(The FASEB Journal. 2002;16:1348-1360.)
© 2002 FASEB
The fetal origins of atherosclerosis: maternal hypercholesterolemia, and cholesterol-lowering or antioxidant treatment during pregnancy influence in utero programming and postnatal susceptibility to atherogenesis
WULF PALINSKI*1 and
CLAUDIO NAPOLI*,
1
* Department of Medicine 0682, University of California San Diego, La Jolla, California, USA; and
Department of Medicine, Federico II University of Naples, 80131 Naples, Italy
1Correspondence: Department of Medicine, 0682, University of California, San Diego, 9500 Gilman Dr., MTF 110, La Jolla, CA 92093-0682, USA. E-mail: wpalinski{at}ucsd.edu and cnapoli{at}ucsd.edu
 |
ABSTRACT
|
|---|
It has long been postulated that pathogenic events during fetal development influence atherosclerosis-related diseases later in life, but the mechanisms involved are unknown. This review focuses on the evidence indicating that maternal hypercholesterolemia during pregnancy is responsible for one cascade of pathogenic events. Maternal hypercholesterolemia is associated with greatly increased fatty streak formation in human fetal arteries and accelerated progression of atherosclerosis during childhood. Recent experiments in genetically more homogeneous rabbits established that temporary diet-induced maternal hypercholesterolemia is sufficient to enhance fetal lesion formation. More important, maternal hypercholesterolemia or ensuing pathogenic events in the fetus increase postnatal atherogenesis in response to hypercholesterolemia. Maternal treatment with cholesterol-lowering agents or antioxidants greatly reduces fetal and postnatal atherogenesis, indicating a pathogenic role of lipid peroxidation and a potential involvement of oxidation-sensitive signaling pathways. Experiments in a murine model showed that differences in arterial gene expression between offspring of normo- and hypercholesterolemic mothers persist long after birth, supporting the assumption that fetal lesion formation is associated with genetic programming, which may in turn affect postnatal atherogenesis. A better understanding of pathogenic programming events in utero may lead to the identification of genes determining the susceptibility to atherosclerosis and define novel preventive approaches.—Palinski, W., Napoli, C. The fetal origins of atherosclerosis: maternal hypercholesterolemia, and cholesterol-lowering or antioxidant treatment during pregnancy influence in utero programming and postnatal susceptibility to atherogenesis.
Key Words: fetal lesion • maternal cholesterol level • vitamin E • postnatal gene expression
|
THE ATHEROGENIC PROCESS MAY BEGIN DURING FETAL DEVELOPMENT
|
|---|
ATHEROSCLEROSIS, the underlying cause of myocardial infarctions and ischemic strokes, is characterized by a long lag time between onset and clinical manifestation. The prodromal stages of atherosclerotic lesions are already formed during fetal development (1
2
3
4
5)
. For example, fatty streaks containing characteristic accumulations of lipids, lipid peroxidation products, and monocyte/macrophages occur in the aorta of premature human fetuses (Fig. 1
A, B). Intimal thickening is also observed in fetal coronary arteries (6)
. In children and young adults, fatty streaks become increasingly prevalent and some of them progress to more advanced stages of atherosclerotic lesions (7
8
9
10
11
12)
. Once initiated, the progression of the atherosclerotic disease is influenced by classical risk factors that promote vascular inflammation and plaque rupture (13
14
15)
.
|
A ROLE FOR MATERNAL HYPERCHOLESTEROLEMIA IN FETAL LESION FORMATION?
|
|---|
Postnatal risk factors of atherogenesis are well defined and the mechanisms contributing to lesion formation are increasingly understood. This is particularly true for hypercholesterolemia, the preeminent role of which has been put in evidence by the marked reduction of atherosclerosis-related clinical events by cholesterol-lowering interventions (16
17
18)
. The observation that maternal hypercholesterolemia is associated with greatly enhanced fatty streak formation in human fetal arteries (Fig. 1C
) (3
, 4)
suggested that hypercholesterolemia may also play a pathogenic role in fetal lesion formation. Although the placenta is impermeable to low density lipoprotein (LDL) particles and term-born children of hypercholesterolemic mothers generally have normal cholesterol levels, a good correlation between maternal and fetal plasma cholesterol levels in 5- to 6-month-old human fetuses (but not in older ones) suggested that during the earlier stages of pregnancy, maternal hypercholesterolemia may promote lesion formation in the fetus (3)
. A strong influence of maternal hypercholesterolemia on the fetal sterol metabolism during some stages of pregnancy has recently been established in animal models (19
, 20)
.
|
IS FETAL LESION FORMATION RELEVANT?
|
|---|
Fetal lesions occur at the same predilection sites as more advanced lesions in adolescents and adults. However, the size of fetal lesions is minute and there is evidence they may partially regress during the final stages of gestation or early infancy, when cholesterol levels are low (11)
. Therefore, interest in the mechanisms contributing to fetal lesion formation would be limited unless these lesions or pathogenic events in the arterial wall induced by or concomitant with their formation influence atherogenesis later in life.
The Fate of Early Lesions in Children (FELIC) study indicated that this may indeed be the case (11)
. This autoptic study of 156 normocholesterolemic children, age 1–14 years, showed that the progression of atherosclerosis was markedly faster in offspring of hypercholesterolemic mothers than in those of normocholesterolemic mothers. Conventional risk factors assessed in children and their mothers could not explain this difference. Clearly, differences in the genetic background may have contributed to a faster progression of the disease in children of hypercholesterolemic mothers, but a number of theoretical considerations made it unlikely that this was the sole explanation. For example, when atherosclerosis was plotted over age, both groups of children were represented by regression lines with high regression coefficients (R=0.87 to 0.98). The slope of the two lines was very different, but few data were scattered between the lines. Such a dichotomous distribution of lesion sizes is hard to reconcile with the assumption that the susceptibility to atherosclerosis was determined by multiple maternal and paternal genes. On the other hand, dominant maternal genes that would not manifest themselves through dyslipidemia or other parameters determined in the FELIC study could not be ruled out as a potential explanation.
