Lactate dehydrogenase (LD; EC 220.127.116.11) isoenzymes are formed by the random combination of two different subunits encoded by two structurally distinct genes, LDHA and LDHB (1). Expression of mammalian LDHA and LDHB is regulated during development and is tissue specific (2)(3); therefore, alterations in the serum LD isoenzyme pattern serve as indicators of pathologic involvement and cancer development (3). In cancer patients, LD isoenzymes originate primarily from tumor tissues and partly from healthy tissues damaged by tumor expansion and invasion. Different phenotypes may originate from expression regulation by other regulatory genes and by the alteration of LDHA or LDHB caused by mutation; chromosomal deletion; duplication, or increase of copy number; and promoter methylation. The increase in LD1 correlates with the total copy number of the short arm of chromosome 12 in tumor cells (4). Recently, we found a high proportion of LD1 in a patient with retinoblastoma. The unique LD isoenzyme pattern was attributable to transcriptional silencing by promoter hypermethylation of LDHA (5).
In mammals, DNA methylation usually occurs at CpG dinucleotides, which are cytosines located 5′ of guanines. Methylation is known to play a role in regulating gene expression during cell development, X chromosome inactivation, genomic imprinting, and carcinogenesis (6)(7). In neoplastic cells, some CpG islands in the promoter region that are usually unmethylated become aberrantly methylated, and this leads to transcriptional silencing. Therefore, an epigenetic event is thought to be one mechanism for the inactivation of tumor suppressor genes (8).
Human LDHB has a CpG-rich region in its promoter that is similar to that of human LDHA and LDHC (9). We found that five cancer cell lines had only LDHA mRNA (10). Most gastrointestinal cancer patients had electrophoretically slow-moving isoenzymes and the LD-A subunit in their sera (3). We predicted that this unique pattern was derived partly from transcriptional silencing attributable to the aberrant promoter hypermethylation of LDHB. We have focused in this work on the aberrant methylation of the promoter region of LDHB in some cancer cell lines and gastrointestinal cancer tissues, with the intention of identifying the relationship between the LD isoenzyme pattern and aberrant promoter methylation.
The present study included 12 cancer cell lines, 20 patients with gastric cancer, and 25 patients with colorectal cancer. The cancer cell lines were the same as those used in our previous studies (10)(11). For the patients, both malignant and nonmalignant tissues were examined. They were obtained from the National Cancer Center Hospital and Hamamatsu University School of Medicine. Each patient consented to the experimental use of specimens and examination of the specimens for pathology. DNA was extracted from the cancer cell lines and the resected tissues by a method described previously (10)(11).
For methylation analysis, bisulfite-PCR single-strand DNA conformation polymorphism (BiPS) analyses were done (12). Briefly, bisulfite treatment was carried out and PCR-single-strand conformation polymorphism analysis was performed with 10% nondenaturing polyacrylamide gels and silver-staining detection (Daiichi Pure Chemicals). The primer sequences for amplification of LDHB were 5′-AGGGAGTGTGTATATTTGAGTT-3′ (sense) and 5′-TCAAACTTACCTATAAACCAAA-3′ (antisense). The promoter sequence of LDHB was taken from GenBank accession no. X13794 and is shown in Fig. 1A⇓ . The region selected for amplification contained exon 0 and has been related to the promoter activity (13). The PCR product was expected to contain 282 bp and 14 CpG sites. The PCR products were sequenced directly by the dideoxy-sequencing procedure with a BigDye Terminator Cycle Sequencing FS Ready Reaction Kit and a PRISM 310 Genetic Analyzer (PE Applied Biosystems) according to the manufacturer’s instructions (12). The amplified products were also cloned into a pDRIVE cloning vector (Qiagen) for sequencing.
Promoter methylation of LDHB was detected in four gastric cancer cell lines and one pancreatic cancer cell line (Table 1⇓ ; Fig. 1B⇓ ). The five cancer cell lines, the LDHB promoters of which were methylated, had only the LD5 isoenzyme at the enzyme activity level and only the LDHA cDNA signal at the mRNA level (10). The methylation status of LDHB in the cell lines was in complete accord with the mRNA and activity expression results. One pancreatic cancer cell line, PSN1, yielded a partially methylated band by BiPS and sequence analysis (Fig. 1⇓ , B and D). This complicated BiPS pattern was reproduced by a repetitive BiPS analysis. The four gastric cancer cell lines yielded only a completely methylated BiPS band, and the other cancer cell lines yielded only an unmethylated BiPS band.
