LXR and ABCA1 control cholesterol homeostasis in the proximal mouse epididymis in a cell-specific manner.

Mammalian spermatozoa undergo important plasma membrane maturation steps during epididymal transit. Among these, changes in lipids and cholesterol are of particular interest as they are necessary for fertilization. However, molecular mechanisms regulating these transformations inside the epididymis are still poorly understood. Liver X receptors (LXRs), the nuclear receptors for oxysterols, are of major importance in intracellular cholesterol homeostasis, and LXR−/−-deficient male mice have already been shown to have reduced fertility at an age of 5 months and complete sterility for 9-month-old animals. This sterility phenotype is associated with testes and caput epididymides epithelial defects. The research presented here was aimed at investigating how LXRs act in the male caput epididymidis by analyzing key actors in cholesterol homeostasis. We show that accumulation of cholesteryl esters in LXR−/− male mice is associated with a specific loss of ABCA1 and an increase in apoptosis of apical cells of the proximal caput epididymidis. ATP-binding cassette G1 (ABCG1) and scavenger receptor B1 (SR-B1), two other cholesterol transporters, show little if any modifications. Our study also revealed that SR-B1 appears to have a peculiar expression pattern along the epididymal duct. These results should help in understanding the functional roles of LXR in cholesterol trafficking processes in caput epididymidis.


Nile Red staining of lipids
Seven-micrometer-thick cryosections were mounted on Super-frost® plus glass slides (Menzel Glaser® Templemars, Paris, France). Cryosections were equilibrated for 5 min in PBS, fi xed for 5 min in 4% paraformaldehyde-PBS, rinsed in PBS, and then incubated at 4°C in a humidifi ed chamber with the Nile Red solution (2 µg/ml in PBS). Sections were counterstained with Hoechst 33342 solution (1 µg/ml; Sigma-Aldrich, St Quentin Fallavier, France) and mounted with coverslips using PBS-glycerol (v/v) as mounting medium. Nile Red is only fl uorescent when incorporated in a hydrophobic environnement. For cells stained with Nile Red and excited with blue (wavelength: 488 nm), membranes fl uoresce in deep red (wavelength: >650 nm) due to PLs. Endosomes containing neutral lipids, such as triglycerides, cholesterol, or CEs, fl uoresced in green (wavelength: ≈ 530 nm).

Quantifi cation of mRNA by real-time quantitative RT-PCR
Real-time quantitative PCR was performed with the Bio-Rad I-Cycler and the IQ TM SYBR® Green Supermix Bio-Rad mix (Bio-Rad, Marne-la-coquette, France). Two microliters of 1:50 diluted cDNA template were amplifi ed by 0.5 units of HotMaster TaqDNA polymerase (Bio-Rad) using SYBR Green dye to measure duplex DNA formation following the manufacturer's instructions. Primer sequences are given in Table1 . Quantifi cation was obtained from the relation between the threshold cycle value and a standard curve ( 16 ).

Immunohistochemistry
Seven-micrometer-thick paraffi n sections were mounted on Superfrost® glass slides and then deparaffi nized with Histoclear for 40 min (National Diagnostic, Merck Eurolab, Fontenay-sous-Bois, France), rehydrated through a graded series of ethanol solutions, and fi nally rinsed in distilled water.
Peroxydase detection. To inhibit endogen peroxidases, slides were placed 30 min in 0.3% H 2 O 2 in water. Sections were then treated using the Vectastain® ABC Kit Rabbit peroxydase IgG (Vector Laboratories, Abcys, Paris, France) according to the manufacturer's instructions. The primary antibody, rabbitpolyclonal anti SR-B1 (1/2,000 in PBS-Normal Goat Serum 0.1%; Novus Biological, Interchim, Montluçon, France), was incubated overnight at 4°C in a humidifi ed chamber. The revelation was obtained with the Vector NovaRED substrate kit for peroxydase (Vector Laboratories) for 10 min. Slides were counterstained with Hematoxylin QS (Vector Laboratories) for 20 s and mounted with coverslips using Cytoseal 60 mounting medium (Electron Microscopy Sciences, Hatfi eld, USA).

