transepithelial transport of poly(amidoamine) dendrimers across caco-2 cell monolayers

Journal of Controlled Release 81 (2002) 355–365www.elsevier.com/ locate / jconrel

Transepithelial transport of poly(amidoamine) dendrimers acrossCaco-2 cell monolayers

a b bMohamed El-Sayed , Mark Ginski , Christopher Rhodes ,a ,*Hamidreza Ghandehari

aUniversity of Maryland School of Pharmacy, Department of Pharmaceutical Sciences, 20 N. Pine Street, Baltimore, MD 21201, USAbGuilford Pharmaceuticals Inc., Department of Pharmaceutics, 6411 Beckley Street, Baltimore, MD 21224, USA

Received 22 December 2001; accepted 22 March 2002

Abstract

The objective of this study was to investigate the influence of physiochemical parameters (such as size, molecular weight,molecular geometry, and number of surface amine groups) of poly (amidoamine) (PAMAM) dendrimers, on theirpermeability across Caco-2 cell monolayers. The permeability of a series of PAMAM dendrimers, generations 0–4 (G0–G4),was investigated across Caco-2 cell monolayers in both the apical to basolateral (AB) and basolateral to apical (BA)directions. The influence of PAMAM dendrimers on the integrity, paracellular permeability, and viability of Caco-2 cellmonolayers was also monitored by measuring the transepithelial electrical resistance (TEER), mannitol permeability, andleakage of lactate dehydrogenase (LDH) enzyme, respectively. G0, G1 and G2 demonstrated similar AB permeabilities,which were moderate several fold higher than the AB permeability of higher generations. The AB and BA permeability ofG0–G4 typically increased with the increase in donor concentration and incubation time. Permeability values are notreported at generations, concentrations or incubation times that the dendrimers were toxic to Caco-2 cells. TEER valuesdecreased and mannitol permeability increased as a function of donor concentration, incubation time, and generation number.LDH results for G3 and G4 indicate that Caco-2 cell viability was reduced with increasing donor concentration, incubationtime, and generation number. The appreciable permeability of G0–G2, coupled with their nontoxic effects on Caco-2 cells,suggest their potential as water-soluble polymeric drug carriers for controlled oral drug delivery. 2002 Elsevier ScienceB.V. All rights reserved.

Keywords: Poly (amidoamine) dendrimers; PAMAM; Caco-2 cells; Transepithelial transport; Drug delivery

1. Introduction limited by their large size and molecular weight.Variation in the mechanism and net permeability of

Oral drug delivery is the most desirable route of polymeric drug carriers has been attributed to theadministration for many drugs. The oral bioavail- difference in their structural features such as hydro-ability of polymeric drug carriers, however, is often dynamic volume, molecular weight, molecular geom-

etry, and charge [1–4]. For example, transport ofneutral poly (ethylene) glycol (PEG) polymers across*Corresponding author. Tel.: 11-410-706-8650; fax: 11-410-rabbit colonic epithelium appears to occur via the706-0346.

E-mail address: [email protected] (H. Ghandehari). paracellular route and depends on the molecular

0168-3659/02/$ – see front matter 2002 Elsevier Science B.V. All rights reserved.PI I : S0168-3659( 02 )00087-1

356 M. El-Sayed et al. / Journal of Controlled Release 81 (2002) 355 –365

Table 1weight of the polymer where permeability is de-Structural features of PAMAM dendrimerscreased as the molecular weight is increased [1].PAMAM Molecular weight Diameter Number of surfaceWhereas, it has been shown that the transport of

a˚dendrimers (Da) (A) NH groups2dendrimers, across everted rat intestinal sacs andb b bMDCK cells, is due to a combination of factors G0 517 15 4b b bG1 1430 22 8including size and charge [2–4].b b bG2 3256 29 16PAMAM dendrimers are a family of water-soluble b b bG3 6909 36 32

polymers characterized by a unique tree-like branch- b b bG4 14,215 45 64ing architecture and a compact spherical shape in

