Hepatic lipidosis, or fatty liver disease, is the progressive accumulation of triacylglycerols in hepatocytes resulting in hepatocellular dysfunction.1 It is exceedingly common in reptiles but particularly in the central bearded dragon (Pogona vitticeps), one of the most common companion reptile species. A prevalence of more than 38% for significant hepatic lipid accumulation was reported in a recent large epidemiological study on necropsy.2 Within bearded dragons reported with hepatic lipid accumulation, the most severe category, characterized by more than 50% of lipid, hepatocellular swelling, and significant fibrosis, was the most prevalent. Female and adult bearded dragons seem predisposed.2
It is currently unclear at which stage hepatic lipid accumulation becomes pathologic in bearded dragons and associated with clinical signs. There is currently no consensus on what amount of hepatic lipid is abnormal in reptiles. A certain percentage of fat seems to be normal in the liver of many reptiles, especially in reproductive females, and may vary tremendously depending on species, season, and other biological and environmental factors.3 The histopathology of bearded dragon hepatic fat accumulation has been thoroughly investigated, and a grading system has been proposed to help standardize the diagnosis and research of this condition.4 Since hepatic lipidosis refers to a disease/dysfunction and is not well defined and connected to clinical signs in bearded dragons, we will mainly refer to hepatic fat accumulation in this manuscript.
The lipid physiology of reptiles is drastically different from mammals,5 and the mechanism and consequences of hepatic lipidosis are better understood in mammals.1,6 The application and extrapolation from knowledge derived in the fields of human and domestic mammal medicine is challenged by unique reptilian physiological processes, such as brumation, intermittent fasting, vitellogenesis in females, triacylglycerol storage in celomic fat pads, and ectothermy.5 The disease is also fundamentally different from cats and cows in that anorexia in an overweight animal in negative energy balance is unlikely to be a major contributor to this disease in an intermittent feeder such as the bearded dragons.2 Hepatic lipidosis is also common in birds and has been the most studied in chickens. In chickens, pathologic hepatic fat accumulation is mainly associated with selection for egg laying, production, feeding high-energy diets, and restricted exercise.7 As extrapolation from endothermic animal groups faces substantial challenges, there is an immediate need for additional reptile-specific research aimed at exploring potential pathogenetic pathways and pathophysiological mechanisms to inform prevention measures as well as diagnostic and therapeutic interventions.
The diagnosis of hepatic fat accumulation and its consequences is challenging. Clinical signs either appear absent or nonspecific, blood biochemistry abnormalities are infrequent unless the disease is very advanced, and routine imaging techniques (radiography, ultrasound) are unhelpful. A preliminary pilot study8 on a small cohort of 14 bearded dragons was recently published, which aimed to explore diagnostic imaging techniques, including celomic ultrasound, coelioscopy, and CT scan. This study8 also explored a broad spectrum of potential biomarkers by means of a mass spectrometry–targeted global metabolomics approach including more than 600 metabolites. The results of this hypothesis-generating study were compatible with some disruptions in primary metabolic pathways, and β-hydroxybutyric acid (also known as 3-hydroxybutyric acid; BHBA) and succinic acid emerged as the most significant biomarkers. However, several potentially confounding factors were present in this study, such as the opportunistic use of a heterogeneous population of bearded dragons in which older and heavier dragons had more severe lesions.8
Based on these initial results, further exploration of primary metabolic pathway disruption was warranted in a larger cohort of bearded dragons with hepatic fat accumulation that varied in severity. Untargeted primary metabolomics by gas chromatography–mass spectrometry is a standard technique to explore large numbers of metabolites, some being volatile, from small quantities of biologic fluids. It has been used in humans and in a variety of veterinary species in relation to fatty liver disease.9–12 The current investigation builds on the results of this earlier study8 and aims to further explore primary metabolic pathways, dyslipidemia, and insulin resistance in relation to hepatic fat accumulation, hepatic fibrosis, and severity classification in bearded dragons with spontaneous changes. We hypothesize that BHBA and succinic acid would be confirmed as useful biomarkers and that evidence of dyslipidemia and insulin resistance would be substantiated, mirroring observations in mammals.
Methods
The research was approved by the UC Davis Institutional Animal Care and Use Committee (protocol #22816).
Animals
A homogeneous cohort of 48 adult bearded dragons were acquired from a local breeder in Chico, California. All animals were originally selected to be culled as part of breeding management considerations. Sample size analysis was performed based on metabolomics data generated in a prior pilot study on 14 bearded dragons,8 considering 10% to 20% changes between severity groups in serial 2-sample tests for selected metabolites with a power of 80%.
