Hyperpolarized δ-[1-¹³C]gluconolactone as a probe of the pentose phosphate pathway

Abstract

The pentose phosphate pathway (PPP) is thought to be upregulated in trauma (to produce excess NADPH) and in cancer (to provide ribose for nucleotide biosynthesis) but simple methods for detecting changes in flux through this pathway are not available. Magnetic resonance imaging of hyperpolarized ¹³C-enriched metabolites offers considerable potential as a rapid, non-invasive tool for detecting changes in metabolic fluxes. In this study, hyperpolarized δ-[1-¹³C]gluconolactone was used as a probe to detect flux through the oxidative portion of the pentose phosphate pathway (PPP_ox) in isolated perfused mouse livers. The appearance of hyperpolarized H¹³CO₃⁻ within seconds after exposure of livers to HP-δ-[1-¹³C]gluconolactone demonstrates that this probe rapidly enters hepatocytes, becomes phosphorylated, and enters the PPP_ox pathway to produce HP-H¹³CO₃⁻ after three enzyme catalyzed steps (6P-gluconolactonase, 6PGDH, and carbonic anhydrase). Livers perfused with octanoate as their sole energy source show no change in production of H¹³CO₃⁻ after exposure to low levels of H₂O₂, while livers perfused with glucose and insulin showed a 2-fold increase in H¹³CO₃⁻ after exposure to peroxide. This indicates that flux through the PPP_ox is stimulated by H₂O₂ in glucose perfused livers but not in livers perfused with octanoate alone. Subsequent perfusion of livers with non-polarized [1,2-¹³C]glucose followed by ¹H NMR analysis of lactate in the perfusate verified that flux through the PPP_ox is indeed low in healthy livers and modestly higher in peroxide damaged livers. We conclude that hyperpolarized δ-[1-¹³C]gluconolactone has the potential to serve as a metabolic imaging probe of this important biological pathway.

Graphical abstract

δ-[1-¹³C]gluconolactone was polarized using standard DNP methods as used as a probe of the oxidative portion of pentose phosphate pathway (PPP_ox). The appearance of hyperpolarized H¹³CO₃⁻ within seconds after exposure of livers to the probe demonstrates that it is rapidly transported into hepatocytes and phosphorylated, enters the PPP_ox and produces H¹³CO₃⁻ after three enzyme catalyzed steps. This new probe has the potential to serve as a metabolic imaging probe of this important biological pathway.

1. Introduction

The pentose phosphate pathway (PPP) is an important pathway in glycolytic tissues. It consists of an oxidative portion (PPP_ox) that produces NADPH in two oxidative steps during the conversion of glucose-6-phosphate to ribulose-5-phosphate followed by a non-oxidative portion (PPP_non) that yields ribose-5-phosphate plus several other carbohydrates needed as biosynthetic precursors. Flux through the PPP_ox has been shown to be stimulated in head trauma, in cancer, and in tissues requiring continuous regeneration of glutathione, but simple methods for measuring flux of glucose through this pathway are limited. Relative flux through the PPP_ox has been measured using radioactive tracer studies,^1–11 mass spectral analyses^12–21 and NMR analyses.^22–27 Net flux through the PPP_ox has been most often assessed by administration of [1,2-¹³C₂]glucose and detection of singly-labeled lactate.^{14,15,17,26,27} δ-Gluconolactone, which exists in equilibrium with γ-gluconolactone in aqueous media (Scheme 1), enters mammalian cells via glucose transporters and phosphorylated by gluconokinase (EC 2.7.1.12) to 6-phospho-δ-gluconolactone, an intermediate in the PPP_ox.²⁸ Based on metabolic data obtained with ¹⁴C labeled glucose and ribose in rabbit liver,²⁹ flux through the PPP_ox appears to be fast enough to allow detection of flux through this pathway within a time frame of 1–2 minutes. This makes δ-[1-¹³C]gluconolactone an attractive tracer for hyperpolarized ¹³C studies of the PPP_ox because the ¹³C atom in this derivative would be released as ¹³CO₂ in the second oxidative step of this pathway. Each intermediate and ¹³CO₂/HCO₃⁻ are predicted to have a reasonably long T₁ values and so potentially can be detected by ¹³C MRS for several minutes after injection of hyperpolarized δ-[1-¹³C]gluconolactone. This presents a unique opportunity to measure the flux through the PPP_ox by monitoring production of HP HCO₃⁻ by ¹³C NMR. It is worth noting that the uptake of HP δ-[1-¹³C]gluconolactone will be in competition with the total glucose uptake but the rate of HP ¹³CO₂ production, as detected by HP H¹³CO₃⁻, should in principle reflect relative flux through the PPP_ox (Scheme 2). It has been shown that hydrolysis of CO₂ to HCO₃⁻ catalyzed by carbonic anhydrase is extremely rapid under physiological conditions^30,31 so cannot be rate limiting in the production of HCO₃⁻ from [1-¹³C]gluconolactone. Here, we examined the use of HP-δ-[1-¹³C]gluconolactone to detect flux through PPP_ox by detection of HP H¹³CO₃⁻ in control livers and in livers pre-exposed to hydrogen peroxide to stimulate flux through this pathway. Since NMR is a non-invasive technique, translation of this method into a clinical setting may allow detection of tissues with an active PPP_ox pathway after a single injection of HP δ-[1-¹³C]gluconolactone.

