"Toxic memory" via chaperone modification is a ... - Wiley Online Library

“Toxic Memory” via Chaperone Modification Is a Potential Mechanism for Rapid Mallory-Denk Body Reinduction Pavel Strnad,1,2 Guo-Zhong Tao,1 Phillip So,1 Kenneth Lau,3 Jim Schilling,3 Yuquan Wei,4 Jian Liao,4 and M. Bishr Omary1 The cytoplasmic hepatocyte inclusions, Mallory-Denk bodies (MDBs), are characteristic of several liver disorders, including alcoholic and nonalcoholic steatohepatitis. In mice, MDBs can be induced by long-term feeding with 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC) for 3 to 4 months or rapidly reformed in DDC-induced then recovered mice by DDC refeeding or exposure to a wide range of toxins for only 5 to 7 days. The molecular basis for such a rapid reinduction of MDBs is unknown. We hypothesized that protein changes retained after DDC priming contribute to the rapid MDB reappearance and associate with MDB formation in general terms. Two-dimensional differential-in-gel-electrophoresis coupled with mass spectrometry were used to characterize protein changes in livers from the various treatment groups. The alterations were assessed by real-time reverse-transcription polymerase chain reaction and confirmed by immunoblotting. DDC treatment led to pronounced charged isoform changes in several chaperone families, including Hsp25, 60, 70, GRP58, GRP75, and GRP78, which lasted at least for 1 month after discontinuation of DDC feeding, whereas changes in other proteins normalized during recovery. DDC feeding also resulted in altered expression of Hsp72, GRP75, and Hsp25 and in functional impairment of Hsp60 and Hsp70 as determined using a protein complex formation and release assay. The priming toward rapid MDB reinduction lasts for at least 3 months after DDC discontinuation, but becomes weaker after prolonged recovery. MDB reinduction parallels the rapid increase in p62 and Hsp25 levels as well as keratin 8 cross-linking that is normally associated with MDB formation. Conclusion: Persistent posttranslational modifications in chaperone proteins, coupled with protein cross-linking and altered chaperone expression and function likely contribute to the “toxic memory” of DDC-primed mice. We hypothesize that similar changes are important contributors to inclusion body formation in several diseases. (HEPATOLOGY 2008;48:931-942.)

I

ntermediate filaments (IFs), together with actin microfilaments and microtubules, are the three filamentous systems of the eukaryotic cytoskeleton.1,2 IFs make up the largest member of the cytoskeleton protein

families and are expressed in a cell-specific manner (e.g., desmin in muscle, neurofilaments in neurons, and keratins in epithelial cells).3,4 The functional importance of IFs is reflected by the growing list of human diseases that

Abbreviations: ATP, adenosine triphosphate; DDC, 3,5-diethoxycarbonyl-1,4-dihydrocollidine; 2D-DIGE, two-dimensional differential-in-gel-electrophoresis; HSF, heat shock factor; Hsp, heat shock protein; IF, intermediate filament; K, keratin; MDB, Mallory-Denk body; RT-PCR, reverse-transcription polymerase chain reaction; SDS, sodium dodecyl sulfate. From the 1Department of Medicine, Palo Alto VA Medical Center, Palo Alto, CA, and Stanford University Digestive Disease Center; the 2Department of Internal Medicine I, University of Ulm, Ulm, Germany; the 3Department of Pediatrics, Stanford University School of Medicine; and the 4State Key Laboratory of Biotherapy, Sichuan University, Chengdu, China. Received February 5, 2008; accepted May 9, 2008. Supported by National Institutes of Health Grant DK52951 and the Department of Veterans Affairs (M. B. O.); National Institutes of Health Grant DK56339; and in part by an EMBO postdoctoral fellowship as well as German Research Foundation grant STR 1095/1-1 (P. S.). Address correspondence to: Bishr Omary, Department of Molecular and Integrative Physiology, 7744 Medical Science Building II, 1301 E. Catherine, Ann Arbor, MI 48109-0622. tel: 734-764-4376; fax: 734-936-8813. Copyright © 2008 by the American Association for the Study of Liver Diseases. Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/hep.22430 Potential conflict of interest: Dr. Omary owns stock in and advises Applied. Supplementary material for this article can be found on the Hepatology Web site (http:interscience.wiley.com/jpages/0270-9139/suppmat/index.html). 931

932

STRNAD ET AL.

are caused by, or predisposed to, IF mutations or variants, respectively.5 In addition, IFs are relatively insoluble proteins that comprise the major constituents of protein inclusions/aggregates that are characteristic of several hepatic, neurodegenerative, and muscular disorders.5-9 These IF-enriched inclusions consist of misfolded and ubiquitinated proteins together with chaperones and the ubiquitin-binding protein p62.7,10,11 Mallory-Denk-bodies (MDBs) are hepatocyte-specific aggregates and are among the most common inclusion bodies because of the relative high prevalence of associated liver diseases.9 MDBs are characteristic of alcoholic steatohepatitis but are also found in other diseases, including nonalcoholic steatohepatitis, primary biliary cirrhosis, and Wilson disease.9,12,13 MDBs consist mainly of the IFs keratin 8 and 18 (K8/K18) that have undergone several posttranslational modifications, including hyperphosphorylation and transamidation.9,13 MDB formation requires an alteration in the normal equimolar K8:K18 ratio (under basal conditions) to disproportional K8⬎K18 levels that ultimately lead to K8 cross-linking and insolubility.9,13 Our understanding of MDB pathogenesis has been facilitated by mouse models whereby feeding mice the hepatotoxic drugs griseofulvin or 3,5-diethoxycarbonyl1,4-dihydrocollidin (DDC) for 2 to 5 months leads to MDB formation. Similar to the human situation, murine MDB formation is reversible, and MDBs largely disappear after switching the animals to a normal diet for 1 month.9 Interestingly, DDC-primed mice (i.e., mice fed DDC long-term, then allowed to recover by feeding a normal diet) posses a remarkable ability to rapidly reform MDBs within 5 to 7 days after exposure to a variety of stress-inducing agents.9,12 This phenomenon is referred to as “toxic memory.”14 Despite the diagnostic importance of MDBs, the precise mechanisms underlying their formation are not completely understood.9,13 Similar to other nonhepatocyte protein aggregates, MDB formation is thought to be facilitated by altered protein conformation that leads to exposure of hydrophobic amino acids.7,15 A change in protein conformation might be caused by a genetic defect or by protein misfolding as a consequence of oxidative stress. In support of the latter possibility, MDB-forming diseases are associated with increased levels of oxidative stress, and rapid MDB reinduction in DDC-primed mice can be reversed by pretreatment with the antioxidant S-adenosylmethionine.16-19 An important mechanism that cells use to prevent protein misfolding is the up-regulation and utilization of chaperones that associate with newly synthesized or damaged proteins and facilitate their correct folding.15 Heat

