Encapsulation of mesenchymal stem cells in glycosaminoglycans‐chitosan polyelectrolyte microcapsules using electrospraying technique: Investigating capsule morphology and cell viability

Abstract Polyelectrolyte microcapsules are modular constructs which facilitate cell handling and assembly of cell‐based tissue constructs. In this study, an electrospray (ES) encapsulation apparatus was developed for the encapsulation of mesenchymal stem cells (MSCs). Ionic complexation between glycosaminoglycans (GAGs) and chitosan formed a polyelectrolyte complex membrane at the interface. To optimize the capsules, the effect of voltage, needle size and GAG formulation on capsule size were investigated. It was observed that by increasing the voltage and decreasing the needle size, the capsule size would decrease but at voltages above 12 kV, capsule size distribution broadened significantly which yields lower circularity. Increase in GAG viscosity resulted in larger microcapsules and cell viability exhibited no significant changes during the encapsulation procedure. These results suggest that ES is a highly efficient, and scalable approach to the encapsulation of MSCs for subsequent use in bioprinting and other modular tissue engineering or regenerative medicine applications.


| INTRODUCTION
Progress in the biofabrication of implantable, engineered tissue is slowed by the challenge of assembling three-dimensional tissue with a fully integrated microvasculature. One strategy for tissue and vessel assembly involves the fusion-based assembly of endothelialized cell spheroids. This approach holds significant promise, but it is hampered by the need to provide substantial mechanical and organizational support to prevent uncontrolled cell aggregation and associated low vas- Unencapsulated cells or cell spheroids are prone to fluid shear damage in dynamic cultures and may also be subject to disruption of organization in vivo due to postimplantation migration. This can diminish both cell metabolic activity and tissue level function, reducing the efficiency of the implanted system.. 1 One approach to overcoming this obstacle is to encapsulate cells, spheroids, or organoids within biopolymer membranes, which degrade after implantation. These membranes require high permeability to nutrients, oxygen, and wastes and therefore the diffusive transport characteristics must be considered when choosing appropriate biopolymers or hydrogels for use in cell encapsulation. 2 As a result, optimizing the parameters influencing the encapsulation process is of great importance. These parameters include the encapsulation chemistry and the biopolymers used. There are multiple methods of cell encapsulation used widely, including microfluidics-based encapsulation, micromolding methods, and droplet/air methods. Microfluidic techniques have multiple advantages including high control over capsule size and morphology, which can generate microcapsules as small as 100 μm. However, the choice of biomaterials to be used is limited in these systems as the solutions used should have high gelation capacity such as alginate/calcium chloride or high interfacial tension such as oil/water. 3,4 Micromolding methods facilitate the formation of cell spheroids and also encapsulation of these spheroids but require specific mold designs and crosslinking methods. 5 Air methods are easy and convenient methods of cell encapsulation, but they lack adequate control over size and uniformity of the microcapsules at the smaller size ranges. 6 Furthermore, a general drawback for these encapsulation methods is an inability to implement mass production of microcapsules. Hence, there is a demand for an encapsulation method that can produce large quantities of capsules and is compatible with a wide range of biomaterials.
Another challenge in many encapsulation techniques is the difficulty in controlling the microcapsule size and uniformity. [7][8][9] We previously reported on the use of cells and cell spheroids encapsulated within glycosaminoglycan (GAG)-chitosan polyelectrolyte membranes as a tool for tissue assembly using modular tissue engineering principles. 2 Improvements to the technique required improvements in the droplet generation method-specifically a reduction in the droplet sizes and hence the resultant capsule sizes to enhance the nutrient transport in the fused tissue construct. Electrospraying (ES) is a method of liquid atomization using high electrical fields to overcome the surface tension of the liquid. 10 The examples of using this technique include cell and particle encapsulation, drug delivery, nanoparticle synthesis, and film deposition. 10,11 This technique operates on the principle of an applied potential difference between two electrodes. 