Brain region-specific immunolocalization of the lipolysis-stimulated lipoprotein receptor (LSR) and altered cholesterol distribution in aged LSR+/− mice

Authors


Abstract

Brain lipid homeostasis is important for maintenance of brain cell function and synaptic communications, and is intimately linked to age-related cognitive decline. Because of the blood–brain barrier's limiting nature, this tissue relies on a complex system for the synthesis and receptor-mediated uptake of lipids between the different networks of neurons and glial cells. Using immunofluorescence, we describe the region-specific expression of the lipolysis-stimulated lipoprotein receptor (LSR), in the mouse hippocampus, cerebellum Purkinje cells, the ependymal cell interface between brain parenchyma and cerebrospinal fluid, and the choroid plexus. Colocalization with cell-specific markers revealed that LSR was expressed in neurons, but not astrocytes. Latency in arms of the Y-maze exhibited by young heterozygote LSR+/− mice was significantly different as compared to control LSR+/+, and increased in older LSR+/− mice. Filipin and Nile red staining revealed membrane cholesterol content accumulation accompanied by significantly altered distribution of LSR in the membrane, and decreased intracellular lipid droplets in the cerebellum and hippocampus of old LSR+/− mice, as compared to control littermates as well as young LSR+/− animals. These data therefore suggest a potential role of LSR in brain cholesterol distribution, which is particularly important in preserving neuronal integrity and thereby cognitive functions during aging.

Abbreviations used
BBB

blood–brain barrier

FFA

free fatty acid

GFAP

glial fibrillary acidic protein

LDL-R

low-density lipoprotein receptor

LpL

lipoprotein lipase

LRP1

low-density lipoprotein receptor-related protein 1

LSR

lipolysis-stimulated lipoprotein receptor

NeuN

neuronal nuclei

PBS

phosphate-buffered saline

PFA

paraformaldehyde

Age-related cognitive decline associated with higher risk of neurodegenerative disorders has become an important public concern as life expectancy increases in industrially developed countries. Although the molecular mechanisms underlying the decrease in memory and other brain-related functions with age are under active investigation, the processes involved are not yet completely understood. Attention has recently focused on lipid status in the CNS, partly because of reports of potential beneficial effects of omega-3 fatty acids (Cederholm and Palmblad 2010; Huang 2010; Calon 2011) and cholesterol-lowering agents such as statins towards lowering risk of neurodegenerative diseases such as Alzheimer's disease, although the results are not clear-cut (Di Paolo and Kim 2011; Shepardson et al. 2011a, b). This latter epidemiologically based issue is of particular interest to public health in view of the numerous patients treated with statins for peripheral hypercholesterolemia. However, no direct link has yet been identified between plasma and cerebral cholesterol metabolism.

Cholesterol is a major component of the brain and is essential for cognitive function because of its role in the formation of functional synapses allowing communication between neurons [for review, (Schreurs 2010)]. Indeed, neuronal requirement for cholesterol is elevated because of the high density of axonal membranes and myelin sheath, of which this sterol is a major component [for review, (Pfrieger and Ungerer 2011)]. Cholesterol is delivered to tissues via lipoprotein particles. However, the brain's access to cholesterol and other lipids in the peripheral circulation is complicated by the presence of the blood–brain barrier (BBB), which serves as a selective low-permeable multicellular barrier. The brain must therefore rely on its own network capable of synthesizing, internalizing and metabolizing these lipids to provide the necessary components for neuronal cell membrane function. The glial cells play a central role towards providing neurons with lipids, and in particular cholesterol, in the form of lipoproteins. Lipoproteins in the CSF are very different from those of the periphery, and have been characterized as high-density lipoprotein-like particles containing primarily apolipoprotein (apo)E and apoJ (Boyles et al. 1985; Pitas et al. 1987; Kim et al. 2009a).

