© 2001 European Society of Cardiology
Discovering altered genomic expression patterns in heart: transcriptome determination by serial analysis of gene expression
Laboratory of Cardiovascular Science, National Institute on Aging, National Institutes of Health Baltimore, MD USA
* Corresponding author. Tel.: +1-4105588095; fax: +1-4105588150 E-mail address: bohelerk{at}grc.nia.nih.gov (K.R. Boheler).
 |
Abstract |
|---|
 Top Abstract 1. Introduction2. Towards a molecular...3. Serial analysis of...4. Application of SAGE...5. Comparative analyses with...6. Potential application of...References |
|---|
The development of cardiovascular diseases such as heart failure involve functional changes that are beneficial short-term, but may be fatal long-term. Current therapeutic approaches are tailored to limit progression of a disease and to maintain quality of life. At a molecular level, these disease processes involve quantitative and qualitative changes in gene expression. Although some changes in mRNA abundance may not have direct protein correlates, analysis of all the mRNAs present in a cell population (the cells transcriptome) has become a focal point of genomic research. The aim is to provide information about the dynamics of total genome expression in response to environmental changes and point to candidate genes responsible for the cascade of events that result in a disease state. One way of performing these analyses utilizes the technique of Serial Analysis of Gene Expression (SAGE). This method evaluates thousands of expressed transcripts both quantitatively and qualitatively in a single assay. In the first of two reviews on transcriptome analysis, we describe the current state of genomic research for determination of the transcriptome by Serial Analysis of Gene Expression, present the first limited SAGE analysis of rodent heart gene expression, and discuss how results generated with this approach can be applied to the study and treatment of cardiovascular diseases.
Key Words: SAGE • Heart Failure • Rodent • Human • Transcriptome
Received October 16, 2000; Revised December 11, 2000; Accepted January 17, 2001
|
1. Introduction |
|---|
|
TopAbstract1. Introduction2. Towards a molecular...3. Serial analysis of...4. Application of SAGE...5. Comparative analyses with...6. Potential application of...References |
|---|
Cardiovascular diseases are among the major causes of morbidity and mortality in developed countries. Of particular concern is progressive heart failure that develops when cardiac load chronically exceeds the capacity of the organ. A significant reduction in cardiac output and an increase in left ventricular end-diastolic pressure activate peripheral compensatory mechanisms to maintain blood pressure and improve stroke volume. Thus, heart failure involves impaired muscle function and the response of the whole body to diminished cardiac output [1]. Altered peripheral responses include constriction of the peripheral circulation, neurogenic and endocrine activation, cytokine activation, structural and functional abnormalities of skeletal muscle, changes in lung function and retention of sodium and water. Short-term, these adaptations are beneficial, but long-term, asymptomatic left ventricular dysfunction may progress to symptomatic heart failure. Therapeutic approaches to treat heart failure (pharmacologic, mechanical devices and surgery) are tailored to limit progression of the disease and to maintain quality of life [1]. Treatment must involve limitation of neuroendocrine and cytokine activation and the reversal of cardiac and extracardiac abnormalities. Current therapies are not always applicable to every patient, the result of which can be catastrophic [1]. Novel approaches are required to ultimately treat the specific underlying molecular mechanisms involved in heart failure and its progression in a given patient.
Genomics, the study of genes and their function, holds the potential to ‘resolve’ long-standing genetic questions. Current estimates are that the human genome contains 80 000–140 000 genes [2], and that any given cell (cardiac myocyte) expresses a combination of 40 000–70 000 genes [3]. Less than 7000 of these gene products have a known or putative function [4]. The genes expressed at a given time and place are a function of development, life-style, aging or disease and the genetic make-up of an individual. A complete understanding of how genetic mechanisms control development, aging and disease thus requires an accurate quantification of gene expression in a global context — not just one gene at a time. More recently, genomics has advanced to what could be called the post-genomic era. Emphasis has therefore shifted from the study of genes to the identification of altered patterns of gene expression (known and unknown) and gene function to understand disease processes. Systematic determination of the transcriptome (identity and expression levels of all mRNA transcripts in a given population of cells) in specific cell types, including those in heart, should revolutionize our understanding of biological processes. In the case of heart disease, identification of all of the gene products with altered abundance will be a critical step in identifying the biological mechanisms leading to cardiovascular disease states. Once identified and the consequences partially understood, novel prognostic and diagnostic indicators of heart disease, and ultimately, new treatments will be developed. The critical question is how to acquire and prioritize information on the transcriptome, how to select relevant disease-related genes, and how to utilize this information for the benefit of man.
