Abstract
Introduction: Brain cholesterol relies on de novo biosynthesis and is crucial for brain development. Cholesterol synthesis is a complex series of reactions that involves more than twenty enzymes to reach the final product and generates a large number of intermediate sterols along two alternate pathways. This is a highly regulated and oxygen-dependent process and thus sensitive to hypoxia. Methods: Using the modified Vannucci procedure, a clinically relevant animal model of neonatal hypoxia ischemia (HI), we characterized the profile of cholesterol and its sterol intermediates, along with the key enzymes on the cholesterol synthetic pathway over a time course of 5 days after HI in the postnatal day 10 mouse brain. Results: Although the total cholesterol levels in the injured cortices appeared to be minimally attenuated at 5 days following HI, there was an overall repression of brain cholesterol biosynthesis. Lanosterol and the downstream sterols in both the Bloch and Kandutsch-Russell (K-R) pathways were consistently reduced for up to 3 days except for desmosterol, which was elevated. Correspondingly, protein expression of the controlling transcription factors sterol regulatory element-binding protein 2 (SREBP-2) and SREBP-1 was decreased at early time points (within 6 h), in parallel with the downregulation of several substrate enzymes for up to 5 days post-HI. HMG-CoA reductase (HMGCR), the first rate-limiting enzyme, was upregulated in the first 24 h after HI. The expression of 24-dehydrocholesterol reductase (DHCR24) that catalyzes the last step to produce cholesterol on the Bloch pathway and bridges the Bloch to K-R pathway was also augmented. Conclusions: Our data suggest perturbed brain cholesterol biosynthesis following neonatal HI. As some sterol intermediates and enzymes have diverse functions in brain development and stress responses other than producing cholesterol, assessment of their dynamic changes after HI is important to understand the lipid responses in rodent HI models and to identify lipid-based targeted therapies in future studies.
Introduction
Cholesterol biosynthesis is central for numerous biological processes in health and disease. There is a great demand for cholesterol throughout brain development as it not only serves as a building block for maximal membrane expansion during rapid brain growth but also plays key roles in central nervous system (CNS) patterning, synaptogenesis, and myelination [1‒4]. After the blood-brain barrier is fully closed, the brain loses access to circulating cholesterol carried in lipoproteins and relies on de novo biosynthesis to meet the needs of this lipid-rich organ [5]. In the postnatal mouse, the rate of cholesterol synthesis in the CNS is highest (300–400 nmol/h) during the first 3 weeks after birth, after which it drops about 6-fold and stabilizes at 10–13 weeks of age [6]. Correspondingly, the key enzymes in the synthetic pathways are significantly higher in the neonatal brain than those in the adult mouse brain [7‒9]. Cholesterol synthesis starts from acetyl-CoA and involves multiple enzymatic reactions with more than 20 intermediates along the pathway (Fig. 1). HMG-CoA reductase (HMGCR) and squalene monooxygenase (SM, encoded by SQLE gene) in the early steps are considered as the main rate-limiting enzymes. Lanosterol is the first sterol committed to cholesterol synthesis from which point it can be diverted into either the Bloch or Kandutsch-Russell (K-R) pathway. These two pathways use the same enzymes and process in parallel with closely related substrates (C24-saturated or C24-unsaturated sterols, respectively), where 24-dehydrocholesterol reductase (DHCR24) catalyzes the reduction of the delta-24 double bond. At the end of the road, cholesterol is generated from desmosterol (mediated by DHCR24 on Bloch pathway) or 7-dehydrocholesterol (by 7-dehydrocholesterol reductase, DHCR7, on the K-R pathway). Cholesterol synthesis must be tightly controlled as it is energy intensive but toxic when made in excess. For one molecule of cholesterol, in addition to high levels of acetyl-CoA, ATP, and NADPH, 11 molecules of oxygen are required at multiple steps [10] (Fig. 1), making this process especially sensitive to oxygen shortages, such as hypoxia or hypoxia ischemia (HI). HI is a significant cause of brain injury in term and late preterm infants with increased risk of mortality and lifelong neurological morbidity [11‒13]. In view of the vital roles that cholesterol plays in brain development, HI is likely to disrupt cholesterol synthesis and metabolism, resulting in debilitating functional outcomes. Understanding how HI affects cholesterol homeostasis in neonatal brain could open new avenue for lipid-based pharmacological interventions for neonatal hypoxic-ischemic encephalopathy (HIE).
