Antigen Presentation by Liver Cells Controls Intrahepatic T Cell ...

8
of April 8, 2018. This information is current as Promote Intrahepatic T Cell Apoptosis Marrow-Derived Cells Preferentially Intrahepatic T Cell Trapping, Whereas Bone Antigen Presentation by Liver Cells Controls Crispe Wajahat Z. Mehal, Francesco Azzaroli and I. Nicholas http://www.jimmunol.org/content/167/2/667 doi: 10.4049/jimmunol.167.2.667 2001; 167:667-673; ; J Immunol References http://www.jimmunol.org/content/167/2/667.full#ref-list-1 , 11 of which you can access for free at: cites 39 articles This article average * 4 weeks from acceptance to publication Fast Publication! Every submission reviewed by practicing scientists No Triage! from submission to initial decision Rapid Reviews! 30 days* Submit online. ? The JI Why Subscription http://jimmunol.org/subscription is online at: The Journal of Immunology Information about subscribing to Permissions http://www.aai.org/About/Publications/JI/copyright.html Submit copyright permission requests at: Email Alerts http://jimmunol.org/alerts Receive free email-alerts when new articles cite this article. Sign up at: Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists All rights reserved. Copyright © 2001 by The American Association of 1451 Rockville Pike, Suite 650, Rockville, MD 20852 The American Association of Immunologists, Inc., is published twice each month by The Journal of Immunology by guest on April 8, 2018 http://www.jimmunol.org/ Downloaded from by guest on April 8, 2018 http://www.jimmunol.org/ Downloaded from

Transcript of Antigen Presentation by Liver Cells Controls Intrahepatic T Cell ...

Page 1: Antigen Presentation by Liver Cells Controls Intrahepatic T Cell ...

of April 8, 2018.This information is current as

Promote Intrahepatic T Cell ApoptosisMarrow-Derived Cells PreferentiallyIntrahepatic T Cell Trapping, Whereas Bone Antigen Presentation by Liver Cells Controls

CrispeWajahat Z. Mehal, Francesco Azzaroli and I. Nicholas

http://www.jimmunol.org/content/167/2/667doi: 10.4049/jimmunol.167.2.667

2001; 167:667-673; ;J Immunol 

Referenceshttp://www.jimmunol.org/content/167/2/667.full#ref-list-1

, 11 of which you can access for free at: cites 39 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists All rights reserved.Copyright © 2001 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on April 8, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

by guest on April 8, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Page 2: Antigen Presentation by Liver Cells Controls Intrahepatic T Cell ...

Antigen Presentation by Liver Cells Controls Intrahepatic TCell Trapping, Whereas Bone Marrow-Derived CellsPreferentially Promote Intrahepatic T Cell Apoptosis1

Wajahat Z. Mehal,* † Francesco Azzaroli,† and I. Nicholas Crispe2*

Systemic activation and proliferation of CD81 T cells result in T cell accumulation in the liver, associated with T cell apoptosisand liver injury. However, the role of Ag and APC in such accumulation is not clear. Bone marrow chimeras were constructedto allow Ag presentation in all tissues or alternatively to restrict presentation to either bone marrow-derived or non-bone marrow-derived cells. OVA-specific CD81 T cells were introduced by adoptive transfer and then activated using peptide, which resultedin clonal expansion followed by deletion. Ag presentation by liver non-bone marrow-derived cells was responsible for most of theaccumulation of activated CD81 T cells. In contrast, Ag presentation by bone marrow-derived cells resulted in less accumulationof T cells in the liver, but a higher frequency of apoptotic cells within the intrahepatic T cell population. In unmodified TCR-transgenic mice, Ag-induced T cell deletion and intrahepatic accumulation of CD81 T cells result in hepatocyte damage, with therelease of aminotransaminases. Our experiments show that such liver injury may occur in the absence of Ag presentation by thehepatocytes themselves, arguing for an indirect mechanism of liver damage.The Journal of Immunology,2001, 167: 667–673.

A ctivated CD81 T cells undergo massive clonal expansion,redistribution via the blood to nonlymphoid organs, andfinally apoptosis. Changes in adhesion molecules allow the

mobilization of activated effector cells from lymphoid organs into theblood, with subsequent flow through all vascular beds. Localization tosites of inflammation depends on up-regulation of adhesion moleculeson vascular endothelium, which results in lymphocyte rolling, fol-lowed by firm adhesion to the endothelium, mediated by integrinssuch as ICAM-1 (1, 2). In parallel, the T cell growth factor IL-2promotes the expression of the proapoptotic factor, Fas ligand (FasL3;CD95L), and the disappearance of the caspase-8 antagonist, Fas-likeIL-1-converting enzyme-inhibitory protein (3). These changes predis-pose the activated T cells to apoptosis through Fas-FasL interactions.The fate of those activated CD81 T cells that do not localize to aninflamed target site is controversial. The expression of FasL has beenreported in a number of nonlymphoid tissues (4), and among these theliver is distinctive in expressing a high density of ICAM-1 (5, 6). It istherefore natural to consider the possibility that the liver has the po-tential both to trap and to kill activated circulating T cells.

We and others have reported the accumulation and apoptosis ofactivated CD81 T cells in the liver during systemic immune re-sponses (7–10). ICAM-1 appears to be important in this process.When a mixture of resting and activated T cells were perfusedthrough a mouse liver, the activated CD81 T were selectively re-

tained (11), but this retention was compromised in an ICAM-1-deficient liver. Most of the retained CD81 T cells were in contactwith Kupffer cells, a highly mobile population of hepatic macro-phages. By 14 h, a proportion of the trapped CD81 T cells hadbegun to undergo apoptosis. These findings support the idea thatthe liver is a trap for activated CD81 T cells and that such trappingresults in their apoptosis.

In many experimental and physiological immune responses, Agpresented on liver cells leads to tolerance, which may be linked to Tcell apoptosis (12–14). To explain the tolerogenic property of liverAgs, we propose the hypothesis that recognition of Ag on the livertissue (i.e., on sinusoidal endothelium and/or hepatocytes) promotes Tcell trapping and apoptotic death. This hypothesis predicts that thecapacity of hepatic non-bone marrow-derived cells to present Ag willcontrol the trapping and apoptosis of CD81 T cells in the liver andthus regulate CD81 T cell removal from the circulating pool.

