Gene Therapy and Autism: Exploring the Current Potential in Monogenic and Polygenic Autism - Introducing Nexus Gene Therapy

 

But, that genetic magic will come with a big price tag!

  

Gene therapy seems to be the way the pharmaceutical industry wants to go when it comes to severe autism. There is a combination of disinterest and dislike when it comes to repurposing cheap old drugs for new uses in the brain. Very high tech, but ultra expensive, gene therapies are attractive to the industry.

As I was writing this post the phase 3 trial for gene therapy for Angelman syndrome was announced to have failed. There were 129 participants, ages 4–17, all with a genetically confirmed full maternal UBE3A deletion (the most common Angelman genotype). The primary endpoint was a change from baseline in the Bayley-4 cognitive raw score at day 338. The secondary endpoint was net response on the Multidomain Responder Index (MDRI), spanning cognition, receptive communication, behavior, gross motor function, and sleep.

So, in this case the placebo was as good as the real gene therapy. This was a trial where all participants had the same single gene autism. Back to the drawing board!

Those 129 had the same affected gene, same broad class of mutation, but not the identical anomaly. Maybe that matters. Maybe the therapy never reached the right neurons. Maybe it just does not work.  

Anyway, back to the post.

 

Could gene therapy eventually be relevant not only to rare single-gene forms of autism, but also to biologically defined subgroups of polygenic or currently idiopathic autism?

For many years, autism research has faced a fundamental problem: we have become increasingly good at finding genes associated with neurodevelopmental disability, but finding a gene is not the same as knowing how to treat it. Gene therapy may eventually change that.

The most obvious strategy is to identify a defective gene and attempt to restore its function. For several genetically defined neurodevelopmental disorders, gene and RNA therapies have already moved beyond theoretical discussion into preclinical and human clinical development. But there may be a bigger—and less obvious—possibility.

What if gene therapy does not always have to target the gene that originally caused the disorder?

What if it could instead target a downstream biological nexus point: a strategically important gene or physiological control system through which the consequences of many different genetic influences converge? That idea could potentially make genetic technologies relevant not only to monogenic disorders, but eventually to some biologically defined forms of polygenic or idiopathic autism.

This does not mean there will be one gene therapy for autism. Probably the opposite. The future may involve a growing toolbox capable of intervening at different points between:

Genetic variation → Altered biology → Physiological dysfunction → Clinical problems

 

First, what do we mean by gene therapy?

Before going further, it helps to understand the basic chain of biology:

DNA → RNA → Protein

• DNA contains genetic instructions.
• RNA carries copied instructions inside cells.
• Proteins perform most functional cellular tasks.

Different therapies can intervene at different points in this chain. They might:

• Alter DNA directly
• Add a functioning gene
• Change how strongly a gene is expressed
• Block a harmful RNA
• Reactivate a silenced gene
• Temporarily provide instructions for making a protein

So gene therapy is not a single technology. It is a family of tools operating at different stages of genetic expression.

 

Autism is not one genetic disease

Autism is extraordinarily heterogeneous. Some people have a relatively well-defined monogenic disorder in which disruption of a single gene has a major effect on neurodevelopment:

• MECP2 — Rett syndrome
• UBE3A — Angelman syndrome
• SHANK3 — Phelan-McDermid syndrome
• TCF4 — Pitt-Hopkins syndrome
• SYNGAP1 and other genetically defined conditions

Autism can be part of the phenotype in these conditions. But these disorders are biologically distinct from autism resulting from the combined influence of many polygenic variants.

For a monogenic disorder, the logic is straightforward:

Identify disrupted gene → Understand biological consequence → Restore or regulate function

For polygenic autism, there is no single defective gene waiting to be repaired. But that does not mean gene therapy will remain irrelevant.

The Obvious Starting Point: Treating the Upstream Genetic Cause

Gene Replacement

If a person has insufficient function of a particular gene, one strategy is delivering a functioning copy using modified viruses, particularly adeno-associated viruses (AAVs).

Conceptually:

Too little functional protein → Provide genetic instructions to make more

However, neurological targets introduce tight constraints. The therapy must reach the correct brain regions, in the correct cells, at precise doses, for the correct duration. Sometimes, too much of the missing protein is dangerous.

