Application of ultrasound-assisted compression and 3D-printing semi-solid extrusion techniques to the development of sustained-release drug delivery systems based on a novel biodegradable aliphatic copolyester

Journal of Drug Delivery Science and Technology 95 (2024) 105652 Available online 9 April 2024 1773-2247/© 2024 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Application of ultrasound-assisted compression and 3D-printing semi-solid extrusion techniques to the development of sustained-release drug delivery systems based on a novel biodegradable aliphatic copolyester C. Ferrero a , * , L. Urpí b , A. Aguilar-de-Leyva a , G. Mora-Casta˜ no a , V. Linares a , M. Mill´ an- Jim´ enez a , A. Martínez de Ilarduya b , I. Caraballo a a Departamento de Farmacia y Tecnología Farmac´ eutica, Facultad de Farmacia, Universidad de Sevilla, C/ Prof. García Gonz´ alez 2, 41012, Sevilla, Spain b Departament d’Enginyeria Química, Universitat Polit` ecnica de Catalunya (ETSEIB), Diagonal 647, 08028, Barcelona, Spain ARTICLE INFO Keywords: Aliphatic copolyester Ultrasound-assisted compression 3D-printing semi-solid extrusion Metronidazole Drug delivery systems Sustained-release ABSTRACT The purpose of this study was to investigate the ability of two advanced hot-processing technologies, Ultrasound- Assisted Compression (USAC) and 3D-printing Semi-solid Extrusion (SSE), to manufacture sustained-release drug delivery systems based on a novel biodegradable aliphatic copolyester. The copolymer was synthesized from ω -pentadecalactone (PDL), 1,4-cyclohexanedimethanol (CHDM) and dimethyl succinate (DMS) (monomer ratio PDL/CHDM/DMS 70/30/30) by enzymatic ring opening polymerization and two-step melt polycondensation processes and has a random microstructure, a high molecular weight (107,100 g mol −1 ) and a relatively low melting point (~65 ◦C). The antibacterial agent metronidazole (MTZ) was chosen as model drug and binary physical mixtures copolymer:drug were prepared at 90:10 and 70:30 w/w ratios. Thermal analysis studies evidenced that the formulations could be processed below their degradation temperatures. Drug delivery devices with dense and meshed structures were manufactured using USAC and SSE techniques, respectively, with USAC devices exhibiting more reproducible physical properties than the SSE systems. Powder X-ray diffraction and scanning electron microscopy studies showed a partial sintering of the copolyester during USAC processing while MTZ remained mostly crystalline. In contrast, the copolymer melted and the drug underwent some amorphization when processed using SSE. In vitro drug release studies in phosphate buffer (pH 6.8) showed that, after an initial burst release of metronidazole, USAC and SSE devices exhibited a prolonged and/or sustained drug release over 20 days. The initial burst release was dependent on the manufacturing technique and the drug/polymer ratio, being minimized for SSE devices containing 10 wt% MTZ. The whole drug release profiles fitted well to the Peppas-Sahlin model, being drug diffusion the predominant release mechanism. After the burst release, the sustained release period of USAC and SSE devices containing 10 wt% MTZ showed a good fit to the zero-order kinetic model, with faster drug release from the SSE devices due to the larger surface area of their meshed structure. In conclusion, this work demonstrates that processing the aliphatic copolyester by both USAC and SSE technologies provides an attractive strategy to manufacture biodegradable drug delivery systems for long-term release of metronidazole. 1. Introduction Aliphatic polyesters are great materials for pharmaceutical and biomedical applications as they are biocompatible or/and biodegradable [1]. Thermal properties such as melting temperature can be controlled by the ratio of ester and methylene segments present in their repeating units [2]. Common used aliphatic polyesters are polyesters type AB, usually obtained by ring opening polymerization (ROP) of their corresponding lactones or dilactones, such as poly(glycolic acid), poly (lactic acid), poly(p-dioxanone) or poly( ε -caprolactone), and polyesters type AABB such as poly(butylene succinate), or poly(butylene adipate), obtained by two-step melt polycondensation process. On the other hand, copolymerization is an easy way to tune both melting and glass transition temperatures [3]. In a recent study carried out by us [4], a series of high molecular weight aliphatic copolyesters derived from ω -pentade- calactone (PDL), dimethyl succinate (DMS) and 1, * Corresponding author. E-mail address: [email protected] (C. Ferrero). Contents lists available at ScienceDirect Journal of Drug Delivery Science and Technology journal homepage: www.elsevier.com/locate/jddst https://doi.org/10.1016/j.jddst.2024.105652 Received 21 November 2023; Received in revised form 26 March 2024; Accepted 7 April 2024 Journal of Drug Delivery Science and Technology 95 (2024) 105652 2 4-cyclohexanedimethanol (1,4-CHDM) were synthesized by a combination of enzymatic ROP and two-step melt polycondensation using a green biocatalyst (Candida antarctica lipase B, CALB). It was demonstrated that the thermal and mechanical properties of these random copolyesters could be tailored by varying the copolyester composition. In the present work, we have made use of this synthetic methodology to prepare, on a larger scale, an aliphatic copolyester poly( ω -pentadeca- lactone-co-cyclohexylenedimethylene succinate) (PPDL 70 CHDMS 30 ) with a selected composition (PDL/CHDMS 70/30) and assess their processability using hot-processing techniques. Due to the relatively low melting temperature range (35–90 ◦C) and good mechanical properties of these copolymers, an ultrasound-assisted compression technique and a 3D-printing technique based on semi-solid extrusion were selected as thermoforming technologies. Ultrasound-Assisted Compression (USAC) is a technology that combines a conventional compression process and high-intensity ultrasound (US) energy (frequencies ≥20 kHz). The application of high-intensity US irradiation on solid materials induces thermal and mechanical effects, changing their physical properties, often permanently [5]. We have recently reviewed the principles of ultrasound-assisted compression and its main applications in pharmaceutical technology [6], including a detailed and illustrated description of the stages of the USAC process. Briefly, the USAC equipment is similar to a conventional small-scale tableting machine with an US generator-transducer-sonotrode coupled to the upper punch. At an initial stage, the pressure applied by the lower punch produces a pre-compaction of the sample, reducing its porosity. Then, ultrasound is applied by the upper punch increasing powder friction and the temperature inside the powder bed. Therefore, plasticity and interparticle bonding increases leading to the sintering of materials [6–8]. After a cooling stage period to reestablish the new solid properties of the sample, the compact is ejected. Physical properties of the materials, mainly glass transition (T g ) and melting (T m ) temperatures, and process parameters such as compaction time, US energy and inter-punch distances have a great influence on this technique [8–10]. As USAC involves a form of thermal processing, thermoplastic polymers with low T g and T m are especially suitable for the manufacture of drug delivery systems using this technology. However, a proper control of the process conditions is necessary to prevent drug degradation [6,8]. Different research papers have dealt with the application of ultrasound-assisted compression in pharmaceutical technology, mainly in the compression of pharmaceutical powders [10–12], preparation of solid dispersions to enhance bioavailability of poorly soluble drugs [8, 11,13,14] and development of sustained-release dosage forms [7, 15–19]. Due to the enhanced mechanical strength of the USAC devices compared to conventional direct-compressed tablets, drug release could be controlled with a smaller amount of excipient resulting in high-efficiency sustained-release systems [6,17,18]. Over the last few decades, three-dimensional (3D) printing technologies have emerged as an innovative and revolutionary tool to manufacture personalized dosage forms. 3D-printing is considered a promising digital alternative to conventional manufacturing processes due to its multiple advantages: ability to design complex geometries, flexibility in the formulation composition and accurate control of drug deposition, combination of multiple drugs in a single dose unit and customized drug release profiles, among others [20–23]. Among the 3D-printing technologies applicable in pharmaceutical drug development, extrusion-based techniques such as Fused Deposition Modeling (FDM), Semi-solid Extrusion (SSE) and Direct Powder Extrusion (DPE) are the most used due to their simplicity, versatility and costeffectiveness [23–31]. SSE technique involves the extrusion at room or moderate temperatures of a semi-solid or semi-molten material (gels, pastes or solids with relatively low melting point) using a syringe-based printhead operated by means of a pneumatic, mechanical (piston or screw) or solenoid system [24,30]. In the present study, a mechanical approach composed of a metal syringe-head nozzle tool and a piston was selected due to its simpler mechanism and greater control over the extrusion flow. The soft material extruded through the nozzle is deposited in sequential layers to print the designed model on the building platform. After printing, solidification occurs by cooling or drying depending on whether heat or solvents were used during the process [24,30,32]. Rheological properties of the materials, mainly viscosity, and printing parameters such as nozzle diameter, printing temperature, flow rate and printing speed should be carefully controlled to achieve a successful printing [22,24,30]. In the pharmaceutical field, SSE technology has found application in the development of oral dosage forms [33–37], rectal formulations [38, 39], implants and patches [31,40–42], among others. Compared with other extrusion-based technologies like FDM, SSE technique offers several advantages. SSE is a single-step process that avoids the need of filament preparation before printing, simplifying the manufacturing process and reducing production time. Moreover, the use of lower printing temperatures and the reduction of the thermal stress minimized the risk of drug degradation, being suitable for thermolabile drugs. The simplicity and printing speed of this technology also facilitates the on-demand production of small and individualized batches, although the low resolution of SSE printers is one of the main limitations for their future implementation at healthcare facilities [22,24,30,43]. To evaluate the performance of PPDL 70 CHDMS 30 copolyester as polymeric carrier in drug delivery systems, Metronidazole (MTZ) was selected as model drug. MTZ, a Class I drug according to the Biopharmaceutics Classification System (BCS), is one of the most widely used antibiotics due to its clinical effectiveness in the prevention and treatment of anaerobic infections [31,44,45]. Metronidazole can be administered both orally and locally, but oral administration causes serious side effects such as gastrointestinal problems and neurotoxicity. Hence, to overcome these adverse effects, local and extended-release formulations are preferred [31,44–46]. We hypothesized that the combination of the biodegradable copolyester PPDL 70 CHDMS 30 with the antimicrobial agent MTZ and their processing by USAC and SSE technologies would result in long-term pharmaceutical devices with potential application in the treatment of localized infections. Therefore, the objective of this research was to develop sustaineddrug delivery devices from the novel biodegradable PPDL 70 CHDMS 30 copolyester using advanced manufacturing technologies such as USAC and SSE. For this purpose, the aliphatic copolyester was synthesized by a combination of enzymatic ROP and two-step melt polycondensation processes and its physicochemical characteristics assessed by 1 H and 13 C NMR spectroscopies and GPC measurements. The antibiotic metronidazole was chosen as model drug. Thermal properties of the raw materials and physical mixtures copolymer:drug (90:10 and 70:30 w/w ratios) were evaluated before processing. Drug delivery devices with cylindrical dense and meshed structures were, respectively, manufactured by USAC and SSE techniques and their characterization performed by PXRD and SEM studies. In vitro drug release behaviour was also analyzed in detail fitting the experimental data to different kinetic models to elucidate the drug release mechanism from the designed devices. 2. Materials and methods 2.1. Materials The aliphatic copolyester PPDL 70 CHDMS 30 was synthesized as previously reported by us [4]. The reaction was scaled up for the synthesis of 50 g of copolyester. Briefly, PPDL 70 CHDMS 30 copolyester was prepared from ω -pentadecalactone (PDL), dimethyl succinate (DMS) and 1, 4-cyclohexanedimethanol (1,4-CHDM) in the stoichiometric amounts to obtain the 70/30 (PDL/CHDMS) composition. This composition was chosen because it has a relatively low melting temperature, which could allow easy processing by the thermoforming techniques used in this work. The synthesis was carried out by a combination of enzymatic ring opening polymerization (ROP) and two-step melt polycondensation C. Ferrero et al. Journal of Drug Delivery Science and Technology 95 (2024) 105652 3 processes in bulk using Novozyme 435 (Candida antarctica Lipase B immobilized on cross-linked polyacrylate beads) as biocatalyst, kindly donated by Novozymes (Bagsværd, Denmark). Although no cytotoxicity studies have been carried out on this copolyester, the use of enzymes in the polymerization avoids the presence of toxic metallic catalysts in the final product. Moreover, the hydrolytic degradation products have been found to be harmless, as reported in the literature [47–49]. Metronidazole (MTZ) PhEur (M w 171.15 g mol −1 , white to yellowish crystalline powder) (batch 210326), a Class I BCS antibacterial agent (water solubility 1 g/100 mL at 20 ◦C), was purchased from Acofarma Distribuci´ on S.A. (Barcelona, Spain). All other chemicals and solvents were of analytical grade. 2.2. PPDL 70 CHDMS 30 copolyester characterization The molecular weights and dispersity of PPDL 70 CHDMS 30 copolyester were determined using a Waters GPC instrument (Foster City, CA, USA) equipped with ultraviolet (UV) and refractive index (RI) detectors. Around 2 mg of polyester was dissolved in 1 mL of chloroform and filtered out with a 0.22 μ m PTFE filter. 100 μ l of this solution was injected to the system and eluted with chloroform at a flow rate of 0.5 mL/min. Two HR5E and HR2 Waters linear Styragel columns (7.8 ×300 mm, pore 10 3 –10 4 Å) packed with crosslinked polystyrene and protected with a precolumn were used. The calibration curve was generated using polystyrene (PS) standards with narrow molecular weight distributions. 1 H and 13 C NMR spectra were registered on a Bruker AMX-300 spectrometer (Billerica, MA, USA) at 25 ◦C. The spectrometer was operated at a frequency of 300.1 MHz for 1 H and at 75.5 MHz for 13 C spectra. 10 or 50 mg of sample were dissolved in 1 mL of a deuterated solvent (CDCl 3 ) for recording each type of spectra, using tetramethylsilane (TMS) as internal reference. A total of 128 scans were acquired for 1 H and around 5000 for 13 C. 2.3. Preparation of physical mixtures The PPDL 70 CHDMS 30 copolyester, in the form of thin flakes, was frozen using liquid nitrogen and milled in a Retsch ZM 200 ultra centrifugal mill (Haan, Germany), equipped with a 0.5 mm output sieve, at 6000 rpm. The fine powder obtained was subsequently dried in an oven at 37 ◦C for 24 h until constant weight. Metronidazole particle size fraction less than 250 μ m was selected after sieving. Binary physical mixtures of copolymer:drug at 90:10 and 70:30 w/w ratios were mixed for 10 min in a Turbula® T2F blender (Willy A. Bachofen, Basel, Switzerland). 