Gastroenterology Volume 135 issue 3 2008 [doi 10.1053%2Fj.gastro.2008.05.055] Pierre–Emmanuel Rautou; Dominique Cazals–Hatem; Richard More -- Acute Liver Cell Damage in Patients

8/13/2019 Gastroenterology Volume 135 issue 3 2008 [doi 10.1053%2Fj.gastro.2008.05.055] PierreEmmanuel Rautou; Dom

1/12

Acute Liver Cell Damage in Patients With Anorexia Nervosa: A PossibleRole of Starvation-Induced Hepatocyte Autophagy

PIERREEMMANUEL RAUTOU,*, DOMINIQUE CAZALSHATEM, RICHARD MOREAU,*, CLAIRE FRANCOZ,*GRARD FELDMANN, DIDIER LEBREC,*, RIC OGIERDENIS, PIERRE BEDOSSA,, DOMINIQUE VALLA,*, andFRANOIS DURAND*,

*Ple des Maladies de lAppareil Digestif, Service dHpatologie, Hpital Beaujon, AP-HP, Clichy, France; INSERM U773, Centre de Recherche Biomdicale Bichat

Beaujon CRB3, Universit Paris 7-Denis-Diderot, Paris, France; Service dAnatomie-Pathologique, Hpital Beaujon, Clichy, France

Background & Aims: Acute liver insufficiency is arare complication of anorexia nervosa. The mecha-nisms for this complication are unclear. The aim ofthis study was to describe patient characteristics and

clarify the mechanisms involved. Methods: Liverspecimens from 12 patients (median age, 24 years;median body mass index, 11.3 kg/m2), with a pro-thrombin index 1.7 and anorexia nervosa as the onlycause for acute liver injury were analyzed. A detailedpathologic examination was performed, includingunder electron microscopy.Results:Liver cell glyco-gen depletion was a constant finding. There was acontrast between the increase in serum alanine ami-notransferase (56 times normal on average; 1,904IU/L) and the absence of significant hepatocyte ne-crosis on histology. Centrilobular changes (trabecu-lar atrophy and/or sinusoidal fibrosis) were observedin 6 patients. There were rare or no (


2/12

encephalopathy at admission. Twelve of these 486 pa-tients had anorexia nervosa as the only identifiable causeof acute liver disease.

The diagnosis of anorexia nervosa was based on thecriteria of the Diagnostic and Statistical Manual of Mental

Disorders, 4th edition.14

These criteria include refusal tomaintain body weight at or above a minimally normalweight for age and height, intense fear of gaining weightor becoming fat even though underweight, disturbancein the way the persons body weight or shape is experi-enced, undue influence of body weight or shape on self-evaluation, or denial of the severity of the existing lowbody weight, and amenorrhea.

Patient Management

All patients were given N-acetylcysteine intrave-nously until the prothrombin index exceeded 50% of

normal (INR1.7). Glucose (5%15%) was infused de-pending on the glucose concentration at admission. Pa-tients received enteral supplementation starting from 0(4 patients) to 500 kcal/day until the prothrombin indexnormalized. Thereafter, calories were gradually increased.Hypokalemia, hypoglycemia, hypophosphoremia, andhypomagnesemia were corrected as needed.

Ten patients underwent a detailed hemodynamic as-sessment by transthoracic echocardiography and/or rightheart catheterization via a Swan-Ganz catheter. Liver bi-opsy was performed within 19 days (median, 2) afterpeak aspartate aminotransferase (AST) owing to the se-verity of liver damage and uncertainties concerning the

cause. One patient underwent ultrasound-guided percu-taneous biopsy and 11 patients underwent transjugularbiopsy.

Histology and Immunostaining

Tissue specimens were fixed in formalin and paraf-fin embedded. We cut 5-mm-thick sections from paraffinblocks and stained them with hematoxylin and eosin, Mal-lory trichrome, Picrosirius red, Perls, and Periodic-acid-Schiff (PAS). Frozen sections were available in 10 patientsand stained with Oil-Red O to evaluate steatosis. The fol-lowing variables were analyzed on histology: fibrosis, lobu-

lar inflammation, centrilobular necrosis or atrophy of he-patocyte, hepatocytic swelling and clarification, steatosis(Oil Red O), PAS staining evaluating glycogen deposition,and ceroid pigments. Apoptotic cells were identified by theterminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) assay, using the Apop Tag Peroxy-dase Kit S7101 (Chemicon International, Temecula, CA).

Because of the extremely poor nutritional status andmajor alteration of liver cell function in these patients,hepatocytes were analyzed for proteins involved in autoph-agy. Genetic studies on yeast have identified 30 AuToph-aGy-related (Atg) genes required for the formation of auto-

phagosomes.15

Most of the Atg genes are conserved inhigher eukaryotes. Immunohistochemical analyses were

performed in 9 patients with anorexia nervosa using theATG5 (AuTophaGy-related 5) antibody (Abgent, San Diego,CA; 1:100 dilution). ATG5 immunostaining identifies auto-phagy protein 5-like, which is a key protein in autophago-some formation, constitutively expressed in all cells.1618

Because the accumulation of newly synthesized unfoldedproteins resulting in endoplasmic reticulum stress can trig-ger autophagy under certain conditions,19 evidence of en-doplasmic reticulum stress was investigated. KDEL is asequence present at the carboxy-terminus of soluble endo-plasmic reticulum resident proteins: glucose-related protein(GRP) 78 (also known as BIP) and GRP 94. BIP is a keyregulator of endoplasmic reticulum function. Immunohis-tochemical analyses were performed in the hepatocytes of 9patients using the KDEL antibody (Calbiochem, Darm-stadt, Germany) at a 1:2,000 dilution according to themanufacturers instructions.

