OSBPL6 protects against demyelination and behavioral disorder via promoting cholesterol transport in oligodendrocyte.

Save

Introduction

Demyelination is associated with behavioral disorder and cognitive impairment in neuropsychiatric diseases, including major depressive disorder (MDD). However, the mechanism underlying myelin damage remains unclear, despite stress maybe a major risk factor.

Objectives

Here, we demonstrate that hippocampal neurons in the chronic stress-induced depressive mice are prone to demyelination due to disrupted cholesterol transport to myelin sheaths in oligodendrocyte.

Methods and Results

Single-nucleus RNA sequencing of lipid metabolism-related signaling pathway revealed significant reduced levels of oxysterol binding protein-like 6 (OSBPL6) in hippocampal oligodendrocytes of depression mouse model. Consistently, specific knockdown of OSBPL6 in oligodendrocytes induced loss of myelin structure, while upregulating OSBPL6 or enhancing OSBPL6 transcription improved these impairments in depressive mice. Furthermore, restoring cholesterol transport with β-cyclodextrin decreased cholesterol accumulation, improved damaged myelin structure, and rescued depression-related behaviors in depressive mice.

Conclusions

These findings suggest that OSBPL6-mediated cholesterol homeostasis in oligodendrocytes can promote myelin production against chronic stress-induced cerebral demyelination and depression with behavioral disorders.

Introduction

Major depressive disorder (MDD) ranks among the most prevalent neuropsychiatric disorders and is projected to evolve into the second − gravest worldwide burden on both public health and the economy by 2030 [1], [2]. Currently, it is hypothesized that the pathogenesis of depression involves various factors including neurotransmitter and receptor dysregulation, hypothalamic–pituitary–adrenal (HPA) axis dysfunction, cytokine imbalances, genetic and epigenetic impacts, along with structural and functional changes in the brain [3], [4]. However, first-line clinical treatments for MDD such as antidepressants, psychotherapy, repeated transcranial magnetic stimulation (r TMS), and light therapy often come with significant side effects.

This suggests that depression may be a heterogeneous disease requiring further elucidation of its detailed pathogenesis[4]. Approximately half of patients lack effective treatment options [5], [6], adding greater challenges to the diagnosis and management of depression.

Recently, metabolic disorders in the central nervous system (CNS) have emerged as major risk factors for depression; however, their underlying mechanisms remain poorly understood. Several studies have indicated a potential association between myelin metabolism impairment and neurodegenerative diseases' pathophysiology[6], [7]. Moreover, alterations in glycolysis and lipid metabolism have been implicated in Alzheimer's disease development[8].

Microglia-mediated disruption of lipid transfer between neural cells leads to lipid droplet accumulation within microglia[9]. A previous study has highlighted the crucial role played by microglia in clearing toxic lipids which are considered pivotal in multiple sclerosis development[10]. Nevertheless, the involvement of metabolic disorders in depression's pathogenesis remains unclear, and further clarification regarding its detailed mechanism is needed.

Recent research has shown the occurrence of abnormal myelin architecture across various brain regions in individuals with MDD. [7], [11]. Additionally, animal models have shown that demyelination and dysfunction of oligodendrocytes contribute to depression-like behaviors[12], [13]. The myelin sheath is created by helically wrapping mature oligodendrocyte (OLs) plasma membrane extensions around specific axon segments[14].

This not only provides electrical insulation for the axons to enhance action potential conduction speed but also offers metabolic and structural support for the axonal compartment[15], [16]. The stability of the myelin sheath plays a crucial role in its functional specificity. One possible molecular basis for myelin stability lies in its lipid composition, which consists of high levels of saturated long-chain fatty acids, sphingolipids, and cholesterol which accounting for approximately 40% molar percentage of total lipids[17], [18].

Cholesterol homeostasis in oligodendrocytes is essential for proper expression of myelin proteins and migration of oligodendrocyte progenitor cells[19]. Therefore, cholesterol metabolism is critical for the development and maintenance of myelin sheath formation within glial cells in vivo [20], [21].

In light of these findings, our study aims to investigate whether dysregulation of lipid metabolism and disturbances of myeline formation within depression-related brain regions contribute to chronic stress-induced behavioral disorders. Furthermore, we seek to elucidate the link between lipid metabolism and demyelination while exploring potential underlying mechanisms.

Material and methods

Mice

The animal experiments and related procedures were endorsed by the Ethics Committee of Shandong University (approval number: ECSBMSSDU − 2022–––2 − 65), and were strictly conducted in accordance with the specifications issued by the Institutional Animal Care and Use Committee. Male C57BL/6J mice, aged 5 to 6 weeks, were sourced from the Shandong University Animal Centre.

Animal models for depression and demyelination

Chronic unpredictable mild stress (CUMS) model

The CUMS model is a widely recognized depression model. After one week of adaptation, male C57BL/6J mice underwent daily exposure to mild and unpredictable stressors for five consecutive weeks., as previously described. The continuous series of mild stressors included a 24-hour light–dark reversal, 24-hour fasting and/or water fasting, 24-hour cage tilting at 45° angle, 24-hour wet bedding exposure, restraint stress lasting for 4–5 h, odor irritation for 24 h, tail clipping for one minute, and noise stimulation for 30 min.

The order of these stressors was randomized with intermittent repetition.

Corticosterone (CORT) model

We adopted a CORT administration approach to induce depressive-like behavior in mice, with modifications based on previously reported methodologies. Specifically, CORT (HY-B1618, MCE) was dissolved in drinking water at concentrations of 25 mg/L, 12.5 mg/L, and 6.25 mg/L. Male C57BL/6J mice were administered daily with CORT-infused drinking water after one week of acclimatization.

The concentration was altered on a weekly basis, and follow-up experiments were conducted after three weeks.

Lysophosphatidylcholine (LPC) model

Focal demyelination is induced by the injection of 1% LPC (L4129, Sigma Aldrich) into the bilateral hippocampal CA1 region. Specifically, after anesthetizing with 1% barbiturates, the mouse's skull was exposed and it was then positioned in a stereotaxic apparatus (Stoelting, USA). The injection was delivered at a flow rate and following needle withdrawal, the overlying skin was sutured, animals were placed on a heat pad for anesthesia recovery.

The mice were subsequently housed for 3 days to facilitate subsequent analyses.

Cuprizone (CPZ) model

In 5-week-old mice, demyelination was induced by administering the toxin copper ketone (bis-cyclohexanone-oxalic dihydrazone) at a concentration of 0.2% (w/w). The toxin was supplied in the form of prefabricated pellets (Aladdin) and mixed into powdered feeds after one week of acclimatization. Behavioral tests were conducted after 12 weeks of rearing.

Drug administration

Between 09:30 and 10:30 each morning, saline or pharmacological agents were administered for consecutive days.

Clemastine treatment

Mice were administered clemastine (YZ-100229, Solarbio) at a dosage of 10 mg/kg/day via intraperitoneal injection during the final two weeks of CUMS and CORT procedures. The selection of this dose was based on previous studies that demonstrated its efficacy in promoting myelination.

Treatment with cyclodextrin ((2-hydroxypropyL)-β-cyclodextrin, CD)

The treatment group received subcutaneous injections of Cyclodextrin in conjunction with the CUMS procedure, administered three days after adeno-associated viral (AAV) infection. Cyclodextrin (HY-101103, Sigma Aldrich) was subcutaneously injected twice a week at a dose of 2 g/kg along with PBS. Behavioral assessments were performed eight weeks following the injection.

T0901317 treatment

The treatment group received subcutaneous injections of T0901317 along with the CUMS procedure. T0901317 (GC10596, GLPBIO) was administered twice a week at a dose of 50 mg/kg via subcutaneous injection, and behavioral assessments were performed after eight weeks.

Mifepristone (RU486) treatment

During the last two weeks of CUMS, the treatment group mice received mifepristone (HY-13683, MCE) via intraperitoneal injection at a dosage of 25 mg/kg.

Behavior testing

Sucrose Preference Test (SPT)

The SPT was primarily utilized for evaluating anhedonia symptoms in animal subjects[22]. Each mouse was housed individually in a separate cage. As part of the 24-hour adaptation phase, two bottles each containing a 1% sucrose solution were provided to them.

Subsequently, the mice were given one bottle of sugar water and one bottle of drinking water next day, with positions alternated every 12 h. After a 24-hour food and water deprivation period, two bottles were respectively filled with a sucrose water and drinking water for another 24 h, again with positions changed every 12 h. Throughout the entire test period, consumption amounts of both sugar water and regular water were recorded.

Sucrose preference was defined as sucrose consumption divided by [water consumption + sucrose consumption], multiplied by 100%.

Open Field Test (OFT)

The OFT was employed to evaluate baseline locomotor activity and anxiety-related behaviors[23]. Each mouse was gently positioned in a 40 × 40 × 40 cm white wooden arena and given 5 min to freely explore. The floor of the arena was sectioned into a 20 × 20 cm central zone and a peripheral zone.

Video tracking system, TopScan, along with its accompanying software, was employed for data recording purposes. SMART analysis software was then utilized to assess the time mice spent in the central area and their total movement distance during the test. Before each new trial, the open − field surface was wiped down with a 75% ethanol solution to get rid of any odors left by previous test subjects.

Elevated Plus Maze (EPM)

The EPM is used to quantify anxiety-related behaviors in mice. Each mouse was placed in the central area of the EPM and allowed to freely explore for five minutes. Movement data was analyzed using the Top scan video tracking system along with its accompanying software.

