Neutralization of IFN-γ reverts clinical and laboratory features in a mouse model of macrophage activation syndrome

Accepted Manuscript Neutralization of interferon-γ reverts clinical and laboratory features in a mouse model of macrophage activation syndrome Giusi Prencipe, PhD, Ivan Caiello, BS, Antonia Pascarella, PhD, Alexei A. Grom, MD, Claudia Bracaglia, MD, Laurence Chatel, DUT, Walter G. Ferlin, PhD, Emiliano Marasco, MD, Raffaele Strippoli, MD, PhD, Cristina de Min, MD, Fabrizio De Benedetti, MD, PhD PII:

S0091-6749(17)31277-0

DOI:

10.1016/j.jaci.2017.07.021

Reference:

YMAI 12958

To appear in:

Journal of Allergy and Clinical Immunology

Received Date: 7 October 2016 Revised Date:

17 July 2017

Accepted Date: 19 July 2017

Please cite this article as: Prencipe G, Caiello I, Pascarella A, Grom AA, Bracaglia C, Chatel L, Ferlin WG, Marasco E, Strippoli R, de Min C, De Benedetti F, Neutralization of interferon-γ reverts clinical and laboratory features in a mouse model of macrophage activation syndrome, Journal of Allergy and Clinical Immunology (2017), doi: 10.1016/j.jaci.2017.07.021. 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 1

Neutralization of interferon-γγ reverts clinical and laboratory features in a mouse model of

2

macrophage activation syndrome Giusi Prencipe1*, PhD, Ivan Caiello1, BS, Antonia Pascarella1, PhD, Alexei A. Grom2, MD, Claudia Bracaglia1, MD, Laurence Chatel3, DUT, Walter G. Ferlin3, PhD, Emiliano Marasco1, MD, Raffaele Strippoli4, MD, PhD, Cristina de Min3, MD, Fabrizio De Benedetti1, MD, PhD

RI PT

Affiliations 1

Division of Rheumatology, Bambino Gesù Children’s Hospital IRCCS, Rome, Italy Division of Rheumatology, ML 4010, Cincinnati Children’s Hospital Medical Center, Ohio, USA 3 NovImmune SA, Geneva, Switzerland 4 Department of Cellular Biotechnologies and Haematology, Sapienza University of Rome, Rome, Italy.

SC

2

*Corresponding author: Giusi Prencipe,

19

Division of Rheumatology,

20

Bambino Gesù Children’s Hospital,

21

Viale di S. Paolo, 15

22

00146 Rome, Italy.

23

Telephone number: +390668592999

24

Fax Number: +390668592904

25

E-mail: [email protected]

26

EP

TE D

18

M AN U

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Funding Source

28

This study was supported by the Italian Ministry of Health (Rome, Italy) Young Investigator Grants

29

(GR-2011-02347874) to G.P., by the FP7 Grant FIGHT-HLH (agreement number 306124) to

30

F.D.B. and C.d.M. and by NIH grants R01-AR059049 to A.A.G.

AC C

27

31 32 33 34 35 1

ACCEPTED MANUSCRIPT 36

Abstract

38

Background. The pathogenesis of macrophage activation syndrome (MAS) is not clearly

39

understood: a large body of evidence supports the involvement of mechanisms similar to those

40

implicated in primary hemophagocytic lymphohistiocytosis. Objective. To investigate the

41

pathogenic role of interferon-γ (IFNγ) and the therapeutic efficacy of IFNγ neutralization in an

42

animal model of macrophage activation syndrome. Methods. We used a MAS model established in

43

mice transgenic for human interleukin-6 (IL-6TG) challenged with LPS (MAS mice). Levels of

44

IFNγ and IFNγ-inducible chemokines were evaluated by real-time PCR in liver and spleen and by

45

ELISA in plasma. IFNγ neutralization was achieved using the anti-IFNγ antibody XMG1.2 in vivo.

46

Results. MAS mice showed a significant upregulation of the IFNγ pathway, as demonstrated by

47

increased mRNA levels of Ifnγ and by higher levels of phospho-STAT1 in liver and spleen and by

48

increased expression of IFNγ-inducible chemokines Cxcl9 and Cxcl10 in liver and spleen as well as

49

in plasma. A marked increase in Il-12a and Il-12b expression was also found in liver and spleen

50

from MAS mice. In addition, MAS mice showed a significant increase in liver CD68 positive

51

macrophages. MAS mice treated with an anti-IFNγ antibody showed a significant improvement in

52

survival and in body weight recovery, associated to a significant amelioration of ferritin, fibrinogen

53

and alanine aminotransferase levels. In MAS mice, treatment with the anti-IFNγ antibody

54

significantly decreased circulating levels of CXCL9, CXCL10 and downstream proinflammatory

55

cytokines. The decrease in CXCL9 and CXCL10 levels paralleled the decrease in the serum levels

56

of proinflammatory cytokines and ferritin. Conclusion. These results provide evidence for a

57

pathogenic role of IFNγ in MAS.

AC C

EP

TE D

M AN U

SC

RI PT

37

58 59 60 2

ACCEPTED MANUSCRIPT 61

Key messages • In a murine model of macrophage activation syndrome (MAS), the IFNγ pathway is

62

markedly activated in liver and spleen in the early phase of the disease.

63

• IFNγ neutralization ameliorates survival and clinical and laboratory features of MAS.

65

• Tissue and circulating levels of the IFNγ inducible chemokines CXCL9 and CXCL10 reflect

66

activation of the pathway and may represent biomarkers of disease in humans with MAS.

RI PT

64

67 68

SC

Key words

Macrophage activation syndrome, hemophagocytic lymphohistiocytosis, IFNγ

74

M AN U

69 70 71 72 73

Abbreviations

76

Chemokine (C-X-C motif) ligand (CXCL)

77

Hemophagocytic lymphohistiocytosis (HLH)

78

Indoleamine 2,3-dioxygenase (IDO)

79

Interferon-γ (IFNγ)

80

Interleukin (IL-)

81

Lipopolysaccharide (LPS)

82

Lymphocytic choriomeningitis virus (LMCV),

83

Macrophage activation syndrome (MAS)

84

Primary Hemophagocytic lymphohistiocytosis (pHLH)

85

Signal Transducer and Activator Of Transcription 1 (STAT1)

86

Systemic juvenile idiopathic arthritis (sJIA)

87

Tumor necrosis factor (TNF-)

AC C

EP

TE D

75

88 89 3

ACCEPTED MANUSCRIPT 90

Capsule summary

91

In a murine model of MAS, the IFNγ pathway is activated and has a pathogenic role, as

92

neutralization of IFNγ reverts the disease. These results provide the rationale for the therapeutic

93

targeting of IFNγ in MAS.

