Abstract
Background & Aims.
Emerging data from pediatric populations suggest that variants in the autophagy-governing Immunity-related GTPase M (IRGM) gene may contribute to nonalcoholic fatty liver disease (NAFLD) susceptibility. We examined the relationship between IRGM rs13361189 variants and NAFLD in a community-based cohort of adults.
Methods.
We included all Framingham Heart Study participants with available data on the IRGM rs13361189 variant, undergoing study-directed computed tomography (CT) scans of the abdomen (2002-2005). Using multivariable linear and logistic regression modeling, we evaluated cross-sectional associations between rs13361189 genotype and hepatic steatosis (HS). Among the subset of participants without baseline HS and who underwent follow-up CT scan between 2008-2011, we used multivariable logistic regression modeling to assess the longitudinal relationship between IRGM rs13361189 genotype and risk for incident HS.
Results.
Among 2,070 participants (50% women; mean age 51±11 years), 332 (16%) had one copy of the variant rs13361189 variant C allele, while 19 (1%) had the CC genotype. Compared to the TT genotype, there was no increased odds of prevalent HS with the CT or CC genotype (multivariable-adjusted odds ratio [OR] 0.93 [95% CI 0.68-128] and 0.86 [95% CI 0.46-1.63], respectively). Among individuals without baseline HS (n=1,052), 19.3% developed incident HS over median 6.1 years. Compared to the TT genotype, neither the CT nor the CC genotype were significantly associated with incident HS (all p>0.05).
Conclusion.
In our community-based, longitudinal cohort of Caucasian adults, variants in the autophagy-governing IRGM gene at the rs13361189 locus were not associated with increased prevalent or incident HS.
Keywords: nonalcoholic fatty liver disease, steatosis, autophagy, gene variants
Introduction
Nonalcoholic fatty liver disease (NAFLD) now represents the most common cause of chronic liver disease in the United States, impacting an estimated 19-24% of the population1. NAFLD is closely linked to diabetes and obesity, including an excess accumulation of metabolically-active visceral adipose tissue (VAT)2. However, not all individuals with diabetes or obesity develop NAFLD, and genetic studies support the heritability of NAFLD3. Genome-wide association studies (GWAS) have identified several NAFLD susceptibility genes, including patatin-like phospholipase domain-containing 3 (PNPLA3)4. However, the proportion of variance in genetic risk explained by these loci remains modest (<5%), and it is possible that some of the “missing heritability” derives from low frequency or rare variants.
Autophagy is a highly conserved metabolic process whereby defective intracellular organelles are targeted for lysosomal destruction5. Defects in autophagy confer susceptibility to chronic inflammatory diseases including obesity and diabetes6, and also to impair ß-oxidation of intrahepatic lipids5. Consequently, it has been hypothesized that impaired autophagy may contribute to the pathogenesis of NAFLD7. In preclinical models, inhibition of autophagy promotes hepatic triglyceride accumulation, while autophagy stimulation with rapamycin accelerates lipid clearance, improving both steatosis and central obesity8. In humans, two cross-sectional studies recently demonstrated that compared to the wild-type (CC genotype), the TT genotype at the rs10065172 locus of the autophagy-governing, immunity-related GTPase M (IRGM) gene is associated with prevalent NAFLD9, 10, and with elevations in alanine aminotransferase (ALT)10. To date, however, the link between functional genetic defects in IRGM and NAFLD susceptibility has not previously been extended to an adult population.
Thus, we examined a functionally-annotated variant in IRGM at the rs13361189 locus, which has been shown in European populations to be in perfect linkage disequilibrium with the rs10065172 locus (r2=1.0)11. In a prior study of a predominantly Han Chinese population, these two loci also were found to be in tight linkage disequilibrium10. Specifically, we evaluated the association between rs13361189 genotype and both prevalent and incident hepatic steatosis (HS), defined by multidetector computed tomography (MDCT), in a well-phenotyped and prospective community-based cohort, the Framingham Heart Study (FHS).
Methods
Detailed methods are outlined in the Supplementary file.
