Adult neurogenesis of the median eminence contributes to structural reconstruction and recovery of body fluid metabolism in hypothalamic self-repair after pituitary stalk lesion

Abstract

Body fluid homeostasis is critical to survival. The integrity of the hypothalamo-neurohypophysial system (HNS) is an important basis of the precise regulation of body fluid metabolism and arginine vasopressin (AVP) hormone release. Clinically, some patients with central diabetes insipidus (CDI) due to HNS lesions can experience recovery compensation of body fluid metabolism. However, whether the hypothalamus has the potential for structural plasticity and self-repair under pathological conditions remains unclear. Here, we report the repair and reconstruction of a new neurohypophysis-like structure in the hypothalamic median eminence (ME) after pituitary stalk electrical lesion (PEL). We show that activated and proliferating adult neural progenitor cells differentiate into new mature neurons, which then integrate with remodeled AVP fibers to reconstruct the local AVP hormone release neural circuit in the ME after PEL. We found that the transcription factor of NK2 homeobox 1 (NKX2.1) and the sonic hedgehog signaling pathway, mediated by NKX2.1, are the key regulators of adult hypothalamic neurogenesis. Taken together, our study provides evidence that adult ME neurogenesis is involved in the structural reconstruction of the AVP release circuit and eventually restores body fluid metabolic homeostasis during hypothalamic self-repair.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00018-022-04457-1.

Introduction

Homeostasis and precise regulation of body fluids and osmolality are essential for survival. In the central nervous system, the hypothalamo-neurohypophysial system (HNS) is the most important functional structure for regulating fluid metabolic balance and electrolyte stability, and it is a component of the hypothalamic neuroendocrine system [1]. The HNS plays a fundamental role in the maintenance of body fluid homeostasis by secreting arginine vasopressin (AVP) in response to internal stimulus signals. Neuroendocrine magnocellular neurons (MCNs) in the supraoptic nucleus (SON) and paraventricular nucleus (PVN) synthesize AVP and oxytocin. Their long axons project through the third ventricle floor and pituitary stalk, forming the hypothalamo-neurohypophysial tract (HNT) and terminating in the neurohypophysis, and these neuroendocrine hormones are released into the bloodstream through the pituitary portal system [2, 3].

The integrity of the hypothalamo-neurohypophysial system is the basis for the precise and timely regulation of body fluid metabolism. Lesions in the HNS can lead to dysfunction in body fluid metabolic regulation, which can be manifested by central diabetes insipidus (CDI) and neuroendocrine hormone disorder [1, 4, 5]. However, in previous clinical and animal studies [6, 7], we found partial compensation and recovery of fluid metabolic function after hypothalamic injury, suggesting that the adult hypothalamus has the potential for self-repair through unknown mechanisms. It is critical to understand the self-repair mechanisms of the adult hypothalamus and develop strategies to improve the outcomes of hypothalamic dysfunction.

Recent studies have revealed that neurogenesis is involved in the repair of the central nervous system after injury [8–10]. Evidence has shown that adult neurogenesis occurs in the mammalian hypothalamus [11]. Adult hypothalamic neurogenesis has important physiological effects on the regulation of hypothalamic-controlled behaviors, such as energy metabolism, aging, and reproduction [12, 13]. However, it remains unclear whether hypothalamic neurogenesis can participate in functional reconstruction and recovery under pathological conditions. In the present study, we used a pre-established animal model of hypothalamic-related fluid metabolic dysfunction to confirm that adult hypothalamic neurogenesis plays a crucial role in the structural reconstruction and functional recovery of body fluid metabolism.

Materials and methods

Animals

Male Sprague–Dawley (SD) rats (6–8 weeks of age) with an average body weight of 200 g (180–220 g) were housed in independent metabolic cages in a temperature-controlled room with a daily light and dark cycle. Food and water were provided without restriction except in the water deprivation experiment. Pregnant 12.5-day-old C57BL/6 female mice for the hypothalamic neural progenitor cell culture were purchased from the Experimental Animal Center of Southern Medical University.

Hypothalamic-related body fluid metabolic disorder model

As described in our previous article, the model was constructed by pituitary stalk electrical lesion (PEL) [7, 14]. In brief, the rats were anesthetized with 5% isoflurane delivered in the air in an enclosed chamber for 2 min. Then, the rats were mounted on a stereotaxic frame (Stoelting). During surgery, 1.5–2% isoflurane was delivered in the air at 0.5 L/min to maintain anesthesia. The skull was opened by removing an approximately 3 × 3 mm square of bone, and a lesion knife with a 2.5 mm wide curved head was lowered in the coronal plane 3.8 mm caudal to the bregma in the sagittal midline until it reached the floor of the skull base. To achieve PEL, a cathodic current of 500 µA was applied for 40 s with a constant power supply output (UGO). For the sham group, the knife was lowered 8 mm beneath the surface of the brain, and no electric current was applied. All models were returned to their metabolic cages after surgery. The daily water intake, urine output and urine specific gravity were monitored daily after surgery.

Culture of hypothalamic neural progenitor cells (NPCs)

The NPCs used in this study were isolated from the hypothalami of embryonic day 12.5 (E12.5) mouse embryos based on previously published methods [15, 36]. Briefly, the hypothalamic region was dissected under a microscope. After enzymatic digestion with digestion solution (2 mL 0.25% trypsin + 200 µl DNase + 2 ml DMEM), we added 1 mL DMEM/10% FBS to stop digestion. After filtering through a 70 µm cell strainer and washing with DMEM, the single-cell suspension was collected and cultured in a 5% CO₂ incubator at 37 °C with neurobasal medium containing 20 ng/ml basic fibroblast growth factor (FGF-2, PeproTech), 20 ng/ml epidermal growth factor (EGF, PeproTech), 2% B27 supplement (GIBCO) and 2 mM l-glutamine.

Central diabetes insipidus (CDI) index

The CDI index was calculated as described in our previous article [7]. In brief, a K-means clustering approach was used to categorize the data as ‘normal condition’ or ‘abnormal condition’. The CDI index was defined as the ratio of the counts in the ‘abnormal’ cluster to the total number of data points.

Bromodeoxyuridine (BrdU) labeling

After the operation, models in the sham and PEL groups received intraperitoneal injection (i.p.) of BrdU (100 mg/kg) to label newborn cells. The specific administration schemes of different experiments were as follows: in the short-term experiments, BrdU was given once a day for the first 3 days after the operation, and models were sacrificed at different time points (3rd/7th/14th day) after the operation; in the long-term experiments, BrdU was given once a day for the first 7 days after the operation, and models were sacrificed on the 28th day after the operation.

