- Email: [email protected]
Marked changes of the hepatic sinusoid in a transgenic mouse model of acute immune-mediated hepatitis
Journal of Hepatology 46 (2007) 239–246 www.elsevier.com/locate/jhep
Marked changes of the hepatic sinusoid in a transgenic mouse model of acute immune-mediated hepatitis Alessandra Warren1, Patrick Bertolino2, Volker Benseler2,3, Robin Fraser4, Geoffrey W. McCaughan2, David G. Le Couteur1,* 1
Centre for Education and Research on Ageing (CERA) and the ANZAC Research Institute, Concord RG Hospital and University of Sydney, Sydney, Australia 2 AW Morrow Gastroenterology and Liver Centre, Centenary Institute of Cancer Medicine and Cell Biology, Royal Prince Alfred Hospital and University of Sydney, Sydney, Australia 3 University of Regensburg Medical Center, Department of Surgery, Regensburg, Germany 4 Department of Pathology, Christchurch School of Medicine and Health Sciences, University of Otago, Christchurch, New Zealand
Background/Aims: The liver sinusoidal endothelial cell (LSEC) is increasingly recognized as having an important role in hepatic immunity. However, the responses of LSECs and the hepatic sinusoid in immune-mediated hepatitis are poorly described. Methods: We studied a transgenic mouse model of acute immune-mediated hepatitis: Met-Kb mice injected with T cells from Des-TCR mice. Results: Hepatitis was characterized by lymphocyte infiltrates causing severe but transient liver damage. There were marked changes in the ultrastructure of the LSEC five days after injection of the T cells that coincided with the peak of the hepatitis. The porosity of fenestrations in the LSEC decreased and the endothelium became thickened. LSECs appeared to be markedly activated. These changes were associated with narrowing of the space of Disse, loss of hepatocellular microvilli and deposition of basal lamina. Lymphocytes were seen passing through fenestrations. Loss of fenestration in the LSEC prevented hepatitis induced by a second injection of lymphocytes on day 5. Conclusions: Structural changes in the LSEC occur during the peak of a mouse model of immune-mediated hepatitis. These changes were associated with attenuation of subsequent liver damage, suggesting that they may influence immunological responses mediated by LSECs or the passage of lymphocytes through LSEC fenestrations. 2006 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved. Keywords: Hepatitis; Hepatocyte; Liver sinusoidal endothelial cell; Fenestration; Mice
1. Introduction Autoimmune hepatitis is a progressive and relatively common disorder [1]. The etiology remains unknown although aberrant autoreactivity is considered to play a significant role and hepatocyte antigens such as CYP2D6 and SLA/LP have been proposed as targets Received 21 April 2006; received in revised form 30 June 2006; accepted 5 August 2006; available online 23 October 2006 * Corresponding author. Tel.: +612 9767 6935; fax: +612 9767 5419. E-mail address: [email protected] (D.G. Le Couteur).
for T-cell-mediated autoimmunity [2]. Recently a transgenic mouse has been developed as a model for the study of immune-mediated hepatitis. Here, Met-Kb mice express MHC class I H-2Kb on hepatocytes which are recognized by adoptively transferred Des-TCR CD8+ T cells expressing a transgenic TCR specific for H-2Kb [3–5]. Following injection of these CD8+ T cells, the recipient transgenic mice develop severe transient liver damage associated with biochemical and histological hepatitis, a result consistent with autoimmune hepatitis. Most pathological studies of hepatitis, including immune-mediated hepatitis, have focussed on
0168-8278/$32.00 2006 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.jhep.2006.08.022
240
A. Warren et al. / Journal of Hepatology 46 (2007) 239–246
hepatocytes and the inflammatory infiltrate. Acute and chronic hepatitis, caused by viral infection, autoimmune disease or toxins, is characterized by prominent infiltration in the parenchyma and perivascular tissue of activated lymphocytes, mainly T cells [6]. The distribution of lymphocyte infiltration is dependent on the type of inflammatory stimulus. In biliary disease the infiltration is distributed around the portal tract, while in viral and autoimmune hepatitis it is more lobular and localised in the parenchyma [7]. The inflammatory infiltrate is usually associated with hepatocyte injury, apoptosis and regeneration. On the other hand, studies on the effects of hepatitis on the hepatic sinusoid are limited. This is surprising given the crucial role of hepatic sinusoidal cells in liver immunology [8]. Furthermore, there is increasing recognition of the importance of the hepatic sinusoid in a broad range of liver diseases such as cirrhosis, fibrosis [9], steatosis [10] and even old age [11]. The liver sinusoidal endothelial cell (LSEC) is of particular interest because of the role of these cells in lymphocyte interactions and antigen presentation [8,12] as well as in facilitating the transfer of substrates between the sinusoidal blood and the hepatocytes [9]. Changes in the hepatic sinusoid under pathological conditions such as viral or drug-induced hepatitis have been reported [13–15]. These include loss of LSEC fenestrations, which are pores in the endothelial cell that facilitate the transfer of substrates between blood and hepatocytes [9,16–18]. In human chronic active hepatitis, it was found that LSECs became swollen and activated and this was associated with deposition of basal lamina and reticulin fibres [19]. Although some studies have described changes in LSEC during hepatitis, very few have investigated changes in these cells during immune-mediated hepatitis. Furthermore, ultrastructural studies in hepatitis have been mostly performed in human samples harvested during the active phase of the disease. This limits the quality of tissue available for scanning electron microscopy and does not allow for analysis of the time course of changes. Therefore we investigated the hepatic sinusoid, particularly the LSEC, in the transgenic Met-Kb mice, which develop severe immune-mediated hepatitis following transfer of Des-TCR CD8+ T cells [4]. This model is ideal to analyse the ultrastructure of sinusoids and their interaction with the inflammatory infiltrate before, during and after the peak of hepatitis.
2. Materials and methods 2.1. Mice The study had the approval of the South Western Sydney Area Health Service Animal Welfare Committee. All the animals were
maintained in the Centenary Institute Animal Facility (University of Sydney, Australia) under specific pathogen-free conditions. B10.BR mice were purchased from the Animal Resources Centre, Perth, WA, Australia. Transgenic Met-Kb mice express the allo-H-2Kb molecule on hepatocytes under the control of the sheep metallothionein promoter [20]. These mice, bred onto the B10.BR background, were kindly provided by Drs. Grant Morahan and Jacques Miller (The Walter and Eliza Hall Institute of Medical Research, Melbourne, Australia). Transgenic Des-TCR mice express on 98% of their CD8+ T cells a TCR recognizing a complex formed by the MHC class I H-2Kb molecule and self-peptides. They were a kind gift of Dr. Bernd Arnold (Deutsches Krebsforschungszentrum, Heidelberg, Germany).
2.2. Adoptive transfer Lymph nodes were removed from Des-TCR mice and a single cell suspension was prepared by pressing the tissue through 100 lm mesh sieves and by washing lymphocytes twice with T-cell medium (RPMI 1640, 10% fetal calf serum, 2 mM L-glutamine and 50 lm 2-b mercaptoethanol: JRH Biosciences, Australia). Fifteen million lymph node cells (0.5 ml), of which 30% are CD8+ T cells expressing Des-TCR, were injected into the lateral tail vein of Met-Kb mice. To determine whether structural changes in the LSEC influenced subsequent inflammatory responses, additional experiments were undertaken where transgenic Des-TCR T cells were injected both on day 0 and day 5.
2.3. Liver samples At 2.5 days, 5 days and 16 days from injection, recipient Met-Kb mice were sacrificed with CO2 and livers were immediately perfusionfixed via a 23 gauge needle inserted into the portal vein. Liver tissue was fixed with 1% glutaraldehyde, 4% paraformaldehyde in PBS (0.1 M sucrose), processed for electron microscopy and paraffin-embedded for light microscopy as described previously [11,21]. Transgenic Met-Kb mice that had not received the Des-TCR lymphocytes were used as controls.
2.4. Light microscopy Part of the liver perfusion-fixed for electron microscopy was paraffin embedded following standard procedure. Five-micrometer thick sections were stained with haematoxylin–eosin and Sirius red for collagen fibres [22].
