The Ethanol Metabolite Acetaldehyde Increases Paracellular Drug Permeability In Vitro and Oral Bioavailability In Vivo
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
0022-3565/10/3321-326–333$20.00
THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 332, No. 1
Copyright © 2010 by The American Society for Pharmacology and Experimental Therapeutics 158642/3543814
JPET 332:326–333, 2010 Printed in U.S.A.
The Ethanol Metabolite Acetaldehyde Increases Paracellular
Drug Permeability In Vitro and Oral Bioavailability In Vivo
Scott J. Fisher, Peter W. Swaan, and Natalie D. Eddington
Department of Pharmaceutical Sciences, University of Maryland, Baltimore, Maryland
Received July 8, 2009; accepted October 9, 2009
ABSTRACT
Alcohol consumption leads to the production of the highly fected. In vivo permeability was assessed in male Sprague-
Downloaded from jpet.aspetjournals.org at ASPET Journals on March 22, 2015
reactive ethanol metabolite, acetaldehyde, which may affect Dawley rats treated for 6 days with ethanol, disulfiram, or saline
intestinal tight junctions and increase paracellular permeability. alone or in combination. Bioavailability of naproxen was not af-
We examined the effects of elevated acetaldehyde within the fected by any treatment, whereas that of paclitaxel was increased
gastrointestinal tract on the permeability and bioavailability of upon acetaldehyde exposure. Although disulfiram has been
hydrophilic markers and drug molecules of variable molecular shown to inhibit multidrug resistance-1 P-glycoprotein (P-gp) in
weight and geometry. In vitro permeability was measured uni- vitro, our data demonstrate that the known P-gp substrate pacli-
directionally in Caco-2 and MDCKII cell models in the presence taxel is not affected by coadministration of disulfiram. In conclu-
of acetaldehyde, ethanol, or disulfiram, an aldehyde dehydro- sion, we demonstrate that acetaldehyde significantly modulates
genase inhibitor, which causes acetaldehyde formation when tight junctions and paracellular permeability in vitro as well as the
coadministered with ethanol in vivo. Acetaldehyde significantly oral bioavailability of low-molecular-weight hydrophilic probes and
lowered transepithelial resistance in cell monolayers and in- therapeutic molecules in vivo even when these molecules are
creased permeability of the low-molecular-weight markers, substrates for efflux transporters. These studies emphasize the
mannitol and sucrose; however, permeability of high-molecu- significance of ethanol metabolism and drug interactions outside
lar-weight markers, polyethylene glycol and inulin, was not af- of the liver.
The metabolism of ethanol leads to the production of range (Nuutinen et al., 1983), and it should be noted that this
acetaldehyde, a highly promiscuous intermediate, that is range is considered pathophysiologically relevant (Atkinson
able to react with various proteins and results in the and Rao, 2001). It has also been estimated that 30 M free,
formation of adducts at concentrations as low as 5 M unreacted, nonadduct acetaldehyde circulates in the blood of
(Salmela et al., 1997). Lysine is the major amino acid heavy drinkers (Brecher et al., 1997).
residue that is associated with adduct formation (Tuma et Emerging evidence suggests that pathophysiologically
al., 1987) via an intermediary Schiff’s base (Braun et al., relevant concentrations of acetaldehyde are also found in
1995). Proteins that preferentially form adducts with ac- the stomach and upper gastrointestinal (GI) tract. Various
etaldehyde include hemoglobin, albumin, tubulin, lipopro- alcohol dehydrogenase (ADH) isoenzymes (I–IV) occur at
teins, collagen, and CYP2E1 (Niemelä, 1999). Although different levels along the GI tract of the rat (Julià et al.,
the rate of protein adduct formation remains unclear, the
1987; Vaglenova et al., 2003). Although ADH IV predomi-
number of nucleophilic amino acid residues (e.g., Lys and
nates on the epithelial surface of the esophagus and stom-
Cys) within the protein apparently increases total adduct
ach, only ADH I is found in the duodenum (Vaglenova et
formation (Mauch et al., 1986).
al., 2003). Upon increasing levels of ethanol consumption,
Upon the administration of ethanol, high concentrations of
ADH I in the gut becomes saturated and it has been
acetaldehyde can be measured in blood. In rats, peak blood
acetaldehyde concentrations are in the millimolar range (Ki- observed that CYP2E1 is induced in both stomach and
noshita et al., 1996); in human blood, after ethanol consump- small intestine to levels 25% above levels in controls
tion, concentrations have been measured in the micromolar (Pronko et al., 2002). Although ADH metabolism of ethanol
in the GI tract is comparatively lower than that in the
liver, it may still contribute significantly to local concen-
Article, publication date, and citation information can be found at
http://jpet.aspetjournals.org. trations of acetaldehyde in the upper GI tract (Pronko et
doi:10.1124/jpet.109.158642. al., 2002).
ABBREVIATIONS: GI, gastrointestinal; ADH, alcohol dehydrogenase; PBS, phosphate-buffered saline; PEG, polyethylene glycol; TEER, trans-
epithelial resistance; ANOVA, analysis of variance; AUC, area under the curve; BCS, biopharmaceutics classification system.
