Phenotypic differences between two Gnmt−/− mouse models for hepatocellular carcinoma
✍ Scribed by Yi-Ming Arthur Chen; Yi-Jen Liao; Shih-Ping Liu; Ting-Fen Tsai
- Publisher
- John Wiley and Sons
- Year
- 2009
- Tongue
- English
- Weight
- 115 KB
- Volume
- 49
- Category
- Article
- ISSN
- 0270-9139
No coin nor oath required. For personal study only.
✦ Synopsis
We would like to express some technical concerns about the results presented in the article by Mato and coworkers in the April 2008 issue of HEPATOLOGY. 1 We have identified the gene for glycine N-methyltransferase (GNMT) as a liver cancer susceptibility gene, generated GnmtϪ/Ϫ mice, and reported on their tendencies toward chronic hepatitis, glycogen storage disease, and liver cancer. [2][3][4][5][6] Although the GnmtϪ/Ϫ mouse models described by Dr. Mato's team and our group share similar strain (129 substrain/B6) backgrounds, they have at least three significant differences in terms of phenotype: (1) Dr. Mato's group reported that their male GnmtϪ/Ϫ mice developed hepatocellular carcinoma (HCC) at 8 months old, whereas our GnmtϪ/Ϫ mice did not develop HCC until a minimum of 14 months of age (up to 24 months); (2) we showed that seven of seven (100%) female GnmtϪ/Ϫ mice develop HCC, whereas Dr. Mato did not mention whether their female GnmtϪ/Ϫ mice had developed HCC; and (3) Dr. Mato's team described global DNA hypermethylation in the liver tissues of their GnmtϪ/Ϫ mice using capillary electrophoresis, whereas we used both [ 3 H]methyl group labeling and a commercially available enzyme immunoassay to identify global DNA hypomethylation in liver tissues from both male and female GnmtϪ/Ϫ mice. 1,6 Published reports indicate that global DNA hypomethylation and promoter gene cytosine-guanosine dinucleotide hypermethylation are both present in human HCC. 7 However, with the exception of Dr. Mato's report, no one has stated that gene knockout mice can develop HCC as early as 8 months of age. One possible explanation for the differences between these two GnmtϪ/Ϫ models is the completeness of the Gnmt gene section being knocked out. The Gnmt gene has six exons; we knocked out exons 1-4 and part of exon 5, 4 whereas Dr. Mato and his collaborator, Dr. Wagner, only knocked out exon 1. 8 Results from a nucleotide sequence analysis indicate that the disrupted allele generated in Dr. Wagner's model contained three potential promoters in intron 1 (Fig. 1). In addition, four potential in-frame ATG codons were located downstream from the potential promoters (three in intron 1 and one in exon 2).
This type of disrupted gene may generate a truncated form of GNMT that lacks the N-terminal 73 amino acids. Previous reports show (1) Gnmt uses its N-terminal 20 amino acids to form a tetramer in the cytosol and (2) the presence of a phosphorylation site in serine-9 of the Gnmt protein. 9,10 We feel there is a need to clarify whether Dr. Mato's GnmtϪ/Ϫ mice generated such a truncated form of Gnmt, which may cause it to act as a dominant-negative protein, thereby implying its involvement in tumorigenesis. If this is the case, then the mouse model they generated is not a complete GnmtϪ/Ϫ mouse model.
In summary, we believe we are the first group to demonstrate that GNMT is a liver cancer susceptibility gene. Although both teams showed HCC development in GnmtϪ/Ϫ mice, without further data, we suspect that Dr. Mato's model is an incomplete Gnmt knockout model.