Introduction

Perfluoroalkyl substances (PFASs) are characterized by stable and strong carbon–fluorine bonds with unique water- and oil-repellent properties, resulting in their wide application in industrial and consumer products (Calafat et al. 2007). However, the strong carbon–fluorine bonds have also resulted in their environmental persistence and bioaccumulation. Among PFASs, perfluorooctanoate (PFOA) is one of the most abundantly and consistently found in the environment and in humans (Barry et al. 2013; Castiglioni et al. 2015; Willach et al. 2016), with a half-life of 3.8 years in human serum (Olsen et al. 2007). The environmental persistence, bioaccumulation, and potential toxicity of PFOA resulted in many countries restricting its industrial production since 2000 (US EPA 2016). However, its persistence as one of the most dominant PFASs in the environment remains unchanged due to its continued production in countries without legal restrictions and via precursor degradation (Prevedouros et al. 2006; Wang et al. 2009).

A large number of studies on the potential health hazards of PFASs are available (Butenhoff et al. 2004; Kennedy et al. 2004; Lau et al. 2004, 2007; Olsen et al. 2009; Olsen and Zobel 2007). Classic laboratory toxicity from PFASs includes liver hypertrophy and tumors, such as those of Leydig cells (Biegel et al. 2001; Kennedy et al. 2004). PFOA is also reported as a potential endocrine disruptor, including reproductive disruption (White et al. 2011). PFOA exposure decreased testosterone level, but increased estradiol content in serum (Biegel et al. 1995; Zhao et al. 2010), which may play an import role in inducing testicular tumors in male rats (Biegel et al. 2001). Low-dose PFOA exposure to prenatal female offspring delayed mammary development from puberty to young adulthood (Tucker et al. 2015; White et al. 2007). Based on human epidemiological data, a Danish study reported that PFOA exposure may be associated with unexplained poor semen quality in young men (Joensen et al. 2009). Study also reported an association between in utero exposure to PFOA and lower sperm concentration and total count in male offspring (Vested et al. 2013). In a study of men attending an in vitro fertilization clinic, luteinizing hormone (LH) and free testosterone were positively correlated with blood PFOA (Raymer et al. 2012). However, epidemiological studies also describe inconsistent, negative, or no association between PFOA levels and reproductive parameters in adult men (Joensen et al. 2009; Raymer et al. 2012).

Our previous study showed that PFOA exposure significantly changed testicular proteomic profiles, in which corticosteroid-binding globulin (CBG, also called transcortin) was increased in PFOA-treated mice compared with the control (Zhang et al. 2014). Mainly secreted by the liver, CBG is a monomeric glycoprotein with binding specificity to anti-inflammatory steroids, such as glucocorticoids and progesterone (Hammond et al. 1991). In humans, its open reading frame (ORF) encodes a polypeptide of 405 amino acids, including a signal peptide, and its final product comprises five N-linked oligosaccharide chains with extensive heterogenicity (Avvakumov and Strelchyonok 1987; Hammond et al. 1987; Robinson et al. 1985). CBG belongs to a non-functional member of the serpin (serine protease inhibitor) superfamily (named SERPINA6). However, little homology has been found between CBG and other steroid-binding proteins, such as sex hormone-binding globulin, vitamin D-binding protein, or alpha-fetoprotein (Pemberton et al. 1988).

CBG is a substrate for neutrophil elastase, and plays a key role in the metabolism and action of glucocorticoids. After proteolytic cleavage by neutrophil elastase, which has a high concentration at inflammation sites, CBG undergoes a conformational transition resulting in the irreversible destruction of its steroid-binding site and release of its ligand, thus significantly increasing local free anti-inflammatory glucocorticoid concentrations at sites of inflammation (Henley and Lightman 2011; Klieber et al. 2007; Lin et al. 2010). Consistent with its regulation of circulating free hormone levels, a substantial increase in free corticosterone concentrations and aggravated response to septic shock have been observed in CBG-null mice, indicating its inability to properly respond to excess free corticosterone (Petersen et al. 2006; Willnow and Nykjaer 2010). Although CBG is primarily synthesized in the liver, its synthesis has been reported in many other tissues, including the lung, kidney, placenta, adipose tissue, and brain; however, the exact role of this locally expressed CBG remains unclear (Caldwell and Jirikowski 2014; Gulfo et al. 2016). In the present study, we explored the location and alteration of testicular CBG under PFOA treatment, as well as its possible effect on steroid hormone synthesis in the testes.

