122 (2007) 338–344www.elsevier.com/locate/jconrel

Journal of Controlled Release

Effect of acidic pH on PLGA microsphere degradation and release

Banu S. Zolnik⁎, Diane J. Burgess

Department of Pharmaceutical Sciences, University of Connecticut, Storrs, CT 06269, United States

Received 2 March 2007; accepted 28 May 2007Available online 5 June 2007

Abstract

Polymer degradation and drug release kinetics from PLGA microspheres were investigated under neutral and acidic pH conditions. Twodifferent Mw formulations (Mw: 25,000 and 70,000) were investigated and both exhibited a triphasic release profile at pH 7.4 as well as at pH 2.4.The initial burst and lag phases were similar for both pH values, while the secondary apparent-zero-order phase was substantially accelerated atpH 2.4. The polymer molecular weight change with time for the microspheres followed first order degradation kinetics for both pH values. Alinear relationship was established between % drug release (post burst release) and Ln (Mw) for both pH conditions. Most significantly,morphological studies showed that the mechanism of polymer degradation changed from “inside-out” degradation at pH 7.4 to “outside-in” atpH 2.4. At pH 7.4, the microspheres followed the usual morphological changes such as surface pitting and pore formation. Whereas, at pH 2.4 themicrospheres maintained smooth surfaces throughout the degradation process and were susceptible to fracturing. The fracturing of themicrospheres was attributed to crystallization of oligomeric degradation products as a consequence of their low solubility at this pH. It alsoappeared that degradation occurred in a more homogeneous pattern at pH 2.4 than is typical of PLGA microspheres at pH 7.4. This may be a resultof the entire microspheres experiencing a close-to-uniform pH at 2.4. However, at pH 7.4, the local micro-environmental pH within themicrospheres has been reported to vary considerably due to a build up of acid oligomers. This heterogeneous degradation results in the randomformation of channels within microspheres degraded at pH 7.4 which was not observed in those degraded at pH 2.4. This is the first time thatmorphological changes during PLGA microsphere degradation have been compared for low and neutral pH and the data shows a change in themechanism of degradation at the low pH.© 2007 Elsevier B.V. All rights reserved.

Keywords: Biodegradable polymers; Dexamethasone; USP 4 apparatus

1. Introduction

Polyesters, such as poly(lactic acid) (PLA) and poly(lactic-co-glycolic acid) (PLGA), have been used as medical resorbablesutures for the past 40 years [1]. These polymers have been usedin the pharmaceutical and biomedical fields as delivery vehicles,(such as, microspheres and nanoparticles) and as scaffoldmaterials as a consequence of their biodegradability and relativebiocompatibility [2]. Polymer properties such as molecularweight (Mw), copolymer composition and crystallinity can betailored to alter polymer degradation and the consequent drug

⁎ Corresponding author. Current Address: Nanotechnology CharacterizationLaboratory, SAIC-Frederick, Inc., National Cancer Institute, Frederick, MD21702, United States. Tel.: +1 860 486 3760; fax: +1 860 486 2076.

E-mail address: d.burgess@uconn.edu (D.J. Burgess).

0168-3659/$ - see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.jconrel.2007.05.034

release profiles. For example, increase in polymer Mw results inlonger degradation times and slower release [3–5].

Release from PLGA microspheres occurs via diffusion,polymer erosion or a combination thereof [6]. It has been shownthat release from low Mw PLGA (e.g., 5 kDa) microspheres isdiffusion controlled whereas, release from highMw PLGA (e.g.,25 kDa) microspheres typically occurs via a combination ofdiffusion and erosion [7]. PLGA erosion occurs via hydrolysis ofthe ester bonds in the polymer backbone. It is widely establishedthat PLGA degradation starts with water uptake, and thathydrolysis leads to the production of acidic oligomers [8,9]. Theretention of the oligomeric degradation byproducts within themicrospheres, which results from their relative hydrophobicity,has been reported to impact the degradation mechanism. Thesemicrospheres typically degrade from the “inside out” due to theoligomeric acid build up in within the microspheres creating an

339B.S. Zolnik, D.J. Burgess / Journal of Controlled Release 122 (2007) 338–344

acidic environment that leads to autocatalysis [10–13]. The pHinside the microspheres has been investigated using confocalmicroscopy with pH sensitive fluorescent dyes, potentiometry,nuclear magnetic resonance and electron paramagnetic reso-nance spectroscopy, as well as with direct microclimate analysisvia microsphere dissolution [11,14–18]. The reported valuesrange from 1.5 to 6.4 and this has been attributed to differencesin the PLGA microsphere formulations investigated. Inparticular, the size and porosity of the PLGA microspheresappeared to be critical to acid build up [11,12].

