Melatonin modulates the gene expression of WEE1 kinase and clock genes: a crosstalk between the molecular clocks of the placenta?

The bioinformatics analysis was designed to identify time-dependent placental DEGs enriched for circadian and cell cycle pathways across independent GEO datasets. We analyzed datasets from the Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo↗) similar to what was reported previously (69, 70) for the terms "placenta and clock", "placenta and circadian", and "trophoblast and culture" (n = 139). We excluded platform data without temporal samples or incomplete incoming data. "GSE86171", "GSE60433", and "GSE40182" include temporal samples between 0 and 48 hours that were visualized using GEO Profiles graphics and the parameter Benjamin and Hochberg false discovery rate methodology with significance thresholds set at log2 fold change (logFC) ≥1 and adjusted p-value <0.05. We utilized the Kyoto Encyclopedia of Genes and Genomes (KEGG) for the functional analysis of cell cycle and circadian rhythms. In the present study, we collected, combined, and identified the gene expression profile using a Venn diagram. p < 0.05 was considered a significant difference by employing DAVID Bioinformatics 6.8, released Oct. 2016. The GO terms were "circadian rhythms", "circadian regulation of gene expression", "regulation of circadian rhythm", "entrainment of the circadian clock by photoperiod", and "cell cycle".

Term placentas from uncomplicated vaginal deliveries were obtained at approximately 07:00 hours at Herminda Martín Hospital (Chillán, Chile) after written informed consent was provided. The Ethics Committee approved the protocols of the Hospital and the University of the Bío Bío. Placentas were maintained at 4°C and processed at 07:00 hours. The tissue was washed three times with ice-cold phosphate-buffered saline (PBS) to eliminate red blood cells and trimmed to obtain a fetal portion of the placenta (chorion). Fifty-four explants of approximately 2 mm (L) × 2 mm (W) × 2 mm (H) and a mass of 45 ± 0.841 mg (wet mass) were used, according to the protocol of Cemerikic et al. (71).

Explants were cultured individually following previously described protocols (71 –73). They were preincubated in M-199 medium (pH 7.2) and maintained in a humidified environment at 37°C and 5% CO₂ for 4 hours. Then, they were transferred to fresh medium either alone (control) or supplemented with 10 nM melatonin (treatment group). The concentration of 10 nM melatonin was selected as a physiologically relevant dose, within the range used in previous studies on peripheral tissues (e.g., 10–100 nM) (74 –79). This lower concentration was chosen to avoid potential pharmacological effects while maintaining biological activity. A sampling of three explants and the supernatant was conducted every 4 hours. All explants were weighed and stored with 1 mL TRIzol reagent (Invitrogen, Invitrogen Corporation, Carlsbad, California, USA). Explants and supernatant were stored frozen at −20°C.

Explants of the human placenta were extracted in two stages: i) by the TRIzol method modified following the manufacturer's instructions (80) (phase separation, precipitation, and washing RNA) to the ethanol phase and later and ii) extraction using Kit SV Total RNA Isolation System modified following the instructions of the manufacturer (Promega Corporation, Madison, Wisconsin, USA) (purification of RNA). The absorbance was measured at 260 and 280 nm using a spectrophotometer to determine the concentration of RNA. Reverse transcription of 20 ng of extracted RNA was performed using the Improm Kit II Reverse Transcription System (Promega, Promega Corporation, Madison, Wisconsin, USA) in a final volume of 20 µL. The reverse transcription was at 70 °C for 5 minutes, 4°C for 5 minutes, 25°C for 5 minutes, 42°C for 60 minutes, and 70°C for 15 minutes.

