|

 

Tera.chem.ut.ee

Journal of Food and Drug Analysis, Vol. 10, No. 2, 2002, Pages 112-119 Determination of Sildenafil Citrate Adulterated in a
Dietary Supplement Capsule by LC/MS/MS
MU-CHANG TSENG AND JER-HUEI LIN
National Laboratories of Foods and Drugs, Department of Health, Executive Yuan 161-2, Kuen Yang Street, Nankang 115, Taipei, Taiwan, R.O.C. (Received: April 18, 2001; Accepted: July 31, 2001) A liquid chromatography/tandem mass spectrometry (LC/MS/MS) method was used to identify and quantify an adulterated substance, in a dietary supplement capsule. The capsule is claimed to be an extract of animal organs and traditional Chinese herbs and indicated forenhancing sexual activity. At first, the sample was extracted with 50% v/v methanol in water and the extract was injected directly into theLC/MS/MS with no separation and a full scan of positive ion eletrospray (ESI +) analysis was performed. The full scan mass spectrum wascompared with the NLFD3/LM library created by Division 3rd of National Laboratories of Foods and Drugs for LC/MS/MS analysis. Asildenafil specific mass spectrum was matched indicating that the capsule may have been adulterated with sildenafil, the active ingredient of VIAGRA . Furthermore, sildenafil was confirmed by carrying out a daughter ion scan and a parent ion scan. The daughter ion spec-trum was compared with the NLFD3/LM library. A negative ion eletrospray (ESI –) analysis was performed for confirming the citrate saltin the capsule. The results indicated that the adulterated ingredient was sildenafil citrate. For quantitative analysis of the sildenafil citrate, multiple reaction monitoring (MRM) was performed. Pirenzepine 2HCl was used as the internal standard. A series of standards solution and sample solution with internal standard were analyzed by LC/ MS/MS. The analyt-ical condition is as following: column: Cosmosil 5C18-AR, 4.6 x 150 mm; mobile phase:methanol:acetonitrile:1% acetic acid (25:17:58);capillary voltage: 3 kV;cone voltage: 80 V; collision energy: 25 eV; source temperature: 120˚C; desolvation temperature: 350˚C. Thedetection limit was 40 ng/mL.
An excellent calibration curve of sildenafil citrate was obtained with the γ2 correlation coefficient of 0.9977. The RSDs of interday and intraday were between 7.56 7.75% and 9.09 11.25%, respectively. The analytical strategy and method is suitable for identification andmeasurement of sildenafil citrate adulterated in capsule. Amount of sildenafil citrate in the capsule was determined as 42.8 mg/capsule. Key words: LC/MS/MS, sildenafil citrate, ESI., MRM, pirenzepine, daughter ion scan, parent ion scan ity and difficulty to ionize. Although HPLC has a good sepa-ration effect, the peaks presented could not provide any direct There are numerous synthetic medicines in the world information about the unknown ingredients adulterated in the and it is very difficult to perform systematic analyses to sample. By using LC/MS to compensate the insufficiency of reveal those compounds if sample mixtures contain unknown GC/MS and HPLC, and applying Tandem mass(LC/MS/MS) ingredients without any information. It is even more difficult to quantify and qualify the unknown ingredients, the process to screen out unknown compounds adulterated in natural is faster and the result is more precise. The present study products such as traditional Chinese medicine (TCM), due to examined samples of food supplement capsule made from the interference caused by the contents of natural compounds animal organ extracts, indicated for enhancing sexuality.
during analysis. Thin Layer Chromatography(TLC) is nor- Since the claimed effect was quite extraordinary, a consumer mally used for separating the target compounds in cases such submitted the sample to the NLFD for inspection. We used as the above and to detect the UV spectrum(1,2) of each TLC LC/MS/MS and by comparing the MS1 mass spectrum with spots. However, ingredients with analogous molecular struc- a library search, found traces of possible adulteration. After ture are of similar UV spectrum and may cause confusion for performing the daughter ion scan, we were able to identify inspection purpose. For many compounds with weak UV the adulterants. No complicated separation and purification absorption, it is difficult to identify them using TLC/UV were required during the analysis processes.
methods. In recent years, the GC/MS analysis method hasgradually become popular. By library searching in massspectrum and comparing the peak of retention time with MATERIALS AND METHODS
library data provided by GC/MS, a more precise inspectionresult can be acquired(3). However, there are limitations to I. Materials GC/MS analysis. For example, GC/MS cannot be adopted inanalyzing compounds of high molecule weight, heat instabil- (I) Reference Standard: sildenafil citrate provided for regis-tration and inspection on ViagraR by Pfizer Pharmaceutical * Author for correspondence. * Author for correspondence. Tel:02-26531239; (Taiwan); pirenzepine 2HCl from Sigma (St. Louis, USA).