Drug release from PLGA microspheres can range from daysto months and therefore, accelerated in vitro drug release testingmethods are often used for manufacturing batch release.However, these methods often involve change in the mecha-nism of drug release from erosion/diffusion control to diffusioncontrol. For example, use of organic solvents solubilizes thepolymer, which can lead to diffusion controlled kinetics. Thischange in the mechanism of release may result in lack ofcorrelation between accelerated release profiles and in vivorelease profiles. In particular, where multiple release phases(such as burst release, lag phase and secondary burst phase)occur, these different phases are often lost in accelerated releasetesting. Correlations have been established for the post burstrelease phase and these can be augmented by real-time tests forthe burst phase [7,19]. In a previous publication from ourlaboratory, it was shown that dexamethasone release followedpolymer degradation under elevated temperature conditions,and obeyed Arrhenius kinetics for the temperatures investigated(37–70 °C) for the post burst release phase [7].

To establish an in vitro–in vivo relationship (IVIVR) forbioequivalence, it may be necessary to develop an in vitrorelease method that mimics the entire in vivo profile. For this, anaccelerated method that does not alter the mechanism of drugrelease may be beneficial. Acid and alkaline conditions catalyzeester hydrolysis and consequently PLGA degradation, and areexpected to accelerate drug release without change in themechanism of release and therefore should enable IVIVRmodeling.

Autocatalysis of PLGA occurs as a result of the generation ofacidic oligomeric units within the microspheres and these play animportant role in PLGA degradation resulting in the formation ofchannels through which drug release occurs [10]. Therefore, wewere interested to investigate the effect of acidic pH on drugrelease from PLGA microspheres to determine how such mediamight affect polymer degradation. In particular, the effect onmicrosphere morphological changes and the consequent drugrelease profile was investigated. Accordingly, a comprehensivemicrosphere characterization (gel permeation chromatography(GPC), differential scanning calorimetry (DSC) and environ-mental scanning electron microscopy (ESEM)) was conducted todetermine factors that may contribute to drug release under acidicconditions.

2. Materials and methods

PLGA Resomer RG503H 50:50 (Mw: 25,000, carboxylicacid end group) and PLGA Medisorb 65:35 DL (Mw: 70,000,

end-capped) were gifts from Boehringer-Ingelheim, and PurduePharma, respectively. Methylene chloride and tetrahydrofuran(THF) were obtained from Fisher Scientific (Pittsburgh, PA).Dexamethasone, and poly(vinyl alcohol) (PVA) were obtainedfrom Sigma (St. Louis, MO).

2.1. Preparation of microspheres

An oil-in-water emulsion solvent extraction/evaporationtechnique was used for dexamethasone microsphere formula-tion. 2 g of PLGAwere dissolved in 8 ml of methylene chloride,and 200 mg of dexamethasone were dispersed in this solutionusing a homogenizer (PowerGen Model 700D homogenizerwith flat-bottom generator, Fisher Scientific, Pittsburgh, PA) at10,000 rpm for 1 min. This organic phase was added slowly to40 ml of a 1 % (w/v) aqueous PVA (Av. Mw 30,000–70,000)solution and homogenized at 10,000 rpm for 3 min. Thisemulsion was added to 500 ml of a 0.1 % (w/v) aqueous PVAsolution and stirred at 250 rpm under reduced pressure for 6 h at25 °C. The resulting microspheres were filtered (DuraporeMembrane Filter, 0.45 μm, Fisher Scientific, Pittsburgh, PA),washed three times and vacuum dried for 24 h.

