Increased T helper type 17 response to pathogen stimulation in patients with primary sclerosing cholangitis


  • Potential conflict of interest: Nothing to report.

  • This work was supported by the German Research Foundation (SFB 841), the Graduate School program “Inflammation and Regeneration,” and the YAEL Foundation.

Address reprint requests to: Christoph Schramm, M.D., Department of Medicine, University Medical Center Hamburg-Eppendorf, Martinistraße 52, 20246 Hamburg, Germany. E-mail:; fax: +49 40-7410-40001.


T helper (Th)17 cells are important for host defense against bacteria and fungi, but are also involved in the pathogenesis of autoimmune diseases. In primary sclerosing cholangitis (PSC), bile fluid is frequently colonized with pathogens and its strong association with inflammatory bowel disease suggests the contribution of pathogen responses to disease pathogenesis. Interleukin (IL)-17A, the signature cytokine of Th17 cells, was recently described to promote inflammation and fibrosis within the liver. Therefore, we investigated Th17 immune response to pathogens in patients with PSC. Bile fluid was obtained by endoscopic retrograde cholangiography, and bacterial and fungal species grew in the majority of samples. In addition, bacterial RNA was stained in liver sections using 16sRNA fluorescence in situ hybridization and was detected in the portal tracts in 12 of 13 tested PSC patients. Bacteria grown from patients' bile fluid were then used to stimulate peripheral blood mononuclear cells (PBMCs) and to assess their Th17 response. Compared to healthy controls or primary biliary cirrhosis patients, PBMCs from PSC patients manifested significantly higher frequencies of Th17 and Th1/Th17 cells after pathogen stimulation. The highest frequencies of Th17 cells were detected after stimulation with Candida albicans, a pathogen that has been linked to disease progression. Immunohistochemically, IL-17A-expressing lymphocytes were detected within the periductal areas of PSC patients. Th17 induction was also noted after stimulation of Toll-like receptor 5 or 7, but not of other pattern recognition receptors tested, pointing to signaling pathways potentially involved in Th17 induction in PSC. Conclusion: We demonstrate an increased Th17 response to microbial stimulation in patients with PSC. These data should prompt further studies investigating the link between pathogen responses, inflammation, and fibrosis in patients with PSC. (Hepatology 2013;53:1084–1093)




autoimmune hepatitis


alcoholic steatohepatitis


biliary epithelial cells


caspase recruitment domain-containing protein 9


endoscopic retrograde cholangiography


flow cytometry


fluorescence in situ hybridization


genome wide association studies


healthy controls


hepatitis C virus


inflammatory bowel disease


interferon gamma






major histocompatibility complex


nonalcoholic steatohepatitis


nucleotide-binding oligomerization domain 2


primary biliary cirrhosis


peripheral blood mononuclear cells


phosphate-buffered saline


primary sclerosing cholangitis




ribosomal RNA


secondary sclerosing cholangitis


T helper


Toll-like receptor


tumor necrosis factor alpha


ursodeoxycholic acid


University Medical Center Hamburg-Eppendorf

Primary sclerosing cholangitis (PSC) is a chronic liver disease characterized by inflammation and fibrosis of the intrahepatic and/or extrahepatic bile ducts, finally leading to liver cirrhosis and end-stage liver disease.[1] To date, there is no medical treatment with a proven benefit on the progressive course of the disease. A key characteristic of PSC is the association with inflammatory bowel disease (IBD), mostly an ulcerative colitis-like disease, in approximately two thirds of the cases.

