Inhibition of inflammatory CD4 T cell activity by murine liver sinusoidal endothelial cells

Inhibition of inflammatory CD4 T cell activity by murine liver sinusoidal endothelial cells

Research Article Inhibition of inflammatory CD4 T cell activity by murine liver sinusoidal endothelial cells Antonella Carambia1, , Christian Frenzel1...

949KB Sizes 0 Downloads 73 Views

Research Article

Inhibition of inflammatory CD4 T cell activity by murine liver sinusoidal endothelial cells Antonella Carambia1, , Christian Frenzel1, , Oliver T. Bruns2, Dorothee Schwinge1, Rudolph Reimer2, Heinrich Hohenberg2, Samuel Huber1, Gisa Tiegs3, Christoph Schramm1, Ansgar W. Lohse1, Johannes Herkel1,⇑ 1

Department of Medicine I, University Medical Centre Hamburg-Eppendorf, Hamburg, Germany; 2Heinrich-Pette Institute, Leibniz Institute for Experimental Virology, Hamburg, Germany; 3Experimental Immunology and Hepatology, University Medical Centre Hamburg-Eppendorf, Hamburg, Germany

Background & Aims: The liver can mitigate the inflammatory activity of infiltrating T cells by mechanisms that are not entirely clear. Here we investigated the role of liver sinusoidal endothelial cells (LSECs) in regulating the activity of inflammatory CD4 T cells. Methods: Interactions between T helper (Th) 1 or Th17 cells and LSEC were studied by intravital microscopy and by in vitro stimulation assays. Results: Circulating CD4 T cells established lasting and repeated interactions with liver endothelium in vivo. Stimulation of Th1 and Th17 cells by LSEC greatly inhibited their capacity to secrete interferon-c or interleukin-17 in vitro; in contrast, stimulation by dendritic cells (DCs) resulted in considerable secretion of both cytokines. Cytokine release by Th1 or Th17 cells seemed to be actively suppressed by LSEC, as indicated by the inhibition of cytokine secretion even in the presence of Th1- and Th17promoting DC. This inhibition of CD4 T cell effector function seemed to depend on the dominance of inhibitory over activating co-stimulatory signals on LSEC, since (1) cytokine secretion could be restored by increased CD28 co-activation; (2) LSEC from interleukin-10/ mice, which manifest increased activating signals, such as MHC II, and decreased inhibitory signals, such as PD-L1, failed to suppress cytokine secretion; and (3) cytokine secretion by Th1 or Th17 cells that lacked PD-1, the ligand for inhibitory PD-L1, could not be suppressed by LSEC. Conclusions: LSEC inhibit inflammatory cytokine secretion of Th1 and Th17 effector CD4 T cells in dependence of interleukin10 and PD-1. Ó 2012 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved.

Keywords: CD4 T cells; Inflammation; Immune regulation; Tolerance; Interleukin-10; PD-1. Received 23 March 2012; received in revised form 4 September 2012; accepted 6 September 2012; available online16 September 2012 ⇑ Corresponding author. Address: Department of Medicine I, University Medical Centre Hamburg-Eppendorf, Martinistr. 52, 20246 Hamburg, Germany. Tel.: +49 40 7410 59736; fax: +49 40 7410 58014. E-mail address: [email protected] (J. Herkel).   These authors contributed equally to this work. Abbreviations: DC, dendritic cell; LSEC, liver sinusoidal endothelial cell; MFI, mean fluorescence intensity; MHC, major histocompatibility complex; Th cell, T helper cell; IL, interleukin; PD-1, programmed death-1; IFN-c, interferon-c.

Introduction The liver features a unique immune microenvironment, which is believed to favour immune tolerance [1–3]. Indeed, antigen delivered to the portal vein [4], or allografts co-transplanted with the allogeneic liver [5,6] are not attacked by the immune system. Hepatic tolerance induction can also be functional outside of the liver, indicated by the finding that the ectopic expression of a neuroantigen in the liver can prevent autoimmune neuroinflammation [7]. However, the mechanisms of hepatic tolerance induction are not entirely clear. A major cell type responsible for the induction of immune tolerance within the liver seems to be the liver sinusoidal endothelial cell (LSEC) [8]. LSEC have been reported to efficiently take up antigens from portal blood, which they cross-present to CD8 T cells, and thereby induce antigen-specific CD8 T cell tolerance [9–12]. The induction of tolerance in CD8 T cells by LSEC seems to depend on signals delivered by PD-L1 expressed on LSEC [13]. Moreover, LSEC may not only tolerize CD8 T cells by direct interaction, but can downmodulate the capability of adjacent dendritic cells to fully activate cytotoxic CD8 T cell responses [14]. In contrast to the established role of LSEC in inducing CD8 T cell tolerance, the capacity of the liver, and LSEC in particular, to induce CD4 T cell tolerance is less well defined. Priming of CD4 T cells by LSEC is rather inefficient and seems to favour Th2 responses instead of Th1 responses [15,16]. Moreover, a recent study has reported that LSEC can prime naive CD4 T cells to become a Foxp3-negative regulatory cell type, which suppresses inflammatory CD4 T cell responses and is protective in a model of T cell-mediated hepatitis [17]. However, the liver also seems to maintain the capacity to control the activity of circulating inflammatory CD4 T cells that had been primed in lymphatic tissue. Indeed, we previously observed that circulating inflammatory CD4 T cells, which recognise a shared autoantigen in nervous tissue and the liver, did not cause hepatitis, although these cells that include both Th1 and Th17 cells, had the capacity to induce severe neuroinflammatory disease [7]. This finding is surprising, because both Th1 and Th17 cells are important mediators of inflammatory liver diseases [18,19]. Thus far, it is not entirely clear how the liver can control inflammatory CD4 T cells that migrate to the liver.