RI PT

94 95 96

SC

97 98

M AN U

99 100 101 102

TE D

103 104 105

EP

106 107

AC C

108 109 110 111 112 113 114 115 4

ACCEPTED MANUSCRIPT INTRODUCTION

117

Macrophage activation syndrome (MAS) is a term used to identify hemophagocytic

118

lymphohistiocytosis (HLH) secondary to rheumatic diseases(1, 2). It is a severe and potentially fatal

119

condition that occurs in the context of rheumatic diseases, particularly systemic juvenile idiopathic

120

arthritis (sJIA)(3). MAS is classified among secondary HLH and, indeed, shares clinical and

121

biochemical features with familial or primary HLH (pHLH)(4). A cytokine storm appears to be the

122

driving feature(5); the ensuing clinical syndrome is characterized by high fever, pancytopenia,

123

hyperferritinemia, hypofibrinogenemia, hepatosplenomegaly and intravascular coagulation(6). The

124

triggering mechanism behind pHLH is a defect in cytotoxicity, caused by mutations in genes

125

encoding proteins required for cytotoxic activity of lymphocytes and natural killer cells(7). The

126

pathogenesis of MAS is not clearly understood(8). A large body of evidence supports the

127

involvement of mechanisms similar to those implicated in pHLH. Indeed, sJIA patients with MAS

128

have been shown to have transient natural killer cell dysfunction, such as decreased cell number and

129

activity, and reduced perforin expression(9-11). Recently, in patients with sJIA and MAS, genetic

130

studies, including whole-exome sequencing, identified rare protein-altering variants in known

131

pHLH-associated genes as well as in new candidate genes, possibly involved in intracellular granule

132

trafficking, further suggesting that MAS predisposition in sJIA could be attributed, at least in part,

133

to variants of genes involved in the cytolytic pathways(12-14). Consistent with these data,

134

Sepulveda et al.(15) demonstrated that the accumulation of monoallelic mutations in HLH-causing

135

genes increases susceptibility to develop HLH immunopathology also in mice.

136

Studies in patients and in animal models of pHLH have suggested a key role of interferon-γ (IFNγ)

137

in the pathogenesis of the disease. Indeed, in a large cohort of children with HLH, highly increased

138

levels of IFNγ, correlated with disease activity, have been reported(16). Moreover, in perforin -/-

139

mice infected with lymphocytic choriomeningitis virus (LMCV), the high amount of IFNγ produced

140

by CD8+ T cells has been demonstrated to be uniquely essential for the development of the

AC C

EP

TE D

M AN U

SC

RI PT

116

5

ACCEPTED MANUSCRIPT disease(17). Similarly, in LMCV-Rab27a-deficient mice, neutralization of IFNγ has been shown to

142

revert central nervous system involvement and to reduce haemophagocytosis(18). Neutralization of

143

IFNγ is able to revert haematological abnormalities also in the mouse model of HLH secondary to

144

infections, mimicked by repeated stimulation of TLR-9 in a normal genetic background(19).

145

Aim of this study was to evaluate the pathogenic role of IFNγ in a mouse model of MAS, based on

146

mice transgenic for human IL-6 (IL-6TG)(20). We have reported that IL-6TG mice, after a single

147

administration of TLR ligands, display an increased fatality rate, associated with higher levels of

148

circulating proinflammatory cytokines and hematologic and biochemical features typically present

149

in MAS(20, 21). In these mice, MAS is induced by mimicking an acute infection with

150

administration of a TLR agonist (i.e. lipopolysaccharide, LPS). This approach recapitulates what

151

occurs in patients with sJIA: an infection is the typical trigger of MAS in the presence of active

152

disease, which is indeed characterized by high levels of IL-6, that appears to play a pivotal

153

pathogenic role in the autoinflammation of sJIA(21, 22).

154

In this study, we investigated whether IFNγ and IFNγ inducible chemokines are increased in MAS

155

induced by LPS in IL-6TG mice and whether administration of an anti-IFNγ antibody improves

156

survival and disease parameters.

TE D

M AN U

SC

RI PT

141

MATERIAL AND METHODS

AC C

158

EP

157

159 160

Mice and in vivo treatments

161

The IL-6TG mouse has been described previously(23). Mice between 10 and 14 weeks of age were

162

administered intraperitoneally a single dose (7.5 or 5 µg/g body weight) of lipopolysaccharide

163

(LPS, E. coli serotype 055:B5; Sigma-Aldrich). For treatment experiments, 100 µg/g body weight

164

of the anti-mouse IFNγ neutralizing antibody XMG1.2 (BioXCell) and of a rat IgG1 isotype-

165

matched control mAb35 (ATCC) were administered to mice intraperitoneally. Mice were 6

ACCEPTED MANUSCRIPT maintained under specific pathogen-free conditions and handled in accordance with the institutional

167

Experimental Ethics Committee guidelines.

168

Cytokine, ferritin, fibrinogen and alanine aminotransferase measurements

169

Mouse plasma cytokine and chemokine levels were determined using MILLIPLEX MAP multiplex

170

immunodetection kits (Merck), according to the manufacturer’s instructions. For all analytes, the

171

lower detection limit was 3.2 pg/ml and the upper detection limit was 10000 pg/ml. CXCL9 and

172

CXCL10 levels were further determined using the mouse Quantikine ELISA KITs (R&D Systems

173

Inc.). Serum ferritin and plasma fibrinogen concentrations were determined using ELISA kits

174

(ALPCO Diagnostics and Abcam, respectively). Alanine aminotransferase levels were determined

175

using an enzymatic assay kit (BiooScientific).

176

RNA Isolation and Quantitative Real-Time PCR

177

Total RNA was extracted from mouse liver and spleen tissues using Trizol (Life technologies) and

178

cDNA was obtained using the Superscript Vilo kit (Invitrogen). Real-time PCR assays were

179

performed using the TaqMan Universal PCR Master Mix (Applied Biosystems) and the following

180

gene-expression assays: mouse Ifnγ, Cxcl9, Cxcl10, Il-12a and Il-12b (Applied Biosystem). Gene

181

expression data were normalized using mouse Hprt (Applied Biosystem) as endogenous control.

182

Data are expressed as arbitrary units (AU), determined using the 2-∆ct method.

183

Protein extraction and Western Blot Analysis

184

Total tissue proteins were extracted using RIPA Buffer (Cell Signalling). For western blotting,

185

protein lysates were resolved by 10% SDS-PAGE. Proteins were then transferred to nitrocellulose

186

membranes (Amersham Life Sciences) and probed with antibodies to phospho-Tyr701 STAT1, total

187

STAT1 and GAPDH (all by Cell Signalling) using standard procedures.

188

Histology and Immunohistochemical analysis

189

Livers from mice were drop-fixed in neutral buffered formalin and then processed for paraffin

190

embedding. 2.5 µm sections where stained with hematoxylin and eosin (H&E). For

191

Immunohistochemical analysis, after moist heat-induced antigen retrieval with EnVision Flex

AC C

EP

TE D

M AN U

SC

RI PT

166

7

ACCEPTED MANUSCRIPT Target Retrieval solutions high pH (Dako Cytomation), 2.5 µm sections were incubated with

193

antibody to CD68 (AbCam) overnight at 4° C. After washing, sections were incubated with

194

appropriate HRP-conjugated secondary antibodies. Chromogenic detection was carried out with the

195

DAB chromogen kit (Dako Cytomation). Nuclei were counterstained with haematoxylin, followed

196

by dehydration and mounting. The number of CD68 positive cells was enumerated in at least 10

197

fields for each tissue at 40X magnification.

198

Statistical Analyses

199

Data are presented as means ± standard error mean, unless otherwise indicated. Group comparisons

200

were performed using the nonparametric Mann-Whitney U test. To assess the correlations between

201

variables, the Spearman rank correlation coefficient was calculated (r). All statistical analyses were

202

performed using the GraphPad Prism IV software (La Jolla, Ca). A value of p<0.05 was considered

203

as statistically significant.

