Analysis of Estrogens and Androgens in Postmenopausal Serum and Plasma by Liquid Chromatography-Mass Spectrometry

Abstract

Liquid chromatography-selected reaction monitoring/mass spectrometry-based methodology has evolved to the point where accurate analyses of trace levels of estrogens and androgens in postmenopausal serum and plasma can be accomplished with high precision and accuracy. A suite of derivatization procedures has been developed, which together with modern mass spectrometry instrumentation provide investigators with robust and sensitive methodology. Preionized derivatives are proving to be useful as they are not subject to suppression of the electrospray signal. Postmenopausal women with elevated plasma or serum estrogens are thought to be at increased risk for breast and endometrial cancer. Therefore, significant advances in risk assessment should be possible now that reliable methodology is available. It is also possible to conduct analyses of multiple estrogens in plasma or serum. Laboratories that are currently employing liquid chromatography/mass spectrometry methodology can now readily implement this strategy. This will help conserve important plasma and serum samples available in Biobanks, as it will be possible to conduct high sensitivity analyses using low initial sample volumes. Reported levels of both conjugated and non-conjugated estrogen metabolites are close to the limits of sensitivity of many assays to date, urging caution in the interpretation of these low values. The analysis of serum androgen precursors in postmenopausal women has not been conducted routinely in the past using liquid chromatography/mass spectrometry methodology. Integration of serum androgen levels into the panel of metabolites analyzed could provide additional information for assessing cancer risk and should be included in the future.

1. Introduction

There is a compelling need for reliable methodology capable of quantifying estrogens in the serum of postmenopausal women because increased levels appear to be associated with increased breast cancer risk [1,2]. Estrogen carcinogenesis arises through a dual mechanism in which estradiol can act either as a hormone to stimulate aberrant cell proliferation or as the precursor to the formation of genotoxic catechol metabolites [3]. Estrogen levels in the breast tissues of postmenopausal women are dependent upon the availability of circulating C-19 androgen precursors, which are converted to estrogens in the tissue (Figure 1). Estrogens can then be released into the circulation, providing biomarkers of tissue estrogen biosynthesis if it is assumed that the circulating levels are reflective of tissue concentrations. This assumption has been questioned because tissue levels of estrogens are significantly higher than the corresponding circulating levels and breast tissue-specific metabolism is known to occur. A pharmacokinetic model has been proposed in which there is rapid equilibrium between tissue and plasma estrogens that may might explain this conundrum [4].

Figure 1.

The formation of estrogens in the tissue postmenopausal women from circulating C-19 androgens and sulfate precursors.

The analysis of circulating androgens concentrations can provide insight into availability of relevant androgen precursors, such as androstenedione and testosterone, which can be taken up into tissue (Figure 1). In postmenopausal women, such an analysis could provide useful additional biomarkers of breast cancer risk. Circulating sulfate conjugates have the potential to provide a source of estrogens in breast tissue through the action of sulfatases, which would release the corresponding non-conjugated steroids [5]. This is particularly relevant to circulating estrone-3-sulfate (a precursor to estrone) and dehydroepiandrosterone (DHEA) sulfate, a precursor to DHEA, which is a substrate for 3β-hydroxysteroid dehydrogenase (HSD)-mediated conversion to androstenedione. The androstenedione can in turn be converted to estrone by aromatase (Figure 1). However, there is little evidence that the conversion of circulating sulfate conjugates to tissue androgens and estrogens actually takes place [4]. Furthermore, the polar nature of the sulfate conjugates suggests that they are not good substrates for passive diffusion from the plasma into breast tissue. However, the ability of multiple drug transporter (MRP)-1 (ABCC1) to transport estrone-3-sulfate [6] and MRP-1 and MRP-4 (ABCC4) to transport DHEA sulfate [7] does provide an alternative mechanism for the conjugated steroids to be taken up by breast tissue. Therefore, the analysis of circulating estrone-3-sulfate and DHEA sulfate in postmenopausal women could also be informative.

