TFAP2E could down-regulate Wnt signaling, however, was still unknown. In order to further understand the detailed DNA methylation status of KCNAB2, we performed hierarchical clustering of beta values from KCNAB2targeted probes on the HM450 array using the top 50% highest standard deviation. In nonfunctional PAs, KCNAB2 was generally hypermethylated across the promoter region, 59UTR, the first exon, and gene body. This was substantially different from the coverage profile observed in functional PAs. Since KCNAB2 is expressed abundantly in the nervous system and T lymphocytes, and appears to play an important role in K + channel activation and subsequent hormone secretion, we further extended our investigation of this pathway. Genome-scale DNA methylation screening of PAs was performed to investigate whether DNA methylation was associated with PA invasion and histopathological subtype classification. To our knowledge, this is the first genome-scale DNA methylation analysis of sporadic PAs, demonstrating that epigenetic modification of key gene substrates may in part account for functional differentiation of PAs. Tumor invasion has been observed to occur in up to 85% of surgically-resected PAs, based on microscopic analysis of dural samples. Consequently, tumor invasion is perhaps the greatest barrier to achieving adequate tumor control in PAs, as complete surgical resection of noninvasive PAs is typically achieved in 70– 95% of noninvasive PA cases, compared with only 20-40% of invasive PAs. In our study, no significant global differences in CpG methylation were identified between invasive and noninvasive PAs. Although CpG island hypermethylation of a series of well-characterized cell cycle regulation genes, including retinoblastoma 1, CDKN2A, and GADD45G have been linked to gene repression and negative regulated cell growth in PA, these genes were not demonstrated to contribute to the invasive PA phenotype in our study. Importantly, no significant differences in CDKN2A methylation were observed between noninvasive and invasive NFAs in our study, a finding which is consistent with prior observations, and may also suggest that CDKN2A inactivation is alternatively related to PA subtype and size. The current findings suggest that genome-scale DNA methylation assessment may be utilized to identify candidate genes involved in PA invasion that would otherwise possibly be concealed in targeted gene studies. In our study, a secondary analysis showed that the angiogenic gene SLIT3 and oncogene FLT1 may be two candidate biomarkers for invasive PAs, if validated in future studies with a larger PA sample size. Meanwhile, although a previous study showed that promoter hypermethylation of CDH13 and CDH1 was detected in PAs but not in normal pituitary tissue, and promoter hypermethylation of CDH13 was observed more frequently in invasive PAs.