M AN U

SC

RI PT

192

204

RESULTS

206

IFNγ and IFNγ inducible chemokine levels are higher in LPS-challenged IL-6TG mice. While

207

before LPS administration Ifnγ mRNA expression levels were not different between WT and IL-6TG

208

mice in both liver (Figure 1A) and spleen (Figure 1B), we found that they were significantly

209

increased in IL-6TG mice compared to WT mice at 12 hours after LPS administration (Figure 1A

210

and B). Moreover, we found that Cxcl9 mRNA levels were significantly higher in liver and Cxcl10

211

mRNA levels were significantly higher both in liver and in spleen compared to WT mice (Figure 1A

212

and B).

213

To further evaluate the activation of the IFNγ pathway in LPS-challenged IL-6TG mice compared to

214

WT mice, liver and spleen protein lysates were tested by Western Blot using a specific antibody for

215

Tyr701-phosphorylated STAT1, that is known to mediate IFNγ effects(24, 25). We did not detect

216

Tyr701-phosphorylation of STAT1 before LPS challenge in both WT and IL-6TG mice tissues.

217

Tyr701-phosphorylated STAT1 was similarly increased both in liver (Figure 1A) and in spleen

218

(Figure 1B) in WT and IL-6TG mice at 6 hours after LPS challenge. Notably, while in WT mice 8

AC C

EP

TE D

205

ACCEPTED MANUSCRIPT phospho-STAT1 levels returned to baseline levels at 12 hours post LPS challenge, in IL-6TG mice

220

they remained markedly higher both in liver and in spleen (Figure 1A and B).

221

Since IL-12, mainly produced by classically activated M1 macrophages(26), is known to stimulate

222

IFNγ production and has been demonstrated to act upstream to induce IFNγ in an animal model of

223

fulminant MAS(27) , we assessed mRNA expression of the Il-12a and Il-12b genes, encoding

224

respectively for the two heterologous chains of IL-12, p35 and p40. We found that these genes were

225

strongly upregulated in the liver from LPS-challenged IL-6TG mice and significantly higher, both in

226

liver and spleen, in IL-6TG mice compared to WT mice (Figure 1C and D). In IL-6TG mice liver,

227

strong positive correlations between Il-12a and Il-12b and Ifnγ mRNA expression were also found

228

(r= 0.93, p=0.001 and r=.083, p=.007 respectively), further suggesting a possible role for IL-12 in

229

inducing IFNγ production in the tissue.

230

When we measured circulating levels of IFNγ, low levels, in the pg/ml range, were detected, with a

231

trend for higher levels in LPS-challenged IL-6TG mice compared to WT mice (Figure 2). These

232

results are consistent with recent observations demonstrating that production of IFNγ typically

233

occurs in peripheral tissues(28). In this respect, circulating levels of IFNγ inducible chemokines,

234

CXCL9 and CXCL10 have been suggested as possible markers of tissue IFNγ production. Indeed in

235

LPS-challenged IL-6TG mice, high levels of CXCL9 and CXCL10 were found in plasma (Figure 2).

236

Interestingly, we found also that circulating levels of CXCL9 were significantly correlated with liver

237

mRNA levels (18 hours post LPS administration) of Ifnγ (r= 0.83, p=0.029) and of Cxcl9 (r=0.89,

238

p=0.017). Similarly, circulating levels of CXCL10 were significantly correlated to liver mRNA

239

levels of Ifnγ (r= 0.82, p=0.030) and of Cxcl10 (r= 0.89, p=0.017). We also found a tight correlation

240

between plasma CXCL9 levels and spleen Cxcl9 mRNA levels (r=1.0, p=0.001). The same analyses

241

were performed in WT mice: no correlations were observed. Furthermore, and consistently with

242

previously published data at 6 hours from LPS challenge(20), at 18 hours, levels of pro-

AC C

EP

TE D

M AN U

SC

RI PT

219

9

ACCEPTED MANUSCRIPT inflammatory cytokines including IL-1β, TNF-α and IL-6 were markedly higher in LPS-challenged

244

IL-6TG mice compared with WT mice (Figure 2).

245

Altogether, our data show that in the mouse model of MAS, Ifnγ expression is increased, STAT1-

246

mediated IFNγ pathway is upregulated and the IFNγ inducible chemokines are upregulated locally

247

and systemically. The significant correlation between Ifnγ and chemokine transcript levels in

248

peripheral tissues and chemokine levels in the blood supports the conclusion that circulating levels

249

of IFNγ inducible chemokines reflect tissue activation of the IFNγ pathway as well as tissue

250

production of the IFNγ inducible chemokines.

251

Incidentally, liver hematoxylin and eosin staining did not reveal major abnormalities in hepatocytes.

252

However, dilated sinusoids populated by mononucleated cells were observed in LPS-challenged IL-

253

6TG mice, but not in WT mice (Figure 3A). Macrophages are most probably key early cells in the

254

pathogenesis of MAS; indeed, immunohistochemical analyses showed a markedly higher number of

255

CD68 (a pan-macrophage marker) positive cells in livers from IL-6TG mice compared to WT mice

256

(Figure 3B), suggesting a role for macrophages in the development of the disease in IL-6TG mice.

257

Neutralization of IFNγ improves survival, body weight recovery and laboratory parameters in

258

LPS-challenged IL-6TG mice. Having demonstrated that the IFNγ pathway is upregulated in

259

murine MAS induced by LPS challenge in IL-6TG animals, the effect of the neutralization of IFNγ

260

was tested by administering an anti-IFNγ antibody. Following control antibody administration , all

261

mice challenged with a dose of LPS known to be lethal in 100% of the animals (7.5 µg/g of body

262

weight) died within 4 days, while 70% of mice treated with the anti-IFNγ antibody survived (Figure

263

4A). Similarly, following control antibody administration, 5/10 mice challenged with a dose of LPS

264

known to be lethal in 50% of the animals (5 µg/g of body weight) died within 4 days, while all mice

265

treated with the anti-IFNγ antibody survived (Figure 4B). No additional deaths occurred in the

266

following days, in both groups of mice, in both experiments (not shown).

AC C

EP

TE D

M AN U

SC

RI PT

243

10

ACCEPTED MANUSCRIPT In mice injected with 5 µg/g of LPS, anti-IFNγ treatment significantly improved body weight

268

recovery, with animals reaching their baseline weight within 3 days (Figure 4C). After 12 hours

269

from LPS administration, a marked elevation of ferritin levels was observed, consistent with

270

previous data (20). While in control antibody-treated IL-6TG mice, ferritin levels did not change, a

271

sharp decrease was observed in mice treated with the anti-IFNγ antibody (Figure 4D). After LPS

272

administration, plasma fibrinogen, an additional laboratory feature of MAS, was significantly lower

273

in IL-6TG mice compared to WT mice (mean ± SD, 4229 µg/ml ± 1885 vs 6654 µg/ml ± 1526, p=

274

0.01), suggesting increased intravascular coagulation activation in this animal model. We found that

275

in LPS-challenged IL-6TG mice fibrinogen levels were significantly higher in mice treated with the

276

anti-IFNγ antibody, compared to mice treated with the control antibody (Figure 4E). In order to

277

biochemically evaluate liver involvement, we measured alanine aminotransferase (ALT) levels. As

278

previously demonstrated, we did not observe a significant difference in circulating ALT levels

279

between LPS-challenged IL-6TG and WT mice both at baseline and 12 hours after LPS (mean ±

280

SD: basal levels, 13.58 ± 6.8 vs 12.2 ± 10.07 U/L; 12 hours after LPS administration, 40.2 ± 21.24

281

vs 38.7 ± 15.30 U/L). In IL-6TG mice following administration of the anti-IFNγ antibody, ALT

282

levels were significantly reduced compared to control-treated animals (Figure 4F).

