| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Review |
Institute for Nutrition Research, School of Medicine, University of Oslo, Oslo, Norway
Published, JLR Papers in Press, August 16, 2002. DOI 10.1194/jlr.R100015-JLR200
1 To whom correspondence should be addressed. e-mail: rune.blomhoff{at}basalmed.uio.no
 |
ABSTRACT |
|---|
 TOP ABSTRACT BackgroundConstructing a classification...The classical RA pathwayLiterature reviewsMETHODSRESULTS AND DISCUSSIONEndnoteREFERENCES |
|---|
We found 27 genes that are unquestionably direct targets of the classical pathway in permissive cellular contexts (Category 3 genes), plus 105 genes that appear to be candidates, pending the results of specific additional experiments (Category 2). Data on another 267 targets are not evocative of direct or indirect regulation either way, although control by retinoic acid through some mechanism is clear (Category 1). Most of the remaining 133 targets seem to be regulated indirectly, usually through a transcriptional intermediary, in the contexts studied so far (Category 0).
Supplementary key words gene regulation • transcription • retinoic acid receptors • tretinoin • RAR • RXR
|
Background |
|---|
|
TOPABSTRACTBackgroundConstructing a classification...The classical RA pathwayLiterature reviewsMETHODSRESULTS AND DISCUSSIONEndnoteREFERENCES |
|---|
It is now known that RA can influence gene expression and protein production in many ways, but in terms of molecular mechanisms, a single, predominant, classical pathway has emerged: all-trans retinoic acid plus a dimer composed of a retinoic acid receptor and a retinoid X receptor (an RAR.RXR dimer) and a more or less regular DNA response element. In this paper, genes that respond through this pathway are called "direct" targets of the classical RA pathway; those that respond to RA through other molecular mechanisms, but do respond, are called "indirect" targets. Since Blalock and Gifford's paper nearly a quarter century ago, more than 532 genes have been put forward as regulatory targets of RA; and while the distinction between direct and indirect regulation is now well entrenched, it is not necessarily germane to every study. Nevertheless, a great deal of suggestive data has been generated and it can be used to construct a tentative classification of RA's targets along these lines.
|
Constructing a classification table |
|---|
|
TOPABSTRACTBackgroundConstructing a classification...The classical RA pathwayLiterature reviewsMETHODSRESULTS AND DISCUSSIONEndnoteREFERENCES |
|---|
It should be stressed that the numeric designations used for the categories are nothing more than tags. With a very few exceptions (which always clearly marked), the Category 0 genes are regulatory targets of RA every bit as much as Category 3 genes. They are simply regulated in different ways. Category 1 and Category 2 genes are also targets, although current research does not allow us to conclude quite so much about the mechanisms employed in these cases. Emphatically, the classification does not mean to impugn the work reported in the any of papers considered. The distinction between direct and indirect regulation is not necessarily relevant to many valid research goals, and a great deal of valuable work has been done in clinical, developmental, and basic science without addressing these questions even obliquely.
Of necessity, the Gene Table is long and complex. However, the genome projects, various proteomic studies, and the preliminary gene ontologies produced over the last few years have made it clear that work on some very interesting biological questions will require dealing with vast amounts of data. Gene expression regulation by RA encompasses a number of such questions and a compilation like the Gene Table would seem to be an economical way to approach some of them.
|
The classical RA pathway |
|---|
|
TOPABSTRACTBackgroundConstructing a classification...The classical RA pathwayLiterature reviewsMETHODSRESULTS AND DISCUSSIONEndnoteREFERENCES |
|---|
It might seem arbitrary, uninformative, or unnecessarily stringent to restrict "direct" regulation to the classical RA pathway and to consign all other regulatory modalities to the catch-all category, "indirect" regulation. However, each alternative regulatory pathway represents a distinct type of genetic event. Perhaps each deserves its own Gene Table. We chose the classical RA pathway as a branch point in the present work, i) because of its preeminent historical position, ii) because the distinction between direct and indirect regulation through this pathway is well established and frequently studied, and iii) because many suggestive and highly relevant studies are available, even though questions of molecular mechanism are not necessarily raised in them.
