Measurement of ¹³C turnover into glutamate and glutamine pools in brain tumor patients

Abstract

Malignant brain tumors are known to utilize acetate as an alternate carbon source in the citric acid cycle for their bioenergetics. ¹³C NMR based isotopomer analysis has been used to measure turnover of ¹³C-acetate carbons into glutamate and glutamine pools in tumors. Plasma from the patients infused with [1,2-¹³C]acetate further revealed the presence of ¹³C isotopomers of glutamine, glucose and lactate in the circulation that were generated due to metabolism of [1,2-¹³C]acetate by peripheral organs. In the tumor cells, [4-¹³C] and [3,4-¹³C] glutamate and glutamine isotopomers were generated from blood-borne ¹³C labeled glucose and lactate which were formed due to [1,2-¹³C[acetate metabolism of peripheral tissues. [4,5-¹³C] and [3,4,5-¹³C] glutamate and glutamine isotopomers were produced from [1,2-¹³C]acetyl-CoA that were derived from direct oxidation of [1,2-¹³C]acetate in the tumor.

Introduction

¹³C NMR spectroscopy is a powerful method to investigate metabolic pathways in malignant human tumors in vivo using intravenous infusion of non-radioactive ¹³C enriched nutrients during the surgical removal of the tumor mass. Previously, it was shown that human glioblastoma (GBM) and metastatic brain tumors were able to oxidize glucose in the citric acid cycle and only a small fraction of glucose carbons was used to derive acetyl-CoA [1,2]. Recently, using ¹³C NMR spectroscopy of surgically resected brain tumor tissues from cancer patients who were infused with [1,2-¹³C]acetate, we have identified that acetate acts as a major bioenergetic substrate in GBM and metastatic brain tumors and contributes up to 48.0% ± 4.0 of the acetyl-CoA pool [3]. Acetate-derived acetyl-CoA enters citric acid cycle via condensation with oxaloacetate (OAA) to form citrate, which further metabolized to form α-ketoglutarate (α-KG) and glutamate [4]. Glutamine is a conditionally essential amino acid that has been shown as a cellular precursor for nucleotide synthesis. Higher glutamine utilization by tumor has been found to support its growth through nucleotide de novo synthesis [5, 6]. In PET imaging, using ¹⁸F-4F-glutamine (¹⁸F-FGln), it was demonstrated that gliomas showed higher uptake of glutamine than that of the normal surrounding brain [7]. Human brain tumor expresses glutamine synthetase (GS) which converts glutamate into glutamine [1–3]. Furthermore, acetate may be metabolized by skeletal muscle, liver and renal tissues to produce key glutamine [5–7]. Although no net synthesis of glucose from acetate carbons occurs in mammalian liver, [1,2-¹³C]acetate influences hepatic gluconeogenesis pathways that leads to ¹³C label incorporation into pyruvate, lactate and glucose in the liver citric acid cycle, which upon entering citric acid cycle in the tumor contributes (via α-ketoglutarate) to glutamate and glutamine pools [8]. Therefore, a comprehensive analysis is required to accurately quantify glutamate and glutamine synthesis in the tumors that includes contributions from circulating ¹³C labeled compounds that are derived from peripheral metabolic activities of [1,2-¹³C]acetate. The purpose of the article was to demonstrate a comprehensive analysis of ¹³C turnover of tumor glutamate and glutamine pools under the presence of blood borne ¹³C-labeled glutamine, ¹³C-lactate and ¹³C-glucose during the infusion of [1,2-¹³C]acetate. This method can be readily applicable to other pathologic and physiologic conditions.

Materials and Methods

Infusion of [1,2-¹³C]acetate in brain tumor patients

Two patients diagnosed with GBM and two patients with metastatic brain tumor (a breast cancer brain metastasis and a non-small-cell lung cancer brain metastasis patient) were enrolled for this study following a clinical protocol approved by University of Texas Southwestern Medical Center Institutional Review Board. As described previously, these patients were infused with sterile and pyrogen free [1,2-¹³C]acetate (Cambridge Isotope Laboratories, Inc. Tewksbury, MA) at 6.0 mg/kg/min for the first 5 min, followed by 3.0 mg/kg/min for 2 hr [3]. Tumor tissue sampling and processing for NMR spectroscopic analysis have been previously described [1–3].

