T lymphocytes interact with hepatocytes through fenestrations in murine liver sinusoidal endothelial cells


  • Alessandra Warren,

    1. Centre for Education and Research on Ageing (CERA) and the ANZAC Research Institute, Concord RG Hospital and University of Sydney, Sydney, Australia
    Search for more papers by this author
  • David G. Le Couteur,

    1. Centre for Education and Research on Ageing (CERA) and the ANZAC Research Institute, Concord RG Hospital and University of Sydney, Sydney, Australia
    Search for more papers by this author
  • Robin Fraser,

    1. Department of Pathology, Christchurch School of Medicine and Health Sciences, University of Otago, Christchurch, New Zealand
    Search for more papers by this author
  • David G. Bowen,

    1. Center for Vaccines and Immunity, Columbus Children's Research Institute, Columbus, Ohio
    Search for more papers by this author
  • Geoffrey W. McCaughan,

    1. AW Morrow Gastroenterology and Liver Centre, Centenary Institute of Cancer Medicine and Cell Biology, Royal Prince Alfred Hospital and University of Sydney, Sydney, Australia
    Search for more papers by this author
  • Patrick Bertolino

    Corresponding author
    1. AW Morrow Gastroenterology and Liver Centre, Centenary Institute of Cancer Medicine and Cell Biology, Royal Prince Alfred Hospital and University of Sydney, Sydney, Australia
    • Centenary Institute of Cancer Medicine and Cell Biology, Locked Bag No 6, Newtown 2042, NSW Australia
    Search for more papers by this author
    • fax: (612) 95 65 6101

  • See Editorial on Page 1083

  • Potential conflict of interest: Nothing to report.


The liver has an established ability to induce tolerance. Recent evidence indicates that this unique property might be related to its distinctive architecture allowing T cells to be activated in situ independently of lymphoid tissues. Unlike lymph node–activated T cells, liver-activated T cells are short-lived, a mechanism that might contribute to the “liver tolerance effect.” Although the potential role of hepatocytes as tolerogenic antigen-presenting cells has been demonstrated, the question as to whether these cells are able to interact with CD8+ T cells in physiological settings remains controversial. Contradicting the immunological dogma stating that naïve T lymphocytes are prevented from interacting with parenchymal cells within non-lymphoid organs by an impenetrable endothelial barrier, we show here that the unique morphology of the liver sinusoidal endothelial cell (LSEC) permits interactions between lymphocytes and hepatocytes. Using electron microscopy, we demonstrate that liver resident lymphocytes as well as circulating naïve CD8+ T cells make direct contact with hepatocytes through cytoplasmic extensions penetrating the endothelial fenestrations that perforate the LSECs. Furthermore, the expression of molecules required for primary T cell activation, MHC class I and ICAM-1, is polarized on hepatocytes to the perisinusoidal cell membrane, thus maximizing the opportunity for interactions with circulating lymphocytes. In conclusion, this study has identified, at the ultrastructural level, a unique type of interaction between naïve T lymphocytes and liver parenchymal cells in vivo. These results hold implications for the pathogenesis of viral hepatitis in which hepatocytes may represent the main antigen-presenting cell, and for the development of immune tolerance as lymphocytes pass through the liver. (HEPATOLOGY 2006;44:1182–1190.)

According to current immunological tenets, vascular endothelial cells form an efficient physical barrier that prevents access by naïve T lymphocytes to the surrounding tissue.1–3 Naïve lymphocytes require activation in peripheral lymphoid tissues by professional antigen-presenting cells (APCs) before they are able to migrate across the endothelium for subsequent interaction with parenchymal cells.

Recent studies provide evidence that the liver may be an exception to this model.4–9 In contrast to other solid organs, naïve CD8+ T lymphocytes can be activated in the liver independently of lymphoid tissues. Although T lymphocyte activation in the liver and lymph nodes (LN) can occur concomitantly, the site of activation appears to drive T lymphocytes toward differing fates. T lymphocyte activation by professional APCs in LN generates an effective immune response. Conversely, T lymphocyte activation in the liver is an inefficient process, leading to diminished cytotoxic activity and reduced T cell survival.8 This model provides a plausible mechanism for the so-called “liver tolerance effect”10 and suggests that the site of primary activation has a profound influence on intrahepatic immunity.9

