WO1998012987A1 - Microparticle formulation and method for inhibiting smooth muscle cell migration and proliferation - Google Patents

Abstract

The present invention contemplates a new formulation for a microparticle used in a drug delivery system. The microparticle is a blend of a biodegradable polymer and a water-soluble polymer. The microparticle also contains a therapeutic compound to be delivered. The present invention also contemplates a method of inhibiting the migration and proliferation of smooth muscle cells using the microparticle of this invention.

Description

MICROPARTICLE FORMULATION AND METHOD FOR INHIBITING SMOOTH MUSCLE CELL MIGRATION AND PROLIFERATION

Related Applications

This application claims priority from provisional application 60/026,631 filed on September 24, 1996.

Technical Field of the Invention

This invention relates to a new microparticle based formulation that can be used as a drug delivery system and a process of using said microparticle to deliver a therapeutic compound that inhibits smooth muscle cell migration and proliferation.

Background of the Invention

Vascular proliferative diseases, such as atherosclerosis, result from an excessive, inflammatory, fibroproliferative response to injury of the endothelial and smooth muscle cells (SMCs) of the arterial wall. The SMCs, which are stimulated to migrate from media to intima, are the principal fibroproliferative component of the lesion (Ross, R., Nature, 362:801 -809 (1993)). The lesion results in the stenosis of the arterial wall. Various invasive treatments are known and used for alleviating the stenosis such as grafting, endarterectomy and percutaneous transluminal angioplasty. Unfortunately, a significant portion of arteries reclose within the first six months after treatment and require repeat procedures. For example, percutaneous transluminal coronary angioplasty has a restenosis rate of 30-50% per year (Beatt, .J., et al., J. Am. Coll. Cardiol., 15, 491 -498 (1990)). The localized delivery of inhibitors of SMCs to an injured artery has been proposed for decreasing restenosis. A number of agents have been reported to inhibit SMC proliferation including heparin, immunosuppressants such as cyclosporin, calcium channel blockers, and angiotensin converting enzyme inhibitors. Additionally, antisense (AS) oligodeoxynucleotides (ODNs) and antibodies are presently being studied as potential therapeutic agents for inhibition of SMC proliferation and migration. Antisense oligodeoxynucleotides are short synthetic DNA molecules whose sequences are complementary to those present in specific target mRNAs within cells. Antisense oligodeoxynucleotide effectiveness against SMC proliferation and migration has been demonstrated in vitro using c-myc AS-ODNs. These c-myc AS-ODNs were found to suppress intimal thickening (Simons, M., et al., Nature, 359:67-70 (1992); Burgess, T.L., et al., PNAS, 92:4051 -4055 (1995); and Shi, Y. et al., Circulation, 90:944-951 (1994)).

The challenge presented with utilizing such agents has been in developing a safe and effective way for localized delivery of these agents to an injured artery. For example, polymer stents or catheters have been used for attempted intraluminal drug delivery of inhibitors (Lincoff, A.M., et al., J. Am. Coll. Cardiol., 23, 18A Abstract (1994); Goldman B., et al., Atherosclerosis, 65:215-255 (1987); and Wolinsky, H., et al., J. Am. Coll. Cardiol. , 15:475-481 (1990)). However, the problem with the stents and catheters is that washout of the drug occurs. Surgical implantation of polymer-based drug delivery systems has also been used. More specifically, ethylene-vinyl acetate copolymer matrices have been used to deliver heparin (Edelman, E.R., et al, PNAS, 87:3773-3777 (1990)). The results demonstrated that continuous delivery decreases intimal hyperplasia. Periadventitial silicone polymer collars have also been used to deliver dexamethasone (Villa, A.E., et al., J. Clin. Invest., 93:1243-1249 (1994)). The results demonstrated that prolonged drug delivery was necessary to prevent restenosis.

Although studies with ethylene-vinyl acetate and silicone polymers have demonstrated the feasibility and benefits of locally controlled adventitial delivery of SMC inhibitors, the long-term implantation of such non-degradable materials in vivo may result in an inflammatory response. Therefore, a need exists for a biodegradable polymer drug delivery system that can be used to deliver a therapeutic compound that inhibits SMC migration and proliferation.

Poly (D,L-lactic-co-glycolic acid) (PLGA) copolymers are synthetic biodegradable polymers that have been approved for human clinical use (Holland, S.J., et al, J. Controlled Rel. , 4:155-180 (1986)). Their degradation in vivo occurs by random non-enzymatic hydrolysis of the polyester bonds along the polymeric backbone at a rate dependent on the copolymer ratio (Reed, A.M., et al., Polymer, 22:494-498 (1981 )). As lactic and glycolic acid form, they are processed through normal metabolic pathways and are ultimately eliminated from the body as carbon dioxide. PLGA copolymers are typically formulated into microparticles and used as a drug delivery system.

Several microencapsulation techniques have been described for preparing PLGA microparticles containing a macromolecule. The most common methods are the solvent evaporation and solvent extraction methods carried out in a double-emulsion systems (such as water-in-oil-in- water) as described in Alonso et al., Vaccine, 12:299-306 (1994). Alonso et al. describe the entrapment of purified tetanus toxin in PLGA microparticles using both of these methods. The release of a macromolecule from a PLGA microparticle has been characterized by a rapid release phase (burst effect) followed by a slow release phase (Alonso et al., Vaccine, 12:299-306 (1994)). For tetanus vaccine formulations, the initial burst tends to be extremely high (more than 50% in 24 hours), due to the closeness of the macromolecule to the microparticle surface. Id. The second phase is characterized by the release of the macromolecule through water-filled pores by diffusion, polymer degradation, and the erosion of the polymer matrix. Id.

In an aqueous environment, microparticle systems formulated from PLGA copolymers undergo a hydration process followed by bulk erosion. During erosion, matrix porosity increases and facilitates macromolecule release by diffusion. Conventional microparticles formulated from PLGA having a molecular weight from about 30,000 to about 75,000 and exhibit a lag time of several days after the initial burst. During that lag time, the release rate of the macromolecule from the microparticle is minimal. The release rate does not increase until the erosion of the polymer matrix is sufficient to release the remaining molecules from the microparticle (Sanders, L.M., et al., J. Pharm. Sci. , 75:356-360 (1986)).

Brief Summary of the Invention

In one aspect, the present invention provides a new microparticle based formulation that can be used as a drug delivery system. The microparticle of this invention is a blend of a biodegradable polymer and a water soluble polymer. The microparticle also contains a therapeutic compound.

The microparticle based formulation of the present invention contains from about 50 to about 99 weight percentage of a biodegradable polymer, from about 0.10 to about 50 weight percentage of a water soluble polymer, and from about 0.10 to about 10 weight percentage of a therapeutic compound. The molecular weight of the biodegradable polymer is from about 2,000 to 100,000, preferably from 30,000 to about 75,000 and the molecular weight of the water soluble polymer is from about 200 to about 100,000. The preferred biodegradable polymer is a poly(D,L lactic-co-glycolic acid) copolymer and the preferred water soluble polymer is poly(ethylene glycol). The preferred therapeutic compound is a smooth muscle cell inhibitor, particularly antisense oligodeoxynucleotides to tenascin mRNA and antibodies for basic fibroblast growth factor.