2.4. Thermal analysis studies Thermal properties of the raw materials and physical mixtures were evaluated using Differential Scanning Calorimetry (DSC) and Thermogravimetric Analysis (TGA). DSC measurements were performed with a DSC Q20 equipment (TA Instruments, New Castle, USA), available at Centro de Investigaci´ on, Tecnología e Innovaci´ on (CITIUS) (Universidad de Sevilla). Accurately weighed samples (4–9 mg) were hermetically sealed in Tzero aluminum pans (40 μ L) and heated from −60 ◦C to 220 ◦C at a heating rate of 10 ◦C/min. Nitrogen was used as an inert purge gas with a flow rate of 100 mL/min. Calibration was performed with indium and bismuth as standards. TGA measurements were conducted on a SDT Q600 equipment (TA Instruments, New Castle, USA), also available at CITIUS (Universidad de Sevilla). Accurately weighed samples (9–11 mg) were introduced in open alumina crucibles (90 μ L) and heated from room temperature to 600 ◦C with a heating rate of 10 ◦C/min. Measurements were conducted under a nitrogen flow of 100 mL/min. The standards used for calibration were bismuth and zinc. Both DSC and TGA data were analyzed using Universal Analysis 2000 V4.5A software (TA Instruments, UK). From DSC data, the melting temperatures and melting enthalpies were calculated. The degradation temperatures (T d 10 % and T d max ) and remaining weight (%) were estimated from the TGA profiles. T d 10 % is defined as the temperature of 10 % weight loss in the TGA curve and T d max is the maximum rate degradation temperature evaluated by the peak in the derivative of the TGA curve. 2.5. Preparation of ultrasound-assisted compression (USAC) devices USAC devices (250 mg) were prepared from the bulk copolymer and the physical mixtures using a laboratory-scale ultrasound-assisted tableting machine (TECNEA USTM L20, Tecnea Engineering Srl, Bologna, Italy) equipped with 11 mm flat cylindrical punches and an US generator coupled to the upper punch. Preliminary studies (US energies in the range of 300–600 J) were performed to define the optimal ultrasound energy for each formulation and to ensure no drug degradation. An ultrasound energy of 500 J was applied to the bulk copolymer at a frequency of 20 kHz. This value was chosen to improve interparticle bonding and promote sintering of the copolyester. US energy values ≤ 300 J resulted in poor compact consolidation and values ≥600 J caused complete melting of the material. For the mixtures containing 10 % and 30 % w/w MTZ, the ultrasound energies were reduced to 400 and 350 J, respectively, to avoid drug degradation. Other parameters selected to achieve a proper compression were as follows: compression pressure 0.3 MPa, compaction time 3 s, cooling time 9 s and detach time 0.9 s. 2.6. Preparation of 3D-printed semi-solid extrusion (SSE) devices 3D-printed devices were manufactured by SSE with a REGEMAT3D V1 modular bioprinter (Regemat 3D S.L., Granada, Spain) customized with a high-temperature (up to 250 ◦C) metallic syringe equipped with a 0.6 mm nozzle. The software REGEMAT3D DESIGNER version 1.5.0 was used to design cylindrical 3D-printed systems of 10 mm diameter and 2 mm height. The 3D-printed devices consisted of 2 solid bottom layers, 1 sealing perimeter and a compartmental mesh size of 4 ×4 mm. The printing settings were as follows: syringe temperature 100 ◦C, nozzle temperature 110 ◦C, layer height 0.2 mm, flow speed 0.1 mm/s, perimeter speed 0.6 mm/s, infill pattern diagonal, infill speed 1 mm/s, retract 10 μ L and retract speed 20 μ L/s. 2.7. Physical characterization of USAC and 3D-printed SSE devices Weight uniformity of the USAC and 3D-printed SSE devices was assessed by individually weighing five tablets per batch in a precision balance Shimadzu AUW120 (Manila, Philippines). Physical dimensions (diameter and thickness) of the devices (n =5) were also measured using a digital micrometer (Comecta, SA, Barcelona, Spain). The crushing strength of the devices could not be determined due to the elastomeric properties of the copolymer. Digital images were also acquired to describe the aspect of the devices. 2.8. Powder X-ray diffraction (PXRD) studies X-ray diffraction patterns of the raw materials and USAC and 3D- printed SSE devices were collected with a Bruker D8 Advance A25 diffractometer (Karlsruhe, Germany) available at CITIUS (Universidad de Sevilla) and equipped with a lineal LynxEye™ detector with an opening of 3.3◦. A Cu anode X-ray tube operating at 40 kV and 30 mA was used to generate the X-rays. The data were recorded in the range of 3◦–70◦(2θ) with an angle step width of 0.015◦and a measurement time of 0.1 s per step. 2.9. Scanning electron microscopy (SEM) studies The surface morphology of the raw materials and the USAC and 3D- printed SSE devices was examined at the Microscopy Service of CITIUS C. Ferrero et al. Journal of Drug Delivery Science and Technology 95 (2024) 105652 4 (Universidad de Sevilla) using a High Resolution Scanning Electron Microscope (FEGSEM) FEI TENEO (FEI Company, Hillsboro, OR, USA), operating at 5 kV and 0.40 nA. Samples were mounted on aluminum stubs using carbon double adhesive tape and then coated with a 10 nmthin Pt layer using a Leica EM ACE600 high vacuum sputter coater (Vienna, Austria). The images were acquired at low and high magnifications. 