16 July 2003 15:00 GMT

by Rabiya Tuma

Washington, DC - Everyone who uses them knows that endless culturing of cells in vitro leads to genetic and behavioral changes in a cell population. Now, one group has examined those changes systematically in melanoma cell lines, and the data may suggest why some melanoma vaccine treatments fail.

Several companies have vaccine therapies in late-stage clinical trial for the treatment of melanoma. In the simplest form, the protocol calls for harvesting cancer cells from patients and either purifying them or some protein component from them, and then immunizing the patient with a cell-derived vaccine. However, most protocols require that the initial cells isolated from patients be grown in culture so that there is sufficient material available for vaccine preparation.

That very process of growing out the cells may drive down the efficacy of the vaccine product, according to Brend Becker from the University of Regensburg in Germany, who was speaking at the annual meeting of the American Association for Cancer Research (AACR), here in Washington.

While he was working with a melanoma cell line, he noticed that the cells changed markedly after several growth passages. Seeing these changes and knowing that vaccine preparations require expansion of the cell population, he began to wonder just how different the primary cancer cells are from the cells used to make the vaccine.

To test this, Becker and his colleagues compared microarray gene expression patterns from primary cancer cells isolated from six melanoma patients and from cell lines derived from each of those samples. RNA from the cell lines was isolated at numerous points between the early passages and up to 34 passages.

Of the 3,200 genes detected on the microarrays, the team used 350 genes for cluster analysis. By the time the cells had gone through just five passages, 2% of the genes already showed significant differences between the primary cells and the cultured cells. This difference increased to 10% after 27 passages.

Additionally, as the cell cultures grew out, it was clear that subclones from the original population became dominant. Thus not all of the cancer cells would be equally represented in a vaccine product derived from such a cell line and some might not be represented at all. And worst yet, the subclones that grow so effectively in a culture dish may not be the most aggressive in the complex microenvironment of a tumor.

"This is hardly shocking," said Paul Meltzer who heads up the molecular genetics section at the National Human Genome Research Institute at the National Institutes of Health in Bethesda, Maryland. "But it is obviously important to see the melanoma cells change in culture."

Some melanoma vaccines work very well, says Becker. There are examples of patients who have lesions all over their body that respond completely so that the disease is no longer detectable, he adds. "But sometimes it works not at all or not very well," he warned.

These new data may explain why some vaccines fail, especially in light of the fact that some primary cancer cell cultures grow very poorly and require a much longer time in culture to produce the same volume of material as some that grow quickly. The slow growing cancers may undergo more changes prior to being used to prepare a vaccine and may make a less effective therapeutic, though that is only speculation at this point.

This is the first systematic analysis of gene expression changes in cancer cell cultures that Becker and his colleagues know of. Another research group has looked at genome amplification and deletion changes in melanoma cells during culturing. Becker plans to compare his gene expression data with those data set to see if the regions of change overlap.