MA nanoparticles on their bioadhesive properties

Journal of Controlled Release 89 (2003) 19–30 www.elsevier.com / locate / jconrel

Influence of the surface characteristics of PVM / MA nanoparticles on their bioadhesive properties ´ a , M.A. Campanero b , M.A. Arangoa a , M.J. Renedo a , J.M. Irache a , * P. Arbos a

´ ´ Farmaceutica ´ , Departamento Farmacia y Tecnologıa , Universidad de Navarra, Irunlarrea, 1, 31080 Pamplona, Centro Galenico Spain b ´ Clınica ´ ´ , Clınica Universitaria de Navarra, 31008 Pamplona, Spain Servicio de Farmacologıa Received 13 September 2002; accepted 2 January 2003

Abstract The aim of this work was to investigate the influence of the cross-linkage of poly(methylvinylether-co-maleic anhydride) (PVM / MA) nanoparticles with increasing amounts of 1,3-diaminopropane (DP) and, eventually, bovine serum albumin (BSA) on their gastrointestinal transit and bioadhesive properties. The fluorescently-labelled formulations were orally administered to rats and, at different times, the amount of nanoparticles in both the lumen content and adhered to the gut mucosa were quantified. The gut transit was evaluated by calculating the gastric (k ge ) and intestinal (k ie ) emptying rates. The adhered fraction of nanoparticles in the whole gut was plotted versus time and, from these curves, the intensity, capacity and extent of the adhesive interactions were estimated. The bioadhesive potential of PVM / MA was much higher when formulated as nanoparticles (NP) than in the solubilised form in water. However, k ge and k ie increased by increasing the extent of cross-linkage of nanoparticles with DP, while the capacity to develop adhesive interactions and the intensity of the adhesive phenomenon were significantly higher for non-hardened than for DP-cross-linked carriers. In contrast, the BSA-coating of cross-linked nanoparticles significantly decreased k ge and k gi , whereas the intensity of the bioadhesive phenomenon was significantly higher than for NP. In summary, the adhesivity of the nanoparticles appears to modulate their gastrointestinal transit profile.  2003 Elsevier Science B.V. All rights reserved. Keywords: Nanoparticles; Bioadhesion; Poly(methylvinylether-co-maleic anhydride); Gastrointestinal transit; Gantrez

1. Introduction Nanoparticles have been proposed as drug carriers for the oral administration of poorly available molecules. Many researchers have reported the capacity of polymeric nanoparticles to improve the oral *Corresponding author. Tel.: 134-948-425600; fax: 134-948425649. E-mail address: [email protected] (J.M. Irache).

bioavailability of drugs with poor absorption characteristics [1–5]. These colloidal carriers are expected to develop adhesive interactions within the mucosa and remain in the gastrointestinal tract, while protecting the entrapped drug from enzymatic degradation, until the release of the loaded drug or their absorption in an intact particulate form. Quantification of bioadhesive forces between particulates and the biological support is a useful indicator for screening drug delivery systems de-

0168-3659 / 03 / $ – see front matter  2003 Elsevier Science B.V. All rights reserved. doi:10.1016 / S0168-3659(03)00066-X

20

´ et al. / Journal of Controlled Release 89 (2003) 19–30 P. Arbos

signed to increase the residence time within the gastrointestinal tract. However, the number of studies dealing with the bioadhesion of particulate systems is limited. Various methods have been developed to assess the bioadhesive properties of particulate dosage forms. Many of them involve the administration of particulates to laboratory animals and quantifying the amount of particles remaining in the gastrointestinal system. Administration methods include forced oral gavage [6–8], surgical stomach implantation [9] and infusion through an in situ loop in the small intestine [10,11]. Tracking has been carried out with radiopaque markers [9], radioactive elements [6,7,12], and fluorescent tags [8,13]. Recently, the analysis of the adhesion curves (cumulative amount of adhered particles in the gut vs. time) has been proposed to quantify (i) the affinity of the material for the mucosa, (ii) the intensity and (iii) relative duration of the bioadhesive phenomenon and (iv) the elimination rate of the adhered particles [14]. This approach permits the estimation of the bioadhesive potential of a given drug delivery system and to make easy comparisons between different formulations. On the other hand, the bioadhesivity of particulates can also be discerned from the determination of the gastric emptying of a single dose orally administered to animals [15,16]. Thus, it is possible to estimate the rate of gastric emptying and the small intestine transit as well as the time-course of distribution of the particles. Poly(methyl vinyl ether-co-maleic anhydride) (PVM / MA) is a biodegradable polyanhydride widely used for pharmaceutical purposes and could be an appropriate copolymer for the preparation of nanoparticulate dosage forms with bioadhesive or mucoadhesive properties [17]. In fact, when polyanhydrides hydrolytically degrade, the product of each cleaved anhydride bond contains two carboxylic acid groups. In accordance with the adsorption theory of adhesion [18], carboxylic groups would enhance the ability of polymers to form hydrogen bonds with components from the mucosa. In this way, poly(fumaric-co-sebacic) anhydride microparticles have demonstrated their high potential to develop bioadhesive interactions with small intestinal tissues in rats [19]. The general aim of this work was to determine the

in vivo behaviour of PVM / MA nanoparticles after the oral administration to laboratory animals. More particularly, we report here the influence of the surface characteristics of different PVM / MA nanoparticle formulations on their gastrointestinal transit profiles and bioadhesive properties. Modifications on the surface of nanoparticles were carried out by their treatment with different amounts of 1,3-diaminopropane and, eventually, by their supplementary coating with albumin.

