Introduction

Type 1 diabetes mellitus is an immune-mediated disease caused by an inflammatory reaction in the pancreatic islets of Langerhans. IL-1β and IFN-γ are pro-inflammatory cytokines implicated in the inflammatory reaction causing beta cell death [1]. IL-1β activates the transcription factor nuclear factor kappa B (NFκB) and the mitogen-activated protein kinases extracellular signal-regulated kinase, p38 and c-Jun N-terminal kinase [2, 3]. IFN-γ activates the transcription factor signal transducers and activators of transcription (STAT)-1 [4]. These transcription factors and mitogen-activated kinases induce the expression of several pro-apoptotic genes, among them genes encoding small chemotactic cytokines (chemokines) [5], that enhance the inflammatory response by recruiting immune-cells to sites of injury or inflammation [6, 7]. Chemokines thereby play an important role in the development of the inflammatory infiltrate in insulitis in the early stages of diabetes [8].

Chemokines can be divided into four subfamilies based on their structure and function. The CC- and CXC-families of chemokines are the largest and the most thoroughly investigated groups. Both sub-families carry four cysteines that are paired by disulphide bonds [9, 10]. In the CC-subfamily of chemokines, the cysteines are adjacent to each other, whereas in the CXC family they are separated by a single amino acid [10]. The CC-family of chemokines includes monocyte chemoattractant protein (MCP)-1 and macrophage inflammatory protein-3α (ST-38). Both MCP-1 and ST-38 are known to attract monocytes, T cells and natural killer cells, and are upregulated in rat beta cells by cytokines [11, 12].

The chemokine macrophage inflammatory protein-2 (MIP-2) belongs to the family of CXC-chemokines and is involved in the attraction of monocytes, T cells, natural killer cells and basophils. Together with the chemokine interferon inducible protein-10 (MOB-1) of the same family, MIP-2 is also upregulated in beta cells upon cytokine-exposure as shown in array studies [12, 13]. Transcription of the genes encoding the chemokines mentioned above is dependent on the transcription factor NFκB [14–17].

Fas (CD95) is a cell-death receptor, which induces apoptosis through an intracellular death domain upon binding by the Fas ligand (FasL) [18, 19]. FasL is mainly expressed on the surface of activated T cells, but is also constitutively expressed by beta cells. FasL belongs to the TNF family of membrane-associated cytokines and Fas is a member of the TNF receptor family [20, 21]. Binding of the ligand to its receptor leads to activation of caspase-8, which subsequently cleaves procaspase-3, resulting in completion of the cell-death programme [22].

In beta cells cytokines induce upregulation of Fas, making the beta cells susceptible to apoptosis upon interaction with FasL-expressing T cells as well as neighbouring beta cells [19, 23]. One study showed that reduced T cell FasL expression prevents diabetes in NOD mice [24]. Several similar studies have focused on the importance of Fas in beta cell death and the development of diabetes [25, 26]. On the other hand, evidence against a role of Fas has also been presented [27, 28], leaving much scope for debate in the field [4].

Cytokines convey their biological information to target cells by binding to cell-surface receptors, thereby activating intracellular signal transduction cascades that induce changes in transcription. However, negative regulators that ensure an appropriate cellular and physiological response to cytokine stimulation are needed to withstand deleterious effects of cytokine signalling. The suppressor of cytokine signalling (SOCS) proteins constitute one such family of negative regulators of cytokine signal transduction that has been shown to downregulate cytokine signalling [29]. The SOCS family consists of eight members, all of which are intracellular proteins: SOCS-1 to -7 and cytokine-inducible SH2-containing protein (CIS). All eight contain a central SH2 domain (of approximately 95 amino acids) and a SOCS-box, i.e. a conserved carboxyl-terminal domain.

Transcripts encoding SOCS-1 to SOCS-3 and CIS are normally present at low or undetectable levels in resting cells, but can rapidly be induced by a broad spectrum of cytokines and other factors [30]. Transcription of Socs genes is induced by the STAT and NFκB transcription factors upon cytokine stimulation, and the SOCS proteins thus generated subsequently inhibit the same pathway that initiated their production. Thus, it is generally accepted that SOCS proteins act in a negative feedback-loop that attenuates cytokine-induced signalling [31].

