The present invention relates to a system for controlled release of active ingredients and especially for controlled drug release. In particular, it relates to a system based on biodegradable polymer microparticles for controlled release of active ingredients such as drugs and therapeutics, including low molecular weight drugs, and biotherapeutics such as DNA, RNA, proteins, antigens and peptides.

Wherever "controlled release" is mentioned, it should be understood as dealing with the release of active compounds in a controlled way and over predetermined periods of time, such as over days, weeks, or months. It encompasses slow and sustained release.

Parenteral dosage forms with slow drug release properties have been developed to answer the need for improving the therapeutic use of drug substances which cannot be or are less suitable to be administered orally due to their physicochemical properties, and which have a relatively short half life because of which they have to be injected frequently. Frequent injections are uncomfortable to patients, and if the injections have to be given by physicians or nurses, they are also rather costly.

Among the biodegradable polymers suitable for preparing injectable microparticles, most experience has been gained with copolymers of lactic and glycolic acid - poly (lactic-co-glycolic acid) (PLGA). These copolymers degrade in physiological conditions through hydrolysis of the ester bonds to form lactic acid and glycolic acid. These degradation products are endogenous compounds and are metabolized via biochemical pathways. The release of entrapped compounds can be controlled by diffusion of the drug through the matrix and/or by degradation of the matrix. The rate of diffusion depends in particular for loaded biotherapeutics essentially on the porosity of the microparticles, while the rate of degradation depends, among other things, on the lactic acid/glycolic acid ratio. The release rate can be tailored according to the desired release pattern of the biotherapeutical, such as a protein, to be incorporated.

Products such as proteins, peptides, pDNA and siRNA are potent molecules for the treatment of patients suffering from different chronic and life-threatening diseases. Since oral administration of these generally labile molecules almost without exception does not result in therapeutic effects, these molecules are almost exclusively administered to patients via injection. In the last 25 years tremendous efforts have been given to the development of systems which give sustained release of therapeutic proteins for a predetermined time. As said, in particular, poly(lactic-co-glycolic acid) (PLGA) has been investigated frequently as protein delivery system, e.g. in the form of implants, microparticles and nanospheres.

Particularly, these PLGA polymers can be tailored to achieve release times from 1-2 months to 1-2 years. For some applications, however, a shorter release time is requested.

However, a short release time of PLGA systems is generally associated with a high initial burst. A high initial burst means a fast release of a drug in a burst stage. While the initial burst is not always harmful and can be utilised in certain specific drug administration strategies, excessive drug release in the burst phase may be toxic, and is also economically inefficient. Further, many drugs that are good candidates for sustained release treatments are not compatible with PLGA and common microparticle processing solvents, and, as a result, suffer from excessive initial burst release. The initial burst and release time of a microparticle loaded with an active compound are dependent, among other things, on the porosity of the PLGA-microparticles. Porous particles are characterised by a high initial burst and a fast release, while non-porous particles allow active compounds to be released slowly and with a minimal initial burst. On the other hand, the release of an active compound from the microparticles is also governed by the degradation rate of the polymer matrix, which is, in turn, dependent on the composition of the polymer.

In addition, instability of the active ingredient, such as protein instability, has become recognized as a major problem of PLGA systems, because it leads to an incomplete release of the entrapped active ingredient. Denaturated or aggregated protein species, for instance, not only become therapeutically inactive, but may also cause unpredictable side effects, such as immunogenicity or toxicity.

An example of a developed product based on PLGA is Nutropin Depot ^® , a sustained release system for human growth hormone which has been withdrawn from the market, due to "uncertainties and limitations in product supply required to meet future patient needs as well as the significant resources necessary for manufacturing". This illustrates the difficulties encountered using this polymer for the design of sustained release formulations. This means that there are severe drawbacks associated with the use of PLGA as protein release system. As said, these include incomplete and difficult-to-tailor release of the entrapped active ingredient such as a protein. Importantly, during degradation of the polymers low molecular weight degradation products accumulate in the matrix which causes a drop in the pH in the matrix. This low pH might induce unfolding/aggregation of the entrapped protein which might contribute to the frequently observed incomplete release of proteins form PLGA particles. Further, the protein aggregates may cause unpredictable and highly unwanted side effects, such as immunogenicity and toxicity. Finally, it has been reported that this low matrix pH catalyzes the formation of amide bonds between carboxyl groups of the degrading polymer and primary amines of proteins. Chemical modification of proteins is of course unwanted.

