Food Emulsifier Glycerin Monostearate Increases Internal Exposure Levels of Six Priority Controlled Phthalate Esters and Exacerbates Their Male Reproductive Toxicities in Rats

Human beings are inevitably exposed to ubiquitous phthalate esters (PAEs). Processed, packaged foods are popular nowadays, in which emulsifiers are frequently added as food additives. It is unclear how emulsifiers affect the bioavailability of ingested PAEs contaminants and their toxicities. The purposes of our study were to explore whether food emulsifier Glycerin Monostearate (GMS) could increase the internal exposure levels of six priority controlled PAEs and affect their reproductive toxicities when male rats are exposed to PAEs mixture (MIXPs). The male rats were exposed to MIXPs by gavage for thirty days in combination with or without given GMS. Phthalate monoesters (MPAEs), primary metabolites of PAEs, in rat urine were used as biomarkers to predict the internal exposure levels of the six PAEs, and their concentrations were determined using UPLC-MS. The reproductive toxicity was evaluated using serum testosterone levels test and histopathology of testes. Results showed that compared to PAEs exposure alone, the internal exposure levels of PAEs increased by 30%-49% in the presence of GMS. PAEs exposure led to the reduction of testosterone level by 23.4%-42.1% in the presence and absence of GMS, respectively, compared to the baseline. Testosterone levels in MIXPs+GMS and DEHP+GMS group were decreased by 9.1% and 13.6%, respectively, compared with MIXPs and DEHP group. Histopathology showed that injuries of testis (deciduous spermatids) were observed, and GMS exacerbated the injuries. The results indicated food emulsifiers chronically taken up might increase safety risks of food PAEs contaminants.


Introduction
In this study, we aimed to investigate the effects of food emulsifier on PAEs' internal exposure levels and male reproductive toxicities using the mixture of six priority controlled PAEs along with the most widely used food emulsifier GMS. Phthalate monoesters (MPAEs), the primary metabolites of PAEs in rat urine, were used as biomarkers to predict the internal exposure levels of the six PAEs, which were quantified using UPLC-MS. The toxicity was evaluated using serum testosterone level test and histopathology of testis. To our knowledge, this study is the first one to evaluate the impact of food emulsifiers on the internal exposure levels and toxicity of food contaminants.  to the standard animal laboratory food for 7 days before experiments at SPF grade Animal Laboratory of Southeast University. Water was supplied ad libitum from glass bottles. Animal experiments were conducted following the "Principles of laboratory animal care" (NIH publication 86-23, revised 1986) and local regulations. Furthermore, all experiments were approved and supervised by the Animal Care and Use Committee and the Animal Ethics Committee at Southeast University (Approval Number: 2015-0123-006).

Ultra-performance liquid chromatography-mass spectrometry analyses
An ultra-performance liquid chromatography (UPLC) method was established for the detection of phthalate monoesters (MMP, MEP, MBP, MBzP, MEHP and MOP), the primary metabolites of PAEs, in rat urine as biomarkers to predict the internal exposure levels of six PAEs. UPLC analysis was carried out using a Thermo Accela system. Sample separation was achieved on Betasil phenyl column (100 mm×2 mm, 3 μm) at a flow rate of 350 μL/min using a mobile phase of 0.1% acetic acid solution (phase A): methanol (phase B). A gradient elution started with 30% B over 4 min, followed by the first linear increase to 40% B over 4 min, the second linear increase to 50% B over 2 min, and the third linear increase to 70% B over 6 min with additional 4 min hold. At the end, a linear decrease to 30% B over 1 min with 1 min hold were applied. The column temperature was set to 30°C. The sample volume was 5 μL. Mass spectrometry was carried out using a triple quadrupole tandem mass spectrometer (Thermo API4000) equipped with an electrospray interface (ESI) and MRM scan mode. Ions were created in the negative ion mode with the sprayer voltage -4.5 kV and the ion source temperature at 500°C. The pressure of collision gas, sheath gas, atomization gas and heat gas were In order to avoiding the migration of PAEs from PVC to urine sample, all the materials and reagents have been kept away from plastic vessels or apparatus, and glass tubes were used in urine sample collection, store and test throughout this study.
The baseline levels of PAEs metabolites were determined using urine samples that were collected before PAEs exposure (namely, blank urine). The blank urine (900 μL) was spiked with 100 μL the quality control solutions with different concentrations, respectively, to make eight calibration standards at the final concentrations of 1, 10, 20, 50, 100, 200, 500 and 1,000 ng/ mL. To prepare quality control samples at the final concentrations of 5, 20 and 100 ng/mL, the blank urine was spiked with the appropriate amount of quality control solutions, respectively. Fresh quality control samples were used to daily evaluate method parameters including extraction recovery, accuracy, precision and matrix effect. All the samples were stored at -20°C to maintain stability.
The matrix effect of each MPAEs was in the range of -11.33% to 5.95%, with relative standard deviation (RSD) <9%. Good recoveries (82%-119%) were observed over a linear range of 5-100 ng/mL, indicating that the method was fairly accurate. Thus, this method was valid for the determination of MPAEs in the urine samples.

