Contents lists available at ScienceDirect

European Journal of Pharmaceutics and Biopharmaceutics

journal homepage: www.elsevier.com/locate/ejpb

Research paper

SPHRINT – Printing Drug Delivery Microspheres from Polymeric Melts

Tal Shpigel, Almog Uziel, Dan Y. Lewitus⁎

Plastics and Polymer Engineering Department, Shenkar College, Ramat-Gan 6262528, Israel

A R T I C L E I N F O

Keywords:SphrintPrintingPolymerMeltSuperoleophobic surfaceMicrospheresDrug delivery

A B S T R A C T

This paper describes a simple, straightforward, and rapid method for producing microspheres from moltenpolymers by merely printing them in an inkjet-like manner onto a superoleophobic surface (microsphereprinting, hence SPHRINT). Similar to 3D printing, a polymer melt is deposited onto a surface; however, incontrast to 2D or 3D printing, the surface is not wetted (i.e. exhibiting high contact angles with liquids, above150°, due to its low surface energy), resulting in the formation of discrete spherical microspheres. In this study,microspheres were printed using polycaprolactone and poly(lactic-co-glycolic acid) loaded with a model activepharmaceutical ingredient—ibuprofen (IBU). The formation of microspheres was captured by high-speed ima-ging and was found to involve several physical phenomena characterized by non-dimensional numbers, in-cluding the thinning and breakup of highly viscous, weakly elastic filaments, which are first to be described inpure polymer melts. The resulting IBU-loaded microspheres had higher sphericity, reproducible sizes and shapes,and superior drug encapsulation efficiencies with a distinctly high process yield (> 95%) as compared to theconservative solvent-based methods used presently. Furthermore, the microspheres showed sustained releaseprofiles.

1. Introduction

In the realm of pharmaceutical products, microparticle-based depotsare applied toward the sustained drug release over long periods, thusfacilitate the controlled delivery of therapeutics. As a result, the fre-quency of injecting drugs composed of small molecules, peptides, orproteins can be significantly reduced in comparison to injections [1,2].Several polymeric excipients exist primarily for this purpose, with mostbeing biodegradable. These include poly(lactic acid) (PLA), poly-glycolic acid (PGA), their copolymer poly(lactic-co-glycolic) acid(PLGA) and polycaprolactone [1,3]. Numerous microparticle-baseddepots products are FDA approved, amongst them are: Zmax® (Azi-thromycin), Decapeptyl®/Trelstar® (Triptorelin), Vivitrol® (Nal-trexone), Arestin® (Minocycline), Risperdal® Consta® (Risperidone),Sandostatin® LAR Depot (Octreotide), Nutropin Depot® (Somatropin),Lupron Depot® (Leuprolide), DepoCyt® (Cytarabine), DepoDur® (Mor-phine), Bydureon® (Exenatide) Somatuline LA (Lanreotide) [1] andrecently approved ZILRETTA™ (triamcinolone acetonide) [4].

Solvent-based methods such as spray-drying and solvent evapora-tion are the most widespread techniques for producing polymeric mi-crospheres [5], but they have fundamental drawbacks owing to theremoval of the solvents and surfactants, resulting in significant API loss,low encapsulation efficiencies, and low process yields [6]. Furthermore,melt-based techniques are less prevalent [7,8], such as spray congealing

and melt solidification, but are solvent-free in essence [7,9,10]. How-ever, they suffer from set-up difficulties and non-uniformity in size andshape of the spheres [7].

In the past few decades, pioneering inkjet printing technologieshave successfully exploit polymers toward printing almost any productthat comes to mind [11,12], including drop-on-demand (DoD) inkjetprinting for pharmaceutical and biomedical device applications[12–14]. One example reviewed by the Schubert group [12] are theinkjet printed Paclitaxel-loaded PLGA microspheres, for controlleddrug-delivery applications. Although are not printed on a substrate,these microspheres were fabricated by jetting polymeric solution into asurfactant bath, followed by solvent evaporation while stirring to formsolid microspheres. The majority of existing printing techniques of anysolid particulates utilize diluted polymer solutions [11–14], andtherefore the properties of viscoelastic fluid jets such as viscosity andsurface tension are of high preeminence for high quality printing andhave been extensively studied. Exceptional inkjet printing techniquesthat utilize pure polymers in 2D and 3D printing are usually wax-basedprinting techniques that are used in the graphical industry, such as theXerox solid ink/wax technology [15] and the wax pattern 3D printer bySolidScape [16]. It is worth mentioning that none of the aforemen-tioned technologies have reported to utilize inkjet printing of purepolymer melts to produce drug-delivery microspheres. Moreover, thefilament thinning mechanism formed through the inkjet printing

https://doi.org/10.1016/j.ejpb.2018.03.006Received 28 December 2017; Received in revised form 11 March 2018; Accepted 14 March 2018

⁎ Corresponding author.E-mail address: Lewitus@shenkar.ac.il (D.Y. Lewitus).

European Journal of Pharmaceutics and Biopharmaceutics 127 (2018) 398–406

Available online 22 March 20180939-6411/ © 2018 Elsevier B.V. All rights reserved.

T

process have been extensively investigated in the literature for solidparticulates and suspensions. Yet, there is no evidence in the literaturefor characterizing the thinning mechanism of a pure filament melt. Thecapillary thinning of the filament is driven by the inherent interfacialtension of the fluid, resulting in an increasing capillary pressure as thefilament diameter decreases (σ/R) [17]. Inertia, viscosity, and elasticityforces could balance the capillary pressure throughout the filamentthinning process, whereas the governing thinning mechanism couldchange with time as a function of the force providing the highest re-sistance to thinning [18]; and can be determined by computing thethinning velocity and three non-dimensional numbers.

