Parabens in urine, serum and seminal plasma from healthy Danish men determined by liquid chromatography tandem mass spectrometry (LC MS/MS)

Transcription

1 Journal of Exposure Science and Environmental Epidemiology (2011) 21, r 2011 Nature America, Inc. All rights reserved /11 Parabens in urine, serum and seminal plasma from healthy Danish men determined by liquid chromatography tandem mass spectrometry (LC MS/MS) HANNE FREDERIKSEN, NIELS JØRGENSEN AND ANNA-MARIA ANDERSSON Department of Growth and Reproduction, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark Parabens are used as anti-microbial preservatives in a range of consumer products, especially in cosmetics. In vitro and animal studies have shown weak estrogenic and other endocrine disrupting effects of parabens, including reduced testosterone levels in exposed male rats. The knowledge of paraben exposure, distribution and excretion in humans is limited. In this study we determined the concentration of five parabens; methyl-, ethyl-, n-propyl-, n-butyl- and benzylparaben in urine, serum and seminal plasma samples from 60 healthy Danish men. To conduct the study a sensitive and specific method using LC MS/MS for simultaneous determination of the five parabens was developed for all three different matrices. Highest concentrations of the parabens were found in urine, wherein methyl-, ethyl-, n-propyl- and n-butyl parabens were measurable in 98%, 80%, 98% and 83% of the men, respectively. Benzyl paraben was only measurable in urine from 7% of the men. Methyl- and n-propyl parabens were also measurable in the majority of serum and seminal plasma samples, whereas the other parabens could only be detected in some of the samples. In all the three matrices significant correlations between the parabens were seen. Furthermore, urinary paraben concentrations correlate to the paraben concentrations in both serum and seminal plasma. Journal of Exposure Science and Environmental Epidemiology (2011) 21, ; doi:138/jes ; published online 10 March 2010 Keywords: endocrine disruptor, paraben, human urine, human serum, seminal plasma, LC MS/MS. Introduction Parabens are a group of industrial chemicals, which during the last decade have been pointed out as possible endocrine disruptors. Some parabens have been shown to have weak estrogenic effects in vitro and in animal studies (Okubo et al., 2001; Byford et al., 2002; Darbre et al., 2002, 2003, 2004; Lemini et al., 2003; Pugazhendhi et al., 2005). Furthermore, studies have shown effects on the male reproductive system, resulting in the reduction of testosterone levels and mature sperm counts, in rats and mice after in utero exposure to some parabens (Oishi, 2001, 2002, 2004; Kang et al., 2002), whereas other studies could not confirm these effects (Hoberman et al., 2008; Taxvig et al., 2008). Recently it has been discussed whether parabens could have an adverse effect on the testis mitochondria function, followed by a decreased reproductive potential (Tavares et al., 2009). In our daily life we are highly exposed to parabens owing to the wide use of parabens as anti-microbial preservatives in 1. Address all correspondence to: Dr. Hanne Frederiksen, Department of Growth and Reproduction, Copenhagen University Hospital, Blegdamsvej 9, Copenhagen DK-2100, Denmark. Tel.: þ Fax: þ hanne.01.frederiksen@rh.regionh.dk Received 6 October 2009; accepted 24 January 2010; published online 10 March 2010 cosmetics such as skin-care and soap products, in pharmaceuticals and in some processed food products. The benefits are many; in general parabens are considered safe with low or no toxic effects, they have no taste or odor and are low-cost products (Soni et al., 2005; CIR Expert Panel, 2009). The most common used parabens are methylparaben (MeP), ethylparaben (EtP), n-propylparaben (n-prp), iso-propylparaben (i-prp), n-butylparaben (n-bup), iso-butylparaben (i-bup) and benzylparaben (BzP), and to increase the activity against microbial contamination in the host product, the parabens are often used in combination (CIR Expert Panel, 2009). Humans are exposed to parabens via absorption through skin (Lobemeier et al., 1996; Akomeah et al., 2004; Janjua et al., 2007, 2008) and via the gastrointestinal tract after oral intake. Parabens are rapidly metabolized into mainly the harmless non-specific p-hydroxybenzoic acid and its respective glucuronic and sulfuric acid conjugates. Several in vitro, animal and a few human studies have shown that after oral intake almost the entire dose of parabens were excreted through urine as p-hydroxybenzoic acid and its conjugates (Soni et al., 2005; CIR Expert Panel, 2009). In contrast, following skin application more of the test doses were excreted in urine as the parent parabens in their free (unconjugated) or conjugated form (glucuronidated and sulphated) (Soni et al., 2005; Janjua et al., 2008; CIR Expert Panel, 2009). In a few human studies urinary, serum and breast tissue levels of the parent parabens have been used as