283

IFNγ inducible chemokines and proinflammatory cytokines are downregulated in mice

284

treated with the anti-IFNγ antibody and are related to changes in ferritin levels during

285

effective treatment. In anti-IFNγ antibody treated IL-6TG mice plasma levels of CXCL9 and

286

CXCL10 were markedly lower than those of mice treated with control antibody (Figure 5A).

287

Moreover, plasma levels of IL-1β, IL-6 and TNF-α, three classical proinflammatory cytokines,

288

were significantly lower in anti-IFNγ antibody treated than in control antibody treated mice (Figure

289

5B). Finally, in LPS-challenged mice treated with the control and the anti-IFNγ antibody, lower

290

circulating levels of CXCL9 and CXCL10 were significantly associated with lower levels of IL-6,

291

IL-1β, TNF-α (Figure 6A), as well as to lower serum ferritin levels (Figure 6B and C). These

AC C

EP

TE D

M AN U

SC

RI PT

267

11

ACCEPTED MANUSCRIPT 292

results support the conclusion that decrease in IFNγ activity, estimated by the decrease in CXCL9

293

and CXCL10 levels, is related to downstream events, such as inflammatory cytokine levels and

294

clinical parameters of disease.

295

DISCUSSION

297

While several observations in murine models and indirect evidence in patients point to a pivotal

298

pathogenic role of IFNγ in pHLH(17, 18), limited data are available on its role in the pathogenesis

299

of the different forms of secondary HLH. In this study, we have used a murine model of MAS in

300

which clinical and laboratory features of MAS are induced by administration of a TLR ligand in IL-

301

6 transgenic mice(20). We found that IL-6TG mice challenged with LPS display an upregulation of

302

the IFNγ pathway, as shown by higher mRNA expression levels of Ifnγ, higher levels of phospho-

303

STAT1 and of the IFNγ inducible genes Cxcl9 and Cxcl10 in liver and spleen, as well as higher

304

levels of circulating CXCL9 and CXCL10. Moreover, we found that treatment of IL-6TG mice with

305

an anti-IFNγ antibody improved survival, body weight and laboratory parameters. Finally,

306

neutralization of IFNγ led also to a marked decrease in circulating levels of CXCL9 and CXCL10,

307

chemokines typically induced by IFNγ, as well as of the downstream inflammatory cytokines, TNF-

308

α, IL-6 and IL-1β.

309

Despite the evidence that in normal mice systemic overexposure to high levels of IFNγ is sufficient

310

to drive cytopenias and hemophagocytosis in a STAT1-dependent manner(29), the data on the role

311

of IFNγ in models of secondary HLH are scarce and, in part, contradictory. In a model of HLH

312

secondary to infections induced by TLR-9 repeated stimulation, on a normal genetic background

313

and in the absence of underlying disease, an important pathogenic role for IFNγ is reported(19).

314

Moreover, using IFNγ overexpressing mice, IFNγ has been identified as a mediator of systemic

315

inflammatory disease, supporting the hypothesis that there is a critical threshold of IFNγ that, when

316

achieved, either locally in tissues or systemically, drives the development of an autoinflammatory-

317

like disease, interestingly with high ferritin(30). In apparent contrast, immune stimulation of IFNγ

AC C

EP

TE D

M AN U

SC

RI PT

296

12

ACCEPTED MANUSCRIPT knock-out (KO) mice with Freund's complete adjuvant produces a systemic inflammatory disease

319

that includes features of sJIA, as well as of MAS, such as anemia, increased numbers of immature

320

blood cells, increased serum levels of IL-6, hemophagocytosis and defective natural killer cell

321

cytotoxicity; it is noteworthy that cytopenias and hyperferritemia were not demonstrated in this

322

model(31). Canna et al.(27) demonstrated that, in mice treated with an IL-10 receptor blocking

323

antibody and a TLR-9 agonist, fulminant MAS and hemophagocytosis arise. When these mice were

324

knocked-out for IFNγ, they developed immunopathology and hemophagocytosis comparable to that

325

seen in WT mice, but did not become anemic and had greater numbers of splenic erythroid

326

precursors(27). Similarly, in a mouse model of HLH associated to cytomegalovirus sepsis, IFNγ

327

KO mice display a more severe spectrum of the disease(32), therefore suggesting a protective role

328

for IFNγ. The interpretation of these results, which appear to be at least in part contradictory, may

329

be particularly difficult, because of compensatory mechanisms secondary to lack or increased

330

expression of IFNγ since prenatal age in these animal models. Although these approaches provide a

331

wealth of information on the pathophysiology of hyperinflammation, these compensatory

332

mechanisms makes it difficult to directly extrapolate the findings into a therapeutic approach in

333

individuals with a “normal” immune response. Indeed, recently two unrelated patients with novel

334

mutations in the IFNγ receptors, highly suggestive of functional receptor deficiency, have been

335

described with a clinical presentation that closely resembled HLH in the context of overwhelming

336

mycobacterial infections (with associated herpes virus infections)(33). This observation has

337

suggested that IFNγ independent pathways may contribute to the development of the clinical

338

syndrome HLH or at least to some of the characteristic features.

339

As previously mentioned, animal models of pHLH caused by deletion of the genes involved in

340

pHLH, on an otherwise normal immune response, have unequivocally demonstrated a pivotal

341

pathogenic role of IFNγ(17, 18). We have used a model in which MAS clinical and laboratory

342

features are triggered in animals with high levels of circulating human IL-6 since early phases of

AC C

EP

TE D

M AN U

SC

RI PT

318

13

ACCEPTED MANUSCRIPT life (not prenatally), with a normally regulated expression of IFNγ(20). Our model has similarities

344

with MAS development in patients with sJIA, where MAS is typically triggered by acute infections

345

on an underlying active sJIA, tipically characterized by increased IL-6 levels.

346

Our data show high mRNA expression of IFNγ, associated to high Il-12a and Il-12b mRNA

347

expression, in liver and spleen, suggesting that IL-12 may be at least one of the factors involved in

348

the increased production of IFNγ in the presence of high in vivo levels of IL-6. The overexpression

349

of IFNγ is biologically relevant, as shown by the prolonged phosphorylation of STAT1 in liver and

350

spleen and by the marked upregulation of the liver and spleen expression and plasma levels of IFNγ

351

inducible chemokines. These observations in liver and spleen, which are target tissues in MAS,

352

appear to complement observations in patients with secondary HLH(34) and MAS. Billiau et al.

353

(35) showed the presence of IFNγ-producing activated CD8+ lymphocytes in 4 liver biopsies of

354

patients with secondary HLH, including one patient with MAS. More recently, high levels of IFNγ

355

and CXCL10 in the sera of 5 patients with HLH, three of them with MAS, have also been

356

reported(36). We have recently found high circulating levels of IFNγ and of IFNγ inducible

357

chemokines in 20 patients with active MAS, but not in patients with active sJIA without MAS(37).

