| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Biology, Massachusetts Institute of Technology (A.Y., S.L.K., M.K.), Cambridge, Massachusetts 02139; Departamento de Gastroenterología, Facultad de Medicina, Pontificia Universidad Católica (M.G.M., L.A., S.Z., A.R.), Santiago 833-0024, Chile; and Biopharmaceuticals Center of Excellence for Drug Discovery, GlaxoSmithKline (M.H.D., K.F.K.), King of Prussia, Pennsylvania 19406
Address all correspondence and requests for reprints to: Dr. Monty Krieger, Department of Biology, Massachusetts Institute of Technology, Room 68-483, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139. E-mail: krieger{at}mit.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
HDL may also have a particularly important role in mammalian female fertility, because in many species, including humans, HDL is the main class of lipoprotein found in substantial amounts in the follicular fluid enveloping oocytes in ovarian follicles (16, 17, 18, 19, 20, 21). This is presumably due to the size limit imposed by the follicular basement membrane, which is known as the blood-follicle barrier (20). The follicular HDL might deliver key lipids to follicular cells or mediate cholesterol efflux from those cells (22). An important mechanism by which HDL cholesterol is delivered to ovarian cells, especially steroidogenic cells, is called selective lipid uptake, which is mediated by the HDL receptor SR-BI (2, 23, 24, 25).
SR-BI mediates cellular selective lipid uptake from HDL by binding to HDL via its apolipoprotein components and subsequently facilitating the net transfer from the core of the particle CE, but not the protein or most of the lipid components of the lipoprotein’s outer shell (2, 24, 25, 26). SR-BI can also mediate the bi-directional movement of UC between lipoproteins and cells (10, 27, 28, 29), and thus, in the presence of a UC gradient between cells and extracellular HDL, can mediate net cellular cholesterol efflux. Therefore, SR-BI, which is most highly expressed in the liver and steroidogenic tissues (24, 30), serves as a cholesterol transporter protein on the surfaces of cells. In the ovary, SR-BI is expressed in thecal cells, luteinized granulosa cells, and interstitial tissue (30, 31, 32, 33). In liver, but not steroidogenic tissues, SR-BI protein expression depends on the presence of an adaptor protein, PDZK1, that binds via one of its four postsynaptic density protein (PSD-95)/Drosophila disc large tumor suppressor (dlg)/tight junction protein (ZO1) (PDZ) domains to the C-terminal cytoplasmic domain of SR-BI (34, 35, 36, 37).
Mice with targeted homozygous null mutations in the SR-BI gene exhibit approximately 2-fold elevated plasma cholesterol carried in both normal size and abnormally large HDL particles (38, 39). The ratio of UC to total cholesterol (TC; UC:TC ratio) in plasma is abnormally high in SR-BI–/– mice (
0.5 vs. 0.25 in controls) (40, 41). When SR-BI–/– mice are crossed into an atherosclerosis-prone genetic background (39, 42) or fed an atherogenic diet (41, 43), they exhibit accelerated atherogenesis. Furthermore, SR-BI–/– females, but not males, are infertile (39, 44). SR-BI–/– females ovulate normal numbers of oocytes that are dead or defective and thus cannot be fertilized to form multicellular embryos. Although SR-BI plays a critical role in mediating the delivery of HDL cholesterol to steroidogenic cells to maintain CE stores and provide substrate for steroidogenesis (2, 45, 46, 47), insufficient ovarian steroid hormone synthesis does not appear to be responsible for the infertility of SR-BI–/– females (39, 44). For example, these animals exhibit a normal estrous cycle and a normal postmating increase in plasma progesterone (39, 44).
