Elevated serum interferon-alpha activity in juvenile dermatomyositis: associations with disease activity at diagnosis and after thirty-six months of therapy

Abstract

Objective: Interferon-alpha (IFNalpha) has been implicated in the pathogenesis of juvenile dermatomyositis (DM). The aim of this study was to examine serum IFNalpha activity in a cohort of children with juvenile DM to determine relationships between IFNalpha and indicators of disease activity and severity.

Methods: Thirty-nine children with definite/probable juvenile DM were included in the study. Serum samples were obtained at the time of diagnosis from 18 untreated patients with juvenile DM. Second samples from 11 of these patients were obtained at 24 months, while they were receiving treatment, and third samples were obtained from 7 of these patients at 36 months. The remaining 21 children were studied 36 months after their initial diagnosis. Serum IFNalpha activity was measured using a functional reporter cell assay.

Results: Patients with juvenile DM had higher serum IFNalpha activity than both pediatric and adult healthy control subjects. In untreated patients, serum IFNalpha activity was positively correlated with serum muscle enzyme levels (P<0.05 for creatine kinase, aspartate aminotransferase, and aldolase) and inversely correlated with the duration of untreated disease (P=0.017). The tumor necrosis factor alpha -308A allele was associated with higher serum IFNalpha levels only in untreated patients (P=0.030). At 36 months, serum IFNalpha levels were inversely correlated with muscle enzyme levels in those patients still requiring therapy and with the skin Disease Activity Score in those patients who had completed therapy (P=0.002).

Conclusion: Serum IFNalpha activity was associated with higher serum levels of muscle-derived enzymes and a shorter duration of untreated disease in patients with newly diagnosed juvenile DM and was inversely correlated with measures of chronic disease activity at 36 months postdiagnosis. These data suggest that IFNalpha could play a role in disease initiation in juvenile DM.

Figure 1

Serum IFN-α activity in JDM patients as compared to controls, and blocking and rule out inhibitor experiments. A. shows serum IFN-α activity in untreated JDM patients (n=18) vs. pediatric healthy controls (n=19) and adult healthy controls (n=129). Each dot represents the IFN-α activity value for one subject (median values: JDM = 3.32, Controls≤18 = 0.29, Controls>18 = 0.43). B. shows blocking experiments in the first two columns, in which samples which had high IFN-α were pre-incubated with anti-IFN-α monoclonal antibody (1ug/mL) and then run in the WISH assay (median values: High IFN-α samples = 4.53, High IFN-α samples + anti-IFN-α = 0.45). Rule out inhibitor experiments are shown in columns 3, 4, and 5 in panel B. Samples which did not show significant IFN-α activity were pre-incubated with recombinant human IFN-α (50U/mL), and then run in the WISH assay. Addition of recombinant IFN-α to the sample resulted in the same IFN-α-induced gene activity in the reporter cells as cells cultured with media and the same amount of IFN-α (median values: Low IFN-α samples = 0.11, Low IFN-α samples + IFN-α = 44.83, Media + IFN-α = 40.25). Lines represent the median and interquartile range, p-values by Mann-Whitney t-test.

Figure 2

Serum IFN-α activity in JDM patients with different disease duration and treatment status. A. compares serum IFN-α activity in samples from patients who are receiving therapy (either 24 or 36 mos after diagnosis) vs. those who are not receiving therapy (either initial untreated sample or those at 36 months who were off therapy). B. shows serum IFN-α in each separate disease/treatment category. Lines represent the median and interquartile range, all comparisons are non-significant, with p-values >0.25 by Mann-Whitney t-test.

Figure 3

Correlation plots showing significant relationships between serum IFN-α activity and clinical variables. A. shows correlation plots for untreated patients between serum IFN-α activity and CPK, AST, aldolase, and duration of untreated disease respectively. B. shows correlation plots for patients still receiving treatment at 36 months between serum IFN-α activity, aldolase and LDH. C. shows correlation plot for patients who were off treatment at 36 months between serum IFN-α activity and skin DAS. Plots are in semi-log format, Y-axis is linear and X-axis is log, best fit semi-log line is shown on each plot.

Figure 4

Serum IFN-α activity in JDM patients stratified by disease duration/treatment status and TNF-α and HLA*DQA 0501 genotypes. A. shows serum IFN-α activity in JDM patients stratified by disease duration/treatment status and TNF-α-308 promoter polymorphism genotype. A- = AA or GA genotype, A is the risk allele. B. shows serum IFN-α activity in JDM patients stratified by disease duration/treatment status and HLA DQA*0501 genotype. Positive indicates a subject carrying the risk allele of HLA DQA*0501, and negative indicates a non-risk allele genotype. 24 month on treatment and 36 month on treatment groups are combined into one group for both of these analyses, and the off treatment group consists of subjects who are 36 months from initial diagnosis and able to successfully discontinue therapy. Lines represent the median and interquartile range, p-values by Mann-Whitney t-test.

Figure 5

Longitudinal data for serum IFN-α, total DAS, and aldolase in children with JDM who had sera available at all time points. Serum IFN-α, total DAS, and aldolase (A, B, and C respectively) are plotted against time in months from diagnosis in seven patients with samples available at these time points. Regression lines and analyses are performed with semi-log (A, log distributed IFN-α data on Y-axis, linear time on X-axis) or linear (B. and C.) models, with p-values calculated from Spearman’s rho.