Studies have revealed the presence of the low-density lipoprotein receptor (LDL-R) in the CNS that is mainly expressed in astrocytes and endothelial cells throughout brain parenchyma (Rapp et al. 2006). ApoE levels in the CSF accumulate in LDL-R−/− mice (Fryer et al. 2005), and are decreased when LDL-R is over-expressed (Kim et al. 2009b), suggesting a role of LDL-R in CNS apoE metabolism. The absence of LDL-R also is related with decreased pre-synaptic bouton density in the hippocampus (Mulder et al. 1993), as well as increased locomotor activity (Elder et al. 2008). This receptor is found throughout the brain, and although highly enriched in the brainstem, no significant changes in cholesterol content was observed in this region of LDL-R−/− mice (Taha et al. 2009).

Other lipoprotein receptors in the LDL-R family also appear to play an important role in lipid homeostasis and cholesterol metabolism of neurons (Ladu et al. 2000; Qiu et al. 2006; Vance and Hayashi 2010). Among them, the LDL-R-related protein (LRP1) expression in the CNS has been the subject of many studies. Its expression is more selective and has been detected specifically in neuronal cell bodies and proximal processes including the hippocampus and the cerebellum (Pitas et al. 1987; Herz and Chen 2006), as well as in the pericytes (Tooyama et al. 1995). LRP1 is also found in astrocytic foot processes and along capillary membranes in a discontinuous manner, reflecting a role in the selective permeability at the level of the BBB (Herz 2003). Indeed, LRP1 has been implicated in the transport and metabolism of the β-amyloid peptide (Fujiyoshi et al. 2011). Deletion of the LRP gene results in embryonic lethality associated with severe malformations of the CNS (Herz et al. 1992) which clearly shows a critical role of this receptor in CNS development. Nevertheless, glial cell-derived apoE-containing lipoprotein-mediated activation of axonal growth to promote neuron communication is not solely LRP dependent (Matsuo et al. 2011). Thus, other alternative or complementary pathways/transporters may exist that contribute towards the maintenance of adequate lipid status and intercellular lipid exchange in the brain.

We have recently demonstrated that the lipolysis-stimulated lipoprotein receptor (LSR), is involved in the hepatic clearance of triglyceride-rich apoB,E-containing lipoproteins during the post-prandial phase (Yen et al. 2008). In vivo studies have shown that this receptor plays an important role in maintaining normal peripheral circulation levels of cholesterol and triglycerides, and in contributing to the regulation of lipid distribution amongst the peripheral tissues (Yen et al. 2008; Narvekar et al. 2009). Absence of both lsr alleles (LSR−/−) is associated with in utero lethality at the embryonic stage, in which both brain-localized hemorrhages and premature dorsal skin detachment are observed (Mesli et al. 2004). During pre-natal development, dorsal skin is initially, along with the brain, part of the same primitive structure that subsequently separates during neural tube formation, therefore suggesting that LSR could be critical for mouse brain development. Our objective therefore was to characterize LSR expression in the brain. Here, we describe for the first time the localization of the lipoprotein receptor LSR in the mouse brain, specifically in neuronal cells and in regions located at the CSF and brain parenchyma interface. In aged LSR+/− mice, we detected an altered cellular brain cholesterol distribution compared with control littermates associated with decreased number of entries in the Y-maze paradigm, suggesting a role of this receptor in the regulation of cholesterol homeostasis in the CNS directly implicated in cognitive function.

Materials and methods

Materials and antibodies

All chemicals and solvents were obtained from Sigma-Aldrich (Saint-Quentin Fallavier, France) unless otherwise indicated. Rabbit anti-LSR antibody was purchased from Sigma-Aldrich and goat anti-actin, mouse anti-lipoprotein lipase (LpL), mouse anti-β-tubulin, and goat anti-calbindin from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Mouse anti-neuronal nuclei (NeuN) and anti-glial fibrillary acidic protein (GFAP) antibodies were purchased from Millipore (Molsheim, France). Alexa488- and Alexa594-conjugated secondary specific anti-mouse and anti-rabbit antibodies were acquired from Molecular Probes (Invitrogen, Cergy Pontoise, France). Secondary anti-mouse and anti-rabbit horseradish peroxidase-conjugated antibodies were purchased from Cell Signaling Technology (Ozyme, Saint-Quentin-en-Yvelines, France).

