Anal Bioanal Chem (2016) 408:2293–2301 DOI 10.1007/s00216-016-9325-2
RESEARCH PAPER
Quantification of low molecular weight selenium metabolites in human plasma after treatment with selenite in pharmacological doses by LC-ICP-MS Konstantina Flouda 1 & Julie Maria Dersch 1 & Charlotte Gabel-Jensen 1 & Stefan Stürup 1 & Sougat Misra 2 & Mikael Björnstedt 2 & Bente Gammelgaard 1
Received: 7 December 2015 / Revised: 5 January 2016 / Accepted: 11 January 2016 / Published online: 1 February 2016 # Springer-Verlag Berlin Heidelberg 2016
Abstract The paper presents an analytical method for quantification of low molecular weight (LMW) selenium compounds in human plasma based on liquid chromatography inductively coupled plasma mass spectrometry (LC-ICP-MS) and post column isotope dilution-based quantification. Prior to analysis, samples were ultrafiltrated using a cut-off value of 3000 Da. The method was validated in aqueous solution as well as plasma using standards of selenomethionine (SeMet), Semethylselenocysteine (MeSeCys), selenite, and the selenosugar Se-methylseleno-N-acetylgalactosamine (SeGal) for linearity, precision, recoveries, and limits of detection and quantitation with satisfactory results. The method was applied for analysis of a set of plasma samples from cancer patients receiving selenite treatment in a clinical trial. Three LMW selenium compounds were observed. The main compounds, SeGal and selenite were tentatively identified by retention time matching with standards in different chromatographic systems, while the third minor compound was not identified. The identity of the selenosugar was verified by ESI-MS-MS product ion scanning, while selenite was identified indirectly as the glutathione (GSH) reaction product, GS-Se-SG. Keywords Selenium . Plasma . LC-ICP-MS . Speciation . ESI-MS * Bente Gammelgaard [email protected]
1
Department of Pharmacy, Faculty of Health and Medical Sciences, University of Copenhagen, Universitetsparken 2, 2100 Copenhagen, Denmark
2
Division of Pathology F46, Department of Laboratory Medicine, Karolinska Institutet, Karolinska University Hospital Huddinge, 141 86 Stockholm, Sweden
Introduction Selenium (Se) is an essential element and exerts its effect via selenoproteins, in which selenocysteine is the active center; the glutathione peroxidases and thioredoxin reductases are examples of selenoproteins [1, 2]. Selenium is introduced to the human body via the food chain, where selenite from soil is transformed to dietary Se represented by selenomethionine (SeMet) and selenocysteine (SeCys) as the main dietary compounds from plants and animals, respectively. The active selenoproteins are synthesized via the intermediate hydrogen selenide, HSe−, which can be methylated to methylselenol— another excretory intermediate in methylation pathways. The selenosugar, Se-methylseleno-N-acetylgalactosamine (SeGal) is the main urinary excretion product. The model was first introduced by Ganther [3], and during the following years further developed by contributions from several research groups as described in several reviews with the increased focus on speciation analysis [4–6]. It is a well-known fact that selenium plays a role in cancer prevention and cancer treatment [7, 8]. Both selenite and MeSeCys are cytotoxic to cancer cells while normal cells are marginally affected at equivalent pharmacological doses [9–11]. These compounds counteract multiple oncogenic signaling pathways that are important for the growth and survival of cancer cells, thereby exhibiting multi-target functions [12], a distinct advantage for cancer chemotherapeutics regarding tumor heterogeneity. A, first in human, phase I clinical trial on the cancer therapeutic effects of selenite has recently been published and the study group will now continue with an improved protocol prior to entering phase II [13]. Another study was initialized in 2014 for identifying the safest most effective selenium compound for chronic lymphocytic leukemia or metastatic prostate cancer patients [14]. Several studies on the cancer preventive effect have
2294
been performed with disputed conclusions [15, 16] but none of these involved Se speciation studies. Hence, information on underlying chemistry was not pursued. The need for speciation analysis is increasingly recognized [17]. The state of the art in selenium speciation analysis is a combination of a separation method with elemental MS for quantification and molecular MS for identification. Several recent reviews have addressed this subject [18–21]. With the introduction of more sensitive and selective molecular MS instruments, more analytical methods based on this technique have emerged, including a method for quantification of low molecular weight (LMW) Se compounds in urine [22], LMW profiling in human serum and urine [23], and liver and kidney from rats [24]. Quantification is usually performed by comparison with standards, however standards are not always available and using one standard for all compounds may be problematic as sensitivity may change during a chromatic elution owing to elution of higher amounts of salts in the beginning and thereby change of ion suppression effects throughout the chromatogram. Isotope dilution analysis (IDA) quantification can be used to overcome this problem. The technique is based on measuring isotope ratios after spiking the sample with a solution enriched with an isotope of minor natural abundance. In species-unspecific IDA, also known as post column IDA, the spike is added to the chromatographic eluent after the separation. A tutorial review on the technique and the equations involved in the calculations has been published by Rodriguez-Gonzáles et al. [25]. Sample treatment needed is highly dependent on the matrix and the following detection method. Common procedures for plasma samples is removal of proteins by precipitation or ultrafiltration in which the centrifugal force is used to separate high molecular weight (HMW) compounds from LMW by passage through a semipermeable membrane with a molecular weight cut-off value. The purpose of this study was to develop an analytical method for quantitative and qualitative analysis of LMW compounds in plasma from patients treated intravenously with selenite. The method should be applied in analysis of plasma samples from patients participating in a clinical trial studying sodium selenite as a cytotoxic agent in advanced carcinoma.
Experimental Instrumentation Chromatography was carried out using an Agilent 1100 series HPLC system (Agilent Technologies, Waldbronn Germany) consisting of a degasser, a pump, a variable wavelength detector, an auto sampler (Agilent Technologies,
K. Flouda et al.
Waldbronn, Germany) and controlled by the ChemStation software (Agilent Technologies, Waldbronn, Germany), the chromatographic conditions are listed in Table 1. The HPLC system was coupled to a Sciex ELAN DRCe or a Sciex Elan 6000 (Perkin Elmer, Norwalk, CT, USA). The ICPMS systems were equipped with a jacketed cyclonic spray chamber and a Micro Mist nebulizer (ARG-1-UM02 from LABsupport, Hillerød, Denmark). ICP-MS sampler and skimmer cones were made of nickel. The plasma and auxiliary gas flow rates were 15 and 1.2 L min−1, respectively. The ELAN DRCe was used in DRC mode with methane as reaction gas at a flow of 0.55 L min−1. Nebulizer gas flow, lens voltage, and RF power were optimized daily on a 100 μg L−1 TMSe solution in mobile phase. The instruments were run using Elan software version 3.4 (Perkin Elmer, Norwalk, CT, USA) and the isotopes 77Se, 80Se, and 82Se were monitored on the ELAN DRCe, while 77Se, 78 Se, and 82Se were monitored on the ELAN 6000. Data acquisition was: dwell time 200 ms, sweeps per reading 1, and readings per replicate 1. The number of replicates was adjusted to chromatographic runtime. For the post column isotope dilution analysis (IDA), the enriched 7 7 Se standard solution in mobile phase (150 μg L−1) was continuously introduced (at 10 μL min−1) using a syringe pump, Harvard Apparatus 11plus (Holliston, MA, USA) with a 5 mL syringe (Hamilton, Bonaduz, Switzerland). The pump rate was controlled on a regular basis by weighing the eluate from 5 min pumping. The mass per minute was converted to volume per minute (flow rate) correcting for the relevant density. The syringe pump was connected to the eluent flow from the HPLC by a static mixing T-piece. The connections before mixing were made of PEEK with an inner diameter of 0.005 mm and after mixing with an inner diameter of 0.010 mm. The isotope ratios were calculated and corrected for mass bias by daily measuring of the experimental isotope ratios of 80Se/77Se using a linear mathematical model [25], while The theoretical isotope ratios were calculated by using the natural abundance [26]. Measurements of blanks showed that the measurements were not completely free of interference. However, the error introduced by interference was considered small and together with the correction for mass bias satisfactory results could be obtained. This was verified by chromatographic analysis of a certified selenite standard. The isotope dilution equation was applied to each data point and the resulting chromatographic peak was integrated. Converting the mass to concentration based on the injected amount, resulted in an accuracy of 99.5 ± 1.1 % (n = 3), which was considered satisfactory. The mass bias was determined daily using a 100 μg L−1 solution of TMSe. The dead time was corrected automatically by the instrument. The isotope dilution equation [27] was
