Cubilin, the endocytic receptor for intrinsic factor-vitamin B(12) complex, mediates high-density lipoprotein holoparticle endocytosis.

Abstract

Receptors that endocytose high-density lipoproteins (HDL) have been elusive. Here yolk-sac endoderm-like cells were used to identify an endocytic receptor for HDL. The receptor was isolated by HDL affinity chromatography and identified as cubilin, the recently described endocytic receptor for intrinsic factor-vitamin B(12). Cubilin antibodies inhibit HDL endocytosis by the endoderm-like cells and in mouse embryo yolk-sac endoderm, a prominent site of cubilin expression. Cubilin-mediated HDL endocytosis is inhibitable by HDL(2), HDL(3), apolipoprotein (apo)A-I, apoA-II, apoE, and RAP, but not by low-density lipoprotein (LDL), oxidized LDL, VLDL, apoC-I, apoC-III, or heparin. These findings, coupled with the fact that cubilin is expressed in kidney proximal tubules, suggest a role for this receptor in embryonic acquisition of maternal HDL and renal catabolism of filterable forms of HDL.

Full text

Page 1

Proc. Natl. Acad. Sci. USA
Vol. 96, pp. 10158–10163, August 1999
Cell Biology
Cubilin, the endocytic receptor for intrinsic factor-vitamin B12
complex, mediates high-density lipoprotein
holoparticle endocytosis
SAMAR M. HAMMAD*, STEINGRIMUR STEFANSSON†, WALEED O. TWAL*, CHRISTOPHER J. DRAKE*, PAUL FLEMING*,
ALAN REMALEY‡, H. BRYAN BREWER, JR.‡, AND W. SCOTT ARGRAVES*§
*Medical University of South Carolina, Department of Cell Biology, Charleston, SC 29425;†Department of Biochemistry, J. H. Holland Laboratory, American
Red Cross, Rockville, MD 20855; and‡Molecular Disease Branch National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD 20892
Communicated by Marilyn Gist Farquhar, University of California at San Diego, La Jolla, CA, July 6, 1999 (received for review May 20, 1999)
ABSTRACT
poproteins (HDL) have been elusive. Here yolk-sac endoderm-
like cells were used to identify an endocytic receptor for HDL.
The receptor was isolated by HDL affinity chromatography
and identified as cubilin, the recently described endocytic
receptor for intrinsic factor-vitamin B12. Cubilin antibodies
inhibit HDL endocytosis by the endoderm-like cells and in
mouse embryo yolk-sac endoderm, a prominent site of cubilin
expression. Cubilin-mediated HDL endocytosis is inhibitable
by HDL2, HDL3, apolipoprotein (apo)A-I, apoA-II, apoE, and
RAP, but not by low-density lipoprotein (LDL), oxidized LDL,
VLDL, apoC-I, apoC-III, or heparin. These findings, coupled
with the fact that cubilin is expressed in kidney proximal
tubules, suggest a role for this receptor in embryonic acqui-
sition of maternal HDL and renal catabolism of filterable
forms of HDL.
Receptors that endocytose high-density li-
High-density lipoprotein (HDL) particles are known to facil-
itate the transport of cholesterol from extrahepatic tissues to
the liver for repackaging into new lipoproteins, bile acid
synthesis,orexcretionintothebile(1,2).Thisprocesshasbeen
proposed to be an important mechanism for the antiathero-
genic effects of HDL. HDL also provides cholesterol to the
ovaries, testes, placenta, and adrenal glands for steroid bio-
synthesis.Moreover,HDLisimportantinmaternal-embryonic
nutrition, providing cholesterol, triglycerides, and lipid-soluble
vitamins to the placenta, yolk sac, and embryo (3, 4).
Characterization of factors involved in the regulation of HDL
catabolism has been the focus of intensive research (5–7). As a
result,itisknownthatthekidneyandliveraremajorsitesofHDL
catabolism and that lipolytic modification of the particle and
structural integrity of its apolipoprotein (apo) components are
significant factors in determining the half life of HDL (5, 8–12).
The mechanism by which HDL is catabolized in the kidney, liver,
and other tissues such as the placenta and yolk sac is not fully
understood. The scavenger receptor SR-BI has been shown to
function as an HDL ‘‘docking’’ receptor that mediates selective
cholesterol uptake (13–15). However, unlike catabolism of low-
density lipoprotein (LDL) by the classic LDL receptor, HDL
catabolism by SR-BI does not involve holoparticle uptake and
lysosomal degradation. This conclusion was supported by the
finding that transgenic mice deficient in SR-BI display elevated
levels of plasma HDL cholesterol yet exhibit no change in the
level of plasma apoA-I (15). Endocytosis and lysosomal degra-
dation of holoparticle HDL is nevertheless known to occur (16),
but endocytic HDL receptors have remained elusive. In the
present study, we have identified a HDL receptor and charac-
terized its capacity to mediate HDL holoparticle endocytosis
leading to lysosomal degradation.
MATERIALS AND METHODS
Proteins. Human apolipoproteins apoA-I, apoA-II, apoC-I,
and apoC-III were purified as described previously (17, 18).
Human apoE3 was purchased from Calbiochem, apoJ from
Quidel (San Diego), and BSA, ovalbumin, and heparinase I from
Sigma.RecombinanthumanRAPwaspurifiedasdescribed(19).
Lipoproteins. Human 1,1?-dioctadecyl-3,3,3?,3?-tetramethylin-
docarbocyanine (DiI)-HDL, 3,3?-dioctadecyloxacarbocyanine,
perchlorate (DiO)-LDL, and rabbit ?VLDL were purchased
fromBiomedicalTechnologies.HumanHDL(density1.063–1.21
g?ml), HDL2, HDL3, delipidated HDL (apoHDL), and LDL
were prepared as described (20, 21). DiI-HDL, HDL, apoHDL,
HDL2, and HDL3were depleted of apoE-HDL and other hep-
arin-binding particles according to Oram (22), dialyzed against
150 mM NaCl, 50 mM Tris pH 7.4 (TBS) containing 0.3 mM
EDTA, and filter sterilized. Human lipoprotein (a) [Lp(a)] was
obtained from Peter Harpel (Mount Sinai Medical Center, New
York), oxidized LDL from Alicia Jenkins (Medical University of
South Carolina, Charleston, SC), and human VLDL (Svedberg
flotationrate100–400)fromDavidChappell(UniversityofIowa
College of Medicine, Iowa City, IA). Lipoprotein concentration
was determined by BCA (bicinchoninic acid) protein assay
(Pierce).
Antibodies.IgGsfromrabbitantimegalin(23,24)werepurified
by protein-G-Sepharose chromatography. Rabbit anticubilin se-
rum and IgG were provided by Pierre Verroust (Hospital Tenon,
Paris, France). Horseradish-peroxidase-conjugated anti-rabbit
IgG was obtained from Amersham Pharmacia.
Cells. Mouse embryonal teratocarcinoma F9 cells (ATCC
CRL1720) were differentiated by treatment with retinoic acid
(RA) and dibutyryl cyclic AMP (Bt2cAMP) for 6 days as
described (25).
125I-HDL Binding, Internalization, and Degradation As-
says. HDL internalization and degradation assays were per-
formed as described for LDL and apoJ (23, 25, 26). ApoE-free
HDL was labeled with [125I]-iodine by using the iodine mono-
chloride method (27). RA?Bt2cAMP-treated and untreated
F9 cells were seeded into gelatin-coated wells (Corning) at 4 ?
104cells?cm2and allowed to grow for 18 h in serum-free
medium (SFM) [DMEM containing ITS (insulin 5 ?g?ml,
transferrin 5 ?g?ml, sodium selenite 5 ng?ml, Boehringer
Mannheim)], and penicillin?streptomycin). Before addition of
125I-HDL, the cells were washed with SFM and incubated in
The publication costs of this article were defrayed in part by page charge
payment. This article must therefore be hereby marked ‘‘advertisement’’ in
accordance with 18 U.S.C. §1734 solely to indicate this fact.
PNAS is available online at www.pnas.org.
Abbreviations: HDL, high-density lipoprotein; LDL, low-density li-
poprotein; RA, retinoic acid; SFM, serum-free medium; CHAPS, 2 ml
of 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate;
DiI, 1,1?-dioctadecyl-3,3,3?,3?-tetramethylindocarbocyanine;
Bt2cAMP, dibutyryl cyclic AMP; dPBS, Dulbecco’s phosphate-
buffered saline; apo, apolipoprotein; Lp(a), lipoprotein (a).
§To whom reprint requests should be addressed.
10158

