Research Interests
Our laboratory works to enable more accurate detection and diagnosis of pancreatic cancer and to identify key factors in cancer progression and dissemination. Our approach involves understanding the nature and functions of changes to carbohydrates in pancreatic cancer.
Glycoprotein biomarkers for pancreatic cancer
A great need exists for better tools to detect and diagnose pancreatic cancer. We are addressing this problem by taking advantage of a frequently observed molecular feature of pancreatic cancer: alterations to the carbohydrate side chains of cell-surface and secreted proteins. Most secreted proteins have carbohydrates (also called glycans) attached to them. The glycans produced by cancer cells can have altered structures relative to those produced by normal cells. The detection of these altered glycans forms the basis of our biomarkers. Conventional biomarkers are based on the detection of a particular protein level; we detect both the protein and the glycan attached to the protein. By measuring both types of information, more accurate detection of cancer can be achieved.
The experimental platform that facilitates these studies is the antibody-lectin sandwich array (ALSA), developed in our laboratory. The method starts with a microarray of antibodies that target various glycoproteins of interest. A complex biological sample is incubated on the array, resulting in the capture of glycoproteins by the antibodies. Then the array is probed with a lectin (a protein with carbohydrate-binding activity), which binds to the captured glycoproteins that bear the lectin’s glycan target. The amount of lectin binding at each antibody indicates the amount of glycan on the proteins captured by that antibody. Diverse lectins can be used to probe a variety of glycans on a given sample. In addition, the captured proteins can be probed with antibodies targeting the core proteins, as in a “sandwich” immunoassay, to obtain the levels of the proteins in parallel assays.
Relative to other technologies, the platform offers a unique combination of capabilities such as reproducible glycan measurements on specific proteins, high-throughput sample processing, and high-sensitivity detection directly from biological samples. These features make the platform ideal for glycoprotein-based biomarker studies. A product based on this technology is now available from GenTel Biosciences (Madison, Wisconsin).
Using this tool, we can now explore the hypotheses that particular glycan structures on specific proteins are found uniquely in certain disease states and that their measurement yields effective detection of cancer. We have characterized the prevalence in pancreatic cancer patients of a variety of glycan structures on several types of proteins. We showed that the glycan levels are altered independently of changes to the protein level, so that measuring both the glycan and protein level gives improved biomarker performance relative to measuring only protein levels in standard immunoassays. Because of that relationship, we have developed several highly effective candidate biomarkers that are moving forward in clinical validation studies.
We work with our clinical collaborators to address several clinical needs. One such need is to help doctors make a more accurate diagnosis of patients with suspected pancreatic cancer. Since pancreatic cancer can be difficult to distinguish from benign conditions of the gastrointestinal tract, highly accurate biomarkers are needed to match patients to the appropriate procedures at the earliest possible time. We also are testing our novel biomarkers for use in drug trials. Biomarkers that give early indications of the effectiveness of a candidate drug could accelerate drug trials and better match patients with the drugs that benefit them most.
A related class of biomarker we are developing is for the diagnosis of patients with pancreatic cysts. Cystic lesions of the pancreas are increasingly being recognized due to the widespread use of high-resolution abdominal imaging. Since certain cyst types are precursors of invasive cancer, this situation presents an opportunity to remove the pre-malignant cysts prior to progression. Unfortunately, accurately differentiating pre-malignant from benign cysts can be difficult. In collaboration with Dr. Diane Simeone at the University of Michigan, we have identified glycan variants of secreted mucins that distinguish benign from pre-cancerous cysts with an 87% accuracy—better than the best current markers. Ongoing work is aimed at validating and building upon these results. Ultimately, we hope to implement a test that could be used to determine which pancreatic cysts should be surgically removed in order to prevent progression to cancer.
New glycan-detection reagents and bioinformatics of protein-glycan interactions
To facilitate the above studies and enable the detection of a wide variety of glycan structures, we are developing new analytical reagents for detecting glycans. Many glycan-binding proteins (lectins) are available commercially, but most do not have the properties required for analytical use, such as high binding affinity and stability. As a consequence, some glycan structures cannot be detected reliably with a lectin. The development of lectins to detect under-explored glycan structures would enable studies of their biological roles or potential use as biomarkers. This work is guided by bioinformatics analyses to characterize and study the specificities of a wide range of lectins and glycan-binding antibodies. The bioinformatics identification of proteins that may specifically detect particular glycan structures leads to experimental work to develop the proteins as analytical reagents. Our bioinformatics work also is valuable for studying the biology of lectins and the diseases in which lectins are involved.
Detecting and targeting pancreatic cancer cell differentiation
Our laboratory also studies a process which contributes to the very poor survival rate associated with pancreatic cancer. The extreme lethality of pancreatic cancer is related to its tendency to metastasize at early stages, prior to diagnosis, and its resistance to chemotherapeutics. We are investigating the concept that the cancer cells that acquire these migratory and drug-resistant traits have undergone a major “differentiation” process accompanied by gross morphological and behavioral changes. Most cancer cells in primary pancreatic tumors are part of structures called ducts that bear resemblance to their counterparts in the normal pancreas. However, a subset of cancer cells may exist that completely change their morphology and behavior so that they are free to migrate and seed new tumor sites. In order for cancer cells to become migratory, they must break away from the epithelial ductal structures and take on characteristics of migratory, mesenchymal cells. This transition involves enormous remodeling of the cell and is likely driven by genetic aberrations, extracellular signals, and the activation of differentiation programs in the cancer cells. We have developed model systems by which we are seeking to identify the key factors that drive or enable that process. We will use these findings to develop new strategies for interfering with cancer cell progression.
Figure 1
 |
Figure 1. Protein and glycan detection using antibody arrays. a) Array-based sandwich assays for protein detection. Multiple antibodies are immobilized on a planar support, and the captured proteins are probed using biotinylated detection antibodies, followed by fluorescence detection using phycoerythrin-labeled streptavidin. b) Antibody-lectin sandwich arrays (ALSA). This format is similar to above, but the detection reagents target the glycans on the capture proteins rather than the core proteins. The glycans on the immobilized antibodies are chemically derivatized to prevent lectin binding to those glycans. c) Example antibody array results for core protein detection (left) and glycan measurement (right). SA-PE, streptavidin-phycoerythrin. |
Figure 2
 |
Figure 2. Distinct changes to glycan levels associated with cell type. Cell lines were treated with various pro-inflammatory signals, including the cytokines IFNg, TNFa, IL-a1, and oxidative stress (H2O2). The cell lines and their treatments are indicated by the column labels. Six cell lines were treated: two bearing cell-surface markers characteristic of tumorigenicity (labeled in red); two not bearing the markers (labeled in black); and two partially bearing the markers (labeled in green). Using the ALSA assay, the levels of various glycans on the mucins MUC1, MUC5AC, and MUC16 in the secretions of the cells were measured before and after treatment. The row labels indicate the lectin used for detection (which determines the glycan detected) and the capture antibody. The color of each square represents the fold-change of the signal after treatment divided by the signal before treatment. The cells bearing markers of tumorigenicity uniquely increased particular glycans, showing a difference from the other cells in their glycan characteristics. See Wu et al., J. Proteome Research, 2009. |