a Number of surface amine groups for unlabeled polymers.solution [5,6]. Their potential as drug carriers arises b Reported by Tomalia et al. [5].from the large number of arms and surface aminegroups that can be utilized to immobilize drugs,enzymes, antibodies or other bioactive agents [7,8]. 2.2. Fluorescence labeling and fractionation ofInitial studies have shown that PAMAM dendrimers PAMAM dendrimerscan permeate across epithelial barriers suggestingtheir potential as oral drug carriers [2–4]. However, PAMAM dendrimers (G0–G4, Table 1) werelimited data are available for a systematic correlation labeled using FITC following a previously reportedof the structure of these polymers with their transport method [9]. Briefly, PAMAM aqueous solutions wereacross biological barriers, specifically the intestinal diluted 100-fold (v /v) in phosphate buffered saline atepithelia. The present work is an attempt to sys- pH 7.4. The corresponding amount of FITC (poly-tematically investigate the influence of physiochemi- mer:FITC molar ratio51:1) was dissolved in acetonecal parameters such as size, molecular weight, (5 mg/ml), added to unlabeled PAMAM solutions,molecular geometry, and number of surface amine and allowed to stir overnight at room temperature.groups of PAMAM dendrimers, on their permeability Fluorescently labeled PAMAM solutions wereacross Caco-2 cell monolayers. Collectively, the dialyzed against deionized water, fractionated on aresults of this research will aid in understanding the Superose 12 HR 16/50 preparative scale columnrelationship between the structural features of using a Fast Protein Liquid Chromatography SystemPAMAM dendrimers and their intestinal permeabili- (FPLC) (Amersham Pharmacia Biotech), a mobilety, with the prospect of designing novel polymeric phase of 30%:70% (v/v) acetonitrile:Tris buffercarriers for oral drug delivery. (pH58.0), and a flow rate of 1.0 ml /min. Eluting

molecules were detected at 280 nm. Fractions corre-sponding to the size and molecular weight of eachpolymer were collected, dialyzed against deionized

2. Materials and methodswater to remove the mobile phase salts, and sub-sequently lyophilized and stored at 4 8C for the

2.1. Materials permeability experiments.

Aqueous solutions of PAMAM dendrimers (G0– 2.3. Caco-2 cell cultureG4), fluorescein isothiocyanate (FITC), and lactatedehydrogenase assay kits were purchased from Sig- Caco-2 cells were grown at 37 8C in T-75 flasks inma-Aldrich Co. (St. Louis, MO). Superose 12 HR an atmosphere of 5% CO and 95% relative humidi-2

10 /30 column, Superose 12 preparative grade beads ty using Dulbecco’s Modified Eagle’s Medium (pHand HR 16/50 column were purchased from Amer- 7.4) supplemented with 10% fetal bovine serum, 1%sham Pharmacia Biotech (Piscataway, NJ). Caco-2 non-essential amino acids, and 0.05% penicillin–cells were purchased from American Type Cell streptomycin. Cells were passaged at 80–90% con-

14Culture (ATCC, Rockville, MD). [ C] Mannitol fluency using a 0.25% trypsin /0.20% ethylene(specific activity 51.5 mCi /mmol) was purchased diamine tetraacetic acid (EDTA) solution. Mediafrom NEN Life Sciences (Boston, MA). were changed approximately every 48 h. Caco-2

M. El-Sayed et al. / Journal of Controlled Release 81 (2002) 355 –365 357

4cells (passages 30–70) were seeded at 6.3310 nomenix, Torrance, CA), a mobile phase of 12:882cells /cm on polycarbonate 6-well or 12-well (v /v%) acetonitrile:phosphate buffer (pH 7.4), and a

Transwell filters (Corning Costar Corporation, flow rate of 1.0 ml /min. G3 and G4 samples wereCambridge, MA) (3.0 mm mean pore size). Caco-2 analyzed using a PolySep GFC-P3000 column (Phe-

cells were cultured on Transwells under the same nomenix), a mobile phase composed of 30:70 (v /incubation conditions and used for transport experi- v%) acetonitrile:Tris buffer (pH 8.0), and a flow ratements 21–28 days after seeding. of 1.0 ml /min. G0–G4 were all detected at an

excitation wavelength of 495 nm and an emissionwavelength of 518 nm.2.4. Caco-2 permeability of PAMAM dendrimers

2.6. Effect of PAMAM dendrimers on TEER acrossThe transport of fluorescently labeled PAMAM

Caco-2 cell monolayersdendrimers (G0–G4) across Caco-2 cell monolayerswas investigated in triplicate in both AB and BA

The effect of PAMAM dendrimers (G0–G4) ondirections at donor concentrations of 1.0 mM (6-well

the integrity of Caco-2 cell monolayers was investi- 2Transwells , 4.7 cm surface area) and 10.0 mMgated by monitoring the TEER across Caco-2 cell 2(12-well Transwells , 1.0 cm surface area). At 1.0monolayers in the presence of PAMAM dendrimersmM, 6-well Transwells were used to permit quanti-at donor concentrations of 0.1, 1.0 and 10.0 mM

tation of receiver samples (by increasing surfaceupon both apical and basolateral incubation. A

area). Permeability experiments were conducted at Millicell ERS meter (Millipore Corporation, Bed-pH 7.0, 37 8C, 5% CO , 95% relative humidity, and2 ford, MA) connected to a pair of chopstick electrodesshaking at 50 rev. /min while maintaining sink

was used to measure TEER values at 30, 60, 90, 120,conditions. Samples were collected at 30, 90, 150,

