Applications of NMR spectroscopy to systems biochemistry

Abstract

The past decades of advancements in NMR have made it a very powerful tool for metabolic research. Despite its limitations in sensitivity relative to mass spectrometric techniques, NMR has a number of unparalleled advantages for metabolic studies, most notably the rigor and versatility in structure elucidation, isotope-filtered selection of molecules, and analysis of positional isotopomer distributions in complex mixtures afforded by multinuclear and multidimensional experiments. In addition, NMR has the capacity for spatially selective in vivo imaging and dynamical analysis of metabolism in tissues of living organisms. In conjunction with the use of stable isotope tracers, NMR is a method of choice for exploring the dynamics and compartmentation of metabolic pathways and networks, for which our current understanding is grossly insufficient. In this review, we describe how various direct and isotope-edited 1D and 2D NMR methods can be employed to profile metabolites and their isotopomer distributions by stable isotope-resolved metabolomic (SIRM) analysis. We also highlight the importance of sample preparation methods including rapid cryoquenching, efficient extraction, and chemoselective derivatization to facilitate robust and reproducible NMR-based metabolomic analysis. We further illustrate how NMR has been applied in vitro, ex vivo, or in vivo in various stable isotope tracer-based metabolic studies, to gain systematic and novel metabolic insights in different biological systems, including human subjects. The pathway and network knowledge generated from NMR- and MS-based tracing of isotopically enriched substrates will be invaluable for directing functional analysis of other 'omics data to achieve understanding of regulation of biochemical systems, as demonstrated in a case study. Future developments in NMR technologies and reagents to enhance both detection sensitivity and resolution should further empower NMR in systems biochemical research.

Keywords: Hyperpolarization; In vivo NMR; Metabolic network and flux; Stable isotope resolved metabolomics; Stable isotope tracers.

Fig. 1

Influence of extraction procedure on NMR spectral quality of tissue extracts. Patient-derived tumor tissues were implanted into Nod/SCID Gamma mice and let grow to ca. 1 cm in diameter before harvest and pulverization in liquid N₂ for metabolite extraction using our CH₃CN:H₂O:CHCl₃ (2:1.5:1 v/v) solvent partitioning method [384]. The resulting polar extract was dissolved in 50% D₂O for 1D ¹H NMR measurement at 14.1 T (black line), followed by 80% acetone treatment to remove residual proteins and analysis again by ¹H NMR (red line). The two ¹H spectra were normalized by the peak area of the H3 resonance of lactate (3-Lac). The spectral region from 0.8 to 4.7 ppm is shown, along with the CH₃ resonance of the DSS-d₆ standard in the inset. The latter was much lower in intensity and broader in linewidth without acetone than with acetone precipitation of proteins, which indicates the influence of DSS binding to residual proteins on resonance quality. In contrast, the signal intensities of most metabolites were enhanced by acetone precipitation (e.g. H1-βglucose, H2-Gly), which is again consistent with signal attenuation by protein binding. In contrast, the reduced glutathione (GSH) resonances (i.e. H2-Cys and H4-Glu resonances of GSH) were absent from the spectrum after acetone treatment, which could be attributed to insolubility of GSH in 80% acetone. Re-extraction of the protein pellet with 60% acetonitrile restored the GSH signals.

Fig. 2

¹³C-lactate satellite patterns in ¹H NMR. Panels (A–D) display the simulated splitting patterns of the ¹H-3 resonance of lactate with these methyl protons coupled to ¹²C-3 (A), to ¹³C-3 or ¹³C-2 (B), to ¹³C₂-2,3 (C), and ¹³C₃-1,2,3 (D). An example ¹H spectrum of cell culture medium after 24 h of lung adenocarcinoma A549 cell growth in U-¹³C₆-glucose is shown in (E), where the complex splitting pattern of ¹³C satellites of H-3 present in U-¹³C₃-lactate is evident.

