Opposite Effects of Diacylglycerol and Phosphatidic Acid in the Modulation of the Plasma Membrane Ca²⁺-ATPase from Kidney Proximal Tubules: A Regulatory Role for Diacylglycerol Kinase in Calcium Homeostasis?

Abstract

Background/Objectives: Kidney proximal tubules reabsorb up to 70% of water and solutes from the glomerular ultrafiltrate, a Ca²⁺-modulated process essential for homeostasis. The plasma membrane Ca²⁺-ATPase (PMCA) in basolateral membranes (BLMs) plays a pivotal role in maintaining intracellular calcium homeostasis and regulating calcium reabsorption. Methods: Here, we investigated the regulatory influence of two key bioactive lipids, diacylglycerol (DG) and phosphatidic acid (PA), on PMCA activity from pig kidney, accompanied by lipidomic assays and transcriptomic data analyses. Results: Biochemical assays revealed dose- and time-dependent inhibition of PMCA by DG, fully reversed by Calphostin C, implicating PKC activation. Conversely, PA significantly stimulated PMCA activity, demonstrating an opposite regulatory effect. Our targeted lipidomics identified multiple DG species in HK-2 cells, suggesting substrate diversity. Analysis of transcriptomic data for hypoxic versus normoxic HK-2 cells revealed dramatic coordinated regulation of DG/PA metabolism genes, with upregulation of DG-producing enzymes (PLCB1, PLDs) and downregulation of DG-consuming kinases (DGKs), predicting enhanced DG accumulation under metabolic stress. ATP2B4 (PMCA4) upregulation indicated compensatory transcriptional responses. Conclusions: Our findings suggest that DG inhibits BLM-associated PMCA via classic and/or atypical PKC-dependent phosphorylation while PA exerts opposing stimulatory effects. Both transcriptional remodeling and post-translational modifications regulate this axis. These findings highlight the DG–Diacylglycerol Kinase–PA axis as a dynamic modulator of Ca²⁺ signaling in the kidney that responds to metabolic stress.

1. Introduction

Calcium (Ca²⁺) is a ubiquitous and versatile second messenger, critically involved in regulating various physiological processes, including muscle contraction [1], neurotransmission [2], secretion [3], cell division [4], differentiation [5], and apoptosis [6]. This versatility stems from the ability of Ca²⁺ to interact with numerous calcium-binding proteins [7] and effector proteins/enzymes, such as the calcium-dependent kinases, PKC [8], and phospholipases [9], which are the most studied. This Ca²⁺-dependent molecular machinery enables the fine-tuned modulation of cell signaling cascades present in virtually all cell types [10].

The intracellular Ca²⁺ concentration is typically maintained in the low nanomolar range (10⁻⁷–3 × 10⁻⁷ M), providing a steep electrochemical gradient between the extracellular milieu and the cytosol. Even a minimal ionic influx produces a rapid and transient increase in intracellular Ca²⁺ [11]. Thus, Ca²⁺-mediated intracellular signaling processes are highly localized and temporally precise. In kidneys and other metabolically active tissues, these dynamics are essential for regulating transport, signaling, and survival pathways.

Dysregulation of intracellular Ca²⁺ homeostasis, as seen in pathological conditions such as metabolic syndrome, obesity, and renal disease, can lead to mitochondrial overload, endoplasmic reticulum stress, and impaired signaling fidelity thereby contributing to disease progression and to lack of vital substances that despite being absorbed through transepithelial fluxes within the nephrons, remains in the filtrate and are driven to urinary excretion [12,13,14].

Since multiple pathways and cellular mechanisms are triggered by Ca²⁺, the ion is rapidly buffered and sequestered by cytosolic proteins, allowing for spatially precise signaling events [10]. In this context, calcium active transporters such as the plasma-membrane Ca²⁺-ATPase (PMCA) and the endoplasmic reticulum Ca²⁺-ATPase (SERCA) are responsible for pumping the ion to the extracellular space [15] or the reticulum [16], helping to fine-tune intracellular Ca²⁺ levels.

The plasma membrane Ca²⁺-ATPases (PMCAs) are high-affinity ATP-driven Ca²⁺ efflux pumps belonging to the P-type ATPase family that maintain cytosolic Ca²⁺ at submicromolar levels across virtually all eukaryotic cells [17,18]. In mammals, four PMCA isoforms (PMCA1–4) encoded by distinct genes (ATP2B1–4) undergo extensive alternative splicing, generating variants with distinct regulatory properties and subcellular targeting preferences [19]. PMCA1 and PMCA4 are broadly expressed “housekeeping” isoforms, whereas PMCA2 and PMCA3 exhibit more tissue-restricted distributions, often enriched in excitable cells [20]. In polarized renal epithelial cells, isoform-specific targeting has been demonstrated, with PMCA1 and PMCA4 predominantly localizing to the basolateral domain [21,22].

Previous work from our group demonstrated that in proximal tubule cells, PMCA activity is functionally compartmentalized in caveolae of the basolateral membrane (BLM), where its integrity is required for pump activity [23]. We also found that ceramide—a key bioactive lipid in membrane microdomains—modulates the pump via spatially and temporally distinct activation of protein kinases [24].

Renal proximal tubule cells can be activated by different hormones, autacoids, and other bioactive molecules, leading to the activation of a complex regulatory machinery harbored in the basolateral membranes that includes ion channels, transporters, receptors, protein kinases, and phospholipases [25,26,27,28,29]. Due to phospholipase C (PLC) activation, phosphoinositide turnover leads to a local rise in membrane DG content, which may be the main factor responsible for different PKC-dependent ion transport modulation. To buffer the elevated DG levels, we had previously proposed the action of a diacylglycerol kinase (DGK) potentially responsible for converting DG into phosphatidic acid (PA) [30]. This pivotal role of DGK switching from DG to PA is the main signaling axis explored in the current work. We aim to demonstrate that both DG and PA can modulate the PMCA in opposing ways. This would represent a fascinating signaling rheostat that not only helps maintain DG levels within a physiological range but also leads to the production of PA. PA may act as a signaling lipid itself or serve as a precursor for various glycerolipids and lyso-glycerolipids, orchestrating, at least in part, the lipid-mediated control of ion fluxes across the BLM from kidney proximal tubule cells. Thus, we aim to evaluate the regulatory effects of DG and PA on PMCA activity in kidney proximal tubules.

