Toxicology Drug Screening in Blood by CE-Msn - a Feasibility Study Application

Toxicology Drug Screening in Blood by CE-MSn - a Feasibility Study Application

Forensics Toxicology

Authors analytical procedure, such as mass spectrometry Doug Knisley, Mark Hetherington, Gordon McKay, and (MS), often involving repeat extraction and Murray Malcom derivatization. Analysis is more efficient if both Pharmalytics, Inc. detection and identification can be completed as Saskatoon part of the same analytical run or, at least, using Saskatchewan the same instrumentation. Canada It was reported that CE-MS is useful in the analysis of drugs in samples of illicit drug Abstract materials [2]. An earlier application note showed that CE-MSn is capable of separating and detecting Preliminary observations indicate the feasibility of certain drugs with sufficient sensitivity to be capillary electrophoresis-mass spectrometry (CE-MS) useful as a possible screening instrument in whole with the Agilent LC/MSD Trap system to be capable of blood analysis [3]. Initial observations were made detecting drugs at <20 ng/mL in whole blood, suggesting using pure drug solutions. CE-MS can only be that it may be a suitable tool for screening whole blood useful in the screening of whole blood, however, if samples for drugs. Additionally, the capability of generat- residues from the matrix do not interfere. The pre- ing MSn data permits reliable identification of detected sent application note describes the analysis of cer- drugs, suggesting that both forensic screening and tain drugs in samples of whole blood, both spiked identification of drugs might be accomplished in one or samples and those from actual toxicology cases. two injections using the same instrument system. In the report of Hudson et al [1], it was noted that the combination of electrophoretic mobility and Introduction the UV spectrum provided an impressive degree of discrimination between analytes. This combina- Capillary electrophoresis (CE) with ultraviolet- tion, however, was not claimed to represent rigor- diode array detection (UV-DAD) is reported to be ous identification. Rather, drugs tentatively an effective tool in the screening of whole blood identified by mobility and UV absorption would be samples for drugs of forensic toxicological subjected to further analysis, probably by gas interest [1]. The first objective of any such screen chromatography/mass spectrometry (GC/MS). This is simply the detection of drugs at toxicologically note shows the feasibility of extending the CZE significant concentrations. Since the primary focus (Capillary Zone Electrophoresis) screening proce- is on detection, drug identification is often dure put forward by Hudson et al [1] to include tentative at the screening stage. Rigorous the collection of +MS, +MS2 and +MS3 data during a identification is frequently deferred to a second single screening run and the subsequent use of the MSn data as evidence of drug identification. Experimental ammonia + 5.0-mL 1-chlorobutane; shake; centrifuge; and evaporate to dryness. To the dry All drug analyses were done using an Agilent residue, add 30 μL of 10-mmol/L phosphate buffer; G1600A 3D CE coupled to an Agilent LC/MSD Trap vortex; centrifuge briefly; transfer to sample cups XCT, using the G1603A CE-MS adapter kit and the and centrifuge again. Each residue was then G1607A CE-MS sprayer. screened by CE-MSn, using the above analytical conditions. For each peak detected, MSn data were CE conditions collected. Capillary: Bare fused silica; 50-μm diameter Capillary length: 21.5 cm to DAD; 84.0 cm to LC/MSD Samples of whole human blood from three actual Trap toxicology cases were previously extracted and Cassette temperature: 25 °C analyzed by ELISA, GC-NPD, GC-ECD, and Run buffer: 100 mmol/L phosphate at pH 2.38 CE-DAD. Drugs detected in the blood samples were Injection: Electrokinetic, 12.0 kV for 16.0 s identified by GC/MS. The dry residue from each of Run voltage: Ramped 0 to 20 kV in 0.15 s; held for these samples was reconstituted and analyzed by n duration of run CE-MS , as above. Run time: 25 min No attempt was made to optimize separation con- Diode array: Wavelength: 200 nm, bandwidth 4 nm ditions in this work. The extraction residues left Reference wavelength: 375 nm, bandwidth 75 nm over from previous analyses were simply analyzed LC/MSD Trap conditions under conditions that approximated those Mass range mode: Ultra Scan reported by Hudson [1]. We observed (and report Ion polarity: Positive here) the coelution of certain drugs that Hudson et al Ion source type: ESI were able to separate. As shown below, how- n Drying gas temp: 130 °C ever, the ability to collect MS data makes these coelutions a less serious problem than might be Drying gas flow: 7.00 L/min supposed since the added selectivity of MSn over- Nebulizer: 8–12 psi comes coelution problems observed with the Trap drive: 27.0 less-selective DAD. Skim 1: 40.0 V Skim 2: –5.0 V Octopole RF amplitude: 131.2 V Results and Discussion Capillary exit: 97.0 V Figure 1 shows the total ion electropherogram Scan range: Typically 50–500 (TIE) of the spiked porcine blood sample, along Averages: 8 spectra with the extracted ion electropherograms (EIEs) Max. accumulation time: 100000 μs for each of the 17 drugs present. These data give ICC target: 100000 preliminary indications on two important points: Charge control: On sensitivity of detection and interference from Sheath liquid: 0.5 % Formic acid in matrix. 50/50 methanol/water Initially, it was thought possible that the analytical Sheath liquid flow: 7.5 μL/min (0.75 mL/min split 100:1) system used here would simply not be sensitive enough for a toxicological screen. From the data Auto MS parameters shown, however, it is clear that useful MS data can Auto MS3:On be collected at drug concentrations of 20 ng/mL in 3 Threshold auto MS : 2500 whole blood. It also appears that a thorough vali- No. of precursors: 1 dation study would show that the limit of detection Fragmentation amplitude: 1.0 V (LOD) for at least some drugs would be substan- Isolation width: 4.0 m/z tially lower than 20 ng/mL. This is probably adequate sensitivity for routine drug screening.