LMW selenium metabolites in human plasma after selenite treatment Table 1 Chromatographic conditions
2295
LC-ICP-MS Reversed phase (RP) Columna Injection volume (μL) Flow rate (μL min−1) Mobile phase (A) Anion exchange (AEX) Columna Injection volume (μL) Flow rate (μL min−1) Mobile phase (B) Size exclusion (SE) Column* Injection volume (μL) Flow rate (μL min−1) Mobile phase (C) LC-ESI-MS Column Column temperature (°C) Injection volume (μL) Flow rate (μL min−1) Mobile phase
Gemini 5 μm C18 (250 × 2 mm) (Phenomenex) 10 200 200 mM Ammonium acetate, 5 %, MeOH, pH 6.7 IonPac AS11-HC (2 × 250 mm) (Dionex) 10 200 5 mM Ammonium citrate, 2 % MeOH, pH 9 Bio SEC-3 (4.6 × 300 mm) (Agilent) 5 350 20 mM Ammonium acetate-2 % MeOH, pH 7 Kinetex C18, 2.6 μm, 100 Å. 2.1 mm ID × 100 mm (Phenomenex) 50 5 500 A: 0.1 % formic acid, 2 % acetonitrile B: 0.1 % formic acid, 95 % acetonitrile Linear gradient: 0–4 min: 0–80 % B, 4–7 min: 100 % A
a
All analytical columns were protected by the respective guard columns
applied to each point of the chromatogram to obtain the corresponding mass flow chromatogram. Peak areas were calculated by integration of the chromatographic peaks using the OriginPro software (version 9.1, OriginLab Corporation, Northampton, MA, USA). LC-ESI-MS analyses were performed using a ThermoFinnigan Q-Exactive mass-spectrometer (Thermo-Finnigan, San Jose, CA), coupled to a Ultimate 3000 UHPLC-system (Thermo-Finnigan, San Jose, CA). The mass spectrometer was equipped with a heated electrospray ionization source operated in the positive ionization mode. All compounds of interest were detected as the [M + H]+ adducts. Ionization settings were optimized using direct infusion of the relevant standard dissolved in the mobile phase (75 % A and 25 % B) at a flow rate of 0.5 ml/min. For MS2 analyses, the normalized collision energy (NCE) was set to 10. For full scan experiments, the instrument resolution was 70,000; in the t-ms2 (targeted ms-ms) mode, the instrument resolution was 17,500. Chemicals and reagents Chemicals Selenomethionine (SeMet), Se-methylselenocysteine (MeSeCys) and sodium selenite were all of analytical grade
and from Sigma (St. Louis, MO, USA). Trimethylselenonium (TMSe) was synthesized according to Foster and Ganther [28]. Se-methylseleno-N-acetylgalactosamine (SeGal) was synthesized at the Department of Medicinal Chemistry (University of Copenhagen). Ammonium acetate (≥98 %), ammonium citrate dibasic and formic acid were from Sigma. Acetonitrile was from Chemsolute (Th. Geyer, Germany). All chemicals were of analytical grade or better. Methanol (for LC-MS, >99.95 %) was from Th. Geyer (Renningen, Germany). Deionized water (18.2 MΩ cm−1) from a Milli-Q plus unit (Millipore, Bedford, MA, USA) was used throughout. Standards Stock standard solutions of TMSe, SeMet, MeSeCys, SeGal, and selenite were prepared in a concentration of 10 mg L −1 in MilliQ water. The stock solutions were standardized against a 1.001 g L−1 Se pure atomic spectroscopy standard (Perkin Elmer, Norwalk, CT, USA). The standards were stored at 4 °C. Spiking standard solutions in a concentration 500 μg Se L−1 were prepared by dilution of stock solutions in MilliQ water.
2296
The certified enriched elemental 7 7 Se (99.2 %, 1,000 μg mL−1 in 5 % HNO3) was donated from Erik H. Larsen (Technical University of Denmark, Søborg, Denmark).
Samples Human plasma samples frozen in aliquots of 250 μL from the SECAR-study [14] were used in the analyses. A human plasma sample from a healthy volunteer was used for validation purposes.
Procedures
K. Flouda et al.
Precision The precision was calculated as the relative standard deviation (% RSD) for triplicate analysis at the levels of 10, 20 and 40 μg Se L−1 in mobile phase (A) as well as human plasma ultrafiltrates. Limit of detection and limit of quantitation (LOQ) Standards of 5 μg Se L−1 concentration in mobile phase (A) and plasma were analyzed in triplicate. Limit of detection (LOD) and limit of quantitation (LOQ) were calculated as 3.3 and 10 times, respectively, the standard deviation of the peak areas divided by the slope of the standard curves.
Plasma sample preparation GS-Se-SG formation The frozen samples were thawed at room temperature before preparation; 200 μL sample aliquots were transferred to Vivaspin® 500 μL centrifugal filter units (Sartorius AG, Göttingen, Germany) with a semi-permeable membrane of molecular weight cut-off (MWCO) 3000 Da and were centrifuged at 14,000×g for 30 min. The filtrates in the lower compartment were transferred to LC vials and analyzed.
An aqueous standard containing 1.5 μM selenite + 0.5 mM GSH was prepared in 10 % formic acid. Selenite was spiked to a blank plasma sample before MWCO filtration to a final concentration of 0.15 μM. After filtration, GSH was added to a concentration of 0.5 mM and formic acid to a final concentration of 10 %. A pooled plasma sample was added GSH to a final concentration of 0.5 mM and a final formic acid concentration of 10 %.
Filter recovery Ten and 20 μg L−1 samples were prepared in triplicates by spiking untreated human plasma as well as ultrafiltrated plasma and analyzed. Recoveries were calculated as the ratios of RP-LC areas of plasma samples spiked before and after filtration.
RP column recovery Standards of MeSeCys, SeMet, SeGal, and selenite were prepared in mobile phase (A) with a concentration of 50 μg Se L−1. All standards were prepared in duplicates and were measured three times by RP-LC-ICP-MS and three times by flow injection analysis (FIA). The dwell time was changed to 100 ms per isotope. Recoveries were calculated as the ratio of RP-LC areas to FIA areas.
Linearity Linearity was examined in mobile phase (A) and plasma. MeSeCys, SeMet, SeGal, and selenite standards were prepared in five different concentrations: 5, 10, 20, 40 and 50 μg Se L−1 in triplicates. The aqueous standards were analyzed directly, while the spiked plasma standards were analyzed after ultrafiltration.
Results and discussion Method development and validation The intended use of the developed method was LMW analysis of plasma samples from cancer patients treated with large doses of selenite. Although the patients were treated with selenite, this compound was not expected to be present in substantial amounts as ingested selenite is supposed to be taken up by red blood cells where it is reduced by GSH, effluxed, and subsequently transferred to the liver bound to albumin [29, 30]. Hence, a reversed phase method was developed based on separation of the standards SeMet, MeSeCys, and SeGal in 200 mM ammonium acetate, pH 7 in 5 % methanol. A chromatogram of the standards is presented in Fig. 1. The method validation was first performed on these three standards, but as the following analysis of plasma samples showed an unidentified peak eluting at 3.3 min in several samples and the most likely identity of this peak was selenite, selenite was also included in the validation. The method was not validated for the excretory metabolite trimethylselenonium (TMSe) as serum background levels and levels after selenium supplementation with different Se species (including selenite) have been reported to be less than 0.1 ng mL−1 and only detected after pre-concentration [22].
LMW selenium metabolites in human plasma after selenite treatment
2297
concentration range of 5–50 μg Se L−1. The correlation coefficient (R) for all species in aqueous and plasma matrix was > 0.99. The intersections with the y-axis were not significantly different from zero (95 % confidence interval). Residual plots showed that a linear regression model was appropriate for the data.
50000
Intensity (cps)
40000 30000 20000 10000
Precision 0 0
1
2
3
4 5 Time (min)
6
7 −1
Fig. 1 Reversed phase chromatogram of a 50 μg Se L MeSeCys, SeMet and SeGal. Monitored isotope: 80Se
8
9
standard of
Initially, the column recovery was examined for the four species. As shown in Table 2, the column recovery varied between 91.8 and 102.4 %, which was considered satisfactory. Plasma proteins should be removed prior to analysis to obtain the LMW fraction. Initial experiments with protein precipitation by addition of organic solvents were not successful owing to poor precision. Instead, it was chosen to remove proteins by ultrafiltration based on a semipermeable membrane with a molecular weight cut-off (MWCO) value of 3000. In a study of Greening and Simpson [31], several commercially available filters of similar MWCO but with different types of membranes were tested. Only the Sartorius Vivaspin® was effective in recovery and enrichment of LMW components of human plasma. The filter of this type was therefore a natural choice for the plasma sample preparation. To determine the filter recovery, spiked plasma samples were filtrated and the areas compared to plasma ultrafiltrates spiked after filtration at the concentration levels 10 and 20 μg Se L−1, respectively. The filter recoveries are presented in Table 2. The obtained filter recoveries were in the range of 97–107 %, which was considered to be satisfactory.
