Isotope tracing reveals glycolysis and oxidative metabolism in childhood tumors of multiple histologies

Abstract

Background: Survival among children with high-risk solid tumors remains poor. Reprogrammed metabolism promotes tumor growth and may contain therapeutic liabilities. Tumor metabolism has been assessed in adults using intra-operative 13C-glucose infusions. Pediatric tumors differ from adult cancers in their low mutational burden and derivation from embryonic tissues. Here we used 13C infusions to examine tumor metabolism in children, comparing phenotypes among tumor types and between childhood and adult cancers.

Methods: Patients recruited to study NCT03686566 received an intra-operative infusion of [U-13C]glucose during tumor resection to evaluate central carbon pathways in the tumor, with concurrent metabolomics to provide a broad overview of metabolism. Differential characteristics were determined using multiple comparison tests and mixed effect analyses.

Findings: We studied 23 tumors from 22 patients. All tumors analyzed by [U-13C]glucose contained labeling in glycolytic and tricarboxylic acid (TCA) cycle intermediates. Labeling in the TCA cycle indicated activity of pyruvate dehydrogenase (PDH) and pyruvate carboxylase (PC), with PDH predominating. Neuroblastomas had high lactate labeling relative to other childhood cancers and lung cancer, and were distinguished by abundant tyrosine catabolites consistent with catecholamine synthesis.

Conclusions: Intra-operative [U13C]glucose infusions are safe and informative in pediatric cancer. Tumors of various histologies use glycolysis and oxidative metabolism, with subtype-selective differences evident from this small cohort. Expanding this cohort may uncover predictive biomarkers and therapeutic targets from tumor metabolism.

Funding: N.C.I grants to P.L. (R21CA220090-01A1) and R.J.D. (R35CA22044901); H.H.M.I. funding to R.J.D.; Children's Clinical Research Advisory Committee funding to K.J.

Keywords: Cancer; glucose; isotopes; metabolism; metabolomics; neuroblastoma; sarcoma.

Figures

Figure 1:. Method for intra-operative [U-13C]glucose infusions in children.

(A) Two weight-based regimens for infusion of [U-13C]glucose. (B) Plasma glucose enrichment levels for pediatric patients on Regimen 1 or Regimen 2.

Figure 2:. 13C labeling in tumor metabolites from two children with neuroblastoma.

(A) Expected labeling of glycolytic and TCA cycle metabolites after infusion with [U-13C]glucose. In this scheme, the two labeled carbons from acetyl-CoA are transferred to citrate positions 1 and 2 and then to positions 4 and 5 of α-KG and carried further around the TCA cycle. Labeled OAA exchanges with labeled aspartate. Further metabolism of labeled OAA into subsequent turns of cycle is not illustrated. (B) Computed tomography of Patient #1, a 9-year old boy with a non-MYCN-amplified neuroblastoma (red box). (C) Time-dependent labeling of glucose, lactate and pyruvate in Patient #1’s plasma during the [U-13C]glucose infusion. (D) Labeling in metabolites extracted from Patient #1’s tumor. Labeling is normalized to the glucose m+6 fraction in plasma at the time of tumor resection. (E) Computed tomography of Patient #3, a 2-year old girl with a non-MYCN-amplified neuroblastoma (red box). (F) Time-dependent labeling of glucose, lactate and pyruvate in Patient #3’s plasma during the [U-13C]glucose infusion. (G) Labeling in metabolites extracted from Patient #3’s tumor. Labeling is normalized to the glucose m+6 fraction in plasma at the time of tumor resection.

Figure 3:. Metabolite labeling reveals evidence of metabolic differences between neuroblastoma and other tumor types.

(A) Fractional enrichment of metabolites extracted from all pediatric tumors in the cohort. Labeling is normalized to the glucose m+6 fraction in plasma at the time of tumor resection. The bar indicates that in neuroblastomas, relative labeling in lactate is higher than relative labeling in 3PG. (B) Fractional enrichment of metabolites extracted from all pediatric tumors in the cohort. These data are the same as in panel (A), but sarcomas are considered as a separate group. The bar indicates the difference in relative lactate labeling between neuroblastoma and other non-sarcoma tumors. (C) Fractional enrichment in tumor lactate, normalized to plasma glucose m+6, in pediatric tumors, adult NSCLC and adult ccRCC. NSCLC and ccRCC data include previously-reported patients , and are displayed here for the purposes of comparing to the new data from pediatric cancers. (D) Lactate m+3/3PG m+3 labeling ratios in the tumors displayed in (C). (E) Summary of labeling surrogates for PC activity (citrate m+3/pyruvate m+3 ratio) and PDH activity (citrate m+2/pyruvate m+3 ratio) for all pediatric tumors subjected to [U-13C]glucose infusion. (F) Comparison of PC and PDH labeling surrogates in different pediatric tumor types. Abbreviations: Low grade fibromyxoid sarcoma (LGF), osteosarcoma (OS), clear cell sarcoma of the kidney (CCS), rhabdomyosarcoma (RMS), synovial sarcoma (SS).

Figure 4:. Metabolomic features differentiate neuroblastomas from other pediatric tumors.

(A) Partial least squares discriminant analysis (PLS-DA) on metabolomics data from neuroblastomas and other pediatric tumors. (B) Volcano plot demonstrating metabolomic differences between neuroblastomas and other pediatric tumors. Several catecholamine-related metabolites accumulated in the neuroblastomas are indicated in red. (C) Metabolism of tyrosine to fumarate and acetoacetate or into catecholamine synthesis. (D) Volcano plot demonstrating gene expression differences between neuroblastomas and other tumors. Tumors used in this analysis include those shown in (E). Genes encoding several enzymes of catecholamine synthesis are indicated in red. (E) Expression of genes involved in tyrosine catabolism and catecholamine synthesis in neuroblastomas and several other forms of cancer. ****p