Effects of Increased Dietary Cholesterol with Carbohydrate Restriction on Hepatic Lipid Metabolism in Guinea Pigs

Abstract

Excessive lipid accumulation within hepatocytes, or hepatic steatosis, is the pathognominic feature of nonalcoholic fatty liver disease (NAFLD), a disease associated with insulin resistance and obesity. Low-carbohydrate diets (LCD) improve these conditions and were implemented in this study to potentially attenuate hepatic steatosis in hypercholesterolemic guinea pigs. Male guinea pigs (n = 10 per group) were randomly assigned to consume high cholesterol (0.25 g/100 g) in either a LCD or a high-carbohydrate diet (HCD) for 12 wk. As compared with HCD, plasma LDL cholesterol was lower and plasma triglycerides were higher in animals fed the LCD diet, with no differences in plasma free fatty acids or glucose. The most prominent finding was a 40% increase in liver weight in guinea pigs fed the LCD diet despite no differences in hepatic cholesterol or triglycerides between the LCD and the HCD groups. Regardless of diet, all livers had severe hepatic steatosis on histologic examination. Regression analysis suggested that liver weight was independent of body weight and liver mass was independent of hepatic lipid content. LCD livers had more proliferating hepatocytes than did HCD livers, suggesting that in the context of cholesterol-induced hepatic steatosis, dietary carbohydrate restriction enhances liver cell proliferation.

Nonalcoholic fatty liver disease (NAFLD) is a pathologic condition that represents a spectrum of histologic features ranging from excessive lipid accumulation in hepatocytes (referred to as hepatic steatosis), inflammatory steatosis or nonalcoholic steatohepatitis, to fibrosis and cirrhosis.²⁶ Once an unnamed disease,²³ NAFLD is now the most common form of chronic liver disease and is currently estimated to affect 30% of the US adult population and nearly 20% of the global adult population.²⁰ Obesity and its associated sequela, namely insulin resistance, are the prime contributors to NAFLD development, given that 75% of obese¹² and virtually all morbidly obese persons¹¹ display NAFLD. This invariable link leads to the characterization of NAFLD as the hepatic manifestation of the metabolic syndrome and therefore a bona fide public health concern.²⁶

The array of histologic abnormalities in NAFLD begins with the excessive cytoplasmic accumulation of hepatocellular fat and the formation of lipid-laden cells (hepatic steatosis).⁶ The presence of steatosis often is explained in terms of the ‘two-hit’ hypothesis,⁹ where conditions such as insulin resistance and obesity lead to the ectopic deposition of lipids in hepatocytes (the ‘first hit’). Susceptible livers are prone to a future ‘second hit’, such as inflammation, and the development of nonalcoholic steatohepatitis, thereby leading to more extensive damage to liver architecture. Several mechanisms are suspected to participate in the pathogenesis of steatosis, including elevated fatty acid delivery from adipose, impaired hepatic fatty acid oxidation, decreased synthesis and secretion of VLDL particles, and enhanced synthesis of fatty acids and triglycerides through hepatic de novo lipogenesis.

The extraction and storage of nutrient-derived energy represents a distinct survival mechanism that is exploited by nearly all organisms. The liver is central to nutrient metabolism and is governed to a great extent by hormonal influences. Insulin, the dominant anabolic hormone, is released from the pancreas in response to dietary carbohydrate, and insulin release signals the need for glucose uptake and storage in tissues. At the level of the liver, storage of newly accumulated glucose is accomplished through the enzymatic conversion of free glucose molecules to large glycogen polymers in a process referred to as glycogenesis. If the intake of carbohydrate eventually exceeds the hepatic glycogen storage capacity, glucose in the glycolytic pathway instead is directed toward the energetically expensive lipogenic pathway for conversion and production of triglycerides. This anabolic setting culminates in the incorporation and assemblage of newly synthesized lipids into spherical VLDL macromolecules for eventual egress from hepatocytes into the circulation. Dietary conditions in which carbohydrate intake is reduced or physiologic conditions such as fasting are characterized by an increased need for cellular energy production, which is primarily achieved through the catabolic actions of glucagon. As the insulin:glucagon ratio begins to decline, the liver initiates the coordinated enzymatic breakdown of glycogen particles (glycogenolysis), releasing free glucose into the circulation. Glucagon also triggers the production of glucose from noncarbohydrate precursors through gluconeogenesis and enhances the oxidation of lipids (such as fatty acids) to provide cellular energy.

