Annals of Burns and Fire Disasters - vol. XII - n. 1 - March 1999

A NEW DELIVERY SYSTEM OF ANTIBIOTICS IN THE TREATMENT OF BURN WOUNDS

Giannola L.I.,* De Caro V.,* Adragna E.,* Giandalia G.,* Giannola G.,* D'Arpa N.,** Napoli B.,** D'Amelio L.M.,** Genovese LM,** Lombardo C.,** Masellis M.**

* Dipartimento di Chimica e Tecnologie Farmaceutiche, Università di Palermo, Palermo, ltaly
** Divisione di Chirurgia Plastica e Terapia delle Ustioni, A.R.N.A.S. Ospedale Civico e Benfratelli *** G. Di Cristina - Maurizio Ascoli", Palermo

SUMMARY. The development of a new antibiotic delivery system suitable for application on burn wounds is described. The system was designed to be administered together with an implant of epidermal cultured keratinocyte stem cells. The antibiotic was entrapped in a reservoir compartment which was capable of releasing the drug at a controlled rate. The desired delivery profile was determined in vitro using a general two-compartment linear time-invariant model. The suitable rate for a release of the antibiotic in sufficient amounts for a therapy of 5-6 days was obtained using a combined mechanism based on diffusion through a membrane and decrease of drug movement properties in the vehicle. The topical release system that is proposed limited the indiscriminate use of broad-spectrum antibiotics, thus reducing the possible incidence of undesirable mulfiresistance. Interruption of the system in the area after five days of treatment was easy as it caused the patient no pain, no trauma to the zone, and no damage to the graft. Eight days after application it was observed that the treatment of patients with cultured keratinocytes had been successful and that the lesions had healed.

It is well known that the burn, compared with other forms of trauma, produces extensive skin barrier disruption together with devitalization of tissues, which causes the formation of large raw areas. The burn wound has a much higher incidence of infection than other forms of trauma, owing to alteration of the cellular immune responses.' Cultured epidermal keratinocyte allografts or autografts have been used as a biological dressing for burn wounds. Flexible natural or synthetic thin films applied to the grafts may be used as barriers for the damaged skin against dust and dirt. However, the contamination of the wound by virulent microflora is often responsible for ineffectiveness of treatment. Infection remains the major problem in the treatment of patients, leading to the indiscriminate use of broad-spectrum antibiotics that transform burn wounds into sites of multiresistant virulent bacteria.
During the past decade a plethora of reports has appeared in the literature regarding dermal and transdertnal delivery of drugs through intact skin with the aim of duplicating the benefits of intravenous drug infusion while avoiding its hazards. An extensive study has been made of the principles, mechanisms, and physicochemical parameters important in the control of percutancous penetration of drugs as well the qualitative and quantitative aspects in the prediction of absorption.
Although a considerable amount of research has been devoted to transdermal drug delivery systems, very little work has been done on applications on skin following thermal injury. The acceptability of a delivery system as a drug administration option can be dramatically altered when burn wounds are involved.
Drugs can be released by dermal and transdermal delivery systems using different mechanisms. The physical and chemical properties of both the drug and the system components can be optimized to reach the desired release profile, although the skin's barrier properties tend to normalize the amount of penetrating drug molecules.
For the topical application of drugs on damaged skin, when there is no barrier, primary resistance to mass transfer depends on the vehicle in which the active ingredient is dispersed. Drug discharge from the vehicle can be regulated by changing the movement properties of the active ingredient in the vehicle itself, by varying the thickness of the layers applied, or by choosing appropriate formulation components. The use of various polymeric vehicles allows optimization of the drug liberation rate from conventional pharmaceutical dosage forms. The selection of polymer and other components thus becomes critical to the success of the formulation. The desired release rate and drug level can also be achieved by using appropriate delivery systems. An excellent liberation rate can also be achieved using films or membranes to match the diffusivity of the drug molecules by selecting useful plasticizers or vehicles, or naturally occurring, sernisynthetic, or synthetic polymers.
The present study reports on a re-evaluation of therapeutic practices and on the development of a new antibiotic delivery system suitable for application in burn patients. The system was designed to be administered after the implantation of epidermal cultured keratinocyte stem cells and was capable of releasing antibiotics with ratecontrolled characteristics. In this type of pre-programmed drug delivery system, the active ingredient is entrapped in a reservoir compartment and covered by a rate-controlling membrane having a specific permeability. The membrane controls the molecular diffusion of drugs crossing the barrier and passing into the surrounding environment at a pre-programmed rate. The resultant equation was C = 0.4159 V'-t~ 0.8649.

