| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Lipid Metabolism Unit (G.P.E., M.W.F.) and Cardiac Unit (A.J.M.), Massachusetts General Hospital, Boston, Massachusetts 02114
Address all correspondence and requests for reprints to: Dr. Mason W. Freeman, Lipid Metabolism Unit, Massachusetts General Hospital, Boston, Massachusetts 02114.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Both HDL and its major apolipoprotein component, apolipoprotein AI (apo AI), have attracted considerable scientific interest because of epidemiological studies that have demonstrated their inverse correlation with the risk of developing coronary artery disease (CAD) (1, 2, 3, 4, 5). The interactions between HDL and tissues necessary to achieve this apparent protective effect against CAD are as yet poorly understood. The study of genetic disorders affecting HDL metabolism would, therefore, be expected to clarify HDL’s role in the atherosclerotic process, perhaps by illuminating some of the mechanisms involved in the reverse cholesterol transport pathway. Many such disorders, involving defects in apo AI structure, have been identified (6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16). These generally result in low HDL cholesterol levels due to decreased production of apo AI and/or increased clearance of an abnormal apo AI protein. Investigations of these disorders have generally failed to shed much light on mechanisms of cellular interactions with HDL lipoproteins.
Unlike the syndromes just described, Tangier disease subjects have low HDL cholesterol levels without a defect in a HDL apolipoprotein. In Tangier disease, the protein structure and gene sequence of apo AI are normal, whereas HDL cholesterol and apo AI levels are low or absent (17, 18). Subjects with Tangier disease have an accumulation of cholesterol esters, mainly in reticuloendothelial cells throughout the body, leading to hyperplastic tonsils, splenomegaly, and neuropathy. Abnormalities in the cellular transport of cholesterol have been described in Tangier disease. Monocyte-derived macrophages and fibroblasts have been found to have decreased cholesterol efflux to acceptor HDL particles and an almost absent cholesterol efflux in response to apo AI (19, 20, 21). Studies differ on whether this is associated with an abnormality in binding of apo AI to the cellular membrane (20, 21). A specific defect causing the alterations in cholesterol transport in cells from patients with Tangier disease has yet to be identified, although a recent paper has identified abnormalities in the coordination of phospholipase C and D activation (22).
Familial HDL deficiencies without known apoprotein or lipoprotein processing enzyme defects, other than Tangier disease, have also been described (23, 24). Cellular cholesterol efflux has not been studied in these unexplained low HDL syndromes, although it has been postulated that such syndromes may be due to defects in cellular cholesterol transport (24). We report here a novel patient with an extremely low HDL cholesterol level who is without clinical Tangier disease and has a decrease in cellular cholesterol efflux to apo AI. Several other alterations in cellular lipid homeostasis also distinguish this patient’s cells from both normal as well as Tangier cells. Identification of genes responsible for the efflux abnormality identified in this patient should provide valuable insights into fundamental mechanisms by which cellular cholesterol is transferred to acceptor lipoproteins for removal.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The proband is a 51-yr-old Caucasian woman who first presented to the Massachusetts General Hospital with atypical chest pain at age 36 yr, when an echo cardiogram established a diagnosis of mitral valve prolapse. She presented to her cardiologist again at age 46 yr when her primary care doctor noted an extremely low HDL cholesterol and was referred to the Lipid Clinic for further evaluation. Laboratory work was significant for an HDL cholesterol level of 5 mg/dL.
After her medical evaluation, the patient was asked to provide blood and skin samples for evaluation of her disorder. All protocols involving the patient were approved by the Institutional Review Board of the Massachusetts General Hospital. Informed consent was obtained for the drawing of blood and performance of the skin biopsy.
Lipid analyses
Total cholesterol and triglycerides in serum were measured enzymatically (Sigma Diagnostics, St. Louis, MO) on a Technicon RA-500 (Technicon Instruments, Tarrytown, NY). HDL cholesterol was separated by precipitation with dextran sulfate and magnesium, then assayed as described above. Low density lipoprotein (LDL) cholesterol was separated by sequential ultracentrifugation (25). Apo AI and apo B were measured by a turbidometric immunoassay (Sigma Diagnostics) on a Technicon RA-500. Lipid values for first degree relatives were determined in the course of previous clinical care and transmitted to the investigators.
