The argument for CRISPR as a modality for rare CNS diseases is not speculative. The genetic targets are identified, the causal variants are characterized, the in vitro editing evidence is substantial, and in some cases the target cells can be edited at useful frequencies with non-viral delivery in culture. What is missing — across nearly every program in this space — is a CNS delivery mechanism that can reach those cells after systemic administration.
This article walks through three CNS rare disease targets where CRISPR editing has genuine therapeutic rationale: Huntington's disease (HTT), Friedreich's ataxia (FXN), and CLN3 Batten disease. For each, the genetic target is well-defined. For each, the delivery problem is the primary obstacle between compelling preclinical data and a credible development path.
Huntington's Disease: The Case for HTT Silencing
Huntington's disease (HD) is caused by a CAG trinucleotide repeat expansion in exon 1 of the HTT gene. Allele counts above approximately 36 CAG repeats are associated with disease; counts above 40 are fully penetrant. The mutant HTT protein (mHTT) contains a polyglutamine expansion that leads to misfolding, protein aggregation, and progressive striatal and cortical neuronal death. The disease follows an autosomal dominant pattern, meaning one mutant allele is sufficient for disease progression — and meaning that allele-selective targeting strategies are at least in principle possible.
The genetic target is unusually tractable for CRISPR approaches. Two broad strategies are under preclinical investigation: selective knockout of the mutant allele using a guide RNA that targets a single-nucleotide polymorphism (SNP) near the expanded repeat region, or non-allele-selective HTT silencing on the premise that reducing total HTT expression is neurologically tolerable. Both approaches have produced in vitro knockdown evidence in patient-derived iPSC-derived neurons, and in vivo HTT reduction has been demonstrated in mouse and rat HD models after intracranial or intrathecal delivery of LNPs or AAV vectors.
The delivery gap is stark. Clinical antisense oligonucleotide (ASO) programs targeting HTT have required intrathecal infusion — repeated dosing via lumbar puncture — because systemic delivery of ASOs does not achieve therapeutically relevant CNS concentrations. CRISPR editing strategies face the same anatomical barrier but with additional complexity: the CRISPR cargo is substantially larger than an ASO (Cas9 mRNA alone is ~4.5 kb; with sgRNA and packaging, delivery vehicles must accommodate substantially more molecular weight) and is not amenable to naked delivery approaches that work for smaller oligonucleotides. The appeal of a single systemic injection that achieves durable HTT reduction in striatal neurons is obvious, but no non-viral delivery system has demonstrated this in any published large animal or primate study to date.
What makes HD particularly compelling for an LNP delivery approach is the disease anatomy. Striatal medium spiny neurons are the primary cell population that degenerates early in HD, and the striatum is anatomically accessible from the systemic circulation via the basal ganglia vasculature. A delivery vehicle that crosses the BBB and shows preferential uptake in the striatum — rather than diffuse whole-brain distribution — would be scientifically interesting and potentially clinically relevant. Whether CNS-targeted LNPs can achieve sub-regional CNS specificity, rather than broad parenchymal distribution, is a question that existing surface engineering work has not yet answered in a rigorous way.
Friedreich's Ataxia: Restoring Frataxin Through Epigenetic Editing
Friedreich's ataxia (FA) is caused by a GAA trinucleotide repeat expansion in the first intron of the FXN gene, which encodes frataxin, a mitochondrial protein essential for iron-sulfur cluster assembly. The repeat expansion — typically 66–1000 GAA repeats on both alleles — causes epigenetic silencing of FXN through heterochromatin formation and DNA methylation changes at the expanded locus, reducing frataxin expression to 4–29% of normal levels. This reduction causes progressive neurodegeneration affecting the dorsal root ganglia, spinocerebellar tracts, and corticospinal neurons, alongside hypertrophic cardiomyopathy.
The therapeutic target is gene reactivation rather than knockout. This positions FA as a particularly interesting case for CRISPR-based epigenetic editing tools — specifically, dCas9-TET3 methylation erasers or CRISPR activation (CRISPRa) approaches that use transcriptional activators fused to catalytically dead Cas9 to upregulate FXN expression by reversing the silencing at the GAA repeat locus without cutting the genome. In vitro evidence for frataxin reactivation through epigenetic editing in patient-derived FA neurons has been published by multiple groups, with frataxin protein restoration to near-normal levels achievable in culture systems using lentiviral or plasmid delivery. The transient expression profile of mRNA-delivered epigenetic editors makes LNPs — rather than viral vectors — conceptually attractive for this application, since sustained dCas9 expression is not needed once the epigenetic mark is erased.
The CNS delivery dimension is particularly challenging for FA because the primary neurodegeneration occurs in the dorsal root ganglia (DRG) — peripheral sensory neurons that reside in the vertebral foramina — in addition to spinocerebellar pathways within the CNS. DRG neurons sit outside the strict BBB but are not easily accessed by systemic nanoparticle delivery either. They are vascularized by fenestrated capillaries, but the tissue architecture restricts LNP access and the ganglia are not anatomically contiguous with the major CNS ventricular compartments. The cardiomyopathy component of FA is actually more accessible to systemic LNP delivery, but the neurological disease dominates the clinical disability score and defines the therapeutic target for most programs.
For a CNS-targeting LNP platform, FA represents a case where the BBB-crossing mechanism matters, but so does the ability to reach DRG neurons after systemic administration — a different anatomical challenge that may require a different surface engineering strategy than pure brain parenchyma targeting via TfR1 or LRP1. It is worth being honest about this: a BBB-crossing LNP optimized for cortical neuron access may not be the right vehicle for FA, and the surface engineering requirements for DRG delivery have received less research attention than BBB-crossing approaches.
