Cytosolic phospholipase A₂α regulates G₁ progression through modulating FOXO1 activity

Abstract

Group IVA phospholipase A₂ [cytosolic phospholipase A₂α (cPLA₂α)] is a key mediator of inflammation and tumorigenesis. In this study, by using a combination of chemical inhibition and genetic approaches in zebrafish and murine cells, we identify a mechanism by which cPLA₂α promotes cell proliferation. We identified 2 cpla₂α genes in zebrafish, cpla₂αa and cpla₂αb, with conserved phospholipase activity. In zebrafish, loss of cpla₂α expression or inhibition of cpla₂α activity diminished G₁ progression through the cell cycle. This phenotype was also seen in both mouse embryonic fibroblasts and mesangial cells. G₁ progression was rescued by the addition of arachidonic acid or prostaglandin E₂ (PGE₂), indicating a phospholipase-dependent mechanism. We further show that PGE₂, through PI3K/AKT activation, promoted Forkhead box protein O1 (FOXO1) phosphorylation and FOXO1 nuclear export. This led to up-regulation of cyclin D1 and down-regulation of p27^Kip1, thus promoting G₁ progression. Finally, using pharmacologic inhibitors, we show that cPLA₂α, rapidly accelerated fibrosarcoma (RAF)/MEK/ERK, and PI3K/AKT signaling pathways cooperatively regulate G₁ progression in response to platelet-derived growth factor stimulation. In summary, these data indicate that cPLA₂α, through its phospholipase activity, is a critical effector of G₁ phase progression through the cell cycle and suggest that pharmacological targeting of this enzyme may have important therapeutic benefits in disease mechanisms that involve excessive cell proliferation, in particular, cancer and proliferative glomerulopathies.—Naini, S. M., Choukroun, G. J., Ryan, J. R., Hentschel, D. M., Shah, J. V., Bonventre, J. V. Cytosolic phospholipase A₂α regulates G₁ progression through modulating FOXO1 activity.

Phospholipase A₂ (PLA₂) comprises a family of enzymes that hydrolyze phospholipids to generate free fatty acids and lysophospholipids. Among these phospholipases, the group IVA cytosolic large molecular mass form of PLA₂ (cPLA₂α) is a major source of arachidonic acid (AA), an important lipid second messenger (1–4) that can be converted by cyclooxygenase, lipoxygenase, and cytochrome p450 enzymes into prostaglandins (PGs), leukotrienes, hydroxyeicosatetraenoic acids, epoxyeicosatrienoic acids, and dihydroxyeicosatrienoic acids, products that are involved in the regulation of a number of cellular processes including inflammation, mitogenesis, and cell differentiation (1–7). Several metabolites produced by cPLA₂α activity, including AA, lysophosphatidic acid, thromboxane A₂, and PGs, have been implicated in the regulation of DNA synthesis and cell proliferation (8–12). However, the exact mechanism by which these eicosanoid products regulate proliferation is poorly understood. Furthermore, many studies addressing this issue have relied upon inhibitors, which cannot be expected to be entirely specific to cPLA₂α. We previously reported the effects of deletion of the cPLA₂α locus in the Apc^Min mouse, one of the most widely used mouse models for colon cancer (13). Mice carrying the Apc^min mutation on a c57BL/6J background spontaneously develop many tumors (adenomas or polyps) due to a dominant germ line mutation in the adenomatous polyposis coli (apc) gene, the mouse homolog of the human APC gene. APC^min/+cPLA₂α^−/− animals show an 83% reduction in tumor number in the small intestine when compared with APC^min/+cPLA₂α^+/− double heterozygotes and Apc^min/+cPLA₂α^+/+ littermates. Interestingly, APC^min/+mPGES-1^−/−, which lack prostaglandin E₂ (PGE₂), showed the same amount of reduction in tumor polyp size and number as the APC^min/+cPLA₂α^−/− mice, suggesting that the effects of cPLA₂α deletion on cell proliferation in this mouse model may be PGE₂ dependent (14).

The aim of the present study was to determine the effect of cPLA₂α on the G₀/G₁ phase of the cell cycle. We used primary cells derived from cPLA₂α^−/− mice and littermate cPLA₂α^+/+ mice for in vitro assays and the zebrafish model for our in vivo studies. The zebrafish has evolved as a facile in vivo model to study human disease because many genes are highly conserved between the 2 vertebrate species, including cyclins, cyclin-dependent kinases (Cdks), and inhibitors of Cdks (15, 16). Expression profiles of cell cycle regulatory genes have shown that genes of major importance to G₁ and S phases of the cell cycle, including orthologs of the retinoblastoma (pRb), cyclin D1, and cyclin E1, were expressed at very low levels early after fertilization and increased markedly between 3 and 6 h postfertilization (hpf), making zebrafish a suitable in vivo model to study early cell division, tissue-specific cellular proliferation, and more broadly, the role of cell cycle genes in development and disease (15).

Here, we identified the cpla₂α gene family in zebrafish, and we show a novel role for cPLA₂α in the regulation of G₁ phase of the cell cycle. Lack of cPLA₂α activity resulted in lower levels of cyclin D1, higher levels of p27^Kip1, a marked decrease in kinase activity associated with Cdk4, and prolongation of G₁ phase. This function of cPLA₂α is dependent on its phospholipase activity and mediated through PGE₂ signaling.

