| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Lipid Research, Vol. 48, 2151-2161, October 2007
Copyright © 2007 by American Society for Biochemistry and Molecular Biology
,
,,,,,,,,**,,**,
* Institut National de la Santé et de la Recherche Médicale, U872, Equipe 6, Paris, F-75006 France
Université Pierre et Marie Curie-Paris 6, UMR S 872, Equipe 6, Paris, F-75006, France
Centre de Recherche des Cordeliers, Université Paris Descartes, UMR S 872, Equipe 6, Paris, F-75006, France
** Ecole Pratique des Hautes Etudes, Laboratoire de Pharmacologie Cellulaire et Moléculaire, Paris, F-75006, France
Institut National de la Santé et de la Recherche Médicale, U538, CHU Saint Antoine, Paris, F-75012, France
Published, JLR Papers in Press, July 24, 2007.
1 To whom correspondence should be addressed. e-mail: athina.kalopissis{at}crc.jussieu.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Supplementary key words high density lipoprotein • cubilin • megalin • proximal tubule • yolk sac cells • mouse kidney cortical cells • transgenic mice
Abbreviations: AMN, amnionless; apoA-II, apolipoprotein A-II; BN, brown Norwegian; DAB, diaminobenzidine; DiI, 1,1'-dioctadecyl-3,3,3',3'-tetramethylindocarbocyanine; hapoA-II, human apolipoprotein A-II; IF, intrinsic factor; IOD, integrated optical density; mapoA-II, murine apolipoprotein A-II; MKC cells, mouse kidney cortical cells; PFA, paraformaldehyde; SR-BI, scavenger receptor class B type I
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Plasma apoA-II content may be regulated by its rates of synthesis and/or catabolism. A kinetic study using stable isotopes has suggested that plasma apoA-II levels are regulated by its rate of synthesis and not by its rate of catabolism (11). The liver is the principal site of apoA-II synthesis (12), but neither the organs catabolizing apoA-II nor the receptor(s) involved have been determined, to our knowledge. We first focused on the kidney as a potential organ of apoA-II catabolism, because it is the major site of apoA-I catabolism in rats, rabbits, and humans (13–17). ApoA-I binds with high affinity to cubilin, a multi-ligand protein associated with the apical membrane of kidney proximal tubules, intestine, and yolk sac (18–20). However, cubilin lacks a classical transmembrane and cytoplasmic domain necessary for endocytosis, and an accessory protein is necessary to internalize cubilin and its ligands. Conclusive evidence has established that megalin, a member of the LDL receptor family expressed in the same tissues, functions in tandem with cubilin as a coreceptor, cubilin binding to megalin with high affinity (21). Indeed, suppression of megalin activity resulted in inhibition of cubilin-mediated HDL endocytosis (22). Cubilin and megalin share some ligands (albumin, receptor-associated protein, vitamin D binding protein), whereas other ligands are specific either for cubilin [intrinsic factor (IF)-vitamin B12, apoA-I, and HDL, transferrin] or for megalin (apoB, apoE, apoJ, lipoprotein lipase, transcobalamine-vitamine B12) (23).
The objective of the present study was to investigate whether apoA-II is catabolized in the kidney. Because iodination has been shown to artifactually alter in vivo apoA-I catabolism by promoting accelerated catabolism and increasing extra-renal degradation pathways (15), we opted for detecting native apoA-II by immunohistochemistry of kidney sections. This technique offers the additional advantage of visualizing internalized apoA-II in specific cell types. To establish a correlation between the plasma concentration of apoA-II and the extent of its catabolism in the kidney, we used three previously established transgenic mouse lines with hapoA-II plasma concentrations ranging from normal (line ß) to two and three times higher (lines 
and 
, respectively) (8). To gain insight into the receptors involved in apoA-II uptake, we studied in vitro apoA-II internalization in rat yolk sac epithelial cells expressing cubilin and megalin and in mouse kidney cortical cells (MKC cells) expressing only megalin.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, and , were backcrossed for at least eight generations to strain C57BL/6J (IFFA-CREDO-Charles River). Transgenic mice were detected as described (7, 8) and were hemizygous for the hapoA-II transgene. Male and female control and transgenic mice over 8 weeks of age were used in equal proportions in all studies. The animals were housed in animal rooms with alternating 12 h periods of light (7 AM–7 PM) and dark (7 PM–7AM), and were fed a chow diet (SAFE) with free access to food and water, unless otherwise specified. Blood was drawn from the retroorbital venous plexus or the abdominal vein between 9 AM and 12 AM. For fasting mice, food was withdrawn at 6 PM on the previous day. The procedures followed were in accordance with institutional regulations for the care and use of laboratory animals.
