Alarmin S100A8/S100A9 as a biomarker for molecular imaging of local inflammatory activity

Abstract

Inflammation has a key role in the pathogenesis of various human diseases. The early detection, localization and monitoring of inflammation are crucial for tailoring individual therapies. However, reliable biomarkers to detect local inflammatory activities and to predict disease outcome are still missing. Alarmins, which are locally released during cellular stress, are early amplifiers of inflammation. Here, using optical molecular imaging, we demonstrate that the alarmin S100A8/S100A9 serves as a sensitive local and systemic marker for the detection of even sub-clinical disease activity in inflammatory and immunological processes like irritative and allergic contact dermatitis. In a model of collagen-induced arthritis, we use S100A8/S100A9 imaging to predict the development of disease activity. Furthermore, S100A8/S100A9 can act as a very early and sensitive biomarker in experimental leishmaniasis for phagocyte activation linked to an effective Th1-response. In conclusion, the alarmin S100A8/S100A9 is a valuable and sensitive molecular target for novel imaging approaches to monitor clinically relevant inflammatory disorders on a molecular level.

ARTICLE

Received 7 May 2014 | Accepted 3 Jul 2014 | Published 6 Aug 2014

Thomas Vogl1,2,*, Michel Eisenblatter3,4,*, Tom Vller1, Stefanie Zenker1, Sven Hermann2,5, Peter van Lent6, Andreas Faust5, Christiane Geyer2,4, Beatrix Petersen1, Kirsten Roebrock1,2, Michael Schafers5,7, Christoph Bremer2,8 & Johannes Roth1,2,7

Inammation has a key role in the pathogenesis of various human diseases. The early detection, localization and monitoring of inammation are crucial for tailoring individual therapies. However, reliable biomarkers to detect local inammatory activities and to predict disease outcome are still missing. Alarmins, which are locally released during cellular stress, are early ampliers of inammation. Here, using optical molecular imaging, we demonstrate that the alarmin S100A8/S100A9 serves as a sensitive local and systemic marker for the detection of even sub-clinical disease activity in inammatory and immunological processes like irritative and allergic contact dermatitis. In a model of collagen-induced arthritis, we use S100A8/S100A9 imaging to predict the development of disease activity. Furthermore, S100A8/S100A9 can act as a very early and sensitive biomarker in experimental leishmaniasis for phagocyte activation linked to an effective Th1-response. In conclusion, the alarmin S100A8/S100A9 is a valuable and sensitive molecular target for novel imaging approaches to monitor clinically relevant inammatory disorders on a molecular level.

1 Institute of Immunology, University of Mnster, 48149 Mnster, Germany. 2 Interdisciplinary Centre for Clinical Research, University of Mnster, 48149 Mnster, Germany. 3 Division of Imaging Sciences and Biomedical Engineering, Kings College London, London SE1 7EH, UK. 4 Department of Clinical Radiology, University of Mnster, 48149 Mnster, Germany. 5 European Institute for Molecular Imaging, University of Mnster, 48149 Mnster, Germany.

6 Department of Rheumatology, Radboud University Medical Centre, 6500 HB Nijmegen, The Netherlands. 7 Cluster of Excellence EXC 1003 Cells in Motion- CiM, University of Mnster, 48149 Mnster, Germany. 8 Department of Radiology, St Franziskus Hospital Mnster, 48145 Mnster, Germany. * These authors contributed equally to this work. Correspondence and requests for materials should be addressed to Th.V. (email: mailto:[email protected]

Web End [email protected] ).

NATURE COMMUNICATIONS | 5:4593 | DOI: 10.1038/ncomms5593 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 1

& 2014 Macmillan Publishers Limited. All rights reserved.

DOI: 10.1038/ncomms5593 OPEN

Alarmin S100A8/S100A9 as a biomarker for molecular imaging of local inammatory activity

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5593

Inammation is the driving force in a vast spectrum of clinically relevant disorders, among others recognized as a major pathological mechanism in malignant and degenerative

diseases, infection and autoimmunity. Current imaging markers mostly reect either metabolism or secondary effects of inammatory reactions, such as increased perfusion or vessel permeability, or are only suitable for a very specic subset of diseases. In addition, all currently established biomarkers widely lack a proven prognostic potential. With biomedical research increasingly discovering the molecular and cellular basis of diseases and highly specic molecular therapies at the same time, both approaches do not provide sufcient diagnostic information. As a result, individually adapted therapy to manage chronic inammatory diseases remains widely elusive despite signicant therapeutic improvements1.

Numerous imaging approaches have been designed to address this issue. In vivo visualization of local inammation has been performed, for example, using F-18-uorodeoxyglucose (18F-FDG)-positron emission tomography (PET) or magnetic resonance imaging (MRI) with or without contrast enhancement2. Although these methods have proven diagnostic value, their implication in clinical practice has not fostered personalized therapy, mostly due to a lack of either desirable specicity (PET) or sensitivity (MRI).

Targeted imaging approaches to overcome these limitations would ideally address a biomarker with high expression/release or accumulation locally at the site of inammation, representative of early inammatory processes and residual disease activity or a prediction of are-ups of disease in remitting-relapsing courses of chronic inammation. In preclinical animal models, non-invasive molecular imaging methods would allow for local and long-itudinal assessment of biomarkers in individual subjects. In the long-term, such biomarkers would facilitate individual adaptation of medication and would lead to a signicant step forward in the concept of personalized medicine.

In recent years, the concept of alarmins or danger-associated molecular pattern molecules (DAMPs) has emerged as a novel mechanism for initiating and promoting inammation and has more recently been recognized as capable of resolving inammation36. Expressed and released during tissue damage or cellular stress reactions, members of this protein family have been shown to be early players in the development of inammatory processes. S100A8 and S100A9, two members of the DAMP-family, are highly expressed in early inltrating phagocytes. During the activation of these cells, S100A8/S100A9 complexes are locally released in virtually all inammatory disorders that are associated with phagocyte activation, like autoimmune diseases, rheumatoid arthritis, allergies, cardiovascular diseases, or local and systemic infections and tumours7, whereas virtually no expression can be found in healthy tissue. We have previously shown that S100A8 and S100A9 promote inammation via the activation of Toll-like

receptor-4 (refs 811). Serum concentrations of S100A8/S100A9 complexes have been shown to be superior over conventional biomarkers for the monitoring of inammatory disorders, especially in the detection of residual disease activity and in the prediction of relapse in arthritis12.

However, biomarkers measured in the blood only reect the systemic state, which is strongly affected by factors like metabolism or blood clearance, limiting the specicity and sensitivity of these approaches. In contrast to systemic measurements, non-invasive imaging should be able to detect the expression of alarmins even at the local site of inammation. Using uorescence reectance imaging (FRI), we now provide the rst evidence that molecular imaging allows for the reliable detection of S100A8 and S100A9 in preclinical models, locally expressed during disease, and that visualization of these proteins in conjunction with further laboratory analysis enables the monitoring of local inammation with unique sensitivity, even allowing for the detection of sub-clinical, residual disease activity. In autoimmune arthritis, we can simultaneously monitor multiple disease foci by in vivo S100A9 imaging and the extent of disease could be determined with high precision and even prognostic value for disease development in independent foci of the same animal. Moreover, by imaging S100A9 expression, we have demonstrated the rst biomarker detecting subclinical differences in phagocyte activation linked to disease outcome in a model of Th1/Th2-dichotomy in response to leishmania infection. We provide evidence that S100A9 may serve as a novel potent biomarker for monitoring local inammatory processes by molecular imaging. The broad applicability and consecutive potential impact on clinical practice is illustrated by assessing the performance of our integrated approach in exemplary models of innate and adaptive immunity, autoimmunity and infection, thus covering representative relevant pathomechanisms of inamma-tory disorders.

