Phosphate buffered saline (PBS), medium-199, rat ELISA kits for GH, horse serum, penicillin, and streptomycin were purchased from Life Technologies (Grand Island, NY, USA). Fura 2-AM, the secondary Alexa Fluor 568 donkey anti-rabbit antibody, and 4′,6-diamidino-2-phenylindole (DAPI) were from Invitrogen (Eugene, OR). Rodent LH ELISA Kit was from Endocrine Technologies (Newark, CA). The primary antibody and standard for the PRL radioimmunoassay as well as the rabbit anti-PRL antibody for immunocytochemistry were from the National Pituitary Agency and Dr. A. F. Parlow (Harbor-UCLA Medical Center, Torrance, CA). ¹²⁵I-Prolactin was from Perkin Elmer Life Sciences (Boston, MA). Bovine serum albumin (BSA) fraction V and saponin were from MP Biomedicals (Solon, OH). The transcriptor first strand cDNA synthesis kit was from Roche Applied Sciences (Indianapolis, IN), Formaldehyde solution was from Thermo Fisher Scientific (Rockford, IL); Bay K8644, PIK93, UNC3230, and Wm were from Tocris (Bristol, UK). ISA2011B was from Med Chem Express (Monmouth Junction, NJ). TRH and dopamine were from Bachem Americas (Vista, CA). Unless stated otherwise, all other chemicals were obtained from Sigma (St. Louis, MO, USA).

Experiments were performed on cultured anterior pituitary cells from normal 75-day-old female Sprague-Dawley rats obtained from Taconic Farms (Germantown, NY). Animals were housed under constant conditions of temperature and humidity, with lights on between 6 AM and 8 PM and water and food ad libitum. Euthanasia was performed by asphyxiation with CO₂, and the anterior pituitary glands were removed after decapitation. Pituitary tissue was cut into 0.5 x 0.5 mm pieces and treated with trypsin, followed by trypsin inhibitor, and mechanically dispersed as previously described (28). Freshly dispersed pituitary cells were cultured in medium 199 containing Earle’s salts, sodium bicarbonate and supplemented with 10% heat-inactivated horse serum, penicillin (100 units/ml), and streptomycin (100 µg/ml). For immunocytochemistry, 50,000 cells/well were plated on poly-L-lysine-coated eight-well multitest slides (MP Biomedicals, Aurora, OH). For secretory studies, cells were cultured in biocoated 24-well plates (Corning, Kennebunk, ME), 0.25 million per well or in 60 mm petri dishes (Falcon/Becton Dickinson, Franklin lakes, NJ) containing pre-swollen Cytodex-1 beads, 12 million per dish. For RNA extraction, cells were seeded on poly-L-lysine-coated 24-well plates, 1.5 million per well. For calcium experiments, cells were plated at a density of 0.7 million per 25 mm glass coverslip (Thomas Scientific, Logan Township, NJ) coated with poly-D-lysine. In all cases, cells were cultured for approximatively 20 h, washed and treated with several PI4K, PIP5K, and PI3K inhibitors or solvents ( Figure 1A ). All experimental procedures were in accordance with the National Institutes of Health Policy Manual 3040-2: Animal Care and Use in the Intramural Program and were approved by the National Institute of Child Health and Human Development, Animal Care and Use Committee (Animal Protocol 19-041).

