Cancer immunotherapists and the media tout monoclonalantibodies as "magic bullets" for scoring bull's-eyes on tumorcells. But no one calls these antigenic sites "magic targets."Instead, oncologists refer to them as "tumor-specific antigens,"or to use immunological argot, "tumor-associated" antigens.
Some of these epitopic cell-surface proteins that lure and bindto monoclonals -- CEA (carcino-embryonic antigen), for instance-- are thickly concentrated on certain solid tumors, but showup thinly on healthy cells all over the body. Still, CEA is auseful, though not totally specific, target for imaging tumors, orlaunching conjugated antibodies to seek and destroy them.
In the June issue of Nature Genetics, British molecularbiologists describe an off-side play to tackle the impossible goalof an antigen that is genuinely unique to its solid-tumor origin.
"Cancer cells," argued the paper's lead author, Terry Rabbitts ofthe Medical Research Council, "arise because of somaticmutations, the most conspicuous of which are chromosomalabnormalities."
These aberrations often take the form of translocationsbetween gene sequences on two different chromosomes.
What makes a chromosome break? By and large, carcinogens --the same factors that cause cancer -- radiation, certainchemicals, tobacco smoke and inheritance. Rabbitts reasonedthat when two chromosomes break and switch DNA sequences,these may form a fusion gene encoding a cancer-causingprotein that, by its very origin, would be unique to the solidtumor involved in the translocation.
The carcinogenic role of these genetic rearrangements is well-known in leukemia, a "soft" tumor, where the long arms of the"Philadelphia chromosome" (No. 22) and chromosome 9 bothbreak, and exchange their loose segments.
This, in fact, is what Rabbitts found in two lines of humanmalignant liposarcoma cells. (Liposarcomas are relativelyinfrequent cancers of fat cells.)
In both cell cultures the British tracked down fusions betweenthe CHOP gene, which is a dominant negative transcriptionregulator on chromosome 12, and a hitherto unknown genesequence on chromosome 16, which they dubbed FUS (forfusion). Sequencing revealed that the fused gene encodes a462-amino-acid protein.
Rabbitts suggested that this chimeric protein could potentiallyserve as the first truly tumor-specific antigen. As such, heproposed, it could provide markers for diagnosis and prognosis,as well as specific targets for immunotherapy or antisenseoligonucleotide treatment strategies.
Cytogen Inc. of Princeton, N.J., is a major developer of cancerantigens for imaging and treatment. What does the company'spresident, immunologist/pathologist Thomas J. McKearn, thinkof the putative British super-specific tumor target?
"It's like unto the Holy Grail," McKearn intoned, the wordsimmunologists have used for at least two decades in their questfor the never-found perfect antigen. "The practical world ofreality in which we live has now taught us that althoughtumor-specific antigens might be highly desirable, one does notneed to have that level of selectivity," he told BioWorld. "Thoseof us who use monoclonal antibodies know that one can achievetargeting to less-exotic molecules. The tumor-associatedantigens are sometimes fully sufficient to meet that need."
McKearn added: "The British group may come up by theirgenetic manipulation with truly useful markers, that could beused to potentially great advantage. This may be a very novel,new and maybe useful way of doing it, so hooray for them, ifthey're going to pursue that."
Medical oncologist Martin Cheever does tumor immunologyresearch at the University of Washington School of Medicine inSeattle. For the past four years, he has been investigatingwhether oncogenic proteins can be used as targets of T celltherapy.
Cheever finds a molecular parallel between the British solid-tumor fusion gene and leukemia's Philadelphia chromosome."What they have shown is that their CHOP-FUS fusion geneencodes what is, in fact, a cancer-specific-protein. Basically, it'sa solid-tumor parallel to the Philadelphia chromosome's BCR-abl, which is also a fusion gene, and encodes a chimeric protein.In that respect, it's obviously very interesting as an example ofa protein expressed only in malignant cells."
BCR-abl, Cheever noted, is being used now to diagnoseleukemia by means of PCR (polymerase chain reaction), todetect the existence of the BCR-abl gene. It's used both forinitial diagnosis and to spot small amounts of residual disease,following therapy.
"In these circumstances," he asked rhetorically, "is BCR-ablchimeric protein a cancer-specific antigen?"
He pointed out that antibodies tend to recognize antigens thatreside on the cell-surface. T cells, on the other hand, bindstructures inside the cell, not on its surface.
Cheever's lab is looking at a parallel system, which is themutated ras oncogene. In this context, he finds an error in theBritish group's statement that "fusion proteins are probably theonly truly tumor-specific antigens."
Cheever demurred: "We've been looking at the rasprotooncogene, activated by point mutation, a single nucleotidesubstitution. The result is a normal protein with a singlesubstituted amino acid." This, he stated, is "another example ofa protein which is absolutely unique to the cancer cell.
"Now we see that the immune system, by recognizing shortpeptides, really may be most adept at recognizing singleamino-acid substitutions," he said.
Citing data not yet published, Cheever told BioWorld: "We cangenerate cytotoxic T cells by immunizing short peptides, whichcan lyse malignant cells that express this protein. Soessentially, killer T cells recognize single-amino-acid-substituted proteins. And these proteins have the potential toserve as tumor-specific-antigens."
-- David N. Leff Science Editor
(c) 1997 American Health Consultants. All rights reserved.