Journal of Analytical Toxicology 2014;38:184 –193 doi:10.1093/jat/bku018 Advance Access publication March 25, 2014
Article
Quantitation of Diethylene Glycol and Its Metabolites by Gas Chromatography Mass Spectrometry or Ion Chromatography Mass Spectrometry in Rat and Human Biological Samples Adam W. Perala1*, Mark J. Filary1, Michael J. Bartels1 and Kenneth E. McMartin2 1
Toxicology and Environmental Research and Consulting, The Dow Chemical Company, Midland, MI, USA, and 2Department of Pharmacology, Toxicology and Neuroscience, Louisiana State University Health Sciences Center, Shreveport, LA, USA
*Author to whom correspondence should be addressed. Email: [email protected]
The misuse of the commonly used chemical diethylene glycol (DEG) has lead to many poisonings worldwide. Methods were developed for analysis of DEG and its potential metabolites; ethylene glycol, glycolic acid, oxalic acid, diglycolic acid and hydroxyethoxy acetic acid in human urine, serum and cerebrospinal fluid samples, collected following a DEG-associated poisoning in the Republic of Panama during 2006. In addition, methods were developed for rat blood, urine, kidney and liver tissue to support toxicokinetic analysis during the conduct of DEG acute toxicity studies in the rat. Sample analysis was conducted using two techniques; ion chromatography with suppressed conductivity and negative ion electrospray ionization with MS detection or with gas chromatography using electron impact ionization or methane negative chemical ionization with MS detection. Stable-isotope-labeled analogs of each analyte were employed as quantitative internal standards in the assays.
Introduction Diethylene glycol (DEG) is an industrial solvent that has been occasionally mis-formulated into pharmaceutical products, resulting in numerous cases of human poisoning (1, 2). Following a suspected DEG poisoning, determination of this chemical can be achieved by direct LC–MS analysis of the commercial product (3). However, DEG is highly metabolized in mammals, primarily to 2-hydroxyethoxyacetic acid (HEAA) with trace concentrations of oxalic acid (OA) (4). To better understand the mechanism of DEG-induced toxicity, it is important to characterize the routes and magnitude of metabolism of this compound in various mammalian species. A set of analytical methods were therefore developed to support the analysis of DEG and proposed metabolites in blood, urine and tissues from an acute toxicity study in the Wistar rat (5), as well as from suspected poisoning victims in the Republic of Panama during 2006 (3). Analytes measured in these assays are DEG and HEAA as well as diglycolic acid (DGA), ethylene glycol (EG), glycolic acid (GA) and OA. Matrices supported for these methods were human urine, serum and cerebrospinal fluid (CSF) and rat whole blood, urine, kidney and liver tissues. Sample analysis was conducted using two techniques. Ion chromatography with suppressed conductivity and negative ion electrospray ionization with MS detection (IC – SCD – ESI-MS) was utilized to qualitatively screen for novel acidic metabolites of DEG, as well as to quantify the higher concentrations of acid metabolites (HEAA, DGA, GA, OA) in rodent samples. Subsequently, chemical derivatization of the four acidic metabolites, followed by gas chromatography with
electron impact (EI) ionization or methane negative chemical ionization (NCI) with MS detection (GC – EI-MS or GC – NCI-MS) was employed for determination of DEG and EG in all samples and the lower concentrations of HEAA, DGA, GA and OA in human samples (6). The GC – MS methods were modifications of our previously reported techniques for EG, GA and OA (7), with appropriate changes made to include determination of DEG, DGA and HEAA. Stable-isotope-labeled analogs of each analyte were employed as quantitative internal standards in the assays.
Material and Methods Chemicals and reagents Acetonitrile (HPLC grade), methanol (HPLC grade), water (HPLC grade), acetic acid (HPLC grade), sodium carbonate (Na2CO3) and sodium hydroxide (NaOH) were obtained from Fisher Scientific (Itasca, IL, USA). N-Methyl-N-(tert-butyldimethylsilyl)trifluoroacetamide (MTBSTFA) was obtained from Thermo Scientific and 2,3,4,5,6-pentafluorobenzoyl chloride (PFBCl) was obtained from Sigma-Aldrich.
Standards Synthetic standards of DEG (100% purity) and OA (99.6% purity) were obtained from Fluka Chemical Corporation (Milwaukee, WI, USA). Standards of EG (99.8% purity), GA (100% purity) and DGA (99.4% purity) were obtained from Sigma-Aldrich Corporation (St. Louis, MO, USA). A standard of HEAA (99% purity) was obtained from Isotec Incorporated (Miamisburg, OH, USA). The internal standards D8-DEG (isotopic purity .99%) and D6-EG (isotopic purity .99%) were obtained from Cambridge Isotope. The internal standards D4-DGA (isotopic purity .99%), D6-HEAA (isotopic purity .99%), 13C1-GA (isotopic purity .99%) and 13C2-OA (isotopic purity .99%) were all obtained from Isotec Incorporated.
Biological sample supplier Control rat urine, blood, kidney tissue and liver tissue were obtained from in house animals. Control human serum, cerebral spinal fluid and urine were obtained from Bioreclamation Inc. (Westbury, NY, USA). All control matrices were pooled by Bioreclamation from male and female donors with all donors giving informed consent and personal details of donors removed from samples prior to shipment to our laboratory.
# The Author [2014]. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]
Stock and standard preparation Preparation of DEG and EG stock, working solutions and calibration standards for rodent sample analysis DEG and EG stock solutions were prepared at a concentration of 1% in Milli-Q Water. Stock solutions were prepared fresh on each analysis day, long-term stability was not assessed. These stock solutions were diluted with Milli-Q water and fortified into control matrices to prepare matrix standards at the desired concentrations of 0.5, 1, 5, 10, 25, 50 and 100 mg/mL for rodent urine, 0.1, 0.5, 1, 5, 10 and 25 mg/mL for rodent blood and 0.1 0.5, 1, 5 and 10 mg/mL for rodent tissues. Note that fortified tissue “spikes” were made by fortifying analytes into control tissue homogenate (0.1 g, made from pulverized, flash frozen tissue þ Milli-Q water, 1:1). Internal standard solutions of both D8-DEG and D6-EG were prepared by dissolving both internal standards in Milli-Q water at a concentration of 500 or 1,000 mg/mL. Preparation of acid metabolites stocks, working solutions and calibration standards for rodent sample analysis GA, HEAA, OA and DGA stock solutions were prepared at a concentration of 1,000 mg/mL in Milli-Q Water. Stock solutions were prepared fresh on each analysis day, long-term stability was not assessed. These stock solutions were diluted with Milli-Q water to obtain solvent standards at the desired concentrations of 0.05, 0.1, 0.5, 1, 5, 10, 25 and 50 mg/mL for analysis of rodent samples. Solvent standards were selected for use due to the large dilutions needed for sample analysis. Fortified samples were prepared by adding appropriate amounts of the standard working solutions into control matrices. Matrices were fortified at three different concentrations: 1, 5 and 25 mg/mL in urine and 5, 50 and 500 mg/mL for blood and tissue samples. Note that fortified tissue “spikes” were made by fortifying analytes into control tissue homogenate (0.1 g, made from pulverized, flash frozen tissue þ MilliQ water, 1:1). The internal standards, 13C-GA, 13 C2-OA, D4-DGA and D6-HEAA, were prepared by dissolving all internal standards in Milli-Q water and added to the final dilutions at concentrations ranging from 1 to 50 mg/mL. The concentrations of internal standard varied per analyte due to different responses on the instrumentation.
Preparation of DEG and EG stock, working solutions and calibration standards for human sample analysis DEG and EG stock solutions were prepared at a concentration of 1% in Milli-Q Water. These stock solutions were diluted with Milli-Q water and fortified into control matrices to prepare matrix standards at the desired concentrations of 0.025, 0.05, 0.1, 0.5, 1, 5 and 10 mg/mL for human urine, 0.02, 0.05, 0.1, 0.5, 1 and 2 mg/mL for human serum and 0.01, 0.05, 0.1, 0.5, 1 and 5 mg/mL for human cerebral spinal fluid. Internal standard solutions of both D8-DEG and D6-EG were prepared by dissolving both internal standards in Milli-Q water resulting in final concentrations of 2 mg/mL of D8-DEG and 4.4 mg/mL of D6-EG. Preparation of acid metabolites stocks, working solutions and calibration standards for human sample analysis GA, HEAA, OA and DGA stock solutions were prepared at a concentration of 1,000 mg/mL in Milli-Q Water. These stock
solutions were diluted with Milli-Q water and spiked into control matrices to prepare matrix standards each analyte at these listed concentrations range: 0.5– 50 mg GA/mL, 0.5 –100 mg HEAA/mL, 1 – 125 mg DGA/mL and 0.5 – 25 mg OA/mL human matrix (n ¼ 6 – 8 concentrations/analyte/matrix). A mixed stock solution of the internal standards 13C-GA, 13C2-OA, D4-DGA and D6-HEAA was prepared by dissolving all internal standards in Milli-Q water resulting in a final concentration of 25 mg/mL of each internal standard.
