Atherosclerosis 203 (2009) 557–562
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/atherosclerosis
Acrolein, IL-6 and CRP as markers of silent brain infarction?
Madoka Yoshidaa,d,1, Hideyuki Tomitorib,1, Yoshiki Machia, Daisuke Katagiria, Shiro Uedaa,
Kentaro Horiguchic, Eiichi Kobayashic, Naokatsu Saekic, Kazuhiro Nishimuraa,
Itsuko Ishiia, Keiko Kashiwagib, Kazuei Igarashia,d,∗
aGraduate School of Pharmaceutical Sciences, Chiba University, 1-8-1 Inohana, Chuo-ku, Chiba 260-8675, Japan
bFaculty of Pharmacy, Chiba Institute of Science, Chiba, Japan
cGraduate School of Medicine, Chiba University, Chiba, Japan
dAmine Pharma Research Institute, Innovation Plaza at Chiba University, Chiba, Japan
a r t i c l e i n f o
Received 24 April 2008
Received in revised form 10 July 2008
Accepted 14 July 2008
Available online 26 July 2008
Silent brain infarction
a b s t r a c t
We found previously that increased levels of polyamine oxidase (PAO) [acetylpolyamine oxidase (AcPAO)
plus spermine oxidase (SMO)], and acrolein (CH2
CHCHO) are good markers of stroke. We then inves-
tigated whether silent brain infarction (SBI) can be detected by measuring acrolein, PAO, or other
biomarkers. Several biomarkers were measured in the plasma of 53 normal subjects and 44 subjects
with SBI. It was found that the levels of protein-conjugated acrolein (PC-Acro), interleukin-6 (IL-6) and
C-reactive protein (CRP) were significantly higher in SBI than in normal subjects. PAO was slightly higher
in SBI than in normal subjects. Since the probability of SBI was increased with age, values were analyzed
including age as a factor. When the combined measurements of PC-Acro, IL-6 and CRP were evaluated
together with age using a receiver operating characteristic curve, SBI was indicated with 89% sensitivity
and 91% specificity. The results indicate that measurement of PC-Acro together with IL-6 and CRP makes
it possible to identify SBI with high sensitivity and specificity.
© 2008 Elsevier Ireland Ltd. All rights reserved.
Polyamines (putrescine, spermidine and spermine) are essen-
tial for normal cell growth and are present in cells at millimolar
concentrations . However, when cells are damaged, the toxic
compounds acrolein (CH2 CHCHO) and H2O2are produced from
polyamines, in particular from spermine, by polyamine oxidase
(PAO) [acetylpolyamine oxidase (AcPAO) and spermine oxidase
(SMO)] . When the toxicities of acrolein and H2O2were com-
Abbreviations: PAO, polyamine oxidase; AcPAO, acetylpolyamine oxidase; SMO,
IL-6, interleukin-6; hsCRP, high sensitive C-reactive protein; MMP-9, matrix
metalloprotease-9; MRI, magnetic resonance imaging; FLAIR, fluid-attenuated
inversion recovery; AUC, area under curve; ELISA, enzyme-linked immunosorbent
assay; ROC, receiver operating characteristic.
An initial report of this study was presented at the annual meeting of the
∗Corresponding author at: Graduate School of Pharmaceutical Sciences, Chiba
University, 1-8-1 Inohana, Chuo-ku, Chiba 260-8675, Japan. Tel.: +81 43 224 7500;
fax: +81 43 224 7500.
E-mail address: firstname.lastname@example.org (K. Igarashi).
