The research team, led by Qing Liu and Fei Huang, began with a deceptively simple question: how do you make a therapeutic vesicle find the right cells? Conventional exosome therapies rely on passive accumulation at injury sites, where inflamed tissue happens to attract circulating nanoparticles. That passive approach rarely delivers enough cargo to the endogenous neural stem cells scattered around a lesion to change their behavior. To solve this, the researchers turned to phage display, a combinatorial screening technique that pits billions of random peptides against a target cell and lets binding affinity do the selecting. They immobilized neural stem cells harvested from embryonic rat spinal cords and ran three successive rounds of biopanning against a random 12-mer peptide library, washing away weak binders and amplifying the survivors in bacterial hosts.

The screen delivered a clear winner. Of 45 sequenced phage clones showing strong binding, 41 carried the same peptide sequence: WTLPPQLMFANI, which the team designated TGP1. When synthesized with a fluorescent tag, TGP1 bound neural stem cell neurospheres with the strongest and most localized signal among all candidates. More importantly, after intravenous injection into rats with spinal cord injuries, the peptide accumulated in the injured spinal cord region, demonstrating that a systemically administered molecule could reach the very tissue where repair is needed. The researchers caution that phage display identifies ligands by binding behavior rather than by receptor identity, so the specific surface molecule that TGP1 recognizes remains unknown—a question they acknowledge will require future affinity-capture studies.

With the targeting peptide in hand, the team built a dual-function delivery vehicle. They used lentiviral vectors to engineer human induced pluripotent stem cells with two genetic payloads: one encoding a fusion protein that displays TGP1 on the exosomal membrane via the abundant membrane protein Lamp2b, and another that boosts production of Apelin, a signaling peptide with known pro-regenerative effects in the nervous system. These engineered cells were then differentiated into neural stem cells, and the exosomes they secreted were harvested by differential ultracentrifugation. Quality control followed the MISEV2018 guidelines of the International Society for Extracellular Vesicles: dynamic light scattering confirmed a hydrodynamic diameter of roughly 100 nanometers, transmission electron microscopy revealed the characteristic cup-shaped morphology, western blots detected the canonical markers Alix, TSG101, and CD81 while excluding endoplasmic reticulum contamination, and endotoxin levels stayed below 1 EU/mL across all preparations.

The engineering worked as designed. ELISA measurements showed that Apelin content in the engineered exosomes was approximately 2.8-fold higher than in unmodified controls. A competitive binding assay provided mechanistic proof that surface TGP1 drives targeting: when free fluorescent TGP1 peptide was mixed with TGP1-displaying exosomes, the two competed for the same binding sites on neural stem cell neurospheres in a ratio-dependent manner. In injured rats, TGP1-modified exosomes accumulated more prominently in damaged spinal cord segments than unmodified vesicles, and confocal microscopy of tissue sections revealed a 4.6-fold increase in co-localization with Nestin- and SOX2-positive endogenous neural stem cells. Co-localization with mature neurons and microglia remained limited, indicating genuine preferential engagement rather than nonspecific tissue penetration.

The functional payload proved equally consequential. In culture, exosome treatment shifted neural stem cell differentiation toward the neuronal lineage, increasing the βIII-tubulin-positive neuronal area and reducing the GFAP-positive astrocytic area. The Apelin-enriched, TGP1-targeted formulation showed the strongest effect on both counts, and also drove the highest levels of Ki67 expression and BrdU incorporation, markers of active proliferation. The real test came under inflammatory stress. The researchers co-cultured neural stem cells with mixed glial cells activated by lipopolysaccharide and interferon-gamma, a Transwell system that recapitulates the cytokine-rich, pro-gliogenic environment of an injured cord. Under these conditions, neuronal differentiation was nearly abolished, with some microscopic fields showing barely detectable βIII-tubulin. TGP1-modified exosomes partially rescued neuronal commitment, and the Apelin-enriched version restored it most effectively, producing the largest βIII-tubulin areas and the smallest GFAP areas at both three and seven days.