Journal of Hepatology 2013 vol. 58 j 112–118

JOURNAL OF HEPATOLOGY In the present study, we demonstrate that LSEC can suppress the secretion of inflammatory cytokines of both Th1 and Th17 cells without inhibiting their proliferation. Inflammatory cytokine secretion of CD4 T cells was actively suppressed by LSEC even in the presence of Th1- and Th17-promoting dendritic cells. We show that LSEC acquire this suppressive phenotype in response to interleukin-10 through regulation of the expression of co-stimulatory molecules. Moreover, we demonstrate that the suppression of cytokine secretion critically depends on a PD-1 signal delivered to T cells. These findings indicate that LSEC are capable of mitigating hepatic inflammatory CD4 T cell activity, which may be relevant for preventing or terminating hepatic inflammation.

ture. On the third day of culture, 50 U/ml recombinant human IL-2 (Chiron Corp., Emeryville, CA, USA) was added. Cell culture supernatants were analysed for levels of IL-10 and IL-12 by ELISA (Peprotech, Hamburg, Germany, and eBioscience, Frankfurt, Germany) at day 4 of culture. To determine CD4 T cell effector function, CD4 T cells were restimulated after 6 days of culture. To that end, the number of vital CD4 T cells was assessed by trypan blue staining, and T cells were restimulated at densities of 5  105 per 500 ll on plate-bound anti-CD3 (3 lg/ml). After one additional day, the secreted cytokines IFN-c and IL-17 were determined by standard ELISA (R&D Systems); the results shown are representative of at least three similar experiments. Alternatively, co-cultures of CD4 T cells together with LSEC and DC were seeded in transwells (Millipore, Darmstadt, Germany), in which LSEC (1  106) were separated from DC (2.5  105) and CD4 T cells (2  106) to prevent direct cell contact. Flow cytometry

Inbred mice of the C57BL/6 or B10.PL strains, B6.129P2-Il10tm1Cgn/J mice of C57BL/6 background that are interleukin-10 deficient [20], and PD-1/ mice of C57BL/6 background [21] were bred and kept at specific pathogen-free conditions. The PD-1 knockout mice were kindly donated by T. Honjo. Age-matched (6–10 week old) mice were used in accordance with governmental animal experimentation guidelines; the experiments were approved by the local animal experimentation committee.

For analysis of T cell proliferation, CD4 T cells were stained with CFSE (2 lM, Life Technologies) before culture. After 3 days of culture, CFSE dilution was measured by flow cytometry. Immunofluorescent surface staining of CD4 T cells was performed with antibodies to CD4, CD44, and CD69. For intracellular cytokine staining, cells were treated with Golgi Plug (1 ll/ml, BD Biosciences) and restimulated with PMA (50 ng/ml) and ionomycin (1 lg/ml, both from Sigma) for 5 h. Fixed cells were then perforated in buffer containing Saponin/BSA (0.5%/2%; Sigma) and stained for IFN-c and IL-17. All antibodies for T cell staining were purchased from BioLegend. LSEC were stained with antibodies to CD146 (Miltenyi Biotec), CD86 (BioLegend), CD80 (BioLegend) MHC-II (BioLegend) and PD-L1 (eBioscience). Flow cytometry data were analysed with FlowJo 7.5 (Tree Star, Ashland, OR, USA) software.

Cell isolation

In vivo CD4 T cell tracking in the liver

For isolation of LSEC, mouse livers were perfused with 0.05% collagenase IV (Sigma–Aldrich, Taufkirchen, Germany) in Gey’s balanced salt solution, mechanically dissected and further digested in 0.05% collagenase IV in Gey’s balanced salt solution for 25 min at 37 °C at constant rotation (240 rpm). Hepatocytes and debris were sedimented twice at 40g and non-parenchymal cells were recovered by centrifugation over a 17% Optiprep (Sigma–Aldrich) gradient at 400g. LSEC were purified from the non-parenchymal cells as described [22] by magnetic sorting with the ME-9F1 antibody (Miltenyi Biotec, Bergisch-Gladbach, Germany). LSEC were seeded on collagen-coated culture plates (Falcon Primaria, BD Biosciences, Heidelberg, Germany). After overnight culture in Iscove’s modified Dulbecco’s medium supplemented with 5% FCS, penicillin (100 U/ml), and streptomycin (100 lg/ml; all from Life Technologies, Darmstadt, Germany), nonadherent cells were removed by medium exchange. As indicated, LSEC were cultured for 1 day in the presence or absence of blocking IL-10 antibody (2.5 lg/ml; R&D Systems; Wiesbaden, Germany), followed by cell lysis in TRIzolÒ (Life Technologies) and reverse transcription of RNA with a cDNA synthesis kit (Roche, Mannheim, Germany), and analysed by quantitative RT-PCR for gene expression using specific TaqMan single tube assays from Life Technologies. CD4+ T cells were isolated from spleen cell suspension by positive magnetic cell separation with anti-CD4 immunomagnetic beads (Miltenyi Biotec) and were further depleted of contaminating APC by Dynabeads (Life Technologies) labelled with anti-CD8, B220, and CD11b. For isolation of DC, cells were labelled with anti CD11c immunomagnetic beads (Miltenyi Biotec) and magnetically separated.

