Nutrient Metabolism

The Bioavailability of Ferulic Acid Is Governed Primarily by the Food Matrix

Rather than Its Metabolism in Intestine and Liver in Rats1
Aline Adam,2 Vanessa Crespy,* Marie-Anne Levrat-Verny,* Fanny Leenhardt,*
Michel Leuillet, Christian Demigné* and Christian Rémésy*
Institut Technique des Céréales et des Fourrages (ITCF), Laboratoire Qualité des Céréales,

Downloaded from https://academic.oup.com/jn/article-abstract/132/7/1962/4687425 by guest on 22 December 2019

75013 Paris, France and *Laboratoire Maladies Métaboliques et Micronutriments,
Centre de Recherches en Nutrition Humaine Auvergne, I.N.R.A. Clermond-Ferrand/Theix,
F-63122 St-Genès-Champanelle, France

ABSTRACT The physiologic importance of ferulic acid (FA), and notably its antioxidant properties, depends upon
its availability for absorption and subsequent interaction with target tissues. Because FA is widely present in
cereals, the aim of the present study was to investigate its intestinal and hepatic metabolism in rats by in situ
intestinal perfusion model (from 10 to 50 nmol/min), and its bioavailability in supplemented diets (from 10 to 250
␮mol/d) or in a complex cereal matrix, i.e., whole flours from Valoris (Triticum aestivum) or Duriac (T. durum)
cultivars and bran or white flour from the Valoris cultivar. In perfused rat intestine, net FA absorption was
proportional to the perfused dose (R2 ⫽ 0.997); once absorbed, FA was completely recovered as conjugated forms
in plasma and bile secretion (representing 5–7% of the perfused dose). In rats fed FA-enriched semipurified diets,
FA absorption was quite efficient because ⬃50% of the ingested dose was recovered in urine. This extensive
elimination by kidneys limited FA accumulation in plasma (typically 1 ␮mol/L in rats fed 50 ␮mol FA/d). In contrast,
in rats fed cereal diets providing 56 – 81 ␮mol FA/d, urine excretion was 90 –95% lower than in rats fed FA-enriched
semipurified diets, and plasma concentrations were ⬃0.2– 0.3 ␮mol/L. Thus, the cereal matrix appears to severely
limit FA bioavailability. This inherently low bioavailability of FA in cereals likely reflects FA association with the fiber
fraction through cross-linking with arabinoxylans and lignins. J. Nutr. 132: 1962–1968, 2002.

KEY WORDS: ● ferulic acid ● bioavailability ● absorption ● cereals ● rats

Prospective studies have found that the habitual intake of
whole-grain foods is associated with reduced coronary heart
disease, total cancer mortality (1,2), incidence of diabetes (3)
and coronary heart disease mortality (4 –7). More specifically,
intake of dietary fiber, especially from grain sources, has also
been linked to reduced risk of coronary heart disease (8) and
diabetes (3). Whole-grain foods contain a variety of biologi-
cally active constituents such as vitamin B, vitamin E, sele-
nium, zinc, copper, magnesium and phytochemicals such as
phenolic compounds, which may contribute synergistically to
the health effects of plant foods (9). Hydroxycinnamic acid
derivatives such as p-coumaric and ferulic acid are present in
cereal cell walls; ferulic acid (FA)3 is the most common
phenolic acid in monocotyledonous cell walls (10). In wheat,
this acid is ester-linked to position O-5 of the arabinosyl side
chains of cell wall arabinoxylans, and occurs in high concen-
trations in the aleurone, pericarp and embryo cells walls, but
only in trace amounts in the starchy endosperm of ripe kernels
1
Supported by A.N.R.T. (Association Nationale pour la Recherche et Tech-
nique), I.N.R.A. (Institut National de la Recherche Agronomique), Univers Céréales
and I.T.C.F. (Institut Technique des Céréales et des Fourrages).
2
To whom correspondence should be addressed.
E-mail: aadam@clermont.inra.fr.
3
Abbreviations used: Bv, bran Valoris diet; FA, ferulic acid; RS, resistant
starch; Valoris, whole wheat flour diet; White Fv, white flour Valoris diet.
(11). The trans-isomer predominates and accounts for 90% of
the total phenolic acids in common flour (12). There was a
very high genetic variability in the ferulic acid contents of the
durum wheat varieties studied. The concentrations ranged
from 0.7 to 2.4 mg/g dry matter, and the levels were higher
than those reported for common wheat, which varies from 0.5
to 1 mg/g dry matter (13). Thus, regular consumption of whole
cereals, for example whole bread, may result in ingestion of
large amounts of ferulic acid.
Ferulic acid has a high antioxidant potential due to its
resonance-stabilized phenoxy radical structure. It is an effec-
tive scavenger of free radicals and it has been approved in
certain countries as a food additive to prevent lipid peroxida-
tion (14). In vitro, it also protects LDL from metmyoglobin-
induced oxidative damage (15). A large proportion of ferulate
dimers are present in a range of plant materials, notably in
wheat bran (5– 8-BendiFA, 8-O-4-diFA and 5–5-diFA). In
vitro, the antioxidant properties of two chemically synthesized
ferulate dimers, 5–5-diFA and 5– 8-BendiFA, have been doc-
umented (16). In the aqueous phase, both dimers were less
effective antioxidants than ferulic acid, although they were
more efficient at inhibiting lipid peroxidation. Moreover, FA
as a free acid had less suppressive effect on LDL oxidation than
ferulic acid sugar esters (feruloyl arabinose) (17).
The physiologic importance of FA, and notably its anti-

0022-3166/02 $3.00 © 2002 American Society for Nutritional Sciences.

Manuscript received 11 February 2002. Initial review completed 10 March 2002. Revision accepted 8 April 2002.

