Between January 2014 and December 2016, liver transplant patients and patients with various liver diseases (who underwent a liver biopsy as part of their diagnostic evaluation or follow up) and healthy volunteers were enrolled in this study. Both female and male patients were included with an age range from 18 to 60 years (Table 1). Patients with human immunodeficiency virus (HIV) infection and hepatocellular carcinoma were excluded due to the immunomodulatory activity of tumor cells. Liver tissue samples were obtained at the time of liver transplantation from healthy donors (n=6) and liver explants (n=11), liver biopsies (n=11) were also obtained from patients with liver disease during follow up. The liver biopsies were divided in two parts, one part for the histological examination and the other part for cell isolation and flow cytometry. All samples were classified at pathology department depending on the severity of fibrosis based on the METAVIR score in fibrotic samples (fibrosis: F1 (mild) and F2 (moderate)) and cirrhotic samples (cirrhosis: F3 and F4). As a healthy control (healthy), we used biopsies obtained from the liver donor during liver transplantation classified as F0 by the METAVIR score. Peripheral blood samples were obtained from patients with liver disease, with liver cirrhosis undergoing transplantation, and healthy volunteers. Our program adheres to the ethical principles of the Istanbul Declaration as well as the document of the Pontifical Academy of Sciences on organ trafficking and transplant tourism. Ethical permission for the study was obtained from the Ethics Committee Board at Favaloro Hospital (DDI(1264)3214), and written informed consent was obtained from each subject included in the study.

Samples of liver tissue were collected in PBS with pH 7.0 containing 0.5mg/mL collagenase (type IV, 312U/mg, Sigma-Aldrich, St. Louis, USA) and then thoroughly dissected using scissors. The sample solution was incubated for 15min at 37°C (enzymatic digestion), and then centrifuged at 30×g for 1min to remove cell clumps and not dissociated tissue. The supernatant was centrifuged at 300×g for 3min, then the pellet was incubated with ACK lysing buffer for red blood cell lysis for another 3min at room temperature. Then, the cells were passed through a 70μm cell strainer and washed with PBS. Finally, the cells were concentrated by centrifugation and resuspended in the final volume to be stained with the antibody cocktail.

Peripheral blood mononuclear cells (PBMC) were isolated freshly from heparinized blood by standard Ficoll-Hypaque (GE Healthcare, USA) density gradient centrifugation. Then, the cells were harvested from the interphase, passed through a 70μm cell strainer, and washed twice with PBS.

Cells were stained with a combination of the following monoclonal antibodies: PE anti-CD127 (clone: hIL-7R; BD Bioscience, New Jersey, USA), FITC anti-CD69 (FN50; BD Bioscience), PerCPeCy5.5 anti-CD294 (BM16; BD Bioscience), APC anti-CD45 (H130; BD Bioscience), APCH7 anti-CD3 (SK7; BD Bioscience), APCCy7 anti-CD19 (BDIB19; Biolegend, San Diego, USA), PECy7 anti-CD117 (104D2; Biolegend). Six-color analysis was performed by FACSCanto II (Becton Dickinson FACS System, Oxford, UK), and the data were analyzed by Flow Jo software (Tree Star, Ashland, USA).

Peripheral blood mononuclear cells were isolated from fresh heparinized blood by standard Ficoll-Hypaque (GE Healthcare) density gradient centrifugation. Then, the cells harvested from the interphase were passed through a 70μm cell strainer and washed twice with PBS. Finally, the cells were concentrated by centrifugation in the final concentration to be stimulated. PBMC (2×106 cells in 200μl) were then stimulated with IL-7 (100ng/mL; Peprotech, NJ, USA); IL-33 (100ng/mL; Biolegend); the synthetic TLR2 ligand Pam3Cys (Novabiochem, Cambridge, UK) at 1μg/mL; or with the TLR4 ligand LPS (1μg/mL; Sigma-Aldrich). Samples were incubated at 37°C and 5% CO2 overnight in RPMI 1640 medium (Thermo Fisher Scientific, Waltham, USA) with 10% FBS (PAA, Pasching, Austria) and an antibiotic-antimitotic cocktail (Thermo Fisher Scientific). Then, cells were centrifuged at 300×g for 5min and stained for flow cytometry analysis.

For the determination of human IL-33 concentration in plasma, specific ELISA Kit from Abcam was used according to the manufacturer's instructions (ab119547-IL-33; Abcam, Cambridge, UK).

