Role of high-mobility group box 1 protein in
post-infarction healing process and left ventricular
Takashi Kohno1, Toshihisa Anzai1*, Kotaro Naito1, Taku Miyasho2, Minoru Okamoto2,
Hiroshi Yokota2, Shingo Yamada3, Yuichiro Maekawa1, Toshiyuki Takahashi1,
Tsutomu Yoshikawa1, Akitoshi Ishizaka1, and Satoshi Ogawa1
1Cardiopulmonary Division, Department of Medicine, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku,
Tokyo 160-8582, Japan;2Department of Veterinary Medicine, School of Veterinary Medicine, Rakuno Gakuen University,
Ebetsu, Japan; and3Central Institute, Shino-Test Corporation, Sagamihara, Japan
Received 27 May 2008; revised 30 September 2008; accepted 24 October 2008; online publish-ahead-of-print 3 November 2008
Time for primary review: 34 days
Aims High-mobility group box 1 protein (HMGB1) is one of the recently defined damage-associated mol-
ecular pattern molecules derived from necrotic cells and activated macrophages. We investigated clini-
cal implications of serum HMGB1 elevation in patients with acute myocardial infarction (MI). Then, we
evaluated the effect of HMGB1 blockade on post-MI left ventricular (LV) remodelling in a rat MI model.
Methods and results Serum HMGB1 levels were examined in patients with ST-elevation MI (n ¼ 35). A
higher peak serum HMGB1 level was associated with pump failure, cardiac rupture, and in-hospital
cardiac death. Then, an experimental MI model was induced in male Wistar rats. The mRNA and
protein expression of HMGB1 were increased in the infarcted area compared with those values observed
in sham-operated rats. We administered neutralizing anti-HMGB1 antibody (MI/anti-H) or control
antibody (MI/C) to MI rats subcutaneously for 7 days. The mRNA levels of tumour necrosis factor-a
and interleukin-1b and the number of macrophages in the infarcted area were reduced on day 3 in
MI/anti-H rats compared with MI/C rats. Interestingly, HMGB1 blockade resulted in thinning and expan-
sion of the infarct scar and marked hypertrophy of the non-infarcted area on day 14.
Conclusion Elevated serum HMGB1 levels were associated with adverse clinical outcomes in patients
with MI. However, HMGB1 blockade in a rat MI model aggravated LV remodelling, possibly through
impairment of the infarct-healing process. HMGB1, a novel predictor of adverse clinical outcomes
after MI, may have an essential role in the appropriate healing process after MI.
Left ventricular (LV) remodelling after myocardial infarction
(MI) is the process of infarct expansion followed by progress-
ive LV dilation and is associated with adverse clinical out-
comes.1Inflammatory response and cytokine elaboration
are integral components of the host response to tissue
injury and play a particularly active role after MI.2–5A well-
orchestrated inflammatory response after MI leads to an
appropriate infarct-healing process and the formation of
a scar with tensile strength, resulting in prevention of
infarct expansion. Despite the importance of the inflamma-
tory response and healing process in post-MI LV remodelling,
the mechanisms that initiate and control these processes
remain to be elucidated.
High-mobility group box 1 protein (HMGB1) was originally
identified as a non-histone DNA-binding nuclear protein pro-
duced by nearly all cell types; it stabilizes nucleosomes and
Recently, it was clarified that HMGB1 is one of the
damage-associated molecular pattern molecules and is
released passively as an endogenous danger signal from necro-
tic, but not apoptotic cells.7,8Extracellular HMGB1 exhibits
inflammatory cytokine-like activity and acts as a potent
mediator of macrophage activation.7,9HMGB1 is also secreted
extracellularly by activated macrophages in response to
pro-inflammatory cytokines.7,10Therefore, HMGB1 has a
unique ability to self-amplify and prolong inflammatory
response and contributes to the pathogenesis of sepsis and
acute lung injury.7,9–12We hypothesized that HMGB1 might be
involved in post-MI inflammatory response and LVremodelling.
Previous clinical and experimental studies have demon-
strated that the use of anti-inflammatory agents after MI
*Corresponding author. Tel: þ81 3 5363 3793; fax: þ81 3 3353 2502.
E-mail address: firstname.lastname@example.org
Published on behalf of the European Society of Cardiology. All rights reserved. & The Author 2008.
For permissions please email: email@example.com.
Cardiovascular Research (2009) 81, 565–573
by guest on December 25, 2015
leads to infarct expansion through the impairment of the
healing process,13,14suggesting that the appropriate inflam-
matory response might be a prerequisite for the prevention
of post-MI LV remodelling. On the other hand, an excessive
and persistent inflammatory response could cause the
myocardium to become vulnerable through additional
recruitment of inflammatory cells and degradation of
myocardium might be susceptible to wall stress, resulting
in infarct expansion. Therefore, it is unclear whether
the attenuation of the post-MI inflammatory response by
HMGB1 blockade could be beneficial or harmful to the
infarct-healing process and post-MI LV remodelling.
This translational study examined the involvement of
HMGB1 in the pathogenesis of post-MI LV remodelling. We
first examined the clinical significance of serum HMGB1
elevation in patients with MI. We then evaluated the
expression level and distribution of HMGB1 in the infarcted
myocardium and assessed whether blockade of HMGB1
could modulate LV remodelling in a rat MI model.
