Cyanobacteria are cosmopolitan photosynthetic prokaryotes, also known as “blue-green algae”. In favorable conditions they can form dense blooms (water discoloration due to extended accumulation of biomass) [1
], most often in freshwater, but also in brackish and marine water environments [1
]. During such blooms, cyanobacteria may release a large variety of bioactive metabolites, including well-known cyanotoxins [1
]. Therefore, they may have negative effects on water safety and quality (drinking, bathing, fishing, and recreational uses) and result in harm to invertebrates and vertebrates including humans [2
]. Different types of cyanobacterial toxins are classified according to their toxicological target as hepatotoxins, neurotoxins, and dermatoxins. The most known and studied cyanotoxins are the microcystins (MCs), potent hepatotoxins produced by several cyanobacterial species, including Microcystis
]. MCs are cyclic heptapeptides [12
], which share a common core structure. Five of the seven amino acids are in general highly conserved while the other two are more variable. Thus, at least 279 different MC-congeners have been reported to date [14
]. This enormous structural diversity within the class makes MCs characterization challenging, especially from a toxicological point of view.
The so-called endocrine disrupting compounds (EDCs), or endocrine disruptors (EDs), represent another increasing environmental and health concern. EDCs are exogenous substances or mixtures named by their capability of interfering with normal endocrine (hormonal) system homeostasis. Such interferences can happen through several different mechanisms and by acting on a number of different targets of the endocrine system’s components (Figure 1
), thereby causing adverse health effects in an intact organism or its progeny, as well as in entire populations [15
Most studies have mainly focused on EDCs of anthropogenic origin without extensively considering those that are naturally-produced in the environment [17
]. However, it has been reported that cyanobacterial compounds (extracts from blooms and exudates) have ED activity, mainly through activation of the estrogen receptor [19
]. It is not entirely clear which of the different compounds present in cyanobacterial blooms exert estrogen activity. The net effect may be the result from the interaction of different cyanobacterial metabolites in the bloom, which constitute in themselves a complex mixture of known and unknown compounds, and/or other substances coexisting in the water [18
In general, available studies on MCs activity mainly focus on just one congener, microcystin-LR (MC-LR), because of its availability, high occurrence, and high hepatotoxicity. It is, however, well known that although one or two MC-variants (not necessarily including MC-LR) are generally dominant in a single cyanobacterial strain [5
], the majority of blooming cyanobacteria produce multiple MC-congeners [23
], in addition to other bioactive compounds.
Additionally, the few studies supporting the ED potential of MCs are based on MC-LR [24
]. Some other studies, with a slightly broader approach, have demonstrated ED activities in extracts of cultures from Microcystis
spp. (which are able to release MCs). However, they seem to contradict, or at least not confirm, the role of MCs in the observed ED activity as the extracts, but not pure MC-LR, had effect in the applied assays, including induction of vitellogenin in zebrafish larvae, an effect associated with exposure to compounds with estrogenic effects [27
]. The estrogenic effects are frequently a result of activation of the estrogen receptor (agonist) or inhibition of the activation of the androgen receptor (antagonist). Reporter gene assays (RGAs) are excellent tools to measure the agonist or antagonist activity at the receptor levels. Furthermore, Stepankova and collaborators (2011) highlighted that cytotoxicity and estrogenic effects from extracts of blooming cyanobacteria containing MCs were greater than what would be expected from the extracts of pure laboratory cultures. These studies indicate that cyanobacterial compounds other than MCs are at least partly involved in the reported estrogenic effects. Furthermore, the potential mechanisms of the estrogenic effects of the cyanobacterial metabolites were likewise not studied.
MCs have previously been reported to cause reproductive toxicity [29
]. Although reproduction is regulated through the endocrine system, it is debated if the reproductive effect of MC exposure is a direct effect on the endocrine system. Furthermore, there are indications that MCs interfere with the endocrine system acting on the hypothalamic–pituitary–thyroid axis in addition to the hypothalamic–pituitary–gonadal axis, which includes the estrogens (endogenous hormones related to reproduction) [32
Based on the previously reported induction of vitellogenin, a clear indication of an estrogenic or anti-androgenic effect, the present study investigated potential ED effects of cyanobacterial extracts from Microcystis and Planktothrix spp. strains and pure MC-LR, with a main focus on estrogenic activity. We used cell-based models to investigate a direct effect at the receptor level (receptors are natural targets for endogenous hormones). In addition, we used a human liver microsomes (HLM) model to investigate possible interferences with cytochrome P450-mediated 17β-estradiol (herein simply referred to as estradiol) biotransformation. Estradiol is an endogenous hormone belonging to the group of estrogens (steroid hormones). Thus, this study aimed to address the following research questions and null hypotheses:
Do cyanobacterial metabolites interfere with hormone receptors, especially with the estrogen receptor?