Several approaches have been investigated in order to tailor release profiles of proteins from PLGA particles and to prevent protein denaturation/aggregation. For instance, pegylated proteins have been entrapped and urea has been added to the formulation both to prevent protein aggregation. Also, additives like, Mg(O H) 2 have been co-entrapped to prevent acidification of the degrading matrix. Although some progress has been made, it is obvious that these routes are not the general solution for the protein/PLGA incompatibility and other routes need to be explored.

Hence there exists a need for developing controlled release systems which do not have the above-mentioned deficiencies. A desired system should release the incorporated drug substances in intact form, and preferably without a considerable initial burst in a short release time. Leemhuis et al. (in Marcomolecules 39 (10), (2006) 2500-2508; and in

Biomacromolecules 8 (2007) 2943-2949) described new hydrophilic polyesters with pendant hydroxyl groups. The synthetic route to prepare specific hydroxylated poly (α-hydroxy acids) is depicted in the scheme here-below and allows the design of a great variety of (co)polymers differing in composition, molecular weight and hydrophilicity. The degradation characteristics of these polymers can be tailored from a few hours to two months by the copolymers composition.

Particularly, Leemhuis et al. use ring-opening polymerization of functionalized dilactones with protected hydroxyl groups, and more specifically ring opening of benzyloxymethylmethylglycolide (BMMG) and of benzoyloxymethylglycolide (BMG), using the benzyloxy group as protecting group.

In preliminary tests, the group of inventors loaded microparticles made of hydroxylated poly (α-hydroxy acids) with lysozyme in the hope to get a system wherein actives would be released in a controlled way over a short period of time. Unfortunately, however, high initial bursts of 25-30% of the loaded lysozyme were observed; the release of protein could not be predicted. On the basis of a lot of research efforts, the inventors succeeded in providing a system for controlled drug delivery, comprising microparticles of low porosity and at least one active compound stably incorporated in these microparticles, said microparticles being based on polymers wherein the system has an initial burst of less than 10%, preferably less than 7%, more preferably less than 5%, and highly preferably less than 3%, and the total release of the active compound after 60 days is more than 50%, preferably more than 60%. These systems, e.g., can be obtained by a double emulsion extraction evaporation method with a volume ratio of dispersed phase to continuous phase equal to or higher than 1/10, preferably higher than 1/5.

Hence, in a first aspect, the present invention relates to a system for controlled release of active compounds or compositions, for instance of pharmaceutical compounds, comprising microparticles having a low porosity and at least one active compound incorporated in the microparticles, said microparticles being based on hydrophilic polyesters having pendant hydroxyl groups, and preferably on hydroxylated poly(alpha-hydroxy acids) obtainable from a water-in-oil-in- water emulsion wherein the volume ratio of the dispersed phase to the continuous phase is higher than or equal to 1/10, preferably higher than or equal to 1/5; said system having an initial burst of active compound of less than 10%, preferably less than 7% and more preferably less than5%, and most preferably less than 3%, and a total release after a period of 60 days of more than 50%, preferably more than 60%. Highly preferred is the embodiment wherein the microparticles are obtained from said emulsion wherein the concentration of the polymer in the organic (the oil) phase of the emulsion is higher than 10%, preferably higher than 15%.

Microparticles are defined herein as substantially solid or semisolid particles, having a weight- and volume average diameter in the region of about 0.1 to about 1000 μm, but usually of about 1 to about 500 μm, regardless of their composition, geometrical shape, or internal structure. For example, spherical microparticles, which are often referred to as microspheres or nanospheres, are included in the term microparticles, just as capsular structures, such as micro- and nanocapsules. Several other synonyms may be used to describe microparticles as defined above. A suitable method for determining the particle size is using an optical particle sizer such as Accusizer ™ 780.

A polymer is defined by IUPAC nomenclature as a substance composed of macromolecules. Macromolecules, or polymer molecules, are individual molecules of high relative molecular mass, the structure of which essentially comprises the multiple repetition of units derived. As used herein, the term polymer may be referred also to a polymeric material which can also be a copolymer, a mixture of polymers, a cross-linked polymer, and mixtures thereof. Often, a polymeric material is simply referred to as a polymer.