Whole porcine blood spiked with 17 drugs, each at It was also considered possible that matrix compo- a concentration of 20 ng/mL of blood was previ- nents extracted from whole blood would over- ously extracted according to the procedure of whelm the analytical system and interfere Hudson [1]. Details are given in the cited reference irreversibly with collection of the MS data. Figure 1 but, briefly, basic drugs were extracted as follows: suggests that matrix components offer no serious 1.0-mL whole blood + 0.2-mL concentrated obstacles to MS analysis.

2 2.0

TIE 1.8 6 10 × 1.6 Intensity

1.4

1.2 0.0 5.0 10.0 15.0 20.0 25.0 Time [min]

0.6 5 5 1.5 0.5 EIE 300 10 10 EIE 241 × × 1.0 Pheniramine 0.4 Codeine 0.3 0.5 0.2 Intensity Intensity 0.0 0.1 5

1.0 EIE 275 5 10 EIE 375 × 4 Chlorpheniramine 10 × 0.5 Hydroxyzine 2 Intensity

0.0 Intensity

4 EIE 319 10 3.0 3 × Brompheniramine 5 EIE 286 10

× 2 Pentazocine 1.5 Intensity 1 Intensity 4 6 EIE 136 0 10 × Amphetamine 3 5 EIE 268

3 10

× 2 Metoprolol Intensity 1 5 2

10 EIE 150 Intensity × Methamphetamine 0 1

4 6 EIE 372 Intensity 10

0 × 4 Trazodone 0

10 EIE 477 2 × 2 Loperamide Intensity 0

Intensity 1 4 6 EIE 376 10 × 5 4 Haloperidol 2 10 EIE 166 Ephedrine and × Pseudoephedrine 2 Intensity 1 Intensity 1.5 4

5 EIE 455 10 × 10 EIE 256 1.0 Verapamil × 4 Diphenhydramine 0.5 2 Intensity

Intensity 0.0 0

5 4 EIE 272 5 1.5 EIE 212 10 10 × Dextromethorphan × 1.0 Methoxamine 2 0.5

Intensity 0 Intensity 0.0 0 5 10 15 20 25 0101520255 Time [min] Time [min]

Figure 1. TIE and EIEs for extracted drug mix.