In vitro experiments: apparatus
The diffusion of antibiotics was determined using Franz-type diffusion cells (Fig. 1). Each cell included four major elements:

1. formulation chamber for the source of diffusing drug to be tested (cap);
2. receptor chamber for the simulated biological medium (body) into which the drug was transferred;
3. membrane allocation site, separating the donor from the receptor compartment; and
4. side arm for withdrawal of samples (sample port).

Accurately weighed quantities of pharmaceutical fon-nulation were placed in the cap to apply the same thickness layer of material in every cell. The receptor chamber of every cell was filled with 42 ml of isotonic phosphate buffer solution (pH 7.4) used as a simulated receptor phase. The temperature was maintained at 37 ± 0.5 'C by means of thermostatically controlled water entering the lower port of the water jacket surrounding the body chamber and circulating out through the upper port. All the cells had an exposed surface area of 10.7 CM2 . The receptor solution was stirred by means of a magnetic follower rotating at 600 rpm, which greatly increased mixing efficiency and reduced the tendency to form a stagnant boundary layer next to the membrane surface. The agitation speed was established to maintain a hydrodynamic condition such that the thickness of the diffusion boundary layer was at a minimum; the primary convective flow below the drugreleasing surface was also minimized. The degree of dispersal of coloration from pennanganate was used to make a qualitative assessment of the stirring rate. At regular time intervals (30 min or I h) samples (I ml) were removed from the centre of the receptor chamber through the sample port, using a long needle and a syringe. To avoid saturation phenomena and to maintain the sink conditions, the sample volume taken was replaced by fresh buffer solution, injected by syringe, ensuring that no air was drawn into the compartment.
Drug permeation was monitored by analysis of the cumulative amount of antibiotic reaching the receptor phase. For each experiment the percentage amount of drug transferred was plotted as a function of time. Each experiment was repeated five times. The results are expressed as the mean ± SEM. Reproducibility was within 3% of the mean. Fig. 2 shows the cumulative amounts of vancomycin permeating from the formulation through VelodermO, which functioned as a diffusion membrane and gave reproducible results. The residual drug content in the formulation and the amount of drug released matched the original content, confirming that no drug was entrapped in Veloderm@.

UV spectrophotornetry
The detection of vancomyein was carried out by UV spectrophotometry with a Shimadzu 1601 UV-VIS instrument, using the appropriate blank and calibration curve at 280.4 mn. The standard solutions had concentration ranges between 0.06 and 0. 15 mg/ml, which were the same as those predicted in the samples from the cells. The calibration curve gave the specific extinction coefficient (E,,, 1 cm = 0.034 at 280.4 nm) of vancomycin hydrochloride with a correlation coefficient 0.9996.

Preparation of vancomycin gel-like dispersion for in vitro experiments
Vancomycin hydrochloride (256 mg), corresponding to 250 mg of vancomycin, was dissolved in 97 ml of a sterile, apyrogenic and isotonic phosphate buffered saline (pH solution, 7.0). Three grams of hydroxyethyleellulose were added to the solution. The resulting mass was sonicated in a Branson 5200 ultrasound bath until a homogeneous-like gel dispersion was obtained.
All manufacturing procedures for pharmaceutical dispersion were performed in order to ensure the product's sterility.

Keratinocyte in vitro culture technique
The skin biopsies were harvested from burn patients, cleaned with 70% ethanol, and processed according to Rheinwald and Green's keratinocyte isolation method.' Skin biopsies were trypsinized overnight at 4 'C in a 0.05% trypsin/0.02% EDTA solution. The cells were isolated from the epidermal/dermal junction and plated onto a feeder monolayer of lethally irradiated 3T3 cells. The medium was made up of three parts Dulbecco's modified with Eagle's medium plus one part Ham's F12 medium. The following were added to the mixture: cholera toxin (0.1 nM), hydrocortisone (0.4 pg/ml), triiodo-l-thyronine (20 pM), insulin (5 pg/ml), transferrin (5 pg/ml), fungizone (250 pg/ml), penicillin (1000 IU/ml), streptomycin (1000 p g/ml), and 10% foetal calf serum. Epidermal growth factor (10 ng/ml) was added with the first medium change. Cells were cultured at 37 'C in a 5% CO atmosphere. The medium was changed every two days.
Subconfluent primary cultures were trypsinized and cells were plated in a secondary culture. The confluent secondary cultures were detached from the surface vessel with the neutral protease Dispasell, washed in serum-free medium, and placed on sterile Veloderm@.

