Atherosclerosis 177 (2004) 291297

Absence of an atheroprotective effect of the garlic powder printanor in

APOE*3-Leiden transgenic mice
Sonia M.S. Espirito Santoa,b , Bart J.M. van Vlijmena,c , Wim van Duyvenvoordea , Erik H.
Offermana , Louis M. Havekesa,b,c , Ingrid Arnaultd , Jacques Augerd , Hans M.G. Princena,
aTNO Prevention and Health, Gaubius Laboratory, P.O. Box 2215, 2301 CE Leiden, The Netherlands
b Department of General Internal Medicine, Leiden University Medical Center, Leiden, The Netherlands
c Department of Cardiology, Leiden University Medical Center, Leiden, The Netherlands
d Institut de Recherche sur la Biologie de lInsecte, University F. Rabelais, Tours, France

Received 10 March 2004; received in revised form 21 June 2004; accepted 22 July 2004
Available online 11 September 2004

Abstract

Numerous animal studies have reported that garlic can protect against atherosclerosis. However, a comparable number of studies do not
support this observation. This contradiction may result from differences in study design, use of different animal models, and use of different
garlic formulations and preparations. Here, we investigated the effect of the chemically well-characterized and production-controlled garlic
powder printanor on atherosclerosis in the APOE*3-Leiden transgenic mouse, a mouse model well suited for evaluating anti-atherosclerotic
properties of drugs and food components under human-like conditions. APOE*3-Leiden mice were fed a Western diet supplemented with
either 5 or 50 g kg1 printanor. As a reference, the commercially available fermented garlic kyolic was included (1.6 g kg1 diet). Treatment
with printanor demonstrated reduced body weight, coinciding with increased feces production and fecal fatty acids excretion. Printanor and
kyolic treatment did not affect plasma lipids, markers of inflammation (serum amyloid A, serum-soluble intercellular adhesion molecule-1,
and blood-leukocytes tumor necrosis factor- (TNF) production) and vascular activation (plasma von Willebrand factor (vWF)). As analyzed
after 28 weeks of treatment, printanor and kyolic did not affect atherosclerotic lesion type, area or composition. Under conditions relevant to the
human situation, the well-characterized and production-controlled garlic powder printanor does not display hypolipidemic, anti-inflammatory
or anti-atherosclerotic properties.
2004 Elsevier Ireland Ltd. All rights reserved.

Keywords: Garlic; Body weight; Inflammation; Atherosclerosis; APOE*3-Leiden transgenic mice

1. Introduction
Garlic has been reported to have beneficial effects on risk
factors associated with cardiovascular disease (CVD) includ-
ing lowering of plasma lipids, systolic blood pressure, de-
crease of pro-inflammatory cytokines production and reduc-
tion of platelet activation state [17]. Through these actions
garlic is thought to protect the vessel wall from progressive
atherosclerotic disease and consequently CVD [14]. Numer-
Corresponding author. Tel.: +31 71 518 1471; fax: +31 71 518 1904.
E-mail address: jmg.princen@pg.tno.nl (H.M.G. Princen).
ous animal studies have been performed showing that garlic
indeed has beneficial effects on risk factors associated with
CVD [1,2]. However, a comparable number of animal stud-
ies do not support these observations [1,2]. Limited number
of animal studies focused on the direct effect of garlic on
atherosclerosis development. The latter studies included rab-
bit and mouse studies that showed a garlic reducing effect
on atherosclerotic lesion area [811]. In contrast, pig and rat
studies showed that garlic did not lower atherosclerotic le-
sion area [12,13]. These conflicting data may be ascribed to
the use of the different animal models, as well as to the use
of different garlic-derived materials, ranging from raw gar-

