Conventional biosensors produce a signal that is in proportion to the concentration of target molecules, so at low concentrations they lose sensitivity and become susceptible to interference from other molecules. For disease biomarkers such as cancer antigens, the ability to differentiate confidently between a zero result and a trace result is critical.

The new sensor developed by Molly Stevens and colleagues at Imperial College London and the University of Vigo, Spain, can detect concentrations that are at least 10 times lower than the best existing ultrasensitive tests. "For many diseases, using current technology to look for early signs can be like finding the proverbial needle in a haystack," says Stevens. "Our new test can actually find that needle."

Nurturing nanostars

The team built its sensors out of tiny gold stars (or nanostars) measuring about 50 nm across. These structures host surface plasmons, which are coherent oscillations of the conduction electrons at the gold surface. Attached to their gold surfaces is the enzyme glucose oxidase (GOx), which acts as a biocatalyst to reduce silver ions in solution. At low concentrations of GOx, silver atoms are deposited so that a silver coating grows around each nanostar (see figure). This causes a shift towards higher frequencies (blueshift) of the nanostar's surface-plasmon resonance. At higher concentrations, the silver crystallizes at a faster rate and tends to nucleate separately in solution, and there is a less pronounced shift of the resonance.

The frequency of the resonance is measured by shining visible/near-infrared light on the nanostars and looking for the frequency where absorbance is greatest. As a result, measuring the frequency before and after GOx is introduced provides a very sensitive measure of the GOx concentration.

The next step is to use the nanosensor to measure the concentration of a biological molecule of interest – in this case a biomarker of prostate cancer called a prostate specific antibody (PSA). To do this, the researchers first coat the gold nanostars with an antibody that grabs the PSA out of the solution. Then, a second antibody – which is bound to the GOx – latches on to the PSA on the nanostar surfaces. Finally, the presence of the GOx initiates the silver-reduction step and shifts the surface-plasmon resonance, which is then measured.

Using this technique, the team could detect PSA at concentrations as low as 10–18 g/ml. This is a billion times more dilute than the limit of the enzyme-linked immunosorbent assay (ELISA) test that is widely used in hospitals.

"Our sensor generates the highest signal at the lowest concentration," says Stevens, "therefore the presence of the target molecule at ultra-low concentrations can be detected with the highest confidence."

"A really neat trick"

The new strategy has made a positive impression on David Duffy, who is head of research at Quanterix – a company developing single-molecule protein-detection technologies that is based in Boston, USA. "It seems like a really neat trick and a different way of approaching it," he says. "It definitely made me stop and think."

A sensitive test for PSA is important because after surgery for prostate cancer, PSA should no longer appear in the body – unless the cancer has spread or if the surgeon has left any affected tissue.

"[Ultra-low levels of PSA] have never been detectable previously, so current diagnostics miss it. All patients are basically in the same boat – they do not know if the surgery was successful in the long term for them," explains Duffy. "Definitely from the gold standard, which is ELISA, this [new method offers] a big jump in sensitivity."

Next step

David Fermín, a nanostructures and electrochemistry expert at the University of Bristol in the UK, agrees that the new results constitute "a very impressive piece of work".

The next step, he suggests, should be a close investigation of how well tiny concentrations of biomarkers can be picked out in the presence of potential interferences. "The researchers mention it briefly in the article, but I think [it will] be a very important aspect to investigate. I am sure that there is a clever chemistry that could be done to make it very specific, so this is an exciting development," he says.

So far, the researchers have only tested the PSA biomarker, but, says Stevens, "We are confident that the test can be adapted to identify many other diseases at an early stage." Of particular interest is p24, a protein related to HIV infection the detection of which could help in diagnosing the infection in its infancy.

The researchers published their work in Nature Materials.

•  This story first appeared on our sister website, physicsworld.com