 

Rett Syndrome Shows the Problem of Dosage

Rett syndrome is caused by insufficient function of MECP2. The simple solution would seem to be adding MECP2. However:

• Too little MECP2 causes Rett syndrome.
• Too much MECP2 causes severe MECP2 duplication syndrome.

This presents a classic Goldilocks problem: not too little, not too much, but just enough. The challenge is not merely delivering a gene—it is controlling biological expression within narrow safety margins.

 

Gene unsilencing: Using a gene that is already there

Sometimes the required gene is present but silenced. In Angelman syndrome, the paternal copy of UBE3A is silenced in relevant neurons.

Instead of adding an exogenous gene, several RNA-based strategies attempt to reactivate the silent paternal copy already in the cell. This represents a major conceptual pivot: changing the regulation of an existing gene rather than delivering a replacement.

 

The monogenic landscape: Where do we stand today?

While the downstream nexus hypothesis remains an emerging paradigm for polygenic autism, gene-targeted therapies for rare monogenic neurodevelopmental syndromes have transitioned from academic theory into pivotal human clinical trials.

Because these single-gene disorders share significant clinical overlap with autism, their clinical progress serves as the proving ground for delivery systems, dosage control, and regulatory frameworks.

 

Monogenic Target	Therapeutic Strategy	Current Clinical Phase
MECP2 (Rett Syndrome)	AAV9 Gene Replacement + Regulated Dosage Control	Phase 1/2 Registrational (NGN-401, TSHA-102)
UBE3A (Angelman Syndrome)	Antisense Oligonucleotides (ASOs) for Paternal Gene Unsilencing	Phase 3 mixed: GTX-102 Aspire trial missed its primary endpoint (Sept 2026); ION582 in Phase 1/2
TCF4 (Pitt-Hopkins Syndrome)	ICV Delivery AAV Gene Therapy (TCF4 Transgene Replacement)	Early Phase 1/2 (MZ-1866)
SHANK3 (Phelan-McDermid)	AAV Delivery Target Replacement & Small-Molecule Modulators	JAG201 in early Phase 1/2; NNZ-2591 (small molecule) advanced to Phase 3 (Koala study)

 

1. MECP2 — Rett Syndrome (The Dosage Control Benchmark)

• Mechanism: Delivering functional copies of the MECP2 gene via AAV9 vectors.
• Current Status: Advanced Phase 1/2 and registrational studies (e.g., Neurogene’s NGN-401 and TSHA-102).
• Key Translation Insight: Early attempts at unconstrained MECP2 gene replacement carried extreme risks of MECP2 duplication toxicity. Leading clinical programs now incorporate engineered dosage-control platforms (such as built-in microRNA binding sites or regulated promoter systems) designed to restrict protein expression to physiological windows.

 

2. UBE3A — Angelman Syndrome (The Unsilencing Benchmark)

• Mechanism: Intrathecally delivered Antisense Oligonucleotides (ASOs) designed to degrade the UBE3A-ATS transcript, thereby un-silencing the healthy paternal copy of UBE3A in neurons.
• Current Status: Mixed. Ionis’s ION582 remains in Phase 1/2. Ultragenyx’s GTX-102 (apazunersen) reached a fully enrolled, pivotal Phase 3 trial (Aspire), but in September 2026 Ultragenyx announced the trial missed its primary endpoint, showing no significant benefit over sham treatment.
• Key Translation Insight: Unlike one-time viral gene therapy, ASOs require periodic intrathecal administration, offering reversibility and dose titratability. But the Aspire result is a sobering data point in its own right: even a mechanistically well-validated, genetically homogeneous target (a confirmed maternal UBE3A deletion) can fail a rigorously controlled pivotal trial. Ultragenyx has pointed to encouraging long-term open-label extension data (KIK-AS) as reason to keep investigating; whether that reflects a real but harder-to-capture effect, a placebo-sensitive endpoint, or a genuinely ineffective mechanism is still unresolved. Either way, it is a live illustration of the same lesson bumetanide already taught: a plausible downstream target can look strong in early, uncontrolled data and still not clear a controlled, blinded trial.