The aim of this review and a future review on DNA arrays is to describe how transcriptome comparisons provide information about the dynamics of total genome expression in response to environmental changes with age. Specifically, we will describe the current state of genomic research for determination of the transcriptome. In this paper, we will focus only on the use of SAGE in this process, and we will show how results generated from SAGE can be applied to the study of cardiovascular diseases.
|
2. Towards a molecular understanding of cardiology |
|---|
|
TopAbstract1. Introduction2. Towards a molecular...3. Serial analysis of...4. Application of SAGE...5. Comparative analyses with...6. Potential application of...References |
|---|
Modern cardiology includes a broad range of allied disciplines: genetics, molecular biology, cell biology, human and animal physiology, pharmacology and clinical interventions. Major developments in cardiovascular molecular biology have fundamentally changed our understanding of the heart and cardiovascular diseases. Less than 20 years ago, the molecular structure of the cardiac myocyte was described only in terms of a fixed phenotype typical of terminally differentiated cells [5]. Recombinant DNA technology led to the view that the phenotype of these cells changes with ontogeny, aging and disease [6], and alterations in steady-state levels of messenger RNA usually paralleled changes in the levels of the corresponding protein products [7]. The mechanisms responsible for these changes in gene expression were largely unknown.
Follow-up studies were undertaken to determine the mechanisms underlying the altered gene expression while attempting to delineate what changes within the myocyte contribute to the phenotype [8]. Most studies have obligatorily focused on the analysis of one gene at a time, often in a controlled in vitro situation, resulting in a plethora of non-overlapping, mutually exclusive functional data. Subsequent studies have taken data from this basic research and applied it to whole-animal models and to the evaluation of human heart disease in the hope that novel developments from basic and clinical research would result in additional insights into heart failure therapy at the organism level. While it was observed that many molecular changes between simple model systems and complex biological systems in humans (increased cardiac ANF with hypertrophy and failure) [9] were in accord, others were discordant, and are presently highly disputed or model-dependent (SR CaATPase expression and function) [7].
Although gene expression is regulated at multiple points, transcription has emerged as a primary target for high throughput genomic analysis. As such, whole genome expression experiments have become important tools in functional genomics. Molecular methods to analyze genome expression include sequencing of expressed sequence tags (ESTs — sequencing of cDNA clones produced from mRNA present in a cell at a given time), subtractive hybridization (isolation of mRNA/cDNA species through competitive hybridization strategies to identify which ones are present at differing abundances in two cell populations), differential display (PCR-based method for identifying mRNAs that are expressed differently between two cell populations), competitive PCR, cDNA, or oligonucleotide arrays (high density arrays consisting of cDNA clones that can be screened by labelled probes generated from total RNA cell populations) and Serial Analysis of Gene Expression (see below). The number of transcripts or mRNAs that can be analyzed by subtractive hybridization, differential display, or competitive PCR is limited, and the relative abundance of mRNA species that can be estimated by determining the proportion of ESTs from a cDNA library is time consuming. When isolated to identify the sequence of ESTs, cDNAs can, however, be printed on arrays or microarrays. DNA microarray technologies (cDNA- or oligonucleotide-based) permit systematic approaches to biological discovery and allow quantitative information to be obtained regarding the complete transcription profile of cells [10]. Although microarray technology can potentially examine the expression patterns of a large number of genes, the method can only examine expressed sequences that have already been identified. SAGE is the only technique that currently promises a quantitative characterization of the complete transcriptome, i.e. both mRNAs that have and have not yet been identified for a cell type or tissue [3].
|
3. Serial analysis of gene expression (SAGE) |
|---|
|
TopAbstract1. Introduction2. Towards a molecular...3. Serial analysis of...4. Application of SAGE...5. Comparative analyses with...6. Potential application of...References |
|---|
SAGE is a molecular technique for evaluating thousands of expressed transcripts both qualitatively and quantitatively [11]. Two major principles underlie SAGE analyses. First, short DNA sequences are sufficient for the identification of individual gene products and second, concatenation (linking together) of the short DNA sequences or tags increases the efficiency of identifying expressed mRNAs in a sequence-based assay. The transcript profile generated by SAGE thus relies on 10–14 base-nucleotide sequences (SAGE tags) for gene identification and cloning. Theoretically, a 14-base tag composed of a combination of four nucleotides (A,C,G and T) present in DNA can define 414 (268 435 456) different transcripts, i.e. a much greater number than the 70 000–140 000 predicted human genes, without any prior knowledge of the gene transcripts present in a cell line or tissue sample.