2 Determination of parabens in human samples Frederiksen et al. biomarkers of exposure to the different parabens (Darbre et al., 2004; Ye et al., 2006a, b, 2008; Janjua et al., 2007, 2008). Despite the rapid metabolization the levels of parent parabens, especially MeP and n-prp, distributed to the different human matrices are not insignificant. Also rat studies have shown that parabens were distributed in their parent form to several different fluids and tissues before excretion (Soni et al., 2005) and even seem to accumulate in amniotic fluid (Frederiksen et al., 2008). Despite the discussion and concern in recent years about the possible endocrine disrupting effects of parabens and development of analytical methods for quantification of parabens in human urine and serum (Ye et al., 2006b, 2008), the knowledge about human exposure, distribution and excretion of parabens are limited. The aim of this study has been to investigate the distribution and excretion of five commonly used parabens (MeP, EtP, n-prp, n-bup and BzP) in human urine, serum and seminal plasma. The study group consisted of 60 young and healthy Danish men, and an analytical method using LC MS/MS was developed for the analyses of the parabens in all three matrices. To our knowledge, the data presented here is the first to report parabens measured in human urine, serum and seminal plasma samples collected from the same individuals. Materials and methods Study Population Urine, blood and semen samples were obtained from 60 young and healthy Danish men participating in a study on male reproductive health. Mean anthropometric data for the population group were: age, 19.7 years (range: years); height, cm (range: cm); weight, 77.0 kg (range: kg); and body mass index (BMI), 23.1 kg/m 2 (range: kg/m 2 ). All participants gave their written consent after having received written and oral information about the study. The study was approved by the ethical committee for the Copenhagen municipality (ref. nos.: KF /96 and KF /98 with amendment on 19 January 2006). All samples were collected in clinic, December 2006, in connection with a physical examination. The three different samples from each of the men were obtained within maximum one hour. Sample Collection and Storage From each participant a blood sample was drawn from the cubital vein and serum was obtained by centrifugation for 10 min at 2000g and subsequently aliquoted for hormone and chemical analyses. Semen samples were obtained by masturbation and preserved at 371C until liquefaction had occurred. After aliquots were taken out for routine clinical assessments of semen quality parameters, the remaining sample was immediately centrifuged for 30 min at 2000g (Jorgensen et al., 2002) and the supernatant consisting of cell-free seminal plasma were used for chemical analyses. For the analyses of parabens aliquots of urine, serum and seminal plasma were transferred into 4 ml glass screw-cap vials. To avoid possible enzyme activity in serum and seminal plasma samples after collection, these samples were added 1.2 M phosphoric acid, 100 ml and 200 ml per 1.0 ml serum, and seminal plasma, respectively, before storing at 201C. Spot urine samples were collected in polyethylene cups and aliquots of each urine sample was transferred to 20 ml glass scintillation vials with tops packed with aluminum foil. All urine samples were stored at 201C until analysis. From surplus human serum and seminal plasma collected over time from different inhouse studies, and human urine sampled from colleagues, who consciously avoided exposure to parabens before the sampling, one pool of each matrices were made and used for the calibration of standards and as control material in the validation of the present method. Reagents and Standards MeP, EtP, n-prp, n-bup and BzP were purchased from Fluka (distributed by Sigma Aldrich, Brndby, Denmark) and deuterium-labeled methylparaben (D 4 -MeP) was obtained from CDN Isotopes (Quebec, Canada). Phosphoric acid, glacial acetic acid, formic acid and sodium dihydrogen phosphate dihydrate were obtained from Merck (Darmstadt, Germany). Acetonitrile, methanol, ammonium acetate and sodium sulfate were all obtained from J.T. Baker. Both Merck and T.J. Baker chemicals were distributed by Bie & Berntsen A/S (Rdovre, Denmark). Ethyl acetate was obtained from BDH Laboratories Supplies (Poole, England and distributed by VWR International, Rdovre, Denmark). 