3 In Case 1, the blood sample was known from previous analyses to contain amphetamine, methamphetamine, and cocaine. Figure 2 shows the TIE and EIEs; Figure 3 shows Auto MSn and library matches from the same sample.

Figure 2 shows one matrix peak (unidentified). As was suggested above, matrix interference does not appear to be a significant problem.

6 TIE 5 Amphetamine + methamphetamine

6 Cocaine

10 4 ×

Intensity Unidentified 2 matrix peak

5 EIE 136 4 5

10 3 ×

2 Intensity

3 EIE 150 6

10 2 ×

Intensity 1

4 EIE 304

6 3 10 ×

2 Intensity

0 010155 20 25 30 Time [min]

Figure 2. Case 1. Amphetamine, methamphetamine, and cocaine.

4 The library searches shown in Figure 3 indicate the correct identification of amphetamine, methamphetamine, and cocaine, in spite of the coelution of the former two compounds. The library search on MS3 data from Peaks 1 and 2 (amphetamine and methamphetamine, respec- tively) showed no matching spectrum. The dominant ion in the library MS3 data from both these compounds was 91.3 m/z. Our analysis showed a dominant ion at m/z 92.2, which was, of course, not interpreted as a match by the search software.

2 4 3 6 3 Auto MS3 10

× X 2 1 1 Intensity

0 0 2 4 6 8 10 12 14 16 18 20 Time [min]

Peak 1 Peak 2 Peak 3

1.5 2.0 5 6

6 +MS 5 +MS 149.9 10 +MS 304.0 136.0 × 10

10 1.5

1.0 × × 119.1 3 1.0 0.5 225.3 0.5 Intensity 0 158.7 Intensity

Intensity 119.0 5 0.0 0.0 1250 10 Library: Cocaine 304.0 × 5 1250 Library: Amphetamine 6 1250 Library: Methamphetamine 750 10 10

× 136.1 × 150.1 750 119.1 750 250 Intensity

5 5 2 Intensity 250 91.4 158.0 181.9 +MS (304.1) Intensity 250 119.1 10 × 3.0 3

5 119.0 5 6 304.0 2 2 1 108.1 150.0 119.0 +MS (136.2) 10 +MS (150.0) 10 Intensity 2.0 × × 4 5 Library: Cocaine 1.0 10 1250 181.9 2 × Intensity

Intensity 750 0 0.0 250 Intensity

5 108.1 150.0 304.0 5 1250 Library: Amphetamine 1250 Library: Methamphetamine 10 4

10 119.1 119.1 × 3 × +MS (304.1→181.9) 750 10 4 750 91.3 × 149.9 108.1 2 Intensity 250 122.1 Intensity 250 91.3 93.1 Intensity 2 0 3 3 → 5 3 3 → +MS (136.2 119.1) 92.2 +MS (150.0 119.1) 4

10 92.2 10

× 1250 Library: Cocaine 10 × 149.9 1 2 × 750 108.1 1

Intensity 119.0 0 107.1 Intensity 250 93.2 Intensity

100 140 180 100 140 180 100 150 200 250 m/z m/z m/z

Figure 3. Case 1. AutoMS3 and library matches.

5 In Case 2, the blood sample was known to contain The library searches shown in Figure 5 identified methylenedioxyamphetamine (MDA) and methyl- both drugs correctly. It is interesting to note, how- enedioxymethamphetamine (MDMA). Figure 4 ever, that the search of MS data on MDA failed to shows the TIE and EIEs. Figure 5 shows the Auto find a match but that the correct match was found MS3 and library matches from the same sample. for MS2 and MS3 data. This was attributed to the presence of background data, such as the frag- Figure 4 shows again the relative absence of ments at m/z 91.1 and 158.8. In the analysis of matrix interference and the coelution of the two MDMA, these background fragments were removed analytes of interest. and correct matches were found on all three searches.