Scheme 1.

Structures of δ-gluconolactone and γ-gluconolactone which interconvert in aqueous solution.

Scheme 2.

The oxidative branch of the pentose phosphate pathway (PPP_ox).

2. Materials and Methods

[1-¹³C]Glucose (99 % enriched) was purchased from Cambridge Isotope Laboratories (Andover, MA) and used without further purification. The trityl radical, tris[8-carboxyl-2,2,6,6-tetra-[2-(1-hydroxyethyl)]-benzo-(1,2-d:4,5-d)-bis-(1,3)-dithiole-4-yl]-methyl sodium salt, was purchased from Oxford Molecular Biotools Ltd. (Abingdon, Oxfordshire, UK) and used without further purification. All other chemicals were obtained from Sigma-Aldrich (St. Louis, MO) at the highest quality available. Female C57BL/6 mice (20–25 g) were obtained from Charles River Laboratories (Wilmington, MA) and fed ad libitum prior to experimentation. The studies were performed under a protocol approved by the Institutional Animal Care and Use Committee at the University of Texas Southwestern Medical Center. Liver enzyme assays, alanine aminotransferase and aspartate aminotransferase, were performed by the Mouse Metabolic Phenotyping Core at the University of Texas Southwestern Medical Center.

δ-[1-¹³C]Gluconolactone was synthesized following the procedure reported for the unlabeled compound.³² A mixture of [1-¹³C]glucose (1.0 g) and Shvo’s catalyst (200 mg) in cyclohexanone (25 mL) was stirred at 50 °C (bath temperature) for 24 hours under argon. The orange reaction mixture was cooled to room temperature and the precipitate was filtered on a Schlenk filter under nitrogen. The precipitate was washed with cyclohexanone (5 mL) and acetone (3 × 10 mL) and dried in a stream of nitrogen to give the desired product, δ-[1-¹³C]gluconolactone, as a white powder (0.70 g, 71 % yield). ¹³C{¹H} NMR (150 MHz, D₂O, δ) 173.80 ppm (C1, s), 81.58 ppm (C3, s), 72.74 ppm (C5, s), 70.95 ppm (C4, s), 67.08 ppm (C2, s), 60.06 ppm (C6, s); MS (ESI) mass calculated for ¹²C₅¹³C₁H₁₀O₆ = 179.1; Found = 178.0 [M–H]^–.

The spin-lattice relaxation times (T₁) of carbon-1 of γ- and δ-[1-¹³C]gluconolactone and [1-¹³C]gluconate were measured at 9.4 T. The T₁ of gluconate was determined on a non-polarized sample using a conventional inversion recovery pulse sequence [d-180º-τ-90º-acquisition] with various τ times. The T₁ values of γ- and δ-[1-¹³C]gluconolactone were determined by fitting the decay of each hyperpolarized signal to:³³

$$\mathrm{S}\left(\mathrm{\tau }\right)={\mathrm{M}}_{\mathrm{p}}{\left(cos\mathrm{\theta }\right)}^{\mathrm{\tau }/\text{TR}}\left(sin\mathrm{\theta }\right){\mathrm{e}}^{-\mathrm{\tau }/\mathrm{T}1}$$

