Tumor necrosis factor?like weak inducer of apoptosis is a ... · Tumor Necrosis Factor–Like Weak...

Tumor Necrosis Factor–Like Weak Inducer ofApoptosis Is a Mitogen for Liver Progenitor Cells

Janina E. E. Tirnitz-Parker,1,2,3 Cornelia S. Viebahn,4 Aniela Jakubowski,5 Borut R. S. Klopcic,1

John K. Olynyk,1,2,3 George C. T. Yeoh,2,4 and Belinda Knight6

Liver progenitor cells (LPCs) represent the cell compartment facilitating hepatic regenerationduring chronic injury while hepatocyte-mediated repair mechanisms are compromised. LPCproliferation is frequently observed in human chronic liver diseases such as hereditary hemo-chromatosis, fatty liver disease, and chronic hepatitis. In vivo studies have suggested that a tu-mor necrosis factor family member, tumor necrosis factor–like weak inducer of apoptosis(TWEAK), is promitotic for LPCs; whether it acts directly is not known. In our murine chol-ine-deficient, ethionine-supplemented (CDE) model of chronic liver injury, TWEAK receptor[fibroblast growth factor-inducible 14 (Fn14)] expression in the whole liver is massively up-regulated. We therefore set out to investigate whether TWEAK/Fn14 signaling promotes theregenerative response in CDE-induced chronic liver injury by mitotic stimulation of LPCs.Fn14 knockout (KO) mice showed significantly reduced LPC numbers and attenuatedinflammation and cytokine production after 2 weeks of CDE feeding. The close associationbetween LPC proliferation and activation of hepatic stellate cells in chronic liver injuryprompted us to investigate whether fibrogenesis was also modulated in Fn14 KO animals.Collagen deposition and expression of key fibrogenesis mediators were reduced after 2 weeksof injury, and this correlated with LPC numbers. Furthermore, the injection of 2-week-CDE-treated wildtype animals with TWEAK led to increased proliferation of nonparenchymal pancytokeratin–positive cells. Stimulation of an Fn14-positive LPC line with TWEAK led to nu-clear factor kappa light chain enhancer of activated B cells (NFjB) activation and dose-de-pendent proliferation, which was diminished after targeting of the p50 NFjB subunit byRNA interference. Conclusion: TWEAK acts directly and stimulates LPC mitosis in an Fn14-dependent and NFjB-dependent fashion, and signaling via this pathway mediates the LPCresponse to CDE-induced injury and regeneration. (HEPATOLOGY 2010;52:291-302)

See Editorial on 3.

Hepatocyte-driven liver regeneration followingacute injury is a highly orchestrated processin which remaining, healthy hepatocytes pro-

liferate under the control of well-understood genetic,cellular, and metabolic networks.1 During chronic liverinjury, however, when the functional liver mass isreduced and coincidentally hepatocyte proliferation iswholly or partly impaired, liver regeneration is

Abbreviations: aSMA, alpha smooth muscle actin; BMOL, bipotential murine oval liver; BMOL-TAT, bipotential murine oval liver–tyrosine aminotransferase;BrdU, 5-bromo-2-deoxyuridine; CDE, choline-deficient, ethionine-supplemented; CK, cytokeratin; CKpan, pan cytokeratin; Cy7, cyanine 7; DAPI, 40,6-diamidino-2-phenylindole; FACS, fluorescence-activated cell sorting; FBS, fetal bovine serum; Fn14, fibroblast growth factor-inducible 14; GFP, green fluorescentprotein; Heps, hepatocytic cell proliferation; IFNc, interferon c; IGF-II, insulin-like growth factor II; IL-6, interleukin 6; KO, knockout; LPC, liver progenitorcell; LTb, lymphotoxin b; MMP, matrix metalloproteinase; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; NFjB, nuclear factor kappa lightchain enhancer of activated B cells; NK, natural killer; NKT, natural killer T; NP, nonparenchymal; PE, phycoerythrin; qPCR, quantitative polymerase chainreaction; rhTWEAK, recombinant human tumor necrosis factor–like weak inducer of apoptosis; RNAi, RNA interference; SEM, standard error of the mean;siRNA, small interfering RNA; SOX-9, SRY (sex determining region Y)-box 9; TIMP, tissue inhibitor of metalloproteinases; TNF, tumor necrosis factor; TRAF,tumor necrosis factor receptor–associated factor; TWEAK, tumor necrosis factor–like weak inducer of apoptosis; WT, wildtype.From the 1School of Medicine and Pharmacology, University of Western Australia, Fremantle, Australia; 2Western Australian Institute for Medical Research, Perth,

Australia; 3Curtin Health Innovation Research Institute, Curtin University of Technology, Perth, Australia; 4School of Biomedical, Biomolecular, and Chemical Sciences,University of Western Australia, Crawley, Australia; 5Biogen Idec, Incorporated, Cambridge, MA; and 6Scottish Centre for Regenerative Medicine, Queens MedicalResearch Institute, Centre for Inflammation Research, Edinburgh, Scotland.Received November 7, 2009; accepted March 3, 2010.Janina E. E. Tirnitz-Parker was supported by an International Postgraduate Research Scholarship and by funds provided by the Western Australian Institute for

Medical Research and the Cancer Council of Western Australia. This research was funded in part by grants from the National Health and Medical Research Council ofAustralia. Belinda Knight was a recipient of a fellowship from the Gastroenterological Society of Australia (until December 2008) and is currently supported by theCentre for Regenerative Medicine at the University of Edinburgh. John K. Olynyk is a National Health and Medical Research Council clinical practitioner fellow.

