One common and early sign of HIV infection is a cancer calledKaposi's sarcoma. Until the AIDS epidemic began, Kaposi'ssarcoma was fairly rare, afflicting mainly elderly men of Jewish orItalian ancestry with disfiguring and disabling, but non-lethal,bruises and welts on legs and feet.

Today, as many as 50 percent of HIV-positive gay and bisexual menbreak out with purplish Kaposi's sarcoma sores on their faces andmouths, which, said molecular biologist Yuan Chang "turn into veryaggressive malignant lesions, disseminating to organs throughout thebody, with massive bleeding."

To Chang, who teaches pathology at Columbia University, in NewYork, this clinical pattern had all the earmarks of an infectiousorigin. She surmised that life-style risk factors specific to AIDSvictims with Kaposi's sarcoma suggested a viral cause, most likely aherpesvirus.

Chang's paper in today's Science tends to confirm, though notprove, this hypothesis. Its title: "Identification of herpesvirus-likeDNA sequences in AIDS-associated Kaposi's sarcoma."

In Kaposi's sarcoma tissues taken at autopsy from men who diedwith AIDS, she and co-workers at Columbia, and Palo-Alto-basedDNAX Research Institute, found three DNA gene sequences highlyhomologous to several herpesviruses, including a possible newhuman type.

To pull out the putative infectious agent from these sequences, theteam employed representational difference analysis, a recentlydeveloped technique that combines polymerase chain reaction (PCR)amplification with subtraction hybridization.

"From the same Kaposi's sarcoma patient," Chang told BioWorldToday, we took diseased and healthy tissue samples. Any exogenoussequences in the former, the only difference between the two, wouldpresumably be the infectious agent. This we subtracted out, enrichedthe material by PCR, then isolated, cloned and sequenced it."

Fully 93 percent of their 27 Kaposi's sarcoma tissue samplesdisplayed this hitherto-unknown herpesvirus sequence. Morefamiliar types, such as Epstein-Barr virus, which causesmononucleosis and Burkitt's lymphoma, also had partially matchinggenes, as did herpesvirus saimiri, which gives fulminant lymphomato Old World monkeys. "Right now," Chang said, "the saimiri typeseems closest to our presumed human viral sequence for Kaposi'ssarcoma."

These viral types, like herpesvirus genitalis, are associated withsexually transmitted diseases.

The researchers also correctly predicted Kaposi's sarcoma in 11tissue samples provided to them blindly from the University ofPittsburgh.

Bolstering this circumstantial evidence is the finding (published inThe Lancet, 1990, Vol. 335: p. 125) several years ago that gay andbisexual men with AIDS are 20 times more likely than hemophiliacsto get Kaposi's sarcoma.

Since submitting their paper to Science two months ago, Chang et al.have performed another telltale test, this one on the Kaposi'ssarcoma limb lesion of an elderly, definitely non-AIDS, patient.Their result, submitted for publication "a couple of weeks ago,"Chang said, is under wraps. But she said, "it would really nail downthe causality, and we're still very optimistic that all types of Kaposi'ssarcoma have the same causality."

Also very recently, Chang's lab, now focussed on the Kaposi'ssarcoma project, has sequenced larger DNA stretches than theoriginal three genes, "and we still stand by our assertion," she said,"that it's a herpesvirus." The next step, she added, "is to isolate thevirus and see if it's oncogenic."

If research now ongoing confirms that the virus in Kaposi's sarcomais causative, it will add one more to the short list of malignanciesdefinitely traceable to a viral pathogen _ notably Burkitt'slymphoma, and liver, cervical and nasopharyngeal cancers.

Although the molecular marker the team discovered links herpes toKaposi's sarcoma, it's not a smoking gun. Columbia infectiousdisease epidemiologist Patrick Moore, who is senior author of theScience paper cautioned: "It is still possible that this agent is acommon latent virus in humans that preferentially colonizesKaposi's sarcoma lesions in immuno-compromised patients." n

-- David N. Leff Science Editor

(c) 1997 American Health Consultants. All rights reserved.