1962
 BIOAVAILABILITY OF FERULIC ACID IN RATS 1963

oxidant property, depends upon its availability for absorption
and subsequent interaction with target tissues. Therefore, it
is important to estimate the bioavailability of this com-
pound to appreciate more fully its real physiologic effect.
Very few studies have been undertaken concerning the
uptake of FA from the diet. Choudhury et al. (18) showed
in rats that 10% of the FA ingested was recovered in urine
24 h after the beginning of gavage, and this urinary excretion
was on the same order as that of humans who had consumed
tomatoes (19).
Because FA is widely present in cereals, the aim of the
present study was to investigate in rats first its intestinal and
hepatic metabolism by an in situ intestinal perfusion model
Animals and diets. Wistar rats (IFFA-CREDO, l’Arbresle,
France), weighing ⬃150 g, were housed, two per cage, in tempera-
ture-controlled rooms (22°C), with a dark period from 800 to 2000 h
and access to food from 800 to 1600 h. For the intestinal perfusion
experiment, rats were fed a standard semipurified diet [73% wheat
starch, 15% casein, 3.5% mineral mixture (AIN 93M formula), 1%
vitamin mixture (AIN 76A formula), 5% corn oil] for 2 wk. The rats
were maintained and handled according to the recommendations of
the Institut National de la Recherche Agronomique Ethics Commit-
tee, in accordance with decree no. 87– 848.
To study the bioavailability of FA in cereals and in FA-supple-
mented diets, rats were fed one of the experimental semipurified diets
distributed as a moistened powder for 21 d (Table 1). The composi-
tion of the purified diets was the same as that of the control diet but

Downloaded from https://academic.oup.com/jn/article-abstract/132/7/1962/4687425 by guest on 22 December 2019

(from 10 to 50 nmol/min), and then its bioavailability in
supplemented diets (from 10 to 250 ␮mol/d) compared with
those in a complex cereal matrix.
MATERIALS AND METHODS
Chemicals. Ferulic acid was purchased from Extrasynthese
(Genay, France). ␤-Glucuronidase/sulfatase from Helix pomatia were
purchased from Sigma (L’Isle D’Abeau, Chesnes, France). Mineral
(AIN 93M) and vitamin (AIN 76A) mixes incorporated into the
diets for the intestinal perfusion experiment were purchased from
ICN (Orcay, France).
Whole wheat flour and milled fractions. Valoris was provided by
ITCF (Institut Technique des Céréales et des Fourrages, Paris,
France). The milled fractions were obtained after the processing of 8
kg of whole grain wheat in a Bühler mill: 22% bran ⫹ remilling, 78%
white flour (0.55 g/100 g of ash). The whole wheat flour diet (Valoris)
was reconstituted with 54.7% white flour and 15.3% bran ⫹ remill-
ing. The white flour diet (White Fv) comprised 54.7% white flour.
The wheat bran diet (Bv) was made up of 15.3% bran ⫹ remilling.
Duriac was provided by INRA Montpellier (Institut National de la
Recherche Agronomique). The milling process was performed and
the milled fractions were blended to produce the whole wheat flour.
enriched in ferulic acid at different concentrations, so that rats
consumed 10, 50 and 250 ␮mol/d ferulic acid (FA 10, FA 50 and FA
250 diets, respectively). The body weight of rats was recorded twice
per week during the experimental period. During the last 7 d of the
experimental period, rats were transferred to metabolic cages and food
intake and fecal excretion were recorded over the last 4 d of the
experiment.
Sampling procedure. For the intestinal perfusion, rats were anes-
thetized with sodium pentobarbital (40 mg/kg body) 18 h after the
beginning of the meal and maintained alive during the perfusion
period. In all experiments, a cannulation of the biliary duct was
performed to test the enterohepatic cycling of ferulic acid. After
cannulation of the biliary duct, a perfusion of jejunal ⫹ ileal segment
of intestine (from 5 cm distal from the flexura duodenojejununalis to
the valvula ileocoecalis) was prepared by installing cannulae at each
extremity. This segment was continuously perfused in situ with a
buffer containing KH2PO4 (5 mmol/L), K2HPO4 (2.5 mmol/L),
NaHCO3 (5 mmol/L), NaCl (50 mmol/L), KCl (40mmol/L), potas-
sium tricitrate (10 mmol/L), CaCl2 (2mmol/L), MgCl2 (2 mmol/L),
pH 6.7, glucose (8 mmol/L) and taurocholic acid (1 mmol/L) at a flow
rate of 1 mL/min, at 37°C and supplemented with 10, 20 and 50
␮mol/L ferulic acid. Aliquots of effluent were collected directly at the
exit of the ileum into plastic tubes (1.5 mL) during the last 5 min of

TABLE 1
Composition of the diets1

Control diet Duriac Valoris White Fv Bv Bv ⫹ RS

g/kg

Digestible wheat starch 717.5 175 175 305 619 469

Potato starch — — — — — 150
Casein 75 75 75 75 75 75
Wheat gluten 75 — — 21 52 52
Mineral mix2 26.3 11.4 17.9 24.9 15.5 15.1
Vitamin mix3 10 8 8 10 8 8
Peanut oil 50 35 35 44 43 43
Whole cereal — 700 700 — — —
White flour — — — 547 — —
Bran — — — — 153 153
Energy, kJ/g 16.4 14.2 14.2 16.0 15.2 15.2

1 The 6 experimental diets consisted of whole wheat flour from Valoris (Triticum aestivum) and Duriac (Triticum durum) cultivars and different
milling fractions of wheat Valoris [white flour (White Fv) and bran (Bv)]. One of the experimental diets was composed of wheat bran Valoris (Bv) with
resistant starch (RS) from potatoes. The Valoris diet had the same quantity of bran as the Bv diet, and the same quantity of white flour as the White
Fv diet. All of the diets were equal in carbohydrate concentration (⬃70%); only the fiber content varied. For the control, Duriac, Valoris, White Fv, Bv
and Bv ⫹ RS diets, the starch concentrations were 71.75, 59.2, 59.2, 69.5, 69.5 and 65%, respectively. The protein concentrations were ⬃15%, and
lipids were ⬃5% for all the diets. The bioavailability of ferulic acid from cereals was compared with the bioavailability of ferulic acid added to a
semisynthetic diet. The composition of the semisynthetic diet was the same as that of the control diet but enriched in ferulic acid at different
concentrations, so that rats consumed 10, 50 and 250 ␮mol/d ferulic acid (FA 10, FA 50 and FA 250, respectively).
2 All diets contained (per kg diet): 4 g Ca, 1.2 g Mg, 5 mg Cu, 36 mg Fe, 38 mg Zn. The mineral content of all diets was determined before the
beginning of the experiment.
3 Vitamin supplied (mg/kg control diet): thiamin, 15; riboflavin, 20; pyridoxine, 10; nicotinamide, 100; pantothenate, 70; folic acid, 5; biotin, 0.3;
cyanocobalamin; 0.05; retinyl palmitate, 1.5; dl-␣-tocopheryl acetate, 125; cholecalciferol, 0.15; menadione, 1.5; ascorbic acid, 50; myo-inositol, 100;
choline, 1.36 g provided by UAR (Villemoisson, Epinay-sur-Orge, France). For the experimental diets, the vitamin concentration of the whole flour was
considered; therefore the vitamin mix supply was reduced to 3 g/kg instead of 10 g/kg.
1964 ADAM ET AL.
perfusion. All volumes were recorded and concentrations were cor-
rected for intestinal absorption of water. Bile was collected through-
out the perfusion step. At the end of the experiment, blood samples
were withdrawn from the abdominal aorta into heparinized tubes.
Plasma, bile and perfusate samples were acidified with 10 mmol/L
acetic acid and then stored at ⫺20°C.
To study the bioavailability of ferulic acid in cereals, rats were
killed at the end of the dark period, when cecal fermentations are still
very active. They were first anesthetized with sodium pentobarbital
(40 mg/kg body) and maintained at 37°C. An abdominal incision was
made and blood was withdrawn from the abdominal aorta (5 mL) into
heparinized tubes. After centrifugation at 10,000 ⫻ g for 5 min, the
plasma was collected, acidified with acetic acid (10 mmol/L) and
stored at ⫺20°C for FA analysis. After blood sampling, the cecum
with its contents was removed and weighed and two samples of cecal
contents were transferred to microfuge tubes and immediately frozen
at ⫺20°C.
Sample treatment. Bile, plasma, urine and perfusate samples were
acidified (to pH 4.9) with 0.1 volume of 0.58 mol/L acetic acid. The
samples were treated for 30 min at 37°C in the absence (unconju-
gated forms) or in the presence (total forms) of 5 ⫻ 106 units/L
␤-glucuronidase and 2.5 ⫻ 105 units/L sulfatase. The reactions were
stopped by the addition of 2.85 volumes of methanol/HCl (200
mmol/L) and the resulting mixtures were centrifuged for 4 min at
14,000 ⫻ g. After this extraction step, 20 ␮L of supernatant was
injected and analyzed by HPLC.
Cereals and fecal samples (200 mg) were saponified for 2 h in the
dark with 10 mL of 2 mol/L NaOH at 35°C (with agitation, in the
presence of N2). 3,4,5 Trimethoxy-trans-cinnamic acid was then
added as an internal standard and the solutions were adjusted to pH
2.0 with 4 mol/L HCl. The phenolic acids were extracted once with
ethyl acetate (5 mL) under agitation and 5 min centrifugation at
3800 ⫻ g. The ether phases were collected in amber test tubes and
evaporated completely in the presence of N2. The dry extracts were
dissolved in 2 mL of methanol. A quantity of the supernatant (20 ␮L)
was injected and analyzed by HPLC.
Chromatographic conditions. The HPLC system used consisted
of an autosampler (Kontron, 360), a UV Detector (set at 320 nm)
and a Software system for data recording and processing. The system
was fitted with a 5-␮m C-18 Hypersil BDS analytical column (150
⫻ 4.6 mm; Life Sciences International, Cergy, France). The mobile
phase consisted of 50 mmol/L CH3COONa (pH 4) (solvent A) and
acetonitrile (solvent B). The chromatographic conditions were as
follows (flow rate 1 mL/min): 0 –3 min: solvent A 85%/solvent B
15%; 3–33 min: solvent A 85%/solvent B 15% 3 solvent A 65%/
solvent B 35%; 33–34 min: solvent A 40%/solvent B 60%; 34 –36
min: solvent A 40%/solvent B 60%; 36 –37 min: return to initial
mobile phase conditions, then equilibration for 3 min.
The plasma samples obtained during the study of ferulic acid
bioavailability in cereals and supplemented diets were analyzed using
a multielectrode coulometric detection (4-electrodes CoulArray, Eu-
rosep, France) with potentials set at 150, 270, 370 and 450 mV.
Statistics. Values are means ⫾ SEM, and the differences between
values were determined by the Kruskal-Wallis test (nonparametric
ANOVA), and Dunn’s Multiple Comparisons Test was used to sep-
arate means.
RESULTS
Intestinal and hepatic metabolism of ferulic acid. Three
different concentrations of FA, ranging from 10 to 50 ␮mol/L,
were perfused through the small intestine in situ, and the
absorption of this compound was directly proportional to the
amount perfused (Fig. 1; Table 2). The percentage absorbed,
corresponding to the net absorption, did not differ at any of
the concentrations (56.1 ⫾ 2.3% of the perfused dose). More-
over, the concentration of ferulic acid recovered in the efflu-
ent was not affected by the enzymatic treatment, suggesting
that this compound was present only as its native form at the
end of the perfusion step.
The HPLC-UV analysis of bile revealed that FA was
FIGURE 1 Absorption of ferulic acid (FA) in the small intestine of
rats after in situ perfusion of three concentrations of this compound for
Downloaded from https://academic.oup.com/jn/article-abstract/132/7/1962/4687425 by guest on 22 December 2019
30 min. Values are means ⫾ SEM, n ⫽ 5.
present only as conjugated forms (glucuronidated and/or sul-
fated forms), and the biliary secretion of these conjugated
forms was quite noticeable at the end of perfusion (Table 2).
Whatever the perfused dose of ferulic acid, the percentage of
biliary secretion was equivalent and represented 5–7% of the
perfused dose.
At the end of the perfusion period, in nonhydrolyzed
plasma, no trace of unconjugated FA was found, indicating
that the circulating forms of this compound were glucurono-
and/or sulfoconjugates. However, these FA metabolites were
not detected in nonhydrolyzed samples, probably because their
response to UV detection was too low under the present
chromatographic conditions. The concentration of FA in hy-
drolyzed plasma depended on the perfused dose and increased
as a function of the perfused dose (Table 2).
A fraction of the ferulic acid perfused dose was not recov-
ered in the effluent and in the bile at the end of the perfusion
step. This unrecovered part could correspond to the conju-
gated ferulic acid distributed in different organs; as a conse-
quence, it could have a potential biological effect. By calcu-
lating the difference between these total outputs from the
perfused intestine (nonabsorbed FA and biliary secretion) and
the perfused dose in the intestine, we estimated the proportion
of conjugated forms available for peripheral tissues (Fig. 2;
Table 2). This percentage of conjugated forms available for
peripheral tissues represented 45–53% of the perfused dose.
Bioavailability of FA in semipurified diet. The level of FA
in feces was negligible whatever the dose of FA ingested
(Table 3). This result suggests that this compound was either
highly absorbed by the gastrointestinal tract and/or extensively
catabolized by the microflora present in the cecum.
In plasma, FA was detected only after enzymatic treatment,
suggesting that FA circulated only as the conjugated forms.
When rats were fed a 0.009% FA diet (10.0 ⫾ 0.3 ␮mol/d),
this compound was not detectable in plasma 18 h after the
beginning of the meal. This suggests that FA is either poorly
absorbed or absorbed but rapidly eliminated in urine. Never-
theless, in rats fed the FA 50 and 250 diets, the plasma
concentration of this compound was also significantly higher
(Table 3).
FA urinary excretion was proportional to the ingested dose
(R2 ⫽ 0.995) (Table 3). This observation supports the view
that this compound was quite highly absorbed. Moreover,
because the plasma concentrations were relatively low or even
undetectable 18 h after the beginning of the meal, it is likely
that absorbed FA was readily eliminated in urine.
Bioavailability of FA in cereals. The presence of whole
cereals or cereal fractions in the diet did not affect daily food
intake or weight gain (Table 4). Rats fed cereal diets had
significantly greater fecal excretions than controls (P ⬍ 0.001)
 BIOAVAILABILITY OF FERULIC ACID IN RATS 1965

