Lipoprotein profiles from cord blood were obtained after sucrose density
gradient ultracentrifugation. The density scale on the x-axis is not a linear
scale because it is determined through mathematical transformation of locations
along the vertical axis of the centrifuge tube (0-35 mm) based on sampled
measured densities after completion of the spin. The density range used to
define VLDL, LDL, and HDL by this method is given by the width of the braces
under the x-axes. HDL indicates high-density lipoprotein; LDL, low-density
lipoprotein; and VLDL, very low-density lipoprotein.
Apo indicates apolipoprotein; MALDI-TOF, matrix-assisted laser desorption/ionization
time-of-flight; and HDL, high-density lipoprotein. Apo C-I–enriched
HDL (A and B) and normal HDL (C and D) were isolated from plasma of a group
0 infant (A and C) and a group 3 infant (B and D) and prepared for MALDI-TOF
mass spectrometry (see “Methods” section).
Plasma lipoproteins were isolated from 4 group 3 infants, 1 group 2
infant, and 3 group 0 infants by ultracentrifugation at density greater than
1.21 g/mL and prepared for electrophoresis. Following electrophoresis, gels
were stained for protein and densitometric scans performed at 596 nm. The
relative amount of the high-density lipoprotein (HDL) subclasses, using an
internal standard, is depicted on the y-axis. Numbers on the curves indicate
HDL particle diameter in nanometers.
The median and 25th and 75th percentiles (box) and 5th and 95th percentiles
(whiskers) for gestational age are shown. The circles represent outliers.
Plots of gestational age vs large (L3) low-density lipoprotein (LDL)
cholesterol and largest (H5) high-density lipoprotein (HDL) cholesterol in
group 3 (elevated; n = 30) and group 0 (undetectable; n = 27) infants. Regression
lines are depicted for each group. For L3 vs age the regression coefficient r = 0.61 for group 3 infants includes the 1 outlier (28-week
old infant with LDL3 of 73 mg/dL). If the outlier is excluded, the regression
coefficient (r = 0.52) is still significant.
For group 0 infants the regression coefficient was r =
0.22. For H5 vs age, the regression coefficients were r = 0.16 for group 3 infants and r = 0.01 for
group 0 infants. To convert from mg/dL to mmol/L, divide by 38.7.
Kwiterovich PO, Cockrill SL, Virgil DG, Garrett ES, Otvos J, Knight-Gibson C, Alaupovic P, Forte T, Zhang L, Farwig ZN, Macfarlane RD. A Large High-Density Lipoprotein Enriched in Apolipoprotein C-IA Novel Biochemical Marker in Infants of Lower Birth Weight and
Younger Gestational Age. JAMA. 2005;293(15):1891-1899. doi:10.1001/jama.293.15.1891
Author Affiliations: The Johns Hopkins Medical
Institutions, Baltimore, Md (Drs Kwiterovich, Garrett, and Zhang and Ms Virgil);
Laboratory for Cardiovascular Chemistry, Texas A & M University, College
Station (Drs Cockrill, Farwig, and Macfarlane); LipoScience Inc, Raleigh,
NC (Dr Otvos); Oklahoma Medical Research Foundation, Oklahoma City (Dr Alaupovic
and Ms Knight-Gibson); and Donner Laboratory, University of California at
Berkeley (Dr Forte).
Context Low birth weight is associated with increased cardiovascular disease
in adulthood, and differences in the molecular weight, composition, and quantity
of lipoprotein subclasses are associated with coronary artery disease.
Objective To determine if there are novel patterns of lipoprotein heterogeneity
in low-birth-weight infants.
Design, Setting, and Participants Prospective study at a US medical center of a representative sample
of infants (n = 163; 70 white and 93 black) born at 28 or more weeks
of gestational age between January 3, 2000, and September 27, 2000. This sample
constituted 20% of all infants born during the study period at this site.
Main Outcome Measures Plasma levels and particle sizes of lipoprotein subclasses and plasma
concentrations of lipids, lipoproteins (high-density lipoprotein [HDL] and
low-density lipoprotein [LDL]), and apolipoproteins.