Recipient mice were narcotized and prepared for intravital liver microscopy [23] on a Nikon A1R microscope (Nikon, Düsseldorf, Germany) in collaboration with the ‘Nikon-Applikationszentrum Norddeutschland (Nikon GmbH)’ at the Heinrich-Pette-Institute. CD4 T cells were purified from the spleen as described above and stained with DiD (Life Technologies). Subsequently, cells were injected (5  106) into the tail vein of recipient mice and CD4 T cell accumulation within the liver sinusoids and interaction with LSEC were recorded at 30 frames per second for up to 1 h, following cell injection, by confocal multicolour video-rate imaging. Alternatively, intravital microscopy was performed 2 days after T cell injection.

Materials and methods Mice

CD4 T cell differentiation To induce inflammatory effector CD4 T cells, primary CD4 T cells (5  105 per well) were cultured in the presence of splenic DC (5  104) and antibody to CD3 (1 lg/ml; BioLegend, Fell, Germany) for 4 days in IMDM/5% FCS. Differentiated CD4 T cells were then re-stimulated twice either by DC or LSEC and antibody to CD3. Twenty-four hours after the first or second restimulation, IFN-c, IL-17, and IL-22 levels in cell culture supernatants were measured by standard ELISA (IFN-c, IL-17: R&D Systems; IL-22: Antigenix America, Huntington Station, NY).

Co-cultures of CD4 T cells, DC, and LSEC CD4 T cells (5  105 per well) were stimulated with 2 lg/ml anti-CD3 antibody (BioLegend) on 5  104 splenic DC and/or 1  105 LSEC in quadruplicates. As indicated, anti-CD28 antibody (2 lg/ml, BioLegend) was added at the start of the cul-

Statistics Differences between experimental groups were assessed for statistical significance by Mann–Whitney test.

Results Interactions of circulating CD4 effector T cells and LSEC in vivo First, we aimed at confirming that circulating CD4 effector T cells do interact to a meaningful degree with LSEC in vivo. Therefore, we performed intravital liver microscopy of mice that were subsequently injected intravenously into the tail with 5  106 DiDlabelled CD4 effector T cells. Indeed, starting from 1 min after injection, CD4 T cells began to appear and accumulate in liver sinusoids (Supplementary movie file 1). The injected T cells rolled along the sinusoids at reduced velocity and several cells attached to the sinusoidal endothelium, typically for about 10–20 min, after which they detached and moved on again. Thus, interactions between LSEC and circulating CD4 T cells seemed to occur frequently and for enough time to allow for functional stimulation. Two days after injection, circulating T cells were still found to adhere to sinusoidal endothelium, indicating that repetetive

Journal of Hepatology 2013 vol. 58 j 112–118

113

Research Article interactions of re-circulating T cells with LSEC are likely to occur (Fig. 1A; upper panel: 3 min after injection; lower panel: 2 days after injection).

the presence of LSEC had almost no effect on DC-induced CD4 T cell proliferation; proliferation levels of CD4 T cells cultured with DC alone were comparable to those of CD4 T cells stimulated by LSEC and DC in co-culture (Fig. 2E).