Aromatase inhibitors have significantly improved the recurrence-free and overall survival rates in breast cancer patients [8]. Unfortunately, only incremental progress has been made over the last decade in preventing breast cancer among postmenopausal women. There is a compelling need to improve this situation in view of the aging world population and the role of aging as an important determinant of breast cancer risk [9,10]. It is clear that implementation of breast cancer prevention programs will require selection of women with high breast cancer risk in order to maximize the benefit/risk ratio [11,12]. It is anticipated that significant advances in risk assessment will be possible if reliable methodology is available to quantify estrogens and androgens in the plasma or serum of postmenopausal women [9]. These measurements can be coupled with other risk factors such as mammographic density [13], bone density [14], body mass index (BMI) [15], and single-nucleotide polymorphisms associated with breast cancer [16] to provide an improved model of breast cancer risk [11]. The present review will focus on the analysis of non-conjugated and conjugated estrogens and androgens using highly specific and sensitive stable isotope dilution liquid chromatography/mass spectrometry methodology that can be used to assess breast cancer risk.

2. Non-conjugated estrogens

Non-conjugated estradiol and its downstream non-conjugated metabolites are present in plasma and serum in the free form (not bound to steroid binding proteins) in postmenopausal women in the fg/mL range, which puts them below the limit of quantitation (LOQ) of routine assays [17,18]. Therefore, estrogens are quantified as a mixture of non-conjugated free and non-conjugated protein-bound forms. Typical serum concentrations of only 2.7-15.9 pg/mL for estradiol and 11.8-37.4 pg/mL for estrone in postmenopausal women [19-26] (Table 1) are still very challenging for most LC-MS-based procedures. Concentrations of non-conjugated free forms are determined by analyzing the amount of plasma steroid binding protein [27] and subtracting the amount of each individual non-conjugated estrogen calculated to non-covalently bind to this protein [28,29]. Clearly, estrogen assays with high sensitivity, specificity, and reproducibility are required in order for meaningful data to be obtained for postmenopausal women [30]. There are three major bioanalytical methods currently in use: radioimmunoassay coupled with chromatography [31], gas chromatography-selected reaction monitoring/mass spectrometry (GC-SRM/MS) [32], and stable isotope dilution liquid chromatography (LC)-SRM/MS [33]. There is increasing reliance on the use of LC-SRM/MS-based methodology because of the relative simplicity of the triple quadrupole instruments that are employed and the potential for future increased specificity by coupling LC with high-resolution ion-trap-based instruments [34,35].

Table 1. Concentrations of non-conjugated estrogens in postmenopausal serum determined by LC-SRM/MS.

Reference	Xu 2007 [19]	Caron 2009 [20]	Gao 2010 [21]	Walsh 2011 [22]	Rangiah 2011 [23]	Rothman 2011 [24]	Fuhrman 2012 [25]	Falk 2013 [26]	Overall
Concentration	Mean pg/mL	Mean pg/mL	Mean pg/mL	Median pg/mL	Mean pg/mL	Mean pg/mL	Median pg/mL	Median pg/mL	Mean pg/mL
E2	15.0	6.6	15.9	3.4	x	3.1	4.2	2.7	7.3
E1	32.7	32.3	x	18.4	12.3	37.4	14.6	11.8	22.8
16α-OH-E2	7.9	x	x	x	x	x	13.5	7.7	9.7
16α-OH-E1	BLQ	x	x	x	BLQ	x	11.2	10.7	10.9
2-OH-E2	BLQ	x	x	x	x	x	BLQ	BLQ	BLQ
2-OH-E1	BLQ	x	x	x	x	x	BLQ	BLQ	BLQ
4-OH-E2	x	x	x	x	x	x	x	x	x
4-OH-E1	BLQ	x	x	x	x	x	BLQ	BLQ	BLQ
2-MeO-E2	BLQ	x	x	x	x	x	1.4	0.7	1.0
2-MeO-E1	BLQ	x	x	x	2.7	x	2.3	4.6	3.2
4-MeO-E2	BLQ	x	x	x	x	x	BLQ	BLQ	BLQ
4-MeO-E1	BLQ	x	x	x	BLQ		BLQ	BLQ	BLQ

x = not analyzed

Abbreviations: BLQ = below lower limit of quantitation; E2 = estradiol; E1 = estrone; OH = hydroxy; MeO = methoxy.

For reliable measurements of multiple non-conjugated estrogens in plasma or serum, it is necessary to employ stable isotope internal standards, which have identical physical properties to the endogenous metabolites, but differ only in mass. Losses that occur during the extraction and chromatographic analysis are then compensated for because the ratio of each endogenous analyte to its internal standard remains the same. Stable isotope analogs also act as carriers to prevent selective losses of trace analytes through binding to active surfaces during extraction and analysis [36]. Until recently, this ideal condition was not possible for estradiol and its metabolites because only deuterated analogs were available for use as internal standards. Deuterated internal standards are not ideal as they can separate from the corresponding endogenous analyte, with a potential for differential suppression of their ESI signals and inaccurate quantification. The availability of many [¹³C₆]-estrogen analogs from Cambridge Isotope Laboratories (Andover, MA) provides internal standards that do not suffer from this potential problem.

