ABSTRACT
Although mRNA expression analysis may have no direct utility in the analysis of protein complexes, combined with proteomics it may yield useful information for additional target validation experiments. Effective integration of nucleic acid-and protein-based technologies may accelerate the selection of the best candidate target. Whereas the target discovery process described above is largely protein driven, knockout technologies and overexpression strategies rely on molecular biology. In a case where a protein is downregulated in the disease, one can envision the use of inducible gene expression vectors as a quick means to restore protein expression and normal cell function, i.e., reintroduction of a tumor suppressor such as p53. More commonly, an overexpressed disease-specific protein becomes the target for a therapeutic inhibitor strategy. If access to preexisting small-molecule inhibitors are not available or cannot be generated easily, the gene can be readily ablated through antisense strategy. An antisense oligonucleotide can be synthesized or an antisense cDNA can be expressed in the cell in an inducible gene expression vector. In some cases, ribozyme constructs (small DNA fragments with intrinsic endonuclease activity that binds the target mRNA through a small DNA sequence complementary to a coding region in the target mRNA) are more useful. However, the effectiveness of these strategies is again monitored most accurately at the protein level. In particular, for the antisense strategy it is important to determine that the antisense selective blocks the expression of the target protein only and doesn’t result in nonspecific effects on protein expression. In the case of cell surface protein targets, quickly advancing phage display technologies now promise functional inhibitory antibodies within months. With these tools one can now establish if interfering at different levels in the pathway will affect the protein expression profile differently and which resembles the protein expression profile of the normal state most closely. This approach is facilitated by the fact that in most cases whole clusters of proteins rather than a single protein will be changed. By using new bioinformatics tools, clusters of proteins can be used in a specific protein expression map to describe a phenotype, thus increasing the predictive value of the analysis.
