Background
Alzheimer’s disease (AD) is a progressive neurodegenerative disorder and the most common form of dementia worldwide. AD is neuropathologically characterized by the deposition of β-amyloid plaques and neurofibrillary tangles in the neocortex of the brain. β-Amyloid plaques are composed of β-amyloid (Aβ) peptides, released from amyloid precursor protein (APP) as a result of sequential proteolytic cleavage by β- (BACE1) and γ-secretases [
1]. According to the amyloid cascade hypothesis, impaired balance between the production and clearance leads to accumulation of Aβ, which triggers downstream events in AD pathology, such as formation of neurofibrillary tangles and induction of neuroinflammation and oxidative stress [
2].
DHCR24 (3-β-hydroxysteroid-Δ-24-reductase), also known as seladin-1 (selective Alzheimer’s disease indicator-1), was initially identified as being expressed at lower levels in the affected as compared to unaffected brain regions in AD patients [
3,
4]. Thus, it was suggested to represent a selective indicator of AD pathogenesis. More recently, the idea that DHCR24 downregulation would selectively indicate AD pathogenesis [
3] has been questioned [
5]. On the other hand, genetic polymorphisms in
DHCR24 may modulate the risk of developing AD in humans [
6], suggesting that
DHCR24 could be genetically associated with AD. Functionally, DHCR24 protein is characterized as a multifunctional enzyme having two distinctive activities: cholesterol-producing activity, i.e., reduction of desmosterol to cholesterol, mediated by the C-terminal region, and H
2O
2-scavenging activity brought about by the oxidoreductase domain near the N-terminus [
7,
8]. An accumulating body of evidence strongly suggests that DHCR24 has neuroprotective properties, which might be associated with these activities. DHCR24 has proven protective in different AD-related stress conditions, including Aβ-induced, oxidative, or endoplasmic reticulum (ER) stress [
3,
9‐
13]. Moreover, DHCR24 has a pivotal role in the de novo synthesis of cholesterol. Brain cholesterol, which is fundamental to the synaptic formation and normal functioning of the brain, is decreased during both AD and normal aging [
14]. In vitro studies suggest that decreased cholesterol levels could contribute to AD in various ways, including increased inflammatory response and increased production or decreased clearance of Aβ [
15‐
17]. Furthermore, decreased levels of DHCR24 lead to the stabilization of BACE1 and consequently to increased β-amyloidogenic processing of APP under apoptotic conditions in vitro [
18]. These findings suggest that augmentation of DHCR24 levels in the affected brain areas might provide a potential therapeutic approach to intervene in AD pathogenesis.
To test if enhanced expression of DHCR24 leads to neuroprotection, we have investigated the effects of DHCR24 overexpression in in vitro and in vivo models upon neuroinflammation. Here, we report for the first time that DHCR24 protects neurons from death upon neuroinflammation induced by lipopolysaccharide (LPS) and interferon γ (IFN-γ) in a neuron-BV2 microglial cell co-culture model. Furthemore, mechanistic elucidation revealed that upon neuroinflammation, the overexpression of DHCR24 did not increase total cellular cholesterol content or APP levels, nor did it affect Akt- or ERK-related neuronal survival pathways or caspase-3 activation in the co-cultures. Importantly, overexpression of DHCR24 increased the total number of dendritic spines and the relative proportion of mushroom spines in mature mouse hippocampal neurons in vitro as well as regionally reduced lesion size in vivo in a mouse model of transient focal cerebral ischemia.
Methods
Lentiviral constructs
Human DHCR24 cDNA in pLenti-III-HA vector (pLenti-CMV-h-DHCR24) and empty pLenti-III-HA plasmid (both obtained from Applied Biological Materials, Richmond, BC, Canada) as well as green fluorescent protein (GFP) under the chicken beta actin (CAG) promoter was cloned into lentivirus transfer (HIV) plasmid and was packed into third-generation self-inactivating lentiviral particles in the BioCenter Kuopio National Virus Vector Laboratory in Kuopio, Finland.
