Dysregulation of Plasmalogen Homeostasis Impairs Cholesterol Biosynthesis*

Background: Physiological significance of plasmalogen homeostasis remains unknown. Results: Elevation of plasmalogens reduced cholesterol synthesis by enhancing degradation of squalene monooxygenase (SQLE), whereas SQLE was stabilized in the absence of plasmalogens. Conclusion: Plasmalogens regulate cholesterol synthesis by modulating the stability of SQLE. Significance: SQLE stability is modulated in response to the cellular level of plasmalogens, in addition to the acute changes of cholesterol level. Plasmalogen biosynthesis is regulated by modulating fatty acyl-CoA reductase 1 stability in a manner dependent on cellular plasmalogen level. However, physiological significance of the regulation of plasmalogen biosynthesis remains unknown. Here we show that elevation of the cellular plasmalogen level reduces cholesterol biosynthesis without affecting the isoprenylation of proteins such as Rab and Pex19p. Analysis of intermediate metabolites in cholesterol biosynthesis suggests that the first oxidative step in cholesterol biosynthesis catalyzed by squalene monooxygenase (SQLE), an important regulator downstream HMG-CoA reductase in cholesterol synthesis, is reduced by degradation of SQLE upon elevation of cellular plasmalogen level. By contrast, the defect of plasmalogen synthesis causes elevation of SQLE expression, resulting in the reduction of 2,3-epoxysqualene required for cholesterol synthesis, hence implying a novel physiological consequence of the regulation of plasmalogen biosynthesis.

Plasmalogen is a subclass of glycerophospholipid harboring vinyl ether-linked alkyl chain at its sn-1 and acyl chain at its sn-2 position. Physiological significance of plasmalogens is highlighted by fatal human diseases defective in plasmalogen synthesis such as Zellweger syndrome and rhizomelic chondrodysplasia punctata (1). Patients with rhizomelic chondrodysplasia punctata manifest clinical features of shortened proximal long bones, rhizomelia, bilateral congenital cataracts, and mental and growth retardation (2), predicting a role of plasmalogens in the development of bone, brain, and eye lens. In addition, studies using plasmalogen-defective mutant Chinese hamster ovary (CHO) cells revealed that plasmalogens are important for the esterification of cholesterol (3). Abnormality of other processes including myelination, paranode organization, and purkinje cell innervation is shown in a knock-out mouse defective in dihydroxyacetone phosphate acyltransferase, an essential enzyme for the synthesis of plasmalogens (4).
Plasmalogen synthesis comprising the seven-step reactions is initiated in peroxisomes. Formation of the ether-bond is catalyzed by the peroxisomal matrix protein, alkyl-dihydroxyacetonephosphate synthase (ADAPS), 2 which replaces the acylchain of 1-acyl-dihydroxyacetone phosphate with a long chain fatty alcohol. This fatty alcohol is generated by a peroxisomal C-tail anchored protein, fatty acyl-CoA reducatase 1 (Far1) (5,6). We earlier demonstrated that Far1 activity is regulated by modulating its stability in response to the cellular level of plasmalogens (6,7). Very recently, a plasmalogen-deficient disorder with intellectual disability, epilepsy, and cataracts was shown in a FAR1-defective patient, implying the essential function of Far1 in plasmalogen synthesis (8). However, the physiological consequence of the homeostasis of plasmalogen in cells remains poorly understood.
In the present study, we assessed the synthesis of lipids by modulating the cellular level of plasmalogens and found that homeostasis of plasmalogen is linked to cholesterol synthesis. Cholesterol synthesis is regulated by posttranslational and transcriptional mechanisms. The third step of cholesterol synthesis catalyzed by HMG-CoA reductase (HMGCR) is generally accepted as a rate-limiting step. HMGCR activity is mainly regulated via a sterol-mediated feedback mechanism at the level of transcription and endoplasmic reticulum-associated degradation of HMGCR (9). Recently, epoxidation of squalene catalyzed by the enzyme squalene monooxygenase (SQLE) is proposed as the second potential rate-limiting step in choles- The intensity of bands was quantified by Multi Gauge version 3.0 software (Fuji Film).
