Detecting altered hepatic lipid oxidation by MRI in an animal model of MASLD

Abstract

Metabolic dysfunction-associated steatotic liver disease (MASLD) prevalence is increasing annually and affects over a third of US adults. MASLD can progress to metabolic dysfunction-associated steatohepatitis (MASH), characterized by severe hepatocyte injury, inflammation, and eventual advanced fibrosis or cirrhosis. MASH is predicted to become the primary cause of liver transplant by 2030. Although the etiology of MASLD/MASH is incompletely understood, dysregulated fatty acid oxidation is implicated in disease pathogenesis. Here, we develop a method for estimating hepatic β-oxidation from the metabolism of [D₁₅]octanoate to deuterated water and detection with deuterium magnetic resonance methods. Perfused livers from a mouse model of MASLD reveal dysregulated hepatic β-oxidation, findings that corroborate in vivo imaging. The high-fat-diet-induced MASLD mouse studies indicate that decreased β-oxidative efficiency in the fatty liver could serve as an indicator of MASLD progression. Furthermore, our method provides a clinically translatable imaging approach for determining hepatic β-oxidation efficiency.

Keywords: MASLD; X-nucleus imaging; deuterium MRI; metabolic flux; octanoate; steatosis; β-oxidation.

Graphical abstract

Figure 1

Changes in mouse body weight, liver weight, and total liver protein over 16 weeks of HFD or LFD (A) Weekly mouse mass of 10-week-old C57BL/6J mice after switching to a high-fat diet (HFD; 60% fat by Kcal) or a low-fat diet (LFD; 10% fat by Kcal; n = 10 for each group). Data are presented as average ± 95% confidence interval (CI; shaded regions). (B and C) Livers of mice on the indicated diets for 16 weeks were collected after ex vivo perfusion. Total liver mass was measured (n = 8 for LFD and HFD) (B) and total liver protein was quantified by Bradford assay (n = 5 for LFD and HFD) (C). Error bars represent the mean ± standard deviation. Statistical significance was determined by analysis of covariance for mouse weight gain and Student’s t test for liver mass and total liver protein (∗p < 0.05).

Figure 2

Measures of oxidative metabolism in an ex-vivo-perfused liver system (A) Diagram of in-magnet perfused mouse liver setup. (B) Representative trace of deuterium magnetic resonance spectroscopy of perfused mouse liver before (red) and 30 min after (blue) perfusion with [D₁₅]octanoate. (C–F) Quantification of HDO and [D₁₅]octanoate peaks presented as average ± 95% CI of individual liver signal in shaded regions (n = 5 for each group). (C) HDO (mmol) produced throughout liver perfusion from [D₁₅]octanoate. (D) mM [D₁₅]octanoate in liver during perfusion. (E) Total HDO per gram of liver protein produced throughout liver perfusion from [D₁₅]octanoate. (F) Ratio of HDO signal intensity to [D₁₅]octanoate signal intensity. (G and H) Total (G) and normalized (H) rates of the indicated metabolic processes were determined by GC-MS and quantified from changes in the concentrations of [D₁₅]octanoate, glucose, and ketones (n = 5 for each group). Statistical significance was determined by analysis of covariance for NMR quantification and Student’s t test (two-tailed) for GC-MS rate comparisons (∗p < 0.05).

Figure 3

Metabolic modeling of flux in an ex-vivo-perfused liver (A) Diagram of INCA metabolic model used to define the rates of [D₁₅]octanoate consumption, electron transport chain activity, and central carbon metabolism. (B) Oxygen consumption was monitored by Hansatech oxygraph+ of afferent and efferent perfusate in [D₁₅]octanoate-perfused livers from mice in LFD and HFD groups. Rates of total oxygen consumption and O₂ consumption normalized to gram of liver protein are shown. (C) Fractional enrichments of 3-hydroxybutyrate to represent ketones, citrate as an indicator of TCA cycle enrichment and acetyl-CoA partitioned into the TCA cycle, glutamate as an exchangeable portion of the TCA cycle with α-ketoglutarate, and succinate to represent the second half of the TCA cycle exchanging with fumarate. (D and E) Predicted flux for the mouse perfused liver by INCA metabolic modeling of (D) total flux per liver and (E) relative flux per gram of liver protein. Error bars represent mean ± standard deviation (n = 5 for each group). Statistical significance was determined by Student’s t test (two-tailed) for rate comparisons (∗p < 0.05).

