doi.org/10.26434/chemrxiv.7771142.v1

A Fast and Sensitive Method Combining Reversed-PhaseChromatography with High Resolution Mass Spectrometry to Quantify2-Fluoro-2-Deoxyglucose and Its Phosphorylated Metabolite forDetermining Glucose UptakeAshley Williams, Deborah Muoio, Guofang Zhang

Submitted date: 26/02/2019 • Posted date: 27/02/2019Licence: CC BY-NC-ND 4.0Citation information: Williams, Ashley; Muoio, Deborah; Zhang, Guofang (2019): A Fast and Sensitive MethodCombining Reversed-Phase Chromatography with High Resolution Mass Spectrometry to Quantify2-Fluoro-2-Deoxyglucose and Its Phosphorylated Metabolite for Determining Glucose Uptake. ChemRxiv.Preprint.

Quantative measurements of the glucose analogue, 2-deoxyglucose (2DG), and its phosphorylated metabolite(2-deoxyglucose-6-phosphate (2DG-6-P)) are critical for the measurement of glucose uptake. While the fieldhas long identified the need for sensitive and reliable assays that deploy non-radiolabled glucose analogues toassess glucose uptake, no analytical MS-based methods exist to detect trace amounts in complex biologicalsamples. In the present work, we show that 2DG is poorly suited for MS-based methods due to interferingmetabolites. We therefore developed and validated an alternative C18-based LC-Q-Exactive-Orbitrap-MSmethod using 2-fluoro-2-deoxyglucose (2FDG) to quantify both 2FDG and 2FDG-6-P by measuring thesodium adduct of 2FDG in the positive mode and deprotonation of 2FDG-6-P in the negative mode. The lowdetection limit of this method can reach 81.4 and 48.8 fmol for both 2FDG and 2FDG-6-P, respectively. Thenewly developed method was fully validated via calibration curves in the presence and absence of biologicalmatrix. The present work is the first successful LC-MS method that can quantify trace amounts of anonradiolabeled glucose analogue and its phosphorylated metabolite and is a promising analytical method todetermine glucose uptake in biological samples.

File list (1)

download fileview on ChemRxivWilliams et al_2DG_2019.v2.pdf (1.64 MiB)

1

A fast and sensitive method combining reversed-phase chromatography with

high resolution mass spectrometry to quantify 2-fluoro-2-deoxyglucose and its

phosphorylated metabolite for determining glucose uptake

Ashley S. Williams1, Deborah M. Muoio1,2,3, and Guo-Fang Zhang1,2*

1Duke Molecular Physiology Institute and Sarah W. Stedman Nutrition and Metabolism Center, Duke University

Medical Center, Durham, NC 27701, USA

2Department of Medicine, Division of Endocrinology, Metabolism, and Nutrition, Duke University Medical Center,

Durham, NC 27710, USA

3Department of Pharmacology and Cancer Biology

*To whom correspondence should be addressed: Guo-Fang Zhang, guofang.zhang@duke.edu, Duke Molecular

Physiology Institute, 300 N Duke St, Durham, NC 27701

Lead contact: Guo-Fang Zhang

2

ABSTRACT

Quantative measurements of the glucose analogue, 2-deoxyglucose (2DG), and its phosphorylated metabolite

(2-deoxyglucose-6-phosphate (2DG-6-P)) are critical for the measurement of glucose uptake. While the field has

long identified the need for sensitive and reliable assays that deploy non-radiolabled glucose analogues to

assess glucose uptake, no analytical mass spectrometry (MS) - based methods exist to detect trace amounts in

complex biological samples. In the present work, we show that 2DG is poorly suited for MS-based methods due

to interfering metabolites. We therefore developed and validated an alternative C18-based LC-Q-Exactive-

Orbitrap-MS method using 2-fluoro-2-deoxyglucose (2FDG) to quantify both 2FDG and 2FDG-6-P by measuring

the sodium adduct of 2FDG in the positive mode and deprotonation of 2FDG-6-P in the negative mode. The low

detection limit of this method can reach 81.4 and 48.8 fmol for both 2FDG and 2FDG-6-P, respectively. The

newly developed method was fully validated via calibration curves in the presence and absence of biological

matrix. The present work is the first successful LC-MS method that can quantify trace amounts of a

nonradiolabeled glucose analogue and its phosphorylated metabolite and is a promising analytical method to

determine glucose uptake in biological samples.