Immunoprecipitation-SQLE-HA 2 , MARCH6-Myc 6 , and MARCH6C9A-Myc 6 were transfected into CHO cells as described (22) and cultured for 2 days in the presence or absence of Etn. The cells were lysed for 5 min on ice with icecold phosphate-buffered saline containing 0.2% Triton X-100 and a mixture of protease inhibitors, and further solubilized at 4°C for 20 min. After centrifugation, cell lysates were subjected to immunoprecipitation using rabbit anti-Myc antibody. Polyclonal antibody to myc peptide was raised in rabbits by injection of the c-Myc peptide, CYILSVQAEEQKLISEEDL.
Data Presentation-Quantitative data were shown as mean Ϯ S.D. from three independent experiments.

Elevation of Plasmalogen Level Reduces Cholesterol
Synthesis-To explore physiological consequences of the homeostasis of plasmalogen, cellular plasmalogens were elevated by supplementing CHO-K1 cells with purified plasmalogens (PlsEtn) or Etn and assessed the biosynthesis of lipids by labeling with either [ 14 C]acetate or [ 14 C]palmitate. Supplementation of Etn increased cellular plasmalogens and phosphatidylethanolamine about 1.5 times more than those in CHO-K1 cells. When cells were cultured in the presence of PlsEtn, plasmalogens were increased nearly by 2-fold, whereas phosphatidylethanolamine levels were not altered (Fig. 1, A and C). By liquid chromatography-electrospray ionization-tandem mass spectrometry (LC-ESI-MS/MS) analysis, plasmalogens containing 18:0 and 18:1, but not 16:0 fatty alcohols at the respective sn-1 position were shown to be increased in the presence of PlsEtn, which was most likely due to the lesser amount of bovine brain-derived plasmalogens containing 16:0 fatty alcohol (data not shown). In contrast, all species of plasmalogens were elevated in the presence of Etn (Fig. 1D). When CHO-K1 cells were metabolically labeled with [ 14 C]acetate in the presence of Etn or PlsEtn, cholesterol synthesis was significantly reduced, whereas synthesis of glycerophospholipids including phosphatidylcholine and serine-and inositol-containing phospholipids (phosphatidylserine and phosphatidylinositol, respectively) were not altered (Fig. 1, A and B). In these conditions, cellular free cholesterol was not altered (Table 1). Similarly, reduction of cholesterol synthesis was also observed in HeLa and HEK293 cells upon culturing with Etn (Fig. 2, A  and B). The cellular level of plasmalogens in HeLa and HEK293 cells was increased about 1.5 times as compared with that in the untreated respective cells (Fig. 2C). Taken together, these results strongly suggest that cholesterol biosynthesis is specifically reduced by elevation of the cellular level of plasmalogens.
Elevation of Plasmalogen Level Down-regulates Cholesterol Synthesis Steps at Post-isoprenoid Biosynthetic Pathway-We further investigated which step of the cholesterol synthesis is affected upon elevation of plasmalogens. Cholesterol is synthe-   sized via the isoprenoid biosynthetic pathway. If the increased level of plasmalogens suppresses the rate of any steps of isoprenoid synthesis including HMGCR, a rate-limiting enzyme of the cholesterol synthesis, farnesylation, and/or geranylgeranylation of proteins such as Pex19p (18,26) and Rab5 (27) would be reduced as reported (28). Therefore, we verified the isoprenylation of Pex19p, Rab5, and Rab4 (Fig. 3A). When cells were treated with lovastatin, an HMGCR inhibitor, the non-farnesylated form of Pex19p and non-geranygeranylated form of Rab5 and Rab4 were recovered in the cytosol fraction, each with a slower mobility than the modified form in SDS-PAGE, consistent with a previous report (28). However, these unmodified Pex19p, Rab5, and Rab4 were not observed when plasmalogen levels were elevated in CHO-K1, hence indicating that isoprenoid synthesis was not altered by the elevation of plasmalogens.