Our previous studies showed that when ovaries from SR-BI–/– mice are transplanted into otherwise SR-BI-positive recipients, they appear to function normally, and that genetic or pharmacological manipulation of lipoprotein metabolism in SR-BI–/– females can partially or fully restore fertility in the absence of SR-BI expression (44). These observations suggested that infertility of SR-BI–/– mice appears to be a direct consequence of abnormal lipoprotein metabolism and not due to ovarian SR-BI deficiency per se. In the current study we explored the hypothesis that hepatic SR-BI expression can play a critical role in murine female fertility by assessing the effects of adenovirus- or stable transgenesis-mediated hepatic expression of wild-type and mutant forms of SR-BI on the lipoprotein metabolism and fertility of SR-BI–/– mice. We found that hepatic SR-BI expression could correct the two main defects in the structure of HDL in SR-BI–/– mice (abnormally large size and UC:TC ratio) and restore female fertility to virtually wild-type levels. Thus, hepatic SR-BI-mediated lipoprotein metabolism influences murine female fertility, raising the possibility that some dyslipidemic disorders might contribute to human female infertility, and targeting lipoprotein metabolism in these conditions might provide a novel approach to complement current assisted reproductive technologies.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Generation of SR-BI-transgenic (SR-BI-Tg) mice
A 1.5-kb fragment spanning the wild-type murine SR-BI cDNA and a cDNA encoding a mutant form of SR-BI, SR-BIQ402R/Q418R (SR-BI-RR) were amplified by PCR from a plasmid (pDT188) that contains the cloned murine SR-BI cDNA (24) and from plasmid VM54 that have been described previously (48), respectively, using primers MunI-SRBI-new (TTATTCCAATTGCCGTCTCCTTCAGGTCCTGAGC) and SRBI-XhoI-rev (CAGCACCTCGAGGGCTTATAGTGTCTTCAGGACCCTA) and a DNA polymerase with proofreading activity, Pfu polymerase (Promega Corp., Madison, WI). A MunI site was introduced within the primer oligo at the 5' of the START codon of the SR-BI gene, and a XhoI site was introduced within the primer oligo at the 3' of the STOP codon of the SR-BI cDNA. The 1.5-kb cDNA fragment encoding the gene in its entirety was subcloned between the MunI and XhoI sites in the pLIV-LE6 plasmid, provided by Dr. John M. Taylor (Gladstone Institute of Cardiovascular Disease, University of California, San Francisco, CA). The SR-BI-coding region was confirmed by DNA sequencing using four primers spanning the gene. The pLiv-LE6 plasmid contains the promoter, first exon, first intron, and part of the second exon of the human apoE gene, and the polyadenylation sequence, and a part of the hepatic control region of the apoE/C-I gene locus (49). The new constructs harboring the wild-type and mutant copies of the SR-BI gene as well as the above-mentioned apoE/apoC-I control locus elements were linearized by NotI/SpeI digestion, and the resulting 6.5-kb fragments were used to generate transgenic mice by standard procedures (50) in C57BL/6 and 50:50 mixed C57BL/6 x 129-S4 background mice. Founder animals were backcrossed to 50:50 mixed C57BL/6 x 129-S4 background mice for three or four generations, and three transgenic mouse lines, SR-BI-Tghigh, SR-BI-Tglow, and SR-BI-RR-Tg, were established. The primers AB2 (GATGGGACATGGGACACGAAGCCATTCT) and AB3 (TCTGTCTCCGTCTCCTTCAGGTCCTGA) were used to amplify solely the genomic SR-BI. The presence of the SR-BI transgene was detected by primers B-Tg1 (TGAAGCTGATGATGACCTT) and B-Tg2 (AGCAGATGCGTGAAACTTGGTGA). Transgene expression was observed mainly in the liver, and some expression was observed in the kidney. The receptor in the kidney is expressed on the apical surfaces of the epithelial cells in the proximal tubules and therefore was found to face the tubular lumen, rather than the plasma, where it is not expected to influence substantially plasma lipoprotein metabolism. Indeed, results using liver-specific adenovirus-mediated transgenesis and these stable transgenic lines were concordant, fully supporting this assumption.
Lethicin:cholesterol acyl transferase (LCAT)-Tg mice that carry the human LCAT gene expressed from the albumin promoter on the C57BL/6 background (51) were purchased from The Jackson Laboratory (Bar Harbor, ME). These mice were crossed to SR-BI–/– males. The resulting offspring were crossed to each other to obtain the animals used in the study. Unless otherwise noted, all experimental animals and controls were on the same mixed C57BL/6:129-S4 background.
Recombinant adenovirus preparation and infection
The adenoviruses encoding wild-type murine SR-BI (Ad.SR-BI) and the control virus without a transgene (Ad.
E1) were described previously (52). The adenovirus encoding SR-BIQ402R/Q418R (Ad.SR-BI-RR) was prepared as previously described (53), where cDNA encoding SR-BIQ402R/Q418R was amplified by PCR from plasmid VM54 (48) and subcloned in the HindIII site of the pShuttle vector, then the recombinant adenoviral genome with the SR-BIQ402R/Q418R under control of the cytomegalovirus promoter was generated by homologous recombination in bacterial cells (65). Large-scale production of recombinant adenoviruses was performed using infected HEK293 cells as described previously (54). On d 0, SR-BI+/+ or SR-BI–/– mice (2–3 months old) on a 50:50 mixed C57BL/6 x 129-S4 background were injected via the femoral vein with PBS or 1 x 1011 viral particles of Ad.E1 for control groups or 1 x 1011 particles of Ad.mSR-BI or Ad.SR-BI-RR for experimental groups. Plasma, hepatic bile, and liver samples were harvested on d 3 after infection.