Identification of dietary programs capable of ameliorating fatty liver and progression of NAFLD are of utmost importance. Currently weight reduction through a combination of dietary modulation and physical activity is the proposed strategy for treatment of NAFLD. However, because diets by nature are a complex mélange of individual compounds variable in their concentrations and quality, ascertaining which dietary component(s) are pertinent to NAFLD remain a difficult task. Given the prominent role of insulin resistance and dyslipidemia in NAFLD, the ideal diet is one that restores insulin sensitivity, reduces adipose mass, and improves the plasma lipid profile. Indeed, a low-carbohydrate diet (LCD) can fulfill all of these needs¹ and may therefore represent a novel dietary intervention in the treatment of NAFLD.

Current knowledge about the pathogenesis of hepatic steatosis was derived mainly from animal models designed to mimic the human conditions of NAFLD. Genetic and dietary models have been developed, each with their particular advantages and caveats; however, no single model fully replicates the human condition. Cavia porcellus, the common guinea pig, has unique characteristics that may most fully resemble human lipid and lipoprotein metabolism, the most pertinent being that guinea pigs primarily transport cholesterol within LDL.¹³ In addition, guinea pigs retain many of the same enzymes and proteins involved in lipoprotein and hepatic lipid metabolism as humans.¹⁴ However, guinea pigs have yet to receive considerable attention as a plausible model for NAFLD. Here we sought to assess the metabolic effects of both LCD and high-carbohydrate diets (HCD) on hepatic steatosis in hypercholesterolemic guinea pigs.

Materials and Methods

Animals.

Pathogen-free male Hartley guinea pigs (n = 20, 18 mo old) were purchased from Charles River Breeding Laboratories (Wilmington, MA) and housed in individual cages (cages were autoclaved prior to use) on a 12:12-h light:dark cycle at 23 °C and approximately 55% average humidity. Guinea pigs were provided water in bottles, and cages were changed every other day. One week prior to the experimental period, all guinea pigs were acclimated to the facility and standard diet. Guinea pigs were randomly assigned (n = 10 per diet group) to consume either LCD (energy distribution of 10% from carbohydrate, 60% from fat, and 30% from protein) or HCD (55% of energy from carbohydrate, 20% from fat, and 25% from protein) for 12 wk; all other components (cellulose, guar gum, vitamin and mineral mixes) were consistent between the diets. Diets were prepared by Research Diets (New Brunswick, NJ). Both the LCD and HCD were supplemented with high dietary cholesterol (0.25 g per 100 g), a level known to induce atherosclerosis in guinea pigs²¹ and equivalent to the consumption of 1800 mg daily for humans.²² Furthermore, this level of dietary cholesterol has been demonstrated to induce fatty liver in guinea pigs within a 12-wk period.³¹ At the end of the experimental period, guinea pigs were euthanized by exanguination under isofluorane anesthesia. Plasma was removed from RBC by centrifugation at 2000 × g for 20 min, and livers were immediately excised, weighed, and snap-frozen and stored for further determinations. All animal protocols were approved by the University of Connecticut IACUC.

Clinical chemistry measures.

Plasma total cholesterol and triglycerides were analyzed enzymatically as described previously.² VLDL cholesterol was isolated through sequential ultracentrifugation in an ultracentrifuge at a density of 1.006 g/mL at 200,000 × g and 10 °C for 45 min. The white uppermost layer of the supernatant was measured enzymatically to determine the cholesterol in VLDL. Plasma HDL cholesterol was measured through dextran-sulphate–magnesium-precipitation after the precipitation of apolipoprotein-B–containing lipoproteins.³³ Plasma LDL cholesterol was calculated according to the following equation: total cholesterol – VLDL cholesterol + HDL cholesterol. Plasma nonesterified fatty acids were analyzed enzymatically (Wako Diagnostics, Richmond, VA) as previously reported.²¹

Activities of the liver enzymes ALT, AST, and ALP were analyzed spectrophotometrically by using a kinetic assay (ThermoScientific, Middletown, VA).