NMR Spectroscopy

¹H and ¹H-decoupled carbon (¹³C) NMR spectra of tumor and plasma extracts were acquired at 600 MHz (¹H frequency)/150 MHz (¹³C frequency) on a Bruker Avance NMR Spectrometer equipped with 10-mm broadband cryogenically-cooled probe (Bruker Biospin, Billerica, MA). The NMR spectra of tumor tissues and plasma used in this work have been published [3]. The lactate C3 carbon resonance at 20.8 ppm was used as an internal chemical shift reference. Peak areas of ¹³C signals of glutamate and glutamine were measured using ACD software (Advanced Chemistry Development, Toronto, Canada). Amounts of various ¹³C-¹³C multiplet (doublet, D45; quartet, Q) and singlet (S) peak areas are reported as a fraction normalized to its total carbon resonance area. ¹³C enrichment of plasma acetate at the end of the infusion period was obtained from ¹H NMR spectra by measuring peak area of ¹³C satellites appearing due to heteronuclear ¹³C-¹H spin-spin coupling (Figure 1A).

Figure 1. Plasma acetate ¹³C enrichment and carbon isotopomers of glutamate and glutamine.

(A) ¹H NMR spectrum of plasma from the breast cancer brain metastasis patient who was infused with [1,2-¹³C]acetate prior to the surgical removal of the tumor, showing methyl (CH₃) protons of acetate, ¹²CH₃ (open circle) refers to the methyl proton signals that are covalently bonded to ¹²C and ¹³CH₃ (filled circle) refers to the methyl protons that are covalently bonded to ¹³C carbon which appear as two sets of doublets separated by ¹J_C-H = 127.10 Hz. The splitting observed in the doublet (5.94 Hz) is due to ²J_C-H (B) illustration of multiple ¹³C isotopomers (a–e) of C4 carbon of glutamate or glutamine used in this analysis. GS, glutamine synthetase.

¹³C-isotopomer analysis

Various ¹³C isotopomers of C4 glutamate or glutamine have been pictorially represented in Figure 1B, and their fractional ¹³C enrichments were determined as described below:

• a = the fraction of glutamine or glutamate that does not have ¹³C in C4 position

• b = the fraction of glutamine or glutamate that has ¹³C only in C4 position

• c = the fraction of glutamine or glutamate that has [3,4-¹³C] labeled isotopomer

• d = the fraction of glutamine or glutamate that has [4,5-¹³C] labeled isotopomer

• e = the fraction of glutamine or glutamate that has [3,4,5-¹³C] labeled isotopomer

By definition, (a + b + c + d + e) = 1, where the variables a–e, are defined as above.

[1,2-¹³C]acetyl-CoA generates the ¹³C-isotopomers, d and e in the citric acid cycle. Isotopomers b and c are derived from [2-¹³C]acetyl-CoA that can be generated from both [1-¹³C]glucose (C1S) and [3-¹³C]lactate (C3S). Multiple mechanisms are responsible for the generation [1-¹³C]glucose and [3-¹³C]lactate that produce [2-¹³C]acetyl-CoA in the tumor. [2-¹³C]acetyl-CoA (Fc2, fraction of acetyl-CoA enriched only in C2 methyl carbon) generates [4-¹³C] glutamate on the first turn and [3,4-¹³C]glutamate on the subsequent turns upon entering citric acid cycle. From the ¹³C NMR spectra of C4 glutamine or glutamate multiplets, the following ¹³C-¹³C spin-spin coupled doublets (D45), quartets (Q) and singlet ¹³C C4 carbon signal (S) areas, are determined. Consequently, the following four equations are derived from the experimental ¹³C NMR peak areas and are given below.

$$\mathit{b}/(\mathit{a}+\mathit{b})=\mathbf{Fc}\mathbf{2}$$ Equation (1)

$$(\mathit{c}/\mathit{b})=\mathbf{D}\mathbf{34}/\mathbf{S}$$ Equation (2)

$$(\mathit{d}/\mathit{b})=\mathbf{D}\mathbf{45}/\mathbf{S}$$ Equation (3)

$$(\mathit{e}/\mathit{b})=\mathbf{Q}/\mathbf{S}$$ Equation (4)

Solutions of the above equations yield fractional contributions of C4 glutamate and glutamine carbon isotopomers (a, b, c, d and e). In the absence of blood-borne ¹³C enriched nutrients due to peripheral metabolism of [1,2-¹³C]acetate, Fc2 should be at the level of natural abundance of ¹³C (1.1%).

