Metabolic signatures of thymomas: potential biomarkers and treatment targets

Abstract

Open in new tabDownload slide

OBJECTIVES

A study of tumour metabolic reprogramming has revealed disease biomarkers and avenues for therapeutic intervention. Metabolic reprogramming in thymoma is currently understudied and largely unknown. This study utilized metabolomics and isotope tracing with ¹³C-glucose to metabolically investigate thymomas, adjacent thymic tissue and benign thymic lesions.

METHODS

From 2017 to 2021, 20 patients with a suspected thymoma were recruited to this prospective Institutional Review Board approved clinical trial. At the time of surgery, 11 patients were infused with ¹³C-glucose, a stable, non-radioactive tracer which reports the flow of carbon through metabolic pathways. Samples were analysed by mass spectrometry to measure the abundance of >200 metabolites.¹³C enrichment was measured in patients who received ¹³C-glucose infusions.

RESULTS

Histological analysis showed that 9 patients had thymomas of diverse subtypes and 11 patients had benign cysts. In our metabolomic analysis, thymomas could be distinguished from both adjacent thymus tissue and benign lesions by metabolite abundances. Metabolites in pyrimidine biosynthesis and glycerophospholipid metabolism were differentially expressed across these tissues.¹³C-glucose infusions revealed differential labelling patterns in thymoma compared to benign cysts and normal thymus tissue. The lactate/3PG labelling ratio, a metabolic marker in aggressive lung tumours correlated with lactate uptake, was increased in thymomas (1.579) compared to normal thymus (0.945) and benign masses (0.807) (thymic tissue versus tumour P = 0.021, tumour versus benign P = 0.013).

CONCLUSIONS

We report metabolic biomarkers, including differential ¹³C labelling of metabolites from central metabolism, that distinguish thymomas from benign tissues. Altered glucose and lactate metabolism warrant further investigation and may provide novel therapeutic targets for thymoma.

INTRODUCTION

Across the different World Health Organization thymic epithelial tumour subtypes, particularly thymomas, it is believed that they represent distinct biological entities with unique molecular expressions, rather than a continuum of the same biology [1]. These tumours are often diagnosed at a later stage in the disease progression unless they present with a paraneoplastic process [2]. Approximately 30–50% of thymomas are considered inoperable at the time of diagnosis [3]. The 5-year survival rates for patients with locally advanced and metastatic disease are 35–50% and 20–30%, respectively [4–6].

Improved diagnostic methods that are noninvasive, a timely assessment of treatment response and more effective treatment options that match or surpass the outcomes of complete resection are needed. Currently, computed tomography and positron emission tomography have limitations in providing disease-specific information from biopsy or surgical resection. The significance of studying thymomas is underscored by the considerable clinical morbidity and mortality associated with these tumours. These factors emphasize the need to identify potential biomarkers for disease detection and explore novel therapeutic options to enhance survival in patients with unresectable disease.

Metabolic reprogramming is a clinically relevant hallmark of cancer [7]. Cancer cells support malignant growth by altering nutrient acquisition and rewiring metabolic pathways, supporting energy generation, biosynthetic activities and redox homeostasis. Importantly, these metabolic alterations can be leveraged for clinical benefit. Increased glucose uptake in most cancer cells is the basis for imaging strategies, such as ¹⁸fluorodeoxyglucose-positron emission tomography (FDG-PET). Reprogrammed metabolism also imposes liabilities that can act as the basis for several chemotherapy strategies, such as anti-folates [8] and isocitrate dehydrogenase inhibitors [9] and others currently under investigation [10]. Identifying unique metabolic features of tumours may continue to yield new diagnostic and therapeutic strategies.

