STAT-5 is activated constitutively in T cells, B cells and monocytes from patients with primary Sjögren's syndrome

Abstract

The expression and phosphorylation of signal transducer and activator of transcription-1 (STAT-1) have been shown to be markedly increased in the salivary gland epithelial cells of patients with primary Sjögren's syndrome (pSS). The present aim was to investigate the activation status of different STAT proteins in peripheral blood (PB) lymphocytes and monocytes, and their correlations with clinical parameters in patients with pSS. To this end, PB samples were drawn from 16 patients with active pSS and 16 healthy blood donors, and the phosphorylation of STAT-1, -3, -4, -5 and -6 proteins was studied in T cells, B cells and monocytes using multi-colour flow cytometry. In addition, mRNA expression of STAT molecules in PB mononuclear cells (PBMC) was studied with quantitative reverse transcriptase–polymerase chain reaction (RT–PCR). Basal phosphorylation of STAT-5 was found to be significantly higher in PB T cells, B cells and monocytes in patients with pSS than in healthy controls. The expression of STAT-5 mRNA was not increased in PBMC. pSTAT-5 levels in B cells and monocytes showed a significant correlation with serum immunoglobulin (Ig)G levels and anti-SSB antibody titres. Constitutive STAT-5 activation in monocytes and CD4⁺ T cells was associated with purpura. There were no major differences in the activation of other STATs between pSS patients and healthy controls. In conclusion, STAT-5 is activated constitutively in PB leucocytes in patients with pSS, and basal STAT-5 phosphorylation seems to associate with hypergammaglobulinaemia, anti-SSB antibody production and purpura.

Introduction

Primary Sjögren's syndrome (pSS) is a many-faceted chronic systemic autoimmune disease characterized by dry eyes and mouth, but also presenting with a wide range of possible extraglandular manifestations 1,2. The risk of non-Hodgkin lymphomas (NHL) has been found to be increased in pSS 3, and this risk has been associated, for example, with the presence of purpura, lymphadenopathy, peripheral neuropathy, cryoglobulinaemia, low C4 levels and increased serum beta-2 microglobulin levels 4–6. Both cell-mediated immune responses and B cell hyperreactivity reflected by hypergammaglobulinaemia and the formation of a plethora of autoantibodies such as anti-Sjögren's syndrome A (SSA) and anti-Sjögren's syndrome B (SSB) antibodies are involved in the pathogenesis of pSS. In recent years the use of biological therapies, in particular those targeting B cells, has also been studied in patients with pSS, with some encouraging results 7–11.

The current view is that the interferon (IFN) signature, both in peripheral blood mononuclear cells (PBMC) and in the salivary glands, is essential in the development of pSS 12,2. Besides types I and II interferons, other cytokines thought to play a role in the pathogenesis of pSS include interleukin (IL)-6, IL-7, IL-10, IL-12, IL-17, IL-21 and tumour necrosis factor (TNF)-α 12. The biological effects of several cytokines involved in the pathogenesis are dependent upon activation of the JAK–STAT (Janus kinase and signal transducer and activator of transcription) signal transduction pathways. This activation leads to the recruitment and phosphorylation of the STAT proteins, resulting in their translocation to the nucleus, where they initiate transcription of the target genes 13. The mRNA and protein of STAT-1 have been shown to be expressed strongly in the cells of labial salivary glands and phosphorylated STAT-1 in the glandular epithelium of patients with pSS 14. In addition, Ramos and associates have demonstrated increased phosphorylation of STAT-3 in T cells from pSS patients 15. The variant haplotype of STAT-4 has been found to predispose to pSS 16–20 and to affect the activation of the type 1 IFN-pathway in pSS 21.

In this study we have analysed the basal activation state of STAT-1, -3, -4, -5 and -6 in fresh peripheral blood (PB) T cells, B cells and monocytes in patients with pSS and healthy volunteers. Further, we have investigated the relationships between STAT phosphorylation and the clinical features of pSS and plasma cytokine concentrations. We also studied the expression of the respective STAT mRNAs in PBMCs by quantitative reverse transcription–polymerase chain reaction (qRT–PCR).

