.

136 Lysosomal proteases are involved in the processing and maturation of each other, and it is known that in mice legumain is involved in the processing of lysosomal cathepsins from the single‐chain form into the two‐chain form. This implies that legumain might regulate cathepsin activity, and thus, indirectly the physiologic processes in which cathepsins are involved. In chapter 4 we describe that human legumain is able to process human active single‐chain cathepsin V into a previously unknown active two‐chain form. Immunoblotting of cathepsin V in epidermal extracts showed that the two‐chain form of cathepsin V exists in vivo. The cleavage site in cathepsin V was determined at Asn290 by N‐terminal sequencing and MALDI‐MS analysis. Although cathepsin L and cathepsin V are closely related to each other, human legumain did not process human cathepsin L since cleavage of human cathepsin L occurs at Asp291, while legumain is specific for hydrolysis of asparaginyl (Asn) bonds. Furthermore, legumain could not be processed by either cathepsin L or cathepsin V. Altogether, our biochemical assays and localization studies, combined with the cystatin M/E deficient mouse model, have revealed a novel biochemical pathway in which cystatin M/E regulates cornification and desquamation of the stratum corneum. From our previous studies we know that cystatin M/E is expressed in hair follicles and probably is involved in the process of hair follicle morphogenesis. In chapter 5 we analyzed the localization of all key molecules of the novel biochemical pathway in the hair follicle, and also in the nail unit. Both are epidermal appendages that can be considered as mini‐organs within the skin as they possess their own stem cell compartments and also show processes of terminal differentiation. Interestingly, cathepsin L and TGM3 showed coordinately expressed co‐localization in the hair bulb and the nail matrix, the regions that provide cells that terminally differentiate into the hair fiber and the nail plate, respectively. Furthermore, co‐localization of cystatin M/E and cathepsin V was found in the inner root sheath of the hair follicle, and was restricted to the infundibulum of the hair follicle around the opening of the sebaceous duct into the hair canal. At this location degeneration of the inner root sheath occurs and the co‐expression of cystatin M/E and cathepsin V suggests that both proteins are involved in this process, possibly through desquamation of the inner root sheath cells. These findings suggest that cystatin M/E and its target proteases are not only involved in terminal differentiation of the epidermis, but also of the hair follicle and the nail unit. Since our previous studies have suggested an important role for cystatin M/E in correct epidermal differentiation, the expression of cystatin M/E, its target proteases, and TGM3 was investigated in inflammatory skin diseases with disturbed skin barrier formation (Chapter 6). mRNA expression levels of cystatin M/E and cathepsin V were found to be decreased in lesional skin of psoriatic and atopic dermatitis patients, whereas increased levels of cathepsin L and TGM3 were found. Decreased mRNA expression of cystatin M/E and cathepsin V was confirmed at the protein level, however, the increase of cathepsin L and TGM3 at the transcriptional level was not reflected by higher protein levels. Recently, our group has adapted an in vitro model of normal human reconstructed skin, in which skin constructs differentiate from 1‐2 basal layers of keratinocytes at day 0 into a completely stratified epidermis on day 12. As this reconstructed skin model could be a

137 valuable tool in future studies, it is important that reconstructed skin closely mimics normal human epidermis. Therefore, the expression patterns of several differentiation markers (involucrin, loricrin, filaggrin) as well as the proteins of the cystatin M/E pathway were studied at different time points during the development of reconstructed skin. The expression of all these proteins in reconstructed skin was qualitatively and quantitatively similar to the in vivo situation (Chapter 6). Our previous work has revealed that human cystatin M/E expression is restricted to skin, lung, and kidney, which was found by Northern blot analysis, reverse transcriptase PCR, and immunohistochemistry. However, real‐time quantitative PCR (qPCR) is currently the most accurate method for detection of quantitative gene expression levels. In chapter 6 we performed qPCR analysis on a large panel of human tissues, which revealed that cystatin M/E tissue distribution is less restricted than previously assumed. In addition to above mentioned tissues, human cystatin M/E expression was also observed in oropharyngeal tissues, esophagus, and aorta. The same panel of tissues was used for qPCR analysis of cathepsin L, cathepsin V, legumain, and TGM3 gene expression. Similar to human cystatin M/E, mouse cystatin M/E is not only expressed in epidermis, but also in various other tissues. Unfortunately, the neonatal lethal skin phenotype of cystatin M/E deficient mice makes it impossible to study the consequences of cystatin M/E deficiency in other tissues and organs in adult animals. In order to overcome this, we devised a strategy to rescue the lethal skin phenotype of cystatin M/E deficient mice by transgenic expression of cystatin M/E in epidermis (Chapter 7). A vector was generated in which the coding sequences of both mouse cystatin M/E and enhanced green fluorescent protein (eGFP) were placed under the control of the involucrin promoter. Expression of the transgene was validated in vitro by transient transfection of the vector into cultured human keratinocytes. Transfected keratinocytes showed eGFP and mouse cystatin M/E protein expression after stimulation to differentiate. As a next step the transgene was used for micro‐injection into mouse zygotes, which has resulted in several transgenic founder mice. The transgenic founder mice will be used for cross‐breeding with cystatin M/E deficient mice to generate “rescued” mice that express cystatin M/E specifically in the epidermis, but remain cystatin M/E deficient in other tissues.

GENERAL DISCUSSION AND FUTURE DIRECTIONS

Altogether, the data presented in this thesis suggest a mechanism in which cystatin M/E is a key molecule in a biochemical pathway that controls stratum corneum homeostasis by regulation of both cornification and desquamation of the stratum corneum (Figure 1). First, cystatin M/E could have a role in desquamation by regulation of cathepsin V protease activity, which is involved in degradation of corneodesmosomal components. Second, cystatin M/E could regulate crosslinking of structural proteins by TGM3 in the cornification process of the epidermis and the hair follicle by controlling cathepsin L and legumain activities. Furthermore, cystatin M/E might also be involved in the regulation of TGM1

138 activity as it is known that cathepsin L can process and activate cathepsin D1,2. It has been reported that cathepsin D, an aspartic protease, is involved in the regulation of TGM13,4. In vitro experiments have shown that TGM1 activity is stimulated by exogenous cathepsin D in cultured keratinocytes, and in cathepsin D deficient mice a reduced TGM1 enzymatic activity and defective TGM1 processing was observed. The regulatory role of cystatin M/E in the cornification process of the epidermis was further supported by the observation that in skin extracts of 6 to 12 day old cystatin M/E deficient mice relatively high levels of proteolyzed cathepsin D and TGM1 are found, whereas no appreciable levels of these proteolyzed proteins could be detected in skin extracts of wild type littermates (Zeeuwen et al. manuscript in preparation). Furthermore, increased levels of mainly membrane bound TGM activity (in situ) were observed in 6 to 12 day old cystatin M/E deficient mice, suggesting that increased TGM1 activity could be involved in the pathology of these mice, which is clearly a direction for future research. In addition, regulation of cathepsin D activity is also important in the desquamation process of the skin as the mature active form of cathepsin D is able to degrade desmosomal components5.

Figure 1: The role of cystatin M/E in epidermal differentiation. A simplified model of the presumed role of cystatin M/E in regulation of processes that control epidermal cornification and desquamation. Cystatin M/E is an inhibitor of legumain and the cathepsins L and V (CTSL, CTSV). Inhibition of legumain activity could regulate processing and activation of (pro)cathepsins, a process that is also under control of intra‐ and extracellular pH changes. Inhibition of CTSV activity could regulate desquamation as CTSV is able to degrade (corneo‐)desmosomal proteins such as desmoglein‐1 (DSG1), desmocollin‐1 (DSC1), and corneodesmosin (CDSN). Inhibition of CTSL activity could regulate cornification as CTSL processes and activates the crosslinking enzyme TGM3. CTSL is also able to process cathepsin D (CTSD), an aspartic protease that can regulate TGM1 activity. Reprinted with permission from: Zeeuwen, P.L., Cheng, T., Schalkwijk, J. (2009). J Invest Dermatol © 2009 The Society for Investigative Dermatology

