Article pubs.acs.org/ac
LC-ESI-MS/MS Analysis of Testosterone at Sub-Picogram Levels Using a Novel Derivatization Reagent Michal Star-Weinstock,* Brian L. Williamson, Subhakar Dey, Sasi Pillai, and Subhasish Purkayastha AB SCIEX, Chemistry and Consumables R&D, 500 Old Connecticut Path, Framingham, Massachusetts 01701, United States S Supporting Information *
ABSTRACT: Testosterone analysis by LC-MS/MS is becoming the analytical method of choice over immunoassays due to its specificity and accuracy. However, neutral steroid hormones possess poor ionization efficiency in MS/MS, resulting in insufficient sensitivity for analyzing samples with trace concentrations of the hormones. The method presented here utilizes a derivatization step involving a novel, permanently charged, quaternary aminooxy (QAO) reagent or MS-tag that reacts to the ketone functionality of testosterone and significantly enhances its ESI-MS/MS sensitivity. This derivatization method enabled quantitation of total testosterone in human serum (200 μL) with a lower limit of quantitation (LLOQ) of 1 pg/mL (3.47 pmol/L), total testosterone in dried blood spots (8−10 μL) with a LLOQ of 40 pg/mL, and free testosterone in serum ultrafiltrate (400 μL) with a LLOQ of 0.5 pg/mL. The linearity of each of the high sensitivity applications was maintained over a broad dynamic range of 1−5000 pg/mL for the serum samples and 40−10 000 pg/mL for the dried blood spots (DBS) with R2 >0.998. The %CV at the LLOQ was 5 pg/mL with or without derivatization.12 In order to reliably measure Te in all female and pediatric samples at the desired level of 108 M−1 for oxime bonds (at room temperature).21 The higher Keq of oxime bond formation ensures better product stability when carbonyl compounds are derivatized with aminooxy reagents compared to hydrazine reagents. Many steroid hydrazones produce common fragment ions that are not viable for specific multiple reaction monitoring (MRM) analysis, resulting in high background noise.17
estosterone (Te) is the principle androgenic hormone in man and responsible for male secondary sex characteristics. Although Te does not directly influence sex-specific characteristics, it is thought to influence important functions such as pubertal development, bone density, muscle mass, erythropoiesis, energy, cognitive function, and mood.1 In women, Te levels are approximately 10-fold lower than levels in men. In infants and children, measurement of Te is important to identify cases of inborn errors of sex steroid metabolism and delayed or precocious puberty.2−4 The diagnosis of androgen deficiency in men and women, as well as the assessment of Te levels in infants and children, requires specific methods with high sensitivity and accuracy. The sensitivity limitations of Te assays are often a problem with serum/plasma samples from female or pediatric sources, low volume dried blood spots (DBS), or free testosterone (FTe) samples. FTe may be a more accurate assessment of androgenic status than total testosterone (TTe), especially in androgen-deficient elderly men5 and female hyperandrogenemia.4 FTe concentrations are only 0.5−3% of the TTe, with lower values found in female than in male serum. Methods with ultrahigh sensitivity that are able to reliably detect trace concentrations of FTe (200-fold higher concentration of androstenedione or DHEA than Te; thus, DHEA or androstenedione do not possess significant interference to the quantitation of Te by the above LC-MS/ MS method with QAO derivatization. Comparison with Other Derivatization Reagents. During the process of reagent design and method development, the sensitivity enhancement observed through QAO derivatization was compared to the commonly used Girard reagents, hydroxylamine, and several other derivatization reagents synthesized in-house. The QAO reagent was chosen over other reagents on the basis of overall sensitivity enhancement and selectivity in MS analysis. The use of permanently charged reagents such as Girard P (GP) and 2-hydrazyno-1-methylpyridine (HMP) to enhance the LC-MS/MS signal of oxosteroids has been previously reported.17,26 Enhancement factors using HMP reagent over nonderivatized Te have been reported to be on the order of 14fold,26 which is much less than the QAO reagent enhancement factor reported here. The GP derivative of Te was found to provide less sensitivity enhancement than the HMP derivative, and the chromatographic peak shape was showing persistent tailing.17 Unlike the product ions of QAO−Te, the fragment ions of the GP or HMP hydrazine reagents are nonspecific (derive only from the fragmentation products of the reagent itself), leading to high background noise in real samples and limiting the enhancement factor. A more commonly used reagent for the analysis of oxosteroids is hydroxyl amine (HA). Previous studies14,22,23 demonstrated sensitivity enhancement for Te with HA derivatization, sufficient to analyze serum from female and pediatric source. Similar to the QAO reagent, HA forms an oxime derivative with oxosteroids; however, the QAO reagent possesses a permanent positive charge that can potentially lead to better sensitivity enhancement in ESI-MS/ MS. The QAO reagent and HA were compared side by side using the same extracted sample with separate derivatization (HA derivatization followed the previously reported reaction conditions22) and analyzed by LC-MS/MS methods optimized for each individual derivative. We found that the QAO reagent provided approximately 10-fold improvement in the S/N ratio over HA for Te analysis (Supporting Information, Figure 5).