MATERIALS AND METHODS

Antibodies and chemicals

The following antibodies were used: anti-cPLA₂α, anti-cPLA₂α (Ser505), anti-AKT, α-phospho-AKT (Ser473), anti-Forkhead box protein O1 (FOXO1), anti-phospho-FOXO1 (Ser256), and anti-phospho-ERK 1/2 (Tyr204) (from Cell Signaling Technology, Beverly, MA, USA). Anti-α-tubulin, anti-EGFP (enhanced green fluorescent protein), anti-cyclin D1, anti-cyclin E, anti-cyclin A, anti-p21Cip1, anti-p27^Kip1, anti-Cdk2, anti-Cdk4, anti-ERK 1/2 anti-glyceraldehyde 3-phosphate dehydrogenase, and anti-lamin A/C were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Anti-BrdU (5-bromo-2′-deoxyuridine) was purchased from Abcam Incorporated (Cambridge, MA, USA). Ionophore A23187 (working concentration 10 μM), BrdU (10 mM), platelet-derived growth factor (PDGF; 10 mg/ml), PD9809 (100 µM), Ly294002 (30 µM), AA (39 pM), AS1842856 (0.1 µM), and PGE₂ (5 nM) were purchased from Sigma-Aldrich (St. Louis, MO, USA, USA). [³H]Thymidine (1 μCi/ml), [³H]AA (0.5 µCi/ml), [γ−³²P]ATP (10 μCi), phosphatidylcholine 1-steratoyl-2-[1-¹⁴C]arachidonyl (0.5 nM), and methyltrienolone (R1881; 100 nM) were purchased from New England Nuclear (Boston, MA, USA). Prostaglandin E₂ receptor 4 (EP₄) antagonist (L-161982; 1 µM) and pyrrophenone (1 µM) were purchased from Cayman Chemicals (Ann Arbor, MI, USA).

Zebrafish husbandry

Wild-type (WT) zebrafish (Danio rerio) were maintained on a 14/10-h light-dark cycle at 28°C and fed twice daily. Fertilized eggs were raised in 30% Danieau’s solution at 28.5°C [58 mM NaCl, 0.7 mM KCl, 0.4 MgSO₄, 0.6 mM Ca(NO₃)₂, and 5 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) (pH 7.2)]. After incubation for 4 h, unfertilized eggs were removed, and fertilized eggs were washed several times with 30% Danieau’s solution. These fertilized eggs were then incubated, 20 eggs per dish, in 30% Danieau’s solution containing 100 mM cadmium chloride. All experiments were carried out in accordance with the guidelines established by the Harvard University Committee on the Use and Care of Animals.

Whole-mount in situ hybridization

In situ hybridization antisense probes for zebrafish cpla₂αa and b were synthesized as described previously (17). Digoxigenin-labeled antisense and sense RNA probes were generated from cDNAs of 24 hpf WT embryos using a digoxigenin-RNA labeling kit (Roche, Mannheim, Germany) according to the manufacturer’s instructions. Each in situ experiment was done at least twice. Embryos were fixed in diluted formalin (1:2.7 in polybutylene terephthalate) at room temperature for 1 h. Alkaline phosphatase-coupled anti-digoxigenin (Roche) was used to localize hybridized probes. NBT/BCIP (Roche) was used as the chromogenic substrate to produce blue precipitates.

Microinjection of mRNA and morpholino oligonucleotides

Antisense morpholino (MO) oligonucleotides (Gene Tools, Philomath, OR, USA) were designed to target the cpla₂αa and b translational start sites (ATG): cpla₂αa MO (5′-AGGTCAGGATGGCACCTTATTTCAA-3′) and cpla₂αb MO (5′-CTCCTTTGGTGACATTTTCAGCCCG-3′). MOs were resuspended in 1 × Danieau’s buffer [58 mM NaCl, 0.7 mM KCl, 0.4 mM MgSO₄, 0.6 mM Ca(NO₃)₂, and 5.0 mM HEPES (pH 7.6)] with 0.1% phenol red (Sigma-Aldrich). Embryos obtained from crosses of adult fish were injected at the 1- or 2-cell stage with an injection volume equal to 2.3 nl MOs per embryo. For the mRNA rescue experiment, human cPLA₂α cDNA was cloned into pcs2+ vector and transcribed in vitro by using the SP6 mMESSAGE mMACHINE Kit (Ambion Corporation, Naugatuck, CT, USA). For phenotype rescue, 100 pg mRNA per embryo was used. Synthesized mRNAs were dissolved in 0.2% phenol red as a tracking dye and then microinjected into 1- to 2-cell stage embryos.

Construction of plasmids

To determine the subcellular localization of the 2 zebrafish cPLA₂α proteins, the open reading frame of each was amplified by PCR from total RNA of 24 hpf zebrafish embryos. Coding sequences were fused to EGFP cDNA in pEGFP-C1 expression vector (Clontech Laboratories, Mountain View, CA, USA). Adenoviral cPLA₂α (Ad-cPLA₂α), the recombinant adenovirus that expresses the human cPLA₂α cDNA, was constructed as described previously (6), propagated in human embryonic kidney (HEK)293 cells and purified by cesium chloride gradient ultracentrifugation. The viral stock was determined to be 2.1 × 10¹² particles/ml and 5 × 10¹⁰ plaque-forming units per cell by plaque assay.

Cell culture

HEK293 and LLC-PK1 cells were purchased from American Type Culture Collection (Manassas, VA, USA). All cells were maintained in DMEM/F12 at 37°C in a 5% CO₂ incubator. Transfection of the constructs was performed using Lipofectamine 2000 (Thermo Fisher Scientific, Waltham, MA, USA) following the manufacturer’s instructions. For starvation experiments, mouse embryonic fibroblasts (MEFs) were kept in serum-free medium for 72 h. For cell cycle reentry, cells were treated with 10% fetal bovine serum (FBS) or 10 mg/ml PDGF. Inhibitors were added to the medium at recommended concentrations 30 min before mitogen stimulation.

Western blotting

Cells were collected and lysed with lysis buffer [50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate, and 1 mM EDTA] supplemented with protease inhibitor cocktail (cOmplete Mini; Roche, Indianapolis, IN, USA). Total cell lysates were resolved by 7.5 and 10% SDS-PAGE, using 25 µg total cellular protein for each sample.

