Doxorubicin (DOX) is a widely used chemotherapeutic agent for both solid and haematological malignancies. DOX’s cardiotoxic side-effects are well known (Moslehi, 2016) and limit the use of DOX despite its excellent anti-cancer action. The mechanism(s) for this toxicity are still poorly understood, although mitochondrial dysfunction has been implicated as a key contributor to cardiac functional decline (Tokarska-Schlattner et al., 2006). Hyperpolarized magnetic resonance imaging (MRI) and MR spectroscopy (MRS) can assess cardiac metabolic fluxes in vivo both in pre-clinical models and in patients (Timm et al., 2018). We have previously shown in a rat model of DOX-cardiotoxicity, that high dose DOX (3 mg kg^–1 for 5 weeks) leads to cardiac systolic dysfunction, which is preceded by a decrease in cardiac carbohydrate metabolism, assessed by hyperpolarized [1-¹³C]pyruvate MRS. However, using a low dose of DOX (2 mg kg^–1 for 6 weeks) we showed that there are no changes in cardiac carbohydrate metabolism and only mild cardiac systolic impairment, which were compensated by increased heart rate and thus led to normal cardiac index, a body-weighed adjusted measure of cardiac output (Timm et al., 2020).

However, cardiotoxicity is not the only side-effect of DOX, and hepatotoxicity associated with DOX-treatment in cancer-patients has also been known for decades (Aviles et al., 1984). DOX is metabolised in the liver by cytochrome P450 and carbonyl reductases (Kassner et al., 2008; Bains et al., 2010). The molecular mechanisms of DOX-induced hepatotoxicity are thought to be due to oxidative stress through reactive oxygen species (ROS) production, which occurs during DOX metabolism in the liver and which leads to inflammation and mitochondrial dysfunction (Prasanna et al., 2020). This eventually causes hepatocyte death and leakage of hepatic enzymes into the circulation. Therefore, DOX-induced hepatotoxicity has previously been assessed by liver serum biomarkers such as serum glutamic oxaloacetic transaminase (GOT) and serum glutamic pyruvic transaminase (GPT) in patients (Damodar et al., 2014) or alanine aminotransferase (ALT) in mice (Nagai et al., 2015). The aspartate aminotransferase (AST):ALT ratio is used in patients as part of the liver function test, and this should be <1 in healthy individuals. Nevertheless, these serum biomarkers indicate cellular damage, and thus a late stage in hepatotoxicity. Since mitochondrial dysfunction is thought to precede those changes, measurements of hepatocyte metabolism may offer an earlier indication of hepatotoxicity or give more assurance that a given dose of DOX is not causing mild liver toxicity.

Doxorubicin has been shown to increase adipose tissue lipolysis and decrease hepatic fatty acid metabolism, thereby inducing hepatosteatitis (Renu et al., 2019), which may lead to oxidative stress and cellular damage. Hyperpolarized magnetic resonance imaging (MRI) and spectroscopy (MRS) can be used to assess hepatic carbohydrate metabolism (Merritt et al., 2011), however, there are currently no markers of long chain fatty acid metabolism available for hyperpolarized ¹³C MRS. We therefore set out to assess if the serum biomarker ALT and hyperpolarized MRS of the in vivo liver can shed light on late and early hepatotoxic effects of DOX, respectively, in our low-dose rat model. We furthermore performed metabolomic analysis of liver tissue to assess both aqueous metabolites and acyl-carnitines, to complement carbohydrate metabolism MRS data with measurements of fatty acyl-CoAs to assess also lipid metabolism. This allowed us to measure not only in vivo real-time carbohydrate metabolism in the rat liver, but also generate a comprehensive list of metabolite changes, including fatty acyl-carnitines, thereby assessing early changes in fatty acid metabolism.

The main limitation of DOX is its severe cardiotoxicity. We have previously shown in a clinically-relevant rat model of DOX-cardiotoxicity, that hyperpolarized [1-¹³C]pyruvate MRS serves as an early detection tool for this toxicity (Timm et al., 2020), and that a low-dose of DOX does not lead to changes in cardiac carbohydrate metabolism. However, cardiotoxicity is not the only side-effect of DOX, with hepatotoxicity also being a major concern (Damodar et al., 2014). We therefore set out to assess liver metabolism in a low-dose DOX model to establish whether we would be able to observe any changes in hepatic metabolism at a dose that shows no cardiac dysfunction in rats. We found that at 2 mg kg^–1 for 6 weeks, there was no significant decrease in cardiac function and that there was no indication of overt liver damage, assessed by plasma ALT activity. We could not detect an effect of low dose DOX on hepatic carbohydrate metabolism, using hyperpolarized [1-¹³C]pyruvate MRS, with the exception of an increased alanine:pyruvate ratio at week 1. However, power calculations determined that 15 rats would be required in each group to determine a difference in alanine:pyruvate ratio at 6 weeks with a power of 80% and α-error of 0.05. A larger cohort of animals may thus allow for small differences in hepatic alanine labelling from hyperpolarized [1-¹³C]pyruvate to be detected. Nevertheless, as MRS data show great variability, this unlikely to be a viable tool to assess early toxic effects of doxorubicin on the liver in patients.