291

facilitated by a compartment of dynamic, heterogene-ous liver progenitor cells (LPCs),2 which in rodentsare often called oval cells. LPCs are rarely detected inhealthy tissue, but once activated, they proliferate andmigrate into the parenchyma and differentiate into ei-ther cholangiocytes or hepatocytes. This process, unlikehepatocyte-mediated liver regeneration, is relativelypoorly understood. Several factors mediating the LPCresponse to injury have been identified, including tu-mor necrosis factor (TNF), lymphotoxin b (LTb), inter-feron c (IFNc), and transforming growth factor b, amongothers3-7; however, the mechanisms by which these medi-ators elicit their effects remain largely unknown. Thus, nodefinitive LPC growth factor has been described to date.In this respect, the 2005 report by Jakubowski and

coworkers8 was a milestone publication; it proposed aTNF family member, tumor necrosis factor–like weakinducer of apoptosis (TWEAK), as a direct LPC mito-gen. Like most TNF superfamily ligands, TWEAK [alsoknown as tumor necrosis factor (ligand) superfamilymember 12, CD255, or previously APO3 ligand] is ini-tially synthesized as a type II transmembrane proteinthat can be cleaved into a soluble cytokine.9 TWEAK, anoncovalent homotrimer, mediates its activity throughbinding to a 14-kDA type I transmembrane receptortermed fibroblast growth factor-inducible 14 (Fn14).10

Fn14 does not possess intrinsic protein kinase activity.Instead, it associates with tumor necrosis factor recep-tor–associated factor (TRAF) adaptor molecules andactivates nuclear factor kappa light chain enhancer ofactivated B cells (NFjB) and the extracellular signal-regulated kinase and c-Jun N-terminal kinase mitogen-activated protein kinase pathways.11,12 TWEAK regu-lates a diverse range of cellular processes, including pro-liferation, differentiation, migration, cell survival, andcell death, and has also been shown to act as a proangio-genic and proinflammatory factor.13

Although TWEAK is almost ubiquitously expressedin adult tissues, including the liver, its cognate receptoris barely detectable in hepatic tissue except duringinjury and repair. Feng et al.14 reported rapid induc-tion of Fn14 expression during the early phases ofregeneration following partial hepatectomy. Theauthors also investigated liver cancer–derived cell lines,

and interestingly, Fn14 overexpression was mainlyfound in the poorly differentiated lines; this hinted atpossible involvement in progenitor cells. This hypothe-sized relationship between LPCs and Fn14 signalingwas strengthened by experiments described by Jaku-bowski and coworkers.8 In their study, it was demon-strated that hepatic TWEAK overexpression, aftertransgenic or adenoviral delivery, induces an LPCresponse; knockout (KO) of TWEAK/Fn14 inhibitsthe appearance of LPCs in an experimental model ofoval cell proliferation. Furthermore, TWEAK directlyinduces proliferation of a biliary epithelial cell linewith progenitor-like characteristics. LPCs consist of aheterogeneous cellular compartment, with different dis-crete subsets potentially being activated under variousinjury contexts. Thus, we wished to further explore therelationship between TWEAK and LPCs with a differentin vivo model of hepatic injury and LPC proliferationdeveloped in our laboratory.15 Additionally, we aimed toinvestigate the effects of TWEAK on LPC lines in vitro.We now demonstrate that TWEAK directly stimulatesLPC mitosis in an Fn14-dependent and NFjB-depend-ent fashion and that this pathway plays a major role inthe rapid growth phase of the LPC response to choline-deficient, ethionine-supplemented (CDE) diet–inducedinjury and re-generation.

Materials and Methods

Animal Experimentation. Fn14 KO mice and re-spective inbred wildtype (WT) animals were used aspreviously described.16 For all other experiments,C57BL/6 mice were obtained from ARC (WesternAustralia). Four-week-old, male mice were adminis-tered a CDE diet.15 Control animals received normallaboratory chow and drinking water. For TWEAKadministration experiments (single injection), recombi-nant human tumor necrosis factor–like weak inducerof apoptosis (rhTWEAK; Biogen Idec, Inc., Cam-bridge, MA17) or vehicle (0.1% bovine serum albuminin normal saline) was injected intraperitoneally into 2-week-CDE-pretreated mice at a final dose of 0.02 lg/g of body weight. All procedures were performed instrict compliance with the guidelines set by the

Address reprint requests to: Janina E. E. Tirnitz-Parker, Fremantle Hospital, T-Block Level 7, Alma Street, Fremantle 6160 WA, Australia. E-mail: [email protected]; fax: þ61-(0)8-9431-2977.Copyright VC 2010 by the American Association for the Study of Liver Diseases.Published online in Wiley InterScience (www.interscience.wiley.com).DOI 10.1002/hep.23663Potential conflict of interest: Nothing to report.Additional Supporting Information may be found in the online version of this article.

292 TIRNITZ-PARKER ET AL. HEPATOLOGY, July 2010

National Health and Medical Research Council ofAustralia and the University of Western Australia Ani-mal Ethics Committee.Primary LPC Isolates. Livers of 2-week-CDE-

treated mice were perfused, digested, and enriched forLPCs by discontinuous Percoll gradient centrifugation,as described previously.18

Cell Lines and Cell Culture. The murine, bipoten-tial LPC lines—bipotential murine oval liver (BMOL)and bipotential murine oval liver–tyrosine aminotrans-ferase (BMOL-TAT)—were established from CDE-treated livers.18 Both lines were maintained as previ-ously reported unless otherwise indicated. For prolifer-ation experiments, serum and growth factor concentra-tions in the media were reduced to 2% fetal bovineserum (FBS), 15 ng/mL insulin-like growth factor II(IGF-II), 10 ng/mL epidermal growth factor (EGF),and 5 lg/mL insulin for three consecutive passagesprior to use. Transfection was performed on 70% to80% confluent cultures with either HiPerfect or Super-Fect reagents (Qiagen, United States) according to themanufacturer’s instructions.Small Interfering RNA (siRNA) Knock-

down. Knockdown of NFjB p50 was performedunder RNA interference (RNAi) conditions as previ-ously described19 with commercial siRNAs (Qiagen)and a human/mouse siRNA starter kit (Qiagen).Cell Proliferation. Proliferation after exposure to

rhTWEAK (Biogen Idec) was assessed in vitro in a me-dium containing 0.5% FBS, 15 ng/mL IGF-II, 10 ng/mL EGF, and 5 lg/mL insulin and was quantifiedwith the colorimetric 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay,20 theCellscreen system,21 or an in vitro 5-bromo-2-deoxyuri-dine (BrdU) incorporation method (Calbiochem, UnitedStates). In situ, proliferation was evaluated via staining forKi67 (Dako, United States), as previously described.19,22