Cholesterol biosynthetic pathway. The green box indicates the steps in the pre-lanosterol pathway. Lanosterol is the first sterol that is committed to cholesterol synthesis and is converted to cholesterol in the Bloch pathway (pink box) or the Kandutsch-Russsell (K-R) pathway (blue box). DHCR24 catalyzes the reduction of the Δ24 double bond of sterol intermediates and thereby links the two post-lanosterol pathways. 24(S),25-EC is produced via a shunt pathway in parallel to cholesterol synthesis. Dashed arrows represent multiple enzymatic steps. The enzymes are in red, and the sterols measured in this study are highlighted in yellow.
Cholesterol biosynthetic pathway. The green box indicates the steps in the pre-lanosterol pathway. Lanosterol is the first sterol that is committed to cholesterol synthesis and is converted to cholesterol in the Bloch pathway (pink box) or the Kandutsch-Russsell (K-R) pathway (blue box). DHCR24 catalyzes the reduction of the Δ24 double bond of sterol intermediates and thereby links the two post-lanosterol pathways. 24(S),25-EC is produced via a shunt pathway in parallel to cholesterol synthesis. Dashed arrows represent multiple enzymatic steps. The enzymes are in red, and the sterols measured in this study are highlighted in yellow.
Intracellular levels of cholesterol are controlled through a feedback regulatory mechanism by the transcription factors, sterol regulatory element-binding proteins (SREBPs, SREBP-1 and SREBP-2), mainly by SREBP-2 [14, 15]. SREBPs are synthesized as ∼122 kDa inactive precursors that are bound to the endoplasmic reticulum membrane, and their activation is regulated by sterols in the endoplasmic reticulum. During cholesterol deficits, SREBP-2 is transported to the Golgi apparatus where it is proteolytically cleaved to ∼65 kDa active form, allowing its translocation to the nucleus to activate transcription of key genes in sterol biosynthesis, including HMGCR, 3-hydroxy-3-methylglutaryl-CoA synthase (HMGCS), SM, and DHCR24, as well as genes involved in cholesterol uptake (LDL receptor) [16, 17]. In turn, excess in sterols inhibits SREBP activity and keeps the synthetic/metabolic pathways in balance. SREBP-2 and other important enzymes are additionally subject to post-translational regulation by E3 ubiquitin ligase-mediated degradation in response to sterol intermediates and/or hypoxia [18]. The activity of several enzymes is altered through phosphorylation [18].
There are very few publications describing cholesterol homeostasis following neonatal HI and the associated consequences on brain damage and repair. In postnatal day 7 (P7) rats, an earlier study [19] showed a loss of hippocampal gangliosides, phospholipids, and cholesterol 7 days to 3 months without any changes immediately (30 min) after HI suggesting a delayed response; however, the data between 30 min and 7 days were missing. Another study in 2009 [20] using P7 rats also demonstrated reduced brain cholesterol in the first 3 days after HI, which preceded hypomyelination in the following 2 weeks, and yet the causal relationship between cholesterol loss and hypomyelination after HI is unclear. Nevertheless, in both studies, only total cholesterol contents were measured without further analysis of cholesterol synthetic precursors and their oxidized products, oxysterols, some of which are bioactive and play important roles in regulating cholesterol homeostasis, especially in the developing brain. A recent study [21] showed changes in sterol and oxysterol concentrations in the brain and plasma in severely injured mice at P9 at 24 h after HI. But without stratification of injury severity, there were minimal changes in brain cholesterol and its intermediates in the first 72 h following neonatal HI, probably indicating a cellular compensatory mechanism to minimize energy and oxygen consumption.
We previously evaluated the effect of HI on the total brain cholesterol levels and its primary metabolite 24S-hydroxycholesterol (24S-HC) following HI in neonatal mice [22]. In this study, we focused on investigating the dynamic changes of the cholesterol precursors and the important enzymes along the two synthetic pathways as some of these sterols not only contribute to cholesterol production but also have their own unique function in brain development.
Materials and Methods
Animals
All animal procedures were performed in accordance with the recommendations outlined in the Guide for the Care and Use of Laboratory Animals of National Institutes of Health, approved by the University of California San Francisco Institutional Animal Care and Use Committee and reported in compliance with the Animal Research: Reporting of In Vivo Experiments guidelines. C57BL/6J mice with litters were purchased from the Jackson Laboratory (strain #000664, Bar Harbor, ME) and housed at the UCSF Laboratory Animal resource Center with ad libitum access to food and water. Both sexes were used on postnatal day 9–10 (P9–P10).
Neonatal Brain Hypoxia Ischemia
Unilateral HI was induced using the modified Vannucci model in pups at P9 or P10 [22‒24]. Through a vertical midline neck incision under isoflurane anesthesia (2–3% isoflurane, balanced oxygen), the left common carotid artery was closed with electrical coagulation. The animals were allowed to recover for 1 h with their dam and then exposed to 60 min of global hypoxia (10% oxygen/balanced nitrogen) in a humidified chamber at 36.5°C. Sham-operated control animals received isoflurane anesthesia and exposure of the left common carotid artery without coagulation and hypoxia.