Our metabolomic data revealed an increase in 6-phosphogluconate (6PG), an intermediate in the Pentose Phosphate Pathway (PPP), and an increase in oxidised glutathione (GSSG). Glutathione is a major intracellular antioxidant (Ballatori et al., 2009), and PPP-derived NADPH keeps the glutathione pool in its reduced (GSH) form (Couto et al., 2016). DOX-induced hepatotoxicity is thought to involve ROS production (Prasanna et al., 2020), and GSH can ‘scavenge’ those ROS by reducing them with glutathione peroxidase, while being itself oxidised to the GSSG disulfide (Couto et al., 2016). An increase in 6PG may thus represent an attempt to counteract oxidation of GSH to GSSG by ROS, and thus reflect early signs of toxic damage on the liver. In fact, supplementation of cystathionine as a glutathione precursor has previously been shown to alleviate DOX-induced hepatotoxicity (Kwiecień et al., 2006).

We furthermore found an increase in hepatic creatine and a decrease in methionine. Creatine is synthesised in the kidney and liver, but the last step, methylation of guanidinoacetate acid, which requires methionine, is mostly performed in the liver (Da Silva et al., 2009). An increase in creatine and a decrease in methionine may therefore reflect increased hepatic creatine synthesis. Livers from DOX-treated rats also had increased ornithine concentrations, which is part of the hepatic urea cycle (nitrogen excretion). Rats on DOX-treatment lose weight over time (Timm et al., 2020), and an increase in hepatic urea cycle may reflect enhanced protein breakdown and nitrogen excretion.

Surprisingly, we found an increase in high energy phosphates in livers from DOX-treated rats (ADP, ATP, GDP, GTP, UDP), as well as an increase in citrate. This may reflect an increase in TCA-cycle activity. Since we saw no changes in PDH flux in these rats using hyperpolarized MRS, suggesting no change in carbohydrate oxidation, which may reflect an increased proportion of fatty acid oxidation contributing to ATP synthesis. Interestingly, we also found that most acyl-carnitine species were increased in the liver. Fatty acids enter the mitochondria for oxidation in the form of acyl-carnitine esters (Longo et al., 2016), and an increase in these acyl-carnitines may explain higher liver energetics and citrate levels. However, elevation of plasma acyl-carnitines (especially long-chain) are also associated with non-alcoholic fatty liver disease (NAFLD) in patients (Enooku et al., 2019), and an increase of long-chain acyl-carnitines in liver tissue is associated with non-alcoholic steatohepatitis (NASH; Pérez-Carreras et al., 2003). An increase in acyl-carnitines in livers of DOX treated rats may therefore also reflect early signs of steatohepatitis. We furthermore found increased concentrations of aceto-acetyl-CoA (AcAcCoA), a metabolite in ketone body synthesis. Ketone bodies are made in the liver when there is excess oxidation of fatty acids, and an increase in AcAcCoA may thus support the notion that increased levels of acyl-carnitines are maladaptive in DOX-treated rats, signalling early signs of fatty liver disease.

Overall, we have shown that a low dose of DOX that shows no cardiac side effects in rats appears to also not cause overt liver cellular damage or changes in carbohydrate metabolism that could be observed with hyperpolarized [1-¹³C]pyruvate MRS. However, we found that low-dose of DOX leads to an increase in antioxidant-defence mechanisms in the liver, reflecting potential early oxidative damage. We also found an increase in most hepatic acyl-carnitine species with DOX-treatment, which may reflect increased adipose lipolysis and hepatic fatty acid uptake. Since we also saw a high energy state and increased citrate concentrations in those livers, this suggests that fatty acid oxidation may be enhanced in the liver upon DOX treatment. This may be compensatory, but since DOX treatment can lead to steatohepatitis, increased hepatic acyl-carnitine species may serve as an early sign of excess lipid accumulation in the liver.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

The animal study was reviewed and approved by the Animal Care and Ethical Review Committee, University of Oxford. All animal experiments conformed to Home Office Guidance on the Operation of the Animals (Scientific Procedures) Act, 1986.

KT conceptualized and designed the study, performed all experiments, analysed the data, and wrote the manuscript. VB helped with all MR imaging experiments. JM helped with some MRI experiments and edited the manuscript. DS helped with some MRI and echocardiography experiments and edited the manuscript. JW helped perform and analyse metabolomics data. JG helped design metabolomics experiments and data analysis. DT contributed to conception and design of the study and edited the manuscript. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.