Given the genetic heterogeneity of human populations and the variability in diets and conventional risk factors, evidence for a pathogenic role of maternal hypercholesterolemia in fetal lesion formation and enhanced susceptibility to postnatal atherogenesis could only be obtained in genetically more homogeneous experimental models. Establishing such causal links would be important not only because it would add to our understanding of the pathogenesis of the disease, but for clinical considerations as well. If maternal hypercholesterolemia is responsible at least in part for a chain of events leading to increased fetal lesion formation, increased susceptibility to atherosclerosis later in life, and ultimately to increased clinical manifestations, then pharmacological interventions in mothers during pregnancy that prevent or reduce the onset of these pathogenic events in the fetus could be expected to provide lifelong benefits.
|
MATERNAL HYPERCHOLESTEROLEMIA AND LIPID PEROXIDATION ENHANCE FETAL LESION FORMATION
|
|---|
Direct evidence for the causal role of maternal hypercholesterolemia and the involvement of oxidative stress has recently been obtained in a rabbit model (21)
. Groups of originally normocholesterolemic female New Zealand white (NZW) rabbits were fed a control chow or one of two hypercholesterolemic diets yielding average plasma cholesterol levels of 
153 and 360 mg/dL during pregnancy. Additional groups received the same two hypercholesterolemic diets supplemented with 1–3% cholestyramine, 100 IU vitamin E, or both.
The extent of lesions in their offspring (15–25 per group) at term birth was then determined in cross sections through the entire aorta. As expected, cholesterol levels significantly increased in mothers fed the hypercholesterolemic diets (Fig. 2
A). Cholestyramine reduced maternal cholesterol, whereas vitamin E had no effect. Plasma cholesterol levels in offspring of all groups at birth were nearly identical (Fig. 2B
), confirming that at the end of a regular pregnancy there is no correlation between the maternal and fetal cholesterol level. In contrast, plasma concentrations of oxidized fatty acids and other measures of lipid peroxidation, such as malondialdehyde, differed significantly between groups (Fig. 2C
). Lipid peroxidation end products increased roughly in proportion to the maternal cholesterol level in offspring of hypercholesterolemic and cholestyramine-treated mothers, and were markedly reduced in offspring of vitamin E-treated mothers. Hypercholesterolemia is known to be accompanied by increased lipid peroxidation in plasma and tissues (22
23
24)
. Because plasma cholesterol levels were similar and very low in all groups of offspring, these differences are likely to reflect placental transfer of oxidized fatty acids from the mother to the fetal circulation (25
, 26)
. Nevertheless, differences in the activity of antioxidant enzymes or other factors influencing the oxidative modification of lipids in the fetus/newborn cannot be ruled out.
View larger version (35K):
[in this window]
[in a new window]
|
Figure 2. Evidence for the causal role of maternal hypercholesterolemia and lipid oxidation in fetal lesion formation. Groups of NZW rabbit mothers were fed normal chow, one of two increasingly hypercholesterolemic diets (Chol 1 or Chol 2) alone, or the same hypercholesterolemic diets supplemented with cholestyramine or vitamin E during pregnancy. A) Average plasma cholesterol level of mothers during pregnancy. Both hypercholesterolemic diets significantly raised maternal cholesterol compared to the control group; cholestyramine, but not vitamin E treatment, reduced it vs. the untreated group on the same diet. *P < 0.0001 vs. control; #P < 0.001 vs. Chol 1; +P < 0.0001 vs. Chol 2; °P < 0.0001 vs. same diet+vitamin E. All P values are Bonferroni corrected. B) Offspring plasma cholesterol at birth. No significant differences were observed between offspring groups. C) Oxidized oleic acid in plasma of offspring at birth. Similar results were obtained for other measures of lipid peroxidation, such as oxidized linoleic acid or malondialdehyde. *P < 0.0001 vs. control; #P < 0.0001 vs. Chol 1; +P < 0.0001 and ++P < 0.001 vs. Chol 2, respectively; °P < 0.0001, °°P < 0.01, and °°°P < 0.05 vs. same diet+cholestyramine; $P < 0.05 vs. Chol 2 + vit E. D) Lesions in the aorta of offspring at birth (expressed as the cumulative area of all lesions per section; average of 90 equidistant cross sections per animal). *P < 0.0001 vs. control; #P < 0.0001 vs. Chol 1; +P < 0.01, ++P < 0.001, and +++P < 0.0001 vs. Chol 2, respectively. Modified from ref 21
, with permission.
|
|
The size of lesions in offspring provided unequivocal evidence for the pathogenicity of maternal hypercholesterolemia (Fig. 2D
). Lesions doubled in the Chol 1 group and quadrupled in the Chol 2 group (P<0.0001). Maternal cholestyramine treatment significantly reduced fetal lesions roughly proportional to the reduction of maternal cholesterol levels. A regression analysis of individual animals in all groups (except those receiving vitamin E) confirmed the linear correlation between maternal cholesterol and lesions at birth (r=0.78, P<0.0001) (21)
. Vitamin E treatment of mothers did not affect maternal hypercholesterolemia but reduced atherosclerosis at birth by 40% (in the Chol 2 group), clearly implicating lipid oxidation or increased intracellular oxidative stress in fetal lesion formation.
These data provide compelling evidence for an atherogenic effect of maternal hypercholesterolemia and the involvement of oxidative stress in fetal lesion formation. It is therefore tempting to assume that fetal fatty streak formation is promoted by the same mechanisms as conventional atherogenesis (28
29
30)
.
