Single neuron models are fundamental for computational modeling of the brain's neuronal networks, and understanding how ion channel dynamics mediate neural function. A challenge in defining such models is determining biophysically realistic channel distributions. Here, we present an efficient, highly parallel evolutionary algorithm for developing such models, named NeuroGPU-EA. NeuroGPU-EA uses CPUs and GPUs concurrently to simulate and evaluate neuron membrane potentials with respect to multiple stimuli. We demonstrate a logarithmic cost for scaling the stimuli used in the fitting procedure. NeuroGPU-EA outperforms the typically used CPU based evolutionary algorithm by a factor of 10 on a series of scaling benchmarks. We report observed performance bottlenecks and propose mitigation strategies. Finally, we also discuss the potential of this method for efficient simulation and evaluation of electrophysiological waveforms.

Epileptic encephalopathies are severe forms of infantile-onset epilepsy often complicated by severe neurodevelopmental impairments. Some forms of early-onset epileptic encephalopathy (EOEE) have been associated with variants in SCN2A, which encodes the brain voltage-gated sodium channel NaV1.2. Many voltage-gated sodium channel genes, including SCN2A, undergo developmentally regulated mRNA splicing. The early onset of these disorders suggests that developmentally regulated alternative splicing of NaV1.2 may be an important consideration when elucidating the pathophysiological consequences of epilepsy-associated variants. We hypothesized that EOEE-associated NaV1.2 variants would exhibit greater dysfunction in a splice isoform that is prominently expressed during early development. We engineered five EOEE-associated NaV1.2 variants (T236S, E999K, S1336Y, T1623N, and R1882Q) into the adult and neonatal splice isoforms of NaV1.2 and performed whole-cell voltage clamp to elucidate their functional properties. All variants exhibited functional defects that could enhance neuronal excitability. Three of the five variants (T236S, E999K, and S1336Y) exhibited greater dysfunction in the neonatal isoform compared with those observed in the adult isoform. Computational modeling of a developing cortical pyramidal neuron indicated that T236S, E999K, S1336Y, and R1882Q showed hyperexcitability preferentially in immature neurons. These results suggest that both splice isoform and neuronal developmental stage influence how EOEE-associated NaV1.2 variants affect neuronal excitability.

The medial prefrontal cortex plays a key role in higher order cognitive functions like decision making and social cognition. These complex behaviors emerge from the coordinated firing of prefrontal neurons. Fast-spiking interneurons (FSIs) control the timing of excitatory neuron firing via somatic inhibition and generate gamma (30-100 Hz) oscillations. Therefore, factors that regulate how FSIs respond to gamma-frequency input could affect both prefrontal circuit activity and behavior. Here, we show that serotonin (5HT), which is known to regulate gamma power, acts via 5HT2A receptors to suppress an inward-rectifying potassium conductance in FSIs. This leads to depolarization, increased input resistance, enhanced spiking, and slowed decay of excitatory post-synaptic potentials (EPSPs). Notably, we found that slowed EPSP decay preferentially enhanced temporal summation and firing elicited by gamma frequency inputs. These findings show how changes in passive membrane properties can affect not only neuronal excitability but also the temporal filtering of synaptic inputs.

G-protein-coupled receptors (GPCRs) initiate a variety of signaling cascades, depending on effector coupling. β-arrestins, which were initially characterized by their ability to "arrest" GPCR signaling by uncoupling receptor and G protein, have recently emerged as important signaling effectors for GPCRs. β-arrestins engage signaling pathways that are distinct from those mediated by G protein. As such, arrestin-dependent signaling can play a unique role in regulating cell function, but whether neuromodulatory GPCRs utilize β-arrestin-dependent signaling to regulate neuronal excitability remains unclear. Here, we find that D3 dopamine receptors (D3R) regulate axon initial segment (AIS) excitability through β-arrestin-dependent signaling, modifying CaV3 voltage dependence to suppress high-frequency action potential generation. This non-canonical D3R signaling thereby gates AIS excitability via pathways distinct from classical GPCR signaling pathways.

Background

Variants in the SCN2A gene that disrupt the encoded neuronal sodium channel Na_V1.2 are important risk factors for autism spectrum disorder (ASD), developmental delay, and infantile seizures. Variants observed in infantile seizures are predominantly missense, leading to a gain of function and increased neuronal excitability. How variants associated with ASD affect Na_V1.2 function and neuronal excitability are unclear.