Journal of Food and Drug Analysis, Vol. 10, No. 2, 2002 Table 1. Parameters of LC/MS screening scan methods by direct sam-
pling for identification of unknown sample Carrier Flow (mL/min) 100 125 150 175 200 225 250 275 300 325 350 375 400 425 450 475 500 525 550 575 600 Source Temp (˚C) Desolvation Temp (˚C) (II) Reagents: LC grade methanol from Labscan (Ireland), acetonitrile from BDH laboratory (England) and acetic acidfrom E-Merck (Germany).
(III) Model TCM sample: Huan Shao Tan from Shen Chung 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 Daughters of 475ES+ (IV) Sample: sample is pink and cream color capsule, con-taining dark brown powder with average weight at 375 (V) Filter: 0.45 µ m Millipore filter film from Gelman Sciences (U.S.A.).
191 Daughters of 191ES- (I) Tandem mass: Micromass Quattro Ultima LC/MS/MS (II) HPLC: Waters 2690 Alliance LC & 996 PDA with 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 Automatic Liquid Sampler and Injector (U.S.A.).
Figure 1. The MS1 mass spectrum (A&B) and daughter ion spectrum
of sample solution (C&D).
(III) Software: MassLynx NT Quattro Data Acquisition(England).
(IV) Mass spectrum data library: NLFD3/LM Library creat- through filter paper and extracted the residue three times.
ed by Division 3rd of National Laboratories of Foods and Collected the filtrate in a 100 mL volumetric flask and added Drugs for LC/MS/MS analysis.
solvent to the mark. Pipette 2 mL of the filtrate and diluted to50 mL with solvent and filtered through Millipore filter to be III. Identification of Unknown Ingredients in the Sample(4 11) used as sample solution.
(I) Preparation of sample solution: Accurately weighed sam- (II) Screening unknown ingredients with MS1 scan: Directly ple powder equivalent to one capsule, put in a 50 mL cen- injected 10 µ L sample solution(5) from Divert/Injection trifugal tube and extracted with 30 mL of 50% v/v methanol.
valve, performed both positive and negative ion electrospray Shook in the ultrasonic shaker for 10 minutes and centrifuged ionization (ESI+ & ESI–) MS1 scan mode based the analysis under 3000 rpm for 10 minutes. Filtrated the supernatant conditions on Table 1 and acquired a total ion current (TIC) Table 2. Parameters of LC/MS/MS daughter ion and parent ion scan by direct sampling for identification of sildenafil citrate
Selected ion (m/z) Daughter ion of 475 Parent ion of 100 Daughter ion of 191 Carrier Flow (mL/min) Collision (eV): Argon Source Temp (˚C) Desolvation Temp (˚C) Journal of Food and Drug Analysis, Vol. 10, No. 2, 2002 chromatogram. After deducting the noise signal process, the (III) Analysis Condition of HPLC and Photodiode Array (11): mass spectrum (Figure 1A&B) was then compared with theNLFD3/LM library database.
1. Column: Cosmosil 5C18-AR, 4.6 x 150 mm.
2. Mobile phase: CH3OH: CH3CN: 1%HOAc (25: 17: 58).
(III) Reconfirming the ingredients by LS/MS/MS analysis: 3. Flow rate: 0.5 mL/min (Split ratio: 2/5).