2. Materials and Methods

2.1. Material

Buffers, BSA, and protease inhibitors were obtained from Sigma Chemical Co. (St. Louis, MO, USA), and Percoll was from Pharmacia (Peapack, NJ, USA). Distilled water, deionized using the Milli-Q system of resins (Millipore Corp., Burlington, MA, USA), was used to prepare all solutions. Diacylglycerol (DG used Dioleoylglycerol, Cat. No. D8894) which consist in a 99% purity mixture of 1,2- and 1,3-diacylglycerol isomers. The product contains primarily 18:1 (oleic acid) fatty acid chains at both sn-1 and sn-2 positions. Calphostin C from Cladosporium cladosporioides (≥90% HPLC purity, Cat#C6303), phorbol 12-myristate 13-acetate (PMA, Cat. No. P8139) and L-α-Phosphatidic Acid sodium salt from egg yolk lecithin (PA, ≥98% purity, Cat. No. P9511), derived from egg phosphatidylcholine by phospholipase D hydrolysis, containing predominantly palmitic (16:0), stearic (18:0), oleic (18:1) and linoleic (18:2) fatty acid chains were all purchased from Sigma-Aldrich (St. Louis, MO, USA). All other reagents were of the highest purity available.

2.2. Obtention of the Basolateral Membrane Fractions (BLM) from Pig Kidney Proximal Tubules

Adult pig kidneys were directly obtained from a slaughterhouse under the supervision of licensed veterinarians, according to the Brazilian law. The BLM-enriched fractions were obtained as previously described [24,29,31]. Each independent experiment utilized BLM prepared from kidneys pooled from 2–3 adult pigs. For all enzymatic assays, each data point represents one independent membrane preparation (biological replicate), with assays performed in technical duplicate and averaged.

2.3. Cell Cultures

Human renal proximal tubule cells (HK-2, ATCC, Manassas, VA, USA: CRL:2190) were cultured in Keratinocyte Serum Free Medium (KSF-M, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 1% pen/strep (Thermo Fisher Scientific) and 2% Fetal Bovine Serum (FBS, Cultilab, Campinas, Brazil). Cells were maintained under 5% CO₂ at 37 °C to achieve 80–90% confluency prior to the experiments.

2.4. Plasma Membrane Ca²⁺-ATPase Activity Determination

Ca²⁺-ATPase activity was assayed in a reaction medium containing 50 mM BisTris propane buffer (pH 7.4), 125 mM KCl, 5 mM MgCl₂, 5 mM ATP, 10 mM NaN₃, and 0.2 mM EGTA. Free Ca²⁺ concentration was buffered to 0.5 µM using CaCl₂, accounting for Ca²⁺, EGTA, ATP, and Mg²⁺ interactions as calculated by the Mcalc software (Version 1.0.3) [32]. Briefly, BLM (0.2 mg/mL protein) were preincubated with 3 mM ouabain for 30 min at room temperature to inhibit Na⁺/K⁺-ATPase. The reaction was initiated by adding ATP and allowed to proceed for 10 min at 37 °C. The reaction was stopped by adding an equal volume of 0.1 N HCl containing activated charcoal. After centrifugation, inorganic phosphate (Pi) released into the supernatant was quantified using the colorimetric method already described [33]. Ca²⁺-ATPase-specific activity was calculated by subtracting the activity in the presence of 2 mM EGTA. Phosphorylation-dependent modulation of enzyme activity was tested following preincubation with bioactive lipids or inhibitors when specified.

2.5. Determination of Protein Kinase C (PKC) Activity

The PKC activity in basolateral membrane (BLM) fractions from kidney proximal tubules was measured as previously described for kinase activity assays [24,34], using histone H8 as a substrate and [γ-³²P]ATP as the phosphate donor. The reaction medium (100 µL final volume) contained 20 mM Hepes-Tris (pH 7.0), 4 mM MgCl₂, histone H8 (1.5 mg/mL), and BLM protein (0.5 mg/mL). Reactions were initiated by the addition of 10 µM [γ-³²P]ATP (7–10 µCi/µmol) and incubated for 10 min at 37 °C. The reaction was stopped with an equal volume of 40% trichloroacetic acid and filtered through 0.45 µm Millipore filters. Filters were washed with 20% trichloroacetic acid and 2 mM phosphate buffer (pH 7.0), and radioactivity incorporated into histone was quantified by liquid scintillation counting. PKC-specific activity was defined as the difference in ³²P incorporation in the presence or absence of 10 nM calphostin C, a PKC inhibitor. Phorbol-12-myristate-13-acetate (PMA, 1 µM) was used to confirm functional PKC activation.

2.6. Mass Spectrometry Analysis of DG Levels in HK-2 Cells

(a)

Sample preparation: HK-2 cells were washed twice with phosphate-buffered saline (PBS) and then dissociated from the culture flasks with trypsin (0.25%). Cells were centrifuged at 1000 rpm for 5 min to remove trypsin, and then resuspended in 400 μL of methanol and 100 μL of water. The mixture was vigorously vortex-mixed for 10 s, followed by centrifugation at 12,000 rpm for 5 min, and the supernatant was transferred to clean vials. The extracts were taken to dryness under a gentle stream of nitrogen and redissolved in a small volume (50 μL) of acetonitrile, isopropanol, and water (1:2:1) for analyses.