Results An elevated lipoprotein peak of a particle with density between 1.062
and 1.072 g/mL was identified using physical-chemical methods. This subclass
of large HDL was enriched in apolipoprotein C-I (apo C-I). Based on the amount
of the apo C-I–enriched HDL peak, 156 infants were assigned to 1 of
4 groups: 0 (none detected), 17%; 1 (possibly present), 41%; 2 (probably present),
22%; 3 (elevated), 19%. Infants in group 3, compared with those in the other
3 groups, had significantly (P<.001) lower mean
birth weight (2683.7 vs 3307.1 g) and younger mean gestational age (36.2 vs
39.3 wk). After correction for age, infants in group 3 had significantly higher
levels of total and large HDL cholesterol and of total and large LDL cholesterol
and LDL particle number. However, infants in group 3 had lower levels of small
HDL, very low-density lipoproteins, and triglycerides than infants in the
other 3 groups. This lipoprotein profile differed from that in infants born
small for gestational age, who had significantly higher triglyceride (P<.001) and apo B (P = .04)
levels, but lower levels of total and large HDL cholesterol (P<.001) and apo A-I (P<.001).
Conclusions Because apo C-I–enriched HDL, and purified apo C-I alone, promotes
apoptosis in vitro, increased amounts of this particle may have physiological
significance and identify a novel group of low-birth-weight infants apparently
distinct from traditionally classified small-for-gestational-age infants.
Low birth weight is associated with cardiovascular risk factors and
death in adulthood.1 Differences in the size,
molecular weight, composition, and quantity of lipoprotein subclasses are
associated with coronary artery disease (CAD).2 In
our initial study of lipoprotein heterogeneity in cord blood, infants born
small for gestational age (SGA) had higher levels of triglyceride-rich very
low-density lipoprotein (VLDL) and intermediate low-density lipoprotein (IDL)
than infants born appropriate for gestational age.3 This
finding extended the observations of others in SGA infants4- 7 and
suggested a link between higher triglyceride-rich VLDL subclasses in SGA infants
and future CAD.
We assessed a large high-density lipoprotein (HDL) subclass in cord
blood8- 15 enriched
in apolipoprotein C-I (apo C-I). Apolipoprotein C-I is a 6.6-kDa apolipoprotein,
and in adults it is a component of VLDL, IDL, and HDL.16 Apolipoprotein
C-I displaces apo E from VLDL and IDL, thereby decreasing their clearance
from plasma.17 Apolipoprotein C-I decreases
the binding of β-VLDL to a remnant receptor, the LDL receptor-related
protein,18,19 and apo E–mediated
binding of VLDL and IDL to the LDL receptor.20,21 Such
inhibitory effects of apo C-I on VLDL and IDL removal from blood may promote
hypertriglyceridemia and atherosclerosis.16 In
that regard, Bjorkegren and coworkers22,23 reported
a significant enrichment of apo C-I in postprandial chylomicron remnants and
VLDL remnants in normolipidemic patients with CAD22 or
early asymptomatic atherosclerosis.23
Apolipoprotein C-I also decreases the transfer of cholesteryl esters
from HDL to VLDL by inhibiting cholesterol ester transfer protein (CETP).24 Apolipoprotein C-I stimulates lecithin cholesterol
acyl transferase,25 an enzyme that esterifies
cholesterol and produces the formation of the mature, spherical HDL from nascent
HDL. The effects of apo C-I on both CETP and lecithin cholesterol acyl transferase
may therefore promote increased amounts of a large HDL, which may be antiatherogenic,
unless the presence of apo C-I on large HDL renders it dysfunctional. For
example, we found that both apo C-I–enriched HDL and purified apo C-I
promote apoptosis of cultured human arterial smooth muscle cells through the
induction of neutral sphingomyelinase and the subsequent steps involved in
apoptosis.26 If such an effect occurs in vivo,
this might promote the rupture of an unstable plaque, leading to myocardial
We studied a group of 163 infants (31 white males, 39 white females;
47 black males and 46 black females), who were previously characterized.3 There were 23 SGA infants, defined as a birth weight
for gestational age of 10% or less. The infants were studied anonymously,
using cord blood that was routinely obtained after birth. The Joint Committee
on Clinical Investigations at Johns Hopkins determined that the study met
the requirements for exempt research. Therefore, informed consent was not
Plasma from cord blood was analyzed for levels of cholesterol, triglycerides,