2.5. Transmission electron microscopy Fixed liver tissue was processed and embedded in Spurr’s resin. Blocks were sampled at random for light microscopic assessment. Three blocks per liver were finally studied, selected randomly from those satisfying requirements for quality of fixation and tissue integrity. Ultra-thin (70–90 nm) sections were taken from each block and ten fields at random were chosen from each liver and photographed for ultrastructural measurement using a Zeiss 902 Transmission Microscope (magnification 12,000·). Transmission electron micrographic measurements (100 per liver) of the thickness of the sinusoidal endothelial cells and the width of the space of Disse were made using the Zeiss KS Image Analysis program. For the basal lamina and collagen deposition analysis 30 sinusoids at random for each time points were studied.
2.6. Scanning electron microscopy For scanning electron microscopy, perfusion-fixed tissue was osmicated (1% OsO4/0.1 mol/l sodium cacodylate buffer), dehydrated in an ethanol gradient to 100% and incubated for 10 min in hexamethyldisilazane. Tissue was then mounted on stubs, splutter-coated with
A. Warren et al. / Journal of Hepatology 46 (2007) 239–246
241
gold and examined using a Jeol Scanning Microscope. Ten images (magnification 25,000·) were taken from each animal for analysis of fenestration diameter and endothelial porosity using the Zeiss KS Image Analysis program. Endothelial porosity was defined as the percentage of the area of the sinusoidal endothelium that was perforated with fenestrations.
2.7. Statistical analysis Statistical analysis was performed with SigmaStat 2.03 (SPSS Inc, Ill). Results are expressed as means ± SEM. Comparisons between groups were undertaken using Kruskal–Wallis test with a post hoc Dunn analysis. P value of <0.05 was considered statistically significant.
3. Results 3.1. Light microscopy of Met-Kb mice injected with DesTCR T cells revealed abundant autoimmune infiltrate and severe liver damage To induce immune-mediated hepatitis, Met-Kb transgenic mice were injected with transgenic Des-TCR T cells as previously published [3,5]. Livers of control Met-Kb mice, not injected with transgenic T cells, were normal, indicating that the transgene had no effect on hepatic or sinusoidal architecture (Figs. 1a and c). Typical features of hepatitis were observed following adoptive transfer of Des-TCR T cells including focal hepatocellular necrosis, fatty change and lobular inflammation. The sinusoidal architecture was disrupted and many sinusoids contained enlarged Kupffer cells and cellular debris. There was an inflammatory parenchymal infiltrate, which was mild at 2.5 days and abundant by 5 days after adoptive transfer. Infiltrating cells were found around the portal tracts and throughout the parenchyma (Fig. 1b). Staining with Sirius red at the peak of hepatitis indicated the presence of collagen fibres within the cellular infiltrate (Fig. 1d). By day 16, much of the parenchymal inflammatory infiltrate had resolved. Consistent with histological observations, serum ALT levels peaked at day 5 and had returned to normal by day 15 (Fig. 1e). Resolution of the disease is associated with peripheral T cell deletion of activated transgenic T cells [3]. 3.2. Electron microscopy of the hepatic sinusoid revealed significant changes in the ultrastructure of the LSEC Significant changes in the ultrastructure of the LSEC were most apparent at the peak of hepatitis within proximity of inflammatory infiltrates. One of the major changes was the loss of fenestrations in the LSECs, a process termed ‘defenestration’ in previous studies [11,23,24] (Fig. 2). Defenestration was only observed at the peak of hepatitis and this was confirmed by image analysis quantification of
e
Fig. 1. Light microscopy of mouse immune-mediated hepatitis. (a) Haematoxylin and eosin (H&E) staining in control transgenic Met-Kb mouse. Original magnification 200·. (b) H&E in transgenic mouse at the peak of hepatitis. There is extensive inflammatory infiltrate in the parenchyma. Original magnification 200·. (c) Sirius red staining for collagen in normal transgenic mouse. Original magnification 200·. (d) Sirius red staining in transgenic mouse at the peak of inflammation (day 5). There are collagen fibres in the infiltrate. Original magnification 400·. (e) Blood ALT levels in Met-Kb mice following injection of T lymphocytes from Des-TCR mice (n = 4).
porosity determined by scanning electron microscopy (Table 1). By day 16 the endothelium had re-fenestrated. Scanning electron microscopy confirmed defenestration and loss of sieve plate formation in the sinusoidal endothelium of animals at the peak of hepatitis (Fig. 2) although these changes were not uniform. The porosity of the endothelium at day 5 was reduced by about 40% while it returned to almost normal levels at day 16 (Table 1). There was no change in the frequency of large endothelial gaps seen in other forms of liver injury. There were many features consistent with activation of LSECs. They were significantly swollen from day 2.5 to day 16 (Table 1, Figs. 2c, e and g). At days 2.5, day 5 and day 16, LSECs contained increased numbers of granules and dense cytoplasmic inclusions around the nucleus. These were associated with occasional small round dark granules present in the cytoplasmic extensions lining the sinusoid (Fig. 3). Macro- and
242
A. Warren et al. / Journal of Hepatology 46 (2007) 239–246
The percentage of sinusoids with underlying basement membrane-like material deposition in the space of Disse increased from about 25% at day 0 and 2.5 days, to about 60% at days 5 and 16 (Table 1). LSECs were not the only cells affected by the immune-mediated hepatitis. At the onset of hepatitis, hepatic stellate cells contained several lipid droplets and were in close proximity to LSECs. Their rough endoplasmic reticulum was abundant and well developed. By day 16, some hepatic stellate cells were free of lipid droplets and displayed collagen fibres near their surfaces. In these mice, collagen deposition in the space of Disse was observed frequently. Kupffer cells were also enlarged and contained multiple granules, cell fragments and apoptotic cells (Fig. 4b). 3.3. Lymphocyte infiltration Inflammatory cell infiltrates were widespread at day 5. Lymphocytes that had traversed the endothelial barrier were partially enveloped by cytoplasmic arms of hepatic stellate cell (Fig. 4b) or hepatocyte or were contained in endothelial pockets (Fig. 4c). Some infiltrating lymphocytes present in the lumen appeared to extend their cytoplasm through endothelial fenestrations (Fig. 4a) suggesting that T cells may undergo transendothelial migration through LSEC fenestrations rather than at the intercellular junction between two neighbouring LSEC. On scanning electron microscopy, inflammatory cells also appeared to migrate into the parenchyma through enlarged gaps on the sinusoidal surface (Fig. 4d).
Fig. 2. Electron microscopy of LSECs in mouse immune-mediated hepatitis. Transmission electron micrographs are shown in (a, c, e and g) and scanning electron micrographs are shown in (b, d, f and h). Day 0 (a and b); day 2.5 (c and d); day 5 (e and f); day 16 (g and h). There are reduced fenestrations and hepatocyte microvilli with the presence of basal lamina-like material (large arrow) in the space of Disse [sD] at day 5. Sinusoidal lumen: [S]; fenestrations (small arrow); Hepatocyte: [H]. Original magnifications 12,000· for transmission electron microscopy and 25,000· for scanning electron microscopy.
3.4. Changes in the hepatic endothelium were associated with protection from further liver damage
micro-pinocytotic vesicles were also observed on the surface and in the cytoplasm of LSECs. Abundant vesicles and enlargement of the Golgi apparatus were also observed occasionally in LSECs at day 16 (Fig. 3b). The width of the space of Disse was significantly reduced at day 5 in areas close to inflammatory infiltrates (Table 1). This was associated with almost complete loss of hepatocyte microvilli and deposition of electron dense material in the space of Disse (Fig. 2e).
The findings above suggest that fenestrations play a critical role in mediating T cell trans-endothelial migration to the parenchyma. To investigate whether reduction in the porosity of the hepatic endothelium associated with the hepatitis influenced subsequent liver damage, we investigated whether mice developing hepatitis had altered susceptibility to a second injection of lymphocytes. Met-Kb mice injected with the first cohort of transgenic Des-TCR T cells on day 0 were injected
Table 1 Quantification of the ultrastructural changes seen in hepatic sinusoid in mouse immune-mediated hepatitis Day 0 Endothelial thickness (nm) Space of Disse width (nm) Porosity (% of endothelial area) Fenestration diameter (nm) Basal lamina (% of sinusoids) Collagen deposition (% of sinusoids)
3*
134 ± 567 ± 13 2.9 ± 0.4 80 ± 2 25 25
*P 6 0.05 using Kruskal–Wallis test with a post hoc Dunn analysis.
