Cationic core-shell nanoparticles for intravesical chemotherapy in tumor-induced rat model: Safety and efficacy

International Journal of Pharmaceutics 471 (2014) 1–9

Cationic core-shell nanoparticles for intravesical chemotherapy intumor-induced rat model: Safety and efficacy

Nazlı Erdogar a, Alper B. _Iskit b, Hakan Eroglu a, Mustafa F. Sargon c, N. Aydın Mungan d,Erem Bilensoy a,*aHacettepe University Faculty of Pharmacy, Department of Pharmaceutical Technology, Sıhhiye-Ankara 06100, TurkeybHacettepe University Faculty of Medicine, Department of Pharmacology, Sıhhiye-Ankara 06100, TurkeycHacettepe University Faculty of Medicine, Department of Anatomy, Sıhhiye-Ankara 06100, TurkeydBülent Ecevit University, Faculty of Medicine, Department of Urology, Kozlu-Zonguldak 67600, Turkey

A R T I C L E I N F O

Article history:Received 17 February 2014Received in revised form 2 May 2014Accepted 7 May 2014Available online 13 May 2014

Keywords:Cationic nanoparticlesCore-shellChitosanPoly-e-caprolactoneMitomycin CBladder cancer

A B S T R A C T

Mitomycin C (MMC) has shown potent efficacy against a wide spectrum of cancers and is clinical firstchoice in superficial bladder tumors. However, intravesical chemotherapy with MMC has been ineffectivedue to periodical discharge of the bladder and instability of this drug in acidic pH, both resulting in highrate of tumor recurrence and insufficiency to prevent progression. Nanocarriers may be a promisingalternative for prolonged, effective and safe intravesical drug delivery due to their favorable size, surfaceproperties and optimum interaction with mucosal layer of the bladder wall. Hence, the aim of this studywas to evaluate and optimize cationic core-shell nanoparticles formulations (based on chitosan (CS) andpoly-e-caprolactone (PCL)) in terms of antitumor efficacy after intravesical administration in bladdertumor induced rat model. Antitumor efficacy was determined through the parameters of survival rateand nanoparticle penetration into the bladder tissue. Safety of the formulations were evaluated byhistopathological evaluation of bladder tissue as well as observation of animals treated with MMC boundto nanoparticles. Results indicated that chitosan coated poly-e-caprolactone (CS-PCL) nanoparticlespresented the longest survival rate among all treatment groups as evaluated by Kaplan–Meier plotting.Histopathological evaluation revealed that cationic nanoparticles were localized and accumulated in thebladder tissue. As intravesical chemotherapy is a local therapy, no MMC was quantified in blood afterintravesical instillation indicating no systemic uptake for the drug which could have subsequently led toside effects. In conclusion, core-shell type cationic nanoparticles may be effective tools for theintravesical chemotherapy of recurrent bladder tumors.

ã 2014 Elsevier B.V. All rights reserved.

Contents lists available at ScienceDirect

International Journal of Pharmaceutics

journal homepage: www.elsev ier .com/locate / i jpharm

1. Introduction

Bladder cancer is one of the most common cancers worldwide,accounting for 3.2% of all cancers (Ferlay et al., 2001). 70% ofbladder cancers are non-muscle invasive and are managed by anendoscopic resection procedure, however 50–70% of non-muscleinvasive tumors come back within five years after transurethralresection (TUR) and 20–30% of recurrent disease progresses tohigher stage or grade resulting in metastasis of the tumor (Miyakeet al., 2004; Lehmann et al., 2006). The most common therapeuticapproach in bladder cancer is intravesical therapy in which drugs

* Corresponding author. Tel.: +90 312 305 12 41; fax: +90 312 305 43 69.E-mail addresses: [email protected] (N. Erdogar),

[email protected] (A.B. _Iskit), [email protected] (M.F. Sargon),[email protected] (N. A. Mungan), [email protected] (E. Bilensoy).

http://dx.doi.org/10.1016/j.ijpharm.2014.05.0140378-5173/ã 2014 Elsevier B.V. All rights reserved.

are directly administered into the bladder to reduce or preventtumor recurrence progression through local chemotherapy(Malmstrom, 2003). The most commonly employed intravesicalagents in patients with superficial bladder cancer are mitomycin C(MMC), thiotepa, ethoglucid, anthracyclines such as doxorubicinand epirubicin, Bacille Calmette Guerin, taxol and the newmitomycin derivative KW-2149 (Heijden van der and Witjes,2003). Intravesical drug delivery may be ineffective due to severalfactors mostly arising from bladder physiology. The most impor-tant drawback is the periodical need of discharge of the bladderapproximately every 2 h. Therefore, even the drug is administeredlocally to the therapeutic site, it is reported to be rapidly dilutedand lost resulting in repeated catheterization of the patient andineffective chemotherapy (Tyagi et al., 2004; Parekh et al., 2006;Kaufman, 2006). Another disadvantage is the very low permeabil-ity of urothelium which is called the bladder permeability barrier(BPB) (Tyagi et al., 2006). Urothelium, is composed of

2 N. Erdogar et al. / International Journal of Pharmaceutics 471 (2014) 1–9

glycosaminoglycan (GAG) mucin layer and this prevents theadhesion of particles to bladder mucosa and limits the mucosalabsorption of molecules into deeper tissues (Tyagi et al., 2004;Parekh et al., 2006). Several approaches to improve intravesicaldrug delivery has been developed aiming to enhance thepermeability of drugs through the bladder wall. Physicalapproaches include electromotive drug administration and ionto-phoresis/electroporation for drug delivery; chemical approachesinclude administration in solvents such as DMSO. Finallytechnological approaches mainly focus on increasing the residencetime of dosage forms in the bladder via bioadhesive colloidalcarriers and nanotechnology for improved permeability of drugsand prolonged residence of drug delivery systems in the bladder(Giannantoni et al., 2006). In our previous studies cationicnanoparticles were developed by either coating PCL nanoparticleswith CS or PLL or directly from CS nanoparticles for the prolongedresidence of MMC in rat bladder (Erdogar et al., 2012) andimproved cellular uptake by MB49 mouse bladder carcinoma cellline (Bilensoy et al., 2009). Cationic nanoparticles for effectiveMMC delivery were optimized for higher drug loading capacity andsmaller particle size by means of different preparation techniquesin our previous studies (Bilensoy et al., 2009; Erdogar et al., 2012)and CS, CS-PCL and PLL-PCL nanoparticles were evaluated forcellular uptake and cytotoxicity properties. In the light of thesestudies, PLL-PCL nanoparticles were not included in animal studiesas they were found to be highly toxic in cell culture studies.Formulations with highest drug loading values and smallest sizewere used in the in vivo studies.