Sample preparation
The frozen urine samples (stored at -80°C freezers) were thawed under room temperature and subsequently filtrated through a 0.45 μm membrane filter. 1 mL urine sample in a 5 mL glass tube was mixed with 40 μL of D4-labelled I.S. mixture, followed by the addition of 250 μL of ammonium acetate buffer solution (pH = 6.5) and 20 μL of β-glucuronidase (140 U/mL, Sigma). The mixture was sealed and incubated in a water bath of 37°C for 90 min to hydrolyze the conjugates of glucuronic acid and phthalate metabolite to release MPAEs. After centrifugation at 4,000 rpm for 10 min, the supernatant was collected, and then diluted with 1 mL phosphate buffer solution (pH = 2.0) for further solid-phase extraction (SPE).
The dilution was passed through the pre-conditioned DVB SPE cartridges (with 1 mL acetonitrile and then 1 mL phosphate buffer solution) at a flow rate of 1 mL/min controlled by Auto-SPE vacuum pumping. 2 mL of formic acid (0.1 mol/L) and 1 mL of purified water was used to clean the cartridge. After cleaning, the pumping was kept on for another 1 min until the cartridge bed was completely dried. The target analytes retained on the cartridges were eluted with 1 mL acetonitrile. The eluent was evaporated to dry in 55°C water bath under nitrogen stream, and finally the dried substance was re-dissolved with 200 μL methanol for LC-MS/MS analysis. Rats in control group received isometrical 0.5% CMC-Na by ig. The treatments continued for 30 days. All rats were monitored each day for their health and mental status during the experimental period. Rats, except those in control group, were raised in metabolism cages to collect 24-hour urine on day 1 to day 7, and then on days 10, 15, 21, 27 and 30. The urine collected before exposure of rat to PAEs was collected to measure the background level of PAEs for selfcontrol. No rat became severely ill during the experiments or died prior to the experimental endpoint. The urine samples were stored at -80°C before UPLC analysis. Urine creatinine was detected by Beckman Coulter DXC800 fully automatic biochemical analyzer.

Internal exposure study
The estimated daily intake of PAEs based on urine MPAEs concentration were regarded as internal exposure dose, which was calculated by formula (1) according to Chen et al [10] and David [32], and relative bioavailability (F rel ) was estimated by formula (3) to compare the internal exposure dose of PAEs in the presence and absence of GMS. where, UE is urinary excretion of MPAEs (mg/g creatinine) and calculated by the formula (2); We calculated CE (creatinine excretion) using the formula: CE = creatinine concentration× urine volumeÄ body weight. CE is a constant and calculated to be 44 mg/kg/d for male rats in this study. F UE is a fractional urinary excretion value of 0.69 for MEP [18] and MBP [33], 0.024 for MEHP [34], 0.73 for MBzP [33], and 0.043 for MOP [18]. DMP and DEP were assumed to have an equal excretion factor as DBP [10,18]. MWd and MWm are molecular weight of PAEs and MPAEs, respectively.

Toxicological effects on testis
At the beginning and end of the experiment, about 0.75 mL blood of each rat was collected from caudal vein by cutting off the tip of their tails, and centrifuged at 3000 rpm for 10 min at 4°C for serum preparation. Serum testosterone (T) level was tested using a commercial assay kit for T (Nanjing JianCheng Bioengineering Institute, Nanjing, China). All rats were fasted overnight (12 h) and euthanized using 4% fluothane in oxygen anesthesia in an airtight container. When the rats had no-blink reflex and no autonomous breath, or no heartbeat, the left testis of each rat was removed for fix in 10% buffered formalin, followed by embedding in paraffin. The 3-5 mm thick sliced sections were stained by Hematoxylin-Eosin. Morphological analyses were performed by an experienced pathologist blinded to the treatment information. The slides were examined by TS100 inverted microscope (Nikon, Japan), and photographed under ×200 magnification. The tissue sections were scored with respect to the degree of injury.

Statistical analyses
All statistical analyses were performed using Statistical Product and Service software, version 17.0 (SPSS Inc., Chicago, IL, USA). Comparison between with and without GMS groups was performed using an independent-samples T test. Multiple comparisons were performed using one-way ANOVA; if the variances are homogenous, the least significant different t test is used, otherwise, a Tammany's T2 test is used. Values are expressed as means and their corresponding standard deviations. Differences with P <0.05 were considered statistically significant.