Plasma ketone bodies (acetoacetate and 3-β-hydroxybutyrate) were determined quantitatively at 404 nm by using a commercial kit (Total Ketone Bodies Kit, Wako Diagnostics).

Plasma insulin was analyzed by using a rat–mouse insulin ELISA (Linco Research, St Charles, MO). This assay is based on the binding of plasma insulin molecules by biotinylated polyclonal antibodies by using microtiter plates coated with mouse monoclonal and antirat insulin antibodies.²¹ Plasma glucose (25 μL) was determined by using an automated analyzer (model 2300 Stat Plus Glucose and Lactate Analyzer, Yellow Springs Instruments, Yellow Springs, OH).

Hepatic total cholesterol was extracted as described previously.¹⁵ After samples were resuspended in 200 μL ethanol and stored in a cold room (4 °C), hepatic total and free cholesterol were quantified by enzymatic analysis. Hepatic cholesteryl ester was calculated after subtracting free cholesterol from total cholesterol. An aliquot of liver tissue was used to extract and quantify hepatic triglycerides as previously described.¹⁵

Hepatic morphology.

To assess hepatic morphology, approximately 1 to 2 g of tissue from the left sublobe of the quadrate lobe adjacent to the gall bladder was removed and immediately suspended in formalin solution. Tissue samples were embedded in paraffin and sliced at 5 μm for hematoxylin and eosin staining to differentiate the nucleus (blue) from the cytoplasm (red). Slides were evaluated on 2 separate occasions by a board-certified pathologist who was blinded to the study design. Slides assigned a grade for fatty infiltration, inflammation, and fibrosis when appropriate. Grading of hepatic steatosis is based on the percentage of lobule and parenchyma involved by using previously developed criteria:⁵ grade 0, clinical significance of less than 5% fat infiltration often ignored in histologic examinations; 1, mild fat infiltration (predominantly macrovesicular lipid droplets) in less than 33% hepatocytes; 2, moderate fat infiltration (typically mixed microvesicular and macrovesicular lipid droplets) in 34% to 66% of hepatocytes; and 3, severe fat infiltration (typically mixed microvesicular and macrovesicular lipid droplets) in greater than 66% of hepatocytes. Slides were examined by light microscopy (BX41 Microscope, Olympus, Irving, TX) at both 40× and 400× magnification and photographed digitally (Coolpix 995, Nikon, Tokyo, Japan).

Hepatic immunohistochemistry.

Hepatocyte proliferation was assessed immunohistochemically by using proliferating cell nuclear antigen (PCNA),¹⁷ an auxiliary protein of DNA polymerase δ; in eukaryotic cells, expression of PCNA is required for DNA replication.⁴ In addition, PCNA is used routinely as a marker of liver regeneration.³⁴ Briefly, 5 μM paraffin-embedded unstained liver sections were incubated with sodium citrate buffer (10 mM sodium citrate, 0.05% Tween 20, pH 6.0) in glass slide holders and placed in a boiling water bath (30 min) to achieve antigen retrieval. Sections then were treated with 3% hydrogen peroxide solution (10 min) to block endogenous peroxidase activity, incubated with avidin and biotin (Vector Laboratories, Burlingame, CA) for 15 min each to block remaining biotin binding sites on the avidin, and treated with 5% normal goat serum in PBS to block nonspecific binding sites. A 1:100-dilution of mouse monoclonal PCNA antibody (Santa Cruz Biotechnology, Santa Cruz, CA) was used as the primary antibody and added to the sections (2 h). A biotinylated goat antimouse secondary antibody (dilution, 1:1000; Santa Cruz Biotechnology) was applied to the sections (30 min) followed by incubation with a streptavidin-conjugated peroxidase (Vectastain Elite ABC kit, Vector Laboratories) for 30 min. Color was developed by exposing the peroxidase to 3,3′-diaminobenzidine (Vector Laboratories) to yield a brown stain. Sections were counterstained with Gill hematoxylin (50:50). The nuclei of hepatocytes in the G₁ phase of the cell cycle were stained light brown. Nuclei in S phase were stained dark brown. Positively stained hepatocyte nuclei in a subset of liver samples were quantified by counting from 5 nonoverlapping fields with a minimum of 100 hepatocytes counted. The data are expressed as a percentage per 100 hepatocytes.