Immunostaining

Brains were processed as either frozen Sects. (40 µm) or paraffin-embedded (4 µm) sections. The indirect streptavidin-peroxidase method was used for immunohistochemistry staining, as previously described [7, 14]. For immunofluorescence staining, the primary antibodies (Table. S1) were incubated at vendor-suggested concentrations at 4 °C overnight. After washing with phosphate buffer (PBS) three times, tissue sections were then incubated at 4 °C overnight with Alexa Fluor-conjugated secondary antibodies. After washing with PBS six times, the nuclei were stained with a mount medium containing 4´,6-diamidino-2-phenylindole (DAPI). For BrdU staining, tissue sections or cells were pretreated with 2 N HCl for 20 min at 37 °C before being incubated with 0.1 M borate buffer for 10 min at 37 °C. Fluorescence images of the sections were captured using a confocal microscope (LSM980, Zeiss). For staining quantification, at least 5 immunofluorescence sections in each group were randomly selected and were reviewed by 2 experimenters blinded to the parameters. The number of immunopositive cells within the anatomical region (e.g., ME) was quantified and a mean cell count per section was obtained.

Measurement of serum AVP

The serum AVP levels were measured as described in our previous article [7, 16]. In brief, blood samples obtained through cardiac puncture were incubated at room temperature for 30 min and centrifuged (4000 rpm × 20 min), and the supernatant was collected. AVP levels were measured with a vasopressin ELISA kit following the manufacturer’s instructions.

Immunolabeling-enabled 3D imaging of solvent-cleared organs (iDISCO)

The tissue samples were treated as described in previous article [17, 18]. In brief, the model brains were dehydrated by gradually transferring them to PBS containing higher percentages of methanol until they were finally transferred to pure methanol. Next, the specimens were bleached overnight at 4 °C in a 1:5 ratio of hydrogen peroxide:methanol. The tissues were gradually rehydrated in PBS by removing the methanol in 20% increments, washed with detergent PBS containing 0.2% Triton X-100, and then incubated at room temperature with blocking serum for 1 day, primary antibodies for 3 days, and secondary antibodies for 3 days. Next, the immunolabeled tissues were cleared overnight with 50% tetrahydrofuran/H₂O (THF) (Sigma). After the samples were cleared in THF, they were transferred to 80% THF/H2O for 1 h, then to 100% THF for 1 h, again to fresh 100% THF for 1 h and finally to dichloromethane (Sigma) until they sank to the bottom of the vial. Finally, the samples were transferred to dibenzyl ether (DBE) (Sigma) until clear and stored in DBE at room temperature. Images were acquired with a multiphoton microscope (Olympus, FV1200MPE). Then, 3D volume files were generated using IMARIS software (Bitplane).

RNA transcriptome sequencing of hypothalamic tissue

The sham and PEL models were sacrificed at 1 day or 7 days after PEL (3 models/group), and hypothalamic tissues were dissected in ice-cold PBS. Hypothalamic AVP nuclei and ME tissues were acquired using forceps under stereomicroscope. RNA Preparation, library construction and RNA transcriptome sequencing of hypothalamic tissue was performed as described in our previous article [16]. Raw and processed RNA-seq data generated in this study are deposited into the Gene Expression Omnibus (GEO) database with accession number GSE187602, GSE167904. Genes with an adjusted p value < 0.05 found by DESeq2 were assigned as differentially expressed. Gene ontology (GO) enrichment analysis of differentially expressed genes was implemented by the clusterProfiler R package [19], in which gene length bias was corrected. GO terms with corrected p value < 0.05 were considered significantly enriched by differential expressed genes.

Electron microscopy

Models were deeply anesthetized with pentobarbital (50 mg/kg, i.p.) and then perfused with 0.1 M phosphate buffer (pH, 7.4) followed by buffer containing 2.5% glutaraldehyde. Tissue embedding and electron microscopy were performed as described [20, 21]. Electron microscopy was performed using a transmission electron microscope (Hitachi H-7500).

Intraventricular drug delivery

To infuse the drug into the ventricle, an osmotic minipump (Model 2001, ALZET) filled with aCSF or specific drugs was implanted through a stereotactic instrument into the right lateral ventricle (− 1.0 mm posterior, − 2.0 mm lateral, − 4.0 mm ventral to bregma) after PEL.

In vivo virus infection

Lenti-CMV-Nkx2.1/Control shRNA was constructed by Obio Technology (Nkx2.1 shRNA targeted sequence: 5´-CCATGTCTTGTTCTACCTT-3´). For in vivo virus infection, models were anesthetized and placed in a stereotaxic instrument for two injections (1.0 μl per injection) of purified virus into the median eminence region (− 3.0/-3.3 mm posterior, − 0.0 mm lateral, − 10.0 mm ventral to the bregma) immediately after PEL. Virus was injected at a rate of 100 nl/min using a Hamilton syringe and microinjection system (KDS Legato 130). The needle syringe was left in place for 10 min before being withdrawn. The in vivo transfection efficiency was measured by western blotting after virus injection. When compared with the noninjected controls, the injection did not significantly increase the death rate or change the normal activity of the models.

Western blotting

Western blotting was performed as described in our previous article [14]. Western blotting was performed using cells or hypothalamic median eminence samples. For hypothalamic extraction, the models were deeply anesthetized before the brains were removed from the skull and the hypothalamic median eminence was separated with forceps under a microscope. Cells and tissues were lysed with RIPA buffer. The protein concentration was determined using a BCA assay kit (Beyotime). Samples were subjected to SDS-PAGE and transferred to PVDF membranes (Millipore). Then, the membranes were blocked with 5% bovine serum albumin and incubated with primary antibodies overnight at 4 °C. The next day, after washing with TBST, membranes were incubated with the corresponding HRP-conjugated secondary antibodies for 1 h at room temperature. Finally, signals were detected using chemiluminescence reagents, and images were captured with a digital camera.

Data analysis of single-cell sequencing (scRNA-Seq).

The hypothalamic single-cell sequencing data were obtained from GEO database, data number: GSE132355. Single-cell sequencing data were downloaded from GEO and followed the original cell type annotation information. The single-cell sequencing expression matrix was analyzed by R4.0.3 software. The "FindAllMarkers" function of the Seurat 3.1.5 package was used to calculate the differentially expressed genes among each group of cells [22]. The Seurat 3.1.5 package was used for cell clustering and differentially expressed gene mapping. UMAP (Uniform Manifold Approximation and Projection for Dimension Reduction) was used to display the results of cell clustering. The Monocle3 package was used for pseudotime analysis [23, 24].