 

3. TCF4 — Pitt-Hopkins Syndrome (Direct CNS Delivery)

• Mechanism: Direct central nervous system delivery (such as intracerebroventricular injection) of AAV vectors carrying functional TCF4 genetic instructions.
• Current Status: Phase 1/2 early-stage safety and dosing trials (e.g., MZ-1866).
• Key Translation Insight: Because TCF4 is a crucial transcription factor across brain development, achieving widespread CNS biodistribution while avoiding systemic viral exposure is the core technical focus.

 

4. SHANK3 — Phelan-McDermid Syndrome (Synaptic Scaffolding)

• Mechanism: AAV-mediated targeted gene replacement (e.g., JAG201) aimed at restoring synaptic density and post-synaptic scaffolding proteins, paired with small-molecule downstream pathway modulators (e.g., NNZ-2591).
• Current Status: JAG201 (AAV gene replacement) is in early Phase 1/2 cohort dosing. NNZ-2591, a small-molecule downstream modulator rather than a gene therapy, has advanced further — a Phase 2 trial reported symptom improvements, and it is now in a pivotal Phase 3 trial (Koala).
• Key Translation Insight: SHANK3 insufficiency directly impacts synaptic structural integrity. Trials here provide a crucial model for whether structural synaptic restoration in older individuals can yield measurable behavioral or functional gains.

 

What these clinical trials teach us about polygenic applications

1.     Delivery Architecture: The shift away from high-dose IV administration toward intrathecal or intracerebroventricular routes in these monogenic trials directly addresses vector safety concerns, establishing lower-dose CNS administration protocols.

2.     Reversibility vs. Durability: The parallel development of one-time AAVs (Rett) and repeatable ASOs (Angelman) illustrates the trade-off between permanent modification and adjustable maintenance dosing. It is not yet a precedent for which approach works better: GTX-102’s pivotal-trial miss is a reminder that reversibility is a safety and flexibility advantage, not a guarantee of efficacy.

3.     Biomarker Validation: Every monogenic success relies on rigorous molecular and electrophysiological endpoints (such as quantitative EEG patterns or specific protein assays)—reinforcing the argument that downstream polygenic targets will fail without equivalent biomarker stratification.

 

The harder question: What about polygenic autism?

In polygenic autism, correcting every contributing upstream genetic variant is unfeasible:

However, upstream genetic complexity does not necessarily produce equal complexity downstream. Multiple upstream variants may converge on shared biological systems, including:

• Synaptic function and plastic adjustments
• Neuronal excitability and E/I balance
• Gene regulation networks and chromatin remodeling
• Protein translation pathways
• Cellular metabolism and mitochondrial function

 

The central thesis: The gene targeted therapeutically does not necessarily have to be the gene that originally caused the disorder.

 

A Remarkable Experiment in Down Syndrome

The clearest preclinical proof-of-concept for downstream nexus targeting comes from Down syndrome.

Down syndrome involves an extra copy of chromosome 21, altering the dosage of hundreds of genes simultaneously. Correcting the primary chromosomal defect across the brain is currently out of reach.

In 2021, Parrini and colleagues targeted a downstream physiological system: neuronal chloride regulation. Using AAV-mediated, neuron-specific RNA interference, they reduced expression of the NKCC1 chloride transporter in the Ts65Dn mouse model of Down syndrome.

 

The researchers did not correct chromosome 21. They targeted a downstream physiological consequence instead. This experiment provides a concrete model for how complex neurodevelopmental disorders might be approached.

 

NKCC1: An Example of a Therapeutic Nexus

NKCC1 (encoded by SLC12A2) imports chloride into cells:

NKCC1 expression → Intracellular chloride level → Neuronal response to GABA → Network inhibition

 

But didn't Bumetanide fail?

Bumetanide, a small-molecule drug that inhibits NKCC1, was studied as a potential autism treatment. Early pilot studies generated interest, but two large Phase III trials enrolled 211 participants each and found no significant benefit over placebo across the overall unselected ASD study population, leading to early trial termination.

This negative result is important, but comparing systemic bumetanide trials to neuron-specific gene therapy highlights three distinct scientific variables:

Parameter	Systemic Bumetanide Trials	Preclinical AAV-NKCC1 Study
Targeting	Systemic drug with limited brain penetration	Targeted CNS neuronal knockdown
Population	Broad, unselected clinical ASD population	Defined biological model of chloride elevation
Mechanism	Transient pharmacological inhibition	Stable genetic recalibration of expression

 

Broad clinical trials can mask efficacy if a drug is tested across a heterogeneous population where only a fraction of participants possess the targeted biological abnormality. Targeted interventions require verified biomarkers.

