---------------------------- 2013 Platform Presentation Winner --------------------------

[1218] Targeted Development of a Specific Biomarker of Endometrial Stromal Cell Differentiation Using Bioinformatics: The IFITM-1 Model

Carlos Parra-Herran, Liping Yuan, Marisa R Nucci, Bradley J Quade. Brigham and Women's Hospital, Boston, MA

Background: Distinction of endometrial stromal tumors from smooth muscle tumors can be difficult. Although immunohistochemistry (IHC) is helpful, care must be taken as the only established stromal differentiation marker, CD10, is expressed diffusely in 20% of leiomyosarcomas. Consequently, we took a bioinformatics approach to identify a new marker by searching the Human Protein Atlas, a database of protein IHC expression profiles.
Design: The search for a new endometrial stromal cell differentiation biomarker began by downloading the raw data file from proteinatlas.org, and importing it into a Microsoft Access database. Our method used Structured Query Language to filter IHC results for those with “strong” staining in endometrial stroma AND “low” staining in endometrial glands and myometrium. Performance of the top candidate antibody was then evaluated with tissue sections from a variety of pathological diagnoses having or related to stromal cell differentiation [Table 1]. IHC was scored in terms of intensity (negative=0 to strong=3) and distribution (absent=0 to diffuse=3).
Results: 790,019 normal tissue-antibody test pairs were filtered by our purely informatic-based method to identify 10 unique candidates. From these 10, interferon induced transmembrane protein 1 (IFITM1) was selected as the top candidate by visually inspecting images of endomyometrium published on the Protein Atlas website. Both the intensity of normal stromal cell staining in proliferative endometrium and the absence of staining in myometrial cells were considered in this evaluation. The selected antibody was then tested against our expanded tissue collection. Table 1 summarizes our IHC results.

Conclusions: Database mining proved to be a rapid, productive method of biomarker discovery. IFITM1 is a specific marker of endometrial stromal differentiation from proliferative endometrium to metastatic stromal sarcoma. Like CD10, IFITM1 is moderate-strongly and diffusely expressed in stromal cells. IFITM1 expression was found in about 25% of cellular leiomyomas, but typically with limited staining intensity and in a patchy distribution. This new marker should be a valuable addition to the IHC panel used in the diagnosis of cellular mesenchymal uterine tumors.