Journal of Chromatography B, 969 (2014) 77–84

Contents lists available at ScienceDirect

Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb

Simultaneous measurements of cortisol and cortisone in urine and hair for the assessment of 11␤-hydroxysteroid dehydrogenase activity among methadone maintenance treatment patients with LC-ESI–MS/MS Zheng Chen a , Jifeng Li a , Guanyi Xu b , Jin Yang a , Jing Zhang a , Huihua Deng a,∗ a b

Research Center for Learning Science, Southeast University, Nanjing 210096, China Center of Methadone Maintenance Treatment, Baixia District Hospital, Nanjing 210004, China

a r t i c l e

i n f o

Article history: Received 17 December 2013 Accepted 3 August 2014 Available online 10 August 2014 Keywords: 11␤-Hydroxysteroid dehydrogenases Ratio of cortisol to cortisone Hair Urine Electrospray

a b s t r a c t The activity of 11␤-hydroxysteroid dehydrogenases (11␤-HSD) is traditionally assessed using the ratio of cortisol to cortisone in urine or saliva. However, these biomarkers only reflect the local activity of 11␤-HSD, and are easily affected by circadian variation of cortisol secretion. The shortcomings might be overcome by hair analysis. The present study aimed to develop an enhanced assay for simultaneous measurements of cortisol and cortisone in both hair and urine samples. The samples were collected from 29 patients under methadone maintenance treatment. The cortisol and cortisone were extracted either by solid phase extraction from a 20-mg milled hair sample after a 14-h incubation in 1 ml of methanol, or by twice liquid–liquid extraction from a 20-fold diluted urine sample. The analyses were performed using high performance liquid chromatography tandem mass spectrometry with electrospray ionization in negative mode. Limits of detection and quantification were 0.5 and 1.25 pg/mg for hair steroids and 0.2 and 0.5 ng/ml for urinary steroids, respectively. The recoveries were more than 97%. The intra- and inter-day coefficients of variation were less than 10%. The ratios of cortisol to cortisone in hair and urine were both less than one, but did not correlate with each other. A possible reason for the lack of correlation was that the ratios in hair and urine might mostly reflect the activity of 11␤-HSD type 2 in the eccrine sweat gland and in the kidney, respectively. Additionally, a significant correlation was observed between results obtained using external standard quantification and internal standard quantification. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Two isozymes of 11␤-hydroxysteroid dehydrogenases (11␤HSD) catalyze the inter-conversion between a physiologically active steroid hormone, cortisol and its inactive metabolite, cortisone. The 11␤-HSD type 1 reversibly catalyze the conversion of cortisone to cortisol, while the 11␤-HSD type 2 irreversibly converts cortisol to cortisone [1,2]. The insufficiency of 11␤-HSD activities results in hypertension, hypokalemia [2,3], apparent mineral corticoid excess syndrome [4] and other physiological or psychological diseases. The activities of 11␤-HSD in vivo can be assessed by the ratio of cortisol to cortisone (Rcc ) [5,6]. The sequential studies have demonstrated that Rcc is related to adrenal diseases

∗ Corresponding author. Tel.: +86 25 83795664; fax: +86 25 83793779. E-mail address: [email protected] (H. Deng). http://dx.doi.org/10.1016/j.jchromb.2014.08.001 1570-0232/© 2014 Elsevier B.V. All rights reserved.

[7], cardiovascular risk [8], chronic fatigue syndrome [9,10] and stress [11]. Most studies utilized Rcc from urine and saliva matrices to assess the activities of 11␤-HSD because urine and saliva collections are non-invasive. Unfortunately, these biomarkers suffer from two shortcomings. First, they show circadian variation [12] and can only reflect the levels in a short term from several minutes to several hours. This means that they could easily be influenced by accidental events and various environmental factors. Second, the 11␤-HSD type 2 is localized in kidney, colon, eccrine sweat gland, salivary gland and other tissues, while the 11␤-HSD type 1 is mostly distributed in liver, lung and adipose tissues [1,2]. The Rcc in urine and saliva may show the local activity of 11␤-HSD type 2 in kidney and salivary glands rather than the overall activity of 11␤-HSD isozymes. Notably, Rcc from hair matrix (Rhcc ) can potentially overcome these shortcomings. Endogenous cortisol and cortisone in hair matrix are thought to be mostly a result of active