(b)

Liquid chromatography-high resolution mass spectrometry (LC-HRMS) analysis: The LC-HRMS system consisted of a Thermo Scientific UltiMate 3000 LC (Waltham, MA, USA) coupled to a Thermo Scientific Q Exactive Plus Quadrupole-Orbitrap Mass Spectrometer (Waltham, MA, USA) equipped with an electrospray (ESI) ion source. Chromatographic separation was obtained using a Waters CSH C₁₈ column (Milford, MA, USA, 150 mm × 2.1 mm × 2.5 μm) at 45 °C. The mobile phase consisted of (A) water:methanol (95:5, v/v) with 0.2% formic acid and (B) methanol:acetone:water (70:25:5, v/v) with 0.2% formic acid. Separation was achieved with the following gradient of B: 0.0 min, 50% B; 10.0 min, 100% B; 10–15 min, 100% B; 15.1 min, 50% B; 20 min, 50% B. The flow rate was 400 μL/min. The mass spectrometer was operated in positive ion mode (ESI+) with a capillary voltage of 3900 V, S-lens RF level of 50 (arbitrary units, a.u.), and capillary temperature of 320 °C. The sheath and auxiliary gas flow rates (nitrogen) were 45 and 20 (a.u.), respectively. Samples were analyzed in the scan range of m/z 120–1200 at the resolution of 70,000 FWHM (full width at half maximum), followed by data-dependent MS/MS analysis (ddMS2 top3 experiment, normalized collision energy (NCE) 25%) using the resolution of 17,500 FWHM. Data acquisition and analysis were performed on Xcalibur 3.0 software (Thermo Scientific).

(c)

Data Processing and Metabolite Annotation: Raw LC–HRMS data were processed using MS-DIAL version 5.2 (RIKEN, Saitama, Japan) for peak detection, deconvolution, alignment, and metabolite identification [35]. Feature extraction parameters were a minimum peak height of 1.0 × 10⁴ and a mass accuracy of 0.05 Da for MS1 and 0.05 Da for MS/MS. Retention time alignment was performed using a tolerance of ±0.1 min, and isotopic/adduct deconvolution was automatically applied. Metabolite annotation was performed by matching experimental spectra with the internal MS-DIAL in silico lipidomics library, considering a tolerance of 10 ppm. The DG lipid class was selected for targeted annotation. The resulting aligned peak table, containing m/z, retention time, and peak area, was exported as a .csv file for downstream statistical analysis.

2.7. Statistical Analysis

Processed data were imported into MetaboAnalyst 5.0 (www.metaboanalyst.ca) for statistical and multivariate analyses. Prior to analysis, missing values were imputed with half the minimum positive value for each variable; data were then normalized by sum, log₁₀-transformed, and auto-scaled. Unsupervised principal component analysis (PCA) was performed to visualize intrinsic clustering patterns and detect outliers.

2.8. Transcriptomic Analysis of HK-2 Cells

The transcriptomic data used in this study were obtained from publicly available datasets deposited in the NCBI’s Gene Expression Omnibus (GEO) database [36]. To validate the physiological relevance of the DG/DGK/PA regulatory axis in human renal proximal tubule cells, we analyzed expression data from the GSE269418 dataset, which profiled HK-2 cells exposed to normoxic (21% O₂) or hypoxic (1% O₂) conditions for 24 h [37]. This dataset was generated using high-throughput RNA sequencing.

Raw count data were downloaded from GEO and processed in R Studio (version 2024.12.1+563). Differential expression analysis was performed using DESeq2 (version 1.46.0) [38], comparing hypoxic versus normoxic conditions. Genes with low counts were filtered (minimum count threshold of 10 reads across all samples), and normalization was performed using the median-of-ratios method implemented in DESeq2. Log₂ fold-change (log₂FC) values and Benjamini–Hochberg false discovery rate (FDR)-adjusted p-values (padj) were extracted for each gene.

A curated gene list was assembled to include key components of DAG metabolism (DGK isoforms, PLDs), PA metabolism (PLPPs, LPIN1), calcium transporters (ATP2B1–4), and lipid-activated signaling kinases (PKC and PLC isoforms). Genes were considered significantly differentially expressed if they met the criteria of |log₂FC| > 0.6 and FDR < 0.05.

For visualization, volcano plots were generated using ggplot2 (version 4.0.0), plotting log₂FC against −log₁₀(FDR), with genes color-coded based on regulation status (upregulated in red, downregulated in blue, non-significant in gray). Additionally, variance-stabilized transformation (VST) values from DESeq2 were Z-score normalized for each gene across samples, and a heatmap was generated using pheatmap (version 1.0.13) with hierarchical clustering (Euclidean distance, complete linkage) applied to genes. All statistical analyses and visualizations were performed in R v4.3.2 (RStudio v2025.09.1 Build 401).

2.9. Statistical Analysis

Statistical analysis was carried out using the One-way ANOVA test and Newman-Keuls post-test. Statistical significance was set at p < 0.05. Data were analyzed using GraphPad Prism 10.1.2 (GraphPad Software, Boston, MA, USA).

