Introduction

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

The ordinary treatment of inflammatory bowel disease requires the frequent intake of anti-inflammatory drugs at high doses, which causes the absorption of those drugs from the small intestine, leading to significant adverse events. Therefore, several strategies have been followed such as the development of prodrugs that deliver drugs specifically in the large bowel after cleaving the active part from the hydrophilic carrier by specific bacterial enzymes in the colon (McLeod et al., 1994; Fedorak et al., 1995) and the development of solid dosage forms that release the drug in the colon in dependence of the physiological environment (Watts and Illum, 1997; Kinget et al., 1998; Tozaki et al., 1999). The administration of drugs by rectal route is also currently used. However, it is not effective when the inflamed tissues are located in the upper parts of the colon.

Although prodrugs lead to reduced adverse effects, a more comfortable dosage frequency cannot be achieved. Sustained drug release devices, e.g., pellets, capsules, or tablets, delivering the drug specifically in the colon for a longer time period have been developed. However, their efficiencies seem to be decreased in many cases due to the diarrhea, a symptom of inflammatory bowel disease that enhances the elimination and reduces the possible drug release time (Hardy et al., 1988; Watts et al., 1992). Thus, a carrier system that delivers the drug specifically and exclusively to the inflamed regions after oral administration for a prolonged period would be desirable. Such a system could reduce side effects significantly in the case of conventional chemical anti-inflammatory compounds.

As reported from previous work, drug carrier systems with a size larger than 200 µm are subjected to the diarrhea symptoms, resulting in a decreased gastrointestinal transit time and therefore to a distinct decrease in efficiency (Hardy et al., 1988; Watts et al., 1992).

Because nanoparticles can be designed to control drug release after oral administration, the development of such nanoparticles seems to be promising primarily to reduce the dosage frequency. In the case of colitis, a strong cellular immune response is known from the inflamed regions, i.e., in general, an increased presence of neutrophils, natural killer cells, mast cells, and regulatory T cells, which have an important role in the pathophysiology of inflammatory bowel disease (Allison et al., 1988; Seldenrijk et al., 1989; Probert et al., 1996). Moreover, it has been reported that microspheres and nanoparticles can be efficiently taken up by macrophages (Tabata et al., 1996). Thus, it may be expected that particle uptake into those immune-related cells or the disruption of the intestinal barrier function (Stein et al., 1998) could allow the accumulation of the particulate carrier system in the desired area. A subsequent increase of residence time that would be postulated for nanoparticles compared with existing drug delivery systems allows a dose reduction. Indeed, it has been demonstrated that microparticles containing dexamethasone showed previously promising results in colitis-induced mice (Nakase et al., 2000). In addition, the successful oral administration of drugs with strong adverse effects such as rolipram, the model drug in our study, may be a new medical approach.

In several different diseases, the proinflammatory cytokine tumor necrosis factor- forms a necessary element in the chain of pathophysiological events leading to inflammation. Among the agents known to inhibit tumor necrosis factor- production rather than block its function, attention has focused on cAMP-elevating phosphodiesterase inhibitors. Rolipram has initially been developed and studied as an antidepressant drug (Wachtel, 1983). Recently, the potential therapeutic use of rolipram in tumor necrosis factor--dependent disease has been demonstrated in several animal models (Nyman et al., 1997; Ross et al., 1997; Hartmann et al., 2000).

The aim of this project was to test in vivo the targeting potential of nanoparticles to the inflamed tissue. A previous study (Lamprecht et al., 2001a) proved an increased nanoparticle deposition in the inflamed tissue of the colon compared with the healthy control. This accumulation allows a potential nanoparticulate drug delivery system to stay in the inflamed area for a longer time. Here, we evaluated the therapeutic efficiency of this drug carrier system using the experimental colitis rat model. To compare the efficiency of the nanoparticle treatment, control rats received the free anti-inflammatory drug as a solution. The mitigating effect of all formulations was determined by a clinical score system, the colon/body weight ratio and the myeloperoxidase activity.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

The biodegradable polymer poly[DL-lactide-co-glycolide] 50/50 (PLGA) (mol. wt. 5,000 or 20,000) was purchased from Wako (Osaka, Japan). Rolipram was received as a gift from Schering AG (Berlin, Germany). Trinitrobenzenesulfonic acid (TNBS) and o-dianisidine hydrochloride were obtained from Sigma Chemical (Deisenhofen, Germany), and hexadecyltrimethylammonium bromide was obtained from Fluka (Deisenhofen, Germany). All other chemical reagents were purchased from Sigma Chemical (Steinheim, Germany), Merck AG (Darmstadt, Germany), or Nacalai Tesque Inc. (Kyoto, Japan) and were of analytical grade.