1Both these authors contributed equally to this work.
polyamines was in fact acrolein . Spermine is either cleaved by
SMO into spermidine and 3-aminopropanal [NH2(CH2)2CHO] or
first acetylated by spermidine/spermine N1-acetyltransferase and
then cleaved by AcPAO into spermidine and 3-acetamidopropanal
[CH3CONH(CH2)2CHO] . It has also been reported that acrolein
is more readily produced from 3-aminopropanal than from 3-
acetamidopropanal . Our hypothesis is that acrolein is produced
from polyamines when cells are damaged, and reacts with protein
and DNA to cause further cell damage [6,7]. Accordingly, the level
of protein-conjugated acrolein (PC-Acro) in plasma should reflect
the magnitude of lesions. We examined whether the levels of PAO
and PC-Acro in plasma are correlated with pathologies that involve
cell damage, and found that levels of PAO and PC-Acro in plasma
are well-correlated with the severity of chronic renal failure 
and stroke . Stroke is a sudden focal neurological deficit caused
by vascular insult, accompanied by cell damage in the central ner-
vous system. Our results indicated that PC-Acro and PAO are useful
as biochemical markers for stroke . It has been reported that
silent brain infarction (SBI) increases the risk of subsequent stroke
[9–11], dementia  and mild cognitive impairment . There
is also a report that SBI is more common in patients with obstruc-
tive sleep apnea . It is, therefore, valuable to detect SBI at an
chemical markers in blood is common and economical compared
0021-9150/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved.
M. Yoshida et al. / Atherosclerosis 203 (2009) 557–562
to diagnostic imaging such as magnetic resonance imaging (MRI)
and computed tomography (CT). In this communication, we stud-
ied whether SBI can be detected by measuring PC-Acro and some
other biochemical markers.
2. Materials and methods
2.1. Subjects and collection of blood
We examined 114 elderly volunteers (61 women and 53 men,
aged 65.5±8.4 years, range 45–88 years). All participants were
healthy volunteers and were living independently at home with-
given by each participant, and our study protocol was approved
by the ethics committees of Graduate School of Medicine, and of
Pharmaceutical Sciences, Chiba University. Experiments were con-
ducted in accordance with the Declaration of Helsinki principles.
Blood containing 3U/mL heparin was centrifuged at 1500×g for
10min at 4◦C, and plasma was carefully collected to avoid contam-
ination by erythrocytes.
2.2. Measurement of PC-Acro in plasma
lysine]) in protein was determined by the method of Uchida et
al.  using ACR-LYSINE ADDUCT ELISA SYSTEM (NOF Corpo-
ration) and 0.01mL plasma. After the reaction was terminated,
absorbance at 450nm was measured by a microplate reader
2.3. Assays for SMO and AcPAO in plasma
The activity of SMO and AcPAO was evaluated by mea-
suring the level of spermidine produced from spermine and
N1-acetylspermine, respectively. The reaction mixture (0.06mL)
To 0.02mL of the reaction mixture, 0.55mL of 5% trichloroacetic
acid was added and centrifuged at 12,000×g for 10min. A 0.01mL
aliquot of the supernatant was used for the polyamine mea-
surement by high-performance liquid chromatography . The
increase per mL plasma.
2.4. Measurement of interleukin-6 (IL-6), C-reactive protein
(CRP), S100B, matrix metalloprotease-9 (MMP-9) and adiponectin
IL-6, high sensitive CRP, S100B, MMP-9 and adiponectin were
measured using Endogen Human IL-6 ELISA Kit (Pierce Biotechnol-
ogy, Inc.), human CRP ELISA Kit (Alpha Diagnostic), CanAg S100BB
EIA (CanAg Diagnostic), Human MMP-9 Biotrak ELISA System (GE
Pharmaceutical Co., Ltd.), respectively, according to the manufac-
turer’s protocol. After the reaction was terminated, absorbance at
450nm (IL-6, CRP, MMP-9 and adiponectin) or 620nm (S100B) was
measured by a microplate reader Hitachi MTP-800AFC. All bio-
blind to the results of the MRI.
attenuated inversion recovery (FLAIR) in parallel with collection of
dilated perivascular spaces. All MRI was performed in 5–10mm
thickness with 1–2mm slice gap with a 1.5-T MRI unit (Signa; GE
Medical Systems). A standard head coil with a receive–transmit
birdcage design was used. The maximum size of focal infarcts was
image. The diagnosis of SBI was made as follows: (1) spotty areas
≥3mm in diameter showing in high intensity in the T2 and FLAIR
images and low intensity in the T1 image, (2) lack of neurological
signs and/or symptoms that can be explained by the MRI lesions,
and (3) no medical history of clinical stroke [9,16]. ‘Normal’ sub-
jects are defined as subjects with: (1) no spotty areas; (2) lack of
neurological signs and/or symptoms and (3) no medical history of
clinical stroke. An interpreter study for evaluating the MRI lesions
was performed by 2 independent medical doctors experienced in
studies of brain disease in a blind fashion. The agreement rate was
more than 99%. Through this diagnosis, 97 subjects were classified
to have less than 3mm infarction, and excluded from SBI.
