Zebrafish model of blood-brain barrier to improve drug delivery to the brain

Division Of Basic Sciences - Nci

Investigators

Linked publications & trials

Abstract

ATP-binding cassette (ABC) transporters are a major component of the blood-brain barrier (BBB) and blood-placental barrier. ABCB1 and ABCG2 are highly expressed at the BBB, where they limit brain penetration of many chemotherapies. We previously developed a murine model for analysis of ABCG2 expression at the blood-brain barrier based on the fact that luciferin is an ABCG2 substrate and its entry into the brain is prevented by transporter expression. In this model, firefly luciferase is under the expression of the GFAP promoter, leading to its expression in the astrocytes. When mice are injected with luciferin, no light signal from the brain is detected due to ABCG2 preventing luciferin from crossing the blood-brain barrier. However, when luciferin is co-administered with an ABCG2 inhibitor, it can cross the blood-brain barrier and react with luciferase expressed in the astrocytes to produce light which can be quantitatively measured. Because studies of the BBB in mice are time-consuming and expensive, we have developed homologous models in the zebrafish, since the zebrafish BBB is very similar to the mammalian BBB. We have developed and characterized a transgenic zebrafish line with NanoLuc, derived by Promega from a deep-sea shrimp, under the control of the GFAP promoter. In this model, NanoLuc is expressed in the developing and adult zebrafish brain and spinal cord. Coelenterazine is one of the substrates for NanoLuc and is transported by both ABCB1 and ABCG2. Furimazine, a coelenterazine derivative with very high yield of light, is an ABCG2 substrate. Thus, this model can be used to study the role of transporters at the blood-brain barrier, but could also be used to screen compounds that might increase permeability of the barrier regardless of the mechanism. We have shown that addition of a substrate, such as furimazine, and an Abcg2 inhibitor added either to the water or injected into the circulation of larval zebrafish result in light generation consistent with penetration of furimazine across the BBB. We have characterized the zebrafish homologs of the human ABCB1 and ABCG2 transporters. We found that zebrafish Abcb4 is nearly identical to human ABCB1 in conferring resistance to 90 known ABCB1 substrates. Zebrafish also have 4 homologs of human ABCG2-Abcg2a, Abcg2b, Abcg2c and Abcg2d. We have functionally categorized these zebrafish ABCG2 orthologs and determined the brain tissue expression. Using RNAscope in situ hybridization and immunolocalization with an anti-zebrafish Abcg2a antibody, we identified Abcg2a as the only ABCG2 homolog expressed at the adult and larval zebrafish BBB, based on its localization to claudin-5 positive brain vasculature. We also found reactivity in the liver and gastrointestinal tract similar to the ABCG2 expression pattern in humans. These results demonstrate the conserved function of zebrafish Abcg2a and suggest that zebrafish may be an appropriate model organism for studying the role of ABCG2 at the BBB. Having identified zebrafish Abcb4 and Abcg2a as the homologous transporters at the zebrafish BBB, we characterized the ability of NanoLuc substrates to be transported by zebrafish and human transporters at the BBB. We examined several coelenterazine derivatives as well as some furimazine derivatives and found that furimazine was the brightest NanoLuc substrate tested and was transported by human ABCG2 and zebrafish Abcg2a. Coelenterazine h was the brightest coelenterazine derivative and was also transported by human ABCG2 and zebrafish Abcg2a. Thus, these compounds could be used to study the role of ABCG2 at the BBB. As context to our own inhibitor studies, and to enable characterization of the BBB in the complete absence of Abcb4 and Abcg2a, we have generated knockout zebrafish models for these genes which are viable and currently under study. In collaboration with Di Xia, the structure of zebrafish Abcb4 was determined by cryoEM and shown to be virtually identical to that of human ABCB1. We have also generated knockout mutants of Abcb4 and Abcg2a and double knock-outs in zebrafish which are viable and fertile and currently being characterized. One approach to treatment of both primary and metastatic brain tumors involves the use of photodynamic therapy which can cross the BBB. In collaboration with Joe Huang at the University of Maryland, we have created an amorphous verteportin nanodrug which also inhibits P-glycoprotein at the BBB, and it is under study as a possible adjunct to treatment of brain tumors. In collaboration with Jing Wu at CCR, NCI, we have explored whether a novel drug zotiraciclib which appears to cross the BBB is a substrate for P-gp and ABCG2 and found that it is not, explaining its ability to enter the brain for treatment of primary IDH-mutant gliomas. Similarly, in collaboration with Sadhana Jackson, at NINDS, we have shown that ibrutinib is a P-gp inhibitor and able to enter the brain for treatment of primary brain tumors.