2.2. Characterization of microspheres

2.2.1. High performance liquid chromatography (HPLC)The concentration of dexamethasone was determined using

HPLC. The HPLC system consisted of a Constametric 4100pump (Thermoseparation), an automatic sample injector (Bio-Rad) and a UVabsorbance detector (Bio-Rad) set at 242 nm. Themobile phase consisted of acetonitrile:water:phosphoric acid(30:70:0.5 v/v/v). The analytical column was a Nova-Pak® C18

(9 mm×150 mm) (Millipore Corp, Waters, Milford, MA). Theflow rate was set at 1 ml/min. The retention time of dexa-methasone was 5 min. The chromatograph was analyzed byPeakSimple Chromatography System (Model 203, software3.29, SRI instruments, Torrance, CA). This method is a stability-indicating HPLC assay [20].

2.2.2. Drug loading10 mg of microspheres were dissolved in 10 ml of THF,

filtered (Millex-HV, 0.45 μm, Fisher Scientific, Pittsburgh, PA)and analyzed using the HPLC method described above fordexamethasone content. Encapsulation efficiency was deter-mined as reported by Fu et al. [21]:

Encapsulationefficiency¼ Experimental drug loading=theoretical drugð Þ � 100k

All measurements were conducted in triplicate and the meanvalues and standard deviations are reported.

2.2.3. Differential scanning calorimeterSamples were analyzed using a TA Instruments 2920 DSC.

Samples were heated to 150 °C and cooled to −30 °C at a rateof 5 °C/min. The second cycle was used to determine theglass transition temperature (Tg) for characterization of the

Table 1Microsphere formulation and physico-chemical characterization

Formulation Ratio ofLA:GA

Mw(kDa)

Tg (°C) Drugloading

Encapsulationefficiency (%)

25 K 50:50 25 44.2 7.4±1.3 71±9.270 K 65:35 70 47.7 7.5±1.1 73±8.6

Fig. 1. Dexamethasone release from PLGA microspheres (formulation 25 K) inpH 7.4 and pH 2.4 PBS buffer using USP apparatus 4 at 37 °C: (□) pH 7.4, and( ) pH 2.4. (Mean±std dev; n=3).

340 B.S. Zolnik, D.J. Burgess / Journal of Controlled Release 122 (2007) 338–344

microspheres as prepared. Data obtained from the first cyclewas used to characterize the degradation profile of themicrospheres. Samples were analyzed in aluminum pans withpinhole lids.

2.2.4. Gel permeation chromatographyThe Mw of both formulations (25 K and 70 K) of PLGAwas

determined using GPC (Waters) with an evaporative lightscattering detector (ELSD, Polymer Laboratories PL-ELS1000). Four Jordi Flash (high-speed) columns were connectedin series. The mobile phase was THF with a flow rate of 3 ml/min at 40 °C. Microspheres (10 mg) were dissolved in 10 ml inTHF and filtered (0.22 μm, Whatman). All glass equipment(vials and syringes) was used to minimize possible contamina-tion from plastic materials. The polymer solution injectionvolume was 100 μl. The data collection and analysis wereperformed using Waters Millennium software. Weight averagemolecular masses were calculated based on polystyrenestandards (1,250,000; 400,000; 200,000; 43,000; 17,600;6930; 2610; 982; and 472 Da). Calibration was performedwith every experiment. All the measurements were conductedin triplicate and the mean values and standard deviations arereported.

2.2.5. Environmental scanning electron microscopy (ESEM)The morphology of dexamethasone loaded microspheres

was obtained using a Philips Electroscan ESEM 2020 at 10 kVaccelerating voltage. Samples were mounted onto carbon-tabbed aluminum stubs and sputtered with a gold/palladiumcoating prior to analysis.