The relative expression of the mRNAs of clock genes BMAL1, PER1-2, and WEE1 was measured in samples of total cDNA. The PCR was performed in a final volume of 10 µL containing 0.33 µL of primers, forward and reverse primers of the genes studied, 3.8 µL of nuclease-free H₂O, and 5.5 µL of Master Mix II SYBR Brilliant Green (Agilent Technologies, Santa Clara, California, USA). The following primers were used: BMAL1, forward, 5′-CTGCATCCTAAAGATATTGCCAAAG-3′, and reverse, 5′-GTCGTGCTCCAGAACATAATCG-3′; PER1, forward, 5′-GGGCAAGGACTCAGAAAGAA-3′, and reverse, 5′-AGGCTCCATTGCTGGTAGAA-3′; PER2, forward, 5′-TGGATGAAAGGGCGGTCCCT-3′, and reverse, 5′-ACTGCAGGATCTTTTTGTGGA-3′; WEE1, forward, 5′-CGCGATGAGCTTCCTGAGCCG-3′, and reverse, 5′-CAGCGCACCGGCGAGAAAGAG-3′; cyclophilin, forward, CTCCTTTGAGCTGTTTGCAG-3′, and reverse, 5′-CACCACATGCTTGCCATCC-3′. For expression from quantitative real-time PCR (qPCR) data, all expression was normalized with cyclophilin for calculating relative gene expression by double delta Ct (ΔΔCt) and transformed to 2^−ΔΔCt.

The supernatant was cleaned with activated charcoal and measured by spectroscopic imaging using Fourier transform infrared (FTIR) spectroscopy associated with Attenuated Total Reflectance (ATR) (ATR–FTIR). Spectral measurements of the melatonin standard curve at 0.3–3,000 nM (Sigma-Aldrich, St. Louis, Missouri, USA.) were conducted, and supernatant samples were measured in triplicate using ATR–FTIR. The sample spectrum of 10 μL was recorded at room temperature in the region 1,000–4,000 cm⁻¹ directly on a JASCO FT/IR-4100 Fourier transform infrared spectrophotometer with a 4.0 cm⁻¹ resolution. A linear relationship was found for melatonin measurement at 1,492 cm⁻¹. The melatonin content was calculated following the methodology described by Filali et al. (81). The inter-assay and intra‐assay coefficients of variation were less than 18%. Endogenous melatonin was quantified in explants maintained without supplementation. Exogenous melatonin levels were evaluated in explants supplemented with 10 nM melatonin. Paired untreated controls and the standard curve were used to differentiate between the hormone secreted by the tissue and the exogenous melatonin added to the medium.

Data were expressed as mean ± SEM and analyzed using repeated-measures ANOVA, followed by Newman–Keuls post-hoc test, or Student's t-test as appropriate. Rhythmicity in gene expression was evaluated using non-linear regression of the sine-wave function expressed as Y = Baseline + Amplitude * Sine (Frequency X + Phase shift). All data were normalized between 0 and 1; the data were analyzed using the GraphPad Prism 5 software, and p < 0.05 was considered statistically significant.

To explore whether circadian and cell cycle pathways were consistently represented in placental gene expression, we first analyzed publicly available transcriptomic datasets (GEO). We first asked whether time of day-dependent transcriptional changes in placental tissue preferentially involve circadian and cell cycle pathways across independent datasets. The expression profiling dataset of mRNA (GEO database) gives the tools for bioinformatics analysis of molecular pathways modified by time hours in the placenta. We performed the identification of DEGs via GO term enrichment and functional classification using DAVID. We selected the GO classification related to the circadian system and cell cycle. We used terms such as "cell cycle", "circadian rhythms", "circadian regulation of gene expression", "regulation of circadian rhythm", and "entrainment of the circadian clock by photoperiod". We detected three complete microarray experiments for analysis: "GSE86171", "GSE60433", and "GSE40182". The periods studied in the microarrays were 0, 3, 12, 24, and 48 hours.

Functional annotation using DAVID identified 391 common DEGs (60.9%) during all time hours studied. Functional enrichment analysis (DAVID/KEGG) of the 391 common DEGs revealed significant overrepresentation of the "circadian rhythm" and "cell cycle" pathways (adjusted p < 0.05). Additional enriched terms included apoptosis and DNA repair, consistent with the central role of circadian regulation in cell proliferation and survival. Figure 1 shows the Venn diagram demonstrating the intersections of genes at different times of the day. Approximately 643, 283, and 1,179 common genes changed their expression level over all the time hours studied. Volcano plots for each dataset (Figures 2A–C) display the distribution of DEGs over time, and the pattern is visualized at every time studied in "GSE86171", "GSE60433", and "GSE40182". Similarly, the data of clock genes and regulators of the cell cycle for log2(fold change) and −log10(p-value) are shown in Table 1 for every time hour. The relative expression values suggest the time variation of clock gene expression in the placenta for at least 24 hours, with a peak for PER2 and CRY1 during the first half of the day. BMAL1 shows a peak early in the morning, and the cell cycle genes TP53, CIPC, and WEE1 show a peak during the interval between early in the morning and noon, suggesting a temporal variation of genes of circadian and cell cycle clocks.

