animal models in idiopathic inflammatory myopathies how to overcome a translational roadblock

Probes of RRV or CHIKV infected patients showed CD74 upregulation on peripheral blood mononuclear cells PBMCs as well as increased serum levels of MIF [62] exemplifying the potential of

PII: S1568-9972(17)30054-X

DOI: doi: 10.1016/j.autrev.2017.03.001

Please cite this article as: Afzali Ali, Ruck Tobias, Wiendl Heinz, Meuth Sven G., imal models in idiopathic inflammatory myopathies: How to overcome a translational

An-roadblock?, Autoimmunity Reviews (2017), doi:10.1016/j.autrev.2017.03.001

This is a PDF file of an unedited manuscript that has been accepted for publication.

As a service to our customers we are providing this early version of the manuscript The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

How to overcome a translational roadblock?

Ali Afzali1,*, Tobias Ruck2,*, Heinz Wiendl2 and Sven G Meuth2

Corresponding Author: Tobias Ruck, Department of Neurology, University of

Mün-ster, Albert-Schweitzer-Campus 1, 48149 MünMün-ster, Germany, Tel +49-251-83-46811,

Fax +49-251-83-46812, tobias.ruck@ukmuenster.de

Abstract

Idiopathic inflammatory myopathies (IIMs) encompass a heterogenic group of rare

muscle diseases with common symptoms including muscle weakness and the

pres-ence of certain histological features Since the pathogenesis remains unclear,

thera-peutic approaches in general comprise unspecific immunosuppression strategies that

have been met with limited success Therefore, a deeper understanding of the

under-lying pathophysiological mechanisms is critically required to assist in development of

targeted therapies Animal models have proven to be tremendously helpful in

mech-anistic studies and allow researchers to overcome the inevitable restrictions of

hu-man research Although the number of different IIM models has drastically increased

over the last few decades, a model that exhibits the phenotypical and

histopathologi-cal hallmarks of IIM is still missing Recent publications have shown promising results

addressing different pathophysiological issues like mechanisms of onset,

chronifica-ACCEPTED MANUSCRIPT

quired in order to improve comparability and transferability among different groups

Here we provide an overview of the currently available IIM models including our own

C-peptide based small-peptide model, critically discuss their advantages and

disad-vantages and give perspectives to their future use

Keywords: Inflammatory myopathies, myositis, animal models, antigens, infections,

transgenic model

Abbreviations

AD, Alzheimer's disease; AMPD, adenosine mono phosphate deaminase; ANA,

antinucle-ar antibodies; ANCA, anticytoplasmatic antibodies; AP-1, activating protein 1; APC, antigen presenting cell; APP, amyloid precursor protein; Aβ, amyloid β; BACE-1, β-APP cleaving

enzyme 1; BiP, binding immunoglobulin protein; BMDC, bone marrow derived DC; C gans, Caenorhabditis elegans; CCL, C-C motif ligand; cdk5, cyclin dependent kinase 5; CFA,

ele-Complete Freund's adjuvant; CHIKV, Chikungunya virus; CIM, C-protein induced myositis;

CK, creatine kinase; CMMM, canine myositis of masticatory muscle; CP2, C-protein fragment 2; CPIM, C-protein peptide induced myositis; CPM, canine polymyositis; CVB, coxsackie virus B; CX3CL, chemokine C-X3-C motif ligand ; CX3CR, chemokine C-X3-C motif receptor; Daf, decay accelerating factor; DC, dendritic cells; DM, dermatomyositis; dPGS, dendritic polyglycerol sulfate; EAE, experimental autoimmune encephalomyelitis; EAM, experimental autoimmune myositis; FOXP3, Forkhead box 3; Grp78, glucose-regulated protein 78; GSK- 3β, glycogen synthase kinase 3β; HMGCR, 3-hydroxy-3-methylglutaryl-coenzyme A reduc- tase; HRS, histidyl transfer RNA synthetase; HSP, heat shock protein; i.m., intramuscular/ly; i.p., intraperitoneal/ly; i.v., intravenous/ly; ICAM-1, intercellular adhesion molecule 1; ICOS, inducible T cell costimulator; IIM, idiopathic inflammatory myopathies;; IVIG, intravenous im-

munoglobulin; KFL, Krüppel-like factor; L infantum, Leishmania infantum; LFA-1, lymphocyte

ACCEPTED MANUSCRIPT

complex; MB, myosin B fraction; MBL, mannose binding lectin; MCK, muscle creatine kinase; MCP, monocyte chemotactic protein; mer, monomer; MHC-I, myosin heavy chain I; MIF, mi- gration inhibitory factor; MIP, macrophage inflammatory protein; MLC, myosin light chain; MPS, Methylprednisolone; mTOR, mechanistic target of rapamycin; MyD88, myeloid differ- entiation primary response gene 88, NF-κB, nuclear factor kappa-light-chain-enhancer of activated B cells; OLE, Oleuropin aglycon; PBMC, peripheral blood mononuclear cells; PKM1/M2, pyruvate kinase M1/M2; PM, polymyositis; poly (I-C), Polyinosinic:polycytidylic acid, PS1, preseniline-1; Ptx, pertussis toxin; Rag, recombination activating gene; rec., re- combinant; ROS, reactive oxygen species; RRV, Ross River virus; RyR, Ryanodine-receptor; s.c., subcutaneous/ly; sIBM, sporadic Inclusion body myositis; spCIM, small peptide CIM;

SR, Sarcoplasmatic reticulum; Syt VII, Synaptotagmin VII; ThS, Thioflavin S; TLR, Toll-like receptor; TNF-Fc, TNF-α receptor fusion protein; TRE, Tetracycline response element; Treg, regulatory T cell; TRIF, TIR-domain-containing adapter-inducing interferon-β; UPR, unfolded protein response; VCAM-1, vascular adhesion molecule 1; VLA-4, very late antigen-4

ACCEPTED MANUSCRIPT

diseases (IIMs include dermatomyositis (DM), polymyositis (PM), sporadic inclusion

body myositis (sIBM) and necrotizing myositis (NM)) that display heterogeneous

clin-ical phenotypes and occur secondary to systemic disorders such as vasculitides or

connective tissue disease [1-5] The clinical characteristics of IIMs include

progres-sive muscle weakness that is accompanied by intact sensitivity and tendon reflexes,

muscle pain and elevated serum creatine kinase (CK) levels [1-3, 6] IIMs can be

fur-ther distinguished by observing the pattern of affected muscles, patient’s age at ease onset, additional involvement of organs, detected autoantibodies and response

dis-to treatment [1, 3, 4] Diagnostic investigations frequently require electromyographic

measurements and histological evaluation of muscle specimens [4, 7]

Patients suffering from DM and PM predominantly present symmetric muscle

weakness in the proximal parts of the extremities in an either chronically progressive

or relapsing-remitting disease course [1, 2, 6, 7] DM is further characterized by

addi-tional skin alterations such as efflorescences, swelling, flush or telangiectasia and the

increased coincidence of malignancies Complement-mediated destruction of

endo-mysial blood capillaries with autoantibodies directed against the endothelium is a

pathological hallmark of DM In response to antibody binding, an immunological

cas-cade is triggered leading to the formation of membrane attack complexes (MAC)

His-tological stainings show a predominant presence of CD4+ T and B lymphocytes in

affected muscle specimens Phenotypes including capillary necrosis, perivascular

damage and loss of muscle fibers also correlate with the inflammatory response [1, 2,

4, 6, 8, 9] The progressive destruction of muscle fibers in PM is assumed to be

main-ly mediated by cytotoxic CD8+ T cells that attack major histocompatibility complex

(MHC) I expressing muscle fibers Muscle biopsies from PM patients show an

endo-ACCEPTED MANUSCRIPT

mysial infiltration of CD8+ T lymphocytes surrounding necrotic and non-necrotic

mus-cle fibers [1-3, 7, 10]