Day 2.5
Day 5
Day 16
149 ± 5 521 ± 12 3.1 ± 0.2 91 ± 1* 26 26
148 ± 4 427 ± 25* 1.9 ± 0.2* 76 ± 1* 64 15
148 ± 4 585 ± 14 3.0 ± 0.3 104 ± 2* 58 33
A. Warren et al. / Journal of Hepatology 46 (2007) 239–246
Fig. 3. Transmission electron micrograph of an activated LSEC at day 5 (a) and day 16 (b). Activated LSECs present abundant Golgi apparatus [G], dense bodies (small arrows), vacuoles [v] and pinocytotic vesicles [*]. There is basal membrane-like material deposition in the space of Disse at day 5 (large arrow). Original magnification 7000· (a); 10,000· (b).
again at the peak of hepatitis (day 5) with a second cohort of CFSE-labelled transgenic Des-TCR T cells (Fig. 5). As expected, all CD8+ T cells of the second cohort proliferated in the lymph nodes of recipient mice as assessed by dilution of the CFSE labelling (Fig. 5b) indicating that the first cohort did not inhibit proliferation of the second cohort. Proliferation was antigen-specific as it was not observed in adoptively transferred B10.BR mice (Fig. 5b). Despite extensive proliferation of the second T cell cohort, Met-Kb mice serially injected with two T cell cohorts did not develop severe subsequent hepatitis in contrast to Met-Kb mice injected with PBS and a single Des-TCR T cell cohort at day 5 (Fig. 5a). These results suggest that the marked changes of the endothelium and the fenestrations associated with the hepatitis limit T cell access to the parenchyma therefore protecting the liver from further damage.
4. Discussion The Met-Kb transgenic mouse model provides a unique opportunity to study the ultrastructure of the LSEC and hepatic sinusoid during immune-mediated hepatitis. We found that there is significant defenestration of the LSEC at the peak of hepatitis at day 5. Interestingly, LSECs had re-fenestrated by day 16, indicating that defenestration is a reversible process at least in this experimental model. In addition, there was some deposition of basal lamina and collagen. The LSECs
243
Fig. 4. Lymphocytes and other inflammatory cells traversing the LSEC during the peak of hepatitis (day 5). (a) Lymphocyte that is migrating from the sinusoid [S] to the space of Disse via an endothelial fenestration. [e] and small arrows: endothelium; [H]: hepatocyte. The nuclei and cytoplasm of the lymphocyte are very attenuated at the points of passage (large arrows). (b) Sinusoid [S] with a hypertrophic Kupffer cell [K] containing an apoptotic body with typical chromatin marginalization [A] and other cellular debris. A lymphocyte [L] traversing through a pore (large arrow) in the space of Disse is partially surrounded by a hepatic stellate cell process [HSC]. Original magnification 4000·. (c) Lymphocyte [L] migrating into the space of Disse via large gaps (large arrow) and another contained in an endothelial pocket. [H]: hepatocyte. Original magnification 3000·. (d) Scanning electron micrograph of two infiltrating cells in the space of Disse facing the sinusoidal lumen through large gaps. The surrounding endothelium appears defenestrated. Original magnification 15,000·.
appeared to be activated, as evidenced by swelling and the appearance of granules and dense cytoplasmic inclusions around the nucleus, macro- and micro-pinocytotic vesicles, and enlargement of the Golgi apparatus. The results are consistent with those reported by Barbadin in human chronic active hepatitis, viral hepatitis and drug-induced hepatitis where on transmission electron microscopy, the LSECs appeared to be activated and there was the formation of complete, incomplete and pseudo-capillaries [13,19]. Although it is not intended to mimic the complexity of hepatitis associated with human diseases or transplantation, the ultrastructural changes of sinusoids and LSECs observed in the transgenic Met-Kb mouse model of immune-mediated hepatitis are very similar to those described in human hepatitis [13,19], confirming that it is suitable for the study of immune-mediated hepatitis. We also observed that the width of the space of Disse was reduced at the peak of the hepatitis and this was associated with retraction of the hepatic microvilli. This
244
A. Warren et al. / Journal of Hepatology 46 (2007) 239–246
Fig. 5. Marked changes in the hepatic endothelium are associated with protection from subsequent liver damage. (a) Kinetics of serum ALT levels in MetKb mice injected with two cohorts of Des-TCR transgenic T cells at day 0 and day 5 (dashed bold line, full square), a single cohort of Des-TCR T cells at day 0 (bold line, full triangle), PBS at day 0 and a cohort of Des-TCR T cells at day 5 (dashed thin line, open white circle). (b) CFSE profile of CD8+ T lymphocytes isolated from the lymph nodes of B10.BR (empty histogram) and Met-Kb (grey histogram) recipient mice injected with a first cohort of DesTCR T cells at day 0 and a second cohort of CFSE-labelled Des-TCR T cells at day 5. Proliferation was assessed 3 days after injection of the second T cell cohort. CFSE-labelled T cells proliferated in Met-Kb mice as assessed by the dilution of CFSE. No proliferation was observed in control B10.BR mice. T cells on the left represent host CD8+ T cells (negative for CFSE).
has been previously described in the process of capillarization in cirrhosis [25]. The mechanism for the retraction or loss of microvilli is unclear, however it may reduce the surface available for antigen presentation to T cells and thus provide protection against further lymphocyte activation and inflammatory damage. Presumably the loss of microvilli would also impair metabolic exchange. We found that lymphocytes appear to migrate between the blood and extravascular space of Disse through fenestrations in the LSEC, which is a novel and intriguing observation. Fenestrations are 50– 150 nm diameter pores in the LSEC that lack a diaphragm and represent a true discontinuity in the LSEC that permits the passage of plasma and various substrates including lipoproteins to travel freely between the sinusoid and space of Disse [9,16]. It is impressive how the cytoplasm and the nucleus of the lymphocytes appear to squeeze through the LSEC fenestrations. Similarly, leukemic myelocytes have been observed passing through fenestrations in LSECs [26]. Consistent with its autoimmune origin, the lymphocytic infiltrate seen in the Met-Kb mice was located within the parenchyma and around the portal tracts [3]. The diversity of infiltration distribution and lymphocyte adhesion appears to be controlled by LSEC expression of adhesion molecules and chemokines such as ICAM1, VCAM-1, VAP-1 activated during inflammation [7]. In a concavalin A mouse model of hepatitis, it was found that the main site of lymphocyte adhesion and extravasation was the sublobular vein, however lymphocyte recruitment via the sinusoids was also demonstrated. Similarly we found LSEC projections extended along adhering lymphocytes, as well as lymphocytes migrating
through the fenestrations in LSECs. Activated lymphocytes and inflammatory cells secrete reactive oxidative species [27,28] and our studies on the effect of oxidative stress on the liver showed that the sinusoidal endothelial cells become swollen and more porous, with large gaps replacing sieve plates [29]. It is possible that activated inflammatory cells also invade the space of Disse via large gaps created by oxidants generated during the inflammatory process. A possible example of this is shown in Fig. 4d. The structural changes in the endothelium seen in hepatitis may have important functional implications. Bardadin and Desmet suggested that activation of sinusoidal endothelial cells and subsequent gradual transformation of sinusoids into capillaries may constitute a protection mechanism for hepatocytes against potentially harmful foreign molecules [13]. Defenestration also could prevent further leukocyte infiltration in the space of Disse. We have confirmed that liver damage induced by autoreactive transgenic CTLs had a protective effect upon adoptive transfer of a potentially harmful second T cell cohort. Protection was not due to an impairment of T cell proliferation as cells from the second cohort proliferated normally in the lymph nodes. It is also unlikely that these changes were induced directly by LSECs in this model as these cells do not express H-2Kb and are therefore unable to activate directly the T cells [5]. The most plausible explanation is that following proliferation in lymph nodes and maturation into effector cells, CTLs have limited access to the parenchyma because of the decrease in fenestration porosity. Loss of LSEC fenestrations also has important implications in liver physiology in particular in lipid
