(T0930-08-52) Biorelevant Imaging and Formulation-Specific Release Prediction for the Rapid Assessment of Long-Acting Injectables

Purpose: Globally, nearly 40 million people live with human immunodeficiency virus (HIV), with over 30,000 new infections occurring in the U.S. every year. Long acting injectables (LAIs) are of particular interest for HIV treatment or pre-exposure prophylaxis because they provide consistent drug delivery without routine treatment adherence. A key challenge in the development of LAIs is assessing variability among depots and developing effective characterization methods that can correlate to downstream clinical outcomes without lengthy testing. In this study, micro computed tomography (µCT) is used to non-invasively investigate the structure of LAI depots injected into subcutaneous tissue. Imaging within excised tissue allows for analysis that is maximally biorelevant while also ensuring the ability to obtain maximal image quality. Through quantification of the depot morphology, in silico dissolution profiles are generated allowing for comparison between different injection modalities without the need for pharmacokinetic studies.

Methods: Six LAI depots were prepared by injecting different volumes of the LAI formulation into porcine tissue via an on-body injector (OBI) (4) or manual injection (2). The depots were imaged via µCT at 120 µm/voxel resolution. 3D tissue volume images were segmented using artificial intelligence-powered image analysis software. Segmentations were used to calculate depot localization, depot volume, surface area, as well as the size distribution and spatial distribution of internal depot porosity. Phase densities were measured using calibration phantoms co-imaged with the ex vivo tissues. Intrinsic dissolution simulations were generated from the individual 3D depot structures and the drug solubility in subcutaneous tissue.

Results: Figure 1 shows 3D renderings from the µCT images. Depots were segmented into the main depot phase, intradepot phase, and pores, as well as the surrounding skin, subcutaneous adipose, and muscle tissue. All the depots appeared ellipsoidal with soft surface ridges. The depots generated from the OBI were 100% within subcutaneous tissue, while the manual injection depots spanned both the subcutaneous and intramuscular space. The portions of depots within the intramuscular tissue had flat, elongated shapes. Manual injection depots had higher surface area-to-volume ratios than OBI depots. This suggests that depot localization is better controlled with OBIs than manual injections. This may be due to several factors, including slower injection rate, reduced needle length and size, and reduced human error. Pore and intradepot phase volume fractions are shown in Table 1 and were consistent between depot type and volume replicates. The 3 mL OBI depots had the highest pore volume fraction, followed by 4.75 mL OBI depots, and finally the manual injection depots. Intradepot phase volume fraction was highest for 4.75 mL OBI depots, followed by 3 mL OBI depots, with manual injection depots having the lowest volume fraction. Additionally, while manual injection depots had significantly smaller pores than OBI depots, intradepot size distributions were similar in each depot and did not have a clear trend by volume or injection type. Intrinsic dissolution simulations were performed based on the depot morphology and drug solubility in subcutaneous tissue and are shown in Figure 2. These simulations predicted all depots to have greater than 90% release at t=180 days. This indicates that the depots have an appropriate shape and size for 6-month targeted release. Depot volume did not produce a clear trend in release profiles. However, depots with higher surface area-to-volume ratios had faster predicted release. In particular, the 1.5 mL manual injection had a predicted release of 100% at t=106 days. This depot had significant differences in tissue localization, morphology, and had a 2-3 times higher surface area-to-volume ratio compared to the other depots. This indicates that depot dissolution is majorly influenced by depot surface area-to-volume ratio.

Conclusion: High-resolution µCT imaging studies revealed critical details of depot morphology and tissue localization for ex vivo LAI depots generated via OBI or manual injection. The depots were localized either fully in the subcutaneous tissue for OBI depots or across subcutaneous and intramuscular tissue for the manual injection depots. Each depot contained pores and two solid phases, which had similar respective spatial distributions in each sample. Intrinsic dissolution simulations of the depots revealed all OBI depots had similar predicted release at t=180 days, but the manual injection depots with higher surface-to-area volume ratio had significantly faster predicted release. This study supports the use of OBIs as a tool to control drug depot morphology, tissue localization, and release profile, which is critical for developing effective long-acting treatments for HIV.

Table 1. Calculated volume, surface area, and surface area-to-volume ratio for each depot, including individual phases and overall depot.

Figure 1. 3D renderings of segmented µCT imaging data for each depot.

Figure 2. Predicted release curves for each depot.