Synthesis of Amorphous Polylactide and Poly ( lactide-co-glycolide ) Containing High L-form Enantiomer for Use in Controlled Release Drug Delivery

In this work, the effect of L-lactide (LL) copolymerization on the properties of poly(D,L-lactide) (PDLL) and poly(D,L-lactide-co-glycolide) (PDLLG) copolymers and its drug release behaviors were determined and discussed. The copolymers were synthesized by ring-opening polymerization of DLL, LL, and G monomer mixtures. The PDLL with DLL/LL ratios of 100/0-50/50 by mole and the PDLLG with DLL/LL/G ratios of 75/0/25-37.5/37.5/25 by mole were investigated. All the copolymers were completely amorphous. The drug-loaded copolymer microparticles with a spherical shape and smooth surface were prepared by the oil-in-water emulsion solvent evaporation method. Indomethacin was used as a poorly-water soluble model drug. The copolymerization of the LL monomer did not change the in vitro drug release profiles of the PDLL and the PDLLG microparticles significantly. It is suggested that these amorphous PDLL and PDLLG copolymers that contain higher L enantiomer amounts have the potential to be developed further as a lower-cost PDLL and PDLLG, respectively, for use as controlled release drug delivery systems.


INTRODUCTION
Controlled release drug delivery systems made from biodegradable particles provide several benefits over traditional formulations. 1Prior to release, the drug is protected from degradation or premature metabolism by the polymeric particle matrix.The release of the drug is sustained over days to months, thereby keeping the drug concentration in the plasma at an effective level for longer periods of therapy time and reducing the toxic side-effects from overdose of the drug.This decreases the frequency of administration and increases patient compliance. 2 Biodegradable microspheres for drug delivery have been widely made from a variety of biodegradable polyesters due to their biodegradability and biocompatibility.The removal of these biodegradable polyester-based microspheres at the end of the therapy is not required.5][6] This is due to the fact that good drug distribution into the amorphous PDLL and PDLLG matrices can be obtained.The semi-crystalline phases in the poly(L-lactide) (PLL) matrix may induce drug aggregation.A good distribution of the entrapped drug into the microparticle matrices could allow a consistent drug release rate.Controllable molecular weights of the PDLL and the PDLLG are usually synthesized by ring-opening polymerization of DLL and DLL/G monomers, respectively.The higher hydrophilic G units in the PDLLG resulted in a faster biodegradation rate than the PDLL.
An equivalent D-lactic acid/L-lactic acid mixture, called D,L-lactic acid, is used to prepare the DLL monomer.However, L-lactic acid, the monomer precursor of LL, is produced at a larger scale for supply to food, cosmetic, and pharmaceutical applications.Therefore, the L-lactic acid is easier to find and cheaper than the D,L-lactic acid.The DLL and LL monomers are prepared from the D,Llactic acid and the L-lactic acid, respectively, by the same procedure.This has the benefit of offering a substantial reduction in cost by the addition of LL to produce lower-cost alternatives to the current commercial PDLL and PDLLG for use in controlled release drug delivery applications.
In the present paper, PDLL and PDLLG copolymers with different LL monomer contents were synthesized by ring-opening polymerization of the DLL/LL and the DLL/LL/G mixtures, respectively.We prepared drug-loaded copolymer microparticles using an oil-in-water emulsion solvent evaporation method.Indomethacin was used as the hydrophobic model drug.The characteristics of the drug-loaded PDLL and PDLLG microparticles containing high LL content and in vitro indomethacin release were determined and compared to the neat PDLL and PDLLG microparticles.

Experimental section Materials
Crude D,L-lactide (DLL) and L-lactide (LL) monomers were synthesized from D,L-lactic acid (85% w/v, 50/50 D-/L-form ratio, Acros Organics) and L-lactic acid (88% w/v, 5/95 D-/L-form ratio, Purac), respectively, by direct polycondensation at 180 °C followed by thermal decomposition at 220 °C under reduced pressure.Crude glycolide (G) monomer was synthesized from glycolic acid (99%, Acros Organics) by the same procedure.The reaction temperatures for the direct polymerization and the thermal decomposition stages were 220 °C and 320 °C, respectively, for preparing the crude G monomer.Crude lactide and glycolide monomers were purified by re-crystallization in ethyl acetate four times.The purified monomers were dried under vacuum at 55°C for 48 h before use in the polymerization.1-dodecanol (98%, Fluka) containing a one-hydroxyl end group was purified by distillation under reduced pressure before use.Stannous octoate [Sn(Oct) 2 , 95%, Sigma], indomethacin (99.95%,Sigma), and Tween80 (Acros Organics) were used without further purification.All reagents used were analytical grade.