ResultsCy5.5-coupled aS100A9 accumulates at sites of inammation. To evaluate the feasibility of monitoring S100A9 expression by optical imaging, we employed irritant contact dermatitis (ICD) as an inammatory model, exclusively driven by innate immunity, independent of the adaptive immune system. ICD was induced in Balb/c mice by the local application of croton oil towards the ear skin13. 18F-FDG-PET, as a clinically established imaging technique to assess inammatory activity14, was used to visualize elevated glucose metabolism in areas of inammation during ICD. Although inammation can be clearly depicted by the maximum intensity PET image (Fig. 1a, red arrow), as well as in the co-registered PET/ computed tomography (CT; Fig. 1b, axial CT slice), the resulting contrast-to-noise ratio (CNR) is low due to 18F-FDG being taken up by all glucose-consuming cells (Fig. 1a).

Figure 1 | In vivo uorescence reectance imaging of mice during ICD. (a) 18F-FDG-PET image and the fused PET/CT (axial CT slice) (b) of ICD in mice treated with croton oil on the right ear. The area of inammation (red arrows) can be depicted in the maximum intensity PET image (a, whole body) with the axial CT slice showing the swelling of the inamed tissue and the local uptake of FDG (ratio inamed ear versus healthy ear 4.8; b). (c) After the

application of a-S100A9-Cy5.5 to Balb/c mice 24 h after elicitation of ICD, optical imaging (OI) was performed at the time points indicated. Strong uorescence intensities were detected only at sites of inammation for up to 96 h. (d) Quantication of CNR shows signicant changes in the affected ears over the observed time period from 24 to 96 h (baseline time point 0). Data are from three independent experiments (each n 5, means.d., *Po0.05,

**Po0.01, ***Po0.001; P values calculated using Students t-test). (e) S100A8/S100A9 serum concentrations 48 h after croton oil application. Data are from ve mice per group (means.d., ***Po0.001; Students t-test). (f) Cryosections of treated and control ears were stained for S100A9-expression. The gure shows representative ear sections of an untreated control ear (left) and treated ears with moderately (middle) and strongly (right) elevated

SNR including the corresponding systemic S100A8/S100A9 level. Scale bar, 100 mm. (g) Application of a-S100A9-Cy5.5 or rabIgG-Cy5.5 to WT or S100A9 / mice 24 h after the elicitation of ICD conrmed the specicity of optical imaging for S100A9 expression in vivo. Data are from ve mice per group (means.d., *Po0.05; P-values calculated using one-way analysis of variance with Bonferronis post test). a.u., arbitrary units.

2 NATURE COMMUNICATIONS | 5:4593 | DOI: 10.1038/ncomms5593 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

& 2014 Macmillan Publishers Limited. All rights reserved.

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5593 ARTICLE

Cy5.5-labelled antibodies against S100A9 (a-S100A9-Cy5.5, 2 nmol of dye/mouse, spectral characteristics: hexc/em 680/

700 nm, band pass 20 nm) were administered intravenously

24 h after croton oil application. Optical imaging was performed

repeatedly 2496 h after tracer application. Local tracer accumulation, as depicted by FRI, showed the highest level at 24 h (Fig. 1c,d). Fluorescence signals reecting the high local release of S100A8/S100A9 correlated well with elevated systemic levels of

Croton

Oil Antibody

OI OI OI OI

24

0

24 72

48 96 h

Baseline

0 1,200

300 600 900

Total photon counts

140 *****

* *

Time after injection (h)

***

120

600

CNR (a.u.)

S100A8/A9 (ng ml1 )

100

80

400

60

40

200

20

0

0

0

24

48

72

96

Control ICD

Local S100A9 SNR

Systemic S100A8/A9

Control ear I Treated I Treated II

12.9 18.5 24.4 74782

AU ng ml1

23242

140 *

*

120

100

CNR (a.u.)

80

60

40

20

0

rabIgG

a-S100A9 a-S100A9 S100A9/

Wild type mice

NATURE COMMUNICATIONS | 5:4593 | DOI: 10.1038/ncomms5593 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 3

& 2014 Macmillan Publishers Limited. All rights reserved.

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5593

S100A8/S100A9 complex in the serum of mice 48 h after croton oil application, as quantied by ELISA (Fig. 1e, Supplementary Fig. 1). Local S100A9 expression could be conrmed using immunohistochemistry (Fig. 1f). In healthy animals, a-S100A9-Cy5.5 showed a biodistribution that is typical for macromolecular substances (Supplementary Fig. 2). To discriminate nonspecic tracer distribution, rabbit IgG of irrelevant specicity was labelled with Cy5.5 (rabIgG-Cy5.5) and injected into a control group of mice. S100A9-decient mice (S100A9 / ) served as additional controls for the specicity of tracer to target binding. Both sets of controls showed only a slight increase in uorescence at local sites of inammation, most likely reecting hyperemia and Fcg

receptor expression/binding in areas of inammation (Fig. 1g). The distribution of rabIgG-Cy5.5 and a-S100A9-Cy5.5 in either healthy wild-type (WT) animals or non-involved organs was virtually identical (Supplementary Fig. 2).

Monitoring local inammation in allergic contact dermatitis (ACD). To assess the capability of S100A9 imaging to reect phagocyte activity in T-cell-dependent adaptive immune reactions, we analysed the expression of S100A9 in a model of ACD13. Two days after the allergen challenge of sensitized mice, a-S100A9-Cy5.5 or rabIgG-Cy5.5 was administered and uorescence intensities were monitored for up to 48 h after tracer application. Disease progression was assessed by monitoring the ear swelling during ACD (Fig. 2a). Already after 3 h, the specic probe accumulated in the area of inammation, peaking at 24 h (Fig. 2b). FRI data (Fig. 2b,c) were again in good accordance with increased S100A8/S100A9 serum levels of 258116 ng ml 1 at day 3 (Po0.01) and of 28991 ng ml 1 at day 4 (Po0.001) compared with control mice (12942 ng ml 1, ve mice per group (two independent experiments, means.d)) and clinical symptoms (ear swelling;

Fig. 2a). Similar uorescence intensities of specic and nonspecic antibodies were obtained in unaffected ears, representing a perfusion background signal (Fig. 2c). To further differentiate the accumulation of labelled specic versus unspecic antibody in the region of inammation, 2 nmol of both Cy5.5-labelled anti-S100A9 and Cy7-labelled rabIgG were injected simultaneously in vivo during ACD. Fluorescence signals obtained after the separate excitation of both probes were acquired and compared with data derived in vitro under dened conditions. The Cy5.5/Cy7 (reecting a-S100A9/rabIgG) ratio in vitro was constant (1.830.15); however, we observed a signicant increase to 4.731.31 in vivo (Fig. 2d). To rule out the effects of the dye properties on probe distribution, a-S100A9-Cy5.5 and a-S100A9-Cy7 were used simultaneously. An identical biodistribution of the differently labelled probes could be observed. Although a direct comparison of absolute signal intensities was obviated by different emission wavelengths and different quantum yields, the signal ratios between the affected and unaffected ear were virtually identical (Fig. 2e). A parallel injection of a-S100A9-Cy5.5 and rabIgG-Cy5.5 did not result in a further increase in the local signal compared with the injection of a-S100A9-Cy5.5 alone. As the S100A9-specic probe is based on a polyclonal antibody and thus presumably contains only a relatively small fraction of antigen-specic antibodies, these data conrm the specicity of our ndings (Supplementary Fig. 3a). Immunohistochemistry showed a signicant inltration by CD11b myeloid cells, Gr-1 granulocytes and F4/80 macrophages as a source of locally released S100A8/S100A9 (Supplementary Fig. 3b).

