Coenzyme Q10 prevents hepatic fibrosis, inflammation, and oxidative stress in a male rat model of poor maternal nutrition and accelerated postnatal growth1

Background: It is well established that low birth weight and accelerated postnatal growth increase the risk of liver dysfunction in later life. However, molecular mechanisms underlying such developmental programming are not well characterized, and potential intervention strategies are poorly defined. Objectives: We tested the hypotheses that poor maternal nutrition and accelerated postnatal growth would lead to increased hepatic fibrosis (a pathological marker of liver dysfunction) and that postnatal supplementation with the antioxidant coenzyme Q10 (CoQ10) would prevent this programmed phenotype. Design: A rat model of maternal protein restriction was used to generate low-birth-weight offspring that underwent accelerated postnatal growth (termed “recuperated”). These were compared with control rats. Offspring were weaned onto standard feed pellets with or without dietary CoQ10 (1 mg/kg body weight per day) supplementation. At 12 mo, hepatic fibrosis, indexes of inflammation, oxidative stress, and insulin signaling were measured by histology, Western blot, ELISA, and reverse transcriptase–polymerase chain reaction. Results: Hepatic collagen deposition (diameter of deposit) was greater in recuperated offspring (mean ± SEM: 12 ± 2 μm) than in controls (5 ± 0.5 μm) (P < 0.001). This was associated with greater inflammation (interleukin 6: 38% ± 24% increase; P < 0.05; tumor necrosis factor α: 64% ± 24% increase; P < 0.05), lipid peroxidation (4-hydroxynonenal, measured by ELISA: 0.30 ± 0.02 compared with 0.19 ± 0.05 μg/mL per μg protein; P < 0.05), and hyperinsulinemia (P < 0.05). CoQ10 supplementation increased (P < 0.01) hepatic CoQ10 concentrations and ameliorated liver fibrosis (P < 0.001), inflammation (P < 0.001), some measures of oxidative stress (P < 0.001), and hyperinsulinemia (P < 0.01). Conclusions: Suboptimal in utero nutrition combined with accelerated postnatal catch-up growth caused more hepatic fibrosis in adulthood, which was associated with higher indexes of oxidative stress and inflammation and hyperinsulinemia. CoQ10 supplementation prevented liver fibrosis accompanied by downregulation of oxidative stress, inflammation, and hyperinsulinemia.


INTRODUCTION
In 1992, Hales and Barker (1) proposed the "thrifty phenotype hypothesis," which postulated that in response to suboptimal in utero nutrition, the fetus alters its organ structure and adapts its metabolism to ensure immediate survival. This occurs through the sparing of vital organs (e.g., the brain) at the expense of others, such as the liver, thus increasing the risk of metabolic disease such as liver dysfunction in later life (2)(3)(4). This risk is exacerbated if a suboptimal uterine environment is followed by rapid postnatal growth (5,6).
Nonalcoholic fatty liver disease (NAFLD) 4 is the hepatic manifestation of the metabolic syndrome. Aspects of the metabolic syndrome, including NAFLD, have been linked to exposure to suboptimal early-life environments (7,8). incidence of NAFLD is high (9), the associated morbidity is low if there is no progression to hepatic fibrosis. Progression to fibrosis is indicative of the clinically important subtype of patients with NAFLD who have a high chance (20%) of developing frank liver cirrhosis and subsequent liver failure (10). At present, it is unknown why this progression occurs only in a subset of individuals. The development of an intervention that prevents these changes from accumulating could improve the prognosis of patients who develop NAFLD later in life.
Increased oxidative stress is a common consequence of developmental programming (11). Increased reactive oxygen species (ROS) have been strongly implicated in the etiology of hepatic fibrosis. Several animal studies have focused on antioxidant therapies to prevent the deleterious phenotypes of developmental programming (12)(13)(14); however, the doses used are not recommended for use in humans. In practice, a suboptimal intrauterine environment is often recognized retrospectively (i.e., after delivery). Thus, it is important to address potential beneficial effects of targeted postnatal interventions.
Coenzyme Q (CoQ 10 ) is a benzoquinone ring linked to an isoprenoid side-chain. The isoform containing 9 isoprenoid units (CoQ 9 ) is most abundant in rodents, whereas CoQ 10 (10 isoprenoid units) is the most common in humans. When oxidized, CoQ 10 shuttles electrons between mitochondrial complexes I and III and complexes II and III. Reduced CoQ 10 is the most abundant endogenous cellular antioxidant (15) and is a safe and effective therapeutic antioxidant (16,17). We have also shown that postnatal CoQ 10 supplementation prevents developmentally programmed accelerated aging in rat aortas (18) and hearts (19).
Liver is one of a few tissues to take up dietary CoQ 10 (20), and CoQ 10 supplementation has been previously investigated as a potential therapy to prevent the progression of NAFLD in humans (21). In this study, we aimed to 1) investigate the effects of poor maternal nutrition and rapid postnatal catch-up growth on hepatic CoQ 9 concentrations and molecular pathways leading to proinflammatory changes and development of fibrosis and 2) determine whether a clinically relevant dose of dietary CoQ 10 could correct any observed hepatic fibrosis.