Prior to each experiment, the accessible sections of the maze were sanitized with 75% ethanol and ensured to be dry.

Forced Swimming Test (FST)

The forced swim test (FST) was employed to evaluate behavioral despair in mice. As reported previously, each mouse was individually placed into a glass tank (18.45 cm in height, 14.5 cm in diameter) filled with water at a depth of 15 cm and maintained at 25 ± 1°C. The immobility time during the last 4 min of the 6-minute test was recorded and analyzed.

Tail Suspension Test (TST)

The TST was used to assess the state of despair in mice, as indicated by their immobile behavior following suspension by the tail. Briefly, individual mice were suspended by paper tape placed abouot 1 cm from the tail tip using a TST apparatus. Their behavior was then recorded for a duration of 4 min after an initial adaptation period of 2 min.

It should be noted that data from any mouse observed grabbing its tail with its paw was excluded from analysis.

Puzzle box

The test site comprised a white plexiglass chamber separated into two sections by a removable partition: a bright starting zone and a smaller covered dark compartment with bedding and food at the base. For each test, mice were placed in the bright starting zone and permitted to access the dark area through a narrow passage (4 x 4 cm) beneath the barrier. A total of nine tests (T1-T9) were conducted over 3 days, progressively increasing the difficulty of barriers set up in the underground passageway for mice.

On day 1, no barriers were present in the underpass, and during T1, there was an open door above it. In T2 and T3, there was no door on the barrier and mice entered through a small underpass. On day 2, T4 replicated conditions from T2 and T3; however, wood shavings surrounded the underpasses that mice had to dig through in T5 and T6.

On day 3, conditions from T5 and T6 repeated in T7 while cardboard blocked access for mice in both T8 and T9, requiring mice to use their teeth and claws to enter the dark area. Mouse performance in this puzzle-box paradigm was assessed by recording the latency time to enter the dark area.

Stereotaxic surgery and virus injection

The virus used for this experiment included: pSLenti-MBP-Osbpl6-3xFLAG-P2A-mCherry-WPRE (1.99 × 108v.g./mL, OBiO), pSLenti-MBP-mCherry-3xFLAG-WPRE (3.06 × 108v.g./mL, OBiO); AAV9-MBP-mCherry-mir30-shOsbpl6 (4.62 × 1013v.g./mL, Vizion) and AAV9-MBP-mCherry (9.69 × 1013v.g./ml, Vizion Bio).

Brain stereotaxic surgery was performed as previously described (7). Firstly, mice were anesthetized with sodium pentobarbital and secured in a stereotaxic apparatus (Stoelting, USA). Subsequently, the virus was bilaterally injected into the hippocampal CA1 regions (AP: −2 mm from the fontanel point; ML: ± 1.5 mm; DV: −1.5 mm) using a micro volume at a rate of 0.1 μL/min (LV) or 0.05 mL /min (AAV) via an electrically operated microinjector pump (Stoelting, USA).

Following injection, the syringe remained at the site for 10 min before slow retraction to prevent viral backflow. The incision was then closed and the mouse placed under a heating pad for anesthesia recovery. Further studies were conducted after 7 days (LV) or 21 days (AAV) postoperatively.

The accuracy of injection sites was verified through immunofluorescence assay, and only data from animals with correctly positioned injections were included in subsequent experiments.

Electron microscopy

The animals were anesthetized and then euthanized. Hippocampal CA1 tissues were immediately immersed in a 2.5% glutaraldehyde solution for overnight fixation. The tissues were subsequently dehydrated through a graded ethanol series, followed by overnight immersion in propylene oxide and embedding in 100% Epon resin.

Ultrathin sections (60 nm thickness) were prepared using an ultrathin microtome (Leica), stained with 2% uranyl acetate for 20 min, f and then with 0.5% lead citrate for 5 min. Images were captured using a transmission electron microscope (Philips Tecnai 20 U-TWIN, The Netherlands) at an accelerating voltage of 120 kV. The inner and outer diameters of myelin fibers were determined using ImageJ software.

The myelin g-ratio in the hippocampus was calculated as the ratio of the inner diameter to the outer diameter of myelin fibers.

Local Field Potential (LFP) recording in freely moving mice

Mice were anesthetized with sodium pentobarbital, followed by the implantation of microdrills into the hippocampus at the following coordinates: 2 mm anteroposterior (AP) and 1.5 mm mediolateral (ML) relative to bregma. The quadrupole tip was advanced to a depth of 1.5 mm. Postoperative mice were housed in home cages for recovery.

Neuronal activity was monitored using an AR recording system (Quanlan, Shanghai), and the power of local field potentials (LFP) was analyzed via AR-analyzer software.

Single nucleus RNA-sequencing library preparation

The concentration of single nucleus suspension was adjusted to 3–4 × 105 nuclei/mL in PBS. Library preparation was performed using a microfluidic chip (GEXSCOPE® Single Nucleus RNA-seq Kit, Singleron Biotechnologies) and single nucleus RNA-seq libraries were constructed following the manufacturer’s instructions (Singleron Biotechnologies). The generated snRNA-seq libraries were sequenced on an Illumina HiSeq X10 instrument with 150 bp paired end reads.

The screening criteria for differentially expressed genes (DEGs) in our study are defined as P < 0.05 (statistical significance) and |avg_log2FC| > 0.

Gene expression profile data collection

Using the GPL10907, GPL16791 and GPL21273 platform, three microarray datasets (GSE24095, GSE248260, GSE176140) comprising depression and control samples were retrieved from the GEO database. The GSE24095 data set contained gene expression profiles of 15 depression patients and 15 normal controls. The GSE248260 data set included 18 samples, with 9 depression patients and 9 were controls.

The GSE176140 dataset consisted of 6 samples, among which 3 were depression model mice.

qPCR

The mice were anesthetized, and bilateral hippocampal CA1 tissue was excised and stored at −80°C for subsequent use. RNA was extracted using the Fast Pure Tissue Total RNA Isolation Kit, and the A260/A280 ratios as well as concentrations of all RNA samples were measured using a spectrophotometer from Thermo Fisher Scientific. Quantitative PCR cDNA synthesis for quantitative PCR was conducted using the Sure Script™ First-Strand cDNA Synthesis Kit (catalog QP056, GeneCopoeia™).

Subsequently, quantitative real-time fluorescent PCR was then performed on the StepOnePlus Real-Time Fluorescent PCR System (Thermo Fisher Scientific) along with BlazeTaq™ SYBR Green qPCR Mix 2.0 (catalog QP031, GeneCopoeia™).

Specific primer sequences were used for the amplify LXR (forward primer: 5′- AGGGATAGGGTTGGAGTC −3′; reverse primer:5′- TTCAAGCGGATCTGTTCT −3′)and GAPDH (forward primer: 5′- TCTCTGCTCCTCCCTGTTC-3′; reverse primer: 5′-ACACCGACCTTCACCATCT-3′). Polymerase chain reaction data were analyzed using CFX Connect software, and quantified mRNA levels were normalized to GAPDH.

Immunofluorescence staining

Mice were anesthetized with sodium pentobarbital and then subjected to transcardial perfusion with ice-cold saline followed by 4% paraformaldehyde. For immunofluorescence staining, prior to hyalinization, brain sections were first washed three times with PBS. They were then blocked in a solution consisting of PBS, 0.2% Triton X-100, 0.25% BSA, and 5% donkey serum for one hour at room temperature.

The sections were subsequently incubated overnight at 4°C with primary antibodies, followed by three PBS washes the next day. Subsequently, fluorescence-conjugated secondary antibodies were applied to sections for 1 h at room temperature in the dark. After three PBS washes, DAPI (Beyotime C1002; 1:1000) counterstaining was performed, and images were acquired using a high-speed confocal platform.

For analysis, 1–3 images per animal (n = 3) were selected with ImageJ 1.8.0. Primary antibodies used in the immunofluorescence experiments included: mouse anti-MBP (Servicebio, GB12226, 1:800); mouse anti-Caspr (Santa Cruz Biotechnology, sc-374489, 1:250); rabbit anti-Caspr (Cell Signaling Technology, 97736, 1:800); rabbit anti-Nav1.6 (Affinity, DF8577, 1:250); rabbit anti-Neurofilament-L (Cell Signaling Technology, 2837, 1:100); mouse anti-OSBPL6 (Gene Tex, GTX107723,1:500); rabbit anti-Olig2 (Proteintech, 13999–1-AP, 1:500); mouse anti-Olig2 (Santa Cruz Biotechnology, sc-293163, 1:250); mouse anti-Perilipin3 (Santa Cruz Biotechnology, sc-390968, 1:250); mouse anti-GR (Santa Cruz Biotechnology,sc-56851, 1:250); EEA (Santa Cruz Biotechnology, sc-137130, 1:250); APP(β-Amyloid, Santa Cruz Biotechnology, sc-28365, 1:250). The following secondary antibodies included: CoraLite488 goat anti-mouse IgG (Proteintech, SA00013-1; 1:1000); CoraLite488 goat anti-rabbit IgG (Proteintech, SA00013-2; 1:500); Alexa Fluor 568 donkey anti-rabbit IgG (Invitrogen, A10042,1:500) Alexa Fluor 647 donkey anti-rabbit IgG (Invitrogen, A32733,1:500) and Alexa Fluor 594 goat anti-mouse IgG (Proteintech, SA00006-3; 1:500), ER-Tracker Blue-White DPX (E12353, Thermo Fisher Scientific).