The latest research in the field of microsphere engineering is a workby the Mano group [19–21] and the Butt group (Deng et al. [22]). Bothgroups manipulate a non-wetting surface, i.e. a surface owing lowsurface energy due to the combination of chemical modifications,hierarchical micrometric structure and a coarse roughness [23]; toproduce polymeric microspheres. Mano et al., describe the formation ofgel capsules by spraying an aqueous gel/polymer/monomer solutiononto superhydrophobic surfaces, followed by solidification via UV ir-radiation (photopolymerization), or through physical crosslinking. Thedescribed process is limited to hydrogels (with water as a solvent),require secondary process (crosslinking), and have a limited controlover particles size. At the Butt group (Deng et al. [22]), they have de-monstrated a method to produce solvent-free microspheres using a non-wetting surface. Their process describes either depositing a droplet ofliquid alkene monomers with a photoinitiator on top of the surface(forming a round droplet), followed by irradiation to radically- poly-merize the material into a sphere. In addition, they have demonstratedthe formation of microspheres from melt through heating of dispersedpolymer powders on a superoleophobic surface. The process paradigmrelies on heating the powders on top of the surface for several minutesto complete melting, forming a molten sphere, which is then cooled.The limitations in the described process are either the use of alkenepolymers only, along with a secondary radiation curing process, whilethe heating of powders is complex to apply and time consuming.

The limitations of these innovative techniques, along with thechallenges of solvent-based methods and inkjet printing techniquespresently used, encouraged us to develop a solvent-free, cost-effective,and versatile technique to rapidly print microspheres of polymer meltsusing non-wetting surfaces with a specialized dispensing valve. In thiswork, we demonstrate for the first time SPHRINT’s capacity to producedrug delivery microspheres from a clinically relevant polymer (PCL)loaded with a model API – Ibuprofen (IBU). The nonsteroidal anti-in-flamatory drug (NSAID) IBU is commonly used as a model API in DDSmainly due to its availability and applicativity (i.e., Salmoria et al. [24]produced IBU-loaded PCL rods through the melt extrusion techniqueand Fernández-Carballido et al. [25] prepared IBU-loaded PLGA mi-crospheres through the solvent-evaporation technique, both for drug-delivery applications). Moreover, its relatively low melting point and itsproximity with the PCL's melting point is of great benefit to the printingsystem, resulting in the reduction of energy consumption. The uniqueprinting process accompanied with an in-depth physical analysis, itsresulting microspheres, and the technology’s potential are describedhereafter. A complementary analysis was conducted for PLGA micro-spheres loaded with IBU, to explore the feasibility to print amorphouspolymeric microspheres, and compare their in vitro drug release profilesto the semi-crystalline, PCL-loaded microspheres.

2. Methods

2.1. Materials

PCL (avg. Mw: approximately 14,000), PLGA (Resomer® RG 502,lactide:glycolide 50:50, ester terminated, Mw: approximately7,000–17,000), IBU (≥98% (GC), Mw 206.28 g/mol), Dulbeco's phos-phate buffered saline (modified without calcium chloride and

magnesium chloride), chloroacetic acid (HPLC grade), and acetonitrile(HPLC grade) were purchased from Sigma-Aldrich, Israel. MiniGeBaFlex-tube dialysis kit (8 kDa MWCO) was purchased from GeneBio-Application LTD, Israel. Dichloromethane (DCM, stab. Absolute,AR-b) was purchased from Bio-Lab LTD, Israel. Ultra-Ever Dry®(UltraTeach Inc.) coating was purchased from Chemitron LTD, Israel.

2.2. Preparation of superoleophobic surfaces

Aluminum surfaces were spray coated with UltraTech Bottom-coat™and Top-coat™ in accordance with the manufacturer's guidelines. Onceprepared, the coated surfaces were placed in an oven at 150 °C for 1 h.

2.3. Preparation of IBU-loaded microspheres

IBU-PCL and IBU-PLGA blends with 30 wt% IBU were prepared byhomogeneously mixing both components molten at 80 °C (PCL) or150 °C (PLGA). The mixtures were inserted into a heated reservoir of apneumatic jetting valve (P-Dot CT, Liquidyn® (Nordson EFD)Oberhaching, Germany). The following settings were used: fluid tem-perature: 80 °C (PCL) or 150 °C (PLGA), fluid pressure: 0.2 bar (PCL) or0.4 bar (PLGA), valve pressure: 2.3–2.8 bar (PCL) or 1.9–3.1 bar(PLGA), spring (tappet) tightness: 0.6–0.9 turns (PCL) or 0.8 turns(PLGA), distance between the nozzle and the surface: 4.3 mm (PCL) or1.3–1.5mm (PLGA). The temperature and pressure were controlledusing Liquidyn's V100 controller. In order to generate microspheres,droplets of the molten blends were jetted through a 150-µm NeedleNozzle (Liquidyn®) onto a superoleophobic surface, and subsequentlycooled at RT (or at −20 °C for reproducibility analysis testing).

2.4. Preparation of neat PCL and PLGA microspheres

PCL flakes and PLGA powder were inserted into a heated reservoirof a pneumatic jetting valve (P-Dot CT, Liquidyn® (Nordson EFD)Oberhaching, Germany). The following settings were used: fluid tem-perature: 150 °C (PCL) or 170 °C (PLGA), fluid pressure: 0.2 bar (PCL) or0.4 bar (PLGA), valve pressure: 2.5 bar (PCL) or 3.1 bar (PLGA), spring(tappet) tightness: 0.9 turns (PCL) or 0.8 turns (PLGA), distance be-tween the nozzle and the surface: 4.3mm (PCL) or 1.5mm (PLGA). Thetemperature and pressure were controlled using Liquidyn's V100 con-troller. In order to generate microspheres, droplets of the molten blendswere jetted through a 150-µm Needle Nozzle (Liquidyn®) onto a su-peroleophobic surface, and subsequently cooled at RT.

2.5. Documentation of microsphere formation process

A digital high-speed camera (120 Phantom v12) equipped with amicroscopic lens (Nikon 10×CFI plan achromat) was used to observeand document the ejection, flight, impact, and stabilization of thedroplet in detail, shooting at 27,000–43,000 fps. The velocity of the jetand radii of the filament at different time lapses were calculated usingPhantom PCC 2.8 and ImageJ processing analysis software [26].