Cell transfection

Lenti-CMV-Nkx2.1/Control shRNA-EGFP-Puro was constructed by Obio Technology. Nkx2.1 shRNA targeted sequence: 5´-CCATGTCTTGTTCTACCTT-3´. The average lentivirus titer was 5.0 × 10⁸ TU/ml. Hypothalamic NPCs were plated on poly-d-lysine/laminin-coated six-well plates and then transfected with lentiviral particles according to the manufacturer's instructions. The transfection efficiency was measured by western blotting after transfection.

Proliferation and differentiation analyses of cultured NPCs

The proliferation and differentiation of NPCs were analyzed using previously published methods [25, 36]. For each analysis, at least 3 independent experiments were performed and used for statistical analyses. To study cell proliferation, we dissociated NPCs with trypsin and plated them in proliferation medium at a density of 50,000 cells/well on poly-d-lysine/laminin-coated confocal dishes. At 20 h postplating, 5 μM BrdU was added to the culture medium for 6 h. NPCs were then washed with PBS and fixed with 4% paraformaldehyde for 30 min at room temperature, followed by immunostaining analysis.

For the differentiation assay, at 24 h postplating, the cells were transferred to differentiation medium and neurobasal medium containing 1 μM retinoic acid, 2% B27 supplement, 2 mM L-glutamine, and 0.5% fetal bovine serum for 5–7 days. Cells were then washed with PBS and fixed with 4% paraformaldehyde for 30 min at room temperature, followed by immunostaining analysis.

Statistical analysis

The sample sizes for experiments were based on previous experience. All data are presented as the mean ± SEM. SPSS 22.0 (IBM) and GraphPad Prism 8.0 (GraphPad Software) software were used for statistical analysis. Data were subjected to Kolmogorov–Smirnov and Shapiro–Wilk normality tests to assess the distribution. The two-tailed Student’s t test or Mann–Whitney U test were used for single comparisons between two groups. One-way ANOVA was used for comparisons between multiple groups. Differences were considered statistically significant at p values < 0.05. Single and double asterisks in the figures indicate statistical significance (*p < 0.05, **p < 0.01, n.s.: no significant differences).

Results

Patterns and partial compensation of body fluid metabolic function after hypothalamic injury

In our previous clinical research [6], it was reported that some patients with CDI due to clinical hypothalamic injury can experience recovery compensation of body fluid metabolism, and the prognosis of this group of patients appears to be related to the structural integrity of the hypothalamic median eminence (ME). A pituitary stalk electrical lesion (PEL) rat model, which avoids excessive damage in the anterior pituitary, was established in our previous study to assess the disorder patterns and mechanism of hypothalamic self-repair in body fluid metabolism [7, 14]. Rats from the PEL and sham model groups were placed in separate metabolic cages, and their daily water intake (DWI), daily urine output (DUO) and urine specific gravity (USG) were monitored for 4 weeks after the operation (Fig. 1A, B).

Fig. 1.

Recovery of body fluid metabolism and pathological changes in the ME after PEL. A, B Schematic representation of experiments. All models were placed back in metabolic cages after the operation, after which BrdU was administered i.p. once a day for 7 days to label newborn cells. The daily water intake, urine output and urine specific gravity were collected daily for 4 weeks. C–E The recovery patterns of body fluid metabolism, including daily water intake (C), daily urine output (D), and urine specific gravity (E). F The CDI index of the PEL model in the first 2 weeks and the 4th week after the operation. G The serum AVP content of PEL models in different stages. H Multiphoton series images of AVP immunolabeling solvent-cleared hypothalami of sham and PEL models. Related to Supplementary Videos 3 and 4. Scale bar, 400 μm. Arrowheads show blood vessels and arrows show anteroposterior AVP axonal projections. I, J Representative images of immunofluorescence staining for AVP (green) and CD34 (red) in sagittal sections of the hypothalamus-pituitary region in sham (I) and PEL (J) models. Scale bar, 500 μm. K Area size quantification of pituitary and neurohypophysis in sham and PEL models. L, M Representative images of immunofluorescence staining for AVP (green) and CD34 (red) in coronal sections of the ME in sham (L) and PEL (M) models. Scale bar, 200 μm. N Quantification of CD34 + cells in the ME of sham and PEL models. Data are shown as the mean ± SEM or medians ± interquartile ranges (n = 4–6 in [C–G], n = 6–7 in [K, N]). *p < 0.05, **p < 0.01 by two-tailed Student’s t test or one-way ANOVA. 3rd V third ventricle, ME median eminence, PPit posterior lobe of pituitary, IPit intermediate lobe of pituitary, APit anterior lobe of pituitary

The overall body fluid metabolic status of the rats from the sham model remained stable during the observation period. The average DWI was 34.0 ± 0.6 mL, the average DUO was 15.6 ± 0.4 mL, and the average USG was 1.0309 ± 0.0003 (Fig. 1C–E). The PEL models presented a classical triphasic pattern of CDI, which can be divided into three periods of fluid metabolic disorder (Fig. 1C–E). The first acute CDI stage (postoperative day [POD] 1), the second temporary remission stage (POD 2–4) and the third persistent CDI stage (POD 5–28) constitute the characteristics of fluid metabolic disorders, which have been described in our previous studies [26].

Interestingly, it was found that partial recovery of body fluid metabolic function occurred at postoperative week 4 rather than in the first 2 weeks. The DWI, DUO, and USG in the first 2 weeks were 79.5 ± 3.5 mL, 48.2 ± 2.7 mL, and 1.0122 ± 0.0005, respectively. The DWI, DUO, and USG in the 4th week were 62.9 ± 2.1 mL, 35.6 ± 1.7 mL, and 1.0139 ± 0.0002, respectively, and all showed significant differences between the two periods. The CDI index, which is calculated by analyzing the three types of data mentioned above, can be used to comprehensively evaluate the body fluid disorder state [7, 16]. The CDI index of the PEL model was significantly lower in the 4th week than in the first 2 weeks after the operation (Fig. 1F). Furthermore, when compared with the previous period (POD 7 and 14), the serum AVP content of PEL models was restored to nearly normal levels on POD 28 (Fig. 1G). All these observations suggest that body fluid metabolic behaviors and AVP neuroendocrine hormone levels experienced functional recovery and compensation after hypothalamic injury.