Within the liver, there is increasing evidence that hepatocytes may function as APCs, a finding highly relevant to some hepatotropic viral infections such as those by the hepatitis C virus. Hepatocytes express high levels of major histocompatibility complex (MHC) Class I molecules,11 as well as intercellular adhesion molecule-1 (ICAM-1) and CD1d,12–14 a ligand recognized by NKT lymphocytes.15 These cells are therefore fully competent to interact and activate lymphocytes via MHC class I receptors. Experimental studies have demonstrated that these cells are indeed efficient APCs in vitro.16, 17 Support for a role of hepatocytes as APCs in vivo comes from studies performed using transgenic mouse models. When Alb-Kb and Met-Kb transgenic mice, expressing the allo-MHC H-2Kb molecule on hepatocytes,18, 19 were adoptively transferred with TCR transgenic T cells specific for H-2Kb, T cell primary activation occurred in the liver independently of lymphoid tissues, thus supporting the thesis that hepatocytes can interact with circulating T cells in vivo.6, 8 Such results suggest that hepatocytes can present antigen to both CD8+ T lymphocytes and possibly also NKT lymphocytes, and thus play a major role in antigen-specific recruitment of these cells to the liver. However, the question arises as to how circulating naïve T lymphocytes could interact with hepatocytes, given the presence of an endothelial barrier. This question remains controversial, as contact between hepatocytes and circulating lymphocytes do not normally occur.20

The unique morphology of the hepatic sinusoid and liver sinusoidal endothelial cells (LSECs) provides a potential mechanism for this unique interaction between T lymphocytes and the liver. Hepatic sinusoids are porous, gossamer-like, cylindrical structures that are slightly narrower than blood cells. They connect afferent portal triads to exiting central hepatic venules, a distance of approximately 1 mm. Approximately one billion sinusoids form the rich capillary network of the liver, which permits the vast hepatic blood flow to course slowly and intimately between the cords of hepatocytes. Because the sinusoidal lumen is narrower than lymphocytes, lymphocytes have been suggested to “massage” the LSECs during their transit, thus facilitating lymph circulation in the space of Disse.21 LSECs occupy a strategic position in the hepatic sinusoid. They are highly specialized endothelial cells that line the wall of the hepatic sinusoid and separate the sinusoidal blood, derived primarily from the portal vein, from hepatocytes. LSECs are perforated by fenestrations, which are pores approximately 100 nm in diameter grouped together in clusters known as liver sieve plates and occupying approximately 5% to 10% of the endothelial surface.22 Fenestrations are true discontinuities in the endothelium, lacking either a diaphragm or underlying basal lamina.

T lymphocytes could interact with hepatocytes by several plausible mechanisms (Fig. 1). First, fenestrations could provide a portal for interaction between hepatocytes and lymphocytes. As a first possibility, extensions of lymphocytes might pass through fenestrations and contact the hepatocellular membrane in the extracellular space of Disse (Fig. 1A). Alternatively, we have previously shown in scanning electron micrographs that cytoplasmic extensions of the hepatocytes (microvilli) extend into the sinusoidal lumen through fenestrations in LSECs.23, 24 Possibly these hepatic microvilli contact circulating lymphocytes within the vasculature (Fig. 1B). Finally, contact between lymphocytes and hepatocytes could occur across gaps between LSECs (Fig. 1C), although the presence of such gaps is controversial because it might be artifactual.

Figure 1.

Different possibilities for lymphocyte–hepatocyte interactions: (A) Through LSEC fenetrations via lymphocyte cytoplasmic extensions penetrating the space of Disse and contacting hepatocyte microvilli in the space of Disse. (B) Through fenestrations via hepatocyte microvilli protruding into the sinusoidal lumen. (C) Through gaps between two LSECs.

Here we used electron microscopy to investigate the intimate interactions between lymphocytes, LSECs, and hepatocytes. Initially we focused on interactions between intrahepatic lymphocytes (IHL) and hepatocytes in the normal liver. However, we also investigated whether naïve T lymphocytes interact with hepatocytes using the Met-Kb and Alb-Kb transgenic mouse models in which their cognate antigen is expressed on hepatocytes. Immunogold labeling was performed to identify donor naïve T lymphocytes and to differentiate them from recipient intrahepatic T lymphocytes.