2.2.6. In vitro release studiesIn vitro release studies were conducted using a modified

USP apparatus 4 (Sotax CE7 smart, or CY 7 piston pump,Sotax, Horsham, PA) with flow-through cells (12mm diameter)packed with glass beads (1 mm) (to prevent microsphereagglomeration and to achieve laminar flow [22]) in a closedsystemmode at 37 °C. 45mg of microspheres were dispersed inthe flow-through cells and 250 ml of 0.1 M phosphate bufferedsaline (PBS) pH 7.4 or pH 2.4 (prepared according to the USPmonograph) with 0.1 % sodium azide was circulated through afiberglass filter (0.45 μm). A flow rate of 20 ml/min was used.1.3 ml samples were withdrawn and replenished with 1.3 ml offresh media. The samples were analyzed by HPLC (asexplained above) and any drug degradation was accountedfor in the cumulative release data reported. When the drugconcentration reached 5% (w/v) of the solubility of dexameth-asone, half of the total media volume was replenished. Thismedia replenishment was taken into account in the calculation

of the cumulative percent release. All measurements wereconducted in triplicate and the mean values and standard de-viations are reported.

2.2.7. In vitro degradation studies20 mg of microspheres were dispersed in 20 ml of 0.1 M

PBS (pH 7.4 or pH 2.4) with 0.1 % sodium azide in flat-bottomed vials and placed in a shaker water bath (C76, NewBrunswick Scientific, Edison, NJ) at 100 rpm at 37 °C. Vialgeometry was chosen to minimize aggregation. At knownintervals, samples were decanted, vacuum dried for 24 h andanalyzed by ESEM, DSC and GPCmethods as described above.All measurements were conducted in triplicate and mean valuesand standard deviations are reported.

3. Results

3.1. Polymer selection and microsphere characterization

To investigate the effect of acidic pH on the mechanismand rate of release of drug from PLGA microspheres, two for-mulations were prepared with different PLGA molecularweights: 25,000 Da and 70,000 Da. These are referred to asformulations 25 K and 70 K, respectively. The microsphereformulations were prepared with similar drug loadings (7.4±1.3% and 7.5±1.1%) to minimize the number of variables thatcan affect drug release [23]. The calculated encapsulationefficiencies were 71±9.2% and 73±8.6% for formulations 25 Kand 70 K, respectively. Table 1 shows the characteristics of thesemicrosphere formulations. As expected, the microsphere Tgvalues increased as the Mw increased and the ratio of lactic acid(LA) to glycolic acid (GA) changed. The Tg values and theirrange (onset-end point) are 44.2 °C (39.7–49.0), and 47.7 °C(43.1–52.0) for formulations 25 K and 70 K, respectively.

3.2. “Real-Time” and acidic pH-accelerated release studies for25 K and 70 K formulations

Drug release kinetics from the 25 K PLGA microspheresfollowed a triphasic release profile typical of PLGAmicrospheres(initial burst release, followed by a lag phase and a secondaryapparent-zero-order phase) for both “real-time” (pH 7.4) and

▪

Fig. 2. Dexamethasone release from PLGA microspheres (formulation 70 K) inpH 7.4 and pH 2.4 PBS buffer using USP apparatus 4 at 37 °C: (▵) pH 7.4, and(▴) pH 2.4 (Mean±std dev; n=3).

Fig. 4. Gel permeation chromatography elution profiles of dexamethasonemicrospheres (formulation 25 K) incubated at 37 °C for 28 days: (a) pH 7.4 PBSbuffer, (b) pH 2.4 PBS buffer.

341B.S. Zolnik, D.J. Burgess / Journal of Controlled Release 122 (2007) 338–344

accelerated conditions (pH 2.4) (Fig. 1). The initial burst releasevalues were 52 and 46% for pH 7.4 and 2.4, respectively. Thisinitial high burst release may be associated with surface poresformed during formulation of themicrospheres [7] aswell as othersurface associated drug. The lag phases were comparable at bothpH conditions. However, the kinetics of the secondary apparent-zero-order phase was faster at pH 2.4 (3.48 day−1) than at pH 7.4(2.13 day−1). The release profiles reached a plateau at 19 days andat 30 days for pH 2.4 and 7.4, respectively.

A similar trend occurred for the 70 K formulation (Fig. 2).The initial burst release and lag phases were similar at both pHvalues. The time to reach 80% of drug release at pH 2.4 was52 days compared to 84 days at pH 7.4. The kinetics of thesecondary apparent-zero-order phase was faster at pH 2.4(5.45 day−1) compared to that at pH 7.4 (1.85 day−1).