Hepatic bile, blood and liver sampling and processing
Mice were anesthetized with ip injection of pentobarbital (4.5 mg/100 g body weight) for adenovirus injections or by 2.5% Avertin for other studies. Hepatic bile was collected for 30 min while mice were kept under anesthesia at 37 C with a heating lamp. Blood was collected by puncture of the inferior vena cava or by heart puncture with a heparinized syringe. Plasma was separated by low speed centrifugation for 10 min at 3000 x g. Hepatic bile flow, calculated by dividing the volume of bile (determined gravimetrically assuming a specific density of 1.0) by the collection time and the liver weight, was expressed as microliters per minute per gram of liver. Bile and some plasma were stored at –20 C for lipid analysis, whereas liver samples were kept at –80 C before analysis by Western blotting.
Analysis of plasma cholesterol and lipoproteins
Plasma samples from single animals or pooled plasma where indicated were size fractionated by fast performance liquid chromatography (FPLC) as previously described (38), and TC and UC in each fraction or in unfractionated total plasma were determined using commercial kits (Wako Chemical USA, Inc., Richmond, VA) or as described previously (55). Slight differences in the peak positions of the HDL from SR-BI+/+ mice are the result of using different FPLC instruments (Fig. 2
vs. Figs. 5 and 7).
|
|
|
Immunoblotting analysis
Total liver membranes (40–50 µg protein/sample) were size-fractionated by 10% SDS-PAGE and immunoblotted on nitrocellulose or polyvinylidene fluoride membranes with either polyclonal antipeptide antibodies for SR-BI (24) or NB400–104 from Novus Biologicals (Littleton, CO) or 
coat protein subunit (-COP) of the coatomer complex COPI (58), used as a protein loading control. Antibody binding to protein samples was visualized by the enhanced chemiluminescence procedure (GE Healthcare/Amersham Biosciences, Arlington Heights, IL) and quantified with a Macintosh Color One scanner (Apple Computer, Cupertino, CA) and National Institutes of Health imaging software version 1.6. Relative levels of hepatic SR-BI protein expression were determined after normalization for -COP protein levels. To quantify the relative amounts of SR-BI in different tissues, tissue lysates were serially diluted and subjected to electrophoresis/immunoblotting, and the relative intensities of the signals were compared by visual examination of protein levels in these mice.
Immunohistochemical localization of hepatic SR-BI
For immunoperoxidase studies, tissues were fixed for 4 h in 4% paraformaldehyde in PBS (pH 7.4) at 4 C, then incubated overnight at 4 C in 30% sucrose in PBS (pH 7.4). Tissues were then frozen in OCT compound (Miles Diagnostics, Elkhart, IN) and stored in liquid nitrogen. Immunoperoxidase staining was performed on 5-µm frozen sections with primary anti-SR-BI antibody (24) and secondary biotinylated antirabbit IgG (1:200 dilution; Vector Laboratories, Inc., Burlingame, CA), and subsequently visualized with Vectastain ABC (Vector Laboratories, Inc.) and diaminobenzidine (Research Genetics, Inc., Huntsville, AL), according to the manufacturer’s protocol.
Fertility assays
Virgin females (6–8 wk of age) were housed continuously with wild-type males, and the numbers of litters and pups were recorded for a period of 4 months.
LCAT assays
We used two assays to measure LCAT activity in plasma samples.
Endogenous LCAT assay.
Trace amounts of [3H]cholesterol were equilibrated with the endogenous UC in the plasma lipoproteins, then LCAT activity was measured as previously described (59). Briefly, 10 µl plasma was diluted by addition of 40 µl TBS buffer [10 mM Tris (pH 8.0), 140 mM NaCl, 0.01% NaN3, and 0.01% EDTA] in a glass tube containing 500,000 cpm [3H]cholesterol (PerkinElmer, Inc., Wellesley, MA), which had been evaporated and dried at the bottom of the tube. The tubes were shaken vigorously overnight at 4 C. The distribution of labeled cholesterol among the plasma lipoproteins was found to be similar to that of the endogenous unlabeled cholesterol by FPLC chromatography (not shown). The samples were divided in half for simultaneous incubations with shaking for 30 min at either 37 C (experimental) or 4 C (controls). The reactions were stopped by the addition of 2 ml ethanol, and lipids were extracted twice with 4 ml hexane containing excess unlabeled cholesterol and cholesteryl oleate (50 µg each) as carriers. The amounts of labeled UC and CE were determined using thin layer chromatography and scintillation counting, and the percent cholesterol esterification rate (CE/TC) and the plasma cholesterol esterification rate (CER; nanomoles of cholesterol converted to cholesteryl ester per hour per milliliter of plasma) were calculated. The CER was calculated based on the specific activity of the unesterified cholesterol in the samples (determined from the mass of unesterified cholesterol in the samples measured by enzymatic methods), which varied significantly among the mouse strains used.
Exogenous LCAT assay.