LSEC mitigate inflammatory cytokine secretion by CD4 effector T cells To study the functional consequences of circulating CD4 T cell stimulation by LSEC, we first induced inflammatory CD4 effector T cells through stimulation of primary CD4 T cells on splenic dendritic cells [24]; and then tested whether these inflammatory CD4 T cells retained their effector function upon repeated restimulation by splenic dendritic cells or LSEC (Fig. 1B–F). Th1 cells, which were restimulated by splenic dendritic cells (dark blue bars), manifested sustained secretion of interferon-c after one (858 pg/ml) or two (1140 pg/ml) subsequent restimulations; in contrast; LSEC (light blue bars) did not sustain interferon-c secretion, resulting in a gradual loss of interferon-c production by Th1 cells after one (273 pg/ml) or two (4 pg/ml) subsequent re-stimulations (Fig. 1B, p = 0.0009). Likewise, Th17 cells lost their ability to produce IL-17 upon restimulation by LSEC (23 and 9 pg/ml after first and second restimulation), but not upon restimulation by DC (66 and 82 pg/ml after the first and second restimulation) (Fig. 1C, p <0.01). Moreover, LSEC-stimulated Th17 cells secreted low amounts of IL-22 (53 and 42 pg/ml after the first and second restimulation, p = 0.6612); in contrast, IL-22 secretion increased after repeated DC stimulation (79 vs. 396 pg/ml after the first or second restimulation, respectively, p = 0.0286) (Fig. 1D). The respective master transcription factors T-bet (Fig. 1E) and RORct (Fig. 1F) were expressed at low levels by LSEC-stimulated T cells, whereas the expression of these transcription factors increased considerably upon repeated DC stimulation. The failure of LSEC to sustain CD4 T cell effector activity was not a consequence of insufficient stimulation of Th1 and Th17 cells, since LSEC-stimulated T cells had upregulated activation marker expression to similar degree as DC-stimulated T cells (MFI CD44: 9031 vs. 10337; and MFI CD69: 468 vs. 669; p >0.05). Active suppression of DC-induced Th1 and Th17 cytokine secretion by co-cultured LSEC To confirm that the mitigation of inflammatory CD4 T cell activity was actively induced by LSEC rather than the result of insufficient stimulation, we stimulated CD4 T cells in co-cultures of splenic DC and LSEC (Fig. 2). We found that LSEC alone could not induce significant IFN-c (209 pg/ml) or IL-17 (17 pg/ml) secretion, and that DC alone induced robust secretion of interferon-c (1011 pg/ml) and IL-17 (73 pg/ml). However, DC-induced cytokine secretion was almost completely abrogated in the presence of co-cultured LSEC (102 pg/ml IFN-c, p = 0.0002; 8 pg/ml IL-17, p = 0.0286). Thus, LSEC indeed seem to be capable of actively suppressing DC-induced Th1 (Fig. 2A) and Th17 (Fig. 2B) activity. Intracellular cytokine staining of co-cultured CD4 T cells further confirmed the suppressive capacity of LSEC on DC-induced CD4 T cell cytokine secretion. Despite intense PMA/ionomycin restimulation, CD4 T cells, which were previously stimulated on DC and LSEC co-cultures, showed impaired IFN-c (10.0% vs. 18.8%; p = 0.0286; Fig. 2C) and IL-17 production (0.7% vs. 1.7%; p = 0.0286; Fig. 2D) compared to CD4 T cells activated by DC alone. We next tested whether the proliferative capacity of DC-stimulated CD4 T cells could also be inhibited by LSEC. Surprisingly,

114

Suppression of DC-induced Th1 and Th17 activity by LSEC depends on cell contact, not on impaired interleukin-12 or excessive IL-10 secretion To determine whether LSEC require cell contact to suppress DC-induced cytokine secretion by inflammatory CD4 T cells, we performed transwell experiments, in which we seeded LSEC at the bottom of culture plates and incubated CD4 T cells and DC within a transwell insert; as control, the CD4 T cell and DC culture was performed in a transwell insert in the absence of LSEC at the bottom of the well. We found that the DC-induced interferon-c and IL-17 secretion by the stimulated CD4 T cells (9891 pg/ml IFN-c; 2444 pg/ml IL-17; Supplementary Fig. 1A and B, black columns) was only marginally reduced by the presence of LSEC in the transwell (8208 pg/ml IFN-c; 2385 pg/ ml IL-17; Supplementary Fig. 1A and B, grey columns). Thus, the suppression of DC-induced interferon-c and IL-17 secretion did not seem to depend on a mediator that was secreted by LSEC, but rather required cell contact. To determine whether LSEC may impair the Th1-promoting function of DC, we assessed the secretion of interleukin-12 in its bioactive heterodimeric form (p70) after 4 days of culture (Supplementary Fig. 2A); interleukin-12 is the major stimulator of Th1 differentiation. Interleukin-12 (p70) levels in cultures of DC alone (385 pg/ml) were only slightly reduced when LSEC were present (315 pg/ml). Thus, the suppressive effect of LSEC on Th1 effector functions appeared not to be mediated by inhibiting interleukin-12 secretion of DC. We tested whether elevated IL-10 levels in LSEC/DC co-cultures might contribute to the suppression of DC-induced CD4 T cell inflammatory cytokine secretion by LSEC. However, IL-10 levels measured in supernatants of CD4 T cells cultured together with LSEC and DC were only slightly higher compared to cultures of CD4 T cells and DC alone (689 vs. 557 pg/ml; p = 0,0571; Supplementary Fig. 2B). Thus, inhibition of DCinduced CD4 T cell effector activity seems not to be mediated by anti-inflammatory IL-10. Taken together, these experiments show that the suppressive capacity of LSEC on CD4 T cell effector function seems to depend on direct cell-contact rather than soluble mediators. LSEC-mediated suppression of DC-induced inflammatory CD4 T cell activity is modified by co-signalling molecules Since the suppression of DC-induced inflammatory cytokine secretion by LSEC seemed to depend on cell contact, it was possible that it was mediated by co-stimulatory molecules. Indeed, LSEC have been reported to express the co-activating molecules CD80 and CD86 at relatively low levels, and the co-inhibitory molecule PD-L1 at relatively high levels [13]. Therefore, we hypothesised that LSEC-mediated suppression of DC-activated inflammatory CD4 T cell cytokine responses could be due to insufficient co-activation of CD4 T cells in the presence of LSEC. To test this hypothesis, we added agonistic anti-CD28 antibody to CD4 T cell cultures with LSEC and/or DC. Indeed, addition of co-activating anti-CD28 antibody restored the IFN-c (1657 pg/

Journal of Hepatology 2013 vol. 58 j 112–118

JOURNAL OF HEPATOLOGY A

LSEC DC

B ***

Restimulation 1

Restimulation 2

250 500 750 1000 1250 IFN-γ [pg/ml]

C

*

0

n.s.