Lipoproteins and apolipoproteins
EDTA-plasma was supplemented with 0.005% gentamycin-1 mM EDTA-0.04% Na-azide and protease inhibitors, and ultracentrifuged to prepare LDL and HDL as described (7, 8). LDL and HDL protein contents were measured (24), and LDL and HDL were labeled with 1,1'-dioctadecyl-3,3,3',3'-tetramethylindocarbocyanine (DiI) (25). Apolipoproteins were analyzed in plasma, HDL, and urine (collected at 4°C, concentrated 100 times and extensively dialyzed using Amicon Ultra-4 tubes with a 5 kDa cutoff) by 18% SDS-PAGE (26) under nonreducing conditions in order to detect the dimeric form of hapoA-II. After protein transfer to nitrocellulose membranes (0.45 µm; BioRad), hapoA-II was detected by Western blotting using a specific rabbit anti-hapoA-II antibody not recognizing murine apolipoprotein A-II (mapoA-II) (5, 6), and bands were visualized with an alkaline phosphatase substrate system (BioRad).
Immunocytochemical analysis of mouse kidney
Kidneys were quickly removed from control and transgenic mice after intracardiac vascular washing with PBS. Kidney samples were then either frozen in liquid nitrogen after 2 h fixation in 2% paraformaldehyde (PFA) or fixed for 24 h in 4% formalin and embedded in paraffin. Both types of sections (frozen or paraffin-embedded) were processed for immunolocalization. Rabbit anti-mouse apoA-I antibody (Biodesign) was diluted 1/1,500, rabbit anti-mouse apoA-II antibody (Biodesign) was diluted 1/2,000, and rabbit and goat anti-hapoA-II antibodies (courtesy of A. Mazur, National Institute for Agricultural Research) were diluted 1/8,000. Rabbit anti-cubilin and rabbit and sheep anti-megalin antibodies have been described previously (27) and were diluted 1/1,000. Secondary antibodies produced in donkey (Jackson Immunolab; diluted 1/200) were labeled with Cy2 (laser excitation at 488 nm) or Cy3 (543 nm). Double fluorescence images were obtained after mixing of primary antibodies, with appropriate controls. Tissue sections were examined with a laser scanning confocal microscope (Zeiss LSM-510). For conventional bright-field immunomicroscopy, paraffin sections were stained with tyramide amplification (NEN Life Science Products; TSA biotin system NEL700), followed by diaminobenzidine (DAB) precipitation without counterstain.
Histological quantification of apolipoprotein tubular reabsorption
This analysis was conducted with the help of a microscope linked to a computer-based image visualization of DAB-stained paraffin sections. Views of random renal cortical fields at x25 magnification were digitalized; total number of tubular sections per field was counted; and the number of tubes stained with DAB product signaling apolipoprotein apical reabsorption was noted. Results were expressed as the percentage of stained tubes/total tubes for each group of animals.
Immunoelectron microscopy
Following intracardiac vascular washing with 0.1 M phosphate buffer and fixation of tissues with 4% PFA in 0.1 M phosphate buffer, kidneys of -transgenic mice were trimmed and immersion-fixed in 4% PFA in 0.1 M phosphate buffer for 4 h at 4°C. Tissues were dehydrated through graded alcohol, infiltrated in a mixture of ethanol-LRWhite resin (AGAR, R 12 81), and then embedded and polymerized in LRWhite resin at 37°C over 5 days. Ultrathin sections (60 nm) were cut on the ultratome Leica Ultracut UCT, and picked up on Formvar/carbon-coated 300-mesh nickel grids. Immunolabeling with rabbit anti-hapoA-II antibody (1/500) was performed overnight at 4°C and visualized by incubation with donkey anti-rabbit-gold (Aurion; 0.8 nm, DAR ultrasmall, 800.311) diluted 1/100 in PBS, 0.8% BSA, 0.1% fish gelatin, and 15 mM NaN3 for 2 h at room temperature. Sections were postfixed for 10 min in 2.5% glutaraldehyde in PBS, and the gold particles were amplified with Silver Enhancement (Aurion kit R-Gent SE-ME) for 30 min at room temperature. Finally, the grids were stained with lead citrate and uranyl acetate according to standard procedures, and photographed in a Jeol CX 100 electron microscope.