H1-0: Extracts from cyanobacterial cultures do not show effects on the estrogen receptor (or other hormone receptors).
Are there other modes of action that may result in endocrine disrupting activity of cyanobacterial metabolites, e.g., effects on cytochrome P450 mediated estradiol biotransformation?
H2-0: Estradiol biotransformation is not affected by the presence of cyanobacterial culture extracts or MCs.
If cyanobacterial metabolites show ED activity, are MCs responsible for any of the observed effects?
H3-0: MCs do not show ED activity.
The different in vitro assays were therefore chosen to take into account that there are a number of processes, receptor-mediated and non-receptor mediated, which could alter endocrine functions [34
Our screening of Microcystis
spp. culture extracts did not show substantial agonist or antagonist estrogen activity at the estrogen receptor level. Two strains showed statistically significant estrogen agonist activity when compared to Z8 medium. Furthermore, two strains showed a weak but statistically significant enhancement of the estradiol response, but we did not investigate further if this was due to synergism. None of these strains produced MCs, showing that the small but significant activity was not related to these toxins. We did not measure biomass, cell count, or the optical density of the cultures prior to extraction, but acknowledge that such measurements should have been performed. However, with respect to literature data investigating estrogen activity of MCs or Microcystis
spp., we expected that extracts from visually dense cultures of the cyanobacteria would give a clear response in the estrogen-responsive RGAs provided that there are cyanobacterial metabolites or constituents that induce effects at the receptor level. Other limitations of the RGA-based screening approach were that the production of cyanobacterial metabolites might change over time depending on the growth stage, and metabolite production is also expected to be different in natural environments compared to laboratory conditions [38
]. The reason for the observed agonist activity of the cyanobacterial culture medium Z8 is unclear as it does not contain any chemicals that are expected to possess estrogenic activities [42
]. We believe that the origin of its weak biological activity could be due to contaminants from equipment used during the production of the medium, e.g., plastics, but did not investigate this further. However, it emphasizes that the activity of metabolites from the studied cyanobacterial strains on the estrogen receptor was minor. We did thus not aim to trace the relatively low activities in the estrogen responsive RGAs to specific cyanobacterial constituents.
While the study of estrogenic effects of cyanobacteria has been the subject of several reports, androgen- and glucocorticoid-type effects have not been studied in depth so far. However, the limited data did not support such effects from cyanobacterial extracts [19
]. We only included androgen- and glucocorticoid-responsive RGAs in preliminary screenings that need further verification. However, the limited data indicated that the cyanobacterial extracts included in this study might induce androgen and glucocorticoid antagonist effects. The replicates variability in these initial screening was high and underlines the need for verification (Tables S1–S3
MCs and extracts of Microcystis
cultures reportedly reduce the reproduction in experimental fish [24
]. The exact mechanism behind the effects on reproduction is not known and may involve effects on several endocrine pathways. Earlier work found that lyophilized M. aeruginosa
but not pure MC-LR increased the expression of vitellogenin genes in zebrafish larvae [28
], frequently interpreted as an indication of exposure to estrogenic compounds. Because of the lack of a clear estrogen activity at the receptor level, we investigated another mechanism that may lead to estrogenic effects, i.e., the modulation of hormone metabolism and biotransformation. Since we focused on effects of the cyanobacterial extracts on the estrogen receptor, it was natural to investigate the effect of MCs and a cyanobacterial extract on the microsomal biotransformation of estradiol. Since most toxicological studies have been carried out using the major MC-congener MC-LR, we also included this compound in our microsome trials, and co-incubated estradiol with 0.5 µM or 5 µM of the toxin. We also co-incubated estradiol with extract from M. aeruginosa
strain PCC7806, and adjusted the extract concentration so that the total MCs assay concentration was about 0.5 µM. The PCC7806 strain has previously been shown to induce ED activity and produces MC-LR as principal MC congener. However, it is unknown if the effects were related to the MCs it produces, or other metabolites, but the data indicated that the ED activity was caused by metabolites other than MC-LR [28
]. Although estradiol depletion was little affected by co-incubation with MC-LR or M. aeruginosa
PCC7806, both the pure toxin and the cyanobacterial extract modulated the production of oxidized metabolites similarly (Table 2
). Thus, the main product from microsomal cytochrome P450-mediated biotransformation, 2-hydroxyestradiol, was markedly reduced when the liver microsomes were co-incubated with either MC-LR or M. aeruginosa
PCC7806. This effect was accompanied by an increase in relative estrone concentrations, and therefore overall estradiol depletion remained unaffected.
