SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Objective

Up-regulation of whole blood type I interferon (IFN)–driven transcripts and chemokines has been described in a number of autoimmune diseases. An IFN gene expression “signature” is a candidate biomarker in patients with dermatomyositis (DM). This study was performed to evaluate the capacity of IFN-dependent peripheral blood gene and chemokine signatures and levels of proinflammatory cytokines to serve as biomarkers for disease activity in adult and juvenile DM.

Methods

Peripheral blood samples and clinical data were obtained from 56 patients with adult or juvenile DM. The type I IFN gene signature in the whole blood of patients with DM was defined by determining the expression levels of 3 IFN-regulated genes (IFIT1, G1P2, and IRF7) using quantitative real-time reverse transcription–polymerase chain reaction. Multiplexed immunoassays were used to quantify the serum levels of 4 type I IFN–regulated chemokines (IFN-inducible T cell α chemoattractant, IFNγ-inducible 10-kd protein, monocyte chemotactic protein 1 [MCP-1], and MCP-2) and the serum levels of other proinflammatory cytokines, including interleukin-6 (IL-6).

Results

DM disease activity correlated significantly with the type I IFN gene signature (r = 0.41, P = 0.007) and with the type I IFN chemokine signature (r = 0.61, P < 0.0001). Furthermore, the serum levels of IL-6 were significantly correlated with disease activity (r = 0.45, P = 0.001). In addition, correlations between the serum levels of IL-6 and both the type I IFN gene signature (r = 0.47, P < 0.01) and the type I IFN chemokine signature (r = 0.71, P < 0.0001) were detected in patients with DM.

Conclusion

These results suggest that serum IL-6 production and the type I IFN gene signature are candidate biomarkers for disease activity in adult and juvenile DM. Coregulation of the expression of IFN-driven chemokines and IL-6 suggests a novel pathogenic linkage in DM.

Adult and juvenile dermatomyositis (DM) are autoimmune disorders characterized by proximal muscle weakness, muscle inflammation, and hyperkeratotic skin rash. Current pathogenetic models suggest that a combination of genetic and environmental factors confers susceptibility to disease (1). Susceptibility to DM is associated with inheritance of the HLA alleles DQA1*0501 (2) and DR3 (3), and polymorphisms in the tumor necrosis factor α (TNFα) gene are associated with disease severity (4). Viral infection has long been suspected to play a causative role in DM (5). Despite advances in the understanding of pathogenetic factors in DM, monitoring of disease activity is heavily dependent on the physician's clinical assessment. Few reliable indicators of prognosis, disease activity, or potential response to treatment have been identified.

Accumulating data from our group and other investigators suggest that cells from the muscle tissue and blood of patients with DM carry distinct immune “fingerprints.” Several studies have demonstrated up-regulation of genes related to antigen processing (6, 7), immunoglobulin-encoding genes (7), and type I interferon-α/β (IFNα/β)–regulated genes in the muscle tissue of patients with DM (6, 8). Type I IFN–inducible proteins (IFNα/β-inducible myxovirus resistance protein [MxA] and IFN-stimulated transcription factor 15 [ISGF-15], and IFN regulatory factor 7 [IRF-7]) were detected in the muscle fibers and capillaries of patients with adult or juvenile DM (8–10). Numerous type I IFN–inducible genes are up-regulated in the peripheral blood, and these form a type I IFN gene “signature” in DM (11, 12). In addition, we have observed elevated serum levels of IFN-inducible chemokines in patients with DM (11). In aggregate, these data documenting type I IFN induction at both the transcriptional and translational levels suggest a major pathogenetic role for type I IFNs in DM.

Histopathologic studies have revealed that both lymphoid and myeloid immune cell infiltrates are present in the muscles of patients with DM. Perivascular and perifascicular infiltrates display B cell and CD4+ T cell predominance in adult DM, and display a combination of B, CD4+, and CD8+ T cells in juvenile DM (13), suggesting that muscle tissue may be a target for both humoral- and cell-mediated autoimmune responses. The functions of both the CD8+ and CD4+ T cell lineages are likely important for causing inflammation in the muscle and other tissues in DM. A number of distinct CD4+ T cell subsets, including Th1, Th2, and Treg, may regulate human autoimmune processes.

