Abnormalities of liver sinusoidal endothelial cells in primary biliary cholangitis

Abstract

BACKGROUND

Liver sinusoidal endothelial cells (LSECs) may have a critical role in the pathogenesis of primary biliary cholangitis (PBC) that has not been investigated.

AIM

To investigate the role of LSECs in PBC.

METHODS

We studied the levels of soluble vascular cell adhesion molecule-1, soluble intercellular adhesion molecule-1 and soluble E-selectin and the tissue plasminogen activator (t-PA) and its inhibitor in the serum of 30 PBC patients before and 25 patients after treatment with ursodeoxycholic acid (UDCA). Moreover, immortalized endothelial cells (EA.hy926) were incubated with serum from patients with PBC, hepatitis C virus (HCV) and normal controls for up to 24 hours. The expression of endothelin (ET) 1, ET2, ET3, and ET receptors A and B were also measured by quantitative polymerase chain reaction.

RESULTS

Vascular cell adhesion molecule and intercellular adhesion molecule were significantly increased in PBC and HCV with the highest values found in PBC patients. UDCA had no effect. Levels were significantly higher in late PBC (stages III-IV), compared with early PBC (stages I-II). t-PA was significantly increased in PBC but not in HCV. Higher values were obtained in late PBC. UDCA decreased t-PA. Plasminogen activator inhibitor-1 levels were similar in all groups. Expression of endothelin 1, endothelin 2, and endothelin 3 significantly varied at different time points. ET receptors A was decreased at 2 hours and 6 hours in PBC, and at 2 hours and 24 hours in HCV. ET receptors B was reduced at 2 hours and 24 hours in both PBC and HCV.

CONCLUSION

Endothelial adhesion molecules are abnormal in PBC particularly in the late fibrotic stages. ET and their receptors are reduced in LSECs after incubation with PBC and HCV sera, findings that might be related to pathogenesis.

Key Words: Primary biliary cholangitis; Sinusoidal endothelial cells; Endothelial markers; Endothelins; Endothelin receptors

Core Tip: Liver sinusoidal endothelial cells (LSECs) are implicated in the pathogenesis of various liver diseases including primary biliary cholangitis (PBC). In the present study, we demonstrated that LSECs biomarkers are significantly altered in the serum of patients with PBC but also in patients with chronic hepatitis C virus. Moreover, we incubated the human endothelial cell line EA.hy926 with sera from patients with PBC and hepatitis C virus and found that the expression of endothelins and their receptors significantly fluctuated at different time points. Collectively, these findings indicate that LSECs may be implicated in the pathogenesis of PBC.

INTRODUCTION

Primary biliary cholangitis (PBC) is a chronic non-suppurative destructive cholangitis, associated with the presence of anti-mitochondrial antibodies and chronic cholestasis ultimately leading to cirrhosis[1]. The prevalence of PBC is high in North America and Europe followed by the Asia-Pacific. Over the last 25 years, disease epidemiology has changed all over the world. The incidence and prevalence of PBC have increased with an incidence rate ranging between 0.23-5.31/100000 population and a prevalence ranging between 1.91-40.2/100000[2,3].

Variations exist within geographical areas. Characteristically, in the Greek island of Crete, where the prevalence and incidence of PBC are among the highest published in Europe, there is a significant variation between the east and the west of the island[4]. The pathogenesis of PBC is still unclear and a number of mechanisms have been proposed. Our earlier studies indicated that endothelins produced by liver sinusoidal endothelial cells (LSECs) may be implicated in the pathogenesis of PBC[5]. Additional studies have indicated that LSECs may be implicated in PBC. Genetic association of endothelial nitric oxide synthase (eNOS) was found in PBC patients. Thus, eNOS intron4 variable number tandem repeats and eNOS exon7 G894T single nucleotide polymorphism were significantly associated with increased risk of PBC[6].

LSECs comprise 20% of the total liver cell population and play a vital role in the regulation of innate and adaptive immunological functions and the progress of liver fibrosis[7,8]. LSECs can be anatomically divided into periportal or zone 1 LSECs, mid-zonal or zone 2 LSECs, and pericentral or zone 3 LSECs. Whether this division has any functional significance is not clear at the time. It should be noted however, that most LSEC-associated genes are differentially expressed along the liver lobule, indicating a possible functional heterogeneity within LSECs[9]. Periportal LSECs mostly express immune regulatory markers such as CD34 and von Willebrand factor, whereas pericentral LSECs mostly express proangiogenic factors such as Wnt2 and Rspo3 and regeneration-associated markers such as Oit3[10,11].

