27 May 2003 17:00 GMT

by Stephani Sutherland TARGETS

Researchers at the Whitehead Institute have screened over 23,000 compounds for anti-cancer properties using an engineered human tumor cell line. Unlike many high-throughput screens, this assay has the power to reveal which specific oncoprotein is targeted by each drug.

Although the screen is, at its core, a simple fluorescence-based viability assay, its strength lies in knowing the genetic identity of the cells, which enabled Brent Stockwell and colleagues to identify just nine compounds that were "synthetically lethal."

The initial screen

Stockwell and his team started with an initial screen of the compounds in two cell lines: normal BJ cells - primary human foreskin fibroblasts - and a genotypically identical cell line containing four engineered oncogenic elements: a genomic construct encoding the Simian Virus 40 large T (LT) and small T (ST) oncoproteins, an oncogenic allele of HRAS (RASv12), and the human catalytic subunit of the enzyme telomerase (hTERT). The result of the four added elements is a continually dividing human tumor cell line.

Human tumor cell lines

The genetically engineered human tumor cell lines were created by Stephen Lessnick and William Hahn of the Dana-Farber Cancer Institute. Hahn explains that the work with human cells grew out of his years of earlier work with rodent tumor cell lines, in which three of the elements were sufficient to transform the cells into cancer cell lines. "It was a matter of revisiting the issue of what it takes to make human cells tumorogenic," he said. The jump to a human cancer cell line came with the addition of the telomerase gene.

The need for telomerase to create tumorigenic human cells begs the question: what is an oncogene? Stockwell notes that there is no clear definition for an oncogene or oncoprotein; rather the label comes after the effects are seen, when a gene "cooperates biologically to make a cell cancerous." Although telomerase has "never been called an oncogene per se," he added, "it is necessary here to allow cells to replicate indefinitely."

Lethal compounds: erastin

The human cancer cell line enabled Stockwell to identify synthetically lethal compounds: that is, small molecules that were lethal to the tumor cells but not the normal BJ cells. Because the two cell lines were otherwise isogenic, he knew that one or more of the added, engineered oncoproteins must somehow be targeted by the effective compounds. Once the initial screen was complete, Stockwell then used other engineered cell lines in which each element was added back alone or in various combinations to pinpoint the specific drug targets. He also examined cells that separately contained either the LT or the ST oncoproteins, as well as the human papillomavirus type 16 (HPV) E6 and E7 oncoproteins. Interestingly, unlike many currently used chemotherapeutic agents, a newly examined compound called erastin seemed to induce a non-apoptotic cell death.

"Though many of the compounds killed,"said Stockwell, "only nine killed cancer cells selectively." Of all the compounds, he says, there was no way to tell a priori that those nine compounds would be selective. "Satisfyingly, though, they did fall into mechanistic classes," he said. Although several of the compounds are already in use as anti-cancer agents, the new work sheds light on their specific mechanism-of-action, and could help to target them to specific cancer cells expressing those targets.

Tailoring to tumor types

That is the hope of identifying synthetic lethal compounds: to find compounds that can be tailored to treat specific tumor types based on which oncoproteins they express. The idea of synthetic lethality is not entirely new; perhaps the most well known examples are Gleevec, which inhibits the breakpoint cluster region-abelsen kinase (BCR-ABL) oncoprotein, and Herceptin, which targets the HER2/NEU oncoprotein. Each of these drugs is now used against specific cancers: Philadelphia chromosome-positive chronic myelogenous leukemia and metastatic breast cancers, respectively.

Brian Druker, of Oregon Health and Science University (OHSU), pioneered the research that led to the development of Gleevec. He says that Stockwell is "right on target," in using the screen to find synthetic lethal compounds. "It's a way to let the tumor cells tell you what compounds will kill them," he said, adding that it makes a good complement to target-based drug design. The weakness of the technique, says Druker, is possibly that "they set the bar too low" in terms of selectivity for cancer cells. (The screen identified compounds with a fourfold increased potency over control cells.) Although an initially low selectivity might be necessary to catch a large pool of therapeutic candidates, a successful drug must eventually work at 10-100-fold selectivity, he says.

Future directions

Although all the identified synthetic lethal compounds are oncogene related, Stockwell envisions that the screening assay might one day be used to find the function of any given gene. As with the newly characterized compounds, one might find a molecule that targets a protein downstream of an overexpressed protein, or a molecule that inhibits a specific interaction between several proteins. The possible mechanisms-of-action of synthetic compounds are numerous, adds Stockwell. "I believe that, in principle, you could use the system to identify the function of any gene," he said.

And, added Druker, "it doesn't have to be cancer." The screen might be useful in identifying the drugs that are best suited to each patient's cancer, each with its own specific molecular profile, he says. "it might help us to use drugs more intelligently," he concluded.

This article was originally published in TARGETS.