To test these predictions, we developed a model in which naiveCD81 T cells from the OT-1 transgenic mouse, expressing a TCRspecific for the OVA peptide SIINFEKL, were adoptively trans-ferred into chimeric mice which had received lethal irradiationfollowed by reconstitution with donor bone marrow. These chi-meras were constructed so that Ag presentation could occur on allcells, or alternatively was restricted either to bone marrow-derivedcells, or to non-bone marrow-derived cells. Systemic activation ofCD81 T cells resulted in accumulation and apoptosis of activatingCD81 T cells in the liver. The ability of non-bone marrow-derivedcells to present Ag was essential for the hepatic accumulation ofCD81 T cells. Chimeras in which these tissues were unable topresent Ag accumulated fewer T cells in the liver, and this wasassociated with impaired deletion of activated CD81 T cells fromthe spleen. This supports a role for the liver in systemic CD81 Tcell homeostasis.

Immune mediated liver injury occurs in a wide variety of sys-temic immune responses (15, 16). Such injury of hepatocytes mayoccur by classical, MHC-restricted recognition of specific anti-genic peptides on hepatocytes, e.g., during immunopathology as-sociated with the CD81 T cell response to hepatitis B virus Ags(17). It has also been proposed that physiological T cell activation

Sections of *Immunobiology and†Digestive Diseases, Yale University School ofMedicine, New Haven, CT 06510

Received for publication December 6, 2000. Accepted for publication May 3, 2001.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby markedadvertisementin accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This work was supported by National Institutes of Health Grant AI37554 (to I.N.C.)and by a Howard Hughes Postdoctoral Fellowship (to W.Z.M.).2 Address correspondence and reprint requests to Dr. I. Nicholas Crispe, The DavidH. Smith Center for Vaccine Biology and Immunology, The Aab Institute for Bio-medical Research. University of Rochester, 601 Elmwood Avenue, Rochester, NY14620. E-mail address: [email protected] Abbreviations used in this paper: FasL, Fas ligand; AST, aspartate aminotransami-nase; ALT, alanine aminotransaminase.

Copyright © 2001 by The American Association of Immunologists 0022-1767/01/$02.00

by guest on April 8, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Page 3: Antigen Presentation by Liver Cells Controls Intrahepatic T Cell ...

can result in hepatic damage simply because of the presence ofactivated cytotoxic cells T cells in the liver, a mechanism termedcollateral damage (18). In the present series of experiments, chi-meric mice in which the hepatocytes were unable to support Agrecognition nevertheless showed evidence of liver damage associ-ated with an activated CD81 T cell influx. This study thus providesevidence for the collateral damage mechanism.

Materials and MethodsAnimals

Nontransgenic C57BL/6J and H2-Kbm1 mice (on the C57BL/6J back-ground) and C57BL/6 mice of the CD45.1 allotype were purchased fromThe Jackson Laboratory (Bar Harbor, ME). A colony of OT-1-transgenicmice was maintained on the C57BL/6J background. All animals werehoused in a specific pathogen-free environment, in accordance with insti-tutional guidelines for animal care.

To construct chimeras, 6- to 8-wk-old recipient mice were irradiatedwith 13 Gy (1300 rad) as a single dose from a cesium irradiator. Mature Tcells were removed from donor bone marrow cell suspensions by comple-ment-mediated lysis of T cells, and between 10 and 153 106 T cell-depleted bone marrow cells were injected i.v. within 4 h of irradiation. Allchimeras were allowed to reconstitute for 2 mo before further experimen-tation. Three types of chimeras were made: C57BL/6J3C57BL/6J; H2-Kbm13C57BL/6J; and C57BL/6J3H2-Kbm1.

Adoptive transfer and in vivo activation

Donor OT-1 mice were killed by CO2 narcosis, and a cell suspension wasobtained by mechanical homogenization of axillary and abdominal lymphnodes. RBC were removed by density gradient separation (LymphocyteSeparation Medium; ICN Biomedicals, Aurora, OH). MHC class II-posi-tive dendritic cells and B cells were removed using a primary Ab (clone212.A1 specific for MHC class II molecules, and clone 2.4-G2 specific forFcRs). Magnetic beads coated with secondary Abs were used to remove thecells coated with primary Ab. CD41 T cells were removed by magneticbeads directly coupled to anti-CD41 Ab (clone MT30). A suspension of5 3 106 lymphocytes, containing.95% CD81 T cells, was injected i.v.into each recipient.

OT-1 T cells were activated after adoptive transfer by daily i.p. injec-tions of 25 nM SIINFEKL peptide for 3 days starting on day 2, as previ-ously described (9). Control mice received saline. To control for any livertoxicity of the SIINFEKL peptide, chimeras without OT-1 cells receivedthe same dosage of the peptide.

Cell isolation, staining, and flow cytometric analysis

At days 3, 5, and 7 after the first peptide or saline injections, mice wereanesthetized with methoxyflurane (Schering-Plough Animal Health, Ken-ilworth, NJ), and exsanguinated by cutting the abdominal aorta and venacava. Blood was collected from the abdominal cavity using a heparinized1-ml syringe, and the serum was assayed for the enzymes aspartate ami-notransminase (AST) and alanine aminotransaminase (ALT) using a mul-tichannel analyzer.

