ORIGINAL ARTICLE
Suppression of T Cell Activation and Collagen Accumulation by an Anti-IFNAR1 mAb, Anifrolumab, in Adult Patients with Systemic Sclerosis Xiang Guo1, Brandon W. Higgs1, Anne C Bay-Jensen2, Morten A. Karsdal2, Yihong Yao1, Lorin K. Roskos1 and Wendy I. White1 Type I IFNs are implicated in the pathophysiology of systemic sclerosis (SSc). Recently, a Phase I open-label trial was conducted with an anti-IFNAR1 receptor antibody (anifrolumab) in adult SSc patients. In this study, we aim to assess the downstream effects of anifrolumab and elucidate the role of type I IFN in SSc. Serum proteins and extracellular matrix (ECM) markers were measured in relation to IFN pathway activation status and SSc disease activity. Our results demonstrated a robust overexpression of multiple serum proteins in SSc patients, particularly those with an elevated baseline type I IFN gene signature. Anifrolumab administration was associated with significant downregulation of T cell–associated proteins and upregulation of type III collagen degradation marker. Whole-blood and skin microarray results also indicated the inhibition of T cell receptor and ECM–related transcripts by anifrolumab. In summary, our study demonstrates suppressive effects of anifrolumab on T cell activation and collagen accumulation through which tissue fibrosis may be reduced in SSc patients. The relationship between these peripheral markers and the clinical response to anifrolumab may be examined in larger double-blind, placebo-controlled trials. Journal of Investigative Dermatology advance online publication, 18 June 2015; doi:10.1038/jid.2015.188
INTRODUCTION Systemic sclerosis (SSc) is a rare, multisystem autoimmune disease of unknown etiology, characterized by vascular alterations, immunological disturbances, and fibrosis of the skin and internal organs. T cell and macrophage infiltration has been demonstrated in skin biopsies of SSc patients preceding fibrosis (Prescott et al., 1992). Endothelial cell activation and adhesion molecule overexpression are also evident in SSc, leading to the migration of immune cells into the extravascular tissue (Sato, 1999). T cell–derived cytokines are important factors that drive the production of extracellular matrix (ECM) components by fibroblasts resulting in excessive fibrosis in SSc patients (O'Reilly et al., 2012). The interaction between T cells and fibroblasts through CD40LG-CD40 signaling is critical for fibroblast activation and proliferation 1
Translational Sciences, MedImmune LLC, Gaithersburg, Maryland, USA and Nordic Bioscience, Herlev, Denmark
2
Correspondence: Xiang Guo, Translational Sciences, MedImmune LLC, One MedImmune Way, Gaithersburg, Maryland 20878, USA. E-mail: [email protected] Abbreviations: BAFF, B cell–activating factor; BH, Benjamin–Hochberg; DLcoP, diffusing capacity for carbon monoxide; ECM, extracellular matrix; FC, fold change; HAQ-DI, Health Assessment Questionnaire-Disability Index; HC, healthy control; MMP, matrix metalloproteinase; mRSS, modified Rodnan skin score; PIIINP, amino-terminal propeptide of type III procollagen; SSc, systemic sclerosis; TGF-β, transforming growth factor-β Received 3 November 2014; revised 30 March 2015; accepted 27 April 2015; accepted article preview online 20 May 2015
© 2015 The Society for Investigative Dermatology
(Yellin et al., 1995; Kawai et al., 2008), which also has a synergistic effect on the profibrotic action of T cell–derived cytokines (Atamas et al., 2002). An open-label trial has demonstrated the association between type III collagen reduction and skin softening observed in several patients on T cell–directed therapy (Frances et al., 1988). IFNs are key cytokines involved in innate and adaptive immunity. The dysregulation of type I IFN signaling in SSc has been demonstrated by genetic linkage, as well as molecular and cellular data (Coelho et al., 2008; Eloranta et al., 2010; Carmona et al., 2013; Liu et al., 2013). Increased levels of IFN–inducible transcripts and chemokines have been associated with the more severe form of the disease (Eloranta et al., 2010; Higgs et al., 2011). IFN therapy has also been implicated in the development of SSc (Solans et al., 2004). However, the mechanisms by which the dysregulated IFN signaling contributes to SSc pathogenesis and severity remain unknown. Anifrolumab (previously known as MEDI-546) is an investigational human IgG1κ mAb directed against subunit 1 of the type I IFN receptor. Recently, a Phase I open-label clinical trial was completed in 34 adult SSc patients who received single (0.1, 0.3, 1.0, 3.0, 10.0, or 20.0 mg kg − 1) or four weekly intravenous doses (0.3, 1.0, or 5.0 mg kg − 1 per week) of anifrolumab (Study MI-CP180). The safety and tolerability profile observed in this study supports further clinical development of anifrolumab (Goldberg et al., 2014). www.jidonline.org
1
X Guo et al.
Effects of Type I IFN Blockade in SSc
Neutralization of a type I IFN gene signature by anifrolumab has also been demonstrated in both blood and skin biopsies (Goldberg et al., 2014). Our current study aims to assess downstream effects of anifrolumab and identify potential biomarkers for anti-IFN therapy in SSc patients. Serum samples were procured from the SSc trial of anifrolumab, along with another Phase I SSc trial (Study MICP200) for an anti-CD19 antibody (MEDI-551) to help establish specificity of IFN involvement. The effects of anifrolumab on SSc–associated proteins were determined relative to IFN gene signature neutralization. ECM turnover markers were evaluated pre- and post administration to assess the potential effects of anifrolumab on fibrosis. Finally, differentially expressed transcripts were identified, as were upstream regulators and enriched pathways in whole-blood and skin biopsies after anifrolumab administration. RESULTS Dysregulated serum proteins and association with clinical assessments in SSc patients
Serum samples were procured from SSc patients enrolled in two Phase I trials for anifrolumab and MEDI-551, respectively. The mean modified Rodnan skin score (mRSS) and the Health Assessment Questionnaire-Disability Index (HAQ-DI) were similar for patients receiving placebo, MEDI-551, or anifrolumab (P40.05). Most participants were female, white, and had diffuse cutaneous SSc and Raynaud’s phenomenon. Positive anti-nuclear, anti-cardiolipin, and anti-SCL-70 antibodies were found in similar percentages of patients among three treatment groups (P40.05). There was no significant difference in the percentage of subjects receiving stable doses of antineoplastic and immunomodulation agents (azathioprine, methotrexate, or mycophenolate), systemic steroid (prednisone, methylprednisolone, or hydrocortisone), or anti-malarials for placebo, MEDI-551, and anifrolumab treatment groups (P40.05) (Table 1). We measured 93 proteins in serum samples at baseline from 26 SSc patients enrolled in anifrolumab trial, 21 SSc patients enrolled in MEDI-551 trial, and 30 healthy controls (HCs) recruited separately. Among 47 SSc patients, 31 had an elevated IFN gene signature (IFN-hi) and 15 did not (IFN-lo), whereas one patient had gene expression data missing. IFN-hi patients showed significantly higher mRSS and HAQ-DI compared with IFN-lo patients (Po0.01), suggesting an association between IFN activity and SSc disease activity. There was no significant difference in age, gender, and race among HCs, IFN-lo, and IFN-hi SSc patients (P40.05). Of the 93 measured proteins, 25 showed serum levels below the lowest limit of quantitation in 470% of all samples. No significant difference was found for the distributions of these negative proteins among HCs, IFN-lo, and IFNhi SSc patients; hence, they were excluded from downstream analysis. Principal component analysis of the 68 remaining proteins demonstrated distinct proteomic features among HCs, IFN-lo, and IFN-hi SSc patients (Supplementary Figure S1 online). Comparison of HCs and all SSc patients identified 37 proteins with significantly higher concentrations in SSc patients (Benjamin–Hochberg (BH) Po0.05). Then, analysis 2
Journal of Investigative Dermatology (2015), Volume 00
of variance analysis followed by pairwise comparison demonstrated higher serum levels of 38 proteins in IFN-hi compared with IFN-lo patients or HCs (BH Po0.05). In total, 41 unique proteins were shown to be upregulated in IFN-hi and/or all SSc patients, among whom 25 are inducible by IFN (Rusinova et al., 2013). Interestingly, 12 proteins had significantly higher concentrations in IFN-hi but not IFN-lo patients compared with HCs, suggesting a unique proteomic feature of IFN-hi SSc patients (Supplementary Table S1 online). We then assessed association of the 41 dysregulated proteins with mRSS, HAQ-DI, percent of predicted diffusing capacity for carbon monoxide (DLcoP), percent of predicted forced vital capacity, and three patient reported outcomes included in the European Disease Activity Index (Δ-skin, Δ-vascular, Δ-heart/lung) (Valentini et al., 2001) in SSc patients by Spearman’s correlation analysis. Although peroxiredoxin-4 showed significant correlation with mRSS, CCL11 may be associated with lung disease evidenced by its negative correlation with DLcoP (BH Po0.05). In addition, five proteins (TIMP1, SPP1, IL-2RA, TNFR2, and CCL21) demonstrated significant correlation with HAQ-DI, and IL-13 was related to Δ-vascular (BH Po0.05) (Table 2). We also evaluated the difference of serum protein levels in the presence of anti-nuclear antibody, anti-Scl70 antibody, or four European Disease Activity Index components (scleredema, digital necrosis, arthritis, hypocomplementaemia), respectively (Valentini et al., 2001). The only association that was uncovered was that MCP1 had significantly higher serum level in the 4 SSc patients with digital necrosis compared with the 43 patients without (fold change (FC) = 1.69, BH P = 0.003). Effects of anifrolumab on serum proteins levels of SSc patients
Effects of placebo, MEDI-551, and anifrolumab on the 41 SScrelated proteins were assessed by paired-comparison of preand post treatment serum samples from patients enrolled in MI-CP200 and MI-CP180 trials. Anifrolumab administration was associated with decreased serum levels of four IFN inducible proteins including B2M, CCL13, CXCL10, and VCAM1, along with CD40LG, IL-2RA, and Ferritin (BH Po0.05) (Supplementary Table S2 online and Figures 1a and b). In contrast, MEDI-551 was associated with increased BAFF (B cell-activating factor of the TNF family) level (BH Po0.05) (Figure 1c), whereas no significant change was observed on other serum proteins. No consistent trend was seen in the placebo group for any serum protein. The lack of a sufficiently sized longitudinal placebo control group and mixing of data from two independent trials prevented a direct comparison of change in anifrolumab-treated patients with placebo- or MEDI-551-treated patients. Although no clear trend could be seen between dosing levels and the effect of anifrolumab on serum proteins because of the small sample size for each dosing regimen, Spearman’s analysis indicated significant correlation between IFN gene signature suppression and downregulation of CXCL10 and CD40LG (Figures 1d and e), suggesting an association between IFN blockade and CXCL10/ CD40LG downregulation after anifrolumab administration.
X Guo et al.