Mitomycin C (MMC) is an antitumor antibiotic exertingtherapeutic activity against many human neoplasms (Carteret al., 1979) and clinical choice in intravesical chemotherapy ofsuperficial bladder tumors. The MMC is known to rapidly degradein acidic environment (Stolk et al., 1986) and cause allergicreactions that are dose-dependent upon systemic uptake and alsochemical cystitis (Thrasher and Crawford, 1992).

Chitosan nanoparticles have shown mucoadhesive propertiesfrom the semi-synthetic cationic polymer chitosan that has a well-known bioadhesive nature, by the establishment of electrostaticinteractions with sialic groups of mucins in the mucus layer. As aresult of this, chitosan promote a structural reorganization of thetight junction-associated proteins and enhance the absorption ofhydrophilic drugs (Bravo-Osuna et al., 2007).

To overcome these problems and to prolong the residence of thechemotherapeutic drug in the bladder, cationic core shell nano-particles of PCL coated with cationic polymers CS was evaluated incomparison to MMC commercial product in solution form inbladder tumor induced rat model as a follow-up study of ourprevious work (Erdogar et al., 2012) comprising in vitro, ex vivoand in vivo evaluation of intravesical chemotherapy with cationicnanoparticles.

2. Materials and methods

2.1. Materials

The following materials were purchased from various compa-nies and then used as received. Chitosan (Protasan UPLC 113viscosity <20 mPa/s Mw< 200 (kDa)2) was obtained from FMCBiopolymers, Norway. Active ingredient (mitomycin C) was a kindgift from the pharmaceutical company Onko-Koçsel, Turkey.Pluronic1 F-68, polyvinylalcohol (PVA), sodium tripolyphosphate(Na TPP) were obtained from Sigma–Aldrich. Poly-e-caprolactone(PCL) (Mn: 42.500) was purchased from Aldrich (St. Louis, MO,United States). Methylene chloride was obtained from J. T. Baker,United States. Ethyl acetate, chloroform, methanol and 2-propanol,N-butyl-N-(4-hydroxybutyl) nitrosamine (BBN) were purchased

from Sigma–Aldrich (St. Louis, MO, United States). All otherchemicals were of reagent grade and solvents were of HPLC grade.

2.2. Preparation and characterization of nanoparticles

2.2.1. Preparation of nanoparticles

2.2.1.1. Core-shell nanoparticles. The MMC-CS-PCL nanoparticleswere prepared by the W/O/W double emulsion technique. Theadjustment of the technique was based on the use of ahomogenizer thus reducing considerably the size of droplets.Briefly, 2 mL 1% PF68 (w/v) solution containing 5 mg MMC (10% ofPCL weight) was emulsified in 10 mL methylene chloridecontaining 0.5% PCL by ultraturrax (IKA T25 basic, Germany) at16,000 rpm for 5 min over an ice bath to form the innerphase. Thisprimary emulsion was immediately injected into 40 mL of aqueoussolution containing 0.1% (w/v) PVA and 0.05% CS in 1% PF68 (w/v)solution for MMC-CS-PCL nanoparticles, using a glass syringe witha needle under agitation. Then the double emulsion was stirred byUltraturrax high speed homogenizer at 16,000 rpm for 5 min overan ice bath. Finally, nanoparticles were obtained in final form afterremoval of organic solvent under vacuum (IKA-Werke-RV06 ML,Germany) at 37 �C to the desired volume (42 mL) (Hasan et al.,2007; Lamprecht et al., 1999).

2.2.1.2. Core nanoparticles. The MMC-CS nanoparticles wereprepared by ionotropic gelation technique based on interactionof oppositely charged groups of chitosan and sodiumtripolyphosphate to form nanoparticles spontaneously. TheMMC was dissolved in TPP solution (20% of polymer weight)and then the TPP solution (0.4 mg/mL) was added to the CSaqueous solution (1.75 mg/mL) and stirred at room temperature.Spontaneously formed nanoparticles were further separated bycentrifugation at 13,500 rpm for 1 h and discarding of thesupernatant and redispersion of the precipitate in water.

2.2.2. Characterization of nanoparticlesParticle size and size distribution, zeta potential, surface

morphology, encapsulation efficiency and in vitro release studieswere preformed within the scope of in vitro characterizationstudies as follows:

Mean diameter (nm) and polydispersity index values of blankand drug loaded nanoparticles were determined by quasi-elasticlight scattering technique (QELS) using Malvern NanoZS (ZetasizerNanoSeries ZS, Malvern Instruments, UK). Analyses were per-formed in triplicate at 25 �C at a 90� angle.

Zeta potential of nanoparticle dispersions were determined toconfirm the surface charge of the particles using Malvern NanoZS(Zetasizer NanoSeries ZS, Malvern Instruments, UK) in triplicate at120� angle and 25 �C.

For the determination of drug loading, nanoparticles wereseparated from the aqueous suspension by ultracentrifugation at13,500 rpm for 1 h. The amount of free MMC in the supernatantwas directly analyzed by an analytically validated HPLC technique(r2 = 0.9995) The HPLC method for the quantification of MMCconsisted of an HP Agilent 1100 HPLC system with a reverse phaseC18 column (150 mm � 4.6 mm, Nucleosil 5C18), a mobile phase ofacetonitrile:water (15:85 v/v), injection volume – 50 mL and flowrate – 1.5 mL/dk. The DAD detector was set at 365 nm. Thepercentage drug loading was calculated according to the followingequation:

%Drugloading ¼ InitialMMC � FreeMMCInitialMMC

� 100

In our previous study, pH 6 and pH 7.8 buffers were selected asrelease media since urine pH is 6 and the MMC is reported to be

N. Erdogar et al. / International Journal of Pharmaceutics 471 (2014) 1–9 3

relatively stable in solutions buffered to a pH of approximately 7.8(Erdogar et al., 2012; Bilensoy et al., 2009). In this studyapproximately 0.08 g lyophilized CS or CS-PCL nanoparticles weretransferred to the dialysis tubing (M.W. cut off 14,000 Da, 25 mm(1.0 in.) flat width, average diameter 16 mm) previously moistenedwith water. Both ends of the bag were clamped and immersed in abeaker containing 40 mL of a pH 6.0 phosphate buffer solution. Thebeaker was placed in a water bath/shaker maintained at 37 � 0.1 �Cat about 200 rpm. One milliliter of the samples were withdrawn at5, 10, 15, 30, 60, 90, 120 and 180 min and subsequently replacedwith the release media after each sampling. Samples were assayedfor MMC by HPLC. The recovery of MMC from the dialysis bag wasdetermined by comparing the amount of MMC added into thedialysis bag at the start of the experiment with the sum of theamount of MMC from the withdrawn samples, the contents of thedialysis bag and the dialysis media at the end of 180 min. The sametechnique was used for the release experiments in pH 7.8 citratebuffer.