326Acetaldehyde Increases Drug Permeability In Vivo 327
In concert with adherens junctions, tight junctions and passage 25 to 35 in an effort to derive monolayers with consistent
their associated actin filaments compose the apical junc- morphological and biological features. All cells were maintained in
tional complex between epithelial cells of the small intestine. 75-cm2 tissue culture flasks at 37°C, in a humidified atmosphere of
The loosening of this complex and the formation of the “leaky 5% CO2 and were cultured in high-glucose Dulbecco’s modified Ea-
gut” in chronic alcoholics has received substantial research gle’s medium, supplemented with 10% fetal bovine serum, 0.1 mM
nonessential amino acids, 2 mM L-glutamine, 1 mM sodium pyru-
interest (Atkinson and Rao, 2001; Keshavarzian et al., 2001;
vate, 100 U/ml penicillin, and 100 g/ml streptomycin. Cells were
Basuroy et al., 2005). Studies demonstrate that acetaldehyde passaged to 80% confluence and were released by trypsinization
increases tyrosine phosphorylation via modulation of ty- (0.25% trypsin and 0.25% EDTA). MDCKII and Caco-2 cells for use
rosine kinases or protein tyrosine phosphatases in three tight in this study were plated at a seeding density of 50,000 and 80,000
junctional proteins, ZO-1, E-cadherin, and -catenin, which cells/cm2, respectively; in 12-well tissue culture-treated Transwell
may be responsible for increased membrane permeability plates. Cell medium was changed every other day, and experiments
(Atkinson and Rao, 2001). In the lower GI tract, anaerobic were performed upon confluence in MDCKII cells (5 days) and upon
bacteria express ADH and have the ability to metabolize differentiation in Caco-2 cells (25 days). At this time each monolayer
ethanol and produce acetaldehyde (Rao, 1998). In both hu- is expected to have polarized, with well developed apical brush
mans and rats, colon acetaldehyde levels may achieve milli- borders and tight junctions. Monolayer integrity was assessed by the
transepithelial resistance (TEER) value as described in the next
molar concentrations (Koivisto and Salaspuro, 1997; Rao,
section.
1998; Atkinson and Rao, 2001). This is an important fact, as Measurement of TEER. TEER of confluent monolayers was
endotoxin is suspected to enter the bloodstream via “leaky” measured as an indicator of monolayer integrity as described previ-
Downloaded from jpet.aspetjournals.org at ASPET Journals on March 22, 2015
tight junctions and play a role in alcohol-induced liver dam- ously (Hidalgo et al., 1989), using a Millicell-ERS ohmmeter (Milli-
age (Keshavarzian et al., 1999). pore Corporation, Billerica, MA). TEER was calculated as ohms 䡠 per
The concentration of acetaldehyde in the upper GI tract square centimeter by multiplying it by the surface area of the insert
becomes significant because this area represents the site of (1.13 cm2) and subtracting the internal resistance of the insert before
absorption for most clinically therapeutic drugs. Higher calculations (25 ⍀ 䡠 cm2). TEER was measured at 0, 60, 120, 180, 240,
paracellular permeability due to disrupted tight junctions and 300 min. Baseline TEER values for Caco-2 and MDCKII mono-
may result in highly variable oral bioavailability; this result layers were between 450 and 600 and 200 and 250 ⍀ 䡠 cm2.
In Vitro Paracellular Permeability Measurements. The
is especially problematic for drugs with narrow therapeutic
transport of radiolabeled paracellular markers was measured in cells
windows. It has been reported that acetaldehyde formation seeded onto Transwell plates in the presence or absence of ethanol
in the intestine is proportional to the ethanol concentration (20% v/v), acetaldehyde (1 mM), or disulfiram (25 mg/ml). The mark-
delivered (Koivisto and Salaspuro, 1997). We hypothesize ers ranged in molecular weight from mannitol (182.2) to sucrose
that the levels of acetaldehyde in the upper small intestine (342.3), to PEG (4000) to inulin (5000). Caco-2 cells were used at day
are sufficient to disturb tight junctional organization, 25 and MDCKII cells at day 5 and upon confluence were washed
thereby affecting drug absorption. In this study, we exam- twice in PBS. To assess effects on permeability, transport of markers
ined changes in paracellular permeability in epithelial cell was measured from apical to basolateral chambers at 60, 120, 180, and
culture models and in a rat model of acetaldehyde formation. 240 min (n ⫽ 6). Cell treatments were agitated orbitally at no greater
By using hydrophilic markers and drug molecules of variable than 50 rpm and were performed in sealed airtight containers at 37°C
to prevent evaporation and in the study with acetaldehyde were re-
molecular weight and geometry, we demonstrate that etha-
placed each hour. Radioactivity was measured using scintillation count-
nol metabolism can significantly affect drug absorption and ing. Apparent permeability (Papp) was calculated as
disposition.
⌬Q 1
P app ⫽ 䡠 (1)
Materials and Methods ⌬t A 䡠 C共0兲
Chemicals and Reagents. Dulbecco’s modified Eagle’s me- where ⌬Q/⌬t is the linear appearance rate of mass in receiver solu-
dium, phosphate-buffered saline (PBS), nonessential amino acids, tion in the lateral chamber, A is the cross-sectional area of the
sodium pyruvate, penicillin, streptomycin, L-glutamine, and tryp- Transwell plate, and C(0) is the initial concentration of the marker in
sin/EDTA were purchased from Invitrogen (Grand Island, NY). the apical chamber.
Fetal bovine serum (0.1 M filtered) was purchased from Gemini Animals. Adult male Sprague-Dawley rats (250 –280 g) were used in
Bio-Products (Woodland, CA). Transwell plates of 12-mm diame- the pharmacokinetic measurement of permeability markers. They were
ter (surface area 1.13 cm2) and 0.4 M pore size were purchased purchased from Harlan (Indianapolis, IN). The study protocols were
from Corning Costar (Cambridge, MA). Acetaldehyde, tetraethyl- approved by the Institutional Animal Care and Use Committee of
thiuram disulfide (disulfiram), D-mannitol[1-3H(N)] (specific ac- the School of Pharmacy (University of Maryland, Baltimore, MD). Rats
tivity 18 Ci/mmol; radiochemical purity ⬎97%), [3H]inulin (212 were housed in the animal facility at a room temperature of 23 ⫾ 1°C.
mCi/g; purity ⱖ95%), and ethanol were purchased from Sigma- Food was withheld 4 h before oral dosing, and rats were allowed free
Aldrich (St. Louis, MO). [U-14C]Sucrose (498 mCi/mmol; purity access to food after dosing (Purina 5001 Rodent Chow; Purina, St.