We next determined the level of sterol intermediates in the cholesterol biosynthetic pathway. Lanosterol, the first sterol intermediate in cholesterol biosynthesis, is synthesized via condensation of isoprene to squalene and subsequent oxidation of squalene to MOS, followed by cyclization of MOS. 24,25-EC is generated de novo by a shunt of the cholesterol biosynthetic pathway through synthesis of 2,3;22,23-diepoxysqualene (DOS) from MOS (29). When CHO-K1 cells were metabolically labeled with [ 14 C]acetate, synthesis of MOS, cholesterol, and 24,25-EC was apparently reduced upon elevation of plasmalogens (Fig. 3B). In HeLa cells, MOS, lanosterol, cholesterol, and 24,25-EC were detected at a reduced level upon the elevation of plasmalogens. When lanosterol synthase (LSS) activity was partially inhibited by Ro48-8071, an inhibitor of LSS (30,31) in HeLa cells, synthesis of lanosterol and cholesterol was lowered, whereas marked accumulation of MOS and slight elevation of 24,25-EC were observed (Fig. 3C), distinct from the results obtained from the elevation of plasmalogens, suggesting that LSS activity is not suppressed upon the elevation of plasmalogens. Moreover, the squalene level was slightly elevated by the increase of plasmalogens in CHO-K1 and HeLa cells (Fig. 3, B and C). Taken together, these results suggest that the oxidation step catalyzed by SQLE is most likely down-regulated by the elevation of plasmalogens.  Elevation of Plasmalogens Causes Degradation of SQLE in a MARCH6-dependent Manner-We next assessed the expression level of SQLE in HeLa cells. Upon elevation of the plasmalogen level, expression of SQLE was reduced to about 50% as compared with the untreated cells, whereas the expression level of P450 reductase (P450R), an endoplasmic reticulum enzyme, was not altered by the elevation of plasmalogens. The reduced SQLE level was fully recovered to that of untreated cells by treatment of the cells with epoxomicin, a proteasomal inhibitor (Fig. 4, A, lane 4). The transcription level of SQLE, LSS, and HMGCR was not altered upon the elevation of plasmalogens (Fig. 4B), where the transcription level of these enzymes was lowered with rapamycin (24). Furthermore, cycloheximide chase experiments using HeLa cells in the presence of Etn revealed that elevation of plasmalogens stimulated degradation of SQLE as compared with the turnover of SQLE in the absence of Etn (Fig. 4C). Taken together, these results suggest that SQLE is specifically degraded upon elevation of plasmalogens in a proteasome-dependent manner.
SQLE is reported to be degraded by membrane-associated RING finger 6 (MARCH6) in a manner dependent on cholesterol (32,33). Therefore, we investigated whether plasmalogendependent degradation of SQLE is also mediated by MARCH6 by expressing MARCH6 or MARCH6C9A, an inactive form of  (Fig. 4D). Expression of MARCH6C9A-Myc 6 elevated the level of SQLE, whereas the expression of MARCH6-Myc 6 slightly reduced the SQLE level, consistent with the earlier studies (32). The expression level of MARCH6 was lower than MARCH6C9A, most likely due to the degradation mediated by autoubiquitination of wild-type MARCH6 (32,34). Degradation of SQLE was stimulated by expression of MARCH6-Myc 6 in the presence of Etn, which was partially inhibited by the expression of MARCH6C9A-Myc 6 . Moreover, plasmalogen-dependent degradation of SQLE was interfered by the treatment of dsRNA against MARCH6 (Fig. 4E), where the transcriptional level of MARCH6 was reduced about 50% (data not shown). In addition, turnover of a cholesterol-dependent degradation of SQLE in plasmalogen-elevated cells had a similar tendency to that in the cells cultured in the absence of Etn, where the cellular free cholesterol level was slightly lowered in plasmalogen-elevated cells (Fig. 4F). Taken together, these results strongly suggested that SQLE is degraded in a MARCH6-dependent manner upon elevation of plasmalogens.