3.3. “Real-Time” and acidic pH-accelerated degradationstudies for 25 K formulation

The Mw change of the PLGA microspheres was monitoredusing GPC for the 25 K formulation. The Mw change appearedto follow first order degradation kinetics (Fig. 3) with theapparent rate constants of 0.0927 and 0.0483 day−1 for pH 2.4and pH 7.4, respectively. Microspheres exhibited a unimodalmolecular weight distribution initially, however as degradationprogressed, a bimodal degradation profile was observed at

Fig. 3. The effect of pH on the change in molecular weight (Mw) of PLGAmicrospheres (formulation 25 K) incubated at 37 °C. Mw was plotted againsttime in days. Samples were incubated in PBS buffer: (□) pH 7.4, and (▪) pH2.4. The polymer molecular weight was measured using gel permeationchromatography (GPC). (Mean±std dev; n=3).

pH 7.4 (Fig. 4a). On the other hand, at pH 2.4, the Mw dis-tribution was bimodal up to day 12, following which a unimodalprofile was observed (Fig. 4b). Unprocessed PLGA and PLGAmicrospheres as prepared before exposure to media are shownin Fig. 4a only.

3.4. Relationship between drug release and Mw change in“Real-Time” and acidic pH-accelerated conditions for 25 Kformulation

In order to investigate whether a relationship could beestablished between change in PLGA Mw and drug release rate(post burst release), the Ln (Mw) was plotted against the % drug

Fig. 5. Correlation between Ln PLGA molecular weight and % dexamethasonerelease from PLGA microspheres. Ln (Mw) of PLGA was plotted againstcumulative % released from PLGA microspheres incubated in PBS buffer: (□)pH 7.4, and (▪) pH 2.4 at 37 °C.

Fig. 6. SEM micrographs of formulation 25 K after exposure to pH 7.4 at 37 °Cover time: (a) initial, (b) day 5, (c) day 10 (d) day 20.

342 B.S. Zolnik, D.J. Burgess / Journal of Controlled Release 122 (2007) 338–344

release with time. Ln (Mw) was selected since the change inMw followed first order degradation kinetics. As shown inFig. 5, a linear relationship existed for both pH conditions

Fig. 7. SEM micrographs of formulation 70 K after exposure to pH 7.4 at 37 °Cover time: (a) initial, (b) day 14, (c) day 30, (d) day 41, (e) day 60, and (f) day 81.

Fig. 8. SEM micrographs of formulation 25 K after exposure to pH 2.4 at 37 °Cover time: (a) initial, (b) day 1, (c) day 3, (d) day 7, (e) day 10, and (f) day 17.

indicating that drug release was dependent upon polymerdegradation (erosion/diffusion release).

3.5. Morphological changes in microspheres in “Real-Time”and acidic pH-accelerated conditions

ESEM micrographs revealed that the 25 K microspheresinitially exhibited a spherical geometry with smooth surfaces(Fig. 6a). As degradation proceeded at pH 7.4 the micros-pheres showed some surface erosion at day 5 and thisincreased with time (Fig. 6b). By day 10, the 25 K micros-pheres appeared to agglomerate and by day 20, they formeda large block of polymer (Fig. 6c and d). The 70 K formu-lation exhibited similar behavior at pH 7.4 but over a slowertime frame. Polymer erosion became evident by day 30(Fig. 7a and c) and this progressed until day 60 (Fig. 7dand e). By day 81, no individual microspheres were observed(Fig. 7f).

Morphological changes in the microspheres with time weredifferent at pH 2.4 compared to pH 7.4 (Fig. 8b and f). Themicrospheres had smooth surfaces throughout the 17-dayperiod. The microspheres started to change from a sphericalgeometry at day 1 and by day 7, it was evident that the

Fig. 9. Differential scanning calorimetry of formulation 25 K after exposure to(a) pH 7.4, and (b) pH 2.4 in PBS buffer with time.

343B.S. Zolnik, D.J. Burgess / Journal of Controlled Release 122 (2007) 338–344

microspheres were fragmenting. Fusion of fragmented micro-spheres appeared to occur around day 10 (Fig. 8e).