LDL and HDL cholesterol, lipoprotein(a), and apo A-I, A-II, B, C-I, C-III,
and E.3,26 Fifteen lipoprotein
subclasses, the number of LDL particles, and the average sizes of VLDL, LDL,
and HDL were determined by nuclear magnetic resonance spectroscopy.3,27 Lipoprotein density profiles for VLDL,
LDL, and HDL were obtained after sucrose density gradient ultracentrifugation
Fractions from the lipoprotein density profile were thawed and a portion
subjected to delipidation.28,29 The
samples were analyzed in duplicate by capillary electrophoresis using the
P/ACE 5510 instrument (Beckman Coulter Inc, Fullerton, Calif) at 17 kV for
The lipoprotein fractions were thawed, centrifuged to pellet particulate
matter, and subjected to solid-phase extraction delipidation.28,29 The
apolipoproteins were eluted and concentrated, and an aliquot was taken for
matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass
spectrometry analysis, using a Voyager Elite XL DE mass spectrometer (PerSeptive
Biosystems, Framingham, Mass). The remaining samples were evaporated to dryness,
reconstituted in 250 μL of 8.0-M urea containing 2% CHAPS (3-[(3-cholamidopropyl)dimethylamino]-1-propanesulfonate),
sonicated, and degassed. Electrophoresis was performed using the IPGphor unit
(Amersham Pharmacia, Uppsala, Sweden) as described.28,29
The relationships among lipids, lipoprotein cholesterols, apolipoproteins,
lipoprotein subclasses, and lipoprotein sizes in 4 groups of infants, sorted
by the amount of the apo C-I–enriched HDL peak, were evaluated, first
using analysis of variance on data that were not adjusted for age, and then
linear regression to correct for influence of gestational age. P values were also estimated using the Kruskal-Wallis test due to the
small size of 2 of the groups (n = 5 each in group 0 and 3). To
evaluate differences in these lipid-related variables between white and black
infants and male and female infants, a χ2 test was performed.
All P values <.05 were considered significant.
All analyses were conducted using STATA software version 7.0 (STATA Corp,
College Station, Tex).
Using lipoprotein density profiles (Figure
1), 156 infants for whom we had adequate plasma samples were classified
into 4 groups, based on the gray intensity scale in the area between LDL and
HDL. Group 0 infants (n = 27) had no inflection above baseline (no
detectable apo C-I–enriched HDL); group 1 infants (n = 64)
had a small inflection (blip) above baseline (1 to 5 on scale) (possible apo
C-I–enriched HDL); group 2 infants (n = 35) had a peak above
baseline (>5 and <50 on scale) (probable apo C-I–enriched HDL); and
group 3 infants (n = 30) had a large peak above baseline (≥50
on scale) (elevated apo C-I–enriched HDL).
A prominent characteristic of the lipoprotein density profiles was the
presence or absence of a distinct peak of a particle with density between
1.062 and 1.072 g/mL between the major peaks for LDL and HDL (Figure 1). The peak density of lipoprotein(a) in adult plasma is
close to 1.055 g/mL, and thus potentially occurring within the density range
between 1.062 and 1.072 g/mL. Lipoprotein(a) levels in cord blood, however,
are very low.3 The mean lipoprotein(a) level
in group 3 infants was 1.2 (SD, 1.3) mg/dL and in group 0 infants was 0.6
(SD, 0.9) mg/dL; these minimal values while statistically different (P<.05) were not quantitatively significant.
One infant each from group 3 and group 0 were first selected for detailed
analyses of this lipoprotein peak. The lipoprotein(a) levels in these 2 infants
were low (3 mg/dL in one and undetectable in the other). Three lipoproteins,
namely, LDL, the lipoprotein with a peak of density 1.062 to 1.072 g/mL, and
HDL, were isolated by sucrose DGU, delipidated, and prepared for the following
Capillary Electrophoresis and Isoelectric Focusing. After capillary electrophoresis was performed, apo A-I (47.7%) was
the major apolipoprotein in the lipoprotein of density 1.062 to 1.072 g/mL
from the group 3 infant, and apo C-I, ordinarily a minor component of HDL,
was the second most prevalent apolipoprotein (37.6%). Negligible amounts of
apolipoproteins were detected in the same lipoprotein density segment from
the group 0 infant. These results were confirmed by isoelectric focusing,
showing clearly apo A-I (isoelectric point [pI], 5.43) and apo C-I (pI 6.70)
bands in the elevated lipoprotein peak from the group 3 infant but not the
group 0 infant.