A large number of estradiol-metabolites have been shown, and they have different estrogenic potency [52
]. The main hepatic phase-I metabolite 2-hydroxyestradiol has a considerably lower estrogenic potential than estradiol, estrone, 4-hydroxyestradiol, and many other tested estrogen metabolites [52
]. Our findings indicate that MC-LR or extracts of a Microcystis
culture may alter the metabolic transformation of estradiol to other metabolites with different estrogenic potential. It is therefore likely that a decreased formation of 2-hydroxyestradiol could lead to increased occurrence of estradiol or other metabolites with a higher estrogenic potency and hence induce an overall estrogenic effect. The degree of modulation of estradiol biotransformation was similar for co-incubation with 0.5 µM and 5 µM MC-LR (Table 2
). However, at present we do not know the underlying mechanism(s) for the apparent modulation. While 2-hydroxylation is one of the principal microsomal biotransformation pathways for estradiol and catalyzed by CYP1A2 and CYP3A4, several other microsomal biotransformation products have been shown [46
]. Future studies should therefore target as many biotransformation products of estradiol as possible, and should also investigate the possible interaction with relevant CYPs in detail. MC-LR and Microcystis
extracts may also affect other biotransformation enzymes like sulphotransferase and steroid sulfatase. These enzymes are key enzymes in the regulation of the equilibrium between the free active form of estradiol and the much less active conjugates. MCs also affect the hypothalamic-pituitary-thyroid axis and the hypothalamic-pituitary-gonad axis in zebrafish [32
]. Alterations in these pathways may interfere with normal homeostasis and hormone levels and hence explain the observed hormonal effects. Even the well-known inhibition of protein phosphatase PP2A [56
] may affect the reproduction as the balance of phosphorylation/dephosphorylation of proteins are crucial for the regulation of intracellular signaling pathways and alterations may affect the development of gametocytes.
5. Materials and Methods
5.1. Chemicals and Reagents
Extraction of cultures: Methanol (MeOH) (gradient quality) was from Romil (Cambridge, UK).
RGAs: Cell culture reagents were supplied by Life Technologies (Paisley, UK). Standards used for RGAs (estradiol, testosterone, progesterone), phosphate buffered saline (PBS), MeOH, dimethyl sulfoxide (DMSO) and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) were purchased from Sigma-Aldrich (Poole, Dorset, UK). Lysis reagents and luciferase assay system were obtained from Promega (Southampton, UK).
Estradiol biotransformation: NADP+
, NADPH, D-glucose 6-phosphate sodium salt, D-glucose-6-phosphate dehydrogenase from baker’s yeast (Saccharomyces cerevisiae), HEPES buffer and ammonium formate were purchased from Sigma-Aldrich Norway (Merck Life Science AS, Oslo, Norway). Estradiol, 2-hydroxyestradiol, 4-hydroxyestradiol, estrone and estriol were purchased from Sigma-Aldrich (Millipore-Sigma; St. Louis, MO, USA). All stock solutions were prepared in MeOH at 1 mg/mL (estradiol, estrone, and estriol) or 0.5 mg/mL (2-hydroxyestradiol and 4-hydroxyestradiol). For estradiol, the working solution at 0.25 µg/mL was prepared by appropriate dilution from stock with MeOH. Stock solutions of 2-hydroxyestradiol, 4-hydroxyestradiol, estrone, and estriol were combined for the preparation of the mixed solution containing 2 µg/mL of each compound in 50% acetonitrile (MeCN). The mixture was used to prepare calibration standards in 50% MeCN at 12.5 ng/mL, 15 ng/mL, 75 ng/mL, 150 ng/mL, and 300 ng/mL for quantitative analysis (Table 3
Inorganic chemicals and organic solvents were reagent grade or better. MC-LR for the multihapten ELISA was from Enzo Life Sciences Inc. (Farmingdale, NY, USA). For further details refer to Samdal et al. (2014) [60
5.2. Cultivation of Cyanobacterial Strains and Extraction
Twenty-five cyanobacterial strains (19 strains from Microcystis
spp. and 6 strains from Planktothrix
spp.) were purchased from The Norwegian Culture Collection of Algae, NORCCA, jointly maintained and owned by the Norwegian Institute for Water Research (NIVA) and the University of Oslo (Table S1
). The M. aeruginosa
PCC7806 strain was purchased from the Pasteur Institute (Paris, France) (Table S1
). All strains were cultivated in Z8 medium [42
] in 100 mL glass Erlenmeyer flasks in an incubator (IPP110plus, Memmert GmbH + Co.KG, Schwabach, Germany) at 18 °C with a 14/10 h light/dark photoperiod, using 1% of maximum light intensity.
