mycophenolate mofetil hepatotoxicity associated with

INFOGRAPHICORIGINAL ARTICLE: HEPATOLOGY

Mycophenolate Mofetil Hepatotoxicity Associated With

Mitochondrial Abnormality in Liver Transplant Recipients

and Mice�Mikako Warren, yTania Mitsinikos, yGeorge Yanni, zMika Sasaki, zAtsuo T. Sasaki, and yDan Thomas

ABSTRACT

Copyright © ESPGHAN and NASPGHAN. All rig

Received July 31, 2020; accepted January 5, 2021.From the �Department of Pathology and Laboratory Medicine, the yDivision

Gastroenterology, and Nutrition, Children’s Hospital Los Angeles, Uni-versity of Southern California Keck School of Medicine, Los Angeles,CA, and the zDivision of Hematology and Oncology, Department ofInternal Medicine, University of Cincinnati College of Medicine, Cin-cinnati, OH.

Address correspondence and reprint requests to Mikako Warren, MD, Mail-stop #43, Children’s Hospital Los Angeles Department of Pathology andLaboratory Medicine, 4650 Sunset Blvd., Los Angeles, CA 90027(e-mail: [email protected]).

Supplemental digital conteappear in the printed teHTML text of this arti

The mouse study was supp(A.T.S.).

The authors report no consible for the content a

Copyright # 2021 by EuHepatology, and NutrGastroenterology, Hep

DOI: 10.1097/MPG.00000

JPGN � Volume 73, Number 4, October 2021

Objectives: Mycophenolate mofetil (MMF) is a widely used immunosup-

pressive agent. MMF hepatotoxicity has been reported in non-transplant and

renal transplant patients with minimal histologic description. This is the first

study describing detailed histology and ultrastructure of MMF hepatotoxicity.

Methods: Four liver-transplant recipients (Cases 1–4) were suspected to

have MMF hepatotoxicity. Cases 1–3 (two females and one male; 4–

17 years) had multiple biopsies for liver function test (LFT) abnormalities.

Case 4 (female; 16 years) had a surveillance biopsy. Electron-microscopic

examination (EM) was requested on Cases 1–3 for unexplained, persistent

LFT elevation and histologic abnormalities despite therapy and Case 4 for

unexplained histologic abnormalities despite a stable clinical course. To

confirm the pathologic changes in the human allografts, livers from MMF-

treated and untreated mice were also reviewed.

Results: While the allograft biopsies showed nonspecific histologic changes,

EM revealed unequivocal mitochondrial abnormalities similar to those seen in

primary and secondary mitochondrial disorders. In Cases 1 and 2, LFTs improved

after stopping and reducing MMF, respectively. In Case 3, pre- and post-MMF

treatment biopsies were performed and only the post-MMF biopsy demonstrated

mitochondrial abnormalities. Mitochondrial abnormality in Case 4 was subclinical.

The mouse study confirmed that MMF caused various stress changes in the

mitochondria; number of mitochondria/cell (mean� standard deviation; untreated

group: 58.25� 8.426; MMF-treated group: 76.37� 18.66), number of lipid

droplets/cell (untreated: 0.9691� 1.150; MMF-treated: 3.649� 4.143) and

sizes of mitochondria (mm, untreated: 0.8550� 0.3409; MMF-treated:

0.9598� 0.5312) were significantly increased in hepatocytes in the MMF-

treated mice compared with the untreated mice (P< 0.0001).

Conclusions: Although MMF is safe for the majority of patients, MMF can

cause mitochondrial stress, which may trigger more severe mitochondrial

abnormalities in a small subset. MMF hepatotoxicity should be considered

for MMF-treated patients with unexplained, persistent LFT abnormalities

and nonspecific histologic findings. EM should be requested for these cases.

Key Words: drug-induced liver injury, histopathology, liver transplant,

mitochondria, mycophenolate mofetil, ultrastructure (electron microscopy)

An infographic is available for this article at: http://links.lww.com/

MPG/C361.

(JPGN 2021;73: 463–470)

What Is Known

� Recognition of drug-induced liver injury (DILI) can bechallenging because DILI displays diverse, often non-specific laboratory and histopathologic changes.

� Rare cases with mycophenolate mofetil (MMF) hepa-totoxicity have been reported in non-transplant andrenal transplant patients.

What Is New

� This is the first study reporting detailed histopathol-ogy and ultrastructure of MMF hepatotoxicity.

� Despite nonspecific histologic abnormalities, electron-microscopic examination (EM) revealed unequivocalmitochondrial abnormalities similar to those seen inprimary and secondary mitochondrial disorders.

� MMF hepatotoxicity should be considered for MMF-treated patients with unexplained, persistent liverenzyme abnormalities and nonspecific histology.

