Advances in immunology, volume 129

Function of Protein Antigens Presented by MHC Class II Molecules 6 3.1 MHC Class II Molecules Present Protein Antigens to B Cells 6 3.2 Misfolded Cellular Proteins Rescued from Protein D

525 B Street, Suite 1800, San Diego, CA 92101-4495, USA

The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK

125 London Wall, London, EC2Y 5AS, UK

First edition 2016

© 2016 Elsevier Inc All rights reserved

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions

This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices

Knowledge and best practice in this field are constantly changing As new research and experience broaden our understanding, changes in research methods, professional practices,

or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.

ISBN: 978-0-12-804799-6

ISSN: 0065-2776

For information on all Academic Press publications

visit our website at http://store.elsevier.com/

Hisashi Arase

Department of Immunochemistry, Research Institute for Microbial Diseases, and Laboratory

of Immunochemistry, WPI Immunology Frontier Research Center, Osaka University, Suita, Osaka, Japan

Marie-Dominique Filippi

Division of Experimental Hematology and Cancer Biology, Cincinnati Children’s Research Foundation, and University of Cincinnati College of Medicine, Cincinnati, Ohio, USA

Institute for Advanced Study, Kyushu University, Higashi-ku, Fukuoka, Japan

Advances in Immunology, Volume 129 # 2016 Elsevier Inc.

ISSN 0065-2776 All rights reserved ix

Jonas Ungerba¨ck

Division of Biology & Biological Engineering, California Institute of Technology, Pasadena, California, USA, and Department of Clinical and Experimental Medicine, Experimental Hematopoiesis Unit, Faculty of Health Sciences, Link€oping University, Link€oping, Sweden Xiaohong Wang

Department of Pathology & Immunology, Baylor College of Medicine, Houston, Texas, USA

Rheumatoid Rescue of Misfolded Cellular Proteins by MHC Class II Molecules: A New Hypothesis for Autoimmune Diseases

2.2 MHC Class II Molecules Function as a Molecular Chaperon to Transport

3 Function of Protein Antigens Presented by MHC Class II Molecules 6 3.1 MHC Class II Molecules Present Protein Antigens to B Cells 6 3.2 Misfolded Cellular Proteins Rescued from Protein Degradation by MHC Class II

3.3 Aberrant MHC Class II Expression on Autoimmune-Diseased Tissues 7

4 Misfolded Proteins Presented on MHC Class II Molecules Are Targets for

4.1 IgG Heavy Chain Presented on MHC Class II Molecules Is a Specific Target for

4.2 β2-Glycoprotein I Associated with MHC Class II Molecules Is a Specific Target

5 Susceptibility to Autoimmune Diseases Is Associated with the Affinity of Misfolded

5.1 MHC Class II Alleles and Autoimmune Disease Susceptibility 11 5.2 Autoantibody Binding to Misfolded Protein/MHC Class II Complex Is

Associated with Autoimmune Disease Susceptibility 12

6 Involvement of Misfolded Protein –MHC Class II Molecule Complexes in

6.2 B Cell Removal Are Effective Treatment for Autoimmune Diseases 13

Advances in Immunology, Volume 129 # 2016 Elsevier Inc.

ISSN 0065-2776 All rights reserved 1

7 Misfolded Cellular Proteins Rescued from Degradation by MHC Class II Molecules

7.1 Misfolded Proteins Associated with MHC Class II Molecules as

7.2 Misfolded Protein Complexed with MHC Class II Molecules as Primary

8 Misfolded Proteins Presented on MHC Class II Molecules as a Therapeutic Target for

1 INTRODUCTION

Specific major histocompatibility complex (MHC) class II alleles affectsusceptibility to many autoimmune diseases Recent genome-wide associa-tion studies have confirmed that the MHC class II loci are the genes moststrongly associated with susceptibility to many autoimmune diseases, includ-ing rheumatoid arthritis (RA) Because MHC class II molecules present pep-tide antigens to helper T cells, an abnormal helper T cell response has beenconsidered to be the main cause of MHC class II gene-associated autoim-mune diseases However, the pathogenic peptide antigens associated withthe autoimmune disease susceptibility conferred by each MHC class II allelehave not yet been identified; therefore, it remains unclear how the MHCclass II gene controls susceptibility to autoimmune diseases On the other

hand, our recent analyses of MHC class II molecules have revealed that teins misfolded in the endoplasmic reticulum (ER) are transported to the cellsurface without being processed to peptides upon association with MHCclass II molecules (Jiang et al., 2013) Furthermore, misfolded proteins pres-ented on MHC class II molecules appear to be involved in autoimmune dis-ease susceptibility as specific targets for autoantibodies (Jin et al., 2014;Tanimura et al., 2015) These novel functions of MHC class II moleculesprovide new insights into the molecular mechanism underlying autoim-mune diseases that will help us answer questions regarding why autoanti-bodies against autoantigens are produced in patients with autoimmunediseases and why MHC class II genes are strongly associated with suscepti-bility to many autoimmune diseases.

pro-2 TRANSPORT OF ER-MISFOLDED PROTEINS TO THECELL SURFACE BY MHC CLASS II MOLECULES

2.1 MHC Class II Molecules Induce Cell Surface Expression

of Misfolded MHC Class I Molecules

The MHC class I molecule comprises a trimolecular complex that includes aheavy chain,β2-microglobulin, and a peptide MHC class I molecules arenot expressed on the cell surface in the absence ofβ2-microglobulin or trans-porter associated with antigen processing (TAP), the latter of which trans-ports proteasome-derived peptides to the ER where they are acquired byMHC class I molecules, indicating that bothβ2-microglobulin and peptide,are required for the cell surface expression of MHC class I molecules On theother hand, some unique monoclonal antibodies (mAbs) that are specific forhuman MHC class I molecules lackingβ2-microglobuin and peptide anti-gens, such as HC10 and L31, have been identified (Giacomini et al., 1997;Stam, Spits, & Ploegh, 1986) These mAbs do not recognize normal MHCclass I molecules associated with β2-microglobulin and peptide antigens(Sibilio et al., 2005) Interestingly, the epitopes recognized by these mAbsare localized on the α1- and α2-domains of MHC class I, which are notlocated within theβ2-microglobulin-binding region, suggesting that thesemAbs recognize a certain unique conformation of MHC class I moleculesthat is induced in the absence of β2-microglobulin and peptide antigens(Arosa, Santos, & Powis, 2007) These incomplete MHC class

I molecules fail to achieve a correct conformation and are not expressed

on the cell surface However, certain cells, such as B cell lines, are recognized

by these mAbs specific for unusual or misfolded MHC class I molecules kingβ2-microglobulin and peptides This phenomenon suggests the exis-tence of a molecular chaperon that transports misfolded MHC class

lac-I molecules lackingβ2-microglobulin and peptides to the cell surface.Expression cloning to identify molecules that would permit expression ofMHC class I proteins on the cell surface unexpectedly revealed that MHCclass II molecules induce the cell surface expression of unusual or misfoldedMHC class I molecules (Jiang et al., 2013) Upon further analysis, someMHC class II alleles induced misfolded MHC class I expression on the cellsurface but others did not, and this difference depended on the amino acidresidues present within the peptide-binding groove of the MHC class IImolecule Furthermore, the MHC class II-induced expression of misfoldedMHC class I molecules was almost completely blocked by a peptide cova-lently attached to MHC class II molecules Indeed, some MHC classII-positive cells express misfolded MHC class I molecules that are recog-nized by HC10 or L31 mAbs, and the direct association of MHC class

I molecules with MHC class II molecules is detectable in these cells fore, MHC class II molecules appear to be involved in the expression ofmisfolded MHC class I molecules on the cell surface (Fig 1)

There-Figure 1 Misfolded major histocompatibility complex (MHC) class I expression tated by MHC class II molecules MHC class I molecules normally are expressed in asso- ciation with β2-microglobulin and peptide antigens (right) In the absence of peptide antigens or β2-microglobulin, MHC class I molecules are not folded correctly and are not expressed on the cell surface However, when the misfolded MHC class

facili-I molecule is associated with an MHC class facili-Ifacili-I molecule in the endoplasmic reticulum,

it is directly transported to the cell surface by the MHC class II molecule without going peptide processing (middle).

under-2.2 MHC Class II Molecules Function as a Molecular Chaperon

to Transport Misfolded Cellular Protein to the Cell SurfaceThe cell surface expression of misfolded MHC class I molecules by MHCclass II molecules raised the possibility that other misfolded proteins mightalso be transported to the cell surface by MHC class II molecules Hen egglysozyme (HEL) is a well-characterized secreted protein, the correct folding

of which requires S–S bonds (Ohkuri, Ueda, Tsurumaru, & Imoto, 2001).When a mutant HEL in which two cysteine residues were substituted withalanine was co-expressed with MHC class II molecules, the mutant HELprotein, which was neither secreted nor expressed on the cell surface, wasinduced on the cell surface in the presence of MHC class II molecules(Jiang et al., 2013) Furthermore, the full-length HEL protein wasco-precipitated with MHC class II molecules These findings, together withthe analyses of MHC class I molecules, suggest that ER-misfolded proteinsare transported to the cell surface by MHC class II molecules upon associ-ation with their peptide-binding grooves In other words, MHC class IImolecules function as a molecular chaperon to transport ER-misfolded pro-teins to the cell surface