Suitable polymers to be used in the invention can be homo- or copolymers of aliphatic polyesters based on alpha-hydroxyacids. In a preferred embodiment the polymer is poly-(lactic-co-hydroxymethyl glycolic acid) (PLHMGA) which is synthesized using 3S-(benzyloxymethyl)-6S-methyl-l,4- dioxane-2,5-dione (BMMG) as a monomer. The monomer BMMG is polymerised alone to form a homopolymer of BMMG, or together with L-lactide or D, L-lactide to form a copolymer with a different lactic/glycolic acid ratio. The synthesis of BMMG and the copolymerisation with lactide and especially L-lactide is described in Leemhuis et al., Macromolecules, 39 (2206) 3500-3508. In a typical procedure, the monomer ratio used is 35/65 % mol/mol for BMMG and L-lactide respectively, but other ratios can also be used. Typically, 25-50 mole% of BMMG is used to obtain microparticles with a degradation time of about 1-2 months. A higher content of BMMG and/or replacing part of BMMG by 3S-(benzyloxymethyl)-l,4-dioxane-2,5-dione (BMG) as a monomer leads to a faster release.

By adjusting the ratio between hydrophilic and hydrophobic monomers control of the degradation rate and the degradation pathway (surface erosion or bulk degradation) can be achieved. It was found by the inventors that the degradation of the homopolymer of BMMG, i.e. without lactide, was governed by surface erosion, while the degradation of copolymers of BMMG with L-lactide (starting from 25 mol% of L-lactide) was mainly driven by bulk erosion. The homopolymer of 3S-(benzyloxymethyl)-l,4-dioxane- 2,5-dione (BMG) and a copolymer with 25 mol% of L-lactide degrade through surface erosion. Higher amounts of L-lactide, e.g. 50 mol% and 75 mol%, lead to the bulk erosion.

In a preferred embodiment, the microparticles are based on the polymer poly(lactic acid-co-hydroxymethyl glycolic acid) and even more preferred on such a polymer containing 50-90 mole.%, preferably 65-85 mole. % lactic acid moieties. Other monomers known in the art to be applied in PLGA systems, can also be used, for example, ε-caprolactone.

Typical molecular weights of the polymer is used are approximately 10-30 kDa (M _n ). Such polymers may have an M _w /M _n ratio of about 2, where M _w is the weight averaged molecular weight and M _n is the number averaged molecular weight. These values can be determined using gel permeation chromatography (GPC). A higher molecular weight generally leads to a slower degradation characteristics.

It was found by the inventors that, when the microparticles are obtained from a water-in-oil-in water (w/o/w) emulsion, their porosity was sufficiently low to be useful in the present invention (that is: high burst releases may be avoided) in case of high volume ratio of dispersed phase (VDP) to continuous phase (VCP): VDP/VCP. In particular, best results were obtained with VDP/VCP of 1/10 and higher. In addition, the best results were obtained when using polymer solutions with polymer concentrations of higher than 10% such as higher than 12%, preferably higher than 14%. Unless otherwise indicated, percentages are in this description percentages by weight drawn to the weight of the final composition in the emulsion.

In the preferred method of the present invention suitable oil phases are organic solvents wherein the polymers used in accordance with the invention are soluble, and which do essentially not miscible with water. Suitable examples are chloroform or dichloromethane.

In a second aspect, the invention relates to a process for preparing a system according to the invention. This process comprises (i) obtaining microparticles of hydrophilic polyesters having pendant hydroxyl groups and preferably of hydroxylated poly(alpha-hydroxy acid) from a water-in-oil-in- water emulsion wherein the concentration of the polymer in the organic phase is higher than 10%, preferably higher than 15%; and wherein the volume ratio of the dispersed phase to the continuous phase is higher than or equal to 1/10, preferably higher than or equal to 1/5; and (ii) loading of the microparticles with at least one active compound.

In a third aspect, the invention is directed to the use of microparticles preparable by the process that forms the second aspect of the invention as system for the controlled release of the active compounds and preferably of drugs. The particle porosity was visually estimated by analysing the surface morphology of the microparticles on the scanning electron microscope pictures.

As test systems, Dextran Blue (as a macromolecular model compound) loaded PLHMGA and PLGA (control formulation) microspheres were prepared by a solvent evaporation technique. The Dextran Blue-loaded PLHMGA microspheres prepared with a 10% polymer solution showed, because of a high porosity, a high burst release (35-75%) and the remaining content was released in a sustained manner for 15-20 days. The microspheres prepared with 15 and 20% polymer solution had a lower porosity and showed a pulsed release after day 8 and in 27 days they released more than 90% of Blue Dextran. It was also found that active ingredients such as proteins incorporated in the system of the invention show high stability and are generally not modified during or otherwise affected by the degradation of the polymer. In addition, the degradation time of these polyesters is much shorter than those of PLGA and could be tailored by the copolymer composition; that is, by the ratio of lactic acid to the α-hydroxy acid, such as the monomer BMMG. The degradation time for, e.g., the different compositions of PLHMGA may range from a few hours to two months, in comparison with 2-4 months for PLGA According to the invention, the polymer microparticles comprise at least one active compound. An active compound is any chemical or biological substance or a mixture of substances which is useful for diagnosis, prevention or treatment of diseases, symptoms, and other conditions of the body, or for influencing body function. To be an active system, the system comprises at least one, and optionally two or more of such active compounds.