� EM should be requested for these cases.

rug-induced liver injury (DILI) represents the leading cause
D of acute liver failure in the United States (1–3) with anestimated incidence of 1 in 10,000 and 1 to 100,000 patients (4).Recognition of DILI can be challenging clinically (3) and histolog-ically (5) because DILI can be present with highly diverse labora-tory and histologic changes. The histologic patterns of DILI includehepatitis, cholestasis, granulomatous inflammation, macro- and/ormicro-vesicular steatosis with or without steatohepatitis, hepato-cellular necrosis ranging from single cell drop-out to broad necrosis,sinusoidal obstruction/veno-occlusive disease, and any combina-tion of these injury patterns. Additionally, DILI frequently displaysnonspecific (unclassifiable) pathologic changes (5). These diverse,
hts reserved.

nt is available for this article. Direct URL citationsxt, and links to the digital files are provided in thecle on the journal’s Web site (www.jpgn.org).orted in part by R21NS100077 and R01NS089815

flicts of interest. The authors alone are respon-nd writing of this article.ropean Society for Pediatric Gastroenterology,

ition and North American Society for Pediatricatology, and Nutrition00000003171

463

http://links.lww.com/MPG/C361

http://links.lww.com/MPG/C361

mailto:[email protected]

http://www.jpgn.org/

http://dx.doi.org/10.1097/MPG.0000000000003171

Warren et al JPGN � Volume 73, Number 4, October 2021

often nonspecific histologic patterns make it difficult to establishpractical diagnostic criteria for DILI.

To treat or prevent acute graft rejection in solid organ recip-ients, azathioprine, 6-mercaptopurine (6-MP), cyclophosphamide,and calcineurin inhibitors, such as cyclosporine and tacrolimus, havebeen used in combination with high-dose corticosteroids. Mycophe-nolate mofetil (MMF) and sirolimus have emerged recently asadditional immunosuppressive agents in managing solid organ trans-plants (6,7). These immunosuppressive agents are often used incombination with other medications and they may be also used totreat patients with pre-existing liver disease, such as autoimmunehepatitis. DILI can be associated with any of these immunosuppres-sive agents. Among these agents, azathioprine- and 6-MP-relatedhepatotoxicity are well-known and account for 1–2% of DILI. Theytypically show cholestatic, hepatocellular and mixed injury patterns(8). Otherwise, the injury is thought to be generally mild and only asmall number of cases have been reported (9). The number of casesmay be undercounted because it is often difficult to diagnose DILIassociated with immunosuppressive agents and determining a par-ticular causative agent is further challenging.

MMF is an immunosuppressive agent commonly used as anadjunctive agent to prevent and/or treat acute cellular rejection(ACR) in solid organ transplant recipients and as a therapeutic agentin non-transplanted patients with various diseases with immunedysregulation. Major adverse effects of MMF include bone marrowsuppression, gastrointestinal, neurological symptoms and teratoge-nicity. Rare sporadic cases with MMF-related hepatotoxicity havebeen reported in the native livers of non-transplant patients andrenal transplant recipients; however, these reports described no oronly minimal histopathology and ultrastructural analysis had notbeen performed (10–16).

Children’s Hospital Los Angeles (CHLA) performs 110–120transplant liver biopsies per year. Liver biopsy is routinely per-formed on transplant recipients with elevated liver function tests(LFTs) and as a surveillance biopsy for patients with normal LFTsin accordance with internal protocols. The predominant indicationfor allograft biopsy is to rule out ACR as it is the most frequentclinical concern and requires prompt treatment. ACR is diagnosedaccording to the standardized histologic criteria defined by theBanff Working Group (Banff criteria) (17); however, other possibleetiologies of liver dysfunction, which are often evident only asnonspecific histologic findings on liver biopsies, are common yetoften receive little consideration in pathology reports.

We herein present four liver transplant patients treated withMMF, who were clinically suspected to have MMF hepatotoxicity.This is the first study that shows the detailed histopathology andultrastructure of MMF hepatotoxicity in human liver allografts andmouse livers treated with MMF.

METHODS

Patients’ Transplant Liver BiopsiesThe study has been approved by our internal Institutional

Review Board (IRB) at CHLA. Liver biopsies were performed byinterventional radiology at CHLA. Biopsies were immediately fixedin 10% formalin (Medical Chemical Cooperation, Torrance, CA,USA) for light-microscopic examination (LM) and 2.5% bufferedglutaraldehyde (BCC Biochemical, Mount Vernon, WA, USA) forelectron-microscopic examination (EM).

All staining and histologic examination were performed atthe Clinical Laboratory Improvement Amendments (CLIA)-certi-fied laboratory at CHLA. For LM, we performed hematoxylin &eosin (H&E) and special staining (Periodic acid-Schiff [PAS], PASwith diastasis [PASD], reticulin, iron and trichrome) per biopsy

Copyright © ESPGHAN and NA

464

according to our operating procedure. Ultrastructural analysis wasperformed as previously described (18,19).

ACR was diagnosed and scored using the rejection activityindex (RAI), which grades: portal inflammation (score 1–3), bileduct damage (score 1–3), and venous endothelial inflammation(score 1–3). Each score was added and the degree of ACR wasscored as follows: RAI¼ 0–9; <3: borderline/indeterminate ACR,3–4: mild ACR, 5–7: moderate ACR and >7: severe ACR (17).