Sporadic IBM starts by affecting the distal parts of the extremities and is

ac-companied by asymmetrically distributed muscle atrophy In contrast to DM and PM,

the pathophysiology is instead marked by the coincidence of neuroimmunological

and neurodegenerative components whereby the causal link still remains unclear

Similar to PM, a predominant endomysial accumulation of CD8+ T cells and presence

of necrotic muscle fibers is seen in immunohistological stainings The

neurodegener-ative component is manifested by intracellular “rimmed vacuoles” consisting of gregated misfolded proteins Interestingly, as seen in Alzheimer’s disease (AD), amy-loid-related and hyperphosphorylated tau proteins are detectable in these pathogno-

ag-monic aggregates [1, 9, 11] Amyloid β (Aβ) has been proposed to be logically relevant Aβ is cleaved from the amyloid precursor protein (APP) by the β- and γ-secretase and is assumed to be (neuro)toxic In contrast, APP is cleaved under physiologic conditions by the a-secretase, releasing neuroprotective metabolites [12,

pathophysio-13]

Despite the typical characteristics of different IIM subtypes, physicians struggle

with the diagnosis due to inconsistent or overlapping histological and serological

find-ings A controversial debate has arisen concerning the diagnostic criteria proposed

by Bohan and Peter almost 40 years ago [14] Several adjustments have been made

over the last few decades that suggest a common IIM subtypes should be diversified

further into smaller entities characterized by specific hallmarks such as circulating

antibodies or the presence of certain immunohistological features [15-18]

Although the incidence (1:100.000) of IIM is relatively low, affected patients

suffer from onerous disabilities limiting quality of life and life expectancy Up to now,

causal and selective therapeutic concepts still show limited efficacy and severe side

ACCEPTED MANUSCRIPT

effects [9, 19, 20] A better understanding of the underlying pathogenic mechanisms

will enable novel therapeutic approaches emphasizing the necessity of intensified

efforts in basic research of IIM Although the usage of animal models in basic

re-search has become a polarizing issue in societal debates, their value in the

develop-ment and risk stratification of promising pharmacological approaches has been

inevi-tably proven over recent decades [21, 22] Unfortunately, in the context of IIM, the

variety and features presented by animal models are currently insufficient and lack

the histological as well as phenotypic properties of IIM, hindering reproducibility and

practicability [23, 24] Over the last 20 years, the number of publications referring to

IIM has increased introducing different kinds of animal models with certain benefits

and limitations (Fig 1A-B) The range of animal models reaches from naturally

oc-curring myositis to nutritional, transgenic, infectious and immunological models that

mimic certain features of IIM first described by Wagner and Unverricht over a century

ago (Fig 1A, Table 1 and 2) [25, 26] In this review, we will discuss important

find-ings as well as advantages and disadvantages of different IIM models and provide

perspectives to improve their relevance for translational research Moreover, we

demonstrate data on our own approach to establish a small-peptide mouse model of

IIM

ACCEPTED MANUSCRIPT

2.1 Natural myositis in animals

Myositis in general is not exclusive to humans and can also occur in different

spe-cies Canine myositis has been investigated over the last 30 years and displays

cer-tain similarities to human IIM Up to now, two types of canine myositis have been

identified: a locally occurring form known as canine masticatory muscles myositis

(CMMM) and a general form that symmetrically affects the extremities called canine

polymyositis (CPM) Both types feature specific characteristics of IIM such as

bilat-eral, symmetric generalized muscle atrophy and weakness, cutaneous lesions,

elec-tromyographic signs of myopathy (e.g positive sharp waves, fibrillation potentials,

high frequency discharges) and circulating immune complexes [27, 28] Recently, a

member of the myosin binding protein-C family specifically expressed within and on

the surface of masticatory muscle type 2M fibers was found to be responsible for the

immune response in CMMM [29] In addition to phenotypical differences, there are

also histological differences that can be used to distinguish between CMMM and

CPM CMMM histologically resembles aspects of DM showing a predominant

infiltra-tion of both B and CD4+ lymphocytes and an increased expression of MHC-II on the

adventitia and endothelium of endo- and perimysial capillaries [30-32] In contrast,

CPM shows features of human PM where CD8+ T cells predominantly infiltrate

mus-cle fibers [30, 31] Shelton and colleagues investigated the expression of genes

in-volved in innate and adaptive immunity in both CMMM and CPM specimens through

microarray analyses and qPCR experiments [32] Genes involved in macrophage and

dendritic cell (DC) activation (CD68, CD40, decorin), migration (CCL4, MCP-2,

MMP19), MHC-I (DLA-A, β2-microglobulin) and MHC-II (e.g., DLA-DR, DQ) antigen presentation as well as in B cell growth (ly86), development, migration (CCL 21,

CXCL 12) and activation (lck, CD40, CD37) were upregulated in both CMMM and

CPM Interestingly, genes regulating molecules of the complement pathway (C1, C1r,

ACCEPTED MANUSCRIPT

C1S, C3, C6 and C7) were found to be upregulated only in CMMM specimens

Fur-thermore, genes encoding pro-inflammatory (IL-18, IL-24, IL-13Rα2, IRF-1, TNFRSF5, caspase 1), anti-inflammatory (TIMP3, IL10R3, IL18BP) or molecules in-

volved in tissue remodeling (different collagen subtypes, laminin and TGFBR3)

showed increased expression levels [32]

The alterations in gene and protein expression found in CMMM and CPM

specimens demonstrate various similarities to human IIM Canine myositis models

also have some major advantages in they do not require disease induction

proce-dures and permit study of the disease under naturally occurring conditions [30] It

would be interesting to see how CMMM or CPM dogs would respond to established

treatment regimens or novel treatment approaches However, experimental settings

for canine-based research are not established in most laboratories, reagents and

an-tibodies are restricted and animal housing is expensive and time-consuming

2.2 Nutritional myositis models

Environmental factors are consistently debated as contributing factors in the etiology

of IIM Research into neurodegenerative diseases such as AD or sIBM has recently

focused more specifically on cholesterol For late onset AD, a diet triggered animal

model featuring specific histological hallmarks of AD was established by feeding

rabbits a cholesterol enriched diet [33] Chen and colleagues evaluated whether this

animal model suitably represents aspects of sIBM by examining histological and

mo-lecular-biological features of muscle specimens [34, 35] Muscle sections from

treat-ed rabbits showtreat-ed characteristic histological features similar to those seen in sIBM

patients These histological features included infiltration of CD11b+ cells,

intramyofi-bril vacuoles as well as increased intramuscular deposition of Aβ, ubiquitin and perphosphorylated tau Although the incidence of these histological findings was rela-

hy-ACCEPTED MANUSCRIPT

tively low (2 out of 6 rabbits), significantly increased expression levels of APP, C99

fragment of APP and β-APP cleaving enzyme 1 (BACE-1) were found in qPCR or western blotting experiments of all treated animals [34] These results are also in

agreement with findings from brain tissue in this animal model [36] Lui and

col-leagues have recently performed microarray experiments on both brain and muscle

samples in order to check for a genetic overlap between brain and muscle tissue [35]

An overlap of twelve upregulated and seven downregulated genes was found

em-phasizing altered expression levels for genes related to hemoglobin synthesis

(in-creased) and mitochondrial oxidative phosphorylation (de(in-creased) It was proposed

that cholesterol compromises the integrity of the erythrocyte cell membrane leading

to hypoxia-mediated upregulation of hemoglobin synthesis [35] Further findings

sup-port the involvement of hemoglobin in Aβ deposition (36) The cholesterol-fed rabbit model predefines the pathogenesis of the presented phenotype Although the diet-

triggered model resembles distinct features of sIBM, the accompanying coincidence

of cardiovascular and neurodegenerative diseases yields a potential bias in the final

evaluation Furthermore, especially in the context of AD the causative connection of

cholesterol has been challenged by contradictory studies especially in the context of