2.6. Determination of encapsulation efficiency

30% IBU-loaded PCL or PLGA microspheres (2 mg) were dissolvedin dichloromethane (DCM, 2ml), precipitated in phosphate buffer so-lution (PBS, pH 7.4), and stirred for 15min with a magnetic stirrer. Themixture was heated at 40 °C to remove the DCM [27], and thereaftercooled and filtered. The aliquots of IBU in the PBS solution were di-rectly analyzed using HPLC-UV in a liquid chromatograph Agilent 1100Series equipped with a Kinetex C18-100 Å column (150mm×4.6mmi.d.) (5 µm) at 25 °C under isocratic conditions. The injection volumewas 20 µL. The mobile phase consisted of acetonitrile:water (60:40)with chloroacetic acid (20mM, pH 3.0). The detection was performedusing a UV detector at 264 nm [28]. The entrapment efficiency was

T. Shpigel et al. European Journal of Pharmaceutics and Biopharmaceutics 127 (2018) 398–406

399

calculated by dividing the actual drug content by the theoretical drugcontent of microspheres [29]. Three replicates from each batch weremeasured against a calibration curve, and the mean encapsulation ef-ficiency (EE%) and standard deviation around the mean were de-termined.

2.7. In vitro drug dissolution studies

30% IBU-loaded PCL or PLGA microspheres were placed in dialysistubes (8 kDa MWCO), and incubated in PBS (pH=7.4) solution at37 °C under mild agitation for 17 days. All the samples tested had aninitially uniform IBU concentration of 0.3mgml−1 (sink condition)[30]. At predetermined time intervals, the aqueous solution was with-drawn and replenished with fresh PBS. Three replicates were measuredfor each type of microsphere. The IBU concentration in the PBS wasmeasured against a calibration curve, and cumulative release profileswere obtained using Eq. (1) [31]:

∑= ×Cumulative drug releaseC

C(%) 100%

t

t

totali

i

(1)

where the cumulative drug release is computed in percentages (%), Ctiis the concentration of IBU measured at time ti, and Ctotal is the totalconcentration of IBU in the tested microspheres obtained from the en-capsulation efficiency (EE%) analysis.

2.8. Sphericity evaluations

The spherical characteristic of the microspheres was determined viacontact angle measurements after solidification using a CA analyzermachine (OCA 15, DataPhysics). The DataPhysics software (SCA20—contact angle) automatically calculated the obtained angles, i.e.,contact angles, between the detected surface and the perimeter of asingle sphere that was manually outlined (as described in ASTMD7334). Ten replicates from each batch were measured, and the meancontact angle and standard deviation around the mean were de-termined.

2.9. Microsphere size, surface morphology and texture

Microsphere size, surface morphology and texture were evaluatedusing SEM. The microspheres were gold–tin coated using a sputtercoater unit (SC7620, Quorum) and imaged using an Aspex ExplorerSEM operated at an acceleration voltage of 13 kV. The size of 20 ran-domly selected microspheres was measured using the SEM softwareroller and determined by calculating the particle mean diameter andstandard deviation around the mean.

2.10. Rheological evaluations

In this work, the melt rheological characterization was conductedfor PCL only. The melt rheological behavior of PCL was characterizedusing an oscillatory shear rheometer (Discovery HR-1, TA instrument)in parallel plate geometry (diameter: 25mm). The measurements werecarried out at a constant shear rate of 0.35 Hz in a temperature ramp inthe range of 50–150 °C, and at a constant temperature of 80 °C in anoscillatory frequency sweep from 0.35 to 100 Hz at a constant strainvalue of 0.2%.

3. Results and discussion

In the printing process (Fig. 1 and SI Video 1), an inkjet-like deviceextends a portion of a molten polymer–API blend through a nozzle ontoa non-wetting surface. The molten material spontaneously forms aspherical droplet and thereafter solidifies in a near-perfect sphericalshape.

Supplementary Video 1. After determining the optimal parameters toproduce small, spherical, and stable microspheres, the sphere formationprocess of molten PCL was documented using high-speed imaging and amagnifying probe (Fig. 2A and SI Video 2). Three major events weredetermined to be responsible for the overall sphere formation process:(1) ejection of polymer-melt jet through the orifice, (2) filament for-mation, thinning, and pinch-off, and (3) melt interaction with the su-peroleophobic substrate and sphere formation. The second and thirdevents occurred simultaneously; i.e., after the jet impacted the surface,it began to adopt its spherical shape as the filament thinned, untilpinching occurred and the tail retracted viscoelastically to form aspherical droplet (see SI Video 3). The implication of using non-wettingsurfaces is shown in Fig. 3, which will be further discussed in detail inthe following sections.

Fig. 1. Process paradigm: a sequence of images (from left to right) captured with a digital camera, showing the evolution of a molten 30% IBU-PCL (IBU: ibuprofen)droplet ejected through an orifice, its impact on a non-wetting surface, and subsequent solidification in a near-perfect spherical shape. Scale bar: 150 µm.

T. Shpigel et al. European Journal of Pharmaceutics and Biopharmaceutics 127 (2018) 398–406

400

Supplementary Video 2.

Supplementary Video 3.

Supplementary Video 4.

Supplementary Video 5. For a better understanding of these threeevents, the rheological, viscometric, and elastic properties of the PCLmelt were measured. Our observations with regard to these phenomenaare detailed below.

3.1. Sphere formation process analysis

3.1.1. Ejection of melt jet through the orificeFirst, the jetting rate and shear rate experienced by the material in

the die owing to pneumatic forces were calculated. The jetting rate (Qp,i.e., pressure flow) was calculated using Eq. (2) [32], and the shear ratederived from Eq. (2) was calculated using Eq. (3). Both Eqs. (2) and (3)are applicable for non-Newtonian fluids, and are determined by thepower law constants n and K of the melt, the pressure difference in thejetting valve ΔP, and the nozzle configuration (diameter D and lengthL).