Statistical analysis.

All data were analyzed by using SPSS 16.0 (SPSS, Chicago, IL). The effect of LCD and HCD diets on body weight were evaluated by using mixed ANOVA. Independent and paired t tests were performed when appropriate, and correlations were performed through linear regression analyses. All data are expressed as mean ± SD. Data were considered statistically significant at an α level of P < 0.05.

Results

Diet effects on body weight, caloric intake, and absolute daily food intake.

Throughout the 12-wk study, guinea pigs from both diet groups displayed similar body weights, with no significant effect of diet and no overall difference in body weight between the LCD and HCD groups.

Food intake was assessed every 2 d by weighing food remaining from the previous day. Results indicated a trend (P = 0.08) for guinea pigs in the LCD group to have a slightly higher daily caloric intake (97.3 ± 10.2 kcal/d) compared with that of HCD guinea pigs (86.3 ± 13.7 kcal/d). Overall, guinea pigs fed LCD consumed 11.3% more calories (P = 0.08) daily than did HCD animals. However, despite the differences in daily caloric intake, the absolute amount of food consumed, in grams per day, was not significantly different between the LCD (22.1 g/d) and HCD (24.7 g/d) groups.

Diet, plasma lipids, and ketone bodies.

Plasma total cholesterol did not differ between groups (Table 1). Plasma LDL concentrations were higher (P < 0.05) in the HCD group than the LDL group, but guinea pigs fed the LCD had higher (P < 0.05) triglyceride and VLDL cholesterol levels than did guinea pigs fed the HCD. In addition, plasma ketone bodies were higher (P < 0.05) in the LCD compared with the HCD group (Table 1), indicating that LCD-fed guinea pigs relied on fatty acids as a preferential fuel source.

Table 1.

Plasma lipids and ketones (mean ± 1 SD) of guinea pigs fed low- or high-carbohydrate diet

	Low	High
Total cholesterol (mg/dL)	98.14 ± 8.08	148.95 ± 24.63
VLDL (mg/dL)	9.01 ± 0.77	6.64 ± 0.55a
LDL (mg/dL)	71.9 ± 11.0	81.7 ± 8.5a
HDL (mg/dL)	16.48 ± 1.74	15.5 ± 1.04
Triglycerides (mg/dL)	75.51 ± 9.26	47.87 ± 4.13a
Nonesterified fatty acids (mEq/L)	0.828 ± 0.093	0.630 ± 0.047
Ketones (µmol/L)	231 ± 118	142 ± 61a

^aSignificant (P < 0.05) difference between groups

Diet and plasma glucose, insulin, and hepatic enzymes.

There were no differences in plasma glucose (LCD, 11.0 ± 0.3 mmol/L; HCD, 11.2 ± 0.6 mmol/L) or insulin (LCD, 232 ± 41 pmol/L; HCD, 251 ± 42 pmol/L) levels between groups. The enzymatic activity of plasma ALT, AST, and ALP did not differ between the LCD and HCD groups (data not shown).

Diet, liver weight, and hepatic lipid content.