To obtain the tag information by SAGE (Fig. 1), poly (A+) RNA (mRNA) is isolated and transcribed into double-stranded cDNA, using biotinylated oligo (dT) as a primer. After digestion with a restriction enzyme (Anchoring Enzyme), the biotinylated 3'-most fragments are purified using streptavidin-coated magnetic beads. The anchoring restriction enzyme (NlaIII) with a 4-bp palindromic recognition site (CATG), on average locates and cuts double-stranded DNA molecules every 256 base pairs (44=256). The isolation step provides short DNA sequences or tags from a defined position within each cDNA. The fragments are divided in half and ligated to two different linkers. Each linker contains a restriction site for the Tagging Enzyme (a type II restriction endonuclease), the Anchoring Enzyme and a priming site for amplification by PCR. A tagging enzyme recognizes specific DNA sequences (e.g. GGGAC(N)10 for BsmFI) and cuts the double-stranded DNA several base pairs away from the recognition site. It is used in the SAGE protocol to release a short tag from the cDNA along with the linker. The linker-tag molecules are ligated tail-to-tail to form ditags consisting of linker-ditag-linker constructs that can be amplified by PCR. After amplification, the ditags are released from the linker molecules by use of the Anchoring Enzyme and ligated together to form concatemers of many ditags. Once propagated in bacteria, the concatemers are sequenced, each tag identified, and the sequence compared to genome/cDNA databases to identify the expressed transcript. SAGE software identifies recognition CATG sequences and subsequently the tag dimers (ditags). Ditags that satisfy the size criteria are extracted and recorded. Sequencing of a single clone consisting of concatemers can thus identify not one EST, as with sequencing cDNA clones to identify ESTs, but multiples of ESTs that can be used to identify many expressed transcripts.
|
Since the introduction of the method in 1995, many improvements have been proposed to increase the number of SAGE tags that can be cloned in a single concatemer [12,13]. The efficiency of the method has increased so that its only limit is the number of reliable sequences generated by automated sequencers. Finally, with the ever rapidly expanding sequence information in public databases, identification of the corresponding sequence on the basis of the expressed tag permits gene matching of a majority of the tags generated [14,15].
|
4. Application of SAGE to transcriptome analysis |
|---|
|
TopAbstract1. Introduction2. Towards a molecular...3. Serial analysis of...4. Application of SAGE...5. Comparative analyses with...6. Potential application of...References |
|---|
Application of SAGE to the cardiovascular system could initially be thought of as a screening tool to determine the expression patterns in specified cell types, until the transcriptome has been well defined for cardiovascular tissues. The major goals of SAGE screening are however, two-fold: 1) to identify all the transcripts present in normal tissues (both known and unknown); and 2) to quantitatively and qualitatively compare the expression profiles between control and diseased (experimental) states (Fig. 2). Once accomplished, patterns and functional correlates can be identified and potential disease-related genes identified for subsequent analysis.
|
To illustrate a transcription profile a limited analysis has been generated from adult mouse heart for the purposes of this review. A C57BL/6 adult mouse heart SAGE library was prepared using standard SAGE protocols [11]. The sequences of 83 clones were analyzed using SAGE 3.04 software to identify 4-bp recognition sequences (CATG for NlaIII enzyme used) and to extract tags located between these identifiers. Excluding duplicate dimers, 1042 SAGE tags were identified. Of these, 41 were excluded from the analysis because they originated from linker sequences, and 1 was determined to be a non-informative poly (A+)-tag possibly having originated from the NlaIII recognition site located immediately upstream of a poly (A+)-tail. The remaining 1000 tags were analyzed further (Fig. 3a,b).
|
For the analysis of rodent sequences, two web-linked sites were accessed to identify the mRNAs corresponding to the sequenced tag. The Non-Redundant (NR) Rodent-Database (Release 116.0, March 3, 2000) and EST-databases (Release 116.0, March 3, 2000) were downloaded from the GenBank ftp site (ftp.ncbi.nlm.nih.gov/genbank), and analyzed using the SAGE software. NlaIII recognition sites (CATG) closest to the poly (A+)-tail were identified in sequences generated from mouse (Mus musculus) samples. All the tags generated from the adult mouse heart SAGE library were compared with these databases to identify sequence matches. As illustrated in Fig. 3b, most SAGE tags directly matched a number of genes in the NR-database or entries in the EST-database (dbEST). A few tags had multiple matches to entries in these databases whereas other tags did not match any known public domain sequence.
In this SAGE analysis of adult mouse heart, 505 individual transcripts were identified. To avoid tags potentially generated by sequencing errors (with single-strand sequencing, sequencing errors range generally from 
0.5–1.0%), only tags that appeared at least twice were analyzed further. One expressed tag (GCTGCCCTCC: EST, similar to cytochrome C oxidase polypeptide I (COX1)) comprised 21% of the total pool of tags (Table 1). Ten other tags were present at levels from 1.0–4.2% of the total. Seven of these ten tags matched genes involved in energy metabolism (ESTs similar to NADH-ubiquinone oxidoreductase chains 2, 4, and 6; COX2 and COX3). One tag with multiple matches probably corresponds to a subunit of cytochrome b, since the second match, TAFII250 transcription factor, has not been reported in heart (UniGene data). Two cardiac-restricted transcripts were highly abundant: one EST, similar to the human atrial natriuretic factor precursor and another corresponding to cardiac myosin heavy chain were both present at a frequency of >1.0% in this pool of 1000-tags. We also found many examples of muscle-associated [Myoglobin; Skeletal muscle 
tropomyosin; EST, similar to human myosin regulatory-light-chain-2, smooth muscle isoform; muscle-specific enolase β (ENO3)] and cardiac-associated [ventricular-alkali-myosin-light-chain (MLC1V); slow/cardiac-troponin-C (cTnC); cardiac troponin-T; myosin light-chain-2] genes. The remaining tags matched a variety of other gene products related to protein synthesis (multiple ribosomal proteins and a few elongation factors), secretion, the cytoskeleton, and a variety of cytoplasmic enzymes.
