Abstract 3949: Deep immunophenotyping using flow cytometry of dissociated cells from Barrett's esophagus and matched adjacent squamous epithelium defines distinct phenotypic clusters

Barrett's esophagus (BE) is defined as metaplasia of the squamous epithelium to a specialized columnar epithelium with risk factors of gastroesophageal reflux and obesity and a predilection for middle-age and older white males. BE progresses through stages of dysplasia (low-grade and high-grade) before developing into esophageal adenocarcinoma. Challenges remain in early detection and predicting which patients may progress to dysplasia. Here we describe a method by which we compare human clinical IRB-approved BE biopsies and adjacent normal squamous epithelium using tissue dissociation and deep immunophenotyping by flow cytometric collection and analysis. A cassette of canonical epithelial or tumor stem cell-associated targets (EpCAM, CD24, CD44, CD49f, Her2/neu, CD133, CD90, CD166, and CD29), immune cell markers (CD3, CD45, CD127, HLA-DR, CD16, CD56, CD4, CD8, CD25, and CD19), as well as targets associated with myeloid derived suppressor cells (CD14, CD15, CD33, CD11b, HLA-DR, CD31 and CD86) were used to discern differences across subjects and between cellular compartments in normal and BE tissue. The Barrett’s samples show a majority population with a characteristic phenotype (EpCAM+CD133lowCD49fhigh) when compared with normal squamous tissue samples (EpCAM-CD133-CD49flow). The samples separate into two discrete groups using hierarchical clustering based on differential surface marker expression of combined epithelial and immune cell markers, but also reveal unexpected, shared phenotypes for some normal and BE samples. Principal component analysis supports this grouping and was used to identify more compelling targets for categorization, such as CD133 and CD49f. The resulting expression and distribution of targets offer a phenotypic fingerprint characterizing both the epithelial cell and immune cell compartment. Besides providing the potential for revealing clinically relevant differences between BE and normal tissue, as well as across subjects, the discovered surface immunophenotypes can be used to target specific subpopulations from dysplastic tissue for further molecular investigation. A deeper understanding of the role of such specific subpopulations should increase the prospects for more complete understanding of BE and its progression. Citation Format: Friedrich G. Hahn, Eileen Snowden, Warren Porter, Mitchell Ferguson, William S. Dillmore, Stephanie S. Yee, Taylor Black, Maureen DeMarshall, Aaron Middlebrook, Smita Ghanekar, Anil Rustgi, Erica L.