Can a stem cell-laden rigid-flexible bioscaffold promote recovery after spinal cord injury?

Three cues combined in one implant

The paper's answer to the repair problem is a single implant that supplies three different kinds of signal at once. The rigid part is a multi-channel conduit built from aligned electrospun polycaprolactone (PCL) nanofibers: three fiber tubes are assembled in a triangular arrangement and wrapped in an outer fiber membrane, so each microchannel presents fibers running lengthwise. The flexible part is a chitosan-based self-healing hydrogel that fills the lumen and carries bone marrow-derived mesenchymal stem cells (BMSCs). The biochemical part is neurotrophin-3 (NT3), a growth factor encapsulated in type I collagen particles that are deposited onto the fibers in a density gradient by masked coaxial electrospraying.

Each component is meant to do a distinct job. The aligned fibers give axons a physical path to follow; the collagen particles protect NT3 and release it gradually instead of in a burst; the hydrogel provides a soft, cell-friendly matrix that can be injected and then re-form. The paper reports that the hydrogel gelled within 20 minutes at room temperature and rejoined into an intact structure within 5 minutes after being cut into segments, and that its storage modulus was 170.15 ± 7.57 Pa, which the authors describe as close to the stiffness of native spinal cord tissue.

A gradient keeps NT3 available for weeks

A growth factor delivered all at once is quickly cleared, so the design puts NT3 inside collagen particles rather than free in solution. In a release test, the Poly_NT3 conduit was immersed in 1 mL of PBS that was replaced at 1, 3, 5, 7, 10 and 14 days, and NT3 in the retrieved solution was measured by ELISA. NT3 was released continuously across days 0 to 14, and the curve had not reached a clear plateau by day 14, which the authors interpret as sustained-release capacity.

The gradient adds direction to that release. Collagen particle density rises along the fiber axis, and in dorsal root ganglion (DRG) explants cultured for 5 days, axons growing toward the higher-density region reached nearly 600 μm, while those on the lower-density side averaged above 400 μm. Axons on uniaxially aligned fibers without the gradient also extended predominantly along the fiber direction, forming polarized outgrowth. These are in vitro measurements of neurite length, not evidence of regenerated circuitry in an animal, but they show that alignment and the NT3 gradient acted together rather than as separate features.

Soft gels keep BMSCs alive and neuron-prone

Because the hydrogel is the cells' immediate environment, the authors tuned its stiffness and tested how BMSCs responded. Four formulations spanned 170.15 ± 7.57 Pa (C2P1), 266.89 ± 27.64 Pa (C3P1), 680.81 ± 56.06 Pa (C2P2) and 1240.21 ± 84.07 Pa (C3P2). Encapsulated BMSCs distributed evenly in three dimensions, kept a spherical shape and showed minimal apoptosis by live/dead staining; both hydrogel components supported viability above 80% in CCK-8 assays.

The differentiation readout is the more decision-relevant result. After neuronal induction, cells in the Poly_NT3_gel group showed the highest percentage of Tuj-1-positive cells and the lowest percentage of GFAP-positive cells among all groups tested, which included hydrogel alone, Poly, Poly_NT3 and Poly_gel. Tuj-1 marks neuronal commitment and GFAP marks astrocytes, so the comparison is about which fate dominated, not about how many cells survived. The authors also describe a time-dependent shift: cells started as spherical clusters in the soft gel and later elongated as the hydrogel degraded and they attached to the rigid fibers.

Five groups in a complete-resection model

The in vivo test used male Sprague-Dawley rats at 6 weeks of age. After a dorsal laminectomy at the 9th thoracic vertebra, a 2 mm segment of the T9 spinal cord was completely resected and replaced with the assigned material. Rats were randomly assigned to five groups: Sham (laminectomy only, no resection), Ctrl (resection with no implant), Poly (PCL tubes only), Poly_NT3 (NT3-loaded PCL tubes) and Poly_NT3_gel (NT3-loaded tubes plus BMSC-laden hydrogel at 2 × 10^7 cells/mL). Animals were housed for 6 weeks before sacrifice, and surviving numbers used in the statistical analyses were 7 for Ctrl, 8 for Poly, 10 for Poly_NT3 and 7 for Poly_NT3_gel.

This design matters for reading the results. The comparison that isolates the full combination is Poly_NT3_gel against Poly_NT3 and Poly, because those groups share the conduit and, in the NT3 group, the growth factor. The Ctrl group shows what happens with resection alone. Because the model is a complete resection rather than a contusion, the findings describe a gap-bridging situation, and the group sizes are small enough that the reported differences should be read as directions of effect supported by the accompanying histological and sequencing data.

Smaller lesions and better walking in treated rats

The paper reports that Poly_NT3_gel reduced lesion volume and fibrosis and improved neural bridging, locomotor function and urinary recovery relative to the comparison groups. Urinary recovery was tracked as a practical outcome: two experienced researchers performed assisted urination twice daily until rats regained autonomous urination, and the interval from surgery to autonomous urination was recorded in days. Locomotor function was assessed in the same six-week window.

These outcomes answer different questions and should not be merged. Lesion volume and fibrosis describe tissue state; neural bridging describes continuity across the resection; locomotor scores describe walking behavior; urinary recovery describes return of bladder control. A scaffold can change one without changing another, and the paper's claim is that the combined implant moved several of them in the same favorable direction. The authors attribute the reduced fibrosis in part to the aligned fibers spatially limiting random fibroblast infiltration and to NT3 acting through TGF-β signaling to inhibit fibroblast proliferation, but that mechanism is their interpretation of the histology and sequencing rather than a directly isolated causal test.

Sequencing shows a less fibrotic cell environment

Single-nucleus RNA sequencing (snRNA-seq) was used to compare cell composition and transcriptional profiles across Sham, Ctrl, Poly_NT3_gel, Poly_NT3 and Poly groups. The analysis profiled neurons, fibroblasts, oligodendrocyte progenitor cells, oligodendrocytes and microglia, and reported that the bioscaffold helped restore transcriptional profiles toward the Sham pattern. It also described a microenvironment that suppressed post-injury fibroblast proliferation while modulating immune responses, including shifts in microglial subpopulations and pseudotime trajectories for fibroblasts.

This layer of evidence is what separates the paper from a purely behavioral report: it offers a cell-level account of why a less fibrotic, more neuron-supportive environment might accompany the functional gains. The limit is that snRNA-seq describes associations between the implant and cell states at the end point; it does not by itself prove that a specific cell population caused the walking improvement. The authors' causal chain runs from material design to cell behavior to tissue repair, and the sequencing supports the middle link rather than closing the whole chain.

A preclinical result, not a clinical protocol

The most important boundary is the model itself. The data come from rats with a complete 2 mm resection at T9, followed for 6 weeks, with group sizes between 7 and 10 animals. Human spinal cord injuries are typically contusions or lacerations with preserved tissue, different geometry and a much longer recovery horizon, so the paper does not establish that the same implant would work in those settings. The hydrogel is also reported to persist in vivo for up to about one week, meaning the cell-delivery phase is short and the later response depends on cells attaching to the rigid fibers as the gel degrades.

Within those limits, the paper's specific priority is the combination rather than any single ingredient. The comparison that carries the argument is Poly_NT3_gel against Poly_NT3 and Poly: adding the BMSC-laden self-healing hydrogel to the NT3-loaded multi-channel conduit is what the authors associate with the highest neuronal differentiation in vitro and with the coordinated tissue, locomotor and urinary improvements in vivo. A reader weighing this work should therefore focus on whether that added cellular and mechanical component, not the conduit or NT3 alone, is what shifts the repair outcome.

Sources used in this answer