78

Z. Chen et al. / J. Chromatogr. B 969 (2014) 77–84

or passive diffusions of blood-related species [13]. Moreover, hair matrix can record the average levels of blood-borne cortisol and cortisone over months [14]. Additionally, 11␤-HSD isozymes were not found in inner hair follicle cells despite the fact that 11␤-HSD type 1 is present in outer hair follicle root sheath cells [15,16] and 11␤-HSD type 2 in eccrine sweat gland in the skin [1,2]. Therefore Rhcc is thought to be more reliable in representing the integrated activity of 11␤-HSD isozymes in various peripheral organs and tissues. Recently, a few studies have investigated the physiological range of Rhcc [17–19]. These studies found that hair cortisone concentration was higher than that of hair cortisol, resulting in Rhcc being less than one on average. Vanaelst et al. reported a median Rhcc level of 0.80 with a range of 0.11–54.4 [18]. Similar results were reported in urine matrix [20], indicating Rhcc could reflect Rcc in urine matrix (Rucc ). As discussed above, Rhcc may represent the integrated activity of 11␤-HSD isozymes, while Rucc may represent the local activity of 11␤-HSD type 2 in the kidney. To date, whether biological mechanisms of Rhcc and Rucc are the same is still unclear. This is because the previous results on Rhcc and Rucc were obtained from different populations. Therefore it is essential to further elucidate the issue on Rhcc and Rucc from the same subjects to examine the association between Rhcc and Rucc . Measuring Rhcc , however, is not without challenges. Cortisol and cortisone contents in hair are very low [17]. Strong matrix effect due to other co-eluting components from hair matrix greatly suppresses the sensitivity and detection limit of assay methods simultaneously measuring hair cortisol and cortisone. For example, Raul et al. reported limits of detection (LOD) and quantitation (LOQ) at 1 and 5 pg/mg for their assay method based on liquid chromatography–tandem mass spectrometry (LC–MS/MS) with atmosphere pressure chemical ionization (APCI) source [17]. We also reported the results of 2 and 5 pg/mg using the method based on LC–MS/MS with electrospray ionization (ESI) source in negative mode [21]. In another study, an ultrahigh performance LC–MS/MS method (UPLC–MS/MS) Vanaelst et al. developed showed LOD and LOQ to be 2 and 5 pg/mg, respectively [18]. As a result, among samples taken from 223 elementary school girls, only 39 and 168 samples showed reliable hair cortisol and cortisone contents (higher than the LOQ), respectively [18]. Among the methods used in the above studies, LC–MS/MS with ESI in negative mode has a wider detection window, a lower background [22] and a higher sensitivity [23,24]. Our earlier study using LC-ESI–MS/MS method in negative mode showed that LOD and LOQ could reach 0.5 and 1.25 pg/mg for hair cortisol and cortisone, respectively. However, it is unknown yet whether the method can be suitable for urinary analysis. The present study aims to develop a simultaneous assay of cortisol and cortisone which is suitable for both hair and urine matrices and then to examine possible correlation between Rhcc and Rucc . The participants were heroin-dependent patients under methadone maintenance treatment (MMT). Earlier studies showed that MMT patients typically had high cortisol levels [25–27]. The higher cortisol levels might be attributed to chronic pain [25], depression [26,27] and sleep disorder [26], but also be affected by oral administration of methadone which could alter the activity of their hypothalamic–pituitary–adrenocortical axis (HPA axis) [28]. It is still unclear if the high cortisol level is attributed to treatment or psychological disorder or both. A study on the association of cortisol and cortisone would be helpful in understanding the issue from the perspective of cortisol metabolism. Additionally, some studies used external standard method (ESM) for quantitation of hair cortisol [29] while most others used internal standard method (ISM). The consistency between these two quantitation methods is also examined here.

