Abstract
Objective
To characterize changes in the hepatic lipid profile and metabolic pathways associated with increasing hepatic fat accumulation in bearded dragons (Pogona vitticeps).
Methods
Untargeted lipidomic analysis was conducted using LC-MS-MS on liver samples from bearded dragons with varying hepatic fat content. Hepatic fat percentage was calculated from digital image analysis of scanned histopathology slides. After data normalization, associations between lipids and hepatic fat percentages were assessed using serial linear models adjusted for false discovery rate, volcano plots, and principal component analysis. Changes in fatty acyl chains of triacylglycerols and phospholipids were characterized graphically using bubble plots. Enrichment and pathway analyses were also performed to examine potential disruptions in lipid metabolic pathways.
Results
36 central bearded dragons were sampled, and 976 lipid molecules were identified and quantified. Triacylglycerols were the most abundant and exhibited significant increases in concentrations and changes in fatty acyl chain characteristics with higher hepatic fat content. Notably, ether-linked glycerolipids were significantly enriched with increasing fat content. Phospholipids, especially phosphatidylethanolamines and phosphatidylinositols, demonstrated a negative association with hepatic fat accumulation, but fatty acyl chains remained stable. Sphingomyelins were also decreased with increasing hepatic fat.
Conclusions
This study shows some significant shifts in the hepatic lipidome of bearded dragons with increased hepatic fat, mainly involving glycerolipids, phospholipids, and sphingolipids.
Clinical Relevance
These findings reveal both shared and unique features when compared to mammalian and avian fatty liver disease and suggest species-specific lipid adaptive mechanisms.
Hepatic fat accumulation is exceedingly common in reptiles and may lead to fatty liver disease, also known as hepatic lipidosis in veterinary medicine.1 The prevalence of increased hepatic fat accumulation is particularly high in central bearded dragons (Pogona vitticeps), one of the most popular companion reptile species, with a prevalence of significant fat accumulation of nearly 40% on necropsy.2 In humans, the disease, referred to as metabolic-associated fatty liver disease (MAFLD), is equally common and is in fact the most prevalent liver disorder.3
The commonly accepted understanding of the pathophysiology of fatty liver in humans is known as the “2-hit hypothesis.” It is characterized by the progressive accumulation of triacylglycerols (TGs), also known as triglycerides, within hepatocytes (first hit), which ultimately leads to oxidative stress, lipotoxicity, and inflammation (second hit), causing fibrosis and hepatic dysfunction over time.3 In bearded dragons and other reptiles, the pathogenesis and pathophysiology of hepatic lipidosis are unclear. A diet high in fat (such as from larval insects) and carbohydrates (including fruits)4 may contribute to increasing hepatic fat content; however, while diet and overfeeding likely play a significant role, this has not been conclusively established. Increased hepatic lipid synthesis and storage during reproduction and vitellogenesis in female reptiles may also be a factor, and female bearded dragons were found to have a higher prevalence of hepatic fat accumulation than males, with 43% higher odds of more severe lesions than males.2 Other factors, such as brumation, environmental temperature, restricted exercise, poor replication of seasonal cycles in captivity, and concurrent diseases, may also have an influence.2
While the disease is well studied in humans and other mammals, direct translation or extrapolation from these species to reptiles is not always feasible due to distinct physiologic and metabolic differences that may impact TG accumulation and resulting clinical signs. For instance, hepatic lipid metabolism and storage differ markedly in reptiles versus in mammals, and hepatic fat deposition tends to be much higher under many physiological conditions.5 When considering the scale of hepatic fat accumulation in bearded dragons, only minimal metabolic changes are detectable, which rarely mirror changes observed in humans and adds to the uncertainty around when such accumulation becomes pathological in this species.6,7 In addition, inflammation and fibrosis are associated with the onset and severity of clinical signs in people, but inflammation is minimally observed in bearded dragons, and while fibrosis is observed, its association with clinical signs or hepatic metabolic dysfunction is not evident.7,8 In humans, MAFLD is strongly associated with dyslipidemia, insulin resistance, type II diabetes, and obesity, but these associations are also not apparent in bearded dragons.6,7
Conversely, fatty liver in reptiles and mammals do share certain characteristics. Fatty acids (FAs) are primarily metabolized in mitochondria by β-oxidation, and mitochondrial dysfunction has been implicated in the pathophysiology of the disease in humans,9,10 bearded dragons, and red-footed tortoises (Chelonoidis carbonaria).7,11 A decrease in β-hydroxybutyric acid has also consistently been linked to hepatic fat accumulation in bearded dragons,6,7 which may be associated with decreasing hepatic FA β-oxidation or ketogenesis impairment as reported in people.12
Exact lipid changes occurring in the liver with increasing fat accumulation have not been characterized in bearded dragons or other reptiles. A comprehensive characterization of hepatic lipids in bearded dragons with increased fat accumulation may provide insights into the pathophysiology and pathways shared with other species. Lipidomics, a branch of metabolomics, is the large-scale study of lipids in biological systems and involves the comprehensive analysis of lipids using liquid chromatography and mass spectrometry. While it has been applied on the plasma of bearded dragons with hepatic fat accumulation for novel biomarker discovery,6,7 it has not been performed directly on the liver. Studying the hepatic lipidome across different stages of hepatic fat accumulation may help uncover underlying disease mechanisms, disruptions in metabolic pathways, and insights into lipid compositional changes that drive disease progression. While there are many such lipidomic investigations in humans with MAFLD,13–15 to the authors’ knowledge this has not been investigated in reptiles before. The phylogenetically closest sauropsid species to squamates in which hepatic lipidomics data are available are poultry birds due to the propensity of egg-laying chickens for fatty liver disease and the diet induction of fatty liver in geese and ducks for the production of foie gras.16–19
The objectives of this study were to comprehensively characterize the hepatic lipid profile in bearded dragons with increasing levels of hepatic fat accumulation; to identify specific lipid classes, subclasses, and species associated with increasing hepatic fat content; to investigate potential changes in the composition of stored TGs; and to explore metabolic pathways and functional lipid alterations that may contribute to hepatic fat accumulation in this species. Our hypotheses were that bearded dragons with higher hepatic fat content would exhibit distinct lipidomic profiles, particularly with elevated levels of TGs, alterations in other lipid classes and subclasses, and shifts toward lipids with shorter fatty acyl chain lengths and lower degrees of unsaturation. We also expected to observe specific disruptions in glycerolipids and phosholipid metabolic pathways.