Subjects and methods

Subjects

Sixteen patients with pSS from the Centre for Rheumatic Diseases in Tampere University Hospital were recruited for the present study. The inclusion criteria were: (1) fulfillment of at least four of the revised American–European consensus group criteria for pSS 22 and (2) active disease with one of the following findings: either the EULAR Sjögren's syndrome disease activity index (ESSDAI) 23 > 11 or one of the following laboratory findings: erythrocyte sedimentation rate (ESR) > 20 mm/h, serum immunoglobulin (Ig)G > 15 mg/l, serum beta-2 microglobulin > 2·2 mg/l or serum complement C4 < 0·10 mg/l. The patients were allowed to have corticosteroid, hydroxychloroquine or immunosuppressive medication such as azathioprine, methotrexate or rituximab in either past or current use. Currently, three patients were treated with low-dose prednisolone (for 11 months, 3 years and 5 years, respectively), one patient with hydroxychloroquine (for 5 years) and one patient with azathioprine (for 4 years). The clinical, immunological and demographic data on the 16 pSS patients are presented in Table1.

Table 1.

Demographic, clinical and immunological characteristics of the 16 patients with primary Sjögren's syndrome fulfilling at least four of the revised American–European consensus group criteria for primary Sjögren's syndrome (pSS) and having active disease

Characteristic	Value
Female/male	14/2
Age, median (range), years	53 (32–80)
Disease duration, median (range), years	11 (0–27)
Duration of sicca symptoms of the eyes, median (range), years	14 (2–40)
Duration of sicca symptoms of the mouth, median (range), years	14 (3–37)
Salivary gland histological grade ≥ 3, n = 13 (%)	12 (92)
ESSPRI, cm, median (IQR)	4·30 (3·01, 5·83)
ESSDAI, median (IQR)	5·00 (3·00, 8·75)
Salivary gland involvement (%)	10 (63)
Purpura (%)	5 (31)
Arthritis (%)	5 (31)
Pulmonary findings (%)	2 (13)
Myositis (%)	1 (6)
CNS symptoms (%)	1 (6)
PNS symptoms (%)	0
Current DMARD use	2 (13)
Current glucocorticoid use	3 (19)
PGH-VAS, cm, median (IQR), n = 15	1·50 (0·10, 2·80)
Pain-VAS, cm, median (IQR), n = 15	0·40 (0·00, 3·00)
HAQ, mean ± s.d. (range 0–3), n = 15	0·33 ± 0·62
Haemoglobin, g/l, mean ± s.d.	132 ± 9
Leucocytes (× 109/l), mean ± s.d.	4·8 ± 1·8
Thrombocytes (× 109/l), mean ± s.d.	249 ± 61
ESR (mm/h), mean ± s.d.	24 ± 23
CRP, mg/l, mean ± s.d.	0·92 ± 1·40
Serum IgG, g/l, mean ± s.d.	18·0 ± 4·0
Serum IgA, g/l, mean ± s.d.	2·2 ± 1·2
Serum IgM, g/l, mean ± s.d.	1·66 ± 1·21
Serum beta-2 microglobulin, mg/l, mean ± s.d.	2·83 ± 0·55
Serum C3, g/l, mean ± s.d.	1·00 ± 0·18
Serum C4, g/l, mean ± s.d.	0·15 ± 0·05
Anti-SSA antibody-positive (%)	16 (100)
Anti-SSB antibody-positive (%)	11 (69)
Plasma IFN-γ, pg/ml, median (IQR)	78·6 (21·4, 123·2)
Plasma IL-1β, pg/ml, median (IQR)	18·8 (3·12, 28·3)
Plasma IL-2, pg/ml, median (IQR)	34·8 (15·9, 107·6)
Plasma IL-4, pg/ml, median (IQR), n = 14	90·9 (41·4, 198·6)
Plasma IL-6, pg/ml, median (IQR), n = 15	11·3 (7·05, 26·5)
Plasma IL-7, pg/ml, median (IQR)	21·3 (10·7, 37·9)
Plasma IL-10, pg/ml, median (IQR), n = 15	101·0 (35·9, 306·3)
Plasma TNF-α, pg/ml, median (IQR)	14·4 (7·2, 17·3)

ESSPRI = EULAR Sjögren's syndrome patient-reported index; IQR = interquartile range; ESSDAI = EULAR Sjögren's syndrome disease activity index; CNS = central nervous system; PNS = peripheral nervous system; DMARD = disease-modifying anti-rheumatic drug; PGH = patient's global health assessment; SSA = Sjögren's syndrome A; SSB = Sjögren's syndrome B; VAS = visual analogue scale; HAQ = Health Assessment Questionnaire; ESR = erythrocyte sedimentation rate; CRP = C-reactive protein; IFN = interferon; Ig = immunoglobulin; IL = interleukin; s.d. = standard deviation; TNF = tumour necrosis factor.