LCAT activity was determined with an exogenous reconstituted discoidal HDL substrate prepared using the sodium cholate method with an initial molar ratio of 0.8:250:12.5 of apoA-I, egg phosphatidylcholine and [14C]cholesterol (60). Briefly, the reaction was initiated by adding 15 µl plasma to reconstituted HDL (22 µg human apoA-I), 0.5% BSA, and 5 mM ß-mercaptoethanol in a total volume of 0.515 ml. After a 20-min incubation at 37 C, the lipids were extracted and analyzed as described above. To ensure linearity, reactions did not proceed beyond 20% conversion of UC to CE. Esterification rates were corrected by subtracting values for reactions conducted without plasma. In addition, the specificity of the reaction was confirmed by complete inhibition with 1.3 mM 5',5'-dithio-bis(2-nitrobenzoic acid), an inhibitor of LCAT. Plasma LCAT activity with the exogenous substrate was determined by calculating the nanomoles of conversion of [14C]cholesterol to CE per hour per milliliter of plasma after correcting for the specific activity of the UC due to dilution by the endogenous UC present in the plasma samples, which was determined enzymatically as described above.
Statistical analysis
The statistical significance of the differences was determined by one-way ANOVA Tukey’s post hoc test, or two-tailed unpaired Student’s t test where appropriate. Differences were considered significant when P < 0.05.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Our initial studies focused on adenovirus-mediated hepatic overexpression of SR-BI. Adenovirus treatment resulted in hepatic expression of the receptors that was approximately 7- to 10-fold greater than in control wild-type mice (see supplemental Fig. 1, published on The Endocrine Society’s Journals Online web site, http://endo.endojournals.org). We determined the effects of hepatic overexpression on plasma cholesterol levels and biliary cholesterol secretion rates (Fig. 1) and the distribution of cholesterol among plasma lipoproteins as determined by FPLC size fractionation (lipoprotein profiles; Fig. 2
). As previously reported (52), Ad.SR-BI-mediated hepatic overexpression in wild-type SR-BI+/+ mice resulted in a dramatic reduction in plasma total cholesterol (primarily in HDL; Figs. 1A and 2A, 
) and a 2.2-fold increase in biliary cholesterol secretion (Fig. 1A). Similar results were observed in SR-BI–/– animals (Fig. 1B). Thus, hepatic overexpression of wild-type SR-BI effectively removes the cholesterol from normal-sized and abnormally large HDL particles in the circulation and/or prevents the formation of these particles by rapidly clearing cholesterol from their precursors. As a consequence of accelerated hepatic HDL cholesterol clearance mediated by SR-BI, biliary secretion of cholesterol increased.
|
), or the lipoprotein profile (Fig. 2A, 
), because the bulk of the plasma cholesterol is in normal-sized HDL particles not recognized efficiently by this mutant receptor. Strikingly, in SR-BI–/– animals (Fig. 1B), hepatic overexpression of SR-BI-RR reduced plasma cholesterol by approximately 60% to the levels observed in control SR-BI+/+ animals, apparently as a consequence of either clearing cholesterol from or preventing the formation of the larger abnormal particles without dramatically influencing normal HDL particles (Fig. 2B, 
). This Ad.SR-BI-RR effect in SR-BI–/– mice was accompanied by a 2.6-fold increase in biliary cholesterol secretion (Fig. 1B). It was somewhat surprising to observe comparable rates of SR-BI-mediated biliary secretion in the Ad.SR-BI- and Ad.SR-BI-RR-treated SR-BI–/– mice, because the wild-type receptor is so much more effective in reducing total plasma cholesterol. This raises the possibility that there may be an upper limit to the rate of biliary cholesterol secretion in these animals that should be explored in future studies. These results indicate that SR-BI-RR behaves in vivo as it did in vitro, in that this mutant receptor cannot efficiently process normal-size HDL, but may be able to mediate lipid uptake from (or prevent the formation of) larger lipoproteins, both LDL (in vitro studies) and the abnormal, large HDL particles in SR-BI–/– mice. Thus, SR-BI-RR may be useful to differentially address the biological consequences of preventing the accumulation of these abnormal particles in the plasma.