Monitoring local activity of inammation in arthritis. S100A8 and S100A9 are highly expressed in synovial tissue in rheumatoid

arthritis and serum concentrations have been shown to be reliable biomarkers reecting local disease activity10,11,15. We performed optical imaging of S100A8 and S100A9 expression in vivo in a murine collagen-induced arthritis (CIA) model and correlated imaging data with the clinical scores of disease activity. CIA was induced by immunization of DBA/jdba1/j mice with type II collagen. Clinical manifestations of arthritis started around day 7 after the last collagen injection. Symptoms prevailed for up to 3 weeks to different degrees in the foot joints. Mice received labelled antibodies 1 week after the last collagen injection (day 28). Optical imaging was performed starting 24 h after dye application.

Owing to the highly variable inammatory response of individual joints in CIA, we performed a clinical scoring for each individual foot on the basis of a three-point scale (CS0, CS1 and CS2) accounting for redness, swelling and deformation. S100A9 expression, as depicted by optical imaging, showed excellent correlation with clinical scoring, clearly discriminating clinically mild from severe joint inammation (Fig. 3a) with a high signal-to-noise ratio (SNR) for severely inamed joints (Fig. 3b). Even single affected small joints could be clearly identied (Fig. 3c). The cumulative disease activity score of all four feet (range 08) correlated well with optical imaging data presented as mean values for all four feet and systemic S100A8/ S100A9 levels of 1,180360 ng ml 1 in mice with mild arthritis (CS2-3) versus 17030 ng ml 1 in healthy control mice and 2,700380 ng ml 1 in mice with high disease activity (CS4-6;

Fig. 3d,e). To assess unspecic tracer distribution, we employed non-targeted Cy5.5-labelled rabIgG and observed only faint tracer signals, even in strongly inamed joints (Fig. 3a). The simultaneous injection of Cy7-labelled antibodies against the S100A8 subunit of the S100A8/S100A9 heterodimer (a-S100A8-Cy7, 2 nmol dye/mouse, spectral characteristics: hexc/em 755/

780 nm, band pass 20 nm) and a-S100A9-Cy5.5 showed an

almost identical in vivo distribution with equally specic accumulation in target areas (Fig. 3f,g). SNR of a-S100A8-Cy7 and a-S100A9-Cy5.5 showed excellent correlation with disease activity scores of individual feet (Fig. 3h). To further assess a potential prognostic value of the presented approach, we conducted the CIA model in C57BL/6 mice. Development of arthritis and outcome in this mouse strain is highly variable and unpredictable. Even during early disease, when rst clinical signs of CIA could only just be detected, scans allowed for the safe delineation of areas of S100A9 expression. All four paws were scored daily and imaging was repeated when manifesting clinical signs of inammation were present. We also demonstrated an excellent correlation between early and late imaging (Fig. 3i,j), as well as between early imaging and the clinical development of individual paws (Fig. 3k).

Phagocyte activity during Leishmania major infection. Experimental leishmaniasis is a well-established model for analysing the Th1/Th2-dependent immune response to infection. C57BL/6 mice show early activation of phagocytes within the rst days after Leishmania major (L. major) inoculation, which results in an effective Th1-response and the clearance of parasites. In contrast, Balb/c mice lack the sufcient activation of phagocytes in the initial phase of infection, nally resulting in the development of a Th2-dominated, ineffective immune response and subsequent parasite dissemination. The consecutive granuloma formation is accompanied by an overwhelming yet inefcient phagocyte activation in the late phase of disease16.

At day 28 after infection, in the late phase of disease, optical imaging of infected animals was performed, showing signicantly higher uorescence signals in the feet of infected Balb/c mice as

4 NATURE COMMUNICATIONS | 5:4593 | DOI: 10.1038/ncomms5593 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

& 2014 Macmillan Publishers Limited. All rights reserved.

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5593 ARTICLE

**

125

a-S100A9

rabIgG

Affected

Unaffected

**

Tracer application

0.12

100

**

Ear swelling (mm)

CNR (a.u.)

0.09

75

**

0.06

50

0.03

25

0.00

0

3 6 24 48 Time (h)

0 1 2 3 4

Days after challenge

a-S100A9-Cy5.5

rabIgG-Cy5.5

3,000

3,000

Fluorescence intensity (a.u.)

Fluorescence intensity (a.u.)

2,500

2,500

2,000

2,000

1,500

1,500

1,000

1,000

500

500

0

0

3 6 24 48

Time (h)

3 6 24 48 Time (h)

Ratio

a-S100A9-Cy5.5/rabIgG-Cy7

a-S100A9-Cy5.5 rabIgG-Cy7

Total photon counts

***

6

a-S100A9-Cy5.5

rabIgG-Cy7

4

2

1

0.75

0.5

0.25

nmol

2

0

300 600 900 1,200 1,500

In vivo CNR (a.u.)

a-S100A9-Cy5.5

a-S100A9-Cy7

In vitro Fl-int (a.u.)

5 NS

4,000

800

Fluorescence intensity (a.u.)

2,000

Affected Unaffected

4

ACD/Control

1,500

3

1,000

1,500

2

2,000

500

1

0

0

a-S100A9-Cy5.5 a-S100A9-Cy7

a-S100A9-Cy5.5 a-S100A9-Cy7

Figure 2 | Monitoring of S100A9 expression during ACD by uorescence reectance imaging (FRI) in vivo. (a) ACD was induced in mice and disease progression was assessed by the increase in ear swelling. (b) FRI was performed at different time points after the application of either a-S100A9-Cy5.5 or rabIgG-Cy5.5, as indicated in the gure. The optimal time point for optical imaging was found to be 24 h after tracer application. (c) Fluorescence intensities of the specic (left image) versus unspecic tracer (right image) of affected (black bars) versus unaffected (white bars) ears allowed for the estimation of Fcg receptor contribution to total FRI signals. (d) Comparison of the ratios of equal amounts of a-S100A9-Cy5.5 and rabIgG-Cy7 in vitro (FL-int uorescence intensities, left side) versus in vivo (CNR, right side). ACD was induced in mice and 2 nmol of Cy5.5-labelled

anti-S100A9 and Cy7-labelled rabIgG were injected simultaneously. Optical imaging was performed 24 h after antibody injection and the region of interest (ROI) of data acquisition was labelled in cyan. Data are from ve mice per group (two independent experiments each, means.d.) *Po0.05, **Po0.01, ***Po0.001; Students t-test. (e) Comparison of the ratios of equal amounts of a-S100A9-Cy5.5 and a-S100A9-Cy7 during ACD. ACD was induced in mice and 2 nmol of Cy5.5- and Cy7-labelled anti-S100A9 each was injected simultaneously. Optical imaging was performed at 24 h after antibody injection and the ROI of data acquisition was labelled in cyan (data are from ve mice, means.d. according to Students t-test). a.u., arbitrary units; NS, not signicant.

compared with C57BL/6 mice and untreated control feet (Fig. 4a, right images). RabIgG-Cy5.5 conrmed the specicity of S100A9-labelling (Fig. 4a, left images). We observed an excellent correlation of optical signals with systemic S100A8/S100A9 levels

(Fig. 4b) and footpad swelling (Fig. 4c) as a clinical sign of inammation. Moreover, in C57BL/6 mice, even at day 28 after infection, uorescence-reected macrophage activity allowed the grading of residual inammation (Fig. 4d).

NATURE COMMUNICATIONS | 5:4593 | DOI: 10.1038/ncomms5593 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 5

& 2014 Macmillan Publishers Limited. All rights reserved.

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5593

a-S100A9-Cy5.5 RabIgG-Cy5.5

CS 1 mild CS 2 severe CS 2 severe

a-S100A9 RabIgG

120 *

100

SNR (a.u.)

80

60

40

20

0

3,000

100

P = 0.003 R2 = 0.79

S100A8/A9 (ng ml1 )

75

2,000

SNR (a.u.)

