NHLBI WORKING GROUP
ON THE ROLE OF CHOLESTEROL AND LIPIDS IN
EMBRYONIC DEVELOPMENT AND CONGENITAL DISEASE

MAY 28, 1998

MEETING SITE: VANDERBILT UNIVERSITY, NASHVILLE TENNESSEE

MINUTES AND CONSENSUS RECOMMENDATIONS:

This document highlights discussions and presentations given by a group assembled by Dr. Lan Hsiang Wang of the National Institute of Health on May 28, 1998 on the campus of Vanderbilt University, Nashville Tennessee.  A major purpose of the meeting was to consider the potential role of cholesterol and lipids, their metabolism, serum carriers and cell surface receptors in normal and abnormal cardiac development.  Through a series of research presentations, ways were identified in which cholesterol and related lipids could affect development in general and the heart in particular.  A second purpose was to extend the discussion of cholesterol to ways in which new and important technological or conceptual opportunities might be used to further understand mechanisms of congenital heart disease.

STATEMENT OF RATIONALE FOR THE WORKING GROUP:

Cholesterol, and the synthetic pathway leading to cholesterol, are central to the formation and function of cell membranes and the synthesis of a number of steroid hormones.  Additionally, cholesterol is used in the synthesis of isoprene moieties that are used to posttranslationally modify proteins (e.g. farnesylation) such that the modified protein can bind to cell membranes, and as a result of membrane association, the function of the protein is changed.

The reputation of cholesterol has traditionally been compromised by its exacerbation of arterial intimal thickening into fatty atherosclerotic plaques.  However, while too much cholesterol may be deleterious, too little can engender birth defects.  An evolution in the understanding of an relationship between birth defects and cholesterol began with the recognition that the Smith-Lemli-Opitz syndrome (SLOS) was a cholesterol deficiency problem resulting from an autosomal mutation in a rate limiting enzyme required for the final step of cholesterol synthesis.  A spectrum of birth defects occur in SLOS including limb, heart, craniofacial and urogenital.  Another common developmental defect with some phenotypic similarities to SLOS is holoprosencephaly (HPE).  The HPE gene was recently found to be the morphogen, sonic hedgehog (Shh).  Knockouts of Shh in mice produced a head phenotype similar to both SLOS and HPE. Similarly inhibitors to cholesterol synthesis produced malformations in mice characteristic of both HPE and SLOS.  The link between Shh and cholesterol appears that the latter is used to covalently transport and anchor the morphogen to target cell receptors indicating a potentially important connection may exist developmentally between the signaling capacity of morphogens and cholesterol.

I. Discussion and Consensus Recommendations Related to Cholesterol and Its Potential Roles in Development:

1. Hedgehog Genes: An overview of the role of the hedgehog genes in mammalian development was presented. Hedgehog genes can act as intermediate and long range signaling molecules that function primarily in pattern formation.  Each is a highly conserved protein which two domains: a signaling domain and an autocatalytic domain. Cholesterol insertion into a binding site of the catalytic domain appears to be required for its enzymatic activity. Cholesterol also binds covalently to the signaling end of hedgehog protein and tethers the morphogen to the cell membrane.  How hedgehog proteins are able to reach their receptors while tethered to cholesterol is unknown.  The key point is that cholesterol modification of the hedgehog class of signaling morphogens is critical to their function.  It is proposed that cholesterol binding of hedgehog is involved in transport into target cells or out of a cell that secretes it.  One intriguing possibility is that hedgehog is simply shed from the surface where it can subsequently serve as intermediate or long range signaling molecule.
   
   Of the three major members of the hedgehog class of morphogens, Sonic hedgehog (Shh) has been most investigated.  It is expressed primarily in the notochord, CNS, gut, limbs, lung and tooth epithelium.  Shh binds to a receptor, patch, which leads to a conformational change in a binding partner, smoothened, that possess kinase signaling activity.  Another downstream target is Gli which encodes a basic helix-loop-helix transcription factor that acts as positive regulator Shh signaling.
   
   Shh has been shown to directly play a role in patterning of the limb and CNS.  It is expressed by the notochord and ventral floor plate where the morphogen appears to function as a ventralizing factor in the formation of neurons in the developing neural tube.  In the limb, Shh is expressed in posterior mesoderm of the limb bud where it conveys the anterior-posterior polarizing activity traditionally associated with the ZPA (zone of polarizing activity).  The expression of Shh helped to open the door to an understanding of molecular asymmetry where its left-sided expression at Hensen's node was linked to the downstream expression of nodal (one of several members of the TGF supergene family associated with breaking symmetry). The role of Shh in regulating laterality may involve a downstream target, ptx-2, frequently used by other genes also expressed asymmetrically.  For example, ptx-2 is primarily expressed in the left heart forming field. What is unclear is why knockouts of Shh do not cause asymmetry.
   
