IL-17A/F enable cholangiocytes to restrict T cell-driven experimental cholangitis via PD-1/PD-L1 interaction

Abstract

Background & Aims:

IL-17A-producing T cells are present in autoimmune cholestatic liver diseases; however, little is known about the contribution of IL-17 to periductal immune responses. Here we investigated the role of IL-17 produced by antigen specific CD8⁺ T cells in a mouse model of cholangitis and in vitro in human cholangiocyte organoids.

Methods:

K14-OVAp mice express a MHC I restricted ovalbumin peptide sequence (SIINFEKL) on cholangiocytes. Cholangitis was induced by the adoptive transfer of transgenic OVA-specific OT-1 CD8⁺ T cells that either had OT-1^wt or lacked IL-17A/F (OT-1^IL17ko). The response of mouse and human cholangiocytes/organoids to IL-17A was assessed in vitro.

Results:

Transfer of OVA-specific OT-1^IL17ko cells significantly aggravated periductal inflammation in K14-OVAp recipient mice, compared to transfer of OT-1^wt T cells. OT-1^IL17ko T cells were highly activated in the liver and displayed increased cytotoxicity and proliferation. IL-17A/F produced by transferred OT-1^wt CD8⁺ T cells induced upregulation of the inhibitory molecule PD-L1 on cholangiocytes, restricting cholangitis by limiting cytotoxicity and proliferation of transferred cells. In contrast, OT-1^IL17ko T cells failed to induce PD-L1 on cholangiocytes, resulting in uncontrolled expansion of cytotoxic CD8⁺ T cells and aggravated cholangitis. Blockade of PD-L1 after transfer of OT-1^wt T cells with anti-PD-L1 antibody also resulted in aggravated cholangitis. Using human cholangiocyte organoids, we were able to confirm that IL-17A induces PD-L1 expression in cholangiocytes.

Conclusion:

We demonstrate an important function of IL-17 in restricting cholangitis and protecting from CD8⁺ T cell mediated inflammatory bile duct injury that was mediated by upregulation of PD-L1 on cholangiocytes. Targeting of IL-17, which is an effective treatment for several autoimmune diseases, for the treatment of cholangitis should be undertaken with caution.

Lay summary

Interleukin 17 (IL-17) is assumed to be a driver of inflammation in several autoimmune diseases, such as psoriasis. IL-17 is also present in inflammatory diseases of the bile duct, but its role in these conditions is not clear, as the effects of IL-17 depend on the context of its expression. We here investigated the role of IL-17 in an experimental autoimmune cholangitis mouse model and we identified an important protective effect of IL-17 on cholangiocytes enabling them to downregulate bile duct inflammation via check point inhibitor PD-L1.

Introduction

Cholangiocytes are epithelial cells that line the intra- and extrahepatic bile ducts and actively modify bile volume and composition. In autoimmune cholestatic liver diseases, such as primary sclerosing cholangitis (PSC) and primary biliary cholangitis (PBC), cholangiocytes are targeted by T cell driven immune attack, leading to cholestasis, ductopenia and finally, to end-stage liver disease^{1, 2}. To date, there is no curative therapy, rendering these diseases major indications for liver transplantation³. Genetic and environmental factors, such as the microbiota in combination with dysregulated adaptive and innate immune responses likely contribute to the pathogenesis of these diseases^{1, 2, 4, 5}.

It has been shown that cholangiocytes are not only passive targets of immune cells, but also actively contribute to the process of periductal inflammation. Cholangiocytes express various Toll-like receptors (TLRs), the ligation of which can induce the recruitment of neutrophils and dendritic cells via CCL2, IL-6, IL-8 and Mip-3a^6–8. Senescence associated cytokines add to a pro-inflammatory periductal environment⁹. Cytokines described to act on cholangiocytes include IFNgamma, IL-1beta, IL-6 and IL-17^{8, 10}. Cholangiocytes activated by IFNgamma express chemokines such as CCL2 and CCL20 and surface proteins, including MHC class II, ICAM-1 and VCAM-1, and thus actively participate in antigen presentation and immune cell recruitment^11–14.

IL-17 is involved in the pathogenesis of many autoimmune diseases and neutralizing IL-17 has been established as a therapy, e.g. of psoriasis^{15, 16}. However, targeting IL-17 in mucosal disease such as Crohńs disease, has been disappointing¹⁷. IL-17 clearly acts in a highly context dependent manner, so it is of paramount importance to gain a better understanding of the role of IL-17, and especially of its most prominent family members IL-17A and IL-17F in mucosal immunology. At mucosal barriers, IL-17 contributes to the protection against extracellular bacterial and fungal pathogens¹⁸. We have previously described the localization of IL-17⁺ T cells around bile ducts of PSC patients and a shift in the balance between IL-17A producing CD4⁺ T (Th17) cells and regulatory T cells (Tregs)^{19, 20}. Th17 cells may be induced by intestinal microbiota and have recently been shown to aggravate murine cholangitis⁴. In addition, we recently reported that the monocyte-cholangiocytes interaction contributes to microbiota-induced Th17 differentiation in PSC patients²¹. Not only Th17 cells, but also IL-17⁺ CD8⁺ T (Tc17) cells were reported to accumulate in inflammatory and autoimmune liver diseases^{11, 22, 23}. Compared to conventional cytotoxic T cells, Tc17 cells were described to exhibit higher pro-inflammatory potential, but reduced secretion of granzyme B, perforins and overall cytotoxicity^23–25. Auto-reactive, cytotoxic CD8⁺ T cells are suspected to promote the pathogenesis of PBC^{26, 27}, but little is known about the functional role of Tc17 cells in the context of autoimmune cholestatic diseases.