2. Material and methods

2.1. Chemicals Poly(methyl vinyl ether-co-maleic anhydride), average molecular weight of 200 000 (Gantrez  AN 119) was kindly gifted by ISP (Barcelona, Spain). Rhodamine B isothiocyanate (RBITC), 1,3diaminopropane (DP) and bovine serum albumin (BSA) were supplied by Sigma (St Louis, MO, USA). All other chemicals used were of reagentgrade and obtained from Merck (Darmstadt, Germany).

2.2. Preparation of nanoparticles and ligand conjugates Poly(methyl vinyl ether-co-maleic anhydride) (PVM / MA) nanoparticles were prepared by a solvent displacement method [17]. In brief, 100 mg PVM / MA copolymer was dissolved in 5 ml acetone and desolvated with 20 ml of an ethanol / water phase (1:1 by volume) under magnetic stirring. The organic solvents were eliminated under reduced pressure ¨ (Buchi R-144, Switzerland) and the freshly prepared PVM / MA carriers were firstly incubated with 1.25 mg RBITC for 5 min at room temperature and, eventually, with a cross-linking agent. For the crosslinkage, the resulting fluorescently-labelled nanoparticles (NP) were hardened by incubation at room temperature for 5 min with either 0.005 (D5NP), 0.01 (D10-NP) or 0.03 (D30-NP) mg DP/ mg bulk polymer. In addition, some D10-NP batches were also incubated at room temperature for 2 h with

´ et al. / Journal of Controlled Release 89 (2003) 19–30 P. Arbos

3 mg BSA / mg bulk polymer (BD-NP). All the nanoparticulate batches were purified by centrifugation at 17 000 rpm for 15 min (Rotor 3336, Biofuge Heraeus, Germany), the supernatants removed and the pellets resuspended in water. The purification procedure was repeated twice and, finally, the formulations were freeze-dried in a Genesis 12EL apparatus (Virtis, USA) using sucrose (5%) as cryoprotector. As control, an aqueous solution of the copolymer was used (Gantrez-sol). For this purpose, NP were dispersed in water and maintained at room temperature until complete dissolution.

21

2.4. Stability of PVM /MA nanoparticles in simulated gastric and intestinal fluids PVM / MA nanoparticles were incubated in simulated gastric (pH 1.2; pepsin 0.32% w / v) and intestinal (pH 7.5; pancreatin 1% w / v) fluids (USP XXIII) at a concentration of 10 mg / ml. At each time interval, samples were withdrawn, diluted with PBS, and the concentration of the remaining particles estimated by turbidimetry at 450 nm, in a spectrophotometer (Hewlett-Packard, USA). In addition, the size of particles was determined by PCS.

2.5. In vitro release of RBITC from nanoparticles 2.3. Characterisation of PVM /MA formulation The particle size and the zeta potential of nanoparticles were determined by photon correlation spectroscopy (PCS) and electrophoretic laser Doppler anemometry, respectively, using a Zetamaster analyser system (Malvern Instruments, UK). Samples were always diluted with 0.05 mM phosphatebuffered saline (PBS, pH 7.4) and measured at 25 8C with a scattering angle of 908. The amount of PVM / MA copolymer transformed into nanoparticles was determined by gravimetry. For this purpose, PVM / MA nanoparticles were freeze-dried, and the yield was calculated as the difference between the initial amount of the polymer used to prepare nanoparticles and the weight of the freeze-dried carriers. The amount of loaded RBITC in the carriers was estimated by spectrofluorimetry (Perkin Elmer, USA) at lex 554 nm and lem 575 nm, as the difference between the initial marker added and the RBITC content determined in the supernatants obtained during the purification step. Finally, the amount of albumin coating D10-NP was calculated by the difference between the total amount of BSA used to prepare the batch and the amount of protein quantified in the aqueous supernatants, and determined using the microbichinchoninic acid (MicroBCA) protein assay (Pierce  , Rockford, USA). Calibration curves were made from the supernatants of blank nanoparticles. Each sample was assayed in triplicate and results were expressed as the amount of albumin per mg nanoparticle.

The release of RBITC from PVM / MA formulations was studied by incubating nanoparticles (10 mg / ml) in simulated gastric and intestinal fluids (USP XXIII). At each time interval, the suspensions of nanoparticles were centrifuged, at 22 500 rpm for 15 min, and the RBITC released quantified by spectrofluorimetry (Perkin Elmer, USA) at lex 554 nm and lem 575 nm.