⬎97%) was purchased from PerkinElmer Life and Analytical Sci- Louis, MO). Water was given ad libitum, and rats were maintained on
ences (Waltham, MA). [14C]PEG-4000 (0.87 mCi/g; purity ⬎95%) a 12-h light/dark cycle (light on from 7:00 AM to 7:00 PM).
and [14C]naproxen sodium (50 mCi/mmol; purity ⬎98%) were Ethanol Dosing and In Vivo Permeability Measurements.
purchased from American Radiolabeled Chemicals (St. Louis, Rats in each group (n ⫽ 5) were treated with saline, disulfiram (100
MO). [14C]Paclitaxel (68 mCi/mmol; purity ⬎97%) was purchased mg/kg), ethanol (5 g/kg b.wt. given as a 33% v/v solution in saline,
from Moravek Radiochemicals (Brea, CA). UniverSol scintillation approximately 7 M), or ethanol and disulfiram [5 g/kg b.wt. (7 M) and
cocktail was purchased from MP Biomedical (Irvine, CA). 100 mg/kg (84 M), respectively]. Treatments were administered
Cell Culture. Human colon adenocarcinoma (Caco-2) cells and twice daily via oral gavage for a period of 6 days. On day 5, all
Madin-Darby canine kidney (MDCKII) cells were purchased from animals were anesthetized and underwent jugular vein cannulation.
the American Type Culture Collection (Manassas, VA). Caco-2 cells On day 6, one of four hydrophilic markers ([3H]mannitol, [14C]su-
were used from passage 25 to 45, and MDCKII cells were used from crose, [3H]inulin, or [14C]PEG-4000) or [14C]naproxen or [14C]pacli-328 Fisher et al.
taxel were given via oral gavage to each group. Serial blood samples ing molecular weight were used in the Caco-2 cell monolayer.
were taken at 0, 20, 40, 60, 90, 120, 240, 480, 720, and 1440 min after [3H]Mannitol (182.2), [14C]sucrose (342.3), [14C]PEG-4000
dosing. Blood was centrifuged (10,000g for 10 min) immediately, and (4000), and [3H]inulin (5000) were evaluated in the presence
plasma was obtained. Scintillation cocktail was added, and plasma and absence of disulfiram, ethanol, and acetaldehyde. Disul-
samples were analyzed for radioactivity by an LS 6500 multipurpose
firam treatment did not result in any significant effect on the
scintillation counter (Beckman Coulter, Fullerton, CA). Noncompart-
mental pharmacokinetic analysis was performed to determine the
permeability of any marker (Fig. 2A). Ethanol treatment
radiomarker absorption in each treatment group using WinNonlin significantly increased the permeability of both [3H]mannitol
version 3.1 (Pharsight, Mountain View, CA). and [3H]inulin (Fig. 2B). Compared with ethanol and disul-
Statistical Analysis. All samples in all experiments were ana- firam, acetaldehyde treatment resulted in the most pro-
lyzed for statistically significant differences between groups using nounced increase in Papp for three of the four permeability
single-factor analysis of variance (ANOVA) with the ␣ value set a markers as well as higher overall transport of [3H]mannitol
priori at p ⬍ 0.05. In vitro and in vivo values are expressed as the in both cell lines (Fig. 2, C and F). Mannitol flux values were
mean ⫾ S.E.M. obtained from six experiments (in vitro) or five
generally higher and more variable (0.1–2.0 ⫻ 10⫺6 cm2/s) in
samples (in vivo).
the present study than in other reports (Volpe, 2008) and
cannot be explained by low radiopurity of the tritiated man-
Results nitol probe. A variety of issues (Volpe, 2008) may have con-
Reversibility of Transepithelial Electrical Resis- tributed to this anomaly, but these factors do not necessarily
tance upon Ethanol and Acetaldehyde Treatment. negate the significance of our data when control and treated
Downloaded from jpet.aspetjournals.org at ASPET Journals on March 22, 2015
TEER measurement is a method used frequently to assess cell monolayers are compared (Fig. 2). It is interesting to note
the integrity of Caco-2 cell monolayers (Hidalgo et al., that the apparent permeability for the high-molecular-
1989; Swaan et al., 1997; Cox et al., 2002; Salama et al., weight marker inulin was significantly increased upon alco-
2003, 2004). TEER was measured each hour over the hol treatment in Caco-2 cells (Fig. 2B) and acetaldehyde
course of the experiment (5 h). Treatment of Caco-2 cell treatment in both cell lines (Fig. 2, C and F), whereas the
monolayers with both ethanol (20% v/v) and acetaldehyde Papp of PEG-4000 was unaffected by any treatment modality.
