Immune stimulation of hepatic fibrogenesis by CD8 cells and attenuation by transgenic interleukin-10 from hepatocytes

B

IA

RW*D

Bcshis(deccCwwtsmcCwdvsmmaaealma

Frve

GASTROENTEROLOGY 2004;127:870–882

ASIC–LIVER, PANCREAS, AND BILIARY TRACT

mmune Stimulation of Hepatic Fibrogenesis by CD8 Cells andttenuation by Transgenic Interleukin-10 From Hepatocytes

IFAAT SAFADI,* MASAYUKI OHTA,* CARLOS E. ALVAREZ,* M. ISABEL FIEL,‡ MEENA BANSAL,*AJAHAT Z. MEHAL,§ and SCOTT L. FRIEDMAN*

Division of Liver Diseases and ‡Department of Pathology, The Mount Sinai School of Medicine, New York, New York; and §Section ofigestive Diseases and Immunobiology, Yale University School of Medicine, New Haven, Connecticut

itaaagfiacItekt

adnCcaIpmcBfiMI

CFmlct

See editorial on page 997.

ackgrounds & Aims: Immunomodulatory cytokines, in-luding interleukin-10 (IL-10), may mediate hepatic fibro-is. Methods: We generated transgenic (TG) mice withepatocyte expression of rat IL-10 (rIL-10) to assess its

mpact on lymphocyte subsets and activation of hepatictellate cells following liver injury from carbon tetrachlorideCCl4) or thioacetamide (TAA). Results: Fibrosis was re-uced in the TG animals in both models, which was notxplained solely by differences in liver injury. By fluores-ence-activated cell sorter (FACS), there were less CD4� Tells in naive TG mice, and, following fibrosis induction,D4� T cells decreased only in wild-type (WT) mice,hereas increases in CD8� T cells seen in WT animalsere significantly attenuated in TG mice. Subtotal irradia-

ion diminished fibrosis equally in both WT and TG groups,uggesting that rIL-10’s antifibrotic effect was lymphocyteediated. To assess the role of lymphocytes on stellate

ell activation, either whole splenic lymphocytes, CD4�, orD8� T-cell subsets from WT animals with CCl4 fibrosisere adoptively transferred to severe combined immuno-eficiency (SCID) recipients, which led to stellate cell acti-ation and fibrogenic stimulation as assessed by expres-ion of transforming growth factor (TGF)-�1 and collagen Iessenger RNA (mRNA) and by immunoblot of �-smoothuscle actin. Moreover, serum aminotransferase levels

nd stellate cell activation mRNA were significantly highermong the CD8� T-cell recipients. Conclusions: Transgenicxpression of rIL-10 in liver leads to reduced fibrosis andlterations in liver lymphocyte subsets both in untreated

iver and following fibrosis induction. In this model, fibrosisay be a CD8� T-cell–mediated disease that is attenu-

ted by rIL-10.

ibrosis is the result of chronic injury to the liver,regardless of etiology. During the hepatic injury

esponse, hepatic stellate cells transdifferentiate, or acti-ate, into proliferative matrix-producing cells that gen-rate fibrosis.1,2 Key markers of activated stellate cells

nclude �-platelet-derived growth factor (PDGF) recep-or, matrix metalloproteinase-2 (MMP-2), intercellulardhesion molecule-1 (ICAM-1), and �-smooth musclectin (�-SMA).1,2 Stellate cell fibrogenesis reflects thectivities of profibrotic cytokines including transformingrowth factor (TGF) �13 and connective tissue growthactor (CTGF)4 and antifibrotic cytokines that includenterleukin-10 and interferon (IFN) �.5–9 An antifibroticctivity of IL-10 has also been suggested by studies ofultured stellate cells in which neutralizing antibodies toL-10 down-regulate matrix accumulation,5 by attenua-ion of fibrosis following its exogenous administration inxperimental hepatitis,9,10 and through analysis of IL-10nockout mice, which have increased fibrosis in responseo toxic injury.7,11

Recent studies have focused not only on cytokines butlso on the inflammatory cells from which they areerived, including hepatic macrophages (Kupffer cells),atural killer (NK) cells, and lymphocytes, includingD4� T helper (Th) and CD8� subsets. The lympho-yte subsets can also be broadly divided into those thatre either Th1 or Th2 predominant.12 For example,FN-�, a component of the Th1 lymphocyte cytokineathway, has potent antifibrotic activity.12,13 C57BL/6ice that exhibit an expansion of IFN-�–producing Th1

ells have comparatively minimal fibrosis, whereasALB/c mice that have a Th2 response develop severebrosis in response to carbon tetrachloride (CCl4).12–15

oreover, BALB/c mice lacking B cells, T cells, andFN-� had less fibrosis than wild-type animals.13 These

Abbreviations used in this paper: �-SMA, �-smooth muscle actin;Cl4, carbon tetrachloride; CTGF, connective tissue growth factor;ACS, fluorescence-activated cell sorter; ICAM-1, intercellular adhesionolecule-1; IL, interleukin; rIL, rat interleukin; MMP-2, matrix metal-

oproteinase-2; PDGF, platelet-derived growth factor; SCID, severeombined immunodeficiency; TAA, thioacetamide; TCR, T-cell recep-or; TG, transgenic; TGF, transforming growth factor.

© 2004 by the American Gastroenterological Association0016-5085/04/$30.00

doi:10.1053/j.gastro.2004.04.062

desihppofiaill

iberoc(gifiw

oqstifibhglsfitiacTf

ffo

w(tbvwbmCE1W1Eaatpefi

lDWoG5pt

pifiT(Kro

ufautw

September 2004 IMMUNE STIMULATION OF HEPATIC FIBROSIS 871

ata suggest that hepatic fibrosis is influenced by differ-nt responses to T-cell subsets and that Th cytokineubsets can modulate the fibrotic response to toxin-nduced liver injury. One apparent paradox is that IL-10as antifibrogenic properties, yet it is produced during arofibrogenic Th2 immune response. The reasons for thisaradox probably lie in the uncharacterized effects ofther Th2 cytokines (e.g., IL-4, 5, 9, and 13) on liverbrosis. Collectively, studies of this type have provokedsearch for genetic loci that correlate with risk of fibrosis

n mice and in humans, with the presumption that suchoci may encode for genes that modulate the immuno-ogic response.16,17

Immune mediated regulation of human liver fibrosis isncreasingly appreciated, and immunosuppression haseen identified as an important permissive stimulus. Forxample, patients with HIV infection are at acceleratedisk of fibrosis when coinfected with HCV, independentf the extent of liver injury.18–20 Similarly, progression toirrhosis in patients undergoing transplantation for HCVwho require long-term immunosuppressive therapy) isreatly accelerated compared with those with nativenfection,21,22 and a subset develop a rapidly progressivebrogenic syndrome that may lead to recurrent cirrhosisithin a year or less.23

Despite this mounting evidence of immune mediationf fibrosis, there is remarkably little known about theualitative and quantitative alterations of lymphocyteubsets during hepatic fibrogenesis or their relationshipo pro- and anti-fibrogenic cytokines. Thus, althoughndirect evidence implicates Th1 and Th2 cytokine pro-les as important determinants of fibrosis, the interplayetween regulatory cytokines and lymphocyte subsetsas not been explored. We approached this issue byenerating a transgenic (TG) mouse secreting rat inter-eukin-10 (rIL-10) in hepatocytes to assess the impact ofustained local expression of the cytokine on hepaticbrogenesis in 2 distinct animal models. Having iden-ified an antifibrotic effect of rIL-10, we next character-zed the influence of this cytokine on lymphocyte subsetsnd identified a specific reduction of CD8� T lympho-ytes. Finally, we tested the fibrogenic activity of CD8�

lymphocytes isolated from animals with liver injuryollowing their adoptive transfer into naive animals.

Materials and Methods

Animals

C57BL/6 mice (wild-type and SCID) were purchasedrom Jackson Laboratories. SCID mice were housed in a barrieracility. Animals received care according to National Institutesf Health guidelines.