Statistical calculations were performed with GraphPad Prism®
or median±interquartile deviation. Groups were compared using
Wilcoxon rank sum test. Difference in each marker or combina-
tion of various markers in SBI vs. no stroke was evaluated using a
receiver operating characteristic (ROC) curve . ROC curve anal-
ysis was performed with artificial neural networks [18,19] using
NEUROSIM/L software (Fujitsu). Candidates of cutoff values and
works and the cutoff value was set up as the closest point on ROC
curve from the P point, that is sensitivity=1 and 1-specificity=0.
Sensitivity, specificity, positive- and negative-predictive values
were calculated with the standard method .
3.1. Correlation between SBI and PC-Acro and/or other
We have previously reported that the size of infarction in stroke
patients was nearly parallel with the multiplied value of PC-Acro
and PAO in plasma . Thus, we tested whether these two mark-
ers and other biochemical markers are correlated with SBI. For this
study, 114 healthy volunteers underwent MRI, and 17 participants
were excluded from the SBI group because they had infarcts of less
than 3mm in diameter. Accordingly, the remaining 97 volunteers
(51 women and 46 men, aged 65.3±8.6 years, range 45–88 years)
were separated into two groups—53 normal subjects and 44 sub-
jects with SBI as detected by MRI (infarcts≥3mm). Locations of
infarctions in 20 randomly chosen subjects with SBI are shown
compiled in Fig. 1; the location and number of infarctions were
similar in the other 24 subjects in this group (data not shown).
The 20 subjects illustrated in Fig. 1 had a total of 37 infarctions, of
which 32 were 3–10mm in size and 5 were 10–20mm. So, most of
infarctions were less than 10mm and were located in the middle of
cerebrum, that is, the subcortical white matter or the basal ganglia,
similar to the size and location of SBI reported in other studies .
Only 2 infarctions were observed in the cerebellum.
Acro were compared between 53 normal subjects and 44 subjects
with SBI. There was a significant difference in average age (61.9 and
69.4 years old, respectively) (Fig. 2), indicating that the probability
M. Yoshida et al. / Atherosclerosis 203 (2009) 557–562
Fig. 1. Location and size of infarctions on brain of 20 subjects with SBI. T1- and T2-weighted MRI was performed in 5–10mm thickness with 1–2mm slice gap with a 1.5-T
MRI unit (a–s). Location and size of infarction of 20 subjects with SBI were described schematically.
of SBI increased with age. The average age of the subjects with less
than 3mm infarction was 66.5 years old. The level of PC-Acro was
in SBI (Fig. 2 and Supplementary Table 1S). The levels of CRP and
IL-6 are reported to increase in the serum of apparently healthy
individuals with SBI . It has also been reported that MMP-9 and
S100B, a protein mainly expressed in astroglial cells, are released
during the early period of stroke . Thus, we measured these
Fig. 2. Comparison of age and level of PC-Acro, IL-6, CRP and adiponectin in the plasma of normal and SBI subjects. Age, PC-Acro, IL-6, CRP and adiponectin were shown with
means (horizontal line). Experiments were performed twice and results were reproducible. The p value was calculated using Wilcoxon rank sum test.
