The epithelial potassium channel Kir7.1 is stimulated by progesterone

Situated in the ventricles of the brain, the choroid plexus (CP) constitutes an important barrier between the peripheral blood plasma and the brain. The CP consists of a layer of cuboidal epithelial cells that are responsible for producing cerebrospinal fluid (CSF; Damkier et al., 2013). In addition to its function as a mechanical buffer, secretion of CSF also provides the nutrients and hormones needed for neurogenesis while, at the same time, removing waste products from the brain (Lun et al., 2015). Tight junctions in the CP epithelium form a barrier between the blood plasma and the CSF, which allow channels and transporters in the CP to establish a highly regulated concentration gradient of ions between the two fluids (Damkier et al., 2013). Therefore, identification of the molecular regulators of CP ion channels is essential to better understand the mechanism behind the healthy ion homeostasis and CP function. One such potential regulator is the steroid hormone progesterone (P4), the levels of which were shown to correlate with ventricular size during the female menstrual cycle (Hagemann et al., 2011) and pregnancy (Oatridge et al., 2002). Interestingly, the expression of the conventional genomic progesterone receptor in CP, as well as other steroid receptors (Table 1), is almost negligible, while the tissue retains a response to P4-induced regulation (Kelley and Mermelstein, 2011; Luoma et al., 2011). Additionally, P4 has been shown to exert a neuroprotective role during brain trauma (Pettus et al., 2005; Stein and Wright, 2010; Deutsch et al., 2013). Intramuscular or intravenous injections of micromolar concentrations of P4 have been shown to alleviate traumatic brain injury (TBI) outcomes and decrease brain edema (Pan et al., 2019), yet the molecular mechanism behind this effect is still unknown. Since P4 is known to regulate cationic channels through nongenomic steroid signaling (Lösel and Wehling, 2003; Miller et al., 2016), we explored whether P4 could modulate CP ion channels in a similar manner. To determine nongenomic signaling events, direct patch-clamp experiments were performed on primary cultures of murine CP cells isolated from the lateral ventricles. These experiments revealed that the inwardly rectifying potassium channel Kir7.1, localized to the luminal side of the CP epithelium, is the major ion channel in CP potentiated by P4 via a G protein–coupled receptor-independent mechanism. Kir7.1 is vital for mammalian physiology, as mice deficient in Kir7.1 die shortly after birth (Yin et al., 2018). The potentiation of Kir7.1 by P4 was recapitulated using recombinantly expressed Kir7.1 in HEK293 cells. Similarly, we found that P4 potentiated the endogenous Kir7.1 expressed in the retinal pigment epithelium (RPE; Shimura et al., 2001), another epithelial tissue with similar ion channel expression as the CP. Our results point to a novel unexpected regulation of Kir7.1 by the steroid hormone P4 in two tissues, CP and RPE.

The C57BL/6N (Charles River) mice and the Abhd2 knockout mice of similar background were kept at the Animal Facility of the University of California, Berkeley, in a room with controlled light (14 h light, 10 h darkness) and temperature (23 ± 0.5°C). The mice were fed a standard chow diet (LabDiet 5053; PicoLab Rodent diet 20) and hyperchlorinated water ad libitum. All experiments were performed in accordance with National Institutes of Health guidelines for animal research and approved by the University of California, Berkeley Animal Care and Use Committee (AUP 2015–07-7742), with every effort made to minimize suffering for the animals. A thorough description of how the Abhd2 full knockout mouse line was generated is provided elsewhere (Björkgren et al., 2019 Preprint).

For gene expression analyses of the CP, the lateral CPs from three adult male and three adult female of either Abhd2^+/+ or Abhd2^−/− genotype were dissected out and snap frozen. The animals were euthanized by CO₂ followed by cervical dislocation. This procedure was performed in the morning when the females were in proestrus. mRNA isolation was performed using the Quick-RNA Microprep Kit (Zymo Research; R1050) according to the manufacturer’s instructions. The mRNA samples were processed for sequencing at QB3 Vincent J. Coates Genomics Sequencing Laboratory, University of California, Berkeley. CP mRNA was enriched for using Oligo dT beads from the Invitrogen Dynabeads mRNA Direct kit (61005). Subsequent library preparation steps of fragmentation, adapter ligation, and cDNA synthesis were done on the enriched mRNA using the KAPA RNA HyperPrep kit (Roche; KK8540). Truncated universal stub adapters were used for ligation, and indexed primers were used during PCR amplification to complete the adapters and to enrich the libraries for adapter-ligated fragments. Samples were checked for quality on an AATI (now Agilent) Fragment Analyzer. Libraries were then quantified using qPCR on a Bio-Rad CFX Connect to measure molarity of only sequenceable molecules using the Kapa Biosystems Universal Illumina quant kit (KK4824), pooled equimolar, clustered at 3 nM with 2% PhiX v3 control library spike-in, and sequenced as a 100-bp paired-end run on an Illumina HiSeq 4000. BCL files were then converted to fastq format and demultiplexed using the innate bcl2fastq v1.19 software.