Unfortunately, endogenous non-conjugated estrogens in postmenopausal serum or plasma cannot be quantified using conventional electrospray ionization (ESI) or atmospheric pressure chemical ionization (APCI) methodology. Therefore, it is necessary enhance the ionization characteristics of estrogens by first converting them to suitable derivatives. Three approaches to enhancing the sensitivity of LC-ESI/MS-based estrogen analysis through derivatization have been reported. The first approach, which we employed originally involves the preparation of an electron capturing pentafluorobenzyl (PFB) derivative of the estrogen coupled with the use of electron capture atmospheric pressure chemical ionization (ECAPC)/MS [37]. The Higashi group has also explored the utility of ECAPC/MS for estrogen analysis by using different electron capturing derivatives [38]. We showed that it was possible to quantify estrogens in the low pg/mL range in plasma using LC-ECAPCI/MS [39]. The second approach, which has much wider utility, involves the use of derivatives that enhance the ESI signal, a strategy that greatly improves sensitivity during LC-ESI/MS analysis. This approach is exemplified by studies from the Singh [40], Tai [41], Ziegler [42], and Kushnir [33] groups who used the dansyl (D) derivative to improve sensitivity of detection of non-conjugated estrogens from human biofluid samples (Figure 2). Alternative derivatives that have been employed include picolinoyl (P) by the Yamashita group [43] and pyridyl-3-sulfonyl (PS) by Spink group [44] group (Figure 2). The third approach involves the preparation of pre-ionized (quaternized) derivatives, so that ionization is not required in the ESI source of the mass spectrometer. This approach was reported in studies by the Chen [45], Adamec [46] and Higashi [47] groups in which N-methyl-2-pyridyl (NMP), N-methyl-nicotinyl (NMN) or 1-(2,4-dinitro-5-fluorphenyl-4,4,-dimethylpiperaziny (MPPZ) derivatives, respectively were attached to the 3-hydroxy phenolic moiety of the estrogen (Figure 2). Our group has also used pre-ionized derivatives to improve sensitivity by adding a Girard P (GP) derivative to the 17-oxo moiety of estrone and its metabolites [23] or by adding a Girard T (GT) derivative to the 17-oxo-moiety of androgens [48] (Figure 2). We have also recently developed the pre-ionized N-methyl-3-sulfonyl-pyridinium (NMPS) derivative that can be used for both estradiol and estrone metabolites (Figure 2).

Figure 2. Derivatives used to enhance the ionization efficiency of estrogens in order to improve sensitivity for LC-MS/MS analysis.

The concentrations of non-conjugated estradiol in the serum of postmenopausal women determined by LC-SRM/MS were reported to be the range of 2.7 to 15.9 pg/mL with a mean value of 7.3 pg/mL (Table 1). Concentrations of serum estrone that were also determined by LC-SRM/MS were reported to be significantly higher – in the range 11.8 to 37.4 pg/mL with a mean value of 22.8 pg/mL (Table 1). These values are in reasonable agreement with those obtained for estradiol in the serum of postmenopausal women (mean 5.1 pg/mL, range 2.9 to 7.3 pg/mL) [32,49-51] and serum estrone (mean 15.2 pg/mL, range 12.7 to 17.6 pg/mL) [50,51] using high sensitivity GC-SRM/MS. This suggests that values in excess of 15 pg/mL for estradiol and 30 pg/mL for estrone should treated with extreme caution. LC-SRM/MS studies that analyzed non-conjugated serum 16α-hydroxy-estradiol reported levels that were quite consistent with a mean value of 9.7 pg/mL and a range of 7.7 to 13.5 pg/mL (Table 1). In contrast, two studies have reported 16α-hydroxy-estrone concentrations to be in a similar range (10.7 to 11.2 pg/mL) whereas two additional studies reported 16α-hydroxy-estrone to be below the limit of quantification (Table 1). It is noteworthy that the highly sensitive method based on a GP derivative, which has a limit of quantification of 0.15 pg/mL for 16α-hydroxy-estrone was unable to detect any of this analyte [23]. This suggests that trace amounts of interfering substances could be responsible for the very low concentrations of the non-conjugated hydroxylated estrogens that have been reported and that the results of these studies should be re-evaluated with assays capable of more sensitive detection. The non-conjugated catechol estrogens (2- and 4-hydroxy-estradiol, 2- and 4-hydroxy-estrone) represent very challenging analytical targets because of their inherent instability. Most of the LC-SRM/MS studies have reported very low or undetectable amounts of these unstable analytes in serum (Table 1). Similarly, the non-conjugated methoxy-estrogens all appear to be present at levels that are below the limit of quantification of most LC-SRM/MS assays (0.7 to 4.6 pg/mL; Table 1).