Effects of Type I IFN Blockade in SSc
Table 1. Clinical and demographic features of systemic sclerosis patients Category
Feature
Demographics
Age (years) (mean ± SD)
Clinical features
Concomitant medications
Placebo (n = 4)
MEDI-551 (n = 24)
Anifrolumab (n = 34) 47.4 ± 12
42.3 ± 14.4
48.1 ± 8.9
Female (%)
50
70.8
79.4
White (%)
100
83.3
73.5
Diffuse cutaneous systemic SSc (%)
75
87.5
94.1
Raynaud’s (%)
100
100
97.1
mRSS (mean ± SD)
22.8 ± 7.5
23.0 ± 8.7
23.3 ± 8.6
HAQ-DI (mean ± SD)
1.44 ± 0.92
1.04 ± 0.83
1.23 ± 0.7
Positive anti-nuclear antibody (%)
50
79.2
70.6
Positive anti-cardiolipin antibody (%)
0
3.6
5.9
Positive anti-Scl-70 antibody (%)
75
37.5
32.4
Antineoplastic and immunomodulating agents (%)
50
37.5
35.3
Corticosteroids for systemic use (%)
0
20.8
35.3
Anti-malarials (%)
0
0
8.8
Abbreviations: HAQ-DI, Health Assessment Questionnaire-Disability Index; mRSS, modified Rodnan skin score; SSc, systemic sclerosis. No significant difference in age, gender, race, clinical features, and concomitant medication profiles among SSc patients who received placebo, MEDI-551, and anifrolumab. respectively (P40.05).
Table 2. Correlation of serum protein levels with disease severity and IFN activity mRSS HAQ-DI DLcoP Δ-Vascular IFN score PRDX4
0.43**
0.32*
− 0.03
0.49**
IL-2RA
0.35*
0.39** − 0.1
− 0.18
0.24
TIMP1
0.29*
0.51** − 0.18
− 0.06
0.39**
SPP1 (OPN)
0.26*
0.43** − 0.03
0.08
0.45**
CCL21 (6Ckine)
0.31*
0.38** − 0.18
0.04
0.21
TNFRSF1B (TNFR2) 0.13
0.39**
0.04
− 0.06
IL-13
0.06
− 0.03
IFN score
0.22
CCL11 (Eotaxin-1)
0.35*
− 0.01
0.03
− 0.11
− 0.46**
− 0.24
0.16 − 0.02
0.47** 0.14
0.26* − 0.09 − 0.08 —
Abbreviations: DLcoP, diffusing capacity for carbon monoxide; HAQ-DI, Health Assessment Questionnaire-Disability Index; mRSS, modified Rodnan skin score. Spearman’s correlation coefficients were used to assess association between serum protein levels and mRSS, HAQ-DI, percent of predicted DLcoP, patient reported worsening of vascular system as compared with the previous month (Δ-vascular), along with whole-blood type I IFN gene signature score. Bold: IFN-inducible protein. **BH Po0.05; *Unadjusted Po0.05.
ECM turnover markers in SSc patients
Serological markers of ECM remodeling hold great potential as noninvasive markers for connective tissue diseases. In this study, we evaluated a set of protein fingerprint markers for ECM turnover (Karsdal et al., 2013), along with two traditional collagen synthesis markers (amino-terminal propeptide of
type I procollagen and amino-terminal propeptide of type III procollagen (PIIINP)) in SSc patients. Comparison of 30 HCs with 53 to 59 SSc patients at baseline demonstrated significantly higher serum levels of nine markers in SSc subjects, including formation/degradation markers of type II, III, IV, V, VI collagens, and matrix metalloproteinase (MMP)– degraded biglycan. ECM accumulation is governed by the dynamic balance of formation and degradation process of each ECM component, and thus we calculated the ratio of formation and degradation markers for bone and type III/ IV collagens. Both PIIINP/C3M and PRO-C3/C3M ratios were significantly higher in SSc patients compared with HCs, suggesting an excess accumulation of type III collagen in SSc patients. In addition, the ratio of bone resorption marker CTX-I and bone formation marker osteocalcin N-MID was significantly higher in SSc patients, suggesting bone loss in these patients (Table 3). As described in the previous section, we assessed association of the 13 dysregulated ECM turnover markers with various clinical and laboratory variables, including mRSS, HAQ-DI, DLcoP, predicted forced vital capacity, Δ-skin, Δ-vascular, Δ-heart/lung, presence of scleredema, digital necrosis, arthritis, hypocomplementemia, anti-nuclear antibody, or anti-Scl70 antibody in SSc subjects. Significant correlation was found between HAQ-DI and type III collagen synthesis markers. The turnover ratios of type III collagen and bone also showed trends of correlation with HAQ-DI (Table 3). The PIIINP/C3M ratio was significantly decreased after treatment of anifrolumab but not MEDI-551 or placebo (Figure 2a). The same was true for the PRO-C3/C3M ratio. www.jidonline.org
3
X Guo et al.
Effects of Type I IFN Blockade in SSc
100
20 0 –20
50
BAFF change (%)
**
40
0
–50
150 100 50 0
–40 –100
–60 Placebo
0
Placebo
0.1 or 0.3 –1 mg kg–1 1.0 mg kg–1 3.0 mg kg 10.0 mg kg–1 20.0 mg kg–1 0.3 mg kg–1 week × 4 1.0 mg kg–1 week × 4 –1 5.0 mg kg week × 4
–40 Cor: 0.66** –80 –60
–40 –20
0
20
Type I IFN gene signature change (%)
MEDI-551 Anifrolumab
Placebo
MEDI-551
Anifrolumab
0.1 or 0.3 mg kg–1 1.0 mg kg–1 3.0 mg kg–1 10.0 mg kg–1 20.0 mg kg–1 0.3 mg kg–1 week × 4 –1 1.0 mg kg–1 week × 4 5.0 mg kg week × 4
60
–20
–100
–50
MEDI-551 Anifrolumab
CD40LG change (%)
CXCL10 change (%)
20
***
300
*
60
CD40LG change (%)
CXCL10 change (%)
360
40 20 0 –20 –40 –60
Cor: 0.38*
–80
–100 –80 –60 –40 –20 0 20 Type I IFN gene signature change (%)
Figure 1. Effects of anifrolumab and MEDI-551 on serum protein markers. (a and b) Soluble CXCL10 and CD40LG levels were significantly suppressed by anifrolumab in systemic sclerosis (SSc) patients enrolled in the MI-CP180 study but not by placebo or MEDI-551 administration in SSc patients enrolled in the MICP200 study. (c) In contrast, the serum BAFF level was significantly increased after MEDI-551 but not placebo or anifrolumab administration. (d and e) The reduction in CXCL10 and CD40LG level was significantly correlated with whole-blood IFN gene signature neutralization in 23 SSc patients who received single (0.1, 0.3, 1.0, 3.0, 10.0, or 20.0 mg kg − 1) or four weekly intravenous doses (0.3, 1.0, or 5.0 mg kg − 1 per week) of anifrolumab. *BH Po0.05, ** BH Po0.01, and *** BH Po0.001.
Separate analysis of three type III collagen markers indicated that anifrolumab was associated with increased C3M level (Figure 2b), whereas no change was observed for two formation markers, suggesting accelerated collagen degradation after anifrolumab treatment. Regulation of peripheral and skin transcripts by anifrolumab
Using whole-blood microarray data, we identified 290 downregulated probe sets and 26 upregulated probe sets at day 56 after anifrolumab administration in comparison with predose transcript levels (false discovery rate o0.01, |FC| 41.5). Ingenuity pathway analysis identified the canonical pathway—activation of IRF by cytosolic pattern recognition receptors—as the most enriched pathway in anifrolumabregulated transcripts (BH P = 0.006). IFN signaling was the second most enriched pathway among these transcripts (BH P = 0.01). Causal analysis applications within ingenuity pathway analysis aim to identify upstream transcriptional regulators that can explain the observed gene expression changes (Kramer et al., 2014). Not surprisingly, multiple IFN subtypes were shown to be inhibited by anifrolumab. More interestingly, CD40LG was predicted to be suppressed by anifrolumab, consistent with the serum proteomics results. T cell receptor was also one of the upstream regulators 4
Journal of Investigative Dermatology (2015), Volume 00
responsible for the observed down/upregulation of target transcripts by anifrolumab (Supplementary Table S3 online). Skin biopsies were collected at predose and day 28 only for 10 SSc subjects who received multiple doses of anifrolumab. Using the same criteria as in whole-blood gene expression analysis, anifrolumab affected the expression of many more genes in the skin with 4,830 downregulated probe sets and 274 upregulated probe sets (false discovery rate o0.01, |FC|41.5). We selected 526 probe sets with more than 2-fold suppression (Supplementary Table S4 online) for Gene Ontology enrichment analysis (Maere et al., 2005). The results demonstrated that epidermis development, cell adhesion, and ECM organization were among the most enriched biological processes in anifrolumab-suppressed skin transcripts (Supplementary Figure S2 online). Interestingly, we observed significant decrease in transforming growth factor-β receptor 1 (TGF-βR1), TGF-βR2, and TGF-βR3 transcripts after administration. Mediators of latent TGF-β activation—thrombospondin and integrin—also showed reduced expression, suggesting the suppression of TGF-β signaling by anifrolumab. Multiple anifrolumab-suppressed transcripts are indeed TGFβ-regulated genes, such as collagens α1 (I) and α1 (III), fibronectin, connective tissue growth factor, and cartilage oligomeric matrix protein (Sargent et al., 2010). Ingenuity
X Guo et al.
Effects of Type I IFN Blockade in SSc
Table 3. Serum markers for the extracellular matrix turnover Mean FC (SSc/HC)
ρ (HAQ-DI)
Assay name
Target
Reference
PINP
N-terminal propeptide of type I collagen
Veidal et al., 2010
1.19
—
C1M
MMP-2/9/13–degraded type I collagen
Leeming et al., 2011
1.29
—
CTX-I
Cathepsin K–degraded type I collagen
Rosenquist et al., 1998
2.59****
0.26*
C2M
MMP–degraded type II collagen
Bay-Jensen et al., 2011
1.58****
0.11
PIIINP
N-terminal propeptide of type III collagen
Trocme et al., 2006
1.74****
0.42**
PRO-C3
N-protease-mediated cleavage of PIIINP
Nielsen et al., 2013
1.51***
0.33** —
C3M
MMP-9–degraded type III collagen
Barascuk et al., 2010
1.11
P4NP 7 S
7 S domain of type IV collagen
Leeming et al., 2012
1.33***
C4M
MMP-2/9–degraded type IV collagen
Veidal et al., 2011a
1.18**
0.03
PRO-C5
N-terminal propeptide of type V collagen
—
1.31**
0.05
C6M
MMP-2/9–degraded type VI collagen
Veidal et al., 2011b
1.39***
0.15
0.07
—
CRPM
MMP–mediated CRP degradation
Skjot-Arkil et al., 2012
0.88
BGN
MMP-9–mediated biglycan degradation
Genovese et al., 2013
1.82****
ELM7
MMP7–degraded elastin
Skjot-Arkil et al., 2010
0.97
—
N-MID
Cathepin-degraded osteocalcin
Skjot-Arkil et al., 2010
1.28
—
VCANM
MMP–mediated versican degradation
Barascuk et al., 2013
0.95
PIIINP/C3M
Type III collagen turnover
—
1.63****
0.28*
PRO-C3/C3M
Type III collagen turnover
—
1.37***
0.27*
P4NP/C4M
Type IV collagen turnover
—
1.14**
0.01
CTX-I/N-MID
Bone turnover
—
2.17****
0.26*
−0.02
—
Abbreviations: FC, fold change of serum marker levels between SSc patients and HCs; HC, healthy control; HAQ-DI, Health Assessment QuestionnaireDisability Index; ρ, Spearman’s correlation coefficient between HAQ-DI and differentially expressed serum markers; SSc, systemic sclerosis. ****BH Po0.001, ***BH Po0.01, **BH Po0.05, and *unadjusted Po0.05.