Results
Baseline demographics and clinical characteristics of the 2,070 participants according to rs13361189 genotype are shown in Table 1. Compared to participants with the rs13361189 TT genotype (n=1719, 83%) or the CT genotype (n=332, 16%), those with the CC genotype (n=19; 1%) were significantly older (mean±standard deviation [SD] age [years] CC vs. CT vs. TT = 54±9 vs. 52±11 vs. 51±11; p=0.016), and more likely to have hypertension (CC vs. CT vs. TT= 16% vs. 34% vs. 27%; p=0.007). However, mean±SD body mass index (BMI) was similar between groups (mean SD BMI for CC vs. CT vs. TT genotypes: 27±−6 vs. 28±5 vs. 28 ±5; p=0.24), and a similar proportion of participants had diabetes (CC vs. CT vs. TT = 10.5% vs. 4.2% vs. 6.8%; p=0.11).
Table 1.
Baseline characteristics of included Framingham Heart Study participants (N=2,070), according to IRGM rs13361189 genotype
| rs13361189 genotype |
||||
|---|---|---|---|---|
| Variable* | TT N=1719 |
CT N=332 |
CC N=19 |
P-value1,2 |
| Age, SD | 50.6+/−10.5 | 52.2+/−11.0 | 54.1+/−9.2 | 0.0162 |
| Female sex, % | 858 (49.9%) | 169 (50.9%) | 8 (42.1%) | 0.746 |
| White, % | 1719 (100%) | 332 (100%) | 19 (100%) | - |
| Alcohol (drinks / week), SD | 3.0+/−3.4 | 2.8+/−3.7 | 3.2+/−4.0 | 0.658 |
| Cohort, % | 0.076 | |||
| • Offspring | 663 (38.6%) | 147 (44.3%) | 10 (52.6%) | |
| • Generation 3 | 1056 (61.4%) | 185 (55.7%) | 9 (47.4%) | |
| BMI, kg/m2, SD | 27.7+/−5.2 | 28.1+/−5.0 | 26.5 +/− 5.7 | 0.241 |
| Waist circumference (cm), SD | 96.7+/−13.9 | 97.9+/−14.5 | 92.7+/−14.6 | 0.167 |
| Diabetes, % | 117 (6.8%) | 14 (4.2%) | 2 (10.5%) | 0.106$ |
| Hypertension, % | 455 (26.5%) | 114 (34.3%) | 3 (15.8%) | 0.007 |
| Hypertriglyceridemia, % | 479 (27.9%) | 103 (31.3%) | 2 (10.5%) | 0.099 |
| HDL cholesterol, mg/d, (SD) | 51.9 (15.5) | 51.0 (16.2) | 53.1 (12.1) | 0.636 |
| Low HDL cholesterol, % | 558 (32.5%) | 122 (36.7%) | 3 (15.8%) | 0.087 |
| Smoking, % | 0.019$ | |||
| • Current | 872 (50.7%) | 164 (49.4%) | 9 (47.4%) | |
| • Former | 643 (37.4%) | 143 (43.1%) | 5 (26.3%) | |
| • Never | 204 (11.9%) | 25 (7.5%) | 5 (26.3%) | |
| Liver phantom ratio, SD | 0.4 (0.1) | 0.4 (0.1) | 0.4 (0.03) | 0.129 |
| Liver phantom ratio ≤ 0.33, % | 304 (17.7%) | 63 (19.0%) | 1 (5.3%) | 0.338 |
| VAT median (range), cm3, SD | 1634.0 (172.7-6410.0) | 1691.0 (227.3-6207.0) | 1335.0 (325.5-5850.0) | 0.635 |
| VAT volume, cm3, SD | 1776.7 (1004.7) | 1827.0 (1010.1) | 1703.2 (1384.2) | 0.666 |
| ALT, IU/L, SD | 25.4 (19.7) | 24.6 (15.7) | 20.6 (6.9) | 0.43 |
| Elevated ALT, % | 613 (35.7%) | 120 (36.1%) | 4 (21.1%) | 0.407 |
| AST, IU/L, SD | 23.5 (14.0) | 22.9 (8.2) | 20.4 (3.9) | 0.472 |
All variables represented as mean (standard deviation, SD) unless stated otherwise.
Abbreviations: IRGM, immunity-related GTPase M gene; VAT, visceral adipose tissue; SAT, subcutaneous adipose tissue; BMI, body mass index; HDL, high-density lipoprotein; LDL, low-density lipoprotein; ALT, alanine aminotransferase; AST, aspartate aminotransferase; SD, standard deviation
P-value is from Fisher’s exact test for categorical data.
Mood’s median test is used to compare 3 medians for continuous variables.