3. Results

3.1. Diacylglycerol Inhibits Ca²⁺-ATPase Activity with High Potency and Early Saturation Manner

To determine whether diacylglycerol (DG) modulates Ca²⁺-ATPase (PMCA) activity, BLM-enriched fractions from renal cortex were incubated (25 min) with increasing concentrations of DG (0.05–100 μM as depicted), and enzyme activity was quantified based on inorganic phosphate (Pi) release. As shown in Figure 1, DG significantly inhibited PMCA activity in a high potency and early saturation manner. At 0.05 μM, DG caused a robust and statistically significant reduction in ATPase activity compared to untreated controls (p < 0.005). Maximal inhibition was achieved at this low micromolar concentration, with enzyme activity remaining suppressed at 0.1, 0.5, and 1 μM DG, demonstrating a plateau effect characteristic of high-affinity, receptor-mediated regulation. Notably, at higher concentrations (10 μM and 100 μM), DG did not maintain the same level of inhibition; activity levels in these groups were intermediate between control and maximally inhibited states. This biphasic response pattern suggests that PMCA inhibition by DG operates through a saturable, specific regulatory mechanism rather than non-specific membrane disruption.

The data presented in Figure 1 clearly demonstrate that very low (submicromolar range) DG concentrations are sufficient to maximally inhibit PMCA, and that this inhibition reflects a specific, high-potency regulatory interaction, not a nonspecific lipid effect. Crucially, the absence of inhibition at 10–100 µM argues against nonspecific detergent-like inhibition of PMCA and instead supports the conclusion that the observed effect at low DG concentrations reflects a specific, regulated signaling process.

3.2. Diacylglycerol Inhibits PMCA Activity in a Time-Dependent Manner

To evaluate the kinetics of DG-mediated modulation of Ca²⁺-ATPase activity, we performed a time-course assay using BLM-enriched fractions from porcine proximal tubules. Membranes were incubated in the presence or absence of 0.05 µM DG, and Ca²⁺-ATPase activity was measured by quantifying inorganic phosphate (Pi) released over time. As shown in Figure 2, DG led to a significant reduction in Ca²⁺-ATPase activity over time. Under control conditions, ATPase activity increased steadily with time, reaching ~1300 nmol Pi·mg⁻¹·min⁻¹ at 50 min. In contrast, DG-treated samples showed a markedly lower Pi accumulation, with activity levels plateauing around 850 nmol Pi·mg⁻¹·min⁻¹ over the same period. The difference in enzyme activity between the groups became statistically significant at 30 and 50 min (p < 0.05, Mann–Whitney test), suggesting a cumulative inhibitory effect over time. The inset in Figure 2 highlights the early kinetic window (0–1 min), where both groups show overlapping activity levels, indicating that DG does not act as an immediate or competitive inhibitor, but rather exerts a delayed effect potentially linked to membrane remodeling, enzyme re-localization, or signal-dependent modulation. Taking together these two initial experimental sets, it is important to mention that Figure 2 defines the time dependence of DG-induced PMCA inhibition, while Figure 1 captures the concentration sensitivity of this effect at the time point (25 min) where inhibition is statistically and biologically evident.

3.3. PKC Inhibition by Calphostin C Prevents DG-Induced Suppression of Ca²⁺-ATPase Activity

To evaluate whether the DG-mediated inhibition of PMCA activity involves PKC activation, BLM-enriched fractions from kidney cortex were incubated with 0.05 µM DG in the presence or absence of Calphostin C (1 nM), a selective inhibitor of DG-sensitive PKC isoforms. As shown in Figure 3, treatment with DG alone significantly reduced Ca²⁺-ATPase activity relative to control membranes, as already demonstrated (p = 0.0459). However, co-treatment with Calphostin C fully prevented the DG-induced inhibition, restoring PMCA activity to levels statistically indistinguishable from the control. Notably, Calphostin C alone had no significant effect on enzyme activity, indicating that PKC inhibition does not modulate basal Ca²⁺-ATPase levels under our experimental conditions. These results suggest that PKC activation is required for DG-mediated suppression of PMCA activity, consistent with a post-translational regulatory mechanism involving kinase-dependent phosphorylation of the pump.

3.4. PMCA Activity Is Increased After Incubation of the BLM-Enriched Fractions with PA

Although PA cannot be directly considered a bioactive lipid, there are some reports in the literature ascribing modulatory effects on different cellular events. PA is recognized as the common precursor for the synthesis of the glycerolipids. It can be produced through the action of a previously described DGK found in the BLM, which phosphorylates DG to form PA [30]. The results shown above indicate that DG-dependent PKC may modulate the PMCA. Therefore, we investigated whether the conversion of DG to PA would alter this modulation. Figure 4A demonstrates that incubating the BLM-enriched fraction with different concentrations of PA as depicted in the abscissa led to a significant increase in PMCA activity. This effect in the pump contrasts sharply with the decrease in activity observed when the membranes were treated with DG. The modulatory effect of PA was not influenced by pre-incubating the BLM-enriched fractions with calphostin-C, as observed for DG treatment. To confirm the functional presence of PKC in our BLM preparations, we directly measured kinase activity using the DAG-mimetic phorbol ester PMA (1 μM). As shown in Figure 4B, PMA treatment induced a robust and highly significant increase in PKC activity (p < 0.0001), validating that functional PKC is present and responsive in our membrane fractions. This supports the conclusion that DG-mediated PMCA inhibition occurs through PKC activation in the BLM, different from the observed for PA. This experimental set indicates that DGK may play a pivotal role in buffering local increases in DG levels, PKC activation, and ion fluxes through the renal epithelia.

3.5. Targeted Lipidomic Analysis Showed Different DG Species in Human Renal Proximal Tubule Cells

To better characterize the modulatory potential of the DG/DGK/PA axis, we decided to use living proximal tubule cells (HK-2 lineage) to targeted lipidomic for DG.

Table 1. Annotated diacylglycerol (DG) species, corresponding m/z [M + Na]⁺ values, and chromatographic retention times (t_r). This table summarizes the DG lipid species annotated identified by LC-HRMS/MS. For each DG species, the measured mass-to-charge ratio (m/z) and the chromatographic retention time (t_r, in minutes) were reported. DG species were annotated according to the total number of carbons and double bonds in the combined acyl chains. The DG specie was annotated by its retention time, mass-to-charge ratio (m/z) and fragmentation spectrum.