⎜ ⎟= ⎛⎝

⎞⎠

⎛⎝ +

⎞⎠

⎛⎝

⎞⎠

Q πD nn

D PLK32

43 1

Δ4p

3 n1

(2)

T. Shpigel et al. European Journal of Pharmaceutics and Biopharmaceutics 127 (2018) 398–406

401

= ⎛⎝

+ ⎞⎠

= ⎛⎝

⎞⎠

γ nn

QπR

D PLK

3 1 Δ4w

p3

n1

(3)

The computed shear rate was determined as 0.35 s−1 according to:K=8351 Pa sn and n=0.759 (power law constants of PCL at 120 °Cobtained from the literature [33]); D= 0.15mm and L=0.2mm;ΔP=0.2 bar (denoted as P1, i.e. material pressure, approximately20,000 Pa). At a calculated shear rate of 0.35 s−1 the neat poly-caprolactone (PCL) exhibited a viscosity of 1.4 Pa s (see SI), resulting ina Reynolds number Re= 0.39 and a Weber number We= 103. Ac-cording to the map illustrated by Derby et al. [34], exhibiting a range ofWe and Re values for inkjet-printable matters, the melt was too viscousto be ejected through the orifice and is considered a non-printable fluid.However, according to Duineveld et al. [35], our process generated

sufficient velocity (6.7 m s−1) to exceed the minimal velocity for dropejection (0.9 m s−1) (Eq. (4), where γ is the surface tension, ρ is the meltdensity, and R is the orifice diameter), thus generating sufficient inertiato surpass the capillary forces in the orifice and enabling the occurrenceof the following event.

≈Uγ

ρR2

min(4)

3.1.2. Melt filament formation, thinning, and pinch-offAs shown in Fig. 2A, approximately 1.8ms after melt ejection, a

slender, cylindrical filament was formed between the orifice and a fixedsurface, fitting the necking shape of highly viscous/elastic fluids [17].

Fig. 2. (A) Frames obtained from high-speed imaging capturing the evolution of a molten PCL jet exiting the nozzle, impact on a superoleophobic surface, formationof a filament, filament breakup, and sphere formation (see SI Video 2). The impact occurs at approximately 1.84ms. With time, the part of the filament that interactswith the surface adopts a round shape, and a filament thinning phenomenon is simultaneously observed. After approximately 40ms, the filament breaks up, and thetail retracts viscoelastically to form a spherical droplet. Scale bar: 500 µm (B) Evolution of the filament diameter for a viscoelastic fluid with a high viscosity melt of neatPCL calculated from (A). The numbers I–IV denote the controlling regimes on filament thinning over time [18]: (I) elastic forces stabilize the filament immediatelyafter exiting the orifice, (II) the radius decreases linearly under viscosity-dominating regime, (III) the decreasing viscosity increases the thinning velocity, and (IV)exponential thinning of the filament under elastic forces, leading to filament break-up. The slope of region (II) and its linearity are denoted.

T. Shpigel et al. European Journal of Pharmaceutics and Biopharmaceutics 127 (2018) 398–406

402

By calculating the following dimensionless group of numbers – the

Ohnesorge number =( )Oh ηργR

, intrinsic Deborah number

⎛⎝

= ⎞⎠

De λ γρR0

2

3 , and elasto-capillary number = =( )Ec DeOh

λγηR

0 , we could

determine which of the forces—viscosity, inertia, or elasticity—weredominant over capillary thinning (symbols are specified in the SI) [18].According to the resulting Oh=26 > > 0.2 and Deo=116 > > 1,it is expected that the viscosity and elasticity both dominate the inertialforces, indicating that the PCL melt is a viscoelastic fluid with high visc-osity [18]. Moreover, with an Ec value of 4.5, it is often impossible toanalytically characterize the thinning evolution of an elastic fluid fila-ment (when 0.5 < Ec < 4.7), according to Clasen et al. [18]; there-fore, it cannot be unequivocally determined whether the viscous forcesdominate elasticity or vice versa. However, according to McKinley [36],with Ec > 1, it is expected that elasticity dominates viscosity. Never-theless, by plotting the thinning of the filament until break-up versustime, as illustrated in Fig. 2B, we observed the overall shape of the plotto be similar to the one reported by Clasen et al. [18], for weakly elasticfluids. The initial thinning process was dominated by an elastic regime,stabilizing the jet immediately after it exited the orifice. It was followedby a second process of constant thinning velocity, independent of thefilament radius [18], typical for a viscosity-dominated regime (Eq. (5)).Eq. (5) represents the Papageorgiou constant thinning behavior of aNewtonian fluid filament under viscous forces, where the radius thin-ning is a function of the surface tension and viscosity of the fluid withtime.

= −R Rγη

t0.07090(5)

The slope of the thinning curve obtained herein for PCL, −0.018γ/μ, was different from the slope calculated by Papageorgiou, −0.0709γ/μ [37]. The latter demonstrated a linear decrease in the filament dia-meter for Newtonian fluids [17] but not for polymer melts. Moreover,extensional viscosity effects became increasingly prevalent in this phaseowing to a coil-stretch transition [17], where polymer chains werestrongly stretched and oriented, thus influencing both the inter- andintra-chain interactions, leading to inconstant viscosity. As the inter-and intra-chain interactions were stronger in polymer melts rather than

polymer solutions, the extensional flow during this phase could have asignificant influence on the molten filament, and could be the reasonfor the difference obtained in the time-scale and slope. As our caseentails filament formation on a fixed surface [36], a “gravitationalstretching” [18] of the viscous filament jet has no effect on its neckinginstability, and therefore is disregarded. The final inflection point (atapproximately 38ms) was the third phase, known as the “extensionalflow field,” [18] where a progressive stretching and orientation oc-curred [17]. It was characterized by a decreasing viscosity (similar to

power law fluids), which promoted an increase in thinning velocity,leading to the final and fourth phase [18]. The fourth phase wasdominated by the elastic forces of the melt, leading to an exponentialthinning of the filament, until the filament broke up after 40ms, but nota linear thinning as expected from diluted polymeric solutions [17]. It isworth mentioning that while only pure PCL was analyzed here, changesin the system's viscosity when using molten IBU-PCL blend were ob-served, which may result in different values in the analyses.

