Page 115 - IJB-5-2
P. 115

Ghalayini S, et al.
           aqueous solution of 50% filtered ethanol was loaded into   flow rates were best for each peptide as several options
           a 10 mL syringe to be pushed through the side channels   were tested to obtain the largest quantity of NPs with the
           of the junctions, as shown in Figure 2.             most uniform size distribution.
             The  ratios  of  the  flow  rates  of  the  ethanol  solution
           to the peptide solution were found to be crucial for NP   2.3 Printed Hydrogel Sample Preparation
           production and thus an optimization process, described   Two vials of CH-01 and CH-02 peptide powders, 18 mg
           in  the  following section,  was employed  to  determine   each, were weighed out and then dissolved in 1  mL
           the ideal ratio. Before starting the actual NP production
           process, the microfluidic chip was stabilized by running   of Milli-Q water  by vortexing  and  sonicating  into  a
                                                               homogenous solution. For the samples containing NPs,
           the syringe pumps at the desired starting flow rates with
           the ethanol solution and with water replacing the peptide   around 0.9 mg of lyophilized NPs were weighed out and
           solution.  This  stabilization  step  ensures  that  the  flow   dissolved in the peptide solution.
           is constant and consistent across  all the channels and   A custom-designed  3D bioprinter  was set up with
           junctions to avoid variations in morphology or decreases   commercial  microfluidic  pumps  as  described  in  our
           in NP yield due to potential blockages. Once the system   previous publications, and a homemade two-inlet nozzle
           started running with the peptide solution loaded, the   was used for extrusion [29,30]   . Structures were printed
           junctions  were  closely  watched  using  the  optical   directly onto 18  mm × 18  mm glass coverslips from
           microscope to ensure that no blockages or disruptions to   Thermo Fischer to facilitate imaging later. Two syringe
           the flow occur. Produced NPs were suspended in ethanol   pumps were loaded for extrusion and the samples were
           solution  which was collected in a 15  mL polystyrene   printed into a grid construct made up of two layers using
           conical falcon tube. The NPs in solution were then frozen   gcode.
           with liquid nitrogen and lyophilized in preparation for the   The  first  syringe  pump  was  loaded  with  the  peptide
           printing process.                                   solution and set to a flow rate of 55 µL/min. The second
                                                               pump was loaded with ×5 PBS and set to a flow rate of
           2.2.2 NP Characterization with Dynamic Light        20 µL/min. Three samples were printed for each condition
           Scattering (DLS)                                    (whether CH-01 or CH-02 and printed with or without
           The NP samples were also characterized using DLS on a   NPs) with a height of two to three layers for each sample
           Zetasizer (Model X) to determine the average size. This   for easier imaging. The same procedure was conducted
           was done during the optimization process to decide which   for both peptides.

                         A                                    B
















                                           C














           Figure 2. Peptide nanoparticles (NPs) preparation. Schematic representation of flow-focusing chip junctions (A), the diameter of the stream
           is started from 4.5 μm to higher than 7.5 μm (B), and image of the setup of the microfluidic platform for peptide NP fabrication (C).

                                       International Journal of Bioprinting (2019)–Volume 5, Issue 2       111
   110   111   112   113   114   115   116   117   118   119   120