Nonspecific (unclassifiable) hepatocellular injury is oftenreferred to ‘‘reactive changes’’. Histologic features of ‘‘reactivechanges’’ include a combination of enlarged hepatocytes withhydropic changes (expanded, pale to clear cytoplasm) and coarseeosinophilic granules (eg, mega-mitochondria), nuclei with aniso-nucleosis and bi-/multi-nucleated forms, cholestasis, steatosis and/or necrosis. Necrosis can range from single cell necrosis (acidophilbodies) to rarely broad necrosis with collapsed lobules.

Mouse Liver Samples

The study has been approved by the Institutional AnimalCare and Use Committee (IACUC) at the University of Cincinnatiand CHLA. All of the animal experiments in the present study werecompliant with relevant ethical regulations regarding animalresearch. Mice were 7–8 months old female, which were house-bred from C57BL/6 mice (the Jackson Laboratory, Bar Harbor, ME,USA). For the MMF-treated group, mice (7–8 months of age) weregiven MMF by oral gavage. MMF was dissolved in 0.5% methyl-cellulose/0.1% Tween 80 solution and 120mg/kg/day was adminis-trated orally. For untreated group, 0.5% methylcellulose/0.1%Tween 80 solution was used for vehicle treatment. Two of the fiveMMF-treated mice died on days 12 and 13 of the treatment and wereexcluded from the experiment. The rest were sacrificed on day 14and necropsy was performed. Livers were harvested and immedi-ately fixed for LM and EM as described above.

H&E and trichrome staining were performed on each liversection from the three MMF-treated mice and the five untreatedmice at CHLA. Ultrastructural analysis was performed on themouse liver tissue from the three MMF-treated mice and threemice randomly selected from the untreated group, according to themethod described above. EM images were captured digitally at thesame magnification (8000�). Thirty hepatocytes per mouse liverwere randomly selected and numbers of the mitochondria and lipiddroplets were counted per hepatocyte using digital images. Thegreatest dimensions of randomly selected 50 mitochondria perhepatocyte from each mouse were measured using image analysissoftware, Cellsens (Olympus, Tokyo, Japan).

Data Analysis (Mouse Livers)

The number of mitochondria and lipid droplets per hepato-cyte (20 cells per mouse, total 60 cells per group) and size of thehepatocellular mitochondria (50 mitochondria per hepatocyte; total60 cells/3000 mitochondria per group) from the MMF treated anduntreated mice were compared by repeated measures mixed modelanalysis, with mice as a random effect and group as fixed, at a 0.005significance level, using Prism8 software (GraphPad Software, SanDiego, CA, USA).

RESULTS

Case ReportsThe study included three female and one male liver trans-

plant recipients. Clinical demographics of the patients are summa-rized in Table 1. Three MMF-treated liver transplant recipients

SPGHAN. All rights reserved.

www.jpgn.org

Copyright © ESPGHAN and NASPGHAN. All rights reserved.

TABLE

1.

Clin

icald

em

og

rap

hic

san

db

iop

syre

sults

of

Case

s1

–4

Bio

psy

EM

Pat

ho

log

yre

po

rt

Cas

en

o.

Ag

eS

exT

ran

spla

nt

Tim

eaf

ter

Tx

Rea

son

for

Tx

Bio

psy

no

.

Tim

ing

of

bio

psy

fro

mth

ein

itia

lb

iop

syo

fth

eev

ent

EM

per

form

edE

Mfe

atu

res

RA

Ire

port

edO

ther

fin

din

gs

rep

ort

edA

dd

itio

nal

test

ing

To

tal

Po

rtal

Du

ctE

nd

oth

elia

lF

ibro

sis

Lo

bu

lar

infl

amm

atio

nA

dd

itio

nal

feat

ure

s

11

2y

FC

adav

eric

wh

ole

liv

er

2m

oB

A,

end

-sta

ge

liver

dis

ease

,p

ort

alh

yp

erte

nsi

on

1D

ay1

No

N/A

31

11

No

ne

Rar

em

ild

N/A

C4

d:

neg

2D

ay8

No

N/A

21

10

No

ne

Fo

cal

neu

tro

ph

ilic

infi

ltra

tes

N/A

Vir

alst

ud

yan

do

ne

HS

V,

EB

ER

.):

CM

V,

EB

ER

,H

SV

1/2

:n

eg):

neg

3D

ay2

0Y

esM

ito

cho

nd

rial

ple

om

orp

his

m,

Cry

stal

loid

incl

usio

ns

21

10

Fo

cal

mil

dp

eris

inu

soid

alan

dp

erip

ort

al

Min

imal

scat

tere

dD

iffu

sere

acti

ve

chan

ges

,m

ild

sin

uso

idal

dil

atat

ion

C4

d:

neg

24

yF

Cad

aver

icw

ho

leli

ver

2m

oH

epat

ob

last

om

a1

Day

1N

oN

/A5

21

2n

on

eO

ccas

ion

alm

ild

N/A

CM

V,

aden

ovi

rus,

HS

V1

/2E

BE

R:

neg

2D

ay1

2N

oN

/A3

11

1n

on

eR

are

mil

dN

/AN

/A

3D

ay2

5N

oN

/A3

11

1M

ild

per

ipo

rtal

(sta

ge

1)