2.6. Gastrointestinal transit studies The gastrointestinal transit studies were carried out using the protocols described previously [8,14], in compliance with the regulations of the responsible committee of the University of Navarra in line with the European legislation on animal experiments (86 / 609 / EU). Male Wistar rats, average weight 225 g (CIFA, Universidad de Navarra, Spain), were housed under normal conditions with free access to food and water. The animals were placed in metabolic cages and fasted overnight to prevent coprophagia but allowing for free access to water. Rats were fed with 1-ml aqueous suspensions of the different formulations, containing 10 mg nanoparticles (around 45 mg particles / kg body weight). The animals were sacrificed by cervical dislocation at 0.5, 1, 3 and 8 h post administration. The abdominal cavity was opened and the gastrointestinal tract removed. Then, the gut was divided into four anatomical regions: stomach, small intestine (cut in four portions of about 15 cm), caecum and colon. Each mucosa segment was opened lengthwise along the mesentery and rinsed with either 20 (intestinal

22

´ et al. / Journal of Controlled Release 89 (2003) 19–30 P. Arbos

portions) or 40 ml (stomach, caecum and colon) of physiological saline (0.9% NaCl) in order to recover the non-adhered fraction. These rinsing liquids were centrifuged at 4000 rpm for 10 min, and the pellets digested in 1 ml 3 M NaOH for 24 h. The remaining residues were diluted to 50 ml with water, and centrifuged again at 4000 rpm for 10 min. In all cases, supernatants and extracts were assayed for RBITC content by spectrofluorimetry in order to estimate the amount of particles in the lumen content. Furthermore, each rinsed mucosa segment was cut into portions of 2-cm lengths and digested in 1 ml 3 M NaOH for 24 h. RBITC was extracted with 2 ml methanol, vortexed for 1 min and centrifuged at 4000 rpm for 10 min. Aliquots (1 ml) of the obtained supernatants were diluted with water (3 ml) and assayed for RBITC content by spectrofluorimetry. These data enabled us to estimate the fraction of nanoparticles adhered to the mucosa. Standard curves, either in rinsing liquids or mucosal segments, were carried out daily to quantify RBITC content.

2.7. Tissue proceeding The presence of RBITC-loaded PVM / MA nanoparticles in the gastrointestinal mucosa was visualised in a microscope (Olympus CH-40, Japan) with fluorescence source (Olympus U-RFLT50, Japan). Mucosa portions of about 2-cm treated with the tissue-proceeding medium O.C.T.姠, were immersed in melting isopentane and frozen with liquid nitrogen. Each sample was cut into 3-mm sections on a cryostat (2800 Frigocut E, Reichert-Jung, Germany), attached to gelatin precoated slides and stored at 220 8C before fluorescence microscopy visualisation.

2.8. Transit parameters The transit of PVM / MA nanoparticles through the gastrointestinal tract was evaluated using a modification of the kinetic model designed by Akiyama et al. [15], and modified by Sakuma et al. [16]. It was assumed that nanoparticles were emptied, from the stomach to the intestine and from this region to the caecum, following a zero-order kinetic. The gastric emptying rate (k ge ) was defined as the terminal

elimination rate of the particles from the stomach to the intestine. The intestinal emptying rate (k ie ) was defined as the terminal elimination rate of the particles from the intestine to the caecum. The amount of nanoparticles in each compartment was estimated from the RBITC measurement as described above. Each rate constant was calculated with the pharmacokinetic software WinNonLin 1.5 (Scientific Consulting, Inc.).

2.9. Bioadhesion parameters For each nanoparticulate formulation, the total adhered fraction in the whole gastrointestinal tract was plotted versus time and, from these curves, the parameters of bioadhesion (Q max , AUC adh , k adh and MRT adh ) estimated as described previously [14]. Q max was defined as the maximal amount of nanoparticles adhered to the gut surface and is related to the capability of the material to develop adhesive interactions. k adh was defined as the terminal elimination rate of the adhered fraction with the gastrointestinal mucosa and calculated using the WinNonlin 1.5 software. AUC adh or the area under the curve of bioadhesion was evaluated by means of the trapezoidal rule up to t z , which denoted the last sampling point, and permitted to quantify the intensity of the bioadhesive phenomenon. Finally, MRT adh was the mean residence time of the adhered fraction of nanoparticles in the mucosa and evaluated the relative duration of the adhesive interactions.

2.10. Statistical methods Parameters were analysed to determine statistical significance. The Pearson test was used to determine the significant differences among bioadhesion parameters. In addition, the Mann–Whitney U-test was carried out on transit and bioadhesion parameters of PVM / MA formulations. In both cases, P,0.05 and P,0.01 were considered to be significant.

