www.elsevier.com/locate/jhep

Journal of Hepatology 46 (2007) 888–898

Hyperthyroidism induces apoptosis in rat liver through activationof death receptor-mediated pathwaysq

Ashok Kumar1, Rohit A. Sinha1, Meenakshi Tiwari1, Rajesh Singh3, Takehiko Koji4,Namratta Manhas5, Leena Rastogi1, Lily Pal2, Ashutosh Shrivastava1,

Ravi P. Sahu1, Madan M. Godbole1,*

1Department of Endocrinology, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Raebareli Road, Lucknow 226 014, India2Department of Pathology, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Raebareli Road, Lucknow 226 014, India

3Indian Institute of Advance Research, Koba Village, Gandhi Nagar 382 007, India4Department of Histology and Cell Biology, Nagasaki University School of Medicine, Nagasaki, Japan

5Division of Pharmacology, Central Drug Research Institute, Lucknow 226 001, India

Background/Aims: The molecular basis of hepatic dysfunction in thyrotoxicosis is not fully understood. Here, we inves-

tigated the effect of altered thyroidal status on death receptor pathways including p75 neurotrophin receptor (p75NTR), a

member of tumor necrosis factor (TNF) receptor superfamily, in rat liver.

Methods: Hyperthyroidism was induced in Sprague–Dawley rats by daily injections of triiodothyronine in a dose of

12.5 lg/100 g body weight for 10 days.

Results: Terminal deoxynucleotide-transferase-mediated dUTP nick end labeling assay and caspase-3 activation data

confirmed apoptosis in hyperthyroid rat liver. We observed the elevated levels of death ligands, TNF-a, Fas ligand andtheir cognate receptors, TNF-receptor-1 and Fas, and 8-fold increase in caspase-8 activation in hyperthyroid rat liver

(p < 0.001). We demonstrated for the first time that hyperthyroidism elevates p75NTR levels and its ligands, pro-nerve

growth factor and pro-brain-derived neurotrophic factor, in rat liver. Further we showed that most of the apoptotic cells

in hyperthyroid liver express p75NTR. We also demonstrated that triiodothyronine administration to rats causes NF-jB

activation, but persistent exposure (10 days) to triiodothyronine deactivates NF-jB leading to sustained c-Jun N-terminal

kinase (JNK) activation.

Conclusions: This study showed that hyperthyroidism-induced apoptosis in rat liver involves the activation of death

receptor-mediated pathways, including p75NTR.� 2007 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved.

Keywords: Thyroid hormone; Hyperthyroidism; Apoptosis; Tumor necrosis factor; Pro-neurotrophins; p75NTR

0168-8278/$32.00 � 2007 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved.

doi:10.1016/j.jhep.2006.12.015

Received 7 August 2006; received in revised form 30 November 2006; accepted 3 December 2006; available online 5 February 2007q The authors who have taken part in this study declared that they have no relationship with the manufacturers of the drugs involved either in the

past or present and did not receive funding from the manufacturers to carry out their research. The authors did not receive funding from any sourceto carry out this study.

* Corresponding author. Tel.: +91 522 2668700x2368; fax: +91 522 2668017.E-mail address: madangodbole@yahoo.co.in (M.M. Godbole).Abbreviations: TH, thyroid hormone; Dwm, mitochondrial trans-membrane potential; T3, triiodothyronine; TNF, tumor necrosis factor; FasL,

Fas ligand; NTR, neurotrophin receptor; NGF, nerve growth factor; BDNF, brain-derived neurotrophic factor; HSC, hepatocyte stellate cells; TNF-R, TNF-receptor; JNK, c-Jun N-terminal kinase; AST, aspartate aminotransferase; ALT, alanine aminotransferase; TUNEL, terminal deoxynuc-leotide-transferase (TdT)-mediated dUTP nick end labeling; EDTA, ethylenediaminetetraacetic acid; ROS, reactive oxygen species; SOD, superoxidedismutase; EMSA, electrophoretic mobility shift assay.

A. Kumar et al. / Journal of Hepatology 46 (2007) 888–898 889

1. Introduction

Thyroid hormones (TH) play an important role innormal physiology of gastrointestinal function. Abnor-mal liver function in thyrotoxicosis has been docu-mented since long back [1–5]. The hepatic injuryassociated with hyperthyroidism varies from mild liverdysfunction to severe central hepatic ischemia. In mostcases, the changes in the liver are characterized bycholestasis [2], some degree of fatty infiltration, cyto-plasmic vacuolization, nuclear irregularity and hyper-chromatism in hepatocytes [6–8]. Until recently, noconclusive evidence has been provided on the natureof hepatic injury in hyperthyroidism [6–8]. Therefore,we initiated the study to understand the molecularbasis of this hepatic dysfunction and demonstratedthat severe hyperthyroidism induces apoptosis in ratliver [9].