A dysregulated response to pathogen stimulation may contribute to the immune activation in PSC, as has been postulated for the associated IBD.[2, 3] From the clinical point of view, it is well known that dominant biliary strictures in patients with PSC are associated with bacterial cholangitis, and that bacterial and, especially, fungal cholangitis may accelerate the progression of PSC.[4, 5] A higher rate of positive bacterial cultures could be obtained from bile of explanted livers from PSC patients, as compared to patients with primary biliary cirrhosis (PBC).[6] In addition, treatment with the antibiotic, metronidazole, resulted in some beneficial effects on liver histology, as compared to treatment with ursodeoxycholic acid (UDCA) alone.[7]

T helper (Th)17 cells, which are characterized by the signature cytokine, interleukin (IL)-17A, play an important role in the defense against extracellular bacteria and fungi,[8, 9] particularly at epithelial and mucosal surfaces.[10] Interestingly, functional as well as genetic evidence suggests a role of Th17 cells for the pathogenesis of IBD.[11] Th17 cells have also been implicated in the pathogenesis of several human autoimmune diseases, such as lupus erythematodes, multiple sclerosis, psoriasis, and rheumatoid arthritis.[12] Therefore, Th17 cells may link pathogen defense to autoimmunity; these mechanisms may also be involved in the development of PSC.

In recent genome-wide association studies (GWAS), several non-MHC (major histocompatibility complex) loci have been found to be associated with PSC.[13, 14] Among these, polymorphisms within the caspase recruitment domain-containing protein 9 (CARD9) and reticuloendotheliosis (REL) genes are of particular interest. These genes code for molecules involved in Th17 differentiation and transduction of signals received by Toll-like receptor (TLR) and dectin-1, which recognize conserved molecules of bacterial and fungal species.[2]

Here, we aimed to investigate the Th17 response to pathogens in patients with PSC. Stimulation of peripheral blood mononuclear cells (PBMCs) with bacteria and, more so, with Candida led to an increased Th17 response in patients with PSC. Bacterial RNA and Th17 cells were both detected within inflamed portal tracts of patients with PSC. These data should prompt further studies investigating the link between pathogen responses and inflammation in the pathogenesis of PSC.

Patients and Methods


All PSC patients attended the liver clinic of the Department of Medicine at the University Medical Center Hamburg-Eppendorf (UKE; Hamburg, Germany) and were diagnosed by generally accepted criteria, including cholangiographies by endoscopy or magnetic resonance imaging.[1] Fifty-eight patients with PSC underwent endoscopic retrograde cholangiography (ERCP), during which bile was acquired and cultured for microbial colonization. Blood of 46 PSC patients was obtained for pathogen stimulation. Exclusion criteria for stimulation experiments were acute inflammatory flares of PSC, overt bacterial cholangitis, or an immunosuppressive therapy with more than 10 mg of prednisolone or 1.5 mg/kg of azathioprine per day. Ten patients with PBC and 26 healthy controls (HCs) were included in the study. PBC was diagnosed according to European Association for the Study of the Liver guidelines.[1] HCs were recruited by the Institute for Transfusion Medicine at UKE in an anonymized fashion. Four patients with secondary sclerosing cholangitis (SSC), 5 with obstructive jaundice resulting from malignancy, 1 with choledocholithiasis, 2 with benign biliary stenoses, and 1 with alcoholic steatohepatitis (ASH) were included in the cholestatic control group. Peri-interventional antibiotics (3 g of sultamicillin intravenously [IV] or 400 mg of ciprofloxacin IV) were given during ERCP as soon as bile samples were obtained. All patients gave written informed consent, and the study was approved by the local ethics committee.

Generation of Heat-Killed Bacteria

Escherichia coli (ATCC25922), Staphylococcus aureus (ATCC25923), Enterococcus faecalis (ATCC29212), and Candida albicans (all from LGC Standards, Wesel, Germany) or patients' own isolates from bile fluid were cultured on blood agar overnight at 37°C. The concentration of bacteria and fungi in phosphate-buffered saline (PBS) was adjusted using McFarland standards. Bacterial suspensions were then heat inactivated, and inactivation efficacy was controlled by overnight culture.

Isolation of PBMCs from Peripheral Blood

PBMCs were obtained from peripheral heparinized blood samples of PSC patients or from buffy coats obtained from blood donors. Blood was layered onto Histopaque-1077 (Sigma-Aldrich, Seeltze, Germany), and PBMCs were isolated after density-gradient centrifugation.