50

1,000

P = 0.001 R2 = 0.86

25

0 0 2 4 6 8

0

Clinical score

0 1,200

300 600 900

Total photon counts

40

*

Control CS 0.5

CS 1.25 CS 2.0

30

*

SNR (a.u.)

20

10

0 CS 0 CS 1 CS 2

**

120

100

SNR (a.u.)

a-S100A8-Cy7

a-S100A9-Cy5.5

80

*

60

40

20

0

CS 0 CS 1 CS 2

Early S100A9 scan Late S100A9 scan

1,500

1,000

500

Late S100A9 scan (a.u.)

5.0105

4.0105

3.0105

2.0105

1.0105

1.0105 2.0105 2.5105 3.0105

P = 0.0001 R2 = 0.74

3 P = 0.0028 R2 = 0.37

Late clinical score

2

1

0

1.0105 2.0105 2.5105 3.0105

Early S100A9 scan (a.u.)

Early S100A9 scan (a.u.)

6 NATURE COMMUNICATIONS | 5:4593 | DOI: 10.1038/ncomms5593 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

& 2014 Macmillan Publishers Limited. All rights reserved.

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5593 ARTICLE

Using S100A9 imaging, even the critical, initial activation of the phagocyte system in L. major-resistant C57BL/6 mice could be detected during the very early phase at day four after infection. Measurements of S100A8/S100A9 expression in sera or local wash-outs by ELISA conrmed the detected imaging signals (Fig. 4e). In contrast, the failure of sufcient early phagocyte activation in Balb/c mice was associated with the absence of a signicant increase in S100A9 expression (Fig. 4f). We could therefore demonstrate that locally expressed S100A9 is the rst imaging marker predicting the development of a Th1 immune response, several weeks before the clinical outcome of infection.

DiscussionInammatory disorders like autoimmune diseases, allergies and acute or chronic infections are important challenges in health care. The continuously growing understanding of the biological basis of inammation stimulates the development of targeted therapies, specically addressing selected steps in the process of pathogenesis. This is accompanied by a growing demand for specic diagnostic approaches, which are capable not only of sensitive detection but also of characterization of the disease stages on a molecular level. Molecular biomarkers that meet the requirements as set out by current therapeutic developmentssensitive reection of disease activity changes under therapy, safe visualization of subclinical disease activity as a sign of therapy failure and a prediction of the outcomeare still missing. However, these requirements are a prerequisite for the adaptation of personalized therapeutic approaches providing optimal therapy.

Analysing S100A8 and S100A9, two members of the DAMP family, our study comprises a novel diagnostic approach: DAMPs or alarmins are widely accepted to represent initial tissue signals in response to cell stress and tissue damage, thus representing ideal candidates for the early and sensitive detection of developing inammation3,5,17. Both proteins are among the most upregulated genes in numerous inammatory diseases7,15. S100A8/S100A9 complexes are secreted during the activation of immigrating phagocytes or released by necrotic cells because of tissue damage at local sites of inammation, acting as endogenous triggers of Toll-like receptor-4 and inducing the expression of pro-inammatory signalling molecules in phagocytes, lymphocytes, endothelial and epithelial cells and osteoclasts8,10,18,19. In this context, S100A8 and S100A9 show typical characteristics of alarmins or DAMPs. Targeting S100A8/S100A9, we therefore monitored an inammatory mechanism that has been

shown as highly relevant in different mouse models of inammation like allergies, autoimmune diseases, arthritis and infection and has also proven to be a valuable marker for discrimination and grading of clinically relevant acute and chronic inammation710,13,18,19.

Targeted imaging of cytokines, for example, interleukin-1b or tumour necrosis factor-a revealed only limited sensitivity and/or specicity and a relatively low SNR2022, probably due to a low gradient of the cytokine concentration between local inammatory exudates and systemic serum. In contrast to most cytokines, chemokines and other pro-inammatory molecules, the local accumulation of S100A8 and S100A9 is very high (up to 100 mg ml 1 and about 50- to 100-fold higher than systemic concentrations), allowing the sensitive and specic detection of local changes in the expression of these molecules15,23. S100A8 and S100A9 comprise about 40% and 5% of the cytosolic proteins in granulocytes and monocytes, respectively. They are specically released during the interaction of phagocytes with inammatory activated endothelial cells and can subsequently bind to heparan sulfate of the endothelial glycocalyx15,2426. The high abundance and matrix binding appear to be responsible for the high local accumulation of these proteins at sites of inammation compared with other inammatory molecules, which allows the detection of these molecules even by MRI27 and accounts for the high target to background ratios that we observed in our in vivo imaging studies.

In contact dermatitis, as a model of local inammation, we could demonstrate that the expression of S100A9 is a very sensitive marker for inammatory processes and reects disease activity independently of the underlying pathomechanism, for example, in toxic or in allergen-induced inammation. In contrast to most imaging studies, we conrmed that our imaging data closely correlated not only to the local expression of S100 proteins by phagocytes in the dermal inltrate, as determined by immunohistochemistry, but also to systemic levels of both proteins and clinical parameters. The specicity of our data was conrmed by the use of nonspecic antibodies (rabbit-derived IgG without relevant specicity) to determine perfusion and unspecic Fcg-receptor binding. Parallel injection of the S100A9-specic probe and rabIgG labelled with different dyes ruled out in vivo competition for the specic target or dye-dependent effects. Moreover, we applied the specic probe a-S100A9-Cy5.5 in ICD in WT and S100A9 / mice. SNR of a-S100A9-Cy5.5 in

WT mice were signicantly higher compared with nonspecic antibodies, as well as compared with data obtained from S100A9 / mice. In this model, WT and S100A9 / mice showed no differences regarding their inammatory phenotype13.

Figure 3 | Detection of single inamed joints in collagen-induced arthritis by optical imaging. (a) CIA was induced in DBA/jdba1/j mice and optical imaging was recorded by FRI using a-S100A9-Cy5.5 or rabIgG-Cy5.5 at day 28. Individual feet were analysed and compared with clinical scores described in the Method section (n 5 mice per group, three independent experiments). (b) Comparison of imaging data of a-S100A9-Cy5.5 and rabIgG-Cy5.5 from

mice feet with CS2 conrms the specicity of our ndings (means.d., *Po0.05, Students t-test, n 5 mice per group, three independent experiments).

(c) Enlarged view of only subclinical inammation of single joints. (d) Correlation of imaging data and systemic S100A8/S100A9 levels of mice with clinical disease severity. Scoring of single feet were added up (maximum score 8 per mouse) and imaging data were calculated as mean values over all four feet. (e) S100A9 immunostaining of paw sections with CS of 0.5, 1.25 and 2 and control at day 28 conrmed correlation of local S100A8/S100A9 expression and severity of inammation. Scale bar, 100 mm. (f,g) Simultaneous injection of a-S100A8-Cy7 (f) and a-S100A9-Cy5.5 (g) shows an almost identical distribution pattern in FRI. (h) Upper panel, representative pictures of paws for the different clinical scores. Lower panels, correlation of clinical scores with

SNR for Cy7-labelled anti-S100A8 and Cy5.5-labelled anti-S100A9 (n 5 mice, means.d., *Po0.05, **Po0.01; one-way analysis of variance with

Bonferronis post test). (i) CIA was induced in C57BL/6 mice and a-S100A9-Cy5.5-driven FRI was performed at the rst signs of arthritis (early time point) and for a second time where disease progression had occurred (late time point, 2 nmol of dye per mouse). All paws were analysed separately and compared with the clinical scoring as described in (a, n 6 mice). Representative images of an early (left image) versus late (right image) S100A9 scan are

shown. White arrows indicate inamed areas. Correlation of optical imaging data of early versus late time point (j) and early time point versus late clinical score (k). To correct for the variable areas of fore- and hind-paws, uorescence intensities were normalized for the ROI size. Two fore-paws of one mouse were excluded from the analysis because of an incorrect position in the scanner (j,k). a.u., arbitrary units.