Synthesis of copolymers
Poly(D,L-lactide) (PDLL) and poly(D,Llactide-co-glycolide) (PDLLG) copolymers were synthesized by ring-opening polymerization of the DLL/LL/G mixtures in bulk at 165 °C for 2.5 h under a nitrogen atmosphere.Sn(Oct) 2 was used as a catalyst at 0.01 mol% and 1-dodecanol was used as an initiator.Copolymers with a theoretical numberaverage molecular weight (M n, theoretical ) of 50,000 g/ mol were prepared.The 1-dodecanol concentrations of 0.28 and 0.27 mol% were used to synthesize the PDLL and the PDLLG, respectively.The crude copolymers were purified by dissolving in chloroform before precipitating in cool n-hexane.They were then dried to a constant weight in a vacuum at 50°C for 48 h.

Characterization of copolymers
The specific optical rotation of the PDLL copolymers was determined in chloroform at a concentration of 1 g/dL at 25°C with a Bellingham and Stanley Polarimeter ADP220 at a wavelength of 589 nm. The chemical compositions of the PDLLG copolymers were measured by 1 H-NMR spectrometry using a Bruker Advance DPX 300 1 H-NMR spectrometer at 25°C.CDCl 3 was used as the solvent, and tetramethysilane was used as the internal standard.
The number-average molecular weight (M n ) and molecular weight distribution (MWD) of the copolymers were determined by Gel Permeation Chromatography (GPC) with a Waters e2695 separations module equipped with PLgel 10 mm mixed B 2 columns operating at 40°C and employing a refractive index (RI) detector.Tetrahydrofuran was used as the solvent at a flow rate of 1.0 mL/min.
The thermal transition properties of the copolymers were determined with a Perkin-Elmer Pyris Diamond differential scanning calorimeter (DSC) under a nitrogen flow.For DSC, copolymers of 5 -10 mg in weight were heated at 10 o C/min over a temperature range of 0 to 200 C (1 st heating scan) to observe their melting temperature (T m ).Then the samples were quenched to 0°C according to the DSC instrument's own default cooling mode before heating from 0 to 200°C (2 nd heating scan) to observe their glass transition temperature (T g ).The T m was measured as the peak value of the endothermal phenomena in the DSC curve.The T g was taken as the midpoint or half of the heat capacity increment, associated with the glass-to-rubber transition.

P r e p a r a t i o n o f d r u g -l o a d e d c o p o ly m e r microparticles
The copolymer microparticles entrapping the indomethacin model drug were prepared by the oil-in-water emulsion solvent evaporation method.Dichloromethane was used as an organic solvent.90 mg of copolymer and 10 mg of indomethacin were dissolved in 2.5 mL of dichloromethane (oil phase).The oil phase was slowly added-drop wise into 400 mL of a 2% w/v Tween80 solution in distilled water (water phase) under magnetic stirring.The organic solvents were evaporated in a fume hood for 6 h.The drug-loaded microparticles suspended in the water phase were obtained.The resulting microparticles were collected by centrifugation before freezedrying.

Characterization of drug-loaded copolymer microparticles
The morphology of the microparticles was observed by scanning electron microscopy (SEM, JEOL JSM-6460LV).The microparticles were sputter-coated with gold to enhance the surface conductivity before scanning.The average size of the microparticles was determined from several SEM images by counting a minimum of 100 particle diameters using the smile view software (version 1.02).
The drug loading of the microparticles was measured by UV-vis spectrophotometry (Lamda 25, Perkin Elmer).For this purpose, the drugloaded microparticles were completely dissolved in dichlorometane before analysis with a UV-vis spectrophotometer at l max = 319 nm. 9 The amount of indomethacin was calculated by comparison with a standard equation of indomethacin solution in dichloromethane.The standard equation and R 2 were y = 0.0181x + 0.0158 and 0.9996, respectively.The theoretical drug loading content (DLC theoretical ), actual drug loading content (DLC actual ), and drug loading efficiency (DLE) were calculated form Equations ( 1) -( 3), respectively.The DLC actual was an average value from three measurements.