An overview of LSECs has been recently published[12]. LSECs secrete chemokines (C-X-C motif chemokine ligands 9, 10, C-C chemokine ligand 25) and express adhesion molecules such as intercellular adhesion molecule 1 (ICAM-1), vascular cell adhesion molecule-1 (VCAM-1) and E selectin, all very important for the recruitment and liver infiltration of lymphocytes[13,14]. Circulating molecules produced by LSECs have been used as markers of endothelial dysfunction in patients with hemophilia and atherosclerosis. They include, among others, markers such as P- and E-selectin, ICAM-1, VCAM-1, tissue-type plasminogen activator (t-PA), and plasminogen activator inhibitor-1 (PAI-1)[15-17].

In the present study, we used circulating LSECs markers in patients with PBC to identify LSECs dysfunction and the possible effect of treatment with ursodeoxycholic acid (UDCA). Moreover, we performed in vitro experiments using an immortalized human endothelial cell line to study the effects of incubation with plasma of PBC patients on the expression of endothelins and their receptors.

MATERIALS AND METHODS

Patients

PBC diagnosis was established on the basis of a cholestatic biochemistry, positive anti-mitochondrial antibodies by immunofluorescence, anti-M2 by enzyme-linked immunosorbent assay and a compatible liver biopsy according to the guidelines of the European Association for the Study of the Liver for PBC[18]. Chronic hepatitis C virus (HCV) was also verified by liver biopsy.

Serum was obtained from 30 patients (27 females, age 21-67 years) with PBC. Eighteen patients were classified as early stage (I-II) and 12 were classified as late stage (III-IV) PBC according to Ludwig et al[19]. Sera from 15 patients with chronic moderate or severe HCV (13 females, age 40-65 years) were used as disease controls and compared with 10 normal controls (9 females, age 32-64 years). After 6 months of UDCA treatment (13-15 mg/kg body weight) serum was available from 25 PBC patients. Demographic characteristics of PBC patients are presented in Table 1.

Table 1 Demographic characteristics of primary biliary cholangitis patients (n = 30), mean ± SD.

Characteristics
Median age	48 (21-67)
Advanced disease stage (III-IV)	40%
AMA positive	100%
ANA positive	47%
IgM (> 170 mg/dL)	86%
IgM (mg/dL)	363 ± 198
ALP at diagnosis (U/L)	263 ± 171
Total bilirubin (mg/dL)	1.06 ± 1.38
Albumin (g/dL)	4.22 ± 0.51
Mayo risk score	4.31 ± 1.18

AMA: Anti-mitochondrial antibodies; ANA: Anti-nuclear antibodies; IgM: Immunoglobulin M; ALP: Alkaline phosphatase.

Methods

All chemicals were from Sigma-Aldrich (St. Louis, MO, United States) unless otherwise stated. ICAM-1, VCAM-1, t-PA, PAI-1 and E selectin were measured using the human Quantikine enzyme-linked immunosorbent assay kits according to manufacturer instructions (all from R&D Systems, Minneapolis, MN, United States). Data were read in a Biotek ELx800 microplate reader (Winooski, VT, United States).

Cell cultures

The human endothelial cell line EA.hy926 was purchased from DSMZ (Braunschweig, Germany). Cells were cultured in RPMI-1640 (Gibco^TM, Thermo Fisher Scientific, MA, United States) for 24 hours at 37 °C and 5% CO₂. Fetal bovine serum was substituted by 10% patient or control serum. Five experiments were performed for each group. Cells were collected at different time-points and washed with PBS. Cell viability was accessed by trypan blue exclusion test at the end of each experiment using wells with cell viabilities greater than 95%.