In situ cell death detection
After deparaffi nization as described above, sections were treat ed with the In Situ Cell Death Detection Kit (Roche female genital tract and that are fundamental for fertilization to occur ( 7 , 8 ). During capacitation, cholesterol depletion appears to be critical for the activation of tyrosine kinases leading to changes in protein conformation of the sperm membrane and cholesterol effl ux leading to Ca 2+ infl ux, which in turns allows the development of motility ( 9 ). Membrane cholesterol is thus a key player in the acquisition of spermatozoa fertilizing potential; paradoxically, molecular events regulating this process are still poorly understood.
Among the various transcription factors that regulate cholesterol homeostasis, liver X receptors (LXRs) have specifi c roles. While LXR ␣ (NR1H3) is expressed in tissues that have an important lipid metabolism, such as testis and liver, LXR ␤ (NR1H2) has a rather ubiquitous expression. They are both activated by oxysterols ( 10 ). LXRs have been associated with various physiological functions (for a review, see Ref. 11 ), such as reverse cholesterol transport by the induction of the ATP-binding cassette proteins ABCA1 and ABCG1. The importance of LXR in the male reproductive physiology was underlined by the LXR Ϫ / Ϫ mice model. These mice showed male infertility starting at around 5 months of age and progressively became sterile after 9 months ( 12 ). In the testis, the loss of germ cells, a decreased testicular testosterone level, and lipid accumulation were described ( 13 , 14 ). In addition, considerable destructuration of the caput epididymidis epithelium was observed, characterized by a reduction of cell height specifi cally located in proximal segments 1 and 2. Moreover, loss of LXR was shown to result in perturbations of caput epididymidis lipid content with cholesteryl ester (CE) accumulations located in the epithelium as well as in peritubular tissue ( 15 ). In addition, spermatozoa from 9-month-old LXR Ϫ / Ϫ male mice and older showed a loss of their fl agella in most of the cases, probably due to excessive fragility of their mid-piece ( 15 ).
This article further investigates the roles of LXR in caput epididymidis physiology. The data presented here show that LXRs play a central role in epididymal cholesterol homeostasis by regulating ABCA1 in a segment and cell-type-specifi c manner, providing new insights about molecular mechanisms regulating sperm cell maturation process and fertility.

Wild-type and LXR
Ϫ / Ϫ male mice were euthanized by decapitation as previously described ( 13 ). These mice came from a hybrid line C57BL6 × 129 SVJ and were housed in an animal facility with controlled environment (temperature = 22°C, 12 h light/12 h dark). In normal conditions, these mice were fed ad libitum with Global-diet-2016S (cholesterol free) from Harlan (Gannat, France). For the Western diet, animals were fed during 7 weeks ad libitum with a 1.25% cholesterol-enriched diet (Safe, Augy, France). Housing and manipulation of these animals were approved by the Regional Ethic Committee in Animal Experimentation (authorization CE2-04).
were characterized in quality and quantity by comparing their retention times to those obtained from a known mixture (MIX37 from Sigma-Aldrich).

Statistical analyses
A Student's t -test was performed to determine signifi cant differences between groups with P value < 0.05.

Age-related CE accumulation in caput epididymidis of LXR ؊ / ؊ mice
Previous studies had shown that homozygous deletion of lxr resulted in CE accumulation in the epididymis of the LXR Ϫ / Ϫ animals ( 15 ). To determine which cells among the different cell types constituting the epididymal epithelium ( Fig. 1A ) are concerned by these accumulations, we used Nile Red staining on caput epididymidis sections of 4-and 9-month-old wild-type and LXR Ϫ / Ϫ mice. No neutral lipid accumulation was observed in caput epididymidis segments 1 and 2 from wild-type mice, whereas lipid droplets were visible in LXR Ϫ / Ϫ mice at 4 and 9 months of age ( Fig. 1B ) in the same segments. These accumulations were mainly localized in interstitial cells surrounding the tubules as well as in one subtype of epithelial cells: the socalled apical cells ( Fig. 1C , arrows). These neutral lipid stainings in apical cells were limited to segments 1 and 2 of the caput epididymides (data not shown).

Cholesterol de novo synthesis does not seem to be responsible for CE accumulation
To check whether increased de novo cholesterol synthesis was responsible for CE increase, the expression of genes encoding cholesterol and CE producing enzymes was measured by quantitative RT-PCR. This included acyl-CoA cholesteryl acyltransferase1 ( acat1 ) and acat2 , which esterify free cholesterol and store CE in lipid droplets; sterol regulatory element binding factor 2 ( srebp2 ), which regulates the de novo synthesis of cholesterol; and 3-hydroxy-3-methyl-glutaryl-coA-reductase ( hmg-coA-red ) and 3-hydroxy-3-methyl-glutaryl-coA-synthase ( hmg-coA-synt ), the rate-controlling enzyme s of the mevalonate pathway that produces cholesterol and other isoprenoids. Expression levels of acat1 , acat2 , and hmg-coA-red were not signifi cantly modifi ed ( Fig. 2 ), while srebp2 and hmg-coA-synt expression was decreased by 42 ± 5% ( P < 0.01) and 48 ± 6.5% ( P < 0.01), respectively, in Diagnostics, Grenoble, France) according to the manufacturer's instructions. Briefl y, the principle of this kit is the labeling of DNA strand breaks by terminal deoxynucleotidyl transferase, which catalyzes polymerization of labeled nucleotides to free 3 ′ -OH DNA ends in a template-independent manner [terminal desoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) reaction].