Intrahepatic lymphocytes were isolated by perfusion of the liver withdigestion buffer consisting of Bruff’s medium containing 0.02% collage-nase IV (Sigma, St. Louis, MO), 0.002% DNase I (Sigma), and 5% FCS.The digestion buffer was infused into the portal vein using a 5-ml syringeand a 21-gauge needle during 1–2 min. Care was taken to minimize injec-tion of air bubbles into the portal vein, and blanching of the whole liver wasused as an indicator of adequate perfusion. After perfusion, the liver wasdissected out of the abdominal cavity and homogenized by forcing througha fine metal strainer. The homogenized liver was incubated with 10 mldigestion buffer at 37°C for 30 min in a shaking water bath. The enzymat-ically digested liver cell suspension was centrifuged at 103 g for 3 min at4°C to remove hepatocytes and cell clumps. The supernatant was thencentrifuged at 1203 g for 10 min to obtain a pellet of cells depleted ofhepatocytes. The volume of the pellet was typically 0.3–0.5 ml, and it wassuspended with Bruff’s medium to a final volume of 1 ml, before beingmixed with 4 ml 30% metrizamide in Bruff’s medium. This procedureresulted in 5 ml cell suspension in 24% metrizamide, which was layeredunder 1 ml serum-free Bruff’s medium and centrifuged at 15003 g for 20min at 4°C in 15-ml conical centrifuge tubes (Falcon, Franklin Lake, NJ).The cells at the interface were collected, washed with PBS, and countedbefore analysis using a FACS.

As a control to discriminate liver-specific effects on T cell accumulationand apoptosis from generic properties of nonlymphoid organs, kidney Tcells were analyzed in parallel. Kidneys were cut into 2- to 3-mm slices,

homogenized by forcing through a metal strainer, and then digested withcollagenase. Renal lymphocytes were isolated using metrizamide, as de-scribed above. The spleen was dissected, homogenized and RBC werelysed using RBC lysing buffer (Sigma). Lymphocytes were washed andcounted before staining for FACS analysis. Sections of liver and kidneywere cut and fixed in 1% paraformaldehyde in PBS for histological anal-ysis before homogenization.

Cells were adjusted to 23 107/ml in staining buffer (saline with 1%bovine albumin). Fifty microliters of the cell suspension were incubatedwith Ab on ice for 30 min, washed with staining buffer, and fixed with 2%paraformaldehyde. FACS data were acquired using a FACSCalibur flowcytometer (BD Biosciences, San Jose, CA), set to acquire all events. TheAbs used for staining were TCRab (clone H57-597), CD45.1 (clone A20),CD45.2 (clone 104), L-selectin (clone MEL-14), LFA-1 (clone 2D7), Va2(clone B20.1), and CD8 (clone 53-6.7). For TUNEL staining of cell sus-pensions and tissue sections, the In Situ Cell Death Detection Kit-Fluores-cein (Boehringer Mannheim, Indianapolis, IN) was used according to themanufacturer’s instructions. FACS data were analyzed using CellQuestsoftware (BD Biosciences).

ResultsHepatic accumulation and apoptosis of CD81 T cells occureven at low precursor frequencies

Peripheral deletion and massive liver accumulation and apoptosisof the responding cells were previously reported using TCR-trans-genic mice (7, 9), leaving open the possibility that these effectsmight be limited to mice with a very high frequency of respondingT cells. To test this and to establish a system to determine the roleof hepatic Ag presentation in this phenomenon, we adoptivelytransferred 53 106 CD81 T cells from OT-1 mice on theC57BL/6J background into unmanipulated nontransgenicC57BL/6J mice (19). The OT-1 mice are transgenic for a TCR thatrecognizes the 8-mer OVA peptide, SIINFEKL, in associationwith the MHC molecule H2-Kb. In all of these experiments, theadoptively transferred T cells were CD45.1/CD45.2 heterozygous,permitting their identification by FACS.

After adoptive transfer, the OT-1 cells were present at low fre-quency (1% or less) in the spleen, liver, and kidney. Daily i.p.injections of saline did not alter the total numbers of lymphocytesor the percentage of OT-1 cells during 7 days (Fig. 1). Daily in-jections of 250ml 100 mM SIINFEKL peptide (i.e., 25 nM pep-tide) beginning on day 2 after adoptive transfer resulted in nosignificant change in cell numbers in the spleen or kidney, but an;7-fold increase in the number of intrahepatic lymphocytes. Thepercentage of OT-1 cells increased in all three organs after peptideadministration, but the largest increase was in the liver at 25-fold,compared with the spleen and kidney where the increase was 7-and 10-fold, respectively (Fig. 1). The number of OT-1 cells ineach organ was calculated by multiplying the total number of lym-phocytes by the percentage of OT-1 cells. The maximum numberof OT-1 cells in the spleen was at day 3 after peptide injection, at12.4 6 3.8 3 106, decreasing to 5.26 3.1 3 106 on day 5,whereas maximum accumulation in the liver was later, with 6.362.4 3 106 OT-1 cells at day 5. By comparison the kidney con-tained relatively few OT-1 cells, 7.46 1.83 104. In all, there wasa 170-fold increase in the total number of OT-1 cells in the liver,compared with 13- and 24-fold in the spleen and kidney,respectively.

Thus, the behavior of OT-1 T cells was very similar in thespleen and the kidney. Clonal expansion on day 3 was followed byloss of the expanded cells over days 5 and 7. In contrast, a muchlarger expansion of the OT-1 cell number in the liver occurred withdelayed kinetics, at day 5. The data show that, although nonquan-titative histological analysis reveal apoptotic CD81 T cells atmany tissue sites including the liver, kidneys, lungs, and intestine(20), the liver is quite distinctive in both the kinetics and the mag-nitude of activated CD81 T cell accumulation.

668 Ag CONTROLS T CELL TRAPPING IN LIVER

by guest on April 8, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Page 4: Antigen Presentation by Liver Cells Controls Intrahepatic T Cell ...

A key finding in earlier studies was the high frequency of ap-optotic CD81 T cells accumulating in the liver (7–10). This wasconfirmed in the present study, in which the percentage of OT-1cells which were TUNEL positive in the spleen, liver, and kidneywere 8.46 2.7%, 18.36 5.2% and 9.66 4.3%, respectively. Inthe PBS-injected control chimeras, there were too few OT-1 cellsto obtain TUNEL information.

Ag presentation by non-bone marrow-derived cells enhancesintraheptic accumulation of activated CD81 T cells

The OT-1-transgenic TCR recognizes the SIINFEKL peptide inassociation with H2-Kb. The H2 molecule H2-Kbm1 differs fromH2-Kb by three amino acids at positions 152, 155, and 156, andthis difference is sufficient to prevent effective presentation of theSIINFEKL peptide (21).