The presentation of whole proteins by MHC class II molecules may seemunusual because it is widely accepted that MHC class II molecules present shortpeptide antigens However, several papers have reported the association oflarge proteins with the peptide-binding grooves of MHC class II molecules(Aichinger et al., 1997; Anderson, Swier, Arneson, & Miller, 1993; Busch,Cloutier, Sekaly, & Hammerling, 1996; Lechler, Aichinger, & Lightstone,1996) When MHC class II molecules were expressed with an invariant chainlacking the endosomal localization signal, the invariant chain was directly trans-ported to the cell surface in association with MHC class II molecules (Anderson

et al., 1993) In addition, association of MHC class II molecules with large teins was observed in the absence of invariant chain (Aichinger et al., 1997;Busch et al., 1996) Therefore, the transport of ER-misfolded proteins tothe cell surface is an intrinsic function of MHC class II molecules In addition,antigens captured by the endocytic pathway form large molecular complexeswith MHC class II molecules in antigen-presenting cells (Castellino,Zappacosta, Coligan, & Germain, 1998) These observations suggest thatMHC class II molecules exhibit the capacity to present large molecular antigensderived not only from misfolded proteins in the ER but also from antigens cap-tured by the endocytic pathway However, the immunological functions oflarge proteins associated with MHC class II molecules were not extensivelyanalyzed; therefore, these functions remained unclear

pro-3 FUNCTION OF PROTEIN ANTIGENS PRESENTED BYMHC CLASS II MOLECULES

3.1 MHC Class II Molecules Present Protein Antigens to

B Cells

The presentation of whole proteins instead of peptides by MHC class II ecules suggests that MHC class II molecules might be involved in an as yetunknown immune response B cells expressing a high-affinity antigen recep-tor can be stimulated with soluble antigens However, these antigens must beassociated with certain cell surface molecules to stimulate B cells with low-affinity antigen receptors, such as those expressed on naı¨ve B cells(Batista & Harwood, 2009; Qi, Egen, Huang, & Germain, 2006) Thepresentation of whole proteins by MHC class II molecules suggests that theseproteins might be involved in B cell activation Indeed, B cells expressing alow-affinity antigen receptor against HEL protein can be stimulated withHEL protein presented on MHC class II molecules but not by solubleHEL protein alone (Jiang et al., 2013) This indicates that MHC class II mol-ecules might be directly involved in the antigen-specific B cell response

mol-3.2 Misfolded Cellular Proteins Rescued from Protein

Degradation by MHC Class II Molecules Might Be

Pathogenic

The invariant chain, which is associated with newly synthesized MHC class

II molecules, transports MHC class II molecules to endolysosomal ments, where they acquire peptide antigens (Germain, 2011) However, theaffinities of MHC class II molecules for the invariant chain are known todiffer in an MHC class II allele-dependent manner (Davenport et al.,1995) It is possible that MHC class II molecules will preferentially associatewith misfolded proteins rather than the invariant chain if the former has astronger affinity for MHC class II molecules than the latter Indeed, the effi-ciency of the invariant chain to block the association of misfolded proteinswith MHC class II molecules differs in an allele-dependent manner, asdescribed above (Jiang et al., 2013; Jin et al., 2014; Tanimura et al.,2015) As most misfolded proteins do not possess a lysosomal targeting sig-nal, MHC class II molecules associated with misfolded proteins instead ofinvariant chain are directly transported to the cell surface without going

compart-to the endolysosomal compartments

The folding of a newly synthesized protein is a complex process that stitutively generates significant amounts of misfolded proteins In certaintypes of cells, more than half of all newly synthesized proteins are foldedincorrectly (Meusser, Hirsch, Jarosch, & Sommer, 2005) However, thesenewly synthesized misfolded proteins typically are promptly degraded inthe cells through various pathways such as ER-associated degradation(ERAD) and therefore not transported outside the cells (Meusser et al.,2005) Accordingly, immune cells are not exposed to these misfolded pro-teins Whereas both the primary structures and conformations of antigens areinvolved in antibody recognition, only the primary structures of antigens areinvolved in T cell recognition because T cell receptor recognizes short pep-tide antigens presented on MHC molecules Therefore, unlike T cells, it ispossible that some B cells do not acquire tolerance to misfolded cellular pro-teins If these misfolded cellular proteins are rescued from protein degrada-tion by MHC class II molecules and subsequently transported extracellular,

con-B cells might recognize these proteins as “neo-self”-antigens and initiate anantibody response (Fig 2)

3.3 Aberrant MHC Class II Expression on

Autoimmune-Diseased Tissues

More than 30 years ago, it was reported that autoimmune-diseased tissuesaberrantly expressed MHC class II molecules (Bottazzo, Pujol-Borrell,Hanafusa, & Feldmann, 1983) Unlike normal thyroid tissues, tissues frompatients with Graves’ disease or Hashimoto’s thyroiditis aberrantly expressMHC class II molecules Similar aberrant MHC class II expression was reported

in various tissues affected by autoimmune diseases such as RA, type I diabetes,primary biliary cirrhosis, and psoriasis (Ballardini et al., 1984; Feldmann et al.,1988; Gottlieb et al., 1986) Because particular MHC class II alleles are associ-ated with autoimmune disease susceptibility, this aberrant MHC class II expres-sion in autoimmune-diseased tissues was considered to be involved in thepathogenicity of autoimmune diseases Indeed, nonimmune cells, such asendothelial cells, strongly express MHC class II molecules in response to stim-ulation from cytokines such as IFN-γ (Jaffe et al., 1989; Pober et al., 1983;Todd, Pujol-Borrell, Hammond, Bottazzo, & Feldmann, 1985) However,these nonimmune cells do not express costimulatory molecules required forthe induction of T cell responses, such as CD80 or CD86 T cells that recognizeantigens presented on MHC class II molecules in the absence of costimulatorysignals are likely to become anergic (Appleman & Boussiotis, 2003) Therefore,aberrant MHC class II expression on nonimmune cells has been considered a

consequence of the inflammation elicited by autoimmunity rather than a cause

of autoimmune disease, and the pathophysiological function of this aberrantMHC class II expression in autoimmune-diseased tissues has not been exten-sively analyzed Unlike T cells, B cells do not require costimulatory signals torespond to antigens, although they require T cell help for Ig class switching

Figure 2 Rheumatoid rescue of misfolded proteins by major histocompatibility plex (MHC) class II molecules In steady state, MHC class II expression is restricted to spe- cific immune cells such as dendritic cells and B cells However, MHC class II molecules are expressed on most cells following stimulation with certain cytokines such as IFN- γ, which is produced in response to infection or inflammation The invariant chain asso- ciates with nascent MHC class II molecules and blocks the association of MHC class II molecules with endoplasmic reticulum (ER)-misfolded proteins However, the affinities

com-of various MHC class II molecules for the invariant chain differ due to allelic phism of MHC class II genes If the avidity of an MHC class II molecule for a misfolded protein is higher than that for the invariant chain, it is possible that misfolded proteins, rather than the invariant chain, will bind to MHC class II molecules As the misfolded proteins do not contain an endolysosomal-targeting signal, they are transported directly to the cell surface by MHC class II molecules Thus, MHC class II molecules func- tion as a molecular chaperon to rescue ER-misfolded proteins from protein degradation Because immune cells normally are not exposed to misfolded proteins and may there- fore be intolerant to them, misfolded proteins rescued by MHC class II molecules may be recognized as “neo-self”-antigens and thus induce autoantibody production In this way, misfolded proteins rescued from protein degradation by MHC class II molecules may be involved in the pathogenesis of autoimmune diseases as autoantibody targets.

polymor-Given this difference between T cells and B cells, it is possible that misfoldedproteins rescued by aberrantly expressed MHC class II molecules can stimulate

B cells to produce autoantibodies

4 MISFOLDED PROTEINS PRESENTED ON MHC CLASS IIMOLECULES ARE TARGETS FOR AUTOANTIBODIES INAUTOIMMUNE DISEASES

4.1 IgG Heavy Chain Presented on MHC Class II Molecules

Is a Specific Target for Autoantibodies in RA

Disease-specific autoantibodies are produced in many autoimmune disorders.Some of these autoantibodies are directly involved in the pathogenicity ofautoimmunity Therefore, it is important to identify the target molecules rec-ognized by autoantibodies in order to understand the pathogenicity of auto-immune diseases If misfolded proteins rescued from protein degradation byMHC class II molecules are targets for autoimmune diseases, autoantibodiesmay recognize these proteins presented on MHC class II molecules.Rheumatoid factor (RF) is a well-known autoantibody discoveredapproximately 75 years ago RF is specific for denatured, but not native,IgG and is detected in approximately 80% of patients with RA (Dorner,Egerer, Feist, & Burmester, 2004) Because RF titers are well correlated withthe clinical symptoms of RA, RF remains an important diagnostic indicator

of RA However, the natural target antigens that induce RF productionremain undefined In addition, it remains unclear why most patients with

RA are RF positive

Antibodies comprise a heavy chain and a light chain; the heavy chain isnot secreted or expressed on the cell surface in the absence of the light chain.However, the heavy chain alone can be expressed well on the cell surface inthe presence of MHC class II molecules (Jin et al., 2014) Furthermore, IgGpresented on MHC class II molecules is recognized by autoantibodies frompatients with RA More importantly, the IgG heavy chain presented onMHC class II molecules was recognized by autoantibodies from patientswith RA but not by those from non-RA patients, including those positivefor RF This suggests that the IgG heavy chain presented on MHC class IImolecules is more specific for autoantibodies from RA patients when com-pared with traditional RF detected by immobilized IgG This finding impli-cates the IgG heavy chain, when presented on MHC class II molecules, as amajor target for RA autoantibodies

4.2 β2-Glycoprotein I Associated with MHC Class II Molecules Is

a Specific Target for Autoantibodies in AntiphospholipidSyndrome

Antiphospholipid syndrome (APS) is an autoimmune disorder associatedwith thrombosis and pregnancy complications (Wilson et al., 1999).Although autoantibodies associated with APS were initially characterized

by reactivity to phospholipids such as cardiolipin, recent analyses have ealed that these autoantibodies are directed mainly against the phospholipid-associatedβ2-glycoprotein I (Bas de Laat, Derksen, & de Groot, 2004; Galli,Barbui, Zwaal, Comfurius, & Bevers, 1993; McNeil, Simpson,Chesterman, & Krilis, 1990) β2-glycoprotein I forms a circular structure

rev-in sera that is lrev-inearized upon brev-indrev-ing to phospholipids, thus exposrev-ing tic autoantibody epitopes onβ2-glycoprotein I (Agar et al., 2010; de Laat,Derksen, van Lummel, Pennings, & de Groot, 2006) However, it hasremained unclear whether phospholipid-boundβ2-glycoprotein I is a nat-ural target for autoantibodies and is involved in the pathogenesis of APS In

cryp-an cryp-analysis of the association betweenβ2-glycoprotein I and MHC class IImolecules, which was similar to that described earlier for the IgG heavychain, intact β2-glycoprotein I was also found to be presented on MHCclass II molecules on the cell surface Furthermore, β2-glycoprotein

I presented on MHC class II molecules was recognized by APS bodies (Tanimura et al., 2015) Anti-β2-glycoprotein I Ab and anti-cardiolipin Ab titers are used clinically to diagnose APS, although somepatients with clinical manifestations of APS do not have detectable theseAbs On the other hand, more than 80% of patients express autoantibodiesagainstβ2-glycoprotein I presented on MHC class II molecules (Tanimura

autoanti-et al., 2015) This suggests thatβ2-glycoprotein I, when presented on MHCclass II molecules, is a major target antigen for autoantibodies in patients withAPS (Fig 3)

The presence of autoantibodies against autoantigens associated with MHCclass II molecules suggests that these autoantigens associate with MHC class IImolecules in certain tissues Indeed, complexes of the IgG heavy chain orβ2-glycoprotein I with MHC class II molecules were detected on synovial mem-branes from patients with RA or uterine decidual tissues from patients withAPS, respectively Similar to the analyses of autoantibodies from patients with

RA and APS, autoantibodies associated with other autoimmune diseases alsospecifically recognize autoantigens complexed with MHC class II molecules(Hui Jin, Ryosuke Hiwa, Satoko Morikami, Noriko Arase, & Hisashi Arase,unpublished observation) Therefore, complexes of misfolded proteins with

MHC class II molecules appear to be major targets of autoantibodies in manyautoimmune diseases.