This clinical investigation was approved by the institutional medical
Ethics Committee and conducted according to the ethical guidelines
outlined in the Declaration of Helsinki. A total of 49 consecutive
patients with first ST-elevation MI (STEMI) were examined. All
patients were admitted to Keio University Hospital between
January 2006 and December 2006. A diagnosis of STEMI was made
on the basis of chest pain lasting ?30 min, the presence of new
ST-segment elevation (at least 1 mV in two or more standard leads
or at least 0.2 mV in two or more contiguous precordial leads),
and an increase in a biochemical marker of myocardial necrosis
[creatine kinase (CK)-MB fraction or troponin T]. We excluded
patients in whom the time elapsed from onset to admission was
.24 h (six patients) and those who died before the determination
of the peak HMGB1 level (three patients). Patients with collagen
disease, advanced liver disease, renal failure, malignancy, or any
infectious disease were also excluded (five patients). Finally, 35
patients were included in this study. For comparison, blood
samples were also collected from 35 patients who were admitted
to the same institute with chronic stable angina (CSA). CSA was
defined as the presence of known coronary atherosclerosis with
typical exertional chest pain relieved by rest and/or nitrates, and
without a change in the frequency or pattern for 3 months before
2.2 Clinical study protocol
We assessed the clinical parameters listed in Table 1. All patients
with revascularization therapy received percutaneous coronary
intervention on admission. We measured the serum HMGB1 level
on admission in patients with MI and CSA, and then 6, 12, 18, 24,
72 h and 7 days after admission in patients with MI. Serum HMGB1
concentration was measured by enzyme-linked immunosorbent
assay (ELISA) (Shino-Test Corporation, Sagamihara, Japan).17We
determined cumulative CK release, which was defined as the area
under the time-activity curve for CK concentration over the first
24 h.18We also measured the serum C-reactive protein level on
admission, and then every 24 h for at least 4 days to determine its
peak value.19The incidence of in-hospital complications, including
pump failure (class II or greater of Killip’s classification or subset II
or greater of Forrester’s classification), cardiac rupture, and
cardiac death was examined. Echocardiography was performed
10–14 days after MI (Sonos 5500, Phillips Medical Systems,
Andover, MA, USA), and LV dimensions and fractional shortening
Myocardial infarcted patient characteristics and peak high-mobility group box 1 protein level
Peak HMGB1 level (ng/mL)
Age ?70 years
Arrival ,6 h
Medication before admission
Medication after admission
Peak C-reactive protein level ?5.5 mg/dL
Data are shown as mean+SD (number of patients).
HMGB1, high-mobility group box 1 protein; ACE, angiotensin-converting enzyme; ARB, angiotensin type-1 receptor blocker.
T. Kohno et al.
by guest on December 25, 2015
were determined. We measured plasma brain natriuretic peptide
(BNP) level 6 months after MI by immunoradiometric assay.20
2.3 Rat model of experimental myocardial
Male Wistar rats, pathogen-free and weighing 200–250 g, were
subjected to left coronary artery ligation or sham operation
under pentobarbital anaesthesia (intraperitoneal pentobarbital
30 mg/kg).21,22To determine serial HMGB1 expression, rats were
sacrificed 1, 3, 7, and 14 days after MI for RNA and protein analyses
(n ¼ 6 rats per time point) and for pathological analyses (n ¼ 6 rats
per time point). Serum HMGB1 level was measured using ELISA to
determine serial changes of serum HMGB1 level. The LV of
sham-operated rats (n ¼ 6 for RNA and protein analyses, n ¼ 6 for
pathological analyses, on day 14) was collected as control. In RNA
and protein analyses, LV tissue with MI was carefully divided into
infarcted (infarcted zone and 1–2 mm of the border zone) and non-
infarcted areas. These tissues were snap-frozen in liquid nitrogen
and then preserved at 2808C. For pathological analyses, hearts
were arrested by infusion of ice-cold saturated potassium chloride,
excised, and placed in ice-cold potassium chloride to achieve
uniform diastolic arrest. The LV was cannulated retrogradely via
ascending aorta, and the hearts were fixed with 4% paraformalde-
hyde at a constant intraventricular pressure. Fixed hearts with 4%
paraformaldehyde were embedded in paraffin. In the HMGB1 block-
ade study, MI rats surviving the operation for 24 h (n ¼ 80) were ran-
domly assigned to two groups: (i) neutralizing polyclonal chicken IgY
anti-HMGB1 antibody (10 mg/kg/day, donated by Shino-Test Corpor-
ation)11,12,23administered subcutaneously for 7 days (MI/anti-H,
n ¼ 40) and (ii) control chicken IgY antibody (MI/C, n ¼ 40).
To prepare neutralizing anti-HMGB1 antibody, IgY class antibody
from the egg yolk of HMGB1-immunized hens was isolated and pur-
ified.24Control IgY antibody was purified from non-immunized egg
yolk. The dosage of neutralizing anti-HMGB1 antibody was deter-
mined according to the previous study.12Rats were sacrificed 3, 7,
and 14 days after MI for mRNA expression (n ¼ 6 rats per time
point) and pathological analyses (n ¼ 6 rats per time point).
Echocardiographic (8.5 MHz linear transducer; EnVisor C, Philips
Medical Systems) and haemodynamic studies (SPC-320, Millar Instru-
ments, Houston, TX, USA) were performed 14 days after MI.21,22All
procedures were performed in accordance with the Keio University
animal care guidelines, which conformed with the Guide for the
Care and Use of Laboratory Animals published by the US National
Institutes of Health (NIH Publication No. 85-23, revised 1996).