Stimulation of PBMCs Using Heat-Killed Bacteria and Fungi and Ligands for Pattern Recognition Receptors

PBMCs were plated onto 96-well, round-bottomed tissue-culture plates (Sarstedt, Nümbrecht, Germany) at 300,000 cells per well in RPMI 1640 GlutaMax-I medium (Invitrogen, Darmstadt, Germany), supplemented with 2% fetal calf serum and 1% penicillin/streptomycin, and cultured at 37°C and 5% CO2. Bacterial stimulation was performed with 108/mL of heat-killed bacteria. On days 5 and 8, 50% of the medium was exchanged, and 30 U of IL-2/mL (Proleukin; Novartis Pharma, Nürnberg, Germany) was added. The following agents were used for pathogen recognition receptor stimulation: lipopolysaccharide (Sigma-Aldrich) as a ligand for TLR-4; peptidoglycan (tlrl-pgnbs; InvivoGen, Toulouse, France) as a ligand for TLR-2 in cooperation with TLR-6; lipoteichoic acid (tlrl-lta; InvivoGen) for stimulation of TLR-2 and TLR-4; flagellin (tlrl-bsfla; InvivoGen) as a ligand for TLR-5 and loxoribine (tlrl-lox; Invivogen) for TLR-7; and as muramyldipeptide (tlrl-mdp; InvivoGen), which binds the intracellular nucleotide-binding oligomerization domain 2 (NOD-2) receptor, and zymosan depleted (tlrl-dzn; InvivoGen) for stimulation of the dectin-1 receptor. Cytokine production was detected by flow cytometry (FCM) after 12 days of culture.

Cell Staining and FCM

After a 4-hour restimulation with 50 ng/mL of phorbol 12-myristate 13-acetate (PMA; P8139; Sigma-Aldrich) and 1 µl/mL of ionomycin (I9657; Sigma-Aldrich) and cytokine release inhibition using 1 µl/mL of GolgiPlug (BD Biosciences, Heidelberg, Germany), PBMCs were washed in PBS and extracellular staining was performed using 7-amino-actinomycin D (BD Biosciences), antihuman CD4 Horizon V450 (560811; BD Biosciences), and antihuman CD8 Horizon V500 (560774; BD Biosciences) at 4°C for 30 minutes. Intracellular staining was performed after washing in PBS and fixation in Cytofix/Cytoperm (BD Biosciences), using antihuman IL-17A/fluorescein isothiocyanate (11-7179; eBioscience, Frankfurt, Germany), antihuman interferon gamma (IFN-γ) /allophycocyanin (554702; BD Biosciences), and antihuman tumor necrosis factor alpha (TNF-α)/phycoerythrin (559321; BD Biosciences) for 40 minutes at 4°C. Cytokine production was analyzed by an LSR-II flow cytometer (BD Biosciences), and data analysis was performed using FACSDiva (BD Biosciences) or FloJo software (Tree Star Inc., Ashland, OR).

Fluorescence In Situ Hybridization of Bacterial 16S rRNA

Fluorescence in situ hybridization (FISH) of bacterial 16S ribosomal RNA (rRNA) on glass slides of liver sections was performed as described previously.[15, 16] The universal eubacterial oligonucleotide probe, EUB-338 (GCT-GCC-TCC-CGT-AGG-AGT), and the control probe, NONEUB-338 (CGA-CGG-AGG-GCA-TCC-TCA), complementary to EUB-338 to exclude nonspecific binding of the probes, were synthesized and 5-prime labeled (Metabion, Planegg/Martinsried, Germany) with the fluorochrome, Cy3. Liver sections (5 µm) were incubated with 25 ng of each oligonucleotide added to 50 μL of hybridization buffer containing 20% formamide for 90 minutes at 46°C before washing with the same stringency. Signal specificity was demonstrated by comparing to the nonrelated Cy3-labeled control NONEUB-338 oligonucleotide.