Increased oxidative stress during fetal development is of particular interest in view of the potential modulation of postnatal atherogenesis. There can be little doubt that arterial cells are exposed to significant oxidative stress. Lesions of human fetuses and term-born rabbits were rich in oxidized LDL (OxLDL) (3
4
5
; Fig. 1B
) that showed a distribution similar to that in early lesions of adults and animal models (30
31
32)
. OxLDL promotes the recruitment of macrophages into the intima (3)
and greatly enhances foam cell formation (28)
. A multitude of mechanisms involving oxidatively modified proteins and other peroxidative compounds can affect the basal machinery of the cell. These include interference with oxidation-sensitive cytoplasmic and/or nuclear signaling pathways that regulate arterial gene expression or transcription (reviewed in refs 33
34
35
). For example, OxLDL modulates the expression of genes involved in cell differentiation and proliferation regulated by the nuclear factor kappa B (36
, 37)
. Mildly and extensively modified OxLDL influence the expression of apoptotic factors activated through Fas and TNF receptors (38
, 39)
, c-Myc-dependent transcription factors (40
, 41)
, as well as genes regulated by the peroxisome proliferator-activated receptor gamma that promote inflammation or reverse cholesterol transport (42
43
44)
. Furthermore, OxLDL triggers extensive immune responses (30
, 45)
, some of which actively modulate atherogenesis (46)
. Given these multiple effects, it is likely that oxidative stress also plays a role in the postulated in utero programming events.
|
MATERNAL HYPERCHOLESTEROLEMIA ENHANCES POSTNATAL ATHEROGENESIS
|
|---|
Experimental evidence that maternal hypercholesterolemia or the ensuing pathogenic events in the fetus indeed enhance the susceptibility to atherosclerosis later in life was recently provided in the same rabbit model (47)
. This study had a design similar to the preceding (21)
, but used only one hypercholesterolemic diet (Chol 2) and a higher dose of vitamin E (1000 IU). Maternal cholesterol levels during pregnancy were very similar to those of the previous study (Fig. 3
A), and lesions at birth confirmed the atherogenic effect of maternal hypercholesterolemia on in utero lesion formation and the protective effect of cholesterol-lowering and antioxidant interventions (Fig. 3B
).
View larger version (58K):
[in this window]
[in a new window]
|
Figure 3. Evidence for the causal role of maternal hypercholesterolemia and lipid oxidation in accelerated postnatal atherogenesis. NZW rabbit mothers were fed normal chow, a hypercholesterolemic diet alone, or the same hypercholesterolemic diet supplemented with cholestyramine or vitamin E during pregnancy. All mothers were switched to normal chow at birth. Offspring were fed a mildly hypercholesterolemic diet until 6 or 12 months of age. A) Average plasma cholesterol level of mothers during pregnancy and terminal levels of their offspring at death. Note that all diets and treatments in the group names refer to mothers, not their offspring. Hyperchol, hypercholesterolemic diet; Vit E, vitamin E; Cholestyr, cholestyramine. *P < 0.0001 vs. control; #P < 0.0001 vs. Hyperchol; +P < 0.0001 vs. Hyperchol and Vit E. B) Average cumulative area of all lesions per section in equidistant aortic cross sections of offspring killed at birth or at 6 and 12 months. Inset: Increase in lesion sizes from 6 to 12 months. Although cholesterol levels were nearly identical in all groups of offspring at any given time, progression of atherosclerosis was clearly accelerated in offspring of untreated hypercholesterolemic mothers compared to offspring of normocholesterolemic mothers. Maternal treatment during pregnancy with a high dose of vitamin E (1000 IU) reduced or abolished this. All P values are Bonferroni corrected. *P < 0.0001, **P < 0.001, ***P < 0.01 vs. control, respectively; #P < 0.0001, ###P < 0.01, ####P < 0.05 vs. Hyperchol, respectively; +P < 0.0001 vs. Hyperchol+Cholestyr. C) Aortic surface area covered by atherosclerosis in offspring at 12 months. An additional group of offspring of mothers treated with both vitamin E and cholestyramine is included in this panel. D) Representative examples of oil red O-stained en face preparations of aortas at 12 months. *P < 0.05 vs. control; #P < 0.0005, ##P < 0.005, and ###P < 0.02 vs. Hyperchol, respectively; +P < 0.005 vs. Hyperchol + Cholestyr. Modified with permission from ref 47
.
|
|
Offspring were then fed a mildly hypercholesterolemic diet containing 0.14% cholesterol for up to a year, which raised cholesterol in all groups to similar levels of 150 mg/dL at 6 months and 270 mg/dL at 12 months (Fig. 3A
). Lesion sizes (again measured as the cumulative lesion area in equidistant sections through the aorta) increased only moderately during the first 6 months, but by up to 8.3-fold during the ensuing 6 months (Fig. 3B
). Absolute differences in lesion sizes between offspring of untreated hypercholesterolemic mothers and offspring of cholestyramine or vitamin E-treated hypercholesterolemic or chow-fed control mothers were significantly greater at 12 months than at birth or 6 months. A linear representation of lesion progression (Fig. 3B
, inset) shows that progression of atherosclerosis was accelerated in offspring of untreated hypercholesterolemic mothers. Determination of another measure of atherosclerosis, i.e., the aortic surface area covered by oil red O-positive lesions, showed analogous differences between groups (Fig. 3C, D
). Remarkably, maternal vitamin E treatment alone or in combination with cholestyramine reduced lesion size to or slightly below that of offspring of normocholesterolemic mothers (Fig. 3C
).