Precision was determined as the relative standard deviation (RSD) at three concentration levels, 10, 20 and 40 μg Se L−1 and the results are summarized in Table 3. The precision in aqueous standards was in the range of 0.3–4.9 % and between 0.5 and 8.3 % in the plasma matrix. The precision was considered acceptable. Accuracy As no certified standards are available for MeSeCys, SeMet and the selenosugar, the accuracy was determined by analyzing a certified selenite standard by the isotope dilution procedure. The accuracy of the isotope dilution procedure was determined to 99.5 ± 1.1 % (n = 3). The filter recoveries, determined as the ratio of RP-LC areas of plasma samples spiked before and after filtration, were in the range 97–107 % as presented in Table 2. The sample preparation by filtration is thus contributing the most to inaccuracy. LOD and LOQ The LOD and LOQ in mobile phase and plasma are given in Table 4. The obtained LODs were comparable
Table 3 Precision given as the relative standard deviation in mobile phase (MP) and plasma matrix
Linearity Linearity was established for MeSeCys, SeMe, SeGal, and selenite in both mobile phase A and plasma matrix, in a
MeSeCys Table 2 C18 column recoveries calculated as the ratio of RP-LC areas and FIA areas. Filter recoveries calculated as the ratio of RP-LC areas of plasma samples spiked before and after ultrafiltration SeMet Column recovery (%) (n = 6) Filter recovery (%) (n = 3) 10 μg Se L−1 20 μg Se L−1 MeSeCys 97.3 ± 3.5 SeMet 102.4 ± 1.4 SeGal 100.7 ± 0.84 Selenite 91.8 ± 0.5
98.0 ± 3.2 100.6 ± 1.9 96.9 ± 3.1 104.5 ± 0.7
106.6 ± 10.7 101.5 ± 9.5 101.6 ± 9.0 101.7 ± 0.4
SeGal
Selenite
C (μg L−1)
MP (n = 3) RSD%
Plasma (n = 3) RSD%
10 20 40 10 20 40 10 20 40 10 20 40
1.46 3.28 0.93 2.55 2.25 1.12 3.57 4.85 1.78 1.50 0.63 0.32
4.47 5.52 0.48 3.41 5.56 5.49 4.69 5.29 7.24 8.29 7.53 3.98
2298
K. Flouda et al.
Table 4 LOD and LOQ, calculated based on the deviation of the peak areas at the 5 μg L−1 level in mobile phase (MP) and plasma matrix
MP (n = 3)
Plasma (n = 3)
LOD (μg L−1)
LOQ (μg L−1)
LOQ (μg L−1)
MeSeCys
0.25
0.77
0.49
1.49
SeMet
0.28
0.84
2.11
6.40
SeGal Selenite
0.35 0.30
1.07 0.92
1.83 0.44
5.55 1.33
Generally, three peaks could be observed in most of the analyzed samples. Peak 3 eluting at 6.7 min was tentatively identified by retention time matching and spiking with SeGal. This peak was observed in most samples after selenite treatment. This is in line with previous studies as SeGal is the major metabolite in mammals [33, 34]. Recently, Kokarnig et al. have reported the presence of traces of selenite, TMSe and MeSeCys in background serum based on retention time matching [23] and in addition the presence of SeGal and SeMet after SeMet supplementation. On the other hand, they reported that total serum Se levels did not increase after selenite supplementation [23]. Peaks 1 and 2 did not match with the retention times of MeSeCys or SeMet. As the unknown peak 2 was only present in very small amounts even when the total Se content was high, identification of this compound was not further pursued. Peak 1 co-eluted with selenite, but this was not considered sufficient evidence for the presence of selenite as it eluted close to the void volume. Analysis of a high Se concentration sample on cation exchange chromatography showed elution of the majority of the sample in the front, thus the unknown peak 1 was not TMSe or any other cationic species; but in a size exclusion
to previously reported values of LMW Se speciation in biological samples [32].
Quantitation of LMW Se species The quantification of Se concentration in the plasma samples by RP-LC-ICP-MS was accomplished by IDA with the ICP-MS detection as described in the BExperimental^ section. The mass bias factor (F) was calculated and the original ICP-MS intensities were imported into an Excel spreadsheet. Once the mass bias factor was estimated the experimental isotope ratios 80Se/77Se were corrected. The dead time was corrected automatically by the instrument. After the corrections, the isotope dilution equation [25] was applied to the whole chromatogram to obtain the mass flow data. Finally, the data were imported to OriginPro software for integration. About 20 samples were injected in duplicate. Figure 2a shows typical mass flow chromatograms from human plasma analyses in duplicate showing identical chromatograms, while Fig. 2b shows the mass flow chromatograms of the standards.
a)
b) 0.0014
MeSeCys
0.0030
3
0.0012
SeMet
0.0025 Mass flow (µg/min
Mass flow (µg/min)
LOD (μg L−1)
0.001 0.0008
1 2
0.0006 0.0004
SeGal
0.0020 0.0015 0.0010 0.0005
0.0002
0.0000
0 0
1
2
3
4
5
6
7
8
Time (min)
Fig. 2 a Examples of mass flow chromatograms from plasma samples. Lower chromatograms: sample before selenite treatment; upper chromatograms (off-set by 0.0002 μg min−1): sample after treatment. The total LMW concentration of the samples were 0.9 and 35.5 μg L−1,
9
0
1
2
3
4
5
6
7
8
9
Time (min)
respectively. Each sample was analyzed in duplicate, blue and red trace, respectively. b Mass flow chromatograms of a 50 μg Se L−1 standard of MeSCys, SeMet, and SeGal
LMW selenium metabolites in human plasma after selenite treatment
system as well as an anion exchange system, the retention time of peak 1 matched selenite retention, which indicated the presence of selenite (Fig. 3). Traces of selenite in serum based on retention time matching by LC-ICP-MS analysis has previously been reported, but at levels hardly distinguished from background [35, 36]. The possibility that selenite is a breakdown product of selenoproteins during storage of samples has also been suggested [37]. However, the presence of selenite in plasma has never been confirmed by molecular MS. The samples in the present study were from patients treated with selenite in doses between 1 and 15.3 mg Se/m2 by infusion corresponding to between 0.5 and 7.7 mg Se, if the body surface is set to 2 m2. These doses are much larger than the doses from supplements, which normally are in the 50–100 μg range. The LMW selenium compounds only constituted a minor fraction of the plasma Se—between 1 and 28 % with a mean of 8 %. Thus, the major Se amount must be accumulated in the proteins either as incorporated in the protein or attached to proteins. Another possibility is elemental selenium. A general pattern in patient samples was increased selenite signals
82
Se Intensity (counts)
Size exclusion chromatography 8000
plasma
7000
plasma spiked
6000 5000 4000
2299
immediately after selenite application followed by a decline over time. These data will be subject of a future study on the pharmacokinetics of selenite in plasma.
Identification with LC-ESI-MS To verify the identity of selenite in the samples, an ultrafiltrate of a pooled plasma sample was exposed to LC-ESIMS in negative ionization mode. To identify selenite in a biological sample, selenite has to be separated from the matrix. In the RP chromatography mode, selenite elutes together with a large amount of other small ions resulting in large ion suppression in the MS. Anion exchange chromatography retained selenite, but the high ionic strength of the citrate eluent caused severe ion suppression. Instead, an indirect approach was taken by formation of the GSH adduct. Selenite is known to form a trisulfide by reaction with GSH following the reaction scheme: 4GSH + SeO32− → GSSG + GS-Se-SG + 2OH− + H2O [38, 39]. Thus, treatment of plasma ultrafiltrate with GSH should result in a compound that could be retained on an RP column and following formation of an ion with the exact mass to charge ratio of 693.076 in ESI-MS positive ionization mode. The extracted ion chromatogram of a pooled plasma sample after addition of GSH is shown in Fig. 4. It appears from the figure, that no signals were obtained in blank plasma or the pooled plasma sample prior to GSH addition, while a large signal corresponding to the GS-SeSG adduct was obtained after GSH addition. The retention
3000 2000 1000 0 0
1
2
3
4
5
6
7
8
9
10
11
12
13
6000000
Time (min)
plasma plasma spiked
12000 10000 8000 6000
Selenite standard + GSH
4000000
Blank plasma + selenite + GSH Blank plasma
3000000
Pooled plasma + GSH
2000000 Pooled plasma
4000
82
Se Intensity (counts)
Anion exchange chromatography
Total Ion Current
5000000
14000
1000000
2000 0 0
1
2
3
4
5
6
7
8
9
10
11
12
Time (min)
Fig. 3 Retention time matching of selenite by spiking plasma samples with standard in size exclusion chromatography and anion exchange chromatography. Chromatograms are offset by 1000 and 2000 counts, respectively
0 0
1 Time / minutes
2
Fig. 4 Extracted ion chromatograms at m/z 693.076 from RP chromatography of selenite standards and samples after addition of GSH. Chromatograms are offset