Mouse primary cortical neuron and BV2 cell co-culture and lentivirus-mediated gene transfer
Mouse primary cortical neurons were harvested from 18-embryonic-day-old JAXC57BL/6J mouse pups. Single-cell suspension was prepared from the dissected cortices by trypsin digestion and trituration, and the cells were plated on poly-d-lysine-coated cell culture plates in Neurobasal feeding medium containing 2% B27, 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. Neurobasal feeding medium and B27 are serum-free and thus do not contain cholesterol. Cells were grown in a humidified incubator at 37 °C in 5% CO2. After 4 days in vitro (DIV), half of the medium was changed to feed the neurons. On DIV4, DHCR24 and control lentiviral particles were used to infect the cells at multiplicity of infection (MOI) 75. After 24 h of lentiviral infection (DIV5), the medium containing the lentiviral particles was removed. BV2 cells were added to the cultures at 1:5 ratio on DIV5 and let to attach for 2 h, after which the neuroinflammation was induced with 200 ng/ml LPS and 20 ng/ml IFN-γ. The anti-inflammatory cytokine interleukin 10 (IL-10, 50 ng/ml, Peprotech) and the inducible nitric oxide synthase (iNOS) inhibitor 1400 W (20 μM, Tocris) were used as positive controls and were added 1 h after seeding of BV2 cells to the co-cultures. Consequently, after the 1-h pre-treatment with IL-10 and 1400 W, neuroinflammation was induced with 200 ng/ml LPS and 20 ng/ml IFN-γ. Samples were collected on DIV7, 48 h after the induction of neuroinflammation.
Immunofluorescence microscopy
The neurons and BV2 cells were plated on poly-d-lysine-coated glass coverslips in 48-well cell culture plates. All the following steps were performed at room temperature unless otherwise noted. The cells were fixed in 4% paraformaldehyde for 20 min and permeabilized in ice-cold methanol for 8 min at − 20 °C. Unspecific antibody binding was blocked by incubation in blocking solution (PBS containing 1% bovine serum albumin and 10% normal goat serum) for 20 min. The cells were double-stained with mouse anti-MAP2 (1:2000, Sigma, M9942; neuronal marker) and rat anti-CD11b (1:300, Serotec, MCA74G; microglial cell marker) or rabbit anti-GFAP (1:300, Dako, Z0334; astrocyte marker) primary antibodies at + 4 °C overnight, followed by staining with anti-mouse-Alexa488 (1:500, Invitrogen) and anti-rabbit-Alexa594 (1:500, Invitrogen,) or anti-rat Alexa594 (1:500, Abcam) secondary antibodies for 1 h. Cells without any of the primary antibodies were used as negative controls for background staining. In between the antibody incubations, the cells were washed three times in PBS for 10 min. The nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI, 1:5000 in PBS, Sigma) for 5 min. The coverslips were mounted on objective slides using GelMount mounting media (Sigma). Cells were imaged using a Zeiss Axio Imager fluorescence microscope at × 10 magnification, and images were prepared using the Zeiss ZEN 2012 program. The number of different cell types (neurons, astrocytes and BV2 cells) in the co-cultures was quantified from five different images taken from three individual wells each in every experiment. All in all, on average 7000 cells were counted in each replicate. Addition of BV2 cells onto the neuronal cultures was observed to decrease the number of neurons by approximately 30% as assessed by the MAP2 immunocytochemistry-based neuronal viability assay (described below). Therefore, the quantified values were normalized for the decrease in neuronal number by dividing with the value of 1.3.
Protein extraction and western blot analysis
Total protein lysates for Western blot were prepared by lysing the cells in T-PER tissue protein extraction reagent (Thermo Scientific) supplemented with protease and phosphatase inhibitor cocktails (× 100, Thermo Scientific) and centrifuging at 10000×g for 10 min. BCA protein assay kit (Thermo Scientific) was used to determine protein concentrations, and 15–30-μg protein samples were separated on NuPAGE 4–12% BisTris gel (Life Technologies) and subsequently blotted onto a PVDF membrane. Following antibodies were used to probe the blots: Akt (1:1000, Cell Signaling Technology), p-Akt recognizing Akt phosphorylated at Ser473 (1:1000, Cell Signaling Technology), APP C-terminus (A8717, 1:2000, Sigma), BACE1 (D10E5, 1:1000, Cell signaling technology), caspase-3 (1:1000, Cell Signaling Technology), DHCR24 (C59D8, 1:800, Cell Signaling Technology), cofilin (D3F9, 1:1000, Cell Signaling Technology), p-cofilin recognizing cofilin phosphorylated at Ser3 (sc-271,921, 1:1000, Santa Cruz Biotechnology, Inc.), ERK 2 recognizing ERK 1 and 2 (1:500, Santa Cruz Biotechnology), p-ERK recognizing ERK 1 and 2 phosphorylated at Tyr204 (1:500, Santa Cruz Biotechnology), CREB (1:1000, Cell Signaling Technology), p-CREB recognizing CREB phosphorylated at Ser133 (1:1000, Cell Signaling Technology), neuroligin-1 (sc-365,110, 1:500, Santa Cruz Biotechnology, Inc.), GAPDH (1:15,000, Abcam), and β-actin (1:1000, Abcam). After incubation with appropriate species-specific horseradish peroxidase (HRP)-linked secondary antibodies (GE Healthcare), the proteins were detected using enhanced chemiluminescence substrates (GE Healthcare) and RT ECL Imager (GE Healthcare) or G:BOX Chemi XRQ (Syngene). Quantity One software (Bio-Rad) was used to quantify the protein levels.
