Vascular cell adhesion molecule 1 expression by biliary epithelium promotes persistence of inflammation by inhibiting effector T-cell apoptosis


  • Simon C. Afford,

    1. National Institute for Health Research Biomedical Research Unit and Center for Liver Research, Medical Research Council Center for Immune Regulation, University of Birmingham, Birmingham, UK
    Search for more papers by this author
  • Elizabeth H. Humphreys,

    1. National Institute for Health Research Biomedical Research Unit and Center for Liver Research, Medical Research Council Center for Immune Regulation, University of Birmingham, Birmingham, UK
    Search for more papers by this author
  • Danielle T. Reid,

    1. Immunology Research Group, Snyder Institute for Chronic Diseases, University of Calgary, Calgary, Alberta, Canada
    Search for more papers by this author
  • Clare L. Russell,

    1. National Institute for Health Research Biomedical Research Unit and Center for Liver Research, Medical Research Council Center for Immune Regulation, University of Birmingham, Birmingham, UK
    Search for more papers by this author
  • Vanessa M. Banz,

    1. Department of Visceral Surgery and Medicine Inselspital, University Hospital and University of Berne, Bern, Switzerland
    Search for more papers by this author
  • Ye Oo,

    1. National Institute for Health Research Biomedical Research Unit and Center for Liver Research, Medical Research Council Center for Immune Regulation, University of Birmingham, Birmingham, UK
    Search for more papers by this author
  • Tina Vo,

    1. Immunology Research Group, Snyder Institute for Chronic Diseases, University of Calgary, Calgary, Alberta, Canada
    Search for more papers by this author
  • Craig Jenne,

    1. Immunology Research Group, Snyder Institute for Chronic Diseases, University of Calgary, Calgary, Alberta, Canada
    Search for more papers by this author
  • David H. Adams,

    1. National Institute for Health Research Biomedical Research Unit and Center for Liver Research, Medical Research Council Center for Immune Regulation, University of Birmingham, Birmingham, UK
    Search for more papers by this author
  • Bertus Eksteen

    Corresponding author
    1. National Institute for Health Research Biomedical Research Unit and Center for Liver Research, Medical Research Council Center for Immune Regulation, University of Birmingham, Birmingham, UK
    2. Immunology Research Group, Snyder Institute for Chronic Diseases, University of Calgary, Calgary, Alberta, Canada
    • Address reprint requests to: Bertus Eksteen, F.R.C.P., Ph.D., Snyder Institute for Chronic Diseases, Health Research and Innovation Center (HRIC), 4AC66-3280 Hospital Drive NW, Calgary, Alberta, Canada T2N 4N1. E-mail:; fax: +1(403) 210-9146.

    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

  • This work was supported by unrestricted grants from The Medical Research Council, British Liver Trust, American Liver Foundation, and the National Institute for Health Research.


Chronic hepatitis occurs when effector lymphocytes are recruited to the liver from blood and retained in tissue to interact with target cells, such as hepatocytes or bile ducts (BDs). Vascular cell adhesion molecule 1 (VCAM-1; CD106), a member of the immunoglobulin superfamily, supports leukocyte adhesion by binding α4β1 integrins and is critical for the recruitment of monocytes and lymphocytes during inflammation. We detected VCAM-1 on cholangiocytes in chronic liver disease (CLD) and hypothesized that biliary expression of VCAM-1 contributes to the persistence of liver inflammation. Hence, in this study, we examined whether cholangiocyte expression of VCAM-1 promotes the survival of intrahepatic α4β1 expressing effector T cells. We examined interactions between primary human cholangiocytes and isolated intrahepatic T cells ex vivo and in vivo using the Ova-bil antigen-driven murine model of biliary inflammation. VCAM-1 was detected on BDs in CLDs (primary biliary cirrhosis, primary sclerosing cholangitis, alcoholic liver disease, and chronic hepatitis C), and human cholangiocytes expressed VCAM-1 in response to tumor necrosis factor alpha alone or in combination with CD40L or interleukin-17. Liver-derived T cells adhered to cholangiocytes in vitro by α4β1, which resulted in signaling through nuclear factor kappa B p65, protein kinase B1, and p38 mitogen-activated protein kinase phosphorylation. This led to increased mitochondrial B-cell lymphoma 2 accumulation and decreased activation of caspase 3, causing increased cell survival. We confirmed our findings in a murine model of hepatobiliary inflammation where inhibition of VCAM-1 decreased liver inflammation by reducing lymphocyte recruitment and increasing CD8 and T helper 17 CD4 T-cell survival. Conclusions: VCAM-1 expression by cholangiocytes contributes to persistent inflammation by conferring a survival signal to α4β1 expressing proinflammatory T lymphocytes in CLD. (Hepatology 2014;59:1932–1943)