Keywords: 2-deoxyglucose, 2-deoxyglucose-6-phosphate, sodium adduct, glucose uptake, LC-MS

3

INTRODUCTION

The gold standard for the determination of glucose uptake is the 2-deoxyglucose (2DG) method. 2DG is

a glucose analogue that lacks one hydroxyl group at C-2 (Scheme 1). Upon entering the cell, 2DG is rapidly

phosphorylated by hexokinase to form 2-deoxyglucose-6-phosphate (2DG-6-P), which is then trapped in the

tissue because the lack of a keto group on C-2 of 2DG prevents the entry of 2DG-6-P into glycolysis (1). A small

fraction of 2DG-6-P can be incorporated into glycogen or dephosphorylated back to 2DG, depending on tissue

type (1-3), although this appears negligible in tissues such as muscle and heart. Therefore, concentrations of

2DG-6-P in tissues or cells can be used as an index of glucose uptake in most of organs, except those with high

activity of glucose-6-phosphatase (G6Pase), such as the liver, kidney, and intestine.

For applications in rodent models, the 2DG method involves bolus administration of trace amounts of

2DG (13uCi 14C-2DG or ~227nmol) followed by serial collection of blood samples to determine rates of 2DG

disappearance from the plasma. The plasma 2DG decay curve is coupled with tissue measurement of 2DG-6-P

to calculate glucose uptake, or more specifically, the glucose metabolic index (Rg) (4). A critical, but often

overlooked, consideration for these assays is that high doses of tracer can inhibit hexokinase in the brain and

perturb whole body glucose metabolism (5). Therefore, determination of glucose uptake in vivo requires a

sensitive and reliable method to quantify trace (nanomolar) amounts of both 2DG and 2DG-6-P in plasma and

tissues.

Measurement of 2DG and 2DG-6-P can be achieved by a number of various techniques, such as NMR

(6), enzymatic assays (7-9), and scintillation counting of radiolabeled metabolites (2,10-14). While each of the

foregoing methods has advantages, notable drawbacks include the requirement for sophisticated

instrumentation and expertise, lack of sensitivity, and logistical and administrative constraints, respectively. By

contrast, a MS-based method would alleviate most of these concerns. However, to our knowledge, an MS-

based assay has not been developed, due in large part to the inherently difficulty of measuring both 2DG and

2DG-6-P using a single method. Moreover, each compound presents its own set of challenges. 2DG-6-P, when

compared to 2DG, is very difficult to quantify using GC-MS because derivatization with TMS or TBDMS causes

partial phosphoryl group hydrolysis and, once derivitized, 2DG-6-P is a relatively large molecule that may

degrade in the GC-MS column (15,16). Despite this challenge, a method to quantify 2DG and 2DG-6-P via GC-

MS has been described; however, this approach requires the conversion of 2DG-6-P to 2DG prior to analysis,

which is less than ideal (17). Second, 2DG-6-P readily forms a negative ion in the electrospray ionization (ESI)

source that can be fragmented in tandem mass spectrometry. Therefore, although 2DG-6-P is a very polar

compound, it is best suited for a LC-MS/MS method run in the negative mode. To this point, a previous study

successfully quantified 2DG-6-P by LC-MS/MS with a linear range from 8 to 30 µM (18), yet no LC-MS method

is available to quantify 2DG, presumably as a consequence of its polarity and poor ionization in the ESI source.

In the current study, we sought to develop and validate a MS-based method to detect trace amounts of

2DG and 2DG-6-P. To this end, we investigated 2DG and 2DG-6-P ionization in the ESI source and its potential

application for quantitation by LC-MS. We found that while 2DG and 2DG-6-P can be measured via LC-MS, the

4

presence of interfering metabolites around the m/z for each compound decreases the accuracy and sensitivity

of the method. To circumvent this problem, we developed a novel LC-MS high resolution Orbitrap MS-based

method to measure 2FDG and 2FDG-6-P without derivatization or extensive sample preparation. The

replacement of 2DG and 2DG-6-P by 2FDG and 2FDG-6-P improved method selectivity and the newly

developed method was validated via calibration curves in distilled water and biological matrix.

MATERIALS AND METHODS

Chemicals and reagents

2DG was purchased from Chem-impex Int'l Inc (Wood Dale, IL). 2-DG-6-P was purchased from Santa Cruz

Biotechnology (Dallas, Texas). 2FDG-6-P was purchased from Omicron Biochemicals Inc. (Southbend, IN).

2FDG, 2ClDG, hexokinase from baker yeast, triethanolamine, formic acid, chloroform, adenosine triphosphate

(ATP), ethylenediaminetetraacetic acid disodium salt (EDTA), and other chemical reagents in analytical grade

or above were purchased from Sigma (St. Louis, MO).

Preparation of 2ClDG-6-P

To obtain 2ClDG-6-P, 150 ul 150 ul of 100 mM 2ClDG, 180 ul of 100 mM ATP, 180 ul of 100 mM MgCl2, 75 ul

of 100 mM EDTA, and 10 U hexokinase were added into 1 ml triethanolamine buffer (50 mM, pH=7.4). The

reaction was maintained at 37C for 30 minutes. The reaction was quenched by adding 1 ml of precooled

methanol. After centrifugation at 800 × g for 15 minutes, the supernatant was dried and dissolved in distilled

water. The concentration of 2ClDG-6-P was estimated based on 2DG-6-P. 2ClDG-6-P was used as the internal

standard (IS) for the external calibration curve, therefore, the absolute concentration was not required.

