Summary Update: Testing the GABA Nucleo-Olivary Hypothesis of Essential Tremor
January 2016 Progress Report: Testing the GABA Nucleo-Olivary Hypothesis of Essential Tremor
Experiment I. Confirmation of AAV vectors functions in Gad2-Cre mice
We plan to test the hypothesis that loss of GABAergic nucleo-olivary (N-O) input into the inferior olive (IO) underlies at least a subset of Essential Tremor. To test this hypothesis, we use AAV vectors expressing genes to impair activity of GABAergic N-O neurons under cre-recombinase regulation, in mice expressing cre specifically in GABAergic neurons (Gad2-Cre mice).
We use three strategies:
(1) optogenetic inactivation of GABA N-O neurons, using AAV expressing the optically activated proton pump, Archaerhodopsin (Arch) to hyperpolarize neurons when stimulated by light delivered locally to these neurons,
(2) inactivation using AAV expressing a mutated human muscarinic receptor (hM4Di), a Designer Receptor Exclusively Activated by administration of a Designer Drug (DREADD), causing hyperpolarization of these neurons, or
(3) lesion of GABAergic N-O neurons using AAV expressing the active subunit A of diphtheria toxin (DtA).
Before those experiments can be performed, however, we first need to demonstrate that the AAV vectors express Arch, hM4Di, DtA, or control fluorescent reporter genes (GFP or mCherry) in GABAergic N-O neurons in Gad2-Cre mice. This is the aim of Experiment I of this IETF-funded study.
As described in our proposal, Gad2-Cre mice (n=3) and non-transgenic wild type (WT) mice (n=3) were injected bilaterally into IO with the following AAV vectors: AAV-Arch-GFP, AAV-GFP, AAV-hM4Di-mCherry, AAV-mCherry or AAV-DtA-mCherry and brain tissue was taken for histological assessment of fluorescent reporter expression 4 weeks post-injection.
Brain sections from Gad2-Cre mice containing IO or DCN showed some expression of fluorescent reporters (GFP for Arch and GFP vectors; mCherry for other vectors) in presumably GABAergic extra-olivary cells in the brainstem, but no reporter expression in fibers within the IO. Also, no expression in cell bodies of neurons in interposed or lateral nuclei of the DCN were observed in Gad2-Cre mice. As expected, no expression was observed in WT mice.
To test the possibility that insufficient time had elapsed to allow retrograde transport from terminals in IO for expression in GABA N-O neurons, an additional n=3 mice per group were injected intra-IO with AAV vectors and tissue was taken 8 weeks post-injection. Again, fluorescence of reporter genes was observed in some extra-olivary neurons within the ventral brainstem, but no fluorescence was observed in GABAergic N-O neurons in Gad2-Cre mice.
We hypothesized that inefficient uptake at GABAergic N-O axonal terminals within the IO, and/or poor retrograde transport of the AAV vector may prevent expression of these vectors. Therefore, we injected n=6 Gad2-Cre mice with AAV-GFP vector directly into the DCN at two sites, unilaterally (in interposed and dentate nuclei). Again, we saw no evidence of expression in DCN cell bodies or fibers in IO. GFP fluorescence was seen within the cerebellar cortex and in the dorsal brainstem, at the penumbra of the vector injection sites in DCN. In addition, fluorescence was observed in fibers ventral to the IO, possibly within GABAergic axons in the medial lemniscus.
The serotype 5 AAV vectors used in our study have been shown to transduce neurons in other GABAergic systems, and our vector methodology has been successful in multiple previous endeavors. It was thus unexpected that GABAergic N-O neurons do not take up these vectors. This appears to reflect an unusual and uncommon lack of tropism for GABAergic N-O neurons, preventing viral vector transduction and expression in these neurons. Fortunately, recent results from another group that emerged as our work was underway indicate that the problem may be solved by using another AAV vector serotype, AAV 9. In a study by Ankri et al. (eLife 2015;10.7554/eLife.06262), cre-regulated AAV serotype 9 vectors showed excellent activity in N-O neurons, with fluorescent reporter expression in Gad2-Cre mice GABAergic N-O neurons after injection into DCN. In addition, Ankri et al. (2015) used a non-cre-regulated AAV serotype 9 vector expressing GFP to achieve retrograde expression of GFP in GABAergic N-O neurons following IO injection.
Therefore, we plan to proceed promptly to the use of AAV serotype 9 vectors to express cre-regulated optogenetic (Arch), DREADD (hM4Di) and toxin (DtA) constructs and controls (GFP, mCherry). AAV vector serotype 9 vectors are available from the University of North Carolina Vector Core. After we have confirmed successful activity of these vectors in GABAergic N-O neurons, we will then proceed with the remainder of the study.
We are confident that once this technical hurdle of AAV vector targeting expression in GABAergic N-O neurons has been overcome, we will have the tools available to test the GABA N-O hypothesis of Essential Tremor in a novel mouse model platform that will provide new insights into the mechanistic basis of tremor generation in this disease.
In addition, we have already successfully overcome a number of other technical barriers, including the difficult stereotaxic surgery technique of reliably targeting the very small IO in mice, successfully establishing a productive breeding colony of Gad2-Cre mice, gaining regulatory approval for all experiments through the Institutional Animal Care and Usage Committee and Institutional Biosafety Committee at UCLA, and setting up the quantitative tremor assessment system.
We are extremely grateful to the IETF and its donors for the ongoing support of this study.