A. Warren et al. / Journal of Hepatology 46 (2007) 239–246
and drug metabolism [9,16]. One of the expected consequences of defenestration would be an increase in the total amount of lipoproteins such as chylomicron remnants circulating in the serum [30]. Although we have not investigated lipid serum levels in the Met-Kb model, a human study of acute hepatitis reported an increase in serum lipoproteins and hypertriglyceridaemia [31]. In conclusion, the LSECs became activated and defenestrated in this transgenic mouse model of immune-mediated hepatitis. Lymphocytes are known to interact extensively with LSECs and we observed lymphocytes migrating through fenestrations in LSECs. This suggests that the changes in the LSEC during hepatitis might have a role in influencing the immune response and ultimately the clinical outcome in autoimmune hepatitis. Acknowledgements The authors wish to thank Grant Morahan and Jacques Miller for providing us with the Des-TCR and Met-Kb mice, as well as Jenny Kingham and the staff of the Centenary Institute animal facility. This work was supported by the National Health and Medical Research Council of Australia (NHMRC) (Project Grant #352342, Program Grant #358308) and the Ageing and Alzheimer’s Research Foundation. A.W. was supported by the Ageing and Alzheimer’s Research Foundation and a NHMRC postgraduate scholarship. V.B. was supported by a German DFG scholarship. References [1] Boberg K. Prevalence and epidemiology of autoimmune hepatitis. Clin Liver Dis 2002;6:635–647. [2] Krawitt E. Autoimmune hepatitis. New Engl J Med 2006;354: 54–66. [3] Bertolino P, Heath W, Hardy C, Morahan G, Miller G. Peripheral deletion of autoreactive CD8+ T cells in transgenic mice expressing H-2Kb in the liver. Eur J Immunol 1995;25:1932–1942. [4] Bertolino P, Bowen DG, McCaughan GW, Fazekas de St. Groth B. Antigen-specific primary activation of CD8+ T cells within the liver. J Immunol 2001;166:5430–5438. [5] Bowen DG, Zen M, Holz L, Davis T, McCaughan GW, Bertolino P. The site of primary T cell activation is a determinant of the balance between intrahepatic tolerance and immunity. J Clin Invest 2004;114:701–712. [6] Volpes R, van den Oort J, Desmet V. Memory T cells represent the predominant lymphocyte subset in acute and chronic liver inflammation. Hepatology 1991;13:826–829. [7] Lalor P, Shields P, Grant A, Adams D. Recruitment of lymphocytes to the human liver. Immunol Cell Biol 2002;80:52–64. [8] Racanelli V, Rehermann B. The liver as an immunological organ. Hepatology 2006;43:S54–S61. [9] Le Couteur DG, Fraser R, Hilmer SN, Rivory LP, McLean AJ. The hepatic sinusoid in aging and cirrhosis: effects on hepatic substrate disposition and drug clearance. Clin Pharmacokinet 2005;44:187–200.
245
[10] Hakamada K, Sasaki M, Takahashi K, Umehara Y, Konn M. Sinusoidal flow block after warm ischemia in rats with dietinduced fatty liver. J Surg Res 1997;70:12–20. [11] Le Couteur DG, Cogger VC, Markus AM, Harvey PJ, Yin ZL, Ansselin AD, McLean AJ. Pseudocapillarization and associated energy limitation in the aged rat liver. Hepatology 2001;33: 537–543. [12] Limmer A, Ohl J, Kurts C, Ljunggren H, Reiss Y, Groettrup M, Momburg F, Arnold B, Knolle P. Efficient presentation of exogenous antigen by liver endothelial cells to CD8+ T cells results in antigen-specific T-cell tolerance. Nat Med 2000;6: 1348–1354. [13] Bardadin KA, Scheuer PJ. Endothelial cell changes in acute hepatitis. A light and electron microscopic study. J Pathol 1984;144:213–220. [14] Bingen A, Martin JP, Klein F, Pessah M. Modification of the amount of cholesterol in hepatic steatosis induced in susceptible and resistant mice infected with MHV3: a biochemical and ultrastructural study. Hepatology 1992;15:1137–1146. [15] Steffan A, Pereira C, Bingen A, Valle M, Martin J, Koehren F, Royer C, Gendrault JL, Kirn A. Mouse hepatitis-virus type-3 infection provokes a decrease in the number of sinusoidal-cell fenestrae both in vivo and in vitro. Hepatology 1995;22: 395–401. [16] Fraser R, Dobbs BR, Rogers GWT. Lipoproteins and the liver sieve: the role of the fenestrated sinusoidal endothelium in lipoprotein metabolism, atherosclerosis, and cirrhosis. Hepatology 1995;21:863–874. [17] Wisse E, Van_Dierendonck JH, De_Zanger RB, Fraser R, McCuskey RS. On the role of the liver endothelial filter in the transport of particulate fat (chylomicrons and their remnants) to parenchymal cells and the influence of certain hormones on the endothelial fenestrae. In: Popper H, editor. Communications of Liver Cells. Lancaster: MTP Press; 1980. p. 195–200. [18] Wisse E, Zanger RBD, Charels K, Smissen PVD, McCuskey RS. The liver sieve: considerations concerning the structure and function of endothelial fenestrae, the sinusoidal wall and the space of Disse. Hepatology 1985;5:683–692. [19] Bardadin KA, Desmet VJ. Ultrastructural observations on sinusoidal endothelial cells in chronic active hepatitis. Histopathology 1985;9:171–181. [20] Morahan G, Allison J, Miller J. Tolerance of class I histocompatibility antigens expressed extrathymically. Nature 1989;339: 622–624. [21] Cogger VC, Warren A, Fraser R, Ngu M, McLean AJ, Le Couteur DG. Hepatic sinusoidal pseudocapillarization with aging in the non-human primate. Exp Gerontol 2003;38: 1101–1107. [22] Lopez-De Leon A, Rojkind M. A simple micromethod for collagen and total protein determination in formalin fixed paraffin-embedded sections. J Histochem Cytochem 1985;33: 737–743. [23] Fraser R, Rogers GW, Bowler LM, Day AJ, Dobbs B, Baxter JN. Defenestration and vitamin A status in a rat model of cirrhosis. In: Wisse E, Knook DL, McCuskey RS, editors. Cells of the hepatic sinusoid. Leiden: The Kupffer Cell Foundation; 1992. p. 195–198. [24] Mori T, Okanoue T, Sawa Y, Hori N, Ohta M, Kagawa K. Defenestration of the sinusoidal endothelial cell in a rat model of cirrhosis. Hepatology 1993;17:891–897. [25] Schaffner F, Popper H. Capillarisation of hepatic sinusoids in man. Gastroenterology 1963;44:239–242. [26] Campbell F. Ultrastructural study of liver sinusoids of mice during invasion by leukemic myelocytes. J Natl Cancer Inst 1977;58:369–376. [27] Devadas S, Zaritskaya L, Rhee S, Oberley L, Williams M. Discrete generation of superoxide and hydrogen peroxide by T cell receptor stimulation: selective regulation of mitogen-activated protein kinase activation and fas ligand expression. J Exp Med 2002;195:59–70.
246
A. Warren et al. / Journal of Hepatology 46 (2007) 239–246
[28] Williams M, Kwong J. T cell receptor stimulation, reactive oxygen species, and cell signaling. Free Radic Biol Med 2004;37: 1144–1151. [29] Cogger VC, Mross PE, Hosie MJ, Ansselin AD, McLean AJ, Le Couteur DG. The effect of acute oxidative stress on the ultrastructure of the perfused rat liver. Pharmacol Toxicol 2001;89:306–311.
[30] Le Couteur DG, Fraser R, Cogger VC, McLean AJ. Hepatic pseudocapillarisation and atherosclerosis in ageing. Lancet 2002;359:1612–1615. [31] Vergani C, Trovato G, Delu A, Pietrogrande M, Dioguardi N. Serumtotal lipids, lipoprotein cholesterol, and apolipoprotein A in acute viral hepatitis and chronic liver disease. J Clin Pathol 1978;31:772–778.