Materials and methods

Animals and PFOA treatment

Male BALB/c mice (aged 6–8 weeks, 20–27 g in weight) were purchased from Beijing Vital River Experimental Animal Technology Co. Ltd. (Beijing, China). The animals were housed in a mass air displacement room. After 1 week of adaptation, they were randomly divided into four groups (n = 15 per group) and treated by oral gavage with Milli-Q water or 1.25, 5, or 20 mg/kg/day of PFOA (Sigma-Aldrich, CAS no. 307-55-1, 99% purity, St. Louis, MO, USA) for 28 consecutive days. The exposure doses were based on earlier research (Yan et al. 2014). After treatment, the mice were killed, and their testes, livers and sera were collected. Part of each testis was fixed in 4% paraformaldehyde for immunofluorescence assay, with the remaining portion immediately frozen in liquid nitrogen and stored at − 80 °C for Western blotting and qPCR detection. All procedures were approved by the Ethics Committee of the Institute of Zoology, Chinese Academy of Sciences.

Cell culture and MTT

A mouse Leydig tumor cell line (mLTC-1 cells) was purchased from the American Type Culture Collection (Manassas, VA, USA) and grown in RPMI-1640 complete culture medium (HyClone, Ogden, UT, USA), containing 10% fetal bovine serum, 100 IU/mL penicillin, and 100 g/mL streptomycin, supplemented with 5% CO2 at 37 °C. The mLTC-1 cells were seeded in 96-well plates with fresh media containing various concentrations of PFOA (50, 100, 200, 300, and 400 µM). After 48 h of exposure, cellular viability was determined using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) (Amresco, Solon, OH, USA) assay. Briefly, 20 µL of MTT [5 mg/mL, dissolved in phosphate-buffered saline (PBS)] was added to each well for 4 h. The supernatants were then removed from the plates and dissolved in 150 µL of DMSO. Corresponding optical density was then measured at 490 nm using a spectrophotometer (BioTek, Synergy H1Hybrid Microplate Reader, USA).

Western blotting

Testes tissue, serum, and mLTC-1 cells were lysed in RIPA buffer replenished with protease inhibitor and phosphatase inhibitor cocktail and were used for Western blotting analysis. Proteins were resolved in 10% separating gel and transferred to a PVDF membrane (Millipore, Billerica, MA, USA). The membranes were then incubated with rabbit anti-mouse cytochrome P450 cholesterol side-chain cleavage enzyme (CYP11A1) (Millipore, Temecula, CA, USA), steroidogenic acute regulatory protein (StAR), 17α-hydroxylase/17,20 lyase (CYP17A1) (Abcam Cambridge, UK), goat anti-mouse 3β-hydroxysteroid dehydrogenase (3β-HSD) (Santa Cruz Biotechnology, Santa Cruz, CA, USA), and rabbit anti-mouse CBG anti-serum (gifted by Prof. Geoffrey L. Hammond from the Departments of Cellular and Physiological Sciences and Obstetrics and Gynaecology, University of British Columbia, Canada). Protein bands were detected by horseradish peroxidase-conjugated secondary antibodies (Boster Biological Technology, Wuhan, China) and chemiluminescent substrates by exposure to X-ray films. These bands were analyzed with QuantityOne software (v4.6.3, Bio-Rad, USA) and the data were normalized to GAPDH levels.

PCR and qPCR

Total RNA was extracted from testes, livers, and mLTC-1 cells using the RNeasy plus Mini Kit (Qiagen, Valencia, CA, USA). cDNA synthesis was performed with the same methods as our previous study (Wang et al. 2015), and the ORF of CBG was amplified by PCR. DNA electrophoresis was performed on 1% agarose gel. The qPCR was performed using the LightCycler®480 qPCR system (Roche Diagnostics GmbH, Switzerland), and data were analyzed with MxPro qPCR software. The comparative CT \(({2^{ - \Delta \Delta {C_{\text{T}}}}})\) method was used to calculate the fold change of mRNA levels (Livak and Schmittgen 2001). β-Actin was chosen as an internal control gene. Primers for ORF amplification of CBG were forward: 5′-tggtcaaccaaaacaatgtcgctc-3′ and reverse: 5′-gacaagcttttaggctggat-3′, and primers for qPCR were forward: 5′-accctcatcctgatcaactacatc-3′ and reverse: 5′-ccctggtctggaagaatgatga-3′.