Sca

tter

edsi

ng

lece

lln

ecro

sis

N/A

CM

V,

aden

ovi

rus,

HS

V1

/2E

BE

R:

neg

4D

ay4

2Y

esM

ito

cho

nd

rial

ple

om

orp

his

mC

ryst

allo

idin

clus

ion

s

31

11

No

ne

No

ne

Rea

ctiv

ech

ang

esw

ith

lip

ofu

scin

,m

ild

sin

uso

idal

dil

atat

ion

CM

VE

BE

R,

C4

d:

neg

31

7y

MC

adav

eric

wh

ole

liv

er

3m

oC

DG

1D

ay1

No

N/A

31

11

Mil

dp

ort

alfi

bro

sis

(sta

ge

1)

Rar

em

ild

N/A

CM

V,

aden

ovi

rus,

HS

V1

/2,

EB

ER

,C

4d

:n

eg2

Day

6Y

esN

orm

al3

–4

1–

21

1P

eris

inu

soid

alfi

bro

sis

(sta

ge

0)

No

ne

Mil

dzo

ne3

dil

atat

ion

C4

d:

neg

3D

ay2

1Y

esM

ito

cho

nd

rial

ple

om

orp

his

m,

Cry

stal

loid

incl

usio

ns

31

11

Per

isin

uso

idal

fib

rosi

s(s

tag

e0

–1

)N

on

eD

iffu

sere

acti

ve

chan

ges

C4

d:

neg

4D

ay3

7N

oN

/A3

11

1N

on

eN

on

eR

eact

ive

chan

ges

and

scat

tere

dac

ido

ph

ilb

od

ies

CM

V,

aden

ovi

rus,

HS

V1

/2,

EB

ER

,C

4d

:n

eg4

16

yF

Cad

aver

icsp

lit

(lef

tla

tera

lse

gm

ent)

13

yB

A,

end

-sta

ge

liv

erd

isea

se1

Day

1N

oN

/A2

10

1P

ort

al,

per

isin

uso

idal

wit

hfo

cal

bri

dg

ing

(Sta

ge

2–

3)

No

ne

N/A

CM

V,

aden

ovi

rus,

HS

V1

/2,

EB

ER

:n

eg

2D

ay3

42

No

N/A

31

11

Po

rtal

,p

eris

inu

soid

alw

ith

foca

lb

rid

gin

g(S

tag

e2

–3

)

Fo

cal,

mil

dF

oca

lh

epat

ocy

ted

rop

ou

tN

on

e

3D

ay5

09

Yes

Mit

och

on

dri

alp

leo

mo

rph

ism

,C

ryst

allo

idin

clus

ion

s

21

01

Po

rtal

,p

eris

inu

soid

alw

ith

foca

lb

rid

gin

g(S

tag

e2

–3

)

No

ne

Mic

rov

esic

ula

rst

eato

sis

(30

%)

EB

ER

:n

eg

No

teth

at‘‘

add

itio

nal

feat

ure

s’’

inth

ep

ath

olo

gy

repo

rts

wer

en

ots

tan

dard

ized

amon

gp

ath

olo

gis

tsan

dd

idn

ota

ffec

tth

efi

nal

dia

gno

ses.

BA¼

bil

iary

atre

sia;

CD

G¼

con

gen

ital

dis

ord

ero

fg

lyco

syla

tio

n;

CM

V¼

cyto

meg

alo

viru

s;D

uct¼

bil

ed

uct

dam

age

sco

re;

EB

ER¼

Ep

stei

n-B

arr

enco

din

gre

gio

nin

situ

hy

bri

dis

atio

n;

En

do

thel

ial¼

ven

ous

end

oth

elia

lin

flam

mat

ion

sco

re;

HS

V¼

her

pes

sim

ple

xv

iru

s;K

asai¼

Kas

aip

roce

du

re;

mo¼

mo

nth

s;N

/A¼

no

tap

pli

cab

le;

neg¼

neg

ativ

e;P

ort

al¼

Po

rtal

infl

amm

atio

nsc

ore

;R

AI¼

reje

ctio

nac

tivit

yin

dex

;T

x¼

tran

spla

nt;

y¼

yea

rs.

JPGN � Volume 73, Number 4, October 2021 Mycophenolate Mofetil Hepatotoxicity

www.jpgn.org 465


(Cases 1–3; two females and one male; 4–17 years of age) hadmultiple allograft biopsies for unexplained, persistently elevatedLFTs. A biopsy was performed on an MMF-treated liver transplantrecipient (Case 4; a female; 16 years of age) for surveillancepurpose as per our internal CHLA protocol. All four patients hada history of good medication compliance.

