(T1430-06-38) Content Uniformity, Drug Quantification and Cross-Reactivity Analysis of Protein Loaded PGSU Implants Using Raman Spectroscopy and Reverse-Phase High Performance Liquid Chromatography (RP-HPLC)

Purpose: Content uniformity and accurate quantification of drug loading (DL) are important parameters in evaluating the performance of medical devices. However, content uniformity and accurate quantification of biological macromolecules such as peptides and proteins in solid polymeric implants remains a major challenge. Our group has developed a flexible elastomer made of poly(glycerol sebacate) urethane (PGSU), for next generation delivery of active pharmaceutical ingredients that allows a steady release of APIs for several months to years. Quantification of small molecule drugs in PGSU has been successfully performed by swelling the polymer in dimethylsulfoxide (DMSO) which increases the polymer network mesh size and results in total extraction of the small molecules within 4 to 24 hours. However, extracting biomolecules such as proteins using DMSO remains challenging, mainly because the swelling of PGSU is still not sufficiently large enough to affect the mesh size, resulting in incomplete recovery of the protein. Therefore, the goal is to develop complementary techniques – one destructive and one non-destructive – that can provide information about the uniformity of DL and also accurately quantify the amount of drug present within these polymeric implants. In order to address these challenges, Raman Spectroscopy was utilized to study the content uniformity and distribution of proteins on the polymeric surface, using intact polymer implants, and>Amino acid content analysis was utilized for quantification of DL and also to provide information regarding any potential cross-reactivity between the protein and crosslinked polymer network, using digested polymer implants.

Methods: Protein-loaded PGSU cylindrical implants (3 mm diameter, 40 mm length) were synthesized from poly(glycerol sebacate) (PGS) by crosslinking it with hexamethylene diisocyanate (HDI), an isocyanate crosslinker, in the presence of a tin catalyst, to form a polyester urethane (PGSU). Bovine serum albumin (BSA) was used as a model protein and four different formulations with drug loadings of 20, 30, 40 and 50% (w/w) were manufactured which were 40 mm in length and 3 mm in diameter. To identify spatial content uniformity of the samples, an inVia™ confocal Raman microscope equipped with a 785 nm laser was utilized to collect spectra of the samples at various points along the surface. A heat map was then generated by quantifying area under the curve (AUC) for the phenylalanine peak at 1004 cm^-1 that is distinct for proteins to obtain information regarding the distribution. Amino acid content analysis was performed using acid hydrolysis and the samples were run on RP-HPLC for drug quantification as well as any cross-reactivity. Digested samples were run using reverse phase and regular phase LC methods for assay quantification and cross-reactivity analysis.

Results: Four different formulations with varying amounts of BSA (20, 30, 40 and 50%) were synthesized using the extrusion process. Figure 1 shows the surface scan of a 40% loaded BSA-PGSU rod using confocal Raman spectroscopy. The Raman microscope objective was selected at 20x magnification and the spot size of the laser was approximately 2.5 µm, and the depth of penetration was approximately 1 µm. A white light image was generated across a length of ~ 8 mm and width of 1 mm, and surface scans by Raman laser (785 nm) were collected with a step size of 20 µm in both x and y directions. Laser power and exposure time for all scans were maintained at 100% and 5 secs respectively. Post processing involved cosmic ray removal, and baseline correction followed by peak quantification at 1004 cm^-1, which is a distinct signal emanating from the phenylalanine residue present in proteins. Figure 1A shows a fairly uniform distribution of BSA across the surface of the implant. The BSA particle size was approximately 20 um, which is larger than the laser spot size, and accordingly the Raman scan shows the boundaries of individual drug particles within the polymer matrix. Any deformities on the surface of the implant such as cracks or air bubbles showed very low protein signal which is expected because of lack of focus. Figure 1B shows the Raman spectrum of a point on the image and the peak that was used for quantification. Figure 2 is a box plot that represents the AUCs of all points that were collected for 20, 30 and 40% loaded BSA rods. While no statistical difference was observed with respect to drug loading on the surface, a linear trend is nevertheless observed between the different formulations. We are currently working on developing methods for accurate quantification of protein loading as well as cross-reactivity in PGSU implants using amino acid content analysis.

Conclusion: Drug uniformity and accurate loading quantification are crucial in solid dosage forms. Our group is currently working on developing techniques that can be used to study uniformity as well as accurately quantify loading of macromolecules in PGSU implants. These techniques can be useful as Quality Control (QC) and Process Analytical Technology (PAT) lot release test methods.

Figure 1. A) Heat map showing the distribution of BSA on the surface of the plant. Corresponding white light image is also shown. B) Raman spectrum of a point on the surface showing the peak that was used for quantification.

Figure 2. Box plot showing the distribution of AUCs of all points collected on the surface for three different formulations. Mean and median values for the different loadings are shown.