Brain Sterols and Oxysterols Measurement
Animals were sacrificed at various time points after HI. The cortices from the sham and HI-injured mice (both ipsilateral and the contralateral hemisphere) were dissected on ice, snap frozen by immersion into dry ice-chilled 2-methylbutane, and stored at −80°C until use. Sterols and oxysterols were extracted and analyzed based on the methods described by McDonald et al. [25]. Briefly, bulk lipids were extracted with methanol: dichloromethane; the sample was hydrolyzed with methanolic KOH to generate a pool of free sterols, which were then isolated and purified with liquid-liquid extraction and solid-phase extraction. Compounds were resolved on C18 core-shell high-performance liquid chromatography columns and by gas chromatography. Sterols were quantitatively measured by isotope dilution using high-performance liquid chromatography-tandem mass spectrometry (MS). Lathosterol was measured using gas chromatography-MS. Besides cholesterol, intermediates along the Bloch pathway, including lanosterol, FF-MAS, zymosterol, 7-dehydrodesmosterol (7-Dehydro Des), and desmonsterol, and those in the K-R pathway, including 24,25-dihydrolanosterol, lathosterol, and 7-dehydrocholesterol (7-Dehydro C), were quantified. In addition, the oxysterol 24(S),25-epoxycholesterol (24(S),25-EC) that is formed in parallel to cholesterol via a shunt of the mevalonate pathway (Fig. 1) was also analyzed. The results were reported as ng per mg tissue weight.
Western Blotting
Protein expression of the key cholesterol biosynthetic enzymes and the transcriptional factors (SREBP-1 and SREBP-2) was evaluated by Western blotting. Protein was extracted with RIPA lysis buffer (R0278, Sigma-Aldrich Inc., St. Louis, MO) with Halt™ proteinase and phosphatase inhibitor cocktail (#78442, Pierce Biotechnology, Rockford, IL). The protein concentrations were measured by BCA assay kit (Pierce). An equal amount of protein samples (40–45 μg) was applied to 4–12% Bis-Tris SDS polyacrylamide gel electrophoresis and transferred to polyvinyl difluoride membrane. The blots were probed with the following primary antibodies overnight at 4°C: SREBP-1 (sc-13551, Santa Cruz Biotechnology, Santa Cruz, CA), SREBP-2 (sc-271616, Santa Cruz), HMGCS (sc-166763, Santa Cruz), HMGCR (ab174830, Abcam, Cambridge, MA), squalene synthase (SQS, sc-271602, Santa Cruz), SM (NBP2-93808, Novus Biologicals, Centennial, CO), lanosterol synthase (LS, sc-514507, Santa Cruz), lanosterol 14α-demethylase (LDM, 13431-1-AP, Proteintech Group Inc., Rosemont, IL), DHCR7 (ab103296, Abcam), DHCR24 (#2033, Cell Signaling Technology Inc., Danvers, MA), and β-actin (sc-47778, Santa Cruz). After washing the membranes with TBS buffer with 0.05% Tween 20 (#T0310, TEKNOVA, Half Moon Bay, CA), HRP-conjugated secondary IgG antibodies (Jackson ImmuoResearch Laboratories Inc.) were used for 1 h at room temperature. Protein signal was visualized on radiographic film with enhanced chemiluminescence (Amersham, GE Healthcare). ImageJ software was used to measure the mean optical densities and the areas of the protein bands. The optical density value of each protein was normalized to that of β-actin to represent the level of protein expression.
Statistical Analysis
For MS experiments, the sterols/oxysterol concentrations were normalized to the wet weight of the cortical samples (ng/mg tissue weight). For Western blotting, the levels of protein expression were normalized to those of sham left cortex (LC) at 1 h. All results were presented as scatter plots with bar (median with interquartile range). Data were evaluated statistically using nonparametric Mann-Whitney test using Prism 9 software. The values of sham LC with sham right cortex (RC) (for sterol/oxysterol data only), sham LC with HI LC, and HI LC with HI RC at each individual time point were compared. Differences were considered significant at p < 0.05.