PLA₂ activity

PLA₂ activity was measured in cells plated on 100 mm tissue culture dishes as previously described (6). Two days after infection, cells were washed, harvested, and lysed by sonication (Heat Systems-Ultrasonics, Incorporated, currently Qsonica, LLC. Newtown, CT, USA). The lysates were centrifuged at 100,000 g for 1 h at 4°C, and protein concentration was determined. cPLA₂α activity was assayed in duplicate at 37°C for 30 min in 100 μl reaction volume that included 30 μg protein, a 0.5 nM concentration of 1,3-phosphatidylcholine 1-steratoyl-2-[1-¹⁴C]arachidonyl, 2.5 mM CaCl₂, 0.2 mM EGTA, 50 mM NaCl, and 75 mM Tris, pH 9.0. The reaction was stopped by adding 800 μl Dole’s reagent (32% isopropyl alcohol, 67% N-heptane, and 1% 1 N H₂SO₄) and vortexing. After centrifugation, 150 μl of the upper phase was transferred to a new tube containing 50 mg silica gel and 800 μl N-heptane. After vortexing and allowing the silica gel to settle, 800 μl supernatant was counted for radioactivity in a liquid scintillation counter.

Assessment of cell proliferation

Cells were trypsinized for 15 min at 37°C and resuspended in a fresh medium containing 0.2% trypan blue, and the number of viable cells was determined in a counting chamber. DNA synthesis was measured as the incorporation of [³H]thymidine into trichloroacetic acid (TCA) insoluble material as previously described (18). After stimulation, cells were pulse labeled with 1 μCi/ml [³H]thymidine for 6 h and fixed on ice for 30 min in 10% TCA. After several washes, the radioactivity incorporated into the TCA-precipitated material was determined by liquid scintillation counting after solubilization in 0.25 M NaOH.

Incorporation and release of [³H]AA

For passive [³H]AA release, the medium was replaced with 0.5 ml DMEM containing 0.1 mCi [³H]AA, and the samples were incubated for an additional 12 h at 37°C. The medium was collected, and cells were then washed 3 times in Ca²⁺- and Mg²⁺-free PBS containing 0.5% bovine serum albumin to remove unincorporated [³H]AA. At the end of each experiment, the cells were scraped off the dishes and lysed. The amount of [³H]AA released into the medium was measured by scintillation counting and was expressed as a percentage of cell-incorporated [³H]AA, which was determined in solubilized cells. Background [³H]AA release from untreated cells was subtracted from all data. For active [³H]AA release, cells were treated with ionophore A23187 for 30 min before the collection of the medium.

Immune complex kinase assay

Immune complex kinase assays for endogenous Cdk activity were performed as described previously (18). Briefly, after stimulation with PDGF (10 ng/ml), cells were lysed in 20 mM HEPES, pH 7.4, 2 mM EGTA, 50 mM β-glycerophosphate, 1 mM DTT, 1 mM Na₃VO₄, 1% Triton X-100, 10% glycerol, 2 μM leupeptin, 10 μM aprotinin, and 400 μM PMSF. After matching the lysates for protein concentration, they were incubated with an antibody against Cdk2 or Cdk4 in the same condition and then incubated with 50 μl settled protein G-Sepharose beads for 2 additional hours. Beads were washed 3 times in lysis buffer and 3 times in assay buffer [20 mM 3-morpholinopropane-1-sulfonic acid (MOPS) (pH 7.2), 2 mM EGTA, 10 mM MgCl₂, 1 mM DTT, and 0.1% Triton X-100]. After the washes, 40 μl of a 1:1 suspension of the protein G-Sepharose beads in assay buffer was incubated for 20 min at 30°C with histone H1 for Cdk2 activity, or with pRb for Cdk4 activity, in the presence of 10 μCi [γ−³²P]ATP in 10 mM MgCl₂. The reactions were stopped with sodium dodecyl sulfate sample buffer, and proteins were separated by SDS-PAGE. The gels were stained with Coomassie Brilliant Blue, exhaustively destained, dried, and subjected to autoradiography. The Coomassie Brilliant Blue-stained histone H1 and pRb bands were excised from the dried gels, and the radioactivity incorporated was determined by scintillation counting. Data are presented as milliunits of activity where 1 mU activity transfers 1 pM/min phosphate to the substrate or as fold increase over the control unstimulated cells.

Quantitative RT-PCR

Total RNAs were extracted with TRIzol (Thermo Fisher Scientific). cDNAs were reverse transcribed using M-MLV reverse transcriptase in the presence of oligo (dT) primers, in accordance with the manufacturer’s instructions. Real-time RT-PCR analysis was performed in the Real-Time PCR System, using SYBR Green qPCR SuperMix (both from Bio-Rad, Richmond, CA, USA) with a reaction volume of 20 μl. The Mus musculus primer pairs were as follows (numbers in parentheses indicate accession number): M. musculus specificity protein 1 (NM_013672) sense primer 5′-acagcaggtggagaaggaga-3′, antisense primer 5′-tgaggctcttccctcactgt-3′; M. musculus nuclear transcription factor Y (NM_001110832) sense primer 5′-agtcagtggaggccagctta-3′, antisense primer 5- ccaggaggcaccaactgtat-3; M. musculus E2F transcription factor 1 (NM_007891) sense primer 5′-gaggctggatctggagactg-3′, antisense primer 5- gaagcgtttggtggtcagat-3; M. musculus breast cancer 1 (NM_009764) sense primer 5′-tgacagtgccaaagaactcg-3′, antisense primer 5′-gatacgctggtgctctcctc-3′; M. musculus V-myc avian myelocytomatosis viral oncogene homolog (NM_001162917) sense primer 5′-tgctgaaaagcatggtgaag-3′, antisense primer 5′-ccagtgccacaagtgagaga-3′; M. musculus ID-3 (NM_008321) sense primer 5′-cgaggagcctcttagcctct-3′, antisense primer 5′-gtctatgacacgctgcagga-3′; M. musculus hairy and enhancer of split-1 (NM_008235) sense primer 5′-acaccggacaaaccaaagac-3′, antisense primer 5′-atgccgggagctatctttct-3′; M. musculus FOXO1 (NM_019739) sense primer 5′-aaccagtccaactcgaccac-3′, antisense primer 5′-tgctcataaagtcggtgctg-3′; M. musculus Cyclin D1 (NM_007631) sense primer 5′-gcgtaccctgacaccaatct-3′, antisense primer 5′-atctccttctgcacgcactt-3′; and mp27^Kip1 (NM_009875) sense primer 5′-aactaacccgggacttggag-3′, antisense primer 5- ccaggggcttatgattctga-3. All primer sets were confirmed to produce a single peak in melting curve as well as a single band by agarose gel electrophoresis in the preliminary study.