procedures for any matrix, including whole blood, for which the preparation of endogenous free matrix is time-consuming. In addition, the endogenous Te is present in the same concentration in all of the calibrators in the surrogate analyte approach, which can be used for determining reproducibility or as a second IS for method validation. A solid−liquid extraction (SLE) method was found to be the most appropriate for the desired volume of 200 μL of serum because it yielded high recovery and reproducibility for Te from the solid phase, was performed in a 96-well plate format, and is amenable to automation. The extraction and derivatization of 96 samples could be completed within 3 h, followed by a short LC-MS/MS run of 5−6 min. The recovery of Te was 100 ± 5%, it was determined by spiking serum with a known amount of d3Te before extraction and comparing it to the same amount spiked after extraction (data not shown). The LLOQ was 1 pg/ mL (3.47 pmol/L), and the dynamic range was 1−5000 pg/mL with R2 >0.99. The reproducibility at the LLOQ was 5 before and after derivatization (data not shown), but in the female samples, Te could not be detected without derivatization with QAO reagent (Figure 5). FTe Analysis. To determine the ability of the QAO derivatization method to accurately analyze FTe in serum samples, 400 μL of ultrafiltrate matrix was spiked with d3Te standards. The spiked ultrafiltrate was processed as described for FTe analysis. The LLOQ for this method was 0.5 pg/mL
good agreement as indicated by the Bland−Altman plot (% difference was within the 95% confidence interval) and a high degree of correlation (r = 0.992, Pearson). The measured and the known concentrations of the CDC samples are shown in the Supporting Information, Table 1. TTe Analysis in DBS. We investigated the feasibility of using the QAO derivatization method to analyze Te from DBS. Te extraction recovery was 79 ± 2%, and linearity was demonstrated (R2 >0.99) over the range of 50−10 000 pg/mL. The LLOQ was approximately 40 pg/mL (138.72 pmol/L), and the %CV at LLOQ was 11 (n = 23) with ±20% accuracy. F
dx.doi.org/10.1021/ac302036r | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
(1.73 pmol/L). The concentration curve was linear (R2 > 0.99) over the tested range of 1−500 pg/mL (physiologically relevant range). Te recovery using SLE 400 mg/well plates was similar to that obtained with the 96-well plates of 200 mg/well (100 ± 5%). At 1 pg/mL concentration, the %CV was 14.7 (n = 15), and the accuracy was 100 ± 15%. The within-day reproducibility was 6.5% CV (n = 6, data not shown). An example of FTe analysis in female serum is shown in Figure 6.
under ambient conditions. The reagent reacts fast with Te, and the derivatization reaction is complete within one hour at ambient temperature. Although the current study focuses on the analysis of Te, the QAO reagent is universal and can potentially form derivatives with any other compound that possesses ketone or aldehyde functionality. Due to the poor ionization efficiency of most ketosteroids in ESI-MS/MS, sensitivity enhancement by derivatization may be the only way to enable specific and accurate analysis. An isotopically enriched QAO reagent can be used to make an IS for any ketosteroid, particularly for those that are rare and for which an internal standard is not commercially available.
