Introduction
Non-alcoholic fatty liver disease (NAFLD) afflicts 25% of the world’s population and is projected to increase due to the increasing prevalence of diabetes and obesity [
1,
2]. The metabolic syndrome (MetS) comprises several disorders that raise the risk of NAFLD [
3]. MetS is identified when at least three of the following criteria are present simultaneously: (a) abdominal obesity, (b) elevated circulating triacylglycerol (TAG), (c) reduced HDL cholesterol levels, (d) hypertension, and (e) increased fasting glucose ± insulin resistance [
4]. Obesity and insulin resistance are factors that contribute to the development of MetS [
3].
Inappropriately elevated angiotensin II (Ang II) and over-activation of its target receptor, AT1, contribute to impaired hepatic lipid metabolism [
5], development of fatty liver [
6], and the development of insulin resistance [
7‐
9]. Thus, elevated Ang II is regarded as a risk factor for NAFLD [
10]. Chronic infusion of Ang II increased circulating insulin, TAG, NEFA, and liver TAG synthesis in rats [
11,
12]. Ang II-mediated activation of AT1 may overload the liver by increasing plasma NEFA, decreasing fatty acid oxidation, and promoting de novo lipogenesis [
5]. Angiotensin receptor blockers (ARB) displace Ang II from its receptor [
13], inhibiting Ang II signaling. While ARBs are widely used to ameliorate MetS-related hypertension [
14], they also improve markers of inflammation [
15,
16], components of the lipid profile, and hepatic lipid accumulation in animal studies and clinical trials [
17‐
21]. However, the mechanisms contributing to these improvements are not fully understood, especially in relationship to AT1 signaling.
The sequestration of hepatic NEFA is mainly dependent on membrane fatty acid transporters, which include the fatty acid transport protein (FATP) family [
22], cluster of differentiation 36 (CD36), and caveolins [
23,
24]. In the liver, FATP5 and FATP2 are the most abundantly expressed of the 6 members of the FATP family. Knockdown [
25] and knockout (KO) [
26] of FATPs ameliorated hepatic steatosis. These studies also demonstrated a positive correlation between FATP levels and NEFA uptake in the liver [
25,
26]. Hepatic expression of CD36 is typically low but increases with fatty liver [
27,
28]. A positive correlation between plasma insulin levels and hepatic CD36 expression was observed in insulin-resistant patients with steatosis [
29].
Recent studies suggest that peroxiredoxin 6 (PRDX6) may provide an anti-steatotic effect during fatty liver disease, mainly through elimination of oxidants [
30]. PRDX6 is a special member of its family because it is bifunctional, having both peroxidase and phospholipase A2 activity [
31]. PRDX6 protected the liver against damage and mitochondrial dysfunction induced by oxidative stress during ischemia-reperfusion [
32]. NAFLD is associated with excessive oxidant production, which is linked with impaired mitochondrial function [
33]. Yet, studies addressing hepatic PRDX6 changes in a MetS model do not exist.
Hepatic lipid accumulation is the hallmark of NAFLD [
3,
34], which may result from an imbalance between lipid acquisition and their disposal or oxidation [
27]. Nonetheless, the mechanisms that promote the development of steatosis, and the potential AT1 signaling involved, are not completely understood [
27,
35]. In the present study we investigated liver lipid substrates and proteins for NEFA uptake, TAG and VLDL cholesterol synthesis, and fatty acid oxidation in the liver of rats treated with ARB and their changes in response to an acute glucose challenge, as well as the potential effects of the removal of ARB treatment.