Dispersed cells from two cell preparations were used for single-cell RNA sequencing (scRNAseq) analysis, as described in Fletcher et al. (32); one preparation was done by dispersion of 30 whole pituitary glands, and the other dispersion was done with 30 separate anterior and intermediate/posterior lobes of the pituitary gland. The whole pituitary sample was run in duplicate, and the anterior and posterior samples were run in separate lanes on a 10X Genomics chromium controller according to manufacturer instructions. The resulting libraries were sequenced on an Illumina HiSeq 2500, and the Cellranger pipeline (10X Genomics) was used for read mapping and transcript counting. A total of 32,455 droplets were recovered from the four lanes. A custom reference genome, cell filtering, differential expression analysis, and ambient RNA removal by SoupX were used as previously described (33). The final data set contained 15,876 cells, for which cell type clusters were identified using known genetic markers: Lum, Col3a1, and C7 for pericytes; Plvap, Kdr, and Flt1 for endothelial cells; Lhx2, Col25a1, and Fgf10 for pituicytes, Mt2A, Aldh3a1, and Capn6 for folliculostellate cells; Pomc for melanotrophs and corticotrophs, with melanotroph-specific Oacyl, Pax7, and Pcsk2, and corticotroph-specific Doc2g, Avpr1b, and Crhr1; Lhb, Gnrhr, and Nr5a1 for gonadotrophs; Pou1f1 for thyrotroph, somatotrophs and lactotrophs, Trhr for thyrotrophs and lactotrophs, with thyrotroph-specific Tshb and Kcnk9, somatotroph-specific Gh1, Ghrhr, and Arhgap36 for somatotrophs, and Prl, Drd2, and Agtr1b for lactotrophs. As expected, the majority of pituicytes and melanotrophs were detected in the posterior pituitary preparation, and pericytes and endothelial cells were present in both the anterior and posterior pituitary preparations. In this study, we use cell-type-specific clustering to examine the expression levels of selected PI3K, PI4K, and PIP5K genes and the percentage of cells of each cell type expressing these genes. Expression of genes of interest was plotted using uniform manifold approximation and projection (UMAP, 34) and percent-expression dot plots using MATLAB (R2018b; The MathWorks, Natick, MA). Single cell RNA sequencing data are accessible at NCBI Gene Expression Omnibus database, accession GSE184319.

After overnight incubation, 8-well multitest slides were washed, cells were treated with Wm or 0.1% DMSO for 1, 2, or 3 h, washed twice with PBS, and fixed with cold 4% formaldehyde solution for 10 min. From this point, every step of immunostaining protocol was followed by washing cells three times with PBS. Cells were incubated with rabbit anti-PRL antibody (1:2000), followed by subsequent incubation with secondary antibody (1:1000) for 30 min at room temperature. All antibodies were diluted in PBS containing 0.2% saponin and 0.5% BSA. Cells were mounted with Flouromount-G, with DAPI. All images were acquired on an inverted confocal laser-scanning microscope (LSM 780; Carl Zeiss GmbH, Jena, Germany), using a 63x oil objective. Micrographs were sized, and their brightness and contrast levels adjusted in Fiji image processing software (35). Cells were counted on 8 tile-scan images from 2 independent experiments (3x3).

Total RNA was extracted from dispersed anterior pituitary cells using RNeasy Plus Mini Kit and reverse transcribed with the Transcriptor First Strand cDNA Synthesis Kit. qRT-PCR was performed in the presence of cDNA (2 ng), TaqMan™ Fast Advanced Master Mix, and TaqMan Gene Expression Assays using the Quantum Studio 6 Flex Real Time PCR System. Target gene expression levels were determined by the comparative 2^-(delta Ct) quantification method using Gapdh as the reference gene, which was previously established to be suitable for anterior pituitary tissue (36). Applied Biosystems pre-designed TaqMan Gene Expression Assays were used: prolactin (Prl, 00561791_m1) and glyceraldehyde 3-phosphate dehydrogenase gene (Gapdh; Rn01462662_g1).

For measurements of intracellular calcium concentration ([Ca²⁺]_i), cultured pituitary cells were washed and bathed in Krebs-Ringer-like medium containing 2.5 μM Fura 2 AM for 1 h at room temperature. The coverslips were washed in Krebs-Ringer-like medium and mounted on the stage of an inverted Observer-D1 microscope (Carl Zeiss, Oberkochen, Germany) with an attached ORCA-ER camera (Hamamatsu Photonics, Hamamatsu City, Japan) and a Lambda DG-4 wavelength switcher (Sutter, Novato, CA). Hardware control and image analysis were performed using Metafluor software (Molecular Devices, Downingtown, PA). Experiments were performed under a 40X oil-immersion objective during exposure to alternating 340 and 380 nm excitation beams, and the intensity of light emission at 520 nm was followed simultaneously in approximately 20 single cells. Changes in [Ca²⁺]_i are represented by the ratio of fluorescence intensities F₃₄₀/F₃₈₀. Single lactotrophs in the mixed population of pituitary cells were identified by their responses to both TRH (100 nM) and dopamine (1 µM), in contrast to thyrotrophs responding only to TRH (100 nM).