 

A new example: Gene editing a downstream metabolic nexus

A striking new example comes from cholesterol rather than autism. In August 2026, researchers reported one-year results from the first-in-human Phase 1a trial of CTX310, a one-time gene therapy treatment designed to switch off ANGPTL3 in liver cells. ANGPTL3 is a regulator of lipid metabolism, and naturally occurring loss-of-function variants in this gene are associated with lifelong reductions in LDL cholesterol and triglycerides. In 15 adults with difficult-to-control lipid disorders, the highest dose of CTX310 produced a mean 52.5% reduction in LDL cholesterol and 47.8% reduction in triglycerides after one year.

The treatment was not correcting the original genetic causes of the patients' lipid disorders; instead, it was deliberately changing a downstream physiological control point. This is conceptually interesting for the hypothesis developed here. It provides a real human example of the principle that a genetic therapy does not necessarily have to repair the original disease-causing mutation: if different upstream causes converge on a measurable biological abnormality, it may sometimes be possible to intervene at the downstream nexus instead.

Of course, CTX310 is a lipid therapy, not an autism treatment, and this small Phase 1 study does not demonstrate that the same strategy will work in autism. But it shows that the broader concept of genetically modifying a downstream physiological regulator rather than correcting the initiating genetic defect has now entered human clinical research

  

Gene therapy does not have to mean permanent DNA editing

"Gene therapy" encompasses tools with varying degrees of reversibility:

  

Controllability as a safety feature

Permanent DNA edits carry risks if the target selection, tissue specificity, or dosing proves suboptimal.

Conversely, reversible tools (such as transient RNA or mRNA therapies) allow dose adjustments, treatment pauses, or modifications over time. For downstream target exploration, controllability is a key safety feature.

 

The economics of gene therapy will shape what gets developed

Monogenic vs. Nexus Economics

• Ultra-rare monogenic targets: Clear molecular mechanisms, but high development costs distributed over small patient populations create commercial challenges.
• Shared downstream targets: If multiple distinct conditions share a downstream biological mechanism, a single intervention could theoretically address a biomarker-defined subgroup spanning several diagnostic categories.

 

 

Durable vs. Repeat-Dose Models

• One-time durable therapies: High upfront costs, long-term regulatory monitoring requirements, and payment structure hurdles for healthcare systems.
• Repeat-dose RNA/mRNA therapies: Distributed costs over time, potential for adjustments, but ongoing administration requirements for patients.

 

Reusable delivery platforms

Developing standardized delivery capsids or lipid nanoparticles capable of carrying different genetic payloads to specific brain cell types could shift the paradigm from one disease, one custom technology toward one platform, multiple biological payloads.

 

Safety: Gene therapy can be dangerous

Gene therapy carries significant physiological risks. In 2025, safety actions were taken following fatal acute liver failure cases in high-dose systemic AAVrh74 trials (Sarepta/Elevidys).

Risks depend heavily on delivery route, vector, and total dose:

• Systemic high-dose IV administration: Requires large viral loads to achieve brain exposure, driving systemic organ exposure and hepatic toxicity.
• Direct CNS-directed administration: Intrathecal or intracerebroventricular delivery targets the central nervous system directly, significantly reducing total vector load and systemic exposure.

 

Evaluating risk

Evaluating risk requires comparing potential treatment adverse events against the natural history of the untreated condition:

 

 

Because autism varies from stable clinical profiles to severe conditions with profound self-injury or treatment-resistant epilepsy, acceptable safety margins will differ depending on individual medical contexts.

 

The translation pathway: A stepwise sequence

Before durable genetic interventions are considered for downstream targets, evidence must be established sequentially:

  

Summary of Evidence Status

Polyvalent mRNA nexus therapy: the equivalent of a "genetic polypill"

If we accept that polygenic autism involves multiple converging biological pathways, targeting a single downstream gene—whether NKCC1, EIF4E, or another physiological node—may rarely capture the full biological picture for a given individual.