5 SUSCEPTIBILITY TO AUTOIMMUNE DISEASES ISASSOCIATED WITH THE AFFINITY OF MISFOLDEDPROTEINS FOR MHC CLASS II MOLECULES

5.1 MHC Class II Alleles and Autoimmune Disease

Susceptibility

Particular MHC class II gene alleles are strongly associated with ity to many autoimmune diseases Extensive analyses of RA susceptibilityaccording to HLA-DR alleles have suggested that specific amino acid resi-dues in the peptide-binding groove of HLA-DR are associated with suscep-tibility to RA (Raychaudhuri et al., 2012) Therefore, certain peptideantigens are thought to be involved in autoimmune diseases However, pep-tide antigens that could explain the susceptibility to autoimmune diseasesconferred by each MHC class II allele have not been identified, and there-fore, the molecular mechanism underlying the control exerted by particular

susceptibil-Figure 3 Recognition of β2-glycoprotein I (β2GPI) presented on MHC class II molecules

by antiphospholipid autoantibodies Native β2GPI forms a circular structure in sera that

is not recognized by antiphospholipid autoantibodies When associated with a pholipid such as cardiolipin, β2GPI forms a linear structure that appears to expose cryp- tic autoantibody epitopes However, autoantibodies from some APS patients do not recognize phospholipid-associated β2GPI On the other hand, β2GPI is also expressed

phos-on cell surfaces in associatiphos-on with MHC class II molecules In additiphos-on, more than 80% of APS patients possess autoantibodies against β2GPI presented on MHC class II molecules, suggesting that β2GPI presented on MHC class II molecules may be a major target antigen for autoantibodies in antiphospholipid syndrome.

MHC class II alleles over susceptibility to autoimmune diseases remainsunknown (Raychaudhuri et al., 2012).

5.2 Autoantibody Binding to Misfolded Protein/MHC Class IIComplex Is Associated with Autoimmune Disease

Susceptibility

In an analysis of autoantibody binding to IgG heavy chains presented onHLA-DR molecules encoded by various alleles, a strong correlation wasobserved between autoantibody binding to IgG heavy chains presented

on HLA-DR and the RA susceptibility conferred by each HLA-DR allele(Jin et al., 2014) The invariant chain only partially blocks the association ofIgG heavy chains with RA-susceptible HLA-DR, but strongly blocks theassociation of IgG heavy chains with RA-resistant HLA-DR Autoanti-bodies fail to bind IgG heavy chains presented on RA-resistant HLA-DR

in the presence of the invariant chain Thus, the IgG heavy chain is the firstmolecule associated with the susceptibility to RA conferred by each HLA-

DR allele Because IgG heavy chain associated with MHC class II moleculescomprises a specific RA autoantibody target, it is possible that the IgG heavychain–MHC class II molecule complex is involved directly in RA pathoge-nicity as an autoantibody target

Similarly, the presentation of self-antigens on MHC class II moleculesencoded by disease-susceptible alleles has been observed in APS(Tanimura et al., 2015) APS-susceptible HLA-DR efficiently presentsβ2-glycoprotein I, and autoantibodies preferentially bind to this complexeven in the presence of the invariant chain Therefore, differences in auto-immune disease susceptibility among the different MHC class II alleles might

be explained by different efficiencies of autoantigen presentation by MHCclass II molecules

6 INVOLVEMENT OF MISFOLDED PROTEIN–MHC CLASS

II MOLECULE COMPLEXES IN

been shown to induce arthritis, suggesting that autoantibodies play an tant role in the pathogenesis of RA, although the target antigens for thesepathogenic autoantibodies remain undefined (Petkova et al., 2006) Autoan-tibodies also play a crucial role in some RA mouse models such as K/BxNmice and the type II collagen-induced arthritis model In K/BxN mice, auto-antibodies against glucose-6-phosphate isomerase are responsible for arthritisdevelopment (Matsumoto et al., 2002; Matsumoto, Staub, Benoist, & Mathis,1999) The adoptive transfer of anti-glucose-6-phosphate isomerase autoan-tibodies from K/BxN mice to healthy mice induces arthritis Similarly, anti-type II collagen autoantibodies induced via immunization with type II colla-gen directly mediate arthritis Similar to the autoantibodies in K/BxN mice,these anti-type II collagen autoantibodies induce arthritis when transferred tohealthy mice (Griffiths & Remmers, 2001) Therefore, autoantibodies appear

impor-to play an important role in the pathogenesis of arthritis not only in RA mousemodels but also in patients with RA

APS autoantibodies are directed mainly against the serum lipoproteinβ2-glycoprotein I and are thought to be involved in the pathogenesis ofAPS However, it remains unclear how these autoantibodies againstβ2-glycoprotein I induce thrombosis or pregnancy complications, as theprotein is not expressed on the surface of healthy blood vascular endothelialcells In addition, it remains unknown why some patients mainly exhibitthrombosis and others predominantly develop pregnancy complications,despite detecting similar autoantibodies in both groups of patients Endothe-lial cells strongly express MHC class II molecules upon IFN-γ stimulation(Jaffe et al., 1989; Pober et al., 1983) Indeed, aberrant MHC class II expres-sion has been observed in uterine decidual tissues from patients with APS.More importantly, complexes ofβ2-glycoprotein I with MHC class II mol-ecules have been detected in uterine decidual tissues from patients with APS(Tanimura et al., 2015) These observations suggest that aberrantly expressedMHC class II molecules on endothelial cells could be targeted by autoanti-bodies, possibly leading to thrombosis in peripheral vessels or the uterus.Therefore, in addition to the presence of autoantibodies, aberrant MHCclass II expression might play an important role in autoantibody-mediatedpathogenesis

6.2 B Cell Removal Are Effective Treatment for AutoimmuneDiseases

The removal of B cells via the administration of an anti-CD20 mAb(rituximab) has been clinically approved as an effective treatment for RA(Jacobi & Dorner, 2010) Anti-CD20 mAb treatment is also effective for

other autoimmune diseases such as APS (Erkan, Vega, Ramon, Kozora, &Lockshin, 2013), systemic lupus erythematosus (Anolik et al., 2004), myas-thenia gravis (Sieb, 2014), Graves’ disease (Heemstra et al., 2008), and pem-phigus (Ahmed, Spigelman, Cavacini, & Posner, 2006; Joly et al., 2007).Furthermore, anti-B lymphocyte stimulator (BlyS) mAb (Belimumab) thatdecreases autoantibody producing B cells is effective for some autoimmunediseases such as SLE (Navarra et al., 2011) Because B cells are involved inantibody production as well as antigen presentation and cytokine secretion,

B cell depletion may affect various aspects in autoimmunity B cell depletion

by anti-CD20 mAb has been reported to correlate with a decrease in antibody levels as well as clinical manifestation, suggesting that autoantibodyproducing B cells play an important role in the pathogenicity of autoim-mune diseases (Cambridge et al., 2006; Thurlings et al., 2008) However,

auto-it is difficult to directly test the pathogenicauto-ity of human autoantibodies fromautoimmune patients in mice because of species differences in autoantigens.Therefore, the pathogenicity of human autoantibodies has been demon-strated only in the context of some autoimmune diseases such as Graves’ dis-ease, myasthenia gravis, and pemphigus As autoantigen–MHC class IImolecule complexes are targeted by autoantibodies, mice expressing bothhuman autoantigens and human MHC class II molecules would be usefulfor testing the pathogenesis of human autoantibodies

7 MISFOLDED CELLULAR PROTEINS RESCUED FROMDEGRADATION BY MHC CLASS II MOLECULES MAYABROGATE IMMUNE TOLERANCE

7.1 Misfolded Proteins Associated with MHC Class IIMolecules as“Nonself”-Antigens

As described above, misfolded proteins, when complexed with MHC class IImolecules, are specific targets for the autoantibodies produced in autoim-mune diseases In addition, a strong association has been observed betweenautoantibody binding to IgG heavy chains presented on MHC class II mol-ecules and the RA susceptibility conferred by each HLA-DR allele Theseobservations suggest that self-antigens presented on MHC class II moleculesare involved in the pathogenicity of autoimmune diseases by serving as tar-gets for autoantibodies In cases involving aberrantly induced or increasedMHC class II expression in response to infection or inflammation in whichER-misfolded proteins have a stronger affinity for MHC class II moleculesthan the invariant chain, misfolded cellular proteins associate with MHCclass II molecules and are subsequently transported to the cell surface

The prompt degradation of ER-misfolded proteins through various cellularmechanisms in steady state (Meusser et al., 2005) prevents the exposure ofthese proteins to immune cells Therefore, ER-misfolded proteins aber-rantly transported to the cell surface by MHC class II molecules might appear

as “neo-self ”-antigens and induce an abnormal immune response (Fig 2).Indeed, XBP-1, a transcription factor induced by the unfolded proteinresponse, was cloned originally from plasma cells in synovial membranes ofpatients with RA (Iwakoshi et al., 2003), suggesting high levels of unfoldedprotein production in these cells Autoantibodies against citrullinated pro-teins are also generated in RA Interestingly, citrullination is known to causeprotein misfolding (Tarcsa et al., 1996) Therefore, citrullination-inducedconformational changes in proteins might augment the association of auto-antigens with MHC class II molecules Furthermore, MHC class II expres-sion is strongly increased in the synovial membranes of patients with RA(Feldmann et al., 1988; Klareskog, Forsum, Scheynius, Kabelitz, &Wigzell, 1982) Because plasma cells expressing low levels of MHC class

II molecules produce large amounts of IgG, misfolded IgG heavy chainsmight associate with MHC class II molecules encoded by RA-susceptiblealleles in certain conditions that induce upregulated MHC class II expression

on plasma cells; the resulting complexes could induce production of antibody against IgG Similarly, β2-glycoprotein I is mainly produced inhepatocytes, which do not express MHC class II molecules in steady state

auto-If APS-susceptible MHC class II molecules that preferentially bind to glycoprotein I are expressed on hepatocytes in response to inflammation

β2-or infection,β2-glycoprotein I will associate with these molecules and thustrigger autoantibody production