Bodipy-Cholesterol staining

For Bodipy staining, brain sections were hyalinized and were incubated overnight at 4°C with 1 μg/mL Bodipy-Cholesterol (MCE, HY-125749) in PBS containing 0.1% TritonX-100, 10% donkey serum albumin and primary antibody. Remaining steps followed the immunofluorescence staining protocol.

ELISA for mouse CORT and adrenocorticotropic hormone (ACTH)

Blood samples were centrifuged for 20 min at a speed of 2500 rpm. The resulting serum was carefully collected for subsequent analysis. Serum levels of CORT and ACTH were measured using an ELISA kit (COIBO BIO, China) following the manufacturer's protocols.

The concentration of each sample was determined using the linear equation generated from the standard calibration curve.

Luxol Fast Blue (LFB) staining of myelin

The LFB Myelin Staining Kit (Firm Blue-Tar Violet Method) is utilized to assess the integrity of myelin sheaths, as well as the extent of degeneration, necrosis, and repair in pathological conditions[6]. The Myelin Staining Kit (Solarbio, G3245) comprises Reagent A (Firm Blue Staining Solution), Reagent B (Firm Blue Differentiation Solution), and Reagent C (Tar Violet Staining Solution). Following the kit instructions with slight modifications, the specific steps included: (1) fixation in 95% ethanol for 1 min; (2) staining in liquid A at 60°C for 30 min; rinsing in 95% ethanol followed by distilled water for 3–5 times; (3) staining in liquid B for 15 s, differentiating color in 75% ethanol twice for 5 s each time, rinsing with distilled water until clear delineation of gray matter and white matter was achieved; (4) staining with solution C twice for 1 min each time; rinsing with absolute ethanol twice for 1 min each time; (5) treatment with xylene I and xylene II for 1 min each, followed by sealing with neutral gum.

Filipin Ⅲ staining

Cholesterol accumulation in neurons was assessed using Filipin III[21], [24]. Tissue slices were incubated with 50 μg/mL Filipin III (Sigma Aldrich, HY-N6718) for 1 h at RT in the dark, followed by three washes and blocking.

Cholesterol quantifications

The fresh hippocampal CA1 tissue was lysed and treated with the lysate from a Tissue total cholesterol Kit (Applygen, E1015) for cholesterol analysis. Specifically, the homogenate mixture was rapidly frozen and enzymatically reacted according to the kit protocol. Cholesterol esterase was excluded when measuring free cholesterol using the same kit.

The measured values were normalized using cholesterol standards with known concentrations provided in the kit. The measurement was performed using an enzyme marker. E1015 and E1016 kits were utilized to determine total cholesterol and free cholesterol respectively, while the difference between these two represented the content of cholesterol esters.

Western Blot

After inducing anesthesia in the mice using sodium pentobarbital, the bilateral hippocampal CA1 regions were promptly dissected on ice, preserved in EP tubes. For protein extraction, RIPA lysate (250 μL per 20 mg of tissue) was utilized, and a protease inhibitor PMSF and phosphatase inhibitor (RIPA: PMSF: PI = 100:1:1) were added and thoroughly mixed prior to grinding and centrifugation. Following centrifugation, the supernatant was collected.

Protein concentration was determined using the BCA Protein Assay Kit (Thermo Fisher Scientific, 23228). Lysates were diluted with 1x Sampling Buffer (1610374, Bio-Rad), boiled at 100°C for 5 min, then separated via 10%-12% SDS-PAGE gel electrophoresis system. Proteins were then transferred to a PVDF membrane and blocked with 5% nonfat milk in TBST buffer or 5% BSA (Solarbio, YZ-100229) for 1 h at RT.

Following this step, the membranes were incubated overnight with primary antibodies. Primary antibodies used were: mouse anti-MBP (Servicebio, GB12226, 1:8000); mouse anti-OSBPL6 (Gene Tex, GTX107723,1:1000); rabbit anti-actin (1:1000); mouse-anti-LXR (Santa Cruz Biotechnology, sc-271064, 1:1000) and mouse-anti-GAPDH (Proteintech, 60004–1-Ig, 1:5000). After three 10-minute washes in TBST, membranes were incubated with secondary antibodies, followed by three additional TBST washes Membranes were developed using ECL reagents (Vazyme, E412) and visualized with a Chemiluminescence Imager (Tanon-4600, China).

All obtained band photos were imported into Image J for analysis.

Luciferase assays

The synthetic sequence was double-digested and ligated into the pGL3-Basic vector to construct the reporter plasmid pGL3-Basic-LXRα promoter (5′-GAATGTTCTGGA-3′). Concurrently, GR overexpression plasmids and control plasmids were also prepared. When 293 T cells reached 60–80% confluence, each group was transfected with the reporter plasmid (pGL3-Basic-LXRα promoter), GR overexpression plasmid, and internal control plasmid PRI-TK.

48 h post-transfection, luciferase activity was measured using a dual luciferase assay system. Relative luciferase activity was calibrated against PRL-TK, and statistical analysis was performed to assess intergroup differences.

Chromatin immunoprecipitation (ChIP)

The interaction between GR and Osbpl6 or LXR was characterized using ChIP. ChIP detection was conducted using the Sonication ChIP Kit (ABclonal, RK20258). Bilateral hippocampal tissues were snap-frozen on dry ice and stored at −80°C until use.

Each ChIP sample consisted of a mixture of anatomical samples from 3 mice, resulting in a total of 2 ChIP samples being analyzed. The tissue mixture was minced on ice and cross-linked with 1% formaldehyde (Thermofisher, YH378157) for 15 min to stabilize the protein-DNA complex. Nuclear extraction and chromatin sonication were then performed using cell swelling buffer, cell sonication buffer, and ultrasonic disruption was applied to break the nuclear membrane and digest DNA into fragments averaging 200–900 bp in length.

Subsequently, the samples were incubated overnight at 4°C with anti-GR antibody (Proteintech, 24050–1-AP). Normal rabbit IgG (ABclonal, RM20712) and positive antibody H3 (ABclonal, RM20711) were included as controls in all experiments. Chromatin-antibody complexes were isolated using ChIP-grade protein A/G magnetic beads (RM002915, ABclonal), followed by sequential washing with low and high salt buffers.

Antibody-bound chromatin was eluted in 1 × ChIP elution buffer, subjected to reverse cross-linking and treated with protease K. DNA purification was carried out using the kit-provided centrifuge column. Purified DNA samples underwent qPCR analysis using specific primers targeting OsbplS6 promoter: forward primer: 5′-AGGTGTGGGTCTGGGTAGTT-3′; reverse primer: 5′-ACACATGGCAGTCTCACCC-3′; as well as LXR promoter: forward primer: 5′- TCAAGCCCTGCTTCTCACAT-3′; reverse primer: 5′- AGGGAGCGCCAAGAGTAAAC-3′[25].

Real-time fluorescence quantitative PCR data obtained from ChIP analysis were analyzed using △△Ct method, normalized to input values. The relative occupancy ratios of genomic regions across experimental groups are shown in the previously reported graph.

Co-Immunoprecipitation (CO-IP)

The classic IP kit from Pierce was employed for CO-IP analysis to determine the interaction between GR and LXR proteins. Hippocampal CA1 protein lysate (total protein content 600–700 μg) was incubated with GR antibody (24050–1-AP, Proteintech), LXR antibody (Santa Cruz Biotechnology, sc-271064), and normal mouse IgG (BD0050, Bioworld) at 4°C for 24 h on a rotating platform. Subsequently, Protein A agarose beads (BD0048, Bioworld) were added and incubated at 4°C for 7 h.

Finally, sample buffer was added and subsequent steps followed the Western Blot protocol.

Statistical analysis

Data comparison between two independent groups was performed using a two-tailed unpaired Student’s t-test (normal distribution). For multiple groups, one-way ANOVA (normal distribution) followed by Tukey’s post hoc test was used. Pearson correlation analysis assessed correlations.

Data were assumed to follow a normal or counterbalanced distribution. All replicates were biological. For data collection and analysis, experimenters were blinded to genotype while performing behavioral testing or image analysis.

Sample sizes were not predetermined statistically but aligned with those in prior publications and a power analysis was conducted using G*Power software. Data are presented as mean ± SEM.

Result

Demyelination is observed in mouse models of depression

To investigate the relevance between depression and demyelination, we utilized the CUMS (chronic unpredictable mild stress) and CORT (Corticosterone) models as previously described[26], [27]. Behavioral tests were conducted on day 35 of the CUMS procedure (Fig. 1A).

CUMS mice displayed depressive-like behaviors: reduced sucrose preference rate in the SPT (Fig. 1B), significantly prolonged immobility time in both TST (Fig. 1C) and FST (Fig.

1D); no significant difference in total distance traveled during OFT, but compared with controls, CUMS mice spent less time in the central area (Fig. 1E-G). In the behavioral tests of CORT mice, comparable depression-like behaviors were observed (Fig.

S1A-G). Myelin basic protein (MBP), which is widely utilized as a cellular marker for myelin [6], showed decreased protein levels within hippocampal regions of depressed mice compared to control animals (Fig. 1H-I, Fig.

1S-T and Fig. S1H-K). These findings indicate that demyelinating phenotype might be correlated with s depression-related behaviors.Fig.