Caspases, the key effectors of apoptosis, can beactivated by two different pathways; extrinsic andintrinsic pathways [10]. The intrinsic pathway involvesthe release of cytochrome c from mitochondria,generation of reactive oxygen species (ROS) and lossof mitochondrial trans-membrane potential (Dwm)[11]. Extrinsic or death receptor-mediated pathway istriggered by extracellular death ligands, such as thetumor necrosis factor (TNF) and Fas ligand (FasL).Binding of TNF-a and FasL to their respective recep-tors induces apoptosis through the apical caspase-8pathway [12,13]. Previously we demonstrated that tri-iodothyronine (T3) acts directly on liver mitochondriato activate intrinsic pathway of apoptosis [9]. Howev-er, the role of death receptor-mediated pathway inhyperthyroidism-induced apoptosis has not beeninvestigated.

p75 neurotrophin receptor (NTR) is a member ofthe death receptor superfamily, which can induce celldeath in neuronal as well as non-neuronal cells uponligand binding [14,15]. It has been demonstrated thatpro-forms of neurotrophins like, nerve growth factor(NGF) and brain-derived neurotrophic factor(BDNF), are more specific and high affinity ligandsfor p75NTR–sortilin receptor complex that can induceapoptosis more efficiently than their mature forms[16–20]. Though neurotrophins and their receptorsare known to express in hepatocytes and hepatocytestellate cells (HSC) [21], however their role in hepaticdysfunction under altered thyroidal status has notbeen studied.

Here, we have investigated the role of death receptorpathway specifically TNF-a/TNF-receptor-1 (TNF-R1),Fas/FasL and caspase-8, and pro-neurotrophins/p75NTR-mediated pathway in hyperthyroidism-in-duced apoptosis in rat liver.

2. Materials and methods

2.1. Antibodies

Antibodies against TNF-a, TNF-R1 and TNF-R2, caspase-8, b-ac-tin, NGF, BDNF, c-Jun N-terminal kinase (JNK) and secondary anti-bodies were purchased from Santa Cruz Biotechnology (Santa Cruz,CA). Antibodies against cleaved-caspase-3 and phosphorylated-JNKwere purchased from Cell Signaling Technology (Beverly, MA).Anti-Fas and FasL sera were prepared, as described previously [22,23].

2.2. Experimental animals

Four-month-old Sprague–Dawley male rats were used in all exper-iments. The animals were housed in a temperature- and light-con-trolled room (21–24 �C, 12-h light/dark cycles); they were allowed achow diet and water ad libitum. Rats were divided into three groups(n = 10 in each group). The first group was given chow diet and ordin-ary drinking water (euthyroid), the second group in addition was givenmethimazole (Sigma, USA) at a concentration of 0.025% (wt/vol) indrinking water for 8 weeks (hypothyroid). Unless otherwise stated,to the rats of third group, T3 (Sigma) was administered subcutaneouslyat a daily dose of 12.5 lg/100 g body weight for 10 days (hyperthyroid)[9]. All animal procedures were undertaken in accordance with theInstitutional Guidelines for Animal Care and Research.

2.3. Serum total T3, aspartate aminotransferase (AST)

and alanine aminotransferase (ALT)

Serum total T3 levels were measured by using standard RIA kits(DPC kit, Los Angeles, CA); AST and ALT were measured using stan-dard kits according to the manufacturer’s instructions (Merck GmbH,Darmstadt, Germany).

2.4. Terminal deoxynucleotide-transferase-mediated

dUTP nick end labeling (TUNEL)

Paraffin-embedded liver sections (4 lm) were processed for TUN-EL assay. TUNEL-positive cells were identified using ApoAlertDNA fragmentation assay kit (BD Bioscience, San Diego, CA) andwere counted in five randomly selected fields. Image-Pro Plus 5.1 soft-ware (Media Cybernetics Inc., Silver Spring, MD) was used for imagecapturing and cell counting.

2.5. Caspase-3 activity

Caspase-3 activity was measured in liver tissue lysates using theApoAlert Caspase-3 Colorimetric Assay Kit (BD Biosciences) follow-ing the manufacturer’s instructions.

2.6. Protein extraction and Western blotting

Liver tissues were minced in 10 volumes of homogenization buffer{10 mM Tris–Cl, (pH 7.5) 150 mM NaCl, 5 mM EDTA, 1% Triton-X100, 0.1% SDS, 10 mM sodium orthovanadate, 1 mM phenylmethyl-sulfonyl fluoride and protease inhibitor cocktail (Sigma)}, homoge-nized and centrifuged at 17,000g for 15 min. The supernatant wasused for Western blotting as described previously [9].

2.7. Immunofluorescence

Paraffin-embedded liver sections (4 lm) were processed for immu-nofluorescence studies. To colocalize p75NTR protein in apoptoticcells, the sections were incubated with rabbit polyclonal cleaved-caspase-3

890 A. Kumar et al. / Journal of Hepatology 46 (2007) 888–898

antibody and mouse monoclonal p75NTR antibody (MC192)overnight at 4 �C. After washing with PBS, the sections were incubatedwith Rhodamine Red-conjugated anti-rabbit IgG (red) and FITC-con-jugated anti-mouse IgG (green) secondary antibodies for 1 h in dark.To study the nuclear translocation of NF-jB, sections were incubatedwith a rabbit polyclonal NF-jB/p65 antibody overnight at 4 �C, fol-lowed by incubation with Rhodamine Red-conjugated anti-rabbitIgG (red) secondary antibody for 1 h. The sections were counterstainedwith Hoechst-33258 (0.5 lg/ml; a nuclear stain) and visualized under afluorescent microscope (Nikon).

2.8. Measurement of Dwm ROS, superoxide dismutase

(SOD) and catalase

The fluorescent probes JC-1 (5 lM) and 2 0,7 0-dichlorofluorescien-diacetate (DCF-DA; 50 lM) (Sigma) were used to measure the Dwm

and ROS, respectively, by flow cytometry in isolated hepatocytes[24,25]. Rat hepatocytes were isolated from euthyroid and T3-treatedrats by collagenase perfusion as described previously [26]. SODand catalase activities were measured in liver homogenate by epi-nephrine auto-oxidation and decomposition of H2O2, respectively[27,28].