Increasing evidence suggests that the interaction of inflammatory T cells and cholangiocytes involves IL-17; however, the effects of IL-17 on cholangiocytes remain unclear. In this study, we focused on the role of IL-17A/F produced by CD8⁺ T cells on cholangiocytes in vitro and murine cholangitis in vivo.

Materials and Methods

Mice

C57BL6/J and OT-1^wt mice (C57BL/6-Tg(TcraTcrb)1100Mjb/J) were obtained from Jackson Laboratory, Maine, USA. K14-OVAp mice were kindly provided by Kirsten Hogquist (Minnesota, USA) and IL-17A/F^−/− mice by Immo Prinz (Hannover, Germany), both on C57BL6/J background. OT-1^IL17ko mice were generated by crossbreeding. All mice were bred and housed under specific pathogen free conditions with 12 h light/dark cycles at the animal care facility of the University Medical Center Hamburg-Eppendorf with access to standard chow diet (1318 rodent diet, Altromin, Germany) and water available ad libitum. The K14-OVAp mouse model was described previously²⁸. Animal care was in accordance with the governmental and institutional guidelines and all experiments comply with the ARRIVE guidelines²⁹ and were approved by the local review board of the State of Hamburg, Germany (G36/16, ORG846 and ORG 979).

Induction of cholangitis in mice

Cholangitis in female K14-OVAp mice was induced as described previously²⁸. Briefly, congenic CD8⁺ T cells recognizing the ovalbumin peptide expressed on K14-OVAp cholangiocytes were isolated freshly from spleens of female OT-1^wt or OT-1^IL17ko donors. Cells were isolated using anti-CD8-FITC antibody (BioLegend, Germany) and anti-FITC immunomagnetic beads (Miltenyi Biotec, Germany) according to the manufacturer’s instructions. Female K14-OVAp recipients were injected i.v. with 200,000 OT-1^wt or OT-1^IL17ko CD8⁺ donor T cells and sacrificed on day 5 after adoptive T cell transfer. 400 µg of isotype control or anti-PD-L1 antibodies (both BioLegend, Germany) were administered one day before and three days after adoptive T cell transfer by i.p. injection as indicated in respective experiments.

Serum transaminases

Mouse alanine (ALAT) and aspartate (ASAT) serum aminotransferase levels were measured using a Cobas Integra 400 plus System (Roche Diagnostics, Switzerland) at the Institute of Experimental Immunology and Hepatology, University Medical Center Hamburg-Eppendorf.

Cell isolation

Spleen cells and liver infiltrating lymphocytes were isolated as described previously²⁸. To isolate cholangiocytes³⁰, female mouse livers were perfused with 0.5 mg/ml collagenase NB 4G (Nordmark, Germany) in PBS and dissected mechanically. Tissue was digested with collagenase (2.5 mg/ml) in PBS for 20 min shaking at 37 °C and filtered with 100 µm strainer. Remaining tissue was further digested with collagenase (5 mg/ml). Finally, tissue was degraded and filtered using 0.05 % trypsin/EDTA (Thermo Fisher Scientific, Germany) for 10 min shaking at 37 °C. Hepatocytes were sedimented two times at 40 g, and debris was depleted using a 35 % percoll gradient (GE Healthcare, UK) at 900 g for 10 min. For mRNA analysis, cholangiocytes were stained with anti-CD326 (EpCAM) (Miltenyi Biotec, Germany) anti-CD45.1-FITC and anti-CD45.2-FITC (both BioLegend, Germany) antibodies for 20 min and subsequently sorted for FITC-negative/APC-positive cells using a BD FACSAria™ III at the Cytometry and Sorting Core Unit, University Medical Center Hamburg-Eppendorf.

Cell cultivation and stimulation

Isolated lymphocytes from spleen or liver were cultivated in Panserin 401-Medium (PAN Biotech, Germany) supplemented with 1 % Penicillin-Streptomycin (Thermo Fisher Scientific, Germany) in flat bottom 96 well plates. Cells were seeded at 500,000 cells per well and stimulated for 24 h using anti-CD3 and anti-CD28 (each 2 µg/ml, BD Biosciences, Germany).

Primary female mouse derived cholangiocytes were cultivated as described previously³¹. Cholangiocytes were grown confluent in 24 well plates and stimulated with fresh medium containing 10 ng/ml IFNgamma and/or IL-17A (PeproTech, USA) for 24 h. For co-culture experiments 500,000 freshly isolated splenic derived CD8⁺ T cells from female OT-1^wt and OT-1^IL17ko donor mice were added for 48 hours to 70 % confluent primary mouse cholangiocytes isolated from female K14-OVAp donors. Supernatants were harvested and analyzed using ELISA for CCL20, IL-6, IL-17A, IFNgamma (all R&D Systems, USA) and granzyme B (Thermo Fisher Scientific, Germany) and cells were harvested for RNA isolation and qPCR analysis.