Preparation Method of Biodegradable Nanoparticles. The preparation of the nanoparticles was achieved by adjusting the oil-in-water emulsion technique (Lamprecht et al., 2001b). Briefly, 40 mg rolipram was dissolved in 4 ml of methylene chloride containing 250 mg of the polymer poly[DL-lactide-coglycolide] 50/50 (mol. wt. 5,000 or 20,000). This solution was thereafter poured into 8 ml of aqueous polyvinyl alcohol solution (1%) and homogenized with an ultrasonifier (Ultrasonic Disruptor model UR-200P; Tomy Seiko Co. Ltd., Tokyo, Japan) in an ice bath for 3 min. After evaporation of the methylene chloride under reduced pressure, the polymer precipitated and the nanoparticles were separated from nonencapsulated drug and free surfactant by centrifugation (14,000g for 5 min).

Nanoparticles Characterization. The nanoparticles were analyzed for their size distribution and their surface potential using a Photal laser particle analyzer LPA 3100 (Otsuka Electronics, Osaka, Japan) and a Zetasizer II (Malvern Instruments, Worcestershire, UK), respectively. The external morphology of NP was analyzed with a JEOL JSM-T330A scanning microscope (Tokyo, Japan). The amount of rolipram entrapped within the NP was determined by measuring the nonentrapped drug amount in the supernatant by the following high-performance liquid chromatography method: GL-Sciences Inc., Inertsil ODS-2 column (Shimadzu, Tokyo, Japan); eluent, acetonitrile/water 40:60; flow rate, 1.0 ml/min; UV detection at 280 nm. The results of this indirect assay were compared with those obtained by the direct assay in a previous study (Lamprecht et al., 2001b). Because a good correlation was found for both methods, the indirect method was chosen for the determination of drug incorporation as well as the drug release assays because it was faster and easier. For the in vitro release profiles, 100 mg of lyophilized NP were resuspended in 200 ml of phosphate buffer, pH 7.4, and incubated into a bath at 37°C under magnetic stirring at 250 rpm. At appropriate intervals, 0.2-ml samples were withdrawn and filtrated through a Millipore 0.1-µm filter. The filtrate was assayed for drug release and replaced by 0.2 ml of fresh buffer. The amount of rolipram in the release medium was determined by the high-performance liquid chromatography assay.

Animal Treatment. Experiments were carried out in compliance with the regulations of the committees of the Gifu Pharmaceutical University (Gifu, Japan) in line with the Japanese legislation on animal experiments. Male Wistar rats (average weight 230-250 g; 12-15 weeks; n = 6/group) were used. To induce an inflammation, the colitis groups were treated by the following procedure: after light narcotizing with ether, the rats were catheterized 8 cm intrarectal and 500 µl of TNBS in ethanol was applied (dose was 150 mg/kg of body weight of TNBS in ethanol, 50% solution). For 3 days the rats were housed without treatment to maintain the development of a full inflammatory bowel disease model. The animals of each group received orally either 0.5 ml of rolipram solution or NP suspension, once daily for five continuous days (rolipram solution: 10 mg/kg of body weight; suspension of nanoparticles: equivalent dose containing). The colitis control group received only saline instead of free drug or drug-containing particles.

Clinical Activity Score System. Colitis activity was quantified with a clinical score assessing weight loss, stool consistency, and rectal bleeding as previously applied by Hartmann et al. (2000). No weight loss was counted as 0 point, 1 to 5% as 1 point, 5 to 10% as 2 points, 10 to 20% as 3 points and >20% as 4 points. For stool consistency, 0 point was given for well formed pellets, 2 points for pasty and semiformed stools that did not stick to the anus, and 4 points were given for liquid stools that stick to the anus. Bleeding was scored as 0 point for no blood, 2 points for positive finding, and 4 points for gross bleeding. The mean of these scores was forming the clinical score ranging from 0 (healthy) to 4 (maximal activity of colitis).