IL-17 Immunohistochemistry

Paraffine liver sections from patients with PSC (n = 18), autoimmune hepatitis (AIH; n = 5), nonalcoholic steatohepatitis (NASH; n = 5), ASH (n = 3), and PBC (n = 2) were stained for IL-17A with antihuman IL-17A antibodies (Abs) (IL-17A [H-132]: sc-7927; Santa Cruz Biotechnology, Heidelberg, Germany) and an Envision kit (EnVision+ System-HRP; Dako, Hamburg, Germany), according to the manufacturer's instructions.

Statistical Analysis

Comparison between groups was performed with one-way analysis of variance followed by Bonferroni's multiple comparison or by Dunn's multiple comparison test, depending on whether or not variables were normally distributed. Normal distribution was assessed by Kolmogorov-Smirnov's test. Significance is indicated as P < 0.05. All horizontal bars represent the median.


Pathogen Detection in Bile and Portal Tracts of Patients With PSC

Because inflammation in PSC is centered around bile ducts, we first investigated whether bile of PSC patients may be colonized with microbes. Therefore, bile was obtained during ERCP from 58 PSC patients. Microbial cultures could be grown from 41 of 58 individual bile specimens. In 19 of 41 cases, more than one microbial species was detectable (Fig. 1). Staphylococci (coagulase negative: 13×; S. aureus: 5×), streptococci (enterococci: 12×; α-hemolytic: 7×), and C. albicans (12×) were the main isolates detected.

Figure 1.

Bile from patients with PSC was obtained during ERCP and cultured for pathogens. In 41 of 58 patients, cultures grew positive. In 19 of 41 cases, more than one bacterial strain was detectable: staphylococci (18×), streptococci (19×), enterococci (12×), α-hemolytic streptococci (7×), and C. albicans (12×).

Because previous ERCP may be a risk factor for biliary bacterial colonization, we determined the rate of biliary interventions before bile sampling. In 23 of 41 cases of positive microbial bile cultures, ERCP had been performed previously. However, 24% of the analyzed patients with PSC had positive microbial cultures without previous manipulation of the biliary tract.

To investigate whether bacteria can be found not only in bile fluid, but also in liver tissue, liver sections from 6 PSC, 5 hepatitis C virus (HCV), and 4 AIH patients were stained for bacterial 16S rRNA using FISH. All 6 PSC patients showed bacterial 16S rRNA within portal tracts, whereas none of the 5 HCV patients and only 1 of 4 AIH patients showed positive staining (Fig. 2).

Figure 2.

Liver samples were analyzed using FISH for bacterial 16S rRNA. Six of six patients with PSC showed positive staining within portal tracts, whereas in only one of four with AIH and zero of five with HCV, bacterial RNA could be detected within the liver. Representative sections of two patients with PSC and one with AIH and HCV are shown. Nuclei are stained in blue.

To exclude an effect of previous endoscopic intervention on these findings, another set of liver sections obtained from 7 patients with PSC in whom previous ERCP could be excluded was investigated, where 6 of 7 stained positive for bacterial 16S rRNA. These findings confirm previous reports that bile of patients with PSC is frequently colonized with pathogens, including Candida, even in the absence of earlier endoscopic intervention. In addition, we could demonstrate that in patients with PSC, bacteria may enter portal tracts where they could potentially induce the stimulation and recruitment of immune cells.

Increased Frequency of Th17 Cells in Patients With PSC in Response to Pathogen Stimulation

After having shown that bacteria can be found in portal tracts of patients with PSC, we were interested in whether pathogen stimulation induces a proinflammatory phenotype in lymphocytes of patients with PSC. To this end, we determined whether stimulation of PBMCs with heat-inactivated bacteria or fungi may affect the expression of proinflammatory cytokines in vitro. Initially, pathogens isolated from patients' own bile were used for stimulation of PBMCs. The results obtained were not different to those obtained using standard pathogens, which were then used in subsequent stimulation assays. The clinical characteristics of PSC and PBC patients and cholestatic controls included in these experiments are shown in Table 1.

Table 1. Clinical Characteristics
Characteristics Median (Range)
  1. Data are presented as median value (ranges).