Twenty-four dragons were 1.5 to 3 years old, and 24 were 4 to 7 years old. There were 25 females and 23 males approximately equally distributed across age groups. Bearded dragons weighed a mean ± SD of 305 ± 79 g. They were temporarily housed at the UC Davis Teaching and Research Animal Care Services headquarters in glass enclosures with mercury vapor bulbs. The dragons were provided water in a bowl and fed a combination of gut-loaded crickets, mealworms, and dark leafy greens supplemented with a calcium carbonate powder (ReptiCalcium; Zoo Med) sprinkled on the food. Physical examination on intake did not reveal any abnormalities, and all animals were in adequate body condition.
CT
Each dragon was sedated using an SC injection of alfaxalone 10 mg/kg (Alfaxan Multidose; Jurox) in the front half of the body. Whole-body CT scans were performed on 8 bearded dragons at a time using a GE LighSpeed16 (GE HealthCare). Slice thickness was 0.625 mm. The field of view was 25 cm, kVP 120, mA 100, and the pitch was 0.935:1 with a 1-second rotation time. Bearded dragons were positioned in 2 adjacent 2 X 2 columns of stackable plexiglass troughs. DICOM files were imported into an open-source imaging software (Horos, version 3.3.6; Horos Project). The hepatic attenuation was measured in Hounsfield units (HU) in multiplanar reconstruction coronal views using standardized 50-mm2 circular regions of interest (ROIs), one drawn on the left and one drawn on the right hepatic lobe, excluding blood vessels if visible. The average of these 2 ROIs was recorded.
Blood sampling
Bearded dragons were fasted for 48 hours prior to sampling. A 2.5-mL blood sample was collected from the caudal tail vein using a 25-gauge needle and a 3-mL syringe and transferred into 4 heparinized 0.5-mL collection tubes without plasma separator and 1 EDTA 0.5-mL collection tube (BD microtainer; Becton Dickinson and Company). Blood tubes were placed on ice until processing for not more than 2 hours after collection. Packed cell volume and total solids were determined, and the blood was centrifuged at 3,000 X g for 7 minutes. Plasma was harvested using plastic micropipettes and transferred into 0.5-mL storage polypropylene tubes (Eppendorf, Hamburg, Germany) and stored at –80 °C until analysis.
Biochemistry, insulin, and lipoprotein analysis
Heparinized plasma was submitted to the University of Miami Avian and Wildlife Laboratory, and a biochemistry panel including aspartate aminotransferase (AST), alanine transaminase (ALT), alkaline phosphatase (ALP), creatine kinase (CK), gamma-glutamyl transferase (GGT), lactate dehydrogenase (LDH), glutamate dehydrogenase (GLDH), uric acid, cholesterol, triglycerides, sodium, potassium, chloride, calcium, glucose, and bile acids was run on a reference biochemistry analyzer (Vitros; Ortho Clinical Diagnostics).
Heparinized plasma was shipped on dry ice to the Endocrinology Section of the Michigan State University Veterinary Diagnostic Laboratory for measurement of plasma insulin concentrations. Insulin was measured with a commercially available radioimmunoassay kit (Hi-14K; MilliporeSigma) utilizing a second-antibody precipitation separation method. The radioimmunoassay was performed with reagent volumes and incubation times described in the manufacturer’s protocol.
Advanced lipoprotein profiling was performed on EDTA plasma at the UC Davis Comparative Clinical Lipidology Research Laboratory using a high-resolution PAGE kit (Lipoprint LDL kit; Quantimetrix) according to the manufacturer instructions. The lipoprotein profile included very-low-density lipoprotein cholesterol, intermediate-density lipoprotein cholesterol, low-density lipoprotein cholesterol 1 through 7 subfractions, and high-density lipoprotein (HDL) cholesterol relative and absolute concentrations. Calculated values were also obtained and included non-HDL cholesterol and HDL:total cholesterol ratio.
Untargeted primary metabolomics
Heparinized plasma was submitted to the UC Davis West Coast Metabolomics Center for untargeted primary metabolomics by gas chromatography–time-of-flight mass spectrometry using a published data acquisition protocol.13 Approximately 200 primary metabolites, including BHBA and succinic acid, were quantified.
Samples were extracted using 1 mL of 3:3:2 acetonitrile (ACN):isopropanol (IPA):H2O (v/v/v). Half of the sample was dried to completeness and then derivatized using 10 μL of 40 mg/mL of methoxyamine in pyridine. Samples were shaken at 30 °C for 1.5 hours. Then, 91 μL of N-methyl-N-trimethylsilyl-trifluoroacetamide and fatty acid methyl esters were added to each sample and shaken at 37 °C for 0.5 hours to finish derivatization. Samples were then vialed, capped, and injected onto the instrument. A 7890 A gas chromatography/autosampler system (Agilent) coupled with a time-of-flight mass spectrometer (Pegasus IV; Leco Corporation) was used. A volume of 0.5 μL of derivatized sample was injected using a splitless method onto a mass spectrometry capillary column with an Integra-Guard (Restek Rtx-5Sil) at 275 °C with a helium flow of 1 mL/min. The gas chromatography oven was set to hold at 50 °C for 1 minute, then ramp to 20 °C/min to 330 °C and then hold for 5 minutes. The transfer line was set to 280 °C, while the electrode ionization ion source was set to 250 °C. The mass spectrometry parameters collected data from 85 m/z to 500 m/z at an acquisition rate of 17 spectra/s.