0021-9150/$ see front matter 2004 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.atherosclerosis.2004.07.024
292 S.M.S.E. Santo et al. / Atherosclerosis 177 (2004) 291297
lic and garlic powders (with variation in production condi-
tions and chemical composition) to isolated sulfur-containing
compounds. In addition, to induce an atherogenic lipopro-
tein profile, previous animal studies were performed using
diets containing a varying and high amount of cholesterol
(12%, w/w), and one can question the relevance of the out-
come of these studies for the human situation. Hence, it re-
mains difficult to conclude whether garlic truly has an athero-
protective effect under conditions relevant to the human
situation.
In the present study, we evaluated the anti-atherosclerotic
properties of a chemically well-characterized and production-
controlled garlic powder derived from the garlic variety print-
anor. This high-quality garlic powder is produced under
high sulfur-fertilization levels during the cultivation, con-
sists of high levels of well-defined bioactive sulfur-containing
compounds including alliin, -glutamyl-S-allylcysteine and
-glutamyl-1-propenylcysteine [14]. Printanor has been pro-
duced as a part of European Union research program enti-
tled Garlic and Health carried out by a consortium of 15
independent research groups from six countries. The main
goals of the Garlic and Health project are to further un-
derstand and improve the production of active compounds
by garlic and to further understand the role of garlic as
diet and as therapy in promoting and sustaining health and
preventing cancer and importantly cardiovascular diseases.
The anti-atherosclerotic properties of printanor were investi-
gated in APOE*3-Leiden transgenic mice. These mice carry
a rare genetic variant of APOE, the APOE*3-Leiden, which
is associated with a dominantly inherited familial dysbe-
talipoproteinemia (FD) in humans. Their lipoprotein pro-
file is very similar to the FD patients in which elevated
plasma cholesterol and triglyceride levels are mainly con-
fined to the VLDL/LDL sized lipoprotein fractions. There-
fore, the APOE*3-Leiden mice display a human-like and
atherogenic lipoprotein profile already when feeding a mild
Western diet (15% saturated fat, 0.2% cholesterol). In addi-
tion, these mice respond comparable to humans to athero-
protective drugs, like statins and fibrates, and food com-
ponents, like stanols and fish oils [15,16]. Recently, these
mice have also been successfully employed to demonstrate
the anti-inflammatory and (plasma cholesterol-independent)
anti-atherosclerotic properties of rosuvastatin [17]. In the cur-
rent study, we investigated printanor at a low dose (5 g kg1
diet), which allowed a garlic dosing comparable to human
studies, and at a high dose (50 g kg1 diet) but still ad li-
bitum tolerated dose in these APOE*3-Leiden mice. As a
reference, the commercially available fermented garlic ky-
olic was included (1.6 g kg1 diet) at the same dose that
demonstrated to have an anti-atherogenic potential at least
in rabbits fed a diet supplemented with 1% cholesterol
[10]. During 28 weeks of treatment, we monitored no ef-
fects of printanor and kyolic on plasma lipids (cholesterol,
triglycerides, lipoproteins), markers of inflammation (serum
amyloid A, serum-soluble intercellular adhesion molecule-1,
lipopolysaccharide-induced tumor necrosis factor- (TNF)
production by blood-derived monocytes) and vascular acti-
vation (plasma von Willebrand factor (vWF)). In addition,
printanor and kyolic treatment did not affect atherosclerotic
lesion type, area, and composition at the level of the aortic
root.
2. Materials and methods
2.1. Mice and experimental diets
Five-month-old female APOE*3-Leiden transgenic mice
backcrossed into a C57Black6/J background were used [18].
Fifty-eight mice were fed a semi-synthetic Western diet
containing 0.2% cholesterol, 15% cocoa butter, 40.5% su-
crose, 1% corn oil, 20% casein, 6.2% cellulose, and 2%
choline chloride (supplied by Hope Farms, Woerden, The
Netherlands) for 3 weeks. Thereafter, mice were assigned to
four different treatment groups (n = 1415 mice per group)
with equal average of plasma cholesterol levels. One control
group continued receiving the Western diet, two experimen-
tal groups received the Western diet supplemented with the
French garlic powder printanor (supplied by INRA, Dijon,
France) at a concentrations of 5 g kg1 (0.5%) and 50 g kg1
(5%) diet, and one group received the Western diet sup-
plemented with the fermented garlic kyolic (kyolic reserve,
Wakunaga, Mission Viejo, CA, USA) at a concentration of
1.6 g kyolic kg1 diet. To mask the possible taste of the in-
corporated garlic-derived material sugar lumps containing
anise (33.3 g kg1 diet) were supplemented to the diets. Print-
anor and kyolic were analyzed by standard high-performance
liquid chromatography (HPLC) procedure for content of
garlic sulfur-containing compounds (alliin, -glutamyl-S-
allylcysteine and -glutamyl-1-propenylcysteine) [14]. All
(experimental) diets and water were given ad libitum. The
institutional committee on animal welfare of TNO approved
all animal experiments.
2.2. Body temperature and total fecal fatty acids
composition
Body temperature was measured using a rectal probed for
mice (RET-3, Physitemp instruments, Clifton, NJ, USA).
Feces were collected during a 3-day period per cage (n
= 34 mice per cage), lyophilized and weighed. To extract
fatty acids from a pool of dried feces, 15 mg of dried feces