One method to investigate tumour metabolism is to compare metabolite abundance between malignant and benign tissue. To date, metabolic investigations into thymoma have been limited to measuring highly abundant metabolites by NMR [11]. Alternative technologies such as mass spectrometry are increasingly used to analyse clinical tumour samples as they offer high sensitivity and a broad coverage of metabolites. Termed ‘metabolomics’, these methods can provide detailed insight into the metabolic composition of biological specimens and can identify unique markers of disease states. While research on genetic and epigenetic alterations in thymomas revealed distinct genetic signatures from normal thymic epithelial cells, few studies have broadly examined metabolic reprogramming in thymomas [12–14].

Stable isotope tracers provide a complementary method to study metabolic activity in cancer patients. While metabolomics affords a thorough measurement of metabolite abundances, it does not report the activity of metabolic pathways. To answer this question, nutrients labelled with stable isotope tracers can be infused to document pathway activity. These nutrients are then metabolized by tissues and can provide unique insight into the fuel choice and fate of specific metabolites. These tracers are safe, non-radioactive isotopes that allow both the tumour and surrounding tissue to take up the labelled nutrient and distribute the label to downstream metabolites. These methods have provided unique insight to metabolic activities of tumours in patients including novel discoveries of metabolic heterogeneity and alternative fuel use in tumours, information that may provide prognostic information and help to identify treatment opportunities [15, 16].

We designed a prospective trial to analyse the metabolic profiles of human thymoma. In this study, we recruited patients who were referred for surgical removal of suspected thymomas. We performed polar and lipid metabolomics analysis, generating a broad coverage of the tissue metabolome. In a subset of patients, we infused [U-¹³C]glucose (i.e. glucose containing ¹³C in all 6 carbons) to analyse potential differences in glycolytic and tricarboxylic acid (TCA) cycle pathways in thymomas. In this work, we found that thymomas have a distinct metabolic profile compared to benign tissues.

PATIENTS AND METHODS

Ethical statement

The Institutional Review Board of the University of Texas Southwestern Medical Center in Dallas approved the study (STU 052012-065 and STU 042017-057; 2012 and 2017). Patients’ written informed consent was obtained. The ClinicalTrials.gov trial registry number is NCT02095808.

Study design and population

From 2017 through 2021, 20 patients with a suspected thymoma were recruited to this institutional review board-approved study. Patients were considered eligible for the study if they had surgically resectable thymic abnormalities or masses. Patients underwent total thymectomy and samples of suspicious malignant and benign tissue after complete resection were immediately snap-frozen in liquid nitrogen for analysis from the specimen after removal.

¹³C glucose infusions

A subset of patients (n = 11) underwent infusion with [U-¹³C]glucose and metabolite tracing analysis, as described previously [16]. On the day of surgery, sterile, pyrogen-free [U-¹³C]glucose (Cambridge Isotope) was administered as a bolus of 8 g over 10 min followed by 4 or 8 g/h infusion through a peripheral intravenous line, generally lasting between 2 and 4 h. A second peripheral intravenous line in the contralateral arm was used to obtain blood to monitor glucose levels and analyse ¹³C enrichment. Resected tissue fragments were rinsed in ice-cold saline and immediately frozen in liquid nitrogen.

Liquid chromatography–mass spectrometry (polar)

Metabolites were separated by hydrophilic interaction liquid chromatography (HILIC) on a Vanquish UHPLC using a Millipore ZIC-pHILIC column and a binary solvent system of 10 mM ammonium acetate in water, pH 9.8 (solvent A) and acetonitrile (solvent B) at a constant flow rate of 0.25 ml/min. Metabolites were measured with a Thermo Scientific QExactive HF-X hybrid quadrupole orbitrap high-resolution mass spectrometer. Individual samples were acquired with a high-resolution mass spectrometer full scan (precursor ion only) method switching between positive and negative polarities. Pooled samples were generated from an equal mixture of all individual samples and analysed using individual positive- and negative-polarity data-dependent tandem mass spectrometry acquisition methods for high-confidence metabolite ID. Metabolites were relatively quantitated by integrating the chromatographic peak area of the precursor ion searched within a 5 ppm tolerance.