Electron Microscopy

Part of the liver specimen from 4 patients in thestudy population was fixed in a 2% solution of glutaralde-hyde buffered with 0.2 mmol/L cacodylate and post fixed inosmium tetroxide before embedding in epoxy resin for elec-tron microscopy. Ultrathin sections stained with uranylacetate and lead citrate were examined with an electronmicroscope Jeol 10 10 (Tokyo, Japan).

Controls

Ten patients with histologically normal livers wereused as controls for immunohistochemical analysis. Detailson these patients are given inSupplementary Table 1(avail-able online atwww.gastrojournal.org). Eleven patients wereused as controls for electron microscopy. Five of these 11patients had histologically normal livers and 6 patients hadvarious causes of liver cell damage characterized by necrosisand/or apoptosis. Details on this group are given inSup-plementary Table 2.Analyses were performed by investiga-tors who were unaware of patient groups. The study con-formed to the ethical guidelines of the 1975 Declaration ofHelsinki.

Results

Patient Characteristics

As shown in Table 1, most patients were youngfemales, with a median BMI of 11.3. BMI at admissioncorresponded in all cases to the lowest BMI the patients hadever had. Six of 12 patients had severe hypoglycemia andcoma at presentation. Concomitant medications are listedinTable 1. None of the patients had any symptoms sug-gesting underlying chronic liver disease.

Detailed results of all diagnostic tests are presented inSupplementary Table 3. Briefly, none of the patients hadmarkers of infection for hepatitis A, B, or C viruses, signif-

icant autoantibody titers, alcohol abuse, diabetes, or hyper-lipidemia. None of the patients had ingested paracetamol

CLINICALLIVER,

P A N C R E A S

A N D

September 2008 ACUTE LIVER INSUFFICIENCY IN ANOREXIA NERVOSA 841
http://www.gastrojournal.org/http://www.gastrojournal.org/http://www.gastrojournal.org/


3/12

(acetaminophen) or aspirin within 7 days before admission.Acetaminophen was undetectable in plasma at admission inall patients. The liver was homogenous in all cases onultrasound (Table 2). Moderate ascites was observed in 4patients. No focal lesions suggesting liver infarction wereobserved. There was no sign of portal hypertension.

Laboratory data are shown inTables 2 and 3. Medianvalues for prothrombin index and factor V at admissionwere 31% and 51%, respectively. Median values for peak ASTand alanine aminotransferase (ALT) were 67 and 56 timesthe upper limit of normal, respectively (the upper limit ofnormal values in women and men are 31 and 35 IU/L,respectively, for AST, and 34 and 45 IU/L, respectively, forALT). At admission serum creatinine was at least twice thebaseline value (before admission) in 8 patients. Serum ureawas also at least twice the baseline value in all patients.

Median serum phosphorus was 0.72 mmol/L (range, 0.151.60); 6 patients had values below the lower limit of normal.

Hemodynamic Status

Individual values of serum troponin I and creatininekinase are listed inTable 3.Four patients (patients 4, 5, 6,and 7) had a significant increase in troponin I (0.1g/L),suggesting some myocardial cell damage, although con-comitant rhabdomyolysis could also have contributed tothis increase.

Two patients (patients 3 and 8) had mild hypoxemia.Nine patients had electrocardiographic changes related tohypokalemia (Table 4). In these patients, median serumpotassium was 3.1 mmol/L (range, 2.73.8). Patients whounderwent echocardiography (n 7) and/or right heartcatheterization (n 7), had altered left ventricular function

Table 1. Clinical Data in 12 Patients With Anorexia Nervosa and Acute Liver Insufficiency

Patient

Age (y)/

gender

Weight

(kg)

Admission

BMI Presenting symptoms

Relevant drug history:

medication/interval between

introduction of treatment and

acute liver insufficiency

1 18/F 34.0 11.9 General asthenia No2 26/M 41.5 13. 0 Hypog lycemic coma/biological assessment dur ing refeeding No

3 22/F 44.0 NA Emesis No

4 31/F 29.6 10. 6 Hypog lycemic coma/biological assessment dur ing refeeding NA

5 28/F 24.4 11.1 Biological assessment during refeeding No

6 23/F 30.0 11.6 Hypoglycemic coma/biological assessment during refeeding Venlafaxine, fluvoxamine/2 mos

7 27/M 30.5 11.3 Hypoglycemic coma Zopiclone/NA

8 47/F 29.0 11.5 Hypoglycemic coma No

9 24/F 24.0 9.6 Biological assessment during refeeding No

10 22/F 34.3 11.6 General asthenia No

11 19/F 27.0 10.3 Syncope Sertraline/2 mos, oxazepam/

6 mos, trimebutine/9 mos

12 21/F 25.6 9. 2 Hypog lycemic coma/biological assessment dur ing refeeding No

BMI, body mass index; F, female; M, male; NA, not available.