2.4 Real-time quantitative reverse transcriptase–
polymerase chain reaction
Total RNA was isolated by acid–phenol extraction in the presence
of chaotropic salts (Trizol, Invitrogen, Carlsbad, CA, USA) and
Real-time quantitative reverse transcriptase–polymerase chain
reaction (RT–PCR) of each sample was carried out with a TaqMan
RNA PCR kit and ABI PrismTM7700 Sequence Detection System
3-phosphate dehydrogenase (GAPDH) was used for normalization.
HMGB1, interleukin-1b (IL-1b), atrial natriuretic factor (ANF), and
GAPDH were purchased as pre-optimized kits [TaqMan probe;
Applied Biosystems: catalogue number Rn02377062_g1 (HMGB1),
Rn00580432_m1 (IL-1b), Rn00561661_m1 (ANF), Rn99999916_s1
(GAPDH)]. The primer pair and probe for tumour necrosis factor-a
(TNF-a) were as follows: forward primer: TGGGCTCCCTCTCATCAGTT;
reverse primer: TGGGCTACGGGCTTGTCA; TaqMan probe: TGGCCC
2.5 Western blotting
Frozen tissue was homogenized in cell lysis buffer (Cell Signaling,
Danvers, MA, USA) containing 1% Triton X-100 and protease inhibi-
tors. After centrifugation at 16 000 g for 30 min at 48C, the super-
polyclonal anti-rat HMGB1 (Pharmingen, San Diego, CA, USA) was
carried out as described previously.22After probing with anti-rat
HMGB1 antibody, membranes were stripped of bound immunoglobu-
lins and reprobed with anti-rat GAPDH (Santa Cruz Biotechnology,
Santa Cruz, CA, USA) to correct for protein loading.
2.6 Histopathological study
Tissues were stained with haematoxylin–eosin. The boundary
lengths of the infarcted and non-infarcted endocardial and epicar-
dial surfaces of all slices were traced using Image J software
(version 1.38x, National Institutes of Health), and the infarct size
was determined as the percentage of infarcted epicardium and
endocardium of the LV.25The area of the LV cavity and total LV,
and the thickness of the septum and scar in sections at the papillary
muscle level were also measured, and then the expansion index was
calculated as described previously: expansion index ¼ (LV cavity
area/totalLV area) ? (non-infarcted
2.7 Immunohistochemical study
Immunohistochemical studies were performed by immunoperoxi-
dase methods.12,21Paraffin section-embedded specimens were cut
into 4 mm-thick sections and stained with antibodies against
HMGB1 (Shino-Test Corporation) and ED-1 (for monocyte-derived
macrophages; Serotec Ltd, Kidlington, Oxford, UK). The number of
macrophages was quantified by counting the total number of posi-
tively stained ED-1 cells in 20 grid fields with a total area of
2.8 Statistical analyses
All continuous data were expressed as the mean value+SD unless
otherwise stated. Comparison between groups was performed
using unpaired t-test or non-parametric analysis for continuous vari-
ables and x2test or Fisher’s exact test for categorical variables.
Statistical significance was defined as a P-value of ,0.05. All stat-
istical analyses were performed using SPSS 13.0 for Windows (SPSS
Inc., Chicago, IL, USA).
3.1 Serum high-mobility group box 1 protein level
in patients with ST-elevation myocardial infarction
and chronic stable angina
Figure 1 shows serum HMGB1 level in patients with CSA and
serial changes in serum HMGB1 level in patients with STEMI.
Serum HMGB1 level in patients with STEMI on admission was
higher than that in patients with CSA (4.5+5.3 vs. 1.0+
1.6 ng/mL, P ¼ 0.0004). Serum HMGB1 level significantly
increased, peaking at 12 h after MI, compared with the base-
line level. On day 7, serum HMGB1 level remained elevated
in patients with STEMI compared with those with CSA (3.3+
5.1 vs. 1.0+1.6 ng/mL, P ¼ 0.011). There were no signifi-
cant differences in age (65+13 vs. 64+13 years old, P ¼
smoking (49 vs. 43%, P ¼ 0.63), hypertension (54 vs. 51%,
P ¼ 0.81), hypercholesterolaemia (63 vs. 60%, P ¼ 0.81),
and diabetes mellitus (26 vs. 17%, P ¼ 0.38) between
patients with STEMI and CSA. Among the patients with
P ¼ 0.55),
Role of HMGB1 in post-MI LV remodelling567
by guest on December 25, 2015
STEMI, peak HMGB1 level did not significantly differ accord-
ing to patient characteristics (Table 1).
3.2 Relationships among peak C-reactive protein
level, cumulative creatine kinase release, and peak
high-mobility group box 1 protein level in patients
with ST-elevation myocardial infarction
The mean peak C-reactive protein level was 7.2+5.2 mg/dL
(0.7–23.2 mg/dL, median 5.5 mg/dL) and cumulative CK
release was 41 205+26 117 h IU/L (7347–102 893 h IU/L,
median 37 070 h IU/L) in patients with STEMI. The mean
elapsed time from the onset of MI to the C-reactive
protein peak was 3+1 days. The peak C-reactive protein
level (r ¼ 0.49, P ¼ 0.003), but not cumulative CK release
(r ¼ 0.25, P ¼ 0.15), showed a significant positive corre-
lation with the peak HMGB1 level. Patients with peak
C-reactive protein level greater than or equal to the
median value had a higher peak HMGB1 level than those
without (Table 1).