Sample preparation Sample preparation for DEG and EG analysis by GC –MS in rodent samples Urine (25 mL) was diluted to 1.25 mL with Milli-Q water, followed by a second dilution of a 40 mL aliquot with 2.0 mL Milli-Q water saturated with NaCl in a 4 mL glass vial. To this diluted sample was added 20 mL of an aqueous internal standard solution (1,000 mL/mL each of D8-DEG and D6-EG). Aliquots of blood (50 mL) or tissue samples (0.1 g of a homogenate made from 0.5 g pulverized, flash frozen tissue þ Milli-Q water, 1:1; Bio-Pulverizer manual impact hammer, Biospec Products, Bartlesville, OK, USA) were added to 1.5 mL of Milli-Q water saturated with NaCl in a 4-mL glass vial and then fortified with 20 –25 mL of an aqueous D8-DEG/D6-EG internal standard solution (500 – 1,000 mg/mL of each). To the diluted samples þ internal standards were added 100 mL of 5 N NaOH, 1.0 mL of toluene and 50 mL of PFBCl. The glass vial was sealed with a Teflon-lined cap and then vortex mixed at 458C for 30 min and then centrifuged at 3,000 rcf for 10 min. Approximately 200 mL of each toluene layer was removed and analyzed by GC– NCI-MS. Sample preparation for acid metabolites by IC– MS in rodent samples Aliquots of urine were initially diluted 50-fold with Milli-Q, followed by a second dilution of a 0.1-mL aliquot with 0.4 mL Milli-Q water and 0.5 mL of Milli-Q containing 2 mg/mL of each of the internal standards (13C-GA, 13C2-OA, D4-DGA and D6-HEAA). Samples were briefly vortex mixed prior to analysis. For blood analysis, a 25-mL sample aliquot was weighed into a vial and 25 mL of acetonitrile was added and the vial, capped with a Teflon-lined cap, was briefly vortex mixed and then sonicated for 10 min. The sample was then centrifuged at 15,000 rcf for 15 min and 20 mL of supernatant was added to a clean 1-mL glass vial. To this vial, 480 mL of Milli-Q water and 10 mL of the internal standard solution (50 mg/mL each of 13C-GA, 13C2-OA, D4-DGA and D6-HEAA) was added, the sample transferred to a 10 kDa spin-filter, centrifuged at 15,000 rcf for 15 min and the resulting filtrate analyzed by IC – MS. Approximately 0.1 g of tissue homogenate (made from 0.5 g pulverized, flash frozen tissue þ Milli-Q water, 1:1; Bio-Pulverizer manual impact hammer) was combined with 0.1 mL of acetonitrile in a 4-mL glass vial, capped with a Teflon-lined cap, briefly vortex mixed, sonicated for 5 min and centrifuged at 15,000 rcf for 15 min. A 50-mL aliquot was transferred to a 2-mL glass autosampler vial containing 400 mL Milli-Q water and 50 mL of an internal standard solution (1 mg/mL each of 13C-GA, 13C2-OA, D4-DGA and D6-HEAA).
Quantitation of Diethylene Glycol and Its Metabolites by GC– MS or IC– MS in Rat and Human Biological Samples 185
Sample preparation for DEG and EG by GC –MS in human samples Varying amounts of urine (0.5 mL), serum (0.25 mL) or CSF samples (0.25 mL) were added to 1.0 mL of Milli-Q water in 4-mL glass vial. A 10-mL aliquot of aqueous internal standard solution (100 mg/mL D8-DEG and 220 mg/mL D6-EG) was added to each sample, followed by 100 mL of 5 N NaOH, 500 mL of toluene and 50 mL of PFBCl. The glass vial was capped with a Teflon-lined cap, vortex mixed at 508C for 1 h and then centrifuged at 3,400 rcf for 10 min. Approximately 200 mL of the toluene layer was removed and analyzed by GC –NCI-MS. Sample preparation for acid metabolites by GC –MS in human samples Serum samples (0.2 mL) were added to acetonitrile at a 1:1 ratio in a centrifuge tube to precipitate proteins, the sample was briefly vortex mixed followed by centrifugation at 15,000 rcf for 10 min. A 200-mL aliquot of urine or serum supernatant þ1.0 mL of 1 N HCl or a 100-mL aliquot of CSF þ 0.9 mL 1 N HCl þ was added to 4-mL glass vial and capped with a Teflon-lined cap. Then 25 mL of the internal standard solution (500 mg/mL each of 13C-GA, 13 C2-OA, D4-DGA and D6-HEAA) was added followed by 1.0 mL of methyl-tert butyl ether (MTBE) containing 0.5% trioctylphosphine oxide. The sample was ortex-mixed for 30 min and the MTBE layer was transferred to a clean 2-mL glass auto sampler vial, the extraction step was repeated and the extracts combined. Samples were blown to dryness under a nitrogen steam and then reconstituted in 450 mL of toluene and 50 mL of MTBSTFA derivatization reagent were added. The vial was capped with a Teflon-lined crimp cap and heated at 608C for 1 h and then analyzed by GC–EI-MS.
Ion chromatography and mass spectrometric conditions for acid metabolites (HEAA, DGA, GA and OA) A Dionex RFIC ion chromatograph model ICS-3000 (Dionex Corp., Sunnyvale, CA, USA), integrated with a Dionex IonPac AG11-HC analytical column (4 250 mm) was used for the chromatographic separation of acid metabolites, and their corresponding internal standards. The mobile phase consisted of HPLC-grade water delivered to an EGC II KOH eluent generator to produce an ion strength gradient which was started at 0.1 mM KOH with a flow rate 1.0 mL/min, held for 5 min and then linearly increased over 35 min to 10 mM KOH, held for 5 min and then linearly increased over 5 – 60 mM and held for 10 min. Finally, the gradient was returned to the initial ion strength of 0.1 mM KOH and a 7-min equilibration time was incorporated between runs. After the eluent exited to the suppressed conductivity detector, 50% of the flow was directed into the ESI-MS system and 50% of the flow was directed across the suppressor. MS detection was performed in the negative ESI mode on an Agilent Technologies (Santa Clara, CA, USA) model G1956B single quadrupole mass spectrometer operated with the following parameters: Fragmentor voltage ¼ 70 V, drying gas temperature ¼ 3508C, drying gas flow ¼ 10.0 L/min, nebulizer pressure ¼ 40.0 psi, and capillary voltage ¼ 4,000 V. The following ions were used for quantitation with the MS operated in the selected-ion monitoring (SIM) mode: GA, m/z 75.0; 13 C-GA, m/z 76.0; HEAA, m/z 119.0; D6-HEAA m/z 125.0, OA, 186 Perala et al.
m/z 89.0; 13C2-OA, m/z 91.0, DGA, m/z 133.0; D4-DGA, m/z 137.0 with dwell time of 146 ms for all ions and internal standards.
Gas chromatography and mass spectrometric conditions for acid metabolites (HEAA, DGA, GA and OA) An Agilent chromatograph model 6890N system was used with a Phenomenex (Torrance, CA, USA) ZB-5 ms column (30 0.25 mm ID 0.50 mm film) for the chromatographic separation of acid metabolites, and their corresponding internal standards (as t-butyldimethylsilyl derivatives). Helium carrier gas was used in constant flow mode. Derivatized extracts of urine, serum and CSF samples were injected using a 2.0-mL splitless injection in a 2758C injection port (helium carrier pressure 10.9 psi; purge flow 50 mL/min; purge time 0.25 min). The temperature gradient started at 408C and linearly increased at 188C/ min to 2808C and held for 2.0 min. MS detection was performed in the EI ionization mode on an Agilent model 5973A single quadrupole mass spectrometer operated with the following parameters: transfer line temperature ¼ 2808C, MS quadrupole temperature ¼ 1508C and MS source temperature ¼ 2308C. The following ions were used for quantitation with the MS operated in the SIM mode: GA, m/z 247.1; 13C-GA, m/z 248.2; HEAA, m/z 291.2; D6-HEAA, m/z 297.2, OA, m/z 261.1; 13C2-OA, m/z 263.1, DGA, m/z 305.2; D4-DGA, m/z 309.2 with dwell time of 100 ms for all ions and internal standards.