Animals

Adult male C57Bl/6J mice (Janvier Breeding, Le Genest Saint Isle, France) were housed in certified animal facilities on a 12-h light/dark cycle with a mean temperature of 21–22°C and relative humidity of 50 ± 20%, and provided rodent chow diet and water ad libitum. LSR+/− mice were produced from a colony in our animal facilities as previously described (Yen et al. 2008). LSR+/+ littermates were used as controls. Animals were handled in accordance with French State Council guidelines for the use and care of laboratory animals.

Immunofluorescence studies

Anesthetized mice were killed by decapitation. Mouse brains were immediately dissected, placed in pre-cooled (−20°C) methylbutane for 30 min at −20°C, removed and then stored at −80°C. Brain sections (14 μm) were prepared using a cryostat HM550 (MicromMicrotech, Francheville, France) and then mounted on gelatin-coated slides.

Brain sections were fixed by treating with 4% (w/v) paraformaldehyde (PFA) in phosphate-buffered saline (PBS) for 15 min at 21°C, and then permeabilized using 0.1% (v/v) Tween-20 in PBS (30 min, 21°C). After fixing, all steps were performed at 21°C. After washing in PBS, the sections were incubated 1 h with 10% (w/v) bovine serum albumin solution, followed by a 2-h incubation with the primary antibody, and washed three times with PBS. For detection, preparations were incubated with the appropriate Alexafluor 488- or Alexafluor 594-conjugated secondary antibody for 2 h at 21°C in the dark. Slices were washed and mounted on slides using Fluormount.

Cholesterol staining

Membrane-bound non-esterified cholesterol was stained using the filipin dye. Briefly, after fixing brain preparations with 4% PFA, the coverslips were incubated in 125 μg/mL filipin in 10% bovine serum albumin for 2 h in the dark, washed in PBS, and then mounted onto slides. For staining of intracellular lipids droplets, PFA-fixed brain sections were incubated with Nile Red dye (10 μg/mL) for 30 min, and then washed in PBS, and mounted onto slides using Prolong Gold antifade reagent, which contains DAPI (Invitrogen). Preparations were visualized and digitized images were taken using identical exposure times of preparations from at least three different mice using either a fluorescence microscope (Nikon, Champigny-sur-Marne, France) or an FV10i-W confocal microscope (Olympus, Rungis, France). Quantitation of the lipid droplets observed in 18-month-old mice was performed using ImageJ (National Institute of Health, Bethesda, MD, USA, http://imagej.nih.gov/ij/, using the option ‘analyze particle’.

Lipid analysis of brain structures

The cerebellum, cortex, hippocampus, and hypothalamus were isolated, rinsed in physiological saline, snap frozen in liquid N2, and stored at −80°C. Lipid extractions were performed on pre-weighed lyophilized tissue samples followed by analysis of triglycerides, total cholesterol, and phospholipids using enzymatic kits (Biomérieux, Marcy-l'Etoile, France), as previously described (Yen et al. 2008).

Immunoblotting

Brain structures homogenates were prepared in radioimmunprecipitation lysis buffer as previously described (Yen et al. 2008). Identical amounts of protein (20–30 μg) were separated by electrophoresis on 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis, and then transferred to nitrocellulose membranes. Immunoblots were performed as previously described (Yen et al. 2008).

Behavioral analysis in Y-maze test

The Y-maze test was performed as previously described (Garcia et al. 2010). The maze was made of opaque Plexiglas, and each arm was 40-cm long, 16-cm high, and 9-cm wide and positioned at equal angles. Mice were placed at the end of one arm before starting the experiment countdown. The series of arm entries were recorded visually, and arm entry was validated when the hind paws of the mouse were completely placed in the arm. Spontaneous alternation behavior was monitored during a 5-min interval. Alternation scores were calculated by analyzing overlapping triplet's sets. The percentage alternation was calculated as the ratio of actual (total alternations) to possible alternations.