These results demonstrate that maternal hypercholesterolemia accelerates the atherogenic response to postnatal exposure to hypercholesterolemia and indicate that pathophysiological events in utero influence the susceptibility to atherosclerosis later in life. A previous experiment in the rabbit model had shown that in the absence of postnatal hypercholesterolemia, significant differences in lesion sizes between groups persisted at 6 months of age, but that the absolute sizes of lesions did not progress compared to birth. This seems to indicate that the pathogenic events occurring during fetal development per se do not induce later lesion formation and that postnatal hypercholesterolemia is a necessary cofactor. However, great caution should be applied in extrapolating this assumption to humans. Cholesterol levels in chow-fed NZW rabbits are very low (50 mg/dL) and may actually result in lesion regression, whereas lesion sizes clearly increase with increasing age in children, even those with ‘normal’ cholesterol levels and lacking other classical risk factors of atherosclerosis (3)
.
Future studies will have to address the question whether the degree of postnatal hypercholesterolemia influences the degree of acceleration of atherogenesis caused by in utero programming. The level of hypercholesterolemia induced in rabbits (47)
was comparable to that seen in the many human subjects and modest by comparison to those seen in most atherosclerosis studies in NZW rabbits fed 1% cholesterol diets. Nevertheless, it represents a fivefold increase compared to the physiological level in this strain. Based on the lack of atherogenesis in chow-fed offspring, one could argue that the effect of fetal programming should increase with increasing postnatal cholesterol levels; conversely, it cannot be ruled out that more extreme hypercholesterolemia may partially mask the effect of in utero programming.
In humans, it remains to be established how much of the accelerated atherogenesis in offspring of hypercholesterolemic mothers (3)
is caused by in utero programming events and how much is due to inherited genetic differences. Nevertheless, the demonstration that fetal pathogenic events increase the postnatal susceptibility to atherosclerosis and that interventions in mothers reduce or prevent this greatly increases interest in the mechanisms responsible.
|
MATERNAL HYPERCHOLESTEROLEMIA MODULATES POSTNATAL GENE EXPRESSION IN THE ARTERIAL WALL
|
|---|
As discussed above, numerous signaling pathways are affected by increased oxidation of LDL or intracellular formation of reactive oxygen species (33
34
35
36
37
38
39
40
41
42
43
44)
. The marked increase of lipid peroxidation in maternal and fetal plasma and the prevalence of OxLDL and other markers of lipid peroxidation in lesions of human fetuses and newborn rabbits make it a foregone conclusion that many genes regulated by oxidation-sensitive pathways will be up- or down-regulated in fetuses of hypercholesterolemic mothers. One would also assume that such regulation disappears once placental passage of oxidized fatty acids (and other, yet unknown effects of maternal hypercholesterolemia) ceases at birth and acute oxidative stress is low for a while. The presence of significant differences in gene expression between offspring of normo- and hypercholesterolemic mothers after a prolonged period on a normocholesterolemic diet would therefore indicate persistent effects of maternal hypercholesterolemia. Some of these could contribute to the enhanced susceptibility to postnatal atherogenesis.
The first evidence for persistent differences in arterial gene expression was obtained recently in LDL receptor-deficient (LDLR-/-) mice (48)
. Hypercholesterolemic mice have become a widely used experimental model of atherosclerosis (49)
. A murine model was selected for the fetal studies because the murine genome is far better characterized than that of the rabbit and because microarrays encompassing a large segment of the murine genome are commercially available. In contrast to most other murine strains, including the parent strain (C57BL/6), extensive hypercholesterolemia can be induced in LDLR-/- mice by dietary means. Thus, cholesterol levels similar to or exceeding those seen in human hypercholesterolemic mothers can be induced in these mice.
The experimental design is shown in Fig. 4
. Groups of female mice were fed regular chow or high-fat diets supplemented with 0.075% or 1.25% cholesterol starting 3 wk before pregnancy. This resulted in cholesterol levels of 1063 and 1299 mg/dL, respectively, compared to 250 mg/dL in the control group. (Maternal cholesterol levels were determined only at the start rather than throughout pregnancy, because blood sampling in mice requires anesthesia that may impair fetuses.) At birth, mothers in the hypercholesterolemic groups were switched to regular chow, and within 1 wk maternal cholesterol levels in all three groups were similar. Offspring were fed regular chow until the age of 3 months and had nearly identical cholesterol levels of 260 mg/dL. Under these conditions, offspring should have enough lesions in the aortic origin to permit one to assess atherogenic effects of maternal hypercholesterolemia, whereas atherogenesis in the aortic tree should be minimal. Hence, determination of aortic gene expression should not be affected by cellular heterogeneity resulting from the presence of various stages of lesions.
View larger version (32K):
[in this window]
[in a new window]
|
Figure 4. Evidence for in utero programming in a murine model of fetal atherogenesis. Modified with permission from ref 48
. See text for a description of the experimental design and results.
|
|
At the age of 3 months, lesions in the aortic origin were indeed markedly greater in male offspring of both hypercholesterolemic groups than in the control group (Fig. 4
, bottom left). Confirmation of an atherogenic effect of maternal hypercholesterolemia in a second animal model lends strength to the assumption that it contributes to enhanced lesion formation in humans. The relative increase was smaller than that observed in humans (Fig. 1C
) or rabbits (Figs. 2D
and 3B)
, but it should be kept in mind that the maternal control group was not normocholesterolemic by murine standards and that some fetal lesion formation may already have occurred in this group.
Visual inspection of all aortas and microscopic examination of serial sections throughout the entire length of two aortas from each group indicated the absence of lesions. After careful in situ removal of the adventitia from the aortas of all other experimental animals, RNA was extracted from individual or pooled aortic segments or entire aortas, as described in detail in ref 49
, and subjected to analysis by Affymetrix gene chips.
Microarrays have been used extensively for studies of cultured cells, but their application to cancer (50
, 51)
and cardiovascular tissues (52
53
54
55)
is only just beginning. In addition to the lack of synchronicity of cell proliferation and differentiation in tissues, experimental strategies using microarray approaches to investigate gene regulation during atherogenesis have to confront the problem that once lesion formation has been initiated, a great variability in cellular composition of the intima exists. The activity of vascular cells, e.g., macrophages, may vary greatly depending on their location within the lesion and their interactions with endothelial cells, SMC, and T cells. For this reason, our first application of this technique focused on the initial changes in arteries before the onset of microscopically detectable intimal thickening.