Animal experimentation
All procedures involving animals were conducted under the British Animals (Scientific Procedures) Act (1986) and underwent ethical review by the University of Cambridge Animal Welfare and Ethical Review Board. Stock animals were purchased from Charles River, and dams were produced from inhouse breeding from stock animals.
Pregnant Wistar rats were maintained at room temperature in specific-pathogen-free housing with the use of individually ventilated cages with environmental enrichment. The dams were maintained on a 20% protein diet (control) or an isocaloric lowprotein (8%) diet, as previously described (22). Access to diets and water was provided ad libitum. All rats used in this study were specific-pathogen-free housed individually at 228C on a controlled 12:12-h light-dark cycle. Diets were purchased from Arie Blok.
The day of birth was recorded as day 1 of postnatal life. Pups born to low-protein-diet-fed dams were cross-fostered to con-trol-fed mothers on postnatal day 3 to create a recuperated litter. Each recuperated litter was standardized to 4 male pups at random to maximize their plane of nutrition. The control group consisted of the offspring of dams fed the 20%-protein diet and suckled by 20% protein-fed dams. Each control litter was culled to 8 pups as a standard. To minimize stress to the pups when cross-fostered, they were transferred with some of their own bedding. Body weights were recorded at postnatal days 3 and 21 and at 12 mo. At 21 d, 2 males per litter were weaned onto standard laboratory feed pellets (Special Diet Services) and the other 2 were weaned onto the same diet supplemented with CoQ 10 to give a dose of 1 mg/kg body weight per day. Diets were given in the home cage. Rat pups were fed these diets until 12 mo of age, at which time all rats were killed by carbon dioxide asphyxiation at w1000. Post mortem, liver tissue was removed, weighed, and snap-frozen in liquid nitrogen and then stored at 2808C until analysis. A further portion of liver tissue (the same area for each sample) was removed and fixed in formalin for histologic assessment.
For all measurements, 1 pup per litter was used, thus "n" represents number of litters. Ten litters per group were used in this study based on power calculation with the use of previous data from our studies of RNA expression in postnatal tissues from programmed animals. Rat numbers were calculated to give 80% power to detect a 20% difference between groups at the P , 0.05 level. Only male rats were used in this study.

CoQ 10 diet preparation
A dose of 1 mg CoQ 10 /kg body weight per day was used in this study, which was administered via the diet (18). This was achieved by appropriate CoQ 10 supplementation of laboratory feed pellets, as we described previously (18,19). Diet was prepared twice a week throughout the study.
CoQ 10, lipid profile, glucose, and insulin analysis Total liver ubiquinone (CoQ 9 and CoQ 10 ) was quantified by reverse-phase HPLC with UV detection at 275 nm, as described previously (18,19). Serum was obtained as detailed previously (18), and blood from the tail vein collected into EDTA tubes and centrifuged for 3 min at 3000 rpm at 48 Celsius to isolate plasma. Fasted blood glucose measurements were obtained by using a glucose analyzer (Hemocue). The serum lipid profile and fasted plasma insulin analyses were performed by using an auto-analyzer (the Wellcome Trust-supported Cambridge Mouse Laboratory). Liver triglyceride concentrations were determined by using the Folch assay (23). Briefly, liver samples were homogenized in a 2:1 ratio of chloroform:methanol. The distinct lipid phase was removed after centrifugation, and lipid weight was quantified after the solvent was removed by evaporation.

