Abstract
Objective:
This study explored the relationship between PNPLA3 rs738409, nutrient intake, and liver histology severity in patients with nonalcoholic fatty liver disease (NAFLD).
Methods:
PNPLA3-rs738409 variant was genotyped in 452 non-Hispanic whites with histologically confirmed NAFLD who completed Food Frequency Questionnaire within 6 months of their liver biopsy. The fibrosis severity on liver histology was the outcome of interest.
Results:
The distribution of PNPLA3 genotypes was CC: 28%, CG: 46% and GG: 25%. High carbohydrate (% of energy/day) intake was positively associated (Adjusted [Adj]. OR: 1.03, P<.01), whereas higher n-3 polyunsaturated fatty acids (n-3 PUFAs) (g/day) (Adj. OR: 0.17, P<.01), isoflavones (mg/day) (Adj. OR: 0.74, P=.049), methionine (mg/day) (Adj. OR: 0.32, P<.01), and choline (Adj. OR: 0.32, P<.01) intakes were inversely associated with increased risk of significant fibrosis (stage of fibrosis ≥2). By using an additive model of inheritance, our moderation analysis showed that PNPLA3 rs738409 significantly modulates the relationship between carbohydrate (%), n-3 PUFAs, isoflavones, methionine, and choline intakes and fibrosis severity in a dose-dependent, genotype manner. These dietary factors tended to have a larger and significant effect on fibrosis severity amongst rs738409 G-allele carriers. Associations between significant fibrosis and carbohydrates (Adj. OR: 1.04, P=.019), n-3 PUFAs (Adj. OR: 0.16, P<.01), isoflavones (Adj. OR: 0.65, P=.025), methionine (Adj. OR: 0.30, P<.01), and choline (Adj. OR: 0.29, P<.01) intakes remained significant only amongst rs738409 G-allele carriers.
Conclusions:
This gene-diet interaction study suggests that PNPLA3 rs738409 G-allele might modulate the effect of specific dietary nutrients on risk of fibrosis in patients with NAFLD.
Keywords: dietary nutrients, carbohydrate, n-3 PUFAs, total isoflavones, methionine, choline
INTRODUCTION
The human patatin-like phospholipase domain-containing 3 gene (PNPLA3) rs738409 C > G (p.I148M) variant, is recognized as the major genetic determinant of nonalcoholic fatty liver disease (NAFLD). Carriers of the G-risk allele present about three time more risk of developing severe NAFLD phenotypes.(1-6) PNPLA3 is a membrane-bound protein that is highly expressed in the liver and adipose tissues. It shows triacylglycerol hydrolase, acyltransferase, and transacylase activities, which promote lipid droplet remodeling in hepatocytes and hepatic stellate cells.(7, 8) In-vitro studies have pointed out PNPLA3 rs738409 variant is associated with reduced fatty acid hydrolysis and impaired mobilization of triglycerides from the hepatic lipid droplets, suggesting a loss of function.(9)
PNPLA3 expression is highly regulated by changes in energy balance and dietary factors. Carbohydrate (CHO) feeding upregulates PNPLA3 expression in human hepatocytes through the carbohydrate response element protein (ChREBP) and the sterol regulatory element-binding protein 1c (SREBP-1c).(10, 11) Recently, it has been proposed that certain unsaturated fatty acids can suppress both lipogenesis and PNPLA3 expression, which may exert beneficial effects on NAFLD severity.(12)
Although several factors may contribute to NAFLD, certain dietary macronutrients have been implicated in the development and progression of NAFLD. Chronic consumption of CHOs, particularly fructose leads to obesity, dyslipidemia, insulin resistance as well as more severe phenotypes of NAFLD (13, 14) In addition to CHOs, deficiency of n-3 polyunsaturated fatty acids (PUFAs) and high n-6/n-3 PUFAs ratio in the diet are associated with the development of NAFLD.(15) Moreover, n-3 PUFAs supplementation improves plasma lipid profile and might reduce liver fat and serum levels of aminotransferase.(16, 17) Most recently, intake of soy isoflavones, which acts as phytoestrogens, has been reported to attenuate NAFLD by the activation of hepatic fatty acid β-oxidation and regulation of hepatic de novo lipogenesis and insulin signaling.(18, 19) Finally, short-term dietary methionine and choline deficiency has been widely used to induce efficiently severe phenotypes of NASH in animal models of NAFLD.(20, 21) However, there are only limited data on the impact of methionine and choline deficiency on the development and progression of NAFLD in humans.
The interplay between genetic and dietary factors might play an important role in the pathogenesis of NAFLD and its progression. The activity of the PNPLA3-148M isoform can be modulated by dietary intake of CHOs, particularly sucrose, and the n-6/n-3 PUFAs ratio in the diet. In-vivo experiments have reported that PNPLA3 148M knock-in mice develop hepatic steatosis when fed a steatogenic high-sucrose diet.(11, 22, 23) Few preliminary human studies suggested a relationship between PNPLA3 rs738409 and foods rich in CHOs or high n-6/n-3 PUFAs ratio on hepatic fat or related phenotypes. (24-26)
Taken together, we hypothesized that there might be a significant relationship between PNPLA3 rs738409 and nutrient intake and the risk for advanced liver histology in individuals with NAFLD. Thus, we first aimed to determine the association between different dietary nutrient intake and the severity of fibrosis among patients with histologically confirmed NAFLD. Next, we tested whether the influence of dietary factors such as carbohydrates, n-3 PUFAs, total isoflavones, methionine, and total choline on fibrosis severity might be modulated by PNPLA3 rs738409.
METHODS
Study design and participants
This analysis includes patients > 18 years of age who were prospectively enrolled in two studies of the NASH Clinical Research Network (NASH CRN), including 1 observational study of NAFLD (NAFLD Database 1, recruiting patients from 2004 to 2009) and 1 randomized controlled trial (PIVENS [NCT00063622]).(27, 28) The institutional review board at each participating center approved these studies, and written consent was obtained from each participant. The diagnosis of NAFLD was based on ≥5% of hepatocytes containing macrovesicular fat and exclusion of significant alcohol consumption (>20 g/d for women and >30 g/d for men) within two years of the initial biopsy, and other causes of chronic liver disease based on clinical history, laboratory studies, and histology. For this study, only non-Hispanic white (Caucasians) individuals who have completed the food questionnaire within six months of their liver biopsy were included for analyses.
Assessment of dietary intake
The dietary intake within the preceding 6 months of study enrollment was assessed with a 109-item modified version of the Block food frequency questionnaire (Block FFQ, version 98.2, NutritionQuest, Berkeley, CA) (29) based on the National Health and Nutrition Examination Survey.(30) The Block 98 FFQ is a self-administered questionnaire that incorporates dietary changes suggested by American national consumption data collected from the third National Health and Nutrition Examination Survey (NHANES III).(29) The intake of energy, macronutrients, and micronutrients were analyzed by the NutritionQuest software. The following variables were analyzed in this study: total energy (kcal) (including energy from fat, CHOs, protein, alcohol and fiber), CHO (% of energy and g/d), fat (% of energy and g/d), protein (% of energy and g/d), saturated fatty acid (SFA) (g/d), monounsaturated fatty acid (MFA) (g/d), polyunsaturated fatty acid (PUFA) (g/d), n-3 PUFAs, trans-fatty acid (g/d), methionine (mg/d), choline (g/d), glycemic index, glycemic load, total isoflavones (mg/d), fiber (g/d), quercetin (mg/d), beta-carotene (ug/d), retinol (ug/d), lycopene (ug/d), cryptoxanthin (ug/d) and betaine (mg/d).
Clinical, laboratory and histological assessment
Participants’ demographics, anthropomorphic measurements, and laboratory tests were collected at each clinical study site as previously described.(28) Liver biopsy specimens were evaluated centrally by the NASH CRN pathology committee using the NASH CRN histologic scoring system (31) and scored as follows: NAFLD activity score (range, 0-8) and its individual components (steatosis [range, 0–3], lobular inflammation [range, 0–3], and hepatocellular ballooning [range, 0–2], and fibrosis (range, 0-4). The presence of significant fibrosis (stage of fibrosis ≥2) or overall fibrosis severity (from stage 0 to 4) were considered as the primary histological phenotypes of interest. We also considered the severity of steatosis, lobular inflammation, and hepatocyte ballooning degeneration as secondary study outcomes.