Certain misfolded proteins may associate constitutively with MHCclass II molecules on B cells or dendritic cells However, autoantibodiesagainst these misfolded proteins are not produced in steady state; immunecells are exposed constitutively to these complexes, a process that appears

to have induced tolerance Indeed, transgenic mice expressing MHC class

II molecules on pancreatic β cells did not exhibit β cell autoimmunity(Lo et al., 1988; Sarvetnick, Liggitt, Pitts, Hansen, & Stewart, 1988).Because MHC class II expression is not inducible in these transgenic mice,even if certain β cell-specific misfolded proteins are presented on MHCclass II molecules, the proteins associated with MHC class II moleculesare always presented to immune cells and may not be recognized as

“neo-self ”-antigens Analyses of mice harboring an inducible MHC class

II transgene might provide valuable information about the function of rantly expressed MHC class II molecules

aber-The CIITA transcription factor is involved in the expression of both MHCclass II molecules and the invariant chain (Reith, LeibundGut-Landmann, &Waldburger, 2005) Therefore, the invariant chain is expressed in mostMHC class II-expressing cells, where it blocks the binding of ER-misfoldedproteins to MHC class II molecules However, MHC class II and invariantchain gene transcription is regulated differentially, and thus the expression ofthese molecules is not always equivalent (Paul et al., 2011) In addition, theaffinity of MHC class II molecules for the invariant chain differs depending

on the encoding MHC class II allele (Patil et al., 2001) Therefore, the amounts

of misfolded proteins, invariant chain, and MHC class II molecules, as well asthe MHC class II allele, seem to determine the efficiency of the associationbetween misfolded proteins and MHC class II molecules

7.2 Misfolded Protein Complexed with MHC Class II Molecules

as Primary Autoantigens for Autoantibodies

It remains unclear how autoantibodies specific for autoantigens presented onMHC class II molecules are produced Given the presence of specific auto-antibodies against these complexes, it is likely that the complexes themselvesinduce autoantibody production However, it is uncertain whether the self-antigens presented on MHC class II molecules initiate local antibodyresponses in nonlymphoid tissues because the germinal center usually isrequired for antibody responses (Klein & Dalla-Favera, 2008) On the otherhand, cell surface MHC class II molecules are known to be released fromcells as exosomes (Thery, Ostrowski, & Segura, 2009) and have beendetected as such in serum (Almqvist, Lonnqvist, Hultkrantz, Rask, &Telemo, 2008; Karlsson et al., 2001; Taylor, Akyol, & Gercel-Taylor,2006) Therefore, when self-antigens are expressed in complex withMHC class II molecules in certain tissues, the complexes may be releasedfrom the cells as exosomes, which might subsequently induce the produc-tion of specific antibodies against the complexes in lymphoid tissues.Most autoantibodies are detected using self-antigens immobilized on plates

or microbeads, suggesting that MHC class II molecules might not be requiredfor autoantibody recognition However, protein immobilization causes signif-icant conformational changes For example, RF does not bind native IgG insera but does bind immobilized IgG Similarly, autoantibodies against β2-glycoprotein I do not bind nativeβ2-glycoprotein I in sera but do bind toβ2-glycoprotein I when immobilized on negatively charged plates In addi-tion, most autoantibodies can detect target antigens when assayed by Westernblot analysis, indicating that they recognize denatured forms of autoantigens

These observations are compatible with the fact that autoantibodies aredirected against misfolded proteins presented on MHC class II molecules.Autoantibodies from most patients with APS recognizeβ2-glycoprotein

I presented on MHC class II molecules, whereas only some patientspossess autoantibodies against plate-bound β2-glycoprotein I Therefore,β2-glycoprotein I presented on MHC class II molecules seems to be the pri-mary target antigen for APS autoantibodies Most autoantibodies are theresult of a somatic hypermutation process that increases the affinity for auto-antigens (Rajewsky, 1996) Because somatic hypermutation is not observed

in naı¨ve B cells, the antigen specificities of autoantibodies that areengineered in vitro to restore the codons present in the germline Ig gene willprovide information about the original antigens that stimulated naı¨ve B cells

to produce autoantibodies A recent analysis indicated that reverted autoantibodies from pemphigus patients did not recognize theautoantigen desmoglein-3, suggesting that this autoantigen did not induceautoantibody production (Di Zenzo et al., 2012) However, somegermline-reverted autoantibodies still recognize autoantigens when pres-ented on MHC class II molecules (Hui Jin & Hisashi Arase, unpublishedobservation) Therefore, autoantigens presented on MHC class II moleculesmight be the primary target antigens that induced autoantibody production.Although most autoantibodies are directed against denatured autoantigens,some autoantibodies recognize native autoantigens Because epitope spread-ing affects antibody diversity, it is possible that autoantibodies raised againstmisfolded proteins presented on MHC class II molecules might haveacquired reactivity against native autoantigens through epitope spreading.The molecular mimicry of autoantigens by microbial antigens is alsoinvolved in the production of some autoantibodies (Munz, Lunemann,Getts, & Miller, 2009) Considering that the antigenicity of misfolded auto-antigens is more similar to that of native autoantigens than of microbial anti-gens, misfolded proteins presented on MHC class II molecules may possibly

germline-be involved in the production of autoantibodies against native autoantigens

8 MISFOLDED PROTEINS PRESENTED ON MHC CLASS IIMOLECULES AS A THERAPEUTIC TARGET FOR

AUTOIMMUNE DISEASES

As misfolded proteins aberrantly rescued from protein degradation byMHC class II molecules might be involved in the pathogenesis of autoim-mune diseases, blocking the association of misfolded proteins with MHC

class II molecules would be a good candidate treatment for autoimmune eases The aberrant MHC class II expression observed in autoimmune-diseased tissues seems to result from stimulation by cytokines such asIFN-γ Blocking the MHC class II expression induced by cytokine stimu-lation would therefore effectively treat autoimmune diseases (Miller,Maher, & Young, 2009) HMG-CoA reductase inhibitors, or statins, areamong the drugs used to reduce serum cholesterol levels Statins have beenreported to reduce inflammation in some autoimmune diseases such as RA,APS, and Sj€ogren’s syndrome (Greenwood, Steinman, & Zamvil, 2006;Khattri & Zandman-Goddard, 2013) Although the exact mechanismsremain unclear, statins appear to reduce IFN-γ-induced MHC class IIexpression on human endothelial cells in vitro (Kwak, Mulhaupt, Myit, &Mach, 2000; Youssef et al., 2002) The anti-inflammatory function of statinsmight also include inhibiting the association of misfolded proteins withMHC class II molecules The development of a strong and specific inhibitorwith which to block the association of misfolded proteins with aberrantlyexpressed MHC class II molecules represents a new target in autoimmunedisease therapy.

dis-9 CONCLUDING REMARKS

The MHC class II locus is the gene most strongly associated with tibility to many autoimmune diseases Extensive analyses of misfolded proteinsaberrantly rescued from protein degradation by MHC class II molecules havenot only revealed that these proteins are specific targets for autoantibodies buthave also suggested that autoantibody binding to these proteins may explain theautoimmune disease susceptibility conferred by certain MHC class II alleles.Therefore, aberrant MHC class II expression on certain tissues or cells mightinduce autoimmune disease An understanding the molecular factors thatinduce aberrant MHC class II expression would be quite important to an under-standing of the causes of autoimmune diseases On the other hand, the physi-ological functions of the aberrantly expressed MHC class II molecules onnonimmune cells remain unclear If aberrant MHC class II expression on non-immune cells is solely responsible for diseases involvement, the pathway bywhich MHC class II is expressed on nonimmune cells will be lost in the course

suscep-of evolution Although MHC class II molecules expressed on nonimmune cellscannot evoke helper T cell responses because of the lack of costimulatory mol-ecule expression, these aberrantly expressed MHC class II molecules mightconfer certain benefits to maintain homeostasis Further analyses of the

misfolded proteins rescued from protein degradation and transported to the cellsurface by MHC class II molecules will reveal currently unknown mechanisms

of immunity in both normal and disease situations

ACKNOWLEDGMENTS

We thank Prof Lewis L Lanier for critical reading of our manuscript This work was partially supported by JSPS KAKENHI Grant Numbers (15K15131, 15H02545, 26117714 and 24115005), the Practical Research Project for Allergic Diseases and Immunology from Japan Agency for Medical Research and development, AMED, The Naito Foundation, The Tokyo Biochemical Research Foundation, The Uehara Memorial Foundation and Terumo Life Science Foundation.

REFERENCES

Agar, C., van Os, G M., Morgelin, M., Sprenger, R R., Marquart, J A., Urbanus, R T.,

et al (2010) β2-glycoprotein I can exist in 2 conformations: Implications for our standing of the antiphospholipid syndrome Blood, 116, 1336–1343.

under-Ahmed, A R., Spigelman, Z., Cavacini, L A., & Posner, M R (2006) Treatment of phigus vulgaris with rituximab and intravenous immune globulin The New England Journal of Medicine, 355, 1772–1779.

pem-Aichinger, G., Karlsson, L., Jackson, M R., Vestberg, M., Vaughan, J H., Teyton, L.,

et al (1997) Major histocompatibility complex class II-dependent unfolding, transport, and degradation of endogenous proteins The Journal of Biological Chemistry, 272,

29127 –29136.

Almqvist, N., Lonnqvist, A., Hultkrantz, S., Rask, C., & Telemo, E (2008) Serum-derived exosomes from antigen-fed mice prevent allergic sensitization in a model of allergic asthma Immunology, 125, 21 –27.

Anderson, M S., Swier, K., Arneson, L., & Miller, J (1993) Enhanced antigen presentation

in the absence of the invariant chain endosomal localization signal The Journal of imental Medicine, 178, 1959–1969.

Exper-Anolik, J H., Barnard, J., Cappione, A., Pugh-Bernard, A E., Felgar, R E., Looney, R J.,

et al (2004) Rituximab improves peripheral B cell abnormalities in human systemic lupus erythematosus Arthritis and Rheumatism, 50, 3580–3590.

Appleman, L J., & Boussiotis, V A (2003) T cell anergy and costimulation Immunological Reviews, 192, 161–180.

Arosa, F A., Santos, S G., & Powis, S J (2007) Open conformers: The hidden face of MHC-I molecules Trends in Immunology, 28, 115–123.

Ballardini, G., Mirakian, R., Bianchi, F B., Pisi, E., Doniach, D., & Bottazzo, G F (1984) Aberrant expression of HLA-DR antigens on bileduct epithelium in primary biliary cirrhosis: Relevance to pathogenesis Lancet, 2, 1009 –1013.

Bas de Laat, H., Derksen, R H., & de Groot, P G (2004) β2-glycoprotein I, the playmaker

of the antiphospholipid syndrome Clinical Immunology, 112, 161 –168.

Batista, F D., & Harwood, N E (2009) The who, how and where of antigen presentation to

B cells Nature Reviews Immunology, 9, 15 –27.