2.3. Antitumor efficacy studies

2.3.1. Tumor induction in the bladderA total of 42 male Sprague Dawley rats (Hacettepe University,

Faculty of Pharmacy, Laboratory Animals Research and BreedingUnit, Turkey) with a median weight of 300–350 g were selected asthe experimental model because of their availability and ease ofcatheterization. The animals were maintained in cages at 22–24 �Cwith a 12-hour/12-hour dark/light cycle and humidity of 55% withfree access to water and food. Rats were fed ad libitum. Allexperiments were performed in accordance with the Turkish Lawfor the Protection of Animals and were approved by the localethical committee of Hacettepe University (Ethical committeeapproval number 2009/6-1). Cages were changed once a week. Theanimals were observed daily and the clinical signs were noted.

Each group included six rats and were treated with thefollowing formulations: MMC solution, the MMC-CS-PCL nano-particles and the MMC-CS nanoparticles, the CS-PCL nanoparticlesand the CS nanoparticles. Bladder tumors were induced by adding0.05% BBN to freely available drinking tap water in dark bottlesthree times a week for 8 weeks. Then after 1 week withouttreatment, rats were given weekly intravesical instillations ofdrugs. The rats were anesthetized for approximately 2–3 h with asingle, weight adjusted intraperitoneal dose of chloralhydrate(400 mg/kg body weight). The MMC commercial product solution,the MMC-CS, the MMC-CS-PCL, the CS and the CS-PCL nanoparticleformulations were administered with a volume of 500 mL intra-vesical instillation to rats using 17 gauge cannula (Dıspocann &Cathula, Turkey) separately during subsequent four weeks. Dosesare given as 33 mg/mL for the MMC-CS nanoparticles and 42 mg/mLfor the MMC-CS-PCL nanoparticles (Ethical committee approvalnumber 2009/6-1).

Anti-tumor efficacy was determined over the followingparameters of survival rate, histopathology of tumor inducedbladders after treatment, bladder weight and the MMC presence inplasma.

2.3.2. Survival rateFollowing administration of the MMC commercial drug solution,

the MMC-CS-PCL nanoparticles, the MMC-CS nanoparticles, the CS-PCL nanoparticles and the CS nanoparticles, everyday all animals (alltreated group, positive control (untreated, tumor-induced group)and negative control (untreated, not tumor-induced group)) werechecked regularly, the time of death and the remaining number ofanimals were recorded. The graphic was drawn with the number ofanimals against time and the nanoparticle formulations wereevaluated in terms of survival time of the rats. The mortality data

were subjected to Kaplan–Meier survival analysis to prepare survivalplots. Statistical analysis were performed by Mantel-Cox log-ranktest (p < 0.05). In addition to this, all the animals were observed interms of parameters of general situation as hematuria, bleeding inextremities and weight loss.

2.3.3. Evaluation of bladders after treatmentFollowing administration of the MMC commercial drug

solution, the MMC-CS-PCL nanoparticles, the MMC-CS nano-particles, the CS-PCL nanoparticles and the CS nanoparticles, ratswere kept in a freezer at �20 �C. After the last administration ofdrugs, the bladders were removed, emptied and bladder weightswere determined for positive control, negative control and alltreatment groups. The difference (%) in bladder weight fromhealthy animals were plotted.

Following drug administration, the rats were anesthetized forapproximately 2–3 h with a single, weight adjusted intraperitonealdose of chloralhydrate (400 mg/kg body weight) and bladders wereremoved. Tissue samples taken from bladders were put into testtubes. For transmission electron microscopic examination; thetissue samples were fixed in 2.5% glutaraldehyde for 24 h, washedin phosphate buffer (pH 7.4), post-fixed in 1% osmium tetroxide inphosphate buffer (pH 7.4) for 2 h and dehydrated in increasingconcentrations of alcohol. Then, the tissues were washed withpropylene oxide and embedded in epoxy-resin embedding media.Semi-thin sections about 2 mm in thickness and ultra thin sectionsabout 60 nm in thickness were cut with a glass knife on a LKB-Nova(LKB-Produkter AB, Bromma, Sweden) ultrotome. The semi-thinsections were stained with methylene blue and examined by aNikon Optiphot (Nikon Corporation, Tokyo, Japan) light micro-scope. Following this examination, the tissue blocks weretrimmed, their ultra thin sections were taken by the sameultrotome and they were stained with uranyl acetate and leadcitrate. Following staining, all the ultra thin sections wereexamined by Jeol JEM 1200 EX (Jeol Ltd., Tokyo, Japan) transmis-sion electron microscope. The electron micrographs were taken bythe same transmission electron microscope.

The samples were then evaluated for the structural propertiessuch as destruction degree, apoptotic cell existence, nanoparticleaccumulation, amount of vacuole, edema and mitochondriadamage.

2.3.4. MMC quantification in plasmaFollowing drug administration, the rats were anesthetized for

approximately 2–3 h with a single, weight adjusted intraperitonealdose of chloralhydrate (400 mg/kg body weight). The bloodsamples were collected intracardially, the plasma was separatedwith centrifugation at 3000 rpm for 30 min at 0 �C and theextraction procedure was examined. 1 mL ultrapure water and7 mL extraction solvent (ethyl acetate:2-propanole:chloroform)(70:15:15) were added into 2 mL plasma and vortexed at 2000 rpmfor 10 min. After centrifugation at 3000 rpm for 20 min at 0 �C, theupper organic phase was transferred into test tube. The organiclayer was evaporated to dryness under a nitrogen stream in waterbath at 31 �C. The residue was dissolved in 500 mL methanol. Afterfiltration with 0.22 mm filter, aliquots were analyzed by ananalytically validated HPLC technique for quantification of MMC(r2 = 0.9995). The HPLC method for MMC is described inSection 2.2.2.

