(T1230-11-60) Characterizing the Effect of Critical Processing Parameters on Critical Quality Attributes of Protein-Loaded Double Emulsion Microparticles

Purpose: Over the past several decades, controlled drug delivery strategies have emerged as a viable solution to address common limitations of standard pharmaceuticals. Our group has demonstrated the efficacy of these systems, specifically poly(lactic-co-glycolic) acid (PLGA) microparticles, in many models of inflammation (1-6). Despite being widely researched in an academic setting, there are limited FDA-approved PLGA microparticle-based delivery systems on the market (7). One potential barrier to translation of these technologies is the scaling of academic lab-based manufacturing protocols to industry-based protocols, which often requires changes in a manufacturing process. Deviations in a manufacturing process, especially pertaining to critical processing parameters (CPPs) or critical material attributes (CMAs), can greatly impact product characteristics, otherwise known as critical quality attributes (CQAs) (8, 9). By understanding the relationship between these variables, researchers can potentially improve translatability of a technology (8). To this end, we sought to elucidate the relationship between CPPs and CQAs of protein-loaded double emulsion PLGA microparticles. More specifically, we explored the impact of CPPs in the primary emulsion process (e.g., inner aqueous phase volume, solvent volume, etc.) on microparticle CQAs, including but not limited to protein release kinetics and microparticle internal structure. Microparticle internal structure, such as inner occlusion size and interconnectivity, was explored through the novel application of a machine learning algorithm provided by our collaborators.

Methods: Prior work using a design of experiments (DOE) approach identified significant CPPs in double emulsion PLGA microparticle fabrication. Building upon this work, protein-loaded MPs, specifically MPs encapsulating recombinant human CCL22 (rhCCL22-MPs), a chemokine for regulatory T cells (Tregs), were fabricated to explore the impact of CPPs on microparticle CQAs. rhCCL22-MPs were fabricated using a well-documented double emulsion solvent evaporation method (1). To assess rhCCL22 release kinetics, MPs were suspended in phosphate buffer saline and incubated at 37°C with end-over-end rotation. Suspended MPs were sampled periodically following centrifugation, enabling supernatants to be extracted. rhCCL22 concentration in supernatants was then quantified using enzyme-linked immunosorbent assays (ELISA). Scanning electron microscopy (SEM, Zeiss) was used to image surface morphology and internal structure. With the captured SEM images, features of the internal structure (e.g., inner occlusion size, % polymer in cross section) were quantified through use of a random forest machine learning classifier through “digiM I2S”, a cloud pharmaceutical image processing and data management platform.

Results: In the first study, the single factor and interaction effects of inner aqueous phase volume and solvent volume were studied. In general, increasing inner aqueous phase volume from 200 uL to 500 or 800 uL was found to accelerate rhCCL22 release across all solvent volumes (Figure 1A). Visually, it appears that changing inner aqueous phase volume alters microparticle internal structure, including inner occlusion size and interconnectivity (Figure 1B). Through a machine learning algorithm, characteristics of the internal microparticle structure were able to be quantified (Figure 1C). Overall, increasing inner aqueous phase volume from 200 uL to 500 uL or 800 uL increases inner occlusion size (Figure 1D). Accordingly, the % polymer between the inner occlusions appears to decrease (Figure 1D). There are limited differences in inner occlusion size and % polymer for 500 uL or 800 uL inner aqueous phase volumes (Figure 1D). In the second study, single factor and interaction effects of polymer amount and solvent volume were studied. When 2 mL of solvent is used, increasing polymer amount does not significantly impact rhCCL22 release (Figure 2A). However, when 4 or 8 mL of solvent is used, increasing polymer amount significantly represses rhCCL22 release (Figure 2A). Microparticle internal structure also changes as these CPPs change (Figure 2B). In general, increasing polymer amount decreases inner occlusion size (Figure 2C). Furthermore, increasing polymer amount increases % polymer between inner occlusions of a cross section (Figure 2C).

Conclusion: Altogether, these results suggest that modulating CPPs in protein-loaded double emulsion microparticle fabrication significantly impact microparticle CQAs, including protein release kinetics and microparticle internal structure. It can be concluded that rhCCL22 release may be heavily influenced by microparticle internal structure, as reduced inner occlusion interconnectivity (or increased % polymer in a cross section) corresponded with accelerated rhCCL22 release. Overall, these results also suggest the importance of quantifying CQAs associated with double emulsion microparticle internal structure and highlight the novel application of a machine learning algorithm to do so.

References: vascularized composite allotransplantation. Sci Adv 6, eaax8429 (2020).
2. J. D. Fisher et al., Treg-inducing microparticles promote donor-specific tolerance in experimental vascularized composite allotransplantation. Proc Natl Acad Sci U S A 116, 25784-25789 (2019).
3. A. J. Glowacki et al., Prevention of inflammation-mediated bone loss in murine and canine periodontal disease via recruitment of regulatory lymphocytes. Proc Natl Acad Sci U S A 110, 18525-18530 (2013).
4. M. L. Ratay et al., Treg-recruiting microspheres prevent inflammation in a murine model of dry eye disease. J Control Release 258, 208-217 (2017).
5. S. C. Balmert et al., In vivo induction of regulatory T cells promotes allergen tolerance and suppresses allergic contact dermatitis. J Control Release 261, 223-233 (2017).
6. M. L. Ratay et al., TRI Microspheres prevent key signs of dry eye disease in a murine, inflammatory model. Sci Rep 7, 17527 (2017).
7. Y. Wang, B. Qin, G. Xia, S. H. Choi, FDA's Poly (Lactic-Co-Glycolic Acid) Research Program and Regulatory Outcomes. AAPS J 23, 92 (2021).
8. L. X. Yu et al., Understanding pharmaceutical quality by design. AAPS J 16, 771-783 (2014).
9. C. Zhang et al., Quality by design thinking in the development of long-acting injectable PLGA/PLA-based microspheres for peptide and protein drug delivery. Int J Pharm 585, 119441 (2020).

Acknowledgements: The first and corresponding authors would like to thank the Nanofabrication and Characterization Facility at the University of Pittsburgh for training on and access to scanning electron microscopy machines used in this study. Furthermore, we would like to thank the team at DigiM for access to software enabling microparticle characterization.


Figure 1. Interaction between inner aqueous phase volume and solvent volume changes release and internal structure of rhCCL22-MPs. (A) Cumulative release (ng rhCCL22/mg MP) profiles of rhCCL22-MPs formulated with each inner aqueous phase volume (200, 500, 800 uL) at each solvent volume (2, 4, 8 mL). (N = 3). (B) Scanning electron microscopy (SEM) images of microparticle cross sections. SEM images taken at 1.5kX. Scale bar = 10 m. (C) Diagram depicting features of microparticle internal structure and associated measurements. (D) Microparticle inner occlusion diameter (m) and polymer matrix composition (% polymer) measurements (n = 3 – 4).


Figure 2. Interaction between polymer amount and solvent volume impacts release kinetics and internal structure of rhCCL22-MPs. (A) Cumulative release (%) profiles of rhCCL22-MPs formulated with each polymer amount (200, 400, 600 mg) at each solvent volume (2, 4, 8 mL) (N = 3). (B) Scanning electron microscopy (SEM) images of microparticle cross sections. SEM images taken at 1.5kX. Scale bar = 10 m. (C) Microparticle inner occlusion diameter (m) and polymer matrix composition (% polymer) measurements (n = 3 – 4).