Gas chromatography
Total lipids from caput epididymides were extracted using chloroform/methanol based on the Folch method and then diluted in 100 µl chloroform. Total lipids were separated into nonpolar lipids or neutral lipids and PLs by the using Sep-pak® column silica cartridges (Sep-Pak, vac 1 cc, 100 mg; Waters, Guyancourt, France) as described by Juaneda and Roquelin (17). Briefl y, after washing the column with 4 ml of chloroform, the samples of total lipid extracts were loaded on the top of the cartridges. Nonpolar lipids were eluted with 4 ml of chloroform, following which the fraction containing the PLs was eluted with 8 ml of methanol. Lipids were then evaporated under a nitrogen fl ux and diluted in 100 µL of toluene and 200 µl methanol. Lipids were methylated for 20 min at 20°C with 100 µl of 2 N sodium methanolate (Sigma-Aldrich), followed by a 20 min incubation with 500 µl 14% BF3/methanol (Sigma-Aldrich). After washing with saturated NaHCO 3 , methyl esters were extracted with 2 ml hexane, vortex and evaporation of the top hexanic phase. Before analysis, the neutral lipid fraction was loaded on a Florisil cartridge (Chromabond® 6 ml/500 mg; Macherey-Nagel, Hoerdt, France) preconditioned with hexane and then eluted with a mixture of hexane:ethylic ether (95:5, v:v). Methyl esters were concentrated, diluted in 200 µl hexane, and stored at Ϫ 80°. Before analysis, samples were fi ltered on a fl orisil column, eluted with a mixture of hexane:ethylic ether (95:5, v:v), evaporated, and diluted in a known volume of hexane. The analysis was performed on a GC trace gas chromatograph (Thermo Electron, Courtaboeuf, France) equipped with a capillary DBWAX column (30 m, 0.25 mm, 0.25 µm thick; JW Scientifi c, Folsom, CA). The injector was a split-splitless type and the detector a fl ame ionization detection type. Fatty acid methyl esters  Abca1  NM_013454  GGAGCTGGGAAGTCAACAAC  ACATGCTCTCTTCCCGTCAG  56°C  176 bp  Abcg1  NM_009593  GCTGTGCGTTTTTGGCTGTT  TGCAGCTCCAATCAGTAGTCCTAA  63°C  84 bp  Acat1  NM_144784  ATTTGCTGATGCTGCCGTAG  TTTTGGGGGTCAATCTCCAG  64°C  177 bp  Acat2  NM_005891.2  CACCCCAGCGAACGCATC  GAGGGTATTGTCTTCCGAGG  59°C  147 bp  Cyclophillin  NM_008907.1  GGAGATGGCACAGGAGGAA  GCCCGTAGTGCTTCAGCTT  62°C  75 bp  hmgcoaR  NM_008255  CTTGTGGAATGCCTTGTGATT  AGCCGAAGCAGCACATGAT  64°C  75 bp  hmgcoaS  NM_145942  TGACATGCTCTCCGAGTACC  CAGGAACATCCGAGCTAGAG  64°C  220 bp  Srebp2  NM_033218  GTTGACGCAGACAGCCAATG  CCCTTACTGGCACTTGAAGG  64°C  241 bp tion pattern. In segment 1, SR-B1 was localized in the vascular endothelium ( Fig. 3B , arrows) and in the apical and basal membranes of the epithelium ( Fig. 3B ). The same localization was observed in segment 2, but apical membrane staining systematically appeared stronger. In segment 3, staining was diffuse in the cytoplasm of epithelial cells, sometimes presenting stronger punctual staining that could have been due to the presence of SR-B1 on intracellular organelles, such as endosomes. No change LXR Ϫ / Ϫ mice compared with the wild-type. Endogenous synthesis of cholesterol was thus not responsible for CE accumulation.
Cholesterol uptake does not seem to be responsible for CE accumulation Cholesterol uptake was investigated by studying the presence of SR-B1 by immunohistochemistry. As shown in Fig. 3A , SR-B1 presented a segment-dependent localiza-