3 D Organoids

Cholangiocyte organoids were isolated and cultured as described previously³². In brief, liver tissue was mechanically dissected followed by collagenase digestion for 30–60 minutes. Digested cells were washed repeatedly and debris was depleted using a percoll gradient (GE Healthcare, UK) at 500 g for 30 min. Cells were washed with DMEM (1% FCS, 1% P/S) and mixed with Matrigel-Culturex RGF BME Type2 (Roche, Germany), plated in a 24-well plate, and allowed to polymerize. After polymerization of the Matrigel, cells were cultured with conditioning medium containing 30% Wnt3a, 10% R-spondin, 1% P/S, 1% L-Glut, 2% B27 (Thermo Fisher Scientific, Germany)), 10 mM nicotinamide (Sigma; Germany)), 1 mM N -acetyl cysteine (Sigma, USA),1 % N2 (Thermo Fisher Scientific, Germany)),10 nM Gastrin (Sigma, USA), 50 ng/ml HGF (PeproTech, Germany)), 50 ng/ml EGF (PeproTech, USA), 5 µM TGFbeta inhibitor (Tocris, Germany)), 100 ng/ml FGF10 (PeproTech, Germany)) and 10 µM Forskolin (Tocris, Germany)). Medium was changed to expansion medium (conditioning medium without Wnt3a) after 3 days and changed every 2 days. Cultures were passaged after 7–10 days. For stimulation, organoids were cultured as described above and stimulated after 3 days with fresh expansion medium containing 100 ng/ml IFNgamma and/or IL-17A (both PeproTech, Germany) for 24 h. Organoids were harvested for RNA isolation and qPCR analysis.

Patients

A total of 76 patients who attended the outpatient service of the YAEL Center for Autoimmune Liver Disease of the I. Department of Medicine, University Medical Center Hamburg-Eppendorf (UKE, Hamburg, Germany), were included in the study. IL-17A cytokine secretion of PBMC was analyzed in blood samples of 30 patients with PSC, 26 patients with PBC, 20 patients with AIH and 20 healthy donors. Cholangiocyte organoids were isolated from liver tissue of two patients with alcoholic liver disease (ALD), two patients with non-alcoholic fatty liver disease (NAFLD) and one patient with cirrhosis of unknown origin. No donor organs were obtained from executed prisoners or other institutionalized persons. This study was approved by the Ethics Committee of Hamburg and written informed consent was obtained from all patients and healthy controls (PV4081).

Real-time qPCR

Total RNA was extracted from liver tissue (NucleoSpin® RNA, Macherey-Nagel, Germany) and reverse-transcribed (High capacity cDNA reverse transcription kit, Thermo Fisher Scientific, Germany) according to the manufacturer’s instructions.

For quantitative real-time PCR analysis, mRNA expression of various genes listed in Table 1 was measured using TaqMan™ Fast Advanced Master Mix and TaqMan®Gene Expression Assays (Thermo Fisher Scientific, Germany). Target gene expression was normalized to Hprt expression and the fold-induction was quantified by normalization to control groups using the ΔΔCt method.

Table 1.

TaqMan^®Gene Expression Assay IDs of genes used in this study

Gene	Assay ID	Gene	Assay ID
Ccl2	Mm00441242_m1	IL6	Hs00174131_m1
CCL2	Hs00234140_m1	Il10	Mm00439614_m1
Ccl20	Mm01268754_m1	Il12a	Mm00434169_m1
CCL20	Hs00355476_m1	Il23a	Mm00518984_m1
Cd274	Mm03048248_m1	Krt19	Mm00492980_m1
CD274	Hs00204257_m1	Ly6c	Mm03009946_m1
Col1a1	Mm00801666_g1	Ly6g	Mm04934123_m1
Col3a1	Mm01254476_m1	Mmp9	Mm00442991_m1
Cxcl9	Mm00434946_m1	Mmp13	Mm00439491_m1
Cxcl10	Mm00445235_m1	Mpo	Mm01298424_m1
Foxp3	Mm00475162_m1	Nfkb1	Mm00476361_m1
Gzmb	Mm00442837_m1	Pdcd1	Mm01285676_m1
Hprt	Mm03024075_m1	Rorc	Mm01261022_m1
HPRT	Hs02800695_m1	Tgfb1	Mm01178820_m1
Ifng	Mm01168134_m1	Tnfa	Mm00443258_m1
Il1b	Mm00434228_m1	Vcam1	Mm01320970_m1

Flow cytometry

Immunofluorescence surface staining was performed with antibodies to CD3, CD4, CD8, CD25, CD45.1, CD45.2, CD274 (PD-L1), CD279 (PD-1) listed in Table 2 (all BioLegend, USA), and CD326 (EpCAM) (Miltenyi Biotec, Germany). Dead cells were stained with Pacific Orange™ Succinimidyl Ester (Thermo Fisher Scientific, Germany) and excluded from further analysis. For intracellular cytokine staining, cells were restimulated with PMA (10 ng/ml) and ionomycin (1 µg/ml, both Sigma-Aldrich, Germany) in the presence of GolgiPlug™ (1 µl/ml, BD Biosciences, USA) for 3 h. Cells were PFA fixed and perforated in PBS containing 0,5% saponine and 2% BSA and stained for IL-17, GzmB, Ki67 (all BioLegend, USA), IFNgamma and IL-2 (both BD Biosciences, USA). Apoptotic cells were stained with Annexin V-FITC in Annexin V Binding Buffer (both BD Bioscience, USA) according to BD protocols. Flow cytometry was performed using a BD LSR II cytometer (BD Biosciences, USA) and data were analyzed with FlowJo software V10.6.0.

Table 2.