Raw data files were imported directly after data acquisition into the associated software (ChromaTOF, version 2.32; Leco Corporation). The data were preprocessed without smoothing and using 3 standard deviations peak width, baseline subtraction just above the noise level, and automatic mass spectral deconvolution and peak detection at signal/noise levels of 5:1 throughout the chromatogram. Apex masses were reported for use in the BinBase algorithm. Result were exported to a data server with absolute spectra intensities and further processed by a filtering algorithm implemented in the metabolomics BinBase database.14
Histopathology and digital image analysis
Bearded dragons were euthanized by IV or intracardiac administration of 1 mL of potassium chloride 2 meq/L. Additional alfaxalone was administered until loss of reflexes if necessary prior to KCl administration. Sex was confirmed on necropsy. The liver was collected approximately 3 to 4 hours later, and the caudal most section of the left liver lobe was fixed in formalin for histopathology. The liver sections were embedded in paraffin, sectioned onto glass slides, and stained with H&E. Masson’s trichrome–stained slides were also obtained for fibrosis evaluation.
All liver glass slides were scanned at 0.25 μm/pixel using a digital scanner (Ocus 40; Grundium). Scans were reviewed by a board-certified veterinary pathologist (LS), and hepatic changes were graded using a previously published classification system developed for bearded dragons.4 Briefly, grading was conducted by semiquantitative assessment of hepatocellular vacuolation (score, 0 to 4) and fibrosis (0 to 4) as well as nuclear count to infer the extent of hepatocellular swelling. To count nuclei, 10 high-power fields images (field of view, 0.02 mm2) for each case were imported into an open-source image analysis software (ImageJ, version 1.53; Rasband WS, US National Institutes of Health) and nuclei were manually counted using the multipoint count tool. Counts from each case were then averaged and stratified to obtain a swelling score (0, 2, 4). The final grade could range from 0 to 12 and was divided into 4 severity tiers (ie, severity classification: no lipid, mild, moderate, severe).3 Slides were also reviewed for the presence of any other pathology, and lizards with non–lipid-related hepatopathies were excluded from the study.
For hepatic lipid content quantification (in percentages), scans were exported into an open-source bioimage analysis software (QuPath, version 0.4.2, 2023),15 and 3 square ROIs (0.5 mm2) were randomly generated with the viewing software (excluding vessels and biliary ducts) and exported to ImageJ. A macro was programmed for automatic fat quantification by successively subtracting and white balancing the background, converting to grayscale, binarization, and calculating the percentage of white (empty) area corresponding to fat. Areas obtained from the 3 ROIs were averaged.
A similar method was used for fibrosis quantification (in percentages). As areas of fibrosis are not as homogeneously distributed throughout the hepatic parenchyma as fat vacuolation, 3 larger ROIs (2.25 mm2) were randomly generated and exported to the image-processing software. A macro was programmed for automatic Masson’s trichrome–positive area (collagen) quantification by successively subtracting and white balancing the background and thresholding the color for positive areas prior to binarization and calculating the positive area. Areas obtained from the 3 ROIs were averaged.
Statistical analysis
The effect of age groups (1.5 to 3 and 4 to 7 years old), sex, weight, and their interactions on hepatic lipid accumulation grade, fat percentage, and fibrosis percentage (determined on digital image analysis) were assessed using an ordinal logit model for grade and an ANOVA for fat and fibrosis percentages. Residuals for the logit model were examined graphically, and ANOVA assumptions of normality and homoscedasticity were assessed using Shapiro-Wilk and Levene’s tests, respectively.
Linear regression analysis was used to explore the association between the amount of fat and fibrosis in the liver, as assessed by digital image analysis, and CT liver attenuation (in HU). Assumptions were assessed on residual plots and residual diagnostics.
Biochemistry measurands and metabolites measured on mass spectrometry were analyzed together. Data were first filtered to remove noninformative variables and to increase statistical power. Variables that had a constant or single value across most samples, variables with more than 20% of missing values, and variables with values above or below the quantification limits were removed. A total of 151 variables were retained. They were then normalized by log transformation and mean centering.
The association between continuous outcome variables (hepatic attenuation in HU, percentage of fat, and percentage of fibrosis measured by digital image analysis) and measurands were assessed using a series of linear models with an alpha of .05 (P < .05) and a false discovery rate of 0.05 (q < .05) for significance. An iterative for-loop was programmed in R to perform 151 linear models and export R2 and P values to a dataset. P values were then adjusted with false discovery rate for the number of linear models. For significant variables, assumptions of linear models were checked on follow-up analysis with residual plots. Lipoprotein values and insulin concentrations were analyzed separately but in a similar manner.