Similar effects of hepatic transgene expression were seen in stable transgenic lines, which are more suitable for the long-term studies of female fertility. We generated Tg lines expressing wild-type SR-BI protein at either 40-fold (SR-BI-Tghigh; see supplemental Fig. 2, published on The Endocrine Society’s Journals Online web site, http://endo.endojournals.org, for immunoblotting data) or approximately 0.7-fold (SR-BI-Tglow; data not shown), and the SR-BI-RR mutant at 40-fold higher than normal (SR-BI-RR-Tg; supplemental Fig. 2). The Q402R/Q418R mutations did not appear to interfere with the ability of the receptor to reach the plasma membrane (see immunohistochemical analysis in supplemental Fig. 2). Table 1 shows the influence of stable Tg expression of these receptors on plasma TC and UC levels and the UC:TC ratio in SR-BI+/– or SR-BI–/– mice (effects on lipoprotein profiles will be described in detail below). As was the case for adenovirus-mediated hepatic receptor overexpression, approximately 40-fold overexpression of wild-type SR-BI dramatically lowered plasma cholesterol levels in both SR-BI+/– and SR-BI–/– mice. Low level (0.7-fold) expression of the wild-type SR-BI transgene also lowered plasma cholesterol levels, but not as dramatically. In contrast, the approximately 40-fold overexpression of the SR-BI-RR transgene resulted in only modest reductions in plasma cholesterol to values similar to those in nontransgenic wild-type control animals (143 to 89 mg/dl for SR-BI+/– mice and 206 to 109 mg/dl for SR-BI–/– mice). The levels of apoA-I, the major apolipoprotein in HDL, in SR-BI–/– mice were not substantially different from those in SR-BI+/+ mice (supplemental Fig. 3) (39). This was also the case for SR-BI–/– [SR-BI-Tglow] and SR-BI–/– [SR-BI-RR-Tg] mice. In contrast, plasma apoA-I levels were very low in SR-BI–/– [SR-BI-Tghigh] mice, consistent with the suggestion that small lipid poor apoA-I particles that may be generated in these mice are rapidly cleared from the circulation (52, 65, 66).
|
|
show that both low-level expression of wild-type SR-BI (SR-BI-Tglow) and approximately 40-fold overexpression of SR-BI-RR in SR-BI–/– mice substantially reduced the amounts of abnormally large HDL-like particles. The low level of wild-type SR-BI was more effective than the approximately 40-fold overexpression of SR-BI-RR. (The effects of stable high-level expression of wild-type SR-BI, almost undetectable levels of cholesterol in all fractions, are not shown.) Strikingly, both low-level expression of wild-type SR-BI and approximately 40-fold overexpression of SR-BI-RR not only reduced the plasma TC values in SR-BI–/– mice to nearly wild-type levels and resulted in lipoprotein profiles similar to those in wild-type mice, but also normalized the UC:TC ratio from its characteristically abnormally high value of 0.52 (40, 41) to 0.23–0.25 (Table 1). Figure 3B shows that the abnormally high UC:TC ratios present throughout the HDL size range in control SR-BI–/– mice were normalized by expression of the SR-BIlow and SR-BI-RR transgenes. (UC levels were below detection limits in SR-BI-Tghigh animals.)
Figure 4 shows that Tg hepatic expression of wild-type SR-BI (both high and low levels) or SR-BI-RR restored nearly normal fertility to female SR-BI–/– mice, enumerated during 4-month mating periods either as the percentage of females producing litters (Fig. 4A) or the number of pups delivered per month per female (Fig. 4B). In addition, the times from the beginning of mating to delivery of the first litter were similar for the nontransgenic control SR-BI+/– and Tg SR-BI–/– females, ranging from 21–23 d. These results are consistent with our previous findings that SR-BI expression in the ovary or other female reproductive organs is not required per se for female fertility and strongly support the proposal that the presence of abnormal lipoproteins in female SR-BI–/– mice is responsible for their infertility (44). These findings were supported by preliminary studies using SR-BI–/– mice treated with adenoviruses. Fertility was restored in a large fraction of the mice injected with Ad.SR-BI encoding wild-type SR-BI and then mated 5 d later (our unpublished observations).
|
Figure 5A shows that the exogenous LCAT activities of control SR-BI+/+ and SR-BI–/– mice were similar (35 and 33 nmol CE formed/h/ml plasma, respectively). Thus, reduced levels of LCAT activity are not responsible for the abnormally high UC:TC ratio in SR-BI–/– mice. Unlike the exogenous LCAT activity, the endogenous LCAT activity in plasma from SR-BI–/– mice was significantly less than that in SR-BI+/+ controls (Fig. 5B). Thus, it appears that the abnormal lipoproteins in SR-BI–/– plasma may be poor substrates for LCAT, and this might contribute to the abnormally high UC:TC ratio in these animals. Low endogenous LCAT activity was also reported by Ma et al. (68).