The presence of inflammatory cells in affected tissue in DM suggests a potential for regulation of the disease process by soluble cytokine networks. Recent work in a mouse model of inflammatory myositis identified the cytokine interleukin-6 (IL-6) as a critical mediator of tissue-destructive processes (14). IL-6 is a proinflammatory cytokine that has been targeted therapeutically for human rheumatic disease. IL-6 plays central roles in the regulation of both innate and adaptive inflammatory and immune responses, as well as both humoral and cell-mediated autoimmune reactions. IL-6 is a multifunctional cytokine that was originally identified as a B cell differentiation factor. IL-6 is produced by various types of lymphoid and nonlymphoid cells, such as T cells, B cells, monocytes, fibroblasts, keratinocytes, endothelial cells, mesangial cells, and several tumor cells. In addition to B cell differentiation activity, IL-6 induces T cell growth and T cell differentiation (15).

Elevated levels of IL-6 have been documented in a variety of rheumatic diseases, including Castleman's disease, primary Sjögren's syndrome, rheumatoid arthritis, colitis, and Crohn's disease (16–19). Blockade of IL-6 and of IL-6 signaling have been shown to be effective in treating several of these inflammatory diseases (e.g., inflammatory arthritis and colitis) (19).

We tested the hypotheses that multiple proinflammatory pathways are active in DM, and that detection of peripheral blood chemokines and proinflammatory cytokines marking these pathways will permit improved disease activity assessments. Using genomic and proteomic approaches, we observed robust activation of the IFNα pathway in the peripheral blood of patients with DM. Prominent type I IFN signatures, manifesting as both transcript up-regulation and elevated levels of serum proteins, correlated strongly with DM disease activity. In addition, elevated IL-6 levels in the patients' serum correlated strongly with both DM disease activity and IFN-driven chemokine levels. These data suggest that genes and proteins in the type I IFN– and IL-6–related pathways are coregulated in DM, and that monitoring dysregulation within these pathways will help guide diagnostic and treatment decisions.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Study subjects and sample collection.

The study protocol was approved by the Human Subjects Institutional Review Boards at the University of Minnesota and the Mayo Clinic, and informed consent was obtained from each participant. Study participants were recruited among attendees at myositis clinics at Mayo Clinic; all patients met the Bohan and Peter criteria for adult DM (n = 37) or juvenile DM (n = 19) (20). There were no inclusion criteria related to disease activity, and therefore all patients who were willing to participate were enrolled. The potential for overlap syndromes (e.g., mixed connective tissue disease–systemic lupus erythematosus [SLE]) was addressed by conducting clinical and laboratory evaluations of the patients during enrollment. Individuals with features of an overlap syndrome were excluded from the study. All blood biomarker data were obtained from patients at the time of study enrollment.

Twenty healthy subjects (mean ± SD age 40 ± 9 years) served as controls for some of the analyses. All control subjects were older than age 18 years. In previous studies of healthy pediatric subjects, IFN-driven gene expression and serum levels of IL-6, IL-8, IL-10, and TNFα were documented (21, 22), and the levels of these analytes were comparable with those obtained in the healthy adult control subjects in the present study.

Disease activity was assessed by study rheumatologists (SRY, SA, AMR) using established disease activity tools for use in myositis clinical trials, as described by the International Myositis Assessment and Clinical Studies (IMACS) Group (23–25). Under the IMACS guidelines, muscle strength was assessed using manual muscle strength testing of 8 muscle groups, to obtain the MMT8 score. Strength was rated on a 10-point scale for each group, yielding a maximal score of 80. Muscle groups tested included the neck flexors and the following dominant-sided muscle groups: shoulder abductors, elbow flexors, wrist dorsiflexors, hip extensors, hip abductors, knee extensors, and ankle dorsiflexors. To assess disease activity based on individual organ systems, the Myositis Disease Activity (MYOACT) portion of the Myositis Disease Activity Assessment Tool was used. The MYOACT assessment utilizes separate 100-mm visual analog scales (VAS) to gauge the physician's evaluation of disease activity in several discrete domains. Involvement of all nonmuscle organ systems (constitutional, cardiac, pulmonary, gastrointestinal, skeletal, and cutaneous) was also evaluated using the composite extraskeletal muscle VAS score. An additional VAS measure, the global VAS score, was used to rate overall disease activity.

Serum levels of creatine kinase (CK), aldolase, aspartate aminotransferase (AST), alanine aminotransferase (ALT), and lactate dehydrogenase (LDH) as well as positivity for antinuclear antibodies (ANAs) and anti–Jo-1 antibodies, the erythrocyte sedimentation rate (ESR), and the C-reactive protein (CRP) level were assayed in the clinical laboratory.

Gene expression measurements.