Lipoprotein and apolipoprotein isolation
LDL, HDL2, and HDL3 were prepared by sequential ultracentrifugation of normal plasma in the density intervals 1.019–1.063, 1.063–1.125, and 1.125–1.21 g/mL, respectively, using standard methods (25). Apo A-I was isolated from delipidated HDL2 as previously described (26). Protein was quantitated by the method of Lowry et al. (27). Trypsin-modified HDL was prepared by incubating HDL with a 1:40 dilution of (wt/wt) trypsin (Life Technologies, Grand Island, NY) at 37 C for 30 min. Phenylmethylsulfonylfluoride (Sigma Chemical Co., St. Louis, MO) was added to a final concentration of 1 mmol/L, and the solution was passed through a Sephadex G-75 (Pharmacia Biotech, Uppsala, Sweden) column (28). The phospholipid content was measured with a commercial kit (Wako Chemicals, Neuss, Germany).
Electrophoretic procedures
SDS-PAGE was performed by the method of Laemmli (29). Various dilutions of plasma from the proband and from a normal control were electrophoresed through a 15% gel or through a 10–20% gradient gel (Z-axis, Hudson, OH), using 25 ng purified apo AI as a standard control. Gels were electrophoretically transferred to nitrocellulose (30), and the membrane was blocked for 2 h in Tris-buffered saline (TBS) with 5% nonfat dried milk and 0.1% Tween at room temperature. The membrane was washed in TBS-0.1% Tween, then incubated for 3 h at room temperature with a monoclonal antibody to apo AI (Cappel, Durham, NC) diluted 1:200 in TBS, 0.1% Tween, and 5% dried milk. After washing, the membrane was incubated at room temperature with a goat antimouse peroxidase conjugate (Sigma Chemical Co.) diluted 1:1000 in TBS, 0.1% Tween, and 5% dried milk for 90 min. The membrane was then visualized with the enhanced chemiluminescence detection kit (ECL, Amersham Life Sciences, Arlington Heights, IL). Fifteen percent SDS-PAGE was performed on aliquots of fractions obtained from passing the patient’s plasma over a Superose 6 column (see below). Nondenaturing Tris-borate-ethylenediamine tetraacetate gradient 10–20% gel (Z-axis) electrophoresis of plasma from the proband and from a normal control and of purified HDL3 was performed and then visualized by immunoblotting using the conditions described above. Isoelectric focusing of delipidated plasma was conducted using gels with 8% polyacrylamide, 0.27% bis-acrylamide, and 2% ampholytes (Bio-Rad, Hercules, CA) in the pH range 4–6 (31). The gel proteins were transferred to nitrocellulose in 0.7% acetic acid and subjected to immunoblotting as described above.
Column chromatography
A 200-µL aliquot of plasma from the patient and from a normal control were chromatographically separated on a Superose 6B fine pressure liquid chromatography column (HR 10/30, Pharmacia LKB, Uppsala, Sweden) using a Pharmacia LKB fine pressure liquid chromatograph with a flow rate of 0.5 mL/min. The first 6 mL were discarded, and the remainder was collected in 1-mL aliquots. The eluate was monitored using absorbance at 280 nmol/L. Two hundred microliters from each fraction were used to measure free and total cholesterol enzymatically by the cholesterol oxidase method (32), with and without cholesterol esterase, respectively. Two microliters of each fraction were subjected to SDS-PAGE on a 15% gel followed by immunoblotting as described above.
PCR amplification
Oligonucleotide primers were designed to amplify portions of the apo AI gene spanning the coding sequences and intron-exon splice junctions. Primers also included EcoRI and BamHI overhangs. Genomic DNA was prepared from the proband and amplified with Taq polymerase (Boehringer Mannheim, Indianapolis, IN) in a reaction volume of 50 µL with 1–2 µg DNA, 0.2 mmol/L deoxy-NTPs, 1.5 mmol/L MgCl, and 10 pmol primers. Cycling was performed on a Perkin-Elmer/Cetus (Norwalk, CT) cycler for 35 cycles employing the following parameters: 95 C for 1 min, 60 C for 45 s, and 72 C for 30 s.
DNA sequencing
PCR products derived from amplifications using the proband’s DNA as a template and the reaction conditions described above were digested with EcoRI and BamHI (New England Biolabs, Beverly, MA), gel purified, and ligated into pUC 18. Subclones were sequenced using universal primers to sequences in pUC 18 and nested internal primers, using T7 polymerase (U.S. Biochemical Corp., Cleveland, OH) according to the manufacturer’s protocol. PCR products were also sequenced directly. Asymmetric PCR amplification was carried out with a single primer in a 100-µL volume, and 70 µL of the reaction volume were spun through a Centricon 100 (Amicon, Danvers, MA) three times. The single stranded PCR product was sequenced with T7 polymerase (U.S. Biochemical Corp.) using a modification of the manufacturer’s directions as follows. Primers were used at 50 pmol, G-label mix was used at a 1:1000 dilution, and deoxy-NTP mixes were diluted in LiCl (Boehringer Mannheim).
