Decreased interferon‐β induced STAT‐4 activation in immune cells and clinical outcome in multiple sclerosis

Interferon‐β (IFN‐β) is used in the treatment of multiple sclerosis (MS). IFN‐β activation of signal transduction and activation of transcription (STAT)‐4 is linked to its immunomodulatory effects. Previous studies suggest a type I IFN deficit in immune cells of patients MS, but data on interferon‐α/β receptor (IFNAR) expression and the relationship with treatment response are conflicting. Here, we compare IFN‐β‐mediated STAT4 activation in immune cells of untreated patients with MS and controls.


| INTRODUCTION
The type I interferon, interferonβ (IFNβ) has been used in the treatment of multiple sclerosis (MS) for many years 1,2 and remains a mainstay disease-modifying treatment for the relapsing forms of the disease. IFNβ, together with IFNα, utilizes a shared heterodimeric receptor, IFNAR, composed of two chains, IFNAR1 and IFNAR2, and signals via Janus kinases Tyk2 and Jak1, resulting in the formation of heterodimers of signal transducer and activator of transcription 1 (STAT1) and STAT2, which on translocation to the nucleus associate with IFN regulatory factor 9 (IRF9) forming the heterotrimeric complex IFN-stimulated gene factor 3 (ISGF3). This complex activates IFNdependent genes. After Janus kinase activation, IFNβ also activates STAT1 homodimers and, primarily in T and NK cells, through docking of STAT2 to STAT4, can induce phosphorylation of STAT4 and activation of gene expression via STAT4. The effect of IFNβ in MS is often partial and variable. Moreover, it is cell type specific, related to differential activation of STATs, sometimes with opposite effects in different cell types. 3 We have shown that IFNβ in T cells leads to the induction of antiinflammatory cytokine IL-10 in a STAT4 activation-dependent manner, explaining in part the immunomodulatory effects of IFNβ. 4 This IL-10 *These authors contributed equally to this work. induction was required for the suppression by IFNβ of IFNγ induction by  Production of type I IFN is deficient in patients with MS, and several studies showed a decreased responsiveness of patients with MS to IFN type I, 5,6 reflected, for example, in reduced expression of IFNstimulated genes and decreased phosphorylation of STAT1. 7 The aim of this study was to determine the effects of IFNβ on the phosphorylation of STAT4 in patients with MS compared with controls as an indication of response to IFNβ, and to determine its relationship with the expression of interferon type 1 receptor (IFNAR). We also aimed to determine whether the effects of IFNβ on phosphorylation of STAT4 in patients with MS correlate with their 5-year clinical course as measured by the change in the expanded disability scale (EDSS) score and clinical relapses.

| MATERIALS AND METHODS
The Nottingham Research Ethics Committee (UK) approved this study.

| Subjects included in the study
Twenty-seven relapsing-remitting patients with MS (17 women, 10 men; age range 29-60, mean age 43; mean expanded disability status scale, EDSS 3.6, range 1-6) and 12 healthy controls (seven women and five men; mean age 42) were included in the study. The patients with MS had not previously received disease-modifying drugs, were clinically stable at baseline (not in relapse) and had had two relapses in the last 2 years. Participants gave written informed consent prior to participation in the study.
In all MS subjects, treatment with IFNβ was commenced as part of their clinical care. Twenty-four patients were treated (13 with IFNβ 1b and 11 with IFNβ 1a), and all patients were followed up in clinic for 5 years. EDSS at 5 years of follow-up was compared with initial EDSS score.

| Cell stimulation
Peripheral blood mononuclear cells were left unstimulated or incubated with 10 ng/ml IFNβ (2.72 U/ml of Rebif ® ) for 30 minutes at 37°C. Varying IFNβ concentrations were analysed for their effect on STAT4 phosphorylation; concentrations used here were those that produced the peak pSTAT4. Optimal duration of stimulation for pSTAT4 results shown here was 30 minutes.

| Quantitative real-time polymerase chain reaction (PCR)
RNA was extracted using RNeasy miniprep kit (Qiagen, Valencia, CA, USA) following manufacturers' instructions. First-strand cDNA synthesis was initiated from 0.5 μg total RNA, using random hexamers triplicate on a MX4000 ® Multiplex System (Stratagene) using standard default thermal cycling conditions. Non-template controls were loaded in triplicate. Quantification of transcripts was carried out using the relative standard curve method as per Applied Biosystems (Foster City, CA, USA; 1997). 9,39 An equal aliquot of undiluted cDNA from each sample was pooled together. 39 This pool was serially diluted (neat, 1:2, 1:5, 1:10, 1:20) to produce standards, from which the Ct value was converted to ng total RNA equivalent used for first-strand synthesis. mRNA expression for each gene is normalized to internal standard (β2 microglobulin expression).