Immunofluorescence

Testes sections (5 µm) were incubated with CBG antibodies at 4 °C overnight. The slides were then incubated with goat anti-rabbit IgG Alexa Fluor 488 (ZSGB-BIO, Beijing, China), and mounted with Vectashield mounting medium with 4′,6′-diamidino-2-phenylindole dihydrochloride (DAPI, ZSGB-BIO, Beijing, China). Fluorescence images were observed using a Nikon Eclipse 90i Fluorescence Microscope (Nikon Instruments Inc., Japan).

RNA-FISH

The probes for Cbg RNA-FISH were commercially synthesized by Biosearch Technologies (Petaluma, CA, USA), and hybridization was performed according to the manufacturer’s protocols. Briefly, testes from the control and PFOA-treated groups (1.25, 5, or 20 mg/kg/day) were fixed in freshly prepared paraformaldehyde (3.7% in DPBS). Following ethanol treatment (70%) for 1 h, the coverslips were then hybridized with QUASAR® 570 dye-labeled Cbg probes overnight at 37 °C in hybridization buffer (90% Stellaris® RNA-FISH hybridization buffer, 10% formamide). The testes were then dyed with DAPI. Finally, the coverslips were mounted in Vectashield® mounting medium (Vector Labs, Burlingame, CA, USA) and observed using a confocal laser-scanning microscope (Zeiss LSM 710, Zeiss, Oberko, Germany).

Serum progesterone, adrenocorticotropic hormone (ACTH), and corticosterone

The progesterone levels in mLTC-1 cells or media were measured using an ELISA Kit (Cayman, Ann Arbor, MI, USA). Total ACTH and corticosterone content in serum were detected by radioimmunoassay using commercial kits (Beijing North Institute of Biological Technology, China).

CBG overexpression and siRNA knockdown

The cDNA encoding CBG was subcloned in eukaryotic expression vector pRc/CMV (Invitrogen, San Diego, CA,) (designated as pRc/CMV-CBG). CBG siRNAs and non-targeting control siRNAs were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). The mLTC-1 cells were seeded in six-well plates with approximately 80% confluence, and on the second day were transfected with 2.5 µg of combination plasmid pRc/CMV-CBG using the Lipofectamine LTX reagent (Invitrogen, Carlsbad, CA, USA) or 80 pmol siRNA-CBG using the Lipofectamine RNAiMAX reagent (Invitrogen, Carlsbad, CA, USA), as suggested by the manufacturer. At 6 h after transfection, cells were incubated with or without 200 µM PFOA in RPMI-1640 complete culture medium for 48 h. At the end of treatment, the media were collected for progesterone detection. The cells were harvested, washed three times with PBS, and lysed for Western blotting or ELISA analysis.

LH stimulation

The mLTC-1 cells were cultured in six-well plates for 24 h. After this, the media was discarded, the cells were washed three times with PBS, and LH in fresh serum-free media (final concentration: 5 and 10 nM) was added to stimulate the cells for 3 h. At the end of incubation, the cells and media were collected for progesterone and CBG detection. Before CBG detection, the media were concentrated with an Amicon® Ultra-4 3K Centrifugal Filter Device (Merck, Millipore, Billerica, MA, USA).

Statistical analysis

Statistical analysis was performed using SPSS software (Version 18, SPSS, Inc., Chicago, IL, USA). Differences between groups were determined using one-way ANOVA followed by Duncan’s multiple range tests or independent sample t tests. All data were presented as means ± SE, and p < 0.05 was considered statistically significant.

Results

PFOA exposure increased CBG levels in mouse testes

To confirm that PFOA exposure increased CBG protein levels in mouse testes, we examined its content in testes in adult male mice treated with PFOA (0, 1.25, 5, or 20 mg/kg BW/day) by oral gavage for 28 consecutive days. Two electrophoretic isoforms of CBG were observed in the testes by Western blotting. Compared with the control group, the total level of CBG significantly increased in the 1.25 and 5 mg/kg/day PFOA treatment groups (Fig. 1a). Immunofluorescent assay was used to further detect the location of CBG in the mouse testes. Results indicated that CBG was located adjacent to the seminiferous tubules in the testes, where Leydig cells reside (Fig. 1b).