Methods
All experimental procedures were reviewed and approved by the institutional animal care and use committees of the Kagawa Medical University (Japan) and the University of California, Merced (USA). Phenotypical data (i.e., body mass, oral glucose tolerance test curves, insulin levels, and blood pressure) has been previously published [
36] with that study focused on redox signaling in the heart in response to a glucose challenge. The current study is unique and complements the previous one by extending the examination of the dynamic changes in hepatic lipid metabolism in response to a glucose challenge, which represents a nutrient overload that may appear in Western diets [
37,
38]. We have included previously published [
36] body mass and plasma angiotensin II levels (Table
1) to better illustrate our model and the efficacy of the ARB treatment. Additionally, we examine the potential detriments of non-compliance with ARB treatment after the initial treatment has stopped (legacy effect) [
39]. Low compliance has proven to be a challenge when treating patients with hypertension, dyslipidemia, and diabetes [
40,
41] and assessing ARB-legacy effects may provide a better understanding of the detriments to removal of treatment [
42,
43].
Table 1
Mean (±SD) end of study body mass and plasma angiotensin II
Body Mass (g) | 465 ± 29 | 610 ± 31a | 598 ± 36a | 602 ± 28a |
Plasma AngII (fmol/ml) | 58.9 ± 3.1 | 68.6 ± 11.6 | 211.3 ± 27.9a,b | 65.2 ± 10.2c |
In this study, we approach the obtained results from two perspectives: (1) the static outcomes following the chronic treatment with ARB and its removal (potential legacy effect) and (2) the dynamic response to a glucose challenge following chronic treatment with ARB and its removal (non-compliance). In all cases, comparisons are made to a lean, strain-control.
Animals
OLETF rats are reported to develop insulin resistance and hyperglycemia by 17 weeks of age [
44,
45]. For that reason, male, age matched, 17-week-old, lean strain-control Long Evans Tokushima Otsuka (LETO; 428 ± 8 g) and obese Otsuka Long Evans Tokushima Fatty (OLETF; 536 ± 6 g) rats (Japan SLC Inc., Hamamatsu, Japan) were used. Rats were assigned to the following groups (
n = 5–7 animals/group/time point): (1) untreated LETO, (2) untreated OLETF, (3) OLETF + angiotensin receptor blocker (
ARB; 10 mg olmesartan/kg/d × 8 weeks) [
46], and (4) OLETF ± ARB (
MINUS; 10 mg olmesartan/kg/d × 4 weeks, then removed the last 4 wks prior to dissection). ARB (Daiichi-Sankyo, Tokyo, Japan) was administered by oral gavage suspended in carboxymethyl cellulose (CMC) to conscious rats and untreated rats were gavaged with CMC only. Animals were maintained in groups of two to three animals per cage, given access to water and standard laboratory chow (MF; Oriental Yeast Corp., Tokyo, Japan), and maintained under controlled temperatures (23–24 °C) and humidity (~55%) with a light-dark cycle of 12–12 h.
For dissections, the 25-week-old rats were fasted for 12 h ± 15 min. To investigate the dynamic response to a glucose challenge, animals were dissected at baseline (T0, fasting) and after 3 (T3) and 6 h (T6) after a glucose load by gavage (2 g glucose/kg mass). Comparisons of the T0 (fasting baseline) data from each group characterized the static changes in response to chronic ARB treatment and its removal. The glucose challenge was performed to evaluate the acute, dynamic changes in metabolism and cellular responses. Initiation of the overnight fasts and the glucose gavages were staggered to meet the exact dissections timepoints. Animals were decapitated and trunk blood collected in vials containing EDTA (Sigma-Aldrich, EDS) and proteinase inhibitor cocktail (Sigma-Aldrich, P2714). Livers were snap frozen in liquid nitrogen and kept at −80 °C until analyzed.
Biochemical analyses & markers of hepatic NAFLD
Insulin, glucose, plasma total cholesterol (TC), TAG, and NEFA concentrations as well as collagen type IV (COL-4) were measured using the commercially available reagents: Insulin Rat ELISA Kit (Thermo Fisher Scientific, ERINS), Autokit glucose (Fujifilm Wako Diagnostics, 997-03001), Total Cholesterol E (Fujifilm Wako Diagnostics, 999-02601), L-Type Triglyceride M (Fujifilm Wako Diagnostics, 994-02891 and 990-02991), HR Series NEFA-HR (2) (Fujifilm Wako Diagnostics, 999-34691, 995-34791, 991-34891, and 993-35191), and CIV ELISA (MyBioSource, MBS732756), respectively, following manufacturer’s instructions. All samples were analyzed in duplicate and only accepted values that fell within a percent coefficient of variability of less than 10% for all measurements.
