Research Article

Experimental and Translational Hepatology

Therapeutic HNF4A mRNA attenuates liver fibrosis in a

preclinical model
Taihua Yang1,2,14, Marion Poenisch3, Rajendra Khanal1,2, Qingluan Hu1,4, Zhen Dai1,2,
Ruomeng Li1,2, Guangqi Song1,5, Qinggong Yuan1,4, Qunyan Yao5, Xizhong Shen5,
Richard Taubert1, Bastian Engel1, Elmar Jaeckel1, Arndt Vogel1, Christine S. Falk6,
Axel Schambach7,8, Daniela Gerovska9, Marcos J. Araúzo-Bravo9,10, Florian W.R. Vondran11,12,
Tobias Cantz1, Nigel Horscroft3,13, Asha Balakrishnan1,4, Frédéric Chevessier3, Michael Ott1,4,
Amar Deep Sharma1,2,*
1
Department of Gastroenterology, Hepatology and Endocrinology, Hannover Medical School, Hannover, Germany; 2Research Group Liver
Regeneration, REBIRTH-Research Center for Translational Regenerative Medicine, Hannover Medical School, Hannover, Germany; 3CureVac
AG, Tübingen, Germany; 4Twincore Centre for Experimental and Clinical Infection Research, Hannover, Germany; 5Department of
Gastroenterology, Zhongshan Hospital of Fudan University, Shanghai, China; 6Institute of Transplant Immunology, Hannover Medical School,
Hannover, Germany; 7Institute of Experimental Hematology, Hannover Medical School, Hannover, Germany; 8Division of Hematology/
Oncology, Boston Children’s Hospital, Harvard Medical School, Boston, USA; 9Group of Computational Biology and Systems Biomedicine,
Biodonostia Health Research Institute, San Sebastián, Spain; 10IKERBASQUE, Basque Foundation for Science, Bilbao, Spain; 11Department of
General, Visceral and Transplant Surgery Regenerative Medicine and Experimental Surgery, Hannover Medical School, Hannover, Germany;
12
German Center for Infection Research Partner Site Hannover-Braunschweig Hannover, Germany; 13Present address of NH: MRM Health NV
Technologie park-Zwijnaarde 94, 9052 Gent; Belgium; 14Present address of TY: Department of Liver Surgery, Renji Hospital, School of
Medicine, Shanghai Jiaotong University, Shanghai, China

Background & Aims: Therapeutic targeting of injuries that
require transient restoration of proteins by mRNA delivery is an
attractive approach that, until recently, has remained poorly
explored. In this study, we examined the therapeutic utility of
mRNA delivery for liver fibrosis and cirrhosis. Specifically, we
aimed to demonstrate the therapeutic efficacy of human hepa-
tocyte nuclear factor alpha (HNF4A) mRNA in mouse models of
fibrosis and cirrhosis.
Methods: We investigated restoration of hepatocyte functions by
HNF4A mRNA transfection in vitro, and analyzed the attenuation
of liver fibrosis and cirrhosis in multiple mouse models, by
delivering hepatocyte-targeted biodegradable lipid nanoparticles
(LNPs) encapsulating HNF4A mRNA. To identify potential mech-
anisms of action, we performed microarray-based gene expres-
sion profiling, single-cell RNA sequencing, and chromatin
immunoprecipitation. We used primary liver cells and human
liver buds for additional functional validation.
Results: Expression of HNF4A mRNA led to restoration of the
metabolic activity of fibrotic primary murine and human hepa-
tocytes in vitro. Repeated in vivo delivery of LNP-encapsulated
HNF4A mRNA induced a robust inhibition of fibrogenesis in 4
independent mouse models of hepatotoxin- and cholestasis-
induced liver fibrosis. Mechanistically, we discovered that
Keywords: mRNA therapeutics; Transcription factors; protein replacement
and cirrhosis.
Received 26 January 2021; received in revised form 30 July 2021; accepted 2 August
2021; available online 25 August 2021
* Corresponding author. Address: Department of Gastroenterology, Hepatology and
Endocrinology, REBIRTH-Research Center for Translational Regenerative Medicine,
Hannover Medical School, Carl-Neuberg Str. 1, 30625 Hannover, Germany; Tel.: 0049
511 532 5255, Fax: 0049 511 532 5234.
E-mail address: sharma.amar@mh-hannover.de (A.D. Sharma).
https://doi.org/10.1016/j.jhep.2021.08.011
paraoxonase 1 is a direct target of HNF4A and it contributes to
HNF4A-mediated attenuation of liver fibrosis via modulation of
liver macrophages and hepatic stellate cells.
Conclusion: Collectively, our findings provide the first direct
preclinical evidence of the applicability of HNF4A mRNA thera-
peutics for the treatment of fibrosis in the liver.
Lay summary: Liver fibrosis and cirrhosis remain unmet medical
needs and contribute to high mortality worldwide. Herein, we
take advantage of a promising therapeutic approach to treat liver
fibrosis and cirrhosis. We demonstrate that restoration of a key
gene, HNF4A, via mRNA encapsulated in lipid nanoparticles
decreased injury in multiple mouse models of fibrosis and
cirrhosis. Our study provides proof-of-concept that mRNA ther-
apy is a promising strategy for reversing liver fibrosis
and cirrhosis.
© 2021 The Author(s). Published by Elsevier B.V. on behalf of Euro-
pean Association for the Study of the Liver. This is an open access
article under the CC BY-NC-ND license (http://creativecommons.org/
licenses/by-nc-nd/4.0/).
Introduction
Transient protein expression by transfer of therapeutic mRNA
considerably limits the risks encountered in viral gene therapy
such as insertional mutagenesis, severe immune responses and
potential activation of oncogenes. Tremendous progress has
already been made with mRNA technology being applied to
vaccine development1–3 and immuno-oncology.4 Lipid nano-
particle (LNP)-formulated mRNAs are shielded from humoral and
cellular immunity and can be produced in large quantities at
lower cost compared to viral vectors.5 mRNA therapeutics are
particularly well-suited to treat injuries not requiring constant
and life-long expression of therapeutic proteins, and offer the

Journal of Hepatology 2021 vol. 75 j 1420–1433

advantage over viral gene therapy of being adaptable to various
medical conditions such as liver fibrosis and cirrhosis. The po-
tential of this new class of mRNA based-drugs in liver fibrosis
and cirrhosis, which contribute to millions of deaths annually
and represent a major healthcare burden worldwide,6 remains to
be harnessed.
To test whether mRNA delivery is able to mitigate liver fibro-
genesis, we selected the human transcription factor HNF4A as a
candidate for attenuation of liver fibrosis. The rationale for select-
ing HNF4A was based on its role as a master regulator of the he-
patocyte phenotype7,8 and as a key factor in xenobiotic
metabolism.9 Moreover, reduced expression of HNF4A has been
reported in liver fibrosis10,11 and viral vector-mediated HNF4A
restoration has previously been demonstrated to reverse features
of chronic liver injury.10,11 Diminished HNF4A-dependent gene
expression has been suggested as a major indicator of hepatocel-
lular failure in alcoholic hepatitis.12 HNF4A, known to inhibit
epithelial-to-mesenchymal transition, which is observed during
wound healing and liver fibrosis,13 is downregulated in liver
fibrosis in HCV-infected patients.14 Inhibition of the HNF4A tran-
scriptional network is often seen with increased matrix stiffness in
cirrhotic livers.15 Therefore, HNF4A has been suggested as a ther-
apeutic target for the treatment of liver fibrosis.16 However, the
question of whether an mRNA-based therapeutic coding for HNF4A
is able to attenuate liver fibrosis and cirrhosis, needs to be
addressed. Therefore, in our present study, we examined the
therapeutic delivery of HNF4A mRNA in several mouse models of
liver fibrosis and cirrhosis, and in human liver cells.
Materials and methods
Ethics Statement
RNA samples from fibrotic human livers were provided by our
prospective biorepository after liver transplantation (Table S1A).
The study was approved by the Ethics Committee of Hannover
Medical School, Hannover, Germany (protocol number 933 for
project Z2 of comprehensive research centre 738). Written
informed consent was obtained from all participants. All exper-
iments were performed in accordance with relevant guidelines
and regulations. No organs or tissues were procured from pris-
oners. All liver transplantations of patients in this study were
performed in Germany, where organ procurement and allocation
are organized by Eurotransplant and the German Organ Trans-
plantation Foundation. For the second cohort, qPCR analyses
were provided by Shanghai Zhongshan Hospital, China
(Table S1B) with ethical approval from the institute. There was
no international transfer of tissue or RNA samples.
Statistical analyses
Statistical significance was determined with the 2-tailed Stu-
dent’s t test or 2-sided Welch’s t test. Statistical significance was
depicted with error bars representing ± SEM. p <0.05 was
considered significant.
For further details regarding the materials and methods used,
please refer to the CTAT table and supplementary information.
Results
Human HNF4A mRNA restoration improves function of
fibrotic hepatocytes
At first, we confirmed reduced endogenous HNF4A mRNA levels
in patients with liver fibrosis, graded by the Ishak score17
(Fig. 1A). In 2 cohorts from Hannover Medical School, Germany,
Journal of Hepatology 2021 vol. 75 j 1420–1433
and Shanghai Zhongshan Hospital, China, liver tissues showed
fibrosis stage-dependent reduction in HNF4A mRNA. Similarly,
HNF4A protein expression was reduced in primary human he-
patocytes (PHHs) from fibrotic human livers (Fig. 1B).
We then examined whether in vitro synthesized HNF4A mRNA
produces functionally active HNF4A protein. We transfected wild-
type (non-codon-optimized) or codon-optimized HNF4A mRNA
into the human cervical carcinoma cell line (HeLa) cells as these
lack endogenous HNF4A expression. Codon-optimized HNF4A
mRNA-transfected cells showed higher HNF4A protein levels than
those transfected with wild-type HNF4A mRNA (Fig. 1C). ZsGreen
mRNA-transfected HeLa cells were our control. Subsequent
transfection in PHHs confirmed functionally active codon-
optimized mRNA as indicated by significant upregulation of
HNF4A-target genes APOA2, APOB and F7 (Fig. 1D). We therefore
used codon-optimized HNF4A mRNA for further experiments.
We determined whether HNF4A restores impaired functions
in PHHs from fibrotic livers. Robust HNF4A protein expression in
PHHs after transfection with HNF4A mRNA was observed
(Figs. 1E-F). qPCR analyses showed significantly increased
expression of hepatocyte markers albumin (ALB), alpha-1 anti-
trypsin (A1AT), transferrin (TF) and transthyretin (TTR) (Fig. 1G).
Importantly, HNF4A mRNA upregulated expression of various
enzymes involved in phase I and phase II drug metabolism and
phase III drug transporters (Fig. 1H). HNF4A mRNA transfection in
hepatocytes from non-fibrotic livers also enhanced expression of
hepatocyte markers (Fig. 1G), enzymes, and transporters of drug
metabolism (Fig. 1H). Significant increases in secreted protein
levels of ALB and A1AT in PHHs transfected with HNF4A mRNA
indicated that restoration of HNF4A in fibrotic PHHs revives
function (Fig. 1I).
To examine drug responsiveness, we treated HNF4A mRNA-
transfected PHHs with b-naphthoflavone, phenobarbital and
rifampicin and measured expression of CYP1A2, CYP2B6, CYP2C9,
CYP2C19 and CYP3A4. Fibrotic PHHs had a reduced response to all
drugs, and HNF4A mRNA delivery restored cytochrome activities
significantly (Fig. 1J).
We next tested whether HNF4A mRNA delivery restores the
function of fibrotic primary mouse hepatocytes (PMHs) isolated
from mice with fibrosis due to carbon tetrachloride (CCl4) in-
jections bi-weekly, for 8 weeks (Fig. S1A). qPCR analyses
showed hepatocyte-specific expression of endogenous Hnf4a
by qPCR (Fig. S1B) and HNF4A protein by co-staining with
markers of different hepatic cells (Fig. S1C). Therefore, we
aimed to restore HNF4A expression using hepatocyte-targeted
LNPs. ZsGreen mRNA transfection led to robust corresponding
protein expression in PMHs (Fig. S2A). HNF4A mRNA trans-
fection induced expression of target genes in PMHs (Fig. S2B-D)
and led to functional restoration in fibrotic PMHs (Fig. S2E-K).
Of note, DNASTAR MegAlign shows 95.8% sequence homology
between human HNF4A and mouse Hnf4a. DNA binding do-
mains of both exhibit similar molecular function due to high
homology. Thus, our in vitro analyses confirm human HNF4A
mRNA delivered into PMHs improves function, as observed
in PHHs.
Targeted delivery of LNP-encapsulated mRNA into
hepatocytes of fibrotic livers
We established efficient LNP-mediated mRNA delivery into he-
patocytes of murine fibrotic livers. LNPs direct cell entry by
interacting with apolipoprotein E, and are subsequently
1421
Research Article Experimental and Translational Hepatology