All patients included in the study presented deep thirddegree flame burns. The wounds were treated by surgical intervention and autoskin grafting, remaining uncovered after treatment.
The preparation of the wound bed for the application of the antibiotic delivery system was of decisive importance for the outcome of the whole procedure. The wound bed had to be clean, free from necrotic tissue residue, and with minimum contamination. Bacteriological surveillance was performed by monitoring the damaged zone at every medication in order to observe any signs of sepsis. Wound care included the application of antibiotic cleanser solutions as topical agents. The cleanser was prepared on the basis of the antibiotic assay results. After 72 h of treatment, the wound bed was ready for placement of the delivery system shown in Fig. 3, which was then protected with a sterile bandage. The system was maintained unchanged on the wound for five days, when it was easily removed without any pain to the patient, trauma to the zone, or damage to the graft.

Optimization of drug release from an appropriate delivery system was obtained in order to evaluate the treatment of deep burns with cultured keratinocytes combined with antibiotic therapy.
The antibiotic vancomycin, which was selected as the antibiotic model on the basis of antibiotic assay results, possesses good water- solubility characteristics, while a variety of commonly used topical formulations, including lipophilic ointments, oil-in-water creams, and oily-type preparations, were not recommended. The administration of simple aqueous sterile solutions would involve intermittent and often subtherapeutic drug levels, a high frequency of application, and other serious obstacles.
To achieve the desired delivery rate we used a combined mechanism based on diffusion through a membrane coupled with the decreased movement properties of the drug molecules in the vehicle. The antibiotic was trapped in a water dispersion of hydroxyethylcellulose used as a vehicle. This polymer is biocompatible, non-toxic, inexpensive, easily available, and capable of forming gellike viscous dispersions. The amount of hydroxyethylcellulose affects the viscosity of the formulation, which is one of the most important parameters affecting the drug liberation rate owing to the reduced molecular movement in the dispersion.
As the limiting step on drug transfer (i.e., the skin) is absent in burn patients, we propose the use of a system in which the skin is replaced by a cultured keratinocyte layer placed on a dressing membrane possessing specific permeability capacities.
The selection of the membrane was made on the basis of the observation of the diffusional behaviour of commercially available materials commonly used in wound dressing: Adaptic@, Trofoprocess@, TransprocessO, KeratoprocessO, and Veloderm@. Of these, Veloderm@ showed the best diffusional characteristics. Permeation performance was assessed by detecting the degree of movement and dispersal of coloration from a permanganate solution passing through the membrane to the acceptor.
Fig. 3 presents a schematic representation of the drug release mechanism from the proposed system.
The main components of the system were:

• the cultured keratinocyte layer;

• the biocompatible permeable membrane, which acted as a physical support for the keratinocyte layer, as a rate-controlling diffusion barrier, and as a dressing material to protect the wound and prevent penetration of environmental agents;

• the drug reservoir, consisting of the hydrophilic drug incorporated in a suitable polymer gel-like sterile dispersion, used as a vehicle, placed on the membrane in a layer about 5 mm thick;

• the external backing, composed of lipophilic occlusive petrolatum gauze, which occluded the system, prevented back diffusion of hydrophilic drug molecules, and eliminated drug losses.

In vitro drug release experiments, in which the skin was replaced by the permeable membrane, were performed in order to test the behaviour of the formulation in conditions approaching in vivo conditions. Drug release was followed by periodic measuring of the amount released in the simulated acceptor fluid. The active ingredient leached from the system and transferred through the membrane into the receptor phase was measured by UV spectrophotometric quantitative analysis. The peaks observed were highly reproducible and were linearly related to concentration over the range of 15 mg/100 ml. Drug flux was calculated by dividing the total amount obtained by the transfer area and sampling interval.
Drug release was evaluated by plotting the percentage amount of drug discharged from the formulation versus time. Fig. 2 shows the liberation profile of vancomycin. The amounts of active ingredient released from the hydroxyethylcellulose dispersion showed time dependence.
To establish the complete release time and the system's shelf life after application we examined the potential mathematical correlation between drug amount released and time.
Since Higuchi's pioneering , studies various mathematical approaches based on differential equations and thermodynamic and kinetic concepts of delivery have been proposed in order to describe the release from topically applied drugs.` In previous publications" we discussed the behaviour and main release models used to describe the drug discharged from multiparticulate drug delivery systems.
We attempted to describe the release profile of vancomycin using a model function. On application of differential rate treatments and linear regression analysis, the evidence indicated that the amount of drug discharged increased linearly with the square root of time, suggesting that the model for diffusion controlled transport should be followed. Fig. 4 shows the discharged drug amounts versus the square root of time. The diffusion equation gave consistently higher values for the correlation coefficient (0.996-0.999) than other equations (0.970-0.990).

Figs. 4a, b - Clinical case: 3rd-degree burn in lumbar region in which the autologous graft did not take owing to Staphylococcus infection. A, B - positioning of autologous keratinocyte sheets mounted on petrolatum gauze (above) and on Veloderm (below).

Fig. 4c - Positioning of autologous keratinocyte sheets mounted on petrolatum gauze (above) and on Veloderm (below). Fig. 4d - Application of sheets on release system containing specific antibiotic (vancomycin).

