PepGen's DM1 Data: Splicing Correction and the Preclinical Questions Ahead

Myotonic dystrophy type 1 is the most common adult-onset muscular dystrophy, with no approved disease-modifying therapy. A central disease mechanism involves an expanded CTG repeat in the 3'untranslated region (3'UTR) of the DMPK gene. This expansion produces RNA containing expanded CUG repeats, which accumulates in nuclear foci and sequesters muscleblind-like (MBNL) proteins, disrupting normal RNA splicing. These abnormalities contribute to clinical manifestations including myotonia, muscle weakness, cardiac conduction defects, and metabolic dysfunction ^[1]. One therapeutic strategy is to reduce the toxic effects of expanded CUG RNA and restore the availability of MBNL proteins for normal splicing.

Figure 1. Myotonic dystrophy type 1 disease mechanisms ^[1].

The clinical development landscape has also highlighted the challenges of translating molecular activity into meaningful functional outcomes.. In September 2026, Novartis reported that del-desiran, the lead asset in its roughly $12B acquisition of Avidity Biosciences, missed its primary endpoint in the Phase 3 HARBOR study ^[2]. Although secondary and exploratory analyses showed evidence of clinical activity, the result underscored the importance of demonstrating that molecular changes translate into meaningful and durable functional benefits.

Splicing correction in the PGN-EDODM1 trial

Then, on November 10, PepGen reported data that reset expectations. Six participants received four monthly intravenous doses of PGN-EDODM1 at 10 mg/kg. Mean splicing correction in muscle reached 46.5% at Week 13 on a 22-gene panel, against 6.8% in a pooled placebo group, and four of the six cleared 50%. Several functional measures also showed directional improvements at Week 16. Video hand opening time improved 3.9 seconds at Week 16 versus 0.2 on placebo, with directional gains in grip strength, Timed Up and Go, and the 10-Meter Walk/Run Test. All TEAEs were mild or moderate, although four mild renal TEAEs and transient albuminuria were observed. Patient-reported outcome measures remained inconclusive at Week 16, highlighting the need for further evaluation in larger cohorts and longer-term follow-up ^[3].

Other approaches are also advancing in DM1. Dyne's z-basivarsen (DYNE-101), which uses a TfR1-targeting Fab to deliver an antisense oligonucleotide, previously demonstrated splicing correction in the ACHIEVE trial and is now being evaluated in the Phase 3 HARMONIA study^[4]. Still, on the same 22-gene readout, PepGen's distribution sits furthest right.

Figure 2. Highest reported splicing correction in a multidose study and improvement in mobility, myotonia, and strength, with favorable safety profile observed at 10 mg/kg ^[5].

One detail of the trial stands out. FREEDOM2 enrolls in Canada, the UK, South Korea, Australia and New Zealand. To understand why the United States is missing from that list, it helps to look back at PepGen's history with the FDA.

Two FDA clinical holds in PepGen's DM1 program

The first hold came in May 2023, when the FDA placed the Phase 1 FREEDOM IND on hold for reasons never disclosed. The agency lifted it that October ^[6].

The second stop is the one still in force. In March 2026 a partial clinical hold landed on FREEDOM2 itself, based in the company's own 10-Q language on "previously submitted preclinical pharmacology and toxicology studies."

The specific nonclinical findings have not been fully disclosed in public regulatory communications. However, analyst reports cited by Fierce Biotech, based on discussions with PepGen management, described decreases in blood pressure and body temperature in a subchronic mouse study. The same reports indicated that comparable findings had not been observed in non-human primates or early clinical data. PepGen reportedly suggested that differences in infusion rates and resulting peak drug exposure may have contributed to the mouse findings ^[7].

Infusion rate is one possible explanation, and not the only one. A conjugate effect that primates and a small Phase 1 cohort were not powered to detect is another. What the mouse cannot do is adjudicate between them: carrying no human target, it offers no on-target arm against which to read the signal.