A MHH, Germany Zhongshan, China B Phalloidin DAPI HNF4A Merge

Fibrotic PHH Control PHH

2.0 *** **
2.0 *
Relative mRNA levels

*** **
**
n.s *
1.5 1.5 *
HNF4A

1.0 1.0

0.5 0.5

0.0 0.0
0 1 2 3 4 0 2 3 4 5 6
APOA2 APOB F7
HNF4A
C D

Relative mRNA levels

2.0
15 3 * 20
** **
A
A

***
F4

**
F4

Normalized by
* 15
HN

1.5 10
HN

2
l
ro

vinculin
nt

PT

***
T

10 ***
Co

1.0
O

***
Vinculin
5 1 **
0.5 5
HNF4A
0.0 U.D. 0 0 0

HN 4A en

HN 4A en

HN 4A en
HN 4 3H
A H
H

HN F4 3H
A H
H

H N 4 3H
A H
H
A
O HN rol
HN 4A

F4 A 6
12

F4 A 6
12

F4 A 6
12
F4

F re

F re

F re
t
PT F
W Con

HN sG

HN sG

HN sG
F

F
Z

Z
T

Control PHH
E Control PHH Fibrotic PHH Control
+ZsGreen mRNA
+HNF4a mRNA
Fibrotic PHH
Control
+ZsGreen mRNA
Control +ZsGreen mRNA +HNF4A mRNA Control +ZsGreen mRNA +HNF4a mRNA +HNF4a mRNA

F G
A

A
RN

RN

A
RN

RN
m

HNF4A ALB A1AT TF TTR

m

m
en

en

Relative mRNA levels

6
4a

4a

5 3 2.5 * 4
re

re
l

*
ro

ro

NF

NF

p = 0.0588
*
sG

sG

* *
Normalized by
nt

nt

4 * 2.0 *
+H

+H
Co

Co

**
+Z

+Z

4 3
*
vinculin

2
3 1.5 **
Vinculin 2
2 * 1.0
2 ** 1
1 ** p = 0.059
0.5 1
HNF4A
0 0 0 0.0 0

H Phase I Phase II
CYP1A2 CYP2B6 CYP2C9 CYP2C19 CYP3A4 UGT1A1 UGT2B7 UGT2B15
Relative mRNA levels

n.s
5 ** 3 2.5 2.5 p = 0.0690
4 * 2.5 p = 0.0921 2.0 *
** * ** n.s
**
p = 0.0714
2.5 n.s

4 ** 2.0 * 2.0 * 2.0 *

* 3 1.5 ** 2.0
2 *
3 1.5 1.5 1.5 1.5 *
* 2 ** * 1.0 *
2 *** 1.0 * 1.0 ** 1.0 ** 1.0
** 1 **
1 ** 0.5 0.5 1 0.5 0.5 0.5
0 0 0.0 0.0 0 0.0 0.0 0.0
Phase III
SLC10A1 SLCO1B3 AHR CAR FXR PXR RXRA RXRB
Relative mRNA levels

p = 0.0527
2.5 ** 2.0 p = 0.0809 2.0 * 2.5 * 2.5 2.5 2.0 2.5
n.s
* ** * * * * n.s n.s

2.0 ** 2.0 2.0 * 2.0 * *

1.5 1.5 1.5 2.0
n.s
1.5 1.5 * 1.5 1.5 * * 1.5 *
1.0 * 1.0 p = 0.0786 * 1.0 * *
* * **
1.0 ** 1.0 1.0 1.0 1.0
0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
I J β-Naphthoflavone Phenobarbital Rifampicin
ALB A1AT CYP1A2 CYP2B6 CYP2C9 CYP2C19 CYP3A4
15 ** 40 2.5
** 2.5 2.5 2.0 *** 2.0
μg/24 h/100,000 cells

** *
μg/24 h/100,000 cells

mRNA fold induction

* ** ** * *
** 2.0 2.0 * 2.0
*
** 30 ** ** 1.5 1.5
10 ** ** * ** *
1.5 1.5 * 1.5 **
20
1.0 1.0 1.0
** 1.0 1.0
**
5 ** **
10 0.5 0.5 0.5 0.5 0.5

0 0 0.0 0.0 0.0 0.0 0.0

Fig. 1. HNF4A mRNA delivery restores function of fibrotic PHHs. (A) qPCR analyses of human HNF4A mRNA in patients with fibrosis; 0 (n = 9), 1 (n = 3), 2 (n = 3),
3 (n = 3) and 4 (n = 3) from MHH, Germany and 0 (n = 6), 2 (n = 4) 3 (n = 3), 4 (n = 5), 5 (n = 4), 6 (n = 5) from Zhongshan Hospital, China. (B) Immunofluorescence
staining show reduced HNF4A protein expression in fibrotic PHHs. Scale bars, 100 lm. (C) Western blot and its quantification for wild-type (non-codon-

1422 Journal of Hepatology 2021 vol. 75 j 1420–1433

A B 8h

1016
1015

Total flux (p/s)

PEG-lipid 1014
WT
1013
1012
Ionizable lipid 1011
1010
109
Structural lipid 108
107

B a Le p s i d
ck ft s de
o e
L e ps nd
ft ide
e
CCl4

U un

gr id

sid
U u
o
gr
Cholesterol

ck
Ba
mRNA 24 h

C ZsGreen/DAPI ZsGreen/DAPI/ALB ZsGreen/DAPI/Desmin 1015

Total flux (p/s)

1014
WT
1013
WT 1012
1011
1010
109

Ba Le psi d
ck ft s de
o e
Le ps nd
ft ide
e
CCl4

U un

gr id

sid
U u
ro
CCl4

c kg
Ba
48 h
ZsGreen/DAPI/CD45 ZsGreen/DAPI/F4/80 ZsGreen/DAPI/SOX9
1014

Total flux (p/s)

1013
WT WT
1012
1011
1010
109
108

Ba Le psi d
ck ft s de
o e
Le ps nd
ft ide
e
CCl4

U un

gr id

sid
CCl4

U u
o
gr
ck
Ba
Fig. 2. Successful mRNA delivery into hepatocytes of fibrotic BALB/c mice by LNP. (A) Illustration of LNPs designed to deliver mRNA into hepatocytes. (B)
Bioluminescence analyses at 8 h, 24 h, and 48 h in WT control and CCl4-induced fibrotic mice (n = 3 mice per group) injected i.v. with 40 lg luciferase mRNA-
encapsulated LNP (Luc/LNP). (C) Immunofluorescence images of ZsGreen expression and co-staining with hepatocyte marker (ALB), HSC marker (desmin),
leukocyte marker (CD45), KC marker (F4/80) and cholangiocyte marker (SOX9) in WT control and CCl4-induced fibrosis in mice, injected i.v. with 40 lg ZsGreen/
LNP. Scale bars: 100 lm for DAPI and Desmin, 50 lm for ALB, CD45, F4/80 and SOX9 images. CCl4, carbon tetrachloride; HSC, hepatic stellate cell; KC, Kupffer cell;
LNP, lipid nanoparticle; WT, wild-type.