Because employed multivariate analysis required a binary outcome, histopathologic grades were simplified into 2 severity classes (no lipid to mild and moderate to severe) for multivariate analysis on the normalized outcome variables (measurands). Clustering was investigated using principal component analysis and partial least square–discriminant analysis (PLS-DA) on the normalized dataset. Metabolites with important contributions in providing class separation were identified on the basis of high variable importance-in-projection scores.
R (version 4.2.0; R Foundation for Statistical Consulting) was used for statistical analysis, GGPLOT2 was used for descriptive plots, and MetaboAnalyst 5.0 was used for multivariate analysis.16,17 The dataset used for statistical analysis was published in a public scientific repository (https://doi.org/10.5061/dryad.pc866t1wd).
Results
Five bearded dragons were excluded due to either infectious/inflammatory/neoplastic hepatopathies or systemic illnesses. The excluded dragons included one animal with a PCV of 6% and hepatic metastasis of neuroendocrine tumor, one with multifocal hepatic granulomas and a mucocele, one with severe gall bladder ectasia and hyperplasia, one with marked hepatic fibrosis and biliary hyperplasia with lymphocyte infiltrates, and one with severe cholangiohepatitis and multifocal heterophilic granulomas with bacteria.
The frequency distribution of bearded dragons within each histologic grade and binary severity classes is depicted (Figure 1). The estimated percentages of fat and fibrosis determined on digital image analysis on liver sections of bearded dragons over histologic grade and CT liver attenuation values are presented (Figure 2).
Bar plot of the distribution of bearded dragons over histopathological grade and severity classes of hepatic lipid accumulation. Note that no dragon has a grade of 0 or 10 to 12.
Citation: American Journal of Veterinary Research 85, 6; 10.2460/ajvr.23.12.0285
Bar plot of the mean ± SE of fat percentage (blue) and fibrosis percentage (red) obtained on digital image analysis of hepatic histology sections over histopathologic grades of hepatic lipid accumulation. Hepatic CT attenuation in Hounsfield units (HU) is superimposed to the plot as a dashed line.
Citation: American Journal of Veterinary Research 85, 6; 10.2460/ajvr.23.12.0285
Age, sex, and body weight had no effect on histologic grade, hepatic fat percentage, hepatic fibrosis percentage, and hepatic attenuation in HU (all P > .05).
Hepatic CT attenuation significantly decreased by 1.2 ± 0.2 HU per 1% increase of fat, as assessed on histology by digital image analysis (P < .001), and significantly increased by 1.7 ± 0.8 HU per 1% increase of fibrosis as assessed on histology by digital image analysis (P = .034) (Figure 3). The R2 for the multiple regression analysis predicting hepatic attenuation based on hepatic fat and fibrosis was 0.60.
Scatter plot of bearded dragon CT hepatic attenuation measured in HU as a function of liver fat content (%) and fibrosis content (%) from histological samples measured by digital analysis and analyzed using multiple linear regression. The different lines represent regression lines (R2 = 0.6) factoring fibrosis content.
Citation: American Journal of Veterinary Research 85, 6; 10.2460/ajvr.23.12.0285
Several measurands were significantly increased with decreasing hepatic CT attenuation and hepatic fat percentage and are included (Table 1).
Statistical results of measurands significantly associated with hepatic lipid accumulation using linear regression analysis.
| Measurands | P | q (FDR) | R2 | Type |
|---|---|---|---|---|
| Hepatic CT attenuation (HU) | ||||
| ALP | < .001 | .043 | 0.19 | Enzyme |
| Maltotriose | < .001 | < .001 | 0.36 | Trisaccharide |
| Aspartic acid | < .001 | .043 | 0.16 | Amino acid |
| Hepatic fat (%) | ||||
| ALP | < .001 | .021 | 0.27 | Enzyme |
| Trehalose | < .001 | .021 | 0.21 | Disaccharide |
| Ribose | < .001 | .026 | 0.23 | Monosaccharide |
| Maltotriose | < .001 | < .001 | 0.40 | Trisaccharide |
| Maltose | < .001 | .021 | 0.23 | Disaccharide |
FDR = False discovery rate. HU = Hounsfield units. q = P value adjusted for false discovery rate. R2 = Linear regression coefficient of determination.
No measurand concentrations were significantly associated with hepatic fibrosis percentage. Succinic acid, BHBA, insulin, and lipoprotein values were not significantly associated with hepatic CT attenuation, hepatic fat, and fibrosis percentages.
Regarding multivariate analysis, binary severity classes did not cluster well on principal component analysis (almost complete overlap of 2 categories), but fair clustering was seen on PLS-DA (minimal overlap of the 2 categories) (Figure 4). The PLS-DA model explained 15.3% of the variance. Using variable importance-in-projection scores for the first component, the most important discriminating variables between the binary severity classes were BHBA, phenylethylamine, 3-aminoisobutyric acid, dihydro-3-coumaric acid, and inosine. All measurand concentrations increased with hepatic histology severity except for BHBA, which was lower with moderate-to-severe hepatic fat changes (Figure 5). Several other variables were important discriminatory variables in the model but in a lesser measure and included other amino acids (cystine, serine, methionine, alanine, glutamic acid), biochemical hepatic parameters (bile acids, LDH, CK), and triglycerides. Lipoprotein values and insulin were not identified as important variables for discrimination between hepatic histology severities.