Samples of whole blood were obtained from subjects and drawn into PAXgene tubes (Qiagen/Becton Dickinson, Franklin Lakes, NJ). Total RNA was isolated from the blood with on-column DNase treatment, according to the manufacturer's protocol. The RNA yield and integrity were assessed using an Agilent Lab-on-a-Chip Bioanalyzer (Agilent Technologies, Palo Alto, CA). The type I IFN gene expression signature was defined in the whole blood by determining the expression levels of 3 IFN-regulated genes (IFIT1, G1P2, and IRF7), as measured by TaqMan quantitative real-time reverse transcription–polymerase chain reaction using the ABI Prism 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA). Relative quantification of the gene expression levels was performed by comparison of the values against a calibrator sample (PAXgene whole blood RNA from a healthy control subject), in accordance with the manufacturer's guidelines, and the results were normalized to the values for GAPDH. For each gene, the 95th percentile of expression levels was calculated. Expression values equal to or greater than the 95th percentile were replaced with the 95th percentile value and then normalized, so that the maximum value for each gene was 1.0 (26). Finally, the normalized expression values for the 3 genes were then summed for each patient, and the sums were adjusted to a 100-point scale, to yield the summary type I IFN gene score.

Serum protein measurements.

Serum was isolated from the blood samples in serum-separator vacutainer tubes (Qiagen/Becton Dickinson). A protease inhibitor (1 μg/ml aprotinin) was added to each sample, and aliquots were immediately frozen at −80°C. Multiplexed sandwich immunoassays (Meso Scale Discovery, Gaithersburg, MD) were used to quantify the serum levels of IFN-regulated chemokines (27) and other proinflammatory cytokines (monokine induced by IFNγ [Mig], macrophage inflammatory protein 1α [MIP-1α], MIP-1β, TNFα, TNF receptor I [TNFRI], IL-10, IL-6, and IL-8). Samples were run in duplicate, and calibrated recombinant proteins were used to generate standard curves. The summary chemokine score (type I IFN chemokine signature) was calculated for each participant in a manner similar to the calculation of the type I IFN gene score. The summary chemokine score was based on the serum levels of IFN-inducible T cell α chemoattractant (I-TAC), IFNγ-inducible 10-kd protein (IP-10), monocyte chemotactic protein 1 (MCP-1), and MCP-2. First, the 95th percentile of serum concentration levels was calculated for each chemokine, and concentration levels equal to or greater than the 95th percentile were replaced with the 95th percentile value (26). Then, for each chemokine, the data were normalized to a maximum value of 1.0. For each participant, the normalized values for the 4 proteins were summed up, and the sum was adjusted to a 100-point scale (26).

Statistical analysis.

Spearman's rank correlation coefficients were calculated to assess the relationship between biomarkers and continuous variables (i.e., disease activity scores and blood biochemical measurements). Group comparisons were performed using Mann-Whitney unpaired tests. Significance levels for both tests were set at P less than or equal to 0.05. P values between 0.05 and 0.01 were defined as indicating modest significance, while P values smaller than 0.01 were deemed to indicate strong significance.

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Correlation of type I IFN–regulated gene expression with DM disease activity.

To begin the search for sensitive and reliable biomarkers in DM, we used established clinical scales to assess disease activity in a cohort of 56 patients with DM (37 with adult DM and 19 with juvenile DM). The demographic and clinical characteristics of the patients with DM are summarized in Table 1. Of the patients with DM, 57% had abnormal (depressed) MMT8 scores, and 78% had cutaneous manifestations, indicating a high prevalence of objective muscle weakness and active disease in the cohort (Table 1). None of the study patients was thought to have a myositis overlap syndrome (e.g., SLE/scleroderma) as defined according to the clinical and/or serologic criteria for the diagnosis of these disorders.

Table 1. Demographic and clinical characteristics of the patient population*
 Adult DM (n = 37)Juvenile DM (n = 19)All patients (n = 56)
  • *

    Except where indicated otherwise, values are the mean ± SD results in patients with adult dermatomyositis (DM) or those with juvenile DM. VAS = visual analog scale; MMT8 = manual muscle strength testing of 8 muscle groups; ANAs = antinuclear antibodies.

  • Patients treated with multiple medications are included in the total numbers for each agent.