were treated with 1 mL of alkaline methanol, as previously
described for bile acids measurement, using pentadecanoic
acid as internal standard [19]. After treatment, the tubes were
cooled to 4 C and the fatty acids were extracted with 5 mL
of 6% potassium carbonate after centrifuging for 10 min at
14,000 g. The fatty acids were separated on a gas chro-
matographic column (CP sil 88, Chrompack, Middelburg,
The Netherlands) in a gas chromatograph (GC) (Varian, Mid-
delburg, The Netherlands) equipped with a flame ionization
detector. The injector and the flame ionization detector were
 S.M.S.E. Santo et al. / Atherosclerosis 177 (2004) 291297
kept at 270 C. The column temperature was programmed
from 170210 C. Quantification was based on the area ratio
of the individual fatty acid to the internal standard [19].
2.3. Blood analysis
After a 4 h fasting period from 7:00 to 11:00, blood sam-
ples were collected from each individual mouse by tail bleed-
ing for isolation of EDTA-plasma or serum. Total plasma
cholesterol and triglyceride levels were measured enzymat-
ically, using commercially available kits (No. C0534 and
No. 337-B, Sigma Diagnostics, Deisenhofen, Germany). Size
fractionation of plasma lipoproteins was performed using a
Smart-system (Pharmacia, Uppsala, Sweden) as previously
described [16].
Plasma alanine aminotransferase (ALAT) was measured
enzymatically per mouse (ALAT test, Roche Diagnostics,
Mannheim, Germany).
Serum amyloid A (SAA), soluble intercellular adhesion
molecule-1 (soluble ICAM1), and plasma vWF were mea-
sured by ELISA specific for SAA [17], soluble ICAM1
(ICAM1, Perbio Science, Etten-Leur, The Netherlands), and
vWF [20].
2.4. Tumor necrosis factor- (TNF) production assay
Whole blood was obtained by tail bleeding and col-
lected in EDTA-coated vials. Blood was diluted 25
times in RPMI medium supplemented with l-glutamine,
penicillin, and streptomycin, (RPMI 1640, BioWhittaker,
Verviers, Belgium), which contained varying concentrations
(0150 ng mL1 ) of lipopolysaccharide (LPS, Re 595, List
Biological Laboratories, Campbell, CA). Following incuba-
tion overnight at 37 C, 50 L of the supernatant was ana-
lyzed for TNF content by ELISA (TNF, BD-Biosciences
Pharmingen, San Diego, CA, USA) [21].
2.5. Atherosclerosis analysis
Hearts and aortas were fixed in phosphate-buffered 4%-
formaldehyde, dehydrated, and embedded in paraffin. Hearts
were cross-sectioned (5-m) throughout the entire aortic root
area. Per mouse, four sections with 30-m intervals were
used for quantification of atherosclerotic lesion area [22].
Sections were routinely stained with hematoxylin-phloxine-
Table 1
Effect of printanor and kyolic on food intake, body weight gain, total dry feces production, total fatty acids excretion in feces and total fatty acid excretion after
28 weeks of study under Western diet conditions
Food intake Body weight
(g/mouse/day) gain (g)
Control 2.8 0.1 6.2 3.1
0.5% Printanor 2.9 0.1 5.0 1.8
5% Printanor 2.8 0.2 2.5 1.6
Kyolic 2.9 0.1 6.3 2.6
Data are presented as mean S.D. for 1415 mice per group. p < 0.05 as compared with the control group.
293
saffron (HPS). Lesion area was determined using Leica Qwin
image analysis software (EIS, Asbury NJ) [22]. The number
of leukocytes attached to the endothelium was counted in the
same four HPS-stained sections with 30-m intervals also
used for the quantification of atherosclerotic lesion area [23].
2.6. Statistical analysis
All data are represented as mean S.D. Data were ana-
lyzed using the MannWhitney U-test. Food intake data was
analyzed with general linear model-repeated measures. p-
values less than 0.05 were regarded as statistically signifi-
cant.
3. Results
3.1. Tolerance of experimental diets
Printanor and kyolic were analyzed for the content of
garlic sulfur-containing compounds (alliin, -glutamyl-S-
allylcysteine and -glutamyl-1-propenylcysteine) by high-
performance liquid chromatography analysis. Printanor and
kyolic contained a comparable amount of alliin (1.61% ver-
sus 1.48%, w/w) and -glutamyl-S-allylcysteine (0.44% ver-
sus 0.54%, w/w). However, printanor had a four-fold higher
content of -glutamyl-1-propenylcysteine as compared with
kyolic (1.27% versus 0.33%, w/w). Thereby, printanor gar-
lic powder contained approximately 30% more of these
bioactive sulfur-rich compounds when compared with kyolic
(3.32% versus 2.35%, w/w).
All experimental diets were well tolerated and no dif-
ferences were observed in food intake for mice that con-
sumed experimental garlic containing diet or a control diet.
All groups of mice appeared healthy and gained body weight
to the same extent during the 28 weeks of the study. How-
ever, from week 12 on, the 5% printanor (highest print-
anor dose)-treated group showed a significantly lower body
weight gain as compared with control group (Table 1). Con-
sequently, the 5% printanor-treated mice had a significantly
lower body weight as compared with controls (22.9 2.3 g
versus 26.6 3.6 g, p = 0.007 for week 28). The 0.5% print-
anor (lowest printanor dose)-treated mice also displayed a
(non-significant) lower body weight gain (Table 1), resulting
in a significant reduction in body weight as compared with
controls (23.1 2.3 g versus 24.9 2.4 g, p = 0.026 for week
Total dry feces production Total fatty acid excretion Total fatty acid excretion
(g feces/100 g mouse/day) in feces (g/g feces) (g/100 g mouse/day)
1.5 0.3 0.027 0.006 0.038 0.007
1.7 0.4 0.032 0.008 0.052 0.012
1.9 0.3 0.029 0.008 0.053 0.011
1.7 0.4 0.023 0.007 0.034 0.013
294 S.M.S.E. Santo et al. / Atherosclerosis 177 (2004) 291297