HEPATOLOGY, September 2008

shock proteins (Hsp) are the best-studied chaperones and are distributed in different subcellular compartments. For example, Hsp60 is found in mitochondria, whereas Hsp70 is a major cytoplasmic chaperone, while other organelles/compartments express different or related chaperone family members.20 However, the protective capacity of chaperones might be depleted or compromised, thereby leading to aggregation of misfolded proteins.7,11,15 We hypothesized that retained molecular changes in DDC-recovered mice are a component of the “toxic memory” that may play a role in the rapid MDB reformation upon rechallenge. To identify such changes, we used an unbiased proteomic approach that revealed that MDB formation is associated with persistent chaperone-charged isoform changes that lead to functional impairment in several chaperones.

Materials and Methods Animal Experiments. MDBs were induced in 2- to 3-month-old male C3H mice (Taconic Farms, Germantown, NY) by feeding a powdered Formulab 5008 diet ad libitum for 3.5 months (Dean’s Animal Feeds, Redwood City, CA) containing 0.1% DDC (Aldrich, St. Louis, MO). To analyze MDB reformation, mice were recovered on a DDC-free Formulab 5008 diet (control diet) for 1 or 3 months followed by refeeding with the 0.1% DDCcontaining diet for 6 to 7 days (Fig. 1A/Fig.5A schematics). Mice were sacrificed by CO2 inhalation followed by blood collection using cardiac puncture, then removal of the liver. Serum was used to measure alanine aminotransferase (IU/L), alkaline phosphatase (IU/L), and total bilirubin (mg/dL). Livers were weighed, cut into pieces, and flash-frozen in liquid nitrogen for biochemical and twodimensional differential-in-gel-electrophoresis (2DDIGE). Alternatively, freshly isolated liver pieces were embedded in optimal cutting temperature compound for immunofluorescence staining or submerged in RNAlater Stabilization Reagent (Qiagen, Valencia, CA) and stored (⫺80°C) for subsequent real-time reverse-transcription polymerase chain reaction (RT-PCR) analysis. The animal experiments protocol was approved by the Institutional Animal Care Committee. 2D-DIGE Analysis. Tissue lysates were prepared by homogenizing the preweighed liver pieces in 2D lysis buffer (30 mM Tris-HCl [pH 8.8]; 7 M urea; 2 M thiourea and 4% CHAPS) to achieve a final protein concentration of 4 to 8 mg/mL. The two liver samples to be compared are labeled with Cy3 or Cy5 dyes and a 1:1 mix is also labeled with Cy2 which was included with the test samples as an internal labeling control. Samples were then analyzed using an Amersham BioSciences 2D-gel system

HEPATOLOGY, Vol. 48, No. 3, 2008

STRNAD ET AL.

933

Fig. 1. DDC treatment causes pronounced molecular changes. (A) Schematic of the mouse treatment protocol. Note that the “primed” mice do not exhibit any MDBs, but are rapidly predisposed to their formation. (B) 2D-DIGE of mouse liver homogenates (labeled with Cy3/Cy5) from control mice (Untreated, Cy3) or mice fed DDC for 3.5 months (Cy5-DDC 3.5M). Altered charged chaperone and cytoskeletal protein isoform changes upon DDC feeding are highlighted. (C) Recovery of DDC-fed mice for 1 month on control diet (Rec 1M) leads to resolution of most but not all of the DDC-induced molecular changes, whereas alterations in chaperones (boxed areas) are retained (see also Supplemental Fig. 1). (D) DDC readministration for 6 days to recovered mice (R1M⫹DDC6d) causes re-emergence of the molecular alterations seen in DDC 3.5 M mice. Three pairs of mice each for panels B-D were compared, and one representative example is shown.

(Amersham BioSciences; Piscataway, NJ). Images were scanned using Typhoon TRIO and analyzed by Image Quant version 5.0 software (Amersham BioScience). An in-gel DeCyder software was used to determine the protein level ratios. Selected spots were collected with an Ettan Spot Picker (Amersham BioSciences), subjected to in-gel trypsinization, peptide extraction, and desalting, followed by MALDI-TOF/TOF (Applied Biosystems,

Foster City, CA) analysis to determine the protein identity. Antibodies. The utilized antibodies are directed toward: K8/K18 (Ab-8592) and K18 (Ab-4668 or Troma2);21 Hsp60 (LK2), transglutaminase-2 (Ab-4), ␣-tubulin (DM1A), and ␤-actin (ACTN05) (Labvision, Fremont, CA); Hsp25 (sc-1049), Hsp/c70 (sc-1060), ubiquitin (sc8017), and p62 (sc-10117) (Santa Cruz Biotechnology,

934

STRNAD ET AL.