2. Experimental 2.1. Chemicals HPLC grade acetone, hexane and methanol were purchased from Dikma, Lake Forest, CA. Formic acid (HCOOH) was from Tedia, Fairfield, OH. Ethyl acetate was from YuWang Chemicals Company, Shandong, China. Analytical grade standards of cortisol and cortisone were from National Institutes for Food and Drug Control, China. Deuterated cortisol (cortisol-9, 11, 12, 12-d4) was from Isotec, Sigma Aldrich, St. Louis, MO. These chemicals were used as received. Water used throughout the experiments was tripledistilled deionized water. Solid phase extraction (SPE) C18 column was purchased from Dikma, Lake Forest, CA. Stock solutions of cortisol and cortisone standards were prepared in methanol at a concentration of 100 ␮g/ml. Cortisol-d4 as internal standard (IS) was prepared in methanol at a concentration of 50 ␮g/ml. The binary mobile phase was 80:20 (V/V) methanol and deionized water containing 0.1% formic acid. Prior to use, the mobile phase was first filtered through a micro porous membrane (0.22 ␮m) and subsequently treated ultrasonically to remove bubbles. 2.2. Participants and sample collection Participants were recruited from a methadone maintenance treatment center in Nanjing, China. Candidates with body mass index (BMI) >28 or 0.05 for ISM). On the other hand, the above results indicated that two quantitation methods, ESM and ISM, could consistently elucidate the same inherent biological mechanism although ISM was a better quantitative method than ESM which was more heavily affected by matrix effect and environmental or operational variations. 4. Discussions This study found that Rucc was less than one (Table 3), which was consistent with previous studies where it was reported to be between 0.2 and 0.8 [9,20,32]. Notably, the cortisol–cortisone relationship in urine was the opposite of that in serum where Rcc was around 10 [33–35]. The significant differences between urine and serum demonstrated the enzyme activity of 11␤-HSD type 2 in kidney giving rise to the inactivation of cortisol to cortisone [1,2]. With regard to hair, the Rhcc was also less than one (Table 4). This result was consistent with that in the previous studies [17–19], but was contrary to that in serum [33–35]. Similarly, the result in hair seemed to be attributed to the 11␤-HSD type 2 compared with those in serum and urine. In fact, the enzyme is located in eccrine sweat glands [1,2], not in hair follicle nor in epidermis or the sebaceous glands. In other words, the enzyme separated from hair follicle could not directly convert cortisol to cortisone in the interior of hair shaft. Alternatively, Raul et al. proposed that conversion of cortisol to cortisone through 11␤-HSD type 2 happened before cortisol’s incorporation into hair [17]. They interpreted that the result might be due to active or passive diffusion of cortisol and cortisone from sweat during the formation of hair shaft. However, Raul et al.’s interpretation was in contrast to the main incorporation pathways into hair of cortisol and cortisone. As summarized in