358

It should be noted that gene expression profile studies of PBMCs of patients with active sJIA, with

359

or without MAS, did not show an IFNγ signature(38-40). However, while gene expression profile

360

studies on PBMCs reveal events that occur in peripheral blood, our animal data in liver and spleen

361

support the hypothesis that, in MAS patients, marked production of IFNγ and of IFNγ inducible

362

proteins occurs first in diseased tissues and then proteins leak into peripheral blood. Supporting this

363

hypothesis, we found that in LPS-challenged IL-6TG mice there was a significant correlation of

364

circulating levels of CXCL9 and CXCL10 with liver/spleen mRNA levels of Cxcl9 and Cxcl10 and,

365

more importantly, with tissue mRNA levels of Ifnγ. Notably, Put et al.(36) reported evident

366

expression of the IFNγ inducible proteins indoleamine 2,3-dioxygenase (IDO) and CXCL10 in one

367

lymph node biopsy from one sJIA patient with MAS. Furthermore, some of us demonstrated in the

368

model of HLH secondary to infection induced by repeated TLR-9 stimulation that total IFNγ levels

AC C

EP

TE D

M AN U

SC

RI PT

343

14

ACCEPTED MANUSCRIPT produced in tissues are 500- to 2,000-fold higher than those measured in blood and identified spleen

370

and liver as major sites of IFNγ production(28). These results are consistent with our finding of a

371

trend for increased circulating levels of IFNγ in our MAS mice. All together, these observations in

372

MAS patients and in MAS animal model support the hypothesis that IFNγ production is markedly

373

increased, that high IFNγ production is biologically relevant and that the over-activation of the

374

IFNγ pathway appears to occur in peripheral tissues which are typically involved by the disease,

375

such as liver and spleen.

376

We also demonstrated the pathogenic role of IFNγ in the murine MAS model. Treatment with an

377

anti-IFNγ antibody led to a significant increase in survival and reverted clinical and laboratory

378

features of MAS, including body weight loss and ferritin, ALT and fibrinogen levels. Furthermore,

379

neutralization of IFNγ led to a significant decrease in classical downstream proinflammatory

380

cytokines. Interestingly, we found that, in LPS-challenged mice, circulating CXCL9 and CXCL10

381

levels were significantly related to IL-1β, IL-6, TNF-α as well as to ferritin levels. Further

382

supporting the relevance of IFNγ neutralization, we showed that the extent of change in upstream

383

events (i.e. CXCL9 and CXCL10 levels) are related to the extent of the decrease in downstream

384

events, such as levels of proinflammatory cytokines, or even more so, levels of ferritin, a classical

385

laboratory parameter of disease activity in clinical practice. Consistently with these data in mice, we

386

found that in in patients with MAS, sampled longitudinally during effective traditional treatment,

387

the circulating levels of both CXCL9 and CXCL10 paralleled the decrease in ferritin during the

388

progressive clinical improvement (data not shown). The data in animals and humans suggest also

389

that serum levels of CXCL9, and possibly of CXCL10, since they reflect tissue production of IFNγ,

390

may represent additional biomarkers of disease activity in MAS. Their relative value compared to

391

standard biochemical parameters, such as ferritin, in identifying patients at the early stages of MAS

392

before it progresses to its life-threatening stage should be investigated in a larger number of

393

patients, possibly in a multicenter study.

AC C

EP

TE D

M AN U

SC

RI PT

369

15

ACCEPTED MANUSCRIPT Our experimental model of MAS has the limitation of having short duration (death occurs in 2-4

395

days after LPS administration), representing therefore an acute model of the disease useful to study

396

early stages (i.e induction mechanisms) of the disease. In contrast, we cannot evaluate improvement

397

in late events, such as tissue infiltration and damage (fibrosis/collagen deposition). Nevertheless,

398

this MAS model has allowed us to deepen our understanding of the early events that are upstream

399

of the development of the disease. Indeed, we showed that in the liver from MAS mice the number

400

of CD68 positive macrophages and the expression of IL-12, mainly produced by M1 macrophages,

401

was increased compared to WT mice.

402

In conclusion, in this study, we have found a role for high levels of IL-6 in inducing an increased

403

macrophage infiltration and an upregulation of Il-12 expression in liver and demonstrated the

404

pivotal role of IFNγ in a murine model of MAS. Our data, together with those obtained in animal

405

models of pHLH and of HLH secondary to infections(17-19, 28), support the hypothesis that HLH

406

forms of different origin share a common inflammatory effector pathway, involving IFNγ. Our

407

results have also significant implications for the treatment of patients with MAS. Encouraging

408

preliminary efficacy and safety data of the phase 2 clinical trial in primary HLH with the anti-IFNγ

409

monoclonal antibody NI-0501 have been recently reported(41).

TE D

M AN U

SC

RI PT

394

EP

410 411

AC C

412 413

Authorship

414

Contribution: G.P. contributed to the research and experimental design, contributed to perform in

415

vivo and in vitro experiments, analyzed data and wrote the manuscript; I.C. performed in vivo and in

416

vitro experiments; A.P., R.S. and E.M. performed in vitro and ex vivo experiments on phospho-

417

STAT1 and analyzed these data; L.C. performed MILLIPLEX MAP assays in humans and mice

418

blood samples; C.B. and A.A.G. collected samples and clinical data from patients, analyzed and

419

wrote the sections on MAS patients; W.G.F. performed MILLIPLEX MAP assays and provided PK 16

ACCEPTED MANUSCRIPT 420

data in mice that guided dosing of the anti-IFNγ antibody; C.d.M. contributed to research design

421

and interpretation and to the writing of the manuscript; F.D.B. designed the research, supervised the

422

research design and laboratory activities and wrote the manuscript.

423

Conflict-of-interest disclosure

425

Dr. De Benedetti’s Institution received unrestricted research grants from Pfizer, AbbVie, Novartis,

426

Novimmune, Roche, SOBI, Sanofi, Gilead and received travel support from Roche and Novartis.

427

Laurence Chatel, Walter G. Ferlin, and Cristina de Min are employees of Novimmune SA. Dr.

428

Alexei Grom received consulting fees and continues research collaboration with Novartis and

429

NovImmune. Other authors declare that they have no competing interests.