Data from six comparisons between offspring of normo- and hypercholesterolemic mothers were analyzed together, using ‘Equalizer’ software (56)
. Statistical significance was assessed by nearest neighbors analysis (57)
and t tests. Results were further restricted by requiring that significantly regulated genes be similarly up- or down-regulated in offspring of both hypercholesterolemic maternal groups. Microarray analysis of the expression of 11,000 murine genes and ESTs in the nonatherosclerotic aortic media and intima indicated that 135 genes/ESTs were significantly up- or down-regulated in offspring of hypercholesterolemic mothers. A graphic representation of these genes is shown at the bottom center of Fig. 4
and a dendrographic analysis (by GeneSpring software) at the bottom right. Regulation of genes was consistently greater in offspring of mothers fed 1.25 cholesterol than in those fed the 0.075% diet, consistent with greater lesion sizes in the latter.
A subset of 12 up-regulated transcripts was subjected to secondary analysis by semiquantitative PCR, using a balanced pool of RNA from all experimental mice of the 0.075% cholesterol and control groups. Four of these genes were found to be up-regulated more than 1.7-fold [fibroblast growth factor binding protein (FGFbp), flavin containing mono-oxidase 3 (FCMo3), NPAS2 (MOP4), and a potassium channel (MERG1)]. Comparison of the expression of one of these genes, FCMo3, in individual mice by quantitative PCR yielded qualitatively similar differences (48)
. Immunocytochemistry of lesion-fee segments of the aortic origin of the same mice indicated consistently more MERG, FGFbp, and NPAS2 protein in the offspring of hypercholesterolemic mothers than in controls (48)
. The degree of regulation of most genes identified by the microarray approach was modest, consistent with the assumption that genes affecting susceptibility to postnatal atherogenesis may exert their influence by acting over prolonged periods rather than by displaying dramatic levels of regulation. However, the joint analysis of medial and endothelial cells may have masked selective regulation of endothelial genes.
Most genes identified by the microarray approach have not previously been linked with atherosclerosis. This is not surprising given that the experiment examined the aortic wall before the onset of visible intimal thickening. This approach reduced tissue heterogeneity considerably, but meant that many genes expressed in the atherosclerotic intima and contributing to lesion progression were unlikely, a priori, to be detected. Future studies investigating the expression of genes affected by fetal programming in animals exposed to hypercholesterolemic diets after birth, in which atherosclerotic lesions are present, will have to rely on laser dissection microscopes to isolate intimal tissues or specific cell types (58)
. This is increasingly attractive due to rapidly diminishing amounts of RNA required for the microarrays and technical advances in laser capture microscopes.
The presence of changes in aortic gene expression long after the end of fetal exposure to hypercholesterolemia establishes, in principle, that maternal hypercholesterolemia results in in utero programming. However, extensive future work is necessary to identify genes that are consistently regulated over time and to narrow the list of candidate genes that may actually contribute to accelerated postnatal atherogenesis. The latter could be attempted, for example, by carrying out maternal interventions similar to those performed in rabbits. Genes not affected by treatments that reduce postnatal atherogenesis could then be ruled out as potential contributors. Interventions with antioxidants may be of particular interest because they may provide clues to the involvement of specific oxidation-sensitive signaling pathways.
Intervention experiments in experimental models may answer the question of whether in utero programming affects genes promoting or inhibiting lipid peroxidation in the arterial wall. A first indication that differences in the vascular activity of antioxidant enzymes (Mn-superoxide dismutase, catalase, or glutathione peroxidase) may influence lesion formation was provided in human fetuses (4)
. In that study, intracranial arteries showed greater antioxidant activity and a much smaller atherogenic response to (maternal) hypercholesterolemia than extracranial arteries. A recent report supports this by showing that when the antioxidant activity in intracranial arteries declines to that of comparable size extracranial arteries in adult and elderly subjects, their atherogenesis accelerates (59)
.
Rabbit or murine models may also be used to test the efficacy of interventions modulating oxidative stress by other pathways and to determine their effects at the molecular level. For example, a multitude of drugs affecting nitric oxide bioactivity are now available that not only have direct effects on endothelial function and vascular tonus, but may contribute to reduced formation of oxygen radical species (60
, 61)
.
|
THE MATERNO/FETAL CHOLESTEROL HYPOTHESIS
|
|---|
Figure 5
provides an overview of the fetal cholesterol hypothesis in the context of other atherogenic mechanisms and highlights some diagnostic and preventive implications of the hypothesis. On the basis of all autoptic studies in humans and the results obtained in animal models, it is clear that maternal hypercholesterolemia and the ensuing oxidative stress exert an atherogenic role in the fetus and later in life. However, it has yet to be established whether maternal hypercholesterolemia actually translates into fetal ‘hypercholesterolemia’ and what levels of fetal cholesterol are atherogenic. Data from experimental animals indicate that even though the placenta is impermeable for large lipoprotein particles and the fetus or placenta synthesize most of the cholesterol required by the fetus (27)
, during the second trimester of pregnancy the fetal cholesterol levels are determined by the maternal sterol metabolism (19
, 20)
. In 5- to 6-month-old human fetuses, fetal cholesterol levels correlate with maternal ones (Fig. 1C
in ref 3
). It is therefore likely that fetuses of hypercholesterolemic mothers have higher cholesterol levels than those of normocholesterolemic mothers. The number of very premature human fetuses examined to date is too low to confirm this with any degree of certainty. Any such comparison in human fetuses of different ages is further complicated by the fact that fetal cholesterol decreases in a linear fashion with increasing fetal age (Fig. 1B
in ref 3
) and that the correlation between fetal and maternal cholesterol levels is lost after the 6th month.