2.9. Electrophoretic mobility shift assays (EMSA)

Nuclear extracts (10 lg/assay) of liver tissues were analyzed byEMSA as described previously [29,30]. The double-stranded NF-jB(5 0-AGTTGAGGG GACTTTCCCAGGC-30) oligonucleotides wereobtained from Promega (Chilworth, UK) and radiolabeled withc32P-ATP (BARC, Mumbai, India) by incubation with 10 U of T4polynucleotide kinase (Promega). After DNA-binding reaction andelectrophoresis, gels were dried and autoradiographed.

2.10. Statistical analysis

The data are presented as means ± SD. Data were compared byone-way ANOVA followed by SNK post hoc test. p value <0.05 wasconsidered statistically significant.

Fig. 1. Changes in serum biochemistry and body weight of experimental

rats. (A) Weight of rats in eu-, hyper- and hypothyroid groups (n = 6 in

each group). Data indicate mean change in body weight (%) over the time

period relative to first day of T3 administration (B and C) Serum ALT

and AST levels: blood was collected from rats at the end of experimental

period (n = 6 in each group). Data are expressed as means ± SD.

Asterisk indicates significant difference compared with euthyroid rats

(*p < 0.001).

3. Results

3.1. Hyperthyroidism elevates serum ALT and AST levels

To assess the thyroidal status of rats, circulating totalT3 levels were measured. Serum T3 levels increased �15-fold in the hyperthyroid group (p < 0.001) over thecontrol group [Serum total T3 nmol/L euthyroid =1.25 ± 0.26; hypothyroid = 0.32 ± 0.13; hyperthyroid =19.5 ± 2.47]. Body weight of hyperthyroid rats decreased14.2 ± 4.21% after 10 days of T3 treatment, whereas itremain unchanged in euthyroid and hypothyroid groups(Fig. 1A). Serum AST and ALT levels increased signif-icantly in the hyperthyroid group (p < 0.001; Fig. 1Band C) as compared to the euthyroid group.

3.2. Hyperthyroidism induces apoptosis in rat liver

To examine apoptosis in rat liver, TUNEL assaywas performed. In the hyperthyroid group, numberof apoptotic cells increased significantly as comparedto the euthyroid group (euthyroid 2.6 ± 0.52% vs.

hyperthyroid 14.7 ± 2.94%; p < 0.001; Fig. 2A and B).To investigate the involvement of caspase-3, Westernblotting was performed using activated-caspase-3antibody which recognizes the p17 and p19 subunits.In the euthyroid group, only a basal level ofcaspase-3 activation was observed which increased

Fig. 2. Hyperthyroidism induces apoptosis in rat liver. (A) Sections (4 lm) were cut from paraffin-embedded liver tissue of eu-, hyper- and hypothyroid rats,

and then TUNEL assay was performed. Scale bar in (a–c), 20 lm. (B) TUNEL-positive cells were counted in five different fields and at least 10,000 cells in

each section were scored. Proportion of apoptotic cells (%) was calculated after counting the number of TUNEL-positive cells/100 nuclei (Hoechst stained).

(C) Tissue homogenate (50 lg) from eu-, hypo- and hyperthyroid (T3 for 10 days) rat liver was run on 15% SDS–PAGE, transferred to nitrocellulose

membranes and probed with an activated-caspase-3 antibody. b-actin was used as a loading control. (D) Fold change in density of cleaved caspase-3 in hyper-

and hypothyroid rats relative to euthyroid rats. Bars represent means of fold change ±SD of five rats in each group. (E) Caspase-3 activity in rat liver. Liver

tissue lysates were prepared from eu-, hypothyroid and T3-treated (1, 3, 5 and 10 days) rats. Caspase-3 activity was measured by colorimetric assay using

DEVD-pNA as substrate. Asterisk indicates significant difference compared with euthyroid rats (*p < 0.001). w = p < 0.001, ** = p < 0.01.

A. Kumar et al. / Journal of Hepatology 46 (2007) 888–898 891

�4-fold in the hyperthyroid group, while in thehypothyroid group, it remained unchanged (Fig. 2Cand D). We also measured caspase-3 activity in livertissue lysates by colorimetric assay, which showed atime-dependent increase after T3 administration(Fig. 2E).

3.3. Hyperthyroidism-induced apoptosis involvesactivation of death receptor pathway

In comparison to euthyroid group, TNF-a and FasLlevels were elevated significantly (p < 0.001) in thehyperthyroid group, whereas in the hypothyroid group,

Fig. 3. Death receptor-mediated pathway is activated in hyperthyroid rat

liver. (A) Tissue homogenate (50 lg) from eu-, hypo- and hyperthyroid

(T3 for 10 days) rat liver was run on 15% SDS–PAGE, transferred to

nitrocellulose membranes and probed with a TNF-a, TNF-R1, TNF-R2,

Fas ligand, Fas, or cleaved subunits of caspase-8. b-actin was used as a

loading control. (B) Fold change in density of TNF-a, Fas ligand, Fas

and cleaved subunit of caspase-8 in hyper- and hypothyroid rats relative

to euthyroid rats. Each bar represents the mean of the respective

individual ratios ±SD (n = 5 rats in each group). Asterisk indicates

significant difference compared with euthyroid rats (*p < 0.001).