internalized into hepatocytes via endocytosis.18–21 However,
whether LNP-formulated mRNA reaches hepatocytes of fibrotic
livers efficiently has remained unknown. LNP carrying mRNA
encoding for Photinus pyralis luciferase (Luc/LNP) or ZsGreen
(ZsGreen/LNP) were injected into CCl4-induced fibrotic and non-
fibrotic control (wild-type) mice i.v. (Fig. 2A). In vivo imaging
revealed robust bioluminescence exclusively in the liver from 8
hours to 120 hours after Luc/LNP injection in both groups
(Fig. 2B). Co-localization studies showed robust ZsGreen protein
expression specifically in hepatocytes of fibrotic and control
mice, but not in other hepatic cells (Fig. 2C). Our results
demonstrated that LNP-formulated mRNA technology could
specifically and efficiently target hepatocytes in murine
fibrotic livers.
Successful mRNA delivery should be therapeutic but with
negligible immune response. Therefore, we first monitored im-
mune response after injecting different doses of ZsGreen/LNP
(Fig. S3A). Measurement of different cytokines and chemokines
indicated that mRNA doses of 1 mg/kg and 2 mg/kg did not elicit
substantial immune responses, whereas 4 mg/kg mRNA induced
a significant immune response (Fig. S3B). Therefore, we used 2
mg/kg mRNA/LNP for subsequent experiments.
We next examined whether repeated i.v. administration of 2
mg/kg mRNA/LNP was safe in fibrotic mice (Fig. S4A). Repeated
injections of 2 mg/kg LNP encapsulating HNF4A mRNA (hence-
forth referred to as HNF4A/LNP) every 5 days induced no sig-
nificant elevations of most cytokines and chemokines (Fig. S4B).
Thus, administration of either single or multiple doses of 2 mg/
kg HNF4A/LNP is well tolerated.
Delivery of human HNF4A mRNA attenuates toxin- and
cholestasis-induced fibrosis
To address whether systemic administration of HNF4A mRNA
inhibits liver fibrosis, we used mice with toxin- (repeated CCl4

=
optimized) and codon-optimized HNF4A mRNA in HeLa cells, 6 hours after HNF4A mRNA delivery. (D) qPCRs for HNF4A target genes. (E) Schematic of transfection
experiments. (F) Western blot (left) and its quantification (right) showing HNF4A protein levels. (G) qPCRs for ALB, A1AT, TF and TTR. (H) qPCRs for multiple CYPs,
other enzymes, and transporters in phase I, phase II and phase III. (I) ELISA for secreted ALB and A1AT. (J) Drug responsiveness of control and fibrotic PHHs upon
HNF4A or ZsGreen mRNA transfection, as determined by expression of CYPs after treatment with indicated drugs. Two-sided Welch’s t test (panel a) or 2-tailed
Student’s t test (C,D and G-J). * p <0.05; ** p <0.01; *** p <0.001. MHH, Hannover Medical School; PHHs, primary human hepatocytes.

Journal of Hepatology 2021 vol. 75 j 1420–1433 1423

Research Article Experimental and Translational Hepatology

A PBS 16x B Hnf4a I Normal diet

J Hnf4a
1.5

Relative mRNA levels

Relative mRNA levels

CCl4 16x
2.0
** DDC diet ***
CCl4 16x HNF4A/LNP 1.5 HNF4A/LNP 1.0
DDC diet
1.0 ZsGreen/LNP
CCl4 16x ZsGreen/LNP
DDC diet 0.5
0.5 AAV-HNF4A
CCl4 16x AAV-HNF4A
DDC diet
0.0 0.0
CCl4 16x AAV-control AAV-control

ice

ice
ice
ice
DDC diet

m
lm
lm

et

et
ro
ro

di

di
nt
nt

al

C
co
P56 P84 P112 P56 P70 P98

co

DD
rm
T

4
l

No
CC
W
PBS injection HNF4A/LNP ZsGreen/LNP HNF4A/LNP ZsGreen/LNP

CCl4 injection AAV-HNF4A AAV-control AAV-HNF4A AAV-control

C 1 2 3 4 5 6
1 WT control
2 CCl4 control
K 1 2 3 4 5 6
1 Normal diet control
2 DDC diet control
3 AAV-control 3 AAV-control
Vinculin Vinculin
4 ZsGreen/LNP 4 ZsGreen/LNP
5 AAV-HNF4A 5 AAV-HNF4A
HNF4A 6 HNF4A/LNP
HNF4A 6 HNF4A/LNP

D ALT E Hydroxyproline assay L Billirubin M Hydroxyproline assay

*** ** * **
*** ** 300 * 0.05 **
1,000 *** 0.25 *** * ***
*** **** * ***
800 0.04
0.20
200
μmol/L 0.03

μg/mg
600
μg/mg

0.15
IU/L

400 0.10 0.02

100
200 0.05 0.01

0 0 0 0.00
1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6
F CCl4 control AAV-control ZsGreen/LNP AAV-HNF4A HNF4A/LNP
N DDC control AAV-control ZsGreen/LNP AAV-HNF4A HNF4A/LNP
H&E

H&E
Sirius red Desmin

Sirius red Desmin

G Desmin Sirius red O Desmin Sirius red

15 **** 20 **** 20 ** 40 ****
**
****
**** **** ***** ** ****
Positive area (%)

Positive area (%)

****
Positive area (%)

Positive area (%)

**** 15 **** **
15 30
10
10 10 20
5
5 5 10

0 0 0 0
2 3 4 5 6 2 3 4 5 6 2 3 4 5 6 2 3 4 5 6
H Col1a1 Col2a1 Acta2 P Col1a1 Col2a1 Acta2
* 10 ** 80 *** ** ** 15 ***
15 * ** *** ** 25 ** **
* ** 50 *** ** ***
Relative mRNA levels

Relative mRNA levels

p = 0.0531
** 8 ** ** ** ***
* 60 40 20
10
10 6
30 15
40
4 20 10
5 5
20
2 10 5
0 0 0 0 0 0
1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6

Fig. 3. Therapeutic HNF4A/LNP delivery inhibits toxin- and cholestasis-induced liver fibrosis. (A, I) Schematic of experiments analysing effect of human
HNF4A mRNA delivery in mouse models of CCl4- (A-H) and DDC (I-P)-induced liver fibrosis (n = 6 mice per CCl4 model, n = 4 mice per DDC model). (B, J) Hnf4a
qPCRs after CCl4 injection (B) or after DDC diet (J). (C, K) Western blots show HNF4A protein expression. (D, L) Liver function tests for ALT and bilirubin show
reduced injury upon HNF4A/LNP administration. (E, M) Hydroxyproline assay. (F, N) Immunohistochemical images of H&E, desmin, and Sirius red stainings. Scale
bars, 100 lm. (G,O) Quantification of Sirius red and desmin stainings. (H, P) qPCRs for Col1a1, Col2a1 and Acta2. *p <0.05; **p <0.01; ***p <0.001; ****p <0.0001. ALT,
alanine aminotransferase; CCl4, carbon tetrachloride; DDC, 3,5-diethoxycarbonyl-1,4-dihydrocollidine; LNP, lipid nanoparticle; P, postnatal day; WT, wild-type.

1424 Journal of Hepatology 2021 vol. 75 j 1420–1433

injection) (Fig. 3A-H) or cholestasis-induced fibrosis (induced fibrotic mice with 1x1011 AAV8 particles, encoding optimized
via 3,5-diethoxycarbonyl-1,4-dihydrocollidine [DDC]-containing human HNF4A under the transthyretin promoter. Liver function
diet)(Fig. 3I-P). HNF4A mRNA levels significantly decreased in test, hydroxyproline assay, histological analyses, desmin and
fibrotic livers from both models (Fig. 3B, 3J). HNF4A/LNP or Sirius red staining, and qPCR analyses of fibrogenic genes
ZsGreen/LNP (henceforth referred to as control) were injected i.v. showed attenuation of liver fibrosis by HNF4A/LNP is comparable
into fibrotic mice at 2 mg/kg per injection (Fig. 3A, 3I). HNF4A to AAV-mediated HNF4A delivery in CCl4- (Fig. 3D-H) or DDC-
protein expression was confirmed in HNF4A/LNP-injected mouse induced fibrosis (Fig. 3L-P).
livers (Fig. 3C, 3K) and was absent in control mice, since the We then investigated whether transient delivery of HNF4A/LNP
HNF4A antibody was only specific to human but not mouse restores endogenous Hnf4a expression. qPCR analyses showed that
HNF4A. HNF4A/LNP-injected mice showed significantly reduced HNF4A/LNP significantly increased endogenous mouse Hnf4a
alanine aminotransferase (ALT) (Fig. 3D) and bilirubin (Fig. 3L) expression (Fig. S5A, B), indicating a long-lasting phenotypic
indicating improved liver function, and significantly reduced change in hepatocytes. We further investigated whether human
levels of collagen, suggesting decreased fibrosis in both CCl4 HNF4A mRNA delivery enhances the expression of other hepatic
(Fig. 3E) and DDC (Fig. 3M) models. Histological analyses, desmin transcription factors, which can restore endogenous mouse Hnf4a.
and Sirius red staining further confirmed reduced fibrosis in CCl4 We transfected fibrotic PMHs with human HNF4A mRNA and
(Fig. 3F-G) and DDC groups (Fig. 3N-O). Additionally, qPCR determined the expression of mouse Hnf1a via qPCR and western
showed significantly decreased expression of fibrogenic marker blot analyses (Fig. S6A). Hnf1a could further induce endogenous
genes, Col1a1, Col2a1 and Acta2 in (Fig. 3H, P). Thus, our data Hnf4a expression via binding to its promoter region. Within 24
together provide evidence that HNF4A mRNA delivery attenuates hours after human HNF4A mRNA transfection, we observed
fibrosis in mouse models of toxin- as well as cholestasis- increased expression of Hnf1a mRNA (Fig. S6B) and protein in he-
induced fibrosis. patocytes (Fig. S6C), and of endogenous mouse Hnf4a (Fig. S6D).
We compared the efficacy of HNF4A/LNP with recombinant To determine whether human HNF4A mRNA influences sta-
adeno-associated virus (AAV) vector serotype 8, by injecting bility of mouse endogenous Hnf4a mRNA, we analyzed mRNA