83

a recent review [14], steroid hormones are possibly incorporated into hair shaft through the following five pathways. The first pathway is the diffusion of blood-borne lipophilic substances including cortisol and cortisone from blood capillaries into hair follicle cells where they are trapped and gradually deposited in the growing hair shaft [13]. The second pathway is the incorporation of cortisol and cortisone in the hair strands during the formation of hair shaft from the deep skin compartments with 11␤-HSD type 1. The third pathway is the incorporation of hormones which might come from the sebum and sweat glands [36]. The fourth pathway is the adhesion of the exogenous cortisol and cortisone from atmosphere after the complete formation of hair shaft [13]. The fifth pathway is the synthesis of cortisol and/or cortisone by hair follicle itself [37]. The first, second and fifth are possibly the main pathways among all the five. Hormones from these three pathways might be distributed inside hair shaft (e.g. cuticle, cortex and medulla) and those from the third and fourth pathways might be on the surface. The Rhcc of less than one might be mainly attributed to the catalysis of 11␤-HSD type 2 in eccrine sweat glands. Considering the above main incorporation pathways, 11␤-HSD type 2 in the eccrine sweat glands near the hair shaft could enter into the cuticle and even the interior of the hair shaft, thereby converting hair cortisol to cortisone. It could be incorporated into the hair shaft either during or after the formation of the hair shaft. For the grown hair shafts, 11␤-HSD type 2, together with sweat, could be dissolved in the sebum layer [1,2] and coated the outmost surface of the grown hair shaft. The enzyme could then enter into the cuticle of the swelled hair shaft by sweat. It emerged into the upper part of the hair root superficial to the duct of sebaceous glands. During the formation of hair shaft, the enzyme could also emerge in the growing hair and enter into the cuticle because of degradation of the inner root sheath of the hair shaft [13]. Sequentially, the enzymes in the cuticle might further drill into cortex and medulla of the hair shaft damaged by sweat and/or environmental factors (e.g. ultraviolet irradiation and frequent washing with shampoo and water [30]). Consequently, the enzymes in cuticle, cortex and medulla of the hair shaft convert cortisol to cortisone. At the same time, hair cortisol is likely dissolved out of a hair matrix frequently exposed to various external factors. If hair structure is intact, cortisol inside the hair is gradually released through a two-stage or multi-stage mechanism [38]. If hair structure is heavily damaged, cortisol in the interior of the hair shaft (e.g. cortex and medulla) is quickly dissolved out. The cortisol possibly reaches the surface of the hair shaft before it is converted into cortisone by 11␤-HSD type 2. In a short brief, 11␤-HSD type 2 in eccrine sweat glands is the possible cause for Rhcc being less than one. Another potential explanation for the Rhcc being less than one is that hair cortisol might be degraded. For example, previous studies reported that the loss of cortisol in saliva was 9.2% per month at room temperature [39] and that in urine about 6% per week at room temperature and a pH of 0.5–7 [40]. The reason for the rate of hair cortisol loss was not clear although hair matrix was considered to be stable [14]. Future studies are necessary to validate some of the mechanisms proposed here. Determining Rhcc in different layers of hair structure would be helpful in elucidating the proposed mechanisms. Additionally, 11␤-HSD type 1 is localized in the skin including dermal fibroblasts [15] and outer root sheath cells of hair follicles [16]. During the formation of hair shaft, cortisol and cortisone could be converted into each other in the outer root sheaths of hair follicles under catalysis of 11␤-HSD type 1 and diffused from the outer to the inner of the hair follicles. After the formation of hair shaft, cortisol and cortisone released from hair shaft and secreted possibly by sebum and sweat glands were converted into each other under catalysis of 11␤-HSD type 1 in the skin. The reversible interconversion between cortisol and cortisone in the skin might change the Rhcc due to the local 11␤-HSD type 1 activity.

84

Z. Chen et al. / J. Chromatogr. B 969 (2014) 77–84

Cortisone, as an inactive metabolite of cortisol, is negligible in adrenal production [1,2]. Circulating cortisone is mostly from the irreversible conversion of cortisol catalyzed by 11␤-HSD type 2 that is mainly localized in mineral corticoid target tissues, such as kidney, colon, eccrine sweat gland and salivary gland [1,2]. Therefore it is inferred that cortisone would be closely associated with cortisol although its content is also modulated by the reversible conversion of cortisone to cortisol under the catalysis of 11␤-HSD type 1. This hypothesis was confirmed by the present finding that urinary free cortisone was significantly and positively correlated with urinary free cortisol (Table 5). This finding indicated that urinary cortisone mostly results from the catalysis of 11␤-HSD type 2 in various peripheral organs and tissues, especially in the kidney. Thus the positive correlation of Rucc with urinary cortisol content (Table 5) demonstrated that Rucc depended strongly on the enzyme-catalyzed conversion of cortisol. Similarly, from the positive correlations of hair cortisone and Rhcc with hair cortisol (Table 5), it can be inferred that hair cortisone is associated with an irreversible conversion of cortisol catalyzed by 11␤-HSD type 2 in various peripheral organs and tissues, and thereby Rhcc might strongly depend on the irreversible conversion of cortisol. The explanation was in accordance with the above incorporation mechanisms whereby most of hair cortisol and cortisone are thought to result from the active and/or passive diffusion of unbound blood species into hair matrix [13]. Furthermore, it can be inferred that hair cortisone might be more dependent on the catalysis of 11␤HSD type 2 in eccrine sweat glands as discussed above. However, we found that hair cortisone showed high correlation with hair cortisol while urinary cortisone showed moderate correlation with urinary cortisol (Table 5). This might be because urinary free cortisol can be converted into other metabolites (e.g. 5␣- and 5␤-dehydrocortisol and 5␣- and 5␤-tetrahydrocortisol) apart from cortisone. Urinary free cortisone can also be converted to dehydrocortisone and then to tetrahydrocortisone [1]. As a result, the cortisone–cortisol correlation in hair was stronger than that in urine. Interestingly, there was no significant difference in the coefficient of Rcc –cortisol correlation between hair and urine in our study. This result implied that 11␤-HSD type 2 in eccrine sweat glands showed the same catalysis activity as that in the kidney which resulted in the no significant difference between Rhcc and Rucc .