Fig. 1. Plasma lipoprotein distribution was determined 28 weeks after garlic dietary treatment. Lipoproteins were size-fractionated by FPLC (1 pool, n = 1415
mice per group), cholesterol (closed squares) and triglycerides (open squares) content of the individual fractions was determined.

20 and 24.6 2.3 g versus 26.6 3.6 g, p = 0.041 for week
28) from week 20 till the end of the experiment (week 28).
Printanor-treated mice had a body temperature compara-
ble to control mice (Control, 38.3 0.2 C; 0.5% printanor,
38.3 0.2 C; 5% printanor, 38.3 0.2 C), indicating that
the lower body weight observed in these groups is possibly
not due to higher mouse metabolic activity. However, the 5%
printanor-treated group did show a significantly (p = 0.006)
21% higher excretion of dry feces as compared with con-
trol (Table 1). Since the fatty acid composition of the dry
feces was not affected by printanor, the net waste of fecal
fatty acids was significantly increased both in the high and
in the low doses of printanor group (Table 1), indicating an
increased fecal fatty acids excretion for the printanor (both
doses)-treated groups. However, these results do not com-
pletely explain the 21% higher excretion of dry feces for the
5% printanor-treated group. Possibly, other compounds are
being excreted. Kyolic-treated mice had body weight, body
temperature, dry feces production and fecal fatty acid excre-
tion comparable to controls.
3.2. Plasma lipids and alanine aminotransferase
Feeding APOE*3-Leiden mice a Western diet resulted
in hyperlipidemia with plasma cholesterol levels rang-
ing from 1013 mmol/L and plasma triglyceride levels of
1.01.5 mmol/L. Fast-performance liquid chromatography
(FPLC) analysis revealed that plasma cholesterol and triglyc-
erides were mainly found in the VLDL/LDL-sized lipopro-
tein fractions and no differences were found between print-
anor, kyolic or control groups (Fig. 1). During 28 weeks of
experiment, printanor (both doses) and kyolic treatment did
not affect plasma cholesterol, triglycerides or lipoproteins.
Hence, cholesterol (and triglyceride) exposure in the garlic-
treated groups is comparable with control group (Table 2).
Printanor (both doses) and kyolic did not affect plasma ala-
nine aminotransferase (ALAT) levels during the study (Con-
trol, 112 70 U L1 ; 0.5% printanor, 93 51 U L1 ; 5%
printanor, 107 60 U L1 ; kyolic, 118 62 U L1 for week
28), indicating a comparable liver function for all experimen-
tal groups.
3.3. Inammation-related parameters
As previously shown, markers of inflammation (serum
amyloid A and serum-soluble intercellular adhesion
molecule-1) and vascular activation (plasma vWF) are good
and sensitive markers to determine the level of inflammation
or vascular activation in APOE*3-Leiden mice [17]. Print-
anor (both doses) and kyolic had no significant reducing ef-
fect on serum amyloid A (SAA), serum-soluble intercellu-
lar adhesion molecule-1 (ICAM1), and plasma vWF levels,
indicating that treatment with these garlic powders did not
modulate the inflammatory and/or vascular activation status
in these mice under Western diet feeding conditions (Table 3).
We also determined the modulatory effects of printanor
(and kyolic) on ex vivo TNF production by blood-derived
monocytes following stimulation with LPS. As illustrated in
Fig. 2, blood isolated from control APOE*3-Leiden mice fed
a high-cholesterol diet (1% cholesterol and 0.5% cholate-
containing diet) produced large amounts of TNF upon
LPS stimulation, significantly higher than mice under reg-