A B 1.5
Hnf4a C

l
ro
CCl4 32x ***
Relative mRNA lelvel

nt

NP
N
co

/L

L
is

en

A/
os
CCl4 32x HNF4A/LNP 1.0

re

F4
rrh

HN
Zs
Ci
CCl4 32x ZsGreen/LNP 1 WT control
0.5 Vinculin 2 Cirrhosis control
3 ZsGreen/LNP
HNF4A 4 HNF4A/LNP
P56 P140 P168 0.0
ice
ice

m
m

CCl4 injection HNF4A/LNP ZsGreen/LNP

is
T
W

os
rrh
Ci

D ALT E Hydroxyproline assay F Body weight G Activity

Cirrhosis control
6,000 * 1.5 34 5 ZsGreen/LNP
* ** 32 4 HNF4A/LNP
*
4,000 1.0 30
Grams

3
μg/mg

Score
U/L

28
2
2,000 0.5 26
24 1

0 0.0 22 0
1 2 3 4 1 2 3 4 140 150 160 170 140 150 160 170
Age Age
H Cirrhosis con. ZsGreen/LNP HNF4A/LNP
I Desmin Sirius Red
J Col1a1 Col2a1 Acta2

60 ** 200 * 100 500

*** 50 * ***
H&E

** **
**
Relative mRNA lelvel

**
80 400
Positive area (%)

40 150 **
40
Sirius red Desmin

30 60 300
100
20 40 200
20
50
10 20 100

0 0 0 0 0
2 3 4 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4

Fig. 4. Mitigation of liver injury in cirrhotic mice by delivery of HNF4A/LNP. (A) Schematic of experimental design. (B) qPCR for endogenous Hnf4a mRNA. (C)
Western blot revealing HNF4A protein expression in livers after HNF4A/LNP injection. Graphs showing (D) Serum ALT levels and (E) Hydroxyproline assay. (F, G)
Body weight and activity score for cirrhotic mice. (H, I) Images of H&E, desmin, and Sirius red staining and quantification. Scale bars, 100 lm. (J) qPCRs for Col1a1,
Col2a1, and Acta2 in cirrhotic mice livers. * p <0.05; ** p <0.01; *** p <0.001. ALT, alanine aminotransferase; CCl4, carbon tetrachloride; LNP, lipid nanoparticle; WT,
wild-type.

Journal of Hepatology 2021 vol. 75 j 1420–1433 1425

Research Article Experimental and Translational Hepatology

decay as a measure of mRNA stability following transcriptional days after bile duct ligation, was observed (Fig. S9B). We per-
inhibition with Actinomycin D treatment in PMHs (Fig. S6E). We formed H&E, TUNEL, Ki67, CD45 and F4/80 staining (Fig. S9C-E).
found that Actinomycin D caused similar decay of mouse Hnf4a HE and F4/80 staining showed similar amounts of necrosis but
mRNA in cells transfected with either human HNF4A mRNA or decreased numbers of macrophages in HNF4A mRNA injected
control ZsGreen mRNA (Fig. S6F). These results support our mice. However, there was no significant difference in cells with
findings that human HNF4A mRNA favors transcription of TUNEL, Ki67 or CD45 staining. Similar results were observed in
endogenous mouse Hnf4a, in part via Hnf1a, without affecting the bile duct ligation model. Thus, HNF4A mRNA delivery may
mRNA stability (Fig. S6G). not affect early phases of acute liver injury induced by either CCl4
We addressed whether alternate HNF4A/LNP-treatment injection or bile duct ligation.
schedules (2 weeks or 6 weeks) and doses (0.5 and 1 mg/kg) are
equally effective. Although both additional treatment schedules Delivery of human HNF4A mRNA inhibits liver injury in a
reduced injury, only mice treated for 6 weeks showed a signifi- mouse model with features of cirrhosis
cant reduction in liver fibrosis (Fig. S7). Of the 2 additional doses, To determine whether HNF4A mRNA inhibits cirrhosis in mice
only 1 mg/kg significantly reduced liver fibrosis (Fig. S8). injected with CCl4 twice weekly for 16 weeks (Fig. 4A), HNF4A/
Together, these results indicate that dose and schedule optimi- LNP or ZsGreen/LNP were administered i.v. starting from 12
zation may increase effectiveness of HNF4A mRNA delivery on weeks, once every 5 days. Cirrhotic mice showed loss of Hnf4a
liver fibrosis attenuation. (Fig. 4B). HNF4A/LNP delivery led to HNF4A protein expression,
We next examined whether HNF4A mRNA reduces acute he- as determined by western blot analyses (Fig. 4C). Injury was
patocyte injury induced by either CCl4 administration or bile evaluated by measurement of ALT, hydroxyproline content, body
duct ligation (Fig. S9A). Significant reduction in endogenous weight changes, activity score, histology, desmin, Sirius red
Hnf4a in hepatocytes of mice with either acute CCl4 injury or 7 staining and qPCR analyses for fibrogenic genes. All these

A B Hnf4a C

l
ro

NP
1.5

NP
Relative mRNA level

nt

/L
HNF4A/LNP

co
*

L
en

A/
2 -/-

re

F4
G
dr

HN
Zs
M
ZsGreen/LNP
1.0
Vinculin

P0 P84 P112 1 Mdr2-/- control

0.5 HNF4A 2 ZsGreen/LNP
ice
ice

3 HNF4A/LNP
HNF4A/LNP ZsGreen/LNP Mdr2-/- mice
lm
lm

ro
ro

nt
nt

co
co
B

2 -/-
FV

dr
M

D Billirubin
E Hydroxyproline assay
F Ck19 Sox9 Epcam
30 2.0 ** 2.0 2.0 *
** * 2.0 *
Relative mRNA level

* * * *
*
1.5 1.5 1.5 1.5
20
μmol/L

μg/mg

1.0 1.0 1.0 1.0

10
0.5 0.5 0.5 0.5

0 0.0 0.0 0.0 0.0

1 2 3 1 2 3 1 2 3 1 2 3 1 2 3
G H I
Mdr2-/- Control ZsGreen/LNP HNF4A/LNP
Desmin Sirius Red Col1a1 Col2a1 Acta2
H&E

25 *** 20 * 2.0 p = 0.0592 2.0 2.0 **

*** * *
Relative mRNA level

*
Positive area (%)

20 *
15 1.5 1.5 1.5
Desmin

15
10 1.0 1.0 1.0
10
5 5 0.5 0.5 0.5
Sirius red

0 0 0.0 0.0 0.0

1 2 3 1 2 3 1 2 3 1 2 3 1 2 3

Fig. 5. HNF4A/LNP delivery attenuates liver injury in Mdr2-/- mice. (A) Schematic of experimental design. (B) qPCR for endogenous Hnf4a mRNA in FVB control
and Mdr2-/- mice. (C) Western blot revealing HNF4A protein expression after HNF4A/LNP injection. Graphs showing (D) Serum bilirubin levels and (E) Hy-
droxyproline assay for collagen content. (F) qPCRs for Ck19, Sox9, and Epcam. (G) Images of H&E, desmin, and Sirius red staining (H) and quantification. Scale bars,
100 lm. (I) qPCRs for Col1a1, Col2a1, and Acta2 in livers from Mdr2-/- mice. *p <0.05; **p <0.01; ***p <0.001. LNP, lipid nanoparticle.

1426 Journal of Hepatology 2021 vol. 75 j 1420–1433

A Control HNF4A
Dmd
B Control HNF4A
Hnf1a
Cntr1
Cyp3a11 Coa6
Gcm1 Slc18a1
Dgka
Frmpd1 P2ry1
Hnf1a Pbld2
B3ga1t1
Gstm1 Plscr2
-4.01
Rnd1 Dmd
Serpinb6b -2.67
Slc22a7 Pon1
Pon1 -1.33
Slc22a7
Duox1 0.00
Po1n Ccl4
Ppp1r1b Arnt1
1.33
Sord
Ell3 Syn1 2.67
Enthd2 Tomm401 4.00
Zkscan5
Reck
Kdf1
Aatk C Chemical carcinogenesis

KEGG Pathway
Ugt3a2 Glutathione metabolism
Sepsecs Drug metabolism
Arl4d
Ces2b Metabolism of xenobiotics by cytochrome P450
Coa6 Fluid shear stress and atherosclerosis
Mgst3
Cat Hepatocellular carcinoma
Slc18a1 Steroid hormone biosynthesis
Gstm4
Ugt2a1 Retinol metabolism
Cyp2c54 Transcriptional misregulation in cancer
Srr
Pbld2 Prostate cancer
P2ry1
Plscr2 Glutathione transferase activity (GO:0004364)
Cdcp1 Complement receptor activity (GO:0004875)

GO Molecular
Luc712
Cited4 NAD(P)H oxidase activity (GO:0016174)
Arid5a Arylesterase activity (GO:0004064)
Maats1
Arnt1
Phospholipid scramblase activity (GO:0017128)
Synl cAMP-dependent protein kinase regulator activity(GO:0008603)
Fprl Aryl hydrocarbon receptor binding (GO:0017162)
Vgll4
Clec1b G-protein coupled nucleotide receptor activity (GO:0001608)
Etv5 Monoamine transmembrane transporter activity (GO:0008504)
Tubb5
Tubb4a G-protein coupled purinergic nucleotide receptor activity (GO:0045028)
Igsf8
Ctsj Glutathione derivate biosynthetic process (GO:1901687)
G6pd2 Glutathione derivate metabolic process (GO:1901685)

GO Biological
Tomm401
Zdhhc2 Xenobiotic catabolic process (GO:0042178)
Itprip -3.01 Organonitrogen compound biosynthetic process (GO:1901566)
Gpr84 -2.01
Ccl4 Complement receptor mediated signaling pathway (GO:0002430)
Ebi3 -1.00 peptide metabolic process (GO:0006518)
Lrrc25
Fmnl1 0.00 Glutathione metabolic process (GO:0006749)
Samsn1 1.00 Sulfur compound biosynthetic process (GO:0044272)
Mefv
Fpr2 2.01 Positive regulation of leukocyte chemotaxis (GO:0002690)
Sdf211
3.01 Serine family amino acid metabolic process (GO:0009069)
Mvcn