Score plot of partial least square–discrimination analysis (PLS-DA) of metabolomics, biochemistry, and lipoprotein data between the 2 first principal components in bearded dragons with 2 simplified classes of hepatic lipid accumulation severities based on histopathology. The right panel shows the variable importance-in-projection score plot for the first component showing the most important variables for classification. BAIBA = 3-Aminoisobutyric acid. BHBA = β-Hydroxybutyric acid. CK = Creatine kinase.
Citation: American Journal of Veterinary Research 85, 6; 10.2460/ajvr.23.12.0285
Side-by-side boxplots of selected metabolites of importance (based on PLS-DA models) measured as peak intensity obtained from mass spectrometry across 2 simplified classes of hepatic lipid accumulation based on histopathology. BAIBA = 3-Aminoisobutyric acid.
Citation: American Journal of Veterinary Research 85, 6; 10.2460/ajvr.23.12.0285
Discussion
This study reports the use of untargeted metabolomics to screen primary metabolites for biomarker discovery purpose and investigation of metabolic pathway alterations in association with a range of hepatic lipid accumulation and histologic changes in bearded dragons quantified by CT, standardized digital histologic section analysis, and histopathological grading. The findings of this study should be considered more robust when compared to a previously published pilot study,8 primarily owing to meticulous planning aimed at minimizing the impact of potential confounding variables, such as age and body weight.
CT was shown to be a useful diagnostic tool to quantify hepatic fat in bearded dragons as well as to diagnose pathologic hepatic lipid accumulation in a previous study8 on a small heterogeneous cohort of 14 bearded dragons. However, while CT hepatic attenuation values were still found to be linearly associated with hepatic fat content, the findings of this larger study demonstrated significant overlaps between different grades and severity classes that would preclude the use of CT alone in the diagnosis and clinical staging of this disease. This is due to the fact that hepatic changes other than simply fat content influence CT liver attenuation values and histopathological grading. For instance, in humans it was shown that livers with cirrhosis generally had higher CT liver attenuation than just steatosis and that hepatic iron content significantly increased hepatic HU values.18 Histopathologic grading of hepatic lipid accumulation in bearded dragons is not only based on the percentage of hepatocellular vacuolation but also on hepatocellular swelling (cell ballooning) and fibrosis, criteria that cannot be quantified on CT.4 Indeed, hepatocyte ballooning, which is related to hepatocellular swelling as defined in reptiles,3 a form of hepatocyte degeneration and injury in people, is a key criteria in diagnosing human nonalcoholic steatohepatitis and its transition from the earlier stage of metabolic associated fatty liver disease.19 It is possible, and likely, that lipid accumulation within bearded dragons’ hepatocytes may be physiologically appropriate to the point of cellular swelling, at which point it would transition to a pathologic accumulation (hepatic lipidosis). Hepatic fibrosis is also strongly associated with prognosis and clinical signs in people.19 Hepatic CT attenuation values were linearly associated with hepatic fat content quantified on digital image analysis; hence, CT could still be used to monitor hepatic fat content in bearded dragons. However, it should not be relied on exclusively for diagnosing hepatic lipid accumulation severity and could potentially misdiagnose severe hepatic lipid accumulation with advanced fibrosis, for instance. Other imaging modalities for the noninvasive diagnosis of this condition in reptiles need to be explored. Examples include imaging techniques targeting hepatic fibrosis, such as ultrasound or magnetic resonance elastography; dual-energy CT; and imaging techniques directly measuring triacylglycerol content, such as MRI with fat-suppression sequences or magnetic resonance spectroscopy.20 In the interim, liver biopsy and histopathology will continue to serve as the preferred diagnostic modality for the identification and grading of hepatic lipid accumulation in bearded dragons and likely most other reptiles.
Regarding using plasma biomarkers for the screening, diagnosis, and monitoring of hepatic lipid accumulation in bearded dragons, similar issues regarding diagnostic reliability were seen. Although the alterations in measured biomarkers observed in this study may provide valuable insights into the pathophysiology of the disease, their diagnostic utility is limited due to substantial overlap among lesion categories.