3.1.3. Melt interaction with the superoleophobic substrate and sphereformation

Fig. 3 illustrates the interaction of a molten jet with a super-oleophobic surface compared to an aluminum surface, signifying thefinal step in the formation of the sphere. The hierarchical micrometricstructure and the coarse roughness allowed the repulsion of the meltfrom the surface, exhibiting high contact angles of more than 150° (asappropriate for lipophilic liquids) [23]; hence, allowing a sphericaldroplet to form. The use of a non-wetting surface in this step is thecornerstone of the entire system to successfully produce sphericaldroplets. Moreover, the adequate filament velocity initially generatedenabled the conservation of the Cassie-Bexter model [38], resulting in aspherical shape of the droplet (high apparent static contact angle) [39].Inversely, when the melt was ejected onto a simple aluminum surface,owing to the high surface tension (γ=33.2mNm−1) similar to theinterfacial tension of the melt (γ=35.9mNm−1) [40], the droplet didnot preserve its spherical shape and a Cassie-Bexter to Wenzel transitionwas observed [38]. It could be observed that, initially, as the filamentimpacted the oleophobic surface, the surface was wetted (Fig. 3, @0ms). The filament exhibited some inertial distortion at approximately10.4 ms after impact (Fig. 2A), and thereafter formed a spherical shapeunder the action of capillary forces [39]. Consequently, owing to thehigh temperature of the melt, its impact and spreading occurred in amuch shorter time scale (approximately 60ms) compared to the soli-dification time (approximately 2 s under ambient conditions). In suchcases, the droplet ended up in a quasi-equilibrium shape, and the soli-dification angle was essentially equal to the apparent static contact angle[39].

3.2. Microspheres characterization and in-vitro drug dissolution studies

3.2.1. Microspheres size, shape, and morphologyOnce the microspheres were solidified, their roundness, i.e.,

“sphericity,” was determined using a contact angle analyzer. PCL and30% IBU-loaded PCL microspheres exhibited high contact angles, ran-ging from 154 ± 3° to 171 ± 4° (Fig. 4A i, ii), respectively. Poly(lactic-co-glycolic acid) (PLGA) and 30% IBU-loaded PLGA micro-spheres had high contact angles as well, ranging from 167 ± 6° to

Fig. 3. Frames obtained from high-speed imaging capturing the evolution of the shape of a molten 30% IBU-PCL blend interacting with either a superoleophobicsurface (top) and preserving its spherical shape after 1.8 s or an aluminum surface (bottom), in which the droplet gradually flattens (See SI Video 4 and SI Video 5,respectively). Scale bar: 500 µm. These frames emphasis the significance of using non-wetting surfaces responsible for the formation of spherical droplets.

T. Shpigel et al. European Journal of Pharmaceutics and Biopharmaceutics 127 (2018) 398–406

403

169 ± 4° (Fig. 4A iii, iv), which were close to 180°—the value of aperfect sphere [41]. The average size of the IBU-loaded microsphereswas determined using scanning electron microscopy (SEM) measure-ments as 372 ± 19 µm and 486 ± 15 µm for PLGA and PCL micro-spheres, respectively. Fig. 4B illustrates the distinguished smooth sur-face morphology of the amorphous PLGA microspheres (loaded with30% IBU) and the apparent spherulite-like structure of the semi-crys-talline PCL microspheres (also loaded with 30% IBU).

3.2.2. Drug encapsulation efficiencyThe mean encapsulation efficiency was determined by HPLC as

95.3 ± 2.7% and 83.3 ± 5.3% for 30% IBU-loaded PCL and PLGAmicrospheres, respectively. IBU-loaded microspheres produced viaemulsification-evaporation have been reported to have the maximumencapsulation efficiencies of 25–84% [30,42] and 30–85% [25] for PCLand PLGA microspheres, respectively. The IBU entrapment efficiency(EE%) values obtained with SPHRINTt are in the highest scale of thosereported in the literature, and even above that in the case of PCL. Thismay be attributed to the followings: (i) adequate miscibility of IBU inPCL or PLGA at the processing temperature (80 or 150 °C), (ii) minimalprocess-derived drug loss through utilization of molten components in aclosed environment, and (iii) eliminating the use of organic solventsinto which the drug can diffuse and can be excluded during solventremoval [6]. Thus stable, consistent, and reproducible results wereobtained. It should be noted that the lower EE% in the PLGA micro-spheres compared with the PCL may be attributed to possible de-gradation of the IBU, or evaporation during the millisecond ejectionthrough the orifice [43].

3.2.3. In vitro drug dissolution studiesThe in vitro IBU cumulative release profiles of 30% IBU-loaded PCL

and PLGA microspheres during 17 days of incubation are illustrated inFig. 5. With a uniform initial amount of IBU (0.3 mg/ml), the 30% IBU-PCL microspheres exhibited an initial rapid IBU release during the firstfour days, followed by a plateaued, slower release profile, whereas the30% IBU-PLGA microspheres presented a linear, constant IBU release.

This conspicuous difference in the release profiles is primarily a resultof the encaptivating polymeric matrix. Here we compare the in vitrodrug release of the semi-crystalline PCL, which considered to as amonolithic material, whilst the drug release rate with the amorphousPLGA matrix is dependent on the rate of its degradation [44,45]. Fur-thermore, the PCL microspheres did not exhibit a burst release, nor-mally observed with microspheres prepared using the solvent-eva-poration method, manifested as approximately 80% within the first24 h [30]. This finding is expected from blends prepared via meltprocessing [46,24]. Burst release is usually an outcome of IBU crystalsappearing on the surface of the microsphere—a phenomenon associatedwith the use of the solvent-evaporation method and is less common inthe case of melt-derived methods [47]. The observed IBU release pro-files obtained for 30% IBU-PLGA microspheres significantly differedfrom those of microspheres prepared using the solvent-evaporationmethod. Fernández-Carballido et al. [25], and Klose et al. [48], re-ported full IBU dissolution (approximately 90–100%) within

Fig. 4. (A) Optical images of (i) neat PCL (154 ± 3°), (ii) 30% IBU-PCL (171 ± 4°), (iii) neat PLGA (167 ± 6°), and (iv) 30% IBU-PLGA (169 ± 4°) microspheres,cooled at room temperature (RT). The values in brackets denote the “sphericity,” expressed as the contact angle values (mean ± S.E.M., n= 10). Scale bar: 200 µm.(B) SEM images of (i) 30% IBU-PCL and (ii) 30% IBU-PLGA microspheres, cooled at RT. Scale bars: 100 µm.