D HNF4A BS1 HNF4A BS2 HNF4A BS3

Paraoxonase 1
5’ 3’
0 644-632 777-763 989-977 2000nt
649-637
E 1.0 PMH 0.5
Hepa 1-6 G 1.0 3T3 cell
BS1
Input control (%)
Input control (%)
Input control (%)

BS2
0.8 0.4 0.8 BS3

0.6 0.3 0.6

0.4 0.2 0.4
0.2 0.1 0.2
0.0 0.0 0.0
IP

IP
IP

IP

IP

IP

IP

IP

IP

IP
Ig IP

Ig P

Ig IP

Ig IP

Ig P

Ig IP
I

A
G

G
A

F4

F4
Ig

Ig
F4

F4

F4

F4

F4

F4

HN

HN
HN

HN

HN

HN

HN

HN

ZsGreen HNF4A
F PMH Hepa1-6 H 3T3 cell
BS1 BS2 BS3 BS1 BS2 BS3 BS1 BS2 BS3
ZsGreen HNF4A ZsGreen HNF4A ZsGreen HNF4A
HN G IP

on IP
P

P
on IP
P

pu F4 P
on IP
P
pu F P
on IP
P

tc 4 P
tro P
P

pu NF IP
on IP
P
HN G IP

on IP
l IP

HN G IP

on IP
l IP

HN G IP

on IP
l IP

IP

on IP

Ig IP
HN G IP

on IP
l IP

HN G IP

on IP
P
lI

lI

lI

HN G I

lI
In HN G I

lI

pu HNF G I
on A I
lI

lI
t c 4A

G
tc A

tc A

t c 4A

G
t c 4A
tc A

tc A

tc A

tc A

tc A

tc A

tro

t ro

tro

tro

tro
tro

tro

tro

tro

t ro

tro

pu F4
pu F4

pu F4

pu F4

pu F4

pu F4

pu F4
Ig

Ig

Ig

Ig

Ig

Ig

Ig
Ig

Ig

Ig

Ig

N
HN

H
pu
In

In

In

In

In

In

In
In

In

In

In

I Control J Paraoxonase 1 K Paraoxonase 1 activity L Paraoxonase 1 activity

RFU (Ex/Em = 368/480 nm)

** 2,000 1.6
Relative Pon 1 activity

PMH Control ***

Relative mRNA levels

2.5 ** ZsGreen mRNA **

HNF4A mRNA
2.0 1,500 1.4
ZsGreen mRNA 1.5
1,000 1.2
PMH 1.0
1.0
0.5 500
0.0 0.8
HNF4A mRNA 0
A
en trol

A
ro

RN
RN

RN

RN
nt

0 20 40 60
n

PMH
Co

Co

m
m

m
A
A

en

Time (min)
F4
F4
re

re
HN
HN
G

G
Zs

Zs

Fig. 6. HNF4A reduces fibrosis by directly targeting the paraoxonase 1 gene. (A) Heat map shows significantly deregulated genes. (B) Selected genes encoding
secretory proteins between Control (ZsGreen/LNP) and HNF4A/LNP treatment groups. (C) KEGG pathway, GO molecular function and GO biological process
enrichment plot. (D) HNF4A binding site prediction within the genomic upstream promoter region of paraoxonase 1. (E-H) qPCRs and gel confirmation in ZsGreen
or HNF4A mRNA-transfected (E, F) PMHs, Hepa1-6, and (G, H) 3T3 cells, after ChIP with 3 different primer pairs surrounding the potential HNF4A BS. Values are
normalized to input control for HNF4A antibody and IgG control ± SEM (n = 3). (I) Schematic of experiment to verify HNF4A mRNA’s effect on paraoxonase 1. (J)
qPCR for paraoxonase 1 expression, (K) dynamic paraoxonase 1 activity, and (L) average paraoxonase 1 activity (n = 3). BS, binding site; ChIP, chromatin
immunoprecipitation; GO, gene ontology; LNP, lipid nanoparticle; PMHs, primary mouse hepatocytes.

Journal of Hepatology 2021 vol. 75 j 1420–1433 1427

Research Article Experimental and Translational Hepatology

A B Paraoxonase 1
C
ZsGreen mRNA CCL2
2.5 ** 1,000
48 h ** *** * **

Relative mRNA levels

2.0 800
Supernatant HSC HSC
post-transfection 1.5 600

pg/ml
HNF4A mRNA
1.0 400
48 h
0.5 200
HNF4A mRNA +
Pon1 siRNA 0.0 0
48 h

en

en

A
F4

F4

N
re

re
R

R
N

N
G

G
si

si
H

H
Zs

Zs
n1

n1
Po

Po
KC KC

A+

A+
F4

F4
N

N
H

H
D HNF4A mRNA + G HNF4A mRNA +
ZsGreen mRNA HNF4A mRNA Pon1 siRNA ZsGreen mRNA HNF4A mRNA Pon1 siRNA
250k F4/80+ 250k F4/80+ 250k F4/80+
88.7% 85.2% 87.5%
200k 200k 200k

Desmin
150k 150k
FSH-H

150k

100k 100k 100k

50k 50k 50k

0 0 0
0 102 103 104 105 0 102 103 104 105 0 102 103 104 105

F4/80 PE a-SMA
HNF4A mRNA +
ZsGreen mRNA HNF4A mRNA Pon1 siRNA
Q1 Q2 Q1 Q2 Q1 Q2
105 105 105
32.3% 34.9% 5.55% 34.6% 12.3% 45.2%
CD11c AF 700

104 104 104

LOX

103 103 103

102 10 2
102
Q4 Q3 Q4 Q3 Q4 Q3
0 0 0
19.7% 13.1% 7.77% 52.1% 7.76% 34.7%
2 3 4
0 102 103 104 105 0 10 2
10 3
10 4
10 5
0 10 10 10 105

CD206 FITC
E M1 percentage M2 percentage H 1.5
40 ** 60 ** ZsGreen mRNA
Relative CTCF per filed

*** * *** *
* HNF4A mRNA
(mean with SEM)

*
Percentage (%)

HNF4A mRNA +
Percentage (%)

30 1.0 *** Pon1 siRNA

40 **
20 *
20 0.5
10

0 0 0.0
Desmin a-SMA LOX
en

en

A
F4

F4

N
re

re
R

R
N

N
G

G
si

si
H

H
Zs

Zs
n1

n1
Po

Po
A+

A+
F4

F4

Acta2 Col1a1 Col2a1

F I
N

N
H

1.5 1.5 1.5

Relative mRNA levels
Relative mRNA levels

Relative mRNA levels

* **
M1 M2 *
* ***
1.5 4 ** * 1.0 1.0 1.0
** **
Relative mRNA levels

Relative mRNA levels

* ***
3
1.0 0.5 0.5 0.5
** ** **
2
0.5 0.0 0.0 0.0
1
A

A
n

en

en
F4

F4

F4

N
e
re

re

re

R
N

N
si

si

si
G

G
H

H
Zs

Zs

Zs
n1

n1

n1
Po

Po

Po

0.0 0
A+

A+

A+
F4

F4

F4

IL-6 iNOS Arg1 Fizz1

N

N
H

Fig. 7. HNF4A-Pon1 pathway induces macrophage polarization and deactivation of HSCs. (A) Schematic of co-culture experiments. (B) qPCR for paraoxonase 1
in PMHs. (C) ELISA showing secreted CCl2 levels in PMHs. (D) FACS analysis of M1- or M2-positive macrophage distribution pattern, (E) quantification of M1/M2
percentage, and (F) qPCRs for M1 (iNOS and IL6)/M2 (Arg1 and Fizz1) related genes in macrophages 48 h after culturing in supernatant from ZsGreen mRNA, HNF4A
mRNA, and, HNF4A mRNA and paraoxonase 1 siRNA co-transfected PMHs. (G) Immunofluorescence stainings for desmin, a-SMA, and LOX (H) quantified corrected
total cell fluorescence after co-culture experiment. (I) qPCRs for Acta2, Col1a1, and Col2a1 (n = 3). HSCs, hepatic stellate cells; KCs, Kupffer cells; PMHs, primary
mouse hepatocytes; siRNA, small-interfering RNA.