View larger version (49K):
[in this window]
[in a new window]
|
Figure 5. The materno/fetal cholesterol hypothesis. See text for detailed description of pathogenic events and diagnostic and therapeutical implications. Paintings: mother: Jan Vermeer, 1632–1675; child: Francisco de Goya, 1746–1828; adult with risk factors: Pierre Auguste Renoir, 1841–1919; announcement of an acute event; Rembrandt van Rijn, 1606–1669; sickness and death: Paul Delaroche, 1797–1885. Diagnostic approaches to assess risk are indicated by yellow arrows, preventive approaches by green arrows.
|
|
The issue of what constitutes hypercholesterolemia in the fetus is also complex. Cholesterol levels were very high at age 5–6 months in all human fetuses, even those of normocholesterolemic mothers. It is therefore likely that these levels are physiological, i.e., reflect an increased need for cholesterol in the rapidly growing fetus. It is nevertheless possible that they lead to some lipid accumulation in the arterial wall and that any further increase greatly enhances lesion formation. The presence of some fatty streaks even in fetuses of normocholesterolemic mothers (Fig. 1)
is consistent with this assumption.
Increased maternal cholesterol levels during pregnancy go hand in hand with increased lipid peroxidation in the mother. Maternal hypercholesterolemia and/or increased maternal oxidative end products directly or indirectly increase lipid peroxidation products in fetal plasma and fatty streak formation in the fetal aorta. Although the causal role of maternal hypercholesterolemia in fetal onset of the atherogenic process has been established and a clear link exists between maternal and fetal cholesterol metabolism during part of the gestation, it is not known how the atherogenic effect is mediated from mother to fetus. Detrimental effects of increased oxidative stress on placental functions and direct passage of oxidized fatty acids from mother to fetus may contribute. Based on extensive in vitro evidence (35
36
37
38
39
40
41
42)
, it can be presumed that accumulation of OxLDL in fatty streaks and enhanced oxidative stress in plasma affect multiple oxidation-sensitive signaling pathways in the arterial wall of the fetus. These in turn modulate the expression of many regulatory genes that affect endothelial function and lesion formation, and thus may influence molecular memory in the arterial wall that determines later atherogenesis in response to classical risk factors (see below).
All of these events are presumed to vary considerably during gestation. Maternal cholesterol levels increase during the third trimester, even in ‘normal’ mothers (62)
, and preliminary data indicate that this increase is much greater in hypercholesterolemic mothers. Placental functions and permeability can also be assumed to change over time, if only as a result of rapid growth. Finally, it has been established that fetal cholesterol levels are high during the second trimester and decline steadily toward birth. We therefore postulate that the pathogenic effect of maternal hypercholesterolemia is not constant throughout pregnancy, but that a window of vulnerability may exist during which high fetal cholesterol levels and strong oxidative stress in both the mother and fetus coincide with a vulnerable immature arterial wall.
Pathogenic events in the fetal artery caused by maternal hypercholesterolemia or the ensuing fatty streak formation then enhance the susceptibility to postnatal hypercholesterolemia and presumably to other conventional risk factors of the disease as well. This may be mediated in part by in utero programming, i.e., persistent up- or down-regulation of genes. Fetal arteries and lesions are very small compared to those in children and adults. It is therefore unlikely, from a quantitative perspective, that persistent fetal lesions contribute significantly to the atherosclerotic burden later in life. However, it is possible that subtle changes of the arterial wall composition (in cells and matrix components) are maintained during growth and contribute to accelerated postnatal atherogenesis (63)
. Whether in utero programming is sufficient to promote atherogenesis in the absence of postnatal risk factors remains to be established, at least in humans.
Animal models have unequivocally shown that maternal hypercholesterolemia promotes fetal lesion formation and accelerates postnatal atherogenesis, but it remains unknown to what extent they contribute to these events in humans. Inherited genetic predisposition is undoubtedly greater in fetuses and children of chronically hypercholesterolemic mothers or mothers that develop temporary hypercholesterolemia during pregnancy and is bound to contribute to lesion formation.
A third factor postulated to influence later atherogenesis is neonatal imprinting of genes. Throughout this review, the terms ‘imprinting’ and ‘in utero programming’ are used in a broad sense to describe a mechanism(s) occurring during a limited prenatal period and leading to permanent changes in gene regulation or other changes in cellular activities that affect the initiation and progression of atherosclerosis. We do not wish to imply that these mechanisms are similar to maternal imprinting, i.e., the deactivation of paternal genes. Evidence for persistent effects of a brief stimulation during the neonatal period has been reported. For example, enhancement of cholesterol degradation by cholestyramine in neonatal guinea pigs and very young white Carneau pigeons conveyed a significant protection against dietary hypercholesterolemia in adulthood (64
, 65)
. This was associated with an increased expression and activity of 7
-hydroxylase in pretreated adult animals upon dietary cholesterol stimulus. Hormonal imprinting in neonatal rats exposed to thyroid-stimulating hormone has been described (66)
. Parental imprinting also occurs during particular stages of fetal development, as shown for the Mas proto-oncogene (67)
. An extensive but controversial body of evidence exists for the effect of dietary factors during lactation and infancy (68
, 69)
. Although not directly pertinent here, the mere possibility that dietary differences influence later atherogenesis emphasizes the importance of not confounding the design of studies in experimental models by administering hypercholesterolemic diets to mothers or newborns during lactation.
During adolescence and in adulthood, atherogenesis is clearly driven by conventional risk factors and becomes an extraordinarily complex process (13
14
15)
. So far, accelerated progression of atherosclerosis in human offspring of hypercholesterolemic mothers has been established only for children and adolescents (11)
. Given the dramatic difference seen throughout childhood and the fact that lesion sizes diverged linearly with increasing age, it is tempting to extrapolate results to adulthood. However, as previously argued for the acceleration of postnatal atherosclerosis in rabbits, it cannot be ruled out that the presence of marked hypercholesterolemia and other risk factors attenuate the effect of fetal programming. Clearly, there is a need to evaluate adult and elderly human populations. It will be necessary to establish a higher incidence of clinical manifestations of acute plaque rupture, arterial stenosis, or death from atherosclerosis-related causes in offspring of hypercholesterolemic mothers.