Table 2. Liver Tests and Hepatic Imaging Data at Admission in 12 Patients With Anorexia Nervosa and Acute Liver

Insufficiency

Patient

Serum AST

level

(ULN)

Serum ALT

level

(ULN)

Prothrombin

index (%)

Coagulation

factor V (%)

Serum bilirubin

(mol/L)

(normal 17

mol/L)

Previous episodes

of increase in

aminotransferase

level

Hepatic imaging

data

1 32 24 50 52 25 No Moderate ascites

2 84 48 30 50 11 Yes Moderate ascites

3 101 181 29 33 46 Yes Normal

4 106 73 44 55 35 Yes Normal

5 71 83 39 28 59 Yes Normal

6 62 41 31 30 22 No Normal

7 46 21 30 69 18 Yes Normal

8 18 10 19 10 4 No Moderate ascites

9 95 80 47 72 14 Yes Moderate ascites

10 31 63 40 63 22 No Normal

11 87 70 30 63 12 Yes Normal

12 55 20 29 38 38 Yes Normal

ALT, alanine aminotransferase; AST, aspartate aminotransferase; ULN, upper limit of normal.

The upper limit of normal for AST was 31 UI/L in women and 35 UI/L in men. The upper limit of normal for ALT was 34 IU/L in women and 45IU/L in men.

CLINICALLIVER,

PANCRE AS

AND

842 RAUTOU ET AL GASTROENTEROLOGY Vol. 135, No. 3


4/12

(decreased cardiac index and/or decreased left ventricularejection fraction). However, none of the patients had evi-dence of right heart failure and/or central venous pressure5 mm Hg. The hepatic venous pressure gradient wasmeasured in 10 patients and was 6 mm Hg in all cases(Table 4).

Histologic Findings and Immunostaining forMarkers of Apoptosis

Histologic data are summarized inTable 5.Noneof the 12 patients had portal fibrosis, portal inflamma-tory infiltrates, or bile duct lesions. Mild lobular inflam-mation was present in only 1 patient. Swelling and clar-ified hepatocytes were observed in all cases. Thesechanges were marked in 5 of 12 patients. PAS stainingshowed decreased glycogen in all patients. Glycogen de-pletion was more pronounced in the 5 patients withmarked hepatocyte swelling (Figure 1A). One of these

patients had mild microvesicular steatosis confirmed bySoudan red staining. No macrovesicular steatosis wasobserved.

Centrilobular liver cell atrophy associated with mildsinusoidal fibrosis was observed in 5 patients (Figure 1B).

Mild centrilobular necrosis without inflammation orcongestion was only observed in 1 patient (patient 3) who

also had microvesicular steatosis. Ceroid pigments wereobserved in all patients, predominating in centrilobular

zone hepatocytes. Perls staining was negative in all but 1

patient (patient 10). Overall, there was no evidence formassive hypoxic hepatitis or congestion.

TUNEL-positive cells were only present in 3 of 12patients, showing that apoptosis was not the leading

mechanism. In these patients, stained hepatocyte nucleiwere limited to centrilobular zones. TUNEL-positive cellswere observed in 5% of liver cells in 1 patient (patient 6)

Table 3. Laboratory Data at Admission in 12 Patients With Anorexia Nervosa and Acute Liver Insufficiency

Patient

Admission K

(mmol/L)

Increase in

serum creatinine

from baselinea

Increase in blood

urea nitrogen

from baselineaCKb

(IU/L)

Troponin I (g/L)

(normal, 0.04)

Lowest glucose

concentration

during hospital

stay (mmol/L)

1 3.4 NA NA 195 0.00 5.02 2.9 2 12.1 1020 0.00 1.1

3 3.1 1.4 3.2 195 0.00 4.7

4 2.8 2.5 14.7 285 0.14 1.5

5 3.2 2 11.9 1770 0.18 1.4

6 2.7 2.1 8.4 1110 0.64 1.0

7 3.8 2.4 12.1 5835 0.41 1.1

8 3.3 2 16.3 510 NA 2.2

9 2.8 3.3 9.3 90 0.00 2.9

10 3.3 2.5 11.0 135 0.00 2.6

11 3.3 1.3 16.4 1254 0.00 3.4

12 2.9 1.8 14.5 150 0.00 1.4

NA, not available.aExpressed in times the baseline value.b

Creatinine phosphokinase (normal value 150 IU/L).

Table 4. Hemodynamic Data and Electrocardiographic Changes in 12 Patients With Anorexia Nervosa and Acute Liver

Insufficiency

Patient

MAP

(mmHg)

LVEF

(%)

CI

(Lmin1m2)

CVP

(mmHg)

PCWP

(mmHg)

HVPG

(mmHg)

QT interval

prolongation

T-wave

flattening

Normal range 5575 2.83.6 110 012 5

1 67 84 3.45 0 7 NA No No

2 90 50 2.06 0 4 4 Yes No

3 70 NA NA 3 NA 1 Yes No

4 NA NA NA NA NA 2 NA NA

5 61 60 NA NA NA NA Yes Yes

6 59 25 1.7 5 10 1 No Yes

7 58 49 NA 0 NA 0 Yes Yes

8 56 15 1.56 5 9 1 No No

9 80 66 NA NA NA 4 Yes No

10 77 NA 1.59 0 6 6 No Yes

11 72 NA 1.84 0 3 2 Yes Yes

12 52 NA 1.73 0 2 2 Yes Yes

CI, cardiac index; CVP, central venous pressure; HVPG, hepatic venous pressure gradient; LVEF, left ventricular ejection fraction; MAP, meanarterial pressure; NA, not available; PCWP, pulmonary capillary wedge pressure.

CLINICALLIVER,

P A N C R E A S

A N D



5/12

and were scarce in the 2 remaining patients (patients 3and 5).