3.6. Thermal analysis of microspheres in “Real-Time” andacidic pH-accelerated conditions

DSC figures (Fig. 9a) revealed that the microspheresexhibited a Tg (obtained from the first cycle) with an onset at45.8 °C at the initial time point. A slight increase in the onset Tgvalues (52.2 °C) was observed at day 1, as degradation of PLGAprogressed at pH 7.4, the Tg decreased to 41.9 °C at day 5. Inaddition, a second transition was observed at day 10, which hadan onset value of approximately 56 °C (Fig. 9a). On the otherhand, at pH 2.4, there was no significant reduction in the Tg andat days 1 and 3, an additional endothermic event was observedat approximately 126 and at 129 °C, respectively (Fig. 9b).

4. Discussion

Acidic pH conditions (pH 2.4) did not affect burst releasefrom PLGA microspheres (Formulations 25 K and 70 K)indicating that diffusion kinetics, which are known to governthe burst phase [7], were independent of the media conditionsinvestigated. However, the rate of the secondary apparent-zero-order phase is increased at pH 2.4 for both formulations. LowpH catalyzes breakage of the ester linkage of the polymer

backbone enhancing polymer erosion which is known to controlthe secondary apparent-zero order phase [6].

In our previous publication, it was shown that the energy ofactivation (Ea) of dexamethasone release was in agreement withpreviously reported Ea of PLGA degradation [7,24]. This andthe insolubility of dexamethasone in the PLGA matrix indicatedthat dexamethasone release followed polymer erosion for thesecondary apparent-zero order phase [7]. As a result, a linearrelationship between % drug released and Ln (Mw) data wasestablished at pH 7.4. A linear relationship is also demonstratedhere for acidic pH (pH 2.4), indicating that drug release waserosion controlled.

Unexpectedly, the morphology of degrading microsphereswas significantly different at the two pH values. At pH 7.4, themicrospheres showed typical erosion behavior with surfacepitting and with an increase in the number and size of chan-nels with time [25]. This behavior is indicative of autocatalysiswhere acidic oligomeric units accumulate and create a porousnetwork within the microspheres, accelerating PLGA degrada-tion. On the other hand, the microspheres incubated at pH 2.4had smooth surfaces throughout the entire study period withno evidence of surface pitting or channels. In addition, thesemicrospheres appeared to become progressively brittle withtime, as evident from the presence of fragmented microspheres.It is speculated that the acidic oligomeric units accumulatewithin the microspheres at pH 2.4 and due to their low solubilityat this pH compared to pH 7.4 (the pKa value of lactic acid is3.8) they crystallize explaining the brittleness of the micro-spheres. The presence of crystalline oligomeric units wasconfirmed by DSC where an additional endothermic meltingpeak was observed at approximately 130 °C for samples in-cubated at pH 2.4 but not for those incubated at pH 7.4. Thisendothermic melting peak occurs at a temperature close to thatpreviously reported for PLGA oligomers [4]. It has beenreported that the initial loss in mass of PLGA microspheres ondegradation was minimal at acidic pH compared to alkalinemedia [26,27], which is consistent with oligomeric degradationproducts remaining within the microspheres at acidic pH due tolow solubility at this pH. On the contrary to what was demon-strated by Faisant et al. [28], an increase in the Tg for samplesincubated at pH 2.4 was not observed.

The smooth appearance of the microspheres incubated atpH 2.4 and of the subsequent microsphere fragments throughoutthe entire degradation process is considered to be a consequenceof the relative homogeneity in pH within these microspheres.Whereas, microspheres incubated at higher pH (above the pKa)are thought to be heterogeneous in pH as a result of pockets ofacidic PLGA oligomers which cause non-uniform degradationcreating a porous microstructure. The change in the molecularweight distribution with time chromatograph exhibits a bimodaldistribution for the 25 K formulation in pH 7.4 media, inicatinga heterogeneous distribution of PLGA oligomers. Whereas, theMw distribution of PLGA degradation for the 25 K formulationin pH 2.4 media indicated more homogenous degradation.Double glass transition peaks (Fig. 8a) were observed forsamples incubated at pH 7.4 from days 10, 15, and 20,indicating heterogeneous degradation within the microspheres.