Results
Loss of Lanosterol and Multiple Cholesterol Precursors on Bloch and K-R Pathway following Neonatal HI
MS-based sterol profiling showed a decrease in lanosterol and several downstream intermediates on the Bloch pathway, including FF-MAS and zymosterol after HI in the ipsilateral cortices (LC), but not in the contralateral side (RC), compared to the values of the sham animals at the time points evaluated (p < 0.05, n = 4–6, Fig. 2). Significant reduction lasted for 3 days, and their levels returned to normal at 5 days (120 h) after HI. The changes in 7-dehydrodesmonsterol were less notable, with slight, but significant decrease at 24 h and 48 h post-HI. We did not quantify other molecules (e.g., squalene) in the early steps (Fig. 1) as lanosterol is the first sterol that is dedicated to cholesterol synthesis. Loss of lanosterol may account for the diminished production of FF-MAS and zymosterol; however, the amount of desmosterol, the immediate precursor of cholesterol on the Bloch pathway, was elevated at 24 h and 72 h with also a trend of increase at 48 h after HI (Fig. 2). On the K-R pathway, we found a significant decrease in 24,25-dihydrolanosterol at 24 h and 72 h after HI in the ipsilateral cortices, as well as less amount of 7-dehydrocholesterol at 24 h and 48 h post-HI (Fig. 3). No alteration in lathosterol was observed in the first 5 days after HI. Overall, the final cholesterol levels were marginally reduced at 5 days after neonatal HI (Fig. 2).
Changes of cholesterol and its sterol precursors on the Bloch pathway following neonatal HI. The total cholesterol levels showed minimal changes over 5 days after HI, while multiple sterol intermediates on the Bloch pathway were consistently reduced for up to 3 days post-HI compared to the values of the sham animals at the time points evaluated, except for desmosterol, which was increased. The amounts of sterols were normalized to the tissue wet weight (ng/mg tissue weight). LC, left cortex (ipsilateral side); RC, right cortex (contralateral side); n = 4 for sham animals at each time point; n = 6 for HI-injured animals (*p < 0.05).
Changes of cholesterol and its sterol precursors on the Bloch pathway following neonatal HI. The total cholesterol levels showed minimal changes over 5 days after HI, while multiple sterol intermediates on the Bloch pathway were consistently reduced for up to 3 days post-HI compared to the values of the sham animals at the time points evaluated, except for desmosterol, which was increased. The amounts of sterols were normalized to the tissue wet weight (ng/mg tissue weight). LC, left cortex (ipsilateral side); RC, right cortex (contralateral side); n = 4 for sham animals at each time point; n = 6 for HI-injured animals (*p < 0.05).
Changes of cholesterol precursors on the K-R pathway following neonatal HI. The levels of 24,25-dihydrolanosterol and 7-dehydrocholesterol (7-Dehydro C) were reduced after HI. Lathosterol was unaffected. The values of sterols were normalized to the tissue wet weight (ng/mg tissue weight). LC, left cortex (ipsilateral side); RC, right cortex (contralateral side); n = 4 for sham animals at each time point; n = 6 for HI-injured animals (*p < 0.05).
Changes of cholesterol precursors on the K-R pathway following neonatal HI. The levels of 24,25-dihydrolanosterol and 7-dehydrocholesterol (7-Dehydro C) were reduced after HI. Lathosterol was unaffected. The values of sterols were normalized to the tissue wet weight (ng/mg tissue weight). LC, left cortex (ipsilateral side); RC, right cortex (contralateral side); n = 4 for sham animals at each time point; n = 6 for HI-injured animals (*p < 0.05).
Decreased Generation of 24(S),25-EC through the Shunt Pathway after HI
24(S),25-EC is formed via a shunt of the mevalonate pathway at early steps, parallel to cholesterol synthesis (Fig. 1). We assessed this oxysterol as it theoretically represents a measure of cholesterol synthesis and is able to regulate intracellular cholesterol hemostasis by inhibiting cholesterol biosynthesis and uptake [26‒28]. Figure 4 shows a considerable reduction in 24(S),25-EC in the ipsilateral injured cortices from 48 h to 5 days after HI (p < 0.05). Its levels remained unchanged at 24 h post-HI.
Production of 24(S),25-EC was diminished following neonatal HI. The levels of 24(S),25-EC that is generated through the shunt pathway were decreased for up to 5 days after HI. The values of the oxysterol were normalized to the tissue wet weight (ng/mg tissue weight). LC, left cortex (ipsilateral side); RC, right cortex (contralateral side); n = 4 for sham animals at each time point; n = 6 for HI-injured animals (*p < 0.05).
Production of 24(S),25-EC was diminished following neonatal HI. The levels of 24(S),25-EC that is generated through the shunt pathway were decreased for up to 5 days after HI. The values of the oxysterol were normalized to the tissue wet weight (ng/mg tissue weight). LC, left cortex (ipsilateral side); RC, right cortex (contralateral side); n = 4 for sham animals at each time point; n = 6 for HI-injured animals (*p < 0.05).