  2. Abbreviations: AST, aspartate aminotransferase; ALT, alanine aminotransferase; GGT, gamma-glutamyl transpeptidase; ALP, alkaline phosphatase.

PSC patients, n46 
Sex, female/male8/38 
IBD, yes/excluded40/6 
Age at time of inclusion, years40(19-71)
Duration of illness, years4(0-18)
Leukocytes, ×10−9/L5.7(2.6-16.7)
Total bilirubin, mg/dL0.7(0.3-15.9)
AST, U/L58(17-261)
ALT, U/L62(12-321)
GGT, U/L104(10-1,734)
ALP, U/L210(69-824)
PBC patients, n10 
Sex, female/male9/1 
IBD, yes/no signs of IBD0/10 
Age at time of inclusion, years60(39-72)
Duration of illness, years2.2(0.1-5.0)
Leucocytes, ×10−9/L7.6(5.5-8.9)
Total bilirubin, mg/dL0.6(0.4-1.1)
AST, U/L30(21-81)
ALT, U/L24(18-55)
GGT, U/L33(18-112)
ALP, U/L111(78-144)
Cholestatic controls, n13 
Sex, female/male5/8 
IBD, yes/no signs of IBD0/13 
Age at time of inclusion, years55(44-78)
Leukocytes, ×10−9/L9.9(5.5-32.5)
Total bilirubin, mg/dL2.6(0.3-28.0)
AST, U/L48(31-216)
ALT, U/L57(6-267)
GGT, U/L213(85-1,374)
ALP, U/L264(116-953)

Stimulation with facultative pathogenic bacteria, such as E. faecalis, induced significantly more IL-17A-producing CD4+ cells in patients with PSC, as compared to HCs (E. faecalis; CD4+IL-17A+: PSC [2.22% ± 1.68%] versus HCs [0.77% ± 0.54%], P < 0.001; Fig. 3A). To investigate whether these findings are specific for PSC, we compared these results to patients with PBC as another autoimmune cholestatic liver disease also treated with UDCA as well as to patients with different diseases leading to cholestasis, including SSC (see above). The increase in IL-17A-producing CD4+ T cells observed in PSC was not noted in patients with PBC or control cholestatic patients (E. faecalis, as compared to PSC: PBC: 0.57% ± 0.25%, P < 0.01; cholestatic controls: 0.53% ± 0.49%, P < 0.001; Fig. 3A). Because PSC-associated IBD may influence Th17 development through an impaired mucosal barrier function of the gut, patients with PSC and no evidence of IBD on colonoscopy were analyzed separately: Patients with PSC only were not different from patients with PSC and associated IBD, demonstrating that the increased frequency of IL-17A+CD4+ T cells is an underlying feature of PSC itself (E. faecalis; PSC only: 3.24% ± 2.21% [P < 0.001], as compared to HCs; Fig. 3A).

Figure 3.

PBMCs were stimulated with heat-killed bacteria or C. albicans for 12 days, then restimulated with PMA/Ionomycin and analyzed for cytokine production using FCM. After stimulation with E. faecalis, patients with PSC showed significantly increased rates of IL-17-producing CD4+ cells, as compared to HCs, PBC patients, or cholestatic controls (A). Patients with PSC, in whom IBD had been excluded (PSC-only), also showed elevated rates of Th17 cells after stimulation with E. faecalis (A). A similar increase in IL-17-producing CD4+ T cells was observed in PSC patients after stimulation with S. aureus (B). The highest rates of CD4+IL-17+ cells were obtained after stimulation with heat-killed C. albicans (C).

S. aureus was found in 9% of bile samples. Stimulation with heat-killed S. aureus also led to an increased rate of IL-17A-producing CD4+ T cells in PSC patients, but not in patients with PBC or HCs (Fig. 3B). As noted above, this effect was independent from the presence of IBD (Fig. 3B).

Interestingly, after stimulation with nonpathogenic heat-killed E. coli, there were no significant differences in IL-17A expression between PSC and HCs (data not shown). Also, rates of CD4+ T cells expressing IFN-γ or TNF-α after bacterial stimulation were similar between HCs and patients with PSC (Fig. 4).