Pathological destruction of the hypothalamo-neurohypophysial system after PEL

Functional compensation is always accompanied by tissue repair. To explore the structural basis of body fluid homeostasis and AVP neuroendocrine hormone recovery in hypothalamic lesions, we analyzed the pathological changes in the AVP hypothalamo-neurohypophysial tract. First, hypothalamic slice AVP staining and multiphoton AVP immunolabeling-enabled three-dimensional imaging of solvent-cleared organs (iDISCO) were used to analyze the upstream nucleus that produces AVP, and the number of AVP neurons and fiber connections in the SON and PVN were both significantly reduced (Fig. S1A and B, Supplementary Videos 1 and 2). Next, the downstream neurohypophysis, an effector organ for the transmission and direct release of AVP hormone, was analyzed by AVP staining of hypothalamic sagittal sections. After PEL, both the total pituitary and the neurohypophysis atrophied, the volume was reduced, and the AVP signal was absent in the neurohypophysis (Figs. 1I, J, S1C). This indicates that in PEL models, the neurohypophysis no longer has the physiological functions of AVP transmission and release. Therefore, we focused on the unknown structural basis for the compensation recovery of the abovementioned AVP release function.

Reconstruction of AVP axons and blood Vessels in the hypothalamic median eminence

Both our and other previous studies have shown that, under physiological conditions, upstream bilateral AVP nuclei emit AVP axons, which converge through the ME and assemble into longitudinal neuroendocrine tracts that extend posteriorly to the pituitary stalk and neurohypophysis [3, 17]. Therefore, when the pathway of AVP fibers from the pituitary stalk to the neurohypophysis is disrupted, the ME becomes the end of the hypothalamic AVP projection. In the ME of the PEL model, the AVP fibers are distributed from the upper middle part to the lower part of the ME (Figs. 1, I, J, S1C). In addition, multiphoton series images of AVP immunolabeling solvent-cleared hypothalami also confirmed the changes in the AVP distribution in the ME, and interestingly, new blood vessel structures appeared in the middle and lower parts of the ME (Fig. 1H, Supplementary Videos 3 and 4). Furthermore, immunofluorescence coronal section staining for AVP and CD34 (marker of vascular endothelial cells) showed increased remodeled vessels in the middle part of the ME in the PEL model (Fig. 1L–N). CD34, MECA-32 (marker of fenestrated vessels) and AVP staining indicated increased interactions between AVP fibers and ME vessels (Fig. S1D). These results suggest that the ME appears to be a hypothalamic self-repair reconstruction tissue in which AVP fibers and vessels cater to each other, which may be the structural basis for partial compensatory recovery in long-term CDI after PEL.

Adult neurogenesis occurs in the median eminence during the hypothalamic self-Repair process after PEL

We hypothesized that the reorganization of AVP fibers and blood vessels in the ME promotes the release of AVP hormones. It has been reported that the regulation of neuroendocrine hormone release in the local hypothalamus requires the participation of functional neurons [27–30], which are relatively lacking in the ME under physiological conditions. To explore the biological mechanism of hypothalamic self-repair, hypothalamic tissues were extracted and transcriptomics sequencing was performed. Gene enrichment analysis showed that, in the short term after PEL, differentially expressed genes were primarily enriched in the biological processes of the cell cycle, cell division, cell proliferation, and chromosome separation (Fig. S2A). As time progressed, differentially expressed genes were primarily enriched in neurogenesis processes, such as neuron generation, neuron development, neuron differentiation, and neuron projection development (Fig. S2B). Together, these results suggest that there is a continuous neurogenesis process in the hypothalamus after PEL, ranging from cell proliferation to neuron maturation.

To confirm the regions and patterns of hypothalamic neurogenesis after PEL, intraperitoneal injections of bromodeoxyuridine (BrdU) were used to label mitotic newborn cells in different groups after PEL (Fig. 2A). SOX2, a well-known neural stem/progenitor cell (NPC) marker, was used for the analysis. NPCs were activated and transformed into the proliferating state. A larger number of newborn NPCs (BrdU + SOX2 +) appeared in the ME of the PEL group than in the ME of the sham group (Fig. S2C-E). The peak proliferation of NPCs occurred primarily in the short term, which was consistent with the sequencing results (Fig. S2F and G). Next, to see if these proliferating NPCs would transform into neurons, we performed fluorescent staining with doublecortin (DCX, a marker of immature neurons) and BrdU. As shown in Fig. 2B–H, the number of immature neurons (DCX +) continued to increase in the ME for two weeks after PEL, and the conversion of BrdU-labeled newborn cells to immature neurons (BrdU + DCX +) occurred in approximately one week.

Fig. 2.

Adult neurogenesis occurs in the hypothalamic median eminence during the hypothalamic self-repair process. A Schematic representation of experiments. PEL or sham operations were performed, after which BrdU was administered i.p. once daily for 3 days. Models were sacrificed on the 3rd or 7th or 14th days after the operation. B–H Representative images of immunofluorescence staining for BrdU (green) and DCX (red) in coronal sections of the ME on the 3rd day (B, C), 7th day (D), and 14th day (E) after Sham or PEL. B–E scale bar, 200 μm. High-magnification images of box areas in (B–E) are shown in (B´–E´), respectively. B´–E´ scale bar, 50 μm. Arrowheads show newborn immature neurons. Quantification of DCX + cells and BrdU + DCX + cells is shown in (F). The relative percentage of BrdU + DCX + /DCX + cells is shown in (G). The relative percentage of BrdU + DCX + /BrdU + cells is shown in (H). I, J Representative images of immunofluorescence staining for BrdU (green) and NeuN (red) in the ME of sham (I) and PEL (J) models on the 28th day after the operation. Arrowheads show newborn mature neurons. Scale bar, 50 μm. K Quantification of BrdU + cells in the ME of sham and PEL models on the 28th day after the operation. L The relative percentage of BrdU + NeuN + , BrdU + GFAP + and BrdU + CD34 + cells in BrdU + cells in the ME of PEL models on the 28th day after the operation. Data are shown as the mean ± SEM or medians ± interquartile ranges (n = 5–12 in [F– H, K and L]). *p < 0.05, **p < 0.01 by two-tailed Student’s t test or one-way ANOVA

Furthermore, to determine the long-term differentiation fate of these newborn cells, a long-term BrdU labeling strategy was applied (Fig. 1B), and markers of mature neurons (NeuN), astrocytes (GFAP) and CD34 were used for immunostaining. There were still many preserved newborn cells in the ME on the 28th day after PEL (Figs. 2I–K, S2H and I). 53.7 ± 1.6%, 23.5 ± 1.7% and 8.36 ± 0.8% of newborn cells had transformed into mature neurons, astrocytes and vessels, respectively (Fig. 2L). Interestingly, when the hippocampal dentate gyrus (a classic neurogenesis region in adults) involved in the PEL needle tract was stained, this neurogenic transformation did not appear (Fig. S2J). In summary, these results suggest that a regionally and functionally specific neurogenesis process emerges in adult hypothalamic ME after PEL.