2.10. In vitro drug release studies The in vitro drug release studies from USAC and 3D-printed SSE devices (n =3) were conducted using a USP Apparatus II (Agilent 708-DS, Singapore, Malaysia) with a paddle rotation speed of 75 rpm. Experiments were run at 37 ±0.5 ◦C using 500 mL of compendial phosphate buffer (pH 6.8) as dissolution medium [50]. Samples (5 mL) were withdrawn at predetermined time intervals, filtered (0.45 μ m) and analyzed by UV–Vis spectrophotometry (Agilent 8453, Waldbronn, Germany) at λ 320 nm. The concentration of released drug was determined using standard calibration curves (r 2 =0.9944–0.9984). Cumulative corrections were made for the previously removed samples when determining the total amount released. The drug release kinetics from the designed devices were also investigated using different mathematical models: Higuchi (1), Korsmeyer–Peppas (2), Peppas–Sahlin (3) and Zero-order (4) equations [51, 52]. Mt M∞ =kH•t0.5(1) Mt M∞ =kK•tn(2) Mt M∞ =kd•tm+kr•t2m(3) Mt M∞ =k0•t(4) where M t /M ∞ is the fractional drug release at time t (M ∞ is considered equivalent to the drug loading); k H is the Higuchi kinetic constant; k K is the Korsmeyer kinetic constant; n is a release exponent that depends on the release mechanism and the geometry of the system; k d and k r are the diffusion and relaxation rate constants, respectively; m is the purely Fickian diffusion exponent for a device of any geometrical shape which exhibits controlled release and k 0 is the zero-order release rate constant. The optimum values for the parameters in each equation were determined by linear or non-linear least-squares fitting methods using SPSS® Statistics 26 software (IBM Corp., NY, USA). The adjusted coefficient of determination (r 2 adj ) and F-ratio probability were used to test the applicability of the release models. 3. Results and discussion 3.1. Synthesis of PPDL 70 CHDMS 30 Fig. 1 displays the chemical structure of the PPDL 70 CHDMS 30 copolyester synthesized by a combination of enzymatic ring opening polymerization and two-step melt polycondensation processes of PDL, CHDM and DMS with the appropriate monomers molar ratio. The copolyester composition and the microstructure were determined by 1 H and 13 C NMR, respectively (Fig. 2). Signals from CH 2 CO in the 1 H NMR spectrum (Fig. 2 bottom) of PDL (2.3 ppm) and succinate (2.6 ppm) units were used for the determination of copolymer composition that was very close to the one used in the feed (70.9/29.1 PDL/ CHDMS mol/mol). On the other hand, signals due to cyclohexylenedimethylene units in the 13 C NMR spectrum (Fig. 2 top) split into different peaks due to both stereochemistry and sequence distributions. By integration of these signals and using the equations reported by Tessier and Fradet [53] the degree of randomness was calculated. This parameter can be defined as the probability of finding hetero-type bonds (PDL-CHDM or PDL-S) divided by the product of the probabilities of finding the corresponding PDL, CHDM and S groups in the copolymer chains. Values close to 0 and exactly 2 are found for block and alternating copolymers, respectively. A value close to 1 (1.05) was obtained for PPDL 70 CHDMS 30 copolyester, indicating a random microstructure, i. e. PDL, CHDM and S units are statistically distributed along the copolymer chains. The weight average molecular weight (M w ) of PPDL 70 CHDMS 30 copolyester determined by GPC was 107,100 g mol −1 . The value of the dispersity (Ð) was very close to 2 (2.2), in agreement with the theoretical value for copolymers obtained by enzymatic ring opening polymerization and polycondensation, where transesterification reactions take place [54]. 3.2. Thermal analysis studies The thermal behaviour of the raw materials and physical mixtures was analyzed using both DSC and TGA. The results are shown in Fig. 3. The original and cryomilled PPDL 70 CHDMS 30 copolymers were subjected to DSC analysis (Fig. 3A) to assess the influence of the milling process in the thermal properties of the copolyester. A broad melting peak was clearly observed during the heating ramp, which indicates the semi-crystalline structure of the copolymer. A melting temperature (T m ) of 64.9 ◦C and a fusion enthalpy (ΔH m ) of 68.0 J/g were estimated for the cryomilled copolyester. The analysis results for the cryomilled fraction were similar to those of the original copolymer (Fig. S1A), which means that the milling process did not modify its thermal properties. The DSC curve of metronidazole showed a sharp endothermic peak at 161.2 ◦C, indicative of the melting process (ΔH m =192.8 J/g) of the crystalline drug. These data were compatible with the data from the manufacturer and the values reported in the literature [31,55–57]. Two melting points could also be distinguished in the DSC curve corresponding to the physical mixtures of copolymer:drug (90:10 and 70:30 w/w ratios): the first one at ≈65 ◦C and the second one at ≈ 160 ◦C. Then, the characteristic thermal events of drug and copolymer were observed in the physical mixtures, although the melting enthalpies were attenuated due to their lower proportion in the mixtures. TGA analysis (Figure⋅3B) was also conducted to assess the degradation profile of the PPDL 70 CHDMS 30 copolyester (original and cryomilled fractions). The cryomilled copolymer exhibited a mass loss step starting at approximately 300 ◦C, with a weight loss of ≈10 % w/w at 391 ◦C (T d 10 % ). The mass loss recorded during the decomposition of the copolymer was continuous, with a maximum degradation temperature of 424 ◦C (T d max ) and a final residue of ≈0.5 % w/w. The TGA curves of the original and cryomilled fractions of the copolymer overlapped Fig. 1. Chemical structure of PPDL 70 CHDMS 30 copolyester. C. Ferrero et al. Journal of Drug Delivery Science and Technology 95 (2024) 105652 5 (Fig. S1B), confirming the negligible influence of the milling process on the thermal stability of the copolyester. The drug metronidazole exhibited lower thermal stability than the copolymer, starting the mass loss at temperatures higher than 160 ◦C. The T d 10 % and T d max were estimated at 224 ◦C and 274 ◦C, respectively, and the final residue after degradation