MALDI-TOF Mass Spectrometry Analyses. The apolipoproteins
in the lipoprotein peak of density 1.062 to 1.072 g/mL from the group 0 infant
were barely detectable (Figure 2A).
In the elevated lipoprotein peak of density 1.062 to 1.072 g/mL (Figure 2B) from the group 3 infant, the intensity
of apo C-I, relative to the intensity of apo A-I, was notably greater than
in HDL (Figure 2D). There was little
difference in the spectra for HDL of usual density between the group 3 (Figure 2D) and the group 0 infants (Figure 2C). These observations were confirmed
in another group 3 infant and in a healthy control. There was no difference
detected in LDL spectra.
Plasma from 4 infants in group 3, 1 in group 2, and 3 in group 0 were
ultracentrifuged at a density of less than 1.21 g/mL and gradient gel electrophoresis
was performed.13,15 As shown in
representative densitometric scans of the gels (Figure 3), group 3 infants differed from group 0 infants. The largest
HDL subclass in group 3 infants had a mean diameter of 11.6 (range, 11.5-11.8)
nm compared with 9.4 (range, 8.8-10.8) nm in group 0 infants and 10.8 nm in
1 group 2 infant.
Apolipoprotein C-I was found in each of different HDL subclasses (Figure 3), as judged by immunoblots of the gradient
gel electrophoresis gels using an anti–apo C-I antibody. These results
are consistent with those from MALDI-TOS mass spectrometry and indicate that
all HDL subclasses contained apo C-I.
In a larger group of infants, the mean (SD) plasma level of apo C-I
of 7.7 (3.2) mg/dL in 17 group 3 infants was significantly higher (P = .04) than that of 5.1 (3.1) mg/dL in 13 infants from
group 3. More extensive apolipoprotein analyses were performed in a smaller
group of 10 infants (Table 1).
The distribution of apo A-I, A-II, B, C-I, and C-III between apo B–containing
lipoproteins (VLDL, IDL, LDL, and lipoprotein[a]) and non–apo B–containing
lipoproteins (HDL) was also determined without prior ultracentrifugation in
infants from group 3 (n = 5) and group 0 (n = 5). The
apo B–containing lipoproteins in 1 mL of plasma were precipitated with
heparin-manganese chloride, and the apolipoprotein levels were measured in
plasma, heparin-manganese supernatants (non–apo B–containing lipoproteins)
and resolubilized precipitates (apo B–containing lipoproteins) using
Apolipoprotein B. The plasma levels of total
apo B were higher in group 3 than in group 0 infants but did not reach statistical
significance (Table 1). All of the apo
B was in the heparin manganese precipitates and none was detected in the supernatants.
Apolipoprotein C-I. The mean levels of apo
C-I in both whole plasma and the heparin-manganese supernatants were about
2-fold higher in group 3 than in group 0 infants (Table 1). All of the apo C-I was in the supernatants and none was
detected in the precipitates, distinctly different than later in life when
a significant portion of apo C-I is associated with the apo B–containing
lipoproteins.30 These immunochemical results
further indicate that the apo C-I–enriched lipoprotein peak is an HDL
subclass rather than a LDL subclass.
Apolipoprotein C-III. Infants in group 3 had
significantly more apo C-III associated with non–apo B–containing
lipoproteins, while infants in group 0 had significantly more apo C-III associated
with apo B–containing lipoproteins (Table
Apolipoproteins A-I and A-II. The apo A-I levels
were higher in both the supernatants and precipitates in the infants in group
3 than in group 0. The mean apo A-II levels between groups 3 and 0 were very
similar for whole plasma, supernatants, and precipitates (Table 1).
Apolipoprotein E. In a separate experiment,
the mean concentration of apo E in group 3 infants was higher than in group
0 infants in pooled whole plasma (12.9 v 5.8 mg/dL) and in heparin-manganese
supernatants (7.9 v 4.8 mg/dL).