Data analysis was performed with GraphPad Prism version 5 software (San Diego, USA). Multiple comparisons were made among the groups using the Kruskal–Wallis with Dunn's post-test. For the in vitro cell analysis, Student's t test was used. Correlations between variables were evaluated with the Spearman rank correlation test. For all tests, a two-sided P value <0.05 was considered significant.

To determine the role of ILC2 in liver fibrosis, we first evaluated the feasibility of detecting ILC2 in hepatic tissue (hepatic ILC2). So far in non-lymphoid human organs, ILC2 has been found in the intestine [18], the skin [12], and the lung [9]. We obtained hepatic tissue from biopsies (either healthy or fibrotic) or samples from liver explants and performed cell analysis by flow cytometry. To identify ILC2 within the lymphocyte gate (CD45+ mononuclear cells with lymphoid morphology), we used the CD3 and CD19 markers to exclude B and T cells. Then, CD127 positive cells were gated and the specific marker prostaglandin D2 receptor (CRTH2, CD294) identified subsequently ILC2 (Fig. 1A). We used the same antibody cocktail to identify ILC2 in the peripheral blood (peripheral ILC2) (Fig. 1E). We detected ILC2 in all hepatic tissue samples within a range of 0.03–0.89% of CD45+ lymphoid cells (Fig. 1B). This range was considerably higher than the percentage of peripheral blood ILC2 (0.001–0.16% of CD45+ lymphoid cells) (Fig. 1F). To elucidate whether ILC2 might contribute to the fibrogenic process, we compared the healthy samples with fibrotic and cirrhotic samples. The percentage of hepatic ILC2 for the healthy group was 0.1±0.02%, the fibrotic group increased to 0.29±0.05% and the cirrhotic group to 0.33±0.07% (Fig. 1B). When we correlated the percentages of hepatic ILC2 with their individual METAVIR fibrotic scores (F0–F4), we found a positive correlation (Fig. 1C). The absolute cell number of ILC2 per gram of hepatic tissue increases with worsening fibrosis. However, differences between the three groups were not statistically significant (Fig. 1D). Also, we investigated whether the ILC2 number of the paired peripheral blood samples correlates with the fibrosis stage of the liver. On the contrary to the hepatic ILC2, cell percentages of peripheral ILC2 did not show significant differences between the three groups or in the correlation with the METAVIR score (Fig. 1F and G). Our results show that ILC2 are present in human liver tissue, and can be detected in small liver biopsies, although in a low percentage. The correlation between ILC2 frequency and METAVIR score also suggests that ILC2 contribute to the development of liver fibrosis and progressive liver disease.

Fig. 1.

Detection of innate lymphoid cells group 2 in human hepatic tissue and peripheral blood. Representative plots of the gating strategy for the identification of ILC2 (CD45+ lymphocyte gate, CD3−, CD19−, CD127+, CD294+) by flow cytometry in (A) hepatic tissue or (E) peripheral blood. (B) Comparison of the percentages of ILC2 isolated from hepatic tissue of healthy individuals (n=7) with histologically confirmed METAVIR score F0, patients with fibrosis (n=9, F1–F2), and with cirrhosis (n=12, F3–F4). (C) A positive correlation between percentages of ILC2 and METAVIR scores is shown. (D) Absolute numbers of ILC2 per gram of hepatic tissue were not significantly different between the three groups. (F) Percentages of ILC2 from peripheral blood of healthy (n=16), fibrotic (n=8), and cirrhotic (n=11) samples were compared and no significant difference was detected. (G) No correlation between percentages of peripheral blood ILC2 and METAVIR score was found. Each data point represents an individual patient and the values are shown as mean±SEM. The groups are compared using Kruskal–Wallis test with Dunn's post-test. Correlation analysis is performed by Spearman rank correlation test (r and P values are indicated in the figure). *P<0.05. ILC2, innate lymphoid cells group 2; FSC, forward scatter; SEM, standard error of the mean.

(0.33MB).

Moreover, we wanted to elucidate whether ILC2 were activated in the fibrotic liver and in the peripheral blood of the patients enrolled in this study. We included the activation marker CD69 in the antibody cocktail for the specific ILC2 detection by flow cytometry (Fig. 2A and C). We observed that ILC2 were significantly more activated in fibrotic and cirrhotic tissue than in healthy tissue (Fig. 2B). Strikingly, the percentage of activated hepatic ILC2 strongly correlated with the METAVIR score (Fig. 2E). On the contrary, the results in peripheral blood were not so conclusive. Although percentages of activated ILC2 in the blood of fibrosis or cirrhosis patients were increased, only the cirrhotic group expressed a significant difference as compared to the healthy group (Fig. 2D). The range of the activated ILC2 level was much higher in the fibrotic and cirrhotic group of hepatic tissue than in the paired samples of peripheral blood (21.1–66.7% vs. 0–24.6%, respectively), supporting our hypothesis that ILC2 are activated locally in the liver (Fig. 2B and D). The level of activated ILC2 in peripheral blood was moderately correlated with the METAVIR score (Fig. 2F). A slight correlation of the activation status of ILC2 between paired samples was observed (Fig. 2G).

