One thing cells have to do to become cancerous is to overturn the normal controls on their growth. As part of this process, the tight controls on things like copying and repairing DNA are starting to break down. As a result, tumors often contain chromosomal rearrangements, which are places where genes are cut and stitched back together in a way they shouldn’t be.
In some cases, the breaks bring two genes together in a way that causes so-called “driver mutations,” forming a fusion protein that pushes cells further down the road to malignancy. For some types of cancer, almost every tumor contains one of these chromosome breaks, making the fusion genes a hallmark of that cancer. A group of researchers at the University of Pittsburgh School of Medicine just took advantage of this specificity by targeting the fusion genes to attack and destroy cancer cells.
The work is based on the CRISPR-Cas system, which is used by bacteria to recognize and chop up foreign DNA (such as that of a virus). Since then, we’ve learned how to target any DNA sequence, making it a formidable tool for gene editing. Here, the researchers used a CRISPR-Cas system to make a nick — a single-stranded cut in the double helix — in the tumor DNA, right at the point where two genes have fused.
This break should activate the cell’s repair pathway to repair that single-stranded nick, so the researchers hijacked this system by supplying it with DNA to use in the repair. With the DNA supplied by the researchers, the repair system placed an enzyme at the site of the gene fusion. The enzyme they chose takes a harmless drug precursor and metabolizes it into its active form. The enzymes therefore target the tumor cell because only it has this DNA rearrangement; then, when the researchers apply the drug, only tumor cells are affected by it. Normal cells are not.
First, the researchers tested the approach by generating cells with the gene rearrangement they wanted; they took a fusion gene that recurs in prostate cancer and put it into a cell line that normally lacks it. Their approach worked: only cells that received all components (fusion gene, CRISPR-Cas, new enzyme and drug) were killed. Cells without the fusion gene could not express the enzyme and so were not affected by the drug.
Let’s test on mice
Next, the researchers took the engineered cells and transplanted them into immunodeficient mice, where the cells grew into tumors. (This is a fairly standard way of generating tumors for drug testing.) If the mice were given all components of the CRISPR/drug system, the tumors shrank by 30 percent. Control mice, whose tumors lacked the fusion gene, showed a 50 percent metastasis rate — even when given the CRISPR/drug components — and died.
Next, the researchers tried it in a human hepatocellular carcinoma cell line that had a different chromosomal rearrangement. This means they had to modify the CRISPR-Cas system to handle this rearrangement. But they saw the same effect: Up to 27 percent of cells expressed the enzyme (this percentage is typical of gene editing using CRISPR-Cas), and up to 27 percent of cells were killed by the drug.
Finally, they put these cells in mice and saw the same thing: mice that also received all the other components had their tumors shrink, but mice that didn’t die.
Current cancer therapies often rely on interrupting the signaling pathways that drive tumor growth, almost begging the tumor to develop resistance. Since fusion gene breakpoints may not be cancer-causing agents, using them to attack tumor cells might not elicit resistance in the same way. And even if using breakpoints does, another piece of DNA damage in the tumors can just be attacked, as long as one can be identified. Because the DNA damage is so highly specific, targeting it with drugs does not generate the nasty side effects that occur when drugs interact with normal cells rather than just cancer cells.
Since chromosomal breakpoints can vary from person to person – and even from different tumors in the same person – this would be a highly individualized therapy. But if it’s technically feasible, it could be promising.
Nature Biotechnology2017. DOI: 10.1038/nbt.3843 (About DOIs).