Page 2

the assay medium (DMEM?20 mM Hepes?ITS?penicillin?
streptomycin?1.5% ovalbumin) alone or in the presence of
HDL (80 ?g?ml), RAP (1 ?M), chloroquine (50 ?M), BSA (1
?M), anticubilin IgG (150 ?g?ml) or control IgG (150 ?g?ml)
and incubated for 30 min at 37°C, 5% CO2. The medium was
then removed and apoE-free
medium alone or in assay medium containing HDL, RAP,
chloroquine,BSA,orIgGsattheaforementionedconcentrations
wasaddedandincubatedwiththecellsfor6.5h.Theconditioned
medium was treated with trichloroacetic acid (10% final) and
centrifuged at 10 K ? g for 10 min. The amount of radioactivity
inthesupernatantwastakentorepresenttheamountofdegraded
HDL (28). The cell layer was washed three times with ice-cold
Dulbecco’s phosphate buffered saline (dPBS) and then treated
with 0.5 mg?ml trypsin, 50 ?g?ml proteinase K, 0.53 mM EDTA
in dPBS (trypsin-K-EDTA) for 3 min at 4°C. The released cells
were pelleted, and the amount of radioactivity in the pellet was
measured and taken to represent the amount of internalized
HDL.
To study HDL binding to RA?Bt2cAMP-treated F9 cells, a
filter assay described previously (29) for measuring binding of
[125I]-LDL to precipitated oocyte membranes was adopted.
Aliquotsofmembraneprecipitatescontaining45 ?gofprotein
were combined with the indicated amounts of125I-HDL and
unlabeled HDL in 100 ?l (25 mM NaCl?2 mM CaCl2?16
mg?ml BSA?12.5 mM Tris?HCL, pH 8.0 final buffer compo-
sition) and incubated for 2 h at 24°C. Receptor-bound125I-
HDL was collected on 0.45-?m cellulose acetate filters (Mi-
cron Separations) as described (29). Binding data were fit to
a single class site model by using LIGAND (30) and assuming a
molecular mass for HDL of 200 kDa.
Measurement of DiI-HDL and DiO-LDL Uptake by Flow
Cytometry. RA?Bt2cAMP-treated and untreated F9 cells (1–
1.5 ? 105cells?cm2) were grown for 18 h in SFM. After the cell
layers were washed with SFM, DiI-HDL or DiO-LDL was added
to a final concentration of 1 ?g?ml and incubated for 2 h at 37°C,
5% CO2. For competition experiments, competitors (HDL,
HDL2, HDL3, LDL, oxidized LDL, VLDL, ?-VLDL, Lp(a) (100
?g?ml),apoA-I(28.3?g?ml,1?M),apoA-II(17.3?g?ml,1?M),
apoC-I (6.6 ?g?ml, 1 ?M), apoC-III (8.8 ?g?ml, 1 ?M), apoJ (70
?g?ml, 1 ?M), ovalbumin (45 ?g?ml, 1 ?M), RAP (39 ?g?ml, 1
?M),apoE(8.5?g?ml,0.25?M),heparin(80?M,250units?ml)
orIgGs(200?g?ml)wereaddedinSFMandincubatedfor45min
before the addition of the fluorescent lipoproteins. After the
incubation, the medium was removed, cell layers washed with
SFM,andthecellswerereleasedwithtrypsin-K-EDTA.Thecells
were washed once with DMEM, twice with dPBS, then subjected
to flow cytometry (FACStar Plus, Becton Dickinson). Plotted
data are from gated cells having fluorescence intensity values ?
the autofluorescence values of 99% of unlabeled cells. For some
experiments, cell layers were treated with heparinase I (10
units?ml) for 2 h before the addition of the DiI-HDL. This
treatment reduced the levels of cell layer-associated [35SO4] by
50% that of controls.
HDL-Sepharose Affinity Chromatography. HDL-Sepharose
affinity chromatography was performed by using detergent ex-
tracts of RA?Bt2cAMP-treated and untreated F9 cells that were
either cell surface [125I]-labeled or unlabeled. For radiolabeling,
cells were washed with dPBS and detached by using 2 mM
EDTA?dPBS. The released cells were washed twice with
DMEM, then twice with dPBS. Cells (1 ? 108) were radioiodi-
nated by using the lactoperoxidase?glucose oxidase method (31).
After labeling, the cells were resuspended in 50 mM octyl-?--D-
glucoside (Calbiochem) in TBS containing 1 mM CaCl2?mM
MgCl2?2 mM PMSF (OG buffer) and passed repeatedly through
a 21-gauge needle. The extract was clarified by centrifugation at
100K?gfor1h.ExtractswereappliedtocolumnsofSepharose-
CL4B (5 ml) equilibrated with the OG buffer. Sepharose CL4B-
absorbed extracts were applied to columns of apoHDL (apoE-
free) coupled to cyanogen bromide-activated-Sepharose (10 mg