(1 mM) significantly lowered TEER values from 1 to 4 h It should be noted that permeability values of the sucrose are
(Fig. 1A); however, the effect of acetaldehyde was more somewhat higher in Caco-2 cells than in MDCKII cells (Fig.
pronounced. The decrease in TEER values for ethanol and 2, A–C, versus D–F, respectively). This result could be attrib-
acetaldehyde treatment in MDCKII cell lines paralleled uted to the presence, albeit at low expression levels, of su-
those observed in Caco-2 cells (Fig. 1B). The maximal crase in Caco-2 but not MDCKII cell lines (Beaulieu and
decrease in TEER was observed upon acetaldehyde treat- Quaroni, 1991). Transport of [U-14C]sucrose hydrolysis prod-
ment at 240 min and was 73.5 ⫾ 2.5% in Caco-2 cells and ucts [14C]glucose and [14C]fructose would be mediated by
49.6 ⫾ 3.0% in MDCKII cells (Fig. 1), returning to starting solute carrier proteins from the Na⫹-dependent glucose
TEER values 60 min after withdrawal of treatment. This transporter and glucose transporter families (Joost and Tho-
reversibility suggests that the opening of tight junctions is rens, 2001; Wood and Trayhurn, 2003; Cheeseman, 2008),
not due to cellular cytotoxicity. These results are consistent potentially resulting in higher apparent permeability mea-
with our previous data showing that the doses of ethanol and surements for sucrose.
acetaldehyde used here are not cytotoxic (Fisher et al., 2008). Acetaldehyde Effects on Paracellular Permeabil-
Ethanol and Acetaldehyde Increase Permeability of ity Markers In Vivo. In vitro studies in both Caco-2 and
Paracellular Markers in Caco-2 and MDCKII Cell Mod- MDCKII cells suggest that low-molecular-weight markers
els. To assess the degree of tight junctional opening, various have significantly increased permeability upon treatment
hydrophilic markers of paracellular permeability of increas- with acetaldehyde. Based on these results, the bioavailability
Fig. 1. Effect of control (PBS, 〫), disulfiram (25 mg/ml, F), ethanol (20%, f), and acetaldehyde (1 mM, Œ) on TEER measured in Caco-2 (A) and
MDCKII (B) cell monolayers. TEER was measured from 0 to 300 min. Treatments were removed at 240 min, and cells were allowed to recover in media
for 60 min. Each data point represents the mean ⫾ S.E.M. of at least eight experiments. Samples were analyzed for any statistically significant
differences between groups using single-factor ANOVA. ⴱ, statistically significant difference between treatment and control, p ⬍ 0.05; ⴱⴱ, p ⬍ 0.001.Acetaldehyde Increases Drug Permeability In Vivo 329
Downloaded from jpet.aspetjournals.org at ASPET Journals on March 22, 2015
Fig. 2. Effect of disulfiram (25 mg/ml) (A and D), ethanol (20%) (B and E) and acetaldehyde (1 mM) (C and F) treatment (⫹) on permeability of
[3H]mannitol, [14C]sucrose, [3H]inulin, and [14C]PEG-4000 versus untreated controls (⫺) in Caco-2 (A–C) and MDCKII (D–F) cells. Permeability was
assessed each hour for a period of 4 h. Each data point represents the mean ⫾ S.E.M. of at least six experiments. Samples were analyzed for any
statistically significant differences between groups using single-factor ANOVA. ⴱ, statistically significant difference between treatment and control,
p ⬍ 0.05; ⴱⴱ, p ⬍ 0.001.
upon oral administration of various molecular weight mark- dehyde in the rat in vivo and found that the dosing scheme
ers were examined in the rat. The four markers were used, as described under Materials and Methods yields approximately
described above, in rats treated with ethanol, disulfiram, or a 80 M acetaldehyde in plasma (Fisher et al., 2008). The
combination of the two to generate in vivo acetaldehyde. Our percentage of the overall dose of [3H]mannitol detected in
laboratory has previously examined the generation of acetal- plasma was significantly higher in rats treated with ethanol/330 Fisher et al.
disulfiram than in rats treated with ethanol alone or disul- availability of a highly bioavailable drug, naproxen, and a
firam alone or control rats. We found similar results with drug with limited bioavailability, paclitaxel. Upon oral dos-
rats who received [14C]sucrose and detected a significant ing with [14C]naproxen, there was no significant difference in
increase in systemic sucrose availability only in rats gener- detected plasma levels in any treatment group (Fig. 5A).
ating in vivo acetaldehyde (Fig. 3B). The systemic availabil- However, there was a significant effect (p ⬍ 0.05) of acetal-
ity of the high-molecular-weight markers, [14C]PEG-4000 dehyde production on the bioavailability of paclitaxel (Fig.
and [3H]inulin, was unchanged compared with that in control 5B), with an approximate 4-fold increase in AUC (7051 ver-
rats for any treatment group (Fig. 3, C and D). sus 1875 mg/ml 䡠 min versus control rats). Although the
Pharmacokinetic analysis for the [3H]mannitol group dem- overall bioavailability of paclitaxel remained very low (less
onstrated that the area under the curve (AUC) profile of the than 6% of the total dose administered) in every treatment
acetaldehyde-generating group was approximately 15-fold group, this result is significant. All of other treatment groups
greater than that of control rats (45,434 ⫾ 5823 Ci 䡠 min/ml showed no significant influence on paclitaxel bioavailability.
versus 3134 ⫾ 642 Ci 䡠 min/ml, respectively). Analysis of the
[14C]sucrose group showed an approximate 7-fold increase in
the AUC of the marker with values of 290.2 ⫾ 74.7 Ci 䡠
Discussion
min/ml for the treatment group versus 41.4 ⫾ 7.3 Ci 䡠 The effect of acetaldehyde, which is generated during the
min/ml for the control group. These data suggest that in vivo course of alcohol consumption, on paracellular permeability
bioavailability is increased in the acetaldehyde-generating and loosening of tight junctions has received increasing at-
Downloaded from jpet.aspetjournals.org at ASPET Journals on March 22, 2015
rat model in a size-dependent fashion, probably as a result of tention in recent years (Rao, 1998; Atkinson and Rao, 2001;
increased intestinal permeability via pharmacologically in- Seth et al., 2004; Sheth et al., 2004; Basuroy et al., 2005).