EM was requested on Cases 1–3 because these patients contin-ued to have LFT abnormalities and nonspecific histologic abnormali-ties despite ongoing medical therapy, and on Case 4 as unexplainednonspecific histologic findings were evident in her surveillance biopsydespite having normal LFTs and a stable clinical course.

Case 1 was a 12-year-old female with a history of biliary atresia,status post-Kasai portoenterostomy with progression to end-stage liverdisease. Her post-transplant course was complicated by elevated LFTs,for which multiple liver biopsies were obtained. Given the findings of


FIGURE 1. Clinical courses and medication history of Cases 1–4. Cases 1surveillancebiopsy as per our internal protocol. EM was requestedon Cases 1–

despite therapy and Case 4 for unexplained nonspecific histology despite n

reduction of MMF improved liver function. After 3 and 2 days after cessation

3 months after cessation and reduction. Case 3 underwent a pair of pre- andmitochondrialabnormalities.Case4 showedasubclinicalmitochondrial abno

to be the cause of the LFTabnormalities. After MMF hepatotoxicity became m

LFTs. Day 0 ¼ initial event or biopsy; EM ¼ electron-microscopic examinati

466

ACR as well as positive donor-specific antibodies and C4d staining, shewas aggressively treated with multiple courses of high dose intravenouscorticosteroids and increasing doses of tacrolimus. Because LFTs werenot normalized despite therapy, a third agent, MMF, was started as anadjunctive agent. However, her LFTs increased after MMF was started(Fig. 1). Given the mitochondrial abnormalities identified by EM,MMF was discontinued. Approximately 3 days after cessation ofMMF, LFTs started to improve (Fig. 1) and normalized at 2 monthsafter cessation (not shown in Fig. 1).

Case 2 was a 4-year-old female with a history of unresectablehepatoblastoma. Her post-transplant course was complicated byhemolytic anemia. Since tacrolimus-induced hemolysis was sus-pected, the medication was changed to cyclosporine, which improvedher hemolysis. Two months following her transplant she developedan acute rise in her LFTs due to ACR. She was aggressively managed


–3 had multiple biopsies for persistently elevated LFTs. Case 4 had a3 for unexplained, persistent LFTabnormalities and nonspecific histology

ormal LFTs and a stable clinical course. In Cases 1 and 2, withdrawal or

and reduction of MMF, LFTs started to improve and normalized at 2 and

post-MMF treatment biopsies and only the post-MMF biopsy showedrmality. ForCase3and4, at the timeof theworkup,MMFwasnot thought

ore certain, MMF was discontinued. The patients currently have normal

on; LFT ¼ liver function test; MMF ¼ mycophenolate mofetil.

www.jpgn.org


with high-dose corticosteroids and tacrolimus, which did not causehemolysis at that time. Her LFTs improved but did not normalize;therefore, MMF was added. However, her LFT abnormalities per-sisted after MMF was started (Fig. 1). Given the mitochondrialabnormalities identified by EM, MMF was reduced. Two days afterreduction of MMF, LFTs started to improve (Fig. 1) and normalizedat 3 months after reduction (not shown in Fig. 1).

Case 3 was a 17-year-old male with a history of congenitaldisorder of glycosylation Type Ib. His post-transplant course wasinitially uneventful, however he developed increased LFTs and giveninitial concerns for tacrolimus toxicity, his tacrolimus dosing wasdecreased and MMF was started; however, mild LFT abnormalitiespersisted after MMF was started (Fig. 1). At the initial diagnosticworkup MMF was not thought to the cause of hepatotoxicity. AfterMMF hepatotoxicity became more certain, it was discontinued andthe graft functions normalized (not shown in Fig. 1).

Lastly, Case 4 was a 16-year-old female with a history ofbiliary atresia status-post Kasai portoenterostomy, who presentedwith acute-on-chronic liver failure leading to end-stage liver dis-ease. Her post-transplant course was notable for EBV viremia, HSVstomatitis and an episode of ACR, all of which resolved before thecurrent presentation. Because her ACR was refractory to standardtherapy, MMF was started. Since then she had been treated withMMF for approximately one year. She underwent a protocolsurveillance liver biopsy as per the institutional protocol (Fig. 1).At the time of workup, MMF was not thought to be a cause ofhistologic abnormalities. After MMF hepatotoxicity became morecertain, it was discontinued. The patient currently has normal LFTs.

Figure 1 shows patients’ biopsy findings along with thelaboratory results and given medications. Multiple biopsies wereperformed for elevated LFTs, and the diagnoses established byboard-certified pediatric pathologists were generally consistentwith low-grade ACR (RAI ranging 2–4), except the initial biopsyof Case 1 with RAI of 5. Initial pulse corticosteroids resulted inlower LFTs; however, LFTs did not normalize despite aggressiveimmunosuppressive therapy.

In Cases 1 and 2, LFTs improved after stopping and reducingMMF, respectively; therefore, MMF hepatotoxicity was suspected.For Case 3, a pair of pre- and post-MMF treatment biopsies wereperformed and only the post-MMF biopsy demonstrated mitochon-drial abnormality. Case 4 was clinically stable but her surveillancebiopsy showed unequivocal mitochondrial abnormality.