RNA extraction and real-time quantitative PCR analysis
RNA was extracted in TRI reagent following manufacturer’s protocol for RNA isolation. cDNA synthesis was carried out using SuperScript III First-Strand Synthesis System for RT PCR (Life Technologies). Target specific PCR primers for mouse DHCR24 (5′-CAAGCCGTGGTTCTTTAAGC-3′ and 5′-CATCCAGCCAAAGAGGTAGC), mouse TNFα (5′-CGAGTGACAAGCCTGTAGCC-3′ and 5′-GTGGGTGAGGAGCACGTAGT-3′), mouse BDNF (5′-TGGCTGACACTTTTGAGCAC-3′ and 5′-GTTTGCGGCATCCAGGTAAT-3′), mouse NQO1 (5′-TAGCCTGTAGCCAGCCCTAA-3′ and 5′-GCCTCCTTCATGGCGTAGTT-3′), mouse HMOX1 (5′-GTCAGGTGTCCAGAGAAGGC-3′ and 5′-GCGTGCAAGGGATGATTTCC-3′), and mouse GAPDH (5′-AACTTTGGCATTGTGGAAGG-3′ and 5′-ACACATTGGGGGTAGGAACA-3′) were obtained from TAG Copenhagen. FastStart SYBR Green Master (Roche) was used for qPCR. The comparative ΔΔCt method was used to calculate GAPDH-normalized expression levels of the target mRNAs.
Lipid extraction and total cholesterol assay
To extract lipids for the total cholesterol measurement, cells were homogenized in chloroform:isopropanol:Igepal (7:11:0.1) mix and centrifuged for 10 min at 15000×g. Supernatant was taken to a clean tube and air-dried at 50 °C and put in a vacuum desiccator for 5 min. Fluorometric Total Cholesterol Assay Kit (Cell Biolabs) was used to measure the total cholesterol levels. The dried lipids obtained from lipid extraction were dissolved in 200 μl assay diluent, and 50 μl dissolved lipid sample was subjected for the analysis. Cholesterol assay was performed according to the manufacturer’s instructions, and the fluorescence signal was measured with excitation wavelength at 560 nm and emission wavelength at 590 nm using Fluorstar Galaxy plate reader. Total cholesterol levels were normalized to total protein levels of each sample.
Neuronal viability assay
Neuronal viability in the mouse primary cortical neuron and BV2 microglial co-cultures was assessed as described earlier [
19]. Briefly, the cells were fixed in 4% paraformaldehyde in PBS for 20 min and then incubated in 0.3% H
2O
2 in methanol to permeabilize the cells and block endogenous peroxidase activity. Incubation in blocking solution containing 1% bovine serum albumin and 10% horse serum for 20 min was used to prevent non-specific staining. Neurons were stained by incubation with mouse anti-MAP2 primary antibody (1:2000, Sigma) overnight at + 4 °C. Next, the cells were incubated with biotinylated horse anti-mouse secondary antibody (1:500, Vector labs) for 1 h and ExtrAvidin-HRP (1:500, Sigma) for another 1 h. The cells were washed three times for 10 min in PBS between the antibody incubations. All antibody dilutions were prepared in the blocking solution. Finally, ABTS Peroxidase Substrate solution (Vector Labs) was prepared following the manufacturer’s instructions and added onto the cells. The absorbance was measured using a microtiter plate reader (ELx808, BioTek or Infinite M200, Tecan) at 405 nm and was directly proportional to the number of neurons in the wells. Six replicate wells per assay were measured for each treatment. The measured background absorbance from co-cultures incubated without the anti-MAP2 primary antibody (negative controls,
n = 6 per assay) were averaged and subtracted from the absorbances in the other wells. The absorbance in the negative control wells was similar to the absorbance measured from wells without any cells. Due to the neuronal loss induced by neuroinflammation, the mean neuronal viability obtained for each group in each experiment was used to normalize BACE1 and Aβ40 levels.