344 B.S. Zolnik, D.J. Burgess / Journal of Controlled Release 122 (2007) 338–344

Other researchers have utilized alkaline pH to enhancedegradation of PLGAmicrospheres and therefore the drug releaserate [27,29]. The morphological changes observed here underacidic conditions were not observed at alkaline pH. This can beexplained by the high solubility of PLGA oligomeric units inalkaline pH as well as the lack of an approximately uniform pHthroughout the degrading microspheres.

5. Conclusions

Acidic pH conditions can be utilized to accelerate drugrelease from PLGA microspheres. At pH 2.4, the mechanism ofdrug release (i.e., dominantly polymer erosion) did not changefrom that under “real-time” (pH 7.4) conditions. Although theextent of acceleration of drug release achieved would not besuitable for an accelerated test for batch release of product, thistype of accelerated test may be appropriate for the establishmentof an IVIVR for a product with a complex release profile. Oftenaccelerated tests change the shape of the release profile andmultiple release phases are lost. For example, the burst releasephase becomes indistinguishable.

Most interestingly, the data reported here show that themorphology of the microspheres during degradation at acidic pHis considerably different from that at pH 7.4. This was consideredto be as a result of homogeneity in pH throughout the micros-pheres at pH 2.4 and the insolubility of the oligomeric PLGAdegradation products at this pH.

Acknowledgments

The authors wish to acknowledge Dr. Fotios Papadimitra-kopoulos, Institute of Materials Science, University of Con-necticut for helpful discussions and insightful comments on thisresearch. This research was supported by US Army MedicalResearch and Material Command (W81XWH-04-1-0779 andW81XWH-05-1-0539); Office of Testing and Research CDER,FDA; CPPR-NSF; and Sotax Corp. The awarding of a USPfellowship to BSZ is greatly appreciated.

References

[1] A.K. Schneider, Polylactide Sutures, U. S. Patent, 3,636,956, United States(1970).

[2] R.K. Kulkarni, E.G. Moore, A.F. Hegyeli, F. Leonard, Biodegradable poly(lactic acid) polymers, J. Biomed. Mater. Res. 5 (1971) 169–181.

[3] T.G. Park, Degradation of poly(D,L-lactic acid) microspheres: effect ofmolecular weight, J. Control. Release 30 (1994) 161–173.

[4] T.G. Park, Degradation of poly(lactic-co-glycolic acid) microspheres:effect of copolymer composition, Biomaterials 16 (1995) 1123–1130.

[5] L. Ramirez, P. Pastoriza, R. Herrero-Vanrell, Biodegradable poly(DL-lactic-co-glycolic acid) microspheres containing tetracaine hydrochloride. in-vitrorelease profile, J. Microencapsul 16 (1999) 105–115.

[6] N. Faisant, J. Siepmann, J.P. Benoit, PLGA-based microparticles:elucidation of mechanisms and a new, simple mathematical modelquantifying drug release, Eur. J. Pharm. Sci. 15 (2002) 355–366.

[7] B.S. Zolnik, P.E.Leary,D.J. Burgess, Elevated temperature accelerated releasetesting of PLGA microspheres, J. Control. Release 112 (2006) 293–300.

[8] R. Baker, Introduction, in: R. Baker (Ed.), Controlled Release of BiologicallyActive Agents, Vol. AWiley-Interscience Publication, 1987, pp. 1–18.

[9] J. Siepmann, A. Gopferich, Mathematical modeling of bioerodible,polymeric drug delivery systems, Adv. Drug Deliv. Rev. 48 (2001) 229–247.

[10] I. Galeska, T.-K. Kim, S. Patil, U. Bhardwaj, D. Chatttopadhyay, F.Papadimitrakopoulos, D.J. Burgess, Controlled release of dexamethasonefrom PLGA Microspheres Embedded within Polyacid-Containing PVAHydrogels, AAPS J. 7 (2005) (Article 22).

[11] K. Fu, D.W. Pack, A.M. Klibanov, R. Langer, Visual evidence of acidicenvironment within degrading poly(lactic-co-glycolic acid) (PLGA)microspheres, Pharm. Res. 17 (2000) 100–106.

[12] I. Grizzi, H. Garreau, S. Li, M. Vert, Hydrolytic degradation of devices basedon poly(DL-lactic acid) size-dependence, Biomaterials 16 (1995) 305–311.