Liquid chromatography–mass spectrometry (non-polar)

Lipids were extracted from tissues using a modified Folch extraction (a 1:1:2 mixture of phosphate-buffered saline:methanol:chloroform). The chloroform layer was isolated and dried under a nitrogen stream. Samples were reconstituted in 1 ml of a 65:35 mix of isopropanol:methanol and submitted for liquid chromatography with tandem mass spectrometry (LC–MS/MS) analysis. Separation of lipids was achieved on a Thermo Vanquish UHPLC system equipped with a Thermo Scientific Accucore C18 column using a binary gradient of 50:50 water:acetonitrile with 5 mM ammonium acetate and 0.1% formic acid (solvent A) and 88:10:2 isopropyl alcohol:acetonitrile:water with 5 mM ammonium acetate and 0.1% formic acid (solvent B) at a constant flow rate of 0.17 ml/min. The column was held at a constant temperature of 60°C throughout the acquisition. Mass spectrometry analysis was performed on a Thermo Scientific Fusion Lumos Tribrid Orbitrap mass spectrometer using a modified data-dependent acquisition method similar to one reported previously [17].

Gas chromatography–mass spectrometry

Blood was obtained prior to infusion and approximately every 45 min during infusion until tissue was removed from the patient. Whole blood was chilled on ice and centrifuged to separate plasma. Aliquots of plasma (25–50µl) were added to 80:20 methanol:water for extraction. Frozen tissue fragments weighing 5–15 mg were added to 80:20 methanol:water and mechanically homogenized. Samples were subjected to a minimum of 3 freeze-thaw cycles and then centrifuged at 16 000 × g for 15 min to precipitate macromolecules. The supernatants were evaporated, then re-suspended in 40 μl anhydrous pyridine with methoxyamine (10 mg/ml) in gas chromatography-mass spectrometry (GC/MS) autoinjector vials and incubated for 10 min at 70°C. Then, 80 μl N-(tert-butyldimethylsilyl)-N-methyltrifluoroacetamide derivatization reagent was added to each vial and incubated at 70°C for 1 h. Aliquots of 1 μl were injected for analysis on either an Agilent 6890 or 7890 gas chromatograph coupled to an Agilent 5973N or 5975C Mass Selective Detector, respectively. Mass isotopologues were corrected for natural abundance.

Quantification and statistical analysis

For metabolomic analysis, patient-matched tumour, non-cancerous tissue samples and benign lesions were analysed by partial least squares discriminant analysis (PLS-DA) and plotted with 95% confidence interval shaded regions. For metabolite tracing analysis, the data were confirmed to be normally distributed by the Shapiro–Wilk test, and statistical significance was assessed by a one-way Analysis of Variance (ANOVA) and Tukey post hoc. Data were presented as average ± standard deviation. All data were considered significant if P < 0.050. Sample sizes were not predetermined based on statistical power calculations.

RESULTS

Patient demographics

This study included 20 patients (70% male) undergoing surgery for a suspected thymoma. 9 patients were determined to have thymomas and 11 patients had a variety of benign cysts and masses. The median age of the patients was 54 years (range, 32–86 years). The median BMI of the patients was 30 (range, 21–41). Of the 20 patients, 7 (35%) had a myasthenia gravis diagnosis and were acetylcholine receptor antibody positive. Three of these 7 myasthenia gravis patients had thymomas and 4 had benign masses. Of the 9 thymomas, there was a variety of different histologic pathologies and clinical stages. The most common World Health Organization (WHO) histologic subtype was type AB (44%), and the most common Masaoka-Koga stage was stage 1 (44%). The mean tumour diameter was 6.2 cm (range, 2–10.5 cm) (Table 1).