Measurands present on the standard reptilian biochemical panel have limited utility for this condition, according to the results of this study and also confirming a previously published report in this species.8 Only ALP was found to be associated with hepatic fat content. ALP is a nonspecific cytosolic enzyme that is found in low amounts in the liver and is more abundant in the kidney, ovary, and spleen in bearded dragons (H Beaufrère, DVM, PhD, UC Davis School of Veterinary Medicine, unpublished data, 2023), but plasma increases typically result from heightened enzyme synthesis with hepatobiliary disease, in particular cholestasis, rather than hepatocellular leakage in domestic mammals.21 Plasma bile acids, triglycerides, and LDH, while identified as more important classifiers than most other measurands, still had relatively modest associations with histologic severity classification. Nevertheless, routine reptilian biochemical panels should not be relied on in the management of this condition in bearded dragons.
Interestingly, decreasing BHBA was found again to be associated with histologic lesion classification, confirming the results of the prior pilot study.8 However, there was no significant association with hepatic fat and fibrosis percentages, and BHBA values showed substantial overlap between the severity of hepatic lesions, which may limit the use of BHBA measurements for diagnostic purposes, unlike previously suggested. In any case, this provides further evidence that hepatic ketogenesis and/or β-oxidation of fatty acids may be impaired with fatty liver disease in bearded dragons as reported in people.22 In chickens, another sauropsid species commonly affected by fatty liver disease, BHBA was also found to be a potential biomarker for the disease, with decreasing concentrations associated with worsening microscopic lesions in the liver.11 Decreasing β-oxidation of fatty acids was proposed as a plausible mechanism to explain this finding in chickens as suggested in this study. Because BHBA concentrations also increase with fasting in sauropsids,23,24 it should ideally be measured after standardized fasting, especially for monitoring purposes.
Succinic acid was initially found to be an important biomarker for hepatic lipid accumulation in bearded dragons in a previous metabolomics pilot study,8 but this was not confirmed in this larger study, invalidating parts of our initial research hypothesis.
Several oligosaccharides (maltose, maltotriose, trehalose) and a monosaccharide (ribose) were found to be associated with hepatic fat content as determined by CT and histology. Maltotriose in particular was the strongest predictor of hepatic fat content. This elevation of various short-chain carbohydrates implies that there may be alterations in pathways related to overall carbohydrate metabolism with increasing hepatic triacylglycerol accumulation. These metabolic pathways are not reported as commonly affected by fatty liver disease in other species, however, and none of these oligosaccharides were important features of the multivariate analysis on histology severity classes. Therefore, the relevance of these changes remains uncertain.
Plasma concentration in one amino acid (aspartic acid) increased with hepatic fat content, and plasma concentrations of multiple other amino acids (3-aminoisobutyric acid, cystine, serine, methionine, alanine, glutamic acid) increased with histologic severity in this study. These increases may be associated with the disruption of various amino acid metabolism pathways. For instance, alanine, glutamic acid, methionine, and serine are major glucogenic amino acids used in gluconeogenesis via the tricarboxylic acid cycle.25 Furthermore, glutamic acid, cystine, and methionine are involved in the metabolism of glutathione, an important endogenous antioxidant.26 Some of these amino acids have also been found to be elevated in the plasma in other species with fatty liver diseases, such as chickens (serine, glutamic acid), cows (serine), and humans (glutamic acid).11,12,27 The nonprotein amino acid 3-aminoisobutyric acid (also known as β-aminoisobutyric acid; BAIBA) was also found to be an important classifying feature in the PLS-DA model for hepatic lipid accumulation histologic severity in this study. An increase in BAIBA in plasma or urine is also associated with fatty liver disease in other species, such as geese, and type 2 diabetes in humans.28–31 BAIBA is mainly secreted by skeletal muscles, but levels are regulated by hepatic mitochondrial enzymes and are associated with important effects on carbohydrate and lipid metabolism.31 An increase in BAIBA may also be related to alterations in other amino acid metabolic pathways.
Other metabolites were found to be important classifiers for the severity of hepatic lipid accumulation, including phenylethylamine, an aromatic amine; dihydro-3-coumaric acid (also known as 3-(3-hydroxyphenyl)propanoid acid), a phenylpropanoic acid; and inosine, a purine nucleoside. Some reports of fatty liver disease in humans mention an increase of some of these metabolites, but the literature is not consistent. Changes in plasma concentrations of these metabolites may also be related to metabolic alterations in the intestinal microbiome because these compounds can be produced by bacteria from the gastrointestinal flora.32 The intestinal microbiome has a significant contribution in the pathophysiology of fatty liver disease in other species.33
It is noteworthy that the many processes highlighted above, such as the tricarboxylic acid cycle, fatty acid β-oxidation, ketogenesis, and amino acid metabolism (alanine, aspartic acid, glutamic acid, BAIBA), all involve hepatic mitochondria or occur within them.34 Interestingly, alterations in hepatic mitochondria have been demonstrated in red-footed tortoises (Chelonoidis carbonaria) with hepatic lipid accumulation.35 Mitochondrial dysfunction and reduced mitochondrial activity have also been associated with human fatty liver disease.33,36 Therefore, considering our results and comparative aspects of fatty liver disease in other species, it is likely that hepatic mitochondrial dysfunction is an important aspect in the pathophysiology of hepatic lipid accumulation and hepatic lipidosis in bearded dragons.