Purification and labeling of hapoA-II and apoA-I
HapoA-II and apoA-I were purified from delipidated human HDL by preparative isoelectric focusing with a pH gradient of 4–6, and purity was checked by SDS-PAGE in 15% gels. ApoA-II was fluorescently labeled with Alexa 488 and apoA-I with Alexa 546 (Alexa Protein Labeling kits A-10237 and A-10235, respectively; Molecular Probes), according to the manufacturer's instructions.
Cell culture studies
Brown Norwegian (BN) rat yolk sac epithelial cells (BN cells) were derived after fetectomy and placental injection with Moloney's sarcoma virus (28). They were grown in DMEM (Life Technologies) supplemented with 10% fetal calf serum, 1% penicillin, and 1% streptomycin, in 5% CO2. MKC epithelial cells were derived from SV40-transgenic mice (29). They were routinely cultured in DMEM medium: Ham's F12 (1:1; v/v), sodium selenite (2 x 10–8 M), nonessential amino acids, glutamine (4 mM), Hepes (33 mM), NaHCO3 (25 mM), penicillin/streptomycin (50 UI/ml), fungizone (0.25 µg/ml), supplemented with insulin (0.5 µg/ml), triiodothyronin (1 µM), transferrin (5 µg/ml), cholera toxin (10 ng/ml), epidermal growth factor (10 ng/ml), dexamethasone (50 nM), and 5% fetal calf serum. They were seeded on semipermeable filters coated with extracellular matrix basement membrane (matrix-derived from Englebreth-Holm-Swarm mouse tumor; Harbor Bioproducts), and incubated in 5% CO2 at 37°C.
Uptake of apoA-II by yolk sac cells
BN cells were grown to 90% confluence on 4-chamber borosilicate slides (Nulge Nunc) in DMEM, washed three times in PBS, and incubated in 250 µl serum-free DMEM supplemented with 0.1% BSA. A first kinetic study was performed to investigate apoA-II uptake: Alexa 488-apoA-II and Alexa 546-apoA-I were separately added to a final concentration of 0.25, 0.5, 1, or 2 µg/ml and incubated at 37°C for 1, 5, 15, or 30 min. The cells were washed three times in PBS and fixed in 4% PFA for 30 min, and apolipoprotein internalization was analyzed by confocal microscopy. Internalization of unlabeled hapoA-II and hapoA-I was studied by incubating cells with 0.25 µg/ml of each apolipoprotein at 37°C for 15 min, followed by washes with PBS, fixation in PFA, and permeabilization with 0.1% Triton X-100 in PBS/3% BSA for 5 min. Cells were then washed with PBS containing 100 mM glycine, incubated with PBS at room temperature for 60 min, first with anti-hapoA-II or anti-hapoA-I antibodies (diluted 1/1,000) and then with Cy2- or Cy3-labeled secondary antibodies (diluted 1/200), and analyzed by confocal microscopy. To study colocalization of Alexa-apoA-II and Alexa-apoA-I, the cells were incubated with both apolipoproteins (2 µg/ml each) at 37°C for 1, 5, or 30 min. After three washes with PBS, cells were fixed in 4% PFA, washed with PBS/glycine, and analyzed by confocal microscopy. To study colocalization of apoA-II and apoA-I with cubilin and megalin, cells were incubated at 4°C for 1 h with either Alexa-apoA-II or Alexa-apoA-I (2 µg/ml of each), washed with PBS, fixed in PFA, washed with PBS/glycine and incubated with rabbit anti-cubilin or anti-megalin antibodies (diluted 1/1,000), followed by incubation with Cy2- or Cy3-labeled secondary antibodies (diluted 1/200) and confocal analysis. To establish nonspecific fluorescence, the cells were incubated with unlabeled apoA-II and apoA-I and then with Cy2- and Cy3-labeled secondary antibodies, omitting the primary antibodies. Optical sections were dual-scanned for both green (Alexa 488-apoA-II) and red (Alexa 546-apoA-I) wavelength.