3. Results

3.1. Characterization of nanoparticle formulations

The mean diameter of nanoparticles were found to be between160 and 320 nm as seen in Table 1. Coating with polymers resulted in

Table 1Particle size distribution, polydispersity index and zeta potential values of coatedand MMC-loaded nanoparticle formulations (n = 3) � SD.

Nanoparticle formulations Particle size(nm) (�SD)

Polydispersityindex

Zeta potential(mV) (�SD)

CS-PCL 319 � 5 0.184 +10.5 � 1CS 166 � 6 0.253 +35.2 � 1

Table 2Drug loading and encapsulation of cationic MMC loaded nanoparticle formulations(n = 3) � SD.

Nanoparticleformulations

Encapsulation efficiency (%) Entrapped drugquantity (mg/mL)

CS-PCL 35 77 � 0.33CS 19 145 � 0.5

4 N. Erdogar et al. / International Journal of Pharmaceutics 471 (2014) 1–9

significant increase in particle size but all of the formulations were atthe nano range suggesting accumulation in cancer tissues and cells.Polydispersity index values of all nanoparticle formulations werelower than 0.3, indicating distribution of particle sizes being narrowand nanoparticle formulations monodisperse in size distribution.

The zeta potential values of chitosan nanoparticles and coatedPCL nanoparticles are found to be between 10 and 35 mV indicatingnet positive surface charges that were aimed for these drug loadedcationic nanoparticle formulations as seen in Table 1.

The SEM and TEM photomicrographs prove the morphologicalproperties of blank cationic nanoparticle formulations. Fig. 1(a) and(b) represents the SEM images and Fig. 1(c) represents the TEMimages of cationic nanoparticles. These images indicate the smoothregular and spherical surfaces of all nanoparticle formulations.

The double emulsion technique is a suitable technique forencapsulation of hydrophilic molecules, leading to an increasedefficacy. The highest encapsulation efficiency was observed withCS-PCL nanoparticles as seen in Table 2 with 35% loading for MMC.

In the release studies, it was observed that all of thenanoparticle formulations have burst effect followed by rapidand complete release due to the high water solubility of MMC and agood interaction of cationic polymers with the release medium.

3.2. Antitumor efficacy studies

3.2.1. Survival rateIn order to evaluate the effect of nanoparticle-bound MMC on

survival rate of the rats in comparison to the MMC in injectablesolution form, 0.05% BBN was added in drinking water to inducebladder tumor (Shimizu et al., 2001; Le Visage et al., 2004; Chenet al., 1998; Ito et al., 1983). Following that, all nanoparticleformulations and MMC solution were given intravesically once aweek during four weeks. Encapsulated dose of the MMC-CSnanoparticles, the MMC-CS-PCL nanoparticles and the MMCsolution are given as 33 mg/mL, 42 mg/mL and 33 mg/mL,respectively. During the experiment, the animals were controlledregularly, health status (diet situation, hematuria, extremities)were recorded. Finally, the blank and MMC loaded nanoparticleformulations and commercial product treated groups wereevaluated with percentage survival rate percentage against thenumber of days with the Kaplan–Meier plots. (p < 0.05)

3.2.2. Histopathological evaluation of bladder specimenAs described in Section 2.3.3, following the last administration

of drugs, the bladders were removed, emptied and weighed. Asseen in Fig. 5, the increase in weight compared to healthy groupmean bladder weights was observed at different percentages intumor induced rats administered with the MMC commercialproduct, the MMC-CS-PCL nanoparticles, the MMC-CS nano-particles, the CS-PCL nanoparticles and the CS nanoparticles.

Fig. 1. SEM images of (a) CS, (b) CS-PCL

Increase in the bladder weight can be attributed to tumorformation and/or inflammation and edema. All treatment groupsshowed approximately 50% increase in weight of bladder as seen inFig. 5. The tumor tissues are fed by formation of new blood vesselsand as a result of this, increase in tumor mass occurs (Zetter, 1998).

As described in Section 2.3.3, the MMC commercial product insolution form, blank and MMC loaded nanoparticle formulationswere administered in the tumor induced rats intravesically once aweek during four weeks. At the end of the experiment, the animalswere sacrificed with a high dose of anesthesia and the bladderswere removed. The histopathological evaluation was performed asdescribed in Section 2.3.3.

3.2.3. MMC systemic uptake evaluationAs described in Section 2.2.3, the blood samples were collected

from rats which received intravesical administration of MMC-CS,MMC-CS-PCL nanoparticle formulations. After performing neces-sary procedure of extraction, the HPLC analysis was carried out andMMC peak was not observed in any of the samples. The MMC inboth solution form and nanoparticle formulation did not pass tothe systemic circulation and therefore no systemic side effectsrelated to the MMC were observed.

4. Discussion

In our previous study, cationic nanoparticles of core or core-shell type were prepared using different preparation techniquesas double emulsion and ionotropic gelation techniques (Erdogaret al., 2012). In vitro characterization data indicated that CSnanoparticles of 160 nm size and +35 mV surface charge wereobtained using the ionotropic gelation technique when core-shell-type approach was tried with PCL being the core polymer

(c) TEM images of CS nanoparticles.

N. Erdogar et al. / International Journal of Pharmaceutics 471 (2014) 1–9 5

and the CS being the shell or coating polymer, the nanoparticleswere somewhat larger around 300 nm since the size wasgoverned by emulsion droplet size used along with thicknessof the layer of coating polymer surrounding the core. The mainfactor affecting particle size was believed to be the preparationmethod. Surface charge was around +10 mV for both core-shellnanoparticles however, the MMC loading was significantlyincreased by three-fold using the CS-PCL nanoparticles. This isbelieved to be a result of using the double emulsion techniqueinstead of nanoprecipitation providing a larger aqueous phasevolume suitable for the encapsulation of the hydrophilic drugMMC. The CS nanoparticles could be prepared uniquely by theionotropic gelation technique thus the loading values of CSnanoparticles remained unchanged.

If we look from the viewpoint of in vivo behavior, the smallparticle size of the drug-loaded nanoparticles is believed to alsohelp in the availability of higher concentration of free drug atmucosal wall via increased nanoparticle permeability betweenlong chains of the glycosaminoglycan layer lining the bladdermucosae and decreased steric hindrance towards the nano-particles by these long chains.

