April 1, 2012 (Vol. 32, No. 7)
Josh P. Roberts
A good assay, they say, is the stuff of good science. Whether it’s a whole new type of test, a twist or a tweak of an old one, or a way of combining things that hasn’t been worked out before, assay development remains an integral—if sometimes behind-the-scenes and underappreciated—part of the discovery process.
Just ask participants of the Assay Development and Screening track of the recent “SLAS 2012” conference. Researchers there discussed the challenges of solubilizing fatty acids while avoiding toxicity, and ways to optimize conditions for high-throughput screens of libraries of molecular inhibitions to generate more reliable hit lists.
Some told of ways to circumvent or directly challenge the idea that some targets may be “undruggable”, while another spoke of uncovering previously unseen mechanisms of action by eliminating an inhibitor built into the standard assay.
Lawrence Wiater, Ph.D., is working on cell-based assays to systematically investigate and characterize metabolism of fatty acids in mammalian cells. Biolog, for which he is a senior scientist and group leader, currently offers a series of 96-well microplates (Phenotype MicroArrays™) that look at metabolic effects of carbon and nitrogen substrates, ions, hormones, metabolic effectors, and anticancer agents. The fatty acid metabolism assays, which are currently in beta testing, will be an extension of that platform.
The assays can help researchers understand what pathways cells use to metabolize substrates that are in the wells, to look at the effects substrates have on cell growth, or to see if they can increase the productivity of a metabolic product in bioprocessing.
“You can just add your cells to hundreds of wells and then you can kinetically monitor each well for increased or decreased productivity of your favorite molecule,” Dr. Wiater pointed out. “You typically don’t know what may affect your pathway of interest. These microplates allow you to screen hundreds of nutritional factors that could modulate that productivity.”
Disease research, too, can benefit from the company’s Phenotype MicroArrays. “Energy metabolism is linked to obesity, diabetes, nutrition, aging, mitochondrial diseases, drug toxicities—especially those that target mitochondria—and then cancer and cachexia,” he said. “Our fatty acid microplates offer additional pathways you can probe to look for relevance in any of these health problems.”
Biolog’s OmniLog™ instrument incubates the microplates at 37°C while it reads and records the linear reduction of a tetrazolium dye to a colored formazan, thereby measuring the rate of metabolism in each well. Currently techniques such as mass spectrometry and liquid chromatography can generate a snapshot of metabolic pathways pools, but Dr. Wiater “doesn’t know of any other technology platform that can measure metabolic rates of fatty acids and other cell energy sources.”
Inhibit the Inhibitor
Reactivation of telomerase allows cells to become immortal, and as such is seen as an attractive anticancer target. Yet the molecular architecture of the catalytic reverse transcriptase (hTERT) of telomerase makes it essentially undruggable. Despite 15 years of trying—and a lot of money, time, and effort—no small molecule-based inhibitors of hTERT have made it to the clinic.
Cancer Research Technology (CRT) medicinal chemist Jon Roffey, Ph.D., uses a more indirect approach to find inhibitors of telomerase in his cell-based assays. He and his collaborators set out to “look for a network of druggable pathways for therapeutic exploitation of a nondruggable target,” he explained.
Using the hTERT promoter cloned by principle investigator W. Nicol Keith, Ph.D., of the University of Glasgow as the basis of a standard luciferase reporter assay, they tested the effects of 79 well-characterized kinase inhibitors. “If you can inhibit anything in that pathway, and you can inhibit promoter activity, you can see a decrease in luciferase activity,” said Dr. Roffey.
Six compounds were found that did just that, three of which were known to inhibit the activity of the enzyme glycogen synthase kinase 3 (GSK3).
But the problem with cell-based assays and reporter-gene assays is that there are multiple pathways that can lead down to the promoter, and so specificity is key. They weeded out compounds that had nonspecific effects on luciferase such as general transcription factor inhibitors, and they utilized viability assays within the cascades because it should take multiple cell cycles before shutting down telomerase would kill the cells.
Ultimately they introduced their inhibitors to the endogenous system. Using both siRNA and small molecules in a panel of cancer cell lines, the researchers were able to show that inhibiting GSK3 led to a reduction in telomerase message, a reduction in telomerase catalytic activity, and ultimately (over the course of many days) a shortening of the telomeres themselves.
The GSK3 work, Dr. Roffey explained, was not the end in itself so much as a proof of principle. It was “basically saying that the promoter assays can be used to find compounds that can modulate the pathways that modulate telomerase expression.”
Some targets are considered “undruggable” because what’s known about their structure says that they lack traditional binding pockets for small molecules or the native ligand binds too tightly to be out-competed. X-ray crystallography, the preferred way of discerning the structure of a protein, freezes a protein into one particular conformation.
But proteins are plastic, points out Joshua Salafsky, Ph.D., CSO of Biodesy: “You’re not able to see all the conformations it’s adopting under physiological conditions, and therefore, just using crystallography, you’re unlikely to find drugs that perturb the conformation in specific ways.
“There are likely transient pockets that open up that aren’t visible in the crystal structure that you would like to develop a molecule to bind to and stabilize a particular conformation of the protein, to render it inactive, for example. Or you’d like to develop an allosteric drug, again that binds to a pocket that may or may not be visible in the crystal structure.”