where S(τ) is the signal intensity, M_p is the Z magnetization at time = 0, θ is the pulse width, τ is time, and TR is the total repetition time. The measured T₁’s of hyperpolarized γ- and δ-[1-¹³C]gluconolactone were 19.9 and 17.8 sec, respectively, while the T₁ of non-polarized [1-¹³C]gluconate was 17.3 sec. All measurements were done at 9.4T. Given that δ-[1-¹³C]gluconolactone and γ-[1-¹³C]gluconolactone were exchanging during the decay process, these values can only be considered “apparent” T₁ values. An early report on the kinetics of this interconversion demonstrated that this exchange reaches equilibrium in water only after ~25–30 s.⁴⁹ It is also interesting to note that [1-¹³C]gluconate does not appear to be an intermediate in this exchange reaction so it is safe to assume that once [1-¹³C]gluconate is formed in tissue, it does not spontaneously convert to either γ- or δ-[1-¹³C]gluconolactone.⁴⁹

δ-[1-¹³C]Gluconolactone was polarized using an Oxford HyperSense® DNP polarizer. This instrument is a “second generation” device designed for routine operation in a research lab setting.³⁴ It is equipped with a 3.35 Tesla superconducting magnet, a microwave source, a solid state nuclear magnetic resonance (NMR) spectrometer for monitoring polarization build up, and a rapid dissolution system for transporting the sample from the HyperSense to a 400 MHz NMR spectrometer for spectroscopy and imaging. ~9 mg of δ-[1-¹³C]gluconolactone was dissolved in a (3:1) water:glycerol mixture doped with the trityl radical at 15 mM. The frozen sample was lowered into the magnetic field, cooled to 1.4 K in a pumped helium bath and the microwave irradiation was turned on. After maximal polarization was achieved (1–2 hours), the irradiation was turned off and the sample was rapidly dissolved using 4 mL of hot (> 190 °C) PBS (10 mM, pH ~ 7.4) and transferred into the bore of an 89 mm vertical 9.4 T NMR spectrometer for spectral acquisition. For tissue experiments, the hyperpolarized solution was introduced into the perfusate chamber directly above a perfused liver via a catheter running to the top of the perfusion rig. The average ¹³C signal enhancement of HP-δ-[1-¹³C]gluconolactone observed a few seconds after dissolution and placement into a 9.4T magnet was 7200; this corresponds to a polarization level of 5.6%.

Livers isolated from C57BL/6 lab mice were perfused in an all glass perfusion system that fits entirely into the bore of an 89 mm vertical bore 9.4 T NMR spectrometer. Mice were anesthetized by an intramuscular injection (0.1 mL) of ketamine (Fort Dodge Animal Health)/xylazine (Boehringer Ingelheim) in a ratio of 85:15. Once anesthetized, a midline laparotomy was performed to expose the hepatic portal vein which was cannulated and injected with heparin (50 IU) to prevent blood clots. The hepatic vein and inferior vena cava were dissected, and flow of perfusate through portal vein was established using a peristaltic pump at ~ 8 mL/min. The excised liver was suspended in a 20 mm NMR tube attached to the perfusion apparatus maintained at 37 °C and perfused at a rate of 8 mL/min with a standard non-recirculating buffered medium (described below) and bubbled continuously with (19:1) O₂/CO₂. Efferent and afferent pO₂ was measured using a blood gas analyzer (GEM Premier 3000, Instrumentation Laboratory, Lexington, MA, USA) to determine oxygen consumption.³⁵ Hepatic viability was evaluated by oxygen consumption measurements and visual inspection. The perfusion apparatus equipped with twin chambers, water jacketed for temperature control and protected by a polycarbonate outer shell was lowered into the bore of the 9.4T wide-bore magnet. The perfusion pressure was adjusted to maintain an adequate perfusion flow through the portal vein. Hyperpolarized substrates were introduced through polyethylene tubing that terminated near the portal vein catheter. A schematic diagram of the apparatus is shown in supplementary materials.

