Nanotech molds promise cheap, fast way to detect cancers

A simple, fast and inexpensive way to detect cancers at an early stage from protein markers in the blood has been developed by a team led by Kobe University Professor Toshifumi Takeuchi.

The team has partnered with Tokyo-based medical equipment company System Instruments in hopes of commercializing the system in as little as five years.

Rather than using antibodies, which are expensive to make, the new detection method tests for the presence of cancer-specific proteins in blood samples using a mold made from an inexpensive polymer material, fabricated in a technique akin to molecular imprinting.

In this technique, the polymer material is packed around the target protein to create a nanometer-sized mold. For detection purposes, a fluorescent material and another substance that acts as a damper on this fluorescent material are set in the bottom of the mold.

The mold interacts in lock-and-key fashion with the target protein, and the fluorescent material becomes activated to emit light only when the target protein fits into the opening.

Using this setup, a blood test for the target cancer marker takes minutes instead of the hours required for antibody-based detection, and Takeuchi expects the cost of the device can be kept down to just a few dollars, which is about a hundredth the cost of antibody-based devices.

Takeuchi collaborated with Kwansei Gakuin University Professor Keiko Tawa to enhance the fluorescence and to develop a chip with an array of countless molds to detect the target protein. When light is shined on the chip, the fluorescent material will emit light that is five to 10 times brighter if the target cancer marker is present in the sample.

In tests on human blood samples to detect AFP protein, which is a liver cancer marker, the method proved to be quick, easy and just as sensitive as the standard antibody-based test.

The mold-making technique could be used to fashion tests for a wide variety of cancer markers, and also to design allergy tests.

Give it a share: