Guest Column | November 16, 2026

Safety, Durability, And Trial Design: Testing 1 Dose To Slow Huntington's

A conversation between P. Peter Ghoroghchian, MD, Ph.D., at Latus Bio and Life Science Connect's Jon O'Connell

Creative cooperation, teamwork-GettyImages-1417554692

A model is only as useful as the therapy that tests it. Latus Bio's is LTS-201, an AAV gene therapy that delivers an engineered microRNA targeting MSH3 to the medium spiny neurons and cortical projection neurons most vulnerable in Huntington's disease (HD).

Working with Children's Hospital of Philadelphia and the University of Pennsylvania, the company dosed nonhuman primates intrastriatally using its engineered AAV-DB-3 capsid and measured a 94% reduction in MSH3 mRNA inside target neurons — against 24% to 39% in bulk tissue, a gap it attributes to the many non-neuronal cells in a tissue sample.

That is a one-time dose against a protein in the DNA repair machinery, measured in two animals per group, with human durability still unknown. In part two of our discussion, Ghoroghchian addresses the long-term safety case, the competition from antisense and small interfering RNA (siRNA), and how a slowly progressing disease gets a readable clinical trial.

MSH3 is part of the cell's DNA repair machinery. You saw no oncogenic activation over the windows studied. But for a one-time dose, how are you thinking about long-term surveillance?

Ghoroghchian: Long-term safety is an important consideration, particularly when targeting a protein involved in DNA repair. However, human genetic studies provide encouraging evidence supporting MSH3 as a therapeutic target. Individuals with naturally occurring variants that reduce MSH3 expression have demonstrated delayed Huntington's disease onset and slower progression. Importantly, individuals with biallelic inherited loss of MSH3 do not demonstrate a higher risk for developing cancers; they do, however, show a higher incidence of a recessive colonic polyp syndrome. Taken together, this human evidence supports the potential therapeutic strategy aimed at MSH3 reduction but supports targeting HD-relevant cells so as to prevent systemic effects. Additionally, mice deficient in MSH3 have a normal lifespan.

This is Part 2 of an article exploring Latus Bio's model for predicting CAG repeat suppression in Huntington's disease therapy. Part 1 discusses the model Latus released to the public.

Click here to read Part 1

Our approach is designed to reduce MSH3 expression in specific neuronal populations. Preclinical studies to date have shown potent HD target-specific reductions with minimal virus levels or MSH3 reduction detected outside the CNS. These studies have also not identified evidence of oncogenic activation within the observation periods evaluated, although those findings cannot definitively rule out longer-term risks.

We recognize that a potentially durable gene therapy requires appropriate long-term monitoring. Our clinical development strategy will incorporate extended safety follow-up consistent with regulatory expectations for AAV therapies, including surveillance for delayed adverse effects.

Several groups are pursuing MSH3 with antisense oligonucleotides and siRNA. What does an AAV approach offer that those do not? Does AAV have any drawbacks where oligos outperform?

Ghoroghchian: We believe the primary advantage of our AAV-based approach is in its potential to deliver durable MSH3 suppression to HD target cells both in the deep brain and in the cerebral cortex, following a single administration. Traditional antisense oligonucleotides and preclinical studies of emergent siRNA approaches support their preferential targeting of cerebral cortex over the deep brain. They also require continued invasive administration (lumbar puncture multiple times per year) for life.

Latus Bio released its Huntington's disease model and webapp on GitHub.

Find it here.

LTS-201 incorporates our engineered AAV-DB-3 capsid, designed to efficiently target medium spiny neurons and layer 5/6 cortical projection neurons that have been implicated in Huntington's disease pathogenesis. This targeted delivery is particularly important because the degree of MSH3 reduction within affected neurons appears to be a critical determinant of therapeutic benefit.

Each modality has advantages and limitations. Antisense oligonucleotides and siRNA offer greater flexibility to adjust or discontinue treatment, whereas AAV therapies present additional considerations around immunogenicity, dose-related toxicity, and the difficulty of reversing or redosing treatment. While LTS-201 has been designed for low-dose, one-time administration directly into the CNS, to offer high HD-target cell activity while aiming to limit systemic immunogenicity and toxicity, antisense oligonucleotides, siRNA, and small molecules distribute throughout the body, raising potential long-term systemic genotoxicity concerns from MSH3 reduction in off-target tissues.

Ultimately, we believe different modalities — alone and in combination — may have important roles in treating Huntington's disease. Our goal is to develop a therapy capable of delivering deep, durable target engagement in the neurons most affected by the disease.

The dose-range study used two animals per group. What does the larger toxicology package look like on the path toward an IND?

Ghoroghchian: The study was designed to establish proof of mechanism, to characterize neuronal targeting, and to understand the relationship between dose and MSH3 reduction. We recognize that a small exploratory study cannot establish the tolerability profile required for clinical development.