Promoter methylation of LDHB was observed in 3 of 20 gastric cancer tissues and in none of the corresponding healthy mucosa (Fig. 1C⇓ ). None of 25 colorectal cancer tissues and corresponding healthy mucosa had promoter methylation in LDHB. Therefore, LDHB promoter was methylated in 5 of 12 cancer cell lines and in 3 of 45 cancers (P = 0.007, Fisher exact test).
The methylation pattern found by BiPS was heterogeneous in the three gastric cancer tissues. Sequence analysis of the PCR products after cloning revealed differences among the three gastric cancer tissues (Fig. 1D⇓ ). GC-1, which had slight methylation in LDHB, was an early gastric cancer (3.5 × 3 cm; stage Ia) with severe intestinal metaplasia located at the proximal portion. GC-2, which showed mild methylation, was an invasive cancer (stage IIIa) with histologic heterogeneity. GC-3, which had heavy methylation, was an invasive cancer (stage II) with histologic heterogeneity.
LDHB promoter was methylated in 5 of 12 cancer cell lines, but in only 3 (7%) of 45 clinical cancer tissues. Therefore, methylation of the promoter is a relatively uncommon mechanism for the frequent increase of cathodal LD isoenzymes in gastric and colorectal cancer patients. The high proportion of cell lines with methylation of the promoter might be attributable to our selection of a small series of cell lines, but the higher proportion in cell lines than in cancer samples from patients might also reflect the biological differences found in the recent report (14).
The presence of partially methylated DNA in the PSN1 cancer cell line is similar to that of hMLH1 in the colorectal cancer cell line C-1 (12). Partial methylation may be caused by the heterogeneous methylation status in the cultured PSN1 cells. The reason LDHB mRNA is not expressed in PSN1 could be that the unmethylated allele in the LDHB promoter has a mutation, a large deletion in a downstream coding region, or a methylation in other regions not examined in this study.
Interestingly, the present study showed that the methylation of the LDHB promoter was a characteristic for the cancers and not a germ-line abnormality of the patients. It means that promoter methylation of LDHB is an epigenetic abnormality but not a genomic alteration.
The three gastric cancer patients with promoter methylation of LDHB did not share any common specific similarities in clinical and pathologic findings except for widely expanded cancer. This issue should be further investigated in the future. The heterogeneous BiPS pattern in the three gastric cancer tissues might be attributable to clonal differentiation with different methylation patterns and histologic heterogeneity. This issue, however, should also be elucidated by a large-scale clinicopathologic study.
Patients with many malignancies combined with increased serum LD activity have a poor prognosis (15). Therefore, additional studies of the regulation of expression of LD and the release of LD from the cancers in patients are warranted.
LDHA expression can be induced by estrogen (16), cyclic AMP (17), hypoxia(18), and c-Myc (19). Induction is caused presumably by the actions of these agents on the LDHA promoter (20). Accordingly, serum LD3, LD4, and LD5 are frequently increased in patients with malignant diseases and reflect increased expression of LDHA by neoplastic cells. We found that the other important reason for the increase in the amounts of electrophoretically slow-moving LD isoenzymes in cancer patients is transcriptional silencing of LDHB expression because of aberrant methylation in the promoter region of LDHB. The frequency of LDHB methylation is not high, but it should be noted that enzyme abnormalities in tumors occasionally originate from aberrant methylation.
This research was supported in part by Grants-in-Aid for Cancer Research (9-13, 13-5) and for the Second Term Comprehensive 10-Year Strategy for Cancer Control from the Ministry of Health, Labour and Welfare, Japan, and by a Grant-in-Aid for Scientific Research (13470518) from the Ministry of Education, Science, Sports, Culture and Technology, Japan. We thank Prof. M. Kanamori for helpful discussion and C. Tatebayashi, M. Ushiama, and N. Fukayama for their technical assistance.
- © 2003 The American Association for Clinical Chemistry