In summary, we found a reduced production of lanosterol, along with its downstream cholesterol precursor sterols on Bloch pathway (FF-MAS, zymosterol, and 7-dehydrodesmonsterol, Fig. 5a) and on K-R pathway (24,25-dihydrolanosterol and 7-dehydrocholesterol, Fig. 5b) within 3 days in the injured cortices (LC) after neonatal HI. Desmostrol on the Bloch pathway was the only sterol that was increased in LC (Fig. 5a), while the lathosterol concentrations on the K-R pathway remained the same (Fig. 5b). The overall cholesterol content was slightly decreased at 120 h. There were no changes in the levels of any intermediate sterols in the contralateral cortices (RC, Fig. 5c, d).
Summarized graphs illustrating the profile of cholesterol and its sterol intermediates over 5 days after neonatal HI. The individual sterol values were normalized to those of the sham animals at the same time points (dashed lines at 1). Desmostrol was the only sterol that was increased in the left cortices (LC, the ipsilateral side) while most sterols on both Bloch and K-R pathways were reduced in the first 3 days following HI (a, b, *p < 0.05 vs. sham values). There were no significant changes in sterol levels in the right cortices (RC, the contralateral side, c, d).
Summarized graphs illustrating the profile of cholesterol and its sterol intermediates over 5 days after neonatal HI. The individual sterol values were normalized to those of the sham animals at the same time points (dashed lines at 1). Desmostrol was the only sterol that was increased in the left cortices (LC, the ipsilateral side) while most sterols on both Bloch and K-R pathways were reduced in the first 3 days following HI (a, b, *p < 0.05 vs. sham values). There were no significant changes in sterol levels in the right cortices (RC, the contralateral side, c, d).
Transcription Factors SREBP-1 and SREBP-2 Protein Expression Is Inhibited Early after HI
To examine how neonatal HI affects cholesterol biosynthesis at the transcriptional levels and on the key enzymes involved in the steps of cholesterol synthesis, especially those that require oxygen, we did Western blotting to measure their protein levels. In the sham animals, the expression of intact SREBP-1 and SREBP-2 precursors at 122 kDa decreased over time from 1 h to 5 days, equally from P9/10 to P14/15 (Fig. 6), suggesting an age-dependent developmental change. The levels of both SREBP-2 precursors and the cleaved active protein were remarkably reduced early after HI (1 h). Expression of SREBP-1 precursors was decreased at 1 h and 6 h, while the cleaved protein was diminished at 1 h with a trend of lower levels at 6 h compared to the sham animals (p = 0.07, Fig. 6). These changes indicate either transcriptional downregulation of SREBPs or accelerated degradation of SREBP precursors. At 24 h, the expression of both SREBP-1 and SREBP-2 was recovered to the sham values (Fig. 6).
Protein expression of the transcription factors SREBP-1 and SREBP-2 after neonatal HI. The representative Western blotting images are shown at the top and the quantification (normalized to β-actin and then to the value of sham 1 h LC) is shown at the bottom. The protein levels of both precursors (at 122 KDa) and the cleaved active form (at 65 KDa) dropped rapidly after HI (within 1 h) (n = 4–5 for each time point, *p < 0.05).
Protein expression of the transcription factors SREBP-1 and SREBP-2 after neonatal HI. The representative Western blotting images are shown at the top and the quantification (normalized to β-actin and then to the value of sham 1 h LC) is shown at the bottom. The protein levels of both precursors (at 122 KDa) and the cleaved active form (at 65 KDa) dropped rapidly after HI (within 1 h) (n = 4–5 for each time point, *p < 0.05).
Impact of Neonatal HI on the Pre-Lanosterol Pathway Enzymes and LDM
The mevalonate pathway is the early step (Fig. 1) leading to lanosterol production and subsequent cholesterol synthesis. The enzymes including HMGCR, HMGCS, and SM are the main targets of SREBP-2, in which HMGCR and SM are considered as the rate-limiting enzymes for cholesterol synthesis [29]. SM is also the first enzyme that needs oxygen; along with LS, it controls the flow of sterols to produce lanosterol/cholesterol or 24(S),25-EC through the shunt pathway. Figure 7 shows that the expression of HMGCS was decreased at 1 h, 24 h, 48 h, and 5 days after HI. The same trend of reduction was found for SQS (downregulated at 1 h, 24, 48, and 72 h), as well as SM (decreased at 6 h, 24 h, and 48 h). The changes in LS and LDM occurred late after HI (LS was decreased at 72 h and 5 days after HI, while LDM was downregulated from 48 h to day 5). The protein expression of HMGCR was upregulated at 6 h and 24 h after HI, which is consistent with what we published previously [22].