In the second aspect, the present invention contemplates a process of inhibiting SMC migration and proliferation. The method involves preparing a microparticle based formulation containing a therapeutic compound that inhibits SMC migration and proliferation, and delivering said microparticle to an area that contains SMCs to inhibit the proliferation and migration of said SMCs.

Description of the Drawings

Figure 1 shows the cumulative mass of antisense oligodeoxynucleotide (AS-ODN) released from PLGA 50:50 microparticles into water as a function of time.

Figure 2 shows the dose-dependent growth inhibition of SMCs after administration of AS-ODN loaded, scrambled oligodeoxynucleotide (SC- ODN) loaded, and blank PLGA 50:50 microparticles over 72 hours (T₇₂).

Figure 3 shows the dose-dependent growth inhibition of SMCs after administration of basic fibroblast growth factor (bFGF) A12 antibody loaded, mouse immunoglobulin (MolgG) loaded (nonspecific control antibody), and blank PLGA 50:50 microparticles over 96 hours.

Figure 4 shows the cumulative mass of antibody for IgG released from PLGA 50:50 microparticles containing various amounts of polyethylene oxide (PEO).

Detailed Description of the Invention

The present invention contemplates a new formulation scheme for a microparticle that can be used as a drug delivery system. The microparticle of this invention is made from a blend of biodegradable polymer and a water soluble polymer. The microparticle also contains a therapeutic compound to be delivered. The blend of the biodegradable polymer and water soluble synthetic polymer improves the release rate of the therapeutic compound from the microparticle. Additionally, the present invention contemplates a process for inhibiting smooth muscle cell migration and proliferation using the microparticle of this invention.

Any biodegradable polymer can be used in the microparticle of this invention. As used herein, the term "polymer" encompasses any type of polymer such as homopolymers, copolymers, terpolymers, and the like. Biodegradable homopolymers that can be used in the formulation include poly(glycolic acid), poly(D-lactic acid), poly(L-lactic acid), poly(D,L-lactic acid), poly(D,L-ethyl glycolic acid), poly(dimethyl glycolic acid), poly(D,L- methyl ethyl glycolic acid) and poly(ε-caprolactone). A biodegradable copolymer that can be used is poly(D,L-lactic-co-glycolic acid) (PLGA). Biodegradable polymers of any molecular weight can be used in the formulation, although it is preferred that the biodegradable polymer have a molecular weight ranging from about 2,000 to about 100,000, most preferably from about 30,000 to about 75,000. The amount of biodegradable polymer in the microparticle is from about 50 to about 99 weight percent, preferably from about 80 to 90 weight percent and most preferably from about 95 to about 97 weight percent.

The preferred biodegradable polymer contemplated for use in the microparticle formulation is a PLGA copolymer. It is further preferred that the PLGA copolymer contain a 50:50 ratio of lactic to glycolic acid. However, PLGA copolymers containing different ratios of lactic to glycolic acid can be used in the formulation as well.

The microparticle also contains a water soluble polymer. This water soluble polymer may be a synthetic polymer. Any water soluble synthetic polymer can be used in the formulation of this invention. Water soluble synthetic polymers that can be used include poly(ethylene glycol) (PEG), poly(vinyl methyl ether), poly(acrylic acid), poly(methacrylic acid), poly(2-hydroxyethyl methacrylate), and poly(N-vinyl pyrrolidone), poly(vinyl alcohol), and the like. The water soluble polymer may also be a natural polymer such as collagen, gelatin, polyamino acids, polysacaharides and the like. The molecular weight of the water soluble polymer may be from about 200 to about 100,000. The preferred water soluble polymer for use in the formulation is PEG. The most preferred water soluble polymer is poly(ethylene oxide) (PEO), which is a high molecular weight ( > 50,000) PEG molecule. The amount of water soluble polymer in the microparticle is from about 0.1 0 to about 50 weight percent, preferably from about 0.50 to about 10 weight percent and most preferably from about 1 to about 5 weight percent. The presence of the water soluble polymer in the formulation is critical for improving the release rate of a therapeutic compound from the microparticle. It is believed that the water soluble polymer forms channels in the microparticle structure that facilitates drug delivery.

The microparticle of this invention also contains a therapeutic compound to be delivered. The amount of therapeutic compound contained in the microparticle is from about 0.10 to about 10 weight percent, preferably from about 1 to about 2 weight percent depending upon the compound employed and its intended use. Any therapeutic compound can be loaded into the microparticle for drug delivery purposes such as proteins, oligonucleotides, antibodies, nucleic acids, glucocorticoids and small molecular weight drugs. As used herein, the term "small molecular weight drugs" means therapeutic drugs having a molecular weight of 500 or less. Examples of small molecular weight drugs include: acetylsalicylic acid, diazepam, ibuprofen and nitroglycerin. The preferred therapeutic compounds are those compounds that inhibit the proliferation and migration of SMCs. A first preferred class of therapeutic compounds for inhibiting smooth muscle cell proliferation and migration are antisense oligodeoxynucleotides. The preferred antisense oligodeoxynucleotides are against tenascin mRNA. Microparticles of this invention loaded with antisense oligodeoxynucleotides against tenascin mRNA inhibit smooth muscle cell proliferation and migration in vitro. The second preferred therapeutic compound for inhibiting SMC proliferation and migration is antibodies against basic fibroblast growth factor (bFGF). Microparticles of this invention loaded with antibodies against bFGF also inhibit smooth muscle cell proliferation and migration in vitro. Microparticles of this invention loaded with either antisense oligodeoxynucleotides to tenascin mRNA or antibodies against bFGF exhibit a small burst effect during the first 24 hours in vitro. This burst effect is small compared to other polymer systems which release most of the loaded drug during the first 24 hours of solvent contact. The relatively small burst effect of the invention herein may be advantageous, since it prevents sudden exposure of the cells to a potentially toxic dose of the therapeutic compound. Additionally, these loaded microparticles do not demonstrate a lag period after the initial burst. Release of protein from conventional microparticles made of high molecular weight PLGA (from about 2,000 to about 100,000) exhibit a lag time after the initial release burst. In contrast, the microparticles of this invention do not exhibit a lag period, but instead directly enter into a biphasic release pattern. This result deviates from previously reported conventional microparticle release kinetics as described in Yan C, et al., J. Controlled Rel. , 32:231 -241 (1 994) and Iwata, M„ et al., Pharm. Res. , 1 0: 1 21 9- 1 227 (1 993) . Additionally, the microparticles of this invention demonstrate controlled release of the oligodeoxynucleotides or antibodies for up to 20 days in vitro.