was ≈4.1 % w/w. Similar thermal behaviour was reported by other authors [31,44]. The TGA curve corresponding to the physical mixture of copolymer: drug (90:10 w/w ratio) showed two degradation steps. The first one corresponded to the drug degradation, with a T d 10 % value of 281 ◦C and a T d max value of 286 ◦C; the second one was indicative of the copolymer decomposition and started at temperatures higher than 368 ◦C (T d max 439 ◦C). Both degradation events were also visible in the TGA curve corresponding to the physical mixture of copolymer:drug (70:30 w/w ratio). However, due to the higher drug percentage, the thermal decomposition events started at lower temperatures. Hence, a T d 10 % value of 240 ◦C and a T d max of 251 ◦C were estimated for the drug degradation step, while copolymer decomposition started at temperatures higher than 358 ◦C (T d max 422 ◦C). The results obtained ensure that the formulations can be processed at 100-110 ◦C, well below their degradation temperatures. 3.3. Physical characterization of USAC devices As stated in section 2.5, USAC cylindrical devices were obtained from the physical mixtures containing 10 % and 30 % w/w MTZ applying ultrasound energies of 400 and 350 J, respectively, through the upper punch of the tableting machine. USAC devices from the bulk copolymer at 500 J were also prepared for comparative purposes in the PXRD studies (see section 3.5). The increase of temperature within the tablet along with the mechanical pressure applied improve the extent of interparticle bonding and promote the sintering of materials [6]. As depicted in Fig. 4A, USAC devices containing 10 % w/w MTZ showed a smooth surface, where isolated crystalline particles corresponding to the drug could be observed. USAC devices containing 30 % w/w MTZ showed similar appearance than their counterparts with 10 % w/w drug but a higher presence of crystalline particles on the surface was noticed Fig. 2. 1 H- (bottom) and 13 C NMR spectra (top) of PPDL 70 CHDMS 30 copolyester with peak assignments. Inset, 13 C NMR signals of CHDM carbons that split due to sequence distributions. Fig. 3. DSC (A) and TGA (B) patterns of raw materials and their corresponding physical mixtures. C. Ferrero et al. Journal of Drug Delivery Science and Technology 95 (2024) 105652 6 (Fig. 4B). Table 1 shows the mean, standard deviation (SD) and relative standard deviation (RSD) of the weight and physical dimensions of the USAC devices. As it can be observed, a high uniformity of weight and dimensions was obtained employing the USAC technology, confirming the reproducibility of this process. A slight decrease of the thickness of the devices was observed for the formulation containing 30 % w/w drug. 3.4. Physical characterization of 3D-printed SSE devices The 3D-printed SSE devices were built onto a glass printing platform, layer-by-layer, according to the digitally designed object. The scaffolds were made up of 10 layers with one sealing perimeter. Firstly, two solid bottom layers were built, i.e., in our design, the first two layers were not a network, but a solid layer made up of parallel strands leaving no gaps between them (Fig. 5B). From 3rd to 10th layer, the printing process followed a diagonal pattern with 4 mm pore size. Hence, a compartmental mesh was obtained (Fig. 5A) to increase the surface area of the systems. All the steps were automatically performed, building the structure without the need of drying steps. Moreover, due to the low melting point of the copolymer (~65 ◦C), processing temperatures around 100 ◦C were used, avoiding drug degradation. SSE devices were successfully printed only from the mixture containing 10 % w/w MTZ. An increase in the drug percentage lead to a high melt viscosity that difficulted the homogeneous flow through the nozzle. Due to the hydrophobicity of the PPDL 70 CHDMS 30 copolymer, it was a challenge to get an easily extrudable mixture with the drug metronidazole, in agreement with the results found by Utomo et al. [31] using the hydrophobic PCL polymer. Results in Table 1 showed that, although the aspect ratio of the devices was similar than those of the USAC systems, the weight of the SSE devices was considerably lower due to their open internal structure. Moreover, the reproducibility of the weight and physical dimensions of the devices worsen when compared with USAC systems obtained from the mixture with the same drug percentage. This was mainly attributed to the low resolution of the SSE technique. Although the use of large size nozzles (0.6 mm diameter in our case) is necessary for high viscosity materials, the resolution and printing accuracy of the 3D-printed device could be compromised. In addition, small variations in the flow of the semi-solid material through the nozzle during printing could result in uneven deposition of material, which could affect the consistency of the thickness of the device and therefore its weight. These technical challenges were also reported in the reviewed literature [20,22,30,40]. 