Livers from either 15 h fasted mice or mice given full access to chow ad libitum were perfused with a modified Krebs-Henseleit (KH) medium containing (in mM): NaHCO₃ (25), NaCl (118), KCl (4.7), MgSO₄ (1.2), KH₂PO₄ (1.2), CaCl₂ (1.25). The perfusate presented to fasted livers also contained sodium octanoate (0.4 mM) as the only carbon energy source. Sodium octanoate was chosen because this intermediate chain-length fatty acid is avidly utilized by tissues, does not require carnitine for transport into mitochondria, and thereby insures that the livers remain in a highly reduced oxidation state. Our hypothesis was that octanoate would supply excess NADH through β-oxidation and any excess reducing equivalents would be cycled into the pool of NADPH via transhydrogenase enzymes (NADPH:NAD⁺ oxidoreductase) to replenish NADP⁺ during exposure of these livers to H₂O₂. Thus, flux through the PPP_ox is predicted to be low in livers perfused with octanoate before and after exposure to H₂O₂. Conversely, in livers from fed mice perfused with 10 mM glucose ± insulin (240 μU/mL) as their only carbon energy source, we hypothesized that flux through the PPP_ox would, by necessity, be stimulated to replenish NADPH in livers exposed to H₂O₂.

After positioning each liver in a 20-mm NMR tube, the viability of the organ preparation was first confirmed by flow rate and O₂ consumption measurements before the liver was placed in the bore of a Varian VNMRS 9.4 T magnet. During a preparation and stabilization time of approximately 20 min, the NMR probe was tuned and the field homogeneity optimized using the ²³Na free induction decay. A ²³Na line width of 15 Hz or less was typically obtained. ¹³C NMR spectra were acquired at ~ 100 MHz in a 25 mm Doty broadband probe using 66° pulses. Livers exposed to H₂O₂ were initially exposed to a perfusate lacking H₂O₂ during a preparation and stabilization period required for O₂ consumption measurements, probe tuning, and sample shimming. Following shimming, the perfusate was switched to one containing H₂O₂ for the remainder of the experiment. After 10–13 min of peroxide exposure, 8 mL of 4 mM HP δ-[1-¹³C]gluconolactone was injected into the perfusion chamber directly above the cannulation site over a period of ~ 30 s. Serial FIDs were acquired immediately after injection of the HP substrate using 16 K complex data points over a ± 32,000 Hz bandwidth without proton decoupling to yield an acquisition time of ~ 1 s and a delay of 4.0 s, giving a repetition time of 5 s for each scan. Typically 20 separate FIDs were collected. These data were zero-filled prior to Fourier transformation and the relative peak areas were measured by integration using the ACDLabs SpecManager version 12.00 (Advanced Chemistry Development, Inc.; Toronto, Canada).

The relative activity of the PPP_ox pathway was also assessed in fed livers perfused with non-polarized 10 mM [1,2-¹³C]glucose and insulin (240 μU/mL), with or without exposure to H₂O₂. These livers were isolated and perfused on the bench using an identical protocol as that described for hyperpolarization experiments. After 30 minutes of perfusion, the livers were freeze-clamped and all water soluble metabolites were extracted into cold perchloric acid. Samples were reconstituted in D₂O and high resolution ¹H spectra were collected at 14.1 T.

All data are reported as mean ± 1 s.d. Statistical significance was determined by a two-tailed t-test using GraphPad Prism version 5.03 (GraphPad Prism Software, Inc, La Jolla, CA). All results are reported at the 95 % confidence level.

3. Results

Figure 1 shows stacked plots of ¹³C NMR spectra of a mouse liver immediately after addition of 4 mM HP δ-[1-¹³C]gluconolactone to an aqueous perfusate feeding the liver. Four resonances were observed and assigned to 6-phospho-[1-¹³C]gluconate (178.6 ppm), 6-phospho-γ-[1-¹³C]gluconolactone (177.0 ppm), 6-phospho-δ-[1-¹³C]gluconolactone (173.8 ppm), and H¹³CO₃⁻ (160.3 ppm) by comparisons with known standards. To evaluate which of these species resulted from metabolism of HP δ-[1-¹³C]gluconolactone by liver, a ¹³C spectrum of HP δ-[1-¹³C]gluconolactone injected into aqueous perfusate (no liver) is compared in Figure 2. In this case, the resonances of δ-[1-¹³C]gluconolactone and γ-[1-¹³C]gluconolactone were both observed but those of [1-¹³C]gluconate and H¹³CO₃⁻ were not observed. This indicates that these two metabolites were generated by metabolism of HP-δ-[1-¹³C]gluconolactone in liver. Both metabolites are products of the PPP_ox pathway. Given that the chemical shifts of these intermediates are likely insensitive to the presence of a phosphate group at position C6 and the fact that the resonance at 178.6 ppm is much larger in the liver spectrum, this resonance must reflect the phosphorylated product, 6-phospho-[1-¹³C]gluconate. Similarly, since 6-phospho-[1-¹³C]gluconate is a metabolic product of 6-phospho-δ-[1-¹³C]gluconolactone and H¹³CO₃⁻ is a metabolic product of 6-phospho-[1-¹³C]gluconate, we conclude that all three carbohydrate intermediates detected by ¹³C NMR in livers must be 6-phospho derivatives of these sugars.