Fig. 1
figure 1

Western blotting (a) and immunofluorescence (b) analyses of CBG in mouse testes after PFOA exposure. For Western blotting analysis, testis homogenates from the control and PFOA-exposed mice (1.25, 5, and 20 mg/kg/day for 28 days) were separated by 10% SDS-PAGE. Results are means ± SE (n = 3). p < 0.05 signifies significant differences, with the same letter between two groups indicating no significant differences. For immunofluorescence analysis, testis sections from all groups were incubated with CBG antibodies and then with Alexa Fluor® 488-conjugated goat anti-rabbit IgG, with the nuclei then dyed with DAPI. Slices were observed using a confocal laser-scanning microscope

Full size image

Cbg mRNA expression in Leydig cells in testes

To confirm that CBG was expressed from interstitial cells in the testes, the ORF of CBG was amplified using cDNA extracted from testes, mLTC-1 cells, and the liver (as a reference) (Fig. 2a). Using qPCR, Cbg mRNA was detected in the testes, though at a very low level compared with that detected in the liver and mLTC-1 cells (Fig. 2b). In addition, Cbg mRNA in the mLTC-1 cells was found at a 2.5-fold lower level than that in the liver. Combined with the above immunofluorescent staining results, we inferred that the lower Cbg mRNA level in testes compared with that in mLTC-1 cells was due to the relatively small proportion of Leydig cells in the testes. RNA in situ hybridization was further carried out to locate Cbg mRNA in testes tissue. Similar to the immunofluorescence results in Fig. 1b, Cbg mRNA was specifically located in the Leydig cells (Fig. 2c).

Fig. 2
figure 2

CBG was expressed in Leydig cells. a Amplification products of the open reading frame (ORF) of CBG in mouse testes, livers, and mLTC-1 cells. b Quantitative real-time PCR (qPCR) analysis showed the relative CBG mRNA levels in mouse testes, livers, and mLTC-1 cells. c RNA-FISH showed that CBG mRNA was mainly expressed in Leydig cells in mouse testes. Relative mRNA levels are presented as means ± SE (n = 3). p < 0.05 signifies significant differences, with the same letter between two groups indicating no significant differences. For PCR in the testes and livers, whole tissues were used. For RNA-FISH assay, testis sections were hybridized with QUASAR® 570 dye-labeled RNA-FISH probes of CBG, and nuclei were dyed with DAPI and observed using a confocal laser-scanning microscope

Full size image

Serum CBG and corticosterone increased after PFOA treatment

Serum CBG levels were detected after PFOA exposure, with elevated levels observed in all three PFOA treatment groups (Fig. 3a). Serum corticosterone levels also increased more than twofold in the 5 and 10 mg/kg/day PFOA groups (Fig. 3b). ACTH showed a decreasing trend after PFOA treatment, with statistical significance in the highest dose group (Fig. 3c).

Fig. 3
figure 3

Serum CBG, corticosterone, and ACTH levels after PFOA exposure. a Serum total CBG levels in the control and PFOA treatment groups (1.25, 5, and 20 mg/kg/day, 28 days) were detected by Western blotting. Serum was diluted at 1:1000 and immunoblotted with CBG antibodies. Serum corticosterone (b) and ACTH (c) levels in the control and PFOA treatment groups were detected with radioimmunoassay. Results are means ± SE (n = 3 for Western blotting, n = 6 for RIA). p < 0.05 signifies significant differences, with the same letter between two groups indicating no significant differences

Full size image

PFOA treatment increased CBG expression in mLTC-1 cells

We next examined whether PFOA treatment changed CBG levels in the mLTC-1 cells. The mLTC-1 cells were treated with increasing concentrations of PFOA (50, 100, 200, 300, and 400 µM) for 48 h. First, cellular viability as an endpoint was determined using MTT assay to assess the toxicity of PFOA on mLTC-1 cells (Fig. 4a). The MTT results showed that PFOA did not influence cellular viability in the 50 and 100 µM PFOA treatment groups, but viability decreased by 20, 35, and 40% in the 200, 300, and 400 µM groups, respectively. The CBG levels were detected by Western blotting under the above treatment doses, and despite the decrease in cellular viability following 200, 300, and 400 µM PFOA treatment, a dose-dependent increase in CBG was observed (Fig. 4b).