Generation and Characterization of RatIL-10 TG Mice

TG mice expressing rIL-10 primarily in hepatocytesere generated by expressing the full-length rIL-10 cDNA

accession no. L02926, nucleotides 1–682) downstream of theransthyretin (TTR) promoter on a mixed C57BL/6 and C3Hackground, using an expression construct generously pro-ided by Dr. T. VanDyke, UNC Chapel Hill.24 Use of rIL-10,hich is fully active in mouse, enabled the discriminationetween endogenous and transgenic IL-10. Two F0 TG ani-als were generated and mated with female wild-type57BL/6 mice, thereby yielding 2 lines. Using a rat-specificLISA, average serum levels in the TG animals were up to000 pg/mL, whereas levels were below detectable levels in allT animals. Serum levels of endogenous mouse IL-10 (mIL-

0) were similar in both WT and TG animals. Rat IL-10LISA levels measured in homogenous samples of liver extractsnd gallbladder bile were also significantly elevated in TGnimals. Detection of rIL-10 mRNA in the liver by semiquan-itative RT-PCR was apparent in both TG lines. The same sizeroduct was also seen in the spleen and intestine, and lessxpression was observed in lungs of line 1 TG animals. Allndings in transgenic line 1 were replicated in line 2.

DNA Extraction and rIL-10 Genotyping

Genomic DNA was extracted from tail clippings fol-owing digestion with Proteinase K using Wizard GenomicNA Purification System (Promega Corporation, Madison,I) according to the manufacturer’s instructions. Oligonucle-

tide primers for rIL-10 used in PCR were: sense, 5=-AGTAA GAC CAG CAA AGG CCA TTC 3= and antisense,=-CAT TTT GAG TGT CAC GTA GGC TTC3=. Theserimers were able to distinguish between rat IL-10 cDNAransgene and mouse IL-10 genomic DNA.

Models of Hepatic Fibrosis

CCl4 (Sigma, C-5331) fibrosis was induced by intra-eritoneal (IP) injections at 0.5 �L CCl4/g body weight (1:10n corn oil) twice each week for 4 weeks. A second model ofbrosis was also used in which thioacetamide (TAA; Sigma,-8531) was administered IP at 200 mg/kg body weight

diluted in water) 3 times each week for 4 weeks. Followingetamine/Xylazine anesthesia, animals were killed, and se-

um, livers, and cells were harvested 3 days after the final dosef either TAA or CCl4.

Tissue RNA Extraction and rIL-10 mRNADetection

Total cellular RNA was extracted from target tissuessing Trizol reagent (GIBCO-BRL, Life Technologies) andollowed by DNase-I digestion. Purified RNA was then used astemplate for reverse transcription into single-strand cDNAsing the Reverse Transcription System (Promega Corpora-ion, Madison, WI). Rat IL-10 cDNA was detected by PCRith the primers listed above.

RimTfpDmac�Sa9appgc(qntiiursott

gAoAGCCG�AlACrwT

luctc

whwnlaoop

NCcsumflaCcesIaFs

fawsfpaDtsbHitn

872 SAFADI ET AL. GASTROENTEROLOGY Vol. 127, No. 3

Real-Time PCR Analysis of FibrogenicMarkers

Synthesis of cDNA was performed using 2 �g of totalNA per sample with random primers and reagents contained

n the Reverse Transcription System kit, according to theanufacturer’s protocol (Promega Corporation, Madison, WI).he reverse transcriptase product was diluted 20� in nuclease-

ree H2O, and 5 �L of each sample was loaded into 96-welllates for real-time PCR in an ABI Prism 7700 Sequenceetection System (Applied Biosystems). �-Actin and �2-acroglobulin served as internal controls, and H2O served asnegative control. Amplification reactions included oligonu-

leotide primers for each target gene, and for �-actin and2-macroglobulin, as well as platinum Taq polymerase andYBR Green DNA-binding dye. Fluorescence signals werenalyzed during each of 40 cycles (denaturation 15 seconds at5°C, annealing 15 seconds at 56°C, and extension 40 secondst 72°C). Denaturation curves of target genes and �-actin,erformed at the end of the PCR, and detection of the PCRroducts by agarose gel electrophoresis confirmed the homo-eneity of the DNA products. Relative quantification wasalculated using the comparative threshold cycle (CT) methodas described in the User Bulletin 2, ABI PRISM 7700 Se-uence Detection System). CT indicates the fractional cycleumber at which the amount of amplified target genes a fixedhreshold within the linear phase of gene amplification and isnversely related to the abundance of mRNA transcripts in thenitial sample. Mean CT of duplicate measurements would besed to calculate �CT as the difference in CT for target andeference. �CT for each sample was compared to the corre-ponding control CT and expressed as �CT. Relative quantityf product was expressed as fold-induction or repression of thearget gene compared with the control primers, according tohe formula 2��CT.

Mouse Oligonucleotide Primers forReal-Time PCR

The sequence of mouse �-Actin (control housekeepingene): forward: GAT GAG ATT GGC ATG GCT TT, reverse:GA GAA GTG GGG TGG CTT TT; �2-macroglobulin (sec-nd control housekeeping gene): forward: ATG CTG AAGAC GGG AAA AA, reverse: CGG CCA TAC TGT CATCT TA; �-smooth muscle actin (�-SMA): forward: TCC TCCTG GAG AAG AGC TAC, reverse: TAT AGG TGG TTTGT GGA TGC; TGF-�1: forward: TGC GCT TGC AGAAT TAA AA, reverse: CTG CCG TAC AAC TCC AGT GA;1() collagen: forward: GTC CCT GAA GTC AGC TGC

TA, reverse: TGG GAC AGT CCA GTT CTT CAT; �-Plate-et Derived Growth Factor receptor: forward: CCG GAA CAACA CAC CTT CT, reverse: TAT CCA TGT AGC CAC CGTA; ICAM 1: forward: ACC CAA CTG GAA GCT GTT TG,

everse: ACC CAA CTG GAA GCT GTT TG; MMP2: for-ard: ACC CAG ATG TGG CCA ACT AC, reverse: TACTT TAA GGC CCG AGC AA.

Cytokine Measurements

For the determination of serum rIL-10 and mIL-10evels, OptEIA ELISA kits (Pharmingen, San Diego, CA) weresed according to the manufacturer’s protocol. A standardurve was generated using recombinant cytokines and concen-rations of samples that were determined by a polynomialurve fit analysis.

Cell Isolation, Staining, and FlowCytometric Analysis

The spleen was homogenized, and lymphocytes wereashed and counted before staining for FACS analysis. Intra-epatic lymphocytes were isolated by perfusion of the liverith digestion buffer. After perfusion, the liver was homoge-ized and incubated at 37°C for 30 minutes. The digestediver cell suspension was centrifuged to remove hepatocytesnd cell clumps. The supernatant was then centrifuged tobtain a pellet of cells depleted of hepatocytes to a final volumef 1 mL. Lymphocytes were then isolated from this cell sus-ension using 24% metrizamide gradient separation.25

Splenic CD4� and CD8� subsets were isolated using DY-AL Biotech kits according to manufacturer’s instructions.ells were adjusted to 2 � 107/mL in staining buffer (in salineontaining 1% bovine albumin). Fifty microliters of the celluspension were incubated with antibody on ice for 30 min-tes, washed with staining buffer, and fixed with 2% parafor-aldehyde. FACS data were acquired using a FACSCalibur

ow cytometer (BD Biosciences, San Jose, CA) set to capturell events. Antibodies used for staining were mouse anti-CD4,D8, T-cell-receptor (TCR)-�/�, and TCR �/ antibodies,onjugated by fluorescein isothiocyanate (FITC), phyco-rythrin (PE), CYCHROME, and allophycocyanin (APC), re-pectively. Intracellular staining of lymphocytes by IFN-� andL-4 (FITC and PE conjugated, respectively) was performedccording to the manufacturer’s protocol (BD Biosciences).ACS data were analyzed using CellQuest software (BD Bio-ciences).