The sequencing results were analyzed by Novogene. Reference genome and gene model annotation files were downloaded from the genome website browser (NCBI/UCSC/Ensembl) directly. Indexes of the reference genome was built using STAR and paired-end clean reads were aligned to the reference genome using STAR (v2.5) and the method of maximal mappable prefix. HTSeq v0.6.1 was used to count the read numbers mapped of each gene. Thereafter, fragments per kilobase of exon model per million reads mapped (FPKM) of each gene was calculated based on the length of the gene and reads count mapped to this gene (Mortazavi et al., 2008). Differential expression analysis between the four groups (Abhd2^+/+ and Abhd2^−/− male versus Abhd2^+/+ and Abhd2^−/− female, three biological replicates per condition) was performed using the DESeq2 R package (2_1.6.3). The resulting P values were adjusted using Benjamini and Hochberg’s approach for controlling the false discovery rate. Genes with an adjusted P value < 0.05, found by DESeq2, were assigned as differentially expressed. The data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus (Edgar et al., 2002) and are accessible through GEO Series accession number GSE181594.

The lateral CPs from two animals of similar age were used for each cell culture. Animals were euthanized by CO₂ followed by cervical dislocation, and the sheets of CP epithelial cells (CPECs) were dissected out from the lateral ventricles into 37°C growth media (high-glucose Dulbecco’s modified Eagle’s medium with L-glutamine [Thermo Fisher Scientific; 11965092], 10% FBS [X&Y Cell Culture; FBS-500-HI], and 1% Penstrep [Thermo Fisher Scientific; 15140–122]), after which the CPs were transferred into 37°C Dulbecco’s PBS (DPBS) without Ca⁺ and Mg⁺ (Thermo Fisher Scientific; 14190144). Enzymatic dissociation of cells from the CP sheets was performed as previously described (Johnson et al., 2018), with some minor modifications. The CP sheets were broken up by incubation with Pronase (0.1 mg/ml in DPBS; Sigma; 10165921001) at 37°C for 30 min. The reaction was inhibited by adding a 5× volume of DPBS after which the cells were allowed to sediment. The cell pellet was further dissociated by trituration in 37°C TrypLE Express (Thermo Fisher Scientific; 12605010). After the cell clusters had sedimented, the supernatant was transferred to a 5× volume of fresh growth media. The trituration was repeated five times until only minor cell aggregates remained. The CPECs were washed in growth media by centrifugation at 800 ×g for 5 min, resuspended in fresh growth media, and plated on poly-D-lysine and laminin (Cell Applications; 127)–coated 5-mm glass coverslip in a 96-well plate. The cells were allowed to attach to the coverslip for 3 h at 37°C and 5% CO₂ before electrophysiology measurements.

RPE cells from two animals of similar age were used for each cell culture. The eyes were processed as previously described (Shang et al., 2018). In brief, the eyes were isolated and incubated in 2% Dispase II (Sigma; D4693) in high-glucose Dulbecco’s modified Eagle’s medium, without glutamine or pyruvate (Thermo Fisher Scientific; 11960044), for 45 min at 37°C and 5% CO₂ to allow the eye tissue to soften. Afterward, the cornea, iris, and lens capsule were removed, and the remaining eye tissue was incubated in RPE culture media (10% Charcoal-Stripped FBS [R&D Systems; S11650], 1% Penstrep, 2.5 mM L-glutamine [GIBCO BRL; 25030–81], and 1X MEM nonessential amino acids [Thermo Fisher Scientific; 11140050]) for 20 min at 37°C and 5% CO₂ to allow the neural retina to detach from the RPE. The RPE sheets were then manually dissected out and placed in an Eppendorf tube where the cells were dissociated from each other by pipetting up and down ∼40 times. The single-cell suspension was plated on laminin (GIBCO BRL; 23017–015)–coated 6.5-mm, 0.4-µm-pore Polyester Transwell membranes (Costar; 3470) in RPE culture media and incubated at 37°C and 5% CO₂ for 24 h before being used for electrophysiology measurements.