892 A. Kumar et al. / Journal of Hepatology 46 (2007) 888–898

TNF-a levels remained unaltered and the FasL levelsincreased (Fig. 3). Expression of TNF-R1 and Fas waselevated �7- and �5-fold, respectively, in the hyperthy-roid group, but their expression did not alter in thehypothyroid group. The expression of TNF-R2 didnot alter in either of the group (Fig. 3A and B). Cleavedsubunit of caspase-8 increased �8-fold in the hyperthy-

roid group compared to the euthyroid group, whereasits levels increased only �1.5-fold in the hypothyroidgroup (Fig. 3).

3.4. Hyperthyroidism elevates p75NTR levels and its

ligands, proNGF and proBDNF

We investigated the effect of altered thyroidal statuson the protein expression of p75NTR, sortilin and theirligands, proNGF and proBDNF, in the rat liver. In theeuthyroid group, two forms of p75NTR protein, 75 kDa(glycosylated) and 50 kDa (nascent non-glycosylated),were detected (Fig. 4A). The expression of both theforms was up-regulated significantly (p < 0.001) in thehyperthyroid group, on the contrary, in the hypothyroidgroup, the p75NTR level decreased significantly(p < 0.01) (Fig. 4A and B). We examined sortilin expres-sion in rat liver and found that the expression of sortilinwas not modulated by altered thyroidal status (Fig. 4A).

To determine whether the liver cells undergoingapoptosis in the hyperthyroid rats expressed p75NTR,we used an antibody against cleaved-caspase-3 [31]together with anti-p75NTR sera for double label immu-nofluorescence. Merged images of p75NTR (green),cleaved-caspase-3 (red) and Hoechst-33258 (blue)showed that majority of the apoptotic cells (shown bywhite arrows) express p75NTR (Fig. 4C). Liver cellsnot expressing p75NTR (red arrows) were negative forcleaved-caspase-3 as well (Fig. 4C, a–b).

In euthyroid rat liver, proNGF (53 kDa) was detect-ed, and its level was up-regulated by �55-fold in thehyperthyroid group. Besides, in the hypothyroid groupa significant increase in proNGF levels was alsoobserved compared to the euthyroid group, but itsincrease was much less as compared to hyperthyroid rats(Fig. 5A and B). Further, mature NGF (14 kDa) wasundetectable in any of the experimental groups. proB-DNF (60 kDa) was present in the euthyroid group,but mature BDNF (14 kDa) was not seen. However,in the hyperthyroid group, proBDNF levels increasedsignificantly (p < 0.001) compared to the euthyroidgroup; whereas in the hypothyroid group its levelsremained unaltered (Fig. 5C and D). Interestingly,mature BDNF was also detected in the hyperthyroidgroup (Fig. 5C).

3.5. Hyperthyroidism-induced apoptosis involves oxidative

stress and dissipation of Dwm

TNF-a is known to induce ROS generation initiatingcell death [11,32]. T3 administration to rats caused sever-al fold increase in ROS generation, as assessed by flowcytometry in isolated rat hepatocytes. ROS generationwas enhanced �6- and 18-fold after 1 day and 3 daysof T3 administration, respectively, but after 5 days, only3-fold increase was observed, which enhanced up to

Fig. 4. p75NTR is elevated in the hyperthyroid rat liver. (A) Tissue homogenate (50 lg) from eu-, hypo- and hyperthyroid (T3 for 10 days) rat liver was

run on 12% SDS–PAGE, transferred to nitrocellulose membranes and probed with a p75NTR or sortilin antibody. b-actin was used as a loading control.

(B) Fold change in density of p75NTR in hyper- and hypothyroid rats relative to euthyroid rats. Bars represent mean values ± SD of five rats in each

group. Asterisks indicate significant difference compared with euthyroid rats (*p < 0.01; **p < 0.001). (C) Representative photomicrographs showing

double immunolabeling for activated caspase-3 and p75NTR in hyperthyroid rat liver. Merged image (d) of p75NTR (a, green), activated caspase-3

(b, red), and Hoechst (c, blue) immunolabeling showed that activated-caspase-3 positive cells express p75NTR (d, white arrows). Red arrows indicate

cells, negative for p75NTR. Scale bar in (a–d), 20 lm.

A. Kumar et al. / Journal of Hepatology 46 (2007) 888–898 893

�7-fold after 10 days of T3 administration (Fig. 6A).Superoxide dismutase (SOD) and catalase play animportant role in quenching of the ROS [11]. SODand catalase activities decreased significantly in thehyperthyroid group compared to the euthyroid group(SOD, p < 0.01; catalase p < 0.001), whereas theiractivities increased significantly in the hypothyroidgroup (SOD, p < 0.05; catalase p < 0.001) (Fig. 6B and C).

Increased ROS levels are often associated with lossof Dwm, an early apoptotic event [11]. To confirm theassociation of Dwm loss with hyperthyroidism-inducedapoptosis, T3 was administered to rats for 1, 3, 5 and10 days and Dwm was measured in isolated hepato-cytes by flow cytometry using JC-1 dye. At highDwm, JC-1 forms aggregates which emits fluorescenceat 590 nm (FL-2 channel), while at low Dwm JC-1remains in monomer form, which emit fluorescenceat 527 nm (FL-1 channel). After T3 administrationto rats, a time-dependent decrease was observed inDwm, which was indicated by increase in FL1/FL2ratio. After 10 days of T3 administration �15-foldincrease in FL1/FL2 fluorescence ratio was observed.(Fig. 6D).