3. Results

3.1. Characterisation of PVM /MA nanoparticles Table 1 summarises the main physico-chemical

´ et al. / Journal of Controlled Release 89 (2003) 19–30 P. Arbos

23

Table 1 Physico-chemical characteristics of PVM / MA nanoparticles (n56)

Gantrez-sol NP D5-NP D10-NP D30-NP BD-NP a

Size (nm)

Zeta potential a (mV)

RBITC content (mg / mg)

BSA bound (mg / mg)

– 27961 28965 28864 30769 308610

– 241.160.5 239.061.8 234.860.5 228.861.8 216.961.0

9.9560.45 10.3360.87 10.2960.65 10.0460.38 3.6060.03 14.4460.18

– – – – – 299.3611.2

Determined in phosphate buffer (pH 7.4; 0.05 M).

characteristics of the different PVM / MA formulations tested. The carriers were prepared by a desolvation process which enabled us to obtain a yield of about 73.862.6% of the copolymer transformed into nanoparticles. Typically, conventional nanoparticles (NP) were of around 280 nm and negatively charged and the amount of RBITC incorporated in these carriers was calculated to be about 10 mg per mg nanoparticle. The cross-linkage of NP with DP slightly modified the size of the resulting carriers. However, zeta potential significantly decreased as a function of the cross-linker concentration used to harden the carriers. Surprisingly, cross-linked nanoparticles with 1,3-diaminopropane (10 mg / mg) and BSA (BD-NP) displayed the lowest zeta potential. Similarly, the cross-linkage of nanoparticles with high amounts of DP (30 mg / mg), dramatically decreased the incorporation of RBITC to the carriers. In contrast, the addition of BSA to coat the particles (BD-NP formulation) enabled us to load about 40% more RBITC when compared with NP.

3.2. Stability of PVM /MA formulations in model gastrointestinal fluids To evaluate the stability and the RBITC release of PVM / MA formulations in the presence of digestive fluids, the different nanoparticulate formulations were incubated in simulated gastric and intestinal fluids (Table 2). In any case, NP formulation was degraded more rapidly than cross-linked nanoparticles and, indeed, the amount of RBITC released after incubation with simulated fluids was always significantly higher for NP than for cross-linked nanoparticles. On the other hand, cross-linkage with increasing amounts of DP decreased the degradation rate of nanoparticles and the fluorescent marker release.

3.3. Distribution of PVM /MA nanoparticles in the GI tract Fig. 1 shows the amounts of the different PVM / MA formulations in the stomach, intestinal segments

Table 2 Degradation of PVM / MA nanoparticles when incubated for 2 h with either gastric or intestinal fluids Gastric fluid

NP D5-NP D10-NP D30-NP BD-NP

Intestinal fluid

Remaining particles (%)

RBITC in particles (%)

Remaining particles (%)

RBITC in particles (%)

58.164.9 65.764.3 70.260.2 92.963.9 84.062.6

70.263.4 70.161.4 78.468.5 84.562.1 72.362.0

63.563.7 82.161.7 78.165.5 95.060.5 85.267.9

60.061.5 80.361.4 85.562.6 86.163.3 80.564.0

24

´ et al. / Journal of Controlled Release 89 (2003) 19–30 P. Arbos

Fig. 1. Distribution of PVM / MA nanoparticles in the gastrointestinal mucosa (adhered fraction) and lumen content (non-adhered fraction), after the oral administration of 1 ml aqueous dispersion containing 10 mg nanoparticles. Each value represents the mean of the results of four experiments. (A) NP, (B) D5-NP, (C) D10-NP, (D) D30-NP, and (E) BD-NP. Plot: x-axis represents the adhered fraction (mg); y-axis represents the different gut segments (Sto: stomach; I1, I2, I3, I4: intestinal portions; Ce: caecum); z-axis represent the time postadministration (h).

´ et al. / Journal of Controlled Release 89 (2003) 19–30 P. Arbos

and caecum at 0.5, 1, 3 and 8 h after the oral administration of an aqueous dispersion containing 10 mg nanoparticles. The first interesting thing was that NP and DP cross-linked nanoparticles displayed an initial tropism for the stomach mucosa and the upper regions of the small intestine (mainly I2 portions). Three hours post-administration, the lumen content and the adhered fractions began to both decrease in the first portions of the gut and move to the distal regions of the small intestine and caecum. In any case, no significant amounts of PVM / MA nanoparticles were found in the colon (data not shown). The cross-linkage of conventional nanoparticles with DP also decreased the adhesive capacity of the resulting carriers. Therefore, 30 min post-administration, the amount of adhered nanoparticles to the stomach diminished from 16%, for conventional nanoparticles, to 7–8% when the particles were treated with DP.

25

On the other hand, cross-linked nanoparticles coated with BSA (BD-NP) showed a more homogenous distribution within the entire gastrointestinal tract and their accumulation in the distal regions of the gut was significantly higher and for longer periods of time than the other formulations. Fig. 2 shows microphotographs concerning the interaction between PVM / MA formulations and the intestinal mucosa. In all cases, they can be visualised in the mucus layer as large fluorescent aggregates of particles. In addition, nanoparticles appeared to be distributed homogeneously along the mucosa segment.

3.4. Transit parameters of nanoparticles in the gastrointestinal tract The gastric (k ge ) and intestinal (k ie ) emptying rates of PVM / MA nanoparticles are listed in Table 3. The cross-linkage of NP with 1,3-diaminopropane in-

Fig. 2. Nanoparticles remaining in the small intestine 3 h after administration to animals. (A) Negative control, (B) NP, (C) D30-NP and (D) BD-NP. Bar: 100 mm.