Studies investigating the effect of maternal hypercholesterolemia, fetal or postnatal imprinting, and postnatal risk factors on atherogenesis from the fetus to old age are complicated by the question of how to best assess the rate of progression of atherosclerosis. In Fig. 5
, a linear progression of atherosclerosis is indicated. Indeed, in human children, linear increases over age were noted for both individual lesion areas and the cumulative area of all lesions corrected for the size of the aorta (specifically, the average cumulative area of lesions in cross sections through the entire aorta divided by the cross-sectional area encompassing the lumen and the arterial wall) (11)
. This was necessary in order to differentiate between lesion growth proportional merely to the growth of the aorta and accelerated atherogenesis and useful when comparing arteries of different caliber, but it must be kept in mind that absolute lesion areas increase exponentially over time.
|
DIAGNOSTIC AND THERAPEUTICAL IMPLICATIONS OF THE MATERNO/FETAL CHOLESTEROL HYPOTHESIS
|
|---|
If it can be established that fetal pathogenic events linked to maternal hypercholesterolemia contribute significantly to atherosclerosis-related morbidity and mortality, an early recognition of the risk would be desirable. Where multiple routine examinations during pregnancy are the rule, the average maternal cholesterol level may be a sufficient indicator of risk. In fact, a differentiation between inherited risk and consequences of fetal exposure to hypercholesterolemia may not be necessary. In both cases, more intense monitoring and perhaps more rigorous reductions of postnatal risk factors seem appropriate. Maternal hypercholesterolemia should therefore be added to the list of risk factors justifying such steps (70
71
72)
. The diagnostic approaches to estimate lesion progression in children and young adults have been the subject of an earlier review (73)
and will not be discussed here.
It is not clear at what maternal cholesterol level increased fetal lesion formation begins, whether the relationship is linear or not (6)
, or what period of fetal development is most vulnerable, if any. All of these uncertainties reduce the prognostic accuracy of cholesterol measurements at a nonstandardized time during pregnancy. Theoretically, an accurate assessment of lesion sizes at birth would be preferable. Because this is not possible by current techniques, surrogate measures may be considered—for example, measures of lipid peroxidation in the plasma of offspring at birth, which correlate reasonably well with fetal lesion formation (Fig. 6
).
View larger version (24K):
[in this window]
[in a new window]
|
Figure 6. Correlation between lipid peroxidation parameters in plasma and aortic lesions in rabbit offspring of normocholesterolemic, hypercholesterolemic, and cholestyramine/vitamin E-treated mothers at the time of birth.
|
|
Of even greater importance are the therapeutical implications of the hypothesis. Cholesterol lowering in hypercholesterolemic mothers by dietary means (74)
, cholestyramine, or other hypocholesterolemic drugs that are safe during pregnancy are obvious candidates. Cholesterol lowering reduces oxidative stress (21
, 47
, 75)
. Antioxidants alone or in combination with cholesterol lowering are also promising (21
, 47)
. In experimental animals, the reduction of fetal lesions and postnatal atherogenesis by vitamin E was remarkable. Again, caution must be taken in extrapolating this to humans. Overwhelming evidence indicates that structurally unrelated antioxidants inhibit conventional atherosclerosis in animal models (reviewed in refs 28
, 29
), but human trials have yielded conflicting results (76
77
78
79
80)
. The majority of these trials measured clinical outcomes in adult subjects with preexisting and often advanced lesions in whom multiple risk factors were present and who were treated for a limited period and often with relatively low doses of antioxidants. Moreover, the time of follow-up (1 to 4 years) was probably too short to assess the clinical outcome of a chronic disease. It is therefore doubtful that they provide useful indications on the efficacy of high doses of vitamin E in human fetuses, where prevention of pathogenic effects on oxidation-sensitive regulatory pathways may be more important than reduction of other atherogenic or thrombogenic effects of OxLDL.
|
RELATIONSHIP TO OTHER HYPOTHESES ON FETAL DETERMINANTS OF CARDIOVASCULAR DISEASE
|
|---|
The idea that environmental influences on uterine development determine later pathologies is not new. Most notable is the Barker hypothesis linking reduced birth weight to hypertension and atherosclerosis-related diseases later in life (81
82
83)
. This challenges the view that predisposition to atherosclerosis is an exclusive consequence of inherited genetic traits and has been the subject of acerbic controversy (84
85
86)
even very large epidemiological studies have not been able to resolve (86
, 87)
. The most convincing evidence supported by a recent report (88)
points to a long-term effect of maternal diet on blood pressure, but the large number of genes, postnatal risk factors (some of which are influenced by inherited genes), and age-dependent factors make it extraordinarily complicated to disentangle genetic from environmental influences on atherosclerosis or to demonstrate a causal role of fetal programming. Another major obstacle has been that many pathogenically distinct factors can lead to a reduced birth weight. Therefore, experimental verification of the Barker hypothesis is only in its infancy (89
, 90)
, and virtually nothing is known on the molecular nature of in utero programming and the pathogenic mechanisms causing it.