upstream regulator analysis also suggested the suppression of TGF-β1 by anifrolumab, with the lowest overlap P-value of 3.05E − 18 and activation z-score of − 4.36. Furthermore, anifrolumab significantly suppressed the expression of TIMP1, TIMP2, and TIMP3 transcripts, which may be responsible for the accelerated degradation of type III collagen described earlier through the activation of matrix metalloproteinases. The suppression of TGF-β-regulated transcripts and TIMPs showed concordant pattern with type I IFNinducible transcripts, suggesting an association between IFN blockade and inhibition of TGF-β signaling after anifrolumab administration (Figure 3). DISCUSSION Our study of the serologic proteome in SSc subjects suggests a distinct proteome profile of IFN-hi SSc subjects compared with IFN-lo SSc patients and HCs (Supplementary Figure S1 online). Overexpressed proteins in SSc subjects are involved in various biological processes, including regulation of T cell proliferation, activation, and movement. The soluble forms of IL-2RA and CD40LG are predominantly released by proteolytic cleavage from the activated CD4+ T cell surface. CXCL10 and CXCL11 are CXCR3 receptor ligands that recruit activated CD4+ and CD8+ T cells to sites of inflammation. In agreement with our results, elevated levels of these proteins have been reported previously in SSc subjects, which
correlate with different disease measures (Degiannis et al., 1990; Steen et al., 1996; Komura et al., 2004; Antonelli et al., 2008; Liu et al., 2013). Immunomodulatory functions of type I IFN include the control of T helper type 1/T helper type 2 function and the regulation of CD8+ cytotoxic T cell activity and memory. Our results demonstrated that anifrolumab administration was associated with peripheral levels of T cell–associated proteins and transcripts regulated by T cell receptor in adult SSc patients, which suggest a suppressive effect of IFN neutralization on T cell activation and/or proliferation in SSc patients. The hallmark of SSc is excessive ECM deposition in the skin, lung, and other internal organs. During the pathological remodeling of ECM, the tissue- and pathology-specific turnover products are released into circulation. Serological markers of collagen synthesis have long been used to assess fibrosis progression in SSc (Hummers, 2010). Increased level of PIIINP was an unfavorable predictor of survival in SSc (Nagy and Czirjak, 2005), and a decrease in PIIINP levels has been reported after the use of immunosuppressant (Heickendorff et al., 1995) or anti-tumor necrosis factor-α body (Denton et al., 2009). ECM accumulation depends on the dynamic balance of synthesis and degradation of each ECM component. Our study indicates increased level of several MMP–degraded neoepitope markers for collagen and biglycan in SSc subjects (Table 3). The ratio of type III www.jidonline.org
5
X Guo et al.
Effects of Type I IFN Blockade in SSc
Color key
160
IFI44L IFI27 IFIT3 MX1 TIMP1 TIMP2 TIMP3 COL1A1 COL1A2 COL3A1 COL5A1 COL5A2 ITGB1 ITGB8 TGFBR1 TGFBR2 TGFBR3 CTGF COMP FGFR3 THBS1 FN1 SPARC
60 PIIINP/C3M change (%)
*** 40
–5 0 log2FC
20 0 –20 –40 –60
0.3 mg kg–1 Placebo
MEDI-551
40 Serum C3M level (ng ml–1)
5.0 mg kg–1
Figure 3. Skin transcripts suppressed by multiple doses of anifrolumab. Significant suppression of type I IFN-inducible transcripts, tissue inhibitors of metalloproteinases (TIMPs), multiple collagens, integrins, transforming growth factor-β (TGF-β) receptors, and other TGF-β-associated transcripts was observed after anifrolumab administration (false discovery rate (FDR) o0.01). The heatmap shows log 2-fold change (FC) of each transcript at day 28 relative to pretreatment for 10 systemic sclerosis (SSc) patient who received four weekly intravenous doses (0.3, 1.0, or 5.0 mg kg − 1 per week) of anifrolumab.
45
35 30 25 20 15 10 5
Anifrolumab
0 Day 0
Day 56
Figure 2. Effects of anifrolumab on type III collagen turnover markers. (a) Significant suppression of amino-terminal propeptide of type III procollagen (PIIINP)/C3M by anifrolumab but not placebo or MEDI-551 administration in systemic sclerosis (SSc) patients enrolled in MI-CP180 and MI-CP200 trials, respectively. ***Po0.001. (b) Serum C3M levels were consistently upregulated by anifrolumab (Po0.001). Red X represents median of the C3M level at day 0 and day 56 after anifrolumab administration, respectively.
collagen synthesis and degradation markers (PIIINP/C3M or PRO-C3/C3M) is associated with disease activity measured by HAQ-DI. Although anifrolumab was not associated with PIIINP change, our results demonstrated consistent suppression of the C3M level associated with administration of the anti-IFNAR antibody, which may result from the reduced levels of MMP inhibitors after treatment. The decreased PIIINP/C3M and PRO-C3/C3M ratios suggest the potential improvement of tissue fibrosis after anifrolumab administration. Furthermore, skin microarray results demonstrated significant suppression of ECM–related transcripts by anifrolumab. Numerous fibrosis-related transcripts were downregulated by anifrolumab. Given that type I IFNs are potent inhibitors of collagen production in vitro (Duncan and Berman, 1987), the suppressive effects of IFN blockade on collagen gene expression might be due to the reduction in inflammation rather than a direct effect to fibroblasts. 6
1.0 mg kg–1
Anifrolumab
Journal of Investigative Dermatology (2015), Volume 00
A complex interplay among activated endothelial cell, T cell, and fibroblast is involved in SSc pathogenesis. Endothelium damage and immune cell infiltration may be alleviated by anifrolumab, evidenced by the decreased concentration of adhesion molecules and chemokines after blockade of IFN signaling. These changes might in turn desensitize SSc fibroblast to autocrine TGF-β signaling through the decreased expression of thrombospondins, integrins, and TGF-β receptors, all of which are involved in latent TGF-β activation and downstream signaling for the stimulation of collagen synthesis (Leask, 2006). Reduced TGF-β signaling and CTGF suppression by anifrolumab may be responsible for the decreased expression of various ECM components and tissue inhibitors of metalloproteinase causing the accelerated degradation of type III collagen. Importantly, our study was not designed to evaluate the effect of anifrolumab on clinical measures of disease activity. Our samples were from a Phase I trial with small sample size and short treatment duration. As no placebo arm was included in this open-label trial, we procured samples from another SSc trial as controls to assess whether or not the observed effects are specific to anifrolumab. Larger clinical trials with randomized placebo-controlled design are required to determine the relationship between changes in peripheral markers and clinical responses of SSc subjects to anifrolumab treatment. Our anti-CD19 antibody, MEDI-551, induced a robust reduction in the B cell and plasma cell gene signatures in both whole-blood and skin biopsies of SSc subjects (Streicher et al., 2014). The increase in the BAFF level after MEDI-551 administration may result from the decrease in its receptors, or a delayed regulation of BAFF mRNA transcription following
X Guo et al.
Effects of Type I IFN Blockade in SSc
B cell depletion (Lavie et al., 2007). Furthermore, plasma cell and B cell gene signature reduction was positively correlated with inhibition of collagen gene expression in skin biopsies post treatment of MEDI-551 (Streicher et al., 2014). Thus, the two antibodies may have an impact on SSc disease activity through different pathways. In summary, our study demonstrates suppressive effects of anifrolumab on T cell activation and collagen accumulation through which tissue fibrosis of SSc subjects may be reduced. Continued clinical studies of anifrolumab in IFN-related diseases including trials in systemic lupus erythematosus will further investigate the relationship between peripheral markers and clinical responses to treatment with anifrolumab.
Whole-genome microarray study PAXGene whole-blood tubes were collected from 29 SSc patients at baseline and day 56 after anifrolumab administration, and skin biopsies were procured from 10 SSc patients at baseline and day 28 after multiple doses of anifrolumab administration. Total RNA was extracted using the PAXgene Blood RNA Kit and the Qiagen RNeasy Fibrous Tissue Mini Kit (Qiagen, Hilden, Germany), respectively. Biotin-labeled cRNA was generated from total RNA and fragmented for hybridization on Affymetrix Human Genome U133 Plus GeneChip arrays (Affymetrix, Santa Clara, CA). Type I IFN gene signature was calculated by whole-blood expression levels of five type I IFN–inducible genes (RSAD2, IFI44, IFI44L, IFI27, and IFI6) and used as a PD marker for anifrolumab (Goldberg et al., 2014). The microarray data has been deposited in Gene Expression Omnibus (GEO) under accession number GSE65336.
MATERIALS AND METHODS Clinical samples
Statistical analysis
In a Phase I, open-label clinical trial, 34 adult SSc subjects received single (0.1, 0.3, 1.0, 3.0, 10.0, or 20.0 mg kg − 1) or a 4-week repeated intravenous doses (0.3, 1.0, or 5.0 mg kg − 1 per week) of anifrolumab (Study MI-CP180). The detailed study design and safety profiles have been reported previously (Goldberg et al., 2014). MEDI-551 is an investigational human IgG1κ mAb that binds to CD19 and depletes B cells. A Phase I, randomized, double-blind, placebo-controlled clinical trial has been conducted to evaluate the safety and tolerability of escalating single doses of MEDI-551 in adult SSc subjects (Study MI-CP200). Identical inclusion/exclusion criteria were used in the two trials. Stable concomitant medication was allowed unless the treatment met the exclusion criterion. Both study protocols were reviewed and approved by Institutional Review Boards or independent ethics committee of each participating center before study initiation. Written informed consent was obtained from each participant before study entry.
Student’s t-test or analysis of variance was used to compare continuous variables among different groups, whereas Fisher’s exact test was used for discrete variables. Multiple comparison problems were controlled by BH adjustment (Benjamini et al., 2001). Differentially expressed genes were identified by significance analysis of microarray (Tusher et al., 2001). All correlations were assessed by Spearman’s correlation tests and the significance was determined via the asymptotic t approximation followed by BH adjustment.
Serum proteomics profiling Sera from 30 HCs, 26 SSc patients at baseline and day 56 after anifrolumab administration, along with 21 SSc patients at baseline and day 85 after either placebo or MEDI-551 administration were procured for the measurement of 93 proteins including chemokines, cytokines, growth factors, soluble receptors, and adhesion molecules (Supplementary Table S2 online). Serum samples were kept frozen at − 80 °C before shipment to Myriad RBM (Austin, TX) for a multiplexed immunoassay based on the Luminex xMAP technology. Samples were processed and analyzed at Myriad RBM according to standard operating procedure.