3.3 Clinical significance of serum high-mobility
group box 1 protein elevation in patients with
ST-elevation myocardial infarction
Peak HMGB1 level was higher in patients with pump failure
(P ¼ 0.0004), cardiac rupture (P ¼ 0.023), and in-hospital
cardiac death (P ¼ 0.013) compared with those without
(Table 2). There was no statistically significant correlation
between peak serum HMGB1 level and LV dimensions (LV
end-diastolic dimension, r ¼ 20.47, P ¼ 0.81; LV end-
systolicdimension,r ¼ 0.11,P ¼ 0.54)orfractionalshortening
(r ¼ 20.22, P ¼ 0.24) 2 weeks after MI. The mean plasma
BNP level 6 months after MI was 85.9+104.5 pg/mL
(6.2–518.0 pg/mL, median
HMGB1 level was positively correlated with plasma BNP
level 6 months after the onset (r ¼ 0.44, P ¼ 0.016).
Patients with plasma BNP level greater than or equal to
than those without (24.3+23.3 vs. 10.2+6.8 ng/mL,
P ¼ 0.037).
48.9 pg/mL).Peak serum
3.4 Serum and myocardial high-mobility group box
1 protein levels in experimental myocardial
Serum HMGB1 level was significantly increased in MI rats
compared with sham-operated rats during the observation
period (Figure 2A). The mRNA expression of HMGB1 in the
infarcted area was elevated on day 3, peaked on day 7 up
to ?5.4-fold compared with that in sham-operated rats,
and remained elevated on day 14 after MI (Figure 2B). The
mRNA expression of HMGB1 in the non-infarcted area of MI
rats was not elevated compared with that in sham-operated
rats during the observation period (day 1; 110+14%, day 3;
105+6%, day 7; 102+18%, day 14; 100+10% mRNA
expression vs. sham). Western blotting showed that the
level of HMGB1 protein in the infarcted area was increased
on day 3, peaked on day 7, and remained elevated on day
14 after MI (Figure 2C).
HMGB1-positive nuclear staining of myocytes was present
in the sham-operated rat myocardium. In MI rats, immuno-
histochemical staining of HMGB1 was enhanced in the
infarcted and border zone myocardium (Figure 3A–H). At
24 h after MI induction, HMGB1 expression was detected in
the cytoplasm of degenerated cardiomyocytes (Figure 3I)
and extracellular fields (Figure 3J) in the infarcted area.
HMGB1 immunoreactivity was also seen in infiltrating inflam-
matory cells 24 h after the induction of MI (Figure 3K).
During the observation period, HMGB1 staining was most
intense on day 3 and was expressed mainly in infiltrating
inflammatory cells (Figure 3L). On day 7, HMGB1 staining
extended over a larger area but weaker labelling was
visible. The number of inflammatory cells with HMGB1
immunostaining was decreased on day 14, and HMGB1 immu-
nostaining was mainly present in fibroblasts (Figure 3M).
3.5 Effect of high-mobility group box 1 protein
blockade in experimental myocardial infarction
To determine whether HMGB1 is involved in post-MI inflam-
matory response and LV remodelling, we treated MI rats
with neutralizing anti-HMGB1 antibody or control antibody.
Neutralizing anti-HMGB1 antibody treatment diminished
MI-induced mRNA expression of TNF-a and IL-1b in the
infarcted area on day 3 (Figure 4A and B). HMGB1 mRNA
expression in the infarcted area was significantly lower in
MI/anti-H than in MI/C on days 3, 7, and 14 (Figure 4C). In
the non-infarcted area, the expression of TNF-a, IL-1b,
and HMGB1 was not different between MI/C and MI/anti-H
during the observation period (Figure 4D–F). The increased
Clinical outcomes and peak high-mobility group box 1
Peak HMGB1 level (ng/mL)
Data are shown as mean+SD (number of patients).
HMGB1, high-mobility group box 1 protein.
level in patients with ST-elevation myocardial infarction (STEMI) compared
with those with chronic stable angina (CSA). Data are mean+SEM.
*P , 0.05 vs. patients with chronic stable angina,
level (on admission).
Serial changes in serum high-mobility group box 1 protein (HMGB1)
†P , 0.05 vs. baseline
T. Kohno et al.
by guest on December 25, 2015
number of ED-1-positive macrophages in the infarcted area
on day 3 was attenuated by the blockade of HMGB1
As shown in Table 3, haemodynamic measurement showed
LV systolic pressure, maximum rate of isovolumic pressure
development (þdP/dtmax), and minimum rate of isovolumic
pressure decay (2dP/dtmin) to be lower and LV end-diastolic
sham-operated rats. Echocardiographic findings indicated
that LV dimensions increased while LV fractional shortening
decreased in the MI/C rats compared with those in
sham-operated rats. No differences existed in left and
theMI/C rats thanin
right ventricular weight per body weight, LV systolic
pressure, and heart rate on day 14 between MI/C and MI/
anti-H. Rats in MI/anti-H exhibited significantly lower LV
fractional shortening and higher LV dimensions compared
with those in MI/C on day 14. LV þdP/dtmax and 2dP/
dtminwere lower and LVEDP was higher in MI/anti-H than
in MI/C on day 14.
LV remodelling was also assessed by LV histomorphometric
analysis. There was no difference in infarct size between MI/
C and MI/anti-H on days 7 (52+2 vs. 53+4%, P ¼ 0.83) and
14 (50+3 vs. 50+2%, P ¼ 0.99). Blockade of HMGB1 exag-
gerated thinning of the infarcted LV wall on days 7 and 14.
serum high-mobility group box 1 protein level by enzyme-linked immunosorbent assay. (B) Quantified data of high-mobility group box 1 protein mRNA expression
in the infarcted area by real-time quantitative reverse transcriptase–polymerase chain reaction. (C) Quantified data of high-mobility group box 1 protein
expression in the infarcted area by western blotting. Data are mean+SEM. *P , 0.05 vs. sham-operated rats.