In an attempt to correct the abnormally high UC:TC ratio in SR-BI–/– mice and possibly restore fertility, we crossed these mice to Tg mice (SR-BI+/+[LCAT-Tg]) expressing a high level of the human LCAT enzyme in addition to the endogenous murine enzyme (51). The exogenous LCAT activities in the LCAT-Tg mice (Fig. 5C) were approximately 7- to 11-fold higher than those in nontransgenic controls (compare to Fig. 5A, note different scales of the y-axes), with the activity in the SR-BI–/–[LCAT-Tg] mice somewhat higher than that in the SR-BI+/+[LCAT-Tg] controls. Indeed, immunoblotting experiments indicated higher LCAT protein levels in the plasma of SR-BI–/–[LCAT-Tg] animals compared with those in SR-BI+/+[LCAT-Tg] controls, possibly due to the higher LCAT-binding capacity of the larger plasma lipoproteins in the receptor-deficient animals (data not shown). As expected from previous studies (51, 69), the LCAT transgene increased the plasma TC level in SR-BI+/+ mice by 60% (Table 2). Although the much smaller transgene-associated increase (18%) in SR-BI–/–[LCAT-Tg] mice was not statistically significant (Table 2), it was accompanied by an increase in the size of the already abnormally large HDL particles (FPLC analysis; shift to the left in Fig. 6). Analysis of the UC:TC ratios (Table 2) showed that expression of the LCAT transgene did not change this already low ratio (0.18) in SR-BI+/+ controls. The transgene significantly reduced the abnormally high UC:TC ratio from 0.44 in SR-BI–/– mice to 0.37 in SR-BI–/–[LCAT-Tg] mice; however, it was still twice that in SR-BI+/+[LCAT-Tg] controls. Thus, an approximately 10-fold increase in plasma LCAT activity was only able to partially overcome the apparent cholesterol esterification defect in SR-BI–/– mice.
|
|
shows the effects of Tg human LCAT expression on the fertility of SR-BI–/– and SR-BI+/– females. There was virtually no effect of the LCAT transgene on the fertility of control SR-BI+/– mice. The LCAT transgene did not significantly restore fertility, as measured by pups per month per female (Fig. 7B). All SR-BI+/–[LCAT-Tg] females (five of five) produced litters, whereas only two of 15 SR-BI–/–[LCAT-Tg] females each produced a single litter, and those were very small (two vs. nine pups per litter). Thus, the very modest decrease in the abnormally high UC:TC ratio induced by the approximately 10-fold increase in plasma LCAT activity in SR-BI–/–[LCAT-Tg] mice was not associated with a substantial increase in female fertility. |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
As we (52) and others (63, 70, 71) previously showed, we found in this study that high level hepatic overexpression of wild-type SR-BI dramatically reduced plasma TC in wild-type and SR-BI–/– mice to nearly undetectable levels and increased biliary cholesterol levels or secretion rate. In contrast, high-level hepatic expression of the SR-BI-RR mutant had virtually no effect on plasma cholesterol levels or biliary cholesterol secretion in SR-BI+/+ mice, presumably because the normal-sized HDL that carries the bulk of the plasma cholesterol in these mice is not an efficient ligand for the mutant receptor. However, in SR-BI–/– mice, hepatic overexpression of SR-BI-RR was able to reduce plasma cholesterol to nearly wild-type levels, lower to essentially wild-type values the abnormally high UC:TC ratio, correct the lipoprotein profile, and induce a substantial increase in biliary cholesterol secretion. Thus, in SR-BI–/– mice high-level hepatic expression of the SR-BI-RR mutant substantially normalizes the otherwise abnormal lipoprotein size and lipid composition by effectively removing cholesterol from abnormally large HDL particles and/or preventing the formation of these particles or their precursors. Therefore, in addition to LDL particles that are larger than normal HDL, and in contrast to normal-sized HDL, the abnormally large HDL particles in SR-BI–/– mice (or their precursors) appear to serve as effective substrates for the SR-BI-RR mutant. The SR-BI–/–[SR-BI-RR-Tg] Tg mice should prove useful for future analysis of the pathophysiological properties of abnormal HDL in SR-BI–/– mice, including defective red blood cell maturation and enhanced susceptibility to atherosclerosis and coronary heart disease (39, 40, 72).
Concurrent with the changes in the plasma lipoprotein profiles, stable Tg hepatic expression of high or low levels of wild-type SR-BI or high levels of the SR-BI-RR mutant restored essentially normal fertility to SR-BI–/– females. Together with our previous studies involving ovary transplantation, inactivation of the apoA-I gene, or treatment with the hypocholesterolemia drug probucol, all of which restored production of functional oocytes by SR-BI-deficient ovaries and thus partially (apoA-I–/–) or fully (ovary transplantation or probucol treatment) restored fertility (44), these results establish that a primary mechanism by which SR-BI maintains murine female fertility is its control of plasma lipoprotein metabolism rather than its expression in reproductive organs.