Figure 1.

A stack plot of ¹³C NMR spectra collected on an isolated mouse liver after continuous addition of hyperpolarized δ-[1-¹³C]gluconolactone to the perfusion media flowing into the portal vein of the liver at 8 mL/min. The ¹³C spectra collected every 5 sec using a 66° pulse. The resonance assignments are described in the text.

Figure 2.

¹³C NMR spectra of hyperpolarized δ-[1-¹³C]gluconolactone injected into a) a NMR tube containing perfusion medium or b) an isolated perfused mouse liver as shown in Figure 1. Both spectra reflect summed 10 summed spectra collected over 50 s. The insets show expansion of the bicarbonate region (155 – 175 ppm).

A comparison of the summed total intensity of the HP-H¹³CO₃⁻ resonances in 20 spectra collected over 50 s for four different liver groups is shown in Figure 3a. These data show than production of HP-H¹³CO₃⁻ in livers perfused with octanoate alone was unaltered by addition of 1% H₂O₂ while in livers perfused with glucose plus insulin alone, HP-H¹³CO₃⁻ production was about 50% lower in the absence of H₂O₂ but returned to basal levels after addition of H₂O₂. In comparison, Figure 3b shows that production of 6-phospho-[1-¹³C]gluconate was constant in all four groups of livers. This indicates that the oxidative reaction that produces NADPH (6PGDH) is rate limiting in this portion of the PPP_ox pathway, while rate of production of 6PG from δ-[1-¹³C]gluconolactone is independent of both physiology (octanoate versus glucose perfusions) and any oxidative tissue damage caused by H₂O₂.

Figure 3.

a) Fraction of the summed areas of the HP-HCO₃⁻ signal over 50 s relative to the total area of all HP ¹³C signals. b) Fraction of the summed area of the HP-6-phosphogluconate signal over 50 s relative to the total area of all HP ¹³C signals. c) The ratio of HP-HCO₃⁻ / HP-6PG signals (AUC) as an index of PPP_ox. The black column represents livers perfused with octanoate without H₂O₂; the dark grey column reflects livers perfused with octanoate plus 1% H₂O₂; the light grey column reflects livers perfused with glucose plus insulin without H₂O₂; the striped column reflects livers perfused with glucose plus insulin plus 1% H₂O₂. All data are reported as the mean ± 1 s.d. for four livers in each group. P = 0.02 for * comparisons in Fig. 3a; P = 0.03 for * comparisons in Fig. 3b.

The kinetic rates of production of 6PG and H¹³CO₃⁻ from δ-[1-¹³C]gluconolactone in liver are compared in Figure 4. While the temporal appearance of HP-6PG did not differ significantly among the four groups (Figure 4b), the signal of HP-HCO₃⁻ appeared more rapidly and reached a higher amplitude in the livers perfused with octanoate compared to livers perfused with glucose plus insulin (Figure 4a). This provides further evidence that flux through 6PGDH is rate limited in livers perfused with glucose plus insulin but not rate limited in livers perfused with octanoate.

Figure 4.

Time-dependent changes in a) the HP-HCO₃⁻ and b) HP-6PG after presenting HP δ-[1-¹³C]gluconolactone to isolated, perfused livers. ( ) perfused with octanoate; ( ) perfused with octanoate plus H₂O₂; ( ) perfused with glucose plus insulin; ( ) perfused with glucose plus insulin plus H₂O₂. Each data point reflects the mean ± 1 s.d. at that time point for n=4 livers in each group.