125I-HDL (2 ?g?ml) in assay
protein?mlresin).ThecolumnswerewashedwithOGbufferand
theboundproteinselutedwithsequentialapplicationsof1,4,and
8 M urea?50 mM Tris, pH 7.4. Peak fractions were pooled and
dialyzed against TBS, 1 mM PMSF and absorbed on wheat germ
lectin-agarose (Amersham Pharmacia). Bound proteins were
eluted with 0.5 M N-acetyl-glucosamine in OG buffer, separated
by SDS?PAGE, and analyzed by Coomassie staining and auto-
radiography.
Nonradiolabeled cells (1 ? 109) were released and washed as
described above, then suspended in 0.25 M sucrose?10 mM
Hepes, pH 7.4, containing an EDTA-free protease inhibitor
mixture (Boehringer Mannheim). The cells were homogenized
on ice by using a Polytron homogenizer (Kinematica, Lucerne,
Switzerland) and the homogenates centrifuged at 2 K ? g for 10
min.Theresultingsupernatantswerecentrifugedat100K?gfor
1 h. The pellets were resuspended in 2 ml of 3-[(3-cholamido-
propyl)dimethylammonio]-1-propanesulfonate (CHAPS) buffer
(20 mM CHAPS?1 mM PMSF?TBS) by repeated passage
through a 21-gauge needle. The extract was clarified by centrif-
ugation at 100 K ? g for 1 h. The concentration of protein in the
supernatantswasdetermined,andequalamountsofproteinfrom
each cell type were applied to columns of Sepharose-CL4B
(Amersham Pharmacia) equilibrated with CHAPS buffer.
Sepharose CL4B-absorbed extracts were applied to columns of
apoHDL-Sepharose and incubated 18 h at 4°C with nutational
movement. Bound proteins were eluted with 8 M urea?50 mM
Tris, pH 7.4, separated by SDS?PAGE, and analyzed by silver
staining and immunoblotting.
Confocal Microscopic Analysis of DiI-HDL Uptake by Cul-
tured Cells and Mouse Embryos. Cells were plated into gelatin-
coated wells of plastic chamber slides (Nalge Nunc) (4 ? 104
cells?cm2) and incubated for 2 h at 37°C. The cells were washed
and incubated in SFM for 18 h. Before addition of fluorescent
lipoproteins, the cells were preincubated with SFM containing
HDL, LDL (40 ?g?ml), or RAP (1 ?M) for 45 min. DiI-HDL (1
?g?ml) and DiO-LDL (1 ?g?ml) were then added in SFM alone
orinthepresenceofHDL,LDL(40?g?ml),orRAP(1?M)and
incubatedwiththecellsfor2h.ThecellswerewashedwithdPBS,
fixed in 3% formaldehyde in dPBS for 20 min, and analyzed by
using a confocal microscope (Bio-Rad).
Embryos (8–8.5 days postcoitum) were removed from timed
pregnant ICR mice so as to leave visceral yolk-sac endoderm
intact and placed in DMEM. The embryos were incubated in
DMEM containing RAP (1 ?M), ovalbumin (1 ?M), rabbit
anticubilin IgG (400 ?g?ml), or normal rabbit IgG (400 ?g?ml)
for 0.5 h at 37°C, 5% CO2. DiI-HDL was added to a final
concentrationof1?g?mlandincubatedfor1h.Themediumwas
removed and the embryos washed with dPBS and fixed with 3%
paraformaldehyde for 15 min. Fixed embryos were placed in
dPBS and examined by confocal microscopy. The intensity of
yolk-sac fluorescence in mouse embryos was evaluated by seven
persons in a double-blinded fashion. The assessments of relative
intensitywereanalyzedbyanonparametricstatisticaltestbyusing
RANK and ANOVA procedures (SAS Institute, Cary, NC).
For immunohistological localization of cubilin, embryos
were fixed as above and permeabilized for 30 min in 0.02%
TritonX-100 in dPBS, 0.1% azide and then blocked for 12 h at
4°C in 1% goat serum, dPBS, azide. The embryos were
incubated with anticubilin IgG (10 ?g?ml) in dPBS?goat
serum?azide for 18 h at 4°C, washed in dPBS?azide for 18 h at
4°C, and incubated 18 h with dichlorotriazinyl aminofluores-
cein-labeled goat anti-rabbit IgG (Jackson ImmunoResearch)
in 1% goat serum. The embryos were washed and examined by
confocal microscopy.
RESULTS
F9 cells that were differentiated to yolk-sac endoderm-like
cells with RA?Bt2cAMP treatment were found to internalize
and degrade [125I]-labeled HDL (Fig. 1). The degradation of
125I-HDL occurred in lysosomes, as evidenced by the fact that
Cell Biology: Hammad et al. Proc. Natl. Acad. Sci. USA 96 (1999)10159