UPLC-MS analysis of MPAEs metabolites in urine
A good linear relationship was obtained for the six MPAEs with the concentrations ranging from 1.0 to 10,000 ng/mL. The linear regression parameters for the calibration curve of each phthalate monoesters are listed in Table 1. The correlation coefficients (R 2 ) of all MPAEs were more than 0.99. The limit of detection (LOD) was 0.5 ng/mL for MMP, MEP, MBP and MBzP, and 0.3 ng/mL for MEHP and MOP. The limit of quantitation (LOQ) was 1.5 ng/mL for MMP, MEP, MBP and MBzP, and 1.0 ng/mL for MEHP and MOP, according to decuple signal-to-noise ratio. The coefficient of variation (CV) for each MPAE to determine inter-and intraday precision of 5, 20 and 100 ng/mL was <14.5%, indicating that the method was linear and reproducible.

GMS increases internal exposure level and relative bioavailability of PAEs
Day 0 PAEs exposure levels in the rats were assigned as the background exposure for self-control. As shown in Table 2, the results showed that they were very low, even under the LOD. After the rats exposure to PAEs, the internal exposure levels of PAEs in the rats of MIXPs+GMS group were significantly higher than those of MIXPs group (P <0.05) (Fig 2). The internal exposure levels of PAEs in rats continuously elevated in both MIXPs group and MIXPs+GMS group with a rapid increase in the first several days (Fig 2A-2F), and the F rel of each PAEs ranged from 130% to 148% (Fig 2H). In addition, the internal exposure level of DEHP in the rats also continuously increased when administered with DEHP in the presence and absence of GMS (Fig 2G), and the DEHP F rel in this group was 139%, obviously higher than DEHP F rel (130%) in rats administered with MIXPs (Fig 2H), indicating the possibility of competitive absorption of PAEs.

Testis toxicity
Serum T levels in experimental rats declined significantly (23.4%-42.1%) at the final stage of the experiment, compared to their self-initial stage and control rats (P <0.05). However, no different change of T level was observed over the period of experiment in control rats. As shown in Fig 3, T level in MIXPs+GMS group was reduced 9.1% comparing to MIXPs group (P <0.05). Similarly, T level in DEHP+GMS group was reduced 13.6% comparing to DEHP group. In addition, T levels in MIXPs group and MIXPs+GMS group were lower than DEHP group and DEHP+GMS group respectively, with notable differences between MIXPs and DEHP group (P <0.05). These results indicated that PAEs might pose reproductive toxicity and GMS could enhance their toxicities. Furthermore, there might have dose addition action among PAEs' toxicological effects of each component. Pathological morphology of testis was examined since T is mainly generated by leydig cells in testis, a main target organ of PAEs. As shown in Fig 4, there were seminiferous tubules (S) with complete arrangement of germinal epithelium in control group with no injury (Fig 4A). Spermatids became sparse and plenty of deciduous spermatids (arrow) were observed in seminiferous tubules in MIXPs+GMS group (Fig 4C), while lots of sperm cells arranged orderly in the tubules and only a few deciduous spermatids were observed in MIXPs group (Fig 4B).
No deciduous spermatids were observed in DEHP+GMS and DEHP group; nevertheless, the spermatids in DEHP+GMS group were sparser than those in DEHP group with disordered arrangement (Fig 4D and 4E). The injury scores of MIXPs+GMS and DEHP+GMS group were higher than MIXPs and DEHP group, respectively (Fig 4F). The results of testicular pathological morphology were correlated with serum T levels, which supported the hypothesis that GMS enhanced PAEs' reproductive toxicity. In addition, we also examined the morphological analyses of liver (S1). Similarly, the toxic effects were also observed in liver.

