Newsdesk

Nanocomputers to diagnose and treat cancer

Scientists have developed a molecular computer, made of DNA, with the potential to analyse and control gene expression. Although years away from clinical application, the device might eventually be used to diagnose some types of cancer and to react by releasing a drug molecule against cancer cells. The minuscule machine, so small that one trillion would fit into a drop of water, has been programmed to identify and analyse mRNA of diseaserelated genes associated with a model of small-cell lung cancer (SCLC) and prostate cancer, and to produce a single-stranded (ss) DNA molecule modelled on a drug with known anticancer activity. The molecular computer has input and output programmable modules as well as software, which are encoded in the four letters of the genetic code. The hardware (the part of the device that does not change) is an enzyme (Fok1) that cuts the strands of DNA in a
Roger Harris/Science Photo Library

Rights were not granted to include this image in electronic media. Please refer to the printed journal.
Small-scale detection aids cancer treatment.

specific way. The automation has two states: positive and negative. The computation starts in the positive state and if it ends in that state, then it is called a positive diagnosis; otherwise it is called a negative diagnosis. The authors of the study (Nature, published online April 28, 2004; DOI: 10.1038/nature02551) programmed the computer to detect the type of mRNA that would be present if specific genes associated with cancer were upregulated or downregulated. In the SCLC model used in the experiments,

the computer was instructed to administer the ssDNA molecule oblimersenpurported to be an antisense drug for treatment of SCLC provided that the genes ASCL1, GRIA2, INSM1 and PTTG1 were overexpressed. In the prostate cancer model, PPAP2B and GSTP1 needed to be underexpressed and PIM1 and HPN overexpressed. The ultimate idea is that these devices could diagnose cancer from within cells and dispense drugs as necessary, says lead author Ehud Shapiro, Weizmann Institute of Science, Rehovot, Israel. Our medical computer might one day be administered as a drug, and be distributed throughout the body by the bloodstream to detect disease markers autonomously in every cell. He adds: In this way, a single cancer cell could be detected and destroyed before the tumour develops.
Xavier Bosch

Six-gene profile predicts survival of lymphoma patients

By screening the activity of as few as six genes, clinicians have predicted the very different outcomes of patients with diffuse large B-cell lymphoma, the most common form of non-Hodgkin lymphoma. Their screen uses RT-PCR technology, making the test manageable for routine use. Keen to find genetic signatures that would predict a patients survival or their response to treatment, scientists have been toying with microarrays for the past 5 years. Their efforts resulted in long lists of candidate genes; however, no routine screen is available. It is also a concern that many clinical laboratories do not have the technology to screen large numbers of genes. Ronald Levy and colleagues at Stanford University (Stanford, CA, USA) therefore wanted to identify a small group of genes whose expression predicts survival in patients with diffuse large B-cell lymphoma, and which can be measured easily. They used RT-PCR to measure the activity of 36 genes in tumour samples from 66 patients with diffuse large B-cell lymphoma. These 36 genes had previously been identified as predictors of positive and negative prognosis. The researchers found that expression of only six genesLM02, BCL6, FN1, CCND2, SCYA3, and BCL2 could predict how long the patients survived (N Engl J Med 2004; 350: 182837). The scientists validated the six-gene model in 298 patients who had participated in two previous microarray studies and concluded that the model could identify patients who do not respond well to standard therapy and who might benefit from new treatments. According to Levy, the group will now check how well the six genes discriminate between good and bad responders in other groups of patients, especially those who are receiving treatments that were not available when the original patients were treated. If all goes well, the scientists want to make the screen widely available. Levy believes an RT-PCR-based test is more likely to be used in routine practice than a microarray assay because medical labs already use RT-PCR for other disease tests. However, Louis Staudt at the National Cancer Institute (Bethesda, MD, USA) emphasises that a diagnostic method that could use more genes to measure the biological features [of tumours] would have greater predictive accuracy. Staudt and colleagues previously published a 17-gene model that was associated with a higher degree of accuracy than the six-gene model (N Engl J Med 2002; 346: 193747). Screening more genes would require the use of microarray technology, but Staudt points out that a miniarray can accommodate more than 2000 genes and could be as cost effective as using RT-PCR. I guess the future will tell which of the particular diagnostic platforms will prove to be most valuable for patients, concludes Staudt.
Martina Habeck

336

THE LANCET Oncology Vol 5 June 2004

http://oncology.thelancet.com

For personal use. Only reproduce with permission from The Lancet.