TABLE 2
Calculations of ferulic acid (FA) absorption and secretion flux related to the FA intestinal perfusion dose in rats1,2,3

Net Flux of conjugated forms

Perfused dose of FA abosrption Biliary flux available for peripheral tissues Plasma concentration

nmol/min ␮mol/L

10 4.75 ⫾ 0.70 0.58 ⫾ 0.15 4.19 ⫾ 0.70 0.79 ⫾ 0.30

20 10.65 ⫾ 0.45 0.98 ⫾ 0.10 9.94 ⫾ 0.50 2.17 ⫾ 0.40
50 28.50 ⫾ 2.05 3.35 ⫾ 0.90 25.20 ⫾ 2.95 2.90 ⫾ 0.30

1 Values are means ⫾ SEM, n ⫽ 5.

2 Values in a column not sharing a superscript differ, P ⬍ 0.05.

Downloaded from https://academic.oup.com/jn/article-abstract/132/7/1962/4687425 by guest on 22 December 2019

3 Calculations of the different fluxes are presented below:
Flux of nonabsorbed aglycone ⫽ flux of aglycone recovered unchanged in the effluent at the end of the perfusion.
Flux of secretion of conjugated forms in the effluent ⫽ (flux of the aglycone in hydrolyzed effluent) ⫺ (flux of the aglycone in nonhydrolyzed effluent).
In this case, no difference was observed in the ferulic acid concentrations before and after enzymatic hydrolysis.
Flux of net absorption ⫽ (flux of perfused dose) ⫺ (flux of aglycone recovered unchanged in the effluent at the end of the perfusion).
Flux of conjugated forms available for tissues ⫽ flux of perfused dose ⫺ [(flux of aglycone recovered unchanged in the effluent at the end of the
perfusion) ⫺ (biliary secretion)].

or rats fed FA enriched diets, except rats fed White Fv. All rats
fed cereal diets had enlarged ceca compared with controls.
However, only rats fed Bv ⫹ resistant starch (RS) had signif-
icantly heavier ceca than controls. In the same way, these rats
and rats fed White Fv had greater short-chain fatty acid
fermentation in the cecum than controls and Bv-fed rats.
When rats were fed Duriac wheat (FA intake close to that
in rats fed FA 50), the fecal excretion of this phenolic acid was
significantly enhanced (Table 3). This suggests that the ab-
sorption of this compound is depressed when it is present in a
complex matrix, such as that of cereals. This observation is
confirmed by the fact that its urinary excretion was also
significantly lower than that in rats fed FA-enriched diets (FA
10, FA 50 and FA 250) and its plasma FA concentration was
undetectable. Moreover, the percentage of FA recovered in
urine in relation to the ingested dose was significantly lower in
rats fed the Duriac diet than in rats fed the FA-enriched
semipurified diets. Whatever the type of wheat (Duriac and
Valoris) in the diet, this FA recovery in urine, in relation to
the ingested dose, was significantly lower in rats fed the diets
enriched in FA.
In the Valoris cultivar, compared with whole wheat flour,
90% of FA was found in the bran (Table 5). In rats fed whole
wheat Valoris and bran (Bv diet), the fecal excretion of FA
was high and the urinary excretion was low. The percentage of
FA recovered in urine, in relation to the ingested dose, was not
different between Valoris- and Bv-fed groups (Table 5). The
addition of RS to the Bv diet did not contribute to an increase
in the fecal excretion of FA. The ingested FA in rats fed white
Fv was significantly lower than in rats fed Valoris or Bv diets,
which explains the low fecal and urinary excretion in rats fed
White Fv (Table 5). However, the urinary excretion in rela-
tion to the ingested dose was significantly higher (Table 5).