NFjB Assay. The activity of NFjB in cell lines wasmeasured with a green fluorescent protein (GFP) re-porter transactivation assay (SABiosciences, UnitedStates) using the recommended methodology. After a6-hour incubation step with the transfection com-plexes, cells were treated with rhTWEAK (BiogenIdec) for 24 hours, and GFP expression was quanti-tated fluorometrically and by fluorescence microscopy.Fluorescence-Activated Cell Sorting (FACS) of

Liver Leukocytes. Leukocytes were prepared fromphosphate-buffered saline–perfused livers as previouslydescribed.23 Cell suspensions were incubated withmonoclonal antibodies [CD3-allophycocyanin, CD4-phycoerythrin (PE), CD8–fluorescein isothiocyanate,natural killer 1.1 (NK1.1)–PE, NK1.1–phycoerythrin

cyanine 7 (PECy7), CD11b-PECy7, and CD11b-PE;all from BD Pharmingen, United States] at 1:50 dilu-tions for 30 minutes on ice, washed with an FACSwash, and sorted with a BD FACSVantage cell sorter(BD Biosciences). Cell duplets and propidium iodide–positive cells were excluded. The population of viableleukocytes was subdivided into CD3þ NK1.1� cells(T cells), CD3þNK1.1þ cells [natural killer T (NKT)cells], and CD3�NK1.1þ cells (NK cells). T cells werefurther separated into CD4þ or CD8þ cells. NK,NKT, CD4þ, and CD8þ T cells were collected inFBS. Macrophages were identified as CD3�

NK1.1�CD11bbright cells. About 10,000 cells per pop-ulation (91%-99% pure) were collected for RNAextraction and real-time polymerase chain reactionanalysis of TWEAK mRNA expression.Gene Expression Analysis. Total RNA was extracted

with TRIzol (Invitrogen, Australia) according to themanufacturer’s instructions and was quantified by spec-trophotometry, and 1 lg of total RNA was reverse-transcribed with the ThermoScript system (Invitrogen).Quantitative polymerase chain reaction (qPCR) wasperformed with Quantitect (Qiagen) or the LightCy-cler 480 system (Roche Diagnostics, United States),and fluorescent output was monitored with the RG-3000 thermal cycler (Corbett Research, Australia) orLightCycler 480, respectively. The qPCR results werenormalized against a housekeeping control gene (pepti-dylpropyl isomerase A, b-actin, or TATA box bindingprotein associated factor 4a) and expressed with respectto controls. Primer sequences are available upon request.Immunofluorescence. Immunofluorescence was per-

formed on methanol/acetone-fixed liver cryosections,cytospins of BMOL cell lines, or primary LPC iso-lates18 according to a standard protocol with antibod-ies diluted in REAL antibody diluent (Dako, Den-mark), and mounting was accomplished with theProLong Gold antifade reagent with 40,6-diamidino-2-phenylindole (DAPI; Invitrogen) for nuclear quantifi-cation. The primary antibodies were rat anti-A6(1:100, Valentina Factor, Bethesda, MD), rat anti–cy-tokeratin 19 (anti-CK19; TROMA-3, 1:200, RolfKemler, Freiburg, Germany), rabbit anti–E-cadherin(24E10, 1:125, Cell Signaling, United States), rat anti-CD45 (Ly-5, 1:200, BD Pharmingen), rat anti-F4/80(1:150, Ruth Ganss, Perth, Australia), rabbit anti–pancytokeratin (anti-CKpan; 1:400, Dako), rat anti-Ki67(TEC-3, 1:100, Dako), rabbit anti-NFjB p65(C22B4, 1:30, Cell Signaling), rat anti-CD68 (FA-11,1:200, Abcam, United States), mouse anti–alphasmooth muscle actin (anti-aSMA; ASM-1, 1:400,Chemicon, United States), and chimeric human Fc/

HEPATOLOGY, Vol. 52, No. 1, 2010 TIRNITZ-PARKER ET AL. 293

mouse Fab anti-mouse Fn14 (chimP2D3, 2 lg/mL, orchimP4A8, 0.5 lg/mL, Biogen Idec). Primary antibod-ies were detected with goat anti-rabbit/rat Alexa Fluor488, goat anti-rat/rabbit/mouse Alexa Fluor 594(1:200, Invitrogen) or biotinylated goat anti-human(1:400, Vector Laboratories, United States), and strep-tavidin-PE (1:200, BD Biosciences) or streptavidin–Alexa Fluor 488 (1:800, Invitrogen). Stainings werequantified by either manual cell counting or with theAnalySIS Life Science Professional program (Olympus,Australia).Sirius Red Staining. Carnoy’s solution–fixed and

paraffin-embedded livers were sectioned, dewaxed, andrehydrated in distilled water before nuclear stainingwith hematoxylin, washing in distilled water, and 30-minute staining with 0.1% Sirius red in saturatedaqueous picric acid.Statistical Analyses. Data are presented as means

and standard errors of the mean (SEMs). Correlationwas assessed by linear regression with PRISM (Graph-Pad Software, Inc., San Diego, CA). Statistical signifi-cance was assessed with one-way analysis of varianceanalysis and the Tukey posttest, Student t test, orMann-Whitney U test (where applicable); this wasfacilitated with GraphPad InSat version 3.0b for Mac-intosh (GraphPad Software).

Results

Fn14 and TWEAK Expression in the Liver. Real-time polymerase chain reaction analysis of whole liverextracts showed low levels of TWEAK mRNA in allsamples, with no significant difference between experi-mental groups (Fig. 1A). In both normal and CDE-injured livers, TWEAK was primarily expressed by nat-ural killer (NK) cells and macrophages (Fig. 1B). Fn14mRNA expression levels were substantially increased inCDE livers by approximately 6-, 17-, and 11-fold with1, 2, and 3 weeks of CDE treatment, respectively (Fig.1C). The correlation of Fn14 mRNA with LPC num-bers (quantified by A6 staining and cell counting of 1-to 3-week CDE livers, as previously described)22 pro-duced a significant (P < 0.0001) positive correlationwith a coefficient of 0.9417 (Fig. 1D). Furthermore,primary nonparenchymal cell isolates from a 2-weekCDE liver had proportions of cells positive for theLPC/biliary marker A6 (36%) and the LPC/epithelialcell marker E-cadherin (34%) similar to those forFn14 (36%; not shown). Double immunofluorescencestaining of these isolates illustrated coexpression ofanother accepted LPC/biliary marker, CKpan24 (Fig.