Age, years59 ± 15.913.4 ± 6.942.1 ± 24.7
No. female/no. male31/612/743/13
Race, no. (%)   
 Caucasian30 (81)14 (74)44 (79)
 Other7 (19)5 (26)12 (21)
Global 100-mm VAS score (no. tested)28.4 ± 25.7 (35)28.1 ± 27.9 (17)28.3 ± 26.2 (52)
Subscale VAS score, no. scoring >0/no. tested (%)   
 Constitutional16/36 (44)5/18 (28)21/54 (39)
 Cutaneous28/36 (78)14/18 (78)42/54 (78)
 Skeletal5/36 (14)5/18 (28)10/54 (19)
 Gastrointestinal10/36 (28)4/18 (22)14/54 (26)
 Pulmonary16/36 (44)2/18 (11)19/54 (35)
 Cardiac3/35 (9)1/18 (6)4/53 (8)
 Composite extraskeletal muscle29/36 (81)14/18 (78)43/54 (80)
 Muscle25/36 (69)6/18 (33)31/54 (57)
MMT8 score, no. scoring <80/no. tested (%)25/34 (74)4/17 (24)29/51 (57)
Antibodies, no. positive/no. tested (%)   
 Anti–Jo-12/30 (7)2/6 (33)4/36 (11)
 ANAs19/34 (56)9/18 (50)28/52 (54)
Disease duration, no. (%)   
 ≤2 years18 (49)10 (53)28 (50)
 2–5 years10 (27)7 (37)17 (30)
 ≥5 years9 (24)2 (10)11 (20)
Medication, no. of patients   
 Prednisone21829
 Methotrexate141024
 Azathioprine707
 Mycophenolate mofetil235
 None8613

Elevated levels of type I IFN–driven genes have been observed to roughly correlate with disease activity in a small set of blood samples from patients with DM (11). We therefore measured the levels of type I IFN–regulated transcripts in whole blood samples obtained from patients with DM. Whole blood levels of 3 type I IFN–regulated transcripts, IFIT1, G1P2, and IRF7, were used to create a composite type I IFN gene score for each patient. The 3 gene-based type I IFN gene scores in the present cohort correlated highly with the microarray-defined full IFN gene signature described in the previous cohort (11) (r = 0.95, P = 2.3 × 10−7). Expression levels of IFIT1, G1P2, and IRF7 were highly correlated with one another (results not shown) and were also correlated with key study analytes (global VAS score, serum IL-6 levels, and type I IFN chemokine score) in the DM cohort (results available from the corresponding author upon request).

The median type I IFN gene score was significantly elevated in patients with DM as compared with healthy controls (Table 2). The type I IFN gene signature was also strongly correlated with disease activity as measured by the physician's global VAS score in patients with DM (r = 0.41, P = 0.007) (Figure 1A). Taken together, these data confirm and extend our previous observations showing that type I IFN–driven genes are up-regulated in the peripheral blood of patients with DM and that the type I IFN gene score correlates with clinical measures of disease activity.

Table 2. Correlations of the expression of type I interferon (IFN)–regulated genes and chemokines with adult or juvenile dermatomyositis (DM)*
 Healthy controls (n = 20)Adult DM (n = 37)Juvenile DM (n = 19)
Median95% CIMedian95% CIP vs. controlsMedian95% CIP vs. controls
  • *

    95% CI = 95% confidence interval; IP-10 = IFNγ-inducible 10-kd protein; I-TAC = IFN-inducible T cell α chemoattractant; MCP-1 = monocyte chemotactic protein 1; Mig = monokine induced by IFNγ; MIP-1α = macrophage inflammatory protein 1α; IL-6 = interleukin-6; TNFα = tumor necrosis factor α.

  • By Mann-Whitney test.

  • Calculated on the basis of 11 samples.

  • §

    Levels of MCP-2 and TNF receptor I were not measured in healthy controls, and therefore are not included. Protein levels are expressed in pg/ml.

IFN gene score5.22.6–9.411.65.9–34.50.00614.96.4–40.50.002
Individual chemokines or cytokine§        
 IP-1031.927.1–36.8317125–851<0.0001529147–1,461<0.0001
 I-TAC24.219.1–32.513563.1–252<0.000192.160.6–482<0.0001
 MCP-1263219–331595440–982<0.0001591531–942<0.0001
 Mig15.712.0–18.789.149.9–115<0.000157.620.3–1130.0001
 MIP-1α41.333.7–47.1212151–289<0.0001233178–465<0.0001
 MIP-1β148134–169162141–2060.3202133–2630.09
 IL-61.21.1–1.53.62.2–5.3<0.00014.62.1–10.1<0.0001
 IL-81511.8–19.17.65.4–13.10.00088.26.0–16.90.02
 IL-102.31.6–3.91.61.2–2.90.052.91.3–5.20.9
 TNFα4.84.5–5.33.32.5–4.90.015.63.5–7.80.5
thumbnail image

Figure 1. Correlations of the type I interferon (IFN) gene signature (A), type I IFN protein (chemokine) signature (B), and serum interleukin-6 (IL-6) levels (C) with disease activity in patients with adult or juvenile dermatomyositis (DM). Disease activity was measured by the global visual analog scale (VAS) score in 52 of 56 patients with DM. Spearman's rank correlation coefficients were used to assess correlations with myositis disease activity.