3. Conjugated estrogens

Two approaches have been employed for the analysis of conjugated estrogens. The first approach involves hydrolysis of the β-glucuronide and sulfate conjugates with β-glucuronidase/arylsulfatase (G/S) such as that purified from helix pomatia, followed by derivatization using one of the derivatives in shown Figure 2 and LC-SRM/MS analysis. Analytical data are often reported as total (T) values, which is the sum of the non-conjugated and conjugated estrogens (Table 2). The second approach involves analysis of the intact conjugate without G/S hydrolysis or derivatization using negative ESI-based LC-MS methodology. Each method has drawbacks and for many of the conjugated estrogens rigorous stable isotope dilution methodology cannot be employed because appropriate standards are not available. The use of G/S hydrolysis for all of the conjugates (except estradiol and estrone where appropriate standards are available) is dependent upon the assumption that quantitative conversion of the conjugates occurs. Non-conjugated heavy isotope standards can compensate for any decomposition of the endogenous estrogens that occurs during the hydrolysis procedure. It is more problematic that no authentic standards are available for many of the potential β-glucuronide and sulfate estrogen conjugates. Furthermore, no systematic studies have been conducted to evaluate which conjugates are present in postmenopausal serum and whether they are completely hydrolyzed by the typical β-glucuronidase/sulfatases that are employed. One way to provide assurance that hydrolysis is complete is to conduct a separate methanolysis of the sample with anhydrous hydrogen chloride in methanol [52]. The concentrations of non-conjugated estrogens that are determined can then be compared with those obtained from G/S hydrolysis [53,54].

Table 2. Concentrations of conjugated and total estrogens in postmenopausal serum determined by LC-SRM/MS.

Reference	Xu 2007 [19]	Labrie 2007 [50]	Labrie 2008 [51]	Masi 2009 [57]	Caron 2009 [20]	Gao 2010 [21]	Patel 2011 [58]	Walsh 2011 [22]	Fuhrman 2012 [25]	Falk 2013 [26]	Dallal 2014 [59]	Overall
Technique	G/S	Intact	Intact	G/S	Intact	G/S	G/S	Intact	G/S	G/S	G/S
Concentration	Mean pg/mL	Mean pg/mL	Mean pg/mL	Mean pg/mL	Mean pg/mL	Mean pg/mL	Mean pg/mL	Median pg/mL	Median pg/mL	Median pg/mL	Median pg/mL	Mean
E2-T	51.5	x	x	31.6	x	48.3	x	x	9.8	6.2	10.5	26.3
E2-3G	x	x	x	x	8.4	x	x	2.5	x	x	x	5.5
E1-T	442.1	x	x	196.7	x	x	x	x	105.5	61.3	77.3	176.6
E1-3G	x	x	x	x	30.9	x	x	22.6	x	x	x	26.8
E1-S	x	232.0	137.0	x	440.0	x	x	170.0	x	x	x	244.8
16α-OH-E2-T	27.9	x	x	741.6	x	x	x	x	126.0	78.4	70.5	208.9
16α-OH-E1-T	8.8	x	x	26.4	x	11.5	19.4	x	11.2	10.7	8.1	13.7
2-OH-E2-T	11.1	x	x	15.9	x	12.7	x	x	10.4	6.6	3.5	10.0
2-OH-E1-T	72.5	x	x	136.4	x	69.4	x	x	19.7	21.2	17.9	56.2
4-OH-E2-T	x	x	x	x	x	x	x	x	x	x	x	x
4-OH-E1-T	11.5	x	x	BLQ	x	x	x	x	6.3	5.0	3.0	6.5
2-MeO-E2-T	BLQ	x	x	19.0	x	x	x	x	2.3	4.7	4.0	7.5
2-MeO-E2-3G	x	x	x	x	6.9	x	x	2.5	x	x	x	4.7
2-MeO-E1-T	BLQ	x	x	32.8	x	x	x	x	3.7	12.4	8.4	14.4
2-MeO-E1-3G	x	x	x	x	6.2	x	x	2.5	x	x	x	4.4
4-MeO-E2-T	BLQ	x	x	9.2	x	x	x	x	0.8	0.6	0.8	2.9
4-MeO-E1-T	BLQ	x	x	16.6	x	x	x	x	1.1	1.1	1.1	5.0