DG Species	m/z [M+Na]+	TR
DG 32:1	589.480	13.69
DG 32:0	591.496	14.84
DG 34:1	617.511	14.89
DG 34:0	619.527	15.71
DG 36:3	641.511	14.14
DG 36:2	643.527	14.98
DG 36:1	645.542	15.75
DG 36:0	647.558	16.26
DG 38:4	667.527	14.84
DG 38:3	669.543	15.51
DG 38:1	673.573	16.29
DG 40:4	695.558	15.37

shows the most abundant DG lipid species, which were identified by LC-HRMS/MS. For each DG species, the measured mass-to-charge ratio (m/z) and the chromatographic retention time (tR, in minutes) are provided. DG species were annotated according to the total number of carbons and double bonds in the combined acyl chains. Each specific DG species was annotated by its respective retention time, mass-to-charge ratio (m/z) and fragmentation spectrum. Figure 5 depicts the different chromatograms for targeted lipidomic analysis of diacylglycerol (DG) species from HK-2 cells. Figure 5A is a total ion chromatogram from a typical proximal tubule renal cell cultivated under standard conditions. Figure 5B–M shows the extracted ion chromatogram from each DG species detected under the conditions and procedures described under Section 2.

3.6. Transcriptomic Profiling Reveals Significant Regulation of DG/PA Metabolism and Calcium Transport Genes in Hypoxic Proximal Tubule Cells

Our biochemical assays demonstrated that DG and PA modulate Ca²⁺-ATPase activity in opposing directions within the BLM of proximal tubules. To establish the physiological context and validate the relevance of this regulatory axis, we analyzed publicly available transcriptomic data from HK-2 cells—a widely used human proximal tubule cell model—exposed to hypoxic (1% O₂) versus normoxic (21% O₂) conditions for 24 h (GSE269418) [37]. Hypoxia is a physiologically relevant stressor in the kidney, particularly in the context of ischemia–reperfusion injury, chronic kidney disease, and metabolic stress [12,39].

Figure 6B presents a volcano plot illustrating genome-wide differential expression, with genes related to DAG/PA metabolism, PKC signaling, PLC isoforms, and PMCA isoforms (ATP2B family) highlighted. The analysis revealed robust and highly significant regulation of multiple key players in this signaling network. Most strikingly, PLCB1—the enzyme responsible for DG generation from phosphoinositide hydrolysis—exhibited the most dramatic upregulation (log₂FC = 1.70, FDR = 6.82 × 10⁻¹⁰⁹), indicating substantially enhanced capacity for DG production under hypoxic stress. This finding directly supports our biochemical observation that DG exerts potent inhibitory effects on PMCA activity (Figure 1). Complementing this, PLD1 and PLD5 were significantly upregulated (log₂FC = 0.82 and 2.34, respectively; FDR < 10⁻³⁵), supporting both phosphatidylcholine-derived DG production and subsequent PA formation pathways.

Conversely, enzymes responsible for DG-to-PA conversion showed coordinated downregulation: DGKQ and DGKG exhibited strong suppression (log₂FC = −0.92 and −1.09; FDR < 10⁻⁵⁷), suggesting reduced capacity to metabolize DG under hypoxia. Combined with elevated DG production via PLCB1, this would result in local DG accumulation, potentially amplifying PKC-mediated PMCA inhibition as observed in our biochemical experiments. The regulation of PA-metabolizing enzymes was complex, with PLPP6 downregulated (log₂FC = −1.09, FDR = 1.53 × 10⁻²²) and PLPP3 upregulated (log₂FC = 0.75, FDR = 6.09 × 10⁻⁵), suggesting dynamic remodeling of PA pools during hypoxic stress.

PKC isoforms displayed differential regulation consistent with isoform-specific roles: PRKCD (PKC-delta) was significantly upregulated (log₂FC = 0.61, FDR = 1.75 × 10⁻²⁸), while PRKCB was downregulated (log₂FC = −0.98, FDR = 3.76 × 10⁻¹²), potentially fine-tuning the balance between PMCA activation and inhibition. Notably, ATP2B4 (PMCA4) showed significant upregulation (log₂FC = 0.70, FDR = 6.99 × 10⁻¹²), likely representing a compensatory transcriptional response to maintain calcium homeostasis when DG-mediated inhibition is enhanced, while ATP2B1 (PMCA1) showed no significant upregulation (log₂FC = 0.23, FDR = 2.88 × 10⁻¹⁰). Finally, ENPP2 (autotaxin) was significantly downregulated (log₂FC = −0.84, FDR = 2.27 × 10⁻⁹). Given our previous demonstration that LPA inhibits PMCA via PKC activation [29], reduced ENPP2 expression may represent an adaptive response limiting excessive PMCA inhibition during hypoxic stress.

Figure 6C provides a comprehensive statistical summary of all analyzed genes, demonstrating the robustness of these findings; Figure 6A displays Z-score normalized expression values across replicates, illustrating consistent patterns and clear separation between experimental conditions. Together, these transcriptomic data provide strong support for the physiological relevance of the DG/DGK/PA axis in regulating calcium homeostasis in human proximal tubule cells, with the coordinated upregulation of DG-producing enzymes and downregulation of DG-consuming enzymes predicting elevated local DG levels under hypoxia—precisely the condition that would enhance DG-mediated PMCA inhibition observed in our biochemical assays.

4. Discussion

The basolateral membrane (BLM) from kidney proximal tubule cells harbors a plethora of cellular signaling systems, including receptors, protein kinases and protein phosphatases, phospholipases, bioactive and precursor lipids strategically co-located with their regulatory targets. Previous works from our group had already demonstrated the importance of bioactive lipids in this scenario. Here, we provide new evidence to include the diacylglycerol (DG)/diacylglycerol kinase (DGK)/phosphatidic acid (PA) axis in this complex regulatory network present in the BLM.