Download figure to PowerPoint

We next investigated associations between the type I IFN gene score and specific clinical manifestations of DM. The type I IFN gene score was strongly correlated with constitutional (r = 0.33, P = 0.03), cutaneous (r = 0.42, P = 0.006), composite extraskeletal muscle (r = 0.35, P = 0.02), and muscle (r = 0.47, P = 0.002) VAS subscales, as well as with the MMT8 score (r = −0.48, P = 0.002) (detailed results available from the corresponding author upon request). In contrast, the type I IFN gene score did not correlate significantly with inflammation indices such as the ESR or the CRP level, nor did it correlate with most of the blood-borne indices of muscle injury (levels of CK, ALT, LDH, or aldolase), with the exception of an association with the levels of AST (results available from the corresponding author upon request). Furthermore, no correlation was observed between the type I IFN gene score and the presence of ANAs or anti–Jo-1 antibodies (results not shown). Taken together, these findings suggest that blood levels of type I IFN–driven transcripts correlate positively with some of the DM disease activity subscales, but show minimal correlation with biochemical indicators of systemic inflammation or muscle injury.

Role of type I IFN–regulated serum chemokine levels as a sensitive biomarker of DM disease activity.

We have previously shown that the serum levels of type I IFN–regulated chemokines are elevated in patients with DM and that chemokine levels are correlated with the type I IFN gene signature (11). In the present cohort, we again observed that the serum levels of type I IFN–driven chemokines (IP-10, I-TAC, and MCP-1) were increased both in the adult DM cohort and in the juvenile DM cohort compared with the levels in healthy controls (Table 2).

To determine whether these IFN-driven protein levels are also correlated with the extent of disease activity in DM, we measured the serum levels of 4 type I IFN–regulated chemokines (I-TAC, IP-10, MCP-1, and MCP-2) in 56 patients with DM and assessed them for correlations with the physician's global assessment of disease activity (global VAS score). Levels of individual chemokines were each strongly correlated with the global VAS score (P < 0.0001 for each) (Table 3). An even stronger correlation (r = 0.61, P < 0.0001) was observed between the type I IFN chemokine score (summation of normalized levels of the 4 chemokines) and the global VAS score (Figure 1B). Similar correlations between the global VAS score and type I IFN chemokine score were observed when the adult and juvenile DM patient groups were evaluated separately (adult DM, r = 0.690, P = 0.0001; juvenile DM, r = 0.532, P = 0.03). Interestingly, when correlations between the global VAS score and the type I IFN gene score were assessed, a significant correlation was observed in the adult DM cohort only (P = 0.003). The type I IFN gene score was strongly correlated with the type I IFN chemokine score in the pooled adult and juvenile DM population (r = 0.53, P = 0.0003). Furthermore, the type I IFN chemokine score was strongly correlated with muscle-specific disease activity indicators (muscle VAS score, r = 0.47, P = 0.0006; MMT8 score, r = −0.44, P = 0.002) (more details available from the corresponding author upon request). The levels of MIP-1α, a type I IFN–inducible chemokine, were also elevated in DM sera, and this was correlated with DM disease activity measured by global VAS score (Tables 2 and 3).

Table 3. Correlations between global VAS score and candidate blood biomarkers in patients with adult or juvenile dermatomyositis*
 Global VAS score
rP
  • *

    Correlations were determined using Spearman's rank correlation coefficients. VAS = visual analog scale; TNFRI = tumor necrosis factor receptor I (see Table 2 for other definitions).

  • By Mann-Whitney test.

  • Component of the type I IFN chemokine score.

IFN gene score0.410.007
IFN chemokine score0.61<0.0001
Individual chemokines or cytokine  
 IP-100.66<0.0001
 I-TAC0.62<0.0001
 MCP-10.470.0008
 MCP-20.530.0003
 Mig0.160.25
 MIP-1α0.560.0001
 MIP-1β0.160.25
 IL-60.450.001
 IL-80.190.17
 IL-100.280.047
 TNFα0.330.023
 TNFRI0.310.045

We also observed that a significant fraction of patients with a global VAS score higher than 10 exhibited IFN chemokine scores lower than 20. We sought to determine whether this VAShigh/IFNlow profile represents a distinct clinical subset. When comparing VAShigh subjects displaying serum IFN chemokine scores >20 (VAShigh/IFNhigh) with those displaying chemokine scores <20 (VAShigh/IFNlow), we found no significant differences in the demographic characteristics, disease duration, most of the subscale disease scores, medication use, or laboratory findings between the 2 groups; however, IL-6 levels were significantly higher (P ≤ 0.05) in the VAShigh/IFNhigh group (results not shown).