|
The mRNA profile generated from this analysis indicates, as expected, that adult myocardium is a specialized tissue with important energetic needs capable of generating force through the use of molecular motors. The mouse heart contracts on average 300–600 beats/min; its requirement for ATP is therefore very high. Seven of the eleven most abundant transcripts are involved in energy metabolism, consistent with the large volume (
50%) of the myocyte occupied by mitochondria. To contract, the myocardium needs cellular motors, hence the abundant expression of myosin light-and heavy-chains. To produce enough protein for energy production and force generation, ribosomal proteins are also very abundant. Most of the identified transcripts can thus be divided up into three functional classes: gene/protein expression, metabolism, and cell structure/motility (Table 1). Standard classifications for such analyses can be found elsewhere [16]. In doing these analyses we found that exactly 50% of the unique tags matched dbEST, 45% matched the NR-database, and 5% did not match either (Fig. 3B). More specifically, two unique tags gave multiple matches to the NR-database, and three more to multiple dbEST entries. One group of tags did not have similarity to any known gene (4 transcripts). Although abundant, these transcripts have yet to be identified/classified in mouse heart and further analyses are required to determine their function. These data illustrate very well the usefulness of SAGE analyses to define the transcriptome of reference tissue samples — SAGE analysis identifies transcripts that are both known and unknown. In fact, the majority of proteins encoded by cardiac mRNAs have an unknown function in heart. For those tags corresponding to identified transcripts, it is a simple matter of classifying and categorizing, on the basis of function, what proteins are likely to be transcribed from the expressed mRNAs. In instances in which the tags do not match known mRNA sequences, the work to determine their role becomes labor intensive and must be tackled at least partially, one gene at a time.
As larger expression profiles are generated it becomes increasingly important to place each transcript into a category and to identify expression patterns (temporal, developmental, topographical, histological, physiological, etc.) for the expressed mRNAs. In our example we provide information on 1000 tags, far to few to generate comprehensive expression profiles. The absence of most transcription factors, intracellular transducers, and signaling molecules indicates that these transcripts are present at a much lower abundance than those needed for the functional requirements of the heart. In disease states, signal transducers and transcription factors may be more informative of potential long-term molecular changes/consequences. Further analysis is therefore required. Sequencing of 
25 000 tags is necessary to identify all the expressed mRNA transcripts in a mammalian cell, like heart myocytes, at a copy number of >30 transcripts/cell (>0.01%) [17]. This is the number of tags that would be necessary to realistically identify and classify products implicated in functional categories or disease-related pathways however, it is insufficient for the determination of the whole transcriptome.
In the analysis presented here from mouse heart — an important genetic model for cardiovascular research [18,19], we have illustrated the types of information that can be generated just from screening normal tissues. Sequencing of the mouse genome is however, far from complete. Only 0.6% of the mouse genome has been ‘finished’, while another 5.5% is in the ‘draft’ stage. Because so little of the mouse genome is complete, most of the unique sequenced tags presented in this review identified only abundant transcripts. The same analysis, but with >50 000 tags, would have indicated that 25% of the unique tag sequences do not correspond to known transcripts (unpublished data). The percentage of unknown unique tags originating from low abundance mRNAs will increase significantly as they are found. Once the mouse and human genomes are completed, the mRNAs from both species are compared and functionally related proteins linked, our understanding of disease and the progression to heart failure will grow exponentially. Although functional categorizations are nearly impossible for unknown transcripts, comparative studies can still be used to identify potential markers or disease-related genes.
To arrive at a comprehensive, quantitative analysis requires a much larger profile. To determine the abundance of a majority of transcripts at 3 copies per cell requires 300 000 SAGE tags [17,20]. To assay transcripts present at an abundance of 1 copy per cell, many more tags are needed. Determination of the entire cell transcriptome becomes labor intensive, relatively expensive, and the data analysis must be automated. Examination of 50–100 000 tags is however, relatively easy; however cataloging, categorizing, and identifying all of the transcripts with altered expression are still formidable tasks. With the near completion of the human genome, this process has become increasingly simple. Research on the human heart poses other problems however, because the availability of ‘normal’ samples to identify ‘abnormal’ gene expression is very limited. Once accomplished, the process of verifying the changes in human tissues and interpreting how and why the changes occur can begin. In non-human animal models in which samples are readily available the tasks of identifying and determining what proteins are encoded will remain daunting for the foreseeable future.