M. Yoshida et al. / Atherosclerosis 203 (2009) 557–562
Detection of SBI with ROC curve of various markers
MarkerAUC Cutoff Sensitivity (%) Specificity (%)PPV (%)NPV (%) LR+LR−
Difference in combination of various markers in SBI vs. no stroke was evaluated using a receiver operating characteristic (ROC) curve  as described in Section 2. AUC, area
under curve. Cutoff values, percentage of sensitivity and specificity, positive- and negative-predictive values (PPV and NPV), and positive- and negative-likelihood (LR+ and
LR−) are shown.
markers in normal and SBI subjects. The levels of IL-6 and CRP in
plasma of SBI subjects were higher than in normal subjects, but
levels of MMP-9 and S100B were not significantly different in nor-
mal and SBI groups (Fig. 2 and Supplementary Table 1S). There are
contradictory reports concerning adiponectin: one is that hypoad-
iponectimia is associated with brain stroke , and the other is
that level of adiponectin is not correlated with stroke . It was
found that the level of adiponectin was high in SBI (Fig. 2). Levels of
other biochemical markers were also measured in blood from nor-
mal and SBI subjects (Supplementary Table 1S). The level of LDH
was high, and that of platelets was low in SBI. Levels of glucose,
total cholesterol and triglyceride, markers of metabolic syndrome
, were not significantly different in normal and SBI subjects.
3.2. Detection of SBI by PC-Acro, IL-6 and CRP
We next determined whether SBI could be detected by altered
technique for assessing diagnostic and predictive accuracy in dis-
ease management . Since age was an important factor among
various markers evaluated for detection of SBI (Fig. 2), all values
were analyzed including age as a factor with artificial neural net-
works using NEUROSIM/L software. The sensitivity and specificity
for detecting SBI by PC-Acro were 82% and 72%, and area under
(Table 1). The ROC curve with PC-Acro is shown in Fig. 3A. Sensitiv-
ity and specificity for detecting SBI by IL-6 and CRP were 82% and
Fig. 3. ROC curve of Age/PC-Acro (A), Age/IL-6 (B), Age/PC-Acro/IL-6 (C), and Age/PC-Acro/IL-6/CRP (D) for SBI vs. no stroke. ROC curve analysis was performed as described
in Section 2. Sensitivity and specificity were evaluated using ROC curve. AUC, area under curve.
M. Yoshida et al. / Atherosclerosis 203 (2009) 557–562
74%, and 73% and 77%, respectively. The AUC for IL-6 and CRP was
0.8310 and 0.7663, respectively (Table 1). The results indicate that
detection of SBI by PC-Acro was nearly equal to that by IL-6, and
higher than that by CRP. ROC curve with IL-6 is shown in Fig. 3B.
To enhance the probability of detecting SBI, sensitivity and
specificity were determined by a combination of two biomarkers
with age. Although the probability of detecting SBI was enhanced
by a combination of two biomarkers, sensitivity and specificity for
detecting SBI were highest with the combination of PC-Acro and
IL-6, and they were 89% and 87%, respectively, and the AUC was
0.8885. ROC curve with these two markers is shown in Fig. 3C. The
positive predictive value (PPV) and negative predictive value (NPV)
were also increased (Table 1). We next determined whether the
probability of detecting SBI increases further with a combination
of three markers—PC-Acro, IL-6 and CRP. Sensitivity and specificity
for detecting SBI became 89% and 91%, respectively, and the AUC
was 0.9196. ROC curve with PC-Acro, IL-6 and CRP is shown in
Fig. 3D. PPV and NPV were 89% and 91%, respectively. Positive-
and negative-likelihood (LR+ and LR−) were 9.40 and 0.13, respec-
tively, which were close to the recommended values in guidelines
for diagnosis .
We have previously reported that plasma levels of PAO and PC-
Acro are good markers for stroke . It has been reported that
the properties of the blood–brain barrier change during cerebral
ischemia such that various molecules can pass into the blood .
This is consistent with the idea that elevated levels of spermidine,
spermine and PAO are derived from damaged cells in the central
nervous system. In this study, we estimated whether SBI can be
detected by the level of PC-Acro and other biomarkers. Data were
analyzed taking into account age because the average age of nor-
mal, SBI, and stroke subjects was 61.9, 69.4 and 70.7 years old ,
respectively. We succeeded in detecting SBI by measuring PC-Acro,
IL-6 and CRP in plasma with 89% sensitivity and 91% specificity
to 61 stroke patients, sensitivity and specificity were 95% and 96%,
respectively (unpublished results). The results indicate that age,
acrolein and inflammation markers are reliable markers for SBI as
well as stroke.