Aβ, TNFα, NO, and ROS measurements
Conditioned media from the cell samples were collected immediately before protein extraction and centrifuged at 10000×g for 10 min. Aβ40 levels in the conditioned media were determined with monoclonal and HRP-conjugated antibody-based Human/Rat β amyloid 40 ELISA kit (Wako, Osaka, Japan). Aβ40 concentrations were normalized to neuronal viability. Mouse TNF alpha ELISA Ready-SET-Go! kit (Affymetrix, San Diego, CA, USA) was used for the detection of tumor necrosis factor α (TNFα) in the conditioned media. Nitric oxide (NO) levels were determined using Griess Reagent Kit for Nitrite Determination (G-7921, Life Technologies) and normalized to neuronal viability determined by the MAP2-ABTS assay described above. All kits were used as instructed by the manufacturers. Reactive oxygen species (ROS) levels in the co-cultures were measured using fluorogenic probe 2′, 7′-Dichlorodihydrofluorescin diacetate (DCFH-DA, Sigma D6883). One hour after adding BV2 cells, samples were labeled with 120 μM DCFH-DA for 30 min. Two hours after adding the BV2 cells, 5 h-neuroinflammation treatment was started. Subsequently, cells were lysed using T-PER lysis buffer (Thermo Scientific), and fluorescence was measured using plate reader at 480 nm/530 nm.
Mouse primary hippocampal neuron culture, transient transfection, and spine morphology analysis
Primary hippocampal neuronal cultures were prepared from 18-day-old mouse JAXC57BL/6J embryos according to the protocol previously described [
20]. Briefly, single-cell solution (240,000 cells/cm
2) was plated on 8-well chamber slides (LabTek) coated with poly-
d-lysine and 30 μg/ml laminin in feeding media composed of Neurobasal medium supplemented with 2% B27, 0.5 mM
l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. Hippocampal neurons were grown in a cell culture incubator at 37 °C in 5% CO
2. Half of the culture media was replaced with fresh feeding media after every 5 days in vitro. On DIV19, mature hippocampal neurons in 8-well chamber slides were transiently co-transfected with a mixture containing 2 μl of Lipofectamine 2000 (Invitrogen) and 0.3 μg of control plasmid DNA (pLenti-III-HA) or pLenti-CMV-h-DHCR24 and 0.3 μg enhanced green fluorescent protein (pEGFP). Hippocampal neurons were fixed in 4% paraformaldehyde 24 h after transfection. Anti-DHCR24 antibody (C59D8, 1:100, Cell Signaling Technology) was used for immunofluorescence staining of DHCR24 in the hippocampal neurons. Hippocampal dendritic spines from GFP-positive neurons were imaged with a Zeiss Axio Observer.Z1 inverted microscope (63× NA 1.4 oil objective) equipped with Zeiss LSM 800 confocal module (Carl Zeiss Microimaging GmbH, Jena, Germany). Serial Z-stacks of optical sections from dendritic segments were captured for spine analysis performed with NeuronStudio software [
21] as described previously [
22].
Animals
Animal experimentation was carried out in accordance with the national regulation and the Council of Europe (Directive 86/609) of the usage and welfare of laboratory animals. Experiments were approved by the Animal Experiment Board of Finland. The JAXC57BL/6J male mice were kept at the National Laboratory Animal Centre at the University of Eastern Finland in a room equipped with 12-h light/dark cycle and controlled humidity. The mice were provided standard laboratory animal chow and water ad libitum. All experiments were carried out during the day light.