1Demyelination was observed in CUMS mice A, Schematic outline of procedure for establishing a CUMS model in mice. B-G, The depression-like behaviors of control and CUMS mice: B, For the SPT, n = 9 mice, P < 0.0001. C, For the TST, n = 9 mice, P = 0.0006.

D, For the FST, n = 9 mice, P = 0.0001. E, Heat-maps of locomotor activity in the OFT. F, Statistical analysis of total distance; n = 9 mice, P = 0.1273.

G, Time spent in the center zone, n = 9 mice, P < 0.0001. H, Immunostaining analysis of MBP protein level in brain from control and CUMS mice. Scale bar: 1 mm.

Panels on the right is higher-magnification images of the hippocampus. Scale bar: 100 μm. I, Quantification of MBP fluorescence intensity in the hippocampus, n = 6 slices from six mice, P = 0.002.

J, Representative images showing Caspr-positive, green-stained paranodal regions and Nav1.6-positive, red-stained paranodal regions in the hippocampal neurons. Scale bar: 10 μm. n = 6 slices from six mice. Panels on the right is high-magnification images of Caspr staining from the left.

Scale bar: 1 μm. K, Nodal lengths were increased in CUMS mice, based on measurements of Caspr-stained regions, n = 20 nodes from six mice, P = 0.0003. L, Histograms showing the frequency distribution of nodal length, which differed between control and CUMS mice.

M, The electron microscopic images showing demyelination in the hippocampus in CUMS mice. Scale bar: 500 μm. Panels at the bottom is higher-magnification images of the hippocampus.

Scale bar: 100 μm. N, The percentage of myelinated axons reveals fewer myelinated fibers in CUMS mice, n = 6 slices from six mice, P < 0.0001. O, Thinner myelin sheaths were observed in CUMS mice as measured by electron microscopy, n = 8 images from six mice, P < 0.0001.

P, The g-ratio of the inner to outer diameter of myelin sheaths plotted against the axon diameter, n = 15 and 14 from six mice, P < 0.0001. Q, Images of APP immunofluorescence for axonal damage. Scale bars, 50 µm.

R, Reduced number of APP + axons in CUMS compared to controls mice, n = 6 slices from six mice, P = 0.0029. S, Representative immunoblots of MBP in the control and CUMS mice. T, Quantification of relative MBP level of the immunoblots, n = 6 mice, P = 0.0097.

Statistical differences were determined by two-tailed unpaired Student’s t-test (F, G, I, N, O, P, R) and unpaired two sample t-test with Welch's correction (B, C, D, K, T). Data are presented as mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Fast nerve conduction depends on the myelination of neuronal axons by glial cells and the aggregation of high-concentration voltage-gated sodium channels (Nav) in the inter-segmental axonal spaces, which are termed nodes of Ranvier[28]. To assess the paranodal junctions in neurons, contactin-associated protein (Caspr) and Nav1.6 were used, with nodal length quantified relative to the boundaries of Caspr-positive regions. Compared to control mice, depressive mice exhibited a significantly longer average nodal length.

(Fig. 1J-K and Fig. S1L-M) and a broader distribution of longer nodal lengths (Fig.

1L and Fig. S1N). Electron microscopy results revealed a reduced number of myelinated axons in depressive mice, accompanied by thinner myelin as quantified by the ratio between myelin thickness and axonal diameter (g-ratio) (Fig.

1M-P and Fig. S1O-R). Additionally, we observed aggregation of app proteins in the hippocampus of depressed mice, indicating neuronal degeneration (Fig.

1Q-R). These findings demonstrate that chronic stress not only impairs the morphology but also compromises the structure of myelin sheaths in depressive mice.

The demyelination mouse model exhibited depressive behavior

To confirm the relationship between demyelination and depression, we subsequently employed the CPZ (Cuprizone) to induce a demyelination model[29], [30]. CPZ is a selective chelating agent of copper ions, which can target damaged mitochondria of oligodendrocytes and induce myelin loss in the central nervous system[31]. The mice were fed a chow diet containing 0.2% (w/w) cuprizone for a 12 weeks,which could induce complete onset of chronic demyelination[15].

Additionally, we utilized the LPC (Lysophosphatidylcholine) model to induce focal demyelination as reported [32], [33] (Fig. 2A,2L). As an additional assessment of learning and executive function associated with demyelinating cognitive dysfunction, we employed the puzzle box test − a classic test in this field[34] (Fig.

S2A). Our results demonstrated that both CPZ and LPC mouse models exhibited significantly prolonged latency time when entering the target area (Fig. 2B,2M).

Interestingly, we observed depression-like behaviors in these demyelinated mice (Fig. 2C-2E and Fig. 2N-2P).

However, no significant alterations were detected in their movement ability or anxiety-like behaviors (Fig. S2B-G). Concurrently, a notable reduction in myelin basic protein levels was observed in the hippocampal region of these mouse models (Fig.

2F-I and Fig. 2Q-T), while LFB staining also revealed destruction of myelin structure within the demyelinated mouse models' brains (Fig. 2J-K).Fig.

2Depression-like behaviors was observed in demyelination mouse model A, Schematic outline of procedure for establishing a 0.2% CPZ-chowing model in mice. B-E, The depression-like and learning-related behaviors of control and CPZ mice: B, For the Puzzle box test, n = 6 mice, P = 0.0083 (T1), P = 0.0872 (T2), P = 0.0519 (T3), P = 0.0319 (T4), P = 0.0008 (T5), P = 0.0234 (T6), P = 0.0165 (T7), P = 0.0379 (T8), P = 0.0337 (T9). C, For the SPT, n = 9 mice, P < 0.0001.

D, For the TST, n = 9 mice, P < 0.0001. E, For the FST, n = 9 mice, P = 0.0002. F, Representative immunoblots of MBP within hippocampus from mice.

G, Quantification of relative MBP level of the immunoblots, n = 6 mice, P = 0.0008. H, Immunostaining analysis of MBP protein level in brain sections from control and CPZ mice. Scale bar: 100 μm.

I, Quantification of MBP fluorescence intensity in the hippocampus, n = 6 slices from 3 mice, P = 0.0013. J, Representative images of LFB (Luxol Fast Blue) histological staining. Scale bar: 1 mm.

Panels at the bottom is higher-magnification images of the hippocampus. Scale bar: 10 μm. K, Quantification results of LFB staining of the hippocampus n = 6 slices from six mice, P < 0.0001.

L, Schematic outline of procedure for establishing a LPC model in mice. M−P, The depression-like and learning-related behaviors of control and LPC mice: M, For the Puzzle box test, n = 6 mice, P = 0.0052 (T1), P = 0.0203 (T2), P = 0.0416 (T3), P = 0.0346 (T4), P = 0.0053 (T5), P = 0.0234 (T6), P = 0.0324 (T7), P = 0.0134 (T8), P = 0.0325 (T9). N, For the SPT, n = 9 mice, P < 0.0001.

O, For the TST, n = 9 mice, P < 0.0001. P, For the FST, n = 9 mice, P = 0.0002. Q, Representative immunoblots of MBP in the control and LPC mice.

R, Quantification of relative MBP level of the immunoblots, n = 6 mice, P < 0.0001. S, Immunostaining analysis of MBP protein level within hippocampus from mice. Scale bar: 100 μm.

T, Quantification of MBP fluorescence intensity in the hippocampus, n = 6 slices from three mice, P < 0.0001. Statistical differences were determined by two-tailed unpaired Student’s t-test (B, C, D, E, G, I, K, M, N, O, P, R, T) and unpaired two sample t-test with Welch's correction (B, M). Data are presented as mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

To further validate the crucial role of myelin homeostasis in depression, a dosage of 10 mg/kg clemastine was administered to a mouse model of depression for a duration of two weeks (Fig. 3A and Fig. 3I).

Clemastine has been widely recognized to promote myelin regeneration[35], and it has been reported to be somewhat effective in saving depressed behavior[6].As anticipated, pharmacological remyelination induced by clemastine effectively ameliorated depressive and anxiety-like behaviors (Fig. 3B and C, J-K and Fig. S3A-D,S3J-K), while also reducing the latency time required to enter the target area in the puzzle box test (Fig.

3D,3L). Concurrently, clemastine treatment led to an upregulation in MBP expression that correlated with the rescue of behavioral abnormalities (Fig. 3E-H and Fig.

3M-P). Importantly, compared with control mice, CUMS mice exhibited reduced power density, indicating diminished neural network synchronization. Crucially, subsequent treatment with clemastine rescued the CUMS-induced reduction in power density, ultimately restoring impaired neural network synchronization to some degree (Fig.

S3E-G). These suggested that restoring the structure of myelin leads to improvements in both its morphology and the function of neurons.Fig. 3Clemastine rescues demyelination and depression-related behaviors in depression mouse model A, Schematic outline of clemastine treatment experiment in CUMS mice.

B–D, Behavioral testing of CUMS mice and clemastine treatment: B, For the SPT, n = 9 mice, P < 0.0001, P = 0.0084, P = 0.9523. C, For the FST, n = 9 mice, P < 0.0001, P = 0.0003, P = 0.9997. D, For the Puzzle box test, n = 6 mice/group, P = 0.9756 (T1), P = 0.8215 (T2), P = 0.9147 (T3), P = 0.3544 (T4), P = 0.0213 (T5), P = 0.0067 (T6), P = 0.0038 (T7), P = 0.0007 (T8), P = 0.0009 (T9).