Previous work has shown that SOCS-3 inhibits IL-1β- and IFN-γ-induced nitric oxide production and apoptosis in the beta cell line insulinoma (INS)-1 [32] through specific inhibition of IL-1β-induced TGF-β activated kinase activity [33]. Furthermore, characterisation of the gene expression profiles in these cells showed that IL-1β-induced pro-apoptotic genes were inhibited by SOCS-3 [13].

In the present study we have investigated whether SOCS-3 affects cytokine-induced beta cell chemokine and Fas expression, thereby potentially reducing the inflammatory response to beta cells and the susceptibility to T cell-mediated killing.

Methods

Cell culture and cytokine exposure

The generation of the Socs3-inducible cell line INS-r3#2 has been described previously [34]. Cells were cultured in RPMI-1640 with glutamax-I (Gibco BRL, Paisley, Scotland, UK) supplemented with 10% (vol./vol.) heat-inactivated TeT System Approved FCS (CLONTECH, Palo Alto, CA, USA), penicillin, streptomycin, 50 μmol/l β-mercaptoethanol (Sigma Aldrich, St Louis, MO, USA), hygromycin (100 μg/ml) and geneticin (100 μg/ml; Gibco BRL). For analysis of the effect of SOCS-3, doxycycline (Sigma Aldrich) was added and the cells cultured for 24 h to allow Socs3 expression. Neonatal rat islets were isolated and single cells were prepared by trypsin treatment as described [35]. Cells (150,000/well) were cultured in extracellular matrix coated 6-well dishes (Biological Industries, Kibbutz Beit, Haemek, Israel) using RPMI 1640 with glutamax-I supplemented with 10% (vol./vol.) heat-inactivated FCS and 1% (wt/vol.) penicillin and streptomycin as described [35]. Recombinant mouse IL-1β was obtained from BD Pharmingen (#554577; San Diego, CA, USA). Recombinant rat IFN-γ was obtained from R&D Systems (585-IF; R&D Systems, Foster City, CA, USA).

Analysis of chemokine and Fas mRNA expression using real-time PCR

For the analysis of chemokine mRNA-expression in INS-1 cells, cells were seeded in six-well plates. The following day 1 μg/ml doxycycline was added to half of the wells, and cells were incubated in medium containing 0.5% (vol./vol.) FCS and cultured overnight. After 48 h cells were exposed to either 150 pg/ml IL-1β for 1, 2, 4, 6, 8, 16 or 24 h or to varying amounts of IL-1β (18.75 pg/ml–2.4 ng/ml) for 6 h. RNA extraction was performed by the TRIzol method (GibcoBRL, Invitrogen, Carlsbad, CA, USA). For analysis of primary rat islet cells, monolayers were transduced with adenovirus encoding Luciferase or SOCS-3 at a concentration of 5 × 108 plaque forming units/ml [36]. The virus titre used was selected by transduction of islet cells with a green fluorescent protein (GFP)-encoding adenovirus and using a concentration giving >95% GFP-positive cells. After 2 days the cells were stimulated with 150 pg/ml IL-1β for 6 h and RNA was isolated. cDNA synthesis was performed by TaqMan Reverse Transcription Reagents (808-0234; Applied Biosystems, Foster City, CA, USA) using 200 ng RNA. Each cDNA sample was subjected to two individual PCR analyses using primer-pairs for the gene in question and Sp-1 for the internal control. Primers used for the reaction were: Mcp-1 (also known as Ccl2): forward primer: 5′-ATC TGT GCT GAC CCC AAT AAG G-3′, reverse primer: 5′-CAC TTG GTT CTG GTC CAG TTT TC-3′; rat Icam1: forward primer: 5′-GCT CAC CTT TAG CAG CTC AAC A-3′, reverse primer: 5′-GTG GAG GCA TGC AGG GAT T-3′; Cx3cl1: forward primer: 5′-ACT TGC ACA GCC CAG ATC ATT-3′, reverse primer: 5′-CTG CGC TCT CAG ATG TAG GAA A-3′; St-38 (also known as Ccl20): forward primer: 5′-GCT TAC CTC TGC AGC CAG TCA-3′, reverse primer: 5′-TGT ACG TGA GGC AGC AGT CAA-3′; Mip2 (also known as Cxcl2): forward primer: 5′-GGA AGA ACA TGG GCT CCT GTA C-3′, reverse primer: 5′-TTC CTG GGT GCA GTT TGT TTC-3′; Mob-1 (also known as Cxcl10): forward primer: 5′-TCC CAC TAC AGC GTG ATG GA-3′, reverse primer: 5′-GCC TTG CTG CTG GAG TTA CTT T-3′; and Sp-1: forward primer: 5′-GGC TAC CCC TAC CTC AAA GG-3′, reverse primer: 5′-CAC AAC ATA CTG CCC ACC AG-3′ (DNA Technology, Århus, Denmark). Each reaction was amplified in a PCR Mastermix supplemented with the DNA binding dye SYBR Green (Applied Biosystems). The samples were run on ABI PRISM 7900 HT Taqman (Applied Biosystems) according to the following program: 95°C 10 min (95°C 15 s, 60°C 1 min) × 40; 95°C 15 s, 60°C 15 s, 95°C 15 s, 4°C hold.