APC, antigen presenting cell; LN, lymph nodes; LSEC, Liver sinusoidal endothelial cell; IHL, Intrahepatic lymphocytes; SEM, scanning electron microscopy; TEM, transmission electron microscopy; CFSE, 5-carboxyfluorescein diacetate succinimidyl ester; TEHLI, trans-endothelial hepatocyte–lymphocyte interactions.

Materials and Methods


All mice were maintained on a B10.BR (H-2k) background in the Centenary Institute animal facility under SPF conditions. B10.BR and C57BL/6 mice were purchased from the Animal Resources Centre (Perth, Australia). Met-Kb and Alb–Kb transgenic mice expressing H-2Kb on hepatocytes under the control of the sheep metallothionein19 or the mouse albumin promoters,18 respectively, were kindly provided by Drs Grant Morahan and Jacques Miller (WEHI, Australia) and Dr Bernd Arnold (DKZ, Germany). Des-TCR, which express an H-2Kb–specific TCR,25 were a gift of Dr. Bernd Arnold. Des-TCR RAG-1−/− mice were generated by backcrossing Des-TCR mice with H–2k RAG-1−/− as previously described.8 All experimental procedures were approved by the University of Sydney Animal Ethics Committee.

Adoptive Transfer.

Pooled LN cells from Des-TCR RAG-1−/− mice were labeled with CFSE (5-carboxyfluorescein diacetate succinimidyl ester), and 5 or 20 million cells were adoptively transferred into recipient transgenic mice as previously described.6

Electron Microscopy.

Five and thirty minutes after the lymphocytes were injected into the portal vein, mice were culled by CO2 narcosis, and livers were immediately perfusion-fixed through the portal vein with a 23-gauge needle at a rate of 3 to 4 mL/min with phosphate-buffered saline and fixative [1% glutaraldehyde, 4% paraformaldehyde, 2 mmol/L CaCl2, 2%(w/v) sucrose 0.1 mol/L cacodylate buffer pH 7.4]. The livers were removed and cut into small pieces before being fixed in the same fixative for 1 hour at 4°C. Fixed tissue was embedded in Spurr's and LR White resin. Liver blocks (at least four per mouse) were analyzed by electron microscopy. For scanning electron microscopy (SEM), fixed tissue was osmicated, dehydrated, and incubated in hexamethyl-disilazane. Tissue was then mounted on stubs, splutter-coated with gold. and examined using a Cambridge S360 Scanning Microscope.

Immunogold and Immunohistochemistry.

Ultra-thin (70-90 nm) sections were cut and collected on formvar-coated nickel slot grids (ProSciTech, Australia). Gold post-embedding immunocytochemistry was then performed to detect LFA-1 or CFSE-labeled cells. Grids were incubated overnight at 4°C with rabbit anti fluorescein/Oregon green primary antibody (Invitrogen, Australia) and for 2 hours at 20°C with gold-conjugated (15 nm or 10 nm, ProSciTech, Australia) secondary antibody. Sections were analyzed for gold-positive lymphocytes, and fields of interest were photographed using a Hitachi H7100FA Transmission Microscope. Part of the fixed tissue was paraffin-embedded, and immunohistochemistry was performed using the same primary antibody. Using a light microscope (magnification 200×), positive cells were counted on 40 mm2 section area for each sample. Immunohistochemistry detecting ICAM-1 and H-2Kb expression was performed on frozen liver sections post-fixation in cold acetone (15 minutes) of recipient mice. Primary antibodies, mouse anti-H-2Kb biotinylated (BD Pharmingen, San Diego, CA) and rat anti-mouse ICAM-1 (Chemicon, Australia) were applied to sections for 2 hours. ICAM-1 sections were incubated with anti-rat fluorescein isothiocyanate (FITC) antibody (Caltag Labs, Carlsbad, CA) or biotinylated anti-ICAM-1 antibody (Sigma Aldrich, Australia) for 1 hour. A solution of streptavidin-horseradish peroxidase (HRPO) conjugate (Dako, Australia) was then applied, and immunolabeling was revealed using 3,3′-diaminobenzidine (Sigma Aldrich).


IHL Interact With Hepatocytes Through Fenestrations in LSECs.