LC-QTOF-MS instrumentation and conditions

An Agilent 1200 HPLC connected to an Agilent 6520 QTOF mass spectrometer was employed for method

optimization including the LC method optimization, ionization, and fragmentation tests. HPLC 1200 was

configured with an isocratic pump, a binary pump, a thermostatted column compartment, and a temperature-

controlled autosampler. The isocratic pump delivered reference solution (5 µM purine and 1.5 µM HP-0921 in

methanol solution containing 5% water.) to the mass spectrometer at a flow rate of 0.5 ml/min with a split of

1/100. The binary pump was used to transport mobile phase (HPLC-MS grade water) at a flow rate of 0.5 ml/min

in isocratic elution mode. The column was a Microsorb-MV C18 column (100 × 4.6 mm, 3 µm) with C18 guard

column and was kept at 40 C in the oven compartment. The autosampler was maintained at 5C, and the

injection volume was 5 µl. The total running time is 7 minutes.

QTOF-MS was operated at electrospray mode. For 2-DG and glucose, QTOF was run in the positive and MS

scan mode with the following parameters: gas temperature, 350C; drying gas, 11 l/min; nebulizer, 50 psig; VCap,

3500V; fragmentor, 130V; skimmer, 65V; Oct 1 RF Vpp, 750 V; collision energy, 0V; mass scan range, 100 –

5

1700. 2DG-6-P and G-6-P were detected in negative mode with same parameters except that ion polarity was

negative.

LC- Q Exactive Hybrid Quadrupole-Orbitrap Mass Spectrometer

LC-Q-Exactive+-Orbitrap-MS was used for final quantitation. The LC method was adopted from the previous LC-

QTOF-MS method. The parameters for Q-Exactive+-MS equipped with a HESI probe: heat temperature: 425 °C;

sheath gas: 30, auxiliary gas, 13; sweep gas, 3; spray voltage, 3.5 kV for positive mode; capillary temperature

was set at 320 °C, and S-lens was 45. A full scan range was set at 60 to 900 (m/z). The resolution was set at 70

000 (at m/z 200). The maximum injection time (max IT) was 200 ms. Automated gain control (AGC) was targeted

at 3 × 106 ions.

Sample preparation for blank plasma 2FDG calibration curves

Two ul of plasma were added to a tube prior to folch extraction using the following solvents: 200 µl methanol,

200 µl distilled H2O, and 200 µl chloroform. The sample mixture was vortexed and centrifuged for 20 minutes at

14,000g at 4°C. The upper phase (~350 µl) containing the extracted matrix from several preparations (~15) was

combined into a single tube. Approximately 350 µl of the upper phase, 50 µl of each 2FDG standard solution,

and 1 nmol 2ClDG (IS) were combined into a tube, vortexed, and dried completely by nitrogen gas at 37C. The

dried residue was resupended in 60 µl distilled water, vortexed, and placed in an autosampler vial for LC-MS

analysis.

Sample preparation for blank tissue 2FDG-6-P calibration curves

Unlabeled mouse skeletal muscle tissue was pulverized in liquid nitrogen. Approximately 10 mg powdered tissue

was weighed prior to folch extraction using the following solvents: 400 μl ice cold methanol, 400 μl chloroform,

and 400 μl distilled water using a Qiagen Tissuelyzer (30 Hz, 1 minute per solvent). After homogenization, the

sample mixture was centrifuged for 20 minutes at 14,000 ×g at 4°C. The upper phase (~750 µl) containing the

extracted matrix from several preparations (~15) was combined into a single tube. Approximately 750 µl of the

upper phase, 50 µl of each 2FDG-6-P standard solution, and 1 nmol 2ClDG (IS) were combined into a tube,

vortexed, and dried completely by nitrogen gas at 37C. The dried residue was resupended in 100 µl distilled

water, vortexed, and placed in an autosampler vial for LC-MS analysis.

Method validation

The linear range, limit of detection (LOD), limit of quantitation (LOQ) were assessed by a serial dilution of

standard stock solutions. The accuracy was assessed by measuring the recovery of low (81.4 fmol for 2FDG

and 48.8 fmol for 2FDG-6-P), middle (1.3 pmol for 2FDG and 1.6 pmol for 2FDG-6-P), and high amounts (10.4

pmol for 2FDG and 50 pmol for 2FDG6P) of standards spiked into blank plasma or blank skeletal muscle samples.

The recovery experiment was repeated over 3 different days.