Fig. 3

¹³C coupling patterns of ¹³C-enriched metabolites in ¹³C NMR. (A) ¹³C NMR spectrum of monoacetone glucose derived from plasma glucose of rat with infusion of ¹³C₂-3,4-glucose and IP injection of [U-¹³C]proprionate and ²H₂O; S: singlet; D34, D12 or D23: doublet from C-3 and C-4, C-1 and C-2 or C-2 and C-3 coupling, respectively; Q: doublet of doublet from C-2 coupling to both C-1 and C-3 [140]. Reprinted from Anal Biochem, 327, E.S. Jin, J.G. Jones, M. Merritt, S.C. Burgess, C.R. Malloy, A.D. Sherry. Glucose production, gluconeogenesis, and hepatic tricarboxylic acid cycle fluxes measured by nuclear magnetic resonance analysis of a single glucose derivative, pp. 149–155, (2004), with permission from Elsevier. (B) ¹³C NMR spectrum of lipids extracted from glioblastoma cultured in [U-¹³C₆]glucose. The inset shows the expansion of the 23 ppm region. ω, ω-1, and ω-2: terminal CH₃, penultimate CH₂, and CH₂ neighbor to ω-1 CH₂; d: doublet; t: triplet; FA-(CH₂)n-: internal CH₂ of fatty acyl chains. Excerpted with permission from R.J. DeBerardinis, A. Mancuso, E. Daikhin, I. Nissim, M. Yudkoff, S. Wehrli, C.B. Thompson, Beyond aerobic glycolysis: transformed cells can engage in glutamine metabolism that exceeds the requirement for protein and nucleotide synthesis, Proc Natl Acad Sci USA, 104 (2007) 19345–19350 [129]. Copyright (2007) National Academy of Sciences, U.S.A.

Fig. 4

2D ¹H TOCSY detects four different isotopomers of glutamate. A549 cells were grown in unenriched or ¹³C₆-glucose for 24 h before extraction for polar metabolites. 2D ¹H TOCSY analysis of unenriched extracts identifies covalent connectivity of Glu-2H to Glu-4H as a cross-peak (A). The corresponding analysis of the ¹³C enriched extract detects 3 pairs of satellite (sat) cross-peaks that are equidistant from the central cross-peak (●) of Glu-2H to Glu-4H (B). , , : satellite cross-peaks of Glu-4H, Glu-2H, and Glu-2,4H, respectively. The four positional isotopomers of Glu that give rise to these cross-peak patterns are shown in (C).

Fig. 5

2-D ¹H–¹³C HCCH–TOCSY analysis detects positional isotopomers with consecutive ¹³C labels. Lung adenocarcinoma A549 cells were grown in ¹³C₆-glucose under control or selenite (Se) treatment. HCCH–TOCSY analysis was performed on the polar cell extracts to enable selective detection of protons attached to consecutively bonded ¹³C atoms, as illustrated in (E). The 2D spectrum in (D) reveals Glu and lactate isotopomers consecutively ¹³C labeled at C-2 and C-3 positions, as well as adenine nucleotide (AXP) isotopomer consecutively ¹³C labeled at C-1 to C-5 of the ribose unit. Also shown are the 1D high-resolution ¹H NMR spectrum of the control cell extract (B) as well as the 2D projection spectra of the control (C) and selenite-treated (A) cell extracts. The ¹³C editing also enhances the detection of minor ¹³C labeled components such as glycerophosphocholine (GPC) and unresolved ¹³C labeled components such as uracil nucleotides (UXP) (denoted by *) (cf. B and C). Such a capability is unique in revealing the decreased synthesis of ¹³C₅-1,2,3,4,5-ribose containing UXP and increased production of ¹³C₃-1,2,3-glycerol containing GPC by the selenite treatment (cf. A and C) (adapted from [108]).

Fig. 6

1D ¹H{¹³C}-HSQC analysis provides an overview of ¹³C positional isotopomers of metabolites in a ¹³C labeled PC9 cell extract. Lung adenocarcinoma PC9 cells were grown in ¹³C₆-glucose or unlabeled glucose for 24 h before extraction. The two polar extracts were subjected to 1D ¹H{¹³C}-HSQC analysis for ¹³C positional isotopomers of various metabolites. Lac: lactate; GSH, GSSH: reduced and oxidized glutathiones; PCholine: phosphocholine; Me: methyl; UXP, CXP, AXP: uracil, cytosine, and adenine nucleotides, respectively; UDPGlcNAc: UDP-N-acetylglucosamine; UDPG: UDP-glucose.