Page 3

chloroquine, a drug that inhibits lysosomal proteinase activity,
effectively blocked the degradation (Fig. 1 a and b). In
addition, RAP, a protein that inhibits ligand binding to LDL
receptor (LDLR) family members, inhibited the endocytosis
anddegradationof125I-HDL(Fig.1aandb).Undifferentiated
F9 cells were unable to mediate HDL internalization and
degradation (Fig. 1 c and d). When differentiated F9 cell
membranes were analyzed for their ability to bind125I-HDL,
saturable and high-affinity (Kd? 77 ? 7 nM, n ? 2) binding
was observed (Fig. 1 e and f).
Incubation of the RA?Bt2cAMP-treated F9 cells with HDL
labeled with DiI (a fluorescent lipid) produced a punctate
subcellular staining pattern consistent with a vesicular local-
ization (Fig. 2a). The cell staining was blocked by addition of
excessunlabeledHDLorRAP(Fig.2bandc).Anoverlapping
pattern of punctate fluorescent staining was observed when
the cells were incubated with both DiO-labeled LDL and
DiI-HDL (Fig. 2 d–f). The colocalization of the two labeled
lipoproteins is an indication that HDL is trafficked through the
same endocytic compartments as LDL. Untreated F9 cells did
not show staining after incubation with DiI-HDL but did show
punctate staining after incubation with DiO-LDL (data not
shown). Based on the results obtained by using radiolabeled
and fluorescent lipid-labeled HDL, it can be concluded that
the differentiated F9 cells express a receptor capable of
mediating HDL endocytosis and lysosomal degradation. Fur-
thermore, the fact that RAP can inhibit HDL endocytosis and
degradation suggests that the HDL-endocytosis receptor is
either a RAP-binding protein, or that a RAP-binding member
of the LDLR family is indirectly involved.
A flow cytometry assay by using DiI-HDL was developed to
evaluate uptake of DiI-HDL by RA?Bt2cAMP-treated F9 cells.
DiI-HDL uptake by treated F9 cells was saturable, with half-
maximal saturation occurring at 25 nM (Fig. 3). This value is in
good agreement with the Kdof 77 nM obtained from Scatchard
analysis (Fig. 1f). Little binding of DiI-HDL to untreated F9 cells
was observed (Fig. 3). The assay was also used to evaluate
lipoprotein and apolipoprotein specificity. HDL subclasses
HDL2,HDL3,andtheirapoE-freeformscompetedforDiI-HDL
uptake (Fig. 4 a and b). HDL3was more effective than HDL2,
irrespective of apoE content. The major HDL apolipoprotein
constituents, apoA-I and apoA-II, as well as apoE competed
effectively for DiI-HDL uptake, whereas apoC-I, apoC-III and
apoJ displayed little or no ability to compete (Fig. 4 c and d).
When lipoprotein specificity was evaluated, LDL, VLDL,
?VLDL, and Lp(a) showed little or no effect on HDL uptake
(Fig. 4 e and f). Similarly, oxidized LDL was unable to compete
for HDL uptake (Fig. 4 g and h). Heparinase treatment has been
shown to block apoE- and hepatic lipase-mediated uptake of
HDL by cultured hepatocarcinoma cells (32). However, neither
heparin nor heparinase I pretreatment of RA?Bt2cAMP-treated
F9 cells inhibited DiI-HDL uptake (Fig. 4 g and h). Taken
together,theresultsdemonstratethattheprocessofHDLuptake
mediated by differentiated F9 cells is highly specific, apparently
involving interactions with apoA-I and?or apoA-II moieties.
HDL-Sepharose chromatography was used to isolate the
HDL holoparticle-endocytosis receptor from extracts of sur-
face-radiolabeled RA?Bt2cAMP-treated F9 cells. A major
Coomassie-stainable radiolabeled polypeptide of ?500 kDa
and several minor polypeptides of ?400, 140, and 45 kDa were
present in the HDL-Sepharose eluates derived from the
FIG. 1.
HDL. (a–d) Amounts of apoE-free125I-HDL internalized and de-
graded by RA?Bt2cAMP-treated (a and b) and untreated F9 cells (c
and d) in the presence of unlabeled HDL, RAP, chloroquine, or BSA.
Plotted values are means ?SD of triplicate values. Data in a–d are
representative of five experiments. (e) Saturation curve for the binding
of125I-HDL to precipitated CHAPS extract of RA?Bt2cAMP-treated
cell membranes. Data in e have been corrected for the amount of
nonspecific binding obtained for each concentration of125I-HDL in
the presence of 2.1 ?M?425 ?g?ml unlabeled HDL. (f) Homologous
liganddisplacementassayinwhich125I-HDL(12nM?464cpm?ng)was
incubated with varying concentrations of unlabeled HDL (9.87–2,400
nM) and 45 ?g of protein precipitated from a CHAPS extract of
RA?Bt2cAMP-treated F9 cell membranes. (Inset) Scatchard plot of
binding data shown in f. Each data point in e and f represents the mean
of duplicate determinations.
RA?Bt2cAMP-treated F9 cells internalize and degrade
FIG. 2.
Bt2cAMP-treated F9 cells incubated with DiI-HDL. RA?Bt2cAMP-
treated F9 cells were incubated for 2 h with apoE-free DiI-HDL alone
or in the presence of unlabeled HDL (b) or RAP (c) and examined by
confocal microscopy. Incubation of RA?Bt2cAMP-treated F9 cells
with both DiI-HDL (d) and DiO-LDL (e) reveals an overlapping
vesicular pattern (f). (Bar ? 25 ?m).
(a) Punctate pattern of fluorescence observed in RA?
FIG. 3.
uptake by RA?Bt2cAMP-treated F9 cells. Treated and untreated F9
cells were incubated with varying concentrations of DiI-HDL (1–500
nM) for 2 h then stripped of surface-associated material and analyzed.
Each data point represents the mean ? range of duplicate determi-
nations (1 ? 104cells?determination).
Flow cytometric analysis showing saturability of DiI-HDL
10160Cell Biology: Hammad et al.Proc. Natl. Acad. Sci. USA 96 (1999)