Real time polymerase chain reaction

Immediately after culture, cells were lysed to obtain mRNA, using a Nucleospin RNA II isolation kit (Macherey-Nagel, Duren, Germany). Total RNA was extracted from 5 × 10⁶ cells. Real time was performed with Applied Biosystems^TM HighCapacity cDNA Reverse Transcription Kit (Foster City, CA, United States). Absence of DNA was verified by polymerase chain reaction (PCR) for glyceraldehyde 3-phosphate dehydrogenase. Quantitative PCR was performed as described previously[20] in an StepOnePlus RealTime PCR System (Foster City, CA, United States) with KAPA-SYBR-FAST qPCR Master-Mix (Wilmington, MA, United States). Relative gene expression was calculated with the ΔΔCT method and each sample was corrected for the control of each time point (untreated sample). Changes were normalized according to cyclophilin-A and expressed as times above or below the control value regarded as 1. All primers were selected from the qPrimer Depot (qPrimerDepot accessed at http://primerdepot.nci.nih.gov, currently PrimerBank https://pga.mgh.harvard.edu/primerbank/). All the primers used are presented in Table 2 and were synthesized by VBC-Biotech (Vienna, Austria).

Table 2 Primers used in quantitative polymerase chain reaction.

Item
Gene name ENDRB	RefSeq ID: NM_000115
Right primer sequence	CCACGGAGGCTTTCTGTATG; Tm: 60.647
Left primer sequence	CACATCGTCATTGACATCCC; Tm: 59.769
cDNA amplicon size	109
Gene name ENDRA	RefSeq ID: NM_001957
Right primer sequence	ATTGTGATCAAAAGGCCAGC; Tm: 60.081
Left primer sequence	GATAGCCAGTCTTGCCCTTG; Tm: 59.836
cDNA amplicon size	100
Gene name ET1	RefSeq ID: NM_001955
Right primer sequence	GACTGGGAGTGGGTTTCTCC; Tm: 60.891
Left primer sequence	TCTCTGCTGTTTGTGGCTTG; Tm: 60.175
cDNA amplicon size	103
Gene name ET2	RefSeq ID: NM_001956
Right primer sequence	TTCCTCCACCTGGAATGTGT; Tm: 60.363
Left primer sequence	AGAGCTTCTCCAAAGGCTGA; Tm: 59.308
cDNA amplicon size	108
Gene name ET3	RefSeq ID: NM_000114
Right primer sequence	GGACAGTCCATAGGGCACC; Tm: 60.342
Left primer sequence	GCACGTGCTTCACCTACAAG; Tm: 59.515
cDNA amplicon size	101
Cyclophilin A	ATGGTCAACCCCACCGTGT; TTCTGCTGTCTTTGGAACTTTGTC

ENDRB: Endothelin receptor B; RefSeq ID: Reference sequence identifier; ENDRA: Endothelin receptor A; ET: Endothelin.

Statistical analysis

The statistical review of the study was performed by a biomedical statistician. Results are expressed as mean ± SD of the mean. Box and whiskers plots were used to demonstrate levels of soluble LSECs markers indicating the first quartile, mean, third quartile, and the inter-quartile range. The Kolmogorov-Smirnov test was used to check the Gaussian data distribution and Bartlett’s test was used to assess the variance of SDs. The non-parametric analysis of variance Kruskal-Wallis’s test was used for comparisons among groups, while the unpaired T-test with Welch correction for Gaussian data distribution or otherwise the Mann-Whitney test for non-Gaussian distribution were used. P ≤ 0.05 was considered statistically significant. Statistical tests were performed with the GraphPad Prism 10.6.0 software.

RESULTS

Soluble VCAM

PBC patients before treatment (111.3 ± 78.4 ng/mL) and HCV patients (66.4 ± 28.6) had higher values than controls (34.5 ± 9.9, Kruskal-Wallis P = 0.0018). There was a significant difference between controls and PBC before (P = 0.0002) and after treatment (89.3 ± 55.9, P = 0.003). There was a reduction of soluble VCAM (sVCAM) values after treatment but the difference was not statistically significant. Similarly, a significant difference existed between controls and HCV (P = 0.004). Importantly, there was a significant difference (P = 0.0001), between early (stages I-II) and late PBC (stages III-IV) (Figure 1A).