Fig. 7

1D ¹H{¹⁵N}-HSQC analysis provides sensitive detection of ¹⁵N-enriched metabolite derivatives. A549 cell extract was derivatized with an ¹⁵N-enriched aminooxy probe ¹⁵N-QDA (N-(2-¹⁵N-aminooxyethyl)-N,N-dimethyl-1-dodecylammonium), which form oxime adducts with carbonyl-containing metabolites such as acetone, pyruvate, and sugar aldehydes (A). The 2 or 3-bond coupling between ¹⁵N of the probe and ¹H of the metabolites enables detection of ¹⁵N via its covalently linked ¹H by ¹H{¹⁵N}-HSQC analysis (B).

Fig. 8

Spin topology pathways for ¹³C₅,¹⁵N₁-glutamate in 2 or 3D NMR experiments. Various 3D NMR experiments that can be performed as the 2D version for ¹³C and ¹⁵N labeled Glu are depicted. Arrows denote magnetization transfer pathways in ¹³C₅,¹⁵N₁-Glu via homonuclear and heteronuclear scalar couplings. C: ¹³C; N: ¹⁵N.

Fig. 9

2D ¹H{¹³C}HR-HSQC analysis provides rigorous identification of ¹³C positional isotopomers of metabolites via ¹³C–¹³C coupling patterns. A polar extract of A549 cells grown in ¹³C₆-glucose was analyzed by high-resolution (hr) ¹H{¹³C}-HSQC at 18.8 T. The 2D HSQC contour plot (A) is shown along with the 1D projection spectrum in the ¹³C dimension (B), where the ¹³C–¹³C coupling patterns are better visualized. The doublet of C-1′, triplet of C-2′, and doublet of C-5′ of ribose attached to the adenine ring (A) of NAD⁺ indicate the presence of ¹³C₅-1,2,3,4,5-ribose-containing NAD⁺ isotopomer, while the triplet of C-2-lactate (Lac) identifies the presence of ¹³C₃-1,2,3-lactate.

Fig. 10

¹H{¹³C}HMBC detects multiple bond ¹H–¹³C connectivity including that between carbons with no directly attached proton(s) or carbons bonded to heteroatoms in the covalent network of given metabolites. The 2D HMBC contour plots of aliphatic (A) and carbonyl/double bonded carbon regions (B) are shown along with the corresponding 1D high-resolution ¹³C NMR spectral regions of a polar extract obtained from lung adenocarcinoma PC9 cells grown in ¹³C₆-glucose for 24 h. The cross-peak for C-5 (C=O) to H-4 of Glu is evident, along with those for C-4 to H-3, as well as C-2 and C-3 to H-4. Also evident are the cross-peaks for C-1 (C=O) to H-3 and C-2 to H-3 of Ala. The expanded 1D ¹³C spectra for C-2, C-3, C-4, and C-5 of Glu as well as C-1 and C-2 of Ala illustrate more clearly the ¹³C–¹³C coupling patterns. C: ¹³C.

Fig. 11

2D ¹H{¹³C}HCACO analysis provides rigorous identification of ¹³C positional isotopomers of carbonyl-containing metabolites in the polar extract of A549 cells. This 2D experiment detects covalent connectivity as cross-peaks from ¹³C-carbonyl carbon to aliphatic protons of metabolites via ¹³C–¹³C coupling, as illustrated in the example structure diagram (top right). The cross-peaks of ¹³C-1 to H-3 and ¹³C-1 to H-2 of lactate (Lac) denote the presence of ¹³C₃-1,2,3-lactate, those of ¹³C-1 to H-2 and ¹³C-4 to H-3 of Asp indicate the presence of ¹³C₂-1,2- and ¹³C₂-3,4-Asp, respectively, and those of ¹³C-1 to H-2, ¹³C-1 to H-3, and ¹³C-5 to H-4 of Glu are consistent with the presence of ¹³C₂-1,2-, ¹³C₃-1,2,3-, and ¹³C₂-4,5-Glu, respectively.

