Potential conflict of interest: Nothing to report.
Supported by a grant from the Academy of Finland (to J.P., Contract no. 120979; 138006; to M.L., Contract no. 124243), the Finnish Diabetes Research Foundation (to J.P. and M.L.), the Finnish Cultural Foundation (to J.P.), the Finnish Heart Foundation (to M.L.), TEKES (to M.L., Contract no. 1510/31/06), and Commission of the European Community (to M.L., Contract no. LSHM-CT-2004-512013 EUGENE2).
Address reprint requests to: Jussi Pihlajamäki, M.D., Ph.D., Professor in Clinical Nutrition, Institute of Public Health and Clinical Nutrition, University of Eastern Finland, 70210 Kuopio, Finland. E–mail: firstname.lastname@example.org fax: +358-17-162792
Dysregulation of the cholesterol synthesis pathway and accumulation of cholesterol in the liver are linked to the pathogenesis of nonalcoholic steatohepatitis (NASH). Therefore, we investigated the association of serum and liver levels of cholesterol precursors with NASH. Liver histology was assessed in 110 obese patients (Kuopio Obesity Surgery Study [KOBS] study, age 43.7 ± 8.1 years [mean ± standard deviation, SD], body mass index [BMI] 45.0 ± 6.1 kg/m2). Serum and liver levels of cholesterol precursors were measured with gas-liquid chromatography. The association between cholesterol precursors and serum alanine aminotransferase (ALT), as a marker of liver disease, was also investigated in a population cohort of 717 men (Metabolic Syndrome in Men Study [METSIM] study, age 57.6 ± 5.8 years, BMI 27.1 ± 4.0 kg/m2). Serum desmosterol levels and the desmosterol-to-cholesterol ratio were higher in individuals with NASH, but not in individuals with simple steatosis, compared to obese subjects with normal liver histology (P = 0.002 and P = 0.003, respectively). Levels of serum and liver desmosterol correlated strongly (r = 0.667, P = 1 × 10−9), suggesting a shared regulation. Both serum and liver desmosterol levels correlated positively with steatosis and inflammation in the liver (P < 0.05). Serum desmosterol had a higher correlation with the accumulation of cholesterol in the liver than serum cholesterol. Serum desmosterol levels (P = 2 × 10−6) and the serum desmosterol-to-cholesterol ratio (P = 5 × 10−5) were associated with serum ALT in the population study. Conclusion: Levels of desmosterol in serum and the liver were associated with NASH. These results suggest that serum desmosterol is a marker of disturbed cholesterol metabolism in the liver. Whether desmosterol has a more specific role in the pathophysiology of NASH compared to other cholesterol precursors needs to be investigated. (Hepatology 2013;53:976–982)
Nonalcoholic fatty liver disease (NAFLD) is the most common liver disease[1, 2] and can be defined as a hepatic component of metabolic syndrome.[3, 4] Hepatic lipid accumulation contributes to known metabolic alterations, such as insulin resistance, hyperglycemia, and hyperlipidemia. NAFLD can be further divided into two major subtypes, which seem to have different outcomes: simple steatosis (NAFL) without liver inflammation or injury and nonalcoholic steatohepatitis (NASH). NASH leads to liver cirrhosis and to increased mortality more often than NAFL.[9, 10]
Cholesterol metabolism is determined by dietary and genetic factors[11, 12] as well as by metabolic alterations in obesity, insulin resistance,[14, 15] and type 2 diabetes mellitus (DM2). In NAFLD, liver steatosis is associated with increased cholesterol synthesis and decreased cholesterol absorption. Interestingly, triglyceride accumulation alone may not induce liver injury or inflammation,[18, 19] whereas the accumulation of free cholesterol and the dysregulation of the cholesterol synthesis pathway relates to NASH.
The purpose of our study was to investigate cholesterol metabolism in obese individuals with NASH. More specifically, we were interested in differences between individuals with simple steatosis and individuals with NASH. To this end, serum and liver levels of three cholesterol precursor sterols, measured as serum surrogate markers of cholesterol synthesis rate, were analyzed in 110 obese individuals with detailed liver histology. The observed association of serum desmosterol with NASH was replicated in a population-based cohort of 717 men. Our results demonstrate that levels of the cholesterol precursor desmosterol in serum and the liver associate with NASH.