HEPATOLOGY, September 2008

Table 1. DDC Administration Leads to Pronounced Isoform Changes in Multiple Proteins Isoform #

Cytoskeletal proteins ␤-Actin Keratin 8 Keratin 18 Chaperones GRP58 GRP75 GRP78 Hsp60 Hsc71 Hsp 84 Detoxification GST mu GST pi Others: Aconitate hydralase Albumin ALDH4A1 Annexin 5 Antioxidant protein 2 Carbonic anhydrase III Demethylglycine Dehydrogenase 10-formyltetrahydrofolate dehydrogenase Heterogenous nuclear ribunucleoprotein 3-Hydroxy-3-methylgluta-rylCoA synthase 2 Major urinary protein Ornithine aminotransferase Transferrin

DDC 3.5M/Untreated (fold change)

Rec 1M/Untreated (fold change)

1

2

3

4

5

1

2

3

4

3.9 2.9 3.2

4.7

1.4 2.0 1.4

1.2 2.1 1.6

2.4 2.9 1.8

2.4 1.8 1.5

1.7 ⫺1.4 1.0 ⫺1.5 ⫺1.5 1.4

1.4 2.1

2.3 2.3

2.6 2.2

3.6 4.1

⫺1.3

2.6

1.5

1.3 ⫺1.0 ⫺1.5 1.4 2.7 ⫺1.3

2.5 2.4 1.0 2.0

1.7 4.6 2.2 2.2

1.4

1.9

2.1 1.3 3.0

2.1 ⫺1.6 3.3

4.4 4.4 5.1

⫺1.2 1.1 ⫺1.2 ⫺1.1

1.4 1.9 1.3 ⫺1.1

1.3

3.7 ⫺1.5

6.2 ⫺1.4

1.6 ⫺1.2

⫺1.7 1.4 ⫺1.2 3.6 1.7 ⫺4.7

1.1 1.4 ⫺1.7 4.6

1.1 2.1 ⫺1.2 7.1

⫺3.6

⫺1.7

⫺1.3

⫺2.5

⫺2.6

⫺2.0

⫺2.2

⫺1.5

⫺1.5

⫺1.1

⫺1.7

⫺1.9

⫺1.5

⫺1.5

⫺2.7

⫺1.9

1.1

1.6

1.4 1.7

1.7 1.6

3.0 ⫺1.1

⫺1.5

R1MⴙDDC6d/Untreated (fold change) 1

2

3

4

5

2.8 2.8 2.2

3.7

1.9 1.2 3.5

1.9 ⫺1.7 3.1

3.2 3.5 3.9

⫺1.1 1.1 ⫺1.0 1.1 1.6

1.5 1.6 1.4 ⫺1.0

1.2

4.5 ⫺2.0

6.5 ⫺1.6

1.5 ⫺1.3

⫺1.5 1.4 ⫺1.6 3.0 1.7 ⫺3.6

1.1 1.4 ⫺2.3 3.8

1.1 2.0 ⫺1.5 4.8

⫺2.7

⫺1.5

1.5

1.2

⫺2.1

1.3 ⫺1.2

⫺2.0

⫺2.1

3.5 2.2

4.8 3.3

1.4

⫺2.3 ⫺3.3 ⫺1.4 ⫺1.4 ⫺67 ⫺61 ⫺39 ⫺19 ⫺2.0 2.1 3.1 3.0 3.1

5

2.1

1.5 1.8

1.6 1.8

2.8 ⫺1.4

⫺1.4

⫺2.4

⫺2.0

⫺2.4

⫺1.8

⫺1.9

⫺3.3

⫺1.7 ⫺2.2 ⫺2.4 ⫺2.1 1.0

⫺1.3 ⫺2.2

⫺1.4 ⫺1.5

⫺1.3

1.7

2.3

3.9

⫺3.4 ⫺1.2 ⫺1.2 ⫺46 ⫺30 ⫺46 ⫺27 ⫺2.5 2.2 2.8 2.6 2.8

Individual spots that showed increased or decreased levels were cut and subjected to proteolysis and characterization as described in Materials and Methods. Three pairs of mice were used for each paired comparison as described in Fig. 1. Values of a representative experiment are shown.

Santa Cruz, CA); K8 (Troma-1) and K19 (Troma-3) (Developmental Studies Hybridoma Bank, Iowa City, IA); and GRP75 (SPS-825), GRP78 (SPA-827), and Hsp72 (SPA-810) (Stressgen Bioreagents, Ann Arbor, MI). Because sc-1060 recognizes both inducible and constitutively-expressed Hsp70 family members, the proteins recognized by this antibody are termed Hsp/c70. Immunofluorescence Staining and Quantitative RealTime RT-PCR. Acetone-fixed liver sections were stained as described,21 then analyzed using a Zeiss LSM510 confocal microscope (Zeiss MicroImaging; Thornwood, NY). Quantitative real-time RT-PCR analysis was performed as described22 using liver tissues preserved in RNAlater Stabilization Reagent (Qiagen). RNA translation was performed with Oligo-dT primers and target gene-specific primers (Supplemental Table 1). Three individual mice were tested for each treatment condition. Immunoblotting and Sucrose Gradient Separation. Total tissue homogenates were generated using a Dounce to solubilize liver pieces in 3% sodium dodecyl sulfate

(SDS) containing sample buffer.21 For sucrose gradient sedimentation, livers were homogenized in 1% Triton X100-containing buffer followed by pelleting.21 The supernatant was incubated (30 minutes) in the presence or absence of 5 mM MgATP (22°C), then immediately layered over a discontinuous sucrose gradient consisting of 6%, 12%, 18%, 24%, 30% sucrose (2 mL each) and centrifuged (38,000 rpm for 18 hours). The bottom of the tube was then punctured with a 22-gauge needle, and 0.6 mL fractions (#1-21) were collected and diluted 1:1 using 4⫻ reducing SDS-containing sample buffer. In order to quantify the amount of chaperones in different fractions, solubilized liver homogenates (in 1% Triton X-100) were incubated with or without 5 mM MgATP (30 minutes), then pelleted. The pellet was homogenized in equal volume of 1% Triton X-100/4⫻ reducing SDS-containing sample buffer. The distribution of chaperones within the fractions was examined by immunoblotting. Protein bands were visualized with an ECLplus kit (PerkinElmer, Boston, MA).21

HEPATOLOGY, Vol. 48, No. 3, 2008

Results DDC Treatment Leads to Persistent Charged Isoform Changes in Several Chaperone Proteins. We employed an unbiased proteomic approach to identify protein changes that associate with MDB formation and with the “toxic memory” phenomenon that is noted in DDC-primed livers. DDC feeding for 3.5 months resulted in abundant MDB formation (not shown but well

STRNAD ET AL.