CLN3 Batten Disease: An Editing Target in a Pediatric Neurodegenerative Disorder
Neuronal ceroid lipofuscinosis type 3 (CLN3 Batten disease) is a lysosomal storage disorder caused by loss-of-function mutations in CLN3, most commonly a 1.02 kb deletion spanning exons 7–8. CLN3 encodes a transmembrane lysosomal protein whose precise molecular function is not fully established, but whose loss leads to progressive accumulation of autofluorescent lipofuscin storage material in neurons, causing vision loss (typically the first symptom, appearing around age 6–8), seizures, progressive cognitive and motor decline, and death typically in the second or third decade of life.
CLN3 Batten is a strong CRISPR candidate for several reasons. The predominant mutation (the common ~1 kb deletion) is present in roughly 85% of CLN3 alleles among affected individuals worldwide, making it a single-target editing problem rather than requiring allele-specific approaches across a heterogeneous mutation landscape. Exon-skipping strategies using CRISPR to delete flanking exon 5 sequence to produce an in-frame partial CLN3 protein with residual lysosomal function have shown preclinical efficacy in mouse models. Base editing approaches targeting specific splice-altering point mutations may also be applicable for a subset of non-deletion alleles. The disease has a well-characterized natural history and is supported by patient advocacy organizations who have driven substantial preclinical investment and biomarker development.
The delivery challenge for CLN3 is that the disease is distributed across the entire brain — not localized to a specific nucleus or circuit. Diffuse neuronal involvement means that a CNS delivery vehicle must achieve broad parenchymal distribution after crossing the BBB, rather than targeted regional delivery. This places a premium on LNP formulations that, after transcytosis across the BBB endothelium, can diffuse into the parenchyma and transduce multiple neuronal populations across cortex, hippocampus, striatum, and cerebellum rather than remaining concentrated near the vasculature.
The pediatric context also raises formulation questions that are underexplored. LNP pharmacokinetics, circulation half-life, and BBB transcytosis efficiency in juvenile rodent models — which better approximate the developmental stage of patients at diagnosis — are substantially understudied compared to adult mouse models. Pediatric brain endothelium may differ in TfR1 expression density and transcytosis competence from adult endothelium; these are empirical questions that require model development rather than extrapolation from adult data.
The Pattern Across All Three
Looking at Huntington's, Friedreich's ataxia, and CLN3 Batten together, a pattern emerges. Each disease has:
- A single causal gene with a well-characterized mutation class — CAG expansion in HTT, GAA expansion in FXN, common deletion in CLN3
- Strong preclinical in vitro editing or silencing evidence in relevant cell models, including patient-derived iPSC-derived neurons
- In vivo efficacy demonstrated in rodent models — but primarily after intracranial injection, intrathecal delivery, or convection-enhanced delivery that bypasses the BBB
- No published evidence of therapeutic CNS gene editing following intravenous administration with a non-viral delivery system in a large animal model
The fourth point is the gap. It is not a target validation problem — all three targets are well-validated by genetics, and in some cases by early clinical work with other modalities. It is not a CRISPR chemistry problem — the editing tools (Cas9 nuclease, base editors, prime editors, CRISPRa) are functionally capable of the required molecular operations in the correct cells. It is a delivery problem: the absence of a non-viral delivery vehicle that can cross the BBB after systemic injection and achieve neuronal transduction at therapeutically relevant levels in a disease-sized animal.
What the Delivery Gap Actually Means for Program Design
The delivery gap has a specific implication for how CNS rare disease CRISPR programs get structured. Programs that cannot solve systemic CNS delivery are forced into one of two positions: either accept intracranial or intrathecal delivery routes with their clinical limitations, or pause the program pending a delivery solution. Neither is a satisfying position for diseases with serious clinical burden and well-validated genetic targets.
Intrathecal delivery is a real clinical option for some CNS indications — ASO programs targeting spinal cord disease have demonstrated it is feasible and repeatable. But intrathecal delivery distributes poorly to forebrain structures, and for diseases like HD where striatal and cortical degeneration are the primary drivers of disability, intrathecal distribution of a CRISPR payload may not achieve sufficient cortical coverage. Intracranial stereotaxic delivery can target specific nuclei but is impractical for diffuse disease and essentially impossible to scale to a repeat-dosing model.
What would change the picture is a non-viral delivery system demonstrating cell-type-resolved CNS biodistribution after intravenous administration in a relevant disease model — with explicit documentation of which cell types (neurons, astrocytes, microglia, endothelial cells) took up the delivery vehicle, at what frequency, and in which brain regions. Most existing LNP CNS biodistribution data reports whole-brain tissue signal, which aggregates across all CNS cell types and cannot distinguish therapeutic neuronal delivery from endothelial cell uptake. Cell-type-resolved data is what the field needs to evaluate whether a CNS LNP platform is actually hitting the right cells.
Biopathio's Position in This Landscape
Biopathio is a delivery platform company. We do not develop gene therapies for specific CNS indications — we engineer the LNP vehicles that CNS gene therapy developers need. Our work focuses on the preclinical evidence base for BBB-crossing LNPs: receptor-mediated transcytosis mechanisms via TfR1 and LRP1, ionizable lipid pKa selection and Gal8-validated endosomal escape in neural cell models, and CRISPR cargo compatibility — including RNP, base editor mRNA, and prime editor configurations — in CNS cell models.
The disease landscape described above defines who we want to work with: teams that have a validated CNS target, a CRISPR editing strategy with in vitro evidence, and are facing the delivery problem as the primary development obstacle. The delivery problem is real and specific. It is solvable with the right formulation and characterization approach. That is the work we are built to do together with CNS gene therapy programs.