Previous data using Met-Kb and Alb-Kb transgenic mouse models in which antigen expression was restricted to liver parenchymal cells suggest that T lymphocytes can interact with hepatocytes within minutes after adoptive transfer.6, 8 The early onset of these events indicates that lymphocytes can access MHC/peptide complexes expressed by liver parenchymal cells. To investigate how hepatocytes and lymphocytes interact, we analyzed the liver sinusoids using both scanning (SEM) and transmission electron microscopy (TEM).

Fenestrations are unevenly distributed and clustered to form structures known as sieve plates (Fig. 2A-B). Fenestrations are complete perforations in the LSEC and therefore expose the underlying hepatocytes and hepatic stellate cells to the sinusoidal lumen (Fig. 2B-C). Furthermore, neither basal lamina nor connective tissue underlie the LSECs.

Figure 2.

Structure of mouse liver sinusoidal endothelial cells. (A) Under scanning electron microscope (SEM), the liver sinusoidal endothelial cells (LSEC) present fenestrations organized in sieve plates (Original magnification 15,000×). (B) Higher magnification SEM picture of a liver sieve in LSEC illustrating the lack of basal lamina between hepatocytes and LSEC. Hepatic microvilli can be noticed under the LSEC layer (Original magnification 20,000×). (C) In ultra-thin sections, the sinusoid is delimited by endothelial cell thin cytoplasmic processes [e] which contain fenestrations (small arrows). They appear as open connections between the sinusoidal lumen [S] and the space of Disse [sD]. In the space of Disse between the LSEC and the hepatocytes, hepatic stellate cells [HSC, marked with an asterix] processes may be observed. Original magnification 7,000×.

Analysis of scanning electron micrographs demonstrated that a large proportion of the IHL (47%) were in direct contact with the luminal surface of LSEC. Some of these lymphocytes were squeezed by the narrow wall of the sinusoids, favoring intimate contact between the two cells (Fig. 3A-B). All lymphocytes demonstrated numerous cytoplasmic extensions with diameters similar to fenestrations (106.87 ± 11.25nm; Fig. 3B-C). Supporting the interaction model displayed in Fig. 1A, such lymphocyte cytoplasmic extensions were seen within the lumen of the fenestrations, extending into the space of Disse and in contact with hepatocyte microvilli (Fig. 3D). In total, 11% of all IHL found in the lumen (total number of lymphocytes examined, n = 60), were found to interact with hepatocyte microvilli by this mechanism. Such lymphocytes were shown to have up to five extensions traversing adjacent fenestrations on single sections examined by TEM (Fig. 3D), although the total number per lymphocyte will obviously be much greater.

Figure 3.

Interactions between intrahepatic lymphocytes (IHL) and liver cells within the sinusoids. (A) Transmission electron microscopy (TEM) picture of an IHL circulating in the lumen of a liver sinusoid. The narrow diameter of liver sinusoids [S] forces the lymphocyte to squeeze through the sinusoids and establish intimate contact with LSEC [e] (Original magnification 12,000×). [H]: hepatocyte (B) and (C) SEM pictures of IHL, which present villi on their surface. (B) IHL in the sinusoidal lumen displaying numerous cytoplasmic extensions with diameters similar to fenestrations (Original magnification 10,000×). In (C) the IHL appear to contact the surface of liver sinusoidal walls using its villi (Original magnification 18,000×) (D) TEM cross-section picture of a lymphocyte in a sinusoid with five cytoplasmic extensions [indicated by arrows] contacting the basal surface of the underlying hepatocyte (original magnification 10,000×)

Hepatocyte Microvilli Extend Into the Sinusoidal Lumen Through Fenestrations But Do Not Contact Lymphocytes.

As suggested in earlier studies,23, 26 hepatocyte microvilli were seen protruding through the fenestrations and into the sinusoidal lumen (Fig. 4A). However, on no occasion (out of the 60 IHL analyzed) were these microvilli seen to be in contact with lymphocytes in the sinusoid. This indicates that this type of interaction (Fig. 1B) is rare, or alternatively that such interactions are unstable and dislodged during the fixation process.

Figure 4.

Liver sinusoidal endothelial cell (LSECs) are not separated by gaps but allow hepatocytes microvilli to protrude through their fenestrations. (A) Hepatocyte microvilli are occasionally observed to protrude into the sinusoidal lumen [S]. Original magnification 40,000. (B) High magnification scanning electron micrograph (SEM) indicates the absence of gaps between two LSECs (→) (Original magnification 35,000×).