The two most distinctive abnormal features of the HDL in SR-BI–/– mice that may influence female fertility are the substantial fraction of HDL that is abnormally large and the high UC:TC ratio in both large and normal-sized particles. However, several lines of evidence indicate that large HDL particles per se are not responsible for female infertility in SR-BI–/– mice. First, treatment of SR-BI–/– mice with the antioxidant and hypocholesterolemic drug probucol, which lowers plasma TC by approximately 50%, normalizes the UC:TC ratio and fully restores fertility, does not substantially alter the abnormally large HDL particle size distribution (40, 44). Second, inactivation of the apoA-I gene in SR-BI–/– mice, which lowers total plasma cholesterol by approximately 50% and partially restores fertility, reduced the amounts of normal, smaller HDL particles relative to the abnormal large particles, resulting in a mean particle size somewhat greater than that in apoA-I-replete controls (44). Third, female PDZK1–/– mice are fertile (34, 73). Hepatic SR-BI protein levels in PDZK1–/– mice are approximately 5% those in wild-type controls, whereas SR-BI protein levels in the ovary and other steroidogenic organs are normal (34). As a consequence of the reduced hepatic SR-BI activity, plasma cholesterol is elevated in abnormally large HDL particles, somewhat reminiscent of those in SR-BI–/– animals. However, the UC:TC ratio is essentially normal in PDZK1–/– mice (34), perhaps as a consequence of the nearly normal extrahepatic SR-BI expression or the normal hepatic expression of the minor, alternatively spliced mRNA isoform called SR-BII that is not present in SR-BI–/– mice. Thus, large HDL particles in PDZK1–/– mice are not associated with reduced fertility.
Although the presence of abnormally large HDLs per se is unlikely to cause female infertility in SR-BI–/– mice, there is a striking correlation between infertility and the abnormal UC:TC ratio. A key determinant of the UC:TC ratio is the activity of LCAT, which is the plasma enzyme responsible for esterification of UC in HDL particles (51, 69, 74). Although the absolute level of LCAT in SR-BI–/– mice appears to be the same as that in control SR-BI+/+ animals when measured using an exogenous reconstituted HDL substrate (this study; also see Ref.68), there was significantly reduced LCAT activity in SR-BI–/– mice (50%) when assessed using the abnormal endogenous lipoproteins as the substrate. Previous studies (38, 39), confirmed here, have shown that the plasma levels of apoA-I, the major activator of LCAT in plasma, are normal in SR-BI–/– and SR-BI–/–/apoE–/– mice. Furthermore, hepatic SR-BI or SR-BI-RR transgene expression, which restored fertility, did not increase plasma apoA-I levels. Thus, reduced apoA-I-mediated activation of LCAT due to lower plasma apoA-I levels is unlikely to be responsible for the high UC:TC ratio. We conclude that the abnormal plasma lipoproteins in SR-BI–/– are poor substrates for LCAT, and that this substantially contributes to the abnormally high UC:TC ratio. While this manuscript was in preparation, similar findings (essentially wild-type levels of LCAT protein, but reduced LCAT activity with endogenous substrate) were reported by Ma et al. (68). In SR-BI–/–/apoE–/– double-knockout mice, the HDL-like particles are even larger than those in SR-BI–/– single knockout mice, forming lamellar/vesicular and stacked discoidal particles, and the UC:TC ratio is higher (0.8 vs. 0.5) (40). It seems likely that the particles in SR-BI–/–/apoE–/– double-knockout mice are especially poor LCAT substrates.
In an attempt to correct the high UC:TC ratio, we crossed SR-BI–/– mice with Tg mice expressing human LCAT in addition to endogenous murine LCAT. Although this resulted in an approximately 10-fold increase in LCAT activity (determined using exogenous substrate), the LCAT transgene only reduced the UC:TC by 16% to 0.37 and failed to restore female fertility. Although species differences in the specificities of the human and murine LCAT enzymes (75) may have influenced the results, our findings suggest that the abnormal lipoproteins in SR-BI–/– mice exhibit a high UC:TC ratio because they are poor LCAT substrates and that SR-BI activity normally helps maintain HDL in a state in which it can serve as an efficient LCAT substrate.
Determination of the precise mechanism by which abnormal lipoprotein metabolism in SR-BI–/– mice results in female infertility will require additional studies. The key lipoprotein-dependent step appears to occur before or immediately during ovulation, because SR-BI–/– females ovulate dysfunctional oocytes that are dead at ovulation or die shortly thereafter (39, 44). There appear to be two general classes of mechanisms by which the abnormal lipoproteins in SR-BI–/– mice might contribute to female infertility. The first involves gain of function mechanisms, in which the abnormal lipoproteins are somehow ovariotoxic. In the absence of ovarian SR-BI, the abnormal lipoproteins might deliver molecules to ovarian cells that are directly toxic (this could include delivery of excessive amounts of cholesterol) (76, 77, 78, 79) or remove essential compounds, the depletion of which would be toxic, resulting in defects in oocyte development/ovulation.