Statistical analysis

All results are presented as mean and SEM. Statistical significance was assessed by the Student's t-test, with significance considered at p values < 0.05.

Results and discussion

LSR is expressed in region-specific neuronal cells, but not in astrocytes

We initially examined the general staining pattern of LSR by immunofluorescence in brain coronal sections of wild-type C57BL6/Rj mice, which appeared to be most marked in the hippocampus, hypothalamus, cerebellum, and areas surrounding the ventricles (data not shown).

A more detailed immunohistochemical study on these different structures was next performed to determine the cell-type localization of LSR using different cell-specific markers (Fig. 1). In the hippocampus (Fig. 1a), LSR was detected in NeuN-positively labeled neurons in the dentate gyrus as well as in the CA1 region, with a staining pattern suggestive of receptor distribution on the cell body membrane and neurites of hippocampal pyramidal neurons. LSR expression was also observed in other pyramidal-shaped neurons in the cerebellum, most particularly in cell bodies and neurites of Purkinje cells (Fig. 1b). Interestingly, fluorescence was not detected in the granular, or in the molecular layer of the cerebellum, but rather only in Purkinje cell bodies and neurites labeled by calbindin, whose extensions penetrated the molecular layer. The yellow regions in the merged picture (Fig. 1b, bottom panel), showed that LSR expression in the cerebellum was limited to these calbindin-expressing cells.

Figure 1.

Immunohistochemical study and immunoblots of LSR in different brain regions. Images of the hippocampus (a), cerebellum (b), third ventricle (c), and choroid plexus (d) are shown immunostained with LSR (top panels, arrows) or the indicated cell-specific markers (middle panels) individually and as a merged image of both markers (bottom panels). LSR was alternatively detected with a red or green fluorochrome-conjugated antibody. Images were taken at 20X magnification and are representative of at least three different experiments in three different animals. DAPI labeling is shown as inserts. (Abbreviations: DG, dentate gyrus; GL, granular layer; ML, molecular layer; VMN, ventro–medial nucleus; CP, choroid plexus; D3V, dorsal third ventricle). e. Immunoblots of LSR and LpL in the indicated brain structures (n = 3 animals) were performed as described in 'Materials and methods'. Either actin or β-tubulin was used as protein loading control. Blots are representative of a total of five animals.

In the third ventricle region, LSR was found to be localized in NeuN-positives neurons, often described as forming the ventro–medial nucleus of the hypothalamus, appearing to be localized at the membrane (Fig. 1c, left panels, close-up). Double staining with an antibody for astrocyte-specific GFAP revealed that all GFAP-positive cells remained negative for LSR expression in the third ventricle cistern (Fig. 1c, right panels, close-up). This was the case in all the examined regions of the mouse brain (data not shown).

These protein-expression studies indicate therefore that LSR was located principally in NeuN-positive neuronal cells, but not in astrocytes. A primary function of glial cells is to provide neurons with lipids, specifically cholesterol, in the form of lipoproteins (Pfrieger and Ungerer 2011). CNS lipoproteins are very different from those found in the periphery, consisting of high-density lipoprotein-like particles, but nevertheless containing apoE (Boyles et al. 1985; Pitas et al. 1987; Kim et al. 2009a). Astrocytes also secrete discoïdal apoE-containing particles that acquire lipids from neurons during their maturation (Pitas et al. 1987; Ladu et al. 2000). The properties of astrocyte-derived apoE-lipoprotein particles are different from those found in CSF, and it has been suggested that they may undergo modifications as they traffic through the brain parenchyma, including neurons, to the CSF (Demeester et al. 2000). As LSR does recognize apoE-containing lipid particles (Bihain and Yen 1992; Yen et al. 1994), it is possible that neurons may acquire lipoproteins and therefore cholesterol via the LSR pathway.