Serum markers for ECM turnover The amino-terminal propeptides of type I procollagen and PIIINP were measured by electrochemiluminescent immunoassay and radioimmunoassay techniques respectively. Furthermore, we analyzed 14 fingerprint markers for ECM turnover by ELISA in SSc subjects. Briefly, serum samples from 30 HCs and up to 59 SSc patients before and after treatments were analyzed by Pacific Biomarkers (Seattle, WA) for amino-terminal propeptides of type I procollagen/PIIINP analysis and Nordic Biosciences (Herlev, Denmark) for the measurement of the additional ECM markers. Detailed experimental procedures for each marker may be found in references listed in Table 3.
CONFLICT OF INTEREST
XG, BWH, YY, LKR, and WIW are full-time employees of MedImmune LLC and have stock in AstraZeneca.
ACKNOWLEDGMENTS We acknowledge Chris Kane and Dominic Sinibaldi for acquiring and managing proteomic data. We thank Chris Morehouse, Zheng Liu, and Liangwei Wang for providing and managing genomic and clinical data. We would also like to thank Stephen Yoo, Gabriel Robbie, and Kathleen McKeever for leading the clinical and translational teams of MI-CP180 and MI-CP200 trials. SUPPLEMENTARY MATERIAL Supplementary material is linked to the online version of the paper at http:// www.nature.com/jid
REFERENCES Antonelli A, Ferri C, Fallahi P et al. (2008) CXCL10 (alpha) and CCL2 (beta) chemokines in systemic sclerosis—a longitudinal study. Rheumatology (Oxford) 47:45–9 Atamas SP, Luzina IG, Dai H et al. (2002) Synergy between CD40 ligation and IL-4 on fibroblast proliferation involves IL-4 receptor signaling. J Immunol 168:1139–45 Barascuk N, Genovese F, Larsen L et al. (2013) A MMP derived versican neoepitope is elevated in plasma from patients with atherosclerotic heart disease. Int J Clin Exp Med 6:174–84 Barascuk N, Veidal SS, Larsen L et al. (2010) A novel assay for extracellular matrix remodeling associated with liver fibrosis: an enzyme-linked immunosorbent assay (ELISA) for a MMP-9 proteolytically revealed neoepitope of type III collagen. Clin Biochem 43:899–904 Bay-Jensen AC, Liu Q, Byrjalsen I et al. (2011) Enzyme-linked immunosorbent assay (ELISAs) for metalloproteinase derived type II collagen neoepitope, CIIM–increased serum CIIM in subjects with severe radiographic osteoarthritis. Clin Biochem 44:423–9
www.jidonline.org
7
X Guo et al.
Effects of Type I IFN Blockade in SSc
Benjamini Y, Drai D, Elmer G et al. (2001) Controlling the false discovery rate in behavior genetics research. Behav Brain Res 125:279–84 Carmona FD, Martin JE, Beretta L et al. (2013) The systemic lupus erythematosus IRF5 risk haplotype is associated with systemic sclerosis. PLoS One 8: e54419 Coelho LF, de Oliveira JG, Kroon EG (2008) Interferons and scleroderma—a new clue to understanding the pathogenesis of scleroderma? Immunol Lett 118:110–5 Degiannis D, Seibold JR, Czarnecki M et al. (1990) Soluble interleukin-2 receptors in patients with systemic sclerosis. clinical and laboratory correlations. Arthritis Rheum 33:375–80
Nagy Z, Czirjak L (2005) Increased levels of amino terminal propeptide of type III procollagen are an unfavourable predictor of survival in systemic sclerosis. Clin Exp Rheumatol 23:165–72 Nielsen MJ, Nedergaard AF, Sun S et al. (2013) The neo-epitope specific PROC3 ELISA measures true formation of type III collagen associated with liver and muscle parameters. Am J Transl Res 5:303–15 O'Reilly S, Hugle T, van Laar JM (2012) T cells in systemic sclerosis: a reappraisal. Rheumatology (Oxford) 51:1540–9
Denton CP, Engelhart M, Tvede N et al. (2009) An open-label pilot study of infliximab therapy in diffuse cutaneous systemic sclerosis. Ann Rheum Dis 68:1433–9
Prescott RJ, Freemont AJ, Jones CJ et al. (1992) Sequential dermal microvascular and perivascular changes in the development of scleroderma. J Pathol 166: 255–63
Duncan MR, Berman B (1987) Persistence of a reduced-collagen-producing phenotype in cultured scleroderma fibroblasts after short-term exposure to interferons. J Clin Invest 79:1318–24
Rosenquist C, Fledelius C, Christgau S et al. (1998) Serum CrossLaps one step ELISA. first application of monoclonal antibodies for measurement in serum of bone-related degradation products from C-terminal telopeptides of type I collagen. Clin Chem 44:2281–9
Eloranta ML, Franck-Larsson K, Lovgren T et al. (2010) Type I interferon system activation and association with disease manifestations in systemic sclerosis. Ann Rheum Dis 69:1396–402 Frances C, Branchet MC, Bletry O et al. (1988) Skin collagen from scleroderma patients before and after cyclosporin A treatment. Clin Exp Dermatol 13:1–3 Genovese F, Barascuk N, Larsen L et al. (2013) Biglycan fragmentation in pathologies associated with extracellular matrix remodeling by matrix metalloproteinases. Fibrogenesis Tissue Repair 6:1536–6 Goldberg A, Geppert T, Schiopu E et al. (2014) Dose-escalation of human antiinterferon-alpha receptor monoclonal antibody MEDI-546 in subjects with systemic sclerosis: a phase 1, multicenter, open label study. Arthritis Res Ther 16:R57 Heickendorff L, Zachariae H, Bjerring P et al. (1995) The use of serologic markers for collagen synthesis and degradation in systemic sclerosis. J Am Acad Dermatol 32:584–8 Higgs BW, Liu Z, White B et al. (2011) Patients with systemic lupus erythematosus, myositis, rheumatoid arthritis and scleroderma share activation of a common type I interferon pathway. Ann Rheum Dis 70:2029–36 Hummers LK (2010) The current state of biomarkers in systemic sclerosis. Curr Rheumatol Rep 12:34–9 Karsdal MA, Nielsen MJ, Sand JM et al. (2013) Extracellular matrix remodeling: The common denominator in connective tissue diseases. Possibilities for evaluation and current understanding of the matrix as more than a passive architecture, but a key player in tissue failure. Assay Drug Dev Technol 11: 70–92 Kawai M, Masuda A, Kuwana M (2008) A CD40–CD154 interaction in tissue fibrosis. Arthritis Rheum 58:3562–73 Komura K, Sato S, Hasegawa M et al. (2004) Elevated circulating CD40L concentrations in patients with systemic sclerosis. J Rheumatol 31:514–9 Kramer A, Green J, Pollard J Jr et al. (2014) Causal analysis approaches in ingenuity pathway analysis. Bioinformatics 30:523–30 Lavie F, Miceli-Richard C, Ittah M et al. (2007) Increase of B cell-activating factor of the TNF family (BAFF) after rituximab treatment: insights into a new regulating system of BAFF production. Ann Rheum Dis 66:700–3 Leask A (2006) Scar wars: Is TGFbeta the phantom menace in scleroderma? Arthritis Res Ther 8:213 Leeming D, He Y, Veidal S et al. (2011) A novel marker for assessment of liver matrix remodeling: an enzyme-linked immunosorbent assay (ELISA) detecting a MMP generated type I collagen neo-epitope (C1M). Biomarkers 16:616–28 Leeming DJ, Nielsen MJ, Dai Y et al. (2012) Enzyme-linked immunosorbent serum assay specific for the 7 S domain of collagen type IV (P4NP 7 S): a marker related to the extracellular matrix remodeling during liver fibrogenesis. Hepatol Res 42:482–93 Liu X, Mayes MD, Tan FK et al. (2013) Correlation of interferon-inducible chemokine plasma levels with disease severity in systemic sclerosis. Arthritis Rheum 65:226–35
8
Maere S, Heymans K, Kuiper M (2005) BiNGO: a cytoscape plugin to assess overrepresentation of gene ontology categories in biological networks. Bioinformatics 21:3448–9
Journal of Investigative Dermatology (2015), Volume 00
Rusinova I, Forster S, Yu S et al. (2013) Interferome v2.0: an updated database of annotated interferon-regulated genes. Nucleic Acids Res 41:D1040–6 Sargent JL, Milano A, Bhattacharyya S et al. (2010) A TGFbeta-responsive gene signature is associated with a subset of diffuse scleroderma with increased disease severity. J Invest Dermatol 130:694–705 Sato S (1999) Abnormalities of adhesion molecules and chemokines in scleroderma. Curr Opin Rheumatol 11:503–7 Skjot-Arkil H, Barascuk N, Register T et al. (2010) Macrophage-mediated proteolytic remodeling of the extracellular matrix in atherosclerosis results in neoepitopes: A potential new class of biochemical markers. Assay Drug Dev Technol 8:542–52 Skjot-Arkil H, Schett G, Zhang C et al. (2012) Investigation of two novel biochemical markers of inflammation, matrix metalloproteinase and cathepsin generated fragments of C-reactive protein, in patients with ankylosing spondylitis. Clin Exp Rheumatol 30:371–9 Solans R, Bosch JA, Esteban I et al. (2004) Systemic sclerosis developing in association with the use of interferon alpha therapy for chronic viral hepatitis. Clin Exp Rheumatol 22:625–8 Steen VD, Engel EE, Charley MR et al. (1996) Soluble serum interleukin 2 receptors in patients with systemic sclerosis. J Rheumatol 23:646–9 Streicher K, Morehouse CA, Groves CJ et al. (2014) The plasma cell signature in autoimmune disease. Arthritis Rheumatol 66:173–84 Trocme C, Leroy V, Sturm N et al. (2006) Longitudinal evaluation of a fibrosis index combining MMP-1 and PIIINP compared with MMP-9, TIMP-1 and hyaluronic acid in patients with chronic hepatitis C treated by interferonalpha and ribavirin. J Viral Hepat 13:643–51 Tusher VG, Tibshirani R, Chu G (2001) Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA 98: 5116–21 Valentini G, Della Rossa A, Bombardieri S et al. (2001) European multicentre study to define disease activity criteria for systemic sclerosis. II. identification of disease activity variables and development of preliminary activity indexes. Ann Rheum Dis 60:592–8 Veidal SS, Karsdal MA, Nawrocki A et al. (2011a) Assessment of proteolytic degradation of the basement membrane: a fragment of type IV collagen as a biochemical marker for liver fibrosis. Fibrogen Tissue Rep 4:22 Veidal SS, Karsdal MA, Vassiliadis E et al. (2011b) MMP mediated degradation of type VI collagen is highly associated with liver fibrosis—identification and validation of a novel biochemical marker assay. PLoS One 6:e24753. Veidal SS, Vassiliadis E, Bay-Jensen AC et al. (2010) Procollagen type I N-terminal propeptide (PINP) is a marker for fibrogenesis in bile duct ligation-induced fibrosis in rats. Fibrogenesis Tissue Repair 3:5 Yellin MJ, Winikoff S, Fortune SM et al. (1995) Ligation of CD40 on fibroblasts induces CD54 (ICAM-1) and CD106 (VCAM-1) up-regulation and IL-6 production and proliferation. J Leukoc Biol 58:209–16