Two types of secretory experiments were performed: static incubations and perfused pituitary cells. After overnight incubation, static cultures were washed with medium 199 supplemented with warm HEPES-containing medium-199 supplemented with 0.1% BSA and penicillin and streptomycin to estimate basal PRL release in the presence of phosphatidylinositol kinase (PIK) inhibitors or solvents used as controls. Cells were incubated at 37°C for 1 h, or in time-course studies for 1-3 h; afterwards, medium was removed to measure secreted hormones. To analyze changes in the intracellular content of PRL in these experiments, 0.5 ml of ice-cold 20 mM sodium carbonate buffer was added to wells, and plates were frozen at -80°C. Next, plates were scraped using a 1 ml pipette tip and contents transferred to individual tubes. The process of scraping was repeated with 0.5 ml of fresh buffer or ethanol and added to the tube containing the first extract. Cell extracts were centrifuged at 3,000 rpm to remove cell debris. Samples were stored at -20°C until analysis. In perfusion experiments, both basal and stimulated (25 mM KCl, 1 µM Bay K8644, or 100 nM TRH) release of PRL were studied. After overnight incubation, beads with attached pituitary cells were transferred to 37°C-heated chambers and continuously perfused with warm HEPES-containing medium-199 supplemented with 0.1% BSA and penicillin and streptomycin for 2 h at a flow rate of 0.5 ml/min to establish stable basal secretion. Basal and stimulated PRL release was studied in the presence of PIK inhibitors or solvents, which were continuously applied to cells, and 1- or 3-min fractions of perfused medium were collected and stored at -20°C until hormone measurements. In most experiments, PRL (released and cell content) was measured by radioimmunoassay (RIA) as previously described (37). In some experiments, released PRL, LH, and GH were measured by ELISA, following the manufacturers’ instructions. Released hormones were expressed as ng/ml and cell content of hormones as µg/well.

Results are presented as representative traces (perfusion experiments) and mean ± SEM values (static incubations and selected perfusion experiments). Statistical analysis was performed by Kruskal-Wallis and Wilcoxon tests, with at least P < 0.01 deemed as statistically significant. Mean and SEM values, statistical analysis, half-maximal inhibitory concentrations (IC₅₀) and time to half-maximal inhibition (τ₅₀) values were calculated using the KaleidaGraph program (Synergy Software, Reading, PA).

To characterize the expression of PIK genes in lactotrophs and other pituitary cell types, we examined the expression levels of these genes using scRNAseq of freshly dispersed anterior and posterior pituitary cells of adult female rats. The UMAP plots shown in Figure 1 indicate that four PI4K genes, Pi4ka, Pi4kb (B) Pi4k2a, and Pi4k2b (C), are expressed in lactotrophs and other hormone-producing cells (somatotrophs, gonadotrophs, thyrotrophs, corticotrophs and melanotrophs) as well as in non-hormonal pituitary cell types (folliculostellate cells, pituicytes, pericytes, and endothelial cells). Among cell types, the percentage of cells expressing these genes varied between 3 and 40, with a mean ± SEM value for all cell types of 18 ± 1.4%. In the lactotroph population, these mRNAs were detected in 14-27% of cells. Among PIP5K1 genes, Pip5k1c was expressed at the highest level, followed by Pip5k1a and Pip5k1b in all cell types, including lactotrophs ( Figure 1D , top panel). Pikfyve and Pip4k2c genes, encoding proteins that can support an alternative synthetic pathway of PI(4, 5)P2, are also expressed in lactotrophs and other pituitary cell types ( Figure 1D , middle panel). Among PI3KC genes, only Pik3ca reached the threshold for detection by scRNAseq, whereas all three regulatory subunit genes (Pik3r1, Pik3r2, and Pik3r3) were detected in lactotrophs and other cell types ( Figure 1D , bottom panel). PI3KCA is tightly associated with a regulatory subunit encoded by the p85alpha gene with three splice forms: p85alpha, p55alpha, and p50alpha (38). However, our scRNAseq system does not discriminate between the various splice forms.

Basal pituitary hormone release comprises regulated exocytosis driven by spontaneous action potential (AP) driven calcium signaling and hormone released by constitutive exocytosis. To study basal hormone release, pituitary cells isolated from adult female rats attached to beads were perfused for 2 h without treatment to wash out previously released hormones, followed by collection of perfusate every 3 min during a 25 min period. This was followed by continuous application of Wm (10 µM in 0.1% DMSO) or solvent (0.1% DMSO) till the end of the experiment ( Figure 2 ). Basal PRL, GH, and LH release, measured from the same samples by ELISA, was: 69 ± 3, 18 ± 2, and 5 ± 0.5 ng/ml, respectively. Application of 10 µM Wm caused a rapid decline in PRL release, with an estimated τ₅₀ of ~20 min ( Figure 2A ), whereas basal GH (B) and LH (C) release were not affected by this inhibitor.