Among the different renal functions, regulating ion and water reabsorption is particularly important. These processes are closely linked to the polarization of tubular cells and the differential distribution of ion transporters in the apical and basolateral membranes [31,40]. Calcium homeostasis is critical for both the entire organism and kidney epithelial cells because this ion modulates diverse cellular events. Thus, in kidney proximal tubule cells, PMCA plays a dual role: while the larger proportion of filtered Ca²⁺ is reabsorbed paracellularly, PMCA is essential for complementary active, transcellular Ca²⁺ reabsorption and, critically, for the fine-tuning of cytosolic Ca²⁺ concentrations during cell signaling events [10,23,41].

The PMCA activity is tightly regulated by a wide array of mechanisms, including Ca²⁺/calmodulin binding, interaction with acidic phospholipids, phosphorylation by protein kinases, limited proteolysis (e.g., by calpain), PDZ-domain-dependent protein scaffolding, and oligomerization [19,20]. These regulatory inputs primarily target the C-terminal cytosolic tail, which acts as an autoinhibitory domain relieved upon calmodulin binding or alternative splicing events. Such structural and regulatory versatility enables PMCA to participate not only in calcium extrusion but also in the compartmentalized control of calcium microdomains, essential for epithelial transport, hormonal signaling, and protection against Ca²⁺-induced cytotoxicity.

Disturbances in intracellular Ca²⁺ homeostasis are definitively recognized as critical drivers of kidney pathology. In polycystic kidney disease, the elevated basal Ca²⁺ levels in renal epithelial cells contribute to increased cell proliferation and cyst growth [42,43]. In ischemia/reperfusion injury, mitochondrial Ca²⁺ overload triggers oxidative stress, ATP depletion, and apoptosis, while enhancing Ca²⁺ transport across mitochondrial membranes reduces Ca²⁺ accumulation and alleviates renal tissue damage [12,39]. Furthermore, ER stress and mitochondrial apoptosis, both Ca²⁺-dependent events, have been implicated in nephrotoxicity induced by contaminants such as 3-MCPD, where Ca²⁺ dysregulation promotes cell death in proximal tubule cells [44]. In addition, lysosomal Ca²⁺ release plays a protective role in crystal nephropathies, where the transcriptional activation of TRPML1 promotes lysosomal clearance and attenuates renal tubular injury [45,46]. Lastly, the integrity of Ca²⁺ signaling is fundamental to preventing glomerular endothelial dysfunction in sepsis-associated acute kidney injury (AKI), where TRPV4-mediated Ca²⁺ influx exacerbates inflammation and barrier disruption [43].

The expression of PMCA1 and PMCA4 in murine kidneys and immortalized cells is also broad, while PMCA2 and PMCA3 have a more distal and inner medullary expression pattern [21,47]. Functional studies in MDCK cells and immortalized proximal tubule cells support the significant contribution of PMCA1 and PMCA4 to transepithelial Ca²⁺ flux, alongside or in place of the Na⁺/Ca²⁺ exchanger [21,45]. Together, these findings highlight the crucial role of PMCA isoforms, particularly PMCA1 and PMCA4, in maintaining renal calcium homeostasis under physiological and pathological conditions.

A large body of work shows that PMCA function is influenced by (i) lipid–protein interactions in the membrane environment and (ii) phosphorylation by kinases, including PKC, with outcomes that depend on PMCA isoform/splice variant, phosphorylation site usage, and membrane microdomain context. Comprehensive reviews and mechanistic syntheses emphasize that PMCA is a highly regulatable Ca²⁺ extrusion system and that kinase-mediated control is not uniform across biological settings [48,49,50,51]. Our concentration–response experiments revealed that DG inhibits PMCA activity in a saturable manner, with maximal inhibition plateauing at concentrations of 0.05 µM and above. The lack of further inhibition at higher concentrations (10–100 µM) strongly suggests a functional, non-toxic, receptor-like or lipid/protein interaction rather than an unspecified detergent-like effect on the BLM. This saturable inhibition profile resembles the isoform-specific activation thresholds of DG-sensitive proteins, where classic/conventional and novel PKC isoforms show differential responses to DG molecular species in the low nanomolar to micromolar range [51]. These findings support the idea that only a subset of DG-responsive effectors is engaged at low concentrations, with higher levels failing to enhance the response any further, possibly due to lipid-binding saturation, allosteric constraints, or negative feedback mechanisms. Conversely, the canonical view is that DG activates PKC, which can phosphorylate PMCA and often increase its activity in several cell types and isoforms (notably PMCA4b) in vascular/other contexts [48,49,52,53,54,55]. At the same time, the literature also supports that PKC signaling can yield different functional outcomes depending on system-level determinants (isoform composition, regulatory domain exposure, basal calmodulin tone, and membrane lipid architecture). This is one reason why the field increasingly frames PMCA regulation as context- and compartment-specific rather than unidirectional [10,49,50,56].

The results presented here support the hypothesis that DG modulates BLM PMCA activity via specific post-translational mechanisms, potentially through the activation of DG-sensitive kinases, such as PKC classic and novel isoforms, already shown in the BLM from pig proximal tubules [57], which are abundantly expressed in kidney cells and capable of phosphorylating PMCA [8,10,29,31,49,50]. On the other hand, the conversion of DG to phosphatidic acid (PA) by DG kinase (DGK) could alter the local lipid environment and indirectly affect PMCA conformation, anchoring, and thus, activity. Independent studies demonstrate that PA can interact with and activate PMCA, including biochemical evidence that PA alters PMCA’s lipid exposure and conformational environment, consistent with a direct stimulatory role [58,59,60]. Our PA-stimulatory results presented here more prominently and explicitly align with this view, emphasizing that our data support the established PA→PMCA activation paradigm while adding kidney membrane-specific context.