Competition studies of apoA-II internalization in yolk sac cells
To study whether apoA-II competed with apoA-I for internalization, BN cells were incubated at 37°C for 5 min with each Alexa-apolipoprotein alone (0.25 µg/ml) or with either hapoA-I (500-fold excess, mol/mol), hapoA-II (500-fold excess, mol/mol), HDL from control mice (52 µg/ml), HDL from transgenic mice (52 µg/ml), or LDL from control mice (52 µg/ml). To study whether the internalization of apoA-II could be inhibited by anti-cubilin antibodies, cells were incubated at 37°C for 10 min in serum-free DMEM containing rabbit anti-cubilin IgG (100 µg/ml) purified from anti-cubilin antiserum by protein G-Sepharose chromatography (Amersham-Pharmacia Biotechnology) or normal goat IgG (100 µg/ml). Alexa 488-apoA-II was then added, to a final concentration of 0.25 µg/ml, and incubated at 37°C for 5 min. At the end of incubation, the cells were washed three times in PBS, fixed in 4% PFA, washed in PBS/glycine, and analyzed by confocal microscopy, as described above. To quantify fluorescent apolipoprotein uptake in the presence of competitors, stack images of the cells (z axis) were taken using the confocal microscope. For each stack, total fluorescence of the field was estimated as the sum of the integrated optical density (IOD) of each slice measured with Scion Image (Scion Corporation). The mean value of four stacks was used to compare the different conditions.
Confocal microscopic analysis of kidney cortical cells
MKC cells were tested for the presence of cubilin and megalin on their surfaces and for uptake of apoA-II, apoA-I, and LDL. MKC cells were incubated at 37°C with Alexa 488-apoA-II or Alexa 546-apoA-I (1 and 2 µg/ml) for 15 min or 2 h. To study uptake of LDL, cells were incubated with DiI-LDL (2 and 4 µg/ml) for 15 min. At the end of incubation, cells were processed as described above and analyzed by confocal microscopy.
Receptor immunoblot analysis
Membrane proteins of kidney, BN, and MKC cells were prepared as described (30). Membrane extracts (10 µg protein for kidney and BN cell extracts, and 86 µg protein for MKC cell extracts) were analyzed by 4–15% SDS-PAGE. After protein transfer to nitrocellulose membranes (0.45 µm; BioRad), cubilin and megalin were detected with specific rabbit anti-cubilin and anti-megalin antibodies diluted 1/2,000, and secondary rabbit peroxydase-conjugated antibodies diluted 1/10,000, and visualized with an ECL kit (RPN 2106; Amersham).
Cubilin and megalin expression analysis
RNA was isolated from kidney using the RNA Instapure Kit (Eurogentec), and RT was carried out. Cubilin, megalin, and 18S were quantified in the same RT reactions by real-time PCR (Light Cycler; Roche). The sequences used for detection of murine cubilin-complementary DNA (cDNA) were as follows: sense 5'-CAGACACCCACCATGGATATGA-3'and antisense 5'-TTCGTTGCCCAAAAGAGTCC-3'; cubilin-cDNA was quantified with the Taqman probe 5'-TCTGGGCCTCCTCTCCTACTGGCTG-3'. The following sequences were used for detection of mouse megalin-cDNA: sense 5'-TCCAAAGTGGGCCTGGC-3' and antisense 5'-TTTTCCACATTTTCTGGTACAGTCA-3'; megalin-cDNA was quantified with the Taqman probe 5'-TGCAGGGACCGTCAGTAAGCTACAG-3'. 18S was quantified using the Taqman probe (Applied Biosystems; 4310893E). Cubilin and megalin expression levels were normalized to that of 18S.
Statistical analysis
Results are given as mean ± SEM, and statistically significant differences were determined with Tukey's multiple comparison test after ANOVA, using GraphPad Prism.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-transgenic mice (Fig. 1A1), but not in proximal tubules of control mice (Fig. 1A2), showing the specificity of the anti-hapoA-II antibody. ApoA-II was present exclusively in proximal tubules, mainly in the first (S1) segment adjacent to the glomerulus, and was not detected in glomeruli and distal nephron (not shown). As a control, mapoA-I, known to be catabolized in the kidney (13–17), was also immunolocalized only in the S1 segment of kidney proximal tubules of the same mice (not shown). To our knowledge, these data are the first report of apoA-II reabsorption in the kidney.