duced opening of the tight junctions. These studies have demonstrated a pronounced effect of
Acetaldehyde Affects on Paclitaxel but Not Naproxen acetaldehyde on the structure of the tight junction, but the
Bioavailability. To further define the potential in vivo rel- extent of tight junctional modulation and its consequent im-
evance of ethanol metabolism in the GI tract of the rat, we pact on drug permeability have not been investigated. There-
studied the effect of acetaldehyde production on the systemic fore, the aim of the present study was to examine the effects
Fig. 3. In vivo bioavailability of low-molecular-weight markers [3H]mannitol (182.2) (A) and [14C]sucrose (342.3) (B) and high-molecular-weight
markers [14C]PEG-4000 (4000) (C) and [3H]inulin (5000) (D). Rats were treated for 6 days via an oral bolus dose with saline control (〫), 5 g/kg ethanol
(䡺), 100 mg/kg disulfiram (E), or 5 g/kg ethanol and 100 mg/kg disulfiram (Œ). Uptake of markers was measured using liquid scintillation as a
percentage of the total starting dose. Each data point represents the mean ⫾ S.E.M. of at least six experiments from time 0 to 1440 min. Samples were
analyzed for any statistically significant differences between groups using single-factor ANOVA. ⴱ, statistically significant difference between
treatment group Vmax and control Vmax (p ⬍ 0.05).Acetaldehyde Increases Drug Permeability In Vivo 331
of acetaldehyde produced from ethanol metabolism in vivo on assessment of PEG probes with radii of 3.5 to 7.4 Å. Parallel
the degree of tight junctional disruption by assessing the to TEER data (Fig. 1), disulfiram had no effect on paracellu-
permeability of various marker molecules of increasing mo- lar permeability of any marker molecules across Caco-2 or
lecular weight and geometry. In addition, we established the MDCKII cell monolayers (Fig. 2, A and D), whereas ethanol
effects of ethanol, its major metabolite, and drugs used to increased mannitol (rh ⫽ 4.1 Å) but not sucrose (rh ⫽ 5.2 Å)
generate in vivo acetaldehyde on the permeability of marker permeability (Fig. 2, B and E). Acetaldehyde treatment (Fig.
molecules in cell culture models in vitro. To validate the 2, C and F) effected the greatest permeability increase for
clinical significance of these findings, we determined the both mannitol and sucrose. MDCKII cells showed a signifi-
effects of ethanol metabolism in the GI tract on the bioavail- cant increase in marker permeability only with acetaldehyde
ability of two drug molecules representing different biophar- treatment, which was somewhat unexpected based on the
maceutics classification system (BCS) classes. Naproxen, an significant decrease in TEER values observed above (Fig. 1).
over-the-counter nonsteroidal anti-inflammatory drug is a This result suggests that ethanol and acetaldehyde have
BCS class I/II compound, displaying high permeability and different mechanisms of paracellular permeability enhance-
high/low solubility dependent on pH (Wu and Benet, 2005); ment in MDCKII cells. It is interesting to note that Caco-2
paclitaxel, on the other hand, is a representative of com- cell monolayers displayed a significant permeability increase
pounds in the relatively rare BCS class IV, which feature of the high-molecular-weight marker, inulin (5000), but not
poor permeability and solubility characteristics. PEG-4000 (4000). This result may be explained by the com-
Cell lines such as Caco-2 and MDCKII form tight junctions pact spherical shape of inulin (rh ⫽ 10 Å versus 15.9 Å for
Downloaded from jpet.aspetjournals.org at ASPET Journals on March 22, 2015
upon differentiation, with apical and basolateral membrane, PEG-4000), which makes it more amenable to paracellular
and are generally considered to be good models for the in transport. This effect was observed previously by Ghande-
vitro study of epithelial permeability. In this study we exam- hari et al. (1997), who demonstrated that the permeability of
ined the effects of ethanol (20% v/v), disulfiram (25 mg/ml), probes via tight junctions is based on molecular weight, along
and acetaldehyde (1 mM) on the reversibility of TEER in both with properties such as geometry, flexibility, hydrophobicity,
Caco-2 (Fig. 1A) and MDCKII cells (Fig. 1B). TEER is an and charge.
established measure of the integrity of cellular monolayers To validate our in vitro results and assess a correlation
and represents a qualitative assessment of the paracellular with in vivo models, we monitored the oral pharmacokinetics
pathway aperture. As anticipated, disulfiram administration and bioavailability of each molecular marker in a rat model
did not affect TEER values because this compound is not exposed to various treatments associated with alcohol metab-
known to affect tight junctions or function as a membrane olism. To simulate the production of acetaldehyde in vivo,
permeabilizer (Balakirev and Zimmer, 2001). Our data show rats were treated with a combination of ethanol and the
a significant decrease in cell monolayer TEER values upon aldehyde dehydrogenase inhibitor disulfiram. Under these
treatment with a “sublethal” dose of ethanol (20% v/v). These conditions we observed a marked increase in the systemic
data are consistent with studies by Ma et al. (1995, 1999), availability upon oral administration of low- but not high-
who showed that ethanol permeabilizes the membrane at molecular-weight probes (Fig. 3). It should be noted that
concentrations less than 10% (v/v); furthermore, they specu- control experiments with ethanol and disulfiram alone did
lated that ethanol reversibly affects tight junctional function not have an effect on probe availability (Fig. 4). Mannitol has
via disassembly of perijunctional actin and myosin filaments. been used as a standard probe in the assessment of junc-
However, the definitive mechanism(s) behind these observed tional integrity (Swaan et al., 1994). Thus, the animal data
effects remain to be studied. Under the experimental condi- presented here correlate well with our in vitro results, indi-
tions applied here, acetaldehyde reversibly and significantly cating an acetaldehyde-induced permeability enhancement
decreases TEER values, a result that is comparable to those for low-molecular-weight molecules, most likely due to its
of studies by our laboratory (Fisher et al., 2008) as well as by disruptive effect on intestinal tight junctions. It has been
other groups (Atkinson and Rao, 2001; Basuroy et al., 2005).