Western blot analyses
A 25 mg piece of frozen liver was homogenized in phosphate buffer for a two-step extraction of cytoplasm and plasma membrane proteins. Briefly, phosphate buffer (50 mM potassium phosphate) (Fisher Scientific, P290 and P285) was used to homogenize the liver, then centrifuged at 15,000 × g to recover the supernatant containing the cytoplasmic fraction. Then, 50 mM potassium phosphate buffer +1% Triton X-100 (Millipore-Sigma, T8787) was used to homogenize the pellet, which contained the plasma membrane fraction. The pellet homogenate was centrifuged at 15,000 × g, and the plasma membrane contents recovered from the supernatant. The buffers contained 3% protease inhibitor cocktail (Sigma-Aldrich, P2714) to help prevent protein degradation. The protein content of the fractions was quantified using the Bradford assay (Bio-Rad Laboratories, 5000203). Total protein (5–10 μg) was resolved in 10% Tris-HCL SDS gels. Proteins were electroblotted onto 0.45-μm polyvinyl difluoride (PVDF) (Millipore-Sigma, IPVH00010) membranes by semi-wet transfer using the Mini Gel Tank and Blot Module Set (Invitrogen, NW2000). Intercept blocking buffer (Li-Cor, 927-60001, 927-70001) was used to block the membranes then incubated 16 h with the corresponding primary antibody (diluted 1:1,000-1:2,000) against glycerol-3-phosphate acyltransferase 1 (GPAM) (abcam, ab69990), FATP5 (Invitrogen, PA5-42028), diacylglycerol O-acyltransferase 1 (DGAT1) (Thermo Fisher, PA5-79150), ApoB (Apolipoprotein B) (Thermo Fisher, PA5-86950), CD36 (Thermo Fisher, PA1-16813), FATP2 (Thermo Fisher, PA5-30420), carnitine palmitoyl transferase I (CPT1A) (Proteintech, 15184-1-AP), Peroxisomal acyl-coenzyme A oxidase 1 (Acox1) (Thermo Fisher, PA5-76341), PRDX6 (Proteintech, 13585-1-AP), and β actin (Cell Signaling Technology, 3700S) (diluted 1:5,000). Membranes were washed, incubated with IRDye 800CW anti-rabbit (Li-Cor, 926-32213) and/or 680RD donkey anti-mouse IgG secondary antibodies (Li-Cor, 926-68072) (diluted 1:20,000), and rewashed. Blots were visualized using the Odyssey system (Li-Cor) and quantified using the Image Studio Lite ver. 5.2 (Li-Cor) using β actin as a loading control. Plasma membrane and cytosolic fractions were tested for purity against Na+/K+ ATPase antibody (Abcam, ab76020) (diluted 1:40,000) and alpha tubulin (Abcam, ab52866) (diluted 1:40,000), respectively.
Statistics
Data was tested for normality using the Shapiro–Wilk test [
47]. Means ± standard deviation (SD) were compared by two-way ANOVAs when analyzing datasets with all timepoints (T0, T3, T6) and groups, and one-way-ANOVA when analyzing datasets with only basal levels (T0 only). Means were considered significantly different at
p < 0.05 using Tukey’s HSD. Correlations were calculated using the Pearson r coefficient [
48,
49], using the means and SD obtained from each group and timepoint, computed using the displayed individual values in each figure. Area under the curve (AUC) analyses were calculated using the area under the concentration curve in batch designs [
50]. Outliers were calculated and removed using the ROUT test [
51] and one outlier was replaced with the mean of the corresponding group [
52]. All statistical analyses were performed with GraphPad Prism 8.4.3 software (GraphPad Prism, La Joya, CA).
