Background & Aims: Metastatic cancers arise from a decades-long succession of increasingly virulent precursor lesions, each of which represents prospective targets for therapeutic intervention. This evolutionary process has been particularly vivid in esophageal adenocarcinoma (EAC), as this cancer and associated precursor lesions, including Barrett's esophagus (BE), low-grade dysplasia (LGD), and high-grade dysplasia (HGD), coexist in an accessible, 2-dimensional pattern in esophageal mucosa. Given the durability of these precursor lesions, it is likely that they, like EAC, rely on stem cells for their regenerative growth. To assess the role of stem cells in the evolution of EAC, we apply technology that selectively clones stem cells from the gastrointestinal tract to patient-matched endoscopic biopsies from each of the precursor lesions implicated in EAC. Methods: Histologically validated, endoscopic biopsy series were used to clone stem cells from each precursor stage, which were then characterized by gene expression, genomic, and functional assays to trace the evolutionary trajectory toward malignancy.

The origin of Barrett's esophagus (BE) and gastric intestinal metaplasia (GIM), obligate precursors of esophageal adenocarcinoma (EAC) and intestinal gastric cancer (iGC), has intrigued investigators for decades and would likely guide preemptive strategies. From endoscopic biopsies of clinically confirmed BE and GIM, we have cloned stem cells committed to intestinal metaplasia in vitro. Remarkably, the gene expression profiles of BE and GIM stem cells are highly related down to broad arrays of transcription factors compared with stem cells of the normal gastric mucosa. Using cell surface markers in common between BE and GIM stem cells, we have identified clusters of cells at the squamocolumnar junction and the distal stomach in mice and have used fluorescence-activated cell sorting (FACS) to clone these cells from both sites. These murine clones can be differentiated in air-liquid interface cultures, recapitulating key features of the human lesions.

We have applied a single-cell cloning technology to patient-matched endoscopic biopsies of Barrett's esophagus and co-existing lesions of low-grade dysplasia (LGD), high-grade dysplasia (HGD), and esophageal adenocarcinoma (EAC). In vitro differentiation and xenografting of these clones yields epithelia with histology corresponding to the origin of these clones, and the EAC clones yield aggressive tumors in vivo. In addition, genomic analyses of these clones reveals their phylogenetic relationships based on common and accumulating mutations. In efforts to identify drugs that would target the Barrett's stem cells for preemptive therapeutics, we performed parallel screens of small molecule libraries against Barrett's stem cells and patient-matched normal esophageal stem cells. Synthetic lethal strategies ultimately yielded drug combinations that showed low nanomolar efficacy against Barrett's stem cells.