In the C57BL/6J (H2-Kb) mice, all cells of the liver were ableto present the SIINFEKL peptide to OT-1 T cells. To study theeffect on liver accumulation of OT-1 T cells in the absence of Agpresentation by non-bone marrow-derived cells, we generated chi-meras in which C57BL/6J bone marrow was infused into B6.C-H2bm1 (H2-Kbm1) hosts (B63bm1). In these chimeras, Ag pre-sentation to OT-1 cells was possible by bone marrow-derived cellsonly. C57BL/6J into C57BL/6J (B63B6) and B6.C-H2bm1 intoC57BL/6 of the CD45.1 allotype (bm13B6.CD45.1) provided the

controls. The negative control of B6.C-H2bm1 into B6.C-H2bm1

chimeras are not suitable recipients for OT-1 cells, because theymount an alloimmune response to the H2-Kb molecule expressedon the T cells. In control experiments, OT-1 cells transferred intointact B6.C-H2bm1 hosts simply disappeared, probably due to al-lorejection. The donors and recipients for B63B6 and B63bm1were CD45.2, but the recipients for bm13B6.CD45.1 chimeraswere C57BL/6J mice with the CD45.1 allotype. This allowed anal-ysis of the percentage of donor bone marrow-derived cells(CD45.2), recipient bone marrow-derived cells (CD45.1), and do-nor OT-1 T cells (CD45.1/CD 45.2 heterozygous) (see Fig. 2).

Preliminary experiments identified day 5 as a time point whenremoval of most of the OT-1 cells from the spleen had occurred,and the liver accumulation was maximal. Fig. 2 shows clonal ex-pansion of OT-1 cells in the spleens of all three groups of chimerasat day 5 after peptide injection. OT-1 T cell expansion in B63B6and B63bm1 chimeras was expected, due to the presence ofH2-Kb on bone marrow cells. Expansion in the bm13B6.CD45.1was likely due to the persistence of recipient bone marrow-derivedcells despite irradiation, and these are identifiable by FACS anal-ysis as CD45.1 positive but CD45.2 negative (Fig. 2). They con-stituted 4.36 1.4% of all the bone marrow-derived cells in thespleens of bm13B6 chimeras injected with PBS. Therefore, thefull activation of OT-1 cells in these chimeras should not be takento imply that activation was independent of bone marrow-derivedcells.

Fig. 3 show the percentage and total number of OT-1 T cells inthe livers of the three groups of chimeras after 5 days of saline orpeptide injection. Peptide injection resulted in accumulation ofOT-1 cells in the livers of mice in all three groups of chimeras.B63B6 and bm13B6.CD45.1 chimeras were able to presentSIINFEKL peptide on non-bone marrow-derived cells, and thenumber of OT-1 cells were similar to those seen in normalC57BL/6 mice. These numbers were 6.43 106 6 1.73 106 in theB63B6 and 5.93 106 6 1.1 3 106 in the bm13B6 chimeras,and these numbers were not significantly different by an unpairedt test (p 5 0.5). In B63bm1 chimeras, hepatocytes and endothe-lial cells were unable to present Ag, and the numbers of OT-1 cellsin the livers of these chimeras was 2.43 106 6 0.9 3 106, whichis significantly different from both the value in B63B6 chimeras( p 5 0.002) and the value in bm13B6 chimeras (p 5 0.001).There were no significant differences among the three groups of

FIGURE 1. TCR-transgenic T cell dynamics in vivo. The percentageand total number of OT-1 cells in the spleen, liver, and kidneys of B6 miceafter PBS (M) or peptide (o) injections. There was an increase in the OT-1population after peptide injection with a peak at day 3 and a rapid declineby day 7 (AandD). A greater increase was evident in the liver, with a peakat a later time point, day 5 (Band E). The kinetics in the kidney weresimilar to the spleen, but of a lower magnitude (CandF). The total num-bers of OT-1 cells (106) follow the same kinetics as the percentages ofOT-1 cells. The very large increase in cell number the liver was due to a7-fold increase in total liver lymphocyte count. The total numbers of lym-phocytes in the spleen and kidney did not change significantly after peptideinjection.

FIGURE 2. T cell expansion in all chimeras. Clonal expansion of OT-1cells (CD45.1/CD45.2 double positive) in the spleens of all three groups ofchimeras at day 5 after peptide injection. The grafted bone marrow cellswere CD45.2 positive in all chimeras, and cells derived from them arevisible in the upper left quadrant. The bone marrow recipients for chimerabm13B6 were CD45.1 positive, and persisting recipient bone marrow-derived cells are visible in the lower right hand quadrants ofC andD. FL2,Fluorescence.

669The Journal of Immunology

by guest on April 8, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Page 5: Antigen Presentation by Liver Cells Controls Intrahepatic T Cell ...

chimeras in the number of OT-1 cells in the kidney after peptideinjection (Fig. 3,E andF, p . 0.8 in all cases).

Ag presentation by non-bone marrow-derived cells enhancesperipheral deletion of OT-1 cells from the spleen

The greater liver accumulation of OT-1 cells in B63B6 chimerasrelative to B63bm1 was associated with lower percentages andtotal numbers of OT-1 cells in the spleen of B63B6 chimeras(Fig. 3, C andD). Thus, the B63B6 chimeras contained 3.223106 6 1.78 3 106 OT-1 cells, whereas the B63bm1 chimerascontained 10.53 106 6 2.0 3 106 OT-1 T cells, a difference that

was significant based on an unpairedt test (p 5 0.001). This sug-gests that the removal of activated OT-1 cells from the spleen wascompromised because of the reduced accumulation of OT-1 cellsin the liver of the B63bm1 chimeras.