In view of the mucoadhesive properties, the nanoparticles arecapable of forming non-covalent bonds such as hydrogen bonds orionic interactions of polymer chains and mucus(Barthelmes et al., 2011). Among mucoadhesive polymers, chitosanhas been extensively exploited due to its capacity to interact withthe negatively charged mucosal surface and to enhance drugabsorption by opening of the tight junctions between mucosalcells. Some studies have demonstrated the promising use ofnanoparticles coated with this polysaccharide for drug delivery. Inthis way, the chitosan-coated nanoparticles have proven to besuitable to incorporate drugs and to be stable, under physiologicalconditions, to increase significantly the penetration of theencapsulated drug. They also present favorable drug loading andrelease profiles as well as a good selectivity to the bladder cancercells (Andrews et al., 2009; Bilensoy et al., 2009; Cui et al., 2006).

Drug-loading capacity is a parameter used to assess thesuitability of a particular drug-carrier system (Ruckmani et al.,2006). The double emulsion technique is a suitable technique forencapsulation of hydrophilic molecules, leading to an increasedefficacy. The highest encapsulation efficiency was observed withthe CS-PCL nanoparticles as seen in Table 2 with 35% loading for theMMC. This could be attributed to the fact that by changing thepreparation method from nanoprecipitation to double emulsiontechnique, the hydrophilic drug MMC was solubilized in the inner

Fig. 2. In vitro release profiles of CS-MMC and CS-PCL-MMC nanoparticle formulations a

aqueous phase and was well dispersed within the nanoparticlesand stabilized by double emulsion technique in a W/O/W emulsionwhich resulted in a higher drug loading.

The MMC is a highly hydrophilic drug so it has a high affinity tothe aqueous phase and the release is very rapid as can beexpected. It was observed that; all of the nanoparticle formula-tions have burst effect followed by rapid and complete release.Nanoparticle formulations are stable at pH 6 and pH 7.8 as seen inFig. 2, while the MMC is unstable at pH 6 as can be indicated in theliterature. In vitro release studies were carried out under twodifferent pH conditions; pH 6.0 mimicking the urine pH and pH7.8 in which the MMC was reported to be the most stable(Quebbeman et al., 1985).

Due to large nanoparticle surface area resulting from the smallparticle size, the MMC was released rapidly and completely underboth pH conditions. The release was completed in an hour. The veryhigh water solubility of the drug facilitated the liberation of MMCfrom cationic nanoparticles. Another factor affecting release ratewas polymer characteristics and low affinity of the drug to thehydrophobic polymer PCL and a favorable interaction of thecationic polymer CS with surrounding release medium.

In order to evaluate the effect of nanoparticle-bound MMC onsurvival rate of rats in comparison to the MMC in injectablesolution form. 0.05% BBN was added in drinking water to inducebladder tumor. After that, all the groups were evaluated withpercentage survival rate percentage against the number of dayswith the Kaplan–Meier plots. The rat model has significantadvantages as large bladder and better developed muscular layerallows better histological assessment of depth of invasion anddecreases the risk of perforation during bladder catheterizationover the murine model (Arentsen et al., 2009). Chemicals areimportant to provide reproducible models necessary for detailedstudies of pathogenesis of bladder cancer (Cohen and Ellwein,1990). Three chemicals have been used in causing bladder tumors:FANFT (N-[4-(5-nitro-2-furyl)-2-thiazolyl] formamide), OH-BBN(N-butyl-N-(4-hydroxybutyl)-nitrosamine) and MNU (N-methyl-N-nitrosurea) (Oliveira et al., 2006). The N-butyl-N-(4-hydrox-ybutyl) nitrosamine (BBN)-induced tumors are very similar tohuman bladder transitional cell carcinomas and are considered tobe an excellent model of the clinical bladder cancer (Ohyama et al.,1997). This animal model is a recognized method of investigatingthe invasive potential of bladder cancer because invasive tumorscan develop within days of the induction (Sagara et al., 2010). Inadditon to this, a high dose of BBN induced a high incidence of PNhyperplasia in a short period. The PN hyperplasia is a preneoplastic

t the pH 6.0 phosphate buffer solution and pH 7.8 citrate buffer solution (n = 6) � SD.

Fig. 3. Survival rates (%) of treated groups upon tumor induction.

6 N. Erdogar et al. / International Journal of Pharmaceutics 471 (2014) 1–9

lesion of rat urinary bladder, and can be used as a marker in studieson the 2-stage process of urinary bladder carcinogenesis(Ito et al., 1983). So in this study, we used the BBN-induced ratbladder tumor to evaluate in vivo bladder cancer.

The survival rate of treatment groups was considered as theprimary parameter in evaluating antitumor efficacy of cationicnanoparticles as this parameter is definitive in antitumor studiesanimal models upon tumor induction. As seen in Kaplan–Meierplot from Fig. 3, all treatment groups were followed up until day 83(week 12) and it is evident that the drug loaded CS-PCL core-shellnanoparticles resulted in the highest survival rate among allgroups with the largest number of animals alive for the longesttime period.

Survival rate plots indicate core-shell nanoparticles withpositive coating with the CS and negative core with the PCL proveto be the optimum intravesical delivery system for the MMC. Thisconfirms the previous findings in in vitro and cell culture studieson cytotoxicity of the MMC loaded nanoparticles against bladdercarcinoma cell line MB49 and cellular uptake studies carried out onhealthy and cancerous bladder cells as well as several other cancerlines indicating higher cytotoxic activity and better cellular uptakewith the CS-PCL core shell nanoparticles (Erdogar et al., 2012;Bilensoy et al., 2009). It was also demonstrated that the MMCloaded CS-PCL nanoparticles resulted in a longer retention time inrat bladder with no significant effect on urine volume (Bilensoyet al., 2009). Therefore, it is in accordance with the previous datathat CS-PCL core-shell nanoparticles encapsulating MMC resultedin the best survival rate findings in tumor induced rat model.