Metabolic syndrome is a hallmark of fatty liver disease in people, and, in terms of blood changes, it is mainly characterized by evidence of insulin resistance and dyslipidemia.33,37 In this study in bearded dragons, there was no significant influence of plasma insulin concentrations, glucose, cholesterol, and lipoprotein fractions and subfractions on hepatic lipid accumulation and the histologic severity classification. Likewise, no significant biomarker was found in association with hepatic fibrosis, which likely would be more predictive of clinical signs and prognosis as in mammals. These suggest that major aspects of the pathophysiology, progression, and diagnosis of fatty liver disease may be different between humans and bearded dragons. Therefore, more reptile-specific data should be obtained to further our understanding of fatty liver disease in susceptible reptiles, especially since pathogenesis and pathophysiology underlie the development of therapeutic options.
Despite using advanced exploration techniques (e.g., mass spectrometry–based untargeted primary metabolomics, advanced imaging, and histopathology digital image analysis), this investigation into hepatic lipid accumulation in bearded dragons remains primarily observational in nature, and various constraints are inherent to this type of research that should be kept in mind when interpreting and applying these results. Experimental induction of hepatic lipidosis in bearded dragons or longitudinal investigations of bearded dragons under different dietary and husbandry parameters may ultimately better be able to decipher the underlying mechanisms and causes of hepatic fat accumulation in this species. Nonetheless, the findings of this study may offer valuable insights into the factors and variables to evaluate and implement when designing such studies.
In conclusion, bearded dragons with various degrees of hepatic lipid accumulation and histologic severity were shown to experience multiple metabolic pathway disruptions, especially as it relates to ketones, amino acids, and oligosaccharides. Unlike the pathophysiology of hepatic lipidosis in mammals, there was no evidence of insulin resistance and dyslipidemia in connection with hepatic lipid accumulation in this species. Furthermore, it is important to reemphasize that current imaging protocols and measurement of plasma biomarkers are insufficient for the diagnosis and staging of hepatic lipid accumulation and hepatic lipidosis in bearded dragons; hepatic biopsy and histopathology remain the recommended approach. Finally, further clinical research into the use of BHBA, a readily measurable plasma metabolite using standard biochemistry analyzers, and CT-scan for disease monitoring once the disease has been histologically graded should be performed.
Acknowledgments
We would like to thank Animal Specialty Inc for generously providing the animals needed for this study.
Disclosures
The authors have nothing to disclose. No AI-assisted technologies were used in the generation of this manuscript.
Funding
This project was partially supported by a Boehringer Ingelheim Veterinary Scholar Fellowship and the UC Davis Students Training in Advanced Research (STAR).
References
- 1.↑
Diehl AM, Day C. Cause, pathogenesis, and treatment of nonalcoholic steatohepatitis. N Engl J Med. 2017;377(21):2063–2072. doi:10.1056/NEJMra1503519
- 2.↑
Barboza T, Susta L, Reavill D, et al. Prevalence and risk factors of hepatic lipid changes in bearded dragons (Pogona vitticeps). Vet Pathol. 2023;60:133–138. doi:10.1177/03009858221105058
- 3.↑
Origgi FC. Lacertilia. In: Terio KA, McAloose D, Leger JS, eds. Pathology of Wildlife and Zoo Animals. Academic Press; 2018:865–889.
- 4.↑
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
- 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.↑
Valtolina C, Favier RP. Feline hepatic lipidosis. Vet Clin North Am Small Anim Pract. 2017;47(3):683–702. doi:10.1016/j.cvsm.2016.11.014
- 7.↑
Crespo R. Developmental, metabolic, and other noninfectious disorders. In: Swayne DE, Boulianne M, Logue CM, eds. Diseases of Poultry. John Wiley & Sons Ltd; 2020:1286–1329.
- 8.↑
Barboza TK, Susta L, zur Linden A, Gardhouse S, Beaufrere 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
- 9.↑
Kalhan SC, Guo L, Edmison J, et al. Plasma metabolomic profile in nonalcoholic fatty liver disease. Metabolism. 2011;60(3):404–413. doi:10.1016/j.metabol.2010.03.006
- 10.