5. Conclusions An assay for the simultaneous measurements of cortisol and cortisone in both hair and urine was developed. The assay was based on HPLC-ESI-MS/MS in negative mode. The method showed good performance with the limits of detection and quantification at 0.5 and 1.25 pg/mg for hair steroids, and 0.2 and 0.5 ng/ml for urinary ones. The recoveries were more than 97% and the intra- and interday coefficients of variation were less than 10% for both matrices. All the urine and hair samples from MMT population can be reliably quantified. This study also found that cortisone levels and Rcc were positively and significantly correlated with cortisol levels for both hair and urine samples. These results showed that cortisone and Rcc levels in urine and hair were associated with the activities of 11␤-HSD type 2 in the kidney and eccrine sweat glands, respectively. Furthermore, both Rhcc and Rucc were found to be less than one, but did not correlate with each other. A possible reason was that the hair cortisol might be irreversibly converted into cortisone by 11␤-HSD type 2 in the eccrine sweat gland through the infiltration of sweat, while urine cortisol by 11␤-HSD type 2 in the kidney. Additionally, ISM could effectively suppress the matrix effect compared to ESM. Nevertheless, concentrations of cortisol and cortisone quantified with ESM were well correlated with those with ISM. These conclusions were drawn on the MMT patients.

Further confirmation of the conclusions above should be pursued in other populations. Acknowledgements The statistical analysis, data interpretation, and the student scholarships were supported by Humanities and Social Science Foundation (11YJAZH019), Ministry of Education, China. The assay of hair cortisol and cortisone with LC–MS/MS and chemical regents was supported by Program for New Century Excellent Talents in University (NCET-08-0122), Ministry of Education, China. References [1] P. Ferrari, Biochim. Biophys. Acta (BBA) – Mol. Basis Dis. 1802 (2010) 1178–1187. [2] M. Quinkler, P.M. Stewart, J. Clin. Endocrinol. Metab. 88 (2003) 2384–2392. [3] N. Glorioso, F. Filigheddu, P.P. Parpaglia, A. Soro, C. Troffa, G. Argiolas, P. Mulatero, Eur. Heart J. 26 (2005) 498–504. [4] M. Palermo, M. Quinkler, P.M. Stewart, Arq. Bras. Endocrinol. Metab. 48 (2004) 687–696. [5] A. Yokokawa, T. Takasaka, H. Shibasaki, Y. Kasuya, S. Kawashima, A. Yamada, T. Furuta, Steroids 77 (12) (2012) 1291–1297. [6] R. Best, B.R. Walker, Clin. Endocrinol. (Oxf.) 47 (1997) 231–236. [7] S. Nomura, M. Fujitaka, K. Jinno, N. Sakura, K. Ueda, Clin. Chim. Acta 256 (1996) 1–11. [8] J.L. Anderson, J.F. Carlquist, W.L. Roberts, B.D. Horne, H.T. May, E.L. Schwarz, M. Pasquali, R. Nielson, M.M. Kushnir, A.L. Rockwood, Am. Heart J. 153 (2007) 67–73. [9] W.K. Jerjes, T.J. Peters, N.F. Taylor, P.J. Wood, S. Wessely, A.J. Cleare, J. Psychosom. Res. 60 (2006) 145–153. [10] W. Jerjes, A. Cleare, S. Wessely, P. Wood, N. Taylor, J. Affect. Disord. 