2300
K. Flouda et al.
Fig. 5 Extracted spectrum from the TIC of the plasma sample peak from RP chromatography of the sample (retention time 1.47– 1.51) and the calculated spectrum
time of this signal matched the retention time of a selenite standard mixed with GSH and a plasma sample spiked with selenite + GSH. The spectrum obtained from the plasma sample and the theoretical spectrum with the characteristic Se isotope pattern of the five most abundant isotopes (76Se 9.4 %, 77Se 7.6 %, 78Se 23.8 %, 80Se 49.6 %, 82 Se 8.7) shown in Fig. 5 are almost identical, which leads to the conclusion that selenite was identified in the sample. The derivatization method was not used for quantitative purposes, thus quantitative derivatization was not a demand. The finding of selenite is not surprising as the patients were treated with large amounts of selenite, however neither increased selenite levels in plasma after selenite treatment, nor evidence based on ESI-MS has not been reported before. A reason for this finding could be that the patients were very ill and simultaneously treated with large amounts of drugs, which could inhibit or exhaust normal metabolic pathways. The identity of the selenosugar was verified by LC-ESI-MS by targeted MS-MS product ion scans on m/z 302, 300, 298,
297, and 296. For clarity, only three of these are shown in Fig. 6, where scans from the RP chromatographic peaks are compared for a selenosugar standard and a plasma sample. It appears from the figure that the product ions 204.087, 186.076, 168.066, 144.066, 138.055, and 126.055 were all produced in the standard as well as the plasma sample. The intensities of the peaks varied, probably owing to the very different matrices. However, the relative contributions from the isotopes were in agreement with relative abundances of the isotopes. Thus, the identity of the selenosugar was considered verified. These ions all represent losses of selenium moieties from the sugar molecule. The formation of these product ions by selective reaction monitoring (SRM) have previously been reported for a selenosugar standard [40], some of them in biological samples as rat liver and kidney cytosol after purification and pre-concentration (204, 186, 144, and 138) [24], human urine (204, 186, 144, 138) [22] and serum (204, 186, 144) [23], but not all of them in one study. The present scans were performed on an Orbitrap instrument, which has the sensitivity to make this possible.
Fig. 6 Product ion scans of a selenosugar standard and b a plasma sample collected from the respective peaks from RP chromatography. The scans are overlaid. Black star Background peak
LMW selenium metabolites in human plasma after selenite treatment
Conclusion A validated analytical method for quantitation of LMW selenium compounds in human plasma using RP-LC-ICP-MS was developed. The method was based on ultrafiltration sample preparation and quantitation by post-column IDA. The validation parameters for recoveries, linearity, precision, LOD, and LOQ, were considered satisfactory for analysis of plasma samples. The method was applied for analysis of plasma samples from patients receiving selenite treatment in a clinical trial. Three selenium species were observed in varying amounts in the plasma samples. Two of these were tentatively identified as SeGal and selenite by retention time matching with standards, while the third minor metabolite was not identified. The identity of selenite was verified by molecular MS as the glutathione adduct GS-Se-SG, while the identity of SeGal was verified by product ion scans on the masses of the sugar containing the five most abundant isotopes of selenium. Acknowledgments The authors wish to thank Erik Huusfeldt Larsen for the donation of the enriched 77Se solution and laboratory technician Camilla Jensen for skillful technical assistance. The SECAR trial was supported by grants from Cancerfonden and Jochnick Foundation, Sweden.
2301 10. 11. 12. 13. 14.
15. 16. 17. 18. 19. 20. 21. 22. 23. 24.
Compliance with ethical standards The SECAR-trial [14] was approved by the Swedish Medical Products Agency and registered in the EU Clinical Trial Register (Eudra CT Number: 2006-004076-13) and by the Ethical Committee of Stockholm (2006/429-31/3). Informed consent was given by the healthy volunteer donating the plasma sample for validation.
25. 26. 27.
Conflict of interest The authors declare that they have no conflict of interest.
References 1.
Kryukov G, Kryukov V, Gladyshev V. J Biol Chem. 1999;274: 33888–97. 2. Ganther HE. Carcinogenesis. 1999;20:1657–66. 3. Ganther HE. J Am Coll Tocicol. 1986;5:1–5. 4. Weekley CM, Harris HH. Chem Soc Rev. 2013;42:8870–94. 5. Suzuki KT. J Health Sci. 2005;51:107–14. 6. Gammelgaard B, Jackson MI, Gabel-Jensen C. Anal Bioanal Chem. 2011;399:1743–63. 7. Rayman MP. Proc Nutrition Soc. 2005;64:527–42. 8. Björnstedt M, Fernandes A. EPMA J. 2010;1:389–95. 9. Nilsonne G, Sun X, Nyström C, Rundlöf AK, Potamitou Fernandes A, Björnstedt M, et al. Free Rad Biol Med. 2006;41:874–85.
28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40.
Menter DG, Sabichi AL, Lippman SM. Cancer Epidemiol Biomarkers Prev. 2000;9:1171–82. Husbeck B, Nonn L, Peehl DM, Knox SJ. Prostate. 2006;66:218– 25. Hanahan D, Weinberg R. Cell. 2011;144:646–74. Brodin O, Eksborg S, Wallenberg M, Asker-Hagelberg C, Larsen EH, Molkert D, et al. Nutrient. 2015;7:4978–94. Randomised trial to identify the safest and most effective selenium compound for cancer patients. Phase Ib randomised double-blind dose-escalation trial to identify the safest most effective selenium compound for use in patients with either chronic lymphocytic leukaemia or metastatic prostate cancer . https://www.anzctr.org.au/ Trial/Registration/TrialReview.aspx?ACTRN=12613000118707. 2015. Lippman S, Klein E, Goodman P, Lucia M, Thompson I, Ford L, et al. JAMA. 2009;301:39–51. Clark LC, Combs GF, Turnbull BW, Slate EH, Chalker DK, Chow J, et al. JAMA. 1996;276:1957–63. Hurst R, Armah CN, Dainty JR, Hart DJ, Teucher B, Goldson AJ, et al. Am J Clin Nutr. 2010;91:923–31. Ogra Y, Anan Y. J Anal At Spectr. 2009;24:1477–88. Dumont E, Vanhaecke F, Cornelis R. Anal Bioanal Chem. 2006;385:1304–23. Pedrero Z, Madrid Y. Anal Chim Acta. 2009;634:135–52. Gammelgaard B, Gabel-Jensen C, Stürup S, Hansen HR. Anal Bioanal Chem. 2008;390:1691–706. Lu Y, Rumpler A, Francesconi KA, Pergantis SA. Anal Chim Acta. 2012;731:49–59. Kokarnig S, Tsirigotaki A, Wiesenhofer T, Lackner V, Francesconi KA, Pergantis SA, et al. J Trace Elem Med Biol. 2015;29:83–90. Gonzalez-Iglesias H, Fernandez-Sanchez ML, Lu Y, Fernandez Menendez S, Pergantis SA, Sanz-Medel A. J Anal At Spectr. 2015;30:267–76. Rodriguez-Gonzalez P, Marchante-Gayon JM, Alonso JIG, SanzMedel A. Spectrochim Acta Atom Spectros. 2005;60:151–207. Berglund M, Frech W, Baxter DC. Spectrochim Acta. 1991;46B: 1767–77. Sariego Muniz C, Marchante Gayon JM, Garcia Alonso JI, SanzMedel A. J Anal At Spectr. 2001;16:587–92. Foster SJ, Ganther HE. Anal Biochem. 1984;137:205–9. Suzuki T, Shiobara Y, Itoh M, Ohmichi M. Analyst. 1998;123:63– 7. Shiobara Y, Ogra Y, Suzuki KT. Analyst. 1999;124:1237–41. Greening DW, Simpson RJ. J Proteomics. 2010;73:637–48. Gammelgaard B, Bendahl L, Jacobsen NW, Stürup S. J Anal At Spectr. 2005;9:889–93. Ogra Y, Ishiwata K, Takayama H, Aimi N, Suzuki KT. J Chromatogr B. 2002;767:301–12. Gammelgaard B, Bendahl L. J Anal At Spectr. 2004;19:135–42. Jitaru P, Goenaga-Infante H, Vaslin-Reimann S, Fisicaro P. Anal Chim Acta. 2010;657:100–7. Palacios O, Ruiz Encinar J, Schaumlöffel D, Lobinski R. Anal Bioanal Chem. 2006;384:1276–83. Palacios O, Lobinski R. Talanta. 2007;71:1813–6. Ganther HE. Biochemistry. 1968;7:2898–905. Arteel GE, Sies H. Environ Toxicol Pharmacol. 2001;10:153–8. Dumont E, Ogra Y, Suzuki KT, Vanhaecke F, Cornelis R. Anal Chim Acta. 2006;555:25–33.