PNPLA3 rs738409 genotyping
DNA samples were received from NASH CRN consortium and were used to genotype PNPLA3 rs738409 variant by a TaqMan genotyping assay (Applied Biosystems, Foster City, CA USA) according to manufacturer’s instructions. The rs738409 in PNPLA3 gene is a missense coding variant (NM_025225.3: c.444C>G), and results in an Ile [ATC] to Met [ATG] substitution. TaqMan probes were used for allelic discrimination (ThermoFisher, Waltham, MA). The genotype of each sample is determined by the fluorescence levels of the reporter dyes, and samples of the same genotype cluster together when plotted. Each reaction contains 5 μl 2X TaqMan Universal PCR Mastermix, No AmpErase UNG, 3.875 μl water, 0.125 μl 80X Assay Mix, and 1 μl sample DNA. Eight controls were included on each 96-well plate: 2 no template controls, 2 replicate heterozygous samples and 2 replicates of each of the homozygous samples. Because genotyping is determined by endpoint reading, the PCR was carried out in standard Applied Biosystems thermocyclers (AB2720, SimpliAmp, and Veriti, ThermoFisher). The PCR products were analyzed in an ABI PRISM® 7300 Sequence Detection System (SDS) instrument (Thermofisher). SDS Software 1.3.1 was used to convert the raw data to pure dye components and to plot the results of the allelic discrimination on a scatter plot of allele C versus allele G. Genotypes were successfully determined for 100% of the participants.
Statistical analysis
Summary results are presented as median (interquartile range [IQR]), or number (percent) of patients. The X2 test was applied for categorical variables. Continuous variables were compared among groups by t test or Mann-Whitney U test, depending on the distribution. Nutrient distributions were tested for normality with the Shapiro–Wilk test; most of the distributions were skewed, thus all of them were log-transformed using the natural logarithm except for CHOs (% of energy). Non-transformed values are displayed in tables for easier interpretation.
Univariate and multivariable binary or ordinal logistic regression models were implemented to estimate odd ratios (ORs), and 95% confidence intervals (CIs) of those variables associated with histological outcomes. All multivariable analyses were adjusted for age, gender, type 2 diabetes mellitus (T2DM), body mass index (kg/m2) and total energy intake (kcal/d).
We further examined whether the effect of specific dietary factors was different across rs738409 genotypes. To check the presence of this effect, we stratified the study population under an additive inheritance model of the rs738409 variant (CC vs GC vs GG). If there was a significant and dose-dependent association between rs738409 GC and GG genotypes and fibrosis severity, moderation effects of rs738409 G-allele were also examined under a dominant genetic model (GG+GC vs CC).
To examine the conditional effect of dietary factors on fibrosis severity via rs738409 genotypes, we performed a moderation analysis.(32) To do so, we tested a moderation model including dietary factors as exposure measures, significant or overall fibrosis as outcomes and rs738409 genotypes as moderators. The conceptual moderation model is shown in Figure 1. The conditional or moderating effect expresses the fraction of the exposure effect (dietary factor) that is modulated through a specific moderator.(32) This model examines whether effects of the exposure (dietary factor) might be significantly different from each other at different levels of the potential moderator (rs738409 genotypes).
Figure 1. Moderation model diagram.
Abbreviations: PNPLA3, patatin-like phospholipase domain-containing protein 3; F, fibrosis.
Dashed lines represent the effect of dietary factors on risk of NAFLD severity at different levels of moderators (rs738409 genotypes). Moderators can strength, weaken, or reverse the nature of a relationship.
Using SPSS (Version 26; Chicago, IL, USA) macro PROCESS (version 3.5) model 1, dietary factors were included as predictors, rs738409 as moderator, and liver histology severity as outcomes.(33) Model 1 assumes a simple moderation model involving one moderator. Significant conditional effects (β coefficients) were determined using 95% bootstrap bias corrected confidence intervals (CIs) based on 10,000 bootstrap samples. If the upper and lower bounds of the 95% bias-corrected CIs do not contain zero, the conditional effect is considered significant, meaning that that effect of exposure on outcome is at least partially contingent on change in the moderator.(32) β coefficient weights (conditional or moderation effect) provide an index of the magnitude of the effect size.(33) All our moderation analyses were adjusted for gender, age,T2DM, BMI (kg/m2) and calorie intake (kcal).
As recommended by Preacher and Hayes,(34) we used a bootstrapping method as it is considered the most powerful, most effective method to use with small samples, and the least vulnerable to Type I errors. In addition, bootstrapping does not assume normal distributions for any variable, and it is also a nonparametric resampling procedure. We resampled the data 10,000 times as recommended by Hayes.(32) Bootstrapping is the most powerful and appropriate method to obtain confidence limits for specific conditional effects under most conditions.(34)
No missing observations were found for those variables included in our main statistical analyses (liver histology scores, PNPLA3 rs738409, age, gender, BMI (kg/m2) and T2DM and dietary components of interest).
All the statistical analyses were performed using statistical packages Stata, release 16 (StataCorp, College Station, TX) and SPSS, version 26 (Chicago, IL, USA). P values less than 0.05 were considered statistically significant.
RESULTS
Baseline characteristics
A total of 452 non-Hispanic whites with biopsy-proven NAFLD and information on their dietary history were included. Two hundred and nineteen (48%) of 452 patients had significant fibrosis on liver histology, 287 (64%) were women, and the median age was 51.1years (range 43.3–57.6). The prevalence of T2DM and hypertension was 24% and 72% respectively. PNPLA3 rs738409 genotypes distribution was as follows: CC, n=128 (28%); GC, n=210 (46%) and GG, n=114 (26%). The genotype distribution did not deviate from Hardy–Weinberg equilibrium (P= .14). Their selected clinical characteristics and dietary intakes are shown in Table 1.
Table 1.
Selected clinical and nutrient intake characteristics of our study population
| Variables | N=452 |
|---|---|
| Age (years) | 51.1 (43.3-57.6) |
| Gender (female) | 287 (64) |
| Body mass index (kg/m2) | 34.25 (30.11-38.53) |
| Type 2 diabetes, n (%) | 107 (24) |
| Hypertension, n (%) | 324 (72) |
| Non-heavy drinkers, n (%) | 156 (35) |
| ALT (U/L) | 59 (40-89) |
| AST (U/L) | 44 (31-63) |
| ALP (U/L) | 80 (66-103) |
| PNPLA3 rs738409G-allele | 324 (72) |
| Histopathological report | |
| Steatosis, n (%) | |
| <5% * | 10 (2.2) |
| 5-33% | 176 (38.9) |
| 33-66% | 159 (35.2) |
| >66% | 107 (23.7) |
| Lobular inflammation, n (%) | |
| No foci | 1 (0.3) |
| <2 foci/200x | 233 (51.5) |
| 2-4 foci/200x | 169 (37.4) |
| >4 foci/200x | 49 (10.8) |
| Ballooning, n (%) | |
| None | 148 (32.7) |
| Few | 119 (26.3) |
| Many | 185 (41) |
| Portal inflammation, n (%) | |
| None | 58 (12.8) |
| Mild | 291 (64.4) |
| > Mild | 103 (22.8) |
| Fibrosis stages, n (%) | |
| 0 | 99 (21.9) |
| 1 | 134 (29.6) |
| 2 | 90 (19.9) |
| 3 | 80 (17.7) |
| 4 | 49 (10.8) |
| Significant fibrosis (F≥2), n (%) | 219 (48) |
| Dietary factors | |
| Total energy (kcal/d) | 1706 (1217-2296) |
| Fat (%) | 38.37 (33.99-43.39) |
| Protein (%) | 15.5 (13.42-17.62) |
| Carbohydrate (%) | 47.49 (41.08-52.75) |
| Saturated fatty acids (g/d) | 21.2 (14-29.25) |
| Monounsaturated fatty acids (g/d) | 26.95 (18.25-38.05) |
| Polyunsaturated fatty acids (g/d) | 17.4 (11.55-23.85) |
| Omega-3 PUFAs (g/day) | 1.57 (1.06-2.37) |
| Trans-Fat (g/day) | 5.41 (3.36-8.21) |
| Methionine (mg/d) | 1425 (980-1965) |
| Total choline (mg/d) | 279.3 (201.1-387.95) |
| Total isoflavones (mg/d) | 1.17 (0.8-1.79) |
| Quercetin (mg/d) | 7.67 (4.02-13.54) |
| Fiber (g/d) | 15.2 (10.8-20.35) |
| Glycemic index | 54.28 (51.53-56.42) |
| Glycemic load | 100.18 (68.73-135.86) |
| Beta-carotene (ug/d) | 2384 (1435-4258) |
| Retinol (ug/d) | 395.27 (234.88-606.19) |
| Lycopene (ug/d) | 4346 (2422-7673) |
| Cryptoxanthin (ug/d) | 51.11 (22.73-125.80) |
| Betaine (mg/d) | 153.95 (97.25-224.7) |
Abbreviations: PNPLA3, patatin-like phospholipase domain-containing 3; ALT, alanine aminotransferase; AST, aspartate aminotransferase; ALP, alkaline phosphatase.
These patients had a diagnosis of borderline or definite steatohepatitis and all of them had a stage of fibrosis ≥2 (5 had cirrhosis, 4 bridging fibrosis and 1 pericellular and portal/periportal fibrosis).
Continuous variables are expressed as median and interquartile range.