Marked changes of the hepatic sinusoid in a transgenic mouse model of acute immune-mediated hepatitis Alessandra Warren1, Patrick Bertolino2, Volker Benseler2,3, Robin Fraser4, Geoffrey W. McCaughan2, David G. Le Couteur1,* 1
Centre for Education and Research on Ageing (CERA) and the ANZAC Research Institute, Concord RG Hospital and University of Sydney, Sydney, Australia 2 AW Morrow Gastroenterology and Liver Centre, Centenary Institute of Cancer Medicine and Cell Biology, Royal Prince Alfred Hospital and University of Sydney, Sydney, Australia 3 University of Regensburg Medical Center, Department of Surgery, Regensburg, Germany 4 Department of Pathology, Christchurch School of Medicine and Health Sciences, University of Otago, Christchurch, New Zealand
Background/Aims: The liver sinusoidal endothelial cell (LSEC) is increasingly recognized as having an important role in hepatic immunity. However, the responses of LSECs and the hepatic sinusoid in immune-mediated hepatitis are poorly described. Methods: We studied a transgenic mouse model of acute immune-mediated hepatitis: Met-Kb mice injected with T cells from Des-TCR mice. Results: Hepatitis was characterized by lymphocyte infiltrates causing severe but transient liver damage. There were marked changes in the ultrastructure of the LSEC five days after injection of the T cells that coincided with the peak of the hepatitis. The porosity of fenestrations in the LSEC decreased and the endothelium became thickened. LSECs appeared to be markedly activated. These changes were associated with narrowing of the space of Disse, loss of hepatocellular microvilli and deposition of basal lamina. Lymphocytes were seen passing through fenestrations. Loss of fenestration in the LSEC prevented hepatitis induced by a second injection of lymphocytes on day 5. Conclusions: Structural changes in the LSEC occur during the peak of a mouse model of immune-mediated hepatitis. These changes were associated with attenuation of subsequent liver damage, suggesting that they may influence immunological responses mediated by LSECs or the passage of lymphocytes through LSEC fenestrations. 2006 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved. Keywords: Hepatitis; Hepatocyte; Liver sinusoidal endothelial cell; Fenestration; Mice
1. Introduction Autoimmune hepatitis is a progressive and relatively common disorder [1]. The etiology remains unknown although aberrant autoreactivity is considered to play a significant role and hepatocyte antigens such as CYP2D6 and SLA/LP have been proposed as targets Received 21 April 2006; received in revised form 30 June 2006; accepted 5 August 2006; available online 23 October 2006 * Corresponding author. Tel.: +612 9767 6935; fax: +612 9767 5419. E-mail address: [email protected] (D.G. Le Couteur).
for T-cell-mediated autoimmunity [2]. Recently a transgenic mouse has been developed as a model for the study of immune-mediated hepatitis. Here, Met-Kb mice express MHC class I H-2Kb on hepatocytes which are recognized by adoptively transferred Des-TCR CD8+ T cells expressing a transgenic TCR specific for H-2Kb [3–5]. Following injection of these CD8+ T cells, the recipient transgenic mice develop severe transient liver damage associated with biochemical and histological hepatitis, a result consistent with autoimmune hepatitis. Most pathological studies of hepatitis, including immune-mediated hepatitis, have focussed on
0168-8278/$32.00 2006 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.jhep.2006.08.022
240
A. Warren et al. / Journal of Hepatology 46 (2007) 239–246
hepatocytes and the inflammatory infiltrate. Acute and chronic hepatitis, caused by viral infection, autoimmune disease or toxins, is characterized by prominent infiltration in the parenchyma and perivascular tissue of activated lymphocytes, mainly T cells [6]. The distribution of lymphocyte infiltration is dependent on the type of inflammatory stimulus. In biliary disease the infiltration is distributed around the portal tract, while in viral and autoimmune hepatitis it is more lobular and localised in the parenchyma [7]. The inflammatory infiltrate is usually associated with hepatocyte injury, apoptosis and regeneration. On the other hand, studies on the effects of hepatitis on the hepatic sinusoid are limited. This is surprising given the crucial role of hepatic sinusoidal cells in liver immunology [8]. Furthermore, there is increasing recognition of the importance of the hepatic sinusoid in a broad range of liver diseases such as cirrhosis, fibrosis [9], steatosis [10] and even old age [11]. The liver sinusoidal endothelial cell (LSEC) is of particular interest because of the role of these cells in lymphocyte interactions and antigen presentation [8,12] as well as in facilitating the transfer of substrates between the sinusoidal blood and the hepatocytes [9]. Changes in the hepatic sinusoid under pathological conditions such as viral or drug-induced hepatitis have been reported [13–15]. These include loss of LSEC fenestrations, which are pores in the endothelial cell that facilitate the transfer of substrates between blood and hepatocytes [9,16–18]. In human chronic active hepatitis, it was found that LSECs became swollen and activated and this was associated with deposition of basal lamina and reticulin fibres [19]. Although some studies have described changes in LSEC during hepatitis, very few have investigated changes in these cells during immune-mediated hepatitis. Furthermore, ultrastructural studies in hepatitis have been mostly performed in human samples harvested during the active phase of the disease. This limits the quality of tissue available for scanning electron microscopy and does not allow for analysis of the time course of changes. Therefore we investigated the hepatic sinusoid, particularly the LSEC, in the transgenic Met-Kb mice, which develop severe immune-mediated hepatitis following transfer of Des-TCR CD8+ T cells [4]. This model is ideal to analyse the ultrastructure of sinusoids and their interaction with the inflammatory infiltrate before, during and after the peak of hepatitis.
2. Materials and methods 2.1. Mice The study had the approval of the South Western Sydney Area Health Service Animal Welfare Committee. All the animals were
maintained in the Centenary Institute Animal Facility (University of Sydney, Australia) under specific pathogen-free conditions. B10.BR mice were purchased from the Animal Resources Centre, Perth, WA, Australia. Transgenic Met-Kb mice express the allo-H-2Kb molecule on hepatocytes under the control of the sheep metallothionein promoter [20]. These mice, bred onto the B10.BR background, were kindly provided by Drs. Grant Morahan and Jacques Miller (The Walter and Eliza Hall Institute of Medical Research, Melbourne, Australia). Transgenic Des-TCR mice express on 98% of their CD8+ T cells a TCR recognizing a complex formed by the MHC class I H-2Kb molecule and self-peptides. They were a kind gift of Dr. Bernd Arnold (Deutsches Krebsforschungszentrum, Heidelberg, Germany).
2.2. Adoptive transfer Lymph nodes were removed from Des-TCR mice and a single cell suspension was prepared by pressing the tissue through 100 lm mesh sieves and by washing lymphocytes twice with T-cell medium (RPMI 1640, 10% fetal calf serum, 2 mM L-glutamine and 50 lm 2-b mercaptoethanol: JRH Biosciences, Australia). Fifteen million lymph node cells (0.5 ml), of which 30% are CD8+ T cells expressing Des-TCR, were injected into the lateral tail vein of Met-Kb mice. To determine whether structural changes in the LSEC influenced subsequent inflammatory responses, additional experiments were undertaken where transgenic Des-TCR T cells were injected both on day 0 and day 5.
2.3. Liver samples At 2.5 days, 5 days and 16 days from injection, recipient Met-Kb mice were sacrificed with CO2 and livers were immediately perfusionfixed via a 23 gauge needle inserted into the portal vein. Liver tissue was fixed with 1% glutaraldehyde, 4% paraformaldehyde in PBS (0.1 M sucrose), processed for electron microscopy and paraffin-embedded for light microscopy as described previously [11,21]. Transgenic Met-Kb mice that had not received the Des-TCR lymphocytes were used as controls.
2.4. Light microscopy Part of the liver perfusion-fixed for electron microscopy was paraffin embedded following standard procedure. Five-micrometer thick sections were stained with haematoxylin–eosin and Sirius red for collagen fibres [22].