Liver Histology

The posterior one third of the liver was fixed in 10%ormalin for 24 hours and then paraffin-embedded in anutomated tissue processor. Seven-millimeter liver sectionsere cut from each animal. Sections (15 �m) were then

tained in 0.1% sirius red F3B in saturated picric acid (bothrom Sigma).26 Hematoxylin and eosin (H&E) staining waserformed for each animal. Additionally, �-smooth musclectin immunohistochemistry was performed using theAKO kit (Cat. no. U7033 EPOS, Monoclonal) according

o the manufacturer’s instructions. Knodell score was as-essed blindly by an expert hepatic pathologist (M.I.F.)ased on H&E and sirius red staining, using the modifiedistological Activity Index (HAI) criteria,27,28 incorporat-

ng semiquantitative assessment of periportal/periseptal in-erface hepatitis (0 – 4), confluent necrosis (0 – 6), focal lyticecrosis/apoptosis and focal inflammation (0 – 4), portal in-

flc5sbm

lscpmtffi

bcdw

acac(a(ws

tigfr

sarc

kcwsl

pWntmtw(tPBcSdq(rhm

qccsM

T

WTWT

Wa

b

c

d

September 2004 IMMUNE STIMULATION OF HEPATIC FIBROSIS 873

ammation (0 – 4), and architectural changes/fibrosis andirrhosis (0 – 6). Data was derived from blinded analysis ofsections from each of 10 animals in each group. Fibrosis

cores were not included in displayed data in Table 1ecause this variable was more thoroughly assessed usingorphometry, as described in the next section.

Fibrosis Quantitation

Relative fibrosis area (expressed as a percentage of totaliver area) was assessed by analyzing 36 sirius red–stained liverections per animal. Each field was acquired at 10� magnifi-ation and then analyzed using a computerized Bioquant mor-hometry system. To evaluate the relative fibrosis area, theeasured collagen area was divided by the net field area and

hen multiplied by 100. Subtraction of vascular luminal arearom the total field area yielded the final calculation of the netbrosis area.

Total Body Irradiation Model

Mice were sublethally irradiated with a single totalody dose of 700 cGy from a dual Cs source (dose rate of 62Gy/min) before induction of fibrosis using CCl4. Sublethaloses were used that suppressed lymphocyte proliferationhile maintaining the animal in an apparently healthy state.29

Adoptive Transfer of Hepatic Fibrosis

Hepatic fibrosis was induced in WT C57BL/6 malenimals using biweekly IP CCl4 injections, whereas negativeontrol animals were administered sterile culture medium IP,nd, for positive control for lymphocyte stimulation, lympho-ytes were isolated from mice immunized with ovalbuminOA) by administering OA via 2 footpad injections containingtotal 200 �g OA emulsified in incomplete Freund’s adjuvant

Sigma) at a 1:1 ratio. On day 10, mice were killed, and spleensere aseptically removed and teased into single-cell suspen-

ions.For adoptive transfer studies, lymphocytes from CCl4-

reated or control animals were transferred to severe combinedmmunodeficiency (SCID) mice from the same genetic back-round, age, and sex. Three million lymphocytes were trans-erred IP weekly from donors to respective different SCIDecipient for 4 weeks. Whole lymphocytes and CD4 and CD8

able 1. Effect of Transgenic Interleukin 10 on Knodell InflamEither Thioacetamide or Carbon Tetrachloride

Inflammatory score

T-TAA (�SD) 3.9 � 0.3 5G-TAA (�SD) 3.8 � 0.4, P � 0.34a 4T-CCI4 (�SD) 3.8 � 0.42, P � 0.31a 7

G-CCI4 (�SD) 3.7 � 0.48, P � 0.3a, 0.3c 6

T, wild type; TG, IL-10 transgenic mice; TAA, thioacetamide; CCI4, cP value using t test.P value between WT animals from both models.P value between WT and TG in each model.P value between TG animals from both models.

ubsets from 3 different donors (CCl4- treated, vehicle treated,nd ovalbumin stimulated) were separately transferred to 9ecipient groups, respectively. Seven SCID animals were in-luded in each subgroup.

For each transfer, 2 animals from each donor group wereilled, starting from day 2 postinitial injection of CCl4 orontrol medium. For each subsequent treatment, lymphocytesere isolated from animals with increasing cumulative expo-

ure to CCl4. SCID recipients were killed 7 days following theast adoptive transfer.

�-Smooth Muscle Actin Immunoblot

Immunoblot analysis of �-SMA in liver extracts waserformed as previously described,30 with modifications.

hole-liver protein extracts were prepared in liver homoge-ization buffer (50 mmol/L Tris-HCl [pH 7.6], 0.25% Tri-on-X 100, 0.15 mol/L NaCl, 10 mmol/L CaCl2, and completeini EDTA-free protease inhibitor cocktail [Roche Diagnos-

ics, Mannheim, Germany]). Then proteins (30 �g per lane)ere resolved on a 10% (wt/vol) SDS-polyacrylamide gel

Novex, Groningen, The Netherlands) under reducing condi-ions. For immunoblotting, proteins were transferred to arotran membrane (Schleicher & Schuell, Dassel, Germany).lots were incubated overnight at 4°C in a blocking bufferontaining 5% skim milk and then incubated with an anti-MA mouse monoclonal antibody (DAKO, cat. no. M0851),iluted 1:2000 for 2 hours at room temperature and, subse-uently, with peroxidase-conjugated goat anti-mouse IgGP.A.R.I.S., Compiègne, France) diluted 1:10,000 for 1 hour atoom temperature. Immunoreactivity was revealed by en-anced chemiluminescence using an ECL kit (Amersham Phar-acia Biotech, Les Ulis, France).

Statistical Analysis

Lymphocyte subsets, serum aminotransferases, fibrosisuantitation, real-time mRNA expression, and cytokine con-entrations in WT vs. TG animals were analyzed for statisti-ally significant differences by Student t test. Then, compari-on between 3 recipient groups was performed according to theann–Whitney Test.

ory and Necroinflammatory Scores in Mice Treated With

croinflammatory score Total score

0.8 9.0 � 0.71.0, P 0.03a 7.7 � 1.4, P 0.01a

1.0, P 0.0009b 11.1 � 1.1, P 0.00007b

1.7 P 0.04*, 0.02d 10.1 � 2.0 P 0.09*, 0.002d

n tetrachloride; SD, standard deviation.

mat

Ne

.8 �

.9 �

.3 �

.2 �

arbo

imSftC�

B(t1(a�

sfiss

siMftwlia9��8f

FDsee

874 SAFADI ET AL. GASTROENTEROLOGY Vol. 127, No. 3

Results

Transgenic rIL-10 Is Antifibrotic in 2 Modelsof Liver Injury

We analyzed fibrosis in WT and TG mice follow-ng induction of fibrosis in 2 mechanistically distinct

odels of injury due to CCl4 and TAA, respectively.erum rIL-10 levels were undetectable in all WT animalsrom both groups and 147 � 222 pg/mL in the un-reated TG animals, whereas concentrations in the TGCl4 and TAA groups were 182.6 � 101.5 and 292.6280 pg/mL, respectively.In the WT CCl4 group the fibrotic area based on

ioquant morphometry was 4.3% (�2.5%) vs. TG 1.5%�0.85%) (P 0.03) (Figures 1A and 1B and 2A). Inhe WT TAA animals, the mean relative fibrosis area was.2% (�0.24%) vs. TG 0.66% (�0.36%) (P 0.01)Figures 1C and 1D and 2A). Reduced fibrosis after TAAnd CCl4 was associated with diminished expression of-SMA as assessed by immunoblot, indicating decreased

igure 1. Reduced fibrosis in IL-10 TG vs. WT animals in 2 models of) animals with 4 weeks of either carbon tetrachloride- CCl4 (A and Birius red, all as described in the Materials and Methods section. Repstablished in WT than IL-10 TG animals, highlighted by arrows (mxperiments with same number of animals in each subgroup.

tellate cell activation in the tissue (Figure 2A). Thisnding correlated with decreased immunohistochemicaltaining of �-SMA in the TG vs. WT fibrotic liverections (data not shown).