To evaluate the diagnostic criteria for pSS, the following data were gathered by interview and from patient charts: duration of pSS, presence of sicca symptoms in the eyes and mouth and the present status of the mouth, a history or presence of recurrent parotid or submandibular gland swellings, the results of the Schirmer 1 test and the results of ophthalmological examinations, salivary flow rate, minor salivary gland histological findings, results of salivary gland ultrasound examinations and the results of anti-SSA and anti-SSB antibody determinations. In addition, the presence of extraglandular manifestations of pSS and their activity was determined by the ESSDAI and the results of the EULAR Sjögren's syndrome patient-reported index (ESSPRI) 24 were also gathered, and data on the routine laboratory tests and immunological findings recorded.

PB samples were obtained from the 16 pSS patients. In addition, PB samples were obtained from 16 (six female, 10 male) anonymous healthy blood donors (Finnish Red Cross Blood Transfusion Service, Tampere), whose median age was 51 years (range 21–63 years), and from 12 patients (eight female, four male) with active [disease activity score in 28 joints (DAS-28) ≥ 3·2] rheumatoid arthritis (RA), whose median age was 70 years (range 55–88 years), who served as controls. The RA patients were receiving various medications, including corticosteroids, traditional disease-modifying anti-rheumatic drugs (DMARDs) and biological therapies.

Flow cytometry

To study the phosphorylation of STAT proteins, five-colour staining and flow cytometric analysis was performed. First, 100 μl aliquots of fresh blood samples were stained with fluorescein isothiocyanate (FITC)-conjugated anti-CD3 (clone UCHT1), allophycocyanin (APC)-eFluor 780-conjugated anti-CD19 (clone HIB19; both from eBioscience, San Jose, CA, USA) and APC-conjugated anti-CD14 (clone TÜK4; Miltenyi Biotec, Auburn, CA, USA) for 20  min on ice. Alternatively, to discriminate between CD45RO⁺ and CD45RO⁻ cells, the cells were stained with FITC-conjugated anti-CD45RO (clone UCHL1; eBioscience) and APC-H7-conjugated anti-CD8 (clone SK1; BD Biosciences, San Jose, CA, USA). The cells were then fixed and red cells lysed with BD Phosflow Lyse/Fix buffer for 15 min at 37 °C, washed and permeabilized in ice-cold methanol overnight at −20 °C. Following two washes, the samples were stained with peridinin chlorophyl (PerCP)-cyanin 5·5 (Cy5·5)-conjugated anti-CD4 (clone SK3) and phycoerythrin (PE)-conjugated phospho-STAT-1 (clone 4a), phospho-STAT-3 (clone 4/P-Stat-3), phospho-STAT-4 (clone 38/p-Stat-4), phospho-STAT-5 (clone 47/Stat-5) or phospho-STAT-6 (clone 18/P-Stat-6) antibodies or mouse isotype control IgG (all from BD Biosciences) for 30 min at RT, followed by two washes. The stainings were analysed using a BD FACSCanto II flow cytometer.

The functionality of phospho-STAT antibodies was tested by stimulating PB with cytokines IFN-γ, IL-6, IL-12, IL-2 and IL-4, which were shown to induce significant phosphorylation of STAT-1, -3, -4, -5 and -6, respectively (data not shown).

Analysis of flow cytometer data was made using Cytobank software (http://www.cytobank.org/). The CD4⁺ and CD4⁻ T cells were gated from the CD3⁺ lymphocyte population. CD19⁺ cells represent B cells and CD14⁺ cells represent monocytes. The level of STAT phosphorylation is given as mean fluorescence intensity ratio (MFIR = MFI of pSTAT staining divided by the MFI of isotype control staining).

Plasma cytokine determinations

Cytokine levels were measured in plasma samples, which were stored at −70 °C. Milliplex MAP high-sensitivity human cytokine panel with magnetic beads (Millipore, Billerica, MA, USA) was used to measure IFN-γ, IL-1β, IL-2, IL-4, IL-6, IL-7, IL-10 and TNF-α levels simultaneously. Duplicate determinations were carried out for each sample.