We next examined whether placental explants maintained circadian oscillations of core clock genes and the cell cycle regulator WEE1 in culture. To validate the in silico observations associated with temporal variations observed in the microarray of the placenta, we cultured human placental explants and measured gene expression every 4 hours for 36 hours. We observed that the BMAL1, PER1, PER2, and WEE1 genes maintain their mRNA expression in the culture of the human placenta for at least 36 hours (Figure 3).

As shown in Figure 3A, clock gene BMAL1 expression increases during daylight hours, showing a rise between 03:00 and 11:00 hours. Also, BMAL1 showed a local peak at 11:00 hours (range 03:00–11:00 hours is different from 15:00–23:00 hours of the second day of culture; p < 0.05, one-way ANOVA and Newman–Keuls post-test), whereas this expression showed a local minimum at 23:00 hours. The relative mRNA expression of PER1 showed no significant changes during the hours studied but exhibited a trend toward higher expression in the evening (Figure 3B).

PER2 expression changed during the hours of culture, showing a peak expression at 19:00 hours on the first day (p < 0.05 ANOVA and Newman–Keuls) and low expression levels in the following hours studied (Figure 3C). Wee1 expression showed no significant differences but trended upward during nighttime hours (Figure 3D).

The temporal data suggest an endogenous oscillation in BMAL1 and PER2 occurring in antiphase with a ~12-hour interval, indicative of a functional circadian clock in placental tissue.

Given that the placenta expresses the capacity to synthesize melatonin and receptors, we tested whether exogenous melatonin modulates the oscillations of BMAL1, PER2, and WEE1. The exposure of placental explants to 10 nM melatonin suppressed the rhythmic peaks of BMAL1 and PER2 expression observed in untreated cultures. Although PER1 and WEE1 did not show statistically significant changes, BMAL1 expression was reduced between 07:00 and 19:00 hours under melatonin treatment (Figures 4A–D). These results suggest that exogenous melatonin can reduce the amplitude of circadian gene expression in placental tissue.

To further capture phase relationships among clock genes, we calculated BMAL1/PER1 and BMAL1/PER2 ratios across timepoints. To further evaluate gene oscillations, we calculated the expression ratios BMAL1/PER1 and BMAL1/PER2 (Figure 5). The circadian oscillation circuits are dependent on the transcriptional/translational feedback loop of clock genes, which act as positive and negative regulators, inducing/inhibiting their expression. BMAL1/PER1 ratios showed non-significant variation but tended to peak at 15:00 and 23:00 hours on the second day of incubation (Figure 5A).

Moreover, BMAL1/PER2 ratios exhibited significant oscillation, peaking between 03:00 and 11:00 hours and declining between 15:00 and 23:00 hours (p < 0.05; Figure 5B), fitting a sine-wave function (r² = 0.7368). In contrast, melatonin treatment inhibited the BMAL1/PER2 peaks (Figures 5C, D).

These findings support the existence of an antiphase rhythm between BMAL1 and PER2, a circadian pattern that is disrupted by melatonin.

To assess circadian gating of the cell cycle, we analyzed the ratio of BMAL1 to WEE1 expression across the culture period. The BMAL1/WEE1 expression ratio revealed a peak at 03:00–11:00 hours, followed by a decline during the night hours of the second day of culture (p < 0.05; Figure 6A). This antiphase relationship between BMAL1 and WEE1 was lost in melatonin-treated explants (Figure 6B). The pattern is consistent with the transcriptional regulation of WEE1 by the CLOCK/BMAL1 complex.

Finally, it was assessed whether placental explants produce melatonin endogenously and whether supplementation alters secretion levels. Endogenous melatonin was quantified in supernatants from untreated cultures (Figure 7A), while apparent exogenous levels were assessed in melatonin-supplemented cultures (Figure 7B). Values in treated conditions were interpreted relative to the standard curve and to paired untreated controls to distinguish secretion from supplemented hormone.

These findings confirm that the human placenta can produce melatonin autonomously and suggest a regulatory feedback loop between melatonin and the placental clock.