High-mobility group box 1 protein (HMGB1) level in serum and infarcted myocardium in rat myocardial infarction (MI) model. (A) Quantified data of
box 1 protein immunohistochemical staining in the infarcted and border zone areas (A–D). The parts of the border zone area indicated by boxes are shown at a
higher magnification in panels (E–H). In the infarcted area, high-mobility group box 1 protein was seen clearly in degenerated myocytes (I), extracellular fields
(black arrows) (J), and inflammatory cells (K) on day 1. High-mobility group box 1 protein was mainly expressed in inflammatory cells on day 3 (L), and in
fibroblasts on day 14 (M). Scale bar indicates 500 mm (A–H) and 100 mm (I–M).
High-mobility group box 1 protein (HMGB1) immunostaining in the infarcted area. Representative low-power micrographs showing high-mobility group
Role of HMGB1 in post-MI LV remodelling 569
by guest on December 25, 2015
Myocardial hypertrophy of the non-infarcted LV wall was
promoted by HMGB1 blockade on day 14 (Figure 6A–D).
The expansion index was significantly higher in MI/anti-H
than in MI/C on days 7 and 14 (Figure 6E).
In parallel with the non-infarcted myocardial hypertrophy,
ANF expression in the non-infarcted area significantly
increased in MI/anti-H compared with MI/C on day 14
We demonstrated that patients with STEMI had a markedly
increased serum HMGB1 level, peaking at 12 h after MI,
and that a higher peak serum HMGB1 level was associated
with pump failure, cardiac rupture, and in-hospital cardiac
death. In an experimental study using a rat MI model, we
found increased myocardial HMGB1 expression in the
infarcted area, which persisted for 14 days after MI. Block-
ade of HMGB1 by a neutralizing anti-HMGB1 antibody atte-
nuated the post-MI inflammatory response; however, it
resulted in marked infarct scar thinning and exaggerated
LV remodelling. These findings suggested that HMGB1 could
be a predictor of adverse clinical outcomes after MI, but
also be a prerequisite for an appropriate healing process
and for preserving the structural integrity of the infarcted
Accumulated evidence indicates that an excessive post-MI
inflammatory response is associated with aggravated LV
B and E), and high-mobility group box 1 protein (HMGB1: C and F) gene expression in the infarcted and non-infarcted areas in myocardial infarcted rats treated
with control (MI/C) or neutralizing anti-high-mobility group box 1 protein antibodies (MI/anti-H). Data are mean+SEM. * P , 0.05 vs. MI/C.
Real-time quantitative reverse transcriptase–polymerase chain reaction analyses of tumour necrosis factor-a (TNF-a: A and D), interleukin-1b (IL-1b:
anti-high-mobility group box 1 protein (HMGB1) antibodies (MI/anti-H: B) on day 3. (C) Graph shows ED-1-immunoreactive cells/mm2. Scale bar indicates
500 mm. Data are mean+SEM. *P , 0.05 vs. MI/C.
Immunohistochemical staining with ED-1 antibody in the infarcted area of myocardial infarcted rats treated with control (MI/C: A) or neutralizing
T. Kohno et al.
by guest on December 25, 2015
remodelling and poor clinical outcomes.2,3,15,19–21It is
known that HMGB1 is quickly released extracellularly after
ischaemic injury.27,28Because extracellular HMGB1 induces
the expression of pro-inflammatory cytokines and adhesion
molecules as an inflammatory mediator,7,9we hypothesized
that serum HMGB1 level could be a predictor of poor clinical
outcomes in patients with MI. Here, we demonstrated that a
marked HMGB1 elevation after MI was associated with pump
failure, cardiac rupture, and in-hospital cardiac death. Peak
serum HMGB1 level was positively correlated with plasma
BNP level 6 months after MI. These findings suggest that
serum HMGB1 level could be a predictor of adverse clinical
outcomes and late-phase LV dysfunction after MI. We pre-
viously reported that an increased peak serum C-reactive
protein level can predict pump failure, cardiac rupture, LV
aneurysm, and 1 year cardiac death after MI;19however,
peak C-reactive protein occurs 2–3 days after MI, which is
not practically useful. Because C-reactive protein is mainly
produced in hepatocytes
pro-inflammatory cytokines,29it is plausible that HMGB1
might be an upstream regulator of serum C-reactive
protein production. We demonstrated that serum HMGB1
level peaked at 12 h after admission, and peak HMGB1
level correlated positively with peak serum C-reactive
protein level. Serum HMGB1 level measurement could be a
useful early predictor of worse clinical outcomes after MI.
Because HMGB1 was released passively from necrotic cells
and secreted actively by inflammatory cells,7,8,10there
might be two possible sources of HMGB1 in MI: necrotic myo-
cardium and infiltrating inflammatory cells. In contrast to
the early elevation of serum HMGB1 level after MI, myocar-
dial HMGB1 mRNA level in the rat MI model did not increase
on day 1. However, immunohistochemical study showed that
HMGB1 immunoreactivity was observed in degenerated myo-
cytes in the infarcted area on day 1 and rapidly disappeared
thereafter, suggesting that the early serum HMGB1 elevation
was mainly due to its passive release from necrotic myo-
cytes. We demonstrated that the elevation of serum
HMGB1 was sustained for 14 days after MI. In addition,
HMGB1 mRNA level in the infarcted area in the rat MI
through stimulation by
data on day 14 after myocardial infarction
Heart weight, haemodynamics, and echocardiographic
Sham MI/C MI/anti-H
7967+559 3748+111* 3218+166*,†
Data are shown as mean+SEM.