RNA isolation and qRT–PCR analysis

PBMCs from 16 patients with pSS and 14 healthy controls were isolated by Ficoll-Paque Plus (Amersham Biosciences, Buckinghamshire, UK) density gradient centrifugation and washed twice with PBS containing 2 mM ethylenediamine tetraacetic acid (EDTA). Total RNAs were isolated from PBMCs using the RNeasy MiniKit (Qiagen, Valencia, CA, USA), according to the manufacturer's instructions. One µg of total RNA was reverse-transcribed using M-MLV reverse transcriptase (Invitrogen, Carlsbad, CA, USA) and random hexamers (Amersham Bioscience) as primer.

The forward and reverse primers recognizing separate exons of STAT genes were designed using the Primer3 program (available at http://primer3sourceforge.net).

The following primers were used: 5′-TCACATTCACATGGGTGGAG-3′ and 5′-CAAAGGCATGGTCTTTGTCA-3′ for STAT-1, 5′-TCACATGCCACTTTGGTGTT-3′ and 5′-GCAATCTCCATTGGCTTCTC-3′ for STAT-3, 5′-GGCAATTGGAGAAACTAGAGG-3′ and 5′-AGGGTGGGTTGGCATACAT-3′ for STAT-4, 5′-GCCAGATGCAGGTGCTGTA-3′ and 5′-GGGATTGTCCAAGTCAATGG-3′ for STAT-5A, 5′-GCGTTATATGGCCAGCATTT-3′ and 5′-CTGGTGCTCTGCCTTCTTCT-3′ for STAT-5B, 5′-GGAAGGGCACTGAGTCTGTC-3′ and 5′-GGCTTTGGCATTGTTGTCTT-3′ for STAT-6 and 5′-GAATATAATCCCAAGCGGTTTG-3′ and 5′-ACTTCACATCACAGCTCCCC-3' for TBP.

The 10-μl real-time PCR reactions were performed in a CFX96 apparatus (Bio-Rad Laboratories, Hercules, CA, USA) using Maxima SYBR Green/ROX qPCR master mix (Thermo Scientific Fermentas, Waltham, MA, USA). Reactions contain 2 μl aliquot of sample or standard cDNA, 0·3 μM of the relevant primers and 1× reaction mixture containing Taq DNA polymerase, SYBR Green I fluorescent dye, deoxynucleotide triphosphates (dNTPs) and reaction buffer.

PCR amplification consisted of an initial incubation step at 95 °C followed by 40 cycles of denaturation (15 s at 95 °C), annealing (30 s at 58 °C) and extension (30 s at 72 °C). The mean STAT expression values from duplicate samples were normalized by dividing them by the mean values obtained for TBP housekeeping gene. Measurements in which the numerical difference between duplicates was more than 30% of the mean value were omitted from the analysis.

Statistical methods

Statistical analyses were performed with IBM spss Statistics version 20. The Mann–Whitney U-test was used for comparisons of continuous variables. Correlations were calculated with Spearman's correlation coefficient. Findings were considered statistically significant at P < 0·05.

Ethical considerations

The study was approved by the Ethical Committee of Tampere University Hospital, Tampere, Finland. All patients gave their written informed consent. The study was conducted according to the principles of the Declaration of Helsinki.

Results

STAT-5 is activated constitutively in PB T cells, B cells and monocytes from pSS patients

We examined whether the basal levels of pSTAT-1, -3, -4, -5 and -6 in unstimulated CD4⁺ T cells (i.e. T helper cells), CD4⁻ T cells (containing mainly CD8⁺ T cells), B cells and monocytes in patients with pSS differ from those in healthy controls. For this purpose, we set up a multi-colour flow cytometric method which permitted analysis of STAT phosphorylation in different cell populations. As fresh PB was used and there was no need for cell purifications or manipulations, the results are likely to reflect well the phosphorylation status of STAT proteins in vivo.

The levels of pSTAT-1, -4 or -6 in unstimulated CD4⁺ T cells, CD4⁻ T cells, B cells and monocytes in patients with pSS did not differ from those in healthy controls (Fig. 1a,c,e). The basal level of pSTAT-3 in CD4⁻ T cells was higher in pSS patients than in healthy controls [median 1·80, interquartile range (IQR) 1·58, 2·10 versus 1·.50, IQR 1·43, 184, respectively, P = 0.029]. However, although barely reaching statistical significance, this difference was very modest (as seen in Fig. 1b), and clearly less prominent than the differences in pSTAT-5 levels between pSS patients and controls.

Figure 1.