protein kinase B1


alcoholic liver disease




B-cell lymphoma 2


bile ducts


biliary epithelial cells


chemokine (C-C motif) ligand


C-C chemokine receptor


chronic liver disease


chemokine (C-X-C motif) ligand


chemokine (C-X-C motif) receptor


enzyme-linked immunosorbent assay


fluorescence-activated cell sorting


Fas ligand


flow cytometry


human epithelial antigen 125




immunoglobulin G






liver-infiltrating lymphocytes


monoclonal Ab


magnetic-assisted cell sorting


mitogen-activated protein kinase


nuclear factor kappa B


primary biliary cirrhosis


polymerase chain reaction


primary sclerosing cholangitis


portal tracts


recombinant human


RAR-related orphan receptor gamma


T helper


tumor necrosis factor alpha


vascular cell adhesion molecule 1



Portal infiltrates are characteristic of not only the biliary diseases primary sclerosing cholangitis (PSC) and primary biliary cirrhosis (PBC), but also autoimmune hepatitis and viral hepatitis.[1] In acute hepatitis, infiltrates resolve by apoptosis and clearance, but in chronic diseases, they persist and drive ongoing fibrogenesis.[2] These observations suggest that the portal microenvironment in chronic hepatitis not only regulates recruitment of T cells, but also promotes their survival.[3] Biliary epithelial cells (BECs) are the focus for immune-mediated liver damage in diseases, such as PBC and PSC.[4]

BECs attract and retain effector leukocytes by expressing adhesion molecules and chemokines, including chemokine (C-X-C motif) ligand (CXCL)9-11 and chemokine (C-C motif) ligand (CCL)2-5, which attract T cells expressing the chemokine receptors, chemokine (C-X-C motif) receptor (CXCR)3 and C-C chemokine receptor (CCR)5, respectively.[5] Selective recruitment of T-cell subsets to portal tracts (PTs) is also mediated by BECs; expression of the transmembrane chemokine, CXCL16, preferentially recruits CD8 T cells expressing CXCR6, whereas CCL20 selectively recruits CCR6+ T helper (Th)17 CD4 cells to position and bind to BECs.[6, 7]

Intercellular adhesion molecule 1 and VCAM-1 are expressed by inflamed cholangiocytes and can potentially support adhesion of T cells expressing the integrins, αLβ2 and α4β1, respectively.[6, 8] In other situations, α4β1 integrin binding to VCAM-1 has been reported to activate “outside-in” signaling and survival pathways that delay or resist apoptosis.[9] This prompted us to investigate whether cholangiocytes through VCAM-1 can induce a survival signal in lymphocytes to promote their persistence in PTs in the setting of liver inflammation.

We report that VCAM-1 is expressed on cholangiocytes in response to inflammatory stimuli and can support T-cell adhesion to BECs. T-cell binding to VCAM-1 reduced formation of activated caspase 3 and enhanced survival through nuclear factor kappa B (NF-κB), protein kinase B1 (Akt1), and p38 mitogen-activated protein kinase (MAPK) activation and B-cell lymphoma 2 (Bcl-2) accumulation. In vivo inhibition of VCAM-1, in a model of biliary inflammation, reduced portal infiltrates and hepatic inflammation. As expected, VCAM-1 inhibition reduced T-cell recruitment, but also increased lymphocyte apoptosis, in periportal areas. Our data suggest a novel mechanism whereby cholangiocyte VCAM-1 not only recruits CD8 and Th17 CD4 T cells in the liver, but also allows them to resist apoptosis and persist. These findings suggest that antiadhesion therapies ameliorate tissue inflammation through multiple modes of action, including apoptosis.