Fig. 12

2D ¹³C–¹³C INADEQUATE analysis provides rigorous identification of ¹³C positional isotopomers of ¹³C-enriched metabolites in the polar extract of PC9 cells grown in ¹³C₆-glucose. This 2D experiment detects carbon covalent connectivity as cross-peaks via ¹³C–¹³C coupling, as illustrated in the example structure diagram (top left). The cross-peaks of ¹³C-1 to ¹³C-2 to ¹³C-3 of Ala and lactate (Lac), ¹³C-1 to ¹³C-2 of Gly, and ¹³C-4 to ¹³C-5 of Glu as well as their splitting patterns unambiguously indicate the presence of ¹³C₃-1,2,3-Ala and -lactate, ¹³C₂-4,5-Glu, and ¹³C₂-1,2-Gly, respectively. Shown above the 2D contour plots are the corresponding 1D high-resolution ¹³C NMR spectral regions.

Fig. 13

Example freeze-clamping devices. A Wollenberger-like device [385] is routinely used in our laboratory for rapid freezing of human and animal tissues to quench metabolism. A commercially available freeze-clamping device for small tissue freezing (BioSqueezer) is also shown (http://www.biosciencetechnology.com/product-releases/2010/12/snap-freeze-tissue-clamp).

Fig. 14

MTBE-based extraction gives comparable ¹H NMR profile of polar metabolites as CHCl₃-based extraction. An NSCLC patient-derived lung cancer cell line (PDLC216) was quenched in cold CH₃CN before extraction of polar and non-polar metabolites by the addition of H₂O and CHCl₃ or MTBE. The solvent composition for the CHCl₃ extraction method was CH₃CN:H₂O:CHCl₃ (2:1.5:1) for the 1st extraction and CHCl₃:methanol:butylated hydroxytoluene (2:1:1 mM) for the 2nd extraction. The solvent composition for the MTBE extraction method was CH₃CN:H₂O:MTBE₃ (2:1:3) for the 1st extraction and MTBE:methanol:H₂O (10:3:2.5) for the 2nd extraction.

Fig. 15

¹⁵N-cholamine tagging of carboxyl groups enables their selective analysis in human serum by ¹H{¹⁵N}-HSQC. The reaction of ¹⁵N-cholamine with carboxylates is shown in (A) and the ¹H{¹⁵N}-HSQC contour plot of a tagged human serum is illustrated in (B). (1) Aconitic acid; (2) adipic acid; (3) alanine; (7) aspartic acid; (8) betaine; (9) citric acid; (11) cystine; (12) formic acid; (15) glutamic acid; (17) glycine; (20) histidine; (21) 3-hydroxybutyric acid; (24) isocitric acid; (28) lactic acid; (29) leucine; (32) malic acid; (37) phenylalanine; (40) pyroglutamic acid; (45) threonine; (46) tryptophan; (47) tyrosine; (48) valine. Reprinted with permission from F. Tayyari, G.A. Gowda, H. Gu, D. Raftery, ¹⁵N-cholamine–a smart isotope tag for combining NMR- and MS-based metabolite profiling, Anal Chem, 85, 8715–8721. Copyright (2013) American Chemical Society [199].

Fig. 16

Carbonyl tagging by ¹⁵N-QDA enables their selective analysis by ¹H{¹⁵N}-HSQC. A549 cells were quenched and reacted in cold acetonitrile containing 0.3 mM ¹⁵N-QDA (N-(2-¹⁵Naminooxyethyl)-N,N-dimethyl-1-dodecylammonium) to form QDA adducts of carbonylated metabolites such as pyruvate in (A). The tagged extract was analyzed by both 1D (B) and 2D ¹⁵N-edited HSQC (C, adapted from Fig. 5 of A.N. Lane, S. Arumugam, P.K. Lorkiewicz, R.M. Higashi, S. Laulhé, M.H. Nantz, H.N.B. Moseley, T.W. M. Fan, Chemoselective detection and discrimination of carbonyl-containing compounds in metabolite mixtures by ¹H-detected ¹⁵N nuclear magnetic resonance, Magnetic Resonance in Chemistry, 53 (2015) 337–343 with permission from John Wiley & Sons [142]) at 14.1 Tesla. DHAP: dihydroxyacetone-3-phosphate; Pyr: pyruvate; Ara: arabinose.