NATURE COMMUNICATIONS | 5:4593 | DOI: 10.1038/ncomms5593 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 7

& 2014 Macmillan Publishers Limited. All rights reserved.

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5593

Analysing CIA, we conrmed an excellent correlation of S100A9 expression, as detected by optical imaging with local disease activity in individual joints. In addition, we demonstrated that the local expression of S100A9 and S100A8 is equally capable

of uncovering sub-clinical disease activity, which is not reected by established clinical parameters. Using blocking antibodies in arthritis models in WT and S100A9 / mice revealed that this alarmin has a pivotal role in the inammatory as well as in the

rabIgG-Cy5.5 a-S100A9-Cy5.5

Infected Healthy

2,300

1,475

Balb/c

C57BL/6

2,025

1,750

1,200

* ***

Footpad swelling (mm)

0

100

1,800

5 Balb/cC57BL/6 *** ***

2,600

1,500

400

4

80

1,500

1,200

3

S100A8/A9 (ng ml1 )

60

900

2

40

CNR (a.u.)

20

600

1

300

0 28 35

Balb/c

C57BL/6

Days after L.m. infection

Systemic S100A8/A9 110 250 480

a.u.

ng ml1

CNR

1.3 30.6 56.9

NS

***

*

800

S100A8/A9 (ng ml1 )

600

**

400

C57BL/6, P=0.047 Balb/c, P=0.19

rabIgG

NS

200

*

50 *

50

40

40

CNR (a.u.)

CNR (a.u.)

0

C57BL/6

Balb/c

30

30

+ +

+ +

L.M. inf.

+

20

20

Serum

+

+

+

+

10

Washout

+

+

+

10

0

0

a-S100A9

a-S100A9

rabIgG

8 NATURE COMMUNICATIONS | 5:4593 | DOI: 10.1038/ncomms5593 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

& 2014 Macmillan Publishers Limited. All rights reserved.

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5593 ARTICLE

destructive process during arthritis10,11,28, indicating that our imaging approach directly reects major pathogenic aspects of arthritis.

In CIA, E-selectin-targeted optical imaging has been demonstrated to reect inammatory activity with similar SNR to our data29. Interestingly, S100A9 has been shown to induce the expression of E-selectin in endothelial cells30, indicating that both methods address different target cells within closely linked inammatory mechanisms. However, the high local level of S100A9 at sites of numerous inammatory diseases makes this molecule more suitable for future imaging approaches in clinical practice. The detection of leukocyte populations with radio-labelled antibodies against macrophages (F4/80), T lymphocytes (CD4, CD3, CD40) or B lymphocytes (CD20), among others, and cell tracking studies with labelled cells have been used to monitor inammatory processes in arthritis31,32. However, all of these approaches suffer from the disadvantage that the inltration of signicant numbers of these leukocytes takes a substantial amount of time and is not the initial event in the inammation cascade. Moreover, leukocytes may persist during the resolution of inammation, limiting the diagnostic value of these approaches for monitoring relapsing remitting courses of inammation. Therefore, most of these studies have identied only low SNR and sometimes entirely failed to show a correlation of imaging results with disease activity31,32. The tissue destruction associated with phagocyte inltration is reected by protease activity and can be detected by small compounds targeting myeloperoxidase, cathepsin or matrix metalloproteinase activity3335. However, these markers, in contrast to S100A8/S100A9, are not sensitive for the early inammatory reaction, but reect later stages of disease only.

Yet another interesting nding in our study was that the expression of S100A8/S100A9 during the initial phase of an inammatory (infectious) process was the rst early and sensitive marker for subclinical phagocyte activation known to be linked to the development of an effective Th1-response in this model. During the very early phase of L. major infection (day 4), we were able to visualize the effective activation of phagocytes in resistant C57BL/6 mice. In susceptible Balb/c mice, the lack of adequate, early phagocyte activation and differentiation into a pro-inammatory M1-phenotype was reected by low S100 expression in our imaging approach associated with priming a Th2 response, not sufcient to ght the infection. Hence, the imaging signal allowed for the prediction of a fatal outcome of disease, even weeks before full manifestation of the clinical phenotype and disease dissemination in this particular disease model.

Our data demonstrate for the rst time that targeting members of the alarmin or DAMP family is a potent strategy for functional

molecular imaging of inammatory processes in general and independent of the underlying pathophysiology. Similar results of local expression of an alarmin family member could be demonstrated using a transgene reporter mice model for HSP-70 expression in ischemic brain injury. However, such transgenic approaches unfortunately only have limited potential for translation into clinical applications36. Owing to their specic mode of expression and release, S100A8 and S100A9 are sensitive biomarkers for the immediate response of innate immune mechanisms to disturbances of tissue homeostasis. The induction and release of S100A8/S100A9 have been shown to correlate very well with disease activity in many clinically relevant disorders, including rheumatoid arthritis, inammatory bowel disease, autoimmune diseases, infections, allograft rejection or chronic processes like atherosclerosis, which underlines the translational potential and high impact of our ndings for future basic research as well as clinical applications3,7,12. Optical imaging driven by uorescently labelled antibodies has a growing translational potential beyond its impact on preclinical research and may be applicable for the examination of supercial lesions such as cutaneous inammation. In addition, uorescence endoscopy is gaining increasing interest and could foster the clinical use of uorescent contrast agents37; for example, in the evaluation of inammatory bowel disease, which is known to exhibit a very high local expression of S100A8/S100A9 (calprotectin)38. The potential of optical imaging of S100A8/ S100A9 for regular clinical use is moreover underlined by the introduction of an optical scanner for visualization of the disease activity in rheumatoid arthritis, another inammatory disorder with high local S100A8/S100A9 expression, currently driven by ICG uorescence39. Finally, S100A8/S100A9 are highly upregulated during tumour development and even in the developing metastatic niche40. Therefore, another specic demand for optical molecular imaging of S100 proteins may arise in the context of intraoperative imaging for the safe delineation of malignant tissue during tumour resection41.

For the translation of target-specic imaging approach into deep tissue imaging, other label strategies may have to be explored. Replacing the uorescent dye with a radionuclide for either single photon emission computed tomography (SPECT) or PET imaging would allow for the examination of virtually all body compartments in humans, with specicity and sensitivity resembling that of optical imaging. With hybrid systems like PET/CT and PET/MRI increasingly applied in clinical imaging, targeted imaging of S100A8/A9 could be combined with high-resolution morphology. Targeted contrast agents for MRI, based on either superparamagnetic iron oxides or Gadolinium, suffer from the low sensitivity of MRI at clinical eld strength andin comparison to optical or radionuclide-driven imaginglarge

Figure 4 | Mouse strain-specic responses during L. major infection monitored by S100A9 imaging in vivo. (a) Right hind legs of C57BL/6 mice and Balb/c mice (three independent experiments, each ve mice per group) were infected with L. major, whereas the left hind legs served as controls. FRI was monitored during the late phase of infection at day 28 after receiving either a-S100A9-Cy5.5 or rabIgG-Cy5.5 (2 nmol of dye per mouse) 24 h earlier. (b) CNR was calculated for both mouse strains at day 28 after L. major infection. Signicant strain-specic differences were found for both local (CNR) and systemic (S100A8/S100A9) parameters (three independent experiments, each ve mice per group, means.d., *Po0.05, ***Po0.001; MannWhitney

U-test). (c) Footpad swelling of infected C57BL/6 mice and Balb/c mice at day 28 and 35 in relation to non-infected contralateral foot pads (means.d., n 5 for each mouse strain, ***Po0.001; t-test) demonstrates the different outcome in both mouse strains. (d) Individual comparison of representative

optical imaging data (CNR) of infected C57BL/6 mice at day 28 after infection shows fairly good accordance, suggesting that systemic S100A8/S100A9 levels resemble disease activities. (e) During early L. major infection (day 4), sera (grey bars) and footpad washouts (black bars) of infected and non-infected mice were collected and analysed for S100A8/S100A9 by ELISA. Systemic and local S100A8/S100A9 levels were already signicantly increased in infected C57BL/6 mice as compared with controls. In Balb/c mice, only a minor, nonsignicant increase in local S100A8/S100A9 expression was observed. Data are from fvie mice per group (means.d., *Po0.05, **Po0.01, ***Po0.001 and NS, not signicant; MannWhitney U-test). (f) At day 4 already, local upregulation of S100A9 expression could be monitored in resistant C57BL/6 mice by optical imaging (P 0.047 by t-test) reecting

phagocyte activation. This was not detectable in susceptible Balb/c mice (P 0.19). Data are from ve mice per group. a.u., arbitrary units.