Confocal microscopy

Cells grown on coverslips were fixed in 2% paraformaldehyde/PBS for 10 min at room temperature. Fixed cells were permeabilized with 0.1% Triton X-100 in PBS for 3 min and blocked in 2% calf serum for 30 min at room temperature. Cells were then incubated with primary antibody for 2 h, washed 3 times with 1× PBS/0.1% Tween 20. Fluorophore-conjugated secondary antibodies were added for 45 min at room temperature. After 3 washes using 1× PBS/0.1% Tween 20, coverslips were mounted with Vectashield (Vector Laboratories, Burlingame, CA, USA) and examined with a confocal Nikon C1 microscope (Nikon, Tokyo, Japan). For quantification studies, ImageJ software plug-in (NIH, Bethesda, MD, USA) was used.

Flow cytometry

Cell cycle profiles were created using ModFit LT (http://www.vsh.com/products/mflt/) and FlowJo v7.6.5 (http://www.flowjo.com/legacy-versions-downloads/) software. For BrdU labeling in zebrafish, dechorionated embryos were incubated with 10 mM BrdU/15% DMSO for 15 min on ice, washed, and incubated for 15 min at 28.5°C. Cells were disaggregated by triturating embryos in 0.25% trypsin/1 mM EDTA using a p200 pipette and fixed in 70% ethanol overnight at −20°C. Suspensions of cells from 40 WT or morphant embryos were pooled and prepared for BrdU/propidium iodide fluorescence-activated cell-sorting analysis. For BrdU labeling in cell culture, BrdU was added to the cells to a final concentration of 10 mM. Ethanol-fixed cells were serially stained with an anti-BrdU antibody and propidium iodide. All fluorescence-activated cell-sorting analysis was conducted using a FACScan instrument (Becton Dickinson, San Diego, CA, USA).

Sequence alignment

Information for the chromosomal locations and names of genes were taken from the UniGene database (www.ncbi.nlm.nih.gov/Uni-Gene/). The zebrafish genome assembly (Zv9, July 2010; www.ensembl.org/Danio_rerio) was used to assign the zebrafish regions bearing the cpla₂α genes. VISTA plot (http://genome.lbl.gov/vista/index.shtml) was used for evaluating exon and intron homology between cpla₂α genes in zebrafish and humans.

Statistics

Multigroup analysis was performed by ANOVA. Student’s t test was used to determine a significant difference between 2 groups. Values of P < 0.05 were considered statistically significant. Significance levels are indicated in the histograms in the figures. The error bars represent mean ± sd values.

RESULTS

Identification of cpla₂α genes in zebrafish

We performed Translated Basic Local Alignment Search Tool Nucleotide (TBLASTN) analysis, using information from human and mouse cPLA2a proteins, against zebrafish public databases and cDNA libraries. We identified 2 distinct zebrafish genes, and we termed them cpla₂αa and cpla₂αb (Fig. 1). cpla₂αa has a putative signal peptide of 12 aa and a mature protein of 742 aa. cpla₂αb has a putative signal peptide of 16 aa and a mature protein of 730 aa. There are 3 known phosphorylation sites in human cPLA₂α: Ser228, Ser505, and Ser727. These sites have been conserved between zebrafish and humans, indicating the possibility of the same regulatory mechanism across species (Fig. 1). Next, we determined the chromosomal positions of the 2 cpla₂α genes and their potential syntenic relations to human chromosomes. Human cPLA₂α maps to Homo sapiens chromosome 1q31.1, and adjacent genes are PG-endoperoxide synthase 2 (PTGS2), phosducin (PDC), nuclear pore complex-associated protein, and proteoglycan 4 in that order toward the centromeric end (Fig. 2A). cpla₂αa and b reside at the 20.7 megabase (Mb) position of D. rerio chromosome 2 and at the 34.1 Mb position of D. rerio chromosome 20, respectively. Both cpla₂α genes are found in a landscape containing ptgs2, pdc, nuclear pore complex-associated protein (tpr), and proteoglycan 4 (prg4), in the same order and direction of their human counterparts (Fig. 2A). These data establish shared synteny between human and zebrafish cpla₂α and associated ortholog genes. They occupy chromosomal segments that maintain conserved gene orders. This level of conservation indicates that the parallel chromosomal segments have been inherited without chromosomal rearrangement from the last common ancestor of these 2 species (19). We performed phylogenetic analysis to place the 2 cpla₂α genes more precisely in the evolutionary history of the cPLA₂α family. This analysis showed slightly more homology between cpla₂αa and the human ortholog, as indicated by higher bootstrap support (0.117 for cpla₂αa versus 0.118 for cpla₂αb) (20) (Fig. 2B). To compare the degree of homology between protein sequences, we analyzed the degree of sequence similarity and divergence between corresponding human and zebrafish orthologs. Of the 2 cPLA₂α proteins, the deduced amino acid sequence of cpla₂αa exhibited a higher homology with the sequences of human cPLA₂α: C2 domain (73% identical, 13% similar), and catalytic domains 1 and 2 (79% identical, 8% similar) (Fig. 2C). We then compared the genomic sequence of a 100 kb segment encompassing the human cPLA₂α gene with the ortholog interval in zebrafish using the AVID/VISTA program. The 2 cpla₂α genes share similar exon/intron organization with the human cPLA₂α gene, all comprising 18 exons and 17 introns (Fig. 2D).

Figure 1.