Injection of viral vectors and stroke surgeries
Total of 27 mice with the age of 3 months were randomized into following treatment groups: (1) lenti-GFP injected sham mice (
n = 7), (2) lenti-GFP injected ischemic mice (
n = 11), and (3) lenti-DHCR24 injected ischemic mice (
n = 9). The mice received lenti-GFP or lenti-DHCR24 injections as described earlier with minor modifications [
23]. Shortly, the anesthesia was induced using 5% isoflurane, and the mice were attached to the stereotactic frame (David Kopf Instruments, Tujunga, CA, USA). Surgical anesthesia was maintained using 1.8% isoflurane. Core body temperature was maintained at 36.5 ± 0.5 °C using a homeothermic unit (PanLab, Harvard Apparatus, Barcelona, Spain) connected to a rectal probe. An incision was made on the skin above the injection site, and the skull was exposed. Two microliters of the lenti-DHCR24 or lenti-GFP with the titer of 9.19 × 10
9 TU/ml and 1.87 × 10
9 TU/ml, respectively, was injected unilaterally into the following coordinates: + 1.8 mm medial/lateral, 0.4 mm anterior/posterior, and − 2.9 mm dorsal/ventral from the bregma. Viruses were injected with a speed of 0.5 μl/min over a period of 4 min using a Hamilton syringe (Hamilton, NV, USA). After the withdrawal of the needle, the skin was sutured and the mice were allowed to wake up. Buprenorphine 0.03 mg/kg, i.p. (Temgesic 0.3 mg/ml), was used as analgesic and was given immediately after the procedure. Four weeks after the injection of the viral vectors, the mice were subjected to transient occlusion of the middle cerebral artery (tMCAO) or sham surgery as described previously [
24]. Briefly, the mice were anesthetized with isoflurane as described above. Silicon-coated monofilament with a diameter of 0.21 ± 0.02 mm (Doccol, MA, USA) was inserted through external carotid artery, advanced through the internal carotid artery to occlude middle cerebral artery. After 45 min of occlusion, the filament was withdrawn and the mice received 1 ml of saline subcutaneously. Mice were then allowed to recover in a heating chamber for 60 min and returned to home cage kept on the heating pad for 24 h. Only half of the cage was placed on heating pad to allow mice to choose their environment. Sham mice underwent similar procedure except that the filament was inserted and withdrawn immediately. Total of six mice died after the surgery from the following treatment groups: lenti-GFP injected ischemic mice (
n = 3) and lenti-DHCR24 injected ischemic mice (
n = 3). In addition, one mouse was excluded due to unsuccessful occlusion.
Magnetic resonance imaging and quantification of lesion volumes
Magnetic resonance imaging (MRI) was performed in anesthetized mice at 1 day post-injury for determination of the lesion volume. MRI was carried out using a horizontal 9.4 T (Agilent technologies, CA, USA) interfaced with Agilent Direct Drive console as previously described [
25]. Multi-slice T2-weighted images were acquired with echo time/repetition time of 40 ms/3000 ms, matrix size 128 × 256, field of view 19.2 × 19.2 mm
2, slice thickness 0.8 mm, and number of slices 12. Images were analyzed blinded to the study groups using Aedes software (Kuopio, Finland) running under MatLab program (Math-works, Natick, USA). The lesion volume was calculated using the formula previously described in [
26]. Specifically, the infarct volumes were obtained by multiplying the pixel size by the slice thickness. The infarct volumes were quantified using the following formula: Infarct volume = [volume of left hemisphere − (volume of right hemisphere − measured infarct volume)]/volume of left hemisphere [
26]. In addition, edema was calculated using the following formula: Edema = (volume of right hemisphere − volume of left hemisphere)/volume of left hemisphere. Both infarct volume and edema were expressed as percentages. The quantification was done by a researcher blinded to the study groups.
Statistical analysis
Statistical significance between groups was tested using independent samples t test or Mann-Whitney U test (two groups) and one-way ANOVA followed by LSD post-test or Kruskal-Wallis test followed by pairwise comparisons with Mann-Whitney U test (three or more groups) depending on whether the data fulfilled the assumptions for parametric tests. Statistical significance level was set at p < 0.05. All statistical analyses were carried out using IBM SPSS Statistics 21.0.
Discussion
DHCR24 has previously been shown to protect neuronal cells in different stress conditions, including oxidative stress, ER stress, Aβ toxicity, and apoptosis [
3,
9‐
13]. Here, we report for the first time that the overexpression of DHCR24 enhances neuronal viability in neuron-BV2 microglial cell co-cultures upon LPS- and IFN-γ-induced neuroinflammation, increases the total number of dendritic spines and the proportion of mushroom spines in mature mouse hippocampal neurons, and is neuroprotective in a mouse experimental model of cerebral stroke. Moreover, our data suggest that the neuroprotection elicited by the overexpression of DHCR24 upon neuroinflammation is not related to significantly altered proinflammatory cytokine response, total cellular cholesterol levels or the activity of proteins linked with neuroprotective signaling, such as Akt, ERK, or CREB. Conversely, the protein levels of NLG1, which is a well-established post-synaptic adhesion protein involved in synaptic plasticity [
30,
38,
39], were increased in neuron-BV2 co-cultures upon the overexpression of DHCR24 particularly in the basal growth conditions. These data suggest that yet undefined neuroprotective mechanism(s) may underlie the improved neuronal survival induced by the overexpression of DHCR24 in the used in vitro and in vivo models.