Tu LN, Showalter MR, Cajka T, et al. Metabolomic characteristics of cholesterol-induced non-obese nonalcoholic fatty liver disease in mice. Sci Rep. 2017;7(1):6120. doi:10.1038/s41598-017-05040-6
- 11.↑
Guo L, Kuang J, Zhuang Y, et al. Serum metabolomic profiling to reveal potential biomarkers for the diagnosis of fatty liver hemorrhagic syndrome in laying hens. Front Physiol. 2021;12:590638. doi:10.3389/fphys.2021.590638
- 12.↑
Zhang X, Liu T, Hou X, et al. Multi-channel metabolomics analysis identifies novel metabolite biomarkers for the early detection of fatty liver disease in dairy cows. Cells. 2022;11(18):2883. doi:10.3390/cells11182883
- 13.↑
Fiehn O, Wohlgemuth G, Scholz M, et al. Quality control for plant metabolomics: reporting MSI-compliant studies. Plant J. 2008;53(4):691–704. doi:10.1111/j.1365-313X.2007.03387.x
- 14.↑
Skogerson K, Wohlgemuth G, Barupal DK, Fiehn O. The volatile compound BinBase mass spectral database. BMC Bioinformatics. 2011;12:321. doi:10.1186/1471-2105-12-321
- 15.↑
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
- 16.↑
Chong J, Soufan O, Li C, et al. MetaboAnalyst 4.0: towards more transparent and integrative metabolomics analysis. Nucleic Acids Res 2018;46(W1):W486–W494. doi:10.1093/nar/gky310
- 18.↑
Fritz GA, Schoellnast H, Deutschmann HA, et al. Density histogram analysis of unenhanced hepatic computed tomography in patients with diffuse liver diseases. J Comput Assist Tomogr. 2006;30(2):201–205. doi:10.1097/00004728-200603000-00006
- 19.↑
Li Y-Y, Zheng T-L, Xiao S-Y, et al. Hepatocytic ballooning in non-alcoholic steatohepatitis: dilemmas and future directions. Liver Int. 2023;43(6):1170–1182.
- 20.↑
Li Q, Dhyani M, Grajo JR, et al. Current status of imaging in nonalcoholic fatty liver disease. World J Hepatol. 2018;10(8):530–542. doi:10.4254/wjh.v10.i8.530
- 21.↑
Hoffmann WE, Solter PF. Diagnostic enzymology of domestic animals. In: Kaneko J, Harvey JW, Bruss M, eds. Clinical Biochemistry of Domestic Animals. 6th ed. Elsevier; 2008:351–378.
- 22.↑
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
- 23.↑
Soehnlein C, Hirsch SE, Myre BL, et al. Method comparison of β-hydroxybutyrate using a point-of-care device and dry chemistry analyzer in three sea turtle species. J Wildl Dis. 2022;58(3):670–674.
- 24.↑
Kaliński A, Glądalski M, Markowski M, Skwarska J, Wawrzyniak J, Banura J. Ketone body levels in wintering great tits parus major in sites differing in artificial food availability. Conserv Physiol. 2022;10(1):coac072.
- 25.↑
Wunschiers R. Metabolism. In: Michal G, Schomburg D, eds. Biochemical Pathways: An Atlas of Biochemistry and Molecular Biology. 2nd ed. John Wiley & Sons Inc; 2012:37–209.
- 26.↑
Wu G, Lupton JR, Turner ND, Lupton JR, Turner ND. Glutathione metabolism and its implications for health. J Nutr. 2004;134(3):489–492. doi:10.1093/jn/134.3.489
- 27.↑
Faquih TO, van Klinken JB, Li-Gao R, et al. Hepatic triglyceride content is intricately associated with numerous metabolites and biochemical pathways. Liver Int. 2023;43(7):1458–1472. doi:10.1111/liv.15575
- 28.↑
Gong Y, Lyu W, Shi X, et al. A serum metabolic profiling analysis during the formation of fatty liver in landes geese via GC-TOF/MS. Front Physiol. 2020;11:581699. doi:10.3389/fphys.2020.581699
- 29.
Katano S, Yano T, Kouzu H, et al. Elevated circulating level of β-aminoisobutyric acid (BAIBA) in heart failure patients with type 2 diabetes receiving sodium-glucose cotransporter 2 inhibitors. Cardiovasc Diabetol. 2022;21(1):285. doi:10.1186/s12933-022-01727-x
- 30.
Miccheli A, Capuani G, Marini F, et al. Urinary 1H-NMR-based metabolic profiling of children with NAFLD undergoing VSL#3 treatment. Int J Obes. 2015;39(7):1118–1125. doi:10.1038/ijo.2015.40
- 31.↑
Tanianskii DA, Jarzebska N, Birkenfeld AL, O’Sullivan JF, Rodionov RN. Beta-aminoisobutyric acid as a novel regulator of carbohydrate and lipid metabolism. Nutrients. 2019;11:524. doi:10.3390/nu11030524
- 33.↑
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
- 34.↑
Spinelli JB, Haigis MC. The multifaceted contributions of mitochondria to cellular metabolism. Nat Cell Biol. 2018;20(7):745–754. doi:10.1038/s41556-018-0124-1
- 35.↑
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
- 36.↑
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. 2023;13(14):2363. doi:10.3390/diagnostics13142363
- 37.↑
Eslam M, Newsome PN, Sarin SK, et al. A new definition for metabolic dysfunction-associated fatty liver disease: an international expert consensus statement. J Hepatol. 2020;73(1):202–209. doi:10.1016/j.jhep.2020.03.039