FIGURE 2 Recapitulative schematic of the fate of ferulic acid (FA)
at the splanchnic level. The diagram shows the FA percentage at the
splanchnic level after perfusion of the jejuno-ileal segment by FA. Under
these conditions, 56.1% of perfused FA enters the enterocytes by an as
yet unidentified mechanism. In these cells, FA is readily conjugated and
the resulting metabolites can leave the intestinal cells only toward the
serosal side because no conjugated forms of FA are detected in the
intestinal lumen. Under such conditions, 56.1% of perfused FA, corre-
sponding to the net absorption, is recovered in the plasma of the
mesenteric vein as conjugated derivatives. A part of these conjugates
enters into the hepatocytes and is secreted in the bile (6%). Under
these conditions, 49.9% of the perfused dose is distributed in the
peripheral tissues and may have biological effects.
DISCUSSION
The aim of this study was to investigate the relative bio-
availability of FA in a supplemented diet and in a complex
cereal matrix in which it is a principal component governing
cell wall integrity, shape and defense against pathogenic in-
gress. In cereals, ferulates and dihydrodiferulates may play an
important role in dietary fiber by influencing the chemical
structure of its components.
Some biological effects have been described in the litera-
ture. It has been shown that FA has strong anti-inflammatory
properties (20), inhibits chemically induced carcinogenesis in
rats and inhibits tumor promotion in mouse skin (21,22).
Oxidation of LDL is important in the pathogenesis of athero-
sclerosis, and recent studies have reported that FA plays a role
as an antioxidant, inhibiting lipid peroxidation (23) and LDL
oxidation (17), and scavenging oxygen radicals (14). Thus, FA
could be useful in the prevention of cardiovascular diseases
and cancers. Nevertheless, its biological properties depend on
its bioavailability, notably in a complex matrix such as cereals.
Because such foods certainly influence polyphenol bioavail-
ability, it is necessary to assess its actual bioavailability in
experimental diets supplemented with FA.
An in situ model of intestinal perfusion was used to deter-
1966 ADAM ET AL.

TABLE 3
Plasma concentrations of ferulic acid (FA) and fecal and urinary recovery of FA after ingestion of Duriac
and supplemented diets by rats1,2

Ingested FA Fecal excretion Urine excretion Urine excretion Plasma concentration

␮mol/d % of intake ␮mol/L

Control 5.3 ⫾ 0.4 0.20 ⫾ 0.0 ND ND ND

FA 103 15.0 ⫾ 0.3 0.52 ⫾ 0.04 6.4 ⫾ 0.5 42.5 ⫾ 2.6 ND
FA 503 62.7 ⫾ 1.3 0.50 ⫾ 0.05 32.2 ⫾ 2.2 51.8 ⫾ 3.1 1.1 ⫾ 0.4
FA 2503 326.4 ⫾ 5.4 0.55 ⫾ 0.07 126.6 ⫾ 4.0 38.6 ⫾ 0.4 7.6 ⫾ 1.9
Duriac 55.7 ⫾ 2.0 15.8 ⫾ 0.6 1.8 ⫾ 0.1 3.2 ⫾ 0.3 ND

Downloaded from https://academic.oup.com/jn/article-abstract/132/7/1962/4687425 by guest on 22 December 2019

1 Values are means ⫾ SEM, n ⫽ 8.
2 Values in a column not sharing a superscript differ, P ⬍ 0.05.
3 The composition of the FA-enriched semisynthetic diets was the same as that of the control diet (for the composition, see Table 1), but enriched
at different FA concentrations, so that rats consumed 10, 50 and 250 ␮mol/d ferulic acid (FA 10, FA 50 and FA 250 respectively).

mine whether the intestinal or hepatic metabolism of FA were
limiting steps for its bioavailability. The effect of FA dose on
absorption was investigated by perfusion at three different
concentrations. FA absorption was directly proportional to the
perfused concentration. This suggests that FA could be ab-
sorbed by passive diffusion or by a facilitated transport which
appears not to be saturated even at a luminal concentration of
50 ␮mol/L. A comparative study showed that cinnamate up-
take is a linear function of substrate concentrations between
25 and 500 ␮mol/L (24). This result is consistent with our
findings that FA absorption is directly proportional to the
perfused concentration (10 –50 ␮mol/L). However, at higher
cinnamate concentrations (0.5–5 mmol/L), absorption of this
compound could be described by a Michaëlis-Menten type
equation with a saturable transport component, and in partic-
ular, by a Na⫹/dependent carrier-mediated transport process
(24). These authors observed a competition between cin-
namate and FA for absorption and concluded that compounds
structurally related to FA are absorbed across the brush border
membrane of rat jejunum by the same transporter. The ab-
sorption of FA seems to depend on a Na⫹/dependent carrier-
mediated transport process that was not saturated at 50
␮mol/L; however, further experiments are required to firmly
identify the transporter involved (24).
In the perfused isolated intestine, Spencer et al. (25)
showed that FA was present on the serosal side as glucu-
ronidated FA, suggesting that this compound could be metab-
olized by intestinal conjugation enzymes. This conjugation
process seems to be common to most polyphenols (26 –29),
and flavonoids in particular. Using the same intestinal model,
67% of quercetin aglycone was shown to be absorbed and
conjugated by the intestine, and 15% of quercetin conjugates
secreted toward the serosal side (28). Under these conditions,
the majority of absorbed quercetin was then reexcreted as
conjugates into the intestinal lumen (28). Intestinal secretion
of conjugated metabolites may constitute an important step for
control of polyphenol availability, but our results suggest that
this efflux of FA conjugated forms does not occur. Under these
conditions, all of the absorbed FA was secreted toward the
serosal side (Fig. 2). This could be due to the nature of the
conjugates synthesized in the enterocytes, which could influ-
ence the channeling of the conjugates, i.e., secreted toward
either the serosal side or the mucosal side.
In bile and in aorta plasma, only FA could be detected after
enzymatic hydrolysis, suggesting that this compound was
present only as the conjugated form. Nevertheless, it is con-
ceivable that FA was also present as the unconjugated form in
these biological fluids, but at concentrations below the limit of
detection (0.5 ␮mol/L).
Whatever the flux of absorbed FA, ⬃6% of the perfused

TABLE 4
Effects of whole wheat flour and its constitutive fractions on growth of rats and on the cecal variables1,2