Real-time PCR was performed to quantify expres-ion of key genes associated with fibrogenesis,1 includ-ng �-SMA, TGF-�1, �-PDGF receptor, collagen I,

MP-2, and ICAM-1. Quantification of these mRNArom liver extracts confirmed an antifibrotic effect ofransgenic rIL-10 after CCl4 (Figure 2B). Comparedith WT untreated livers, all mRNAs increased fol-

owing fibrosis induction in the CCl4 model. However,ncreases were significantly higher in WT vs. TGnimals: �-SMA was increased 2.2-fold � 0.6-fold vs..9-fold � 3.4-fold (P 0.0002), TGF-�1: 2.7-fold

1.4-fold vs. 25.8-fold � 5.7-fold (P 0.0001),-PDGF receptor: 10.3-fold � 2-fold vs. 34.4-fold �.7-fold (P 0.0001), collagen I: 42.2-fold � 9.6-old vs. 36.4-fold �6.9-fold (P 0.04), matrix met-

injury. Fibrosis was induced in either WT (A and C) or IL-10 TG (B andhioacetamide- TAA (C and D), and tissue sections were stained withtative tissue sections are shown, with fibrotic septa, which are morecation �10). This finding was reproducible for at least 5 different

liver) or tresenagnifi

af2sn

Iriaw�(iW(w0

ibaHw

(ladsta�2W1T

FeREfiTectweeg2

September 2004 IMMUNE STIMULATION OF HEPATIC FIBROSIS 875

lloproteinase 2 (MMP2): 23.3-fold � 6-fold vs. 36.3-old � 7.6-fold (P 0.02), ICAM-1: 7.8-fold �.9-fold vs. 36.4-fold � 6.9-fold (P 0.0001). Aimilar pattern was also seen in the TAA model (dataot shown).

rIL-10 Only Modestly Diminishes theSeverity of Liver Injury

Despite its potential hepatoprotective activity,L-10’s antifibrotic effect was not exclusively due toeduced liver injury because ALT and AST were sim-lar between WT and TG animals: In the CCl4-treatednimals, the mean serum ALT levels in WT vs. TGere 127.6 (�43.8) IU/L vs. 142.5 (�72.9) IU/L (P

0.293), and mean serum AST levels were 57.2�28.7) IU/L vs. 63 (�56.1) IU/L (P � 0.38). Sim-larly, in the TAA model, mean serum ALT levels in

T vs. TG were 144.7 (�58.8) IU/L vs. 131.5�52.4) IU/L (P � 0.26), and mean serum AST levelsere 55.1 (�30.2) IU/L vs. 60.2 (�16.3) IU/L (P �.42). Also, the H&E histologic inflammatory scoring

igure 2. Reduced relative fibrosis area in IL-10 TG vs. WT animalsxtracts of WT and IL-10 TG mice. (A) Relative fibrosis and �-SMA expreelative fibrosis area (expressed as a percentage of total liver area) wach field was acquired at �10 magnification and then analyzed usinbrosis area in the livers of TG animals in both models. P values refeAA, respectively. Lower panel: Whole liver protein lysates were extracxpression. Decreased �-SMA expression is depicted in lysates from 2ompared with 2 WT animals also administered TAA and CCl4 (first anhe same result obtained. (B) Reduced expression of fibrogenic mRNAith CCl4 for 4 weeks as described previously, then mRNA was extxpression of a panel of genes associated with hepatic fibrogenesisxpression of fibrogenic mRNAs in IL-10 TG mice after CCl4 comparedrowth factor receptor (PDGFR), collagen �1 (I), ICAM-1, and MMP2.separate experiments.

n the WT vs. TG subgroups (Table 1) was similar inoth models: 3.9 � 0.3 vs. 3.8 � 0.4 in TAA modelnd 3.8 � 0.4 vs. 3.7 � 0.5 in CCl4 (P � NS).owever, the necroinflammatory scoring was some-hat attenuated in the TG mice (Table 1).

IL-10 Modulates Lymphocyte Subsets in TGMice

FACS analyses of TCR-�/� (�/�)�, TCR-�/�/)�, CD4�, CD8� T lymphocytes and the intracel-ular cytokines IL-4 and IFN-� were performed on spleennd liver lymphocytes before and following fibrosis in-uction. Following TAA induction, analyses revealedignificant differences between naive WT and TG micehat were primarily confined to the liver (Figure 3) butlso seen in spleen (data not shown). Intrahepatic TCR-/� and CD4� T cells were significantly reduced from9% � 2.6% and 30.7% � 6.7%, respectively, in naiveT mice to 8.7% � 0.4% (P 0.001) and 5.6% �

.7% (P 0.00001) in rIL-10 TG animals. Both hepaticCR-�/� and CD4� T cells were significantly decreased

models of liver injury and reduced �-SMA expression in whole livern: Upper panel: Animals were treated as described in Figure 1 legend.ssessed by analyzing 36 sirius red–stained liver sections per animal.omputerized Bioquant morphometry system. There was less relativeomparison between wild-type and transgenic animals after CCl4 and

and 30 �g total protein were loaded per lane and analyzed for �-SMATG mice receiving TAA and CCl4, respectively (third and fourth lanes),ond lanes). The experiment was performed in 2 sets of animals with

10 TG mice during liver injury: Either WT or IL-10 TG mice were treatedd from whole liver lysates and real-time PCR performed to quantifyescribed in the Materials and Methods section. There was reducedWT mice, including �-smooth muscle actin, TGF-�1, platelet-deriveddepicted represent triplicate values, with similar results obtained in

in 2ssioas ag a cr to cted,IL-10d secin IL-racte, as d

withData

i8wd

i1aibicmiIfIfaw1cfifsInsi0ni(

m0

FlbCfiar

FmiclccTwcdtoocCTcnti

876 SAFADI ET AL. GASTROENTEROLOGY Vol. 127, No. 3

n WT mice to 18% � 2.6% (P 0.004) and 16.3% �.7% (P 0.04), respectively. No significant changesere found in subset distribution following fibrosis in-uction among TG mice.Of note, hepatic CD8� T cells were increased only

n WT animals treated by TAA from 5.6% � 1.4% to2.4% � 7% (P � 0.09), whereas this increase wasttenuated in TG mice. In contrast, hepatic TCR-�/ncreased in both groups following TAA induction,ut they were higher in the TG animals. Secretion ofntrahepatic IL-4 and IFN-� by both CD4 and CD8 Tells were evaluated in naive and fibrotic WT and TGice. Following TAA fibrosis induction, IL-4 secret-

ng CD4 T cells were significantly decreased in naiveL-10 TG mice compared with naive WT animals,rom 41% � 13% to 10% � 3.6% (P 0.01).nterferon-� secreting CD4 T cells were also decreasedrom 45% � 8% in naive WT animals to 11% � 3%mong naive TG animals (P 0.001); 22% � 10%ithin fibrogenic WT animals (P � 0.01) and 14% �2% from fibrogenic TG mice. In contrast, IL-4 se-reting CD8 T cells were significantly increased inbrotic WT mice compared with naive WT animals,rom 20% � 9% to 51% � 18% (P 0.03). Theecretion by mixed splenic lymphocytes of IL-4 andFN-� were similar in naive WT and TG mice (dataot shown). Following fibrosis induction, however,ignificant alterations were seen only with IL-4, whichncreased from 7.2% � 5.3% to 14.9% � 8.5% (P .02) in WT animals. Although IL-4 in TG mice didot significantly change following fibrosis induction,ts relative expression was lower than in WT animals9.3% � 7.4%; P 0.06).