Fig. 2.

Elevated level of activated ILC2 in liver fibrosis. (A) Representative plots of the gating strategy for the identification of activated ILC2 with the commonly used activation marker CD69 (CD45+, CD3−, CD19−, CD127+, CD294+, CD69+). (B) Percentages of activated ILC2 of hepatic tissue samples from healthy individuals (n=7) with a METAVIR score F0; fibrotic patients (n=9) with F1–F2; and cirrhotic patients (n=12) with F3–F4. (C) Representative plots of activated ILC2 from peripheral blood. (D) Percentages of activated ILC2 from peripheral blood of healthy individuals (n=16); patients with fibrosis (n=8); patients with cirrhosis (n=11). (E) A strong correlation between activation status of ILC2 and fibrotic METAVIR score in hepatic tissue is shown, whereas (F) this correlation is weaker in peripheral blood. (G) The activation level of ILC2 correlates between the paired samples of hepatic tissue and peripheral blood. Each data point represents a patient and the values are shown as mean±SEM. The groups were compared using Kruskal–Wallis test with Dunn's post-test. Correlation analyses were performed by Spearman rank correlation test. *P<0.05, **P<0.01.

(0.25MB).

To confirm that IL-33 induces CD69 expression on ILC2, we performed an in vitro experiment with freshly isolated peripheral blood cells. After 24h in vitro stimulation, we found that the mean percentage of CD69+ ILC2 was increased 4-fold in the IL-33 group compared to the control group without IL-33 (4.7±1.2 (SEM) vs. 19.2±2.6%) (Fig. 3A). In addition, when we included TLR2 stimulus PamC3 and the TLR4 stimulus LPS in the IL-7 and IL-33 cocktail, we observed that both stimuli could not significantly further increase the percentage of CD69+ ILC2. These different stimuli did not induce upregulation of CD69 on CD3+ CD69+ CD294+ Th2 cells under the same conditions (Fig. 3B). These results show that hepatic ILC2 are activated locally in fibrotic tissue and the frequency of activated ILC2 increases with progression of liver fibrosis, supporting our hypothesis that ILC2 actively participate in the pathophysiology of the fibrotic process in the liver.

Fig. 3.

ILC2 express elevated level of CD69 after IL-33 stimulation in vitro. Isolated peripheral blood mononuclear cells from freshly blood of healthy volunteers were stimulated with IL-7, IL-33, the synthetic TLR2 ligand Pam3Cys, or with the TLR4 ligand LPS. (A) Levels of the activation marker CD69 were measured by flow cytometry and results expressed as mean percentages of CD69+ ILC2. (B) Mean of percentages of activated CD3+ CD294+ Th2 cells exhibit no statistical difference between the study groups. The values are shown as mean±SEM and the groups are compared using unpaired Student's t test. *P<0.05. IL, interleukin; TLR, Toll-like receptor. Data are shown of three independent experiments, except for the LPS group with two experiments.

(0.07MB).

IL-33 is the main cytokine activating ILC2. IL-33, located in different parenchymal cells, is released locally into the tissue microenvironment after cell damage and most probably is active only locally. However, IL-33 might reach peripheral blood and be detectable in serum. We therefore measured plasma level of IL-33 and correlated the amounts with the METAVIR score. A significant increase of plasma IL-33 was observed only in cirrhotic patients with METAVIR 4 (Fig. 4A) and interestingly, the IL-33 level positively correlates with the METAVIR score (Fig. 4B).

Fig. 4.

Increased plasma levels of interleukin-33 in cirrhotic patients. (A) Levels of IL-33 (pg/mL) in plasma from healthy individuals (n=4), fibrotic (n=5; METAVIR F1–F2) or cirrhotic (n=6; F3–F4) patients estimated by ELISA. (B) Correlation between IL-33 plasma level (pg/mL) and fibrosis METAVIR score. Each symbol represents a patient; lines indicate mean values and SEM. Data assessed using Kruskal–Wallis with Dunn's post-test and Spearman rank correlation test. *P<0.05.

(0.05MB).