The microparticle of this invention can be made using any process known in the art such as the solvent evaporation or solvent extraction technique as described in Alonso et at., Vaccine, 1 2:299-306 (1 994). The preferred method for making the microparticle of this invention involves a modification of the solvent-extraction technique described in Alonso et al. The method involves mixing the requisite amounts of a biodegradable polymer and water soluble polymer in a solvent, such as dichloromethane. This mixture is then placed into a test tube, such as a flint glass test tube. Next, a therapeutic compound is dissolved in water and placed into the test tube with the biodegradable polymer and water soluble synthetic polymer mixture. The contents of the test tube are then emulsified with a vortexer. Blanks may then be prepared using water. If blanks are prepared, the solution is then reemulsified with an emulsifier, such as polyvinyl acetate, to form a double emulsion. Next, an alcohol, such as isopropanol, is added and the solution maintained on a magnetic stirrer to extract the solvent to the external alcoholic phase and precipitate the dissolved polymer to form microparticles. The system is stirred for at least one hour to assure total extraction of the solvent. The microparticles are then sieved to form particles having an appropriate size for drug delivery purposes. The microparticles of this invention have a size smaller than 100 micrometers regardless of their method of formulation. Preferably, the microparticle have a diameter of from about 1 to about 100 micrometers, most preferably from about 10 to about 80 micrometers.

The present invention also contemplates a process for inhibiting SMC migration and proliferation. The first step involves preparing a microparticle having the formulation of this invention. The therapeutic compound loaded into the microparticle is a compound that inhibits the proliferation and migration of SMCs. The second step involves delivering said microparticle to an area known to contain SMCs. Once the microparticle is delivered to this area, the therapeutic compound is released from the microparticle in a controlled fashion and inhibits the proliferation and migration of SMCs. The microparticle of this invention may be delivered to an area containing SMCs in several ways. For example, if the microparticle is to be delivered to smooth muscle cells in vitro, then the microparticle may be delivered by placing the microparticle in cell culture medium. If the microparticle is to be delivered in vivo, such as to a patient suffering form atherosclerosis, then the microparticle may be delivered either adventitially or directly inside an artery by an injection catheter.

The following Examples illustrate particular embodiments of the present invention and are not limiting of the specification and claims in any way. EXAMPLES

EXAMPLE 1 : Formulation of Microparticles Loaded with Fluorescein

Isothiocyanate-Labeled Dextran and Fluorescein Polymers, Reagents and Drugs

Polymers and Reagents

PLGA of 50:50 copolymer ratio of lactic to glycolic acid was supplied by Medisorb. The polymer weight average molecular weight (Mw) was measured by gel permeation chromatography as 42,000. PEG with nominal Mw of 4,600 (Milwaukee, Wl) and PVA, 88% mole hydrolyzed, with a Mw of 25,000 were purchased from Aldrich. The two model drugs were utilized were fluorescein isothiocyanate-labeled dextran (FITC-dextran) of Mw = 1 9,600 (Sigma Chemical, St. Louis, MO) and FITC conjugated rabbit gamma immunoglobulin (FITC-lgG) of Mw = 1 50,000 (Sigma) . Bicinchoninic acid (BCA) protein assay kit to measure the FITC- IgG concentration in an aqueous solution was purchased from Pierce Chemical (Rockford, IL).

Fabrication of Microparticles

Microparticles were manufactured by a modification of a double- emulsion-solvent-extraction technique. 2.5 mg of model drug compound was dissolved in 1 25 μl water and injected into a flint glass test tube containing 247.5 mg total mixture of PLGA and PEG dissolved in 1 ml dichloromethane and emulsified with a vortexer. The solution was reemulsified in 100 ml of 0.3% aqueous PVA resulting in a double emulsion that was poured into 100 ml of 2% aqueous isopropanol maintained on a magnetic stirrer. The extraction of the dichloromethane to the external alcoholic phase precipitated the dissolved polymer which in turn resulted in formation of microparticles. The system was stirred for 1 hour to assure total extraction of the solvent. The formed microparticles were finally sieved to sizes smaller than 1 25 μm, rinsed in water, and collected by centrifugation. Several different formulations were prepared using this method by varying the PLG A/PEG ratio.

Size distribution of the microparticles was measured with a Coulter

Multisizer (Coulter Electronics, Hialeah, FL) after microparticles were redispersed in Isoton II (Coulter Electronics). This analysis was carried out with a 200 μm aperture orifice tube and the results were reported as mean microparticle diameter.

The entrapment efficiency of the model drug was determined by comparing the amount of starting FITC-labeled compound with the quantity actually entrapped. Approximately 10 mg of microparticles were added to 1 ml dichloromethane and allowed to dissolve over 6 hours. The

FITC-labeled compound was extracted by adding 1 ml water and shaken every 6 hours over 24 hours. The concentration of FITC-labeled compound in the aqueous phase was determined by absorption at 496 nm in a UV-VIS Beckman spectrophotometer (Fullerton, CA).

Nuclear Magnetic Resonance (NMR)

NMR spectroscopy was utilized to determine the percentage of PEG remaining in the microparticles after preparation. Microparticles containing 1 , 2, and 5 weight percentage of PEG and weighing at least 10 mg were placed into glass NMR sample tubes with 1 ml of deuterated chloroform solution (Acros Organics, NJ) containing 1 % v/v tetramethylsilane. The proton NMR spectra were acquired on a NMR spectrometer (Bruker AC250, Germany). In the spectrum of PLGA/PEG blends, the hydrogens of the methyl group of lactic acid resonated at 5.2 ppm while the hydrogens of the glycolic acid appeared as a sharp peak at 4.8 ppm, and the hydrogens of the homopolymer PEG appeared as a sharp peak at 3.6 ppm. A comparison of the relative integration values of the peaks indicated the composition of the blend in terms of mole fractions. The residual PEG weight percent was calculated by multiplying each mole fraction by its respective repeating unit molecular weight and dividing the value obtained for PEG by that for PLGA plus PEG times one hundred.

Differential Scanning Calorimetrv (DSC)

A differential scanning calorimeter (TA Instruments, Series 2920, New Castle, DE) equipped with a mechanical cooling accessory was utilized to evaluate the compatibility between the blended polymers. Samples (5 mg) were equilibrated to 80°C, cooled at 1 °C/min to -20°C, and heated at a ramp rate of 10°C/min. to 80°C. The slow cooling period allows the formation of PEG crystallites from any regions in which PEG has phase separated from PLGA within the microparticles.

Gel Permeation Chromatography

Microparticles containing 0, 1 , 2, and 5 weight percent PEG of approximately 30 mg were placed into 1 ml water in 1 .5 ml microvials and maintained at 37°C. Every 7 days the samples were decanted off and fresh water was added. At 9, 18, and 27 days samples of each formulation were collected, frozen and vacuum-dried for 24 hours.