The complete open reading frame of Kcnj13 (the gene encoding Kir7.1) was amplified from a cDNA library, generated by standard procedures from a C57BL/6N male murine small intestine, an organ known to express high levels of Kcnj13. Primers used for the amplification bound to the untranslated region of the gene (in capital letters) with added restriction sites for SacI and SalI (Kir7.1 forward, SacI: 5′-ata​aga​gct​cCA​CAA​AGG​AAC​CGA​GAA​ACA​C-3′; Kir7.1 reverse, SalI: 5′-ata​agt​cga​cGG​ACG​GTG​TAG​ATG​AGT​CTT​A-3′). After restriction digest of the sequence-confirmed full-length Kcnj13 and the mammalian expression vector pIRES2-EGFP (Clonetech; 6029–1) by standard procedures, Kcnj13 was inserted into the vector using the Quick Ligation Kit (New England BioLabs; M2200) according to the manufacturer’s instructions. For recombinant expression, cells of the human embryonic kidney cell line (ATCC; HEK293, CRL-1573) were plated on collagen (Roche; 11179179001)–coated coverslips and transfected using Lipofectamine 2000 reagent (Life Technologies; 11668027) according to the manufacturer’s recommendation. Overexpression was verified by the presence of Kir7.1 in immunocytochemical staining and Western blot analysis as mentioned below. The EGFP of the pIRES2-EGFP vector was used to visualize the transfected cells for whole-cell patch-clamp experiments.

For staining of mouse brain sections, C57BL/6N, Abhd2^+/+, and Abhd2^−/− adult male mice were euthanized by perfusion with 4% paraformaldehyde (PFA; Electron Microscopy Sciences; 15714S) after which the brains were dissected out, placed in a sucrose gradient, and frozen in Tissue-Tek O.C.T. compound (Sakura; 4583). The brains were sectioned using a cryotome into 8-µm-thick coronal sections. For Kir7.1 staining, the sections were permeated by incubation with 0.5% Triton X-100 (Fisher BioReagents; BP151-100) for 15 min. For the detection of α/β hydrolase domain–containing protein 2 (ABHD2), antigen retrieval was performed by incubation with 1% SDS (Sigma; L4509-100G) for 5 min. Thereafter, the sections were washed three times in PBS, blocked with 5% BSA (Sigma; A9647) for 1 h at room temperature, and incubated with primary antibody in PBS supplemented with 1% BSA overnight at 4°C. After washing three times in PBS, the sections were incubated with secondary antibody for 1 h, washed three times in PBS, and mounted.

For the detection of both ABHD2 and Kir7.1 in CPECs isolated from Abhd2^+/+ and Abhd2^−/− mice, the cells were fixed with 2% PFA in PBS for 15 min and washed three times with PBS followed by antigen retrieval for 5 min using 0.75% SDS. For blocking, the cells were incubated overnight in 3% BSA and 0.1% Saponin (Sigma; 47036-50G-F) in PBS at +4°C, followed by overnight incubation with the primary antibody in blocking solution at +4°C. After washing three times in 0.1% saponin/PBS solution, the cells were incubated with secondary antibody in PBS supplemented with 1% BSA and 0.1% saponin for 1 h, washed three times in 0.1% saponin/PBS solution, and mounted. For detection of Kir7.1 in HEK293 cells (transfected as previously mentioned and cultured overnight on poly-D-lysine (Millipore; P1024) coated coverslips), the cells were fixed with 4% PFA in PBS for 10 min, permeated, stained, and mounted according to the protocol for Kir7.1 staining of brain sections.

For detection of Kir7.1 in cultured RPE cells (isolated and cultured overnight on poly-D-lysine– and laminin-coated coverslips), cells were fixed with 2% PFA in PBS for 15 min and washed three times with PBS. Next, permeabilization and blocking of the cells were performed simultaneously by 1-h incubation in blocking solution (1% BSA and 0.1% saponin in PBS). The cells were then incubated with the primary antibody in the blocking solution overnight at 4°C. After washing three times in 0.1% saponin/PBS, the cells were incubated with secondary antibody for 1 h, washed three times in 0.1% saponin/PBS solution, and mounted.

The following primary antibodies were used: rabbit polyclonal anti-ABHD2 (Proteintech Group; 1:300, 14039–1-AP) and mouse monoclonal anti-Kir7.1 (Santa Cruz Biotechnology; 1:300, sc-398810). The following secondary antibodies conjugated with different fluorescent dyes were used: Dylight 488–conjugated goat anti-rabbit IgG (1:2,000; Molecular Probes; A11008; Jackson ImmunoResearch; 111–485-144) and Cy5-conjugated donkey anti-mouse (Jackson ImmunoResearch; 1∶2,000, 715–175-150). For mounting, we used ProLong Gold antifade reagent with DAPI (Invitrogen; P36935). The cells were visualized with an inverse microscope (Olympus; IX71) equipped with lumencor light engine (SPECTRAX; 6-LCR-SA) while the brain sections were visualized using a confocal laser scanning microscopy (Olympus; Fluoview FV1000).

Lateral CP slices, isolated as described above, were used for ISH. RNA was extracted from the murine ovary using a Qiagen RNAeasy mini kit followed by cDNA synthesis with a Phusion RT-PCR kit (Finnzymes). The specific translated region of Abhd2 was amplified using the following primers: (forward, 5′-GTC​GGA​TGG​TGC​CAC​TTC-3′; reverse, 5′-CCT​CCA​TCT​GCT​CCG​TGT-3′) and subcloned into a vector containing Sp6/T7 promoters (Promega). Single-strand digoxigenin-labeled RNA probes were synthesized (Roche), and color ISH was performed as in Ishii et al. (2004). A sense probe was used as a negative control.