Page 4

treated F9 cell extracts but were not in eluates from the
untreated cells (Fig. 5 a and b).
Megalin?LRP-2 represented a candidate for the ?500-kDa
HDL-Sepharosebindingprotein,givenitssimilarsizeandthatits
expression is augmented by RA?Bt2cAMP treatment of F9 cells
(25) (Fig. 5c). However, immunoblot analysis showed that the
500-kDa HDL-Sepharose-binding protein was not reactive with
antibodies capable of detecting mouse megalin (data not shown).
Another candidate was cubilin, a recently identified 460-kDa
RAP-binding endocytic receptor for intrinsic factor-vitamin B12
(33, 34). In addition, cubilin had been shown to be expressed by
yolk-sac endoderm cells (35), the phenotype of RA?Bt2cAMP-
treated F9 cells. Immunoblotting with cubilin antibodies showed
that cubilin was expressed by RA?Bt2cAMP-treated F9 cells but
not by untreated F9 cells (Fig. 5c). Furthermore, the 500-kDa
HDL-Sepharose binding protein reacted with cubilin antibodies
(Fig. 5 d and e). The antibodies also reacted with an ?800-kDa
polypeptide present in the HDL-Sepharose eluate that corre-
sponds to the previously described 800-kDa cubilin dimer (33).
Cubilin antibodies did not react with polypeptides present in the
HDL-Sepharose eluate derived from the untreated F9 cells. The
resultsindicatethatcubilinisthemajorconstituentpresentinthe
profile of HDL-Sepharose binding proteins isolated from RA?
Bt2cAMP-treated F9 cells.
To directly test the role of cubilin as an endocytic receptor
for HDL, cubilin antibodies were evaluated for their ability to
block DiI-HDL and125I-HDL internalization mediated by the
RA?Bt2cAMP-treated F9 cells. As shown in Fig. 6, cubilin
antibodies effectively inhibited cellular uptake of DiI-HDL (a
and b) but not DiO-LDL (c). Similarly, cubilin antibodies were
found to effectively inhibit both the internalization and lyso-
somal degradation of125I-HDL (Fig. 6 e and f).
The yolk-sac endoderm mediates uptake of maternal-
derived HDL (3, 4) and is a major site of cubilin expression
(35). To determine whether yolk-sac uptake of HDL is medi-
ated by cubilin, mouse embryos having intact visceral yolk sacs
were incubated in medium containing DiI-HDL with or with-
out antagonists of cubilin activity, RAP, or anticubilin IgG. In
the absence of an antagonist, DiI-HDL became incorporated
exclusively within the visceral extraembryonic yolk-sac
endoderm (Fig. 7 a, b, and d). Little or no DiI-HDL was
detected within the intraembryonic endoderm. Consistent
with these findings, immunostaining of embryos with cubilin
antibodies revealed that cubilin is expressed by the visceral
extraembryonic endoderm but not the intraembryonic
endoderm (Fig. 7f). When RAP was coincubated with DiI-
HDL (Fig. 7c), compared with the control protein ovalbumin
(Fig. 7 a and b), there was a significant reduction (P ? 0.01)
in the incorporation of the DiI-HDL into extraembryonic
endoderm, as indicated by fluorescence intensity RANK scores
FIG. 4.
Bt2cAMP-treated F9 cells in the presence of competitors. (a and b)
CellswereincubatedwithapoE-freeDiI-HDL(1?g?ml)aloneorwith
unlabeledHDL,HDL2,HDL3,apoE-freeHDL2,andapoE-freeHDL3
(100 ?g?ml). (c and d) Cells were incubated with DiI-HDL alone or
with apolipoproteins apoA-I, apoA-II, apoC-I, apoC-III, apoJ, apoE,
or ovalbumin. (e and f) Cells were incubated with DiI-HDL alone or
with unlabeled HDL, and LDL, VLDL, ?VLDL, or Lp(a). (g and h)
Cells were incubated with DiI-HDL alone or with unlabeled HDL,
LDL, oxidized LDL, heparin, or heparinase. (a, c, e and g) Percentage
of cells that internalized DiI-HDL in the presence or absence of the
indicated agents. (b, d, f, and h) Mean fluorescence intensity values of
the cells that internalized DiI-HDL in the presence or absence of the
indicated agents. Cells (1 ? 104) were analyzed for each plotted value.
A mean of 24% of the cells in the RA?Bt2cAMP-treated F9 cell
cultures show measurable levels of internalized DiI-HDL. Data pre-
sented is representative of three experiments.
Flow cytometric analysis of DiI-HDL uptake by RA?
FIG. 5.
from [125I]-surface-labeled cells, whereas c–e were from unlabeled cells. (a) Coomassie-stained gel of proteins eluted from HDL-Sepharose by using
1, 4, and 8 M urea-containing buffer (lanes 3–8). Lanes 1 and 10 contain aliquots of detergent extracts from each cell type. Molecular weight
standards are in lanes 2 and 9. (b) Autoradiograph of the gel shown in a. (c) Immunoblot analysis of detergent extracts of RA?Bt2cAMP-treated
(lane 1) and untreated (lane 2) F9 cells by using megalin and cubilin antibodies. (d) Silver-stained SDS?PAGE profile of sequential fractions eluted
from HDL-Sepharose by using 8 M urea-containing buffer. Lane 1 contains molecular weight standards, lanes 2–5 correspond to HDL-Sepharose
eluted fractions from extracts of RA?Bt2cAMP-treated F9 cells, lane 6 is blank, and lanes 7–10 are HDL-Sepharose eluted fractions from untreated
F9 cells. (d) Anticubilin immunoblot analysis of the same fractions shown in c.
HDL-Sepharose affinity chromatography of extracts of RA?Bt2cAMP-treated and untreated F9 cells. Profiles shown in a and b were
Cell Biology: Hammad et al. Proc. Natl. Acad. Sci. USA 96 (1999) 10161