Restimulation 1

***

Restimulation 2

3 min

LSEC DC

E

0.0 0.2 0.4 0.6 0.8 1.0 TBX21 expression normalized to HPRT

LSEC DC **

Restimulation 1

2d

**

Restimulation 2 20

D

40 60 80 IL-17 [pg/ml]

*

Restimulation 2 0

Restimulation 2

**

LSEC DC

n.s.

Restimulation 1

Restimulation 1

**

0

LSEC DC

F

0.00 0.05 0.10 0.15 RORC expression normalized to HPRT

100 200 300 400 IL-22 [pg/ml]

Fig. 1. Functional interactions between liver endothelium and circulating CD4 T cells. (A) Intravital microscopy of mouse livers 3 min or 2 days after intravenous injection of labelled CD4 T cells (green); nuclei stained blue, autofluorescence in red. Secretion of (B) IFN-c, (C) IL-17 and (D) IL-22 was determined after restimulation of Th1 and Th17 cells on LSEC or DC. Expression of master transcription factor genes (E) Tbx21 and (F) RORC was determined after restimulation of Th1 and Th17 cells on LSEC or DC. ⁄⁄⁄p <0.0005; ⁄⁄p <0.005; ⁄p <0.05; n.s., not significant.

ml) and IL-17 (649 pg/ml) cytokine responses of CD4 T cells activated by DC in the presence of LSEC to levels comparable to those of CD4 T cells stimulated by DC alone (IFN-c: 1221 pg/ml, IL-17: 832 pg/ml) (Fig. 3A and B). Moreover, in cultures of CD4 T cells with LSEC alone, supplementation of CD28-co-activation led to the induction of substantial IFN-c (947 pg/ml) and IL-17 (287 pg/ml) cytokine responses (Fig. 3A and B). These results indicated that the impaired inflammatory cytokine production by CD4 T cells was induced by dominance of co-inhibition over co-activation supplied by LSEC. Since the relatively low expression by LSEC of the activating molecules CD80 and CD86 has been reported to be a consequence of their exposure in vivo to interleukin-10 [25], we compared expression levels of activating and inhibitory molecules on LSEC from IL-10 knockout and wild type mice. Indeed, we found that LSEC isolated from IL-10 deficient mice manifested higher mean fluorescence intensity (MFI) levels of CD40 (837 vs. 620; p = 0.0286), CD80 (2657 vs. 1309; p = 0.0286) and CD86 (1265 vs. 571; p = 0.0286) than LSEC from wild type mice. Even more pronounced differences could be seen in the expression levels of MHC II molecules, which were considerably upregulated on LSEC from IL-10 deficient mice, and in those of PD-L1 molecules, which were downregulated on LSEC from IL-10 deficient mice (Fig. 3C). Thus, exposure of LSEC to IL-10 seemed to shift expression of positive and negative co-stimulators to a rather inhibitory than activating phenotype. To confirm this notion, we cultured LSEC from wild type mice for 1 day in the presence or absence of a blocking antibody to IL-10 and then determined by quantitative RT-PCR the expression levels of the MHC II gene H2Aa and of the CD274 gene, which encodes PD-L1 (Fig. 3D). We found that inhibition of IL10 signaling significantly increased expression of MHC II and significantly decreased expression of PD-L1, confirm-

ing that exposure to IL-10 is responsible for the dominance of coinhibitory signals on LSEC. Therefore, we hypothesised that IL-10-conditioning was essential for the suppressive activity of LSEC. To address this issue, we co-cultured LSEC from mice deficient in interleukin10 together with CD4 T cells and DC. Indeed, IL-10/ LSEC were not able to suppress interferon-c secretion by DC-stimulated CD4 T cells (Fig. 3E, white column) compared to CD4 T cells stimulated by DC alone (dark blue column). However, when IL-10/ LSEC were preincubated with recombinant IL-10 prior to the coculture, the capacity of LSEC to suppress DC-mediated IFN-c secretion by CD4 T cells was at least partly restored (Fig. 3E, white column + IL-10). Taken together, these results show that IL-10 signalling modulates expression of accessory molecules on LSEC and is therefore required for the suppressive capacity of LSEC. LSEC-mediated suppression of DC-induced CD4 T cell effector activity depends on PD-1 signalling To confirm the dependence on co-inhibitory molecules, such as PD-L1, of LSEC-induced inhibition of cytokine responses, we tested whether suppression of CD4 T cell effector activity may depend on a PD-1 signal; PD-1 is the major receptor for PD-L1 and a negative regulator of T cell responses [21]. Therefore, we isolated CD4 T cells from PD-1 knockout mice or wild type C57BL/6 mice and stimulated these cells on wild-type DC and LSEC, alone or in co-culture. The DC-induced interferon-c secretion of wild type CD4 T cells (Fig. 3F, dark blue column) was significantly reduced in the presence of LSEC (white column) (521 pg/ml vs. 1347 pg/ml; p = 0.0286); in contrast, the DCinduced interferon-c secretion by PD-1 knockout CD4 T cells

Journal of Hepatology 2013 vol. 58 j 112–118

115

Research Article LSEC LSEC + DC DC

**

* ***

400

D

60

LSEC

LSEC + DC

DC

*

20

LSEC LSEC + DC DC

40 20

10

* * 1.2 0.8 0.4 0.0

0 IL-17 LSEC

LSEC + DC

DC % Undivided

E

*

LSEC LSEC + DC DC

IFN-γ

200 0

DC

0

80

IL-17 [pg/ml]