For data analysis the ΔΔC(T) method was applied as described previously [37].

For assessment of Fas mRNA expression, cells were cultured in 60 mm dishes, 1 × 106 cells/3 ml medium. The next day 1 μg/ml doxycycline was added to the relevant cultures and after 24 h cells were exposed to 150 pg/ml or 1 ng/ml of IL-1β or a combination of IL-1β and IFN-γ (150 pg/ml IL-1β, 10 ng/ml IFN-γ,) for 6 or 24 h. Primers used for this PCR reaction were: forward primer: 5′-TGC ACC TCG TGT GGA CTT GA-3′, reverse primer: 5′- GGA ACT TTG TTT CTT GCA TT′-3 (DNA Technology, Århus, Denmark). Sp-1 was used as internal control as described above.

Transient transfection and Luciferase assay

INS-r3#2 cells were seeded in 24-well dishes. After 2 days cells were transiently transfected using SuperFect (Qiagen, Valencia, CA, USA) with a total of 2 μg DNA (0.2 μg of an internal control [pRL-TK]; Promega, Madison, WI, USA), 0.8 μg empty vector (pcDNA; Invitrogen, Carlsbad, CA, USA) and 1 μg promoter construct (pGL-Fas or p-Mcp-1 [14]). After 4 h the transfection mixture was removed and the cells were incubated in the presence or absence of 1 μg/ml doxycycline. After 24 h, cells were exposed to cytokines for the times indicated and Luciferase assay was performed according to the manufacturer’s instructions using the Dual-Luciferase Reporter Assay System (Promega). Luciferase activity was measured using a Luminometer (Berthold Technologies, Bad Wildbad, Germany).

Statistical analysis

All values are presented as means ± SEM. Statistical analysis was done by a Student’s t test. Significance was assumed at a p value of less than 0.05.

Results

The effect of SOCS-3 on IL-1β-induced expression of chemokines

To determine whether the upregulation of chemokines by IL-1β was inhibited by SOCS-3, expression of six chemokine mRNAs was investigated. INS-r3#2 cells were cultured in the presence or absence of doxycycline for 24 h to induce Socs3 expression and subsequently exposed to 150 pg/ml IL-1β for 6 h. Chemokine mRNA expression level was measured by real-time PCR. All investigated chemokines were upregulated following IL-1β stimulation for 6 h (Fig. 1a). The fold induction by IL-1β varied from 3.5-fold for Cx3cl1 to about 100-fold for St-38. Socs3 significantly inhibited IL-1β-induced expression of the chemokines Mcp-1 and Mip-2, but not of Cx3-c1, Mob-1, rat Icam1 or St-38. In order to test whether the observed effects of IL-1β and SOCS-3 on chemokine expression could also be observed in primary islet cells, primary neonatal rat islet cultures were transduced with recombinant adenovirus encoding SOCS-3. Neonatal rat islets were used because of the high efficiency of adenoviral transduction. The level of Socs3 mRNA expression achieved using adenoviral transduction was found to be two to three times higher than doxycycline-inducible systems. We analysed the expression levels of the three chemokines showing the highest induction by IL-1β in INS-1 cells. These chemokines Mcp-2, Mip-2 and St-38 were all induced by IL-1β in primary neonatal rat islet cells, with induction being reduced in cells producing SOCS-3 by 60% (Mcp-2), 75% (Mip-2) and 70% (St-38) (Fig. 1b) compared with primary cells transduced with a control Luciferase-expressing adenovirus. Next we analysed the effect of SOCS-3 on IL-1β-induced chemokines in INS-1 cells at different time points and concentrations. However, only data from chemokines that were inhibited by SOCS-3 are shown.