Table 2
Effect of printanor and kyolic on plasma cholesterol and triglycerides at 28 weeks of study under Western diet conditions. Cholesterol and triglycerides exposure
during the 28 weeks of the study
Cholesterol Triglycerides Cholesterol exposure Triglycerides exposure
(mmol/L) (mmol/L) (mmol/L weeks) (mmol/L weeks)
Control 11.4 2.5 1.0 0.3 346.3 40.5 40.4 8.9
0.5% Printanor 12.7 2.5 1.0 0.3 359.8 47.0 39.7 6.8
5% Printanor 12.6 4.4 0.9 0.2 379.1 67.3 40.5 5.6
Kyolic 10.3 2.1 0.9 0.3 321.3 32.2 37.4 7.9
Data are presented as mean S.D. for 1415 mice per group. p < 0.05 as compared with the control group.
 S.M.S.E. Santo et al. / Atherosclerosis 177 (2004) 291297 295

Fig. 2. Lipopolysaccharide (LPS)-induced tumor necrosis factor- (TNF) production by blood-derived monocytes was determined for control, 0.5% prin-
tanor-, 5% printanor-, and kyolic-treated groups (n = six mice per group) in APOE*3-Leiden mice at week 26. In addition, we included six mice on regular
chow diet (negative control) and six mice on 0.5% cholate-containing high-cholesterol diet (positive control). Significant differences as compared with control
on Western diet are indicated p < 0.05.