SAGE analyses of human tissues are thus much more robust than those performed on mouse and on other species. The reason for this is quite simple: the International Human Genome Sequencing Consortium will soon complete sequencing of the human genome [21]. Recently it was announced that a ‘working draft’ of the human genome was complete. It consists of overlapping fragments covering >97% of the human genome, and is assembled for approximately 85% of the genome. Approximately 24% of the entire genome, including whole chromosomes 21 and 22 have already reached a ‘finished’ form. The imminent completion of the human genome sequence, particularly when combined with expression analyses like SAGE, will provide researchers in almost every field of modern medicine with the ability to measure genetic risk factors, perfect diagnostics, design, and ultimately select drugs for patients.
|
5. Comparative analyses with SAGE |
|---|
|
TopAbstract1. Introduction2. Towards a molecular...3. Serial analysis of...4. Application of SAGE...5. Comparative analyses with...6. Potential application of...References |
|---|
An important advantage of SAGE is that it can be used to compare expression profiles from any number of SAGE libraries (Fig. 2). Such comparisons can be performed on libraries generated within a single laboratory or via the SAGEmap website where comparisons with any other publicly deposited SAGE library can be performed, simply by clicking on a ‘mouse’. Comparison of samples can reliably detect transcripts with changes in expression with great sensitivity. While a majority of transcripts are expressed at almost equal levels (
90–95%) in most cells, changes in expression of certain genes can be verified with statistical analyses [14,22]. These data are quantitative and provide valuable insight into physiological and patho-physiological states. SAGE has been used successfully for transcription profiling and identification of transcripts with altered expression between a number of normal or disease states [20,23–33]. Application of SAGE to the study of cancer has yielded the most comprehensive comparative data. The first large-scale SAGE study was performed to identify novel p53 tumor suppressor-regulated genes in transformed rat embryo fibroblasts [34]. Of approximately 15 000 genes analyzed, only 14 were found to increase in p53-expressing cells compared to controls. Many of the genes were predicted to encode proteins that either could generate or respond to oxidative stress, including one implicated in apoptosis. Based on these analyses of the transcriptome, the authors concluded that p53 led to apoptosis through a three-step process involving transcriptional induction of redox-related genes, formation of reactive oxygen species, and oxidative degradation of mitochondrial components. The result was cell death. Zhang et al. [20] generated five SAGE libraries from human colorectal and pancreatic cancers and normal colorectal epithelium. Once analyzed, >500 transcripts had altered expression in cancer cells. While a number of these were known to be dysregulated in cancer, multiple factors previously unassociated with malignancy were also detected. These findings are now the subject of further investigation. Recently, in another study on colorectal cancer, St. Croix et al. [35] clearly demonstrated that tumor and normal endothelium are distinct at the molecular level, the findings of which were predicted to be important for the development of anti-angiogenic therapies. An extensive summary of the use of SAGE to identify specific genes involved in malignant processes and cancer has been reviewed by [3].
Our group recently used SAGE to identify novel transcription factors involved in early cardiac development (Anisimov et al., submitted). Over 150 000 tags were generated from three libraries, derived from undifferentiated and differentiating embryonic cells induced to generate a high percentage of cardiac myocytes. We observed a developmental profile where genes are activated or repressed in a differentiation-specific manner, and we identified a number of transcription factors involved in early cardiac development, including GATA-4, TEF-1, and HAND1 expression. A novel set of transcripts, both known and unknown, changed significantly in early development, indicating a pattern of expression potentially required for the development of cardiac cells (Fig. 4). We are nearing completion of a large SAGE profile examining adult mouse heart. These data are currently being used for comparison purposes with the developmental study of cardiac myocyte differentiation and should be useful to other investigators evaluating the transcriptome of the heart.
|
Results such as these underscore the validity and power of transcriptome analysis in identifying proteins important to heart physiology and pathophysiology. By making SAGE data from heart publicly available at SAGEmap, the discovery and identification of gene products implicated in biological pathways important to the cardiovascular system, disease-related or not, will be accelerated. Because SAGE is the only technique currently capable of a quantitative characterization of the complete transcriptome for a cell type or tissue, thorough application of this technique to cardiovascular diseases should be given a high priority.
|
6. Potential application of SAGE to cardiovascular diseases |
|---|
|
TopAbstract1. Introduction2. Towards a molecular...3. Serial analysis of...4. Application of SAGE...5. Comparative analyses with...6. Potential application of...References |
|---|
The potential impact of SAGE on the understanding and treatment of heart failure is profound. To date, we have only performed SAGE analyses on cells isolated from culture and rodent models. Transcription profiles of human heart failure, including those with known or unknown genetic causes, should be generated. Such results could rapidly lead to better diagnostic and prognostic indicators, and where the genetic cause is unknown (e.g. most forms of idiopathic dilated cardiomyopathies), may point to the originating event(s) responsible for the progression to disease. In those cases where the genetic causes are known (mutations in myosin heavy chain for hypertrophic cardiomyopathies), the signaling events downstream of the lesion can be determined, potentially leading to better therapeutic interventions. Because heart failure affects tissues other than the myocardium, profiles should also be made from other tissues (lung, skeletal muscle, kidney, etc.) to better understand the wide-ranging effects of heart failure on human physiology.