139

Cystatin M/E and disease The past few years have taught us a great deal on the role of cystatin M/E and its cognate proteases in epidermal biology. Currently no human disease is known that is caused by genetic alterations of the proteins in this newly discovered pathway. Nevertheless, all unresolved heritable human skin disorders that show faulty cornification, desquamation, and/or hair follicle morphogenesis are candidates for cystatin M/E deficiency, and the identification of such a disorder remains a major challenge. As cathepsin V and cathepsin D are involved in breakdown of corneodesmosomal proteins, they could also be candidate genes for ichthyosiform skin diseases that show disturbed desquamation by retention of scales. In addition to excessive TGM activity due to cystatin M/E deficiency, as seen in ichq mice, lack of TGM activity can also lead to disturbed cornification as witnessed by the autosomal recessive lamellar ichthyosis in which mutations in TGM1 are responsible for the disease6. Although TGM3 accounts for more than 75% of the total TGM activity in the epidermis7, it cannot compensate for cornified envelope formation in patients with lamellar ichthyosis in which TGM1 activity is absent8, probably because crosslinking by TGM1 precedes crosslinking by TGM3 according to the model for assembly of the cornified envelope (described in Chapter 1). Although this might seem paradoxical at first glance, both loss and inappropriate gain of TGM activity could explain ichthyotic changes, although the mechanisms could be different. Excessive TGM1 and TGM3 activity could lead to retention of scales by hyper‐crosslinked corneocytes, whereas deficiency for TGM1 could lead to irregular scaling due to lack of attachment of involucrin to the ω‐hydroxyceramides. There are, however, other forms that are clinically similar to lamellar ichthyosis but are not linked to the locus of the TGM1 gene. The redundancy of TGM3 is indicated by knockout mice that show no particular skin phenotype9. This leaves open the possibility that mutations in genes that are involved in the processing of TGMs into the active form might be causative for other forms of ichthyosis. In conclusion, we suggest that cystatin M/E is a candidate gene for heritable human skin disorders that show faulty cornification and desquamation.

Epigenetic regulation of epidermal cystatin M/E expression Recent epigenetic studies have provided evidence that regulation of cystatin M/E gene expression is regulated by methylation of the CpG islands present in the cystatin M/E proximal promoter region and exon 110‐13. Methylation of CpG islands is a postreplicative modification thought to play a role in gene transcription in eukaryotes, either through interference with transcription factor binding or through the recruitment of repressors that specifically bind sites containing methylated CpG dinucleotides. A growing number of human diseases was found to be associated with aberrant DNA methylation14,15. It was reported that cystatin M/E expression is frequently silenced in breast carcinomas by epigenetic inactivation related to DNA‐methylation (see Chapter 1 for details). Cystatin M/E is expressed in the differentiating layers of the epidermis, but absent in the basal layer, suggesting that during epidermal differentiation cystatin M/E is transcriptionally activated.

140 Alternatively, cystatin M/E might be epigenetically silenced in basal keratinocytes by DNA methylation of the promoter region. Initiation of epidermal differentiation could trigger demethylation of the cystatin M/E promoter and result in transcriptional activity of the cystatin M/E gene. Investigating the degree of methylation of the CpG island in the cystatin M/E promoter in non‐differentiated vs. differentiated keratinocytes might provide clues for epigenetic regulation of epidermal cystatin M/E expression. For these experiments keratinocytes could be used from in vitro cell cultures in which primary keratinocytes are stimulated to differentiate, or ex vivo from biopsies of normal human skin. Keratinocytes can be sorted by flowcytometry based on their expression of a specific differentiation marker, e.g. keratin 10. It would be interesting to investigate if the cystatin M/E promoter is methylated in non‐differentiated keratinocytes and demethylated in differentiated keratinocytes. This could indicate if epidermal expression of cystatin M/E is epigenetically regulated. Epigenetic changes in DNA and chromatin have been implied in ageing and longevity16. Studies have revealed that genome‐wide levels of DNA methylation, usually of repetitive sequences, decrease with age, while on the other hand hypermethylation of CpG islands in gene promoters is increased17,18. Interestingly, ageing skin shows thinning of the epidermis, dehydration (dry skin), and susceptibility to pruritus (itching) and infection19, which implies impaired skin barrier formation and function. Since the cystatin M/E gene is prone to promoter hypermethylation, small changes in DNA methylation that accumulate with age might ultimately lead to decreased cystatin M/E expression, which could result in the above mentioned symptoms of aged skin. Determination of the degree of methylation of the cystatin M/E promoter in keratinocytes from children vs. elderly could provide support for this hypothesis.

In vivo and in vitro models The cystatin M/E deficient mouse model has provided more insight in the physiological processes that are involved in epidermal differentiation, and has led to the discovery of a novel biochemical pathway in stratum corneum homeostasis that we have presented in this thesis. Fortunately, we were able to obtain mice deficient for cathepsin L, legumain or TGM3, which have been previously generated by other groups9,20,21. Generation of double knock‐out mice that are deficient for cystatin M/E and either cathepsin L, legumain or TGM3 might provide in vivo evidence for the role of cystatin M/E in epidermal differentiation. These studies are in their final stages, and preliminary data have already revealed that cathepsin L‐cystatin M/E double knockout mice are viable, suggesting that the balance between the protease cathepsin L and its inhibitor cystatin M/E is highly important in epidermal barrier formation and function (Zeeuwen et al., manuscript in preparation). One disadvantage of using mouse models is that results cannot automatically be extrapolated to humans. Recently, our group has adapted an in vitro model of normal human reconstructed skin, in which adult abdominal keratinocytes are seeded on de‐ epidermized dermis and cultured into a complete stratified epidermis22. This reconstructed skin model could be a valuable tool to study the novel biochemical pathway in human tissue.

141 For this purpose it is important that in vitro human reconstructed skin closely mimics normal human epidermis. Our study, as presented in Chapter 6, on the expression patterns of several differentiation markers and the key molecules of the novel biochemical pathway has shown that the expression of all these molecules in reconstructed skin was qualitatively and quantitatively similar to the in vivo situation, suggesting that human reconstructed skin could be used to study the cystatin M/E pathway in vitro. Gene knockdown or overexpression of certain genes in human reconstructed skin could be achieved by lentiviral delivery of constructs that drive expression of siRNA or genes under stratum specific promoters in keratinocytes. Lentiviral transduction results in the incorporation of the transgene into the genomic DNA of keratinocytes. Knockdown of cystatin M/E in reconstructed skin could lead to impaired epidermal differentiation resulting in a phenotype characterized by hyperkeratosis, stratum corneum abnormalities, and excess protease activity, as observed in ichq mice. Subsequently, co‐transduction of lentiviral constructs targeted to gene knockdown of cystatin M/E and either cathepsin L, cathepsin V, legumain or TGMs, could be performed to reverse the abnormal phenotype. Alternatively, the use of cell‐permeable protease inhibitors could also restore epidermal barrier formation and function in reconstructed cystatin M/E deficient skin. Furthermore, overexpression of TGM1 or TGM3 could possibly also result in a phenotype as described above. These experiments will provide more insight in the role of the cystatin M/E pathway in human epidermis. In view of the fact that expression of cystatin M/E is not restricted to epidermis, cystatin M/E could be involved in physiologic and pathophysiologic processes in other tissues and organs. Disturbed cystatin M/E expression or function could speculatively result in various pathologies: abnormal keratinization in hair follicles and nails, impaired barrier formation and function of oropharyngeal epithelia, altered lung function, cardiovascular disease, renal dysfunction, and corneal abnormalities. The generation of cystatin M/E deficient mice with transgenic cystatin M/E expression in epidermis could be a first step towards the elucidation of cystatin M/E function in other tissues and organs besides epidermis. This transgenic mouse will be “rescued” from the neonatal lethal skin phenotype and offers opportunities to study the effects of cystatin M/E deficiency in later stages of life. Since cystatin M/E could also act as tumor suppressor, it would also be interesting to investigate the effects of cystatin M/E deficiency on tumorigenesis, tumor growth, and metastasis.

CONCLUDING REMARKS

Epidermal differentiation is an intricate process that involves structural, enzymatic, and signaling functions of a variety of proteins. Elucidation of the molecular processes that underlie epidermal differentiation is of great importance for understanding skin barrier formation and function. The studies in this thesis have revealed the role of the cysteine protease inhibitor cystatin M/E as a key molecule in a biochemical pathway that controls stratum corneum homeostasis by regulation of processes in cornification and desquamation.

142 In a more general context, the discovery of this novel biochemical pathway illustrates the importance of balance between proteases and their inhibitors in physiologic processes, since a shift to either side of the scale could have dramatic effects and may result in pathology.