Directly injected 10 µL of the above sample solution to per- 4. Injection volume: 10 µL.
form ESI+ daughter ion scan (Table 2). The base peak of frag- 5. Photodiode array: Scan range: 200 350 nm.
ment (m/z 475) revealed in Figure 1 was selected and a 6. Running time: 15 min.
daughter ion TIC chromatogram was acquired. After deduct-ing the noise signal process, a mass spectrum was obtained V. Evaluation of Recovery (Figure 1C) and compared with the NLFD3/LM library data-base. Took daughter ion fragment m/z 100 to perform ESI+ Took Huan Shao Tan model sample powder, adulterated parent ion scan and acquired result as the above to confirm sildenafil citrate reference standard and mixed thoroughly to whether it was produced from m/z 475 ion fragment (Figure make 10, 5, 3% w/w sildenafil citrate content in each sample (label as A, B, C adulterated samples). Accurately weighed100 mg of each group in 50 mL centrifugal tube. Prepared the (IV) Confirming organic salt: In order to confirm whether test solution by following the steps of III-(I), and then quan- sildenafil was combined with citrate, m/z 191 ion fragment titated the content of sildenafil citrate by following the above was selected to perform ESI– daughter ion scan following the method and calculated the recovery.
Table 2 conditions (Figure 1D). Compared the spectrum withthe NLFD3/LM library database.
VI. Evaluation of Precision IV. Quantitating the Content of Sildenafil Citrate in the Prepared 2.64 µg/mL and 13.20 µg/mL of sildenafil cit- Sample(4 10) rate standard solution, performed intra- and inter-day analy-sis each for five times by following the MRM quantitative (I) Preparation of standard solution and calibration curve: methods. Calculated the results by calibration equation Accurately weighed sildenafil citrate reference standard and derived from previous experiment and calculated the stan- dissolved with 50% v/v methanol to make a series standard dard deviation (SD) and relative standard deviation(RSD).
solution with concentrations at 1.76, 4.40, 8.80, 13.20, 17.60µg/mL, each containing 1.28 µg/mL pirenzepine 2HCl inter- VII. Detection Limit Test nal standard solution. Based on the following HPLC/Photodiode array conditions, Multiple Reaction Monitoring Took 1.76 µg/mL standard solution, diluted with 50% (MRM) analysis was performed with the combined analysis v/v methanol to 1, 5, 10, 25 and 50 times. Inspected with the condition in Table 2 and 3. Plotted the peak area ratio MRM quantitative method mentioned previously and between sildenafil citrate reference standard and the internal acquired individual signal/noise ratio (S/N) at the injection standard as the Y axis, and the concentration of standard volume of 10 µL. The detection limit was defined when the solution as the X axis to derive the calibration curve equation S/N ratio was 3.
of sildenafil citrate Y= m*X + b., the relative coefficientswere also calculated.
RESULTS AND DISCUSSION
(II) Preparation of sample solution and quantitating the con-tent of sildenafil citrate in the sample: Accurately weighed I. Source of NLFD3/LM Library Database & Establishment five portions of sample powder each equal to one capsule andplaced in 50 mL centrifugal tube, respectively. Extracted and The LC/MS/MS analysis technology was found to be filtered as III-(I) procedures. Pipetted 2 mL of the filtrate and very useful. For one compound, setting the parameter on the 2 mL of 32 µg/mL pirenzepine 2HCl internal standard solu- instrument under different ionization mode, capillary volt- tion, and added the solvent to make exact 50 mL. Filtered age, collision energy, cone voltage and source temperature through Millipore filter, used the filtrate as the sample solu- would result in different mass spectra. The analytic parame- tion. Performed MRM analysis as described previously to ters of LC/MS/MS are more complex than these of GC/MS.
determine the content of sildenafil citrate. GC/MS has already had library databases Wiley, PMW and Table 3. Parameters of MRM set for quantitative analysis of sildenafil citrate
sildenafil citrate Journal of Food and Drug Analysis, Vol. 10, No. 2, 2002 NIST ready to use, but not LC/MS/MS. In order to analyze 2: Daughters of 475ES+ unknown ingredients by LC/MS/MS, the important thing is to establish a mass spectra library database for every ingredi- ent. The mass spectra of sildenafil, citric acid, and piren-zepine were established by NLFD3/LM with the methods described as follows: (I) Standard Solution: Took the standard of the target ingredi- 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 ents, using appropriate solvents such as water, methanol or 2: Daughters of 191ES– 1% acetic acid etc to dissolve into a standard solution of 5 (II) Selection of ionization mode: Selected ESI+, ESI–, APCI+ or APCI– ionization mode based on the nature ofcompounds, and tuned the instrument by altering some of the instrument analytic parameters with the standard solutions in order to acquire an obvious (M+H)+ or (M-H)– ion fragment 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 and some cleaved fragments.