While a post-doc at Columbia University, Dr. Salafsky developed a novel way to detect biomolecules, based on a technique used in physics and physical chemistry research called second harmonic generation (SHG), by labeling them with SHG-active dyes.
Subsequently, at Biodesy, the company he founded, he developed this advance into a tool to monitor a protein as its conformation changes in real time. Upon excitation, immobilized biomolecules labeled with SHG-active dyes will re-radiate two photons of red light as a single photon of blue light (the “second harmonic”).
The key, he said, is that the amount of blue light produced is very sensitive to the orientation of the dye, and so “we can detect a very small shift in the average orientation of the probe due to protein conformational change.”
By labeling a protein at a specific amine or cysteine, different parts of the protein can be monitored.
Biodesy recently used this technique to identify activators and inhibitors of Ras activity. “We not only were able to show that these compounds in fact did change the conformation of Ras directly, but that the label site also told you whether conformation changed at that particular site,” Dr. Salafsky said.
“It was really exciting that the two active compounds in our hands that changed conformation of Ras were also the two inhibitors that other people had found in cell-based assays, and that the third compound that had no effect in the cell-based assays had no effect on the conformation, in our case, either when Ras was labeled at the cysteine or at the amines.”
Two for the Price of One
But sometimes a new assay is literally just two established techniques wrapped up together.
Due to the high cost associated with high-throughput screening, “people usually do not run two assay platforms for the same target with the same library,” said Yuhong Du, Ph.D., associate director of the Chemical Biology Discovery Center at Emory University. “You have to select one.”
Both fluorescence polarization (FP) and time resolved fluorescence resonance energy transfer (TR-FRET) techniques have been widely used in high-throughput screening for many years. Each has its drawbacks, each generates hit lists that tend to only partially overlap, and each acquires it own set of data: FP measures the rotation of the molecule (which is slowed by interaction with another molecule), while FRET measures the proximity of the FRET donor and acceptor molecules. Many benchtop multilabel readers can read both.
Because the two assays do not interfere with one another, Dr. Du and her colleagues decided to combine them together “in one well at the same time, to monitor the same interaction, to give us an information-rich primary HT screening,” she explained. “You don’t need to run two assays in parallel, instead you run two assay platforms in the same well.”
Of course, if it was just a matter of taking two readings, someone would have come up with the Dual-Readout F2 assay a long time ago. To satisfy the requirements for both, components in the reaction require tedious optimization, generally involving bidirectional titrations of two binding partners. And sometimes compromise: for example, more protein may be needed to boost the signal for one readout, yet it saturates that of the other.
As a model study, F2 was used to screen a library of 100,000 compounds that disrupt the interaction between the Mcl-1 anti-apoptosis protein and its known inhibitor peptide Noxa.
By taking only the compounds that were positive in both assays, they were able to efficiently prioritize their hits based on behaviors of positive compounds in these two platforms for a follow-up study. And, Dr. Du pointed out, simply repeating the same screening with a single readout would not add significant information to the hit list, as was achieved in a single step with the F2 assay.
Enzymology Is a Competitive Business
When Eduard Sergienko, Ph.D., was looking to set up appropriate assays to identify modulators of tissue-nonspecific alkaline phosphate (TNAP), he realized that buffers such as diethanolamine (DEA), traditionally used in alkaline phosphatase (AP) assays, can actually participate as a substrate in the reaction, re-directing it toward transphosphorylation.
This could pose a problem since a high concentration of DEA routinely employed for boosting the assay sensitivity would saturate the reaction, making it nearly impossible to find competitive inhibitors.
“The physiological significance of the transphosphorylation reaction and the identity of its alcohol substrate are still unknown,” the director of assay development at Conrad Prebys Center for Chemical Genomics (CPCCG) at Sanford-Burnham Medical Research Institute said. “We wanted to find compounds that would inhibit different potential enzyme activities and thus wanted to have compounds with various mechanisms of action.”
Dr. Sergienko and his colleagues adapted CDP-star, a chemiluminescence substrate widely used for detection of AP in blotting, allowing them to eliminate DEA if they so chose. The assay was sensitive enough, and with a large enough dynamic range, to screen at low enzyme concentration and with the substrate at Km. “One can emphasize a certain mechanism of action by playing with the concentration of substrate relative to the Km level,” he explained.
They screened TNAP against the Molecular Library Screening Center Network (MLSCN) collection containing 64,394 compounds and found three major categories of inhibitor structural scaffolds along with some minor categories and singletons. One of these scaffolds contained the first competitive inhibitors for TNAP, and for APs in general.
Only four of the 55 hits identified by the high-throughput luminescent assay screen could be confirmed in a colorimetric assay performed with high DEA concentration—giving evidence that it is perhaps directly competing for the enzyme’s binding site.
To support TNAP lead-optimization efforts at CPCCG, Dr. Sergienko and his team have gone on to develop a biomarker assay for AP activity within blood plasma that can be run at physiological pH using small-volume aliquots of minimally diluted plasma samples—making it amenable to high-throughput methodology.
This approach allowed testing of structure-activity relationship compounds and predicting their efficacy in vivo. In addition, the biomarker assay allows monitoring activity of TNAP in the samples taken from animals after dosing them with TNAP inhibitors in pharmacokinetic studies, providing pharmacodynamic information for animal models relevant to TNAP.