Hypothalamic Hnscr regulates glucose balance by mediating central inflammation and insulin signal

Abstract Objectives Hypothalamic dysfunction leads to glucose metabolic imbalance; however, the mechanisms still need clarification. Our current study was to explore the role of hypothalamic Hnscr in glucose metabolism. Materials and Methods Using Hnscr knockout or htNSC‐specific Hnscr overexpression mice, we evaluated the effects of Hnscr on glucose metabolism through GTTs, ITTs, serum indicator measurements, etc. Immunofluorescence staining and Western blotting were performed to test inflammation levels and insulin signalling in hypothalamus. Conditioned medium intervene were used to investigate the effects of htNSCs on neuronal cell line. We also detected the glucose metabolism of mice with htNSCs implantation. Results Hnscr expression decreased in the hypothalamus after high‐fat diet feed. Hnscr‐null mice displayed aggravated systematic insulin resistance, while mice with htNSC‐specific Hnscr overexpression had the opposite phenotype. Notably, Hnscr‐null mice had increased NF‐κB signal in htNSCs, along with enhanced inflammation and damaged insulin signal in neurons located in arcuate nucleus of hypothalamus. The secretions, including sEVs, of Hnscr‐deficient htNSCs mediated the detrimental effects on the CNS cell line. Locally implantation with Hnscr‐depleted htNSCs disrupted glucose homeostasis. Conclusions This study demonstrated that decreased Hnscr in htNSCs led to systematic glucose imbalance through activating NF‐κB signal and dampening insulin signal in hypothalamic neurons.

were reported in multiple cell types located in the mediobasal hypothalamus (MBH), including neurons, astrocytes, and htNSCs. [6][7][8] And enhanced hypothalamic inflammation finally contributes to the development of pre-diabetes. 9 HtNSCs were reported to play a leading role in the regulation of the aging process and metabolic disorders. 10,11 Inflamed htNSCs had impaired neurogenesis capacity, which contributed to the loss of certain neurons, 8,12 and also had altered secretory patterns, 13 finally resulting in a metabolic imbalance.
Here, we discovered the lncRNA Hnscr participated in the regulation of high-fat diet-induced insulin resistance. Hnscr was decreased in the hypothalamus after high-fat diet feeding. Knockout of Hnscr aggravated high-fat diet-induced glucose clearance and tolerance, contributing to peripheral insulin resistance. Conversely, htNSC-specific Hnscr overexpression ameliorated this phenotype. Mechanically, Hnscr stabilized trim56 from degradation, leading to decreased NF-κB phosphorylation. Through conditioned medium intervention and htNSCs implantation, we found Hnscr-knockout htNSCs further enhanced inflammation and dampened insulin signal in neurons located in the mediobasal hypothalamic region (MBH) through secretions, finally damaged systematic insulin sensitivity.

| Hnscr knockout aggravates high fat diet-induced glucose disorders
In previous study, we found the pivotal roles of Hnscr, a lncRNA specially enriched in htNSCs, in regulating aging process, and aged Hnscr null mice had aggravated aging-related physiological phenotype. 10 Since hypothalamus is reported as the headquarter of aging development and whole-body metabolism, 5

we further inquired whether
Hnscr participated in modulating glucose metabolism. To this end, we first detected Hnscr level in mice with a normal chow diet (NCD) and high-fat diet (HFD) and found decreased hypothalamic Hnscr level after HFD feed ( Figure 1A). Hnscr null mice and their age-matched wild-type mice were challenged with NCD and HFD for 3 months.
Hnscr depletion increased fasting blood glucose level after HFD feeding ( Figure 1B), with no significant difference in random blood glucose compared with their controls ( Figure 1C). Long-term HFD feed elevated circulating insulin level ( Figure 1D). And Hnscr null mice had higher level of HOMA-IR index ( Figure 1E), indicating dampened AKT, and GSK3β phosphorylation in liver. Data are presented as mean ± SEM (n = 5-6). Statistical significance was calculated by two-tailed Student's t test or two-way ANOVA (*p < 0.05, **p < 0.01, ***p < 0.001, #p < 0.05, ##p < 0.01) insulin sensitivity induced by Hnscr depletion. We also conducted glucose tolerance tests (GTTs) and insulin tolerance tests (ITTs) to evaluate their capacity for glucose tolerance and clearance. Hnscr null mice had worse performance in both GTTs and ITTs with or without highfat diet feed, revealed by higher glucose level after glucose or insulin stimulation ( Figure 1F-I). Additionally, the hepatic insulin signalling was also damaged after Hnscr knockout, demonstrated by decreased level of phospho-IR, phospho-AKT, and phospho-GSK3β after insulin stimulation ( Figure 1J,K). We also detected parameters related to lipid metabolism. Hnscr null mice had aggravated liver steatosis, with increased liver weight and greater liver size ( Figure S1a