Methods
The research was approved by the University of California-Davis IACUC (protocol #22816).
Animals and tissue samples
Forty-one bearded dragons, originally selected to be culled for breeding management reasons, were acquired from a local breeder in Chico, California. The cohort included 20 males and 21 females. Seventeen were 1.5 to 3 years old, and 24 were 4 to 7 years old. The weight, which was normally distributed, had a mean ± SD of 312 ± 81 g. Animals were deemed healthy based on a complete physical examination and normal appetite. At the breeding facility, they were fed a diet consisting of greens and insects (mealworms and crickets) supplemented with calcium carbonate.
Animals were humanely euthanized, also as part of concurrent studies,7 using 1 mL of IV or intracardiac potassium chloride 2 mEq/L under alfaxalone (Alfaxan Multidose; Jurox) 10 to 20 mg/kg SC anesthesia.
Bearded dragons were necropsied about 2 to 4 hours after euthanasia. Sex was confirmed on necropsy. The caudal most section of the left liver lobe was sampled and fixed in formalin for histopathology. The rest of the liver was harvested and placed in a disposable polyethylene sampling bag (Whirl-Pak; Nasco Healthcare) and stored at –80 °C until processing.
Digital image analysis
The liver sections were embedded in paraffin, sectioned onto glass slides, and stained with H&E. Slides were reviewed by a veterinary board–certified pathologist (Leonardo Susta, Ontario Veterinary College, University of Guelph) for the presence of any pathology other than hepatic lipidosis, and lizards with non–lipid-related hepatopathies were excluded from the study. Livers were also graded by histopathology according to a published scoring system.8
All liver glass slides were scanned at 0.25 μm/pixel using a digital scanner (Ocus 40; Grundium) for the visual quantification of hepatic fat by digital image analysis.
For hepatic lipid content quantification (in percentages), scans were exported into an open-source bioimage analysis software (QuPath version, 0.4.2),20 and 3 square regions of interest (0.5 mm2) were randomly generated with the viewing software (excluding vessels and biliary ducts) and exported to ImageJ (Version 1.53k, National Institutes of Health, Bethesda, MD, USA, available at https://imagej.nih.gov/ij/). A macro was programmed for automatic fat quantification by successively subtracting and white balancing the background, converting to grayscale, binarizing, and calculating the percentage of white (empty) area corresponding to fat. Areas obtained from the 3 regions of interest were averaged. In addition, livers were arbitrarily classified into 3 categories (0% to 20% fat, 20% to 40% fat, and 40% to 60% fat) for statistical analysis.
Lipidomics
Stored livers were thawed on ice, and 10-to-30-mg samples were harvested from the left liver lobe, measured using a high-precision scale, and placed in Eppendorf 1.5-mL polypropylene tubes and refrozen at –80 °C until analysis.
Samples were submitted to the University of California-Davis West Coast Metabolomics Center for a complex lipid analysis by untargeted lipidomics using LC-MS-MS.
Samples were extracted using a previously published extraction procedure, which included methyl-tert-butyl ether, methanol, and water.21 The organic (upper) phase was evaporated to dryness and resuspended in 110 uL of a solution consisting of 90% methanol, 10% toluene, 50 ng/mL of 1,2-didecanoyl-sn-glycero-3-phosphocholine, and other lipid internal standards (Splash Lipidomix; Avanti Polar Lipids Inc). The mixture was then shaken for 20 seconds, sonicated for 5 minutes at room temperature, and centrifuged for 2 minutes at 16,100 X g. The samples were then placed in 54-vial trays.
The samples were then loaded onto an HPLC system (1290 Infinity; Agilent). For positive ionization mode analysis, 2 μL was injected onto a liquid chromatography column (Acquity Premier BEH C18; Waters Corp) with 1.7-μm particle size, 2.1-mm internal diameter, and 50-mm length. The mobile phase consisted of 60% acetonitrile, 40% water, 10 nM ammonium formate, and 0.1% formic acid (phase A) and 90% isopropanol, 10% acetonitrile, 10 mM ammonium formate, and 0.1% formic acid (phase B). A hybrid quadruple-orbitrap mass spectrometer (Q Exactive HF; Thermo Scientific) was used. The gradient used was 0 minutes at 15% (B), 0.75 minutes at 30% (B), 0.98 minutes at 48% (B), 4.00 minutes at 82% (B), 4.13 to 4.50 minutes at 99% (B), and 4.58 to 5.50 minutes at 15% (B) with a flow rate of 0.8 mL/min. For negative ionization mode analysis, 5 μL were injected onto the same type of liquid chromatography column and analyzed using a quadrupole-orbitrap mass spectrometer (Exploris; Thermo Scientific).
Data processing was performed using MS-DIAL software (version 5) to conduct peak detection, alignment, and annotation of lipid species.22 To ensure data accuracy, duplicates and isotopes were checked and deleted using MS-FLO software.23 Approximate concentrations of all identified species were obtained by comparing them with internal standards, including 1 internal standard per lipid class. Lipid species were reported according to the short-hand notation of the LIPID MAPS classification and nomenclature (Table 1).24
Lipids identified in the liver of bearded dragons by LC-MS-MS lipidomics.