The PLGA molecular weight distribution was determined by gel permeation chromatography (Waters, Milford, MA) equipped with a differential refractor (Waters, Series 410). The samples were dissolved in chloroform and eluted in a series configuration through a Phenogel guard column (model 494386, 50x7.8mm, 5 μm particle diameter, Phenomenex, Torrance, CA) and a Phenogel column (linear 0 - 100,000 mixed bed, 7.8x300 mm, 5 μm particle diameter) at a flow rate of 1 ml/min. The molecular weight distribution curve representing the PLGA was selected on the chromatogram and the molecular weight was determined relative to polystyrene standards (Tosoh, Tokyo, Japan). Scanning Electron Microscopy (SEM)

The morphology of degrading microparticles at 0, 3, and 9 days was observed by scanning electron microscopy (Model JSM-5300, JEOL, Tokyo, Japan) at 25 kV. Microparticles (0.025 g) containing 1 % FITC- dextran or FITC-lgG and 0 or 5 weight percent of PEG, were placed into 1 ml water in 1.5 ml microvials and maintained at 37°C. Every 3 days the samples were decanted off and fresh water was added. At 3 and 9 days microparticle samples were collected, frozen and vacuum-dried for 24 hours. Before viewing on SEM, samples were freeze-dried, mounted on glass coverslips with nickel print (GC Electronics, Rockford, IL), and coated with a Au-Pd film of 300-600 A thickness.

In Vitro Release Studies

A sample of 25 ( ± 0.09) mg FITC-lgG and 20 ( ±0.04) mg FITC- dextran loaded microparticles were weighed, placed into 1 ml water in 1 .5 ml microvials, and maintained at 37°C on a shaker table (60 rpm). At appropriate intervals, after centrifugation, the water was collected and replaced. The removed water was filtered and stored at 4°C until analyzed. The FITC-dextran concentration was measured at 490 nm using a UV-VIS Beckman spectrophotometer. FITC-lgG concentration was measured using the microplate BCA assay at 470 nm.

Statistics

All samples were run in triplicate, except for the determination of PEG weight percentage from NMR spectroscopy (n = 1 ), and expressed as mean ± standard deviation (SD). Different groups were compared at a given level using the unpaired two group T-test. Single-factor analysis of variance (ANOVA) was employed to assess statistical significance of the PLGA 50:50 half-lives of degradation studies. A significance level of 0.05 was used in all the statistical tests performed. Initial Microparticle Characterization

SEM revealed that smooth spherical microparticles were produced from all formulations whether or not PEG was included. The concentration of the polymer organic mixture was chosen so as to fix the solution viscosity in the range where microparticles slightly less than 100 μm in diameter would be produced. The particle size of FITC-loaded microparticles decreased with increasing PEG content for 1 , 2, and 5 weight percent samples as determined by single factor ANOVA (p<0.05), as shown in Table 1 below.

Table 1

Entrapment Efficiency and Diameter of FITC-dextran and FITC-lgG Loaded Microparticles

Initial PEG FITC-dextran FITC-lgG

Weight Percent Entrapment Diameter Entrapment Diameter

(wt%) Efficiency, %* m Efficiency, %* m

0 77±2 81.5±4 85±4 77.6±4

1 76 ±3 87.2±4 83±3 82.2 ±3

2 64±1 83.8±1 87 ±2 77.0±3

5 67 ±2 77.0±3 92±3 64.0 ±4

*Theoretical loading was 1 wt%.

The microparticle size was also dependent on the model drug used. FITC- lgG loaded microparticles containing 2 and 5 weight percent of PEG were smaller than their FITC-dextran counterparts (p<0.05). Differences in loading were also noted as shown above in Table 1. The entrapment efficiency of FITC-dextran loaded microparticles decreased with increasing PEG content (p<0.05). The FITC-lgG loaded microparticles entrapment efficiency did not depend on PEG content as determined by single factor ANOVA (p>0.05). NMR studies showed that a substantial fraction of the PEG originally dissolved in the organic phase of the emulsion was extracted into the aqueous phase and not incorporated into the final microparticle as shown below in Table 2.

TABLE 2 Residual PEG in Prepared Microparticles

Initial PEG Final PEG Percent of Initial

Weight Percent Weight Percent PEG Remaining

(wt%) (wt%) After Fabrication*

1 0.49 49

2 0.65 33

5 1 .49 30

* final PEG wt% divided by initial PEG wt%.

Although, the weight percent of PEG incorporated into the final microparticle and the corresponding percentage remaining were a function of the theoretical PEG weight percent. Microparticles manufactured with 1 , 2, and 5 weight percent PEG exhibited 0.49, 0.65, and 1 .49 weight percent PEG in the final microparticles, and the corresponding percentage of original PEG incorporated was 49, 33, and 30 weight percentage. Clearly, this demonstrates that the higher the theoretical weight percentage PEG the lower the percentage of remaining PEG in the final microparticle.

Microparticles were tested by DSC for evidence of any phase separation by the appearance of PEG crystallites following heat treatment. Two polymers with different structures and molecular weights may not necessarily form homogenous blends, because partial phase separation can occur during precipitation from solution. The PEG utilized has a higher affinity fro the organic/aqueous interface and even for the aqueous phase than the PLGA, so it is possible that PEG could migrate preferentially to the microparticle surface and precipitate. Moreover, the higher molecular weight PLGA might precipitate earlier within the organic globule due to a lower solubility. However, no PEG crystallites were noted attesting to the homogeneity of the blends.

In Vitro Degradation of Microparticles The decrease in the weight average molecular weight of the PLGA forming microparticles containing 0, 1 , 2, and 5 weight percent PEG occurs during incubation over a 27 day period. A reduction in the initial weight average molecular weight of approximately 10 fold was observed after 27 days of degradation for all samples. The half-life of PLGA for each microparticle formulation was calculated assuming that the degradation is random and is represented by an exponential decay. The half-life for PLGA 50:50 in microparticles containing 0, 1 , 2, and 5 weight percent PEG was calculated to be 7.5 ± 0.3, 8.2 ± 0.4, 9.2 ± 0.7, and 9.0 ± 0.8 days, respectively. The PEG content had a significant effect on the PLGA half-life (p < 0.05).

The microparticles of this invention were prepared by a double emulsion technique which is known to produce internal pores. These internal pores, however, are not interconnected because microparticles prepared by standard double emulsion solvent extraction techniques produce isolated pores. The use of PEG may produce microparticles in which the internal pores created by the initial aqueous phase are interconnected. Aqueous fluid-filled pores that are interconnected within the microparticle would eliminate degradation products more efficiently, and diminish the autocatalytic effect in polymer degradation effectively extending the half-life. Microparticles formed without PEG and thus with no interconnected pores exhibited a reduced half-life because of the inability of such microparticles to eliminate degradation products. In microparticles with isolated pores, the degradation process was heterogeneous and proceeded more rapidly in the center than at the surface. SEM of IgG-FITC loaded PLGA 50:50 microparticles with 0 and 5 weight percent PEG incubated for 0, 3, and 9 days showed the shape and integrity of the microparticles were maintained throughout the 9 day study, but significant morphological changes were exhibited as early as day 3. The microparticles at day 0 exhibited a smooth surface and a spherical shape while at day 3 surface pores became apparent and by day 9 the pores comprised a large percentage of the surface area revealing the internal structure of the microparticles for both formulations. Notably, microparticles containing 5 weight percent PEG exhibited a similar structure to 0 weight percent PEG. This suggests that the changes resulting from the addition of PEG are molecular and not macroscopic as the morphology of the different formulations appears the same.

