Professional Documents
Culture Documents
Therapeutic HNF4A mRNA Attenuates Liver Fibrosis in A Preclinical Model
Therapeutic HNF4A mRNA Attenuates Liver Fibrosis in A Preclinical Model
Therapeutic HNF4A mRNA Attenuates Liver Fibrosis in A Preclinical Model
Background & Aims: Therapeutic targeting of injuries that paraoxonase 1 is a direct target of HNF4A and it contributes to
require transient restoration of proteins by mRNA delivery is an HNF4A-mediated attenuation of liver fibrosis via modulation of
attractive approach that, until recently, has remained poorly liver macrophages and hepatic stellate cells.
explored. In this study, we examined the therapeutic utility of Conclusion: Collectively, our findings provide the first direct
mRNA delivery for liver fibrosis and cirrhosis. Specifically, we preclinical evidence of the applicability of HNF4A mRNA thera-
aimed to demonstrate the therapeutic efficacy of human hepa- peutics for the treatment of fibrosis in the liver.
tocyte nuclear factor alpha (HNF4A) mRNA in mouse models of Lay summary: Liver fibrosis and cirrhosis remain unmet medical
fibrosis and cirrhosis. needs and contribute to high mortality worldwide. Herein, we
Methods: We investigated restoration of hepatocyte functions by take advantage of a promising therapeutic approach to treat liver
HNF4A mRNA transfection in vitro, and analyzed the attenuation fibrosis and cirrhosis. We demonstrate that restoration of a key
of liver fibrosis and cirrhosis in multiple mouse models, by gene, HNF4A, via mRNA encapsulated in lipid nanoparticles
delivering hepatocyte-targeted biodegradable lipid nanoparticles decreased injury in multiple mouse models of fibrosis and
(LNPs) encapsulating HNF4A mRNA. To identify potential mech- cirrhosis. Our study provides proof-of-concept that mRNA ther-
anisms of action, we performed microarray-based gene expres- apy is a promising strategy for reversing liver fibrosis
sion profiling, single-cell RNA sequencing, and chromatin and cirrhosis.
immunoprecipitation. We used primary liver cells and human © 2021 The Author(s). Published by Elsevier B.V. on behalf of Euro-
liver buds for additional functional validation. pean Association for the Study of the Liver. This is an open access
Results: Expression of HNF4A mRNA led to restoration of the article under the CC BY-NC-ND license (http://creativecommons.org/
metabolic activity of fibrotic primary murine and human hepa- licenses/by-nc-nd/4.0/).
tocytes in vitro. Repeated in vivo delivery of LNP-encapsulated
HNF4A mRNA induced a robust inhibition of fibrogenesis in 4 Introduction
independent mouse models of hepatotoxin- and cholestasis- Transient protein expression by transfer of therapeutic mRNA
induced liver fibrosis. Mechanistically, we discovered that considerably limits the risks encountered in viral gene therapy
such as insertional mutagenesis, severe immune responses and
Keywords: mRNA therapeutics; Transcription factors; protein replacement potential activation of oncogenes. Tremendous progress has
and cirrhosis. already been made with mRNA technology being applied to
Received 26 January 2021; received in revised form 30 July 2021; accepted 2 August vaccine development1–3 and immuno-oncology.4 Lipid nano-
2021; available online 25 August 2021
particle (LNP)-formulated mRNAs are shielded from humoral and
* Corresponding author. Address: Department of Gastroenterology, Hepatology and
Endocrinology, REBIRTH-Research Center for Translational Regenerative Medicine, cellular immunity and can be produced in large quantities at
Hannover Medical School, Carl-Neuberg Str. 1, 30625 Hannover, Germany; Tel.: 0049 lower cost compared to viral vectors.5 mRNA therapeutics are
511 532 5255, Fax: 0049 511 532 5234.