Establishing fibroblast cell cultures
Cultured fibroblasts were initiated by explant cultures from a 3-mm punch biopsy at a 1-mm skin thickness obtained from the proband under sterile conditions from the medial aspect of the inner thigh and placed in Hanks’ Balanced Salt Solution without bicarbonate. The tissue was transferred to a 100-mm sterile plastic petri dish and diced. Using a forceps, pairs of moist explants were transferred to 25-cm2 flasks in DMEM with 15% FBS, 50 U/mL penicillin, and 50 µg/mL streptomycin. Explant culture was performed in the Cytogenetics Laboratory of the Brigham and Women’s Hospital. Primary fibroblast cell lines from two normolipidemic controls from our laboratory and from a patient with Tangier disease (gift from Dr. John F. Oram, University of Washington, Seattle, WA; these cells are characterized in Ref. 21 and referred to in that report as TG2 cells) were also maintained for comparison. All cells were used between the 5th and 12th passages.
Cholesterol loading and labeling of cells
Fibroblasts were seeded into culture dishes and grown to
60–80% confluence in DMEM with 10% FBS. Cells were labeled by
including 0.2–0.5 µCi [3H]cholesterol/mL (New England
Nuclear, Boston, MA) in the medium and incubating until confluent
(
72 h). Labeled cells were cholesterol enriched by incubation in
DMEM with 2 mg/mL fatty acid-free BSA (DMEM-FA-BSA) and the indicated
concentrations of nonlipoprotein cholesterol from an ethanol stock
solution for 24 h. Cells were incubated for 48 h in DMEM with
1 mg/mL FA-BSA to allow cellular cholesterol pools to equilibrate (33).
Other labeled cells were enriched with cholesterol by incubating the
cells for 48 h in DMEM with lipoprotein-deficient FBS to which 100
µg/mL LDL protein had been added. Cells were rinsed and incubated
overnight in DMEM-FA-BSA (1 mg/mL).
Measuring cellular cholesterol efflux
The efflux of radiolabeled cholesterol from cells was measured as previously described (34). Briefly, cholesterol-loaded cells were incubated in DMEM-FA-BSA with increasing levels of apo AI for 16 h at 37 C or with increasing levels of either HDL or trypsin-modified HDL for 6 h at 37 C. After incubation, efflux medium was removed and centrifuged to remove cells and debris, and an aliquot of medium was removed for measurement of radioactivity. After washing with PBS, cell lipids were extracted from culture dishes with hexane-isopropanol (3:2, vol/vol), then evaporated to dryness under nitrogen gas, reconstituted in chloroform, and subjected to TLC. Cell proteins were dissolved in 0.1 N NaOH, and aliquots were quantified by the method of Lowry et al. (27).
TLC of the cellular extracts were performed on silica gel plates (Whatman International, Maidstone, UK) developed in hexane-ether-acetic acid (130:40:1.5, vol/vol/vol). Lipid spots were visualized by staining with I2 vapor and identified by their comigration with standards. Cholesterol mass was determined by scraping lipid spots and extracting them with CHCl3-CH3OH (2:1). Cholesterol esters were saponified with ethanolic 1 mol/L KOH for 1 h at 80 C before extraction. Extracts were evaporated, reconstituted in isopropanol, and assayed using the cholesterol oxidase method.
Measuring esterification of cellular cholesterol
Esterification of cellular cholesterol was measured as incorporation of [14C]oleate (New England Nuclear) into cellular cholesterol ester (34). After incubations with experimental medium as described above, cells were rinsed once with PBS, then incubated for 1 h with DMEM containing 9 µmol/L [14C]oleate and 3 µmol/L BSA at 37 C. Cells were washed and extracted with hexane-isopropanol. Cell proteins were solubilized in 0.1 N NaOH and quantified by the method of Lowry. Cell lipids were separated by TLC on silica G plates developed in hexane-ether-acetic acid, and individual spots were taken for counting. Cholesterol esterification is expressed as [14C]oleate incorporated in [14C]esters per mg cell protein.