Although a significant increase in the level of PAO (AcPAO plus
SMO) was observed in stroke , a similar change was not seen
in SBI. An increase in AcPAO occurred first, followed by increases
in SMO and then PC-Acro after the onset of stroke . In this con-
text it is also notable that the turnover rate of enzymes such as
PAO is thought to be faster than that of the products such as PC-
Acro. At this moment, however, it remains to be clarified when
SBI occurred. This may be clarified by determining the half-life of
PC-Acro. Acrolein produced by lipid peroxidation may also con-
tribute to the increase in the level of PC-Acro , although the
level of acrolein produced by lipid peroxidation was much lower
than that produced by oxidation of polyamines by PAO (data not
produced from spermine using an inhibitor of PAO.
We omitted from the SBI group 17 subjects who had infarctions
less than 3mm in diameter. If these subjects were included in the
SBI group (total 61 subjects), the average age was 68.6 years old,
the sensitivity and specificity by measuring PC-Acro and IL-6 were
95% and 77%, respectively, and the AUC was 0.8916. However, the
level of CRP was not significantly different between normal and
this expanded SBI group. Thus, sensitivity and specificity by mea-
the AUC was 0.8764, indicating that accuracy to detect SBI slightly
decreased if subjects with infarctions less than 3mm in diameter
were included in the SBI group (see Supplementary Table 2S, Figs.
1S and 2S).
With regard to adiponectin, more detailed analysis is necessary
since there are contradictory reports (Fig. 2 and Refs. [23,24]). Our
data also indicate that LDH is high in SBI. There is also a report that
LDH increases in cerebrospinal fluid in case of stroke . Through
a detailed analysis, it may be possible to find another biochemical
marker of stroke.
In conclusion, it is possible to identify SBI subjects with high
probability by measuring PC-Acro, IL-6 and CRP. It has been
infarctions can be detected at the period of SBI, it becomes possible
to delay or reduce aggravation of the infarction by suitable therapy.
Future studies are necessary to develop how to treat SBI patients
for the improvement of quality of life.
We thank Drs. K. Williams and A.J. Michael for their help in
preparing this manuscript. This work was supported by a Grant for
Fund Method from New Energy and Industrial Technology Devel-
opment Organization (NEDO), Japan.
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at doi:10.1016/j.atherosclerosis.2008.07.022.
 Igarashi K, Kashiwagi K. Breakthroughs and views. Polyamines: myste-
rious modulators of cellular functions. Biochem Biophys Res Commun
 Tomitori H, Usui T, Saeki N, et al. Polyamine oxidase and acrolein as novel
biochemical markers for diagnosis of cerebral stroke. Stroke 2005;36:2609–13.
 Sharmin S, Sakata K, Kashiwagi K, et al. Polyamine cytotoxicity in the
presence of bovine serum amine oxidase. Biochem Biophys Res Commun
 Wang Y, Casero Jr RA. Mammalian polyamine catabolism: a therapeutic target,
a pathological problem, or both? J Biochem (Tokyo) 2006;139:17–25.
 Wood PL, Khan MA, Moskal JR. The concept of “aldehyde load” in neurode-
generative mechanisms: cytotoxicity of the polyamine degradation products
hydrogen peroxide, acrolein, 3-aminopropanal, 3-acetamidopropanal and 4-
aminobutanal in a retinal ganglion cell line. Brain Res 2007;1145:150–6.
 Uchida K, Kanematsu M, Sakai K, et al. Protein-bound acrolein: potential mark-
ers for oxidative stress. Proc Natl Acad Sci USA 1998;95:4882–7.
 Kawai Y, Furuhata A, Toyokuni S, Aratani Y, Uchida K. Formation of acrolein-
derived 2?-deoxyadenosine adduct in an iron-induced carcinogenesis model. J
Biol Chem 2003;278:50346–54.