935

Table 2. DDC Treatment Results in Altered Chaperone Messenger RNA Expression HSF1 HSF2 GRP75 GRP78 Hsp72 Hsc71 Hsp60

Untreated

DDC 3.5M

Rec 1M

R1MⴙDDC6d

1 ⫾ 0.17 1 ⫾ 0.24 1 ⫾ 0.13* 1 ⫾ 0.17 1 ⫾ 0.29* 1 ⫾ 0.20 1 ⫾ 0.17

1.05 ⫾ 0.12 0.78 ⫾ 0.07 0.71 ⫾ 0.03* 0.86 ⫾ 0.09 0.15 ⫾ 0.10* 0.92 ⫾ 0.21 0.70 ⫾ 0.05

1.33 ⫾ 0.12 0.80 ⫾ 0.09 0.81 ⫾ 0.11 0.95 ⫾ 0.13 0.59 ⫾ 0.25 0.86 ⫾ 0.04† 0.82 ⫾ 0.02

1.20 ⫾ 0.16 0.78 ⫾ 0.15 0.75 ⫾ 0.08 0.75 ⫾ 0.16 0.28 ⫾ 0.11 1.18 ⫾ 0.08† 0.82 ⫾ 0.04

Messenger RNA levels were estimated as described in Materials and Methods. Data were averaged from the amplification of four separate livers and normalized to the levels of untreated mice, which were set as 1. The values are expressed as the mean ⫾ standard deviation. *P ⬍ 0.05 (when comparing the two marked conditions). †P ⬍ 0.01 (when comparing the two marked conditions).

Fig. 2. DDC treatment leads to altered protein levels of chaperones and cytoskeletal proteins. Immunoblot analysis of total mouse liver lysates depicts alterations in several proteins after DDC administration. Mice were fed DDC for 3.5 months (DDC 3.5M) then recovered for 1 month on a control diet (Rec 1M) followed by DDC refeeding for 6 days (R1M⫹DDC6d). Long-term DDC feeding or short refeeding causes upregulation of the cytoskeletal proteins K8, K18, actin, and, to a lesser extent, tubulin. In contrast, Hsp72 and GRP75 become significantly down-regulated after long-term DDC feeding. Two independent mice per condition are shown, but similar findings were noted using at least three mice per condition.

described by Zatloukal et al.9) and pronounced charged isoform changes in multiple protein categories including chaperones, detoxifying, and cytoskeletal proteins (Fig. 1A,B) (for quantification, see Table 1). The recovery of DDC-exposed mice after switching to control diet for 1 month led to MDB disappearance, and many of the molecular alterations returned to normal (Fig. 1C). However, prominent among the charged isoform changes that did not revert to baseline are those involving chaperones Hsp60/Hsc71 and GRP58/75/78 (Fig. 1; Table 1; Supplemental Fig. 1). To study the rapid MDB reappearance upon DDC rechallenge, recovered mice were refed DDC for 6 days and, as already well established,9 DDC readministration caused abundant MDB reformation (data not shown). Notably, the accompanying molecular alterations were similar to those induced by long-term DDC feeding (Table 1 and compare Fig. 1B with Fig. 1D and left and right panels in Supplemental Fig. 1). DDC Treatment Results in Altered Chaperone Levels. We examined the DDC-induced changes in chaperones by immunoblotting and real-time RT-PCR. As described previously, DDC administration increases the levels of cytoskeletal proteins.9 However, most chaperone protein levels were either unaltered by DDC treatment (Hsp60, Hsp/c70, GRP78) or decreased (Hsp72, GRP75) (Fig.2), with the caveat that this assessment is based on antibody reactivity as contrasted with 2D-DIGE (Fig. 1) based on protein identification by mass spectrometry. Of the tested chaperones, only Hsp25 was DDCinducible and was most significantly up-regulated after DDC refeeding (Fig. 2). Quantitative real-time RT-PCR revealed that the lowered GRP75 and Hsp72 levels were at least in part due to transcriptional down-regulation. HSF1/2 levels were not significantly different (Table 2). The constitutive Hsc71 messenger RNA expression was slightly up-regulated after DDC rechallenge when com-

936

STRNAD ET AL.

HEPATOLOGY, September 2008

Fig. 3. DDC treatment results in altered chaperone function. Livers were removed from age- and sex-matched control (Untreated), 3.5-month DDC-fed, DDC-fed then 1- or 3-month recovered (Rec 1M/3M) mice. Tissues pieces were homogenized in 1% Triton X-100 containing buffer. The supernatants were incubated for 30 minutes in the presence or absence of 5 mM MgATP (⫹ATP and ⫺ATP) followed by sucrose gradient (0%-30%) sedimentation. Fractions (#1-21) were then collected followed by immunoblotting using antibodies to (A,B) Hsp/c70 and (C,D) Hsp60. Note that DDC feeding causes a decrease in Hsp association with high molecular weight complexes as evidenced by a minimal signal in fractions 1 through 5 (boxes in ⫺ATP panels). In contrast, similar chaperone distribution is noted in livers from 1 month of recovery and the untreated mice (panels A, C). However, ATP treatment induces a more prominent release of chaperone complexes from their substrates in “Rec 1M” as compared with “untreated” livers (based on loss of signal in fractions 5 through 8; ellipses in panels B, D).