150, 180, and 210 min. All experiments wereand 210 min. The 30-min sample was discarded to

conducted in triplicate at pH 7.0, 37 8C, 5% CO and2ensure steady-state transport. The integrity of Caco-295% relative humidity while shaking at 50 rev. /min.

cell monolayers was monitored by measuring TEERControl experiments, in HBSS, were also conducted

before and after each experimentat the same experimental conditions yielding mean

Permeability coefficients (P ) of G0–G4 were 2eff TEER values in the range of 800–1200 V.cm .calculated as previously reported [10]. Control trans-port experiments were also conducted across

2.7. Effect of PAMAM dendrimers on mannitolTranswell filters without Caco-2 cells to determinepermeability across Caco-2 cell monolayers

the filter permeability (P ). The permeability offilter

Caco-2 cell monolayers (P ) was estimated by 14m The permeability of [ C] mannitol (3.2 mM)correcting the effective permeability (P ) for filtereff across Caco-2 cell monolayers was investigated in21 21permeability (P ) according to: P 5 P 1filter eff m triplicate in both AB and BA directions at donor21P . After permeating across Caco-2 cell mono-filter concentrations of 0.1, 1.0, and 10.0 mM PAMAMlayers, the dendrimers were randomly examined for 14dendrimers. Control permeability of [ C] mannitolthe absence of low molecular weight fragments using

was studied in triplicate in blank HBSS in both thesize exclusion chromatography techniques (as de-

AB and BA directions. Permeability experimentsscribed in Section 2.2).

were conducted at pH 7.0, 37 8C, 5% CO , 95%2

relative humidity, and shaking at 50 rev. /min while2.5. HPLC analysis of PAMAM dendrimers maintaining sink conditions. Samples were collected

at 30, 60, 90, 120, 150, 180, and 210 min. TheFluorescently labeled PAMAM samples were ana- 30-min sample was discarded to ensure steady-state

lyzed using a HPLC system (Hewlett Packard, Palo transport. The integrity of Caco-2 cell monolayersAlto, CA) with fluorescence detection (Shimadzu was also monitored by measuring TEER before andCorporation, Kyoto, Japan). G0, G1 and G2 samples after each experiment. Samples were analyzed bywere analyzed using a LUNA C-18 column (Phe- liquid scintillation counting (Beckman Coulter, Ful-


lerton, CA). The monolayer permeability (P ) of order G0.G1.G2.G3.G4 and size exclusionm14[ C] mannitol, in the presence and absence of chromatograms ruled out the presence of free FITC

PAMAM dendrimers, was calculated as described and any small or large molecular weight polymerearlier for PAMAM dendrimers. fragments which could have potentially influenced

permeability results (Fig. 1).2.8. Lactate dehydrogenase (LDH) leakage assay

3.2. Caco-2 permeability of PAMAM dendrimersLDH is a cytosolic enzyme that is not normally

secreted outside the cell. However, upon damage to At 1.0 mM, PAMAM dendrimers demonstrated ancell membranes, it has been proven to leak into the incubation time- and generation-dependent Caco-2culture medium [11]. LDH leakage into the apical permeability profile (Fig. 2). Results indicate G0, G1compartment was used to measure the effect of and G2, demonstrated similar AB permeabilities thatPAMAM dendrimers (G0–G4) on the viability of were up to six fold higher than the AB permeabilityCaco-2 cell monolayers. Caco-2 cell monolayers (12- of larger G3 dendrimers. Under similar conditions,

well Transwells ) were treated with unlabeled G0– the BA permeability of each dendrimer was typicallyG4 at apical donor concentrations of 0.1, 1.0 and higher than its corresponding AB permeability. Per-10.0 mM. Caco-2 cell monolayers were also treated meability of G3 and G4 are not reported at incuba-with blank HBSS and 1% Triton X-100 as negative tion time points where cytotoxicity may have con-and positive controls, respectively. The effect of tributed to the observed permeability values asincubation time of each dendrimer on LDH leakage denoted by LDH results (Fig. 2 and Table 2). Atwas investigated at 90, 150, and 210 min. Experi- 10.0 mM, PAMAM dendrimers demonstrated anments were conducted in triplicate at pH 7.0, 37 8C, incubation time- and generation-dependent per-5% CO and 95% relative humidity while shaking at meability profile which was similar to that observed2

50 rev. /min to mimic the conditions of PAMAM at 1.0 mM (Fig. 3). However, both AB and BApermeability experiments. LDH leakage in the donor permeabilities for a given generation at 10.0 mMcompartment was quantified using an LDH assay kit were typically higher than their corresponding valuesfollowing the manufacturer’s specifications. A one- at 1.0 mM. At longer incubation times, the AB

tailed t-test (Microsoft Excel ) (P,0.01) was usedto identify significant differences between LDHresults for PAMAM dendrimers and the negativecontrol, blank HBSS.