List of antibodies

Antibody	Fluorochrome	Clone	Company	Order no
Murine PD-L1	-	10F.9G2	Biolegend	124301
Rat IgG2b,kappa	-	RTK4530	Biolegend	400601
anti-mouse CD3	-	145–2C11	Biolegend	100331
anti-mouse CD28	-	37.51	Biolegend	102112
anti-mouse CD3	FITC	17A2	Biolegend	100204
anti-mouse CD4	PE-Dazzle	RM4–5	Biolegend	100566
anti-mouse CD8	V450	53–6.7	BD Bioscience	560469
anti-mouse CD45.1	APC	A20	Biolegend	110714
anti-mouse CD45.1	FITC	A20	Biolegend	110706
anti-mouse CD45.2	FITC	104	Biolegend	109806
anti-mouse GzmB	FITC	GB11	Biolegend	515403
anti-mouse PD-1	BV421	29F.1A12	Biolegend	135217
anti-mouse PD-L1	PE	MIH5	eBioscience	12–5982-82
anti-mouse IFNgamma	AF700	XMG.1	Biolegend	557998
anti-mouse IL-17A	PE	eBio17B7	eBioscience	12–7177-81
anti-human CD3	AF700	UCHT1	Biolegend	300424
anti-human CD4	PacificBlue	RPA-T4	Biolegend	300521
anti-human IL17A	AF647	BL168	Biolegend	512310
anti-human IFNgamma	PE	4S.B3	Biolegend	502509
Murine PD-L1	-	PD-L1/B7-H1	R&D	AF1019
CK19	-	Troma III	DSHB	-
goat-anti Rat	Cy5	-	Invitrogen	A10525
goat-anti Rat	AF546	-	Invitrogen	A11081

Histology

For mouse derived liver tissue formalin-fixed and paraffin-embedded sections were stained with Hematoxilin/Eosin (H&E, Carl Roth, Germany) according to standard procedure. CK19 (Troma III, DSHB, USA), mouse PD-L1 (R&D Systems, Germany) and CD45.1 (BioLegend, Germany) stainings were performed on cryo-frozen liver tissue from K14-OVAp mice.

CK19 and IL-17A RNAscope® staining was carried out on cryo-frozen tissue from K14-OVAp mice. In situ RNA hybridization was performed using the RNAscope® Fluorescent Multiplex Detection Reagent Kit (Advanced Cell Diagnostic, Italy) according to manufactureŕs instruction. Briefly, cryo-frozen tissue was fixated and endogenous peroxidase activity was blocked. Slides were stained with CK19 (Troma III, DSHB, USA) and goat-anti rat-Cy5 (Invitrogen, Germany) followed by an incubation for 2 hours at 40 °C with target C1 probe specific for IL-17A (Mm-Il17a). The signal was amplified and developed with fluorescent label combination (Amp4 B) for fluorescent detection.

For human CD8/ IL-17 double staining formalin-fixed paraffin-embedded human liver sections were dehydrated, followed by antigen retrieval and endogenous peroxidase block. CD8 was stained using mouse anti-human CD8 (Dako, Germany) and anti-mouse HRP polymer. Peroxidase activity was visualized using 3.3’-Diaminobenzidine (DAB). IL-17A was stained using anti-human IL-17A (R&D Systems, Germany) polyclonal rabbit anti-goat IgG (DAKO, Germany) and anti-rabbit AP-complex (Polap kit, Zytomed, Germany). AP was used to visualize IL-17 using the POLAP-Kit (Zytomed Systems, Germany) according to manufactureŕs instructions. Slides were the counterstained with Haemalaun and mounted with mounting media (both Roth, Germany). All pictures were taken with a Biorevo BZ-9000 fluorescence microscope (Keyence, Japan) at RT. Histological scoring was performed by a pathologist in a blinded fashion and liver inflammation was assessed using the modified hepatitis activity index (mHAI)³².

Statistical analysis

Statistical analysis was performed with GraphPad Prism (V.8.4.3) software. Data are presented as mean ± SD. Differences between two groups were assessed for statistical significance using the Mann-Whitney test. Comparisons between more than two groups were performed by ordinary one-way ANOVA (Analysis of Variance) Test and Tukey`s Post Test if not mentioned otherwise. Significance levels are indicated by asterisks: **** indicate p < 0.0001; *** indicate p < 0.001; ** indicate p < 0.01; * indicates p < 0.05.

Results

Antigen-specific CD8⁺ T cells deficient in IL-17A/F aggravate experimental autoimmune cholangitis

OT-1^IL17ko cells were isolated from crossbred mice. To exclude inherent functional differences, the activation status between OT-1 CD8⁺ T cells proficient and deficient in IL-17A/F was investigated ex vivo. Proliferative capacity, cytokine secretion and cytotoxicity were found to be unaffected by the lack of IL-17A/F. In addition, the expression of inhibitory molecules (such as Tim3), activation marker (CD69, CD25, CD44) or effector memory marker (CCR7, CD62L) were similar on CD8⁺ T cells isolated from OT-1^wt compared to OT-1^IL17ko mice (Suppl. Fig. 1+2).

For cholangitis induction, OVA-specific OT-1^wt CD8⁺ T cells or OT-1^IL17ko CD8⁺ T cells were adoptively transferred into K14-OVAp recipient mice that present the MHC-I restricted OVA-peptide SIINFEKL on cholangiocytes²⁸. On day 5 following adoptive T cell transfer, recipient mice showed periportal inflammation and increased levels of serum transaminases compared to PBS treated controls (Fig. 1A, B). The transferred antigen-specific T cells located in the periductal area of recipient livers (Fig. 1C). We observed increased cholangitis severity in mice receiving OT-1^IL17ko T cells, with more severe histological inflammation and increased levels of serum transaminases (Fig. 1A, B).

Fig. 1. Lack of IL-17A/F in antigen-specific CD8⁺ T cells aggravates cholangitis in K14-OVAp animals.