MI, myocardial infarction; Sham, sham-operated rats; MI/C, MI rats
treated with control antibody; MI/anti-H, MI rats treated with neutraliz-
ing anti-HMGB1 antibody; LVW, left ventricular weight; BW, body weight;
RVW, right ventricular weight; LVSP, left ventricular systolic pressure; HR,
heart rate; LVEDP, left ventricular end-diastolic pressure; LV þdP/dtmax,
left ventricular maximum rate of isovolumic pressure development; LV
2dP/dtmin, left ventricular minimum rate of isovolumic pressure decay;
LVEDD, left ventricular end-diastolic dimension; LVESD, left ventricular
end-systolic dimension; FS, fractional shortening.
*P , 0.05 vs. Sham,†P , 0.05 vs. MI/C.
expression in the non-infarcted area. Representative haematoxylin–eosin-stained cross-sections of infarcted hearts treated with control (MI/C: A) or neutralizing
anti-high-mobility group box 1 protein antibodies (MI/anti-H: B) on day 7. Effect of high-mobility group box 1 protein blockade as measured by the thickness of
the infarcted and non-infarcted segments (C and D) and expansion index (E). (F) Quantified data of atrial natriuretic factor mRNA expression in the non-infarcted
area on day 14 by real-time quantitative reverse transcriptase–polymerase chain reaction. Data are mean+SEM. *P , 0.05 vs. MI/C.
Effects of high-mobility group box 1 protein (HMGB1) blockade on myocardial histological characteristics and atrial natriuretic factor (ANF) mRNA
Role of HMGB1 in post-MI LV remodelling 571
by guest on December 25, 2015
model was significantly increased on day 3, peaked on day 7,
and remained elevated on day 14 after MI, and its expression
was detected mainly in infiltrating inflammatory cells. Thus,
the sustained serum HMGB1 elevation after MI might be
caused by active secretion from infiltrating inflammatory
cells. Whereas our clinical findings implicate HMGB1 as an
early mediator of the post-MI inflammatory response,
HMGB1, released from activated inflammatory cells, acts
late as a downstream mediator of inflammation, and its
expression increases after the initial rise in TNF-a and
IL-1b and persists after the disappearance of these
pro-inflammatory cytokines.7,10,23We revealed that the kin-
etics of HMGB1 mRNA expression in the infarcted area is also
delayed relative to that of TNF-a and IL-1b, as reported pre-
viously.5,22These findings led to speculation that passive
HMGB1 release from necrotic myocardium might initiate
the post-MI inflammatory response, and subsequent active
release of HMGB1 from infiltrating inflammatory cells
might contribute to the prolongation of the post-MI inflam-
An excessive and persistent inflammatory response after
MI could be unfavourable, leading to LV aneurysm, cardiac
rupture, or chronic LV dilatation.3,15,19–21We have reported
that peripheral monocytosis after reperfused MI was associ-
ated with LV dysfunction and LV aneurysm,20and that gra-
nulocyte–macrophage colony-stimulating factor induction
in a rat MI model resulted in exaggerated LV remodelling
with increased monocyte-derived macrophage infiltration
and impaired reparative fibrosis in the infarcted area.21
We initially hypothesized that the attenuation of the
post-MI inflammatory response by HMGB1 blockade might
prevent LV remodelling. However, HMGB1 blockade failed
to demonstrate any beneficial effect on post-MI LV function.
On the contrary, it resulted in the exaggeration of post-MI LV
remodelling in association with an inhibited inflammatory
response. Although neutralizing anti-HMGB1 antibody inhib-
ited the HMGB1 self-amplification pathway and reduced
pro-inflammatory cytokine expression and macrophage infil-
tration in the infarcted myocardium, the alteration of the
inflammatory response by HMGB1 blockade might be
harmful to infarct healing.
Infarct healing can be divided into three overlapping
phases: inflammatory phase, proliferative phase, and matu-
ration phase.3Infiltrated macrophages during the inflamma-
tory phase contribute to the removal of necrotic tissue,
release of proteolytic enzymes, and production of fibrino-
genic and angiogenic mediators that are important for the
formation of granulation tissue, leading to the proliferative
and maturation phases of infarct healing. van Amerongen
et al. demonstrated that macrophage depletion impairs
wound healing and increases LV remodelling after myocar-
dial injury in mice, suggesting that macrophages play an
important role in the infarct-healing process.30HMGB1 is a
potent stimulator of macrophages, and its effect on macro-
phages includes the promotion of pro-inflammatory cytokine
secretion.9A recent study demonstrated that a lower tissue
HMGB1 level in diabetic skin was associated with impaired
wound healing, and this defect was overcome by topical
application of HMGB1.31It is possible that specific blockade
of HMGB1 might impair the post-MI healing process, result-
ing in adverse LV remodelling. HMGB1 during the early
phase of MI might be a key mediator of the healing
process, which preserves the structural integrity of the
infarcted tissue and protects against increased dilatation
of the LV.
Recent experimental studies enhance our understanding
regarding the role of HMGB1 in post-MI LV remodelling.