Our previous studies showed that acetaldehyde in concentra-
tions up to 2 mM are not significantly cytotoxic (Fisher et al.,
2008). The effect of acetaldehyde on monolayer resistance
was more pronounced in MDCKII cell monolayers than in
Caco-2 cells (75% versus 50% of their starting values, respec-
tively). Because MDCKII cell and Caco-2 cells originate from
the kidney and the intestine, respectively, these data may
highlight the differential effect of ethanol and its metabolites
on paracellular integrity in different organs.
To further assess the effect of acetaldehyde on membrane
permeability, we used several molecular markers of various
sizes that cross membranes exclusively via a paracellular
route. The permeability of these probes was tested on cell
Fig. 4. In vivo AUC determinations in the in vivo rat model as analyzed
monolayers with fully differentiated cells and junctional com- by WinNonlin. Male Sprague-Dawley rats (n ⫽ 6) were exposed to saline
plexes. Studies by Knipp et al. (1997), using a series of control (C), 33% ethanol (E), 100 mg/kg disulfiram (D), or ethanol/disul-
structurally unrelated compounds with variable hydrody- firam (A) and given one of four molecular probes, [3H]mannitol, [14C]su-
namic radii (rh) estimated the pore radius of Caco-2 cells to be crose, [14C]PEG-4000, and [3H]inulin, respectively. Each data point rep-
resents the mean ⫾ S.E.M. of at least six experiments from time 0 to 1440
approximately 5.2 Å, whereas Watson et al. (2001) accurately min. Samples were analyzed for any statistically significant differences
established this pore radius at 4.5 Å using permeability between groups using single-factor ANOVA.332 Fisher et al.
Fig. 5. In vivo bioavailability of low-molecular-weight drug [14C]naproxen (252.2) (A) and medium-molecular-weight drug [14C]paclitaxel (853.9) (B).
Rats were treated for 6 days via an oral bolus dose with saline control (䡺), 5 g/kg ethanol (F), 100 mg/kg disulfiram (⽧), or 5 g/kg ethanol and 100
mg/kg disulfiram (f). Uptake of drug was measured as a percentage of the total starting dose using liquid scintillation. Each data point represents
the mean ⫾ S.E.M. of at least six experiments from time 0 to 1440 min. Samples were analyzed for any statistically significant differences between
groups using single-factor ANOVA.
Downloaded from jpet.aspetjournals.org at ASPET Journals on March 22, 2015
suggested that acetaldehyde disrupts tight junctions via a remained low (⬍6%). It should be noted that disulfiram has
tyrosine kinase-dependent mechanism (Atkinson and Rao, been shown to be an inhibitor of P-glycoprotein in vitro (Loo
2001). Evidence exists for a dramatic reduction of the zonula et al., 2004), probably by covalent modification of specific
occludens protein ZO-1 within tight junctions upon applica- cysteine residues in both substrate and ATP-binding do-
tion of pathophysiologically relevant levels of acetaldehyde mains (Loo and Clarke, 2000). However, metabolism of di-
(⬃650 M) (Rao, 1998; Atkinson and Rao, 2001). Extending sulfiram, which is rapidly converted upon ingestion in the
these in vitro observations to our studies, we conclude that stomach (Johansson, 1992), would make this result difficult
acetaldehyde generated in vivo directly affects the intestinal to observe in vivo, especially because the studies were per-
tight junctional complex, thereby causing a temporary size- formed at disulfiram concentrations much higher than those
and geometry-dependent increase in permeability. Overall, achieved clinically. It is noteworthy that coadministration of
these results indicate that acetaldehyde, not ethanol, is re- disulfiram or ethanol with paclitaxel did not lead to in-
sponsible for the observed modulation of tight junctions both creased paclitaxel plasma levels (Fig. 5B), thereby excluding
in vitro and in vivo. the possibility of an indirect effect of disulfiram or ethanol on
To further assess the clinical relevance of the above obser- P-glycoprotein activity that would influence intestinal pacli-
vations, we determined the oral bioavailability of two drug taxel permeability. Therefore, the increase in paclitaxel bio-
molecules upon administration of saline (control), ethanol, availability observed here upon administration of disulfiram
disulfiram, or acetaldehyde (i.e., coadministered disulfiram/ and ethanol cannot be attributed to inhibition of P-glycopro-
ethanol). We selected the nonsteroidal anti-inflammatory tein; however, an effect of in vivo-generated acetaldehyde on
drug naproxen as a representative control drug with a low P-glycoprotein activity cannot be excluded. These results
molecular weight (252), high solubility at neutral to basic pH, highlight the fact that perturbation of tight junctions can
and reportedly high permeability and almost complete enhance the permeability of drug molecules and even over-
(⬎95%) oral bioavailability. These features place naproxen in come an opposing force created by efflux transporters.