2.0

6 TIE 1.8 10 × 1.6 1.4

Intensity 1.2 1.0 8 1 3 6 EIC 179.9 10 × 4

Intensity 2

0 6 5 2 3 EIC 193.9

10 4 × 3 2 Intensity 1 0 5.0 10.0 15.0 20.0 Time [min]

Figure 4. Case 2: MDA and MDMA.

6 2 2.0

1.5 6 10 ×

1.0 Intensity

0.5

0.0 2 2 4 6 8 10 12 14 Time [min]

Peak 1 Peak 2 2.0

2.0 +MS 5 +MS 193.9 5 91.1 158.8 1.5 10 × 10 × 1.0 1.0 0.5 164.8 Intensity 162.9 Intensity 142.8 180.0 0.0 121.0 0.0

5 1250 Library: MDMA 193.9 10

6 × 3 +MS2(180.0) 750

10 162.9 × 4 Intensity 250 91.1 162.9 2 Intensity 1.50 +MS2(194.0)

180.7 5 162.9 0 10 × 1.00

1250 0.50 2 162.9 Library: MDA Intensity 10

× 134.9 0.00 750 Library: MDMA

5 1250

Intensity 162.9 10

250 × 750 8 +MS3(180.0→162.8) Intensity 250 134.9 2 6 135.0 10 × 106.3 4 +MS3(194.0→163.0) 4 5 134.9 3 10 ×

Intensity 2 150.7 2 105.1 0 1 Intensity 0 1250

2 Library: MDA 134.9 Library: MDMA 5 10 1250 × 134.9 10

750 × 750 105.1

Intensity 105.1

250 Intensity 250

60 100 140 180 220 60 100 140 180 220 m/z m/z

Figure 5 Case 2. Auto MS3 and library matches.

7 In Case 3, initial analysis of the blood sample by Second, erythrohydrobupropion and threohy- Hudson et al showed several forensically signifi- drobupropion are known to be major metabolites cant compounds. Quinine, cocaine, cocaethylene, of bupropion [4]. While Figure 6 suggests that bupropion, paroxetine, and lidocaine were identi- these two compounds are potentially separable fied by GC/MS. As before, our analysis of this (Peaks 8 and 9), no attempt was made here to fur- sample attempted to replicate this work, given dif- ther resolve the system. As it was, the compounds ferent instrument configurations. Figure 6 shows were not separated by Auto MS3 and, accordingly, the TIE for this sample. The library matches from Table 1 shows an entry only for Peak 8 thus indi- the Auto MS3 analysis are shown in Table 1. There cating that the two compounds were indistinguish- are several points to note in the following data. able on the basis of MS data. None of the metabolites of bupropion were previously First, for a variety of reasons, Peaks 2, 3, 4, and 5 identified by GC/MS. were not included in Table 1. Peak 2 appeared upon later searching to be methylecgonine (see Finally, with Peak 6, the known drug was not iden- below) and Peak 3 appeared to be a matrix peak tified on the library search of MS data because of (unidentified); Peak 4 was consistent with the the presence of background fragments. As with the ISTD (methoxamine) and Peak 5 was consistent MDA/MDMA example in Case 2, however, identifi- with the lidocaine metabolite, monoethylglycinexy- cation was correct on the searches of MS2 and MS3 lidide (MEGX). data.

8 8

Key: 1 Quinine/quinidine 6 2 methylecgonine 3 Unknown 4 Methoxamine (IS)

6 9 5 MEGX

10 10

× 6 Cocaine 7 Cocaethylene 4 8 Bupropion metabolite

Intensity 9 Bupropion metabolite 10 OH-bupropion 7

1 3 2 4 2 5 6

0.0 5.0 10.0 15.0 20.0 25.0 Time [min]

Figure 6. TIE of Case 3.