Page 5

9.77 ? 0.22 SD and 4.16 ? 0.15 SD, respectively. When cubilin
antibodieswerecoincubatedwithDiI-HDL(Fig.7e),therewas
also a significant decrease (P ? 0.01) in the incorporation of
the labeled HDL into extraembryonic endoderm compared
with coincubation of DiI-HDL with control IgG (Fig. 7d), as
indicated by RANK scores of 9.16 ? 0.41 SD and 3.83 ? 0.41
SD, respectively.
DISCUSSION
Here we report the identification of cubilin as a cell-surface
receptor that mediates endocytosis and lysosomal degradation
of holoparticle HDL. A key to this identification was the use
of an in vitro model of holoparticle HDL uptake that permitted
isolation of the receptor and characterization of salient aspects
of the holoparticle uptake process, including lipoprotein and
apolipoprotein specificity. The role of cubilin as a holoparticle
HDL endocytosis receptor was confirmed through the use of
cubilin antibodies to inhibit internalization of HDL labeled on
either its apolipoprotein or lipid constituents. The endocytic
activity of cubilin reported here is consistent with recent
studiesshowingcubilintobeanendocyticreceptorforintrinsic
factor-vitamin B12and Ig light chain (33, 36).
Given the lack of a membrane spanning element, cubilin is
considered to be a peripheral membrane protein (34). The
absence of a transmembrane and cytoplasmic domain in
cubilin raises the question of how it can mediate HDL endo-
cytosis. An answer could be that cubilin-mediated clearance
requires a coreceptor. Such is the case with the endocytosis of
urokinase type plasminogen activator (uPA), which involves its
phosphoinositol-linked receptor uPAR and coreceptors of the
LDL receptor family (37). Megalin is an obvious candidate for
a cubilin coreceptor because of its reported interaction with
cubilin (34). Whereas megalin antibodies capable of inhibiting
endocytic functions of megalin (23, 25, 26), including DiO-
LDL uptake (Fig. 6c), had little or no inhibitory effect on
DiI-HDL uptake (Fig. 6d), a megalin antiserum obtained from
Pierre Verroust inhibited HDL uptake by 25% (data not
shown). The variability in these results leaves open the possi-
bility that megalin plays a role as a coreceptor in cubilin-
mediated uptake of HDL. Moreover, the fact that RAP and
apoE are potent competitors of HDL uptake is consistent with
possible involvement of members of the LDL receptor family.
However, RAP has also been shown to bind cubilin (33). The
role of other minor polypeptides present in the HDL-
Sepharose chromatography elution profile must also be con-
sidered.
The extraembryonic visceral endoderm of the yolk sac is a
major site of cubilin expression (35). Studies have shown this
endoderm mediates uptake of maternal HDL (3, 4, 38);
however, the significance of maternal HDL to embryonic
development is not fully understood. The vital role that
yolk-sac cubilin plays in embryonic development is indicated
by the teratogenic effects observed when cubilin antibodies are
administered to pregnant rats (35). In light of our findings, we
speculate that the cubilin antibody-induced defects may be the
result of inhibition of cubilin-mediated transport of maternal
HDL by the extraembryonic visceral endoderm, perhaps lead-
ing to embryonic deficiency in HDL-associated cholesterol,
lipid-soluble vitamins (A, E, and K), phospholipids, and?or
triglycerides. Consistent with this hypothesis, developmental
abnormalities induced by the cubilin antibodies, including
neurodevelopmental?craniofacial defects, are similar to de-
velopmental defects caused by mutations in genes involved in
cholesterol biosynthesis (7-dehydrocholesterol-?7-reductase),
lipoprotein trafficking (apoB and megalin), and in the choles-
terol-modified signaling protein, sonic hedgehog (39).
The expression of cubilin along the apical surfaces of kidney
proximal tubule cells (35) highlights its potential to clear HDL
FIG. 6.
dation. (a–c) RA?Bt2cAMP-treated F9 cells were incubated with
apoE-free DiI-HDL alone or in the presence of unlabeled HDL, RAP,
anticubilin IgG, antimegalin?LRP-2, or control IgG. (d) Treated F9
cells were incubated with DiO-LDL alone or in the presence of
unlabeled LDL, HDL, RAP, antimegalin?LRP-2 IgG, or control IgG.
(a, c, and d) Percentage of cells that internalized DiI-HDL or
DiO-LDL in the presence or absence of the indicated agents. (b) Mean
fluorescence intensity values of the cells (in a) that internalized
DiI-HDL in the presence or absence of the indicated agents. Cells (1 ?
104) were analyzed for each value plotted in a–d. Data in a–d are
representative of three experiments. (e and f) Amounts of125I-HDL
internalized (e) and degraded (f) by RA?Bt2cAMP-treated F9 cells
incubated with apoE-free125I-HDL in assay medium alone or in assay
medium containing unlabeled HDL, RAP, chloroquine, BSA, anti-
cubilin IgG, or control IgG. Values in e and f are means ? SD of
triplicate values.
Cubilin antibodies inhibit HDL internalization and degra-
FIG. 7.
the extraembryonic visceral endoderm of mouse embryos. (a–e)
Confocal images of mouse embryos (8 days postcoitum) incubated for
1 h ex utero in DMEM containing DiI-HDL plus ovalbumin (a and b),
RAP (c), normal rabbit IgG (d), or rabbit anticubilin IgG (e). (f and
g) Mouse embryos stained with rabbit anticubilin IgG (f) or normal
rabbit IgG (g). i.e., intraembryonic endoderm; ee, extraembryonic
endoderm; ec, ectoplacental cone. (Bar ? 200 ?m.)
Cubilin antagonists inhibit incorporation of DiI-HDL into
10162Cell Biology: Hammad et al. Proc. Natl. Acad. Sci. USA 96 (1999)