NATURE COMMUNICATIONS | 5:4593 | DOI: 10.1038/ncomms5593 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 9

& 2014 Macmillan Publishers Limited. All rights reserved.

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5593

amounts of the agent are required to incur a measurable change in relaxivity42. Although the visualization of tracer accumulation in areas of inammation could be demonstrated for Gadolinium-loaded nanoparticles in selected experimental models27, a convincing approach with translational potential has yet to be developed. Moreover, the combinatory application of contrast enhanced MRI and targeted imaging in hybrid systems to acquire even more diagnostic information in a single examination would be negated by the use of targeted MRI probes. As a paradigm for such future developments in molecular imaging, we provide the rst fully integrated diagnostic approach on a member of the DAMP/alarmin family in various preclinical models of different modes of inammation with clear potential for translation into clinical practice.

Materials and methods

Mice and reagents. C57BL/6 mice, Balb/c mice (Harlan Laboratories), DBA/ jdba1/j mice (Janvier-Elevage) and S100A9-decient mice (S100A9 / , back-crossed to C57BL/6 or Balb/c background (F10 generation))43 were used at the age of 812 weeks, sex matched for each set of experiments and housed under specic pathogen-free conditions. All experiments with mice were performed with the approval of the State Review Board of Nordrhein-Westfalen (Germany) according to the German law for animal welfare (Permit Number: 84-02.04.2012.A058) or by the Ethics Committee of University Hospital Nijmegen (Permit Number: DEC 2014-044). All reagents were purchased from Sigma at the highest purity grade available, unless indicated otherwise.

ELISA. We used an in-house ELISA to determine the concentrations of S100A8/ S100A9 in sera and washouts of footpads, as described earlier (Supplementary Fig. 1)10. We calibrated our ELISA against puried S100A8/S100A9 heterodimer as complexes have been shown to be the predominant form of these proteins.

Antibodies and antibody labelling. Rabbit-derived antibodies addressing S100A9 or S100A8 were puried via protein G-sepharose and labelled with the uorochromes Cy5.5 or Cy7 according to the manufacturers instructions (GE Healthcare). Cy5.5-labelled rabbit IgG without relevant specicity in mice served as a control. Briey, 5 mg of the antibody was dialysed towards 100 mM Na2CO3 buffer, pH 8.0 and a 20-fold excess of the uorochrome was added for 90 min at RT. The resulting tracer was puried from unbound dye using size exclusion chromatography (PD10 column). The labelling efcacy (dye/antibody ratio) was determined on the basis of ultraviolet-spectra of the puried dyeantibody compound using PBS as a reference buffer. Typically, the labelling resulted in 2.53.0 uorochrome molecules per antibody, irrespective of the precursors.

In vivo imaging. Mice were held under isourane inhalation anaesthesia forthe duration of the scan and the imaging chamber was heated to 30. As the total scan time was usually under 1 min, physiological effects due to a signicant decrease of body temperature were not to be expected. Mice were intravenously injected with either the specic Cy5.5-labelled S100A9 antibody (a-S100A9-Cy5.5, 2 nmol of Cy5.5 B100 mg antibody in total) or Cy5.5-labelled antibody of irrelevant specicity (rabIgG-Cy5.5), unless specied otherwise. In selected experiments, a-S100A9-Cy5.5- and Cy7-labelled S100A8 antibodies (a-S100A8-Cy7) or a-S100A9-Cy5.5- and Cy7-labelled S100A9 antibodies (a-S100A9-Cy7) or a-S100A9-Cy5.5 and rabIgG-Cy5.5 antibodies were administered in parallelin the same animal for simultaneous detection.

FRI. FRI was performed using the Carestream FX Pro Imaging Station (Carestream Health). For imaging of Cy5.5-labelled antibodies, excitation light was set to630 nm using an appropriate bandpass ltre. Emission at 700 nm was recorded using a ltre-equipped high-sensitivity (4-million-pixel) cooled charge-coupled device camera. Acquisition time was 30 s for each image, followed by a photography-style white light image or conventional X-ray for image fusion and co-registration of anatomical information.

For the imaging of Cy7-labelled compounds, excitation and emission wavelengths were 730 and 790 nm, respectively; acquisition time was 30 s.

For the time of examination, mice were held under isourane inhalation anaesthesia (2.5% isourane in air).

For each region of interest measured for imaging analysis, the mean uorescence intensity (SI) and resulting standard deviation (s.d.) were determined.

From the acquired uorescence signals of the target region (SItarget), SNRs were

calculated as SNR SI

target/s.d.background signal.

If possible, the comparison of affected organs (target) and healthy organs (control) in the same animal CNRs were calculated as CNR (SI

target

PET and CT scanning. Animals were anaesthetized with isourane, and 10 MBq of

18F-FDG in 100 ml 0.9% saline was injected intravenously 1 h before each PET analysis. For PET acquisition, animals were placed on a heat-controlled multi-modal scanning bed and PET list mode data were acquired for 15 min using the 32-module quadHIDAC scanner (Oxford Positron Systems), dedicated to small animal imaging. The scanner has an effective resolution of 0.7 mm (full-width at half-maximum) in the transaxial and axial directions when using an iterative resolution recovery reconstruction algorithm. Subsequently, the scanning bed was transferred to the computed tomography scanner (Inveon, Siemens Medical Solutions) and a medium resolution (25 mm) CT acquisition was performed for each mouse. PET data were reconstructed into a single image volume for each mouse with a voxel size of 0.4 0.4 0.4 mm3. CT was reconstructed into a

volume data set with a voxel size of 0.007 0.007 0.007 mm3. Image data sets

were co-registered using extrinsic markers attached to the multimodal scanning bed and commercially available image analysis software (Inveon Research Workplace, Siemens Medical Solutions).

Immunohistochemistry. Immunohistochemistry of ear sections (cryo) or paw sections (parafn) was performed as described earlier using puried rabbit antisera against murine S100A9 (refs 10,13). Briey, after inhibition of endogenous peroxidase activity in frozen tissue sections Fc receptors were blocked by incubating in PBS/1% BSA including 50% normal goat serum (NGS). Slides were immunostained in a two-step procedure of incubation of primary antibody or isotype control followed by a horseradish peroxidase-conjugated secondary antibody using AEC as chromogen. Images were acquired by using an upright microscope (Axioskop, Zeiss). Paws from arthritis experiments were xed in 4% formaldehyde and joints were decalcied with 5% formic acid in PBS during 7 days. After dehydration and embedment in parafn, sections of the paws were cut (7 mM) in a standardized manner and processed for S100A9 staining. Sections were treated with 1% H2O2 to inhibit endogenous peroxidase and 0.1% Triton/PBS for antigen retrieval and additionally incubated with rabbit anti-S100A9 antibodies followed by a goat-anti-rabbit biotinylated antibody and 3,30 diaminobenzidine (DAB).

