Cell based reporter assays: misleading approach in drug discovery?
Posted October 25, 2010on:
Cell based reporter and in vitro homogeneous assays that are used in drug/drug target screening are considered to be powerful and convenient tools in primary screening of compounds or drug molecules. Research reagent and pharmaceutical/biotechnology companies are heavily involved in developing new technologies and methods for primary screening using homogeneous or cell based reporter assays. We are skeptical about the actual benefits of currently used cell based reporter or homogeneous assays that are vulnerable to false drug/drug target identification. The logical reasons for our skepticism are explained in following paragraphs.
In vitro homogeneous assays may not identify all potential drug candidates: The concept of homogeneous assays to “add, mix and read”, looks very elegant. One can screen thousands of compounds in a day without involving cumbersome and error-prone multistep methods. However, the question that remains unanswered is: -Are these homogeneous assays just for convenience and not really meant for “true” discovery of potential drug candidates? Is there a huge possibility to miss most promising drug candidates by adopting homogeneous assays in primary screening? Let us take an example and try to analyze whether these questions are relevant or not. Luciferase based assays are the most sensitive methods and are being widely used in various drug discovery assays. We will take kinase assay as an example to demonstrate potential advantages and disadvantages of homogeneous assays. Currently, there are few reagent companies who sell kinase assay products that are based on ATP quantitation (ATP depletion based assay) using luciferase reagent. At a glance, this assay looks very promising. The assay is done by performing a kinase assay. ( kinase + substrate + ATP+ buffer components + test compounds (potential kinase inhibitors)) followed by the detection of ATP using luciferase reagent (luciferase + luciferin). The ATP in the kinase reaction drives a luminescent reaction catalyzed by luciferase. Luciferase uses ATP to convert luciferin into oxyluciferin and light. In the presence of kinase activity, ATP is depleted from the reaction; thus, there will be less ATP to catalyze the luciferase reaction resulting in less emitted light. In screening applications for kinase inhibitors, emission of higher light output is scored as positive hits. A careful analysis of this assay raises several questions. The ATP, luciferase and luciferin are the moving targets in this assay. Possibly, test compounds can affect any one of these moving targets, in addition to the target kinase, since test compounds are not removed during the detection step. . If these happens, it will result in false positives or false negatives. Possibly, the effect of test compounds can be tested using control experiments using compounds alone with luciferase reagents. However, if test compounds are hydrolyzed or modified in presence of kinases and if these modified products or bi-products affect any one of the luciferase detection components, this will drastically affect the outcome. It is not possible to speculate that all existing or future compounds will not affect luciferase assay reagents, including non-ATP based Renilla luciferase assay. At the same time, one can argue that tens of thousands of compounds have already been screened using luciferase based assays and there were no reports on false positives or false negatives. How could we confirm this when the assay cannot differentiate between true or false positives/negatives? It is not possible to speculate that compounds that were negative in these assays are truly “negative” and vice versa. This is the major issue with any reporter (enzymes) based assays such as luciferase or beta-galactosidase or beta-lactamase or GFP etc. There are even kinase assays in the market that utilize multiple enzymes in addition to luciferase (e.g. ADP based kinase assays). These assays are more vulnerable to false results because in addition to luciferase reagents additional enzymes present in this assay can become the target for test compounds. Several reporter enzyme based assays are available for various drug screening assays such as cAMP, proteases, caspases etc. There is huge likelihood of getting false positive hits while using two enzyme systems in which inhibitors may find luciferase instead of kinase. We believe, enzyme based reporter assays may not be suitable for homogeneous primary screening, which may result in selecting wrong compounds for further screening by leaving most promising compounds undetected, unless all test compounds were tested parallel using other more predictive assays such as radioisotope based assay for kinases. The convenience of homogeneous assays in primary screening of thousands of compounds can lead to the selection of a wrong drug candidate, which may or may not end up as a true drug.