Parents of 100ES+ (III) Acquisition of MS1 mass spectrum: Adopted the para-meters from the above tune, injected 10 µL standard solutionfrom the divert injection valve, using 50% methanol as mobile phase to perform MSI scan and acquired the TICchromatogram. MS 1 mass spectra was acquired afterdeducting the noise signal processes. If the same spectrums were acquired after repeating the processes for three times, 100 125 150 175 200 225 250 275 300 325 350 375 400 425 450 475 500 525 550 575 600 the spectra would then be included into NLFD3/LM library with ionization mode, capillary voltage, collision energy, 1:Daughters of 352ES+ cone voltage, molecular weight and molecular formula.
(IV) Acquisition of MS/MS mass spectrum: Selected target fragment of (M+H)+ or (M-H)– in general and performeddaughter ion scan with different collision energy until the daughter ion spectrum was obvious and stable. If the same mass spectrums were achieved after repeating the processes 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 for three times, the mass spectrum would then be included Figure 2. The daughter ion spectrum of sildenafil(A), citrate(B), piren-
zepine(D) and parent ion spectrum of sildenafil(C).
into the NLFD3/LM library with ionization mode, capillaryvoltage, collision energy, cone voltage, molecular weight andmolecular formula.
After extracting with 50% methanol, the solution went II. Reason for Adopting LC/MS/MS Analysis through a MS 1 scan to acquire its spectrum (Figure 1A).
The matching quality of this result was 91.8% compared to When trying to screen out the unknown adulterants from that of sildenafil as recorded in the NLFD3/LM library data- TCM or food, the complexity of the natural ingredients base. we therefore suspected the sample was adulterated with would interfere with the results due to the enormous cate- sildenafil citrate. We further performed a daughter ion scan gories of the medications. The best method would be to on the fragment m/z 475 (M+H)+ of sildenafil to generate a adopt the MS analysis and by compare with the library data- daughter ion mass spectrum (Figure 1C) and compared with base. In this study, the sample claimed to be a food capsule the NLFD3/LM library database (Figure 2A). The matching with content extracted from animal organs for enhancing sex- quality was at 92.5%, which certified the adulterant was uality. Since the product label did not specify the content sildenafil. The result was further confirmed by performing a detail, in order to screen out the ingredients, the GC/MS parent ion scan which proved that the m/z 100 daughter ion method was adopted but failed. This was because the silde- was cleaved from the m/z 475 parent ion (Figure 2C) without nafil citrate could not be inspected under the routine GC/MS interference by other ingredients and could be used to set up analysis conditions(13) and therefore, LC/MS/MS was adopt- the MRM quantitative parameter. The accurate result was achieved by using only the 50% methanol, without the com-plicated extraction and purification process. The analytical III. Process of Screening Sildenafil Citrate time required was shortened as well.
Journal of Food and Drug Analysis, Vol. 10, No. 2, 2002 Figure 3. Fragmentation of citric acid by ESI– analysis.
Figure 5. Fragmentation of pirenzepine by ESI+ analysis.
sample solution with color, and it was difficult to identify thecitrate salt. Our study used LC/MS/MS to perform ESI– ionanalysis and chose (M-H)– ion fragment m/z 191 (C of the citrate acid to perform the daughter ion scan (Figures1D&2B. According to the general rules for predicting promi-nent fragmentation, one of the three carboxyl groups of cit- Figure 4. Fragmentation of sildenafil by ESI+ analysis.
rate acid will be easily eliminated and trigger chemicaldegradation. It could also easily eliminate water and form a IV. Identification of Citrate stable structure in the neighboring hydroxyl group afterdehydration and cyclization to acquire m/z 111 major frag- Through MS1 mass scanning, we suspected that the ment (Figure 3).