Metabolite profiles distinguish thymomas from adjacent tissues and benign lesions

Patients underwent this study as designed (Fig. 1) to determine if we could identify metabolic signatures that distinguish thymomas from benign tissue. All resected tissues, including thymoma, benign masses and adjacent thymus were collected intra-operatively and immediately snap-frozen after resection. Samples were first analysed by mass spectrometry after a polar extraction which resulted in measurement of ∼250 metabolites. PLS-DA was used to assess metabolic differences between thymomas and non-malignant tissues, including adjacent normal thymus and benign masses and cysts (Fig. 2A). The thymoma cluster is distinct from the 2 control groups which indicates that this tissue class can be categorized by differential metabolite abundances. In other words, thymomas have consistent metabolic differences that distinguish them from benign tissue. We further investigated which metabolites differentiate thymomas from benign lesions via heatmap (Fig. 2B). Some metabolites were enriched in thymomas (e.g. glycerol 3-phosphate, glutaminyl-glutamic acid, phosphoethanolamine) while others were less abundant than in benign tissues (e.g. creatinine, ornithine, glycerol cyclic phosphate). We found that metabolites in pyrimidine biosynthesis and glycerophospholipid metabolism pathways were consistently different between these tissue types.

We noted that multiple lipid species were differentially expressed between tumour and benign tissues in our initial assessment of the first few patients. To investigate this more specifically, we performed a lipidomics analysis by liquid chromatography–mass spectrometry. Here, non-polar metabolites are extracted and analysed in an analogous workflow. PLS-DA of the lipid metabolome was partially able to differentiate thymomas from benign or adjacent tissues (Fig. 3A). To assess which metabolite species account for this difference, we analysed the samples with a variable important in projection which highlights the metabolites that drive the differences between groups. Multiple lipids are differentially expressed between benign and tumour samples, but we noticed an enrichment of lysophosphatidylcholine (LPC) and phosphatidylcholine species while abundant triglyceride species were decreased (Fig. 3B) in thymomas compared to benign tissues, suggesting that thymomas have broadly altered lipid metabolism compared to healthy, adjacent tissues.

Specific metabolic activities differentiate thymomas from benign tissues

While metabolomics is a powerful tool to broadly measure metabolite abundances, it does not directly provide information about pathway activity or how pathways are supplied by different nutrients. To assess if thymomas can be distinguished from benign tissues based on pathway activity, we performed [U-¹³C]glucose infusions in patients with thymic lesions.

Tissue samples obtained from patients infused with [U-¹³C]glucose were analysed for enrichment of ¹³C in glycolytic and TCA cycle metabolites (Fig. 4A). All tissues displayed a progressive dilution of ¹³C enrichment in glycolytic and TCA cycle metabolites with comparable enrichment across the tissues (Fig. 4B). Ratios of metabolite enrichment can also provide insight into nutrient use. For instance, lactate is generated downstream from glucose but also can be directly taken into the tumour from the bloodstream [16] (Fig. 4C). Comparing the ¹³C enrichment of lactate to 3-phosphoglycerate (3PG), a metabolite usually derived from glucose within the tissue, has been used to assess the degree in which lactate is taken up from the blood. The results of this study show that thymomas have a significantly elevated lactate/3PG ratio (1.579) compared to normal thymus tissue (0.945) and benign masses (0.807) (Fig. 4D, thymic tissue versus tumour, P = 0.021, tumour versus benign, P = 0.013). This result implies that part of the thymoma lactate pool may arise from outside the tumour rather than from local glucose metabolism within the tumour. Of note, the lactate/3PG labelling ratio was no different between benign lesions and non-malignant thymic tissue adjacent to the tumours.

DISCUSSION

This study combined metabolomic analyses with ¹³C-glucose tracing to determine the unique ways that thymoma metabolism is altered compared to benign tissues. We demonstrated that thymomas are metabolically distinct from benign tissue both in metabolite abundance and the way that they utilize glucose-derived carbon. These findings are consistent with the emerging consensus that altered metabolism is a hallmark of neoplastic processes [7, 18–20].