Staining of LSR was also detected in ependymocytes, which was most visible in the area surrounding the third ventricle cistern (Fig. 1c). Close examination revealed that LSR staining was localized on the apical side of ependymocytes, which is directly in contact with the CSF. This polarized expression further supports our hypothesis that LSR plays a direct role in CSF-circulating lipoprotein metabolism. Lipoprotein particles in the brain migrate from the interstitial space surrounding neuronal cells to the CSF, allowing lipid transport in the whole CNS. The presence of LSR at the apical side of ependymocytes, could therefore be consistent with its function of lipoprotein receptor as described for the liver (Yen et al. 2008; Narvekar et al. 2009). The absence of LDL-R did not significantly alter brain polyunsaturated fatty acid composition (Chen et al. 2008), suggesting that the brain may acquire these lipids through other pathways independent of the LDL-R. LDL-R−/− mice crossed with LSR+/− mice led to a significant decrease in docosahexaneoic acid/arachidonic acid ratio in the hippocampus (unpublished observations). As docosahexaneoic acid in the brain is acquired primarily from the periphery, this suggests that LSR may be involved in lipid transport in the brain. Further investigations are underway to identify the specific role of LSR in lipoprotein transport at the apical membrane of ependymocytes.

LSR is localized in the choroid plexus of the mouse brain

The choroid plexus is an epithelium-like structure located on the inner side of the lateral ventricles and is responsible for continuous CSF secretion. It is also present in the third ventricle and the roof of the fourth ventricle, and appears as a detached structure located in the dorsal part of the third ventricle lumen. The leptin receptor, originally cloned from this region (Tartaglia et al. 1995) was used as marker for this structure and found to colocalize with LSR expression (Fig. 1d), thus indicating that LSR was also expressed in choroidal cells. Interestingly, we had previously reported that leptin regulates LSR protein levels through the pErk canonical signaling pathway induced upon interaction of leptin with its receptor (Stenger et al. 2010). It could therefore be speculated that a similar regulatory mechanism is present in the CNS.

A second, more permeable barrier is present inside the brain at the level of the ventricle and choroid plexus, called the blood–CSF barrier. CSF is secreted from the choroid plexus surrounding the inner side of cerebral ventricle, and its components are essential for neurons to function where its chemical composition reflects brain function and status. The choroid plexus has also been shown to be an important site for the clearance of β-amyloid peptide, and expresses both megalin and LRP1, two receptors thought to be involved in the clearance of leptin and beta amyloid, respectively (Dietrich et al. 2008; Fujiyoshi et al. 2011). Ependymocytes represent the main cellular component of the choroid plexus and are responsible for CSF secretion. As the choroid plexus is a region rich in capillaries and membrane-associated shuttle systems for communication between brain parenchyma and CSF, this would represent another potential region in which LSR may play a role in the transport of lipids. Interestingly, a recent report by Daneman's laboratory identified the presence of lsr gene transcript in the endothelial cells of the brain which form the BBB (Daneman et al. 2010). Furthermore, a recent report demonstrated that LSR is critical for the tricellular junctions of epithelial cells (Masuda et al. 2011), and may therefore play a similar role in the integrity of the cell barriers in the brain. Taken together, these data suggest a potential role of LSR in the transport of lipids or lipoproteins at the interface between the periphery and the brain (BBB), as well as between the brain parenchyma and CSF.

LSR expression in total protein extracts from different brain structures

The antibody used for the immunofluorescence studies was previously shown to be LSR specific (Yen et al. 2008; Stenger et al. 2010). Nevertheless, immunoblots were performed on homogenates prepared from the hippocampus, cerebellum, and hypothalamus, and revealed a band with an apparent molecular mass between 58 and 61 kDa, corresponding to that observed for hepatic LSR (Fig. 1e) (Yen et al. 1999). Although immunohistochemical studies did not detect a strong signal of LSR in the frontal cortex, immunoblotting revealed the presence of LSR in this brain region as well (data not shown).