4-Methylumbelliferone, 4-methylumbelliferyl-b-D-glucuronide and 4-methylumbelliferyl sulfate were obtained from Sigma Aldrich (Brndby, Denmark). b-glucuronidase (Escherichia coli K12) and aryl-sulfatase were obtained from Roche Diagnostics (Mannheim, Germany). Milli-Q water was cleaned in a Millipore System (Synthesis A10). Solidphase extraction (SPE) cartridges (Strata XL; 200 mg, 3 ml) were obtained from Phenomenex (Allerd, Denmark). All chemicals were of analytical or HPLC grade and all chemicals, solutions and laboratory wares were checked for contamination with parabens before use. Preparation of Standards Stock solutions of each of the native paraben standards and the labeled standard D 4 -MeP were prepared in volumetric flasks by dilution with 50% acetonitrile to a concentration of 100 mg/ml. Mixtures of all native parabens were prepared from the stock solutions and diluted with 10% acetonitrile to a range of working solutions ( mg/ml). D 4 -MeP was diluted with 10% acetonitrile to a final working solutions of 200 ng/ml. Journal of Exposure Science and Environmental Epidemiology (2011) 21(3) 263

3 Frederiksen et al. Determination of parabens in human samples Enzymatic Hydrolysis and SPE of Samples After thawing, samples were mixed and aliquots of 500 ml were added 1.5 ml 2 M ammonium acetate buffer, ph 6.5 each. 20 ml of the 200 ng/ml D 4 -MeP solution were then added to all samples as internal standard. Subsequently, 20 ml 100 ng/ml 4-methylumbelliferyl b-d-glucuronide and 20 ml 100 ng/ml 4-methylumbelliferyl sulfate were added to all samples for the control of enzyme reaction (Ye et al., 2006b). Immediately before the incubation 15 ml of a freshly prepared enzyme mixture containing 5 ml b-glucuronidase and 10 ml aryl-sulfatase was added to all the samples, mixed and incubated for 90 min at 371C in a shaking water bath. The reaction was terminated by adding 1 ml of 3.6 M phosphoric acid. All samples were purified by the automated SPE (Aspec XL, Gilson, Middleton, USA). The SPE cartridges were preconditioned with 1 ml acetonitrile, 1 ml milli-q water and 1 ml 0.15 M ammonium acetate buffer, ph 2 3. Subsequently samples were applied to the columns and the tubes were rinsed with 3 ml 0.15 M ammonium acetate also applied tothecolumns.beforeelutionthecartridgeswerewashed with 2 ml of 1% phosphoric acid followed by 2 ml of 10% acetonitrile. The compounds were eluted with 1.5 ml pure acetonitrile followed by 1.5 ml ethyl acetate. The eluates were evaporated to dryness (451C, 5 15 p.s.i. N 2 ) and the residues of total metabolites were resuspended in 200 ml of resuspension solution (1% formic acid in 60% acetonitrile/water). Preparation of Calibration and Quality Control Samples All calibration and quality control material were prepared in the three pools of respectively urine, serum and seminal plasma and treated exactly as the real samples as described above. The pools were dispensed into 500 ml aliquots. Pool aliquots for blank (unspiked) samples and calibration samples were deconjugated and purified by SPE and subsequently the calibration samples were spiked with the native paraben solutions to a final concentration of 0.5, 2.0, 10, 50, 100 and 500 ng/ml. Before deconjugation and SPE purification, pool aliquots for quality control were added the native paraben solutions to final concentrations of 5, 10 and 50 ng/ml. An extra set of quality control samples was prepared to determine the SPE recovery; these samples were not added D 4 -MeP to final concentration of 10 ng/ml until after deconjugation and SPE. All calibration and quality control material were stored at 201C until use. As solvent blank samples, 2000 ml 2 M ammonium acetate buffer was used and treated as described above. Instrumental Analysis For all detection and quantification of analytes, a high performance liquid chromatograph (LC) (Surveyor, Thermo- Finnigan, San Jose, CA, USA) coupled with a tandem mass spectrometer (MS/MS) (Finnigan TSQ Quantum Ultra triple quadrupole mass spectrometer, Thermo Electron Table 1. Solvent gradient: A, 0.1% acetic acid in water; B, 0.1% acetic acid in acetonitrile. Time (min) Solvent A (%) Solvent B (%) Corporation, San Jose, CA, USA) was used in combination with the X-calibur software program. The LC MS/MS system was equipped with an electrospray ion source (ESI) and was running in negative mode. The injection volume was 20 ml on a Synergi 4U fusion-rp 80A column ( mm 4 mm) (Phenomenex, Anschaffenburg, Germany) equipped with a filter (frit ss blk, A 120,5mm pore size from Upchurch Scientific distributed by Mikrolab, Aarhus, Denmark) in front of the column. The flow rate was 300 ml/min and the column temperature was 251C. Solvents were A: 0.1% acetic acid in water and B: 0.1% acetic acid in acetonitrile. Table 1 shows the solvent programming and Figure 1 shows MS transitions. The optimized MS/MS interphase settings used were: spray voltage, 3000 V; sheath gas (N 2 ) pressure, 55 psi; auxiliary gas (N 2 ) pressure, 10 psi; capillary temperature, 3501C; collision gas (Ar) pressure, 1.0 mtorr; tube lens offset, 80 V; and for the formation of the product ions the collision energy