   In the developing lung primordium, Shh is expressed in the epithelium derived from endoderm. This epithelium interacts with adjacent mesenchyme to promote branching which in Shh knockout mice was severely restricted.  These findings suggest that Shh secreted by the epithelium may induce adjacent mesenchyme to secrete factors that promote branching.  This observation suggests an inductive paradigm that may prove useful for future investigations of other epithelial mesenchymal interactions.
   
   Desert hedgehog (Dhh) appears to be the only known hedgehog gene to be expressed in the heart (AV canal).  Its function in heart development is unknown.  The major known function for Dhh appears to be in spermatogenesis; within the testis, Sertoli cells appear to be the principal site of expression.  Knockouts of Dhh result in small testes producing infertile sperm (females appear normal). Additionally, Leydig cells differentiation appears abnormal leading to decreased androgen production.
   
   Indian hedgehog (Ihh) appears to be required for growth, not patterning.  It is expressed in the hypertrophic region of cartilage, particularly in the growth (epiphyseal) plates.  Knockouts lead to dwarfism.
   
   
2. SLOS and Other Inborn Errors of Cholesterol Biosynthesis: The recent discovery of abnormal cholesterol biosynthesis in children with SLOS was discussed by many attendees because it clearly demonstrated the need for cholesterol in development.  In the three documented inborn errors of cholesterol synthesis, mevalonic aciduria, desmosterolosis and SLOS, each results in dysmorphic phenotypes that range in severity, mental retardation and growth disorders. These defects emphasize the intrinsic role of cholesterol in development and, even though frequency of these syndromes is on the order of 1 in 20,000 (highest for SLOS), it raises the question of whether hypocholesterolemia occurs in other birth defects involving similar dysmorphogies and neurological or growth disorders but whose etiology is virtually unknown.  It is possible that there remain inherited deficiencies of other enzymes of cholesterol biosynthesis that have not been identified.
   
   In analyzing the phenotype of the three disorders, it is of interest that mevalonic aciduria which is a proximal defect of sterol metabolism and does not result in cholesterol deficiency, also does not present with significant malformations, while SLOS and desmosterolosis which are distal defects and result in deficiency of cholesterol are associated with significant malformations of all organ systems.  Earlier animal studies using pharmacological inhibitors of cholesterol synthesizing enzymes revealed similar anomalies to those seen in the human syndromes. Some of these teratogenically induced anomalies could be prevented by cholesterol supplementation to the mother.  These data raised questions as to whether an adequate amount of cholesterol is needed early in development and may be more important with regard to the causation of anomalies than the presence of the proper cholesterol precursors.  A consensus of opinion was that future studies of investigation should include offering sterol analysis to patients with developmental delays or mental retardation, particularly of the heart, limbs, palate and genitalia. Similarly, tissue of fetuses with intrauterine demise and stillborns should be analyzed to further define the spectrum of hypocholesterolemia in human development.

3. Cholesterol, Protein Lipidation and Intracellular Signaling: The cholesterol synthetic pathway gives rise to intermediates termed isoprenes that are used to lipidate proteins which frequently leads to modifications in the function of these proteins.  One question discussed by the group was how changes in the cholesterol pathway could affect isoprenylated proteins associated with intracellular signaling.  One answer is the outcome observed for the farnesylation (a form of isoprenylation) of ras in cardiac muscle.  Farnesylation of ras increased its association with cell surface membranes.  This association between membrane and lipidated-ras was found to be reversible and essential for ras activity.  Farnesylation has been suggested to be a regulatable process in embryonic chick myocytes. Normally, chick myocytes contain a predominance of unfarnesylated ras in the cytoplasm, however, when myocytes were cultured in lipoprotein-depleted serum, cholesterol synthesis increased and ras was farnesylated and shifted the membrane.  Movement of ras to the membrane lead to the activation of genes associated with response to autonomic nerve signaling.  Thus, the observation with ras indicates that increased cholesterol metabolism may alter the posttranslational modification of a protein that mediates intracellular signaling.
   