REFERENCES

1. Wille, A., Gerber, A., Heimburg, A., Reisenauer, A., Peters, C., Saftig, P., Reinheckel, T., Welte, T., Buhling, F. (2004) Cathepsin L is involved in cathepsin D processing and regulation of apoptosis in A549 human lung epithelial cells. Biol.Chem. 385, 665‐670 2. Laurent‐Matha, V., Derocq, D., Prebois, C., Katunuma, N., Liaudet‐Coopman, E. (2006) Processing of human cathepsin D is independent of its catalytic function and auto‐activation: involvement of cathepsins L and B. J.Biochem. 139, 363‐371 3. Egberts, F., Heinrich, M., Jensen, J. M., Winoto‐Morbach, S., Pfeiffer, S., Wickel, M., Schunck, M., Steude, J., Saftig, P., Proksch, E., Schutze, S. (2004) Cathepsin D is involved in the regulation of transglutaminase 1 and epidermal differentiation. J.Cell Sci. 117, 2295‐2307 4. Higuchi, K., Kawashima, M., Takagi, Y., Kondo, H., Yada, Y., Ichikawa, Y., Imokawa, G. (2001) Sphingosylphosphorylcholine is an activator of transglutaminase activity in human keratinocytes. J.Lipid Res. 42, 1562‐1570 5. Horikoshi, T., Igarashi, S., Uchiwa, H., Brysk, H., Brysk, M. M. (1999) Role of endogenous cathepsin D‐like and chymotrypsin‐like proteolysis in human epidermal desquamation. Br.J.Dermatol. 141, 453‐459 6. Huber, M., Rettler, I., Bernasconi, K., Frenk, E., Lavrijsen, S. P., Ponec, M., Bon, A., Lautenschlager, S., Schorderet, D. F., Hohl, D. (1995) Mutations of keratinocyte transglutaminase in lamellar ichthyosis. Science 267, 525‐528 7. Kim, H. C., Lewis, M. S., Gorman, J. J., Park, S. C., Girard, J. E., Folk, J. E., Chung, S. I. (1990) Protransglutaminase E from guinea pig skin. Isolation and partial characterization. J.Biol.Chem. 265, 21971‐21978 8. Candi, E., Melino, G., Lahm, A., Ceci, R., Rossi, A., Kim, I. G., Ciani, B., Steinert, P. M. (1998) Transglutaminase 1 mutations in lamellar ichthyosis. Loss of activity due to failure of activation by proteolytic processing. J.Biol.Chem. 273, 13693‐13702 9. John, S. (2006) Die Inaktivierung des TGM3 Gens in der Maus und ihre Auswirkung auf die Haarmorphogenese. Thesis. University of Cologne, Germany. 10. Ai, L., Kim, W. J., Kim, T. Y., Fields, C. R., Massoll, N. A., Robertson, K. D., Brown, K. D. (2006) Epigenetic silencing of the tumor suppressor cystatin M occurs during breast cancer progression. Cancer Res. 66, 7899‐7909 11. Rivenbark, A. G., Jones, W. D., Coleman, W. B. (2006) DNA methylation‐dependent silencing of CST6 in human breast cancer cell lines. Lab Invest 86, 1233‐1242 12. Schagdarsurengin, U., Pfeifer, G. P., Dammann, R. (2007) Frequent epigenetic inactivation of cystatin M in breast carcinoma. Oncogene 26, 3089‐3094 13. Qiu, J., Ai, L., Ramachandran, C., Yao, B., Gopalakrishnan, S., Fields, C. R., Delmas, A. L., Dyer, L. M., Melnick, S. J., Yachnis, A. T., Schwartz, P. H., Fine, H. A., Brown, K. D., Robertson, K. D. (2008) Invasion suppressor cystatin E/M (CST6): high‐level cell type‐specific expression in normal brain and epigenetic silencing in gliomas. Lab Invest 88, 910‐925 14. Robertson, K. D. (2005) DNA methylation and human disease. Nat.Rev.Genet. 6, 597‐610

143 15. Feinberg, A. P. (2007) Phenotypic plasticity and the epigenetics of human disease. Nature 447, 433‐440 16. Fraga, M. F., Esteller, M. (2007) Epigenetics and aging: the targets and the marks. Trends Genet. 23, 413‐418 17. Issa, J. P. (2003) Age‐related epigenetic changes and the immune system. Clin.Immunol. 109, 103‐108 18. So, K., Tamura, G., Honda, T., Homma, N., Waki, T., Togawa, N., Nishizuka, S., Motoyama, T. (2006) Multiple tumor suppressor genes are increasingly methylated with age in non‐ neoplastic gastric epithelia. Cancer Sci. 97, 1155‐1158 19. Farage, M. A., Miller, K. W., Elsner, P., Maibach, H. I. (2008) Intrinsic and extrinsic factors in skin ageing: a review. Int.J.Cosmet.Sci. 30, 87‐95 20. Roth, W., Deussing, J., Botchkarev, V. A., Pauly‐Evers, M., Saftig, P., Hafner, A., Schmidt, P., Schmahl, W., Scherer, J., Anton‐Lamprecht, I., Von Figura, K., Paus, R., Peters, C. (2000) Cathepsin L deficiency as molecular defect of furless: hyperproliferation of keratinocytes and pertubation of hair follicle cycling. FASEB J. 14, 2075‐2086 21. Shirahama‐Noda, K., Yamamoto, A., Sugihara, K., Hashimoto, N., Asano, M., Nishimura, M., Hara‐Nishimura, I. (2003) Biosynthetic processing of cathepsins and lysosomal degradation are abolished in asparaginyl endopeptidase‐deficient mice. J.Biol.Chem. 278, 33194‐33199 22. Tjabringa, G., Bergers, M., van Rens, D., de Boer, R., Lamme, E., Schalkwijk, J. (2008) Development and validation of human psoriatic skin equivalents. Am.J.Pathol. 173, 815‐823

144

Nederlandse samenvatting

Tsing Cheng

145

146 Nederlandse samenvatting

De huid is het grootste orgaan van het menselijk lichaam en heeft een belangrijke functie, aangezien het de barrière vormt die de dieper gelegen weefsels bescherming biedt tegen verwonding, micro‐organismen, schadelijke stoffen en uitdroging. Tevens vervult de huid een esthetische functie die van belang is in sociale en psychologische context. Huid bestaat hoofdzakelijk uit twee componenten: de epidermis, die zichzelf constant vernieuwt en het stratum corneum (hoornlaag) vormt; daaronder bevindt zich de dermis, bestaande uit bindweefsel, bloedvaten en zenuwen. Een juiste ontwikkeling van de epidermis is van essentieel belang voor een goed functionerende huidbarrière en een verstoorde vorming van het stratum corneum kan tot ernstige huidziekten leiden. Terminale differentiatie van de epidermis is een complex proces van verhoorning en afschilfering van de keratinocyten (huidcellen) dat beschreven wordt in hoofdstuk 1. Bestudering van de moleculaire processen die betrokken zijn bij epidermale differentiatie vergroot onze kennis over de vorming en functie van de huidbarrière en is van belang voor de ontwikkeling van nieuwe therapieën voor huidziekten. Vele eiwitten met diverse functies blijken betrokken te zijn bij epidermale differentiatie waaronder proteases en hun remmers. Proteases zijn enzymen die andere eiwitten in stukken kunnen knippen door interne verbindingen te verbreken via hydrolyse. Eiwitten kunnen door proteases zowel worden afgebroken als geactiveerd. De classificatie, specifieke kenmerken en functies van de verschillende proteases in het algemeen worden in hoofdstuk 1 besproken, waarna de aandacht vervolgens wordt gevestigd op lysosomale cysteine proteases. De maturatie en het intracellulaire transport van de cathepsines wordt behandeld en tevens worden specifiek cathepsine L, cathepsine V en legumaine besproken. Het onderzoek beschreven in dit proefschrift richt zich op cystatine M/E, een remmer van cysteine proteases. Hoofdstuk 1 bevat een samenvatting van alle voorgaande studies die betrekking hebben op cystatine M/E, waarin onder meer aandacht wordt besteed aan de cystatine M/E deficiënte muis en de rol van cystatine M/E in de ontwikkeling van de huid, de rol van cystatine M/E als tumor suppressor molecuul in tumorbiologie en de epigenetische regulatie van cystatine M/E expressie. Omdat cystatine M/E sterk tot expressie komt in de epidermis en cystatine M/E deficiënte muizen een letaal huidfenotype vertonen waarin de huidbarrière verstoord is, zal cystatine M/E waarschijnlijk een belangrijke rol hebben in de ontwikkeling van de epidermis. Echter, in welke fysiologische processen cystatine M/E een functie vervult tijdens epidermale differentiatie, is vooralsnog onduidelijk. Het doel van dit promotieonderzoek was hierin meer duidelijkheid te verkrijgen. Eerder onderzoek heeft aangetoond dat cystatine M/E een remmer is van de lysosomale asparaginyl endopeptidase legumaine, wiens activiteit in de epidermis en haarfollikel door cystatine M/E gereguleerd zou kunnen worden. Cystatines zijn echter hoofdzakelijk remmers van lysosomale cathepsines die tot een andere klasse behoren dan legumaine. Welke cathepsine nu precies door cystatine M/E wordt geremd was nog onduidelijk aan het begin van dit promotieonderzoek. De identificatie van deze cathepsine