Cholesterol Synthesis in Plasmalogen-deficient Cells-We next investigated cholesterol synthesis in plasmalogen-deficient cells such as fibroblasts from an ADAPS-deficient patient and an ADAPS-defective CHO mutant, adaps ZPEG251 (6, 13). Cholesterol was more efficiently synthesized than 24,25-EC in human fibroblasts from a normal control (Fig. 5A). In contrast, 24,25-EC was more effectively synthesized than cholesterol in fibroblasts from an ADAPS-deficient patient. In addition, there was the same tendency in the sterol synthesis in plasmalogendeficient CHO cell mutant, adaps ZPEG251 (Fig. 5B), as indicated in the ratio the 24,25-EC/cholesterol (Fig. 5B, lower  panel). These results suggest that the abnormally low level of plasmalogens causes suppression of cholesterol synthesis.
In the cholesterol synthetic pathway, MOS is converted to either cholesterol or further epoxidized to DOS that leads to the synthesis of 24,25-EC (35,36). Interestingly, MOS was under the detectable level in ZPEG251, which was restored to the normal level in ZPEG251/ADAPS-HA 2 cells stably expressing ADAPS-HA 2 (Fig. 5, B and C). Fatty alcohols accumulate in plasmalogen-deficient cells (6,37). Suppression of fatty alcohol synthesis by lowering the FAR1 expression did not restore MOS synthesis, implying that the absence of plasmalogens, but not the accumulation of fatty alcohols, results in the reduction of MOS synthesis (Fig. 5C). Together, SQLE more likely prefers to synthesize DOS in plasmalogen-deficient cells, thereby giving rise to the reduced level of MOS utilized for the synthesis of cholesterol.
To investigate the synthesis of DOS in plasmalogen-deficient cells, plasmalogen-deficient and -replete cells were treated with Ro48-8071, an inhibitor of LSS. Synthesis of 24,25-EC was mildly inhibited, whereas synthesis of cholesterol was efficiently suppressed, hence suggesting that Ro48-8071 partially inhibited the LSS activity (31). Upon such treatment, DOS accumulated more in ZPEG251 cells than CHO-K1 and ZPEG251/ADAPS-HA 2 cells (Fig. 5D). Taken together, these results suggest that the distinct reduction of cholesterol synthesis in plasmalogen-deficient cells is most likely due to the limited amount of MOS available for the synthesis of cholesterol.
Expression of SQLE Is Elevated in Plasmalogen-deficient Cells-Synthesis of 24,25-EC is modulated by the relative activities of SQLE and LSS (36). Partial inhibition of LSS activity or overexpression of SQLE stimulates 24,25-EC synthesis (31,38). However, accumulation of DOS was not observed in the absence of Ro48-8071 in ZPEG251 (data not shown), suggesting that the reduced activity of LSS is unlikely for the preferential DOS production in plasmalogen-deficient cells. Therefore, we investigated the expression level of SQLE in plasmalogendeficient cells including ADAPS-defective fibroblasts and ZPEG251 and found that SQLE was expressed at about a 2-fold higher level in plasmalogen-deficient cells than that in control fibroblasts and CHO-K1 cells (Fig. 6, A and B). Quantitative RT-PCR analysis showed that the mRNA expression level of SQLE in plasmalogen-deficient cells was not altered as compared with that in control cells (Fig. 6, A and B, lower panels). Furthermore, cycloheximide chase experiments revealed that SQLE in plasmalogen-deficient fibroblasts was more stable than that in control cells (Fig. 6C). Collectively, these results suggest that SQLE is stabilized in plasmalogen-deficient cells more likely by a post-translational mechanism.