Histologic and Ultrastructural Features of thePatients’ Liver Biopsies

The histologic findings of each case are summarized inTable 1. Biopsies demonstrated features of mild ACR with patchyportal inflammation, some of which were accompanied by mildductal damage and/or mild subendothelial lymphocyte infiltrates. Inaddition, they showed mild, nonspecific hepatocellular injury(‘‘reactive changes’’). These ‘‘reactive changes’’ included mildlyenlarged hepatocytes with granular cytoplasm, anisonucleosis, andfocal areas with predominantly microvesicular steatosis (Fig. 2Aand B). For all of the four cases, EM revealed similar ultrastructuralfeatures including prominent mitochondrial pleomorphism (vari-ability in size and shape) and crystalloid inclusions (Fig. 2C and D),except for the pre-MMF biopsy in Case 3.

Histologic and Ultrastructural Features ofMouse Livers

Histology of the livers from the MMF-treated and untreatedmice showed no significant differences. MMF-treated mouse livers


www.jpgn.org

showed only minimal ‘‘reactive changes’’ with rare anionucleosis,granular cytoplasm and no fibrosis (Fig. 2E and F). Analysis usingEM images revealed various mitochondrial ‘‘stress changes’’ in thelivers from the MMF-treated mice. Number of mitochondria per cell(mean � standard deviation; untreated group: 58.25� 8.426;MMF-treated group: 76.37� 18.66), number of lipid droplets percell (untreated group: 0.9691� 1.150; MMF-treated group:3.649� 4.143) and sizes of mitochondria (mm, untreated group:0.8550� 0.3409; MMF-treated group: 0.9598� 0.5312) were sig-nificantly increased in hepatocytes in the MMF-treated groupcompared with those in the untreated group (P< 0.0001; Fig. 3).

DISCUSSIONMycophenolic acid and its pro-drug form (MMF), are potent

inhibitors of inosine monophosphate dehydrogenase (IMPDH), arate-limiting enzyme of the de novo GTP biosynthesis. Lympho-cytes depend on IMPDH activity for their proliferation andimmune-related activity and therefore, are sensitive to IMPDHinhibitors. MMF has been used as an adjunct immunosuppressiveagent for solid organ transplanted patients (20,21). Compared withthe other previously-developed immunosuppressants, MMF has anequivalent or superior immunosuppressive ability and feweradverse effects that enable long-term use (22,23). MMF has beenwidely used not only for solid organ transplant recipients but alsofor large numbers of patients with various immune-dysregulationdiseases and for treating and/or preventing graft-versus-host diseasein bone marrow transplant recipients. MMF is a critically importantdrug for human diseases associated with undesirable immunity (24).

MMF hepatotoxicity is thought to be rare (9,16). Rare caseshave been reported in non-transplant patients treated with MMF,including patients with focal segmental glomerulosclerosis (11),atopic dermatitis (12), ANCA-positive vasculitis (13), and scleritisof the eyes (14). Balal et al (15) reported 79 renal transplantrecipients treated with MMF. Of these 11 patients (13.9%) hadrelatively mildly elevated LFTs, which were normalized afterreduction or withdrawal of MMF. The mechanism of the hepato-toxicity is unclear and these reports described no or minimalhistologic findings. EM had not been performed for any of the cases.

Mitochondria are cellular powerhouses and regenerate ATPfrom ADP through the oxidative phosphorylation system. Mito-chondria also provide intermediate metabolites, which are criticalfor cellular functions and proliferation (25) and regulate cell func-tions via reactive oxygen species (ROS) generated during theoxidative phosphorylation process (26). Mitochondria also havean important role in cell death, which is triggered by mitochondrialmembrane disruption and then apoptosis and/or necrosis of the cell(27). Recent evidence suggests that hepatotoxic drugs can causemitochondrial dysfunction in the liver through diverse mechanisms,such as direct inhibition of mitochondrial respiration and beta-oxidation and damage to mitochondrial DNA, mitochondrial tran-scripts, and mitochondrial protein synthesis (27–30).

Our results (Cases 1 and 2) and the previously reported cases,in which withdrawal or reduction of MMF improved liver function,are supportive evidence that MMF hepatotoxicity occurring in asubset of MMF-treated patients, either with native or transplantedliver. Ultrastructural findings in Case 3 with a pair of pre- and post-MMF treatment biopsies reinforces the link between MMF treat-ment and hepatocellular mitochondrial abnormalities. Case 4 dem-onstrated that mitochondrial abnormality can be subclinical. Themouse study did not show as dramatic morphological changes asthose seen in the patients’ allografts, but it confirmed various stresschanges in mitochondria associated with MMF treatment.

MMF is safe to use majority of patients; however, it may stressmitochondria and may trigger more severe mitochondrial abnormality


467

Copyright © ESPGHAN and NASPGHAN. All rights reserved.

