3D-Printed Scaffolds Loaded With Stem Cells Restore Function After Spinal Cord Transection

Spinal cord injury (SCI) affects an estimated 302,000 people in the United States alone, and no restorative treatments currently exist. The central challenge in SCI repair is re-establishing neural circuitry across the lesion gap — a goal that cell transplantation alone struggles to achieve due to poor structural support, limited cell survival, and insufficient directional axonal guidance. This study from the University of Minnesota directly addresses these barriers by combining 3D bioprinting, human iPSC-derived regionally specific spinal neural progenitor cells (sNPCs), and organoid biology into a single implantable construct.

The team used an extrusion-based multi-material 3D printer to fabricate silicone scaffolds measuring approximately 1.6 mm wide, 0.65 mm tall, and 2 mm long (for transplantation), each containing three parallel microscale channels approximately 200 µm wide and 440 µm tall. A Pluronic hydrogel sacrificial layer was printed first to allow clean scaffold detachment from the glass substrate without damage. A cell-laden bioink — composed of human iPSC-derived sNPCs, Matrigel, and neural media with growth factors — was dispensed at 4°C into the channels using point dispensing after the silicone was cured for at least 5 hours. Two assembled scaffolds were transplanted into a 1.8 mm gap created by 2 mm thoracic transection injuries in rats.

In vitro characterization revealed highly ordered neuronal development within the scaffold channels over time. By day 15, SMI312-positive axonal projections were already guided along the channel length. By day 30, axonal networks extended over the tops of the channels. At 40 days, immunostaining confirmed the presence of regionally specific ventral interneuron subtypes: FOXP2-positive V1 neurons, Chx10-positive V2a neurons, and Evx1-positive V0 neurons. At 140 days, MAP2-positive neurons continued to grow along the scaffold. Quantification at 170 days showed a mixed cellular composition including MAP2-positive neurons, APC-positive oligodendrocytes, and GFAP-positive astrocytes. Remarkably, scaffolds maintained neuronal identity — confirmed by MAP2 and SMI312 co-expression — for at least 365 days post-printing, demonstrating extraordinary long-term stability.

In vivo, the transplanted organoid scaffolds produced significant functional recovery in the rat transection model. At 12 weeks post-transplantation, the majority of cells within the scaffolds had differentiated into neurons and integrated with host spinal cord tissue. Critically, the transplanted cells formed synaptic connections in both rostral and caudal directions relative to the scaffold, consistent with establishing a functional neural relay across the lesion. This bidirectional integration is mechanistically important — it suggests the organoid scaffold can serve as a bridge reconnecting descending motor and ascending sensory pathways interrupted by injury.

The study's implications extend beyond SCI. The approach demonstrates that combining 3D printing precision, organoid biology, and clinically relevant human iPSC-derived cells can produce complex neural constructs with defined architecture, regional identity, and long-term viability. The use of silicone — a non-degradable, biocompatible, gas-permeable material with established medical-grade use — avoids confounds introduced by scaffold degradation and supports oxygen delivery to metabolically demanding neural cells. While this is currently a rat model proof-of-concept, the scalability of the iPSC platform and 3D printing methodology makes clinical translation a realistic longer-term goal. Key remaining challenges include scaling scaffold dimensions to human lesion sizes, confirming immune compatibility in non-immunosuppressed large-animal models, and establishing long-term safety of non-degradable implants.