Fig. 1
figure 1

SOCS-3 inhibits IL-1β-induced Mip-2, Mcp-1 and St-38 expression. a INS-r3#2 cells were cultured for 24 h in the presence (white bars) or absence (black bars) of doxycycline (Dox; 1 μg/ml) to induce SOCS-3 expression. Cells were subsequently exposed to 150 pg/ml IL-1β for 6 h and total RNA was isolated. b–d Monolayers of rat islet cells were exposed for 48 h to adenovirus encoding Luciferase (Luc) or SOCS-3 and subsequently stimulated with (white bars) or without (black bars) 150 pg/ml IL-1β for 6 h, after which total RNA for Mcp-1 (b), Mip-2 (c) and St-38 (d) was isolated. Chemokine expression was quantified using real-time PCR and results are presented as mean fold induction by IL-1β. *p < 0.05; n = 7 for Mcp-1 and Cx3cl1, n = 5 for rat Icam1, n = 6 Mob-1, n = 9 for Mip-2 and St-38 (a), n = 2 (b–d)

Full size image

SOCS-3 inhibits IL-1β-induced Mip-2 mRNA expression in a time- and dose-dependent manner

In time-response experiments (Fig. 2a), transient Mip-2 induction by IL-1β was observed with mRNA-expression increasing by a maximum of 14-fold after 6 h, returning to threefold after 16 to 24 h. SOCS-3 significantly inhibited Mip-2 expression after 4, 6, 8 and 24 h of IL-1β exposure. Maximal inhibition (26%) was observed after 8 h. A dose-dependent increase in Mip-2 mRNA expression was observed (Fig. 2b) at all concentrations of IL-1β tested, with the maximum increase (70-fold) occurring at 1.2 ng/ml of IL-1β. SOCS-3 significantly inhibited Mip-2 expression in response to 18.75, 150 and 300 pg/ml, and 1.2 and 2.4 ng/ml of IL-1β, with maximal inhibition (45%) being registered at 18.75 pg/ml IL-1β.

Fig. 2
figure 2

SOCS-3 inhibits IL-1β-induced Mip-2 expression in a time- and dose-dependent manner. INS-r3#2 cells were cultured for 24 h in the presence (white bars) or absence (black bars) of doxycycline (Dox; 1 μg/ml) to induce SOCS-3-expression. Cells were subsequently exposed to 150 pg/ml IL-1β for the time indicated (a) or to the indicated concentrations of IL-1β for 6 h (b) and total RNA was isolated. Mip-2-expression was quantified using real-time PCR and results are shown as fold stimulation by IL-1β. *p < 0.05, n = 4–9

Full size image

SOCS-3 inhibits IL-1-induced Mcp-1 mRNA expression in a time- and dose-dependent manner

IL-1β induced transient expression of Mcp-1 (Fig. 3a). A 46-fold maximal increase in stimulation was reached after 6 h. SOCS-3 inhibited IL-1β-stimulated Mcp-1 mRNA expression at all time points except 2, 16 and 24 h. Maximum inhibition (86%) was seen at 1 h. In dose–response experiments IL-1β concentrations above 37.5 pg/ml significantly induced Mcp-1 mRNA expression (Fig. 3b) with a maximum induction of 160-fold at 2.4 ng/ml. SOCS-3 significantly inhibited expression of Mcp-1 mRNA at all concentrations except 37.5 pg/ml. The greatest inhibition (95%) was observed at 18.75 pg/ml.