AD [37] However, cholesterol metabolism seems to be a potential pathogenic target

in inflammatory myopathies since autoantibodies directed against

3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMGCR), an enzyme critically involved in

cholesterol synthesis, has been associated with necrotizing myositis Moreover,

pharmacological inhibition of HMGCR with statins leads to muscle pain and

weak-ness in up to 5 % of treatment and is suspected to be associated with an increased

risk of necrotizing myositis [38-43] Therefore, future animal models for necrotizing

myositis might target HMGCR

ACCEPTED MANUSCRIPT

2.3 Infectious animal models

Infectious diseases have been widely observed in the past to affect skeletal muscle

and hence we proposed to be involved in the initiation of myositis Certain mouse

strains inoculated with Coxsackie virus B1 (CVB) develop a dose-dependent, rapidly

progressing myositis of the proximal hind limbs with mortality rates of up to 30 %

[44-46] The first clinical and histological signs of myositis become evident four days after

inoculation followed by disease progression over three weeks that is accompanied by

myofiber necrosis, endomysial infiltration and adipose tissue reorganization after

an-other three weeks [44] The virus is detectable until two weeks after inoculation and

the virus RNA until four weeks, pointing to a self-sustaining immune reaction [44,

47-49] Yitterberg and colleagues showed that T cell deficient athymic nude (nu/nu) and

thymectomized mice did not develop chronic CVB myositis arguing for the

involve-ment of T cells in disease chronification [50, 51] Although increased expression

lev-els of circulating immune complexes and antinuclear (ANA) or anticytoplasmatic

anti-bodies (ANCA) suggest an involvement of humoral immune mechanisms, there was

no evident correlation to muscle lesion development [46, 52] Interestingly, strains

expressing the MHC haplotype H-2d were more susceptible to CVB myositis [53]

hinting towards an indirect pathomechanism for skeletal muscle destruction including

T cell activation

A less severe disease course was found in mice inoculated with the

alpha-viruses Ross River (RRV) or Chikungunya (CHIKV) Both alpha-viruses present similar

phenotypes with myositis, arthritis and tendosynovitis phenotypes [54, 55] CHIKV

titers persisted for up to 3 weeks post-injection [54] Infection of recombination

acti-vating gene 1 deficient mice (Rag1-/-), a transgenic mouse line model unable to

pro-duce mature lymphocytes, with both RRV and CHIKV showed no difference in the

disease course excluding a fundamental role of the adaptive immune system in

dis-ACCEPTED MANUSCRIPT

ease development [55, 56] In contrast, less severe clinical signs were observed in

complement C3 or C3 receptor deficient mice upon infection with both alphaviruses

However, a reduction of inflammatory infiltrates was not observed hinting at an

influ-ence of the complement system in the effector phase rather than the induction phase

of the disease [57, 58] Further investigations have identified the lectin pathway as

the underlying mechanism of complement activation as increased levels of mannose

binding lectin (MBL) in sera of RRV infected patients In accordance, mice deficient in

MBL developed less severe phenotypes upon RRV infection [59] Ameliorated RRV

myositis was also seen after treatment with the monocyte chemotactic protein (MCP)

inhibitor binderit [60] Recent work has identified macrophage migration inhibitory

factor (MIF) and its receptor as potential therapeutic targets for RRV and CHIKV

myositis Mild disease courses were reported in MIF deficient (MIF-/-) or receptor (CD74) deficient (CD74-/-) mice, even though viral titers or the extent of in-

MIF-flammatory infiltration remained unaltered [61, 62] Probes of RRV or CHIKV infected

patients showed CD74 upregulation on peripheral blood mononuclear cells (PBMCs)

as well as increased serum levels of MIF [62] exemplifying the potential of animal

models to improve the understanding of pathogenic mechanisms in myositis

Mice inoculated with the protozoan Trypanosoma cruzi present mild myositis

of skeletal muscle cells and cardiomyocytes that reaches a peak six weeks after

in-oculation and is followed by a rapid recovery Parasites can be detected up to twelve

weeks after inoculation Endomysial infiltrates mainly consist of lymphocytes with a

predominance of CD8+ T cells [63, 64] Inoculation of decay accelerating factor 1

de-ficient mice (Daf-/-) with Trypanosoma cruzi resulted in a more severe disease course

accompanied by increased histological scores.Daf1 inhibits cell destruction mediated

by complement attack; therefore, Daf1 deficiency exacerbates

complement-dependent immune reactions [65]

ACCEPTED MANUSCRIPT

Visceral leishmaniasis caused by protozoan parasites usually affects inner

or-gans like the spleen, liver or bone marrow Dogs suffering from CMMM have tested

positive for Leishmania infantum infections, encouraging further analysis [66, 67]

Paciello and colleagues, therefore, inoculated syrian hamsters intraperitoneally (i.p.)

with Leishmania infantum and performed immunohistological analysis on isolated

skeletal muscles They observed weight loss, reduced activity and loss of strength

accompanied by increased CK levels Endomysial infiltrates consisted mainly of T

cells with an increased ratio of CD8+ T cells invading necrotic myofibers at the sites

of expressed MHC-I molecules [68] Unfortunately, details of the effects on other

or-gans and the disease course were not given

When considering these models together, inoculation with infectious agents

triggered in most cases acute progressive myositis with chronification that enabled

the study of mechanisms of acute destruction and maintenance of chronic muscle

disease This especially permitted investigation of the mechanism of muscle fibrosis

and fatty degeneration Nonetheless, infectious animal models seem to be very

harmful to the treated animals resulting in increased mortality that occurs during the

process of disease induction Furthermore, inoculation with infectious agents

resem-bles a systemic inflammatory response The effects on organs other than the skeletal

muscle should also be checked in treated animals to evaluate the specificity of

mus-cle inflammation Additionally, further questions concerning the pathophysiology need

to be addressed: i.e does the muscle destruction result from an immediate effect of

the infectious agents or as a consequence of a molecular mimicry or a completely

different approach? Moreover, strict requirements for animal housing and

experi-mental setups prevent broad use of these models

ACCEPTED MANUSCRIPT

2.4 Immunological animal models

An excess of particular antigens has the potential to trigger autoimmune responses

against host tissue under certain environmental conditions that increase the

suscep-tibility to autoimmunity In immunological animal models this particular environment

needs to be induced by the use of adjuvants such as complete Freund’s adjuvant (CFA) and pertussis toxin (Ptx) in order to increase the efficiency of immunization

Both CFA and Ptx have been shown to modulate interaction of the innate and

adap-tive immune system by enhanced antigen presentation, cytokine secretion and T cell

priming [69, 70] All immunological animal models described here make use of

anti-gens emulsified in CFA followed by simultaneous i.p injections of Ptx However, it

should be mentioned that in IIM, it is still unknown whether these diseases are

anti-gen-driven Currently, three myositis models have been established that use antigens

isolated from skeletal muscle tissue (Fig 2): two types of experimental

autoim-mune myositis (EAM) induced by a partially purified myosin B (MB) fraction or

by purified myosin (>95 % purity) and C-protein induced myositis (CIM) They

can be either induced either by injection of the particular antigen or by transfer of

an-tigen-experienced splenocytes isolated from cocultures with muscle cells [71] or

im-munized mice (Fig 1A)

In 1956, Pearson and colleagues published their first pioneering efforts in the

establishment of a rodent antigen-mediated myositis model by subcutaneously (s.c.)

injecting rats with muscle homogenates emulsified with Mycobacterium tuberculosis

(M tuberculosis) (Fig 1A) They observed an unspecific inflammatory effect on

myo-fibers and surrounding tissues that resulted in an arthritic phenotype [72] Later,

Dawkins and colleagues performed s.c and intramuscular (i.m.) injections of

synge-neic muscle homogenates into the necks of guinea pigs resulting in an increased

number of inflammatory infiltrates, myofiber necrosis, and phagocytosis, first coining