3. Results

3.1. Fluorescence labeling and fractionation ofPAMAM dendrimers

PAMAM dendrimers were fluorescently labeled ata 1:1 molar ratio with FITC. Fractionation of fluores-cently labeled PAMAM dendrimers was based on theprinciples of size exclusion chromatography wherebymolecules with larger hydrodynamic volumes elute Fig. 1. Size exclusion chromatogram of fluorescently labeled

PAMAM dendrimers (G0–G4) after fractionation on a Superoseearlier than those with smaller hydrodynamic vol-12 HR preparative scale column (Amersham Pharmacia Biotech)umes. The elution volumes of the fluorescentlyusing a mobile phase composed of 30:70 (v /v) acetonitrile:Tris

labeled PAMAM dendrimers (G0–G4) were used as buffer (pH58) at a flow rate of 1 ml /min. Detection of elutingindicators of their relative sizes. As expected the molecules was done using a UV detector at a fixed wavelengthelution volumes of PAMAM dendrimers were in the (l5280 nm).


26permeabilities of G0–G2 were moderate (131025and 1310 cm/s) and the BA permeabilities were

25high (greater than 1310 cm/s), respectively,when compared to the permeability of a low perme-able marker compound such as mannitol. Size exclu-sion chromatograms of the randomly selected re-ceiver samples demonstrated elution volumes whichwere similar to those of the original fluorescentlylabeled PAMAM dendrimers without small molecu-lar weight fragments or impurities (data not shown).

3.3. Effect of PAMAM dendrimers on Caco-2 cellmonolayer integrity

TEER values were measured in the presence ofdonor concentrations of 0.1, 1.0, and 10.0 mMPAMAM dendrimers (G0–G4). To summarize thislarge amount of data, however, TEER values at allconcentrations are only illustrated for G0 (Fig. 4A)and G4 (Fig. 4B). The decline in TEER for eachgeneration was concentration-dependent in both theAB and BA directions. The onset and the extent ofdecline in TEER values also changed as a function ofgeneration. The incubation of larger PAMAM den-drimers on the apical side (Fig. 4C) resulted in afaster, more pronounced, and extended decline inTEER values than smaller ones. For example, TEERin the presence of G0 declined slowly for the first 30min and continued to decrease slowly with increasein incubation time. On the other hand, TEERdropped sharply within the first 30 min followingexposure to larger PAMAM generations such as G3and remained constant throughout the rest of theexperiment. Basolateral incubation of PAMAM den-drimers also resulted in a faster and more pro-nounced decline in TEER compared to apical incuba-tion under similar conditions (Fig. 4D).

3.4. Effect of PAMAM dendrimers on paracellularpermeability

Mannitol permeability was investigated acrossCaco-2 cell monolayers in both AB and BA direc-

Fig. 2. AB (h) and BA (j) Caco-2 permeability of PAMAM tions at donor concentrations of 0.1, 1.0 and 10.0dendrimers (G0–G4) at a donor concentration of 1.0 mM and mM of each generation. Even though results indicateincubation times of: (A) 90 min, (B) 150 min, and (C) 210 min.

that both AB and BA mannitol permeability in-AB and BA permeability values are not reported (**) at toxiccreased with incubation time, data are only reportedincubation time points. Results are reported as mean 6 standard

error of the mean (S.E.M.). at the last incubation time point, 210 min (Fig.