E, Representative immunoblots of MBP in the CUMS and clemastine treatment mice. F, Quantification of relative MBP level of the immunoblots, n = 6 mice, P < 0.0001, P = 0.0032, P = 0.9859. G, Immunostaining analysis of MBP protein level in brain sections.

Scale bar: 100 μm. H, Quantification of MBP fluorescence intensity in the hippocampus, n = 6 slices from three mice, P = 0.0007, P = 0.0239, P = 0.9925. I, Schematic outline of clemastine treatment experiment in CORT mice.

J–L, Behavioral testing of CORT mice and clemastine treatment: J, For the SPT, n = 9 mice, P < 0.0001, P = 0.0005, P > 0.9999. K, For the FST, n = 9 mice, P < 0.0001, P = 0.0007, P = 0.7973. L, For the Puzzle box test, n = 6 mice, P = 0.0368 (T1), P = 0.1084 (T2), P = 0.3741 (T3), P = 0.0062 (T4), P = 0.009 (T5), P = 0.0018 (T6), P = 0.0007 (T7), P = 0.0129 (T8), P < 0.0001 (T9).

M, Representative immunoblots of MBP in the CORT and clemastine treatment mice. N, Quantification of relative MBP level of the immunoblots, n = 6 mice, P = 0.0002 and 0.0001, P = 0.9926. O, Immunostaining analysis of MBP protein level in brain sections.

Scale bar: 100 μm; original magnification. P, Quantification of MBP fluorescence intensity in the hippocampus, n = 6 slices from three mice, P = 0.0009, P = 0.0124, P = 0.5963. Statistical differences were determined by one-way ANOVA with Turkey’s post hoc comparison (B, C, D, F, H, J, K, L, N, P) and Brown-Forsythe ANOVA with Tukey's multiple comparisons test (D, L).

Data are presented as mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

The expression of OSBPL6 in oligodendrocyte is down-regulated in the hippocampus of a mouse model of depression

In the CNS, myelin sheath is a multilayer lipid membrane formed by oligodendrocytes surrounding nerve axons[16]. Cholesterol in the myelin sheath has been demonstrated to play a vital role in myelin sheath growth during brain maturation[19]. To further identify the specific molecular mechanism that may underlying demyelination in depression, we performed single-nucleus RNA sequencing (snRNA-seq) of hippocampi of CUMS and control mice (Fig.

4A-C and Fig. S4A-F). Notably, there was a high correlation between oligodendrocytes and neurons (Fig.

4D and Fig. S4G). Pathway analysis of DEGs (different expression genes) between ctrl and CUMS hippocampal Oligodendrocyte included cholesterol homeostasis and intracellular lipid transport, which may relate to the structure of myelin sheath (Fig.

4E). Further analysis of DEGs revealed both up-regulated and down-regulated genes. Volcano plot filtering displayed differential expression levels in the oligodendrocytes of hippocampus, with a significant decrease observed for oxysterol binding protein-like 6 (Osbpl6) in CUMS mice (Fig.

4F and Fig. S4H-I). We focused on the change of other genes related to lipid metabolism, including Sms, Gnpat, Ugcg, Agps (related to synthesis of sphingomyelin and sphingolipid); Fasn, Pltp, Pex5 (related to synthesis of fatty acid); Hmgcs1, Fdft1, Fdps, Mvk (related to synthesis of cholesterol) and Apoe, Abca1, Abcg1(related to transport of lipid), and found the change of Osbpl6 is most significantly (Fig.

4F). OSBPL6, as member of the oxidative sterol-binding protein family, belongs to the intracellular lipid receptor category [36]. Gene Ontology (GO) analysis indicated that OSBPL6 was enriched in biological processes associated with lipid transport, cholesterol transport, intermembrane lipid transfer pathways, and cholesterol binding (Fig.

4G), suggesting its critical role in regulating intracellular cholesterol transport and efflux. Then, we further determined the reduced protein levels of OSBPL6 in hippocampal oligodendrocytes of mice exposed to CUMS. mice hippocampal oligodendrocytes subjected to CUMS (Fig. 4H-I).

Immunostaining analyses also provided additional evidence that OSBPL6 expression was downregulated, along with reduced MBP expression in the hippocampus of depression-induced mice (Fig. S5A-B). These results demonstrate a potential association between demyelination and deficiency of OSBPL6.Fig.

4SnRNA-seq of hippocampus reveals the deficiency of osbpl6 in oligodendrocytes of CUMS mice A, Uniform manifold approximation and projection (UMAP) representation of nuclear transcriptomes (each dot = one nucleus). Sequenced nuclei clustered into broad cell classes and were subsequently mapped onto specific cell types. B, Proportion of each cell type detected across the different groups.

C, The numbers of different expression genes (DEGs) within each cell type. D, The interaction network map between all cell types showing oligodendrocytes is closely linked to neurons, network nodes are cell types, network edge thickness is the total number of ligand and receptor pairs, and line color is consistent with the type of ligand cells. E, GO (Gene Ontology) analysis of DEGs between control and CUMS mice hippocampal oligodendrocytes.

F, Volcano plot showing DEGs as a function of log2 fold change and statistical significance (two-sided P value adjusted for multiple comparisons) between control and CUMS mice (n = 8 mice / group). The down-regulated Osbpl6 gene and other genes related to lipid metabolism has been labelled, P = 0.03.G, The bar plot of GO analysis for OSBPL6 showed correlation to the regulation of lipid metabolism. BP, biological process.

MF, molecular function. H, Immunostaining analysis of OSBOL6 and Olig2 in control and CUMS mice brain hippocampus. Scale bar: 20 μm.

I, Relative quantification of OSBPL6+ and Olig2+ cells in the hippocampus, n = 5 slices from three mice, P = 0.0105. Statistical differences were determined by two-tailed unpaired Student’s t-test (I). Data are presented as mean ± s.e.m. *P < 0.05.

Since myelin sheath formation heavily relies on cholesterol in oligodendrocytes[37], we hypothesized whether disturbances in cholesterol transport are associated with depression. In CUMS mice, both esterified and unesterified cholesterol were elevated compared to controls (Fig. S5C-D), with subcellular localization revealing increased cholesterol accumulation in early endosomes and the endoplasmic reticulum (ER) (Fig.

S5E-G)—indicating disrupted cholesterol transport between endosomes, ER and the plasma membrane. Bodipy-cholesterol staining further showed enhanced cholesterol accumulation in the cell body of oligodendrocytes in CUMS mice, with less accumulation along axon filaments relative to controls (Fig. S5H-K).

Consistent with this, Filipin III staining of hippocampal slices and lipid droplet staining both confirmed increased lipid accumulation within oligodendrocytes in CUMS mice (Fig. S6A-D). Together, these findings indicate disrupted cholesterol homeostasis in the cellular membrane system of CUMS mice.

OSBPL6 regulates cholesterol homeostasis and myelin structure in a mouse model of depression

To investigate the function of OSBPL6 deficiency in the development of depression, specifically its impact on cholesterol homeostasis in oligodendrocytes and subsequent damage to the myelin sheath, we employed both genetic knockdown and upregulation strategies. We hypothesized that increased OSBPL6 expression is essential for cholesterol homeostasis and myelination, thereby preventing demyelination and alleviating mouse depression-like behaviors.

To validate this, we selectively upregulated OSBPL6 within oligodendrocytes of the bilateral hippocampal CA1 region in CUMS mice (Fig. 5A). The upregulation of OSBPL6 alleviated depression-like behaviors (Fig.

5B-C and Fig. S7A-C, S7E-G), as well as improved learning and memory function in CUMS mice (Fig. S7D).

It also restored reduced hippocampal myelin levels (Fig. 5D-E) and immunostaining showed substantially more axons encapsulated by myelin sheaths in CUMS mice (Fig. 5F-G).

Additionally, OSBPL6 upregulation significantly reduced intracellular bodipy-cholesterol droplets (Fig. 5H-I).Fig. 5OSBPL6 in oligodendrocytes regulates cholesterol transport, myelin homeostasis and behaviors in mice.

A, Schematic demonstrating injection of pSLenti-MBP-Osbpl6-mCherry into oligodendrocytes in hippocampal CA1 to upregulate OSBPL6 in the oligodendrocytes in hippocampal CA1 of CUMS mice. Upper, the representative fluorescent images showing red fluorescent protein (mCherry) and green fluorescent protein (Olig2) expression in the hippocampal CA1. Scale bar: 20 μm.

B-C, The depression-like behaviors of mice: B, For the SPT, n = 9 mice, P < 0.0001. C, For the FST, n = 9 mice, P < 0.0001, P = 0.003. D, Representative immunoblots of OSBPL6 and MBP.

E, Quantification of relative MBP and OSBPL6 level of the immunoblots, n = 6 mice, P < 0.0001, P = 0.0007, P = 0.0004. F, Representative images showing MBP and NF immunoreactivity in mice. Scale bar: 20 μm.

G, The area of neuronal axon (NF+) for MBP was calculated using Image J, n = 5 slices from three mice, P = 0.0088. H, Immunofluorescence detection of cholesterol accumulation in Olig2+ cell. Scale bar: 10 μm.

I, Quantification of immunofluorescence detection of cholesterol in Olig2+ cell in mice, n = 5 slices from three mice, P = 0.0382. J, Schematic demonstrating injection of AAV9-MBP-mCherry-shOsbp16 into hippocampal CA1 to knockdown OSBPL6 in the oligodendrocytes in hippocampal CA1 of WT mice. Upper, the representative fluorescent images showing red fluorescent protein (mCherry) and green fluorescent protein (Olig2) expression in the hippocampal CA1.