2.5. Transmission electron microscopy Fixed liver tissue was processed and embedded in Spurr’s resin. Blocks were sampled at random for light microscopic assessment. Three blocks per liver were finally studied, selected randomly from those satisfying requirements for quality of fixation and tissue integrity. Ultra-thin (70–90 nm) sections were taken from each block and ten fields at random were chosen from each liver and photographed for ultrastructural measurement using a Zeiss 902 Transmission Microscope (magnification 12,000·). Transmission electron micrographic measurements (100 per liver) of the thickness of the sinusoidal endothelial cells and the width of the space of Disse were made using the Zeiss KS Image Analysis program. For the basal lamina and collagen deposition analysis 30 sinusoids at random for each time points were studied.
2.6. Scanning electron microscopy For scanning electron microscopy, perfusion-fixed tissue was osmicated (1% OsO4/0.1 mol/l sodium cacodylate buffer), dehydrated in an ethanol gradient to 100% and incubated for 10 min in hexamethyldisilazane. Tissue was then mounted on stubs, splutter-coated with
A. Warren et al. / Journal of Hepatology 46 (2007) 239–246
241
gold and examined using a Jeol Scanning Microscope. Ten images (magnification 25,000·) were taken from each animal for analysis of fenestration diameter and endothelial porosity using the Zeiss KS Image Analysis program. Endothelial porosity was defined as the percentage of the area of the sinusoidal endothelium that was perforated with fenestrations.
2.7. Statistical analysis Statistical analysis was performed with SigmaStat 2.03 (SPSS Inc, Ill). Results are expressed as means ± SEM. Comparisons between groups were undertaken using Kruskal–Wallis test with a post hoc Dunn analysis. P value of <0.05 was considered statistically significant.
3. Results 3.1. Light microscopy of Met-Kb mice injected with DesTCR T cells revealed abundant autoimmune infiltrate and severe liver damage To induce immune-mediated hepatitis, Met-Kb transgenic mice were injected with transgenic Des-TCR T cells as previously published [3,5]. Livers of control Met-Kb mice, not injected with transgenic T cells, were normal, indicating that the transgene had no effect on hepatic or sinusoidal architecture (Figs. 1a and c). Typical features of hepatitis were observed following adoptive transfer of Des-TCR T cells including focal hepatocellular necrosis, fatty change and lobular inflammation. The sinusoidal architecture was disrupted and many sinusoids contained enlarged Kupffer cells and cellular debris. There was an inflammatory parenchymal infiltrate, which was mild at 2.5 days and abundant by 5 days after adoptive transfer. Infiltrating cells were found around the portal tracts and throughout the parenchyma (Fig. 1b). Staining with Sirius red at the peak of hepatitis indicated the presence of collagen fibres within the cellular infiltrate (Fig. 1d). By day 16, much of the parenchymal inflammatory infiltrate had resolved. Consistent with histological observations, serum ALT levels peaked at day 5 and had returned to normal by day 15 (Fig. 1e). Resolution of the disease is associated with peripheral T cell deletion of activated transgenic T cells [3]. 3.2. Electron microscopy of the hepatic sinusoid revealed significant changes in the ultrastructure of the LSEC Significant changes in the ultrastructure of the LSEC were most apparent at the peak of hepatitis within proximity of inflammatory infiltrates. One of the major changes was the loss of fenestrations in the LSECs, a process termed ‘defenestration’ in previous studies [11,23,24] (Fig. 2). Defenestration was only observed at the peak of hepatitis and this was confirmed by image analysis quantification of
e
Fig. 1. Light microscopy of mouse immune-mediated hepatitis. (a) Haematoxylin and eosin (H&E) staining in control transgenic Met-Kb mouse. Original magnification 200·. (b) H&E in transgenic mouse at the peak of hepatitis. There is extensive inflammatory infiltrate in the parenchyma. Original magnification 200·. (c) Sirius red staining for collagen in normal transgenic mouse. Original magnification 200·. (d) Sirius red staining in transgenic mouse at the peak of inflammation (day 5). There are collagen fibres in the infiltrate. Original magnification 400·. (e) Blood ALT levels in Met-Kb mice following injection of T lymphocytes from Des-TCR mice (n = 4).
porosity determined by scanning electron microscopy (Table 1). By day 16 the endothelium had re-fenestrated. Scanning electron microscopy confirmed defenestration and loss of sieve plate formation in the sinusoidal endothelium of animals at the peak of hepatitis (Fig. 2) although these changes were not uniform. The porosity of the endothelium at day 5 was reduced by about 40% while it returned to almost normal levels at day 16 (Table 1). There was no change in the frequency of large endothelial gaps seen in other forms of liver injury. There were many features consistent with activation of LSECs. They were significantly swollen from day 2.5 to day 16 (Table 1, Figs. 2c, e and g). At days 2.5, day 5 and day 16, LSECs contained increased numbers of granules and dense cytoplasmic inclusions around the nucleus. These were associated with occasional small round dark granules present in the cytoplasmic extensions lining the sinusoid (Fig. 3). Macro- and
242
A. Warren et al. / Journal of Hepatology 46 (2007) 239–246
The percentage of sinusoids with underlying basement membrane-like material deposition in the space of Disse increased from about 25% at day 0 and 2.5 days, to about 60% at days 5 and 16 (Table 1). LSECs were not the only cells affected by the immune-mediated hepatitis. At the onset of hepatitis, hepatic stellate cells contained several lipid droplets and were in close proximity to LSECs. Their rough endoplasmic reticulum was abundant and well developed. By day 16, some hepatic stellate cells were free of lipid droplets and displayed collagen fibres near their surfaces. In these mice, collagen deposition in the space of Disse was observed frequently. Kupffer cells were also enlarged and contained multiple granules, cell fragments and apoptotic cells (Fig. 4b). 3.3. Lymphocyte infiltration Inflammatory cell infiltrates were widespread at day 5. Lymphocytes that had traversed the endothelial barrier were partially enveloped by cytoplasmic arms of hepatic stellate cell (Fig. 4b) or hepatocyte or were contained in endothelial pockets (Fig. 4c). Some infiltrating lymphocytes present in the lumen appeared to extend their cytoplasm through endothelial fenestrations (Fig. 4a) suggesting that T cells may undergo transendothelial migration through LSEC fenestrations rather than at the intercellular junction between two neighbouring LSEC. On scanning electron microscopy, inflammatory cells also appeared to migrate into the parenchyma through enlarged gaps on the sinusoidal surface (Fig. 4d).
Fig. 2. Electron microscopy of LSECs in mouse immune-mediated hepatitis. Transmission electron micrographs are shown in (a, c, e and g) and scanning electron micrographs are shown in (b, d, f and h). Day 0 (a and b); day 2.5 (c and d); day 5 (e and f); day 16 (g and h). There are reduced fenestrations and hepatocyte microvilli with the presence of basal lamina-like material (large arrow) in the space of Disse [sD] at day 5. Sinusoidal lumen: [S]; fenestrations (small arrow); Hepatocyte: [H]. Original magnifications 12,000· for transmission electron microscopy and 25,000· for scanning electron microscopy.
3.4. Changes in the hepatic endothelium were associated with protection from further liver damage
micro-pinocytotic vesicles were also observed on the surface and in the cytoplasm of LSECs. Abundant vesicles and enlargement of the Golgi apparatus were also observed occasionally in LSECs at day 16 (Fig. 3b). The width of the space of Disse was significantly reduced at day 5 in areas close to inflammatory infiltrates (Table 1). This was associated with almost complete loss of hepatocyte microvilli and deposition of electron dense material in the space of Disse (Fig. 2e).
The findings above suggest that fenestrations play a critical role in mediating T cell trans-endothelial migration to the parenchyma. To investigate whether reduction in the porosity of the hepatic endothelium associated with the hepatitis influenced subsequent liver damage, we investigated whether mice developing hepatitis had altered susceptibility to a second injection of lymphocytes. Met-Kb mice injected with the first cohort of transgenic Des-TCR T cells on day 0 were injected
Table 1 Quantification of the ultrastructural changes seen in hepatic sinusoid in mouse immune-mediated hepatitis Day 0 Endothelial thickness (nm) Space of Disse width (nm) Porosity (% of endothelial area) Fenestration diameter (nm) Basal lamina (% of sinusoids) Collagen deposition (% of sinusoids)
3*
134 ± 567 ± 13 2.9 ± 0.4 80 ± 2 25 25
*P 6 0.05 using Kruskal–Wallis test with a post hoc Dunn analysis.
