Scientific Overview: Cancer Immunodiagnostics

Research Interests

Members of the Haab laboratory identify protein and carbohydrate abnormalities in the blood of cancer patients and investigate the significance and potential clinical usefulness of those abnormalities. We develop novel experimental methods to facilitate this work, and we collaborate with both clinicians and basic scientists to pursue research on pancreatic and prostate cancers.

Low-volume, high-throughput antibody and protein arrays

We have developed the ability to probe multiple proteins or carbohydrate structures using low sample volumes, which provides a powerful tool for identifying and measuring protein and carbohydrate abnormalities in cancer. Antibody and protein arrays immobilized on the surface of a microscope slide are the key to such a capability. A biological sample such as blood serum can be incubated on an array to investigate interactions between the immobilized molecules and the proteins or antibodies in the sample. Those interactions can be probed to obtain information such as protein abundance, glycosylation level, or protein-protein interaction level.

The routine use of these tools was made possible by the development of a practical method for processing multiple arrays on a microscope slide (Fig. 1). A stamp imprints a wax pattern onto the surface of a slide, creating hydrophobic partitions that segregate various samples. Distinct stamp designs can be used to form differing sizes and numbers of partitions. A design that imprints 48 arrays on one slide requires only 6 μl of sample per array, with each array composed of 144 distinct spots of immobilized molecules. Such a design enables the efficient processing of many samples or testing of many conditions in parallel, as demonstrated in the projects described below. The device for creating these slides is commercially available from The Gel Company, San Francisco.

Figure 1. High-throughput sample processing using a novel slide partitioning method. A) Wax is imprinted onto a microscope slide to form borders around multiple arrays. Wax is melted by the hotplate under the bath, and a slide is inserted upside-down into the holder. Bringing the lever forward raises a stamp out of the wax bath to touch the slide, imprinting the design onto the slide. Two stamps are shown in front of the machine. B) Loading samples onto a slide containing 48 arrays. The arrays are spaced by 4.5 mm, which is compatible with the 9 mm spacing of standard multichannel pipettes. C) Samples loaded onto slides containing 12 (top), 48 (middle), and 192 (bottom) arrays (96 samples loaded).

Glycans in pancreatic cancer

One of the major interests of the lab is characterizing and studying the changes in carbohydrate structures (glycans) on particular proteins from pancreatic cancer patients. A novel technique developed in our laboratory enables the measurement of specific glycans on multiple proteins in biological samples (Fig. 2A, B). We use lectins—proteins that bind specific glycan structures—as well as glycan-binding antibodies to probe the levels of particular glycans on the proteins captured on the antibody arrays. Several types of lectins, each with its own carbohydrate binding specificity, can be used to identify the carbohydrate structures associated with each protein. We can analyze many different patient samples or cell culture conditions, looking at associations between glycan levels and disease states or at the effects of certain perturbations on glycan structures. This method is in development for commercial use by GenTel Biosciences (Madison, WI).

Mucins are long-chain, heavily glycosylated proteins on epithelial cell surfaces that have roles in cell protection, interaction with the extracellular space, and regulation of extracellular signaling. Screening studies in collaboration with Randall Brand and Diane Simeone have revealed a variety of glycan alterations on mucin molecules from pancreatic cancer patients (a representative example is shown in Fig. 2C). Altered carbohydrates on mucins can affect critical processes in cancer such as cell migration or extracellular signaling to the immune system. We are characterizing the glycan structural variation on mucins secreted from cancer cells and other cells, and we are using cell culture systems to study the origins and effects of those variations. We are pursuing hypotheses about the effects of extracellular stress from an inflammatory tumor environment on mucin carbohydrate structures and the resulting interactions of those structures with inflammatory proteins and host cells.

Figure 2. Complementary antibody array formats for protein and glycan detection. A) Sandwich assay with fluorescence detection to measure protein abundance. B) Antibody-lectin assay. The biotinylated lectin binds to glycans on the proteins captured by the immobilized antibodies. The antibodies are first chemically derivatized to prevent lectin binding to the glycans of the immobilized capture antibodies. C) Detecting protein and glycan variation in cancer and control sera. Sandwich detection of the MUC1 and CEA proteins showed similar levels in serum samples from a cancer patient and a control subject (left images). The anti-CA19-9 antibody, which targets a glycan structure, detected a significant glycan increase on MUC1 and CEA in the cancer serum (right images).