Preferred active compounds are those which are used in chronic and long-term treatments and/or which have a low oral bioavailability, such as hormones, growth factors, hormone antagonists, antipsychotics, antidepressants, cardiovascular drugs, and the like. In another aspect, a preferred class of active compounds is that of peptides and proteins. Examples of (recombinant) proteins, which are considered very interesting from a pharmacological point of view, are cytokines, such as interleukines, interferons, HgH, EPO, IgG's, tumor necrosis factor (TNF), insulin, proteins for use in vaccines, growth hormones and other known pharmaceutically interesting proteins as described in, e.g., Pharmaceutical Biotechnology, Ed. D.J.A. Crommelin, R.D. Sinclair and B. Meiborn, Informa, THM Editiion (2008). Other preferred active compounds are polysaccharides, and oligo- or polynucleotides, DNA, RNA, iRNA, antibiotics, and living cells.

The following non-limiting examples illustrate the invention and do not limit its scope in any way. In the examples and throughout this specification, all percentages, parts and ratios are by weight unless indicated otherwise.

In the drawings, Fig. 1 shows scanning electron micrographs of PLHMGA and PLGA microspheres prepared from a 15% polymer solution in dichloromethane (DCM) and different VDP/VCP ratios: (a,b) PLHMGA microspheres with a VDP/VCP of 1/100; (c,d) PLHMGA microspheres with a VDP/VCP of 1/10; and (e,f) PLGA microspheres with a VDP/VCP of 1/10; Fig. 2 (a) shows the relative mass decrease of PLHMGA microspheres, and Fig. 2 (b) shows the number average molecular weight (M _n ) decrease of PLHMGA as a function of time; the microspheres were prepared with an initial polymer concentration in DCM of (■) 10%; (A) 15% and (•) 20% (w/w); degradation studies were done in PBS at 37 ⁰ C; Fig. 3 shows Η-NMR spectra pf PLHMGA (top) and of the insoluble residues isolated after 50 days of degradation (bottom); samples were dissolved in

DMSO;

Fig. 4 shows DSC thermograms of PLHMGA microspheres (made from a 15%

(w/w) polymer solution) before degradation and after 3, 7, 15 and 31 days of degradation; ΔH values of the melting peak are 19,39 and 50 J/g for day 7, 15 and 31, respectively;

Fig. 5 shows the Dextran Blue release (a) and lysozyme release (b) from

PLHMGA microspheres in PBS (VDP/VCP of 1/10, n=3) prepared from a 10 (■),

15 (A) and 20% (T) polymer solution; PLGA microspheres (♦) ; and Fig. 6 shows the release of lysozyme (n = 3) from PLHMGA microspheres prepared from a 15% polymer solution in DSM at VDP/VCP = 1/100 (■) and

VDP/VCP = 1/10 (A).

Example 1. Synthesis of copolymers of 3S-(benzyloxymethyl)-6S-methyl-l, 4- dioxane-2, 5-dione with L-lactide

3S-(benzyloxymethyl)-6S-methyl-l, 4-dioxane-2, 5-dione (BMMG, Fig. 2) was synthesized according to Leemhuis et al. Macromolecules 39 (2006) 3500-3508

H ₃

Synthesis of hydrophilic aliphatic polyesters based on lactic acid and glycoljc acid with pendant hydroxy! groups, poly (lactic-co-hydroxymethyl glycoϋc acid)