The CD4:CD8 ratio significantly decreased in WTice following CCl4 fibrosis induction, from 1.92 �

.33 to 1.01 � 0.33 (P 0.0002). This finding was not

igure 4. Similar relative fibrosis area in IL-10 TG vs. WT animals iniver injury following irradiation: Relative fibrosis area was assessedy morphometry in TG and WT animals after induction of liver injury byCl4. Prevention of lymphocyte expansion by irradiation diminishedbrosis equally in both WT and TG groups, suggesting that rIL-10’sntifibrotic effect was mediated by bone marrow–derived cells. Dataepresents 36 fields from each of 8 mice in each group.

igure 3. Altered intrahepatic lymphocyte subsets in WT and IL-10 TGice following liver injury. Using FACS, intrahepatic lymphocytes were

solated, and their composition was analyzed to define patterns thatorrelated with reduced fibrosis. (A) FACS analysis of intrahepaticymphocytes: The Y-axis expresses the percentage of �/� and �/ Tells from all lymphocytes and of CD4�/CD8� T cells from �/� Tells. Analyses revealed significant reduction of �/� TCR and CD4�cells in naive IL-10 TG compared with WT animals. Following 4

eeks of TAA to induce fibrosis, CD8� T cells increased and CD4� Tells decreased in WT animals. However, both subsets were furtherecreased after TAA administration in IL-10 TG mice (B) Characteriza-ion of intrahepatic lymphocytes by cell staining: Intracellular stainingf both intrahepatic CD4� and CD8� for interferon-� and IL-4 dem-nstrates that IL-4–secreting CD4� T cells are significantly de-reased in naive IL-10 TG mice compared with naive WT animals.ompared with naive WT animals, interferon (IFN)-�–secreting CD4�cells were decreased in all other groups. IL-4–secreting CD8� T

ells were significantly increased in fibrotic WT mice compared withaive WT animals and were reduced in the fibrotic TG animals. Unlikehe fibrosis-specific alterations, intrahepatic lymphocytes from micemmunized with ovalbumin (OA) showed only CD4� secretion of IFN-�.

srt

iaCwcwlmta

ceoeiab0li

FesofiNaf

September 2004 IMMUNE STIMULATION OF HEPATIC FIBROSIS 877

een in TG animals and was confined to the liver. Theesults in animals treated with CCl4 results were similaro those treated with TAA (data not shown).

Taken together, naive IL-10 TG mice showed signif-cant reduction of TCR-�/� CD4� T cells. In WTnimals following fibrosis induction, TCR-�/� andD4� T cells decreased, and CD8� T cells and IL-4ere increased. However, these increases in CD8� T

ells and IL-4 were attenuated in TG mice comparedith WT mice. These findings raised the possibility that

oss of CD8� T cells associated with elevated IL-10ight account for IL-10’s antifibrotic effect, a possibility

ested directly using adoptive transfer of CD8� T cells,s described below.

igure 5. Adoptive transfer of lymphocytes induces early changes of hither naive or CCl4-treated mice and were transferred via IP administrubsequent treatment, lymphocytes were isolated from animals with if lymphocytes from CCl4-treated (A and B) or control (C) mice. Liver sebrosis by the appearance of slender septa (arrows) in animals followio parenchymal disruption or inflammation is apparent. (D) �-Actin imnimals were analyzed for �-smooth muscle actin expression by immurom animals treated with CCl4 compared with 2 mice receiving lymp

IL-10’s Antifibrotic Effect DisappearsFollowing Irradiation

Sublethal irradiation was used to reduce lympho-yte activity and thereby define whether the antifibroticffect of IL-10 was attributable to lymphocyte alterationsr other factors. Both WT and IL-10 TG mice (7 animalsach group) were sublethally irradiated, then fibrosis wasnduced by 4 weeks of CCl4 in both WT and TGnimals. In irradiated animals, fibrosis was attenuated inoth groups to a similar extent: Relative fibrosis area was.33% � 0.23% in WT and 0.32% � 0.19% in TGiver (Figure 4). Analysis of �-SMA by immunoblot,mmunohistochemistry, and real-time PCR analyses

ic fibrosis in naive SCID mice. Mixed lymphocytes were isolated fromweekly from WT donors to each SCID recipient for 4 weeks. For eachsing cumulative exposure to CCl4. (A–C) Histology of SCID recipientss were stained with sirius red, which demonstrates evidence of early

e transfer of lymphocytes from CCl4-treated but not from control mice.blot: Whole liver lysates (30 �g of total protein per lane) of recipientt. Expression was greater in 2 SCID recipients receiving lymphocyteses from untreated animals.

epatationncreaction

ng thmunonoblohocyt

wlsg

CfsSCsw

C5filoicAI1crawCwt

wT0phrihmCnii3mnpAso

osfitsbpnpT

FoirftiCelndFar

878 SAFADI ET AL. GASTROENTEROLOGY Vol. 127, No. 3

ere also similar in both groups (data not shown). Fol-owing irradiation and fibrosis induction, ELISA meanerum rIL-10 levels were 188 � 121 pg/mL in the TGroup and were undetectable in WT mice.

Adoptive Transfer Model Indicates ThatHepatic Fibrosis Can Be Mediated by CD8�T Cells

Because the data pointed to a potential role ofD8� T cells in mediating fibrosis, we adoptively trans-

erred lymphocyte subsets from animals with CCl4 fibro-is to determine their effect on naive (nontransgenic)CID animals. Remarkably, normal animals receivingD8� T cells from animals with CCl4 fibrosis demon-

trated mild fibrotic septa based on sirius red staining,ithout obvious inflammatory cell infiltration in all

igure 6. Increased liver injury and fibrosis following adoptive transferf CD8� T lymphocytes from animals with CCl4 fibrosis. (A) Livernjury: serum aminotransferase levels were assayed in SCID miceeceiving either whole mixed lymphocytes, CD4�, or CD8� T cellsrom CCl4-treated animals for 4 weeks to assess whether adoptiveransfer induced hepatocellular damage. Although all animals receiv-ng lymphocytes had elevated AST, values were significantly greater inD8� T-cell recipients compared with SCIDs that were administeredither whole lymphocytes or CD4� T cells. There was no effect of

ymphocytes from untreated control animals. All control recipients hadormal ALT and AST values. (B) Fibrosis: livers of SCID animalsescribed in panel A were analyzed by morphometry as described inigure 2 legend, and the relative fibrosis area is depicted. SCIDnimals receiving CD8� T cells had significantly more fibrosis thanecipients of either whole mixed lymphocytes or CD4� T cells.

Cl4-stimulated lymphocyte recipients (Figure 5A andB). In contrast, control animal groups did not developbrosis, including SCID mice adoptively transferredymphocytes from animals that were immunized withvalbumin (Figure 5C). These findings correlated withncreased �-SMA expression by immunoblot in treatedompared with control animals (Figure 5D). Mean serumST/ALT values (Figure 6) were 58.5 � 15.3/21.5 � 1

U/L in animals receiving whole lymphocytes, 57.4 �9.8/19.4 � 2.9 IU/L in animals receiving CD4� Tells, and 141.5 � 50.2/27.3 � 5.6 IU/L in animalseceiving CD8� T cells. Thus, CD8� T-cell recipientsppeared to have more hepatocellular injury comparedith SCIDs that received either whole lymphocytes orD4� T cells (P � 0.02, P � 0.018, respectively). Thisas associated with increased fibrosis as assessed by quan-

itative morphometry (Figure 6B).Increased mRNA expression of fibrogenic target genes

as associated with early septa formation in CD8�-cell recipients (Figure 7), including TGF-�1: 1.5- �.04-fold increase in SCID mice receiving whole lym-hocytes from CCl4-treated mice vs. recipients from ve-icle-treated controls, 1.9- � 0.05-fold in CD4� T-cellecipients, and 2.4- � 0.1-fold in CD8� T-cells recip-ents. TGF-�1 mRNA expression was significantlyigher in CD8� T-cell recipients compared with ani-als receiving whole lymphocytes (P 0.0001) orD4� T cells (P 0.0004). Collagen I expression wasot increased (1-fold � 0.04-fold) in SCID mice receiv-ng whole lymphocytes from CCl4-treated animals butncreased 1.6- � 0.1-fold in CD4� T-cell recipients and.3- � 0.05-fold in CD8� T-cell recipients. Collagen IRNA expression in CD8� T-cell recipients was sig-

ificantly higher than in animals receiving whole lym-hocytes (P 0.0001) or CD4� T cells (P 0.0001).s described earlier, these CD8 cells were mainly IL-4

ecretors (Figure 3B), whereas this capacity was absent invalbumin-stimulated lymphocytes.