Cecum Feces

Diet Food intake3 WEight gain3 Total weight4 Wall weight4 pH4 SCFA4 Fecal excretion3

g/d g/d g g ␮mol/cecum g DM/d

Control 17.7 ⫾ 1.2 5.8 ⫾ 0.6 2.33 ⫾ 0.21a 0.77 ⫾ 0.06a 7.04 ⫾ 0.05 175 ⫾ 27 0.5 ⫾ 0.04a
Duriac5 18.6 ⫾ 0.7 5.9 ⫾ 0.3 2.86 ⫾ 0.37a 0.91 ⫾ 0.07a 6.53 ⫾ 0.05 288 ⫾ 44 2.6 ⫾ 0.06bc
Valoris5 19.3 ⫾ 0.7 6.0 ⫾ 0.4 3.57 ⫾ 0.24a,b 1.02 ⫾ 0.04a,b 6.14 ⫾ 0.07 451 ⫾ 45 1.7 ⫾ 0.1b
White Fv5 19.3 ⫾ 0.9 6.2 ⫾ 0.5 3.35 ⫾ 0.20a,b 0.94 ⫾ 0.04a 6.30 ⫾ 0.08 403 ⫾ 35 0.7 ⫾ 0.05a,b
Bv5 18.8 ⫾ 1.0 5.8 ⫾ 0.6 2.91 ⫾ 0.14a 0.90 ⫾ 0.03a 6.75 ⫾ 0.04 263 ⫾ 19 1.7 ⫾ 0.1b
Bv ⫹ RS5 20.3 ⫾ 1.1 6.4 ⫾ 0.3 5.04 ⫾ 0.49b 1.36 ⫾ 0.11b 6.03 ⫾ 0.04 488 ⫾ 27 3.3 ⫾ 0.1c