To detect Kir7.1 in HEK293 cells transfected with the pIRES2-EGFP/Kir7.1 construct, the cells were cultured for 24 h after transfection and harvested by standard procedures. The samples were analyzed by Western blotting using the mouse monoclonal antibody against Kir7.1 (1:10,000 dilution) and a peroxidase-conjugated goat anti-mouse secondary antibody (EMD-Millipore; 1:15,000 dilution, AP181P). To ensure equal sample loading, the membrane was stripped by incubation with OneMinute Plus Strip (GM Biosciences; GM6011) according to the manufacturer’s instructions. Thereafter, the membrane was rehybridized with mouse monoclonal anti-actin antibody (Abcam; 1:5,000 dilution, ab3280-500) and peroxidase-conjugated goat anti-mouse secondary antibody (1:15,000 dilution).

KMeSO₃ was purchased from Alfa Aesar (39505), P4 from Millipore (5341), and VU590 (7,13-bis[(4-nitrophenyl)methyl]-1,4,10-trioxa-7,13-diazacyclopentadecane) dihydrochloride from Fisher (389110). All other compounds were from Sigma. Testosterone (Sigma; T1500) was purchased in accordance with the controlled substance protocol CSUA #123 approved by the University of California, Berkeley, Office of Environment, Health & Safety. For patch-clamp measurements, all bath solutions contained a concentration of 1:1,000 EtOH and DMSO, the solvents used to dissolve the steroid hormones and the antagonist VU590, respectively.

CPECs and HEK293 cells, plated on 5-mm glass coverslips, were placed in a perfusion chamber (Warner Instruments; RC-24E), and RPE cells plated on Transwell membranes were placed in an open bath recording chamber (Warner Instruments; RC22) and secured by a slice anchor (Warner Instruments; SHD-22L/15) for electrophysiological measurements at ambient temperatures. Data were acquired using Clampex 10.5 software (Molecular Devices), which controlled an AXOPATCH 200B amplifier and an AxonTM Digidata 1550A digitizer (both from Molecular Devices) with an integrated Humbug noise eliminator. Data were not corrected for liquid junction potential changes. Gigaohm seals were established in dissection media (in mM: 140 NaCl, 15 HEPES, 15 d-glucose, 5 KCl, 1 CaCl₂, and 1 MgCl₂, [pH 7.3 adjusted with NaOH, 310 mOsm/liter]). The patch pipette was filled with either a Cs⁺- or a K⁺-based solution (in mM: 130 CsMeSO₃ or KMeSO₃, 20 HEPES, 5 BAPTA (1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid), and 1 MgCl₂, [pH 7.3 adjusted with CsOH or KOH, 295 mOsm/liter]) and had a resistance of 3–6 MΩ. For measurements of the difference in monovalent currents, the bath solutions contained, in mM: 43 HEPES, 1 MgCl₂, and one of the following: 130 CsMeSO₃, 130 KMeSO₃, 130 RbCl, or 130 NaMeSO₃ [pH 7.3, 310 mOsm/liter]). All buffers, including those containing steroid hormones, were applied under constant perfusion. The Kir7.1 antagonist VU590 (100 µM), in combination with P4, was applied directly to the bath solution. To block G-protein signaling, 1 mM guanosine 5′-O-(2-thiodiphosphate; GDPβS) was applied to the K⁺-based pipette solution. Measurements were performed at 10 kHz and filtered at 1 kHz. Cells were stimulated every 5 s by voltage ramps from −80 mV to +80 mV with a holding potential of 0 mV or by 500-ms voltage steps from −120 mV to +40 mV in 20 mV increments with a holding potential of −40 mV. Access resistance and membrane capacitance were between 4 and 10 MΩ and 33 ± 1.09 pF for CPECs, 7–13 MΩ and 19.3 ± 1.08 pF for HEK293 cells, and 10.5–14 MΩ and 12.7 ± 2.17 pF for RPE cells. Membrane capacitance served as a proxy for the cell surface area and used for normalization of current amplitudes (i.e., current density). Capacitance artifacts were graphically removed in OriginPro 8.6 (OriginLab). For current-clamp measurements, the pipette solution contained (in mM) 40 NaCl, 100 KMeSO₃, 20 HEPES, 1 BAPTA, 5 Mg-ATP, and 0.5 Na-GTP (pH 7.2 was adjusted with KOH before adding the ATP and GTP stocks, 295 mOsm/liter). The bath solution for current clamp contained (in mM) 140 NaCl, 5 KMeSO₃, 1 MgCl₂, 20 HEPES, and 10 d-glucose (pH 7.3 adjusted with NaOH, 310 mOsm/liter). Measurements were performed at 10 kHz and filtered at 10 kHz under a gap-free protocol with a 0 pA input signal.