1E,G), and Fn14 (Fig. 1F,G) in all CKpanþ cells. Todetermine the phenotype of the small number ofCKpan�/Fn14þ cells, additional nonparenchymal cellmarkers were tested. Of these, only aSMA showed def-inite cellular colocalization with Fn14 in a small pro-portion of aSMAþ cells (Fig. 1H). Inflammatory cells(CD68þ, CD45þ) did not express Fn14 (SupportingFig. 1).LPC Expansion Is Reduced in CDE-Fed Fn14

Null Mice. The absence of Fn14 abrogated the linearexpansion of LPCs in response to CDE diet–inducedinjury, and this led to a significant reduction in LPCnumbers at 2 but not 3 weeks of treatment. Quantita-tion of sections stained with the biliary LPC markersA6 (Fig. 2A,D-I) and CK19 (Fig. 2B and SupportingFig. 2) or the epithelial cell marker E-cadherin (Fig.2C and Supporting Fig. 2), which was recently recog-nized as a marker for identification of LPCs in chroni-cally injured livers,25 illustrated that this effect is inde-pendent of the marker used.Inflammation and Fibrosis Correlate with

Reduced LPC Expansion in Fn14 Null Mice. Ourgroup and others have previously described coregula-tion of inflammatory and fibrogenic responses withLPC proliferation in the CDE model.4,22,26,27 Thus,we wished to determine whether the reduction of LPCproliferation observed in Fn14 null mice correlatedwith these other important regenerative pathologies.Numbers of both CD45þ leukocytes and F4/80þ mac-rophages were significantly reduced in Fn14 KO liversat 2 but not 3 weeks (Fig. 3A,B). In accordance,expression of the key inflammatory cytokines TNF(Fig. 3C), interleukin 6 (IL-6; Fig. 3D), IFNc (Fig.3E), and LTb (Fig. 3F) also fell at 2 weeks. To assessfibrogenesis, livers were stained with Sirius red, andimages were digitally quantified. Collagen deposition(Fig. 4A-C) and collagen-1 mRNA levels (Fig. 4E)correlated with LPC numbers in WT and Fn14 KOmice (Fig. 2), as observed in previous studies.4 CDE-induced mRNA expression of tissue inhibitor of met-alloproteinases 1 (TIMP-1; Fig. 4F) and TIMP-2(Fig. 4G), but not SRY (sex determining region Y)-box 9 (SOX-9; Fig. 4D), matrix metalloproteinase 2(MMP-2; Fig. 4H), or MMP-9 (Fig. 4I), wasreduced in Fn14-deficient animals at both 2 and 3weeks.TWEAK Is a Direct Mitogen to LPCs In

Vitro. The clonal LPC lines BMOL and BMOL-TAT18 were confirmed to be strongly positive for theTWEAK receptor Fn14 in the majority of cells (Fig.5A). Treatment of BMOL and BMOL-TAT cellswith rhTWEAK led to activation of the transcription


factor NFjB, which was an expected outcome afterstimulation of Fn14.11,12 NFjB activation was dose-dependent, however, only at lower concentrations(Fig. 5B).Stimulation of growth factor- and serum-starved

cells with rhTWEAK induced proliferation in a dose-dependent and time-dependent fashion, as determinedby Cellscreen analysis (Fig. 5C) and confirmed byMTT assay (not shown). In agreement with the NFjBactivation data, this effect was also dose-dependentonly at doses up to 50 ng/mL (Fig. 5D).

TWEAK Stimulates LPC Proliferation In Vivo. Toaddress whether TWEAK can stimulate LPC prolif-eration in vivo, we administered rhTWEAK to micewith a preexisting LPC response induced by 2 weeksof CDE feeding and then measured proliferationin situ, 24 or 48 hours later, by staining for the prolif-erating cell marker Ki67. Hepatocyte proliferationdid not differ between placebo and rhTWEAK-injected groups at either time point. In contrast, 24hours after rhTWEAK administration, a 2-foldincrease was observed in the number of Ki67-labeled

Fig. 1. TWEAK and Fn14 expres-sion in normal and regenerating liv-ers. (A) qPCR analysis of the wholeliver for TWEAK showed no signifi-cant change in the expression ofTWEAK mRNA in CDE diet–fed miceversus control diet–fed controls (n¼ 6). (B) The isolation of liver leu-kocytes by flow cytometry and sub-sequent TWEAK qPCR showed thatNK cells and macrophages werethe major sources of TWEAK inboth control liver and CDE liver (n¼ 2), with each replicate repre-senting the combined RNA fromtwo mice. (C) In contrast toTWEAK, whole liver Fn14 mRNAlevels were massively increased inCDE-treated mice versus healthyanimals (n ¼ 3). (D) Linear regres-sion analysis of the CDE liver ondays 7, 14, and 21 showed thatFn14 mRNA levels correlated withthe average number of A6þ LPCspresent in the same liver (n ¼ 3).Data represent means 6 SEM,with statistical significance repre-sented as *P < 0.05, **P <0.01, and ***P < 0.001. Thedashed line signifies mRNA levelsof control diet–fed WT mice. (E)Percoll gradient isolates of primaryLPCs from 14-day CDE liver con-tained clusters of cells positive forCKpan. (F,G) These clusters werealmost always copositive for Fn14.(H) Double staining for Fn14 andaSMA showed both Fn14� (toppanel) and Fn14þ (bottom panel)myofibroblasts.


