Structural biologists at Wageningen University and the Van Andel Institute in Grand Rapids reported in Nature on April 15 that they had persuaded a CRISPR enzyme to read chemistry as location. The enzyme is ThermoCas9, first isolated from a bacterium in a Wageningen compost heap. Unlike the textbook Cas9, ThermoCas9 is inhibited by a methyl group at a specific position in its target sequence. In the cell, that methyl group is the flag for "healthy." Remove the flag and the enzyme cuts. [1]

The practical consequence is that ThermoCas9 cuts tumor-cell DNA and leaves healthy cells alone. The team delivered the enzyme into MCF-7 breast cancer cells and MCF-10A non-tumorigenic breast epithelial cells, targeting sites known to be hypomethylated — under-flagged — in breast cancer patients. [2] The enzyme worked in the cancer cells and was inactive in the healthy controls. "ThermoCas9 uses methylation like an address to precisely target cancer cells while leaving healthy cells untouched," said Hong Li of the Van Andel Institute, a senior author. [3]

Two caveats ship with the paper. It does not show that cutting tumor DNA kills the tumor cell — that is the next round of work, funded by a January European Research Council grant to John van der Oost at Wageningen. And methylation is not binary: some tumor sites carry methyl flags, and some healthy sites do not, so the enzyme will miss some cancer cells and occasionally clip a healthy one. That is also true of chemotherapy. The paper's natural companion is the FDA CRISPR safety framework the paper covered yesterday, which will have to decide whether epigenetic targeting fits the agency's existing on-target and off-target bands.

ThermoCas9's 2017 characterization was a tool. Its methylation sensitivity was reported as a curiosity. Nine years later, the curiosity is the feature.

-- KENJI NAKAMURA, Tokyo