Hepatic LSR activity as a lipoprotein receptor requires the presence of free fatty acids (FFAs), which causes a conformational change of the receptor exposing a binding site for apoB or apoE (Bihain and Yen 1992; Yen et al. 1994). As FFAs are lipolytic products, immunoblots were also performed to determine if lipoprotein lipase (LpL) protein could be detected on the same structures in which LSR was expressed. LpL protein was present to different degrees in all structures tested (Fig. 1e). This lipase appeared quite strongly expressed in the hippocampus (Fig. 1e), as had been previously reported (Goldberg et al. 1989). LpL was also detected in the frontal cortex (data not shown). The presence of LpL in these regions could therefore potentially provide the source of FFA that would be needed for LSR to function as a lipoprotein receptor. Interestingly, a study did report significant learning and memory deficits in mice lacking LpL, suggesting that LpL function is somehow connected to lipid status in the brain through as yet undefined mechanisms (Xian et al. 2009). We would propose that the presence of LpL in the different brain structures and most particularly the pyramidal neurons of the hippocampus (Goldberg et al. 1989), provides the means to produce FFA necessary for LSR to function as a lipoprotein receptor. If LpL is deficient, the resulting lack of FFA would lead to suboptimal activity of LSR, thus preventing the maintenance of appropriate lipid status in the neurons in this region. We next sought for evidence of cognition dysfunction in mice with reduced LSR expression.

Aged LSR+/− mice demonstrate decreased exploratory activity in Y-maze behavior test

The absence of both alleles of LSR is embryonic lethal (Mesli et al. 2004). However, heterozygote LSR+/− mice are viable, and exhibit decreased lipoprotein clearance during the post-prandial phase (Yen et al. 2008). Initial observations of young LSR+/− mice did not reveal any grossly evident phenotype relating to potential neurological problems. We did, however, observe a significant impact of reduced LSR expression in aging LSR+/− mice on peripheral lipid homeostasis in the form of higher body mass and potential peripheral leptin resistance that was most marked in female mice (Stenger et al. 2010). In view of LSR's region-specific distribution in neurons of brain structures related to learning and memory, we questioned if potential changes in the CNS could also be observed as a result of reduced LSR expression with age. To avoid potential effects of hormonal variations for the behavioral assessment, we used young (10 weeks) and aged (18 months) male LSR+/− mice as compared to control LSR+/− littermates on standard diets. Neither group displayed a significant difference in body weight (data not shown). Our previous studies have shown no difference in food intake in LSR+/− mice on a standard diet (Yen et al. 2008; Stenger et al. 2010). The Y-maze behaviorial test is widely used to detect potential cognitive impairments, including spatial memory and short-term memory in rodents, and primarily reflects hippocampus function. Furthermore, while the cerebellum is associated with locomotor activity, it can also be involved in cognition, specifically via the Purkinje cells (Burguiere et al. 2010; Rochefort et al. 2011). We found that while spontaneous alternance did not differ between the LSR+/+ and LSR+/− of both young and old mice (Fig. 2a), the number of entries into Y-maze arms was significantly lower in 10-week- and 18-month-old LSR+/− mice as compared to age-matched controls (Fig. 2b). This may have been primarily because of a significantly longer latency in the initial arm entry (Fig. 2c). Closer examination revealed that while the total time in the arm 2–8 entries was not different in the two groups of young mice, it was significantly different in the older LSR+/− mice as compared to age-matched controls (Fig. 2d). As this was the first time that the mice had been exposed to the Y-maze, these results suggest that reduced LSR may be associated with cognitive disturbances related to reactivity to novel surroundings. Interestingly, a previous study suggests that Purkinje cells of the cerebellum may play an important role in determining spatial factors for the hippocampus (Rochefort et al. 2011), both regions of which express LSR.

Figure 2.