Adult neurogenesis of the ME is involved in the reconstruction of AVP regulated neural circuits and affects the outcome of body fluid metabolism after PEL

Structure is the basis of function. The ability of residual hypothalamic AVP MCNs to produce neuroendocrine vesicles and transport them to the reconstructed ME after PEL was first confirmed by electron microscopy (Fig. S3A-D). Nerve fiber connections and synapse formation are the structural basis for the interaction and regulation of different types of neurons. To explore the integration of ME newborn neurons and AVP fibers after PEL, markers of mature axons and dendrites (microtubule-associated protein 2, MAP2) and of presynaptic (synaptophysin 1, SYN1) and postsynaptic (postsynaptic density protein-95, PSD-95) structures were used for multiple immunofluorescence analysis. Newborn mature neurons (BrdU + MAP2 +) formed numerous and close neural connections with redistributed AVP fibers in the ME (Fig. 3A). Moreover, synaptic structures were found to be formed in the local microenvironment in newborn mature neurons (BrdU + MAP2 +) and were dominated by presynaptic regulatory structures in the ME after PEL (Fig. 3B, C). To verify the effect relationship between these newborn neurons and the metabolic state of body fluids, we conducted a water deprivation experiment on the PEL models (Fig. S3E). ME newborn neurons can be activated (BrdU + c-Fos +) by the altered metabolic state of body fluids, and serum AVP increases after water deprivation (Fig. S3F–H). The results further indicate that newborn neurons in the ME have functions related to body fluid regulation.

Fig. 3.

Adult neurogenesis of the ME is involved in the reconstruction of AVP regulated neural circuits and affects the outcome of body fluid metabolism after PEL. A Representative images of immunofluorescence staining for MAP2 (red), AVP (green) and BrdU (white) in the ME on the 28th day after Sham or PEL. Scale bar, 100 μm. High-magnification images of box areas are shown in (A1–A4). Arrowheads show BrdU + MAP2 + neurons that have connections with AVP axons. B Representative images of immunofluorescence staining for MAP2 (red), SYN1 (green) and BrdU (white) in the ME on the 28th day after PEL. Scale bar, 50 μm. High-magnification images of box areas in B are shown in (B1, B2). Arrowheads show BrdU + MAP2 + neurons that express the presynaptic marker SYN1. C Representative images of immunofluorescence staining for MAP2 (red), PSD-95 (green) and BrdU (white) in the ME on the 28th day after PEL. Scale bar, 50 μm. High-magnification images of box areas in (C) are shown in (C1, C2). Arrowheads show BrdU + MAP2 + neurons that express the postsynaptic marker PSD-95. D Schematic representation of experiments. aCSF, Ara-C or IGF1 was delivered i.c.v. for 7 days after the operation. BrdU was administered i.p. once a day for 7 days to label newborn cells. Models were sacrificed on the 28th day after the operation. E, F Representative images of immunofluorescence staining for BrdU (green) in the ME of different treatment groups (E). Scale bar, 200 μm. Quantification of the number of BrdU + cells in different treatment groups on the 28th day (F). G–I The biological parameters of body fluid metabolism, including daily water intake (G), daily urine output (H), and urine specific gravity (I), in the different treatment groups. J The CDI index of different treatment groups in the 4th week after the operation. K The serum AVP content of different treatment groups on the 28th day after the operation. Data are shown as the mean ± SEM or medians ± interquartile ranges (n = 4–6 in [F–K]). *p < 0.05, **p < 0.01 by two-tailed Student’s t test or one-way ANOVA

To clarify the role of adult neurogenesis in the ME on the recovery of AVP-related body fluid metabolism, we performed the intraventricular (i.c.v.) administration of previously reported mitotic inhibitor (cytarabine, Ara-C) and neurogenesis activator (insulin-like growth factor 1, IGF1) and control artificial cerebrospinal fluid (aCSF) to regulate the neurogenesis level of the ME after PEL (Fig. 3D) [31–33]. The number of newborn cells and adult neurogenesis level of the ME was affected by different treatments (Figs. 3E, F, S3I and J). The inhibition of ME proliferation after PEL resulted in worse body fluid metabolism outcomes. In contrast, the enhancement of ME neurogenesis after PEL resulted in a relatively better body fluid metabolism outcome (Fig. 3G–K). Different treatment effects on the redistribution of AVP fibers and vasculature in the ME were ruled out (Fig. S4A, B). Overall, these results suggest that adult neurogenesis in the ME is crucial for the functional recovery of body fluid metabolism following PEL.

Analysis of single-cell sequencing data suggests Nkx2.1 is involved in the regulation of hypothalamic neuronal development

Adult hypothalamic NPCs are a continuation of embryonic hypothalamic NPCs and retain the potential for multidirectional differentiation after hypothalamic development [11, 13, 34]. Some researchers have used single-cell sequencing (scRNA-Seq) technology to conduct preliminary analysis of the developing hypothalami of mice [35]. Therefore, we performed an in-depth analysis targeting processes of hypothalamic neuronal development based on published single-cell sequencing data (GSE132355) to screen the key regulators of hypothalamic neurogenesis.

Data from 74,146 cells were analyzed, and a cell type clustering analysis was performed to classify them into eight hypothalamic cell types: hypothalamic NPCs, hypothalamic neurons, astrocytes, oligodendrocytes, microglia, endothelial cells, meningeal cells, and immune cells (Fig. 4A). Next, pseudotime analysis of the development of hypothalamic neurons was performed, and the pseudotime trajectory of the neurons was found to be consistent with the temporal pattern of hypothalamic tissue sampling (Fig. 4B, C), which confirmed the reliability of the pseudotime analysis of the hypothalamic neurons. We also screened out genes that fit the pseudotime trajectory of neuron development and found several genes in the Nkx family of homeodomain transcription factors. Cells expressing the transcription factor NK2 homeobox 1 (Nkx2.1, also called Nkx2-1 and Ttf-1) accounted for the greatest proportion of these cells, corresponding to 12.35% of the total number of neurons.

Fig. 4.