87 (2005) 299–304. [11] D. Atlaoui, M. Duclos, C. Gouarne, L. Lacoste, F. Barale, J.C. Chatard, Med. Sci. Sports Exerc. 36 (2004) 218–224. [12] I. Perogamvros, L.J. Owen, J. Newell-Price, D.W. Ray, P.J. Trainer, B.G. Keevil, J. Chromatogr. B 877 (2009) 3771–3775. [13] F. Pragst, M.A. Balikova, Clin. Chim. Acta 370 (2006) 17–49. [14] E. Russell, G. Koren, M. Rieder, S. Van Uum, Psychoneuroendocrinology 37 (2012) 589–601. [15] A. Tiganescu, E.A. Walker, R.S. Hardy, A.E. Mayes, P.M. Stewart, J. Invest. Dermatol. 131 (2010) 30–36. [16] M. Terao, H. Murota, A. Kimura, A. Kato, A. Ishikawa, K. Igawa, E. Miyoshi, I. Katayama, PLoS ONE 6 (2011) e25039. [17] J.S. Raul, V. Cirimele, B. Ludes, P. Kintz, Clin. Biochem. 37 (2004) 1105–1111. [18] B. Vanaelst, N. Rivet, I. Huybrechts, B. Ludes, S. De Henauw, J.S. Raul, Anal. Methods 5 (2013) 2074–2082. [19] J. Zhang, J. Li, Y. Xu, J. Yang, Z. Chen, H. Deng, Clin. Chim. Acta 426 (2013) 25–32. ˛ ˛ [20] A. Plenis, L. Konieczna, I. Oledzka, P. Kowalski, T. Baczek, Mol. BioSyst. 7 (2011) 1487–1500. [21] Z. Chen, J. Li, J. Zhang, X. Xing, W. Gao, Z. Lu, H. Deng, J. Chromatogr. B 929 (2013) 187–194. [22] T. Henriksen, R.K. Juhler, B. Svensmark, N.B. Cech, J. Am. Soc. Mass Spectrom. 16 (2005) 446–455. [23] P. Labadie, E.M. Hill, J. Chromatogr. A 1141 (2007) 174–181. [24] L. Salste, P. Leskinen, M. Virta, L. Kronberg, Sci. Total Environ. 378 (2007) 343–351. [25] R.N. Jamison, J. Kauffman, N.P. Katz, J. Pain Symptom Manag. 19 (2000) 53–62. [26] D. Pud, C. Zlotnick, E. Lawental, Addict. Behav. 37 (2012) 1205–1210. [27] E. Peles, S. Schreiber, Y. Naumovsky, M. Adelson, J. Affect. Disord. 99 (2007) 213–220. [28] J.H. Schluger, L. Borg, A. Ho, M.J. Kreek, Neuropsychopharmacology 24 (2001) 568–575. [29] Q. Xie, W. Gao, J. Li, T. Qiao, J. Jin, H. Deng, Z. Lu, Clin. Chem. Lab. Med. 49 (2011) 2013–2019. [30] J. Li, Q. Xie, W. Gao, Y. Xu, S. Wang, H. Deng, Z. Lu, Clin. Chim. Acta: Int. J. Clin. Chem. 413 (2012) 434–440. [31] B. Sauvé, G. Koren, G. Walsh, S. Tokmakejian, S.H.M. Van Uum, Clin. Invest. Med. 30 (2007) E183–E191. [32] R. Gatti, E. Cappellin, B. Zecchin, G. Antonelli, P. Spinella, F. Mantero, E.F. De Palo, J. Chromatogr. B 824 (2005) 51–56. [33] G. Morineau, A. Boudi, A. Barka, M. Gourmelen, F. Degeilh, N. Hardy, A. AlHalnak, H. Soliman, J.P. Gosling, R. Julien, Clin. Chem. 43 (1997) 1397–1407. [34] M. Vogeser, J. Briegel, R. Zachoval, Clin. Biochem. 35 (2002) 539–543. [35] M.M. Kushnir, R. Neilson, W.L. Roberts, A.L. Rockwood, Clin. Biochem. 37 (2004) 357–362. [36] J.S. Meyer, M.A. Novak, Endocrinology 153 (2012) 4120–4127. [37] N. Ito, T. Ito, A. Kromminga, A. Bettermann, M. Takigawa, F. Kees, R.H. Straub, R. Paus, FASEB J. 19 (2005) 1332–1334. [38] X. Xing, Z. Chen, J. Li, J. Zhang, H. Deng, Z. Lu, Clin. Chim. Acta 421 (2013) 62–72. [39] A. Garde, Å. Hansen, Scand. J. Clin. Lab. Invest. 65 (2005) 433–436. [40] K. Miki, A. Sudo, Clin. Chem. 44 (1998) 1759–1762.