Fig. 4
figure 4

CBG protein increased in mLTC-1 cells after PFOA treatment. mLTC-1 cells were treated with various concentrations of PFOA for 48 h, and cellular viability (a) and (b) protein levels of CBG were detected with MTT assay and Western blotting, respectively. Results are means ± SE (n = 3 for Western blotting analysis, n = 6 for MTT assay). p < 0.05 signifies significant differences, with the same letter between two groups indicating no significant differences

Full size image

CBG increased progesterone and steroid hormone synthesis proteins in mLTC-1 cells

As a steroid hormone-binding protein, we examined whether CBG influenced steroid hormone production in the mLTC-1 cells. We overexpressed CBG in the mLTC-1 cells, and found higher progesterone levels in the cells compared to that in the pRc/CMV vector group (Fig. 5a). Exposure of mLTC-1 cells to PFOA (200 µM for 48 h) decreased progesterone levels; however, these levels were partly rescued with CBG transfection. To explore whether CBG affected progesterone levels in vitro through steroidogenesis metabolism, the levels of proteins involved in steroidogenesis metabolism were determined in mLTC-1 cells. CBG overexpression increased StAR, CYP11A1, CYP17A1, and 3β-HSD levels in mLTC-1 cells (Fig. 5b). Exposure to 200 µM of PFOA for 48 h resulted in significant down-regulation of CYP11A1 and CYP17A1 levels, as determined by Western blotting, whereas CBG overexpression, at least partly, rescued their expressions. Furthermore, we knocked down CBG in mLTC-1 cells using its siRNA, and found a significant decrease in the progesterone level compared with that in the control group (Fig. 5c), with proteins related to steroid hormone synthesis also decreased (Fig. 5d). Exposure to 200 µM of PFOA for 48 h resulted in the significant down-regulation of CYP11A1 and CYP17A1, with siRNA-CBG reducing them further.

Fig. 5
figure 5

CBG increased progesterone and steroid hormone synthesis protein levels in mLTC-1 cells. Overexpression of CBG increased progesterone secretion (a) and steroid hormone synthesis protein levels in mLTC-1 cells (b). CBG knockdown decreased progesterone secretion (c) and steroid hormone synthesis protein levels in mLTC-1 cells (d). mLTC-1 cells were transfected with pRc/CMV-CBG or its siRNA. After 6 h transfection, cells were treated with or without 200 µM of PFOA for 48 h, with progesterone secreted from the mLTC-1 cells into the supernatant then detected with ELISA assay, and cellular steroid hormone synthesis proteins, including steroidogenic acute regulatory protein (StAR), cytochrome P450 cholesterol side-chain cleavage enzyme (CYP11A1), 17α-hydroxylase/17,20 lyase (CYP17A1), and 3β-hydroxysteroid dehydrogenase (3β-HSD), detected by Western blotting analysis. Results are means ± SE (n = 3 for Western blotting analysis, n = 6 for ELISA assay). p < 0.05 signifies significant differences. *, #, and & means significant differences between indicated groups

Full size image

LH stimulated CBG production and release from mLTC-1 cells

LH treatment (5 and 10 nM, 3 h) dramatically stimulated the production of progesterone in mLTC-1 cells, as well as its release into the supernatant (Fig. 6a, b). Similar to progesterone, CBG levels increased in the mLTC-1 cells after LH stimulation (Fig. 6c), and its secretion into the media was enhanced (Fig. 6d).These results showed that the production and secretion of CBG in mLTC-1 cells were, at least partly, controlled by the LH signaling pathway.

Fig. 6
figure 6

Luteinizing hormone (LH) triggered the production and release of CBG in mLTC-1 cells. LH stimulus (5 and 10 nM for 3 h) triggered the production of progesterone in mLTC-1 cells (a) and its release (b) into culture media, as well as the production (c) and release (d) of CBG protein. Progesterone and CBG were detected using ELISA and Western blotting, respectively. Results are means ± SE (n = 4 for Western blotting analysis, n = 6 for ELISA assay). p < 0.05 signifies significant differences, with the same letter between two groups indicating no significant differences

Full size image

Discussion

Interference on serum CBG and corticosterone by PFOA as an endocrine-disrupting effect

The hypothalamic–pituitary–adrenal (HPA) axis is a major system involved in physiological and psychological stress responses (Dallman et al. 2006), in which hypothalamic corticotrophin-releasing hormone (CRH) stimulates ACTH secretion in the pituitary, and ACTH triggers glucocorticoid release from adrenal glands into the blood; free glucocorticoids, in turn, negatively regulate themselves by inhibiting CRH and ACTH release (Richard et al. 2010). The plasma CBG level plays an important role in HPA axis activity and regulation because CBG determines the levels of free and biologically active glucocorticoids.