sample was adulterated with sildenafil citrate. As reportedby USP(12), Chp(13), and JP.(14), citrate was generally identi- V. Choosing Internal Standard fied by a chemical coloring reaction. However, this methodwas easily affected by the complicated natural ingredients or The best internal standard was isotope isomer of the Journal of Food and Drug Analysis, Vol. 10, No. 2, 2002 1: MRM of 2 Channels ES 1: MRM of 2 Channels ES 2.9': pirenzepine7.3': sildenafil citrate sildenafil citrate standard UV spectrum sample preparation UV spectrum Figure 6. The MRM chromatograms of standard (A), sample solution (B) and the UV spectrums of LC/MS/MS analysis with PDA detector (C&D).
Table 4. Recoveries of sildenafil citrate adulterated in TCM
Table 5. Intraday and interday precision analysis of sildenafil citrate
Adulterated amount Mean ± SD (RSD,%) standard (µg/mL) Mean ± S.D.
2.58 ± 0.20 13.36 ± 1.01 2.49 ± 0.28 12.76 ± 1.16 subject when carrying out the quantitative process(13). Sincethe compound was difficult to acquire in the market, we calibration curve was Y = 0.0957X + 0.0231, and the corre- chose pirenzepine instead as both sildenafil and pirenzepine lation coefficient γ2 was 0.9977.
could derive 4-methyl-piperazine during the cleavageprocess (Figures 4 & 5), and could produce intensity frag- VII. Evaluation of Recovery ment under the analytic conditions. In addition, both theanalysis subject and the internal standard while being collid- Three groups of model TCM sample were adulterated ed by argon in the collision cell of the LC/MS/MS, the m/z with sildenafil citrate. The results showed that the average 475 of sildenafil (M+H)+ would cleavage into 4-methyl- recovery was above 97.4% with relative standard deviation values at 3.0 50% (Table 4). Although the ingredients con- 5H12N2 ) fragment (m/z 100), when the frag- ment m/z 352 of pirenzepine (M+H)+ ion would cleavage tained in the sample may not be the same as the model TCM into 1-methylene-4-methyl-piperazine (C samples, the result indicated that quantitative and qualitative 6H13N2 , m/z 113) ion fragment (Figure 4). The fragments of the two com- sildenafil citrate by using LC/MS/MS could block out the pounds were distinctive and therefore, during the MRM disturbance caused by natural ingredients and could achieve quantitative process, the fragments would not interfere with reliable precision.
each other. Also under the current LC separation conditions,the retention time of sildenafil citrate and pirenzepine were at VIII. Precision 7.3 and 2.9 minutes (Figure 6). There was no mutual distur-bance problem.
The relative standard variations from intra- and inter- day testing were at 7.53 7.64% and 9.11 11.14%, respec- VI. Plotting the Calibration Curve tively. These results indicated that the precision was withinacceptable range (Table 5).
To quantify sildenafil citrate by MRM, the standard solution concentration range was at 1.76 17.60 µg/mL. The IX. Detection Limit Test Journal of Food and Drug Analysis, Vol. 10, No. 2, 2002 In this study, we found that the S/N ratio was at 2.7 J. H. 2001. Application of quantitative chemometric when the standard solution was diluted to 50 times (concen- analysis techniques to direct sampling mass spectrome- tration at 35.2 ng/mL). Should S/N ratio at 3 be the detection try. Anal. Chem. 73: 596-605.
limit, 40 ng/mL would then be the relative detection limit 6. Peng, S. X., Branch, T. M. and King, S. L. 2001. Fully automated 96-well liquid-liquid extraction for analysis ofbiological samples by liquid chromatography with tan- X. Quantitative Results dem mass spectrometry. Anal. Chem. 73: 708-714.