SAGE analyses of tissues from animal models will ultimately provide the most poignant information about early and subsequent events that lead to heart failure. Ischemic heart disease, heart failure and hypertrophy models have all been established in a variety of animal models (rats, dogs, rabbits, etc.) using surgical techniques. Indeed, a large number of models of heart failure are still (and will continue to be) performed in rats. A number of rodent models also spontaneously develop cardiac pathologies [36,37] or can be manipulated genetically to mimic a disease state [38–41]. Genetic manipulations have successfully generated animals with myocardial hyperplasia, hypertrophy, cardiomyopathy, heart failure, and myocarditis. These have become valuable tools for the study of specific pathways and functional abnormalities [36,42–44]. Although the genetic initiation event is known in transgenic models, the down-stream events necessary to evoke heart failure remain to be determined. Because SAGE profiles can be generated in animal models from both cross-sectional and longtitudinal study designs, down-stream events can be identified and examined in great detail, with and without pharmacological interventions. We will ultimately even be able to distinguish between changes in gene expression due to aging, from those due to progressive disease processes. Finally, animal models will be of great use in developmental studies because of the ethical difficulties of research on humans. Together, these temporal and spatial profiles of the changes in gene expression as a function of the genetic anomaly, or more likely, the adaptive response to congestive heart failure with or without drug intervention will provide critical information relevant to the development and treatment of heart failure. Application of SAGE to rat and mouse models may therefore be particularly apt for future transcriptome analyses of heart failure.
Although SAGE holds the promise of accurately quantifying changes in cardiac gene expression, the power of this method can lead to false conclusions. Ideally, SAGE libraries should be made from pure cell populations — a SAGE library from heart contains cardiac myocytes, endothelial, epithelial, neuronal, smooth muscle and numerous other cell types. Because SAGE accurately quantifies all transcripts present in a sample, changes in gene expression may reflect mRNA composition between cell types and correspondingly may not be implicated in a disease process. When great care has been taken to analyze a specific cell population however, under two conditions (experimental- or disease-related), statistical analyses are very powerful in identifying candidate genes for subsequent studies. These candidate mRNAs should however, be subjected to independent analyses both at the mRNA level (PCR, Northern analysis, in situ analysis) and, where possible, for the protein (immunocytochemistry, Western analysis, epitope analysis). While current estimates suggest that 60% of all changes in mRNA abundance reflect changes in protein abundance, the remainder involve other mechanisms. Thus, results from comparative SAGE analyses must be confirmed at the protein level, where possible, to ensure that corresponding changes in protein abundance and/or function are actually implicated in the experimental or disease-related process being examined.
Although we have begun analyzing numerous gene transcription profiles from rodent models (developmental-, cardiac- and disease-related), comparisons with human heart transcript profiles present unique challenges. The difficulty is not so much in obtaining samples and when necessary using microSAGE or PCR-SAGE to generate profiles from human biopsies, but in interpreting data of samples composed of multiple cell types, taken from patients of different ages, state of failure, drug regimens, etc. Comparisons between animal models and human occurrences of heart failure should help overcome these limitations to reveal signaling events and pathways that lead to conserved end-points in a disease process. This will require extensive data-analysis, identification of which models most appropriately mimic human diseases, increased financial resources, and long-term commitments by researchers involved in modern cardiology. The ultimate goal is to tailor treatments to individuals based on diagnostic gene profiling of patients.
In the field of oncology, such results are already being realized. Specifically, strong correlations have been demonstrated between certain pathological conditions and the presence of expressed genes. These data have been confirmed by complementary techniques, and the corresponding protein changes have been identified. A few of these have already been used for advanced diagnostics, and their identification is yielding valuable information about the possible mechanisms underlying these disease processes [45,46]. With enough samples and appropriate comparisons, similar benefits can be realized in the cardiovascular realm.
In conclusion, the scientific basis for the transition from cardiac hypertrophy to heart failure or the development of any other cardiovascular disease requires research that takes an encompassing view of the cardiac milieu to elucidate specific mechanisms. Gene expression analysis is a critical first step in this process and Serial Analysis of Gene Expression can detect transcripts with altered levels of expression in a highly effective, comprehensive and reliable way. An alternative method for transcriptome analysis employs the use of cDNA arrays or microarrays (Chips) and will be the subject of a subsequent review. Regardless of the technique utilized, most of the initial transcriptome data may have limited relevance to biological mechanisms underlying the disease processes. The identification of specific similarities in altered gene expression patterns in animal models and human heart failure will however, provide critical insight into future directions for animal research relevant to human heart failure. Ultimately these results provide the basis for novel and more effective treatment regimens for congestive heart failure that will be specifically tailored to each patient.
|
References |
|---|
|
TopAbstract1. Introduction2. Towards a molecular...3. Serial analysis of...4. Application of SAGE...5. Comparative analyses with...6. Potential application of...References |
|---|
- The treatment of heart failure. Task Force of the Working Group on Heart Failure of the European Society of Cardiology. Eur Heart J. 1997;18:736–53.
