Van Eyk, Dunn - Proteomic and Genomic Analysis of Cardiovascular Disease - 2003 (522919), страница 13
Текст из файла (страница 13)
tTA is a transcription factor that regulates the expression ofany gene downstream of its cognate promoter sequence (tetOp). Therefore, a second mouse line containing the desired transgene downstream of tetOp is bred tothe first mouse line, creating a so-called ‘tet-off’ system, in which the interactionof tTA and tetOp and thereby the expression of the gene of interest is inhibited inthe presence of tetracycline.
If a mutant form of tTA, reverse tTA (rtTA), which isactivated (i.e., will bind to the operator DNA) rather than repressed by tetracycline, is used, tetracycline can be administered to the mice only for the period of31322 Cardiac Disease and The Transcriptometime when expression of the transgene is desired (‘tet-on’) [33].
The inducible system can be combined with the Cre/lox in order to provide precise control of cre expression and has been successfully used to control gene ablation and transgeneexpression in mouse hearts [27, 31, 32]. The current technology, while somewhatflawed in terms of its reproducibility and “leakiness” nevertheless offers the potential of improving dramatically our ability to precisely affect and control the transcriptome, decreasing the inherent biological noise of globally based RNA analyses.2.3Is The Mouse A Valid Model for Human-Based Disease Transcriptome Studies?When considering the validity of in vitro models for studies of human diseasemechanisms, the limitations are clear.
In vitro models are designed to reduce acomplex situation to the supposedly most relevant components, which can thenbe tightly controlled under the chosen experimental conditions. This simplification is also the major shortcoming of all in vitro approaches and determines thevalidity for a particular model. However, the question of validity cannot be answered as easily when genetic manipulations are discussed.
With increasingly sophisticated technologies, we can now study the consequences of alterations in thegenome in the whole animal. The genomic alteration is clearly defined and controlled – but we study the consequences in a complex system, in which many potentially interfering variables are unknown. Consequently, the only criterion of validity is the generation of the ‘correct’ phenotype. Hence, the conclusions dependto a large extent on the thoroughness and quality of the phenotypic characterization. As far as the new miniaturized technologies allow us to detect, mice appearto be a valid model system for a number of human disorders [34].
However, themouse heart clearly differs from the human: the basic motor protein complement, calcium handling and the electrophysiology are all distinctly different. Another source of difficulty in interpreting murine data are the dramatic strain-dependent phenotypes. Depending on the genetic background, both penetrance andthe characteristics of a particular mutation can vary considerably [35, 36]. Thisstrain dependency is thought to be due to modifier loci [37, 38] but little is knownabout these elements. Does the lack of a phenotype after genetic manipulationdemonstrate that the altered gene does not play a critical role for the disease of interest? Or does it merely indicate that the response to the genetic alteration canonly be detected in an alternate mouse strain? Similarly, the positive result of aparticular study alone does not necessarily prove its general validity, as the strainused could be predisposed to a pathology that normally would not present.
Thedifferent strains have, under baseline conditions, different transcriptome profiles,which add further to the difficulties in comparing data sets between labs, evenwhen supposedly identical strains are being used. Therefore, every individualmouse model has to be carefully tested by designing appropriate internal and external controls.2.4 Arrays and Cardiovascular Disease2.4Arrays and Cardiovascular DiseaseThe Good: Why carry out transcriptome analysis? Altered gene expression hasbeen a focus of biomedicine ever since it became apparent that a cell responds toexternal and internal stimuli by altering its genetic output.
In the past, experiments were often designed to monitor the appearance/disappearance of a particular transcript or set of transcripts in response to a well-defined stimulus. Thelarge-scale interrogation of cardiac mRNAs offers a relatively easy way of carryingout a first pass at understanding gene function in the heart. Instead of being restricted to the parallel processing of 1–20 gene transcripts via traditional hybridization techniques such as Northern analysis or dot blotting, the experimentalistcan assess the expression profile of an entire genome in one experiment.The Bad: However, the inherent limitations of the transcriptome-based studiesare daunting.