This raises a compelling long-term question: Could the clinical principle of a personalized multi-target "Polypill" eventually be realized through programmable RNA platforms?

 

The concept: modular, multi-payload delivery

In immunology, polyvalent vaccines combine distinct antigen sequences into a single formulation to address multiple viral strains simultaneously. A polyvalent mRNA platform for central nervous system (CNS) targets would operate on a similar architecture:

1.     A universal delivery platform: Utilizing advanced delivery vehicles—such as cell-type-specific lipid nanoparticles (LNPs) or engineered viral capsids—designed to cross the blood-brain barrier and target specific neural cells.

2.     Modular payload cocktails: Rather than delivering a single transcript, the platform carries a custom-blended payload ratio tailored to an individual’s specific biological profile:

o    Payload A: Recalibrating neuronal chloride regulation (e.g., modulating NKCC1/KCC2 ratios).

o    Payload B: Normalizing hyperactive translation initiation pathways (e.g., dampening EIF4E or mTOR signaling).

o    Payload C: Supporting mitochondrial cellular bioenergetics or resolving neuroinflammatory responses.

 

The intervention becomes genuinely personalized and pathway-driven, rather than label-based.

 

Key advantages: reversibility and developmental adaptability

Unlike permanent DNA editing, mRNA-based interventions are transient. In complex neurodevelopmental conditions, this transience provides two critical clinical safeguards:

• Age-adaptive formulations: Human brain biology is not static. A young child at age four may require a payload formulation optimized for synaptic maturation and circuit refinement, whereas the same individual at age sixteen might benefit from a revised ratio targeting network excitability or cellular metabolism.
• Titratability and safety: If an individual experiences an adverse response or if an objective biomarker fails to shift as predicted, the dosage of a specific payload component can be adjusted, swapped, or stopped entirely in subsequent administrations. Controllability serves as a primary safety mechanism.

 

A long-term horizon, not an imminent reality

To be clear polyvalent mRNA therapy represents a long-term conceptual horizon—a vision spanning decades—rather than an imminent treatment. Moving from theory to practice requires overcoming profound technical hurdles:

• Validating reliable, non-invasive biomarkers capable of measuring downstream pathway activity in living human brain tissue.
• Achieving cell-type specificity (e.g., excitatory neurons vs. inhibitory interneurons vs. astrocytes) without off-target toxicity.
• Determining precise therapeutic windows for multi-target RNA dosing in developing neural circuits.

Yet, while the engineering required to deliver multi-target brain therapies remains futuristic, the underlying principle is already relevant today, moving away from broad diagnostic categories toward precise, biomarker-guided combinations that address an individual’s actual biology.

ARBs and ACE inhibitors for Autism, an old Peter idea finally explored in a research model

 

 A home run? Certainly worth further consideration. 

When I was doing my review of unexplored potential autism therapies several years ago, I did look at two closely related classes of drugs. ARBs and ACE inhibitors.

I wrote about it in blog posts and set out why I thought the ARB telmisartan was the best one to trial first.

 

           Targeting Angiotensin in Schizophrenia and Some Autism          

Just when you thought we had run out hormones to connect to autism and schizophrenia, today we have Angiotensin.

Angiotensin is a hormone that causes vasoconstriction and a subsequent increase in blood pressure. It is part of the renin-angiotensin system, which is a major target for drugs (ACE inhibitors) that lower blood pressure. Angiotensin also stimulates the release of aldosterone, a hormone that promotes sodium retention which also drives blood pressure up.

Angiotensin I has no biological activity and exists solely as a precursor to angiotensin II.

Angiotensin I is converted to angiotensin II  by the enzyme angiotensin-converting enzyme (ACE).  ACE is a target for inactivation by ACE inhibitor drugs, which decrease the rate of Angiotensin II production.  

It turns out that Angiotensin has some other properties very relevant to schizophrenia, some autism and quite likely many other inflammatory conditions. 

Blocking angiotensin-converting enzyme (ACE) induces those potent regulatory T cells that are lacking in autism and modulates Th1 and Th17 mediated autoimmunity.  See my last post on Th1,Th2 and Th17. 

In addition, Angiotensin II affects the function of the NKCC1/2 chloride cotransporters that are dysfunctional in much autism and at least some schizophrenia.

Then I wrote another post and made a trial of Telmisartan.