Subjects and Methods
Liver Biopsy Study of Morbidly Obese Subjects
Obese individuals were selected from an ongoing study recruiting all subjects undergoing bariatric surgery at Kuopio University Hospital (35 men and 75 women, age 43.7 ± 8.1 years, body mass index [BMI] 45.0 ± 6.1 kg/m2; for other characteristics see Supporting Table 1).[24, 25] Every subject participated in a 1-day visit including an interview on the history of previous diseases and current drug treatment, and an evaluation of glucose tolerance and cardiovascular risk factors. Fasting blood samples were drawn after 12 hours. All patients with alcohol consumption of more than 2 doses per day were excluded from the study. One individual had gradus 4/4 fibrosis in a liver biopsy but liver function tests were normal and there were no signs of portal hypertension in recruitment or at follow-up. Chronic hepatitis B and C virus (HBV, HCV) were tested using serology if alanine aminotransferase (ALT) levels were elevated prior to surgery. In general, HCV and HBV infections are rare in Finland compared to many other countries (incidence in 2011: chronic HBV infections 4.2/100 000 and all HCV infections 21.7/100 000; Statistical Database of Infectious Disease Register, Finland). If there were signs of iron accumulation in liver biopsy, hemochromatosis was ruled out with serum iron balance measurements and an HFE gene mutation test, when appropriate.
A total of 717 men from the population-based cross-sectional METSIM Study (Metabolic Syndrome in Men Study) were included in the study. Subjects, age 45 to 70 years, were randomly selected from the population register of the town of Kuopio, eastern Finland (population of 95,000). Their age was 57.6 ± 5.8 years and BMI 27.1 ± 4.0 kg/m2. Individuals on statin or fibrate treatments were excluded from the analysis. Informed consent was obtained from each participant and the study protocol was approved by the Ethics Committee of Northern Savo Hospital District and was in accordance with the Helsinki Declaration.
Liver Biopsies and Liver Histology
Liver biopsies were obtained using Trucut needles (Radiplast, Uppsala, Sweden) during elective gastric bypass operations (n = 110). Overall histological assessment of liver biopsy samples was performed by one pathologist according to the standard criteria[26, 27] and histological diagnosis was originally divided into three categories: 1) not NASH; 2) possible NASH; and 3) definite NASH (Supporting Table 2). In addition, steatosis was graded into four categories (<5%, 5%-33%, 33%-66%, and > 66%); fibrosis was scored 0-4 in analysis; inflammation was defined as an unweighted sum of lobular inflammation and portal inflammation (details in Supporting Table 2); NAFLD activity score was defined as an unweighted score of steatosis, lobular inflammation, and hepatocellular ballooning, according to the NASH clinical research networking scores and definitions. To specifically compare cholesterol metabolism in simple steatosis versus NASH we divided subjects into categories based on liver phenotype: 1) Normal liver without any steatosis, inflammation, ballooning, or fibrosis; 2) Simple steatosis (steatosis >5%) without evidence of hepatocellular ballooning, inflammation, or fibrosis; and 3) definitely NASH (see above for histological diagnosis, Supporting Table 2).
Laboratory Determinations and Calculations
Plasma glucose was measured by enzymatic hexokinase photometric assay (Konelab Systems Reagents, Thermo Fischer Scientific, Vantaa, Finland). Serum insulin was determined by immunoassay (ADVIA Centaur Insulin IRI, no 02230141, Siemens Medical Solutions Diagnostics, Tarrytown, NY). Insulin resistance index was calculated based on homeostasis model assessment (HOMA-IR). In the population study, the Matsuda insulin sensitivity index was also calculated. Cholesterol and triglycerides from serum and from lipoprotein fractions were assayed by an automated enzymatic method (Roche Diagnostics, Mannheim, Germany). Serum (all 110 participants) and liver (62 samples available) cholesterol precursors cholestenol, desmosterol, and lathosterol, which reflect whole-body cholesterol synthesis,[30, 31] were quantified with gas liquid chromatography on a 50-m long capillary column (Ultra 2; Agilent Technologies, Wilmington, DE) using 5α-cholestane as internal standard. The values are given as μg/dL and also expressed as ratios to cholesterol (102 mmol/mol cholesterol).