Materials and Methods

Male wild-type C57BL/6, OT1, OT2 transgenic TCR mice, RORγT-GFP (6-8 weeks) were purchased from The Jackson Laboratory (Bar Harbor, ME). Ova-Bil mice[10] were a gift from Dr. Marion Peters (University of California San Francisco, San Francisco, CA). All procedures were carried out in accord with the Animals (Scientific Procedures) Act of 1986 and University of Calgary (Calgary, Alberta, Canada), Alberta Assessment Consortium approval (M11025). Liver tissue was obtained from patients undergoing liver transplantation and used to isolate intrahepatic T cells and primary cholangiocytes. All samples were collected with informed patient consent and local research ethics committee approval (local ethical approval 04/Q2708/41 and REC 2003/242).

Prevention of Apoptosis in Liver-Infiltrating Lymphocytes After Stimulation With Recombinant Human VCAM-1 or BEC Coculture

Liver-infiltrating CD3, CD8, and CD4 T cells were used to determine whether coculture with BEC or recombinant human (rh)VCAM-1 had an effect on lymphocyte survival. Isolation and culture of liver-infiltrating lymphocytes (LILs) and BECs were performed as previously described[7, 11] and are summarized in the Methods section of the Supporting Information. LIL subsets were incubated in media alone, with rhVCAM-1 Fc (10 µg/mL; R&D Systems, Minneapolis, MN) attached to human immunoglobulin G (IgG) microbeads (Miltenyi Biotech, Cambridge, MA), rhVCAM-1 Fc plus benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone (z-vad-fmk; 20 µM; Promega, Madison, WI), or rhVCAM-1 Fc plus α4β1 blocking antibodies (Abs; 10 µg/mL) each for 24 hours. Jurkat cells were used as a positive control of cell death and were cultured in (1) media alone, (2) Fas ligand (FasL; 50 ng/mL; R&D Systems), and (3) FasL plus z-vad-fmk for 24 hours. Similarly, LIL subsets were cocultured with cytokine-activated BECs ± α4β1 blocking Abs. Further details are contained in the Methods section of the Supporting Information. T cells were then harvested into ice-cold phosphate-buffered saline and western blotting or fluorescence-activated cell sorting (FACS) performed as described in the Methods section of the Supporting Information.

In Vivo Detection of Apoptosis

To detect in vivo apoptosis, we used the Ova-Bil model, as previously described,[10] and summarized in the Methods section of the Supporting Information. Ova-bil and Ova-Bil × RORγTGFP mice received 1 × 107 OT1 and 1 × 107 OT2 splenocytes and on day 7 after injection with 50 or 100 ug of anti-VCAM-1 monoclonal Ab (mAb) or IgG2a control. After 24 hours, mice received a single injection of apolipoprotein (Apo)-Trace (Sigma-Aldrich, St. Louis, MO), according to the manufacturer's instructions, to label apoptotic cells. Animal were subsequently anesthetized by intraperitoneal injection of 200 mg/kg of ketamine (Bayer Inc. Animal Health, Toronto, Ontario, Canada) and 10 mg/kg of xylazine (Bimeda-MTC, Cambridge, Ontario, Canada). Apoptotic cells labeled with Apo-Trace (488-nm filter set) were detected in the liver by inverted spinning-disk confocal microscopy, performed as described previously.[12] T-cell subsets were isolated, sorted, and analyzed by FACS, based on CD8 or RORγTGFP expression, after in vivo Apo-Trace injection. Apoptosis was quantified by manual counting of apoptotic cells per high power field (100×), expression of Apo-trace, and cleavage of active caspase 3 by isolated T-cell subsets. The mean of three representative fields were recorded. In addition, liver tissue was harvested and stained for ubiquitin expression. Further details are contained in the Methods section of the Supporting Information.

Figure 1.

VCAM-1 expression by human BDs in vivo and proximity of T cells to BDs. (A) Tissue sections from diseased explanted livers were used to confirm VCAM-1 expression on BDs by IHC (PSC liver section shown; brown-red chromogen pigment; 200× magnification). (B) Expression of VCAM-1 by biliary epithelium was recorded for individual diseases by manual counts of positive staining BDs per three high power views and is expressed as a percentage of total BDs counted (n = 6 per individual disease). (C) Expression of VCAM-1 by biliary epithelium was confirmed by coimmunofluorescence using anti-HEA-125 mAb as a specific cholangiocyte marker (labeled with Texas Red) and anti-VCAM-1 mAb (labeled with FITC green). Merged images confirmed coexpression (yellow) of HEA-125 and VCAM-1 by cholangiocytes. Significant amounts of VCAM-1 were also detected in nearby areas of portal fibrosis (white arrow). (D) IHC with anti-CD3, CD4, CD8, or isotype control Abs were used to demonstrate the location of T cells in relation to BD (brown-red pigment; 400× magnification). FITC, fluorescein isothiocyanate.