ACCEPTED MANUSCRIPT

the term EAM Clinical signs were absent; however, side effects were observed in

spleens as hyperplasia of the reticuloendothelial system and in the lungs as diffuse

granulomatosis [73] Further adjustments to the protocol such as transferring isolated

lymph gland cells from EAM mice to wild type (wt) mice or reducing the amount of

adjuvant used were not sufficient to improve the relevance of the model [74, 75] It

was more then ten years later in 1987 that a group from Tokyo purified the muscle

homogenate to a “partially purified MB fraction” consisting of different muscle cific proteins such as myosin, actin and tropomyosin that induced a mild skeletal

spe-muscle specific myositis in guinea pigs Furthermore, they observed endomysial

infil-tration consisting mainly of macrophages and to a lesser extent lymphocytes that was

accompanied by complement C3 accumulation and elevated CK serum levels [76,

77] The disease course was self-limiting starting with histological abnormalities

sev-en days after immunization that reached a peak at day 14 and recovered by day 25

[78] Subsequent adjustments to this model included the use of xenogeneic MB, the

modification of the number as well as the frequency of injections and the choice of

the target animal [78-83] The SJL/J mouse strain was found to be the only strain

susceptible to this type of EAM after four immunizations with MB in weekly intervals

[84] Rabbit and human MB demonstrated the highest myositogenic potential for

mu-rine EAM compared to mumu-rine and rat MB [82] There were no differences when

au-tologous or xenogeneic myosin was used [79] and the disease severity correlated

with the number of injections [78, 79] Ito and colleagues performed detailed

im-munohistochemical stainings on muscle specimens from EAM rats after four

immun-izations with rabbit MB that were delivered on a weekly basis They observed a

pre-dominance of CD8+ T cells among the endomysial infiltrates that invaded necrotic

myofibers The sarcolemma of non-necrotic muscle fibers expressed the adhesion

molecules ICAM-1 (intercellular adhesion molecule 1) and VCAM-1 (vascular cell

ACCEPTED MANUSCRIPT

adhesion molecule 1) Mononuclear infiltrates expressed the corresponding ligands

lymphocyte function-associated antigen 1 (LFA-1) and very late antigen-4 (VLA-4)

[80] Adhesion molecules have been proposed to be critically involved in antigen

presentation in inflammatory myopathies by stabilizing the immunological synapse

(79-80) A group from Paris was able to reproduce this myositis model in SJL/J mice

and observed a significant loss of muscle strength as measured by an inverted

screen test which measures the time a mouse can hold on an inverted wire grid They

also investigated the role of regulatory T cells (Treg) in the context of EAM showing

that Treg depletion resulted in increased histological myositis scores without affecting

other tissues In a second approach, in vitro expanded polyclonal

CD4+CD25+CD62Lhigh Tregs were injected intravenously (i.v.) the day before

immun-ization with MB resulting in amelioration of histological scores of EAM [79] The latest

adjustment to the model comprises an immunization scheme in guinea pigs with six

successive injections of rabbit MB followed by four successive injections of defibrase

in weekly intervals Defibrase is a serine protease with the substrate fibrinogen, an

acute-phase protein EAM induction following this protocol achieved more severe

histological scores [85] However, EAM induced with MB was not able to fully mimic

all features of IIM The disease course was mild, a predominance of macrophages

with a rather small number of lymphocytes was observed among the infiltrates and

clinical features such as muscle weakness or loss of body weight were seen only

rarely [86]

In order to address these issues Kojima and colleagues improved the

purifica-tion of myosin in the MB fracpurifica-tion to a purity of up to 95 % and induced

myosin-mediated myositis in rats These rats developed EAM with a more severe,

self-limiting disease course Although clinical signs were mostly absent, striking features

became evident in histological experiments After comparing different injection

proto-ACCEPTED MANUSCRIPT

cols and different donor species for myosin, severe EAM was induced with a scheme

of four weekly injections of either human or guinea pig myosin Two types of

inflam-matory lesions were described; those filled only with CD11b+ cells and others

charac-terized by presence of both CD11b+ and TCRαβ+ cells Most TCRαβ+ cells were fiber invading CD8+ T cells The transfer of EAM spleen and lymph node cells or iso-

myo-lated immunoglobulin G (IgG) to wt rats did not induce EAM [87] Suzuki and

col-leagues subsequently reproduced these findings in SJL/J mice Immunoreactivity to

CX3CL1 (chemokine C-X3-C motif ligand 1), also known as fractalkine, was shown in

endomysial infiltrates and vascular endothelial cells as well as to its receptor

CX3CR1 on CD8+ T cells and macrophages I.p injection of an anti-CX3CL-1

mono-clonal antibody led to decreased endomysial infiltration accompanied by reduced

mRNA levels of TNF-α, IFN-γ and perforin mRNA [88] Increased levels of soluble fractalkine were also detected in sera of myositis patients and correlated with disease

activity [89] Scuderi and colleagues were able to induce EAM with purified rabbit

myosin in C57BL/6 mice after four immunizations in weekly intervals in order to

in-vestigate the influence of IL-6 on EAM in transgenic IL-6 deficient mice (IL-6-/-) They

demonstrated a complete resistance to EAM in IL-6-/- mice showing an absence of

inflammatory lesions in histological stainings accompanied by regular weight gain

and increased muscle strength compared to immunized C57BL/6 mice [90] pointing

towards an essential role of IL-6 in EAM pathology Recently, treatment of myositis

patients with tocilizumab, a monoclonal antibody directed towards the IL-6 receptor,

led to improvement of clinical and MRI parameters [91] The latest adjustment of

puri-fied rabbit myosin EAM was published by Kang and colleagues in 2015 The

estab-lished immunization protocol consisting of four injections of 1 mg myosin delivered in

weekly intervals was compared to a high dose regime of two injections of 1.5 mg

my-osin administered over two weeks Although both immunization protocols produced

ACCEPTED MANUSCRIPT

no differences in histological scores, the high dose regime was able to achieve a

more severe clinical disease courses with reduced muscle strength and weight gain

[92]

In previous studies, EAM was used to investigate the effects of established

human IIM therapies as well as novel therapeutic approaches, demonstrating the

importance of suitable animal models in understanding the mechanisms of action

(Table 3) Glucocorticosteroids and intravenous immunoglobulins (IVIG) are well

es-tablished as first line therapies in mild and moderate disease courses of IIM [93]

Ac-cordingly, Schneider and colleagues were able to show an effect of i.v

methylpredni-solone (MPS) treatment on T cell apoptosis in EAM rats Infusion with MPS (twice, 12

hours apart) one week after the final immunization for EAM induction led to higher

levels of apoptotic T cells with more than 50 % being CD8+ T cells [94] In addition, a

dose-dependent reduction of histological scores in EAM was observed after IVIG

treatment for five consecutive days SJL/J mice immunized with myosin or transfused

with splenocytes isolated from EAM mice and pretreated in vitro with both IL-2 and

myosin developed less severe scores after IVIG transfusions [95] Severe cases of

IIM require more aggressive therapies such as cytostatic agents or monoclonal

anti-bodies like rituximab [93] Calcineurin-inhibition has proven its efficacy in transplant

rejection and different autoimmune diseases through immediate impairment of T cell

development and proliferation [83] Nemoto and colleagues treated EAM SJL/J mice

daily with i.v injections of tacrolimus (day 14 to day 35) resulting in reduced

histolog-ical scores and decreased immunoreactivity to ICAM-1 [83] Similar results were

achieved by supplementation with the synthetic retinoid Am80 Retinoids have shown

to modulate differentiation and proliferation of different cells including lymphocytes

and macrophages by binding various nuclear receptors Ohyanagi and colleagues

have treated SJL/J mice immunized with rabbit myosin simultaneously with Am80 in

ACCEPTED MANUSCRIPT

a preventive (day 1 to day 21) and therapeutic (day 15 to day 28) regime Although

both approaches showed no effects on the incidence of inflammatory lesions or the

amount of necrotic myofibers, a reduction in the number of infiltrating cells was found

Further immunohistochemical stainings revealed that the decline in infiltrating cells is

accompanied by reduced levels of TNF-α and IL-1β mRNA in Am80 treated EAM

muscle tissues Additionally, in vitro proliferation assays were used to measure levels

of cytokines secreted by myoblasts IL-17 levels did not change after AM80

treat-ment, whereas significantly higher levels of IL-4, IL-10 and IFN-γ were found hinting

at the enhanced proliferation of Th1 and Th2 cells Interestingly, as shown using in

vitro experimental setup, Am80 efficiently impaired the chemotactic capability of both

myoblasts and macrophages as indicated by decreased levels of chemokine C-C

motif ligand (CCL) 2 (also known as MCP-1) and CCL 5 (also known as RANTES)