Discussion
In human and animal body, PAEs are rapidly metabolized to their corresponding monoesters, and some of which can be further metabolized to the oxidation products of their lipophilic aliphatic side chain. For example, DMP is metabolized to MMP, while DEHP is metabolized to MEHP, which can be further metabolized to MEHHP, MEOHP and so on [5]. The simple monoester metabolites of PAEs are still the most common biomarkers to evaluate the exposure levels of their parent compounds [18,33,34]. There are some reports on the determination of MPAEs in urine, serum and milk by either HPLC or LC [35][36][37]. We have established HPLC method for analyzing MEHP and DEHP in plasma [38]. In this study, we developed a simple and rapid UPLC-MS method to determine the concentrations of MPAEs in rat urine and to evaluate the effects of food emulsifier GMS on PAEs internal exposure level. Theoretically, the UPLC-MS method has significant advantages in speed, resolution, and sensitivity of analysis, especially time saving and solvent consumption [39]. The UPLC-MS method developed in this study is validated with respect to the linearity range and the limit of detection. The retention time of MMP, MEP, MBP, MBzP, MEHP and MOP is 2.14, 7.54, 12.87 13.21, 18.58 and 19.20 min, respectively, with no interference from endogenous urine constituents at or near the retention time of those targeted analytes. The validated method has been applied to successfully determine the concentrations of MPAEs in urine sample in this study. The usage of emulsifiers, such as GMS and polysorbate 20, in insoluble drugs is to improve their oral bioavailability by maintaining their stability and facilitating their release [29,30,40]. Emulsifiers are often used in food manufacture with higher concentrations, which may increase the oral bioavailability of food contaminants, such as PAEs. It has been reported that the oral bioavailability of DEHP in rats following single oral administration without emulsifier was 6.74% in the study of Chang-Liao et al [41] and 13.6% in the study of Pollack et al [42]. In our previous study, we found that polysorbate 80 could increase intestinal absorption of DEHP in rats, and the relative bioavailability of DEHP significantly increased by 1.9 folds [38]. In the present study, we found that food emulsifier GMS could increase the internal exposure level and oral bioavailability of six priority controlled PAEs, and the relative bioavailability of DMP, DEP, DBP, BBP, DEHP and DNOP increased by 1.48, 1.42, 1.34, 1.35, 1.30, and 1.39 folds, respectively.
Food emulsifiers are surfactants, which can reduce surface hydrophobicity, increase colonic permeability and make intestines sensitive to chemical aggression [43,44]. P-glycoprotein (Pgp), locating in cytomembrane, plays an important role in protecting the body from harmful substances by limiting their entry into enterocytes and/or pump them out of enterocytes into the intestinal lumen [44,45]. Zhu et al [30] found that polyoxyethylene (40) stearate (PS), a non-ionic surfactant, inhibited P-gp-mediated efflux by modulating substrate-stimulated P-gp ATPase activity. Similarly, another commonly used food emulsifier polysorbate 80, has inhibitive effect on P-gp [46]. Our previous study has demonstrated that polysorbate 80 could damage the structure and function of mitochondria in enterocytes, decrease mitochondrial respiration rate, and then inhibit the function of P-gp ATPase activity, leading to increased absorption of DEHP [38]. Barta et al [47] reported that GMS suppressed the activity and expression of P-gp in Caco-2 monolayer cells. Therefore, the potential mechanisms underlying the enhancement of PAEs' internal exposure by GMS may include increased intestinal permeability and suppressed P-gp in enterocytes. An ongoing study is being conducted to understand the underlying mechanisms of the synergetic function of GMS in PAEs bioavailabilities.
PAEs are endocrine disruptors, and have reproductive and developmental toxicity [6,48]. Liver and testis are main target organs of PAEs [49,50]. Helal et al [48] reported that DEHP exposure at 1,000 mg/kg/d for 6 weeks could decrease testis weight and degenerate the germinal epithelium and leydig cells in rat testis. Diisobuty phthalate (DIBP), diisoheptyl phthalate (DIHP), diisononyl phthalate (DINP) and DEHP exposure induced the reduction of fetal testicular T production in SD and Wistar rat by inhibiting the expression of steroidogenesis genes such as steroidogenic acute regulatory protein (StAR) and Cyp11a [51]. DBP also decreased testicular T level in rat by down-regulating the expression of StAR, P450scc (P450 side-chain cleavage enzyme), 3β-HSD (3β-hydroxysteroid dehydrogenase), P450c17, which are key genes in steroidogenesis [52]. Phthalate esters have the similar structure of benzene ring and two adjacent aliphatic chains. They exert similar toxicological effects, such as disrupting the function of endocrine and male reproductive toxicity [17]. Howdeshell et al [27] found that exposure to the mixture of DEP, DBP, BBP, DEHP, and diisobuty phthalate (DiBP) inhibited fetal testicular T production in SD rat in a cumulative, dose-additive manner. Such effects are commonly reported for exposure to very high doses of PAEs (>500 mg/kg/d). In our study, with the exposure level of 100 mg/kg/d, deciduous spermatids were observed in MIXPs+GMS group, and spermatids became sparse in DEHP+GMS group, as well as T declined. These findings suggest that GMS exacerbated PAEs' toxicity to testis. In addition, we also observed that GMS exacerbated PAEs' toxicity to liver (S1). The exacerbated damage of testis and liver may result from increased internal exposure level of PAEs in the presence of GMS.
One of the pitfalls in most studies including ours, the experimental PAEs exposure levels are much higher than the everyday human exposure doses, which are usually low but chronic, continuous, and cumulative. Although the experimental doses cannot accurately reflect actual human exposure scenarios, widely used emulsifiers in foods may significantly increase the toxicities of PAEs by enhancing their bioavailabilities, thereby leading to human health risks. Several previous studies have reported that the elevated incidence of autoimmune diseases is associated with increased consumption food emulsifiers [44,53,54]. It will be very interesting to explore how food emulsifiers modify the toxicities of the chronic exposure to PAEs at low dose based on actual human exposure levels.