Association between nutrient intake and fibrosis severity
Univariate associations of dietary macronutrients and significant fibrosis are detailed in Table 2. Carbohydrate (% of energy/day) intake was significantly higher in patients with significant fibrosis (48.25, IQR: 42.1-53.95) than those with none or mild fibrosis (46.66, IQR: 40.48-51.77) (P= .047). n-3 PUFAs (g/day) (1.41, IQR: 0.98-2.24 vs 1.75, IQR: 1.19-2.45; P= .007), total isoflavones (mg/day) (1.1, IQR: 0.75-1.67 vs 1.25, IQR: 0.84-1.86; P= .032), and methionine (mg/day) (1280, IQR: 900-1940 vs 1470, IQR: 1050-2000, P= .044) intake levels were lower in patients with significant fibrosis compared to those without or mild fibrosis. There was a trend towards lower levels of total choline intake in those with significant fibrosis (272.7, IQR: 186.9-379.1) compared with those with none or mild fibrosis (291.7, IQR: 222.7-391) (P= .117).
Table 2.
Association between selected baseline features and significant fibrosis (F≥2)
| Variables | Presence of significant fibrosis (F≥2) |
||
|---|---|---|---|
| No N=233 |
Yes N=219 |
P value* |
|
| Age (years) | 48.6 (40.4-55) | 53.4 (46.1-60.1) | <.001 |
| Gender (female) | 138 (59) | 149 (68) | .052 |
| Non-heavy drinkers, n (%) | 95 (41) | 61 (28) | .004 |
| Type 2 diabetes mellitus (yes), n (%) | 29 (12) | 78 (36) | <.001 |
| Body mass index (kg/m2) | 34.04 (29.76-37.92) | 34.51 (30.48-38.98) | .271 |
| PNPLA3 rs738409, n (%) | .031 | ||
| CC | 76 (32.6) | 52 (23.7) | |
| CG | 105 (45.1) | 105 (47.9) | |
| GG | 52 (22.3) | 62 (28.4) | |
| Lab tests | |||
| ALT (U/L) | 60 (41-85) | 59 (40-92) | .822 |
| AST (U/L) | 39 (28-55) | 50 (35-75) | <.001 |
| Dietary features | |||
| Energy (kcal/d) | 1743 (1312-2257) | 1622 (1080-2317) | .559 |
| Fat (% of energy) | 38.7 (34.5-43.58) | 37.89 (33.47-43.05) | .197 |
| Protein (% of energy) | 15.51 (13.56-17.7) | 15.48 (13.12-17.44) | .297 |
| Carbohydrate (% of energy) | 46.66 (40.48-51.77) | 48.25 (42.1-53.95) | .047 |
| Saturated fatty acids (g/d) | 22.5 (15.1-30.3) | 19.6 (12.6-28.2) | .341 |
| Fiber (g/d) | 15.4 (11.3-20) | 14.9 (10.1-20.7) | .469 |
| Glycemic index | 54.28 (51.96-56.29) | 54.28 (51.1-56.48) | .815 |
| Glycemic load | 102.87 (71.36-133.57) | 96.19 (63.98-139.73) | .886 |
| Monounsaturated fatty acids (g/d) | 28.7 (20.1-37.5) | 24.8 (16.4-38.7) | .657 |
| Trans-Fat (g/d) | 5.68 (3.69-7.78) | 5.27 (3.19-8.58) | .485 |
| n-3 PUFAs (g/d) | 1.75 (1.19-2.45) | 1.41 (0.98-2.24) | .007 |
| Methionine (mg/d) | 1470 (1050-2000) | 1280 (900-1940) | .044 |
| Total choline (mg/d) | 291.7 (222.7-391) | 272.7 (186.9-379.1) | .117 |
| Total isoflavones (mg/d) | 1.25 (0.84-1.86) | 1.1 (0.75-1.67) | .032 |
| Quercetin (mg/d) | 8.23 (4.23-14.31) | 7.02 (3.92-12.92) | .302 |
| Beta-carotene (ug/d) | 2398 (1440-4387) | 2374 (1425-4212) | .548 |
| Retinol (ug/d) | 391.5 (238.5-606.3) | 396.12 (232.61-605.7) | .974 |
| Lycopene (ug/d) | 4475 (2641-7674) | 4255 (2363-7671) | .867 |
| Cryptoxanthin (ug/d) | 54.99 (22.23-126.58) | 46.75 (23.3-123.79) | .815 |
| Betaine (mg/d) | 154.3 (103.3-208.2) | 153.5 (92.9-242) | .560 |
Abbreviations: PNPLA3, patatin-like phospholipase domain-containing 3; PUFAs, polyunsaturated fatty acids.
Chi-square tests for categorical variables. T test or Mann-Whitney U test for continuous variables.
All dietary variables were log-transformed except carbohydrate (% of energy) which was normally distributed.
Continuous variables are expressed as median and interquartile range.
Significant associations between dietary factors and significant fibrosis persisted in the fully adjusted models that included controlling for total calorie intake (kcal/day), age, gender, BMI (kg/m2) and T2DM (Table 3). The risk was elevated nearly 1.67 (95% CI: 1.001-2.95), and 1.94 (1.09-3.44) folds for the 2 highest % carbohydrates quartiles. The risk of significant fibrosis was lower in the 2 highest quartiles of n-3 PUFAs (Adj. ORs: 0.35 [95% CI: 0.19-0.67] and 0.31 [95% CI: 0.14-0.71]) and total isoflavones (Adj. ORs: 0.55 [95% CI: 0.30-0.99] and 0.63 [95% CI: 0.31-0.99]) intake. Likewise, the risk of significant fibrosis was significantly lower in the 3 highest quartiles of methionine (Adj. ORs: 0.46 [95% CI: 0.26-0.84], 0.33 [95% CI: 0.17-0.64] and 0.23 [95% CI: 0.09-0.57]), and total choline (Adj. ORs: 0.33 [95% CI: 0.18-0.60], 0.39 [95% CI: 0.20-0.76] and 0.26 [95% CI: 0.11-0.63]) intake. Table 3 depicts adjusted risks of significant fibrosis across quartiles of carbohydrates (% of energy/day), n-3 PUFAs, total isoflavones, methionine, and total choline intake.
Table 3.
Associations between selected dietary factors and significant fibrosis (F≥2). Results based on univariate and multivariable analysis.
| Variables | Presence of significant fibrosis (F≥2) |
|||
|---|---|---|---|---|
| No N=233 |
Yes N=219 |
Adj. Odds Ratio (95% CI) |
P value* | |
| Carbohydrate (% of energy) | 46.66 (40.48-51.77) | 48.25 (42.1-53.95) | 1.03 (1.01-1.06) | <.001 |
| Quartiles, n (%) | ||||
| ≤ 41.07, n=113 | 63 (56) | 50 (44) | Reference | - |
| 41.08-47.49, n=113 | 64 (57) | 49 (43) | 0.98 (0.56-1.72) | .945 |
| 47.50-52.78, n=113 | 55 (49) | 58 (51) | 1.67 (1.000-2.95) | .048 |
| > 52.78, n=113 | 51 (45) | 62 (55) | 1.94 (1.09-3.44) | .024 |
| n-3 PUFAs (g/d) | 1.75 (1.19-2.45) | 1.41 (0.98-2.24) | 0.17 (0.06-0.47) | .001 |
| Quartiles, n (%) | ||||
| ≤ 1.052, n=113 | 50 (44) | 63 (56) | Reference | - |
| 1.053-1.575, n=113 | 48 (42) | 65 (58) | 0.88 (0.49-;1.57) | .676 |
| 1.576-2.372, n=113 | 72 (64) | 41 (36) | 0.35 (0.19-0.67) | .001 |
| >2.372, n=113 | 63 (56) | 50 (44) | 0.31 (0.14-0.71) | .006 |
| Methionine (mg/d) | 1470 (1050-2000) | 1280 (900-1940) | 0.32 (0.15-0.64) | .001 |
| Quartiles, n (%) | ||||
| ≤ 980, n=112 | 67 (40) | 45 (60) | Reference | - |
| 981-1425, n=114 | 55 (52) | 59 (48) | 0.46 (0.26-0.84) | .012 |
| 1426-1967, n=113 | 47 (58) | 66 (42) | 0.33 (0.17-0.64) | .001 |
| > 1967, n=113 | 50 (56) | 63 (44) | 0.23 (0.09-0.57) | .001 |
| Total isoflavones (mg/d) | 1.25 (0.84-1.86) | 1.1 (0.75-1.67) | 0.74 (0.53-0.99) | .049 |
| Quartiles, n (%) | ||||
| ≤ 0.80, n=114 | 47 (41) | 67 (59) | Reference | - |
| 0.81-1.17, n=113 | 60 (53) | 53 (47) | 0.71 (0.40-1.25) | .245 |
| 1.18-1.79, n=113 | 63 (56) | 50 (44) | 0.55 (0.30-0.99) | .050 |
| > 1.79, n=112 | 63 (56) | 49 (44) | 0.63 (0.31-0.99) | .049 |
| Total choline (mg/d) | 291.7 (222.7-391) | 272.7 (186.9-379.1) | 0.32 (0.15-0.69) | .003 |
| Quartiles, n (%) | ||||
| ≤ 210, n=113 | 45 (19.3) | 68 (31.1) | Reference | - |
| 211-279, n=113 | 65 (27.9) | 48 (21.9) | 0.33 (0.18-0.60) | <.001 |
| 280-386, n=113 | 62 (26.6) | 51 (23.3) | 0.39 (0.20-0.76) | .005 |
| >386, n=113 | 61 (26.2) | 52 (23.7) | 0.26 (0.11-0.63) | .003 |
Abbreviations: PUFAs, polyunsaturated fatty acids.