Discussion
NAFLD afflicts 25% of the world’s population [
1]. Hepatic TAG accumulation is the hallmark of NAFLD [
3,
34], with lipid accumulation in the liver resulting from an imbalance between lipid sequestration and its disposal or metabolism [
27]. Nonetheless, the mechanisms promoting steatosis in MetS [
27,
35,
75] in relation to Ang II signaling [
5] are not completely understood. In our study, we investigated the effects of chronic AT1 blockade statically and the dynamic responses to a glucose challenge on hepatic lipid accumulation and proteins mediating NEFA uptake, TAG and VLDL cholesterol synthesis, and fatty acid oxidation in rats with MetS. Additionally, we investigated the legacy-effect of ARB treatment following its removal. We found a marked decrease on liver TAG that may be achieved by modulating NEFA uptake, through CD36, and increased TAG export via ApoB. The inverse relationship between PRDX6 abundance and hepatic TAG and the increasing levels of PRDX6 in response to the glucose challenge when hepatic TAG levels are decreasing suggests that PRDX6 may protect the liver from steatosis derived from TAG accumulation. Furthermore, our results are unique as they highlight the detrimental effects of treatment non-compliance through many contrasts between the ARB and MINUS groups.
OLETF rats are characterized by elevated Ang II before the onset of insulin resistance [
76], suggesting that elevated Ang II may contribute to the development of insulin resistance [
7]. Liver-specific deletion of the Ang II receptor (AT1) reduces hepatic steatosis [
63], evidencing the contribution of AT1 signaling on hepatic lipid accumulation. TAG accumulation is a hallmark of NAFLD [
3] and NEFA can be a precursor for TAG formation [
5], while VLDL cholesterol can be assembled into larger molecules that may not leave the hepatocyte [
72]. In the liver, ARB decreased basal NEFA and TAG indicative of an improvement in liver lipid metabolism in MetS. Conversely, hepatic TAG, NEFA, and TC were increased in the MINUS group during the glucose challenge, suggesting that non-compliance is associated with hepatic lipid accumulation in response to a glucose load, consistent with a Western diet. Thus, while most of the ARB-mediated benefits remained statically, they were lost during the glucose challenge.
Hepatic expression of CD36 is typically low although it increases with fatty liver disease [
27,
28]. In our study, hepatic CD36 levels in LETO were remarkably lower than all OLETF groups and remained relatively low throughout the glucose challenge. ARB treatment decreased CD36 abundance and even after its removal the levels remained lesser than OLETF. Furthermore, fasting plasma insulin and basal CD36 membrane protein abundance were strongly correlated, consistent with the data in patients with steatosis [
29] suggesting that the hyperinsulinemia associated with MetS contributes to hepatic lipid accumulation and dysregulation of lipid metabolism via up-regulation of hepatic CD36.
CD36 protein abundance was increased in the myocardium of hyperglycemic mice [
77] and vascular lesions of hyperglycemic patients [
78], suggesting that elevated glucose may also up-regulate the abundance of CD36. In the present study, liver CD36 was greater in ARB, ending the glucose challenge. The lack of a detectable changes CD36 in response to the glucose challenge in LETO suggests that, during healthy conditions, CD36 protein levels are maintained despite the hyperglycemia. Additionally, CD36 levels decrease initially at 3 h post-glucose in OLETF, ARB and MINUS, increasing further only in ARB at 6 h post-glucose, when elevated glucose levels have cleared circulation, suggesting that hepatic CD36 may not be stimulated by hyperglycemia during MetS. Thus, the changes in CD36 abundance in response to hyperglycemic conditions may be tissue-specific.
The FATP family facilitates the transport of NEFA into the cell. FATP5 and FATP2 are the most abundant in the liver [
22]. Knockdown and KO studies of hepatic FATP proteins are correlated with NEFA uptake [
25,
26,
79]. ARB treatment did not change the basal membrane abundance of either protein, but FATP2 abundance decreased at 6 h post-glucose in ARB, while CD36 was elevated in the same group, suggesting that increased CD36 may compensate for the decrease in FAT proteins in the liver. Non-compliance (MINUS) increased the FATPs expression, suggesting that the potential to sequester NEFA is increased, supported by its greater liver NEFA AUC.