The Gene Table is intended to cover every gene now known to be regulated by retinoic acid. The last attempt at delineating a complete set of such genes was published by Chytil and Raiz-ul-Haq in 1990 (3). They listed more than 125 proteins that we now take to be monogenic, plus a number of other proteins of less clear provenance. Gudas et al. took a slightly different starting point 4 years later, and wrote detailed descriptions of most RA targets known at the time. They categorized them primarily along functional or homology lines (4).
|
Literature reviews |
|---|
|
TOPABSTRACTBackgroundConstructing a classification...The classical RA pathwayLiterature reviewsMETHODSRESULTS AND DISCUSSIONEndnoteREFERENCES |
|---|
The retinoid receptors are members of a much larger group of transcription factors, the so-called nuclear receptors. An encyclopedic overview of this large and important class of proteins is Gronemeyer and Laudet's 1995 monograph (16). It remains invaluable even though its publication preceded some of the more recent work on co-regulators, intermediary factors, and the chromatin connection. For an update in those areas, see Rosenfeld and Glass (17). Chawla et al. (18) recently reviewed the connection between the nuclear receptors and lipid physiology, and both RARs and RXRs play roles in this. Finally, two collections of particularly noteworthy reviews appeared in the mid-1990s: one covering various aspects of the nuclear receptors and the other, various aspects of the retinoids. See (19) and (20), respectively, and the articles that accompany them.
|
METHODS |
|---|
|
TOPABSTRACTBackgroundConstructing a classification...The classical RA pathwayLiterature reviewsMETHODSRESULTS AND DISCUSSIONEndnoteREFERENCES |
|---|
Although we made every effort to identify and follow up on "novel" genes identified in differential display-type experiments, we have not included any genes so totally uncharacterized that they have not yet even been named. See (21) for some examples. Nor have we included fragments so far identified only as ESTs. See (22) and (23) for examples of these.
An analysis of this sort would ideally be limited to work done in "normal" cells or individuals; the activities of RA and its receptors in aberrant cell types would then be handled separately as exceptions. We have tried to do this up to a point. Work on cells that have suffered catastrophic DNA events that are likely to have affected RA's activity, certain viral integrations, extraordinary recombinations, engineering experiments, and the like, have been excluded except to make occasional special points. In particular, work on acute promyelocytic leukemia (APL) cells, which generally express oncogenic RAR
fusions, have been largely excluded on this ground. Nevertheless, a great deal of research has been done on RA's activities in APL cells and we refer the reader to (24) for a review. Of course, many common cell lines contain genomic anomalies that are not likely to have affected RA's activity overall: HepG2 and Caco-2 lines, for example. For the purposes of this work, such cell lines are considered normal.
As a rule, we did not consider experiments in which RA was used in conjunction with another treatment, although we tried to take note of any controls using RA alone. The exception to this is where some form of external "activation" seems to be required for any expression of the target gene, for example, the interleukins. It should be stressed that by excluding combo-treatments we automatically ruled out many studies using RA plus cAMP (or RA plus cAMP and theophylline) rather than RA alone. We did, however, consider these experiments if they confirmed points suggested elsewhere by RA alone. This is an admitted limitation of the present work, but the complexity of regulatory interactions in these cases is still overwhelming.
Constructing a database of papers and genes
Using various free text and MeSH (Medical Subject Headings) strategies at the United States National Library of Medicine's PubMed gateway, we created a database of more than 4,000 papers relevant to the regulation of gene expression by retinoic acid. We identified the gene or genes considered in each paper, and, based on abstracts, selected what appeared to be the most relevant studies for each gene. Using this set of abstracts and the associated MEDLINE coding, we determined which species had been investigated, located the gene's official name at LocusLink (25), and performed supplementary searches based on official nomenclature, curated aliases, and any novel names or aliases applied to orthologs. This process was iterated as necessary, and eventually led to a list of relevant papers for each gene. These entries were then re-evaluated at the abstract level and the most promising papers (for our purposes) were gathered and consulted for data, discussions, and further citations. New candidate genes went through the same process as they turned up. By the end of the project, nearly 8,000 papers (not including reviews) had been considered to one degree or another.