No Gaps Between LSECs.

Electron micrographs also showed that although LSECs do not form tight junctions, they do form a relatively continuous endothelial barrier. No gaps were seen between adjacent LSEC on any occasion, nor were any interactions visualized between lymphocytes and hepatocytes through gaps between adjacent lymphocytes (Fig. 4B). Therefore, contacts between T cells and hepatocytes are unlikely to occur between LSECs (Fig. 1C).

Naïve T Lymphocyte Interact With Hepatocytes Through Fenestrations in LSECs.

As primary activation of antigen-specific T cells by hepatocytes has been evidenced in vivo6, 8; we next investigated whether naïve T cells also establish interactions with hepatocytes through LSEC fenestrations. For this purpose, LN T cells from Des-TCR RAG-1−/− mice, containing a pure and monoclonal population of naïve CD8+ T cells specific for H-2Kb, were labeled with CFSE and injected into control B10.BR mice or Met-Kb and Alb-Kb transgenic mice expressing H-2Kb on hepatocytes.6, 8 Analysis was performed at very early stages of T lymphocyte activation (5 and 30 minutes after injection of lymphocytes).

By counting CFSE-labeled T cells by immunohistochemistry, twice as many labeled donor T lymphocytes were detected in the livers of Met-Kb compared with the control B10.BR mice (Fig. 5A), a result consistent with flow cytometric and radiolabeling studies6, 27; V. Benseler and P. Bertolino unpublished result). Most of the donor T lymphocytes (93%) were found in the sinusoids rather than in portal tracts (Fig. 5B).

Figure 5.

Naïve T lymphocytes can contact hepatocytes through fenestrations. (A) Most naïve T cells injected into Met-Kb and B10.BR animals were localized along the sinusoids rather than around the portal tracts. 5-Carboxyfluorescein diacetate succinimidyl ester (CFSE)-labeled T cells detected by immunohistochemistry using an anti-fluorescein antibody coupled to HRPO. The histograms show the average number of CFSE+ cells in a tissue area of 1 mm2 by light microscopy in B10.BR and Met-Kb mice at 5 minutes and 30 minutes after T cell transfer. An area of 40mm2 was analyzed for each liver. Most positive cells (93%) were located in the parenchyma rather than around portal tracts (B). (C) SEM micrograph of purified naïve T cells. As shown for IHL, naïve T cells isolated from the LN of Des-TCR RAG-1−/− mice also display cytoplasmic extensions (Original magnification 8,000×). (D) LR-White embedded ultra-thin TEM section of an immunogold labeled naïve T lymphocyte establishing one TEHLI (indicated by the large arrow). Liver was perfusion-fixed 5 minutes after adoptive transfer of CFSE+ cells (Original magnification 30,000×). (E) Another example of an immunogold labeled cell contacting both LSEC [*] and hepatocyte (Original magnification 12,000×). The TELHI is indicated by the large arrow. In D and E, 10 nm gold beads, indicated by small arrows and used to identify the donor T cells, are shown in the enlarged detail of the section. No gold labeling was seen in liver cells or in control mice not injected with CFSE-labeled T cells (data not shown).

Naïve T lymphocytes displayed villi similarly to IHL (Fig. 5C) and were recognized by the presence of CFSE detected using immunogold staining. Naïve T lymphocytes (2/17 cells in B10BR mice and 8/52 cells in Met-Kb mice) were observed that had cytoplasmic extensions traversing the fenestrations and in direct contact with hepatocytes (Fig. 5D-E). However, these were less frequently seen compared with the IHL, with only one such interaction seen per TEM image of each naïve lymphocyte (compared with five for IHL). These experiments suggest that naïve T cells are also able to interact with hepatocytes through LSEC fenestrations.

Expression of MHC Class I Molecules and ICAM-1 on Hepatocytes Is Polarized Toward the Sinusoidal Lumen.