Multivariable binary logistic regression. All multivariable analyses were adjusted for age, gender, type 2 diabetes, body mass index (kg/m2) and calorie intake (kcal).
Dietary variables were log-transformed except carbohydrate (% of energy) which was normally distributed.
Carbohydrates (% of energy): PCT 25th ≤ 41.07; PCT 25-50th 41.08-47.49; PCT 50-75th 47.50-52.78; PCT >75th >52.78.
n-3 PUFAs (g/d): PCT 25th ≤ 1.052; PCT 25-50th 1.053-1.575; PCT 50-75th 1.576-2.372; PCT >75th >2.372.
Total isoflavones (mg/d): PCT 25th ≤0.80; PCT 25-50th 0.81-1.17; PCT 50-75th 1.18-1.79; PCT >75th >1.79.
Methionine (mg/d): PCT 25th ≤980; PCT 25-50th 981-1425; PCT 50-75th 1426-1967; PCT >75th >1967.
Total choline (mg/d): PCT 25th ≤210; PCT 25-50th 211-279; PCT 50-75th 280-386; PCT >75th >386.
Continuous variables are expressed as median and interquartile range.
Supplemental Table 1 depicts associations between carbohydrate (% of energy), n-3 PUFAs (g/d), methionine (mg/d), total isoflavones (mg/d) and total choline (mg/d) intake and overall fibrosis severity (from stage 0 to 4). Interestingly associations between these specific dietary factors and overall fibrosis remained statistically significant and in the same direction even after controlling for the same potential confounders.
Neither the intake of protein, and fat, including SFA, MAF, trans-fatty acids and overall PUFAs, nor that of quercetin, fiber, glycemic index or glycemic load, betaine, retinol, lycopene, cryptoxanthin, and beta-carotene was associated with risk for significant fibrosis (Table 2).
Finally, we did not find significant associations between our selected dietary factors and other NAFLD-related histological features including steatosis, lobular inflammation, and hepatocyte ballooning degeneration (supplemental Tables 2-6).
PNPLA3 rs738409 as a modifier of the relationship between dietary factors and fibrosis severity.
Results based on an additive genetic model (CC vs GC vs GG)
We performed single moderation analysis with model 1 (Process macro) to examine the effect of PNPLA3 rs738409 on the relationship between each dietary factor and fibrosis severity under an additive (Table 4) genetic model. Each moderation analysis was adjusted for age, gender, T2DM, BMI (kg/m2) and calorie intake (kcal).
Table 4.
Role of PNPLA3 rs738409 as a moderator between dietary factors and fibrosis severity. Conditional effects of the dietary factors at values of the moderator using an additive genetic model.
| Bootstrapped estimates (n=10,000) | Bootstrapped estimates (n=10,000) | |||||||
|---|---|---|---|---|---|---|---|---|
| Dietary factors | Unadjusted β coefficients (conditional effects) † |
Standard errors |
95% confidence intervals |
P value | Adjusted β coefficients (conditional effects) † |
Standard errors |
95% confidence intervals |
P value |
| Significant fibrosis (F≥2) | ||||||||
| Carbohydrate (% of energy) | ||||||||
| PNPLA3-CC | 0.004 | 0.019 | −0.032 to 0.040 | .820 | 0.021 | 0.020 | −0.019 to 0.061 | .302 |
| PNPLA3-GC | 0.020 | 0.011 | −0.002 to 0.042 | .075 | 0.032 | 0.013 | 0.008 to 0.057 | .010 |
| PNPLA3-GG | 0.036 | 0.019 | −0.002 to 0.073 | .062 | 0.044 | 0.021 | 0.003 to 0.084 | .035 |
| n-3 PUFAs (g/day) ‡ | ||||||||
| PNPLA3-CC | −0.092 | 0.453 | −0.980 to 0.797 | .840 | −1.332 | 0.645 | −2.595 to −0.068 | .039 |
| PNPLA3-GC | −0.513 | 0.271 | −1.044 to 0.019 | .059 | −1.728 | 0.510 | −2.727 to −0.729 | .001 |
| PNPLA3-GG | −0.934 | 0.465 | −1.845 to −0.023 | .045 | −2.124 | 0.644 | −3.386 to −0.862 | .001 |
| Total isoflavones (mg/day) ‡ | ||||||||
| PNPLA3-CC | −0.021 | 0.253 | −0.517 to 0.476 | .935 | 0.044 | 0.284 | −0.512 to 0.600 | .877 |
| PNPLA3-GC | −0.299 | 0.134 | −0.561 to −0.038 | .025 | −0.301 | 0.165 | −0.625 to −0.022 | .048 |
| PNPLA3-GG | −0.578 | 0.254 | −1.076 to −0.081 | .023 | −0.647 | 0.286 | −1.207 to −0.086 | .024 |
| Methionine (mg/day) ‡ | ||||||||
| PNPLA3-CC | −0.019 | 0.323 | −0.653 to 0.615 | .953 | −0.886 | 0.470 | −1.806 to 0.034 | .059 |
| PNPLA3-GC | −0.316 | 0.185 | −0.679 to 0.046 | .087 | −1.104 | 0.364 | −1.817 to −0.391 | .002 |
| PNPLA3-GG | −0.614 | 0.325 | −1.251 to 0.023 | .059 | −1.322 | 0.457 | −2.217 to −0.417 | .004 |
| Total choline (mg/d) ‡ | ||||||||
| PNPLA3-CC | 0.029 | 0.334 | −0.625 to 0.683 | .931 | −0.869 | 0.483 | −1.815 to 0.078 | .072 |
| PNPLA3-GC | −0.266 | 0.199 | −0.657 to 0.125 | .182 | −1.113 | 0.388 | −1.873 to −0.352 | .004 |
| PNPLA3-GG | −0.561 | 0.348 | −1.243 to 0.121 | .107 | −1.357 | 0.491 | −2.320 to −0.394 | .006 |
| Overall fibrosis severity (stage 0-4) | ||||||||
| Carbohydrate (% of energy) | ||||||||
| PNPLA3-CC | 0.014 | 0.012 | −0.009 to 0.037 | .240 | 0.023 | 0.011 | 0.001 to 0.044 | .040 |
| PNPLA3-GC | 0.020 | 0.007 | 0.006 to 0.034 | .004 | 0.026 | 0.007 | 0.013 to 0.039 | <.001 |
| PNPLA3-GG | 0.027 | 0.012 | 0.003 to 0.050 | .025 | 0.029 | 0.011 | 0.007 to 0.051 | .009 |
| n-3 PUFAs (g/day) ‡ | ||||||||
| PNPLA3-CC | −0.133 | 0.286 | −0.695 to 0.430 | .643 | −0.803 | 0.342 | −1.475 to −0.130 | .019 |
| PNPLA3-GC | −0.406 | 0.170 | −0.741 to −0.071 | .018 | −1.060 | 0.270 | −1.590 to −0.529 | <.001 |
| PNPLA3-GG | −0.679 | 0.288 | −1.246 to −0.112 | .019 | −1.317 | 0.348 | −2.002 to −0.633 | <.001 |
| Total isoflavones (mg/day) ‡ | ||||||||
| PNPLA3-CC | 0.011 | 0.157 | −0.299 to 0.320 | .946 | 0.057 | 0.154 | −0.246 to 0.360 | .713 |
| PNPLA3-GC | −0.213 | 0.081 | −0.372 to −0.054 | .009 | −0.187 | 0.088 | −0.360 to −0.013 | .035 |
| PNPLA3-GG | −0.437 | 0.155 | −0.741 to −0.132 | .005 | −0.430 | 0.154 | −0.733 to −0.128 | .005 |
| Methionine (mg/day) ‡ | ||||||||
| PNPLA3-CC | −0.026 | 0.204 | −0.427 to 0.375 | .899 | −0.547 | 0.253 | −1.045 to −0.049 | .031 |
| PNPLA3-GC | −0.270 | 0.116 | −0.499 to −0.042 | .021 | −0.737 | 0.195 | −1.121 to −0.353 | <.001 |
| PNPLA3-GG | −0.514 | 0.201 | −0.909 to −0.119 | .011 | −0.927 | 0.245 | −1.409 to −0.444 | <.001 |
| Total choline (mg/d) ‡ | ||||||||
| PNPLA3-CC | −0.015 | 0.212 | −0.431 to 0.401 | .943 | −0.561 | 0.260 | −1.072 to −0.049 | .032 |
| PNPLA3-GC | −0.248 | 0.126 | −0.496 to 0.000 | .050 | −0.759 | 0.207 | −1.167 to −0.352 | <.001 |
| PNPLA3-GG | −0.481 | 0.217 | −0.909 to −0.054 | .027 | −0.957 | 0.265 | −1.479 to −0.436 | <.001 |
Abbreviations: PNPLA3, patatin-like phospholipase domain-containing 3; PUFAs, polyunsaturated fatty acids.