´ et al. / Journal of Controlled Release 89 (2003) 19–30 P. Arbos

26

Table 3 Kinetic parameters (gastric, k ge , and intestinal, k ie , emptying rates) of nanoparticles in the gut Formulation

k ge (h 21 )

k ie (h 21 )

NP D5-NP D10-NP D30-NP BD-NP

0.6260.20 0.7360.17 0.9860.41 0.9360.22 0.2160.07**

0.2460.07 0.2360.03 0.3960.07** 0.4060.03** 0.1160.04**

**P,0.01 versus conventional nanoparticles (NP).

creased both rates, whereas the coating nanoparticles with BSA significantly decreased the gastric and intestinal emptying rates. For Gantrez-sol, unfortunately, these parameters could not be calculated because of the great variability observed when analysing the rinsing liquids containing the lumen content.

3.5. Bioadhesion of nanoparticles in the gastrointestinal tract Fig. 3 shows, for the different formulations tested, the evolution of the cumulative amount of adhered particles (expressed in mg) on the whole gastrointestinal tract over time. The aqueous solution of the copolymer (Gantrez-sol) showed the lowest initial capacity to interact with the mucosa; although, a similar amount of RBITC was recovered in the gut mucosa 3 h post-administration. On the other hand, NP and DP-cross-linked nanoparticles displayed similar adhesive profiles with a maximum of adhesion 30 min post-administration and a rapid decline in the adhered fraction over time. In contrast, BD-NP displayed a lower initial capacity to interact with the gastrointestinal mucosa; although, for this formulation, a plateau of adhesion was observed during the first 3 h post-administration. Table 4 summarises the parameters used to quantify the in vivo bioadhesive characteristics of the different formulations tested. The formulation of the copolymer as nanoparticles allowed an increase in the bioadhesive properties. The maximal amounts adhered to mucosa were about 2.3-times higher for NP than for the copolymer dissolved in water. In the same way, the AUC adh significantly increased when the copolymer was folded as nanoparticles. On the other hand, the cross-linkage of nanoparticles with

Fig. 3. Evolution of the adhered fraction of PVM / MA formulations in the whole gastrointestinal tract with the time, after a single oral administration of 10 mg nanoparticles. (A) NP and polymer solution (Gantrez-sol). (B) D5-NP, D10-NP, D30-NP and BD-NP formulations.

DP decreased both the AUC adh and MRT adh of the resulting carriers. This fact was enhanced by increasing the amount of cross-linking agent used during the hardening process and was confirmed by the dramatic increase in k adh . Similarly, the coating with BSA of the cross-linked nanoparticles also diminished the initial capacity of these carriers to develop adhesive interactions (Q max ) within the gastrointesti-

´ et al. / Journal of Controlled Release 89 (2003) 19–30 P. Arbos

27

Table 4 Parameters of bioadhesion for the different formulations tested

Gantrez-sol NP D5-NP D10-NP D30-NP BD-NP

AUC adh (mg h)

MRT (h)

k adh (h 21 )

Q max (mg)

7.12 10.49 10.93 6.60 5.58 16.58

5.3 5.23 5.19 5.11 4.58 9.70

0.2860.04 0.2960.03 0.3760.02** 0.8060.08** 0.8860.24** 0.2560.05

1.5360.17* 3.6460.34 3.4360.11 2.3160.79* 2.2560.06* 2.8461.05*

*P,0.05 and **P,0.01 versus conventional nanoparticles (NP).

nal mucosa. However, BD-NP formulation displayed the lowest elimination rate and its AUC adh and MRT adh were about 1.6- and 2-times higher than for conventional nanoparticles.

3.6. Transit-bioadhesion correlation Fig. 4 shows, for the five formulations tested, the correlation between the gastrointestinal parameters, expressed as either stomach or intestine emptying rates, and the intensity of the bioadhesive phenomenon (expressed as AUC adh ). In both cases, good correlation was obtained between the kinetic parameters and the bioadhesivity of nanoparticles. For k ge , the correlation coefficient was calculated to be 0.9733, whereas for k ie , r was .0.9867. Furthermore, low gastric or intestinal emptying rates were related to high adhesive capacities in the gastrointestinal mucosa.

Fig. 4. Relationship between the gastrointestinal transit properties (expressed as k ge and k ie ) of the PVM / MA nanoparticles and their bioadhesive potential (expressed as AUC adh ).

4. Discussion RBITC-loaded PVM / MA nanoparticles can be easily prepared by a solvent displacement method, which enabled us to prepare homogenous and reproducible batches. However, the resulting carriers (NP) did not show a high stability in aqueous media and, for instance, their complete dissolution in water took place in less than 24 h. One possible solution to stabilise these carriers may be their cross-linkage with molecules possessing hydroxyl or amine residues. In this work, conventional nanoparticles were cross-linked with 1,3-diaminopropane (D5-NP, D10NP and D30-NP) and, eventually, coated with BSA (BD-NP). Cross-linkage of PVM / MA nanoparticles clearly modified the physico-chemical characteristics of the resulting carriers, especially their surface properties and the RBITC loading. In the first case, crosslinkage of NP with increasing amounts of DP significantly decreased the negative zeta potential of nanoparticles. This is in agreement with other studies describing significant reductions in the zeta potential of nanoparticles cross-linked with glutaraldehyde [20,21]. The lowest negative value found for BD-NP could be attributed to the presence of the BSA coating layer, which was further corroborated by the microBCA assay. In fact, coating nanoparticles with macromolecules may shift the shear plane of the diffusion layer to a greater distance from the nanoparticles, thus resulting in a decrease in the absolute value of zeta potential [22]. In the present study, nanoparticles incorporating RBITC were used to study the gastrointestinal transit and bioadhesive properties in the gut. Concerning the encapsulation of RBITC in the nanoparticles, crosslinkage with concentrations higher than 0.01 mg