pared with recovered mice, whereas the remaining tested chaperones did not exhibit significant changes (Table 2). DDC Administration Impairs Chaperone–Substrate Association. To test the effect of DDC treatment on the ability of chaperones to associate with their substrates, we devised a sedimentation-based assay that examined Hsp/c70 and Hsp60 chaperone complex formation and relative size in nonionic detergent-soluble liver lysates. This assay demonstrated that DDC exposure interferes with the ability of Hsp/c70 and Hsp60 to form large-sized complexes as demonstrated by a right-to-left shift in the presence of these chaperones from large to smaller-sized complexes (see boxed areas, Fig. 3A,C). No significant shift was noted in the distribution of Hsp/c70 and Hsp60 after 1 month recovery (Fig. 3A,C). Another important indication of appropriate chaperone function is their ability to release their bound substrates via chaperone ATPase activity in the presence of MgATP.23 This chaperone property was examined using our sedimentation assay in the presence of adenosine triphosphate (ATP). Under basal conditions, the distri-

bution profile of Hsp/c70 and Hsp60 switches to lowersized complexes after ATP exposure, although the pattern of shifting differs for Hsp/c70 (compare Fig. 3A with Fig. 3B) versus Hsp60 (compare Fig. 3C with Fig. 3D). This ATP-induced release is even more pronounced in DDCfed mice (Fig. 3B,D) which implies that DDC results in unstable chaperone-substrate binding. Notably, 1-month recovery normalizes the complex distribution profile in the absence of ATP (Fig. 3A,C) but not in the presence of ATP (Fig. 3B,D). A prolonged 3-month recovery of DDC-fed mice further normalizes the spread of Hsp/c70 distribution after ATP treatment (relative complex stability) to make it more similar to the non–DDC-treated spread (Fig. 3B) and similar findings were observed for Hsp60 (not shown). We also examined the effect of DDC and ATP on the distribution of Hsp/c70 and Hsp60 in the detergent-soluble (the fraction we used for the complex formation sedimentation assay) and the detergent-insoluble (the remaining pellet) fractions. Both Hsp60 and Hsp/c70 are predominantly soluble proteins and their distribution is

HEPATOLOGY, Vol. 48, No. 3, 2008

STRNAD ET AL.

Fig. 4. Hsp60 and Hsp/c70 are primarily Triton-soluble proteins and their solubility is not altered after ATP treatment. Liver homogenates from control mice and mice fed DDC for 3.5 months were prepared using 1% Triton X-100 – containing buffer and incubated with ⫾5 mM ATP (30 minutes) followed by pelleting. The supernatant (S) was collected and the pellet (P) was suspended in reducing Laemmli sample buffer. Equal fractions were separated by SDS–polyacrylamide gel electrophoresis, then analyzed by Commassie staining and immunoblotting with antibodies to Hsp/c70 and Hsp60.

not affected by DDC or ATP, although DDC and ATP do change the overall protein profile based on Coomassie blue staining of the entire fraction as may be expected (Fig. 4). DDC-feeding generates an additional Hsp/c70 antibody-reactive species of slightly slower migration

937

(Fig. 4, arrowhead), which was seen only in Triton-insoluble fractions and likely reflects an uncharacterized insoluble form of Hsp/c70. Priming Toward Rapid MDB Reinduction Becomes Weaker with Time, but Lasts for at Least 3 Months. We examined the persistence of the “toxic memory” by feeding mice DDC for 3.5 months followed by recovery on a normal diet for 1 or 3 months (Fig. 5A). In both cases, DDC challenge for 7 days led to rapid MDB reformation, albeit to nearly 3-fold less in the 3-month versus the 1-month recovery duration (Fig. 5Bd-i). In contrast, no MDBs were noted in naı¨ve mice fed DDC for 7 days (Fig. 5Ba-c). Therefore, priming toward rapid MDB reinduction lasts for at least 3 months but becomes weaker with time. This supports prior reports showing that rapid reinduction can occur 2 months after recovery.14 The 3-month recovery also leads to less liver injury upon DDC readministration as determined by the decreased rise in alanine aminotransferase levels (Table 3). Aside from the ability to form MDBs rapidly, primed mice retain slightly elevated alkaline phosphatase levels and repeated DDC exposure results in higher alkaline phosphatase levels when compared with first-time DDC treatment (Table 3). Also, the liver size does not normalize even after 3 months of recovery on control diet and repeated DDC exposure induces pronounced liver hypertrophy (Table 3). Repeated DDC Exposure Elicits Molecular Changes Different to Those Induced by One-Time DDC Feeding. In addition to comparing long-term DDC with DDC rechallenge as described above, we also analyzed the molecular differences in response to short-term (7-day) single administration versus short-term DDC refeeding. Repeat DDC exposure resembles the single short-term course of DDC feeding in terms of the induction of K8 and K18 and to a lesser extent tubulin, as well as the induction of transglutaminase-2 (Fig. 6). However, one distinct difference is that DDC refeeding leads to prominent K8 cross-linking (based on the presence of K8 species in the stacking [S] gel) which is not seen in control livers,24 and to elevated Hsp25, K19, and p62 (Figs. 6 and 7). Although most of K8 cross-linking disappears during recovery, remnant cross-linked K8-contain-

Table 3. DDC Induces Cholestatic and Hepatic Damage and Liver Hypertrophy

Bilirubin (mg/dL) ALT (IU/L) ALP (IU/L) Liver weight/body weight (%)

Untreated

DDC 7d

DDC 3.5M

Rec 1M

R1M ⴙ DDC7d

Rec 3M

R3Mⴙ DDC7d

0.5 ⫾ 0.4 49 ⫾ 24 90 ⫾ 12* 4.4 ⫾ 0.2

1.3 ⫾ 1.4 1,342 ⫾ 1,237 251 ⫾ 8† 5.4 ⫾ 1.3

0.6 ⫾ 0.7 1,170 ⫾ 90 654 ⫾ 77 16.6 ⫾ 0.2

0.4 ⫾ 0.2 51 ⫾ 13 115 ⫾ 6* 7.1 ⫾ 0.4

1.1 ⫾ 1.1 2,006 ⫾ 452* 479 ⫾ 50† 12.0 ⫾ 0.9

ND ND ND 6.1 ⫾ 0.1

0.5 ⫾ 0.5 987 ⫾ 462* 496 ⫾ 54† 8.1 ⫾ 0.4

Sex- and age-matched mice were used (4 to 6 mice per condition). Values are expressed as the mean ⫾ standard deviation. Abbreviations: ALT, alanine aminotransferase; ALP, alkaline phosphatase; ND, not determined. *P ⫽ 0.05. †P ⬍ 0.02 (DDC7d versus DDC refeeding).