Statistical Analysis

Statistical analysis was performed as described in individual experiments using GraphPad Prism software.


VCAM-1 Expression on Inflamed Bile Ducts in Human Liver Diseases

Normal liver tissue did not express detectable amounts of VCAM-1; however, the majority of bile ducts (BDs) in diseased livers in alcoholic cirrhosis, hepatitis C virus, and, particularly, in the biliary diseases, PSC and PBC, expressed VCAM-1 (Fig. 1B). Expression of VCAM-1 by BDs was confirmed by coimmunofluorescence using Abs to human epithelial antigen 125 (HEA-125; a biliary epithelial marker) and VCAM-1. Although strong VCAM-1 expression was observed on BDs, there was also significant VCAM-1 detectable in surrounding connective tissue in the PTs (Fig. 1C). Immunohistochemistry (IHC) revealed that CD3, CD8 and CD4 positive lymphocytes were closely positioned to BD epithelium in the diseased liver (Fig. 1D, example from a PSC liver shown).

Control of VCAM-1 Expression by Cholangiocytes in Culture

To determine factors that control VCAM-1 expression in cholangiocytes, we cultured primary human biliary epithelium (BECs) from PSC, PBC, and alcoholic liver disease (ALD) with tumor necrosis factor alpha (TNF-α), interferon-gamma (IFN-γ), interleukin (IL)−17, and CD40L, either alone or in combination, and measured VCAM-1 by immunocytochemistry, flow cytometry (FCM), and reverse-transcriptase polymerase chain reaction (PCR). TNF-α alone or in combination with CD40L induced the strongest expression of VCAM-1 (Fig. 2A). VCAM-1 was expressed by 55% of unstimulated cholangiocytes on the cell surface, which increased to 80% after 12 hours of stimulation with TNF-α. PCR confirmed TNF-α, alone or in combination with CD40L or IFN-γ, as a potent inducer of VCAM-1 in cultured human BECs (Supporting Fig. 2). Because VCAM-1 can also be shed as a soluble molecule after cleavage on the cell surface by metalloproteinase ADAM17, we measured soluble VCAM-1 in cholangiocyte supernatants by enzyme-linked immunosorbent assay (ELISA) and demonstrated increased levels in response to TNF-α alone or in combination with IFN-γ, CD40L, and IL-17 (Fig. 2B). There were no detectable differences in VCAM-1 expression between BECs from PSC, PBC, or ALD; hence, BECs from these sources were interchangeably used in subsequent assays.

Figure 2.

Induction of VCAM-1 expression in cultured cholangiocytes. Primary human cholangiocytes were grown in media alone or with cytokines (IL-17, TNF-α, and IFN-γ) and CD40L, alone or in combination, for 6 hours to assess messenger RNA expression or 12 hours for protein expression. (A) Cholangiocytes in culture were washed and fixed after 12 hours and cyto-IHC performed using anti-VCAM-1 mAb or isotype controls. Stimulation with TNF-α or TNF-α with CD40L is shown (brown-red pigment; 400× magnification). (B) Supernatants were collected from cultured cholangiocytes after 12 hours, and soluble VCAM-1 expression was measured by ELISA. Experiments were performed in triplicate wells per donor (n = 6 different cholangiocyte donors per condition). *P < 0.05; **P < 0.01, as compared to untreated controls.

T Cells Adhere to Primary Cholangiocytes in Culture Using α4β1 Integrins

We confirmed expression of α4 and β1 integrins by peripheral blood and liver-derived lymphocytes by FCM (Supporting Fig. 3) and showed that lymphocytes demonstrated increased binding to TNF-α-activated cholangiocytes. When VCAM-1 was blocked by a specific Ab, lymphocyte binding was significantly reduced, but not by an isotype-matched control. Lymphocyte binding was reduced by the addition of pertussis toxin to lymphocytes to prevent G-protein-coupled chemokine receptor-mediated binding (negative control). Lymphocyte binding was significantly increased by incubating lymphocytes with manganese chloride, which induces integrin conformational activation directly (positive control; Fig. 3).