| Hypothalamic Hnscr overexpression attenuates systematic insulin resistance
To confirm the regulatory effects of hypothalamic Hnscr in glucose metabolism, we generated a mice model with htNSC-specific Hnscr overexpression also contributed to activated insulin signalling, with increased phospho-IR, phospho-AKT, and phospho-GSK3β

| Hnscr knockdown in htNSCs enhances inflammation in hypothalamus
Increased hypothalamic inflammation contributed to aging process and metabolic dysfunction. 9 RNA-sequencing was conducted on htNSCs of Hnscr null mice and wild-type mice in relevant study, 10 and GO analysis demonstrated the altered inflammatory response pathway based on the DEGs (differentially expressed genes) in Hnscr-null htNSCs. In light of these evidences, we inquired whether Hnscr deficiency-related glucose metabolic disorder was mediated by hypothalamic inflammation. Increased phospho-NF-κB was found in the hypothalamus of Hnscr-null mice with HFD, compared with controls ( Figure 3A Besides inflammation, the hypothalamic action of insulin signalling also affects glucose homeostasis. 23 We further investigated whether insulin signal in hypothalamus were affected by Hnscr.

| Hnscr attenuates NF-κB activation by stabilizing Trim56
To investigate the underlying mechanism, we isolated htNSCs from tein are reported as important regulators of immunity and inflammation, and some of them were implicated in NF-κB activation. [14][15][16][17] Reduced trim56 expression at the protein level was found after Hnscr knockout, rather than mRNA level ( Figure 4I-K). Meanwhile, the suppressive effects could be abolished by protease inhibitors treatment ( Figure 4L,M), suggesting that Hnscr protected Trim56 from proteasome degradation. We further investigated whether trim56 mediated the effects of Hnscr on NF-κB activation, and transfected both Ad-Hnscr and Ad-shTrim56 into 293 T cell line. Interestingly, decreased phospho-NF-κB was partly reversed by trim56 ablation (Figure 4N,O). Hence, reduced Hnscr contribute to NF-κB activation, and this effect can be mediated by decreased trim56 expression.

| Secretions of Hnscr-null htNSCs lead to inflammation and insulin resistance in neurons
HtNSCs can exert their effects through endocrine ability. 18 Based on these findings, we hypothesized it was the secretion of htNSCs that led to enhanced inflammation and dampened insulin signal in neurons. Therefore, we extracted the conditioned medium from htNSCs of secretion. 18 We further inquired whether the changed neuronal phenotype was mediated by htNSC-derived sEVs. Accordingly, sEVs were isolated from CM-HnscrKO and CM-wildtype, subsequently treating CAD cell line. Interestingly, sEVs treatment also

| Hnscr-null htNSCs implantation impairs systematic insulin sensitivity
To elucidate the effects of htNSCs in systematic glucose metabolism, we further implanted Hnscr-null htNSCs and control htNSCs into MBH of adult wild-type mice ( Figure 6A). Exogenous htNSCs survived in the MBH area of recipient mice 5 days after injection ( Figure 6B). In line with previous findings, Hnscr-null htNSCs implantation led to increased TNFα level in the hypothalamus ( Figure 6C), indicating the pro-inflammatory effects of Hnscr-null htNSCs. After intra-MBH injection, mice were fed with a one-month high-fat diet to investigate whether Hnscr-depleted htNSCs affected the early stage of high-fat diet-induced glucose disorder. Mice with Hnscr-depleted htNSCs implantation had dampened ability of glucose clearance, as revealed by ITTs ( Figure 6D,E), while no significant difference was found in the GTTs ( Figure 6F,G). Taken together, local supplement of Hnscrdepleted htNSCs aggravated central inflammation, and partly exacerbating the high-fat diet-induced glucose disorder.