nonparenchymal cells. By 48 hours post-rhTWEAKdelivery, the proliferative wave had ended (Fig. 6A).Morphologically, many cells of the Ki67-labeled, non-parenchymal cell fraction had the appearance of LPCsand were located periportally. In accordance, doublestaining for CKpan24 and Ki67 illustrated that the ma-jority of nonparenchymal cells stimulated by rhTWEAKinjection were LPCs (Fig. 6B,C). Immunofluorescentstaining for leukocytes revealed a time-dependentincrease in CD45þ inflammatory cell numbers inrhTWEAK-injected animals compared to placebo-treated animals (Fig. 6D), and this is consistent with thepreviously demonstrated coregulation of the inflamma-tory response with LPC proliferation (Fig. 3).NFjB Mediates TWEAK/Fn14-Induced LPC Pro-

liferation. Regenerating CDE livers showed theexpected up-regulation of NFjB signaling in compari-

son with WT controls. Interestingly, numerous cellsstrongly positive for cytoplasmic and nuclear (active)NFjB were localized in periportal liver regions. Dou-ble staining of sections for A6 and NFjB demon-strated that many NFjBþ cells were represented byLPCs (Fig. 6E-G); this suggests that this transcriptionfactor mediates the pro-proliferative effects of TWEAKin the CDE model. To test this hypothesis, RNAi wasused to knock down the p50 NFjB subunit in theLPC lines BMOL and BMOL-TAT (which were iso-lated from a CDE-treated liver) before measurementof the mitotic S phase by BrdU incorporation. Trans-fection of p50 siRNA reduced TWEAK-induced LPCproliferation in a dose-dependent fashion. Dose-de-pendent proliferation of LPCs following TWEAKstimulation was not affected by transfection of ascrambled control siRNA vector (Fig. 6H).

Fig. 2. LPC expansion is attenuated in Fn14-deficient animals. Fn14 KO or WT mice were subjected to a CDE diet to induce hepatocyte injuryand stimulate the regenerative response of LPCs. After 2 weeks on the CDE diet, LPC numbers in WT livers were significantly increased versuscontrol diet. In contrast, Fn14 KO mice exhibited only a modest enrichment of LPCs after 14 days of CDE treatment. Interestingly, however, thenumbers of LPCs in the Fn14 KO livers normalized within 3 weeks on the diet, and they did not differ from the WT response at the 21-day timepoint. This pattern of LPC induction was evident, regardless of whether (A,D-I) A6 (A6 in red and DAPI for nuclear quantitation in blue), (B)CK19 (see Supporting Fig. 2), or (C) E-cadherin (see Supporting Fig. 2) was used to quantify the response. Data represent means 6 SEM (n ¼3-4). Statistical significance is represented as *P < 0.05. The dashed line signifies cell counts of control diet–fed WT mice.


Discussion

Once considered by many to be little more than ahistological artifact, adult LPCs (oval cells) are nowthe subject of much attention because of their poten-tial use as vectors or targets for cell or drug-based ther-apy28 and their clear association with chronic hepaticdisease progression.6,19,29 With respect to the former,it is interesting to note that despite considerableefforts, no LPC-specific growth or differentiation fac-tors have been identified. The characterization ofTWEAK as a mitogen for LPCs not only is importantfor our understanding of how regenerative responsesare mediated in chronic damage but also draws us onestep closer to achieving selective modulation of LPCbehavior for the purposes of cell therapy and regenera-tive medicine.The study by Jakubowski et al.8 showed definitively

that the TWEAK ligand and its cognate receptor Fn14mediate the behavior of at least a subset of LPCs. Thiswas demonstrated by several approaches, including (1)gene-targeted mice, which showed a reduced presenceof A6-expressing cells in an experimental model ofductular reaction, and (2) increased numbers of LPCsin transgenic mice, which overexpressed TWEAK.LPCs are recognized as being highly heterogeneousand dynamic, and several previous studies from ourlaboratory have illustrated selective modulation of dis-crete subpopulations of LPCs by cytokines.3,30,31

Because the LPC compartment induced in variousliver injury models may be different, we aimed to fur-ther explore the role of the TWEAK pathway in anadditional in vivo model of LPC expansion and tostudy its effects directly on LPC lines in vitro.Using the CDE model of hepatocellular injury and

regeneration, which our laboratory pioneered for usein mice,15 we now show that TWEAK signaling via itsreceptor Fn14 is essential for the early expansion ofLPCs in response to CDE-induced liver injury. In liv-ers from CDE diet–fed mice, LPCs proliferate outwardfrom periportal regions, not in ducts or pseudoducts,but as streams of cells. We have extensively studied thedynamics of this response and shown that the LPCproliferation in this model has two phases: lineargrowth and maintenance. Linear growth beginsbetween days 3 and 5 and is completed around days12 to 16. During this phase, LPCs have a doublingtime of approximately 18 hours, a cycling time typicalof a stem or progenitor cell.32 At the end of the lineargrowth phase, LPC numbers plateau, and only smallnumbers of LPCs are seen dividing (B. Knight,unpublished observations, 2009). Thus, it is interestingthat Fn14 KO mice displayed significantly fewer LPCson day 14 but not day 21; this suggests that TWEAKis important in mediating the rapid expansion of theLPC pool immediately after the onset of liver damageand is less significant in affecting their slow turnover

Fig. 3. Inflammation is attenuated in Fn14-deficient animals. (A,B) Quantification of leukocytes (CD45þ) or macrophages (F4/80þ) in WT liv-ers showed a modest (1.5- to 2-fold) enrichment in both cell types after 2 or 3 weeks of CDE feeding. This increase was attenuated in Fn14 KOanimals at 14 days but rebounded by 21 days. Accordingly, mRNA expression of the proinflammatory cytokines (C) TNF, (D) IL-6, (E) IFNc, and(F) LTb, which increased in response to the CDE diet in WT mice, was suppressed in animals lacking Fn14, but again only at 2 weeks. Data rep-resent means 6 SEM (n ¼ 3-4). Statistical significance is represented as *P < 0.05. The dashed line signifies cell counts or mRNA levels ofcontrol diet–fed WT mice.