Only a small fraction of glucocorticoid hormones are free in circulation, and the largest fraction of these hormones is bound to CBG. Therefore, CBG serves as a reservoir of biologically inactive protein-bound steroids, and controls the amount of free hormones for target tissues (Mendel 1989). Studies have revealed that CBG possesses an exposed elastase-sensitive loop, known as the reactive center loop (RCL), which, upon cleavage by neutrophil elastase, adopts a “relaxed” conformational change in tertiary protein structure, leading to a loss of glucocorticoid binding affinity and the delivery of cortisol to sites of inflammation (Klieber et al. 2007; Lewis and Elder 2011; Pemberton et al. 1988). Previous studies suggest fine control interaction between plasma CBG and free glucocorticoids, in which CBG modulates dynamic equilibrium between CBG-bound and unbound glucocorticoids (Mattos et al. 2013). Under PFOA treatment, blood CBG levels increased significantly in all three PFOA-treated groups. PFOA changed the serum CBG concentration, and thereby may influence corticosteroid bioavailability. Various hormone treatments can cause fluctuation in CBG levels and, therefore, influence the levels of free glucocorticoids in plasma (Smith and Hammond 1992). CBG thus represents a possible target for endocrine disruptors (Hampl et al. 2016), and interference in blood CBG in our present study may be considered as an endocrine-disrupting effect of PFOA.

PFOA not only increased blood CBG levels in mice, but also increased blood corticosterone levels two- to threefold in the 5 and 20 mg/kg/day PFOA-treated groups. A decreasing trend in serum ACTH was also observed in the PFOA exposure groups, with statistical significance in the 20 mg/kg/day group, which can be due to the negative feedback by increase serum corticosterone. Positive association between CBG and total cortisol levels has been observed in people following Trier Social Stress Tests (TSST) (Kumsta et al. 2007), and an increase in circulating CBG levels has also been reported in acute stress, which might restrain the rapid rise in free circulating corticosterone concentrations (Qian et al. 2011). However, other studies have reported decreased CBG expressions in patients and animals undergoing acute physiological stress, resulting in an increase in free glucocorticoids (Fleshner et al. 1995; Garrel 1996; Neufeld et al. 1994; Savu et al. 1981; Spencer et al. 1996).

There is likely a complex regulation between CBG and glucocorticoids. HPA axis regulation by CBG has been observed in rare CBG-deficient families, with coding sequence polymorphisms of CBG leading to reduced affinity to cortisol or inactive CBG (Lin et al. 2012). Most deficient patients show no symptoms of illness, though occasionally hypotension and fatigue, with very low total blood glucocorticoids, but normal free glucocorticoids in basal conditions (Brunner et al. 2003; Emptoz-Bonneton et al. 2000; Gagliardi et al. 2010; Henley and Lightman 2011; Hill et al. 2012; Lin et al. 2012; Perogamvros et al. 2010; Torpy et al. 2001). CBG-deficient mice (Cbg−/−) exhibit normal levels of gluconeogenetic enzymes, but aggravated sensitivity to LPS-induced septic shock, such as poor control of cytokine reactions and despair-like behavioral responses (Petersen et al. 2006; Richard et al. 2010). Conversely, although no glucocorticoid-response elements have been found in the Cbg promoter, putative binding sites for CCAAT/enhancer-binding protein beta (C/EBP β), which can tether to glucocorticoid receptors, are present, and chromatin immunoprecipitation shows increased co-recruitment of C/EBP β and glucocorticoid receptors to Cbg promoters after dexamethasone treatment, suggesting an inhibition effect of CBG by glucocorticoids (Verhoog et al. 2014). Furthermore, CBG levels following synthetic ACTH (synacthen) challenge also show a modest but significant decline in humans (Lewis et al. 2003). In addition, adrenalectomy and glucocorticoid administration can alter the rate of CBG production and secretion (Feldman et al. 1979). Therefore, the probability that the higher serum CBG levels observed in PFOA-treated mice were induced by increased total corticosterone, as a secondary effect of PFOA exposure, cannot be excluded.