Protein expression of the key enzymes in the pre-lanosterol pathway and LDM. The representative Western blotting bands are shown at the top and the quantification (normalized to β-actin and then to the value of sham 1 h LC) is shown at the bottom. The protein levels of HMGCS, SQS, SM, LS, and LDM were all reduced, while the expression of HMGCR was upregulated at 6 h and 24 h after HI (n = 4–6 for each time point, *p < 0.05).
Protein expression of the key enzymes in the pre-lanosterol pathway and LDM. The representative Western blotting bands are shown at the top and the quantification (normalized to β-actin and then to the value of sham 1 h LC) is shown at the bottom. The protein levels of HMGCS, SQS, SM, LS, and LDM were all reduced, while the expression of HMGCR was upregulated at 6 h and 24 h after HI (n = 4–6 for each time point, *p < 0.05).
Different Effects of HI on the Expression of the Terminal Enzymes DHCR7 and DHCR24
DHCR7 and DHCR24 catalyze the final steps of cholesterol synthesis via the K-R and Bloch pathways, respectively (Fig. 1). An in vitro study suggested that they interact physically and functionally [30]. We found that DHCR7 was rapidly turned over at 1 h in response to HI, and its expression returned to normal at 6 h. In contrast, the DHCR24 levels were upregulated from 1 h to 72 h following HI (Fig. 8).
Different effects of HI on the expression of the terminal enzymes DHCR7 and DHCR24. The representative Western blotting bands are shown at the top and the quantification (normalized to β-actin and then to the value of sham 1 h LC) is shown at the bottom. The protein levels of DHCR7 were attenuated early after HI (at 1 h), while the expression of DHCR24 was upregulated for up to 3 days post-HI (n = 4–6 for each time point, *p < 0.05).
Different effects of HI on the expression of the terminal enzymes DHCR7 and DHCR24. The representative Western blotting bands are shown at the top and the quantification (normalized to β-actin and then to the value of sham 1 h LC) is shown at the bottom. The protein levels of DHCR7 were attenuated early after HI (at 1 h), while the expression of DHCR24 was upregulated for up to 3 days post-HI (n = 4–6 for each time point, *p < 0.05).
Discussion
In this study, we focused on investigating the effects of neonatal HI on cholesterol precursor sterols and pathway enzymes using P10 C57BL/6 mice. We found that following HI, there was a global repression of brain cholesterol biosynthesis within 3 days with reduced levels of intermediates on both Bloch and K-R pathways, except for desmosterol, which was increased. Total cholesterol levels, however, showed minimal changes indicating it was imprecise to represent alterations in intracellular sterol dynamics. Correspondingly, protein expression of the key enzymes in the early pre-lanosterol pathway (HMGCS, SQS, SM, and LS) was decreased, whereas HMGCR and the terminal enzyme DHCR24 were upregulated. These changes were observed only in the ipsilateral hemisphere, the regions where ischemia is induced when combined with hypoxia resulting in neuronal death and brain damage, but not in the contralateral side that experiences hypoxia without ischemia and with no histological injury.
The interhemispheric differences in lipid alterations may be attributed to the differential effects of HI and hypoxia alone on the cellular oxidative and energy metabolism, on substrate or energy supply for cholesterol synthesis, which have been described in studies using the Vannucci model [31‒33]. For example, in the ipsilateral hemisphere, ATP and phosphocreatine are irreversibly depleted during HI without a complete recovery for at least 18 h, and a secondary depletion occurs at 24 and 48 h. In contrast, high-energy phosphates are well preserved in the contralateral hemisphere [31‒33]. Further investigation is required to ascertain the exact underlying mechanisms.
Shutdown or diminished cholesterol biosynthesis in the injured cortices after HI could be part of the cellular adaptive or defense responses for oxygen and energy sparing. The pathway was switched off rapidly as SREBP-2, the master transcriptional factor that activates genes in cholesterol synthesis, was downregulated within 1 h following HI. Expression of SREBP-1 was also markedly attenuated for at least 6 h. While SREBP-1 is primarily responsible for activation of genes for fatty acid synthesis, one of the isoforms SREBP-1a also promotes cholesterol synthetic genes [34]. With no overt alterations in cholesterol levels in our model, SREBP-2 may not be processed by the canonical sterol-mediated feedback regulation, but presumably underwent accelerated protein degradation under hypoxia by the E3 ubiquitin ligase membrane associated ring-CH-type finger 6 (MARCH6), in a hypoxia-inducible factor-independent manner as reported in cell culture studies [35]. The rapid reduction of both the precursors and the cleaved active form of SREBPs after HI supports this idea, suggesting that the regulation is unlikely to occur at the mRNA level or through an increase in precursor cleavage.