SQLE Interacts with MARCH6 in a Plasmalogen-dependent
Manner-Knockdown of MARCH6 stabilizes SQLE under normal culture conditions and abolishes cholesterol-dependent degradation of SQLE (32,33). Therefore, we investigated if the absence of plasmalogens results in degradation or inactivation of MARCH6, giving rise to the elevation of the expressed levels of SQLE. However, addition of cholesterol in wild-type CHO-K1 and ZPEG251 similarly promoted degradation of SQLE (Fig. 7A), suggesting that expression and activity of MARCH6 was not altered in the absence of plasmalogens.
We further assessed whether plasmalogens regulate interaction of SQLE with MARCH6 by a coimmunoprecipitation assay. HA-tagged SQLE (SQLE-HA 2 ) plus MARCH6-Myc 6 or MARCH6C9A-Myc 6 were coexpressed in ZPEG251 and CHO-K1 cells, and subjected to immunoprecipitation using anti-Myc antibody. SQLE-HA 2 was coimmunoprecipitated with MARCH6C9A-Myc 6 in ZPEG251 and CHO-K1 (Fig. 7C), consistent with the earlier study (32). Moreover, a larger amount of SQLE-HA 2 was coimmunoprecipitated with MARCH6C9A-Myc 6 upon the elevation of plasmalogens in CHO-K1 cells. Similarly, the elevation of plasmalogens stimulates the interaction of MARCH6-Myc 6 with SQLE-HA 2 (Fig.  7B), where the interaction of SQLE-HA 2 with MARCH6-Myc 6 in CHO-K1 and ZPEG251 was less efficient. Together, these results suggest that plasmalogens regulate the stability of SQLE by modulating interaction of SQLE with MARCH6.

Discussion
In the present study, we show that the cellular plasmalogen level regulates cholesterol synthesis by modulating SQLE stability. Cholesterol synthesis is shown to be controlled at multiple steps, including sterol regulatory element-binding protein (SREBP)-mediated transcriptional regulation and post-translational regulation of HMGCR, a rate-limiting enzyme of cholesterol synthesis (39,40). In addition, recent studies revealed that the activity of SQLE is controlled at a post-translational level through the cholesterol-dependent ubiquitination and proteasomal degradation (12,32,33). In the present study, our finding that elevation of cellular plasmalogens also causes suppression of cholesterol synthesis in several cell lines (Figs. 1 and 2), implies that plasmalogen-dependent degradation of SQLE is a conserved mechanism for the regulation of cholesterol synthesis.
MARCH6 is an E3 ligase responsible for the degradation of SQLE in response to an exogenous cholesterol influx (32,33). SQLE expression is elevated by reduction of MARCH6 (32, 33) (Fig. 4E), expression of an inactive form of MARCH6 (33) (Fig.  4D), or treatment of cells with epoxomycin (Fig. 4A), hence suggesting that SQLE is constitutively degraded in a MARCH6dependent manner. Indeed, SQLE is constitutively degraded in SRD-1 cells (32), HeLa (Fig. 4C), and normal control fibroblasts (Fig. 6C). Moreover, SQLE interacts with MARCH6 without addition of external cholesterol (Fig. 7B) (32), where the interaction is increased upon elevation of cellular plasmalogens. Taken together, we suspect that the constitutive degradation of SQLE is more likely stimulated by the elevation of plasmalogens, but not cholesterol. Exogenous supplementation of cholesterol to the culture cells induces degradation of SQLE (12, 32, 33) by enhancing MARCH6-mediated ubiquitination (32). However, cholesterol-dependent degradation of SQLE in plasmalogen-deficient cells was likewise detected as that in wildtype cells (Fig. 7A). Moreover, elevation of plasmalogens did not synergistically augment the cholesterol-dependent degradation of SQLE in HeLa cells (Fig. 4F). Accordingly, we suggest that plasmalogens augment the interaction of SQLE with MARCH6 by a mechanism distinct from that involving cholesterol, although the mechanism underlying cholesterol-dependent interaction between SQLE and MARCH6 is not defined.