3.6. Hyperthyroidism activates JNK

NF-jB negatively regulates TNF-a-mediated apopto-tic signaling [32] by inhibiting the sustained activation ofJNK [33]. Using EMSA, basal level of NF-jB DNA-binding activity was observed in euthyroid rat liver,which was enhanced �7- to 10-fold after 1, 3 and 5 daysof T3 administration, but after 10 days of T3 administra-tion, its level declined sharply even below the basal level(Fig. 7A and B). Deactivation of NF-jB in the hyper-thyroid group (10 days after T3 administration) was fur-ther demonstrated by double immunofluorescence usingan antibody against RelA/p65 subunit of NF-jB andHoechst-33258. Merged images of NF-jB (red) andHoechst (blue) demonstrated a considerable amount ofnuclear localization of NF-jB (pink) in the euthyroidrat liver (Fig. 7C, upper panel). In contrast, in the hyper-thyroid group, minimal translocation of NF-jB to thenucleus was observed, and was predominantly localizedin cytosol, suggesting the deactivation of NF-jB underthis condition (Fig. 7C, lower panel). Lastly, immuno-blotting with phospho-JNK specific antibody showed�3-fold increase in phospho-JNK levels in the hyperthy-

Fig. 5. Hyperthyroidism increased the proNGF and proBDNF levels in rat liver. (A and C) Tissue homogenate (50 lg) from eu-, hypothyroid- and

hyperthyroid (T3 for 10 days) rat liver was run on 15% SDS–PAGE, transferred to nitrocellulose membranes and probed with a NGF antibody (A) or

BDNF antibody (C). b-actin was used as a loading control. (B and D) Fold change in density of proNGF (B) and proBDNF (D) in hyper- and hypothyroid

rats relative to euthyroid rats. Bars represent mean values ± SD of five rats in each group. Asterisk indicates significant difference compared with

euthyroid rats (*p < 0.001).

894 A. Kumar et al. / Journal of Hepatology 46 (2007) 888–898

roid group (Fig. 7D and E), whereas in the hypothyroidgroup its levels were not altered. Further, the levels ofnon-phosphorylated JNK remained unaltered in eitherof the group (Fig. 7D).

4. Discussion

The hypermetabolic state in hyperthyroidism resultsin oxidative injury in some of the tissues, including liver,heart and muscles. There are mounting evidences thatshow liver damage in hyperthyroid patients [1–5], how-ever the detailed mechanism of hepatic dysfunction dur-ing hyperthyroid conditions has not been elucidated yet.Previously we demonstrated that severe hyperthyroid-ism induces apoptosis in rat liver through mitochon-dria-mediated pathway [9]. Here we have shown theinvolvement of death receptor-mediated pathway ofapoptosis in hyperthyroid rat liver. We demonstratefor the first time that hyperthyroidism increases the lev-els of p75NTR and its cell death inducing ligands, proN-

GF and proBDNF, in rat liver. Further, we observedthat the activation of death receptors includingp75NTR can induce apoptotic signaling by activatingJNK. These signaling pathways may converge on mito-chondria that result in enhanced ROS generation, lossof Dwm and finally the activation of proteolytic cascade.

Weight loss despite good appetite is a classical mani-festation of thyrotoxicosis. Indeed, we have also found asignificant decrease in body weight of hyperthyroidgroup rats. Corroborating with clinical studies [1–5], sig-nificant increase in serum AST and ALT levels in thehyperthyroid group indicated that hyperthyroidisminduced hepatic dysfunction in rats. Haematoxylin andeosin stained sections of hyperthyroid liver showed thefocal loss of bilaminar pattern with loss of cellulardetails (data not shown). Patchy lobular disarray alongwith focal hepato-cellular necrosis was also observed,indicating hepatic injury under hyperthyroid condition.To understand the nature of cell death in hyperthyroidrat liver, TUNEL and caspase-3 activity assays wereperformed that confirmed our earlier findings [9].

A. Kumar et al. / Journal of Hepatology 46 (2007) 888–898 895

Several members of death receptors and their ligandslike TNF-a and FasL/Fas play an important role inpatho-physiology of hepatic dysfunction [13]. In fact,liver is very susceptible to FasL-induced apoptosis[34]. We observed elevated Fas and FasL levels in thehyperthyroid group, suggesting the involvement of

Fas/FasL-mediated pathway. In this study, a hugeincrease in the TNF-a level was observed in hyperthy-roid rat liver, which is in agreement with a previousreport [30]. TNF-a exerts a variety of effects that aremediated by TNF-R1 and TNF-R2, but its apoptoticeffects are only mediated by TNF-R1 [35]. Enhancementof TNF-R1 expression in the hyperthyroid group indi-cates the activation of TNF-a-induced cell death path-ways. Initiation of death receptor pathway activatescaspase-8, which subsequently leads to activation ofeffector caspases [12]. A remarkable increase incleaved-caspase-8 levels in hyperthyroid rat liver indicat-ed the activation of death receptor pathway.

The ultimate fate of cells exposed to TNF-a is deter-mined by signal integration between its different effec-tors, including NF-jB, JNK and caspases. NF-jBnegatively regulates TNF-a-induced apoptosis by up-regulating the expression of anti-apoptotic genes viz.Bcl-2, SOD and JNK-phosphatases and thereby inhibit-ing sustained JNK activation [32,36]. NF-jB activationin rat liver, exposed to T3 for 1, 3 and 5 days is in agree-ment with previous studies [30,37], on the contrary fur-ther exposure of rats to T3 leads to the deactivation ofNF-jB that may lead to sustained JNK activation andapoptosis [32].