Livers from guinea pigs fed LCD were markedly (P < 0.001) larger and weighed more (approximately 40%) than those from HCD guinea pigs (Table 2). Despite the difference in liver weights, there were no differences between diet groups in hepatic total cholesterol, free cholesterol, esterified cholesterol, or nonesterified fatty acids (Table 2). In addition, regression analysis demonstrated a lack of correlation between liver weight and hepatic triglyceride, nonesterified fatty acids, and total cholesterol in both the LCD and HCD groups combined as well as within groups, suggesting that the variation in liver weight cannot fully be explained by hepatic lipid content. A correlation between body weight and liver weight was observed in the HCD group (r = 0.708, P = 0.022) but not in the LCD group (r = 0.519, P = 0.124; Figure 1 A and B). Moreover, the ratio of liver weight to body weight was significantly (P < 0.001) greater in the LCD (0.051 ± 0.007 g) than the HCD (0.069 ± 0.009 g) group, suggesting that the increased body weight in LCD guinea pigs does not fully account for the greater liver weight of these animals. Furthermore, at the conclusion of this study, there was no difference in final body weight, as demonstrated by mixed ANOVA (P = 0.231), thus suggesting that factors other than body weight contributed to the increased liver weight in LCD animals. Lastly, regression analyses were performed to determine the extent to which hepatic triglyceride, nonesterified fatty acids, and total cholesterol contributed to the variation in liver weight between animals fed the LCD and HCD. The results demonstrated that for the LCD group (r² = 0.154, r = 0.392) this variation in hepatic lipids explains only 15.4% of the variation in liver weight, whereas in the HCD group (r² = 0.451, r = 0.671) the difference in total hepatic lipids explains 45.1% of the variation in liver weight. Collectively, these analyses suggest that hepatic lipid accumulation is not the major contributor to the increased liver weight in LCD animals. In fact, hepatic lipid content appears a more important determinant of liver weight for the HCD group.

Table 2.

Liver parameters (mean ± 1 SD) of guinea pigs fed low- or high-carbohydrate diet

	Low	High
Liver weight (g)	71.4 ± 11.1	50.5 ± 9.9a
Total cholesterol (mg/g)	11.1 ± 3.3	9.1 ± 1.7
Free cholesterol (mg/g)	4.7 ± 2.3	3.1 ± 2.6
Cholesterol ester (mg/g)	6.3 ± 4.9	7.4 ± 1.8
Triglycerides (mg/g)	17.1 ± 6.1	17.4 ± 6.3
Nonesterified fatty acids (mEq/g)	234 ± 111	236 ± 166

^aSignificant (P < 0.05) difference between groups

Figure 1.

Correlation between liver weight and final body weight in guinea pigs fed (A) LCD or (B) HCD.

Hepatic morphology.

Liver sections from each guinea pig were examined by a board-certified pathologist who was blinded to treatment group to assess the degree of fatty infiltration and presence of progressive liver damage, including inflammation and fibrosis. Regardless of diet, livers from all guinea pigs were considered to be severely steatotic and were classified as grade 3, with a full spectrum of microvesicular and macrovesicular lipid droplets being present. Therefore, all livers displayed mixed hepatic steatosis. The centrolobular and midzones of all livers were equally steatotic, with variation (if present) restricted to the peripheral sections of the lobule; therefore the fatty change displayed a lobular pattern (Figure 2 A through D).

Figure 2.

Representative examples of liver tissue from guinea pigs fed (A, B) LCD or (C, D) HCD. Hematoxylin and eosin stain; magnification: 40× (A, C); 400× (B, D).

All livers had rare inflammatory foci and were classified as having grade 1 (mild) inflammation and occasional scattered degenerate cells. There was no histologic evidence of Mallory bodies, eosinophilic inclusions located within the perinuclear cytoplasm of ballooned hepatocytes in the pericentral parenchyma that are associated with increased inflammation and pericellular fibrosis.⁷ In addition, there was no evidence of fibrosis; therefore all livers in the current study were purely steatotic and represented the initial stage of NAFLD.

Assessment of hepatocyte proliferation.

Hepatocyte proliferation was assessed by using PCNA immunohistochemistry, to elucidate the potential mechanisms responsible for the differences in liver weight between the LCD and HCD groups. Light microscopy (magnification, 400×) revealed that livers from the LCD group (Figure 3 A) displayed more positively brown stained nuclei than did those from the HCD group (Figure 3 B). The percentage of positively stained nuclei in LCD livers was 44% ± 7% compared with 36% ± 6% (P < 0.05) in HCD liver samples.