Figure 4

Metabolic characterization of ad-libitum-fed 12-week-old C57BL/6J mouse torso through magnetic resonance imaging and spectroscopy (A–C) Images are represented using a ¹H volume coil (column 1) and a ²H saddle coil (column 2) with ¹H (grayscale) and ²H (red) merges of the two (column 3). The liver is circled in yellow, and the 0.05% D₂O standard is circled in magenta. Axial FLASH images through the mouse vertical cross-section are taken with respect to the liver in ¹H and ²H modes. The images in (A) represent the signal prior to (pre-injection) and 24 min post- (post-injection) [D₁₅]octanoate administration. Overlays of the ¹H and ²H axial images demonstrate localization of the signal to the liver after administration of [D₁₅]octanoate tracer. Representative 2PD images obtained 52 min after [D₁₅]octanoate injection are shown in (B). In (C), for the coronal images taken 80 min post-injection, the heart and shoulder are circled in teal, and the liver is circled in yellow. (D and E) Spectral analysis of fed mouse torso post-tail vein injection of [D₁₅]octanoate. Data shown represent a single representative fed mouse. In (D), the spectral overlay of whole-volume, deuterium, single-pulse acquisition pre-injection and 3, 15, and 90 min post-injection are colored with blue, red, green, and magenta, respectively. (E) shows a line plot with 95% CI of the continuous spectral timeline of deuterium signal for HDO (green) and [D₁₅]octanoate (pink) for 3 separate mice.

Figure 5

Representative metabolic imaging of C57BL/6J mouse liver for the dietary timeline from 8 to 44 weeks of age (A and B) All images represent an overlay of the respective deuterium image with the proton axial for the hepatic region. Overlays are ordered from top to bottom per image as follows: ²H-FLASH before injection of [D₁₅]octanoate, 27 min 2PD acquisition starting at 17 min post-injection (HDO and [D₁₅]octanoate components), and 27 min ²H-FLASH acquisition starting at 62 min post-injection. For each given time point, the 18-h-fasted mice are in the left column, and the ad-libitum-fed mice are in the right column. Images were arranged with respect to timeline in the order of (A) baseline and 17, 24, and 36 weeks on chow diet and (B) 8.5, 17, 24, and 36 weeks on HFD. (C) Percentage of steatosis score in 18-h-fasted (left) and ad-libitum-fed (right) C57BL/6J mice on either HFD (pink bar, blue dots) or chow diet (green bar, orange dots). The 5% steatosis cutoff for MASLD/NAFLD is represented as a red dashed bar. Error bars represent the mean ± standard deviation. B refers to significance relative to baseline. C refers to significance when comparing the HFD mice to their chow diet counterparts for the same age. F refers to significance when comparing fed and fasted animals within the same diet and age. The numbers of fasted animals used were n = 5, 4, 4, 4, 4, 4, 3, and 4 at baseline, 8.5 weeks HFD, 17 weeks HFD, 17 weeks chow, 24 weeks HFD, 24 weeks chow, 36 weeks HFD, and 36 weeks chow, respectively. The numbers of fed animals used were n = 4, 3, 4, 4, 5, 4, 3, and 5 at baseline, 8.5 weeks HFD, 17 weeks HFD, 17 weeks chow, 24 weeks HFD, 24 weeks chow, 36 weeks HFD, and 36 weeks chow, respectively.