28

´ et al. / Journal of Controlled Release 89 (2003) 19–30 P. Arbos

DP/ mg polymer dramatically decreased the marker content; probably because the cross-linker agent would possess a higher affinity for the carboxylic groups (formed by the opening of anhydride groups) of the copolymer than isothiocyanate residues of RBITC. In fact, isothiocyanates can also react with carboxylic groups but need stronger conditions (i.e. acidic pH values) than primary amines or hydroxyl residues [23]. In contrast, the coating of nanoparticles with BSA significantly increased the amount of the marker loaded in the resulting carriers. This is in consistency with Schreiber and Haimovich, who described a stronger and non-labile interaction between RBITC and albumin by incubation in aqueous media [24]. When PVM / MA formulations were administered by the oral route to fasted rats, their distribution within the gut and bioadhesivity to the mucosa appeared to be influenced by the carrier shape (copolymer solubilised in water or folded in nanoparticles), extent of cross-linkage and presence of the BSA coating layer. Thus, the mucoadhesive potential of NP appeared to be much stronger than that observed with Gantrezsol. In fact Q max and AUC adh were found to be 2.3-fold and 1.5-fold higher, respectively, for NP than for the copolymer in the dissolved form. These results appear to be in accordance with previous studies suggesting that the nanoparticulate form would facilitate both the initial contact and the establishment of adhesive interactions between the pharmaceutical dosage form and the components of the mucosa [25,26]. However, regarding k adh and MRT adh , both parameters were of the same order for NP and Gantrez-sol (Table 4). These results enable us to hypothesise that a spherical shape may facilitate the penetration in the mucus layer and the development, in a more rapid way, of stronger adhesive interactions with components of the mucosa than the solubilised form of the copolymer. Once in the mucosa, both formulations (NP and Gantrez-sol) would be eliminated by the mucus turnover. The cross-linkage of nanoparticles with increasing amounts of DP raised the stability of the resulting carriers and minimised the release of the loaded marker (Table 2). Nevertheless, the gastric and intestinal emptying rates increased by increasing the extent of cross-linkage (from NP to D30-NP), while

the capacity to develop adhesive interactions and the intensity of the adhesive phenomenon were higher for conventional nanoparticles than for cross-linked carriers. The high ability of NP to develop adhesive interactions within the gastrointestinal tract may be related to the formation of carboxylic groups from the polyanhydride residues of the copolymer. These carboxylic groups would develop hydrogen bonds with components of the mucosa, such as mucins. This adhesive mechanism has been described for poly(fumaric-co-sebacic acid) microparticles [9] and nanoparticles [27]. Indeed, cross-linkage of PVM / MA nanoparticles with DP in an aqueous medium would block carboxylic groups, conferring a higher stability to the resulting carriers against dissolution; however, at the same time, decreasing their capacity to establish adhesive interactions with the mucosa. These results agree well with those obtained with conventional gliadin nanoparticles, which displayed a higher capacity to interact with the mucosa than particles cross-linked with glutaraldehyde [8]. On the other hand, cross-linkage of nanoparticles was revealed from a significant decrease in the negative zeta potential, directly related to the intensity of the hardening process with DP. Higher amounts of DP yielded less negative nanoparticles, which were less efficient to establish adhesive interactions with the mucosa. In this context, D30-NP showed the lowest adhesive intensity (AUC adh ) and the highest elimination rate of the adhered fraction (k adh ), which may be a probe that hydrophobicity is a major hindrance for penetration in the mucus layer. This was also suggested by Durrer et al., who found that the hydrophilicity of latexes increased their adsorption to rat intestinal mucosa [28]. More recently, it has been described that the hydrophobicity of polystyrene nanoparticles was the main factor influencing the strongest reduction in their association with mucus-secreting MTX-E12 cells [29]. The BSA coating of D10-NP nanoparticles also influenced the transit and the bioadhesive pattern of PVM / MA nanoparticles. The emptying stomach and intestine rates diminished at least twice in comparison with conventional nanoparticles (P,0.01), whereas the intensity of the bioadhesive phenomenon was significantly higher, as revealed by AUC adh and MRT adh , than for NP. Similarly, another interesting