Figure 1 Primary biliary cholangitis before (30 patients) and after ursodeoxycholic acid treatment (25 patients) compared to 12 normal controls and 15 patients with chronic hepatitis C virus. Values for early (18 patients) and late primary biliary cholangitis (12 patients) are also presented. A: Soluble vascular cell adhesion molecule levels; B: Soluble intercellular adhesion molecule levels; C: Soluble E selectin levels; D: Serum tissue plasminogen activator; E: Serum plasminogen activator inhibitor-1. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001. NS: Not significant; sVCAM: Soluble vascular cell adhesion molecule; HCV: Hepatitis C virus; PBCb: Primary biliary cholangitis before treatment with ursodeoxycholic acid; PBCa: Primary biliary cholangitis after treatment with ursodeoxycholic acid; PBC: Primary biliary cholangitis; sICAM: Soluble intercellular adhesion molecule levels; eSelectin: Soluble E-selectin; t-PA: Tissue plasminogen activator; PAI: Plasminogen activator inhibitor 1.

Soluble ICAM

PBC patients before treatment (38.8 ± 13.9 ng/mL) and HCV patients (25.6 ± 4.2) had higher values than controls (16.7 ± 5.2, Kruskal-Wallis P < 0.0001). There was a significant difference between controls and PBC before (P < 0.0001) and after treatment (35.4 ± 14.1, P < 0.0001). There was no reduction of soluble ICAM (sICAM) values after treatment. Similarly, a significant difference existed between controls and HCV (P = 0.00038). Importantly, there was a significant difference (P = 0.007), between early (stages I-II) and late PBC (stages III-IV) (Figure 1B).

E selectin

There was no significant difference between controls (24.2 ± 7.2 ng/mL) and either PBC (32.9 ± 12.1) or HCV patients (32.8 ± 12.3, Kruskal-Wallis P = 0.15). PBC treatment slightly decreased E selectin levels (27.3 ± 8.7), but the difference was not significant. However, there was a significant difference between early (29.1 ± 8.2) and late PBC (43 ± 13.6, P < 0.0038) (Figure 1C).

t-PA

PBC patients before treatment (5.5 ± 2.9 ng/mL) had higher values and HCV patients had similar values (2.2 ± 1.8) compared to controls (2.6 ± 0.7, Kruskal-Wallis P < 0.004). There was a significant difference between controls and PBC before treatment (P = 0.007). There was a reduction of t-PA values after treatment (P = 0.026). Importantly, there was a significant difference (P = 0.002), between early (stages I-II) and late PBC (stages III-IV) (Figure 1D).

PAI-1

There were no significant differences between controls (19.6 ± 5.5 ng/mL) and HCV (18.1 ± 4.8) or PBC before (21.8 ± 8.5) and after treatment (17.6 ± 9.1 Kruskal-Wallis P = 0.27). A decrease was noted between PBC before and after treatment, but the difference was not statistically significant (P = 0.12). No difference was also found between early and late PBC (P = 0.31) (Figure 1E).

Endothelin 1

Figure 2A shows the results of endothelin 1 expression after incubation of EA.hy926 cells with sera from PBC and HCV patients expressed as times over the control. There was a significant reduction after incubation for 2 hours (Kruskal-Wallis P < 0.001) for both PBC (0.76 ± 0.25, P < 0.001) and HCV (0.47 ± 0.03, P < 0.001) but no differences after 6 hours (PBC: 1.08 ± 0.37; HCV: 1.53 ± 0.68) and 24 hours (PBC: 0.55 ± 0.21; HCV: 0.66 ± 0.06).

Figure 2 Time points expression in EA.hy926 cells after incubation with serum from primary biliary cholangitis and chronic hepatitis C virus patients compared to controls considered as 1.0. Each bar represents the mean ± SD of 5 experiments. A: Endothelin 1 expression; B: Endothelin 2 expression; C: Endothelin 3 expression; D: Endothelin receptor A expression; E: Endothelin receptor B expression. ^aP < 0.05, ^bP < 0.01, ^cP < 0.001. NS: Not significant; ET: Endothelin; PBC: Primary biliary cholangitis; HCV: Hepatitis C virus; ENDRA: Endothelin receptor A; ENDRB: Endothelin receptor B.

Endothelin 2

Figure 2B demonstrates the results for endothelin 2. There was a significant reduction for both PBC (0.54 ± 0.18, P = 0.004) and HCV (0.51 ± 0.10, P = 0.002) after incubation for 2 hours (Kruskal-Wallis P = 0.02) and 24 hours compared to controls (PBC: 0.55 ± 0.19, P = 0.001; HCV: 0.53 ± 0.13, P = 0.001). At 6 hours, there was a non-significant reduction for PBC (0.90 ± 0.29) and a slight increase for HCV (1.27 ± 0.12, P = 0.04) compared to controls.

