January 1, 2016 (Vol. 36, No. 1)
Phenotypic Screening Is Being Paired With a Succession of Complementary Technologies
As its image processing and memory capabilities keep improving, high-content analysis (HCA) is emerging as a formidable—and flexible—innovation platform. HCA is being partnered with three-dimensional (3D) cell culture, whole-animal models, systems biology, and genetic editing to drive discovery in basic research and clinical medicine.
For example, various combinations of HCA and other technologies are emerging that look particularly promising in the assessment of cancer risk. Finally, HCA is beginning to incorporate multiwell formats that can also be used with genetic tools to dissect a lead candidate’s mechanism of action.
With the wealth of information available from whole-genome sequencing, one would think that determining a person’s cancer risk would be simple. However, this is not necessarily the case, for several reasons.
One complication is cost. While a genome can be sequenced for $1,000, the interpretation of that genome is still extremely expensive, not to mention unreimbursed in most medical plans. Next, not all genetic mutations involved in cancer are necessarily elucidated to the point of determining cancer risk for a patient. Finally, other elements such as lifestyle, epigenetics, and genetic penetrance come into play.
“One widely accepted method of determining cancer risk is DNA damage repair. Phenotypically, this can be seen over time, looking at the rate of recovery from double-stranded DNA breaks,” states Sylvain Costes, Ph.D., a principal investigator at Lawrence Berkley National Laboratory and the CTO at Exogene Biotechnology.
“By measuring DNA damage over time, we can determine an individual’s cancer risk,” expounds Dr. Costes. “This physiological measurement takes into account the genetic background and environmental factors, such as diet, of the person being tested. It does not address why a person has problems repairing DNA, just how severe the repair problem is.”
Blood samples from persons who have been told that they may be at higher risk of cancer are exposed to chemicals and radiation to induce double-stranded DNA breaks. The damaged cells are then allowed to repair themselves for various lengths of time, before the cells are fixed and stained for double-stranded DNA breaks.
With a special custom-built microscope, Exogene can scan 384,000 wells in three dimensions in approximately 20 hours.
“While a typical genetic test sequences a specific gene, using a phenotypic assessment takes into account both genetic background and environmental issues,” continues Dr. Costes. “Ideally, we see this test being used to assess cancer risk for people who have been identified at a higher risk of cancer by their physicians.
“Just think of this as a test, like one for high cholesterol—if your doctor finds you have high cholesterol, then she can work with you to lower it. No need for expensive genomic testing to determine why, especially when scientists still do not know all the gene polymorphisms leading to poor DNA repair.”
At present, Exogene is working on developing clinical trials to test this idea. Ultimately, this may prove to be a less expensive way of determining overall risk of cancer for an individual without having to look at specific genetic mutations at numerous loci.
HCA + 3D Cultures for Cancer Drug Discovery
Another innovator in HCA is Daniel LaBarbera, Ph.D., associate professor of drug discovery and medicinal chemistry at the University of Colorado. His lab specializes in using 3D tissue culture and organoids to screen compounds for drug discovery. One area of particular interest for him is the reversible process known as the epithelial-to-mesenchymal transition (EMT).
“When certain cancers metastasize, the cells must transform from an epithelial state (which is localized) to a mesenchymal phenotype (which is motile and invasive). This also is true in a variety of other conditions such as wound healing, diabetes, and liver and lung fibrosis,” explains Dr. LaBarbera. “We have been screening natural products libraries for compounds that inhibit or reverse EMT.
“Finding a compound that inhibits the mechanisms promoting EMT could be applied to several situations, in addition to preventing cancer metastasis. For example, age or diabetes can lead to cataract development, which in turn can relapse within a few years of cataract surgery, a complication known as posterior capsule opacification, or PCO. Having a medicine to prevent PCO would be a tremendous health benefit to millions of cataract patients worldwide.”
“Using a 3D model to screen against EMT is a better representation than a 2D system of what goes on biologically,” maintains Dr. LaBarbera. “In 3D cultures, microenvironment components are present that can help recapitulate the tumor. A 3D spheroid can consist of multiple cells types with various zones of oxygen and nutrient gradients, just as a tumor in humans.”
Screening a library of natural compounds has led to a good candidate. “We found a candidate that induces reversion of EMT and correlates with potent antitumor and anti-metastatic activity both in vitro and in animals,” exclaims Dr. LaBarbera. “This is based on a novel biological action against a well-characterized protein complex that regulates EMT genes. My lab will be publishing a paper with details in early 2016. At present, we are working on optimizing this lead through medicinal chemistry.
“This is very promising for treating advanced cancers using a combination drug treatment approach: tumor cells that have undergone EMT are very resistant to traditional cancer drugs. If we can force the tumor cells to revert back to a more vulnerable state (that is, the epithelial state) using drugs that target EMT, then we can also employ established cancer drugs to effectively eradicate tumors.”
HCA + Whole Animal Models for Neurodegenerative Diseases
Zebrafish is the answer for a whole-animal model that can be used in a high-throughput context for screening compounds’ effect against neurodegenerative diseases. “Assessing what is going on deep inside the brain is a challenge,” expounds Su Guo, Ph.D., professor of bioengineering and therapeutic sciences at the University of California, San Francisco. “Utilizing a transparent animal that can grow inside a 96-well format permits easy evaluation of cells within the brain.”