1428 Journal of Hepatology 2021 vol. 75 j 1420–1433

analyses provided evidence that HNF4A/LNP inhibits cirrhosis in
mice (Fig. 4D-J).
Human HNF4A mRNA delivery suppresses cholestasis and
fibrosis in Mdr2-/- mice
Severe forms of cholestasis are frequent causes of cirrhosis.22
Hence, we addressed whether HNF4A mRNA suppresses injury
in multidrug resistance gene 2 knockout mice (Mdr2-/-), a sur-
rogate mouse model of progressive familial intrahepatic chole-
stasis.23,24 We injected 12-week old Mdr2-/- mice with 2 mg/kg
HNF4A/LNP or ZsGreen/LNP, once every 5 days, before sacrificing
them at 16 weeks of age (Fig. 5A). We observed lower expression
of Hnf4a mRNA in Mdr2-/- mice compared to age-matched FVB
background control mice (Fig. 5B). HNF4A/LNP-injected mice
showed HNF4A protein expression (Fig. 5C), significantly
decreased levels of bilirubin, hydroxyproline content, reduced
expression of Ck19, Sox9 and Epcam, reduced injury by histology,
desmin and Sirius red staining, and decreased expression of
fibrogenic genes, Col1a1, Col2a1 and Acta2 (Fig. 5D-I). Thus, our
experiments in Mdr2-/- mice demonstrate that HNF4A mRNA
delivery attenuates features of severe cholestasis in mice.
Mechanisms of HNF4A mRNA-mediated inhibition of
fibrogenesis: paraoxonase 1 contributes to antifibrotic effects
of HNF4A mRNA
To elucidate mechanisms of HNF4A mRNA-mediated fibrosis in-
hibition, we performed transcriptome analyses on cirrhotic livers
from HNF4A/LNP- and control ZsGreen/LNP-treated mice. We
found 67 genes with significant changes in expression (p <0.01;
Fig. 6A). We examined if HNF4A mRNA delivery in hepatocytes
influences other hepatic cells via altered gene expression of
secretory proteins. Thus, we selected 13 significantly deregulated
secreted protein-encoding genes (Fig. 6B). Gene ontology and
KEGG pathway analyses indicated enrichment of drug and
glutathione metabolism pathways (Fig. 6C). To identify direct
targets of HNF4A, among these 13 genes, we used the TRANSFAC
transcription factor database, and found paraoxonase (Pon1) has 4
binding sites for HNF4A in its promoter region (Fig. 6D). Secreted
protein PON1, primarily produced by hepatocytes, has an impor-
tant role in the antioxidant system and a hepatoprotective func-
tion in liver injury.25 Chromatin immunoprecipitation and qPCR
confirmed binding of HNF4A with the Pon1 promoter in PMHs and
Hepa 1-6 cells (Fig. 6E,F), as well as with the mouse Pon1 pro-
moter, in murine 3T3 cells that lack endogenous Hnf4a expression
(Fig. 6G,H). Further, we confirmed elevation in Pon1 mRNA levels,
and enhanced PON1 activity upon transfection of PMHs with
human HNF4A mRNA (Fig. 6I-L). We also confirmed increased
Pon1 expression upon HNF4A mRNA delivery in 4 models of
chronic liver injury (CCl4-induced fibrosis and cirrhosis, Mdr2-/-,
and DDC-induced cholestasis) (Fig. S10). Thus, we identified the
Pon1 promoter as a direct target of HNF4A and demonstrated that
expression and activity of PON1 was enhanced after delivery of
HNF4A mRNA, in vitro and in vivo.
Human liver bud model confirms that PON1 contributes to
HNF4A-mediated attenuation of fibrogenesis
We next established a human liver bud model (Fig. S11A,B),
generated by aggregation of 4 hepatic cell types: functional PHH-
derived organoids, human Kupffer cells (transduced with a GFP-
encoding lentiviral vector), primary human hepatic stellate cells
(HSCs: transduced with a dsRed-encoding lentiviral vector) and
Journal of Hepatology 2021 vol. 75 j 1420–1433
endothelial cells. Transfection of HNF4A mRNA in hepatocyte
organoids reduced secreted levels of C-C motif chemokine ligand
2 (CCL2) and tissue inhibitor of metalloproteinase 1 (TIMP1)
(Fig. S11C) in the liver bud and decreased numbers of Kupffer
cells, as indicated by reduced GFP fluorescence (Fig. S11B).
Importantly, expression of fibrotic markers ACTA2, COL1A1 and
LOX also decreased (Fig. S11D). Inhibition of PON1 by small-
interfering RNA (siRNA) diminished the antifibrogenic effects
observed upon HNF4A mRNA transfection. These results
confirmed our findings that PON1 contributes to HNF4A-
mediated attenuation of fibrogenesis in human cells as well.
HNF4A-Pon1-mediated reduction in macrophage infiltration
and modulation of M1 to M2 polarization contributes to
reduced fibrogenesis
To further verify if PON1 contributes to HNF4A-mediated inhi-
bition of fibrogenesis, we co-transfected PMHs with HNF4A
mRNA and siRNA against Pon1 (Fig. 7A). Pon1 siRNA-transfected
PMHs showed robust Pon1 knockdown (Fig. 7B). 48 hours after
transfection, we collected cell culture medium and examined its
effect on Kupffer cells and HSCs (Fig. 7A). Pon1 has been reported
to degrade CCL2,26 a pro-inflammatory chemokine and a critical
regulator of liver fibrosis.27 We confirmed strongly suppressed
CCL2 secretion in medium collected from HNF4A mRNA-
transfected PMHs (Fig. 7C). Medium from cells transfected with
only HNF4A mRNA increased the number of M2 macrophages,
which are considered as anti-inflammatory macrophages
(Fig. 7D,E). Of note, transition into M2 macrophages has been
shown to suppress fibrosis in mice.28 Medium collected from
PMHs co-transfected with Pon1 siRNA and HNF4A mRNA,
induced the M1 phenotype (Fig. 7D,E). Gene expression analyses
of IL6, iNOS, Arg1 and Fizz1 confirmed that HNF4A drives mac-
rophages into the M2 phenotype, whereas inhibition of Pon1
suppressed transition into M2 macrophages (Fig. 7F). Likewise,
exposure of HSCs to medium collected from PMHs transfected
with HNF4A mRNA reduced mRNA and protein levels of fibro-
genic markers such as desmin, aSMA, LOX (Fig. 7G-I). Further, the
co-transfection of PMHs with Pon1 siRNA and HNF4A mRNA
diminished the HNF4A-mediated reduction in HSC activation
(Fig. 7G-I). Thus, the HNF4A-PON1 pathway suppresses fibro-
genesis in HSCs and induces transition of macrophages into the
anti-inflammatory macrophage phenotype, which together
contribute to the inhibition of fibrogenesis.
To confirm the effect of HNF4A mRNA delivery in chronic liver
injury models, we analyzed M1 and M2 macrophages, and pro-
fibrogenic markers in established chronic liver injury models.
qPCR analysis showed decreased expression of Tnfa, Ccl2 and
iNos, and increased expression of Arg1 and Fizz after HNF4A
mRNA/LNP treatment in all 4 chronic liver injury models
(Fig. S12A,C,E,G). F4/80 immunohistochemistry staining indi-
cated reduced macrophage infiltration after HNF4A mRNA/LNP
treatment (Fig. S12B,D,F,H). Further, expression of pro-fibrogenic
genes Tgfb1, Pdgf, and Timp1, and protein levels of aSMA
decreased after HNF4A mRNA/LNP treatment (Fig. S12).
Single-cell RNA sequencing analyses upon HNF4A
mRNA delivery
To gain further insights into the molecular and cellular changes
induced by HNF4A, we first injected mice with a single adminis-
tration of CCl4 (8 ll/g of 10% CCl4), thus inducing acute liver injury.
24 hours later, we treated mice once with HNF4A/LNP or with
1429
Research Article Experimental and Translational Hepatology

A CCl4
ZsGreen/LNP Hepatocytes
Control
Non-parenchymal cells
Day 2 Day 3
CCl4 Non-parenchymal cells HNF4A
10x Genomics
Hepatocytes ScRNA-seq
HNF4A/LNP
B C

ll
ll
do

C
ol

ce
ce

ac
p

p
DC
HS
He

Ch

Ne
En

Ku
M
B
T
B cell
Neu Hep
Endo Chol
Endo
T cell
UMAP2

HSC
Hep Chol T cell
DC Fraction of cells
Mac B cell in group (%)
DC
Neu
Kup Mac 20 40 60 80100
Kup Mean expression
HSC in group

Kr 1
So 9
x9

Pt t7

Ig 1
Ac 7
C o ta 2
a1
x
T r l5
Cd 2
Bc 3 d
R u 1a
2
r9

Cd c

Cs b

S1 nlg
S1 0 a 8
a4
r2

C1 2
C1 a
Cs b
fr1
Aq b

Cd 22
M tr

Re f3r
Lo

k
Al

up
t1

p
p

bc

nx

79

z
q
q
pr
T

Cc
Kr

Cc

Cc
Ig

Ly
fb

l1

l1

00
UMAP1

t
0 5

0
D E F CD68 control CD68 HNF4A CD68 Mac
Chol 4 4 80
Kup
70
HSC

Number of cells
3 3 60
B cell
UMAP2

UMAP2

UMAP2
DC 50
Neu 2 2 40
T cell
Endo 30
Mac 1 1 20
Control Hep 10
HNF4A 0 10 20 30 40
0 0 0
UMAP1 Number of cell types in control and HNF4A (%)
UMAP1 UMAP1 2 4 6

G 0.15
H K
CCL2 control CCL2 HNF4A CCL2 Mac
6 6
50
0.06
5 5

Number of cells
0.10 40
Density

4 4
Density

UMAP2

UMAP2

0.04
30
3 3
0.05
0.02 20
2 2

10
0.00 1 1
0.00
-10 -5 0 5 10
-20 -10 0 10 20 0 0 0
MPI AMDI UMAP1 UMAP1 2 4 6 8

I J L
Control:9.7% Control:23.3%
Arg1 control Arg1 HNF4A Arg1 Mac
M2-like cells M1-like cells 7
20 HNF4A:35.8% HNF4A:16.3%
10 4 4
6
Number of cells

10 5
3 3
UMAP2

UMAP2
AMDI
AMDI

0 4
0 2 2 3

2
-10 -10 1 1
Pre-activation Tansitional 1
Control:16.7% Control:50.4%
M0-like cells M1-like cells HNF4A:24.4% HNF4A:23.6% 0 0 0
-20 2 4 6
-10 -5 0 5 10 -10 -5 0 5 UMAP1 UMAP1
MPI MPI
M 428
N Apoa2 131.2 145.4
Apoa2
199 Mup3 23.6 32.1
Mup3
229
Mup20 Mup20 3.0 7.6
60
Fah Fah 3.0 3.0
40
Cyp2a12
99 Cyp2a12 2.4 2.7
Cyp3a11
Cyp3a11 3.4 5.5
Control (1795 Hep cells) HNF4A (2370 Hep cells) Control HNF4A
O 373
P
Col3a1 Col3a1 22.0 22.6
249
Tagln Tagln 18.5 13.0
6
C3 C3 1.1 1.0
34
Jund Jund 4.2 2.9
16
Egr1
15 Egr1 2.0 1.3
Klf2
Klf2 1.9 1.5
Control HNF4A
Control (160 HSC cells) HNF4A (117 HSC cells)

Fig. 8. Single-cell RNA sequencing confirms HNF4A/LNP treatment reduces macrophage infiltration and HSC activation. (A) Schematic overview of designed
scRNA-seq experiment for hepatocytes and non-parenchymal cells after ZsGreen/LNP and HNF4A/LNP treatment in high-dose CCl4-injected mice. (B) UMAP of
Leuven clustering of cells from control and HNF4A groups. (C) Dotplot of cluster marker genes. Radii of circles are proportional to number of cells expressing a