was set to 20 V. Operation and Quality Control Procedure For validation of the method accuracy the recoveries of the SPE and recoveries of native standards spiked in pools of urine, serum and seminal plasma were determined. Analytical batch for urine, serum or seminal plasma, respectively, included five repeated sequences of a solvent blank sample, a blank pool sample and the quality control pools spiked with 5, 10 and 50 ng/ml of the native paraben solution. In the same five sequences pool samples spiked with 10 ng/ml of the native parabens and to which the D 4 -MeP solution (20 ml, 200 ng/ml) was added after SPE were also included to enable the determination of SPE recovery. The method accuracy was expressed as the percentage of expected levels measured in the quality control samples spiked before deconjugation and purification (spike recovery). Furthermore the SPE recovery was determined by calculating the percentage difference in concentration between quality control samples added the labeled standard D 4 -MeP before deconjugation and purification and quality control samples to which D 4 -MeP was added after deconjugation and purification. For validation of the inter- and intra-day precision of the method, each analytical batch included two solvent blank samples, and for each of the three matrices six pool samples; two blank pools samples, two pools spiked with 5 ng/ml and two pools spiked with 50 ng/ml of the native parabens. These 264 Journal of Exposure Science and Environmental Epidemiology (2011) 21(3)

4 Determination of parabens in human samples Frederiksen et al. Relative Abundance MeP m/z EtP m/z n-prp m/z n-bup m/z BzP m/z typical daily sample batch contained from samples including calibration samples. Data Analysis and Statistics The X-calibur software was used for automatic selection and integration of ions of analytes and internal standards in all chromatograms. Peaks were corrected manually if necessary. Only peaks with a signal-to-noise ratio Z3 were integrated. The ratios between of the peak areas in samples (calibration, control and unknowns) and the corresponding internal standards were calculated and the average calibration curve weighted by the reciprocal of the standard amount (1/x) was constructed in the X-calibur program. For the quantification of control and unknown samples the y-intercept of the linear calibration curve were set to 0 in order to adjust for possible background levels present in the serum pool used to prepare the calibration samples. Thus the concentration in unknown samples were calculated as the area ratios divided by the slope of calibration curve. All validation data were expressed as means±sd or as the relative SD (RSDB(SD/mean) 100). The limit of detection (LOD) was calculated as 3S 0,whereS 0 was the value of the SD as the concentration approaches zero. S 0 was determined as the point where the best-fit line of the SD of five repeated sets of the lowest four standards versus the known standard concentration cross the y-intercept (Blount et al., 2000). To describe the levels of parabens in the urine, serum and seminal plasma samples, 25th, median (50th), 75th percentiles and the maximum concentration were computed. To analyze for correlations between the different parabens in each matrices and between the single parabens in different matrices Spearman correlations were computed. In some of the samples the paraben concentration was below LOD and in others the concentration was not detectable. For correlation analyses all data were included and not detectable data were all set to, which was half of the lowest LOD. P-valueso0.05 were considered as statistically significant. All statistical analyses were carried out using SPSS (SPSS for Windows, version 17.0, SPSS Chicago, IL, USA) Time (min) Figure 1. LC-MS/MS extracted ion chromatograms of product ions in negative mode [M-H] of parabens spiked in a sample of pooled urine. Spike level for all analytes was 50 ng/mlb1.0 ng on column. quality control samples were evenly distributed among the unknown urine, serum and seminal plasma samples. In all, daily urine, serum or seminal plasma sample batches, series of calibration samples were analyzed before and after the unknown samples and the quality control samples. In addition, a solution of 4-methylumbelliferone (20 ng/ml) was analyzed once or twice in every batch. A Results Method Characteristics and Validation Extracted ion chromatograms of a urine pool sample spiked with 50 ng/ml of the native paraben standards, their retention times and precursor/product ions in negative mode [M H] are shown in Figure 1. The precursor/product ion pair for D 4 -MeP (not shown) was m/z and D 4 -MeP had retention time analog to MeP. The concentrations of all parabens were calculated in relation to D 4 -MeP, as this was the only commercially available internal standard of the parabens at the time when analyses were carried out. Owing to matrix effects in especially serum and seminal plasma, all calibration samples Journal of Exposure Science and Environmental Epidemiology (2011) 21(3) 265