Liver Gene Expression
All samples for gene expression analysis were immediately frozen in liquid nitrogen. Total RNA from liver tissue was extracted using Tri-Reagent (Applied Biosystems [ABI] Foster City, CA) and reverse-transcribed using the High Capacity cDNA Reverse Transcription Kit (ABI) according to the manufacturer's protocol. Quantitative real-time polymerase chain reaction (PCR) was carried out with the Applied Biosystems 7500 Real Time PCR System using KAPA SYBR FAST qPCR Universal Master Mix (Kapa Biosystems, Woburn, MA). Primers are available by request (J.P. and D.K.).
Genotyping of PNPLA3
PNPLA3 was genotyped using the TaqMan SNP Genotyping Assay (rs738409, Applied Biosystems) according to their protocol.
Data are presented as mean ± standard deviation (SD). Nonparametric Kruskal-Wallis or Mann-Whitney tests were used in the liver biopsy study and in the population cohort to compare the study groups. The General Linear Model (GLM) was used in the liver biopsy study to evaluate the effect of statin medication and insulin resistance (HOMA-IR) on differences between study groups. Spearman's rank correlation was used for correlation analyses. Analyses were conducted with the SPSS v. 17 program (Chicago, IL). P < 0.05 was considered statistically significant.
Clinical Characteristics of Obese Individuals
To compare cholesterol metabolism between individuals with normal liver, simple steatosis, and NASH we created groups of individuals with a distinct histological phenotype (as defined in the Methods). Seventy-one obese subjects had distinct phenotypes as follows: normal liver (n = 21), simple steatosis (n = 17), and NASH (n = 23) (Table 1). The clinical characteristics of all 110 patients based on histological diagnosis, including those without clear histological diagnosis, are presented in Supporting Table 1. There were no statistically significant differences in gender distribution, age, or BMI between the phenotype groups (Table 1), or between individuals in these subgroups and those individuals that could not be categorized into these groups (n = 39; Supporting Table 1). As expected, insulin resistance (HOMA-IR) was higher in individuals with simple steatosis or NASH compared to those with normal liver (Fig. 1A). In contrast, total and low-density lipoprotein (LDL)-c levels were not elevated in individuals with simple steatosis groups but were significantly higher in individuals with NASH compared to those with a normal liver (P = 0.008) or simple steatosis (P = 0.009; Fig. 1A). Similar results were obtained after adjustment for the use of statins and HOMA-IR (Table 1).
Table 1. Clinical Characteristics (Mean ± SD) of Obese Individuals by Liver Phenotype
Levels of Serum Desmosterol Are Associated With NASH
As expected, based on earlier findings that liver steatosis is associated with increased cholesterol synthesis, we observed higher levels of serum desmosterol, a precursor of cholesterol in the cholesterol biosynthesis pathway, in individuals with NASH (P = 0.009; Fig. 1B). Interestingly, the levels of two other cholesterol precursors, cholestenol and lathosterol, were not significantly altered between the study groups (Table 1, Fig. 1B). Importantly, serum desmosterol was significantly elevated only in individuals with NASH (P = 0.002), not in individuals with simple steatosis (P = 0.289), compared to individuals with normal liver (Fig. 1B). The ratio of serum desmosterol to serum cholesterol was also higher in subjects with NASH (P = 0.003). The results remained essentially unchanged when subjects using statins (n = 30) were excluded from the analysis (Supporting Fig. 1, characteristics shown in Supporting Table 3).
Next we investigated the correlation of serum desmosterol levels with specific histopathological changes. All 110 obese individuals were included in this analysis (Table 2). Serum levels of desmosterol correlated positively with steatosis (r = 0.256, P = 0.006), fibrosis (r = 0.372, P < 0.001), inflammation (r = 0.383, P < 0.001), and NAFLD activity score (r = 0.338, P < 0.001) (Table 2). More important, the correlation with steatosis (r = 0.288, P = 0.004), fibrosis (r = 0.283, P = 0.003), and NAFLD activity score (r = 0.323, P = 0.001) was also significant for the desmosterol/cholesterol ratio, suggesting a more specific association of desmosterol with NASH compared to serum levels of total cholesterol or other markers of cholesterol synthesis. Although we had fewer men in the study, we also analyzed the data separately in men and women. The correlation of serum desmosterol with liver inflammation was significant in women (r = 0.474, P < 0.001, n = 75) and the same trend was observed in men (r = 0.289 P = 0.092, n = 35).