In Vitro Release Studies

The cumulative mass and normalized mass release profiles for both compounds examined in this study depended on the PEG content of the microparticles. Microparticles prepared with FITC-lgG showed an initial burst from 3.4 to 57.5% over the first 24 hours, while microparticles prepared with FITC-dextran showed a generally smaller burst between

16.2 and 27.4% over the first 24 hours, for PEG content of 0 and 5 weight percentage as shown below in Table 3.

Table 3 Burst Effect of FITC-dextran and FITC-lgG For The First 24 Hours

Initial PEG Percent of Loading Percent of Loading

Weight Percent FITC-dextran FITC-lgG

(wt%)

0 16.2± 1.9 3.4 ± 1.2

1 16.5 ± 0.4 15.5 ± 1.3*

2 27.4 ± 5.7* 54.0 ± 8.0*

5 27.1 ± 3.1 57.5 ± 3.8

* Statistically different from value corresponding to immediately smaller initial PEG content. The larger initial burst of FITC-lgG was a nominal effect. Much of the FITC-dextran was removed in the manufacturing process, as evidenced by lower loading (see Table 1 ), and consequently less was available to be released during the first 24 hours of the in vitro release studies. In addition, since the molecular weight of FITC-dextran (19,000) is much smaller than that of FITC-lgG (150,000), the FITC-dextran diffuses more rapidly out of molecular pores created by PEG.

Following the initial burst over the first 24 hours, FITC-dextran loaded microparticles displayed a 12 day lag phase. A single factor ANOVA test (p <0.05) showed that the rates of release over this time interval for 0, 1 , 2, and 5 weight percent of PEG, as shown below in Table 4, were significantly different.

Table 4

Release Rates For FITC-lgG And FITC-Dextran From 1 wt% Loaded Microparticles of Varying Initial PEG Content Calculated Using

Linear Regression.

Initial PEG Weight Percent (wt%)

Release Rate 0 1 2 5

FITC-lgG 0.021 0.068 0.132 0.214

(μg/day/mg microparticles) ±0.001 ±0.01 1 ±0.030 ±0.020

Days 1 -28

FITC-dextran 0.082 0.098 0.071 0.089

(μg/day/mg microparticles) ± 0.018 ±0.017 ±0.006 ± 0.012

Days 1 -12

FITC-dextran 0.289 0.346 0.604 0.656

(μg/day/mg microparticles) ± 0.025 ± 0.034 ±0.123 ±0.192

Days 12-28*

* Linear regression for 2 and 5 wt% PEG formulation was performed up to 22 days at which time at least 95% of FITC-dextran was released. Following the initial lag phase, FITC-dextran loaded microparticles demonstrated a later release phase which occurred at a greater rate (p < 0.05) for all equivalent PEG formulations until they became depleted. The rates of release during this later phase were determined to be significantly different (p <0.05) as determined by a single factor ANOVA. Microparticles containing FITC-lgG showed no lag phase over the time interval investigated. One of the most significant observations is that the release profiles appear to be linear after the initial burst for microparticles containing PEG and FITC-lgG. The correlation coefficients for all linear regressions were greater than 0.98. Furthermore, the rate constants significantly increased with increasing percentage of PEG (p <0.05). Notably, the release rate of FITC-dextran during the initial rate was significantly higher than that of the FITC-lgG for microparticles containing 0 and 1 weight percentage PEG, whereas the release rate of 2 and 5 weight percentage formulations was lower (p <0.05). In comparing later phases of release, FITC-dextran was released at a much greater rate for all formulations (p<0.05).

The entrapped model drugs were released from microparticles containing different amounts of PEG at different rates. The higher the PEG weight percent, the greater the rate of release. These observations are as expected and in keeping with the hypothesis of PEG forming molecular pores for drug release. Moreover, microparticles prepared with higher percentages of PEG have more molecular pores created during the manufacturing process that contribute to the initial release rate. In addition, microparticles made with higher PEG contents have more PEG remaining after manufacturing as shown in Table 1 to subsequently form more molecular pores and channels for release at later times. The presence of PEG renders the release mechanism diffusion controlled rather than degradation controlled because the PEG content has an increasing effect on PLGA half-life. EXAMPLE 2: Formulation of Microparticles Loaded with Antisense

Oligodeoxynucleotides and their Use in Inhibiting the Migration and Proliferation of Smooth Muscle Cells In Vitro.

Polymers and Reagents

PLGA of 50:50 copolymer ratio of lactic to glycolic acid was purchased from Medisorb (Cincinnati, OH). The polymer weight average molecular weight (Mw) of 45,000 was measured by gel permeation chromatography. PEO with a nominal Mw of 4,600 was purchased from Aldrich (Milwaukee, Wl). Poly(vinyl alcohol) (PVA), 88% mole hydrolyzed, with a Mw of 25,000 was also supplied by Aldrich. All other reagents were of analytical grade. Distilled deionized water was utilized in all studies.

Oligodeoxynucleotides Phosphorothioated ODNs were purchased from Oligos Etc., Inc.

(Wilsonville, OR) and were shipped as lyophilized powder. The ODN were resuspended in Tris-EDTA buffer ( 10 mM Tris-CI, l/mM EDTA, pH 8.0). The AS-ODN sequence (5'ACC ATG GGG GCC GTG ACC TGG CTA'3) was complementary to the translation initiation start site of the rat tenascin mRNA and inhibited SMC proliferation (see Denner, L.A., et al., Circulation 90: 1 592 (1994)). The scrambled (SO sequence (5" ATC AGC TCT GTC AGC GCG CCA GCG'3) was chosen to contain the same base composition the AS-ODN sequence but in a random order, and screened through Genbank (Genetics Computer Group, Inc., Madison, Wl) to ensure that the SC-ODN did not match rat mRNA for tenascin or other sequences.

Fabrication of ODN Loaded PLGA Microparticles

Microparticles were manufactured by a modification of a double- emulsion-solvent-extraction technique as described in Alonso et al., Vaccine, 12:299-306, (1 994). 0.0025 g of ODN were dissolved in 50 μl water and injected into a flint glass test tube containing 0.245g PLGA and 0.0025 g PEO dissolved in 1 ml dichloromethane and emulsified with a vortexer. Blanks were prepared using 50 μl water. The solution was reemulsified in 100 ml of 0.3% aqueous PVA resulting in a double emulsion that was poured into 100 ml of 2% aqueous isopropanol maintained on a magnetic stirrer. The extraction of the dichloromethane to the external alcoholic phase precipitated the dissolved polymer which in turn resulted in formation of microparticles. The system was stirred for 1 hour to assure total extraction of the solvent. The formed microparticles were finally sieved to sizes smaller than 100 micrometers, rinsed in water, and collected by centrifugation. AH microparticles were sterilized with ethylene oxide before use.

The entrapment efficiency was determined by comparing the amount of starting ODN with the quantity actually entrapped. Approximately 5 mg of microparticles were added to 1 .5 ml dichloromethane and allowed to dissolve over 6 hours. The ODN was extracted by adding 1 ml water, periodically shaking over 24 hours, and analyzed at 260 nm in UV-VIS Beckman (Fullerton, CA) spectrophotometer.