IFN-γ [pg/ml]

600

LSEC + DC

*

1000 800

LSEC

CD4

1200

C

% IFN-γ

B

% IL-17

LSEC LSEC + DC DC

CD4

A

40 30 20 10 0

*** *** LSEC LSEC + DC DC

Fig. 2. LSEC actively suppress DC-induced inflammatory cytokine production of CD4 T cells. Secretion of (A) IFN-c or (B) IL-17 by CD4 T cells stimulated by LSEC, DC or LSEC/DC co-cultures. Intracellular (C) IFN-c and (D) IL-17 staining of CD4 T cells stimulated by LSEC and/or DC. Representative stainings and mean and SEM of four samples are shown. (E) DC-induced proliferation of CFSE-labelled CD4 T cells is not mitigated by LSEC. Representative stainings of T cells stimulated by LSEC, LSEC/DC co-culture or DC, and mean and SEM of four samples of each culture are shown.

(dark blue column -/-) was not significantly impaired by the presence of LSEC (white, -/- striped column) (1222 pg/ml vs. 1319 pg/ ml; p = 0.6857). Thus, the suppression of DC-induced CD4 cytokine secretion by LSEC seemed to be mediated through a PD-1 signal.

Discussion Inflammatory CD4 T cell responses have been implicated in the pathogenesis of a variety of liver diseases such as viral hepatitis or autoimmune disorders [1–3]. However, in healthy homoeostasis, the outcome of CD4 T cell stimulation in the liver or by liver cells is often tolerance [8,26]. It has been reported that Kupffer cells [27] and hepatic stellate cells [28,29] may contribute to hepatic tolerance by prostaglandin secretion or induction of CD4 T cell apoptosis, respectively. In this study, we examined the role of LSEC in regulating the inflammatory activity of CD4 T cells. When we tracked labelled CD4 T cells after intravenous injection by intravital liver microscopy, we could observe rapid accumulation of CD4 T cells in the sinusoids that seemed to establish repeated and lasting interactions with the sinusoidal endothelium (Supplementary movie file 1; Fig. 1A). To test the functional relevance of repeated inflammatory CD4 T cell stimulation by LSEC, we restimulated Th1 and Th17 effector cells either on LSEC or on DC. We found that repeated stimulation by LSEC, but not by DC, led to an abrogation of inflammatory interferon-c and IL-17 responses in differentiated effector CD4 T cells (Fig. 1B and C). These findings are in concordance with previous observations, that LSEC do not prime inflammatory Th1 responses [15,17], and do not support the expansion of Th1 cells [16,17]. However, in contrast to the reports by others [16,17], we could not observe an outgrowth of IL-4 producing CD4 T cells (data not shown). This 116

discrepancy might be explained by the different genetic background of the mouse strains used (C57BL/6 vs. Balb/c). The failure of LSEC to sustain secretion of interferon-c and IL17 by CD4 T cells was not a result of T cell death, as assessed by trypan blue staining before restimulation (not shown). Moreover, CD4 T cell activation markers CD44 and CD69 were upregulated to similar degrees after repeated stimulation by LSEC or DC, indicating that the absence of inflammatory cytokine production in LSEC-stimulated CD4 T cell cultures was not due to poor CD4 T cell activation by LSEC. In addition, we show that LSEC not only fail to sustain inflammatory CD4 T cell responses, but may also actively suppress CD4 T cell effector function in co-cultures with Th1- and Th17-promoting DC (Fig. 2A–D). These findings are in agreement with a recent publication, in which LSEC have been shown to suppress DC-induced cytotoxic CD8 T cell responses [14]. However, in contrast to the reported suppression of CD8 T cell proliferation and cytotoxicity by LSEC, LSEC seemed to inhibit only CD4 T cell effector function, but not DC-induced CD4 T cell proliferation (Fig. 2E). In our experiments, LSEC-mediated suppression of Th1 and Th17 responses did not appear to depend on soluble factors such as IL-12, but rather on cell contact (Supplementary Fig. 1 and 2). Indeed, inhibition of cytokine secretion seemed to be constrained by the balance between inhibitory and activating co-stimulatory molecules, as indicated by the restoration of cytokine secretion when adding agonistic antibody to CD28 to our cultures (Fig. 3A and B), or when using CD4 T cells that lack the inhibitory receptor PD-1 (Fig. 3F). This finding is in accordance with a recent study reporting a key role of PD-L1 on LSEC in inducing CD8 T cell tolerance [13]. It was previously reported, that surface expression of co-stimulatory molecules as well as MHC class II could be downregulated on LSEC by IL-10 [13]. We also find increased expression of CD40, CD80, CD86 and MHC class II molecules on LSEC

Journal of Hepatology 2013 vol. 58 j 112–118

JOURNAL OF HEPATOLOGY B

1500 1200 900 600 300 0 Anti-CD28

**

- +

E

PD-L1

1250 1000 750 500 250 0 IL-10 - +

*

250

Conflict of interest - +

0.02 0.01

0.00 IL-10 Ab

-

F

- +

- +

PD-L1 *** 2.0

The authors who have taken part in this study declared that they do not have anything to disclose regarding funding or conflict of interest with respect to this manuscript.