938

STRNAD ET AL.

HEPATOLOGY, September 2008

Fig. 5. The “toxic memory” is attenuated after 3 months of recovery. (A) Schematic of the protocol used to compare 1-month with 3-month recovery after DDC feeding for 3.5 months. (B) Refeeding of recovered mice with DDC leads to rapid MDB formation as noted with double immunofluorescence staining using antibodies to K8 and ubiquitin. MDBs were quantified using immunostained livers from three animals per condition and were found to be nearly 3-fold more abundant in the 1-month (R1M⫹DDC7d) as compared with the 3-month (R3M⫹DDC7d) recovery period that was followed by 7 days of rechallenge. Note that DDC administration to previously untreated mice for 7 days (DDC7d) does not lead to MDB formation. Bar ⫽ 50 ␮m.

ing species do remain up to 3 months after recovery. Similar findings, in terms of increased K19 (Fig. 7) and Hsp25 (Supplemental Fig. 2) were noted by immunofluorescence staining. For example, prominent K19-expressing cells persist 1 or 3 months after recovery and increase significantly after rechallenge (Fig. 7c-f). Interestingly, the chaperone Hsp25 is present at minimal levels in control mice but its levels increase slightly after one-time DDC feeding as compared with the dramatic increase and colocalization with MDBs after DDC rechallenge (Supplemental Fig. 2).

Discussion The major findings of our study are as follows. (1) DDC treatment leads to pronounced charged isoform changes in several chaperones including Hsp25/60/ GRP58/GRP75/GRP78/Hsc71 which are retained at least 1 month after discontinuation of DDC feeding, whereas changes in numerous other proteins become less prominent upon recovery. (2) DDC feeding results in altered expression of Hsp72, GRP75, and Hsp25

HEPATOLOGY, Vol. 48, No. 3, 2008

Fig. 6. DDC refeeding results in a different molecular response when compared with first-time DDC feeding. The levels of the indicated proteins were compared by immunoblotting of liver homogenates obtained from untreated mice (Untreated); mice fed DDC for 7 days (DDC7d); mice fed DDC for 3.5 months, then recovered for 1 or 3 months (Rec 1M and Rec 3M, respectively); and mice refed DDC for 7 days after a 1-month (R1M⫹DDC7d) or 3-month (R3M⫹DDC7d) recovery.

and in functional impairment of Hsp60 and Hsp/c70 chaperone function. Hence, the well-described phenomenon of rapid MDB reappearance9 that was termed “toxic memory” by Denk and colleagues14 corresponds to, at least in part, specific changes in chaperones and their functions. (3) The “toxic memory” phenotype lasts for at least 3 months after DDC discontinuation but becomes weaker as recovery becomes longer. (4) Long-term followed by short-term DDC readministration differs from a single short-term DDC feeding in the extent of K19, Hsp25, and p62 induction as well as K8 cross-linking. The precise nature of the chaperone charge-altering isoform-generating alterations remain to be defined, but deamidation is one likely candidate modification that was noted in several proteins (not shown). Within chaperones, deamidation of crystallin, a member of the Hsp27 family, occurs during aging and increases during cataract formation.25,26 Interestingly, crystallin deamidation is in part catalyzed by tissue transglutaminase, an enzyme known to be involved in mouse MDB formation.24,27 Notably, protein deamidation may lead to protein aggregation as noted for the spontaneous deamidation of the amylin peptide that induces the formation of amyloidlike aggregates.28 Chaperone alterations have been ob-

STRNAD ET AL.

939

served after exposure to alcohol and fatty diet,29-31 which in humans can lead to alcoholic steatohepatitis and nonalcoholic steatohepatitis, respectively. Another striking feature of the observed chaperone changes is their longevity. Chaperone alterations are retained for at least one month after discontinuation of DDC-feeding, while isoform changes in many other proteins normalize. Further studies are needed to distinguish whether the observed persistent changes are due to an irreversible alteration coupled with a remarkably slow chaperone turnover or due to an ongoing chaperone modification despite withdrawal of the drug. In addition to chaperone modification, DDCprimed and DDC-refed mice posses numerous other alterations in gene expression and protein levels (present study),32 which may influence MDB formation in DDC-primed livers. In addition to chaperone function, chaperone expression levels are also important. Chronic DDC-feeding diminishes the levels of Hsp72 and GRP75 at the protein and messenger RNA levels (Fig. 2 and Table 2). Although HSF1/HSF2 messenger RNA levels were unaltered in DDC-fed mice (Table 2), impaired HSF transcriptional activity might be responsible, because heat shock factor (HSF) transcriptional activity is regulated at multiple posttranslational levels.33 Similar to DDC, the human MDB inducer ethanol leads to no or limited chaperone induction,34,35 whereas liver chaperone levels are regularly elevated in multiple acute and chronic stress conditions.36-38 Therefore, the low chaperone levels in DDC-fed or chronic alcohol-consuming individuals may facilitate aggregate formation by decreasing chaperone function (Fig. 3) and levels (Fig. 2). Similarly, depleted chaperone levels were observed in aggregate-forming mouse models of Huntington disease39 and amyotrophic lateral sclerosis,40 and chaperone overexpression successfully suppressed aggregate generation.41,42 In contrast, Hsp25 was up-regulated, particularly in DDC-refed mice. Although both chaperones potentially protect cells from protein aggregation,43 Hsp70 may be more important than Hsp25 in MDB formation given its interaction with K8/1844 and its easier detectability in MDBs.10,45 We propose an important role for chaperones during the stepwise process of MDB formation (Fig.8). Initially, several early changes (such as keratin overexpression with an increased K8⬎K18 ratio) occur that are not sufficient for MDB formation.24 In parallel, elevated oxidative stress16,17,46 is initially “absorbed” via chaperone stabilization of misfolded proteins. As chaperone function becomes compromised and insufficient (findings herein), progressive keratin misfolding

940

STRNAD ET AL.