Scale bar: 20 μm. K-L, The depression-like behaviors of mice: K, For the SPT, n = 9 mice, P = 0.0003. L, For the FST, n = 9 mice, P = 0.0045.

M, Representative immunoblots of OSBPL6 and MBP within hippocampus of mice. N, Quantification of relative MBP level of the immunoblots, n = 6 mice, P < 0.0001, P = 0.0411. O, Representative images showing MBP and NF immunoreactivity in mice.

Scale bar: 20 μm. P, The area of neuronal axon (NF+) for MBP was calculated using Image J, n = 5 slices from three mice, P = 0.0058. Q, Immunofluorescence detection of cholesterol accumulation in Olig2+ cell.

Scale bar: 10 μm. R, Quantification of immunofluorescence detection of cholesterol in Olig2+ cell in mice, n = 4 slices from three mice, P = 0.0005. Statistical differences were determined by two-tailed unpaired Student’s t-test (G, I, K, N, P, R), unpaired two sample t-test with Welch's correction (L) and by one-way ANOVA with Turkey’s post hoc comparison (B, C, E).

Data are presented as mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.001.

In contrast, specific knockdown of OSBPL6 in oligodendrocytes located in the hippocampal CA1 region of normal mice was associated with the emergence of depression-related behaviors (Fig. 5J-L and Fig. S7H-J, S7L-N) and impaired learning and memory (Fig.

S7K). Western blotting confirmed the knockdown of OSBPL6 (Fig. 5M-N).

Importantly, knockdown of OSBPL6 in oligodendrocytes also decreased the number of myelinated axons (Fig. 5O-P) while increasing cholesterol accumulation within these cells (Fig. 5Q-R).

These results suggest that deficiency in OSBPL6 within oligodendrocytes is associated with enhanced cholesterol accumulation and compromised myelination processes; however, normalization of OSBPL6 facilitate improvements in cholesterol transport capacity and mitigate stress-induced structural abnormalities in myelin sheaths.

Improved cholesterol transport rescues behavioral disorders by promoting remyelination in depressive mice

2-hydroxypropyL-β-cyclodextrin (CD) has been extensively used to alleviate cholesterol accumulation phenotypes by lowering the amount of cholesterol stored inside cells and promoting its transfer/relocation to other cellular membranes[38]. During the CUMS procedure, CD was administered to mice via hypodermic injection at a dose of 10 mg/kg twice weekly for 8 weeks (Fig. 6A), with the dose and duration based on previous reports[39], [40].

In the treatment group, administration of CD effectively ameliorated depression-like behaviors (Fig. 6B-E). However, anxiety-like behaviors assessed via the OFT showed no improvement (Fig.

S8A-C). Meanwhile, damaged myelin sheaths were restored after CD treatment (Fig. 6F-I) and significantly reduced cholesterol accumulation in oligodendrocytes, as quantified by bodipy-cholesterol staining (Fig.

6J-K). Moreover, it rescued depression-related behaviors and demyelination caused by OSBPL6-knockdown (Fig. 6L-V and Fig.

S8D-F). These results provide further evidence that cholesterol transportation is essential for maintaining myelination homeostasis and regulating depression-related behaviors.Fig. 6Cyclodextrin, which promotes cholesterol transportation, can rescue demyelination and depression-like behavior in mice A, Schematic outline of cyclodextrin (CD) treatment experiment in CUMS mice.

B-E, The depression-like and learning-related behaviors of CUMS and CUMS mice with cyclodextrin treatment: B, For the SPT, n = 9 mice, P = 0.0057. C, For the FST, n = 9 mice, P < 0.0001. D, For the TST, n = 9 mice, P = 0.0002.

E, For the Puzzle box test, n = 6 mice, P = 0.8864 (T1), P = 0.5862 (T2), P = 0.3624 (T3), P = 0.0638 (T4), P = 0.0448 (T5), P = 0.0565 (T6), P = 0.0006 (T7), P = 0.0344 (T8), P = 0.0438 (T9). F, Representative images showing MBP and NF immunoreactivity in CUMS and CUMS with cyclodextrin treatment. Scale bar: 20 μm.

G, The area of neuronal axon (NF+) for MBP was calculated using Image J, n = 5 slices from three mice, P = 0.011. H, Representative immunoblots of MBP in the CUMS and cyclodextrin treatment mice. I, Quantification of relative MBP level of the immunoblots, n = 6 mice, P = 0.0007.

J, Immunofluorescence detection of cholesterol accumulation in Olig2+ cell. Scale bar: 10 μm. K, Quantification of immunofluorescence detection of cholesterol in Olig2+ cell in mice, n = 5 slices from three mice, P = 0.0026.

L, Schematic outline of cyclodextrin treatment experiment in knockdown of Osbpl6 mice. M−P, The depression-like and learning-related behaviors of mice: M, For the Puzzle box test, n = 6 mice, P = 0.1221 (T1), P = 0.3996 (T2), P = 0.9823 (T3), P = 0.0193 (T4), P = 0.1072 (T5), P = 0.0892 (T6), P = 0.0152 (T7), P = 0.2729 (T8), P = 0.0016 (T9). N, For the SPT, n = 9 mice, P = 0.0013.

O, For the FST, n = 9 mice, P = 0.0075. P, For the TST, n = 9 mice, P = 0.0004. Q, Representative images showing MBP and NF immunoreactivity in mice.

Scale bar: 20 μm. R, The area of neuronal axon (NF+) for MBP was calculated using Image J, n = 5 slices from three mice, P = 0.0007. S, Representative immunoblots of MBP in mice.

T, Quantification of relative MBP level of the immunoblots, n = 6 mice, P = 0.0052. U, Immunofluorescence detection of cholesterol accumulation in Olig2+ cell. Scale bar: 10 μm.

V, Quantification of immunofluorescence detection of cholesterol in Olig2+ cell in mice, n = 5 slices from three mice, P = 0.0025. Statistical differences were determined by two-tailed unpaired Student’s t-test (B, C, D, E, G, L, K, M, N, O, P, R, V) and unpaired two sample t-test with Welch's correction (E, M, T). Data are presented as mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

The expression of OSBPL6 and cholesterol transport are inhibited by GR in a mouse model of depression

To elucidate the mechanisms behind the disorder of cholesterol transport in depression mice, we further investigated the potential signaling pathways responsible for the downregulation of OSBPL6 following chronic stress. We analyzed two independent human proteomic datasets and one independent mouse proteomic dataset and identified overlapping genes, notably including OSBPL6, which is downregulated in depression patients and aligns with our findings in mice (Fig. 7A and Fig.

S9A-C). Additionally, an independent human proteomic dataset highlighted overlapping genes, including CORT (Fig. S9A).

Pathway analyses of the overlapping genes from human proteomic datasets revealed differences in pathways associated with axonogenesis, hormone transport and secretion, and lipid metabolism (Fig. 7B). Similarly, pathway analyses of the overlapping genes from both human and mouse proteomic datasets indicated differences in pathways related to cholesterol transport and hormone secretion (Fig.

7C). Considering that hyperactivity of the hypothalamic–pituitary–adrenal (HPA) axis has been linked to depression, [41], we confirmed elevated serum CORT and upstream ACTH in CUMS mice (Fig. 7D-E).

Furthermore, we observed a moderate correlation between the relative expression of OSBPL6 protein and the levels of CORT and ACTH in the serum of mice (Fig. 7F-G), as well as between OSBPL6 expression and sucrose preference—supporting a close association with depression-like behaviors (Fig. 7H).Fig.

7Treatment of GR antagonist mifepristone rescue the demyelination and depression-like behaviors A, Venn diagram of Differentially Expressed Genes (DEGs) between depressive and control human and mice hippocampi from public database GSE24095, GSE248260 and GSE176140. B, Gene ontology pathway analysis (GO) of overlapped DEGs from human hippocampi from public database. C, Gene ontology pathway analysis of overlapped DEGs from human and mice hippocampi from public database.

D, The serum corticosterone (CORT) level of mice, n = 6 mice, P = 0.0027. E, The serum ACTH level of control and CUMS mice. n = 6 mice, P = 0.0057. F, Correlations of the serum CORT level and the relative expression of OSBPL6, for correlations, a simple linear regression was used for analysis, n = 12 mice, P = 0.0001, r2 = 0.7824.

G, Correlations of the serum ACTH level and the relative expression of OSBPL6, for correlations, a simple linear regression was used for analysis, n = 12 mice, P = 0.0003, r2 = 0.7465. H, Correlations of the preference of sucrose and the relative expression of OSBPL6, for correlations, a simple linear regression was used for analysis, n = 18 mice, P < 0.0001, r2 = 0.8219. I, Schematic outline of mifepristone (RU486) treatment experiment in CUMS mice.

J-M, The depression-like and learning-related behaviors of CUMS and CUMS mice with RU486 treatment: J, For the SPT, n = 8 mice, P < 0.0001, P = 0.0021. K, For the FST, n = 8 mice, P < 0.0001, P = 0.0059. L, For the TST, n = 8 mice, P < 0.0001, P = 0.0048.