Methods

We examined the properties of 11 ASD-associated SCN2A variants in heterologous expression systems using whole-cell voltage-clamp electrophysiology and immunohistochemistry. Resultant data were incorporated into computational models of developing and mature cortical pyramidal cells that express Na_V1.2.

Results

In contrast to gain of function variants that contribute to seizure, we found that all ASD-associated variants dampened or eliminated channel function. Incorporating these electrophysiological results into a compartmental model of developing excitatory neurons demonstrated that all ASD variants, regardless of their mechanism of action, resulted in deficits in neuronal excitability. Corresponding analysis of mature neurons predicted minimal change in neuronal excitability.

Conclusions

This functional characterization thus identifies SCN2A mutation and Na_V1.2 dysfunction as the most frequently observed ASD risk factor detectable by exome sequencing and suggests that associated changes in neuronal excitability, particularly in developing neurons, may contribute to ASD etiology.

Loss-of-function variants in the gene SCN2A, which encodes the sodium channel Na_V1.2, are strongly associated with autism spectrum disorder and intellectual disability. An estimated 20%-30% of children with these variants also suffer from epilepsy, with altered neuronal activity originating in neocortex, a region where Na_V1.2 channels are expressed predominantly in excitatory pyramidal cells. This is paradoxical, as sodium channel loss in excitatory cells would be expected to dampen neocortical activity rather than promote seizure. Here, we examined pyramidal neurons lacking Na_V1.2 channels and found that they were intrinsically hyperexcitable, firing high-frequency bursts of action potentials (APs) despite decrements in AP size and speed. Compartmental modeling and dynamic-clamp recordings revealed that Na_V1.2 loss prevented potassium channels from properly repolarizing neurons between APs, increasing overall excitability by allowing neurons to reach threshold for subsequent APs more rapidly. This cell-intrinsic mechanism may, therefore, account for why SCN2A loss-of-function can paradoxically promote seizure.

Background

The membrane potential of individual neurons depends on a large number of interacting biophysical processes operating on spatial-temporal scales spanning several orders of magnitude. The multi-scale nature of these processes dictates that accurate prediction of membrane potentials in specific neurons requires the utilization of detailed simulations. Unfortunately, constraining parameters within biologically detailed neuron models can be difficult, leading to poor model fits. This obstacle can be overcome partially by numerical optimization or detailed exploration of parameter space. However, these processes, which currently rely on central processing unit (CPU) computation, often incur orders of magnitude increases in computing time for marginal improvements in model behavior. As a result, model quality is often compromised to accommodate compute resources.

New method

Here, we present a simulation environment, NeuroGPU, that takes advantage of the inherent parallelized structure of the graphics processing unit (GPU) to accelerate neuronal simulation.

Results & comparison with existing methods

NeuroGPU can simulate most biologically detailed models 10-200 times faster than NEURON simulation running on a single core and 5 times faster than GPU simulators (CoreNEURON). NeuroGPU is designed for model parameter tuning and best performs when the GPU is fully utilized by running multiple (> 100) instances of the same model with different parameters. When using multiple GPUs, NeuroGPU can reach to a speed-up of 800 fold compared to single core simulations, especially when simulating the same model morphology with different parameters. We demonstrate the power of NeuoGPU through large-scale parameter exploration to reveal the response landscape of a neuron. Finally, we accelerate numerical optimization of biophysically detailed neuron models to achieve highly accurate fitting of models to simulation and experimental data.

Conclusions

Thus, NeuroGPU is the fastest available platform that enables rapid simulation of multi-compartment, biophysically detailed neuron models on commonly used computing systems accessible by many scientists.

Autism spectrum disorder (ASD) is strongly associated with de novo gene mutations. One of the most commonly affected genes is SCN2A. ASD-associated SCN2A mutations impair the encoded protein Na_V1.2, a sodium channel important for action potential initiation and propagation in developing excitatory cortical neurons. The link between an axonal sodium channel and ASD, a disorder typically attributed to synaptic or transcriptional dysfunction, is unclear. Here we show that Na_V1.2 is unexpectedly critical for dendritic excitability and synaptic function in mature pyramidal neurons in addition to regulating early developmental axonal excitability. Na_V1.2 loss reduced action potential backpropagation into dendrites, impairing synaptic plasticity and synaptic strength, even when Na_V1.2 expression was disrupted in a cell-autonomous fashion late in development. These results reveal a novel dendritic function for Na_V1.2, providing insight into cellular mechanisms probably underlying circuit and behavioral dysfunction in ASD.