6

Data analysis

Mass spectrometric data was acquired by MassHunter Data Acquisition (B.06.01) software, MassHunter

Quantitative Analysis for QTOF (B.07.01) software, and Thermo Xcalibur software. Analyte concentrations were

calculated using a 1/x weighted linear regression analysis of the standard curve.

RESULTS AND DISCUSSION

Ionization and fragmentation of 2DG and 2DG-6-P

The goal of this study was to develop a LC-MS or LC-MS/MS method, therefore we started by injecting 2DG and

2DG-6-P into the QTOF to check the ionization. From the mass spectra of the two compounds (Figure 1), we

could not find the protonated 2DG (2DG-6-P) whose m/z is expected to be 165.0757 (C6H13O5) (Figure 1A).

Instead we identified a strong sodium adduct of 2DG at m/z 187.0580 (theoretical m/z of C6H12O5Na is

187.0577) (Figure 1A). As anticipated, the ionization of 2DG-6-P was a strong negative ion at m/z = 243.0310

(Figure 1B).

Next we investigated the fragmentation of 2DG and 2DG-6-P to determine if a MS/MS method could be

developed using the triple quadrupole (Q-TOF) mass spectrometer. In order to identify the fragments, we ran

the fragmentation with 2DG and 2DG-6-P with the increasing collision energy (CE) from 10 to 40 V. 2DG-6-P

was easily fragmented even with 10 V of CE (Figure 2A). Two fragments at m/z 96.709 and m/z 78.9607 were

likely to be [phosphoric acid]- and [phosphoric acid-H2O]-. With the increasing CE, [phosphoric acid-H2O]- became

the dominant fragment (Figures 2B-D). However, 2DG remained largely intact as a sodium adduct ion at

CE=10V (Figure 3A). Thus we increased the CE and found with increasing CE, 2DG was fragmented into

several small fragments (Figures 3B-D), particularly at CE=40V (Figure 3D). At high collision energy, 2DG was

fragmented at multiple locations and there was no a major fragment that could be used for tandem mass

spectrometry. This is supported by previous work on sodium adducts showing they are difficult to measure due

to insufficient fragmentation and poor reproducibility (19). Thus, for the remainder of the study, we focused our

efforts on the development of a LC-MS method to determine both [2DG-Na]+ and [2DG-6-P-H]-. For all

subsequent assays, mass spectrometry was run in the positive mode for the [2DG-Na]+ assay and in the negative

mode for the [2DG-6-P-H]- assay.

HPLC method development

Sugar and phosphorylated sugars are very polar compounds and this creates several challenges for LC-MS

method development including: (i) poor retention on reversed-phase columns which impedes method

development and reproducibility and (ii) peak tailing for phosphorylated sugars due to the metal tubing system.

Therefore, we tested different reversed-phase columns and mobile phases. A Microsorb-MV C18 column (100

× 4.6 mm, 3 µm) was found to have very good retention of both sugars and phosphorylated sugars. Organic

solvents (methanol and acetonitrile) in the mobile phase worsen peak tailing, thus we selected an isocratic mobile

7

phase containing 98% H2O, 2% methanol and 0.01% formic acid based on optimization to ensure: (i) low peak

tailing, (ii) reproducible retention time of biological samples, and (iii) long term use of the column. Using these

conditions, 2DG and 2DG-6-P were well separated, and all of them came out before 3 minutes (data not shown)

and the total running time was 10 minutes per sample.

Sample preparation

Since the present work employed an isocratic LC method with aqueous as mobile phase, it was necessary to

remove all lipid during the sample preparation to avoid altering the LC column’s performance by lipid

accumulation. To this point, we employed a Folch wash approach for both tissue and plasma sample preparation.

In addition, a 1-µl injection volume was used to decrease the amount of biological matrix loaded onto the column.

The effect of sodium ion concentration on the relative abundance of [2DG-Na]+ in ESI

The lack of protonated 2DG in our Q-TOF extracted ion chromatograms and the presence of the sodium adduct

is supported by the phenomenon that neighboring hydroxy groups can form a stable triangular cyclic ion after

binding with one metal ion. In addition, glucose possesses a strong affinity to sodium and other metal ions in the

ESI source (20) and the formation efficiency of sodium adducts is higher for oxygen bases. Glucose and 2DG

have 6 and 5 oxygen atoms, respectively, and the strong sodium affinity indicated that the 2DG sodium adduct

could be used for quantitation since a glucose or 2DG sodium adduct would be more resistant to ionization

condition changes. Thus we hypothesized that a MS-based method for the quantitation of a sodium adduct is

required although it is highly unusual due to the difficulty of fragmentation and delicate ionization conditions (21).