Lipoprotein density profiling was performed in 156 of the 163 infants
(95.7%) previously reported3 to determine the
frequency of appearance and degree of enrichment of the lipoprotein density
1.062 to 1.072 g/mL peak.
The levels of the lipid-related variables were determined in the 4 groups
of infants (Table 2). Because of the
influence of gestational age on the apo B– and apo A-I–containing
lipoproteins in this population,3 the P values were determined using data both nonadjusted and
adjusted for gestational age. Before age adjustment, all the variables except
apo B and small VLDL were significantly different. After age adjustment, the
only LDL variables that remained significantly higher in group 3 were LDL
subclasses L3 and L1 and LDL size (Table 2).
In contrast, all the HDL and VLDL related variables remained significantly
different after correction for gestational age. Large HDL levels were higher
while small HDL and the VLDL related variables were lower in group 3. We also
examined age-corrected means, which were very similar to the measured mean
levels shown in Table 2 for the HDL-
and VLDL-related variables. Differences between large LDL and large HDL among
all groups were independent of triglycerides and VLDL. Despite the fact that
the gestational ages were very similar in groups 0, 1, and 2, there were also
impressive dose-response relationships for the levels of all 6 of the HDL-related
variables from group 0 through groups 1, 2, and 3 infants (Table 2). These analyses indicate further that the differences for
the HDL subclasses shown in Table 2 were
independent of age.
The mean (SD) gestational ages of the infants were 39.7 (1.8) weeks
in group 0; 39.3 (1.3) weeks in group 1; 38.8 (1.7) weeks in group 2; and
36.2 (4.2) weeks in group 3 and differed significantly (P<.001). The mean gestational age in group 3 infants was not only
younger but had a distribution that was clearly broader than that in groups
0, 1, and 2 (Figure 4).
The mean (SD) birth weights of the infants were 3268.6 (631.9) g in
group 0; 3412.2 (548.3) g in group 1; 3240.6 (609.2) g in group 2; and 2683.7
(783.3) g in group 3 and differed significantly (P<.001),
being particularly low in group 3. After correcting for gestational age, the
birth weights were no longer significant (P = .15).
There were no significant differences in the numbers of male and female (P = .38), white and black (P = .88), or SGA and appropriately sized (P = .34) infants among the 4 groups.
Multivariable linear regression models showed that after correcting
for gestational age and race, levels of total and large HDL and apo A-I were
significantly higher in the group 3 infants than in the SGA infants without
elevated apo C-I–enriched HDL (Table 3).
Conversely, the total triglycerides and apolipoprotein B levels, indicative
of higher triglyceride-rich lipoproteins, were significantly higher in the
SGA infants than in the group 3 infants (Table
3). There were no significant differences in LDL particle number,
LDL levels, and LDL subclasses between the 2 groups. We further directly compared
the group 3 infants who had elevated apo C-I–enriched HDL (n = 24)
with infants who were SGA (n = 13); 6 infants who were both in group
3 and the SGA group were excluded from the analyses. Results similar to those
in Table 3 were found.