3.5. Powder X-ray diffraction (PXRD) studies Fig. 6 shows the X-ray diffraction patterns of the raw materials, USAC and 3D-printed SSE devices. As it can be observed, the PPDL 70 CHDMS 30 copolyester exhibited a semi-crystalline structure with only two peaks: a more intense peak at 21.4◦and a weaker peak at 23.8◦(2θ). These peaks are due to 110 and 200 planes of the pseudo-orthorhombic unit cell of the crystallized PDL segments of the copolyester, in agreement with the results reported by Urpí et al. [4]. Metronidazole showed a highly crystalline structure since sharp intensity diffraction peaks at characteristic 2θ angles could be visualized. The most intense peak appeared at 12.3◦followed by a less intense peak at 24.7◦(2θ). Other peaks of less intensity could be observed at 13.9, 16.2, 17.2, 18.0, 19.5, 20.4, 21.5, 23.3, 24.0, 25.4, 27.4, 28.0, 29.3, 29.8, 33.3 and 33.9◦(2θ). This characteristic pattern has been previously reported by other authors [44,55,57,58]. The diffraction patterns of the raw materials were in good correlation with the results of the thermal analysis studies. The two characteristic peaks of the copolymer could also be observed in the diffractogram corresponding to the USAC device containing 100 % w/w copolymer, although a slight decrease in the intensity of both peaks could be appreciated. This fact might be attributed to the partial sintering of the copolymer particles due to the mechanical pressure applied during compression and the temperature rise caused by ultrasounds. Nevertheless, the ultrasound energy of 500 J was not enough to reach a complete transition of the material. Fig. 4. Digital pictures of the USAC devices (top view) with 10 % w/w MTZ (A) and 30 % w/w MTZ (B) (the image contrast was adjusted to improve drug particles visualization). Table 1 Physical characteristics of the USAC and 3D-printed SSE devices. Formulation Weight (mg) Diameter (mm) Thickness (mm) USAC (10 % w/w MTZ) 248.3 ±0.7 10.948 ±0.091 2.408 ±0.009 RSD 0.27 % RSD 0.83 % RSD 0.35 % USAC (30 % w/w MTZ) 245.2 ±2.4 10.874 ±0.186 2.244 ±0.037 RSD 0.98 % RSD 1.71 % RSD 1.67 % SSE (10 % w/w MTZ) 104.6 ±15.3 10.009 ±0.219 2.073 ±0.155 RSD 14.67 % RSD 2.18 % RSD 7.47 % Data represent mean ±SD (n =5); RSD is the relative standard deviation. Fig. 5. Virtual design and digital pictures of the 3D-printed SSE devices: (A) top view (meshed internal structure), (B) bottom view (solid bottom layers). C. Ferrero et al. Journal of Drug Delivery Science and Technology 95 (2024) 105652 7 In the case of the USAC devices containing 10 and 30 % w/w MTZ, the diffractograms showed the characteristic X-ray pattern of the drug, indicating that metronidazole remained mostly crystalline during USAC process. However, the intensity of the peaks was attenuated accordingly to the proportion of MTZ in the mixture, i.e., for the larger peak at 12.3◦, the intensity decreases from 420402 (pure drug) to 120826 (formulation with 30 % w/w drug) or 37835 (formulation with 10 % w/w drug). It could also be noticed that the diffraction peaks of the copolymer overlapped with those of metronidazole at the same 2θ angle. Conversely, in the 3D-printed device containing 10 % w/w MTZ, some amorphization of the drug could be observed since a change in the X-ray pattern (light blue line in Fig. 6) compared with the same formulation prepared by the USAC technique (dark blue line in Fig. 6) could be appreciated, with an important decrease in the intensity of the peaks at 12.3◦and 24.7◦. Mirank´ o et al. [56] also reported the presence of metronidazole in both crystalline and amorphous states in solid dispersions obtained by nano spray drying. The peak broadening in the XRD profiles was attributed to the formation of small drug crystallites. 3.6. Scanning electron microscopy (SEM) studies Fig. 7 shows SEM images of the raw materials. Two types of irregular particles could be distinguished in the microphotographs corresponding to PPDL 70 CHDMS 30 copolymer (Figs. 7A1 and A2). Some particles had a rough surface and rounded edges while others presented flat and smoother surfaces, probably due to the cryomilling process. A wide particle size distribution could be observed in the SEM images corresponding to metronidazole, with large and irregular particles and smaller ones adhered to their surface (Fig. 7B1, B2). This morphology was in accordance with the literature [45]. SEM images of the USAC devices are collected in Fig. 8. USAC devices containing 10 wt% MTZ showed a smooth surface (Figs. 8A1 and A2) due to the thermoplastic deformation and partial sintering of the copolymer. Like digital pictures in Fig. 4, more discontinuities were observed on the surface of the systems containing 30 wt% MTZ (Fig. 8B1, B2), due to their higher load of crystalline drug particles. These observations were in good agreement with the higher crystallinity detected in the PXRD scans. The cross-sectional microphotographs of the devices containing 10 wt% MTZ (Fig. 8C) also confirmed the existence of a quasi-continuum Fig. 6. PXRD patterns of the raw materials, USAC and 3D-printed SSE devices. Fig. 7. SEM microphotographs of the raw materials: PPDL 70 CHDMS 30 copolymer (A1, 40x) (A2, 250x); Metronidazole (B1, 40x) (B2, 250x). C. Ferrero et al. Journal of Drug Delivery Science and Technology 95 (2024) 105652 8 medium, due to the application of ultrasounds and subsequent thermoplastic deformation of the copolymer. The insoluble copolyester almost surrounded the active agent particles. In contrast, the higher presence of crystalline drug particles in the formulation with 30 wt% MTZ leaded to a more heterogenous internal structure (Fig. 8D). Fig. 9 shows the SEM images of the 3D-printed SSE devices as well as the strands extruded through the nozzle for comparative purposes. It could be observed that the extruded strands obtained from the mixture containing 10 %wt MTZ (Figs. 9A1 and A2) had a smooth surface due to the copolymer melting. On further magnification, small drug crystalline particles spread over the surface of the melted polymer were