Hyperthyroidism is known to enhance ROS genera-tion and induce oxidative stress in liver tissue by disturb-ing the endogenous pro-oxidant/anti-oxidantequilibrium [38]. Hence several fold increase in ROS,decreased levels of SOD and catalase activities underhyperthyroid condition shift the ‘rheostat’ towardspro-oxidants, resulting in loss of Dwm and activationof intrinsic pathway of apoptosis, as observed hereand reported previously [9,11,39]. An early increase inBcl-2, manganese-SOD and activation of NF-jB in liverin response to T3 may constitute a protective mechanismagainst ROS toxicity [37], but further exposure to toxi-cants or T3 can exacerbate the liver injury by deactivat-ing NF-jB [40]. Furthermore, persistent TNF signalingleading to enhanced ROS generation under NF-jB

Fig. 6. Hyperthyroidism enhances ROS generation, and decreases Dwm

and anti-oxidant enzyme activities. (A and D) Hepatocytes were isolated

from euthyroid and T3-treated rat (1, 3, 5 and 10 days) livers by

collagenase perfusion method. Cell viability of the isolated hepatocytes

was more than 90% in all the groups, as assessed by trypan blue dye

exclusion method ROS (A) and Dwm (D) were measured by flow

cytometry using DCF-DA and JC-1 dye, respectively. ROS levels (A)

are expressed as fold change in DCF fluorescence. Increase in the ratio of

FL1/FL-2 fluorescence of JC-1 indicates loss of Dwm (D) whereas an

increase in the FL-1 fluorescence of DCF indicates enhanced ROS

production (A). Data shown are representative of four rats in each group.

Superoxide dismutase (B) and catalase (C) activities were measured in

tissue homogenate of rat liver. Data are expressed as mean values ± SD

of four rats in each group. Asterisks and psi (w) indicate significant

difference compared with euthyroid rats (*p < 0.05; **p < 0.01;wp < 0.001).

b

Fig. 7. DNA-binding activity of NF-jB and its localization, and JNK activation in rat liver. (A) T3 (12.5 lg/100 g body weight) was administered to rats

for 1 (lane 2), 3 (lane 3), 5 (lane 4) and 10 days (lane 5) and nuclear lysates were prepared from above T3-treated and euthyroid (lane 1) rat livers and

analyzed by EMSA. DNA–protein complex was electrophoresed on 4% polyacrylamide gel in Tris–borate buffer, then gel was dried and

autoradiographed. N = negative control. (B) Densitometric quantitation of relative DNA-binding activity of NF-jB. Each data point represents the mean

of the respective individual ratios ± SD (n = 4 rats in each group). (C) Photomicrographs showing the localization of NF-jB in euthyroid (a–c) and

hyperthyroid (T3 for 10 days) (d–f) rat liver. Paraffin-embedded sections were labeled with a NF-jB/p65 antibody (red) and counterstained with Hoechst-

33258, a nuclear stain (blue). Merged image (c and f) of NF-jB/p65 (a and d) and Hoechst (b and e) shows the localization of NF-jB in nuclei (arrows) of

euthyroid rat liver (c), whereas in hyperthyroid liver, NF-jB was localized in cytosol (f). Scale bar in (a–f) 20 lm. Data shown are representative of four

rats in each group. (D) Tissue homogenate (50 lg) from euthyroid, hyperthyroid (T3 for 10 days) and hypothyroid rat liver was run on 12% SDS–PAGE,

transferred to nitrocellulose membranes and probed with a phospho-specific JNK (upper panel) and non-phosphorylated-JNK (lower panel) antibody. (E)

Fold change in density of phospho-JNK (p54 and p46) in hyper- and hypothyroid rats relative to euthyroid rats. Each bar represents the mean of the

respective individual ratios ± SD (n = 4 rats in each group). Asterisks and psi indicate significant difference compared with euthyroid rats (*p < 0.05,wp < 0.001).

896 A. Kumar et al. / Journal of Hepatology 46 (2007) 888–898

A. Kumar et al. / Journal of Hepatology 46 (2007) 888–898 897

inhibition may result in sustained JNK activation thatcan promote apoptosis in the hyperthyroid liver [32].

p75NTR is also a member of TNF-receptor super-family, which induces apoptosis in a cell context-depen-dent manner upon neurotrophin binding [14]. The roleof neurotrophins and their receptors in liver patho-phys-iology is still not clear. Previous studies have demon-strated that activated HSC express p75NTR andundergo apoptosis upon binding with NGF [41,42].Our observation showing elevated levels of p75NTRand its ligands, proNGF and proBDNF, in hyperthy-roid rat liver suggests its implication in apoptosis induc-tion. This was further supported by the findings thatmajority of apoptotic cells in the hyperthyroid liverexpress p75NTR. Sortilin is required for formation ofactive ternary complex comprising, proNGF/proB-DNF, sortilin and p75NTR, that promote apoptotic sig-naling [19,20]. Thus, presence of sortilin and high levelsof proNGF and proBDNF under hyperthyroid condi-tion may activate the p75NTR-signaling to induce apop-tosis in rat liver. The activation of JNK is essential forp75NTR-mediated apoptosis [43]. Hence JNK activa-tion observed here further supports our hypothesis thatpro-neurotrophins-p75NTR signaling may be involvedin hyperthyroidism-induced apoptosis in rat liver.

Significant decrease in levels of AST, cleaved caspase-3, p75NTR and phospho-JNK in the hypothyroidgroup, as compared to the euthyroid group, suggeststhat hypothyroidism may have a protective role in liverinjury. Surprisingly, the increased levels of FasL andproNGF were also observed in the hypothyroid liver,but due to low levels of their cognate receptors, Fasand p75NTR, apoptosis is diminutive.