Myeloperoxidase Activity. The measurement of the myeloperoxidase activity was performed to quantify the severity of the colitis. It is a reliable index of inflammation caused by infiltration of activated neutrophils into the inflamed tissue. Activities were analyzed according to Krawisz et al. (1984). Briefly, distal colon specimen was minced in 1 ml of hexadecyltrimethylammonium bromide buffer (0.5% in 50 mM phosphate buffer) on ice and homogenized. The homogenate was sonicated for 10 s, freeze-thawed three times, and centrifuged at 10,000 rpm for 3 min. Myeloperoxidase activity in the supernatant was measured spectrophotometrically. Supernatant (0.1 ml) was added to 0.167 mg/ml of o-dianisidine hydrochloride and 0.0005% hydrogen peroxide, and the change in absorbance at 460 nm was measured. One unit of myeloperoxidase activity was defined as the amount that degraded 1 µmol of peroxidase per minute at 25°C.

Characterization of Systemic Adverse Effect. The neurotropic effects of rolipram after oral administration in rats were described in detail by Wachtel (1983). Following these procedures, the adverse effect of rolipram was determined after administration of the drug. One hour after the drug or particles were administered, the behavior of rats was observed through transparent cages. The incidence of forepaw shaking and grooming was counted for 60 min by an observer unaware of the animal's treatment. The following criteria were used: rapid repetitive forepaw shaking and grooming.

Statistical Analysis. The results were expressed as mean values ± S.D. The Mann-Whitney-Wilcoxon U test was used to investigate differences statistically because the number of animals in each group was relatively low. However, in all statistical analysis normality and equal variance was passed. Therefore, the Student's t test was also applied to examine significance of differences. In all cases, P < 0.05 was considered to be significant and the marked significant differences are valid for both statistical test systems.

	Results

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

All nanoparticles were prepared with poly[DL-lactide-coglycolide], a biocompatible and biodegradable polymer that is now well established for use in humans. Nanoparticles were characterized in terms of size, polydispersity, surface potential, encapsulation efficiency, and drug release (Table 1). As it can be seen in Fig. 1, nanoparticles prepared by the sonication method had a spherical shape, submicrometer size, and were relatively monodispersed. Figure 2 illustrates the in vitro drug release profiles obtained for the two formulations later used in the in vivo experiments, by representing the percentage of rolipram release with respect to the amount of rolipram encapsulated. Drug release occurred in two phases: a first initial burst release and a sustained release of the drug over 1 week resulting from the diffusion of the drug through the polymer. The repeated washing steps after the preparation reduced the adsorbed drug amount on the nanoparticle surface to a minimum, subsequently, the initial burst release was lower than 30% of drug released within the first 2 h for both formulations (Fig. 2, inset).

View larger version (74K): [in this window] [in a new window]   Fig. 1.   Scanning electron microscopic images of nanoparticles prepared by PLGA (mol. wt. 20,000) (a) or PLGA (mol. wt. 5,000) (b).

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Release profiles of rolipram nanoparticles in phosphate buffer, pH 7.4, at 37°C during 168 h. NP were prepared by oil-in-water emulsion technique using ultrasonication; , PLGA nanoparticles (mol. wt. 20,000); , PLGA nanoparticles (mol. wt. 5,000). Data are shown as mean ± S.D.

To evaluate the therapeutic value of rolipram-containing nanoparticles, the effect of the carrier system was studied on preexisting colitis. On day 3, all animals received an intrarectal application of TNBS except the healthy control group. Before this time point, animals showed no clinical problems. After inducing the experimental colitis the clinical score increased rapidly and consistently for the next 3 days for all groups (Fig. 3). The inflamed tissue showed an extremely increased mucus production in the area of distal colon compared with the histology of healthy gut sections from the control group. Significant damages of the intestinal tissue, e.g., ulceration, have been observed (Fig. 4).

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Clinical activity score during the whole experimental period always determined for n = 6 animals (, healthy control group; , colitis control group; , rolipram solution; , PLGA nanoparticles (mol. wt. 20,000); , PLGA nanoparticles (mol. wt. 5,000). *P < 0.05 compared with colitis control rats given saline.

View larger version (92K): [in this window] [in a new window]   Fig. 4.   Microscopical images of a colon section through a tissue sample of the colitis group showing the strong increased mucus layer on the mucosa (1, mucus; 2 + 3, epithelium; 4, muscularis) (a) and a typical ulceration in the colon (b). Scale bars, 500 µm.