There are many questions yet to be answered: a) if we are aware of these limitations why the research community is still using in vitro homogeneous assays for drug discovery research? b) Are we using the homogeneous assays since these assays are simply more convenient, need less resources/reagents and less expensive than multistep approaches. The truth is, these assays are not cost-effective in the long term. Reagent companies are more interested in selling their products or marketing their technologies, with or without worrying about the long-term scientific impact of their products.
Cell based reporter assays may not be ideal for drug discovery: Cell based reporter assays that are used in drug/drug target screening are more vulnerable to false identification of drug molecules. In a reporter enzyme based assays, it is not possible to confirm that the positive readout is really due to a) the test compound or b) hydrolyzed products of test compounds or c) induction of alternative signaling pathways by test compounds, which trigger the activation of reporter enzymes. For example, fusion protein based stable cell lines are used in GPCR assays. In a primary screen, a compound may be selected based on its effect on GPCR activation. It is very hard to confirm whether the GPCR activation is not due to the effect of the compound on fusion protein, rather than the actual GPCR protein? The GPCR protein fused with a reporter protein does not behave or have the structural/surface integrity of the same un-tagged GPCR protein. A compound or protein that can recognize the untagged-GPCR in cells may not always recognize GPCR protein fused with another reporter protein and vice versa. In addition to this, expression of foreign proteins, like GPCR fusion proteins, can induce genetic and physiological changes within a designer stable cell line, which is different from normal cell lines. If so, how could we differentiate the effect of a drug on GCPR activation through normal cell signaling pathway, not through new proteins or metabolites that are induced in genetically modified stable cell lines? One should also take account of the genetic modifications that can occur during continuous culture of stable cell line selection procedure. The drug discovery screening assays using a drug target fused with a reporter enzyme or protein may not always mimic the effect of a drug on that particular wild type drug target (without fusion protein). Successful examples can be cited, but, these are not universally applicable. Cellular regulation is highly conserved and tightly controlled, otherwise a single mutation in a protein or a defect in protein-protein interaction would not have led to the onset of human diseases. The fundamental question we need to ask whether a drug molecule selected as a “positive drug candidate” based on its effect on a drug target protein fused with another reporter enzyme will have similar effect or binding characteristics on wild type drug target protein (non-fusion). Researchers have to give more emphasis on the physiological and genetic effects of fusion proteins in cells and how do these changes affect cell signaling pathways, which in turn affect the efficacy of the screening assay?
Unlike in vitro homogeneous assays, it is nearly impossible to have real negative control in cell based assays (considering the effect of cellular components on hydrolysis or degradation of test compounds). This makes even harder to identify potential false positives or negatives in these assays. There are several cell based reporter assays in the market for drug discovery research and these assays should be carefully analyzed for long-term benefits in identifying true drug candidates that are safer to use. On the other hand, cell based assays (without reporter enzymes) can be a powerful tool in analyzing the global effect of any given drugs. If we can analyze total changes within a cell, at protein or nucleic acid or metabolite level, due to the effect of a particular drug, possibly we can understand more about the effect of drugs at physiological or cell signaling level that can pin point some clues to the potential side effects. Innovative technologies that can thoroughly analyze these changes in cell based assays and in vivo animal model studies will have great potential in future drug discovery research. For example, it is possible to develop techniques for measuring cAMP using non-invasive and non-reporter enzyme based methods. If we can measure glucose molecules real-time in human blood using non-invasive methods, similar methods can also be feasible in measuring metabolites such as cAMP, phosphoinositides, P450 actiavtion/inhibition etc. in a cell based system. Along with the activation of cAMP, if we can identify molecular and physiological changes that are induced by test compounds will provide critical insights into the possible side effects of drugs. Such innovative approaches in drug discovery research will give us hope in developing safer drugs in future.