in the latter phase of the diet response. One interpreta-tion is that the linear LPC expansion phase is delayedby the lack of TWEAK signaling, and other pathwayscompensate subsequently. Many cytokines orchestratethe LPC response, and it is therefore also not surpris-ing that a deficiency of a single cytokine inhibits butdoes not fully block LPC proliferation, as has beenshown for CDE-fed mice deficient for TNFR1, TNF/LTa, IFNc, or LTb.3,5,6 In this respect, it is importantto note that Jakubowski et al.8 also used the 14-daytime point to examine the response of Fn14 KO miceto 3,5-diethoxycarbonyl-1,4-dihydrocollidine–inducedinjury. Preliminary data detailing the dynamics of theLPC response in this model suggest that rapid growthalso occurs between approximately days 5 and 16(B. Knight, unpublished observations, 2009), withpeak Ki67 labeling of LPCs on day 12 (Boulter et al.,

unpublished data, 2009). In our in vivo studies, wehave not detected an increase in TWEAK mRNAexpression in response to injury. Expression ofTWEAK in CDE cell isolates was predominantly inNK cells and macrophages, which are known to associ-ate physically with proliferating LPCs in CDE andother models.33 In contrast, Fn14 expression was mas-sively up-regulated in CDE samples and correlatedpositively with LPC numbers. In primary cell isolatesfrom a CDE liver, all cells positive for the LPC/biliarymarker CKpan coexpressed Fn14. We initially sus-pected that the small proportion of CKpan�/Fn14þ

cells were LPCs reflecting a more hepatocytic geneexpression profile and thus negative for the biliaryantigen CKpan. However, double staining for othernonparenchymal cell markers with Fn14 yielded anunexpected result: the myofibroblast marker aSMA

Fig. 4. Fibrosis correlates with the LPC response in Fn14-deficient animals. (A,B) Sirius red staining and (C) digital quantification showed aspatiotemporal pattern of induction similar to that seen previously with LPC numbers. More specifically, after 2 weeks on the CDE diet, collagendeposition was clearly evident in the periportal areas, where LPC expansion also took place. (E) Associated with this was an increase in expres-sion in collagen-1 mRNA. Additional fibrogenic mediator mRNA expression levels were also increased in the WT CDE liver compared to WT con-trols: (D) SOX-9, (F) TIMP-1, (G) TIMP-2, and (H) MMP-2 but not (I) MMP-9. Fn14-deficient mice showed a clear attenuation of fibrosis at 2weeks as measured by (B,C) Sirius red, (E) collagen-1, and (F,G) suppression of TIMP expression. (C,E-G,I) By 3 weeks, this response appearedto have partially rebounded but remained mostly lower than that of WT mice. Data represent means 6 SEM (n ¼ 3-4). Statistical significance isrepresented as *P < 0.05 and **P < 0.01. The dashed line signifies mRNA levels of control diet–fed WT mice.


colocalized with Fn14 in a subset of cells. The impli-cation of this, at this time, is not entirely clear as a lin-eage relationship between LPCs and hepatic stellatecells has been proposed.34 Regardless, we propose thata paracrine TWEAK-Fn14 signaling system operates inLPC-mediated liver regeneration, in which localizedsecretion of the TWEAK ligand by NK cells and mac-rophages regulates LPC behavior by binding and acti-vating Fn14 on LPCs.Another novel finding from our study is the coregu-

lation of inflammation and collagen deposition withLPC responses in TWEAK signaling–deficient mice.Previous studies have suggested that these responses gohand in hand,4,22,35,36 and again in the present study,we have robustly demonstrated this correlation. Wepreviously hypothesized that inflammatory infiltrationprecedes the progenitor cell response in CDE-fedmice,32 and we more recently proposed that initiationof LPC proliferation stimulates activation of hepaticstellate cells and thus matrix remodeling.4 Interestingly,Van Hul et al.26 disagree with our proposed sequence;they have suggested that the fibrosis response occursbefore LPC expansion. In CDE-fed Fn14 null mice,the hepatic inflammatory response was delayed, andthis is consistent with reduced inflammation in cardio-toxin-treated Fn14 null animals in a model of skeletalmuscle regeneration.16 Collagen deposition correlatedwith the LPC response, with a reduction evident at 2but not 3 weeks of CDE feeding. If TWEAK is adirect and selective mitogen for LPCs, this implies

that fibrosis follows the LPC response and not theother way around. However, as inflammation was core-gulated in the Fn14 KO mice, it is not possible fromthis alone to resolve the relative sequence of the twoevents.To demonstrate that TWEAK is a definitive LPC

mitogen, development of an in vitro system was essen-tial. In this respect, it is important to note thatJakubowski et al.8 used a well-differentiated biliary epi-thelial cell line. We have previously established immor-talized LPC lines from WT or transgenic CDE-treatedmouse livers and characterized as bipotential murineoval liver (BMOL) cells.18 As early as 24 hours afterTWEAK exposure, growth of the BMOL lines wasnotably stimulated; this continued for up to 6 dayswith no plateau apparent. Proliferation was dose-de-pendent up to 50 ng/mL rhTWEAK; at 100 ng/mL,however, growth was slightly reduced, as was NFjBactivation. RNAi experiments have demonstrated thecentral role of NFjB in mediating Fn14 signaling.When TWEAK binds Fn14, it recruits adaptor mole-cule TRAF2 (or TRAF5), which activates the NFjBsignaling complex and thus results in a long-lastingprosurvival, pro-proliferative response.12 In accordancewith an important role for NFjB in mediating LPCexpansion, we found numerous cells expressing bothinactive (cytoplasmic) and active (nuclear) NFjB inCDE-treated livers. Considering all this, we suggestthat TWEAK signaling, via Fn14 to NFjB, mediatesthe rapid growth of LPCs in the first 14 days of CDE

Fig. 5. LPCs are induced to pro-liferate by TWEAK. (A) The murineLPC lines BMOL and BMOL-TAT (iso-lated from a chronically injured CDEliver) were confirmed to be TWEAK-responsive by strong positivity forthe receptor Fn14 in the vast major-ity of cells. (B) Treatment of BMOLcells with rhTWEAK induced activa-tion of NFjB (as judged by a GFPreporter assay) in a dose-dependentfashion up to 10 ng/mL. (C,D)Accordingly, proliferation, as deter-mined by either Cellscreen or MTTassay (not shown), was also dose-dependently induced by TWEAK, butagain only at the lower end of theconcentration ranges. Data repre-sent means 6 SEM (n ¼ 4).