Bottazzo, G F., Pujol-Borrell, R., Hanafusa, T., & Feldmann, M (1983) Role of aberrant HLA-DR expression and antigen presentation in induction of endocrine autoimmunity Lancet, 2, 1115–1119.

Busch, R., Cloutier, I., Sekaly, R P., & Hammerling, G J (1996) Invariant chain protects class II histocompatibility antigens from binding intact polypeptides in the endoplasmic reticulum The EMBO Journal, 15, 418–428.

Cambridge, G., Leandro, M J., Teodorescu, M., Manson, J., Rahman, A., Isenberg, D A.,

et al (2006) B cell depletion therapy in systemic lupus erythematosus: Effect on autoantibody and antimicrobial antibody profiles Arthritis and Rheumatism, 54, 3612–3622.

Castellino, F., Zappacosta, F., Coligan, J E., & Germain, R N (1998) Large protein ments as substrates for endocytic antigen capture by MHC class II molecules Journal of Immunology, 161, 4048–4057.

frag-Davenport, M P., Quinn, C L., Chicz, R M., Green, B N., Willis, A C., Lane, W S.,

et al (1995) Naturally processed peptides from two disease-resistance-associated DR13 alleles show related sequence motifs and the effects of the dimorphism at position

HLA-86 of the HLA-DRβ chain Proceedings of the National Academy of Sciences of the United States of America, 92, 6567–6571.

de Laat, B., Derksen, R H., van Lummel, M., Pennings, M T., & de Groot, P G (2006) Pathogenic anti-β2-glycoprotein I antibodies recognize domain I of β2-glycoprotein

I only after a conformational change Blood, 107, 1916 –1924.

Di Zenzo, G., Di Lullo, G., Corti, D., Calabresi, V., Sinistro, A., Vanzetta, F., et al (2012) Pemphigus autoantibodies generated through somatic mutations target the desmoglein-3 cis-interface Journal of Clinical Investigation, 122, 3781 –3790.

Dorner, T., Egerer, K., Feist, E., & Burmester, G R (2004) Rheumatoid factor revisited Current Opinion in Rheumatology, 16, 246–253.

Erkan, D., Vega, J., Ramon, G., Kozora, E., & Lockshin, M D (2013) A pilot open-label phase II trial of rituximab for non-criteria manifestations of antiphospholipid syndrome Arthritis and Rheumatism, 65, 464–471.

Feldmann, M., Kissonerghis, A M., Buchan, G., Brennan, F., Turner, M., Haworth, C.,

et al (1988) Role of HLA class II and cytokine expression in rheumatoid arthritis dinavian Journal of Rheumatology Supplement, 76, 39–46.

Scan-Galli, M., Barbui, T., Zwaal, R F., Comfurius, P., & Bevers, E M (1993) Antiphospholipid antibodies: Involvement of protein cofactors Haematologica, 78, 1–4.

Germain, R N (2011) Uncovering the role of invariant chain in controlling MHC class II antigen capture Journal of Immunology, 187, 1073–1075.

Giacomini, P., Beretta, A., Nicotra, M R., Ciccarelli, G., Martayan, A., Cerboni, C.,

et al (1997) HLA-C heavy chains free of β2-microglobulin: Distribution in normal tissues and neoplastic lesions of non-lymphoid origin and interferon- γ responsiveness Tissue Antigens, 50, 555 –566.

Gottlieb, A B., Lifshitz, B., Fu, S M., Staiano-Coico, L., Wang, C Y., & Carter, D M (1986) Expression of HLA-DR molecules by keratinocytes, and presence of Langerhans cells in the dermal infiltrate of active psoriatic plaques The Journal of Experimental Med- icine, 164, 1013–1028.

Greenwood, J., Steinman, L., & Zamvil, S S (2006) Statin therapy and autoimmune disease: From protein prenylation to immunomodulation Nature Reviews Immunology, 6, 358–370.

Griffiths, M M., & Remmers, E F (2001) Genetic analysis of collagen-induced arthritis in rats: A polygenic model for rheumatoid arthritis predicts a common framework of cross- species inflammatory/autoimmune disease loci Immunological Reviews, 184, 172–183 Heemstra, K A., Toes, R E., Sepers, J., Pereira, A M., Corssmit, E P., Huizinga, T W.,

et al (2008) Rituximab in relapsing Graves’ disease, a phase II study European Journal of Endocrinology, 159, 609–615.

Iwakoshi, N N., Lee, A H., Vallabhajosyula, P., Otipoby, K L., Rajewsky, K., & Glimcher, L H (2003) Plasma cell differentiation and the unfolded protein response intersect at the transcription factor XBP-1 Nature Immunology, 4, 321–329.

Jacobi, A M., & Dorner, T (2010) Current aspects of anti-CD20 therapy in rheumatoid arthritis Current Opinion in Pharmacology, 10, 316 –321.

Jaffe, E A., Armellino, D., Lam, G., Cordon-Cardo, C., Murray, H W., & Evans, R L (1989) IFN-γ and IFN-α induce the expression and synthesis of Leu 13 antigen by cul- tured human endothelial cells Journal of Immunology, 143, 3961–3966.

Jiang, Y., Arase, N., Kohyama, M., Hirayasu, K., Suenaga, T., Jin, H., et al (2013) Transport

of misfolded endoplasmic reticulum proteins to the cell surface by MHC class II cules International Immunology, 25, 235–246.

mole-Jin, H., Arase, N., Hirayasu, K., Kohyama, M., Suenaga, T., Saito, F., et al (2014) antibodies to IgG/HLA class II complexes are associated with rheumatoid arthritis sus- ceptibility Proceedings of the National Academy of Sciences of the United States of America, 111, 3787–3792.

Auto-Joly, P., Mouquet, H., Roujeau, J C., D’Incan, M., Gilbert, D., Jacquot, S., et al (2007).

A single cycle of rituximab for the treatment of severe pemphigus The New England nal of Medicine, 357, 545–552.

Jour-Karlsson, M., Lundin, S., Dahlgren, U., Kahu, H., Pettersson, I., & Telemo, E (2001) erosomes” are produced by intestinal epithelial cells European Journal of Immunology, 31,

Lechler, R., Aichinger, G., & Lightstone, L (1996) The endogenous pathway of MHC class

II antigen presentation Immunological Reviews, 151, 51–79.

Lo, D., Burkly, L C., Widera, G., Cowing, C., Flavell, R A., Palmiter, R D., et al (1988) Diabetes and tolerance in transgenic mice expressing class II MHC molecules in pancre- atic beta cells Cell, 53, 159–168.

Matsumoto, I., Maccioni, M., Lee, D M., Maurice, M., Simmons, B., Brenner, M.,

et al (2002) How antibodies to a ubiquitous cytoplasmic enzyme may provoke joint-specific autoimmune disease Nature Immunology, 3, 360 –365.

Matsumoto, I., Staub, A., Benoist, C., & Mathis, D (1999) Arthritis provoked by linked

T and B cell recognition of a glycolytic enzyme Science, 286, 1732 –1735.

McNeil, H P., Simpson, R J., Chesterman, C N., & Krilis, S A (1990) Anti-phospholipid antibodies are directed against a complex antigen that includes a lipid-binding inhibitor

of coagulation: β2-glycoprotein I (apolipoprotein H) Proceedings of the National Academy

of Sciences of the United States of America, 87, 4120–4124.

Meusser, B., Hirsch, C., Jarosch, E., & Sommer, T (2005) ERAD: The long road to destruction Nature Cell Biology, 7, 766–772.

Miller, C H., Maher, S G., & Young, H A (2009) Clinical use of interferon-γ Annals of the New York Academy of Sciences, 1182, 69–79.

Munz, C., Lunemann, J D., Getts, M T., & Miller, S D (2009) Antiviral immune responses: Triggers of or triggered by autoimmunity? Nature Reviews Immunology, 9, 246–258 Navarra, S V., Guzman, R M., Gallacher, A E., Hall, S., Levy, R A., Jimenez, R E.,

et al (2011) Efficacy and safety of belimumab in patients with active systemic lupus erythematosus: A randomised, placebo-controlled, phase 3 trial Lancet, 377, 721–731 Ohkuri, T., Ueda, T., Tsurumaru, M., & Imoto, T (2001) Evidence for an initiation site for hen lysozyme folding from the reduced form using its dissected peptide fragments Protein Engineering, 14, 829 –833.

Patil, N S., Pashine, A., Belmares, M P., Liu, W., Kaneshiro, B., Rabinowitz, J.,

et al (2001) Rheumatoid arthritis (RA)-associated HLA-DR alleles form less stable complexes with class II-associated invariant chain peptide than non-RA-associated HLA-DR alleles Journal of Immunology, 167, 7157–7168.

Paul, P., van den Hoorn, T., Jongsma, M L., Bakker, M J., Hengeveld, R., Janssen, L.,

et al (2011) A Genome-wide multidimensional RNAi screen reveals pathways ling MHC class II antigen presentation Cell, 145, 268–283.

control-Petkova, S B., Konstantinov, K N., Sproule, T J., Lyons, B L., Awwami, M A., & Roopenian, D C (2006) Human antibodies induce arthritis in mice deficient in the low-affinity inhibitory IgG receptor Fc gamma RIIB The Journal of Experimental Medicine, 203, 275–280.

Pober, J S., Gimbrone, M A., Jr., Cotran, R S., Reiss, C S., Burakoff, S J., Fiers, W.,

et al (1983) Ia expression by vascular endothelium is inducible by activated T cells and by human γ interferon The Journal of Experimental Medicine, 157, 1339–1353.

Qi, H., Egen, J G., Huang, A Y., & Germain, R N (2006) Extrafollicular activation of lymph node B cells by antigen-bearing dendritic cells Science, 312, 1672 –1676 Rajewsky, K (1996) Clonal selection and learning in the antibody system Nature, 381,

751 –758.

Raychaudhuri, S., Sandor, C., Stahl, E A., Freudenberg, J., Lee, H S., Jia, X., et al (2012) Five amino acids in three HLA proteins explain most of the association between MHC and seropositive rheumatoid arthritis Nature Genetics, 44, 291–296.

Reith, W., LeibundGut-Landmann, S., & Waldburger, J M (2005) Regulation of MHC class II gene expression by the class II transactivator Nature Reviews Immunology, 5, 793–806.

Sarvetnick, N., Liggitt, D., Pitts, S L., Hansen, S E., & Stewart, T A (1988) dependent diabetes mellitus induced in transgenic mice by ectopic expression of class

Insulin-II MHC and interferon-gamma Cell, 52, 773–782.