Figure 3.

Representative examples of liver tissue from guinea pigs fed (A) LCD or (B) HCD. Black arrows indicate brown-stained nuclei of hepatocytes in the S phase of the cell cycle. PCNA immunohistochemistry; magnification, 400×.

Discussion

The current study suggests a unique role of carbohydrate restriction in modulating the hepatic lipid and cholesterol metabolism of guinea pigs during the development of hepatic steatosis. Guinea pigs that consumed high dietary cholesterol in conjunction with either LCD or HCD for 12 wk developed severe hepatic steatosis, as evidenced by histologic examination. Plasma ketones were substantially higher in the LCD group, indicating that these guinea pigs preferentially utilized fatty acids as an energy source.¹⁹ The combination of LCD or HCD with high cholesterol resulted in differences in plasma lipids that were somewhat unexpected, with LCD resulting in higher plasma triglycerides and HCD in higher LDL cholesterol.

No differences were observed between groups in the concentrations of hepatic triglyceride, total cholesterol, or nonesterified fatty acids, yet the liver weight of LCD-fed guinea pigs was greater than that of guinea pigs fed HCD. Further, liver weight and hepatic contents of the LCD group were independent of body weight. These unexpected findings redirected our research toward elucidating the mechanisms responsible for the disparity in liver weight and whether the presence or absence of dietary carbohydrate ultimately facilitated these events. Because plasma ketone bodies were greater in the LCD group compared with the HCD group, it is unlikely that hepatic glycogen could contribute substantially to the increased liver weight in LCD-fed guinea pigs. Furthermore, all guinea pigs displayed a similar degree of liver injury, evidenced by no differences in plasma levels of hepatic enzymes or histology. We subsequently hypothesized that the livers of LCD-fed guinea pigs may have been prone to a proliferative process to increase the hepatocyte population, ultimately in an attempt to store excess lipid, thereby avoiding the cytotoxic effects of free cholesterol and fatty acids. Indeed, qualitative analysis of hepatic PCNA levels demonstrated an increased expression of this cell-cycle protein in LCD livers, suggesting an intrinsic effect of this diet in enhancing hepatocyte growth.

The strong association between insulin resistance, obesity, metabolic syndrome, and NAFLD suggest dietary modulation as a critical means for treatment; however, the type of diet that is most effective in attenuating NAFLD is largely unknown. Patients with NAFLD tend to consume diets rich in simple carbohydrates, often in the presence of elevated saturated fatty acids,³⁵ a combination that may exacerbate the degree of liver damage. Therefore, the general consensus³⁵ is that excessive intakes of these macronutrients are harmful and should be decreased. The specific contribution of carbohydrate and fat to NAFLD progression, however, remains enigmatic and has only recently been investigated. To date, few long-term controlled studies have investigated the role of carbohydrate restriction on NAFLD, although several lines of evidence suggest that limited carbohydrate consumption may be beneficial for improved liver function. A 1-y prospective study²⁸ found that increased carbohydrate intake was significantly associated with higher odds of hepatic inflammation, whereas increased fat intake was associated with significantly lower odds of inflammation in morbidly obese patients with NAFLD. In addition, the authors²⁸ observed a trend toward lower odds of steatosis and fibrosis in the highest tertile of fat intake, suggesting that the presence of fat may improve NAFLD histopathology. The mechanisms underlying carbohydrate-induced hepatic inflammation have yet to be clarified. A more recent study³⁰ found that a low-carbohydrate, ketogenic diet (less than 20 g carbohydrate daily) decreased steatosis, inflammation, and fibrosis in NAFLD patients within a 6-mo period, concomitant with weight loss and lowered plasma insulin and glucose levels.