Loss of ABCA1 is related to CE accumulation
Since de novo cholesterol synthesis and cholesterol uptake did not seem to be at stake to explain CE accumulation in the caput epididymidis of LXR Ϫ / Ϫ mice, we analyzed the localization and distribution of two membrane cholesterol transporters involved in cholesterol effl ux: ABCA1 and ABCG1. In wild-type mice, ABCA1 was localized in the apical part of the apical cells in segments 1 and 2 at 4 ( Fig. 4A , a-e) and 9 months (Fig. 4A, c-g). In LXR Ϫ / Ϫ mice both at 4 and 9 months of age, a complete loss of the was observed at 9 months of age. The LXR Ϫ / Ϫ mice presented the same pattern in segment 3. The main difference was observed in segments 1 and 2 where both staining and epithelium height decreased signifi cantly. This reduced signal was not correlated with Western blot analysis of SR-B1 ( Fig. 3C ). These results suggested that SR-B1 was not directly responsible for the neutral lipid accumulation observed. In addition, LDL receptor (LDLR) localization was studied, and no difference observed (data not shown). animals versus an arbitrary value of 1 in the wild-type animals. Each value is the mean ± SEM of three measurements performed on three different animals using cyclophillin as internal standard. ** P < 0.01 compared with wild-type mice. Fig. 3. SR-B1 expression and localization do not change with the genotype or the age. A: SR-B1 immunoperoxydase staining in caput epididymidis from wild-type mice. Magnifi cation ×25. B: SR-B1 immunoperoxydase staining in segment1 (s1), segment 2 (s2), and segment 3 (s3) caput epididymidis from wild-type and LXR Ϫ / Ϫ mice at 4 and 9 months of age.
Each photomicrograph is representative of three different experiments made on three different individuals. Inset shows negative control. Original magnifi cation ×630. C: Relative level of SR-B1 protein in caput and cauda epididymidis from wild-type and LXR Ϫ / Ϫ mice at 4 and 9 months of age. Histograms are expressed as mean ± SEM of three different experiments made on three different animals using ␤ -actin as an internal standard for quantifi cation.
found for ABCA1 at 4 months of age in the wild-type and LXR Ϫ / Ϫ mice, whereas a signifi cant 1.6-fold decrease was found in the LXR Ϫ / Ϫ mice compared with wild-type mice ( P < 0.05) at 9 months of age. This apparent discrepancy ( Fig. 4A versus Fig. 4B ) refl ects that ABCA1 protein levels in the caput epididymides of LXR Ϫ / Ϫ animals at 4 months of age are modifi ed in a discrete manner (i.e, solely in apical cells of segments 1 and 2) that is not suffi cient at that ABCA1 staining was observed ( Fig. 4A, b-d and f-h). ABCG1 presented staining both in the nucleus of epithelial cells and on the apical membrane of the epithelium, in segments 1 and 2 irrespective of cell type and genotype ( Fig. 4A, i-p). Accumulation of ABCA1 and ABCG1 was determined by Western blot in protein extracts made with entire caput epididymides ( Fig. 4B ). Contrary to the histochemical data presented in Fig. 4A , no difference was (a-h) and ABCG1 (i-p) immunofl uorescence staining in segment 1 (s1) and segment 2 (s2) caput epididymidis from wild-type and LXR Ϫ / Ϫ mice at 4 and 9 months of age. Bars = 10 µm. Each photomicrograph is representative of three different experiments made on three different individuals. Inset shows negative control. B: Relative proteins levels of ABCA1 and ABCG1 in caput epididymidis at 4 and 9 months from wild-type and LXR Ϫ / Ϫ mice. Histograms are expressed as mean ± SEM of three different experiments using ␤ -actin for relative quantifi cation made on three different animals. C: Relative level of abca1 mRNAs in caput epididymides at 4 and 9 months. Histograms are expressed as a normalized value of the expression level in lxr Ϫ / Ϫ animals versus an arbitrary value of 1 in the wild-type animals. Each value is the mean ± SEM of three measurements performed on three different animals using cyclophillin as internal standard. * P < 0.05 and ** P < 0.01 compared with wild-type mice.
sented in Fig. 6 demonstrate that ABCA1 accumulation was clearly increased in the caput epididymides of wildtype mice fed the Western diet ( Fig. 6 ; P < 0.05). The epididymal epithelium thus responds to the cholesterol enriched diet by enhancing ABCA1 protein via LXR stimulation.