12 To generate electrosprayed droplets, the polymer solution is extruded through a metal needle and the tip of the needle is maintained at a high voltage relative to the counter electrode, which can be a ring electrode through which droplets pass, or a grounded solution into which droplets are collected. These droplets are highly charged and can be in the range of nanometers to micrometers. In the ES technique, several interdependent parameters influence droplet size, size distribution, encapsulation efficiencies, and loading capacities. [13][14][15] These parameters include physical properties of the liquids, voltage, needle gauge, distance to collector/counter electrode, solution flowrate, and surfactant concentration. 16 In general, droplets formed will be in one of these modes: dripping, pulsating, cone-jet, and multi-jet mode. To obtain uniform droplets, the affecting parameters should be adjusted in a way that a cone-jet is formed at the tip of the needle. 17 In cone-jet mode, the liquid meniscus forms an axisymmetric, uniform cone termed as the Taylor cone. 18 As soon as the charge accumulated on the droplet overcomes the surface tension of the liquid, a uniform jet is formed. It has been seen experimentally that ES in the cone-jet mode happens when the liquid conductivity is in the range of 10 −5 to 10 −11 S/m, a range within which all the semiconducting liquids fall. If the conductivity is either lower or higher than this range, droplets will be formed in dripping or multi-jet mode, which are unfavorable with regard to uniformity. 19 To date, the characteristics of electrosprayed droplets containing cells are not completely understood and it is important to proceed in a step-wise manner to understand the relationship between processing parameters and characteristics of electrospray-generated capsules before progressing to the inclusion of high value, fragile cells, and bioactive molecules. 20 In addition to the cell type, the choice of biomaterial used for encapsulation is of great importance. 21 Electrospray droplet generation can be adapted as a technique to encapsulate cells using the GAG-chitosan complex method. 2 In this research, microcapsules were formed by the complex coacervation between chitosan and GAG. 22,23 Chitosan is the second most abundant polysaccharide after cellulose. 24 In dilute acidic solution, chitosan amino groups protonate and the polymer can subsequently form ionic complexes with a wide variety of natural or synthetic anionic species. 25 Other specific characteristics of the chitosan include: antibacterial, antifungal, mucoadhesive, analgesic, and haemostatic properties. 26 GAGs are also a family of highly sulfated, complex polysaccharides that play a variety of important biological roles in the body. All GAGs are negatively charged due to the presence of carboxyl and/or acidic sulphate groups. GAGs are widely distributed in animals and are essential for maintaining the integrity of the connective tissues. In solution, they are highly viscous and have low compressibility. Hence, they function as lubricating agents in articulating joints. 27 They also bind and modulate the biological activity of many peptide growth factors and extracellular matrix proteins. 28 Therefore, their use as components of the microcapsule structure may have beneficial effects on cell metabolic activity and functionality. 2,28,29 In this study, capsule formation from ES-generated droplets was investigated together with the viability and growth of mesenchymal stem cells (MSCs). MSCs were used as the model cell type due to their multipotency and broad use in tissue engineering systems. 29,30 Although there are many researches on using the ES technique for fabricating microparticles and microcapsules for drug delivery purposes and delivering different cells on different scaffolds, there are only a few researches concentrated on using this method as an encapsulation technique using biomaterial solutions and stem cells together. 14,21,[31][32][33][34] To the best of our knowledge, this research is the first report on encapsulation of MSCs in the chitosan and GAG microenvironment using the ES technique. Capsules provide structural organization, zonation, shear protection, and scalability. In general, the novelty of this work lies in the following features. First, the ES method allows us to make uniform microcapsules at a large scale. This is one of the advantages of this technique. The only competitor to the ES method with the same capabilities are the microfluidic encapsulation methods, but as the interfacial tension of the solutions used in this research is low (below 0.1 mNm −1 ), reliable formation of microcapsules via microfluidics is extremely difficult. 35,36 Furthermore, microfluidic encapsulation using low interfacial tension materials requires additional mechanical actuators or the addition of organic liquids to stimulate phase separation or increase interfacial tension between two liquids, changes that interfere with the overall biocompatibility of the system. [37][38][39] Finally, the biomaterials used for encapsulation in this research have not been previously used in the ES method. GAG-chitosan capsules made by the ES method provide improved cell handling, structural organization, zonation, shear protection, and scalability. We propose these capsules as a modular tissue engineering platform for use as building blocks of 3D tissue structures.