Immunofluorescence analysis revealed no significant differences in the percentage of lactotrophs during 3 h incubation with 10 µM Wm, when compared to cultures treated with solvent: controls, 1-3 h = 33.7 ± 2.1 (n = 22); Wm, 1 h = 35.7 ± 1.77 (n = 7), Wm, 2 h = 34.9 ± 3.2 (n = 7), and Wm, 3 h = 32.1 ± 2.7 (n = 8). Figure 3 shows PRL-positive cells in controls and cells treated with Wm for 3 h. Therefore, the cell-type specific action of Wm on basal hormone release in cultured pituitary cells was not caused by selective loss of lactotrophs to Wm but was a genuine effect on PRL secretion.

In further experiments, cultured cells in static incubations were treated with various Wm concentrations for 3 h, and PRL was measured in medium by RIA. Figure 4A illustrates that Wm inhibits basal PRL release in the 1 – 10 µM concentration range, with an estimated IC₅₀ of ~5.5 µM, and was ineffective in the nanomolar concentration range. In perfused pituitary cells, Wm was ineffective at 0.1 µM but inhibited basal PRL release when administrated at 1 and 10 µM ( Figure 4B ). Therefore, inhibition of basal PRL release induced by Wm at concentration of 10 µM was detected by both methods, ELISA and RIA, confirming the specificity of PRL measurements. The inhibitory Wm action on the basal PRL release was consistently observed, as shown in Figures 4C , 5B, C , 7C, D , and 9D . Basal PRL release in pituitary cells derived from adult male rats was also inhibited by 10 µM Wm (data not shown).

Wm inhibits Class I PI3Ks in the low nanomolar concentration range (19) and PI4Ks in the low to high micromolar concentration range (20). Among the four known PI4Ks, only PI4KA and PI4KB are inhibited by Wm (20). PI4KA is specifically inhibited by GSK-A1 and PI4KB by PIK93 ( Figure 1A ). In static cultures of anterior pituitary cells, GSK-A1 inhibited basal PRL release in a concentration-dependent manner, with an IC₅₀ of ~10 nM ( Figure 5A ). In perfused pituitary cells, the rate and level of inhibition of basal PRL release by 100 nM GSK-A1 was highly comparable to the inhibition induced by 10 µM Wm ( Figure 5B ). In contrast, 1 µM PIK93 was not able to replicate the inhibitory action of Wm on basal PRL release ( Figure 5C ). Likewise, the PI3K inhibitor LY294002, did not affect basal PRL release in perfused pituitary cells when applied in 1 and 50 µM for 35 min ( Figure 5D ). Furthermore, ISA2011B (10 µM) and UNC3230 (1 µM), inhibitors of PIP5K1A and PIP5K1C, respectively (26, 27), were ineffective to inhibit basal PRL release when applied alone or together during 2 h ( Figure 5E ) and 3h ( Figure 5F ) incubations. In the same experiments, GSK-A1 significantly (P < 0.01) inhibited PRL release during 2 and 3 h of incubation. Therefore, the inhibitory effects of Wm and GSK-A1 at such concentration ranges suggest a role of PI4KA, but not PI4KB, PI3Ks or PIP5Ks, in basal PRL release, reflecting the effects of the drug on the exocytotic pathway and/or on de novo PRL synthesis.

To elucidate whether the effects of Wm and GSK-A1 on basal PRL release reflect a direct action on the exocytotic pathway or an indirect action through blockade of de novo PRL synthesis, we performed three experiments. In the first experiment, we compared the rates of effects of Wm and cycloheximide, a protein synthesis inhibitor, on basal PRL release in perfused pituitary cells. These experiments, shown in Figure 4C , illustrate that the rate of Wm-induced reduction in basal PRL release was faster than the rate of reduction induced by cycloheximide, treatment, suggesting that the exocytotic pathway, rather than PRL synthesis, was blocked by Wm. In the second experiment, we examined the time-course of 10 µM Wm and 100 nM GSK-A1 effects on PRL release ( Figure 6A ) and intracellular PRL content ( Figure 6B ) in static incubation. Both compounds inhibited basal PRL release in a similar time-dependent manner, but the blockade of PRL release was accompanied by increase in intracellular PRL cell content. In third experiment, we examined Prl expression ( Figure 6C ) in Wm- (top) and GSK-A1 (bottom) treated cells. Neither treatment significantly affected the expression of this gene during 1, 2, or 3 h incubation. These results indicate that inhibition of PI4KA blocks the exocytotic pathway without affecting de novo PRL synthesis.