ular chow or Western diet feeding conditions (Fig. 2). Blood sions) to fibrous plaques with a cap and a lipid core with
drawn from Western diet fed mice produced significantly some cholesterol clefts and extracellular lipid (Type IV le-
more TNF following LPS (concentrations 2.3, 9.4, 75, and sions). In our study, qualitative assessment of atherosclerotic
150 ng mL1 of LPS) stimulation as compared with blood lesions type showed that 0.5% printanor or kyolic treatment
drawn from mice fed a regular chow diet. Although this as- had comparable percentage of mild (Type IIII) and severe
say is sufficient to detect a modulatory effect of diet with (Type IVV) lesions when compared with control mice (mild
respect to inflammation, we did not detect any effect of print- lesions: 75 12%, 72 18%, and 81 10%; severe lesions:
anor (both doses) or kyolic on LPS-induced blood TNF 25 10%, 28 7%, and 19 5% for control, 0.5% print-
production (Fig. 2). anor, and kyolic, respectively) (Fig. 3A). The 5% printanor-
treated group had a tendency towards a higher percentage of
3.4. Atherosclerosis analysis severe lesions as compared with the control group; however,
this did not reach statistical significance (38 17% versus
Printanor (both doses) and kyolic treatment did not result 25 10%, p = 0.073) (Fig. 3A). Quantitative assessment of
in any reducing effect on risk factors associated with cardio- atherosclerotic lesions area revealed that 0.5% printanor- and
vascular disease, i.e. plasma cholesterol, cholesterol expo- kyolic-treated groups had an atherosclerotic lesion area com-
sure, and markers of the level of inflammation. We next de- parable to the control-treated group (Control, 42.3 35.6
termined if printanor and kyolic treatment affects atheroscle- 103 m2 ; 0.5% printanor, 43.5 34.1 103 m2 ; ky-
rosis development in APOE*3-Leiden mice. To this end, olic, 30.5 27.7 103 m2 ) (Fig. 3B). Although the 5%
we evaluated the atherosclerotic lesion type and quantified printanor-treated group had a tendency towards a higher le-
atherosclerotic lesion area in the aortic root after 28 weeks sion area as compared with the control group (79.9 65.6
of treatment under Western diet conditions Table 3. For all 103 m2 versus 42.3 35.6 103 m2 , p = 0.227), this did
experimental groups, atherosclerotic lesions were found in not reach statistical significance (Fig. 3B). Printanor (both
the aortic intima, ranging from mild foam cell (Type III le- doses) and kyolic had comparable number of leukocytes at-
tached to the endothelium of the atherosclerotic aortic root
Table 3 (Fig. 3C). In addition, gross visual inspection of HPS-stained
Effect of printanor and kyolic on markers of inflammation (serum amyloid sections yielded no differences in saffron-positive collagen,
A and soluble intercellular adhesion molecule-1) and vascular activation
macrophage or smooth muscle content of the lesions of the
(plasma von Willebrand factor) after 28 weeks of study under Western diet
conditions garlic-treated mice. Hence, printanor (both doses) and kyolic
did not (significantly) affect atherosclerosis lesion type, area,
SAA Soluble ICAM1 vWf (%PP
(g/mL) (g/mL) rata ) or composition.
Control 6.7 6.4 35.2 5.4 117.5 46.7
0.5% Printanor 6.5 5.6 33.5 3.7 154.0 90.8
5% Printanor 4.3 3.0 35.7 5.1 142.4 88.6
Kyolic 6.4 5.4 31.7 5.3 124.2 62.2 4. Discussion
Data are presented as mean S.D. for 1415 mice per group. p < 0.05 as
compared with control group. In the present study, we evaluated the anti-atherosclerotic
a Percentage of pooled rat plasma.
properties of a chemically well-characterized and production-
296 S.M.S.E. Santo et al. / Atherosclerosis 177 (2004) 291297

Fig. 3. Twenty-eight weeks after garlic treatment, atherosclerosis was qualitatively and quantitatively assessed by determining: (A) type of lesions, (B) extent
of atherosclerosis (total lesion area), (C) number of monocytes attached to the endothelium in the control, 0.5% printanor-, 5% printanor- and kyolic-treated
groups at the level of the aortic root. For the total lesion area each data point represents the mean total lesion area per mouse.