Fig. 4e - Medication on day 4 - comparison of sheets mounted on petrolatum gauze and on Veloderm. Fig. 4f - Appearance on day 8. Fig. 4g - Final result.

On the basis of these observations it was possible, using geometric extrapolation, to predict that the drug administered by the system would be sufficient for a therapy of 5-6 days with only one application. During the clinical investigations it was observed that five days after a single application of the antibiotic delivery system, when the bandage was removed for a check, the system was found to have adhered perfectly to the treated surface. The whole system was then easily removed and the wound area appeared dry and without any moist or serous exudate. This result was probably a consequence of interactions between the exudate and the components of the formulation applied, involving osmotic phenomena in the transfer through the membrane. Eight days after application it was observed that the treatment of patients with the cultured keratinocytes had been successful and that the lesions had healed.
Clinical cheeks showed no sign of sepsis, proving that the wounds had not been not colonized by micro-organisms, which may cause inhibition of keratinocyte growth during the period of application.

The topical release system limited the indiscriminate use of broad-spectrum antibiotics, thus reducing the possible incidence of undesirable multiresistance.
The modest dose of antibiotics administered produced constant drug levels in amounts sufficient to inhibit sepsis and prolong action. Suspension of the system from the treated area after five days was easy and caused no pain to the patient, no trauma to the zone, and no damage to the graft. Eight days after application it was observed that treatment of patients with the cultured keratinocytes had been successful and that the lesions had healed.

RESUME. Les Auteurs décrivent un nouveau système pour le transport des antibiotiques utilisable dans le traitement des brûlures Le nouveau système a été conçu pour être employé avec l'implantation des cellules-souches des kératinocytes cultivés épidermiques. L'antibiotique a été bloqué dans un compartiment réservoir capable de libérer le médicament à vélocité contrôlée. Le profil de transport désiré a été déterminé in vitro en utilisant un modèle général linéaire à deux compartiments invariant dans le temps. La vélocité appropriée pour le transport de l'antibiotique en quantité suffisante pour une thérapie de 5-6 jours a été calculée moyennant un mécanisme combiné basé sur la diffusion à travers une membrane et la diminution des propriétés de mouvement du médicament du véhicule. Le système proposé pour la libération topique a limité l'emploi sans discernement des antibiotiques à spectre large et réduit la possibilité de la manifestation d'une multirésistance indésirable. L'interruption de l'emploi du système dans la zone traitée après cinq jours n'a pas causé aucune douleur pour les patients ni traumatismes localisés ni détérioration de la greffe. Les Auteurs ont observé, huit jours après l'application, que le traitement des patients avec les kératinocytes cultivés a été effectué avec succès et que les lésions se sont cicatrisées.

BIBLIOGRAPHY

1. Allgower M., Schoenenberger G.A., Sparkes B.G.: Burning the largest immune organ. Burns, 21 (Suppl. 1): S7-S47, 1995.
2. Odessey R.: Multicenter experience with cultured epidermal autograft for treatment of burns. J. Burn Care Rehabil., 13: 174-80, 1992.
3. Banakar U.V.: Percutancous absorption and principles of drug delivery. In: "Transdermal drug product development", Banakar UN., Osborne D.W. (eds), Technomic Publishing AG, 15-48, 1995.
4. Rheinwald J.G., Green H.: Cell serial cultivation of strains of human epidermal keratinocytes: the formation of keratinizing colonies from single cells. Cell, 6: 331-44, 1975.
5. Higuchi T.: Permeation of mechanical barriers by chemical agents: Medical Laboratories Contract Report 32, Chemical Corp. Medical Laboratories, Army Chem. Center, NM, 1954 and J. Pharm. Sci., 50: 874-5, 1961.
6. Bunge A.L.: Predicting release rates from topical formulations containing drugs in suspension. In: "Prediction of percutaneous penetration, methods, measurements, modeling", vol. 2, Scott R.C., Guy R.H., Hadgraft J., Bodd6 H.E. (eds), IBC Technical Services, 577-84, 1991.
7. Giannola L.I., De Caro V., Rizzo M.C.: Preparation of white beeswax microspheres loaded with valproic acid and kinetic study of drug release. Drug Dev. Ind. Pharm., 21: 793-807, 1995.
8. Giannola L.I., De Caro V., Severino A.: Camauba wax microspheres loaded with valproic acid: preparation and evaluation of drug release. Drug Dev. Ind. Pharm., 21: 1563-72, 1995.
9. Giannola L.I., De Caro V.: Entrapment of phenytoin into microspheres of oleaginous materials: process development and in vitro evaluation of drug release. Drug Dev. Ind. Pharm., 23: 114552, 1997.