Gas chromatography and mass spectrometric conditions for DEG and EG An Agilent chromatograph model 6890þ system was used with a Phenomenex ZB-5 ms column (30 m 0.25 mm ID 0.50 mm film) for the chromatographic separation of DEG and EG, and their corresponding internal standards. Helium carrier gas was used in constant flow mode. Derivatized urine samples were injected using a 2.0-mL split injection in a 2758C injection port (helium carrier pressure 11.95 psi; split flow 11.1 mL/ min; split ratio 10:1). Derivatized serum and CSF samples were injected using a 2.0-mL pulsed splitless injection in a 2758C injection port (helium carrier pressure 11.95 psi; pulse pressure 25 psi; pulse time 0.3 min; purge flow 50 mL/min; purge time 0.25 min). The temperature gradient started at 1008C held for 0.5 min and then linearly increased at 208C/min to 3208C and held for 1.0 min. MS detection was performed in the NCI mode using methane gas on an Agilent model 5973N single quadrupole mass spectrometer operated with the following parameters: Transfer line temperature ¼ 3008C, MS quadrupole temperature ¼ 1808C and MS source temperature ¼ 2008C. The following ions were used quantitation with the MS operated in the SIM mode: DEG, m/z 494.0; D8-DEG m/z 502.0; EG, m/z 450.0; D6-EG, m/z 454.0 with dwell time of 100 ms for all ions and internal standards.
Data analysis Integration of Selected-Ion Monitoring mass spectral datafiles SIM datafiles collected on both the Agilent G1956B LC/MSD and the Agilent 5973B mass spectrometers were converted to “.wiff”
datafiles, compatible with the AB Sciex Analyst v1.5 software (Framingham, MA, USA). This conversion was done with the “Translat.exe” utility, available with the Analyst software package. The resulting Analyst-format SIM-chromatogram files were processed with Analyst v1.5 to obtain chromatographic peak areas for the analytes and respective internal standards. Chromatograms obtained from the IC-suppressed conductivity detector were collected and displayed with the Dionex Chromeleon software (v 6.8). Linearity Linearity of calibration was tested by analyzing solvent calibration standards or matrix standards prepared above. Calibration curves in the concentration range for all analytes were constructed by plotting the corresponding peak area ratios of analyte/internal standard versus the concentrations. Weighted (1/x) linear regression analysis, available in the Analyst v1.5 application, was used to determine the slope, intercept and correlation coefficient (R 2) for all calibration curves. Method quantitation limits The limit of quantitation (LOQ) was set at the concentration below which the method could not operate with acceptable precision and accuracy (+15% of original value), and at which the signal-to-noise ratio was .8.
Results and discussion Chromatography and selectivity The samples from the DEG acute toxicity study in rats contained relatively high concentrations of metabolites, as the dose concentrations employed were 2 – 10 g/kg body weight (5). The IC – MS method was therefore utilized for quantitation of acid metabolites in these rat study samples because of ease of sample preparation, as the analytes required no derivatization prior to analysis. In addition, the simultaneous use of suppressed conductivity detection allowed for the qualitative analysis of any other substantial peaks that arise from DEG, beyond the targeted acidic metabolites of HEAA, DGA, GA and OA. Preliminary IC – MS analysis of HEAA, DGA, GA and OA in 50-fold diluted urine samples afforded detection limits of 1 – 5 mg/mL (data not shown). However, matrix suppression of the GA analyte and corresponding internal standard required further dilution to 500-fold to obtain acceptable recovery of GA from the urine matrix. For quantitation of acid metabolites in blood and tissue samples, the samples were first mixed with acetonitrile to precipitate proteins followed by a high-speed centrifuge and final dilution in Milli-Q water. Total dilution factors for blood and tissue samples were 50-fold. For human samples, the IC – MS method did not have sufficiently low detection limits to quantitate the acid metabolites in various matrices (Table I). To support the analysis of human poisoning-case samples (6), a GC – MS method based on our previous technique for GA and OA analysis (7) was used for analysis of these metabolites, as well as the additional DEG metabolites of HEAA and DGA, in serum, urine and CSF samples, as the PFBCl derivatives. GC –MS was also used for DEG and EG analysis for all rat and human matrices. Sample preparation was slightly modified between the rat and human methods, which involved removing
Table I Limits of Quantitation Analyte
Limits of quantitation (mg/g or mL) Rat matrices
DEG EG GA OA DGA HEAA
Human matrices
Blood
Urine
Tissue
Serum
Urine
CSF
5.02 5.02 15.6a 3.17a 1.6a 3.88a
10.1 50.0 50.5a 25.6a 25.7a 25.3a
2.06 5.14 25.7a 10.7a 2a 2a
0.0203 0.0201 0.51 1.03 10.1 0.513
0.0507 0.101 0.255 – 1.26 0.257
0.0507 0.101 0.51 – 1.01 0.513
a
Analysis conducted by IC –MS (all others via GC –MS).
the sodium chloride from the diluent water of the human urine, serum and spinal fluid samples. Quantitation limits were better for human samples, but the major factor contributing to this was the larger sample size that was available for analysis. Typical SIM ion chromatograms for IC–MS analysis of GA, OA, HEAA and DGA are shown in Figure 1; the overall chromatography run times were within 55 min. Under the IC – MS conditions used here, the analytes of OA, GA, HEAA and DGA have retention times of 13.6, 14.2, 35.5 and 43.9 min, respectively (Figure 1A). The corresponding internal standards for each analyte have almost the same retention times as the analytes (Figure 1B). Control water did not show a SIM response at the retention time of DGA (Figure 2). A minor, broad peak (11–16 min) was present in blank water that coeluted with the well-resolved peaks for HEAA and GA. This interference was 20 –100 lower than the stated LOQ for these analytes in the various rat matrices, and did not impact the precision of the methods, as shown in the standard refits, discussed below. A second minor peak was present in blank water that coeluted with OA. Again, this peak was 3–25 lower than the stated LOQ for OA in the various rat matrices, and did not impact method precision. Typical SIM ion chromatograms for GC – MS analysis of the t-butyldimethylsilyl derivatives of GA, OA, HEAA and DGA are shown in Figure 3; the overall chromatography run times were within 13 min. Under the GC–EI-MS conditions used here, the analytes of GA, OA, HEAA and DGA have retention times of 9.1, 9.1, 10.93 and 11.5 min, respectively (Figure 3A–D). The corresponding internal standards for each analyte have almost the same retention times as the analytes (Figure 3E–H) and no isotopic cross-over was observed. Control serum contained low, endogenous concentrations of GA, OA and DGA, but no detectable HEAA. Typical SIM ion chromatograms of DEG and EG are shown in Figure 4; the overall chromatography run times were within 13 min. Under the GC – NCI-MS conditions used here, the analytes of DEG and EG have retention times of 8.3 and 9.6 min (Figure 4A). The corresponding internal standards for each analyte have almost the same retention times as the analytes (Figure 4B) and no isotopic cross-over was observed. Control serum did not show a SIM response for any of those analytes, as shown in Figure 5. IC – MS or GC – MS spectra of the various analytes and internal standards are shown in Figures 6 –8.
Calibration curve, linearity and limit of quantitation A series of solvent standards were prepared for acidic metabolite analysis by IC –MS, or matrix calibration standards for the GC–MS
Quantitation of Diethylene Glycol and Its Metabolites by GC– MS or IC– MS in Rat and Human Biological Samples 187
Figure 1. Example summed, selected-ion IC–MS chromatograms of an acid metabolites standard containing 1 mg/mL of each analyte (GA, OA, DGA and HEAA) and 1 mg/mL of each internal standard (13C-GA, 13C2-OA, D4-DGA and D6-HEAA).
Figure 2. Example summed, selected-ion IC–MS chromatograms of a water blank containing 1 mg/mL of each internal standard (13C-GA, 13C2-OA, D4-DGA and D6-HEAA). 188 Perala et al.