In vitro drug release tests
An in vitro drug release test with the microparticles was performed.About 10 mg of drugloaded microparticles were placed in a pretreated dialysis bag before being incubated in a flask containing 200 mL of 0.02 M phosphate buffer saline (PBS, pH 7.4).The flasks were kept in a shaker incubator at 37°C and 100 rpm for 48 h.At each desired time, some supernatant was withdrawn and replaced with an equal volume of fresh PBS medium.The release concentration of the indomethacin in the supernatant was determined by a UV-vis spectrophotometer at l max = 319 nm 9 .
The amount of indomethacin model drug was calculated by comparison with a standard equation of indomethacin solution in PBS.The standard equation and R 2 were y = 0.02x + 0.0047 and 0.9994, respectively.The cumulative release

Characterization of copolymers
The yields of all the copolymers obtained from the precipitation method were higher than 85%.Table 1 reports the number-average molecular weight (M n ) and molecular weight distributions (MWD) of the copolymers obtained from the GPC curves.All the GPC curves were unimodal.The M n and MWD values were in the ranges of 47,100-56,400 g/mol and 1.5-2.7,respectively.The M n values obtained from the GPC were close to the values of the theoretical M n (50,000 g/mol).The M n of the copolymers was then controlled by the 1-dodecanol concentration.
The PDLL copolymers with different DLL/ LL ratios were polymerized from the mixtures of DLL and LL monomers.The feed DLL/LL ratios were 100/0, 90/10, 80/20, 70/30, 60/40, and 50/50 by mole, which corresponded to the feed X D /X L ratios  2. The final X D /X L ratios of the PDLL copolymers determined from the polarimetry are also reported in Table 2.
They were very close to the values of the feed X D /X L ratios.The results suggest that the PDLL copolymers with different X D /X L ratios can be synthesized from the DLL/LL mixtures.
The PDLLG copolymers with different DLL/ LL/G ratios were synthesized from the mixtures of DLL, LL, and G momomers.The DLL/LL/G ratios were 70/0/25, 62.5/7.5/25,60/15/25, 52.5/22.5/25,45/30/25, and 37.5/37.5/25by mole, as reported in Table 2.The feed DLL-LL/G ratio was constant at 75/25 by mole.The final DLL-LL/G ratios were determined from the 1 H-NMR spectrum, an example of which is shown in Figure 1 for the 62.5/7.5/25DLL/ LL/G copolymer including peak assignments.It can be seen that the half DLL and LL units exhibited the same peaks as the methine protons (-CH; peak a) and methyl protons (-CH 3 ; peak b) at 5.2 and 1.5 ppm, respectively. 10The DLL/LL ratios of the PDLL copolymers could not be determined from the 1 H-NMR spectra.The half G units showed peaks of methylene protons (-CH 2 ; peak c) in the range 4.6- 5.0 ppm. 11The final DLL-LL/G ratios of the PDLLG copolymers calculated from the integral peak areas of the peaks a and c are also reported in Table 2.They were nearly the same values as with the feed DLL-LL/G ratio (75/25 by mole).
From the 1 st heating scan DSC curves, the melting peaks of all the PDLL and the PDLLG copolymers were not found (DSC thermograms not shown).This suggests that the copolymers with the DLL/LL and the DLL/LL/G ratios in the ranges of this study were completely amorphous.It is well known that amorphous polyesters are more appropriate for use in drug delivery.The entrapped drug could be uniformly distributed into the amorphous matrix better than the semi-crystalline matrix.This enhances the consistent drug release.The glass transition temperatures (T g ) of the copolymers obtained from the 2 nd heating scan DSC curves were similar and in the ranges of 53-54°C and 48-49°C for the PDLL and the PDLLG copolymers, respectively.The results indicate that the DLL/LL and DLL/LL/G ratios do not affect the T g values of the copolymers.The T g values of the PDLLG copolymers were slightly lower than the PDLL copolymers because the T g of the homopolymers of polyglycolide (37°C) was lower than the polylactide (55°C).