SREBP-2 reduction led to a parallel decrease in the expression of its transcriptional targets, including HMGCS, SQS, SM, LS, and LDM [17], with the exception of HMGCR, which was elevated. As the first rate-limiting enzyme for cholesterol synthesis, HMGCR is also subject to another level of post-translational regulation, an event that is sensitive to oxygen availability. In CHO-7 cells and in mouse liver, HMGCR can be degraded in hypoxia through a mechanism that involves activation of HIF-1α and accumulation of lanosterol and 24,25-dihydrolanosterol [36‒38]. The increased HMGCR in our model was possibly due to a lack of normal feedback inhibition from these sterols and the cells responded by developing a compensatory increase in HMGCR levels. The third oxygen-sensitive enzyme is SM, which is also rate limiting and the first in the pathway to require oxygen. It has been reported that hypoxia reduces the levels of full-length SM through both transcriptional downregulation and accelerated post-translational degradation (by MARCH6), which results in a truncated but constitutively active form of the enzyme [39]. This cannot be confirmed in our study because SM activity was not analyzed.
The second enzyme in the pathway that requires oxygen, and thus is susceptible to hypoxia, is LDM (encoded by Cyp51a1 gene). We previously published that the repressed Cyp51a1 expression under hypoxia is mediated by the binding of SREBP-2/HIF-2α/FoxO4 ternary complex on its promoter [40]. Notably, LDM is also post-translationally degraded by MARCH6, same as SREBP-2 and SM. Nitric oxide, but not hypoxia, can promote LDM turnover [41].
DHCR24 and DHCR7 act at the final steps to produce cholesterol on the Bloch and K-R pathway, respectively. DHCR24 is also positioned at the gateway to divert the intermediates from Bloch to the K-R pathway by catalyzing the reduction of the delta-24 double bond. Although DHCR24 is one of the SREBP-2 targets, unlike the other enzymes that were downregulated after HI, we observed an elevated DHCR24 protein level from 1 h to at least 72 h. It is possible that other regulatory inputs predominated over the influence of SREBP-2. For example, DHCR24 expression could be induced in response to oxidative stress [42]. Again, it is unclear whether DHCR24 enzymatic activity was affected, which can be modified by phosphorylation, oxysterols (24(S),25-EC), liver X receptors, steroid hormones (progesterone) [43‒45], or micro RNAs [46]. DHCR24 has been ascribed diverse functions, such as in cell signaling [47, 48], being antioxidant [49‒51], anti-apoptotic [51, 52], anti-inflammatory [53], and is neuroprotective [54] in adult mouse stroke model [55]. Thus, its elevation in neonatal HI might be beneficial.
With DHCR24 upregulation after HI, it seems paradoxical that its substrate desmosterol was increased. However, this enzyme may be substrate induced. The mechanism underlying desmosterol accumulation needs to be defined, especially when the entire Bloch pathway seems to be shut down. Desmosterol amount peaks in early postnatal days in rodent brain and in human fetal brain before the period of myelination and is not detectable in brains of adult humans and animals [44, 56‒60]. The developmental changes in DHCR24 activity in rat brain follow the same trajectory [58]. Desmosterol is considered as a major source of cholesterol to support CNS myelin formation highlighting its unique contribution to brain development [56, 58, 59]. It is unknown whether HI-induced temporal accumulation of desmosterol is detrimental or beneficial for brain recovery, but studies have demonstrated its role in regulating oxidative stress [61] and inflammation [62‒64].
The decreased synthesis of lanosterol from the early steps limited sterol flux through the two divided downstream pathways, which may explain why FF-MAS, zymosterol, and 7-Dehydro Des on the Bloch, 24,25-dihydrolanosterol, and 7-Dehydro C on the K-R pathway were consistently reduced. In addition, nine oxygen molecules are consumed during the removal of the 4α, 4β, and 14α methyl groups in lanosterol and 24,25-dihydrolanosterol; therefore, hypoxia could slow down the demethylation of lanosterol and 24,25-dihydrolanosterol with diminished production of FF-MAS and zymosterol. Some cholesterol precursors (and oxysterols) are more reactive than cholesterol and have potent biological effects that differ from cholesterol. For example, 8,9-unsaturated sterols including lanosterol, FF-MAS, zymostenol, and 14-dehydrozymostenol are all able to enhance formation of oligodendrocytes from their progenitor cells and induce remyelination [65]. Inhibition of the related enzymes (LDM, DHCR14, SC4MOL, HSD17B7, or EBP) leading to accumulation of these sterols facilitated the development of drugs for multiple sclerosis [65‒68]. Targeting this portion of pathway (LDM to EBP) will probably offer new treatment options for HIE in which oligodendrocyte maturation arrest is common [69], or for white matter injury in premature infants [70]. Furthermore, accumulation of lanosterol and desmosterol in macrophages was found to modulate immune responses by increasing their phagocytic activity and anti-infection ability [62, 64]. It is likely that the changes in these intermediates, especially in microglia or other brain immune cells, are linked to the inflammatory responses following neonatal HI.