Single-cell sequencing analyses reveal that Nkx2.1 is involved in the regulation of hypothalamic neuronal development. A Cell type clustering analysis of 74,146 hypothalamic cells with color-coded identity. B, C Pseudotime analysis of hypothalamic neurons (B) and individual neurons displayed at a series of developmental timepoints (C). D, E The Nkx2.1 gene expression status in different hypothalamic cell types. (F) Expression dynamics of the Nkx2.1 gene in NPCs and neurons during hypothalamic development. Data shown as violin plots. G, H Relative expression level of the Nkx2.1 gene in NPCs and neurons. **p < 0.01 by Mann–Whitney U test

To verify the specific role of Nkx2.1 in the differentiation and transformation of NPCs to neurons, we first targeted all cells expressing Nkx2.1 and found that Nkx2.1 was almost exclusively expressed in developing hypothalamic NPCs and neurons (91.4%) (Fig. 4D, E). Moreover, the temporal expression pattern of the Nkx2.1 gene in these developing hypothalamic NPCs and neurons was consistent with the neurogenesis pattern. As the hypothalamus developed, the number of Nkx2.1-expressing cells and their expression levels increased during the period of NPC activation (E12–E14). During the following period of neuron generation (E16-E18), the number of Nkx2.1-expressing cells appeared to increase significantly, but the gene expression level appeared to decrease over time (Fig. 4F). The overall analysis of the average expression of the Nkx2.1 gene in hypothalamic NPCs was higher than that in hypothalamic neurons (Fig. 4G, H), and the Nkx2.1 gene ranked third among all differential genes among these positive NPCs and neurons (Table. S2). Taken together, these results suggest that Nkx2.1 is involved in the developmental regulation of hypothalamic neurons. It was found that some Nkx2.1-positive cells were still retained in the adult hypothalamus ME using the mouse brain atlas in the Allen Brain Map (https://portal.brain-map.org/). Moreover, Nkx2.1 shows a high degree of conservation and homology in rats and mice (http://asia.ensembl.org/).

NKX2.1 regulates neuronal differentiation in hypothalamic ME neurogenesis

BrdU labeling was used to confirm the role of NKX2.1 in the neurogenesis of hypothalamic ME (Fig. 2A), and well-known NPC markers (Nestin), NKX2.1, and BrdU were used for immunofluorescence staining analysis. After PEL, the activated NPCs (BrdU + Nestin +) in the ME expressed NKX2.1 (Fig. 5A–C), and the number of NKX2.1-positive and newborn NKX2.1-positive (BrdU + NKX2.1 +) cells increased within a week (Fig. 5A–D. The proportion of BrdU-labeled NKX2.1-positive NPCs (BrdU + Nestin + NKX2.1 +) decreased after 3 days (Fig. 5E, F), which is consistent with the differentiation pattern. The NKX2.1 protein expression pattern of the ME was also confirmed by western blot (WB) analysis (Fig. S6A). To confirm the role of NKX2.1 in neuronal maturation and body fluid metabolism recovery in the ME after PEL, NKX2.1 expression in the ME was inhibited by stereotactic injection of lentivirus into the PEL model (Fig. 6A). WB analysis of ME tissue confirmed the effectiveness of the interference in vivo (Fig. S6B). Compared with the control group, inhibiting the expression of NKX2.1 in the ME after PEL resulted in poorer body fluid metabolism and lower serum AVP in the long term (Fig. 6B–F). Although the number of newborn cells in the ME was not significantly altered, the number of newborn mature neurons was significantly reduced after NKX2.1 expression intervention (Fig. 6G–J. The effect of NKX2.1 intervention on AVP fibers and vascular redistribution was also ruled out (Fig. S4C, D). These results suggest that NKX2.1 is involved in the transformation of hypothalamic NPCs into mature neurons in the ME following PEL.

Fig. 5.

NKX2.1 expressed in activated newborn NPCs in the ME after PEL. A–C Representative images of immunofluorescence staining for Nestin (green), NKX2.1 (red) and BrdU (white) in the ME of sham models (A) or on the 3rd day (B) and 7th day (C) after PEL. Scale bar, 200 μm. High-magnification images of box areas in A-C are shown in (A´–C´). Scale bar, 50 μm. Arrowheads show BrdU + NKX2.1 + cells. D Quantification of the number of NKX2.1 + cells (left) and NKX2.1 + BrdU + cells (right) in the ME of different groups. E Quantification of the number of NKX 2.1 + BrdU + Nestin + cells in the ME of different groups. F Quantification of the percentage of NKX2.1 + BrdU + Nestin + cells in ME NKX2.1 + BrdU + cells for different groups.Data are shown as the mean ± SEM or medians ± interquartile ranges (n = 8–9 in [D–F]). *p < 0.05, **p < 0.01 by two-tailed Student’s t test or one-way ANOVA

Fig. 6.

NKX2.1 expression interference affects body fluid metabolism prognosis and neuronal differentiation of hypothalamic NPCs. A Schematic representation of experiments for shRNA-mediated Nkx2.1 knockdown. Lentiviruses expressing control shRNA or Nkx2.1 shRNA were administered in the ME immediately after PEL operation, followed by BrdU administration i.p. once daily for 7 days. Models were sacrificed 28 days after the operation. B–D The biological parameters of body fluid metabolism, including daily water intake (B), daily urine output (C), and urine specific gravity (D) in different groups. E The CDI index of different groups in the 4th week after the operation. F The serum AVP content of different groups on the 28th day after the operation. G, H Representative images of immunofluorescence staining for BrdU (green) and NeuN (red) in coronal sections of the ME in the two groups. Scale bar, 50 μm. Arrowheads show BrdU + NeuN + neurons. I Quantification of BrdU + cells and BrdU + NeuN + cells in the two groups. J Quantification of the percentage of BrdU + NeuN + cells in BrdU + cells for different groups. K Western blot analysis of NKX2.1 interference efficiency in vitro. L, M BrdU proliferation experiment of hypothalamic NPCs after Nkx2.1 knockdown in vitro. Representative images of immunofluorescence staining for BrdU (L). Quantification of the percentage of BrdU + cells in different groups (M). N–Q Differentiation induction experiment of hypothalamic NPCs after Nkx2.1 knockdown in vitro. Representative images of immunofluorescence staining for TUJ1 (N). Quantification of the percentage of TUJ1 + cells in different groups (O). Quantification of the number (P) and length (Q) of TUJ1 + cell dendrites in different groups.Data are shown as the mean ± SEM or medians ± interquartile ranges (n = 5–6 in [B–J, M], O and [P], n = 11–14 in [Q]). *p < 0.05, **p < 0.01 by two-tailed Student’s t test

Furthermore, to demonstrate that NKX2.1 regulates hypothalamic neural differentiation, we used our previously established embryonic hypothalamic NPCs primary culture technology to test NKX2.1 function in vitro [36]. NKX2.1 was found to be highly expressed in hypothalamic NPCs, which was consistent with the single-cell sequencing results (Fig. S5A, B). WB was used to confirm that the lentiviral transfection interfered with NKX2.1 expression in hypothalamic NPCs (Fig. 6K). The BrdU proliferation experiment confirmed that NKX2.1 inhibition had no effect on the proliferation of hypothalamic NPCs (Fig. 6L, M). However, in the differentiation induction experiment, NKX2.1 intervention in hypothalamic NPCs resulted in a decrease in the conversion rate of neurons (TUJ +) and a decrease in length of neural dendrites in vitro (Fig. 6N–Q. Together, the results of in vivo and in vitro experiments show that NKX2.1 regulates neuronal differentiation and maturation of hypothalamic NPCs.