Cholangitis severity in K14-OVAp recipient mice on day 5 after adoptive transfer of OT-1^wt or OT-1^IL17ko CD8⁺ T cells: (A) mHAI histological activity index of H&E-stained liver sections and (B) levels of serum transaminases ASAT and ALAT. (C) Representative immunofluorescent staining of congenic cell marker CD45.1 (red, transferred T cells), CK19 (green, cholangiocytes) and nuclei (blue) performed on cryo-sections of K14-OVAp liver after adoptive cell transfer. Data are expressed as mean ± SD; n=10–17, representing pooled data from three independent experiments. Levels of significance: *p≤0,05; **p≤0,01; ***p≤0,001; ****p≤0,0001 (ordinary one-way ANOVA).

Lack of IL-17A/F in antigen-specific CD8⁺ T cells affects gene expression in cholangiocytes and T cells themselves

We next aimed to elucidate the mechanisms leading to the increase in cholangitis severity induced by OT-1 CD8⁺ T cells deficient in IL-17A/F. To exclude parenchymal or endothelial cell apoptosis, TUNEL staining of liver sections was performed. This did not reveal any significant differences between the groups (Suppl. Fig. 3). However, transfer of antigen-specific T cells significantly induced the hepatic expression of pro-inflammatory, chemotactic and immune-regulating genes compared to PBS treated controls (Fig. 2A). Importantly, the transfer of OT-1^IL17ko CD8⁺ T cells resulted in significantly increased hepatic mRNA expression of genes regulating T cell recruitment (Vcam1 and Cxcl9), inflammation (Ifng) and cytotoxicity (Gzmb) compared to transfer of OT-1^wt CD8⁺ T cells proficient in IL-17 (Fig. 2B). However, the expression of Ccl20, Il1b and Il6 was significantly reduced. These genes were described to be expressed by cholangiocytes, the target cells of T cell driven inflammation in this model^{10, 11}. To confirm that a direct interaction between CD8⁺ T cells and cholangiocytes contributes to these findings, primary cholangiocytes were isolated from healthy K14-OVAp mice and co-cultured with OVA-specific OT-1^wt or OT-1^IL17ko CD8⁺ T cells (Fig. 2C). Co-culture of OT-1^IL17ko CD8⁺ T cells with primary cholangiocytes resulted in a decreased secretion of CCL20 and IL-6 by cholangiocytes and a greatly increased secretion of T cell-derived IFNgamma. Since this was not a feature of OT-1^IL17ko T cells per se, these data pointed to a bidirectional interaction between cholangiocytes and CD8⁺ T cells dependent on IL-17.

Fig. 2. IL-17A/F-deficiency in OT-1^IL17ko CD8⁺ T cells leads to differential effects on T cell and cholangiocyte activation.

Whole liver mRNA expression analysis of pro-inflammatory, chemotactic and immune-regulating genes from K14-OVAp recipient mice 5 days after transferring OVA-specific OT-1^wt cells or OT-1^IL17ko cells and PBS as control. (A) Log2-fold change of mRNA expression of all measured genes normalized to PBS treated controls. (B) X-fold mRNA expression of Ifng, Gzmb, Pdcd1, Vcam1, Ccl20, Cxcl9, Il1b and Il6 normalized to PBS treated control mice. (C) Primary cholangiocytes isolated from K14-OVAp mice were cultured alone or in the presence of OT-1^wt or OT-1^IL17ko CD8⁺ T cells for 48 h. Chemokine and cytokine secretion were analyzed in supernatants using ELISA. Data are expressed as mean ± SD; n=10–12, representing pooled data from two independent experiments. Levels of significance: *p≤0,05; **p≤0,01; ***p≤0,001; ****p≤0,0001 (ordinary one-way ANOVA).

Antigen-specific CD8⁺ T cells lacking IL-17A/F show increased proliferation and activation during cholangitis in vivo

We previously demonstrated that liver damage in our model was primarily driven by the transferred antigen-specific OT-1 CD8⁺ T cells^{28, 34}. In a detailed flow cytometric analysis of liver infiltrating lymphocytes in K14-OVAp animals following transfer of OT-1^wt compared to OT-1^IL17ko CD8⁺ T cells we here confirmed that the most prominent differences in T cell populations were observed regarding the frequency of transferred CD8⁺ T cells by (Suppl. Fig 4). Therefore, we next focused on the phenotype of transferred OT-1^wt and OT-1^IL17ko T cells during cholangitis. IL-17A/F-deficiency in transferred OT-1^IL17ko CD8⁺ T cells resulted in their increased expression of the cytotoxicity marker granzyme B, whereas high expression of IFNgamma was detected in both IL-17A/F-competent and -deficient OT-1 CD8⁺ T cells (Fig. 3A). Of note, increased expression of IL-17A in livers of K14-OVAp recipient animals was observed following transfer of OT-1^wt CD8+ T cells (Fig. 3B). Confirming our previous findings, the majority of cytokines were produced by the transferred OT-1 CD8⁺ T cells and not by recruited endogenous T cells (Fig. 3C). Stimulation of MACS-purified T cells confirmed the results observed by flow cytometry (Fig. 3D).

Fig. 3. Enhanced cytotoxicity and cell expansion of OT-1^IL17ko CD8⁺ T cells in experimental cholangitis.

