ICR mice were purchased from Japan SLC. Dcx-EGFP mice (Gong et al., 2003) were provided by the Mutant Mouse Research Resource Center (MMRRC. RRID: MMRRC_000244-MU). All animals were maintained at a temperature of 22 ± 1°C with a 12-h light/dark cycle (lights on between 08:00 and 20:00) and had ad libitum access to food and water. All animal experiments were performed in accordance with the guidelines and regulations of Nagoya City University (Approval No. 21–028).

Recombinant human TGF-β1 (rhTGF-β1) was purchased from Abcam (cat# ab50036), reconstituted in 10 mM citric acid (pH 3.0) to a concentration of 5 μg/100 μL, and stored at −80°C. The following commercially available primary antibodies were used in the present study: rat anti-green fluorescent protein (GFP; 1:500, Nacalai Tesque, cat# 04404–84); rabbit anti-GFP (1:1000, MBL, cat# 598); guinea pig anti-Dcx (1:200, Millipore, cat# AB2253); rat anti-Ki67 (1:500, eBioscience, cat# 14–5698-82); goat anti-Iba1 (1:500, Abcam, cat# ab5076); rabbit anti-TGF-β1 (1:500, Abcam, cat# ab215715); rat anti-TGF-βRI (1:100, Santa Cruz Biotechnology, cat# sc-101574), mouse anti-TGF-βRII (1:100, Santa Cruz Biotechnology, cat# sc-17791); rat anti-BrdU (1:500, Abcam, cat# ab6326); and rabbit anti-Pax6 (1:100, Covance, cat# 901301). Nuclei were stained with Hoechst 33342 (1:10000, Sigma-Aldrich).

Postnatal day 2 mice were subjected to cerebral cortical cryogenic injury as described previously (Jinnou et al., 2018; Nakajima et al., 2024). Briefly, the mice were placed in a stereotaxic instrument (David Kopf Instruments) and anesthetized with 1% isoflurane in oxygen (0.3 L/min). Then, the skull was exposed through a scalp incision. A pre-chilled hexagonal 1.2 mm-wide wrench was used as a probe and prepared using liquid nitrogen. The metal probe was placed on the exposed right skull (0.5 mm anterior and 1.2 mm right lateral to the true lambda) for 10 s. The procedure was repeated three times with an interval of 5 s to re-chill the probe. The scalp was sutured, and the mice were placed in a heated recovery box for 15 min and then returned to their home cage. At 7 days post-injury, mice were fixed by transcardiac perfusion with 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer and postfixed 4 h or overnight in the same fixative at 4°C.

In vitro culture of V-SVZ-derived neuroblasts was performed as described previously (Sawada et al., 2020). Briefly, the V-SVZ tissues were dissected from postnatal day 0–1 pups and dissociated using trypsin–EDTA (Invitrogen). The cells were washed twice in L-15 medium (Invitrogen) containing 40 μg/mL DNaseI (Roche Diagnostics) and seeded on Matrigel (BD Biosciences)-coated cover glasses (Matsunami) with Neurobasal medium (Invitrogen) containing 2% B-27 Supplement (Invitrogen), 2 mM GlutaMAX (Invitrogen), 50 U/mL penicillin–streptomycin (Invitrogen), and 5 μM BrdU (Sigma-Aldrich). rhTGF-β1 was added to the culture medium at a final concentration of 0, 0.5, or 5 ng/mL. At 3 days in vitro, cells were fixed with 4% PFA in 0.1 M phosphate buffer for 30 min at room temperature (RT) and subjected to immunocytochemistry.