147 zou een belangrijke bevinding zijn om de rol van cystatine M/E in epidermale differentiatie op te kunnen helderen. Hoofdstuk 2 beschrijft de identificatie van twee cathepsines die door cystatine M/E geremd kunnen worden. Door middel van biochemische experimenten werd aangetoond dat in vitro cystatine M/E een hoge affiniteit heeft voor de cathepsines L en V en dat de activiteit van deze proteases geremd wordt. De dissociatie constanten (Ki) van de protease‐protease remmer complexen waren laag, namelijk tussen de 0.5 en 2 nM. Door middel van recombinante varianten van cystatine M/E met mutaties in de bindingsdomeinen is aangetoond dat cystatine M/E twee verschillende actieve domeinen bevat: één voor de remming van de cathepsines en één voor de remming van legumaine. Vervolgens werd voor de biologische relevantie van deze resultaten de lokalisatie van cystatine M/E, cathepsine L, cathepsine V en legumaine in normale humane epidermis bestudeerd met behulp van immunohistochemie. Cystatine M/E en cathepsine V werden aangetroffen in de bovenste lagen van het stratum granulosum en het stratum corneum, terwijl expressie van cathepsine L en legumaine in de gehele epidermis met uitzondering van het stratum corneum werd gevonden. Eén van de bevindingen in de cystatine M/E deficiënte muis was voortijdige en verhoogde activatie van transglutaminase 3 (TGM3) in de epidermis. Transglutaminases zijn enzymen die structurele eiwitten aan elkaar koppelen tijdens de verhoorning van de keratinocyten. Inactief TGM3 met een grootte van 77 kDa wordt door een tot dan toe onbekende protease geknipt in de actieve vorm bestaande uit 30 en 47 kDa fragmenten. Hoofdstuk 2 beschrijft de betrokkenheid van cathepsine L bij de activatie van TGM3. Uit experimenten met verschillende cathepsines (B, D, L, S en V) bleek dat alleen cathepsine L in staat was om inactief TGM3 te knippen in de 30 en 47 kDa fragmenten, die daarna ook daadwerkelijk biologische activiteit vertoonde. Met behulp van N‐terminale sequencing werd de knipplaats van cathepsine L in TGM3 bepaald op Ala466. In hoofdstuk 3 wordt de co‐lokalisatie van cystatine M/E met cathepsine L, cathepsine V en legumaine alsmede de cellulaire compartmentalisatie van deze eiwitten bestudeerd. Met behulp van immunofluorescentie microscopie werd co‐lokalisatie van cystatine M/E met cathepsine L en legumaine aangetoond in het stratum granulosum, terwijl co‐lokalisatie van cystatine M/E met cathepsine V gevonden werd in de bovenste lagen van het stratum granulosum en het stratum corneum. De compartmentalisatie van deze vier eiwitten is vervolgens bestudeerd met behulp van immuno‐elektronen microscopie. Cystatine M/E en cathepsine V bleken apart van elkaar door de cel getransporteerd te worden in lamellaire granules. Vervolgens kunnen beide eiwitten worden uitgescheiden in de extracellulaire ruimte van het stratum corneum, waar zij lokaliseren rondom corneodesmosomen, de verbindingen tussen de verhoornde keratinocyten die de cellen bij elkaar houden. Cathepsine L en legumaine werden niet gevonden in lamellaire granules, maar in het cytoplasma, wat opmerkelijk was aangezien beide eiwitten beschouwd worden als lysosomale proteases. Echter, lysosomale proteases blijken ook betrokken te zijn in processen die zich buiten de lysosomen afspelen, zoals apoptose, een proces dat enige gelijkenissen vertoont met terminale differentiatie van de epidermis.

148 De co‐lokalisatie van cystatine M/E en cathepsine V rondom de corneodesmosomen leidde tot de suggestie dat beide eiwitten mogelijk een rol spelen in de afschilfering van de dode huidcellen van het stratum corneum. Voor afschilfering is het namelijk van belang dat de corneodesmosomen afgebroken worden en dit gebeurt door de proteolyische activiteit van verschillende uitgescheiden proteases. Daarom werd het vermogen van cathepsine V om in vitro de (corneo‐)desmosomale eiwitten desmogleine‐1, desmogleine‐3 en corneodesmosine af te breken bestudeerd in hoofdstuk 4. Cathepsine V bleek deze eiwitten snel te kunnen afbreken bij een pH van 5,5, welke ook heerst in de bovenste lagen van het stratum corneum. Bij een neutrale pH van 7,0 was er nauwelijks afbraak van de corneodesmosomale eiwitten door cathepsine V. Deze resultaten passen in de theorie dat de afschilfering van de huid een pH‐afhankelijk proces is, waarin de geleidelijke pH‐ verandering in het stratum corneum, van neutraal in de onderste lagen naar zuur in de bovenste lagen, bepaalt welke proteases hoofdzakelijk actief zijn in de verschillende lagen van het stratum corneum. Het is bekend dat lysosomale proteases elkaar kunnen knippen en activeren en uit studies met legumaine deficiënte muizen is gebleken dat legumaine dit kan doen met cathepsines. In hoofdstuk 4 is onderzocht of dit ook voor humaan legumaine, cathepsine L en cathepsine V geldt. Legumaine bleek in staat te zijn om “single‐chain” cathepsine V te knippen in een tot nu toe onbekende actieve “two‐chain” vorm van cathepsine V. Deze nog niet bekende vorm van cathepsine V is vervolgens door middel van immunoblotting aangetoond in humane epidermale eiwitextracten, wat aangeeft dat dit een natuurlijk voorkomende vorm is. Met behulp van N‐terminale sequencing en MALDI‐MS analyse werd de knipplaats van legumaine in cathepsine V bepaald op Asn290. Ondanks dat cathepsine V en cathepsine L zeer verwant aan elkaar zijn, bleek legumaine “single‐chain” cathepsine L niet naar de “two‐chain” vorm te kunnen knippen, waarschijnlijk omdat cathepsine L de knipplaats op Asp291 heeft en legumaine specifiek knipt na een Asn residu. Verder onderzoek leerde dat zowel cathepsine L als cathepsine V niet in staat zijn om legumaine te knippen of te activeren. Eerder onderzoek heeft aangetoond dat cystatine M/E ook tot expressie komt in haarfollikels en waarschijnlijk betrokken is bij haarfollikel morfogenese. In hoofdstuk 5 wordt de (co‐)lokalisatie van cystatine M/E, cathepsine L, cathepsine V, legumaine en TGM3 in zowel de haarfollikel als de nagel bestudeerd met behulp van immunofluorescentie microscopie. De haarfollikel en de nagel zijn een soort mini‐organen in de huid met hun eigen stamcellen en zij vertonen tevens terminale differentiatie. Co‐lokalisatie van cathepsine L en TGM3 werd aangetoond in de haar bulbus en de nagel matrix, de locaties die de terminaal differentiërende cellen leveren die uiteindelijk de haar en de nagelplaat vormen. Tevens wordt in deze studie de co‐lokalisatie van CTSL en TGM3 in het stratum granulosum van de epidermis aangetoond. De dubbelkleuring laat een geleidelijke transitie van cathepsine L expressie naar TGM3 expressie zien, wat duidt op een gecoördineerde expressie van de activerende protease (cathepsine L) en het substraat (TGM3). Verder werd co‐lokalisatie van cystatine M/E en cathepsine V aangetoond in de binnenste haarwortelschede van de haarfollikel, specifiek ter hoogte van het infundibulum nabij de