5 Frederiksen et al. Determination of parabens in human samples MeP MeP MeP EtP EtP EtP Relative Abundance 9.52 n-prp n-bup 9.52 n-prp n-bup 9.49 n-prp n-bup n.d BzP BzP BzP Time (min) Figure 2. LC-MS/MS extracted ion chromatograms of paraben product ions in a urine sample (a), a serum sample (b) and a seminal plasma sample (c), all samples from one young Danish man. All samples were obtained within one hour in December were prepared in the same matrix as the unknown samples, urine, serum and seminal plasma, respectively. The endogenous levels of the analytes in the pooled matrices were for most of the analytes o1 ng/ml except the concentration of MeP in the serum pool which was 11.0 ng/ml. However, the calibration curves of standards spiked in urine, serum or seminal plasma all gave good linearity (correlation coefficients (r 2 )40.99) in the full measurement range ( ng/ml). The LOD was determined for the parabens in all three matrices (Table 2). Under the condition that 500 mlofsample were purified and resolved in 200 ml after purification (resulting in a 2.5-fold concentration), LODs were r0.41 ng/ml for all the parabens in all matrices. Very good SPE and spiked recoveries were obtained for nearly all parabens in all three matrices (Table 2). The method accuracy ranged from 85% to 110% with SD below 10 for most of the parabens. Most important was that acceptable recoveries could be obtained in the lowest spike level of the parabens (5 ng/ml) in serum and seminal plasma, because very low concentrations of the parabens were expected in real-life serum and seminal plasma samples. In the lowest spike level only the spike recovery of EtP (79.7%) and n-prp (82.7%) in serum and EtP (78.0%) in seminal plasma were lower than 85%. The precision of the method was determined by repeated measurements of 24 urine, 18 serum and 14 seminal plasma quality control samples spiked in a low and a high level of the analytes. The spiked quality control samples were measured over a period of 2 months in respectively 12, 9 and 7 different batches (Table 3). The RSD expressed in percentage reflecting the intra- and inter-day variability, ranged from 2.8% to 29.2%. In general, the highest RSD values were found among the parabens in the low spike level, which could be caused by the use of D 4 -MeP as labeled standard for all the parabens. However, the RSD values were below 10% for most of the analytes in all three matrices, which demonstrated a good precision of the method over time. Parabens Levels in Human Samples Urine, serum and seminal plasma from 60 young Danish men were analyzed and the total levels of parabens (sum of unconjugated, deglucuronidated and desulfated parabens) were determined (Figure 2). All parabens except BzP were measurable in most of the urine samples, with MeP excreted in urine in highest amounts (median ¼ 17.7 ng/ml) (Table 4). The urinary excretion of n-prp, EtP and n-bup were about 5, 10 and 100-fold lower, respectively. In a few of the urine samples the concentration of MeP was higher than the linear 266 Journal of Exposure Science and Environmental Epidemiology (2011) 21(3)

6 Determination of parabens in human samples Frederiksen et al. Table 2. Method accuracy for urine serum and seminal plasma: limits of detection (LOD), solid-phase extraction (SPE) recoveries (n ¼ 5) and spiked standard concentration recoveries (n ¼ 5). Matrix Paraben LOD (ng/ml) SPE recovery 10 ng/ml (%±SD) Spiked recovery 5 ng/ml (%±SD) 10 ng/ml (%±SD) 50 ng/ml (%±SD) Urine MeP ± ± ± ±2.5 EtP ± ± ± ±3.0 n-prp ± ± ± ±6.4 n-bup ± ± ± ±3.2 BzP ± ± ± ±6.7 Serum MeP ± ± ± ±5.6 EtP ± ± ± ±5.4 n-prp ± ± ± ±5.5 n-bup ± ± ± ±5.4 BzP ± ± ± ±6.8 Seminal plasma MeP ± ± ± ±2.9 EtP ± ± ± ±3.9 n-prp ± ± ± ±5.1 n-bup ± ± ± ±6.8 BzP ± ± ± ±11.5 Abbreviation: n, number of samples. Table 3. Inter- and intra-day precision of spiked quality control samples. Matrix Paraben Low-quality control High-quality control Mean (ng/ml) RSD (%) Mean (ng/ml) RSD (%) Urine (n ¼ 24) MeP EtP n-prp n-bup BzP Serum (n ¼ 18) MeP EtP n-prp n-bup BzP Seminal plasma (n ¼ 14) MeP EtP n-prp n-bup BzP Abbreviations: n, number of samples; RSD, relative standard deviation. Inter-day precision was determined over a period of 2 months. range of the calibration curves. These samples were reanalyzed after two or four-fold dilution to ensure that the concentration was correctly determined. In serum and seminal plasma the levels of all the parabens were much lower than in urine (Table 4). In serum MeP and n-prp were detected in most of the samples, whereas the levels of the other parabens only were above LOD in few of the samples. Apart from BzP, the other four parabens were detected in several of the seminal plasma samples but with very low median levels below ng/ml. Also in serum and seminal plasma was MeP observed in highest concentration followed by n-prp. Correlations Between Parabens and Matrices Highly significant positive correlations were observed between all four parabens excreted in urine, all with P-levels o In Figure 3a the correlation between MeP and the three other parabens are shown. In serum MeP and n-prp Journal of Exposure Science and Environmental Epidemiology (2011) 21(3) 267