In humans, MARCH6 is expressed in several different tissues including heart, brain, kidney, and liver (41). However, only three proteins, including SQLE, HMGCR, and type 2 iodothyronine deiodinase, have been so far identified as a substrate for MARCH6-mediated degradation in mammals (32,33,42). However, MARCH6-mediated degradation of HMGCR seems to be independent of plasmalogens because isoprenylation of Pex19p, Rab5, and Rab4 (Fig. 3A), and squalene synthesis (Fig.  3, B and C) were not reduced by the increment of plasmalogens. Further studies including identification of more potential substrates of MARCH6 and the effects of plasmalogens on the degradation of such substrates are clearly required. Sterol-dependent degradation of SQLE seems to be a conserved mechanism in higher eukaryotes (43). Sterol-dependent degradation of ERG1, the SQLE homologue, is likewise mediated by MARCH6 homologue Doa10 in Saccharomyces cerevisiae (33). A similar mechanism was postulated (32, 44) from the findings that accumulation of squalene in the SQLE-defective plant was restored when a mutant allele of the plant homologue of MARCH6, SUD1, was crossed into this background (32,44). However, plasmalogens are not synthesized in plant (2,45), whereas only trace amounts of plasmalogens are detected in S. cerevisiae (46). Given these findings, we suggest that plasmalogen-mediated degradation of SQLE is specific for mammals. SQLE apparently spans the endoplasmic reticulum membrane via its hydrophobic segment located in the N-terminal region (12,47), although the membrane topology of SQLE is not yet defined. Interestingly, the region encompassing the 100-amino acid sequence including a potential transmembrane domain of human SQLE is sufficient for cholesterol-dependent degradation (12). However, this region is absent from yeast ERG1, and only a few amino acids are conserved between plant and human   1-4, 6, and 7) and adaps ZPEG251 (lanes 5 and 8), and cultured for 2 days in the presence (lanes 3 and 6) or absence (lanes 1, 2, 4, 5, 7, and 8)  SQLE. Accordingly, we propose that plasmalogens affect the interaction of SQLE with MARCH6 via the transmembrane domain in the N-terminal region of SQLE.
SQLE is proposed as the second rate-limiting enzyme in cholesterol synthesis (10 -12). Inhibition of SQLE efficiently reduces cholesterol synthesis (48,49), where accumulation of squalene does not cause any major adverse effects (50). Furthermore, the reduced level of cholesterol synthesis in plasmalogen-elevated cells is consistent with the lowered level of HMGCR activity (51) and cholesterol synthesis (12) in the cholesterol-elevated cells. Therefore, suppression of SQLE expression by elevation of plasmalogens might be an alternative potential way to reduce cholesterol synthesis without affecting the synthesis of physiologically important metabolites in the mevalonate pathway, including dolichol, ubiquinone, heme A, and prenylated proteins.
In the present study, we found that synthesis of cholesterol was specifically reduced in plasmalogen-deficient cells such as adaps CHO mutant ZPEG251, whereas 24,25-EC synthesis was elevated in plasmalogen-deficient ZPEG251 and fibroblasts derived from an ADAPS-deficient patient when cultured in the presence of FCS (Fig. 5). Our result is not compatible with that of the earlier study addressing newly synthesized cholesterol in the presence of lipoprotein-deficient serum (52). The HMGCR activity (51) and synthesis of cholesterol (52) are dramatically increased in the presence of lipoprotein-deficient serum, suggesting that the ablation of cholesterol synthesis in plasmalogen-deficient cells is hindered by stimulating mevalonate synthesis. Interestingly, endoplasmic reticulum stress in the PEX2 Ϫ/Ϫ mouse liver causes an elevation of cholesterol synthesis via activation of the SREBP-2 pathway (53-55), whereas cholesterol synthesis in the PEX5 Ϫ/Ϫ mouse is not affected (56). Transcription of SQLE and LSS, targets in the SREBP-2 pathway (36,57), were not elevated in adaps ZPEG251 and fibroblasts derived from a patient defective in ADAPS (Fig. 6, A and  B), suggesting that elevation of cholesterol synthesis in the PEX2 Ϫ/Ϫ mouse liver is caused by multiple peroxisomal dysfunctions as well as the defect in plasmalogen synthesis.