Sterol biosynthesis
Sterol biosynthesis was measured by incubating fibroblasts in DMEM with 10% lipid-deficient FBS for 48 h, then incubating for 2 h with DMEM-FA-BSA with 2 µCi [14C]acetate/mL (New England Nuclear) at 37 C. Cells were then extracted. Aliquots of lipids were evaporated to dryness and saponified in 1 mL 1 mol/L KOH in 80% ethanol for 1 h. Nonsaponified lipids were reextracted with 1.5 mL water and 5 mL hexane, and aliquots were taken for TLC and quantitation of radioactivity.
Phospholipid biosynthesis
The incorporation of [3H]choline into phosphatidylcholine was measured by incubating confluent noncholesterol-enriched fibroblasts for 4 h in DMEM containing 1 mg/mL BSA and 2 µCi/mL [3H]choline (New England Nuclear). After incubation, cell layers were chilled, then washed five times with PBS, extracted with isopropanol for 18 h, and subjected to TLC, as previously described, to isolate phosphatidylcholine (33).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The patient is a 51-yr-old Caucasian woman who was referred for an
abnormal lipid profile. Past medical history at that time was
significant for treated hypertension diagnosed when the patient was in
her mid-20s, hypothyroidism, and mitral valve prolapse. Medications at
presentation to the Lipid Clinic included estrogen and thyroid
replacements. On physical exam, blood pressure was 160/90 mm Hg. No
xanthoma, xanthelasma, or corneal clouding were noted. Liver and spleen
were of normal size by palpation. Tonsils were present and not
hypertrophied or orange. An exercise stress test to stage 4 Bruce
protocol performed at age 47 yr was negative. A slit lamp eye exam
performed at age 49 yr was normal. Abdominal ultrasound at age 51 yr
showed that the liver and spleen were not enlarged. Laboratory work was
significant for a total cholesterol level of 161 mg/dL, a HDL
cholesterol level of 5 mg/dL, a LDL cholesterol level of 98
mg/dL, and a triglyceride level of 390 mg/dL (see Table 1
). The apo AI level was 11 mg/dL. The
apo B level was 111 mg/dL.
|
). The patient has two brothers, neither of whom
has a history of CAD. The oldest brother, age 46 yr, has a HDL
cholesterol value of 24 mg/dL. The second brother, age 40 yr, had onset
of hypertension in his 20s. He has a HDL cholesterol level of 39 mg/dL.
The patient does not have any children. DNA analyses
A restriction digest of the patient’s genomic DNA with PvuII, EcoRI, PstI, and SalI was probed with a PstI fragment that included all of the coding sequences of the apo AI gene. This showed restriction fragments of the expected length (data not shown). We were able to amplify all the exons of the proband’s apo AI gene and found products of the expected length (data not shown). Exons 2, 3, and 4 of the apo AI gene were amplified across the coding sequences and splice junctions, and ligated into pUC 18. Two clones from each of these amplifications were sequenced. Sequences obtained from all clones matched that of the published sequence for normal apo AI. To confirm that the individual PCR subclones were representative of the population of amplified products, we chose also to sequence the PCR products directly, without subcloning. The direct sequence obtained was normal for all primer pairs.
Apo AI protein and HDL particle studies
Protein bands separated by SDS-PAGE of whole plasma on both 15%
and 10–20% gradient gels were visualized by Western blotting with an
anti-apo AI monoclonal antibody. These indicated that the proband’s
apo AI migrated similarly to the apo AI in plasma taken from a normal
control and to apo AI purified from pooled plasma. These findings
suggested that the proband’s apo A1 was of normal size (Fig. 1a). However, the SDS-PAGE gels showed a
significantly reduced level of apo AI, at about 10% of normal.
Delipidated whole plasma from the proband was compared to that from a
normal control by analysis on an isoelectric focusing gel, pH range
4–6 (Fig. 1b). Pro-apo AI and the two major mature isoforms of apo AI
were present in both the proband and control plasma and migrated
similarly. Qualitatively, the ratio of pro-apo AI to mature apo AI was
greatly increased in the proband compared to normal. These data suggest
that the proband makes a mature apo AI that is grossly of normal size
and charge, but present in significantly reduced quantities. Pro-apo AI
appears also to be of normal charge, but it is present in amounts
similar to that found in the normal control.
|
) in a pattern similar to that seen in
the HDL3 (data not shown). HDL particle size was also
investigated using a Superose 6B column. In the patient, cholesterol
was present in fractions correlating with the elution positions of very
low density lipoprotein and LDL. However, much less cholesterol was
present in fractions corresponding to the elution position of HDL. The
HDL cholesterol eluted off the column later than that from normal
control plasma (Fig. 2b). The percentage of free cholesterol was
measured in each of the lipoprotein subfractions. The percentage of
free cholesterol overall was 29%, and it did not differ substantially
among the VLDL, LDL, and HDL subfractions. Immunoblotting of the
fractions with an anti-apo AI antibody demonstrated the presence of apo
AI only in the proband’s fractions corresponding to HDL (Fig. 2c).