After fixation in 4% PFA and postfixation for 4 h, brain samples were incubated in 15 and 30% sucrose solutions, encapsulated in OCT compound (Sakura Finetek Japan), and then cooled using dry ice and isopentane. Serial brain slices (around 0.5 mm anterior to the bregma, slices span: 280 μm) were obtained at a thickness of 10 μm using a cryostat (CryoStar NX70, Epredia). The PIC technique was then performed as described previously (Honda et al., 2022) Briefly, reverse transcription was performed by adding ultraviolet (UV)-responsive 6-nitropiperonyloxymethyl-caged reverse transcription primers containing the T7 promotor, unique molecular identifiers (UMIs), multiple barcodes, and a polyT sequence onto the sections. Immunofluorescence was then performed to visualize regions of interest for subsequent UV irradiation. To cleave 6-nitropiperonyloxymethyl moieties from reverse transcription primers, the regions of interest were irradiated with UV light for 3 min with a Digital Micromirror Device (Polygon 1,000-G; Mightex Systems). The total tissue lysate was then collected and purified using 20 mg/mL proteinase K. Second-strand DNA was synthesized using the nick translation method. For in vitro transcription reactions, synthesized cDNAs were transcribed into RNAs using a T7 transcription kit. The amplified RNAs were further reverse transcribed, followed by paired-end sequencing on an Illumina platform. The sequences were separated by the sample barcodes using UMI-tools and mapped to the reference genome using HISAT2. UMI-tools and feature Counts were used to generate the UMI count data assigned to genes. Differentially expressed genes (DEGs) were extracted using DESeq2 [false discovery rate (FDR) = 0.1]. DESeq2 was also used to transform the count data into regularized log data before performing principal component analysis using the R prcomp function. Gene Ontology (GO) analyses were performed using Metascape (Zhou et al., 2019).¹

The fixed brain was sectioned coronally into 60-μm serial slices (around 0.5 mm anterior to the bregma, slices span: 300 μm) using a vibratome (VT-1200S, Leica) or into 10-μm slices using a cryostat and incubated for 30 min at RT in blocking solution (10% normal donkey serum and 0.2% TritonX-100 in PBS). The brain sections or fixed cells were then incubated overnight at 4°C with the primary antibodies. Bound primary antibodies on samples were visualized with Alexa Fluor-conjugated secondary antibodies (1:1000, Invitrogen) or biotin-conjugated antibodies (1:1000, Jackson ImmunoResearch) along with the Vectastain Elite ABC kit (Vector Technology) and Tyramide Signal Amplification (Thermo Fisher Scientific) according to the standard procedure.

Fluorescence in situ hybridization was performed using the RNAscope multiplex assay (Advanced Cell Diagnostics) according to the manufacturer’s protocol (Wang et al., 2012). Briefly, after OCT compound removal, primary antibodies were applied to 10-μm thick cryostat serial sections (around 0.5 mm anterior to the bregma, slices span: 280 μm) for immunohistochemistry. Then, the sections were fixed using 4% PFA at RT for 30 min, dehydrated with ethanol, treated with 1% hydrogen peroxide at RT for 30 min, and boiled in RNAscope Target Retrieval Reagent (Advanced Cell Diagnostics) at 98°C for 15 min. Next, the sections were hybridized with the probes for 2 h at 40°C. The following commercially available primary antibodies were used in the present study: Ccnd2 probe (Mm-Ccnd2, Advanced Cell Diagnostics, cat# 433211); Sez6l probe (Mm-Sez6l, Advanced Cell Diagnostics, cat# 492631); and Dner probe (Mm-Dner, Advanced Cell Diagnostics, cat# 413951). The sections were then washed and hybridized with amplifiers and visualized using Opal620 (1:1000, Akoya Bioscience). Immediately after this process, secondary antibodies were applied as described in the immunohistochemistry and immunocytochemistry section.

Images were acquired using an FV3000 confocal laser-scanning microscope (Evident) with a 20× objective lens (NA 0.8) or 40× objective lens (NA 1.4). To count the Ccnd2-, Sez6l-, and Dner- mRNA puncta in nuclei, Dcx-positive cells were randomly selected from the dorsal corner of the V-SVZ and migratory zone. All Dcx-positive cells in the field of view were selected to count BrdU-incorporating cells among Dcx-positive cells in vitro.

All statistical tests were two-tailed and analyzed using EZR (Kanda, 2013). The data distribution was analyzed using the Kolmogorov–Smirnov test. Comparisons between two groups were performed using an unpaired t-test. Comparisons among multiple groups were performed using the Kruskal–Wallis test followed by a post-hoc Bonferroni test or one-way ANOVA followed by a post-hoc Bonferroni test. All numerical data are presented as the mean and SEM, and p-values less than 0.05 were considered statistically significant.