Figure 3.

Adhesion of T cells to cholangiocytes. Cholangiocytes were grown in media alone or stimulated with TNF-α for 12 hours. Fresh CD3, CD8, and CD4 liver-infiltrating T cells were isolated by MACS. Thirty minutes before the addition of lymphocytes to cholangiocyte monolayers, blocking Abs to VCAM-1, α4, or β1 alone or in combination were added to cholangiocytes or lymphocytes. Lymphocytes were also stimulated with the positive control, MnCl2, or the negative control, Pertussis toxin. Lymphocytes were added to cholangiocytes and allowed to adhere for 1 hour. Fixed cells were identified using CD3, CD4, or CD8 Abs. (A) Panel demonstrates representative photomicrographs of lymphocyte adhesion to cholangiocytes (400× magnification) and (B) the number of adherent lymphocytes per field of view. Gray bars are CD3 T cells, black bars = CD8 cells, and white bars = CD4 (n = 3, with mean ± standard error of the mean). *P < 0.05; **P < 0.01; ***P < 0.001 or ns (not significant), as compared to untreated controls.

VCAM-1/α4β1-Dependent Interactions Inhibit Intrahepatic T-Cell Apoptosis

Coculture of freshly isolated human liver CD3 T cells with TNF-α-activated primary BECs inhibited activation of caspase 3 and suggested that VCAM-1/α4β1 interactions between BECs and T cells led to improved T-cell survival (Fig. 4). To determine whether this observation is specifically dependent on VCAM-1 and because of the technical challenges to prevent contamination of T-cell apoptosis data by biliary apoptosis, we elected to subsequently use rhVCAM-1 Fc bound to microbeads, instead of BECs, in cultures with freshly isolated human intrahepatic T-cell populations (CD3, CD8, and CD4) after magnetic-assisted cell sorting (MACS; Fig. 4). Lymphocyte apoptosis measured by caspase 3 cleavage was markedly reduced when cells were stimulated with rhVCAM-1 Fc for 24 hours, and this effect was blocked by the addition of α4β1 blocking Abs. Both CD4 and CD8 T-cell populations derived a survival benefit from interactions with VCAM-1 (Fig. 4). Neither BEC coculture nor rhVCAM-1 Fc induced T-cell proliferation (not shown).

Figure 4.

Survival and apoptosis LILs after incubation with VCAM-1. (A and B) Freshly isolated human liver CD3 T cells were cocultured for 24 hours with TNF-α-activated primary BECs. T cells were harvested, and active caspase 3 was measured by FCM. (A) Representative FACS plots are shown, and (B) data are summarized as mean ± standard error of the mean. **P < 0.01; n = 7. (C) CD3, CD4, or CD8 liver-derived lymphocytes were cultured in media alone (1), VCAM-1 Fc chimera bound to IgG microbeads (2), VCAM-1 Fc plus z-vad-fmk (3), or VCAM-1 Fc plus α4β1 blocking Abs (4) for 24 hours. Jurkat cells were used as a control for caspase 3 activation in the following conditions: unstimulated (5); FasL (positive control) (6); or FasL plus z-vad-fmk (negative control) (7). (C) Representative western blottings for CD3, CD4, CD8, and Jurkat cleaved caspase 3 Asp 175 and their corresponding β-actin blots. (D) Summarizes optical density data for cleaved caspase 3 Asp 175 activity after normalization against β-actin (n = 3, with mean ± standard error of the mean). *P < 0.05; **P < 0.01 or ns (not significant), as compared to untreated controls.

VCAM-1-Induced T-Cell Survival Is Mediated by NF-κB, MAPK, and Akt1 Pathways

We assessed the ability of VCAM-1 to activate known pathways associated with cell survival (Fig. 5).[9] Lymphocytes were MACS sorted into CD3, CD8, and CD4 populations and cultured with rhVCAM-1 Fc beads and activation of NF-κB, Akt1, and p38 MAPK pathways measured by western blotting.

Figure 5.