M AN U

SC

RI PT

424

430 431 432

TE D

433 434 435

EP

436

AC C

437 438 439 440 441 442 443 444 445 17

ACCEPTED MANUSCRIPT 446

REFERENCES

448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496

1. Ravelli A, Grom AA, Behrens EM, Cron RQ. Macrophage activation syndrome as part of systemic juvenile idiopathic arthritis: diagnosis, genetics, pathophysiology and treatment. Genes and immunity. 2012;13(4):289-98. 2. Grom AA, Horne A, De Benedetti F. Macrophage activation syndrome in the era of biologic therapy. Nature reviews Rheumatology. 2016;12(5):259-68. 3. Schulert GS, Grom AA. Macrophage activation syndrome and cytokine-directed therapies. Best practice & research Clinical rheumatology. 2014;28(2):277-92. 4. Freeman HR, Ramanan AV. Review of haemophagocytic lymphohistiocytosis. Arch Dis Child. 2011;96(7):688-93. 5. Schulert GS, Grom AA. Pathogenesis of macrophage activation syndrome and potential for cytokine- directed therapies. Annual review of medicine. 2015;66:145-59. 6. Vastert SJ, Prakken BJ. Paediatric rheumatic disease: Diagnosing macrophage activation syndrome in systemic JIA. Nature reviews Rheumatology. 2014;10(11):640-2. 7. Risma K, Jordan MB. Hemophagocytic lymphohistiocytosis: updates and evolving concepts. Current opinion in pediatrics. 2012;24(1):9-15. 8. Strippoli R, Caiello I, De Benedetti F. Reaching the threshold: a multilayer pathogenesis of macrophage activation syndrome. The Journal of rheumatology. 2013;40(6):761-7. 9. Grom AA, Villanueva J, Lee S, Goldmuntz EA, Passo MH, Filipovich A. Natural killer cell dysfunction in patients with systemic-onset juvenile rheumatoid arthritis and macrophage activation syndrome. The Journal of pediatrics. 2003;142(3):292-6. 10. Wulffraat NM, Rijkers GT, Elst E, Brooimans R, Kuis W. Reduced perforin expression in systemic juvenile idiopathic arthritis is restored by autologous stem-cell transplantation. Rheumatology. 2003;42(2):375-9. 11. Zhang M, Behrens EM, Atkinson TP, Shakoory B, Grom AA, Cron RQ. Genetic defects in cytolysis in macrophage activation syndrome. Current rheumatology reports. 2014;16(9):439. 12. Kaufman KM, Linghu B, Szustakowski JD, Husami A, Yang F, Zhang K, et al. Whole-exome sequencing reveals overlap between macrophage activation syndrome in systemic juvenile idiopathic arthritis and familial hemophagocytic lymphohistiocytosis. Arthritis & rheumatology. 2014;66(12):3486-95. 13. Vastert SJ, van Wijk R, D'Urbano LE, de Vooght KM, de Jager W, Ravelli A, et al. Mutations in the perforin gene can be linked to macrophage activation syndrome in patients with systemic onset juvenile idiopathic arthritis. Rheumatology. 2010;49(3):441-9. 14. Zhang K, Biroschak J, Glass DN, Thompson SD, Finkel T, Passo MH, et al. Macrophage activation syndrome in patients with systemic juvenile idiopathic arthritis is associated with MUNC13-4 polymorphisms. Arthritis and rheumatism. 2008;58(9):2892-6. 15. Sepulveda FE, Garrigue A, Maschalidi S, Garfa-Traore M, Menasche G, Fischer A, et al. Polygenic mutations in the cytotoxicity pathway increase susceptibility to develop HLH immunopathology in mice. Blood. 2016;127(17):2113-21. 16. Xu XJ, Tang YM, Song H, Yang SL, Xu WQ, Zhao N, et al. Diagnostic accuracy of a specific cytokine pattern in hemophagocytic lymphohistiocytosis in children. The Journal of pediatrics. 2012;160(6):984-90 e1. 17. Jordan MB, Hildeman D, Kappler J, Marrack P. An animal model of hemophagocytic lymphohistiocytosis (HLH): CD8+ T cells and interferon gamma are essential for the disorder. Blood. 2004;104(3):735-43. 18. Pachlopnik Schmid J, Ho CH, Chretien F, Lefebvre JM, Pivert G, Kosco-Vilbois M, et al. Neutralization of IFNgamma defeats haemophagocytosis in LCMV-infected perforin- and Rab27a-deficient mice. EMBO molecular medicine. 2009;1(2):112-24. 19. Behrens EM, Canna SW, Slade K, Rao S, Kreiger PA, Paessler M, et al. Repeated TLR9 stimulation results in macrophage activation syndrome-like disease in mice. The Journal of clinical investigation. 2011;121(6):2264-77. 18

AC C

EP

TE D

M AN U

SC

RI PT

447

ACCEPTED MANUSCRIPT

EP

TE D

M AN U

SC

RI PT

20. Strippoli R, Carvello F, Scianaro R, De Pasquale L, Vivarelli M, Petrini S, et al. Amplification of the response to Toll-like receptor ligands by prolonged exposure to interleukin-6 in mice: implication for the pathogenesis of macrophage activation syndrome. Arthritis and rheumatism. 2012;64(5):1680-8. 21. De Benedetti F, Brunner HI, Ruperto N, Kenwright A, Wright S, Calvo I, et al. Randomized trial of tocilizumab in systemic juvenile idiopathic arthritis. N Engl J Med. 2012;367(25):2385-95. 22. de Benedetti F, Martini A. Targeting the interleukin-6 receptor: a new treatment for systemic juvenile idiopathic arthritis? Arthritis and rheumatism. 2005;52(3):687-93. 23. De Benedetti F, Alonzi T, Moretta A, Lazzaro D, Costa P, Poli V, et al. Interleukin 6 causes growth impairment in transgenic mice through a decrease in insulin-like growth factor-I. A model for stunted growth in children with chronic inflammation. The Journal of clinical investigation. 1997;99(4):643-50. 24. Hu X, Ivashkiv LB. Cross-regulation of signaling pathways by interferon-gamma: implications for immune responses and autoimmune diseases. Immunity. 2009;31(4):539-50. 25. Schindler C, Levy DE, Decker T. JAK-STAT signaling: from interferons to cytokines. The Journal of biological chemistry. 2007;282(28):20059-63. 26. Trinchieri G. Interleukin-12 and the regulation of innate resistance and adaptive immunity. Nature reviews Immunology. 2003;3(2):133-46. 27. Canna SW, Wrobel J, Chu N, Kreiger PA, Paessler M, Behrens EM. Interferon-gamma mediates anemia but is dispensable for fulminant toll-like receptor 9-induced macrophage activation syndrome and hemophagocytosis in mice. Arthritis and rheumatism. 2013;65(7):1764-75. 28. Buatois V, Chatel L, Cons L, Lory S, Richard F, Guilhot F, et al. Use of a mouse model to identify a blood biomarker for IFNgamma activity in pediatric secondary hemophagocytic lymphohistiocytosis. Translational research : the journal of laboratory and clinical medicine. 2016. 29. Zoller EE, Lykens JE, Terrell CE, Aliberti J, Filipovich AH, Henson PM, et al. Hemophagocytosis causes a consumptive anemia of inflammation. J Exp Med. 2011;208(6):1203-14. 30. Reinhardt RL, Liang HE, Bao K, Price AE, Mohrs M, Kelly BL, et al. A novel model for IFN-gammamediated autoinflammatory syndromes. Journal of immunology. 2015;194(5):2358-68. 31. Avau A, Mitera T, Put S, Put K, Brisse E, Filtjens J, et al. Systemic juvenile idiopathic arthritis-like syndrome in mice following stimulation of the immune system with Freund's complete adjuvant: regulation by interferon-gamma. Arthritis & rheumatology. 2014;66(5):1340-51. 32. Brisse E, Imbrechts M, Put K, Avau A, Mitera T, Berghmans N, et al. Mouse Cytomegalovirus Infection in BALB/c Mice Resembles Virus-Associated Secondary Hemophagocytic Lymphohistiocytosis and Shows a Pathogenesis Distinct from Primary Hemophagocytic Lymphohistiocytosis. Journal of immunology. 2016;196(7):3124-34. 33. Tesi B, Sieni E, Neves C, Romano F, Cetica V, Cordeiro AI, et al. Hemophagocytic lymphohistiocytosis in 2 patients with underlying IFN-gamma receptor deficiency. The Journal of allergy and clinical immunology. 2015;135(6):1638-41. 34. Takada H, Takahata Y, Nomura A, Ohga S, Mizuno Y, Hara T. Increased serum levels of interferongamma-inducible protein 10 and monokine induced by gamma interferon in patients with haemophagocytic lymphohistiocytosis. Clinical and experimental immunology. 2003;133(3):448-53. 35. Billiau AD, Roskams T, Van Damme-Lombaerts R, Matthys P, Wouters C. Macrophage activation syndrome: characteristic findings on liver biopsy illustrating the key role of activated, IFN-gamma-producing lymphocytes and IL-6- and TNF-alpha-producing macrophages. Blood. 2005;105(4):1648-51. 36. Put K, Avau A, Brisse E, Mitera T, Put S, Proost P, et al. Cytokines in systemic juvenile idiopathic arthritis and haemophagocytic lymphohistiocytosis: tipping the balance between interleukin-18 and interferon-gamma. Rheumatology. 2015;54(8):1507-17. 37. Bracaglia C, de Graaf K, Pires Marafon D, Guilhot F, Ferlin W, Prencipe G, et al. Elevated circulating levels of interferon-gamma and interferon-gamma-induced chemokines characterise patients with macrophage activation syndrome complicating systemic juvenile idiopathic arthritis. Annals of the rheumatic diseases. 2017;76(1):166-72. 38. Fall N, Barnes M, Thornton S, Luyrink L, Olson J, Ilowite NT, et al. Gene expression profiling of peripheral blood from patients with untreated new-onset systemic juvenile idiopathic arthritis reveals