To compare gene expression profiles of neuroblasts migrating through the V-SVZ (hereafter referred to as “V-SVZ neuroblasts”) and those migrating in the peri-injured cortex (hereafter referred to as “cortical neuroblasts”), the cerebral cortex of Dcx-EGFP mice, in which neuroblasts were labeled with EGFP, was cryoinjured, and brain sections were prepared for spatial transcriptome analysis (see Materials and Methods; Figure 1A). We identified V-SVZ neuroblasts or cortical neuroblasts as GFP-positive cells localized in the V-SVZ (Figure 1B) or peri-injured cortex (Figure 1C), respectively, and irradiated their Hoechst-labeled nuclei with UV light. GFP-positive nuclei were randomly selected when irradiating V-SVZ neuroblasts (Figure 1D), but among cortical neuroblasts, GFP-positive nuclei that did not overlap with surrounding nuclei and were oval in the direction of migration were selected (Figure 1E; Supplementary Figure 1).

V-SVZ and cortical neuroblasts were collected from four brain slices per mouse, for a total three mice each (50, 50, and 50 nuclei from the V-SVZ; 44, 45, and 50 nuclei from the peri-injured cortex). There was no significant bias in the number of reads and assigned reads/UMIs (Supplementary Figures 2A,B). In the V-SVZ neuroblast-derived samples, 12,959, 11,350, and 14,314 genes (average 12,874 genes) were detected, with UMI counts per cell of 2,788, 1,966, and 4,051 (average 2,935), respectively (Supplementary Figures 2C,D). In contrast, in the cortical neuroblast-derived samples, 10,419, 3,544, and 8,313 genes (average 7,425 genes) were detected, with UMI counts per cell of 1,489, 205, and 728 (average 807), respectively (Supplementary Figures 2C,D). The principal component analysis results of the gene expression profiles showed that the clusters of cortical neuroblast-derived and V-SVZ neuroblast-derived samples were clearly separable according to their origin (Figure 2A). DEG analysis showed that a total of 215 genes were significantly differentially expressed (FDR = 0.1), with 97 genes upregulated in cortical neuroblasts and 118 genes upregulated in V-SVZ neuroblasts (Figure 2B). To confirm the veracity of the DEG data obtained, immunostaining and FISH were performed on the three molecules that showed moderate changes in expression. Consistent with the DEG dataset, immunoreactivity of the Pax6 protein (Kohwi et al., 2005) was significantly downregulated in cortical neuroblasts compared to V-SVZ neuroblasts (Supplementary Figures 3A–C). In contrast, the numbers of fluorescent puncta for Sez6l- and Dner-mRNA were significantly increased in cortical neuroblasts (Supplementary Figures 3D–I). These results indicate clear differences in the gene expression profiles between cortical and V-SVZ neuroblasts.

GO analysis was then performed on the 97 DEGs that had significantly higher expression and the 118 DEGs that had significantly lower expression in cortical neuroblasts than in V-SVZ neuroblasts. Genes that were downregulated in cortical neuroblasts showed terms related to neurogenesis, maintenance of the immature state, and the Wnt and BMP signaling pathways (Figure 2C). However, genes that were upregulated in cortical neuroblasts had terms related to adjusting the direction of migration, resistance to cell death, and the integrin and TGF-β signaling pathways (Figure 2D). For example, Ccnd2 (which was associated with the GO term neurogenesis) was significantly downregulated in cortical neuroblasts (Figure 2E). However, Lpar1 and Mtor (Figure 2F, associated with the GO term migratory orientation) and Gclm, Scl1a1, and Mef2c (Figure 2G, associated with the GO terms inflammation and resistance) were significantly upregulated in cortical neuroblasts. These results suggest that cortical and V-SVZ neuroblasts have distinct gene expression properties, and thus these cells exhibit different characteristics.

To verify the downregulation of Ccnd2 in cortical neuroblasts detected by PIC, Ccnd2-mRNA expression was visualized by fluorescence in situ hybridization. In brains at 7 days after injury, Ccnd2-mRNA signaling puncta in the nuclei of Dcx-positive neuroblasts were significantly reduced in cortical neuroblasts compared with those in V-SVZ neuroblasts (Figures 3A,B). The proliferative activity of neuroblasts was also assessed by immunohistochemical staining for Ki67, which is expressed during all active phases of cell cycle progression (G1, S, G2, and M), but not in the G0 resting phase (Zacchetti et al., 2003). Approximately 37% of V-SVZ neuroblasts and 24% of RMS neuroblasts were Ki67-positive, whereas only approximately 2% of cortical neuroblasts were Ki67-positive (Figures 3C,D). These results suggest that cortical neuroblasts have suppressed proliferative activity.