| Lipid category/subclass | N | Median total concentration (ng/mg) | Abundant lipid speciesa |
|---|---|---|---|
| Fatty acyls | 61 | ||
| FA | 55 | 103 | FA 18:1; FA 18:2; FA 20:4 |
| CAR | 6 | 0.2 | CAR 16:0; CAR 18:2; CAR 18:1 |
| Glycerolipids | 311 | ||
| DG O | 1 | 0.02 | DG O-19:1_18:0 |
| DG | 22 | 316 | DG 18:1_18:2; DG 34:2; DG 36:2 |
| TG O | 91 | 2,236 | TG 16:0_18:1_18:2,O |
| TG | 197 | 90,592 | TG 52:3; TG 52:2; TG 52:4 |
| Glycerophospholipids | 484 | ||
| LPC | 17 | 84 | LPC 16:0; LPC 18:2; LPC 18:1 |
| LPC O | 4 | 9 | LPC O-16:0 |
| LPC P | 1 | 1 | LPC P-18:0 |
| LPE | 14 | 82 | LPE 18:0; LPE 20:4; LPE 18:1 |
| LPE O | 6 | 60 | LPE O-16:1; LPE O-16:0 |
| LPI | 5 | 7 | LPI 18:0; LPI 18:2 |
| LPS | 3 | 4 | LPS 18:0; LPS 20:3 |
| PC | 99 | 7,119 | PC 34:2; PC 36:2; PC 36:4 |
| PC O | 79 | 495 | PC 16:0_18:2,O2 |
| PC P | 23 | 332 | PC P 34:1; PC P 36:3 |
| PE | 51 | 1,033 | PE 36:2; PE 36:3; PE 18:0_20:4 |
| PE O | 75 | 449 | PE O-38:2; PE O-16:1_18:2 |
| PE P | 12 | 59 | PE P-36:5 |
| PEtOH | 5 | 10 | PEtOH 16:1_18:1 |
| PG | 15 | 251 | PG 18:1_18:1; PG 18:1_18:2 |
| PG O | 6 | 9 | PG 18:1_18:2,O2 |
| PI | 29 | 665 | PI 36:2; PI 36:3; PI 38:4 |
| PI O | 15 | 2 | PI 18:0_18:2,O2 |
| PS | 17 | 527 | PS 18:0_18:2; PS 16:0_22:4 |
| PS O | 8 | 32 | PS O-19:0_18:1 |
| Sphingolipids | 117 | ||
| Cer | 38 | 34 | Cer d34:1; Cer d32:1 |
| SM | 79 | 838 | SM d34:1; SM d32:1; SM d30:1 |
| Sterol lipids | 3 | ||
| CE | 3 | 606 | CE 18:2; CE 18:1 |
| Total | 976 |
CAR = Acyl carnitines. CE = Cholesteryl esters. Cer = Ceramides. DG = Diacylglycerols. DGO = Alkylacylglycerols. FA = Free fatty acids. LPC = Lysophosphatidylcholines. LPCO = Alkyl-lysophosphatidylcholines. LPCP = Lysophosphatidylcholine plasmalogens. LPE = Lysophosphatidylethanolamines. LPEO = Alkyl-lysophosphatidylethanolamines. LPI = Lysophosphatidylinositols. LPS = Lysophosphatidylserines. PC = Phosphatidylcholines. PCO = Alkyl-phosphatidylcholines. PCP = Phosphatidylcholines plasmalogens. PE = Phosphatidylethanolamines. PEO = Alkyl-phosphatidylethanolamines. PEP = Phosphatidylethanolamines plasmalogens. PEtOH = Phosphatidylethanols. PG = Phosphatidylglycerols. PGO = Alkyl-phosphatidylglycerols. PI = Phosphatidylinositols. PIO = Alkyl-phosphatidylinositols. PS = Phosphatidylserines. PSO = Alkyl -phosphatidylserines. SM = Sphingomyelins. TG = Triacylglycerols. TGO = Alkyldiacyl/dialkylacyl-glycerols.
Complex lipids are reported either by their sum composition or their full fatty acid chain analysis whenever determined on mass spectrometry.
The major lipid species within each subclass is given as well as the median total concentrations in ng/mg of liver for a cohort of 36 bearded dragons to compare magnitudes of concentrations between subclasses.
Statistical analysis
Different statistical analyses were performed on the lipidomics dataset to investigate associations between identified individual lipids and increasing hepatic fat accumulation. Functional analysis and more in-depth characterization of chain length and degree of unsaturation of TGs and select phospholipids across the hepatic lipid accumulation spectrum were also performed.
A series of linear regression models were performed with the concentration of the different lipid species as outcome variables and hepatic fat content and sex as dependent variables. Sex was included in the models as a covariable because of the influence of female sex on lipid physiology in reptiles in general and hepatic fat accumulation in particular.2,5 Lipid concentrations were normalized for all lipids by log-transformation and mean centering before statistical analysis. Obvious duplicates between positive and negative mass spectrometry analytical modes were removed. An iterative for-loop was programmed in R (version 4.3.2; R Foundation for Statistical Computing) to perform approximately 1,000 linear models and export parameter estimates and P values to a dataset. A false discovery rate of 0.05 (q < 0.05) was applied for significance based on the total number of models performed.
The most important changes in the hepatic lipidome were additionally identified on a volcano plot, which plotted the linear model coefficients for hepatic fat (parameter estimates from normalized data as measures of the magnitude of effect) against the statistical significance (log of the P values) of the various lipids by lipid subclasses. The top 10% of lipid species with the highest magnitudes of effect (above the 90th percentile of the absolute values of coefficients) were emphasized as they exhibited the strongest associations with increasing hepatic fat content. In addition, a principal component analysis was performed on the 3 fat categories on normalized lipid species concentrations. Metabolites with important contributions to class separation were identified based on their loadings.
Functional analysis was performed by using an unweighted lipid class enrichment analysis for all lipid subclasses as well as a pathway analysis only for fatty acyls, glycerolipids, and glycerophospholipids (phospholipids) at the sum composition level of characterization using BioPan.25 The enrichment ratio was calculated as the ratio of the number of significant lipid species by class over the number of detected lipid species by class divided by the ratio of the total number of significant lipids over the total number of detected lipids. An enrichment ratio over 1 means that a lipid class/subclass has more significant lipids than expected if the frequency of significance was equal among lipid classes/subclasses. Significant pathways were assessed using t tests in BioPan.
For TG and some phospholipids, the degree of unsaturation and chain length of FAs were graphically assessed using bubble plots over hepatic fat accumulation categories. Only median percentages were plotted for comparison purposes and to account for sample size differences between categories. The phosphatidylcholine (PC):phosphatidylethanolamine (PE) ratio over hepatic fat percentages was also assessed using linear regression.