Figure 3. Example selected-ion GC–MS chromatograms of an acid metabolites serum matrix standard containing 10 mg analyte/mL of each analyte and 25 mg/mL of each internal standard (13C-GA, 13C2-OA, D4-DGA and D6-HEAA).
analyses, at generally 6 –8 concentrations with a fixed amount of their corresponding internal standards. These solutions were processed and analyzed at least three times on separate days.
Overall, the calibration curves obtained were linear over the concentration ranges required for the assays. The LOQ’s for each analyte can be seen in Table I. The method quantitation limits and
Quantitation of Diethylene Glycol and Its Metabolites by GC– MS or IC– MS in Rat and Human Biological Samples 189
Figure 4. Example selected-ion GC – MS chromatograms of a serum matrix standard containing DEG and EG (0.5 mg/mL of each analyte) and 2 mg/mL of D8-DEG and 4.4 mg/mL D6-EG internal standards.
Figure 5. Example selected-ion GC–MS chromatograms of control serum sample by GC–MS.
190 Perala et al.
Figure 6. Representative GC–NCI-MS mass spectra of DEG (A), D8-DEG (B), EG (C) and D6-EG (D) as PFBCl derivatives.
Figure 7. Representative IC–MS mass spectra of GA (A), 13C-GA (B), OA (C), 13C2-OA (D), DGA (E), D4-DGA (F), HEAA (G) and D6-HEAA (H).
Quantitation of Diethylene Glycol and Its Metabolites by GC– MS or IC– MS in Rat and Human Biological Samples 191
Figure 8. Representative GC–MS mass spectra GA (A), 13C-GA (B), OA (C), 13C2-OA (D), DGA (E), D4-DGA (F), HEAA (G) and D6-HEAA (H) þ MTBSTFA derivative.
accuracy were also acceptable for the analysis of rat or human samples, with standard refits to the calibration curve generally within 15% of nominal values (see Supplementary data).
Effect of matrix on recovery In the GC– MS assays, matrix standards were utilized so fortified samples were not prepared. In the IC –MS assays solvent standard along with fortified matrix spikes were used to determine acid metabolite concentrations in rat matrices. However, relative recoveries (analyte vs. internal standard) were quite acceptable, with matrix standard refits, as discussed above, generally within 15% of nominal values. Recovery of the four acidic metabolites via the IC –MS method was somewhat more variable ( +30%), but was sufficient for the determination of these analytes in this initial, acute toxicity study (Tables II–IV).
Advantages of various methods The IC –MS method for acid metabolite analysis showed acceptable chromatographic resolution with adequate sensitivity for the high concentrations of metabolites found in the rat acute toxicity study samples. The sample preparation for this method was quite rapid, requiring only dilution and blood protein precipitation prior to analysis. Incorporation of suppressed conductivity detection also allowed for screening of non-targeted analytes. The major disadvantages were long sample analysis times (.40 min per sample) and large dilutions of samples, which limited method quantitation limits. 192 Perala et al.
Table II Linearity for DEG and EG by GC –EI-MS Rat analyte DEG EG Human analyte DEG EG
Linearity range (mg/mL) 0.5 –100 5 –100 0.02 –2 0.02 –2
R2 0.9992 0.9988 0.9984 0.9989
Table III Linearity for Acid Metabolites by IC –MS Analyte
Linearity range (mg/mL)
R2
GA OA HEAA DGA
0.1 –25 0.05 –5 0.05 –50 0.05 –5
0.9999 0.9899 0.9952 0.9990
Table IV Linearity for Acid Metabolites by GC –NCI-MS Analyte
Linearity range (mg/mL)
R2
GA OA HEAA DGA
0.05 –51 0.5 –25 0.5 –100 1 –100
0.9990 0.9907 0.9985 0.9922
The GC– MS methods for metabolite analysis for DEG/EG in all samples, and the acidic metabolites in the human poisoning-case samples showed acceptable chromatographic resolution with adequate sensitivity. The analysis run times (12 min per sample) were quick and minimum sample dilution required. The major disadvantages were a complex sample preparation (extraction/ derivatization) and somewhat noisier chromatography baselines vs. the IC –MS method.
Specificity The methods described in this manuscript are considered to be specific for the determination of DEG and the five potential metabolites (ED, GA, OA, HEAA and DGA) in the various biological matrices. In the GC –MS analysis, method additional confirmation ions were not incorporated into the assay, due to the lack of significant, compound-specific fragment ions for some analytes and based on the extremely low concentrations of metabolites (EG, GA, OA, HEAA and DGA) present in actual samples. The use of stable-isotope-labeled internal standards, however, confirmed analytes’ retention times for each sample analyzed. Also, the retention time and mass spectral response for each analyte was as a chemical derivative, providing further specificity to these analyses. Slight differences in retention times between the deuterated internal standard and the analyte are a common occurrence (8, 9).
Conclusion Analytical methods were developed for the analysis of DEG and five potential metabolites in biological matrices of rat and human. The GC –MS method used for human sample analysis is a modification of previously published techniques for EG, GA and OA. The IC – MS method developed for the four acidic metabolites of DEG is novel and has been applied to rat blood, rat urine, rat tissues, human serum, human urine and human CSF. The GC – EI-MS method for acid metabolite analysis has better detection limits but requires more sample preparation vs. the IC – MS method. The IC –MS method is adequate for samples with higher concentrations of OA, GA, HEAA and GA, requiring very little sample preparation and affords screening of non-targeted analytes (via suppressed conductivity detection). The GC – NCI-MS analysis method for DEG and EG is robust and has acceptable detection limits. Future research will continue to use the GC– MS methods for DEG and its metabolites; however, we will explore the possibility of using GC with triple quadrupole mass spectrometry detection to improve selectivity, sensitivity and provide additional confirmation of analyte structure. These sensitivity improvements will support future toxicokinetic animal studies that may evaluate a dose-response in DEG metabolism.
Supplementary data Supplementary data are available at Journal of Analytical Toxicology online.
Funding Funding for this work was provided by the Ethylene Glycol/ Ethylene Oxide Panel of the American Chemistry Council and the U.S. Centers for Disease Control and Prevention.
Acknowledgment The authors thank Dan Markham for critical review of this manuscript.
References 1. Wax, P.M. (1995) Elixirs, diluents, and the passage of the 1938 Federal Food, Drug and Cosmetic Act. Annals of Internal Medicine, 122, 456– 461. 2. Wax, P.M. (1996) It’s happening again – another diethylene glycol mass poisoning. Journal of Toxicology – Clinical Toxicology, 34, 517– 520. 3. Barr, D.B., Barr, J.R., Weerasekera, G., Wamsley, J., Kalb, S.R., Sjodin, A. et al. (2007) Identification and quantification of diethylene glycol in pharmaceuticals implicated in poisoning epidemics: an historical laboratory perspective. Journal of Analytical Toxicology, 31, 295–303. 4. Mathews, J.M., Parker, M.K., Matthews, H.B. (1991) Metabolism and disposition of diethylene glycol in rat and dog. Drug Metabolism and Disposition: The Biological Fate of Chemicals, 19, 1066– 1070. 5. Besenhofer, L.M., Adegboyega, P.A., Bartels, M., Filary, M.J., Perala, A.W., McLaren, M.C. et al. (2010) Inhibition of metabolism of diethylene glycol prevents target organ toxicity in rats. Toxicological Sciences, 117, 25– 35. doi:10.1093/toxsci/kfq167. 6. Schier, J.G., Hunt, D.R., Perala, A., McMartin, K.E., Bartels, M.J., Lewis, L.S. et al. (2013) Characterizing concentrations of diethylene glycol and suspected metabolites in human serum, urine, and cerebrospinal fluid samples from the Panama DEG mass poisoning. Clinical Toxicology (Philadelphia), doi:10.3109/15563650.2013.850504. 7. Pottenger, L.H., Carney, E.W., Bartels, M.J. (2001) Dose-dependent nonlinear pharmacokinetics of ethylene glycol metabolites in pregnant (GD 10) and nonpregnant Sprague-Dawley rats following oral administration of ethylene glycol. Toxicological Sciences, 62, 10 –19. 8. Baillie, T.A. (1981) The use of stable isotopes in pharmacological research.Pharmacological Reviews, 33, 81 –132. 9. Turowski, M., Yamakawa, N., Meller, J., Kimata, K., Ikegami, T., Hosoya, K. et al. (2003) Deuterium isotope effects on hydrophobic interactions: the importance of dispersion interactions in the hydrophobic phase. Journal of the American Chemical Society, 125, 13836 – 49, doi:10.1021/ja036006 g.