These results indicate that the proband’s HDL is primarily present as
a smaller, denser HDL particle than that seen in normal controls.
|
Cellular cholesterol efflux and metabolism
The ability of apo AI to promote cellular
[3H]cholesterol efflux was examined using primary skin
fibroblasts from the proband, two normal controls, and a patient with
Tangier disease (Fig. 3a). For these
studies, cells were enriched with cholesterol by incubation with
nonlipoprotein cholesterol and then incubated with either albumin alone
or increasing levels of apo AI. The percentage of
[3H]cholesterol appearing in the medium after incubation
with apo AI is a measure of the total cellular cholesterol efflux to
apo AI. Both the normal cell lines and the proband’s cell line were
able to support [3H]cholesterol efflux in a
dose-responsive manner. However, maximal efflux of cholesterol from the
proband’s cells was reduced by nearly 50% (8 ± 0.2% for
proband vs. 14 ± 0.3% for combined controls;
P = 0.00007). This decrease in efflux was not as
profound as that seen in the Tangier cells, which demonstrated nearly
absent apolipoprotein-stimulated [3H]cholesterol efflux,
as previously reported (21).
|
, b and c). The cellular cholesterol ester content
decreased by an average of 25 ± 6% of baseline in the control
cells lines and by 19 ± 6% of baseline in cells from the proband
(P = 0.178). The unesterified cholesterol content
decreased an average of 8 ± 3% from the baseline in the control
cell lines and 4 ± 1% from baseline in the proband’s cells
(P = 0.075). Tangier cells did not measurably decrease
their cellular esterified or unesterified cholesterol content after
incubation with apo AI. Although the data did not reach statistical
significance, the proband’s cells appeared to have a diminished
ability to promote efflux of both free and esterified cholesterol,
rather than a selective inability affecting just one cholesterol
pool.
After cholesterol enrichment, cells from the proband contained a lesser
percentage of [3H]cholesterol found in esterified
cholesterol (17 ± 1%) compared to an average of 24 ± 1.6%
for the two control cell lines (P = 0.0003) and 36
± 1.4% for the Tangier cell line (P = 0.0003; Fig. 3b). Differences in cholesterol esterification rates were seen in a
separate experiment in which cells from one control and from the
proband were loaded with varying levels of either free cholesterol or
LDL, and cholesterol esterification was measured by the incorporation
of [14C]oleate into cellular cholesterol esters.
Growth-arrested cells from the proband, which were not enriched with
cholesterol, esterified 162 ± 29 pmol oleate/mg cellular protein,
which was lower than that in the controls (579 ± 26;
P = 0.00048; Fig. 4a).
This difference persisted after loading with varying levels of LDL, and
after loading with 5 µg/mL free cholesterol, but was not apparent
after incubation with 10, 20, or 30 µg/mL free cholesterol (Fig. 4b).
The attainment of maximal esterification rates when cells were enriched
with nonlipoprotein cholesterol suggests that the difference in
esterification rates is not due to a defect in acyl cholesterol acyl
transferase (ACAT), but, rather, may be due to differences in substrate
provision. The data suggest that more free cholesterol is available in
an ACAT accessible pool after loading with nonlipoprotein cholesterol
than after loading with LDL.
|
). Efflux
from the proband’s fibroblasts was compared to that from one of the
normal control cell lines. In cells that had not been cholesterol
enriched, the proband’s cells had a lower percentage of
3H-labeled esterified cholesterol (1.2 ± 0.3%
compared to 3.2 ± 0.3% in the control cell line). In cells
loaded with LDL, this difference was again apparent (15.4 ± 0.7%
in the proband vs. 21.0 ± 0.5% in the control cell line).
However, after loading with free cholesterol, the proband’s cells did
not differ in [3H]cholesterol ester content from the
normal cell line. Regardless of the degree of cholesterol ester
accumulation in the cells, under all conditions tested, the proband’s
efflux of [3H]cholesterol remained approximately half
that found in control cells (Fig. 6).