DISCUSSION
This study investigated the role of LXR nuclear receptors in the regulation of cholesterol homeostasis in caput epididymidis, using a comparison between wild-type and LXR Ϫ / Ϫ male mice. Nile Red staining allowed us to show that CEs were mainly accumulated in the apical cells of caput epididymidis segments 1 and 2 leading to increased cell apoptosis. This accumulation of CEs was linked to the loss of ABCA1 expression in segments 1 and 2. ABCA1 has already been described to be important for male fertility ( 20 ). Our data also showed precisely the ABCA1 distribution in the proximal epididymis. In caput segments 1 and 2, ABCA1 was expressed in the apical cells, while in the more distal caput (segment 3 and downwards) ABCA1 staining was found in principal cells. Recently, Morales et al. ( 21 ) also reported the detection of ABCA1 in the principal cells of an undefi ned mouse caput segment, most likely a segment posterior of segment 2. Interestingly, ABCA1 accumulation was only lost in the segments 1 and 2 of LXR Ϫ / Ϫ animals, while posterior of segment 2, ABCA1 distribution was unchanged, suggesting a specifi c regulation of abca1 by LXR along the epididymis. Confi rming this pattern of regulation, segments 1 and 2 were the sole segments showing disturbed epithelia in LXR Ϫ / Ϫ mice ( 12 ). The fact that CE accumulations were mainly due to a defect in ABCA1 representation was confi rmed by quantitative RT-PCR analysis of mRNAs involved either in de novo cholesterol synthesis or cholesterol esterifi cation. Expression levels of srebp-2 , hmg-CoA-red , hmg-coA-synt , acat1 , and acat2 were measured, and a signifi cant difference (decrease) was only observed for srebp-2 and hmg-coAsynt . These results suggest that neither endogenous cholesterol synthesis nor esterifi cation was modifi ed in LXR Ϫ / Ϫ mice. The decrease recorded in srebp-2 and hmg-coA-synt expression could be a consequence of CE accumulations age to modify the overall quantity of ABCA1 protein in the entire caput. For ABCG1, no signifi cant differences in accumulation of the four variants, the 60, 70, 100, and 120 kDa protein bands, was found at 4 and 9 months of age in wild-type and LXR Ϫ / Ϫ mice. However, there was a clear change in the relative representation of three of four variants since we recorded a decrease of the 60 kDa ABCG1 variant and an increase of the 100 and 120 kDa ABCG1 variants. Western blot data were correlated with quantitative RT-PCR analysis ( Fig. 4C ) showing that abca1 mRNA accumulation did not change in the caput epididymides of 4-month-old LXR Ϫ / Ϫ animals, while a 70% decrease was recorded in the caput epididymides of 9-month-old animals when compared with controls ( P < 0.001). These data lead us to suggest that ABCA1 is involved in neutral lipid accumulations.

CE accumulation enhances apoptosis in segment 1 epithelial cells
Since neutral lipid accumulations are known to cause apoptosis in some cell types ( 18 ), we carried out TUNEL staining experiments on caput epididymidis from wildtype and LXR Ϫ / Ϫ mice at 4 months. The number of apoptotic cells was signifi cantly increased (2.3-fold) in the caput epididymidis of LXR Ϫ / Ϫ mice compared with controls ( Fig. 5 ; P < 0.01) but only in segment 1 and was not restricted to a particular cell type. However, apoptosis was greater in apical cells (11.4-fold increase compared with the wild-type) compared with the other epithelial cell type (4.8-fold increase compared with the wild-type), suggesting a correlation between CE accumulations and apoptosis in apical cells.