It is still controversial how long and how frequently instillationsof intravesical chemotherapy have to be given. From a systematicreview of the literature of randomized clinical trials, that havecompared different schedules of intravesical chemotherapyinstillations, one can only conclude that the ideal duration andintensity of the schedule remains undefined because of theconflicting data (Babjuk et al., 2012). However, the most widelyused MMC protocol is weekly instillations for the first 8 weeks thanmonthly instillations up to one year. That protocol is an intensiveschedule, requires a great number of invasive procedures such asreplacement urethral catheter besides the systemic toxicity. Todecrease the number of instillations with successful clinicalresponse will result in decreased toxicity and increase in qualityof life of cancer patient. 83 days effective CS-PCL-MMC nano-particles compared with weekly single MMC solutions could allowto reach this aim because it requires fewer instillations with lowerMMC dose.

It was very interesting to observe that the unloaded CSnanoparticles had a very positive effect on the survival rate ofbladder tumor bearing rats. This could be attributed to theintrinsic antitumor properties of chitosan reported as dose-dependent tumor-weight inhibition against S-180 tumor ang H-22 (Mouse hepatoma) tumor models upon oral and iv adminis-tration routes (Qi and Xu, 2006). The chitosan nanoparticles withpositive charge and small diameter were suggested to bepromising agents for antitumor efficacy. The chitosan was alsoreported to show a growth inhibitory effect on tumor cells,inhibition of tumor angiogenesis and inhibitory effect on tumormetastasis (Tokoro et al., 1988; Carreno-Gomez, 1999; Murataet al., 1989, 1991). In another study, growth-inhibitory effect ofchitosan on human bladder tumor cells was found. It was foundthat chitosan activates caspase-3 which is a common mediator ofthe apoptotic pathway that plays a dominant role in the deathsignaling iniated by various antitumor agents. It was suggestedthat chitosan may interact with the cell membrane to triggerapoptosis (Hasegawa et al., 2001).

Surprisingly, the MMC loaded chitosan nanoparticles were notas effective as the blank CS nanoparticles. This is believed to be as aresult of the strong physicochemical interaction between MMCand sodium tripolyphosphate (TPP) which is used as a macromol-ecule to form nanoaggregates with chitosan during preparation bythe ionic gelation technique due to its strong charge. Thisphenomenon was in fact demonstrated in our previous studywith the DSC analysis (Bilensoy et al., 2009). The DSC thermogramsof MMC and nanoparticle components used in our previous studywere also given. As can be seen in DSC thermograms, MMC shows astrong interaction with TPP suggested by the disappearance of TPPmelting endotherms around 110 �C. It is noteworthy that MMCdoes not have the same type of interaction with CS since CSdecomposition endotherm is still present when it is analyzedtogether with MMC. During the optimization studies of the CSnanoparticles loaded with the MMC, it was found that when thedrug is dissolved in TPP aqueous solution drug loading issignificantly higher than when the drug is dissolved CS solution(unpublished results). This confirms the fact that the MMCinteracts strongly with the TPP. It was also believed that chitosanmay interact with NaCl present in the MMC used in this study,therefore altering encapsulation and release properties of theactive ingredient as well as diminishing its intrinsic anticanceractivity. Actually this reduction in intracellular uptake of chitosannanoparticles when loaded with a charged active molecule wasdemonstrated in cellular uptake studies performed by Bilensoy

Fig. 4. Percentage of increase in weight of rats treated with MMC loaded and blanknanoparticle formulations compared with healthy bladder weight (�SD).

Fig. 5. Histopathological images of rats treated with nanoparticle formulations and MMMMC commercial product solution Fig. 5c – MMC-CS-PCL nanoparticle formulation Fig.Fig. 5f – CS nanoparticle formulation applied groups) (n – nucleus, v – vacuole, m – m

N. Erdogar et al. / International Journal of Pharmaceutics 471 (2014) 1–9 7

et al. (2009). In these studies, uptake of fluorescent marker loadedCS nanoparticles were less than CS-PCL or PLL-PCL nanoparticles invarious cell lines including MCF7, Hela, L929, G/G and MB49 cellsproving a strong interaction between MMC and NaTPP. The MMChas a positive charge itself and in acidic urine pH, was believed tohave a higher positive charge due to protonable amine groups inacid media. Therefore, strong charge interaction of the drug withthe CS is believed to mask the cytotoxic effect of the drug whenencapsulated in the CS nanoparticles. There exists no suchinteraction with CS-PCL core material and therefore this groupis proved to give the highest survival rate (Bilensoy et al., 2009).Koppolu and colleagues utilized precipitation–coacervation tech-nique to comprehensively characterize the effects of variousformulation factors such as precipitant salt concentration.Preliminary studies indicated that the salt concentration hadnegligible effect on the chitosan particle size, polydispersity andencapsulation efficiency. NaCl was used as precipitant salt was notable to form particles (Koppolu Prasanth et al., 2014). In anotherstudy; they have shown that the rate of TPP binding to chitosan can

C commercial product solution (Fig. 5a – untreated, tumor-induced group Fig. 5b –

5d – CS-PCL nanoparticle formulation Fig. 5e – CS-MMC nanoparticle formulationitochondria, c – collagen fibers, g – golgi apparatus, * – nanoparticle).

8 N. Erdogar et al. / International Journal of Pharmaceutics 471 (2014) 1–9

be reduced through the addition of monovalent salt as NaCl. Slowionotropic gelation kinetics achieved at high NaCl and low TPPconcentrations can facilitate the elucidation of the micro- andnanogel formation mechanism (Huang and Lapitsky, 2012).

These reported anticancer activities of chitosan helped improvethe anticancer efficacy of nanoparticles both as core and as shellmaterial. All in vitro, cell culture and in vivo studies showed CS-PCLloaded with MMC is the most promising formulation among theevaluated.

As a result, survival rate data revealed that the PCL nano-particles coated with CS emerge as promising systems due to acombination of their positive surface charge in the shell given byCS coating and negative core charge by PCL resulting in improvedpassage of nanoparticles through the mucus gel layer of thebladder wall and better penetration, good biocompatibility andloading/release properties of the core polymer PCL and intrinsicanticancer, urothelial barrier penetration characteristics andmucoadhesive properties of CS coating (Fig. 4).

Fig. 5a–f show histopathological images of bladder biopsysamples obtained from different groups. Some of the importantaspects found in the histopathological evaluation is as follows:Fig. 5a (positive control-untreated tumor induced rat) and Fig. 5btreated MMC commercial product both displayed atypical cellswith giant vacuoles indicating progression of tumor cells in thebiopsies. Fig. 5a–f demonstrates very well the formation of tumorin all groups with different degrees of healing or progression oftumor according to the treatments received during the study.