controlled garlic powder printanor in APOE*3-Leiden trans-
genic mice. Printanor treatment did not affect plasma lipids,
markers of general inflammation and vessel wall activa-
tion, and atherosclerotic lesion type, area or composition.
Hence, under conditions relevant to the human situation, gar-
lic powder printanor does not display hypolipidemic, anti-
inflammatory or anti-atherosclerotic properties.
Treatment of mice with printanor resulted in a reduction
in body weight. Loss of body weight was independent of
food intake, and coincided with increased feces production
and fecal fatty acid excretion. The disposals of fatty acids, i.e.
energy via feces can at least partly explain why the printanor-
treated mice gain less body weight. How printanor and its
constituents affect fecal fatty acid disposal is subject to spec-
ulation, but this may be at the level of printanor-mediated
inhibition of food absorption including free fatty acids. Inter-
estingly, Elkayam et al. also observed reduced body weight
in rats when feeding the garlic-active compound allicin, sug-
gesting that garlic-active compounds are responsible for the
weight reduction [24]. As in our study, reduced body weight
gain was independent of food intake and independent of any
toxic side effects of allicin. Via an unknown mechanism, the
garlic supplementation to food may be effective in weight
control in obesity. Interestingly, since the effects were already
observed at low amounts (i.e. printanor 0.5%), this may have
implications for the human situation.
Previously, we demonstrated that garlic and garlic sulfur-
containing compounds do not display any hypolipidemic
properties as evaluated in detail in APOE*3-Leiden mice
[25], these data were confirmed in the present study (Fig. 1,
Table 2). In vitro studies, using endothelial cells showed
that high concentrations of garlic sulfur-containing com-
pounds (S-allylcysteine, aged garlic extract, and garlic pow-
der extracts) decrease cytokine production (interleukin (IL)-
1, IL-6, and TNF), indicating that garlic may have anti-
inflammatory properties [57]. In the present study, we ob-
serve that printanor and kyolic does not affect serum amyloid
A, intercellular adhesion molecule-1, lipopolysaccharide-
induced tumor necrosis factor- production by blood-derived
monocytes, or plasma von Willebrand factor. Hence, in con-
trast to the in vitro data, our in vivo data suggest that printanor
and kyolic, and possibly garlic in general, do not display anti-
inflammatory properties. The reason for this discrepancy may
be the use of unrealistic high concentrations of garlic-derived
materials in the in vitro studies, which are not achieved in vivo
even at a dosage of 50 g kg1 (5%) garlic powders in the diet.
In the present study we included kyolic (1.6 g kg1 ) as ref-
erence, since it was shown to exert an anti-atherogenic effect
at least in rabbits (64% less fatty streak-like lesions) [10]. In
our hands, kyolic treatment at the same dose had no effect
on atherosclerosis (p = 0.477, n = 15 mice). One major dif-
ference between the two studies is the type of animal model
i.e. rabbits versus APOE*3-Leiden transgenic mice and the
type of lesion, i.e. fibro foamy lesions (rabbits) versus human-
like lesions (APOE*3-Leiden mice). More importantly, to get
the rabbits hypercholesterolemic, the diet was supplemented
with a high amount of cholesterol (i.e. 1% w/w). In contrast,
APOE*3-Leiden mice easily obtain plasma cholesterol levels
of 1013 mmol/L when the diet is supplemented with only
0.2% (w/w) cholesterol. We have no mechanistic explanation
for the difference in response to kyolic between rabbits and
APOE*3-Leiden transgenic mice. However, one can ques-
 S.M.S.E. Santo et al. / Atherosclerosis 177 (2004) 291297
tion the relevance of the findings of the kyolic rabbit study
for the human situation, since a human diet contains far less
cholesterol.
In conclusion, this study shows an absence of hypolipi-
demic, anti-inflammatory and anti-atherogenic effect of the
well-characterized and production-controlled garlic powder
printanor in a humanized animal model. As such we conclude
that printanor has no beneficial effect on CVD. Several studies
have been performed showing that garlic may have beneficial
effects on risk factors associated with CVD [1,2]. However,
numerous studies, including ours, do not support these obser-
vations [1,2]. Recently, a meta-analysis of the epidemiologic
literature on the association between garlic consumption and
risk of various cancers has shown the presence of publication
bias in this field [26]. Since there is a possibility of publica-
tion bias for the current garlic and CVD topics, we propose
detailed and extensive meta-analysis of the literature on the
association between garlic and CVD. Such analysis might
help to clarify whether garlic truly has beneficial effects in
cardiovascular disease in humans.
Acknowledgements
This work was supported by a European Union research
project QLK1-CT-1999-498 and by the fellowship of the