While the role of CBG in the binding and transport of glucocorticoids in circulation is well documented, many additional functions of CBG have also been proposed. For example, plasma CBG not only binds cortisol and is involved in the regulation of the HPA axis, but it also has a relatively high affinity to progesterone (Hammond 1990). During pregnancy, high progesterone levels can displace cortisol from the steroid-binding sites of CBG in circulation and at the maternal-fetal interface (Benassayag et al. 2001). In addition, it has been reported in human populations that circulating CBG can serve as a marker of insulin secretion and its level correlates with markers in metabolic syndrome (Fernandez-Real et al. 1999, 2000, 2001, 2002, 2005). Metabolism disturbance is one of the most common toxic effects of PFOA observed in rodents (Lau et al. 2007), and alteration of blood CBG may contribute to the disturbance effect in PFOA-treated groups.

PFOA increased intrinsic CBG levels in testicular Leydig cells

By binding steroid hormones, especially glucocorticoids (cortisol in humans and corticosterone in rodents), and partly progesterone, CBG transports and influences the circulating concentrations and biological actions of these steroids. Circulating CBG is primarily synthesized in the liver, but immunoreactive studies have also shown it to be located in other steroid hormone target tissues (Gulfo et al. 2016). Consistent with our previous proteomic study, using immunoblotting we confirmed increased CBG levels in the testes after PFOA treatment in mice. Further fluorescent study showed the cellular presence of CBG in Leydig cells, which are adjacent to the seminiferous tubules. However, these results do not exclude the possibility that the fluorescent signal was not due to the initial proteins synthesized from testicular Leydig cells, but came from a CBG-steroid complex that entered Leydig cells by transportation or endocytosis from the blood. Previous studies support the existence of specific receptors for CBG on membranes of steroid response cells (Hryb et al. 1986), though work in identifying these specific cell-surface receptors has been unsuccessful. However, weak expression of Cbg in the testes has been reported previously by Northern blotting, although no work was carried out on its spatial expression patterns or role in testes (Hammond et al. 1987). We carried out amplification of the Cbg ORF in testes, mLTC-1 cells, and the liver (as a control). Our results revealed that Cbg mRNA was expressed in the testes and mLTC-1 cells, though at very low levels compared with that detected in the liver, the main tissue of CBG synthesis. A higher Cbg mRNA level was found in the mLTC-1 cells than in the testes, supporting the result that testicular Cbg was expressed dominantly in Leydig cells, which comprise only a very low percentage of cell mass in the testes. The specific expression of Cbg in Leydig cells, but not other cells, was further confirmed with in situ RNA-FISH. Taken together, our results confirmed, for the first time, that CBG is produced intrinsically in the Leydig cells of mice. For females, Cbg expression is also reported in uterine endometrial tissue, predominantly in the secretory phase, and its mRNA level is negatively correlated with progesterone levels in serum, though its role is not clear (Misao et al. 1994).

Glycosylation of CBG varied in different tissues and serum

Two electrophoretic isoforms were observed in the mouse testes by Western blotting, though only bands with high or low molecular weights were revealed in serum and mLTC-1 cells, respectively, indicating glycosylation heterogeneity of CBG in various tissues. Structural and functional characterization of CBG N-glycosylation using glycomics and glycoproteomics have shown that N-glycosylation sites in CBG are differently glycosylated, with mostly bi- and tri-antennary branching, suggesting different N-glycosylation under different hormonal states (Sumer-Bayraktar et al. 2011). Carbohydrates in CBG may exert diverse biological functions and influence the folding, secretion, and half-life of CBG, as well as the targeting of steroid hormones in circulation, although inconsistent results have been reported (Avvakumov and Hammond 1994; Avvakumov et al. 1993).