Eliciting irritant and ACD. ICD was induced by the application of 1% croton oil in olive oil-acetone (1:4) to the dorsal surface of the right ear of mice (n 5 per

group) for 24 h, whereas the left ear served as a control. FRI was performed at 24, 48, 72 and 96 h after tracer application, corresponding to 48, 72, 96 and 120 h after croton oil treatment, respectively. Subsequently, mice were killed and ears were snap-frozen in liquid nitrogen and transferred for immunohistochemistry.

To elicit ACD, C57BL/6 or Balb/c mice were sensitized by the application of 25 ml of 0.5% 2,4-dinitrouoro-1-benzene (Sigma) in olive oil/acetone (1:4) to the shaved abdominal wall on two consecutive days. Six days later, mice were challenged with 15 ml of 0.4% 2,4-dinitrouoro-1-benzene in olive oil/acetone on the dorsal surface of the right ear. At day 2, mice were injected with either a-S100A9-Cy5.5 or rabIgG-Cy5.5 or a combination of a-S100A9-Cy5.5/rabIgGCy7, a-S100A9-Cy5.5/a-S100A9-Cy7 or a-S100A9-Cy5.5/rabIgG-Cy5.5. FRI was performed at different time points after antibody application, as indicated in Fig. 2. The disease severity of ACD was monitored by measuring the ear swelling.

Induction and imaging of CIA. Arthritis was induced in DBA/jdba1/j or C57BL/6 mice by immunization using bovine collagen type II as described in detail earlier44. Briey, bovine collagen type II (bCII, MD Biosciences) was dissolved in 0.05 M acetic acid at a concentration of 2 mg ml 1. Mice were injected subcutaneously at the tail base with 100 mg bCII emulsied in Complete Freunds Adjuvant (Difco). The animals were boosted at day 21 with an intraperitoneal injection of 100 mg bCII. The onset of polyarthritis occurred around 45 days later. Mice were regularly inspected from day 14 after disease induction and scored for swelling, erythema and deformation of each joint three times a week (CS0 no swelling,

CS1 slight swelling and erythema, CS2 pronounced oedema including joint

rigidity). Scoring of single joints was added up to a maximum possible score of eight per mouse (two per paw). Regarding the susceptibility of C57BL/6 mice, arthritis was evaluated using an extended scoring three-point scale: 0 normal;

0.5 erythema, light oedema; 1 mild but denitely visible, erythema and oedema

of one digit/toe or limb; 2 erythema and moderate oedema of at least two digits/

toe or limb; 3 erythema, severe oedema of the entire paw and/or rigidity. Imaging

was performed after arthritis was clinically detectable in the majority of treated animals 24 h after tracer application of either a-S100A9-Cy5.5 or rabIgG-Cy5.5 or a combination of a-S100A9-Cy5.5/a-S100A8-Cy7. At the end of the experiment, mice were killed and serum was collected for S100A8/S100A9 quantication by ELISA. Hindpaws were excised and processed for histology.

Experimental leishmaniasis. Cutaneous leishmaniasis was initiated in two different mouse strains (C57BL/6 and Balb/c) by the subcutaneous application of 2 107 promastigotes (stationary phase) of L. major in 50 ml PBS into the right

hind footpad. Footpad thickness of the infected in relation to the healthy foot was assessed for clinical monitoring of disease. FRI was performed at days 4 and 28 after infection (representative for an early and a late phase immune response), 24 h after tracer injection. In separate experiments, sera and footpad washouts were

SIcontrol)/

s.d.background.

10 NATURE COMMUNICATIONS | 5:4593 | DOI: 10.1038/ncomms5593 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

& 2014 Macmillan Publishers Limited. All rights reserved.

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5593 ARTICLE

collected at days 4 or 28 after L. major inoculation for S100A8/S100A9 quantication by ELISA. Briey, L. major infected and non-infected hindpaws were washed out by s.c. injection of 250 ml PBS. Without applying any force, in order to obtain secreted protein, approximately 250 ml of the draining liquid was collected and used for further analysis.

Statistical analysis. Results are presented throughout as mean valuesstandard deviation (s.d.). P-values are given in the gure legends and values of P40.05 were considered not to be signicant. Statistical analyses were performed by parametric tests (t-test or one-way analysis of variance) and the MannWhitney U-test.

References

1. Chan, A. C. & Behrens, T. W. Personalizing medicine for autoimmune and inammatory diseases. Nat. Immunol. 14, 106109 (2013).

2. Gotthardt, M., Bleeker-Rovers, C. P., Boerman, O. C. & Oyen, W. J. Imaging of inammation by PET, conventional scintigraphy, and other imaging techniques. J. Nucl. Med. 51, 19371949 (2010).

3. Chan, J. K. et al. Alarmins: awaiting a clinical response. J. Clin. Invest. 122, 27112719 (2012).

4. Harris, H. E., Andersson, U. & Pisetsky, D. S. HMGB1: a multifunctional alarmin driving autoimmune and inammatory disease. Nat. Rev. Rheumatol 8, 195202 (2012).

5. Rock, K. L., Latz, E., Ontiveros, F. & Kono, H. The sterile inammatory response. Annu. Rev. Immunol. 28, 321342 (2010).

6. Yang, D., de la Rosa, G., Tewary, P. & Oppenheim, J. J. Alarmins link neutrophils and dendritic cells. Trends Immunol. 30, 531537 (2009).

7. Ehrchen, J. M., Sunderkotter, C., Foell, D., Vogl, T. & Roth, J. The endogenous Toll-like receptor 4 agonist S100A8/S100A9 (calprotectin) as innate amplier of infection, autoimmunity, and cancer. J. Leukoc. Biol. 86, 557566 (2009).

8. Vogl, T. et al. Mrp8 and Mrp14 are endogenous activators of Toll-like receptor 4, promoting lethal, endotoxin-induced shock. Nat. Med. 13, 10421049 (2007).

9. Loser, K. et al. The Toll-like receptor 4 ligands Mrp8 and Mrp14 are crucial in the development of autoreactive CD8 T cells. Nat. Med. 16, 713717 (2010).

10. van Lent, P. L. et al. Myeloid-related proteins S100A8/S100A9 regulate joint inammation and cartilage destruction during antigen-induced arthritis. Ann. Rheum. Dis. 67, 17501758 (2008).

11. van Lent, P. L. et al. Active involvement of alarmins S100A8 and S100A9 in the regulation of synovial activation and joint destruction during mouse and human osteoarthritis. Arthritis Rheum. 64, 14661476 (2012).

12. Foell, D. et al. Methotrexate withdrawal at 6 vs. 12 months in juvenile idiopathic arthritis in remission: a randomized clinical trial. JAMA 303, 12661273 (2010).

13. Petersen, B. et al. The alarmin Mrp8/14 as regulator of the adaptive immune response during allergic contact dermatitis. EMBO J. 32, 100111 (2013).14. Haroon, A., Zumla, A. & Bomanji, J. Role of uorine 18 uorodeoxyglucose positron emission tomography-computed tomography in focal and generalized infectious and inammatory disorders. Clin. Infect. Dis. 54, 13331341 (2012).

15. Frosch, M. et al. Myeloid-related proteins 8 and 14 are specically secreted during interaction of phagocytes and activated endothelium and are useful markers for monitoring disease activity in pauciarticular-onset juvenile rheumatoid arthritis. Arthritis Rheum. 43, 628637 (2000).

16. Sacks, D. & Noben-Trauth, N. The immunology of susceptibility and resistance to Leishmania major in mice. Nat. Rev. Immunol. 2, 845858 (2002).

17. McDonald, B. et al. Intravascular danger signals guide neutrophils to sites of sterile inammation. Science 330, 362366 (2010).