| DISCUSSION
In our study, we found the regulatory role of Hnscr, a htNSC-abundant lncRNA, in glucose homeostasis. Hnscr-knockout mice had aggravated systematic insulin resistance after HFD feed, while htNSC-specific Hnscr F I G U R E 6 Hnscr-null htNSCs implantation impaired systematic insulin sensitivity. (A) Scheme of Hnscr-null htNSCs implantation. HtNSCs from Hnscr-null and wildtype mice were injected into MBH of recipient mice, followed by high-fat diet-feed. Five days after injection, hypothalamus were isolated to detect survival of implanted htNSCs and inflammatory levels. One month later, GTTs and ITTs were conducted. (B) Immunofluorescence of Dil-labelled htNSCs in MBH. Scale bar: 50 μm. (C) qPCR analysis of TNFα, IL-1β and IL-6 in hypothalamus of mice injected with Hnscr-null htNSCs and wildtype htNSCs. (D-G) Glucose tolerance tests and insulin tolerance tests of mice with Hnscr-null htNSCs implantation. Statistical significance was calculated by two-tailed Student's t test (*p < 0.05) overexpression revealed the opposite phenotype. Further investigation revealed that Hnscr depletion in htNSCs enhanced NF-κB activation and altered secretory pattern, therefore leading to aggravated inflammation and decreased insulin signal in neurons located in the arcuate nucleus of the hypothalamus (ARH). And Hnscr-deficient htNSCs implantation aggravated whole-body insulin resistance.
HtNSCs are a group of neural stem cells capable of self-renew and differentiation, which locate in the mediobasal hypothalamus and the third ventricle wall. 5 HtNSCs are implicated in the regulation of glucose metabolism. IKKβ/NF-κB activation in htNSCs impaired their multi-directional differentiating capacity, 8,19,20 and contributes to decreased POMC neurons, further aggravating obesity and pre-diabetes phenotype. 8 HtNSCs also exert their function through the secretion function. Zhang et al. reported that altered miRNA secretory patterns from inflammatory htNSCs accelerated the aging process. 18 And decreased parathymosin secretion from htNSCs led to enhanced senescence in neighbouring neurons. 21 In our study, we found that Hnscr deficiency in htNSCs increased the vulnerability of NF-κB activation after TNFα exposure, and vice versa. Besides htNSCs, we also noticed that Hnscr ablation Since hypothalamic insulin signal was reported to be implicated in the regulation of obesity and insulin resistance, 22 we also investigated the neuronal insulin signalling and found the reduced phospho-AKT level in neurons after Hnscr knockout in htNSCs. Two main populations of neurons, POMC and AgRP, express leptin receptor and insulin receptor, 23,24 and insulin action in these neurons affected systematic insulin resistance. 25 Inhibition of AKT phospholation in POMC neurons induced age-and diet-related insulin resistance, 26-28 while, increasing PI3K activity, the upstream of AKT, in POMC neurons improved insulin and glycemic responses. 29 Thus, we would investigate the implicated neurons in following studies.
In our study, we started from the findings that dampened systematic insulin sensitivity were induced by global Hnscr knockout.
Inspired by hepatic insulin resistance in Hnscr-null mice, we also constructed transgenic mice with liver-specific Hnscr overexpression to inquire whether Hnscr expression in hepatocytes mediated the effects. Unfortunately, hepatic Hnscr overexpression failed to alter peripheral insulin sensitivity or liver steatosis after HFD feed ( Figure S3). Thus, we turned our attention towards htNSCs, since

| Cells
HtNSCs were isolated according to our protocol described previously. 10  Experiments were performed with htNSCs that had been passaged to 3-7 generations.

| Conditional medium intervention
For conditional medium intervention, the conditioned medium was harvested from htNSCs of wild-type or Hnscr-null mice, and filtered through 0.22 μm ultrafiltration filters. CAD cells were seeded in 24-well plates at the density of 2 Â 10 6 cells per well 1 day before the intervention, and treated with a mixed conditioned medium (CM: DMEM/F-12 medium = 3:1) for 48 h, followed by detection of insulin signal and inflammatory level.
Mice were anaesthetized with 1% sodium pentobarbital (0.15 ml/20 g) and fixed to the locator. The X, Y, and Z axis values were read using the Bregma point coordinates as the origin. The MBH was located 5.8 mm below the skull surface, 1.7 mm behind the Bregma horizontal line, and 0.25 mm lateral to the midline of the brain. 10 The cranial surface at the hypothalamic localization was drilled with a miniature handheld cranial drill. The needle was slowly introduced into the hypothalamus, and the micro-injection pump was fixed to pump the AAVs at a rate of 0.2 μl/min.
The cell implantation method is as previously described. 8,18 Briefly, cultured htNSCs were labelled with DiI following the instruction of the cell plasma membrane staining kit (Beyotime, China). DiI-labelled htNSCs were then suspended in 0.5 μl phosphate buffer saline and bilaterally injected into the MBH (9000 cells on each MBH side) using the coordinates. The same injection site as above.

| Glucose tolerance test (GTT) and insulin tolerance test (ITT)
The levels of blood glucose were measured by a glucometer monitor (Sinocare). Mice were intraperitoneal injection with glucose (Sigma, 1 g/kg) or insulin (Sigma, 1 U/kg). Blood was taken from the tail vein at 0 min, 15 min, 30 min, 60 min, 90 min, and 120 min to measure blood glucose values respectively.

| Blood glucose, serum insulin, and homeostasis model assessment of insulin resistance index (HOMA-IR)
The levels of blood glucose were measured by a glucometer monitor (Sinocare). The levels of serum insulin were detected by an ELISA kit (Cusabio, China). Homeostasis model assessment of insulin resistance (HOMA-IR) index was calculated using the following formula: fasting glucose levels (mmol/L) Â fasting serum insulin (mIU/L)/22.5.

| Western blot analysis
The Western Blot analysis was conducted as previously described. 31
The primer pairs used for qRT-PCR were listed.