1 Values are means ⫾ SEM, n ⫽ 8.

2 Values in a column not sharing a superscript differ, P ⬍ 0.05.
3 Variables measured during the study; DM, dry matter.
4 Variables measured postmortem; SCFA, short-chain fatty acids.
5 See Table 1 for diet composition.
 BIOAVAILABILITY OF FERULIC ACID IN RATS
TABLE 5
Plasma concentrations of ferulic acid (FA) and fecal and urinary recovery of FA after ingestion of cereals meals by rats1,2
Ingested FA Fecal excretion
␮mol/d
Control 5.3 ⫾ 0.4a 0.2 ⫾ 0.0a
Valoris3 81.2 ⫾ 3.3b 16.8 ⫾ 1.6b
White Fv3 5.0 ⫾ 0.2a 0.3 ⫾ 0.0a
Bv3 73.7 ⫾ 4.0b 28.3 ⫾ 1.9b
Bv ⫹ RS3 74.7 ⫾ 4.0b 23.9 ⫾ 1.4b
1 Values are means ⫾ SEM, n ⫽ 8.
2 Values in a column not sharing a superscript differ, P ⬍ 0.05.
3 See Table 1 for diet composition.
dose was recovered in bile (Fig. 2). This suggests that the
biliary secretion was proportional to the perfused dose and not
saturated. Under these conditions, the fraction available for
the tissues was ⬃50%, whatever the dose absorbed (Fig. 2).
This value is quite substantial because, for example, only 9%
of perfused quercetin was found available for these tissues (28).
The fractional availability of FA is relatively important, sug-
gesting that the intestinal and hepatic metabolism of this
compound is not a limiting step for its bioavailability.
The bioavailability of FA was studied in rats fed a diet
supplemented with FA for 3 wk to achieve a daily intake of 10,
50 or 320 ␮mol FA/d. The control diet contained 0.003% FA,
present in the purified wheat starch (Table 3). In plasma, FA
was present only as conjugated forms. This result is in accor-
dance with Azuma et al. (30), showing that phenolic acids,
such as caffeic and chlorogenic acids are present in plasma as
glucuronided and/or sulfated forms, with only a minor part of
these compounds identified as the unconjugated form. We did
not detect unconjugated FA, possibly because they were
present in concentrations below of the detection limit. How-
ever, Azuma et al. (30) used a procedure (gavage) delivering a
large amount of compound in a short time, leading to a direct
diffusion of the compound administered through the intestinal
wall; this could explain the presence of unconjugated caffeic
acid in plasma.
In the present work (postabsorptive conditions), the plasma
concentrations of FA were very low in rats fed FA diets.
Azuma et al. (30) showed that the Tmax of caffeic acid was 2 h
and rapidly returned to undetectable values within 6 h after
gavage. This observation suggests that the enterohepatic cycle
did not allow the plasma concentrations of this compound to
remain high. This agrees with the result obtained by in situ
intestinal perfusion of FA because only 6.2 ⫾ 0.9% of the
perfused dose was secreted into the bile (Fig. 2).
The urinary excretion of FA was between 38 and 51% of
the ingested dose. When rats were fed 15 ␮mol FA/d, the
plasma concentration of this compound was not detectable
18 h after the beginning of the meal, suggesting that absorbed
FA was completely eliminated in urine. These results are
consistent with those found in in situ intestinal perfusion
because 50% of the perfused dose was available for peripheral
tissues and thus for urinary excretion. This substantial recov-
ery is noteworthy compared with that found by Choudhury et
al. (18), showing that only 10.5% of FA ingested was present
in urine 24 h after the beginning of the meal. Despite the
substantial urinary recovery of FA, a part of the ingested dose
of this compound was not recovered because a weak fraction
was detected in the feces. This unrecovered fraction could be
due to the degradation of FA by the microflora. Decarboxyl-
1967
Urine excretion Urine excretion Plasma concentration
% of intake ␮mol/L
ND ND ND
3.4 ⫾ 0.2b 3.6 ⫾ 0.6a 0.31 ⫾ 0.05
0.4 ⫾ 0.0a 8.4 ⫾ 0.9b ND
2.9 ⫾ 0.1a,b 3.9 ⫾ 0.1a 0.22 ⫾ 0.03
4.1 ⫾ 0.4b 5.5 ⫾ 0.4b 0.27 ⫾ 0.02
Downloaded from https://academic.oup.com/jn/article-abstract/132/7/1962/4687425 by guest on 22 December 2019
ation of phenolic acids occurs readily when they ware incu-
bated anaerobically with extracts of cecum or colon contents
or feces (31).
When FA was ingested in whole cereal (1.2 mg/g) or bran
(4.6 mg/g), the recovery of ferulic acid in urine was quite
limited. This suggests that its absorption is much lower than
when FA is present in the diet as a free acid. In fact, the
presence of ester-linked monomeric and dimeric phenolics
leads to reduced biodegradability of the cell wall polysaccha-
ride, and limits the release of ferulic acid. Diferulates have a
dramatic effect on both the rate and extent of polysaccharide
degradation. Cross-linking through diferulates may inhibit the
binding of endoxylanases, thus limiting the extent of arabi-
noxylan degradation and consequently the release of esterified
ferulic acid (32). Furthermore, cross-linking may prevent lo-
calized areas from swelling, excluding key enzymes for dietary
fiber degradation (33). The form of the acid and its location in
the plant affect its fate after ingestion by mammals. Covalently
bound phenolic acids are released only after extensive micro-
bial attack and thus at sites of maximum microbial activity in
the colon (34). Degradative enzymes with molecular weights
⬎ 20 kDa would be totally excluded from most of the pores in
the wheat samples and even a 20- to 30-kDa enzyme would
have difficulty diffusing across the small numbers of pores with
radii of 2.5– 4 nm (35). As a result, these enzymes with a
Stokes radius of ⬎20 kDa are unable to diffuse into the
hydrated wall surface. All of these observations may explain
why FA present in whole wheat or bran is highly recovered in
feces and poorly recovered in urine.
The percentage of FA recovered in urine in relation to the
ingested dose was significantly higher in rats fed white Fv than
in rats fed Bv. The bran fraction presents a well-defined pore
structure compared with the endosperm fraction. The latter,
which is made up chiefly of starch, presents a much more
diffuse pattern and little evidence of structure, thus yielding a
more easily degradable substance (35). Moreover, approxi-
mately one third of unlignified endosperm arabinoxylans are
soluble in water in contrast to the acid arabinoxylan from
pericarp, which is mainly insoluble. This explains why non-
starch polysaccharides in primary cell walls from endosperm
are more extensively broken down than fiber in secondary
lignified cell walls from the pericarp (cf. Table 4). Therefore,
FA linked to arabinoxylan in the endosperm is more easily
released after enzymatic and microbial degradation.
The rate of release of esterified FA from a complex matrix
in the digestive tract appears dependent on the nature of the
fibers to which it is linked, but also on its concentration in the
matrix. Therefore, the greater the concentration of ferulic acid
in the matrix, the more probable the dimer formation, which
1968 ADAM ET AL.
implies a modification of the physical and chemical properties
of the fiber. However, when administered in meals as a free
acid, FA is easily absorbed and its intestinal and hepatic
metabolism do not limit its bioavailability.
In plasma, FA was present only as conjugated forms; under
these conditions, the biological effects of its metabolites re-
quire investigation. In humans, little is known about the
biological protective effects of FA metabolites. It has been