In theory, inhibition of PI4KA could cause blockade of spontaneous voltage-gated calcium influx, leading to inhibition of PRL release, or may inhibit the exocytotic pathway downstream of voltage-gated calcium signaling. To clarify this issue, we facilitated voltage-gated calcium influx by the addition of 1 µM Bay K8644, an L-type calcium channel agonist ( Figure 7A ), or by depolarization of the cells by 25 mM KCl ( Figure 7B ) in control cells (left panels) and cells pretreated with 10 µM Wm for 45 min (right panels). In the presence of Bay K8644 or high potassium, 100 nM TRH was added at the end of recording to estimate the potential effects of Wm on calcium mobilizing action of this agonist, which is dependent on the presence of PI (4,5)P₂ in the plasma membrane (39). These experiments revealed that stimulation of calcium influx was not affected by 45 min application of 10 µM Wm. The experiments further showed no difference in the pattern of TRH-induced calcium mobilization in Wm-treated cells ( Figures 7A, B ). This is an important finding, as it indicates that Wm treatment did not decrease PI (4,5)P2 levels, in agreement with findings in other cell types that Wm and GSK-A1 treatments do not decrease PI (4,5)P2 levels in the absence of phospholipase C activation (40).

Consistent with calcium-secretion coupling in lactotrophs, facilitation of voltage-gated calcium influx by Bay K8644 rapidly increased PRL release in perfused control cells, followed by a prolonged decay in secretion to basal levels ( Figure 7C ). Depolarization of control cells by high potassium also increased PRL release but was followed by a faster return to basal secretion levels ( Figure 7D ). However, in cells treated with 10 µM Wm there was a progressive decay in PRL release, and application of Bay K8644 only slightly recovered PRL release, with a peak in response below the basal levels in controls ( Figure 7C ). High potassium was practically ineffective in elevating PRL release in Wm-treated cells ( Figure 7D ).

Similarly, 100 nM GSK-A1 administrated over 45 min did not affect Bay K8644-stimulated voltage-gated calcium influx and TRH-induced calcium mobilization. ( Figure 8A ). TRH receptors are expressed in lactotrophs and thyrotrophs, but only lactotrophs respond to application of dopamine administration ( Figure 8A , left). In contrast to calcium signaling, GSK-A1 dramatically reduced Bay K8644-stimulated PRL release in perfused pituitary cells ( Figure 8B , left). When Wm and Bay K8644 were added at the same time, however, no inhibition of the early spike of PRL release was observed ( Figure 8B , right), suggesting that blockade of calcium-secretion coupling takes time to develop, likely because it takes time for PI4P levels to decrease after blockade of PI4KA (41).

It has been well established that TRH-stimulated calcium mobilization from the intracellular pool increases the release of PRL in vitro (18). As mentioned earlier, TRH-stimulated calcium mobilization was not affected by prolonged (45 min) treatment with 10 µM Wm and 100 nM GSK-A1 ( Figures 7 and 8 ). In general, TRH-induced calcium signaling consist of an early spike response, which reflects calcium mobilization from intracellular pools, and sustained calcium transients, reflecting extracellular calcium influx (18, 39). Both calcium signaling phases were unaffected in cells treated with GSK-A1 for 45 min ( Figures 9 A, B ). Dopamine abolished sustained intracellular calcium transients, confirming that they were driven by voltage-gated calcium channels and that the selected cells were lactotrophs ( Figures 9A, B ). Additional administration of TRH in the presence of dopamine shows the independence of the spike phase from calcium influx in the controls ( Figure 9A ) and that the blockade of PI4KA does not affect the repeated mobilization of calcium from intracellular stores ( Figure 9B ).

Figure 9C shows that the profile of TRH-induced PRL release in perfused control cells is biphasic, composed of an early spike phase and a sustained plateau phase, and is therefore comparable to the TRH-stimulated calcium signaling profile. In contrast to calcium signaling, however, TRH-induced PRL release was dramatically reduced in GSK-A1 ( Figure 9C ) and Wm ( Figure 9D ) treated cells. Administration of 1 µM PIK93, a PI4KB inhibitor, was ineffective ( Figures 9C, D ). Therefore, functional PI4KA is also critical for the release of PRL induced by calcium mobilization.