The chemotactic capability was attributed to the impaired activity of activating protein

1 (AP-1) in cultured myoblasts upon Am80 treatment Interestingly, MCP-1 and

RANTES were found to be upregulated in inflamed muscle tissues of inflammatory

myopathy specimens and were assumed to be involved in the chemotaxis of

ob-served infiltrates [96-98] Furthermore, autoantibodies directed against myosin were

detected in EAM mice, and treatment of Am80 led to a reduction of anti-myosin

IgG2a and IgG2b levels [99]

Prevel and colleagues investigated the effects of the mTOR (mechanistic

tar-get of rapamycin) inhibitor rapamycin (also known as sirolimus) in EAM mice

Ra-pamycin was administered in a preventive (day 1 to day10) or therapeutic regime

The preventive setting showed beneficial effects on muscle strength and histological

scores in a dose-dependent manner significantly reducing the number of infiltrating

and circulating lymphocytes T cell lymphopenia was sustained by a reduction of

CD4+ as well as CD8+ T cells, whereas the amount of B cells remained unchanged

ACCEPTED MANUSCRIPT

Interestingly, Treg numbers were augmented in rapamycin treated EAM mice

Addi-tionally, the amount of proliferating Ki-67+ T cells was reduced and the transcription

factor Krüppel-like factor (KLF) 2, a transcription factor critically involved in

lympho-cyte homing, was found to be upregulated in nạve and activated T cells Thus, the

authors assume that mTOR inhibition leads to decreased levels of proliferation and

the maintenance of T cells in lymph nodes by KFL2 enhancement Tregs are spared

by rapamycin since the mTOR pathway is not relevant for Treg functions In a

cura-tive regime rapamycin was administered one day after the last injection for ten days

leading to comparably reduced histological scores and increased muscle strength

accompanied by higher Treg levels However, there was no effect on the KLF2

path-way [100]

Although purified myosin for EAM induction helped to overcome certain

limita-tions of MB EAM, there are still objeclimita-tions to designating it as the “perfect” myositis model Indeed, the disease course and the pattern of infiltrating cells resemble cer-

tain aspects of IIM However, EAM provides only moderate incidences and lacks

clin-ical signs of muscle impairment A major concern lies in the limitations of murine

EAM as SJL/J mice spontaneously develop a form of myopathy with growing age that

is related to a defect in the dysferlin gene Some researchers object to the claim that

the antigen-mediated effects are biased by the spontaneous myopathy of SJL/J mice

[101, 102] Up until now, Scuderi and colleagues were the only group to successfully

induce EAM in C57BL/6 mice [84, 90] Most transgenic models are established on

the C57BL/6 background, meaning susceptibility to C57BL/6 mice would be of great

value

C-protein induced myositis (CIM) demonstrates many histopathological and

immunological features of human IIM In experiments inducing autoimmune

myocar-ditis with cardiac C-protein, Kohyama and colleagues purified C-protein from skeletal

ACCEPTED MANUSCRIPT

muscle to investigate its myositogenic capability in rats C-protein was injected 4

times in weekly intervals into multiple sites of the back In contrast to EAM that uses

purified myosin as an antigen, CIM achieved reproducibly higher incidences of

in-flammation with more severe disease courses (histological scores: purified myosin

0.23 ± 0.05 vs purified C-protein 1.41 ± 0.22 with p < 0.001) They observed

endo-mysial infiltrates that were enriched with macrophages, CD4+ and CD8+ T cells

Alt-hough CD4+ T cells were predominant among the infiltrates, CD8+ Tcells were

par-ticularly found to invade necrotic myofibers as seen in human IIM specimens [103]

The purification of myosin or C-protein is a complicated procedure, which is

biased by the protocol itself and the muscle tissue used for purification It has been

shown, that different samples of muscle tissue were able to achieve different

inci-dence and histological scores [104] Sugihara and colleagues established a protocol

to produce four recombinant (rec.) protein fragments in Escherichia coli; rec

C-protein fragment 2 (CP2) showed the highest myositogenic potential They were able

to induce moderate CIM in C57BL/6 mice with almost 100 % incidence and a

self-limiting, T cell driven disease course Histology of muscle sections revealed a

dis-ease maximum at day 14 with an enrichment of macrophages, CD8+ and CD4+ T

cells that fully recovered after day 21 Interestingly, a particular distribution pattern of

CD4+ and CD8+ T cells was described with CD4+:CD8+ ratios of 1:1 in endomysial

sites and 3.5:1 in perivascular sites, supporting the concept of the interplay of CD4+

and CD8+ T cells in inflammatory reactions In this context, they investigated the role

of CD4+ T, CD8+ T and B cells in CIM focusing on incidence and histological scores

Depletion of CD4+ or CD8+ Tcells was achieved by i.p application of anti-CD4 or

an-ti-CD8 mAb for three days, ten days before immunization Although depletion of both

T cell subsets led to decreased incidences and histological scores, the effects on

CD8 depletion were more pronounced (histological scores: CD4+ depletion 0.7 ± 0.7

ACCEPTED MANUSCRIPT

vs control 2.0 ± 0.4 with p < 0.05; CD8+ depletion 0.6 ± 1.0 vs control 1.9 ± 0.6 with

p < 0.05) [101] This finding was supported by ameliorated CIM courses in microglobulin-deficient mice lacking mature CD8+ T cells or perforin-deficient mice

β2-with impaired cytotoxicity of CD8+ T cells In another approach, splenocytes isolated

from CIM mice and sorted to CD4+ or CD8+ Tcells were cocultured with bone marrow

derived dendritic cells (BMDC) isolated from nạve mice and preincubated with CP2

The transfer of preactivated CD4+ T cells to wt mice led to mild myositis with a

mod-erate incidence (up to 40 %) Transfer of the same CD4+ T cells to mice depleted of

CD8+ Tcells produced the same result arguing for the capability of CD4+ Tcells to

induce myositis Transfer of pretreated CD8+ T cells resulted in severe histological

scores and high incidences [105] Immunoglobulin heavy chain deficiency showed no

influence on CIM considering incidence and histological scores precluding a central

role of B cells in CIM pathology [101] Furthermore, CIM was used to investigate the

role of cytokines, chemokines and costimulatory molecules for disease development

with a special focus on representatives already detected in IIM specimens [106-109]

IL-1α, IL-1β and TNF-α have been found to be upregulated in muscle tissue from CIM mice [110] Additionally, CIM was induced in TNF-α (TNFα-/-) and IL-1α/β double (IL-1α/β-/-) knockout mice CIM in TNFα-/-

occurred with an incidence of around 80 %,

whereas IL-1α and IL-1β deficiency significantly reduced CIM incidence (14 %) [101] Interestingly, IFN-γ deficient (IFN-γ-/-) mice developed more severe CIM disease courses than the control group with an additional infiltration of lymphocyte antigen 6

complex (Ly-6G) expressing neutrophils This effect was attributed to the enhanced

cytotoxicity of CD8+ T cells [111] Moreover, IL-6 was found to be expressed on

infil-trating macrophages in CIM mice Only 20 % of IL-6-/- mice developed CIM with

at-tenuated histological scores [112] IL-17A- as well as IL-4-deficiency had no effect on

CIM [111, 112] Immunoreactivity and elevated serum levels of C-X-C motif

chemo-ACCEPTED MANUSCRIPT

kine 10 (CXCL10) were found in CIM mice accompanied by increased expression of

chemokine receptor CXCR3 on macrophages and CD8+ T cells [113] Further

inves-tigations revealed expression of inducible T cell costimulator (ICOS) on infiltrating T

cells in muscle sections from rats immunized with rec C-protein three times on a

weekly basis These CIM rats were treated i.v with anti-ICOS mAb twice per week

during immunization resulting in an amelioration of CIM with decreased numbers of