The initiation of primary T cell activation requires TCR and LFA-1 cross-linking. TEM immunogold staining of LN cells demonstrates that LFA-1 is expressed on the lymphocyte villae and potentially able to interact with its ligand ICAM-1 (Fig. 6). Both MHC/peptide complex and ICAM-1 have been shown to be expressed at relatively high levels on purified hepatocytes11, 28 (Bertolino P., unpublished results). To investigate whether these molecules were physiologically accessible to circulating T cells, we investigated the in situ localization of both H-2Kb and ICAM-1 in the liver using light and fluorescence microscopy. Interestingly, expression of ICAM-1 in the liver was not evenly distributed but was strongly concentrated toward the surface lining the sinusoids (Fig. 7A-B). Polarized expression of H-2Kb toward the basolateral surface of the hepatocyte was also evidenced in C57BL/6 mice in which all liver cells express H-2Kb (Fig. 7E) but also in Met-Kb and Alb-Kb mice where H-2Kb expression is restricted to hepatocytes (Fig 7C-D).

Figure 6.

Transmission electron micrograph of immunogold staining of LFA-1 on the membrane of a representative lymph node cell. Lymph node cells with a naive morphology isolated from a wild-type C57BL/6 mouse were stained using an anti-LFA-1 antibody and an anti-rat antibody coupled to gold particles. Gold particles (indicated by arrow heads) were distributed relatively evenly on the membrane, including on microvilli. Control sections (no primary anti-LFA-1 antibody) did not display any staining (data not shown).

Figure 7.

ICAM-1 and H-2Kb expression are polarized and localized predominantly toward the lumen. (A) ICAM-1 expression in the liver of B10.BR mice using immunofluorescence. (B) ICAM-1 expression in Met-Kb mice by immunohistochemistry on liver frozen sections. (C-E) H-2Kb expression by immunohistochemistry in Met-Kb mice (C), Alb-Kb (D), and C57BL/6 mice (E). [S]: sinusoid (Original magnification 400×). (F) Retention of CFSE-labeled Des-TCR T cells in the liver of Met-Kb mice is inhibited by blocking ICAM-1 and ICAM-2. CFSE-labeled LN Des-TCR cells were injected into Met-Kb and B10.BR mice as previously described27 in the absence or in the presence of 300 μg of blocking Fab fragments specific for ICAM-1 (KAT-1) and ICAM-2 (3C4), provided by Prof A. Hamann. Liver retention of donor CD8+ T cells was measured at 1h after T cell transfer by calculating the total number of CD8+ CFSE+ T cells harvested from this organ using flow cytometry as previously described.8 Histograms represent the mean and standard error of 2 to 4 mice per group.

These results suggest that expression of two of the molecules required for primary T cell activation, ICAM-1 and MHC Class I, is polarized toward the basolateral surface of the hepatocyte, thus maximizing the primary activation signal that hepatocytes might provide to T cells

To test whether ICAM-1 expressed by liver cells is critical for the retention of antigen-specific T cells when H-2Kb is expressed by hepatocytes, Met-Kb and B10. BR mice were injected with CFSE-labeled Des-TCR LN T cells in the absence or in the presence of Fab fragments specific for ICAM-1 and ICAM-2 that block the interactions of these molecules with their common ligand LFA-1. As shown in Fig. 5A, the total number of donor CD8+ T cells isolated from the liver of Met-Kb mice at 1 hour after T cell transfer was approximately twice the number isolated from B10.BR mice. Anti–ICAM-1 and ICAM-2 treatment inhibited this antigen-specific retention to control levels observed in B10.BR mice (Fig. 7F). These results suggest that T cell retention mediated by hepatocytes is ICAM-1/LFA-1 dependent.


The electron micrographs presented in this study support the model of interaction depicted in Fig. 1A. and indicate that both IHL and naïve T lymphocytes insert cytoplasmic extensions through fenestrations in the LSEC and that such extensions make direct contact with the hepatocellular membranes. On two-dimensional TEM of ultra-thin sections, we observed up to five such extensions on each IHL and one on naïve T lymphocytes. This indicates that these interactions are quite common in each lymphocyte, and they were seen in at least 10% to 15% of naïve T lymphocytes in the liver. Our study probably underestimates the prevalence of these interactions because TEM is limited in the number of cells that can be studied and provides only a single section of each lymphocyte. We propose to term these interactions trans-endothelial hepatocyte–lymphocyte interactions (TEHLI).