Angiotensin II in the Brain & Therapeutic Considerations

I was pleased to see that some researchers have recently published a paper on this subject. They chose an ACE inhibitor called Captopril.

 

Captopril restores microglial homeostasis and reverses ASD-like phenotype in a model of ASD induced by exposure in utero to anti-caspr2 IgG

Microglia play a crucial role in brain development, including synaptic pruning and neuronal circuit formation. Prenatal disruptions, such as exposure to maternal autoantibodies, can dysregulate microglial function and contribute to neurodevelopmental disorders like autism spectrum disorder (ASD). Maternal antibodies targeting the brain protein Caspr2, encoded by ASD risk gene Cntnap2, are found in a subset of mothers of children with ASD. In utero exposure to these antibodies in mice leads to an ASD-like phenotype in male but not in female mice, characterized by altered hippocampal microglial reactivity, reduced dendritic spine density, and impaired social behavior. Here, we studied the role of microglia in mediating the effect of in utero exposure to maternal anti-Caspr2 antibodies and whether we can ameliorate this phenotype. In this study we demonstrate that microglial reactivity emerges early in postnatal development and persists into adulthood following exposure in utero to maternal anti-Caspr2 IgG. Captopril, a blood-brain barrier permeable angiotensin-converting enzyme (ACE) inhibitor, but not enalapril (a non-BBB permeable ACE inhibitor) ameliorates these deficits. Captopril treatment reversed microglial activation, restored spine density and dendritic arborization in CA1 hippocampal pyramidal neurons, and improved social interaction. Single-cell RNA sequencing of hippocampal microglia identified a captopril-responsive subcluster exhibiting downregulated translation (eIF2 signaling) and metabolic pathways (mTOR and oxidative phosphorylation) in mice exposed in utero to anti-Caspr2 antibodies treated with saline compared to saline-treated controls. Captopril reversed these transcriptional alterations, restoring microglial homeostasis. Our findings suggest that exposure in utero to maternal anti-Caspr2 antibodies induces sustained neuronal alterations, microglial reactivity, and metabolic dysfunction, contributing to the social deficits in male offspring. BBB-permeable ACE inhibitors, such as captopril, warrant further investigation as a potential therapeutic strategy in a subset of ASD cases associated with microglial reactivity.

 

So here is an update that incorporates all these ideas and the new study.

 ___ 

Targeting the Brain Renin-Angiotensin System: From Schizophrenia to Autism (2025 Update)

By Peter Lloyd-Thomas, Epiphany ASD Blog

In 2017, I wrote about the idea that drugs targeting the renin–angiotensin system (RAS)—ACE inhibitors and ARBs—might have therapeutic effects beyond blood pressure, including in schizophrenia and autism. At that time, the discussion was mostly mechanistic. Today, new evidence strengthens the rationale and provides translational plausibility.

 

Why the Brain RAS Matters

While angiotensin II is best known for regulating blood pressure, the brain has its own RAS, which regulates:

·         AT₁ receptors → oxidative stress, neuroinflammation, microglial activation

·         AT₂ and Mas receptors → neuroprotection, mitochondrial function, anti-inflammatory signaling

·         ACE → converts Angiotensin I → II and degrades bradykinin, affecting cerebral blood flow

Shifting the balance from AT₁-dominated to AT₂/Mas signaling can normalize microglial function, improve neuronal energy metabolism, and support synaptic plasticity.

 

New Autism-Relevant Evidence (2025)

A recent study (Spielman et al., Molecular Psychiatry, 2025) used a mouse model of maternal anti-Caspr2 antibodies, a risk factor for some forms of autism. Male offspring showed:

·         Hyperactive microglia

·         Reduced hippocampal dendritic spines

·         Impaired social behavior

Captopril, a BBB-penetrant ACE inhibitor, reversed these deficits. In contrast, enalapril, which poorly crosses the BBB, was ineffective. Single-cell RNA sequencing revealed captopril restored microglial metabolic homeostasis (mTOR, oxidative phosphorylation, eIF2 signaling), linking microglial function directly to behavioral outcomes.