Overall, we recorded 368 cases of prevalent HS, including 304 (18%) among participants with the rs13361189 TT genotype, 63 (19%) among those with the CT genotype, and 1 (5%) among those with the CC genotype (Table 1). Compared to the rs13361189 TT genotype, no significant difference in HS was found with the CT genotype (adjusted ß= −0.001, 95% CI −0.006,0.005) or the CC genotype (adjusted ß = −0.002, 95% CI −0.013,0.009; Table 2). These associations were not materially altered after accounting for BMI, diabetes, smoking status and alcohol intake. Next, we evaluated the relationship between IRGM variants and prevalent HS, defined dichotomously by LPR≤0.33 (Table 2). Compared to the TT genotype, neither the CT nor the CC genotype were associated with prevalent HS (adjusted OR 0.93 [95% CI 0.68-128] and 0.86 [95% CI 0.46-1.63], respectively. There was no evidence of effect modification by BMI (Table S1; p-interaction>0.05) or by VAT volume (Table S2; p-interaction>0.05). We also assessed the relationship between IRGM variants and elevations in serum ALT (Table 2). However, compared to the TT genotype, no significant association with elevated ALT was found with either the CT or the CC genotypes (adjusted OR 0.97 [95% CI 0.77-1.23] and 0.95 [95% CI 0.59-1.51], respectively).
Table 2.
Cross-sectional association between the IRGM rs13361189 variant (C) allele1 and hepatic steatosis2 or elevated ALT3, among Framingham Heart Study participants (n=2,070)
|
IRGM Genotype |
||||
|---|---|---|---|---|
| Outcome | rs13361189 TT (Ref.) n=1719 |
rs13361189 CT n=332 |
rs13361189 CC n=19 |
P-value |
| Difference in mean Liver Phantom Ratio2 | -- | |||
| Model 1*; β' [95% CI]◊ | 1 | 0.002 [−0.004, 0.009] | 0.005 [−0.008, 0.018] | 0.444 |
| Multivariable model 2¥◊; β' [95% CI] | 1 | 0.003 [−0.003, 0.009] | 0.007 [−0.006, 0.019] | 0.288 |
| Hepatic Steatosis2 | -- | |||
| Model 1*; OR [95% CI] | 1 | 0.93 [0.67, 1.28] | 0.86 [0.45, 1.65] | 0.653 |
| Multivariable Model 2¥; OR [95% CI] | 1 | 0.93 [0.68-1.28] | 0.86 [0.46, 1.63] | 0.653 |
| Elevated ALT (IU/L)3 | -- | |||
| Model 1*; OR [95% CI] | 1 | 0.97 [0.77, 1.22] | 0.94 [0.59, 1.50] | 0.796 |
| Multivariable Model 2¥; OR [95% CI] | 1 | 0.97 [0.77-1.23] | 0.95 [0.59-1.51] | 0.822 |
Abbreviations: OR, Odds Ratio; CI, confidence interval; ALT, alanine aminotransferase; IRGM, Immunity-Related GTPase M
Variant allele assessed using an additive model.
Measures of hepatic steatosis included (1) continuous liver phantom ratio, in which a lower ratio corresponds to increased hepatic fat accumulation, and (2) hepatic steatosis defined dichotomously by liver-phantom ratio ≤ 0.33.
Elevated ALT level defined as ALT > upper limit of normal (ULN), which was defined as ALT > 19U/L in women and ALT >30U/L in men.
Beta coefficient representing the difference in mean liver attenuation between genotype of interest compared to the rs13361189 TT genotype (referent)
Model 1: adjusted for age (years), sex, cohort (Third Generation vs. Omni), hypertension, hypertriglyceridemia, low-HDL-C, the number of PNPLA3 rs738409 variant alleles, and eigenvectors #1-#3.
Model 2: Model 1 + smoking status (current, prior, never), alcohol intake (drinks/week), body mass index (BMI, kg/m2) and diabetes (yes/no)
In secondary analyses, we explored the association between rs13361189 genotype and prevalent HS using a dominant genotype model that compared the CT/CC genotypes to the TT genotype, and our findings were null (adjusted OR for HS=0.96, 95% CI 0.68-1.36). Similarly, there was no significant association between rs13361189 genotype and elevated ALT using a dominant model (adjusted OR=0.95, 95% CI 0.59-1.51).
Finally, in the exploratory longitudinal analysis (n=1,052), we identified 204 cases of incident HS between paired MDCT scans, over a median of 6.1 years. In both the minimally-adjusted and the full multivariable models, there were no significant associations between IRGM rs13361189 variants and incident NAFLD (Table S3).