Electron Microscopy

Results of electron microscopy in the liver speci-mens from 4 patients (patients 1, 5, 11, and 12) showeda low density of glycogen particles in swollen hepatocytes(Figure 1C). The endoplasmic reticulum was slightly di-lated in some cells.

Numerous hepatocytes in all 4 specimens containedautophagosomes, a hallmark of autophagy. Autophago-somes are membranous vesicular compartments seques-tering cytoplasmic components, including mitochondriaand endoplasmic reticulum (Figure 1D). During the au-tophagic process, cytosolic material is sequestered by anexpanding membrane, the phagophore, resulting in adouble-membrane vesicle or autophagosome. The nas-cent autophagic vacuole or autophagosome is thenloaded with various acid hydrolases by fusion with pre-existing lysosomes to form autolysosomes.12 In the liver

specimens in this series, numerous mature lysosomescontaining dense osmiophilic bodies as a result of theautophagic process were visible in most hepatocytes (Fig-ure 1E) and, occasionally, in Kupffer cells. As a conse-quence of this process, a low density of organelles wasfound in the hepatocyte cytoplasm in all patients. Mostmitochondria seemed normal, except for rare giant or-ganelles. As shown in Figure 1F, abnormal, irregularlyshaped nuclei with disorganized chromatin were alsovisible in some hepatocytes with autophagosomes in thecytoplasm. These changes, associated with the loss inelectron density from the breakdown of cytoplasmic

components, as well as the decreased number of mito-chondria (although those present in the cytoplasm ap-

peared to be normal) are evidence of autophagic celldeath.20

No autophagosomes were observed in the hepatocytesin any of the 11 control patients with electron micros-copy (Supplementary Table 1).

Immunostaining for Proteins Involved inAutophagy and Endoplasmic Reticulum Stress

Similar ATG5 expression was found in the hepa-tocytes from the study population and the control sub-jects (Table 5 and Supplementary Table 2) with immu-nohistochemistry, suggesting that ATG5 was availablefor the autophagic process in patients with anorexianervosa as well as in those without.

The KDEL sequence was detected in all patients in thestudy population versus only 1 out of 10 controls ( Table5and Supplementary Table 2). BIP contains the KDELsequence and is a key regulator of endoplasmic reticulumfunction. It is involved in protein folding and in main-tenance of the permeability barrier of the endoplasmic

reticulum, and it functions as a folding sensor.21 Thesedata are consistent with the existence of some endoplas-mic reticulum stress in our patients.

Outcome

None of the 12 patients developed encephalopa-thy. Renal function returned to normal within 6 days inall patients. Within 1 month, all patients with initialcardiac dysfunction recovered with a normal ejectionfraction. Only patient 6 required dobutamine infusionfor 6 consecutive days.

A 50% decrease in AST and ALT levels from peak values

was observed within 2 and 5 days, respectively(Figure 2).The prothrombin index was 50% of normal in all pa-

Table 5. Histologic Data in 12 Patients With Anorexia Nervosa and Acute Liver Insufficiency

Patients

Time from

admission

to biopsy

(days) Fibrosis

Lobular

inflammation

Centrilobular

lesions

Hepatocytic

swelling

Oil red

O (%)

PAS:

glycogen

deficit

PAS: ceroid

pigments

Percentage of

TUNEL-positive

nuclei per

lobule (%) ATG5a KDELb

1c

1 0 0 0 Marked 0 Marked Mild 0 NA NA2 2 CL 0 Atrophy Mild 0 Mild Mild 0 NA NA

3 2 0 0 Necrosis Marked 80 Marked Mild 2 (CL) 0 Mild

4 9 0 0 0 Moderate 0 Mild Marked 0 NA NA

5c 3 0 Mild 0 Moderate 0 Moderate Marked 2 (CL) Mild CL Marked

6 2 CL 0 Atrophy Moderate 0 Moderate Mild 5 (CL) Moderate

CL

Mild

7 2 0 0 0 Marked 0 Marked Mild 0 Mild CL Mild

8 2 CL 0 Atrophy Marked NA Marked Marked 0 Mild CL Marked

9 6 0 0 Atrophy Mild NA Moderate Marked 0 Mild CL Mild

10 2 CL 0 0 Marked 0 Marked Marked 0 0 Marked

11c 2 0 0 0 Moderate 10 Mild Mild 0 Mild CL Moderate

12c 1 CL 0 Atrophy Mild 0 Mild Marked 0 Mild CL Moderate

CL, centrilobular; NA, not available; TUNEL, terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling.a

Detection of autophagy-related 5 using immunohistochemistry.bDetection of KDEL sequence using immunohistochemistry.cElectron microscopy analysis.CL

INICALLIVER,

PANCRE AS

AND



6/12

Figure 1. Light (A,B) and electron microscopy (C,D,E,F). (A) Diffuse hepatocytic swelling due to a marked glycogen deficit according to the PASstaining (original magnification, 250). (B) Centrilobular fibrosis and atrophy better identified after reticulin staining (original magnification, 250;Gordon Sweet stain). (C) Hepatocytes observed by electron microscopy show low density of organelles and absence of glycogen in the cytoplasm(original magnification,7500).(D) Partial view of a hepatocyte. Disse space (D) with collagen (C). In thecenter of thefigure, a typical autophagosome(arrows) sequestering cytoplasmic components such as a fragment of mitochondrion (M) and circular membranes (original magnification,20,000).(E) Low-magnification image showing a hepatocyte containing numerous mature lysosomes (L) (original magnification, 3000). N, nucleus. (F) Ahepatocyte with features of autophagic cell death, including several autophagosomes (arrows) visible in the cytoplasm, the irregular shape of thenucleus (N), and disorganized chromatin. Note that most mitochondria (m) appear normal (original magnification, 5000).