BCS class I (because pH influences naproxen solubility, low These studies further demonstrate the pronounced effects
pH conditions would constitute a class II categorization). As of acetaldehyde on intestinal permeability and its impact on
expected, the addition of ethanol or production of acetalde- drug absorption in chronic alcoholics. Furthermore, data pre-
hyde in the animal during drug administration had no sig- sented here add to mounting evidence that acetaldehyde is
nificant effect on naproxen bioavailability (Fig. 5A). The che- responsible for the loosening of tight junctions rather than
motherapeutic drug paclitaxel, on the other hand, is a ethanol. Although this series of experiments displays
relatively high-molecular-weight drug (854) that is insoluble changes in functional assays, it is clear that continued exam-
in aqueous solution. It displays poor permeability character- ination of the mechanism by which acetaldehyde affects tight
istics partly because of its affinity for the efflux transporter junctions, adherens junctions, proteins within these junc-
P-glycoprotein, abundantly expressed along the GI tract. As tional complexes, or some combination thereof, is necessary.
a result, paclitaxel has low and variable bioavailability and is Overall, our data illustrate that modulation of the paracel-
not traditionally given via oral routes of delivery. Owing to lular permeation pathway by acetaldehyde affects drug ab-
its poor permeability and solubility, it is a member of BCS sorption and bioavailability, depending on drug size, geome-
class IV, which comprises drug molecules that are typically try, and BCS classification.
difficult to formulate. Thus, we used paclitaxel as a repre-
sentative class IV drug (Varma and Panchagnula, 2005) to References
observe the effects of alcohol and it metabolites on the open- Atkinson KJ and Rao RK (2001) Role of protein tyrosine phosphorylation in acetal-
dehyde-induced disruption of epithelial tight junctions. Am J Physiol Gastrointest
ing of tight junctions. After acetaldehyde production, plasma Liver Physiol 280:G1280 –G1288.
paclitaxel levels were significantly higher on oral adminis- Balakirev MY and Zimmer G (2001) Mitochondrial injury by disulfiram: two differ-
ent mechanisms of the mitochondrial permeability transition. Chem Biol Interact
tration compared with values for controls (saline, ethanol, 138:299 –311.
and disulfiram) (Fig. 5B), even though overall bioavailability Basuroy S, Sheth P, Mansbach CM, and Rao RK (2005) Acetaldehyde disrupts tightAcetaldehyde Increases Drug Permeability In Vivo 333
junctions and adherens junctions in human colonic mucosa: protection by EGF and Mauch TJ, Donohue TM Jr, Zetterman RK, Sorrell MF, and Tuma DJ (1986)
L-glutamine. Am J Physiol Gastrointest Liver Physiol 289:G367–G375. Covalent binding of acetaldehyde selectively inhibits the catalytic activity of ly-
Beaulieu JF and Quaroni A (1991) Clonal analysis of sucrase-isomaltase expression sine-dependent enzymes. Hepatology 6:263–269.
in the human colon adenocarcinoma Caco-2 cells. Biochem J 280 (Pt 3):599 – 608. Niemelä O (1999) Aldehyde-protein adducts in the liver as a result of ethanol-
Braun KP, Cody RB Jr, Jones DR, and Peterson CM (1995) A structural assignment induced oxidative stress. Front Biosci 4:D506 –513.
for a stable acetaldehyde-lysine adduct. J Biol Chem 270:11263–11266. Nuutinen H, Lindros KO, and Salaspuro M (1983) Determinants of blood acetaldehyde
Brecher AS, Hellman K, and Basista MH (1997) A perspective on acetaldehyde level during ethanol oxidation in chronic alcoholics. Alcohol Clin Exp Res 7:163–168.
concentrations and toxicity in man and animals. Alcohol 14:493– 496. Pronko P, Bardina L, Satanovskaya V, Kuzmich A, and Zimatkin S (2002) Effect of
Cheeseman C (2008) GLUT7: a new intestinal facilitated hexose transporter. Am J chronic alcohol consumption on the ethanol- and acetaldehyde-metabolizing sys-
Physiol Endocrinol Metab 295:E238 –E241. tems in the rat gastrointestinal tract. Alcohol Alcohol 37:229 –235.
Cox DS, Raje S, Gao H, Salama NN, and Eddington ND (2002) Enhanced perme- Rao RK (1998) Acetaldehyde-induced increase in paracellular permeability in Caco-2
ability of molecular weight markers and poorly bioavailable compounds across cell monolayer. Alcohol Clin Exp Res 22:1724 –1730.
Caco-2 cell monolayers using the absorption enhancer, zonula occludens toxin. Salama NN, Fasano A, Lu R, and Eddington ND (2003) Effect of the biologically
Pharm Res 19:1680 –1688. active fragment of zonula occludens toxin, ⌬G, on the intestinal paracellular
Fisher SJ, Lee IJ, Swaan PW, and Eddington ND (2008) Evaluation of the effect of transport and oral absorption of mannitol. Int J Pharm 251:113–121.
ethanol’s toxic metabolite acetaldehyde on the gastrointestinal oligopeptide trans- Salama NN, Fasano A, Thakar M, and Eddington ND (2004) The effect of ⌬G on the
porter, PEPT1: in vitro and in vivo studies. Alcohol Clin Exp Res 32:162–170. transport and oral absorption of macromolecules. J Pharm Sci 93:1310 –1319.
Ghandehari H, Smith PL, Ellens H, Yeh PY, and Kopecek J (1997) Size-dependent Salmela KS, Sillanaukee P, Itälä L, Väkeväinen S, Salaspuro M, and Roine RP
permeability of hydrophilic probes across rabbit colonic epithelium. J Pharmacol (1997) Binding of acetaldehyde to rat gastric mucosa during ethanol oxidation.