Fig. 6. NFjB mediates the pro-proliferative effects of TWEAK on LPCs in vivo and in vitro. (A) The injection of recombinant TWEAK into mice withan existing LPC response, which was invoked by 2 weeks of CDE feeding, caused an increase in nonparenchymal (NP) cell proliferation 24 hours afterinjection but not 48 hours after injection. Hepatocytic cell proliferation (Heps) was not affected. (B,C) Double immunofluorescent staining showedthat the majority of proliferating NP cells were CKpanþ LPCs. (D) CD45þ inflammatory cell numbers were significantly increased 48 hours afterTWEAK treatment in comparison with placebo-injected controls. Staining of CDE liver sections 24 hours after TWEAK treatment for (E) the LPC/biliarymarker A6 and (F) NFjB showed (G) numerous LPCs with active (nuclear) NFjB. (H) Knockdown of NFjB by RNAi reversed the pro-proliferativeeffects of TWEAK on LPC lines in vitro. Data represent means 6 SEM (n ¼ 4-6). Statistical significance is represented as *P < 0.05.


feeding. These findings are important in the context ofmanipulating cells for possible therapeutic applicationsand for future experiments investigating the pro-prolif-erative and anti-proliferative effects of growth factorsand cytokines on LPCs in vitro and in vivo.In our model, we propose that local production of

TWEAK by NK cells and macrophages mediates theexpansion of LPCs in chronic injury models, but it issignificant that we can show that delivery of exogenousrhTWEAK induces proliferation of periportal CKpan-expressing nonparenchymal cells showing LPC mor-phology. We detected a mean doubling in proliferatingnonparenchymal cells 24 hours after TWEAK adminis-tration. By 48 hours, numbers had returned to the base-line. The majority of Ki67þ nonparenchymal cells(�60%) were LPCs (as judged by their CKpan positiv-ity24). It should be emphasized, however, that this is aconservative estimate, as we have shown previously thatin the CDE model, only a subset of LPCs is A6/CKpan-positive.18 It is therefore likely that some of the Ki67þ/CK� cells are nonbiliary LPCs. This suggests that itmight be possible to augment LPC-mediated liverregeneration with TWEAK, and there are implicationsfor its use in patients with chronic liver disease. How-ever, such studies need to proceed with care, given ourfindings that a subset of myofibroblasts is also Fn14þ

and thus may also respond to TWEAK stimulation.In summary, demonstrating definitively that

TWEAK is a mitogen for LPCs in vitro and in vivo,this study has confirmed and extended the previouswork of Jakubowski et al.8 On the basis of concurrentanalysis of isolated and in situ LPCs, we conclude thatthe TWEAK signaling system is activated in thechronically injured liver by up-regulated expression ofthe Fn14 receptor on existing and newly producedLPCs. TWEAK protein is produced and secreted byneighboring NK cells and macrophages and, oncebound to Fn14, activates NFjB and, in turn, expres-sion of numerous cell cycle regulatory factors.

Acknowledgments: The authors are grateful to BenDwyer for technical assistance and to Dr. ValentinaFactor, Associate Professor Ruth Ganss, and ProfessorRolf Kemler for providing antibodies for this study.

References1. Fausto N, Campbell JS, Riehle KJ. Liver regeneration. HEPATOLOGY

2006;43:S45-S53.

2. Bird TG, Lorenzini S, Forbes SJ. Activation of stem cells in hepaticdiseases. Cell Tissue Res 2008;331:283-300.

3. Akhurst B, Matthews V, Husk K, Smyth MJ, Abraham LJ, Yeoh GC.Differential lymphotoxin-beta and interferon-gamma signaling during

mouse liver regeneration induced by chronic and acute injury. HEPATO-

LOGY 2005;41:327-335.

4. Knight B, Lim R, Yeoh GC, Olynyk JK. Interferon-gamma exacerbatesliver damage, the hepatic progenitor cell response and fibrosis in amouse model of chronic liver injury. J Hepatol 2007;47:826-833.

5. Knight B, Yeoh GC. TNF/LTalpha double knockout mice displayabnormal inflammatory and regenerative responses to acute and chronicliver injury. Cell Tissue Res 2005;319:61-70.

6. Knight B, Yeoh GC, Husk KL, Ly T, Abraham LJ, Yu C, et al.Impaired preneoplastic changes and liver tumor formation in tumor ne-crosis factor receptor type 1 knockout mice. J Exp Med 2000;192:1809-1818.

7. Nguyen LN, Furuya MH, Wolfraim LA, Nguyen AP, Holdren MS,Campbell JS, et al. Transforming growth factor-beta differentially regu-lates oval cell and hepatocyte proliferation. HEPATOLOGY 2007;45:31-41.

8. Jakubowski A, Ambrose C, Parr M, Lincecum JM, Wang MZ, ZhengTS, et al. TWEAK induces liver progenitor cell proliferation. J ClinInvest 2005;115:2330-2340.

9. Chicheportiche Y, Bourdon PR, Xu H, Hsu YM, Scott H, Hession C,et al. TWEAK, a new secreted ligand in the tumor necrosis factor familythat weakly induces apoptosis. J Biol Chem 1997;272:32401-32410.

10. Meighan-Mantha RL, Hsu DK, Guo Y, Brown SA, Feng SL, PeifleyKA, et al. The mitogen-inducible Fn14 gene encodes a type I trans-membrane protein that modulates fibroblast adhesion and migration.J Biol Chem 1999;274:33166-33176.

11. Brown SA, Richards CM, Hanscom HN, Feng SL, Winkles JA. TheFn14 cytoplasmic tail binds tumour-necrosis-factor-receptor-associatedfactors 1, 2, 3 and 5 and mediates nuclear factor-kappaB activation.Biochem J 2003;371:395-403.

12. Saitoh T, Nakayama M, Nakano H, Yagita H, Yamamoto N, YamaokaS. TWEAK induces NF-kappaB2 p100 processing and long lastingNF-kappaB activation. J Biol Chem 2003;278:36005-36012.

13. Burkly LC, Michaelson JS, Hahm K, Jakubowski A, Zheng TS.TWEAKing tissue remodeling by a multifunctional cytokine: role ofTWEAK/Fn14 pathway in health and disease. Cytokine 2007;40:1-16.