The 2 methods in this study, metabolomics and metabolite tracing, both revealed altered metabolism in thymomas. The metabolomic analysis provided insight into the overall metabolic phenotype of thymomas including a snapshot of the abundances of major metabolites and how those abundances differed from benign tissues. This analysis indicates that pyrimidine metabolism and glycerophospholipid metabolism may be enhanced in thymomas. Our data also show that glycerol 3-phosphate, a metabolite derived from glycolysis, is more abundant in thymomas as compared to benign tissues. This elevated presence of a product of glucose catabolism may indicate enhanced glucose utilization, a phenomenon observed in numerous malignancies [18]. In the context of thymomas, increased glucose uptake is consistent with the ability to detect these malignancies through fluorodeoxyglucose-PET imaging. Another finding of note was the increased abundance of phosphoethanolamine in thymomas compared to benign tissues. Increased phosphoethanolamine has previously been posited as a mechanism for glucocorticoid treatment resistance in mouse models of thymoma [21]. Further studies may reveal more survival advantages associated with the alteration of other metabolic pathways that can reveal a new understanding of thymoma biology.

Lipidomic analysis revealed multiple LPC species elevated in thymomas relative to benign cysts. LPC is typically generated from the phospholipase-mediated cleavage of phosphatidylcholine and induces cell signalling responses by binding to G protein-coupled receptors and Toll-like receptors. In cancer, the relevance of LPC remains unclear. In some cancer types, elevated plasma levels of LPCs (particularly 18:0) correlate with better patient prognosis [22]. In addition to elevated LPCs, our data also show a significant decrease in abundant triglyceride species. Together, these observations suggest that thymomas may undergo substantial plasma membrane remodelling during growth. The action of enzymes such as LPC acyltransferase and the Lands cycle in general protects many cancers and is a potential therapeutic target [23].

The metabolite tracing analysis indicated that thymomas are capable of using multiple nutrients. In patients infused with ¹³C-glucose, several thymic tumours displayed an increased lactate/3PG labelling ratio, a marker we previously correlated with increased lactate uptake [16]. While future studies could determine the functional mechanisms and clinical relevance of this in thymoma, the use of lactate as fuel has been suggested in other neoplasms [24] and supports aggressive tumour features in other cancer models. Studies in pre-clinical models of melanoma demonstrated that inhibition of the lactate transporter monocarboxylate transporter 1 abrogated the metastatic capacity of these tumours [25]. Given the apparent upregulation of lactate uptake in thymomas, the findings of this study suggest that similar investigations into thymoma may be warranted. If thymomas upregulate lactate uptake for use as a fuel, then this pathway may prove to be a novel treatment target for thymomas.

Another study recently characterized thymoma metabolism with NMR spectroscopy [11], comparing the abundance of 37 metabolites between less aggressive thymomas (WHO stages A, AB, and B1) to more aggressive thymomas (WHO stages B2 and B3) and thymic carcinomas. Here, NMR identified increased metabolites, such as lactate, in more aggressive thymomas. While we did not observe similar results in our comparison of thymomas vs benign tissue, our tracing data implicate lactate as potential fuel source in thymoma, and it will be important to evaluate if this feature increases further in aggressive thymomas. Further analysis with a greater sample size may reveal more metabolic biomarkers that distinguish high- and low-stage tumours, potentially leading to diagnostic insight for early detection of aggressive thymomas.

One promising development in tumour detection comes in the form of hyperpolarized ¹³C MRI. This imaging modality uses MRI to detect ¹³C in vivo, with applications in cancer, cardiovascular disease, and inflammatory processes [26]. Metabolite tracing studies such as this one lay the groundwork for establishing how to interpret such imaging in tumours. For instance, if more aggressive thymomas take up lactate as a fuel source, then this may be detectable non-invasively with hyperpolarized ¹³C MRI. Ultimately, such a breakthrough could have significant clinical impact and help delineate operable versus non-operable tumours and response to treatment in non-operable tumours. These potential applications underscore the importance of this research as a catalyst for future breakthroughs.