For each gene, we then studied the scientific evidence presented in the selected papers and evaluated the degree to which a direct regulatory pathway had been demonstrated, suggested, or brought into question. This information was distilled into several short standardized phrases and incorporated into the Gene Table, along with species information, any alternative names and symbols used in the selected studies, and references to the most essential papers.
Concordance of working and official gene names
Most genes have several names. By "official nomenclature" we mean names and symbols approved by (or pending before) the Human Genome Organization Nomenclature Committee, the Mouse Genome Informatics Nomenclature Committee, the International Rat Genetic Nomenclature Committee, or the Zebrafish Nomenclature Committee. We have followed official nomenclature whenever possible. This can be confusing when the official name of a gene is either uninformative, uncommon, or simply designed for a purpose that is not one's own. For example, most readers probably would not recognize Nr2f1 as the name of the gene that encodes COUP-1. However, while understandable from a historical perspective, the proliferation of trivial names (for both genes and proteins) has been scientifically unhelpful and using official names solves the problem. The lists of alternatives and aliases kept by the nomenclature committees and at LocusLink should quickly resolve any questions.
It is not always easy to determine which gene has been studied in a given paper, or which papers deal with the same gene; and this is not limited to older papers. It can be particularly problematic when several species, or several apparently unrelated scientific questions, have been studied in different papers. In a number of cases, we had to align published primer sequences with groups of homologs, follow LinkOuts to cited sequences at the National Center for Biotechnology Information's Entrez system, or even BLAST nucleotides strings taken from journal figures.
As a rule, the Gene Table uses the gene symbol from the species discussed in the earliest paper cited; when no approved, pending, or interim name was available for the gene in that species, we generally chose the mouse version. The nomenclature committees try to keep symbols and leading phrases invariant over vertebrate species (except for orthographic differences) so this is little more than a matter of choice. In order to save space, only symbols, not full names, are used in the first column of the table.
Trivial names from cited papers
The second column of the Gene Table, "Name in refs," lists only the gene or protein designations used in the papers cited. It does not include other aliases, no matter how common they may be in the literature. Table 1 provides a concordance between these working names (or abbreviations) and the symbols used in the Gene Table, but only for cases where the two are very different. Throughout, we have suppressed the distinction between genes and the proteins they encode.
|
The regulatory directions column
For each gene, we have noted the predominant regulatory direction attributed to RA, up or down. This can be problematic in situations where, intuitively, RA can effect opposite actions in different cellular contexts: up during differentiation, for example, and down during growth inhibition. Again, we concentrated on what was most frequently reported. Genes are marked ‘vrs’ (various) when there is no obvious predominant direction. For all such genes, it should be clear from our comments whether the category rating is based on a single regulatory direction or on the data taken as a whole. For example, there are several clear demonstrations that the rapid down-regulation of Myc is indirect in the cell types in which this has been investigated. This seems likely to apply whenever Myc is down-regulated. Its rapid up-regulation in other contexts, however, has not convincingly been shown to be indirect anywhere. Myc's Category 2 rating therefore refers to its rapid up-regulation following a moderate dose of RA in certain situations. The comments column should make this clear. Everything in the table is based on currently available data, of course, and as additional contexts are studied, more cell types, different developmental stages, unusual environmental situations, and so forth, the picture will only get more complex.