shown that FA in the glucuronidated form conserves its ability
to protect LDL oxidation in an in vitro system (17). These
protective effects could occur only during the postprandial
period because FA conjugates were rapidly eliminated in urine
after a meal. Human studies should be pursued to evaluate the
bioavailability of FA and its biological effects after ingestion of
a cereal matrix, which is a common food.
LITERATURE CITED
1. Jacobs, D. R., Jr., Marquart, L., Slavin, J. & Kushi, L. H. (1998) Whole-
grain intake and cancer: an expanded review and meta-analysis. Nutr. Cancer 30:
85–96.
2. Jacobs, D. R., Pereira, M. A., Meyer, K. A. & Kushi, L. H. (2000) Fiber
from whole grains, but not refined grains, is inversely associated with all-cause
mortality in older women: the Iowa women’s health study. J. Am. Coll. Nutr. 19:
326S–330S.
3. Meyer, K. A., Kushi, L. H., Jacobs, D. R., Jr., Slavin, J., Sellers, T. A. &
Folsom, A. R. (2000) Carbohydrates, dietary fiber, and incident type 2 diabetes
in older women. Am. J. Clin. Nutr. 71: 921–930.
4. Pietinen, P., Rimm, E. B., Korhonen, P., Hartman, A. M., Willett, W. C.,
Albanes, D. & Virtamo, J. (1996) Intake of dietary fiber and risk of coronary
heart disease in a cohort of Finnish men. The Alpha-Tocopherol, Beta-Carotene
Cancer Prevention Study. Circulation 94: 2720 –2727.
5. Fraser, G. E., Sabate, J., Beeson, W. L. & Strahan, T. M. (1992) A
possible protective effect of nut consumption on risk of coronary heart disease.
The Adventist Health Study. Arch. Intern. Med. 152: 1416 –1424.
6. Liu, S., Stampfer, M. J., Hu, F. B., Giovannucci, E., Rimm, E., Manson,
J. E., Hennekens, C. H. & Willett, W. C. (1999) Whole-grain consumption and
risk of coronary heart disease: results from the Nurses’ Health Study. Am. J. Clin.
Nutr. 70: 412– 419.
7. Anderson, J. W., Hanna, T. J., Peng, X. & Kryscio, R. J. (2000) Whole
grain foods and heart disease risk. J. Am. Coll. Nutr. 19: 291S–299S.
8. Wolk, A., Manson, J. E., Stampfer, M. J., Colditz, G. A., Hu, F. B., Speizer,
F. E., Hennekens, C. H. & Willett, W. C. (1999) Long-term intake of dietary fiber
and decreased risk of coronary heart disease among women. J. Am. Med. Assoc.
281: 1998 –2004.
9. Slavin, J. L., Jacobs, D., Marquart, L. & Wiemer, K. (2001) The role of
whole grains in disease prevention. J. Am. Diet. Assoc. 101: 780 –785.
10. Smith, M. M. & Hartley, R. D. (1983) Occurrence and nature of ferulic
acid substitution of cell wall polysaccharides in gramineous plants. Carbohydr.
Res. 118: 65– 80.
11. Fulcher, R. G. (1983) Fluorescence microscopy of cereals. In: New
Frontiers in Food Microstructure, pp. 167–175. American Association of Cereal
Chemists, Ottawa, Canada.
12. Sosulski, F., Krygier, K. & Hogge, L. (1982) Free esterified and extract-
able-bond phenolics acids. III. Composition of phenolic acids in cereal and potato
flours. J. Agric. Food Chem. 30: 337–340.
13. Lempereur, I., Rouau, X. & Abecassis, J. (1997) Genetic and agro-
nomic variation in arabinoxylan and ferulic acid contents of durum wheat (Triticum
durum L.) grain and its milling fractions. J. Cereal Sci. 25: 103–110.
14. Graf, E. (1992) Antioxidant potential of ferulic acid. Free Radic. Biol.
Med. 13: 435– 448.
15. Castelluccio, C., Bolwell, G. P., Gerrish, C. & Rice-Evans, C. (1996)
Differential distribution of ferulic acid to the major plasma constituents in relation
to its potential as an antioxidant. Biochem. J. 316: 691– 694.
16. Garcia-Conesa, M. T., Plumb, G. W., Kroon, P. A. & Wallace, G. (1997)
Antioxidant properties of ferulic acid dimers. Redox Rep. 3: 239 –244.
17. Ohta, T., Nakano, T., Egashira, Y. & Sanada, H. (1997) Antioxidant
activity of ferulic acid beta-glucuronide in the LDL oxidation system. Biosci.
Biotechnol. Biochem. 61: 1942–1943.
18. Choudhury, R., Srai, S. K., Debnam, E. & Rice-Evans, C. A. (1999)
Urinary excretion of hydroxycinnamates and flavonoids after oral and intravenous
administration. Free Radic. Biol. Med. 27: 278 –286.
19. Bourne, L. C. & Rice-Evans, C. (1998) Bioavailability of ferulic acid.
Biochem. Biophys. Res. Commun. 253: 222–227.
20. Chawla, A. S., Singh, M., Murthy, M. S., Gupta, M. P. & Singh, H. (1987)
Downloaded from https://academic.oup.com/jn/article-abstract/132/7/1962/4687425 by guest on 22 December 2019
Anti-inflammatory action of ferulic acid and its esters in carrageenan-induced rat
paw edema model. Indian J. Exp. Biol. 25: 187–189.
21. Imaida, K., Hirose, M., Yamaguchi, S., Takahashi, S. & Ito, N. (1990)
Effects of naturally occurring antioxidants on combined 1,2-dimethylhydrazine-
and 1-methyl-1-nitrosourea-initiated carcinogenesis in F344 male rats. Cancer
Lett. 55: 53–59.
22. Huang, M. T., Smart, R. C., Wong, C. Q. & Conney, A. H. (1988)
Inhibitory effect of curcumin, chlorogenic acid, caffeic acid, and ferulic acid on
tumor promotion in mouse skin by 12-O-tetradecanoylphorbol-13-acetate. Can-
cer Res. 48: 5941–5946.
23. Uchida, M., Nakajin, S., Toyoshima, S. & Shinoda, M. (1996) Antioxi-
dative effect of sesamol and related compounds on lipid peroxidation. Biol.
Pharm. Bull. 19: 623– 626.
24. Wolffram, S., Weber, T., Grenacher, B. & Scharrer, E. (1995) A Na(⫹)-
dependent mechanism is involved in mucosal uptake of cinnamic acid across the
jejunal brush border in rats. J. Nutr. 125: 1300 –1308.
25. Spencer, J. P., Chowrimootoo, G., Choudhury, R., Debnam, E. S., Srai,
S. K. & Rice-Evans, C. (1999) The small intestine can both absorb and
glucuronidate luminal flavonoids. FEBS Lett. 458: 224 –230.
26. Walle, U. K., Galijatovic, A. & Walle, T. (1999) Transport of the fla-
vonoid chrysin and its conjugated metabolites by the human intestinal cell line
Caco-2. Biochem. Pharmacol. 58: 431– 438.
27. Donovan, J. L., Crespy, V., Manach, C., Morand, C., Besson, C., Scalbert,
A. & Rémésy, C. (2001) Catechin is metabolized by both the small intestine
and liver of rats. J. Nutr. 131: 1753–1757.
28. Crespy, V., Morand, C., Manach, C., Besson, C., Demigné, C. & Rémésy,
C. (1999) Part of quercetin absorbed in the small intestine is conjugated and
further secreted in the intestinal lumen. Am. J. Physiol. 277: G120 –G126.
29. Andlauer, W., Kolb, J., Stehle, P. & Furst, P. (2000) Absorption and
metabolism of genistein in isolated rat small intestine. J. Nutr. 130: 843– 846.
30. Azuma, K., Ippoushi, K., Nakayama, M., Ito, H., Higashio, H. & Terao, J.
(2000) Absorption of chlorogenic acid and caffeic acid in rats after oral admin-
istration. J. Agric. Food Chem. 48: 5496 –5500.
31. Scheline, R. R. (1966) The decarboxylation of some phenolic acids by
the rat. Acta Pharmacol. Toxicol. (Copenh). 24: 275–285.
32. Grabber, J. H., Hatfield, R. D. & Ralph, J. (1998) Diferulate cross-links
impede the enzymatic degradation of non-lignified maize walls. J. Sci. Food Agric.
77: 193–200.
33. Bunzel, M., Ralph, J., Marita, J., Hatfield, R. & Steinhart, H. (2001)
Diferulates as structural components in soluble and insoluble cereal dietary fibre.
J. Sci. Food Agric. 81: 653– 660.
34. Chesson, A., Provan, G., Russell, W., Scobbie, L., Richardson, A. &
Stewart, C. (1999) Hydroxycinnamic acids in the digestive tract of livestock
and humans. J. Sci. Food Agric. 79: 373–378.
35. Chesson, A., Gardner, P. & Wood, T. (1997) Cell wall porosity and
available surface area of wheat straw and wheat grain fractions. J. Sci. Food
Agric. 75: 289 –295.