´ et al. / Journal of Controlled Release 89 (2003) 19–30 P. Arbos

point was that the profile of the curve of bioadhesion (Fig. 3) was different from that observed for the other formulations. In this case, Q max was significantly reduced (P,0.05), when compared with NP formulation, although it maintained a constant value for at least 3 h. The high retention in the stomach observed for BD-NP may be related to the use of BSA as coating agent. On the one hand, BSA displays a high affinity with gastric mucin, mainly in acidic media (pH 1.0) [30]. On the other hand, BSA—as with every protein—can be degraded in the stomach and, therefore, induce both the digestion process and the enterogastric reflex, that delays the gastric emptying [31]. In spite of BD-NP showing the less negative zeta potential and a low capacity to establish adhesive interactions with the mucosa (as revealed by Q max ), this formulation displayed the highest AUC adh . This fact may be related to a high residence and a more homogenous distribution along the gut (Fig. 2), and a low gastric emptying rate (Table 3). Both phenomena may facilitate the adsorption to the mucus layer by minimising the saturation of the free mucosa surface available to interact with the carriers. This is consistent with Durrer et al., who found that small particles (up to 670 nm) can be considered as adsorbates which penetrate into a porous adsorbent (the mucus layer) until the internal area available for adsorption is saturated [28,32]. Irrespective of the differences observed for both modalities to modify the surface of PVM / MA nanoparticles (cross-linkage with diamine and, eventually, coating with BSA), the most significant conclusion that could be drawn from these data is that the presence of a BSA coating layer around the particles leads to a greater adsorption of the resulting nanoparticles. These results agree well with previous reports which show significant decreases in the gastric emptying rates of particulates when coating with hydrophilic molecules such as polyglycerol ester of fatty acids [15] or poly(vinylamine) [16]. In addition, a good relationship between the gastrointestinal transit of the different nanoparticle formulations and their adhesivity within the mucosa was found (Fig. 4). In all cases, low gastric or intestinal emptying rates correlated well to high intensities of the bioadhesive interactions developed by PVM / MA nanoparticles.

29

5. Conclusions It has been demonstrated that the hardening process modified the surface characteristics and, therefore, modulated the in vivo behaviour of PVM / MA nanoparticles. In fact, a high amount of DP agent decreased the adhesivity as well as increased the transit through the gastrointestinal tract. However, BD-NP formulation displays the highest bioadhesive potential and the lowest emptying rates. Finally, it was demonstrated that either transit and bioadhesion parameters described the influence of the cross-linkage. Moreover, gastric and intestinal emptying rates were correlated with the intensity of the adhesion.

Acknowledgements ´ was supported by a fellowship grant Pau Arbos ´ de Amigos de la Universidad de from ‘Asociacion Navarra’. This research is supported by grants from ´ in Spain the ‘Ministerio de Ciencia y Tecnologıa’ (AGL2000-0299-C03 and SAF2001-0690-C03-01).

References [1] J. Kreuter, Peroral administration of nanoparticles, Adv. Drug Deliv. Rev. 7 (1991) 71–86. [2] H. Takeuchi, H. Yamamoto, T. Niwa, T. Hino, Y. Kawashima, Enteral absorption of insulin in rats from mucoadhesive chitosan-coated liposomes, Pharm. Res. 13 (1996) 896–901. ´ H. Vranckx, P. Balschmidt, P. Couvreur, Poly[3] C. Damge, (alkyl cyanoacrylate) nanospheres for oral administration of insulin, J. Pharm. Sci. 86 (1997) 1403–1409. [4] S. Yang, J. Zhu, Y. Lu, B. Liang, C. Yang, Body distribution of camptothecin solid lipid nanoparticles after oral administration, Pharm. Res. 16 (1999) 751–757. [5] A. Dembri, M.J. Montisci, J.C. Gantier, H. Chacun, G. Ponchel, Targeting of 39-azido-39deoxythymidine (AZT)loaded poly(isohexylcyanoacrylate) nanospheres to the gastrointestinal mucosa and associated lymphoid tissues, Pharm. Res. 18 (2001) 467–472. ¨ [6] L. Araujo, M. Sheppard, R. Lobenberg, J. Kreuter, Uptake of PMMA nanoparticles from the gastrointestinal tract after oral administration to rats: modification of the body distribution after suspension in surfactant solutions and in oil vehicles, Int. J. Pharm. 176 (1999) 209–224. [7] M.J. Montisci, A. Dembri, G. Giovannuci, H. Chacun, D. ˆ Duchene, G. Ponchel, Gastrointestinal transit and mucoadhesion of colloidal suspensions of Lycopersicon esculentum L.

30

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18] [19]