NKX2.1 regulates adult hypothalamic neurogenesis via the sonic hedgehog signaling pathway after PEL

Previous studies have shown that the biological effects of NKX2.1 are associated with several signaling pathway mechanisms, such as the sonic hedgehog (SHH) signaling pathway, the Notch signaling pathway, and the FGF signaling pathway [37–39]. To determine the signaling pathway mechanism by which NKX2.1 regulates adult neurogenesis of the ME during hypothalamic self-repair, we extracted ME tissue after PEL to perform immunoblotting analysis. Related SHH signaling pathway molecules (SHH, SMO, and GLI1) appeared to change in the same pattern as NKX2.1 (Fig. S6A) and were similarly consistent with NKX2.1 expression suppression of the ME in the PEL model (Fig. S6B). These results suggest that the role of NKX2.1 in regulating ME adult neurogenesis after PEL is likely linked to the SHH signaling pathway.

To further confirm the regulatory relationship between NKX2.1 and the SHH signaling pathway as well as its effect on body fluid function restoration, previously reported SHH signaling inhibitors (vismodegib, Vis), SHH signaling activators (smoothened agonist, SAG) and control aCSF were administered intraventricularly after PEL (Fig. 7A) [40, 41]. WB analysis of ME tissue confirmed the effectiveness of the interference in vivo (Fig. S6C). SHH signaling inhibition of the ME after PEL resulted in poorer body fluid metabolism and lower serum AVP in the long term (Fig. 7B–F). The number and proportion of newborn mature neurons was also significantly reduced (Fig. S7A–C). Moreover, activation of SHH signaling reversed the negative effects of NKX2.1 interference on body fluid metabolism (Fig. 7B–F) and neuronal transformation after PEL (Fig. S7A–C). The effects of SHH signaling intervention on AVP fibers and vascular redistribution were also ruled out (Fig. S7D, E). Consistent with all these results, it was confirmed that NKX2.1 regulates adult neurogenesis in the ME via the SHH signaling pathway during the hypothalamic self-repair process after PEL.

Fig. 7.

NKX2.1-mediated SHH signaling pathways plays a role in functional recovery of body fluid metabolism after PEL. A Schematic representation of experiments. aCSF, Vismodegib or SAG was delivered i.c.v. for 7 days starting on the 3rd day after the operation. BrdU was administered i.p. once a day for 7 days to label newborn cells. Models were sacrificed on the 28th day after the operation. B–D The biological parameters of body fluid metabolism, including daily water intake (B), daily urine output (C), and urine specific gravity (D) in different groups. E The CDI index of different groups in the 4th week after the operation. F The serum AVP content of different groups on the 28th day after the operation. Data are shown as the mean ± SEM or medians ± interquartile ranges (n = 5 in [B–F]). *p < 0.05, **p < 0.01 by one-way ANOVA

Discussion

The central nervous system, including the cerebral cortex, hippocampus, and spinal cord, shows structural plasticity and self-repair abilities under injury and pathological conditions [9, 10, 42–44]. Structural plasticity and self-repairing abilities are the basis for functional compensation and recovery after injury. However, less attention has been paid to whether the hypothalamus, which is the central regulator of a wide range of homeostatic and instinctive physiological processes, has the potential for structural reconstruction [13]. One important reason for this could be the difficulty of developing hypothalamic injury models. Here, we demonstrated the process and neurogenesis mechanisms of hypothalamic functional reconstruction in self-repair using a preconstructed rat model of hypothalamic-related fluid metabolic dysfunction (Fig. 8). We found that the hypothalamic ME reconstructed into a new neurohypophysis-like structure, regulated the release of AVP neuroendocrine hormones, and restored body fluid metabolic function after the failure of neurohypophysis function. In humans, a newly formed ectopic posterior pituitary has been found in the region of the infundibular recess of the third ventricle in the patients with pituitary stalk interruption syndrome (PSIS), and the assumption that the HNT axons establish new vascular connection during development resulting in normal levels of AVP and absence of CDI also indicates the structural plasticity of hypothalamus [45–47].

Fig. 8.

Schematic diagram of the hypothesis that adult ME neurogenesis contributes to structural reconstruction and recovery of body fluid metabolism after PEL. ME appears to be a hypothalamic self-repair reconstruction tissue in which AVP fibers and vessels cater to each other after PEL. Activated and proliferating adult NPCs differentiate into mature neurons, which then integrate with remodeled AVP fibers to reconstruct the local AVP hormone release neural circuit in the ME after PEL. Mechanistically, NKX2.1 regulates adult neurogenesis via SHH signaling pathway in the ME after PEL

Structural reconstruction of the ME is an efficient and economical way for the body to repair the AVP neuroendocrine release function of the HNS after PEL. First, the ME is a component of the AVP hypothalamic pituitary axonal transmission pathway. Previous research has shown that the ME is the confluence of the long axons of AVP neurons in the bilateral upstream SON and PVN, and that it is the hypothalamic converging channel before the AVP HNT projects to the neurohypophysis [3, 17]. In addition, under physiological conditions, the capillaries on the ventral side of the ME are part of the portal system and release hypothalamic neuroendocrine hormones, such as thyroid stimulating hormone releasing hormone, gonadotropin releasing hormone, growth hormone releasing hormone, and a small amount of AVP [48, 49]. Thus, we found that remodeling the distribution of AVP axons and vessels near each other in the ME may be a relatively quick way to repair AVP release pathways after PEL. The possible mechanisms of axon regeneration and angiogenesis of ME still need to be further studied. Moreover, in our study, the serum AVP content was restored to near-normal levels after PEL. This implies that the residual AVP MCNs in the upstream nucleus require a more powerful AVP hormone releasing capacity. These AVP neurons require more input signals that modulate the release of AVP hormone. The reconstruction of AVP fibers and blood vessels in the ME reestablishes AVP transmission, whereas the local neural signal serves as the "power motor" that drives the release of AVP neuroendocrine hormones. It was determined that adult neurogenesis of the hypothalamic ME is involved in this important process after PEL.

