Tumour-Specific Markers: The Holy Grail of Cancer Diagnostics

---

By Amnon Gonenne

Medical literature abounds with examples of the benefits of early cancer detection. Cure rates are always dramatically higher before the tumour has spread and while surgery is still an option. For example, in cervical cancer, detection at the very earliest stages of the disease is associated with a 99 per cent five-year survival rate. Similarly encouraging statistics may be found for cancers of the breast, ovaries, colon, skin and other sites.

Cancer detected through physical examination or medical imaging is usually too advanced for hope of a cure, which has led to an explosion of research into molecular diagnostics. Immunoassays, the leading category of such tests, are exquisitely sensitive and have the potential to diagnose cancer at the stage of just a few cells.

Unfortunately commercial molecular diagnostics for cancer suffer from a lack of specificity, sensitivity and overall robustness.

Three biomarkers (or so-called “cancer markers”) that immediately come to mind are prostate-specific antigen (PSA) for prostate cancer, CA-125 for ovarian malignancies and Her2-neu in breast cancer.

PSA is found in normal prostate tissue and at elevated levels when the gland is inflamed. Since PSA is not specific to prostate malignancy, diagnosis of prostate cancer based on rising PSA levels results in a high rate of false positives – up to 75 per cent according to some studies. False positives are highly undesirable because they trigger costly, invasive medical interventions that divert healthcare resources that could be better spent.

Biopsies of 100 suspected prostate cancer patients with PSA readings of 3 ng/ml or higher will return only 25 confirmed cases.

In addition, 40 per cent of PSA-negative readings are false. PSA was approved by the U.S. Food and Drug Administration at a time when prostate cancer diagnostics were essentially non-existent. Were the PSA test to come up for regulatory review today it is unlikely it would be approved.

One finds an analogous situation with Cancer Antigen-125 (CA-125) which is not a true tumour marker but a product of female genital mucosa and other endothelial tissues. CA-125 is detected in many ovarian cancers, as well as in tumours of the breast, colon, and other organs. CA-125 was approved by the FDA for monitoring recurrent disease in ovarian cancer patients, is ineffective in detecting early-stage ovarian cancer, and its blood levels are a poor predictor of patient outcome. First-line chemotherapy for ovarian cancer often knocks CA-125 readings back to baseline, but the majority of patients relapse within a year with aggressive disease.

Similarly Her2-neu, a natural biomarker over-expressed in approximately 25 per cent of breast cancer patients, is unsuitable as a diagnostic test. Because the gene is also expressed in many somatic cells, it is not surprising that Herceptin, the cancer drug recommended for women with Her2-neu-positive breast cancer, has been associated with cardiac and other toxicities.

Clearly a useful cancer test will possess high diagnostic specificity and sensitivity, be expressed exclusively by tumours of one type, and detect cancer early to provide a reasonable hope of cure. The ability to predict both outcomes and response to therapy could be considered added benefits.

Cancer-Specific Antigens
The first true antigenic surface marker specific to cancer cells was EGFRvIII, a mutation in the epidermal growth factor receptor (EGF) found in a variety of different cancers but not in normal cells.

This true cancer marker is uniquely different from the native EGFR that has served as a target for highly touted non-specific cancer therapy (e.g. the monoclonal antibody Avastin), in those cancers in which EGFR is over-expressed.

Discovered about ten years ago, EGFRvIII was originally dismissed by scientists but has shown usefulness as a possible therapeutic target. But while this receptor is cancer-specific, it is not organ-specific as it appears in several cancer types. Another potentially useful cancer-specific marker is TMPRSS2:ERG, the fusion product of the prostate-specific marker TMPRSS2 and the transcription factor ERG. TMPRSS2:ERG levels accurately predict the likelihood that prostate cancer patients will relapse after surgery. However, just half of prostate cancer patients express TMPRSS2:ERG, which does not bode well for using this marker as a screening test or general diagnostic.

So despite an avalanche of knowledge in molecular biology in recent years, not a single cancer-specific cell surface antigen has yet been discovered. We know genes mutate inside cancer cells, but the molecular manifestation of these events as surface antigens has thus far eluded detection and characterization. This is not surprising as cancer cells have evolved numerous “escape” mechanisms for eluding not just pharmacologic intervention, but the body’s own defense mechanisms. Masking of tumour-specific antigens (TSAs) by a cell’s native surface antigens, which prevents direct interaction with TSAs, is just one such mechanism. Another is the relatively poor immunogenicity of TSAs compared with neighbouring antigens.

The standard approach to discovering new proteins involves explicit knowledge of their existence and the ability to isolate them. Barring a first-in-kind discovery, it is quite unlikely that TSAs will be characterized sufficiently given the current state of technology.

Scientists have developed a technique for generating antibodies against TSAs without a priori knowledge of their location or structures. This technology is then used to generate monoclonal antibodies (MAbs) against melanoma, colorectal, prostate and ovarian cancers in animal models.

The key to the technique is hybridoma generation using methods similar to phage display. We generate very large hybridoma libraries containing MAbs to every available antigen on the surface of the cancer cell, including rare and poorly-immunogenic TSAs.

The MAbs are then back-reacted with a panel of cells that includes cancer cells from which the antibodies were generated. This screening process is repeated many times against different unrelated cancer cells and normal cells. Antibodies that bind only a single type of cancer cells must be specific for TSAs, and are termed “universal” to that cancer.

Discovering TSAs and antibodies that bind to them would be tremendously difficult using conventional hybridoma techniques, which create libraries of relatively modest size in four to six months. This technique generates tens of thousands of hybridomas in four to six weeks. The change is not merely quantitative but qualitative as well. By preserving the antigen’s three-dimensional structure, our methodology assures that if TSAs are present, they will reach antibody-producing immune system cells and create antibodies that bind strongly to their targets.

Mabcure has validated its own MAb-generating technology through two proof-of-concept studies. In the first test an anti-melanoma MAb generated from a library of more than 6,000 hybridomas correctly identified all samples from ocular melanoma. The second study used an anti-ovarian cancer MAb to identify disease from the blood of 13 of 13 patients with advanced ovarian cancer. Six of these patients, who had recently completed their first course of chemotherapy, were deemed to be in full remission by CA-125 testing, but our test indicated they still had active disease.

This fact was, unfortunately, borne out by relapses in these patients. The ovarian cancer MAb did not cross-react with serum from healthy patients.

Although our cancer-specific antigens will require development work before approval and commercialization, cancer-specific antigens provide the scientific basis for designing numerous products for cancer diagnosis, imaging, and therapy. They represent a sea-change in the molecular diagnosis and treatment of cancer. If produced at reasonable cost such antibodies will provide highly improved sensitivity and specificity for cancer screening and diagnosis, and monitoring of therapy as well. Similarly tumor-specific MAbs, when coupled with the appropriate radionuclide, might one day pinpoint micrometastases for diagnostic purposes and/or direct radiotherapy.

Finally, studies involving the MAbs and their cancer cell targets will lead to the discovery and characterization of TSAs, which themselves might be the targets of pharmacologic intervention.