Because TGF-β is a representative inhibitory factor of proliferation that has been extensively studied (Morikawa et al., 2016), we focused on TGF-β signaling pathway-related genes, among which expression was upregulated in cortical neuroblasts in our GO analysis (Figure 2D). In a neonatal mouse model of cortex injury, Iba1-positive microglia/macrophages localized in areas away from the injured area, such as the RMS and corpus callosum, exhibited non-activated ramified morphology (Vidal-Itriago et al., 2022), and did not express TGF-β1 (Figure 4A). In contrast, in the peri-injured cortex, Iba1-positive cells had activated amoeboid morphology and expressed TGF-β1 protein (Figure 4A). To determine whether neuroblasts could respond to secreted TGF-β1, cortical neuroblasts were stained for TGF-β receptors. Dcx-EGFP-positive neuroblasts expressed both type I and type II TGF-β receptors (Figures 4B,C). These results suggest that TGF-β is present in the peri-injured cortex and that it can affect neuroblasts.

Finally, to investigate the effect of TGF-β on the proliferative activity of neuroblasts, V-SVZ-derived neuroblasts were cultured with various concentrations (0, 0.5, and 5.0 ng/mL) of rhTGF-β1. The proliferative activity of neuroblasts was assessed by BrdU incorporation into the nucleus. In the control group, more than 4% of Dcx-positive neuroblasts were proliferating cells labeled with BrdU. In contrast, in the groups supplemented with 0.5 and 5.0 ng/mL of rhTGF-β1, only less than 2% of Dcx-positive neuroblasts were positive for BrdU, indicating that proliferating cells were significantly reduced compared with those in the control group (Figures 5B,C). These results suggest that the proliferative activity of neuroblasts is suppressed by TGF-β1.

Because lesions and their surrounding areas in the injured/diseased brain, including a variety of injured or dead cells and recruited immune cells, show more complex structures than those in intact areas (Meirelles et al., 2017; Yu et al., 2022), high spatial resolutions are required for analyses targeting these lesions. Previously, several spatial transcriptomics analyses of the injured/diseased brain have been performed using spatial capture technology (e.g., 10x Visium) or laser microdissection (Boone et al., 2013; Li et al., 2023; Zheng et al., 2023). However, these techniques had several disadvantages, such as the large size of the analysis unit, which contains multiple cells, and burned regions that cannot be analyzed. In contrast, the PIC method we used can capture samples with nanometer-level spatial resolution and without loss of tissue (Honda et al., 2022). Moreover, PIC can detect gene expression with higher sensitivity than single-cell RNA sequencing or other spatial transcriptome technologies (Supplementary Figure 2; Honda et al., 2022). Using PIC, we succeeded in obtaining small differences in the gene expression profiles of V-SVZ and cortical neuroblasts in the injured brain by precisely segmenting the nuclei of cells of interest based on the labeled proteins and cell morphology (Figures 1, 2; Supplementary Figure 1). Thus, PIC is a useful technique for spatial transcriptome analysis, with high spatial resolution and deep sensitivity without tissue loss, even in environments showing complex structures such as injured or diseased tissues.

Neuroblasts migrating in the RMS maintain an immature state during their long journey to the OB (Francis et al., 1999). However, our GO analysis (Figure 2) showed that migrating neuroblasts in the peri-injured cortex exhibited downregulated expression of genes involved in regulation of the Wnt and BMP signaling pathways, which are known to suppress neuronal differentiation (Wodarz and Nusse, 1998; Le, 2022). In fact, several genes associated with GO terms relevant to the negative regulation of nervous system development and differentiation were significantly decreased in cortical neuroblasts (Figure 2), suggesting that differentiation of neuroblasts is promoted in the injured cortex. In support of this finding, transplantation of NSCs derived from human-induced pluripotent stem cells in the injured brain results in rapid differentiation into neurons and glial cells (Jensen et al., 2013). Similar to our results (Figure 3), studies have reported that approximately 15–20% of neuroblasts migrating in the V-SVZ/RMS of young adult mice are Ki67-positive and are actively proliferating (Ikeda et al., 2010). In this study, we found that the expression of cell proliferation markers such as Ccnd2-mRNA and Ki67 protein was downregulated in cortical neuroblasts (Figures 2, 3). In general, a negative relationship exists between cell maturity and proliferative ability in neuronal lineage cells (Miyashita et al., 2021). Taken together, these findings suggest that cortical neuroblasts have a decreased ability to maintain their immature state, with a concomitant decrease in proliferative activity.