CD11b+ and CD8+ T cells as well as reduced mRNA levels of IL-1α and MCP-1 [114]

In contrast to human IIM CIM was inducible in MHC-I deficient (MHC-I-/-) mice without

changes in the disease course, whereas L-selectin (CD62L) deficiency (L-selectin-/-)

showed ameliorated disease courses with a reduction of immune cell infiltrates

argu-ing for an essential role of CD62L in immune cell homargu-ing to the muscle Infusion of T

cells isolated from wt mice 24 hours before immunization was able to restore myositis

incidence and severity Inflamed muscle from L-selectin-/- showed reduced mRNA

levels of the inflammatory cytokines and chemokines IL-6, IL-10, IL-12, IFN-γ and MCP-1 compared to inflamed muscle from wt CIM mice Treatment with dendritic

polyglycerol sulfate (dPGS), an inhibitor of L-selectin, was able to ameliorate severity

of CIM [115] However, CIM induction in MHC-I-/- somehow questions the pathogenic

importance of MHC-I expression in human IIM muscle

Okiyama and colleagues were interested in the mechanisms of CIM

remis-sion since disease courses end with a regresremis-sion 21 days after the immunization

Therefore, they modified the injection protocol In a first step, they proved that there

is no tolerance induction against CP2 with repeated injections However, reinjection

of only CFA without CP2 was able to induce a milder myositis potentially by a

reacti-vation of circulating, autoreactive T cells Reinjection of the Toll-like receptor (TLR)

ligands lipopolysaccharide (LPS) or polyinosinic:polycytidylic acid (poly (I-C)) instead

of CP2 resulted likewise in mild reinduction of CIM proving the involvement of local

ACCEPTED MANUSCRIPT

innate immunity Pretreatment of muscle tissue with CFA, LPS or poly (I-C) was also

able to induce myositis upon transfer of in vitro primed splenocytes isolated from CIM

mice and co-cultured with CP2-pulsed BMDCs Thus, a complex interplay of primed

muscle tissue, components of innate and adaptive immunity is involved in the

mech-anisms after reinduction of CIM Interestingly, Tregs were not involved in the

mecha-nisms of remission since experiments with anti-CD25 mAb or IL-10 deficiency

dis-played no differences in the CIM disease course [116]

CIM was also used to test therapeutic approaches (Table 3) Administration

of IVIG for five days three days after immunization of C57BL/6 mice with CP2

attenu-ated disease course, incidence and histological scores [101] Since targeted therapy

with specific antibodies has shown beneficial effects in small numbers of IIM patients,

CIM has been used for the investigation of the underlying mechanisms [91, 117,

118] Treatment of CIM C57BL/6 mice with an antibody against IL-6 receptor at day 0

and day 7 showed comparable suppression of CIM incidence and severity [112]

Im-munomodulation of both IL-1 and TNF-α led to diminished histological scores tinuous s.c injection of an IL-1 receptor antagonist (IL-1Ra) showed superior efficacy

Con-compared to intermittent i.p application of anti-IL-1Ra mAb Both a rec TNF-α tor fusion protein (TNFR-Fc) and an anti-TNFα mAb sufficiently reduced histological scores [110] Similar results were obtained upon treatment of CIM mice with anti-

recep-CXCL10 i.p on a daily basis from day 8 to day 20 [113]

Compared to antigens used in EAE (e.g MOG35-55), CP2 comprises a length

of 300 amino acid monomers (mer) CP2 isolation or synthesis is complex,

time-consuming and expensive Therefore, the group of Sugihara as well as our group

both attempted to identify the myositogenic loci of CP2 to establish a feasible, reliable

and cost-effective animal model for IIM The first attempts were carried out on rats

with twelve overlapping 29 to 30 mer peptides identifying two T cell and one B cell

ACCEPTED MANUSCRIPT

epitope with special emphasis on peptide 5 with the sequence

ELTREDSFKARYR-FKKDGKRHILIFSDVV; immunization with these peptides was not able to fully

repro-duce CIM, hinting that other factors may be involved in the complex immune reaction

[104] Our group followed this idea and dissected CP2 into 36 overlapping peptides

for a “small peptide CIM” (spCIM) We performed single immunizations and

differ-ent protocols with up to four consecutive immunizations on a weekly basis resulting in

mild to severe myositis depending on the peptide used and the number of

immuniza-tions (Fig 3-4) For histological grading, we used a muscle inflammation score

pro-posed by Sugihara and colleagues, which is mainly based on the number of necrotic

muscle fibers [101] After a run with all 36 peptides candidate peptides were identified

due to histological scores and immunological effects in in vitro experiments (Fig 3-5)

In vitro experiments comprised proliferation assays and IL-2 secretion (Fig 5)

Out-come parameters concerning the clinical phenotypes were assessed with the

Rotarod-Test and body weight gain (Fig 6A-B), histological scores and in vitro

im-munological evaluations including activation status of CD4+ and CD8+ T cells isolated

from spleen and lymph nodes (Fig 6D) Although a few peptides were able to induce

myositis clinically and histologically, peptide 13 with the sequence

KDGVEL-TREDSFKARYRF exceeded all the other peptides in its myositogenic potential

In-terestingly, the sequence of peptide 13 is congruent with a part of peptide 5 used by

Matsumoto and colleagues in their approaches Multiple immunizations with peptide

13 led to a severe spCIM disease course without body weight gain (Fig 6A) and

di-minished test time and speed in the Rotarod-Test (Fig 6B) These mice showed

signs of inflammation with enlarged lymph nodes and fatty tissue reorganization of

the muscle (Fig 6C) Moreover, CD4+ as well as CD8+ T cells isolated from lymph

nodes showed a tendency for increased activation with upregulated levels of CD25,

CD40L and CD69 (Fig 6D) Kawachi and colleagues followed a different approach

ACCEPTED MANUSCRIPT

Considering CD8+ T cells and MHC-I as key players in the pathomechanism of IIM,

they investigated the fit of different skeletal muscle peptides to the binding anchor

motif of the MHC-I molecule of BALB/C mice Among the candidate peptides only

pyruvate kinase M1/M2 (PKM1/M2) was able to induce myositis in BALB/C mice The

immunization was performed with injections of BMDCs primed in vitro with the

partic-ular peptide into inguinal lymph nodes [119] Okiyama and colleagues followed this

idea and used three different internet-based prediction systems to identify putative

fragments of CP2 binding to MHC-I Among 24 calculated candidates the peptide

with the sequence HILIYSDV was superior in in vitro upregulation of MHC-I in target

cells Surprisingly, immunization with HILIYSDV was not sufficient to induce CIM

Also our group was not able to induce myositis upon immunization with a similar

fragment However, Okiyama and colleagues observed severe myositis after

transfu-sion of BMDCs pulsed with HILIYSDV and simultaneous CFA injections They called

their model C-protein peptide induced myositis (CPIM), which was performed on

both CD4+ and CD8+ T cell depleted mice, respectively CD4+ depletion had no

influ-ence on the CPIM disease course, whereas CD8+ depletion resulted in ameliorated

histological scores Nonetheless, in vitro assays investigating the cytotoxic capability

of CD8+ T cells isolated from CPIM mice could not confirm a HILIYSDV-specific

cyto-toxic reaction The authors argue that the lack of specific cytocyto-toxicity could result

from technical limitations or missing interplay of CD4+ and CD8+ T cells [120] Further

improvement could be achieved by using a combination of different candidate

pep-tides or analysis of the conformational structure and its binding region

Patients suffering from IIM often present manifestations in other organs raising

the question as to whether the antigen is strictly limited to skeletal muscle proteins

Nakano and colleagues proposed laminin as a possible antigen due to its

wide-spread expression (Fig 2) Analogous to CIM, they performed immunizations in rats

ACCEPTED MANUSCRIPT

with laminin emulsified in CFA simultaneously with Ptx injections Although clinical

signs were absent, they were able to observe a significant increase of necrotic fibers

accompanied by endomysial infiltration enriched with macrophages, CD4+ and CD8+

T cells The distribution pattern was similar to those seen in EAM and CIM with CD8+