In conclusion, the present study and our earlierreport clearly demonstrate that excess TH acts at multi-ple levels to induce apoptosis. TH excess enhances theexpression of several death receptors and their ligands,resulting in activation of apical caspase-8, which is fur-ther amplified through the activation p75NTR-mediatedpathways. However, the cell types showing elevated lev-els of cell death inducing ligands in hyperthyroid liverare less understood. In hyperthyroid liver, TNF-a isproduced by the activated Kupffer cells [44]. WhereasNGF and BDNF are expressed on hepatocytes andHSCs [21,41,42]. The production of these ligands in aparticular cell type under hypothyroid condition needsfurther investigation. Moreover, future studies are alsowarranted to understand whether T3 directly induceapoptosis in hepatocyte through the enhanced expres-sion of cell death inducing ligands like TNF-a, FasL,proNGF and proBDNF.

Acknowledgements

Dr. M.M. Godbole was supported by grants fromDepartment of Science and Technology, Government

of India (SR/FIST/LS-II-018/2003). UGC fellowshipto Ashok Kumar is duly acknowledged. We are indebtedto Dr. M.V. Chao (Skirball Institute of BiomolecularMedicine, New York), Dr. C.M. Petersen (Aarhus Uni-versity, Denmark), Dr. Brigitte Pettmann (Marseille,France) and Dr. M.K. Ernst (Frederick, Maryland)for their generous gifts of antibodies against p75NTR(9992), sortilin, p75NTR (MC 192) and NF-jB/p65,respectively. We thank Dr. Sanghamitra Bandyopadhy-ay for critical reading of the manuscript.

References

[1] Kobayashi C, Sasaki H, Kosuge K, Miyakita Y, Hayakawa M,Suzuki A, et al. Severe starvation hypoglycemia and congestiveheart failure induced by thyroid crisis, with accidentally inducedsevere liver dysfunction and disseminated intravascular coagula-tion. Intern Med 2005;44:234–239.

[2] Hasan MK, Tierney WM, Baker MZ. Severe cholestatic jaundicein hyperthyroidism after treatment with 131-iodine. Am J Med Sci2004;328:348–350.

[3] Bellassoued M, Mnif M, Kaffel N, Rekik N, Rebai T, Tahri N,et al. Thyrotoxicosis hepatitis: a case report. Ann Endocrinol2001;62:235–238.

[4] Choudhary AM, Roberts I. Thyroid storm presenting with liverfailure. J Clin Gastroenterol 1999;29:318–321.

[5] Gavin LA, Cavalieri RR. Interrelationship between thyroid glandand the liver. In: Zakim D, Boyer TD, editors. Hepatology: Atextbook of liver disease. Philadelphia: Saunders; 1982. p.508–515.

[6] Dooner HP, Parada J, Aliaga C, Hoyl C. The liver in thyrotox-icosis. Arch Intern Med 1967;120:25–32.

[7] Klion FM, Segal R, Schaffner F. The effect of altered thyroidfunction on the ultrastructure of the human liver. Am J Med1971;50:317–324.

[8] Sola J, Pardo-Mindan FJ, Zozaya J, Quiroga J, Sangro B, PrietoJ. Liver changes in patients with hyperthyroidism. Liver1991;11:193–197.

[9] Upadhyay G, Singh R, Kumar A, Kumar S, Kapoor A, GodboleMM. Severe hyperthyroidism induces mitochondria-mediatedapoptosis in rat liver. Hepatology 2004;39:1120–1130.

[10] Lavrik IN, Golks A, Krammer PH. Caspases: pharmacologicalmanipulation of cell death. J Clin Invest 2005;115:2665–2672.

[11] Newmeyer DD, Ferguson-Miller S. Mitochondria: releasingpower for life and unleashing the machineries of death. Cell2003;112:481–490.

[12] Shakibaei M, Schulze-Tanzil G, Takada Y, Aggarwal BB. Redoxregulation of apoptosis by members of the TNF superfamily.Antioxid Redox Signal 2005;7:482–496.

[13] Yin XM, Ding WX. Death receptor activation-induced hepato-cyte apoptosis and liver injury. Curr Mol Med 2003;3:491–508.

[14] Nykjaer A, Willnow TE, Petersen CM. p75NTR-live or let die.Curr Opin Neurobiol 2005;15:49–57.

[15] Chao MV. Neurotrophins and their receptors: a convergencepoint for many signalling pathways. Nat Rev Neurosci2003;4:299–309.

[16] Lee R, Kermani P, Teng KK, Hempstead BL. Regulation of cellsurvival by secreted proneurotrophins. Science2001;294:1945–1948.

[17] Beattie MS, Harrington AW, Lee R, Kim JY, Boyce SL, LongoFM, et al. ProNGF induces p75-mediated death of oligodendro-cytes following spinal cord injury. Neuron 2002;36:375–386.

[18] Hempstead BL. Dissecting the diverse actions of pro- and matureneurotrophins. Curr Alzheimer Res 2006;3:19–24.

898 A. Kumar et al. / Journal of Hepatology 46 (2007) 888–898

[19] Nykjaer A, Lee R, Teng KK, Jansen P, Madsen P, Nielsen MS,et al. Sortilin is essential for proNGF-induced neuronal celldeath. Nature 2004;427:843–848.