In the current study, guinea pigs in both diet groups displayed severe hepatic steatosis, contrary to the initial hypothesis that carbohydrate restriction would decrease steatosis compared with that in the carbohydrate enriched group. Several explanations may help clarify this observation. First, both diet groups were supplemented with a high level of dietary cholesterol (0.25 g per 100 g diet), a value equivalent to the consumption of 1800 mg cholesterol daily for humans.²² A recently emerging view implicates free cholesterol as a central molecule in the progression of steatosis to nonalcoholic steatohepatitis in both animal models²⁴ and humans,²⁷ with lipidomic analyses of liver tissue in NAFLD patients demonstrating an increased content of free cholesterol in both steatotic and nonalcoholic steatohepatitis livers.²⁷ Hepatic total cholesterol concentration, including free and esterified cholesterol, was increased in both dietary groups of guinea pigs, albeit at levels lower than that of hepatic triglyceride.

The most prominent and unanticipated finding of the current study was the marked difference in liver weight between the groups. Despite the difference in liver mass, there were no differences in hepatic lipids between the LCD and HCD groups, nor were hepatic lipids correlated with liver weight in the LCD group. Furthermore, the data indicate that liver weight was independent of body weight. One possible explanation is that in the presence of the elevated dietary fat and cholesterol, hepatocytes were stimulated to undergo a proliferative process to accommodate the storage of these lipids. Indeed, an abundance of evidence exists that highlights an inextricable link between fat and hepatocyte proliferation. For instance, a well-recognized, albeit poorly understood phenomenon, during rodent liver regeneration is the enhanced release of peripheral adipose free fatty acids, which are uptaken by the liver, thereby leading to increased hepatic de novo lipogenesis and the accumulation of intracellular triglyceride (transient steatosis).^29,32 Early inquiries into the mechanistic details of liver regeneration in rats revealed that lean livers indeed become transiently steatotic.¹⁰ Congruous with these observations, it has been demonstrated that regenerating rat liver hepatocytes normally accumulate lipid microdroplets for a 24- to 72-h period after removal of approximately two thirds of liver mass.²⁵ This phenomena is believed to reflect a metabolic adaptation of hepatocytes to changing cellular conditions, where fats are used as a readily available energy source and are available for incorporation into lipid cellular membranes. Moreover, hepatic acinar zone 1 cells, which are believed to be the initial site for hepatocyte proliferation, appear to prefer fatty acids as an energy source.¹⁶

Interestingly, glucose administration has been observed to inhibit hepatic cell division and tissue repair in animals with partial hepatectomy and in humans recovering from liver injury.¹⁸ Although the mechanisms by which glucose administration inhibits liver cell growth are still largely unknown, one potential explanation is that due to the presence of insulin and the potent suppressive effect on hormone-sensitive lipase, less adipocyte hydrolysis occurs and less fatty acids are mobilized to the liver.¹⁸ A more specific hypothesis has also been proposed⁸ based on the observation that glucose depresses hepatocyte DNA synthesis by directly reducing the expression of c-Myc, a nuclear phosphoprotein with a crucial role in the progression of cells through the G₁ and G₂ phases of the cell cycle. However, other observations³ revealed that glucose inhibition of hepatocyte growth was abrogated when glucose was administered simultaneously with various amino acids, a situation that is relevant to dietary and clinical interventions. The expression of PCNA, a routine marker for hepatocyte growth,³⁴ was greatest in the livers of the LCD group and tended to be lower in the livers of the HCD group. These results indicate that the presence of dietary fat, in the absence of carbohydrate, resulted in a more pronounced proliferation of hepatocytes and a concomitant increase in liver mass in the current study.

The evidence we present here collectively indicates that excessive dietary cholesterol ingestion by guinea pigs results in hepatic steatosis, irrespective of the carbohydrate content. In addition, LCD livers were heavier than were HCD livers, despite no differences in hepatic triglycerides, total cholesterol, or nonesterified fatty acids between groups. The higher liver weight of LCD guinea pigs may be due to enhanced hepatocyte proliferation, given that LCD livers tended to display more actively proliferating hepatocytes than did HCD livers. In the future, additional in-depth studies must be undertaken to investigate the potential interactions between carbohydrate restriction and liver physiology, particularly in the context of NAFLD and without the confounding effects of excessive dietary cholesterol.