These data suggest that the decrease in efflux was present regardless
of the loading conditions.
|
|
).
The absolute difference in cholesterol efflux to nonmodified HDL over
trypsin-modified HDL in the proband (4.1 ± 1%) was approximately
half that seen in the control cells (P = 0.0009). These
data are consistent with our prior observations showing that efflux to
apo AI is decreased in the proband to about half the control values. In
the Tangier’s cells, efflux to nonmodified HDL increased minimally
over trypsin-modified HDL, consistent with the work of Francis et
al. (21). Overall, efflux of cholesterol to nonmodified HDL was
significantly greater from the normal cells than from the proband’s
cells (45% decrease) and the Tangier’s cells (51% lower than control
values). Cholesterol efflux to trypsin-modified HDL was more similar
for all cell lines, and in a separate study, cholesterol efflux to
trypsin-modified HDL in the proband’s cells fell within the range of
normal established by two control cell lines (data not shown). These
results and those examining efflux to purified apo A-I are consistent
with the interpretation that apolipoprotein-mediated efflux is impaired
in the proband’s cells.
|
). Sterol synthesis was
down-regulated appropriately in all three cell lines after incubation
with LDL (data not shown). The cellular content of esterified and
unesterified cholesterol in the proband’s cells was intermediate to
that seen in the two control cell lines (Fig. 8a). Cholesterol
esterification rates based on the incorporation of labeled oleate were
measured in fibroblasts that were cholesterol enriched by incubation
with 15 µg/mL free cholesterol. Esterification rates in the
proband’s cells fell within the boundaries established by the two
normal control cell lines (Fig. 8a) and down-regulated appropriately
after exposure to apo AI (data not shown).
|
). The amount of
phosphatidylcholine synthesized in cells from the proband (94,093
± 12,545 cpm [3H]choline/mg cell protein) was
significantly greater than that measured in both control (66,761
± 4,045 and 44,868 ± 5,233 cpm) and Tangier cell lines
(45,962 ± 5,160 cpm; P < 0.005 for all). |
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In addition to an atypical lipid profile, this proband has none of the classic phenotypic findings of Tangier disease, such as hypertrophied orange tonsils or history of tonsillectomy, hepato-splenomegaly, or peripheral neuropathy. The proband also lacks clinical findings that would establish a diagnosis of LCAT deficiency, either of fish eye disease (massive corneal clouding) or of classical LCAT deficiency (anemia, renal failure, corneal clouding). These clinical findings in concert with a normal percentage of esterified cholesterol in plasma effectively exclude any known abnormality involving the LCAT enzyme. The patient thus presents with a clinical syndrome whose features do not appear to correspond to any other previously characterized low HDL syndrome. Data from the family suggest that the low HDL trait might be heritable, with one brother having an approximately half-normal HDL cholesterol value.
Several probands with initially unexplained low HDL syndromes subsequently have been found to have an absent or defective apo AI gene. Two such families had gene defects involving both the apo A-I and apo CIII genes (6, 7), whereas others have had asynthetic mutations effecting the apo A-I gene alone (8, 9, 10). Many structural variants of apo A-I have also been described. Only a few of these of these have been associated with altered lipid profiles (11, 12, 13, 14, 15, 16). Two of these, apo AI Milano and apo AI Iowa, have been shown to result in hypercatabolism of apo AI in vivo (11, 12, 13).
The DNA and protein data in our proband do not suggest a defect in apo AI as the cause of her low HDL cholesterol level. Southern blotting and agarose gel electrophoresis showed apo AI restriction fragments and PCR products, respectively, to be of expected lengths, ruling out a gross gene deletion. Sequencing of two subclones for each of four PCR primer pairs spanning the coding regions and splice junctions of the apo AI gene showed her sequence to be the wild type. Direct sequencing of PCR products derived from amplification of the AI gene also failed to reveal any alteration in the proband’s sequence. The DNA data are corroborated by SDS-PAGE 15% and 10–20% gels showing the proband’s apo AI to be of a normal size, although decreased to about 10% of the amount found in normal plasma. An isoelectric focusing gel showed the pro-apo AI and the two major mature isoforms of apo AI to be of normal charge. The ratio of pro-apo AI to mature apo AI in plasma appeared strikingly increased in the proband. The absolute amount of pro-apo AI did not appear decreased compared to that found in normal plasma. This increase in the ratio of pro-apo AI to mature apo AI is similar to that reported in diseases in which the fractional catabolic rate of apo AI is increased, such as Tangier disease (23).