R (version 4.3.2; R Foundation for Statistical Computing) was used for statistical analysis, and ggplot2 (version 3.5.1) was used for statistical graphing.26 A significance level of .05 was used.
Results
Five livers were excluded from statistical analysis due to concurrent lesions other than hepatic fat accumulation and included multifocal hepatic granulomas, gall bladder ectasia and hyperplasia, cholangiohepatitis, and neuroendocrine tumor metastasis. A total of 36 livers were included in the study. The mean ± SD of hepatic fat content measured on digital image analysis of histologic sections was 32.6% ± 14.1%, which was normally distributed according to a Shapiro-Wilk test. There were 5 livers with 0% to 20% fat (3 females, 2 males), 22 livers with 20% to 40% fat (10 females, 12 males), and 9 livers with 40% to 60% fat (5 females, 4 males). In terms of hepatic fat accumulation histologic grades, 1 liver was scored as grade 1, 6 livers grade 2, 9 livers grade 3, 8 livers grade 4, 5 livers grade 5, 4 livers grade 6, 2 livers grade 8, and 1 liver grade 9. Overall, 24 livers were considered with mild severity, 9 livers with moderate, and 3 livers with severe hepatic lipid changes.
A total of 976 distinct lipid molecules from 5 lipid classes were identified by LC-MS-MS across the livers (Table 1). Glycerolipids were by far the most abundant lipids in the liver on a weight basis (Figure 1), and C52 TGs predominated. Glycerophospholipids were the second most abundant lipids on a weight basis, with C34 to C36 glycerophosphocholines (PC) and C36 glycerophosphoethanolamines (PE) being the most common. Identified sphingolipids mainly included sphingomyelins (SMs) and ceramides with similar sphingoid bases and fatty acyl chains. Cholesteryl esters were mainly dominated by cholesteryl linoleate (18:2) at more than 87% of the total on a weight basis. Fatty acyls were in the lowest amount and were mainly composed of free FAs: oleic acid (FA 18:1), linoleic acid (FA 18:2), and arachidonic acid (FA 20:4). There was a clear shift in the hepatic lipidome with increasing fat accumulation with marked increases in glycerolipids (mainly TG) and overall decreases in phospholipids.
Stacked bar plot of the median concentrations of lipid categories and subclasses in bearded dragon livers across 3 different amounts of hepatic fat accumulation. CAR = Acyl carnitines. CE = Cholesteryl esters. Cer = Ceramides. DG = Diacylglycerols. DGO = Alkylacylglycerols. FA = Free fatty acids. LPC = Lysophosphatidylcholines. LPCO = Alkyl-lysophosphatidylcholines. LPCP = Lysophosphatidylcholine plasmalogens. LPE = Lysophosphatidylethanolamines. LPEO = Alkyl-lysophosphatidylethanolamines. LPI = Lysophosphatidylinositols. LPS = Lysophosphatidylserines. PC = Phosphatidylcholines. PCO = Alkyl-phosphatidylcholines. PCP = Phosphatidylcholines plasmalogens. PE = Phosphatidylethanolamines. PEO = Alkyl-phosphatidylethanolamines. PEP = Phosphatidylethanolamines plasmalogens. PG = Phosphatidylglycerols. PGO = Alkyl-phosphatidylglycerols. PI = Phosphatidylinositols. PIO = Alkyl-phosphatidylinositols. PS = Phosphatidylserines. PSO = Alkyl-phosphatidylserines. SM = Sphingomyelins. TG = Triacylglycerols. TGO = Alkyldiacyl/dialkylacyl-glycerols.
Citation: American Journal of Veterinary Research 86, 4; 10.2460/ajvr.24.10.0316
On serial linear models of normalized concentration data controlling for sex, 115 lipids were significantly associated with increasing hepatic fat accumulation (all q < 0.05; all P < .006). Lipids, whose concentrations increased with hepatic fat content, included diacylglycerol 38:3, PC 40:1, 3 SMs, 34 TGs (in C34 to C54 with 0 to 8 double bonds), and 29 alkyldiacyl/dialkylacylglycerols (TG Os) (in C39 to C55 with 1 to 7 double bonds and all being alkyldiacylglycerols). Lipids, whose concentrations decreased with hepatic fat content, included FA 44:1, 9 PCs (in C34 to C38 with 2 to 6 double bonds), PC plasmalogen 34:0, 12 PEs (in C34 to C40 with 2 to 6 double bonds, most including linoleic acid), 3 alkyl-phosphatidylethanolamines (PE Os), 1 phosphatidylethanol (PE tOH 16:0_16:1, 6), phosphatidylinositols (PI; in C35 to C37 with 2 to 4 double bonds), 2 phosphatidylserines (PS; PS 34:2 and PS 40:5), 11 SMs, and TG O 47:6.
The volcano plot (Figure 2) shows the general trend in the hepatic lipidome with increasing fat content. As hepatic fat increased, glycerolipids generally increased, whereas phospholipids, sphingolipids, and sterol lipids decreased. The most impactful lipids most associated with increasing hepatic fat content were TG and TG O lipids with 2 SMs, and those most associated with decreasing hepatic fat content were PE with 1 PC and 1 SM. Most of the TG and TG O that increased had shorter acyl chain lengths and were characterized by medium-chain FAs in the sn-1 position, such as FA 8:0 (caprylic acid), FA 10:0 (capric acid), and FA 12:0 (lauric acid).
Volcano plot with parameter estimates from linear models on normalized hepatic lipid concentrations on the x-axis and the log of P values on the y-axis. The vertical dashed lines represent the 90th percentile limits (5% on either side). The horizontal dashed line represents a P value of .05, and the horizontal dotted line has a q value of 0.05 (P values adjusted for false discovery rate). Lipids on the top right corner were significantly and positively associated with hepatic fat accumulation and ranked within the top 10% of those most strongly associated with hepatic fat content. Lipids on the top left corner were significantly and negatively associated with hepatic fat accumulation and ranked within the top 10% of those most strongly associated with hepatic fat content. Lipids are color-coded by lipid categories and subclasses.
