(T1530-10-65) Artificial Extracellular Matrices Containing Human Collagen for In Vitro Analysis of Formulations Intended for Subcutaneous Injection

Purpose: No preclinical animal model has been identified that can predict the pharmacokinetic properties of biopharmaceutical formulations intended for subcutaneous (SC) injection in man.¹ To fill this gap, in vitro tools have been developed to investigate distinct events that can affect the uptake properties of a material from the SC injection site to reach systemic circulation, an overall outcome that defines its bioavailability. Typically, these methods of analysis employ a membrane, hydrogel, and/or cellular components to mimic the in vivo environment, chemically discriminating between the diffusion of a formulation’s components and/or between formulations.² We have developed an in vitro method to monitor the release of injectable pharmaceutical following injection into a dialysis membrane encased glycosaminoglycan-based hydrogel.^1,4 Herein, we look to probe the influence collagen (type I and III @ 0, 0.05, and 0.1 mg/mL) has on the material and chemical characteristics that dictate the in vitro bioavailability responses of previously described artificial extracellular matrices. Here, we examine the influence collagen (type I and III) has on the viscoelastic properties of the hydrogel format and compare collagen’s impact on drug release properties in this in vitro model of the SC injection site

Methods: Material characteristics of artificial extracellular matrix - The rheological characteristics of the novel and previously described artificial extracellular matrices (ECMs)^1,3 (Table 1) were analyzed using an Anton-Paar® MCR102e rheometer. Samples were deposited between 25mm parallel plates. frequency scans were conducted from 100-0.01 s^-1 at 1% strain and time-dependent responses were taken at 1 s^-1 at 1% strain for 2 hours in a hydrating environment. All samples were analyzed in an environment where 2-dimensional diffusion off of the sample stage was allowed. SCISSOR assay of therapeutics - Multiple formulations including caffeine, rapid & basal insulins, denosumab, and bevacizumab were analyzed using the SCISSOR® system (Pion Inc.).^1,3,4 Concentrations of injectates in the receiving chambers were monitored in real-time using in situ fiber optic dip probes connected to the Rainbow® UV-Vis spectrometer (Pion Inc.) or offline analysis was carried out with an Agilent A1100 HPLC after sampling.

Results: Material characteristics – Rheometric analysis of fresh ECMs showed no significant difference between the ECM and 0.05 mg/mL ECM-ColF (ECM with collagen from human fibroblast, mix of type I and III) complex viscosity, while 0.1 mg/mL ECM-ColF showed a more viscous (p < 0.05), and more stable behavior. After ≥100 hour assays ECM showed a ~99% loss in complex viscosity, from 0.74 ± 0.01 to 0.004 ± 0.001 Pa·s. The addition of 0.05 and 0.1 mg/mL collagen resulted in a viscosity loss of 0.72 ± 0.01 to 0.013 ± 0.004 Pa·s, and 0.76 ± 0.01 to 0.035 ± 0.04 Pa·s, respectively. Optical Density - Spectroscopic analysis showed that ECM with 0.1 mg/mL collagen type I had an optical density of 0.014 ± 0.016 absorbance units (AU), whereas ECM with 0.1 mg/mL collagen type II was 0.006 ± 0.010 AU, compared with a baseline of -0.002 ± 0.002 AU for ECM. Caffeine injections – Caffeine release profiles were collected using all variations of artificial ECMs (N=3). The caffeine release from all models indicated complete release at 24 hours. Insulin injections – Rapid and basal insulin (50-200uL, N=2-3) were injected and monitored over 4 days. 0.05 mg/ml ECM-ColF allowed rapid insulin to reach a maximum of 73 ± 2% release within 12 hours, while basal insulin plateaued at 67 ± 9% release over the duration of the experiment. The ECM had complete release of rapid insulin within 20 hours. Denosumab injections – Release of denosumab formulations (50-200 uL, N=3) was monitored over 1 week. ECM allowed 100% release within 72 ± 6 hrs, whereas the 0.05 mg/mL ECM-ColF resulted in a plateau at 62 ± 3%, while the ECM-ColF with 0.1 mg/mL collagen plateaued at 58 ± 5%. (Figure 1)

Conclusion: Artificial ECMs containing collagen types I and III and concentrations (0, 0.05, 0.1 mg/mL) were evaluated using five model injectables within Pion’s subcutaneous injection site simulator (SCISSOR). The increased retention of viscosity over time indicates that collagen complexes with the artificial ECM components in an attractive manner, resulting in increased stability without an initial increase in viscosity. Caffeine injections into each model showed analogous behavior over short time scales. However, the collagen-containing ECMs exhibited a more complex release profile, with multiple points of inflection and increased lag time. Commercially available formulations of insulin analogs and denosumab were injected to elucidate peptide and monoclonal antibody release behavior with and without collagen’s inclusion within the artificial ECM. In conclusion, collagen not only plays a role in supporting the architecture of these artificial ECMs, but it can directly affect the release profiles of all classes of injectables without significant deviation to the initial material characteristics.

References: 1. Kinnunen, H. M., Sharma, V., Contreras-Rojas, L. R., Yu, Y., Alleman, C., Sreedhara, A., . . . Mrsny, R. J. (2015). A novel in vitro method to model the fate of subcutaneously administered biopharmaceuticals and associated formulation components. Journal of Controlled Release, 94-102.
2. Bender, Christian, et al. Evaluation of in Vitro Tools to Predict the in Vivo Absorption of Biopharmaceuticals Following Subcutaneous Administration. Journal of Pharmaceutical Sciences, vol. 111, no. 9, 2022, pp. 2514–2524,
3. Bown, H. K., Bonn, C., Yohe, S., Yadav, D. B., Patapoff, T. W., Daugherty, A., & Mrsny, R. J. (2018). In vitro model for predicting bioavailability of subcutaneously injected monoclonal antibodies. J Control Release, 13-20.
4. “Scissor N3 - Biorelevant System for SC Formulations.” Pion, pion-inc.com/scientific-instruments/in-vivo-predictive-tools/subcutaneous/scissor.

Acknowledgements: Pion Inc. would like to thank Bath University for their continued collaboration on these projects.

Relative denosumab release profiles from ECM (green), ECM-ColF @ 0.05 mg/mL (orange) and 0.1 mg/mL (yellow) collagen

Table 1. Time-dependent responses of ECMs with 0 mg/mL collagen, 0.1 mg/mL collagen from human fibroblasts (ECM-ColF – 0.1), and 0.05 mg/mL collagen from human fibroblasts (ECM-ColF – 0.05), taken at 1 s-1 at 1% strain for 2 hours in a hydrating environment before assay