Cytoplasmic extensions, or pseudopods, on lymphocytes are present in both naïve and activated T lymphocytes, and their size is similar to LSEC fenestrations, which provides opportunity for the extensions to pass through the fenestrations and interact with hepatic microvilli. Recent reports indicate that only 10 peptide/MHC complexes are necessary to reach a maximal Ca2+ T cell response and form a stable synapse,29 suggesting that the sizes of the extensions and fenestrations are large enough to initiate and accommodate nascent immunological synapses. T lymphocyte activation through the TEHLI seems therefore possible. We also have shown that naïve T lymphocytes establish TEHLI. Although these results do not provide direct evidence that TEHLI mediate T cell activation, speculating that naïve T lymphocytes might be activated by hepatocytes via TEHLI is tempting. If this is the case, this provides a pivotal observation underpinning the potential role of hepatocytes as APCs in vivo. Naïve T lymphocytes established relatively few TEHLI, perhaps suggesting that they are more random or less stable than those established by activated lymphocytes, where adhesion molecules and integrins recognition is likely to facilitate the formation of immunological synapses and enhance the opportunity for, and stability of, TEHLI.

The number of TEHLI established by naïve T cells at 5 and 30 minutes after T cell transfer in Met-Kb and B10.BR mice was similar, suggesting that naïve T cells scan the hepatocyte basolateral surface independently of antigen expression, similarly to dendritic cell/T cell interactions.30 Thus serial contacts with peptide/MHC complexes may lead to a cumulative activation signal in the T cell, upregulation and affinity changes of LFA-1 resulting in T cell arrest. We propose that these changes allow activated T cells to establish more stable TEHLI. This model explains why IHL, known to be mostly activated cells, displayed more TEHLI than naïve T cells.

Confirming our previous observations, we have shown here, using both SEM and TEM, the presence of hepatic microvilli extending through the fenestrations into the sinusoidal lumen. Although we have not found any clear evidence that these microvilli interact with circulating lymphocytes (Fig. 1B), we cannot totally exclude that such brief interactions occur and contribute to T cell activation.

Other factors might facilitate TEHLI. The low blood flow rate within the liver sinusoids31–33 favors interactions between blood lymphocytes and liver cells.10 This might allow primary activation to occur in the absence of selectins.8, 34 In addition, lymphocytes are larger than some sinusoids and therefore can massage the LSEC as they traverse the liver21 (Fig. 3A). This increases the possibility of cytoplasmic extensions on the lymphocytes passing through fenestrations. The absence of basal lamina and collagen in the space of Disse also would facilitate access to the hepatocellular membrane. Finally, we found that the expression of ICAM-1 and MHC Class I molecules, which are required for primary T cell activation, is limited to the basolateral membrane of the hepatocyte, which would maximize opportunity for receptor recognition and immunological synapse.

Many implications follow the discovery of TEHLI. This is the first ultrastructural demonstration of any interaction between naïve T lymphocytes and parenchymal cells in vivo. The general immunological paradigm is that naïve T lymphocytes must be activated by professional APCs in LN before they develop the capacity to traverse any endothelial barrier and invade the surrounding tissue. However, here we have clearly observed direct contact between naïve T lymphocytes and hepatocytes in the absence of inflammation, suggesting that the liver is an exception to this paradigm. Contact between hepatocytes and naïve T cells is suggested by our previous studies using transgenic mouse models. Naïve T lymphocytes expressing a transgenic TCR specific for the allo-MHC class I molecule H-2Kb are activated within minutes after adoptive transfer in the liver of Met-Kb and Alb-Kb transgenic mice expressing the H-2Kb molecule on hepatocytes. Although Met-Kb mice do display some ectopic expression of H-2Kb in the LN on some professional APCs, the expression of the antigen in the liver is restricted to hepatocytes.8 Similarly, in Alb-Kb mice, H-2Kb expression is detected on hepatocytes but not on LSECs and Kupffer cells.8 As the transgenic TCR recognizes the intact H-2Kb class I molecule rather than a peptide derived from H-2Kb, cross-presentation is excluded in this model. Antigen-specific retention and activation of naïve T lymphocytes in the liver within minutes of adoptive transfer implies that H-2Kb expressed by hepatocytes is readily available and recognized by naïve T lymphocytes. We confirmed that the lymphocytes that were seen on electron microscopy were naïve T lymphocytes by using immunogold to detect the marker CFSE, which is present only in the naïve T lymphocytes derived from the donor mice.