For statistical analyses of the activation and inhibition of Kir7.1 by different steroid hormones and the antagonist VU590, in measurements from CPECs, HEK293, and RPE cells, Clampfit 10.3 (Molecular Devices) and GraphPad Prism 5 (GraphPad Software) software was used. Unpaired and paired t test were used to determine statistical significance, assigning P ≤ 0.05 as the limit. All results are shown with SEM.

Fig. S1 shows the relative gene expression of the lateral CPs isolated from Abhd2^+/+ and Abhd2^−/− male and female mice. Fig. S2 shows expression of Abhd2 in murine CPs. Fig. S3 shows a summary of cumulative current densities recorded during the study and the presence of I_TRPM3 in CPEC. Data S1 lists the genes expressed in the lateral CPs from wild-type and Abhd2^−/− male and female mice.

To study the nongenomic effect of steroid hormones on CP ion conductance, whole-cell patch-clamp experiments were performed on CPECs isolated from the murine lateral ventricle. To avoid possible changes in gene expression during prolonged cell culture, all analyses were performed on the same day of tissue collection, within 3–5 h of cell extraction. When the cells were placed in a bath solution containing Cs⁺ and Mg²⁺, a mostly outward current with a small inward component was observed in response to the voltage ramp from −80 mV to +80 mV applied from a holding potential of 0 mV (Fig. 1 A). This inward current was further significantly potentiated by exposure to P4 with a 15.5 ± 2.4-fold and 5.7 ± 0.8-fold increase in cells isolated from female and male CP, respectively (Fig. 1, A and B), developing a strong inward rectification. Interestingly, while the initial response to P4 application was statistically stronger in females, repetitive P4 applications increased the current fold change of male CPECs from the initial 5.7 ± 0.8-fold for the first application to 7.8 ± 1.9-fold for the second and eventually to 11.9 ± 2.3-fold for the third application of P4 (Fig. 1 B). The female CPECs still retained stronger response to P4 for the second (18.2 ± 3.6-fold), as well as for the third (17.6 ± 4.1-fold), applications of this steroid. The cellular response to P4 was also initially slow and required almost 6 min to achieve saturation. However, after subsequent washout and readministrations, the response was almost immediate, with only a slight desensitization of the current (Fig. 1 C). The initial slow activation component was similar among females and males, with average activation kinetics of 4.8 ± 1.3 min for females and 4.5 ± 0.9 min for males (data from three independent experiments).

To determine the ion channel responsible for the observed current, we analyzed mRNA expression data from adult murine CPs. Interestingly, one of the most abundantly expressed ion channels in CPECs is the inwardly rectifying potassium channel Kir7.1 (encoded by Kcnj13; Fig. S1 A, Table 2, and Data S1), which is also the only potassium channel that conducts Cs ions and is not blocked by them (Krapivinsky et al., 1998). To test the hypothesis that P4 activates Kir7.1, the Kir7.1/K_ir1.1–specific antagonist VU590 was used. VU590 significantly reduced the inward current and further prevented P4 to potentiate it (Fig. 2 A). Since the expression of Kcnj1 (encoding Kir1.1) is not detected in CPECs (Table 2 and Data S1), and given the presence of Cs⁺ in the recording pipette that should inactivate any potassium channels except Kir7.1, it is highly likely that the observed reduction in current is caused by blocking the activity of Kir7.1. Furthermore, Kir7.1 was detected via immunohistochemical staining in the apical membrane of the murine CP and was also present after lateral CP cell isolation and culture (Fig. 2, B–D). Kir7.1 has a signature ion selectivity with the following order of preference Rb⁺>>Cs⁺>K⁺>>Na⁺ (Shimura et al., 2001), which was indeed observed in our CPEC recordings, as is evident from representative I-V curves and current density profiles (Fig. 2, E–G). Since Kir7.1 mainly conducts potassium ions in its native state, further experiments were performed using bath and pipette solutions where Cs⁺ was replaced with K⁺. In addition, 100 µM VU590 was applied at the end of each recording to confirm that the measured potentiated current was not a byproduct of a leak current.