It is worth emphasizing that the saturable inhibition profile observed in our Figure 1, with maximal effect at 0.05 µM DG and plateau through 1 µM, is consistent with specific engagement of DG-binding proteins rather than non-specific membrane effects. The reduced inhibition at very high DG concentrations (10–100 µM) likely reflects: (i) sequestration of DG into micelles or non-productive membrane domains, (ii) disruption of the caveolar microdomains essential for PMCA regulation [23], or (iii) compensatory activation of protective signaling pathways. This concentration-response pattern resembles the bell-shaped curves reported for PKC activation by DG and phorbol esters [8,51,56], where optimal signaling occurs within a narrow concentration window. Importantly, the effective concentration range (0.05–1 µM) aligns with estimated local DG concentrations in activated membrane microdomains following phospholipase C stimulation, supporting the physiological relevance of our findings.

The preservation of PMCA activity by Calphostin C pre-treatment strongly supports the view that DG inhibits BLM-Ca²⁺-ATPase activity through a PKC-dependent pathway (Figure 3). This is in line with the canonical role of DG as a second messenger that activates conventional and novel PKC isoforms [8,51,52,56] several of which (e.g., PRKCA, PRKCB, PRKCE) were confirmed to be robustly expressed in the presented transcriptomic analysis of both HK-2 and HPTEC cells (Figure 6). Pre-incubation of BLM fractions with Calphostin C alone does not alter PMCA activity, suggesting that PKC is not constitutively active under basal conditions, but becomes engaged upon DG stimulation.

These findings reinforce the hypothesis that the regulation of PMCA in the proximal tubule is dynamic and lipid-sensitive, integrating metabolic cues such as DG accumulation into a signaling cascade that fine-tunes calcium handling. This is especially relevant under conditions like hypoxia, where the transcriptomic data showed preserved expression of both DG-metabolizing enzymes and PMCA isoforms, supporting a model where post-translational signaling rather than transcriptional remodeling governs Ca²⁺-ATPase modulation (Figure 6).

The interplay between DG and PA resembles a crucial regulatory axis in lipid-mediated signaling, particularly in epithelial tissues like the kidney proximal tubule. DG is generated via phospholipase C (PLC)-mediated hydrolysis of glycerolipids, mainly phosphoinositides, and serves as a prototypical lipid second messenger that activates a wide array of effectors, including conventional and novel PKC isoforms, protein kinase D (PKD), and other C1-domain containing proteins [27,57,61,62].

To prevent sustained DG accumulation and terminate PKC signaling, DG can be phosphorylated by diacylglycerol kinases (DGKs), forming PA—another bioactive lipid with distinct signaling roles such as modulating mTOR, PI4P5K, Ras, and phospholipase D activity [30,62,63,64]. Notably, recent evidence has demonstrated that DGK isozymes selectively utilize distinct molecular species of DG, often derived from phosphatidylcholine-specific PLC pathways, suggesting a phosphoinositide turnover-independent regulation of PA production that could confer unique functional outputs depending on the cellular context [51]. In fact, our targeted lipidomic analysis for DG revealed a great variety of DG molecular species (Table 1), which strongly suggest simultaneous DG formation from different glycerolipids pools resulting in the heterogeneous composition of cellular DG (Figure 6), as already demonstrated by other authors [65,66,67]. The lipidomic analysis defines DG pool remodeling, not single-species causality; the exogenous DG represents a mechanistic probe of DG–PKC signaling capacity; and the biological insight lies in demonstrating that DG enrichment—contrasting with PA—drives PKC-dependent PMCA inhibition in renal membranes, irrespective of precise acyl-chain composition. The balance between DG and PA pools, rather than the dominance of a single DG species, is a critical determinant of PMCA regulation in proximal tubule membranes and highlights the role of diacylglycerol kinase in such an important renal fate. This lipid rheostat model would not be evident without integrating species-resolved lipidomics with targeted functional assays, as done here.

As a limitation of the study, we could not test the different DG species in the experiments shown in Figure 1, Figure 2, Figure 3 and Figure 4. Further studies are necessary to clarify the specificity of these different DG species to PMCA modulatory activity and other PKC-triggered actions.

In the BLM, a unique DGK activity has already been identified, displaying insensitivity to Ca²⁺ and resistance to R59949 (a classical DGK inhibitor), while being potently inhibited by sphingosine. These characteristics allowed us to suggest the presence of a non-classical DGK isoform, potentially DGKθ, that operates independently of intracellular Ca²⁺ signals [30]. Intriguingly, this DGK appears functionally coupled to membrane-bound PLC, enabling rapid PA production following DG generation. This coupling might spatially constrain DG availability and control PKC activation, particularly in caveolin-rich lipid microdomains where PMCA is localized [23,30,31,61]. The downstream effects of DG and PA interconversion extend to the regulation of membrane calcium transporters.

The PMCA activity in renal BLM is not only inhibited by DG-mimicking stimuli (e.g., phorbol esters) but also potently modulated by lysophosphatidic acid (LPA), a PA-derived metabolite. Our group had recently demonstrated that LPA activates G protein-coupled receptors, initiating a PLC–PKC signaling cascade that results in inhibitory phosphorylation of PMCA [29]. Conversely, ceramide, another bioactive lipid present in the same membrane domain, exerts a stimulatory effect on PMCA via a protein kinase A (PKA)-dependent phosphorylation pathway, counteracting PKC-mediated inhibition and supporting calcium clearance under stress conditions [24]. These findings highlight how the DG/PA axis and associated kinases establish a tightly regulated lipid signaling environment that directly affects PMCA activity and, consequently, calcium homeostasis in proximal tubule cells. This biochemical network may be particularly relevant during episodes of ischemia or hormonal stimulation, where phosphoinositide turnover and lipid kinase activity are heightened. The spatial and temporal integration of these signals in BLM microdomains provides a mechanistic basis for lipid-kinase-mediated fine-tuning of calcium transporter function in the kidney.