RESEARCH PAPER
Quantification of low molecular weight selenium metabolites in human plasma after treatment with selenite in pharmacological doses by LC-ICP-MS Konstantina Flouda 1 & Julie Maria Dersch 1 & Charlotte Gabel-Jensen 1 & Stefan Stürup 1 & Sougat Misra 2 & Mikael Björnstedt 2 & Bente Gammelgaard 1
Received: 7 December 2015 / Revised: 5 January 2016 / Accepted: 11 January 2016 / Published online: 1 February 2016 # Springer-Verlag Berlin Heidelberg 2016
Abstract The paper presents an analytical method for quantification of low molecular weight (LMW) selenium compounds in human plasma based on liquid chromatography inductively coupled plasma mass spectrometry (LC-ICP-MS) and post column isotope dilution-based quantification. Prior to analysis, samples were ultrafiltrated using a cut-off value of 3000 Da. The method was validated in aqueous solution as well as plasma using standards of selenomethionine (SeMet), Semethylselenocysteine (MeSeCys), selenite, and the selenosugar Se-methylseleno-N-acetylgalactosamine (SeGal) for linearity, precision, recoveries, and limits of detection and quantitation with satisfactory results. The method was applied for analysis of a set of plasma samples from cancer patients receiving selenite treatment in a clinical trial. Three LMW selenium compounds were observed. The main compounds, SeGal and selenite were tentatively identified by retention time matching with standards in different chromatographic systems, while the third minor compound was not identified. The identity of the selenosugar was verified by ESI-MS-MS product ion scanning, while selenite was identified indirectly as the glutathione (GSH) reaction product, GS-Se-SG. Keywords Selenium . Plasma . LC-ICP-MS . Speciation . ESI-MS * Bente Gammelgaard [email protected]
1
Department of Pharmacy, Faculty of Health and Medical Sciences, University of Copenhagen, Universitetsparken 2, 2100 Copenhagen, Denmark
2
Division of Pathology F46, Department of Laboratory Medicine, Karolinska Institutet, Karolinska University Hospital Huddinge, 141 86 Stockholm, Sweden
Introduction Selenium (Se) is an essential element and exerts its effect via selenoproteins, in which selenocysteine is the active center; the glutathione peroxidases and thioredoxin reductases are examples of selenoproteins [1, 2]. Selenium is introduced to the human body via the food chain, where selenite from soil is transformed to dietary Se represented by selenomethionine (SeMet) and selenocysteine (SeCys) as the main dietary compounds from plants and animals, respectively. The active selenoproteins are synthesized via the intermediate hydrogen selenide, HSe−, which can be methylated to methylselenol— another excretory intermediate in methylation pathways. The selenosugar, Se-methylseleno-N-acetylgalactosamine (SeGal) is the main urinary excretion product. The model was first introduced by Ganther [3], and during the following years further developed by contributions from several research groups as described in several reviews with the increased focus on speciation analysis [4–6]. It is a well-known fact that selenium plays a role in cancer prevention and cancer treatment [7, 8]. Both selenite and MeSeCys are cytotoxic to cancer cells while normal cells are marginally affected at equivalent pharmacological doses [9–11]. These compounds counteract multiple oncogenic signaling pathways that are important for the growth and survival of cancer cells, thereby exhibiting multi-target functions [12], a distinct advantage for cancer chemotherapeutics regarding tumor heterogeneity. A, first in human, phase I clinical trial on the cancer therapeutic effects of selenite has recently been published and the study group will now continue with an improved protocol prior to entering phase II [13]. Another study was initialized in 2014 for identifying the safest most effective selenium compound for chronic lymphocytic leukemia or metastatic prostate cancer patients [14]. Several studies on the cancer preventive effect have
2294
been performed with disputed conclusions [15, 16] but none of these involved Se speciation studies. Hence, information on underlying chemistry was not pursued. The need for speciation analysis is increasingly recognized [17]. The state of the art in selenium speciation analysis is a combination of a separation method with elemental MS for quantification and molecular MS for identification. Several recent reviews have addressed this subject [18–21]. With the introduction of more sensitive and selective molecular MS instruments, more analytical methods based on this technique have emerged, including a method for quantification of low molecular weight (LMW) Se compounds in urine [22], LMW profiling in human serum and urine [23], and liver and kidney from rats [24]. Quantification is usually performed by comparison with standards, however standards are not always available and using one standard for all compounds may be problematic as sensitivity may change during a chromatic elution owing to elution of higher amounts of salts in the beginning and thereby change of ion suppression effects throughout the chromatogram. Isotope dilution analysis (IDA) quantification can be used to overcome this problem. The technique is based on measuring isotope ratios after spiking the sample with a solution enriched with an isotope of minor natural abundance. In species-unspecific IDA, also known as post column IDA, the spike is added to the chromatographic eluent after the separation. A tutorial review on the technique and the equations involved in the calculations has been published by Rodriguez-Gonzáles et al. [25]. Sample treatment needed is highly dependent on the matrix and the following detection method. Common procedures for plasma samples is removal of proteins by precipitation or ultrafiltration in which the centrifugal force is used to separate high molecular weight (HMW) compounds from LMW by passage through a semipermeable membrane with a molecular weight cut-off value. The purpose of this study was to develop an analytical method for quantitative and qualitative analysis of LMW compounds in plasma from patients treated intravenously with selenite. The method should be applied in analysis of plasma samples from patients participating in a clinical trial studying sodium selenite as a cytotoxic agent in advanced carcinoma.
Experimental Instrumentation Chromatography was carried out using an Agilent 1100 series HPLC system (Agilent Technologies, Waldbronn Germany) consisting of a degasser, a pump, a variable wavelength detector, an auto sampler (Agilent Technologies,
K. Flouda et al.
Waldbronn, Germany) and controlled by the ChemStation software (Agilent Technologies, Waldbronn, Germany), the chromatographic conditions are listed in Table 1. The HPLC system was coupled to a Sciex ELAN DRCe or a Sciex Elan 6000 (Perkin Elmer, Norwalk, CT, USA). The ICPMS systems were equipped with a jacketed cyclonic spray chamber and a Micro Mist nebulizer (ARG-1-UM02 from LABsupport, Hillerød, Denmark). ICP-MS sampler and skimmer cones were made of nickel. The plasma and auxiliary gas flow rates were 15 and 1.2 L min−1, respectively. The ELAN DRCe was used in DRC mode with methane as reaction gas at a flow of 0.55 L min−1. Nebulizer gas flow, lens voltage, and RF power were optimized daily on a 100 μg L−1 TMSe solution in mobile phase. The instruments were run using Elan software version 3.4 (Perkin Elmer, Norwalk, CT, USA) and the isotopes 77Se, 80Se, and 82Se were monitored on the ELAN DRCe, while 77Se, 78 Se, and 82Se were monitored on the ELAN 6000. Data acquisition was: dwell time 200 ms, sweeps per reading 1, and readings per replicate 1. The number of replicates was adjusted to chromatographic runtime. For the post column isotope dilution analysis (IDA), the enriched 7 7 Se standard solution in mobile phase (150 μg L−1) was continuously introduced (at 10 μL min−1) using a syringe pump, Harvard Apparatus 11plus (Holliston, MA, USA) with a 5 mL syringe (Hamilton, Bonaduz, Switzerland). The pump rate was controlled on a regular basis by weighing the eluate from 5 min pumping. The mass per minute was converted to volume per minute (flow rate) correcting for the relevant density. The syringe pump was connected to the eluent flow from the HPLC by a static mixing T-piece. The connections before mixing were made of PEEK with an inner diameter of 0.005 mm and after mixing with an inner diameter of 0.010 mm. The isotope ratios were calculated and corrected for mass bias by daily measuring of the experimental isotope ratios of 80Se/77Se using a linear mathematical model [25], while The theoretical isotope ratios were calculated by using the natural abundance [26]. Measurements of blanks showed that the measurements were not completely free of interference. However, the error introduced by interference was considered small and together with the correction for mass bias satisfactory results could be obtained. This was verified by chromatographic analysis of a certified selenite standard. The isotope dilution equation was applied to each data point and the resulting chromatographic peak was integrated. Converting the mass to concentration based on the injected amount, resulted in an accuracy of 99.5 ± 1.1 % (n = 3), which was considered satisfactory. The mass bias was determined daily using a 100 μg L−1 solution of TMSe. The dead time was corrected automatically by the instrument. The isotope dilution equation [27] was