abca1 is regulated by LXR in the caput epididymidis
To determine whether abca1 was physiologically regulated by LXR in the caput epididymidis, 8-month-old wildtype mice were given a Western diet for 7 weeks. This diet has already been shown to activate LXR ( 19 ). Data pre- accumulation was found between wild-type and LXR Ϫ / Ϫ mice. In addition, our data also showed specifi c SR-B1 localization in the various epididymal segments. The LXR Ϫ / Ϫ mice also present abnormalities in their epididymal fatty acid metabolism. We showed earlier that the expression level of srebp-1c in the caput epididymidis was downregulated in these mice and that little differences were seen in the fatty acid content of both the PL and neutral lipid fractions ( 15 ). More refi ned fatty acid evaluations realized here on caput epididymides of LXR Ϫ / Ϫ animals show an increase in the relative proportions of palmitoleic acid (C16:1 n-7) in the neutral lipid fraction, correlated with a decrease of palmitic acid (C16:0). These changes are in accordance with intracellular CE accumulations, as the principal fatty acid esterifying the cholesterol moieties is palmitoleic acid.
The association of CE accumulations and apoptosis that we revealed by TUNEL assays in the apical cells of segment 1 has already been reported elsewhere ( 18 ). Apoptosis of apical cells could be the starting point of epididymal epithelium destructuration. This point raises the question of the physiological role of this particular cell type in the proximal region of the epididymal duct. Apical cells are as previously demonstrated in hamster ( 22 ). In parallel, apoE expression, an apoprotein acceptor of effl uxed cholesterol, was also investigated and no change at all was recorded (data not shown).
At 4 months of age, only ABCA1 staining in apical cells of segments 1 and 2 in LXR Ϫ / Ϫ mice was lost. Since mRNA extractions were carried out on the whole caput, it is likely that the variations seen in abca1 expression were not detectable. However, only ABCA1 protein level was altered, whereas for ABCG1, we have observed only discrete changes in the representation of its splicing variants [( 23 ); see above]. To our knowledge, neither different cellular distribution nor specifi c roles have yet been ascribed to these distinct abcg1 variants.
To determine cholesterol fl uxes in caput epididymidis more precisely, we investigated the cellular location of lipoprotein receptors SR-B1 and LDL-R. As no evidence of LDL-R was shown, it appeared that the main receptor represented in the caput epithelium was SR-B1, which mediates selective uptake of CEs from HDL to cells ( 24 ), thus being a possible candidate in the CE accumulations observed in LXR Ϫ / Ϫ male mice. However, our results overruled this hypothesis, since no difference in SR-B1 Fig. 6. ABCA1, a target gene of LXR, is overexpressed in caput epididymidis under a cholesterol-enriched diet. Relative proteins levels of ABCA1 in caput epididymidis in 8-month-old wild-type mice fed during 7 weeks with a Western diet (1.25% cholesterol). Histograms are expressed as mean ± SEM of three different experiments using ␤ -actin for relative quantifi cation made on three different animals. * P < 0.05 compared with wild-type mice known to be rich in mitochondria and to show endo cytic activity ( 25 ). It was suggested that they are involved in fl uid acidifi cation because of their high carbonic anhydrase content ( 26 ) as well as in sodium transport and chloride fl uxes ( 27 ). Our data suggest that these cells are also critical for cholesterol traffi cking in the proximal epididymis. Based on our data, we propose a model of cholesterol fl uxes in this organ. In segment 1, lipoproteins could reach the epithelium through fenestrated blood capillaries and be processed by SR-B1. Cholesterol could then be transferred to the lumen via ABCA1 in the apical cells and ABCG1 in the principal cells. This would imply cooperation between these two cell types, as ABCA1 transfers cholesterol and PLs to lipid-poor apoproteins, such as apoA1, apoE, and apoJ, whereas ABCG1 effl uxes cholesterol to mature HDL but not to lipid-poor apoproteins [as reviewed in ( 28 )]. Such cooperation in these two ABC transporters has already been shown in vitro ( 28 ). This hypothesis is supported by the fact that apoE has already been shown to be expressed in the caput epididymidis [this study and ( 29 )]. SR-B1 apical staining in the caput segment 1 suggests that exchanges might occur between the epithelium and testicular fl uid entering the organ. The stronger staining for SR-B1 obtained on the microvilli of segment 2 caput epithelial cells suggests an intense cholesterol reabsorption process in that particular segment. Downstream, the subsequent intracytoplasmic cellular localization of SR-B1 (segment 3) is here in favor of intracellular traffi cking of cholesterol and interaction of late endosomes with lysosomes ( 30 ). Intercellular exchanges in the epithelium might also occur since SR-B1 was also detected on the lateral membranes (from caput segment 4 to cauda; data not shown). Such intercellular exchanges have been shown elsewhere ( 31 ).
In conclusion, this study showed for the fi rst time that regulation of cholesterol homeostasis in mice caput epididymidis is a cell and segment-specifi c process regulated in part by LXR. Our data offer new perspectives for study of the molecular mechanisms related to cholesterol maturation of sperm cells and emphasize the already important role devoted to the proximal caput epididymidis in the posttesticular sperm cell maturation process. It has demonstrated a new role played by the so-called apical cells of the proximal caput epithelium in cholesterol homeostasis and epididymis physiology.