Day 2.5
Day 5
Day 16
149 ± 5 521 ± 12 3.1 ± 0.2 91 ± 1* 26 26
148 ± 4 427 ± 25* 1.9 ± 0.2* 76 ± 1* 64 15
148 ± 4 585 ± 14 3.0 ± 0.3 104 ± 2* 58 33
A. Warren et al. / Journal of Hepatology 46 (2007) 239–246
Fig. 3. Transmission electron micrograph of an activated LSEC at day 5 (a) and day 16 (b). Activated LSECs present abundant Golgi apparatus [G], dense bodies (small arrows), vacuoles [v] and pinocytotic vesicles [*]. There is basal membrane-like material deposition in the space of Disse at day 5 (large arrow). Original magnification 7000· (a); 10,000· (b).
again at the peak of hepatitis (day 5) with a second cohort of CFSE-labelled transgenic Des-TCR T cells (Fig. 5). As expected, all CD8+ T cells of the second cohort proliferated in the lymph nodes of recipient mice as assessed by dilution of the CFSE labelling (Fig. 5b) indicating that the first cohort did not inhibit proliferation of the second cohort. Proliferation was antigen-specific as it was not observed in adoptively transferred B10.BR mice (Fig. 5b). Despite extensive proliferation of the second T cell cohort, Met-Kb mice serially injected with two T cell cohorts did not develop severe subsequent hepatitis in contrast to Met-Kb mice injected with PBS and a single Des-TCR T cell cohort at day 5 (Fig. 5a). These results suggest that the marked changes of the endothelium and the fenestrations associated with the hepatitis limit T cell access to the parenchyma therefore protecting the liver from further damage.
4. Discussion The Met-Kb transgenic mouse model provides a unique opportunity to study the ultrastructure of the LSEC and hepatic sinusoid during immune-mediated hepatitis. We found that there is significant defenestration of the LSEC at the peak of hepatitis at day 5. Interestingly, LSECs had re-fenestrated by day 16, indicating that defenestration is a reversible process at least in this experimental model. In addition, there was some deposition of basal lamina and collagen. The LSECs
243
Fig. 4. Lymphocytes and other inflammatory cells traversing the LSEC during the peak of hepatitis (day 5). (a) Lymphocyte that is migrating from the sinusoid [S] to the space of Disse via an endothelial fenestration. [e] and small arrows: endothelium; [H]: hepatocyte. The nuclei and cytoplasm of the lymphocyte are very attenuated at the points of passage (large arrows). (b) Sinusoid [S] with a hypertrophic Kupffer cell [K] containing an apoptotic body with typical chromatin marginalization [A] and other cellular debris. A lymphocyte [L] traversing through a pore (large arrow) in the space of Disse is partially surrounded by a hepatic stellate cell process [HSC]. Original magnification 4000·. (c) Lymphocyte [L] migrating into the space of Disse via large gaps (large arrow) and another contained in an endothelial pocket. [H]: hepatocyte. Original magnification 3000·. (d) Scanning electron micrograph of two infiltrating cells in the space of Disse facing the sinusoidal lumen through large gaps. The surrounding endothelium appears defenestrated. Original magnification 15,000·.
appeared to be activated, as evidenced by swelling and the appearance of granules and dense cytoplasmic inclusions around the nucleus, macro- and micro-pinocytotic vesicles, and enlargement of the Golgi apparatus. The results are consistent with those reported by Barbadin in human chronic active hepatitis, viral hepatitis and drug-induced hepatitis where on transmission electron microscopy, the LSECs appeared to be activated and there was the formation of complete, incomplete and pseudo-capillaries [13,19]. Although it is not intended to mimic the complexity of hepatitis associated with human diseases or transplantation, the ultrastructural changes of sinusoids and LSECs observed in the transgenic Met-Kb mouse model of immune-mediated hepatitis are very similar to those described in human hepatitis [13,19], confirming that it is suitable for the study of immune-mediated hepatitis. We also observed that the width of the space of Disse was reduced at the peak of the hepatitis and this was associated with retraction of the hepatic microvilli. This
244
A. Warren et al. / Journal of Hepatology 46 (2007) 239–246
Fig. 5. Marked changes in the hepatic endothelium are associated with protection from subsequent liver damage. (a) Kinetics of serum ALT levels in MetKb mice injected with two cohorts of Des-TCR transgenic T cells at day 0 and day 5 (dashed bold line, full square), a single cohort of Des-TCR T cells at day 0 (bold line, full triangle), PBS at day 0 and a cohort of Des-TCR T cells at day 5 (dashed thin line, open white circle). (b) CFSE profile of CD8+ T lymphocytes isolated from the lymph nodes of B10.BR (empty histogram) and Met-Kb (grey histogram) recipient mice injected with a first cohort of DesTCR T cells at day 0 and a second cohort of CFSE-labelled Des-TCR T cells at day 5. Proliferation was assessed 3 days after injection of the second T cell cohort. CFSE-labelled T cells proliferated in Met-Kb mice as assessed by the dilution of CFSE. No proliferation was observed in control B10.BR mice. T cells on the left represent host CD8+ T cells (negative for CFSE).
has been previously described in the process of capillarization in cirrhosis [25]. The mechanism for the retraction or loss of microvilli is unclear, however it may reduce the surface available for antigen presentation to T cells and thus provide protection against further lymphocyte activation and inflammatory damage. Presumably the loss of microvilli would also impair metabolic exchange. We found that lymphocytes appear to migrate between the blood and extravascular space of Disse through fenestrations in the LSEC, which is a novel and intriguing observation. Fenestrations are 50– 150 nm diameter pores in the LSEC that lack a diaphragm and represent a true discontinuity in the LSEC that permits the passage of plasma and various substrates including lipoproteins to travel freely between the sinusoid and space of Disse [9,16]. It is impressive how the cytoplasm and the nucleus of the lymphocytes appear to squeeze through the LSEC fenestrations. Similarly, leukemic myelocytes have been observed passing through fenestrations in LSECs [26]. Consistent with its autoimmune origin, the lymphocytic infiltrate seen in the Met-Kb mice was located within the parenchyma and around the portal tracts [3]. The diversity of infiltration distribution and lymphocyte adhesion appears to be controlled by LSEC expression of adhesion molecules and chemokines such as ICAM1, VCAM-1, VAP-1 activated during inflammation [7]. In a concavalin A mouse model of hepatitis, it was found that the main site of lymphocyte adhesion and extravasation was the sublobular vein, however lymphocyte recruitment via the sinusoids was also demonstrated. Similarly we found LSEC projections extended along adhering lymphocytes, as well as lymphocytes migrating
through the fenestrations in LSECs. Activated lymphocytes and inflammatory cells secrete reactive oxidative species [27,28] and our studies on the effect of oxidative stress on the liver showed that the sinusoidal endothelial cells become swollen and more porous, with large gaps replacing sieve plates [29]. It is possible that activated inflammatory cells also invade the space of Disse via large gaps created by oxidants generated during the inflammatory process. A possible example of this is shown in Fig. 4d. The structural changes in the endothelium seen in hepatitis may have important functional implications. Bardadin and Desmet suggested that activation of sinusoidal endothelial cells and subsequent gradual transformation of sinusoids into capillaries may constitute a protection mechanism for hepatocytes against potentially harmful foreign molecules [13]. Defenestration also could prevent further leukocyte infiltration in the space of Disse. We have confirmed that liver damage induced by autoreactive transgenic CTLs had a protective effect upon adoptive transfer of a potentially harmful second T cell cohort. Protection was not due to an impairment of T cell proliferation as cells from the second cohort proliferated normally in the lymph nodes. It is also unlikely that these changes were induced directly by LSECs in this model as these cells do not express H-2Kb and are therefore unable to activate directly the T cells [5]. The most plausible explanation is that following proliferation in lymph nodes and maturation into effector cells, CTLs have limited access to the parenchyma because of the decrease in fenestration porosity. Loss of LSEC fenestrations also has important implications in liver physiology in particular in lipid
