Win by a nose

An interdisciplinary team of scientists at

is developing a molecular nose, a multiplexed sensor platform that can deconstruct a cell and distinguish its individual components.

It could help in the understanding of cellular networks and assist in the development of drugs.

According to Prof Walter Kolch, lead investigator on the project, only one of 30 drug compounds that go through clinical trials enters the commercial market.

'With the molecular nose, you could take a number of drugs that have known side effects and establish the signature patterns for the side effects. The nose will also identify the efficacy signature patterns in a drug, so you could look for others with similar patterns.

'You could then devise a quite efficient pre-screening system for drug development,' said Kolch.

Readouts from the nose will reveal signature patterns of cellular networks, which will allow the researchers to see how the components interact with each other.

'When you stimulate a cell with a growth effect or a hormone, the cell will do the exact same thing every time. We know the components of the cell, but we don't know how the components respond to stimulation,' said Kolch.

'The response is mediated by intercellular networks that we call signal construction networks. The big problem with the signal construction networks is not every signal shares its own network, or own pathway, but they are being used and shared by many signals.

'The molecular nose allows us to tie an input with an output simply by the response pattern it gives us.'

Prof Mark Girolami, an investigator working on the signal processing side of the project, said: 'This project is really an analogy to the electronic nose. In these you would train the nose with individual signals. When it takes a "sniff", it should then be able to deconvolve the complex signal into its constituent parts.

'It's the same here. We are going to look at the multiple readouts that measure protein concentrations in a cell and, from the readouts, we will be able to deconvolve those into previously understood and defined responses from proteins.'

The nose consists of sensor hardware — the microengineered nanoparticle used to probe the cells — and software that will implement the various inference and signal processing methods. This would enable it to monitor separate outputs of several hundred network components simultaneously in cell populations, or in a single cell, which the scientists said is not possible now.

'We will take the readouts and from them we will be able to infer from computerised models what is happening on a biological systems-wide level,' said Girolami.

'We don't want to just measure one protein in isolation but how it interacts with other proteins to produce the overall system level of response.'

The team plans to build three versions of the nose. The first will be developed using standard molecular biology techniques, which involves analysing a cell to assess its contents.

An array of sensors, with 1,000 individual probes, is first trained with individual stimuli, for example, by stressing the cell with heat. The cell's response is then observed and catalogued to establish a library of 1,000 response patterns. Artificial transcriptional reporters developed by the scientists, small pieces of DNA that bind to particular genes, will act as a reporter of the amount of transcriptional activity in that gene.

In this version, however, the cells need to be killed for the outputs to be analysed, while the second version will allow the researchers to monitor the readouts in living cells by using only up to 30 of the most informative sensors.

'Microengineering comes into the second version, where we put the components of the molecular nose on tiny silver beads that we can put into the cells and then monitor the readouts using Raman spectroscopy,' said Kolch.

'While the first version will simply use libraries of sensors, the second will need to be slimmed down because you cannot fit hundreds of sensors on a nanobead. Based on the initial results, we will use only those sensors that give us the best differentiation.'

The scientists plan to use a third version to analyse the cellular networks in a transgenic mouse.

As well as being able to monitor response patterns in living cells, the molecular nose will enable the researchers to monitor single cells.

'Most studies, so far, have been done on cell population, measuring the average of 10 million cells,' said Kolch. 'But research has shown that individual cells will respond differently when exposed to the same stimulus.'