In the frame of histopathological evaluation, treatment groupswere evaluated in terms of tissue damage. Tumor formation wascharacteristic with large number of microvili in luminal surface astumor markers. In addition to this, dent nuclei, sharp-thorn likestructures and vacuoles clearly show the formation of tumor tissuein untreated group.

The MMC loaded CS-PCL nanoparticles treatment groupdisplayed a large section of mitochondria observed as normalstructural appearance. No edema was observed and cell nucleiwere smooth. the vacuoles were small. Generally cells werehealthy and a large number of nanoparticles were observed in thebladder tissues (Fig. 5c). Nanoparticle accumulation was observedin the bladder tissues as indicated with arrows for both MMCloaded and blank CS-PCL nanoparticles.

If an evaluation is made in terms of destruction; maximaldestruction was observed for the group treated with MMC loadedCS nanoparticles. The shortest survival time was also seen for thisgroup treated with MMC loaded CS nanoparticles. As a result,survival time is compatible with destruction degree.

The most promising formulation was the MMC loaded CS-PCLnanoparticles in terms of nanoparticle accumulation in tissue anddegree of destruction. The maximum amount of nanoparticleaccumulation in tissue and lowest tissue damage were observedwith this group. At the same time, this group had the longestsurvival period.

Purpose of this study was to increase local effect on bladdersurface with developed nanoparticle formulations which havebioadhesive properties. However, if an active agent passes onto thesystemic circulation causing a systemic effect is undesirable forlocal applications. The active agent MMC which is used in our studyhas a large number of systemic side effects. The most commonserious side effect is bone marrow suppression, particularlythrombocytopenia and leukopenia. In addition, tiredness andshortness of breath due to decrease in the number of red bloodcells (anemia), infection due to decrease in the number of whiteblood cells are reported. As a result of blood analysis, the MMC inblood circulation was not determined in all animal groups and itwas concluded that this situation is in correlation with our basicpurpose.

5. Conclusions

In this paper, we describe the in vivo antitumor efficacy ofmucoadhesive and cationic nanoparticles as nano-sized drugcarriers for local chemotherapy of bladder tumors. The MMCloaded CS-PCL core-shell nanoparticles in the tumor tissue showedimproved antitumor efficacy compared to other formulations, asshown by improvement in survival rates, histopathologicalevaluation and bladder weight. In this study, it was clearlyobserved that the MMC loaded CS-PCL cationic nanoparticlesprovide a significantly improved perspective in chemotherapy ofbladder tumors and preventing recurrence and therefore should befurther investigated.

Conflict of interest

The authors state no conflict of interest.

Acknowledgements

Authors would like to acknowledge TUBITAK Scientific ResearchProject (109S172) for financial support of this study. Alper B. Iskithas been supported by the Turkish Academy of Sciences, in theframework of the Young Scientist Award Program (EA-TUBA-GEBIP/2001-2-11).

References

Andrews, G.P., Laverty, T.P., Jones, D.S., 2009. Mucoadhesive polymeric platforms forcontrolled drug delivery. Eur. J. Pharm. Biopharm. 71, 505–518.

Arentsen, C.H., Hendricksen, K., Oosterwijk, E., Witjes, J.A., 2009. Experimental ratbladder urothelial cell carcinoma models. World J. Urol. 27, 313–317.

Babjuk, M., Oosterlinck, W., Sylvester, R., Kaasinen, E., Böhle, A., Palou, J., Rouprêt, M.,2012. Guidelines on Non-muscle-invasive Bladder Cancer (TaT1 and CIS). http://www.uroweb.org/gls/pdf/05_TaT1_Bladder_Cancer.pdf.

Barthelmes, J., Perera, G., Hombach, J., Dünnhaupt, S., Bernkop-Schnürch, A., 2011.Development of mucoadhesive nanoparticulate drug delivery system for atargeted drug release in the bladder. Int. J. Pharm. 416, 339–345.

Bilensoy, E., Sarisozen, C., Esendagli, G., Dogan, A.L., Aktas, Y., Sen, M., Mungan, N.A.,2009. Intravesical cationic nanoparticles of chitosan and polycaprolactone forthe delivery of mitomycin C to bladder tumors. Int. J. Pharm. 371, 170–176.

Bravo-Osuna, I., Vauthier, C., Farabollini, A., Palmieri, G.F., Ponchel, G., 2007.Mucoadhesion mechanism of chitosan and thiolated chitosan-poly(isobutylcyanoacrylate) core-shell nanoparticles. Biomaterials 28, 2233–2243.

Carreno-Gomez, B.D.R., 1999. Evaluation of the biological properties of solublechitosan and chitosan microspheres. Int. J. Pharm. 148, 231–240.

Carter, S.K., Crooke, S.T., Alder, N.A., 1979. Mitomycin C: Current Status and NewDevelopments. Academic Press, New York.

Cui, F., Qian, F., Chunhua, Y., 2006. Preparation and characterization ofmucoadhesive polymer-coated nanoparticles. Int. J. Pharm. 316, 154–161.

Cohen, S.M., Ellwein, L.B., 1990. Cell proliferationin carcinogenesis. Science 249,1007–1011.

Erdogar, N., Iskit, A.B., Mungan, N.A., Bilensoy, E., 2012. Prolonged retention and invivo evaluation of cationic nanoparticles loaded with mitomycin C designed forintravesical chemotherapy of bladder tumours. J. Microencapsul. 29, 576–582.

Chen, T.X., Yamamoto, S., Gen, H., Murai, T., Mori, S., Oohara, T., Makino S, WanibuchiH., Fukushima S, 1998. Infrequent involvement of microsatellite instability inurinary bladder carcinomas of the NON/Shi mouse treated with N-butyl-N-(4-hydroxybutyl)nitrosamine. Cancer Lett. 123, 41–45.

Ferlay, J., Bray, F., Pisani, P., Parkin, D.M., 2001. Cancer incidence. MortalityPrevalence Worldwide. .

Giannantoni, A., Di Stasi, S.M., Chancellor, M.B., Costantini, E., Porena, M., 2006. Newfrontiers in intravesical therapies and drug delivery. Eur. Urol. 50, 1183–1193.

Hasan, A.S., Socha, M., Lamprecht, A., Ghazouani, F.E., Sapin, A., Hoffman, M.,Maincent Ubrich, P.N., 2007. Effect of the microencapsulation of nanoparticleson the reduction of burst release. Int. J. Pharm. 344, 53–61.