Page 6

from the glomerular filtrate. The process of glomerular filtra-
tion of HDL, although acknowledged to take place, is poorly
characterized. Neutral substances with diameters ?8 nm and
anionic substances ?6 nm are not filtered by the glomerulus
(40). HDL particles range in diameter from 7.2–12.9 nm, with
HDL3 ranging from 7.2–8.8 nm and HDL2 ranging from
8.8–12.9 nm (41). Thus, HDL3is a likely HDL subclass for
cubilin to encounter in the postglomerular filtrate. In addition,
lipid-poor apoA-I, which represents as much as 8% of total
serum apoA-I (42), may also be cleared from the filtrate by
cubilin. Factors such as selective cholesterol uptake by SR-BI
and lypolysis act to reduce HDL particle size and lead to
accelerated HDL clearance presumably via the kidney (14, 43,
44). An expected consequence of renal clearance of HDL3or
lipid-poor apoA-I by cubilin is catabolism of these particles.
Catabolism is consistent with our results and with results from
nephron microperfusion studies showing that125I-HDL3 is
lysosomally degraded within proximal tubule cells (45). An-
other possible consequence could be recycling of HDL?HDL
apolipoproteins, perhaps through a process in which HDL
particles and?or apolipoprotein constituents bypass degrada-
tion and are delivered to the blood for reutilization. Such a
consequence may be related to our finding that chloroquine
treatment inhibited125I-HDL degradation yet did not lead to
intracellular accumulation of internalized HDL (Fig. 1a).
Typically, chloroquine treatment leads to intracellular accu-
mulation of ligands targeted for lysosomal degradation such as
LDL (25). It is therefore possible that HDL?HDL apolipopro-
teins can be trafficked via at least two pathways after cubilin-
mediated endocytosis, one that leads to the lysosomal degra-
dation and another that leads out of the cell (e.g., transcytosis).
This latter process may be similar to ‘‘retroendocytosis’’ of
HDL, in which internalization is followed by unloading of the
cholesteryl ester and retrograde secretion of the lipid-depleted
particle (46). It also may be similar to transintestinal epithe-
lium transport of IgG and A (47) and transcytosis of LDL
across the blood-brain barrier (48).
In summary, RA?Bt2cAMP-treatment of a mouse terato-
carcinoma cell line induces the formation of yolk-sac
endoderm-like cells capable of endocytosing and lysosomally
degrading HDL. HDL binding to these cells is saturable and
of high affinity. The receptor responsible for mediating the
binding and endocytosis of HDL was identified as cubilin.
Uptake of HDL by mouse embryo yolk-sac endoderm was
shown to be mediated by cubilin. The significance of cubilin-
mediated uptake of HDL to the processes of maternal-
embryonic lipoprotein transport and embryonic development
remains to be established, as does the potential of cubilin to
mediate renal uptake of plasma-derived HDL and conse-
quences of its activity on HDL homeostasis.
We thank Dr. Pierre Verroust for providing cubilin antibodies. This
work was supported by an American Heart Association Grant to
W.S.A. S.M.H. and W.O.T. are recipients of fellowships from the
American Heart Association.
1.
2.
3.
4.
5.
Eisenberg, S. (1984) J. Lipid Res. 25, 1017–1058.
Tall, A. R. (1990) J. Clin. Invest. 86, 379–384.
Woollett, L. A. (1996) J. Lipid Res. 37, 1246–1257.
Wyne, K. L. & Woollett, L. A. (1998) J. Lipid Res. 39, 518–530.
Glass, C., Pittman, R. C., Civen, M. & Steinberg, D. (1985) J. Biol.
Chem. 260, 744–750.
Horowitz, B. S., Goldberg, I. J., Merab, J., Vanni, T. M., Ramakrish-
nan, R. & Ginsberg, H. N. (1993) J. Clin. Invest. 91, 1743–1752.
Silver, D. L., Jiang, X. C. & Tall, A. R. (1999) J. Biol. Chem. 274,
4140–4146.
Le, N. A. & Ginsberg, H. N. (1988) Metabolism 37, 614–617.
6.
7.
8.
9.Brinton, E. A., Eisenberg, S. & Breslow, J. L. (1991) J. Clin. Invest. 87,
536–544.
Deeb, S. S., Cheung, M. C., Peng, R. L., Wolf, A. C., Stern, R., Albers,
J. J. & Knopp, R. H. (1991) J. Biol. Chem. 266, 13654–13660.