Exp Ther 280:747–753. J Lab Clin Med 129:627– 633.
Hidalgo IJ, Raub TJ, and Borchardt RT (1989) Characterization of the human colon Seth A, Basuroy S, Sheth P, and Rao RK (2004) L-Glutamine ameliorates acetalde-
carcinoma cell line (Caco-2) as a model system for intestinal epithelial permeabil- hyde-induced increase in paracellular permeability in Caco-2 cell monolayer. Am J
ity. Gastroenterology 96:736 –749. Physiol Gastrointest Liver Physiol 287:G510 –G517.
Johansson B (1992) A review of the pharmacokinetics and pharmacodynamics of Sheth P, Seth A, Thangavel M, Basuroy S, and Rao RK (2004) Epidermal growth
disulfiram and its metabolites. Acta Psychiatr Scand Suppl 369:15–26. factor prevents acetaldehyde-induced paracellular permeability in Caco-2 cell
Joost HG and Thorens B (2001) The extended GLUT-family of sugar/polyol transport monolayer. Alcohol Clin Exp Res 28:797– 804.
Downloaded from jpet.aspetjournals.org at ASPET Journals on March 22, 2015
facilitators: nomenclature, sequence characteristics, and potential function of its Swaan PW, Hillgren KM, Szoka FC Jr, and Oie S (1997) Enhanced transepithelial
novel members (review). Mol Membr Biol 18:247–256. transport of peptides by conjugation to cholic acid. Bioconjug Chem 8:520 –525.
Julià P, Farrés J, and Parés X (1987) Characterization of three isoenzymes of rat Swaan PW, Marks GJ, Ryan FM, and Smith PL (1994) Determination of transport
alcohol dehydrogenase. Tissue distribution and physical and enzymatic properties.
rates for arginine and acetaminophen in rabbit intestinal tissues in vitro. Pharm
Eur J Biochem 162:179 –189.
Res 11:283–287.
Keshavarzian A, Choudhary S, Holmes EW, Yong S, Banan A, Jakate S, and Fields
Tuma DJ, Newman MR, Donohue TM Jr, and Sorrell MF (1987) Covalent binding of
JZ (2001) Preventing gut leakiness by oats supplementation ameliorates alcohol-
acetaldehyde to proteins: participation of lysine residues. Alcohol Clin Exp Res
induced liver damage in rats. J Pharmacol Exp Ther 299:442– 448.
11:579 –584.
Keshavarzian A, Holmes EW, Patel M, Iber F, Fields JZ, and Pethkar S (1999) Leaky
Vaglenova J, Martínez SE, Porté S, Duester G, Farrés J, and Parés X (2003)
gut in alcoholic cirrhosis: a possible mechanism for alcohol-induced liver damage.
Expression, localization and potential physiological significance of alcohol dehy-
Am J Gastroenterol 94:200 –207.
drogenase in the gastrointestinal tract. Eur J Biochem 270:2652–2662.
Kinoshita H, Ijiri I, Ameno S, Fuke C, Fujisawa Y, and Ameno K (1996) Inhibitory
Varma MV and Panchagnula R (2005) Enhanced oral paclitaxel absorption with
mechanism of intestinal ethanol absorption induced by high acetaldehyde concen-
trations: effect of intestinal blood flow and substance specificity. Alcohol Clin Exp vitamin E-TPGS: effect on solubility and permeability in vitro, in situ and in vivo.
Res 20:510 –513. Eur J Pharm Sci 25:445– 453.
Knipp GT, Ho NF, Barsuhn CL, and Borchardt RT (1997) Paracellular diffusion in Volpe DA (2008) Variability in Caco-2 and MDCK cell-based intestinal permeability
Caco-2 cell monolayers: effect of perturbation on the transport of hydrophilic assays. J Pharm Sci 97:712–725.
compounds that vary in charge and size. J Pharm Sci 86:1105–1110. Watson CJ, Rowland M, and Warhurst G (2001) Functional modeling of tight
Koivisto T and Salaspuro M (1997) Effects of acetaldehyde on brush border enzyme junctions in intestinal cell monolayers using polyethylene glycol oligomers. Am J
activities in human colon adenocarcinoma cell line Caco-2. Alcohol Clin Exp Res Physiol Cell Physiol 281:C388 –C397.
21:1599 –1605. Wood IS and Trayhurn P (2003) Glucose transporters (GLUT and SGLT): expanded
Loo TW, Bartlett MC, and Clarke DM (2004) Disulfiram metabolites permanently families of sugar transport proteins. Br J Nutr 89:3–9.
inactivate the human multidrug resistance P-glycoprotein. Mol Pharm 1:426 – 433. Wu CY and Benet LZ (2005) Predicting drug disposition via application of BCS:
Loo TW and Clarke DM (2000) Blockage of drug resistance in vitro by disulfiram, a transport/absorption/elimination interplay and development of a biopharmaceu-
drug used to treat alcoholism. J Natl Cancer Inst 92:898 –902. tics drug disposition classification system. Pharm Res 22:11–23.
Ma TY, Hollander D, Tran LT, Nguyen D, Hoa N, and Bhalla D (1995) Cytoskeletal
regulation of Caco-2 intestinal monolayer paracellular permeability. J Cell Physiol Address correspondence to: Dr. Peter W. Swaan, Department of Pharma-
164:533–545. ceutical Sciences, University of Maryland, 20 Penn St., Baltimore, MD 21201.
Ma TY, Nguyen D, Bui V, Nguyen H, and Hoa N (1999) Ethanol modulation of E-mail: pswaan@rx.umaryland.edu
intestinal epithelial tight junction barrier. Am J Physiol 276:G965–G974.You can also read