To determine the effect of NaCl concentration on the ionization of 2DG and glucose, the relative abundance of

2DG and glucose was assessed in the presence of increasing concentrations of NaCl (Figure 4). Ionization of

2DG progressively increased with [NaCl] from 0 to 10 mM and slightly dropped at 100 mM NaCl. Notably,

intensity of the glucose sodium adduct remained constant from 0 to 10 mM NaCl and, similar to 2DG, 100 mM

NaCl suppressed the ionization of glucose presumably due to suppression of the ionization source. In sum, these

results show that NaCl concentration differently affects the relative abundance of 2DG and glucose using ESI. It

is plausible that differences between glucose and 2DG exist as (1) the loss of one –OH from 2DG could decrease

the sodium affinity compared to glucose and (2) formation of the glucose sodium adduct is maximized even at

trace amounts of sodium ion.

Endogenous interference in biological samples and the selection of a monohalogenated 2DG compound

Although the high resolution Orbitrap Q Exactive MS dramatically improves differentiation and separation of

several isobaric compounds, interference from complicated biological matrices remain a challenge due to the

presence of structural isomers and isotopologues. To assess whether unlabeled 2DG could be utilized as a

metabolic tracer in rodents, we sought to determine if there were any potential interfering metabolites around the

m/z of 2DG in untreated (a.k.a. blank) mouse plasma. As shown in Figure 5A, the extracted chromatogram of

[2DG-Na]+ (m/z = 187.0577) from a blank plasma sample demonstrates the existence of a highly expressed,

8

interfering metabolite with a very similar m/z and retention time that would impede method sensitivity. Next we

investigated whether other monohalogenated glucose analogues with different m/z, such as 2FDG and 2ClDG,

could be used as a tracer and internal standard, respectively. Similar to 2DG, 2FDG (or [18F]FDG) is an

established glucose analogue which is often used to in humans and rodents to determine glucose uptake via

Positron Emission Tomography (PET) (22-25). We extracted the ion chromatograms of 2FDG and 2ClDG from

blank plasma samples (Figures 5B and 5D) and established that no interference was observed at a similar

retention time when the m/z of 2FDG or 2ClDG was extracted. 2FDG or 2ClDG were only present in the extracted

chromatogram when blank plasma was spiked with 2FDG (Figure 5C) or 2ClDG (Figure 5E). Notably, similar

interference was also observed for 2DG-6-P (m/z = 243.0264) in blank, untreated skeletal muscle (Figure 6A).

2FDG-6-P (m/z = 261.0180) or 2ClDG-6-P (m/z = 276.9885) was not found in the blank skeletal muscle sample

(Figures 6B and 6D) except when the blank skeletal muscle sample was spiked with 2FDG-6-P (Figure 6C) or

2ClDG-6-P (Figure 6E). While an ideal internal stand for quantitation is the stable isotope labeled compound,

we chose 2ClDG (and its phosphorylated metabolite, 2ClDG-6-P) as our internal standard(s) because stable

isotope labeled 2DG ([1-13C]2DG) and 2DG-6-P ([1-13C]2DG-6-P) are also subjected to the same interference

from the M+1 isotopologue of the endogenous compound at m/z 187.0577 and cannot be used as internal

standards. 2ClDG and 2ClDG-6-P do not have interfering metabolites in plasma and muscle tissue, therefore

they were chosen as internal standards for quantifying 2FDG and 2FDG-6-P, respectively.

Method validation and mass accuracy

After initial method optimization, all method validation was conducted using our UPLC-Q-Exactive-MS+ platform

which has the merit of high sensitivity and mass resolution. The high resolution of Q-Exactive-MS+ mass

spectrometry improves the selectivity of assay and decreases the baseline of mass scan. The measured m/z

values of analytes and internal standards in Q-Exactive-MS+ were compared to their theoretical values. The

detailed data is shown in Table 1. The mass accuracy were within 2.7 ppm among all analytes and internal

standards (see Table 1).

Linearity, detection and quantification limits for the method

Next we ran an external calibration curve to check the sensitivity and linearity of the method. Figure 7 shows

the chromatogram of 11 different concentrations of 2FDG (Figure 7A) and 12 different concentrations of 2FDG-

6-P (Figure 7C) with equal concentrations of the internal standards (2ClDG or 2ClDG-6-P) (Figures 7C and

7D). The peak shapes of 2FDG (Figure 7A), 2FDG-6-P (Figure 7C), and 2ClDG-6-P (Figure 7D) were sharp

(peak width < 0.2 minutes) and symmetric. Only 2ClDG (Figure 7B) showed a slight peak split which might be

caused by the spatial isomerism of Cl. However, the spatial isomerism of Cl in the relatively larger molecule,

2ClDG-6-P, is minimized and comes out as a single peak (Figure 7D).

To determine whether biological matrix affects the slope of the calibration curves, calibration curves of 2FDG

and 2FDG-6-P were prepared in distilled water and compared to those prepared in blank plasma/skeletal muscle

9

extracts (Figure 8). Skeletal muscle extract had little effect on the calibration curve of 2FDG-6-P. In contrast,

plasma matrix increased the slope of calibration curve of 2FDG compared to the one generated in distilled water.