We next plotted the levels of large LDL (L3) and largest HDL (H5) against
gestational age for the group 3 and group 0 infants (Figure 5). Group 3 infants had higher values of L3 than group 0
infants, but these L3 levels decreased dramatically with increasing gestational
age. When all 30 infants in group 3 were included, the slope of the line for
LDL3 vs age had a β of –2.55 (95% CI, –1.32 to –3.78); P<.001 and correlation coefficient of r = 0.61. After excluding the one outlier (a 28-week infant with L3
of 73 mg/dL), the slope of the line for L3 vs age had a β coefficient
of −1.99 (95% CI, −3.20 to −0.78; P = .003) and correlation coefficient of r = 0.53. In distinct contrast, the higher H5 levels in group 3
did not decrease with gestational age, indicating strongly that the elevated
amount of apo C-I–enriched HDL in group 3 persisted and was not simply
a consequence of younger gestational age (Figure
We report here the novel observation that the large HDL subclass in
cord blood8- 15 is
enriched in apo C-I. Infants with elevated apo C-I–enriched HDL (group
3) were further unique in that they had notably lower birth weights and younger
gestational ages and significantly different plasma levels of lipids, lipoproteins,
apolipoproteins, and lipoprotein subclasses and lipoprotein size than infants
with undetectable (group 0), possible (group 1), or probable (group 2) apo
An elevated lipoprotein peak in density range 1.062 to 1.072 g/mL, identified
in about 1 in 5 infants in this well-characterized population,3 was
a large HDL particle enriched in apo C-I, expressed in all infants regardless
of sex or race. The existence of an apo C-I–enriched HDL was supported
by the congruent results of a number of physical-chemical and immunochemical
methods. At birth all the plasma apo C-I was found in HDL, and group 3 infants
had higher apo C-I levels than did group 0 infants. The distributions of the
quantity of apo C-III and apo A-I in the apo B– and apo A-I–containing
particles were also significantly different in the group 3 and group 0 infants,
suggesting other differences between their HDL particles. Consistent with
our prior work26 and that of others,9,10 the apo E content was higher in HDL
from group 3 infants. Using immunoaffinity chromatography, an apo C-I–enriched
HDL and an apo C-I–poor HDL were isolated and characterized from pooled
plasma of group 3 and of group 0 infants, respectively.26 Such
apo C-I–enriched HDL, but not apo C-I–poor HDL, promoted apoptosis
in human cultured arterial smooth muscle cells.26 These
observations support further the enriched apo C-I content of HDL from group
3 infants, and indicate its potential physiologic and clinical significance.
Apolipoprotein C-I influences the activities of CETP, lecithin cholesterol
acyl transferase, and hepatic lipase.16 Apolipoprotein
C-I in HDL inhibits CETP.24 Adults deficient
in CETP31 have large HDL particles, but in
a single study, the large HDL in cord blood did not appear to be due to a
deficiency of CETP.5 We assessed indirectly
the activity of CETP. In a subset of 40 infants, there was no relationship
between CETP activity and apo C-I levels (data not shown). The association
of the D442G (aspartic acid→glycine at codon 442) mutation of the CETP gene with increased HDL cholesterol in adults was
not seen in cord blood.32 Although lecithin
cholesterol acyl transferase activity is low in cord blood,33,34 cholesterol
esterification is normal and does not appear to account for the large HDL
species in cord blood.5 Additional studies
in a larger number of group 3 and 0 infants are needed to determine the role
of apo C-I in the production of the large, apo C-I–enriched HDL particle
and its effect on CETP and lecithin cholesterol acyl transferase activities.
Our hypothesis is that infants with elevated apo C-I–enriched
HDL at birth will have higher apo C-I levels in childhood. After the infant
exits the intrauterine environment, where there is no ingestion of dietary
fat, to the postprandial state, a transfer of apo C-I from HDL to chylomicrons
occurs. A significant enrichment of apo C-I in chylomicrons is associated
with exaggerated postprandial triglyceridemia,22 delayed
chylomicron remnant removal,22,23 and
enrichment of apo C-I in VLDL remnants in normolipidemic patients with CAD22 or early asymptomatic atherosclerosis.23 Thus,
in children older than 2 years and in adults who have elevated apo C-I levels,
a low-fat diet may be indicated to decrease the formation of chylomicrons
and VLDL remnants. If apo C-I–enriched HDL were also present in children
and adults with elevated apo C-I levels (as it is in apo C-I–transgenic
mice), both the delayed removal of triglyceride-rich remnants and the effect
of an apparently dysfunctional HDL could promote atherosclerosis and CAD.
Conde-Knape et al35 described an apo
C-I–enriched HDL in a moderately expressing APOC-I transgenic mouse model on an APOE null background.
Apolipoprotein C-I–enriched HDL (but not VLDL) inhibited hepatic lipase
but not lipoprotein lipase. We found no relationship between a common HpaI restriction fragment-length polymorphism36 in
the promotor of APOC-I in all infants. A –514
C→T restriction fragment-length polymorphism in the promotor of the hepatic
lipase gene (LIPC) explains about 30% of hepatic
lipase activity,37 but we found no relationship
between this variant and apo C-I–enriched HDL.
We examined whether elevated apo C-I HDL might be a normal concomitant
of the earlier gestational age observed in group 3 infants. Sixty percent
of the group 3 infants were born older than 36 weeks’ gestational age,
and their large HDL levels were as elevated as those in the younger age groups,
indicating that the expression of elevated apo C-I–enriched HDL was
not simply a normal consequence of younger gestational age.