AC C

497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547

19

ACCEPTED MANUSCRIPT molecular heterogeneity that may predict macrophage activation syndrome. Arthritis and rheumatism. 2007;56(11):3793-804. 39. Sikora KA, Fall N, Thornton S, Grom AA. The limited role of interferon-gamma in systemic juvenile idiopathic arthritis cannot be explained by cellular hyporesponsiveness. Arthritis and rheumatism. 2012;64(11):3799-808. 40. Ogilvie EM, Khan A, Hubank M, Kellam P, Woo P. Specific gene expression profiles in systemic juvenile idiopathic arthritis. Arthritis and rheumatism. 2007;56(6):1954-65. 41. Jordan M, Allen C, De Benedetti F, Grom AA, ; Ballabio M, Ferlin WG, et al. A Novel Targeted Approach to the Treatment of Hemophagocytic Lymphohistiocytosis (HLH) with an Anti-Interferon Gamma (IFNγ) Monoclonal Antibody (mAb), NI-0501: First Results from a Pilot Phase 2 Study in Children with Primary HLH. . 57th Annual Meeting & Exposition LBA-3 2015.

RI PT

548 549 550 551 552 553 554 555 556 557 558 559 560

SC

561 562

M AN U

563 564 565 566 567 568

TE D

569 570 571 572

EP

573 574

AC C

575 576 577 578 579 580 581 582 583 584 585 20

ACCEPTED MANUSCRIPT 586 587 588

FIGURE LEGENDS

589

SC

RI PT

Figure 1. LPS-challenged IL-6TG mice display upregulation of the IFNγ pathway and increased mRNA expression of Il-12a and Il-12b. Wild type (WT) and IL-6TG mice were challenged with LPS (5 µg/g of body weight) and, at different times after the challenge, mice were sacrificed. mRNA expression levels of Ifnγ and of the IFNγ regulated genes Cxcl9 and Cxcl10 were assayed in liver (A) and spleen (B). The results are expressed as arbitrary units (AU) and obtained after normalization with the housekeeping gene Hprt. Tyr701 Phosphorylated STAT1 protein levels were assessed by Western Blot Analysis in liver (A) and spleen (B). In A and B, on the right panels, densitometric analyses confirmed the increased levels of Tyr701 Phosphorylated STAT1 in liver and spleen from IL-6TG mice compared to WT mice (n=4 or 5 in each group). mRNA expression levels of Il-12a and Il-12b were also evaluated in liver (C) and spleen (D) Data in A-D (mean ± SEM) are representative of at least 2 independent experiments with at least 3 mice in each group. Statistical analyses were performed using one-tailed Mann-Whitney U test. * p<0.05, ** p<0.01, *** p<0.001

M AN U

590 591 592 593 594 595 596 597 598 599 600 601 602 603

Figure 2. Circulating levels of IFNγ inducible chemokines and proinflammatoy cytokines are increased in LPS-challenged IL-6TG mice. Plasma levels of IFNγ, IFNγ inducible chemokines CXCL9 and CXCL10 and of proinflammatory cytokines IL-1β, IL-6 and TNF-α were measured in WT and IL-6TG mice at 18 hours after the challenge with LPS (5 µg/g of body weight). Data (mean ± SEM) are representative of at least 2 independent experiments with at least 3 mice in each group. Statistical analyses were performed using one-tailed Mann-Whitney U test. * p<0.05, ** p<0.01

TE D

604 605 606 607 608 609 610

Figure 3. Presence of dilated sinusoids and of infiltration by CD68 positive macrophages

EP

in liver from LPS-challenged IL-6TG mice. Liver sections from WT and IL-6TG mice, at 12 hours after the challenge with LPS (5 µg/g of body weight), were stained with H&E. Dilated sinusoids populated by mononucleated cells are indicated by arrows (A). Representative sections are shown at a magnification of ×40. Representative images of liver sections from WT and IL-6TG mice stained for total macrophages using the CD68 antibody at 12 hours after the challenge with LPS (5 µg/g of body weight) (B). CD68 positive cells are indicated by arrows. Scale bars: 200 µm. On the right panel, the number of CD68 positive cells per field was reported. Data (mean ± SEM) are representative of at least 2 independent experiments with at least 2 mice in each group. Statistical analyses were performed using one-tailed Mann-Whitney U test. * p<0.05

AC C

611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627

Figure 4. Treatment with a neutralizing anti-IFNγ antibody improves survival and clinical and laboratory parameters of MAS in LPS-challenged IL-6TG mice. IL-6TG mice were challenged with 7.5 µg/g of body weight of LPS (A) or with 5 µg/g of body weight of LPS (B) at day 0 and, 12 hours after the challenge, mice were treated with the IFNγ neutralizing antibody XMG2.1 (100 µg/g of body weight) or with the control antibody mAb35 and were monitored for survival. Survival rates were determined every day for 4 days (n=10 mice per group). Fisher’s exact 21

ACCEPTED MANUSCRIPT 628 629 630 631 632 633 634

test was performed. (C-F) IL-6TG mice were challenged with LPS (5 µg/g of body weight) at day 0, treated as before described with the XMG2.1 or the mAb35 control antibody, and clinical and laboratory parameters were evaluated: body weight variation (C), serum ferritin (D), plasma fibrinogen (E) and plasma ALT levels (F). In A, average body weight loss data are presented with last observation carried forward for mice that succumbed to LPS challenge. Data (mean ± SEM) are representative of at least 2 independent experiments with at least 3 mice in each group. Statistical analyses were performed using one-tailed Mann-Whitney U test. * p<0.05, ** p<0.01