Table 2Lactate dehydrogenase (LDH) assay of PAMAM dendrimers

a a aIncubation time LDH leakage % LDH leakage % LDH leakage %(min) (90 min) (150 min) (210 min)

HBSS 14.662.1% 16.161.2% 6.360.8%PAMAM-G0 0.1 mM – – 6.760.3%

1.0 mM 11.961.6% 8.360.5% 5.161.8%10.0 mM 16.361.5% 8.060.4% 6.261.4%

PAMAM-G1 0.1 mM – – 5.461.0%1.0 mM 15.963.0 9.462.1% 4.760.2

10.0 mM 16.561.1 10.361.4% 7.960.5%PAMAM-G2 0.1 mM – – 8.760.5%

1.0 mM 19.961.2% 6.963.8% 8.560.7%10.0 mM 42.263.1% 20.961.5% 33.462.0%

PAMAM-G3 0.1 mM – – 12.160.7%1.0 mM 16.361.2% 14.062.1% 11.060.5%

10.0 mM 25.261.1% 19.860.7% 19.261.9%PAMAM-G4 0.1 mM – – 14.961.9%

1.0 mM 46.063.0% 39.561.5% 29.660.9%10.0 mM 221.0614.8% 210.9611.1% 141.462.7%

Triton X-100 100.0634.6% 100.0623.2% 100.0622.2%

Italicized cells denote a significant increase in leakage of LDH compared to that in HBSS based upon a P-value ,0.01 using a one tailt-test.

a Results are reported as % of LDH leakage observed upon incubation with the positive control, Triton X-100.

5A–C). Mannitol permeability in the presence of was no significant LDH leakage upon incubation ofPAMAM dendrimers is reported in reference to Caco-2 cell monolayers with G0 and G1 at con-control AB and BA mannitol permeabilities, which centrations of 0.1, 1.0 and 10.0 mM for an incuba-

26 26were 1.26310 cm/s and 1.96310 cm/s, re- tion time of up to 210 min. G2 induced a statisticallyspectively. At 0.1 mM, both AB and BA mannitol significant leakage of LDH only at 10.0 mM at 210permeability increased as much as six fold (Fig. 5A). min. G3 induced a statistically significant LDHAt 1.0 mM, AB and BA mannitol permeability leakage at 210 min at donor concentrations of 0.1,increased as much as 4- and 16-fold, respectively 1.0, and 10.0 mM. G4 induced a statistically signifi-(Fig. 5B). At 10.0 mM, AB and BA mannitol cant LDH leakage at all the concentrations andpermeability increased as much as 8- and 13-fold, incubation times investigated.respectively (Fig. 5C). Results indicate that the BAmannitol permeability was typically higher than itscorresponding AB permeability under similar ex- 4. Discussionperimental conditions. Both AB and BA permeabilityof mannitol increased with the increase in PAMAM Considering the reported molecular weight /diam-generation number. Only the permeability values at eter and the size exclusion profiles of fluorescentlynon-toxic PAMAM concentrations, as evident by labeled PAMAM dendrimers (Table 1, Fig. 1), oneLDH results (Table 2), are reported. would expect their Caco-2 permeability to decrease

with the increase in generation number. However,3.5. Effect of PAMAM dendrimers on Caco-2 cell the Caco-2 permeability of G0–G4 did not fit aviability typical size- and/or molecular weight-restricted per-

meability profile [1]. The AB permeability of smallerThe influence of PAMAM dendrimers on the PAMAM dendrimers, G0, G1 and G2, showed

viability of Caco-2 cell monolayers, as denoted by similar and moderate permeability values despite theLDH leakage results, is reported in Table 2. There incremental increase in their molecular weights. The


moderate permeability of these dendrimers may beattributed to structural features such as charge,molecular weight and geometry. Compared to neutralmolecules (such as urea and mannitol), or anionicmolecules (such as formate and lactate), positivelycharged molecules (such as methylamine andatenolol) have been reported to permeate at a higherrate across Caco-2 cell monolayers due to a favor-able electrostatic interaction with the negativelycharged epithelial surfaces [12]. Similarly, positivelycharged polymers, such as chitosan derivatives, havebeen shown to enhance paracellular permeabilityacross Caco-2 cell monolayers through modulationof tight junctions [13–17]. The extent of this modu-lation has also been shown to depend on the numberof positively charged groups available to interactwith the negatively charged epithelium [17]. Accord-ing to the Henderson–Hasselbalch equation, 41.5%of the surface primary amine groups in PAMAMdendrimers are positively charged at pH 7.0, whilethe interior tertiary amine groups are almost un-ionized. Therefore, it is possible that the cationicnature of PAMAM dendrimers could have facilitatedtheir interaction with Caco-2 cell epithelia, resultingin modulation of the tight junctions and a subsequentincrease in their permeability, probably via theparacellular route. The interaction between PAMAMdendrimers and Caco-2 epithelial surface would beexpected to increase with generation number, due tothe increase in the number of positively chargedsurface amine groups (Table 1). Since the per-meability of mannitol increased with increase ingeneration number (Fig. 5), results would support adirect interaction of PAMAM dendrimers with Caco-2 epithelia.