HEPATOLOGY, September 2008

Fig. 7. Double immunofluorescence staining of K8/K18 (red) and K19 (green) highlights the increased number of K19-positive cells in Rec 1M, R1M⫹DDC7d and R3M⫹DDC7d mice when compared with the other treatment conditions. Bar ⫽ 100 ␮m.

occurs47,48 together with K8 cross-linking24,49 that ultimately leads to keratin aggregation and MDB formation (Fig. 8). Upon mouse recovery after switching to a normal diet, keratin expression normalizes and K8 cross-linking disappears except for some remnants (Fig. 6) together with MDB resolution. However, the retained chaperone dysfunction predisposes to rapid and relatively unprotected keratin misfolding, keratin cross-linking, p62 accumulation, and MDB reformation upon repeated DDC exposure (Fig. 6). Persistent chaperone damage may therefore provide one explanation for why rapid MDB reformation can be induced

by a variety of stresses in primed mice that would otherwise not induce MDBs in unprimed mice. Further studies using animals with altered chaperone levels/ function are needed to clarify their role in MDB formation. While our study focused on Hsp60/70, our findings do not exclude the possibility that other chaperones may play similar or even more important roles. Acknowledgment: We are grateful to Evelyn Resurrection for assistance with immune staining, Kris Morrow for help with figure preparation, and Dr. Rainer Muche (University of Ulm) for advice in statistical analysis.

HEPATOLOGY, Vol. 48, No. 3, 2008

STRNAD ET AL.

941

Fig. 8. Persistent chaperone dysfunction associates with rapid MDB reformation upon DDC refeeding. The schematic highlights the major molecular events that occur during MDB formation, recovery, and rapid MDB reformation. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com]

References 1. Bershadsky AD, Vasiliev JM. Cytoskeleton. New York: Plenum Press, 1988:133-154. 2. Ku N-O, Zhou X, Toivola DM, Omary MB. The cytoskeleton of digestive epithelia in health and disease. Am J Physiol 1999;277: G1108-G1137. 3. Herrmann H, Ba¨r H, Kreplak L, Strelkov SV, Aebi U. Intermediate filaments: from cell architecture to nanomechanics. Nat Rev Mol Cell Biol 2007;8:562-573. 4. Kim S, Coulombe PA. Intermediate filament scaffolds fulfill mechanical, organizational, and signaling functions in the cytoplasm. Genes Dev 2007; 21:1581-1597. 5. Omary MB, Coulombe PA, McLean WH. Intermediate filament proteins and their associated diseases. N Engl J Med 2004;351:2087-2100. 6. Goebel HH. Congenital myopathies with inclusion bodies: a brief review. Neuromuscul Disord 1998;8:162-168. 7. Ross CA, Poirier MA. Protein aggregation and neurodegenerative disease. Nat Med 2004;10:S10-S17. 8. Der Perng M, Su M, Wen SF, Li R, Gibbon T, Prescott AR, et al. The Alexander disease-causing glial fibrillary acidic protein mutant, R416W, accumulates into Rosenthal fibers by a pathway that involves filament aggregation and the association of alpha B-crystallin and HSP27. Am J Hum Genet 2006;79:197-213. 9. Zatloukal K, French SW, Stumptner C, Strnad P, Harada M, Toivola DM, et al. From Mallory to Mallory-Denk bodies: what, how and why? Exp Cell Res 2007;313:2033-2049. 10. Zatloukal K, Stumptner C, Fuchsbichler A, Heid H, Schnoelzer M, Kenner L, et al. p62 Is a common component of cytoplasmic inclusions in protein aggregation diseases. Am J Pathol 2002;160:255-263. 11. Macario AJ, Conway de Macario E. Sick chaperones, cellular stress, and disease. N Engl J Med 2005;353:1489-1501. 12. Jensen K, Gluud C. The Mallory body: morphological, clinical and experimental studies (Part 1 of a literature survey). HEPATOLOGY 1994;20:10611077.

13. Ku NO, Strnad P, Zhong BH, Tao GZ, Omary MB. Keratins let liver live: mutations predispose to liver disease and crosslinking generates MalloryDenk bodies. HEPATOLOGY 2007;46:1639-1649. 14. Denk H, Stumptner C, Zatloukal K. Mallory bodies revisited. J Hepatol 2000;32:689-702. 15. Bukau B, Weissman J, Horwich A. Molecular chaperones and protein quality control. Cell 2006;125:443-451. 16. Mehta K, Van Thiel DH, Shah N, Mobartan S. Nonalcoholic fatty liver disease: pathogenesis and the role of antioxidants. Nutr Rev 2002;60:289293. 17. Dey A, Cederbaum AI. Alcohol and oxidative liver injury. HEPATOLOGY 2006;43(Suppl):S63-S74. 18. Farrell GC, Larter CZ. Nonalcoholic fatty liver disease: from steatosis to cirrhosis. HEPATOLOGY 2006;43(Suppl):S99-S112. 19. Li J, Bardag-Gorce F, Dedes J, French BA, Amidi F, Oliva J, et al. Sadenosylmethionine prevents mallory denk body formation in drugprimed mice by inhibiting the epigenetic memory. HEPATOLOGY 2008;47: 613-624. 20. Georgopoulos C, Welch WJ. Role of the major heat shock proteins as molecular chaperones. Annu Rev Cell Biol 1993;9:601-634. 21. Ku NO, Toivola DM, Zhou Q, Tao GZ, Zhong B, Omary MB. Studying simple epithelial keratins in cells and tissues. Methods Cell Biol 2004;8: 489-517. 22. Nakamichi I, Toivola DM, Strnad P, Michie SA, Oshima RG, Baribault H, et al. Keratin 8 overexpression promotes mouse Mallory body formation. J Cell Biol 2005;171:931-937. 23. Gething MJ, Sambrook J. Protein folding in the cell. Nature 1992;355: 33-45. 24. Strnad P, Harada M, Siegel M, Terkeltaub RA, Graham RM, Khosla C, et al. Transglutaminase-2 regulates Mallory body inclusion formation and liver size. Gastroenterology 2007;132:1515-1526. 25. Van Kleef FS, De Jong WW, Hoenders HJ. Stepwise degradations and deamidation of the eye lens protein alpha-crystallin in ageing. Nature 1975;258:264-266.