CLINICALLIVER,

P A N C R E A S

A N D



7/12

tients and the INR was 1.7 5 days after the peak AST.One month after the peak AST, the median prothrombinindex was 97% (range, 74118).

Ten of the 12 patients were still alive with a meanfollow-up of 46 42 months (range 1138). Patient 7died 8 months after discharge for unknown reasons whilethe BMI was 10.4. Patient 1 was lost to follow-up.

Discussion

This series shows that extremely poor nutritionalstatus from anorexia nervosa can lead to acute and pro-found alteration of liver function. Severe undernutritionwas constant in this population with the BMI 13 in allcases. Marked glycogen depletion was a constant findingon liver biopsies. Certain patients also had severe hypo-glycemia at admission, which is rarely reported duringanorexia nervosa.1,22 Our results indicate that the alteredliver function in these patients was a result of an acuteprocess rather than chronic underlying liver insufficiency.Indeed, none of the patients had clinical or morphologic(either macroscopic or microscopic) evidence of chronicliver disease. The hepatic venous pressure gradient, whenmeasured, was low. More important, liver function rap-idly returned to normal after early resuscitation and

refeeding.One could imagine that acute liver cell necrosis was the

main mechanism in these patients. In particular, ourresults seem to suggest the presence of hypoxic hepatitis-like lesions.23 Indeed, there was a rapid increase in serumtransaminases with AST levels above the ALT levels andassociated impairment in renal function was observed inmost patients. Seven patients who underwent cardiovas-cular investigations had a low cardiac index. Finally, bothserum transaminases and coagulation factors rapidly re-turned to normal or near-normal values, which occurs incases of hypoxic hepatitis after the cause has been cured.

However, this mechanism was not confirmed becausebiopsy did not show any significant centrilobular hepa-

tocellular necrosis or congestion, which are the hallmarksof hypoxic hepatitis. Focal necrosis resulting in smallliver infarcts, which might not have been seen with corebiopsy, is highly unlikely because the liver parenchymawas homogeneous on imaging in all patients. No areas of

decreased echogenicity were observed on ultrasound.24,25Moreover, the portal and hepatic veins and the hepaticartery were patent for physiologic blood flow on Dopplerultrasound.

Interestingly, markers of apoptosis, a mechanism fre-quently involved in various acute liver diseases,26,27 wereabsent or very scarce in this population. TUNEL stainingwas frankly positive in only 1 patient. Moreover, therewere no ultrastructural changes suggesting apoptosisin the liver specimens from 4 patients with electronmicroscopy.

Because our findings showed that neither liver cell

necrosis nor apoptosis was the leading mechanism, wespeculated that autophagy, an alternative mechanism ofcell damage, could be involved in these severely starvedpatients. Interestingly, we found numerous autophago-somes, a hallmark of autophagy, in the hepatocytes of the4 patients on electron microscopy.28 In contrast, no au-tophagosomes were observed in any of the control pa-tients, including those with evidence of liver necrosisand/or apoptosis. ATG5, a key protein for autophago-some formation, was also found in the study populationdespite the altered liver cell function (as reflected bydecreased coagulation factors). As mentioned, some liver

cells presented with the characteristic features of auto-phagic cell death that have been documented in bothanimals and humans.20 Owing to the rapidity of theterminal events resulting in cell death during autophagy,different cells engaged in the same process may presentdifferent patterns at a given time. This may explain whywe observed liver cells at different stages, namely, auto-phagy without signs of cell death and autophagy withsigns of cell death.29 During apoptosis, nuclei are con-densed and membrane integrity is preserved. Conversely,during autophagic cell death, the permeability of theplasma membrane is altered.30 Increased permeability of

the hepatocyte plasma membrane could explain, at leastin part, the contrast between a marked increase in serumtransaminases in these patients and the absence of exten-sive liver cell necrosis on histology. Slight dilatation ofthe endoplasmic reticulum was observed in some cells onelectron microscopy. Furthermore, KDEL expression wasconstant in the study population, whereas it was presentin only 1 out of the 10 controls. These observationssuggest the presence of some endoplasmic reticulumstress. This pathway, which is known to be influenced bystarvation,21 has also been shown to trigger autophagy.19

These findings suggest that starvation-induced authoph-

agy was the principal mechanism involved in liver celldamage in this population.

Figure 2. Course of AST elevations in 12 patients with anorexia ner-vosa and acute liver insufficiency.