In addition to polymer charge, the change inmolecular weight of the polymer has an impact onCaco-2 permeability. Results indicate that for smallerPAMAM dendrimers (G0–G2), Caco-2 permeabilitydid not decrease with an increase in generationnumber, despite the relatively high molecular weightof G1 and G2. This may be attributed to the increasein the number of positively charged amine groups as

Fig. 3. AB (h) and BA (j) Caco-2 permeability of PAMAM the generation number increased. In fact, the ABdendrimers (G0–G4) at a donor concentration of 10.0 mM and permeability of G0–G2 was higher than the per-incubation times of: (A) 90 min, (B) 150 min, and (C) 210 min.

meability of commonly used neutral polymeric drugAB and BA permeability values are not reported (**) at toxiccarriers of corresponding molecular weights, such asincubation time points. Results are reported as mean 6 standard

error of the mean (S.E.M.). PEG 900 and PEG 4000 [1], although the per-


Fig. 4. Effect of PAMAM dendrimers: G0 (A), G4 (B) at 0.1 mM (♦), 1.0 mM (m), and 10.0 mM (d), and HBSS (j) on transepithelialelectrical resistance (TEER) across Caco-2 cell monolayers upon apical incubation. Effect of PAMAM generation: G0 (♦), G1 (m), G2 (d),G3 (3), G4 (*), and HBSS (j) on TEER values across Caco-2 cell monolayers at 10.0 mM on the apical side (C), and on the basolateralside (D).

meability of these polymers was assessed across Incubation time-dependent permeability may be duerabbit colonic epithelium. Despite the higher number to a prolonged interaction of these dendrimers withof positively charged surface amine groups, at non- the Caco-2 epithelia, resulting in a more pronouncedtoxic donor concentrations, permeability of G3 was modulation of tight junctions. This hypothesis ismuch lower than G0–G2 probably due to its higher supported by the observed decrease in TEER withmolecular weight. time for all generations (Fig. 4). Concentration-de-

In addition to the observed generation-dependent pendent permeability is not normally observed forpermeability of PAMAM dendrimers, AB and BA passively absorbed molecules. Even though thepermeability also demonstrated dependence on incu- mechanism of permeation of PAMAM dendrimersbation time and concentration (Figs. 2 and 3). has not yet been fully investigated, it is possible that


increasing the concentration of positively chargedamine groups in solution increased the degree ofinteraction of PAMAM with Caco-2 epithelia, re-sulting in the higher observed permeability at higherconcentrations. This interpretation is supported bysimilar observations that increasing the concentrationor the degree of quaternization of N-trimethylchitosan, a positively charged water-soluble polymer,results in greater modulation of the tight junctionsand a subsequent increase in the mannitol per-meability [14,17]. The BA permeabilities ofPAMAM dendrimers were also typically higher thantheir corresponding AB permeabilities under similarconditions (Figs. 2 and 3). This may be attributed toeither differences in tight junctional structure at theapical and basolateral membranes or efflux by secret-ory transporters. Caco-2 cell monolayers possesstight junctional proteins which are present at theapical but not at the basolateral membrane. Theseproteins not only serve to connect adjacent cells, butare also responsible for maintaining the integrity oftight junctions [18]. Similar observations have beenreported upon application of EDTA to the apical andbasolateral membranes of Caco-2 cell monolayers,whereby basolateral application of EDTA resulted ina more pronounced decline in TEER and higherparacellular permeability compared to the apicalapplication of equivalent EDTA concentrations [18].This directional permeability could also be due to thepotential affinity of the studied dendrimers for effluxtransporters, such as P-glycoprotein (P-gp), since ithas been demonstrated that highly cationic moleculestend to have high affinity for P-gp [19]. However,efflux cannot be the only factor contributing to thedirectional permeability of these dendrimers sincemannitol, which is not a substrate for P-gp, alsoexhibited a higher BA versus AB permeability undersimilar experimental conditions (Fig. 5). Additionalstudies are warranted to evaluate the mechanism oftransport of these dendrimers across Caco-2 cellmonolayers.