´ et al. / Journal of Controlled Release 89 (2003) 19–30 P. Arbos and Lotus tetragonobulus lectin-PLA microsphere conjugates in rats, Pharm. Res. 18 (2001) 829–837. M.A. Arangoa, M.A. Campanero, M.J. Renedo, G. Ponchel, J.M. Irache, Gliadin nanoparticles as carriers for the oral administration of lipophilic drugs. Relationships between bioadhesion and pharmacokinetics, Pharm. Res. 18 (2001) 1521–1527. D.E. Chickering, J.S. Jacob, T.A. Desai, M. Harrison, C.N. Morrell, P. Chaturvedi, E. Mathiowitz, Bioadhesive microspheres: III. An in vivo transit and bioavailability study of drug-loaded alginate and poly(fumaric-co-sebacic anhydride) microspheres, J. Controlled Release 48 (1997) 35–46. C.M. Lehr, J.A. Bouwstra, J.J. Tukker, H.E. Junginger, Intestinal transit of bioadhesive microspheres in an in situ loop in the rat—A comparative study with copolymers and blends based on poly(acrylic acid), J. Controlled Release 13 (1990) 51–62. B. Tirosh, A. Rubinstein, Migration of adhesive and nonadhesive particles in the rat intestine under altered mucus secretion conditions, J. Pharm. Sci. 87 (1998) 453–456. J. Kreuter, U. Mulle, K. Munz, Quantitative and microautoradiographic study on mouse intestinal distribution of polycyanoacrylate nanoparticles, Int. J. Pharm. 55 (1989) 39–45. ¨ A. Lamprecht, U. Schafer, C.M. Lehr, Size-dependent bioadhesion of micro- and nanoparticulate carriers to the inflamed colonic mucosa, Pharm. Res. 18 (2001) 788–793. ´ M.A. Arangoa, M.A. Campanero, J.M. Irache, P. Arbos, Quantification of the bioadhesive properties of protein-coated PVM / MA nanoparticles, Int. J. Pharm. 242 (2002) 129–236. Y. Akiyama, N. Nagahara, T. Kashihara, S. Hirai, H. Toguchi, In vitro and in vivo evaluation of mucoadhesive microspheres prepared for the gastrointestinal tract using polyglycerol esters of fatty acids and a poly(acrylic acid) derivative, Pharm. Res. 12 (1995) 397–405. S. Sakuma, R. Sudo, N. Suzuki, H. Kikuchi, M. Akashi, M. Hayashi, Mucoadhesion of polystyrene nanoparticles having surface hydrophilic polymeric chains in the gastrointestinal tract, Int. J. Pharm. 177 (1999) 161–172. P. Arbos, M. Wirth, M.A. Arangoa, F. Gabor, J.M. Irache, Gantrez  AN as a new polymer for the preparation of ligand-nanoparticle conjugates, J. Controlled Release 83 (2002) 321–330. A.J. Kinloch, The science of adhesion: I. Surface and interfacial aspects, J. Mater. Sci. 15 (1980) 2141–2166. D.E. Chickering, E. Mathiowitz, Bioadhesive microspheres I.

[20]

[21]

[22]

[23] [24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

A novel electrobalance-based method to study adhesive interactions between individual microspheres and intestinal mucosa, J. Controlled Release 34 (1995) 251–261. M.A. Arangoa, G. Ponchel, A.M. Orecchioni, M.J. Renedo, ˆ D. Duchene, J.M. Irache, Bioadhesive potential of gliadin nanoparticulate systems, Eur. J. Pharm. Sci. 11 (2000) 333– 341. C. Weber, C. Coester, J. Kreuter, K. Langer, Desolvation process and surface characterisation of protein nanoparticles, Int. J. Pharm. 194 (2000) 91–102. R. Gref, A. Doub, P. Quellec, T. Blunk, R.H. Muller, J.M. Verbavatz, R. Langer, The controlled intravenous delivery of drugs using PEG-coated sterically stabilized nanospheres, Adv. Drug Deliv. Rev. 16 (1995) 215–233. S. Patai, Z. Rappoport, The Chemistry of Sulphur-containing Functional Groups, Wiley, Chichester, UK, 1993. A.B. Schreiber, J. Haimovich, Quantitative fluorimetric assay for detection and characterisation of Fc receptors, Methods Enzymol. 93 (1983) 147–155. ˆ D. Duchene, G. Ponchel, Bioadhesion of solid oral dosage forms, why and how?, Eur. J. Pharm. Biopharm. 44 (1997) 15–23. J. Shimoda, H. Onishi, Y. Machida, Bioadhesive characteristics of chitosan microspheres to the mucosa of rat small intestine, Drug Dev. Ind. Pharm. 27 (2001) 567–576. G.P. Carino, J.S. Jacob, E. Mathiowitz, Nanosphere based oral insulin delivery, J. Controlled Release 65 (2000) 261– 269. ˆ C. Durrer, J.M. Irache, F. Puisieux, D. Duchene, G. Ponchel, Mucoadhesion of latexes II. Adsorption isotherms and desorption studies, Pharm. Res. 11 (1994) 680–683. D.A. Norris, P.J. Sinko, Effect of size, surface charge and hydrophobicity on the translocation of polystyrene microspheres through gastrointestinal mucin, J. Appl. Polym. Sci. 63 (1997) 1481–1492. E.E. Hassan, J.M. Gallo, A simple rheological method for the in vitro assessment of mucin-polymer bioadhesive bond strength, Pharm. Res. 7 (1990) 491–495. J.B. Kirsner, C.S. Winans, in: W.A. Sodeman, T.M. Sodeman (Eds.), Sodeman’s Pathologic Physiology, Saunders, Philadelphia, 1979, p. 798. G. Ponchel, M.J. Montisci, A. Dembri, C. Durrer, D. ˆ Duchene, Mucoadhesion of colloidal particulate systems in the gastrointestinal tract, Eur. J. Pharm. Biopharm. 44 (1997) 25–31.