Research Article: AMPA GluA1-flip targeted oligonucleotide therapy reduces neonatal seizures and hyperexcitability

Date Published: February 8, 2017

Publisher: Public Library of Science

Author(s): Nicole M. Lykens, David J. Coughlin, Jyoti M. Reddi, Gordon J. Lutz, Melanie K. Tallent, Giuseppe Biagini.


Glutamate-activated α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (AMPA-Rs) mediate the majority of excitatory neurotransmission in brain and thus are major drug targets for diseases associated with hyperexcitability or neurotoxicity. Due to the critical nature of AMPA-Rs in normal brain function, typical AMPA-R antagonists have deleterious effects on cognition and motor function, highlighting the need for more precise modulators. A dramatic increase in the flip isoform of alternatively spliced AMPA-R GluA1 subunits occurs post-seizure in humans and animal models. GluA1-flip produces higher gain AMPA channels than GluA1-flop, increasing network excitability and seizure susceptibility. Splice modulating oligonucleotides (SMOs) bind to pre-mRNA to influence alternative splicing, a strategy that can be exploited to develop more selective drugs across therapeutic areas. We developed a novel SMO, GR1, which potently and specifically decreased GluA1-flip expression throughout the brain of neonatal mice lasting at least 60 days after single intracerebroventricular injection. GR1 treatment reduced AMPA-R mediated excitatory postsynaptic currents at hippocampal CA1 synapses, without affecting long-term potentiation or long-term depression, cellular models of memory, or impairing GluA1-dependent cognition or motor function in mice. Importantly, GR1 demonstrated anti-seizure properties and reduced post-seizure hyperexcitability in neonatal mice, highlighting its drug candidate potential for treating epilepsies and other neurological diseases involving network hyperexcitability.

Partial Text

Glutamate-activated α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (AMPA-Rs) mediate the majority of brain fast excitatory neurotransmission. The four AMPA-R subunits, GluA1-4, typically assemble as tetrameric cation channels from heterodimers of GluA1, 3 or 4 combined with GluA2, or GluA1 homodimers [reviewed in[1]]. Further GluA subunit diversity is generated by alternative splicing which critically regulates AMPA-R properties including intracellular trafficking, glutamate sensitivity, and desensitization kinetics [2,3]. Ionotropic AMPA-R GluA subunits are alternatively spliced at the flip/flop cassette, and GluA “flip” or “flop” splice variant composition confers substantially different channel properties in the brain [4,5]. Flip and flop are tandem 115 nucleotide cassette exons which code for a portion of transmembrane region 4 and confer only 9–11 amino acid difference between isoforms to regulate channel kinetics and glutamate sensitivity [4]. GluA2-4 flip containing channels have much slower desensitization kinetics with fast resensitization compared to those containing the respective flop isoforms [2,6,7,8]. In contrast, GluA1-flip and GluA1-flop containing channels have similar desensitization kinetics, but GluA1-flip confers greater glutamate sensitivity, giving rise to larger amplitude responses to given concentrations of glutamate than GluA1-flop [4,5]. These characteristics make AMPA-Rs containing GluA-flip isoforms higher gain channels compared to GluA-flop. However, the contributions of GluA flip and flop isoforms to synaptic transmission and neuronal network properties are unknown, due to a lack of experimental compounds that specifically modulate splice variants of individual GluAs.

The SMO developed in this study provided the opportunity to evaluate the impact of direct manipulation of GluA flip/flop isoform expression on AMPA-R mediated synaptic transmission in vivo, as a first step to understanding how GluA alternative splicing can be harnessed as an anti-seizure or neuroprotective therapeutic. Compared to AMPA-R’s containing a native mixture of GluA1-flip and flop subunits, our results showed that reducing GluA1-flip levels decreased amplitudes of CA1 aEPSCs, while decay kinetics were unaffected. This suggests GluA1-flip subunits mediate larger synaptic responses than GluA1-flop containing AMPA-Rs. Importantly, these findings confirm in vitro studies, that unlike other AMPA-R subunits, flip and flop variants of GluA1 differ in sensitivity to glutamate and resulting synaptic gain, but not in desensitization kinetics [52,53,54]. An increase in GluA1-flip expression, such as occurs in epileptic brain [12,13,14,15], would therefore increase network excitability, leading to increased susceptibility to seizures. The marked decrease in both aEPSCs and CTZ sensitivity seen following a reduction in GluA1-flip expression is also contrary to the current dogma that GluA1 does not contribute significantly to synaptic transmission under basal conditions, but rather is inserted at the synapse only after high-frequency input [55,56]. In support of our results, GluA1 knockout mice showed reduced responsiveness to uncaged glutamate at CA1 synaptic sites [57]. Together this data indicates that GluA1-flip mediates a much larger component of the synaptic response early in development than previously thought, making this particular AMPA-R subunit isoform a prime target to prevent hyperexcitability in neonatal epilepsy.