observed. These small size crystals could also be appreciated in the SEM images corresponding to the solid bottom layer (Fig. 9B1, B2) and the perimeter (Fig. 9C1, C2) of the SSE devices. Although the perimeter appeared as a continuous melted phase, parallel lines could also be visualized due to the layer-by-layer 3D-printing process. The presence of small regular polyhedral particles between 1 and 5 μ m on the surface of the strands and SSE devices suggests that MTZ particles remained mostly crystalline although a marked habit modification of MTZ crystals (Fig. 9B2) could be appreciated when compared with the raw material (Fig. 7B2). It is hypothesized that, during the extrusion process, the crystalline drug suffers a partial amorphization and a subsequent recrystallization on the surface of the strand. This behaviour was also supported by the different crystalline pattern exhibited by the 3D-printed devices in the PXRD studies. Other authors also found small MTZ crystals dispersed on the surface of devices obtained by semi-solid extrusion [31], solvent casting [55] or nano spray drying [56] techniques. 3.7. In vitro drug release studies The cumulative drug release profiles from USAC devices (10 and 30 wt% MTZ) and 3D-printed SSE devices (10 wt% MTZ) in phosphate buffer (pH 6.8) are illustrated in Fig. 10. All formulations exhibited a prolonged and/or sustained drug release that could be mainly attributed to the hydrophobicity and slow degradation rate of the PPDL 70 CHDMS 30 copolyester. After a burst release of 15.4 ±2.1 %, USAC devices containing 10 wt % drug showed a sustained drug release behaviour, with 32.1 ±5.7 % drug released at physiological pH after 21 days. As expected, the burst release (23.1 ±0.6 %) increased in the case of the USAC devices containing 30 wt% MTZ and the drug release was faster due to the higher drug loading in the formulation [59]. Then, a cumulative amount of 84.6 ±8.0 % was released after 21 days. A sustained drug release behaviour was also observed for the SSE devices, with a lower burst release (5.9 ±0.8 %) than the value found for the USAC devices containing the same drug loading (10 wt%). These results could be explained by the influence of the manufacturing technique. The SSE 3D printing technique uses heat and mechanical pressure to extrude the physical mixture of polymer and drug. After the melting and 3D printing processes, the structure solidified as a dense and continuum medium that difficulted water uptake. However, the USAC technique combines a conventional compression process and US irradiation [6]. Although ultrasounds caused a temperature rise within the compact that increased the interparticle bonding and improved the mechanical strength of the device, the ultrasound energy was not enough to cause the complete sintering of the copolymer. Thus, some porosity could remain, favouring the access of water and consequent Fig. 8. SEM microphotographs of the USAC devices: 10 wt% MTZ -axial view- (A1, 50x) (A2, 400x); 30 wt% MTZ -axial view- (B1, 50x) (B2, 400x); 10 wt% MTZ -cross-sectional view- (C, 46x); 30 wt% MTZ -cross-sectional view- (D, 50x). C. Ferrero et al. Journal of Drug Delivery Science and Technology 95 (2024) 105652 9 dissolution of the hydrophilic drug metronidazole. After the 4th day of release, the effect of the device’s surface area becomes predominant over the influence of the manufacturing technique. Therefore, due to the larger surface area of the 3D-printed system and partial amorphization of the crystalline drug by this technique, the drug release increased up to 49.6 ±6.5 % after 21 days. Utomo et al. [31] also found that the release rate of MTZ from intravaginal devices of the biodegradable polymer poly(caprolactone) (PCL) was dependent on the shape of the device. Hence, faster release was observed from mesh systems (larger surface area) when compared with flat discs. To further elucidate the drug release mechanisms, dissolution data (M t /M ∞ <0.6) were fitted to Higuchi, Korsmeyer-Peppas, Peppas-Sahlin Fig. 9. SEM microphotographs of the 3D-printed SSE devices containing 10 wt% MTZ and the corresponding strands: Extruded strands (A1, 100x) (A2, 400x); SSE device surface (B1, 200x) (B2, 3000x); SSE device perimeter (C1, 42x) (C2, 400x). Fig. 10. Cumulative in vitro drug release profiles from USAC devices (10 and 30 wt% MTZ) and 3D-printed SSE devices (10 wt% MTZ) in phosphate buffer (pH 6.8) at 37 ◦C (mean ±SD, n =3). Table 2 Mathematical modeling and drug release kinetics from USAC and SSE devices. Kinetic model Parameters USAC 90:10 USAC 70:30 SSE 90:10 Higuchi (1) k H (h −0.5 ) 0.0072 0.0272 0.0196 r 2 adj 0.9878 (F = 1460) 0.9897 (F = 1255) 0.9863 (F = 1510) Korsmeyer- Peppas (2) n 0.11 0.19 0.36 k K (h -n ) 0.1351 0.1920 0.0417 r 2 adj 0.9009 (F = 165) 0.9189 (F = 148) 0.9576 (F = 476) Peppas-Sahlin (3) k d (h −0.45 ) 0.0055 0.0166 0.0117 k r (h −0.90 ) 0.0003 0.0016 0.0009 r 2 adj 0.9935 (F = 1371) 0.9953 (F = 1374) 0.9987 (F = 7811) Zero-order (4) k 0 (h −1 ) 0.0003 0.0017 0.0009 r 2 adj 0.9685 (F = 554) 0.9818 (F = 702) 0.9771 (F = 899) Zero-order (4) (1st-21th day period) k 0 (h −1 ) 0.0003 0.0015 0.0008 r 2 adj 0.9967 (F = 3291) 0.9846 (F = 383) 0.9910 (F = 1548) k H , Higuchi kinetic constant; n, release exponent; k K , Korsmeyer kinetic constant; k d , diffusion kinetic constant; k r , relaxation kinetic constant; k 0 , zero-order release rate constant; r 2 adj , adjusted coefficient of determination; F, F distribution for residual variance analysis (p =0.000). C. Ferrero et al.