0

10

20

30

40

50

60

0 5 10 15 20

Cum

ulat

ive

IBU

rele

ase

/ %

Time / days30%IBU-PCL 30%IBU-PLGA

Fig. 5. Cumulative release of IBU from 30% IBU-loaded PCL and PLGA mi-crospheres, incubated at 37 °C with mild agitation up to 17 days. Quantified IBUamount is expressed as cumulative release, according to Eq. (1). Symbols re-present mean ± S.E.M. (n= 3).

T. Shpigel et al. European Journal of Pharmaceutics and Biopharmaceutics 127 (2018) 398–406

404

approximately six days of incubation in a first-order fashion, whereasthe SPHRINT printed microspheres exhibited a linear, zero-order re-lease profile, releasing 23.4 ± 0.4% of the IBU after 17 days of in-cubation under similar conditions. This unambiguous difference mayalso be attributed to the use of a melt-based technique, which is alsoobserved in the PCL-based microspheres [49].

4. Conclusions

The process of jetting drug-loaded polymeric melt droplets on su-peroleophopbic surfaces, or SPHRINT (SI Video 6), enables us to pro-duce near-perfect microspheres. While the process of SPHRINT ap-peared straightforward initially, we discovered intricate physicalphenomena governing the mechanism of sphere-formation; besideprocess and performance efficiencies, which in turn render microsphereproducts more accessible [50]. SPHRINT printing eliminates the use oforganic solvents and surfactants; it offers microspheres with re-producible size, shape, and morphology within and between the bat-ches (see SI); and the produced microspheres can be easily collectedowing to their spherical shape (SI Video 7). Furthermore, we believethat SPHRINT may turn microsphere production ubiquitous, allowingfor desktop manufacturing of microspheres easily scalable to industrialquantities (with a production rate of 25 Hz, in 1 h, 4.7 g of PCL mi-crospheres may be printed from a single printing head). Finally,SPHRINT may provide a new dimension in reservoir-injectable drugdelivery technologies, enabling the employment of multifarious poly-mers for microsphere production and tuned release profiles.

Supplementary Video 6.

Supplementary Video 7.

Acknowledgments

We thank The Innovation Center ACT at Shenkar for partiallyfunding this project.

Declaration of interest

The authors declare no competing financial interests.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, in theonline version, at 10.1016/j.ejpb.2018.03.006.

References

[1] A.C. Anselmo, S. Mitragotri, An overview of clinical and commercial impact of drugdelivery systems, J. Control. Release 190 (2014) 15–28, http://dx.doi.org/10.1016/j.jconrel.2014.03.053.

[2] C. Berkland, M.J. Kipper, B. Narasimhan, K. Kim, D.W. Pack, Microsphere size,precipitation kinetics and drug distribution control drug release from biodegradablepolyanhydride microspheres, J. Control. Release 94 (2004) 129–141, http://dx.doi.org/10.1016/j.jconrel.2003.09.011.

[3] M.A. Woodruff, D.W. Hutmacher, The return of a forgotten polymer – poly-caprolactone in the 21st century, Prog. Polym. Sci. 35 (2010) 1217–1256, http://dx.doi.org/10.1016/j.progpolymsci.2010.04.002.

[4] V.B. Kraus, P.G. Conaghan, H.A. Aazami, P. Mehra, A.J. Kivitz, J. Lufkin, J. Hauben,J.R. Johnson, N. Bodick, Synovial and systemic pharmacokinetics (PK) of triamci-nolone acetonide (TA) following intra-articular (IA) injection of an extended-releasemicrosphere-based formulation (FX006) or standard crystalline suspension in pa-tients with knee osteoarthritis (OA), Osteoarthr. Cartil. 26 (2018) 34–42, http://dx.doi.org/10.1016/j.joca.2017.10.003.

[5] A. Gharsallaoui, G. Roudaut, O. Chambin, A. Voilley, R. Saurel, Applications ofspray-drying in microencapsulation of food ingredients: an overview, Food Res. Int.40 (2007) 1107–1121, http://dx.doi.org/10.1016/j.foodres.2007.07.004.

[6] R. Deshmukh, P. Wagh, J. Naik, Solvent evaporation and spray drying technique formicro- and nanospheres/particles preparation: a review, Dry. Technol. 34 (2016)1758–1772, http://dx.doi.org/10.1080/07373937.2016.1232271.

[7] B. Albertini, N. Passerini, F. Pattarino, L. Rodriguez, New spray congealing atomizerfor the microencapsulation of highly concentrated solid and liquid substances, Eur.J. Pharm. Biopharm. 69 (2008) 348–357, http://dx.doi.org/10.1016/j.ejpb.2007.09.011.

[8] R. Kamble, A. Kumar, K. Mahadik, A. Paradkar, Ibuprofenglyceryl monostearate(GMS) beads using melt solidification technique: effect of HLB, Int. J. Pharm.Pharm. Sci. 2 (2010) 100–104.

[9] H. Reithmeier, J. Herrmann, A. Göpferich, Development and characterization oflipid microparticles as a drug carrier for somatostatin, Int. J. Pharm. 218 (2001)133–143, http://dx.doi.org/10.1016/S0378-5173(01)00620-2.