Endothelin 3

Figure 2C shows that there were no statistical differences after 2 hours of incubation (PBC: 1.85 ± 1.95; HCV: 0.97 ± 0.78). After 6 hours, there was a small but significant increase in PBC (1.50 ± 0.52, P = 0.04) and a higher increase in HCV (3.08 ± 0.38, P = 0.0001) compared to controls (Kruskal-Wallis P = 0.0002). At 24 hours of incubation, there was a significant decrease for both PBC (0.54 ± 0.19, P = 0.003), and HCV (0.58 ± 0.12, P = 0.002) compared to controls (Kruskal-Wallis P = 0.002).

Endothelin receptor A

Figure 2D presents the results for endothelin receptor A. A significant decrease at 2 hours for both PBC (0.67 ± 0.19, P = 0.014) and HCV (0.47 ± 0.15, P = 0.0003) compared to controls (Kruskal-Wallis P = 0.0009) was observed. Results were different for PBC and HCV at 6 hours and 24 hours. Thus, at 6 hours there was a significant reduction for PBC (0.43 ± 0.18, P = 0.0008) but no change for HCV (0.95 ± 0.47). On the contrary, there was no change for PBC (1.05 ± 0.35) and a significant reduction for HCV (0.51 ± 0.33, P = 0.042) at 24 hours.

Endothelin receptor B

Figure 2E presents the results for endothelin receptor B. At 2 hours a significant reduction for both PBC (0.76 ± 0.25, P = 0.0037) and HCV (0.43 ± 0.03, P = 0.0004) compared to controls (Kruskal-Wallis P = 0.0004) was observed. Similar reductions were observed at 24 hours for PBC (0.56 ± 0.21, P = 0.006) and HCV (0.66 ± 0.06, P = 0.002). At 6 hours no changes were observed (PBC: 1.08 ± 0.37; HCV: 1.53 ± 0.68).

DISCUSSION

In the present study, we used soluble markers produced by LSECs to identify abnormalities of these cells in PBC. We measured sICAM-1, sVCAM-1, soluble E-selectin and the constituents of the coagulation pathway t-PA and its inhibitor PAI-1. Furthermore, we incubated the immortalized human endothelial cell line EA.hy926 with serum from PBC patients to detect possible effects on expression of endothelins and their receptors. In all measurements sera from patients with chronic moderate-severe HCV were used as disease controls.

LSECs are implicated in the pathogenesis of liver diseases collaborating with other sinusoidal cells[21]. They are important for the regulation of immune responses and liver fibrosis. Adhesion molecules such as ICAM, VCAM and E selectin expressed by LSECs, are involved in lymphocyte migration and infiltration of liver parenchyma[22,23]. Based on our previous findings of endothelin abnormalities in PBC, we have proposed a pathogenetic hypothesis where LSECs have a vital role[24].

Our findings showed that sICAM-1 and sVCAM-1 were significantly elevated in patients with PBC but also in patients with HCV compared to controls. Levels were significantly increased in late stages of PBC compared with the early stages. Elevations of soluble adhesion molecules are consistent with endothelial activation and represent endothelial dysregulation as suggested in a recent meta-analysis of rheumatoid arthritis[25]. An upregulation of adhesion molecules at the surface of LSECs, including ICAM-1 and VCAM-1 were also considered as compatible with dysfunctional LSECs[22,26].

Earlier immunohistochemical reports in PBC found that ICAM-1 and VCAM-1 are upregulated on sinusoidal endothelium[27,28]. VCAM-1 was rarely identified on bile ducts of patients[27] indicating that LECs are probably the source of sVCAM. There is also evidence that sICAM does not originate from activated T lymphocytes[29].

In agreement with our study, increased sICAM-1 levels were reported in patients with PBC compared to normal controls and were significantly associated with the histological stage[29-31]. Earlier reports have indicated that treatment with UDCA reduced by 20% the levels of sICAM. In our study, there was no significant reduction after treatment with UDCA. This may be due to the fact that we measured adhesion molecule levels after 6 months of treatment, while earlier studies treated the patients for over a year[29,30]. Interestingly, higher levels of sICAM-1 were reported in female patients, indicating that female patients may have more severe LSECs dysregulation compared to males. Indeed, most of the patients in our study were females, but this was also the case for normal and disease controls[32].