CBG increased progesterone and steroid hormone synthesis and responded to LH stimulation

CBG serves as a reservoir of plasma steroid hormones, particularly glucocorticoids, and regulates the amount of free hormones in target tissues. However, recent discovery suggests that CBG may play an additional role to glucocorticoid carrier. Genetic variants of CBG are associated with fatigue–pain syndromes and hypotension (Henley and Lightman 2011). In addition to that synthesized in the liver and secreted into circulation, CBG has also been located in intracellular compartments. Although the role of intrinsically expressed CBG from plasma is not clear, extrahepatic CBG synthesis suggests additional unknown functions (Hryb et al. 1986). Our study confirmed the specific expression of CBG in Leydig cells, located between Sertoli cells, which maintain proper levels of androgens, particularly testosterone, to stimulate and support spermatogenesis. We then explored the role of intrinsically expressed Leydig cell CBG using mLTC-1 cells, which maintain the properties of the original Leydig cells to some degree, and are commonly used in steroidogenesis-related research (Rebois 1982). As observed in our in vitro study, PFOA exposure increased the CBG protein in the mLTC-1 cells in a dose-dependent manner, even at doses that resulted in a 40% loss in cellular viability. Progesterone, rather than testosterone, was the dominant steroid hormone produced in the mLTC-1 cells. Consistent with our previous study, PFOA treatment decreased progesterone in the mLTC-1 cells (Zhao et al. 2017). PFOA is a suspected endocrine disruptor, resulting in lower testosterone levels in laboratory animals as well as in isolated Leydig cells, and some studies indicate that the structural lesions in testes in PFOA-treated mice are associated with the decrease of testosterone concentrations (Biegel et al. 1995; Kim et al. 2001; Shi et al. 2007; Sinha Hikim et al. 1997; Young and Nelson 2001; Zhao et al. 2010). Leydig cells synthesize testosterone with precursor cholesterol through a series of reactions catalyzed by enzymes such as CYP11A1, CYP17A1, and 3β-HSD. The mitochondrial enzyme CYP11A1 converts cholesterol to pregnenolone, the first rate-limiting step in steroid hormone production; CYP17A1 and 3β-HSD are involved in the conversion of many intermediates, especially 3β-HSD converting pregnenolone to progesterone in steroid hormone synthesis (Payne and Youngblood 1995). We explored alterations in CYP11A1, CYP17A1, and 3β-HSD enzymes, as well as StAR, which transport cholesterol to the inner mitochondrial membrane for steroidogenesis initiation, under CBG overexpression. Previous studies suggest that PFOA treatment dramatically damages steroidogenesis in Leydig cells (Zhang et al. 2014). Corresponding with the reduction in progesterone production, both CYP11A1 and CYP17A1 under 200 µM PFOA treatment for 48 h were significantly reduced. However, CBG overexpression increased the levels of all four proteins in steroidogenesis and partly rescued their levels under the damage caused by PFOA treatment. The effect of CBG on progesterone and steroidogenesis-related proteins was further confirmed using siRNA assay. The results implied that CBG can increase steroid hormone production in Leydig cells by enhancing protein levels involved in steroidogenesis.

LH increases cAMP through the G protein-mediated signaling pathway, which activates steroidogenesis in Leydig cells (Abdou et al. 2014). To further explore whether intrinsic CBG in Leydig cells responded to LH stimulation, we incubated the mLTC-1 cells with LH for 3 h, and found a significant increase in progesterone in the cells. Therefore, CBG responded to LH stimulation, and its induction by LH may contribute to steroidogenesis by LH stimulation. Furthermore, a significant increase in secretory CBG in the supernatants was also observed after LH stimulation, implying that the CBG produced in testicular Leydig cells not only played a cellular role in steroidogenesis, but also contributed to circulating CBG levels.

Although some companies aimed to phase out PFOA utilization, emission and production (Hoke et al. 2015), current PFOA is still one of the most dominant PFASs in general population (Averina et al. 2018). Reports supported that PFOA exposure decreased the proportion of morphologically normal spermatozoa in men (Joensen et al. 2009), and increased the risk of testicular cancer in humans (Barry et al. 2013). Mode of action for PFOA-induced testicular cancer in rodents suggested that hormone modulation, especially testosterone in Leydig cells, may play an important role in promotion of cell proliferation and carcinogenesis (Klaunig et al. 2012). The CBG modulation effect of PFOA observed in our study, as well as its potential effect on steroid hormone synthesis, supports the hormone interference mode of action for PFOA, and a long exposure of PFOA may further lead to testicular cancer.

Conclusions

In summary, we unambiguously demonstrated that testicular Leydig cells synthesized CBG, which could be induced by PFOA treatment. We also revealed the role of CBG in mLTC-1 Leydig cells, which promoted the levels of progesterone and steroidogenesis-related proteins, such as CYP11A1 and CYP17A1. In addition, the production of CBG responded to LH stimulation in mLTC-1 cells.