Cerebrospinal fluid drainage to the lymphatic system in rats and nonhuman primates

Objectives: The availability of quantitative mechanistic data on cerebrospinal fluid (CSF) drainage from the subarachnoid space is important for understanding pathological processes involving the CNS as well as for development of efficient and safe methods of delivering drugs to the CNS and meningeal targets. Multiple controversies regarding the mechanisms of CSF turnover and its role in the transport of endogenous and exogenous solutes still exist despite fifteen decades of research. It is generally (but not unanimously) agreed that CSF is predominantly produced by the choroid plexus. It has long been known that lymphatic vessels in the dura collect dye particles injected into the CSF (G. Schwalbe, 1869). H. Quincke (1872) proposed that CSF drainage occurs at the nerve roots while L.H. Weed (1914) concluded that CSF predominantly drains through arachnoid villi to the blood and that lymphatics are only a minor mechanism for CSF reabsorption. The role of dural lymphatics in antigen sampling from the CSF, which does not require uptake of a large fraction of the CSF, was studied by R. Weller’s group (R. Carare et al, 2013). However, “integration” of subarachnoid space and lymphatics was proposed by L. Koh et al in 2005, and it was proposed by A. Aspelund et al (2015) that meningeal lymphatics play a major role in draining macromolecules from the brain. The goal of the present study was to estimate the lymphatic uptake of macromolecules from the CSF in rats and nonhuman primates using quantitative PET imaging.

Methods: Full body imaging PET data obtained over multiple studies on the pharmacokinetics of macromolecules administered to the lumbar and cerebro-cervical CSF of rats and monkeys were reprocessed to determine the fraction of dose that accumulated in lymph nodes. Five human recombinant proteins (prospective enzyme replacement therapeutics), phage M13 particles, two different virions and model nanoparticles were labeled with either ¹²⁴I or ⁸⁹Zr. The macromolecules and particles were administered either through direct injection to cisterna magna (rats) or through pre-implanted intrathecal injection ports (rats, monkeys). Dynamic data acquisition was carried out for 30 minutes, followed by static acquisition of whole body images for at least 48 hours. Images were reconstructed using Siemens OSEM3D protocol with correction for attenuation and scattering, and analyzed to identify lymph nodes accumulating the radioactive labels and, where identified, to determine the fraction of the administered dose in the nodes.

Results: No lymphatic uptake attributable to CSF drainage at nerve roots was detected in either of the two species provided that there was no damage to the dura. Even minor perforations, e.g., caused by direct needle injection, showed significant CSF leakage into the respective sentinel lymph node(s). Therefore, only animals with pre-implanted injection ports outside of the study regions were used to estimate the lymphatic uptake of the solutes. In animals with undamaged dura, the lymphatic uptake from the CSF was found to be low (0.32%ID±0.28%ID) in monkeys in all regions of the body. Lymphatic uptake of macromolecules was more significant in rats, but only in the anterior cervical area (2.4%ID±1.8%ID in deep lymph cervical nodes) with no accumulating lymph nodes in any other areas. The kinetics of label uptake in the liver and other organs were in agreement with direct transfer from the CSF to the blood and did not show delays that would have been caused by prior lymphatic drainage.

Conclusion: The data suggests that direct drainage to the blood is the primary pathway by which macromolecules are cleared from the CSF in animals with undamaged dura. The higher label uptake in the deep cervical nodes of rats may be related to lymphatic drainage from either the olfactory epithelium to which the CSF reportedly drains into in rodents, or from the area of the foramen magnum. Research Support: This study was supported by NIH grants R21NS090049, R01NS092838 and R21CA191979