Fig. 17

Metabolomics-edited transcriptomic analysis for selenite-induced perturbations of glycolysis, pentose phosphate pathway, TCA cycle, glutaminolysis, and fatty acid metabolism in A549 cells. Panel (A) shows the pathways perturbed in selenite-treated A549 cells, based on SIRM analysis, which was used to interrogate parallel transcriptomic data for relevant gene expression changes. The signal log ratio of selenite treated to control mRNA expression is listed below each protein into which the mRNA was translated. Negative and positive values indicate a decrease and increase in expression, respectively. → indicates down regulation of genes, inhibition of protein activity, or depletion of ¹³C-labeled metabolites (denoted by *). → indicates up regulation of genes, activation of protein activity, or accumulation of labeled metabolites. → indicates trace reactions, transport processes, and changes in concentration of unlabeled metabolites. The open green arrows indicate glutaminolysis of unlabeled Gln taken up from the medium and subsequent oxidation to generate unlabeled malate. Dashed arrows indicate multiple reaction steps. Acetyl-CoA C: Acetyl-CoA carboxylase (ACC); AMPK: 5′-AMP-activated protein kinase; DAG: diacylglycerol; DHAP: dihydroxyacetone phosphate; FPBase: fructose bisphosphatase; Glnase IP: glutaminase interacting protein; G-6-P DH: glucose-6-phosphate dehydrogenase; GAP, glyceraldehyde-3-phosphate; GPC, glycerophosphorylcholine; α-KG DH: α-ketoglutarate dehydrogenase; LDH: lactate dehydrogenase; LysoPLase: lysophospholipase; ME: malic enzyme; Malate DH: malate dehydrogenase; N transporter: neutral amino acid transporter; OAA: oxaloacetate; PC: phosphorylcholine; PFK2: phosphofructokinase 2; PLases: phospholipases; Pyruvate DH: pyruvate dehydrogenase (adapted from [386] with permission). Selenite-reduced fatty acyl synthesis is evident in the 1D ¹³C projection spectrum of 2D ¹H{¹³C}-HSQC analysis of A549 lipids shown in (B). PCL: phosphatidylcholine lipids; PEL: phosphatidylethanolamine lipids. The Western blot analysis for the protein level of phosphorylated or inactivated ACC (pACC) and AMPK (pAMPK) in response to selenite treatment is shown in (C), where selenite-induced increase in pACC and decrease in pAMPK levels relative are evident. This data suggests inactivation of AMPK by phosphorylation underlies enhanced ACC phosphorylation or inactivation by AMP.

Fig. 18

2D ¹H{¹³C}-HSQC NMR analyses reveal in vivo metabolic reprogramming in human lung tumor tissues resected from an NSCLC patient infused with ¹³C₆-glucose (adapted from Fan et al. [97]). U-¹³C₆-glucose was infused into the patient 2–3 h prior to resection and cryopreservation of cancer and surrounding benign (normal) tissues. The liquid N₂-frozen tissues were pulverized and extracted for polar metabolites, which were analyzed by 2D ¹H{¹³C}-HSQC at 14.1 T. The 1D ¹³C projection spectrum is shown in (A), where the increase in ¹³C abundance of various metabolites at different carbon positions is evident in cancer versus paired benign tissues. The pathway scheme in (A) traces ¹³C atoms from ¹³C₆-glucose into intermediates of glycolysis, pentose phosphate pathway (PPP), the Krebs cycle, Ser-Gly pathway, and uracil biosynthesis. Based on this scheme and ¹³C abundance data in (A), the cancer tissue displays enhanced capacity for all of these pathways. ●: ¹²C; , : respective ¹³C tracing of pyruvate dehydrogenase (PDH) or pyruvate carboxylase (PCB)-initiated Krebs cycle reactions; Lac: lactate; GSH/GSSG: reduced/oxidized glutathiones; Glc: glucose; G6P: glucose-6-phosphate; UDPGGlcNAc: UDP-N-acetylglucosamine; UDPG: UDP-glucose; UXP: uracil nucleotides; AXP: adenine nucleotides; SHMT: serine hydroxymethyl transferas.