Alternatively, female infertility could be a consequence of a loss of function mechanism in which the abnormal lipoproteins are unable to perform the function(s) of wild-type lipoproteins. In the absence of SR-BI, the abnormal lipoproteins might be unable to deliver, or remove by efflux, key compounds to/from the ovary, for example, HDL-facilitated vitamin E uptake (80) or HDL-mediated efflux of cholesterol. Particularly, the abnormal lipoproteins in SR-BI–/– mice might be unable to facilitate removal of excess cellular cholesterol, because their high UC:TC ratio would prevent thermodynamically driven cellular cholesterol efflux down a cholesterol concentration gradient (10, 28, 29, 81). The pathogenic accumulation of intracellular cholesterol, due to either defective cholesterol efflux or abnormal enhanced delivery, appears to contribute to defective red blood cell maturation in SR-BI–/–/apoE–/– double-knockout mice (79). UC efflux from the cells in the ovary may be particularly important during the preovulatory period when granulosa cells produce progesterone after the LH surge. Progesterone has been shown to inhibit cholesterol esterification in vitro (82, 83) at concentrations that are present in the ovary (84). Thus, it has been suggested that cholesterol efflux may help cells avoid UC accumulation (85). In humans, HDL in follicular fluid is in the cholesterol-poor pre-ß form (86), a good acceptor for cellular efflux of excess UC from the ABCA1 transporter (8). In addition, LCAT is present in follicular fluid and may contribute to efficient efflux by preventing UC accumulation in the HDL (85, 87). Because of the limited size of murine follicles, it has been difficult to characterize their lipoprotein composition and metabolism.
The role of HDL-mediated cholesterol efflux may be especially relevant in SR-BI–/– mice, because SR-BI can directly mediate cholesterol efflux (10, 28, 81) as well as selective lipid uptake (24) and thus contribute to the bidirectional flow of lipids to maintain a local balanced environment of cholesterol. The relationship between female infertility in SR-BI–/– mice and the functions of two other cell surface proteins that mediate cholesterol efflux, ABCA1 (88, 89) and ABCG1 (90), is unclear. However, it is noteworthy that although ABCG1–/– mice exhibit normal fertility (90), ABCA1–/– females exhibit impaired fertility (placental malformation) (91) and have very low plasma HDL levels (92, 93). It is possible that the normal ovarian expression of SR-BI itself, even if it were to occur in the context of the abnormal lipoproteins present in SR-BI–/– mice, might mitigate the deleterious influence of these lipoproteins on female fertility (e.g. by facilitating lipid transport between the lipoproteins and ovarian cells). Thus, ovarian SR-BI expression might contribute to the fertility of PDZK1–/– female mice even though hepatic SR-BI expression is dramatically reduced, and abnormally large HDLs accumulate in the circulation (34). Our observation that Tg hepatic expression of SR-BI or SR-BI-RR in SR-BI–/– females at least partially corrects their lipoprotein abnormalities and restores fertility is compatible with either class of mechanism. It is clear that not all variations in HDL structure or metabolism will substantially reduce female fertility. For example, there is no evidence of female infertility in species with naturally low HDL levels (e.g. guinea pigs) or in animals with experimentally generated low plasma HDL, such as apoA-I-deficient or Tg, hepatic SR-BI-overexpressing mice (63, 70, 94). Thus, even small steady-state levels of plasma HDL, provided these particles have the appropriate size and composition or metabolic fates, can be sufficient to prevent ovarian toxicity (e.g. accumulation of toxic cholesterol levels).
In summary, the current studies confirmed the importance of the hepatic expression of SR-BI for murine lipoprotein metabolism and support our conclusion that the structure of HDL and its role in cholesterol transport and metabolism may be critical in determining mammalian female fertility. These findings raise the possibility that dyslipidemia might contribute to some cases of human female infertility, and in such cases, targeting dyslipidemia might complement current assisted reproductive technologies.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
All authors have nothing to declare.
First Published Online January 12, 2006
Abbreviations: Ad, Adenovirus; apoA-I, apolipoprotein A-I; apoE, apolipoprotein E; CE, cholesteryl ester; CER, cholesterol esterification rate; FPLC, fast performance liquid chromatography; HDL, high-density lipoprotein; LCAT, lethicin:cholesterol acyl transferase; LDL, low-density lipoprotein; PDZ, postsynaptic density protein (PSD-95)/Drosophila disc large tumor suppressor (dlg)/tight junction protein (ZO1); SR-BI, scavenger receptor, class B, type I; TBS, Tris-buffered saline; TC, total cholesterol; Tg, transgenic; UC, unesterified cholesterol; VLDL, very low-density lipoprotein.
Received October 11, 2005.
Accepted for publication January 4, 2006.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
; n = 5), SR-BI–/–[SR-BI-Tghigh] (
; n = 7), SR-BI–/–[SR-BI-RR-Tg] (
, n = 10). A, Fertility, expressed as the percentage of females producing litters (the numbers of females that produced litters relative to the number of females studied are shown above the bars). B, Average number of pups delivered per month per female mated (error bars represent the SEM). *, Statistically significant difference from the other values by ANOVA Tukey’s post hoc test (P < 0.05). The mean litter sizes, which include only those females that delivered pups, are shown below.