1.5 1.0

Supplementary data

0.5 0.0

+

-

+

LSEC LSEC + DC DC

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jhep.2012.09. 008. References

n.s.

n.s. * *

- +

**

500

H2Aa ** 0.03

IFN-γ [pg/ml]

IFN-γ [pg/ml]

LSEC (IL10-/-) LSEC (IL10-/-) + DC DC

H2Aa expression normalized to HPRT

D

MHC II

1000 750

0 Anti-CD28

- +

C

This study was supported by the Deutsche Forschungsgemeinschaft (SFB 841) and by the Federal Ministry of Education and Research (TOMCAT).

n.s.

n.s.

- +

Financial support

LSEC LSEC + DC DC

CD274 expression normalized to HPRT

IFN-γ [pg/ml]

LSEC LSEC + DC DC ***

IL-17 [pg/ml]

A

-

1750 1500 1250 1000 750 500 250 0 PD-1

* *

*

+/+

-/-

+/+ -/-

+/+ -/-

Fig. 3. The suppressive capacity of LSEC depends on IL-10 exposure and coinhibitory molecules. Secretion of (A) IFN-c and (B) IL-17 by LSEC- or LSEC/DCstimulated CD4 T cells was restored by addition of stimulating anti-CD28 antibody. (C) Comparison of accessory molecules MHC II and PD-L1 on LSEC from IL-10 deficient and wild type mice. (D) Comparison of MHC II and PD-L1 gene expression by wild type LSEC after in vitro culture in the presence or absence of blocking IL-10 antibody. (E) LSEC from IL-10/ mice failed to inhibit DC-induced cytokine secretion by CD4 T cells; pre-incubation of IL-10/ LSEC with recombinant IL-10 (10 ng/ml) for 1 day partially restored inhibition of cytokine secretion. (F) LSEC inhibit DC-induced cytokine secretion of CD4 T cells from wild type mice, but not of CD4 T cells from PD-1 deficient mice. ⁄⁄⁄p <0.0005; ⁄⁄ p <0.005; ⁄p <0.05; n.s., not significant.

from IL-10 deficient mice. Moreover, expression of PD-L1, the major ligand for PD-1, is decreased on LSEC from IL-10 deficient mice (Fig. 3C). Furthermore, we show that exposure of LSEC from wild-type mice to IL-10 is instrumental for the dominant expression of inhibitory over activating co-stimulatory signals by LSEC (Fig. 3D). These latter findings are in agreement with a recent report that links IL-10 signalling to the expression of PDL1 on tolerogenic APC [30]. Moreover, these findings seem to mechanistically explain the link between the incapability of IL10 deficient LSEC to inhibit CD4 T cell cytokine secretion (Fig. 3E) and the insensitivity of PD-1 deficient CD4 T cells to LSEC-induced inhibition of cytokine secretion (Fig. 3F). Thus, IL10, which is abundantly produced in the liver, e.g., by Kupffer cells [31] or stellate cells [32], seems to be a major inducer of a tolerogenic phenotype in LSEC that allows them to inhibit the inflammatory activity of CD4 T cells. Therefore, hepatic tolerance seems to result from the coordinated activity of different liver cell types, including LSEC, Kupffer cells and stellate cells, each of which featuring complementary tolerance mechanisms.

[1] Adams DH, Eksteen B, Curbishley SM. Immunology of the gut and liver: a love/hate relationship. Gut 2008;57:838–848. [2] Racanelli V, Rehermann B. The liver as an immunological organ. Hepatology 2006;43:S54–S62. [3] Selmi C, Mackay IR, Gershwin ME. The immunological milieu of the liver. Semin Liver Dis 2007;27:129–139. [4] Cantor HM, Dumont AE. Hepatic suppression of sensitization to antigen absorbed into the portal system. Nature 1967;245:744–745. [5] Calne RY, Sells RA, Pena JR, Davis DR, Millard PR, Herbertson BM, et al. Induction of immunological tolerance by porcine liver allografts. Nature 1969;223:472–476. [6] Sriwatanawongsa V, Davies HS, Calne RY. The essential roles of parenchymal tissues and passenger leukocytes in the tolerance induced by liver grafting in rats. Nat Med 1995;1:428–432. [7] Lüth S, Huber S, Schramm C, Buch T, Zander S, Stadelmann C, et al. Ectopic expression of neural autoantigen in mouse liver suppresses experimental autoimmune neuroinflammation by inducing antigen-specific Tregs. J Clin Invest 2008;118:3403–3410. [8] Thomson AW, Knolle PA. Antigen-presenting cell function in the tolerogenic liver environment. Nat Rev Immunol 2010;10:753–766. [9] Smedsrod B. Clearance function of scavenger endothelial cells. Comp Hepatol 2004;3:S22. [10] Schurich A, Böttcher JP, Burgdorf S, Penzler P, Hegenbarth S, Kern M, et al. Distinct kinetics and dynamics of cross-presentation in liver sinusoidal endothelial cells compared to dendritic cells. Hepatology 2009;50:909–919. [11] von Oppen N, Schurich A, Hegenbarth S, Stabenow D, Tolba R, Weiskirchen R, et al. Systemic antigen cross-presented by liver sinusoidal endothelial cells induces liver-specific CD8 T-cell retention and tolerization. Hepatology 2009;49:1664–1672. [12] Limmer A, Ohl J, Kurts C, Ljunggren HG, Reiss Y, Groettrup M, et al. Efficient presentation of exogenous antigen by liver endothelial cells to CD8+ T cells results in antigen-specific T-cell tolerance. Nat Med 2000;6:1348–1354. [13] Diehl L, Schurich A, Grochtmann R, Hegenbarth S, Chen L, Knolle PA. Tolerogenic maturation of liver sinusoidal endothelial cells promotes B7homolog 1-dependent CD8+ T cell tolerance. Hepatology 2008;47:296–305. [14] Schildberg FA, Hegenbarth S, Schumak B, Limmer A, Knolle PA. Liver sinusoidal endothelial cells veto CD8 T cell activation by antigen-presenting dendritic cells. Eur J Immunol 2008;38:957–967. [15] Knolle PA, Schmitt E, Jin S, Germann T, Duchmann R, Hegenbarth S, et al. Induction of cytokine production in naive CD4(+) T cells by antigenpresenting murine liver sinusoidal endothelial cells but failure to induce differentiation toward Th1 cells. Gastroenterology 1999;116:1428–1440. [16] Klugewitz K, Blumenthal-Barby F, Schrage A, Knolle PA, Hamann A, Crispe IN. Immunomodulatory effects of the liver: deletion of activated CD4+ effector cells and suppression of IFN-gamma-producing cells after intravenous protein immunization. J Immunol 2002;169:2407–2413. [17] Kruse N, Neumann K, Schrage A, Derkow K, Schott E, Erben U, et al. Priming of CD4+ T cells by liver sinusoidal endothelial cells induces CD25low

