Precise rewiring of corticospinal axons and spinal interneurons via near-infrared optogenetics for spinal cord injury treatment

Abstract

To date, precise restoration of proper connections between posttrauma axons and neurons following spinal cord injury (SCI) remains a substantial challenge. Here, we developed glutamate-linked upconversion nanoparticles (Glu-UCNP) to facilitate optogenetic control of axonal sprouting in SCI mice. After being specifically uptaken by the postsynaptic interneurons innervated by corticospinal tract (CST) axons, Glu-UCNP not only serves as internal light transducers that convert near-infrared light to visible light but also acts as nanobeacons that guide axonal sprouting toward postsynaptic neurons of glutamatergic synapses. This in situ optogenetic modulation successfully demonstrated the restoration of spinal motor circuits by rebuilding functional connections between CST axons and postsynaptic interneurons. It was corroborated by live-cell recording, immunofluorescence staining, in vivo Ca²⁺ imaging, and pellet-reaching tests. Transcriptome sequencing further elucidated the molecular network changes underlying this optogenetic modulation. These findings highlight the potential therapeutic applications of optogenetic modulation in the reassembly of neural circuits after SCI.

Fig. 1.. Characterizations of UCNP.

(A) Schematic diagram of in situ optogenetic modulation–mediated reconnections between CST axons and glutamatergic interneurons. (B) TEM images and elemental component analyses of PEG-UCNP. Scale bars, 50 nm [B(a) and B(c) to B(l)] and 5 nm [B(b)]. HAADF, high-angle annular dark field. (C) TEM images and elemental component analyses of Glu-UCNP. Scale bars, 50 nm [C(a) and C(c) to C(l)] and 5 nm [C(b)]. (D) The XRD patterns of free-UCNP (black line), PEG-UCNP (blue line), Glu-UCNP (red line), and the standard card (orange line, JCPDS: 28-1192) of hexagonal β-NaYF₄ crystals. (E) The particle size distribution of free-UCNP, PEG-UCNP, and Glu-UCNP. (F) Fluorescent emission spectra of free-UCNP, PEG-UCNP, and Glu-UCNP under excitation of a 980-nm laser. (G) Zeta potential of free-UCNP, PEG-UCNP, and Glu-UCNP. Data were represented as mean ± SD. Error bars: SD. a.u., arbitrary units.

Fig. 2.. The specific uptake and axonal guidance of Glu-UCNP in vitro.

(A) Schematic representation of none of specific uptake of PEG-UCNP by MAP2⁺ PSD-95⁻ and MAP2⁺ PSD-95⁺ neurons. (B) In vitro absorption and distribution of PEG-UCNP in MAP2⁺ PSD-95⁺ and MAP2⁺ PSD-95⁻ neurons. Scale bars, 20 and 5 μm. (C) Schematic representation of specific uptake of Glu-UCNP by MAP2⁺ PSD-95⁻ and MAP2⁺ PSD-95⁺ neurons. (D) In vitro absorption and distribution of Glu-UCNP in MAP2⁺ PSD-95⁺ and MAP2⁺ PSD-95⁻ neurons. Scale bars, 20 and 5 μm. (E) Semiquantitative analysis of the fluorescence intensity of UCNP in MAP2⁺ PSD-95⁻ neurons. (F) Semiquantitative analysis of the fluorescence intensity of UCNP in MAP2⁺ PSD-95⁺ neurons. (G) Schematic diagram of the axonal guidance mediated by Glu-UCNP in VChR1-expressed neurons under the illumination of a 980-nm laser. (H) Fluorescence images of VChR1-expressed neurons. (I to L) Live-cell recording of the growth of axons in VChR1-expressed neurons. A transparent agar piece that encapsulated with Tm³⁺- or Er³⁺-doped UCNP was affixed to one side of the bottom of the culture dish. Upon the agar piece was exposed to the 980-nm laser, the blue or green upconversion light it emitted was a potential attractor for axon growth. The growth of axons was recorded for 6 hours by a live-cell workstation with or without the illumination of a 980-nm laser. (M) Statistical analysis of axon length among different groups. (N) Quantitative analysis of angular variation among different experimental groups. [(A), (C), (G), (I) to (L), (M), and (N)] Created in BioRender. Ji, Z. (2025) https://BioRender.com/qmny57u. Data were represented as mean ± SD. Error bars: SD. P values in (E), (F), (M), and (N) were determined using an unpaired two-tailed Student’s t test. P < 0.05 was considered to denote statistical significance. n.s., not significant.