Phosphorylated signal transducer and activator of transcription (pSTAT)-1), -3, -4, -5 and -6 levels in unstimulated peripheral blood (PB) CD4⁺ T cells, CD4⁻ T cells, B cells and monocytes in patients with primary Sjögren's syndrome (pSS) (n = 14–16) and healthy blood donor controls (n = 8–16) (a–e) and pSTAT-5 levels in patients with RA (n = 6–12) and healthy controls (f). The exact patient numbers are given in parentheses after the descriptions of the cell populations studied. The data are presented as mean fluoresence intensity ratio [MFIR = mean fluoresence intensity (MFI) of pSTAT staining divided by MFI of isotype control staining]. Horizontal lines represent mean values within groups. Statistically significant differences are marked with an asterisk (*P < 0·05; **P < 0·01; ***P < 0·001).

We found the basal pSTAT-5 levels in all cell types studied to be significantly higher in patients with pSS when compared with healthy controls (Fig. 1d). The highest pSTAT-5 levels were observed in CD4⁺ T cells (Fig. 1d). The differences in STAT-5 phosphorylation remained significant in all cell types when only female pSS patients and female control subjects were compared (data not shown). The gating strategy and example histograms of pSTAT-5 stainings from two pSS patients and controls are shown in Fig. 2.

Figure 2.

(a)The dot-blots indicate the gating strategy for different cell populations. (b) Example histograms of isotype control and phosphorylated signal transducer and activator of transcription-5 (pSTAT-5) stainings of two healthy volunteers and two patients with primary Sjögren's syndrome (pSS). Mean fluorescence intensity ratios [MFIR = mean fluoresence intensity (MFI) of pSTAT staining divided by MFI of isotype control staining] of pSTAT-5 stainings are presented in the table below the histograms.

In order to investigate whether STAT-5 phosphorylation is increased preferentially in antigen-activated memory T cells in pSS, we performed analysis with a memory T cell marker CD45RO (data not shown). Surprisingly, we observed that the phosphorylation of STAT-5 was somewhat higher in CD45RO⁻, CD4⁺ T cells [MFIR 5·26 ± 0·47 (mean ± standard deviation (s.d.), n = 3] when compared with CD45RO⁺, CD4⁺ memory T cells (MFIR 3·82 ± 0·68). A similar tendency was observed between CD45RO⁻, CD8⁺ T cells (MFIR 2·83 ± 0·40) and CD45RO⁺, CD8⁺ memory T cells (MFIR 2·43 ± 0·18).pSTAT-5 levels in CD4⁺ and CD4⁻ T cells showed a significant correlation with each other (r = 0·748, P = 0.001) and with pSTAT-5 levels in monocytes (r = 0·509, P = 0·044 and r = 0·561, P = 0·024, respectively). In addition, pSTAT-5 levels in B cells and in monocytes showed a strong correlation (r = 0·857, P < 0·0001) with each other.

Previous data from our laboratory have shown increased phosphorylation of STAT-3 in circulating leucocytes from patients with RA25. To investigate whether STAT-5 activation is specific for pSS, we analysed baseline STAT-5 phosphorylation in patients with RA. As shown in Fig. 1f, pSTAT-5 levels in CD4⁺ and CD4⁻ T cells were also up-regulated in RA patients, whereas pSTAT-5 levels in B cells and monocytes were comparable to those in healthy controls.

pSTAT-5 levels in B cells and monocytes correlate with serum IgG levels and anti-SSB titres

As basal pSTAT-5 levels were up-regulated in all cell types studied, we analysed the clinical data on the pSS patients in relation to pSTAT-5 activation.pSTAT-5 levels in B cells and monocytes correlated significantly with serum IgG concentrations and anti-SSB antibody titres (Table2), and inversely with the duration of xerostomia. In addition, STAT-5 phosphorylation in B cells correlated inversely with the ESSPRI and patients’ global health-visual analogue scale (VAS) and pain-VAS. No correlation was found with the ESSDAI (Table2). There were no significant correlations between STAT-5 phosphorylation in CD4⁺ T cells and CD4⁻ T cells and any of the clinical findings listed in Table2 (data not shown).

Table 2.