149 monding van de talgklier in het haarkanaal. Op deze plek vindt degeneratie van de binnenste haarwortelschede plaats en de co‐lokalisatie van cystatine M/E en cathepsine V wekt de suggestie dat beide eiwitten een rol spelen in dit proces, mogelijk in de afschilfering van de cellen van de binnenste haarwortelschede. Deze studie suggereert dat cystatine M/E en de cathepsines L en V niet alleen betrokken zijn in terminale differentiatie van de epidermis, maar ook in die van de haarfollikel en de nagel. Hoofdstuk 6 behandelt studies waarin gebruik is gemaakt van kwantitatieve meetmethoden om de gen‐ en eiwitexpressie van cystatine M/E, cathepsine L, cathepsine V, legumaine en TGM3 te bestuderen. Allereerst worden de resultaten over de genexpressie van cystatine M/E, cathepsine L, cathepsine V, legumaine en TGM3, verkregen met real‐time quantitative PCR (qPCR), in een groot aantal verschillende humane weefsels beschreven. Soortgelijke studies zijn al eens eerder uitgevoerd met behulp van Northern blot en reverse‐ transcriptase PCR, maar qPCR analyse is op dit moment de nauwkeurigste methode voor het bepalen van kwantitatieve genexpressie. In huid, long en nier is expressie van cystatine M/E al eerder aangetoond; in de qPCR analyse werd naast de eerder genoemde weefsels ook expressie van cystatine M/E aangetroffen in oropharyngeale weefsels, slokdarm en aorta. Verder is in hoofdstuk 6 de expressie van cystatine M/E, cathepsine L, cathepsine V, legumaine en TGM3 in psoriasis en atopisch eczeem onderzocht. Psoriasis en atopisch eczeem zijn twee veel voorkomende ontstekingsziekten van de huid met een verstoorde huidbarrière. De kwantitatieve expressie werd bepaald op zowel mRNA als eiwitniveau met behulp van respectievelijk qPCR en kwantitatieve analyse van immunofluorescente kleuringen. In lesionale huid van patiënten met psoriasis en atopisch eczeem bleek, vergeleken met gezonde vrijwilligers, de mRNA expressie van cystatine M/E en cathepsine V te zijn verminderd, terwijl de mRNA expressie van cathepsine L en TGM3 verhoogd was. De verminderde genexpressie van cystatine M/E en cathepsine V komt overeen met verminderde eiwitexpressie. Echter, de verhoogde mRNA expressie van cathepsine L en TGM3 kon niet worden bevestigd op eiwitniveau. Hoofdstuk 6 beschrijft tevens de expressie en lokalisatie van cystatine M/E, cathepsine L, cathepsine V, legumaine en TGM3 alsmede enkele differentiatie markers (involucrine, loricrine en filaggrine) in een in vitro kweekmodel van humane gereconstrueerde normale huid. Uit analyses van deze gereconstrueerde huid bleek dat het expressiepatroon van alle onderzochte eiwitten in de humane gereconstrueerde huid zowel kwalitatief als kwantitatief overeen kwam met humane normale huid. Hieruit kan worden verondersteld dat dit in vitro huidmodel in de toekomst kan worden gebruikt om de rol van cystatine M/E in epidermale differentiatie te onderzoeken. Net als humaan cystatine M/E komt muis cystatine M/E niet alleen tot expressie in de huid, maar ook in diverse andere weefsels. Helaas zorgt het letale huidfenotype van de cystatine M/E deficiënte muis ervoor dat deze muis kort na de geboorte overlijdt, waardoor het niet mogelijk is om het effect van cystatine M/E deficiëntie in andere (volwassen) weefsels te bestuderen. Om dit toch mogelijk te maken wordt in hoofdstuk 7 de ontwikkeling van een transgeen muismodel besproken. In dit muismodel wordt geprobeerd

150 het letale huidfenotype van cystatine M/E deficiënte muizen te voorkomen door middel van transgene expressie van cystatine M/E in de epidermis. Hiervoor is een vector gecreëerd waarin de coderende DNA sequenties voor muis cystatine M/E en enhanced green fluorescent protein (eGFP) onder de controle van de involucrine promoter zijn geplaatst. Expressie van het transgen werd in vitro getest door gekweekte humane keratinocyten te transfecteren met de vector. In getransfecteerde keratinocyten die gestimuleerd werden om te differentiëren werd zowel eGFP als muis cystatine M/E eiwit tot expressie gebracht. Vervolgens zijn enkele transgene “founder” muizen verkregen door middel van micro‐ injectie van het transgen in muis zygoten. Deze transgene founder muizen zullen gekruist worden met cystatine M/E deficiënte muizen waaruit uiteindelijk levensvatbare cystatine M/E deficiënte muizen worden geboren met transgene expressie van cystatine M/E in de epidermis en met cystatine M/E deficiëntie in de overige weefsels. De resultaten van het onderzoek beschreven in dit proefschrift hebben geleid tot een model dat de rol van cystatine M/E in epidermale differentiatie beschrijft, zoals dat gepresenteerd wordt in hoofdstuk 8. Tijdens epidermale differentiatie heeft cystatine M/E een sleutelrol in een nieuwe biochemische route die de homeostase van het stratum corneum beïnvloedt door zowel de verhoorning en de afschilfering van het stratum corneum te reguleren. Cystatine M/E is betrokken bij de afschilfering door de activiteit van cathepsine V te controleren, een protease die in staat is om corneodesmosomale eiwitten af te breken. Cystatine M/E is betrokken bij de verhoorning doordat het, via controle over cathepsine L activiteit, de activatie van TGM3 en daarmee het aan elkaar koppelen van structurele eiwitten door TGM3 reguleert. Eventueel zou cystatine M/E ook de activiteit van TGM1 kunnen reguleren, aangezien cathepsine L tevens cathepsine D kan activeren, een protease dat weer betrokken is bij het knippen en activeren van TGM1. Tot slot wordt in hoofdstuk 8 vooruit gekeken hoe de rol van cystatine M/E in de nieuwe biochemische route verder bestudeerd kan worden. Tot nu toe zijn er geen huidziekten bekend waarvan de oorzaak ligt in genetische mutaties van de eiwitten die onderdeel zijn van deze biochemische route. Het cystatine M/E gen lijkt een goede kandidaat voor mutatie‐analyse in erfelijke huidaandoeningen met verstoorde verhoorning en afschilfering. Verder zou bestudering van DNA‐methylatie in de cystatine M/E gen promoter in ongedifferentieerde en gedifferentieerde keratinocyten kunnen aantonen of epidermale expressie van cystatine M/E epigenetisch gereguleerd wordt. Dubbele “knock‐out” muismodellen worden momenteel ontwikkeld om de hierboven genoemde biochemische route die leidt tot epidermale differentiatie, in vivo te bestuderen. Tevens zou de nieuwe biochemische route onderzocht kunnen worden in humaan weefsel door middel van lentivirale “knock‐down” of overexpressie van genen in in vitro humane gereconstrueerde huid. De ontwikkeling van de cystatine M/E deficiënte muis met transgene expressie van cystatine M/E in de epidermis biedt mogelijkheden om de functie van cystatine M/E in andere weefsels dan de huid te bestuderen, evenals de effecten van cystatine M/E deficiëntie op de ontwikkeling van tumoren en metastase.

151

152

Dankwoord Curriculum Vitae List of Publications Color illustrations

Tsing Cheng

153 Dankwoord

Hier ligt het dan voor je neus: het boek waar ik de afgelopen jaren voor heb gewerkt. Waarschijnlijk ben je gelijk doorgebladerd naar dit dankwoord, maar als je het straks een keertje hebt door genomen, dan heb je enig idee van wat ik op het lab heb uitgevoerd. Of het is alleen maar ingewikkelder en onbegrijpelijker geworden, maar dat is niet erg. Dit proefschrift en de promotie zijn natuurlijk niet tot stand gekomen zonder de hulp en steun van anderen, onderzoek doe je tenslotte niet alleen. Daarom wil ik hier mijn dank betuigen aan de vele mensen die op de een of andere manier een positieve bijdrage hebben geleverd aan mijn promotietijd in Nijmegen.