Figure 4.

PBMCs were stimulated as described in the legend to Fig. 3. Expression of IFN-γ as well as TNF-α after stimulation with E. faecalis (A and D), S. aureus (B and E), and C. albicans (C and F) was similar in CD4+ T cells between groups.

C. albicans was cultured from 12 of 58 individual bile samples and was previously described to have a negative effect on progression of disease, including time to liver transplantation.[5] After stimulation with C. albicans, up to 30% of CD4+ T cells produced IL-17A, which were the highest rates observed in our experiments (C. albicans; mean values of CD4+IL-17A+: PSC [7.02% ± 8.46%] versus HCs [1.55% ± 1.78%], P < 0.01; Fig. 3C). The rate of IL-17 expression varied in every single individual, according to the pathogen used for stimulation, and did not seem to correlate with clinical patient characteristics. As noted with bacterial stimulation, no differences were detected for overall expression of IFN-γ or TNF-α within the population of CD4+ T cells (Fig. 4).

Th1/Th17 T Cells Are Increased in Patients With PSC

Th17 cells coexpressing IFN-γ (Th1/Th17 cells) have been reported to be of pathogenetic relevance in autoimmune diseases[17] and for immune response to C. albicans.[18] Therefore, we determined the rate of CD4+ T cells from peripheral blood expressing both IL-17A and IFN-γ after pathogen stimulation. Patients with PSC had higher rates of Th1/Th17 cells after stimulation with bacteria and especially after stimulation with C. albicans, as compared to patients with PBC (C. albicans; CD4+IL- 17A+-IFNγ+: PSC [4.5% ± 5.82%] versus HCs [1.10% ± 1.53%], P < 0.05; versus PBC [0.18% ± 0.15%], P < 0.01; Fig. 5B). These results further support the notion that PSC is associated with an increased Th17 response.

Figure 5.

PBMCs were stimulated as described in the legend to Fig. 3. Patients with PSC showed elevated rates of CD4+IL-17+IFN-γ+ (Th1/Th17) cells after stimulation with E. faecalis (A) and C. albicans (B), as compared to patients with PBC, healthy, or cholestatic controls. PBMCs were stimulated for 12 days with the TLR-5 ligand, flagellin (C), or the TLR-7 ligand, loxoribine (D), together with concanavalin A and were restimulated with PMA/Ionomycin. Upon stimulation of TLR-5 and TLR-7, PSC patients showed significantly higher rates of Th17 cells than HCs, patients with PBC, or cholestatic controls.

Increased Frequency of Th17 Cells in Patients With PSC in Response to TLR-5 and TLR-7 Stimulation

Pathogen detection is mediated by pattern recognition receptors, such as TLR. To define the pathways involved in pathogen-induced Th17 differentiation in PSC, we stimulated PBMCs with various TLR ligands (for the various ligands used, see above). After stimulation with the TLR-5 and TLR-7 ligands, flagellin and loxoribine, PBMCs of PSC patients showed a significant increase in the rate of CD4+IL-17A+ T cells, as compared to HCs and patients with PBC (TLR-5; PSC [8.28% ± 6.44%] versus HCs [3.77% ± 2.71%], P < 0.05; versus PBC [2.06% ± 1.04%], P < 0.05; Fig. 5C; TLR-7: PSC [4.64% ± 2.96%] versus HCs [1.98% ± 1.64%], P < 0.01; versus PBC [0.83% ± 0.37%], P < 0.01; versus cholestatic controls [1.08% ± 1.18], P < 0.001; Fig. 5D). Stimulation of TLR-2, TLR-4, TLR-6, as well as NOD-2 and dectin-1 receptor, did not lead to significant differences in IL-17 expression.

To summarize these results, PSC patients seem to have an increased Th17 response after stimulation with heat-inactivated pathogens, which are present in bile fluid of the majority of patients. Similar results obtained with selective TLR ligands may guide future studies investigating signaling pathways involved in this response.