The B63B6 and B63bm1 chimeras do not differ only in Agpresentation by hepatocytes and endothelial cells, because non-bone marrow-derived cells of other organs in the B63bm1 chi-meras would also be unable to present the SIINFEKL peptide toOT-1 cells. To study the effect of Ag presentation by non-bonemarrow-derived cells in another organ, the percentage and totalnumber of OT-1 cells in the kidneys were determined in chimerasof all three groups. There was an;15- to 24-fold increase in thenumber of OT-1 cells in the kidney in peptide-injected mice com-pared with PBS-injected controls (from;2 3 103 to 40 3 103).There were, however, no significant differences between differentgroups of chimeras in the total number of OT-1 cells in the kid-neys. The kidneys of B63B6 chimeras contained 483 103 618 3 103 OT-1 cells, the kidneys of B63bm1 chimeras contained49 3 103 6 21 3 103 OT-1 cells, and the kidneys of bm13B6chimeras contained 453 103 6 21 3 103 OT-1 cells. None ofthese differences was significant by an unpairedt test (p . 0.8 inevery case).

Thus, in the absence of an inflammatory stimulus, the presenceof Ag on tissue cells did not enhance the localization of activatedT cells to the kidney. An experiment conducted in transgenic micethat expressed hepatitis B Ags in many tissues supports this inter-pretation, because the introduction of Ag-specific CTL into thesemice resulted in immunopathology only in the liver (22). The livermay be unique in its capacity to allow such interactions, becauseboth Kupffer cells and sinusoidal endothelial cells express a highresting level of ICAM-1 (5, 6). In our mice, the systemic vascu-lature was not inflamed, and based on this argument plus the kid-ney data, we would argue that the lack of Ag presentation bynon-bone marrow-derived cells in B63bm1 chimeras most likelydid not affect the localization of OT-1 cells to organs other than theliver.

Increased apoptosis of activated CD81 T cells retained in liverswith Ag presentation limited to bone marrow-derived cells

The percentage of TUNEL-positive OT-1 cells was consistentlyhigher in the liver, compared with the spleen and kidney, in allthree groups of chimeras (Fig. 4). This percentage was similar tothat observed in nonchimeric mice (7–10). The total number ofTUNEL-positive OT-1 cells was, however, lower in B63bm1 chi-meras (0.56 0.3 3 106) compared with chimeras B63B6 andbm13B6.CD45. (1.26 0.49, 1.16 0.4 3 106). This differencewas because there were many fewer OT-1 cells in the livers ofB63bm1 chimeras than B63B6 or bm13B6.CD45 (Fig. 3).

FIGURE 3. Ag presentation controls liver accumulation. Percentagesand total numbers of OT-1 cells in the spleens, livers and kidneys ofB63B6, bm13B6, and B63bm1 chimeras after PBS (M) or peptideinjections (o). There were significantly more OT-1 cells in the spleens ofchimera B63bm1 (AandB), and this was associated with fewer OT-1 cellsin the liver (CandD). The percentage and numbers of OT-1 cells in thekidney were not significantly different between the three groups of chi-meras (Eand F). Data are percentage of OT-1 cells (A, C, and E) orabsolute number in millions (B, D,andF).

FIGURE 4. T cell apoptosis in chimeras. Per-centages of OT-1 cells that were TUNEL positive(1ve) in the spleens, livers, and kidneys of chimerasB63B6, bm13B6, and B63bm1 after peptide in-jections. The frequency of TUNEL-positive cellswas higher for OT-1 cells in the livers of all threegroups of chimeras, compared with the spleen andkidney. Comparison between the three groups ofchimeras revealed more TUNEL-positive cells in theliver in B63bm1 chimeras. This difference was sig-nificant across four experiments (p5 0.04). FL1,Fluorescence.

670 Ag CONTROLS T CELL TRAPPING IN LIVER

by guest on April 8, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Page 6: Antigen Presentation by Liver Cells Controls Intrahepatic T Cell ...

However, although there was significant variation in the percent-age of TUNEL-positive OT-1 cells between experiments, the meanpercentage of TUNEL-positive cells was elevated 1.6-fold (range,1.1–2.0-fold) in the B63bm1 chimeras, relative to the B63B6chimeras. This increase was statistically significant (pairedt test,n 5 4 experiments,p 5 0.04).

Immune mediated hepatic injury occurs in the absence of Agpresentation by hepatocytes

The B63bm1 chimeras allowed us to test the requirement forhepatocyte Ag presentation in the induction of hepatocyte injuryby CD81 T cells. Serum aminotransaminases were elevated in allthree groups of chimeras after peptide injection (Table I). Theaminotransaminase levels were at least 10 times greater than in thePBS injected controls, and this elevation was statistically signifi-cant by an unpairedt test (p 5 0.001). The serum aminotransami-nases in B63bm1 chimeras after peptide injection appeared to betwofold lower than in the B63B6 and bm13B6 chimeras, anddespite the wide variation in individual values, this difference wasstatistically significant (p 5 0.02). Injection of peptide into chi-meras that had not received OT-1 cells did not result in significantelevation of aminotransaminases (AST 346 10.5 U/L, ALT 2367.2 U/L). Fig. 5 shows liver histology in chimeras injected withPBS and SIINFEKL peptide. In all three groups of chimeras, therewas a mononuclear infiltrate around the portal tract, around andcentral veins, and in liver lobules. This was associated with theeosinophilic bodies that are characteristic of hepatocyte apoptosis.

To determine whether the elevated aminotransaminase levelswere associated with hepatocyte apoptosis, liver sections werestained using the TUNEL technique (Figs. 5 and 6). There were noTUNEL-positive nuclei in the livers of PBS-injected mice (Fig. 5,b, f, and j). However, TUNEL-positive nuclei were evident in thelivers of all three groups of chimeras injected with peptide (Fig. 5,d, h,andl). The majority of TUNEL-positive nuclei were of small

mononuclear cells, but TUNEL-positive hepatocyte nuclei werealso present (Fig. 6,b, d, and f). Identical TUNEL staining ofkidney tissues from the same chimeras revealed few TUNEL-pos-itive nuclei (Fig. 7,b, d, and f). This is consistent with publisheddata showing CD81 T cell apoptosis in the kidney (20), but ourquantitative analysis of OT-1 T cells shows that in absolute num-bers very few OT-1 cells are undergoing apoptosis in this organ.