- Dickson D. Gene estimate rises as US and UK discuss freedom of access. Nature (1999) 401:311.[Web of Science][Medline]
- Velculescu V.E., Madden S.L., Zhang L., et al. Analysis of human transcriptomes [letter]. Nat Genet (1999) 23:387–388.[Web of Science][Medline]
- Fields S. The future is function [news]. Nat Genet (1997) 15:325–327.[CrossRef][Web of Science][Medline]
- Schwartz K., Carrier L., Mercadier J.-J., Lompré A.-M., Boheler K.R. Molecular phenotype of the hypertrophied and failing myocardium. Circ (1993) 87(Suppl. VII):5–10.
- Boheler K.R., Schwartz K. Gene expression in cardiac hypertrophy. Trends Cardiovasc Med (1992) 2:176–182.[CrossRef][Web of Science]
- Barton P.J.R., Boheler K.R., Brand N.J., Thomas P.S. Molecular Biology of Cardiac Development and Growth (1995) Austin TX: R.G.Landes.
- Chien K.R., Zhu H., Knowlton K.U., et al. Transcriptional regulation during cardiac growth and development. Annu Rev Physiol (1993) 55:77–95.[CrossRef][Web of Science][Medline]
- Chien K.R., Knowlton K.U., Zhu H., Chien S. Regulation of cardiac gene expression during myocardial growth and hypertrophy: molecular studies of an adaptive physiologic response. FASEB Journal. (1991) 5:3037–3046.[Abstract]
- Young R.A. Biomedical discovery with DNA arrays. Cell (2000) 102:9–15.[CrossRef][Web of Science][Medline]
- Velculescu V.E., Zhang L., Vogelstein B., Kinzler K.W. Serial analysis of gene expression. Science (1995) 270:484–487.
[Abstract/Free Full Text] - Powell J. Enhanced concatemer cloning-a modification to the SAGE (Serial Analysis of Gene Expression) technique. Nucleic Acids Res (1998) 26:3445–3446.
[Abstract/Free Full Text] - Kenzelmann M., Muhlemann K. Substantially enhanced cloning efficiency of SAGE (Serial Analysis of Gene Expression) by adding a heating step to the original protocol. Nucleic Acids Res (1999) 27:917–918.
[Abstract/Free Full Text] - Lash A.E., Tolstoshev C.M., Wagner L., et al. SAGEmap: A public gene expression resource. Genome Res (2000) 10:1051–1060.
[Abstract/Free Full Text] - Wheeler D.L., Chappey C., Lash A.E., et al. Database resources of the National Center for Biotechnology Information. Nucleic Acids Res (2000) 28:10–14.
[Abstract/Free Full Text] - Hwang D.M., Dempsey A.A., Wang R.-X., et al. A genome-based resource for molecular cardiovascular medicine toward a compendium of cardiovascular genes. Circulation (1997) 96:4146–4203.
[Abstract/Free Full Text] - Bertelsen A.H., Velculescu V.E. High-throughput gene expression analysis using SAGE. DDT (1998) 3:152–159.
- Chien K.R. Genes and physiology: molecular physiology in genetically engineered animals. J Clin Invest (1996) 97:901–909.[Web of Science][Medline]
- Chien K.R. Molecular medicine in genetically engineered animals: series introduction [editorial]. J Clin Invest (1996) 97:2.[CrossRef][Web of Science][Medline]
- Zhang L., Zhou W., Velculescu V.E., et al. Gene expression profiles in normal and cancer cells. Science (1997) 276:1268–1272.
[Abstract/Free Full Text] - Hattori M., Fujiyama A., Taylor T.D., et al. The DNA sequence of human chromosome 21. The chromosome 21 mapping and sequencing consortium. Nature (2000) 405:311–319.[CrossRef][Medline]
- Audic S., Claverie J.M. The significance of digital gene expression profiles. Genome Res (1997) 7:986–995.
[Abstract/Free Full Text] - Madden S.L., Galella E.A., Riley D., Bertelsen A.H., Beaudry G.A. Induction of cell growth regulatory genes by p53. Cancer Res (1996) 56:5384–5390.
[Abstract/Free Full Text] - Hibi K., Liu Q., Beaudry G.A., et al. Serial analysis of gene expression in non-small cell lung cancer. Cancer Res (1998) 58:5690–5694.
[Abstract/Free Full Text] - Lal A., Lash A.E., Altschul S.F., et al. A public database for gene expression in human cancers. Cancer Res (1999) 59:5403–5407.
[Abstract/Free Full Text] - Nacht M., Ferguson A.T., Zhang W., et al. Combining serial analysis of gene expression and array technologies to identify genes differentially expressed in breast cancer. Cancer Res (1999) 59:5464–5470.