CLINICALLIVER,

PANCRE AS

AND



8/12

Previous studies have shown that autophagy is a majorcellular pathway for the degradation of long-lived pro-teins, constituents of cytoplasm and organelles.12,31 Au-tophagy results in the formation of autophagosomes,16

which then frequently fuse with lysosomes where degra-

dation occurs. Autophagy can be involved in cell ho-meostasis. Indeed, basal levels of autophagy can be ob-served under nutrient-rich conditions.16 However, one ofthe best known roles of autophagy is its action duringnutrient starvation.16 In this context, the macromole-cules resulting from autophagy are presumably releasedfrom the vacuoles for reuse in the cytosol. These macro-molecules supply a source of continued biosynthesis.32

Autophagy may also play a role in mammalian cells todelay apoptosis during nutrient deprivation.33 Mice lack-ing Atg5, a key gene for autophagy, die during the neo-natal period, because autophagy is required for recycling

intracellular material after the interruption of the pla-cental blood supply.34 Autophagy is a well-known sur-vival strategy under stress conditions. However, recentstudies have shown that autophagy can effectively kill acell if the components are rapidly consumed.12,35 Thisseries does not show that autophagy resulted in massiveliver cell death. In these patients, there was a contrastbetween relatively mild liver cell damage on liver histol-ogy and a profound alteration in liver function. There-fore, these findings suggest that, in severe anorexianervosa, autophagy may result in major hepatocyte dys-function, even without massive cell death.

Although there was evidence that autophagy played a

pivotal role in this population, other mechanisms result-ing in liver cell injury might have been involved. Inparticular, a transient decrease in cardiac output couldhave been a contributing factor. Interestingly, in vivoanalysis of the response to nutrient starvation in animalshas shown that autophagy occurs in the heart.16,17Auto-phagy has also been described in the heart of patientswith dilated cardiomyopathy.36 Whether autophagy alsoplayed a role in the circulatory changes observed in somepatients requires further investigation.

In conclusion, anorexia nervosa with severe undernu-trition should be added to the list of conditions causing

acute liver insufficiency. This study shows that neitherliver cell necrosis nor apoptosis, which are the mainmechanisms leading to cell death during acute liver fail-ure, played a pivotal role in this specific population. Ourfindings strongly suggest that in these severely under-nourished patients, autophagy was the leading pathway,resulting in significant alteration in liver cell function.Further studies are needed on the role of autophagyduring other conditions in humans.

Supplementary Data

Note: To access the supplementary material ac-companying this article, visit the online version of

Gastroenterology at www.gastrojournal.org, and at doi:10.1053/j.gastro.2008.05.055

References

1. Treatment of patients with eating disorders, 3rd ed. American

Psychiatric Association. Am J Psychiatry 2006;163(7 suppl):4

54.

2. Umeki S. Biochemical abnormalities of the serum in anorexia

nervosa. J Nerv Ment Dis 1988;176:503506.

3. Komuta M, Harada M, Ueno T, et al. Unusual accumulation of

glycogen in liver parenchymal cells in a patient with anorexia

nervosa. Intern Med 1998;37:678 682.

4. Nordgren L, von Scheele C. Hepatic and pancreatic dysfunction in

anorexia nervosa: a report of two cases. Biol Psychiatry 1977;

12:681686.

5. Ratcliffe PJ, Bevan JS. Severe hypoglycaemia and sudden death

in anorexia nervosa. Psychol Med 1985;15:679 681.

6. Rivera-Nieves J, Kozaiwa K, Parrish CR, et al. Marked transami-

nase elevation in anorexia nervosa. Dig Dis Sci 2000;45:1959

1963.7. Ozawa Y, Shimizu T, Shishiba Y. Elevation of serum aminotrans-

ferase as a sign of multiorgan-disorders in severely emaciated

anorexia nervosa. Intern Med 1998;37:3239.

8. Di Pascoli L, Lion A, Milazzo D, et al. Acute liver damage in

anorexia nervosa. Int J Eat Disord 2004;36:114117.

9. Yamada Y, Fushimi H, Inoue T, et al. Anorexia nervosa with

recurrent hypoglycemic coma and cerebral hemorrhage. Intern

Med 1996;35:560563.

10. De Caprio C, Alfano A, Senatore I, et al. Severe acute liver

damage in anorexia nervosa: two case reports. Nutrition 2006;

22:572575.

11. Furuta S, Ozawa Y, Maejima K, et al. Anorexia nervosa with

severe liver dysfunction and subsequent critical complications.

Intern Med 1999;38:575579.

12. Baehrecke EH. Autophagy: dual roles in life and death? Nat Rev

Mol Cell Biol 2005;6:505510.

13. Durand F, Bernuau J. Acute liver failure. In Rodes J, Benhamou

JP, Blei A, et al, eds. The textbook of hepatology: from basic

science to clinical practice, 3rd ed. Boston: Blackwell; 2007:

12891312.

14. American Psychiatric Association. Diagnostic and statistical man-

ual of mental disorders, 4th ed, text revision. Washington, DC:

American Psychiatric Association; 2000.

15. Xie Z, Klionsky DJ. Autophagosome formation: core machinery

and adaptations. Nat Cell Biol 2007;9:11021109.

16. Mizushima N, Yamamoto A, Matsui M, et al. In vivo analysis of

autophagy in response to nutrient starvation using transgenic

mice expressing a fluorescent autophagosome marker. Mol Biol

Cell 2004;15:11011111.

17. Nakai A, Yamaguchi O, Takeda T, et al. The role of autophagy in

cardiomyocytes in the basal state and in response to hemody-

namic stress. Nat Med 2007;13:619624.

18. Hara T, Nakamura K, Matsui M, et al. Suppression of basal

autophagy in neural cells causes neurodegenerative disease in

mice. Nature 2006;441:885889.

19. Yorimitsu T, Nair U, Yang Z, et al. Endoplasmic reticulum stress

triggers autophagy. J Biol Chem 2006;281:30299 30304.

20. Bursch W. The autophagosomal-lysosomal compartment in pro-

grammed cell death. Cell Death Differ 2001;8:569581.