TEER values (Fig. 4) and mannitol permeability(Fig. 5) were used to assess the effect of PAMAM

14Fig. 5. AB (h) and BA (j) Caco-2 permeability of [ C] dendrimers (G0–G4) on the integrity and paracellu-mannitol at an incubation time of 210 min in the presence of: (A) lar permeability of Caco-2 cell monolayers, respec-0.1 mM; (B) 1.0 mM; (C) 10.0 mM solution of PAMAM

tively. Results indicate that TEER values decreasedendrimers (G0–G4). AB and BA permeability values are notwith PAMAM concentration and generation number.reported (**) at toxic concentrations. Results are reported as

mean6standard error of the mean (S.E.M.). Low TEER values and high mannitol permeabilities


across Caco-2 cell monolayers are typically observed suggests their potential utility in controlled oral drugto be indicative of cell monolayer disruption via delivery. Results suggest that by adjusting the geom-cytotoxicity. PAMAM dendrimers have been previ- etry, molecular weight, surface charge, and residenceously identified to demonstrate generation-dependent time of PAMAM dendrimers in the GI tract it isin vitro hemolytic [20] and myotoxic effects [21]. possible to control their transepithelial transport.Consistent with these findings, LDH leakage resultsindicate that small PAMAM dendrimers (G0–G2)were generally non-toxic to Caco-2 cell monolayers References(except for G2 at 10.0 mM at 210 min), while largerdendrimers (G3 and G4) exhibited toxicity (Table 2).

[1] H. Ghandehari, P.L. Smith, H. Ellens, P.-Y. Yeh, J. Kopecek,PAMAM dendrimers caused a decrease in TEER Size-dependent permeability of hydrophilic probes acrossvalues and increase in mannitol permeability at rabbit colonic epithelium, J. Pharmacol. Exp. Ther. 280

(1997) 747–753.concentrations that were shown to be non-toxic to[2] R. Wiwattanapatapee, B. Carreno-Gomez, N. Malik, R.Caco-2 cell monolayers. The effects of these den-

Duncan, Anionic PAMAM dendrimers rapidly cross adult ratdrimers on TEER and mannitol permeability couldintestine in vitro: a potential oral delivery system?, Pharm.

be attributed to tight junctional modulation due to the Res. 17 (2000) 991–998.cationic nature of PAMAM dendrimers. The progres- [3] R. Wiwattanapatapee, B. Carreno-Gomez, N. Malik, R.

Duncan, PAMAM dendrimers as a potential oral drugsive decline in TEER values or increase in mannitoldelivery system: uptake by everted rat intestinal sacs in-vitro,permeability with the increase in incubation time,J. Pharm. Pharmacol. 50 (Suppl.) (1998) 99.would also support this hypothesis due to the pro-

[4] F. Tajarobi, M. El-Sayed, B. Rege, J. Polli, H. Ghandehari,longed interaction of the positively charged amine Transepithelial transport of poly amidoamine dendrimersgroups with the Caco-2 epithelial surface. It is across Madin–Darby Canine Kidney cells, Int. J. Pharm. 215currently unclear how these results would correlate (2001) 263–267.

[5] D.A. Tomalia, A.M. Naylor, W.A. Goddard III, Starburstto intestinal permeability in humans, since it hasdendrimers: molecular-level control of size, shape, surfacebeen shown that the comparative biology of the tightchemistry, topology, and flexibility from atoms to macro-

junctions of Caco-2 monolayers and the human small scopic matter, Angew. Chem. Int. Ed. Engl. 29 (1990)intestine are different [22,23]. More specifically, 138–175.Caco-2 monolayers are generally recognized as [6] D.A. Tomalia, StarburstE / cascade dendrimers: fundamental

building blocks for a new nanoscopic chemistry set, Aldrich-epithelium with more ‘restrictive’ overall tight junc-im. Acta 26 (1993) 91–101.tional structure than human small intestinal epi-

[7] D.S. Wilbur, P.M. Pathare, D.K. Hamlin, K.R. Buhler, R.L.thelium [22,23]. Clearly additional studies to identify Vessella, Biotin reagents for antibody pretargeting. 3. Syn-the mechanism of permeation for these dendrimers thesis, radioiodination, and evaluation of biotinylated star-would assist in interpreting these results. burst dendrimers, Bioconjug. Chem. 9 (1998) 813–825.

[8] N.A. Peppas, Star polymers and dendrimers: prospects oftheir use in drug delivery and pharmaceutical applications,Control. Release Soc. Newslett. 12 (1995) 12–13.

5. Conclusions [9] M. El-Sayed, M.F. Kiani, M.D. Naimark, A.H. Hikal, H.Ghandehari, Extravasation of poly (amidoamine) (PAMAM)

The permeability of PAMAM dendrimers across dendrimers across microvascular network endothelium,Pharm. Res. 18 (1) (2001) 23–28.Caco-2 cell monolayers appears to depend on a

[10] D. Fleisher, Gastrointestinal transport of peptides: ex-combination of structural features such as their size,perimental systems, in: M.D. Taylor, G.L. Amidon (Eds.),

molecular weight, geometry and number of surface Peptide-based Drug Design: Controlling Transport and Me-amine groups. PAMAM dendrimers enhanced their tabolism, American Chemical Society, Washington, DC,own permeability, as well as the permeability of 1995, pp. 501–523.