Citation: American Journal of Veterinary Research 86, 4; 10.2460/ajvr.24.10.0316
On multivariate analysis, no clear clustering was seen between the 3 categories of hepatic fat content on the principal component analysis model, which explained 49.1% of the cumulative variance (Figure 3). However, a clustering trend could be observed that was mainly driven by some TG O, TG, PC, PE, and PS, but cumulative loadings for the first and second components of the most impactful variables were only around 0.15 to 0.18.
Principal component analysis plot for the 2 first components of the hepatic lipidome of bearded dragons across 3 different amounts of hepatic fat accumulation based on normalized lipid concentrations.
Citation: American Journal of Veterinary Research 86, 4; 10.2460/ajvr.24.10.0316
On enrichment analysis (Figure 4), TG O was the lipid subclass most impacted by increasing hepatic fat content, with 3 times more significant lipids than other detected subclasses. Other subclasses that were more impacted than average were PE, PI, SM, and TG. The PC:PE ratio also linearly increased with increasing hepatic fat (P = .003).
Enrichment plot of plasma lipid subclasses in the hepatic lipidome of bearded dragons with increasing hepatic fat accumulation based on normalized lipid concentrations.
Citation: American Journal of Veterinary Research 86, 4; 10.2460/ajvr.24.10.0316
Pathway analysis from the 0% to 20% through 20% to 40% hepatic fat content categories found that the reaction from PE to lysophosphatidylethanolamine (LPE) was the most upregulated (P = .02). When comparing the 0% to 20% through 40% to 60% hepatic fat content, the most upregulated reaction chains were PE→PC→DG→TG (P = .008) and PE→PC→LPC (P = .026).
Regarding TG characteristics, TG in C52 with 2 to 4 double bonds predominated in all livers regardless of hepatic fat content (Figure 5). These TG predominantly included palmitic acid (FA 16:0) in the sn-1 position, oleic (FA 18:1) or linoleic (FA 18:2) acid in the sn-2 position, and oleic, linoleic, or linolenic (FA 18:3) acid in the sn-3 position. As hepatic fat content increased, there was a shift in the distribution of FAs, with FA profiles becoming less concentrated and expanding to FAs with shorter chains and lower degrees of unsaturation. Phosphatidylcholine 34:2 and PC 36:2 were the most common PCs, and PE 36:2 and PE 36:3 were the most common PEs. No noticeable shifts in fatty acyl chain length and unsaturation with increasing hepatic fat categories were appreciated graphically (plots not shown).
Bubble plots of the fatty acyl chain length and unsaturation distribution of TGs across 3 different amounts of hepatic fat accumulation. Circle size and color intensity reflect lipid median concentrations.
Citation: American Journal of Veterinary Research 86, 4; 10.2460/ajvr.24.10.0316
Discussion
In this study, we comprehensively analyzed the hepatic lipidome of bearded dragons with varying amounts of hepatic fat by LC-MS-MS untargeted lipidomics and detected and quantified, using internal standards, close to 1,000 lipid molecules from 5 lipid categories and 29 lipid subclasses at the species level (sum composition) to the molecular species level of characterization, depending on lipids. Our results showed a quantitative shift in the relative proportions of hepatic lipids and qualitative changes within lipid subclasses. These findings may shed some light on the pathogenesis and pathophysiology of hepatic fat accumulation in bearded dragons, especially as compared to nonreptilian species in which fatty liver is also common.
As fatty liver is defined by the accumulation of TGs, as the primary lipid storage form, we found a sharp increase in these lipids with escalating fatty accumulation as expected. There were also some qualitative changes to TG with overall shorter fatty acyl chains and lower degrees of unsaturation. In particular, TG with medium-chain FAs (in C8 to C12) in the sn-1 position were the lipids with the highest magnitude of changes seen on the volcano plot. Because all the bearded dragons had the same origin and diet, a dietary reason for these observed shifts is less likely. These TG changes are somewhat similar or even more pronounced to what is seen in humans with MAFLD and poultry in some studies13–15,19,27,28 and could inform potential pathophysiological mechanisms of fatty liver in bearded dragons. First, this could indicate increased de novo lipogenesis, which produces mainly saturated FAs, albeit mostly palmitic (FA 16:0) and oleic acid (FA 18:0).29 The increased presence of caprylic, capric, and lauric acids (FA 8:0 to 12:0) in TG in the liver with increased fat is puzzling as medium-chain FAs are not the dominant FA produced by de novo synthesis and are also uncommon in dietary items fed to bearded dragons, such as insects and greens.30,31 Also, the opposite trend was seen in an experimental study in mice.32 This may point to differences with mammals. During de novo FA synthesis, FAs are synthesized by FA synthase by sequentially adding 2-carbon units from acetyl coenzyme A to a growing FA chain. During this process, chain termination mostly occurs when the FA chain reaches C16 (palmitic acid), which triggers thioesterase to cleave it from the FA synthase enzymatic complex.29 Medium-chain thioesterases, such as those found in lactating mammary glands, are more selective for medium-chain FAs and produce FAs in C8 to C12. Whether such medium-chain thioesterases are present and more upregulated in fatty livers in bearded dragons or that long-chain thioesterases have altered FA chain termination is unknown but should be further explored. Other enzymes involved in the generation of polyunsaturated FAs include elongases (elongation of very long chain FAs) and desaturases (FA desaturase), whose activities may be altered during hepatic fat accumulation in bearded dragons as observed in humans and poultry.13,16,19 In addition, medium-chain FAs are typically metabolized more quickly than long-chain FAs, and their increased presence may simply indicate impaired and decreased FA oxidation.
Another group of glycerolipids found to be an important driver of hepatic fat accumulation in bearded dragons was TG O, which largely included alkyldiacylglycerols. Alkyldiacylglycerols are a type of ether lipids in which 1 of the fatty acyl chains is attached to the glycerol backbone through an ether link instead of an ester link like in TGs.33 Enrichment analysis identified TG O as the most impacted lipid subclass, with numerous individual TG O species also prominently highlighted on the volcano plot. The significance of this finding is challenging to interpret as studies measuring TG O in the liver of animals or humans with fatty liver are lacking. However, TG O levels are increased in the plasma of children with MAFLD, in which TG O lipid species were the most important lipids differentiating stages of fatty liver disease.34 Considering these findings, TG O clearly plays an important role in the pathophysiology of hepatic fat accumulation in bearded dragons and needs further investigation.