Conditional or moderation effects were calculated while adjusting for age, gender, type 2 diabetes, BMI (kg/m2) and calorie intake (kcal).
Log-transformed variables.
Overall, these analyses showed that PNPLA3 rs738409 significantly modulates the relationship between carbohydrate (%), n-3 PUFAs, total isoflavones, methionine, and total choline intakes and fibrosis severity in a dose-dependent, genotype manner. These dietary factors tended to have a larger and significant effect on significant or overall fibrosis in those individuals carrying rs738409 G-allele.
The conditional effect of carbohydrates (%) intake on significant fibrosis was in a positive direction and reached its highest level in patients with GG (β = 0.044, 95% CI: 0.003 to 0.084) and GC (β = 0.032, 95% CI: 0.008 to 0.057) genotypes. In contrast, the effect of carbohydrates (%) consumption was smaller and not significant (β = 0.021, 95% CI: −0.019 to 0.061) for CC genotype.
The effects of other dietary factors including n-3 PUFAs, total isoflavones, methionine and total choline on significant fibrosis were also moderated via the rs738409 variant in a dose-dependent, allele manner, but in a negative direction. Overall, the conditional effects of the above-mentioned dietary factors reached its highest level in GG and GC genotypes. For instance, the conditional effect of n-3 PUFAs intake on significant fibrosis via rs738409 was significantly higher at GG (β = −2.124, 95% CI: −3.386 to −0.862) and GC genotypes (β = −1.728, 95% CI: −2.727 to −0.729) as compared to CC (β = −1.332, 95% CI: −2.595 to −0.068). Likewise, the protective effect of high consumption of total isoflavones (β = −0.647, 95% CI: −1.207 to −0.086 and β = −0.301, 95% CI: −0.625 to −0.022), methionine (β = −1.322, 95% CI: −2.217 to −0.417 and β = −1.104, 95% CI: −1.817 to −0.391), and total choline (β = −1.357, 95% CI: −2.320 to −0.394 and β = −1.113, 95% CI: −1.873 to −0.352) on significant fibrosis was significantly higher at GG and GC genotypes, respectively. Interestingly, the protective effect of these dietary factors on risk of significant fibrosis was negligible or null amongst rs738409 CC carriers.
Further moderation analysis, including overall fibrosis as a co-primary study outcome showed larger effect sizes associated with carbohydrate (%), n-3 PUFAs, total isoflavones, methionine, and total choline intakes in individuals with rs738409 GC and GG than those with CC genotypes. These findings indicate that changes in levels of consumption of certain dietary factors might predict greater change in the risk of fibrosis severity among individuals carrying the rs738409 G-allele.
Finally, we did not find significant moderation effects of rs738409 on the relationship between other dietary factors and significant or overall fibrosis (supplemental Tables 7 and 8).
Results based on a dominant genetic model (GC+GG vs CC)
Since moderation analysis under an additive model showed larger effect sizes of dietary factors on fibrosis severity among individuals with rs738409 GC and GG genotypes, we sought to estimate the effect of dietary nutrients on risk of fibrosis in carriers of G-allele via moderation and logistic regression analysis while controlling for potential confounders.
We present the results of the moderation analysis under a dominant genetic model in Table 5. As seen in the table, higher dietary consumptions of carbohydrate (%), n-3 PUFAs, total isoflavones, methionine or total choline had significant effects on risk of fibrosis severity in carriers of rs738409 G-allele. This analysis also confirms a null or weak effect of those dietary factors and fibrosis severity in homozygous CC individuals.
Table 5.
Role of PNPLA3 rs738409 as a moderator between dietary factors and fibrosis severity. Conditional effects of the dietary factors at values of the moderator using a dominant genetic model.
| Bootstrapped estimates (n=10,000) | Bootstrapped estimates (n=10,000) | |||||||
|---|---|---|---|---|---|---|---|---|
| Dietary factors | Unadjusted β coefficients (conditional effects) † |
Standard errors |
95% confidence intervals |
P value | Adjusted β coefficients (conditional effects) † |
Standard errors |
95% confidence intervals |
P value |
| Significant fibrosis (F≥2) | ||||||||
| Carbohydrate (% of energy) | ||||||||
| PNPLA3-CC | 0.009 | 0.021 | −0.033 to 0.050 | .679 | 0.027 | 0.023 | −0.018 to 0.073 | .239 |
| PNPLA3-GC+GG | 0.025 | 0.013 | 0.000 to 0.051 | .054 | 0.034 | 0.014 | 0.006 to 0.062 | .018 |
| n-3 PUFAs (g/day) ‡ | ||||||||
| PNPLA3-CC | −0.030 | 0.520 | −1.049 to 0.989 | .954 | −0.239 | 0.684 | −2.579 to 0.102 | .070 |
| PNPLA3-GC+GG | −0.694 | 0.317 | −1.315 to −0.074 | .028 | −1.844 | 0.540 | −2.902 to −0.786 | .001 |
| Total isoflavones (mg/day) ‡ | ||||||||
| PNPLA3-CC | 0.227 | 0.306 | −0.373 to 0.826 | .458 | 0.313 | 0.350 | −0.373 to 0.999 | .371 |
| PNPLA3-GC+GG | −0.429 | 0.154 | −0.731 to −0.126 | .005 | −0.453 | 0.184 | −0.814 to −0.092 | .014 |
| Methionine (mg/day) ‡ | ||||||||
| PNPLA3-CC | 0.092 | 0.382 | −0.657 to 0.840 | .810 | −0.741 | 0.514 | −1.748 to 0.266 | .149 |
| PNPLA3-GC+GG | −0.448 | 0.211 | −0.862 to −0.034 | .034 | −1.210 | 0.378 | −1.950 to −0.469 | .001 |
| Total choline (mg/d) ‡ | ||||||||
| PNPLA3-CC | 0.122 | 0.385 | −0.634 to 0.877 | .316 | −0.729 | 0.514 | −1.736 to 0.279 | .156 |
| PNPLA3-GC+GG | −0.402 | 0.233 | −0.859 to 0.054 | .084 | −1.231 | 0.413 | −2.040 to −0.422 | .003 |
| Overall fibrosis severity (stages 0-4) | ||||||||
| Carbohydrate (% of energy) | ||||||||
| PNPLA3-CC | 0.018 | 0.013 | −0.008 to 0.044 | .164 | 0.025 | 0.012 | 0.005 to 0.053 | .020 |
| PNPLA3-GC+GG | 0.021 | 0.008 | 0.005 to 0.038 | .011 | 0.029 | 0.008 | 0.010 to 0.041 | .001 |
| n-3 PUFAs (g/day) ‡ | ||||||||
| PNPLA3-CC | −0.105 | 0.328 | −0.750 to 0.539 | .749 | −0.813 | 0.366 | −1.533 to −0.093 | .027 |
| PNPLA3-GC+GG | −0.521 | 0.199 | −0.913 to −0.130 | .009 | −1.176 | 0.293 | −1.752 to −0.600 | <.001 |
| Total isoflavones (mg/day) ‡ | ||||||||
| PNPLA3-CC | 0.097 | 0.192 | −0.281 to 0.475 | .614 | 0.138 | 0.190 | −0.234 to 0.511 | .466 |
| PNPLA3-GC+GG | −0.283 | 0.089 | −0.458 to −0.108 | .002 | −0.253 | 0.095 | −0.440 to −0.067 | .008 |
| Methionine (mg/day) ‡ | ||||||||
| PNPLA3-CC | 0.011 | 0.240 | −0.461 to 0.484 | .962 | −0.523 | 0.277 | −1.067 to 0.022 | .060 |
| PNPLA3-GC+GG | −0.364 | 0.133 | −0.625 to −0.103 | .006 | −0.824 | 0.205 | −1.228 to −0.421 | <.001 |
| Total choline (mg/d) ‡ | ||||||||
| PNPLA3-CC | 0.040 | 0.243 | −0.437 to 0.517 | .869 | −0.518 | 0.278 | −1.064 to 0.028 | .063 |
| PNPLA3-GC+GG | −0.353 | 0.148 | −0.643 to −0.063 | .017 | -−0.864 | 0.224 | −1.304 to −0.425 | <.001 |
Abbreviations: PNPLA3, patatin-like phospholipase domain-containing 3; PUFAs, polyunsaturated fatty acids.