Starting from day 6, rats received orally either rolipram solution or rolipram nanoparticles daily for five consecutive days, only the colitis control group received saline instead. The clinical activity score was used to evaluate the severity of the colonic inflammation and the colitis control group proved to be an excellent model of inflammation as evidenced by the highly increased clinical activity. All drug-receiving groups showed a decrease of inflammation severity after a lag time of 24 to 48 h. The difference between drug-treated groups and colitis controls became significant on day 9. During the whole rolipram treatment period the clinical activity was lowered by free drug and by the two drug carrier formulations as well. After the 5 days without drug treatment, the free drug group showed a strong relapse, whereas for the NP groups continuously reduced clinical activity scores were observed.

Because rolipram undergoes a very strong first-pass effect (Krause and Kühne, 1988) the plasma concentration of the drug could not readily be determined. Therefore, the strong visible neurotropic effect has been used as an indicator as reported from previous work (Wachtel, 1983). Grooming and forepaw shaking indicated the intensity of the neurotropic adverse effects as a proof of systemic drug absorption. This showed an enormous difference between rats receiving rolipram as free drug in solution and those receiving rolipram entrapped in nanoparticles (Fig. 5).

View larger version (39K): [in this window] [in a new window]   Fig. 5.   Neurotropic adverse effects after rolipram administration during administration (day 7 + 8) or relapse phase (day 13 + 14). Colitis control group receiving saline (), rolipram solution-receiving group (), PLGA nanoparticles (mol. wt. 20,000)-receiving group (), and PLGA nanoparticles (mol. wt. 5,000)-receiving group (). *P < 0.05 compared with colitis control rats given rolipram solution.

On day 11 (24 h after the last drug administration), the first series of animals was sacrificed and colon/body weight ratio and myeloperoxidase activity were determined to quantify the inflammation. The drug-treated groups showed a distinct decrease in the colon/body weight ratio compared with the colitis control group (Fig. 6a). The differences between free drug and the nanoparticle formulations were not significant. Furthermore, the myeloperoxidase activity in samples obtained from the inflamed colonic tissue was examined (Fig. 6b). Here also, an enormous difference between rolipram-treated and control groups was found. As observed previously, no significant differences for the mitigating effect of all rolipram-receiving groups were observed.

View larger version (38K): [in this window] [in a new window]   Fig. 6.   Determination of colon/body ratio (a) and myeloperoxidase activity (b) after final drug administration (day 11) and washout phase (day 15). Healthy control group (), colitis control group (), rolipram solution-receiving group (), PLGA nanoparticles (mol. wt. 20,000)-receiving group (), and PLGA nanoparticles (mol. wt. 5,000)-receiving group (). Data are shown as mean ± S.D. *P < 0.05 compared with colitis control rats given saline. #P < 0.05 compared with rolipram solution group on day 11.

The remaining animals were sacrificed on day 15, i.e., 5 days after the last drug administration. Whereas the colitis control group showed a continuously strong colonic inflammation reflected by a high clinical activity score, the rolipram groups responded differently to the lack of drug treatment. The group initially treated with free drug showed a distinct relapse of the inflammation, manifested by clinical activity score as well as an increased colon/body weight ratio and myeloperoxidase activity. This relapse was almost as severe as the inflammation of the nontreated colitis control group regarding clinical activity and colon/body weight ratio. However, the myeloperoxidase activity was still significantly lower than in the colitis control group. In contrast, the nanoparticle groups showed rather no deterioration in colon/body weight ratio and myeloperoxidase activity after 5 days without drug treatment. Only the clinical score showed a slight increase during this period. The neurotropic adverse effect was observed to be slightly higher than in the colitis control group, but no significant difference between free drug and nanoparticle groups was found.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

The behavior of the proposed nanoparticulate drug delivery system was examined with a specific view to therapeutic activity after oral administration. Owing to its simplicity and reproducibility the TNBS colitis model was selected. Furthermore, it is a relevant model because it involves the use of an immunological hapten and develops a chronic inflammation rather than an acute mucosal injury (Yamada et al., 1992). The experimental colitis model in rat after the intrarectal administration of TNBS (Morris et al., 1989) should allow an in vivo characterization of the particulate carrier systems under the influence of chronic inflammation symptoms.

As expected, significant damage of the intestinal tissue, e.g., ulceration, have been observed. In these areas a high amount and infiltration activity of immune-related cells was found (Elson et al., 1995). Thus, an enhanced uptake of administered particles by these cell types could be expected, which resulted in an advantageous accumulation of the carrier system in the inflamed area. Moreover, the inflamed distal colon tissue showed an extremely increased mucus production in the areas surrounding the ulceration compared with the histology of healthy gut sections from the control group. This observation supports the hypothesis of particle accumulation in the inflamed area because small polymeric particles were found to be attached to the healthy intestinal mucus (McClean et al., 1998).