In a typical procedure, a copolymer of BMMG and L-lactide (monomer ratio 35/65 % mol/mol) was synthesized by melt copolymerization as follows: BMMG (960 mg, 3.8 mmol) and L-lactide (1027 mg, 7.1 mmol) were transferred into a dried Schlenk tube under a dry nitrogen atmosphere. Benzyl alcohol (11.86 mg; 50 μL from a 237 mg/ml toluene stock) and Snθct2 (22.2 mg; 100 μL from a 222 mg/ml toluene stock solution) were added. Toluene was removed by putting the tube under vacuum for 2 h. The tube was closed and heated to 110 ⁰ C using an oil bath for 16 h while stirring. After cooling to room temperature, the formed copolymer was dissolved in around 5 ml chloroform, precipitated into 250 ml of cold methanol and vacuum dried after filtration to give poly (lactic acid- ran -benzyloxymethyl glycolic acid) (PLBMGA) as a white solid (1.95 g). The copolymer was then dissolved in distilled THF (300 ml) and 10% w/w of Pd/C was added to remove the protecting benzyl group. After stirring at room temperature for 16 h under hydrogen atmosphere, the catalyst was removed by using a Hyflo filter. Evaporation in vacuo yielded 1.55 g of poly (lactic-co-hydroxymethyl glycolic acid) (PLHMGA).

PLBMGA was obtained in almost quantitative yield and the copolymer composition as determined by NMR matched the feed ratio (Table 1).

Table I. Characteristics of PLBMGA and PLHMGA

Feed Copolymer Yield M _n M _n ratio composition b

% theoretical meusured M _w /M _n Tg

(NMR

MJV analysis) (kg/mol) (kg/mol) ( ⁰ C)

PLBMGA 35/65 38/62 95 19 14 1.4 35

PLHMGA 35/65 32/68 79 15 1 1 1.3 38

³ M = monomer BMMG, L = lactide; ^b Determined with GPC using the following calibration standards: polystyrene for PLBMGA (solvent: THF) and PEG for PLHMGA (solvent: DMF/LiCI)

Further, DSC analysis showed that PLBMGA was completely amorphous with a T _g of 35 ⁰ C. The protecting benzyl group of BHMMG was removed by catalytic hydrogenation using Pd/C as catalyst. ¹ H NMR analysis confirmed complete removal of the protecting group by the disappearance of the peak at ~ 7.3 ppm. The deprotected copolymer (abbreviated as PLHMGA) was obtained in a high yield (~ 80%) with a T _g of 38 ⁰ C and the copolymer compositions (determined by NMR analysis) were close to the feed ratio. The M _n of the deprotected copolymer was around the same as for the protected polymer, indicating that no chain scission had occurred during deprotection. ¹ H NMR measurements of PLBMGA and PLHMGA dissolved in CDCl ₃ were performed at 298 K on a Varian Gemini-300 NMR machine, operating at 300 MHz.

Differential scanning calorimetry (DSC) analysis was done on a TA Instruments DSC Q2000 machine. PLBMGA and PLHMGA (approximately 5 mg) were loaded into aluminum pans and heated from room temperature to 200 ⁰ C at a heating rate of 5 °C/min. next, the samples were cooled down to -50 ⁰ C and heated to 200 ⁰ C with a rate of 5 °C/min. the melting temperature ™ was determined from the endothermic peak of the DSC thermogram in the first heating scan and the glass transition temperature (T _g ) was determined from the endothermic peak of the DSC thermogram in the first heating scan and the glass transition temperature (T _g ) was determined from the thermogram recorded in the second heating scan.

Gel permeation chromatography (GPC) was carried out on a Waters Aliance system, with a Waters 2695 separating module and a Waters 2414 Refractive Index detector. Two PL-gel 5 μm mixed-D columns fitted with a guard column (Polymer Labs, M _w range 0.2-400 kDa) were used. Calibration was done with polystyrene standards using THF as the mobile phase (1 ml/min) for PLBMGA. For the GPC analysis of PLHMGA, a 10 mM solution of LiCl in DMF with the flow rate of 0.7 ml/min was used as eluent and the columns were calibrated with PEG standards (see the above-cited article of Leemhuis et al.).

Example 2. Preparation of the microspheres.

Dextran Blue loaded microspheres of PLHMGA of PLGA were prepared essentially as described by Wang et al. in J. Control ReI. 82 (2002), 289-307. In short, 200 μl of Dextran Blue solution in water (50 mg/ml) were mixed either with 800 μl of a solution of 10, 15 and 20% (w/w) PLHMGA solution in dichloromethane (DCM) or 15% (w/w) PLGA in the same solvent. The water/DCM two phase system was emulsified by using an IKA homogenizer (IKA Labortechnik Staufen, Germany) for 30 s at the highest speed.