DiscussionThe adhesion of activated T lymphocytes to vascular endotheliumnormally depends on selectin-addressin-mediated rolling, whichslows the cells down, then firm adhesion based on the increasedexpression of endothelial adhesion molecules, which occurs as aresult of local inflammation (1, 2). However, this does not holdtrue for the liver, in which slow, intermittent blood flow through ananastomosing system of sinusoids is accompanied by the consti-tutive expression of adhesion molecules (5, 6, 25–27). The normalliver retains activated T cells introduced via the portal vein, andthis retention is mediated primarily by ICAM-1, which is presentboth on liver sinusoidal endothelium and on Kupffer cells (11).Hepatic retention of activated T cells has a strong preference forCD81 T cells and is associated with apoptosis of a proportion ofthe retained cells. The liver has a large blood flow and occupies aunique position in the body, because all lymphocytes leaving boththe spleen and the intestine pass through the liver before reachingthe systemic circulation. This plus the unique vasculature of theliver raise the question whether intrahepatic T cell trapping andapoptosis have any implications for either gastrointestinal or sys-temic T cell responses.

It is not possible to test the effect of liver T cell retention on asystemic immune response by removing or bypassing the liver. Wetherefore chose to determine whether the capacity of the liver topresent Ag regulates hepatic trapping and apoptosis of activatedCD81 T cells. If such regulation were present, it would allow us

Table I. Elevation of hepatocyte aminotransaminase enzymes in peptide-injected chimerasa

Chimeras

PBS-Injected Peptide-Injected

AST ALT AST ALT

B63B6 226 9 20 6 8 3216 58 2866 105bm13B6 276 11 266 9 2436 77 2156 105B63bm1 186 9 19 6 9 2176 95 1256 85

a Injection of OT-1 T cells plus antigenic peptide resulted in liver damage in all groups of chimeras. In particular, there wasdamage in B63bm1 chimeras, in which the hepatocytes were unable to present Ag. Thus, the mechanism of damage was notdependent on specific CTL-mediated killing of the hepatocytes but was indirect. AST and ALT are in units per liter.

FIGURE 5. T cell apoptosis in situ in livers. Light mi-croscopy with TUNEL staining of liver tissues from thethree types of chimeras after PBS and peptide injection.b, d, f, h, j,and l, TUNEL-positive nuclei in green. Thelivers of all three chimeras after peptide injection containnumerous TUNEL-positive cells. Magnification,360.

671The Journal of Immunology

by guest on April 8, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Page 7: Antigen Presentation by Liver Cells Controls Intrahepatic T Cell ...

to address whether there were any consequences for the behaviorof T cells during a systemic CD81 T cell response. Our experi-mental approach to this issue was to construct three sets of radi-ation bone marrow chimeras, which allows us to determinewhether the presentation of the SIINFEKL peptide on bone mar-row-derived cells, non-bone marrow-derived cells or both contrib-utes to intrahepatic trapping of CD81 T cells. The three maincellular components of the liver are the hepatocytes, the sinusoidalendothelial cells, and the bone marrow-derived Kupffer cells. Inaddition, there are bone marrow-derived intrahepatic lymphocytes,which are unlikely to play a key role in Ag presentation, and den-dritic cells, which could play such a role. Previous studies haveshown that isolated hepatocytes (28, 29), sinusoidal endothelialcells (30, 31), and Kupffer cells (31, 32) are all competent topresent Ag in vitro. We therefore base our interpretation on theextrapolation that all of these cell types can present antigenic pep-tide in vivo.

The data in Fig. 3 show that Ag presentation by the non-bonemarrow-derived cells of the liver is important in the trapping ofCD81 T cells, whereas our previous studies implicate ICAM-1 andby implication its main ligand, LFA-1. In the intact liver, these twosystems are likely to work together. It seems unlikely that thebinding affinity of Ag-specific TCR-mediated recognition wouldcontribute directly to adhesion between T cells and either endo-thelium or hepatocytes. Instead, such interactions would come intoplay after an activated T cell has formed an initial adhesion toendothelium or hepatocytes, most likely through a nonspecific ad-hesion molecule such as ICAM-1. The effect of TCR ligation onLFA-1/ICAM-1 interactions has been studied extensively in vitro.Stimulation of T cells by receptor cross-linking, or phorbol esters,increases both the level of LFA-1 expression and the affinity ofLFA-1 for ICAM-1 (33, 34). Thus, in activated CD81 T cellstraversing liver sinusoids, Ag recognition would be expected toincrease the avidity of LFA-1, promoting firm adhesion and theretention of the T cell in the liver.

The sinusoidal endothelium resembles inflamed vascular endo-thelium, because it constitutively expresses a high level of

ICAM-1, but unlike inflamed vascular endothelium there is noexpression of the B7-1 and B7-2 costimulatory molecules. Thishigh level of expression of ICAM-1 on sinusoidal endotheliummay contribute to the apoptosis of CD81 T cell seen in the liver.Stimulation of naive and memory CD81, but not CD41, T cells byanti-TCR Ab and ICAM-1 has been shown to result in both pro-liferation and apoptosis of the responding T cells (35). This con-trasts with anti-TCR and B7-1 stimulation, which results in a sim-ilar level of proliferation but reduced apoptosis. It is of interest thatthe costimulatory effect of ICAM-1 preferentially acts on CD81 Tcells (35–37). Thus, it is possible that ICAM-1 is the basis of bothintrahepatic CD81 T cell retention and CD81 T cell apoptosis.However, other potential proapoptotic mechanisms are present inthe liver. Our data suggest that bone marrow-derived Kupffer cellsmay be important in the induction of apoptosis, and in addition toICAM-1 these cells express several proapoptotic molecules, in-cluding membrane bound FasL and soluble TNF-a, whereas sinu-soidal endothelium expresses galectin-1 (38–41). Furthermore,systemic T cell activation promotes FasL expression in severaltissues, including liver (4). The contribution of these various celltypes and proapoptotic molecules is not yet defined.