Results and Discussion

Figure 1A shows the portion of ¹H NMR spectrum of plasma from a GBM patient who was infused with [1,2-¹³C]acetate prior to the surgical removal of the tumor. In the spectrum, ¹²CH₃ refers to the methyl proton signals of acetate that are covalently bonded to ¹²C carbon (resonating at 1.91 ppm) and ¹³CH₃ refers to the methyl protons that are covalently bonded to ¹³C carbon which appear as two sets of doublets (at ~1.80 and ~2.10 ppm), separated by one-bond with a ¹³C-¹H spin-spin coupling, ¹J_C-H = 127 Hz. The splitting observed within the doublet (5.94 Hz) is due to ²J_C-H. This doublet patterns confirms the presence of [1,2-¹³C]acetate in the plasma. ¹³C-enrichment of plasma acetate was 88.0 ± 6.0% and its concentration at the end of the infusion period was 1.27 ± 0.22 mM. Figure 2 shows schematic diagram illustrating metabolism of infused [1,2-¹³C]acetate in a tumor cell. By the action of nucleocytosolic acetyl-CoA synthetase (ACSS2), [1,2-¹³C]acetate is converted to [1,2-¹³C]acetyl-CoA which enters the citric acid cycle yielding ¹³C-labeling at carbons 4 and 5 of α-ketoglutarate (α-KG), glutamate and glutamine during the first turn of the cycle. After multiple turns of the cycle, carbons 3, 4 and 5 of α-KG, glutamate and glutamine are labeled with ¹³C. Glutamine synthetase (GS), the key enzyme responsible for the conversion of glutamate to glutamine, is highly expressed in human brain tumors, supporting in situ glutamine synthesis from glutamate [1–3,11]. We have observed similar levels of doublets (D45) and quartets (Q) in glutamate and glutamine C4 carbon signals in ¹³C NMR spectra of tumor extracts (Figs. 3A and 3B) which provide evidence for de novo glutamine synthesis from glutamate in the tumor cells. ¹³C multiplet peak areas of C4 glutamate and glutamine in the tumor are given in Fig. 3C. These peak areas were used to determine fractional contributions of various C4 ¹³C isotopomers of glutamate and glutamine as described in the methods and are summarized in the Tables 1A and B. Previously, we observed the presence of ¹³C-labeled glucose and lactate isotopomers in the circulation that were generated due to systemic metabolism of [1,2-¹³C]acetate [8]. Additionally, ¹³C labeled glutamine was also detected in the plasma of these tumor patients. Figures 3D and 3E show ¹³C NMR spectra of C4 carbon multiplet signals of plasma glutamate and glutamine of a GBM patient. Detection of prominent doubly labeled 4 and 5 carbons of glutamine ¹³C isotopomer (D45) in the plasma of a GBM patient (Fig. 3E) indicates that the presence of ¹³C labeled glutamine in the circulating blood generated due to [1,2-¹³C]acetate metabolism by peripheral organs. As illustrated in Fig 3D, C4 glutamate signals are very weak in the plasma. Plasma C4 glutamine (Fig. 3E) ¹³C-multiplet peak areas are given in Fig. 3F and are dominated by D45 with weaker quartet (Q) signals. The following are the summary of the results from these spectral data: (i) some of the blood-borne ¹³C labeled glutamine from the metabolism of [1,2-¹³C]acetate by peripheral tissues may have entered the tumor and both glutamine pools show very similar D45 and Q values, (Fig. 3F and supplementary Table S1). (ii) ¹³C fractional enrichment of C4 glutamate pool in the tumors was 35.90 ± 11.25% ((b+c+d+e) of tumor C4 GLU in Table 1A) whereas, this value in C4 glutamine was only 28.10 ± 6.44% ((b+c+d+e) of tumor C4 GLN in Table 1B). If the ¹³C-glutamine