21. Schroder M, Kaufman RJ. The mammalian unfolded protein re-

sponse. Annu Rev Biochem 2005;74:739789.

22. Rich LM, Caine MR, Findling JW, et al. Hypoglycemic coma in

anorexia nervosa. Case report and review of the literature. ArchIntern Med 1990;150:894 895.

CLINICALLIVER,

P A N C R E A S

A N D

http://www.gastrojournal.org/http://www.gastrojournal.org/http://dx.doi.org/10.1053/j.gastro.2008.05.055http://dx.doi.org/10.1053/j.gastro.2008.05.055http://dx.doi.org/10.1053/j.gastro.2008.05.055http://dx.doi.org/10.1053/j.gastro.2008.05.055http://www.gastrojournal.org/


9/12

23. Henrion J, Schapira M, Luwaert R, et al. Hypoxic hepatitis: clinical

and hemodynamic study in 142 consecutive cases. Medicine

(Baltimore) 2003;82:392406.

24. Lev-Toaff AS, Friedman AC, Cohen LM, et al. Hepatic infarcts:

new observations by CT and sonography. AJR Am J Roentgenol

1987;149:8790.

25. Cook GJ, Crofton ME. Hepatic artery thrombosis and infarction:evolution of the ultrasound appearances in liver transplant recip-

ients. Br J Radiol 1997;70:248251.

26. Sun F, Hamagawa E, Tsutsui C, et al. Evaluation of oxidative

stress during apoptosis and necrosis caused by D-galactosamine

in rat liver. Biochem Pharmacol 2003;65:101107.

27. Jaeschke H, Lemasters JJ. Apoptosis versus oncotic necrosis in

hepatic ischemia/reperfusion injury. Gastroenterology 2003;125:

12461257.

28. Mizushima N. Methods for monitoring autophagy. Int J Biochem

Cell Biol 2004;36:24912502.

29. Lockshin RA, Zakeri Z. Apoptosis, autophagy, and more. Int

J Biochem Cell Biol 2004;36:24052419.

30. Yu L, Alva A, Su H, et al. Regulation of an ATG7-beclin 1 program

of autophagic cell death by caspase-8. Science 2004;304:

15001502.31. Yoshimori T. Autophagy: a regulated bulk degradation process

inside cells. Biochem Biophys Res Commun 2004;313:453

458.

32. Mizushima N, Ohsumi Y, Yoshimori T. Autophagosome formation

in mammalian cells. Cell Struct Funct 2002;27:421429.

33. Boya P, Gonzalez-Polo RA, Casares N, et al. Inhibition of macro-

autophagy triggers apoptosis. Mol Cell Biol 2005;25:

10251040.

34. Kuma A, Hatano M, Matsui M, et al. The role of autophagy during

the early neonatal starvation period. Nature 2004;432:10321036.

35. Martinet W, De Meyer GR, Andries L, et al. In situ detection of

starvation-induced autophagy. J Histochem Cytochem 2006;54:

8596.

36. Shimomura H, Terasaki F, Hayashi T, et al. Autophagic degener-

ation as a possible mechanism of myocardial cell death in dilated

cardiomyopathy. Jpn Circ J 2001;65:965968.

Received November 16, 2007. Accepted May 21, 2008.

Address requests for reprints to: Franois Durand, MD, Service

dHpatologie, Hpital Beaujon, 100 Boulevard du Gnral Leclerc, 92

110 Clichy, France. E-mail: [email protected]; fax: 33 1

47 30 44 55.

The authors thank Nolle Bendersky, MD and Caroline de Kergue-nec, MD for their help in the collection of data; Laurent Ledain, MD for

his cardiologic advice, and Alain Grodet for technical assistance in

electron microscopy.

CLINICALLIVER,

PANCRE AS

AND

mailto:[email protected]:[email protected]:[email protected]:[email protected]


10/12

Supplementary Table 1. Clinical, Biological, and Histologic Data From 10 Controls With Normal Liver Biopsy in Whom ATG5

and KDEL Staining Were Performed

Control

Age (y)/

Gender Indication for liver biopsy

Serum ALT

level

(ULN)a

Serum GGT

level

(ULN)

Liver biopsy

data ATG5b KDELc

1 65/F Liver nodule NA NA Normal Moderate CL 02 24/F Unexplained

aminotransferase level

elevation

3.1 3.8 Normal Mild CL Mild

3 79/F Liver nodule 0.8 0.5 Normal Moderate CL 0

4 52/F Unexplained biological

cholestasis

2.0 18.0 Normal Mild CL 0

5 66/F Hepatocellular carcinoma

developed on normal

liver. Assessment of

liver fibrosis before

hepatectomy

0.6 1.9 Normal 0 0

6 66/F Previous history of ovarian

and pancreatic cancers;

diffuse liver changes

(metastasis?)


7 62/M Unexplained arterial

hypoxemia. Search for

cirrhosis


8 52/M Assessment of liver

fibrosis before total

pancreatectomy (chronic

alcoholic pancreatitis)


9 41/M Eliminate contralateral

tumor involvement

before hepatectomy for

cholangiocarcinoma


10 25/M Unexplained liver tests

abnormalities

1.2 2.1 Normal 0 0

CL, centrilobular; F, female; GGT, -glutamyl transferase; M, male; NA: not available; ULN, upper limit of normal.aULN is 35 IU/L in women and 45 IU/L in men.bDetection of autophagy-related 5 using immunohistochemistry.cDetection of KDEL sequence using immunohistochemistry.