Ang II-mediated signaling can increase hepatic lipid-overload by decreasing hepatic fatty acid oxidation [
5]. Acox1 is a rate-limiting enzyme in peroxisomal lipid β-oxidation [
67]. Acox1-deficient rodents exhibit sudden steatosis [
80]. On the other hand, CPT1A converts fatty acyl-CoA to fatty acyl-carnitine for subsequent β-oxidation [
81]. Human CPT1 expression is reduced during NAFLD [
82], suggesting that decreased β-oxidation may be an important factor in the development of steatosis. In our study, ARB treatment did not change Acox1 or CPT1A protein abundance; however, during the glucose challenge, Acox1 abundance was greater in MINUS than OLETF and ARB, while at the end of the challenge, Acox1 abundance was lesser than OLETF and ARB. These changes suggest that the removal of ARB may have changed the sensitivity of Acox1 regulation to glucose. On the other hand, glucose induced an increase in CPT1A in the healthy LETO, which remained elevated after 6 h, suggesting that under normal conditions a healthy liver may increase lipid oxidation to help reduce TAG and NEFA accumulation in response to a glucose load. CPT1A levels in the OLETF, ARB and MINUS groups were substantially lower than LETO, however, they increased in the OLETF and ARB groups at 6 h post-glucose, while MINUS levels remained suppressed. In MINUS, the decreasing trend (
p < 0.07) in CPT1A, but not Acox1, abundance suggests that an acute glucose load may reduce the hepatic β-oxidation capabilities and that CPT1A abundance may be more sensitive to a glucose load, following non-compliance.
Collectively, these data suggest that: (1) under normal, healthy conditions the potential to increase β-oxidation may be primarily via CPT1-mediated mechanisms, (2) MetS may be associated with impaired CPT1A-mediated β-oxidation, which is supported by the lack of CPT1A abundance increase during the glucose challenge in the OLETF, ARB and MINUS groups, and (3) that non-compliance (MINUS) may be associated with impaired β-oxidation via CPT1A to a greater extent than untreated (OLETF) conditions.
During insulin resistance, hyperinsulinemia may increase lipogenesis and TAG accumulation [
83,
84]. Insulin can increase activation of the liver-X-receptor to promote hepatic lipogenesis [
68]. GPAM [
70] and DGAT [
71] participate in the synthesis of TAG. Hepatic GPAM expression was elevated in patients with steatosis or NASH [
85]. In our study, plasma insulin AUC was lower with ARB treatment, which corresponded with lower liver TAG AUC, yet GPAM abundance remained unchanged. This may be due to GPAM activity being regulated, more so than its abundance. On the other hand, DGAT deficiency in primary hepatocytes protected against increased lipid deposition by decreasing TAG synthesis [
86]. Also, inhibition of DGAT1 protected against fatty liver in mice on a high-fat diet [
87]. Yet during lipodystrophy, increased de novo hepatic fatty acid synthesis caused steatosis, independent of changes in DGAT1 [
88], suggesting that static changes in DGAT1 may not have a significant contribution to TAG synthesis. The increasing trend (
p < 0.07) in GPAM abundance in response to glucose in OLETF suggests that high glucose-loads may worsen hepatic lipid accumulation by stimulating TAG synthesis. Alternatively, the measures of GPAM and DGAT protein abundance may not be enough to accurately reflect changes in the activity of these enzymes [
88,
89]. The maintenance of these relatively higher levels of hepatic TAG in the OLETF and MINUS groups may be a consequence of chronically elevated TAG synthesis, which may not be stimulated by the glucose load used in this study.