Sibilio, L., Martayan, A., Setini, A., Fraioli, R., Fruci, D., Shabanowitz, J., et al (2005) Impaired assembly results in the accumulation of multiple HLA-C heavy chain folding intermediates Journal of Immunology, 175, 6651–6658.

Sieb, J P (2014) Myasthenia gravis: An update for the clinician Clinical and Experimental Immunology, 175, 408–418.

Stam, N J., Spits, H., & Ploegh, H L (1986) Monoclonal antibodies raised against tured HLA-B locus heavy chains permit biochemical characterization of certain HLA-C locus products Journal of Immunology, 137, 2299 –2306.

dena-Tanimura, K., Jin, H., Suenaga, T., Morikami, S., Arase, N., Kishida, K., et al (2015) glycoprotein I / HLA class II complexes are novel autoantigens in antiphospholipid syn- drome Blood, 125, 2835–2844.

β2-Tarcsa, E., Marekov, L N., Mei, G., Melino, G., Lee, S C., & Steinert, P M (1996) Protein unfolding by peptidylarginine deiminase Substrate specificity and structural relationships

of the natural substrates trichohyalin and filaggrin The Journal of Biological Chemistry, 271, 30709–30716.

Taylor, D D., Akyol, S., & Gercel-Taylor, C (2006) Pregnancy-associated exosomes and their modulation of T cell signaling Journal of Immunology, 176, 1534–1542.

Thery, C., Ostrowski, M., & Segura, E (2009) Membrane vesicles as conveyors of immune responses Nature Reviews Immunology, 9, 581–593.

Thurlings, R M., Vos, K., Wijbrandts, C A., Zwinderman, A H., Gerlag, D M., & Tak, P P (2008) Synovial tissue response to rituximab: Mechanism of action and iden- tification of biomarkers of response Annals of the Rheumatic Diseases, 67, 917–925 Todd, I., Pujol-Borrell, R., Hammond, L J., Bottazzo, G F., & Feldmann, M (1985) Inter- feron- γ induces HLA-DR expression by thyroid epithelium Clinical and Experimental Immunology, 61, 265 –273.

Wilson, W A., Gharavi, A E., Koike, T., Lockshin, M D., Branch, D W., Piette, J C.,

et al (1999) International consensus statement on preliminary classification criteria for definite antiphospholipid syndrome: Report of an international workshop Arthritis and Rheumatism, 42, 1309–1311.

Youssef, S., Stuve, O., Patarroyo, J C., Ruiz, P J., Radosevich, J L., Hur, E M.,

et al (2002) The HMG-CoA reductase inhibitor, atorvastatin, promotes a Th2 bias and reverses paralysis in central nervous system autoimmune disease Nature, 420, 78–84.

University of Cincinnati College of Medicine, Cincinnati, Ohio, USA

1 Corresponding author: e-mail address: marie-dominique.filippi@cchmc.org

Contents

4.2 Neutrophil-Intrinsic Control of Transcellular Migration 46

at the endothelial cell junction (paracellular) or directly through the endothelial cell body (transcellular) The extravasation cascade is controlled by series of engagement

of various adhesive modules, which result in activation of bidirectional signals to trophils and endothelial cells for adequate cellular response This review will focus on recent advances in our understanding of mechanism of leukocyte crawling and diape- desis, with an emphasis on leukocyte –endothelial interactions and the signaling

neu-Advances in Immunology, Volume 129 # 2016 Elsevier Inc.

ISSN 0065-2776 All rights reserved 25

pathways they transduce to determine the mode of diapedesis, junctional or junctional I will also discuss emerging evidence highlighting key differences in the two modes of diapedesis and why it is clinically important to understand specificity

non-in the regulation of diapedesis.

1 INTRODUCTION

Neutrophils are the first line of cellular defense against invadingmicroorganisms and play a central role in innate immunity and inflammatoryprocesses (Ley et al., 2007; Phillipson & Kubes, 2011) These white bloodcells circulate into the blood stream but must cross the endothelial barrier toreach inflamed tissues This rapid migration from the blood to site of infec-tions is critical for pathogen elimination and tissue repair in response to acuteinflammation However, when uncontrolled, excessive accumulation ofactivated neutrophils into tissue leads to tissue damage during hyper-inflammatory disorders, including acute lung injury, multiple organ failuresyndrome, vascular inflammation, or arthritis

The initial step of the inflammatory response is a reorganization of theendothelial cell surface to capture floating neutrophils The release of inflam-matory cytokines and bacteria-derived peptides stimulates the upregulation

of adhesive molecules, on the endothelial luminal surface, which locally mote weak and transient adhesive interactions between neutrophils and theendothelium, known as “rolling.” The deposition of chemokines on theendothelial luminal surface, then, triggers the activation of leukocyteintegrins that promote their firm adhesion and arrest via interactions withtheir ligand counterpart expressed on the endothelial surface Subsequently,the activated neutrophils further respond to chemokines and undergo adrastic cell shape change from round to flat and highly polarized—defining

pro-a cell “front” pro-and “repro-ar or uropod.” The pro-adoption of this polpro-arized shpro-apepermits the cells to migrate or crawl on the endothelial lumen surface tofind a nearby site to cross the endothelial cells lining the blood vessels Thislatter process is called diapedesis Once passed the endothelial barrier,the cells must cross the pericyte layer within the venular basal membranebefore to reach the inflamed interstitial tissues This sequence of events rep-resents the paradigm of the extravasation cascade and is summarized inFig 1(Ley et al., 2007; Muller, 2011; Nourshargh et al., 2010) Althoughlong debated, it is now accepted that leukocytes can breach the endothelialbarrier by two distinct routes (Carman et al., 2007; Feng et al., 1998;

Mamdouh, Mikhailov, & Muller, 2009; Marchesi, 1961; Millan et al., 2006;Muller, 2011; Phillipson et al., 2006; Yang et al., 2005) The leukocytes canfind their way between two endothelial cells (the paracellular route) This

is facilitated by the disruption of endothelial vascular endothelial cadherin contacts, which form a paracellular gap through which the cellsmigrate Alternatively, the leukocytes can transmigrate directly throughindividual endothelial cell (the transcellular route) In this case, the endothe-lial cell junctions remain intact Instead, the membrane of neutrophils andendothelial cells fuse and remodel into a transcellular channel, forming

(VE)-a p(VE)-ath for leukocytes The leukocyte extr(VE)-av(VE)-as(VE)-ation is (VE)-a highly regul(VE)-atedprocess that involves the engagement of complex interactions betweenthe leukocyte and the endothelium, including via selectins, integrins,intercellular adhesion molecule (ICAM), junctional adhesion molecule(JAM), and platelet endothelial cell adhesion molecule (PECAM) Theseinteractions are well coordinated and are known to occur in a sequentialmanner (Muller, 2011) Beyond promoting cell–cell interactions, adhesive

Blood vessel

Tissue Endothelial cells

Leukocytes

Pericytes Basementmembrane

Capture rolling

Adhesion Crawling Diapedesis

Paracellular/transcellular

Floating

LFA-1–ICAM1/2 VLA4–VCAM L-Selectin

E/P-Sepectin-PSGL1 Mac-1–ICAM1

Mac-1–ICAM PECAM CD99 JAM ESAM

Adhesive molecules and ligands

Figure 1 The leukocyte extravasation cascade is controlled by sequential adhesive interactions between leukocytes and endothelial cells This schema depicts various steps and the adhesive molecules that are involved at each step The neutrophil extrav- asation cascade involves a sequence of tethering and rolling along the endothelium, followed by firm adhesion and arrest onto the endothelium Subsequently, neutrophils undergo lateral migration or crawling on endothelial cells to find a permissive site for transmigration It should be noted that subsequent to moving across the endothelial barrier, leukocytes undergo abluminal crawling between endothelial cells and pericytes before crossing the basement membrane and migrating within interstitial tissues ( Nourshargh, Hordijk, & Sixt, 2010 ).

molecules send bidirectional signaling from the leukocytes to the endothelialcells and vice versa that participate in the establishment of leukocyte polarity,their ability to crawl on the endothelium, and that are instrumental inguiding the mode of leukocyte diapedesis (Herter & Zarbock, 2013;Muller, 2011).

This review will summarize key mechanisms and leukocyte signalingpathways that control the extravasation cascade It will focus on emergingevidences of new pathways that specifically control transcellular migration,underscoring that, after all, paracellular and transcellular are regulated byseparate mechanisms It will then discuss the impact transmigration routemay have on the immune response and why it is clinically important tounderstand specificity in the regulation of diapedesis

2 LEUKOCYTE INTERACTIONS WITH THE ENDOTHELIUMThe extravasation cascade has been well studied, in particular in thecontext of paracellular migration It is mediated by a series of complexand sequential interactions between the leukocytes and the endothelial api-cal surface via various adhesion receptors These receptors have been exten-sively reviewed elsewhere (Ley et al., 2007; Luo, Carman, & Springer, 2007;Muller, 2013; Nourshargh & Alon, 2014) Endothelial (E)- and platelet(P)-selectin that are expressed on the endothelial apical surface upon inflam-matory insults capture leukocytes and mediate their rolling onto theendothelium via leukocyte (L)-selectin Subsequently, firm adhesion iscontrolled by adhesion receptors of the immunoglobulin family, namelyleukocyte integrins (LFA-1 [lymphocyte function-associated antigen-1 also

αLβ2integrin or CD11a/CD18], Mac-1 [macrophage-1 antigen alsoαMβ2integrin or CD11b/CD18], and VLA-4 [very late antigen-4 also α4β1integrin]), which bind to their endothelial ligands, including ICAM(ICAM-1 and -2) and vascular cell adhesion molecule 1 (VCAM-1), respec-tively Following firm adhesion, leukocytes adopt a polarized shape andcrawl onto the endothelial apical surface in search for a permissive site ofextravasation Locomotion of leukocytes is strictly dependent onβ2integrins(Phillipson et al., 2006; Schenkel, Mamdouh, & Muller, 2004) In neutro-phils, which express both LFA-1 and Mac-1, genetic ablation of LFA-1 andMac-1 has established that LFA-1 and Mac-1 play sequential roles in theextravasation cascade LFA-1 mediates neutrophil firm adhesion whereasMac-1 controls their crawling onto the endothelial apical surface(Phillipson et al., 2006; Sumagin et al., 2010) Subsequently, the leukocytes

engage a sequence of interactions to cross the endothelial barrier, whichinvolve JAM-1/A/C (junctional adhesion molecule-1/A/C), PECAM-1(platelet endothelial cell adhesion molecule), CD99, and ESAM (endothelialcell adhesion molecule; Muller, 2013; Nourshargh et al., 2010) In thischapter, I will mostly focus on mechanisms of leukocyte crawling anddiapedesis.