| Measurements
Peripheral blood mononuclear cells were evaluated on Epics XL flow cytometer (Beckman Coulter, Fullerton, CA, USA), and the results were analysed using software WINMDI 2.8. Statistical analysis employed the paired t-test. R 2 was used to assess correlation between IFNAR expression and STAT4 activation. Spearman's rank correlation was used to correlate EDSS change and pretreatment IFNAR expression. P≤.05 was considered significant.

| Activation of STAT4 by IFNβ
We first confirmed previous results showing that IFNβ increases pSTAT4. IFNβ stimulation increased pSTAT4 compared with unstimulated cells for MS and controls ( Figure 1). No difference in total STAT4 was observed (not shown).
We then compared STAT4 activation in patients and controls.

| IFNAR expression in patients and controls
We next investigated upstream of STAT4, at the receptor level.

We then assessed IFNAR1 and IFNAR2 mRNA expression in
PBMCs. Patients with MS and controls had similar IFNAR2 expression; however, IFNAR1 expression was significantly lower in patients with MS when compared to controls (P=.028) (Figure 3).
We also found a positive correlation between IFNAR expression and IFNβ-induced pSTAT4 in patients with MS (P=.03) (Figure 4).

| DISCUSSION
Here, we show that untreated patients with MS have reduced IFN responses as reflected by reduced STAT4 activation due to reduced IFNAR expression.
In our treatment-naive patients with MS, IFNAR1 expression was significantly lower than in controls. This confirms previous reports on IFNAR subtypes mRNA pointing to lower IFNAR1 mRNA levels in untreated patients with MS than in healthy controls. 10,11 The mechanism may involve downregulation of IFNAR by inflammatory cytokines, as we showed. 4 High levels of gene expression of IFNAR correlate with greater biological response to IFNβ as measured by STAT4 signalling has been associated with Th1 development, inflammation and autoimmunity; however, its roles are not completely understood. STAT4 Th1 signalling suppresses Foxp3-positive T regulatory cells (Treg) and favours autoimmune disease. 13 The transition of a subset of Th17 precursors to Th1-like cells is contingent upon STAT4 expression. 14 STAT4 represents a core signalling hub in MS. 4  and IFNβ are considered to have biologically opposing effects. IL-12 is increased 19 and considered detrimental in MS, 20 but IFNβ treatment does not appear to potentiate IL-12 proinflammatory effects. We have shown that prior exposure to IFNβ enhances STAT4 activation by IL-12, but the net effect is not proinflammatory because of parallel induction of IL-10 by IFNβ. 4 In progressive stages of MS, IL-10 loses its ability to suppress IFNγ and IL-12, while IFNβ treatment loses efficacy considerably in this phase of the disease. Therefore, the beneficial immunomodulatory effects of IFNβ may in part be a result of its ability to suppress IL-12. 4 The type I IFN pathway of STAT4 activation is more important in human immune responses than in mouse. 21 Type I IFN-induced STAT4 activation requires interactions of STAT4 with STAT2, and there is genetic dissimilarity between murine and human STAT2. 22 Here, we show a decreased IFNβ STAT4 signalling and IFNAR expression in whole, unseparated PBMCs from patients with MS. There are major differences in the responses of primary human leucocyte subsets to IFNβ. 23 Although the activation of STAT4 by type I IFN was thought initially to occur only in T and NK cells, STAT4 activation is also shown in endothelial cells, activated monocytes and dendritic cells (DCs). 24  Immune priming can change, positively or negatively, the quantity and quality of signal transduction by type I IFNs. 24 The decreased STAT4 signalling in response to IFNβ found in the PBMC in the group of untreated patients with MS reported here could therefore reflect the baseline MS immunological milieu in the absence of therapeutic immunomodulation.
Because of the role of STAT4 in mediating inflammation in MS and, at the same time, in mediating IFNβ anti-inflammatory effects, we planned to look at long-term clinical correlates of STAT4 activation in untreated patients with MS just prior to initiating disease-modifying treatment with IFNβ. The patient group was then followed up for 5 years. Disability progression at 5 years is often used as a measure of long-term response to treatments in MS. 27,28 Of note, the patients in this study were early participants in the UK Multiple Sclerosis Risk Sharing Scheme. 29 This large observational clinical cohort study recently showed that treatment with IFNβ reduces disability progression measured by EDSS scores over 6 years of treatment. 29 Early inclusions in the Risk Sharing Scheme (and access to treatment with IFN) consisted predominantly of patients with longer disease duration and F I G U R E 3 IFNAR mRNA expression. IFNAR1 and IFNAR2 mRNA expression measured using real-time PCR (relative to internal control: β2 microglobulin). Peripheral blood mononuclear cell obtained from untreated patients with MS (n=27) and controls (n=12) had similar IFNAR 2 expression; however, IFNAR1 expression was significantly lower in patients with MS when compared to controls (P=.028). Bars represent the normalized median value in arbitrary units, and error bars represent interquartile range. ★P=.028; NS, not significant F I G U R E 4 Correlation between IFNAR protein expression and IFNβ-induced STAT4 activation in patients with MS. A positive correlation is observed between IFNAR mRNA in unstimulated peripheral blood mononuclear cell from untreated patients with MS (n=20) and pSTAT4 mRNA induced after in vitro stimulation with IFNβ 10 ng/ml for 30 minutes (R=.43; P=.03) thus more advanced disease, more disability and higher EDSS scores.