Fig. 3
figure 3

SOCS-3 inhibits IL-1β-induced Mcp-1 expression in a time- and dose-dependent manner. INS-r3#2 cells were cultured for 24 h in the presence (white bars) or absence (black bars) of doxycycline (Dox; 1 μg/ml) to induce SOCS-3-expression. Cells were subsequently exposed to 150 pg/ml IL-1β for the time indicated (a) or to the indicated concentrations of IL-1β for 6 h (b) and total RNA was isolated. Mcp-1-expression was quantified using real-time PCR and results are shown as fold stimulation by IL-1β. *p < 0.05, n = 4–9

Full size image

SOCS-3 inhibits cytokine-induced Mcp-1 promotor activity in INS-1 cells

To investigate whether the regulation of chemokine mRNA expression by SOCS-3 was exerted at the transcriptional level, the ability of SOCS-3 to influence the activity of a cytokine-responsive Mcp-1 promoter construct was analysed. The chemokine Mcp-1 promoter was chosen for these experiments because of the marked induction of mRNA by IL-1β. When cells were exposed for 6 h to 150 pg/ml or 1 ng/ml IL-1β, a 1.6- and 2.6-fold increase, respectively was seen in promoter activity. SOCS-3 expression significantly inhibited the Mcp-1 promoter activity in response to 150 pg/ml or 1 ng/ml IL-1β (Fig. 4).

Fig. 4
figure 4

SOCS-3 inhibits cytokine-induced Mcp-1 promotor activity in INS-1 cells. INS-r3#2 cells were transfected with a Mcp-1 promoter–Luciferase plasmid and a Renilla control plasmid for 4 h and subsequently incubated overnight in the presence (white bars) or absence (black bars) of doxycycline (Dox; 1 μg/ml). The transfected cells were exposed to IL-1β for 6 h and lysed. The lysates were subjected to Dual-Luciferase assay to measure the activity of the promoter. Results are presented as promoter activity normalised to the Renilla internal control. *p < 0.05, n = 7

Full size image

SOCS-3 inhibits cytokine-induced Fas expression in beta cells

IL-1β at 150 pg/ml or 1 ng/ml caused significant induction of Fas mRNA expression at 6 or 24 h, but SOCS-3 did not affect this induction (Fig. 5). The combination of IL-1β and IFN-γ enhanced stimulation of Fas mRNA expression, which in turn was significantly reduced by SOCS-3.

Fig. 5
figure 5

SOCS-3 inhibits cytokine-induced Fas expression. INS-r3#2 cells were cultured in the presence (white bars) or absence (black bars) of doxycycline Dox; 1 μg/ml) to induce SOCS-3-expression. Cells were subsequently exposed to the cytokines indicated for 6 or 24 h and total RNA was isolated. Fas mRNA expression was quantified using real-time PCR. Results are shown as fold stimulation by IL-1β. *p < 0.05, n = 6

Full size image

SOCS-3 inhibits cytokine-induced Fas promoter activity in beta cells

To study whether the regulation of Fas expression by SOCS-3 was caused by a change at the transcriptional level, a Fas promoter–reporter analysis was performed (Fig. 6). Cells were exposed to either 1 ng/ml of IL-1β or a mixture of IL-1β and IFN-γ (1 ng/ml and 10 ng/ml, respectively) for 2, 4 or 6 h in the presence or absence of SOCS-3. IL-1β and IL-1β in combination with IFN-γ induced Fas promoter activity in a time-dependent manner. There was no difference in induction between IL-1β alone and the combination of IL-1β with IFN-γ. At all time points the promoter activity induced by IL-1β plus IFN-γ was reduced by SOCS-3. At two of the three time points (2 and 6 h) SOCS-3 also inhibited transcription induced by IL-1β alone.

Fig. 6
figure 6

SOCS-3 inhibits cytokine-induced Fas promotor activity. INS-r3#2 cells were transfected with a Fas promoter–Luciferase plasmid and a Renilla control plasmid for 4 h and subsequently incubated overnight in the presence (white bars) or absence (black bars) of doxycycline (Dox; 1 μg/ml). The transfected cells were exposed to IL-1β 150 pg/ml alone or IL-1β in combination with IFN-γ (1 ng/ml) as indicated for 2, 4 and 6 h and were then lysed. The lysates were subjected to Dual-Luciferase assay to measure the activity of the promoter. Results are shown as promoter activity normalised to Renilla internal control. *p < 0.05, n = 5

Full size image

Discussion

In this study, we investigated the expression patterns of six chemokines upon cytokine exposure and the ability of SOCS-3 to regulate this expression. First, we were able to show an increase in mRNA expression in response to cytokine exposure for all six chemokines in INS-1 cells. Regulation by IL-1β was verified in primary rat islet cells for Mcp-1, Mip-2 and St-38. Second, we showed that SOCS-3 inhibited the expression of two of these, Mcp-1 and Mip-2 in INS-1 cells, and in addition inhibited IL-1β-induced production of ST-38 in primary neonatal rat cells, with the reduction in mRNA expression of Mcp-1 being associated with reduced promoter activity. However, we were not able to show that SOCS-3 inhibited the cytokine-induced expression of chemokines Cx3cl1, Mob-1 and rat Icam1 in INS-1 cells.