A. Warren et al. / Journal of Hepatology 46 (2007) 239–246
and drug metabolism [9,16]. One of the expected consequences of defenestration would be an increase in the total amount of lipoproteins such as chylomicron remnants circulating in the serum [30]. Although we have not investigated lipid serum levels in the Met-Kb model, a human study of acute hepatitis reported an increase in serum lipoproteins and hypertriglyceridaemia [31]. In conclusion, the LSECs became activated and defenestrated in this transgenic mouse model of immune-mediated hepatitis. Lymphocytes are known to interact extensively with LSECs and we observed lymphocytes migrating through fenestrations in LSECs. This suggests that the changes in the LSEC during hepatitis might have a role in influencing the immune response and ultimately the clinical outcome in autoimmune hepatitis. Acknowledgements The authors wish to thank Grant Morahan and Jacques Miller for providing us with the Des-TCR and Met-Kb mice, as well as Jenny Kingham and the staff of the Centenary Institute animal facility. This work was supported by the National Health and Medical Research Council of Australia (NHMRC) (Project Grant #352342, Program Grant #358308) and the Ageing and Alzheimer’s Research Foundation. A.W. was supported by the Ageing and Alzheimer’s Research Foundation and a NHMRC postgraduate scholarship. V.B. was supported by a German DFG scholarship. References [1] Boberg K. Prevalence and epidemiology of autoimmune hepatitis. Clin Liver Dis 2002;6:635–647. [2] Krawitt E. Autoimmune hepatitis. New Engl J Med 2006;354: 54–66. [3] Bertolino P, Heath W, Hardy C, Morahan G, Miller G. Peripheral deletion of autoreactive CD8+ T cells in transgenic mice expressing H-2Kb in the liver. Eur J Immunol 1995;25:1932–1942. [4] Bertolino P, Bowen DG, McCaughan GW, Fazekas de St. Groth B. Antigen-specific primary activation of CD8+ T cells within the liver. J Immunol 2001;166:5430–5438. [5] Bowen DG, Zen M, Holz L, Davis T, McCaughan GW, Bertolino P. The site of primary T cell activation is a determinant of the balance between intrahepatic tolerance and immunity. J Clin Invest 2004;114:701–712. [6] Volpes R, van den Oort J, Desmet V. Memory T cells represent the predominant lymphocyte subset in acute and chronic liver inflammation. Hepatology 1991;13:826–829. [7] Lalor P, Shields P, Grant A, Adams D. Recruitment of lymphocytes to the human liver. Immunol Cell Biol 2002;80:52–64. [8] Racanelli V, Rehermann B. The liver as an immunological organ. Hepatology 2006;43:S54–S61. [9] Le Couteur DG, Fraser R, Hilmer SN, Rivory LP, McLean AJ. The hepatic sinusoid in aging and cirrhosis: effects on hepatic substrate disposition and drug clearance. Clin Pharmacokinet 2005;44:187–200.
245
[10] Hakamada K, Sasaki M, Takahashi K, Umehara Y, Konn M. Sinusoidal flow block after warm ischemia in rats with dietinduced fatty liver. J Surg Res 1997;70:12–20. [11] Le Couteur DG, Cogger VC, Markus AM, Harvey PJ, Yin ZL, Ansselin AD, McLean AJ. Pseudocapillarization and associated energy limitation in the aged rat liver. Hepatology 2001;33: 537–543. [12] Limmer A, Ohl J, Kurts C, Ljunggren H, Reiss Y, Groettrup M, Momburg F, Arnold B, Knolle P. Efficient presentation of exogenous antigen by liver endothelial cells to CD8+ T cells results in antigen-specific T-cell tolerance. Nat Med 2000;6: 1348–1354. [13] Bardadin KA, Scheuer PJ. Endothelial cell changes in acute hepatitis. A light and electron microscopic study. J Pathol 1984;144:213–220. [14] Bingen A, Martin JP, Klein F, Pessah M. Modification of the amount of cholesterol in hepatic steatosis induced in susceptible and resistant mice infected with MHV3: a biochemical and ultrastructural study. Hepatology 1992;15:1137–1146. [15] Steffan A, Pereira C, Bingen A, Valle M, Martin J, Koehren F, Royer C, Gendrault JL, Kirn A. Mouse hepatitis-virus type-3 infection provokes a decrease in the number of sinusoidal-cell fenestrae both in vivo and in vitro. Hepatology 1995;22: 395–401. [16] Fraser R, Dobbs BR, Rogers GWT. Lipoproteins and the liver sieve: the role of the fenestrated sinusoidal endothelium in lipoprotein metabolism, atherosclerosis, and cirrhosis. Hepatology 1995;21:863–874. [17] Wisse E, Van_Dierendonck JH, De_Zanger RB, Fraser R, McCuskey RS. On the role of the liver endothelial filter in the transport of particulate fat (chylomicrons and their remnants) to parenchymal cells and the influence of certain hormones on the endothelial fenestrae. In: Popper H, editor. Communications of Liver Cells. Lancaster: MTP Press; 1980. p. 195–200. [18] Wisse E, Zanger RBD, Charels K, Smissen PVD, McCuskey RS. The liver sieve: considerations concerning the structure and function of endothelial fenestrae, the sinusoidal wall and the space of Disse. Hepatology 1985;5:683–692. [19] Bardadin KA, Desmet VJ. Ultrastructural observations on sinusoidal endothelial cells in chronic active hepatitis. Histopathology 1985;9:171–181. [20] Morahan G, Allison J, Miller J. Tolerance of class I histocompatibility antigens expressed extrathymically. Nature 1989;339: 622–624. [21] Cogger VC, Warren A, Fraser R, Ngu M, McLean AJ, Le Couteur DG. Hepatic sinusoidal pseudocapillarization with aging in the non-human primate. Exp Gerontol 2003;38: 1101–1107. [22] Lopez-De Leon A, Rojkind M. A simple micromethod for collagen and total protein determination in formalin fixed paraffin-embedded sections. J Histochem Cytochem 1985;33: 737–743. [23] Fraser R, Rogers GW, Bowler LM, Day AJ, Dobbs B, Baxter JN. Defenestration and vitamin A status in a rat model of cirrhosis. In: Wisse E, Knook DL, McCuskey RS, editors. Cells of the hepatic sinusoid. Leiden: The Kupffer Cell Foundation; 1992. p. 195–198. [24] Mori T, Okanoue T, Sawa Y, Hori N, Ohta M, Kagawa K. Defenestration of the sinusoidal endothelial cell in a rat model of cirrhosis. Hepatology 1993;17:891–897. [25] Schaffner F, Popper H. Capillarisation of hepatic sinusoids in man. Gastroenterology 1963;44:239–242. [26] Campbell F. Ultrastructural study of liver sinusoids of mice during invasion by leukemic myelocytes. J Natl Cancer Inst 1977;58:369–376. [27] Devadas S, Zaritskaya L, Rhee S, Oberley L, Williams M. Discrete generation of superoxide and hydrogen peroxide by T cell receptor stimulation: selective regulation of mitogen-activated protein kinase activation and fas ligand expression. J Exp Med 2002;195:59–70.
246
A. Warren et al. / Journal of Hepatology 46 (2007) 239–246
[28] Williams M, Kwong J. T cell receptor stimulation, reactive oxygen species, and cell signaling. Free Radic Biol Med 2004;37: 1144–1151. [29] Cogger VC, Mross PE, Hosie MJ, Ansselin AD, McLean AJ, Le Couteur DG. The effect of acute oxidative stress on the ultrastructure of the perfused rat liver. Pharmacol Toxicol 2001;89:306–311.
[30] Le Couteur DG, Fraser R, Cogger VC, McLean AJ. Hepatic pseudocapillarisation and atherosclerosis in ageing. Lancet 2002;359:1612–1615. [31] Vergani C, Trovato G, Delu A, Pietrogrande M, Dioguardi N. Serumtotal lipids, lipoprotein cholesterol, and apolipoprotein A in acute viral hepatitis and chronic liver disease. J Clin Pathol 1978;31:772–778.