Conditional or moderation effects were calculated while adjusting for age, gender, type 2 diabetes, BMI (kg/m2) and calorie intake (kcal).
Log-transformed variables.
Table 6 shows associations between mean levels of dietary nutrients and risk of significant fibrosis according to rs738409 genotypes (GC+GG vs CC). Higher carbohydrate (% of energy) intake was associated with higher risk of significant fibrosis among individuals carrying the rs738409 G-allele (Adj. OR: 1.04, 95% CI: 1.01-1.07, P= .019). Conversely, higher intakes of n-3 PUFAs (Adj. OR: 0.16, 95% CI: 0.05-0.53 per log 1SD increased, P= .003), total isoflavones (Adj. OR: 0.65, 95% CI: 0.44-0.95 per log 1SD increased, P = .025), methionine (Adj. OR: 0.30, 95% CI: 0.13-0.70 per log 1SD increased, P= .005) and total choline (Adj. OR: 0.29, 95% CI: 0.11-0.73 per log 1SD increased, P = .009) were associated with lower risk of significant fibrosis among individuals carrying the rs738409 G-allele. Importantly, these associations remained significant after controlling for important confounders, including age, gender, BMI (kg/m2), T2DM and total calorie intake.
Table 6.
Associations between significant fibrosis and mean levels of selected dietary factors. The influence of PNPLA3 rs738409 genotypes based on a dominant genetic model
| Dietary factors | PNPLA3 rs738409 genotypes | ||||
|---|---|---|---|---|---|
| CC | |||||
| F0-F1 | F≥2 | P value† |
P value‡ |
||
| Median (IQR) | Median (IQR) | OR (95% CI) ‡ | |||
| Carb (% of energy) | 46.09 (40.12-51.78) | 46.91 (39.12-52.95) | .682 | 1.02 (0.97-2.53) | .383 |
| n-3 PUFAs (g/d) * | 1.71 (1.31-2.44) | 1.67 (1.21-2.73) | .856 | 0.24 (0.04-1.51) | .128 |
| Total isoflavones (mg/d) * | 1.25 (0.83-1.83) | 1.19 (0.84-1.7) | .959 | 1.58 (0.72-3.49) | .255 |
| Methionine (mg/d) * | 1530 (1205-2155) | 1635 (1120-2205) | .959 | 0.49 (0.13-1.92) | .311 |
| Total choline (mg/d) * | 311.85 (235.1-415.8) | 328.15 (238.95-421.4) | .754 | 0.48 (0.12-1.87) | .295 |
| Dietary factors | CG + GG | ||||
| F0-F1 | F≥2 | P value† |
P value‡ |
||
| Median (IQR) | Median (IQR) | OR (95% CI) ‡ | |||
| Carb (% of energy) | 46.96 (40.58-51.76) | 48.44 (42.62-54.49) | .049 | 1.04 (1.01-1.07) | .019 |
| n-3 PUFAs (g/d) * | 1.75 (1.11-2.45) | 1.34 (0.89-2.02) | .005 | 0.16 (0.05-0.53) | .003 |
| Total isoflavones (mg/d) * | 1.24 (0.88-1.92) | 1.01 (0.7-1.58) | .016 | 0.65 (0.44-0.95) | .025 |
| Methionine (mg/d) * | 1440 (1030-1990) | 1240 (830-1900) | .044 | 0.30 (0.13-0.70) | .005 |
| Total choline (mg/d) * | 285.9 (210.2-381) | 255.9 (178.9-372.3) | .084 | 0.29 (0.11-0.73) | .009 |
Abbreviations: SD, standard deviation; PUFAs, polyunsaturated fatty acids; IQR, interquartile range; OR, odd ratio.
P values correspond with log-transformed values.
T test or Mann–Whitney U test when appropriate.
Simple logistic regression. Associations between dietary factors and significant fibrosis were adjusted for calorie intake (kcal), age, gender, BMI (kg/m2), and type 2 diabetes.
ORs and their CIs are given per log 1 SD increased.
DISCUSSION
In our biopsy-proven NAFLD cohort of non-Hispanic whites, higher % carbohydrates intake was associated with an increased severity of fibrosis. Our results are in alignment with previous studies reporting that diets rich in carbohydrates not only increase insulin resistance and liver fat content but also are associated with fibrosis severity.(13, 14) Our data also show that high n-3 PUFAs, total isoflavones, methionine and total choline intakes were associated with a reduced fibrosis severity, and these associations remained significant after controlling for well-known confounders.
Low dietary consumption of n-3 PUFAs have been associated with exacerbated risk of NAFLD (16, 17), and clinical trials using n-3 PUFAs supplementation have shown certain beneficial effects on triglycerides and aminotransferase levels as well as fatty liver content and inflammation.(35, 36) However, a few prospective trials have failed to demonstrate an association between n-3 PUFAs administration and fibrosis improvement. (37-39)
The influence of isoflavones intake on NAFLD features in humans is unknown. Experimental animal or in vitro studies have documented that isoflavones supplementation might have positive effects on lipid metabolism, insulin resistance, inflammation, and oxidative stress, the most important pathophysiological pathways in NAFLD.(19) Flavonoids have been found to upregulate PPARγ (40) and downregulate SREBP-1c (41) gene expressions. In addition, flavonoids might reduce the production of radicals and other reactive species of oxygen as well as inhibit lipoxygenases and cyclooxygenases activity and block NF-κB pathway.(42)
Short-term dietary methionine and choline deficiency has been widely used to induce efficiently severe phenotypes of NASH in animal models of NAFLD.(20) However, there are only limited data on the impact of methionine and choline deficiency on the development and progression of NAFLD in humans. Both methionine and choline might play key roles in hepatic mitochondrial β-oxidation and very low-density lipoprotein (VLDL) synthesis.(21)
To date, little is known whether the influence of dietary factors on fibrosis severity might be modulated by PNPLA3 rs738409. Our data show that individuals with PNPLA3 rs738409 have altered susceptibility to high carbohydrates consumption; higher CHOs (%) intake, particularly among G-allele carriers was associated with higher risk of fibrosis severity. This result is in line with previous reports showing that the effects of the PNPLA3 rs738409 variant on triglycerides levels and hepatic fat accumulation can be modulated by high CHOs intake, but interestingly this effect is only observed among G-allele carriers.(24, 26) Experimental studies have consistently reported that PNPLA3 gene expression is strongly regulated by changes in energy balance, especially those induced in response to carbohydrates feeding through SREBP-1,(11, 43) which is a key inducer of carbohydrate-stimulated de novo lipogenesis (DNL). Interestingly, among carriers of the rs738409 G- but not C-allele, the DNL increases in proportion to liver fat.(44, 45) These findings might suggest that in presence of carbohydrate overfeeding, homozygous rs738409 CC individuals might be protected against NAFLD and its progression. However, individuals who carry the G-allele might be more susceptible to increased NAFLD severity when carbohydrate intake is high.(24) Most recently, some studies have reported that low-carb diets seem to reduce hepatic fat to a greater extent in patients with PNPLA3 GG genotype as compared to CC genotype, despite similar weight loss in both genotype groups.(46, 47) We hypothesized that increased carbohydrate intake in carriers of the rs738409 G-allele might stimulate the PNPLA3 expression, which may elicit increased expression of a protein with reduced capacity to hydrolyze triglycerides in the liver, favoring its accumulation and subsequent development of inflammation and fibrosis.(48)
It has been recently suggested that carriers of the PNPLA3 rs738409 are at higher risk of liver damage due to imbalanced intake of omega-6:omega-3 PUFAs.(25, 26, 49) More promising findings have shown that PUFAs, but not SFA and MFA may be increased in hepatic triglycerides droplets (TD) but depleted in TD of very low-density lipoproteins among homozygous carriers of the PNPLA3 rs738409 variant compared with non-carriers.(50, 51)These findings may be in alignment with a loss of acylglycerol transacetylase and triacylglycerol hydrolase function in PNPLA3 rs738409-GG homozygous carriers.(52) Our data show that higher intake of n-3 PUFAs was associated with lower fibrosis severity in PNPLA3 rs738409 GC/GG than CC genotype. Omega-3 PUFAs are known to downregulate the expression of SREBP-1c, a key regulator of hepatic lipogenesis and fatty acid oxidation. Thus, it would be expected that n-3 PUFAs supplementation can provide significant benefits to patients with NAFLD.
Our study also shows that individuals who consumed higher amount of total isoflavones, methionine and total choline had less risk of significant fibrosis; however, this protective effect seemed to be more remarkable among rs738409 G-allele carriers. In animal models, isoflavones intake has been associated with positive effects on hepatic steatosis and delayed progression of NAFLD through downregulation of several pathways involving SREBP-1c- and FAS-mediated lipogenesis, among others.(18) To date, no experimental model has explored interactions between PNPLA3 gene and isoflavones, methionine, and choline intakes and its impact on NAFLD. Thus, further RCT may confirm the effects of supplementation of nutrients such as isoflavones, methionine and choline on NAFLD severity.