7 Frederiksen et al. Determination of parabens in human samples Table 4. Levels of parabens (ng/ml) in urine, serum and seminal plasma from young Danish men (n ¼ 60). Matrix Paraben nolod a Percentiles Maximum 25th Median 75th Urine MeP EtP n-prp n-bup BzP 56 olod olod olod 2.06 Serum MeP EtP 42 olod olod n-prp n-bup 58 olod olod olod 0.87 BzP 58 olod olod olod 0.29 Seminal plasma MeP 27 olod EtP 22 olod n-prp n-bup 29 olod BzP 55 olod olod olod 1.48 Abbreviations: LOD, limit of detection; n, number of samples. a Number of samples below LOD. were 4LOD in most of the samples and a significant correlation between them was observed (Figure 3b). EtP was only 4LOD in 18 samples; however, the concentration of EtP in serum also correlates significantly to the MeP concentration (Figure 3b). No other correlations between the parabens in serum were observed. Similar to urine, MeP, EtP, n-prp and n-bup in seminal plasma all were significantly correlated to each other. In Figure 3c the correlations of MeP to the other measurable parabens in seminal plasma are shown. Urine, blood and seminal fluid were sampled from each of the men within 1 h and for those parabens that could be measured the concentration correlates significantly between the three different matrices (Figure 4a d). Discussion All of the analyzed parabens except BzP were found in most of the 60 urine samples from the young Danish men. BzP is the most hydrophobic of the parabens we analyzed for and it is possible that the lack of BzP in urine could be due to it being excreted in urine as the water soluble but non-specific metabolite p-hydroxybenzoic acid rather than in its parent form. However, also in serum and seminal fluid were the BzP levels olod in the majority of the samples indicating that the exposure to this parabens is much lower than the other parabens. For MeP, EtP and n-prp the maximum levels measured were more than fold higher than the average levels measured suggests a wide range of exposure among participants. Compared with previous studies of urinary concentration of parabens in US male and female adults (Ye et al., 2006a) the median urinary concentration of the parabens were in general about 2.5-fold lower in our Danish men, except of EtP, which was twice as high in the Danish men. This may represent a country difference in use of parabens between the USA and Denmark. However, the US study also included women, whereas our study did not, and thus the difference in excretion pattern may also reflect a difference in exposure between women and men. As far as we know parabens have never before been measured in human seminal plasma. Traces of parabens in seminal plasma were expected, because our previous rat study showed that parabens were distributed to several fluids and tissues (Frederiksen et al., 2008). However, it was a surprise that several of the parabens could be detected in seminal plasma, and for n-prp and n-bup the levels were even higher in seminal plasma than in serum. Adverse effects of PrP and BuP on male reproduction such as decreased serum testosterone, sperm concentration and motility of the spermatozoa have been observed in animal studies (Oishi, 2001, 2002, 2004; Kang et al., 2002; CIR Expert Panel, 2009), whereas MeP and EtP do not seem to have the same effects. We cannot say whether the parabens we measured in seminal plasma derived from fluids coming from the testis together with the spermatozoa and thus reflect a direct exposure of the testis or whether they derived from seminal fluid coming from the accessory glands. But irrespective of the route, it is concerning that we in seminal plasma observed relatively high concentration of n-prp and that a high number of samples also contained measurable levels of n-bup, as these two parabens are suspected to affect male reproduction (Oishi, 2001, 2002). In a previous study we found a higher concentration EtP and n-bup in rat amniotic fluid compared to maternal serum. Furthermore the con- 268 Journal of Exposure Science and Environmental Epidemiology (2011) 21(3)