Are there any solutions? Obviously, one can ask that we have questioned the use of assays that are being widely accepted by the drug discovery research community. Are we right? If so, what are our solutions? From our point of view the questions we raised are relevant to current drug discovery approaches. Some existing solutions: a) instead of homogeneous or cell based reporter assays (exceptions can be based on targets), researchers may try to adopt techniques or methods that involve two or more step procedures for compound screening. It is true that this approach is not convenient like homogeneous assays, but multiple step methods can be automated like homogeneous assays. For example, radioisotope based kinase assays are widely used to confirm the kinase activity. However, this assay is not amendable for high throughput format because of safety and waste disposal issues. Peptide or solution array (glass slide or 96/384 wells) based assays coupled with antibody or phosphospecific detection reagent can identify “true” potential drug candidates than homogeneous assays. In these assays test compounds and other components are removed prior to the addition of detection reagents. Moreover, these are direct assays, phosphospecific, rather than currently used indirect kinase assays. Also, innovative technologies need to be developed for the detection of phosphate groups without using antibodies.
b) We need to develop new methods or modify existing methods for the applications in drug discovery. A very good example is ADME/Tox assays. Mass spectrometry (MS) assays are widely used with accuracy and reliability. These technologies can be adopted for various drug discovery assays rather than using reporter enzyme based assays. Mass spectrometry based assays can detect direct effects of drugs without manipulating cells or in vitro conditions. Similar MS assays can be developed for several drug discovery screening assays, which can quantitatively identify true changes in metabolites or proteins due to the drug treatment. HTS assay methods needs to be developed for most of the MS based drug discovery assays, which include cAMP, phosphoinositides, lipids, proteases, caspases etc. Multistep antibody based assays, not homogeneous assays, or antibody arrays are also very powerful tools in addition to MS assays. Non-specificity and cost are the major hurdles in using antibody based HTS assays.
c) Non-invasive methods will be the most powerful future tool in drug discovery research. These assays do not need manipulation of cells or the use of multiple regents for the detection. Recent developments in Raman spectroscopy and related areas will open up new avenues for non-invasive methods for the detection of metabolites, nucleic acids and proteins within a cell or tissue or organ or even in whole animals. These assays can be accurate, predictive and cost-effective. Reagent and instrument companies should invest their innovative minds to develop such non-invasive methods for drug discovery research. Academic and industrial researchers should come up with innovative non-invasive drug screening assays that can help in generating safer drugs in most efficient way.
End Note: It is not our intention to invalidate all innovative technologies that have been developed by very creative and talented scientists. These technologies have helped in finding new ways to screen drugs molecules that led to several successful drugs in the market. However, it is also true that currently used homogeneous or cell based reporter assays (that are used in drug/drug target screening) do not identify all potential drug candidates. Adoption of these assays by more and more researchers/laboratories worldwide will have negative impact on new drug discoveries. The selection of drug screening assays should be made based on long-term scientific impact of these assays. It is also the responsibility of research reagent companies, who develop and market these kinds of products, to analyze the “true” drug discovery value of their products. Based on recent reports on life threatening side effects of so many valuable drugs it is the right time to argue about existing pre-clinical and clinical drug screening assays and approaches. For example, antidiabetic drugs can cause cancer, heart attacks and so forth. It is really a night mare when we think that a drug for a particular disease can cause several other diseases and side effects. On the basis of these observations, we need to accept the fact that there are issues with current pre-clinical and clinical drug screening assays or methods. In other words, current drug screening assays do not provide sufficient information on potential life threatening side effects of drugs. It is possible to develop novel cell based or animal based drug screening technologies and tools that can predict fatal side effects, possibly by identifying cell based biomarkers (proteins, metabolites, miRNA etc.). We believe radical changes are needed in pre-clinical or clinical drug screening assays. Both industrial and academic researchers need to come up with innovative ideas to address these problems in current drug discovery approaches. Recent HTS initiatives by NIH are excellent avenues for addressing true drug discovery value of existing assays. Regulatory agencies like FDA might need to revisit the relevance and applications of existing pre-clinical and clinical assays in context to potential side effects of drugs.