T cells invading myofibers and an accumulation of CD4+ Tcells in the endomysium

[121] Histidyl-transfer RNA synthetase (HRS), also known as Jo-1, was proposed

as a potential antigen since antibodies to Jo-1 are regularly found in IIM patients and

have been shown to correlate with disease activity (Fig 2) [122] Although the first

attempts at immunization with a full-length rec protein were not path-breaking,

im-munization with a fusion protein emulsified in CFA expressing a fragment of HRS

with specific B and T cell epitopes was able to induce inflammation in muscle and

lung tissue, two distinct hallmarks of antisynthetase syndrome [114, 123] This model

was modified by intramuscular injections of the very same antigen in the absence of

CFA and Ptx resulting in endomysial infiltration of macrophages, CD4+ and CD8+ T

cells both positive for CD44 and CCR5 The first signs of immune cell infiltration in

the muscle became evident at day 7 after immunization and lasted for 7 weeks

Fur-thermore, B and T cells isolated from these immunized mice showed strong

respons-es to the antigen by enhanced autoantibody production and increased proliferation

Nonetheless, the immune reaction was triggered independently of the T cell and B

cell receptor due to successful immunization of Rag2-/- mice [124] This is promising

since adjuvants bias results by inducting unspecific immune responses However, the

authors do not mention whether the reaction occurs only in immunized muscle tissue

as in focal myositis or if it also affects untreated muscles Since the chemotactic

po-tential of HRS was recently shown in in vitro experiments, its abilities to activate TLR

pathways were also investigated [124, 125] T cell depleted splenocytes isolated from

wt mice were able to produce autoantibodies upon application of HRS

fragment-ACCEPTED MANUSCRIPT

fusion protein [126] In vitro experiments showed the capability of the fusion protein to

directly stimulate TLR2 and TLR4 Although myositis was inducible in both TLR2-

(TLR2-/-) and TLR4-deficient (TLR4-/-) mice, the antibody response to the antigen was

diminished in both mice [126, 127] More precisely, IgM autoantibodies against the

fusion protein were still present, whereas an antibody switch to IgG was absent [127]

In order to elucidate the underlying mechanism, further immunizations of

TIR-domain-containing adapter-inducing interferon-β (TRIF)- and myeloid differentiation primary response gene 88 (MyD88)-deficient mice were performed hinting at an in-

volvement of both molecules in the activation of the innate immune system by HRS

As adaptor proteins both molecules are linked to intracellular signaling pathways

in-cluding activation of TLR [126, 127] However, despite the promising data provided

by investigating HRS-induced myositis, several questions remain unanswered

In-vestigators have focused mainly on inflammation in muscle tissue and have not

in-vestigated lung tissue or mentioned whether only immunized muscles or

unimmun-ized muscles were affected Additionally, information about the clinical status of the

mice and a comparison to the human disease is missing Nonetheless, induction of

myositis with an antigen relevant in human disease offers a great opportunity to

gen-erate results with therapeutic implications

2.5 Transgenic myositis models

Research on the pathophysiology of sIBM has not yet answered the question

wheth-er the disease starts with inflammation, degenwheth-eration, or whethwheth-er these processes

take place in parallel Animal models established for sIBM over the last 20 years

have focused on the neurodegenerative component of its pathology following

re-search on AD Most transgenic models aim to enrich intracellular levels of Aβ in cle fibers The first models lacked muscle specificity as muscle deposits were rather

mus-ACCEPTED MANUSCRIPT

found as a secondary product in AD models A murine model overexpressing the C99

fragment of APP showed unselective Aβ deposits in different tissues [128] Further adjustment of the C99 gene with a substitution of lysine to valine at the α-secretase cleavage position resulted in more severe histological tissue alterations including Aβ- and thioflavin S-positive (ThS) inclusions, lymphocytic infiltration and sarcoplasmic

vacuoles These findings were again not restricted to muscle tissue, but they were

more frequent in older mice [129] Sugarman and colleagues firstly described muscle

specific Aβ enrichment in two independent mouse lines by combining the APP

gene with a mouse muscle creatine kinase (MCK) promoter (MCK-APP mice) [130]

They were able to detect the same age-related histological changes with

immunore-activity against both Aβ40 and Aβ42, whereby Aβ40 was markedly higher expressed than Aβ42 in fast twitching type II fibers [131] However, the preponderance of Aβ40does not conform to findings from human sIBM sections where Aβ42 is assumed to be the central cytotoxic metabolite [131, 132] In order to solve this issue, the same

group extended their model by introducing a knock-in mutation in the preseniline-1

(PS1) gene, a component of the γ-secretase complex, producing significantly higher intracellular levels of Aβ42 [132] Peak levels of Aβ42 were reached at 14 months of age and were accompanied by motor function impairment and histological changes

such as centralized nuclei intracellular aggregates and surrounding infiltrates

en-riched with CD8+ T lymphocytes [132] Although PCR experiments were not able to

show significantly increased mRNA levels of CD8 compared to controls, increased

levels of phosphorylated tau were detected Further experiments investigating

possi-ble underlying signaling pathways of phosphorylation revealed increased enzyme

activities of glycogen synthase kinase 3β (GSK-3β) and cyclin dependent kinase 5 (cdk5) [132] In accordance to this model, histological stainings of human sIBM mus-

cle samples revealed that fast twitching type IIB fibers are more affected than other

ACCEPTED MANUSCRIPT

fiber types [131, 133] A different group from Boston focused on this observation and

developed a mouse model where APP overexpression is targeted exclusively to

fast twitching type II fibers by linking the APP gene with the type II fiber specific

myosin light chain (MLC) 1/3 promoter/enhancer (MLC-APP) [133] They performed

multiple tests including histological stainings, muscle strength analysis,

electromyog-raphy and CK measurements The MLC-APP transgenic mice were significantly

weaker in strength tests and presented global atrophy of the extremities with growing

age accompanied by myopathic alterations in electromyography and a four-fold

in-crease in CK levels Histological changes included intracellular Aβ42 and ThS gation, intramyofibril vacuoles and centralized nuclei [133] In contrast to the MCK-

aggre-APP transgenic model, endomysial infiltrates were not found Interestingly,

microelec-trode recordings revealed an 2-fold increase in the concentration of intracellular free

calcium [Ca2+]i before the first histological changes were observed [133, 134] Further

experiments on isolated muscle fibers from both MLC-APP and MCK-APP transgenic

mice revealed an indirect interaction between intracellular Aβ42 deposition and [Ca2+]i [135] The effect of Aβ42 on [Ca2+]i is attributed to increased mobilization from intracel-lular (by Ryanodine-receptor (RyR) activation) as well as extracellular (by increased

membrane permeability) Ca2+ stores [134, 135] Both pathways result in increased

[Ca2+]idiminishing the driving force for Ca2+ entry with the consequence of reduced

contractility Boncompagni and colleagues have found structural and functional

alter-ations in mitochondria from muscle fibers of both transgenic mouse models with

de-creased intracellular pH and inde-creased levels of reactive oxygen species (ROS)

[136] Recent findings revealed an enhancement of RyR activity upon oxidative

stress [137] Thus, they propose a cascade starting with Aβ overexpression and tochondrial degeneration leading to increased ROS release and a consequently en-

mi-hanced activity of RyR [136] However, despite all the remarkable achievements with

ACCEPTED MANUSCRIPT

this sIBM model, recent publications indicate that there have been difficulties in

trans-ferring this transgenic model to other laboratories So far, other groups have not been

able to reproduce the previously reported histological changes Lower copy numbers

of the transgene, failure of incorporation, genome rearrangement or loss of copy

number were proposed as possible explanations [138] Furthermore, both transgenic

models lack characteristic immunological and histopathological features such as

mononuclear cell infiltration

During the first international conference of myositis in 2015 a group from

Mu-nich introduced a novel transgenic animal model for sIBM that follows a different

ap-proach They detected increased levels of lymphotoxins (LT) and downstream genes

associated with LT receptor activity in samples of human IIM patients LTs have been

implicated in lymphocyte - target cell interactions rendering them as novel therapeutic

targets in inflammatory disorders like renal inflammation [139] With regard to these

findings they generated a transgenic murine model that constitutively expresses LT α and β specifically in skeletal muscle cells upon attachment to human α-skeletal actin