ODN Release

A sample of 20 ( ±0.05) mg PLGA 50:50 microparticles was placed into 1 ml water in 1 .5 ml microvials, and maintained at 37°C on a shaker table (60 rpm). The water was collected and replaced at 6, 12, and 24 hours, daily thereafter until day 8, then every two days until day 20. The removed water was filtered and stored at 4°C until spectrophotometrically analyzed at 260 nm. SMC Culture

Vascular SMCs were harvested from the carotid arteries of Sprague-Dawley rats by enzymatic digestion (Chamley-Campbell, J.H., et al., J. Cell Bio I. , 89:379-383 (1 981 )). Primary cells were cultured in Dulbecco's Modified Eagle's Media (DMEM) supplemented with 20% fetal bovine serum (FBS), 1 % glutamine, 100 U/ml penicillin, 1 00 mcg/ml streptomycin, and 50 mcg/ml neomycin. Cultures were maintained in a humidified incubator at 37 °C with 5% CO₂. All cells beyond the first passage were cultured in DMEM with 10% FBS, 1 % glutamine, 100 U/ml penicillin, 100 mcg/ml streptomycin, and 50 mcg/ml neomycin.

SMC Proliferation

Second and third passage SMCs were seeded into 24-well cluster plates (Falcon, Lincoln Park, NJ) at a density of 30,000 cells per well (corresponding to 14,900 cells/cm²). Twenty-four hours after plating the original medium was replaced with growth arrest medium (DMEM containing 0.1 % FBS). After 72 hours the arrest medium was replaced by 1 ml of medium containing 10% FBS. At that time, PLGA 50:50 microparticles with or without ODN were also placed into the appropriate wells. After an additional 72 hours, the cells were photographed, trypsinized, and counted on a Coulter Counter (Coulter, model Z_F. Hialeah, FL) . To achieve different dosages, 4 ( ± 0.05), 8 ( ± 0.05), 1 2 ( ± 0.02), 1 6 ( ± 0.04) and 20 ( ± 0.07) mg of microparticles were used (n = 6). Control wells contained DMEM/FBS alone.

SMC Migration Third and fourth passage SMCs were seeded at 60,000 cells² per well into 3/1 6 inch diameter Teflon fences (corresponding to 336,000 cells/cm) in 6-weII cluster plates, and allowed to attach in 10% FBS containing DMEM. After 24 hours, the Teflon fence was removed, the culture was rinsed 3 times with phosphate buffered saline (PBS) to remove any unattached cells, and 2 ml of fresh media (10% FBS) and 32 mg of microparticles were added to wells. SMCs were allowed to migrate over the surface. After 72 hours, the cultures were rinsed with PBS, fixed with 10% neutral buffered formalin, and stained with toluidine blue- 0 (1 % w/v). Morphometric analysis was performed to determine the area covered by the migrating and proliferating SMCs. Digitized images of the stained cultures were taken using a JVC Tk- 10700 color video camera attached to the photographic port of an Askimina SMC4 (Jena, Micro- Tech Instruments, Dallas, TX) microscope and interfaced to a computer. The culture area occupied by SMCs was traced and calculated by calibrating the software (NIH Image 1 .55) with a known standard.

Statistical Analysis

All samples were run in triplicate, except for dose response samples and control group in migration study (n = 6), and expressed as mean ± standard deviation (SD). Single-factor analysis of variance (ANOVA) was employed to assess the statistical significance of dose response data. If ANOVA was found to be significant by the global F-test, Scheffe's test was performed to compare pairs of sample sets. Different groups were compared at a given dosage using the unpaired two group T-test. A significance level 0.05 was used in all of the statistical tests performed.

Results ODN Entrapment Efficiency and Release Kinetics

The entrapment efficiency of ODN loaded microparticles was 74 ( ± 4)% for AS-ODN (n = 6) and 77 ( ± 5)% for SC-ODN (n = 3) . Therefore, the actual ODN loading of the microparticles was 0.74 and 0.77 weight percent for AS-ODN and SC-ODN, respectively. The cumulative mass of AS-ODN released from 20 mg samples of PLGA microparticles exhibited a small burst effect at day 1 (1 7% of loaded drug) followed by a sigmoidal AS-ODN release profile with approximate release rates of 0.36 ( ± 0.02) and 0.09 ( ± 0.01 ) μg/day/mg PLGA for days 1 -5 and 5-20, respectively. After 20 days, 55% of total loaded AS-ODN was released. Release kinetics for SC-ODN were similar to AS-ODN. Figure 1 shows the cumulative mass of AS-ODN released from the microparticles (with 0.74 weight % AS-ODN loading) into water at 37°C as a function of time. The error bars represent mean ± standard deviation (SD) for n = 3.

ODN Effects on SMC Proliferation

AS-ODN released from PLGA microparticles inhibited SMC proliferation in a dose dependent manner. Figure 2 shows the dose- dependent growth inhibition of SMCs after administration of AS-ODN loaded, scrambled oligodeoxynucleotide (SC-ODN) loaded, and blank PLGA 50:50 microparticles over 72 hours (T₇₂). Error bars represent mean ± standard deviation for n = 6. Statistical significance between microparticles containing AS-ODN and SC-ODN at a given dosage is indicated by (*), whereas ( §) designates statistical significance between values containing either AS-ODN or SC-ODN and blank microparticles with no ODN (p< 0.05). T₀ represents attached cell number 24 hours after plating.

SMCs exposed to 4, 8, 1 2, 1 6 and 20 mg of AS-ODN loaded microparticles exhibited decreases of 1 2, 36, 43, 61 , 75%, respectively, relative to control wells which contained no microparticles. As shown in Figure 2, for 20 mg of PLGA microparticles loaded with AS-ODN, the final average cell count was not significantly lower than the attached cell number 24 hours after plating. A single factor ANOVA test showed the dose responsiveness to be statistically significant (p < 0.05) for 4, 8, 12, 1 6, and 20 mg dosages. The data were further analyzed with Scheff's F- test to determine statistical significance of dosages from one another. All other dosages, except the 8 mg compared to the 1 2 mg and 1 6 compared to 20 mg dose, were statistically significant from one another. SC-ODN loaded microparticles also inhibited SMC proliferation in a dose dependent fashion determined by a single factor ANOVA involving al dosages (p<0.5). However, the inhibition was less than that seen in Figure 2, PLGA microparticles without ODN had no effect on SMC proliferation. The PLGA microparticles remained suspended in the medium, with no apparent cell attachment; they did not seem to interfere with the cultured SMCs.

ODN Effects on SMC Migration

A migration front was obtained by seeding SMCs into Teflon fences. Following removal of the fences 24 hours after seeding, SMCs migrated radially and concurrently proliferated resulting in an increased surface area. Wells containing AS-ODN loaded PLGA microparticles showed smaller increases in cell surface area (see Table 5 below) 72 hours after fence removal <p < 0.05), compared to wells containing no microparticles, SC-ODN loaded microparticles, or blank microparticles (0% ODN). The area covered by migrating SMCs in wells with AS-ODN loaded microparticles was 22% of the corresponding area in control wells with no microparticles. Moreover, the measured areas of wells with SC-ODN loaded microparticles and blank microparticles showed no significant difference form those for control wells containing only SMCs.