Massive liver retention and apoptosis of activated CD81 T cellsis accompanied by pathology in the liver. In the original study ofHuang et al. (7), we observed histological evidence of liver dam-age, and this was subsequently confirmed in another experimentalmodel by elevation in serum aminotransaminases (9). In these ex-periments, CD81 T cell activation was induced by the injection ofantigenic peptide in TCR-transgenic mice, and such peptides mayhave been presented by hepatocytes so that hepatocytes would betargets for the cytotoxic effector function of CD81 T cells trappedin the liver. Hepatocyte injury by CD81 T cells after MHC-re-stricted recognition of hepatocyte Ags occurs, and it requires both

FIGURE 7. Minimal intrarenal apoptosis. Light microscopy withTUNEL staining of kidney tissues from the three chimeras after peptideinjection showing few scattered green TUNEL-positive cells. Magnifica-tion, 360.

FIGURE 6. Lymphoid cell and hepatocyte apoptosis. High power lightmicroscopy of liver tissues from the three types of chimera after peptideinjection showing green TUNEL-positive lymphocytes (arrowheads) andhepatocytes (arrows). Magnification,3200.

672 Ag CONTROLS T CELL TRAPPING IN LIVER

by guest on April 8, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Page 8: Antigen Presentation by Liver Cells Controls Intrahepatic T Cell ...

perforin and CD95L cytotoxic mechanisms (23). Because acti-vated CD81 T cells express FasL and because hepatocytes can bekilled by Fas ligation (24), it has been hypothesized that T cellactivation can result in hepatocyte injury in the absence of hepa-tocyte Ag presentation, a mechanism termed collateral damage(18). In the present study, the hepatitis induced by systemic CD81

T cell activation when hepatocytes are unable to present theSIINFEKL peptide proves that activation of a small percentage ofT cells can result in such “collateral” liver injury. The relativecontributions of direct and collateral damage are very difficult todetermine from these limited data. The fact that the mean serumALT was ;2-fold lower in the B63bm1 chimeras could be takento imply that about one-half of the liver damage was direct,whereas one-half was collateral. However, individual variation be-tween experimental mice makes this difference hard to evaluate.Even if the difference proves to be real in a more extensive seriesof experiments, it could simply be because the absolute number ofOT-1 cells in the liver is less in these chimeras. Whatever therelative contributions of the two mechanism in intact mice, thepresent data provide clear evidence for a novel mechanism of col-lateral liver injury, which may be involved in some disease states.

AcknowledgmentsWe thank the Yale Liver Center for the use of its morphology core facility.

References1. Butcher, E. C., and L. J. Picker. 1996. Lymphocyte homing and homeostasis.

Science 272:60.2. Salmi, M., and S. Jalkanen. 1997. How do lymphocytes know where to go?

Current concepts and enigmas of lymphocyte homing.Adv. Immunol. 64:139.3. Rafaeli, Y., L. Van Parijs, C. A. London, J. Tschopp, and A. Abbas. 1998. Bio-

chemical mechanisms of IL-2-regulated Fas-mediated T cell apoptosis.Immunity8:615.

4. Bonfoco, E., P. M. Stuart, T. Brunner, T. Lin, T. S. Griffiths, Y. Gao,H. H. Nakajima, T. A. Ferguson, and D. R. Green. 1998. Inducible nonlymphoidexpression of Fas ligand is responsible for superantigen-induced peripheral de-letion of T cells.Immunity 9:711.

5. Scoazec, J. Y., and G. Feldmann. 1994. The cell adhesion molecules of hepaticsinusoidal endothelial cells.J. Hepatol. 20:296.

6. Scoazec, J.-Y. 1996. Adhesion molecules in normal human liver.Hepato-Gastroenterology 43:1103.

7. Huang, L., G. Soldevila, M. Leeker, R. A. Flavell, and I. N. Crispe. 1994. Theliver eliminates T cells undergoing antigen-triggered apoptosis in vivo.Immunity1:741.

8. Iwakoshi, N. N., I. Goldschneider, F. Tausche, J. P. Mordes, A. A. Rossini, andD. L. Greiner. 1998. High frequency apoptosis of recent thymic emigrants in theliver of lymphopenic diabetes-prone BioBreeding rats.J. Immunol. 160:5838.

9. Russell, J. Q., G. J. Morrissette, M. Weidner, C. Vyas, D. Aleman-Hoey, andR. C. Budd. 1998. Liver damage results preferentially from CD81 T cells trig-gered by high affinity peptide antigens.J. Exp. Med. 88:1147.

10. Belz, G. T., J. D. Altman, and P. C. Doherty. 1998. Characteristics of virus-specific CD81 T cells in the liver during the control and resolution phases ofinfluenza pneumonia.Proc. Natl. Acad. Sci. USA 95: 13812.

11. Mehal, W. Z., A. E. Juedes, and I. N. Crispe. 1999. Selective retention of acti-vated CD81 T cells by the normal liver.J. Immunol. 163:3202.

12. Calne, R. Y., R. A. Sells, J. R. Pena, D. R. Davis, P. R. Millard, B. M. Herbertson,R. M. Binns, and D. Davies. 1969. Induction of immunological tolerance byporcine liver allografts.Nature 223:472.

13. Qian, S., L. Lu, F. Fu, Y. Li, W. Li, T. E. Starzl, J. J. Fung, and A. W. Thomson.1997. Apoptosis within spontaneously accepted mouse liver allografts.J. Immu-nol. 158:4654.

14. Bertolino, P., W. R. Heath, C. L. Hardy, G. Morahan, and J. F. A. P. Miller. 1995.Peripheral deletion of autoreactive CD81 T cells in transgenic mice expressingH-2Kb in the liver.Eur. J. Immunol. 25:1932.

15. Runyon, B. A., D. R. LaBrecque, and S. Anuras. 1980. The spectrum of liverdisease in systemic lupus erythematosus: report of 33 histologically-proved casesand review of the literature.Am. J. Med. 69:187.

16. Shiffman, M. L. 1997. Immune disorders and the liver. InThe Liver and SystemicDisease,N. Gitlin, ed. Churchill Livingstone, New York, p. 147.

17. Nakamoto, Y., L. G. Guidotti, V. Pasquetto, R. D. Schreiber, and F. V. Chisari.1997. Differential target cell sensitivity to CTL-activated death pathways in hep-atitis B virus transgenic mice.J. Immunol. 158:5692.