 Sakata K, Kashiwagi K, Sharmin S, et al. Increase in putrescine, amine oxidase,
and acrolein in plasma of renal failure patients. Biochem Biophys Res Commun
 Kobayashi S, Okada K, Koide H, Bokura H, Yamaguchi S. Subcortical silent brain
infarction as a risk factor for clinical stroke. Stroke 1997;28:1932–9.
 Vermeer SE, Hollander M, van Dijk EJ, et al. Silent brain infarcts and white
matter lesions increase stroke risk in the general population: The Rotterdam
Scan Study. Stroke 2003;34:1126–9.
 Vermeer SE, Longstreth Jr WT, Koudstaal PJ. Silent brain infarcts: a systematic
review. Lancet Neurol 2007;6:611–9.
 Lopez OL, Jagust WJ, Becker JT, et al. Risk factors for mild cognitive impair-
ment in the Cardiovascular Health Study Cognition Study, part 2. Arch Neurol
 Minoguchi K, Yokoe T, Tazaki T, et al. Silent brain infarction and platelet activa-
tion in obstructive sleep apnea. Am J Respir Crit Care Med 2007;175:612–7.
 Uchida K, Kanematsu M, Morimitsu Y, et al. Acrolein is a product of lipid per-
oxidation reaction. Formation of free acrolein and its conjugate with lysine
residues in oxidized low density lipoproteins. J Biol Chem 1998;273:16058–66.
 Igarashi K, Kashiwagi K, Hamasaki H, et al. Formation of a compensatory
polyamine by Escherichia coli polyamine-requiring mutants during growth in
the absence of polyamines. J Bacteriol 1986;166:128–34.
 Hoshi T, Kitagawa K, Yamaguchi H, et al. Relations of serum high-sensitivity
C-reactive protein and interleukin-6 levels with silent brain infarction. Stroke
M. Yoshida et al. / Atherosclerosis 203 (2009) 557–562
 Hanley JA, McNeil BJ. The meaning and use of the area under a receiver oper-
ating characteristic (ROC) curve. Radiology 1982;143:29–36.
 Kazmierczak SC, Catrou PG, Van Lente F. Diagnostic accuracy of pancre-
atic enzymes evaluated by use of multivariate data analysis. Clin Chem
 Ellenius J, Groth T, Lindahl B, Wallentin L. Early assessment of patients with
suspected acute myocardial infarction by biochemical monitoring and neural
network analysis. Clin Chem 1997;43:1919–25.
 Brenner H, Gefeller O. Variation of sensitivity, specificity, likelihood ratios and
predictive values with disease prevalence. Stat Med 1997;16:981–91.
 Hougaku H, Matsumoto M, Handa N, et al. Asymptomatic carotid lesions and
silent cerebral infarction. Stroke 1994;25:566–70.
 Foerch C, Otto B, Singer OC, et al. Serum S100B predicts a malignant course
of infarction in patients with acute middle cerebral artery occlusion. Stroke
 Chen M-P, Tsai JC-R, Chung F-M, et al. Hypoadiponectinemia is associ-
ated with ischemic cerebrovascular disease. Arterioscler Thromb Vasc Biol
 Söderberg S, Stegmayr B, Stenlund H, et al. Leptin, but not adiponectin, predicts
stroke in males. J Intern Med 2004;256:128–36.
 Ninomiya JK, L’Italien G, Criqui MH, et al. Association of the metabolic
syndrome with history of myocardial infarction and stroke in the Third
National Health and Nutrition Examination Survey. Circulation 2004;109:
 Linden A. Measuring diagnostic and predictive accuracy in disease manage-
ment: an introduction to receiver operating characteristic (ROC) analysis. J Eval
Clin Pract 2006;12:132–9.
 Ballabh P, Braun A, Nedergaard M. The blood–brain barrier: an overview:
structure, regulation, and clinical implications. Neurobiol Dis 2004;16:
 Lampl Y, Paniri Y, Eshel Y, Sarova-Pinhas I. Cerebrospinal fluid lactate dehy-
drogenase levels in early stroke and transient ischemic attacks. Stroke