Prior studies have linked the neuroprotective effects of DHCR24 directly to its cholesterol-synthesizing activity, showing cholesterol-dependent protection from oxidative stress [
11] or Aβ toxicity [
9] and maintenance of lipid raft integrity [
10]. Here, a moderate, but statistically non-significant increase in total cellular cholesterol levels was detected, and this coincided with enhanced neuronal survival in the DHCR24-overexpressing neuron-BV2 co-cultures after the induction of neuroinflammation. However, it is possible that the moderately increased total cholesterol levels may simply reflect the increased number of cholesterol-rich neurons rather than increased DHCR24-mediated local synthesis of cholesterol. Also, owing to the fact that only the total cellular cholesterol levels were determined, it is impossible to specifically define whether the overexpression of DHCR24 specifically affected the levels of free cholesterol or cholesterol esters upon neuroinflammation. We also detected increased total and mushroom spine density in the mature mouse hippocampal neurons overexpressing DHCR24 upon basal growth conditions. This is an important finding as dendritic spines are dynamic structures tightly regulated by membrane lipid composition [
40,
41]. Thus, it could be hypothesized that a local increase in cholesterol synthesis by transient DHCR24 overexpression may facilitate spine formation in dendrites. Recently, it was suggested that cholesterol levels modulate the NMDA receptor (NMDAR) activity [
42], while the activation of NMDAR in turn triggered downstream signaling, promoting either cell survival or cell death [
43]. In the present study, the overexpression of DHCR24 did not significantly modulate PI3K/Akt, ERK1/2-mediated MAP kinase, or CREB signaling pathways associated with neuronal survival and growth, suggesting that these pathways do not play a major role in the DHCR24-mediated neuroprotection upon LPS- and IFN-γ-induced neuroinflammation. Conversely, upon basal growth conditions, the overexpression of DHCR24 significantly augmented the protein levels of NLG1 in neuron-BV2 microglial cell co-cultures, and after the induction of neuroinflammation, a similar trend was observed upon the overexpression of DHCR24. This is an important finding given that the augmented levels of NLG1 have been shown to increase the spine and synapse growth in cultured neurons [
38,
39], emphasizing the central role of NLG1 in retaining the synaptic plasticity and integrity. Recently, it was further established that NLG1 regulates the dendritic spines and synaptic plasticity via LIMK1/cofilin-mediated actin reorganization [
30]. More specifically, it was shown that NLG1 activates LIM-protein kinase (LIMK1), which in turn promotes the activation of cofilin through the phosphorylation of cofilin at Ser 3. Consequently, the Ser 3 phosphorylated form of cofilin promotes actin assembly, which then facilitates spine and synapse formation as well as long-term potentiation. In the present study, however, we did not observe a significant increase in the cofilin phosphorylation at Ser 3 upon overexpression of DHCR24 in neuron-BV2 microglial co-cultures, suggesting that molecular mechanism(s) other than NLG1/LIMK1/cofilin-mediated actin reorganization underlies the observed neuroprotection upon the overexpression of DHCR24 in co-cultures under neuroinflammation.
Increased plasma cholesterol content has been previously shown to inhibit the maturation of APP during cellular aging [
44], suggesting that the age-associated alterations in relation to cholesterol content are instrumental for cellular functions. Here, we observed enhanced maturation of the APP695 isoform in the control co-cultures upon the induction of neuroinflammation, while a similar increase was not detected in the DHCR24-overexpressing co-cultures. Importantly, APP maturation changes were evident only in the stress-induced, but not in the basal growth conditions, and were not associated with significantly altered levels of BACE1 or Aβ. Several lines of evidence suggest that cholesterol is a central component regulating synaptic function and plasticity [
14,
40]. Cholesterol and sphingolipids are enriched in the membranes of dendritic spines, and alterations in the levels of these lipids modulate spine morphology and synaptic activity via affecting the arrangement of actin cytoskeleton and the trafficking of NMDA receptors [
40]. Moreover, it has been shown that the reduced synthesis of total brain cholesterol is a prominent feature in aging individuals and AD patients [
45‐
47]. This is in line with the recent findings showing that constitutive loss of hippocampal cholesterol impairs cognition in old rats through an Akt-mediated molecular mechanism, leading to reduced hippocampal long-term potentiation [
41].