Clearly, the observation of TEHLI between naïve T lymphocytes and hepatocytes consolidates the evidence for the role of hepatocytes as APCs. In vitro experiments have demonstrated that hepatocytes can act as very efficient APCs for high-affinity transgenic T lymphocytes and that this activation is ICAM-1 dependent.16 The observation that antigen-specific T cell activation occurs in both Met-Kb and Alb-Kb transgenic mice indicates that hepatocytes also can act as efficient APCs in situ. This property would be favored by the high local expression of MHC Class I and ICAM-1 on the basolateral surface of the hepatocytes. Based on a large number of immunohistochemical studies in mice and humans, expression of MHC molecules on hepatocytes has been generally regarded as very low.35, 36 However, a recent report11 has shown that these cells express abundant and conformationally stable MHC class I/peptide complexes with surface densities that are nearly as high as on splenocytes. Using quantitative flow cytometry techniques and calibration standards to adjust for the differences in cell size and autofluorescence between hepatocytes and splenocytes, it was demonstrated that, on a per cell basis, mouse hepatocytes express 7- to 16-fold higher levels of MHC class I molecules than splenocytes, whereas membrane densities were at least 30% to 75% as high as those estimated on splenic lymphocytes.11 Expression of both ICAM-1 and MHC class I is normally polarized. Consequently, it is likely that local densities of these molecules are higher at the basolateral surface of hepatocytes. Freshly isolated hepatocytes would be expected to redistribute expression of these molecules on the whole plasma membrane surface after dissociation, thus decreasing overall membrane ligand density. Because peptide/MHC and ICAM-1 density influences T cell activation,37 we therefore predict that in vitro experiments using hepatocytes underestimate the APC capability of these cells under physiological conditions.

Hepatocytes may act as APCs only in situations in which they express the relevant antigen, for example, the Met-Kb transgenic mouse model of autoimmune hepatitis and the early phase of viral hepatitis. Intrahepatic T lymphocyte activation during early hepatitis C infection could contribute to the impaired immune response observed in chronic hepattis C virus.9, 38 Using the Met-Kb and Alb-Kb models, we have demonstrated that T lymphocyte activation in the liver is an inefficient process, leading to diminished cytotoxic activity and reduced cell survival,8 a type of response consistent with immune tolerance. Based on these results, we propose that TEHLI play a critical role in mediating activation of naïve T lymphocytes by hepatocytes and in the development of tolerance. By transplanting mouse livers expressing the ovalbumin antigen into C57BL/6 mice, a recent study39 has confirmed our initial finding that antigen-specific naïve transgenic T cells are activated intrahepatically. However, in their model, T cells activated in the liver mature into full effector cells rather than being deleted, thus contradicting the concept that primary activation in the liver leads to tolerance. Whether these discrepancies reflect differences between transgenic models or, as we have recently proposed, whether inflammation occurring during transplantation alters the fate of intrahepatically activated T cells, is unclear.9

As well as providing insight into the normal immune system, our observations might have implications for liver conditions associated with altered LSEC morphology and in particular those conditions associated with loss of fenestrations such as cirrhosis and old age. We have shown that old age is associated with dramatic reductions in the fenestrations of LSECs23, 40, 41; therefore, the altered immune responses of older people might in part be mechanistically linked to reduced opportunity for TEHLI in old age.

In conclusion, the results of this study indicate that direct contact occurs between lymphocytes and hepatocytes via processes we term TEHLI. This report provides ultrastructural evidence of interaction between naïve T lymphocytes and parenchymal cells in vivo. We propose that activation of naïve T lymphocytes by antigens presented on the hepatocellular membrane occurs via TEHLI. If they result in T cell activation, TEHLI provide a new mechanism that impacts on the immune response; provides a plausible mechanism for altered immune responses in conditions associated with loss of fenestrations such as aging; and generates a potential therapeutic target for modulating the immune response in viral hepatitis and autoimmune disease.


The authors thank Dr Alf Hamann for providing the anti-ICAM-1 and anti-ICAM-2 Fab fragments, Drs Bard Smedsrod and Randi Olsen for their precious help with the immunogold staining of LFA-1, Drs Bernd Arnold, Grant Morahan, and Jacques Miller for providing us with the Alb-Kb, Des-TCR, and Met-Kb mice, as well as Jenny Kingham and the staff of the Centenary Institute animal facility.