M, For the Puzzle box test, n = 6 mice, P = 0.8638 (T1), P = 0.9722 (T2), P = 0.6931 (T3), P = 0.6586 (T4), P = 0.0004 (T5), P = 0.0097 (T6), P = 0.0016 (T7), P = 0.0083 (T8), P = 0.0097 (T9). N, Representative immunoblots of OSBPL6 and MBP in mice. O, Quantification of relative OSBPL6 and MBP level of the immunoblots, n = 6 mice, P = 0.0112, P = 0.0033, P = 0.0028, P = 0.007.

P, Immunofluorescence detection of cholesterol accumulation in Olig2+ cells. Scale bar: 20 μm. Q, Quantification of immunofluorescence detection of cholesterol in Olig2+ cell in mice, n = 5 slices from three mice, P = 0.0007, P = 0.008.

R, Immunofluorescence detection of cholesterol accumulation in MBP+ axons. Scale bar: 20 μm. S, Quantification of immunofluorescence detection of cholesterol in MBP+ axons in mice, n = 5 slices from three mice, P = 0.0002, P = 0.0042.

Statistical differences were determined by two-tailed unpaired Student’s t-test (E), unpaired two sample t-test with Welch's correction (D), one-way ANOVA with Turkey’s post hoc comparison (J, K, L, M, O, Q, S). Data are presented as mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, **** P < 0.0001.

Consistently, the antagonist of the glucocorticoid receptor (GR), mifepristone (RU486), significantly ameliorated depression-related behaviors and cognitive impairments (Fig. 7I-M and Fig. S9D-F).

Notably, the downregulation of OSBPL6 and MBP in CUMS mice was markedly reversed upon GR blockade with RU486 (Fig. 7N-O). Furthermore, the disruption of cholesterol transport induced by CUMS was also restored following GR inhibition (Fig.

7P-S). Collectively, these findings suggest that the decreased OSBPL6 expression and disrupted cholesterol homeostasis in CUMS mice are likely attributable to the hyperactivity of glucocorticoids (GCs) binding to their receptor GR in the hippocampal regions.

GR downregulated OSBPL6 expression via inhibiting LXR transcription in oligodendrocytes

Considering that GR functions as a nuclear transcription factor, it is proposed that elevated release of GCs during stress promotes the translocation of GR into the nucleus, subsequently inhibiting Osbpl6 transcription and leading to disruptions in cholesterol homeostasis and demyelination. Immunofluorescence staining revealed an increased nuclear translocation of GR in oligodendrocytes under chronic stress compared to control mice, which was significantly reduced by RU486 treatment (Fig. 8A).

However, ChIP analysis revealed no direct binding between GR and Osbpl6 promotor (Fig. 8B). Therefore, we investigated the interaction between GR and Liver X Receptor (LXR), a transcription factor known to directly regulate OSBPL6 expression[42].

Notably, the luciferase reporter assay and ChIP analysis demonstrated that GR can bind to the promoter of LXR directly and inhibit its transcription. (Fig. 8B and Fig.

S10A-B). Furthermore, RU486 treatment significantly reversed the decreased mRNA and protein levels of LXR caused by CUMS (Fig. 8C and Fig.

S10C-D). Additionally, Co-IP analysis indicated no protein binding between GR and LXR (Fig. S10E).Fig.

8GR downregulated OSBPL6 via inhibiting LXR transcription in oligodendrocytes and rescue the demyelination and depression-like behaviors A, Immunocytochemical detection of GR in Oligodendrocyte (Olig2+) shows nuclear localization in control, CUMS and CUMS with mifepristone (RU486) treatment mice. Scale bars: 10 μm. B, ChIP assays performed with the indicated antibodies (anti-GR) and hippocampus of CUMS mice.

Data for the Osbpl6 and LXR promoter are shown here, n = 3 independent experiments, P = 0.1552 and P = 0.0003. C, The mRNA level of LXR within hippocampus from mice, n = 6 mice, P = 0.0017 and P = 0.0007. D, Schematic outline of LXR agonist T0901317 treatment experiment in CUMS mice.

E-H, The depression-like and learning-related behaviors of mice: E, For the SPT, n = 9 mice, P = 0.0001. F, For the TST, n = 9 mice, P = 0.0001. G, For the TST, n = 9 mice, P < 0.0001.

H, For the Puzzle box test, n = 6 mice, P = 0.2511 (T1), P = 0.785 (T2), P = 0.7757 (T3), P = 0.5814 (T4), P = 0.2737 (T5), P = 0.0137 (T6), P = 0.0181 (T7), P = 0.0324 (T8), P = 0.0207 (T9). I, Representative immunoblots of OSBPL6 and MBP in mice. J, Quantification of relative OSBPL6 and MBP level of the immunoblots, n = 6 mice, P < 0.0001, P = 0.0007.

K, Representative images showing MBP and NF immunoreactivity in mice. Scale bar: 50 μm. L, The area of neuronal axon (NF+) for MBP was calculated using Image, n = 5 slices from three mice, P = 0.0007.

M, Immunofluorescence detection of cholesterol accumulation in Olig2+ cells. Scale bar: 20 μm. N, Quantification of immunofluorescence detection of cholesterol in Olig2+ cell in mice, n = 5 slices from three mice, P = 0.0056.

O, The less OSBPL6 transported intracellular cholesterol, resulting in the aggregation of both intracellular free cholesterol and esterified cholesterol in the endoplasmic reticulum, which led to demyelination and behavioral disorders. Statistical differences were determined by two-tailed unpaired Student’s t-test (B, F, G, H, J, L, N), unpaired two sample t-test with Welch's correction (E, H) and one-way ANOVA with Turkey’s post hoc comparison(c). Data are presented as mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

To investigate the regulatory effects of LXR on Osbpl6 transcription and its impact on cholesterol homeostasis, we administered T0901317, an LXR agonist, to CUMS mice for 8 weeks (Fig. 8D) as previewed reported[43]. Behavioral tests revealed that T0901317 treatment significantly improved behavioral disorders in mice (Fig.

8E-H and Fig. S10F-H). Furthermore, T0901317 also upregulated the expression of OSBPL6 and MBP (Fig.

8I-J), attenuated hippocampal demyelination (Fig. 8K-L), and reduced cholesterol accumulation in oligodendrocytes within the hippocampus (Fig. 8M-N).

These findings demonstrate that chronic stress-induced downregulation of OSBPL6 is likely attributed to negative regulation of LXR by GR, subsequently disrupting cholesterol and myelination homeostasis.

Discussion

Disorders of cholesterol metabolism have been increasingly reported in certain neurodegenerative disorders [44], [45], but has been little studied in depression. The present study revealed that the deficiency of OSBPL6, an intracellular lipid binding/transporter protein, leads to a systemic failure of intracellular cholesterol homeostasis and gradient maintenance, affecting multiple organellar compartments,which resulted in the abnormal distribution of cholesterol within myelinating oligodendrocytes, eventually leading to demyelination and depression-like behaviors in mice.

Lipids are critical regulators for brain activity that shape plasticity and maintain brain homeostasis[46], [47]. Imbalanced lipid metabolism, often involving maladaptive myelin synthesis, is a hallmark of neurological and neurodegenerative diseases[46]. For instance, inflammatory stimuli promote the expression of cholesterol-derived 25-hydroxycholesterol (25HC) in microglia, which disrupts LTP induction and inhibits learning[48].

Recent studies also highlight the oligodendrocytes’ role in cholesterol synthesis for myelin regeneration in chronic myelin disease models and multiple sclerosis patients[49]. In the CNS, oligodendrocytes enable rapid action potential conduction via myelin sheaths and support neuronal metabolism by clustering sodium channels at axonal nodes of Ranvier (a prerequisite for saltatory conduction).[50]. Beyond their role in neurodegenerative diseases like amyotrophic lateral sclerosis (ALS), oligodendrocytes also modulate depression, sleep-wake cycles, and glioblastoma[50], [51].

Cholesterol, a vital lipid molecule, is indispensable for brain development and physiology[19], [52], [53], with its homeostasis tightly regulated systemically and cellularly. Critically, intracellular cholesterol homeostasis relies on maintaining a cholesterol gradient, with concentration highest in the plasma membrane and lowest in ER, the primary site for cholesterol synthesis and esterification[54], [55], [56]. Typically, excess intracellular free-cholesterol is esterified in ER and stored in lipid droplets[57], [58].

Our results showed impaired cholesterol homeostasis in the hippocampus of CUMS mice, which was associated with aberrant cholesterol accumulation in oligodendrocytes and impaired myelination. Single-nucleus RNA sequencing revealed reduced Osbpl6 expression in hippocampal oligodendrocytes of CUMS mice, and OSBPL6-knockdown mice exhibited the same phenotypes—indicating this cytoplasmic sterol-binding protein is crucial for intracellular cholesterol homeostasis.

OSBPL6 belongs to the oxysterol-binding protein-like (OSBP) family, a group of intracellular lipid receptors that have been poorly studied. In vivo, it regulates intracellular cholesterol transport and efflux to maintain cholesterol homeostasis[59], [60]. It is expressed in the CNS[60], [61] and involvement in cholesterol transport between intracellular compartments.

Studies have shown that OSBPL6 is linked to premature coronary artery disease[62] and fluctuations in high-density lipoprotein cholesterol (HDL-C) levels[63], implying a potential role in cholesterol metabolism. Specifically, OSBPL6 localizes to endolysosomes and facilitates the delivery of cholesterol to the ER and other cytoplasmic membranes[64]. Our findings demonstrate that there is an accumulation of cholesterol within oligodendrocyte cell bodies but less distribution within myelin sheaths.