FIGURE 2. Histologic and ultrastructural features of the transplant biopsies (A and B) H&E (A: 100�, B: 400�); (C and D) electron micrograph (C:

8000�, D: 18,000�) and the untreated mouse livers (E) and livers treated with MMF (F) and untreated mice (F) (E1 and F1: H&E [400�]; E2 and F2:

electron micrograph [8000�]). In addition to the features categorized into mild ACR, the histology showed mild, nonspecific (unclassified)hepatocellular abnormalities (‘‘reactive changes’’, A and B). EM revealed prominent mitochondrial pleomorphism (variability in size and shape) and

crystalloid inclusions (C and D), except for the biopsy taken before stating MMF in Case 3. Red arrowheads indicate extremely large ones among the

pleomorphic mitochondria. Histologically, the livers from the untreated (E1) and MMF-treated mice (F1) showed no recognizable differences. They

had only minimal ‘‘reactive changes’’ with mild anionucleosis and granular cytoplasm. Ultrastructurally, hepatocellular mitochondria of the MMF-treated mice (F2) showed more pleomorphism (variability in size and shape) and lipid droplets compared with the livers from the untreated mice (E2).

Red and blue arrowheads indicate large pleomorphic mitochondria and lipid droplets, respectively. ACR ¼ acute cellular rejection; EM ¼ electron-

microscopic examination; H&E ¼ hematoxylin & eosin; MMF ¼ mycophenolate mofetil.


468 www.jpgn.org

A

B C

Untreated 1 Unteated 2 Untreated 3 MMF Treated 1 MMF Treated 2 MMF Treated 3

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

S iz e

of M

itoch

o ndr

i a (µ

m)

Samples: 3000 mitochondria Mean: 0.8550 SD: 0.3409

Samples: 3000 mitochondria Mean: 0.9598 SD: 0.5312

* p < 0.0001

FIGURE 3. Image analysis using EM revealed mitochondrial ‘‘stress changes’’; numbers of mitochondria and lipids and degree of mitochondrialpleomorphism (variability in size) were significantly increased in hepatocytes from the MMF-treated group compared with the untreated group

(P<0.0001). EM ¼ electron-microscopic examination; MMF ¼ mycophenolate mofetil.


in a small subset. One should note that MMF’s pharmacokinetics andpharmacodynamics vary among individuals (31,32).

Diagnosing mitochondrial disorders is challenging by rou-tine LM alone, even for cases of genetically proven primarymitochondrial disorders with severe clinical manifestations,because the histologic features are diverse, ranging from normal,reactive changes, hepatitis, to various degrees of cellular necrosis,and are often nonspecific (18). EM plays an important role inidentifying morphological mitochondrial abnormalities at the ultra-structural level. Our EM results strongly suggest a correlationbetween MMF hepatotoxicity and mitochondrial abnormalities inhuman and mouse livers treated with MMF. The next important stepwould be to investigate a larger number of patients and to look forrisk factor(s), such as genetic alteration(s), that may make patientsmore susceptible to MMF hepatotoxicity. The role of IMPDH inmitochondrial morphology and function also needs to be studied.

In summary, this is the first study describing detailed histo-logic and ultrastructural features of MMF hepatotoxicity. Despitemild nonspecific histologic abnormalities, the allograft biopsiesrevealed unequivocal mitochondrial abnormalities similar to thoseseen in primary and secondary mitochondrial disorders. The mouse


www.jpgn.org

study confirmed that MMF caused various stress changes in themitochondria. Although MMF is safe for the majority of patients,MMF may stress mitochondria and the stress may trigger moresevere mitochondrial abnormality in a small subset. MMF hepato-toxicity should be considered for MMF-treated patients, either withnative or transplanted livers, who have unexplained, persistent LFTabnormalities and nonspecific histologic changes. EM can play acritical role in diagnosing these cases.

Acknowledgments: We would like to thank Alexander Navarrofor processing the ultrastructural materials. The mouse study wassupported in part by R21NS100077 and R01NS089815 (A.T.S.).

REFERENCES1. Navarro VJ, Senior JR. Drug-related hepatotoxicity. N Engl J Med

2006;354:731–9.2. Ghabril M, Chalasani N, Bjornsson E. Drug-induced liver injury: a

clinical update. Curr Opin Gastroenterol 2010;26:222–6.3. Fontana RJ, Hayashi PH, Gu J, et al., DILIN Network. Idiosyncratic drug-

induced liver injury is associated with substantial morbidity and mortalitywithin 6 months from onset. Gastroenterology 2014;147:96–108.


469


4. Ostapowicz G, Fontana RJ, Schiødt FV, et al., Schiødt FV, et al., AcuteLiver Failure Study Group. Results of a prospective study of acute liverfailure at 17 tertiary care centers in the United States. Ann Intern Med2002;137:947–54.

5. Fisher K, Vuppalanchi R, Saxena R. Drug-induced liver injury. ArchPathol Lab Med 2015;139:876–87.

6. Mukherjee S, Mukherjee U. A comprehensive review of immunosup-pression used for liver transplantation. J Transplant 2009;2009:701464.doi: 10.1155/2009/701464.