18. van Zoelen, M. A. et al. Expression and role of myeloid-related protein-14 in clinical and experimental sepsis. Am. J. Respir. Crit. Care Med. 180, 10981106 (2009).

19. Huckins, D. S. et al. A novel biomarker panel to rule out acute appendicitis in pediatric patients with abdominal pain. Am. J. Emerg. Med. 31, 13681375 (2013).

20. Barrera, P. et al. Radiolabelled interleukin-1 receptor antagonist for detection of synovitis in patients with rheumatoid arthritis. Rheumatology 39, 870874 (2000).

21. Barrera, P., Oyen, W. J., Boerman, O. C. & van Riel, P. L. Scintigraphic detection of tumour necrosis factor in patients with rheumatoid arthritis. Ann. Rheum. Dis. 62, 825828 (2003).

22. Roimicher, L. et al. (99m)Tc-anti-TNF-alpha scintigraphy in RA: a comparison pilot study with MRI and clinical examination. Rheumatology (Oxford) 50, 20442050 (2011).

23. Achouiti, A. et al. Myeloid-related protein-14 contributes to protective immunity in gram-negative pneumonia derived sepsis. PLoS Pathog. 8, e1002987 (2012).

24. Rammes, A. et al. Myeloid-related protein (MRP) 8 and MRP14, calcium-binding proteins of the S100 family, are secreted by activated monocytes via a novel, tubulin-dependent pathway. J. Biol. Chem. 272, 94969502 (1997).

25. Robinson, M. J., Tessier, P., Poulsom, R. & Hogg, N. The S100 family heterodimer, MRP-8/14, binds with high afnity to heparin and heparan sulfate glycosaminoglycans on endothelial cells. J. Biol. Chem. 277, 36583665 (2002).

26. Teigelkamp, S. et al. Calcium-dependent complex assembly of the myeloic differentiation proteins MRP-8 and MRP-14. J. Biol. Chem. 266, 1346213467 (1991).

27. Maiseyeu, A. et al. In vivo targeting of inammation-associated myeloid-related protein 8/14 via gadolinium immunonanoparticles. Arterioscler. Thromb. Vasc. Biol. 32, 962970 (2012).

28. Cesaro, A. et al. An inammation loop orchestrated by S100A9 and calprotectin is critical for development of arthritis. PLoS ONE 7, 18 (2012).

29. Gompels, L. L. et al. In vivo uorescence imaging of E-selectin: quantitative detection of endothelial activation in a mouse model of arthritis. Arthritis Rheum. 63, 107117 (2011).

30. Viemann, D. et al. Myeloid-related proteins 8 and 14 induce a specic inammatory response in human microvascular endothelial cells. Blood 105, 29552962 (2005).

31. Gaal, J. et al. 99m Tc-HMPAO labelled leukocyte scintigraphy in patients with rheumatoid arthritis: a comparison with disease activity. Nucl. Med. Commun. 23, 3946 (2002).

32. Mountz, J. M., Alavi, A. & Mountz, J. D. Emerging optical and nuclear medicine imaging methods in rheumatoid arthritis. Nat. Rev. Rheumatol. 8, 719728 (2012).

33. Gross, S. et al. Bioluminescence imaging of myeloperoxidase activity in vivo. Nat. Med. 15, 455461 (2009).

34. Ntziachristos, V., Ripoll, J., Wang, L. V. & Weissleder, R. Looking and listening to light: the evolution of whole-body photonic imaging. Nat. Biotechnol. 23, 313320 (2005).

35. Peterson, J. D. et al. Optical tomographic imaging discriminates between disease-modifying anti-rheumatic drug (DMARD) and non-DMARD efcacy in collagen antibody-induced arthritis. Arthritis Res. Ther. 12, R105 (2010).

36. de la Rosa, X. et al. In vivo imaging of induction of heat-shock protein-70 gene expression with uorescence reectance imaging and intravital confocal microscopy following brain ischaemia in reporter mice. Eur. J. Nucl. Med. Mol. Imaging 40, 426438 (2013).

37. Kim, S. Y. & Myung, S. J. Optical molecular imaging for diagnosing intestinal diseases. Clin. Endosc. 46, 620626 (2013).

38. Leach, S. T. & Day, A. S. S100 proteins in the pathogenesis and diagnosis of inammatory bowel disease. Expert Rev. Clin. Immunol. 2, 471480 (2006).

39. Werner, S. G. et al. Inammation assessment in patients with arthritis using a novel in vivo uorescence optical imaging technology. Ann. Rheum. Dis. 71, 504510 (2012).

40. Hiratsuka, S. et al. The S100A8-serum amyloid A3-TLR4 paracrine cascade establishes a pre-metastatic phase. Nat. Cell Biol. 10, 13491355 (2008).

41. Orosco, R. K., Tsien, R. Y. & Nguyen, Q. T. Fluorescence imaging in surgery. IEEE Rev. Biomed. Eng. 6, 178187 (2013).

42. Huang, C. H. & Tsourkas, A. Gd-based macromolecules and nanoparticles as magnetic resonance contrast agents for molecular imaging. Curr. Top. Med. Chem. 13, 411421 (2013).

43. Manitz, M. P. et al. Loss of S100A9 (MRP14) results in reduced interleukin-8-induced CD11b surface expression, a polarized microlament system, and diminished responsiveness to chemoattractants in vitro. Mol. Cell Biol. 23, 10341043 (2003).

44. Koenders, M. I. et al. Tumor necrosis factor-interleukin-17 interplay induces S100A8, interleukin-1beta, and matrix metalloproteinases, and drives irreversible cartilage destruction in murine arthritis: rationale for combination treatment during arthritis. Arthritis Rheum. 63, 23292339 (2011).

Acknowledgements

We thank Eva Nattkemper, Heike Hater, Ina Winkler and Heike Berheide for their excellent technical support. This work was supported by Grants from the Interdisciplinary Center of Clinical Research of the University of Muenster (Vo2/014/09, PIX), the German Research Foundation (DFG) EI 878 1-1 to M.E., CRC 656 A09, CRC 1009 B8 and B9 from the European Unions Seventh Framework Program under EC-GA No. 305266 MIAMI to Th.V. and J.R., by the Federal Ministry of Education and Research (BMBF), project AID-NET to J.R. and CR-UK (C1519/A6906) and the KCLUCL Comprehensive Cancer Imaging Centre CR-UK and EPSRC, in association with the MRC and DoH to M.E.

Author contributions

Th.V. designed and supervised the study and experiments, performed animal studies and wrote the manuscript. M.E. designed the study and the experiments, performed optical imaging and wrote the manuscript. T.V., B.P., S.Z., K.R. and P.v.L. performed animal experiments and analysed the data. S.H. and A.F. performed PET and PET/CT experiments and analysed the data. C.G. and C.B. designed and supervised the optical imaging experiments and edited the manuscript. M.S. supervised the PET imaging experiments

NATURE COMMUNICATIONS | 5:4593 | DOI: 10.1038/ncomms5593 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 11

& 2014 Macmillan Publishers Limited. All rights reserved.

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms5593

How to cite this article: Vogl, T. et al. Alarmin S100A8/S100A9 as a biomarker for molecular imaging of local inammatory activity. Nat. Commun. 5:4593 doi: 10.1038/ ncomms5593 (2014).

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/by/4.0/

and edited the manuscript. J.R. designed the study and experiments and wrote the manuscript.

Additional information

Supplementary Information accompanies this paper at http://www.nature.com/naturecommunications

Web End =http://www.nature.com/ http://www.nature.com/naturecommunications

Web End =naturecommunications

Competing nancial interests: The authors declare no competing nancial interests.

Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/

Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions/

Web End =reprintsandpermissions/

12 NATURE COMMUNICATIONS | 5:4593 | DOI: 10.1038/ncomms5593 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

& 2014 Macmillan Publishers Limited. All rights reserved.