x = not analyzed

Abbreviations: BLQ = below lower limit of quantitation; E2 = estradiol; E1 = estrone; OH = hydroxy; MeO = methoxy; T= total (non-conjugated + conjugated); G/S= β-glucuronidase/arylsulfatase; -S = sulfate; -G = glucuronide

In general, very few heavy isotope internal standards are available for rigorous quantification of estrogen conjugates. A heavy isotope internal standard for estrone-3-sulfate (2,4,16,16-[²H₄]-estrone-3-sulfate) is commercially available from C/D/N isotopes (Pointe-Claire, CA) so that reliable quantitative assays can be conducted for this important analyte [22,55]. It is surprising that this standard has not generally been used for quantitative determinations of estrone-3-sulfate in the serum of postmenopausal women. When conducting quantitative determinations, care has to be taken with calibration standards because both unlabeled and labeled estrone-3-sulfate contain Tris as a stabilizer as well as significant amounts of water [55]. Quantifying the actual amount of estrone-3-sulfate in solutions requires UV spectrophotometry, in which the λ_max of estrone at 270 nM (ε 2000) is measured [56]. Three regioisomeric heavy isotope labeled estradiol β-glucuronides ([²H₄]-estradiol-3β-glucuronide, [²H₄]-estradiol-17-β-glucuronide, [²H₄]-estradiol-3,17-bis-β-glucuronide) have been synthesized using rat liver microsomes and used as internal standards in LC-SRM/MS assays for the corresponding endogenous β-glucuronides [20,22]. No heavy isotope internal standards are available for any of the other estrogen conjugates that have been analyzed (Table 2).

The very low concentrations of non-conjugated estradiol in the serum of postmenopausal women are reflected in the low level of total (non-conjugated + conjugated) serum estradiol with a mean value of 26.3 pg/mL and a range of 6.2 to 51.5 pg/mL [19-22,25,26,50,51,57-59] (Table 2). The reported mean levels of estradiol-3β-glucuronide of 5.5 pg/mL suggest that the sulfate conjugate is probably also present in the serum. Recent studies have found total mean estrone concentrations to be 176.6 pg/mL with a range of 61.3 to 442.1 pg/mL (Table 2). However, specific analyses of estrone-sulfate (137 to 440 pg/mL) and estrone-3β-glucuronide (22.6 to 33.9 pg/mL) suggest that some of the total estrone conjugate values obtained after G/S hydrolysis might be an underestimate of the true values (Table 2). Therefore, future studies should focus on the analysis of intact estrone-sulfate using stable isotope LC-SRM/MS methodology [22,55]. This approach would be particularly useful for monitoring the effect of aromatase inhibitors [60].

Mean levels of total 16α-hydroxy-estradiol have been reported to be 208.9 pg/mL with a wide range of 27.9 to 741.6 pg/mL. However, more recent reports suggest that the actual range may be closer to 70.5 to 126.0 pg/mL (Table 2). Levels of total 16α-hydroxy-estrone were reported as being much lower (mean 13.7 pg/mL, range 8.1 to 26.4 pg/mL) (Table 2). These values are similar to the non-conjugated levels (Table 1) adding further concern that these low levels values might simply arise from quantification of trace amounts of interfering substances. Levels of the other total estrogen conjugates except for 2-hydroxy-estrone were all very close to the LOQs (8 pg/mL) that have been reported for most of the assays (Table 2). Therefore, care should be exercised in interpreting these values.

