SMA Organoids Separate SMN Restoration from Neuromuscular Functional Recovery
A neuromuscular organoid preprint compares risdiplam and branaplam across neuronal and muscle compartments. Both increased SMN protein and junction-associated structures, while branaplam produced a stronger muscle-maturation and contraction response in the principal patient model.
A Paired Neural and Muscle Model
Researchers at the Max Delbrück Center and collaborating academic institutions developed an automation-compatible neuromuscular organoid model of spinal muscular atrophy (SMA). The organoids form separate spinal neural and skeletal muscle compartments connected through neuromuscular contacts. The preprint uses that arrangement to examine how restoration of survival motor neuron protein, or SMN, affects the two lineages and their combined function.
Three SMA type 1 patient-derived iPSC lines were compared with healthy pluripotent stem cell controls. Patient organoids retained recognizable neural–muscle organization, but their neuromuscular junction-associated structures matured poorly and their muscle fibers remained smaller. Functional impairment differed by patient line; the most affected line, Pt1, showed markedly reduced spontaneous contractions.
The authors quantified postsynaptic acetylcholine-receptor clusters using α-bungarotoxin staining and normalized these measurements to myofiber number. These structures were associated with neuronal projections. Curare abolished contractions, supporting the contribution of neuromuscular transmission to the recorded movement.
Two Splicing Modulators Produce Different Downstream Responses
Risdiplam and branaplam both modify SMN2 splicing. Dose selection identified 250 nM risdiplam and 40 nM branaplam as compatible with organoid viability and structure. Both increased SMN protein across the three patient lines. Subsequent lineage-resolved analysis focused on Pt1 organoids treated from day 20 to day 30.
Single-nucleus RNA sequencing found broadly comparable cell-type composition between treatment groups. The compounds nevertheless produced different transcriptional responses within those populations. Both affected neuronal programs associated with synaptic signaling and neuronal function. Branaplam induced a broader response, including muscle genes associated with myogenic differentiation and contractile maturation, such as TNNI2 and MYL1.
The analysis distinguished genes shifting toward control expression from genes whose existing disease-associated change became stronger after treatment. That distinction matters because a larger transcriptional response can include both reversal and further enhancement of a pre-existing program. The authors connect some enhanced muscle programs with a possible compensatory response, rather than classifying every expression change as normalization.
Junction Counts and Contraction Give Different Information
Later measurements at approximately day 60 followed treatment begun at day 20. Both drugs increased junction-associated cluster numbers: roughly twofold for risdiplam and fourfold for branaplam in the reported Pt1 comparison. Risdiplam produced little improvement in contraction, whereas branaplam partly restored contractile activity and significantly increased myofiber diameter.
These findings connect tissue-level function with the muscle-compartment response. An increase in postsynaptic structures accompanied both treatments, but stronger contraction was associated with additional muscle growth and maturation. Three independent experiments with three organoids per condition supported the main reported structural and functional comparisons; individual organoids and sequenced nuclei remain different sampling levels.
A further experiment combined risdiplam with prednisolone. The combination increased junction-associated structures and myofiber diameter compared with risdiplam alone. This result provides a separate experimental route for investigating muscle-directed contributions alongside SMN restoration.
Supporting Reproducible Candidate Comparisons
The organoid workflow includes progenitor banking and robot-assisted differentiation. Comparisons with manual generation showed preserved neural–muscle organization, while automated production reduced early size variability. This production work supports the practical use of paired-lineage models for repeated drug-response experiments.
The pharmacological findings are specific to the tested concentrations, developmental stages and patient backgrounds, with the detailed drug-response analysis concentrated in Pt1. The authors use branaplam as a functional probe and note its lack of clinical viability. Within that scope, the model brings molecular response, tissue structure and integrated contraction into the same evaluation framework, making their agreement or divergence experimentally visible.