| RESULTS AND DISCUSSION
One of the main goals of this research was to make uniform, small (200-500 μm) microcapsules using the ES technique ( Figure 1).
Although it has been shown that this method is able to produce uniform and monodispersed droplets, it is still a challenge to achieve this goal because of high interrelations between parameters affecting the process. 14,33 In this section, the effects of these parameters on capsule size, uniformity, shape, and cell viability were evaluated.

| Microcapsule size evaluation
In the first set of experiments, the microcapsule size in each test was analyzed and optimized conditions were determined. It was observed that the capsule diameter was decreased by increasing the voltage from 10 to 14 kV in all three GAG types with both needle sizes. However, the reduction in capsule diameter was more obvious with the first GAG formulation (0.5% hyaluronic acid [HA] + 4% chondroitin 4-sulfate [CSA]) compared to the two others. This was likely due to the lower viscosity of the solution when using a 0.5% HA concentration. Solutions with higher viscosity require higher voltages in order for electrostatic forces to overcome the surface tension of the liquid. 15 Therefore, it is harder for them to detach from the tip of the needle compared to lower viscosity solutions, resulting in larger microcapsules (Figures 2 and 3).
As can be seen in Figures 2 and 3, decreasing the needle diameter decreases the average capsule size. This trend was seen in all three GAG formulations. Among the three GAG formulations, the 2% HA formulation had the highest viscosity compared to the other two. The effect of viscosity can be also seen in the lower uniformity of the microcapsules tension results in more difficult detachment of the droplets from the needle tip and results in a droplet elongation that contributes to tail formation in this GAG formulation. The higher viscosity retards the tendency of free droplets to restore a spherical shape during free fall, and the teardrop shape is ultimately immobilized upon formation of the capsule membrane. Capsules with the higher HA content also showed thicker walls.
Quantitative results for the capsule size assessment are shown in Apart from the capsule size, the uniformity of the microcapsules is another factor that affects diffusive transport in implantation procedures. 40 The results for the microcapsules formed by an 18 G (OD) needle are shown in Figure 5a, with similar trends being seen for the 22 G (OD) needle size in Figure 5b. These results suggest that increasing the voltage from 12 to 14 kV results in a reduction in uniformity across the whole range of microcapsules as the circularity deviates from unity. Circularity deviation from unity was also more visible for the higher viscosity formulations compared to lower viscosities. These results also suggest that using solutions with higher viscosities, in addition to yielding larger capsules, also generates less uniform microcapsules across the whole range of voltages and needle sizes. Moreover, by comparing the results for two needle sizes, it can be clearly seen that the effect of needle size on circularity of microcapsules is negligible compared to the effects of voltage and GAG formulation. Similar results have been reported in other ES studies where deviations from circularity have been reported as being related to higher voltages and higher viscosities of the electrosprayed solution. 41 It should be noted that deviations from circularity may actually have desirable effects on encapsulated cells, as the surface to volume ratio of a given capsule increases with decreasing circularity. The net effect may be a reduction in diffusion distances within the capsules and an overall enhancement of nutrient and oxygen availability. Use of higher voltages in the ES method results in smaller microcapsules, but this has been reported to affect the cell viability and metabolic activity to a significant degree, as higher voltages can potentially inflict greater cell membrane damage leading to cell death. 32,34,42 Hence, a voltage of 12 kV with a 22-G needle size was taken as the optimized condition to achieve both smaller and more uniform microcapsules compared to the other voltages and needle size combinations. The capsules formed under these optimized conditions were later used to analyze the wall morphology and cell viability. Although not evaluated here, thermal effects at high voltages have also been reported to damage the cell wall, an effect that occurs at high voltages. 43