4. Non-conjugated and conjugated androgens

The importance of analyzing androgens stems from their potential conversion to estrogens in breast and endometrial tissue (Figure 1). Consequently, a number of studies have reported the analysis of non-conjugated testosterone, DHEA, and androstenedione in serum samples from postmenopausal women (Table 3) [50,51,61]. Levels of non-conjugated testosterone determined by LC-SRM/MS have been reported as a mean level of 173 pg/mL with a range of 109 to 248 pg/mL [22,24,50,51,61-65] (Table 3). This is slightly higher than the mean value of 107 pg/mL (range 90-130 pg/mL) that was reported using GC-SRM/MS [22,24,50,51,62,63,65]. It is noteworthy that when modern very high sensitivity triple quadrupole instrumentation was employed, the concentration of serum testosterone of 187 pg/mL was still higher than the values reported by GC-SRM/MS [64]. The higher values for serum non-conjugated testosterone reported by the LC-SRM/MS methods could be overestimating the actual serum concentrations, so these values should be re-evaluated. Doing so could involve use of derivatization procedures to improve sensitivity and specificity [48] and/or use of highresolution MS [34,35]. The mean level of serum androstenedione was reported to be 380 pg/mL with a range of 354 to 440 pg/mL (Table 3). These consistent values suggest that serum androstenedione concentrations in postmenopausal women are truly in the range of 380 pg/mL. The mean level of non-conjugated DHEA was reported to be 1790 pg/mL with a range 1670 pg/mL to 1910 pg/mL (Table 3). In contrast to non-conjugated serum testosterone, this value agrees well with the mean level of 1203 pg/mL obtained by GC-SRM/MS (range 720 to 1800 pg/mL) [50,51,65]. The mean level of DHEA sulfate was reported to be 499 ng/mL (range from 355 to 600 pg/mL), which is approximately five orders of magnitude higher than the mean value of 1790 pg/mL reported for the non-conjugated form of DHEA (Table 3). If the sulfate conjugate can serve as a precursor to the formation of estrogens (Figure 1) this represents an enormous pool that could potentially be eliminated by the use of sulfatase inhibitors [5]. Surprisingly, no studies have been reported on the use of LC-SRM/MS for the analysis of conjugated testosterone in the serum of postmenopausal women. This would be a worthwhile endeavor for the future if indeed the sulfate conjugate can be transported into tissues and undergo hydrolysis to provide an additional source of non-conjugated testosterone for conversion to estrone (Figure 1).

Table 3. Concentrations of non-conjugated and conjugated androgens in postmenopausal serum determined by LC-SRM/MS.

Reference	Labrie 2007 [50]	Labrie 2008 [51]	Kushnir 2010 [61]	Bui 2010 [62]	Walsh 2011 [22]	Rothman 2011 [24]	Haring 2012 [63]	Methlie 2013 [64]	Labrie 2013 [65]	Overall
Concentration	Mean	Mean	Median	Mean	Mean	Mean	Mean	Median	Mean	Mean
T (pg/mL)	x	x	180	248	140	109	x	187	x	173
AD (pg/mL)	x	x	360	x	440	x	354	367	x	380
DHEA (pg/mL)	x	x	1670	x	1910	x	x	x	x	1790
DHEA-S (ng/mL)	502	355	x	x	600	x	x	x	540	499

x = not analyzed

Abbreviations: AD, androstenedione; T, non-conjugated testosterone, DHEAS, DHEA-sulfate.

Summary and Future Directions

The availability of a suite of derivatization procedures makes it possible to quantify nonconjugated estrogens by LC-SRM/MS (Table 1) with sensitivity comparable to that which can be obtained by GC-SRM/MS [32]. Pre-ionized derivatives are also proving to be useful for the quantification of androgens [48], although this methodology has not yet been applied to postmenopausal serum samples. This suggests that in the future it will be possible to conduct LC-MS/MS assays on multiple estrogen and androgen metabolites in serum and plasma at an order of magnitude lower than current methodology (Table 1). The availability of high sensitivity high-resolution ion trap instrumentation such as the Thermo Q-Exactive (San Jose, CA) should make it possible to conduct analyses with further increases in sensitivity and specificity. Preliminary results are very encouraging with high-resolution instruments [34,35]. Improved specificity could also arise from the use of improved chromatographic separations such as that which can be obtained with supercritical fluid chromatography (SFC) [66]. The availability of modern triple quadrupole mass spectrometers such as the Waters Xevo (Milford, MA), which are integrated with SFC (Acquity UPC2) could prove to be very useful for routine nonconjugated estrogen analyses. The high sensitivity that can be obtained with modern LCSRM/MS will also permit the use of smaller volumes of biofluids to help conserve important plasma and serum samples. This will make it possible to use plasma and serum samples available from existing Biobanks without significantly depleting the total volume available. This could permit additional studies to be conducted on the same samples in order to help understand the factors that cause an increase in breast cancer risk.