8 Determination of parabens in human samples Frederiksen et al. Paraben conc. in urine (ng/ml) EtP, r=0,664. p<0.001 n-prp, r=0.719, p<0.001 n-bup, r=0.572, p< MeP concentration in urine (ng/ml) Paraben conc. in serum (ng/ml) EtP, r=0.381, p<0.005 n-prp, r=0.606, p<0.001 n-bup, r=0.124, p>0.05 Paraben conc. in seminal plasma (ng/ml) EtP, r=0.402, p<0.005 n-prp, r=0.505, p<0.001 n-bup, r=0.422, p<0.005 MeP conc. in serum (ng/ml) MeP conc. in seminal plasma (ng/ml) Figure 3. Correlations (Spearman) between the concentration of MeP and other parabens (EtP, n-prp and n-bup) detected in urine (a), serum (b) and seminal plasma (c). All samples (n ¼ 60 from each matrices) were included in the plot, the concentration olod and not detectable samples were in the present plot set to Bhalf of the lowest LOD. Journal of Exposure Science and Environmental Epidemiology (2011) 21(3) 269

9 Frederiksen et al. Determination of parabens in human samples MeP in serum, r=0.866, p<0.001 MeP in sem. pl., r=0.390, p<0.005 EtP in serum, r=0.629, p<0.001 EtP in sem. pl., r=0.496, p<0.001 ng/ml ng/ml MeP in urine (ng/ml) EtP in urine (ng/ml) n-prp in serum, r=0.618, p<0.001 n-prp in sem. pl., r=0.657, p<0.001 n-bup in serum, r=0.449, p<0.001 n-bup in sem.pl., r=0.414, p<0.005 ng/ml ng/ml n-prp in urine (ng/ml) n-bup in urine (ng/ml) Figure 4. Correlations (Spearman) between the urinary concentration of MeP (a), EtP (b), n-prp (c) and n-bup (d) and the concentration of the respective parabens in serum and seminal plasma. All samples (n ¼ 60 from each matrices) were included in the plot, concentration olod and not detectable samples were in the present plot set to Bhalf of the lowest LOD. centration in amniotic fluid was proportional with dose, whereas the concentration in maternal serum were similar in all dose groups suggesting that continuous exposure during pregnancy maybe lead to an accumulation of the parabens in amniotic fluid (Frederiksen et al., 2008). We speculate whether parabens also accumulate in seminal plasma and whether seminal plasma paraben concentrations, especially n-pup and n-bup, may have any impact on sperm quality in adult men as seen in rodents (Oishi, 2001, 2002). Further studies of the correlation between paraben exposure and male reproductive parameters such hormone levels and semen quality are needed to answer this questions. Significant positive correlations between the different parabens were observed, indicating that men, who are exposed to high levels of one of the parabens, are likely to also have relatively high exposure to the other parabens. This may be owing to the fact that parabens often are used in combination in consumer products (CIR Expert Panel, 2009). However, the question remains whether a general higher urinary excretion of parabens seen in some men is truly due to higher exposure. Considering that parabens are believed to rapidly metabolize into the non-specific metabolite p-hydroxybenzoic acid, an alternative explanation for a general higher urinary excretion of parent parabens in some men might be that these men have a less effective paraben metabolism. Also between the three different matrices we did observe significant positive correlations between the paraben concentrations. This indicates that urinary paraben levels can be used as a biomarker for the circulating levels of parabens. As urine in general contained higher levels of parabens and is easier to sample, it appears to be the preferred matrix to study human exposure. Our results show, however, that 270 Journal of Exposure Science and Environmental Epidemiology (2011) 21(3)

10 Determination of parabens in human samples Frederiksen et al. serum or seminal plasma parabens levels also could be used as biomarker of paraben exposure. The levels of parabens we measured in urine, serum and seminal plasma from healthy, young Danish men presumably reflect only a fraction of human parabens exposure, as most of the parabens taken up are believed to be rapidly metabolized into harmless and unspecific metabolites. The fraction of the parent compound of parabens excreted in urine may depend on the route of exposure. Uptake of parabens through skin may thus result in more parent compound avoiding degradation in the liver before urinary excretion compared with oral uptake. The presence of parent compounds in the three different human matrices nevertheless indicates that parabens-uptake results in the exposure of human tissues to the potentially harmful parent compounds. Conflict of interest The authors declare no conflict of interest. Acknowledgements For skilled technical assistance, the authors thank Ole Nielsen, Department of Growth and Reproduction, Rigshospitalet, Copenhagen University Hospital, Copenhagen, Denmark. The instrumental equipment was financially supported by Velux Fonden and this work was supported by the Danish Research Agency (Grant no ) and the Lundbeck Foundation (j.nr. 124/05). References Akomeah F., Nazir T., Martin G.P., and Brown M.B. Effect of heat on the percutaneous absorption and skin retention of three model penetrants. Eur J Pharm Sci 2004: 