Limitations

This study should be interpreted in the context of several limitations. The normal thymus tissue was only removed from patients who had a thymoma or other mass which may have influenced the metabolome of the normal thymus tissue. This limitation was inherent to the study design as only patients with a suspected thymoma were enrolled in the study due to ethical considerations. The normal thymus tissue in this study thus approximates but may not fully represent a perfectly normal thymus. However, given the positive study result showing a metabolomic distinction between thymomas and the purported normal thymus tissue, it is likely that this limitation had minimal impact on the validity of the results.

Another limitation of this study is the sample size. Although we only studied 20 patients, each sample was subjected to a broad and detailed metabolic analysis. For every sample, 250 metabolites were examined, creating a robust comparative database of 5500 unique metabolite abundances. This is substantial when compared to similar studies, which used even smaller sample sizes, measured fewer metabolic features, yet still established metabolic differences between malignant and benign tissue [15]. For these reasons, the sample size was sufficient to compare thymomas to benign tissues. Although sufficient tissue was obtained to compare thymomas to benign tissues, there were insufficient samples of each grade and stage of thymoma to compare metabolomes of more and less aggressive tumours or thymomas from patients with myasthenia gravis to those without myasthenia gravis. The sample of thymomas in this study was representative of various thymoma grades and stages which increases confidence in the comparison of thymomas to benign tissues. However, given the increasing awareness that thymomas of different grades are quite different, the sample of thymomas in this study should be considered in that context.

CONCLUSION

This study broadly characterized the ways in which the metabolic phenotype of thymomas is different from benign tissues. The study serves as a foundational endeavour aimed at sparking more in-depth research into metabolic biomarkers and potential therapeutic targets for thymoma. It provides a starting point, shedding light on the metabolic distinctiveness of thymomas, especially their increased lactate uptake, when compared to benign tissues. Future work should be directed at investigating the most likely avenues for practical applications in terms of diagnosis and therapy including upregulation of lactate uptake and identification of clinically relevant metabolic features at different stages of the disease which might be leveraged for improved disease detection.

ACKNOWLEDGEMENTS

Kemp H. Kernstine is the Robert Tucker Hayes Chair in Cardiothoracic Surgery. Ralph J. DeBerardinis is the Joel B. Steinberg, M.D. Distinguished Chair in Pediatrics and Robert L. Moody Sr Faculty Scholar at University of Texas Southwestern and a Howard Hughes Medical Investigator.

FUNDING

This work was supported by the National Cancer Institute (R35-CA-220449) of the National Institutes of Health. Kemp H. Kernstine and Ralph J. DeBerardinis received a Translational Research Award from the V Foundation.

Conflict of interest: There are no conflicts of interest to report pertaining to this study. Ralph J. DeBerardinis is a founder and advisor for Atavistik Bioscience and an advisor for Agios Pharmaceuticals, Vida Ventures and Droia Ventures which were not involved in the study.

DATA AVAILABILITY

The data supporting this study are available on request.

Author contributions

James W. Miller: Data curation; Formal analysis; Investigation; Writing—original draft; Writing—review & editing. Brandon M. Faubert: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Writing—review & editing. Thomas P. Mathews: Investigation; Methodology; Writing—review & editing. John K. Waters: Investigation; Methodology; Writing—review & editing. Ralph J. DeBerardinis: Conceptualization; Data curation; Funding acquisition; Project administration; Resources; Supervision; Writing—review & editing. Kemp H. Kernstine: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Supervision; Writing—review & editing.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Hiroshi Date, Clemens Aigner, Luca Voltolini and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

Presented at the ESTS Annual Meeting, June 6, 2023, Milan, Italy.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• LPC

  Lysophosphatidylcholine

 
• PLS-DA

  Partial least squares discriminant analysis

 
• TCA

  Tricarboxylic acid

 
• WHO

  World Health Organization