In contrast to glycerolipids, phospholipids, including most subclasses, particularly PE and PI, showed a negative association with increasing hepatic fat accumulation in bearded dragons. This trend aligns with human, mouse, and poultry study findings,13–17,19,27,32,35,36 suggesting some shared pathophysiological features across species. Phospholipids are essential components of cell membranes, forming the lipid bilayer that maintains cellular structure.37 Quantitative and qualitative changes in phospholipids, especially variations in FA chain length and unsaturation, can impact membrane integrity and fluidity, potentially affecting cellular function. The fatty acyl chain lengths and unsaturation level of phospholipids did not seem to be significantly altered in bearded dragons, contrary to what is observed in some studies14,19 in mammals and birds. In reptiles, phospholipids generally have shorter fatty acyl chain lengths and a lower degree of polyunsaturation than nonhibernating mammals, likely due to adaptations associated with ectothermy.38 Therefore, taxon-specific phospholipid acyl composition may partly explain the differences observed in this study. In addition, the PC/PE ratio increased with hepatic fat accumulation in bearded dragons, whereas the opposite is observed in humans.13,14 The PC/PE ratio serves as an indicator of PE N-methyltransferase (PEMT) activity, the primary enzyme involved in PE synthesis from PC. This ratio is also associated with cellular membrane integrity as PC is mainly located in the outer monolayer of the phospholipid bilayer of cell membranes, whereas PE is mainly located in the inner monolayer. A low PC:PE ratio is associated with a loss of membrane integrity and higher permeability to proinflammatory factors.13,36 Our observations in bearded dragons are consistent with an increase rather than a decrease in PEMT activity. Pathway analysis was also compatible with an upregulation of PE conversion of PE to lysophosphatidylethanolamine or PC and PC to glycerolipids, which are associated with the enzymes PEMT and diacylglycerol O-acyltransferase 2 in mammals. Plasmalogens or PE O, another group of ether lipids, are also important membrane phospholipids, and a decrease is frequently implicated in MAFLD pathogenesis as they protect membranes from oxidative stress and have other beneficial anti-inflammatory and antisteatosis properties.13,33,36 While some PE O species concentrations significantly decreased in bearded dragons with higher hepatic fat, there was no unique trend beyond the general decreasing pattern consistent with other phospholipids. Overall, these findings suggest that while there is a global decrease in phospholipids in the liver of bearded dragons and shifts toward glycerolipids with increasing hepatic fat, the composition and proportion of membrane-associated phospholipids remain sound. This may indicate that bearded dragons respond differently than mammals and birds under high cellular lipid loads and may be better able to maintain cell membrane integrity. Finally, PI were also prominent on enrichment analysis and are the third most common phospholipid group found in the liver. Phosphatidylinositols are unique in that, in addition to their role in cell membrane structure, they are also more bioactive and at the center of multiple signaling reactions as precursors of phosphoinositides and other metabolites.39 A reduction in PI may, therefore, be associated with metabolic pathway and cell function disruption. A specific decrease in PI was not reported in similar studies in humans.
Sphingomyelins, a group of structural lipids that belong to the sphingolipid category and are common in cell membranes, decreased overall in bearded dragon livers with increasing fat. In some studies, this pattern in fatty livers was also seen in mammals and poultry, but an increase was most commonly encountered in other studies.13,14,40,41 Ceramides, the simplest types of sphingolipids and products of SM catabolism, which differ from SMs by the lack of an additional head group attached to the sphingoid base, were not found to be significantly altered in this study overall. It is surprising as increased plasma or hepatic ceramide levels were associated with MAFLD and steatohepatitis and connected to increased lipotoxicity, decreased insulin sensitivity, mitochondrial dysfunction, and other deleterious effects in multiple studies.13,36,40 Sphingolipids were not included within the pathway analysis as BioPan did not classify them within any specific metabolic reactions, but it is evident that the observed decrease in SM in bearded dragons with higher hepatic fat content may indicate a disruption in overall sphingolipid metabolism.
There are several limitations to our study. First, in terms of the dataset, lipids identified on negative and positive mass spectrometry ionization modes were analyzed together. While we removed obvious duplicates, some lipid duplicates were likely still present due to differences in the depth of lipid characterization (sum composition to molecular species level) across categories between the 2 modes, potentially resulting in the inclusion of isobaric and isomeric lipid species in the analyzed dataset. Second, our bearded dragon cohort was unbalanced for hepatic fat groups, with the middle group (20% to 40%) being overrepresented. Data were mainly analyzed using continuous data (percentage of hepatic fat), but the general abundance plot, bubble plots, and principal component analysis were still generated using hepatic fat categories, and interpretation should consider this skewed study population. Lastly, our results were heavily compared with human data and other animal studies that reported differences in hepatic lipidome from well-defined categories, typically 2 or 3, such as normal liver, hepatic steatosis, and steatohepatitis.
Given that the threshold at which hepatic lipid accumulation becomes pathological in bearded dragons remains unclear, we preferred to study the alterations of the hepatic lipidome based on serial linear modeling of continuous data (hepatic fat percentage on histopathology) rather than categorizing livers into potentially biased severity groups. We thought this offered a more objective and precise characterization of how hepatic lipids change with the progressive accumulation of fat. In addition, including sex as a covariable in all models allowed the exploration of changes independently from sex effects.