LMW selenium metabolites in human plasma after selenite treatment Table 1 Chromatographic conditions
2295
LC-ICP-MS Reversed phase (RP) Columna Injection volume (μL) Flow rate (μL min−1) Mobile phase (A) Anion exchange (AEX) Columna Injection volume (μL) Flow rate (μL min−1) Mobile phase (B) Size exclusion (SE) Column* Injection volume (μL) Flow rate (μL min−1) Mobile phase (C) LC-ESI-MS Column Column temperature (°C) Injection volume (μL) Flow rate (μL min−1) Mobile phase
Gemini 5 μm C18 (250 × 2 mm) (Phenomenex) 10 200 200 mM Ammonium acetate, 5 %, MeOH, pH 6.7 IonPac AS11-HC (2 × 250 mm) (Dionex) 10 200 5 mM Ammonium citrate, 2 % MeOH, pH 9 Bio SEC-3 (4.6 × 300 mm) (Agilent) 5 350 20 mM Ammonium acetate-2 % MeOH, pH 7 Kinetex C18, 2.6 μm, 100 Å. 2.1 mm ID × 100 mm (Phenomenex) 50 5 500 A: 0.1 % formic acid, 2 % acetonitrile B: 0.1 % formic acid, 95 % acetonitrile Linear gradient: 0–4 min: 0–80 % B, 4–7 min: 100 % A
a
All analytical columns were protected by the respective guard columns
applied to each point of the chromatogram to obtain the corresponding mass flow chromatogram. Peak areas were calculated by integration of the chromatographic peaks using the OriginPro software (version 9.1, OriginLab Corporation, Northampton, MA, USA). LC-ESI-MS analyses were performed using a ThermoFinnigan Q-Exactive mass-spectrometer (Thermo-Finnigan, San Jose, CA), coupled to a Ultimate 3000 UHPLC-system (Thermo-Finnigan, San Jose, CA). The mass spectrometer was equipped with a heated electrospray ionization source operated in the positive ionization mode. All compounds of interest were detected as the [M + H]+ adducts. Ionization settings were optimized using direct infusion of the relevant standard dissolved in the mobile phase (75 % A and 25 % B) at a flow rate of 0.5 ml/min. For MS2 analyses, the normalized collision energy (NCE) was set to 10. For full scan experiments, the instrument resolution was 70,000; in the t-ms2 (targeted ms-ms) mode, the instrument resolution was 17,500. Chemicals and reagents Chemicals Selenomethionine (SeMet), Se-methylselenocysteine (MeSeCys) and sodium selenite were all of analytical grade
and from Sigma (St. Louis, MO, USA). Trimethylselenonium (TMSe) was synthesized according to Foster and Ganther [28]. Se-methylseleno-N-acetylgalactosamine (SeGal) was synthesized at the Department of Medicinal Chemistry (University of Copenhagen). Ammonium acetate (≥98 %), ammonium citrate dibasic and formic acid were from Sigma. Acetonitrile was from Chemsolute (Th. Geyer, Germany). All chemicals were of analytical grade or better. Methanol (for LC-MS, >99.95 %) was from Th. Geyer (Renningen, Germany). Deionized water (18.2 MΩ cm−1) from a Milli-Q plus unit (Millipore, Bedford, MA, USA) was used throughout. Standards Stock standard solutions of TMSe, SeMet, MeSeCys, SeGal, and selenite were prepared in a concentration of 10 mg L −1 in MilliQ water. The stock solutions were standardized against a 1.001 g L−1 Se pure atomic spectroscopy standard (Perkin Elmer, Norwalk, CT, USA). The standards were stored at 4 °C. Spiking standard solutions in a concentration 500 μg Se L−1 were prepared by dilution of stock solutions in MilliQ water.
2296
The certified enriched elemental 7 7 Se (99.2 %, 1,000 μg mL−1 in 5 % HNO3) was donated from Erik H. Larsen (Technical University of Denmark, Søborg, Denmark).
Samples Human plasma samples frozen in aliquots of 250 μL from the SECAR-study [14] were used in the analyses. A human plasma sample from a healthy volunteer was used for validation purposes.
Procedures
K. Flouda et al.
Precision The precision was calculated as the relative standard deviation (% RSD) for triplicate analysis at the levels of 10, 20 and 40 μg Se L−1 in mobile phase (A) as well as human plasma ultrafiltrates. Limit of detection and limit of quantitation (LOQ) Standards of 5 μg Se L−1 concentration in mobile phase (A) and plasma were analyzed in triplicate. Limit of detection (LOD) and limit of quantitation (LOQ) were calculated as 3.3 and 10 times, respectively, the standard deviation of the peak areas divided by the slope of the standard curves.
Plasma sample preparation GS-Se-SG formation The frozen samples were thawed at room temperature before preparation; 200 μL sample aliquots were transferred to Vivaspin® 500 μL centrifugal filter units (Sartorius AG, Göttingen, Germany) with a semi-permeable membrane of molecular weight cut-off (MWCO) 3000 Da and were centrifuged at 14,000×g for 30 min. The filtrates in the lower compartment were transferred to LC vials and analyzed.
An aqueous standard containing 1.5 μM selenite + 0.5 mM GSH was prepared in 10 % formic acid. Selenite was spiked to a blank plasma sample before MWCO filtration to a final concentration of 0.15 μM. After filtration, GSH was added to a concentration of 0.5 mM and formic acid to a final concentration of 10 %. A pooled plasma sample was added GSH to a final concentration of 0.5 mM and a final formic acid concentration of 10 %.
Filter recovery Ten and 20 μg L−1 samples were prepared in triplicates by spiking untreated human plasma as well as ultrafiltrated plasma and analyzed. Recoveries were calculated as the ratios of RP-LC areas of plasma samples spiked before and after filtration.
RP column recovery Standards of MeSeCys, SeMet, SeGal, and selenite were prepared in mobile phase (A) with a concentration of 50 μg Se L−1. All standards were prepared in duplicates and were measured three times by RP-LC-ICP-MS and three times by flow injection analysis (FIA). The dwell time was changed to 100 ms per isotope. Recoveries were calculated as the ratio of RP-LC areas to FIA areas.
Linearity Linearity was examined in mobile phase (A) and plasma. MeSeCys, SeMet, SeGal, and selenite standards were prepared in five different concentrations: 5, 10, 20, 40 and 50 μg Se L−1 in triplicates. The aqueous standards were analyzed directly, while the spiked plasma standards were analyzed after ultrafiltration.
Results and discussion Method development and validation The intended use of the developed method was LMW analysis of plasma samples from cancer patients treated with large doses of selenite. Although the patients were treated with selenite, this compound was not expected to be present in substantial amounts as ingested selenite is supposed to be taken up by red blood cells where it is reduced by GSH, effluxed, and subsequently transferred to the liver bound to albumin [29, 30]. Hence, a reversed phase method was developed based on separation of the standards SeMet, MeSeCys, and SeGal in 200 mM ammonium acetate, pH 7 in 5 % methanol. A chromatogram of the standards is presented in Fig. 1. The method validation was first performed on these three standards, but as the following analysis of plasma samples showed an unidentified peak eluting at 3.3 min in several samples and the most likely identity of this peak was selenite, selenite was also included in the validation. The method was not validated for the excretory metabolite trimethylselenonium (TMSe) as serum background levels and levels after selenium supplementation with different Se species (including selenite) have been reported to be less than 0.1 ng mL−1 and only detected after pre-concentration [22].
LMW selenium metabolites in human plasma after selenite treatment
2297
concentration range of 5–50 μg Se L−1. The correlation coefficient (R) for all species in aqueous and plasma matrix was > 0.99. The intersections with the y-axis were not significantly different from zero (95 % confidence interval). Residual plots showed that a linear regression model was appropriate for the data.
50000
Intensity (cps)
40000 30000 20000 10000
Precision 0 0
1
2
3
4 5 Time (min)
6
7 −1
Fig. 1 Reversed phase chromatogram of a 50 μg Se L MeSeCys, SeMet and SeGal. Monitored isotope: 80Se
8
9
standard of
Initially, the column recovery was examined for the four species. As shown in Table 2, the column recovery varied between 91.8 and 102.4 %, which was considered satisfactory. Plasma proteins should be removed prior to analysis to obtain the LMW fraction. Initial experiments with protein precipitation by addition of organic solvents were not successful owing to poor precision. Instead, it was chosen to remove proteins by ultrafiltration based on a semipermeable membrane with a molecular weight cut-off (MWCO) value of 3000. In a study of Greening and Simpson [31], several commercially available filters of similar MWCO but with different types of membranes were tested. Only the Sartorius Vivaspin® was effective in recovery and enrichment of LMW components of human plasma. The filter of this type was therefore a natural choice for the plasma sample preparation. To determine the filter recovery, spiked plasma samples were filtrated and the areas compared to plasma ultrafiltrates spiked after filtration at the concentration levels 10 and 20 μg Se L−1, respectively. The filter recoveries are presented in Table 2. The obtained filter recoveries were in the range of 97–107 %, which was considered to be satisfactory.