[10] M. Maheshwari, A.R. Ketkar, B. Chauhan, V.B. Patil, A.R. Paradkar, Preparation andcharacterization of ibuprofen-cetyl alcohol beads by melt solidification technique:effect of variables, Int. J. Pharm. 261 (2003) 57–67, http://dx.doi.org/10.1016/S0378-5173(03)00292-8.

[11] P. Calvert, Inkjet printing for materials and devices, Chem. Mater. 13 (2001)3299–3305, http://dx.doi.org/10.1021/cm0101632.

[12] B.-J. de Gans, P.C. Duineveld, U.S. Schubert, Inkjet printing of polymers: state of theart and future developments, Adv. Mater. 16 (2004) 203–213, http://dx.doi.org/10.1002/adma.200300385.

[13] R.D. Boehm, P.R. Miller, J. Daniels, S. Stafslien, R.J. Narayan, Inkjet printing forpharmaceutical applications, Mater. Today. 17 (2014) 247–252, http://dx.doi.org/10.1016/j.mattod.2014.04.027.

[14] H.-J. Trost, T. Chen., M.E. Grove, R. Hoenigman, D. Radulescu, D. Silva, D. Taylor,A.-K.C. Nallani, J.-B. Lee, Ink-jetting polymers for electronic, optical and biomedicalmicro-devices, ASPE Spring Top. Meet. Precis. Micro/Nano Scale Polym. BasedCompon. Device Fabr. (2005) 44–49.

[15] Xerox, Xerox ColorQube Solid Ink Printers and Multifunction Printers, (n.d.). < http://www.office.xerox.com/latest/SLDBR-02> .

[16] https://www.solidscape.com/, solidscape, (n.d.).[17] J. Dinic, M. Biagioli, V. Sharma, Pinch-off dynamics and extensional relaxation

times of intrinsically semi-dilute polymer solutions characterized by dripping-onto-substrate rheometry, J. Polym. Sci. Part B Polym. Phys. 55 (2017) 1692–1704,http://dx.doi.org/10.1002/polb.24388.

[18] C. Clasen, P.M. Phillips, L. Palangetic, J. Vermant, Dispensing of rheologicallycomplex fluids: the map of misery, AIChE J. 58 (2012) 3242–3255, http://dx.doi.org/10.1002/aic.13704.

[19] A.M.S. Costa, M. Alatorre-Meda, N.M. Oliveira, J.F. Mano, Biocompatible polymericmicroparticles produced by a simple biomimetic approach, Langmuir 30 (2014)4535–4539, http://dx.doi.org/10.1021/la500286v.

[20] A.M.S. Costa, M. Alatorre-Meda, C. Alvarez-Lorenzo, J.F. Mano, Superhydrophobicsurfaces as a tool for the fabrication of hierarchical spherical polymeric carriers,Small 11 (2015) 3648–3652, http://dx.doi.org/10.1002/smll.201500192.

[21] M.I. Rial-Hermida, N.M. Oliveira, A. Concheiro, C. Alvarez-Lorenzo, J.F. Mano,Bioinspired superamphiphobic surfaces as a tool for polymer- and solvent-in-dependent preparation of drug-loaded spherical particles, Acta Biomater. 10 (2014)4314–4322, http://dx.doi.org/10.1016/j.actbio.2014.06.009.

[22] X. Deng, M. Paven, P. Papadopoulos, M. Ye, S. Wu, T. Schuster, M. Klapper,D. Vollmer, H.J. Butt, Solvent-free synthesis of microparticles on superamphiphobicsurfaces, Angew. Chem. – Int. Ed. 52 (2013) 11286–11289, http://dx.doi.org/10.1002/anie.201302903.

[23] M.K. Chaudhury, G.M. Whitesides, Correlation between surface free energy andsurface constitution, Science (80-.) 255 (1992) 1230–1232. 10.1126/science.255.5049.1230.

[24] G.V. Salmoria, F. Sibilia, V.G. Henschel, S. Fare, M.C. Tanzi, Structure and prop-erties of polycaprolactone/ibuprofen rods prepared by melt extrusion for im-plantable drug delivery, Polym. Bull. 74 (2017) 1–15, http://dx.doi.org/10.1007/s00289-017-1999-x.

[25] A. Fernández-Carballido, R. Herrero-Vanrell, I.T. Molina-Martínez, P. Pastoriza,Biodegradable ibuprofen-loaded PLGA microspheres for intraarticular

T. Shpigel et al. European Journal of Pharmaceutics and Biopharmaceutics 127 (2018) 398–406

405

administration: effect of Labrafil addition on release in vitro, Int. J. Pharm. 279(2004) 33–41, http://dx.doi.org/10.1016/j.ijpharm.2004.04.003.

[26] J. Schindelin, I. Arganda-Carreras, E. Frise, V. Kaynig, M. Longair, T. Pietzsch,S. Preibisch, C. Rueden, S. Saalfeld, B. Schmid, J.-Y.J.-Y. Tinevez, D.J. White,V. Hartenstein, K. Eliceiri, P. Tomancak, A. Cardona, K. Liceiri, P. Tomancak, Fiji:an open source platform for biological image analysis, Nat. Methods 9 (2012)676–682, http://dx.doi.org/10.1038/nmeth.2019.Fiji.

[27] S. Tamilvanan, B. Sa, Effect of production variables on the physical characteristicsof ibuprofen-loaded polystyrene microparticles, J. Microencapsul. 16 (1999)411–418, http://dx.doi.org/10.1080/026520499288870.

[28] D. Klose, F. Siepmann, K. Elkharraz, J. Siepmann, PLGA-based drug delivery sys-tems: importance of the type of drug and device geometry, Int. J. Pharm. 354(2008) 95–103, http://dx.doi.org/10.1016/j.ijpharm.2007.10.030.

[29] N. Bolourtchian, K. Karimi, R. Aboofazeli, Preparation and characterization ofibuprofen microspheres, J. Microencapsul. 22 (2005) 529–538, http://dx.doi.org/10.1080/02652040500161941.