To our knowledge, the present study is the first to examine and quantify the effect of PNPLA3 rs738409 and dietary nutrients on risk of significant fibrosis within a moderation model.(32) The current results highlight the modulating role that PNPLA3 rs738409 have on the relationships between diet and risk of significant fibrosis among NAFLD patients. The strength of this pathway seems to be conditional on rs738409 allelic distribution. The current findings do suggest that individuals carrying the G allele of rs738409, when compared to the C allele, can moderate the effect of certain dietary factors on risk of significant fibrosis.
Some study limitations need to be highlighted. Our study was limited by its cross-sectional nature that did not allow us to establish causal relationships. The recall-based nature of our questionnaire may have led to an underestimation of macronutrient intakes. Our analyses may have been limited by the relatively small sample size, albeit our moderation analyses use bootstrapping method that is considered the most powerful, most effective method to use with small samples, and the least vulnerable to type I errors. Results rely on non-Hispanic whites; thus, future studies in ethnically mixed NAFLD populations may be considered. Similarly, relationships between nutrients and other gene variants related with NAFLD such as in TM6SF2, HSD17B13, FADS1/S2 and others should be explored. Finally, we lack a non-NAFLD liver disease control group with genetic, dietary, and liver histology data, so further studies might confirm whether PNPLA3 rs738409-diet interactions seen in NAFLD patients are reproducible in the context of other chronic liver diseases. Further interventional trials incorporating the assessment of macronutrient concentrations in plasma would be required to unravel gene-by-nutrients interactions and its influence on NAFLD.
In conclusion, our data indicate that the PNPLA3 rs738409 variant may modify the association between dietary intakes of carbohydrates, n-3 PUFAs, total isoflavones, methionine, and total choline and fibrosis severity among NAFLD patients. Carriers of the PNPLA3 rs734809 G-allele with a high carbohydrate intake may be prone to more aggressive histological phenotypes of NAFLD, supporting the hypothesis of a severe functional defect in the hydrolysis of triglycerides in carriers of the G allele. In addition, high dietary intakes of n-3 PUFAs, total isoflavones, methionine and total choline may attenuate liver fibrosis specially in individuals carrying PNPLA3 rs734809 G-allele. Our results may have important implications for future targeted nutritional intervention in NAFLD patients based on PNPLA3 rs738409 genotype. Individuals with rs738409 G-allele might get greater benefits of low-carbohydrate diets enriched with omega-3 supplements and high content in polyphenols, methionine, and choline than those carrying the CC genotype. However, further prospective interventional studies might corroborate the long-term benefits of diets or specific nutrients in individuals with NAFLD and specific gene variants.
Supplementary Material
Study Highlights:
What is known
Either PNPLA3 rs738409 or certain dietary factors are associated with risk of fibrosis severity
PNPLA3 gene expression is regulated by some nutrients, but relationships between rs738409, nutrients intake, and liver histology in NAFLD is not well understood.
What is new here
Dietary carbohydrate, n-3 PUFAs, total isoflavones, methionine, and total choline intake are associated with risk of fibrosis severity.
The relationship between specific dietary nutrients and risk of fibrosis is different for individuals carrying the rs738409 G-allele.
High carbohydrate intake associates to increased risk of fibrosis severity amongst carriers of rs738409 G-allele.
High intakes of n-3 PUFAs, total isoflavones, methionine, and total choline relate to reduced risk of fibrosis severity amongst carriers of rs738409 G-allele.
Acknowledgements
This study was approved by the NIDDK funded Nonalcoholic Steatohepatitis Clinical Research Network as an Ancillary Study (NASH CRN AS # 93). Biosamples and phenotype and histological data were provided by the NASH CRN.
Financial support:
This study was supported in part by 1 R01 DK106540 to Dr. Liu and by David W Crabb Endowed Professorship funds to Dr. Chalasani.
Abbreviations
- PNPLA-3
Patatin-like phospholipase domain-containing 3
- NAFLD
Nonalcoholic fatty liver disease
- CHO
carbohydrates
- ChREBP
carbohydrate response element binding protein
- SREBP-1c
sterol regulatory element-binding protein-1c
- PUFAs
polyunsaturated fatty acids
- NASH
Nonalcoholic steatohepatitis
- NHANES III
The Third National Health and Nutrition Examination Survey
- SFA
saturated fatty acid
- MFA
monosaturated fatty acids
- SD
standard deviation
- OR
odd ratio
- CI
confidence interval
- BMI
body mass index
- RCT
randomized controlled trial
- DNL
de novo lipogenesis
- VLDL
very low-density lipoprotein
- HDL
High-density lipoprotein
- LDL
Low-density lipoprotein
- AST
Aspartate aminotransferase
- ALT
Alanine aminotransferase
- Hb1Ac
Hemoglobin A1c
- HOMA-IR
Homeostatic Model Assessment of Insulin Resistance
- T2DM
Type 2 diabetes mellitus
- PCT
percentile
Footnotes
Conflicts of Interests:
There are none for this paper. For full disclosure, Dr. Chalasani has ongoing paid consulting activities (or had in preceding 12 months) with NuSirt, Abbvie, Afimmune (DS Biopharma), Allergan (Tobira), Madrigal, Coherus, Siemens, La Jolla, Foresite labs, and Genentech. These consulting activities are generally in the areas of nonalcoholic fatty liver disease and drug hepatotoxicity. Dr. Chalasani receives research grant support from Exact Sciences, Intercept, and Galectin Therapeutics where his institution receives the funding. Over the last decade, Dr. Chalasani has served as a paid consultant to more than 35 pharmaceutical companies and these outside activities have regularly been disclosed to his institutional authorities. Drs. Vilar-Gomez, Sookoian, Pirola, Wilson, Belt, Liu and Tiebing have nothing to disclose.
REFERENCES
- 1.Romeo S, Kozlitina J, Xing C, et al. Genetic variation in PNPLA3 confers susceptibility to nonalcoholic fatty liver disease. Nat Genet 2008;40:1461–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Sookoian S, Pirola CJ. Meta-analysis of the influence of I148M variant of patatin-like phospholipase domain containing 3 gene (PNPLA3) on the susceptibility and histological severity of nonalcoholic fatty liver disease. Hepatology 2011;53:1883–94. [DOI] [PubMed] [Google Scholar]
- 3.Rotman Y, Koh C, Zmuda JM, et al. The association of genetic variability in patatin-like phospholipase domain-containing protein 3 (PNPLA3) with histological severity of nonalcoholic fatty liver disease. Hepatology 2010;52:894–903. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Speliotes EK, Butler JL, Palmer CD, et al. PNPLA3 variants specifically confer increased risk for histologic nonalcoholic fatty liver disease but not metabolic disease. Hepatology 2010;52:904–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Sookoian S, Castano GO, Burgueno AL, et al. A nonsynonymous gene variant in the adiponutrin gene is associated with nonalcoholic fatty liver disease severity. J Lipid Res 2009;50:2111–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Unalp-Arida A, Ruhl CE. Patatin-Like Phospholipase Domain-Containing Protein 3 I148M and Liver Fat and Fibrosis Scores Predict Liver Disease Mortality in the U.S. Population. Hepatology 2020;71:820–834. [DOI] [PubMed] [Google Scholar]
- 7.He S, McPhaul C, Li JZ, et al. A sequence variation (I148M) in PNPLA3 associated with nonalcoholic fatty liver disease disrupts triglyceride hydrolysis. J Biol Chem 2010;285:6706–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Kumari M, Schoiswohl G, Chitraju C, et al. Adiponutrin functions as a nutritionally regulated lysophosphatidic acid acyltransferase. Cell Metab 2012;15:691–702. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.BasuRay S, Smagris E, Cohen JC, et al. The PNPLA3 variant associated with fatty liver disease (I148M) accumulates on lipid droplets by evading ubiquitylation. Hepatology 2017;66:1111–1124. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Dubuquoy C, Robichon C, Lasnier F, et al. Distinct regulation of adiponutrin/PNPLA3 gene expression by the transcription factors ChREBP and SREBP1c in mouse and human hepatocytes. J Hepatol 2011;55:145–53. [DOI] [PubMed] [Google Scholar]