It is stated that expression of VCAM-1, promotes lymphocyte recruitment in PBC[33]. However, VCAM-1 on LSECs is implicated not only in lymphocyte recruitment, but also in sinusoidal vascularization and progress of fibrosis through the Hippo/Yap1 pathway[23], or through direct activation of hepatic stellate cells[23]. LSEC-specific VCAM deletion considerably reduced experimental liver fibrosis in mice[23]. Therefore, increased VCAM indicates increased vascularization and development of liver fibrosis which is a possible explanation for the increased levels in late PBC. Interestingly, in a LSEC line, deficiency in autophagy was associated with increased VCAM-1 expression[34,35]. Dysregulated autophagy has been reported in PBC[36,37]. It may well be therefore that increased VCAM may be due to deficient autophagy of LSECs in PBC.

The increased levels of sVCAM and sICAM in patients with HCV indicate that this finding is not specific for PBC. Our results are in agreement with previous reports for different liver diseases. sICAM-1 and sVCAM-1 levels were increased in patients with all stages of alcoholic liver disease[38]. Increased expression of endothelial activation markers including VCAM and ICAM was observed in alcoholic hepatitis[32], and was associated with a worse prognosis[39]. In obese metabolic dysfunction-associated steatohepatitis patients, VCAM-1 levels were also associated with the degree of liver fibrosis[40,41].

In a study of HBV and HCV associated chronic hepatitis and cirrhosis patients, sICAM and sVCAM were increased compared with the controls. Levels were higher in cirrhosis, in analogy with our findings in late PBC[42]. In chronic HCV hepatitis patients, these adhesion molecules are localized on sinusoidal cells[43]. Additional studies of HCV patients report increased serum levels of VCAM and ICAM in agreement with our study[44-46]. However, in a recent study of chronic HCV, increased serum VCAM-1 was found only in cirrhosis[47]. In HCV patients, sICAM-1 levels were decreased after 6 months of interferon alpha and ribavirin therapy in contrast to sVCAM-1 levels[48,49]. Similar results have been reported in human immunodeficiency virus/HCV co-infected patients with elevated levels of sVCAM-1, sICAM-1[50].

Tumor necrosis factor-α and possibly other pro-inflammatory cytokines upregulate ICAM-1 and VCAM on LSECs[43,47,51]. These reports provide an interesting explanation for the similarity of findings between PBC and HCV in our study. It may be that, the pro-inflammatory environment found mostly in HCV is responsible for the increased LSEC markers, in contrast to PBC where the increase may be due to dysregulated autophagy. This assumption requires further investigation.

Plasma soluble E-selectin is a frequently used biomarker of systemic endothelial dysfunction in liver diseases[52]. High serum levels of E-selectin have been reported in chronic hepatitis and cirrhosis, where immunohistochemical localization identified intense membrane staining on LSECs in severe chronic hepatitis[50,53]. E-selectin is inactive in resting endothelial cells, but it increases in response to pro-inflammatory cytokines such as interleukin-1 and tumor necrosis factor-α[54-56]. In contrast to VCAM and ICAM, levels of E selectin were not significantly elevated in either PBC or HCV patients in our study. However, there was a trend towards an increase in both conditions. An explanation for the discrepancy may be the reported in vitro findings, where an endothelial human cell line demonstrated reduced E-selectin expression under the influence of hepatocyte growth factor. This may be operative in vivo studies of patients[57]. Unlike our findings, increased levels of soluble E-selectin were reported in an earlier study of PBC. In the same study however, higher levels correlated with advanced PBC stages in agreement with our study where a significant increase of soluble E-selectin was found in late PBC stages[58].