Table 2. Spearman's Correlations of Serum and Liver Levels of Cholesterol and Desmosterol With Liver Histology
NAFLD Activity Score
aCorrelation is significant at the 0.001 level (2-tailed).
bCorrelation is significant at the 0.01 level (2-tailed).
cCorrelation is significant at the 0.05 level (2-tailed).
To investigate potential mechanisms between serum desmosterol and NASH, we measured total cholesterol and desmosterol in liver tissue as well (available from 62 subjects not differing from the total study group in age, gender distribution, and BMI, Supporting Table 4). As expected, liver cholesterol correlated with steatosis (r = 0.353, P = 0.005), inflammation (r = 0.421, P = 0.001), and NAFLD activity score (r = 0.378, P = 0.002). The correlation of liver desmosterol with steatosis and inflammation was also significant, but of smaller magnitude (Table 2). Levels of serum and liver desmosterol correlated strongly (r = 0.667, P = 1 × 10−9; Fig. 2A), suggesting a shared regulation. Importantly, serum desmosterol levels correlated with liver cholesterol (r = 0.483, P = 7 × 10−5; Fig. 2B) more strongly than with serum cholesterol (r = 0.330, P = 0.009).
We also investigated the relationship between serum desmosterol and the expression of selected liver genes regulating cholesterol and triglyceride metabolism (available from 80 subjects not differing from the total study group in age, gender distribution, and BMI, characteristics shown in Supporting Table 4). Serum desmosterol correlated positively with the expression of SREBP1c (r = 0.328, P = 0.003, n = 80) but not significantly with SREBP1a (r = 0.199, P = 0.076). A similar trend was observed with respect to liver desmosterol levels that correlated with SREBP1c expression (r = 0.366, P = 0.024) but not with SREBP1a (r = 0.085, P = 0.602). In contrast, serum levels of cholestenol and lathosterol correlated with both SREBP1a and 1c messenger RNA (mRNA) expression (all P < 0.05; Supporting Table 5), suggesting differential roles for desmosterol and cholestenol/lathosterol in the liver. Finally, PNPLA3 genotype did not associate with markers of cholesterol synthesis in the Kuopio Obesity Surgery Study (KOBS) (P > 0.1) or in the METSIM study (P > 0.2; data not shown).
Desmosterol Correlates With ALT Levels in a Population Study
To investigate the significance of serum desmosterol at the population level we measured the levels of serum desmosterol and ALT in 717 men not using cholesterol-lowering medication. To this end, the population was divided into quartiles according to serum ALT. The strongest association of ALT was observed with obesity (BMI, P = 2 × 10−17) and insulin sensitivity (Matsuda Index, P = 3 × 10−26), most likely due to the strong correlation between liver steatosis and obesity/insulin resistance. However, the association of desmosterol levels (P = 1 × 10−12) and the desmosterol/cholesterol ratio (P = 4 × 10−10) with ALT (Fig. 3A) was stronger than that of total cholesterol, LDL cholesterol, and HDL cholesterol (P = 1 × 10−4, P = 1 × 10−3, and P = 0.547, respectively). Levels of desmosterol were higher in individuals with increased body weight, central obesity, and insulin resistance (Fig. 3B). Moreover, desmosterol levels also correlated with serum levels of interleukin 1 receptor antagonist (IL1-RA) (r = 0.157, P = 2 × 10−5), a marker of lobular inflammation and NAFLD activity score in NASH.
In this study we demonstrate that both serum and liver levels of desmosterol associate with NASH in obese individuals (Fig. 1, Table 2). This association was related to cholesterol accumulation in the liver (Fig. 2). In addition, serum desmosterol levels and the desmosterol/cholesterol ratio were associated with ALT in a random population-based sample of 717 men. The increased cholesterol synthesis in liver steatosis and the dysregulation of the cholesterol synthesis pathway in NASH have been shown in earlier studies. Our findings extend these earlier models by suggesting a more specific role of desmosterol metabolism in NASH.
