The inability to efficiently deliver small interfering RNAs to target organs hinders their therapeutic application. So a demonstration of siRNA delivery to a notoriously difficult organ — the brain — is very exciting indeed.
The process of RNA interference (RNAi) uses small RNA triggers to decrease gene expression in a sequence-specific manner. In most applications, a small interfering RNA (siRNA) sequence — 21–23 nucleotides long — silences a target messenger RNA with a complementary sequence by either directing site-specific cleavage or inhibiting its translation into a protein1,2. This technique is a powerful and versatile tool for inhibiting gene expression, and has elicited a great deal of interest in harnessing it to treat human disease. Although during the past two years several significant findings have enabled systemic delivery of siRNAs to various tissues and organs1, one tissue that has been difficult to target from the periphery is the brain; this is because traversing the blood–brain barrier poses a great challenge. On page 39 of this issue, Kumar et al.3 describe a method to meet this challenge.
The junctions between the endothelial cells that line the brain capillaries prevent the passage of most molecules4. Chemically modified liposomes conjugated to monoclonal antibodies raised against epidermal growth factor can penetrate the mouse brain, as can neuropeptides, but these approaches do not directly target neurons4.
Similarly, viral vectors can be designed to express short hairpin-structured RNA sequences, which, once expressed, are processed to siRNAs. Such vectors have been used to introduce siRNAs directly into the brain tissue using the neurosurgical method of stereotaxis; thus, it has been shown that a decrease in the transcription of certain genes alleviates symptoms of neuronal diseases5,6. But direct injection within the cranium (the protective upper portion of the skull) is invasive, and systemic delivery of brain-targeting agents is a highly desirable alternative.
Kumar et al.3 exploited the fact that neurotropic viruses — such as the rabies virus — that preferentially infect the nervous system can penetrate the brain. The rabies virus achieves this through glycoproteins on its lipid envelope. To test whether rabies viral glycoprotein (RVG) could promote transport of a non-rabies viral vector to the brain, the authors used the technique of pseudotyping, whereby an envelope glycoprotein from a specific virus is used to confer tropism for specific cell types to a viral vector. They showed that RVG-expressing lentivirus vectors could specifically infect neurons in culture, but not HeLa cells.
They then showed that a RVG-pseudotyped lentiviral vector, encoding a short hairpin RNA designed to target Japanese encephalitis virus (JEV), provided complete protection against this virus when both the vector and JEV were intracranially injected into the brains of mice. By contrast, the same lentiviral vector pseudotyped with the vesicular stomatitis G glycoprotein (VSV G) envelope did not afford protection, even though — as shown in previous work7,8 — it reached neuronal cells. This is because retrograde spread of RVG occurs in axonal processes, facilitating its transynaptic spread to neighbouring neurons, whereas this doesn't happen for VSV G.
To transfer siRNAs into neural cells, Kumar et al. identified a 29-residue peptide from the RVG envelope that selectively binds to the acetylcholine receptor. They then fused this peptide with a sequence of nine arginine (Arg) residues that binds to negatively charged nucleic acids, including siRNAs8. They found that, when it was injected into mice, the peptide-conjugated siRNA (RVG–Arg–siRNA) specifically reached neurons (Fig. 1). This was detected through its effect on a synthetic protein known as enhanced green fluorescent protein (EGFP) that had previously been introduced into the mice.
The authors then questioned whether RVG–Arg–siRNA could be used to deliver siRNAs to brain cells in vivo after intravenous injection. They found that injection of mice expressing EGFP with RVG–Arg–siRNA complexes induced selective silencing of protein expression only in brain tissue. Injection of a control complex comprising the siRNA bound to the arginine sequence and the rabies virus matrix peptide (RV-MAT–Arg), however, did not affect EGFP expression. Similar results were obtained with an siRNA directed against the mRNA of the protein SOD1 — a natural target that is expressed in several cell types of the brain.
Finally, the authors used three intravenous injections of RVG–Arg–siRNA on successive days to protect mice from a JEV infection. Untreated mice, mice treated with the RV-MAT–Arg–siRNA complex, or those treated with a RVG–Arg–siRNA complex, in which the siRNA was a nonspecific sequence, all died within 10 days of JEV infection. However, 80% of mice treated with a siRNA complex specifically directed against JEV survived. By assaying for the levels of type I interferon, the authors ruled out the involvement of immune mediators, such as interferons, in protection against JEV infection conferred by RVG–Arg–siRNA complexes.
The widespread distribution of acetylcholine receptors in the thymus, the lungs and lymphocytes9 doesn't seem to affect either the bioavailability or the specific delivery of the RVG peptide to the central nervous system. But it is unclear which brain cells the RVG peptide reaches, because microglia, astrocytes, endothelial cells and neurons all express acetylcholine receptors9. Whether the RVG peptide traverses the blood–brain barrier to enter cells in the functional regions of the brain also remains unresolved. And which cells the siRNAs ultimately reside in is not known.
The exact mechanism by which the RVG peptide traverses the blood–brain barrier and enters cells is yet to be elucidated. However, the capacity of the gp120 protein of HIV to bind the acetylcholine receptor, and the ability of α-bungarotoxin and HIV Tat proteins to cross the barrier8, suggest that peptides derived from other viruses that infect the brain — including HIV, the herpes simplex virus and flaviviruses — should also be explored as siRNA delivery agents.
Given the tremendous potential of RNAi as a therapy, it is tempting to speculate that the findings of Kumar and colleagues3 could open the door for the treatment of neurological diseases. More immediately, however, it is hoped that the application of this approach to mice and large animal models of human diseases will demonstrate its usage, and indicate its feasibility, for human applications.
Kim, D. H. & Rossi, J. J. Nature Rev. Genet. 8, 173–184 (2007).
Martin, S. E. & Caplen, N. J. Annu. Rev. Genomics Hum. Genet. doi: 10.1146/annurev.genom.8.080706.092424 (2007).
Kumar, P. et al. Nature 448, 39–43 (2007).
Pardridge, W. M. Adv. Drug Deliv. Rev. 59, 141–152 (2007).
Davidson, B. L. & Boudreau, R. L. Neuron 53, 781–788 (2007).
Xia, H. et al. Nature Med. 10, 816–820 (2004).
Wong, L.-F. et al. Hum. Gene Ther. 17, 1–9 (2006).
Wender, P. A. et al. Proc. Natl Acad. Sci. USA 97, 13003–13008 (2000).
Gotti, C. & Clementi, F. Prog. Neurobiol. 74, 363–396 (2004).
The authors declare no competing financial interests.
About this article
Molecular Pharmaceutics (2017)
Experimental Neurology (2012)
A brain-targeted rabies virus glycoprotein-disulfide linked PEI nanocarrier for delivery of neurogenic microRNA