Exogenous administration of HMGB1 in the peri-infarcted
LV might have therapeutic potential for the attenuation of
LV remodelling in a permanent MI model through a mechan-
ism that involvesthe activation
cells.32,33Kitahara et al. reported that HMGB1 enhances
angiogenesis and restores cardiac function in a permanent
MI model using transgenic mice with cardiac overexpression
of HMGB1.34These findings demonstrated that HMGB1 has
beneficial effects on the heart after MI through biological
responses related to healing process. Because HMGB1 is
also associated with post-MI inflammatory response, the
caution might be needed when applying these results to
practical use. Multiple points of control may exist to
ensure that the inflammatory response is contained (both
topographically and temporally) according to the area and
time of injury.35,36Because an excessive and persistent
inflammatory response after MI could be unfavourable,
HMGB1 administration during the delayed phase might
have different results. It is also expected that modest
HMGB1 blockade in a situation with an excessive inflamma-
tory response might have some therapeutic potential. The
inflammatory response is much more severe in reperfusion
injury than in permanent MI. In fact, Andrassy et al. recently
showed that systemic administration of HMGB1 causes an
increase in the inflammatory responses and worsens LV
remodelling after myocardial ischaemia–reperfusion injury.
Conversely, HMGB1 inhibition with a functional antagonist of
HMGB1 conferred protection against ischaemia–reperfusion
injury.37HMGB1 might act as a double-edged sword in
post-MI inflammatory response and have bidirectional
effects on LV remodelling depending on the site, extent,
and timing of HMGB1 modulation.38
The present study has some limitations. First, the number
of study participants and events was limited; therefore, it is
likely that the statistical power is too low. To confirm the
clinical significance of serum HMGB1 elevation, a large pro-
spective clinical cohort study is needed. Second, in the
experimental study, we did not evaluate the exact level of
HMGB1 blockade in serum and myocardium. Because the
results could be altered with the extent of HMGB1 blockade
as discussed earlier, further studies including the quantifi-
cation of HMGB1 blockade are needed to clarify the exact
significance of HMGB1 in post-MI LV remodelling.
We demonstrated that the elevation of serum HMGB1 level is
in-hospital cardiac death, in association with an increased
serum C-reactive protein level. These findings suggest that
the overexpression of HMGB1 is associated with poor clinical
outcomes through an excessive inflammatory response in
patients with MI. However, blockade of HMGB1 in a rat MI
model resulted in worsening of LV remodelling through
impaired infarct healing and marked scar thinning. These
findings indicate that HMGB1, a novel predictor of adverse
clinical outcomes after MI, might have an essential role in
the appropriate healing process and in preserving the struc-
tural integrity of the infarcted LV.
T. Kohno et al.
by guest on December 25, 2015
Conflict of interest: none declared.
The Japanese Ministry of Education, Culture, Sports, Science, and
Technology (18790510 to T.K); Keio University Medical Science
1. Pfeffer MA, Braunwald E. Ventricular remodeling after myocardial infarc-
tion. Experimental observations and clinical implications. Circulation
2. Frangogiannis NG, Smith CW, Entman ML. The inflammatory response in
myocardial infarction. Cardiovasc Res 2002;53:31–47.
3. Nian M, Lee P, Khaper N, Liu P. Inflammatory cytokines and postmyocar-
dial infarction remodeling. Circ Res 2004;94:1543–1553.
4. Ertl G, Frantz S. Healing after myocardial infarction. Cardiovasc Res
5. Ono K, Matsumori A, Shioi T, Furukawa Y, Sasayama S. Cytokine gene
expression after myocardial infarction in rat hearts: possible implication
in left ventricular remodeling. Circulation 1998;98:149–156.
6. Bustin M. Regulation of DNA-dependent activities by the functional motifs
of the high-mobility-group chromosomal proteins. Mol Cell Biol 1999;19:
7. Lotze MT, Tracey KJ. High-mobility group box 1 protein (HMGB1): nuclear
weapon in the immune arsenal. Nat Rev Immunol 2005;5:331–342.
8. Scaffidi P, Misteli T, Bianchi ME. Release of chromatin protein HMGB1 by
necrotic cells triggers inflammation. Nature 2002;418:191–195.
9. AnderssonU, WangH, Palmblad
Erlandsson-Harris H et al. High mobility group 1 protein (HMG-1) stimu-
lates proinflammatory cytokine synthesis in human monocytes. J Exp
10. Wang H, Bloom O, Zhang M, Vishnubhakat JM, Ombrellino M, Che J et al.
HMG-1 as a late mediator of endotoxin lethality in mice. Science 1999;
11. Ueno H, Matsuda T, Hashimoto S, Amaya F, Kitamura Y, Tanaka M et al.
Contributions of high mobility group box protein in experimental and
clinical acute lung injury. Am J Respir Crit Care Med 2004;170:
12. Suda K, Kitagawa Y, Ozawa S, Saikawa Y, Ueda M, Ebina M et al.
Anti-high-mobility group box chromosomal protein 1 antibodies improve
survival of rats with sepsis. World J Surg 2006;30:1755–1762.
13. Roberts R, DeMello V, Sobel BE. Deleterious effects of methylpredniso-
lone in patients with myocardial infarction. Circulation 1976;53:
14. Hammerman H, Kloner RA, Hale S, Schoen FJ, Braunwald E. Dose-
dependent effects of short-term methylprednisolone on myocardial
infarct extent, scar formation, and ventricular function. Circulation
15. Sun M, Dawood F, Wen WH, Chen M, Dixon I, Kirshenbaum LA et al. Exces-
sive tumor necrosis factor activation after infarction contributes to sus-
ceptibility of myocardial rupture and left ventricular dysfunction.