By measuring Kir7.1 potentiation with different concentrations of P4 the half maximal effective concentration for P4 was determined to be 7.02 ± 2.3 µM, with the maximal effective concentration reached ∼50 µM (Fig. 3, A and B). To test whether other steroid hormones are able to increase the activity of Kir7.1, CPECs were stimulated with 10 µM testosterone, estradiol, pregnenolone sulfate (PS) and the synthetic progestin levonorgestrel (Fig. 3, C and D). Surprisingly, none of these steroids were able to significantly change the activity of Kir7.1 (Fig. 3, C and D), although the addition of 10 µM P4 at the end of each experiment still caused a 3.08 ± 0.20-fold increase in the inward current at −80 mV (Fig. 3 D). The results indicate that P4 activation of Kir7.1 is specific for this steroid. The significance of P4 activation was also analyzed by the current-clamp method, where the solutions used had an ion concentration and osmolarity comparable to that of CSF and the intracellular environment of CPECs. To measure the native membrane potential of CPECs, no current was applied to the cells, and the average initial V_m measured was −57.2 ± 4.6 mV (n = 3; Fig. 3 E). Under such conditions, the application of 10 µM P4 caused hyperpolarization of the cells, with an average decrease in membrane potential of 9.13 ± 3.56 mV (Fig. 3, E and F). Similar to the whole-cell patch-clamp recordings, applying 100 µM of the antagonist VU590 caused a significant inhibition of Kir7.1 conductance, which led to depolarization of the cell, with a significant increase in membrane potential of 29.95 ± 6.51 mV, even in the presence of P4 (Fig. 3 E).

The CP epithelium contains several membrane-bound P4 receptors that could potentially mediate the cell response to P4. Interestingly, we found that the recently characterized P4 receptor ABHD2 (Miller et al., 2016) is also highly expressed in the CP (Figs. S1 and S2 and Table 1) and cultured CPECs (Fig. 4). ABHD2 has been shown to regulate ion channel activation in human sperm cells by breaking down the endocannabinoid 2-arachidonoylglycerol in the cell membrane (Miller et al., 2016). However, CPECs isolated from global Abhd2 knockout mice that were generated in our laboratory (Björkgren et al., 2019 Preprint) showed activation of Kir7.1 similar to that of cells from wild-type animals (Fig. 4, E and F), thus negating any possible effect of the receptor on Kir7.1 activation. However, CPECs isolated from Abhd2 knockout female and male mice showed an altered Kir7.1 slow activation component in response to P4 stimulation, with faster activation kinetics of 2.5 ± 0.43 min for females and 1.96 ± 0.36 min for males (data from six to nine independent experiments). The kinetics is significantly faster than the activation kinetics shown by wild-type littermates, with 4.8 ± 1.3 min for wild-type females and 4.5 ± 0.9 min for wild-type males (n = 3). While it is unlikely that ABHD2 is required for Kir7.1 activation by P4, it is possible that its presence changes the bioactive lipid landscape of the cells and alters the channel activation.

To further validate a direct interaction between P4 and Kir7.1, the full-length mRNA sequence of murine Kcnj13 was cloned into an expression vector and subsequently expressed in HEK293 cells. The cells do not display any endogenous expression of the channel, but, as was demonstrated in a previous study (Carrington et al., 2018), the products of the transfected construct were glycosylated and transported to the cell membrane, as shown by Western blot and immunocytochemistry (Fig. 5, A–D). Whole-cell patch-clamp recordings of the transfected HEK293 cells show a similar potentiation of I_Kir7.1 after applying 10 µM P4 to the bath solution (Fig. 5, A and E; and Fig. S3 A), and the fast activation of Kir7.1 by P4 in HEK293 cells resembled the activation of CPECs upon second and third P4 application.

Furthermore, our mRNA-sequencing data provided evidence of a possible G protein–coupled progesterone receptor in the CP (Table 1), a signaling pathway previously shown to be involved in Kir7.1 regulation (Carrington et al., 2018). To test whether or not these signaling pathways would lead to the observed potentiation of Kir7.1 activation, 1 mM GDPβS, a competitive inhibitor of GTP activation of G-protein signaling, was applied to the pipette solution while recording Kir7.1 from either CPECs (Fig. 5, B and E) or HEK293 cells (Figs. 5 E and S3 A). In both cases, the presence of the inhibitor had no effect on P4- mediated Kir7.1 potentiation (Fig. 5, B and E), which further confirmed that G-protein signaling is not involved in P4 regulation of Kir7.1. It also must be mentioned that our intracellular (pipette) solutions did not contain ATP, GTP, or any other essential nucleotides required to power G-protein signaling. Recent reports suggest that Kir7.1 could be coupled to the melanocortin-4 receptor in a G protein–independent manner (Ghamari-Langroudi et al., 2015); however, melanocortin-4 receptor is not expressed in the murine CP (Data S1). Together, these results make a compelling argument for the activation of Kir7.1 to be caused by a direct interaction of steroid P4 with the channel.

Kir7.1 is also found in the RPE (Shimura et al., 2001), where the channel works in concert with the Na⁺-K⁺-ATPase to regulate K⁺ homeostasis and photoreceptor excitability (Shahi et al., 2017). To determine if Kir7.1 in RPE cells shows similar activation by P4 as in CPECs, whole-cell patch-clamp recordings were performed. RPE cells that had been isolated and cultured overnight were used for the experiments (Fig. 6). The currents measured from these cells were considerably smaller than the currents of CPECs (Fig. 6, A–C; and Fig. S3 A), and the slow kinetics of the onset of P4 response as recorded from CPECs (Fig. 1 C) was absent from RPE recordings (Fig. 6 D). However, a statistically significant potentiation of inward I_Kir7.1 was observed after stimulation with 10 µM P4 (Figs. 6 C and S3 A). Thus, the effect of P4 on Kir7.1 activity is not limited to the CP but could, in addition, have important implications in other tissues.