2.1 Docking Structures and Crawling

Arrest of leukocytes on the endothelium is mediated by a shift from mediate affinity to high-affinity (HA) β2-integrins (Shaw et al., 2004).Adhesion molecules including leukocyte integrins and endothelial ICAMredistribute into dense clusters located at the leukocyte–endothelial cellinterface and surrounding the cells (Shaw et al., 2004) These dense clustersstabilize and strengthen leukocyte–endothelial cell interactions Followingarrest and firm anchorage onto apical endothelial surface, the leukocytes flat-ten and adopt a highly polarized shape enabling their lateral migration orcrawling for several microns on the vascular endothelium in search for per-missive site of transmigration (Phillipson et al., 2006) Leukocyte motility orcrawling depends on asymmetric rearrangement of the leukocyte cytoskel-eton in response to chemokines, which is coordinated with a dynamic cycle

inter-of assembly and disassembly inter-of adhesive points binding the leukocyte to theendothelium During this process, filamentous actin (F-actin) polymerizesasymmetrically forming the cell leading edge, and providing the protrusiveforces to propel the cell membrane forward, whereas a network of actomy-osin assembles along the cell sides and the trailing edge or uropod, and pre-vents lateral membrane protrusions to occur (Ridley et al., 2003; Stephens,Milne, & Hawkins, 2008; Williams et al., 2011) Maintaining cell polarityand a single leading edge are critical for persistent migration in one direction

At the same time, an active reorganization of the cell plasma membraneoccurs and involves the polarized redistribution of membrane receptors,including integrins In migrating neutrophils, the plasma membranebecomes organized into lipid-rich domains that are different at the frontand at the rear (Bodin & Welch, 2005; Pierini et al., 2003) The transmem-brane receptor CD45 accumulates at the front, whereas the uropod and sides

of the cells are enriched in CD44, L-selectin, heavily glycosylated proteins—e.g., PSGL-1 and integrins (Barreiro et al., 2007; Bodin & Welch, 2005;Pierini et al., 2003; Zhang et al., 2006) During active migration on endo-thelial cells, LFA-1 and Mac-1 are being excluded from the protrusive

leading edge; instead, they actively redistribute into punctuated regions ofclustered integrins that are located underneath the cells as well as alongthe sides and at the uropod of the leukocytes (Cinamon et al., 2004;Kumar et al., 2012; Smith et al., 2005; Zhang et al., 2006) These spatialchanges are also accompanied by changes in both affinity and avidity ofthe integrins for their ligands In lymphocytes, HA-LFA-1 can be seenenriched toward the uropod In addition, dense clusters of HA-LFA-1–ICAM develop in the ventral part of the leukocytes in close contact withthe endothelial apical surface (Shulman et al., 2009; Smith et al., 2005).The clusters of HA-LFA-1, and Mac-1, generated by crawling leukocytesare dynamic focal assemblies to modulate the strength of leukocyte/endothelial interactions and are necessary for leukocyte crawling on vascularendothelium (Shulman et al., 2009; Smith et al., 2005) Invitro studies haveshown that blocking Mac-1 or CD18 with monoclonal antibodies signifi-cantly blocked monocyte crawling onto HUVEc (Shulman et al., 2009).Intravital microscopy in vivo demonstrated that CD11b-null neutrophilsfailed to crawl in the vessel lumen (Phillipson et al., 2006).

Interestingly, integrins not only provide dynamic adhesion points, theyare also important to regulate the intracellular cytoskeleton and to maintainleukocyte polarization during crawling Live imaging indicated thatmonocytes treated with functional blocking antibody to CD18 wouldadhere, polarize, and extent protrusions; they would then often retractexisting protrusions and extend new protrusions in several directions(Schenkel et al., 2004) Monocytes would rotate on their uropod, unable

to reach endothelial junctions (Schenkel et al., 2004) In neutrophils,CD11b-deficiency caused the cells to extent inappropriate lateral protru-sions, which induced a systematic change in direction (Szczur, Zheng, &Filippi, 2009) This inability to travel in one direction was due to defectiveassembly of the actomyosin network at the uropod, indicating that Mac-1plays a specialized role in maintain the cell polarity axis (Szczur et al.,2009) Hence, ICAM-β2 integrin clusters are critical for maintaining leu-kocyte polarity and efficient crawling These studies underscore the criticalrole integrins play during leukocyte locomotion that goes well beyondtheir role in attachment

In addition to forming dense clusters underneath the cells, ICAM clustersare also consistently seen surrounding the cells prior diapedesis (Fig 2A;Barreiro et al., 2002; Carman et al., 2003; Millan et al., 2006; Shaw

et al., 2004) These clusters are regulated by endothelial cell cortical skeleton, which regulates the membrane localization of adhesion receptors

cyto-through ezrin, radixin, and moesin (ERM) proteins, to anchor the cytes to the endothelial surface (Barreiro et al., 2002) The recruitment ofVCAM or ICAM to cell–cell contact on the apical endothelial membrane

leuko-is dependent on endothelial tetraspanin (Barreiro et al., 2002, 2005) and tactin (Yang et al., 2006) Some studies have also reported the existence ofactin-rich microvilli-containing ICAM clusters arising from the endothelial

cor-Transcellular Paracellular

B

A

Mac-1/LFA-1 Endothelial F-actin ICAM

PECAM JAM

VVO Caveola

MMP?

LBRC

Src WASp Rap1b

PI3K Akt

Rap1b

PI3K Akt

Mac-1

Membrane fusion

Figure 2 The leukocyte diapedesis (A) Representation of the transcellular cup made of clusters of leukocyte integrins and endothelial ICAM Some studies have observed the formation of actin-microvilli embracing the transmigrating leukocyte (B) It is now accepted that leukocytes can transmigrate at the junction between two endothelial cells (paracellular migration depicted in left panel) or directly though endothelial cells (transcellular migration depicted in right panel) Paracellular migration is accompanied

by the disruption of the endothelial cell junction to form a gap through which the cells migrate This is accompanied by the reorganization of an adhesive platform and the recycling of adhesive molecules via the LBRC On the other hand, during transcellular migration, the endothelial cell junctions remain intact Instead, neutrophil –endothelial cell contacts fuse (represented in blue) and remodel into a transcellular channel forming

a path for leukocytes This necessitates the recruitment of actin-rich membrane, enriched caveola and vesicle, vesicular vacuolar organelles as well as the recruitment of various adhesive molecules via the LBRC In addition, the involvement of MMP activity is likely and may help remodeling the leukocyte –endothelial cell interaction to facilitate the formation of the transcellular channel Several signaling mechanisms important for invasive protrusions and transcellular have been identified High ICAM density, high integrin signaling, low Rap1b, and subsequent high PI3K/Akt signaling trigger neutro- phil invasive protrusions and transcellular migration.

ICAM-surface, which seem to embrace the leukocytes (Barreiro et al., 2002;Carman et al., 2003) These docking structures are also known as

“transmigratory cup.” They require intracellular calcium, intact actin, andmicrotubule filaments in endothelial cells (Carman et al., 2003) and aremediated by endothelial RhoG signaling (van Buul et al., 2007) The for-mation of the transmigratory cup-containing actin-rich microvilli was seen

on firmly adherent leukocytes, and preceding both the paracellular and scellular route It remains unclear whether this structure represents an adhe-sive platform that is formed to firmly anchor the leukocytes onto theendothelial surface prior to emigration or whether it actively participates

tran-in leukocyte crawltran-ing (Carman & Sprtran-inger, 2004) Nevertheless, these ies consistently support the essential functions of dynamic integrin–ICAMbonds during leukocyte crawling in vitro and in vivo

stud-2.2 Leukocyte Invasive Protrusions

During lateral crawling, leukocytes extend highly dynamic membrane trusions, constantly protruding and retracting onto the endothelial cell sur-face prior to emigration Initially observed in neutrophils byCinamon et al.(2004), they also occur during lymphocyte crawling (Carman et al., 2007;Millan et al., 2006; Shulman et al., 2009) Live-cell imaging combining withimmunofluorescence demonstrated that crawling leukocytes generatednumerous finger-like protrusions that extended underneath the cell and atthe cell periphery, concentrated at the uropod (Carman et al., 2007;Millan et al., 2006; Shulman et al., 2009) These projections are assumed

pro-to be cell-aupro-tonomous as they are equally observed in leukocyte crawling

on EC-free substrate but are stimulated under shear stress conditions(Shulman et al., 2009) They create deep invaginations onto the endothelialcells away from the junctions, and through endothelial junctions; and, sowere named “invasive protrusions” (Carman et al., 2007; Millan et al.,2006; Shulman et al., 2009) Although one study observed invasive protru-sions preceding transcellular migration only (Carman et al., 2007), othershave demonstrated their occurrence during the initial stage of transmigrationboth at and away from the junction (Martinelli et al., 2014; Shulman et al.,2009) These protrusions may coincide with the integrin-enriched focalzones described above Immunofluorescence indicated HA-LFA-1 situated

at the base of individual invasive filopodia; and they were observed both inthe ventral part of the leukocyte and at the cell periphery (Shulman et al.,2009) Carman et al reported that these structures closely resemble

podosomes classically seen in myeloid cells, as these protrusions were rich inactin, and were surrounded by rings of integrins (Carman et al., 2007).Due to their dynamic nature, constantly protruding and retracting ontothe endothelial surface, it was hypothesized that they are important to guidetransmigration by scanning the endothelial surface to find a site permissivefor transmigration This hypothesis was recently demonstrated (Martinelli

et al., 2014) Carman and colleagues propose that these podosomes serve

as “mechanosensors” to “probe” the endothelial cell surface in order to findpermissive sites for transcellular migration (Carman, 2009; Carman et al.,2007) Using atomic force microscopy-enabled nanoindentation along withelectron and fluorescence microscopy, they show that lymphocyte protru-sions sense the levels of resistance of endothelial cell junctions and stiffness ofendothelial cells, and, as a result, can identify area of weak endothelial actindensity where the cells then transmigrate (Martinelli et al., 2014)

2.3 Paracellular Diapedesis

Paracellular or junctional diapedesis is itself a multistep process, which iscontrolled by the sequential involvement of ICAM-1/2, VCAM-1,JAM-1/A/C, PECAM-1, CD99, and ESAM (Muller, 2013; Nourshargh

et al., 2010) One essential component of the paracellular route is the ing of the endothelial junction It has been established that leukocyte–endothelial cell interactions via ICAM-β2 integrin trigger the activation

open-of signals to endothelial cells, which lead to the phosphorylation open-ofVE-cadherin—a necessary step for loosening the adherent endothelial celljunctions and facilitating the passage of leukocytes (Vestweber, 2008) Then,leukocytes migrate and cross the endothelial junction via sequential interac-tions with several adhesive molecules JAM-A/C (Woodfin et al., 2007,2009) and PECAM (Muller et al., 1993) are critical for leukocyte diapedesis(Fig 2B) The use of genetic deletion mouse models combined with intra-vital microscopy to identify the exact location where leukocyte transmigra-tion was blocked established that heterophilic interactions betweenendothelial JAM-A/C and leukocyte β2 integrins control transmigrationupstream of PECAM Indeed, the main site of arrest of JAM-A-deficientneutrophils was found to be at the level of the endothelium In contrast,PECAM-deficient neutrophils were mostly arrested between endothelialcells and below the endothelial cell basement membrane (Woodfin et al.,

2007, 2009) Elegant experiments using sequential addition and removal

of anti-PECAM and anti-CD99 blocking antibody or vice versa further

demonstrated that CD99 is required at a later stage of the transmigrationprocess than PECAM (Lou et al., 2007; Schenkel et al., 2002) Interestingly,PECAM-1 interactions stimulate the recruitment of unligated adhesionmolecules (e.g., PECAM, JAM-A, CD99) that leukocytes can interact withwithin the endothelial junction, likely to guide leukocytes moving across thejunction Unligated molecules are recruited to the endothelial cell border viacertain types of vesicles called the endothelial lateral border recycling com-partment (LBRC;Mamdouh et al., 2003) At the same time, the LBRC isthought to allow high-density adhesive interactions to be pushed aside toremove structural barrier to transmigration and open the endothelial junc-tion This compartment is trafficked to the site of transmigration by kinesinmolecular motor along microtubules (Mamdouh, Kreitzer, & Muller, 2008).