|
ApoA-II colocalizes with apoA-I, cubilin, and megalin, and is targeted to lysosomes of kidney proximal tubules
In these experiments, kidney sections were prepared from the higher-expressing mice, and immunostaining of hapoA-II, mapoA-I, cubilin, and megalin was performed. ApoA-II (Fig. 2A2), as apoA-I (Fig. 2A1), was present in the apical cytoplasm in punctate form, corresponding to small vesicles, and both apolipoproteins partially colocalized in the same vesicles (Fig. 2A3 and Fig. 2A4
). Cubilin (Fig. 2B1) and megalin (Fig. 2C1) were abundantly expressed and lined the entire brush border apical membrane of the tubules, where they colocalized (not shown). Of note, apoA-II partially colocalized with cubilin (Fig. 2B3) and somewhat with megalin (Fig. 2C3) only in the brush border and was present alone in small vesicles in epithelial cells just below the membrane. Similar results were obtained for apoA-I, which served as a positive control (not shown).
|
mice. Immunogold apoA-II was present near the brush border membrane, in endocytic invaginations and vacuoles, and was concentrated in lysosomes (Fig. 2D1-D4). The intense apoA-II immunostaining in lysosomes is suggestive of apoA-II degradation in kidney proximal tubules.
Renal catabolism of apoA-II is a function of its plasma concentration
To study the relationship between renal catabolism and plasma level of hapoA-II, we compared the amount of hapoA-II immunostaining in kidney sections from the three hapoA-II transgenic lines ß, , and with plasma hapoA-II concentrations of 20, 40, and 60 mg/dl, respectively (Table 1
). Furthermore, we modulated plasma hapoA-II by overnight fasting all transgenic lines, because fasting lowers plasma hapoA-II up to 40% in the mice (8). Indeed, hapoA-II is lower in total plasma and HDL fractions of fasted as compared with fed transgenic mice (Fig. 1C). In the fed state, the percentage of tubules immunostained for hapoA-II was lowest in ß mice and similar between and mice, despite the higher plasma content of hapoA-II in compared with mice (Table 1 and Fig. 1B). A possible explanation is that lysosomal degradation of apoA-II proceeds at similar rates in and mice. In each transgenic line, immunostaining of hapoA-II was consistently lower in fasted compared with fed mice. Immunostaining of cubilin and megalin showed that both proteins were abundantly present in brush border membranes of fed and fasted transgenic and control mice. We also verified that kidney mRNA levels of cubilin and megalin were similar among all genotypes in both the fed and fasted states, indicating that the lower renal uptake of apoA-II in fasted mice could not be due to decreased cubilin and/or megalin expression levels (Table 2
). Rather, renal apoA-II catabolism is related to circulating apoA-II concentration, suggesting an association between hapoA-II concentration in plasma and reabsorption in the kidney.
|
|
|
|
Lack of apoA-II and apoA-I internalization in megalin-expressing MKC cells
To assess whether megalin is sufficient to mediate apoA-II and apoA-I endocytosis, we used MKC cells expressing megalin but not cubilin. We first established that the megalin mRNA amount of MKC cells was comparable to that of kidney, whereas cubilin mRNA was not detected (not shown). Western blot analysis of total membrane proteins revealed the presence of megalin and the complete absence of cubilin in MKC cells (Fig. 4A). These results were confirmed by immunohistochemistry of MKC cells using anti-megalin and anti-cubilin antibodies (not shown). We then showed that DiI-LDL was taken up efficiently, establishing that megalin was functional (Fig. 4B1). However, neither Alexa-apoA-II nor Alexa-apoA-I was internalized in MKC cells (Fig. 4B2 and 4B3, respectively). Thus, megalin alone cannot internalize apoA-II and apoA-I.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This study shows that apoA-II shares renal clearance with apoA-I, the major HDL apolipoprotein, and provides physiologically relevant data for in vivo apoA-II catabolism in the kidney. Indeed, endogenous unlabeled mapoA-II and hapoA-II were detected by immunohistochemistry in proximal tubules of control and hapoA-II-transgenic mice, respectively. As a positive control, native mapoA-I was immunolocalized in proximal tubules of all mice. Previous studies of in vivo apoA-I catabolism after injection of 125I-apoA-I-HDL into rats, rabbits, and mice detected 125I-apoA-I in tissue homogenates, either primarily in kidney (13) or in liver (14). However, it was shown in rabbits that 125I-apoA-I did not behave in vivo as unlabeled apoA-I in terms of tissue uptake: endogenous apoA-I was catabolized predominantly in the kidney, whereas about half of 125I-apoA-I was catabolized in the kidney and half in the liver (15). We therefore avoided labeling apoA-II with 125I in this study, and instead, opted for immunolocalization of native endogenous apoA-II in kidney sections. The observation that apoA-II and apoA-I were reabsorbed in the first segment of proximal tubules and were totally absent in urine suggests a very efficient uptake, as in the case of IF (23). Internalization of apoA-II was followed by its targeting to lysosomes, suggestive of degradation, and similarly to transferrin, albumin, and IF (23).