For extract preparation, 3 mL of a cyanobacterial culture was transferred to a glass tube and stored at −20 °C overnight, then allowed to thaw at room temperature, and 3 mL of MeOH was added. The tube was then vortex-mixed for 20 s, sonicated for 5 min and centrifuged for 10 min at 1,000 × g. The supernatant was transferred to screw-cap liquid-chromatography (LC) glass vials and stored at −20 °C until use. For concentration of extracts, 1ml aliquots were dried under a gentle stream of nitrogen and re-dissolved in 50 µL of 50%MeOH in screw-cap LC glass vials and stored at −20 °C until use.
5.3. Microcystin ELISA
MCs production was assessed using an indirect competitive ELISA as described by Samdal et al. (2014) [60
]. All samples and standard dilutions were done in duplicate and performed at ~20 °C. Absorbances were measured at 450 nm using a SpectraMax i3x plate reader (Molecular Devices, Sunnyvale, USA). For further details refer to Samdal et al. (2014) [60
5.4. Cell Culture
Three reporter gene assay (RGA) cell lines were cultivated and used: the MMV-Luc (estrogen responsive), TARM-Luc (androgen responsive), and TGRM-Luc (glucocorticoid responsive). They were previously developed by transforming human mammary gland cell lines using the luciferase gene under the control of a steroid hormone inducible promoter, as reported by Willemsen et al. (2004) [61
]. The cells were routinely grown in 75 cm2
tissue culture flasks (Nunc, Roskilde, Denmark), incubated at 37 °C in an atmosphere containing 5% CO2
and at 95% humidity. The MMV-Luc cell culture medium was made up of: Dulbecco’s Modified Eagle Medium (DMEM), supplemented with 10% fetal bovine serum (FBS) and 1% L-glutamine (2 mM). The TARM-Luc and TGRM-Luc cell culture medium was made up of DMEM Glutamax, supplemented with 10% FBS. The MMV-Luc culture medium was free from phenol red because of its weak estrogenic activity that may interfere with the assay.
5.5. Reporter-Gene Assays (RGAs)
Reporter gene assays (RGAs) are produced by transfecting cell lines with relevant receptors. A transactivation step with a signaling protein (the luciferase) is included. This step allows the measurement of receptor activation, distinguishing at the same time between agonist and antagonist behaviour [61
]. The RGA procedures have previously been described by Frizzell et al. (2011) [62
]. Briefly, cells were seeded at a concentration of 4 × 105
cells/mL, 100 mL/well in white walled, clear, and flat bottomed 96-well plates (Greiner Bio-One, Fricken-hausen, Germany). RGA medium differed from that used for maintaining cell cultures in that hormone depleted serum was used instead of common FBS. After 24 h pre-incubation, cyanobacterial extract solutions as well as steroid hormone standards were added in triplicate to the cells at a final methanol concentration of 0.1%. Positive controls for the estrogen-, androgen-, and glucocorticoid-responsive RGAs were estradiol (1.36 ng/mL or 5 nM), testosterone (14.5 ng/mL) and cortisol (181 ng/mL), respectively. A solvent control (0.25% MeOH in deionized water, v/v) was also included. Antagonist tests included co-incubation of test compounds with the positive controls. The cells were incubated for 48 h. Supernatants were then discarded, and the cells were washed once with PBS. Then, 25 µL cell lysis buffer was added to each well. Reading was performed by adding 100 µL of luciferase substrate to each well and measuring luciferase activity with a Mithras Multimode Reader (Berthold, Other, Germany). Agonist response was measured relative to the solvent or cyanobacterial growth medium, and antagonist response was measured relative to the positive controls.
Estrogen-responsive RGAs and connected MTT cell viability assays were performed in triplicate and repeated in three independent exposures, while androgen-responsive and glucocorticoid-responsive RGAs and connected MTT cell viability assays were performed in triplicate, but in only one exposure each.
