Automating wide-scale attacks on disease: In a person with spinal muscular atrophy, fibroblasts could suffer from a deficiency in a protein called SMN (labeled green), and scientists at Merck use 1,538-well plates  

High Content Is King

A yellow robot arm moves with a combination of inhuman speed and delicate grace, loading a microscope with a plate split up into 1,536 wells, each filled with ovary cells from Chinese hamsters. Here the classic image of scientists peering into microscopes to laboriously scan slide after slide is gone. In an instant, the automated microscope at Merck Research Laboratories in North Wales, Penn., digitally images all of the specimens simultaneously.

Welcome to the brave new world of high-content screening (HCS), which is dramatically scaling up the number of experiments on cells that researchers can run, just as DNA microarrays helped revolutionize molecular biology. Using robots, scientists can now grow most any kind of tissue in a variety of conditions, dose those cells with any of the millions of molecules that pharmaceutical and biotechnology companies keep in their libraries, image the effects these molecules have on cell function using microscopes and high-resolution digital cameras and run computer algorithms to mine the resulting avalanche of data. Operating 24 hours a day, just one HCS system at Merck can test and image 4.3 million wells per week.

"We're no longer just sitting there in front of our microscopes preparing one slide at a time," says Jeremy Caldwell, head of Merck's automated biotechnology unit. "Automation frees us up to think much more deeply about the design of experiments, how we might find what's important."

While previous high-throughput methods focused on a limited number of chemical markers to figure out what effect specific molecules have on cells, high-content screening pays attention to more: the movement of proteins in cells; how cells, nuclei or organelles change size and shape; how cell-membrane permeability varies; and other characteristics. "Instead of what might be under a lamppost, you're illuminating what is going on in the rest of the entire city," Caldwell says.

Cost-Saving Content

In the past five years, HCS has increasingly found its way into drug discovery, to see which potential therapies are worth developing for clinical trials. "One can see if potential anti-angiogenesis drugs actually do shrink blood vessels feeding tumors, or see how many neurites a potential neuronal-regeneration drug causes neurons to grow," explains Mark Collins, senior marketing manager for Thermo Fisher Scientific’s cellular imaging and analysis group in Pittsburgh, Penn. The first high-content screening device was introduced a decade ago from cellular-imaging firm Cellomics, now part of Thermo Fisher Scientific.

On the flip side, HCS is also employed in toxicology screens to weed out molecules not worth developing. "Instead of putting drugs in animals for tests or getting to humans and getting a situation like Vioxx after, perhaps, investing hundreds of millions of dollars, there is an awful lot of interest in running in vitro toxicology screens on cell-based models of liver, heart, and every other organ of interest for signs of toxicity," Collins says. "By looking at the entire cell and measuring multiple events, one can see the whole picture and hopefully not miss key details—think of it like a ‘cellular autopsy.’"

Basic researchers also use HCS. For instance, Merck is learning what genes do by using small interfering RNA molecules to knock out specific genes and seeing what happens. "In the old days of the genomics era, the idea was that DNA microarrays could provide everything you needed to find out what genes should be targeted to treat a disease, but they just tell you which genes are activated and which are not, and to find out which genes are causal in disease, you really have to probe cellular function," Caldwell says. "High-content screening can help us follow what is happening in the cell over space and time to dissect how genes work in relationship to each other."

Emerging research proves that HCS can be effective. For example Collins notes that researchers at the Hospital for Sick Children in Toronto used HCS to discover new drugs that could reverse the effects of cystic fibrosis, "including the well-known clinically approved anti-cancer drug Velcade." He adds that another study suggested that a typical pharmaceutical company with about 10 drug discovery projects per year could save $350 million annually by using HCS to help determine which potential therapies are toxic early in the drug-discovery process.

Even More Content

Analyzing the mountains of data from faster and more complete technologies creates today's HCS bottleneck. "It's a challenge to manage the terabytes of data that a typical high content–screening experiment can generate or analyze 200 multivariate parameters per cell," Collins says. Another challenge that Collins says "keeps me up at night" is finding a way to make it easy enough to be accessible to a broad spectrum of researchers. "How can we make sophisticated technology that a scientist doesn't have to devote their life learning how to use without dumbing it down to the point where it's no longer valuable?" he asks.

If scientists can put HCS's gigantic amounts of data to use and keep the techniques easy enough to run, a great deal of potential might lie in wait. For instance, although an HCS system looks at an entire cell, it typically only follows a few kinds of molecules at once. Instead Merck hopes to soon monitor the behavior of up to dozens of kinds of molecules in cells simultaneously using a novel "poly-target robot system" from the Genomics Institute of the Novartis Research Foundation. This innovation could help researchers begin to really understand organisms as complex systems of interacting processes, the goal of the field of systems biology as it was originally envisioned. "All this rich information could prove invaluable not just for drug discovery but for basic research," Caldwell says.

Perhaps the ultimate goal of high-content screening is to gather enough data on cells "to build myself a virtual cell," Collins says. "I could then—in silico—answer questions about biology." Walking through a lab festooned with the electronic innards of automated biotechnology systems, Caldwell agrees. "I would love to do that," he says. "It's something we could potentially begin to see in our lifetimes with the aid of high-content screening."