Alignment of human (h), mouse (m), and 2 zebrafish (z) cPLA₂α protein sequences. Identical amino acid residues are highlighted in black box shading, whereas similar amino acid replacements are presented in gray box shading. Gaps are introduced for optimal alignment. Horizontal lines indicate different domains of cPLA₂α: C2 domain (blue), catalytic domain 1 (green), and catalytic domain 2 (red). Black arrowheads point to the phosphorylation sites known to be critical for cPLA₂α activation (Ser228, Ser505, and Ser727).

Figure 2.

Characterization of cpla₂α genes in zebrafish. A) Genomic analysis of conserved synteny for human (H) and zebrafish (zf) cpla₂α genes. The 2 cpla₂α genes are located on chromosomes (ch) 2 and 20 in a region near orthologs of human PTGS2 and PDC, which map near human cPLA₂α on chromosome 1 at q 31.1. B) Structural and phylogenic analysis shows the evolutionary relationship between cPLA₂α genes in different species. C) Percentage of identical (ID) and similar (SI) base pairs between human and 2 zebrafish cpla2α genes. Sequences were aligned with the National Center for Biotechnology Information Basic Local Alignment Search Tool program. D) VISTA plot displays the alignment of the human cPLA₂α locus with the ortholog zebrafish region. There is a considerable homology between human exons (red vertical lines) and zebrafish exons. Horizontal lines show the identicality scores (50 and 100%). E) LLC-PK1 cells were transfected with plasmids expressing cPla₂αa or b proteins tagged with EGFP. Upon ionophore activation, both proteins localized to perinuclear membranes, compatible with activated proteins. Immunoblotting confirmed expression of the proteins in cellular lysates. Scale bar, 10 µm. F) Passive and active [³H]AA release from proliferating LLC-PK1 cells transfected with EGFP, EGFP-tagged human or zebrafish cPLA₂α proteins. Cells in specified experiments were treated with the cPLA₂α inhibitor, pyrrophenone.

To verify that cpla₂αa and b have conserved phospholipase activity, LLC-PK1 cells were transiently transfected with either cPla2αa or cPla2αb EGFP-fusion constructs, and the cellular localization and phospholipase activity of the 2 proteins were compared. Our immunofluorescence data showed that both proteins were localized to the cytoplasm and were mobilized to the nuclear membrane after treatment with calcium ionophore A23187 (21) (Fig. 2E). This perinuclear localization upon activation is a characteristic of all PLA₂ group IVA family members (21). Western blotting confirmed the expression of EGFP-fusion proteins in transfected cells (Fig. 2E). To compare the phospholipase activity of the 2 proteins, [³H]AA release was measured. Short-term (30 min) incubation with ionophore was sufficient to maximally induce [³H]AA release. Both proteins showed phospholipase activities comparable to the human counterpart (Fig. 2F). Of note, ionophore treatment selectively liberates AA derived from pools accessible to the Ca²⁺-dependent isoforms of cPLA₂α and can thus be considered selective for the activation of cPLA₂α (21). Preincubation with 1 µM pyrrophenone (22), a specific inhibitor of cPLA₂α, was sufficient to significantly inhibit phospholipase activity.

Temporospatial expression patterns of cpla₂αa and b during zebrafish development

Both transcripts were readily detectable at 12 hpf and continued to be expressed at similar levels until 72 h, after which the expression of cpla₂αa declined dramatically, whereas transcription of cpla₂αb was sustained (Fig. 3A). Western blotting on whole lysates was performed using an antibody raised against human cPLA₂α. Consistent with RT-PCR data, protein expression was detectable at 12 h, with peak expression occurring between 24 and 34 hpf (Fig. 3B). To examine the distribution of the 2 cpla₂α transcripts, whole-mount in situ hybridization of WT zebrafish embryos was performed at 24 hpf. Both cpla₂αa and b were expressed in the developing brain; however, with different and distinct expression patterns. cpla₂αa was moderately expressed in a cell cluster restricted to the forebrain and hindbrain with a clearer and more intense signal in the otic vesicles (Fig. 3C). cpla₂αb, on the other hand, was expressed mainly in the midbrain. Thus, based on the spatiotemporal expression patterns, we infer that these 2 genes are likely to be implicated in the formation of the CNS and neuronal development in zebrafish.

Figure 3.

Temporospatial expression of cPla₂αa and b during early zebrafish development. A) cpla₂αa and b mRNAs were detected by RT-PCR as early as 12 hpf until 96 hpf. β-Actin was used as the loading control. B) Immunoblotting shows the expression of the cPla₂αa and cPla₂αb proteins in the zebrafish embryo whole lysates using an Ab raised against human cPLA₂α. C–F) Antisense RNA probes were used for whole-mount in situ hybridization on zebrafish embryos at 24 hpf. cpla2αa transcripts are detected in forebrain and hindbrain (white arrowheads in C) and otic vesicles (white arrows in C and D), whereas cpla2αb transcripts are mainly detected in midbrain (black arrow in E). Sense probe was used as a negative control (F).

Growth defects in embryos with reduced levels of either cpla₂αa or cpla₂αb

The early onset and sustained expression suggested that cpla₂α genes play an important role in the development of the brain region. Morphologic analyses were carried out to compare the cpla₂α MO-injected and control MO-injected embryos. At 24 hpf, we were able to detect a significant decrease in the size of the otic vesicles of cpla₂αa MO-injected embryos using starmarker riboprobes (∼60% decrease) (23) (Fig. 4A). In cpla₂αb MO-injected embryos, we observed subtle differences in the size of the midbrain at 24 hpf (data not shown), but this morphologic change was more prominent at 48 hpf (Fig. 4B, C). The shape of the contours of forebrain remained relatively unaffected. We analyzed transverse sections to quantify these differences by measuring the average cross-sectional area of the tectal region and observed a reduction of ∼40% in the cpla₂αb morphants compared with controls (Fig. 4D–F).

Figure 4.