Cancer biomarkers

Improved methods of detecting and diagnosing cancer could significantly improve outcomes for many patients. We are seeking to identify and validate protein biomarkers that could form the basis of clinical cancer diagnostics. The antibody-based assays that we are using are valuable for this work because they are very reproducible, inexpensive, and high-throughput. In addition, the use of miniaturized arrays of antibodies allows us to efficiently test many antibodies and samples and to rapidly develop new assays. We are applying these capabilities in novel approaches to biomarker discovery and validation.

Mouse models of cancer may provide a good resource for biomarker discovery because the genetic and experimental variation between samples can be closely controlled, thus making the identification of abnormal protein levels easier than with human clinical specimens. Mass spectrometry studies performed by other members of an NCI-sponsored consortium have identified candidate biomarkers in mouse models of ovarian and pancreatic carcinomas. Using newly generated antibodies that target those proteins, we are developing assays to determine the levels of these candidate biomarkers in the mouse models and to assess their diagnostic value for human cancer. Low-volume methods are crucial for these studies because only a small sample is available from each mouse. These studies could establish a new paradigm for biomarker discovery and validation.

Longitudinal biomarkers

An NCI-sponsored project in our laboratory focuses on the hypothesis that the diagnostic performance of particular biomarkers can be improved by using measurements collected on multiple occasions (longitudinal measurements) rather than at just a single point in time. By looking at changes over time, it may be possible to more accurately distinguish abnormal levels in a given individual, since that person’s normal level could be used as a reference point. In a collaboration with Robert Vessella and William Catalona, we are investigating this question for the detection of prostate cancer recurrence. By using various formats of antibody arrays, we can explore different data types and multiple proteins, which we hope will establish the extent of diagnostic improvement using longitudinal information. Another collaborator, Ziding Feng, is developing the statistical methods for analyzing the data, which may have value for other applications of this approach.

Tumor-reactive antibodies

We and others have investigated measurements of tumor-reactive antibodies as biomarkers. Certain tumor proteins elicit an antibody-based immune response in a high percentage of cancer patients. In collaboration with Samir Hanash, Gilbert Omenn, and others, we have further developed the experimental methods for identifying tumor-reactive antibodies using protein arrays. We are applying this method to the detection of prostate cancer and prostate cancer recurrence. The changes in the tumor-reactive antibodies are being assessed using the longitudinal approach described above, which may improve the diagnostic performance of those biomarkers and give insight into the role of immune response in determining the likelihood of cancer recurrence.

Pancreatic cancer biomarkers

Other biomarker studies in our lab are focused on pancreatic cancer in collaboration with Anna Lokshin, Michael Hollingsworth, and others in the Early Detection Research Network (EDRN), which is an NCI-sponsored consortium dedicated to discovering and validating cancer biomarkers. We use the glycan and protein detection technologies described above to identify and study biomarkers for the early detection or more accurate diagnosis of pancreatic cancer. We have shown that, in certain cases, the measurement of a glycan on a protein is more accurate for detecting cancer than the measurement of the protein alone in traditional antibody assays. We are now seeking to define which protein and glycan alterations have the highest diagnostic and prognostic significance.

External Collaborators

• Philip Andrews, University of Michigan, Ann Arbor
• Randall Brand, Evanston Northwestern Healthcare, Evanston, Illinois
• William Catalona, Northwestern University, Evanston, Illinois
• Ziding Feng, Fred Hutchinson Cancer Research Center, Seattle, Washington
• Irwin Goldstein, University of Michigan, Ann Arbor
• Samir Hanash, Fred Hutchinson Cancer Research Center, Seattle, Washington
• Michael A. Hollingsworth, University of Nebraska, Omaha
• Anna Lokshin, University of Pittsburgh, Pennsylvania
• Gilbert Omenn, University of Michigan, Ann Arbor
• Alan Partin, Johns Hopkins University, Baltimore, Maryland
• Diane Simeone, University of Michigan, Ann Arbor
• Robert Vessella, University of Washington, Seattle