8 Table 1. Library Searches of Auto MS from Case 3 Identification +MS +MS2 +MS3 Fit Rfit Purity Identification Peak 1 325.2 307.1 Library search 325.1 940 669 645 Quinine/Quinidine 307.2 975 975 972 Quinine/Quinidine NC Peak 6 226.8 304.1 181.9 149.9 (Bkgd) Library search 304.1 No match 181.9 995 995 995 Cocaine 149.9 930 933 930 Cocaine Peak 7 318.1 195.9 149.9 Library search 318.1 994 992 992 Cocaethylene 195.9 997 997 997 Cocaethylene 149.9 997 998 997 Cocaethylene Peak 8 241.9 167.9 150.9 Library search 242.0 995 995 994 Amino OH bupropion* 167.9 997 997 997 Amino OH bupropion* 150.9 993 993 993 Amino OH bupropion* Peak 10 256.0 237.9 138.9 Library search 256.0 994 998 992 Hydroxybupropion 237.9 998 998 998 Hydroxybupropion 138.9 994 994 994 Hydroxybupropion

NC Not completed. *Metabolites of bupropion exist in erythro- and threo- forms, not completely resolved here.

The AutoMS3 analysis shown above did not detect Conclusions certain compounds, such as bupropion, that were known to be present. Whether a particular peak is CE-MS with the Agilent LC/MSD Trap system is detected by AutoMS3 or not depends upon several capable of detecting drugs at concentrations less factors, chief among them the AutoMSn threshold than 20 ng/mL in whole blood. It is, therefore, a that is set and the resolution that is possible with suitable tool for screening whole blood samples for the electrophoretic conditions used. Note that it is drugs as part of routine toxicological analyses. In possible to prepare an Include List specifying pre- addition to relatively sensitive detection, however, cursor ions that the analyst specifically wants to the capability of generating MSn data permits reli- search for. That was not done in this work. able identification of detected drugs. This means that both screening and identification of drugs However, later manual searching of the electro- might be accomplished in one or two injections in pherogram (Figure 6) on the basis of selected a single instrument system. Finally, we show that, masses indicated the presence of lidocaine and its through the use of the Auto MSn feature, it may be metabolite, MEGX, in the region of Peak 5. Bupro- feasible to automate much of the drug screening pion, metabolites of which were detected by procedure, thereby potentially increasing sample AutoMS3 in Peaks 8 and 10, appeared to be part of throughput significantly. We emphasize that these the unresolved complex at Peak 8. While both are preliminary observations indicating feasibility cocaine and the metabolite, cocaethylene, were only. More development and validation work is detected by AutoMS3, methylecgonine was not. A required. search based on its mass showed methylecgonine likely to be present in Peak 2. It was not detected in the AutoMS3 analysis because we had no entry for it in our library.

9 www.agilent.com/chem References 1. J. C. Hudson, M. Golin, M. Malcolm, and C. F. Whiting, (1998), Can. Soc. Forens. Sci. J. 31: 1–29. 2. D.T. Phan and P.B. Harrsch, Agilent Technologies, publication 5968-9221E www.agilent.com/chem 3. D. Knisley, M. Hetherington, and G. McKay, (Pharmalytics, Inc.), Analysis of Drugs by CE-MSn, Agilent Technologies, publication 5989-2910EN www.agilent.com/chem 4. Compendium of Pharmaceuticals and Special- ties, 35th edition, Canadian Pharmacists Association, Ottawa, ON, 2000.

Acknowledgements

The authors acknowledge J. Hudson for the gift of extraction residues, and the staff of RCMP Foren- sic Laboratory Services laboratories in Regina, Saskatchewan and Winnipeg, Manitoba for their cooperation and helpful advice.

For More Information

For more information on our products and services, visit our Web site at www.agilent.com/chem.

Agilent shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material.

Information, descriptions, and specifications in this publication are subject to change without notice.

Printed in the USA 6j\jhi'&!'%&* 5989-2911EN

For Forensic Use.