Quantitation of Diethylene Glycol and Its Metabolites by GC– MS or IC– MS in Rat and Human Biological Samples 193
Article
Quantitation of Diethylene Glycol and Its Metabolites by Gas Chromatography Mass Spectrometry or Ion Chromatography Mass Spectrometry in Rat and Human Biological Samples Adam W. Perala1*, Mark J. Filary1, Michael J. Bartels1 and Kenneth E. McMartin2 1
Toxicology and Environmental Research and Consulting, The Dow Chemical Company, Midland, MI, USA, and 2Department of Pharmacology, Toxicology and Neuroscience, Louisiana State University Health Sciences Center, Shreveport, LA, USA
*Author to whom correspondence should be addressed. Email: [email protected]
The misuse of the commonly used chemical diethylene glycol (DEG) has lead to many poisonings worldwide. Methods were developed for analysis of DEG and its potential metabolites; ethylene glycol, glycolic acid, oxalic acid, diglycolic acid and hydroxyethoxy acetic acid in human urine, serum and cerebrospinal fluid samples, collected following a DEG-associated poisoning in the Republic of Panama during 2006. In addition, methods were developed for rat blood, urine, kidney and liver tissue to support toxicokinetic analysis during the conduct of DEG acute toxicity studies in the rat. Sample analysis was conducted using two techniques; ion chromatography with suppressed conductivity and negative ion electrospray ionization with MS detection or with gas chromatography using electron impact ionization or methane negative chemical ionization with MS detection. Stable-isotope-labeled analogs of each analyte were employed as quantitative internal standards in the assays.
Introduction Diethylene glycol (DEG) is an industrial solvent that has been occasionally mis-formulated into pharmaceutical products, resulting in numerous cases of human poisoning (1, 2). Following a suspected DEG poisoning, determination of this chemical can be achieved by direct LC–MS analysis of the commercial product (3). However, DEG is highly metabolized in mammals, primarily to 2-hydroxyethoxyacetic acid (HEAA) with trace concentrations of oxalic acid (OA) (4). To better understand the mechanism of DEG-induced toxicity, it is important to characterize the routes and magnitude of metabolism of this compound in various mammalian species. A set of analytical methods were therefore developed to support the analysis of DEG and proposed metabolites in blood, urine and tissues from an acute toxicity study in the Wistar rat (5), as well as from suspected poisoning victims in the Republic of Panama during 2006 (3). Analytes measured in these assays are DEG and HEAA as well as diglycolic acid (DGA), ethylene glycol (EG), glycolic acid (GA) and OA. Matrices supported for these methods were human urine, serum and cerebrospinal fluid (CSF) and rat whole blood, urine, kidney and liver tissues. Sample analysis was conducted using two techniques. Ion chromatography with suppressed conductivity and negative ion electrospray ionization with MS detection (IC – SCD – ESI-MS) was utilized to qualitatively screen for novel acidic metabolites of DEG, as well as to quantify the higher concentrations of acid metabolites (HEAA, DGA, GA, OA) in rodent samples. Subsequently, chemical derivatization of the four acidic metabolites, followed by gas chromatography with
electron impact (EI) ionization or methane negative chemical ionization (NCI) with MS detection (GC – EI-MS or GC – NCI-MS) was employed for determination of DEG and EG in all samples and the lower concentrations of HEAA, DGA, GA and OA in human samples (6). The GC – MS methods were modifications of our previously reported techniques for EG, GA and OA (7), with appropriate changes made to include determination of DEG, DGA and HEAA. Stable-isotope-labeled analogs of each analyte were employed as quantitative internal standards in the assays.
Material and Methods Chemicals and reagents Acetonitrile (HPLC grade), methanol (HPLC grade), water (HPLC grade), acetic acid (HPLC grade), sodium carbonate (Na2CO3) and sodium hydroxide (NaOH) were obtained from Fisher Scientific (Itasca, IL, USA). N-Methyl-N-(tert-butyldimethylsilyl)trifluoroacetamide (MTBSTFA) was obtained from Thermo Scientific and 2,3,4,5,6-pentafluorobenzoyl chloride (PFBCl) was obtained from Sigma-Aldrich.
Standards Synthetic standards of DEG (100% purity) and OA (99.6% purity) were obtained from Fluka Chemical Corporation (Milwaukee, WI, USA). Standards of EG (99.8% purity), GA (100% purity) and DGA (99.4% purity) were obtained from Sigma-Aldrich Corporation (St. Louis, MO, USA). A standard of HEAA (99% purity) was obtained from Isotec Incorporated (Miamisburg, OH, USA). The internal standards D8-DEG (isotopic purity .99%) and D6-EG (isotopic purity .99%) were obtained from Cambridge Isotope. The internal standards D4-DGA (isotopic purity .99%), D6-HEAA (isotopic purity .99%), 13C1-GA (isotopic purity .99%) and 13C2-OA (isotopic purity .99%) were all obtained from Isotec Incorporated.
Biological sample supplier Control rat urine, blood, kidney tissue and liver tissue were obtained from in house animals. Control human serum, cerebral spinal fluid and urine were obtained from Bioreclamation Inc. (Westbury, NY, USA). All control matrices were pooled by Bioreclamation from male and female donors with all donors giving informed consent and personal details of donors removed from samples prior to shipment to our laboratory.
# The Author [2014]. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]
Stock and standard preparation Preparation of DEG and EG stock, working solutions and calibration standards for rodent sample analysis DEG and EG stock solutions were prepared at a concentration of 1% in Milli-Q Water. Stock solutions were prepared fresh on each analysis day, long-term stability was not assessed. These stock solutions were diluted with Milli-Q water and fortified into control matrices to prepare matrix standards at the desired concentrations of 0.5, 1, 5, 10, 25, 50 and 100 mg/mL for rodent urine, 0.1, 0.5, 1, 5, 10 and 25 mg/mL for rodent blood and 0.1 0.5, 1, 5 and 10 mg/mL for rodent tissues. Note that fortified tissue “spikes” were made by fortifying analytes into control tissue homogenate (0.1 g, made from pulverized, flash frozen tissue þ Milli-Q water, 1:1). Internal standard solutions of both D8-DEG and D6-EG were prepared by dissolving both internal standards in Milli-Q water at a concentration of 500 or 1,000 mg/mL. Preparation of acid metabolites stocks, working solutions and calibration standards for rodent sample analysis GA, HEAA, OA and DGA stock solutions were prepared at a concentration of 1,000 mg/mL in Milli-Q Water. Stock solutions were prepared fresh on each analysis day, long-term stability was not assessed. These stock solutions were diluted with Milli-Q water to obtain solvent standards at the desired concentrations of 0.05, 0.1, 0.5, 1, 5, 10, 25 and 50 mg/mL for analysis of rodent samples. Solvent standards were selected for use due to the large dilutions needed for sample analysis. Fortified samples were prepared by adding appropriate amounts of the standard working solutions into control matrices. Matrices were fortified at three different concentrations: 1, 5 and 25 mg/mL in urine and 5, 50 and 500 mg/mL for blood and tissue samples. Note that fortified tissue “spikes” were made by fortifying analytes into control tissue homogenate (0.1 g, made from pulverized, flash frozen tissue þ MilliQ water, 1:1). The internal standards, 13C-GA, 13 C2-OA, D4-DGA and D6-HEAA, were prepared by dissolving all internal standards in Milli-Q water and added to the final dilutions at concentrations ranging from 1 to 50 mg/mL. The concentrations of internal standard varied per analyte due to different responses on the instrumentation.
Preparation of DEG and EG stock, working solutions and calibration standards for human sample analysis DEG and EG stock solutions were prepared at a concentration of 1% in Milli-Q Water. These stock solutions were diluted with Milli-Q water and fortified into control matrices to prepare matrix standards at the desired concentrations of 0.025, 0.05, 0.1, 0.5, 1, 5 and 10 mg/mL for human urine, 0.02, 0.05, 0.1, 0.5, 1 and 2 mg/mL for human serum and 0.01, 0.05, 0.1, 0.5, 1 and 5 mg/mL for human cerebral spinal fluid. Internal standard solutions of both D8-DEG and D6-EG were prepared by dissolving both internal standards in Milli-Q water resulting in final concentrations of 2 mg/mL of D8-DEG and 4.4 mg/mL of D6-EG. Preparation of acid metabolites stocks, working solutions and calibration standards for human sample analysis GA, HEAA, OA and DGA stock solutions were prepared at a concentration of 1,000 mg/mL in Milli-Q Water. These stock
solutions were diluted with Milli-Q water and spiked into control matrices to prepare matrix standards each analyte at these listed concentrations range: 0.5– 50 mg GA/mL, 0.5 –100 mg HEAA/mL, 1 – 125 mg DGA/mL and 0.5 – 25 mg OA/mL human matrix (n ¼ 6 – 8 concentrations/analyte/matrix). A mixed stock solution of the internal standards 13C-GA, 13C2-OA, D4-DGA and D6-HEAA was prepared by dissolving all internal standards in Milli-Q water resulting in a final concentration of 25 mg/mL of each internal standard.