Allereerst wil ik mijn copromotor Patrick bedanken voor zijn begeleiding, inzet, advies en steun bij dit promotieonderzoek. Dit proefschrift borduurt voort op het onderzoek dat jij hebt uitgevoerd voor jouw promotie; het was dan ook een absolute meerwaarde dat jij als post‐doc, en nu als junior PI, aan het cystatine M/E onderzoek kon blijven meewerken. Je bent zeer georganiseerd, nauwkeurig en kritisch, wat vaak wel van pas kwam, en ja, af en toe kun je behoorlijk bits zijn, maar vaak ook wel terecht. Met jou zo naast de deur en lab‐ tafel kon ik altijd bij je terecht met vragen en voor feedback. Ook was je zeer behulpzaam bij experimenten, artikelen, figuren en het proefschrift. Ik waardeer het zeer en ben je dankbaar voor al het werk dat je hebt gedaan. Daarnaast ben je sociaal ook een prima kerel en kunnen we goed met elkaar overweg; het was altijd erg gezellig tijdens borrels, congressen, stapavonden, etc., en onze gedeelde waardering (samen met Joost) voor goede televisieseries zoals Six Feet Under zal ik ook niet vergeten. Ik heb het toch maar getroffen met een begeleider zoals jij. Natuurlijk wil ik ook Joost, mijn promotor, bedanken voor de begeleiding en advies over mijn onderzoek, carrière en subsidieaanvragen. In het bijzonder wil ik je bedanken voor dat ene mailtje dat je mij stuurde, waarin je mij wees op een vacature die jullie hadden en mij vroeg te solliciteren als ik interesse had. Ik had die vacature zelf nooit opgemerkt, zo goed verstopt was hij, maar hij sprak me wel ontzettend aan. Ik ben blij dat Patrick en jij uiteindelijk mij boven de andere kandidaat (met de twee enorme pluspunten) hebben verkozen. Mijn eerste indruk van jou was: een normale, vriendelijke en relaxte professor, daar voel ik mij wel prettig bij. Vriendelijk en relaxed ben je zeker, maar normaal? Dat weet ik nog niet zo helemaal als ik soms je verhalen hoor, maar ik vind het heel prettig dat jij mijn promotor bent. Verder wil ik hier Ivonne bedanken voor de enorme bijdrage die zij heeft geleverd aan dit proefschrift. Bedankt voor al het histologische werk dat jij hebt verricht, dankzij jou zit dit boek vol met prachtige foto’s, waarvan er twee zelfs de covers van de JID sieren!

154 Het lab zou het lab niet zijn zonder al die leuke mensen die er werken en voor de gezellige sfeer zorgen, zowel binnen als buiten het lab. Een leuke werksfeer draagt immers bij aan een goed resultaat. En natuurlijk stonden ze ook klaar voor advies en hulp bij experimenten. Piet wil ik bedanken voor de flowcytometrie, zijn input en alle hulp bij vragen en problemen over computers en programma’s. Graag wil ik mijn (oud‐) kamergenoten bedanken voor de gezelligheid en de goede gesprekken: Gys (geniet van je pensioen), John (de man van de gadgets met o.a. een slaapcyclus lamp), Roelie, Jeroen en Heleen. De man van de foute muziek en slechte woordgrappen, PJ, verdient een aparte vermelding. Jij hebt mij in Parijs op slinkse wijze aan het hardlopen gekregen en uiteindelijk zijn we zelfs Refter‐ buddies geworden! Bedankt voor de vele leuke momenten en mijn nieuwe hobby, en veel succes met je eigen promotie. Dan zijn er nog de andere mede‐promovendi, allemaal met een verschillende achtergrond, maar uiteindelijk met hetzelfde doel (en dezelfde stress): Manon Zw. (bedankt dat ik je oude kamer kon overnemen), David (en zijn pesterijtjes), Judith (altijd giechelig en vrolijk), Marijke (respect voor deze drukke mama), Rosanne, Rieke, Jorn, Maartje, Annechien, Tim, Marloes, Haike en Michelle. Natuurlijk ook een bedankje aan de rest van het lab: Diana (altijd fijn om mee te roddelen), Mieke, Sandra, Desiree en Corien. Mijn studenten Christiaan en Mohammed wil ik bedanken voor het werk dat zij verricht hebben onder deze strenge begeleider, succes met jullie verdere opleiding en carrière. I would like to thank all the co‐authors abroad: Kiyotaka Hitomi, Thomas Reinheckel, Colin Watts, Akemi Ishida‐Yamamoto, Cécile Caubet, and Michel Simon, for their support and contribution to this research project. Graag wil ik Wiljan Hendriks bedanken voor zijn advies en hulp. Connie Verbaas, Maikel School, Iris Lamers‐Elemans, Claudia Lagarde en Debby Smits van het CDL: helaas is er nog geen “rescued” muis, maar bedankt voor alle moeite. A big ‘thank you’ goes out to Angela Christiano for offering me the opportunity to work at her lab and continue my academic career in skin biology. Alle overige mensen bij Dermatologie (staf, secretariaat, assistenten, administratie, verpleging en fotografie) wil ik bedanken voor hun interesse en gezelligheid tijdens pauzes en dagjes uit. Speciale bedankjes gaan uit naar Diny, Jan B., Manon Z.‐F., Wynand en Bas.

En dan missen we nog één persoon bij Dermatologie die ik nog niet bij naam genoemd heb, maar die vermelding zeker waard is: Jeanette, mijn maatje op de afdeling. Vanaf het begin, die eerste keer met de junior onderzoekers naar de Matrixx, klikte het tussen ons en hebben we een goede vriendschap opgebouwd. Jij moest dan ook wel één van mijn paranimfen worden. Bedankt dat jij mijn paranimf wil zijn en succes met het afronden van de opleiding. En dan mijn andere paranimf: mijn oud‐huisgenote in Utrecht en goede vriendin Yvette. Op de dag van mijn tweede sollicitatiegesprek bij Dermatologie tekende jij voor mij een klavertje vier ter succes. Die dag werd ik aangenomen en heb jij de tekening ingelijst. De afgelopen jaren heeft het klavertje vier op mijn lab‐tafel gestaan; het heeft zeker succes opgeleverd! En nu gaan we weer in dezelfde stad wonen: de Grote Appel, ik kijk er naar uit. Dank je wel dat je voor mijn promotie overkomt. Bij deze wil ik ook jullie partners, Jeroen en Bert, bedanken voor de leuke tijden in de afgelopen jaren.

155 Nu de link met Utrecht is gelegd, wil ik graag mijn oud‐studiegenoten uit Utrecht bedanken voor alle feestjes en de bezoekjes aan Nijmegen: Martin, Roel, Bert, Judith, Monique, Jurgen, Marnix en Yen. We verspreiden ons steeds meer over de wereld, maar hopelijk blijven we elkaar regelmatig zien. Verder wil ik nog mijn oud‐huisgenoot Michael (succes met de laatste loodjes van je promotie) en Richard en Tara bedanken voor alle leuke momenten in de afgelopen jaren. Wat ook over heel Nederland en de wereld verspreid zit, maar zijn basis nog in Utrecht heeft: FatCat, de chill‐out room van het werk. Alle FatCatters worden bedankt voor de soms zeer welkome afleiding en de nodige onzin. Het is geen geheim dat ik Nijmegen om diverse redenen eigenlijk niet zo’n leuke stad vind. Gelukkig heb ik er toch een mooi sociaal leven op kunnen bouwen, zodat er nog best te leven valt. In Nijmegen wonen namelijk ook leuke mensen en heb ik goede vrienden gemaakt. In het bijzonder wil ik Silvo, Pascal, Jos, Mart, Peter, Joost, Loek, Robert, Joris, Ivo en Berend bedanken voor hun steun, interesse en de nodige ontspanning in de vorm van filmavonden, feesten, weekendjes weg, etc. Natuurlijk kan ik de mensen van het atletiek niet vergeten met voorop trainer Richard Riley en zijn dames (o.a. Hanna, Kim, Merel, Suzanne, Hilde, Gonneke) en heren (o.a. Maurice, Alex, Hidde, John) van het RRRT. Daarnaast bedank ik Marcel en Erik voor de pokeravondjes en natuurlijk Klaas, die ik niet alleen op de baan, maar als collega junior onderzoeker ook regelmatig via het werk tegenkom. Het is altijd wel lachen met jou, ik noem maar even een bepaalde NCMLS workshop waar we elkaars competenties moesten beoordelen en natuurlijk de PhD student retraites. Ook jou wens ik veel succes toe met de laatste loodjes van het promoveren.