Suppression of T Cell Activation and Collagen Accumulation by an Anti-IFNAR1 mAb, Anifrolumab, in Adult Patients with Systemic Sclerosis Xiang Guo1, Brandon W. Higgs1, Anne C Bay-Jensen2, Morten A. Karsdal2, Yihong Yao1, Lorin K. Roskos1 and Wendy I. White1 Type I IFNs are implicated in the pathophysiology of systemic sclerosis (SSc). Recently, a Phase I open-label trial was conducted with an anti-IFNAR1 receptor antibody (anifrolumab) in adult SSc patients. In this study, we aim to assess the downstream effects of anifrolumab and elucidate the role of type I IFN in SSc. Serum proteins and extracellular matrix (ECM) markers were measured in relation to IFN pathway activation status and SSc disease activity. Our results demonstrated a robust overexpression of multiple serum proteins in SSc patients, particularly those with an elevated baseline type I IFN gene signature. Anifrolumab administration was associated with significant downregulation of T cell–associated proteins and upregulation of type III collagen degradation marker. Whole-blood and skin microarray results also indicated the inhibition of T cell receptor and ECM–related transcripts by anifrolumab. In summary, our study demonstrates suppressive effects of anifrolumab on T cell activation and collagen accumulation through which tissue fibrosis may be reduced in SSc patients. The relationship between these peripheral markers and the clinical response to anifrolumab may be examined in larger double-blind, placebo-controlled trials. Journal of Investigative Dermatology advance online publication, 18 June 2015; doi:10.1038/jid.2015.188
INTRODUCTION Systemic sclerosis (SSc) is a rare, multisystem autoimmune disease of unknown etiology, characterized by vascular alterations, immunological disturbances, and fibrosis of the skin and internal organs. T cell and macrophage infiltration has been demonstrated in skin biopsies of SSc patients preceding fibrosis (Prescott et al., 1992). Endothelial cell activation and adhesion molecule overexpression are also evident in SSc, leading to the migration of immune cells into the extravascular tissue (Sato, 1999). T cell–derived cytokines are important factors that drive the production of extracellular matrix (ECM) components by fibroblasts resulting in excessive fibrosis in SSc patients (O'Reilly et al., 2012). The interaction between T cells and fibroblasts through CD40LG-CD40 signaling is critical for fibroblast activation and proliferation 1
Translational Sciences, MedImmune LLC, Gaithersburg, Maryland, USA and Nordic Bioscience, Herlev, Denmark
2
Correspondence: Xiang Guo, Translational Sciences, MedImmune LLC, One MedImmune Way, Gaithersburg, Maryland 20878, USA. E-mail: [email protected] Abbreviations: BAFF, B cell–activating factor; BH, Benjamin–Hochberg; DLcoP, diffusing capacity for carbon monoxide; ECM, extracellular matrix; FC, fold change; HAQ-DI, Health Assessment Questionnaire-Disability Index; HC, healthy control; MMP, matrix metalloproteinase; mRSS, modified Rodnan skin score; PIIINP, amino-terminal propeptide of type III procollagen; SSc, systemic sclerosis; TGF-β, transforming growth factor-β Received 3 November 2014; revised 30 March 2015; accepted 27 April 2015; accepted article preview online 20 May 2015
© 2015 The Society for Investigative Dermatology
(Yellin et al., 1995; Kawai et al., 2008), which also has a synergistic effect on the profibrotic action of T cell–derived cytokines (Atamas et al., 2002). An open-label trial has demonstrated the association between type III collagen reduction and skin softening observed in several patients on T cell–directed therapy (Frances et al., 1988). IFNs are key cytokines involved in innate and adaptive immunity. The dysregulation of type I IFN signaling in SSc has been demonstrated by genetic linkage, as well as molecular and cellular data (Coelho et al., 2008; Eloranta et al., 2010; Carmona et al., 2013; Liu et al., 2013). Increased levels of IFN–inducible transcripts and chemokines have been associated with the more severe form of the disease (Eloranta et al., 2010; Higgs et al., 2011). IFN therapy has also been implicated in the development of SSc (Solans et al., 2004). However, the mechanisms by which the dysregulated IFN signaling contributes to SSc pathogenesis and severity remain unknown. Anifrolumab (previously known as MEDI-546) is an investigational human IgG1κ mAb directed against subunit 1 of the type I IFN receptor. Recently, a Phase I open-label clinical trial was completed in 34 adult SSc patients who received single (0.1, 0.3, 1.0, 3.0, 10.0, or 20.0 mg kg − 1) or four weekly intravenous doses (0.3, 1.0, or 5.0 mg kg − 1 per week) of anifrolumab (Study MI-CP180). The safety and tolerability profile observed in this study supports further clinical development of anifrolumab (Goldberg et al., 2014). www.jidonline.org
1
X Guo et al.
Effects of Type I IFN Blockade in SSc
Neutralization of a type I IFN gene signature by anifrolumab has also been demonstrated in both blood and skin biopsies (Goldberg et al., 2014). Our current study aims to assess downstream effects of anifrolumab and identify potential biomarkers for anti-IFN therapy in SSc patients. Serum samples were procured from the SSc trial of anifrolumab, along with another Phase I SSc trial (Study MICP200) for an anti-CD19 antibody (MEDI-551) to help establish specificity of IFN involvement. The effects of anifrolumab on SSc–associated proteins were determined relative to IFN gene signature neutralization. ECM turnover markers were evaluated pre- and post administration to assess the potential effects of anifrolumab on fibrosis. Finally, differentially expressed transcripts were identified, as were upstream regulators and enriched pathways in whole-blood and skin biopsies after anifrolumab administration. RESULTS Dysregulated serum proteins and association with clinical assessments in SSc patients
Serum samples were procured from SSc patients enrolled in two Phase I trials for anifrolumab and MEDI-551, respectively. The mean modified Rodnan skin score (mRSS) and the Health Assessment Questionnaire-Disability Index (HAQ-DI) were similar for patients receiving placebo, MEDI-551, or anifrolumab (P40.05). Most participants were female, white, and had diffuse cutaneous SSc and Raynaud’s phenomenon. Positive anti-nuclear, anti-cardiolipin, and anti-SCL-70 antibodies were found in similar percentages of patients among three treatment groups (P40.05). There was no significant difference in the percentage of subjects receiving stable doses of antineoplastic and immunomodulation agents (azathioprine, methotrexate, or mycophenolate), systemic steroid (prednisone, methylprednisolone, or hydrocortisone), or anti-malarials for placebo, MEDI-551, and anifrolumab treatment groups (P40.05) (Table 1). We measured 93 proteins in serum samples at baseline from 26 SSc patients enrolled in anifrolumab trial, 21 SSc patients enrolled in MEDI-551 trial, and 30 healthy controls (HCs) recruited separately. Among 47 SSc patients, 31 had an elevated IFN gene signature (IFN-hi) and 15 did not (IFN-lo), whereas one patient had gene expression data missing. IFN-hi patients showed significantly higher mRSS and HAQ-DI compared with IFN-lo patients (Po0.01), suggesting an association between IFN activity and SSc disease activity. There was no significant difference in age, gender, and race among HCs, IFN-lo, and IFN-hi SSc patients (P40.05). Of the 93 measured proteins, 25 showed serum levels below the lowest limit of quantitation in 470% of all samples. No significant difference was found for the distributions of these negative proteins among HCs, IFN-lo, and IFNhi SSc patients; hence, they were excluded from downstream analysis. Principal component analysis of the 68 remaining proteins demonstrated distinct proteomic features among HCs, IFN-lo, and IFN-hi SSc patients (Supplementary Figure S1 online). Comparison of HCs and all SSc patients identified 37 proteins with significantly higher concentrations in SSc patients (Benjamin–Hochberg (BH) Po0.05). Then, analysis 2
Journal of Investigative Dermatology (2015), Volume 00
of variance analysis followed by pairwise comparison demonstrated higher serum levels of 38 proteins in IFN-hi compared with IFN-lo patients or HCs (BH Po0.05). In total, 41 unique proteins were shown to be upregulated in IFN-hi and/or all SSc patients, among whom 25 are inducible by IFN (Rusinova et al., 2013). Interestingly, 12 proteins had significantly higher concentrations in IFN-hi but not IFN-lo patients compared with HCs, suggesting a unique proteomic feature of IFN-hi SSc patients (Supplementary Table S1 online). We then assessed association of the 41 dysregulated proteins with mRSS, HAQ-DI, percent of predicted diffusing capacity for carbon monoxide (DLcoP), percent of predicted forced vital capacity, and three patient reported outcomes included in the European Disease Activity Index (Δ-skin, Δ-vascular, Δ-heart/lung) (Valentini et al., 2001) in SSc patients by Spearman’s correlation analysis. Although peroxiredoxin-4 showed significant correlation with mRSS, CCL11 may be associated with lung disease evidenced by its negative correlation with DLcoP (BH Po0.05). In addition, five proteins (TIMP1, SPP1, IL-2RA, TNFR2, and CCL21) demonstrated significant correlation with HAQ-DI, and IL-13 was related to Δ-vascular (BH Po0.05) (Table 2). We also evaluated the difference of serum protein levels in the presence of anti-nuclear antibody, anti-Scl70 antibody, or four European Disease Activity Index components (scleredema, digital necrosis, arthritis, hypocomplementaemia), respectively (Valentini et al., 2001). The only association that was uncovered was that MCP1 had significantly higher serum level in the 4 SSc patients with digital necrosis compared with the 43 patients without (fold change (FC) = 1.69, BH P = 0.003). Effects of anifrolumab on serum proteins levels of SSc patients
Effects of placebo, MEDI-551, and anifrolumab on the 41 SScrelated proteins were assessed by paired-comparison of preand post treatment serum samples from patients enrolled in MI-CP200 and MI-CP180 trials. Anifrolumab administration was associated with decreased serum levels of four IFN inducible proteins including B2M, CCL13, CXCL10, and VCAM1, along with CD40LG, IL-2RA, and Ferritin (BH Po0.05) (Supplementary Table S2 online and Figures 1a and b). In contrast, MEDI-551 was associated with increased BAFF (B cell-activating factor of the TNF family) level (BH Po0.05) (Figure 1c), whereas no significant change was observed on other serum proteins. No consistent trend was seen in the placebo group for any serum protein. The lack of a sufficiently sized longitudinal placebo control group and mixing of data from two independent trials prevented a direct comparison of change in anifrolumab-treated patients with placebo- or MEDI-551-treated patients. Although no clear trend could be seen between dosing levels and the effect of anifrolumab on serum proteins because of the small sample size for each dosing regimen, Spearman’s analysis indicated significant correlation between IFN gene signature suppression and downregulation of CXCL10 and CD40LG (Figures 1d and e), suggesting an association between IFN blockade and CXCL10/ CD40LG downregulation after anifrolumab administration.
X Guo et al.