   Other examples of isoprenylated proteins include nuclear lamins, rho protein, rab proteins, and cdc42.  All these are among a class of proteins involved in G protein signaling, cytoskeletal rearrangements and vesicle transport.  The understanding of the regulation of the enzymes involved in modification of these proteins, the involvement of these proteins in developing heart cells, and the effect of flux through the cholesterol synthetic pathway all need to be considered. A consensus of opinion was that studies targeted at determining the role of protein lipidation and isoprenylation as a mechanism to regulate cell differentiation should be a high priority. These processes are likely to be critical regulators of embryogenesis by establishing or altering growth factor responsiveness in developing cells.
   
   
4. Lipoprotein Secretion and Development: As carriers of cholesterol, apolipoproteins potentially should play an important role in development. Indeed, knockouts that variably reduced levels of ApoB have convincingly shown that lipoproteins are essential for development, particularly of the CNS and the GI tract.  In many ways, the phenotype of variable ApoB knockouts resembled SLOS.  A reduction of 25-30% in lipoprotein was sufficient to engender maldevelopment.  In ApoB -/- mice, all ApoB synthesis was blocked; most embryos died by day 12.  The primary cause of death appeared to be arrested yolk sac formation.  As a consequence, lipid transport to the developing embryo was deficient and cholesterol stores were much reduced.  One hypothesis proposed was that any gene that might inhibit lipoprotein production by the yolk sac would lead to developmental abnormalities.  As a test of the hypothesis, a gene which is expressed in the yolk sac and is required for apoB-lipoprotein assembly was knocked out.  Mice homozygous for this gene [microsomal triglyceride transfer protein (Mttp)] died around day 10.5; their yolk sacs were pale compared to wild.ypes owing to an apparent reduction in blood island formation.  These and other results indicate a link between lipoprotein synthesis and hematopoiesis.
   
   Information was presented that differences in apolipoprotein secretion and esterification of cholesterol occur in adult Down's patients which, in some unknown way, protect them from atherosclerosis.  It is also well known that Down's patients are high risk for Alzheimer's disease possibly due to a gene, amyloid precursor protein, expressed on chromosome 21.  When this is combined with the recently described association of apolipoprotein E 4 allele with increased risk of Alzheimer's disease, the implication for future study is that novel insights into mental retardation and cardiovascular disease could come from studying cholesterol and lipoproteins in children born with trisomy 21 or in animal models of Down syndrome (e.g. murine trisomy 16).
   
   A consensus opinion was that future studies were needed to determine if yolk sac lipoprotein secretion is needed in humans as well as mice, and, if so, whether it is important simply as means of delivering triglycerides, fat soluble vitamins, or cholesterol, or if apolipoproteins are required for morphogenesis in some other way like hematopoiesis or normal vasculogenesis.  For example, do serum lipoproteins have a mitogenic effect that disrupts normal growth patterns of cells critical to vascular and heart development, like the neural crest.  Another future direction would be to explore the developmental role of receptors for lipoproteins. Virtually nothing is known regarding these receptors during development and if the primary role of these receptors is merely to take up cholesterol or lipids or if binding of ligands to these receptors mediates intracellular signaling.

II. Programmatic Recommendations:

A series of recommendations with some supportive justification was also made regarding future directions for research in congenital cardiovascular diseases with or without implication for cholesterol or lipids.  These are briefly listed below.

1. Human Molecular Genetics: As exemplified by SLOS, human genetic models (syndromes) provide a powerful reminder of the critical importance of molecular genetics in understanding complex birth defects.  For example, mutations in proteins and enzymes encoded by mitochondrial or nuclear DNA that are essential for cardiac energy production can cause sudden death in infants, dilated cardiomyopathy and hypertrophic cardiomyopathy.  Similarly, single gene defects are now being identified for specific forms of congenital cardiovascular disease.  Atrial septal defects have been linked to mutations in transcription factors such as Tbx5 (Holt-Oram syndrome) and tinman (Nkx-2.5).  Vasculopathies secondary to mutations of fibrillin-1 (Marfan's syndrome), elastin (Williams' syndrome and supravalvular aortic stenosis), tenascin (Ehlers-Danlos syndrome) and jagged (Alagille's syndrome) have been defined.  Also identified are loci for a dominant form of atrial septal defect, AV septal defect and total anomalous pulmonary return.  In animal models, mutations in right-left dynein have been linked to heterotaxy syndromes, asplenia and polysplenia.  Genes have been traced to specific loci on chromosome 21 which when duplicated as in Down's syndrome engender AV septal defects and conotruncal malformations.  Specific candidates at these loci have recently been identified such as DS-CAM (Down syndrome cell adhesion molecule).  Thus, because it is now clear that a single gene modification can result in life-threatening heart or vessel defects, it is recommended that as the human genome project matures, both mapping and gene localization studies should be encouraged.  In turn, such studies should be used to develop animal models in which the mechanisms of human congenital heart disease can be studied morphologically and physiologically.