1430 Journal of Hepatology 2021 vol. 75 j 1420–1433

ZsGreen/LNP (Fig. 8A). We chose the acute injury model rather
than chronically developed liver fibrosis since this enabled the
identification of direct targets of HNF4A. 24 hours after HNF4/LNP
injection, we isolated all liver cells by Percoll-density gradient
centrifugation and subjected them to 10x Genomics single-cell
RNA sequencing (scRNA-seq). Our scRNA-seq analyses, per-
formed on 13,880 liver cells, identified different populations of
liver cells, including hepatocytes, cholangiocytes, HSCs, macro-
phages, Kupffer cells, neutrophils, LSECs, dendritic cells, plasma B
cells and T cells (Fig. 8B,C). Hepatocytes and macrophages showed
substantial transcriptomic differences (Fig. 8D,E). HNF4A/LNP-
injected mice showed decreased expression of pan-macrophage
marker CD68 (Fig. 8F). MacSpectrum computational analysis of
scRNA-seq data was reported to define macrophage polariza-
tion.29 Our MacSpectrum analyses suggested transition into M2-
macrophages (Fig. 8G-J). Additionally, we observed decreased
expression of pro-inflammatory M1 marker Ccl2 and increased
anti-inflammatory M2 marker Arg1 (Fig. 8K,L) in HNF4A/LNP-
injected mice. We observed increased expression of hepatocyte
genes Apoa2, Mup3, Mup20, Fah, Cyp2a12 and Cyp3a11 in the he-
patocyte cluster, (Fig. 8M,N) and decreased expression of fibrotic
genes Tagln, C3, Jund, Egr1 and Klf2 in the HSC cluster (Fig. 8O,P).
Together, scRNA-seq analyses confirmed our findings that HNF4A
mRNA delivery attenuates liver fibrosis by modulating expression
of genes in cellular compartments including hepatocytes, mac-
rophages and HSCs.
Discussion
Taken together, our work provides the first evidence that ther-
apeutic mRNA delivery restores intracellular HNF4A transcrip-
tion factor levels, induces endogenous Hnf4a expression, and
improves metabolic activity of targeted hepatocytes. Our data
show that hepatocyte-specific delivery of HNF4A mRNA in
injured livers attenuates fibrosis and cirrhosis in multiple inde-
pendent mouse models of liver diseases.
Our study identified a new mechanism showing restoration of
HNF4A modulates non-parenchymal cells such as macrophages
and HSCs. Our transcriptome analyses of the whole liver and
scRNA-seq led to identification of Pon1 as a novel direct target of
HNF4A. Increased PON1 promoted an anti-inflammatory M2
phenotype in macrophages and suppressed the pro-fibrogenic
activity of HSCs. However, the role of macrophages in chronic
liver diseases is complex,30 with the emergence of a new
macrophage phenotype in the fibrotic liver of rodents and human
samples.31 Thus, further classification and evaluation of
macrophage functions after HNF4A mRNA/LNP treatment are
required to understand how HNF4A mediates its antifibrotic effect
through macrophage modulation. Other unknown mechanisms
may further contribute to the antifibrotic properties of HNF4A.
How does human HNF4A mRNA restore expression of mouse
Hnf4a in the liver? Exogenous human HNF4A mRNA may increase
stability of endogenous mouse Hnf4a. However, our results from
mRNA stability analyses excluded this. Exploring another possi-
bility that exogenous human HNF4A may bind to and enhance
expression of other mouse transcription factors, we found HNF4A
mRNA delivery increases mouse Hnf1a expression, which can
bind to promoter regions of multiple hepatocyte genes, including
Hnf4a. Therefore, we speculate that using a feedback loop
mechanism, exogenous human HNF4A first induces endogenous
Hnf1a expression, which in turn binds to the endogenous Hnf4a
promoter, thus restoring mouse Hnf4a levels.
A major strength of our study is the lack of any adverse immune
response due to mRNA or LNP following repeated HNF4A mRNA
administration, even though all in vivo experiments were per-
formed with immunocompetent mice. This is particularly impor-
tant in case of fibrosis relapse, frequently observed in patients.
Multiple injections of viral vectors may not be feasible due to
development of viral capsid-specific antibodies after the first in-
jection and, hence, may preclude repeated use in case of fibrosis
relapse and cirrhosis. Another strength of HNF4A mRNA delivery is
that despite a short half-life of delivered mRNA, it led to cyclic
expression of endogenous Hnf4a. Therefore, this may be a safer
therapeutic approach than continuous expression via viral vectors.
Additionally, using LNP-mediated mRNA delivery enables large-
scale manufacturing at low costs, within a short duration.
Another relevant factor in our study is that we achieved attenuation
of injury in fibrotic and cirrhotic mice using human HNF4A mRNA.
This suggests human HNF4A mRNA is suitable to exert antifibrotic
effects in vivo. However, it is possible that these antifibrotic effects
could be more robust when tested in patients with fibrosis. Our
in vitro experiments demonstrating restoration of lost functions in
fibrotic PHHs, and the robust antifibrotic effects in the human liver
bud after human HNF4A mRNA delivery are encouraging and
highlight the potential of our findings for future clinical translation.
HNF4A may regulate expression of multiple CYP enzymes,
thereby influencing liver damage induced by toxins such as CCl4
and DDC. We indeed observed increased expression of multiple
CYP enzymes. However, this in principle should aggravate, rather
than mitigate toxin-induced damage. Activities of CYP enzymes
are essential to generate the trichloromethyl radical leading to

=
gene in a cluster/group of cells. The color codifies mean expression of a gene in a cluster. Redder color indicates higher expression. Dendrograms show tran-
scriptomics similarity between cell clusters. (D) UMAP of control (Blue) and HNF4A (Red) groups. (E) Percentage of number of cells in hepatocyte and non-
parenchymal cell clusters in relation to total number of cells in the control (Blue) and HNF4A (Red) groups. (F) UMAPs and histogram of distribution of gene
expression of the total macrophage marker gene CD68 in the macrophage (Mac) cluster, in control and HNF4A groups. In the histogram, blue and red bars indicate
control and HNF4A groups, respectively. The ordinate shows number of cells, and the abscissa, the expression level. Mean values of each distribution are marked
with green frames. (G) Macrophage distributions of control and HNF4A along the MPI scale and (H) AMDI scale calculated by the MacSpectrum method. (I)
Macrophage subsets designated as M2-like, M1-like, transitional M1-like and pre-activation cells, on the MacSpectrum plot and (J) percentages calculated for
each subset. (K) UMAPs and histogram of relative gene expression distribution of the macrophages cluster (Mac) typical M1 marker Ccl2 and (L) M2 marker Arg1,
in control and HNF4A groups. In the histogram, blue and red bars correspond to control and HNF4A groups, respectively. The ordinate shows number of cells and
the abscissa, the level of expression. The mean values of each distribution are marked with green frames. (M-P) Tracksplots and heatmaps of expression of
hepatocyte marker genes in the hepatocytes cluster (M-N), and expression of fibrotic genes in HSCs (O-P) in control and HNF4A groups. Tracksplots show level of
expression of each gene in each single cell. Number of cells in each cluster and group are in parentheses. The range of expression of each gene is on the right of its
associated trackplot. Heatmaps show mean expression of each gene across all cells represented in the trackplots. AMDI, activation-induced macrophage dif-
ferentiation index; CCl4, carbon tetrachloride; Chol, cholangiocytes; DCs, dendritic cells; Endo, liver sinusoidal endothelial cells; Hep, hepatocytes; HSCs, hepatic
stellate cells; Kup, Kupffer cells; LNP, lipid nanoparticle; Mac, macrophages; MPI, macrophage polarization index; Neu, neutrophils; scRNA-seq, single-cell RNA
sequencing; UMAP, uniform manifold approximation and projection.