Increased clearance of a normal apo AI in Tangier disease is associated with abnormalities in cellular cholesterol metabolism. The efflux of cholesterol from monocyte-derived macrophages and primary skin fibroblasts is decreased in response to incubation with HDL or, more dramatically, with apo AI (19, 20, 21). Francis et al. attributed this to an abnormal binding of apo AI to high affinity binding sites (21), whereas Rogler et al. found no abnormalities in the binding of HDL (20). The latter investigators attributed the cholesterol efflux defect to impaired activation of protein kinase C, as it was reversed by incubation with a membrane-permeable activator of protein kinase C. Walter et al. recently described a dysregulation of phosphatidylcholine-specific phospholipases C and D in Tangier fibroblasts. In these cells, the HDL-induced formation of phosphatidic acid by phospholipase D was greatly reduced, whereas the formation of diacylglycerol by phospholipase C was enhanced (22). Increasing the levels of phosphatidic acid pharmacologically could overcome the efflux defect in the Tangier’s cells. The researchers speculate that the molecular cause of the efflux defect is in an upstream effector responsible for the G protein-dependent regulation of these phospholipases. The increased clearance of apo AI in Tangier’s disease may be related to this inability of cells to donate lipid to apo AI, as lipid-poor HDL is thought to be more rapidly cleared by the kidney (37).
The preponderance of small dense apo AI-containing particles seen on nondenaturing PAGE and Superose 6B column chromatography of our proband’s serum suggested that a defect in the removal of cellular cholesterol to apo AI might be present. Such a defect has been postulated to be a potential cause of unexplained low HDL syndromes (24), but has not previously been demonstrated in cells other than those from patients with classic Tangier disease. The data presented in this report demonstrate that the proband’s fibroblasts do indeed have an abnormality in cellular cholesterol metabolism, measured as a decrease in apo AI-mediated cholesterol efflux.
Efflux of cholesterol in response to apo AI was reduced by approximately 50% in primary skin fibroblasts from the proband compared to that in fibroblasts from two normal controls. This defect was seen when cells were not cholesterol enriched and when cells were cholesterol enriched by either lysosomal or nonlysosomal pathways. Efflux of cholesterol to HDL was also diminished compared to that in both normal controls, but the cholesterol efflux to trypsin-modified HDL from the proband’s cells fell within the range of normal established by the two control cell lines. These data suggest that the efflux defect is due to abnormalities in specific, apo AI-mediated events rather than to a defect in diffusional, apolipoprotein-independent efflux.
The fibroblasts from the proband also exhibited a decrease in partitioning of labeled cholesterol to esterified cholesterol pools. This is in contradistinction to the Tangier fibroblasts, which had increased appearance of labeled cholesterol in esterified cellular cholesterol pools compared to both normal controls. This difference may explain the divergent clinical phenotypes, with Tangier patients having deposition of cholesterol esters in tissues throughout the body leading to the tonsillar hypertrophy, hepato-splenomegaly, and neuropathy not seen in our proband. The reason for the proband’s decreased partitioning of labeled cholesterol to esterified pools is not clear. The normal rate of incorporation of labeled oleate in the proband’s cells suggests that this is not a defect in ACAT activity, but, rather, may represent differences in ACAT-accessible cholesterol pools or in cellular cholesterol distribution. The absolute mass of esterified and unesterified cholesterol in the proband’s cells, however, did not differ from that in the two controls.
Cholesterol synthesis in the proband’s fibroblasts was not different from that in the two control cell lines; phosphatidylcholine synthesis in the proband’s cells was increased compared to that in both two control fibroblast cell lines and the Tangier cell line. Phosphatidylcholine synthesis in the Tangier cell line was similar to the lower level in the two control cell lines. This accords with previous studies by Francis et al. (21) showing decreased phosphatidylcholine synthesis in this Tangier cell line. It has been suggested that apolipoprotein-mediated cholesterol efflux is tightly coupled to phospholipid efflux, resulting in the formation of pre-ß HDL (38). Altered phospholipid metabolism in the proband could also be linked to the cholesterol efflux defect by altering membrane composition (39), which could affect cellular cholesterol transport or distribution, or by an effect on the coordinate regulation of the phosphatidylcholine-specific lipases. Alternatively, the apparent increase in phosphatidylcholine synthesis could itself be a secondary phenomenon resulting from such a dysregulation. It has been shown that macrophages increase their phosphatidylcholine synthesis in response to cholesterol enrichment (40), raising the possibility that increased phosphatidylcholine synthesis could be a response to a sensed increase in intracellular cholesterol. This observation may be a fruitful one for further study.
