Patient-Customized Oligonucleotide Therapy for a Rare Genetic Disease

Abstract

Genome sequencing is often pivotal in the diagnosis of rare diseases, but many of these conditions lack specific treatments. We describe how molecular diagnosis of a rare, fatal neurodegenerative condition led to the rational design, testing, and manufacture of milasen, a splice-modulating antisense oligonucleotide drug tailored to a particular patient. Proof-of-concept experiments in cell lines from the patient served as the basis for launching an "N-of-1" study of milasen within 1 year after first contact with the patient. There were no serious adverse events, and treatment was associated with objective reduction in seizures (determined by electroencephalography and parental reporting). This study offers a possible template for the rapid development of patient-customized treatments. (Funded by Mila's Miracle Foundation and others.).

Figure 1.. Study Timeline and Genetic Diagnosis.

Panel A shows the timeline from the initial clinical diagnosis to the initiation of treatment. ASO denotes antisense oligonucleotide. Panel B shows an IGV (Integrative Genomics Viewer) image of the patient’s whole-genome sequencing (WGS) read alignments near MFSD8 intron 6. Characters in the mapped sequencing reads indicate mismatched or unaligned (i.e., soft-clipped) bases. Vertical dashed lines indicate chimeric read breakpoints. (Additional details are provided in Fig. S2 in the Supplementary Appendix, available with the full text of this article at NEJM. org.) Panel C shows splicing and translational effects of the SVA (SINE–VNTR–Alu) insertion in MFSD8. The abbreviation “i6” indicates the upstream region of the SVA insertion site in intron 6 that is misspliced with exon 6 as a result of the activation of the i6.SA cryptic splice site by the SVA insertion. (See Fig. S5 for additional details.)

Figure 2.. Antisense Oligonucleotide Drug Development.

Panel A shows the location and chemistry of the ASOs that were designed to block the i6.SA splice acceptor site or exonic splice enhancer (ESE) elements. (Additional details are provided in Table S1.) The ESE elements were predicted with RESCUE-ESE and ESEfinder.^, 2′-MOE denotes 2′-O-methoxyethyl, and 2′-OMe 2′-O-methyl. Panel B shows the ratio of the normal exon 6–exon 7 (E6–E7) splicing to the abnormal exon 6–intron 6 (E6–i6) splicing (normalized to a no-transfection control), measured in patient fibroblasts that were transfected (for 24 hours at 100 nmol per liter) as indicated. To measure splice isoform-specific levels, multiplex reverse-transcriptase polymerase chain reactions were conducted with isoform-specific primer sets, and then the intensity of the isoform-specific bands was quantified by gel electrophoresis (Fig. S6). “Scrambled” indicates a nontargeting oligonucleotide (TY772). I bars indicate 95% confidence intervals of the means. P values were calculated by two-sided t-test. Panel C shows RNA sequencing (RNA-seq) analysis validation of the splice-correcting effect of milasen (TY777). For the calculation of the fraction of normal splicing (exon 6–exon 7), three other splicing events that are mutually exclusive with the normal splicing were considered. Splicing events supported by only one read are not shown. P values were calculated by Fisher’s exact test. Panel D shows intracellular vacuoles, visualized by electron microscopy, in control fibroblasts (MFSD8 wild-type human foreskin fibroblast; BJ cell line) and in patient fibroblasts that are either untreated or transfected with the indicated oligonucleotide. Scoring was performed on a scale of 0 to 5, with 0 representing the lowest and 5 representing the highest level of vacuole accumulation.

Figure 3.. N-of-1 Clinical Study.

Panel A shows the dosing schedule. (Additional details are provided in Fig. S14A.) Panel B shows the concentration of milasen in cerebrospinal fluid (CSF) before each administration (trough). An additional measurement of the concentration in CSF was obtained at day 174 (without concurrent dose administration). I bars indicate the minimum and maximum values of duplicate measurements. Trough levels rose steadily in a dose-proportional fashion until day 40, at which point they dropped to 1.7 ng per milliliter and then resumed their rise with repeated dosing up to a plateau of 18 to 27 ng per milliliter. The dip at day 40 may have been due to a CSF leak, given its coincident timing with a post–lumbar puncture headache after the previous dose. A similar plateauing of CSF trough levels was observed in a previous study of intrathecally delivered nusinersen (9 to 11 ng per milliliter after four repeated doses of 12 mg). Panel C shows the trends in seizure frequency and duration as reported in a seizure diary recorded by the parents. Seizures were all of the same type: sudden startle followed by uncontrollable, untriggered laughter that was different from the patient’s natural laugh, at times accompanied by an increase in the nonspecific repetitive hand movements she had at baseline. Panel D shows the trends in seizure activity as detected by electroencephalography (EEG). In a comparison of the means of the initial two recordings and the subsequent three recordings, the daily seizure count, seizure duration, and percent cumulative time spent in seizure decreased by 63% (from 31.5 to 11.7 per day), 52% (from 108 seconds to 52 seconds), and 85% (from 3.9% to 0.6%), respectively.