It all goes back to the first half of the 20th century when a French dermatologist, Albert Sézary (1880-1956), gave his eponym to a rare malignancy now known as cutaneous T-cell lymphoma (CTCL). The Sézary syndrome (SS) is a weird leukemic variant of CTCL.

But that's not all. The more common form of this uncommon cancer, mycosis fungoides (MF), initially masquerades as ordinary, benign eczema. It's a skin-associated variant of CTCL.

"MF is a disease of T cells," observed molecular biologist Louise Showe at the Wistar Institute in Philadelphia. "MF, if detected and diagnosed quite early, is quite treatable. That stage of the disease is rather indolent in its course, so it can go on for 10 years or even decades. But Sézary syndrome, which is the leukemic form, is more aggressive than MF, and patients there have a three-year prognosis of survival.

"One of the problems with MF," Showe continued, "is that it looks very similar to benign skin disease, so it cannot be diagnosed readily. People with it frequently go for years to the doctor, who diagnoses their condition as simple eczema. Eventually a real dermatologist looks at the lesions and finds out that it's actually this MF malignancy. There is no good therapy for MF yet. Chemotherapy is not useful and most of the treatments being used now are biological response modifiers. These are cytokines that induce the immune system. Since MF is such a rare disease it's probably not been studied so well as many other cancers. But as it progresses so slowly, we researchers should be able to do something to help these people. Both Sézary versions total 1,500 to 2,000 new cases reported in the U.S. each year.

"What we wanted to do in the first instance," she noted, "was learn something about the Sézary syndrome, which is the more aggressive form - a little more difficult to study - and then employ the same strategy with the skin form."

Showe is senior author of a paper in the current Journal of Experimental Medicine (Rockefeller University Press), dated June 2, 2003. Its title: "Classification and prediction of survival in patients with the leukemic phase of cutaneous T cell lymphoma."

"The major thrust of this paper," Showe told BioWorld Today, "is that one can use microarrays - gene chips - to find characteristics of cancer cells. This can provide for better diagnosis, and a prognostic index for the cancer."

Complementing Inadequate Cancer Diagnostics

"Essentially, expression patterns of genes tell us: This patient is likely to progress rapidly; that patient not so likely.' So this patient should be treated differently than that patient, while on the surface both could look very similar," she said. "This work," she continued, "was undertaken with the goal of being able to develop molecular criteria to complement the existing diagnostic tools for cancer. These are not adequate in many cases to differentiate different classes of people with the same malignancy. That was basically our goal. So in a sense," she went on, "the Sézary syndrome is an advance approach, which could be extended to other cancers.

"To find out whether or not we could identify gene expression patterns that correlated with disease progression, we first looked at patient samples with high tumor burdens, identified genes associated with prognosis, then went back and looked at patients with low tumor burdens. What we found out is that there's definitely a gene expression signature that correlates with poor prognosis, but doesn't depend on tumor burden. So some of the patients with very low tumor burdens had very low prognoses, and were identified by the same genes that identified poor prognosis in patients with high tumor burdens.

"We generated our own microarrays for this work," she recounted. "Most people use DNA chips that are either on glass or the Affymetrix platforms. We chose a filter-based array, because of its sensitivity. What we have on these arrays are essentially spot DNA for cloned genes. So on a filter we put little dots of DNA, each corresponding to a different gene. Then we took our cancer cells and our normal cells and extracted the RNA - the product of those genes. The total library of RNA being expressed in those cells is then labeled with a radioactive tag, and incubated with the filter that has the spots of DNA for the genes on it.

"The amount of RNA that sticks to each spot is labeled, so the amount of radioactivity at each spot reflects how abundant that message is in that cell. Then we did that with cells from cancer patients and ones from normal controls, analyzed the data and asked: What's the difference between these two populations?'

"And we found 385 genes out of 6,600 were significantly different between those two populations. They were either overactive or underactive in blood samples taken from 45 cancer patients as compared to 20 normal controls. A panel of 10 genes was sufficient to identify a class of patients who will succumb to their disease within six months, regardless of their tumor burden. So then we asked, How few of these genes do we need to really do diagnosis?' For the high-tumor-burden patients we could go down to two or four genes and still diagnose them, using our particular analysis program. Then when we wanted to look at patients with lower tumor burdens, we needed more genes, eight or 20 more on the array."

Five-Center, Three-Year Clinical Trials Ongoing

"Six months ago," Showe noted, "five U.S. comprehensive cancer centers launched a concerted clinical trial of SS. They are the universities of Pennsylvania, Johns Hopkins, Northwestern and Michigan, plus M.D. Anderson Cancer Center in Houston. Each phase of therapy will be 13 weeks on varying regimens of interleukin-12 and IL-2. Then patients will go on for another 13 weeks. This three-year trial," Showe pointed out, "will focus primarily on MF - mycosis fungoides.

"So the general hope is after the trial what will the clinical application outcome be - whether patients do or don't respond to cytokine treatment. But what we're interested in is being able to analyze those clinical samples with arrays, so we can determine," Showe concluded, "why some cases respond and others do not."