Stock phrases used in the summary column
Every phrase in the Gene Table and every category rating should be read with the implicit qualification, "in the cell types or at the developmental stages studied." Even the paradigm of classical RA regulation, RARß, is not under RA control at all times or in all cell types. To keep the table as concise as possible, and to make comparisons easier, we used the following stock phrases when applicable: 1) "No good d/t data" means we found no experiments using dose and time conditions within our limits for suggestive data. The phrase does not impugn the work referred to but was chosen for its brevity. In particular, minute-by-minute observations using physiological doses of ligand are only rarely relevant in clinical research or developmental work. In fact, pharmacological doses may be the only effective therapies in certain clinical situations and teratological doses have been indispensable in some truly seminal developmental studies. In ordinary circumstances, however, it is generally assumed that direct transcriptional modulation is rapid and that it can be initiated with a physiologically moderate dose of ligand. Ideally, unless there is a transport problem, one would like to see experiments using nanomolar concentrations of RA and making observations within minutes. However, the number of experiments conforming to these standards is very small, so we set 1 µM x 6 h as the upper limit for "suggestive" data. This was a necessary compromise given the range of scientific questions addressed in the papers consulted. 2) "Specific ligands" refers to either receptor-selective ligands or ligands that do not have the full complement of biological effects associated with all-trans retinoic acid (for example, ligands that help sort out AP-1 events). 3) The phrase "functional binding site" implies that a whole range of thoughtfully-designed tests has been performed, and that a more or less recognizable response element has been identified. The phrase is distinguished from such other notations as "functional motifs" (for which no dimer binding or native transcriptional verification has been made), "binding sites" (from which heterologous promoters can be driven), "motifs" (which are supported by sequence analysis only), and so forth. The phrase "no motif found" says that a promoter or other presumed control region was inspected in at least one of the papers cited, but that no candidate motif was found. 4) "Other NRs" indicates that other nuclear receptors are known to be involved in the gene's regulation in some cells. The importance of noting this stems from the cross-talk that can occur between nuclear receptors, and from the similarity of nuclear receptor binding sites (which can be confounding when extreme dose conditions are used). 5) "During differentiation" (or a similar phrase) indicates that the gene has only been studied during differentiation, growth control, proliferation control, cell cycle arrest, apoptosis, wound healing, hypertrophy, or any of the other wholesale cellular or phenotypic changes RA can effect. We did not always include such an annotation. 6) " ... not for RA ... " or " ... not for RA alone ... " means that the referenced experiments, or parts of them, have been done with ligands other than all-trans RA, usually 9-cis or a synthetic, or with RA plus an additional factor. 7) "d/t borderline" signals that while at least some data fall within our dose/time limits, they are right on the borderline. This is meant to draw attention to the compromise inherent in the limits imposed for "suggestive" data. 8) "Probably indirect" is more specific than it sounds. It indicates that a transcriptional intermediary, as opposed to another indirect mechanism, is most likely involved: RA regulates X and X regulates Y. The particular intermediary is noted in some cases.
The citations column
We have attempted to evaluate RA's role in the control of 532 genes and could not possibly cite every relevant paper. Each paper we do cite makes a point directly connected to the Gene Table: a first assertion of RA control, a regulatory direction, a time or dose curve, a binding site, a species, or something else. In addition, we have cited a (very) few papers of particular historical importance even though the research described may have preceded the experimental techniques or genetic models that underpin today's RA work. To save space, we have used PubMed Unique Identifiers (PMIDs) rather than traditional citations.
PMIDs are the unique record numbers assigned to journal articles at the National Library of Medicine. They can be used in PubMed, National Library of Medicine Gateway, and other National Institutes of Health databases to retrieve citations, abstracts, cross-links to GenBank sequences or other sequence-based information, external links to full-text articles where available, and so forth. Unmodified identifiers are valid queries, type or paste the number(s) into the Search Box, at all appropriate National Library of Medicine front ends, but in complex queries or in other databases the tag "[PMID]" may be required.
The category ratings column
The ratings reflect overall assessments. Experimental evidence varies from gene to gene and there is no algorithm that can assign a category automatically. Investigators use different techniques and have different scientific questions in mind; the quality of figures varies, and the threshold of "proof" varies from lab to lab. For each gene then, the rating expresses our overall reading of the evidentiary situation based on all the work considered. Again, not all of the studies were designed to investigate mechanisms, so we are imposing extrinsic considerations in some cases.
Category 0. There is no particular reason to believe that this gene is directly regulated through the classical RA pathway.
Case 1
Indirect regulation has been demonstrated in a context that seems likely to apply generally and no other data suggest that direct regulation is likely in other contexts. Indirect regulation can include the existence of RA-regulated transcriptional intermediaries, non-transcriptional or post-transcriptional effects, and so forth.