In addition to the cholesterol-synthesizing activity, previous studies have demonstrated that DHCR24 has a direct H
2O
2-scavenging activity, which is able to protect cells from oxidative stress [
3,
11,
12]. Also, DHCR24 overexpression-mediated protection from ER-stress-induced apoptosis was recently linked to decreased ROS levels [
12]. Thus, it is possible that the protection from neuroinflammation-induced neuronal death is also linked to ROS-scavenging feature by DHCR24. Here, the measurements of ROS levels 6 h after the induction of neuroinflammation in neuron-BV2 cell co-cultures did not reveal any quantifiable ROS production, suggesting that ROS are not produced in the early phases of neuroinflammation in this model. Neuroprotective effects of DHCR24 have also previously been linked to the reduced activation of caspase-3 [
3,
12], and increased caspase-3 activation was observed when DHCR24 was downregulated in the neuroblastoma cells [
18]. Here, the neuronal viability was increased in DHCR24-overexpressing neuron-BV2 cell co-cultures, but DHCR24 did not alleviate caspase-3 activation, indicating that the neuroprotective effect of DHCR24 was not mediated via inhibition of caspase-3 activity. This was also reflected by the unaltered levels of caspase-cleaved ~ 60 kDa APP fragment between DHCR24- and control-transduced neuron-BV2 cell co-cultures upon neuroinflammation.
We also observed that the lentivirus-mediated overexpression of DHCR24 in striatum reduced the ischemia-induced lesion size in a mouse model of transient focal ischemia. MRI images taken at 1 day after stroke revealed a very local, yet significant protection against ischemia-induced cell death (infarct volume) at the close proximity of the lentivirus injection site without any significant effect on edema. Although the protection was very local as the overexpression of the construct did not spread through the injected hemisphere, it provides a proof of principle that DHCR24 is able to confer protection also in vivo. Furthermore, it is expected that the local DHRC24 levels are higher near the injection site similar to that seen with GFP, which in turn could explain why the analyses conducted from the whole cell extracts showed a significant increase only in the mRNA levels of DHCR24, but not in the protein levels. Owing to the fact that the increased levels of NLG1 in the neuron-BV2 cell co-cultures upon the overexpression of DHCR24 were observed, a similar assessment of NLG1 was done in mouse striatal brain tissue. The total protein levels of NLG1 were not significantly altered, while the maturation of NLG1 (measured as the increased ratio of NLG1-m vs. NLG1-im) showed a trend towards an increase in the DHCR24-transduced striatal samples. NLG1 is known to be N- and O-glycosylated [
48], but the exact role of this maturation process in the context of synaptic plasticity is not well-established. Thus, it remains to be determined whether enhanced maturation of NLG1 is related to neuroprotective functions upon different acute stress conditions. Finally, elucidation of the expression profile of TNFα, BDNF, HMOX1, and NQO1 did not reveal alterations in mice overexpressing DHCR24 as compared to control mice, suggesting that DHCR24 is able to locally mitigate the ischemia-induced damage in a mouse model of transient focal cerebral ischemia without significantly affecting the inflammatory, neurotrophic, or oxidative stress responses. Collectively, the findings in the present study are in line with a previous study showing that DHCR24 exerts cholesterol-dependent neuroprotection in an experimental stroke model in mice, in which DHCR24 was genetically downregulated [
49]. More specifically, it was suggested that the underlying molecular mechanism of DHCR24 is linked to the maintenance of lipid raft integrity in astrocytes by assuring the EAAT2-mediated uptake of glutamate excess upon ischemic stress [
49]. This is an interesting mechanistic finding also in the context of the present study, as we observed that the number of astrocytes significantly increased after the addition of BV2 microglial cells to the neuronal cultures before the induction of neuroinflammation. Thus, increased expression of DHCR24 also in the astrocytes, and not exclusively in the neurons, might also contribute to the improved neuronal survival upon neuroinflammation in our co-culture model. Taken together, the two in vivo studies conducted in different experimental stroke models reinforce the key role of DHCR24 in molecular processes underlying neuroprotection.