An unexpected finding was that group 3 infants had a birth weight that
was 584.8 g (1.29 lb) lower than group 0 infants. The lipoprotein profile
of group 3 infants was characterized by higher total and large HDL cholesterol
but lower total and VLDL triglycerides, in distinct contrast to that of our
relatively hypertriglyceridemic SGA infants.3 Furthermore,
of the 30 infants in group 3, only 6 were SGA, and their mean gestational
age of 37.5 weeks was actually higher than that of 35.9 weeks in the 24 non-SGA
group 3 infants. The low mean birth weight in group 3 infants was present
in both the SGA (2402.1 g) and the non-SGA (2754.0 g) infants. Thus, elevated
apo C-I–enriched HDL identified a new group of low-birth-weight infants,
apparently distinct from SGA.
The relationship between low birth weight and adult cardiovascular disease
is attributed to intrauterine effects on fetal tissue development1 but might also be explained by genes that influence
birth weight and cardiovascular risk in later years. There is substantial
evidence that genes influence birth weight.38 Infants
born SGA are a heterogeneous group and may be constitutionally small or have
either symmetrical or asymmetrical pathological intrauterine growth restriction.1 The higher triglyceride-rich lipoproteins and apo
B levels found in SGA infants may provide a link to future cardiovascular
disease.3- 7 The
distinct lipoprotein phenotype of elevated apo C-I–enriched HDL, if
it persists throughout childhood into adulthood, may be a novel risk factor
for cardiovascular disease. There are other possible effects of a dysfunctional
HDL beyond apoptosis26 that may promote atherosclerosis.
Some patients with CAD and elevated HDL39 have
a proinflammatory HDL because it fails to prevent the formation of, or inactivation
of, oxidized phospholipids.40 A dysfunctional
HDL might also be less efficient at mediating efflux of cholesterol from macrophages.
How might apo C-I influence the expression of low birth weight in infants
with the elevated apo C-I–enriched HDL phenotype? Given our data on
the promotion of apoptosis in vitro by apo C-I, we were intrigued by the observation
that apoptosis, impoverished villous development, and fetoplacental angiogenesis
were significantly higher in placentas from pregnancies with low birth weight
due to intrauterine growth restriction.41- 44 Clearly,
detailed genetic, clinical, and biochemical studies of the apo C-I–enriched
HDL phenotype will be necessary to elucidate the role of apo C-I in the pathogenesis
Corresponding Author: Peter O. Kwiterovich,
Jr, MD, 550 N Broadway, Suite 308, Baltimore, MD 21205 (firstname.lastname@example.org).
Author Contributions: Dr Kwiterovich had full
access to all of the data in the study and takes responsibility for the integrity
of the data and the accuracy of the data analysis.
Study concept and design: Kwiterovich.
Acquisition of data: Kwiterovich, Cockrill,
Virgil, Otvos, Knight-Gibson, Alaupovic, Forte, Farwig, Macfarlane.
Analysis and interpretation of data: Kwiterovich,
Cockrill, Garrett, Otvos, Knight-Gibson, Alaupovic, Forte, Zhang, Farwig,
Drafting of the manuscript: Kwiterovich.
Critical revision of the manuscript for important
intellectual content: Kwiterovich, Cockrill, Virgil, Garrett, Otvos,
Knight-Gibson, Alaupovic, Forte, Zhang, Farwig, Macfarlane.
Statistical analysis: Virgil, Garrett, Zhang.
Obtained funding: Kwiterovich, Macfarlane.
Administrative, technical, or material support:
Kwiterovich, Cockrill, Virgil, Knight-Gibson, Alaupovic, Forte.
Study supervision: Kwiterovich, Alaupovic,
Financial Disclosures: None reported.
Funding/Support: This work was supported by
The Thomas Wilson Foundation and by grants from the National Institutes of
Health (M01-RR-00052, HL 18574, HL 54566, and HL 68794).
Role of the Sponsors: Neither sponsor participated
in the design and conduct of the study; collection, management, analysis,
and interpretation of the data; or in the preparation, review, and approval
of the manuscript.