Molecule that Inflames Cancer May Also Dampen Spread of Disease

Some Therapies Against Akt Could Inadvertently Encourage Metastasis

A villainous, death-defying molecule that stokes the growth of many different cancers has surprising redeeming qualities, report two HMS research teams. The molecule, an overactive protein known as Akt, seems to rein in the trouble it creates by discouraging the growing tumors from invading surrounding tissues.

Split personality. In three-dimensional cultures, breast epithelial cells normally organize themselves into hollow spheres, mimicking milk glands (above left). The same cells engineered to overexpress a growth factor receptor implicated in breast cancer grow into larger, solid cancerous masses (above center). When Hanna Yoko Irie shut down Akt2 (above right), this prevented the aberrant behavior induced by the growth factor receptor, indicating that Akt2 is required for the cancerous effect. In contrast, blocking Akt1 (right) unleashed a more aggressive behavior involving the formation of spikes and cords that protruded into the surrounding gel, suggesting that Akt1 normally suppresses this invasive activity induced by hyperstimulation of the growth factor receptor.		Image courtesy of Hanna Yoko Irie

The findings may lead to more precisely targeted anticancer agents even as they raise concerns about a new generation of promising drugs in the pipeline that might inadvertently promote the spread of disease.

Mischief Maker
Akt is the central node of cellular machinery commonly hot-wired in cancer cells. Mutations in its partner proteins kick the PI3K/Akt signaling network into overdrive, which scientists believe both creates and nurtures a wide variety of cancers. The corrupted network has many ways to accelerate cell growth and proliferation, flout apoptosis, and enable chemotherapy resistance.

Deep pockets in its atomic structure make Akt appear vulnerable to small-molecule inhibitors. More than 20 companies are developing novel therapeutics or redeploying existing drugs to shut down the hyperactive molecule, according to a review article in the December Nature Reviews Drug Discovery. None of the compounds have progressed to clinical trials, but the approach has been reinforced by news that several recent celebrated successes in molecular therapy for cancer ultimately shut down this signaling pathway, including imatinib (Gleevec) for myelogenous leukemia, trastuzumab (Herceptin) for breast cancers with amplified HER-2/neu gene activity, and erlotinib (Tarceva) and gefitinib (Iressa) for cancers with EGFR mutations.

Now, independent studies from the laboratories of Alex Toker, HMS associate professor of pathology at Beth Israel Deaconess Medical Center, and Joan Brugge, HMS professor of cell biology, show that at least one version of Akt blocks the migration and invasion of human breast cancer cells in culture.

“Both papers are unexpected results,” said Lewis Cantley, HMS professor of medicine at BID and a member of the HMS Systems Biology Department, whose lab first discovered PI3K. Cantley and others showed how PI3K activates Akt and identified the many mechanisms by which Akt, in turn, inactivates important tumor suppressors, turns on cell growth, and sidesteps cellular suicide. Last year, his group showed that gefitinib kills lung cancer cells by blocking this pathway.

“These discoveries raise cautions about the consequences of administering a drug inhibitor of Akt,” Cantley said. “It may make cancer more metastatic. But if you want my bet, I would say there will likely still be a subset of cancers that can be dramatically improved by treatment with Akt inhibitors. We have to be cautious about teasing out that subset.”

Protective Akt
The two new studies reveal two different mechanisms Akt employs to discourage invasion. And the study from Brugge’s group further teases out functional differences between two of the three forms of Akt.

Photo by Graham Ramsay

One version of a protein known to stimulate cancer cell growth and proliferation also blocks migration and invasion, according to research by (clockwise from left) Merav Yoeli-Lerner, Joan Brugge, Hanna Yoko Irie, and Alex Toker.

In Toker’s lab, postdoctoral fellow Merav Yoeli-Lerner started testing cancer-related genes to learn which controlled motility of human breast cancer cells. Activated Akt genes she inserted into the cells blocked their movement through a nylon filter to the bottom of the lab dish more potently than any other gene she added. This was a surprise because other studies had shown that activated Akt in fibroblast cells promoted invasion of nearby cancer cells through the connective tissue.

“We spent a year trying to eliminate possible artifacts,” Toker said. “As we delved deeper, we realized this is how things are. For the subsequent year, we tried to understand the mechanism.”

