Dimitri Michael Kullmann, MD, PhD, professor of neurology at University College London (UCL), has spent years finding an approach to treat neurons affected by neurodevelopmental and neuropsychiatric disorders, such as epilepsy. “There’s an enormous unmet need,” Kullmann told GEN. “A lot of people with epilepsy don’t respond well to medication, and then some of these people are offered surgery that is invasive and limited by the risk of irreversible neurological deficit.”

But now, Kullmann and colleagues including Gabriele Lignani at UCL may have picked up a strategic trail to the gene therapy solution for epilepsy. In a research article published in Science, the team, led by PhD student Yichen Qiu, propose a gene therapy strategy that down-regulates the excitability of overactive neurons, and tested it in models of epilepsy with mice and human mini-brains. The AAV-based gene therapy approach reduced neuronal excitability by harnessing biochemical signals that increase in response to seizure-related activity, leading to a persistent anti-epileptic effect without interfering with normal behaviors. These results show that activity-dependent gene therapy is a promising on-demand cell-autonomous treatment for brain circuit disorders.

Genetic therapies have tantalized researchers like Kullmann as promising strategies to treat this seizure disorder and other electrically atypical pathologies by modulating neuronal excitability in a region-specific and cell-type-specific manner. However, this approach is limited in discriminating between neurons involved in circuit pathologies and “healthy” surrounding or intermingled neurons. In the case of epilepsy, subpopulations of neurons exhibit stereotypical patterns of pathological discharges during seizures. Targeting these neurons would be an essential step toward a rational treatment with minimal side effects. “The challenge was to develop a treatment that could discriminate between neurons participating in the seizures and bystander or intermingled neurons,” said Kullmann.

The study “On-demand cell-autonomous gene therapy for brain circuit disorders” was published in the latest issue of Science (VOL 378 ISSUE 6619).

A solution to seizures

In the past, the gene therapies Kullmann and his colleagues had worked with used a constitutive promoter to drive the transgene expression. Kullmann said that the key to this gene therapy problem was to build in some components regulatable by electrical activity. “For instance, with chemogenetics, a gene therapy can be permanent but the effect size controlled with a drug to have more room in the therapeutic window not to be dictated by the viral vector dose. However, chemogenetics does not achieve the goal of making the therapy switch itself on and off autonomously,” Kullmann told GEN.

Kullmann says his colleague Gabriele Lignani, PhD, associate professor at UCL, proposed using the promoters of immediate early genes (IEGs) to tag neurons. IEGs evolved to regulate the expression of other genes in response to activity. Several biochemical pathways converge on IEGs. These activity-dependent genes switch on in response to intense neuronal activity via calcium or neurotransmitters, and the IEGs modulate the transcription of other genes. IEG promoters have previously been used to drive the expression of optogenetic tools to manipulate memory

Kullmann, Lignani, and PhD student Yichen Qiu sought to identify an IEG with a promoter that could be used to drive the transcription of an ion channel to stabilize neuron excitability. “We were not interested in the gene product of the IEG, in this case, cFos,” said Kullman. “We’re taking the promoter for cFos and then harnessing that to control the transcription of a therapeutic transgene.”

An activity-dependent approach

So, the UCL research team created a construct where the cFos promoter drove the expression of an engineered potassium channel (EKC) that, when overexpressed, decreases both neuronal excitability and synaptic neurotransmitter release. In this Science research article, Qiu and colleagues reported that their activity-dependent, cell-autonomous, on-demand gene therapy that follows network dynamics drives the expression of a transgene that decreases neuronal excitability.

“[With IEGs], if the seizures go away, the transgene expression will drift down to baseline and back to the ground state, and it’s as if the neurons hadn’t been modified, which is another way of mitigating concerns about irreversible manipulation of the brain,” Kullmann told GEN. “For the neurons participating in the seizures, this approach is agnostic as to whether they are triggering the seizure or downstream neurons propagating it.”

“The critical thing here is that the dynamic range over which these IEGs switch on as seizures are associated with the extremely intense activity of neurons,” said Kullmann. “We are often asked whether the [IEGs] are ‘immediate’ enough. They switch on for over half an hour, and a typical seizure lasts less than that. We’re not going to be able to stop that initial seizure. The strategy is really to prevent the next seizure.”

Using many different, sophisticated behavioral tests, the University College London researchers demonstrate that this approach is specific for neurons that participate in pathological network activity but is also time-limited in that transgene expression persists only for as long as neurons are hyperactive. In a chronic model of epilepsy, the gene therapy approach achieved a greater decrease in seizure frequency than previously reported for gene therapy with constitutive promoters or with widely used antiseizure drugs and without deleterious effects on normal behaviors.

Schematic of the Activity-dependent Gene Therapy
Schematic of the Activity-dependent Gene Therapy. [Gabriele Lignani]

Climbing to the clinic

Qiu thinks there’s still quite a way to go before this gene therapy can go into the clinic. “We’ve collected some really good data to show there are no behavioral off-target effects; but then again, it is a long way from where we are now to the clinic,” said Qiu.

Promisingly, IEGs appear to be conserved. “There is data that if you take tissue from people undergoing epilepsy surgery, you can detect cFos,” said Kullmann. “A study shows that the pathological activity in these people’s brains before surgery correlates with the amount of cFos expression. So, that’s quite conserved.”

The translational potential is underlined by the preliminary evidence of effectiveness in human cortical assembloids (hCAs). At the end of the paper, Qiu and colleagues successfully applied the gene therapy system in human mini-brains in a dish. “Admittedly, it’s a very reduced system, but it is human. We can harness the same cFos-EKC channel construct, which switches on in response to a chemical stimulus that triggers an electrical event that looks much like a seizure. And indeed, that attenuates the effect of a second stimulus compared to the control virus. In principle, this should work in a human brain. Of course, we have to be very cautious in making that claim because it’s a model that is very far from the normal brain.”

Kullmann speculates that epilepsy is just one neuropsychiatric or neurological disease characterized by the overactivity of populations of neurons. “There is evidence in Parkinson’s disease for overactivity in specific circuits; same with early-stage schizophrenia or first-episode psychosis,” said Kullmann. “Whether suppressing that would prevent a patient from developing the whole disease course in schizophrenia is an open question. But, we already have some diseases treated with deep brain stimulation, which is thought to work paradoxically by suppressing specific neurons, such as Parkinson’s.”

Diving deep with delivery

One of the other issues concerning translation to the clinic is the delivery. But, given that the components of activity-dependent gene therapy are typical mammalian genetic elements packaged in a well-tolerated viral vector already in the clinic, there is a relatively straightforward path to first-in-human studies. What’s more, Kullmann asks how viral vectors will reach the right neurons. “It’s easy enough with mice because the brain is quite small; just do one injection, and you get quite good coverage,” said Kullmann. “Of course, it’s more of a challenge for a human, but that hasn’t stopped people use viral vectors similar to this to treat other neurological diseases. More and more is known about how to get these AAVs into the human brain. It’ll be an exciting journey.”