Discussion

Our study has explored the potential interactionsf humoral factors and lymphocyte subsets by first as-essing the impact of hepatic transgenic rIL-10 on liverbrosis and lymphocyte responses in vivo then exploitinghis response to dissect the contributions of lymphocyteubsets on fibrogenesis. IL-10 was chosen because of itsroad immunomodulatory and antifibrotic properties,roviding a useful reagent to perturb the homeostasis oformal and injured liver. IL-10 is a regulatory cytokineroduced by numerous cell types including activatedh0, Th1 (in human), and Th2 CD4� T-helper cells;

ccipracisemKTon

IIacg(tisaen

tmatp

Thwaflfiita

ooiiiacCpTtlSi

�CCmpcitC

Flmh1

September 2004 IMMUNE STIMULATION OF HEPATIC FIBROSIS 879

ytotoxic CD8� T cells; monocytes; B cells; Kupfferells; hepatocytes31; and hepatic stellate cells.32 IL-10nhibits antigen-specific activation, proliferation, androduction of cytokines by Th0, Th1, and Th2 clones byeducing the antigen-presenting capacity of monocytes,ssociated with down-regulation of MHC class II mole-ules and B7 expression on their surface.33–35 In contrast,n B lymphocytes, IL-10 stimulates immunoglobulinecretion.7,36 IL-10 also exerts potent anti-inflammatoryffects by down-regulating the synthesis of proinflam-atory cytokines and chemokines by monocytes andupffer cells stimulated by endotoxin, including IL-1,NF-�, IL-6, IL-8, and IL-12 and by inducing synthesisf the IL-1R antagonist.37–39 IL-10 also down-regulateseutrophil chemotaxis and chemokine expression.40

In addition to its broad immunoregulatory effects,L-10 has specific activities in liver fibrosis. Expression ofL-10 by stellate cells increases during cellular activationnd is associated with reduced matrix accumulation inulture.5,7 Moreover, levels of IL-10 are reduced in pro-ressive human fibrosis because of hepatitis C virusHCV),41 implying that loss of IL-10 might contributeo increased fibrogenesis. Animal models of liver injury,ncluding IL-10 knockout mice,7,11,42,43 have also sub-tantiated IL-10’s antifibrotic activity. Such findings innimals have led to a human trial of IL-10, which yieldedncouraging early results after 3 months,44 which wereot sustained at 1 year.45

We chose to generate a transgenic mouse with hepa-ocyte secretion of rIL-10 rather than systemically ad-inistering the cytokine to enhance its local hepatic

ctivity while minimizing systemic effects. As expected,ransgenic rIL-10 exerted an antifibrotic effect in 2 ex-erimental models of liver injury and fibrosis, CCl and

igure 7. Adoptive transfer of CD8� T lymphocytes induces TGF-�1 anysates of naive SCID mice after receiving 4 weeks of either whole mRNA levels of TGF-�1 and collagen �1 (I) mRNA by real-time PCR.igher in CD8� T-cell recipients. There were 7 animals in each group, aof these 2 experiments is shown). Recipients of lymphocytes from

4

AA. However, this effect was not solely due to aepatoprotective effect of rIL-10 because AST and ALTere similar between TG and WT mice following toxin

dministration, although a modest increase in necroin-ammation was seen histologically. Moreover, the anti-brotic effect of rIL-10 disappeared when mice wererradiated prior to induction of fibrosis, suggesting thathe response was mediated by alterations in lymphocytectivity and/or distribution.

Overexpression of IL-10 may have a number of effectsn the hepatic immune system. Direct effects on T cellsf IL-10 include suppression of a variety of cytokines,ncluding IL-2 tumor necrosis factor-� and IL-15, andnhibition of T-cell proliferation.46–48 Indirect effectsnclude inhibition of the antigen-presenting capacity ofntigen-presenting cells and down-regulation of MHClass II and costimulatory molecules including CD54,D80, and CD86.49 The priming of CD4� T cells in theresence of IL-10 also results in the development of ther1 subset of regulatory T cells. These have the charac-

eristics of producing IL-10, and TGF-� upon restimu-ation (similar to Th2 cells), but do not produce IL-4.uch regulatory T cells induce local antigen nonspecificmmune suppression after activation.50

Fibrosis induction was associated with decreased TCR-/� CD4�, CD4� interferon-�� T cells, and CD4/D8 ratio and increased numbers of total CD8� T cells,D8� IL-4� T cells, and secretion of IL-4 by totalixed lymphocytes. Moreover, these changes were more

rominent in intrahepatic lymphocytes compared withirculating lymphocytes. In TG animals, however, thencrement of CD8� T cells and IL-4 was less prominenthan in WT animals, raising the possibility that loss ofD8� T cell induction in TG mice might contribute to

lagen I mRNA in recipient livers. mRNA was extracted from whole liverymphocytes, CD4�, or CD8� T cells and analyzed for expression oftranscripts were elevated, with expression of collagen I significantlyseparate experiments were performed with similar results (data from

rol animals had no change in either transcript.

d colixed lBothnd 2cont

raln

jcstoallttbieittcacc

eHpCiwtwilvttphc

wcobirbfr

1

1

1

1

1

1

1

1

1

880 SAFADI ET AL. GASTROENTEROLOGY Vol. 127, No. 3

educed fibrogenesis. These observations led to a directssessment of fibrogenic activity conferred by differentymphocyte subsets following their adoptive transfer intoaive SCID animals.Evidence of early fibrogenic stimulation and liver in-

ury was apparent following adoptive transfer of lympho-ytes from CCl4 treated donors, with CD8� T-cell sub-ets harboring the greatest fibrogenic activity. Althoughhe relative effects may appear modest, the brief exposuref recipient animals to lymphocyte subsets followingdoptive transfer through intermittent administration isikely to be only a small fraction of the extent to whichymphocytes reside in liver during native injury. None-heless, adoptive transfer of fibrogenic activity of thisype has not been previously reported, and these findingsegin to uncover a potential role of lymphocyte subsetsn mediating liver injury and fibrosis, a feature not yetxplored despite the prominent lymphocytic infiltrationn many chronic liver diseases, especially viral hepati-is.51 The finding significantly expands earlier observa-ions implicating CD8� T cells in biliary fibrosis asso-iated with murine graft vs. host disease52 and inlcoholic cirrhosis.53,54 Significant reductions of CD4� Tells was also observed in alcoholic cirrhosis and evenorrelates with severity of liver cirrhosis.55

Our findings are also consonant with the recognizedffect of immunosuppression on accelerating fibrosis inIV/HCV infection and following orthotopic liver trans-

lantation for HCV.18–20 A relative decrease ratio ofD4/CD8 in these settings parallels the association seen

n mice following fibrogenic induction in our studies,hereas transgenic rIL-10 increased this ratio, primarily

hrough a relative depletion of CD8� T cells. Consistentith our data, a decreased CD4/CD8 ratio was reported

n chronic hepatitis C associated with increased CD8�ymphocytes.56 An intriguing report has described re-ersibility of hepatic fibrosis following bone-marrowransplantation for treatment of thalassemia.57 Althoughhe primary therapeutic effect of this intervention wasrobably to reduce hepatic iron, another benefit mightave been the elimination of a fibrogenic CD8� T-celllone from the host bone marrow by replacement.

Although our data are derived from a murine modelhose immunoregulation in fibrosis is likely to be far less

omplex than in humans, the findings emphasize the rolef lymphocyte subsets in mediating not only liver injuryut also the fibrogenic response. Admittedly, these stud-es do not address several key aspects of lymphocyteesponses in fibrosis, including their apoptotic activity,58

ut they nonetheless provide an important starting pointor further refining our understanding of the immuneegulation of fibrosis.

References1. Friedman SL. Molecular regulation of hepatic fibrosis, an inte-

grated cellular response to tissue injury. J Biol Chem 2000;275:2247–2250.

2. Friedman SL. Liver fibrosis—from bench to bedside. J Hepatol2003;38(Suppl. 1):S38–S53.

3. Gressner AM, Weiskirchen R, Breitkopf K, Dooley S. Roles ofTGF-� in hepatic fibrosis. Front Biosci 2002;7:D793–D807.

4. Paradis V, Dargere D, Vidaud M, De Gouville AC, Huet S, MartinezV, Gauthier JM, Ba N, Sobesky R, Ratziu V, Bedossa P. Expres-sion of connective tissue growth factor in experimental rat andhuman liver fibrosis. Hepatology 1999;30:968–976.