[13] M. Vert, S. Li, H. Garreau, More about the degradation of LA/GA-derivedmatrixes in aqueous media, J. Control. Release 16 (1991) 15–26.

[14] A. Shenderova, T.G. Burke, S.P. Schwendeman, The acidic microclimatein poly(lactide-co-glycolide) microspheres stabilizes camptothecins,Pharm. Res. 16 (1999) 241–248.

[15] A. Shenderova, A.G. Ding, S.P. Schwendeman, Potentiometric method fordetermination of microclimate pH in poly(lactic-co-glycolic acid) films,Macromolecules 37 (2004) 10052–10058.

[16] P.A. Burke, Determination of internal pH in PLGA microspheres using31P NMR spectroscopy, Proceedings of the International Symposium onControlled Release of Bioactive Materials 23rd, 1996, pp. 133–134.

[17] A. Brunner, K. Mader, A. Gopferich, pH and osmotic pressure insidebiodegradablemicrospheres during erosion, Pharm.Res. 16 (1999) 847–853.

[18] K. Mader, B. Gallez, K.J. Liu, H.M. Swartz, Non-invasive in vivocharacterization of release processes in biodegradable polymers by low-frequency electron paramagnetic resonance spectroscopy, Biomaterials 17(1996) 457–461.

[19] D.J. Burgess, D.J.A. Crommelin, A.J. Hussain, M.-L. Chen, EUFEPSworkshop report, assuring quality and performance of sustained andcontrolled release parenterals, Eur. J. Pharm. Sci. 21 (2004) 679–690.

[20] M. Spangler, E.Mularz, Avalidated, stability-indicatingmethod for the assayof dexamethasone in drug substance and drug product analyses, and the assayof preservatives in drug product, Chromatographia 54 (2001) 329–334.

[21] Y.-J. Fu, S.-S. Shyu, F.-H. Su, P.-C. Yu, Development of biodegradable co-poly(-lactic/glycolic acid) microspheres for the controlled release of 5-FU bythe spray drying method, Colloids Surf., B Biointerfaces 25 (2002) 269–279.

[22] B.S. Zolnik, J.L. Raton, D.J. Burgess, Application of USP apparatus 4 andin situ fiber optic analysis to microsphere release testing, DissolutionTechnol. 12 (2005) 11–14.

[23] R.T. Liggins, H.M. Burt, Paclitaxel loaded poly(-lactic acid) (PLLA)microspheres: II. The effect of processing parameters on microspheremorphology and drug release kinetics, Int. J. Pharm. 281 (2004) 103–106.

[24] M. Miyajima, A. Koshika, J. Okada, M. Ikeda, Effect of polymer/basicdrug interactions on the two-stage diffusion-controlled release from a poly(-lactic acid) matrix, J. Control. Release 61 (1999) 295–304.

[25] L. Li, S.P. Schwendeman, Mapping neutral microclimate pH in PLGAmicrospheres, J. Control. Release 101 (2005) 163–173.

[26] B. Bittner, C. Witt, K. Mader, T. Kissel, Degradation and protein releaseproperties of microspheres prepared from biodegradable poly(lactide-co-glycolide) and ABA triblock copolymers: influence of buffer media onpolymer erosion and bovine serum albumin release, J. Control. Release 60(1999) 297–309.

[27] K. Makino, H. Ohshima, T. Kondo, Mechanism of hydrolytic degradationof poly(lactide) microcapsules: effects of pH, ionic strength and bufferconcentration, J. Microencapsul 3 (1986) 203–212.

[28] N. Faisant, J. Akiki, F. Siepmann, J.P. Benoit, J. Siepmann, Effects of thetype of release medium on drug release from PLGA-based microparticles:experiment and theory, Int. J. Pharm. 314 (2006) 189–197.

[29] B.C. Clark, P.A. Dickinson, I.T. Pyrah, Case study: in vitro/in vivo releasefrom injectable microspheres, in: D.J. Burgess (Ed.), Injectable DispersedSystems: Formulation, Processing and Performance, vol. 149, Taylor &Francis, Boca Raton, 2005, pp. 543–570.