Growth defect in zebrafish lacking cPla₂αa or cPla₂αb. A) Fertilized eggs were injected with control (Ctrl) or cpla₂αa MOs. Whole-mount in situ hybridization with riboprobes for starmarker (Stm) was performed. Dorsal view, anterior to the top. Area of the otic vesicle was quantified using ImageJ software. A.U, arbitrary unit. B–E) Fertilized eggs were injected with control (B, D) or cpla₂αb (C, E) MOs. Gross morphologies and cross-sectional area of the midbrain on coronal plastic sections are represented. Red dotted lines indicate the presumptive tectal region. Black arrowheads in (B) and (C) indicate the position of the representative coronal sections in (D) and (E), respectively. F) Cross-sectional area of the midbrain on coronal plastic sections (D and E) was compared between the 2 groups using ImageJ software. G) One-cell embryos were injected with control or cpla₂αa/b double MOs, and phenotypes of each embryo were observed at different time points. Embryos in specified experiments were coinjected with human cPLA₂α mRNA. Indicated in the right-upper corners is the ratio of embryos with presented morphology to the total number of injected embryos. All embryos were coinjected with 1.5 ng p53 MO. H) Embryos from (G) were subjected to immunoblotting using an anti-human cPLA₂α antibody. Black arrowhead indicates the band corresponding to the human cPLA₂α protein in extracts of fish injected with human cPLA2α mRNA. I) One-cell embryos were injected with control or cpla₂αa/b double MOs. Embryos in specified experiments were coinjected with AA or PGE₂. The number of embryos with the illustrated phenotype and the total number of injected embryos are indicated. **P < 0.01.

To eliminate any possibility of functional redundancy, we injected 1-cell embryos with increasing concentrations of both MOs and analyzed the morphology at 24, 48, and 72 hpf. Down-regulation of both genes simultaneously resulted in head shrinkage (smaller and incomplete head and otolith) and central axis shortening (Fig. 4G). The body trunk, tail, and yolk extension were slightly shortened. This evidence suggests that cpla₂α is implicated in the development of the body plan in zebrafish, particularly in the neural system. All aspects of the phenotype were more severe with increased concentrations of MOs. Closer examination revealed that these changes consisted largely of growth inhibition as evidenced by the reduced circumference of the otolith and the reduced size of the head region without overt evidence for necrosis or gross malformation (Fig. 4G). Coinjection of 100 pg human cPLA₂α mRNA completely rescued the phenotype in double morphants, indicating that the observed phenotype is the result of a reduction in cPLA₂α protein expression (Fig. 4G). Immunoblotting confirmed the efficiency of knockdown in MO-injected embryos (Fig. 4H). We further performed rescue experiments by coinjecting double-morphant embryos with AA or PGE₂. Coinjection of either compound normalized the head and axis deformities at 48 and 72 hpf. These data indicate that the observed developmental defects are dependent on the phospholipase activity of cPLA₂α (Fig. 4I).

Lack of cpla₂α results in a G₁/S transition defect in zebrafish

Next, we investigated whether the growth defects in cpla₂α morphant zebrafish could be due to cell cycle deregulation. Embryos were injected with control or cpla₂αa/b double MOs. A total of 40 embryos in each group were digested at 24 hpf, and cells were pooled for cell cycle analysis. Analysis of the cell cycle profile showed that cpla₂α double morphants had a significantly higher percentage of G₀/G₁ cells (Δ ∼20%), and a significantly lower percentage of S phase cells (Δ ∼18%), compared with embryos injected with control MO (Fig. 5A). We further evaluated the ability of each embryo to incorporate BrdU by flow cytometry and immunostaining (Fig. 5B). In the control embryos, 35.2 ± 3.4% of cells were BrdU⁺. This is indicative of the high rate of cell proliferation at this early stage of zebrafish development. In contrast, the double-morphant embryos incorporated less BrdU (13.2 ± 1.8%; P < 0.01). Coinjection of AA or PGE₂ rescued the phenotype produced by cpla₂α knockdown (Fig. 5A, B). Altogether, these data indicate a requirement of cpla₂α in promoting the G₁/S transition in zebrafish. Viability of these cells was not compromised. cpla₂α-morphant embryos did not show a significant increase in the levels of active caspase-3 or in the vital incorporation of acridine orange (data not shown).

Figure 5.

Lack of cPla₂α reduces G₁/S transition during zebrafish development. A) One-cell embryos were injected with control or double cpla₂α MOs (upper panel). Cell cycle distribution was determined at 24 hpf. Embryos in specified experiments were coinjected with AA or PGE₂. Percentage of cells in each stage of the cell cycle was quantified (n = 3; lower panel). B) Control or cpla₂α double morphants were pulse labeled with BrdU for 30 min. BrdU uptake was quantified by flow cytometry (n = 3). As shown on the right-lower corners, embryos from each condition were stained with anti-BrdU antibody. *P < 0.05; **P < 0.01.

cPLA₂α regulates G₁ progression in murine cells

Next, we assessed cell cycle progression in MEFs and mesangial cells (MCs) isolated from WT and cPLA₂α^−/− mice. cPLA₂α activity was assessed in WT and cPLA₂α^−/− MEFs. cPLA₂α activity was not detectable in cPLA₂α^−/− MEFs (Fig. 6A). Infecting cells with Ad-cPLA₂α at increasing multiplicity of infections increased cPLA₂α activity in cPLA₂α^−/− cells to levels comparable to those seen in WT cells, whereas infection with the control virus, Ad-LacZ, did not change cPLA₂α activity. The increase in cPLA₂α activity in Ad-cPLA₂α-infected cells paralleled the level of increase in cPLA₂α protein expression (Fig. 6A). Under unstimulated conditions at 30 min, there was no difference in [³H]AA release between WT and cPLA₂α^−/− MEFs (Fig. 6B). After 2 h, and before stimulation, the percentage of [³H]AA release by WT cells was higher when comparing with cPLA₂α^−/− cells infected with Ad-LacZ (P < 0.01). Infection with Ad-cPLA₂α increased [³H]AA passive release from cPLA₂α^−/− cells to the level of WT cells. Treatment of the cells with ionophore A23187 for 30 min induced a greater increase in [³H]AA release from WT and cPLA₂α^−/− cells infected with Ad-cPLA₂α (Fig. 6B). These results suggest that cPLA₂α is the major source of Ca²⁺ ionophore-induced release of AA in MEFs.

