1f and data not shown). Although NEFH was expressed in KYSE30 at baseline, it was further increased by 5-Aza-dC (Fig. 1e, middle), indicating that its expression was at least partially suppressed by methylation. These results suggest that promoter hypermethylation nearly of NEFH is one of the causal factors of loss of NEFH expression. We then performed RT-PCR and real time RT-PCR in cDNA prepared from tumor and normal tissues, and found that NEFH was significantly down-regulated in ESCC (Fig. 1e, right and Fig. 1g). To investigate protein expression of NEFH, we performed immunohistochemical (IHC) staining using ESCC tissue arrays. Among 26 pairs of ESCC (PT) with matched normal adjacent tissues (PN) analyzed, decreased NEFH expression was observed in 10 cases (38.5%) (Table S1 and S2).

NEFH expression was detected in 24 out of 26 cases (92.3%) of non-malignant esophageal tissues (normal, inflammation and hyperplasia), whereas absent or faint expression of NEFH was observed in 35 out of 47 cases (74.5%) of neoplastic tissues (Table 1 and Table S3). When compared to normal esophagus tissue, the negative expression of NEFH in adenocarcinoma (75%, 15/20) (P<0.001), squamous cell carcinoma (66.6%, 12/18) (P=0.002), and metastatic cancer (88.8%, 8/9) (P<0.001) was statistically significant but this was not the case in hyperplasia (20%, 2/10) (P=0.477) (Fig. 1h). Table 1 Immunohistochemical analysis of NEFH in esophageal cancer tissue microarray with normal tissue controls (ES804).

Knockdown of NEFH Increases Tumor Growth Since the NEFH promoter was specifically methylated in tumor tissues and its expression was down-regulated in ESCC, NEFH might function as a tumor suppressor in esophageal cancer. In order to test this hypothesis, we first examined whether knockdown of NEFH increased cell growth after establishing NEFH shRNA stable clones in KYSE30 cells (Fig. S2a�CS2c). A significant increase in clonogenic cell growth was observed in cells with diminished NEFH expression (N12 and N20) (Fig. 2a). Increased cell proliferation was also observed in N12 and N20 cells compared to control cells (C2) (Fig. 2b). Because cell proliferation is closely linked to progression of the cell cycle, cell cycle analysis was performed by flow cytometry. The population of N12 and N20 cells residing in the G0/G1 phase yet decreased in the G2 phase compared to C2 cells (Fig.

2c), indicating that cell cycle progression through G0/G1 and then a block at G2 occurred from loss of NEFH. In addition, in vitro invasive activity was increased in N12 and N20 cells (Fig. 2d). When cells were Carfilzomib incubated in 10% serum medium, the number of cells passing into the chamber increased about 2.5- and 7-fold in N12 and N20 cells respectively, compared to C2 cells (Fig. 2d, left). Under HGF treatment, the invasive activity of N12 and N20 cells increased about 3.