DM diagnosis and/or disease activity scores showed varying associations with the levels of other inflammatory chemokines and cytokines. Mig, an IFN-driven chemokine, showed increased levels in the sera of patients with DM (Table 2), but this did not correlate with the extent of disease activity (Table 3). IL-8 levels were lower in patients with DM compared with healthy controls, but this did not correlate with disease activity. IL-10 levels were lower in the adult DM cohort, but not in the juvenile DM cohort, in comparison with the levels in healthy controls, and this showed a modest correlation with DM disease activity. Modest correlations between DM disease activity and serum levels of TNFα and TNFRI were also found (Table 3). With the exception of IL-10 and TNFα, we observed no differences in the levels of any other chemokines or cytokines between the adult DM and juvenile DM groups.

The type I IFN chemokine score was also modestly or strongly correlated with VAS scores for several specific manifestations of DM, including constitutional (r = 0.31, P = 0.03), cutaneous (r = 0.30, P = 0.03), gastrointestinal (r = 0.34, P = 0.02), skeletal (r = 0.38, P = 0.006), composite extraskeletal muscle (r = 0.43, P = 0.002), and muscle (r = 0.43, P = 0.01) VAS subscales (results available from the corresponding author upon request). Unlike the type I IFN gene score, the type I IFN chemokine score was modestly correlated both with indices of overall inflammation (ESR) and with muscle injury as assessed by the levels of AST, ALT, and aldolase (results available from the corresponding author upon request). Taken together, these data indicate that serum levels of type I IFN–regulated chemokines are strongly correlated both with the expression levels of type I IFN–driven genes and with the extent of DM disease activity.

Correlation between myositis IL-6 serum levels and DM disease activity.

We next investigated whether serum markers of leukocyte function hold promise as candidate biomarkers for DM disease activity. Serum levels of 8 proteins implicated in T cell– and myeloid cell–dependent inflammatory processes were compared in serum samples from patients with DM and 20 healthy controls. Serum levels of several cytokines showed strong positive correlations with myositis disease activity (Table 3). Intriguingly, the levels of IL-6 were highly correlated with disease activity (r = 0.45, P = 0.001). These data suggest that IL-6 is a candidate biomarker for disease activity in adult and juvenile DM.

Coregulation of IL-6 and type I IFN–driven chemokines in DM.

Since the serum levels of type I IFN–regulated chemokines and the serum levels of IL-6 are each highly correlated with DM disease activity, we next investigated whether these 2 pathways may be coregulated in DM. Using an unsupervised hierarchical clustering analysis, we identified a tight cluster of proteins displaying similar patterns of regulation, which included several type I IFN–regulated chemokines (I-TAC, IP-10, MCP-1, MCP-2, and MIP-1α) as well as IL-6 (Figure 2). Serum levels of IL-6 were significantly correlated with both the type I IFN gene score and the type I IFN chemokine score (Figure 3). Particularly striking were the strong correlations between the type I IFN chemokine score and IL-6 levels (r = 0.71, P < 0.0001). Interestingly, the levels of other cytokines that correlated with disease activity (IL-10, TNFα, and TNFRI) did not cluster within the IFN/IL-6 group (Figure 2). Taken together, these data suggest that specific coregulation of IFN-driven protein levels and IL-6 production occurs in many patients with DM.

thumbnail image

Figure 2. Hierarchical clustering analysis of serum levels of 12 proteins (rows) in 52 of 56 patients with adult or juvenile DM (columns), showing coregulation of IL-6 (in blue) and type I IFN–induced chemokines (macrophage inflammatory protein 1α [MIP-1α], IFN-inducible T cell α chemoattractant [I-TAC], IFNγ-inducible 10-kd protein [IP-10], monocyte chemotactic protein 1 [MCP-1], and MCP-2) (in red). Expression data are shown as the log2 ratios relative to the mean value in patients with inactive disease (those having a global VAS score of 0 [n = 9]). The color gradient key reflects the relative expression on a log2 scale (range of linear fold change, −8 to +8). TNFα = tumor necrosis factor α; TNFR-1 = tumor necrosis factor receptor I; MIG = monokine induced by IFNγ (see Figure 1 for other definitions).