Subsequently, 800 μl of 1% PVA solution were added and the mixture was vortexed for 30 s at the maximum speed. The resulting w/o/w emulsion was then transferred into 8 ml of an aqueous PVA (0.5% w/w) solution and stirred for 1 hour at room temperature to evaporate DCM. After that the microspheres were separated by centrifugation (Laboratory centrifuges, 4K 15, Germany) at 5000 g for 5 min, washed 3 times with 100 ml reversed osmosis water and freeze dried overnight.

Lysozyme loaded microspheres were prepared using the same method as the Dextran Blue-loaded microspheres except that the resulting w/o/w emulsion was transferred into either 80 ml or 8 ml of an aqueous PVA (0.5% w/w) solution (VDP/VCP were 1/100 and 1/10, respectively; VDP = volume of the dispersed phase: VDP = the volume of the continuous phase) and stirred for 1 hour at room temperature to evaporate DCM. In addition, blank microspheres of PLHMGA were prepared with a VDP/VCP of 1/10. The average size and size distribution of the microspheres were measured using an Accusizer ^th 780 (Optical particle sizer, SantaBarbara, Calif ormia, USA). The morphology of the microspheres was studied by scanning electron microscopy (Phenom™, FEI Company, the Netherlands). Microspheres were glued on 12 mm diameter aluminum sample holder using conductive carbon paint (Agar scientific Ltd., England) and coated with palladium under vacuum using an ion coater.

The loading efficiency (LE) of Dextran Blue loaded microspheres was investigated by dissolving 20 mg of microspheres in DMSO and measuring the absorbance at 620 mm. Calibration was done with Dextran Blue dissolved in DMSO (concentration ranging from 10-200 μg/ml).

The lysozyme loading efficiency was determined by a micro-BCA method (Hongkee, J. Pharm. SCI-US, 86.(1997), 1325-1318). Briefly, 20 mg of freeze- dried microspheres were dissolved in 2 ml DMSO and 10 ml of a 0.05 M NaOH solution containing 0.5% (w/v) SDS were added and incubated at room temperature for 1 hour. Calibration was done with lysozyme dissolved in DMSO/0.05 M NaOH (volume ratio 1/5, lysozyme concentration ranging from 10-200 μg/ml) also containing 0.5% (w/v) SDS. Twenty five μl of standards and samples were pipetted into a 96 well plate and 200 μl of a BCA working reagent (a freshly prepared solution of 50 parts of a BCA stock solution with 1 part of 4% CUSO4.5H2O) were added. The plates were incubated for 2 hours at 37 ⁰ C. The absorbance of the solutions in the wells was measured by a Novapath Microplate Reader at a wavelength of 550 nm.

Dextran Blue and lysozyme-loaded PLHMGA as well as blank microspheres were thus prepared using a double emulsion extraction- evaporation method. Table 2 summarizes the characteristics of obtained microspheres. ,. _ , concentration in Volume weight _τ _ ,„,., _{τ n /n/\} i- Yield % ingredient Polymer „„.,.., , . . ° LE (%) LC (%) t b ^J DCM %(w/w) mean(μm) ^x ' ^x '

10 5.0 ±0.5 50 ±4 5.5 ±0.5 75.8 ±0.8 extran Blue PLHMGA 15 7.0 ±0.5 84 ±4 6.0 ±0.1 73.3 ± 1.6

20 8.0 ±1.0 85 ±5 4.3 ±0.2 73.3 ±2.4 extran Blue PLGA 15 13.0 ±2.0 72 ±3 4.7 ±0.1 78.9 ± 1.2 lysozyme ^¥ PLHMGA 15 10.5 ± 0.2 78 ±3 5.0 ± 1.0 76.0 ±0.6

10 6.6 ± 1.1 66 ±1 7.5 ±0.2 73.3 ±2.3 lysozyme PLHMGA 15 7.7 ±0.3 81 ±5 5.1 ±0.4 78.7 ±1.7

20 12.4 ±0.1 85 ± 1 3.7 ±0.1 83.0 ± 1.2

lvsozvme PLGA 15 19.6 ± 1.6 67 ± 7 4.3 ±0.2 78.3 ±3.8

^¥ VDP/VCP = 1/100

^* LE (Loading Efficiency) expressed as encapsulated Dextran Blue or protein divided by the total amount of Dextran Blue or protein used for encapsulation. t LC (loading Capacity) expressed as encapsulated amount of Dextran Blue or protein divided by the total dry weight of the microspheres.