Precision was determined as the relative standard deviation (RSD) at three concentration levels, 10, 20 and 40 μg Se L−1 and the results are summarized in Table 3. The precision in aqueous standards was in the range of 0.3–4.9 % and between 0.5 and 8.3 % in the plasma matrix. The precision was considered acceptable. Accuracy As no certified standards are available for MeSeCys, SeMet and the selenosugar, the accuracy was determined by analyzing a certified selenite standard by the isotope dilution procedure. The accuracy of the isotope dilution procedure was determined to 99.5 ± 1.1 % (n = 3). The filter recoveries, determined as the ratio of RP-LC areas of plasma samples spiked before and after filtration, were in the range 97–107 % as presented in Table 2. The sample preparation by filtration is thus contributing the most to inaccuracy. LOD and LOQ The LOD and LOQ in mobile phase and plasma are given in Table 4. The obtained LODs were comparable
Table 3 Precision given as the relative standard deviation in mobile phase (MP) and plasma matrix
Linearity Linearity was established for MeSeCys, SeMe, SeGal, and selenite in both mobile phase A and plasma matrix, in a
MeSeCys Table 2 C18 column recoveries calculated as the ratio of RP-LC areas and FIA areas. Filter recoveries calculated as the ratio of RP-LC areas of plasma samples spiked before and after ultrafiltration SeMet Column recovery (%) (n = 6) Filter recovery (%) (n = 3) 10 μg Se L−1 20 μg Se L−1 MeSeCys 97.3 ± 3.5 SeMet 102.4 ± 1.4 SeGal 100.7 ± 0.84 Selenite 91.8 ± 0.5
98.0 ± 3.2 100.6 ± 1.9 96.9 ± 3.1 104.5 ± 0.7
106.6 ± 10.7 101.5 ± 9.5 101.6 ± 9.0 101.7 ± 0.4
SeGal
Selenite
C (μg L−1)
MP (n = 3) RSD%
Plasma (n = 3) RSD%
10 20 40 10 20 40 10 20 40 10 20 40
1.46 3.28 0.93 2.55 2.25 1.12 3.57 4.85 1.78 1.50 0.63 0.32
4.47 5.52 0.48 3.41 5.56 5.49 4.69 5.29 7.24 8.29 7.53 3.98
2298
K. Flouda et al.
Table 4 LOD and LOQ, calculated based on the deviation of the peak areas at the 5 μg L−1 level in mobile phase (MP) and plasma matrix
MP (n = 3)
Plasma (n = 3)
LOD (μg L−1)
LOQ (μg L−1)
LOQ (μg L−1)
MeSeCys
0.25
0.77
0.49
1.49
SeMet
0.28
0.84
2.11
6.40
SeGal Selenite
0.35 0.30
1.07 0.92
1.83 0.44
5.55 1.33
Generally, three peaks could be observed in most of the analyzed samples. Peak 3 eluting at 6.7 min was tentatively identified by retention time matching and spiking with SeGal. This peak was observed in most samples after selenite treatment. This is in line with previous studies as SeGal is the major metabolite in mammals [33, 34]. Recently, Kokarnig et al. have reported the presence of traces of selenite, TMSe and MeSeCys in background serum based on retention time matching [23] and in addition the presence of SeGal and SeMet after SeMet supplementation. On the other hand, they reported that total serum Se levels did not increase after selenite supplementation [23]. Peaks 1 and 2 did not match with the retention times of MeSeCys or SeMet. As the unknown peak 2 was only present in very small amounts even when the total Se content was high, identification of this compound was not further pursued. Peak 1 co-eluted with selenite, but this was not considered sufficient evidence for the presence of selenite as it eluted close to the void volume. Analysis of a high Se concentration sample on cation exchange chromatography showed elution of the majority of the sample in the front, thus the unknown peak 1 was not TMSe or any other cationic species; but in a size exclusion
to previously reported values of LMW Se speciation in biological samples [32].
Quantitation of LMW Se species The quantification of Se concentration in the plasma samples by RP-LC-ICP-MS was accomplished by IDA with the ICP-MS detection as described in the BExperimental^ section. The mass bias factor (F) was calculated and the original ICP-MS intensities were imported into an Excel spreadsheet. Once the mass bias factor was estimated the experimental isotope ratios 80Se/77Se were corrected. The dead time was corrected automatically by the instrument. After the corrections, the isotope dilution equation [25] was applied to the whole chromatogram to obtain the mass flow data. Finally, the data were imported to OriginPro software for integration. About 20 samples were injected in duplicate. Figure 2a shows typical mass flow chromatograms from human plasma analyses in duplicate showing identical chromatograms, while Fig. 2b shows the mass flow chromatograms of the standards.
a)
b) 0.0014
MeSeCys
0.0030
3
0.0012
SeMet
0.0025 Mass flow (µg/min
Mass flow (µg/min)
LOD (μg L−1)
0.001 0.0008
1 2
0.0006 0.0004
SeGal
0.0020 0.0015 0.0010 0.0005
0.0002
0.0000
0 0
1
2
3
4
5
6
7
8
Time (min)
Fig. 2 a Examples of mass flow chromatograms from plasma samples. Lower chromatograms: sample before selenite treatment; upper chromatograms (off-set by 0.0002 μg min−1): sample after treatment. The total LMW concentration of the samples were 0.9 and 35.5 μg L−1,
9
0
1
2
3
4
5
6
7
8
9
Time (min)
respectively. Each sample was analyzed in duplicate, blue and red trace, respectively. b Mass flow chromatograms of a 50 μg Se L−1 standard of MeSCys, SeMet, and SeGal
LMW selenium metabolites in human plasma after selenite treatment
system as well as an anion exchange system, the retention time of peak 1 matched selenite retention, which indicated the presence of selenite (Fig. 3). Traces of selenite in serum based on retention time matching by LC-ICP-MS analysis has previously been reported, but at levels hardly distinguished from background [35, 36]. The possibility that selenite is a breakdown product of selenoproteins during storage of samples has also been suggested [37]. However, the presence of selenite in plasma has never been confirmed by molecular MS. The samples in the present study were from patients treated with selenite in doses between 1 and 15.3 mg Se/m2 by infusion corresponding to between 0.5 and 7.7 mg Se, if the body surface is set to 2 m2. These doses are much larger than the doses from supplements, which normally are in the 50–100 μg range. The LMW selenium compounds only constituted a minor fraction of the plasma Se—between 1 and 28 % with a mean of 8 %. Thus, the major Se amount must be accumulated in the proteins either as incorporated in the protein or attached to proteins. Another possibility is elemental selenium. A general pattern in patient samples was increased selenite signals
82
Se Intensity (counts)
Size exclusion chromatography 8000
plasma
7000
plasma spiked
6000 5000 4000
2299
immediately after selenite application followed by a decline over time. These data will be subject of a future study on the pharmacokinetics of selenite in plasma.
Identification with LC-ESI-MS To verify the identity of selenite in the samples, an ultrafiltrate of a pooled plasma sample was exposed to LC-ESIMS in negative ionization mode. To identify selenite in a biological sample, selenite has to be separated from the matrix. In the RP chromatography mode, selenite elutes together with a large amount of other small ions resulting in large ion suppression in the MS. Anion exchange chromatography retained selenite, but the high ionic strength of the citrate eluent caused severe ion suppression. Instead, an indirect approach was taken by formation of the GSH adduct. Selenite is known to form a trisulfide by reaction with GSH following the reaction scheme: 4GSH + SeO32− → GSSG + GS-Se-SG + 2OH− + H2O [38, 39]. Thus, treatment of plasma ultrafiltrate with GSH should result in a compound that could be retained on an RP column and following formation of an ion with the exact mass to charge ratio of 693.076 in ESI-MS positive ionization mode. The extracted ion chromatogram of a pooled plasma sample after addition of GSH is shown in Fig. 4. It appears from the figure, that no signals were obtained in blank plasma or the pooled plasma sample prior to GSH addition, while a large signal corresponding to the GS-SeSG adduct was obtained after GSH addition. The retention
3000 2000 1000 0 0
1
2
3
4
5
6
7
8
9
10
11
12
13
6000000
Time (min)
plasma plasma spiked
12000 10000 8000 6000
Selenite standard + GSH
4000000
Blank plasma + selenite + GSH Blank plasma
3000000
Pooled plasma + GSH
2000000 Pooled plasma
4000
82
Se Intensity (counts)
Anion exchange chromatography
Total Ion Current
5000000
14000
1000000
2000 0 0
1
2
3
4
5
6
7
8
9
10
11
12
Time (min)
Fig. 3 Retention time matching of selenite by spiking plasma samples with standard in size exclusion chromatography and anion exchange chromatography. Chromatograms are offset by 1000 and 2000 counts, respectively
0 0
1 Time / minutes
2
Fig. 4 Extracted ion chromatograms at m/z 693.076 from RP chromatography of selenite standards and samples after addition of GSH. Chromatograms are offset