On day 5 after disease induction, transferred OT-1^wt or OT-1^IL17ko CD8⁺ T cells were isolated from livers and analyzed. (A) Cytokine expression in endogenous and transferred CD8⁺ T cells was analyzed by flow cytometry. (B) Immunohistochemical RNA scope staining of IL-17A (yellow) and cholangiocytes (purple) in livers of K14-OVAp animals on day 5 following transfer of OT-1^wt or OT-1^IL17ko CD8⁺ T cells. (C) IL-17A expression of total living cells and of total CD45.1⁻ and CD45.1⁺ (transferred) T cells. (D) Cytokine secretion of stimulated liver infiltrating lymphocytes was quantified by ELISA. Flow cytometry was used to analyze (E) the percentages and total numbers and (F) surface expression of CD25 of OT-1^wt and OT-1^IL17ko CD8⁺ T cells within recipient livers. (G) Expression of IL-2, Ki67 and annexin V was analyzed in liver infiltrating OT-1^wt and OT-1^IL17ko CD8⁺ T cells using flow cytometry. Data are expressed as mean ± SD; (A), (F), (G) Data from one representative experiment of three independent experiments is shown; (C), (E) n=10–12, representing pooled data from two independent experiments. Levels of significance: *p≤0,05; **p≤0,01; ***p≤0,001; ****p≤0,0001 (ordinary one-way ANOVA (A, C, D, E), multiple t test (E, F), Mann-Whitney U test (G)).

Moreover, significantly increased frequencies and absolute numbers of OT-1 CD8⁺ T cells were observed in livers of recipient mice on day 5 after transfer of OT-1^IL17ko T cells (Fig. 3E). In order to differentiate between increased recruitment and proliferation of transferred T cells, time kinetic experiments were performed. On day 3 following disease induction the numbers of recruited OT-1 CD8⁺ T cells were not different between cells proficient or deficient in 17A/F and the expression of CD25, the IL-2 receptor alpha chain which is required for T cell proliferation, was similar in both groups. However, on day 5 after cell transfer, IL-17A/F-competent OT-1^wt CD8⁺ T cells showed a marked decrease in expression of CD25, whereas this downregulation was not observed in OT-1^IL17ko CD8⁺ T cells (Fig. 3F). In addition, lack of IL-17A/F in OT-1^IL17ko T cells resulted in significantly increased T cell proliferation, indicated by elevated intracellular IL-2 and Ki67 expression and in reduced T cell apoptosis, demonstrated by reduced annexin V staining on day 5 after transfer (Fig. 3G). These results suggested that IL-17 produced by CD8⁺ T cells served to regulate T cell expansion and contraction in the later phase of inflammation.

IL-17A/F expression by CD8⁺ T cells induces upregulation of PD-L1 in cholangiocytes

It has previously been shown that cholangiocytes can be activated by pro-inflammatory cytokines. Thus, upon stimulation with IFNgamma, the cytokine highly expressed by transferred OT-1 cells, cholangiocytes upregulated the expression of Ccl2 and Ccl20 mRNA (Suppl. Fig. 5A). Since we now demonstrated that, in our cholangitis model, T cell activation, proliferation and contraction depend on IL-17 and that IL-17 regulates cholangiocyte activation, we next asked how IL-17A/F produced by CD8⁺ T cells modulates the interaction between T cells and cholangiocytes. The inactivation of effector T cells can be mediated by PD-1/PD-L1 interaction³⁵ and remarkably high PD-1 expression was found on transferred OT-1 CD8⁺ T cells in the livers of recipient mice independent of their IL-17 expression (Fig. 4A). Thus, we hypothesized that IL-17 might induce the expression of co-inhibitory molecules such as PD-L1 on cholangiocytes. Immunofluorescence staining confirmed the expression of PD-L1 on cholangiocytes in K14-OVAp recipient mice on day 5 after disease induction (Fig. 4B). We next aimed to quantify the expression of PD-L1 on mRNA and protein level and analyzed isolated cholangiocytes from K14-OVAp recipient mice on day 5 after disease induction. After transfer of OT-1^wt CD8⁺ cells competent in IL-17A/F, cholangiocytes showed upregulation of Cd274 mRNA encoding for PD-L1 (Fig. 4C). Significantly lower levels of Cd274 mRNA were found in cholangiocytes targeted by IL-17A/F-deficient OT-1^IL17ko CD8⁺ T cells. At the protein level, flow cytometric analysis confirmed these observations (Fig. 4D). To further determine the effect of IL-17 on the expression of PD-L1 by cholangiocytes, primary mouse cholangiocytes were examined after stimulation with pro-inflammatory cytokines in vitro. IFNgamma, but not IL-17A alone, strongly induced the expression of Cd274 mRNA. However, the expression of Cd274 mRNA in cholangiocytes was highest when cells had been stimulated with a combination of IFNgamma and IL-17A (Fig. 4E; Supp. Fig. 6). In summary, these results show that IL-17 acts as an enhancer of IFNgamma-induced expression of inhibitory PD-L1 in cholangiocytes in vitro and in vivo.

Fig. 4. PD-L1 expression on cholangiocytes is enhanced in the presence of IL-17.

5 days after induction of experimental cholangitis, flow cytometry was used to analyze (A) the surface expression of PD-1 on OT-1^wt and OT-1^IL17ko CD8⁺ T cells isolated from liver. (B) Immunofluorescence staining of PD-L1 (red) on cholangiocytes (green) in K14-OVAp mice on day 5 following transfer of OT-1^wt and OT-1^IL17ko CD8⁺ T cells. (C) Cd274 mRNA expression and (D) surface expression and MFI of CD274 (PD-L1) on cholangiocytes determined in cholangiocytes isolated on day 5 after cell transfer and cell sorted based on the expression of CD326. Regulation of Cd274 mRNA expression was measured in (E) primary mouse cholangiocytes following cytokine stimulation for 24 h. Data were analyzed using qPCR and normalized to the control group. Bars represent mean ± SD. (A), (D), (E) Data from one representative experiment of three independent experiments is shown; (C) n=10–14, representing pooled data from two independent experiments. Levels of significance: *p≤0,05; **p≤0,01; ***p≤0,001; ****p≤0,0001 (ordinary one-way ANOVA).