IL-17A-Expressing T Cells Are Detected in Livers of Patients With PSC

Because Th17 cells are important both for pathogen defense and for autoinflammatory responses, we aimed to determine their localization within livers of patients with PSC. Therefore, we stained liver sections of 18 patients with PSC with Abs to IL-17A. Indeed, in all patients, IL-17A+ lymphocytes were detected in frequencies from 0.5% to 5% of all lymphocytes. IL-17A+ lymphocytes aggregated around bile ducts and in areas of neoductular proliferation, whereas in fibrotic septae and liver lobules, very few IL-17A+ lymphocytes were detected (Fig. 6). The numbers of IL-17 expressing cells per portal tract were significantly increased in PSC, as compared to control patients with AIH (P < 0.01), NASH (P < 0.01), or ASH (P < 0.05).

Figure 6.

Liver tissue from 10 PSC patients was stained for IL-17A. All PSC patients showed IL-17A-producing cells located mainly in areas of neoductular proliferation (A) and around bile ducts (B and C). Between 0.5% and 5.0% of all lymphocytes stained positive for IL-17A. Fibrotic septae contained only a few IL-17+ cells (D). Representative stainings are shown.


The pathogenesis of PSC is unknown. Clinical observations and the association with IBD suggest that dysregulated immune responses upon microbial stimulation may be involved in disease pathogenesis. This is supported by recent GWAS demonstrating polymorphisms in genes relevant to pathogen defense and in genes involved in the generation of Th17 cells.[14, 19, 20] Th17 cells are important players in bacterial and fungal defense.[21] Furthermore, Th17 cells have been implicated in autoimmune inflammation in various murine models as well as human autoimmune diseases.[22] Here, we report that patients with PSC show increased Th17 responses toward pathogen stimulation in vitro, which was independent from the presence of IBD. Furthermore, IL-17A-expressing lymphocytes as well as bacterial RNA were found within portal tracts of PSC livers.

In PSC, autoimmunity is discussed as one of the pathogenetic mechanisms,[23] supported by recent observations that genetic polymorphisms may alter the binding capacity of human leukocyte antigen class II molecules in patients with PSC.[24] Th17 cells have emerged as a major proinflammatory Th cell subset, which can induce autoimmunity in mouse models.[25] In humans, the presence of Th17 cells in inflamed tissue has been described in several autoimmune and immune-mediated diseases, such as psoriasis, rheumatoid arthritis, multiple sclerosis, asthma, and IBD.[26] In addition, treatment trials investigate the role of blocking IL-17A in several human diseases.[27] Besides their role in promoting tissue inflammation, Th17 cells are induced by pathogens to aid their elimination. This has been described for the clearance of Candida and bacterial pathogens in mice[28] and for Candida in humans, as shown in patients with hyperimmunoglobulin E syndrome.[12, 29] It is tempting to speculate that an increased exposure to pathogens or a change in the microbial community in bile[20] may induce a dysregulated Th17 response, which may then contribute to uncontrolled portal and biliary inflammation in PSC.

In clinical practice, recurrent bacterial cholangitis leads to the rapid progression of PSC. This has been supported by data demonstrating that the culture of Candida species in bile is a risk factor for the progression to end-stage disease.[5] Candida was present in 20% of bile cultures reported on here. Earlier studies have suggested that the rate of bacterial colonization is high in patients with PSC.[4, 6] Here, for the first time, we describe that bacterial RNA can be found within portal tracts of PSC patients, but not so of patients with chronic HCV or AIH as controls. This suggests that pathogens may either pass the biliary epithelial barrier or may enter the liver through portal blood flow. It has to be critically stated that in spite of precautions taken during ERCP, such as peri-interventional antibiotic treatment, it is not possible to perform biliary cannulation in a sterile manner. Therefore, pathogens could be introduced by the procedure itself. Although bile was not obtained by ERCP in the study by Olsson et al., they reported previous ERCP as a risk factor for positive bile cultures.[6]