2300
K. Flouda et al.
Fig. 5 Extracted spectrum from the TIC of the plasma sample peak from RP chromatography of the sample (retention time 1.47– 1.51) and the calculated spectrum
time of this signal matched the retention time of a selenite standard mixed with GSH and a plasma sample spiked with selenite + GSH. The spectrum obtained from the plasma sample and the theoretical spectrum with the characteristic Se isotope pattern of the five most abundant isotopes (76Se 9.4 %, 77Se 7.6 %, 78Se 23.8 %, 80Se 49.6 %, 82 Se 8.7) shown in Fig. 5 are almost identical, which leads to the conclusion that selenite was identified in the sample. The derivatization method was not used for quantitative purposes, thus quantitative derivatization was not a demand. The finding of selenite is not surprising as the patients were treated with large amounts of selenite, however neither increased selenite levels in plasma after selenite treatment, nor evidence based on ESI-MS has not been reported before. A reason for this finding could be that the patients were very ill and simultaneously treated with large amounts of drugs, which could inhibit or exhaust normal metabolic pathways. The identity of the selenosugar was verified by LC-ESI-MS by targeted MS-MS product ion scans on m/z 302, 300, 298,
297, and 296. For clarity, only three of these are shown in Fig. 6, where scans from the RP chromatographic peaks are compared for a selenosugar standard and a plasma sample. It appears from the figure that the product ions 204.087, 186.076, 168.066, 144.066, 138.055, and 126.055 were all produced in the standard as well as the plasma sample. The intensities of the peaks varied, probably owing to the very different matrices. However, the relative contributions from the isotopes were in agreement with relative abundances of the isotopes. Thus, the identity of the selenosugar was considered verified. These ions all represent losses of selenium moieties from the sugar molecule. The formation of these product ions by selective reaction monitoring (SRM) have previously been reported for a selenosugar standard [40], some of them in biological samples as rat liver and kidney cytosol after purification and pre-concentration (204, 186, 144, and 138) [24], human urine (204, 186, 144, 138) [22] and serum (204, 186, 144) [23], but not all of them in one study. The present scans were performed on an Orbitrap instrument, which has the sensitivity to make this possible.
Fig. 6 Product ion scans of a selenosugar standard and b a plasma sample collected from the respective peaks from RP chromatography. The scans are overlaid. Black star Background peak
LMW selenium metabolites in human plasma after selenite treatment
Conclusion A validated analytical method for quantitation of LMW selenium compounds in human plasma using RP-LC-ICP-MS was developed. The method was based on ultrafiltration sample preparation and quantitation by post-column IDA. The validation parameters for recoveries, linearity, precision, LOD, and LOQ, were considered satisfactory for analysis of plasma samples. The method was applied for analysis of plasma samples from patients receiving selenite treatment in a clinical trial. Three selenium species were observed in varying amounts in the plasma samples. Two of these were tentatively identified as SeGal and selenite by retention time matching with standards, while the third minor metabolite was not identified. The identity of selenite was verified by molecular MS as the glutathione adduct GS-Se-SG, while the identity of SeGal was verified by product ion scans on the masses of the sugar containing the five most abundant isotopes of selenium. Acknowledgments The authors wish to thank Erik Huusfeldt Larsen for the donation of the enriched 77Se solution and laboratory technician Camilla Jensen for skillful technical assistance. The SECAR trial was supported by grants from Cancerfonden and Jochnick Foundation, Sweden.
2301 10. 11. 12. 13. 14.
15. 16. 17. 18. 19. 20. 21. 22. 23. 24.
Compliance with ethical standards The SECAR-trial [14] was approved by the Swedish Medical Products Agency and registered in the EU Clinical Trial Register (Eudra CT Number: 2006-004076-13) and by the Ethical Committee of Stockholm (2006/429-31/3). Informed consent was given by the healthy volunteer donating the plasma sample for validation.
25. 26. 27.
Conflict of interest The authors declare that they have no conflict of interest.
References 1.
Kryukov G, Kryukov V, Gladyshev V. J Biol Chem. 1999;274: 33888–97. 2. Ganther HE. Carcinogenesis. 1999;20:1657–66. 3. Ganther HE. J Am Coll Tocicol. 1986;5:1–5. 4. Weekley CM, Harris HH. Chem Soc Rev. 2013;42:8870–94. 5. Suzuki KT. J Health Sci. 2005;51:107–14. 6. Gammelgaard B, Jackson MI, Gabel-Jensen C. Anal Bioanal Chem. 2011;399:1743–63. 7. Rayman MP. Proc Nutrition Soc. 2005;64:527–42. 8. Björnstedt M, Fernandes A. EPMA J. 2010;1:389–95. 9. Nilsonne G, Sun X, Nyström C, Rundlöf AK, Potamitou Fernandes A, Björnstedt M, et al. Free Rad Biol Med. 2006;41:874–85.
28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40.
Menter DG, Sabichi AL, Lippman SM. Cancer Epidemiol Biomarkers Prev. 2000;9:1171–82. Husbeck B, Nonn L, Peehl DM, Knox SJ. Prostate. 2006;66:218– 25. Hanahan D, Weinberg R. Cell. 2011;144:646–74. Brodin O, Eksborg S, Wallenberg M, Asker-Hagelberg C, Larsen EH, Molkert D, et al. Nutrient. 2015;7:4978–94. Randomised trial to identify the safest and most effective selenium compound for cancer patients. Phase Ib randomised double-blind dose-escalation trial to identify the safest most effective selenium compound for use in patients with either chronic lymphocytic leukaemia or metastatic prostate cancer . https://www.anzctr.org.au/ Trial/Registration/TrialReview.aspx?ACTRN=12613000118707. 2015. Lippman S, Klein E, Goodman P, Lucia M, Thompson I, Ford L, et al. JAMA. 2009;301:39–51. Clark LC, Combs GF, Turnbull BW, Slate EH, Chalker DK, Chow J, et al. JAMA. 1996;276:1957–63. Hurst R, Armah CN, Dainty JR, Hart DJ, Teucher B, Goldson AJ, et al. Am J Clin Nutr. 2010;91:923–31. Ogra Y, Anan Y. J Anal At Spectr. 2009;24:1477–88. Dumont E, Vanhaecke F, Cornelis R. Anal Bioanal Chem. 2006;385:1304–23. Pedrero Z, Madrid Y. Anal Chim Acta. 2009;634:135–52. Gammelgaard B, Gabel-Jensen C, Stürup S, Hansen HR. Anal Bioanal Chem. 2008;390:1691–706. Lu Y, Rumpler A, Francesconi KA, Pergantis SA. Anal Chim Acta. 2012;731:49–59. Kokarnig S, Tsirigotaki A, Wiesenhofer T, Lackner V, Francesconi KA, Pergantis SA, et al. J Trace Elem Med Biol. 2015;29:83–90. Gonzalez-Iglesias H, Fernandez-Sanchez ML, Lu Y, Fernandez Menendez S, Pergantis SA, Sanz-Medel A. J Anal At Spectr. 2015;30:267–76. Rodriguez-Gonzalez P, Marchante-Gayon JM, Alonso JIG, SanzMedel A. Spectrochim Acta Atom Spectros. 2005;60:151–207. Berglund M, Frech W, Baxter DC. Spectrochim Acta. 1991;46B: 1767–77. Sariego Muniz C, Marchante Gayon JM, Garcia Alonso JI, SanzMedel A. J Anal At Spectr. 2001;16:587–92. Foster SJ, Ganther HE. Anal Biochem. 1984;137:205–9. Suzuki T, Shiobara Y, Itoh M, Ohmichi M. Analyst. 1998;123:63– 7. Shiobara Y, Ogra Y, Suzuki KT. Analyst. 1999;124:1237–41. Greening DW, Simpson RJ. J Proteomics. 2010;73:637–48. Gammelgaard B, Bendahl L, Jacobsen NW, Stürup S. J Anal At Spectr. 2005;9:889–93. Ogra Y, Ishiwata K, Takayama H, Aimi N, Suzuki KT. J Chromatogr B. 2002;767:301–12. Gammelgaard B, Bendahl L. J Anal At Spectr. 2004;19:135–42. Jitaru P, Goenaga-Infante H, Vaslin-Reimann S, Fisicaro P. Anal Chim Acta. 2010;657:100–7. Palacios O, Ruiz Encinar J, Schaumlöffel D, Lobinski R. Anal Bioanal Chem. 2006;384:1276–83. Palacios O, Lobinski R. Talanta. 2007;71:1813–6. Ganther HE. Biochemistry. 1968;7:2898–905. Arteel GE, Sies H. Environ Toxicol Pharmacol. 2001;10:153–8. Dumont E, Ogra Y, Suzuki KT, Vanhaecke F, Cornelis R. Anal Chim Acta. 2006;555:25–33.