2. Cellular Mechanisms: As the genes with relevance to heart disease become identified, the question then becomes how the encoded proteins affect heart development.  If knockouts or knockins produce phenotypes indicative of functional significance, the challenge then becomes to determine the morphogenetic mechanism by which the missing, truncated or overexpressed protein actually modified development.  The problem becomes even more daunting if lethality occurs early in development. Several centrally defining, morphogenetic mechanisms were proposed that should be studied.  These were epithelial mesenchymal transformations and neural crest migration and signaling.  The former is the process by which cushion tissues form and fuse to form valvuloseptal primordia.  In particular, the inducers and mediators of transformation processes and the transcription factors that regulate them were highlighted for study. While neural crest are widely recognized for their role in septating the aortic sac into the roots of the aorta and pulmonary trunk, their role within the outlet region of the heart tube is unknown. What is clear is that two prongs of neural crest migrate into the truncus and conus regions of the outflow tract of the heart from the septum which divides the aortic sac.  If prevented from reaching the outflow tract following experimental ablation, Pax3 mutations (Splotch mouse) or monosomy of 22q11 (DiGeorge syndrome), misalignment of conus septum with the ventricular septum occurs (e.g. double outlet right ventricle). Yet nothing is known about the morphogenetic mechanisms by which neural crest might influence septal alignments.  It is recommended that genes known to produce misalignments (e.g. TGF beta2) or cushion defects (DSCAM) be evaluated for potential roles in epithelial transformations or cell migrations etc.  Rescue of cellular or tissue phenotypes should be considered in future studies as the ultimate assay of gene function.
   
   A strong recommendation was also made for the use of non-mammalian animal models.  It was noted how extensively the field of cardiovascular developmental biology has been advanced by studies in xenopus, zebrafish, C. elegans and fruit flies.  That effort could be enhanced by continued development of experiments using these systems.  The development of organisms, particularly like Xenopus, provide high throughput screening systems for genes and gene activities and may also provide a sensitive way to examine potential mechanisms of oligonucleotide mediated gene conversion and gene control.
   
   
3. Technological Considerations: Several opinions were expressed that technologies for the analysis of genetically altered animal models had not kept pace with the very rapid increase in the production of animals with mutant phenotypes.  Careful classification and characterization of phenotypes is essential for ultimately mapping the genes responsible for normal and abnormal development.  Researchers who are well trained at generating animal models are not often well trained to assess the results of their labors.  To this end, several ideas were proposed.  Many of these concerned electronic mechanisms (bioinformatics) to store and assemble data.  It was suggested that such data could enhance development of mathematical models that could lead to a better understanding of the biophysics of developmental processes.  Also the development of imaging modalities that include both live as well as fixed tissue observations were urged.  The use confocal microscopy in a variety of modes including the tunneling microscope was seen as a way to image large amounts of tissues, even older whole embryos.
   
   A major consensus was expressed for a center for Magnetic Resonance Imaging.  MRI generates non-distorted 3-dimensional data of either fixed or living embryos that can be manipulated interactively in real time and in a fraction of the time required for traditional optical imaging and allows for creation of visual models for fast and easy interpretation of complex data. Moreover, using multiple pulse sequences, tissues with inherent differences in contrast can be readily visualized without staining and even if differences are slight quantification of image densities is possible. With use of targeted contrast agents, the potential for increased resolution is enormously expanded.  It was recommended and endorsed by consensus that to enhance the effectiveness of studying growth and development of living and fixed embryos that dedicated centers of magnetic resonance microscopy (MRM) be established that can provide 3-dimensional quantitative techniques, in utero imaging techniques and finally targeted mechanisms for use of contrast agents.

CONCLUSION AND SUMMARY RECOMMENDATIONS:

There was consensus that the discussions on cholesterol and its relationship to embryogenesis be forwarded to NIH staff as a recommendation for a potential future RFA whereas other recommendations were proposed as potential areas NIH staff may wish to consider as relevant to programmatic issues on the study of congenital heart disease.

Last Updated April 2011