In conclusion, this study on the hepatic lipidomics of bearded dragons improves our understanding of the intricate lipid metabolic changes occurring with increasing fat accumulation. Overall, our results could suggest that, in addition to the likely increased uptake of FAs from overfeeding in captivity, there is evidence of increased de novo lipogenesis contributing to overall TG increase. Several lipid metabolic pathways, including sphingolipid and phospholipid metabolism, also seemed to be altered. This is to consider along with previously identified evidence on plasma metabolomics studies6,7 in bearded dragons of disruptions in other metabolic pathways involving ketones, amino acids, and oligosaccharides potentially coupled to mitochondrial dysfunction. On the other hand, many changes strongly associated with fatty liver and hepatocyte injury in homeotherms, some being considered central to the pathophysiology of MAFLD, were not observed here, such as an increase in ceramides and other sphingolipids, a decrease in the PC:PE ratio, and a decrease in plasmalogens, whereas unique changes were seen, such as increases in TG with medium-chain FAs increasing in concentrations, an important increase in TG O, and a decrease in SMs. Though challenging to interpret, these findings may indicate a slower progression of metabolic disruption and a generally higher tolerance in bearded dragons to increased hepatic fat levels. This somewhat aligns with the fact that hepatic inflammation is seldom observed in bearded dragons, whereas it is the normal progression of the disease in humans; that clinical markers of hepatic function, injury, or insulin resistance are seldom changed; and, lastly, that the amount of hepatic fat associated with disease in humans is much lower than bearded dragons by a factor of 3-fold to 10-fold. Further research is necessary to deepen our understanding of the pathophysiology associated with high hepatic fat accumulation in bearded dragons. Utilizing other omics approaches, such as transcriptomics and genomics, could help uncover the underlying mechanisms driving the patterns observed in this study.
Acknowledgments
The authors want to thank Animal Specialty Inc for generously providing the animals needed for their study, the West Coast Metabolomics Center at UC Davis for lipidomic analysis, Lisa Pacumio for her help in collecting the liver samples and Leonardo Susta for performing histopathology.
Disclosures
The authors have nothing to disclose. No AI-assisted technologies were used in the generation of this manuscript.
Funding
The authors have nothing to disclose.
ORCID
H. Beaufrere https://orcid.org/0000-0002-3612-5548
References
- 1.↑
Origgi FC. Lacertilia. In: Terio KA, McAloose D, Leger JS, eds. Pathology of Wildlife and Zoo Animals. Academic Press; 2018;865–889.
- 2.↑
Barboza T, Susta L, Reavill D, Beaufrère H. Prevalence and risk factors of hepatic lipid changes in bearded dragons (Pogona vitticeps). Vet Pathol. 2023;60(1):133–138. doi:10.1177/03009858221105058
- 3.↑
Boccatonda A, Andreetto L, D’Ardes D, et al. From NAFLD to MAFLD: definition, pathophysiological basis and cardiovascular implications. Biomedicines. 2023;11(3):883. doi:10.3390/biomedicines11030883
- 4.↑
Barboza TK, Abood SK, Beaufrère H. Survey of feeding practices and supplement use in pet inland bearded dragons (Pogona vitticeps) of the United States and Canada. J Herpetol Med Surg. 2022;32(3).
- 5.↑
Price ER. The physiology of lipid storage and use in reptiles. Biol Rev Camb Philos Soc. 2017;92(3):1406–1426. doi:10.1111/brv.12288
- 6.↑
Barboza TK, Susta L, Linden A, Gardhouse S, Beaufrère H. Association of plasma metabolites and diagnostic imaging findings with hepatic lipidosis in bearded dragons (Pogona vitticeps) and effects of gemfibrozil therapy. PLoS One. 2023;18(2):e0274060. doi:10.1371/journal.pone.0274060
- 7.↑
Beaufrère H, Pacumio L, Susta L, Tarbert D, Ammersbach M, Keel K. Hepatic lipid accumulation is associated with multiple metabolic pathway alterations but not dyslipidemia and insulin resistance in central bearded dragons (Pogona vitticeps). Am J Vet Res. 2024;85(6):ajvr.23.12.0285.
- 8.↑
Barboza T, Beaufrère H, Reavill D, Susta L. Morphological features of hepatic lipid changes in bearded dragons (Pogona vitticeps), and a proposed grading system. Vet Pathol. 2023;60(1):123–132. doi:10.1177/03009858221128921
- 9.↑
Loomba R, Friedman SL, Shulman GI. Mechanisms and disease consequences of nonalcoholic fatty liver disease. Cell. 2021;184(10):2537–2564. doi:10.1016/j.cell.2021.04.015
- 10.↑
Garrafa E, Segala A, Vezzoli M, et al. Mitochondrial dysfunction in peripheral blood mononuclear cells as novel diagnostic tools for non-alcoholic fatty liver disease: visualizing relationships with known and potential disease biomarkers. Diagnostics (Basel). 2023;13(14):2363. doi:10.3390/diagnostics13142363
- 11.↑
Sartori MR, Navarro CDC, Castilho RF, Vercesi AE. Aggravation of hepatic lipidosis in red-footed tortoise Chelonoidis carbonaria with age is associated with alterations in liver mitochondria. Comp Biochem Physiol B Biochem Mol Biol. 2022;260:110731. doi:10.1016/j.cbpb.2022.110731
- 12.↑
Fletcher JA, Deja S, Satapati S, Fu X, Burgess SC, Browning JD. Impaired ketogenesis and increased acetyl-CoA oxidation promote hyperglycemia in human fatty liver. JCI Insight. 2019;5(11):e127737. doi:10.1172/jci.insight.127737
- 13.↑
Kartsoli S, Kostara CE, Tsimihodimos V, Bairaktari ET, Christodoulou DK. Lipidomics in non-alcoholic fatty liver disease. World J Hepatol. 2020;12(8):436–450. doi:10.4254/wjh.v12.i8.436
- 14.↑
Béland-Bonenfant S, Rouland A, Petit JM, Verges B. Concise review of lipidomics in nonalcoholic fatty liver disease. Diabetes Metab. 2023;49(3):101432. doi:10.1016/j.diabet.2023.101432
- 15.↑
Vvedenskaya O, Rose TD, Knittelfelder O, et al. Nonalcoholic fatty liver disease stratification by liver lipidomics. J Lipid Res. 2021;62:100104. doi:10.1016/j.jlr.2021.100104
- 16.↑
Hong L, Sun Z, Xu D, et al. Transcriptome and lipidome integration unveils mechanisms of fatty liver formation in Shitou geese. Poult Sci. 2024;103(2):103280. doi:10.1016/j.psj.2023.103280
- 17.↑
Wei R, Ning R, Han C, et al. Lipidomics analysis reveals new insights into the goose fatty liver formation. Poult Sci. 2023;102(3):102428. doi:10.1016/j.psj.2022.102428
- 18.