Although biological matrix had little effect on 2FDG-6-P calibration curve, subsequent calibration curves were

generated in corresponding biological matrix and the performance of the analytical method was further assessed

by investigating the linearity, slope, limit of detection (LOD), and limit of quantitation (LOQ) (Table 2). The

accuracy and reproducibility of the method was determined using three different concentrations of analytes from

the calibration curves (low, middle, and high) assayed over three different days (Table 3). Notably, no changes

were observed for either compound (2FDG and 2FDG-6-P) after storage for 2 weeks in the 4C LC sampler. As

demonstrated in Tables 2 and 3, the present method for 2FDG and 2FDG-6-P quantitation is sensitive, robust,

accurate, and reproducible.

CONCLUSIONS

The 2DG method is the gold-standard for the determination of glucose uptake. However, a single MS-

based method to measure both 2DG and 2DG-6-P does not exist. To fill this gap, we developed and validated a

simple, fast, and sensitive C18-based LC-Q-Exactive-Orbitrap-MS method to quantify both 2DG and 2DG-6-P

using a monohalogenated form of 2DG, 2FDG. Our method is unique because it leverages the inherent nature

of glucose analogues, such as 2FDG, to form sodium adducts in combination with the high resolution and

sensitivity of the Orbitrap Q-Exactive to reliably quantify trace amounts of 2FDG and 2FDG-6-P.

ACKNOWLEDGEMENTS

Financial support for this work was provided by the NIDDK Mouse Metabolic Phenotyping Centers (National

MMPC, RRID: SCR_008997, www.mmpc.org) under the MICROMouse program, grants DK076169 and National

Institutes of Health grants F32DK105922 (ASW) and R01DK089312 (DMM).

10

REFERENCES

1. Virkamaki, A., Rissanen, E., Hamalainen, S., Utriainen, T., and Yki-Jarvinen, H. (1997) Incorporation of [3-3H]glucose and 2-[1-14C]deoxyglucose into glycogen in heart and skeletal muscle in vivo: implications for the quantitation of tissue glucose uptake. Diabetes 46, 1106-1110

2. Doenst, T., and Taegtmeyer, H. (1998) Complexities underlying the quantitative determination of myocardial glucose uptake with 2-deoxyglucose. J. Mol. Cell Cardiol 30, 1595-1604

3. Sweet, I. R., Najafi, H., Li, G., Grodberg, J., and Matschinsky, F. M. (1997) Measurement and modeling of glucose-6-phosphatase in pancreatic islets. Am. J. Physiol 272, E696-E711

4. Kraegen, E. W., James, D. E., Jenkins, A. B., and Chisholm, D. J. (1985) Dose-response curves for in vivo insulin sensitivity in individual tissues in rats. Am. J. Physiol. 248, E353-362

5. Borg, W. P., Sherwin, R. S., During, M. J., Borg, M. A., and Shulman, G. I. (1995) Local ventromedial hypothalamus glucopenia triggers counterregulatory hormone release. Diabetes 44, 180-184

6. Hopkins, J. C., Radda, G. K., Veech, R. L., and Clarke, K. (2004) Accumulation of 2-deoxy-D-glucose-6-phosphate as a measure of glucose uptake in the isolated perfused heart: a 31P NMR study. Metab Eng 6, 36-43

7. Sasson, S., Oron, R., and Cerasi, E. (1993) Enzymatic assay of 2-deoxyglucose 6-phosphate for assessing hexose uptake rates in cultured cells. Anal. Biochem 215, 309-311

8. Yamamoto, N., Kawasaki, K., Sato, T., Hirose, Y., and Muroyama, K. (2008) A nonradioisotope, enzymatic microplate assay for in vivo evaluation of 2-deoxyglucose uptake in muscle tissue. Anal Biochem 375, 397-399

9. Tong, H., Chen, W., London, R. E., Murphy, E., and Steenbergen, C. (2000) Preconditioning enhanced glucose uptake is mediated by p38 MAP kinase not by phosphatidylinositol 3-kinase. J Biol Chem 275, 11981-11986

10. Thompson, A. L., Lim-Fraser, M. Y., Kraegen, E. W., and Cooney, G. J. (2000) Effects of individual fatty acids on glucose uptake and glycogen synthesis in soleus muscle in vitro. Am J Physiol Endocrinol Metab 279, E577-584

11. Halseth, A. E., Bracy, D. P., and Wasserman, D. H. (1999) Overexpression of hexokinase II increases insulinand exercise-stimulated muscle glucose uptake in vivo. Am J Physiol 276, E70-77