The use of three hapoA-II-transgenic mouse lines with a gradient of hapoA-II expression allowed us to show that apoA-II reabsorption in the kidney is related to its plasma concentration. The decrease in plasma hapoA-II resulting from fasting was also accompanied by a decrease in apoA-II content in the kidney. On the other hand, cubilin and megalin were abundantly present in brush border membranes of fed and fasted mice of all genotypes, and their mRNAs were similar. Taken together, these observations suggest that plasma apoA-II concentration is regulated by its rate of synthesis and not by its rate of catabolism, in accordance with apoA-II turnover studies (11).
A major question arising from the present data is whether apoA-II is cleared in the kidney in lipid-poor form or associated with HDL. Because of its high hydrophobicity, hapoA-II binds tightly to the surface phospholipids of HDL, and it is not known whether part of apoA-II circulates in plasma in lipid-poor form, as does apoA-I (17). Furthermore, we did not detect apoA-II in the lipoprotein-deficient fraction (d > 1.21 g/l) of human or murine plasma (unpublished observations). In our transgenic mice, hapoA-II associates with both large (10 nm) and small (7.8 nm) HDL particles, the latter predominant in the high-expressing mice (7, 8). It remains indeterminate whether in vivo HDL holoparticles filter through the glomerulus. At present, we can only speculate that some small HDL particles carrying hapoA-II may filter through the glomerulus and be reabsorbed in the proximal tubules. Regarding apoA-I, it is probably reabsorbed in kidney in lipid-poor form (13), and an inverse relationship has been demonstrated between serum-free apoA-I levels and glomerular filtration rates in patients with renal failure (17).
An important issue concerns the proteins/receptors mediating apoA-II internalization in kidney proximal tubules. Based on several lines of evidence obtained in vivo and in vitro, the present data indicate that cubilin may be an apoA-II receptor in the kidney, although they do not exclude the existence of other apoA-II receptors. Native apoA-II taken up in vivo was detected in kidney sections by immunostaining, and colocalized to a great extent with apoA-I, an established cubilin ligand, in endocytic vesicles below the brush border membrane of tubular epithelial cells. ApoA-II was internalized only in proximal tubules, mainly in the S1 segment adjacent to the glomeruli, where cubilin and megalin are abundantly expressed, and was not detected in distal or collecting tubules, where cubilin and megalin are not expressed (23). This pattern of internalization is similar to that reported for known ligands of the cubilin-megalin complex (23). Conversely, cubilin and megalin were not detected in intracellular vesicles by confocal microscopy, suggesting that they rapidly recycle to the apical membrane from an early endosomal compartment. Indeed, electon microscopic studies clearly showed cubilin and megalin colocalization on the luminal membrane, in endocytic compartments, and in the recycling compartment dense apical tubules (23). In yolk sac epithelial cells with abundant cubilin expression, apoA-II colocalized to a great extent with cubilin at the cell surface, as did apoA-I. More importantly, both apolipoproteins competed equally efficiently with each other for uptake, and uptake of apoA-II was very efficiently blocked by HDL. Moreover, apoA-II internalization was inhibitable by anti-cubilin antibodies, to the same extent as apoA-I internalization. Previously, Hammad et al. (19) showed that HDL endocytosis by cubilin- and megalin-expressing endoderm-like F9 cells was inhibitable by apoA-II. Conversely, MKC cells that express megalin but not cubilin do not internalize apoA-II and apoA-I, in agreement with the Hammad et al. (22) report that megalin did not bind HDL, delipidated HDL, or apoA-I.