Guo Y, Liao JH, Liang ZL, Balasubramanian B, Liu WC. Hepatic lipid metabolomics in response to heat stress in local broiler chickens breed (Huaixiang chickens). Vet Med Sci. 2021;7(4):1369–1378. doi:10.1002/vms3.462
- 19.↑
You M, Zhang S, Shen Y, et al. Quantitative lipidomics reveals lipid perturbation in the liver of fatty liver hemorrhagic syndrome in laying hens. Poultry Sci. 2023;102(2):102352. doi:10.1016/j.psj.2022.102352
- 20.↑
Bankhead P, Loughrey MB, Fernández JA, et al. QuPath: open source software for digital pathology image analysis. Sci Rep. 2017;7(1):16878. doi:10.1038/s41598-017-17204-5
- 21.↑
Matyash V, Liebisch G, Kurzchalia TV, Shevchenko A, Schwudke D. Lipid extraction by methyl-tert-butyl ether for high-throughput lipidomics. J Lipid Res. 2008;49(5):1137–1146. doi:10.1194/jlr.D700041-JLR200
- 22.↑
Tsugawa H, Cajka T, Kind T, et al. MS-DIAL: data-independent MS/MS deconvolution for comprehensive metabolome analysis. Nat Methods. 2015;12(6):523–526. doi:10.1038/nmeth.3393
- 23.↑
DeFelice BC, Mehta SS, Samra S, et al. Mass spectral feature list optimizer (MS-FLO): a tool to minimize false positive peak reports in untargeted liquid chromatography–mass spectroscopy (LC-MS) data processing. Anal Chem. 2017;89(6):3250–3255. doi:10.1021/acs.analchem.6b04372
- 24.↑
Liebisch G, Fahy E, Aoki J, et al. Update on LIPID MAPS classification, nomenclature, and shorthand notation for MS-derived lipid structures. J Lipid Res. 2020;61(12):1539–1555. doi:10.1194/jlr.S120001025
- 25.↑
Lopez-Clavijo AF, Gaud C, Sousa BC, et al. BioPAN: a web-based tool to explore mammalian lipidome metabolic pathways on LIPID MAPS. F1000Res. 2021;10:4. doi:10.12688/f1000research.28022.2
- 27.↑
Gorden DL, Myers DS, Ivanova PT, et al. Biomarkers of NAFLD progression: a lipidomics approach to an epidemic. J Lipid Res. 2015;56(3):722–736. doi:10.1194/jlr.P056002
- 28.↑
Puri P, Wiest MM, Cheung O, et al. The plasma lipidomic signature of nonalcoholic steatohepatitis. Hepatology. 2009;50(6):1827–1838. doi:10.1002/hep.23229
- 29.↑
Gurr MI, Harwood JL, Frayn KN. Fatty acid metabolism. In: Gurr MI, Harwood JL, Frayn KN, eds. Lipids: Biochem, Biotech and Health. 6th ed. Wiley Blackwell. 2016:44–123.
- 30.↑
Almazan AM, Adeyeye SO. Fat and fatty acid concentrations in some green vegetables. J Food Compost Anal. 1998;11(4):375–380. doi:10.1006/jfca.1998.0596
- 31.↑
Finke MD. Complete nutrient content of four species of commercially available feeder insects fed enhanced diets during growth. Zoo Biol. 2015;34(6):554–564. doi:10.1002/zoo.21246
- 32.↑
Eisinger K, Krautbauer S, Hebel T, et al. Lipidomic analysis of the liver from high-fat diet induced obese mice identifies changes in multiple lipid classes. Exp Mol Pathol. 2014;97(1):37–43. doi:10.1016/j.yexmp.2014.05.002
- 33.↑
Magnusson CD, Haraldsson GG. Ether lipids. Chem Phys Lipids. 2011;164(5):315–340. doi:10.1016/j.chemphyslip.2011.04.010
- 34.↑
Draijer LG, Froon-Torenstra D, van Weeghel M, et al. Lipidomics in nonalcoholic fatty liver disease: exploring serum lipids as biomarkers for pediatric nonalcoholic fatty liver disease. J Pediatr Gastroenterol Nutr. 2020;71(4):433–439. doi:10.1097/MPG.0000000000002875
- 35.↑
Rodríguez-Calvo R, Samino S, Girona J, et al. Hepatic lipidomics and molecular imaging in a murine non-alcoholic fatty liver disease model: insights into molecular mechanisms. Biomolecules. 2020;10(9):1275. doi:10.3390/biom10091275
- 36.↑
Di Sessa A, Riccio S, Pirozzi E, et al. Advances in paediatric nonalcoholic fatty liver disease: role of lipidomics. World J Gastroenterol. 2021;27(25):3815–3824. doi:10.3748/wjg.v27.i25.3815
- 37.↑
Gurr MI, Harwood JL, Frayn KN. Important biological lipids and their structures. In: Gurr MI, Harwood JL, Frayn KN, eds. Lipids: Biochemistry, Biotechnology and Health. 6th ed. Wiley Blackwell. 2016:13–43.
- 38.↑
Mitchell TW, Ekroos K, Blanksby SJ, Hulbert AJ, Else PL. Differences in membrane acyl phospholipid composition between an endothermic mammal and an ectothermic reptile are not limited to any phospholipid class. J Exp Biol. 2007;210(pt 19):3440–3450. doi:10.1242/jeb.007286
- 39.↑
Zhang X, Majerus PW. Phosphatidylinositol signalling reactions. Semin Cell Dev Biol. 1998;9(2):153–160. doi:10.1006/scdb.1997.0220
- 40.↑
Simon J, Ouro A, Ala-Ibanibo L, Presa N, Delgado TC, Martinez-Chantar ML. Sphingolipids in non-alcoholic fatty liver disease and hepatocellular carcinoma: ceramide turnover. Int J Mol Sci. 2020;21(1):40. doi:10.3390/ijms21010040
- 41.↑
Régnier M, Polizzi A, Guillou H, Loiseau N. Sphingolipid metabolism in non-alcoholic fatty liver diseases. Biochimie. 2019;159:9–22. doi:10.1016/j.biochi.2018.07.021