The transcriptomic analysis presented in Figure 6 strongly validates our biochemical findings, revealing that genes encoding DG-metabolizing enzymes, PA-producing pathways, PKC isoforms, and PMCA pumps undergo significant transcriptional remodeling in response to hypoxic stress. Analysis of the GSE269418 dataset showed dramatic upregulation of PLCB1, the key enzyme generating DG from phosphoinositide hydrolysis, suggesting hypoxic proximal tubule cells substantially increase DG production capacity. This would enhance PKC activation and subsequent PMCA inhibition, exactly as observed in our biochemical assays. Importantly, the coordinated downregulation of DGKQ and DGKG indicates reduced DG-to-PA conversion during hypoxia. Combined with elevated DG production via PLCB1, this would result in local DG accumulation in membrane microdomains where PMCA resides [23,30], amplifying the DG-PKC-PMCA inhibitory axis and providing a mechanistic explanation for calcium overload in ischemic kidney injury [12,39].

Conversely, significant upregulation of PLD1 and PLD5 represents an alternative PA generation pathway independent of DGK, as PLDs hydrolyze phosphatidylcholine directly to PA [64]. Given our findings that PA stimulates PMCA activity (Figure 4), PLD-derived PA could serve as a protective mechanism counterbalancing DG-mediated inhibition. The differential regulation of PKC isoforms—PRKCD upregulated while PRKCB was downregulated—further supports isoform-specific PMCA regulation reflecting distinct subcellular localizations and substrate specificities [8,51]. Most notably, ATP2B4 (PMCA4) upregulation likely represents a compensatory mechanism to maintain calcium extrusion capacity when existing PMCA is post-translationally inhibited, consistent with previous observations that PMCA expression is maintained during renal stress despite compromised pump activity [20,21,24,27]. Additionally, ENPP2 downregulation would reduce lysophosphatidic acid (LPA) production, limiting LPA-mediated PMCA inhibition via PLC-PKC signaling previously demonstrated [29]. The presented gene expression analysis is complementary and contextual, it does not claim direct functional causality for individual genes on PMCA activity, and the primary mechanistic evidence for PMCA regulation in this study derives from direct enzymatic assays combined with pharmacological and lipidomic data.

Collectively, these transcriptomic data reveal a coordinated transcriptional program that remodels the lipid signaling landscape during hypoxic stress, balancing DG production for signaling with compensatory mechanisms (PMCA4 transcription, PLD-mediated PA production, reduced LPA synthesis) to maintain calcium homeostasis. Our biochemical experiments demonstrating opposing effects of DG (inhibition) and PA (stimulation) on PMCA activity are thus situated within a dynamic regulatory network where both lipid levels and enzyme expression are actively modulated. These findings emphasize that the DG/DGK/PA axis is a physiologically relevant regulatory system responding to metabolic stress, particularly hypoxia—a clinically relevant stressor in ischemic kidney injury, acute kidney injury, and chronic kidney disease. Future studies should investigate whether pharmacological modulation of this axis (e.g., DGK activation, PLD modulation, or PKC inhibition) could preserve calcium homeostasis and protect proximal tubule cells during ischemic or metabolic stress.

5. Conclusions

The data presented here underscore how localized and systemic alterations in Ca²⁺ dynamics contribute to the onset and progression of kidney disease. Understanding Ca²⁺ cell homeostasis and signaling not only reveals fundamental mechanisms involved in renal physiology but also sheds light on potential targets for therapeutic intervention. The present study provides evidence that DG inhibits PMCA activity in renal proximal tubule membranes via PKC-dependent mechanisms, while PA exerts opposite stimulatory effects, establishing these interconvertible lipids as reciprocal regulators of calcium homeostasis. The reversal of DG’s effect by Calphostin C confirms a role for PKC isoforms, which are constitutively expressed in renal epithelial cells. In our renal proximal tubule basolateral membrane preparation, DG produces a PKC-dependent net inhibition of PMCA activity, which contrasts with frequently cited PKC-mediated activation paradigms. Importantly, the kidney tubule membrane context is known to exhibit specialized PLC/PKC lipid signaling and DGK coupling at basolateral membranes, which supports biological plausibility for compartment-specific DG/PA signaling outcomes.

We also integrate the opposing functional actions of DG versus PA as a coherent signaling axis (DG↔DGK↔PA) relevant to stress physiology, rather than treating these lipids as isolated modulators. Our targeted lipidomic analysis revealed multiple DG molecular species in HK-2 cells, indicating complex substrate diversity that may confer specificity to DGK-mediated regulation. Transcriptomic data analysis of hypoxic HK-2 cells revealed dramatic and coordinated regulation of genes encoding DG-producing enzymes (PLCB1, PLDs), DG-metabolizing kinases (DGKs), PA-modifying enzymes (PLPPs), PKC isoforms, and PMCA pumps. The coordinated upregulation of DG-producing pathways alongside downregulation of DG-consuming enzymes under hypoxic stress suggests that local DG accumulation is a regulated response contributing to calcium dysregulation during ischemic injury. The compensatory upregulation of ATP2B4 (PMCA4) indicates that proximal tubule cells attempt to maintain calcium extrusion capacity even when PMCA is post-translationally inhibited. These findings support a model in which the DG/DGK/PA axis serves as a dynamic, lipid-mediated rheostat that integrates metabolic signals with calcium homeostasis in the kidney, with both transcriptional remodeling and post-translational modifications working in concert to fine-tune calcium handling under physiological and pathological conditions.