Characterization of copolymer microparticles
The drug-loaded copolymer microparticles of PDLL and PDLLG were prepared by the oil-inwater emulsion solvent evaporation method.The yields of the microparticles, based on the weights of the feed copolymer and drug, were in the range of 82.4-93.7%.The morphology of the drug-loaded microparticles was observed from the SEM images, as shown in Figures 2 and 3 for the PDLL and the PDLLG, respectively.They were spherical in shape with a smooth surface.The average particle sizes determined from several SEM images are summarized in Table 3.They were in the range of 89-113 mm.The morphology results suggest that the monomer ratio of the copolymers did not significantly affect their particle morphology and size.
The theoretical drug loading content (DLC theoretical ) of all the copolymer microparticles calculated from Equation (1) was 10 wt %.The actual drug loading content (DLC actual ) and the drug loading efficiency (DLE), as summarized in Table 3, were in the ranges of 5.39-6.60%and 53.9-66.0%,respectively.Both DLC actual and DLE did not change significantly with the DLL/LL and the DLL/LL/G ratio.

In vitro drug release
The in vitro drug release test was performed in PBS at 37°C for 48 h.The release profiles of indomethacin are illustrated in Figure 4 for the PDLL microparticles.All PDLL microparticles with different DLL/LL ratios showed similar sustained drug release patterns.An initial burst release of the drug near the particle surfaces was detected within the first 12 h followed by a slower drug release.The initial burst release of the PDLL microparticles was in the range of 30-40%.The drug release was in the range of 60-70% at 48 h of release time.
Figure 5 shows the drug release profiles of the PDLLG microparticles.All PDLLG microparticles with different DLL/LL/G ratios exhibited similar sustained drug release profiles; an initial burst release effect within the first 12 h followed by slower drug release.However, the ranges of the initial burst release and drug release at 48 h of the PDLLG microparticles were 50-60% and 80-90%, respectively.It can be concluded that the DLL/LL and DLL/LL/G ratios did not significantly influence the drug release behaviors of the PDLL and the PDLLG microparticles, respectively.However the PDLLG microparticles exhibited a faster drug release than those of the PDLL microparticles.The drug release results suggest that the PDLL and PDLLG copolymers with different monomer mole ratios showed potential for use as controlled release drug delivery systems.The concentrations of the drug in the plasma could be maintained at the therapeutic level for longer periods of time.Therefore, the frequency of drug administration could be reduced.
The predominant drug release mechanism of the PDLL microparticles was proposed to be the drug diffusion process.This was supported by the SEM image of the PDLL microparticles in Figure 6(a).The PDLL microparticles after 48 h release were still spherical in shape and had a smooth surface.Meanwhile the PDLLG microparticles had surface erosion, as shown in Figure 6(b).Thus the drug release mechanisms of the PDLLG microparticles within 48 h may include both drug diffusion and surface erosion. 10This may be due to the higher hydrophilicity and the lower T g of the PDLLG microparticle matrices that gave easier water penetration into the particle matrices.Therefore, higher water penetration induces faster surface erosion.

CONCLUSIONS
PDLL and PDLLG copolymers with different DLL/LL and DLL/LL/G mole ratios, respectively, were successfully synthesized via ring-opening polymerization of the mixtures of DLL, LL, and G monomers.They were completely amorphous.The drug-loaded copolymer microparticles of the PDLL and the PDLLG were prepared by the oil-inwater emulsion solvent evaporation method.These amorphous microparticles could be used to entrap a poorly-water soluble model drug, indomethacin with a 89-113 mm size and a 53.9-66.0%loading efficiency for drug delivery.All the copolymer microparticles were spherical in shape and had a smooth surface.The copolymerization of the LL units was as high as 50 and 37.5 mol% for the PDLL and the 75/25 PDLLG copolymers, respectively, did not induced PLLA crystallization, and did not affect the drug release profiles significantly.The PDLLG microparticles showed faster drug release than the PDLL microparticles.
In conclusion, the results presented here show that the amorphous PDLL and PDLLG copolymers synthesized in this work have the potential to be developed further as drug delivery systems.The amorphous PDLL and PDLLG copolymers with a high L enantiomer content could provide a viable lower-cost alternative to the commercial PDLL and PDLLG.

Table 2 : Enantiomer ratios and chemical compositions of copolymers Comonomer ratio Feed Final Feed Final (by mole) X D /X L ratio X D /X L ratio DLL-LL/G ratio DLL-LL/G ratio (by mole) a (by mole) b (by mole) c (by mole) d
bDetermined from polarimetry method.c Calculated from feed DLL/LL/G ratio.d Calculated from 1 H-NMR spectra.

Table 3 : Characteristics of drug-loaded copolymer microparticles
a Determined from several SEM images b Calculated from Equation (2) c Calculated from Equation(3)