7. D'Arienzo, C., Wang, T. D. and Gale, P. J. 1998. High- After quantitation, sildenafil citrate contained in the five speed direct analysis of pharmaceuticals in plasma using sample solutions were at 39.2, 45.0, 43.3, 40.4 and 46.1 turbulent-flow chromatography and ESI-MS/ MS. p. 562.
mg/capsule, average at 42.8 mg/capsule, with relative devia- Proc. 46th ASMS conf. mass spectrmetry. N. J., U.S.A. tion at 6.8%. According to the instruction label(16), the sug- 8. Ackloo, S. Z., Smith, R. W., Marvin, C. H., Terlouw, gested dosage of VIAGRA as suggested by Pfizer Labs., is J.K.and McCarry, B. E. 1998. Electrospray(ES) liquid at one dose of 50 mg per day with maximum dose intake at chromatography/mass spectrometry(LC/MS) and 100 mg. Sildenafil citrate, as recorded in pharmacology, LC/MS/MS analysis of Ginseng saponins(Ginsenosides) could induce acute hypotension to cardiovascular patients extracted from Panax qinquefolium L. (American gin- taking nitrate compounds. A dosage of 3 or more of the tes- seng) root. p.569. Proc. 46th ASMS conf. mass spectrme- tion TCM capsules adulterated with sildenafil citrate could try. Allied Topics, Orlando. F. L., U.S.A. be dangerous.
9. Weigl, P., Liang, Z. and Bansal, S. 1998. LC/MS/MS method for simultaneous quantitation of protease inhibitors (saquinavir, indiavir, nelfinavir and ritonavir)in mouse plasma. p.1428. Proc. 46th ASMS conf. mass We thank Pupa Cicada for her translation.
spectrometry. Allied Topics, Orlando. F. L., U.S.A.
10. Li, L. Y., Vestal, C. H. and Kyranos, J. N. 1997.
Screening combinatiorial library compounds for a target receptor by automated high throughput mass spectrome- 1. Wen, K. C. and Tsai, M. J. 1995. Series of Testing try and tandem mass spectrometry. p.897. Proc. 45th Methods for Chinese Herbal Medicines (VII)-Analysis ASMS conf. mass spectrum. Allied Topics, Palm and Its Data, Spectra and Chromatograms of Synthetic Drugs as Adulterants in Chinese Medicinal Preparations.
11. Qualitative control of VIAGRA tablets in factory spec- National Laboratories of Foods and Drugs. Taipei, ifications, distributed by Pfizer Labs. NY., U.S.A.
Taiwan. (in Chinese) 12. USPXX II & NFX VII. 1990. p.315 & p.1518. The 2. Wen, K. C. and Tsai, M. J. 1995. Series of Testing United States Pharmacopeial Convention, Inc. meeting at Methods for Chinese Herbal Medicines (X)-Analysis and Washington, D. C., U.S.A.
Its Data, Spectra and Chromatograms of Synthetic Drugs 13. The Pharmacopoeia of R.O.C. 4th. 1995. p.194 & appen- as Adulterants in Chinese Medicinal Preparations.
dix p.20. Department of Health, Executive Yuan. R.O.C.
National Laboratories of Foods and Drugs. Taipei, Taiwan. (in Chinese) 14. The Pharmacopoeia of Japan. 1996. 13th ed. B302-B304.
3. Tseng, M. C., Tsai , M. J., Lin, J. H. and Wen, K. C. 2001.
Hirokawa publishing Co. Tokyo, Japan. (in Japanese) Determination of anorectics adulterated in traditional 15. Wang, P. Y., Tai, S. T., Huang, B. C., Liu, R. H. and Suen, Chinese medicines by GC/MS. J. Food Drug Anal. 8: E. T. T. 1996. GC/MS analysis of amphetamine in urine with amphetamine-d5 (side chain) as an internal stan- 4. Hoffmann, E. 1996. Tandem mass spectrometry: A dard. J. Food Drug Anal. 4: 123-130. (in Chinese) primer. J. Mass Spectrometry 31: 129-137.
16. Instruction pamphlet for VIAGRA (Sildenafil citrate) 5. Gardner, W. P., Shaffer, R. E., Girard, J. E. and Callahan, tablets. 2000. Distributed by Pfizer Labs. NY., U.S.A.
Journal of Food and Drug Analysis, Vol. 10, No. 2, 2002 sildenafil citrate sildenafil citrate Cosmosil 5C18-AR, 4.6 × 150 mm MRM, Multiple Reaction sildenafil citrate sildenafil citrate sildenafil citrate

Source: http://tera.chem.ut.ee/~jpenchuk/documents/kursused/Mass-spec/ViagraRsildenaficitrateGC_MS.pdf