Mechanism of increased survival of T cells after incubation with VCAM-1. To determine how T cells gain a survival advantage by binding to VCAM-1, we assessed activation of known survival pathways: NF-κB; Akt1; or MAPK. CD3, CD4, or CD8 cells were assessed in culture media alone, with soluble recombinant VCAM-1 Fc chimera bound to IgG microbeads or VCAM-1 Fc plus NF-κB, Akt1, or MAPK inhibitors or positive controls CD3/CD28 beads (NF-κB and Akt1) or anisomycin (MAPK) for 1 hour and harvested for western blotting. Jurkat cells were used as a positive control for apoptosis in the following conditions: unstimulated; FasL; or FasL plus the pan-caspase inhibitor, z-vad-fmk. Graphs show optical density data for (A) phospho-NF-κB serine 536, (B) phospho-Akt1 serine 473, or (C) phospho-p38 MAPK Thr 180/Tyr 182 activation at 1 hour after normalization with β-actin. (D) Active caspase 3 measurements were used as experimental controls. *P < 0.05; **P < 0.01; ***P < 0.001, as compared to untreated controls.

sVCAM-1 Fc significantly increased NF-κB phosphorylation (phospho-serine 536), Akt1 phosphorylation (phospho-Akt1 serine 473), and p38 MAPK phosphorylation (phospho-p38 MAPK Thr 180/Tyr 182). CD3, CD4, and CD8 cells showed the same patterns of signaling through NF-κB, Akt1, and p38 MAPK. Phosphorylation changes to these pathways in response to rhVCAM-1 Fc correlated with decreased active caspase 3 activity.

VCAM-1/T-Cell Interactions In Vivo

To test the effects of VCAM-1 in vivo and on specific T-cell subsets, we used the Ova-bil model of biliary inflammation.[10] This is an antigen-driven model of biliary injury that requires both CD4 and CD8 responses and can be used to model immune-driven biliary injury as part of several diseases, including PSC.[13] First, we confirmed that VCAM-1 expression is up-regulated on BDs in this model. IHC detected intense VCAM-1 expression on BDs, peribiliary tissue, and surrounding hepatic sinusoids (Fig. 6A).

Figure 6.

In vivo effect of VCAM-1 neutralization in a model of hepatobiliary inflammation. We transferred 1 × 107 OT1 and 1 × 107 OT2 splenocytes to Ova-Bil mice. (A) Expression of VCAM-1 in this model was visualized by IHC using anti-VCAM-1 Abs (representative slide using brown-red chromogen pigment; 200× magnification). Narrow arrow highlights VCAM-1 expression by a BD, whereas large arrows points to VCAM-1 expression in hepatic sinusoids. Control slide shown using negative isotype control Abs. After transfer of OT1/OT2 cells, mice were treated with 50 or 100 ug of anti-VCAM-1 mAb intravenously to block VCAM-1 or control IgG2a on days 0, 3, and 6. After 8 days, hepatic inflammation was measured by (B) hematoxylin and eosin staining of tissue sections (arrows denote inflammatory infiltrates; 200×) and (C) alanine aminotransferase (ALT) levels (n = 6; mean ± standard error of the mean). *P < 0.01, compared to control.

Next, we assessed, using the Ova-Bil model, whether Ab-mediated inhibition of VCAM-1 had potential therapeutic benefits and whether these effects were mediated by altered T-cell recruitment and/or T-cell apoptosis. After 8 days, there was a decrease in liver inflammation (Fig. 6B,C) and cholestasis, as measured by serum bile acids (Supporting Fig. 3).

This reduction was only partially explained by a reduction in T-cell recruitment to the liver (Supporting Fig. 4). To determine whether T-cell apoptosis also contributed to the reduced liver inflammation as a result of anti-VCAM-1 treatment, liver injury was induced in Ova-Bil mice and, on day 7, mice were injected with 50 or 100 ug of anti-VCAM-1 mAb or control. After 24 hours, mice received a single injection of Apo-Trace to label apoptotic cells and their livers were imaged by spinning disk intravital microscopy (Fig. 7A,B). The addition of either 50 or 100 ug of anti-VCAM-1 mAb significantly increased apoptosis in the PTs and in the surrounding hepatic sinusoids, whereas isotype negative control Abs had no effect (Fig. 7C). Peribiliary inflammatory cells stained strongly for ubiquitin, providing further support that targeting VCAM-1 limits liver inflammation by inducing apoptosis of inflammatory lymphocytes.

Figure 7.