7. Moini M, Schilsky ML, Tichy EM. Review on immunosuppression inliver transplantation. World J Hepatol 2015;7:1355–68.

8. Bjornsson ES, Gu J, Kleiner DE, et al., DILIN Investigators. Azathiopr-ine and 6-mercaptopurine-induced liver injury: clinical features andoutcomes. J Clin Gastroenterol 2017;51:63–9.

9. LiverTox. Clinical and research information on drug-induced liver injuryBethesda (MD): National Institute of Diabetes and Digestive and KidneyDiseases. https://www.ncbi.nlm.nih.gov/books/NBK547852/. 2012. [Ac-cessed 31 July 2020].

10. Chalasani N, Fontana RJ, Bonkovsky HL, et al. Causes, clinicalfeatures, and outcomes from a prospective study of drug-induced liverinjury in the United States. Gastroenterology 2008;135:1924–34.

11. Loupy A, Anglicheau D, Mamzer-Bruneel MF, et al. Mycophenolatesodium-induced hepatotoxicity: first report. Transplantation 2006;82:581.

12. Hantash B, Fiorentino D. Liver enzyme abnormalities in patients withatopic dermatitis treated with mycophenolate mofetil. Arch Dermatol2006;142:109–10.

13. Dourakis SP, Boki K, Soultati A, et al. Acute hepatitis followingmycophenolate mofetil administration for ANCA-positive vasculitis.Scand J Rheumatol 2007;36:237–9.

14. Sen HN, Suhler EB, Al-Khatib SQ, et al. Mycophenolate mofetil for thetreatment of scleritis. Ophthalmology 2003;110:1750–5.

15. Balal M, Demir E, Paydas S, et al. Uncommon side effect of MMF inrenal transplant recipients. Ren Fail 2005;27:591–4.

16. Hernandez N, Bessone F, Sanchez A, et al. Profile of idiosyncratic druginduced liver injury in Latin America: an analysis of published reports.Ann Hepatol 2014;13:231–9.

17. No author listed, Banff schema for grading liver allograft rejection: aninternational consensus document. Hepatology 1997;25:658–63.


470

18. Warren M, Shimada H. Cytologic and ultrastructural findings ofbronchoalveolar lavage in patients with chronic granulomatous disease.Pediatr Dev Pathol 2018;21:347–54.

19. Warren M, Shimura M, Wartchow EP, Yano S. Use of electron micro-scopy when screening liver biopsies from neonates and infants: experi-ence from a single tertiary children’s hospital (1991-2017). UltrastructPathol 2020;44:32–41.

20. Allison AC, Kowalski WJ, Muller CD, et al. Mechanisms of action ofmycophenolic acid. Ann NY Acad Sci 1993;696:63–87.

21. Bentley R. Mycophenolic acid: a one hundred year odyssey fromantibiotic to immunosuppressant. Chem Rev 2000;100:3801–26.

22. Molnar AO, Fergusson D, Tsampalieros AK, et al. Generic immuno-suppression in solid organ transplantation: systematic review and meta-analysis. BMJ 2015;350:h3163. doi: 10.1136/bmj.h3163.

23. El Hajj S, Kim M, Phillips K, et al. Generic immunosuppression intransplantation: current evidence and controversial issues. Expert RevClin Immunol 2015;11:659–72.

24. Downing HJ, Pirmohamed M, Beresford MW, et al. Paediatric use ofmycophenolate mofetil. Br J Clin Pharmacol 2013;75:45–59.

25. Owen OE, Kalhan SC, Hanson RW. The key role of anaplerosis andcataplerosis for citric acid cycle function. J Biol Chem 2002;277:30409–12.

26. Sena LA, Chandel NS. Physiological roles of mitochondrial reactiveoxygen species. Mol Cell 2012;48:158–67.

27. Pessayre D, Fromenty B, Berson A, et al. Central role of mitochondria indrug-induced liver injury. Drug Metab Rev 2012;44:34–87.

28. Labbe G, Pessayre D, Fromenty B. Drug-induced liver injury throughmitochondrial dysfunction: mechanisms and detection during preclini-cal safety studies. Fundam Clin Pharmacol 2008;22:335–53.

29. Begriche K, Massart J, Robin MA, et al. Drug-induced toxicity onmitochondria and lipid metabolism: mechanistic diversity and deleter-ious consequences for the liver. J Hepatol 2011;54:773–94.

30. Ramachandran A, Visschers RGJ, Duan L, et al. Mitochondrial dys-function as a mechanism of drug-induced hepatotoxicity: currentunderstanding and future perspectives. J Clin Transl Res 2018;4:75–100.

31. Kiang TKL, Ensom MHH. Population pharmacokinetics of mycophe-nolic acid: an update. Clin Pharmacokinet 2018;57:547–58.

32. Staatz CE, Tett SE. Pharmacology and toxicology of mycophenolate inorgan transplant recipients: an update. Arch Toxicol 2014;88:1351–89.


www.jpgn.org

https://www.ncbi.nlm.nih.gov/books/NBK547852/

mycophenolate mofetil hepatotoxicity associated with

Documents