5.6. MTT Cell Viability Assay
Flat-bottomed 96-well plates (Nunc) were seeded with 4×105 cells/well of the appropriate RGA cell line. Cyanobacterial extracts were added to the cells after a pre-incubation period of 24 h at the desired concentration and maintaining a solvent concentration of 0.1% MeOH. Controls and extracts were incubated for 48 h. Supernatants were then discarded, and 50 µL of MTT solution (MTT stock solution/assay media ratio 1:6) added to each well and the cells incubated for another 3 h. Viable cells would convert the soluble and yellow dye MTT into insoluble and purple formazan. Supernatant were removed and 200 µL of DMSO added to each well to solubilize the formazan crystals. Optical density was measured at 570 nm (630 nm as reference) using a Sunrise spectrophotometer (TECAN, Switzerland). Cell viability was expressed as the relative absorbance of samples compared to the solvent control.
5.7. Human Liver Microsome (HLM) Incubation
In vitro metabolism trials were performed using commercially available human liver microsomes (HLM; No. X008067, Lot YAO; Celsis In Vitro Technologies, Baltimore, MD, USA). Estradiol (5 µM assay concentration) was incubated either alone or in presence of MC-LR solutions (0.5 µM and 5 µM assay concentration), as well as in presence of a M. aeruginosa PCC7806 extract. The extract was concentrated to obtain an in-assay concentration of MCs (sum of MC-LR and [D-Asp3]MC-LR) of approximately 0.5 µM. Individual incubations were carried out with 2 mg/mL of microsomal protein in a final volume of 1 mL containing a NADPH-generating system (0.91 mM NADPH, 0.83 mM NADP+, 19.4 mM glucose-6-phosphate, 1 U/mL glucose-6-phosphate dehydrogenase and 9 mM magnesium chloride hexahydrate) as well as an incubation buffer (45 mM Hepes, pH 7.4) at 37 °C for 60 min. Enzymatic reactions were stopped by adding aliquots from the individual incubations to ice-cold MeCN (1:1). All samples were vortex-mixed and placed on ice. They were centrifuged at 2,000 × g (Eppendorf, Hamburg, Germany) to precipitate proteins, filtered (0.22 µm Corning Costar Spin-X centrifuge tube filters, Corning Inc., NY, USA) and transferred to chromatography vials.
5.8. Quantification of Estradiol and Biotransformation Products Using LC–MS/MS
Estradiol and estradiol-related oxidative metabolites were determined by using an Agilent 1290 Infinity Binary UHPLC System with vacuum degasser and thermostatted column compartment (30 °C) coupled via an electrospray ionisation interface (ESI) to an Agilent 6470 TripleQ mass spectrometer (Agilent Technologies, Santa Clara, CA, USA). The instrument’s MassHunter Optimizer software (version B 8.0; Agilent) was used to evaluate the fragmentation patterns of the target compounds. The precursor-ion to product-ion transition, mass-to-charge ratio (m/z
), fragmentor voltage (FV), collision energy (CE), and cell accelerator voltage (CAV) for multiple reaction monitoring (MRM) of all analytes (Table 3
) were optimized by infusing a solution of the compounds isocratically with a flow rate of 0.4 mL/min (MeCN:water, 1/1, both containing 0.1% formic acid). Two transitions were monitored for each compound using dynamic multiple reaction monitoring (dMRM) according to Table 3
The ES interface was operated in positive ion mode including a drying gas temperature of 300 °C, a drying gas flow of 10 L/min, a nebulizer gas pressure of 25 psi, a sheath gas temperature of 375 °C, a sheath gas flow of 12 L/min, a nozzle voltage of 1000 V, and a capillary voltage of 4 kV. A novel separation method was developed using a 150 × 2.1 mm i.d. Kinetex F5 (2.6 μm particles) HPLC column including a 0.5 μm KrudKatcher Ultra pre-column filter (Phenomenex). The target compounds were eluted using a mobile phase gradient consisting of water (A) and MeCN (B), both containing 0.1% formic acid. Chromatographic separation was achieved using isocratic elution at 2% B for 6 min followed by a linear gradient to 35% B over 2 min, and followed by a second linear gradient to 48% B over 10 min. After flushing the column for 3 min with 95% B, the mobile phase was returned to the initial conditions, and the column was eluted isocratically for an additional 2 min. The total run time was 24 min. The flow rate was 0.25 mL/min, the injection volume was 1 µL and the column temperature was maintained at 30 °C.
The biotransformation of estradiol to 2-hydroxyestradiol and estrone in the different groups (i.e., 0.5 µM MC-LR + 5 µM estradiol, 5 µM MC-LR + 5 µM estradiol or M. aeruginosa
PCC7806 + 5 µM estradiol) was compared to the reference group (5 µM estradiol) using One-Way Analysis of Variance (ANOVA) followed by Dunnett’s test (JMP 14.0, SAS Institute Inc., Cary, NC, USA) (Figure S2, Table S4