Rader, D. J., Castro, G., Zech, L. A., Fruchart, J. C. & Brewer, H. B.
(1991) J. Lipid Res. 32, 1849–1859.
Jackle, S., Rinninger, F., Lorenzen, T., Greten, H. & Windler, E.
(1993) Hepatology 17, 455–465.
Acton, S., Rigotti, A., Landschulz, K. T., Xu, S., Hobbs, H. H. &
Krieger, M. (1996) Science 271, 518–520.
Kozarsky, K. F., Donahee, M. H., Rigotti, A., Iqbal, S. N., Edelman,
E. R. & Krieger, M. (1997) Nature (London) 387, 414–417.
Rigotti, A., Trigatti, B. L., Penman, M., Rayburn, H., Herz, J. &
Krieger, M. (1997) Proc. Natl. Acad. Sci. USA 94, 12610–12615.
Steinberg, D. (1996) Science 271, 460–461.
Brewer, H. B., Ronan, R., Meng, M. & Bishop, C. (1986) Methods
Enzymol. 128, 223–246.
Jackson, R. L. & Holdsworth, G. (1986) Methods Enzymol. 128,
288–296.
Williams, S. E., Ashcom, J. D., Argraves, W. S. & Strickland, D. K.
(1992) J. Biol. Chem. 267, 9035–9040.
Osborne, J. C. (1986) Methods Enzymol. 128, 213–222.
Kelly, J. L. & Kruski, A. W. (1986) Methods Enzymol. 128, 170–180.
Oram, J. F. (1986) Methods Enzymol. 129, 645–659.
Hammad, S. M., Ranganathan, S., Loukinova, E., Twal, W. O. &
Argraves, W. S. (1997) J. Biol. Chem. 272, 18644–18649.
Kounnas, M. Z., Haudenschild, C. C., Strickland, D. K. & Argraves,
W. S. (1994) In Vivo 8, 343–351.
Stefansson, S., Chappell, D. A., Argraves, K. M., Strickland, D. K. &
Argraves, W. S. (1995) J. Biol. Chem. 270, 19417–19421.
Kounnas, M. Z., Loukinova, E. B., Stefansson, S., Harmony, J. A.,
Brewer, B. H., Strickland, D. K. & Argraves, W. S. (1995) J. Biol.
Chem. 270, 13070–13075.
Goldstein, J. L., Basu, S. K. & Brown, M., S. (1983) Methods Enzymol.
98, 241–260.
Goldstein, J. L. & Brown, M. S. (1974) J. Biol. Chem. 249, 5153–5162.
George,R.,Barber,D.L.&Schneider,W.J.(1987)J.Biol.Chem.262,
16838–16847.
Munson, P. J. & Rodbard, D. (1980) Anal. Biochem. 107, 220–239.
Ashcom, J. D., Tiller, S. E., Dickerson, K., Cravens, J. L., Argraves,
W. S. & Strickland, D. K. (1990) J. Cell Biol. 110, 1041–1048.
Zhong-Sheng, J., Dichek, H. L., Miranda, R. D. & Mahley, R. W.
(1997) J. Biol. Chem. 272, 31285–31292.
Birn, H., Verroust, P. J., Nexo, E., Hager, H., Jacobsen, C., Chris-
tensen, E. I. & Moestrup, S. K. (1997) J. Biol. Chem. 272, 26497–
26504.
Moestrup, S. K., Kozyraki, R., Kristiansen, M., Kaysen, J. H.,
Rasmussen, H. H., Brault, D., Pontillon, F., Goda, F. O., Christensen,
E. I., Hammond, T. G., et al. (1998) J. Biol. Chem. 273, 5235–5242.
Sahali, D., Mulliez, N., Chatelet, F., Dupuis, R., Ronco, P. &
Verroust, P. (1988) J. Exp. Med. 167, 213–218.
Batuman, V., Verroust, P. J., Navar, G. L., Kaysen, J. H., Goda, F. O.,
Campbell, W. C., Simon, E., Pontillon, F., Lyles, M., Bruno, J., et al.
(1998) Am. J. Physiol. 275, F246–54.
Strickland, D. K., Kounnas, M. Z. & Argraves, W. S. (1995) FASEB
J. 9, 890–898.
Hatzopoulos, A. K., Rigotti, A., Rosenberg, R. D. & Krieger, M.
(1998) J. Lipid Res. 39, 495–508.
Farese, R. V., Jr. & Herz, J. (1998) Trends Genet. 14, 115–120.
Bohrer, M. P., Baylis, C., Humes, H. D., Glassock, R. J., Robertson,
C. R. & Brenner, B. M. (1978) J. Clin. Invest. 61, 72–78.
Nichols, A. V., Krauss, R. M. & Musliner, T. A. (1986) Methods
Enzymol. 128, 417–431.
Neary, R. H. & Gowland, E. (1987) Clin. Chem. 33, 1163–1169.
Levak-Frank, S., Weinstock, P. H., Hayek, T., Verdery, R., Hofmann,
W., Ramakrishnan, R., Sattler, W., Breslow, J. L. & Zechner, R.
(1997) J. Biol. Chem. 272, 17182–17190.
Ueda, Y., Royer, L., Gong, E., Zhang, J., Cooper, P. N., Francone, O.
& Rubin, E. M. (1999) J. Biol. Chem. 274, 7165–7171.
Peterson, D. R., Hjelle, J. T., Carone, F. A. & Moore, P. A. (1984)
Kidney Int. 26, 411–421.
DeLamatre, J. G., Sarphie, T. G., Archibold, R. C. & Hornick, C. A.
(1990) J. Lipid Res. 31, 191–202.
Rodewald, R. (1973) J. Cell Biol. 58, 189–211.
Dehouck, B., Fenart, L., Dehouck, M. P., Pierce, A., Torpier, G. &
Cecchelli, R. (1997) J. Cell Biol. 138, 877–889.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
Cell Biology: Hammad et al.Proc. Natl. Acad. Sci. USA 96 (1999) 10163