To validate that flux through the PPP_ox was indeed active in livers perfused with glucose, a separate group of livers from fed mice were perfused on the bench with [1,2-¹³C₂]glucose plus insulin for 30 min followed by freeze-clamping, tissue extraction and ¹H NMR. In this series, direct evidence for flux through the PPP_ox was shown by the ratio of [3-¹³C]lactate to [2,3-¹³C₂]lactate in the ¹H NMR spectrum of tissue extracts. [2,3-¹³C₂]lactate reflects direct conversion of [1,2-¹³C]glucose to lactate without passage through PPP_ox while [3-¹³C]lactate reflects metabolism of [1,2-¹³C₂]glucose through PPP_ox to produce [1-¹³C]ribulose-5-phosphate followed by carbon-carbon rearrangements in the PPP_non to produce either [1-¹³C]- or [3-¹³C]fructose-6-phosphate and ultimately [1-¹³C] or [3-¹³C]lactate. Representative ¹H spectra of tissue extracts of control and H₂O₂ treated livers perfused with [1,2-¹³C₂]glucose indicate that [2,3-¹³C₂]lactate and [3-¹³C]lactate are present in both groups of livers. The ratio of [3-¹³C]lactate/[2,3-¹³C]lactate is relatively small in both spectra consistent with about ~5% of glucose passing through the PPP_ox. Although the amount of [2,3-¹³C]lactate appears to be slightly lower in spectra of livers perfused with H₂O₂ (consistent lower flux through the PPP_ox as reported by the HP studies), this difference did not reach statistical significance.

4. Discussion

Flux through the oxidative portion of the pentose phosphate pathway (PPP_ox) has been reported to be as low as 1% of glycolytic flux in glial cells²⁰ to as high as 23% in rat livers.^4,5 Absolute PPP_ox flux has been reported to be between 0.20 – 0.40 μmol/min/g wet weight in the rat liver^4,5,36,37 under basal conditions, about 20 % of glycolytic flux.^4,5,38 After an oxidative insult, flux through the PPP_ox typically increases to produce NADPH to replenish the reduced form of glutathione (GSH). For example, Dusick et al infused [1,2-¹³C]glucose into patients suffering from a traumatic brain injury to determine PPP_ox flux by collecting blood samples and analyzing the lactate by gas-chromatography/mass spectrometry.¹⁷ An increase of 7–20% of total glycolytic flux through the PPP_ox was observed in patients with brain injury compared to non-injured patients. Traditionally, measurement of the PPP_ox flux has been determined by measuring ¹⁴CO₂ production or from the labeling patterns in lactate caused by the co-infusion of [1-^14/13C]glucose and [6-^14/13C]glucose or infusion of [1,2-¹³C₂]glucose.^{3–5,7–11,13–15,17–21,26,27,39,40} These studies typically need hours to produce sufficient label through the PPP_ox to properly assay its flux. Hence, a HP probe that could potentially provide a quick readout of increase flux through the PPP_ox would be of interest.