Sample preparation Sample preparation for DEG and EG analysis by GC –MS in rodent samples Urine (25 mL) was diluted to 1.25 mL with Milli-Q water, followed by a second dilution of a 40 mL aliquot with 2.0 mL Milli-Q water saturated with NaCl in a 4 mL glass vial. To this diluted sample was added 20 mL of an aqueous internal standard solution (1,000 mL/mL each of D8-DEG and D6-EG). Aliquots of blood (50 mL) or tissue samples (0.1 g of a homogenate made from 0.5 g pulverized, flash frozen tissue þ Milli-Q water, 1:1; Bio-Pulverizer manual impact hammer, Biospec Products, Bartlesville, OK, USA) were added to 1.5 mL of Milli-Q water saturated with NaCl in a 4-mL glass vial and then fortified with 20 –25 mL of an aqueous D8-DEG/D6-EG internal standard solution (500 – 1,000 mg/mL of each). To the diluted samples þ internal standards were added 100 mL of 5 N NaOH, 1.0 mL of toluene and 50 mL of PFBCl. The glass vial was sealed with a Teflon-lined cap and then vortex mixed at 458C for 30 min and then centrifuged at 3,000 rcf for 10 min. Approximately 200 mL of each toluene layer was removed and analyzed by GC– NCI-MS. Sample preparation for acid metabolites by IC– MS in rodent samples Aliquots of urine were initially diluted 50-fold with Milli-Q, followed by a second dilution of a 0.1-mL aliquot with 0.4 mL Milli-Q water and 0.5 mL of Milli-Q containing 2 mg/mL of each of the internal standards (13C-GA, 13C2-OA, D4-DGA and D6-HEAA). Samples were briefly vortex mixed prior to analysis. For blood analysis, a 25-mL sample aliquot was weighed into a vial and 25 mL of acetonitrile was added and the vial, capped with a Teflon-lined cap, was briefly vortex mixed and then sonicated for 10 min. The sample was then centrifuged at 15,000 rcf for 15 min and 20 mL of supernatant was added to a clean 1-mL glass vial. To this vial, 480 mL of Milli-Q water and 10 mL of the internal standard solution (50 mg/mL each of 13C-GA, 13C2-OA, D4-DGA and D6-HEAA) was added, the sample transferred to a 10 kDa spin-filter, centrifuged at 15,000 rcf for 15 min and the resulting filtrate analyzed by IC – MS. Approximately 0.1 g of tissue homogenate (made from 0.5 g pulverized, flash frozen tissue þ Milli-Q water, 1:1; Bio-Pulverizer manual impact hammer) was combined with 0.1 mL of acetonitrile in a 4-mL glass vial, capped with a Teflon-lined cap, briefly vortex mixed, sonicated for 5 min and centrifuged at 15,000 rcf for 15 min. A 50-mL aliquot was transferred to a 2-mL glass autosampler vial containing 400 mL Milli-Q water and 50 mL of an internal standard solution (1 mg/mL each of 13C-GA, 13C2-OA, D4-DGA and D6-HEAA).
Quantitation of Diethylene Glycol and Its Metabolites by GC– MS or IC– MS in Rat and Human Biological Samples 185
Sample preparation for DEG and EG by GC –MS in human samples Varying amounts of urine (0.5 mL), serum (0.25 mL) or CSF samples (0.25 mL) were added to 1.0 mL of Milli-Q water in 4-mL glass vial. A 10-mL aliquot of aqueous internal standard solution (100 mg/mL D8-DEG and 220 mg/mL D6-EG) was added to each sample, followed by 100 mL of 5 N NaOH, 500 mL of toluene and 50 mL of PFBCl. The glass vial was capped with a Teflon-lined cap, vortex mixed at 508C for 1 h and then centrifuged at 3,400 rcf for 10 min. Approximately 200 mL of the toluene layer was removed and analyzed by GC –NCI-MS. Sample preparation for acid metabolites by GC –MS in human samples Serum samples (0.2 mL) were added to acetonitrile at a 1:1 ratio in a centrifuge tube to precipitate proteins, the sample was briefly vortex mixed followed by centrifugation at 15,000 rcf for 10 min. A 200-mL aliquot of urine or serum supernatant þ1.0 mL of 1 N HCl or a 100-mL aliquot of CSF þ 0.9 mL 1 N HCl þ was added to 4-mL glass vial and capped with a Teflon-lined cap. Then 25 mL of the internal standard solution (500 mg/mL each of 13C-GA, 13 C2-OA, D4-DGA and D6-HEAA) was added followed by 1.0 mL of methyl-tert butyl ether (MTBE) containing 0.5% trioctylphosphine oxide. The sample was ortex-mixed for 30 min and the MTBE layer was transferred to a clean 2-mL glass auto sampler vial, the extraction step was repeated and the extracts combined. Samples were blown to dryness under a nitrogen steam and then reconstituted in 450 mL of toluene and 50 mL of MTBSTFA derivatization reagent were added. The vial was capped with a Teflon-lined crimp cap and heated at 608C for 1 h and then analyzed by GC–EI-MS.
Ion chromatography and mass spectrometric conditions for acid metabolites (HEAA, DGA, GA and OA) A Dionex RFIC ion chromatograph model ICS-3000 (Dionex Corp., Sunnyvale, CA, USA), integrated with a Dionex IonPac AG11-HC analytical column (4 250 mm) was used for the chromatographic separation of acid metabolites, and their corresponding internal standards. The mobile phase consisted of HPLC-grade water delivered to an EGC II KOH eluent generator to produce an ion strength gradient which was started at 0.1 mM KOH with a flow rate 1.0 mL/min, held for 5 min and then linearly increased over 35 min to 10 mM KOH, held for 5 min and then linearly increased over 5 – 60 mM and held for 10 min. Finally, the gradient was returned to the initial ion strength of 0.1 mM KOH and a 7-min equilibration time was incorporated between runs. After the eluent exited to the suppressed conductivity detector, 50% of the flow was directed into the ESI-MS system and 50% of the flow was directed across the suppressor. MS detection was performed in the negative ESI mode on an Agilent Technologies (Santa Clara, CA, USA) model G1956B single quadrupole mass spectrometer operated with the following parameters: Fragmentor voltage ¼ 70 V, drying gas temperature ¼ 3508C, drying gas flow ¼ 10.0 L/min, nebulizer pressure ¼ 40.0 psi, and capillary voltage ¼ 4,000 V. The following ions were used for quantitation with the MS operated in the selected-ion monitoring (SIM) mode: GA, m/z 75.0; 13 C-GA, m/z 76.0; HEAA, m/z 119.0; D6-HEAA m/z 125.0, OA, 186 Perala et al.
m/z 89.0; 13C2-OA, m/z 91.0, DGA, m/z 133.0; D4-DGA, m/z 137.0 with dwell time of 146 ms for all ions and internal standards.
Gas chromatography and mass spectrometric conditions for acid metabolites (HEAA, DGA, GA and OA) An Agilent chromatograph model 6890N system was used with a Phenomenex (Torrance, CA, USA) ZB-5 ms column (30 0.25 mm ID 0.50 mm film) for the chromatographic separation of acid metabolites, and their corresponding internal standards (as t-butyldimethylsilyl derivatives). Helium carrier gas was used in constant flow mode. Derivatized extracts of urine, serum and CSF samples were injected using a 2.0-mL splitless injection in a 2758C injection port (helium carrier pressure 10.9 psi; purge flow 50 mL/min; purge time 0.25 min). The temperature gradient started at 408C and linearly increased at 188C/ min to 2808C and held for 2.0 min. MS detection was performed in the EI ionization mode on an Agilent model 5973A single quadrupole mass spectrometer operated with the following parameters: transfer line temperature ¼ 2808C, MS quadrupole temperature ¼ 1508C and MS source temperature ¼ 2308C. The following ions were used for quantitation with the MS operated in the SIM mode: GA, m/z 247.1; 13C-GA, m/z 248.2; HEAA, m/z 291.2; D6-HEAA, m/z 297.2, OA, m/z 261.1; 13C2-OA, m/z 263.1, DGA, m/z 305.2; D4-DGA, m/z 309.2 with dwell time of 100 ms for all ions and internal standards.