12. Manchester, J., Kong, X., Nerbonne, J., Lowry, O. H., and Lawrence, J. C., Jr. (1994) Glucose transport and phosphorylation in single cardiac myocytes: rate-limiting steps in glucose metabolism. Am J Physiol 266, E326-333

13. Hansen, P. A., Gulve, E. A., and Holloszy, J. O. (1994) Suitability of 2-deoxyglucose for in vitro measurement of glucose transport activity in skeletal muscle. J Appl Physiol (1985) 76, 979-985

14. Chi, M. M., Manchester, J. K., Carter, J. G., Pusateri, M. E., McDougal, D. B., and Lowry, O. H. (1993) A refinement of the Akabayashi-Saito-Kato modification of the enzymatic methods for 2-deoxyglucose and 2-deoxyglucose 6-phosphate. Anal Biochem 209, 335-338

15. Yang, L., Kombu, R. S., Kasumov, T., Zhu, S. H., Cendrowski, A. V., David, F., Anderson, V. E., Kelleher, J. K., and Brunengraber, H. (2008) Metabolomic and mass isotopomer analysis of liver gluconeogenesis and citric acid cycle. I. Interrelation between gluconeogenesis and cataplerosis; formation of methoxamates from aminooxyacetate and ketoacids. J Biol Chem 283, 21978-21987

16. Yang, L., Kasumov, T., Yu, L., Jobbins, K. A., David, F., Previs, S. F., Kelleher, J. K., and Brunengraber, H. (2006) Metabolomic assays of the concentration and mass isotopomer distribution of gluconeogenic and citric acid cycle intermediates. Metabolomics 2, 85-94

17. Cuthbertson, D. J., Babraj, J. A., Mustard, K. J., Towler, M. C., Green, K. A., Wackerhage, H., Leese, G. P., Baar, K., Thomason-Hughes, M., Sutherland, C., Hardie, D. G., and Rennie, M. J. (2007) 5-aminoimidazole-4-carboxamide 1-beta-D-ribofuranoside acutely stimulates skeletal muscle 2-deoxyglucose uptake in healthy men. Diabetes 56, 2078-2084

18. az-Moralli, S., Ramos-Montoya, A., Marin, S., Fernandez-Alvarez, A., Casado, M., and Cascante, M. (2012) Target metabolomics revealed complementary roles of hexose- and pentose-phosphates in the regulation of carbohydrate-dependent gene expression. Am. J. Physiol Endocrinol. Metab 303, E234-E242

19. Lopes, N. P., Stark, C. B., Gates, P. J., and Staunton, J. (2002) Fragmentation studies on monensin A by sequential electrospray mass spectrometry. Analyst 127, 503-506

11

20. McIntosh, T. S., Davis, H. M., and Matthews, D. E. (2002) A liquid chromatography-mass spectrometry method to measure stable isotopic tracer enrichments of glycerol and glucose in human serum. Anal Biochem 300, 163-169

21. Kruve, A., Kaupmees, K., Liigand, J., Oss, M., and Leito, I. (2013) Sodium adduct formation efficiency in ESI source. J Mass Spectrom 48, 695-702

22. Sprinz, C., Altmayer, S., Zanon, M., Watte, G., Irion, K., Marchiori, E., and Hochhegger, B. (2018) Effects of blood glucose level on 18F-FDG uptake for PET/CT in normal organs: A systematic review. PLoS One 13, e0193140

23. Aoyama, R., Takano, H., Kobayashi, Y., Kitamura, M., Asai, K., Amano, Y., Kumita, S. I., and Shimizu, W. (2017) Evaluation of myocardial glucose metabolism in hypertrophic cardiomyopathy using 18F-fluorodeoxyglucose positron emission tomography. PLoS One 12, e0188479

24. Broecker-Preuss, M., Becher-Boveleth, N., Bockisch, A., Duhrsen, U., and Muller, S. (2017) Regulation of glucose uptake in lymphoma cell lines by c-MYC- and PI3K-dependent signaling pathways and impact of glycolytic pathways on cell viability. J Transl Med 15, 158

25. Honka, M. J., Latva-Rasku, A., Bucci, M., Virtanen, K. A., Hannukainen, J. C., Kalliokoski, K. K., and Nuutila, P. (2018) Insulin-stimulated glucose uptake in skeletal muscle, adipose tissue and liver: a positron emission tomography study. Eur J Endocrinol 178, 523-531

12

Table 1. Mass accuracy of compounds measured by Q-Exactive-MS+

Compound Ion Formula Theoretical m/z Measured m/z Difference(ppm)

