A critical evaluation of drug delivery from ligand modified nanoparticles: Confounding small molecule distribution and efficacy in the central nervous system.

In this work, we sought to test how surface modification of poly(lactic-co-glycolic acid) (PLGA) nanoparticles with peptide ligand alters the brain specific delivery of encapsulated molecules. For biodistribution studies, nanoparticles modified with rabies virus glycoprotein (RVG29) were loaded with small molecule drug surrogates and administered to healthy mice by lateral tail vein injection. Mice were perfused 2h after injection and major anatomical regions of the CNS were dissected (striatum, midbrain, cerebellum, hippocampus, cortex, olfactory bulb, brainstem, and cervical, thoracic, lumbar and sacral spinal cord). For functional studies, surface modified nanoparticles were loaded with the chemotherapeutic camptothecin (CPT) and administered to mice bearing intracranial GL261-Luc2 gliomas. Outcome measures included tumor growth, as measured by bioluminescent imaging, and median survival time. We observed that small molecule delivery from PLGA nanoparticles varied by as much as 150% for different tissue regions within the CNS. These differences were directly correlated to regional differences in cerebral blood volume. Although the presence of RVG29 enhanced apparent brain delivery for multiple small molecule payloads, we observed minimal evidence for targeting to muscle or spinal cord, which are the known sites for rabies virus entry into the CNS, and enhancements in brain delivery were not prolonged due to an apparent aqueous instability of the RVG29 ligand. Furthermore, we have identified concerning differences in apparent delivery kinetics as measured by different payloads: nanoparticle encapsulated DiR was observed to accumulate in the brain, whereas encapsulated Nile red was rapidly cleared. Although systemically administered CPT loaded nanoparticles slowed the growth of orthotopic brain tumors to prolong survival, the presence of RVG29 did not enhance therapeutic efficacy compared to control nanoparticles. These data are consistent with a model of delivery of hydrophobic small molecules to the brain that does not rely on internalization of polymer nanoparticles in target tissue. We discuss an important risk for discordance between biodistribution, as typically measured by drug surrogate, and therapeutic outcome, as determined by clinically relevant measurement of drug function in a disease model. These results pose critical considerations for the methods used to design and evaluate targeted drug delivery systems in vivo.