Discussion
To our knowledge, this represents the first epidemiological study of the relationship between IRGM rs13361189 variants and HS risk in a prospective, community-dwelling adult population. Despite appropriate statistical power and careful accounting for important confounders, including PNPLA3 variants and population stratification, the IRGM rs13361189 gene variant did not influence hepatic fat content, nor was it associated with risk of prevalent or incident steatosis, in this population.
A body of recent data suggests that defects in autophagy may influence the pathogenesis of HS and steatohepatitis (NASH), through hepatic lipid dysregulation and inflammatory activation5. While the precise mechanisms remain uncharacterized, IRGM is known to influence mitochondrial function, where it plays a direct role in the co-assembly of core autophagy machinery, including ULK1, Beclin-1 and ATG16L112. In vitro, the knockdown of IRGM expression decreases cellular levels of phospho-ULK1, Beclin-1, and LC3-II, which leads to accumulation of intrahepatic lipid droplets10. In contrast, this phenotype is rescued when IRGM is overexpressed, or when cells are treated with rapamycin, an inducer of autophagy10. Furthermore, preclinical models of IRGM-knockout mice show that impaired autophagy increases susceptibility to diabetes6 as well as to NAFLD13.
Clinical evidence regarding a relationship between IRGM gene variants and NAFLD susceptibility is scarce9, 10. Within this community-based cohort of adults, we found no association between the IRGM rs13361189 variant and risk of either prevalent or incident steatosis. This is supported by previous population-based adult cohort studies, in which neither the IRGM rs10065172 or rs13361189 variants conferred susceptibility to NAFLD at GWAS significance thresholds4, despite a MAF of at least 10% in European populations14. Although our findings appear to contrast with two recent cross-sectional candidate gene studies9, 10, there are several potential explanations for these discrepant findings. First, the prior studies were limited by small sample size, cross-sectional design, and included highly-selected, pediatric populations, with limited generalizability. Second, candidate gene studies carry a well-recognized risk for false positive findings and lack of reproducibility, particularly when conducted in small, selected populations15. While a role remains for classical candidate gene studies in NAFLD, particularly for mechanistic or functional associations, our findings highlight the continued need for GWAS, exome-, and whole genome-wide association studies as well as investigations of epigenetic and transcriptional regulation of NAFLD.
Finally, it is also possible that the influence of IRGM gene variants on NAFLD risk is specific to pediatric populations. Pediatric NASH remains a challenging diagnosis, as it is marked by unique histopathological features rarely found in adults16. In a retrospective study of 100 pediatric NASH liver biopsy specimens, only 17% of children had adult-type NASH, defined by steatosis, lobular inflammation and hepatocyte ballooning1, while more than 50% had “Type 2” NASH, with macrovesicular steatosis, portal inflammation with or without fibrosis, and little to no ballooning degeneration16. Children with Type 2 NASH are more likely to be male, younger in age, and obese, compared to children with adult-type NASH16. To date, it remains unknown whether these two distinct patterns reflect underlying differences in the etiopathogenesis or evolution of disease; similarly, it is not known whether variants in IRGM could account for a proportion of these differences. To test this hypothesis, future prospective natural history studies are needed in well-phenotyped adult and pediatric populations.
We highlight several important limitations. First, our study population was limited to Caucasians, thus our results may not be directly translatable to persons of other ancestries. Second, only 19 participants in our cohort had the rs13361189 variant CC genotype; this highlights the need for future large-scale studies in diverse populations, particularly given the potential impact of ethnic variability in IRGM genotypes on NAFLD susceptibility9, 10, as well as the possibility of effect modification by PNPLA3 genotype9. Finally, liver biopsy remains the gold standard for the assessment of nonalcoholic steatohepatitis (NASH) and fibrosis, and future studies will be needed to characterize the impact of IRGM variants on NAFLD histology. Nevertheless, CT is well-validated for the assessment and quantification of hepatic steatosis, which has demonstrated clinical and prognostic significance for persons with NAFLD.
In conclusion, within a large, prospective community-based population of U.S. adults, the IRGM rs13361189 variant was not significantly associated with prevalent or incident HS. Additional prospective studies are needed to validate our findings, and to determine if this association is unique to pediatric NAFLD/NASH.