942

STRNAD ET AL.

26. Takemoto L, Boyle D. Increased deamidation of asparagine during human senile cataractogenesis. Mol Vis 2000;6:164-168. 27. Boros S, Wilmarth PA, Kamps B, de Jong WW, Bloemendal H, Lampi K, et al. Tissue transglutaminase catalyzes the deamidation of glutamines in lens betaB(2)- and betaB(3)-crystallins. Exp Eye Res 2008;86:383-393. 28. Nilsson MR, Driscoll M, Raleigh DP. Low levels of asparagine deamidation can have a dramatic effect on aggregation of amyloidogenic peptides: implications for the study of amyloid formation. Protein Sci 2002;11:342-349. 29. Carbone DL, Doorn JA, Kiebler Z, Sampey BP, Petersen DR. Inhibition of Hsp72-mediated protein refolding by 4-hydroxy-2-nonenal. Chem Res Toxicol 2004;17:1459-1467. 30. Carbone DL, Doorn JA, Kiebler Z, Ickes BR, Petersen DR. Modification of heat shock protein 90 by 4-hydroxynonenal in a rat model of chronic alcoholic liver disease. J Pharmacol Exp Ther 2005;315:8-15. 31. Moon KH, Hood BL, Kim BJ, Hardwick JP, Conrads TP, Veenstra TD, et al. Inactivation of oxidized and S-nitrosylated mitochondrial proteins in alcoholic fatty liver of rats. HEPATOLOGY 2006;44:1218-1230. 32. Bardag-Gorce F, Dedes J, French BA, Oliva JV, Li J, French SW. Mallory body formation is associated with epigenetic phenotypic change in hepatocytes in vivo. Exp Mol Pathol 2007;83:160-168. 33. Voellmy R. Feedback regulation of the heat shock response. Handb Exp Pharmacol 2006;172:43-68. 34. Calabrese V, Renis M, Calderone A, Russo A, Barcellona ML, Rizza V. Stress proteins and SH-groups in oxidant-induced cell damage after acute ethanol administration in rat. Free Radic Biol Med 1996;20:391-397. 35. Rakonczay Z Jr, Boros I, Ja´rmay K, Hegyi P, Lonovics J, Takacs T. Ethanol administration generates oxidative stress in the pancreas and liver, but fails to induce heat-shock proteins in rats. J Gastroenterol Hepatol 2003;18: 858-867. 36. Fujio N, Hatayama T, Kinoshita H, Yukioka M. Induction of mRNAs for heat shock proteins in livers of rats after ischemia and partial hepatectomy. Mol Cell Biochem 1987;77:173-177. 37. Broome´ U, Scheynius A, Hultcrantz R. Induced expression of heat-shock protein on biliary epithelium in patients with primary sclerosing cholangitis and primary biliary cirrhosis. HEPATOLOGY 1993;18:298-303. 38. Lee KJ, Terada K, Oyadomari S, Inomata Y, Mori M, Gotoh T. Induction of molecular chaperones in carbon tetrachloride-treated rat liver: implications in protection against liver damage. Cell Stress Chaperones 2004;9:58-68.

HEPATOLOGY, September 2008

39. Hay DG, Sathasivam K, Tobaben S, Stahl B, Marber M, Mestril R, et al. Progressive decrease in chaperone protein levels in a mouse model of Huntington’s disease and induction of stress proteins as a therapeutic approach. Hum Mol Genet 2004;13:1389-1405. 40. Maatkamp A, Vlug A, Haasdijk E, Troost D, French PJ, Jaarsma D. Decrease of Hsp25 protein expression precedes degeneration of motoneurons in ALS-SOD1 mice. Eur J Neurosci 2004;20:14-28. 41. Cummings CJ, Mancini MA, Antalffy B, DeFranco DB, Orr HT, Zoghbi HY. Chaperone suppression of aggregation and altered subcellular proteasome localization imply protein misfolding in SCA1. Nature Genet 1998; 19:148-154. 42. Vacher C, Garcia-Oroz L, Rubinsztein DC. Overexpression of yeast hsp104 reduces polyglutamine aggregation and prolongs survival of a transgenic mouse model of Huntington’s disease. Hum Mol Genet 2005; 14:3425-3433. 43. Meriin AB, Sherman MY. Role of molecular chaperones in neurodegenerative disorders. Int J Hyperthermia 2005;21:403-419. 44. Liao J, Lowthert LA, Ghori N, Omary MB. The 70-kDa heat shock proteins associate with glandular intermediate filaments in an ATP-dependent manner. J Biol Chem 1995;270:915-922. 45. Janig E, Stumptner C, Fuchsbichler A, Denk H, Zatloukal K. Interaction of stress proteins with misfolded keratins. Eur J Cell Biol 2005; 84:329-339. 46. Tephly TR, Coffman BL, Ingall G, Ziet-Har MS, Goff HM, Tabba HD, et al. Identification of N-methylprotoporphyrin IX in livers of untreated mice and mice treated with 3, 5-diethoxycarbonyl-1, 4-dihydrocollidine: source of the methyl group. Arch Biochem Biophys 1981; 212:120-126. 47. Cadrin M, French SW, Wong PT. Alteration in molecular structure of cytoskeleton proteins in griseofulvin-treated mouse liver: a pressure tuning infrared spectroscopy study. Exp Mol Pathol 1991;55:170-179. 48. Kachi K, Wong PT, French SW. Molecular structural changes in Mallory body proteins in human and mouse livers: an infrared spectroscopy study. Exp Mol Pathol 1993;59:197-210. 49. Zatloukal K, Fesus L, Denk H, Tarcsa E, Spurej G, Bock G. High amount of epsilon-(gamma-glutamyl)lysine cross-links in Mallory bodies. Lab Invest 1992;66:774-777.