- 11.Huang Y, He S, Li JZ, et al. A feed-forward loop amplifies nutritional regulation of PNPLA3. Proc Natl Acad Sci U S A 2010;107:7892–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hao L, Ito K, Huang KH, et al. Shifts in dietary carbohydrate-lipid exposure regulate expression of the non-alcoholic fatty liver disease-associated gene PNPLA3/adiponutrin in mouse liver and HepG2 human liver cells. Metabolism 2014;63:1352–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Abdelmalek MF, Suzuki A, Guy C, et al. Increased fructose consumption is associated with fibrosis severity in patients with nonalcoholic fatty liver disease. Hepatology 2010;51:1961–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Jensen T, Abdelmalek MF, Sullivan S, et al. Fructose and sugar: A major mediator of non-alcoholic fatty liver disease. J Hepatol 2018;68:1063–1075. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Araya J, Rodrigo R, Videla LA, et al. Increase in long-chain polyunsaturated fatty acid n - 6/n - 3 ratio in relation to hepatic steatosis in patients with non-alcoholic fatty liver disease. Clin Sci (Lond) 2004;106:635–43. [DOI] [PubMed] [Google Scholar]
- 16.Parker HM, Johnson NA, Burdon CA, et al. Omega-3 supplementation and non-alcoholic fatty liver disease: a systematic review and meta-analysis. J Hepatol 2012;56:944–51. [DOI] [PubMed] [Google Scholar]
- 17.Guo XF, Yang B, Tang J, et al. Fatty acid and non-alcoholic fatty liver disease: Meta-analyses of case-control and randomized controlled trials. Clin Nutr 2018;37:113–122. [DOI] [PubMed] [Google Scholar]
- 18.Liu H, Zhong H, Leng L, et al. Effects of soy isoflavone on hepatic steatosis in high fat-induced rats. J Clin Biochem Nutr 2017;61:85–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Xin X, Chen C, Hu YY, et al. Protective effect of genistein on nonalcoholic fatty liver disease (NAFLD). Biomed Pharmacother 2019;117:109047. [DOI] [PubMed] [Google Scholar]
- 20.Machado MV, Michelotti GA, Xie G, et al. Mouse models of diet-induced nonalcoholic steatohepatitis reproduce the heterogeneity of the human disease. PLoS One 2015;10:e0127991. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Anstee QM, Goldin RD. Mouse models in non-alcoholic fatty liver disease and steatohepatitis research. Int J Exp Pathol 2006;87:1–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Smagris E, BasuRay S, Li J, et al. Pnpla3I148M knockin mice accumulate PNPLA3 on lipid droplets and develop hepatic steatosis. Hepatology 2015;61:108–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Li JZ, Huang Y, Karaman R, et al. Chronic overexpression of PNPLA3I148M in mouse liver causes hepatic steatosis. J Clin Invest 2012;122:4130–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Davis JN, Le KA, Walker RW, et al. Increased hepatic fat in overweight Hispanic youth influenced by interaction between genetic variation in PNPLA3 and high dietary carbohydrate and sugar consumption. Am J Clin Nutr 2010;92:1522–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Santoro N, Savoye M, Kim G, et al. Hepatic fat accumulation is modulated by the interaction between the rs738409 variant in the PNPLA3 gene and the dietary omega6/omega3 PUFA intake. PLoS One 2012;7:e37827. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Stojkovic IA, Ericson U, Rukh G, et al. The PNPLA3 Ile148Met interacts with overweight and dietary intakes on fasting triglyceride levels. Genes Nutr 2014;9:388. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Sanyal AJ, Chalasani N, Kowdley KV, et al. Pioglitazone, vitamin E, or placebo for nonalcoholic steatohepatitis. N Engl J Med 2010;362:1675–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Neuschwander-Tetri BA, Clark JM, Bass NM, et al. Clinical, laboratory and histological associations in adults with nonalcoholic fatty liver disease. Hepatology 2010;52:913–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Boucher B, Cotterchio M, Kreiger N, et al. Validity and reliability of the Block98 food-frequency questionnaire in a sample of Canadian women. Public Health Nutr 2006;9:84–93. [DOI] [PubMed] [Google Scholar]
- 30.Block G, Woods M, Potosky A, et al. Validation of a self-administered diet history questionnaire using multiple diet records. J Clin Epidemiol 1990;43:1327–35. [DOI] [PubMed] [Google Scholar]
- 31.Kleiner DE, Brunt EM, Van Natta M, et al. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology 2005;41:1313–21. [DOI] [PubMed] [Google Scholar]
- 32.Hayes AF. Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York, NY, US: Guilford Press; 2013. [Google Scholar]
- 33.Hayes AF. An Index and Test of Linear Moderated Mediation. Multivariate Behav Res 2015;50:1–22. [DOI] [PubMed] [Google Scholar]
- 34.Preacher KJ, Hayes AF. Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods 2008;40:879–891. [DOI] [PubMed] [Google Scholar]
- 35.Scorletti E, Byrne CD. Omega-3 fatty acids and non-alcoholic fatty liver disease: Evidence of efficacy and mechanism of action. Mol Aspects Med 2018;64:135–146. [DOI] [PubMed] [Google Scholar]
- 36.Nogueira MA, Oliveira CP, Ferreira Alves VA, et al. Omega-3 polyunsaturated fatty acids in treating non-alcoholic steatohepatitis: A randomized, double-blind, placebo-controlled trial. Clin Nutr 2016;35:578–86. [DOI] [PubMed] [Google Scholar]
- 37.Scorletti E, Bhatia L, McCormick KG, et al. Effects of purified eicosapentaenoic and docosahexaenoic acids in nonalcoholic fatty liver disease: results from the Welcome* study. Hepatology 2014;60:1211–21. [DOI] [PubMed] [Google Scholar]
- 38.Nobili V, Carpino G, Alisi A, et al. Role of docosahexaenoic acid treatment in improving liver histology in pediatric nonalcoholic fatty liver disease. PLoS One 2014;9:e88005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Sanyal AJ, Abdelmalek MF, Suzuki A, et al. No significant effects of ethyl-eicosapentanoic acid on histologic features of nonalcoholic steatohepatitis in a phase 2 trial. Gastroenterology 2014;147:377–84 e1. [DOI] [PubMed] [Google Scholar]
- 40.Medjakovic S, Mueller M, Jungbauer A. Potential health-modulating effects of isoflavones and metabolites via activation of PPAR and AhR. Nutrients 2010;2:241–79. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Ahn TG, Yang G, Lee HM, et al. Molecular mechanisms underlying the anti-obesity potential of prunetin, an O-methylated isoflavone. Biochem Pharmacol 2013;85:1525–33. [DOI] [PubMed] [Google Scholar]
- 42.Garcia-Lafuente A, Guillamon E, Villares A, et al. Flavonoids as anti-inflammatory agents: implications in cancer and cardiovascular disease. Inflamm Res 2009;58:537–52. [DOI] [PubMed] [Google Scholar]
- 43.Perttila J, Huaman-Samanez C, Caron S, et al. PNPLA3 is regulated by glucose in human hepatocytes, and its I148M mutant slows down triglyceride hydrolysis. Am J Physiol Endocrinol Metab 2012;302:E1063–9. [DOI] [PubMed] [Google Scholar]
- 44.Sevastianova K, Santos A, Kotronen A, et al. Effect of short-term carbohydrate overfeeding and long-term weight loss on liver fat in overweight humans. Am J Clin Nutr 2012;96:727–34. [DOI] [PubMed] [Google Scholar]
- 45.Mancina RM, Matikainen N, Maglio C, et al. Paradoxical dissociation between hepatic fat content and de novo lipogenesis due to PNPLA3 sequence variant. J Clin Endocrinol Metab 2015;100:E821–5. [DOI] [PubMed] [Google Scholar]
- 46.Shen J, Wong GL, Chan HL, et al. PNPLA3 gene polymorphism and response to lifestyle modification in patients with nonalcoholic fatty liver disease. J Gastroenterol Hepatol 2015;30:139–46. [DOI] [PubMed] [Google Scholar]
- 47.Sevastianova K, Kotronen A, Gastaldelli A, et al. Genetic variation in PNPLA3 (adiponutrin) confers sensitivity to weight loss-induced decrease in liver fat in humans. Am J Clin Nutr 2011;94:104–11. [DOI] [PubMed] [Google Scholar]
- 48.Dongiovanni P, Valenti L. A Nutrigenomic Approach to Non-Alcoholic Fatty Liver Disease. Int J Mol Sci 2017;18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Arendt BM, Comelli EM, Ma DW, et al. Altered hepatic gene expression in nonalcoholic fatty liver disease is associated with lower hepatic n-3 and n-6 polyunsaturated fatty acids. Hepatology 2015;61:1565–78. [DOI] [PubMed] [Google Scholar]
- 50.Mitsche MA, Hobbs HH, Cohen JC. Patatin-like phospholipase domain-containing protein 3 promotes transfer of essential fatty acids from triglycerides to phospholipids in hepatic lipid droplets. J Biol Chem 2018;293:6958–6968. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Luukkonen PK, Nick A, Holtta-Vuori M, et al. Human PNPLA3-I148M variant increases hepatic retention of polyunsaturated fatty acids. JCI Insight 2019;4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Huang Y, Cohen JC, Hobbs HH. Expression and characterization of a PNPLA3 protein isoform (I148M) associated with nonalcoholic fatty liver disease. J Biol Chem 2011;286:37085–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.