Our results clearly show that megalin is not an apoA-II and apoA-I receptor. The similarities of apoA-I and apoA-II uptake suggest that cubilin is their common receptor. However, cubilin requires megalin as a coreceptor for ligand internalization; and a second accessory protein, amnionless (AMN), that binds to its amino-terminal third, ensures cubilin trafficking to the cell surface, and directs subcellular localization and endocytosis of cubilin and its ligands (32, 33). Fyfe et al. (32) have therefore proposed the existence of a functional complex of cubilin/AMN named "cubam." AMN is expressed in the same tissues as cubilin and megalin and is also expressed in the BN yolk sac cells used in the present study (P. Verroust, unpublished observations). Thus, our results are equally well explained by binding of apoA-II to cubilin or to cubam, whereas internalization necessitates megalin, as demonstrated by several studies.
This study did not address the relative uptake of apoA-II in the kidney compared with any other organ. If apoA-II is mainly catabolized as an HDL component, it may share with apoA-I other sites of catabolism. When 125I-apoA-I-HDL was injected in vivo, 125I-apoA-I was detected primarily in tissue homogenates of kidney, ovary, and adrenal, and in lesser proportions in liver and spleen (13). It is noteworthy that the relative proportions of 125I-apoA-I-HDL taken up by kidney and liver vary among studies; in some studies, uptake is greater in kidney (13, 15, 34), and in other studies, in liver (14, 35). A study using independently generated hapoA-II-transgenic mice showed an increased 125I-apoA-I-HDL catabolism in kidney in the high-expressing 11.1 line, compared with the low-expressing 25.3 transgenic line or with control mice (34). The greater apoA-I catabolism in 11.1 transgenic mice was attributed to the displacement of mapoA-I from HDL by the more hydrophobic hapoA-II, and to the generation of smaller HDL particles (reproduced in vitro). On the other hand, HDL-apoA-I catabolism in liver and kidney was comparable between wild-type and scavenger receptor class B type I (SR-BI) knockout mice (35). A 95 kDa high-affinity HDL binding protein, which binds HDL3, apoA-I, and apoA-II, has been characterized in hepatic and nonhepatic cell lines (36). Another important tissue for HDL uptake is the developing embryo, in which cholesterol is essential for growth as well as for steroid hormone synthesis by embryonic and extraembryonic tissues. HDL-cholesterol is particularly important in rodents that have low plasma levels of remnants/LDL. The visceral yolk sac epithelium is the functional fetal-maternal barrier and takes up maternal lipoproteins via several receptors: LDL-receptor, cubilin, megalin, and AMN, SR-BI, with finely tuned temporal expression patterns for each receptor (37–40). Cubilin is expressed very early on the apical surface of yolk sac, before SR-BI [expressed on day 10.5 (37)], and its deficiency in cubilin-knockout embryos completely inhibited DiI-HDL uptake at day 8.0 (40). At day 10.5, HDL-apoA-I and HDL-cholesteryl esther were cleared at equally very high rates in the yolk sac, indicating that the whole HDL particle was internalized (38). HDL holoparticle uptake may thus be mediated by cubilin, already expressed in the yolk sac on day 8.0 (40), whereas SR-BI mediates selective HDL-cholesterol uptake (37). Because cubilin has a restricted pattern of expression (kidney proximal tubules, distal small intestine, and visceral endoderm of the embryo), other receptor pathways for apoA-II are expected to be operative in other organs. Further in vivo studies are needed to investigate apoA-II catabolism in organs besides kidney, to identify the receptors involved, and to estimate the relative contributions of individual organs to apoA-II catabolism.
In conclusion, the present data demonstrate for the first time a physiologically relevant clearance of apoA-II in the kidney, which occurred in relation to plasma apoA-II concentration. ApoA-II reabsorption in kidney proximal tubules is probably mediated by cubilin, like that of apoA-I, and is probably followed by degradation in lysosomes. Future studies will address the question of whether apoA-II is cleared in the kidney in a lipid-poor form or as a component of HDL subclasses, as well as the role of other organs in apoA-II catabolism. Our transgenic mouse models expressing hapoA-II at various levels will be useful for gaining insight into the in vivo role of apoA-II in HDL remodeling and catabolism, which are important parameters underlying the protective actions of HDL on cardiovascular artery disease.
|
ACKNOWLEDGMENTS |
|---|
Manuscript received February 16, 2007 and in revised form July 5, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