Royal Netherlands Academy of Arts and Sciences attributed
to Dr. B.J.M. van Vlijmen. We thank all the people involved
in the EU project Garlic and Health for their collaboration
and Ria van den Hoogen for technical assistance.
References
[1] Orekov AN, Grunwald J. Effects of garlic on atherosclerosis. Nutri-
tion 1997;13:65663.
[2] Berthold HK, Sudhop T. Garlic preparations for prevention of
atherosclerosis. Curr Opin Lipidol 1998;9:5659.
[3] Brace LD. Cardiovascular benefits of garlic (Allium sativum L). J
Cardiovasc Nurs 2002;16:3349.
[4] Banerjee SK, Maulik SK. Effect of garlic on cardiovascular disor-
ders: a review. Nutr J 2002;1:114.
[5] Ide N, Lau BH. Garlic compounds minimize intracellular oxida-
tive stress and inhibit nuclear factor-kappa b activation. J Nutr
2001;131:1020S6S.
[6] Hodge G, Hodge S, Han P. Allium sativum (garlic) suppresses leuko-
cyte inflammatory cytokine production in vitro: potential therapeu-
tic use in the treatment of inflammatory bowel disease. Cytometry
2002;48:20915.
[7] Keiss HP, Dirsch VM, Hartung T, et al. Garlic (Allium sativum
L.) modulates cytokine expression in lipopolysaccharide-activated
human blood thereby inhibiting NF-kappaB activity. J Nutr
2003;133:21715.
[8] Bordia A, Verma SK. Effect of garlic feeding on regression of ex-
perimental atherosclerosis in rabbits. Artery 1980;7:42837.
297
[9] Mand JK, Gupta PP, Soni GL, Singh R. Effect of garlic on experi-
mental atherosclerosis in rabbits. Indian Heart J 1985;37:1838.
[10] Efendy JL, Simmons DL, Campbell GR, Campbell JH. The effect of
the aged garlic extract, kyolic, on the development of experimental
atherosclerosis. Atherosclerosis 1997;132:3742.
[11] Durak I, Ozturk HS, Olcay E, Can B, Kavutcu M. Effects of garlic
extract on oxidant/antioxidant status and atherosclerotic plaque for-
mation in rabbit aorta. Nutr Metab Cardiovasc Dis 2002;12:1417.
[12] Lee WM, Lee KT. Lack of effect of phenformin, pyridinolcarbamate
combined with a decalcifying agent, and an oniongarlic mixture on
growth and necrosis of aortic atherosclerosis in ballooned swine.
Exp Mol Pathol 1980;33:3458.
[13] Heinle H, Betz E. Effects of dietary garlic supplementation in a rat
model of atherosclerosis. Arzneimittelforschung 1994;44:6147.
[14] Arnault I, Christides JP, Mandon N, Haffner T, Kahane R, Auger J.
High-performance ion-pair chromatography method for simultaneous
analysis of alliin, deoxyalliin, allicin and dipeptide precursors in
garlic products using multiple mass spectrometry and UV detection.
J Chromatogr A 2003;991:6975.
[15] van Vlijmen BJ, Mensink RP, van t Hof HB, Offermans RF, Hofker
MH, Havekes LM. Effects of dietary fish oil on serum lipids and
VLDL kinetics in hyperlipidemic apolipoprotein APOE*3-Leiden
transgenic mice. J Lipid Res 1998;39:11818.
[16] Volger OL, van der Boom H, de Wit EC, et al. Dietary plant
stanol esters reduce VLDL-cholesterol secretion and bile saturation
in APOE*3-Leiden transgenic mice. Arterioscler Thromb Vasc Biol
2001;21:104652.
[17] Kleemann R, Princen HM, Emeis JJ, et al. Rosuvastatin reduces
atherosclerosis development beyond and independent of its plasma
cholesterol-lowering effect in APOE*3-Leiden transgenic mice: ev-
idence for anti-inflammatory effects of rosuvastatin. Circulation
2003;108:136874.
[18] Van Vlijmen BJ, van den Maagdenberg AM, Gijbels MJ, et al. Diet-
induced hyperlipoproteinemia and atherosclerosis in apolipoprotein
E3-Leiden transgenic mice. J Clin Invest 1994;93:140310.
[19] Post SM, De Crom R, Van Haperen R, Van Tol A, Princen HM.
Increased fecal bile acid excretion in transgenic mice with ele-
vated expression of human phospholipid transfer protein. Arterioscler
Thromb Vasc Biol 2003;23:8927.
[20] Tranquille N, Emeis JJ. The simultaneous acute release of tissue-type
plasminogen activator and von Willebrand factor in the perfused rat
hindleg region. Thromb Haemost 1990;63:4548.
[21] Von Der Thusen JH, Kuiper J, Fekkes ML, et al. Attenuation of
atherogenesis by systemic and local adenovirus-mediated gene trans-
fer of interleukin-10 in LDLr/ mice. FASEB J 2001;15:27302.
[22] Delsing DJ, Offerman EH, van Duyvenvoorde W, et al. Acyl-CoA:
cholesterol acyltransferase inhibitor avasimibe reduces atherosclero-
sis in addition to its cholesterol-lowering effect in ApoE*3-Leiden
mice. Circulation 2001;103:177886.
[23] Volger OL, Mensink RP, Plat J, et al. Dietary vegetable oil and
wood derived plant stanol esters reduce atherosclerotic lesion size
and severity in apoE*3-Leiden transgenic mice. Atherosclerosis
2001;157:37581.
[24] Elkayam A, Mirelman D, Peleg E, et al. The effects of allicin on
weight in fructose-induced hyperinsulinemic, hyperlipidemic, hyper-
tensive rats. Am J Hypertens 2003;16:10536.
[25] Espirito Santo SM, Van Vlijmen BJ, Buytenhek R, et al. Well-
characterized garlic-derived materials are not hypolipidemic in
APOE*3-Leiden transgenic mice. J Nutr 2004;134:15003.
[26] Fleischauer AT, Poole C, Arab L. Garlic consumption and cancer
prevention: meta-analyses of colorectal and stomach cancers. Am J
Clin Nutr 2000;72:104753.