Fig. 19

NSCLC tissue slices maintain distinct metabolic reprogramming. Paired cancerous (CA) and non-cancerous (NC) lung tissues (A) freshly resected from an NSCLC patient were treated with U-¹³C₆-glucose for 24 h at 37 °C/5% CO₂. The tissue extracts obtained were analyzed by 1D ¹H{¹³C}-HSQC at 14.1 T (B). The elevated buildup of various ¹³C metabolites in CA versus NC tissues is consistent with enhanced glycolysis (cf. ¹³C-Lac), the Krebs cycle initiated by PDH (¹³C-4-Glu) or PCB (¹³C-2 or 3-Glu), onecarbon metabolism (¹³C-2-Gly), and pentose phosphate pathway/nucleotide biosynthesis (¹³C-1′-AXP, (¹³C-N1′ or A1′-NAD⁺), as depicted in ¹³C tracing of these pathways in (B). Such distinct metabolic reprogramming in CA lung tissue slices recapitulates that of human lung cancer tissues in vivo (cf. Fig. V.B.2). See Fig. V.B.2 for abbreviations.

Fig. 20

In vivo 2D ¹H{¹³C}ge-HMQC analysis of cat brain. The ¹³C-edited 2D ¹H spectrum acquired in vivo from a cat brain resolves resonances of ¹³C-4-Glu (4a) from ¹³C-4-Gln (4b), which cannot be achieved with 1D ¹H methods. Also resolved are resonances of ¹³C-3-Glu/Gln (5), ¹³C-2-Glu/Gln (3), ¹³C-1-α-glucose, and ¹³C-1-β-glucose; the latter three would be difficult to detect without the better water suppression afforded by the 2D method [260]. Excerpted from MAGMA, 17 (2004) 317–338, In vivo 2D magnetic resonance spectroscopy of small animals, P. Meric, G. Autret, B.T. Doan, B. Gillet, C. Sebrie, J.C. Beloeil, Fig. 14a with kind permission from Springer Science and Business Media.

Fig. 21

In vivo 1D ¹³C NMR analysis of human brain metabolism. The ¹³C spectra in (A) were acquired in vivo from the midline occipital/parietal lobe of a human volunteer infused with ¹³C₁-1-glucose, ¹³C₁-3-lactate, or ¹³C₁-2-acetate. The differences in ¹³C incorporation from these tracer substrates into Glu and Gln are evident. With glucose and lactate as precursors, a major fraction of ¹³C appeared in Glu C4, which is consistent with primary location of glucose and lactate metabolism in brain neurons. In contrast, the acetate tracer led to much higher ¹³C enrichment in Gln C4, which supports acetate metabolism via the Krebs cycle in brain astrocytes. Based on the time course analysis of in vivo ¹³C NMR data such as those in (A), the rate of neuronal Krebs cycle (V_TCAn) and Glu/Gln cycle (V_cyc) can be estimated as a function of brain electrical activity. Panel (B) plots V_TCAn versus V_cyc from 11 studies and shows a 1:1 linear relationship. These results indicate that the Glu/Gln cycle for the Gln neurotransmitter production constitutes a major metabolic flux for brain energy metabolism and that >80% of neuronal oxidative ATP production is coupled to neuronal signaling [262]. Reproduced from Figs. 2 and 3 of D.L. Rothman, H.M. De Feyter, R.A. de Graaf, G.F. Mason, K.L. Behar, ¹³C MRS studies of neuroenergetics and neurotransmitter cycling in humans, NMR in biomedicine, 24 (2011) 943–957 with permission from John Wiley & Sons Inc.

