ABSTRACT
Purpose: Folate receptor alpha (FOLR1) is a membrane bound receptor involved in the transport of folate as well as other regulatory cellular processes. The purpose of this study was to examine the expression of FOLR1 in uterine cancers and to identify changes in gene expression that are associated with overexpression of FOLR1.
Experimental Design: Fifty eight frozen uterine cancer specimens were stained for FOLR1 using immunohistochemistry and results were correlated with transcript expression noted on quantitative PCR. Total RNA from 16 cases of uterine serous carcinoma (USC) was analyzed for gene expression using the Affymetrix HG-U133A and HG-U133B [GeneChip]. USCs overexpressing FOLR1 were compared to cancers with an absence of FOLR1 using binary comparison and template matching of data was used to identify genes that correlate with FOLR1 expression. Selected targets from this analysis were evaluated by quantitative PCR as well as in an independent set of USC represented in quadruplicate on a tissue microarray (TMA).
Results: Overexpression of FOLR1 was observed in 11/16 (69%) of USC and 0/10 normal endometrium cases using frozen tissue specimens. Binary comparison between FOLR1 positive and negative cases identified 121 genes altered by 2-fold at p < 0.01 of which 45 are well correlated with FOLR1 expression pattern. Using quantitative PCR, both mesothelin (MSLN) and PTGS1 (COX1) were significantly increased in FOLR1 overexpressing tumors (p=0.014 and p=0.006 respectively). TMA confirmed that overexpression of FOLR1 and MSLN respectively occurred in 23/48 (48%) and 17/54 (32%) of pure USC.
Conclusion: Both FOLR1 and MSLN are cell surface targets that are co-expressed at high levels in USC and are appealing targets for biologic therapy.
Dataset
In order to assess if a group of genes distinguishes USC with increased FOLR1 expression from those with decreased FOLR1 expression, we performed a binary comparison of the expression data from these two groups. This analysis indicated that 245 array features were differentially expressed between cases with FOLR1 overexpression and those without (parametric student t-test p-value at < 0.005) and that 121 of the 245 array features were differentially expressed by at least 2-fold. The complete data reflective of this analysis is provided.