Case 2
Hexamer motifs have been found in a location that might represent a regulatory unit, but no other evidence of RA involvement has been offered in any paper we know of.
Case 3
An historical correction has been made and the gene is no longer thought to be under RA control.
Category 1. There is solid evidence that the gene is controlled by RA and no indirect mechanism has been demonstrated experimentally. At the same time, the available data do not justify a prediction, or even suggest which way a prediction should go: direct or indirect regulation.
Case 1
Induction or suppression has been shown, but the dose and/or time conditions exceed our limits for "suggestive" data.
Case 2
Physiological, clinical, or dietary information (or evidence from transgenics, knock-ins, or knockouts) strongly implicates RA, but there is no particular reason to posit direct regulation through the classical RA pathway.
Case 3
mRNA studies are lacking but protein studies or other evidence suggests that further work should be done.
Category 2. The gene is a strong candidate for direct regulation, but specific data are lacking.
Case 1
Transcriptional effects have been demonstrated under suggestive dose and time conditions but i) no binding site connection has been made, or ii) the involvement of an RAR.RXR dimer is not clear.
Case 2
There is highly promising binding site information plus basic inductive or suppressive data.
Category 3. A persuasive case has been made, or can be made based on currently available data, that the gene is directly regulated by RA in at least one genetically "normal" cell type.
Requirement 1
Transcription-based induction or suppression (within the limits of 1 µM or less x 6 h or less) has been confirmed in some reasonably general context.
Requirement 2
Evidence of RAR.RXR involvement has been produced or strongly implied.
Requirement 3
A functional binding site, preferably conserved, has been found and tested in a broad panel of experiments.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTBackgroundConstructing a classification...The classical RA pathwayLiterature reviewsMETHODSRESULTS AND DISCUSSIONEndnoteREFERENCES |
|---|
Category 0 encompasses two cases. First, there are 124 genes that seem to be regulated indirectly in the contexts studied. We are aware of no data or arguments suggesting that these genes might be directly regulated through the classical RA pathway in other cellular contexts. Nine other genes (Adh1, BTK, FSCN2, Htf9c, IBSP, Itgb7, Lpl, Ranbp1, and Slc9a2) were also put into Category 0. They are discussed in the literature, but there is no strong reason to believe that they are regulated by RA at the transcriptional level. In two cases, Adh1 and Lpl, suspected or predicted mRNA changes were not confirmed, and while most of the others contain motifs resembling RA response elements, there is no evidence suggesting that these motifs, which can be highly ambiguous in the best of circumstances, represent biologically active retinoic acid response elements.
The remaining 267 genes, slightly more than half of those we evaluated, fall into Category 1. They are regulated by RA in some way, but the data available at present do not allow us to predict direct versus indirect control. Most have not yet been studied except in long-term or high-dose contexts, and for many, the ultimate interest has been clinical, developmental, or diagnostic rather than mechanistic. Additional work will need to be done to push these genes into more informative categories.
In fact, future research may change the classification status of any gene in the table. The method used to select Category 3 genes was designed to be sufficiently rigorous that no gene would easily be struck from the group, but there is no reason why any one of them might not be regulated indirectly in other contexts as well. Beyond that, we expect future research to find that many of the Category 2 genes are direct targets, and that some of the Category 1 genes are as well. In fact, some of the Category 0 genes may turn out to be direct targets too, but in contexts that have not yet been studied.
Regulatory direction
In terms of regulatory direction, 311 genes are always or almost always up-regulated in the contexts studied, 109 are always or almost always down-regulated, and the rest are quite variable. Most investigators now believe that direct regulation through the classical RA pathway is always inductive, although there is no theoretical reason why this should be so (and it is not true of some other transcription factors). Nevertheless, all the Category 3 genes are up-regulated and only three of the Category 2 genes are usually down-regulated. One Category 3 gene, Hoxb1, is marked "various" because it can be directly up-regulated in some contexts, but down-regulated, probably indirectly, in others. Given that many transcriptional events seem to be regulated cyclically, a "various" regulatory direction should probably be much more common than the data imply; most likely this is due to a lack of measurements taken along a fine enough time continuum. Table 2 summarizes category and direction data for the 532 genes. (The reader is reminded that gene expression in the presence of RA is the topic here. The repression of basal transcription by RAR.RXR in the absence of RA is an entirely different matter.)