Histologic assessment
Liver samples were fixed in formalin, paraffin-embedded, and sectioned to a 6-mm thickness by using a microtome. Picro Sirius Red staining was used to stain for fibrosis. Cell-D software (Olympus Soft Imaging Solutions) was used to quantify the thickness of fibrosis around all visible hepatic vessels (including all arteries and veins) from 1 section per sample. This sample was taken at the same point (20 sections for each sample) by using a nonbiased grid sampling method. All analyses were performed at 103 magnification by using an Olympus microscope (Olympus Soft Imaging Solutions). All analyses were performed blinded.

Gene expression
RNA was extracted by using a miReasy mini kit (Qiagen) following the manufacturer's instructions, as detailed previously (19). A DNase digestion step was performed to ensure no genetic DNA contamination. RNA (1 mg) was used to synthesize complementary DNA by using oligo-dT-adaptor primers and Moloney murine leukemia virus reverse transcriptase (Promega). Gene expression was determined by using customdesigned primers (Sigma) and SYBR Green reagents (Applied Biosystems). Primer sequences are presented in Table 1. Quantification of gene expression was performed with the use of a Step One Plus reverse transcriptase-polymerase chain reaction machine (Applied Biosystems). Equal efficiency of the reverse transcription of RNA from all groups was confirmed through quantification of expression of the housekeeping gene b-actin (Actb). The expression of Actb did not differ between groups (effect of maternal diet, P = 0.9; effect of CoQ 10 supplementation, P = 0.8; control: 153 6 32; recuperated: 144 6 24; control CoQ 10 : 143 6 12; and recuperated CoQ 10 : 157 6 21 average copy numbers).

Mitochondrial DNA copy number
Total DNA was extracted using a phenol/chloroform extraction protocol (24). Mitochondrial DNA copy number analysis was performed as described previously (25).

4-Hydroxynonenal and 3-nitrotyrosine analysis
Protein nitrotyrosination was assayed by using a Nitrotyrosine ELISA kit (MitoSciences), according to the manufacturer's instructions. Lipid peroxidation was analyzed by using an Oxi-Select HNE Adduct ELISA kit (Cambridge Biosciences), according to the manufacturer's instructions.

Statistical analysis
Data were analyzed by using a 2-factor ANOVA with maternal diet and CoQ 10 supplementation as the independent variables. Post hoc testing was carried out when appropriate and is indicated in the text accordingly. Data are represented as means 6 SEMs. All statistical analyses were performed with the use of Statistica 7 software (Statsoft), and for all tests, P values ,0.05 were considered significant. Data were checked for normal distribution. In all cases, "n" refers to the number of litters (with 1 rat used from each litter).

Anthropometric measurements
Recuperated offspring were born smaller than control rats (6.3 6 0.2 compared with 7.4 6 0.2 g; P , 0.001) and underwent rapid postnatal catch-up growth so that their weights were similar to those of the control offspring at postnatal day 21 (52.2 6 0.9 compared with 50.7 6 1.2 g). At 12 mo of age, there was no effect of maternal diet or CoQ 10 supplementation on body weights or absolute liver weights ( Table 2).
Dietary CoQ 10 supplementation leads to greater hepatic CoQ 9 and CoQ 10 concentrations Recuperated hepatic CoQ 9 and CoQ 10 concentrations were unaltered compared with those in control rats ( Table 3). However, CoQ 9 and CoQ 10 concentrations were greater (P , 0.01) when supplemented with CoQ 10 (Table 3).