September 2008 ACUTE LIVER INSUFFICIENCY IN ANOREXIA NERVOSA 848.e1


11/12

Supplementary Table 2. Clinical, Biological, Histologic Data in 11 Controls With Electron Microscopy Analysis

Control

Age (y)/

Gender

Indication for liver

biopsy

Serum ALT

level

(ULN)a

Serum GGT

level

(ULN)b Liver biopsy data

Percentage of

TUNEL-

positive

nuclei per

lobule

Percentage of

necrotic cells

per lobule

Electron microscopy

analysis

11 22/F Mauriac syndrome 35.4 27.0 Glycogen

accumulation No

necrosis

0 0 Glycogen

accumulation

12 56/F Unexplained l iver

tests

abnormalities

1.2 8.0 Normal 0 0 Normal

13 21/M Increase in alkaline

phosphatase

level (context of

total parenteral

nutrition)

1.3 4.2 CL steatosis (40%) 0 0 Steatosis

14 30/M Liver tests

abnormalities in

a context of total

parenteralnutrition

NA NA Normal 0 0 Normal

15 50/M Increase in alkaline

phosphatase

level (context of

total parenteral

nutrition)

1.0 1.0 Steatosis (50%) 0 0 Steatosis

16 55/M Liver

retransplantation

for ischemic

cholangiopathy

0.6 9.3 Cirrhosis

Intrahepatic

venous

thrombosis

5% CL and

2% PP

2% Apoptotic and

necrotic cells

17 49/M Liver

transplantation

for alcoholic

cirrhosis and

portal veinthrombosis

1.0 2.0 Cirrhosis 10% CL and

1% ML

5% Phagolysosomes

18 37/F Liver

transplantation

for drug-induced

acute liver failure

30.3 1.3 Abundant necrosis;

inflammatory

infiltrate

2% 20% Numerous necrotic

cells. Altered

mitochondria

19 22/F Autoimmune

hepatitis

2.0 3.7 Periportal fibrosis;

inflammatory

infiltrate

2% 2% Few necrotic cells

20 30/F Ac ut e fatt y liver of

pregnancy

7.0 1.0 Steatosis (50%) 0 5% Phagolysosomes

21 17/M Acute liver

insufficiency due

to autoimmune

hepatitis

25.0 NA Abundant necrosis;

inflammatory

infiltrate

2% CL 15% Numerous necrotic

cells. Dilated

endoplasmic

reticulum

ALT, alanine aminotransferase; CL, centrilobular; GGT, -glutamyl transferase; ML, mediolobular; NA, not available; PP, periportal; ULN, upper

limit of normal.aULN is 35 IU/L in women and 45 IU/L in men.bULN is 38 UI/L in women and 55 UI/L in men.

848.e2 RAUTOU ET AL GASTROENTEROLOGY Vol. 135, No. 3


12/12

Supplementary Table 3. Results of all Diagnostic Tests Performed in 12 Patients With Anorexia Nervosa and Acute Liver

Insufficiency

Patients 1 2 3 4 5 6 7 8 9 10 11 12

Anti-Hepatitis A virus IgM 0 0 0 0 0 0 0 0 0 0 0 0

Hepatitis B virus surface antigen 0 0 0 0 0 0 0 0 0 0 0 0

Anti-hepatitis B virus core Ab 0 0 0 0 0 0 0 0 0 0 0 0Anti-hepatitis C virus Ab 0 0 0 0 0 0 0 0 0 0 0 0

Hepatitis C virus RNA NA 0 0 0 NA 0 0 NA NA 0 0 NA

Anti-VCA IgM NA 0 0 0 0 NA NA NA 0 0 0 NA

Anti-CMV IgM NA 0 0 NA 0 NA NA NA 0 0 0 NA

Anti-herpes simplex virus Ab NA 0 NA 0 NA NA NA NA 0 NA NA 0

Anti-HIV Ab 0 0 0 0 0 0 0 0 0 0 0 0

Serum anti-nuclear Ab 0 0 0 0 0 0 0 0 0 0 0 0

Serum anti-smooth muscle Ab 0 0 0 0 0 0 0 0 0 0 0 0

Serum Ig level (normal, 616 g/L) 6.9 4.8 11.5 NA 8.2 NA 6.4 4.9 7.1 7.3 5.6 6.4

Serum alcohol level 0 0 0 0 0 0 0 0 0 0 0 0

Serum acetaminophen level 0 0 0 0 0 0 0 0 0 0 0 0

Serum ceruloplasmin level

(normal, 220610 mg/L)

200 274 230 300 NA 233 217 NA 201 150 NA 157

Serum ammonia level (normal,

1050 mol/L)

NA 25 93 59 92 65 NA 62 74 39 73 53

Arterial lactate level (normal,

0.52.2 mol/L)

2.6 0.9 2.3 NA 0.8 1.1 NA 1.2 1.1 1.8 0.8 0.3

Ab, antibody; CMV, cytomegalovirus; Ig, immunoglobulin; NA, not available; VCA, viral capsid antigen of Epstein-Barr virus.

September 2008 ACUTE LIVER INSUFFICIENCY IN ANOREXIA NERVOSA 848.e3

Gastroenterology Volume 135 issue 3 2008 [doi 10.1053%2Fj.gastro.2008.05.055] Pierre–Emmanuel Rautou; Dominique Cazals–Hatem; Richard More -- Acute Liver Cell Damage in Patients

Documents