Effects of Type I IFN Blockade in SSc
Table 1. Clinical and demographic features of systemic sclerosis patients Category
Feature
Demographics
Age (years) (mean ± SD)
Clinical features
Concomitant medications
Placebo (n = 4)
MEDI-551 (n = 24)
Anifrolumab (n = 34) 47.4 ± 12
42.3 ± 14.4
48.1 ± 8.9
Female (%)
50
70.8
79.4
White (%)
100
83.3
73.5
Diffuse cutaneous systemic SSc (%)
75
87.5
94.1
Raynaud’s (%)
100
100
97.1
mRSS (mean ± SD)
22.8 ± 7.5
23.0 ± 8.7
23.3 ± 8.6
HAQ-DI (mean ± SD)
1.44 ± 0.92
1.04 ± 0.83
1.23 ± 0.7
Positive anti-nuclear antibody (%)
50
79.2
70.6
Positive anti-cardiolipin antibody (%)
0
3.6
5.9
Positive anti-Scl-70 antibody (%)
75
37.5
32.4
Antineoplastic and immunomodulating agents (%)
50
37.5
35.3
Corticosteroids for systemic use (%)
0
20.8
35.3
Anti-malarials (%)
0
0
8.8
Abbreviations: HAQ-DI, Health Assessment Questionnaire-Disability Index; mRSS, modified Rodnan skin score; SSc, systemic sclerosis. No significant difference in age, gender, race, clinical features, and concomitant medication profiles among SSc patients who received placebo, MEDI-551, and anifrolumab. respectively (P40.05).
Table 2. Correlation of serum protein levels with disease severity and IFN activity mRSS HAQ-DI DLcoP Δ-Vascular IFN score PRDX4
0.43**
0.32*
− 0.03
0.49**
IL-2RA
0.35*
0.39** − 0.1
− 0.18
0.24
TIMP1
0.29*
0.51** − 0.18
− 0.06
0.39**
SPP1 (OPN)
0.26*
0.43** − 0.03
0.08
0.45**
CCL21 (6Ckine)
0.31*
0.38** − 0.18
0.04
0.21
TNFRSF1B (TNFR2) 0.13
0.39**
0.04
− 0.06
IL-13
0.06
− 0.03
IFN score
0.22
CCL11 (Eotaxin-1)
0.35*
− 0.01
0.03
− 0.11
− 0.46**
− 0.24
0.16 − 0.02
0.47** 0.14
0.26* − 0.09 − 0.08 —
Abbreviations: DLcoP, diffusing capacity for carbon monoxide; HAQ-DI, Health Assessment Questionnaire-Disability Index; mRSS, modified Rodnan skin score. Spearman’s correlation coefficients were used to assess association between serum protein levels and mRSS, HAQ-DI, percent of predicted DLcoP, patient reported worsening of vascular system as compared with the previous month (Δ-vascular), along with whole-blood type I IFN gene signature score. Bold: IFN-inducible protein. **BH Po0.05; *Unadjusted Po0.05.
ECM turnover markers in SSc patients
Serological markers of ECM remodeling hold great potential as noninvasive markers for connective tissue diseases. In this study, we evaluated a set of protein fingerprint markers for ECM turnover (Karsdal et al., 2013), along with two traditional collagen synthesis markers (amino-terminal propeptide of
type I procollagen and amino-terminal propeptide of type III procollagen (PIIINP)) in SSc patients. Comparison of 30 HCs with 53 to 59 SSc patients at baseline demonstrated significantly higher serum levels of nine markers in SSc subjects, including formation/degradation markers of type II, III, IV, V, VI collagens, and matrix metalloproteinase (MMP)– degraded biglycan. ECM accumulation is governed by the dynamic balance of formation and degradation process of each ECM component, and thus we calculated the ratio of formation and degradation markers for bone and type III/ IV collagens. Both PIIINP/C3M and PRO-C3/C3M ratios were significantly higher in SSc patients compared with HCs, suggesting an excess accumulation of type III collagen in SSc patients. In addition, the ratio of bone resorption marker CTX-I and bone formation marker osteocalcin N-MID was significantly higher in SSc patients, suggesting bone loss in these patients (Table 3). As described in the previous section, we assessed association of the 13 dysregulated ECM turnover markers with various clinical and laboratory variables, including mRSS, HAQ-DI, DLcoP, predicted forced vital capacity, Δ-skin, Δ-vascular, Δ-heart/lung, presence of scleredema, digital necrosis, arthritis, hypocomplementemia, anti-nuclear antibody, or anti-Scl70 antibody in SSc subjects. Significant correlation was found between HAQ-DI and type III collagen synthesis markers. The turnover ratios of type III collagen and bone also showed trends of correlation with HAQ-DI (Table 3). The PIIINP/C3M ratio was significantly decreased after treatment of anifrolumab but not MEDI-551 or placebo (Figure 2a). The same was true for the PRO-C3/C3M ratio. www.jidonline.org
3
X Guo et al.
Effects of Type I IFN Blockade in SSc
100
20 0 –20
50
BAFF change (%)
**
40
0
–50
150 100 50 0
–40 –100
–60 Placebo
0
Placebo
0.1 or 0.3 –1 mg kg–1 1.0 mg kg–1 3.0 mg kg 10.0 mg kg–1 20.0 mg kg–1 0.3 mg kg–1 week × 4 1.0 mg kg–1 week × 4 –1 5.0 mg kg week × 4
–40 Cor: 0.66** –80 –60
–40 –20
0
20
Type I IFN gene signature change (%)
MEDI-551 Anifrolumab
Placebo
MEDI-551
Anifrolumab
0.1 or 0.3 mg kg–1 1.0 mg kg–1 3.0 mg kg–1 10.0 mg kg–1 20.0 mg kg–1 0.3 mg kg–1 week × 4 –1 1.0 mg kg–1 week × 4 5.0 mg kg week × 4
60
–20
–100
–50
MEDI-551 Anifrolumab
CD40LG change (%)
CXCL10 change (%)
20
***
300
*
60
CD40LG change (%)
CXCL10 change (%)
360
40 20 0 –20 –40 –60
Cor: 0.38*
–80
–100 –80 –60 –40 –20 0 20 Type I IFN gene signature change (%)
Figure 1. Effects of anifrolumab and MEDI-551 on serum protein markers. (a and b) Soluble CXCL10 and CD40LG levels were significantly suppressed by anifrolumab in systemic sclerosis (SSc) patients enrolled in the MI-CP180 study but not by placebo or MEDI-551 administration in SSc patients enrolled in the MICP200 study. (c) In contrast, the serum BAFF level was significantly increased after MEDI-551 but not placebo or anifrolumab administration. (d and e) The reduction in CXCL10 and CD40LG level was significantly correlated with whole-blood IFN gene signature neutralization in 23 SSc patients who received single (0.1, 0.3, 1.0, 3.0, 10.0, or 20.0 mg kg − 1) or four weekly intravenous doses (0.3, 1.0, or 5.0 mg kg − 1 per week) of anifrolumab. *BH Po0.05, ** BH Po0.01, and *** BH Po0.001.
Separate analysis of three type III collagen markers indicated that anifrolumab was associated with increased C3M level (Figure 2b), whereas no change was observed for two formation markers, suggesting accelerated collagen degradation after anifrolumab treatment. Regulation of peripheral and skin transcripts by anifrolumab
Using whole-blood microarray data, we identified 290 downregulated probe sets and 26 upregulated probe sets at day 56 after anifrolumab administration in comparison with predose transcript levels (false discovery rate o0.01, |FC| 41.5). Ingenuity pathway analysis identified the canonical pathway—activation of IRF by cytosolic pattern recognition receptors—as the most enriched pathway in anifrolumabregulated transcripts (BH P = 0.006). IFN signaling was the second most enriched pathway among these transcripts (BH P = 0.01). Causal analysis applications within ingenuity pathway analysis aim to identify upstream transcriptional regulators that can explain the observed gene expression changes (Kramer et al., 2014). Not surprisingly, multiple IFN subtypes were shown to be inhibited by anifrolumab. More interestingly, CD40LG was predicted to be suppressed by anifrolumab, consistent with the serum proteomics results. T cell receptor was also one of the upstream regulators 4
Journal of Investigative Dermatology (2015), Volume 00
responsible for the observed down/upregulation of target transcripts by anifrolumab (Supplementary Table S3 online). Skin biopsies were collected at predose and day 28 only for 10 SSc subjects who received multiple doses of anifrolumab. Using the same criteria as in whole-blood gene expression analysis, anifrolumab affected the expression of many more genes in the skin with 4,830 downregulated probe sets and 274 upregulated probe sets (false discovery rate o0.01, |FC|41.5). We selected 526 probe sets with more than 2-fold suppression (Supplementary Table S4 online) for Gene Ontology enrichment analysis (Maere et al., 2005). The results demonstrated that epidermis development, cell adhesion, and ECM organization were among the most enriched biological processes in anifrolumab-suppressed skin transcripts (Supplementary Figure S2 online). Interestingly, we observed significant decrease in transforming growth factor-β receptor 1 (TGF-βR1), TGF-βR2, and TGF-βR3 transcripts after administration. Mediators of latent TGF-β activation—thrombospondin and integrin—also showed reduced expression, suggesting the suppression of TGF-β signaling by anifrolumab. Multiple anifrolumab-suppressed transcripts are indeed TGFβ-regulated genes, such as collagens α1 (I) and α1 (III), fibronectin, connective tissue growth factor, and cartilage oligomeric matrix protein (Sargent et al., 2010). Ingenuity
X Guo et al.
Effects of Type I IFN Blockade in SSc
Table 3. Serum markers for the extracellular matrix turnover Mean FC (SSc/HC)
ρ (HAQ-DI)
Assay name
Target
Reference
PINP
N-terminal propeptide of type I collagen
Veidal et al., 2010
1.19
—
C1M
MMP-2/9/13–degraded type I collagen
Leeming et al., 2011
1.29
—
CTX-I
Cathepsin K–degraded type I collagen
Rosenquist et al., 1998
2.59****
0.26*
C2M
MMP–degraded type II collagen
Bay-Jensen et al., 2011
1.58****
0.11
PIIINP
N-terminal propeptide of type III collagen
Trocme et al., 2006
1.74****
0.42**
PRO-C3
N-protease-mediated cleavage of PIIINP
Nielsen et al., 2013
1.51***
0.33** —
C3M
MMP-9–degraded type III collagen
Barascuk et al., 2010
1.11
P4NP 7 S
7 S domain of type IV collagen
Leeming et al., 2012
1.33***
C4M
MMP-2/9–degraded type IV collagen
Veidal et al., 2011a
1.18**
0.03
PRO-C5
N-terminal propeptide of type V collagen
—
1.31**
0.05
C6M
MMP-2/9–degraded type VI collagen
Veidal et al., 2011b
1.39***
0.15
0.07
—
CRPM
MMP–mediated CRP degradation
Skjot-Arkil et al., 2012
0.88
BGN
MMP-9–mediated biglycan degradation
Genovese et al., 2013
1.82****
ELM7
MMP7–degraded elastin
Skjot-Arkil et al., 2010
0.97
—
N-MID
Cathepin-degraded osteocalcin
Skjot-Arkil et al., 2010
1.28
—
VCANM
MMP–mediated versican degradation
Barascuk et al., 2013
0.95
PIIINP/C3M
Type III collagen turnover
—
1.63****
0.28*
PRO-C3/C3M
Type III collagen turnover
—
1.37***
0.27*
P4NP/C4M
Type IV collagen turnover
—
1.14**
0.01
CTX-I/N-MID
Bone turnover
—
2.17****
0.26*
−0.02
—
Abbreviations: FC, fold change of serum marker levels between SSc patients and HCs; HC, healthy control; HAQ-DI, Health Assessment QuestionnaireDisability Index; ρ, Spearman’s correlation coefficient between HAQ-DI and differentially expressed serum markers; SSc, systemic sclerosis. ****BH Po0.001, ***BH Po0.01, **BH Po0.05, and *unadjusted Po0.05.