[Abstract/Free Full Text] - Takano T., Amino N. Molecular-based diagnosis of thyroid carcinomas by detecting cancer-specific mRNAs] [In Process Citation]. Rinsho Byori (2000) 48:149–154.[Medline]
- Neilson L., Andalibi A., Kang D., et al. Molecular Phenotype of the Human Oocyte by PCR-SAGE. Genomics (2000) 63:13–24.[CrossRef][Web of Science][Medline]
- Virlon B., Cheval L., Buhler J.M., Billon E., Doucet A., Elalouf J.M. Serial microanalysis of renal transcriptomes. Proc Nat Acad Sci USA (1999) 96:15286–15291.
[Abstract/Free Full Text] - Welle S., Bhatt K., Thornton C.A. Inventory of high-abundance mRNAs in skeletal muscle of normal men. Genome Res (1999) 9:506–513.
[Abstract/Free Full Text] - Hashimoto S., Suzuki T., Dong H.Y., Nagai S., Yamazaki N., Matsushima K. Serial analysis of gene expression in human monocyte-derived dendritic cells. Blood (1999) 94:845–852.
[Abstract/Free Full Text] - Hashimoto S., Suzuki T., Dong H.Y., Yamazaki N., Matsushima K. Serial analysis of gene expression in human monocytes and macrophages. Blood (1999) 94:837–844.
[Abstract/Free Full Text] - Yamashita T., Hashimoto S., Kaneko S., et al. Comprehensive gene expression profile of a normal human liver [In Process Citation]. Biochem Biophys Res Commun (2000) 269:110–116.[CrossRef][Web of Science][Medline]
- Madden S.L., Galella E.A., Zhu J., Bertelsen A.H., Beaudry G.A. SAGE transcript profiles for p53-dependent growth regulation. Oncogene (1997) 15:1079–1085.[CrossRef][Web of Science][Medline]
- St Croix B., Rago C., Velculescu V., et al. Genes expressed in human tumor endothelium. Science (2000) 289:1197–1202.
[Abstract/Free Full Text] - Hasenfuss G. Alterations of calcium-regulatory proteins in heart failure. Cardiovasc Res (1998) 37:279–289.
[Free Full Text] - Boluyt M.O., Bing O.H. Matrix gene expression and decompensated heart failure: the aged SHR model. Cardiovasc Res (2000) 46:239–249.
[Abstract/Free Full Text] - Chien K.R. Molecular advances in cardiovascular biology. Science (1993) 260:916–917.
[Free Full Text] - Gottshall K.R., Becker K.D., Hunter J.J., Chien K.R. A genetic based model of cardiac hypertrophy in MLC-ras mice. J.Card Fail (1996) 2:S28–S34.
- Tanaka N., Dalton N., Mao L., et al. Transthoracic echocardiography in models of cardiac disease in the mouse. Circulation (1996) 94:1109–1117.
[Abstract/Free Full Text] - Chien K.R. Cardiac muscle diseases in genetically engineered mice: evolution of molecular physiology. Am J Physiol (1995) 269:H755–H766.[Web of Science][Medline]
- Baker D.L., Hashimoto K., Grupp I.L., et al. Targeted over expression of the sarcoplasmic reticulum Ca2+-ATPase increases cardiac contractility in transgenic mouse hearts. Circ Res (1998) 83:1205–1214.
[Abstract/Free Full Text] - Huang X., Pi Y., Lee K.J., et al. Cardiac troponin I gene knockout: a mouse model of myocardial troponin I deficiency. Circ Res (1999) 84:1–8.
[Abstract/Free Full Text] - Wang J., Wilhelmsson H., Graff C., et al. Dilated cardiomyopathy and atrioventricular conduction blocks induced by heart-specific inactivation of mitochondrial DNA gene expression. Nat Genet (1999) 21:133–137.[CrossRef][Web of Science][Medline]
- Bittner M., Meltzer P., Chen Y., et al. Molecular classification of cutaneous malignant melanoma by gene expression profiling. Nature (2000) 406:536–540.[CrossRef][Medline]
- Vaarala M.H., Porvari K., Kyllonen A., Vihko P. Differentially expressed genes in two LNCaP prostate cancer cell lines reflecting changes during prostate cancer progression. Lab Invest (2000) 80:1259–1268.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Steenman, Y.-W. Chen, M. Le Cunff, G. Lamirault, A. Varro, E. Hoffman, and J. J. Leger Transcriptomal analysis of failing and nonfailing human hearts Physiol Genomics, January 15, 2003; 12(2): 97 - 112. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. Juhasz, Y. Zhu, R. Garg, S. V. Anisimov, and K. R. Boheler Analysis of altered genomic expression profiles in the senescent and diseased myocardium using cDNA microarrays Eur J Heart Fail, December 1, 2002; 4(6): 687 - 697. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. R. Boheler, J. Czyz, D. Tweedie, H.-T. Yang, S. V. Anisimov, and A. M. Wobus Differentiation of Pluripotent Embryonic Stem Cells Into Cardiomyocytes Circ. Res., August 9, 2002; 91(3): 189 - 201. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||