[30] K.J. Zhu, Y. Li, H.L. Jiang, H. Yasuda, A. Ichimaru, K. Yamamoto, P. Lecomte,R. Jerome, Preparation, characterization and in vitro release properties of ibu-profen-loaded microspheres based on polylactide, poly(∋-caprolactone) and theircopolymers, J. Microencapsul. 22 (2005) 25–36, http://dx.doi.org/10.1080/02652040400026350.

[31] F. Yao, K. Weiyuan John, Drug release kinetics and transport mechanisms of non-degradable and degradable polymeric delivery systems, Expert Opin. Drug Deliv. 7(2010) 429–444, http://dx.doi.org/10.1517/17425241003602259.

[32] H. Yang, Y. He, C. Tuck, R. Wildman, R. Hague, High viscosity jetting system for 3Dreactive inkjet printing, in: Twenty Forth Annu. Int. Solid Free. Fabr. Symp. – AnAddit. Manuf. Conf., 2013, pp. 505–513.<https://sffsymposium.engr.utexas.edu/Manuscripts/2013/2013-40-Yang.pdf > .

[33] S. Kalambur, S.S.H. Rizvi, Rheological behavior of starch–polycaprolactone (PCL)nanocomposite melts synthesized by reactive extrusion, Polym. Eng. Sci. 46 (2006)650–658, http://dx.doi.org/10.1002/pen.20508.

[34] B. Derby, Inkjet printing of functional and structural materials: fluid property re-quirements, feature stability, and resolution, Annu. Rev. Mater. Res. 40 (2010)395–414, http://dx.doi.org/10.1146/annurev-matsci-070909-104502.

[35] P.C. Duineveld, M.M. De Kok, M. Buechel, A.H. Sempel, K.A.H. Mutsaers, P. Van deWeijer, I.G.J. Camps, T.J.M. Van den Biggelaar, J.-E.J.M. Rubingh, E.I. Haskal, Ink-jet printing of polymer light-emitting devices, Proc. SPIE – Int. Soc. Opt. Eng. 4464(2002) 59–67, http://dx.doi.org/10.1117/12.457460.

[36] G.H. McKinley, Dimensionless groups for understanding free surface flows ofcomplex fluids, Soc. Rheol. Bull. (2005) 8 http://web.mit.edu/nnf/publications/GHM84.pdf.

[37] D.T. Papageorgiou, On the breakup of viscous liquid threads, Phys. Fluids 7 (1995)1529–1544, http://dx.doi.org/10.1063/1.868540.

[38] M. Nosonovsky, Materials science: slippery when wetted, Nature 477 (2011)412–413, http://dx.doi.org/10.1038/477412a.

[39] F. Gao, A.A. Sonin, Precise deposition of molten microdrops: the physics of digitalmicrofabrication, Proc. R. Soc. A Math. Phys. Eng. Sci. 444 (1994) 533–554, http://dx.doi.org/10.1098/rspa.1994.0037.

[40] B.B. Sauer, G.T. Dee, Surface tension and melt cohesive energy density of polymermelts including high melting and high glass transition polymers, Macromolecules35 (2002) 7024–7030, http://dx.doi.org/10.1021/ma0202437.

[41] Y.C. Jung, B. Bhushan, Wetting behavior of water and oil droplets in three-phaseinterfaces for hydrophobicity/philicity and oleophobicity/philicity, Langmuir 25(2009) 14165–14173, http://dx.doi.org/10.1021/la901906h.

[42] N. Carreras, V. Acuña, M. Martí, M.J. Lis, Drug release system of ibuprofen in PCL-microspheres, Colloid Polym. Sci. 291 (2013) 157–165, http://dx.doi.org/10.1007/s00396-012-2768-x.

[43] S. Ramukutty, E. Ramachandran, Reaction rate models for the thermal decom-position of ibuprofen crystals, J. Cryst. Process Technol. 4 (2014) 71–78, http://dx.doi.org/10.4236/jcpt.2014.42010.

[44] J. Siepmann, F. Siepmann, Mathematical modeling of drug delivery, Int. J. Pharm.364 (2008) 328–343, http://dx.doi.org/10.1016/j.ijpharm.2008.09.004.

[45] A.N. Ford Versypt, D.W. Pack, R.D. Braatz, Mathematical modeling of drug deliveryfrom autocatalytically degradable PLGA microspheres – a review, J. Control.Release 165 (2013) 29–37, http://dx.doi.org/10.1016/j.jconrel.2012.10.015.

[46] H. Lian, Z. Meng, Melt electrospinning vs. solution electrospinning: a comparativestudy of drug-loaded poly (ε-caprolactone) fibres, Mater. Sci. Eng. C. 74 (2017)117–123, http://dx.doi.org/10.1016/j.msec.2017.02.024.

[47] R. Bodmeier, H. Chen, Preparation and characterization of microspheres containingthe anti-inflammatory agents, indomethacin, ibuprofen, and ketoprofen, J. Control.Release 10 (1989) 167–175, http://dx.doi.org/10.1016/0168-3659(89)90059-X.

[48] D. Klose, F. Siepmann, J.F. Willart, M. Descamps, J. Siepmann, Drug release fromPLGA-based microparticles: effects of the “microparticle:bulk fluid” ratio, Int. J.Pharm. 383 (2010) 123–131, http://dx.doi.org/10.1016/j.ijpharm.2009.09.012.

[49] R. Bhardwaj, J. Blanchard, In vitro evaluation of poly (d,l-lactide-co-glycolide)polymer-based implants containing the α-melanocyte stimulating hormone analog,Melanotan-I, J. Control. Release 45 (1997) 49–55, http://dx.doi.org/10.1016/S0168-3659(96)01544-1.

[50] K. Kim, D. Pack, Microspheres for drug delivery, in: M. Ferrari, A.P. Lee, J. Lee(Eds.), BiolBiomed. Nanotechnol. Springer, 2006, pp. 18–50 http://www.bookme-trix.com/detail/book/c35674c6-7c47-4d15-ab67-d308dd115141#downloads.

T. Shpigel et al. European Journal of Pharmaceutics and Biopharmaceutics 127 (2018) 398–406

406