upstream regulator analysis also suggested the suppression of TGF-β1 by anifrolumab, with the lowest overlap P-value of 3.05E − 18 and activation z-score of − 4.36. Furthermore, anifrolumab significantly suppressed the expression of TIMP1, TIMP2, and TIMP3 transcripts, which may be responsible for the accelerated degradation of type III collagen described earlier through the activation of matrix metalloproteinases. The suppression of TGF-β-regulated transcripts and TIMPs showed concordant pattern with type I IFNinducible transcripts, suggesting an association between IFN blockade and inhibition of TGF-β signaling after anifrolumab administration (Figure 3). DISCUSSION Our study of the serologic proteome in SSc subjects suggests a distinct proteome profile of IFN-hi SSc subjects compared with IFN-lo SSc patients and HCs (Supplementary Figure S1 online). Overexpressed proteins in SSc subjects are involved in various biological processes, including regulation of T cell proliferation, activation, and movement. The soluble forms of IL-2RA and CD40LG are predominantly released by proteolytic cleavage from the activated CD4+ T cell surface. CXCL10 and CXCL11 are CXCR3 receptor ligands that recruit activated CD4+ and CD8+ T cells to sites of inflammation. In agreement with our results, elevated levels of these proteins have been reported previously in SSc subjects, which
correlate with different disease measures (Degiannis et al., 1990; Steen et al., 1996; Komura et al., 2004; Antonelli et al., 2008; Liu et al., 2013). Immunomodulatory functions of type I IFN include the control of T helper type 1/T helper type 2 function and the regulation of CD8+ cytotoxic T cell activity and memory. Our results demonstrated that anifrolumab administration was associated with peripheral levels of T cell–associated proteins and transcripts regulated by T cell receptor in adult SSc patients, which suggest a suppressive effect of IFN neutralization on T cell activation and/or proliferation in SSc patients. The hallmark of SSc is excessive ECM deposition in the skin, lung, and other internal organs. During the pathological remodeling of ECM, the tissue- and pathology-specific turnover products are released into circulation. Serological markers of collagen synthesis have long been used to assess fibrosis progression in SSc (Hummers, 2010). Increased level of PIIINP was an unfavorable predictor of survival in SSc (Nagy and Czirjak, 2005), and a decrease in PIIINP levels has been reported after the use of immunosuppressant (Heickendorff et al., 1995) or anti-tumor necrosis factor-α body (Denton et al., 2009). ECM accumulation depends on the dynamic balance of synthesis and degradation of each ECM component. Our study indicates increased level of several MMP–degraded neoepitope markers for collagen and biglycan in SSc subjects (Table 3). The ratio of type III www.jidonline.org
5
X Guo et al.
Effects of Type I IFN Blockade in SSc
Color key
160
IFI44L IFI27 IFIT3 MX1 TIMP1 TIMP2 TIMP3 COL1A1 COL1A2 COL3A1 COL5A1 COL5A2 ITGB1 ITGB8 TGFBR1 TGFBR2 TGFBR3 CTGF COMP FGFR3 THBS1 FN1 SPARC
60 PIIINP/C3M change (%)
*** 40
–5 0 log2FC
20 0 –20 –40 –60
0.3 mg kg–1 Placebo
MEDI-551
40 Serum C3M level (ng ml–1)
5.0 mg kg–1
Figure 3. Skin transcripts suppressed by multiple doses of anifrolumab. Significant suppression of type I IFN-inducible transcripts, tissue inhibitors of metalloproteinases (TIMPs), multiple collagens, integrins, transforming growth factor-β (TGF-β) receptors, and other TGF-β-associated transcripts was observed after anifrolumab administration (false discovery rate (FDR) o0.01). The heatmap shows log 2-fold change (FC) of each transcript at day 28 relative to pretreatment for 10 systemic sclerosis (SSc) patient who received four weekly intravenous doses (0.3, 1.0, or 5.0 mg kg − 1 per week) of anifrolumab.
45
35 30 25 20 15 10 5
Anifrolumab
0 Day 0
Day 56
Figure 2. Effects of anifrolumab on type III collagen turnover markers. (a) Significant suppression of amino-terminal propeptide of type III procollagen (PIIINP)/C3M by anifrolumab but not placebo or MEDI-551 administration in systemic sclerosis (SSc) patients enrolled in MI-CP180 and MI-CP200 trials, respectively. ***Po0.001. (b) Serum C3M levels were consistently upregulated by anifrolumab (Po0.001). Red X represents median of the C3M level at day 0 and day 56 after anifrolumab administration, respectively.
collagen synthesis and degradation markers (PIIINP/C3M or PRO-C3/C3M) is associated with disease activity measured by HAQ-DI. Although anifrolumab was not associated with PIIINP change, our results demonstrated consistent suppression of the C3M level associated with administration of the anti-IFNAR antibody, which may result from the reduced levels of MMP inhibitors after treatment. The decreased PIIINP/C3M and PRO-C3/C3M ratios suggest the potential improvement of tissue fibrosis after anifrolumab administration. Furthermore, skin microarray results demonstrated significant suppression of ECM–related transcripts by anifrolumab. Numerous fibrosis-related transcripts were downregulated by anifrolumab. Given that type I IFNs are potent inhibitors of collagen production in vitro (Duncan and Berman, 1987), the suppressive effects of IFN blockade on collagen gene expression might be due to the reduction in inflammation rather than a direct effect to fibroblasts. 6
1.0 mg kg–1
Anifrolumab
Journal of Investigative Dermatology (2015), Volume 00
A complex interplay among activated endothelial cell, T cell, and fibroblast is involved in SSc pathogenesis. Endothelium damage and immune cell infiltration may be alleviated by anifrolumab, evidenced by the decreased concentration of adhesion molecules and chemokines after blockade of IFN signaling. These changes might in turn desensitize SSc fibroblast to autocrine TGF-β signaling through the decreased expression of thrombospondins, integrins, and TGF-β receptors, all of which are involved in latent TGF-β activation and downstream signaling for the stimulation of collagen synthesis (Leask, 2006). Reduced TGF-β signaling and CTGF suppression by anifrolumab may be responsible for the decreased expression of various ECM components and tissue inhibitors of metalloproteinase causing the accelerated degradation of type III collagen. Importantly, our study was not designed to evaluate the effect of anifrolumab on clinical measures of disease activity. Our samples were from a Phase I trial with small sample size and short treatment duration. As no placebo arm was included in this open-label trial, we procured samples from another SSc trial as controls to assess whether or not the observed effects are specific to anifrolumab. Larger clinical trials with randomized placebo-controlled design are required to determine the relationship between changes in peripheral markers and clinical responses of SSc subjects to anifrolumab treatment. Our anti-CD19 antibody, MEDI-551, induced a robust reduction in the B cell and plasma cell gene signatures in both whole-blood and skin biopsies of SSc subjects (Streicher et al., 2014). The increase in the BAFF level after MEDI-551 administration may result from the decrease in its receptors, or a delayed regulation of BAFF mRNA transcription following
X Guo et al.
Effects of Type I IFN Blockade in SSc
B cell depletion (Lavie et al., 2007). Furthermore, plasma cell and B cell gene signature reduction was positively correlated with inhibition of collagen gene expression in skin biopsies post treatment of MEDI-551 (Streicher et al., 2014). Thus, the two antibodies may have an impact on SSc disease activity through different pathways. In summary, our study demonstrates suppressive effects of anifrolumab on T cell activation and collagen accumulation through which tissue fibrosis of SSc subjects may be reduced. Continued clinical studies of anifrolumab in IFN-related diseases including trials in systemic lupus erythematosus will further investigate the relationship between peripheral markers and clinical responses to treatment with anifrolumab.
Whole-genome microarray study PAXGene whole-blood tubes were collected from 29 SSc patients at baseline and day 56 after anifrolumab administration, and skin biopsies were procured from 10 SSc patients at baseline and day 28 after multiple doses of anifrolumab administration. Total RNA was extracted using the PAXgene Blood RNA Kit and the Qiagen RNeasy Fibrous Tissue Mini Kit (Qiagen, Hilden, Germany), respectively. Biotin-labeled cRNA was generated from total RNA and fragmented for hybridization on Affymetrix Human Genome U133 Plus GeneChip arrays (Affymetrix, Santa Clara, CA). Type I IFN gene signature was calculated by whole-blood expression levels of five type I IFN–inducible genes (RSAD2, IFI44, IFI44L, IFI27, and IFI6) and used as a PD marker for anifrolumab (Goldberg et al., 2014). The microarray data has been deposited in Gene Expression Omnibus (GEO) under accession number GSE65336.
MATERIALS AND METHODS Clinical samples
Statistical analysis
In a Phase I, open-label clinical trial, 34 adult SSc subjects received single (0.1, 0.3, 1.0, 3.0, 10.0, or 20.0 mg kg − 1) or a 4-week repeated intravenous doses (0.3, 1.0, or 5.0 mg kg − 1 per week) of anifrolumab (Study MI-CP180). The detailed study design and safety profiles have been reported previously (Goldberg et al., 2014). MEDI-551 is an investigational human IgG1κ mAb that binds to CD19 and depletes B cells. A Phase I, randomized, double-blind, placebo-controlled clinical trial has been conducted to evaluate the safety and tolerability of escalating single doses of MEDI-551 in adult SSc subjects (Study MI-CP200). Identical inclusion/exclusion criteria were used in the two trials. Stable concomitant medication was allowed unless the treatment met the exclusion criterion. Both study protocols were reviewed and approved by Institutional Review Boards or independent ethics committee of each participating center before study initiation. Written informed consent was obtained from each participant before study entry.
Student’s t-test or analysis of variance was used to compare continuous variables among different groups, whereas Fisher’s exact test was used for discrete variables. Multiple comparison problems were controlled by BH adjustment (Benjamini et al., 2001). Differentially expressed genes were identified by significance analysis of microarray (Tusher et al., 2001). All correlations were assessed by Spearman’s correlation tests and the significance was determined via the asymptotic t approximation followed by BH adjustment.
Serum proteomics profiling Sera from 30 HCs, 26 SSc patients at baseline and day 56 after anifrolumab administration, along with 21 SSc patients at baseline and day 85 after either placebo or MEDI-551 administration were procured for the measurement of 93 proteins including chemokines, cytokines, growth factors, soluble receptors, and adhesion molecules (Supplementary Table S2 online). Serum samples were kept frozen at − 80 °C before shipment to Myriad RBM (Austin, TX) for a multiplexed immunoassay based on the Luminex xMAP technology. Samples were processed and analyzed at Myriad RBM according to standard operating procedure.