5. Wang SC, Tsukamoto H, Rippe RA, Schrum L, Ohata M. Expres-sion of interleukin-10 by in vitro and in vivo activated hepaticstellate cells. J Biol Chem 1998;273:302–308.

6. Mathurin P, Xiong S, Kharbanda KK, Veal N, Miyahara T, Mo-tomura K, Rippe RA, Bachem MG, Tsukamoto H. IL-10 receptorand coreceptor expression in quiescent and activated hepaticstellate cells. Am J Physiol Gastrointest Liver Physiol 2002;282:G981–G990.

7. Thompson K, Maltby J, Fallowfield J, McAulay M, Millward-SadlerH, Sheron N. Interleukin-10 expression and function in experi-mental murine liver inflammation and fibrosis [In Process Cita-tion]. Hepatology 1998;28:1597–1606.

8. Shi Z, Steward TA, Rockey DC. Liver injury and fibrosis in inter-feron gamma knockout mice. Hepatology 1995;22:276A.

9. Louis H, Le Moine O, Peny MO, Gulbis B, Nisol F, Goldman M,Deviere J. Hepatoprotective role of interleukin 10 in galac-tosamine/lipopolysaccharide mouse liver injury. Gastroenterol-ogy 1997;112:935–942.

0. Louis H, Le Moine O, Peny MO, Quertinmont E, Fokan D, GoldmanM, Deviere J. Production and role of interleukin-10 in concanavalinA-induced hepatitis in mice. Hepatology 1997;25:1382–1389.

1. Louis H, Van Laethem JL, Wu W, Quertinmont E, Degraef C, Vanden Berg K, Demols A, Goldman M, Le Moine O, Geerts A,Deviere J. Interleukin-10 controls neutrophilic infiltration, hepa-tocyte proliferation, and liver fibrosis induced by carbon tetrachlo-ride in mice. Hepatology 1998;28:1607–1615.

2. Heinzel FP, Sadick MD, Holaday BJ, Coffman RL, Locksley RM.Reciprocal expression of interferon � or interleukin 4 during theresolution or progression of murine leishmaniasis. Evidence forexpansion of distinct helper T-cell subsets. J Exp Med 1989;169:59–72.

3. Shi Z, Wakil AE, Rockey DC. Strain-specific differences in mousehepatic wound healing are mediated by divergent T-helper cyto-kine responses. Proc Natl Acad Sci U S A 1997;94:10663–10668.

4. Wang ZE, Zheng S, Corry DB, Dalton DK, Seder RA, Reiner SL,Locksley RM. Interferon �-independent effects of interleukin 12administered during acute or established infection due to Leish-mania major. Proc Natl Acad Sci U S A 1994;91:12932–12936.

5. Wang ZE, Reiner SL, Zheng S, Dalton DK, Locksley RM. CD4�effector cells default to the Th2 pathway in interferon �-deficientmice infected with Leishmania major. J Exp Med 1994;179:1367–1371.

6. Bataller R, North KE, Brenner DA. Genetic polymorphisms and theprogression of liver fibrosis: a critical appraisal. Hepatology2003;37:493–503.

7. Hillebrandt S, Goos C, Matern S, Lammert F. Genome-wide anal-ysis of hepatic fibrosis in inbred mice identifies the susceptibilitylocus Hfib1 on chromosome 15. Gastroenterology 2002;123:2041–2051.

8. Sanchez-Quijano A, Andreu J, Gavilan F, Luque F, Abad MA, SotoB, Munoz J, Aznar JM, Leal M, Lissen E. Influence of humanimmunodeficiency virus type 1 infection on the natural course of

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

4

4

4

4

4

4

4

4

5

September 2004 IMMUNE STIMULATION OF HEPATIC FIBROSIS 881

chronic parenterally acquired hepatitis C. Eur J Clin MicrobiolInfect Dis 1995;14:949–953.

9. Benhamou Y, Bochet M, Di Martino V, Charlotte F, Azria F,Coutellier A, Vidaud M, Bricaire F, Opolon P, Katlama C, PoynardT. Liver fibrosis progression in human immunodeficiency virusand hepatitis C virus coinfected patients. The Multivirc Group.Hepatology 1999;30:1054–1058.

0. Benhamou Y, Di Martino V, Bochet M, Colombet G, Thibault V,Liou A, Katlama C, Poynard T. Factors affecting liver fibrosis inhuman immunodeficiency virus-and hepatitis C virus-coinfectedpatients: impact of protease inhibitor therapy. Hepatology 2001;34:283–287.

1. Marcellin P, Asselah T, Boyer N. Fibrosis and disease progres-sion in hepatitis C. Hepatology 2002;36:S47–S56.

2. Poynard T, Mathurin P, Lai CL, Guyader D, Poupon R, TainturierMH, Myers RP, Muntenau M, Ratziu V, Manns M, Vogel A, CapronF, Chedid A, Bedossa P. A comparison of fibrosis progression inchronic liver diseases. J Hepatol 2003;38:257–265.

3. Schluger LK, Sheiner PA, Thung SN, Lau JY, Min A, Wolf DC, FielI, Zhang D, Gerber MA, Miller CM, Bodenheimer HC Jr. Severerecurrent cholestatic hepatitis C following orthotopic liver trans-plantation. Hepatology 1996;23:971–976.

4. Wu H, Wade M, Krall L, Grisham J, Xiong Y, Van Dyke T. Targetedin vivo expression of the cyclin-dependent kinase inhibitor p21halts hepatocyte cell-cycle progression, postnatal liver develop-ment and regeneration. Genes Dev 1996;10:245–260.

5. Rizzo LV, Morawetz RA, Miller-Rivero NE, Choi R, Wiggert B, ChanCC, Morse HC III, Nussenblatt RB, Caspi RR. IL-4 and IL-10 areboth required for the induction of oral tolerance. J Immunol1999;162:2613–2622.

6. Chevallier M, Guerret S, Chossegros P, Gerard F, Grimaud JA. Ahistological semiquantitative scoring system for evaluation ofhepatic fibrosis in needle liver biopsy specimens: comparisonwith morphometric studies. Hepatology 1994;20:349–355.

7. Ishak K, Baptista A, Bianchi L, Callea F, De Groote J, Gudat F,Denk H, Desmet V, Korb G, MacSween RN, Phillips MJ, PortmannBG, Poulsen H, Scheuer PJ, Schmid M, Thaler H. Histologicalgrading and staging of chronic hepatitis. J Hepatol 1995;22:696–699.

8. Knodell RG, Ishak KG, Black WC, Chen TS, Craig R, Kaplowitz N,Kiernan TW, Wollman J. Formulation and application of a numer-ical scoring system for assessing histological activity in asymp-tomatic chronic active hepatitis. Hepatology 1981;1:431–435.

9. Miranda SR, Erlich S, Visser JW, Gatt S, Dagan A, Friedrich VL Jr,Schuchman EH. Bone marrow transplantation in acid sphingomy-elinase-deficient mice: engraftment and cell migration into thebrain as a function of radiation, age, and phenotype. Blood1997;90:444–452.

0. Rockey DC, Boyles JK, Gabbiani G, Friedman SL. Rat hepaticlipocytes express smooth muscle actin upon activation in vivoand in culture. J Submicrosc Cytol Pathol 1992;24:193–203.

1. Alfrey EJ, Most D, Wang X, Lee LK, Holm B, Krieger NR, Sibley RK,Huie P, Dafoe DC. Interferon-� and interleukin-10 messengerRNA are up-regulated after orthotopic liver transplantation intolerant rats: evidence for cytokine-mediated immune dysregula-tion. Surgery 1995;118:399–404; discussion 404–405.

2. Thompson KC, Trowern A, Fowell A, Marathe M, Haycock C, ArthurMJP, Sheron N. Primary rat and mouse hepatic stellate cellsexpress the macrophage inhibitor cytokine interleukin-10 duringthe course of activation in vitro. Hepatology 1998;28:1518–1524.