(HSA) These HSA-LTab mice showed histological signs of myositis with infiltration

of immune cells, expression of cytokines and chemokines as well as upregulation of

MHC-I Reduced body weight and loss of muscle strength were also described

Inter-estingly, protein aggregates were detected in myofibers, which correlated to

autoph-agy Although they did not clarify the composition of these aggregates, they

de-scribed them as reminiscent of human IBM [140] However, in case the aggregates

resemble deposits found in human sIBM specimens, the HSA-LTab myositis model

seems to be promising since it also includes mononuclear cell infiltration in contrast

to the MCK-APP model

Different researchers have left the field of rodent models and are now using

Caenorhabditis elegans (C elegans) as potential model for sIBM Although C

ele-ACCEPTED MANUSCRIPT

gans is fundamentally different to humans, its usage in a scientific context is widely

acknowledged, and the Nobel Prize in Physiology or Medicine for 2002 was awarded

to Sydney Brenner, Robert Horvitz and John Sulston for their work in C elegans

re-search [141] Link and colleagues have described a transgenic C elegans model

expressing human Aβ under the control of a muscle specific gene [142] Muscles from these worms show immunoreactivity to ThS and Aβ Further work has shown genetic alterations hinting at induced oxidative stress and activation of stress re-

sponses such as increased levels of chaperones or heat shock proteins (HSP)

[143-145] Minniti and colleagues have investigated the role of metal ions for Aβ tion demonstrating a connection between copper (Cu2+) concentration and intracellu-

deposi-lar aggregates [146] Although Cu2+ supplementation increased the number of

amy-loid deposits, aggregation of Aβ oligomers was significantly reduced resulting in tenuated locomotor impairment [146-148] They propose aggregation of intracellular

at-Aβ deposits as a resistance mechanism to toxic effects of metal ions like Cu2+

[146]

Diomede and colleagues used this model to investigate the effects of phenols on

in-tracellular Aβ deposition [149] Transgenic worms fed with the phenol oleuropin con (OLE) developed less Aβ aggregates and showed reduced paralysis compared

agly-to untreated worms [149] However, when taken agly-together, these findings show that

the model suffers from limitations concerning the neuroinflammatory component of

sIBM Inflammatory infiltrates or the expression of inflammatory molecules were not

described

Endomysial infiltration by mononuclear cells is a histological hallmark of

hu-man IIM assumed to be closely linked to muscle fiber loss [2] A group in London

hy-pothesized that the loss of muscle fibers begins before immune cells invade This

hypothesis is based on findings from IIM specimens expressing MHC-I on and inside

muscle fibers without contact to infiltrating mononuclear cells (Fig 2) [150-152] They

ACCEPTED MANUSCRIPT

propose that an unknown stimulus induces MHC-I overexpression leading to SR

stress by an accumulation of MHC-I components [153] Following this hypothesis,

they established a murine model with constitutive MHC-I expression linked to a

tetracycline response element (TRE) and a MCK promoter as seen in the MCK-APP

model [154] TRE serves as a regulatory element and supplementation of food with

tetracycline induces MHC-I expression These transgenic mice referred as H+T+

mice present severe disease signs with growing age like reduced activity, hind limb

weakness, up to 50 % weight loss and increased CK levels Histology features show

typical signs of myositis like atrophy, centralized nuclei and mononuclear cell

inva-sion were found However, the H+T+ model has limitations in the transfer to the

hu-man disease since B and T cells are absent They detected immunoreactivity for

ICAM 1 as well as increased mRNA levels of macrophage inflammatory protein (MIP)

1α, MCP 1 and IL-15 8 out of 23 H+T+ mice showed reactivity to HRS without lation to disease course, histological scores or lung involvement [154, 155] Extensive

corre-investigation of the proteasome showed significant alterations Molecules involved in

SR responses like heat shock proteins or chaperones were found to be upregulated

while proteins important for muscle cell structure and function like myosin, actinins or

titin were downregulated [153, 156] A special focus was set on SR and its crucial

role in protein synthesis and apoptosis In this context, both the unfolded protein

re-sponse (UPR) and overload rere-sponse were found to be activated The chaperone 78

kDa Glucose-regulated protein (Grp78, also known as binding immunoglobulin

pro-tein (BiP)) and caspase 12 both critically involved in UPR showed increased

expres-sion levels as well as the transcription factor NF-κB (nuclear factor enhancer of activated B cells) [153] Nagaraju and colleagues assume that SR reacts

kappa-light-chain-on the overexpressikappa-light-chain-on of MHC-I by inducing these pathways NF-κB initiates the pression of proinflammatory molecules like cytokines and MHC-I Interestingly, even

ex-ACCEPTED MANUSCRIPT

after removing tetracycline from the food supply, MHC-I expression remained high

Moreover, Grp78 participates in the control of protein folding and aggregation An

accumulation of unfolded proteins is a signal for cell stress by which Grp78 activates

caspase 12 introducing further events leading to apoptosis [153] Rayavarapu and

colleagues treated H+T+ mice with the proteasomal inhibitor bortezomib resulting in

attenuated disease courses and decreased Grp78 levels [156] Cell stress is also

evident in metabolic processes Accordingly, Coley and colleagues showed reduced

activity of adenosine monophosphate deaminase (AMDP) 1 However treatment of

the H+T+ mice with the AMDP1 metabolite D-Ribose had no effect on disease

pro-gression [157, 158] In order to prove, the concept that mononuclear cell infiltrates

are not critically important for muscle fiber loss a group from France has extended

the H+T+ model by crossing it with Rag2-/- They were able to observe similar

pheno-typical and histological changes as described for the H+T+ mice confirming that

mus-cle damage occurs as a consequence of MHC-I overexpression [159] The H+T+

model offers a completely new concept in contrast to immunological animal models

searching for candidate antigens The consecutive expression of MHC-I is in

agree-ment with human IIM [109, 150, 160]

Coincidentally, Chakrabati and colleagues have found myositis-like changes in

synaptotagmin VII deficient mice (Syt VII -/- ) including endomysial infiltration in

af-fected muscle tissue, fiber degeneration, increased CK levels and reduced grip

strength Except for fibrotic skin lesions, other organs were not affected Infiltrates

showed immunoreactivity to CD3, CD11b and Ly-6G Interestingly, antinuclear

anti-bodies were detected in the serum of Syt VII-/- mice [161] Synaptotagmins are Ca2+

binding membrane proteins essentially involved in vesicular and lysosomal

exocyto-sis (Fig 2) Wounded cells use lysosomal exocytoexocyto-sis to reseal their cell membrane

[162] Chakrabati and colleagues propose that tissues frequently exposed to

me-ACCEPTED MANUSCRIPT

chanical stress depend on lysosomal exocytosis A deficiency would lead to a

dys-function of membrane integrity and hence to exposure of intracellular molecules

serv-ing as antigens [161] In order to prove this concept they developed a complicated

passive EAM model by transferring lymphocytes isolated from a double transgenic

mouse model deficient in Syt VII and forkhead box P3 (FOXP3) into Rag1-/- mice

[163] The idea was to enhance autoimmunity in Syt VII-/- by depleting Tregs and

transferring the emerging auto-reactive T cells into lymphocyte-deficient mice This

passive EAM led to a severe phenotype with higher histological scores and increased

CD4+ and CD8+ T cell muscle infiltration Supplementation of isolated wild type Tregs

was able to attenuate the severity of the passive EAM [163] The Syt VII-/- model

fol-lows an interesting approach in the context of antigen representation However, it

would be interesting to know whether isolated lymphocytes show reactivity to certain

skeletal muscle molecules and whether the degree of mechanical stress is able to

induce the phenotype Furthermore, the mentioned skin lesions need to be

ad-dressed with a special focus on complement activation in order to check similarities

to DM