In this study, we found that low levels of the three circulating miRNAs (miR-20a, miR-27a, and miR-126) were related to severe NAFLD for all subjects. Specifically, lower circulating miR-20a and miR-27a levels were found in severe NAFLD subjects. Down-regulated circulating miR-20a and miR-27a were associated with severe disease in NAFLD subjects, whereas down-regulated circulating miR-126 was associated with severe disease in male NAFLD subjects.
It is well known that levels of circulating miRNAs in humans vary widely according to several factors such as age, gender, or race29,30,31. Given that circulating miRNAs are affected by several elements, the existence of confounding factors should be considered. That is to say, we need to investigate the levels of circulating miRNAs in the general population. Our logistic regression results show an association between down-regulated circulating miR-20a, miR-27a, and miR-126 levels and disease severity in NAFLD subjects. Although we adjusted for lifestyle risk factors, such as smoking and exercise, circulating miR-20a and miR-27a were significantly associated with NAFLD (Tables 2 and 3). These results suggest that circulating miR-20a and miR-27a levels may represent novel factors for NAFLD that are independent of well-known lifestyle-related factors.
Over the past few years, several studies have been carried out that have focused on circulating miRNAs in NAFLD patients. Specifically, Liu et al. reported that some circulating miRNAs, such as miR-122, are approximately 2-fold higher in NAFLD samples. In addition, Tan et al. have showed that the levels of serum miRNAs, such as miR-122, are altered in NAFLD patients32. In contrast to previous results, the current study considered the general population. To date, no study has attempted to investigate the relationships between circulating miRNAs and NAFLD in general population except for our research. It should be noted that samples from the general population were able to provide clarity regarding lower circulating miRNA levels, indicating that future studies that consider the general population may be more successful when screening for NAFLD biomarkers.
To date, several studies have demonstrated that NAFLD may change circulating miRNA expression14,15,33,34. For example, NAFLD patients show the up-regulated expression of circulating miRNAs, such as miR-21, miR-34a, miR-122, and miR-4517. Another study has also shown that NAFLD increases with circulating miR-34a and miR-192 expression35. Additionally, Salvoza N et al. showed up-regulated miR-34a and miR-122 expression in NAFLD patients36. It should be noted that the current study found lower levels of miRNAs, such as miR-20a, miR-27a, and miR-126. Generally, circulating miRNA expression is normalized by U6 levels or spiked-in cyn-39 levels due to the lack of an established internal standard for circulating miRNA. Thus, results regarding variations in circulating miRNA levels may be affected depending on the selection of internal standards. As such, the ratio of down-regulated to up-regulated miRNA may be more useful in distinguishing between normal and NAFLD patients in the absence of an internal standard. Indeed, Deng Y et al.reported that the ratio-based method can solve the normalization problem for circulating miRNAs37.
FIB-4 is one of biomarkers for liver fibrosis38. In order to elucidate whether the levels of circulating miR-20a, 27a and 126 are not associated with liver fibrosis, we examined the association of these circulating miRNAs levels and FIB-4 score. As a result, there was a weak inverse correlation between circulating miR-126 level and FIB-4 score. There is a report showing that hepatic miR-126 level is decreased in patients with NASH39. On the other hand, circulating miR-20a and 27a levels were not correlated with neither FIB-4 score nor advancing liver fibrosis. To the best of our knowledge, there were no reports showing that decreased circulating miR-20a and 27a levels are associated with liver fibrosis. Therefore, we suggested that decreased circulating miR-20a and 27a levels could reflect intrahepatic steatosis, but not liver fibrosis.
In the present study, the levels of serum miR-20a and miR-27a, which are associated with obesity, MetS, and T2DM14,22,23,24, were found to be significantly lower in NAFLD subjects. It has been reported that NAFLD shows ectopic fat accumulation that is usually accompanied by decreased glycogen synthesis40. Hepatic glycogen synthesis is stimulated by miR-20a41. Down-regulated circulating miR-20a levels were associated with reduced hepatic glycogen synthesis. However, Dan et al. reported that circulating miR-20a was up-regulated in T2DM patients with NAFLD complication42. The reason for the difference between their report and our results is unclear at present. In the present study, NAFLD subjects had not been diagnosed with T2DM, although the NAFLD subjects had significantly higher (although slight) fasting serum glucose and HbA1c levels than normal subjects. Therefore, the differences regarding circulating miR-20a levels in NAFLD patients between their report and our study may be due to differences in T2DM severity.
It has been demonstrated that miR-27a is related to lipid metabolism in the liver43. Zhang M et al. reported that hepatic miR-27a plays a critical role in lipid homeostasis of the liver and that the overexpression of miR-27a attenuated NAFLD development44. In addition, it has been reported that the expression of miR-27a decreases PPARγ expression45. PPARγ regulates lipogenic and adipogenic gene expression and is related to the pathophysiology of obesity. In this study, we indicated that low levels of circulating miR-27a were associated with severe NAFLD. Given the results of previous research, we suggest that the decreased expression of circulating miR-27a may result in increased PPARγ expression in the liver.
In the present study, we found that the low levels of three circulating miRNAs (i.e., miR-20a, miR-27a, and miR-126) were associated with NAFLD. Nevertheless, there are some limitations to our study. NAFLD in the recruited subjects was diagnosed by US, because we recruited subjects from health examination participants and because it is known that US is a reliable and accurate method to be able to detect moderate-severe fatty liver as much as histological examination46. Liver biopsy is the gold standard for NAFLD diagnosis. Therefore, further studies are required to establish the association of the levels of serum miRNAs with the severity of NAFLD diagnosed by liver biopsy. Published studies, including our present study, lack long-term outcome data regarding disease activity, prognostic expectations, and pathogenesis. In addition, we recruited only Japanese subjects in this study, which may have potentially biased our results towards a particular ethnic group. The inclusion of other ethnic groups in future studies may result in different circulating miRNA profiles. Therefore, further studies are required to elucidate whether down-regulated circulating miR-20a and miR-27a may be used as novel biomarkers for severe NAFLD and whether a combination of circulating miR-20a and/or miR-27a with other circulating miRNAs enhances the sensitivity of the disease severity-dependent detection of NAFLD.
In conclusion, the results of the present study indicate that down-regulated circulating miR-20a, miR-27a, and miR-126 are associated with severe NAFLD in the Japanese general population, although the association of down-regulated circulating miR-20a and miR-27a levels with severe NAFLD seem to be superior to that of down-regulated circulating miR-126 levels. We suggest that down-regulated circulating miR-20a and miR-27a may be useful biomarkers for severe NAFLD.