16. SivasubramanianN, CokerML,
DeMayo FJ, Spinale FG et al. Left ventricular remodeling in transgenic
mice with cardiac restricted overexpression of tumor necrosis factor.
17. Yamada S, Inoue K, Yakabe K, Imaizumi H, Maruyama I. High mobility
group protein 1 (HMGB1) quantified by ELISA with a monoclonal antibody
that does not cross-react with HMGB2. Clin Chem 2003;49:1535–1537.
18. Shiraki H, Yoshikawa T, Anzai T, Negishi K, Takahashi T, Asakura Y et al.
Association between preinfarction angina and a lower risk of right ventri-
cular infarction. N Engl J Med 1998;338:941–947.
19. Anzai T, Yoshikawa T, Shiraki H, Asakura Y, Akaishi M, Mitamura H et al.
C-reactive protein as a predictor of infarct expansion and cardiac
rupture after a first Q-wave acute myocardial infarction. Circulation
20. Maekawa Y, Anzai T, Yoshikawa T, Asakura Y, Takahashi T, Ishikawa S et al.
Prognostic significance of peripheral monocytosis after reperfused acute
myocardial infarction:a possible role for left ventricular remodeling.
J Am Coll Cardiol 2002;39:241–246.
21. Maekawa Y, Anzai T, Yoshikawa T, Sugano Y, Mahara K, Kohno T et al.
Effect of granulocyte–macrophage colony-stimulating factor inducer
on left ventricular remodeling after acute myocardial infarction. J Am
Coll Cardiol 2004;44:1510–1520.
22. Sugano Y, Anzai T, Yoshikawa T, Maekawa Y, Kohno T, Mahara K et al. Gra-
nulocyte colony-stimulating factor attenuates early ventricular expan-
sion after experimental myocardial infarction. Cardiovasc Res 2005;65:
23. Kim JY, Park JS, Strassheim D, Douglas I, Diaz del Valle F, Asehnoune K
et al. HMGB1 contributes to the development of acute lung injury after
hemorrhage. Am J Physiol Lung Cell Mol Physiol 2005;288:L958–L965.
24. Hassl A, Aspock H. Purification of egg yolk immunoglobulins. A two-step
procedure using hydrophobic interaction chromatography and gel fil-
tration. J Immunol Methods 1988;110:225–228.
25. Fishbein MC, Maclean D, Maroko PR. Experimental myocardial infarction
in the rat: qualitative and quantitative changes during pathologic evol-
ution. Am J Pathol 1978;90:57–70.
26. Alhaddad IA, Tkaczevski L, Siddiqui F, Mir R, Brown EJ Jr. Aspirin enhances
the benefits of late reperfusion on infarct shape. A possible mechanism of
the beneficial effects of aspirin on survival after acute myocardial infarc-
tion. Circulation 1995;91:2819–2823.
27. Tsung A, Sahai R, Tanaka H, Nakao A, Fink MP, Lotze MTet al. The nuclear
factor HMGB1 mediates hepatic injury after murine liver ischemia–
reperfusion. J Exp Med 2005;201:1135–1143.
28. Kim JB, Sig Choi J, Yu YM, Nam K, Piao CS, Kim SW et al. HMGB1, a novel
cytokine-like mediator linking acute neuronal death and delayed
neuroinflammation in the postischemic brain. J Neurosci 2006;26:
29. Dinarello CA. Interleukin-1 and the pathogenesis of the acute-phase
response. N Engl J Med 1984;311:1413–1418.
30. van Amerongen MJ, Harmsen MC, van Rooijen N, Petersen AH, van
Luyn MJ. Macrophage depletion impairs wound healing and increases
left ventricular remodeling after myocardial injury in mice. Am J
31. Straino S, Di Carlo A, Mangoni A, De Mori R, Guerra L, Maurelli R et al.
High-mobility group box 1 protein in human and murine skin: involvement
in wound healing. J Invest Dermatol 2008;128:1545–1553.
32. Limana F, Germani A, Zacheo A, Kajstura J, Di Carlo A, Borsellino G et al.
Exogenous high-mobility group box 1 protein induces myocardial regener-
ation after infarction via enhanced cardiac C-kitþ cell proliferation and
differentiation. Circ Res 2005;97:e73–e83.
33. Chavakis E, Hain A, Vinci M, Carmona G, Bianchi ME, Vajkoczy P et al.
High-mobility group box 1 activates integrin-dependent homing of endo-
thelial progenitor cells. Circ Res 2007;100:204–212.
34. Kitahara T, Takeishi Y, Harada M, Niizeki T, Suzuki S, Sasaki Tet al. High-
mobility group box 1 restores cardiac function after myocardial infarction
in transgenic mice. Cardiovasc Res 2008;80:40–46.
35. Dewald O, Zymek P, Winkelmann K, Koerting A, Ren G, Abou-Khamis T
et al. CCL2/monocyte chemoattractant protein-1 regulates inflammatory
responses critical to healing myocardial infarcts. Circ Res 2005;96:
36. Nathan C. Points of control in inflammation. Nature 2002;420:846–852.
37. Andrassy M, Volz HC, Igwe JC, Funke B, Eichberger SN, Kaya Z et al. High-
mobility group box-1 in ischemia–reperfusion injury of the heart. Circula-
38. Takahashi M. High-mobility group box 1 protein (HMGB1) in ischemic
heart disease: beneficial or deleterious? Cardiovasc Res 2008;80:5–6.
Role of HMGB1 in post-MI LV remodelling573
by guest on December 25, 2015