Hasegawa, M., Yagi, K., Iwakawa, S., Hirai, M., 2001. Chitosan induces apoptosis viacaspase-3 activation in bladder tumor cells. Jpn. J. Cancer Res.: Gann 92, 459–466.

Heijden van der, A.G., Witjes, J.A., 2003. Intravesical chemotherapy: an update-newtrends and perspectives. EAU Update Series 1, pp. 73–79.

Huang, Y., Lapitsky, Y., 2012. Salt-assisted mechanistic analysis of chitosan/tripolyphosphate micro- and nanogel formation. Biomacromolecules 13, 3868–3876.

Ito, N., Fukushima, S., Shirai, T., Nakanishi, K., 1983. Effects of promoters on N-butyl-N-(4-hydroxybutyl)nitrosamine-induced urinary-bladder carcinogenesis in therat. Environ. Health Perspect. 50, 61–69.

N. Erdogar et al. / International Journal of Pharmaceutics 471 (2014) 1–9 9

Kaufman, D.S., 2006. Challenges in the treatment of bladder cancer. Ann. Oncol. 5,V106–V112.

Koppolu Prasanth, B., Smith, S.G., Ravindranathan, S., Jayanthi, S., Kumar, K.,Zaharoff, D.A., 2014. Controlling chitosan-based encapsulation for protein andvaccine delivery. Biomaterials 35, 4382–4389.

Lamprecht, A., Ubrich, N., Hombreiro Perez, M., Lehr, C., Hoffman, M., Maincent, P.,1999. Biodegradable monodispersed nanoparticles prepared by pressurehomogenization–emulsification. Int. J. Pharm. 184, 97–105.

Lehmann, J., Retz, M., Sidhu, S.S., Suttmann, H., Sell, M., Paulsen, F., Harder, J.,Unteregger, G., Stöckle, M., 2006. Antitumor activity of the antimicrobialpeptide magainin II against bladder cancer cell lines. Eur. Urol. 50, 141–147.

Le Visage, C., Rioux-Leclercq, N., Haller, M., Breton, P., Malavaud, B., Leong, K., 2004.Efficacy of paclitaxel released from bio-adhesive polymer microspheres onmodel superficial bladder cancer. J. Urol. 171, 1324–1329.

Malmstrom, P.U., 2003. Intravesical therapy of superficial bladder cancer. Crit. Rev.Oncol. Hematol. 47, 109–126.

Miyake, H., Eto, H., Hara, I., So, A., Li, D., Gleave, M.E., 2004. Synergistic antitumoractivity by combined treatment with gemcitabine and antisense oligodeox-ynucleotide targeting clusterin gene in an intravesical administration modelagainst human bladder cancer kotcc-1 cells. J. Urol. 171, 2477–2481.

Murata, J., Saiki, I., Makabe, T., Tsuta, Y., Tokura, S., Azuma, I., 1991. Inhibition oftumor-induced angiogenesis by sulfated chitin derivatives. Cancer Res. 51, 22–26.

Murata, J., Saiki, I., Nishimura, S., Nishi, N., Tokura, S., Azuma, I., 1989. Inhibitoryeffect of chitin heparinoids on the lung metastasis of B16-BL6 melanoma. Jpn. J.Cancer Res.: Gann 80, 866–872.

Ohyama, C., Kawamura, S., Satoh, M., Saito, S., Yoshikawa, K., Hoshi, S., Orikasa, S.,1997.Endoscopic observation for detection and monitoring of N-butyl-N-(4-hydrox-ybutyl) nitrosamine-induced bladder tumor in rats. Urol. Res. 25, 183–186.

Oliveira, P.A., Colaço, A., De la Cruz, L.F., Lopes, C., 2006. Experimental bladdercarcinogenesis-rodent models. Exp. Oncol. 28, 2–11.

Parekh, D.J., Bochner, B.H., Dalbagni, G., 2006. Superficial and muscle-invasivebladder cancer: principles of management for outcomes assessments. J. Clin.Oncol. 24, 5519–5527.

Quebbeman, E.J., Hoffman, N.E., Ausman, R.K., Hamid, A.A.R., 1985. Stability ofmitomycin admixtures. Am. J. Hosp. Pharm. 42, 1750–1754.

Ruckmani, K., Sivakumar, M., Ganeshkumar, P.A., 2006. Methotrexate loaded solidlipid nanoparticles (SLN) for effective treatment of carcinoma. J. Nanosci.Nanotechnol. 6, 2991–2995.

Qi, L., Xu, Z., 2006. In vivo antitumor activity of chitosan nanoparticles. Bioorg. Med.Chem. Lett. 16, 4243–4245.

Sagara, Y., Miyata, Y., Nomata, K., Hayashi, T., Kanetake, H., 2010. Green teapolyphenol supresses tumor invasion and angiogenesis in N-butyl-(4-hydox-ybutyl) nitrosamine-induced bladder cancer. Cancer Epidemiology 34, 350–354.

Shimizu, H., Akasaka, S., Suzuki, S., Akimoto, M., Shimada, T., 2001. Preferential genetransfer to BBN-induced rat bladder tumor by simple instillation of adenoviralvector. Urology 57, 579–584.

Stolk, L.M., Fruijtier, A., Umans, R., 1986. Stability after freezing and thawing ofsolutions of mitomycin C in plastic minibags for intravesical use. Pharmaceu-tisch Weekblad Scientific Edition 8, 286–288.

Tokoro, A., Tatewaki, N., Suzuki, K., Mikami, T., Suzuki, S., Suzuki, M., 1988. Growth-inhibitory effect of hexa-N-acetylchitohexaose and chitohexaose against meth-A solid tumor. Chem. Pharm. Bull. (Tokyo) 36, 784–790.

Tyagi, P., Li, Z.H., Chancellor, M., De Groat, W.C., Yoshimura, N., Huang, L., 2004.Sustained intravesical drug delivery using thermosensitive hydrogel. Pharm.Res. 21, 832–837.

Tyagi, P., Wu, P.C., Chancellor, M., Yoshimura, N., Huang, L., 2006. Recent advances inintravesical drug/gene delivery. Mol. Pharm. 3, 369–379.

Thrasher, J.B., Crawford, E.D., 1992. Complications of intravesical chemotherapy.Urol. Clin. North America 19, 529–539.

Zetter, B.R., 1998. Angiogenesis and tumor metastasis. Annu. Rev. Med. 49, 407–424.