Figure 6

Quantitative measurements of liver mass as well as total hepatic β-oxidation and hepatic β-oxidation relative to liver mass HFD-fed mice are in pink, while baseline and chow-fed mice are in dark green. (A and B) Liver mass in grams from ad libitum-fed mice (n = 4, 3, 4, 4, 5, 5, 3, and 5 at baseline, 8.5 weeks HFD, 17 weeks HFD, 17 weeks chow, 24 weeks HFD, 24 weeks chow, 36 weeks HFD, and 36 weeks chow, respectively) (A) or 18-h-fasted mice that underwent deuterium MRI at the start of the diet (8 weeks of age) until the end of imaging (44 weeks of age) (n = 5, 4, 4, 4, 4, 4, 3, and 4 at baseline, 8.5 weeks HFD, 17 weeks HFD, 17 weeks chow, 24 weeks HFD, 24 weeks chow, 36 weeks HFD, and 36 weeks chow, respectively) (B). Weeks on diet was determined by access to the MRI system. (C–F) MRI quantification of 2PD-derived HDO produced per mg of [D₁₅]octanoate injected (top) and HDO produced per mg of [D₁₅]octanoate and g of liver mass (bottom). See Table S7 for statistical evaluation. The quantification of liver images is ordered as follows: ad-libitum-fed mouse groups on (C) chow diet (n = 4, 4, 5, and 5 at 0, 17, 24, and 36 weeks of diet, respectively) or (D) HFD (n = 4, 3, 4, 5, and 3 at 0, 8.5, 17, 24, and 36 weeks of diet, respectively), followed by 18-h-fasted mouse groups on (E) chow diet (n = 5, 4, 4, and 4 at 0, 17, 24, and 36 weeks of diet, respectively) or (F) HFD (n = 5, 4, 4, 4, and 3 at 0, 8.5, 17, 24, and 36 weeks of diet, respectively). Error bars represent the mean ± standard deviation. B refers to significance relative to baseline. C refers to significance when comparing the HFD mice to their chow diet counterparts for the same age. F refers to significance when comparing fed and fasted animals within the same diet and age.

Figure 7

Correlation of HDO measures to histological score (MAS score) and liver inflammatory marker ALT during HFD HFD fed mice are in pink, while baseline and chow fed mice are in dark green. (A–D) Boxplot of (A) total β-oxidative capacity or (B) β-oxidative efficiency normalized to MAS score during diet timeline in ad-libitum-fed mice (n = 4, 2, 4, 4, 5, 5, 2, and 2 at baseline, 8.5 weeks on HFD, 17 weeks on HFD or chow, 24 weeks on HFD or chow, and 36 weeks on HFD or chow, respectively) or 18-h-fasted mice (C and D) (n = 4, 3, 4, 4, 3, 4, 2, and 2 at baseline, 8.5 weeks on HFD, 17 weeks on HFD or chow, 24 weeks on HFD or chow, and 36 weeks on HFD or chow, respectively). (E–H) Correlation plot with shaded area represented 95% CI of (E) total β-oxidative capacity or (F) β-oxidative efficiency to MAS score during diet timeline in ad-libitum-fed mice or 18-h-fasted mice (G and H). (I–L) Boxplot of total (I) or relative (J) β-oxidative capacity normalized to ALT score during diet timeline in ad-libitum-fed mice (n = 4, 3, 4, 3, 5, and 5 at baseline, 8.5 weeks on HFD, 17 weeks on HFD or chow, and 24 weeks on HFD or chow, respectively) or 18-h-fasted mice (n = 5, 4, 3, 2, 3, and 4 at baseline, 8.5 weeks on HFD, 17 weeks on HFD or chow, and 24 weeks on HFD or chow, respectively) (K and L). (M–P) Correlation plot with shaded area represents 95% CI of (M) total β-oxidative capacity or (N) β-oxidative efficiency to MAS score during diet timeline in ad-libitum-fed mice or 18-h-fasted mice (O and P). See Table S7 for statistical evaluation. Error bars represent the mean ± standard deviation. B refers to significance relative to baseline. C refers to significance when comparing the HFD mice to their chow diet counterparts for the same age.