DISCUSSION
Our findings that a maternal low-protein diet in utero followed by accelerated postnatal growth (recuperated) confers a higher risk of oxidative damage, proinflammatory changes, and liver fibrosis suggest that the early environment is an important determinant of an individual's risk of developing complications of fatty liver disease. The potential for progression from NAFLD to nonalcoholic steatohepatitis and finally to hepatitis, fibrosis, and cirrhosis is well described. However, it has not previously been possible to identify patients who are at high risk of these changes. Our finding that recuperated rats developed more hepatic fibrosis than did controls indicates that the early environment plays a central role in the risk of liver disease in later life. Identifying the early environment as influential in the propensity to hepatic inflammation and fibrosis provides valuable new insight into predetermining individuals at increased risk from hepatic manifestations of the metabolic syndrome.
HSCs are the main source of extracellular matrix formation during hepatic fibrosis (26). The key step in inducing fibrosis during liver injury is the transformation of quiescent HSCs to activated HSCs, which differentiate into myofibroblasts (26). We found increased expression of Acta2 and Des in recuperated offspring.
Acta2 is expressed in myofibroblasts of damaged liver (27) and hence is a good marker of HSC activation. In rats, HSC activation and proliferation correlate with a high expression of Des and are Values are as means 6 SEMs; n = 10/group; 10/10 rats used. * , ***C compared with R and R compared with RQ: *P , 0.05, ***P , 0.001. Interaction between in utero protein restriction, accelerated postnatal growth, and CoQ 10 supplementation (statistical interaction): P = 0.001 (A and F). Data were analyzed by using 2-factor ANOVA and Duncan's post hoc testing, where appropriate. C, control; Col1a1 = collagen type 1, a1; CoQ 10 , coenzyme Q; CQ, control CoQ 10 ; R, recuperated; RQ, recuperated CoQ 10 . found in HSCs of acutely injured liver (27). KC infiltration and activation play a prominent role in HSC activation. Increased KC infiltration coincides with the activation of HSC markers such as Acta2 (28). Cle4f (a unique KC receptor for glycoproteins and therefore a good marker of KC activation) was increased in livers of recuperated offspring.
Increased concentrations of proinflammatory cytokines are also crucial in initiating HSC activation. Hepatic protein expression of Tnf-a and Il-6 was greater in recuperated offspring, suggesting that inflammation plays a role in the HSC activation and consequent hepatic fibrosis observed in our model. Tgfb1 mRNA levels were not changed in recuperated offspring; however, we cannot discount the possibility that Tgfb1 is upregulated at the protein level. Mcp1 (a chemokine that acts as a chemoattractant for HSCs) was also upregulated in recuperated livers. Mmps are also associated with hepatic fibrosis (29). Mmp9 expression was upregulated in recuperated livers. Mmp9 is prominent in scar areas of active fibrosis, and treatment with a profibrotic agent can increase its expression, with peak expression coinciding with induction of inflammatory cytokines (29). Mmp2 expression was unaltered between groups. Because Mmp expression is an early event in wound healing, the time window for Mmp2 elevation may have been missed and that difference would be observed only in younger animals.
A further driving factor in HSC activation and fibrosis is increased ROS, which can be generated by Tnf-a, Il-6 (30), KCs (28), and the mitochondrial electron transport chain. We found increased ROS in the context of increased lipid peroxidation [which is known to increase in liver disease (31)] and greater expression of the NOX-2 components (Gp91 phox , P22 phox , P47 phox , and P67 phox ), a major source of hepatic ROS production, which have been observed in hepatic fibrosis (32,33). Complex I activity, a predominant generator of ROS (34), was greater in recuperated livers. Decreased antioxidant defense capacity was evidenced by a reduction in Gpx1, a peroxidase responsible for the conversion of H 2 O 2 into H 2 O and O 2 . Increased concentrations of cellular H 2 O 2 , due to Gpx1 depletion, could cause accumulation of the hydroxyl anion, a free radical that can directly increase lipid peroxidation.