Eight months later the hold stands, and on the November earnings call CEO James McArthur declined to offer a timeline. The remedy under way is not a clinical dataset or a primate study; it is, per the 10-Q, "additional nonclinical mouse studies designed to generate data to respond to FDA feedback."

What can those additional studies tell PepGen about the signal it saw in mice?

How Avidity, Dyne and PepGen deliver oligonucleotides to muscle

Part of the answer depends on how each drug enters muscle cells and what it binds once inside. On both counts, Avidity and Dyne share a strategy, while PepGen takes a different approach.

Novartis (AOC Platform)

Avidity's AOC platform, now part of Novartis, uses a monoclonal antibody to deliver a small interfering RNA (siRNA). For del-desiran, the antibody binds transferrin receptor 1 (TfR1) on muscle cells, triggering receptor-mediated uptake. Once inside, the siRNA guides the cell's RNA interference machinery to degrade DMPK RNA, reducing the pool of transcripts that trap the splicing factor MBNL1 ^[8].

Dyne (FORCE Platform)

Z-basivarsen uses the same TfR1-mediated route, with an antibody fragment carrying an antisense oligonucleotide (ASO). The ASO binds DMPK RNA in the nucleus and recruits RNase H1, an enzyme that cuts the RNA ^[9]. For both AOC and FORCE, the choice of mouse model therefore depends on whether the antibody or fragment recognizes the mouse receptor and whether the payload recognizes the target RNA.

Figure 3. Dyne's FORCE platform ^[9].

PGN-EDODM1 uses a cell-penetrating peptide to carry a phosphorodiamidate morpholino oligomer (PMO) into muscle cells and help it escape the endosome. The PMO binds directly to expanded CUG repeats in DMPK RNA, displacing MBNL1 and freeing it to regulate splicing while leaving the transcript intact ^[5].

Figure 4. PGN-EDODM1 is designed to target pathogenic DMPK RNA to restore splicing using EDO technology to deliver to a broad range of tissues ^[5].

For PepGen, the relevant question is whether the mouse model expresses repeat-containing RNA that the PMO can bind. This matters when assessing whether target engagement could have contributed to the blood pressure and temperature changes that prompted the second clinical hold.

What PepGen's mouse safety findings mean for preclinical study design

Three features of PepGen's experience matter for other teams planning preclinical studies of muscle-directed oligonucleotides.

• A pharmacologically relevant target is needed to interpret on-target effects. When a sequence-directed oligonucleotide does not recognize the endogenous target in a conventional animal species, that model may still provide useful information about target-independent effects, including exposure-related toxicity, formulation effects, and immune responses. However, it cannot fully characterize pharmacological activity or potential toxicity resulting from interaction with the intended human target. A target-compatible model or an appropriate surrogate approach may therefore be needed to complement conventional nonclinical studies.
• Exposure conditions can influence nonclinical safety findings. Bolus-rate-dependent cardiovascular and thermoregulatory responses in mice are well described across oligonucleotide and peptide-conjugate chemistries, and scale poorly. A design without a matched slow-infusion arm invites a finding the sponsor will spend years unwinding.
• Nonclinical findings remain relevant throughout clinical development. The mouse data at issue reached the agency in 2024 and passed without objection through a Phase 1 dosing up to 20 mg/kg. They resurfaced in 2026 and stopped a Phase 2. A dossier stays live for the life of a program, which argues for building the definitive model before the IND rather than after a hold.

Together they point at one design decision: for a muscle-directed oligonucleotide, the nonclinical species should carry the human target sequence. Otherwise safety and efficacy are read in two different animals, and only one is in the regulatory file.

Target engagement in humanized DMPK mice

A humanized model can help researchers test whether an oligonucleotide engages its intended RNA target. Cyagen's huDMPK mouse (C001882) carries the human DMPK gene and its flanking sequences. For PepGen's PGN-EDODM1, the target is an expanded CUG repeat in DMPK RNA. The current model has no engineered repeat expansion, but Cyagen can assess the feasibility of introducing one. Compared with expanding the repeat in the native mouse Dmpk gene, placing it within human DMPK would preserve the human RNA sequence surrounding the repeat. This could reduce one source of uncertainty when the FDA evaluates how well mouse target-engagement data apply to humans.