The volume weight mean diameter of the microspheres slightly increased from 5 to 8 μm for Dextran-Blue loaded microspheres and from 6 to 12 μm for lysozyme-loaded microspheres with increasing PLHMGA concentration in DCM. Likely, a higher viscosity of the oil phase (due to higher polymer concentration) results in larger emulsified droplets, which in turn yields larger microspheres. The loading efficiency (LE) of Dextran Blue and lysozyme increased from 50-60% to 85% with increasing polymer concentration. The VDP/VCP ratio had no significant effect on the LE but the particle size was slightly increased when this ratio was decreased from 1/10 to 1/100. The observed increase in LE with increasing polymer concentration can be explained as follows. First, at higher polymer concentration a faster precipitation of the polymer occurs during the solvent evaporation process, which in turn will retard diffusion of the protein into the continuous phase. Secondly, a higher viscosity of the polymer solution in DCM will reduce the protein mobility in the w/o emulsion droplets.

The surface morphology of lysozyme-loaded PLHMGA microspheres prepared at VDP/VCP 1/10 and 1/00 and with 15% w/w PLHMGA in DCM was studied by SEM analysis. Microspheres made a VDP/VCP of 1/10 were rather porous (Fig. 1 a,b), whereas micropsheres with a VDP/VCP of 1/10 were essentially non-porous (Fig. 1 c, d) The large pores in the microspheres prepared at a low VDP/VCP ratio can likely be explained by a rapid extraction of dichloromethane from the dispersed phase into the continuous phase, resulting in a faster polymer precipitation PLGA microspheres prepared at a VDP/VCP of 1/10 showed a surface morphology comparable to those of PLHMGA microspheres (Fig. 1, e, f). Example 3

Microspheres of PLHMGA incubated at 37 ⁰ C and pH 7.4 showed a continuous weight loss in time; after 50 days less than 20% of their initial weight remained (Fig. 2 a) This Fig. 2 shows that the polymer mass loss was independent of the polymer concentration in DCM used to prepare the

PLHMGA microspheres, which suggests that the microspheres degrade via bulk erosion. The M _n of the non- dissolved polymer as determined by GPC, showed a gradual decrease in time (Fig. 2b)

The degradation of the microspheres was not entirely complete in the time frame studied (60 days) and some insoluble residues remained. ¹ H NMR analysis of the residues showed that the duplet peak around 4 ppm attributed to the protons of the methylene group of hydroxymethyl glycolic acid had disappeared, and oligomers of lactic acid remained (Fig. 3). This means that during degradation preferential removal of the HMGA units occurred during degradation. Likely, due to the hydration of the hydroxyl groups of these units, hydrolysis preferentially occurs at the hydroxyl enriched sites in the polymer.

DSC analysis of PLHMGA microspheres showed that they were completely amorphous before degradation with a T _g of 25 ⁰ C (Fig. 4), while crystalline fragments of lactid acid, the degree of polymerization of these lactic acid oligomers can be estimated to be around 10.

Example 4. In vitro release of Dextran Blue and lysozyme from PLHMGA microspheres.

About 20 mg of Dextran Blue-loaded microspheres were suspended in 1.5 ml of phosphate buffered saline (PBS) and incubated at 37 ⁰ C under mild agitation. At different time points, the samples were centrifuged (5000 g for 1 min) and 0.5 ml of supernatant was removed and replaced by 0.5 ml of fresh buffer. The absorbance at 620 nm was measured and calibration was done with Dextran Blue in PBS (100-300 μg/ml).

For lysozyme loaded microspheres, 20 mg of microspheres were suspended in 1.5 ml of PBS and the sampling was done as for the Dextran Blue-loaded microspheres. The protein content in the supernatant was determined by ultra performance liquid chromatography (Acquity UPLC ^® ) using a BEH300 C18 1.7 μm column. Elution was performed at a flow rate of 0.25 ml/min, using a gradient starting with 75% eluent A (95%, H ₂ O, 5% CAN and 0.1% TFA) and 25% eluent B (100% CAN and 0.1% TFA). The eluent linearly changed to 60% A / 40% B during 4 min and the volume ratio was set back at 75% A / 25% B after 5 minutes. The injection volume was 5 μl and detection was done at 280 nm. A series of lysozyme standards (10-500 μg/ml) was used for calibration.