TABLE 5

SMC culture Fractional Increase % Area Covered by Number of Culture Area* Migrating SMCs experiments Relative to Control

Control 1.64 ± .019 100 6

Blank 1 .81 ±0.06 127 3

SC-ODN¹ 1 .70 ± 0.08 109 3

AS-ODN¹ 1 .14 ±0.06* § t 22 3 t Mean ± SD

1 Statistical significance between microparticles containing AS-ODN and SC-ODN is indicated by (*), AS-ODN and blank by (§), and AS-ODN and control by (t)(p < 0.05). Summary

These results demonstrate that a 24-mer AS-ODN to the rat tenascin mRNA can be successfully incorporated into PLGA microparticles for controlled release and that AS-ODN delivered by such microparticles can inhibit SMC proliferation and migration in vitro.

AS-ODN release results demonstrated a relatively small burst effect during the first 24 hours. However, unlike conventional microparticles made of PLGA having a high molecular weight ranging from about 30,000 to about 75,000, the AS-ODN microparticles did not show a lag period after the initial burst of drug and directly entered into a biphasic release pattern.

The delivery rate of the AS-ODN microparticle system, on a per unit polymer mass basis, was 0.09 to 36 μg/day/mg PLGA. This was significantly higher than the delivery rate of the ethylene vinyl acetate matrices containing heparin, which delivered heparin at a release rate of 0.001 2 to 0.0024 μg/day/mg EVAc (See Edelman, E.R., et al., PNAS, 87:37673-3777 (1 990)). The release rate for the EVAc was normalized by ethylene vinyl acetate weight to allow comparison on a per mass bases to PLGA microparticles. The SC-ODN microparticles served as controls to examine AS-ODN sequence specific effects on smooth muscle cell proliferation and migration. The SC-ODN has the same base composition the AS-ODN, but in a randomized sequence such that it could not hybridize with any sequence contained in Genbank, including the tenascin sequence. It is believed that the SC-ODN may function as the AS-ODN, through non- antisense mechanisms to inhibit smooth muscle cell proliferation (See Burgess, T.L., et al., PNAS, 92:4051 -4055 (1 995)). These include binding to specific target proteins (See Beltinger, C, et al., J. Clin. Invest. , 95:1814-1823 (1 995); Stein, C.A., et al., Antisense Res. Devei , 3: 1 9-31 (1 993); and Gao, W.Y., et al., Mol. Pharmacol., 41 :223-229 (1 992)) and secondary effects of nucleosides and nucleotides produced from degradation of the ODN by nucleases (See Kamano, et al., Biochem. Int., 26:527-543 (1992); and Rathbone, M.P., et al., In Vitro Cell Develop. Biol., 28A, 529-536 (1992)).

EXAMPLE 3: Formulation of Microparticles Loaded with an Antibody to Basic Fibroblast Growth Factor and Their Use in Inhibiting the Migration and Proliferation of Smooth Muscle Cells In Vitro.

Polymers and Reagents

Same as in Example 2.

Fabrication of Microparticles Same as in Example 2.

Antibody to Fibroblast Growth Factor

The basic fibroblast growth factor antibody can be made using standard procedures well known in the art. Briefly, BALB/c mice were immunized by intraperitoneal and subcutaneous injections of human recombinant bFGF, which was emulsified with an equal volume of complete Freund's adjuvant. As measured by the ELISA technique, all mice produced serum antibodies to bFGF. An ELISA was performed with assay plate coated with a solution of bFGF diluted in phosphate buffered saline (PBS) by incubating overnight at 4°C. Plates were blocked with bovine serum albumin (BSA) in PBS for 1 hour at room temperature (RT) and then incubated with optimally diluted mouse anti-sera to bFGF on a shaker. Plates were washed with Tris- buffered saline (TBS) and alkaline phosphatase conjugated goat anti- mouse -IgG diluted 1 to 30,000 in TBS supplemented with 0.1 % BSA. After a 45 minute incubation on a shaker at RT, the plates were washed in TBS and pNPP phosphatase substrate system for ELISA was added to each well. The plates were incubated for 30 minutes at RT. The phosphatase reaction was measured by the increase in absorbency at 420 nm.

Mice with the highest serum level of anti-bFGF were given an additional 30 μg of the immunogen intravenously in PBS 21 days after the last immunization. Three days later, spleen cells of the mice were harvested for production of hybridomas.

Hybridomas to human recombinant bFGF were produced using standard established techniques (G. Koehler, C. Milstein, Eur. J. Immunol. 6:51 1 -51 9 (1976) and R.J. Bjercke, et al., J. immunol. Meth. 90:203- 21 3 (1 986)). Minced immune spleens were passed through a stainless steel screen and diluted to 2 x 10⁷ splenocytes/ml in serum-free Dulbecco's Modified Eagles medium (DMEM). Equal volumes of splenocytes and non-lgG secreting murine myeloma cells (P3 x 63 - Ag.8.753; 5 x 1 0⁶ cells/ml) were pelleted and suspended in polyethylene glycol (PEG) that was then diluted to a final 5% PEG concentration over 10 minutes with serum-free DMEM. After centrifugation, the cell pellet was resuspended in complete DMEM containing fetal calf serum, NCTC 109, nonessential Minimum Essential Medium amino acids and sodium pyruvate (1 00x) to give 5x10⁶ cells/ml. Cells were plated at 5 x 10⁵ cells/well into microtiter plates and incubated in a humidified CO₂ atmosphere overnight at 37°C. On each of the next 3 days, 0.1 ml complete DMEM containing hypoxanthine, aminopterin, and thymidine (HAT) was added to each well, which substances prevent the growth of the nonfused, parental cells which are hypoxanthine guanine phosphoribosyltransferase (HGPRT or HPRT) deficient. After 1 0-14 days post-fusion, when hybridomas reached near confluency, supernatants were screened for antibodies to bFGF by ELISA.

Supernatants that had absorbency readings > 8 times background (0.065) were considered positive, consisting of 232 wells positive out of 960 total wells (24%). Positive hybridomas were expanded in 24-well cluster plates, retested as positive to bFGF absorbency readings of 0.5 to 1 .6 and negative readings against acidic FGF of background absorbency. Of the wells tested positive, the cultures with the highest absorbency values were expanded, cloned, and aliquots of cells (5x10⁶) frozen in liquid nitrogen at each stage of expansion. The cloned hybridomas (10⁷) were injected intraperitoneally into BALB/c mice that had been primed with pristane 2 weeks earlier. After 7-10 days, ascites fluid was collected, pooled for each hybridoma cell line, and stored at -80°C.

The isotype and light chain composition of cloned supernatant monoclonal antibodies were determined using ELISA, substituting secondary antibody reagents with alkaline phosphatase conjugated rabbit IgG antibodies specific for IgG,, lgG_2a, lgG_2b, lgG₃, IgM or mouse K or λ light chains.