First and most importantly, the transcriptome is not representativeof the proteome, as many transcripts are not being actively translated. Second,contextual information is lost, as the phenotype of the cell is highly dependentupon the protein-protein interactions: information that is not present in the transcriptome data.
Third, in terms of cell function, a protein’s value, like that of realestate is often dependent upon location and its state of being. There are ample examples of protein activity and concomitant biological function being highly oreven completely dependent upon post-translational modifications, which eithermodify protein-protein interactions or are responsible for shuttling the proteinfrom one part of the cell to another [39]. Thus, the fact that a transcript is presentor upregulated does not necessarily translate into altered biological activity at theprotein level and conversely, the lack of any observed regulation, when the transcript(s) in question is compared to an unstimulated control, does not mean thatthe cognate protein’s activity has not been significantly modified, either by achange in its subcellular location, or modification of its inherent catalytic activity.Finally, it is now abundantly clear that the “1 gene-1 protein” dogma deducedfrom studies in simple procaryotic systems no longer holds.
Rather, “1 gene” cangive rise to many polypeptides of radically different function, through alternativesplicing of the primary transcript and/or post-translational processing [40]. Thus,interrogation of a transcript is often confounded by questions of whether a particular exon is represented in the array or if the probe sequences can distinguish between closely related, but functionally distinct isoforms.The Ugly: The multiple experimental designs of different experimentalists, whileperhaps logical for each group’s purposes, make it impossible to meaningfullycompare results between them. Thus, widely disparate numbers of up- or downregulated genes in various disease models of the heart exist, with only marginaloverlap apparent between the data sets. This is to be expected during the earlyphases of a technology’s development, as the laboratories struggle to assimilatethe technique into their own experimental programs.
Nevertheless, a certain degree of rigor has been lacking as laboratories leap headfirst into the “bag ofchips,” with experiments not taking into account both the inherent noise of the33342 Cardiac Disease and The Transcriptometechnique, which dictates multiple, independent sample preparations and experimental duplication, and biological noise, which leads to a significant amount ofscatter in the data. Both sources of error can be significant but solvable, throughthe use of multiple experiments carried out along a well-defined experimentaltimeline as the phenotype develops. Thus, rather than depending upon a snapshot of data at a single time point, a collection of points during a period of timewill be much more valuable in actually determining whether a particular transcript’s up- or down-regulation is associated with an aspect of a pathological state.An obvious application of the technology is to uncover the changes that precede, accompany and result from the expression of a primary genetic insult.
Numerous catalogues now exist, both in real [41, 42] and virtual space (http://www.nhlbi.nih.gov/resources/pga/). The first point to consider is that even at thisearly stage of cardiovascular disease-based global genomics, we are drowning indata. At the authors’ institution (http://www.cincinnatichildrens.org/Research/Research_Cores/Pediatric_Informatics/BioSoftware/default.htm), approximately 10gigabytes of data are collected and archived every week. Second, because of thedifferent methodologies, the different investigators and filters that are being applied both before and after data acquisition, there is depressingly little agreementconcerning which genes are up and down regulated.
As alluded to above, the lackof detail in published experimental protocols and post-data acquisition algorithmsmake it impossible to actually compare cardiovascular data sets that are posted inthe literature or in virtual space. Third, very few, if any of the analyses are trulycomplete. Rather, most consist of a very limited number of points in the diseasemodel and the interrogation is rarely comprehensive. Fourth, validation is oftenlacking or incomplete, consisting, for the most part, of confirming the general upor down-regulation of a gene either by in situ hybridization or Northern analysiswith the cognate cDNA or oligonucleotide probe. Fifth, gene annotation is also incomplete, with the data presented either as parsed, column based lists with an abbreviated one phrase description, or placed into functional groups on the basis ofrather ill-defined criteria. Only very rarely are comprehensive analyses presentedand the data filtering rigorously annotated, so that the overall quality of the categorization can be objectively determined.In short, we are for the most part at the level of preparing catalogues, with thebiological values waiting to be assigned in the future as the initial filters identifycandidates for further validation.
![](https://s.studizba.com/z.php?f=/templates/si/i/noavatar.png&w=100&h=100&t=1"){kind=avatar}