This suggests that chronic stress-induced OSBPL6 downregulation drives oligodendrocyte cholesterol transport defects; ER cholesterol accumulation and elevated esterified cholesterol may stem from failed OSBPL6-mediated multiple organellar compartments intracellular cholesterol homeostasis and gradient maintenance.

However, we have not yet validated the potential compensatory relationships between OSBPL6 and other OSBP/ORP members through targeted experiments several points require further clarification and validation. Notably, prior studies have highlighted functional overlaps among OSBP/ORP family proteins in lipid homeostasis regulation. For instance, OSBP1 has been shown to modulate cholesterol transport and APP processing in neuronal cells[65]; this implies that OSBP1 may share redundant or complementary roles with OSBPL6 in maintaining organelle-specific lipid balance—an interaction that warrants systematic investigation.

Another case in point is the ORP9-ORP11 heterodimers participate in myelin-associated lipid synthesis by regulating ER-Golgi lipid exchange, yet their compensatory relationship with OSBPL6 in cholesterol transport remains unresolved and requires subsequent targeted validation.

Oligodendrocyte myelin sheath formation requires transcriptional activation of myelin-related and cholesterol/lipid biosynthesis genes[66], but the regulatory mechanisms of this transcription in mature oligodendrocytes remain unclear. Our study addresses this by investigating OSBPL6 downregulation in chronic stress-induced depressive mice. Given the elevated glucocorticoid levels observed in the serum of both CUMS mice and major depressive disorder patients, we investigated the localization and function of the GR.

GR is highly expressed in the hippocampus[67], where it binds GCs to form a nuclear-translocating complex that regulates transcription[68]. Interestingly, our findings indicate that upon entering the nucleus, this GR-GC complex does not directly regulate Osbpl6 transcription; instead, it modulates OSBPL6 levels by inhibiting LXR (a sterol-responsive nuclear receptor that activates cholesterol transport/efflux genes.[69], [70], consistent with prior reports[36]. Using the GR antagonist mifepristone (RU486), we confirmed GR blockade increases LXR expression, which promotes Osbpl6 transcription, reduces cholesterol accumulation, and rescues axon demyelination.

Therefore, our findings suggest that chronic stress-induced hyperactivity of the HPA axis results in elevated levels of glucocorticoids. These glucocorticoids bind to GR and form a complex undergoing nuclear translocation, inhibit LXR activity, and decrease OSBPL6 expression. Correspondingly, mifepristone ameliorated demyelination and depressive behaviors in CUMS mice—partially elucidating the antidepressant mechanisms of HPA axis inhibition[71], [72].

Gender differences are critical for generalizing findings from depression and myelination studies to broader populations. The decision to focus exclusively on male mice in the present research was driven by practical considerations inherent to the study design. Specifically, the demyelination-induced depression model utilized requires an extended experimental cycle, and female mice typically exhibit greater variability in behavioral and emotional phenotypes over this period.

This variability is attributable to hormonal fluctuations associated with the estrous cycle, which introduces a significant degree of uncertainty when investigating molecular mechanisms and complicates the interpretation of preliminary data, particularly during the validation of core pathways linking demyelination to depressive phenotypes. It should be noted that this focus on male mice does not reflect a disregard for sex-specific mechanisms. Notably, this focus on male mice does not reflect a disregard for sex-specific mechanisms.

On the contrary, exploring gender disparities in demyelination-associated depression remains a central priority in our ongoing and future research agenda, with plans to systematically address these gaps in subsequent work.

One limitation of our current study is the lack of direct quantification of (2-hydroxypropyl)-β-cyclodextrin (HP-β-CD) bioavailability in the CNS. Nevertheless, preliminary experimental, pathological, and literature evidence collectively supports the potential of HP-β-CD to access the CNS under chronic stress conditions: first, 14C-labeled tracer experiments have confirmed that HP-β-CD crosses the blood–brain barrier (BBB) via a slow, non-saturable transcellular diffusion mechanism, a pathway that ensures its accumulation in the brain parenchyma at biologically effective concentrations[73]; second, the chronic stress model we employed induces neuroinflammation, endothelial dysfunction, and tight junction protein degradation[74], which disrupt BBB integrity and thereby enhance HP-β-CD penetration beyond its baseline physiological capacity; third, there has been reported that systemic HP-β-CD administration has been shown to delay disease onset, reduce intraneuronal storage, and extend lifespan in APOE4/4-TR mice[75], while recent mechanistic studies in NPC1 models further demonstrated that high-dose peripheral administration achieves moderate CNS exposure sufficient to ameliorate specific neuropathological phenotypes despite restricted BBB permeability[40]. Notably, like our current work, these precedent studies lack direct BBB tracking data but consistently validate functionally relevant CNS effects of HP-β-CD.

Future research should use radiolabeled HP-β-CD or in vivo imaging techniques to measure its concentration in the central nervous system under chronic stress conditions, which will clarify the pharmacokinetic basis of its neuroprotective effect.

In addition, it should be noted that the regulatory effects of OSBPL6 on cholesterol transport and myelination are primarily limited to the hippocampus, a critical brain region associated with depression development[76], [77], [78]. However, it cannot be excluded that OSBPL6 may also be altered in other brain regions associated with depression. Additionally, this study mainly focused on mature oligodendrocytes.

In reality, during neurodevelopment, oligodendrocytes proliferate and differentiate from oligodendrocyte precursor cells (OPCs) to form the myelin sheath[16]. Previous evidence has shown consistent alterations in OLs, including OPCs and mature oligodendrocytes, in depression and other psychiatric disorders[79]. Chronic stress may also lead to morphological damage and excessive oxidative stress in OPCs[79].

Finally, we cannot rule out the possibility that other pathways related to cholesterol metabolism are aberrant in depression. Considering the results of Gene Ontology analysis, we examined mRNA levels of several genes screened for association with cholesterol synthesis as well as the classical apolipoprotein ApoE4 but found no significant changes after chronic stress. However, it is still possible that alterations in relevant metabolic pathways may also be involved in other models of depression.

The present study only provides a preliminary elucidation of OSBPL6′s mechanistic roles in animal models, and translating these findings to clinical practice undoubtedly faces multiple challenges—for instance, it remains unclear how our preclinical observations relate to human physiology or could be clinically translated, even beyond the neurobiological context (e.g., to peripheral metabolism).

That said, existing research offers initial support for linking our animal findings to human major depressive disorder (MDD): prior postmortem studies confirm abnormal myelination (including myelination-related synaptic homeostasis disruption) in depressed patients—consistent with our preclinical results on OSBPL6-mediated myelination regulation, which establishes a neuropathological connection between mouse models and human MDD. Additionally, neuroimaging evidence shows white matter tract alterations in depressed individuals, and 40–50% of patients with demyelinating diseases (e.g., multiple sclerosis) experience significant depressive symptoms—suggesting myelination abnormalities may contribute to depression, further laying the groundwork for translating our OSBPL6-focused findings.

Conclusion

In conclusion, our findings validate that hyperactivity of the HPA axis under chronic stress leads to nuclear translocation of GR which prevents transcription of LXR thereby inhibiting expression of OSBPL6. The down-regulation of OSBPL6 leads to abnormal transport and progressive accumulation of cholesterol in oligodendrocytes resulting in axon demyelination and depression-like behaviors. We provide new evidence for the important role played by OSBPL6 underlying the pathogenesis of depression through regulating cholesterol homeostasis within oligodendrocytes within the central nervous system.

Furthermore, we propose pharmacologically promoting cholesterol transportation restored myelin homeostasis in vivo and improved behavior disorders in CUMS mice, which may provide novel insights for the treatment of depression.

Data availability

Further information and data that support the findings of this study are available from the corresponding author Shuyan Yu (shuyanyu@sdu.edu.cn) upon reasonable request. Source data are provided with this paper.

Code data availability

No custom code was used in the study.

The animal and experimental procedures were approved by the Ethics Committee of Shandong University (ECSBMSSDU-2022–2-65) and strictly adhered to the guidelines set by the Institutional Animal Care and Use Committee.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Related articles

PubMed

Fasting glucose-to-HDL cholesterol ratio and risk of MASLD in nonobese Chinese adults: A retrospective cohort study.

Metabolic dysfunction-associated steatotic liver disease (MASLD) is increasingly recognized in nonobese individuals; however, reliable markers for early risk identification remain limited. The fasting glucose-to-high-density lipoprotein cholesterol (FHR) ratio, which reflects integrated glucolipid metabolism, has emerged as a potential indicator. This study aimed to evaluate the association between FHR and the incidence of MASLD in nonobese adults. This retrospective cohort study included 16,173...

PubMed

Remnant Cholesterol Inflammatory Index and its Association With Cardiometabolic Multimorbidity in Middle-Aged and Advanced-Aged Individuals: Evidence From the China Health and Retirement Longitudinal Study.

Background The association between remnant cholesterol inflammatory index (RCII) and cardiometabolic multimorbidity (CMM) remains unclear. RCII is defined as: {Total Cholesterol − (High-Density Lipoprotein Cholesterol + Low-Density Lipoprotein Cholesterol)] × High-Sensitivity C-Reactive Protein/10}. This study examined the association between baseline RCII and CMM among middle-aged and older adults in China. Material/Methods This study analyzed data from the China Health and Retirement Longit...