Recently accumulated evidence has indicated that the hypothalamus is an important area of adult neurogenesis in the central nervous system [11, 13]. Adult hypothalamic neurogenesis is involved in a variety of homeostatic processes and physiological behaviors, such as energy metabolism, reproduction, and body temperature homeostasis. In the physiological state, adult neurogenesis in the hypothalamus is lower than that in classic adult neurogenesis areas, including the subgranular zone of the hippocampal dentate gyrus and the subventricular zone of the lateral ventricle [13, 50]. This was also verified in our experimental sham group (Figs. 2I and S2C). Moreover, increased levels of adult neurogenesis in the hypothalamic ME often require stimulation by pathological factors, which may be related to the location of the ME. The ME, which is located outside of the blood–brain barrier, is also the primary site for sensing changes in circulating metabolite and hormone levels and is thus subtly positioned to undergo plastic changes in response to environmental conditions [51, 52]. Previous studies indicate adult neurogenesis can play a role in postinjury repair processes in the central nervous system [8, 10, 43]. Therefore, in our research, we used transcriptome sequencing, BrdU incorporation, drug regulation and other strategies to confirm that a complete adult neurogenesis process of NPC proliferation, differentiation and maturation appeared in the ME after PEL. Newborn neurons eventually integrated with remodeled AVP fibers to reconstruct the local AVP hormone release neural circuit. We found that the phenomenon of body fluid metabolism-related adult neurogenesis is hypothalamic region-specific and is a purposeful and functionally corresponding repair process. Previous studies have shown that the regulation of hypothalamic AVP hormone release is related to various molecules and channels in AVP neurosecretory terminals [27–30]. For example, ATP, P2X receptor family, calcium channel, purinergic, opioid, dopamine, etc. We hypothesized that the newborn ME neurons interact with AVP fibers via axons and form synaptic regulatory structures after PEL. These neurons alter the ion concentration within AVP neurosecretory terminals through the aforementioned molecular channels, affecting the release of AVP vesicles and promoting AVP hormone secretion. This process requires further study by electrophysiological methods.

Tanycytes located on the sidewall and bottom of the third ventricle of the hypothalamus are thought to be the main source of adult hypothalamic neurogenesis. Based on their distribution location, morphology and gene expression, tanycytes are classified into four main subtypes: α1, α2, β1 and β2 [13, 53]. In our study, when the distribution regions of proliferating NPCs in the ME were compared, the activated newborn NPCs were mostly found in the upper layer of the ME, and subsequently, these labeled cells appeared in the middle layer of the ME (Fig. S2D and E), suggesting that the source of NPCs in the ME after PEL may be β2 tanycytes. The change in cell distribution could be due to local migration of NPCs, which is consistent with the redistribution trend of AVP and blood vessels. Additionally, it is worth noting that in this study, not all newborn cells in the ME were converted into mature neurons. Gliogenesis and angiogenesis also appeared in ME after PEL, suggesting that the repair of the hypothalamic AVP release circuit is a systemic process. The increase of new blood vessels may be intended to allow for more efficient AVP release. Therefore, in the pharmacological intervention experiment, due to the antimitotic effect of Ara-c, both neurogenesis and angiogenesis were inhibited, so the long-term recovery of CDI in the Ara-c treated group was significantly affected. The functional roles of these other types of cells need to be investigated further.

At the mechanistic level, we focused on differentiation fate regulators that activate the conversion of neural stem cells into mature neurons during the ME adult neurogenesis process. Adult hypothalamic NPCs are a population of cells that have retained their multidirectional differentiation potential after hypothalamic development [11, 13]. Thus, based on the analysis of single-cell sequencing data from the developing hypothalami in mice [35], we selected Nkx2.1 as a research candidate because of the specificity of its cell expression distribution and the match between its expression pattern and the neural pseudotime trajectory and hypothalamic neurogenesis. In addition, Nkx2.1 has high gene order conservation (GOC) and whole genome alignment (WGA) scores for ortholog pairs between rats and mice in Ensembl (http://asia.ensembl.org/), indicating the high-confident orthologies and conservation of Nkx2.1 in these two species. It has been previously reported that NKX2.1 plays a role in neuronal development processes, such as classification, migration, and differentiation, in the embryonic hypothalamus [39, 54, 55]. NKX2.1 has been linked to the regulation of body fluid homeostasis in previous studies [56]. In our study, we demonstrated that NKX2.1 plays a key role in adult hypothalamic neurogenesis by regulating the differentiation decision of activated NPCs toward neurons, and that this action is achieved via the downstream SHH signaling pathway.

In conclusion, our study reveals that adult hypothalamic neurogenesis is an important pathophysiological change in hypothalamic self-repair. After a hypothalamic lesion, adult neurogenesis in the corresponding local area of the hypothalamus is triggered to promote the structural–functional reconstruction of the organism. Hypothalamic neurogenesis is the basis for the stabilization and reconstruction of hypothalamic function, which has a direct effect on disease prognosis. Therefore, our research provides basic evidence for the protection of key structures (hypothalamus and ME) in clinical neurosurgery. For some patients with hypothalamic-sellar region surgery (such as craniopharyngiomas, pituitary tumors, etc.), when the dissection of pituitary stalk is unavoidable, the patient may benefit in the postoperative course by dissecting the pituitary stalk as distally as possible and preserving the intact third ventricle floor and ME structures. In addition, the application of pharmacological or biological strategies to promote endogenous hypothalamic neurogenesis may be the potential therapeutic direction for the treatment of hypothalamic-related diseases.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

YO and MZ designed the experiments, generated data, and drafted the manuscript; YO, MC, HD, GW and JP carried out majority of the experiments; HD, KL and XW contributed to sample collection; RY analyzed single-cell sequencing data; XZ and YL co-advised the study and guided experiments; ZF and SQ designed the experiments, supervised the experiments, and co-wrote the manuscript.

Funding

This work was supported by National Natural Science Foundation of China (81900709), Guangdong Basic and Applied Basic Research Foundation (2021A1515110290, 2020A1515110028, 2020A1515110564), Science and Technology Planning Project of Guangzhou (201902020017, 202102020977).

Data availability

The datasets generated during the current study are available from the corresponding author on reasonable request.

Declarations

Conflict of interest

The authors declare no competing interests.

Ethical approval and consent to participate

All animal experiments were performed in accordance with the approved protocols by the ethics committee of Nanfang Hospital, Southern Medical University. No human samples were included in this study.

Consent for publication

All authors agree to publish the current finding.