Evaluating the effect of injection protocols on intrathecal solute dispersion in non-human primates: an in vitro study using a cynomolgus cerebrospinal fluid system.

Abstract

Background: Achieving effective drug delivery to the central nervous system (CNS) remains a challenge for treating neurological disorders. Intrathecal (IT) delivery, which involves direct injection into the cerebrospinal fluid (CSF), presents a promising strategy. Large animal studies are important to assess the safety and efficacy of most drugs and treatments and translate the data to humans. An understanding of the influence of IT injection parameters on solute distribution within the CNS is essential to optimize preclinical research, which would potentially help design human clinical studies.

Methods: A three-dimensional (3D) in vitro model of a cynomolgus monkey, based on MRI data, was developed to evaluate the impact of lumbar injection parameters on intrathecal solute dispersion. The parameters evaluated were (a) injection location, (b) bolus volume, (c) flush volume, (d) bolus rate, and (e) flush rate. To simulate the CSF flow within the subarachnoid space (SAS), an idealized CSF flow waveform with both cardiac and respiratory-induced components was input into the model. A solution of fluorescein drug surrogate tracer was administered in the lumbar region of the 3D in vitro model filled with deionized water. After injection of the tracer, the CSF system wide-solute dispersion was imaged using high-resolution cameras every thirty seconds for a duration of three hours. To ensure repeatability each injection protocol was repeated three times. For each protocol, the average spatial-temporal distribution over three hours post-injection, the area under the curve (AUC), and the percent injected dose (%ID) to extra-axial CSF (eaCSF) at three hours were determined.

Results: The changes to the lumbar injection parameters led to variations in solute distribution along the neuro-axis. Specifically, injection location showed the most impact, enhancing the delivery to the eaCSF up to + 10.5%ID (p = 0.0282) at three hours post-injection. Adding a post-injection flush of 1.5 ml at 1 ml/min increased the solute delivery to the eaCSF by + 6.5%ID (p = 0.0218), while the larger bolus volume resulted in a + 2.3%ID (p = 0.1910) increase. The bolus and flush rates analyzed had minimal, statistically non-significant effects.

Conclusion: These results predict the effects of lumbar injection parameters on solute distribution in the intrathecal space in NHPs. Specifically, the choice of injection location, flush, and bolus volume significantly improved solute delivery to eaCSF. The in vitro NHP CSF model and results offer a system to help predict and optimize IT delivery protocols for pre-clinical NHP studies.