Our study also shows induction of Fas mRNA expression upon cytokine exposure, both after 6 and 24 h, and that this expression was inhibited by SOCS-3.

Fas promoter activity was significantly inhibited when expression of SOCS-3 was induced in the cells. SOCS-3 inhibited the activity of the promoter when induced by IL-1β alone or by IL-1β + IFN-γ.

In an mRNA array study, it has previously been shown that the chemokines Cx3cl1, St-38, Mob-1, Mip-2 and the intracellular adhesion molecule rat Icam-1 are upregulated in response to cytokine exposure [13]. The array study also showed that this upregulation was significantly inhibited by SOCS-3. In the present study, using real-time PCR, we were able to confirm this inhibitory effect of SOCS-3 for the chemokine Mip-2.

The chemokine MCP-1 has previously also been shown to be upregulated by IL-1β [38]. The transcriptional regulation of Mcp-1 by cytokines has been characterised [39] showing an IL-1β-responsive enhancer region in rat Mcp-1 gene between −2,180 and −2,478. This region contains two NFκB sites, and mutation in either of these abrogated IL-1β-induced Mcp-1 promoter activity [40]. In the present study we show that SOCS-3 did indeed suppress IL-1β-induced Mcp-1 promoter-activity. All chemokines investigated in the present study are dependent on the transcription factor NFκB, and since SOCS-3 blocks this transcription factor, this might be the mechanism by which SOCS-3 inhibits chemokine expression.

Fas has been shown to be upregulated in rat beta cells following IL-1β stimulation, but not by IFN-γ [40]. Also in human [23] and in mouse islets [19] IL-1β stimulation resulted in increased production of FAS and associated beta cell apoptosis. In line with this, we observed here that Fas mRNA expression was increased upon cytokine exposure after both 6 and 24 h. Because it is known that NFκB is necessary for the transcription of Fas, it was not surprising that in the present study SOCS-3 inhibited expression of Fas as well as the activity of the Fas promoter.

Since all six chemokines examined in this study are dependent on the transcription factor NFκB for their expression, we would expect them all to be similarly inhibited. The differential effect of SOCS-3 on expression of the six chemokines could have been caused by different mechanisms. First, the number of experiments carried out for each chemokine might not have been large enough to show a small but statistically significant inhibition by SOCS-3. Thus, increasing the number of experiments might increase the likelihood of identifying a slight inhibition. Another possible explanation could be that the level of SOCS-3 produced by the cells upon doxycycline exposure is insufficient to inhibit the cytokine-induced chemokine expression of the four chemokines that did not show a statistically significant inhibition by SOCS-3 in INS-1 cells. In support of this hypothesis, we found that in primary islet cells Socs expression induced by adenoviral transduction was able to inhibit IL-1β-stimulated expression of Mip-2, Mcp-2 and St-38. With the concentration of adenovirus used in these studies, a two- to threefold higher expression of Socs is achieved compared with the levels observed using the doxycycline-inducible system in INS-1 cells. Finally, other pathways are involved in regulating expression of these chemokines in addition to the NFκB pathway. IL-1β also signals via protein kinase C and small G-proteins, and it is possible that some of these pathways influence expression of the chemokines even in the presence of SOCS-3.

All of the experiments in the present study only addressed expression of chemokines at the mRNA level. It could be interesting to investigate whether this inhibition seen at the mRNA level translates to the protein level. Additionally, using motility assays, it could be elucidated whether the number of attracted immune cells in a chamber can be reduced by inducing SOCS-3 expression in beta cells.

The presented data show that SOCS-3 is able to inhibit IL-1β-induced expression of the chemokines Mcp-1, St-38 and Mip-2 as well as of the death receptor Fas. The perspectives of such an inhibition are interesting, as the inflammatory response in beta cells and the susceptibility to T cell-mediated killing could potentially be inhibited, thus preventing the development of type 1 diabetes.