Supplementary Material
Acknowledgments
Funding Sources:
This work was supported by National Heart, Lung, and Blood Institute contracts N01-HC-25195 and HHSN268201500001l. XD is supported by NIH R01 DK089256. ML is supported by the National Institute of Diabetes and Digestive and Kidney Diseases K23 DK113252 and the Boston University School of Medicine Department of Medicine Career Investment Award. TS is supported by the American Association for the Study of Liver Diseases (AASLD) Clinical and Translational Research Award.
List of Abbreviations:
- IRGM
Immunity-related GTPase M
- NAFLD
nonalcoholic fatty liver disease
- CT
computed tomography
- HS
hepatic steatosis
- OR
odds ratio
- CI
confidence interval
- VAT
visceral adipose tissue
- GWAS
Genome-wide association studies
- PNPLA3
patatin-like phospholipase domain-containing 3
- ALT
alanine aminotransferase
- MDCT
multidetector computed tomography
- FHS
Framingham Heart Study
- BMI
body mass index
- LPR
liver-phantom ratio
- MAF
mean allelic frequency
- NASH
nonalcoholic steatohepatitis
Footnotes
Conflict of Interest: The authors report no conflicts of interest.
REFERENCES
- 1.Chalasani N, Younossi Z, Lavine JE, et al. The Diagnosis and Management of Nonalcoholic Fatty Liver Disease: Practice Guidance from the American Association for the Study of Liver Diseases. Hepatology 2017. [DOI] [PubMed] [Google Scholar]
- 2.Speliotes EK, Massaro JM, Hoffmann U, et al. Fatty liver is associated with dyslipidemia and dysglycemia independent of visceral fat: the Framingham Heart Study. Hepatology 2010;51:1979–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Schwimmer JB, Celedon MA, Lavine JE, et al. Heritability of nonalcoholic fatty liver disease. Gastroenterology 2009;136:1585–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Speliotes EK, Yerges-Armstrong LM, Wu J, et al. Genome-wide association analysis identifies variants associated with nonalcoholic fatty liver disease that have distinct effects on metabolic traits. PLoS Genet 2011;7:e1001324. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Singh R, Kaushik S, Wang Y, et al. Autophagy regulates lipid metabolism. Nature 2009;458:1131–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Lee HY, Kim J, Quan W, et al. Autophagy deficiency in myeloid cells increases susceptibility to obesity-induced diabetes and experimental colitis. Autophagy 2016;12:1390–403. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Singh R, Xiang Y, Wang Y, et al. Autophagy regulates adipose mass and differentiation in mice. J Clin Invest 2009;119:3329–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Lin CW, Zhang H, Li M, et al. Pharmacological promotion of autophagy alleviates steatosis and injury in alcoholic and non-alcoholic fatty liver conditions in mice. J Hepatol 2013;58:993–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Bellini G, Miraglia Del Giudice E, Nobili V, et al. The IRGM rs10065172 variant increases the risk for steatosis but not for liver damage progression in Italian obese children. J Hepatol 2017;67:653–655. [DOI] [PubMed] [Google Scholar]
- 10.Lin YC, Chang PF, Lin HF, et al. Variants in the autophagy-related gene IRGM confer susceptibility to non-alcoholic fatty liver disease by modulating lipophagy. J Hepatol 2016;65:1209–1216. [DOI] [PubMed] [Google Scholar]
- 11.Broad Institute Single Nucleotide Polymorphism Annotation and Proxy Search (SNAP). http://archive.broadinstitute.org/mpg/snap/ldsearch.php; accessed September 4, 2017.
- 12.Chauhan S, Mandell MA, Deretic V. IRGM governs the core autophagy machinery to conduct antimicrobial defense. Mol Cell 2015;58:507–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Hur JH, Park SY, Dall’Armi C, et al. Phospholipase D1 deficiency in mice causes nonalcoholic fatty liver disease via an autophagy defect. Sci Rep 2016;6:39170. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.McCarroll SA, Huett A, Kuballa P, et al. Deletion polymorphism upstream of IRGM associated with altered IRGM expression and Crohn’s disease. Nat Genet 2008;40:1107–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Border R, Keller MC. Commentary: Fundamental problems with candidate gene-by-environment interaction studies - reflections on Moore and Thoemmes (2016). J Child Psychol Psychiatry 2017;58:328–330. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Schwimmer JB, Behling C, Newbury R, et al. Histopathology of pediatric nonalcoholic fatty liver disease. Hepatology 2005;42:641–9. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.