We observed a sustained decrease in 24(S),25-EC for at least 5 days after HI. This is a unique oxysterol in that it is formed not from cholesterol but via a shunt pathway from 2,3(S)-epoxysqualene using SM and the same downstream enzymes for cholesterol synthesis. Being formed in parallel to cholesterol, it serves as a measure of de novo cholesterol synthesis [28, 71]. Loss of 24(S),25-EC might be a surrogate marker of depressed cholesterol synthetic pathway following neonatal HI; meanwhile, its reduction indicates an impairment of acute feedback regulation of de novo cholesterol synthesis as 24(S),25-EC can inhibit SREBP activation and accelerate HMGCR degradation to protect the cells against accumulation of newly synthesized cholesterol [27, 71]. It is also the most potent ligand for liver X receptors to facilitate cholesterol efflux [72]. It is therefore a strong feedback regulator to fine-tune intracellular cholesterol homeostasis [27]. The ratio of 24(S),25-EC/cholesterol is high during brain development for its role to promote midbrain dopaminergic neurogenesis [73‒76]. Same as 8,9-unsaturated sterols, 24(S),25-EC can drive differentiation of oligodendrocyte precursor cells to oligodendrocytes [77].
In summary, our work provides insight on how HI interferes with pathways of cholesterol biosynthesis in the P10 neonatal mouse brain. Although descriptive, this information has not been documented in the field of perinatal brain injury and is useful to our understanding of lipid responses to HI. Identification of the primary driver and net outcome of all these changes are challenging as some enzymes are exquisitely regulated by different inputs and there is interplay between different regulatory mechanisms. Additionally, there are coordinated networks of cholesterol transportation and exchange between neurons and glia cells in different cellular contexts [78], which are different between the developing and adult brain. It is our goal to shift focus to explore in depth the cholesterol homeostasis at the level of individual brain cell population and study how cells interact with each other to keep the overall brain cholesterol amount in check while assisting each other to survive and fulfill their key functions.
Limitations of the Study
(1) We did not perform flux analysis using stable isotope labeling to trace flux through the early and the two post-lanosterol pathways, which would represent the rate of de novo cholesterol synthesis [79]. This approach could give us a better idea of at which stop(s) the traffic was blocked along the pathways. (2) We did not measure the activity of the enzymes, which could be regulated post-translationally and can better pinpoint which steps were inhibited or accelerated. (3) Our results were not stratified with the severity of brain injury, which accounted for the variability of the data. (4) Although cholesterol synthetic pathways are similar in all cell types, the capability of synthesis, the preference to use Bloch or K-R pathway or a hybrid pathway, the relative abundance of the enzymes and intermediates, the regulation mechanisms, and how synthesis is coupled to the pathway of cholesterol intake and efflux are drastically different among neurons and glia cells [80‒82]. Our results were limited by only reporting the overall/net changes in the cortical regions, not at the cell-specific level. Nevertheless, defining sterol changes after HI brain injury and the mechanistic insights would allow for the development of novel lipid-based therapeutic targets for HIE, as reported recently that inhibition of DHCR7 with increased 7-Dehydro C improves neuronal survival, while attenuating brain ferroptosis and tissue injury after neonatal HI [83].
Statement of Ethics
The animal study protocol was reviewed and approved by University of California San Francisco Institutional Animal Care and Use Committee (Approval No. IACUC # AN202238-00).
Conflict of Interest Statement
The authors have no conflicts of interest to declare.
Funding Sources
This work was supported by the National Institute of Neurological Disorders and Stroke (R56NS114563 and R21NS133533 to Dr. Jiang, R35NS097299 to Dr. Ferriero) and Research Council of Finland (decision No. 355256) to Dr. Manninen.
Author Contributions
F.L., N.R.S., and C.D.C.: investigation; C.Y.: formal analysis; C.M.Z.: writing – review and editing; T.M.: formal analysis and writing – review and editing; J.G.M.: methodology, resources, and investigation; D.M.F.: writing – review and editing and funding acquisition; X.J.: funding acquisition, conceptualization, project administration, supervision, validation, and writing – original draft preparation.
Data Availability Statement
Raw data are publicly available upon publication of this manuscript. Further inquiries can be directed to the corresponding author.