Multi-step aberrant CpG island hyper-methylation is associated with the progression of adult T-cell leukemia/lymphoma

Abstract

Aberrant CpG island methylation contributes to the pathogenesis of various malignancies. However, little is known about the association of epigenetic abnormalities with multistep tumorigenic events in adult T cell leukemia/lymphoma (ATLL). To determine whether epigenetic abnormalities induce the progression of ATLL, we analyzed the methylation profiles of the SHP1, p15, p16, p73, HCAD, DAPK, hMLH-1, and MGMT genes by methylation specific PCR assay in 65 cases with ATLL patients. The number of CpG island methylated genes increased with disease progression and aberrant hypermethylation in specific genes was detected even in HTLV-1 carriers and correlated with progression to ATLL. The CpG island methylator phenotype (CIMP) was observed most frequently in lymphoma type ATLL and was also closely associated with the progression and crisis of ATLL. The high number of methylated genes and increase of CIMP incidence were shown to be unfavorable prognostic factors and correlated with a shorter overall survival by Kaplan-Meyer analysis. The present findings strongly suggest that the multistep accumulation of aberrant CpG methylation in specific target genes and the presence of CIMP are deeply involved in the crisis, progression, and prognosis of ATLL, as well as indicate the value of CpG methylation and CIMP for new diagnostic and prognostic biomarkers.

Figure 1

MSP and UMSP analysis of normal PBMC, HTLV-1 carriers, and ATLL patients. A: Representative results of MSP/UMSP of ATLL specimens; PBMC (SssI) was used as a positive control; healthy volunteer PBMC used as negative control; U, unmethylated DNA; and M, methylated DNA. PCR products’ sizes are shown on the right. The number on each lane indicates patients’ specimen number. B: To evaluate the sensitivity of MSP/UMSP in the present condition, various template DNA ratio of ATLL tumor cell line ED40515(−), showing hypermethylation of the SHP1 gene, to normal PBMC without methylation of that gene from 100, 10, 1, 0.1, 0.01, 0.001, to 0%, analyzed by MSP/UMSP. C: The results of a bisulfite sequencing analysis of ED40515(−) and healthy PBMC: filled box is methylated and open box is unmethylated. Horizontal axis shows the CpG sites and vertical axis shows the patient specimen. D: Representative data of bisulfite sequencing. E: To evaluate the effects of HTLV-1 infected cells and/or neoplastic ATLL cells content on MSP/UMSP, specimens of acute type ATLL were diluted to about 1% content of neoplastic ATLL cells with normal PBMCs, which was less than the content of HTLV-1 infected cells in carrier specimens. HTLV-1 viral load, white blood cell counts, and content of atypical lymphocytes (%) of each carrier and acute type ATLL specimen were summarized. F: The typical results of flow cytometry showed that CD4⁺CD25⁺ T cells, which was enriched with HTLV-1 infected cells and/or neoplastic ATLL cells, were contained in acute type ATLL patient #11 (85.52%) and in HTLV-1 carrier #9 (6.85%) specimens. G: Summary of the clinical data including HTLV-1 viral load, white blood cell counts, and content of atypical lymphocytes in HTLV-1 carriers and acute type ATLL specimen. H: Results of a bisulfite sequencing analysis of the SHP1 gene with HTLV-1 carrier specimen, which showed faint signal in MSP analysis. Filled box, methylated; open box, unmethylated. Horizontal axis shows the CpG sites and vertical axis shows the patient specimen.

Figure 2

Distribution of the number of methylated genes in normal PBMC, HTLV-1 carrier, smoldering-, chronic-, acute-, and lymphoma-type ATLL.

Figure 3

SHP1 and p16 protein expression in smoldering, chronic, and lymphoma type ATLL specimens. Top three rows: A typical case with lymph node involvement of ATLL cells in lymphoma type ATLL (A, H&E staining, ×400). Immunohistochemical staining showed that atypical ATLL lymphocytes were positive for CD4 (B, ×400), CD25 (C; ×400), and negative for SHP1 (F, ×400) and p16 (G, ×400) in lymphoma type ATLL specimen. The normal reactive lymph node was used for SHP1 positive control (D, ×400). The endocervical squamous cell carcinoma specimen was used as positive control for p16 (E, ×400). Scale bars = 50 μm. Bottom three rows: Triple-staining of fluorescent immunohistochemistry showed that the expression of SHP1 protein was completely abolished in CD25⁺ lymphoma type ATLL specimen (N, O, P ×400), which was clearly contrast to the strong expression of SHP1 protein in interfollicular lymphocytes of reactive lymphoid hyperplasia (H, I, J ×400). Expression of SHP1 protein in the skin of smoldering type ATLL was intermediate (K, L, M ×400). Alexa fluor 488 (green): anti-SHP1 Ab (H, K and N); Alexa fluor 555 (red): anti-CD25 Ab (I, L and O) Hoechst33258 (blue). Merged pictures were shown in (J, M and P). Scale bar = 20 μm.

Figure 4

The overall survival analyzed by the Kaplan-Meyer method. A: The overall survival of ATLL patients according to the number of methylated genes with Kaplan–Meyer analysis. The overall survival of the ATLL patients significantly decreased as the number of methylated genes increased. B: The overall survival of ATLL patients in terms of the CIMP presence. The CIMP-positive ATLL patients showed a significantly poor prognosis.

Figure 5

Specific genes that show statistically significant increase in the CpG island hypermethylation incidence during the progression of ATLL. Genes associated with arrow indicate significant increase of methylation frequency from arrow start-pointed status to end-point status, which does not always indicate the direction of disease progression. The SHP1 and p73 genes were significantly methylated after HTLV-1 infection (P < 0.05). The hypermethylation of the SHP1 gene was associated with the progression to aggressive type (acute and lymphoma type) ATLL from carrier and smoldering type ATLL, with statistical significance. The HCAD gene was significantly highly methylated at the onset of every type of ATLL, on the other hand the DAPK, MGMT, and p16 genes were significantly hypermethylated at the progression to lymphoma type-ATLL. Statistical data were originated from Table 1.