[2DG-Na]+ Positive C6H12O5Na 187.0576 187.0578 1.1

[2FDG-Na]+ Positive C6H11FO5Na 205.0482 205.0480 1.0

[2ClDG-Na]+ Positive C6H11ClO5Na 221.0187 221.0184 1.4

[2DG6P-H]- Negative C6H12O8P 243.0270 243.0274 1.6

[2FDG6P-H]- Negative C6H11FO8P 261.0170 261.0177 2.7

[2ClDG6P-H]- Negative C6H11ClO8P 276.9875 276.9881 2.2

Table 2. Calibration curve, linearity, LOD, and LOQ of the method

Compound Slope of calibration curve

for 1/x (10 points, n = 3) Linear range(fmol) R2 LOQ(fmol) LOD(fmol)

2FDG 5.70.2 81.4 - 10416 0.9955 81.4 40.7

2FDG-6-P 0.900.03 48.8-50000 0.9998 48.8 24.4

Table 3. Accuracy and reproducibility of the method at various concentrations of analytes

2FDG 2FDG-6-P

Low

(81.4fmol)

Medium

(1.3 pmol)

High

(10.4 pmol)

Low

(48.8fmol)

Medium

(1.6 pmol)

High

(50 pmol)

Replicate 1 – Recovery (%) 103.6 106 98 122 106 99

Intra deviation (SD) 4.8 1.6 3.2 6.7 5.6 4.5

Replicate 2 - Recovery (%) 119 110 100 117 107 99

Intra deviation (SD) 22.3 2.1 1.2 6.8 0.16 4.5

Replicate 3 -Recovery (%) 103 111 101.4 85 105 99

Intra deviation (SD) 5.9 2.2 3.9 105 2.7 1.4

Inter-day average 115 109 99 108 106 99

Inter-day deviation (SD) 10.4 2.5 1.2 20 1.2 0.04

13

FIGURES

Scheme 1. Chemical structure of analytes and internal standards. The chemical structures of all compounds

utilized in this study including: 2DG, 2DG-6-P, 2FDG, 2FDG-6-P, 2ClDG, and 2ClDG-6-P.

14

Figure 1. Ionization of 2DG and 2DG-6-P in the Q-TOF. Ionization of (A) 2DG and (B) 2DG-6-P in the Q-TOF.

15

Figure 2. Fragmentation of [2DG-6-P-H]- in the Q-TOF. Fragmentation of M0 [2DG-6-P-H]- in the Q-TOF at

(A) CE = 10V, (B) CE = 20V, (C) CE = 30V, and (D) CE = 40V.

16

Figure 3. Fragmentation of [2DG-Na]+ in the Q-TOF. Fragmentation of the sodium adduct of 2DG, [2DG-Na]+,

in the Q TOF at (A) CE = 10V, (B) CE = 20V, (C) CE = 30V, and (D) CE = 40V.

17

Figure 4. Effect of sodium ion concentration on the relative abundance of 2DG and glucose using ESI.

Relative abundance of 2DG and glucose using ESI with varying concentrations of NaCl. Data represent mean

±SD. N=3 replicates.

18

Figure 5. Endogenous plasma interferences to 2DG, 2FDG, and 2ClDG. Extracted ion chromatograms for

(A) 2DG m/z at 187.0577 in blank plasma, (B) 2FDG m/z at 205.0481 in blank plasma, (C) 2FDG m/z at 205.0481

in blank plasma spiked with 2.44pmol 2FDG, (D) 2ClDG m/z at 221.0185 in blank plasma, and (E) 2ClDG m/z

at 221.0185 in blank plasma spiked with 1nmol 2ClDG.

19

Figure 6. Skeletal muscle interferences to 2DG-6-P, 2FDG-6-P, and 2ClDG-6-P. Extracted ion

chromatograms for (A) 2DG-6-P m/z at 243.0264 in blank skeletal muscle, (B) 2FDG-6-P m/z at 261.0180 in

blank skeletal muscle, (C) 2FDG-6-P m/z at 261.0180 in blank skeletal muscle spiked with 2.44 pmol 2FDG-6-

P, (D) 2ClDG-6-P m/z at 276.9885 in blank skeletal muscle, and (E) 2ClDG-6-P m/z at 276.9885 in blank skeletal

muscle spiked with 1 nmol 2ClDG-6-P.

20

Figure 7. Chromatograms of 2FDG, 2ClDG, 2FDG-6-P, and 2ClDG-6-P from different concentrations of

analytes and equal concentrations of internal standard(s). Extracted ion chromatograms of 11-12 different

concentrations of (A) 2FDG and (C) 2FDG-6-P with equal concentrations of (B) 2ClDG (internal standard for

2FDG) and (D) 2ClDG-6-P (internal standard for 2FDG-6-P).

21

Figure 8. Calibration curves in distilled water and biological matrices. Calibration curves for (A) 2FDG-6-P

in distilled water, (B) 2FDG in distilled water, (C) 2FDG-6-P calibration curve in skeletal muscle matrix and (D)

2FDG calibration curve in plasma matrix.

download fileview on ChemRxivWilliams et al_2DG_2019.v2.pdf (1.64 MiB)