In vivo detection of apoptosis after VCAM-1 neutralization. To detect apoptosis in vivo, we adoptively transferred 1 × 107 OT1 and 1 × 107 OT2 splenocytes to Ova-Bil mice and, on day 7, injected with 50 or 100 ug of anti-VCAM-1 mAb or IgG2a control. After 24 hours, mice received a single injection of Apo-Trace to label apoptotic cells and livers were imaged by spinning disk intravital microscopy. Apoptotic cells were detected as bright fluorescein (white) positive cells and highlighted by white arrows in the PTs. (A) Representative black and white still photos are shown. (B) Apoptotic cells were quantified by manual cell counting of three representative fields used per mouse (n = 5; mean ± standard error of the mean). **P < 0.001, compared to control. (C) Twenty-four hours after 100 ug (intravenously) of anti-VCAM-1 mAb, liver tissue was harvested from Ova-Bil mice and further evidence of peribiliary apoptosis was obtained by immunochemistry with anti-ubiquitin mAb. Representative slide demonstrating ubiquitin-positive peribiliary inflammatory cells (dark brown chromogen). Control slide shown using negative isotype control Abs (n = 5).

Our previous work has demonstrated that Th17 CD4 cells are actively recruited to BDs; hence, we evaluated whether biliary VCAM-1 expression also contributed to their survival.[7] Subset ex vivo analysis of sorted intrahepatic T cells revealed that treatment with VCAM-1Fc significantly reduced apoptosis in CD8 and CD4 RORγt+ subsets, as measured by Apo-Trace and cleavage of caspase 3 (Fig. 8A-C). Ex vivo experiments with isolated CD4 RORγt+ (Fig. 8D) and CD8 (not shown as similar) from inflamed Ova-Bil livers demonstrated that coculture with VCAM-1-coated beads or TNF-α-activated BECs led to a reciprocal accumulation of mitochondrial Bcl-2 to promote survival. FACS analysis demonstrated high levels of expression of α4 integrins by CD8 T cells and α4 expression by up to 20% by RORγT+ CD4 cells (Supporting Fig. 5).

Figure 8.

VCAM-1 induces murine CD8 and Th17 CD4 survival. After induction of Ova-Bil injury, hepatic cells were isolated, sorted into CD8 and CD4 RORγt subsets, and cultured ex vivo with VCAM-1 Fc-Beads or IgG beads. Apoptosis was detected by Apo-Trace and FACS. (A) Representative FACS plots. (B) Summary of mean Apo-Trace-positive cells. (C) Western blotting revealed reduced cleavage of active caspase 3 in response to VCAM-1 Fc beads coculture. (D) Ex vivo culture with CD4 RORγt in combination with BEC or VCAM-1 Fc-beads increased mitochondrial fraction Bcl-2 levels, as measured by western blotting (n = 6; mean ± standard error of the mean). *P < 0.01 or **P < 0.001, compared to IgG control or unstimulated BEC.


We have demonstrated that cholangiocyte VCAM-1 can promote lymphocyte survival by binding to α4β1 and suggest this is a novel mechanism by which VCAM-1 promotes persistence of inflammation in chronic hepatitis.

VCAM-1 expression has been shown to increase in chronic liver diseases (CLDs), including ALD[14] and PBC.[15] The majority of studies have focused on VCAM-1 expression by vascular endothelium and its role in immune cell recruitment[16] as well as its expression and function on BDs has been underappreciated. VCAM-1 expression by BDs was first reported in PBC by IHC[15] and confirmed by PCR microarrays.[17] VCAM-1 expression in the liver is induced by multiple pathways, including Toll-like receptor 4 stimulation with lipopolysaccharide,[18] IFN-γ/signal transducer and activator of transcription 1 pathway,[19] farnesoid X receptor signaling,[20] and by bile acids.[21] Conversely, therapies that target these pathways to limit liver inflammation, such as norursodeoxycholic acid,[22] are associated with reduced biliary VCAM-1 expression. Data on factors that specifically regulate VCAM-1 expression by BECs are limited.[5] We report that TNF-α is a potent inducer of VCAM-1 expression in cholangiocytes, particularly when combined with INF-γ or CD40L, suggesting that T cells contribute to VCAM-1 expression and their own survival by producing CD40L, IFN-γ, or IL-17 to enhance biliary VCAM-1 levels. BECs from PSC, PBC, and ALD did not differ in their ability to express VCAM-1 and suggests that VCAM-1 expression by BECs is not disease specific, but a generic response to hepatic inflammation.