In the normal brain, the migratory direction of V-SVZ neuroblasts is controlled by attractants such as prokineticin-2, netrin, GDNF, and HGF, which are secreted from the OB (Ng et al., 2005; Paratcha et al., 2006; Wang et al., 2011; Nakajima et al., 2021). However, cortical neuroblasts in the peri-injured areas are recruited by activated microglia/macrophage-derived factors in the lesion, such as osteopontin and CXCL12 (Miller et al., 2005; Yan et al., 2009). These lesion-derived attractants function through activation of integrin β1 signaling (Yan et al., 2009). Consistent with previous reports, our GO analysis showed that cortical neuroblasts exhibit upregulated expression of genes involved in chemotaxis and integrin signaling pathways (Figure 2). In addition, cytoskeletal reorganization is required to control the direction of neuroblast migration. The DEG analysis in this study (Figure 2) showed that cortical neuroblasts had upregulated expression of Lpar1 and Mtor, which are involved in reorganization of the actin cytoskeleton (Sarbassov et al., 2004). These results suggest that cortical neuroblasts undergo a dynamic cytoskeletal reorganization, followed by a change in migratory direction.

In the lesion, activated glial cells secrete neurocytotoxic cytokines. For example, IL-1β induces delayed neuronal death by ferroptosis, a non-apoptotic type of cell death caused by iron-dependent reactive oxygen species accumulation after stroke (Dixon et al., 2012; Karuppagounder et al., 2016; Zille et al., 2017). Removal of reactive oxygen species by glutathione can effectively block ferroptosis (Yang and Stockwell, 2016; Alim et al., 2019). Gclm and Slc1a1, which showed upregulated expression in cortical neuroblasts in our DEGs analysis (Figure 2), are molecules involved in the glutathione synthesis pathway and prevent ferroptosis through glutathione production (Gipp et al., 1995; Aoyama et al., 2006; Stockwell et al., 2017). Another gene with upregulated expression, Mef2c (Figure 2), also contributes to neuronal survival by inducing the expression of PGC1a, which is involved in energy metabolism in mitochondria (Ambasudhan et al., 2013). Together, these findings suggest that cortical neuroblasts have acquired the ability to migrate toward the site of injury while resisting the surrounding environment, where inflammation makes survival difficult.

The GO analysis revealed that the TGF-β signaling pathway is enhanced in neuroblasts in the peri-injured cortex (Figure 2), suggesting that these neuroblasts are exposed to TGF-β in the peri-injured cortex. Previous studies have reported that TGF-β1 production is enhanced after stroke and traumatic injury and that its main source is activated microglia/macrophages (Lindholm et al., 1992; Lehrmann et al., 1998). Similarly, in the neonatal mouse model of cryogenic cortical injury used in this study, Iba1-positive microglia/macrophages in the injured cortex expressed TGF-β1 (Figure 4).

TGF-β binds to its receptor, the type I/II TGF-β receptor complex, which subsequently induces phosphorylation of the intracellular signaling molecules Smad2/3. Phospho-Smad2/3 form a complex with Smad4, which translocates into the nucleus and induces expression of genes such as p15/CDK inhibitor-2B and p21/CDK inhibitor-1, thereby inhibiting cell proliferation (Li et al., 1995; Datto et al., 1995; Hachana and Larrivée, 2022). TGF-β has been reported to negatively regulate neural stem/progenitor cell proliferation. For example, TGF-β reduces BrdU- and Ki67-positive proliferative cells in cultured neurons derived from the embryonic hippocampus and cortex (Vogel et al., 2010) and inhibits the formation of neurospheres derived from adult rat V-SVZ cells (Wachs et al., 2006). However, the effects of TGF-β on neuroblasts migrating in the postnatal brain were completely unknown. Our immunohistochemical studies revealed that neuroblasts express type I and type II TGF-β receptors (Figure 4). Furthermore, our in vitro experiments directly demonstrated that neuroblast proliferation is suppressed by TGF-β (Figure 5). These findings suggest that TGF-β secreted by cells in the injured cortex may be involved in the inhibition of proliferative activity of cortical neuroblasts.

In conclusion, we found that migrating neuroblasts exhibit slightly but distinctly different properties depending on the environment of their journey. The information provided by this study on gene expression profiles, comparing neuroblasts migrating in the V-SVZ and in the peri-injured cortex, could contribute to further understanding of the mechanisms of neuronal regeneration and the development of central nervous system regenerative medicine.