■
ASSOCIATED CONTENT
S Supporting Information *
Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Notes
The authors declare no competing financial interest. For research use only. Not for use in diagnostic procedures. The trademarks mentioned herein are the property of AB Sciex Pte. Ltd. or their respective owners. AB SCIEX is being used under license.
■
ACKNOWLEDGMENTS The authors thank Lisa Sapp, Hua-Fen Liu, and Keith Goodman for scientific discussion and providing us with the CDC HoSt project samples. We would also like to thank Jeanette Hill from Spot On Sciences for providing us HemaSpot fan-shaped dried blood spot samples.
Figure 6. Determination of percent free Te (%FTe) in a serum sample. For this sample, the TTe concentration was measured to be 20.3 pg/mL in 200 μL of serum. The FTe concentration was measured to be 0.6 pg/mL in 400 μL of ultrafiltrate (30 kDa MWCO membrane) from the same serum. A shallower LC gradient was used for this analysis (30−55% buffer B, 1−4 min); therefore, the retention time is slightly later.
■
REFERENCES
(1) Mazer, N. A. Int. J. Fertil. Women's Med. 2002, 47, 77−86. (2) Migeon, C. J.; Berkovitz, G. D.; Brown, T. R. Sexual differentiation and ambiguity. In The Diagnosis and Treatment of Endocrine Disorders in Childhood and Adolescence; Kappy, M. S., Blizzard, R. M., Migeon, C. J., Eds.; Charles C. Thomas: Springfield, IL, 1994; pp 573−715. (3) Kulin, H. E.; Finkelstein, J. W.; D’Arcangelo, M. R.; Susman, E. J.; Chinchilli, V.; Kunselman, S. J. J. Pediatr. Endocrinol. Metab. 1997, 10, 395−400. (4) Rosner, W.; Auchus, R. J.; Aziz, R.; Sluss, P. M.; Hershel, R. J. Clin. Endocrinol. Metab. 2010, 92, 405−413. (5) Yeap, B. B.; Almeida, O. P.; Hyde, Z.; Norman, P. E.; Chubb, S. A. P.; Jamrosik, K. Eur. J. Endocrinol. 2007, 156, 585−594. (6) Chen, Y.; Yazdanpanah, M.; Wang, Y. W.; Hoffman, B. R.; Diamandis, E. P.; Wong, P. Y. Clin. Biochem. 2012, 43, 490−496. (7) Rauh, M. Mol. Cell. Endocrinol. 2009, 301, 272−281. (8) Moal, V.; Mathieu, E.; Reynier, P.; Malthiéry, Y.; Gallois, Y. Clin. Chim. Acta 2007, 386, 12−19. (9) Cawood, M. L.; Field, H. P.; Ford, C. G.; Gillingwater, S.; Kickman, A.; Cowan, D.; Barth, J. H. Clin. Chem. 2005, 51, 1472− 1479. (10) Demers, M. L. Steroids 2008, 73, 1333−1338. (11) Shiraishi, S.; Lee, P. W. N; Leung, A.; Goh, V. H. H.; Swerdloff, R. S.; Wang, C. Clin. Chem. 2008, 54, 1855−1863. (12) Kushnir, M. M; Rockwood, A. L.; Roberts, W. L.; Yue, B.; Bergquist, J.; Meikle, A. W. Clin. Biochem. 2011, 44, 77−88.
In this sample, the TTe and FTe concentrations were measured as 20.3 and 0.6 pg/mL, respectively, and the %FTe was therefore calculated as 2.97, which is in accordance with the expected range of approximately 2−3%.27
■
CONCLUSIONS Derivatization with novel QAO reagent enabled ultrahigh sensitivity analysis of Te at