If expression of the expanded RNA were confirmed in a customized model, researchers could assess target engagement alongside changes in blood pressure and temperature. Varying the infusion rate or repeating the dose could help characterize the safety signal. Linking that signal to target engagement would require appropriate controls, and assessing splicing correction would require a measurable baseline splicing defect. For the TfR1-directed approaches described above, Cyagen also offers a huTFRC/huDMPK dual-humanized model.

Figure 5. Gene Expression Detection in Heart, Tibialis Anterior (TA), Liver, Cerebral Cortex, and Hippocampus Tissues of huDMPK Mice and Wild-Type (WT) Mice (8 weeks old, homozygous, n=3).

Although recent clinical findings have demonstrated measurable splicing correction, the relationship between molecular improvement, durable functional benefit, and long-term safety remains to be established. These questions highlight the importance of selecting preclinical models that can provide relevant evidence on target engagement, tissue distribution, and treatment-related effects.

Choose a model with a verified target sequence and a measurable response that addresses the specific preclinical question. Cyagen's neurology CRO services can help you select or develop a mouse model and choose molecular or functional assessments suited to your research question.

👉 [Explore the full validation data for huDMPK mouse model]

👉 [Discuss Your DMPK-Targeted Preclinical Study with Our Scientists]

Reference

[1] Rahm L, Hale MA, Raaijmakers RHL, Marrero Quiñones A, Patki T, Johnson NE, van Bokhoven H, Mul K. Myotonic dystrophy type 1: clinical diversity, molecular insights and therapeutic perspectives. Nat Rev Neurol. 2025 Nov;21(11):623-641. doi: 10.1038/s41582-025-01139-x. Epub 2025 Sep 22. PMID: 40983775.

[2] Novartis. Novartis provides update on delpacibart etedesiran (del-desiran) Phase III HARBOR study for the treatment of myotonic dystrophy type 1 (DM1) [Internet]. Basel (Switzerland): Novartis; 2026 Sep 8 [cited 2026 Nov 12].

[3] PepGen Inc. PepGen announces positive Phase 2 FREEDOM2-DM1 10 mg/kg results demonstrating highest reported splicing correction and improvements across multiple functional measures [Internet]. Boston (MA): PepGen Inc.; 2026 Nov 10 [cited 2026 Nov 12].

[4] Dyne Therapeutics, Inc. Dyne Therapeutics announces additional one-year clinical data from Phase 1/2 ACHIEVE trial of z-basivarsen (DYNE-101) for myotonic dystrophy type 1 (DM1) [Internet]. Waltham (MA): Dyne Therapeutics, Inc.; 2026 Sep 29 [cited 2026 Nov 12].

[5] PepGen Inc. FREEDOM2-DM1 10 mg/kg MAD cohort data update [Internet]. Boston (MA): PepGen Inc.; 2026 Nov 10 [cited 2026 Nov 12]. 26 slides.

[6] PepGen Inc. PepGen Inc. announces FDA has lifted the clinical hold on its investigational new drug application for FREEDOM-DM1 Phase 1 study of PGN-EDODM1 for myotonic dystrophy type 1 (DM1) [Internet]. Boston (MA): PepGen Inc.; 2023 Oct 12 [cited 2026 Nov 12].

[7] Taylor NP. PepGen hit with partial clinical hold, analyst blames ‘understaffed FDA'. Fierce Biotech [Internet]. 2026 Mar 5 [cited 2026 Nov 12].

[8] Avidity Biosciences. Platform overview [Internet]. San Diego (CA): Avidity Biosciences [cited 2026 Nov 12].

[9] Dyne Therapeutics, Inc. Our FORCE™ platform [Internet]. Waltham (MA): Dyne Therapeutics, Inc. [cited 2026 Nov 12].