Tot slot wil ik mijn lieve familie bedanken voor alle steun, begrip en interesse van de afgelopen jaren. Het is niet altijd makkelijk geweest om uit te leggen waarmee ik zo druk bezig was, maar dit proefschrift bewijst dat ik toch wel iets heb uitgevoerd naast het werk in het restaurant. Mijn broers Djindji en Siao, grote neef Feng, schoonzusjes Hsu‐Sia, Shie Yee en Jinting, en natuurlijk kleine Jamie, met jullie heb ik de afgelopen jaren de vele mooie, gelukkige en bijzondere momenten van het leven mogen meemaken, welke vaak een goede afleiding van het onderzoek en reden voor zelfreflectie waren. De laatste regels zijn voor mijn ouders, die mij altijd hebben gestimuleerd om te studeren en hard te werken. Zelf hebben jullie het goede voorbeeld gegeven van hoe je door hard werken je doelen kunt bereiken. De verplichte weekendjes Limburg kwamen niet altijd goed uit, maar ik heb er ook veel voordeel van gehad. Lieve papie en mamie, bedankt voor alles. Ondanks dat ik misschien niet helemaal in het ideale plaatje pas, weet ik dat jullie van mij houden en trots op mij zijn. Ik bof toch maar met zulke ouders en ik hou ontzettend veel van jullie.

156 Curriculum Vitae

Tsing Cheng werd geboren op 3 augustus 1979 te Heerlen. Zijn jeugd bracht hij samen met zijn broers en neef door in Born, waar zijn ouders een Chinees restaurant hebben. In 1997 werd het Gymnasium diploma aan het College Sittard behaald, waarna hij in hetzelfde jaar begon aan de studie Medische Biologie aan de Universiteit Utrecht. De opleiding omvatte twee stages waarin onderzoekservaring werd opgedaan. Bij de afdeling Medische Genetica van het UMC Utrecht heeft hij meegewerkt aan een genome‐wide scan in type 2 Diabetes Mellitus in de groep van prof. dr. Cisca Wijmenga. Tevens werd in het kader van deze stage een prijswinnende scriptie geschreven. Voor de tweede stage vertrok hij naar Edmonton, Canada, waar hij bij de afdeling Medische Genetica van de University of Alberta onderzoek deed naar eiwitten betrokken bij stoornissen in het koper‐metabolisme onder begeleiding van prof. dr. Diane W. Cox. Tijdens de studie was hij actief binnen de Medisch Biologen Vereniging Mebiose, waar hij een jaar lang een bestuursfunctie vervulde. De studie werd in 2003 succesvol afgerond met het behalen van het doctoraal diploma Medische Biologie. In 2004 is hij als junior onderzoeker aan zijn promotieonderzoek begonnen bij de afdeling Dermatologie van het UMC St. Radboud Nijmegen onder leiding van dr. Patrick L.J.M. Zeeuwen en prof. dr. Joost Schalkwijk. De resultaten van het verrichte onderzoek zijn beschreven in dit proefschrift. Tevens heeft hij zijn resultaten op diverse nationale en internationale congressen gepresenteerd en heeft hij een NCMLS Award en een ESDR Travel Grant ontvangen. Vanaf juli 2009 zal hij werkzaam zijn als postdoctorale onderzoeker in de groep van prof. dr. Angela M. Christiano bij de afdeling Dermatologie van de Columbia University in New York, Verenigde Staten.

157 List of Publications

Cheng, T., Hitomi, K., van Vlijmen‐Willems, I.M., de Jongh, G.J., Yamamoto, K., Nishi, K., Watts, C., Reinheckel, T., Schalkwijk, J., and Zeeuwen, P.L. (2006) Cystatin M/E is a high affinity inhibitor of cathepsin V and cathepsin L by a reactive site that is distinct from the legumain‐binding site. A novel clue for the role of cystatin M/E in epidermal cornification. J.Biol.Chem. 281, 15893‐15899

Zeeuwen, P.L., Ishida‐Yamamoto, A., van Vlijmen‐Willems, I.M., Cheng, T., Bergers, M., Iizuka, H., and Schalkwijk, J. (2007) Colocalization of cystatin M/E and cathepsin V in lamellar granules and corneodesmosomes suggests a functional role in epidermal differentiation. J.Invest.Dermatol. 127, 120‐128

Cheng, T., van Vlijmen‐Willems, I.M., Hitomi, K., Pasch, M.C., van Erp, P.E., Schalkwijk, J., and Zeeuwen P.L. (2009) Colocalization of cystatin M/E and its target proteases suggests a role in terminal differentiation of human hair follicle and nail. J.Invest.Dermatol. 129, 1232‐1242

Zeeuwen, P.L., Cheng, T., and Schalkwijk, J. (2009) The biology of cystatin M/E and its cognate proteases. J.Invest.Dermatol. In press, doi:10.1038/jid.2009.40

Cheng, T., Tjabringa, G.S., van Vlijmen‐Willems, I.M., Hitomi K., van Erp, P.E., Schalkwijk, J., and Zeeuwen, P.L. (2009) A biochemical pathway that controls skin barrier formation: expression of its key components in inflammatory skin diseases. Br.J.Dermatol. In press

Cheng, T., Dolk, C., Caubet, C., Simon M., Watts, C., Kamsteeg, M., Serre, G., Schalkwijk, J., and Zeeuwen, P.L. (2009) A regulatory role for cystatin M/E in epidermal desquamation. Identification of a two‐chain form of cathepsin V. Submitted

158 ABSTRACTS

Cheng, T., Hitomi, K., Ishida‐Yamamoto, A., van Vlijmen‐Willems, I.M., de Jongh, G.J., Bergers, M., Schalkwijk, J., and Zeeuwen, P.L. (2006) Novel targets for cystatin M/E: a new route in epidermal differentiation. J.Invest.Dermatol. 126, Supplement 3, S3 (abstract for the 36th annual meeting of the European Society for Dermatological Research (ESDR), 7 – 9 September 2006, Paris, France: oral presentation)

Cheng, T., Hitomi, K., Ishida‐Yamamoto, A., van Vlijmen‐Willems, I.M., de Jongh, G.J., Schalkwijk, J., and Zeeuwen, P.L. (2007) Novel targets for cystatin M/E: a new pathway in epidermal differentiation (abstract for the 10th Gordon Research Conference (GRC) on Barrier Function of Mammalian Skin, 5 – 10 August, 2007, Newport, Rhode Island, USA: poster presentation)

Cheng, T., Dolk, C., Caubet, C., Simon, M., Watts, C., Kamsteeg, M., Serre, G., Schalkwijk, J., and Zeeuwen, P.L. (2009) A potential regulatory role for cathepsin V and its cognate inhibitor cystatin M/E in epidermal desquamation. J.Invest.Dermatol. 129, Supplement 1, S74 (abstract for the 69th annual meeting of The Society for Investigative Dermatology (SID), 6 – 9 May 2009, Montréal, Canada: poster presentation)

159

160 Color illustration – Chapter 1, page 13

Figure 3: Cornified envelope assembly. Initiation of cornified envelope assembly starts in the spinous layers with the attachment of envoplakin, periplakin, and involucrin to the plasma membrane by transglutaminases. This is followed by incorporation of lipids to form the lipid envelope for water barrier function. In the granular layers additional structural proteins including loricrin and SPRs are crosslinked to the scaffold for reinforcement of the cornified envelope. Eventually, other proteins are also crosslinked to the cornified envelope. Reprinted with permission from: Candi, E., Schmidt, R., Melino, G. (2005) Nat.Rev.Mol.Cell Biol. 6, 328‐340. ©2005 Nature Publishing Group

161 Color illustration – Chapter 2, page 54

Figure 1: Predicted three‐dimensional representation of cystatin M/E. The crystal structure of cystatin D was used as a basis for this model. The regions of the inhibitory sites for papain‐ like cysteine proteases and for legumain are marked. The locations of the amino acids substituted in the cystatin M/E variants are shown in white.

162 Color illustration – Chapter 2, page 55

Figure 3: Immunostaining for cystatin M/E, cathepsin V, cathepsin L, and legumain in normal human skin and its appendages. Note the stratum‐specific expression of cystatin M/E and CTSV in the stratum granulosum of the epidermis. CTSV is only present in the inner layer of the outer root sheet of the hair follicle, whereas cystatin M/E is also expressed in many more layers of the outer root sheet of the hair follicle. The secretory coil epithelium of the sweat glands is positively stained for cystatin M/E (arrowheads), but the ductal part is almost negative (the asterisk is surrounded by ducts). No expression of CTSV was found in the sweat glands. CTSL and legumain are also expressed in the stratum granulosum. Furthermore, positive staining was also found in the other layers of the epidermis and its appendages. The sweat gland ducts as well as the secretory coils are both positively stained for CTSL and legumain. Scale bar = 100 μm.