This case of an unexplained low serum HDL cholesterol level appears to be associated with a defect in cellular cholesterol efflux in response to apo AI. Several lines of evidence provided in this report suggest that this defect differs from that present in patients with Tangier disease. This proband’s clinical phenotype differs from that typical in either homozygous or heterozygous Tangier disease. Although the half-normal cholesterol efflux value is consistent with efflux values seen in cells from Tangier disease obligate heterozygotes (Eberhart, G., and M. Freeman, unpublished observation), the patient’s HDL cholesterol level is far below any described for a Tangier heterozygote. In addition, differences in phosphatidylcholine synthesis and cholesterol partitioning distinguish the patient’s cells from Tangier patient’s cells, further arguing that the patient does not represent a case of heterozygous Tangier disease. It is possible, of course, that different mutations affecting the same genetic locus could give rise to different cellular biochemical and clinical phenotypes. As the gene responsible for Tangier disease has not been identified, we were unable to investigate that possibility. Studies of other Tangier patients’ cells and those of obligate heterozygotes are currently underway in our laboratory to further clarify similarities and differences in their handling of cellular lipids.
The discovery of a cholesterol efflux defect in a proband without apparent Tangier disease raises the possibility that others of the unexplained low HDL syndromes encountered in clinical practice may be associated with a defect in cholesterol efflux. In addition, differences in cellular cholesterol efflux could underlie some of the unexplained population variability in HDL cholesterol levels, as it appears that only a small amount of the observed variability in HDL cholesterol levels in the general population can currently be accounted for by known genetic factors (41, 42, 43, 44, 45). Investigation of the molecular genetic basis for cholesterol efflux-deficient phenotypes seen in rare individuals, such as the patient in this report, may help elucidate mechanisms responsible for the more common low HDL cholesterol syndromes, as well as provide a rationale for newer therapeutic modalities aimed at the prevention of coronary artery disease.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received July 28, 1997.
Revised November 20, 1997.
Accepted December 1, 1997.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  R. S. Kiss, N. Kavaslar, K.-i. Okuhira, M. W. Freeman, S. Walter, R. W. Milne, R. McPherson, and Y. L. Marcel Genetic Etiology of Isolated Low HDL Syndrome: Incidence and Heterogeneity of Efflux Defects Arterioscler Thromb Vasc Biol, May 1, 2007; 27(5): 1139 - 1145. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Haidar, R. S. Kiss, L. Sarov-Blat, R. Brunet, C. Harder, R. McPherson, and Y. L. Marcel Cathepsin D, a Lysosomal Protease, Regulates ABCA1-mediated Lipid Efflux J. Biol. Chem., December 29, 2006; 281(52): 39971 - 39981. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Haidar, S. Mott, B. Boucher, C. Y. Lee, M. Marcil, and J. Genest , Jr. Cellular cholesterol efflux is modulated by phospholipid-derived signaling molecules in familial HDL deficiency/Tangier disease fibroblasts J. Lipid Res., February 1, 2001; 42(2): 249 - 257. [Abstract] [Full Text]
|
|
|
|
|
|
M. E. Brousseau, G. P. Eberhart, J. Dupuis, B. F. Asztalos, A. L. Goldkamp, E. J. Schaefer, and M. W. Freeman Cellular cholesterol efflux in heterozygotes for Tangier disease is markedly reduced and correlates with high density lipoprotein cholesterol concentration and particle size J. Lipid Res., July 1, 2000; 41(7): 1125 - 1135. [Abstract] [Full Text]
|
|
|
|
|
|
M. E. Brousseau, E. J. Schaefer, J. Dupuis, B. Eustace, P. Van Eerdewegh, A. L. Goldkamp, L. M. Thurston, M. G. FitzGerald, D. Yasek-McKenna, G. O'Neill, et al. Novel mutations in the gene encoding ATP-binding cassette 1 in four Tangier disease kindreds J. Lipid Res., March 1, 2000; 41(3): 433 - 441. [Abstract] [Full Text]
|
|
|
|
 |
|
 M. W. Freeman Effluxed lipids: Tangier Island's latest export PNAS, September 28, 1999; 96(20): 10950 - 10952. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