Fig. 3.. Axon sprouting mediated by in situ optogenetic modulation in SCI mice.

(A) Schematic diagram of the experimental procedure. (B to E) Representative immunofluorescence images of sagittal spinal cord sections of mice in the groups of PEG-UCNP (B), Glu-UCNP (C), PEG-UCNP + NIR (D), and Glu-UCNP + NIR (E). Each yellow dashed line denoted the lesion border of the spinal cord. Scale bar, 400 μm (B to E). (F) Quantification of the fluorescence intensity of CST axons at corresponding distances from the lesion borders in (B) to (E). (G) The enlarged view of the white boxed area in (E). [G(a)] VChR1-GFP channel of (G). (H to K) Enlarged views of the white boxed areas in (G). The channels of NeuN (blue), VChR1-GFP (green), and GFAP (red) were applied to label neurons, VChR1-expressed CST axons, reactive astrocytes, respectively. [H(a) and K(a)] Merged images of VChR1-GFP and UCNP channels. [H(b) and K(b)] Merged images of NeuN and UCNP channels. The fluorescence images of UCNP (white) were acquired under the illumination of an externally coupled NIR (980 nm) laser. Scale bars, 100 μm (G) and 25 μm [(H) to (K)]. (A) Created in BioRender. Ji, Z. (2025) https://BioRender.com/a7df18f. Data were represented as mean ± SEM. Error bars: SEM. P values in (F) was calculated by one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test. *P < 0.05, **P < 0.01, and ***P < 0.001; n = 11; two-way ANOVA followed by Fisher’s least significant difference test.

Fig. 4.. In situ optogenetic modulation promoted CST axon sprouting toward postsynaptic neurons of glutamatergic synapses.

(A) Immunofluorescence of the Glu-UCNP + NIR group shows CST axons (GFP, green), neurons (NeuN, blue), and excitatory interneurons (PSD-95, red). [A(a)] A magnified image of the white boxed area in (A). [A(b)] A three-dimensional reconstructed image of the white boxed area in [A(a)], demonstrating spatial contact between CST axons and excitatory interneurons that absorbed Glu-UCNP. Scale bars, 500 μm (A) and 10 μm [A(a) and A(b)]. (B) TEM of the PEG-UCNP group showing PEG-UCNP and mitochondria (red arrowheads). Scale bars, 500 nm. (C) TEM images of Glu-UCNP group showing presynaptic CST axon terminal (yellow), excitatory interneuron (purple), and synaptic structures (red arrowheads). Scale bars, 500 nm. (D to G) Representative immunofluorescence images of sagittal spinal cord sections from mice in the groups of PEG-UCNP (D), Glu-UCNP (E), PEG-UCNP + NIR (F), and Glu-UCNP + NIR (G). Scale bars, 500 μm. (H) Enlarged view of the white boxed area in (G). (I to K) Enlarged views of the white boxed areas in (H). Chx10 (blue), VChR1-GFP (green), and GFAP (red) costaining was used to label VChR1-GFP–expressing CST axons, reactive astrocytes, and glutamatergic V2a interneurons, respectively. I(a) to K(a) were the merged images of VChR1-GFP and UCNP channels. I(b) to K(b) were the merged images of Chx10 and UCNP channels. The fluorescence images of UCNP (white) were acquired under the illumination of an externally coupled NIR (980 nm) laser. Scale bars, 500 μm (H) and 20 μm [(I) to (K)]. (L) Statistical analysis of the percentage of Chx10-positive neurons that ingested UCNP. (M) Statistical analysis of the percentage of Chx10-positive neurons in contact with CST axons. Data were presented as mean ± SD. Error bars: SD. P values in (L) and (M) were calculated by one-way ANOVA followed by Tukey’s post hoc test. P < 0.05 was considered to denote statistical significance.

Fig. 5.. Increase of Ca²⁺ flux in in situ optogenetic modulation reassembled neuronal circuits.