In other experiments, endogenous Akt stimulated by insulin-like growth factor-1 (IGF-1) also blocked migration and invasion of cells. The researchers could restore the metastatic behavior of breast cancer cells by using small interfering RNAs to specifically eliminate Akt—much like many pharmaceutical companies want to do. In a different assay, activated Akt prevented cells from closing a wound efficiently, in contrast to inactive Akt–expressing cells.

To figure out the mechanism, the researchers started with the familiar. Three years ago, Toker and his colleagues discovered that the NFAT (nuclear factor of activated T cells) family of transcription factors promotes carcinoma invasion in human breast and colon carcinoma cell lines. And Yoeli-Lerner found that activated Akt sweeps away NFAT by recruiting a molecular janitor that tags the transcription factors with ubiquitin, marking them for disposal in the proteasome. No one yet knows which genes the transcription factors turn on to put the cells in motion, Toker said.

The test tube studies, in the Nov. 23 Molecular Cell, are consistent with a recent mouse study of tumorigenesis, Toker said. Canadian researchers found mice with activated Akt have more primary breast tumors, but fewer metastatic cancers. “The challenge now is to figure out if this is what is going on in human cancers,” he said. “This study may challenge other people to take into consideration an array of responses that cancer cells display in the PI3K signaling pathway.”

Distinctive Paths
Toker’s group worked exclusively with Akt1. There are two other forms of Akt made by genes on different chromosomes, but there has been no clear functional distinction among them. A related paper in the Dec. 19 Journal of Cell Biology from Brugge’s lab teased out distinct roles of Akt1 and Akt2 in migration and invasion.

Postdoctoral fellow Hanna Yoko Irie began her experiments seeking to understand how IGF-1 and other oncogenes induce cancerous changes in the human breast. Several years ago, a report from the Nurses Health Study by HMS researchers found that higher levels of IGF-1 were associated with a slightly increased risk of breast cancer in premenopausal women.

Irie worked with three-dimensional cell cultures. In these 3-D model systems, normal breast epithelial cells organize themselves into small hollow spheres similar in many ways to milk glands. Cells overexpressing the receptor for IGF-1 grew into large hyperproliferating masses that lacked the glandlike structure. Akt stood out in mutation studies to dissect the molecular events leading to the architectural distortions. Metabolic studies from other labs hinted at differences in the two most common Akt isoforms, so she grew the cancerlike breast cells and suppressed the Akt1 or Akt2 proteins with small interfering RNAs.

As with Yoeli-Lerner’s first experiments, Irie was surprised by the results. Without active Akt1, the IGF-hyperstimulated cells formed substantially different structures featuring spikelike protrusions and cordlike arrays of cells that invaded the surrounding environment. In contrast, dampening the expression of Akt2 prevented the hyperproliferative phenotype induced by IGF-1 hyperstimulation, and these cells grew into nearly normal spheres with hollow lumens (see images above).

“One important implication relates to the conversion from noninvasive to invasive tumors,” Brugge said. “Tumors have different stages—hyperplasia, carcinoma-in-situ, and invasive carcinoma. Yoko’s study suggests that the loss of a protein like Akt1 can dramatically convert a noninvasive behavior to invasive and represent a novel mechanism for tumor progression, as opposed to progression induced by a cytokine or dominant mutation that can activate invasive pathways.”

Both versions of Akt play a role in the oncogene-induced proliferation, Irie found. And Akt2 must be expressed for the invasive phenotype to flourish in the absence of Akt1. In a different mechanism of invasion, silenced Akt1 leads to more ERK, which belongs to another major signaling pathway and tends to be activated by growth factors and associated with the transition from solid to invasive cancer.

“Pharmaceutical companies are developing Akt inhibitors that will be useful in evaluating the effects of isoform-specific inhibition in various mouse tumor models,” Irie said. Based on the in vitro studies, Brugge said, it may be safer to inhibit both.

Further complicating matters, both research teams believe that Akt may play distinct roles in various cancers and that other signaling networks may interact with the pathway in complex ways. “Different cell types have different sets of genes, express different levels of isoforms, and are wired differently,” Brugge said. “We can’t make generalizations, even about other epithelial cancers.” The two groups are collaborating on further studies.

—Carol Cruzan Morton

top