It is distinct from caveola and vesiculo-vacuolar organelles (VVO) Finally,once past the endothelial cell layer, neutrophils transmigrate through peri-cytes and the vascular basement membrane in ICAM-1/Mac-1–LFA-1- andPECAM-1-dependent manners (Dangerfield et al., 2002; Proebstl et al.,2012; Voisin & Nourshargh, 2013)

2.4 Transcellular Diapedesis

A number of studies have now provided convincing evidence for the rence of transcellular migration in vivo, as reviewed in Sage and Carman(2009) Earlier studies using transmission electron microscopy of tissue sec-tions demonstrated that neutrophils migrated almost exclusively via the tran-scellular route in skin tissues in response to the bacterial chemoattractantformyl-Met-Leu-Phe (fMLP), in vivo (Feng et al., 1998) More recently,serial-section confocal fluorescence microscopy indicated that 15% of neu-trophils migrated transcellularly in macrophage inflammatory protein2-alpha (MIP2-alpha)-challenged cremaster muscle, in vivo (Phillipson

occur-et al., 2006) Finally, the transcellular migration seems to prevail whenthe endothelial cell junctions are too tight, such as the blood–brain barrier(Lossinsky & Shivers, 2004; Wolburg, Wolburg-Buchholz, & Engelhardt,2005) Hence, it has become clear that the transcellular route is a regulatedprocess invivo In this regards, several factors have recently been shown tofavor transcellular migration, including the stiffness of endothelial cells, thetightness of endothelial cell junctions, or the density of integrin ligands at theendothelial apical surface; these factors will be discussed later (Martinelli

et al., 2014; Schaefer et al., 2014; Yang et al., 2005)

Transcellular migration is a fascinating process enabling leukocytes tocross the endothelial cell barrier away from the endothelial cell junctions.For this, the membrane of leukocytes and endothelial cells fuses to form atranscellular channel between the apical and basal membrane facilitating leu-kocyte transmigration while leaving the endothelial cell junctions intact(Carman et al., 2007) Surprisingly, the adhesive molecules and mechanismsthat guide transcellular migration are very similar to those controlling junc-tional migration Like for paracellular migration, transcellular diapedesis isalways preceded by ICAM-dependent lateral leukocyte crawling onto theendothelial surface during which the cells extend “scanning/invasive” pro-trusions (Carman et al., 2007; Gorina et al., 2014; Martinelli et al., 2014;Shulman et al., 2009); the formation of a transmigratory cup made ofICAM-1 clusters and of docking structures as well as the recruitment ofPECAM-1, CD99, and JAM-A to leukocyte–endothelial cell contact viathe LBRC are also necessary for transcellular diapedesis (Carman et al.,

2003, 2007; Mamdouh et al., 2009; Millan et al., 2006)

Detailed epifluorescence and total internal reflection fluorescencemicroscopy time-lapse imaging provided important information on thetranscellular process (Carman et al., 2007; Millan et al., 2006) First, thesestudies confirmed that transmigrating leukocytes extended active protru-sions and were surrounded by rings enriched in LFA-1 ICAM-1 andVCAM-1 on endothelial cells localize to F-actin-rich docking structuresaround adherent leukocytes (Carman et al., 2007; Millan et al., 2006) Inter-mediate filaments, such as vimentin, participate in forming a robust dockingstructure at the interface between the leukocyte and endothelial cells(Nieminen et al., 2006) Interestingly, endothelial caveolin-1 was found dis-tributed in ICAM-1-rich areas at the endothelial cell periphery and sur-rounding actively transmigrating leukocytes away from the endothelialjunction (Millan et al., 2006) As the lymphocytes extended dynamic pro-trusions into endothelial cells, ICAM-1 clusters then internalized intocaveolin-1 and F-actin-rich membrane invaginations and vesicles and trans-located to the basal plasma membrane, such that the cells seemed to glidethrough the LFA-1 ring leaving a small cluster of LFA-1 on the endothelialapical surface (Millan et al., 2006) Interestingly, caveolin-1 localized morefrequently around lymphocytes taking the transcellular route than para-cellular Knockdown of caveolin-1 in endothelial cells specifically reducedtranscellular migration suggesting a specialized role for caveolin-1 in thenonjunctional migration (Millan et al., 2006) Consistently, another group

reported high levels of caveolin-1 in endothelial cells favored the scellular path whereas its downregulation promoted the paracellular route(Marmon et al., 2009) Hence, once the site for transcellular migrationhas been decided, podosomes/protrusions can extend into long “invasive-like” protrusions to facilitate the transcellular channel likely via recruitment

tran-of specialized endothelial cell vesicles providing cellular membrane as well ascytoskeleton components and adhesive molecules (Carman et al., 2007;Millan et al., 2006) Indeed, EM studies have shown the recruitment ofmembrane enriched vesicles at site of transcellular migration as well as occur-rence of membrane fusion between the leukocyte and endothelial cells.Membrane fusion depended on SNARE-containing membrane fusioncomplexes and involved the recruitment of actin and lipid raft-rich mem-branes via displacement of endothelial cell caveloa and vesicular vacuolarorganelles (Carman et al., 2007; Millan et al., 2006) Caveola and vesicularvacuolar organelles were not observed during paracellular migration,although the LBRC was involved during both junctional and nonjunctionalmigration Hence, these studies highlight some key differences between thetwo modes of diapedesis The nature of the vesicles to be recruited to the site

of migration differs between the two routes Caveola- and VVO-mediatedmembrane fusion between leukocytes and endothelial cells appear to beunique to transcellular migration (Fig 2B) The mechanisms controllingthese events are likely key determinant factors of paracellular and tran-scellular migration and require further investigations

3 SIGNALING MECHANISM

3.1 Signaling in Leukocyte Transmigration

3.1.1 Regulation of Integrin Activation

The extravasation cascade is mainly regulated by a coordinated cellularresponse to chemokines and adhesive molecules (Gambardella &Vermeren, 2013; Mocsai, Walzog, & Lowell, 2015; Williams et al.,2011) Here, I will mostly focus on firm adhesion, crawling, and transmigra-tion In resting state, leukocytes are freely floating in the blood stream owing

to their low affinity for the endothelial apical surface This is due to a lowexpression of integrin ligands on the endothelial apical surface, and to thebent conformation of the leukocyte integrin, which offers low binding affin-ity for ligands (Herter & Zarbock, 2013; Luo et al., 2007) In response toinflammatory insult, chemokines are released and immobilized onto theendothelial apical surface In addition, the expression of endothelial integrin

ligand increases Cells initially respond to immobilized endothelialchemokines through leukocyte heterotrimeric guanine nucleotide-bindingregulatory proteins (G-proteins)-coupled receptors (GPCR), which thentransmit intracellular signals that coordinate active rearrangement of thecytoskeleton and integrin activation These intracellular signals trigger achange in integrin conformation, which increases integrin affinity forligands, known as “inside-out signaling.” The intracytoplasmic tail ofintegrins is bound to the cytoskeleton, and integrin activation requires activerearrangement of these bounds via interaction with cytoskeletal proteinsalpha-actinin, talin-1, and kindlin-3 (Herter & Zarbock, 2013; Luo et al.,2007) In addition, the reorganization of single integrin molecule into clus-ters of several molecules via lateral movement of integrin within the plasmamembrane enables multiple integrin ligand interactions, known as aviditychanges, which strengthens leukocyte–endothelial cell interactions Binding

of ligands to integrin in turn triggers signaling cascades, called “outside-insignaling,” further regulating leukocyte behavior Ras proximity 1 (Rap1)

is an evolutionary conserved protein of the Ras-like GTPase superfamilythat cycles between GTP-bound active and GDP-bound inactive formsthrough guanine exchange factors (GEFs) and GTPase-activating proteins(GAPs; Caron, 2003; M’Rabet et al., 1998) The mammalian genomeencodes two Rap1 genes, Rap1a and Rap1b, which are highly homologous,although they have both redundant and specific functions (Caron, 2003;Chrzanowska-Wodnicka et al., 2005; Li et al., 2007; Wittchen,Aghajanian, & Burridge, 2011) Rap1 has emerged as a key regulator ofintegrin activation through inside-out signaling (Katagiri et al., 2003;Sebzda et al., 2002) In immune cells, Rap1 promotes lymphocyte adhesionand migration Rap1 is activated at the plasma membrane and recruits itseffector RAPL to the integrinα subunit tail (Katagiri et al., 2003) In addi-tion, Rap1 recruits talin to the integrinβ subunit tail, which is sufficient toopen integrins into high ligand binding affinity Another actin binding pro-tein kindlin-3 binds to the integrin tail and directly participates in integrinconformational changes (Svensson et al., 2009) These tensions further con-trol lateral mobility of the integrin within the plasma membrane and playcritical role in integrin clustering The integrin inside-out signaling cascadeinvolves other signaling molecules, including notably Src kinases, PLC, andPI3K, which will not be described here This canonical pathway has mostlybeen described and validated using genetic models for LFA-1 activation andfirm adhesion Its involvement in Mac-1 activation and neutrophil adhesion

is less clear Recent work suggested that Rap1b may be dispensable for