People with fatty liver are more likely to have the PEMT rs7946 SNP, yet populations with the mutant allele do not have fatty liver

Abstract

We reported that individuals with nonalcoholic fatty liver were significantly more likely to have a single nucleotide polymorphism (SNP) in the gene encoding for phosphatidylethanolamine N-methyltransferase (PEMT). This SNP (rs7946, +5465 G. → A) leads to a V175M substitution and is a loss of function mutation (1). Drs. Romeo, Cohen, and Hobbs conducted an interesting study further investigating our observation. They asked whether individuals with the mutant allele (A) were more likely to have fatty liver, and concluded that this is not the case. At first glance, the two reports appear to be arriving at opposite conclusions. Populations with fatty liver are more likely to have the mutant allele (Song et al.), yet populations with the mutant allele do not, on average, have fatty liver (Romeo et al.). However, there is an important difference in the questions being asked. When the underlying mechanisms are considered, it is apparent that both observations can be true. PEMT catalyzes de novo synthesis of phosphatidylcholine in liver (2, 3). Loss of function of this enzyme can result in fatty liver because phosphatidylcholine is required for hepatic secretion of triacylglycerol in very low density lipoproteins (VLDL) (4–6). Thus, the rs7946 SNP in PEMT slows the rate of removal of fat from the liver. If the rate of production of triacylglycerol exceeds the rate of removal, fatty liver occurs. It is possible that under normal circumstances the rate of triacylglycerol synthesis is slow enough that even the mutant PEMT activity is sufficient to sustain adequate export of this fat. Only under conditions of rapid triacylglycerol production (i.e., excess calorie intake) would the defect in PEMT become rate-limiting and cause fatty liver. Thus, the rs7946 SNP in PEMT could be necessary but not sufficient to cause fatty liver. If this were the case, then individuals with fatty liver would be more likely to have the PEMT SNP, but not every individual with the PEMT SNP would have fatty liver. Romeo et al. report data from 2349 subjects. We studied only 87 subjects. Romeo included many more African American and Hispanic subjects, and in these two populations a third are wild type, a third heterozygous for the rs7946 SNP, and a third homozygous for the rs7946 SNP. This is very different from the distribution in Caucasians. Our study included mainly Caucasian subjects, so it is most appropriate to compare our results with those observed in Romeo's Caucasian subjects. These results (Table 1) are remarkably similar. To ascertain whether the A allele is over-represented in subjects with fatty liver, we can use the Fisher's Exact Test (7) (Table 2). We conclude that the study of Romeo et al., in a very large population, supports our finding that the rs7946 SNP in PEMT is more common in Caucasians with fatty liver. The Romeo et al. study provides important additional information: 1) Caucasians have a different distribution of this SNP than do Hispanics and African Americans; 2) having this SNP is not sufficient to cause fatty liver, as many individuals with the SNP have normal liver fat. It is likely that this SNP slows export of fat from liver, and results in fatty liver only when triacylglycerol synthesis is accelerated.