18. Crispe, I. N., T. Dao, K. Klugewitz, W. Z. Mehal, and D. P. Metz. 2000. The liveras a site of T-cell apoptosis: graveyard, or killing field?Immunol. Rev. 174:47.

19. Kim, S. K., K. S. Schluns, and L. Lefrancois. 1999. Induction and visualizationof mucosal memory CD8 T cells following systemic virus infection.J. Immunol.163:4125.

20. Wack, A., P. Dellaporta, I. N. Crispe, and D. Kioussis. 1997. Multiple sites ofpost-activation CD81 T cell disposal.Eur. J. Immunol. 27:577.

21. Nikolic-Zugic, J., and F. R. Carbone. 1990. The effect of mutations in the MHCclass I peptide binding groove on the cytotoxic T lymphocyte recognition of theKb-restricted ovalbumin determinant.Eur. J. Immunol. 20:2431.

22. Ando, K., L. G. Guidotti, A. Cerny, T. Ishikawa, and F. V. Chisari. 1994. CTLaccess to tissue antigen is restricted in vivo.J. Immunol. 153:482.

23. Nakamoto, Y., L. G. Guidotti, V. Pasquetto, R. D. Schreiber, and F. V. Chisari.1997. Differential target cell sensitivity to CTL-activated death pathways in hep-atitis B virus transgenic mice.J. Immunol. 158:5692.

24. Ogasawara, J., R. Watanabe-Fukunaga, M. Adachi, A. Matsuzawa, T. Kasugai,Y. Kitamura, N. Itoh, T. Suda, and S. Nagata. 1993. Lethal effects of the anti-Fasantibody in mice.Nature 346:608.

25. Jones, A. L. 1996. Anatomy of the normal liver. InHepatology, a Textbook of LiverDiseases.Zakim, D., and Boyer, T. D., eds. W. B. Saunders, Philadelphia, p.1.

26. MacPhee, P. J., E. E. Schmidt, and A. C. Groom. 1995. Intermittence of bloodflow in liver sinusoids, studied by high resolution in vivo microscopy.Am. J. Physiol. 269:G692.

27. McNab, G., J. L. Reeves, M. Salmi, S. Hubscher, S. Jalkanen, and D. H. Adams.1996. Vascular adhesion protein 1 mediates binding of T cells to human hepaticendothelium.Gastroenterology 110:522.

28. Lee, Y. C., L. Lu, F. Fu, W. Li, A. W. Thomson, J. J. Fung, and S. Qian. 1999.Hepatocytes and liver nonparenchymal cells induce apoptosis in activated T cells.Transplant. Proc. 31:784.

29. Bertolino, P., M. C. Trescol-Biemont, and C. Rabourdin-Combe. 1998. Hepato-cytes induce functional activation of naive CD81 T lymphocytes, but fail topromote survival.Eur. J. Immunol. 28:221.

30. Limmer, A., J. Ohl, C. Kurts, H. G. Ljunggren, Y. Reiss, M. Groettrup,F. Momburg, B. Arnold, and P. A. Knolle. 2000. Efficient presentation of exog-enous antigen by liver endothelial cells to CD81 T cells results in antigen-specificT-cell tolerance.Nat. Med. 6:1348.

31. Lohse, A. W., P. A. Knolle, K. Bilo, A. Uhrig, C. Waldmann, M. Ibe, E. Schmitt,G. Gerken, and K. H. Meyer Zum Buschenfelde. 1996. Antigen-presenting func-tion and B7 expression of murine sinusoidal endothelial cells and Kupffer cells.Gastroenterology 110:1175.

32. Roland, C. R., L. Walp, R. M. Stack, and M. W. Flye. 1994. Outcome of Kupffercell presentation to a cloned murine Th1 lymphocyte depends on the inducibilityof nitric oxide synthase by IFN-g. J. Immunol. 153:5453.

33. Springer, T. A., M. L. Dustin, T. K. Kishimoto, and S. D. Marlin. 1987. Thelymphocyte function-associated LFA-1, CD2, and LFA-3 molecules, cell adhe-sion receptors of the immune system.Annu. Rev. Immunol. 5:223.

34. Dustin, M. L., and T. A. Springer. 1989. T-cell receptor cross-linking transientlystimulates adhesiveness through LFA-1.Nature 341:619.

35. Deeths, M. J., and M. F. Mescher. 1999. ICAM-1 and B7-1 provide similar butdistinct costimulation for CD81 T cells, while CD41 T cells are poorly costimu-lated by ICAM-1.Eur. J. Immunol. 29:45.

36. Bachmann, M. F., K. McKall-Faienza, R. Schmits, D. Bouchard, J. Beach,D. E. Speiser, T. W. Mak, and P. S. Ohashi. 1997. Distinct roles for LFA-1 andCD28 during activation of naive T cells, adhesion versus co-stimulation.Immu-nity 7:549.

37. Kuhlman, P., V. T. Moy, B. A. Lollo, and A. A. Brian. 1991. The accessoryfunction of murine intercellular adhesion molecule-1 in T lymphocyte activation:contributions of adhesion and co-activation.J. Immunol. 146:1773.

38. Muschen, M., U. Warskulat, T. Peters-Regehr, J. G. Bode, R. Kubitz, andD. Haussinger. 1999. Involvement of CD95 (Apo-1/Fas) ligand expressed by ratKupffer cells in hepatic immunoregulation.Gastroenterology 116:666.

39. Bradham, C. A., J. Plumpe, M. P. Manns, D. A. Brenner, and C. Trautwein. 1998.Mechanisms of hepatic toxicity. I. TNF-induced liver injury.Am. J. Physiol.275:G387.

40. Perillo, N. L., K. E. Pace, J. J. Seilhamer, and L. G. Baum. 1995. Apoptosis ofT cells mediated by galectin-1.Nature 378:736.

41. Lotan, R., P. N. Belloni, R. J. Tressler, D. Lotan, X. C. Xu, and G. L. Nicholson.1994. Expression of galectins on microvessel endothelial cells and their involve-ment in tumour cell adhesion.Glycoconjugate J. 11:462.

673The Journal of Immunology

by guest on April 8, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from