E-mail address: sharma.amar@mh-hannover.de (A.D. Sharma). particularly well-suited to treat injuries not requiring constant
https://doi.org/10.1016/j.jhep.2021.08.011 and life-long expression of therapeutic proteins, and offer the
*** **
**
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
***
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
m
en
en
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
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
** *
μg/24 h/100,000 cells
* ** ** * *
** 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
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-
1016
1015
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
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
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, (Fig. S3A). Measurement of different cytokines and chemokines
whether LNP-formulated mRNA reaches hepatocytes of fibrotic indicated that mRNA doses of 1 mg/kg and 2 mg/kg did not elicit
livers efficiently has remained unknown. LNP carrying mRNA substantial immune responses, whereas 4 mg/kg mRNA induced
encoding for Photinus pyralis luciferase (Luc/LNP) or ZsGreen a significant immune response (Fig. S3B). Therefore, we used 2
(ZsGreen/LNP) were injected into CCl4-induced fibrotic and non- mg/kg mRNA/LNP for subsequent experiments.
fibrotic control (wild-type) mice i.v. (Fig. 2A). In vivo imaging We next examined whether repeated i.v. administration of 2
revealed robust bioluminescence exclusively in the liver from 8 mg/kg mRNA/LNP was safe in fibrotic mice (Fig. S4A). Repeated
hours to 120 hours after Luc/LNP injection in both groups injections of 2 mg/kg LNP encapsulating HNF4A mRNA (hence-
(Fig. 2B). Co-localization studies showed robust ZsGreen protein forth referred to as HNF4A/LNP) every 5 days induced no sig-
expression specifically in hepatocytes of fibrotic and control nificant elevations of most cytokines and chemokines (Fig. S4B).
mice, but not in other hepatic cells (Fig. 2C). Our results Thus, administration of either single or multiple doses of 2 mg/
demonstrated that LNP-formulated mRNA technology could kg HNF4A/LNP is well tolerated.
specifically and efficiently target hepatocytes in murine
fibrotic livers. Delivery of human HNF4A mRNA attenuates toxin- and
Successful mRNA delivery should be therapeutic but with cholestasis-induced fibrosis
negligible immune response. Therefore, we first monitored im- To address whether systemic administration of HNF4A mRNA
mune response after injecting different doses of ZsGreen/LNP 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.
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
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
μg/mg
600
μg/mg
0.15
IU/L
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
****
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
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.
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
os
rrh
Ci
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
** **
**
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.
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
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
*
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
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.
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
BS2
0.8 0.4 0.8 BS3
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
** 2,000 1.6
Relative Pon 1 activity
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.
A B Paraoxonase 1
C
ZsGreen mRNA CCL2
2.5 ** 1,000
48 h ** *** * **
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
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
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 (%)
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
N
H
* **
M1 M2 *
* ***
1.5 4 ** * 1.0 1.0 1.0
** **
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
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.
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
=
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.
subsequent liver damage. The absence of a few CYP enzymes analyses. DG and MJA-B performed the computational analysis of
have been reported to protect mice against CCl4-induced fibrosis; the scRNA-seq data. MP, AB, AS, AV, NH, FC, TC and MO provided
for example, Cyp2e1-/- mice are resistant to CCl4-induced hepa- conceptual evaluation of the project. All authors commented on
totoxicity.32 More recently, loss of Cyp2a5 was shown not to and approved the manuscript.
affect CCl4-induced liver fibrosis.33 Similarly, reduced levels of
Cyp3a enzymes, which are required for DDC metabolism, can Data availability statement
suppress DDC-induced liver fibrosis.34 Therefore, it is unlikely The authors declare that data supporting the findings of this
that mice showing amelioration of liver fibrosis upon HNF4A study are available within the article and its supplementary in-
mRNA treatment were protected from the toxin itself. formation files. The scRNA sequencing data and microarray data
In summary, our findings provide the first preclinical evi- generated in this study have been deposited in NCBI’s Gene
dence that therapeutic HNF4A mRNA delivery via LNP attenuates expression Omnibus and are accessible through GEO Series
liver fibrosis and cirrhosis. Our study may serve as a novel accession number GSE165277.
paradigm for application of mRNA-based therapeutics for the
treatment of liver fibrosis.
Acknowledgments
LNP formulations used in this study were kindly provided by
Abbreviations
Ying Tam and Paulo Lin (Acuitas Therapeutics). We thank Anto-
A1AT, alpha-1 antitrypsin; AAV, adeno-associated virus; ALB,
nia Hapke, Kerstin Beushausen and Jana Keil for excellent tech-
albumin; ALT, alanine aminotransferase; CCL2, chemokine C-C
nical assistance.
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, Supplementary data
liver sinusoidal endothelial cells; Mdr2, multidrug resistance Supplementary data to this article can be found online at https://
gene 2; MPI, macrophage polarization index; PHHs, primary doi.org/10.1016/j.jhep.2021.08.011.
human hepatocytes; PMHs, primary mouse hepatocytes; PON1,
paraoxonase 1; scRNA-seq, single-cell RNA sequencing; siRNA, References
small-interfering RNA; TIMP1, tissue inhibitor Author names in bold designate shared co-first authorship
of metalloproteinase 1.