Gas chromatography and mass spectrometric conditions for DEG and EG An Agilent chromatograph model 6890þ system was used with a Phenomenex ZB-5 ms column (30 m 0.25 mm ID 0.50 mm film) for the chromatographic separation of DEG and EG, and their corresponding internal standards. Helium carrier gas was used in constant flow mode. Derivatized urine samples were injected using a 2.0-mL split injection in a 2758C injection port (helium carrier pressure 11.95 psi; split flow 11.1 mL/ min; split ratio 10:1). Derivatized serum and CSF samples were injected using a 2.0-mL pulsed splitless injection in a 2758C injection port (helium carrier pressure 11.95 psi; pulse pressure 25 psi; pulse time 0.3 min; purge flow 50 mL/min; purge time 0.25 min). The temperature gradient started at 1008C held for 0.5 min and then linearly increased at 208C/min to 3208C and held for 1.0 min. MS detection was performed in the NCI mode using methane gas on an Agilent model 5973N single quadrupole mass spectrometer operated with the following parameters: Transfer line temperature ¼ 3008C, MS quadrupole temperature ¼ 1808C and MS source temperature ¼ 2008C. The following ions were used quantitation with the MS operated in the SIM mode: DEG, m/z 494.0; D8-DEG m/z 502.0; EG, m/z 450.0; D6-EG, m/z 454.0 with dwell time of 100 ms for all ions and internal standards.
Data analysis Integration of Selected-Ion Monitoring mass spectral datafiles SIM datafiles collected on both the Agilent G1956B LC/MSD and the Agilent 5973B mass spectrometers were converted to “.wiff”
datafiles, compatible with the AB Sciex Analyst v1.5 software (Framingham, MA, USA). This conversion was done with the “Translat.exe” utility, available with the Analyst software package. The resulting Analyst-format SIM-chromatogram files were processed with Analyst v1.5 to obtain chromatographic peak areas for the analytes and respective internal standards. Chromatograms obtained from the IC-suppressed conductivity detector were collected and displayed with the Dionex Chromeleon software (v 6.8). Linearity Linearity of calibration was tested by analyzing solvent calibration standards or matrix standards prepared above. Calibration curves in the concentration range for all analytes were constructed by plotting the corresponding peak area ratios of analyte/internal standard versus the concentrations. Weighted (1/x) linear regression analysis, available in the Analyst v1.5 application, was used to determine the slope, intercept and correlation coefficient (R 2) for all calibration curves. Method quantitation limits The limit of quantitation (LOQ) was set at the concentration below which the method could not operate with acceptable precision and accuracy (+15% of original value), and at which the signal-to-noise ratio was .8.
Results and discussion Chromatography and selectivity The samples from the DEG acute toxicity study in rats contained relatively high concentrations of metabolites, as the dose concentrations employed were 2 – 10 g/kg body weight (5). The IC – MS method was therefore utilized for quantitation of acid metabolites in these rat study samples because of ease of sample preparation, as the analytes required no derivatization prior to analysis. In addition, the simultaneous use of suppressed conductivity detection allowed for the qualitative analysis of any other substantial peaks that arise from DEG, beyond the targeted acidic metabolites of HEAA, DGA, GA and OA. Preliminary IC – MS analysis of HEAA, DGA, GA and OA in 50-fold diluted urine samples afforded detection limits of 1 – 5 mg/mL (data not shown). However, matrix suppression of the GA analyte and corresponding internal standard required further dilution to 500-fold to obtain acceptable recovery of GA from the urine matrix. For quantitation of acid metabolites in blood and tissue samples, the samples were first mixed with acetonitrile to precipitate proteins followed by a high-speed centrifuge and final dilution in Milli-Q water. Total dilution factors for blood and tissue samples were 50-fold. For human samples, the IC – MS method did not have sufficiently low detection limits to quantitate the acid metabolites in various matrices (Table I). To support the analysis of human poisoning-case samples (6), a GC – MS method based on our previous technique for GA and OA analysis (7) was used for analysis of these metabolites, as well as the additional DEG metabolites of HEAA and DGA, in serum, urine and CSF samples, as the PFBCl derivatives. GC –MS was also used for DEG and EG analysis for all rat and human matrices. Sample preparation was slightly modified between the rat and human methods, which involved removing
Table I Limits of Quantitation Analyte
Limits of quantitation (mg/g or mL) Rat matrices
DEG EG GA OA DGA HEAA
Human matrices
Blood
Urine
Tissue
Serum
Urine
CSF
5.02 5.02 15.6a 3.17a 1.6a 3.88a
10.1 50.0 50.5a 25.6a 25.7a 25.3a
2.06 5.14 25.7a 10.7a 2a 2a
0.0203 0.0201 0.51 1.03 10.1 0.513
0.0507 0.101 0.255 – 1.26 0.257
0.0507 0.101 0.51 – 1.01 0.513
a
Analysis conducted by IC –MS (all others via GC –MS).
the sodium chloride from the diluent water of the human urine, serum and spinal fluid samples. Quantitation limits were better for human samples, but the major factor contributing to this was the larger sample size that was available for analysis. Typical SIM ion chromatograms for IC–MS analysis of GA, OA, HEAA and DGA are shown in Figure 1; the overall chromatography run times were within 55 min. Under the IC – MS conditions used here, the analytes of OA, GA, HEAA and DGA have retention times of 13.6, 14.2, 35.5 and 43.9 min, respectively (Figure 1A). The corresponding internal standards for each analyte have almost the same retention times as the analytes (Figure 1B). Control water did not show a SIM response at the retention time of DGA (Figure 2). A minor, broad peak (11–16 min) was present in blank water that coeluted with the well-resolved peaks for HEAA and GA. This interference was 20 –100 lower than the stated LOQ for these analytes in the various rat matrices, and did not impact the precision of the methods, as shown in the standard refits, discussed below. A second minor peak was present in blank water that coeluted with OA. Again, this peak was 3–25 lower than the stated LOQ for OA in the various rat matrices, and did not impact method precision. Typical SIM ion chromatograms for GC – MS analysis of the t-butyldimethylsilyl derivatives of GA, OA, HEAA and DGA are shown in Figure 3; the overall chromatography run times were within 13 min. Under the GC–EI-MS conditions used here, the analytes of GA, OA, HEAA and DGA have retention times of 9.1, 9.1, 10.93 and 11.5 min, respectively (Figure 3A–D). The corresponding internal standards for each analyte have almost the same retention times as the analytes (Figure 3E–H) and no isotopic cross-over was observed. Control serum contained low, endogenous concentrations of GA, OA and DGA, but no detectable HEAA. Typical SIM ion chromatograms of DEG and EG are shown in Figure 4; the overall chromatography run times were within 13 min. Under the GC – NCI-MS conditions used here, the analytes of DEG and EG have retention times of 8.3 and 9.6 min (Figure 4A). The corresponding internal standards for each analyte have almost the same retention times as the analytes (Figure 4B) and no isotopic cross-over was observed. Control serum did not show a SIM response for any of those analytes, as shown in Figure 5. IC – MS or GC – MS spectra of the various analytes and internal standards are shown in Figures 6 –8.
Calibration curve, linearity and limit of quantitation A series of solvent standards were prepared for acidic metabolite analysis by IC –MS, or matrix calibration standards for the GC–MS
Quantitation of Diethylene Glycol and Its Metabolites by GC– MS or IC– MS in Rat and Human Biological Samples 187
Figure 1. Example summed, selected-ion IC–MS chromatograms of an acid metabolites standard containing 1 mg/mL of each analyte (GA, OA, DGA and HEAA) and 1 mg/mL of each internal standard (13C-GA, 13C2-OA, D4-DGA and D6-HEAA).
Figure 2. Example summed, selected-ion IC–MS chromatograms of a water blank containing 1 mg/mL of each internal standard (13C-GA, 13C2-OA, D4-DGA and D6-HEAA). 188 Perala et al.
Figure 3. Example selected-ion GC–MS chromatograms of an acid metabolites serum matrix standard containing 10 mg analyte/mL of each analyte and 25 mg/mL of each internal standard (13C-GA, 13C2-OA, D4-DGA and D6-HEAA).
analyses, at generally 6 –8 concentrations with a fixed amount of their corresponding internal standards. These solutions were processed and analyzed at least three times on separate days.