The adult murine CP represents a novel model to study nongenomic steroid signaling, because P4 is able to regulate this tissue (Lindvall-Axelsson and Owman, 1990; Kelley and Mermelstein, 2011) even under negligible expression levels of the nuclear progesterone receptor (Table 1 and Data S1). Since P4 is known to affect ion channels in neurons, either through a direct interaction between the steroid and the channel or through an indirect response that occurs following steroid binding to its receptor (Kelley and Mermelstein, 2011; Luoma et al., 2011; Majeed et al., 2011; Miehe et al., 2012; Miller et al., 2016), we aimed to determine whether P4 can influence ion conductance in CPECs via one of those mechanisms. By performing electrophysiological studies on CPECs, we showed a novel, P4-specific activation of an inwardly rectifying potassium channel, Kir7.1. This P4 activation was inhibited by the addition of VU590, a specific inhibitor of Kir1.1 and Kir7.1 channels. Since only Kir7.1 is highly expressed in CPECs, and all other potassium conductances were blocked by intracellular Cs ions, these experiments further confirm the channel identity. Furthermore, the recorded conductance displayed Kir7.1-specific hallmarks, such as the lack of an internal Mg²⁺ block, a high Rb⁺ conductivity, and a low sensitivity to Cs⁺ (Krapivinsky et al., 1998; Shimura et al., 2001; Sackin et al., 2010). Eventually, we were able to recapitulate Kir7.1 potentiation by P4 with recombinantly expressed Kir7.1.

Interestingly, the potentiation of I_Kir7.1 by P4 was independent of ABHD2, a known P4 membrane receptor. This was confirmed by recordings from Abhd2-deficient murine CPECs, which retained a similar activation of Kir7.1 by P4. The involvement of G protein–coupled receptors in this process was ruled out, as addition of GDPβS, a competitive inhibitor of G-protein signaling, did not alter the potentiation of I_Kir7.1. Additionally, recordings of a recombinant Kir7.1 expressed in HEK293 cells, as well as recordings in the absence of intracellular messengers, such as GTP, displayed similar potentiation of Kir7.1 by P4. We, therefore, suggest that P4 activates Kir7.1 directly.

There were clear differences in stronger up-regulation of Kir7.1 by P4 recorded from female CPECs compared with those of age-matched male mice. Based on the gene expression profile (Table 2), Kcnj13 expression is not statistically different between sexes; however, it is possible that a difference in the lipid composition of the plasma membranes of CP cells between males and females would be the cause of these distinct responses. This important phenomenon requires more in-depth studies to reveal the molecular basis behind the observed sex differences. Interestingly, previous studies on the effect of membrane fluidity on Kir channels have shown their sensitivity to cholesterol-driven regulation. Although the activity of most K_ir channels is suppressed with increased cholesterol levels in the membrane, Kir7.1 is an exception to the rule and, instead, showed a slight activation of ion flux (Rosenhouse-Dantsker et al., 2010). In silico analyses stated that this activation was caused by a weak interaction between Kir7.1 and cholesterol rather than a change in membrane fluidity (Fürst et al., 2014). Similarly, steroids are known to change the cell membrane fluidity (Whiting et al., 2000; McDonnel et al., 2003; Korkmaz and Severcan, 2005). If Kir7.1 activation by P4 was due only to an altered membrane composition, then other lipophilic hormones would have had a similar effect on Kir7.1. However, none of the tested steroids (testosterone, estradiol, levonorgestrel, and PS) were able to influence I_Kir7.1. Thus, Kir7.1 displays a specific affinity for P4, and the mechanisms of such an interaction will be interesting to explore further. However, we cannot exclude the possibility that other steroid molecules can exhibit suppressing or inhibitory effect on Kir7.1 activity, and it is possible to assume that their basal concentration might be different in plasma membrane of the CP between sexes.