A size-dependent particle deposition in the gastrointestinal tract of healthy rats was reported in previous studies (Jani et al., 1989, 1990; Desai et al., 1996). These studies reported an increase of recovery for smaller particles in the whole gut: we observed the same phenomenon, but the size dependence was less pronounced. On the contrary, there was a strong influence of the particle size in the experimental colitis model (Lamprecht et al., 2001a). In our study the highest deposition amount was found for nanoparticles (particle diameter 100 nm) inside the inflamed tissue of the colon (14.5 ± 6.3% of the total administered particle mass). Two major reasons can be stated for the more distinct size-dependent deposition in the case of colitis. First, smaller particles are taken up more easily by macrophages in the area of active inflammation. Second, the strong increased mucus production leading to a thicker mucus layer in the inflamed areas allows a higher amount of particle adherence. Smaller particles can be better bound to the mucus layer due to an easier penetration into the layer with respect to their relatively small size. In addition, both phenomena may occur at the same time.

The choice of an optimal particle size for the design of a particulate carrier system has to be discussed, based on two major influencing factors. It has to be kept in mind that by increasing the particle size, a higher drug-loading capacity can be reached, which allows the transport of higher drug amounts with less polymer. On the contrary, we observed a tendency of a higher deposition rate and a better targeting index for smaller particles. Combining these two effects, we prepared nanoparticles with a diameter between 200 and 500 nm allowing the production of optimal nanoparticles including efficient drug loading and still having a size range that promises a high targeting efficiency.

The use of biodegradable polymers for nanoparticle preparation was preferable for this application to prevent complications with long-term deposition of nanoparticles or any residual component inside the ulcerated tissue. Moreover, polyvinyl alcohol-coated nanoparticles were found to be very efficient in delaying nanoparticles degradation in the gastrointestinal tract (Landry et al., 1998).

Rolipram has proven to have an anti-inflammatory potential in our experimental colitis model as reported previously by Hartmann et al. (2000).

The clinical activity score, colon/body weight ratio, and myeloperoxidase activity decreased significantly after the oral administration of rolipram nanoparticles or free drug. Nanoparticles formulations proved to be as efficient as the free drug in mitigating the experimental colitis. The local anti-inflammatory effect was achieved by the controlled drug release during the nanoparticle deposition period in the inflamed colon areas.

After the final rolipram administration, the group receiving free drug showed a strong relapse, whereas this was not observed in animals fed with rolipram nanoparticles. This suppressed relapse might be due to an accumulated deposition of nanoparticles observed in the first part of this study, which would retain the drug carrier in the inflamed regions of the colon for up to several days.

On the other hand, the reduction of neurotropic effects that was observed after the administration of the rolipram nanoparticles is based on the reduced availability of rolipram during the gastrointestinal passage by the surrounding polymeric carrier. An additional advantage seems to be that only a low uptake of nanoparticles into Peyer's patches and their translocation has been reported (Florence et al., 1995), which can prevent an uncontrollable spreading of the nanoparticles during their transportation through the gut.

The remaining adverse effect that was observed for both nanoparticle preparations might be due to the initial burst effect as observed in vitro (30% released in 2 h; Fig. 2). This burst release leads to some drug release in the stomach and small intestine, which is in favor of drug absorption from the upper regions of the gut and subsequent adverse effects of rolipram.

Charge interactions are reported to further enhance binding of macromolecules to the inflamed tissue because it has been shown that ulcerated tissues contain high concentrations of positively charged proteins that increased the affinity to negatively charged substances (Nagashima, 1981). Thus, this also could explain an enhanced attachment of nanoparticles to the inflamed mucus areas. However, a decreased surface potential as the here-tested formulation NP-5005 did not lead to significant pharmacological differences based on this nanoparticle property.

	Conclusions

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

Polymeric particulate carrier systems are expected to target the inflamed tissue in inflammatory bowel diseases. This new delivery system allows the desired drug to accumulate in the inflamed tissue with high efficiency. Compared with approaches from previous works, this includes two major advantages. The drug is concentrated at its site of action, which reduces possible adverse effects and enhances the effect of the administered dose. Moreover, the sustained drug release allows pharmacological effects to be extended due to the prolonged presence time of the carrier system at the targeted inflamed area. This deposition of NP in the inflamed tissue should be given particular consideration in the design of new carrier systems for the treatment of inflammatory bowel disease.