After having shown the presence of pathogens in bile and portal tracts of patients with PSC, we investigated whether PSC patients may manifest an increased Th17 response after pathogen stimulation. Here, we report that in patients with PSC, but not in patients with PBC, stimulation of PBMCs with heat-inactivated bacteria led to a marked induction of Th17 responses. There was no difference between patients' own pathogens and standard pathogens, but it should be noted that nonpathogenic E. coli strains were not able to elicit an increase in IL-17A expression. Of note, and in line with the deleterious effects of Candida cholangitis on the progression of disease, stimulation of PBMCs with inactivated C. albicans led to the highest expression of IL-17A in up to 30% of CD4+ T cells (Fig. 3C). In addition, more Th17 cells were found to coexpress IFN-γ (Th1/Th17 cells) after stimulation with E. faecalis or C. albicans (Fig. 5). These cells may be especially relevant for the development of autoimmune diseases[30, 31] and for memory functions involved in the defense against S. aureus and C. albicans.[28]

IL-17A-expressing lymphocytes could be detected around bile ducts of patients with PSC. Whereas these cells can be found in other inflammatory liver diseases as well,[32] it is interesting to note that IL-17A receptors are expressed on biliary epithelial cells (BECs), and that upon stimulation with IL-17, BECs produce proinflammatory cytokines, such as IL-1β, IL-6, and IL-23.[33] These cytokines could, in turn, promote the survival of Th17 cells. It is tempting to speculate that these cytokines could not only increase the survival of Th17 cells themselves,[34] but also stimulate fibroblasts to induce periductular fibrosis.[35, 36]

Selective stimulation of the TLR-5 and −7 ligands, but not stimulation with other TLR ligands, also led to the induction of IL-17A in PSC, but not in the control groups. Further elucidation of signaling pathways involved may help to understand this aberrant response to pathogens.

These results are reminiscent of data obtained in IBD patients, where the IL-23/Th17 axis has been reported to shape inflammatory response in the gut.[11] In children with IBD, an aberrant Th17 response to TLR stimulation and stimulation with Candida has been demonstrated previously.[37] Additionally, polymorphisms in genes relevant for the generation and maintenance of Th17 cells, such as the IL-23 receptor, were highly associated with IBD in GWAS.[38] Of note, in the patients reported on here, aberrant Th17 response was independent from the presence or absence of IBD, strongly suggesting that this is a feature of PSC itself and not the associated IBD.

How might these findings relate to the growing body of evidence for genetic associations in PSC?[13, 19, 20] Of the non-MHC loci associated with PSC, most genes are involved in immune reactions. Of relevance for this study are the associations within the CARD9 and REL loci.[14] CARD9 is involved in the signaling cascade subsequent to the stimulation of dectin-1 by selective ligands or Candida species and modulates activation of the nuclear factor kappa B subunit, c-REL, and the activation of p38 and c-Jun N-terminal kinase.[39-41] It has recently been described that dectin-2 may also be involved in the induction of IL-17 expression, which may explain why we could not detect an increase in IL-17 expression using the dectin-1 ligand, depleted zymosan.[41] CARD9 has been described in the activation of dendritic cells during fungal infections and their capacity to induce Th17 cells by producing proinflammatory cytokines, particularly IL-23.[41, 42] Future studies will have to investigate whether the polymorphisms described in patients with PSC result in an altered Th17 response.

In conclusion, we report here an increased Th17 and Th1/Th17 response toward pathogen stimulation in patients with PSC, which was independent of the presence of associated IBD. The highest IL-17A expression was observed after stimulation with C. albicans, a pathogen associated with disease progression in PSC. Th17 response could be induced by the selective stimulation of TLR-5 and −7, enabling us to explore the signaling pathways involved in this response. Because Th17 cells may also have beneficial effects in the complex pathogenesis of PSC, the exact roles of IL-17A and other Th17 cell-associated cytokines, such as IL-22, need to be clarified before IL-17 could be regarded as a therapeutic target in PSC.


The authors thank Agnes Malotta, Marko Hilken, Lars Tharun, and Gerlinde Apitzsch for their excellent technical assistance.