[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
Financial support open-label, non-randomised, prospective, first-in-human phase 1 clinical
The work was supported in part by funding from CureVac, DFG; trial. Lancet 2017;390(10101):1511–1520.
SFB738 project C12, Lower Saxony Ministry of Science and Cul- [2] Sullenger BA, Nair S. From the RNA world to the clinic. Science
2016;352(6292):1417–1420.
ture (MWK)-funded REBIRTH Center for Translational Regener- [3] Abbasi J. First phase 1 trial of a prophylactic mRNA vaccine reported.
ative Medicine, and Plus3 program of Boehringer Ingelheim to JAMA 2017;318(22):2173.
ADS. LNP formulations used in this study were kindly provided [4] Sebastian M, Schroder A, Scheel B, Hong HS, Muth A, von Boehmer L, et al.
by Ying Tam and Paulo Lin (Acuitas Therapeutics). CF received 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
funding support from the DFG; SFB738 project B3. EJ received
Immunother 2019;68(5):799–812.
funding from DFG, SFB738, project Z2. R.T. received funding from [5] Berraondo P, Martini PGV, Avila MA, Fontanellas A. Messenger RNA
transplantation center from Hannover Medical School. therapy for rare genetic metabolic diseases. Gut 2019;68(7):1323–1330.
[6] Tsochatzis EA, Bosch J, Burroughs AK. Liver cirrhosis. Lancet
Conflict of interest 2014;383(9930):1749–1761.
[7] Parviz F, Matullo C, Garrison WD, Savatski L, Adamson JW, Ning G, et al.
MP, NH and FC are employees of CureVac AG, Tuebingen Ger- Hepatocyte nuclear factor 4alpha controls the development of a hepatic
many, a publicly listed company developing mRNA-based vac- epithelium and liver morphogenesis. Nat Genet 2003;34(3):292–296.
cines, cancer immunotherapeutics and mRNA-based protein [8] Berasain C, Avila MA. Regulation of hepatocyte identity and quiescence.
replacement therapies. All authors may hold shares or stock 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
options in the company. MP, NH and FC are inventors on several
nuclear receptor HNF4alpha determines PXR- and CAR-mediated xeno-
patents on mRNA-related technology and use thereof. Other biotic induction of CYP3A4. Nat Med 2003;9(2):220–224.
authors declare no conflict of interest. [10] Nishikawa T, Bell A, Brooks JM, Setoyama K, Melis M, Han B, et al.
Please refer to the accompanying ICMJE disclosure forms for Resetting the transcription factor network reverses terminal chronic he-
further details. 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.
Authors’ contributions [12] Argemi J, Latasa MU, Atkinson SR, Blokhin IO, Massey V, Gue JP, et al.
ADS conceived the idea, designed the study and provided the Defective HNF4alpha-dependent gene expression as a driver of hepato-
conceptual framework for the study. TY, RK, QH, ZD and RL cellular failure in alcoholic hepatitis. Nat Commun 2019;10(1):3126.
[13] Santangelo L, Marchetti A, Cicchini C, Conigliaro A, Conti B, Mancone C,
performed all experiments and analyzed the data. MP, NH and FC
et al. The stable repression of mesenchymal program is required for he-
provided the mRNA and LNP. QYu assisted during animal ex- patocyte identity: a novel role for hepatocyte nuclear factor 4alpha.
periments. CF contributed to measurement of cytokines and Hepatology 2011;53(6):2063–2074.
chemokines. F.WR V. provided primary human hepatocytes. ADS [14] Aydin Y, Kurt R, Song K, Lin D, Osman H, Youngquist B, et al. Hepatic stress
wrote the manuscript with the help of TY, MP, MO, FC, and AB. response in HCV infection promotes STAT3-mediated inhibition of
HNF4A-miR-122 feedback loop in liver fibrosis and cancer progression.
RT, BE and EJ provided RNA samples from MHH cohort. GS, QYa Cancers (Basel) 2019;11(10).
and XS helped with the collection and analyses of HNF4A mRNA [15] Desai SS, Tung JC, Zhou VX, Grenert JP, Malato Y, Rezvani M, et al.
from Zhongshan Hospital cohort. AB helped with the in silico Physiological ranges of matrix rigidity modulate primary mouse