Target Discovery and Validation

In pharmaceutical proteomics the questions related to drug targets can be distinguished in two main types: compound mode of action on the one hand and disease mechanisms on the other. Sometimes an interesting substance is discovered in a phenotypic screening process, as it is used in the search for cytotoxic or cytostatic drugs, e.g., in oncology or anti-infectives research. The output of such a screening is a compound or a small family of compounds with the desired phenotype, i.e., it kills tumor cells or bacteria, but the mode of action, in other words the relevant cellular target, is unknown. A typical case could be a natural product, either purified or even as component in an extract from traditional medicine. Without knowledge about the target, medicinal chemistry efforts to improve the properties of the compound are not very fruitful. There are several ways of addressing this question. One is the chemical proteomics

Biomarker activities Focus on throughput

Research Preclinical Phase I Phase II/III

Target discovery activities Focus on information content

Figure 5 Illustration of the different proteomics activities in relation to the drug discovery process.

approach: the compound itself is used as an affinity bait to capture the target protein from an extract of the appropriate cell or tissue. This strategy will be discussed in more detail in the next section. More often, expression profiling is employed in the hope that the observed changes in protein expression will lead to the target protein. An illustration of how this can work is the identification of methionine aminopeptidase (MetAp) as the target of bengamides, a study carried out in the authors' laboratory.94 Bengamides are natural products originally isolated from marine sponges with potent anti-tumor activity in vitro and in vivo. Their pattern of inhibition in a panel of cancer cell lines turned out to be quite distinct from other compound classes, suggesting that the activity of bengamides was due to the interaction with a novel target. A protein expression analysis of the H1299 small lung cell carcinoma cell line treated with the lead bengamide derivative LAF389 revealed alterations in a subset of about 20 proteins by 2D electrophoresis, which were readily identified by PMF In-depth biochemical and mass spectrometric analysis94 showed that LAF389 blocked the removal of the N-terminal methionine on one of the identified proteins, 14-3-3g. This pointed to a direct or indirect inhibition of the involved enzyme, MetAp. Follow-up experiments showed that the N-terminal modification was indeed specific for this compound class and the ultimate proof was provided by the observation that the compound inhibited and cocrystallized with the purified MetAp enzyme. The discovery of MetAp as the target for bengamides was a major step forward for the medicinal chemists, who could now further optimize the compound, starting from the crystal structure of the target instead of a cell proliferation assay. At the same time, it is clear that much more effort would be required to unravel the actual mechanism of cytotoxicity by the bengamides. Parallel proteomic approaches could be employed, e.g., whether the different forms of 14-3-3g interact with different proteins, or whether there is a larger set of proteins with altered N-termini, among which the true effector of cell cycle arrest could be found.

Probably the foremost expectation that accompanied the advent of large-scale expression-profiling methods was that completely novel biological insights would be generated at a much higher speed, including a better understanding of disease mechanisms, especially in complex diseases, where current therapies are often unspecific or only symptomatic. A good example of such a disease is schizophrenia.105 Schizophrenia has a notable genetic disease component and hence a significant number of chromosomal 'hot spots' have been identified in families with a distinct prevalence of the disease. However, because of the complexity and heterogeneity of the disease, none of these hot spots has yielded a clear connection to a disease mechanism. A plausible reason for these inconsistencies is that those linkage studies are limited to the genetic predisposition for the disease, not taking the important environmental aspect into account. And of course schizophrenia is just one, albeit highly complex, example out of a long list of diseases that result from the interaction of predisposition and environment, including cancer, cardiovascular disease, and depression. Therefore it is not surprising that for many investigators expression profiling represented a paradigm shift, the first 'top-down approach' to complex disease: the investigation of the disease phenotype(s) at the molecular level, more or less decoupled from all the factors that led to that disease state.

Consequently, many proteomic studies have been undertaken to characterize psychiatric disease,105 cardiovascular disease,106 and, most of all, cancer,107 just to name the areas of human disease, where arguably the highest need exists for novel therapeutic approaches. If all these studies have anything in common, it is probably the fact that virtually none of them has gone beyond the stage of describing differences. Hence the biggest lack is the validation of the long lists of potential biomarkers or even targets to prove that they are consistently (dys-)regulated and thus might be part of a disease mechanism or at least a useful diagnostic. There are several reasons why it is so difficult to turn protein expression data into relevant biological information. The choice of samples is one of them. Tissue samples from patients are clearly the most relevant but unfortunately also the most difficult to deal with, because of the inherent variability. And, as discussed before, expression profiling is most powerful with homogeneous samples. At the protein level that is even more pronounced than at the mRNA level, because the proteome is at least an order of magnitude larger, while at the same time the number of variables (i.e., proteins) that can actually be measured is much smaller. Moreover, if the number of proteins is not enough of a problem in its own right, there are many levels of additional heterogeneity, both with respect to the patients (age, diagnosis, comorbidities) as well as to the samples themselves (dissection, postmortem effects, mixture of different cell types). In other words, at the current level of coverage and throughput, a proteomics analysis of a disease phenotype in patient tissue is still a major challenge. In animal models many of the problems of heterogeneity are significantly reduced. However, the caveat is that if one is to extrapolate the molecular phenotype of an animal model to human disease, there has to be a certain level of overlap between the pathways involved. And that is hard to achieve if there is little knowledge about the disease mechanism in the first place. As more knowledge on biological pathways becomes available,98 it will become easier to derive such animal models and more rewarding to study them with expression-profiling methods.

In conclusion, target discovery is probably the one area where proteomics has not yet fulfilled the sometimes unrealistic expectations. With current technology, a comprehensive analysis of protein expression is feasible in simple homogeneous systems, such as prokaryotes, and a substantial coverage can still be obtained in cell culture models derived from higher organisms. However, there is still some ground to cover before we will be able to perform an in-depth investigation of patient material to discover novel therapeutic targets. Moreover, analysis of protein expression levels is only the starting point for proteomics, because most of the functionally relevant events occur posttranslationally. Therefore, expression proteomics experiments should be combined with approaches that aim to elucidate relevant biological pathways in order to deliver a meaningful contribution to target discovery.

Your Heart and Nutrition

Your Heart and Nutrition

Prevention is better than a cure. Learn how to cherish your heart by taking the necessary means to keep it pumping healthily and steadily through your life.

Get My Free Ebook

Post a comment