Naïve T cells express low levels of α4β1 integrins, whereas memory and effector T cells express higher levels of these molecules.[23] The majority of liver-infiltrating T cells are antigen experienced, express high levels of α4β1, and bind to VCAM-1 on sinusoidal endothelial cells during recruitment to the liver.[5] Our data suggest that similar pathways mediate adhesion to BDs because, in vitro and in vivo, this was reduced by blocking Abs against VCAM-1 or α4β1.

Inhibition α4 integrins with natalizumab is efficacious in treating chronic inflammation in inflammatory bowel disease and multiple sclerosis through presumed inhibition of proinflammatory immune cell recruitment.[24] However, it is possible that disruption of α4-VCAM-1 interactions not only limits recruitment of proinflammatory immune cells, but also causes apoptosis of those already resident in the tissue and involved in immune damage. Evidence in support of a role for VCAM-1 as a survival signal during inflammation comes predominantly from the joint in rheumatoid arthritis, where both lymphocytes and neutrophils gain a survival advantage by engaging VCAM-1.[9, 25]

The fate of lymphocytes is determined by a balance of prosurvival pathways (NF-κB, Akt1, and MAPK) and increases of molecules, such as Bcl-2 and Bcl-extra large, on the one hand and the availability of death signals (FasL) and activation of the caspases on the other.[26] The liver is normally a hostile environment for T cells, with high levels of apoptosis,[27] and thus it is not surprising that we observed high levels of cleaved caspase 3 activity in LILs once isolated. This was significantly reduced during coculture with BECs or with the addition of VCAM-1 Fc to cells or with the pan caspase inhibitor, z-vad-fmk, thus suggesting that VCAM-1 causes a net shift from apoptosis and death toward T-cell survival and is confirmed by accumulation of mitochondrial Bcl-2. These effects of VCAM-1 appear to be limited to the inflamed liver, because minimal VCAM-1 expression was observed in the normal liver and pretreatment with anti-VCAM-1 mAb in mice before induction of liver injury failed to ameliorate liver inflammation.

Effector T cells, in particular, CD8 and RAR-related orphan receptor gamma (RORγt) CD4 T cells, appeared to require binding to VCAM-1 to activate known survival pathways to ensure their persistence in the liver. The mammalian Rel/NF-κB family of transcription factors play a central role in the immune system by regulating several processes, including survival of immune cells.[28] Previous studies have shown that when antigen receptors and costimulatory signals from CD28 on T and B lymphocytes are engaged, these act on NF-κB[26] by Akt, phosphoinositide 3-kinase and Ras-MAPK pathways.[28] In addition to activating the classical NF-κB pathway, Akt may also phosphorylate p65, causing enhanced transcription of NF-κB.[29] In our system, there was an up-regulation of NF-κB p65 activation in response to sVCAM-1 Fc or stimulation with CD3/CD28 beads after 1 hour and a decrease in p65 phosphorylation when the NF-κB activation inhibitor was added, confirming that integrin binding to VCAM-1 can signal through NF-κB to increase cell survival, as was shown using neutrophils and α9β1 integrins.[9]

Akt has been shown to promote T-cell survival against apoptotic stimuli by up-regulating Bcl-2.[30] Similarly, we observed activation of Akt1 and increased mitochondrial accumulation of Bcl-2. This is consistent with a previous report where integrin engagement activated Akt, and this pathway converged on NF-κB.[31] VCAM-1 interactions similarly activated the p38 MAPK pathway,[32] which suggests that multiple pathways converge to promote CD8 and CD4 Th17 survival in the liver. These events occurred in the absence of significant T-cell proliferation.

In summary, we propose an important role for VCAM-1 in mediating survival of T cells in human and murine inflammatory liver disease. Thus, increased expression of VCAM-1 during inflammation not only regulates T-cell recruitment, but also allows inflammation to persist by giving lymphocytes a survival signal and retaining them at sites such as the BDs in diseases such as PBC and PSC.


The authors acknowledge technical and infrastructure support from Dr. Pina Collarusso, Dr. Paul Kubes, and the Live Cell Imaging core at the Snyder Institute, University of Calgary.