Correlations of phosphorylated signal transducer and activator of transcription-5 (pSTAT-5) levels in B cells and monocytes with clinical and immunological characteristics in 16 patients with primary Sjögren's syndrome (pSS)

	B cells		Monocytes
Variable	r for pSTAT-5 P-value		r for pSTAT-5 P-value
Age	−0·286	0·301	−0·062	0·820
Duration of dryness of the eyes	−0·405	0·135	−0·391	0·135
Duration of xerostomia	−0·787	<0·0001	−0·512	0·043
Duration from diagnosis	−0·438	0·103	−0·180	0·506
ESR	0·092	0·261	0·193	0·474
CRP	0·032	0·910	0·176	0·515
Serum IgG	0·600	0·018	0·524	0·037
Serum IgA	−0·246	0·376	−0·397	0·128
Serum IgM	−0·182	0·516	−0·047	0·863
Serum beta-2 microglobulin	0·032	0·909	0·145	0·592
Anti-SSB titre (n = 13)	0·696	0·012	0·579	0·038
C3	0·184	0·511	0·211	0·434
C4	−0·054	0·849	−0·054	0·849
ESSPRI	−0·732	0·002	−0·326	0·217
ESSDAI	0·050	0·859	0·169	0·533
PGH-VAS	−0·683	0·007	−0·371	0·173
Pain-VAS	−0·632	0·015	−0·367	0·178
HAQ	−0·503	0·067	−0·264	0·341

ESR = erythrocyte sedimentation rate; r = correlation coefficient; CRP = C-reactive protein; ESSPRI = EULAR Sjögren's syndrome patient-reported index; ESSDAI = EULAR Sjögren's syndrome disease activity index; SSA = Sjögren's syndrome A; SSB = Sjögren's syndrome B; Ig =immunoglobulin; PGH = patient's global health assessment; VAS = visual analogue scale; HAQ = Health Assessment Questionnaire

pSTAT-5 levels in CD4⁺ T cells and monocytes are increased in pSS patients with purpura

In addition, we compared pSTAT-5 levels in pSS patients yielding distinct extraglandular findings from the ESSDAI with pSTAT-5 levels in pSS patients without these findings. Comparisons were made only regarding extraglandular manifestations, which were present in at least three of the patients (see Table1 for frequencies of the extraglandular findings). These extraglandular findings were: current or previous purpura, current or previous salivary gland swelling and current or previous arthritis.pSS patients with a history of purpura (n = 5) had significantly higher median (IQR) pSTAT-5 levels in their CD4⁺ T cells compared with those without purpura (n = 11) [5·59 (IQR = 4·28, 7·38) versus 3·88 (IQR = 3·31, 4·50), P = 0·020]. pSTAT-5 levels in CD4^– T cells did not differ significantly between pSS patients with and without purpura [3·74 (IQR = 3·30, 5·06) versus 3·05 (IQR = 2·56, 3·73), P = 0·061], nor did the pSTAT-5 levels in B cells differ in the respective groups [4·51 (IQR = 3·24, 5·66) versus 3·38 (IQR = 2·59, 3·56), P = 0·090]. The levels of pSTAT-5 in monocytes of pSS patients with a history of purpura were significantly higher than in pSS patients without purpura [3·86 (IQR = 3·24, 5·66) versus 2·92 (IQR = 2·16, 3·48), P = 0·027].

No significant differences were observed in pSTAT-5 levels between pSS patients with a history of salivary gland swellings or with a history of arthritis and those without these manifestations (data not shown).

Relationship between pSTAT-5 levels and plasma cytokine concentrations

We next investigated whether any of the plasma cytokine concentrations determined (IFN-γ, IL-1β, IL-2, IL-4, IL-6, IL-7, IL-10, TNF-α) correlated with pSTAT-5 levels in the pSS patients. The plasma IL-1β concentrations were significantly higher in the 16 pSS patients compared with 14 healthy controls (median 18·8 pg/ml, IQR = 3·12, 28·3 versus 2·07 pg/ml, IQR = 1·15, 5·27, P = 0·005), as were also the plasma IL-2 concentrations (34·8 pg/ml, IQR = 15·9, 107·6 versus 12·6 pg/ml, IQR = 7·12, 20·0, P = 0·010), respectively. Other plasma cytokine levels did not differ significantly between the pSS patients and the healthy controls (data not shown).

Of the cytokines tested, IL-2 and IL-7 function by activating STAT-5 26,27. However, the levels of IL-2 and IL-7 did not correlate significantly with pSTAT-5 in any of the cell types. Interestingly, pSTAT-5 levels in CD4⁺ T cells correlated significantly with IFN-γ (r = 0·500, P = 0·049), IL-4 (r = 0·560, P = 0·037, n = 14) and TNF-α concentrations (r = 0·535, P = 0·033). pSTAT-5 levels in CD4⁻ T cells correlated significantly with IFN-γ (r = 0·570, P = 0·021), IL-10 (r = 0·536, P = 0·039, n = 15) and TNF-α levels (r = 0·596, P = 0·015). STAT5 phosphorylation in B cells correlated significantly only with IL-4 (r = 0·653, P = 0·011, n = 14). None of the cytokines correlated with pSTAT-5 levels in monocytes.