3. Del Prete G, De Carli M, Almerigogna F, Giudizi MG, Biagiotti R,Romagnani S. Human IL-10 is produced by both type 1 helper(Th1) and type 2 helper (Th2) T cell clones and inhibits theirantigen-specific proliferation and cytokine production. J Immunol1993;150:353–360.

4. de Waal Malefyt R, Haanen J, Spits H, Roncarolo MG, te Velde A,Figdor C, Johnson K, Kastelein R, Yssel H, de Vries JE. Interleukin10 (IL-10) and viral IL-10 strongly reduce antigen-specific human Tcell proliferation by diminishing the antigen-presenting capacity ofmonocytes via downregulation of class II major histocompatibilitycomplex expression. J Exp Med 1991;174:915–924.

5. Willems F, Marchant A, Delville JP, Gerard C, Delvaux A, Velu T,de Boer M, Goldman M. Interleukin-10 inhibits B7 and intercel-lular adhesion molecule-1 expression on human monocytes. EurJ Immunol 1994;24:1007–1009.

6. Rousset F, Garcia E, Defrance T, Peronne C, Vezzio N, Hsu DH,Kastelein R, Moore KW, Banchereau J. Interleukin 10 is a potentgrowth and differentiation factor for activated human B lympho-cytes. Proc Natl Acad Sci U S A 1992;89:1890–1893.

7. Defrance T, Vanbervliet B, Briere F, Durand I, Rousset F, Banche-reau J. Interleukin 10 and transforming growth factor � cooperateto induce anti-CD40-activated naive human B cells to secreteimmunoglobulin A. J Exp Med 1992;175:671–682.

8. de Waal Malefyt R, Abrams J, Bennett B, Figdor CG, de Vries JE.Interleukin 10(IL-10) inhibits cytokine synthesis by human mono-cytes: an autoregulatory role of IL-10 produced by monocytes. JExp Med 1991;174:1209–1220.

9. Fiorentino DF, Zlotnik A, Mosmann TR, Howard M, O’Garra A.IL-10 inhibits cytokine production by activated macrophages.J Immunol 1991;147:3815–3822.

0. Kasama T, Strieter RM, Lukacs NW, Burdick MD, Kunkel SL.Regulation of neutrophil-derived chemokine expression by IL-10.J Immunol 1994;152:3559–3569.

1. Napoli J, Bishop GA, McGuinness PH, Painter DM, McCaughanGW. Progressive liver injury in chronic hepatitis C infection corre-lates with increased intrahepatic expression of Th1-associatedcytokines. Hepatology 1996;24:759–765.

2. Arai T, Hiromatsu K, Kobayashi N, Takano M, Ishida H, Nimura Y,Yoshikai Y. IL-10 is involved in the protective effect of dibutyrylcyclic adenosine monophosphate on endotoxin-induced inflam-matory liver injury. J Immunol 1995;155:5743–5749.

3. Gazzinelli RT, Wysocka M, Hieny S, Scharton-Kersten T, CheeverA, Kuhn R, Muller W, Trinchieri G, Sher A. In the absence ofendogenous IL-10, mice acutely infected with Toxoplasma gondiisuccumb to a lethal immune response dependent on CD4� Tcells and accompanied by overproduction of IL-12, IFN-� andTNF-�. J Immunol 1996;157:798–805.

4. Nelson DR, Lauwers GY, Lau JY, Davis GL. Interleukin 10 treat-ment reduces fibrosis in patients with chronic hepatitis C: a pilottrial of interferon nonresponders. Gastroenterology 2000;118:655–660.

5. Nelson DR, Tu Z, Soldevila-Pico C, Abdelmalek M, Zhu H, Xu YL,Cabrera R, Liu C, Davis GL. Long-term interleukin 10 therapy inchronic hepatitis C patients has a proviral and anti-inflammatoryeffect. Hepatology 2003;38:859–868.

6. de Waal Malefyt R, Yssel H, de Vries JE. Direct effects of IL-10 onsubsets of human CD4� T cell clones and resting T cells. Spe-cific inhibition of IL-2 production and proliferation. J Immunol1993;150:4754–4765.

7. Schandene L, Alonso-Vega C, Willems F, Gerard C, Delvaux A,Velu T, Devos R, de Boer M, Goldman M. B7/CD28-dependentIL-5 production by human resting T cells is inhibited by IL-10.J Immunol 1994;152:4368–4374.

8. Bejarano MT, de Waal Malefyt R, Abrams JS, Bigler M, BacchettaR, de Vries JE, Roncarolo MG. Interleukin 10 inhibits allogeneicproliferative and cytotoxic T-cell responses generated in primarymixed lymphocyte cultures. Int Immunol 1992;4:1389–1397.

9. Moore KW, de Waal Malefyt R, Coffman RL, O’Garra A. Interleu-kin-10 and the interleukin-10 receptor. Annu Rev Immunol 2001;19:683–765.

0. Roncarolo MG, Bacchetta R, Bordignon C, Narula S, Levings MK.Type 1 T regulatory cells. Immunol Rev 2001;182:68–79.

5

5

5

5

5

5

5

5

RN(

h

aCA

882 SAFADI ET AL. GASTROENTEROLOGY Vol. 127, No. 3

1. Koziel MJ. Cytokines in viral hepatitis. Semin Liver Dis 1999;19:157–169.

2. Inada S, Suzuki K, Kimura T, Hayashi A, Narita T, Yui R, AsakuraH, Fujiwara M. Concentric fibrosis and cellular infiltration aroundbile ducts induced by graft-versus-host reaction in mice: a role ofCD8� cells. Autoimmunity 1995;22:163–171.

3. Laso FJ, Iglesias-Osma C, Ciudad J, Lopez A, Pastor I, Orfao A.Chronic alcoholism is associated with an imbalanced productionof Th-1/Th-2 cytokines by peripheral blood T cells. Alcohol ClinExp Res 1999;23:1306–1311.

4. Laso FJ, Iglesias-Osma C, Ciudad J, Lopez A, Pastor I, Torres E,Orfao A. Alcoholic liver cirrhosis is associated with a decreasedexpression of the CD28 costimulatory molecule, a lower ability ofT cells to bind exogenous IL-2, and increased soluble CD8 levels.Cytometry 2000;42:290–295.

5. Lombardo L, Capaldi A, Poccardi G, Vineis P. Peripheral bloodCD3 and CD4 T-lymphocyte reduction correlates with severity ofliver cirrhosis. Int J Clin Lab Res 1995;25:153–156.

6. Panasiuk A, Prokopowicz D, Zak J, Wysocka J. Peripheral blood T,B, and NK cells in relation to histological hepatitis activity andfibrosis stage in chronic hepatitis C. Hepatogastroenterology2003;50:178–182.

7. Muretto P, Angelucci E, Lucarelli G. Reversibility of cirrhosis inpatients cured of thalassemia by bone marrow transplantation.Ann Intern Med 2002;136:667–672.

8. Kobayashi S, Seki S, Kawada N, Morikawa H, Nakatani K, UyamaN, Ikeda K, Nakajima Y, Arakawa T, Kaneda K. Apoptosis of Tcells in the hepatic fibrotic tissue of the rat: a possible inducingrole of hepatic myofibroblast-like cells. Cell Tissue Res 2003;311:353–364.

Received September 1, 2003. Accepted April 22, 2004.Address requests for reprints to: Scott L. Friedman, M.D., Box 1123,

oom 11-70F, Mount Sinai School of Medicine, 1425 Madison Avenue,ew York, New York 10029. e-mail: [email protected]; fax:

212) 849-2574.The authors thank Francis Eng, Michael P. Cooreman, Suhair She-

ada, and Alex Yui Hui for technical assistance.Supported by grants from NIDDK (DK37340 and DK56621 to S.L.F.

nd DK02965-01 to W.Z.M.), Feld Family Trust, Artzt Primary Biliaryirrhosis Program, and Shirley Fiterman Basic Science Researchward (to W.Z.M.).