Our broader IND-enabling program has evaluated tolerability, biodistribution, dose-related effects, and potential toxicities through appropriately designed studies, including longer-term assessments relevant to AAV gene therapy; we will provide a complete update of this work at a future date.

The current published findings provide important evidence of biological activity, but they represent one component of the overall development package. We will work with regulators to ensure that the necessary nonclinical studies support the proposed clinical dose and trial design.

Bulk tissue showed MSH3 dropped by a range of 24% to 39% at the high dose, but inside the target neurons you measured 94% reduction, and you attribute the gap to dilution. Does that explanation sufficiently support a number carrying so much weight in the model?

Ghoroghchian: The distinction is between measuring bulk MSH3 expression across an entire tissue sample, which contains neurons but also a multitude of non-neuronal cell types, and measuring it within neurons vulnerable to HD, which our therapy is designed to specifically target.

In bulk tissue, MSH3 levels are the aggregate of multiple cell types, including cells that are not relevant to HD and that may not be efficiently transduced by our novel AAV capsid variant. As a result, bulk measurements can substantially underestimate the degree of MSH3 suppression within the relevant neuronal populations for HD. Our novel  capsid has a marked preference for transducing neurons compared with astrocytes.

We therefore used cell-specific RNAscope measurements to quantify MSH3 reduction in medium spiny neurons, the HD-target neurons whose loss leads to disease onset and progression. At the highest dose of LTS-201, we observed 94% MSH3 mRNA reduction in this vulnerable cell population (as assessed in more than 90% of the cells assayed).

Importantly, we did not rely solely on that molecular measurement. We also demonstrated dose-dependent reductions in somatic instability in an HD mouse model, providing independent evidence of the intended biological mechanism. Nevertheless, the relationship between these measurements and clinical efficacy remains to be validated in patients.

AAV durability in the human brain is still an open question. How does the model address it?

Ghoroghchian: The model allows researchers to explore how different levels of MSH3 reduction and therapeutic cell coverage influence predicted disease outcomes. However, the published treatment simulations assume sustained target engagement and do not establish how long MSH3 suppression will persist in human patients.

One potential advantage of AAV gene therapy is that neurons are nondividing cells, with longitudinal studies from first-generation therapies supporting long-term, and potentially lifelong, expression. Nevertheless, durability for any experimental agent must be demonstrated clinically. We will need to assess the persistence of therapeutic activity over time and determine whether sustained MSH3 suppression translates into anticipated clinical benefits.

On the subject of time, HD moves slowly, and your framework sets a five-year delay for minimum meaningful effect. In a trial, how would you demonstrate that within a reasonable follow-up period?

Ghoroghchian: The published five-year threshold represents a modeled delay in disease milestones that can be considered clinically meaningful. It does not mean that a clinical trial would necessarily need to follow patients for five years before detecting any effect. Importantly, the model has a multitude of predicted outputs, including time to biomarker changes, advancement through each of the Huntington's Disease Integrated Staging System (HD-ISS) stages, changes in the composite Unified Huntington's Disease Rating Scale (cUHDRS), as well as in total functional capacity (TFC). For clinical trial development, cUHDRS and TFC are established endpoints that have been accepted by the EMA and FDA, respectively. The model output can, thus, be used to potentially inform selection of patient populations that could show early biomarker changes. That might also demonstrate statistical differences between treated and untreated natural history patients at various timepoints. These analyses could prospectively aid in both patient selection and appropriate powering of clinical studies for different endpoints.    

Our clinical development approach would seek to combine appropriate clinical assessments with biomarkers that may provide earlier evidence of biological activity and disease modification. Potential measures include changes in cerebrospinal fluid (CSF) neurofilament light (NfL) levels, imaging-based assessments of neurodegeneration, somatic CAG repeat expansion, and validated clinical outcomes such as cUHDRS and TFC.

We recognize that establishing a meaningful treatment effect in a slowly progressing disease presents challenges. That's precisely why quantitative models that connect target engagement, biomarkers of disease activity, and longer-term clinical outcomes may be valuable in informing trial design, patient selection, and the interpretation of emerging clinical data.

Any proposed surrogate or intermediate endpoint would, of course, require appropriate validation and regulatory alignment.

About The Expert:

P. Peter Ghoroghchian, MD, Ph.D., is the CEO of Latus Bio. Previous roles include cofounder, president, and CEO of Ceptur Therapeutics; senior vice president and head of therapeutic development at Repertoire Immune Medicines, a Flagship Pioneering company; chief technology officer at Poseida Therapeutics; and cofounder and chief scientific officer at Vindico NanoBioTechnology. He received his MD and Ph.D. in bioengineering from the University of Pennsylvania.