This explains the relatively high EDSS scores at baseline in our patient group (mean EDSS 3.5). Generally, patients with MS who reach EDSS 3 may already enter secondary progression. 30 This may explain why this group showed evidence of disability progression over 5 years; this progression is difficult to quantify and compare to the overall Risk Sharing Scheme results, or what would be expected from the natural history of MS. However, in this cohort, clinical progression was not predicted by the pretreatment in vitro responsiveness to IFNβ. Whether the disease-modifying treatment altered the responsiveness and the expression of STAT4 and IFNAR in the short or medium term is unknown.
The lack of correlation of pSTAT4 expression before treatment with IFNβ to disability status at 5 years suggests the lack of a specific defect in STAT4 signalling that would link with the progression of disability. Of note, it was shown that, although the baseline expression of genes involved in IFNβ-related pathways can differ between different untreated patients with MS, and treatment with IFNβ can differentially induce gene expression (strong induction in those patients with low endogenous IFN-like activity, but only weak induction in patients with high endogenous IFN-like activity before treatment); patients with high and low endogenous IFN-like activity at baseline show similar clinical long-term courses of disease. 31 We found a negative, albeit marginal, correlation of relapses with IFNAR1 mRNA expression (r=-.4; P=.05). This may suggest that IFNAR expression can be a predictor of relapses under treatment.
Nevertheless, a reliable documentation of relapses was possible only in the 22 of patients followed (81%), and under-reporting of MS relapses cannot be excluded. In a multicentre, retrospective study including 103 patients with MS from 7 UK centres, 46% of patients did not report at least one relapse in the past and 28% of most recent relapses were unreported. 32 Several reasons may contribute to this, including rare review schedules. 32 The occurrence of relapses remains the principal factor for starting disease-modifying treatments (DMTs) under current MS treatment guidance. 33 The risk of severe MS relapses is lower in patients treated with DMTs. 34 While cost-effectiveness analyses in MS tend to focus on long-term disability, individual relapses have a major impact on pa-tients‫׳‬ working and family lives, which is often overlooked. 32 This highlights relapse reduction as a worthy treatment goal in MS. 32 Regular MRI follow-up data were not available for the group of patients in this study. MRI after the first year of treatment with IFNβ can be a good predictor of the outcome at 4 years, 35 and the occurrence of two or more new T2 lesions or new gadolinium-enhancing lesions in patients with variable responses to IFNβ associates a significantly increased risk of future relapses and progression. 36 In this study, we intended to determine whether pretreatment activation of pSTAT4 predicts longer term outcome. Therefore, we did not assess the medium-or long-term effects of IFNβ treatment on STAT4 signalling and IFNAR expression. Subnormal responses to type I IFN in MS can be corrected by long-term IFNβ therapy which is able to prime the subnormal IFN response and elevate STAT signalling levels. 37 Gene expression profiling studies show that the IFNβ deficiency exhibited by patients with MS can be corrected over time by repeated IFNβ administration 38 ; however, IFN signalling and gene activation pathways are partially resistant to IFN therapy in MS. 2 Of note, the decreased IFNβ response in patients with MS can be improved by combination therapy with other treatments with synergistic effects.
We have previously shown that prior exposure of PBMC of healthy controls and patients with MS to dexamethasone results in increased sensitivity of the cells to IFNβ, to an increase in pSTAT4 generation and in upregulation of its receptor expression. 39 We showed similar enhancement of IFNβ signalling in cells from healthy controls and patients with MS by curcumin. 40 In conclusion, here we show for the first time that IFNβ-induced STAT4 signalling is reduced in untreated patients with MS as compared to healthy subjects. The reduced pSTAT4 induction by IFNβ in patients with MS compared with controls indicates a decreased ability to activate IFNβ signalling pathway via STAT4. This, combined with the reduction in IFNAR expression, suggests that untreated patients with MS have decreased IFNβ responsiveness. Therapeutic strategies to improve this responsiveness, for example, by synergistic combinations may be beneficial in the treatment for MS in the future.