Expression of STAT mRNA in PBMC

The expression of STAT-1 and STAT-3 mRNA in PBMC was significantly higher in pSS patients than in healthy controls, STAT-1 being more clearly up-regulated (Fig. 3). In contrast, there were no significant differences in STAT-4, -5A, -5B or -6 mRNA expression between pSS patients and healthy controls (Fig. 3).

Figure 3.

The levels of signal transducer and activator of transcription (STAT)-1, -3, -4, -5A, -5B and -6 mRNA in peripheral blood mononuclear cells (PBMC) from patients with primary Sjögren's syndrome (pSS) (n = 11–15) and healthy controls (n = 8–11). The data are presented as relative STAT expression divided by TATA-binding protein (TBP) levels, and the horizontal lines represent the median expression levels in the groups. Statistically significant differences between pSS patients and healthy volunteers are marked with an asterisk (*P < 0·05; **P < 0·01).

Discussion

The main finding in this study was that STAT-5 is activated constitutively in PB T cells, B cells and monocytes in patients with pSS compared with healthy controls. In contrast, the basal activation of STAT-1, -4 or -6 did not differ between pSS patients and healthy volunteers. The analysis was conducted with minimal cell manipulations and should thus reflect the situation in vivo. STAT-3 activation was modestly higher in pSS patients than in healthy controls in CD4^– T cells, but not in any other cell types. This observation is in line with previous findings by Ramos and associates, who observed that there is constitutive activation of STAT-3 in T cells from pSS patients 15. However, they only investigated STAT-3 phosphorylation, not the phosphorylation of other STATs. It has to be emphasized that in our study the difference in STAT-3 phosphorylation in CD4^– T cells between pSS patients and controls was minor, in particular compared with the respective findings regarding pSTAT-5.

The phosphorylation of STAT-4, -5 and -6 in PB leucocytes has not been investigated previously in patients with pSS. However, in patients with systemic lupus erythematosus (SLE), increased activation of STAT-5 in B cells and T cells has been reported 28, as found here in patients with pSS. Constitutively activated STAT-5 has also been found in several forms of lymphoid leukaemia and lymphomas, both in mice and in humans 26, and over-expression of constitutively active STAT-5 has also been identified in chronic myeloid leukaemia 29. The concordance between our current findings of constitutively activated STAT-5 phosphorylation in pSS patients and previously reported findings in lymphomas (reviewed in [26]) is in line with the increased risk of NHL in pSS 3–6. Regarding the risk of lymphoproliferative disorders, the constitutive activation of pSTAT-5 in B cells could be important.

Activation of STAT-5 by IL-2 in naive CD4⁺ T cells is essential for the differentiation of regulatory T cells (T_reg), which are thought to suppress autoimmune diseases, and constitutively active STAT-5b has been shown to enforce T_reg cell development 30. In addition, activation of STAT-5 inhibits the differentiation of Th17 cells, which are pathogenic in the context of autoimmune diseases 31. Increasing evidence points to the instability of Th cell subtypes, and already differentiated Th cells can be redirected to another subtype in the presence of optimal signals 32. It may therefore be speculated that the increased constitutive STAT-5 phosphorylation found here in CD4⁺ T cells from the pSS patients could enhance T_reg cell development, while inhibiting the generation of Th17 cells. Indeed, the proportion of CD4⁺25^+high T_reg cells has been found to be increased in the PB of pSS patients 33. However, contrasting findings have also been reported 34, and currently the role of T_reg cells in the pathogenesis of pSS remains controversial 12.

It may be argued that increased STAT-5 phosphorylation in T cells from patients with pSS could be related to differences in the T cell subtypes between patients and controls. For example, the proportion of memory T cells is increased in the PB of pSS patients 35, and Th1 and Th17 subtypes seem to predominate in pSS 12. However, in our study STAT-5 phosphorylation in CD45RO⁺ memory T cells from pSS patients was lower than in naive T cells. In addition, STAT-5 phosphorylation was increased in the whole T cell population, not only in a certain proportion of T cells. Finally, the up-regulation of pSTAT-5 levels in B cells and monocytes also suggests that this finding is a general phenomenon observed in immune cells from pSS patients, and not specific for a certain subtype of cells.