| Capsule wall morphology
To analyze the morphology of the microcapsule membranes in different GAG formulations, scanning electron microscopy (SEM) images were captured from the surface and interior of ruptured microcapsules.
These images clearly show the wall thickness and the porous microstructure of the capsule walls ( Figure 6). All microcapsules were hollow and had a porous polyelectrolyte complex membrane. The wall porosity

| Biopolymer materials
The materials used in the encapsulation were: chitosan (90% deacetylation) with medium molecular weight, CSA from bovine tra-   (1)) shows the effect of these parameters:  Table 1 displays the set of FIGURE 8 Encapsulated cell viability in the three formulations immediately before and after encapsulation, as well as 30 days post-encapsulation. Data are mean and standard deviation from three independent culture runs. Viability percentages were obtained by counting viable and nonviable cells after fluorescence imaging of capsules exposed to the viability probes Calcein-AM and Ethidium homodimer. All results are capsules formed at 12 kV using the 22G needle

| Microcapsule morphology and cell viability analysis
The original formulation for the encapsulation technique was developed by Matthew et al. 22  ties than larger sizes. 46 The microcapsules formed in this range were used later for cell viability tests.
For SEM imaging, microcapsules were fixed in 2% glutaraldehyde in a cacodylate buffer, washed with water, dehydrated in ethanol, and flash frozen in liquid nitrogen. The frozen mass was lyophilized until dry and then sputter coated with gold for SEM imaging at an acceleration voltage of 15 kV on a JEOL JSM-7600 microscope. In addition to size comparisons, the shape of the microcapsules was analyzed as an indicator of the capsule uniformity. Here, the circularity (Equation (2)) was used as a tool to evaluate the uniformity of microcapsules formed in each voltage with the corresponding needle size and GAG formulation. "A" and "P" represent area and perimeter of the microcapsules, respectively. Values equal to 1 represent a perfect circle while smaller values suggest deviations from circularity, and a higher surface-tovolume ratio of the microcapsules. All results were calculated for a minimum of 100 capsules in each experiment.
Cell viability was investigated using Calcein-AM/ethidium homodimer method (Cytotoxity kit L3224; Invitrogen). As the microcapsule wall acts as a diffusion barrier, high concentrations of the dyes were used. In brief, 4 μL of Calcein-AM stock and 4 μL of ethidium homodimer stock solutions were added to 1 mL of PBS and the solution was mixed thoroughly. About 600 μL of the prepared dye solution was added to each well of a 12 well plate containing the microcapsules.
The microcapsules were then incubated for 30 min and washed afterward with PBS to remove the background dye. Microcapsules were then imaged under a fluorescence microscope (Nikon Diaphot 300) and the number of dead and live cells in a sample population of~100 microcapsules were counted.

| CONCLUSIONS
In this study, the effects of electrospray droplet generation on characteristics of GAG-chitosan microcapsules and encapsulated MSCs were investigated. The ES technique yielded fairly uniform microcapsules at sizes that were typically subject to large size dispersion using the traditional air atomization methods. Higher voltages produced smaller capsules with a narrow size distribution, but also resulted in less spherical capsule shapes. Solution viscosity was determined to be an important variable in that smaller, more circular capsules could be generated easier using lower solution viscosities than by either increasing voltages or reducing needle diameters. HA-based capsules were found to support greater cell adhesion to capsule walls than the carboxymethyl cellulose-based formulation. These results demonstrated that electrospray droplet generation is a viable and superior alternative to previous air flow methods for encapsulating MSCs within glycosaminoglycan-chitosan polyelectrolyte membranes.