Another interesting observation is the slow kinetics of the onset of P4 responses that is only evident upon initial application of P4 to CP cells and is absent from consequent P4 applications or even the initial application of P4 to RPE cells. This could be explained by a presence of a still-unknown lipid factor, such as an inhibitory lipid, another steroid, or a potential additional Kir7.1 regulatory subunit, that must be eliminated or decoupled from the membrane or ion channel complex before P4 can exert its full effect. All of these important phenomena require follow-up studies that may shed a light on sex differences in CP physiology. As mentioned above, Kir7.1 is also found on the apical side of RPE cells, where it works in concert with the Na⁺,K⁺ pump to maintain the ion homeostasis of the subretinal space (Shimura et al., 2001). Kir7.1 is vital for normal RPE physiology, as RPE-specific deletion of the channel leads to photoreceptor degradation (Roman et al., 2018). Mutations in Kir7.1 gives rise to snowflake vitreoretinal degeneration and Leber congenital amaurosis, retinal dystrophy disorders that also cause blindness (Hejtmancik et al., 2008). Similar to CPECs, RPE cells use Kir7.1 to balance the ion content of the cell through secretion of K⁺ ions taken up by the Na⁺,K⁺-ATPase. Our results showed that Kir7.1 expressed in RPE also responded to P4, which points to a potential regulatory role for P4 in the function of RPE. The effect of sex hormone signaling on RPE functions has not been well studied, although enzymes required for synthesis of P4 are found in the retina (Guarneri et al., 1994; Cascio et al., 2015). Changes in hormone production during aging are also thought to cause a difference in eye disorder prevalence among men and women (Cascio et al., 2015). Further studies on the connection between P4 production in the eye and Kir7.1 activity could give new insight into these discrepancies.

We are just beginning to observe the full impact of Kir7.1 on brain functions, especially since a recent study by Papanikolaou et al. (2019) showed expression of Kir7.1 throughout the brain in both neuronal and glial structures. Our data prompt future studies on the relationship between P4-driven Kir7.1 regulation and CSF production, as well as the regulation of ventricle size during brain trauma. It also would be important to explore the role of P4-activation of Kir7.1 in TBI pathology. At high enough concentrations, P4 could lead to hyperpolarization, altering ion exchange and inhibiting water secretion into the ventricles, which ultimately would benefit TBI patients. While the low micromolar concentration range required for Kir7.1 activation appears to be supraphysiological for CSF, which is known to have a low composition of steroid carriers, the concentration of steroid hormones within the CPEC plasma membrane is not known. The filtration rate of the CP is ∼0.5 liters per day, which allows for a total exchange of the 150 ml CSF within the ventricular system three or four times daily (Marques et al., 2017). As filtration progresses, it leads to accumulation of lipophilic compounds and locally produced neurosteroids in the CPEC plasma membrane, which can ultimately reach the concentration levels required for channel modulation. The physiological concentrations of P4 in CSF fluctuate between 0.1 nM (Datta et al., 1986) and 10 nM (Datta et al., 1986; during pregnancy); however, circulating P4 concentrations can reach high micromolar concentrations upon intramuscular or intravenous injections of 1–80 mg P4 during the course of TBI treatment (Pan et al., 2019), which potentially could lead to Kir7.1 modulation.

CP represents a convenient tissue to study the physiology of ion channels, particularly their regulation by steroid hormones via nongenomic mechanisms (Haoui et al., 2021), because of the low expression of genomic steroid receptors in the adult CP (Table 1). As is evident from CP gene expression analyses (Data S1), several transient receptor potential ion channels, such as TRPM3 and TRPV4, were also found to be expressed in CP (Millar et al., 2007). In fact, TRPM3 regulation by PS (Majeed et al., 2012) can be recorded from CP cells upon inhibition of Kir7.1 activity by VU590 (Fig. S3 B). Additionally, TRPM3 conductance could be partially responsible for the outward currents that were observed in this study.

In conclusion, the inwardly rectifying potassium channel Kir7.1 is emerging as the major ion channel in the CP potentiated by P4 via a G protein–coupled receptor-independent mechanism. As few endogenous regulators of Kir7.1 are known, its P4 activation represents a novel significant regulatory modality for this channel. The potentiation of Kir7.1 by P4 was steroid specific and took place regardless of what cell type the channel was expressed in. The impact of P4 regulation on Kir7.1 may, therefore, not be limited to the CP and RPE but extend to all tissues where the channel is found. Future studies should focus on the possible impact of increased P4 levels on these tissues.

Joseph A. Mindell served as editor.

We thank Dr. Xiaoquan Li for help with the pilot experiments involving in situ staining for ABHD2, Dr. Terry Machen for help with equipment and advice on epithelial biology, Elena Solonenko for help with cell preparations, and Dr. Lin He laboratory for advice on initial isolation of the CP.

This work was supported by National Institutes of Health grant R01GM111802, a Pew Innovation Fund 2020 award from Pew Charitable Trusts, a Rose Hills Foundation award, and Packer Wentz Endowment Will. This work was also funded in part by the Global Consortium for Reproductive Longevity and Equality at the Buck Institute, made possible by the Bia-Echo Foundation, award GCRLE-0920 (P.V. Lishko). This work also used the Vincent J. Coates Genomics Sequencing Laboratory at University of California, Berkeley, which is supported by National Institutes of Health instrumentation grant S10 OD018174.

The authors declare no competing financial interests.

Author contributions: I. Björkgren and P.V. Lishko contributed to the conception and design of this research and wrote the manuscript together with M. Haoui. All authors performed experiments, analyzed data, and reviewed and approved the final version of the manuscript.