Download figure to PowerPoint

thumbnail image

Figure 3. Correlations of serum IL-6 levels with clinical activity (A) and with the type I IFN signature (B) in 54 of 56 patients with adult DM and those with juvenile DM (JDM). A, Measures of disease severity (rows) in patients with DM (columns) included the global VAS score and VAS subscores for specific manifestations of DM, including muscle strength and composite extraskeletal (ES) muscle scores, as well as VAS scores for cardiac, pulmonary, gastrointestinal, skeletal, and cutaneous manifestations. B, The scores for the expression levels of type I IFN–regulated genes and chemokines (I-TAC, IP-10, MCP-1, and MCP-2) (rows) in patients with DM (columns) are shown as the log2 ratios relative to the mean value in patients with inactive disease (those having a global VAS score of 0 [n = 9]). The color gradient key reflects the relative expression on a log2 scale (range of linear fold change, −8 to +8). Spearman's rank correlation coefficients were used to assess correlations between IL-6 levels and each of the other features. Gray bars indicate missing data. ∗ = P ≤ 0.05; ∗∗ = P ≤ 0.01; ∗∗∗ = P ≤ 0.001; ∗∗∗∗ = P ≤ 0.0001. See Figure 1 for other definitions.

Download figure to PowerPoint

Association of IL-6 levels with specific DM disease manifestations.

Since the serum levels of IL-6 showed a strong correlation with the extent of DM disease activity, we next investigated the correlation between this cytokine and other indicators of disease activity. IL-6 levels were modestly or strongly correlated with VAS scores for several individual disease manifestations, including constitutional, cutaneous, skeletal, gastrointestinal, composite extraskeletal muscle involvement, and muscle strength (Figure 3, and results available from the corresponding author upon request), and with MMT8 scores (r = −0.39, P = 0.007). Serum IL-6 levels were also modestly correlated with the ESR and levels of AST and ALT (results available from the corresponding author upon request). These results suggest that, similar to the levels of IFN-driven chemokines, serum IL-6 levels may serve as an indicator of activity within key DM clinical domains such as the muscle and skin.

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Our results demonstrate that the levels of IL-6 and type I IFN–regulated transcripts and proteins are elevated in the peripheral blood of patients with adult or juvenile DM. Type I IFN gene and protein signatures and serum IL-6 levels are correlated significantly with indicators of DM disease activity and with each other. Our results suggest the possibility that coordinated dysregulation of type I IFN signaling and IL-6 production may be a characteristic pathogenic feature of DM.

Dysregulated type I IFN has also been implicated in the pathogenesis of other autoimmune diseases, including human SLE (28, 29), Sjögren's syndrome (30), and scleroderma (31). A key aspect of the current model for IFN production in autoimmunity is the triggering of IFN production after ligation, by viral and bacterial nucleic acids, of Toll-like receptors (TLR-7 and TLR-9) in plasmacytoid dendritic cells (DCs) (32, 33). Plasmacytoid DCs comprise a DC subset that has important functions in the regulation of innate and adaptive immune responses, particularly those in response to viral infections. Plasmacytoid DCs are relatively infrequent in the peripheral blood, but they are found in the lesional skin of patients with SLE (34).

Both plasmacytoid DCs and type I IFN–induced proteins (MxA, ISGF-15, and IRF-7) are found in the affected muscle and skin of patients with DM (10, 35–37), suggesting a potential pathophysiologic function for plasmacytoid DCs. Indeed, CD4+ plasmacytoid DCs comprise a large proportion (30–90%) of what had previously been thought to be CD4+ T cells in DM muscle (8, 37). Viral infection has long been suspected as an inciting factor in DM (5), but the specific pathogen-derived TLR ligands that might signal elaboration of IFN by plasmacytoid DCs remain unknown.

Type I IFNs have pleiotropic functions that may contribute to the pathogenesis of DM. Type I IFNs up-regulate class I major histocompatibility complex (MHC) expression, activate natural killer cell cytotoxicity, promote activated T cell survival, and support DC maturation (36). The regulatory effect of IFN on class I MHC expression may be particularly relevant in human DM, since class I MHC levels are elevated in the muscle cells of patients with DM (36). Furthermore, animals with enforced muscle-specific overexpression of class I molecules exhibit inflammation and muscle necrosis (38). IFN-regulated proteins have been shown to play a role in recruitment of lymphocytes to sites of inflammation in the muscle (39, 40) and skin (41). These proteins (IP-10, I-TAC, MCP-1, and MCP-2) were highly expressed in our DM cohort, and this comprised the type I IFN chemokine score.