21: Blount B.C., Milgram K.E., Silva M.J., Malek N.A., Reidy J.A., Needham L.L., and Brock J.W. Quantitative detection of eight phthalate metabolites in human urine using HPLC-APCI-MS/MS. Anal Chem 2000: 72: Byford J.R., Shaw L.E., Drew M.G., Pope G.S., Sauer M.J., and Darbre P.D. Oestrogenic activity of parabens in MCF7 human breast cancer cells. JSteroid Biochem Mol Biol 2002: 80: CIR Expert Panel. Final amended report on the safety assessment of methylparaben, ethylparaben, propylparaben, isopropylparaben, butylparaben, isobutylparaben, and benzylparaben as used in cosmetic products. Int J Toxicol 2009: 27(Suppl 4): Darbre P.D., Aljarrah A., Miller W.R., Coldham N.G., Sauer M.J., and Pope G.S. Concentrations of parabens in human breast tumours. J Appl Toxicol 2004: 24: Darbre P.D., Byford J.R., Shaw L.E., Hall S., Coldham N.G., Pope G.S., and Sauer M.J. Oestrogenic activity of benzylparaben. J Appl Toxicol 2003: 23: Darbre P.D., Byford J.R., Shaw L.E., Horton R.A., Pope G.S., and Sauer M.J. Oestrogenic activity of isobutylparaben in vitro and in vivo. J Appl Toxicol 2002: 22: Frederiksen H., Taxvig C., Hass U., Vinggaard A.M., and Nellemann C. Higher levels of ethyl paraben and butyl paraben in rat amniotic fluid than in maternal plasma after subcutaneous administration. Toxicol Sci 2008: 106: Hoberman A.M., Schreur D.K., Leazer T., Daston G.P., Carthew P., Re T., Loretz L., and Mann P. Lack of effect of butylparaben and methylparaben on thereproductivesysteminmalerats.birth Defects Res B Dev Reprod Toxicol 2008: 83: Janjua N.R., Frederiksen H., Skakkebaek N.E., Wulf H.C., and Andersson A.M. Urinary excretion of phthalates and paraben after repeated whole-body topical application in humans. Int J Androl 2008: 31: Janjua N.R., Mortensen G.K., Andersson A.M., Kongshoj B., Skakkebaek N.E., and Wulf H.C. Systemic uptake of diethyl phthalate, dibutyl phthalate, and butyl paraben following whole-body topical application and reproductive and thyroid hormone levels in humans. Environ Sci Technol 2007: 41: Jorgensen N., Carlsen E., Nermoen I., Punab M., Suominen J., Andersen A-G., Andersson A-M., Haugen T.B., Horte A., Jensen T.K., Magnus Ø., Petersen J.H., Vierula M., Toppari J., and Skakkebæk N.E. East-West gradient in semen quality in the Nordic-Baltic area: a study of men from the general population in Denmark, Norway, Estonia and Finland. Hum Reprod 2002: 17: Kang K.S., Che J.H., Ryu D.Y., Kim T.W., Li G.X., and Lee Y.S. Decreased sperm number and motile activity on the F1 offspring maternally exposed to butyl p-hydroxybenzoic acid (butyl paraben). JVetMedSci2002: 64: Lemini C., Jaimez R., Avila M.E., Franco Y., Larrea F., and Lemus A.E. In vivo and in vitro estrogen bioactivities of alkyl parabens. Toxicol Ind Health 2003: 19: Lobemeier C., Tschoetschel C., Westie S., and Heymann E. Hydrolysis of parabenes by extracts from differing layers of human skin. Biol Chem 1996: 377: Oishi S. Effects of butylparaben on the male reproductive system in rats. Toxicol Ind Health 2001: 17: Oishi S. Effects of propyl paraben on the male reproductive system. Food Chem Toxicol 2002: 40: Oishi S. Lack of spermatotoxic effects of methyl and ethyl esters of p-hydroxybenzoic acid in rats. Food Chem Toxicol 2004: 42: Okubo T., Yokoyama Y., Kano K., and Kano I. ER-dependent estrogenic activity of parabens assessed by proliferation of human breast cancer MCF-7 cells and expression of ERalpha and PR. Food Chem Toxicol 2001: 39: Pugazhendhi D., Pope G.S., and Darbre P.D. Oestrogenic activity of p-hydroxybenzoic acid (common metabolite of paraben esters) and methylparaben in human breast cancer cell lines. J Appl Toxicol 2005: 25: Soni M.G., Carabin I.G., and Burdock G.A. Safety assessment of esters of p-hydroxybenzoic acid (parabens). Food Chem Toxicol 2005: 43: Tavares R.S., Martins F.C., Oliveira P.J., Ramalho-Santos J., and Peixoto F.P. Parabens in male infertility-is there a mitochondrial connection? Reprod Toxicol 2009: 27: 1 7. Taxvig C., Vinggaard A.M., Hass U., Axelstad M., Boberg J., Hansen P.R., Frederiksen H., and Nellemann C. Do parabens have the ability to interfere with steroidogenesis? Toxicol Sci 2008: 106: Ye X., Bishop A.M., Reidy J.A., Needham L.L., and Calafat A.M. Parabens as urinary biomarkers of exposure in humans. Environ Health Perspect 2006a: 114: Ye X., Kuklenyik Z., Bishop A.M., Needham L.L., and Calafat A.M. Quantification of the urinary concentrations of parabens in humans by online solid phase extraction-high performance liquid chromatography-isotope dilution tandem mass spectrometry. J Chromatogr B Analyt Technol Biomed Life Sci 2006b: 844: Ye X., Tao L.J., Needham L.L., and Calafat A.M. Automated on-line columnswitching HPLC-MS/MS method for measuring environmental phenols and parabens in serum. Talanta 2008: 76: Journal of Exposure Science and Environmental Epidemiology (2011) 21(3) 271