t-PA which converts plasminogen into plasmin is produced from vascular endothelial cells[59]. t-PA levels in the circulation are regulated by two distinct mechanisms. T-PA is inhibited directly by PAI-1, or it is cleared by the liver. In both mechanisms PAI-1, has a vital role[60]. PAI-1 is also produced by endothelial cells[61,62]. PAI-1 and t-PA are implicated in fibrosis. Plasmin is a fibrolytic agent. Therefore, t-PA is fibrolytic, while PAI-1 is pro-fibrotic[63]. PAI-1 deficiency ameliorated fibrosis progression in a mouse model of fibrosis[64,65], and a PAI-1-specific inhibitor decreased transforming growth factor-β-mediated myofibroblast differentiation[66]. Reduction of PAI-1 by siRNA and PAI-1 inhibition decreased myofibroblasts characteristics of cancer-associated fibroblasts[67]. On the other hand, t-PA deficiency increased apoptosis of interstitial myofibroblasts in a mouse model[68]. This may be not so in every situation. It has been suggested that PAI-1 deficiency increases myofibroblasts[69], while, PAI-1-deficient mouse cardiac endothelial cells were more sensitive to transforming growth factor-β promoting thus fibrosis[70].

In our study, t-PA was significantly increased in PBC patients, but not in those with chronic HCV. Late PBC had significantly elevated t-PA levels compared to early stages. On the contrary, PAI-levels were not different between either PBC or HCV compared to controls. An earlier study has reported that the plasma concentrations of t-PA, and PAI-1 were significantly increased in patients with PBC. Our findings are in agreement for t-PA, but not for PAI-1[71]. Our findings indicate that the balance between progress of fibrosis and fibrosis resolution favors fibrinolysis in PBC. Endothelins and their receptors are implicated in the progress of liver diseases. A detailed description has been recently reviewed[72].

To further assess the effect of patient serum on the expression of endothelins and their receptors by sinusoidal endothelial cells, we used incubation experiments with the endothelial cell line EA.hy926. Several genes that are expressed by the other well established human umbilical vein endothelial cell line such as genes of endothelial cell markers, are also expressed by EA.hy926 cells making them suitable for in vitro experiments. On the other hand, the two cell lines differ in their production of cytokines after activation with bacterial lipopolysaccharide[73].

Our study showed that incubation with PBC and HCV serum had significant effects in different endothelins and their receptors at different time points. Endothelin 1, endothelin 2 and endothelin 3 had an expression that was similar in both groups. However, endothelin 1 was only reduced at the early time of 2 hours while endothelin 2 was also reduced at 24 hours after a temporary recovery at 6 hours. Endothelin 3 on the other hand, was increased at 6 hours and reduced at 24 hours. The reason for these fluctuations is not clear. This is more evident in the results of endothelin receptor A. Its expression was reduced at 2 hours and 6 hours and recovered at 24 hours in PBC. On the contrary in HCV, after a more intense reduction at 2 hours and a temporary recovery at 6 hours, there was a significant reduction at 24 hours. These findings may indicate that the unknown serum constituents that affect endothelial cells may be different in PBC and HCV. It is tempting to assume that the pro-inflammatory environment and the pro-inflammatory cytokines involved in HCV and the increased serum bile acids that predominate in cholestatic diseases may be responsible for the different results. This assumption requires further investigation.

One drawback of our study is the lack of identification of the cellular origin of LSEC markers. Several endothelial cells produce these markers such as endothelial cells of the cardiovascular system that might influence the results. However, the majority of our patients were women of relatively younger age without history of cardiovascular problems. Therefore, it is plausible that the origin of these markers is indeed the liver and most possibly the sinusoidal endothelin cells. A similar drawback exists for the endothelin findings as we do not know the contents of the patient serum used for incubation experiments. Endothelins in serum may originate from different endothelial cells. Thus, apart from LSECs, it has been shown that biliary epithelial cells and peribiliary mast cells may participate in serum levels that possibly influence our findings[74].

CONCLUSION

In the present study, significant abnormalities were observed in the functional status of LSECs of PBC patients as detected by abnormalities in a series of serum endothelial markers. Although these findings are not disease specific, as they were observed in the disease control group of patients with chronic HCV, they support the hypothesis that LSECs are participating in the pathogenesis of PBC. This is further supported from the findings of the in vitro incubation experiments. An endothelial cell line was incubated with serum from patients with PBC and HCV. Significant abnormalities of the expression of endothelins and their receptors were identified. Importantly, there were differences in expression between PBC and HCV at different time points possibly indicating different serum circulating factors that are responsible for the results.

ACKNOWLEDGEMENTS

We acknowledge the assistance of Tsakou A, MD, MSc in biostatistics for helping in the statistics of the present study.