In SLE patients a positive association has been found between pSTAT-5 levels and the lupus activity index 28. We found no association between pSTAT-5 and the ESSDAI, i.e. the new activity index of systemic manifestations of pSS. However, in monocytes and CD4⁺ T cells and, to a certain extent, in B cells, there was an association between constitutive STAT-5 activation and purpura, a manifestation which has been associated with hypergammaglobulinaemia and has, in some studies, been found to be an adverse predictor for NHL 5. Nevertheless, it should be borne in mind that the number of pSS patients with purpura was low, and therefore strong conclusions cannot be drawn.

Moreover, in B cells and monocytes pSTAT-5 levels correlated significantly with serum IgG levels and anti-SSB antibody titres, whereas they correlated inversely with the duration of xerostomia. STAT-5 activation in B cells also correlated inversely with ESSPRI and patients’ global health-VAS and pain-VAS. However, when interpreting these data, it should be taken into account that only a limited number of pSS patients were included in the present study. It would, nevertheless, appear that pSTAT-5 levels are not associated with the sicca symptoms and arthralgias, but rather with signs of immunological activity. Such a conception is particularly plausible, as STAT-5 is a key regulator of immune functions and the lymphoid system (reviewed in 26,27,36). That the association between STAT-5 phosphorylation and clinical parameters depends upon the cell type studied would imply that the cell types in which STAT-5 is activated vary in different patients.

Previous data from our laboratory indicate that STAT-3 is constitutively active in T cells and monocytes from patients with RA, and STAT-3 activation correlates with IL-6 levels 25. In contrast, pSTAT-3 levels are not increased in CD4⁺ T cells, B cells or monocytes from pSS patients. According to the present results, pSTAT-5 levels in T cells from RA patients are also up-regulated. However, in contrast to findings in pSS patients, pSTAT-5 levels in B cells and monocytes from RA patients are normal. Therefore, it seems that there are disease-specific profiles in the constitutive activation of STATs. A plausible explanation would be that circulating leucocytes are exposed to different cytokines in pSS and RA.

STAT-5 is activated by a variety of cytokines, including several interleukins, erythropoietin (EPO) and prolactin 37. Of the cytokines tested in the present study, IL-2 and IL-7 function by activating STAT-5 26,27. However, we found no significant correlation between IL-2 or IL-7 levels and constitutive pSTAT-5. Conversely, pSTAT-5 levels correlated with IFN-γ, IL-4, IL-10 and/or TNF-α, depending on cell type. Because these cytokines do not activate STAT-5, it seems likely that both they and pSTAT-5 are associated with the inflammatory response in pSS, but that these cytokines are not directly responsible for STAT-5 activation in PB cells.

As STAT-5 was activated constitutively, we also investigated whether its expression in PBMC was up-regulated. The expression was similar in pSS patients and controls, indicating that increased expression of STAT-5 does not explain the constitutive activation of pSTAT-5 in pSS patients. Conversely, we found increased expression of STAT-1 and -3 in PBMC from pSS patients. Expression of mRNA and protein for STAT-1 and -3 has been shown to be increased in the salivary glands of patients with pSS 14,38,39. Our results using PB samples are in accord with these previous findings.

We present here a novel finding of constitutive STAT-5 activation in PB leucocytes from pSS patients. Although STAT-5 is constitutively active, its expression level in PBMCs is unaltered in pSS. Importantly, STAT-5 phosphorylation seems to correlate with clinical manifestations such as purpura, anti-SSB antibody production and hypergammaglobulinaemia in pSS patients. These findings should encourage further studies in this context, including studies on the effects of cytokine stimulations on STAT phosphorylation, particularly as the therapeutic potential of the JAK–STAT inhibitors is currently being investigated for various autoimmune diseases.

Author contributions

All authors (M. P., O. S., P. I.) took part in the design of the study and in the interpretation of the data. M. P. was responsible for patient interviews and examinations, and wrote the first draft of the manuscript. P. I. planned the laboratory experiments. P. I. and O. S. were responsible for the laboratory work. O. S. provided the reagents and laboratory equipments. All authors participated in the revision of the manuscript, and have read and approved the final manuscript.

Disclosure

The authors declare that they have no conflicts of interest.