Fig. 22

Two site exchange. Magnetization time courses were calculated for two-site exchange A ⇔ B with k₁ = k₋₁ = 0.05 s⁻¹ and T₁ values for A and B of 20 s. Blue squares = M_A; red circles = M_B. The equilibrium and initial magnetizations of A were set to unit and 10,000, respectively. The modified Bloch equations (3A) and (3B) were solved as described in the text, and the curves were generated using Kaleidagraph (Synergy Software). The solutions to this system are: M_B(t) = k₁〈M_A〉[exp(λ₁t) − exp (λ₂t)]/(λ₁ − λ₂). $M_{\mathrm{A}}(t)=[(\epsilon -{\lambda }_2(M_{\mathrm{A}}^0-\langle M_{\mathrm{A}}\rangle ))exp({\lambda }_{1}t)+({\lambda }_1(M_{\mathrm{A}}^0-\langle M_{\mathrm{A}}\rangle -\epsilon )\text{exp}({\lambda }_{2}t)]/({\lambda }_1-{\lambda }_2)$. 2λ_1,2 = −ρ_A + ρ_B + k₁ + k₋₁ ± [(ρ_A + ρ_B + k₁ + k₋₁)² − 4(ρ_Aρ_B + ρ_Ak₁ + ρ_Bk₋₁)]. $\epsilon =-({\rho }_{\mathrm{A}}+k_1)M_{\mathrm{A}}^0+k_{-1}\langle M_{\mathrm{B}}\rangle +{\rho }_{\mathrm{A}}\langle M\rangle$. where $M_{\mathrm{A}}^0$ is the initial (hyperpolarized) magnetization of A.

Fig. 23

Central Metabolism probed by hyperpolarized ¹³C enriched precursors. Metabolites in red have been ¹³C hyperpolarized at one or more atoms, and those in italics have been observed as products of hyperpolarized substrates via different pathways. These include glucose (Glc) oxidation to pyruvate (Pyr) via glycolysis, subsequent pyruvate metabolism via the Krebs cycle and lactic fermentation, anaplerosis via pyruvate carboxylation to oxaloacetate (OAA), and glucose metabolism via the pentose phosphate pathway from glucose-6-phosphate (G6P), 6-phosphogluconate (6PG) to ribulose-5-phosphate (Ru5P). Also shown are anaplerosis via glutaminolysis or Gln conversion to Glu and then to α-ketoglutarate (αKG) for entry into the Krebs cycle and ethanol oxidation to acetate via alcohol dehydrogenase (ADH) and aldehyde dehydrogenase (ALDH). DHAP: dihydroxyacetone phosphate; GAP: glyceraldehyde-3-phosphate; 3PGA: 3-phosphoglycerate; PEP: phosphoenolpyruvate; Lac: lactate, Succ succinate; DES: diethylsuccinate; fum: fumarate; mal: malate; HK: hexokinase; LDH: lactate dehydrogenase; ALT: alanine pyruvate aminotransferase; PDH: pyruvate dehydrogenase; SDH: succinate dehydrogenase; FH: fumarate hydratase; MDH: malate dehydrogenase; AST: aspartate/oxalacetate aminotransferase; PC: pyruvate carboxylase; GLS glutaminase.

Fig. 24

Magnetic resonance imaging of three co-hyperpolarized ¹³C agents in mouse brain. ¹³C-Urea, ¹³C-1-hydroxymethylcyclopropane (HMCP), and ¹³C-2-t-butanol were simultaneously hyperpolarized before infusing into normal and transgenic TRAMP mice bearing prostate cancer. A multiband frequency encoding and balanced steady-state free precession (SSFP) MRI method was used to acquire ¹³C images of the dynamic distributions of these three agents in mouse tissues. The axial tripolarized images of mouse brain (color) are overlaid on the T₂ images (gray), which show the remarkable speed at which t-butanol crosses the blood–brain barrier, unlike urea and HMCP A parallel set of dynamic images of prostate tumor tissues compared with those of normal tissues enabled measurement of tissue permeability for the three agents. Tumor tissues were shown to have a higher permeability to urea and HMCP than normal brain and liver. Such a capability can be readily extended to investigating the efficiency of therapeutic drug delivery to target tumor tissues [343]. Reproduced from Fig. 5 in C. von Morze, R.A. Bok, G.D. Reed, J.H. Ardenkjaer-Larsen, J. Kurhanewicz, D.B. Vigneron, Simultaneous multiagent hyperpolarized (13)C perfusion imaging, Magn Reson Med, 72 (2014) 1599–1609. With permission from John Wiley & Sons, Inc.