Fig 5 (a) shows the release of Dextran Blue from different PLHMGA microspheres in PBS. Microspheres made with 10% polymer solution in DCM, showed a high burst release (>70%) followed by a gradual release, reaching 90% of the loaded amount, over the next 14 days. Increasing the initial polymer concentration to 15 and 20 % resulted in microspheres showing a decreased initial burst of 20 and 10%, respectively. This lower burst release can be attributed to the lower porosity of these microspheres. After this small burst, hardly any Dextran Blue started to be released, reaching completeness (>90% of the loaded amount) in 27 days. In contrast, after a low burst release, PLGA microspheres showed hardly any release of Dextran Blue during 27 days. After 40 days, as polymer degradation progressed, they started to release Dextran Blue up to 65% in 65 days. The difference between Dextran Blue release from PLHMGA and PLGA microspheres maybe explained by the different degradation rate of PLHMGA and PLGA, which is more rapid for PLHMGA. Fig. 5 (b) shows that the release of lysozyme from PLHMGA microspheres also depended in the polymer concentration in the DCM phase used for the preparation of microspheres. The microspheres prepared from a 10% polymer solution showed a burst release of around 35% followed for the microspheres prepared from a 15 and 20% polymer solution, respectively, and these microspheres started to release the protein after a certain delay time (around 8-10 days). Thereafter, as microsphere degradation progressed, the protein was released with almost zero-order kinetics from the next 25 days. PLGA microspheres showed an initial burst release (7%) followed by a very slow release reaching 11% of the loaded amount after 60 days (Fig. 5 b).

Several reasons for the incomplete release of encapsulated proteins from PLGA microspheres may be given. Particularly, aggregation of the entrapped protein, chemical degradation or interaction between protein and polymer may occur. The effect of the rate of solvent removal on lysozyme release from the resulting PLHMGA microspheres (prepared from a 15% polymer solution) was investigated by comparing the release from microspheres prepared with a VDP/VCP of 1/10 and 1/100 (Fig. 6).

Microspheres prepared with a low VDP/VCP=1/100 showed a burst release of around 20% followed by a sustained release of lysozyme up to 70% of the loaded dose for 30 days. Microspheres with a higher VDP/VCP (1/10), in agreement with the Dextran Blue loaded microspheres, showed a low burst release (<5%) and 60% of the loaded dose was released in the next 30 days. This difference in release kinetics may be attributed to the porosity of the microspheres. Microspheres prepared at a VDP/VCP of 1/100 were rather porous (Fig. 1 a,b) in contrast to the microspheres prepared with a VDP/VCP of 1/10 (Fig. 1 c,d).

Fig. 5 (b) and 6 show that the release of lysozyme from PLHMGA microspheres prepared from PLHMGA solutions of 10, 15 and 20% in DCM and with different VDP/VCP ratios was not complete (reaching 50-70% of loaded amount). To investigate whether lysozyme was present in the insoluble residues present in the release samples after 60 days, the residues were washed twice with RO water and after freeze- drying, they were dissolved in DMSO (200 μl). After 1 hour, 1 ml of 0.05 M NaOH containing 0.5% (w/v) SDS was added and the amount of lysozyme was determined by the micro-BCA assay both in the water fraction and the DMSO/NaOH/SDS solution. Additionally, the presence of protein was studied by the FTIR analysis of the freeze dries residues using a BIO-RAD FTS6000 FT-IR (BIO-RAD, Cambridge, MA, USA) and KBr (H. Susi, D. Michael Byler, Biochem. and Biophys. Res. Comm., 115 (1983) 391). The FTIR spectra were measured at room temperature and a total of 32 scans at a resolution of 2 cm ¹ were averaged. The protein amount in both the wash fractions and the DMSO solutions was measured (Table 3).

specific lysozyme activity of polymer released from lysozyme extracted recovery polymer concentration microspheres from insoluble (%) in DCM %(w/w) after 60 days residues (%)

(%) (%)

10 73 ± 4 10 ± 4 S3 ± 8 1 11 ± 13

15 59 ± 1 20 ± 4 80 ± 4 103 ± 13

PLHMGA

15 ^* 72 ± 1 IS ± 4 91 ± 1 105 ± 6

20 52 ± 9 26 ± 1 78 ± 9 1 18 ± 3

PLGA 15 1 1 ± 1 68 ± 2 79 ± 2 102 ± 5

' VDP/V _CP = 1/100 (others with V _w /V _C p = 1/10)

Table 3 shows that the biological activity of the released lysozyme was fully preserved for all investigated samples. Further, the far-UV CD spectrum as well as the fluorescence emission spectrum of the released lysozyme overlapped almost completely with that of native lysozyme (results not shown), indicating that the secondary and tertiary structure of released its structural and functional integrity.