Expression of SQLE is at a very low level in most non-cholesterolgenic tissues, whereas SQLE is highly abundant in liver, followed by gut, skin, and neural tissue (58). Interestingly, the cellular plasmalogen level is very low in several tissues such as liver and small intestine (59). Therefore, the limited amount of plasmalogens in liver may contribute to the high level of SQLE expression by suppressing the MARCH6-mediated degradation of SQLE, thereby resulting in the efficient synthesis of 24,25-EC as observed in plasmalogen-deficient cells (Fig. 5).
Synthesis of 24,25-EC is regulated by the relative activity of SQLE and LSS (36). LSS activity is high in cholesterogenic and non-cholesterogenic tissues (58). Therefore, it is likely that synthesis of 24,25-EC is dependent on the activity of SQLE. However, the regulation mechanism of SQLE activity remains unknown. Synthesis of DOS and 24,25-EC is elevated in plasmalogen-deficient cells (Fig. 5) and 24,25-EC synthesis is increased by overexpression of SQLE (60). Taken together, the synthesis rate of 24,25-EC more likely depends on the expression level of SQLE, although we cannot exclude the possibility that unidentified post-translational modification of SQLE contributes to the regulation of SQLE activity.
24,25-EC is a physiological ligand for liver X receptor (LXR) in liver (61), playing a role in the reverse transport of cholesterol by stimulating transcription of ABCA1 encoding ATP-binding cassette transporter A1 (62)(63)(64)(65) and IDOL coding for the inducible degrader of the LDL receptor (66). Because 24,25-EC synthesis is elevated in plasmalogen-deficient cells, we suspect that a limited amount of plasmalogens in liver manipulates the synthesis of 24,25-EC toward the effective reverse cholesterol transport. Plasmalogen deficiency causes the impaired highdensity lipoprotein (HDL)-mediated cholesterol efflux from murine macrophage-like cells (67), the altered transport of internalized cholesterol to the endoplasmic reticulum (3), and abnormal cellular distribution of cholesterol (68). Moreover, addition of cis-(Ϯ)-2-O-docosahexaenoyl-1-O-hexadecylglycerol increases cholesterol esterification by raising the sterol-Oacyltransferase 1 expression level in HEK293 cells (69) and elevation of plasmalogens reduces the cholesterol level in HeLa cells cultured in the presence of exogenously added cholesterol (Fig. 4F). Taken these together with our findings, we suggest that the cellular plasmalogen level plays an important role at multiple steps in cholesterol homeostasis.
Plasmalogens are the major glycerophospholipids in brain and nerve tissue (70 -72). Cholesterol, which is required for the proper function of neuronal cells, is mainly provided by glial cells in the brain, because the blood-brain barrier separates the brain from the circulating cholesterol (73). Therefore, the regulation of plasmalogen synthesis most likely plays a pivotal role in the homeostasis of cholesterol especially in the central nervous system. Neurons do not efficiently synthesize cholesterol and mainly take up the cholesterol produced by astrocytes (74), hence suggesting that regulation of plasmalogen homeostasis unequivocally contributes to regulation of cholesterol synthesis in astrocytes. Our findings reported here open a way to address cholesterol homeostasis involving plasmalogen physiology.