163 Color illustration – Chapter 3, page 62

Figure 1: Colocalization of cystatin M/E and CTSV in human normal skin. Immunofluorescence staining of (green color; a, d, g, j) cystatin M/E and (red color; b, e, h, k) CTSV, and (yellow‐orange merge color; c, f, i, l) double staining for both proteins in (a–c) the epidermis, (d–f) the hair follicle, (g–i) the sweat glands, and (j–l) the sebaceous glands. The rectangular area marked in (c) is shown enlarged in (m). Note that the SC is detached from the SG. The rectangular areas marked in (d and e) are shown enlarged in (n and o). DNA staining by 4’,6‐diamidine‐2’‐phenylindole dihydrochloride is colored in blue. Bar = (a–i) 50 μm; (j–l) 100 μm; and (m–o) 12.5 μm.

164 Color illustration – Chapter 3, page 63

Figure 2: Colocalization of cystatin M/E and CTSL in human normal skin. Immunofluorescence staining of (green color; a, d, g, j) cystatin M/E and (red color; b, e, h, k) CTSL, and (yellow‐orange merge color; c, f, i, l) double staining for both proteins in (a–c) the epidermis, (d–f) the hair follicle, (g–i) the sweat glands, and (j–l) the sebaceous glands. DNA staining by 4’,6‐diamidine‐2’‐phenylindole dihydrochloride is colored in blue. Bar = (a–i) 50 μm and (j–l) 100 μm.

165 Color illustration – Chapter 3, page 64

Figure 3: Colocalization of cystatin M/E and AEP in human normal skin. Immunofluorescence staining of (green color; a, d, g, j) cystatin M/E and (red color; b, e, h, k) AEP, and (yellow‐orange merge color; c, f, i, l) double staining for both proteins in (a–c) the epidermis, (d–f) the hair follicle, (g–i) the sweat glands, and (j–l) the sebaceous glands. DNA staining by 4’,6‐diamidine‐2’‐phenylindole dihydrochloride is colored in blue. Bar = (a–i) 50 μm and (j–l) 100 μm.

166 Color illustration – Chapter 5, page 90

Figure 2: Colocalization of cystatin M/E and its target proteases CTSV, CTSL, and legumain in human hair follicle. Immunofluorescence staining of (a, d, g, j, m, and p) cystatin M/E (green color), and (b, e, h, k, n, and q) CTSV, CTSL, or legumain (red color), and (c, f, i, l, o, and r) double staining (yellow–orange merge color) for both proteins on (c, i, and o) longitudinal and (f, l, and r) transversal sections of the hair follicle. Nuclei are blue (DNA staining by 4’,6‐diamine‐2’‐phenylindole dihydrochloride (DAPI)). Bar = 100 μm.

167 Color illustration – Chapter 5, page 91

Figure 3: Colocalization of TGase 3 and CTSL in human epidermis and hair follicle. Immunofluorescence staining of (a and d) TGase 3 (green color), and (b and e) CTSL (red color), and (c and f) double staining (yellow–orange merge color) for both proteins on epidermis (a–c), and both longitudinal (d–f) and transversal (insets d–f) sections of the hair follicle. The arrows (f) show the position of the transversal section (inset). Immunofluorescence staining of legumain was found in the (h) hair bulb, whereas no positive staining could be detected for (g) cystatin M/E or (i) CTSV in this part of the follicle. Nuclei are blue by DAPI staining for DNA. Bar = 100 μm.

168 Color illustration – Chapter 5, page 92

Figure 4: TGase 3 and its presumed substrates in human hair follicle. Immunofluorescence staining of (a, d, g, j, and m) TGase 3 (green color), and (b, e, h, k, and n) loricrin, involucrin, THH, or TGase 1 (red color), and (c, f, i, l, and o) double staining (yellow–orange merge color) for these proteins on transversal (a–i) and longitudinal (j–o) sections of the hair follicle. Nuclei are blue by DAPI staining for DNA. Bar = 100 μm.

169 Color illustration – Chapter 5, page 93

Figure 6: Colocalization of cystatin M/E and CTSL in the human nail unit. Immunofluorescence staining of (a, d, g, j, and m) cystatin M/E (green color), and (b, e, h, k, and n) CTSL (red color), and (c, f, i, l, and o) double staining (yellow–orange merge color) for both proteins in epidermis (a–c), proximal nail fold (d–f), nail matrix (g–i), nail bed (j–l), and hyponychium (m–o). Nuclei are blue by DAPI staining for DNA. Bar = 100 μm.

170 Color illustration – Chapter 5, page 94

Figure 7: Colocalization of TGase 3 and CTSL in the human nail unit. Immunofluorescence staining of (a, d, g, j, and m) TGase 3 (green color), and (b, e, h, k, and n) CTSL (red color), and (c, f, i, l, o) double staining (yellow–orange merge color) for both proteins in epidermis (a–c), proximal nail fold (d–f), nail matrix (g–i), nail bed (j–l), and hyponychium (m–o). Nuclei are blue by DAPI staining for DNA. Bar = 100 μm.

171 Color illustration – Chapter 6, page 108

Figure 4: Protein expression of cystatin M/E and its target proteases in inflamed skin. Representative immunofluorescence stainings (each group n=6) of cystatin M/E, CTSV, CTSL, LGMN, and TGM3 protein expression in normal human skin (left column), and in skin of patients with AD (middle column) and PS (right column). Nuclei are blue (DNA staining by 4’,6‐diamine‐2’‐phenylindole dihydrochloride (DAPI)). Bar = 100 µm, CysM = cystatin M/E.

172 Color illustration – Chapter 6, page 110

Figure 6: Expression of epidermal differentiation markers in human reconstructed skin. Development of normal human skin equivalents. Adult human keratinocytes were seeded on DED, and cultured for 3 days submerged, followed by 0, 3, 6, 9 and 12 days culturing at the air‐liquid interface. Morphology of the skin constructs was studied by H&E staining. Differentiation of the skin constructs was studied by immunohistochemical stainings for involucrin, loricrin and filaggrin. Bar = 100 µm.

173 Color illustration – Chapter 6, page 112

Figure 8: Expression of cystatin M/E and its target proteases in reconstructed human skin. Adult human keratinocytes were seeded on DED, and cultured for 3 days submerged, followed by 0, 3, 6, 9 and 12 days culturing at the air‐liquid interface. Representative immuno‐fluorescence stainings (in duplo) showed the expression of Cystatin M/E, CTSV, CTSL, LGMN, and TGM3 in these reconstructed skin cultures. Nuclei are blue (DNA staining by 4’,6‐diamine‐2’‐phenylindole dihydrochloride (DAPI). Bar = 100 µm, CysM = cystatin M/E.

174 Color illustration – Chapter 7, page 126

B

Figure 2: Expression of eGFP in transfected keratinocytes. A) eGFP protein expression analysis by flowcytometry in non‐transfected and pBS‐INV‐Cst6‐IRES‐eGFP transfected keratinoctyes. The transfected sample shows an increased number of fluorescent keratinocytes as compared to non‐transfected cells. B) Fluorescence microscopy shows eGFP protein expression in a pBS‐INV‐Cst6‐IRES‐eGFP transfected keratinocyte.

175 Color illustration – Chapter 8, page 139

Figure 1: The role of cystatin M/E in epidermal differentiation. A simplified model of the presumed role of cystatin M/E in regulation of processes that control epidermal cornification and desquamation. Cystatin M/E is an inhibitor of legumain and the cathepsins L and V (CTSL, CTSV). Inhibition of legumain activity could regulate processing and activation of (pro)cathepsins, a process that is also under control of intra‐ and extracellular pH changes. Inhibition of CTSV activity could regulate desquamation as CTSV is able to degrade (corneo‐) desmosomal proteins such as desmoglein‐1 (DSG1), desmocollin‐1 (DSC1), and corneodesmosin (CDSN). Inhibition of CTSL activity could regulate cornification as CTSL processes and activates the crosslinking enzyme TGM3. CTSL is also able to process cathepsin D (CTSD), an aspartic protease that can regulate TGM1 activity. Reprinted with permission from: Zeeuwen, P.L., Cheng, T., Schalkwijk, J. (2009). J Invest Dermatol © 2009 The Society for Investigative Dermatology

176