Journal of Hepatology 2021 vol. 75 j 1420–1433 1431

Research Article
subsequent liver damage. The absence of a few CYP enzymes
have been reported to protect mice against CCl4-induced fibrosis;
for example, Cyp2e1-/- mice are resistant to CCl4-induced hepa-
totoxicity.32 More recently, loss of Cyp2a5 was shown not to
affect CCl4-induced liver fibrosis.33 Similarly, reduced levels of
Cyp3a enzymes, which are required for DDC metabolism, can
suppress DDC-induced liver fibrosis.34 Therefore, it is unlikely
that mice showing amelioration of liver fibrosis upon HNF4A
mRNA treatment were protected from the toxin itself.
In summary, our findings provide the first preclinical evi-
dence that therapeutic HNF4A mRNA delivery via LNP attenuates
liver fibrosis and cirrhosis. Our study may serve as a novel
paradigm for application of mRNA-based therapeutics for the
treatment of liver fibrosis.
Abbreviations
A1AT, alpha-1 antitrypsin; AAV, adeno-associated virus; ALB,
albumin; ALT, alanine aminotransferase; CCL2, chemokine C-C
motif ligand 2; CCl4, carbon tetrachloride; DC, dendritic cells;
DDC, 3,5-diethoxycarbonyl-1,4-dihydrocollidine; HNF4A, hepa-
tocyte nuclear factor 4 alpha; HSC, hepatic stellate cell; LSECs,
liver sinusoidal endothelial cells; Mdr2, multidrug resistance
gene 2; MPI, macrophage polarization index; PHHs, primary
human hepatocytes; PMHs, primary mouse hepatocytes; PON1,
paraoxonase 1; scRNA-seq, single-cell RNA sequencing; siRNA,
small-interfering RNA; TIMP1, tissue inhibitor
of metalloproteinase 1.
Financial support
The work was supported in part by funding from CureVac, DFG;
SFB738 project C12, Lower Saxony Ministry of Science and Cul-
ture (MWK)-funded REBIRTH Center for Translational Regener-
ative Medicine, and Plus3 program of Boehringer Ingelheim to
ADS. LNP formulations used in this study were kindly provided
by Ying Tam and Paulo Lin (Acuitas Therapeutics). CF received
funding support from the DFG; SFB738 project B3. EJ received
funding from DFG, SFB738, project Z2. R.T. received funding from
transplantation center from Hannover Medical School.
Conflict of interest
MP, NH and FC are employees of CureVac AG, Tuebingen Ger-
many, a publicly listed company developing mRNA-based vac-
cines, cancer immunotherapeutics and mRNA-based protein
replacement therapies. All authors may hold shares or stock
options in the company. MP, NH and FC are inventors on several
patents on mRNA-related technology and use thereof. Other
authors declare no conflict of interest.
Please refer to the accompanying ICMJE disclosure forms for
further details.
Authors’ contributions
ADS conceived the idea, designed the study and provided the
conceptual framework for the study. TY, RK, QH, ZD and RL
performed all experiments and analyzed the data. MP, NH and FC
provided the mRNA and LNP. QYu assisted during animal ex-
periments. CF contributed to measurement of cytokines and
chemokines. F.WR V. provided primary human hepatocytes. ADS
wrote the manuscript with the help of TY, MP, MO, FC, and AB.
RT, BE and EJ provided RNA samples from MHH cohort. GS, QYa
and XS helped with the collection and analyses of HNF4A mRNA
from Zhongshan Hospital cohort. AB helped with the in silico
1432 Journal of Hepatology 2021 vol. 75 j 1420–1433
Experimental and Translational Hepatology
analyses. DG and MJA-B performed the computational analysis of
the scRNA-seq data. MP, AB, AS, AV, NH, FC, TC and MO provided
conceptual evaluation of the project. All authors commented on
and approved the manuscript.
Data availability statement
The authors declare that data supporting the findings of this
study are available within the article and its supplementary in-
formation files. The scRNA sequencing data and microarray data
generated in this study have been deposited in NCBI’s Gene
expression Omnibus and are accessible through GEO Series
accession number GSE165277.
Acknowledgments
LNP formulations used in this study were kindly provided by
Ying Tam and Paulo Lin (Acuitas Therapeutics). We thank Anto-
nia Hapke, Kerstin Beushausen and Jana Keil for excellent tech-
nical assistance.
Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.jhep.2021.08.011.
References
Author names in bold designate shared co-first authorship
[1] Alberer M, Gnad-Vogt U, Hong HS, Mehr KT, Backert L, Finak G, et al.
Safety and immunogenicity of a mRNA rabies vaccine in healthy adults: an
open-label, non-randomised, prospective, first-in-human phase 1 clinical
trial. Lancet 2017;390(10101):1511–1520.
[2] Sullenger BA, Nair S. From the RNA world to the clinic. Science
2016;352(6292):1417–1420.
[3] Abbasi J. First phase 1 trial of a prophylactic mRNA vaccine reported.
JAMA 2017;318(22):2173.
[4] Sebastian M, Schroder A, Scheel B, Hong HS, Muth A, von Boehmer L, et al.
A phase I/IIa study of the mRNA-based cancer immunotherapy CV9201 in
patients with stage IIIB/IV non-small cell lung cancer. Canc Immunol
Immunother 2019;68(5):799–812.
[5] Berraondo P, Martini PGV, Avila MA, Fontanellas A. Messenger RNA
therapy for rare genetic metabolic diseases. Gut 2019;68(7):1323–1330.
[6] Tsochatzis EA, Bosch J, Burroughs AK. Liver cirrhosis. Lancet
2014;383(9930):1749–1761.
[7] Parviz F, Matullo C, Garrison WD, Savatski L, Adamson JW, Ning G, et al.
Hepatocyte nuclear factor 4alpha controls the development of a hepatic
epithelium and liver morphogenesis. Nat Genet 2003;34(3):292–296.
[8] Berasain C, Avila MA. Regulation of hepatocyte identity and quiescence.
Cell Mol Life Sci 2015;72(20):3831–3851.
[9] Tirona RG, Lee W, Leake BF, Lan LB, Cline CB, Lamba V, et al. The orphan
nuclear receptor HNF4alpha determines PXR- and CAR-mediated xeno-
biotic induction of CYP3A4. Nat Med 2003;9(2):220–224.
[10] Nishikawa T, Bell A, Brooks JM, Setoyama K, Melis M, Han B, et al.
Resetting the transcription factor network reverses terminal chronic he-
patic failure. J Clin Invest 2015;125(4):1533–1544.
[11] Yue HY, Yin C, Hou JL, Zeng X, Chen YX, Zhong W, et al. Hepatocyte nuclear
factor 4alpha attenuates hepatic fibrosis in rats. Gut 2010;59(2):236–246.
[12] Argemi J, Latasa MU, Atkinson SR, Blokhin IO, Massey V, Gue JP, et al.
Defective HNF4alpha-dependent gene expression as a driver of hepato-
cellular failure in alcoholic hepatitis. Nat Commun 2019;10(1):3126.
[13] Santangelo L, Marchetti A, Cicchini C, Conigliaro A, Conti B, Mancone C,
et al. The stable repression of mesenchymal program is required for he-
patocyte identity: a novel role for hepatocyte nuclear factor 4alpha.
Hepatology 2011;53(6):2063–2074.
[14] Aydin Y, Kurt R, Song K, Lin D, Osman H, Youngquist B, et al. Hepatic stress
response in HCV infection promotes STAT3-mediated inhibition of
HNF4A-miR-122 feedback loop in liver fibrosis and cancer progression.
Cancers (Basel) 2019;11(10).
[15] Desai SS, Tung JC, Zhou VX, Grenert JP, Malato Y, Rezvani M, et al.
Physiological ranges of matrix rigidity modulate primary mouse
 hepatocyte function in part through hepatocyte nuclear factor 4 alpha.
Hepatology 2016;64(1):261–275.
[16] Ji D, Chen GF, Wang JC, Cao LH, Lu F, Mu XX, et al. Identification of TAF1,
HNF4A, and CALM2 as potential therapeutic target genes for liver fibrosis.
J Cell Physiol 2019;234(6):9045–9051.
[17] Ishak K, Baptista A, Bianchi L, Callea F, De Groote J, Gudat F, et al. Histo-
logical grading and staging of chronic hepatitis. J Hepatol
1995;22(6):696–699.
[18] Thess A, Grund S, Mui BL, Hope MJ, Baumhof P, Fotin-Mleczek M, et al.
Sequence-engineered mRNA without chemical nucleoside modifications
enables an effective protein therapy in large animals. Mol Ther : J Am Soc
Gene Ther 2015;23(9):1456–1464.
[19] Schlake T, Thran M, Fiedler K, Heidenreich R, Petsch B, Fotin-Mleczek M.
Messenger RNA, a novel avenue to antibody therapy? Mol Ther 2019.
[20] Thran M, Mukherjee J, Pönisch M, Fiedler K, Thess A, Mui BL, et al. mRNA
mediates passive vaccination against infectious agents, toxins, and tu-
mors. EMBO Mol Med 2017;9(10):1434–1447.
[21] Pardi N, Tuyishime S, Muramatsu H, Kariko K, Mui BL, Tam YK, et al.
Expression kinetics of nucleoside-modified mRNA delivered in lipid
nanoparticles to mice by various routes. J Controlled Release
2015;217:345–351.
[22] Lazaridis KN, LaRusso NF. Primary sclerosing cholangitis. N Engl J Med
2016;375(12):1161–1170.
[23] Smit JJ, Schinkel AH, Oude Elferink RP, Groen AK, Wagenaar E, van
Deemter L, et al. Homozygous disruption of the murine mdr2 P-glyco-
protein gene leads to a complete absence of phospholipid from bile and to
liver disease. Cell 1993;75(3):451–462.
[24] Popov Y, Patsenker E, Fickert P, Trauner M, Schuppan D. Mdr2 (Abcb4)-/-
mice spontaneously develop severe biliary fibrosis via massive dysregu-
lation of pro- and antifibrogenic genes. J Hepatol 2005;43(6):1045–1054.
Journal of Hepatology 2021 vol. 75 j 1420–1433
[25] Ferre N, Camps J, Cabre M, Paul A, Joven J. Hepatic paraoxonase activity
alterations and free radical production in rats with experimental
cirrhosis. Metabolism 2001;50(9):997–1000.
[26] Camps J, Hernandez-Aguilera A, Garcia-Heredia A, Cabre N, Luciano-
Mateo F, Arenas M, et al. Relationships between metformin, paraoxonase-
1 and the chemokine (C-C motif) ligand 2. Curr Clin Pharmacol
2016;11(4):250–258.
[27] Ehling J, Bartneck M, Wei X, Gremse F, Fech V, Mockel D, et al. CCL2-
dependent infiltrating macrophages promote angiogenesis in progres-
sive liver fibrosis. Gut 2014;63(12):1960–1971.
[28] Bala S, Csak T, Saha B, Zatsiorsky J, Kodys K, Catalano D, et al. The pro-
inflammatory effects of miR-155 promote liver fibrosis and alcohol-
induced steatohepatitis. J Hepatol 2016;64(6):1378–1387.
[29] Li C, Menoret A, Farragher C, Ouyang Z, Bonin C, Holvoet P, et al. Single cell
transcriptomics based-MacSpectrum reveals novel macrophage activa-
tion signatures in diseases. JCI Insight 2019;5.
[30] Guillot A, Tacke F. Liver macrophages: old dogmas and new insights.
Hepatol Commun 2019;3(6):730–743.
[31] Ramachandran P, Dobie R, Wilson-Kanamori JR, Dora EF, Henderson BEP,
Luu NT, et al. Resolving the fibrotic niche of human liver cirrhosis at
single-cell level. Nature 2019;575(7783):512–518.
[32] Wong FW, Chan WY, Lee SS. Resistance to carbon tetrachloride-induced
hepatotoxicity in mice which lack CYP2E1 expression. Toxicol Appl
Pharmacol 1998;153(1):109–118.
[33] Hong F, Si C, Gao P, Cederbaum AI, Xiong H, Lu Y. The role of CYP2A5 in
liver injury and fibrosis: chemical-specific difference. Naunyn Schmie-
debergs Arch Pharmacol 2016;389(1):33–43.
[34] Nantasanti S, Toussaint MJ, Youssef SA, Tooten PC, de Bruin A. Rb and p53
liver functions are essential for xenobiotic metabolism and tumor sup-
pression. PLoS One 2016;11(3):e0150064.
1433