The materno/fetal cholesterol hypothesis differs from the Barker hypothesis in that there is little evidence for a significant role of reduced birth weight. The FELIC study noted an inverse correlation between birth weight and atherosclerosis in children, but only in offspring of normocholesterolemic mothers (11)
. A slight trend toward lower birth weights in rabbit offspring of hypercholesterolemic mothers also failed to reach statistical significance. Nevertheless, experimental evidence for the materno/fetal cholesterol hypothesis supports the basic assumption of the Barker hypothesis, i.e., that conditions during fetal development profoundly influence atherogenesis later in life. By providing a specific hypothesis of the pathogenic mechanisms involved and identifying genetically relatively homogeneous experimental models in which their impact can be determined in the absence of the majority of confounding factors vexing human studies, we may answer the fundamental question of whether fetal programming can be pathogenetically and clinically relevant.
|
OUTLOOK AND CLINICAL PERSPECTIVES
|
|---|
Future studies will have to address several important questions. The first is the relationship between the degree of maternal hypercholesterolemia and the increase in atherosclerosis in their offspring. So far, enough data for this purpose have been accumulated only for normocholesterolemic mothers and mothers who had or developed extensive hypercholesterolemia during pregnancy. Virtually no data are available about the effects of borderline or moderate hypercholesterolemia on lesion progression. A second focus of future studies should be the effects of maternal hypercholesterolemia on placental functions in order to better understand the pathophysiological mechanisms through which maternal hypercholesterolemia enhances fetal atherogenesis. A third important issue is the correlation between maternal hypercholesterolemia and clinical manifestations of atherosclerosis already discussed above. Patients with or without clinically established atherosclerosis-related disease should be compared with regard to their maternal cholesterol status. Finally, the benefits and risks of more effective hypocholesterolemic and antioxidant treatments should be thoroughly explored to exclude potential teratogenic effects.
|
ACKNOWLEDGMENTS
|
|---|
We dedicate this work to our mothers, Gertrud Palinski and Graciela Concha Molinari Napoli, for their lifelong love and encouragement. This review was supported by NHLBI grant HL-56989 and ISNIH grant 5825/2000.
|
REFERENCES
|
|---|
- Buja, L. M., Kita, T., Goldstein, J. L., Watanabe, Y., Brown, M. S. (1983) Cellular pathology of progressive atherosclerosis in the WHHL rabbit: an animal model of familial hypercholesterolemia. Arteriosclerosis 3,87-101[Abstract/Free Full Text]
- D’Armiento, F. P., Di Gregorio, F., Ciafre, S. A., Posca, T., Liguori, A., et al (1993) Histological findings and evidence of lipid conjugated dienes and malonyldialdehyde in human fetal aortas. Acta Paediatr 82,823-838[Medline]
- Napoli, C., D’ Armiento, F. P., Mancini, F. P., Witztum, J. L., Palumbo, G., Palinski, W. (1997) Fatty streak formation occurs in human fetal aortas and is greatly enhanced by maternal hypercholesterolemia: intimal accumulation of LDL and its oxidation precede monocyte recruitment into early atherosclerotic lesions. J. Clin. Invest. 100,2680-2690[Medline]
- Napoli, C., Witztum, J. L., de Nigris, F., Palumbo, G., D’Armiento, F. P., Palinski, W. (1999) Intracranial arteries of human fetuses are more resistant to hypercholesterolemia-induced fatty streak formation than extracranial arteries. Circulation 99,2003-2010[Abstract/Free Full Text]
- Palinski, W., Napoli, C. (1999) Pathophysiological events during pregnancy influence the development of atherosclerosis in humans. Trends Cardiovasc. Med. 9,205-214[CrossRef][Medline]
- Ikari, Y., McManus, B. M., Kenyon, J., Schwartz, S. M. (1999) Neonatal intima formation in the human coronary artery. Arterioscler. Thromb. Vasc. Biol. 19,2036-2040[Abstract/Free Full Text]
- Stary, H. C. (1989) Evolution and progression of atherosclerotic lesions in coronary arteries of children and young adults. Arteriosclerosis 9(Suppl. 1),19-32
- Berenson, G. S., Wattigney, W. A., Tracy, R. E., Newman, W. P., Srinivasan, S. R., et al (1992) Atherosclerosis of the aorta and coronary arteries and cardiovascular risk factors in persons aged 6 to 30 years and studied at necropsy (The Bogalusa Heart Study). Am. J. Cardiol. 70,851-858[CrossRef][Medline]
- . Pathobiological Determinants of Atherosclerosis in Youths (PDAY) Research Group (1993) Natural history of aortic and coronary atherosclerotic lesions in youth. Findings from the PDAY study. Arterioscler. Thromb. 13,1291-1298[Abstract/Free Full Text]
- Berenson, G. S., Srinivasan, S. R., Bao, W., Newman, W.P., 3rd, Tracy, R. E., Wattigney, W. A. (1998) Association between multiple cardiovascular risk factors and atherosclerosis in children and young adults. The Bogalusa Heart Study. N. Engl. J. Med. 338,1650-1656[Abstract/Free Full Text]
- Napoli, C., Glass, C. K., Witztum, J. L., Deutsch, R., D’Armiento, F. P., Palinski, W. (1999) Influence of maternal hypercholesterolaemia during pregnancy on progression of early atherosclerotic lesions in childhood: Fate of Early Lesions in Children (FELIC) study. Lancet 354,1234-1241[CrossRef][Medline]
- Stary, H. C. (2000) Lipid and macrophage accumulation in arteries of children and the development of atherosclerosis. Am. J. Clin. Nutr. 72,1297S-1306S[Abstract/Free Full Text]
- Ross, R. (1999) Atherosclerosis—an inflammatory disease. N. Engl. J. Med. 340,115-126[Free Full Text]
- Napoli, C., Ignarro, L. J. (2001) Nitric oxide and atherosclerosis. Nitric Oxide 5,88-97[CrossRef][Medline]
- Libby, P. (2001) Current concepts of the pathogenesis of the acute coronary syndromes. Circulation 104,365-372[Free Full Text]
- . Scandinavian Simvastatin Survival Study Group (1994) Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: the Scandinavian Simvastatin Survival Study (4S). Lancet 344,1383-1389
TOP
ABSTRACT
THE ATHEROGENIC PROCESS MAY...