 

ACE Inhibitors vs ARBs: CNS and Immune Effects

Feature	ACE inhibitors (e.g., captopril)	ARBs (BBB-permeable, e.g., telmisartan)
↓ Ang II	Yes	No (blocks AT₁ receptor)
↑ Bradykinin / NO	Yes	No
BBB penetration	Variable — captopril high, enalapril low	Most low; telmisartan high
Microglial activation	↓ via less Ang II & more NO	↓ via AT₁ blockade
NKCC1/2 chloride cotransporters	Normalized via ↓ Ang II	Normalized via AT₁ blockade
Regulatory T cells (Tregs)	Strong ↑	Moderate ↑ (telmisartan strongest among ARBs)
Th1/Th17 autoimmunity	Modulated ↓	Modulated ↓
PPAR‑γ activation	No	Yes (telmisartan)
Evidence in ASD model	Captopril reversed phenotype (2025)	Mechanistically promising; anecdotal human benefit

Both classes modulate neuroinflammation, chloride signaling, and immune function, but ACE inhibitors and ARBs differ in mechanisms and potency.

 

Clinical Evidence in Schizophrenia

Telmisartan has been trialed in adults with schizophrenia (NCT00981526), primarily for metabolic side effects of antipsychotics (clozapine, olanzapine). Secondary observations included:

·         Improvement in negative symptoms

·         Modest cognitive benefits

·         Good tolerability over 12 weeks

This demonstrates CNS activity in humans, beyond metabolic effects, supporting translational plausibility for neuropsychiatric conditions.

 

Personal Observation in Autism

Years ago, I trialed telmisartan in my son. The effect was striking: he began singing spontaneously—something no other therapy had achieved. Singing engages emotion, motivation, and executive coordination, all dependent on healthy microglial and neuronal metabolism. While anecdotal, this observation aligns with mechanistic insights from both the mouse autism model and schizophrenia trials.

 

Safety and Accessibility

ACE inhibitors and ARBs are:

·         Widely prescribed globally for hypertension and heart protection

·         Generic, inexpensive, and safe in adults

·         Typically well-tolerated (ACE-i cough, hypotension, mild electrolyte changes)

This makes them practical candidates for drug repurposing in neurodevelopmental and neuropsychiatric disorders.

 

Mechanistic Summary

1.     Microglial hyperactivation contributes to synaptic and behavioral deficits in some autism subtypes.

2.     Brain RAS modulation (ACE-i or ARB) restores microglial homeostasis, improves energy metabolism, and supports synaptic plasticity.

3.     NKCC1/2 chloride cotransporter regulation: By reducing Ang II (ACE-i) or blocking AT₁ (ARB), these drugs normalize intracellular chloride, restoring proper GABAergic inhibition.

4.     Immune regulation: ACE inhibition induces regulatory T cells (Tregs) and modulates Th1/Th17 autoimmunity. BBB-penetrant ARBs like telmisartan also modulate these pathways, enhanced by PPAR‑γ activation.

5.     Behavioral outcomes: In mice, captopril reverses ASD-like phenotypes; anecdotal human reports suggest telmisartan may improve engagement, motivation, and communication.

 

Next Steps for Research

·         Carefully designed biomarker-driven pilot trials in humans, selecting individuals with evidence of neuroinflammation or maternal autoantibody exposure.

·         CNS-focused outcome measures (microglial imaging, inflammatory markers, synaptic function).

·         Behavioral endpoints relevant to autism (social interaction, expressive communication).

Or skip that and maybe make an n=1 trial?

 

Take-Home Message

Drugs long used for cardiovascular health may have untapped potential in neurodevelopmental and neuropsychiatric disorders. BBB-penetrant ACE inhibitors and ARBs, particularly telmisartan, can modulate:

·         Microglial activity

·         Neuronal chloride gradients

·         Immune regulation

Recent mouse data (Spielman et al., 2025) and human observations in schizophrenia support mechanistic plausibility and safety, making these drugs promising candidates for further study in selected autism subgroups.

 

References and Further Reading:

Spielman et al., Molecular Psychiatry, 2025: Captopril restores microglial homeostasis in anti-Caspr2 ASD model

NCT00981526, Telmisartan in schizophrenia (Fan X, 2018)

Lloyd-Thomas, 2017: Angiotensin II in the Brain

Lloyd-Thomas, 2017: Targeting Angiotensin in Schizophrenia and Some Autism