Print

Motivation

Any disease is implicitly defined as a deviation from disease-free (baseline) states, yet our knowledge of the disease-free state is shockingly limited.

One of the fundamental performance requirements of a diagnostic test is that it must not give a positive signal in subjects without the disease of interest. While searches for molecular markers of cancer typically focus exclusively on comparing cancer cells to the corresponding normal cells (e.g. breast tumor versus normal breast cells), this comparison is far from adequate to ensure biomarker specificity. A comprehensive molecular portrait of the normal human and a map of the range of normal human variation are essential foundations for development of disease diagnostics.

By building a database of gene and protein expression in the blood and other tissues of cancer-free men and women, we can rationally select biomarker candidates that are likely to be specific to a cancer. Because biomarker validation is such a costly endeavor, such up front vetting of candidate biomarkers provides a huge savings in time, cost, and resources. This core resource offers extremely high leverage, as it enables efficient and effective prioritization and evaluation of candidate biomarkers across all of our cancer research programs.

A critical requirement for a test that accurately diagnoses cancer is that it is not found in other confounding non-disease conditions, ranging from stress to benign disease. Thus, a significant component of the Baseline Program is ensuring availability of control specimens to assess the specificity of leading candidate biomarkers across the wide range of non-diseased conditions that the marker must overcome.

Research

Baseline Program seeks to characterize the 'normal' human and 'normal' human variation, as a foundation for rational biomarker discovery and prioritization.

In order for a biomarker to be adequately specific, it must not only be absent from the normal counterpart cells but also from any other organs or tissues in the body that may contribute to background signal. There are several different types of information that can help us to determine this element of specificity. These data will allow us to determine which candidate biomarkers should be rapidly eliminated and which should be aggressively pursued.

Normal Blood Studies

• Presence/Absence - Candidate biomarkers that are frequently found in the blood of cancer-free controls are unlikely to be useful as markers of cancer. While traditional assays that precisely measure the amount of individual proteins are rarely available for novel biomarkers, current proteomic methods can be used to roughly estimate the abundance of large numbers of proteins at once. We thus supported a project to aggregate data from large numbers of proteomic studies in which blood from cancer-free individuals was analyzed as well as a project to directly profile the blood of a selected set of healthy individuals. Candidate biomarkers that are frequently found in the blood of cancer-free controls can thus be identified and de-prioritized for assay development and validation.
• Variation - Biomarkers whose levels vary to the same extent within an individual over time as they vary between individuals are unlikely to provide a robust test for cancer. Thus, we are supporting a project to characterize this variation component for large numbers of proteins as an additional filter for prioritizing novel candidate biomarkers.
• Cancer-free blood specimens - Validation of biomarker specificity requires measuring the biomarker in patients with cancer as well as in large numbers of cancer-free controls. The more rare the disease, the greater the number of controls specimens we must evaluate; in the case of ovarian and pancreatic cancer, several thousand specimens must be tested to ensure that the rate of false positives is not too high. It is a major undertaking to not only recruit patients but to also process, store, annotate, and manage such large specimen banks. To address this need, we supported a project to collect blood from healthy women in British Columbia, which served both as a pilot for a large population-level, prospective blood collection effort and an immediate source of control specimens. We are also actively investigating other sources of normal and control specimens for use across all Canary cancer research programs.

Normal Tissue Studies

• Tissue Expression - Biomarkers that are uniquely expressed by cancer tissues have the greatest potential utility for both blood and imaging tests. We are supporting a project that will use DNA microarrays to survey the expression patterns of thousands of different genes in sixty different human organs and tissues. The resulting database will be used to rapidly assess the tissue specificity and corresponding likelihood of candidate biomarker utility for uniquely identifying cancer.
• Cellular Localization - The specific location of a protein within the cell is a valuable indicator of whether it is likely to be useful as a secreted biomarker (ideal for blood tests) or a cell surface biomarker (ideal for imaging tests). Tissue microarrays provide a tool for rapidly determining the cellular localization of a given biomarker in many different tissues at once. We are supporting the construction of a tissue microarray resource that can be used to rapidly characterize candidate biomarkers across a wide range of tissues.
• Incidental cancers - In order to develop tests that specifically detect lethal cancers, people harboring these target cancers must be distinguished from patients with non-lethal tumors. In some cancers, such as prostate cancer, non-lethal cancers are extremely common and are therefore a major confounding factor in developing a test that specifically detects lethal cancers. For other cancers, the range and frequency of non-lethal cancers is not well understood. We are thus supporting a pilot project to investigate the feasibility of using autopsy studies to characterize the range and frequency of non-lethal, occult cancers in specific organs of interest.

View Canary Funded Projects