Y-maze test in young and old LSR+/− mice as compared with LSR+/+ control littermates. Young and old LSR+/− mice (10 weeks, n = 5; 18 months, n = 18) and control littermates (LSR+/+, 10 weeks, n = 4, 18 months, n = 14) were subjected to a Y-maze test as described in the 'Materials and methods'. Results are shown as mean and SEM of the (a)% alternance, (b) number of entries, (c) time in the first arm, and (d) accumulated time in the subsequent 2–8 arms with the p-value indicated when statistically significant.

Cholesterol distribution profile in young and old LSR+/− mice

We then examined brain cholesterol distribution in LSR+/− mice as compared to LSR+/+. As cell-specific distribution of LSR was observed in the hippocampus as well as in Purkinje cells of the cerebellum, brain sections with these regions from the same young and old LSR+/− and LSR+/+ mice were labeled with Nile Red, which stains intracellular lipid droplets. There were no significant differences in Nile Red staining of the hippocampus or cerebellum in the two groups of young mice (Fig. 3a, b, upper panels). Indeed, the staining appeared rather homogeneous and diffuse in the different brain areas. No lipid droplets were evidenced even at the yellow-gold fluorescence level, which has been used for the detection of low levels of lipid droplets (Greenspan et al. 1985). However, in the 18-month-old mice, distinct areas of lipid droplets were observed (Fig. 3a, b, lower panels), which we were able to quantitate using ImageJ software. While a tendency towards a decrease of these droplets in the dentate gyrus of the hippocampus in 18-month-old LSR+/− was observed (Fig. 3a and c), there were significantly lower levels of intracellular lipid in the Purkinje cells of the cerebellum (Fig. 3b and c) in these mice as compared to LSR+/+ controls. In view of brain cholesterol content, we would speculate that this most likely represents cholesteryl esters rather than triglyceride accumulation in these regions.

Figure 3.

Nile Red staining of brain regions of young and old LSR+/− and LSR+/+ control littermates. Brain cryosections of 10-week- (upper two rows) and 18-month-old (lower two rows) LSR+/− and LSR+/+ mice were stained with Nile Red, and visualized by fluorescence microscopy as described in the Materials and Methods. Representative images are shown for n = 3 for the hippocampus (a), and the cerebellum (b). (c) Image analysis of signal intensity of the lipid droplets observed in the 18-month-old mice was performed using Image J, and shown as mean ± SEM (n = 3) for LSR+/+ (□) or LSR+/− (■). Statistical significance comparing LSR+/+ and LSR+/− are indicated.

Brain sections from LSR+/+ and LSR+/− mice were also labeled with filipin, which stains membrane-associated free cholesterol. In 10-week-old mice (Fig. 4a, b, c, upper panels), a diffuse and homogenous labeling of cholesterol was observed in the different regions of the hippocampus (CAI, 4a), the third ventricle (4b), and the lateral ventricle (4c), with some increased intensity in the ependymal cell layer. However, the filipin staining of 18-month-old LSR+/− mice versus controls was much different (Fig. 4a, b, c, lower panels). Whereas membrane-associated cholesterol appeared diffuse and homogenous in the 18-month-old controls, in LSR+/− mice, an accumulation of filipin-labeled cholesterol was observed, both at the neuronal level in the CA1 of the hippocampus (Fig. 4a), and at the level of the ependymal cell layer surrounding the third ventricle (Fig. 4b). A similar phenomenon was observed in the lateral ventricle (Fig. 4c, left panels). Filipin staining of cholesterol was markedly increased in the cell layer surrounding the lateral ventricle, and the cell-layer morphology itself appeared to be disorganized. Interestingly, LSR staining in this area was strikingly altered, from an apical uniform bright fluorescence in the control mice, to a faint punctiform, dispersed labeling in the aged LSR+/− mice (Fig. 4c, lower panels). These results clearly show that cholesterol distribution was significantly altered in aged LSR+/− mice. The lipid composition of the different structures was measured in the old mice (Fig. 4d). In the cerebellum, hippocampus, and hypothalamus, which are the three regions where we found LSR to be detected, total cholesterol content showed a tendency to increase, but this did not reach statistical significance. We cannot rule out, however, biochemical changes in cholesterol esters or other cholesterol metabolites. Interestingly, a similar trend was observed for phospholipid content, but this only achieved statistical significance in the hippocampus of the 18-month-old LSR+/− mice. Cholesterol:phospholipids ratios themselves were not significantly modified (data not shown), suggesting a tight relationship between cholesterol and phospholipid levels. Triglyceride levels were not significantly changed. Gross morphology of the brains of the LSR+/− versus controls did not seem to be different, nor were the weights of the brain structures significantly different; calculation of total amount of lipid per structure did not reveal any difference (data not shown). Taken together, this suggests that LSR+/− mice display modified cholesterol distribution with age, in which higher concentrations of cholesterol are located in the membranes. As cholesterol levels are not significantly different, the decreased intracellular lipid may be because of increased cholesterol movement into the membranes. It is also possible that the lipid changes observed in Figs 3 and 4 are localized in specific regions, and are not easily detectable by biochemical analysis. The differences observed in cholesterol distribution were accompanied by changes in membrane morphology as shown by the filipin staining. The disorganized staining observed suggests that decreased LSR expression in these mice could lead to structural disturbances in the cellular layer in some regions, such as the ependymocyte layer. In view of this receptor's implication in intercellular junctions (Masuda et al. 2011), this disruption of cellular organization could lead to an eventual loss in cell polarization.