|
Although no regulatory or evolutionary theory formally justifies it so far, it is tempting to see a certain logic in several other genes directly regulated by RA: HSD17B1 is involved in the function of other nuclear receptors; H1F0 is activated at differentiation and points of development; one of SFTPB's functions is developmental; IL2RA is involved in apoptosis; Ucp1 is expressed only in brown adipose tissue (and is therefore connected to dietary lipids); ETS1 ultimately derives from the E26 virus (and a number of viral control regions contain sequences that can respond to RA); Foxa1 and Egr1 are expressed early in differentiation. The other Category 3 genes are CD38 (which was originally identified as a differentiation antigen), Tgm2, and Pck1.
We found 105 Category 2 genes that can be more or less rapidly up- or down-regulated at the transcriptional level in the presence of RA. Some of these genes are probably regulated directly. It would be surprising if there were a common thread among them, and there is not. They encode proteins of almost every imaginable type.
However, several domain architectures turn up a number of times among the Category 2 and 3 genes and should probably be mentioned. Taking the 132 genes in these two categories together, 11 contain homeobox domains (Cdx1, GBX2, Hoxa1, HOXA4, Hoxb1, Hoxb4, Hoxd4, LHX1, Meis1, NCX, and Pit1) and six encode zinc finger proteins (NR2C2, NR4A3, RARA, RARB, RARG, and Egr1). Of those six, five are nuclear receptors with both c4 zinc finger domains and nuclear receptor ligand-binding domains. Five of the genes in the two categories are from the lipocalin/cytosolic fatty-acid binding protein family (APOD, Crabp1, CRABP2, Rbp1, and RBP4); and five contain tyrosine kinase catalytic, or eukaryotic protein kinase, domains (CSF1R, EGFR, LYN, Tgfbr1, and Tgfbr2). Three of the genes encode helix-loop-helix DNA-binding domain sequences (MYC, MYCN, and Srebf1); three encode short chain dehydrogenases (HSD17B1, HSD17B2, and RDHL); and three contain TGF-ß propeptide domains (Ebaf, Gdf5, and Tgfb3).
State of the science
Intuitively, the number of Category 3 genes found in this work is surprisingly small, given the conservation of three RAR genes plus a triad of RXRs and multiple isoforms of all. The largest cohesive group of Category 2 or 3 genes consists of those somehow connected to the retinoids or nuclear receptors, the "infrastructure" of the regulatory system itself. And while evolution may not be particularly parsimonious, one suspects that the machinery of the classical RA pathway with all its complexities and autoregulatory loops has been conserved, not to regulate itself, but because it is uniquely useful in controlling, directly or indirectly, a particular range of genetic events in various cells and at different times of life. This suggests that the group of Category 3 genes will grow as new data become available on genes already in the table, and as new targets are discovered. There is circumstantial evidence for this, too. Since at least the mid-1980s, subtraction or differential-display experiments using RA have been turning up "novel" genes and there is no sign that this is slowing down. Many of these genes have not been investigated beyond the original paper mentioning them, and most are probably cases of indirect regulation. Nevertheless, this adds an exciting dimension to the RA field and points to quite a few experiments waiting to be done.
In works that deal with a large number of genes, it has become customary to summarize functions, family memberships, and other quiddities, "ontologies" as they are now called in a puzzling use of the word. This is done as a first step in finding underlying biological regularities, and we have done it for that reason in this paper. However, its significance should not be overplayed. Duplications of whole genes, coding plus regulatory and non-coding regions, do not endure evolutionary time unchanged, and it is by now perfectly clear that non-coding regions are far more labile than coding regions. While some progress has been made in identifying regulatory elements analytically, see (26) or (27), for example, intervening sequences seem to be highly variable. Indeed, the evolutionary comings and goings of regulatory signals remain almost completely mysterious, and RA response elements, which are almost always found in traditional promoters or extended, multi-function enhancers, are short, degenerate, ambiguous signals ripe for evolutionary experimentation. One would therefore expect only coincidental functional or formal resemblances among the complete set of genes controlled by RA. What this tells us is that many interesting and surprising results remain to be found: genes whose regulation by retinoic acid is not a priori predictable.