Figure 6.

cPLA₂α regulates G₁ progression in murine cells. A) PLA₂ activity was determined in WT and cPLA₂α^−/− MEFs after infection with varying concentrations [multiplicity of infection (MOI)] of Ad-cPLA₂α. Infection with an adenovirus expressing LacZ was used as a negative control. Immunoblotting represents cPLA₂α protein expression 48 h after infection at varying MOIs. B) [³H]AA release was quantified in MEFs after infection with varying MOIs of Ad-LacZ or Ad-cPLA₂α. Cells were exposed to vehicle (serum-free medium) for 30 and 120 min or calcium ionophore for 30 min at 37°C. Ctrl, control; pfu, plaque-forming unit. C) Cell cycle distribution of unsynchronized WT and cPLA₂α^−/− MEFs. D) The percentage of cells in each phase of the cell cycle from (C) was determined (n = 3). E) Unsynchronized WT and cPLA₂α^−/− MEFs remained untreated or were treated with the specified CM from cPLA2α^+/+ or cPLA2α^−/− MEFs or PGE₂ for 24 h. Cells were pulse labeled with BrdU for 30 min, and BrdU⁺ population was quantified. F and G) Serum-starved WT and cPLA₂α^−/− MEFs were treated with PDGF to reenter the cell cycle. Cell cycle profiles at specified time points were created using FlowJo software (F). Percentage of cells in each phase of the cell cycle from each time point was quantified (G) (n = 3). H) [³H]AA release from proliferating WT and cPLA₂α^−/− MCs. Passive and ionophore-stimulated release of [³H]AA in the supernatant was quantified. Cells in specified groups were infected with LacZ or cPLA₂α adenoviruses. Protein lysates from cells were separated by SDS-PAGE and assessed with a cPLA₂α antibody. I, J) Serum-starved WT and cPLA₂α^−/− MCs were stimulated for 24 h with different concentrations of FBS. Total cell number (I) and [³H]thymidine incorporation (J) were determined. K) cPLA₂α^−/− MCs were infected with Ad-LacZ or Ad-cPLA₂α and serum starved before stimulation for 24 h with 0 or 2% FBS. [³H]Thymidine incorporation was determined. L) Serum-starved WT and cPLA₂α^−/− cells remained untreated or were treated with PDGF. [³H]Thymidine incorporation was determined after 24 h. *P < 0.05; **P < 0.01.

Analysis of the cell cycle profile in unsynchronized cPLA₂α^−/− cells showed a significant increase in the number of G₀/G₁ cells (Δ ∼20%) and a significant decrease in the number of cells in S (Δ ∼13%) and G₂/M (Δ ∼4%) phases compared to WT cells (Fig. 6C, D). Ad-cPLA₂α, but not Ad-LacZ, rescued the cell cycle phenotype in cPLA₂α^−/− cells. Next, WT and cPLA₂α^−/− MEFs were incubated with conditioned medium (CM) collected from WT or cPLA₂α^−/− cells or with PGE₂, and BrdU incorporation was quantified by flow cytometry. Incubation with PGE₂ or WT CM significantly increased the BrdU⁺ population in cPLA₂α^−/− cells (Fig. 6E). These data confirm our rescue experiments with AA and PGE₂ in zebrafish and suggest that the observed cell cycle phenotype is dependent on cPLA₂α phospholipase activity and mediated through PGE₂ signaling.

Next, we examined the role of cPLA₂α in mitogen-induced cell proliferation. Various mitogens, including PDGF, epidermal growth factor, and serum, are capable of stimulating cell cycle entry and G₁ progression (24, 25). We chose PDGF because it is a well-known activator of cPLA₂α (26, 27). WT and cPLA₂α^−/− MEFs were synchronized by serum starvation and treated with 10 mg/ml PDGF to reenter cell cycle. Cell cycle progression was assessed by flow cytometry at indicated times (Fig. 6F). We observed a delay in entry into S phase in cPLA₂α^−/− cells compared to WT (24 h in cPLA₂α^−/− versus 20 h in WT), indicating a prolonged G₁ phase in cPLA₂α^−/− cells. In addition, there was a significant decrease in the peak S and G₂/M populations in cPLA₂α^−/− cells compared to WT (∼40% in cPLA₂α^−/− versus ∼60% in WT) (Fig. 6G). These data suggest that lack of cPLA₂α prolongs the G₁ phase and reduces the proliferating capacity in response to mitogen stimulation without a significant change in the duration of the other phases of the cell cycle.

Next, cPLA₂α activity was assessed in WT and cPLA₂α^−/− MCs by measuring basal and Ca²⁺ ionophore-stimulated release of [³H]AA (Fig. 6H). When treated with ionophore for 30 min, WT MCs released a large amount of the [³H]AA, whereas there was no increased release in the cPLA₂α^−/− MCs. Thus, the stimulated [³H]AA release resulted entirely from cPLA₂α activation. We performed a rescue experiment by infecting the cPLA₂α^−/− MCs with Ad-LacZ or Ad-cPLA₂α. In contrast to LacZ-expressing cells, cells expressing Ad-cPLA₂α released a substantial amount of [³H]AA after stimulation with ionophore. Having established that the WT MCs have functional cPLA₂α activity, we then confirmed the presence of cPLA₂α protein by immunoblotting (Fig. 6H).