Figure 4.

Filipin staining of brain regions of young and old LSR+/− and LSR+/+ control littermates. Brain cryosections from 10-week- (upper two rows) and 18-month-old (lower two rows) LSR+/− and LSR+/+ mice were stained with filipin, and visualized by fluorescence microscopy as described in the 'Materials and methods'. Representative images are shown for n = 3 for the hippocampus (a), third ventricle (b), and the lateral ventricle (c, left panels). Imunofluorescence staining (green fluorochrome) of LSR of the lateral ventricle is also shown (c, right panels). (d) Lipid composition of brain structures in 18-month-old LSR+/− and LSR+/+ control littermates. Concentrations of total cholesterol, triglycerides, and phospholipids were measured in dissected brain structures as described in the 'Materials and methods', and are shown as mean ± SEM (n = 4) (*p = 0.022, vs. LSR+/+).

Whether these modifications in cholesterol distribution are related to the observed increased latency in older LSR+/− mice still remains to be determined. Previous studies have reported data supporting the role of cholesterol in influencing neuronal integrity and synaptic plasticity (Krisanova et al. 2012), as well as cognitive behavior (Camargo et al. 2012; Uranga and Keller 2010). It has also been reported that cholesterol influences APP processing and production of the beta-amyloid peptide (Gamba et al. 2012) because of its distribution in the membrane rather than the actual level in neurons (Burns et al. 2006).

In conclusion, we have shown data revealing the region- and cell-specific distribution of LSR in the brain. As for two other peripheral lipoprotein receptors, LDL-R and LRP1, that have been shown to play important roles in brain function; the results reported here point towards a potential role of LSR in the maintenance of cholesterol homeostasis in mice during aging. Interestingly, neuron-specific knockout LRP1 mice display modifications in brain lipid levels that lead to age-related progressive cognitive dysfunction (Liu et al. 2010). Investigation is actively underway to determine the mechanisms involving LSR in the regulation of cholesterol and lipid homeostasis in the brain, as well as the consequences in terms of cognition. This in turn may lead toward the development of novel therapeutic and preventive strategies targeting LSR with the aim toward maintaining appropriate CNS lipid status and thus normal brain cognitive functions during aging.

Acknowledgements

This work was funded in part by the European League against Alzheimer's disease (LECMA), and by local grants provided by la Région Lorraine, and the University of Lorraine (BQR). CS was recipient of a post-doctoral fellowship from LECMA. AP is on a thesis scholarship provided by the French Ministry of Higher Education and Research. The authors state that there is no conflict of interest to disclose.

Ancillary