Cancer specific promoter CpG Islands hypermethylation of HOP homeobox (HOPX) gene and its potential tumor suppressive role in pancreatic carcinogenesis

Abstract

Background: We have recently identified HOP hoemobox (HOPX) as a tumor suppressor gene candidate, characterized by tumor-specific promoter DNA hypermethylation in human cancers, and it can remarkably inhibit tumors' aggressive phenotypes. In this current study, we for the first time examined methylation level of HOPX and tested the functional relevance in pancreatic cancer (PC).

Methods: Clinical features of HOPX promoter hypermethylation was investigated in 89 PC tissues, and immunohistochemistry was added. We also examined its functional relevance in phenotype assays such as soft agar, proliferation, invasion, and cell cycle analysis.

Results: PC tissues had HOPX gene hypermethylation as compared to the corresponding normal pancreas tissues, and its uniqueness was robust to discriminate tumor from normal tissues (AUC = 0.85, P < 0.0001). Unexpectedly, HOPX was increased in expression in tumor tissues, and immunohistochemistry revealed its predominant expression in the Langerhans islet cells, where HOPX was reduced in expression for PC cells with promoter hypermethylation. HOPX transfectants exhibited G1 arrest with subG1 accumulation, and inhibited tumor forming and invasive ability.

Conclusion: Defective expression of HOPX which is consistent with promoter DNA hypermethylation may explain aggressive phenotype of pancreatic cancer, and intense expression of HOPX in the Langerhans cells may in turn uniquely contribute to pancreatic carcinogenesis.

Figure 1

Analysis of HOPX-β methylation and expression in pancreatic cancer cell lines-1. (A) Schematic diagram of the 3 spliced transcript variants and common transcript core in HOPX (middle panel) and of CpG islands (gray area) in the 5′-flanking region of HOPX gene (bottom panel). Vertical bars indicate the dinucleotides CpG. Arrows indicate the sequences for bisulfite sequencing analysis or Q-MSP, respectively. F1, F2 and F3 represent forward primers for HOPX-α (331 bp) and HOPX-γ (456 bp), HOPX-β (376 bp), and HOPX-core (254 bp) in RT-PCR or Q-RT-PCR; R, common reverse primer; P, probe for Q-MSP; TSS, transcription start site; ATG, translation start codon. (B) Expression level of HOPX in PC cell lines was examined by RT-PCR (left panel) and Q-RT-PCR (HOPX-β and core/β-actin x 100, (right panel). (C) Expression level of HOPX in PC cell lines was examined by WB (top panel) and IP/WB (bottom panel). Transfectants we performed had the V5 epitope and polyhistidine region in the C-terminal peptide, and so, added approximately 5 kDa to original protein.

Figure 2

Analysis of HOPX-β methylation and expression in pancreatic cancer cell lines-2. (A) Representative bisulfite sequencing results in 5 PC cell lines and TE15. Arrowhead indicates dinucleotide CpG. (B) Cloned PCR products from PC cell lines. White and black circles denote unmethylated and methylated CpG sites, respectively. X means seven nucleotide deletion; AGGCCGG. (C) mRNA expression by RT-PCR (top panel) and Q-RT-PCR (bottom panel) after treatment with the demethylation agent, 5-aza-dC, in the presence or absence of TSA, a histone deacetylase inhibitor. 1A and 5A, 1 and 5 μM 5-aza-dC; T, TSA.

Figure 3

HOPX expression status in PC. (A) Expression level of HOPX in PC was tested in RT-PCR (top panel) and Q-RT-PCR (bottom panel). T, primary tumor; N, corresponding pancreatic tissue. (B) Expression level of HOPX in PC was examined by western blotting. (C) Immunohistochemical staining for HOPX in primary tumor (top panel) and normal tissue (bottom panel), with hematoxylin eosin staining (original magnification, X40). These immunohistochemical stainings were performed by short term exposure of DAB. (D) In this condition, islet cells only stained (original magnification, X400). scale Bars, 100 μm.

Figure 4

Immunohistochemistry and Quantitative methylation analysis in 89 samples. (A) Representative immunohistochemical staining for HOPX in normal tissues and primary tumor with or without HOPX-β hypermethylation (original magnification, left X200, right X400). scale Bars, 100 μm. (B) Frequency of HOPX-β hypermethylation by Q-MSP. Dashed line indicates the optimal cut-off value (1.5). (C) ROC curve of HOPX-β methylation for detection of PC. Area under the curve (AUC) represents the accuracy in discriminating normal from tumor in term of sensitivity and specificity (P < 0.0001). (D) Methylation value of HOPX-β in JPS stage III, IVa and IVb. Data are expressed as mean ± SD. (E) Identification of an optimal cut-off value for the prognosis using the log rank prognostic analysis.

Figure 5

Functional analysis of HOPX in PC cells. (A) HOPX expression level in HOPX expressing stable cell lines was determined by mRNA expression (lower panel) and protein expression (upper panel). HOPX protein was detected by WB with HOPX antibody (3D6) and the flag V5 antibody. β-actin was shown as a loading control. (B) Proliferation assay was performed for 5 days. Data are shown as absorbance at 450 nm. error bars, SD. (C) Anchorage-independent colony formation assay was performed. After 3 weeks of cell culture, colonies were counted and photographed at 40× magnification under a microscope. Colonies were also visualized by ethidium bromide staining. error bars, SD. (D) Image of cell cycle assay. Thick black bars, subG1 phase. (E) Matrigel invasion assay. After fixation and staining, invading cells were photographed and counted at 100x magnification. error bars, SD.