in the tumor was mainly synthesized from ¹³C-glutamate (Tables 1A and B), then the total ¹³C fractional enrichment of the C4 carbon of glutamine is expected to be less than or equal to that of its precursor, glutamate C4 carbon, assuming similar pool sizes. Our current results of lower ¹³C fractional enrichment of C4 glutamine (Table 1B) compared to C4 glutamate in the tumor suggests that ¹³C labeling from infused [1,2-¹³C]acetate entered glutamate pool in the tumor before it reached glutamine. ((d+e)/FE) given in Tables 1A and 1B corresponds to the portion of C4 ¹³C fractional enrichment of glutamate and glutamine that was derived from direct utilization of [1,2-¹³C]acetate by tumor cells. As illustrated in Figure 2, amount of [2-¹³C]acetyl-CoA gives the readout of contributions from peripheral metabolism of infused [1,2-¹³C]acetate to the tumor. Through the following mechanisms, [2-¹³C]acetyl-CoA may have attributed to ¹³C labeling at C4 of glutamate and glutamine (C4S), in addition to background signal due to natural abundance of ¹³C(1.1%): (i) as we reported earlier, ¹³C label from [1,2-¹³C]acetate entering the liver citric acid cycle may enter gluconeogenesis pathways via oxaloacetate pool which led to the detection of multiple ¹³C labeled isotopomers including [1-¹³C]/[1,2-¹³C]glucose in the plasma [8]. [3-¹³C]lactate isotopomer was also detected in the plasma in these patients due to the systemic metabolism of infused acetate. When these ¹³C labeled isotopomers of glucose and lactate enter the tumor, subsequently they generate [2-¹³C]acetyl-CoA, which further yield [4-¹³C]glutamate/glutamine and [3,4-¹³C]glutamate/glutamine after first and multiple turns of the cycle respectively (Figure 2). (ii) increased flux through oxidative branch of pentose phosphate pathway (PPP) was reported in human orthotopic mouse models of GBM and renal cell carcinoma [11–13]. Here, peripherally-derived [1,2-¹³C]glucose in the plasma was used to measure PPP flux relative to glycolysis using the following approach: C3 lactate singlet (C3S) is produced through the combination of flux through oxidative branch of PPP and a minor contribution from natural abundance ¹³C signal of C3 lactate carbon (1.1%). The singlet (C3S) to doublet (D23) ratio of lactate C3 multiplet gives the measure of PPP flux relative to glycolysis (4.81 ± 0.62, Supplementary Table S2). This result indicates that activity of PPP may contribute to ¹³C labeling at C3 lactate which produces [2-¹³C]acetyl-CoA via [3-¹³C]pyruvate in these patient tumors. As described earlier, [2-¹³C]acetyl-CoA yields [4-¹³C]glutamate/glutamine and [3,4-¹³C]glutamate/glutamine after first and multiple turns of the cycle respectively. (iii) pyruvate recycling mechanism in the tumor cells generates [3-¹³C] and [1,2-¹³C]pyruvate isotopomers from [3,4-¹³C] and [1,2-¹³C]OAA isotopomers respectively, after the first turn of the citric acid cycle during the metabolism of [1,2-¹³C]acetate [3,14–16]. [3-¹³C]pyruvate further metabolized to generate [2-¹³C]acetyl-CoA via pyruvate dehydrogenase (PDH) which yielded [4-¹³C]glutamate/glutamine and [3,4-¹³C]glutamate/glutamine in the citric acid cycle. [1,2-¹³C]pyruvate generates [1-¹³C]acetyl-CoA, which upon entering citric acid cycle yield [5-¹³C]glutamate and glutamine isotopomer. Contributions from peripheral metabolism of acetate to overall C4 ¹³C fractional enrichment of glutamate and glutamine were 6.60% ± 0.02% and 9.08 ± 0.03% respectively.