[11] S.-J. Wu, J. Robinson, Transcellular and lipophilic complex-mannitol. At higher concentrations and incubationenhanced intestinal absorption of human growth hormone,times, G3 and G4 demonstrated toxic effects toPharm. Res. 16 (1999) 1266–1272.

Caco-2 cell monolayers. The moderate permeability [12] G.T. Knipp, N.F.H. Ho, C.L. Barsuhn, R.R. Borchardt,of PAMAM generations G0, G1 and G2 across Caco- Paracellular diffusion in Caco-2 cell monolayers: effect of2 cell monolayers, and the lack of cellular toxicity, perturbation on the transport of hydrophilic compounds that


vary in charge and size, J. Pharm. Sci. 86 (1997) 1105– [18] A.B.J. Noach, Y. Kurosaki, M.C.M. Blom-Roosemalen,1110. A.G.d. Boer, D.D. Breimer, Cell-polarity dependent effect of

[13] G. Borchard, H.L. Lueben, A.B.G.d. Boer, J.C. Verhoef, chelation on the paracellular permeability of confluent Caco-C.-M. Lehr, H.E. Junginger, The potential of mucoadhesive 2 cell monolayers, Int. J. Pharm. 90 (1993) 229–237.polymers in enhancing intestinal peptide drug absorption. III. [19] A. Seelig, X.L. Blatter, F. Wohnsland, Substrate recognitionEffects of chitosan glutamate and carbomer on epithelial by P-glycoprotein and the multidrug resistance-associatedtight junctions in vitro, J. Control. Release 39 (1996) 131– protein MRP1: a comparison, Int. J. Clin. Pharmacol. Ther.138. 38 (2000) 111–121.

[14] A.F. Kotze, H.L. Lueben, B.J.d. Leeuw, A.B.G.d. Boer, J.C. [20] N. Malik, R. Wiwattanapatapee, R. Klopsch, K. Lorenz, H.Verhoef, H.E. Junginger, N-Trimethyl chitosan chloride as a Frey, J.W. Weener, E.W. Meijer, W. Paulus, R. Duncan,potential absorption enhancer across mucosal surfaces: in Dendrimers: relationship between structure and biocom-vitro evaluation in intestinal epithelial cells (Caco-2), Pharm. patability in vitro, and preliminary studies on the biodistribu-

125Res. 14 (1997) 1197–1202. tion of I-labelled polyamidoamine dendrimers in vivo, J.[15] A.F. Kotze, M.M. Thanou, H.L. Lueben, A.B.G.d. Boer, J.C. Control. Release 65 (2000) 133–148.

Verhoef, H.E. Juninger, Enhancement of paracellular drug [21] G.A. Brazeau, S. Attia, S. Poxon, J.A. Hughes, In vitrotransport with highly quaternized N-trimethyl chitosan chlo- myotoxicity of selected cationic macromolecules used inride in neutral environments: in vitro evaluation in intestinal non-viral gene delivery, Pharm. Res. 15 (1998) 680–684.epithelial cells (Caco-2), J. Pharm. Sci. 88 (1999) 253–257. [22] P. Artursson, A. Ungell, J. Lofroth, Selective paracellular

[16] A.F. Kotze, H.L. Lueben, B.J.d. Leeuw, A.B.G.d. Boer, J.C. permeability in two models of intestinal absorption: culturedVerhoef, H.E. Junginger, Comparison of the effect of differ- monolayers of human intestinal epithelial cells and ratent chitosan salts and N-trimethyl chitosan chloride on the intestinal segments, Pharm. Res. 10 (1993) 1123–1129.permeability of intestinal epithelial cells, J. Control. Release [23] A. Collett, E. Sims, D. Walker, Y.L. He, J. Ayrton, M.51 (1998) 35–46. Rowland, G. Warhurst, Comparison of HT-29-18-C1 and

[17] M.M. Thanou, B.I. Florea, M.W.E. Langemeyer, J.C. Ver- Caco-2 cell lines as models for studying intestinal paracellu-hoef, H.E. Junginger, N-Trimethylated chitosan chloride lar drug absorption, Pharm. Res. 13 (1996) 216–221.(TMC) improves the intestinal permeation of the peptidedrug buserelin in vitro (Caco-2 cells) and in vivo (rats),Pharm. Res. 17 (2000) 27–31.

transepithelial transport of poly(amidoamine) dendrimers across caco-2 cell monolayers

Documents