Alternatively, hepatic lipids can be mobilized through the secretion of VLDLs [
72], which are a primary vehicle to transport synthesized TAG to circulation for utilization in peripheral tissues. Each VLDL particle contains one molecule of ApoB [
73], suggesting that ApoB abundance is a reliable surrogate measure for relative changes of VLDL in the liver. Secretion of VLDL-TAG is increased in patients with NAFLD [
90]. Also, it is suggested that VLDL particles from individuals with NAFLD may be larger and contain more TAG [
72,
91], preventing them from leaving the cell. Therefore, the differences in particle size cannot be excluded as a contributing factor to the observed accumulation of hepatic TAG. In our study, ApoB abundance increased in OLETF and was decreased in ARB. This suggests that the elevated levels in OLETF may be necessary to bind the greater amount of free TAG, increasing the potential for shuttling it out of the cell. Alternatively, the increased levels of ApoB in OLETF may reflect accumulation due to large particles of VLDL, which cannot be shuttled out of the liver [
72].
In response to the glucose challenge, ApoB abundance increased over time in ARB. This may be due to the increased synthesis of VLDL to further decrease TAG accumulation in the liver of the ARB group. This is further supported by increased plasma TAG, which translates to elevated circulating VLDL, since VLDL accounts for ~20% of the measured plasma TAG [
92,
93]. Conversely, the low abundance of ApoB in MINUS at 6 h post-glucose suggests that the potential to shuttle TAG out of the cell in response to glucose may be impaired, and that the increased plasma TAG may come mainly from diet and increased lipolysis, contributing to greater levels of hepatic TAG, reflecting the detrimental effect of treatment removal. Ultimately, these data provide an additional mechanism by which hepatic TAG is regulated in response to nutrient loads.
Induction of CD36 with palmitic acid induced hepatocyte activation dependent on oxidative stress pathways, while CD36 KO reduced these adverse effects [
24], suggesting that changes in CD36-mediated NEFA transport are associated with changes in redox balance in the liver. However, these relationships are not well-defined during MetS. PRDX6 is member of the PRDX family of proteins, which may protect against obesity-related pathologies, mainly through elimination of oxidants [
94‐
96]. The C57BL/6J‐Tg transgenic mice with increased PRDX6 expression prevented the liver from developing steatosis [
30], suggesting that PRDX6 may contribute directly to the regulation of hepatic lipid accumulation. PRDX6 KO mice on a high-fat diet increased levels of circulating alanine aminotransferase (ALT), a marker of hepatic injury that is associated with development of NASH [
74]. In our study, basal PRDX6 abundance was greater in OLETF, while ARB treatment reduced this increase, suggesting that ARB treatment improved the redox status in the liver, and that increased levels of hepatic PRDX6 may reflect the need for improved redox balance in the liver. Glucose stimulated a trend (
p < 0.06) for linear increase in the ARB group, suggesting that a glucose load may stimulate redox gene transcription factors to initiate antioxidant mechanisms in response to increased glucose levels. PRDX6 and liver TAG levels were negatively correlated in the ARB group, suggesting that either: (a) increasing hepatic TAG accumulation may decrease PRDX6 levels or (b) increasing PRDX6 reduces hepatic TAG accumulation. Yet, very limited information is available regarding the relationship between PRDX6 and liver TAG levels. Therefore, in the current study, deciphering which is the independent and dependent variable is not possible, though the novel finding is that there is a negative relationship between these two variables that has not been previously reported in a model of MetS. Although the negative correlation between PRDX6 and liver TAG can be thought as isolated, it may be meaningful for future research linking PRDX6 and NAFLD, where literature on this topic is scarce. Although the link between NAFLD and PRDX6 remains unclear [
30,
74], our data supports the idea that PRDX6 may participate in protecting the liver from TAG accumulation.
In summary, these results demonstrate that chronic blockade of AT1, through ARB treatment, protects the liver from TAG accumulation, especially during a glucose load. This may be achieved by decreasing NEFA uptake and increasing TAG export via CD36 and FATP2, and ApoB, respectively. Additionally, treatment non-compliance reverted many of the potential benefits observed in ARB, which may leave the liver more susceptible to further lipid accumulation and injury over time. Finally, frequent acute glucose loads may contribute to increased hepatic lipid accumulation through the maintenance of TAG synthesis and impaired β-oxidation and cellular lipid export, to ultimately develop NAFLD in MetS.