Serum markers for ECM turnover The amino-terminal propeptides of type I procollagen and PIIINP were measured by electrochemiluminescent immunoassay and radioimmunoassay techniques respectively. Furthermore, we analyzed 14 fingerprint markers for ECM turnover by ELISA in SSc subjects. Briefly, serum samples from 30 HCs and up to 59 SSc patients before and after treatments were analyzed by Pacific Biomarkers (Seattle, WA) for amino-terminal propeptides of type I procollagen/PIIINP analysis and Nordic Biosciences (Herlev, Denmark) for the measurement of the additional ECM markers. Detailed experimental procedures for each marker may be found in references listed in Table 3.
CONFLICT OF INTEREST
XG, BWH, YY, LKR, and WIW are full-time employees of MedImmune LLC and have stock in AstraZeneca.
ACKNOWLEDGMENTS We acknowledge Chris Kane and Dominic Sinibaldi for acquiring and managing proteomic data. We thank Chris Morehouse, Zheng Liu, and Liangwei Wang for providing and managing genomic and clinical data. We would also like to thank Stephen Yoo, Gabriel Robbie, and Kathleen McKeever for leading the clinical and translational teams of MI-CP180 and MI-CP200 trials. SUPPLEMENTARY MATERIAL Supplementary material is linked to the online version of the paper at http:// www.nature.com/jid
REFERENCES Antonelli A, Ferri C, Fallahi P et al. (2008) CXCL10 (alpha) and CCL2 (beta) chemokines in systemic sclerosis—a longitudinal study. Rheumatology (Oxford) 47:45–9 Atamas SP, Luzina IG, Dai H et al. (2002) Synergy between CD40 ligation and IL-4 on fibroblast proliferation involves IL-4 receptor signaling. J Immunol 168:1139–45 Barascuk N, Genovese F, Larsen L et al. (2013) A MMP derived versican neoepitope is elevated in plasma from patients with atherosclerotic heart disease. Int J Clin Exp Med 6:174–84 Barascuk N, Veidal SS, Larsen L et al. (2010) A novel assay for extracellular matrix remodeling associated with liver fibrosis: an enzyme-linked immunosorbent assay (ELISA) for a MMP-9 proteolytically revealed neoepitope of type III collagen. Clin Biochem 43:899–904 Bay-Jensen AC, Liu Q, Byrjalsen I et al. (2011) Enzyme-linked immunosorbent assay (ELISAs) for metalloproteinase derived type II collagen neoepitope, CIIM–increased serum CIIM in subjects with severe radiographic osteoarthritis. Clin Biochem 44:423–9
www.jidonline.org
7
X Guo et al.
Effects of Type I IFN Blockade in SSc
Benjamini Y, Drai D, Elmer G et al. (2001) Controlling the false discovery rate in behavior genetics research. Behav Brain Res 125:279–84 Carmona FD, Martin JE, Beretta L et al. (2013) The systemic lupus erythematosus IRF5 risk haplotype is associated with systemic sclerosis. PLoS One 8: e54419 Coelho LF, de Oliveira JG, Kroon EG (2008) Interferons and scleroderma—a new clue to understanding the pathogenesis of scleroderma? Immunol Lett 118:110–5 Degiannis D, Seibold JR, Czarnecki M et al. (1990) Soluble interleukin-2 receptors in patients with systemic sclerosis. clinical and laboratory correlations. Arthritis Rheum 33:375–80
Nagy Z, Czirjak L (2005) Increased levels of amino terminal propeptide of type III procollagen are an unfavourable predictor of survival in systemic sclerosis. Clin Exp Rheumatol 23:165–72 Nielsen MJ, Nedergaard AF, Sun S et al. (2013) The neo-epitope specific PROC3 ELISA measures true formation of type III collagen associated with liver and muscle parameters. Am J Transl Res 5:303–15 O'Reilly S, Hugle T, van Laar JM (2012) T cells in systemic sclerosis: a reappraisal. Rheumatology (Oxford) 51:1540–9
Denton CP, Engelhart M, Tvede N et al. (2009) An open-label pilot study of infliximab therapy in diffuse cutaneous systemic sclerosis. Ann Rheum Dis 68:1433–9
Prescott RJ, Freemont AJ, Jones CJ et al. (1992) Sequential dermal microvascular and perivascular changes in the development of scleroderma. J Pathol 166: 255–63
Duncan MR, Berman B (1987) Persistence of a reduced-collagen-producing phenotype in cultured scleroderma fibroblasts after short-term exposure to interferons. J Clin Invest 79:1318–24
Rosenquist C, Fledelius C, Christgau S et al. (1998) Serum CrossLaps one step ELISA. first application of monoclonal antibodies for measurement in serum of bone-related degradation products from C-terminal telopeptides of type I collagen. Clin Chem 44:2281–9
Eloranta ML, Franck-Larsson K, Lovgren T et al. (2010) Type I interferon system activation and association with disease manifestations in systemic sclerosis. Ann Rheum Dis 69:1396–402 Frances C, Branchet MC, Bletry O et al. (1988) Skin collagen from scleroderma patients before and after cyclosporin A treatment. Clin Exp Dermatol 13:1–3 Genovese F, Barascuk N, Larsen L et al. (2013) Biglycan fragmentation in pathologies associated with extracellular matrix remodeling by matrix metalloproteinases. Fibrogenesis Tissue Repair 6:1536–6 Goldberg A, Geppert T, Schiopu E et al. (2014) Dose-escalation of human antiinterferon-alpha receptor monoclonal antibody MEDI-546 in subjects with systemic sclerosis: a phase 1, multicenter, open label study. Arthritis Res Ther 16:R57 Heickendorff L, Zachariae H, Bjerring P et al. (1995) The use of serologic markers for collagen synthesis and degradation in systemic sclerosis. J Am Acad Dermatol 32:584–8 Higgs BW, Liu Z, White B et al. (2011) Patients with systemic lupus erythematosus, myositis, rheumatoid arthritis and scleroderma share activation of a common type I interferon pathway. Ann Rheum Dis 70:2029–36 Hummers LK (2010) The current state of biomarkers in systemic sclerosis. Curr Rheumatol Rep 12:34–9 Karsdal MA, Nielsen MJ, Sand JM et al. (2013) Extracellular matrix remodeling: The common denominator in connective tissue diseases. Possibilities for evaluation and current understanding of the matrix as more than a passive architecture, but a key player in tissue failure. Assay Drug Dev Technol 11: 70–92 Kawai M, Masuda A, Kuwana M (2008) A CD40–CD154 interaction in tissue fibrosis. Arthritis Rheum 58:3562–73 Komura K, Sato S, Hasegawa M et al. (2004) Elevated circulating CD40L concentrations in patients with systemic sclerosis. J Rheumatol 31:514–9 Kramer A, Green J, Pollard J Jr et al. (2014) Causal analysis approaches in ingenuity pathway analysis. Bioinformatics 30:523–30 Lavie F, Miceli-Richard C, Ittah M et al. (2007) Increase of B cell-activating factor of the TNF family (BAFF) after rituximab treatment: insights into a new regulating system of BAFF production. Ann Rheum Dis 66:700–3 Leask A (2006) Scar wars: Is TGFbeta the phantom menace in scleroderma? Arthritis Res Ther 8:213 Leeming D, He Y, Veidal S et al. (2011) A novel marker for assessment of liver matrix remodeling: an enzyme-linked immunosorbent assay (ELISA) detecting a MMP generated type I collagen neo-epitope (C1M). Biomarkers 16:616–28 Leeming DJ, Nielsen MJ, Dai Y et al. (2012) Enzyme-linked immunosorbent serum assay specific for the 7 S domain of collagen type IV (P4NP 7 S): a marker related to the extracellular matrix remodeling during liver fibrogenesis. Hepatol Res 42:482–93 Liu X, Mayes MD, Tan FK et al. (2013) Correlation of interferon-inducible chemokine plasma levels with disease severity in systemic sclerosis. Arthritis Rheum 65:226–35
8
Maere S, Heymans K, Kuiper M (2005) BiNGO: a cytoscape plugin to assess overrepresentation of gene ontology categories in biological networks. Bioinformatics 21:3448–9
Journal of Investigative Dermatology (2015), Volume 00
Rusinova I, Forster S, Yu S et al. (2013) Interferome v2.0: an updated database of annotated interferon-regulated genes. Nucleic Acids Res 41:D1040–6 Sargent JL, Milano A, Bhattacharyya S et al. (2010) A TGFbeta-responsive gene signature is associated with a subset of diffuse scleroderma with increased disease severity. J Invest Dermatol 130:694–705 Sato S (1999) Abnormalities of adhesion molecules and chemokines in scleroderma. Curr Opin Rheumatol 11:503–7 Skjot-Arkil H, Barascuk N, Register T et al. (2010) Macrophage-mediated proteolytic remodeling of the extracellular matrix in atherosclerosis results in neoepitopes: A potential new class of biochemical markers. Assay Drug Dev Technol 8:542–52 Skjot-Arkil H, Schett G, Zhang C et al. (2012) Investigation of two novel biochemical markers of inflammation, matrix metalloproteinase and cathepsin generated fragments of C-reactive protein, in patients with ankylosing spondylitis. Clin Exp Rheumatol 30:371–9 Solans R, Bosch JA, Esteban I et al. (2004) Systemic sclerosis developing in association with the use of interferon alpha therapy for chronic viral hepatitis. Clin Exp Rheumatol 22:625–8 Steen VD, Engel EE, Charley MR et al. (1996) Soluble serum interleukin 2 receptors in patients with systemic sclerosis. J Rheumatol 23:646–9 Streicher K, Morehouse CA, Groves CJ et al. (2014) The plasma cell signature in autoimmune disease. Arthritis Rheumatol 66:173–84 Trocme C, Leroy V, Sturm N et al. (2006) Longitudinal evaluation of a fibrosis index combining MMP-1 and PIIINP compared with MMP-9, TIMP-1 and hyaluronic acid in patients with chronic hepatitis C treated by interferonalpha and ribavirin. J Viral Hepat 13:643–51 Tusher VG, Tibshirani R, Chu G (2001) Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA 98: 5116–21 Valentini G, Della Rossa A, Bombardieri S et al. (2001) European multicentre study to define disease activity criteria for systemic sclerosis. II. identification of disease activity variables and development of preliminary activity indexes. Ann Rheum Dis 60:592–8 Veidal SS, Karsdal MA, Nawrocki A et al. (2011a) Assessment of proteolytic degradation of the basement membrane: a fragment of type IV collagen as a biochemical marker for liver fibrosis. Fibrogen Tissue Rep 4:22 Veidal SS, Karsdal MA, Vassiliadis E et al. (2011b) MMP mediated degradation of type VI collagen is highly associated with liver fibrosis—identification and validation of a novel biochemical marker assay. PLoS One 6:e24753. Veidal SS, Vassiliadis E, Bay-Jensen AC et al. (2010) Procollagen type I N-terminal propeptide (PINP) is a marker for fibrogenesis in bile duct ligation-induced fibrosis in rats. Fibrogenesis Tissue Repair 3:5 Yellin MJ, Winikoff S, Fortune SM et al. (1995) Ligation of CD40 on fibroblasts induces CD54 (ICAM-1) and CD106 (VCAM-1) up-regulation and IL-6 production and proliferation. J Leukoc Biol 58:209–16