Figure 2. Metabolic model.

Schematic diagram illustrating the fate of the individual ¹³C carbon atoms from infused [1,2-¹³C]acetate. [4,5-¹³C] and [3,4,5-¹³C]GLU/GLN (green filled circles) were generated from acetate-derived [1,2-¹³C]acetyl-CoA. [4-¹³C] and [3,4-¹³C]GLU/GLN (purple filled circles) were mainly generated from blood-borne ¹³C labeled glucose/lactate that originated from [1,2-¹³C]acetate metabolism of peripheral organs. ¹³C labeling patterns in α-KG, GLU and GLN are shown along with the numbers representing the carbon atom positions. Abbreviations: α-KG, α-ketoglutarate; GLU, glutamate; GLN, glutamine; OAA, oxaloacetate; PDH, pyruvate dehydrogenase; ACSS2, nucleocytosolic acetyl-CoA synthetase 2. Filled circle (green or purple) refers to ¹³C and open circle refers to ¹²C carbon nuclei. Green pattern filled circles correspond to ¹³C isotopomers that generate ¹³C labeled C5 glutamate isotopomer and do not influence C4 ¹³C fractional enrichment.

Figure 3. ¹³C NMR spectral data.

(A,B) and (D,E) show the ¹³C NMR spectra of C4 carbon of glutamate and glutamine of tumor and plasma from a GBM patient respectively. In E, * refers to an overlap of unknown carbon signal with one of the D45 GLN doublet peaks. The patient was infused with [1,2-¹³C]acetate prior to the resection of the tumor. Charts C and F show normalized ¹³C peak areas of C4 carbon multiplets of glutamine and glutamate from four brain tumor patients participated in the study. Chart C compares the peak areas between glutamate and glutamine C4 carbon multiplets in the tumor. Chart F shows the comparison of C4 GLN peak areas between plasma and tumor. S, singlet is due to both natural abundance of ¹³C and Fc2; D34, doublet due to ¹³C-¹³C J coupling between the carbons 3 and 4; D45, doublet arises due to ¹³C-¹³C J coupling between the carbons 4 and 5; Q, quartet is due to ¹³C-¹³C J couplings between the carbons 3,4 and 5.

Table 1A.

¹³C fractional enrichments of GLU C4 in the tumors. Calculation of a–e are from ¹³C NMR spectra of C4 GLU as described in the text (Materials and Methods). d and e correspond to the fractions that have [4,5-¹³C] and [3,4,5-¹³C] isotopomers respectively. a is the fraction that has no ¹³C. (b+c) together represents the contributions from the metabolism of [1,2-¹³C]acetate by peripheral organs to tumor GLU C4 ¹³C pool. FE is equal to (b+c+d+e) and corresponds to fractional enrichment of GLU C4 in the tumor.

Tumor type	a	b	c	d	e	FE	(b+c)/FE	(d+e)/FE
GBM	0.758	0.014	0.002	0.188	0.038	0.242	0.066	0.934
GBM	0.717	0.024	0.003	0.222	0.035	0.283	0.094	0.906
Breast metastasis	0.537	0.017	0.006	0.354	0.087	0.464	0.050	0.950
Lung metastasis	0.554	0.019	0.006	0.339	0.083	0.446	0.056	0.944

Table 1B.

¹³C fractional enrichments of GLN C4 in the tumors. Calculation of a–e are from ¹³C NMR spectra of GLN C4 as described in the text (Materials and Methods). d and e correspond to the fractions that have [4,5-¹³C] and [3,4,5-¹³C] isotopomers respectively. a is the fraction that has no ¹³C. (b+c) together represents the contributions from the metabolism of [1,2-¹³C]acetate by peripheral organs to tumor GLN C4 ¹³C pool. FE is equal to (b+c+d+e) and corresponds to fractional enrichment of GLN C4 in the tumor.

Tumor type	a	b	c	d	e	FE	(b+c)/FE	(d+e)/FE
GBM	0.762	0.014	0.002	0.196	0.027	0.238	0.065	0.935
GBM	0.787	0.026	0.002	0.171	0.015	0.213	0.131	0.869
Breast metastasis	0.659	0.021	0.003	0.276	0.041	0.341	0.071	0.929
Lung metastasis	0.669	0.023	0.009	0.271	0.028	0.331	0.096	0.904

Conclusion

In summary, we have illustrated an analysis to determine ¹³C turnovers in glutamate and glutamine pools in four brain tumor patients from the infusion of [1,2-¹³C]acetate prior to surgical removal of the tumor. This analysis included the contributions from ¹³C labeled nutrients detected in the plasma that originated from systemic metabolism of [1,2-¹³C]acetate. ¹³C-glutamine is also detected in the plasma and it was shown to have similar ¹³C isotopomer populations of the glutamine present in the tumor tissues. The current analysis revealed that major portion of C4 ¹³C fractional enrichment of glutamate and glutamine (93.4 ± 0.02% in glutamate and 90.9 ± 0.03% in glutamine) was derived from [1,2-¹³C]acetate-derived acetyl-CoA. The analysis described here provides a simple method to quantity the relative contributions from direct utilization of acetate by tumor cells mediated via ACSS2 and peripheral metabolism of [1,2-¹³C]acetate through ¹³C labeled circulating nutrients to ¹³C fractional enrichment of glutamate and glutamine pool in the tumor. In addition, the presence of ¹³C isotopomers of glucose was utilized to measure relative flux of PPP. This current methodology can be easily extended to similar scenarios for investigating in situ metabolism of other citric acid cycle intermediates and also to other pathologic and physiologic conditions.