Our novel finding is that serum and liver desmosterol are related to inflammation in NASH. All markers of cholesterol synthesis correlated with histological steatosis in our study (data not shown) as described earlier in a study measuring steatosis with magnetic resonance imaging (MRI). However, only serum levels of desmosterol associated with NASH (Fig. 1B). Our findings support the findings of a previous small study (n = 20) indicating that the serum desmosterol to cholesterol ratio (a marker of cholesterol synthesis) was significantly elevated in NASH. Results of other precursors were not reported in that study.
The association of serum desmosterol with NASH cannot be explained solely by increased cholesterol synthesis because other markers of cholesterol synthesis were not associated with NASH (Fig. 1B). This suggests that desmosterol may have an independent role in the pathogenesis of NASH. Desmosterol is a precursor of cholesterol in the cholesterol biosynthesis pathway. Thus, its levels could relate either to direct effects of desmosterol or reflect changes in other components of the cholesterol synthesis pathway. Desmosterol strongly activates LXR target genes in vivo[35, 36] and in a mouse model deficient for the gene coding the desmosterol reductase enzyme (DHCR24), which catalyzes the conversion of desmosterol into cholesterol.[37, 38] Our gene expression results support the view that desmosterol may have a specific role in activating, e.g., LXR target genes, as compared to cholestenol and lathosterol (Supporting Table 5). Interestingly, the DHCR24 gene function has also been associated with apoptosis and with protective responses to oxidative stress, all phenomena also important in NASH. Furthermore, HCV infection induces desmosterol accumulation and overexpression of DHCR24 in cell lines. These potential similarities in desmosterol metabolism between NASH and HCV infection are of particular interest because HCV infection is also associated with serum lipid abnormalities and liver steatosis. There are several potential direct and indirect mechanisms that need to be investigated in further experimental studies to clarify the link between desmosterol metabolism and NASH.
We acknowledge that serum cholesterol precursors are only surrogate markers of the cholesterol synthesis pathway. However, it is not feasible to measure cholesterol synthesis directly in a large population. In addition, our results suggest distinct roles for cholesterol precursors and these differences could not have been observed if only the cholesterol synthesis rate had been measured. Moreover, we carefully controlled for the treatment with statins (Table 1) and analyzed the data after excluding subjects using statins (Supporting Table 3 and Supporting Fig. 1). With respect to the population study, we also recognize that ALT is an unspecific marker of liver disease in a population. However, it is not possible to obtain liver biopsies in a large random population cohort, as was used in our study. One limitation of the population study was that all participants were men. Therefore, the results with respect to the association between serum ALT and desmosterol in the population cannot be generalized to women. In obese individuals we found a correlation between serum desmosterol and liver inflammation in women but not significantly in men, probably due to the lower number of men. However, we cannot exclude that gender would modify the association between desmosterol and NASH. Finally, the role of desmosterol as a noninvasive marker for NASH should be evaluated in a larger study with an independent histological validation cohort to test the diagnostic accuracy. Although our gene expression analysis suggests a differential role for desmosterol, as compared to cholestenol and lathosterol, we acknowledge that the analysis is not conclusive. Thus, we are currently pursuing a larger study investigating the role of cholesterol precursors in the liver.
In summary, serum and liver levels of desmosterol are associated with NASH in obese individuals. The association with liver disease was also confirmed in a large random population-based cohort by showing an association between serum desmosterol and ALT. The association of serum desmosterol with liver desmosterol, and with cholesterol accumulation in liver, suggests that serum desmosterol is a marker of disturbed cholesterol metabolism in the liver. However, a more specific role of desmosterol metabolism in NASH is also possible, as suggested in HCV.[42, 43]
We thank Päivi Turunen, Tiina Sistonen, and Matti Laitinen for their careful work in patient recruitment and laboratory analyses, and Leena Kaipiainen for the sterol analyses.
Author Contributions: M.S. researched data and wrote the article with the help of V.M. and J.L. S.V., P.K., H.G., and J.P. conducted the Kuopio Obesity Surgery Study (KOBS). H.G. was also responsible for the analysis of cholesterol precursors. D.K. performed gene expression analyses. J.K. and M.L. were responsible for the population study METSIM (Metabolic Syndrome in Men Study). J.P. was responsible for the clinical and molecular studies, researched data, and had full access to all the data to take responsibility for the integrity and the accuracy of the analyses.