As zebrafish develop from the egg, they spend a week in the larval stage, living off the nutrients from the egg yolk. Dr. Guo has pioneered a chemical-genetics model that permits the destruction of specific neurons within the developing brain. Coupling this with a fluorescent reporter allows her to test a large number of compounds for their protective activity.
“The limiting factor in the screening experiments is getting enough eggs,” notes Dr. Guo. “We don’t have to worry about feeding the larvae as they develop. Adding the compounds to be screened directly to the wells makes this a simple, easy-to-use format.”
Traditionally, screening compounds for activity in brain diseases has been a challenge, as getting tissue culture cells to mimic neurons is problematic. Dr. Guo’s zebrafish system allows a readout of the substances in vivo, directly assessing neuronal survival in the phenotype. “This type of screening experiment would be prohibitively expensive to do in mice,” comments Dr. Guo. “Cost-wise, employing zebrafish is much more doable.”
The Guo lab has pioneered a genetic system that can be turned on or off to produce a neurotoxin in specific neurons. Small molecules are added to the wells containing the developing zebrafish larvae to determine if they can offer protective value against the induced neurotoxin. Survival of the neurons is determined by looking for a fluorescent signal from them.
Dr. Guo’s lab can image over 10,000 animals in one day. Currently, the Guo lab is validating hits from a screen looking at small compounds the prevent cell death in neurons that produce dopamine. This is a model for Parkinson’s disease. Papers are in progress, describing the system in detail.
With 3,000 compounds screened to far, two hits have been validated.
HCA + Systems Biology for Cardiac Medicine
Another lab innovating in HCA is that of Jeff Saucerman, Ph.D., associate professor of biomedical engineering at the University of Virginia. Bringing together two robust approaches, HCA and systems biology, Dr. Saucerman illuminates novel pathways and targets relevant to cardiac medicine.
“We study how the heart remodels in response to stress,” states Dr. Saucerman. “Employing systems biology allows us to fit all the pieces together in networks. This back and forth between computational and experimental models is especially fruitful.
“By developing our own system, with an automated microscope and image analysis, we experimentally validate the predictions from our computational models. We integrate this data to infer new molecular networks.”
“Applying this approach enables us to tease apart the signaling pathways involved in cardiac hypertrophy,” continues Dr. Saucerman. “We found a number of surprising new correlations between factors controlling cell shape and proliferation. To show causation, we used a genetic knockdown method. The results identified a new pathway involved in cardiac hypertrophy.
“Using this approach of high-content screens to find correlations, genetic knockdowns for causation, and then proteomics to map out pathways guided our lab to several interesting discoveries. A high-content approach permits us to characterize many phenotypes at once, as opposed to a single readout on a plate reader; hence the power of HCA.”
More information can be gained by tracking an individual cell in culture up to a week with automated live imaging. Experiments can become very large, with millions of images.
“Fortunately, we could simply pick up our server, carry it over to the University of Virginia supercomputer center, and connect it directly to their system,” informs Dr. Saucerman. “Transferring all that information over the Internet would have taken weeks.”
HCA + Genetic Editing for Fast Analysis of Toxicity Pathways
Another innovation in HCA is combining it with DNA-editing technology. Krishanu Saha, Ph.D., assistant professor of biomedical engineering and bioethics at the University of Wisconsin-Madison, is employing this tactic to discern the mechanisms of action for several substances.
“In collaboration with the EPA center for Human Models for Analysis of Pathways (HMAP), we are maturing reprogrammed pluripotent stem cells to get heart, brain, and liver cells,” elucidates Dr. Saha. “By then screening these cell types with the EPA TOXcast library, we can get a signature of toxicity for each of the substances. These capabilities will allow future experiments to determine the toxic nature of unknown compounds.”
Machine learning is employed to find patterns of toxicity in the cells, based on cell morphology. Additionally, RNA from the cells can be quantitated to add information to the images. “With the EPA HMAP center team, we are compiling images of cell health,” states Dr. Saha.
Finding the mechanism of action in cell toxicity can be difficult. Using the CRIPSR-Cas9 technique to knockdown specific genes and then screen for HCA is a robust approach.
“After gene editing has been carried out, it is technically challenging to characterize the phenotypes of cells,” explains Dr. Saha. “We look at cells in an automated fashion across multiple days, after gene knockdown, with HCA. Our system employs a screening-compatible multiwell plate specially modified to sequester cells within a mini-well, or a ‘well within a well.’
“This forced localization allows us to do CRISPR-Cas9 edits in 400 localized colonies in one 96-well plate. Then we monitor the colonies every day for several days, gaining information on how each of these edited cells behave.”
“Earlier work took several months to find one clone harboring an edit on one allele,” continues Dr. Saha. “Our results show that we routinely get one-third of the clones with both alleles edited. Moreover, we can then use the automation to isolate a particular clone of interest.
“By streamlining this process, and integrating other technologies we have greatly increased the speed at which potential mechanisms of action can be elucidated in human cells. Application of this technology to determine the mechanism of action for drug candidates is promising. Given the high-throughput nature of the screen, this traditionally long process could be greatly sped up.”