Over the last quarter century, a substantial body of knowledge has been built up concerning gene expression regulation by RA. That work has contributed significantly to our understanding of context-regulated transcription, vertebrate development, and a host of important clinical issues. From the particular perspective of this paper, much of the work we consulted was tantalizingly close to helping answer the direct-versus-indirect question even though it was not originally designed to address that question at all. In other cases, elucidating a molecular pathway was a primary research goal and a clear answer was determined; and in a few cases, intriguing scientific issues have turned up when regulatory mechanisms do not seem to be as clear-cut as originally expected, as with LAMB1 (28–32). Of course, many RA studies seek clinical or nutritional information, and the poignant need for such studies is beyond question; yet in the larger scheme, knowing which regulatory events are direct and which are indirect can perhaps lead to superior pharmacological and nutritional protocols as well as to progress in basic science.
|
Endnote |
|---|
|
TOPABSTRACTBackgroundConstructing a classification...The classical RA pathwayLiterature reviewsMETHODSRESULTS AND DISCUSSIONEndnoteREFERENCES |
|---|
We think of this paper as a working document and hope that our errors and oversights will generously be pointed out by our colleagues so that the table can be updated, improved, and maintained, by us or by another group, as an evolving assessment of RA's genetic workings.
Gene Table—Continued
c Called 17-ß-HSD-II in PMID 0008013376.
d The Hs symbol and name is POU1F1: POU domain, class 1, transcription factor 1 (Pit1, growth hormone factor 1).
e It is not yet clear exactly what the active binding site(s) are. Orthologous control regions are definitely involved and there appears to be some degree of conservation in Mm and Rn.
f Rn promoter in Hs cells.
g See PMID 0010194513 for a brief review of RA and ApoA1.
h The sequence appears to come from chromosome 7 but may contain a large Line1 repeat.
i Figure 2B in PMID 0001700780 appears to show a data point which would satisfy our dose and time criteria. However, it is not discussed in the text.
j This is the Hs name. There is a 94% aa identity to rabbit OC2 according to OMIN.
k There has been some controversy about the metabolic products of the the gene(s) in different spp; also, Cyp26 may not be RA-inducible in some cells that nevertheless metabolize RA.
l It is not clear whether Stra7 and Gbx2 are different genes. The GB entries are virtually identical where they overlap. The Stra7 clone is effectively included in the Gbx-2 Ref Seq.
m The allelic variant GGTP1*C used in some studies is thought not to effect the generality of the RA work.
n Site from 2nd intron and flanking exon more or less conserved in Hs, Rn, Ma, Oc, Cf, Ss, Gg, and cats.
o Mm symbol and name.
p This assumes the Ggal gene RIHB (NCBI GI 434357) is orthologous to Mm Mdk.
q Interim Hs name; no Rn assignment.
r Rn data mentioned but not shown.
s The ability of RA to counteract estrogen through the OTX ERE is discussed in PMID 0001655807, and the ERE was used as a "negative RARE" in combination with transfected RARa, JUN, and ER.
t The figure demonstrating this is not easy to interpret.
u Rn promoter and exogenous RAR/RXR in Hs cells.
v Interim symbol and name.
w Earlier papers that do not distinguish enzyme forms are not considered here.
x Interim symbol and name.
y Interim symbol and name.
z Mm symbol and name.
aa Name by analogy to mammalian crystallins.
bb We assume DDX1 is the gene in question; there are other DEAD box proteins, of course, but the paper does not clearly distinguish them.
cc Interim symbol and name.
dd To us, the figure showing rapid induction is unconvincing; no dose is given, either.
ee It is not clear what has happened between times 0 and 24 hours in Figure 2b of PMID 0010674883.