[20] Teng HK, Teng KK, Lee R, Wright S, Tevar S, Almeida RD,et al. ProBDNF induces neuronal apoptosis via activation of areceptor complex of p75NTR and sortilin. J Neurosci2005;25:5455–5463.

[21] Cassiman D, Denef C, Desmet VJ, Roskams T. Human and rathepatic stellate cells express neurotrophins and neurotrophinreceptors. Hepatology 2001;33:148–158.

[22] Koji T, Kobayashi N, Nakanishi Y, Yoshii A, Hashimoto S,Shibata Y, et al. Immunohistochemical localization of Fasantigen paraffin sections with rabbit antibodies against humansynthetic Fas peptides. Acta Histochem Cytochem1994;27:459–463.

[23] Suda T, Takahashi T, Golstein P, Nagata S. Molecular cloningand expression of the Fas ligand, a novel member of the tumornecrosis factor family. Cell 1993;75:1169–1178.

[24] Salvioli S, Maseroli R, Pazienza TL, Bobyleva V, Cossarizza A.Use of flow cytometry as a tool to study mitochondrial membranepotential in isolated, living hepatocytes. Biochemistry1998;63:235–238.

[25] Andres D, Alvarez AM, Diez-Fernandez C, Zaragoza A, CascalesM. HSP 70 induction by cyclosporine A in cultured hepatocytes:effect of vitamin E succinate. J Hep 2000;33:570–579.

[26] Seglen PO. Incorporation of radioactive amino acids into proteinin isolated rat hepatocytes. Biochim Biophys Acta1976;442:391–404.

[27] Misra HP, Fridovich I. The generation of superoxide radicalduring the autooxidation of hemoglobin. J Biol Chem1972;247:6960–6962.

[28] Aebi H. Catalase in vitro. Methods Enzymol 1984;105:121–126.[29] Deryckere F, Gannon F. A one-hour minipreparation technique

for extraction of DNA-binding proteins from animal tissues.Biotechniques 1994;16:405–406.

[30] Tapia G, Fernandez V, Varela P, Cornejo P, Guerrero J, VidelaLA. Thyroid hormone-induced oxidative stress triggers nuclearfactor-kappaB activation and cytokine gene expression in rat liver.Free Radic Biol Med 2003;35:257–265.

[31] Hu BR, Liu CL, Ouyang Y, Blomgren K, Siesjo BK. Involvementof caspase-3 in cell death after hypoxia-ischemia declinesduring brain maturation. J Cereb Blood Flow Metab2000;20:1294–1300.

[32] Kamata H, Honda S, Maeda S, Chang L, Hirata H, Karin M.Reactive oxygen species promote TNFalpha-induced death and

sustained JNK activation by inhibiting MAP kinase phosphatases.Cell 2005;120:649–661.

[33] Deng Y, Ren X, Yang L, Lin Y, Wu X. A JNK-dependentpathway is required for TNFalpha-induced apoptosis. Cell2003;115:61–70.

[34] Ogasawara J, Watanabe-Fukunaga R, Adachi M, Matsuzawa A,Kasugai T, Kitamura Y, et al. Lethal effect of the anti-Fasantibody in mice. Nature 1993;364:806–809.

[35] Varfolomeev EE, Ashkenazi A. Tumor necrosis factor: anapoptosis JuNKie?. Cell 2004;116:491–497.

[36] Karin M, Lin A. NF-kappaB at the crossroads of life and death.Nat Immunol 2002;3:221–227.

[37] Fernandez V, Tapia G, Varela P, Castillo I, Mora C, Moya F,et al. Redox up-regulated expression of rat liver manganesesuperoxide dismutase and Bcl-2 by thyroid hormone is associatedwith inhibitor of kappaB-alpha phosphorylation and nuclearfactor-kappaB activation. J Endocrinol 2005;186:539–547.

[38] Fernandez V, Videla LA. Influence of hyperthyroidism onsuperoxide radical and hydrogen peroxide production by rat liversub-mitochondrial particles. Free Radic Res Commun1993;18:329–335.

[39] Kalderon B, Hermesh O, Bar-Tana J. Mitochondrial permeabilitytransition is induced by in vivo thyroid hormone treatment.Endocrinology 1995;136:3552–3556.

[40] Valencia C, Cornejo P, Romanque P, Tapia G, Varela P, VidelaLA, et al. Effects of acute lindane intoxication and thyroidhormone administration in relation to nuclear factor-kappaBactivation, tumor necrosis factor-alpha expression, and Kupffercell function in the rat. Toxicol Lett 2004;148:21–28.

[41] Trim N, Morgan S, Evans M, Issa R, Fine D, Afford S, et al.Hepatic stellate cells express the low affinity nerve growth factorreceptor p75 and undergo apoptosis in response to nerve growthfactor stimulation. Am J Pathol 2000;156:1235–1243.

[42] Oakley F, Trim N, Constandinou CM, Ye W, Gray AM, FrantzG, et al. Hepatocytes express nerve growth factor during liverinjury: evidence for paracrine regulation of hepatic stellate cellapoptosis. Am J Pathol 2003;163:1849–1858.

[43] Bhakar AL, Howell JL, Paul CE, Salehi AH, Becker EB, Said F,et al. Apoptosis induced by p75NTR overexpression requires Junkinase-dependent phosphorylation of Bad. J Neurosci2003;23:11373–11381.

[44] Fernandez V, Videla LA, Tapia G, Israel Y. Increases intumor necrosis factor-alpha in response to thyroid hormone-induced liver oxidative stress in the rat. Free Radic Res2002;36:719–725.