Inhibition of PD-L1 signalling aggravates experimental autoimmune cholangitis through uncontrolled T cell expansion

We next aimed to confirm our hypothesis that CD8⁺ T cell activation and expansion during experimental cholangitis can be regulated via the PD-1/PD-L1 axis. To that end, the development of cholangitis was investigated after adoptive transfer of OT-1^wt CD8⁺ T cells in combination with a blocking antibody to PD-L1. The inhibition of PD-L1-signalling aggravated cholangitis in K14-OVAp recipient animals compared to isotype-treated mice with highly increased histopathological inflammation scores (Fig. 5A) and elevated levels of transaminases (Fig. 5B) on day 5 after adoptive transfer. Analysis of OT-1^wt T cells in anti-PD-L1-treated mice showed significantly increased intrahepatic frequencies of transferred antigen-specific CD8⁺ CD45.1⁺ T cells (Fig. 5C) with elevated surface expression of CD25 and intracellular expression of IL-2 compared to cells isolated from isotype treated controls (Fig. 5D). Anti-PD-L1 treatment reduced apoptosis of transferred OT-1^wt CD8⁺ T cells, indicated by lower frequencies of annexin V positive OT-1 CD8⁺ T cells compared to OT-1 CD8⁺ T cells from isotype-treated mice. Moreover, the expression of pro-inflammatory Ifng and cytotoxic Gzmb mRNA was increased in liver tissue of antibody-treated K14-OVAp mice on day 5 after disease induction (Fig. 5E).

Fig. 5. Inhibition of PD-L1 induces aggravated cholangitis in K14-OVAp recipient animals.

Cholangitis severity was assessed 5 days after adoptive transfer of OT-1^wt CD8⁺ T cells into K14-OVAp mice treated with anti-PD-L1 or isotype-control. (A) mHAI histological activity index of H&E stained liver sections and (B) serum levels of liver transaminases ASAT and ALAT. (C) The percentage of transferred OT-1^wt CD8⁺ CD45.1⁺ T cells was determined within recipient livers and (D) the expression of CD25, IL-2 and annexin V on OT-1^wt CD8⁺ T cells isolated from livers using flow cytometry. (E) Expression of Gzmb, Ifng and Pdcd1 mRNA was analyzed in K14-OVAp whole liver tissue by qPCR. Data are expressed as mean ± SD; n=4–5, Data from one representative experiment of two independent experiments is shown. Levels of significance: *p≤ 0,05; **p≤ 0,01; ****p≤0,0001 (ordinary one-way ANOVA (A, B, C, E), Mann-Whitney U test (D)).

CD8⁺ IL-17A producing T cells in patients with cholestatic liver diseases

We next investigated the presence of IL-17A producing CD8⁺ T cells in human blood and liver from patients with autoimmune cholestatic liver disease. To this end we performed immunohistochemical double-staining for CD8 and IL-17A in human explant liver tissue (Fig. 6A). We observed CD8⁺ T cells in all analyzed liver sections mainly within portal tracts and liver lobules. Only in the livers of PSC patients we observed CD8⁺ T cells infiltrating the bile duct epithelial layer and thus close interaction between CD8⁺ T cells and cholangiocytes. In addition, IL-17A protein expression was detected around bile duct infiltrating CD8⁺ T cells in PSC. In peripheral blood, we observed significantly increased expression of IL-17A in CD8⁺ T cells from PSC and PBC patients compared to AIH patients and healthy controls (Fig. 6B). Using human cholangiocyte organoids derived from explant liver tissue, we were able to confirm that the highest expression of CD274 mRNA in cholangiocytes was induced by the combination of IFNgamma and IL-17A (Fig. 6C).

Fig. 6. Presence of CD8⁺IL17A⁺ producing T cells in cholestatic liver diseases.

(A) Immunohistochemical double staining for CD8 (brown) and IL-17A (red) in explant liver sections from NASH and PSC, and from unaffected tissue derived from tumour margin. (B) Peripheral blood of patients with PSC (n=30), PBC (n=26), AIH (n=20) and healthy donors (n=20) was stimulated with PMA/Ionomycin for 4 h and analyzed for cytokine production using flow cytometry. (C) CD274 mRNA expression was measured using qPCR in human liver derived organoids generated from ALD (n=2), NAFLD (n=2) and cirrhosis of unknown origin (n=1) following cytokine stimulation for 24 h. (B) Data were analyzed using the Kruskal-Wallis test, followed by Dunn’s Multiple Comparison **p≤0.01, ***p≤0.001. Bars represent mean ± SD. Levels of significance: *p≤0,05; ***p≤0,001; ****p≤0,0001 (ordinary one-way ANOVA).

Discussion

There is no curative treatment for autoimmune cholestatic liver diseases such as PBC and PSC, thus they often progress to end stage liver disease^{1, 2}. Targeted treatment options are therefore urgently needed. PBC and PSC are diseases of the bile ducts, but mucosal immunology of the bile ducts is poorly understood. IL-17 plays a major role in the pathogenesis of several autoimmune diseases outside the gastrointestinal tract¹⁵. Even though IL-17-producing T cells have been implicated in PBC and PSC pathogenesis it remains unclear whether IL-17 presents such a therapeutic target due to its highly context-dependent activities, including its antimicrobial functions¹⁸. We aimed to dissect the interaction of T cells and cholangiocytes by investigating the effects of IL-17 expressed by CD8⁺ T cells on cholangiocytes in vitro and in vivo in autoimmune experimental cholangitis.