14. Feng SL, Guo Y, Factor VM, Thorgeirsson SS, Bell DW, Testa JR,et al. The Fn14 immediate-early response gene is induced during liverregeneration and highly expressed in both human and murine hepato-cellular carcinomas. Am J Pathol 2000;156:1253-1261.

15. Akhurst B, Croager EJ, Farley-Roche CA, Ong JK, Dumble ML,Knight B, et al. A modified choline-deficient, ethionine-supplementeddiet protocol effectively induces oval cells in mouse liver. HEPATOLOGY

2001;34:519-522.

16. Girgenrath M, Weng S, Kostek CA, Browning B, Wang M, Brown SA,et al. TWEAK, via its receptor Fn14, is a novel regulator of mesenchy-mal progenitor cells and skeletal muscle regeneration. EMBO J 2006;25:5826-5839.

17. Jakubowski A, Browning B, Lukashev M, Sizing I, Thompson JS,Benjamin CD, et al. Dual role for TWEAK in angiogenic regulation.J Cell Sci 2002;115:267-274.

18. Tirnitz-Parker JE, Tonkin JN, Knight B, Olynyk JK, Yeoh GC. Isola-tion, culture and immortalisation of hepatic oval cells from adult micefed a choline-deficient, ethionine-supplemented diet. Int J BiochemCell Biol 2007;39:2226-2239.

19. Knight B, Tirnitz-Parker JE, Olynyk JK. C-kit inhibition by imatinibmesylate attenuates progenitor cell expansion and inhibits liver tumorformation in mice. Gastroenterology 2008;135:969-979.

20. Denizot F, Lang R. Rapid colorimetric assay for cell growth and sur-vival. Modifications to the tetrazolium dye procedure giving improvedsensitivity and reliability. J Immunol Methods 1986;89:271-277.

21. Viebahn CS, Tirnitz-Parker JE, Olynyk JK, Yeoh GC. Evaluation ofthe ‘‘Cellscreen’’ system for proliferation studies on liver progenitorcells. Eur J Cell Biol 2006;85:1265-1274.

22. Knight B, Akhurst B, Matthews VB, Ruddell RG, Ramm GA,Abraham LJ, et al. Attenuated liver progenitor (oval) cell and fibrogenicresponses to the choline deficient, ethionine supplemented diet in theBALB/c inbred strain of mice. J Hepatol 2007;46:134-141.


23. Bertolino P, Heath WR, Hardy CL, Morahan G, Miller JF. Peripheraldeletion of autoreactive CD8þ T cells in transgenic mice expressingH-2Kb in the liver. Eur J Immunol 1995;25:1932-1942.

24. Kofman AV, Morgan G, Kirschenbaum A, Osbeck J, Hussain M, SwensonS, et al. Dose- and time-dependent oval cell reaction in acetaminophen-induced murine liver injury. HEPATOLOGY 2005;41:1252-1261.

25. Ueberham E, Aigner T, Ueberham U, Gebhardt R. E-cadherin as areliable cell surface marker for the identification of liver specific stemcells. J Mol Histol 2007;38:359-368.

26. Van Hul NK, Abarca-Quinones J, Sempoux C, Horsmans Y, LeclercqIA. Relation between liver progenitor cell expansion and extracellularmatrix deposition in a CDE-induced murine model of chronic liverinjury. HEPATOLOGY 2009;49:1625-1635.

27. Ruddell RG, Knight B, Tirnitz-Parker JE, Akhurst B, Summerville L,Subramaniam VN, et al. Lymphotoxin-beta receptor signaling regulateshepatic stellate cell function and wound healing in a murine model ofchronic liver injury. HEPATOLOGY 2009;49:227-239.

28. Knight B, Matthews VB, Olynyk JK, Yeoh GC. Jekyll and Hyde:evolving perspectives on the function and potential of the adult liverprogenitor (oval) cell. Bioessays 2005;27:1192-1202.

29. Lowes KN, Brennan BA, Yeoh GC, Olynyk JK. Oval cell numbers inhuman chronic liver diseases are directly related to disease severity. AmJ Pathol 1999;154:537-541.

30. Davies RA, Knight B, Tian YW, Yeoh GC, Olynyk JK. Hepatic ovalcell response to the choline-deficient, ethionine supplemented model of

murine liver injury is attenuated by the administration of a cyclo-oxy-genase 2 inhibitor. Carcinogenesis 2006;27:1607-1616.

31. Lim R, Knight B, Patel K, McHutchison JG, Yeoh GC, Olynyk JK.Antiproliferative effects of interferon alpha on hepatic progenitor cellsin vitro and in vivo. HEPATOLOGY 2006;43:1074-1083.

32. Knight B, Matthews VB, Akhurst B, Croager EJ, Klinken E, AbrahamLJ, et al. Liver inflammation and cytokine production, but not acutephase protein synthesis, accompany the adult liver progenitor (oval) cellresponse to chronic liver injury. Immunol Cell Biol 2005;83:364-374.

33. Zhang W, Chen XP, Zhang WG, Zhang F, Xiang S, Dong HH, et al.Hepatic non-parenchymal cells and extracellular matrix participate inoval cell-mediated liver regeneration. World J Gastroenterol 2009;15:552-560.

34. Yang L, Jung Y, Omenetti A, Witek RP, Choi S, Vandongen HM,et al. Fate-mapping evidence that hepatic stellate cells are epithelial pro-genitors in adult mouse livers. Stem Cells 2008;26:2104-2113.

35. Clouston AD, Powell EE, Walsh MJ, Richardson MM, Demetris AJ,Jonsson JR. Fibrosis correlates with a ductular reaction in hepatitis C:roles of impaired replication, progenitor cells and steatosis. HEPATOLOGY

2005;41:809-818.

36. Richardson MM, Jonsson JR, Powell EE, Brunt EM, Neuschwander-Tetri BA, Bhathal PS, et al. Progressive fibrosis in nonalcoholic steato-hepatitis: association with altered regeneration and a ductular reaction.Gastroenterology 2007;133:80-90.


Tumor necrosis factor?like weak inducer of apoptosis is a ... · Tumor Necrosis Factor–Like Weak...

Documents

Transcript of Tumor necrosis factor?like weak inducer of apoptosis is a ... · Tumor Necrosis Factor–Like Weak...