Journal of Hepatology 2013 vol. 58 j 112–118

117

Research Article [18] [19]

[20] [21] [22]

[23]

[24]

[25]

118

forkhead box protein 3-regulatory T cells suppressing autoimmune hepatitis. Hepatology 2009;50:1904–1913. Lafdil F, Miller AM, Ki SH, Gao B. Th17 cells and their associated cytokines in liver diseases. Cell Mol Immunol 2010;7:25024–25025. Tiegs G. Cellular and cytokine-mediated mechanisms of inflammation and its modulation in immune-mediated liver injury. Z Gastroenterol 2007;45: 63–70. Kuhn R, Lohler J, Rennick D, Rajewsky K, Muller W. Interleukin-10-deficient mice develop chronic enterocolitis. Cell 1993;75:263–274. Iwai Y, Terawaki S, Ikegawa M, Okazaki T, Honjo T. PD-1 inhibits antiviral immunity at the effector phase in the liver. J Exp Med 2003;198:39–50. Schrage A, Loddenkemper C, Erben U, Lauer U, Hausdorf G, Jungblut PR, et al. Murine CD146 is widely expressed on endothelial cells and is recognized by the monoclonal antibody ME-9F1. Histochem Cell Biol 2008;129:441–451. Thiberge S, Blazquez S, Baldacci P, Renaud O, Shorte S, Ménard R, et al. In vivo imaging of malaria parasites in the murine liver. Nat Protoc 2007;2: 1811–1818. Wiegard C, Wolint P, Frenzel C, Cheruti U, Schmitt E, Oxenius A, et al. Defective T helper 1 response by hepatocyte-stimulated CD4 T cells impairs anti-viral CD8 response and viral clearance. Gastroenterology 2007;133: 2010–2018. Knolle PA, Uhrig A, Hegenbarth S, Löser E, Schmitt E, Gerken G, et al. IL-10 down-regulates T cell activation by antigen-presenting liver sinusoidal

[26] [27] [28]

[29]

[30]

[31]

[32]

endothelial cells through decreased antigen uptake via the mannose receptor and lowered surface expression of accessory molecules. Clin Exp Immunol 1998;114:427–433. Carambia A, Herkel J. CD4 T cells in hepatic immune tolerance. J Autoimmun 2010;34:23–28. You Q, Cheng L, Kedl RM, Ju C. Mechanism of T cell tolerance induction by murine hepatic Kupffer cells. Hepatology 2008;48:978–990. Ichikawa S, Mucida D, Tyznik AJ, Kronenberg M, Cheroutre H. Hepatic stellate cells function as regulatory bystanders. J Immunol 2011;186: 5549–5555. Dangi A, Sumpter TL, Kimura S, Stolz DB, Murase N, Raimondi G, et al. Selective expansion of allogeneic regulatory T cells by hepatic stellate cells: role of endotoxin and implications for allograft tolerance. J Immunol 2012; 188:3667–3677. Wölfle SJ, Strebovsky J, Bartz H, Sähr A, Arnold C, Kaiser C, et al. PD-L1 expression on tolerogenic APCs is controlled by STAT-3. Eur J Immunol 2011;41:413–424. Knolle P, Schlaak J, Uhrig A, Kempf P, Meyer zum Büschenfelde KH, Gerken G. Human Kupffer cells secrete IL-10 in response to lipopolysaccharide (LPS) challenge. J Hepatol 1995;22:226–229. Lee WC, Yu MC, Chiang YJ, Wang HC, Lu L, Qian S. Liver stellate cells suppress dendritic cells through IL-10. Transplant Proc 2005;37:10–11.

Journal of Hepatology 2013 vol. 58 j 112–118