ff It is not clear what has happened between times 0 and 24 hours in Figure 2b of PMID 0010674883.
gg F3 is frequently studied in APL cells because it is thought to be involved in the pathology of the disease. Some of the work cited here is in APL lines.
hh Interim symbol and name.
ii Probably Ins2 in Rn.
jj KRT6A seems to be the predominantly expressed K6 gene in Hs; the paper cited for Bt (in whom there are 3 K6 genes) is concerned with K6b; the motifs in PMID 0009326392 BLAST identically (and with the same single mismatch) to the provisional refseqs for both Hs K6 genes, KRT6A and KRT6b; the Hs AP-1 work is on K6b.
kk Most investigations so far have dealt with Erk activation, not message induction.
ll Interim symbol and name.
mm The gene studied now appears to be the ortholog of Msx2, not Msx1 (as thought at the time).
nn Mm symbol and name.
oo Interim symbol and name.
pp Induced in 3-dimensional systems but not in 2-dimensional cultures of keratinocytes and fibroblasts.
qq The effects of 9-cis, which are not covered here, have also been investigated. Cf. PMID 10403834 and PMID 0009717711 for example.
rr Interim Hs symbol and name.
ss SPRR1A, SPRR2, and SPRR3 are covered in some of these papers; the RA situation is basically the same.
tt Probable name, see PMID 11416019.
uu Hs DNA in Rn cells.
vv Suppression at 8 hours (100 nM) is discussed in PMID 10502285, but Figure 1B suggests it is significant by 4 h.
ww Both TOP2A and TOP2B have been studied, but most of the RA work has concentrated on 2A.
xx The statistical significance of a slight decrease at 6 h in PMID 11146166 is not clear.
yy An RARE half site seems to be marked in a GenBank entry but neither the site nor RA is mentioned in the associated paper.
zz No official name or symbol; no curated orthologs.
aaa Aggrecanase-1 is an alias for ADAMTS4; some of the papers listed here cover ADAMTS5 (aggrecanse-2) as well. MMP3 and MMP13 (q.v.) may also be involved.
bbb Probable Mm ortholog; no Rn assignment.
ccc High consentrations of retinoic acid inhibit BGLAP induction by vitamin D, but a well-characterized AP-1 response element is contained in the VDRE. Some experiments found neither induction nor suppression by RA alone.
ddd There is no evidence that RA has different effects on the expression of the splicing alternates, calcitonin and calcitonin gene related peptide (CGRP).
eee Several studies have also been done in Rn and Hs using 9-cis. No good d/t data there, either.
fff Rn sequences in HepG2 cells; no RA regulation seen in hamster.
ggg Interim symbol and name.
hhh An altered transcription rate could not be shown in nuclear run-on experiments (although controls worked as expected); this was attributed to the highly stable mRNA.
iii Symbol and name pending.
jjj Interim symbol and name.
kkk Mm symbol and name.
lll Interim symbol and name.
continued
Gene Table—Continued
mmm Very long exposure may induce expression in some systems.
nnn Promoter constructs from Hs used in Rn cells.
ooo Mm symbol and name.
ppp The older articles study enzyme activity without distinguishing OAS1, -2, and -3.
qqq An altered transcription rate could not be shown in nuclear run-on experiments (although controls worked as expected); this was attributed to the highly stable mRNA.
rrr Many studies have looked at mechanisms by which RA influences RB phosphorylation. They are not included here.
sss Interim symbol and name.
ttt An altered transcription rate could not be shown in nuclear run-on experiments (although controls worked as expected); this was attributed to the highly stable mRNA.
uuu An altered transcription rate could not be shown in nuclear run-on experiments (although controls worked as expected); this was attributed to the highly stable mRNA.
vvv Interim Mm symbol and name.
www Probable Mm ortholog; no Rn assignment.
xxx It is not clear whether the repressed gene was Shmt1 or Shmt2.
yyy An altered transcription rate could not be shown in nuclear run-on experiments (although controls worked as expected); this was attributed to the highly stable mRNA.
zzz Interim symbol and name.
aaaa Studies do not necessarily distinguish members of the TIMP family.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
