Immunohistochemistry (IHC) has become a key technique to assess expression as well as subcellular localization of proteins of interest in tissues.1,2
IHC plays an essential role both in clinical medicine and in basic research. Tissue microarrays (TMAs) have been successfully incorporated as a high-throughput platform for cancer research.3
For instance, Kampf et al. generated TMAs, which contained 48 normal human tissues along with 216 tumors representing the 20 most common cancer types.4
TMAs are useful because they can provide diagnostic information on a large number of patient samples while minimizing the number of slides to be processed.
Biomedical studies are often carried out on model systems employing cell lines to address the functional importance of molecules of interest. However, it still remains difficult to select rapidly and accurately appropriate cell line(s) for functional studies. In 2005, Ferrer et al. developed and immunophenotyped a panel of paraffin-embedded cell lines using five human prostate cancer cell lines and a human cervical adenocarcinoma cell line.5
Waterworth et al. published a cell array containing 23 cancer cell lines, of which two were derived from pancreatic cancer (MiaPaCa and Panc-1).6
In 2006, Andersson et al. described cell TMAs (cells and tissues) as a tool for antibody-based proteomics using 46 frequently used cell lines in addition to 12 patient cell samples.7
Therefore, microarrays generated from cultured cell lines which are formalin fixed and embedded in paraffin can serve as a platform for in vitro
analysis of protein expression profiles. To our knowledge, no cell microarrays (CMAs) have yet been reported that were designed to evaluate protein expression and subcellular localization for a single cancer type in a comprehensive fashion using a panel of a large number of cell lines.
As a pilot, we choose pancreatic cancer, which is the fourth leading cause of cancer-related deaths in the United States, with a 5-year survival rate <6%.8,9
A number of studies have reported potential biomarkers for this deadly disease. We have previously developed a compendium of potential biomarkers in pancreatic cancer cataloguing altered genes curated from the published literature.10
Despite the availability of this long list of potential biomarkers, no systematic studies have yet been reported to validate them across frequently studied cell lines. In this study, a comprehensive CMA for human pancreatic cancers was generated that contains all pancreatic cancer cell lines available from the American Type Culture Collection (ATCC, http://www.atcc.org
) in addition to several others that were generated at Johns Hopkins. We have tested these pancreatic cancer CMAs by immunocytochemical labeling for Ep-CAM (epithelial cell adhesion molecule) to determine the expression of this epithelial marker. To address the general applicability of the pancreatic cancer CMA, we also performed immunocytochemical labeling for CA19-9 (carbohydrate antigen 19-9) and CD44, molecules that are known to be overexpressed in some pancreatic cancers.11
In addition to these, we also examined two potential biomarkers, HLA class I and TMPRSS4 (transmembrane protease, serine 4), and observed strong labeling of the large majority of the pancreatic cancer cell lines present on the CMA. Overall, our results indicate that CMA is a high-throughput platform for antibody-based screening in cancers for rapid characterization of cell lines for antigens of interest.