Not all patients with active DM according to the clinical disease activity score displayed significant elevation in the type I IFN chemokine score (Figure 1B). In fact, we observed a significantly large population of patients in the VAShigh/IFNlow subset, but saw a near absence of patients in the converse VASlow/IFNhigh population. This trend may suggest that 1) IFN gene and chemokine scores are less sensitive than the global VAS score for detecting clinically important disease, 2) the global VAS score overestimates disease activity in DM, 3) a subset of DM disease is driven by an IFN-independent pathophysiologic mechanism, or 4) treatment may have been recently initiated. The level of IL-6 was also elevated in those with the VAShigh/IFNhigh profile (details available from the corresponding author upon request), suggesting that activation of IL-6 and type I IFN may describe a clinical subset of patients with DM whose disease has a unique pathophysiologic pathway. Distinguishing between these possibilities and establishing the degree to which IFN-driven events may serve as disease activity biomarkers for the DM patient population as a whole will require longitudinal studies that prospectively evaluate correlations between disease flare and type I IFN gene or chemokine score elevation.

IL-6 is implicated in the regulation of both innate and adaptive immune processes in a number of human autoinflammatory and autoimmune diseases (16–19). Elevated serum IL-6 levels are found in Castleman's disease, a rare disorder characterized by fevers, anorexia, anemia, hypergammaglobulinemia, and plasma cell accumulation in enlarged lymph nodes (18). In rheumatoid arthritis, IL-6 levels are elevated in both the serum and synovial fluid (42). IL-6 is a major contributor both to the constitutional symptoms of fatigue and fever, and, via activation of osteoclasts, to the bony erosions that characterize the rheumatoid joint. Antagonism of IL-6 function through genetic or pharmacologic manipulation has been shown to be beneficial in the treatment of experimental arthritis in murine models, as well as in the treatment of human Castleman's disease, colitis, rheumatoid arthritis, and systemic-onset juvenile arthritis (15, 17, 18).

Evidence in the literature concerning a pathogenic role for IL-6 in inflammatory myositis is sparse. In the experimental mouse model of myosin-induced myositis, deficiency of IL-6 leads to marked amelioration of the clinical signs and pathologic findings of muscle injury (14). However, there are no descriptions of IL-6 overexpression at the level of skeletal muscle in humans with DM.

Few reports to date have suggested the utility of IL-6 as a biomarker for disease activity in the idiopathic inflammatory myopathies (IIMs), including DM. Interestingly, Tucci et al reported no significant differences in the serum levels of IL-6 between controls and patients with active IIM (43). The contrast between the latter results and the present findings of elevated IL-6 levels may be due to the more limited, yet more diverse patient population in that study (n = 8 patients, comprising both those with polymyositis and those with DM). Our findings suggest that IL-6 levels are positively correlated with overall disease activity and with specific manifestations of DM, as well as with the levels of type I IFN–regulated chemokines (Figure 3). The latter finding suggests that there is coregulation of the type I IFN– and IL-6–related pathways, and possibly a regulatory cross-talk between the pathways. Notably, IL-6 and MCP-1, a type I IFN–regulated chemokine, have both been shown to be potent inducers of inflammation, acting through mononuclear cell accumulation at a site of tissue injury (19), and both are coregulated by the same transcription factors (44). Whether the level of IL-6 is a sufficiently sensitive clinical marker for current or impending disease flare is a question that must be addressed by longitudinal, prospective studies.

The majority of patients with DM in this study were receiving treatment with antimetabolite, antiproliferation, or corticosteroid agents at the time of enrollment. We observed a nonsignificant trend toward both a lower global VAS score and lower levels of candidate biomarker analytes (the type I IFN chemokine score) in the treatment groups compared with those not taking immunosuppressive medication; for example, IL-6 levels were lower in the treated group compared with the untreated group (median 3.32 pg/ml versus 8.05 pg/ml; P = 0.04). Previously, Walsh et al reported a marked reduction in type 1 IFN signature levels in patients with DM after treatment with prednisone, intravenous immunoglobulin, mycophenolate mofetil, methotrexate, or azathioprine (12). Although the low sample size does not allow definite conclusions, we speculate that the higher disease activity and higher values for laboratory parameters among the untreated subjects were due to a high prevalence of new-onset, treatment-naive disease in that group. However, the cross-sectional design of the present study did not allow an assessment of analyte fluctuation in response to treatment in individual subjects over time. Longitudinal studies are required to elucidate the exact relationship between immunosuppressive agents, disease activity, and the expression of candidate biomarkers in DM.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Peterson had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Bilgic, Ytterberg, McNallan, Koeuth, Bauer, Peterson, Baechler, Reed.

Acquisition of data. Bilgic, Ytterberg, Amin, McNallan, Wilson, Koeuth, Ellingson, Newman, Bauer, Peterson, Reed.

Analysis and interpretation of data. Bilgic, Koeuth, Newman, Bauer, Peterson, Baechler, Reed.

Acknowledgements

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

We thank the Minnesota Partnership for Biotechnology and Medical Genomics for providing the funding for these studies.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES