Ciphergen Diagnostics (Fremont, California), through its research with Johns Hopkins University (Baltimore), has determined what it hopes to be a new route for detecting cancer at very early stages by focusing on host response proteins — proteins that previously have been found to instigate the inflammatory cascade in, for example, coronary disease, Alzheimer's disease or infection.

Ciphergen reported on the publication of its findings in the International Journal of Cancer, those findings suggesting that Surface Enhanced Laser Desorption/Ionization Time of Flight Mass Spectrometry (SELDI-TOf-MS)-based immunologic and chromatographic assays can be used to detect and quantitatively assay for modified forms of host response proteins. Such modification may aid in a much more precise classification of cancers.

Ciphergen said that while all cancer types generate a host response, this paper demonstrates that different cancers activate different subsets of host response proteins.

"As we started doing this work, and as we started getting our data, what we discovered that many of the markers that we found were not what you would call 'classical' tumor markers, but what you would call host response proteins," Eric Fung, vice president of medical and clinical affairs at Ciphergen, told Diagnostics & Imaging Week.

"We believe that the host response protein amplification cascade may be a viable alternative to searching for the elusive, highly specific, classical tumor marker," Fung said. "While classical tumor markers depend on adequate tumor bulk to be detected, HRPAC [host response protein amplification cascade] amplifies signals generated by even the smallest tumors. Because most other diseases also elicit a host response, we are studying whether this phenomenon can be used to detect other conditions such as Alzheimer's disease, infectious disease and cardiovascular disease in their earliest stages."

Ciphergen and the Johns Hopkins University School of Medicine published in Cancer Research in August 2004 a 503-patient study that revealed three biomarkers which could distinguish patients with early stage ovarian cancer from control individuals. The three biomarkers were apolipoprotein A1, a modified form of transthyretin and a fragment of ITIH4.

Ciphergen said the parent forms of those proteins are each host response proteins.

The new publication now demonstrates that "the relative abundance of the modified forms of those host response proteins is tumor-type specific. Multivariate analysis can be applied to those variable modifications as detected by SELDI-TOF-MS, permitting better classification of cancer."

Fung told D&IW that classical tumor markers have "reasonably good sensitivity for late-stage disease, where there is a lot of cancer, but much less sensitivity for early stage cancer, which is our ability to pick up cancers, particularly very small cancers. Because these markers are produced by the cancer itself, the bigger the tumor, the more of the marker that is expressed."

But there's another problem with classical tumor markers, he said — that they are also made by non-cancer cells.

For example, the PSA test for prostate cancer not only detects levels of the prostate-specific antigen when cancer is present, but also in prostatitis and benign prostatic hyperplasia — in other words, non-cancerous forms of the disease — thus offering "less than optimal specificity."

Fung said there have been two approaches to looking for more markers: looking for additional "classical" tumor markers, and combining "different types of markers." Ciphergen initially set out to pursue both strategies, but it found something a little different.

What the company "came to understand is that it's not as much a matter that the host response proteins are present where there's disease, [but] the fact that the body can modify these proteins," he said. This, he said, "is really the field that we work in quite extensively using pattern recognition in combination with proteomics."

The company's technology allows it to identify changes in post-translational modification to proteins, meaning the different ways that proteins are modified by enzymes that are made by a specific disease tissue. Those enzymes, he said, can be "truncations, where an enzyme cuts off either the front end or the back-end of the protein. There can be other types of modifications, where sugar is added, or a lipid molecule is added. And there are over 200 types of these post-translational modifications that have been described.

"This is really what I like to call the 'a-ha' moment, that is, it's not just the fact that we're combining markers, but we're actually looking at how proteins are modified by these diseases processes and combining our ability to measure these changes in post-translational modifications that's really going to help us get the types of diagnostic accuracy that we are looking for."

Fung said he thinks this finding is "really going to change the way that we diagnose disease," and it may play a role, not only in determining new diagnostic tests for such diseases as ovarian cancer — the specific target of the latest study — but also in developing therapeutics.

There's more work to be done.

But Gail Page, president of Ciphergen, suggested to D&IW that the company will work toward a multi-marker panel using host response proteins or the post-translational modifications.

She noted that there are 10 million women that are identified as being high-risk for ovarian cancer.

"If you think about the unmet clinical needs that are out there in the marketplace, I can't think of anything more important to work on than being able to diagnose disease at an early stage," Page said, "because obviously the benefits are tremendous [for] quality of life and life years saved."