Accumulation of hepatic triglycerides also plays a role in hepatic fibrosis (35); however, neither liver nor plasma triglycerides were altered in recuperated rats. This may be explained by the fact that recuperated offspring are fed a feed pellet diet and do not display an obesogenic phenotype. This in itself is interesting, because it shows that the observed deleterious liver phenotypes develop in a physiologic environment that had been influenced only by developmental programming per se, and not by obesity.
Insulin resistance and hyperinsulinemia are also major contributors to liver fibrosis (36) and are inherently linked to increased oxidative stress. Recuperated offspring had wholebody insulin resistance as indicated by hyperinsulinemia. The hyperinsulinemia was associated with hepatic insulin signaling protein dysregulation, as shown by the downregulation of IR-b, IRS-1, and Akt-1.
CoQ 10 supplementation also prevented hyperinsulinemia in recuperated rats; however, hepatic insulin signaling protein dysregulation was not normalized by CoQ 10 supplementation. The whole-body insulin sensitivity may therefore be improved through other mechanisms such as improvements in muscle and/or adipose tissue insulin sensitivity. CoQ 10 may exert antifibrotic effects through the activation of the Nrf2/antioxidant response element (Nrf2/ARE) pathway. Nrf2 is a transcription factor that responds to oxidative status and regulates transcription of genes involved in antioxidant defense. CoQ 10 treatment in a model of hepatic fibrosis ameliorates liver damage via suppression of Tgfb1 and upregulation of Nrf-ARE-associated genes (40). Although Nrf2 expression was not affected by maternal diet, CoQ 10 supplementation increased Nrf2 by 4-fold. The antioxidant genes involved in the Nrf2/ARE pathway-Hmox1, Gst, and Gpx1-were increased by CoQ 10 supplementation. This suggests that CoQ 10 supplementation upregulates the Nrf2/ARE pathway via suppression of Tgfb1 (40). These observations support a role for CoQ 10 supplementation in increasing antioxidant defenses to a protective level in animals that have experienced detrimental catch-up growth (18,19). Because Nqo1 activity is known to prevent 1 electron reduction in quinones, it is plausible that because hepatic CoQ 10 concentrations are elevated by CoQ 10 supplementation, Nqo1 expression is not required and thus suppressed.
In conclusion, a suboptimal early-life environment combined with a mismatched postnatal milieu predisposes offspring to increased hepatic ROS, inflammation, and hyperinsulinemia, leading to hepatic fibrosis. This recapitulates changes seen in patients in whom benign NAFLD progresses to cirrhosis and ultimately liver failure. We suggest that the early environment life is crucial in identifying the subgroup of patients at highest risk of such progression; however, this should be tested in humans. A clinically relevant dose of CoQ 10 reversed liver fibrosis via downregulation of ROS, inflammation, and hyperinsulinemia and upregulation of the Nrf2/ARE antioxidant pathway. Because fibrosis contributes to up to 45% of deaths in the industrial world (41), CoQ 10 supplementation may be a cost-effective and safe way of reducing this global burden in at-risk individuals, before the development of an NAFLD phenotype. Values are means 6 SEMs; n = 10/ group; 10/10 rats used. * , ***C and R compared with CQ and RQ: *P , 0.05, ***P , 0.001. *C compared with R and C compared with CQ, P , 0.05. ***R compared with RQ, P , 0.001. Statistical interactions: Nrf2, P = 0.9; Hmox1, P = 0.6; Nqo1, P = 0.4; Gst, P = 0.1; Gpx1, P = 0.01. Data were analyzed by using 2-factor ANOVA and Duncan's post hoc testing, where appropriate. C, control; CoQ 10 , coenzyme Q; CQ, control CoQ 10 ; Gpx1, glutathione peroxidase 1; Gst, glutathione synthetase; Hmox1, heme oxygenase 1; Nqo1, NAD(P)H dehydrogenase, quinone 1; NRF, nuclear erythroid 2-related factor; Nrf2, nuclear factor, erythroid 2-like 2; R, recuperated; RQ, recuperated CoQ 10 .
We thank K Phillips for technical expertise. The authors' responsibilities were as follows-JLT-A: conducted research, analyzed data; DSF-T: provided essential materials; IPH and VN: conducted research; CEA: wrote the manuscript; MSM-G: provided essential materials; JMM: designed the research; SEO: designed the research; JLT-A and SEO wrote the manuscript and had primary responsibility for final content; and all authors: read and approved the final manuscript. None of the authors declared a conflict of interest.