Recent LC-SRM/MS assays have revealed that true serum levels of 16α-hydroxy-estrone are likely to be lower than previously reported. This should lead to re-evaluation of the importance of this metabolite, as it has been proposed to be involved in breast cancer progression [67-69]. Several studies were unable to detect non-conjugated 16α-hydroxy-estrone, while other studies found levels of the metabolite to be very close to the reported limits of quantification (Table 1). Reported levels of non-conjugated 16α-hydroxy-estradiol are similarly very close to the limits of quantification of the assays employed. This suggests that when these analytes are analyzed with greater sensitivity and specificity, serum concentration will actually be closer to 1 pg/mL. Intriguingly, very low levels of the conjugated forms of both 16α-hydroxy-estrone and 16α-hydroxy-estradiol have also been reported (Table 2). No methods have been developed to detect the intact β-glucuronide and sulfate conjugates of the 16α-hydroxy estrogens and so it is conceivable that the lack of detection could be due to incomplete hydrolysis by the G/S-based procedures normally employed. Therefore, there is a compelling need to confirm these findings using alternative methodology such as hydrolysis of conjugates with anhydrous hydrogen chloride [52] rather than by G/S.

The quantification of estrone-sulfate is particularly important as it is the majorcirculating form of estrone [60] and can potentially serve as a precursor to estrone in tissues through the action of sulfatases [5]. Furthermore, estrone-sulfate could potentially serve as a biomarker for the effectiveness of aromatase inhibitors [60]. It should be possible to detect low pg/mL levels by stable isotope dilution LC-SRM/MS, which would be approximately 1 % of the original circulating form, confirming complete inhibition of estrogen biosynthesis. Stable isotope dilution LC-SRM/MS assay methodology should be as specific as possible. Unfortunately, the only heavy isotope internal standard available for estrone-sulfate is the tetradeuterated form. Therefore, there is a critical need to synthesize the corresponding [¹³C]-analog of estrone-sulfate in order to overcome the problems inherent to use of deuterated internal standards. The [¹³C]-analog would additionally be stable to acid hydrolysis, overcoming any additional concerns that deuterium could exchange for protium during the analytical procedure.

Significant advances have been made in the development of LC-SRM/MS assays over the lastdecade, allowing increasingly sensitive and reliable quantification of serum estrogens and androgens. These advances in analytical methodology will facilitate the development of improved breast cancer risk models that incorporate serum concentrations of a comprehensive panel of estrogen and androgen metabolites. [30]. Previous studies have shown that such models have the potential to significantly improve breast cancer prevention [11,12]. The LCSRM/MS assays have potential utility for discovering biomarkers for the treatment and early detection of endometrial cancer as exemplified in the study of Audet-Walsh et al. [22]. The ability to routinely analyze serum and plasma estrogens and androgens with very high sensitivity and specificity by stable isotope dilution LC-SRM/MS is a promising avenue towards saving a large number of women from these devastating diseases [30,70].

Abbreviations used

2-hydroxy-estradiol

2,3-dihydroxy-17β-estradiol

2-methoxy-estradiol

2-methoxy-3-hydroxy-17β-estradiol

4-methoxy-estradiol

3-hydroxy-4-methoxy-17β-estradiol

4-hydroxy-estradiol

3,4-dihydroxy-17β-estradiol

16α-hydroxy-estradiol

3,16α-dihydroxy-17β-estradiol

estradiol

17β-estradiol

APCI

atmospheric pressure chemical ionization

BMI

body mass index

CYP

cytochrome P-450

D

dansyl

DHEA

dehydroepiandrosterone

ECAPCI

electron capture atmospheric pressure chemical ionization

ESI

electrospray ionization

GP

Girard P

GT

Girard T

G/S

β-glucuronidase/arylsulfatase

LC

liquid chromatography

HSD

hydroxysteroid dehydrogenase

MPPZ

1-(2,4-dinitro-5-fluorphenyl-4,4,-dimethylpiperazinyl

MS

mass spectrometry

NMN

N-methyl-nicotinyl

NMP

N-methyl-2-pyridyl

NMPS

N-methylpyridynium-3-sulfonyl

P

picolinoyl

PFB

pentafluorobenzyl

PS

pyridyl-3-sulfonyl

SRM

selected reaction monitoring

SULT

sulfotransferase