In vivo results followed the same pattern. Rats received a standardized compression injury at the T9-T10 vertebral level, followed by weekly tail vein injections of exosomes at 0.5 mg/kg. Eight weeks of repeated dosing produced no detectable toxicity: histology of major organs was unremarkable, and serum markers of liver and kidney function showed no treatment-related deterioration. At the lesion site, BrdU labeling revealed that TGP1-targeted exosomes increased the proliferating neural stem cell population by roughly 1.8-fold compared with untreated injured animals, and Apelin enrichment added a further 1.6-fold increase on top of that. Newly generated immature neurons, marked by DCX, rose approximately 1.5-fold with targeted exosomes and a further 2.3-fold with Apelin enrichment, while later-stage NF200-positive neuronal lineage cells increased 3.3-fold and then 1.5-fold respectively. Three-dimensional reconstruction showed BrdU-labeled cells acquiring NeuN and βIII-tubulin signals, with synaptophysin-positive and VGLUT2-positive puncta in close spatial association—early evidence of synapse-associated neuronal maturation.

The treatment also remodeled the lesion environment itself. Immunofluorescence analysis showed that Apln-TGP1 exosomes most strongly suppressed the accumulation of C3-positive neurotoxic astrocytes and iNOS-positive pro-inflammatory microglia, the two cellular hallmarks of the inhibitory glial response. Anterograde tracing with an AAV-delivered fluorescent reporter revealed greater extension of corticospinal tract fibers in treated animals, and markers of axonal integrity, serotonergic and catecholaminergic innervation, and myelin reconstruction all pointed toward improved structural preservation. MRI confirmed smaller lesions and better tissue continuity. Functionally, the benefits were sustained: BBB locomotor scores improved progressively through 56 days post-injury, motor-evoked potential amplitudes increased, footprint patterns and toe-spreading normalized, thermal withdrawal latency recovered, and hindlimb muscle preserved more of its oxidative slow-twitch fiber character.

Transcriptomic profiling of treated spinal cords tied these effects to molecular programs. RNA sequencing showed differential expression of inflammation-related genes such as Cxcl1 and Ccl19, enrichment of ion transport and synaptic signaling processes, and activation of Apelin and calcium signaling pathways. Gene set enrichment analysis highlighted the PPAR signaling pathway, serotonergic synapse signaling, and ion channel activity. Western blotting confirmed increased expression of APJ, the Apelin receptor, and of PPAR-gamma specifically, with no change in other PPAR subtypes. When the researchers blocked PPAR-gamma with the antagonist GW9662 in cultured neural stem cells, neuronal markers fell and astrocytic markers rose; co-treatment with Apln-TGP1 exosomes reversed these trends and restored proliferative activity. The authors are careful to frame these findings appropriately: the newly generated cells show structural features consistent with neuronal maturation, but functional integration into circuits will require further electrophysiological validation, and longer-term safety studies remain to be done. Still, by separating the problem of targeted delivery from the problem of fate regulation—and solving both within a single nanoscale vesicle—the platform offers a coherent blueprint for coaxing the spinal cord’s own stem cells into doing what they have always been capable of, but rarely choose, to do.

Subject of Research: Engineered exosome-based targeted delivery to endogenous neural stem cells for spinal cord injury repair

Article Title: An engineered exosome-based platform for targeted delivery to endogenous neural stem cells and fate modulation after spinal cord injury

Article References: An engineered exosome-based platform for targeted delivery to endogenous neural stem cells and fate modulation after spinal cord injury. (n.d.). https://doi.org/10.1016/j.mtbio.2026.103718

Image Credits: AI Generated

DOI: 10.1016/j.mtbio.2026.103718

Keywords: spinal cord injury, exosomes, neural stem cells, phage display, Apelin, Lamp2b, neurogenesis, PPAR-gamma, targeted drug delivery, glial scar, neuroinflammation, regenerative medicine