SC

Figure 5. In mice treated with the anti-IFNγ antibody, circulating levels of proinflammatory cytokines are downmodulated. IL-6TG mice were challenged with LPS (5 µg/g of body weight) and, 12 hours after the challenge, mice were treated with the IFNγ neutralizing antibody XMG2.1 (100 µg/g body weight) or with the mAb35 control antibody. 18 hours after the treatment with antibodies, mice were sacrificed and levels of CXCL9 and CXCL10 (A) and of proinflammatory cytokines (B), including IL-1β, IL-6 and TNF-α were assayed in plasma samples collected at time of sacrifice. Data in A and B (mean ± SEM) are representative of at least 2 independent experiments with at least 3 mice in each group. Statistical analyses were performed using one-tailed MannWhitney U test. * p<0.05, ** p<0.01

M AN U

636 637 638 639 640 641 642 643 644

RI PT

635

645

Figure 6. Circulating levels of IFNγγ inducible chemokines CXCL9 and CXCL10 are correlated to changes in proinflammatory cytokine and ferritin levels during effective treatment in IL-6TG mice. IL-6TG mice were challenged with LPS (5 µg/g of body weight) and, 12 hours after the challenge, mice were treated with the IFNγ neutralizing antibody XMG2.1 (100 µg/g body weight) (filled square) or with the mAb35 control antibody (open square). 18 hours after the treatment with antibodies, mice were sacrificed and levels of CXCL9 and CXCL10 were measured in plasma and related to circulating levels of proinflammatory cytokines IL-6, IL-1β and TNF-α (A) as well as to serum ferritin levels (B and C). In A-C, the Spearman rank correlation coefficient was calculated (r) to assess the correlations between variables.

TE D

646 647 648 649 650 651 652 653 654

EP

655

AC C

656

22

Figure 1 A A 0.02 0.01 0.00

0

B

Hours

B

80

Liver Cxcl9 mRNA (AU)

**

60

40ACCEPTED

MANUSCRIPT

20 4 2 0

12

0

Hours

5 2 1 0

0

Hours

12

0

Hours

6 4 2 0

0

Hours

12

STAT1

pSTAT1/STAT1 relative density (AU)

pSTAT1

18

***

0.04 0.03 0.02 0.01 0.00

0

Hours

12

Liver Il-12b mRNA (AU)

WT IL-6TG

0.05

0.20

***

0.15 0.10 0.05 0.00

0

25

18

*

20 15 10 5 0

0

Hours

12

Hours

12

1.5

F

*

1.0 0.5 0.0

Spleen Il-12a mRNA (AU)

12

IL-6TG

WT

WT

6

IL-6TG

IL-6TG

WT

WT

IL-6TG

GAPDH

0

6 12 Hours

WT IL-6TG

AC C

D

12

0

0

6 12 Hours

18

WT IL-6TG

0.5

***

0.4 0.3 0.2

*

0.1 0.0

0

Hours

12

Spleen Il-12b mRNA (AU)

0.05

D

8

TE D

0.10

0.00

10

Spleen Cxcl10 mRNA (AU)

Spleen Cxcl9 mRNA (AU)

*

0.15

0.5 0.0

18

WT IL-6TG

0.20

**

1.0

SC

WT

12

1.5

M AN U

6

IL-6TG

WT

IL-6TG

WT

0

EP

Spleen IfnJ mRNA (AU)

IL-6TG

WT Hours

IL-6TG

GAPDH

Liver Il-12a mRNA (AU)

10

RI PT

STAT1

E

*

15

WT IL-6TG

pSTAT1

C C

20

12

pSTAT1/STAT1 relative density (AU)

Liver IfnJ mRNA (AU)

*

0.03

Liver Cxcl10 mRNA (AU)

WT IL-6TG

0.04

0.08

***

0.06 0.04 0.02 0.00

0

Hours

12

Figure 2

ACCEPTED MANUSCRIPT

8000

10 5

6000 4000

M AN U

15

2000

0

4000 2000 0

TE D EP

9000 6000 3000 0

120

*** TNF-D (pg/ml)

50

12000

IL-6 (pg/ml)

IL-1E (pg/ml)

100

*

AC C

WT IL-6TG

150

*

6000

0

0

200

RI PT

20

8000

*

CXCL10 (pg/ml)

10000

SC

25 CXCL9 (pg/ml)

IFN-J (pg/ml)

WT IL-6TG

90 60 30 0

**

Figure 3 WT

B

WT

IL-6TG

WT IL-6TG

IL-6TG CD68 positive cells per field

AC C

CD68

EP

TE D

M AN U

SC

RI PT

A

ACCEPTED MANUSCRIPT

20 15 10 5 0

*

Figure 4 A

B

LPS (7,5 Pg/g body weight)

40 20

D

TE D

Ab

95

EP

85

0

1

2 3 Days

2 3 Days

15000

10000 5000

F

Ab LPS

5000

0 6 12 18 24 30 Hours

**

*

LPS

0 6 12 18 24 30 Hours

CTRL Ab Anti-IFN J Ab

60

*

10000

Ab

15000

0

4

CTRL Ab Anti-IFN J Ab

0

1

20000

ALT (U/L)

Weight variation (%)

25000

*

90

0

CTRL Ab Anti-IFN J Ab

LPS

100

20 0

4

CTRL Ab Anti-IFN J Ab

105

40

SC

2 3 Days

60

M AN U

1

Ferritin (ng/ml)

0

80

RI PT

60

0

Fibrinogen (Pg/ml)

Surviving mice (%)

**

80

0

E

*

100

AC C

Surviving mice (%)

100

C

LPS (5 Pg/g body weight)

CTRL Ab Anti-IFN J Ab ACCEPTED MANUSCRIPT

CTRL Ab Anti-IFN J Ab

Ab

*

40 LPS

20 0

0 6 12 18 24 30 Hours

4

Figure 5 AA CTRL Ab Anti-IFN J Ab

3000 2000 1000

3000 2000 1000

0

0

10 0

10000 8000 6000 4000 2000 0

*

25 TNF-D (pg/ml)

20

*

IL-6 (pg/ml)

IL-1E (pg/ml)

30

AC C

CTRL Ab Anti-IFN J Ab

EP

TE D

B B

40

**

SC

*

M AN U

4000

CXCL10 (pg/ml)

CXCL9 (pg/ml)

5000

RI PT

ACCEPTED MANUSCRIPT

20 15 10 5 0

*

Figure 6 A

ACCEPTED MANUSCRIPT

r=0.73 p=0.006

CXCL10

r=0.95 p<0.0001

r=0.68 p=0.01

Anti-IFNJ Ab

TE D EP AC C

r= 0,83 p= 0,0008

Ferritin (ng/ml)

1000

r= 0,76 p= 0,003

Ferritin (ng/ml)

00 00 10

0 00 10

00 10

0 10

00 00 10

0 00 10

00 10

0

10000

100

10

10

CXCL9 (pg/ml)

10000

100

r=0.92 p<0.0001

M AN U

C

CTRL Ab

1000

CXCL10

r=0.88 p=0.0002 r=0.83 p=0.008

CXCL10 (pg/ml)

B

TNF-D

RI PT

IL-1E

r=0.92 p<0.0001

SC

IL-6

CXCL9