Fabrication of Microparticles Loaded With an Antibody for Basic Fibroblast Growth Factor Microparticles were manufactured by a modification of a double- emulsion-solvent-extraction technique described in Example 2. 2.5 mg of antibody for bFGF was dissolved in 125 μl water and injected into a flint glass test tube containing 0.245 g PLGA and 0.0025 g PEO dissolved in 1 ml dichloromethane and emulsified with a vortexer. Blanks were prepared using 50 μl water. The solution was re-emulsified in 100 ml of 0.3% aqueous PVA resulting in a double emulsion that was poured into 100 ml of 2% aqueous isopropanol maintained on a magnetic stirrer. The extraction of the dichloromethane to the external alcoholic phase precipitated the dissolved polymer which in turn resulted in formation of microparticles. The system was stirred for 1 hour to assure total extraction of the solvent. The formed microparticles were finally sieved to sizes smaller than 100 μm, rinsed in water, and collected by centrifugation. All microparticles were sterilized by ethylene oxide before use. SMC Culture

Vascular SMCs were harvested as described in Example 2.

SMC Proliferation

Fourth and fifth passage SMCs were seeded into 24-well cluster plates (Falcon, Lincoln Park, NJ) at a density of 20,000 cells per well. Twenty-four hours after plating, the original medium was replaced and transwells containing PLGA 50:50 microparticles with or without Ab were placed into the appropriate wells. After 24 hours bFGF was added to the medium ( 1 mg/ml). After an additional 72 hours, the cells were photographed, trypsinized, and counted on a Coulter Counter (Coulter, model Z_F Hialeah, FL). To achieve different dosages, 5 ( ± 0.05), 1 5 ( ± 0.05) and 30 ( ± 0.08) mg of microparticles were used (n = 3) . Control wells contained DMEM/FBS alone.

Resul s Antibody Effects on SMC Proliferation

Ab-bFGF released from PLGA microparticles inhibited bFGF stimulated SMC proliferation in a dose dependent manner. Figure 3 shows the dose-dependent growth inhibition of SMCs after administration of antibody to bFGF (A1 2) loaded, MolgG (nonspecific antibody) loaded, and blank PLGA 50:50 microparticles over 96 hours. Smooth muscle cells were stimulated with the addition of 1 mg/ml of basic fibroblast growth factor (bFGF) after 24 hours of incubation of the smooth muscle cells with the microparticles. Error bars represent mean ± SD for n = 3 unless noted otherwise. Statistical significance between microparticles containing A1 2 and MolgG at a given dosage is indicated by ( §), whereas (*) designates insignificant statistical difference between control containing no bFGF and A1 2 loaded microparticles at 30 mg (p < 0.05) . As shown in Figure 3, for 30 mg of PLGA microparticles loaded with Ab-bFGF, the final SMC counts were statistically the same as the control wells containing no bFGF. A single factor ANOVA test showed the dose responsiveness to be statistically significant (p <0.05) for 5, 15, and 30 mg dosages. The data were further analyzed with Scheffe's F- test to determine statistical significance of dosages from one another. All other dosages were statistically significantly different from one another. No toxicity effects were evident at any of the dosages.

Claims

CLAIMS:

1 . A microparticle based formulation for use in drug delivery comprising a biodegradable polymer and a water soluble polymer.

2. The microparticle based formulation of claim 1 further comprising a therapeutic compound.

3. The microparticle based formulation of claim 1 wherein the biodegradable polymer has a molecular weight of from about 2,000 to about 100,000 and the water soluble polymer has a molecular weight of from about 200 to about 100,000.

4. The microparticle based formulation of claim 1 wherein the microparticle contains from about 50 to about 99 weight percent of the biodegradable polymer and from about 0.10 to about 50 weight percent of the water soluble polymer.

5. The microparticle based formulation of claim 1 wherein the microparticle has a diameter of from about 10 to about 100 micrometers.

6. The microparticle based formulation of claim 1 wherein the biodegradable polymer is poly(glycolic acid), poly(D-lactic acid), poly(L- lactic acid), poly(D,L-lactic acid), poly(D,L-ethyl glycolic acid), poly(dimethyl glycolic acid), poly(D,L-methyl ethyl glycolic acid), poly(ε- caprolactone) or poly(D,L-lactic-co-glycolic acid).

7. The microparticle based formulation of claim 1 wherein the water soluble polymer is poly (ethylene glycol), poly (ethylene oxide), poly (vinyl methyl ether), poly (acrylic acid), poly (methacrylic acid), poly (2- hydroxy ethyl methacrylate), poly (N-vinyl pyrrolidone), poly (vinyl alcohol), collagen, gelatin, a poly(amino acid) or a polysaccharide.

8. The microparticle based formulation of claim 2 wherein the therapeutic compound is present in the amount of from about 0.10 to about 10 weight percent.

9. The microparticle based formulation of claim 2 wherein the therapeutic compound is an oligonucleotide, protein, nucleic acids, glucocorticoids or small molecular weight drugs.

10. A method for inhibiting smooth muscle cell migration, the method comprising the steps of: a. preparing a microparticle based formulation comprising a biodegradable polymer and a water soluble polymer; b. loading a therapeutic compound into the microparticle that inhibits the migration of smooth muscle cells; and c. delivering the microparticle based formulation containing the therapeutic compound to an area containing smooth muscle cells.

1 1 . The method of claim 10 wherein the biodegradable polymer is poly(glycolic acid), poly(D-lactic acid), poly(L-lactic acid), poly(D,L-lactic acid), poly(D,L-ethyl glycolic acid), poly(dimethyl glycolic acid), poly(D,L- methyl ethyl glycolic acid), poly(ε-caprolactone) or poly(D,L-lactic-co- glycofic acid).

12. The method claim 10 wherein the water soluble polymer is poly (ethylene glycol), poly (ethylene oxide), poly (vinyl methyl ether), poly (acrylic acid), poly (methacrylic acid), poly (2-hydroxy ethyl methacrylate), poly (N-vinyl pyrrolidone), poly (vinyl alcohol), collagen, gelatin, poly(amino acids) or polysaccharides.

1 3. A method for inhibiting smooth muscle cell proliferation, the method comprising the steps of: a. preparing a microparticle based formulation comprising a biodegradable polymer and a water soluble polymer; b. loading a therapeutic compound into the microparticle that inhibits the proliferation of smooth muscle cells; and c. delivering the microparticle based formulation containing the therapeutic compound to an area containing smooth muscle cells.

14. The method of claim 1 3 wherein the biodegradable polymer is poly(glycolic acid), poly(D-lactic acid), poly(L-lactic acid), poly(D,L-lactic acid), poly(D,L-ethyl glycolic acid), poly(dimethyl glycolic acid), poly(D,L- methyl ethyl glycolic acid), poly(ε-caprolactone) or poly(D,L-lactic-co- glycolic acid) .

1 5. The method claim 1 3 wherein the water soluble polymer is poly (ethylene glycol), poly (ethylene oxide), poly (vinyl methyl ether), poly

(acrylic acid), poly (methacrylic acid), poly (2-hydroxy ethyl methacrylate), poly (N-vinyl pyrrolidone), poly (vinyl alcohol), collagen, gelatin, poly(amino acids) or polysaccharides.