Overall, the calibration curves obtained were linear over the concentration ranges required for the assays. The LOQ’s for each analyte can be seen in Table I. The method quantitation limits and
Quantitation of Diethylene Glycol and Its Metabolites by GC– MS or IC– MS in Rat and Human Biological Samples 189
Figure 4. Example selected-ion GC – MS chromatograms of a serum matrix standard containing DEG and EG (0.5 mg/mL of each analyte) and 2 mg/mL of D8-DEG and 4.4 mg/mL D6-EG internal standards.
Figure 5. Example selected-ion GC–MS chromatograms of control serum sample by GC–MS.
190 Perala et al.
Figure 6. Representative GC–NCI-MS mass spectra of DEG (A), D8-DEG (B), EG (C) and D6-EG (D) as PFBCl derivatives.
Figure 7. Representative IC–MS mass spectra of GA (A), 13C-GA (B), OA (C), 13C2-OA (D), DGA (E), D4-DGA (F), HEAA (G) and D6-HEAA (H).
Quantitation of Diethylene Glycol and Its Metabolites by GC– MS or IC– MS in Rat and Human Biological Samples 191
Figure 8. Representative GC–MS mass spectra GA (A), 13C-GA (B), OA (C), 13C2-OA (D), DGA (E), D4-DGA (F), HEAA (G) and D6-HEAA (H) þ MTBSTFA derivative.
accuracy were also acceptable for the analysis of rat or human samples, with standard refits to the calibration curve generally within 15% of nominal values (see Supplementary data).
Effect of matrix on recovery In the GC– MS assays, matrix standards were utilized so fortified samples were not prepared. In the IC –MS assays solvent standard along with fortified matrix spikes were used to determine acid metabolite concentrations in rat matrices. However, relative recoveries (analyte vs. internal standard) were quite acceptable, with matrix standard refits, as discussed above, generally within 15% of nominal values. Recovery of the four acidic metabolites via the IC –MS method was somewhat more variable ( +30%), but was sufficient for the determination of these analytes in this initial, acute toxicity study (Tables II–IV).
Advantages of various methods The IC –MS method for acid metabolite analysis showed acceptable chromatographic resolution with adequate sensitivity for the high concentrations of metabolites found in the rat acute toxicity study samples. The sample preparation for this method was quite rapid, requiring only dilution and blood protein precipitation prior to analysis. Incorporation of suppressed conductivity detection also allowed for screening of non-targeted analytes. The major disadvantages were long sample analysis times (.40 min per sample) and large dilutions of samples, which limited method quantitation limits. 192 Perala et al.
Table II Linearity for DEG and EG by GC –EI-MS Rat analyte DEG EG Human analyte DEG EG
Linearity range (mg/mL) 0.5 –100 5 –100 0.02 –2 0.02 –2
R2 0.9992 0.9988 0.9984 0.9989
Table III Linearity for Acid Metabolites by IC –MS Analyte
Linearity range (mg/mL)
R2
GA OA HEAA DGA
0.1 –25 0.05 –5 0.05 –50 0.05 –5
0.9999 0.9899 0.9952 0.9990
Table IV Linearity for Acid Metabolites by GC –NCI-MS Analyte
Linearity range (mg/mL)
R2
GA OA HEAA DGA
0.05 –51 0.5 –25 0.5 –100 1 –100
0.9990 0.9907 0.9985 0.9922
The GC– MS methods for metabolite analysis for DEG/EG in all samples, and the acidic metabolites in the human poisoning-case samples showed acceptable chromatographic resolution with adequate sensitivity. The analysis run times (12 min per sample) were quick and minimum sample dilution required. The major disadvantages were a complex sample preparation (extraction/ derivatization) and somewhat noisier chromatography baselines vs. the IC –MS method.
Specificity The methods described in this manuscript are considered to be specific for the determination of DEG and the five potential metabolites (ED, GA, OA, HEAA and DGA) in the various biological matrices. In the GC –MS analysis, method additional confirmation ions were not incorporated into the assay, due to the lack of significant, compound-specific fragment ions for some analytes and based on the extremely low concentrations of metabolites (EG, GA, OA, HEAA and DGA) present in actual samples. The use of stable-isotope-labeled internal standards, however, confirmed analytes’ retention times for each sample analyzed. Also, the retention time and mass spectral response for each analyte was as a chemical derivative, providing further specificity to these analyses. Slight differences in retention times between the deuterated internal standard and the analyte are a common occurrence (8, 9).
Conclusion Analytical methods were developed for the analysis of DEG and five potential metabolites in biological matrices of rat and human. The GC –MS method used for human sample analysis is a modification of previously published techniques for EG, GA and OA. The IC – MS method developed for the four acidic metabolites of DEG is novel and has been applied to rat blood, rat urine, rat tissues, human serum, human urine and human CSF. The GC – EI-MS method for acid metabolite analysis has better detection limits but requires more sample preparation vs. the IC – MS method. The IC –MS method is adequate for samples with higher concentrations of OA, GA, HEAA and GA, requiring very little sample preparation and affords screening of non-targeted analytes (via suppressed conductivity detection). The GC – NCI-MS analysis method for DEG and EG is robust and has acceptable detection limits. Future research will continue to use the GC– MS methods for DEG and its metabolites; however, we will explore the possibility of using GC with triple quadrupole mass spectrometry detection to improve selectivity, sensitivity and provide additional confirmation of analyte structure. These sensitivity improvements will support future toxicokinetic animal studies that may evaluate a dose-response in DEG metabolism.
Supplementary data Supplementary data are available at Journal of Analytical Toxicology online.
Funding Funding for this work was provided by the Ethylene Glycol/ Ethylene Oxide Panel of the American Chemistry Council and the U.S. Centers for Disease Control and Prevention.
Acknowledgment The authors thank Dan Markham for critical review of this manuscript.
References 1. Wax, P.M. (1995) Elixirs, diluents, and the passage of the 1938 Federal Food, Drug and Cosmetic Act. Annals of Internal Medicine, 122, 456– 461. 2. Wax, P.M. (1996) It’s happening again – another diethylene glycol mass poisoning. Journal of Toxicology – Clinical Toxicology, 34, 517– 520. 3. Barr, D.B., Barr, J.R., Weerasekera, G., Wamsley, J., Kalb, S.R., Sjodin, A. et al. (2007) Identification and quantification of diethylene glycol in pharmaceuticals implicated in poisoning epidemics: an historical laboratory perspective. Journal of Analytical Toxicology, 31, 295–303. 4. Mathews, J.M., Parker, M.K., Matthews, H.B. (1991) Metabolism and disposition of diethylene glycol in rat and dog. Drug Metabolism and Disposition: The Biological Fate of Chemicals, 19, 1066– 1070. 5. Besenhofer, L.M., Adegboyega, P.A., Bartels, M., Filary, M.J., Perala, A.W., McLaren, M.C. et al. (2010) Inhibition of metabolism of diethylene glycol prevents target organ toxicity in rats. Toxicological Sciences, 117, 25– 35. doi:10.1093/toxsci/kfq167. 6. Schier, J.G., Hunt, D.R., Perala, A., McMartin, K.E., Bartels, M.J., Lewis, L.S. et al. (2013) Characterizing concentrations of diethylene glycol and suspected metabolites in human serum, urine, and cerebrospinal fluid samples from the Panama DEG mass poisoning. Clinical Toxicology (Philadelphia), doi:10.3109/15563650.2013.850504. 7. Pottenger, L.H., Carney, E.W., Bartels, M.J. (2001) Dose-dependent nonlinear pharmacokinetics of ethylene glycol metabolites in pregnant (GD 10) and nonpregnant Sprague-Dawley rats following oral administration of ethylene glycol. Toxicological Sciences, 62, 10 –19. 8. Baillie, T.A. (1981) The use of stable isotopes in pharmacological research.Pharmacological Reviews, 33, 81 –132. 9. Turowski, M., Yamakawa, N., Meller, J., Kimata, K., Ikegami, T., Hosoya, K. et al. (2003) Deuterium isotope effects on hydrophobic interactions: the importance of dispersion interactions in the hydrophobic phase. Journal of the American Chemical Society, 125, 13836 – 49, doi:10.1021/ja036006 g.
Quantitation of Diethylene Glycol and Its Metabolites by GC– MS or IC– MS in Rat and Human Biological Samples 193