We previously reported that IL-17⁺ cells localized around the bile ducts of PSC livers¹⁹. Pathogen-induced Th17 differentiation was increased in PSC patients and we recently observed that an increased Th17 differentiation in PSC already occurs in vivo²¹. The presence of IL-17-producing cells has been associated with a more severe disease phenotype and a harmful role in autoimmune liver diseases^{11, 36–38}. IL-17 is produced by several T cell subsets, including CD8⁺ T cells. IL-17⁺ CD8⁺ T (Tc17) cells have mainly been assigned a pro-inflammatory role, but could also contribute to immune regulation^{23, 25}. The distinct role of IL-17 and Tc17 cells in the pathogenesis of autoimmune liver diseases remains unclear.

In previous work, we described enhanced recruitment of endogenous Th17 cells towards the site of inflammation after disease induction with OT-1^wt CD8⁺ T cells in the K14-OVAp mouse model used in this study²⁸. We hypothesized, that transferred, antigen-specific OT-1^wt CD8⁺ T cells induce initial cholangiocyte activation and subsequent recruitment of endogenous CD4⁺ T cells to the liver. By inducing cholangitis with antigen-specific OT-1 CD8⁺ T cells lacking IL-17A/F (OT-1^IL17ko), we now demonstrate communication between CD8⁺ T cells and cholangiocytes. In line with a regulatory function of Tc17 cells^{23, 25}, we observed highly activated and cytotoxic OT-1 T cells in the setting of IL-17A/F deficiency, and increased expansion of these cells.

Mechanistically, CD8⁺ T cell-derived IL-17 induced the expression of PD-L1 on antigen-presenting cholangiocytes that limited the expansion of self-reactive T cells in cholangitis. PD-L1 presented by APCs is well known to inhibit T cell proliferation after binding to the PD-1 receptor³⁵ and cholangiocytes have been previously described to upregulate expression of PD-L1 and PD-L2 in vitro, protecting themselves by reducing CD8⁺ T cell cytotoxicity^{35, 39, 40}.

Indeed, imbalanced expression of PD-1/PD-L1 has previously been reported in inflamed livers of PBC and AIH patients^{41, 42}. Moreover, recent case reports on the development of secondary sclerosing cholangitis in patients receiving checkpoint inhibitor treatment highlight the functional importance the PD-L1/PD-1 axis for biliary immune homeostasis^{43, 44}.

Inflammatory cytokines, such as IFNgamma and TNF have already been described as potent inducers of PD-L1 expression^{45, 46}. By activation of STAT3^47–49 or by its influence on Cd274 mRNA translation⁴⁶, IL-17 was described to support PD-L1 expression in different epithelial and immune cells. Our in vivo and in vitro data clearly demonstrate that IL-17 significantly enhances the expression of PD-L1 on cholangiocytes, thereby providing mechanistic evidence for the bidirectional communication between CD8⁺ T cells and cholangiocytes.

We and others have shown that IL-17A induces the expression of Ccl20 mRNA in cholangiocytes in vitro (Suppl. Fig. 6). Secretion of CCL20 by cholangiocytes leads to the recruitment and positioning of CCR6⁺ Th17 and Tc17 cells around bile ducts¹¹. Furthermore, activated cholangiocytes were shown to promote Th17 and Tc17 differentiation by providing IL-1beta and IL-6¹⁰. Importantly, we found increased numbers of Tc17 cells in recipient animals after disease induction with IL-17-competent OT-1 CD8⁺ T cells and the observed reduced cytotoxicity of Tc17 cells may have contributed to the milder inflammation observed in these animals ^23–25. Therefore, antigen-specific Tc17 cells seem to engage in a feedback loop with cholangiocytes, which limits auto-reactive T cell expansion in autoimmune experimental cholangitis.

It is tempting to speculate whether IL-17⁺ T cells seen in PBC and PSC patients also engage in such a feedback loop in human disease. In light of the recently reported pathogenetic role of Th17 cells in PSC, our data underline the need to better define the role of IL-17 in liver inflammation in a cell dependent manner and may help to explain, why broad neutralization of IL-17 may not be effective treatment of cholangitis.

Highlights:

• IL-17 induces the expression of PD-L1 in mouse and human cholangiocytes

• CD8⁺ T cell-derived IL-17 induces the expression of PD-L1 on antigen-presenting cholangiocytes in an antigen driven, CD8⁺ T cell mediated experimental cholangitis model

• PD-L1 expression by cholangiocytes restricts the expansion of self-reactive T cells in experimental cholangitis

• caution may be warranted using broad neutralization of IL-17 for the treatment of T cell mediated cholangitis.

Abbreviations

IL

Interleukin

PD

programmed cell death protein

PDL-1

Programmed cell death 1 ligand

OVA

Ovalbumin

OT

Ovalbumin transgene

PSC

primary sclerosing cholangitis

PBC

primary biliary cholangitis

TLR

Toll like receptor

Treg

regulatory T cells

ALAT

alanine serum aminotransferase

ASAT

aspartate serum aminotransferase

ALD

alcoholic liver disease

NAFLD

non-alcoholic fatty liver disease

CK

Cytokeratin

MFI

mean fluorescence intensity

Data availability statement:

Data are available upon request.