(A) Schematic of the fiber photometry setup for recording calcium transients in interneurons that have uptaken Glu-UCNP and are in contact with CST axons in pellet-reaching mice. (B and C) A schematic representation of Ca²⁺ dynamics on the rostral (B) and caudal (C) sides of the lesion site, respectively. Heatmaps [B(a) and C(a)] and time curves [B(b) and C(b)] of Ca²⁺ signaling from in vivo Ca²⁺ imaging experiments. Mean values were indicated by solid lines, and SEM values were presented by shaded areas. (D and E) Statistical analysis of the changes in fluorescence intensity of Ca²⁺ (D) and the area under curves (E) at both rostral and caudal of the spinal cord. Groups A to F represented PEG-UCNP, Glu-UCNP, PEGUCNP + NIR, Glu-UCNP + NIR, and Glu-UCNP + NIR + KYNA, respectively. (A) Created in BioRender. Ji, Z. (2025) https://BioRender.com/ilw0335. Data were expressed as mean ± SD. Error bars representing SD. P values in (D) and (E) were determined using an unpaired two-tailed Student’s t test. P < 0.05 was considered to denote statistical significance.

Fig. 6.. In situ optogenetic modulation promoted the performance of SCI mice in pellet-reaching tests.

(A) Schematic diagram of the pellet-reaching test and categories of test results, including miss, no grasp, drop, and success. (B) Statistical analysis of time coursed success rate of the pellet-reaching tests. (C) The trajectory formed by the motion of shoulder, elbow, wrist, and digits. (D) The trajectories (from lifting paw to contacting the pellet) of mice in all groups during the pellet-reaching tests. Blue lines, PEG-UCNP; red lines, Glu-UCNP; green lines, PEG-UCNP + NIR; purple lines, Glu-UCNP + NIR. (E and F) Statistical analysis of the trajectory length (E) and velocity (F) during the pellet-reaching tests at 42 and 84 d.p.i.. (G) Statistical analysis of the rate of miss, no grasp, drop, and success during the pellet-reaching tests at 42 and 84 d.p.i.. Data were represented as mean ± SEM. Error bars: SEM. P values in (B) was calculated by one-way ANOVA followed by Tukey’s post hoc test. *P < 0.05, **P < 0.01, and ***P < 0.001; n = 11. Data are represented as mean ± SD. Error bars: SD. P values in (E) to (G) were calculated by two-way ANOVA followed by Tukey’s post hoc test. P < 0.05 was considered to denote statistical significance.

Fig. 7.. Molecular networks underlying in situ optogenetic modulation.

(A) Gene Ontology (GO) terms were generated for genes that were up-regulated or down-regulated in the Glu-UCNP + NIR (FG⁺GFP⁺) groups compared to sham (FG⁺GFP⁺), PEG-UCNP (FG⁻GFP⁺) group, PEG-UCNP + NIR (FG⁻GFP⁺) group, Glu-UCNP (FG⁻GFP⁺) group, or Glu-UCNP + NIR (FG⁻GFP⁺) group. Differentially expressed genes related to neuronal activity, synapse formation, neuronal differentiation, and axonal growth were color-coded from red (high expression) to blue (low expression). ATP, adenosine triphosphate. (B) KEGG enrichment analysis was between the Glu-UCNP + NIR (FG⁻GFP⁺) and Glu-UCNP + NIR (FG⁺GFP⁺) groups. The enrichment factor represents the ratio of differentially expressed genes to the total genes, with dot size indicating genes count involvement in specific signaling pathways and color showing false discovery rate (FDR) [−log₁₀ (FDR)] ranging from 0.75 (blue) to 0.25 (red). TGF-β, transforming growth factor–β. (C) Western blots of PI3K-AKT, JAK-STAT3, NEFH, and GAP43 signaling pathway related proteins at 84 d.p.i. (D to J) Semiquantitative analyses of NEFH, GAP43, and proteins related to PI3K-AKT-mTOR and JAK-STAT signaling pathways based on Western blots. (K to N) Coimmunostaining analysis of phosphorylated and total PI3K/JAK2 in corticospinal neurons. Representative confocal micrographs showed immunostaining of p-PI3K (K), PI3K (L), p-JAK2 (M), and JAK2 (N) in the FG⁻GFP⁺ and FG⁺GFP⁺ corticospinal neurons of the Glu-UCNP + NIR group. Scale bar, 50 μm. (O to R) Semiquantitative analysis of the fluorescence intensity of p-PI3K, PI3K, p-JAK2, and JAK2 in the FG⁻GFP⁺ and FG⁺GFP⁺ corticospinal neurons of the Glu-UCNP + NIR group. n = 3. Data were presented as mean ± SD. Error bars: SD. P values were calculated by one-way ANOVA followed by Tukey’s post hoc test [(D) to (J)] or unpaired two-tailed Student’s t test [(O) to (R)]. P < 0.05 was considered to denote statistical significance.