In the current study, exposure of HP δ-[1-¹³C]gluconolactone to livers resulted in generation of HP H¹³CO₃⁻ as detected by ¹³C NMR. In this case, HP H¹³CO₃⁻ could only have been generated via decarboxylation of HP 6-phospho-[1-¹³C]gluconate, a reaction catalyzed by 6PGDH in the PPP_ox pathway (Scheme 2). The bicarbonate fraction of total ¹³C HP signal in livers perfused with octanoate was identical to that seen in livers perfused with glucose and insulin after exposure to H₂O₂ (Figure 3) but the temporal dependence of the HP signals differed (Figure 4). The initial rate of appearance of HP HCO₃⁻ in livers from fasted animals perfused with octanoate with or without H₂O₂ was more rapid, reached higher maximum intensity, then dropped off more rapidly compared to all other groups. This can be ascribed to less competition between HP δ-[1-¹³C]gluconolactone and any endogenous glucose for metabolism in the PPP_ox. The fact that flux through PPP_ox in livers perfused with octanoate did not increase with addition of H₂O₂ suggests that this medium-chain length fatty acid supplies sufficient reducing equivalents to maintain cellular redox levels without stimulating the PPP_ox. This is reflected by the identical ratio of HP-bicarbonate/HP-6PG measured in livers perfused with octanoate ± H₂O₂ (Figure 3c). The data show that the kinetics of appearance of 6PG was identical in all four groups indicating that the activity of 6-phosphogluconolactonase is insensitive to perfusion conditions. 6PG is known to be an allosteric activator of PFK^41–44 and PK^45,46 so in livers perfused with octanoate but no glucose, the amount of available 6PG would be limited to that formed from HP-6-phospho-δ-gluconolactone plus small amounts generated from endogenous stores of glucose. In livers perfused with glucose plus insulin, one can assume that additional 6PG would be produced from non-polarized glucose in addition to that produced from HP-6-phospho-δ-[1-¹³C]gluconolactone, so higher levels of 6PG would stimulate glycolysis without necessarily stimulating flux through the PPP_ox pathway. In fact, in those livers, net flux through the PPP_ox as measured by the ratio of HP-bicarbonate/HP-6PG was lower by ~50% compared to the other groups. This was the only metabolic condition that resulted in an increase in flux through the PPP_ox after addition of H₂O₂ as detected by generation of HP-HCO₃⁻. In experiments designed to validate the HP results using non-polarized [1,2-¹³C]glucose plus insulin (Figure 5), PPP_ox flux was also found to be low (< 5%) under basal perfusion conditions. Although addition of H₂O₂ to those livers tended to decrease the amount of [2,3-¹³C]lactate compared to [3-¹³C]lactate, this decrease did not reach statistical significance. This suggests that the HP measurement is more sensitive to small changes in PPP_ox than is the steady-state measurement of flux from [1,2-¹³C]glucose to lactate. One should note that other mono-labeled glucose derivatives could in principle be formed in these experiments as a result of ¹³C-enriched intermediates leaving the TCA cycle via pyruvate cycling. This contribution would reduce the estimate of PPP_ox flux activity even lower than suggested by the ¹H NMR measurements.

Figure 5.

¹H spectra of isolated mouse livers perfused with [1,2-¹³C₂]glucose plus insulin. Top spectrum: Liver exposed to 1 mM H₂O₂. Bottom spectrum: Control liver. The majority of lactate produced during the perfusion period was not enriched in ¹³C. The small wings around the large unlabeled lactate doublet at 1.33 ppm show that small amounts of [2,3-¹³C]lactate and [3-¹³C]lactate are both present.

From a biological perspective, an imaging method for detecting flux through 6-phosphogluconate dehydrogenase (6PGDH) should reflect flux through the entire oxidative portion of the pentose phosphate pathway (PPP_ox). Although the HP method presented here has been shown to be sensitive to changes in relative flux rates through the PPP_ox, one cannot obtain absolute flux values in units of μmol/min/gww of tissue nor can it report the percent of total glucose contributing to glycolysis that bifurcates through the PPP_ox pathway first. We made no attempt to fit the kinetic curves of Figure 4 to a kinetic model largely because the conversion of HP-δ-[1-13C]gluconolactone to HP-bicarbonate involves four enzyme-catalyzed steps, gluconokinase, 6P-gluconolactonase, 6PGDH, and carbonic anhydrase so obtaining quantitative kinetic constants for each of these steps would require may assumptions that we cannot currently justify. Given that other investigators have demonstrated model-free area-under-the-curve (AUC) measurements may work equally well for analysis of kinetic HP data,⁴⁷ our initial observations here suggest that that the ratio of AUC’s for bicarbonate/6PG appears to be sensitive to changes in flux through the PPP_ox at least in this limited data set. Further studies will be necessary to test whether HP-δ-[1-13C]gluconolactone is a generally useful probe of the PPP_ox pathway in other tissues and in vivo.

Abbreviations

PPP

pentose phosphate pathway

PPP_ox

oxidative branch of the PPP

MRS

magnetic resonance spectroscopy

MRI

magnetic resonance imaging

DNP

dynamic nuclear polarization

HP

hyperpolarized/hyperpolarization

ppm

parts per million

KH

Krebs-Henseleit

6PG

6-phosphogluconate

6PGDH

6-phosphogluconate dehydrogenase

PDH

pyruvate dehydrogenase

PEPCK

phosphoenolpyruvate carboxykinase

PFK

phosphofructokinase

PK

pyruvate kinase