To assess cell cycle progression in cPLA₂α^−/− MCs, cell count and [³H]thymidine incorporation were examined in response to different concentrations of FBS. FBS increased cell numbers (Fig. 6I) and DNA synthesis as reflected by [³H]thymidine incorporation (Fig. 6J) to a greater extent in WT cells compared with cPLA₂α^−/− cells. In order to confirm that the reduced cell proliferation in cPLA₂α^−/− MCs was due to the absence of cPLA₂α, serum-starved cPLA₂α^−/− cells were infected with Ad-LacZ or Ad-cPLA₂α and stimulated with 0 or 2% FBS for 24 h. Thymidine incorporation was greater in cells infected with Ad-cPLA₂α both in the presence and the absence of FBS (P < 0.01) (Fig. 6K), providing further evidence that the difference in proliferation between WT and cPLA₂α^−/− cells is directly related to the level of cPLA₂α expression. Likewise, stimulation with 10 ng/ml PDGF induced a significantly greater increase in DNA synthesis in WT compared to cPLA₂α^−/− MCs (P < 0.01) (Fig. 6L).

cPLA₂α regulates expression of G₁ mediators cyclin D1 and p27^Kip1

The ratio of cyclin-Cdks:Cdk inhibitors is an important factor that drives G₁ progression (28). We observed a substantial increase in p27^Kip1 levels and a significant decrease in cyclin D1 levels in cpla₂α double-morphant zebrafish embryos compared to control embryos (Fig. 7A). Coinjection of AA or PGE₂ normalized the levels of cyclin D1 and p27^Kip1 in double-morphant embryos. Next, we examined the expression of G₁ cyclins and Cdk inhibitors in WT and cPLA₂α^−/− MEFs. Serum-starved WT and cPLA₂α^−/− MEFs (G₀ synchronized) were treated with PDGF to reenter the cell cycle. We explored the temporal pattern for changes in cyclin D1, cyclin E, Cdk4, Cdk2, p27^Kip1, and p21^Cip1 levels following serum stimulation at the indicated times (0, 4, 8, 12, and 16 h). Cyclin D1 is undetectable in quiescent cells, but its expression is first observed at 4 h in WT cells and peaks at 16 h (Fig. 7B). In cPLA₂α^−/− cells, the level of cyclin Dl is very low, its induction is markedly delayed, and the level is reduced compared with WT cells. p27^Kip1 levels were highest at 0 h. In cPLA₂α-expressing cells, there was a rapid reduction of p27^Kip1 at 4 h, which remained low until 16 h. In contrast, in cPLA₂α^−/− cells, p27^Kip1 remained unchanged until 8 h where there was a discrete reduction of the protein amount. There was no observable difference in protein levels of cyclin E, p21^cip1, Cdk2, or Cdk4 during the time course between WT and cPLA₂α^−/− cells (Fig. 7B).

Figure 7.

cPLA₂α regulates cyclin D1 and p27^Kip1 expression in vivo and in vitro. A) Embryos injected with control (Ctrl) MO or cpla₂αa/b double MOs were assessed for specified proteins. Rescue experiments were performed using AA or PGE₂. α-Tubulin was used as a loading control. B) G₀-synchronized WT and cPLA₂α^−/− MEFs were treated with PDGF. Cells were then lysed at indicated time points and evaluated by SDS-PAGE for specified proteins. C) Serum-starved MEFs were stimulated with PDGF for 0 (control), 4, 8, and 12 h. Cdk2 and Cdk4 activities were measured by immune complex kinase assay using histone H1 (Cdk2) or pRb (Cdk4) as substrates. Western blots reveal equal levels of protein at each time point. D) WT MEFs were pretreated with vehicle (control) or cPLA₂α inhibitor (pyrrophenone) before PDGF stimulation. Protein extracts were evaluated at indicated time points using specified antibodies. **P < 0.01.

In the regulation of the transition from G₀ through G₁ phase, cyclin D1 and cyclin E are associated with Cdk4 and Cdk2, respectively (28). To evaluate the functional significance of decreased cyclin Dl levels in cPLA₂α^−/− MEFs, we determined the kinase activity of immunoprecipitated cyclin-Cdk4 complex on pRb. pRb phosphorylation was reduced markedly in cPLA₂α^−/− cells compared with the WT cells from 0 to 12 h following serum stimulation (Fig. 7C). These data indicate that reduced cyclin D1 expression in cPLA₂α^−/− MEFs is accompanied by diminished Cdk4 activity. In contrast, we did not observe any significant difference in Cdk2 activity in WT and cPLA₂α^−/− MEFs, as measured by immune complex kinase assay using histone H1. Total expression of Cdk2 and Cdk4 is not affected by PDGF and is similar in both cell lines as shown by Western blotting (Fig. 7C).

Next, serum-starved WT MEFs were treated with vehicle or 1 µM pyrrophenone before stimulation with PDGF. Cell lysates were collected at indicated time points and analyzed by Western blotting (Fig. 7D). We observed reduced cyclin D1 and increased p27^Kip1 protein expression in pyrrophenone-treated cells, further confirming that this function of cPLA₂α is phospholipase dependent.

cPLA₂α regulates FOXO1 phosphorylation

We sought to determine the mechanism through which cPLA₂α regulates cyclin D1 and p27^Kip1 expression. Several transcription factors including specificity protein 1 (29, 30), nuclear transcription factor Y (31), breast cancer 1 (32), DNA-binding protein inhibitor ID-3 (33), hairy and enhancer of split-1 (34), FOXO1 (35, 36), V-myc avian myelocytomatosis viral oncogene homolog (37), and E2F transcription factor 1 (38) are known to regulate the expression of cyclin D1 and p27^Kip1. Using quantitative PCR, we analyzed the expression of each transcription factor in untreated, unsynchronized MEFs and MCs. FOXO1 was the only transcription factor up-regulated in both cPLA₂α^−/− MEFs and MCs compared to WT cells (3-fold the basal expression) (Fig. 8A). Western blotting confirmed the increased levels of FOXO1 in cPLA₂α^−/− cells compared to WT cells (Fig. 8B). Intriguingly, immunostaining revealed a more prominent cytoplasmic localization of FOXO1 in WT MEFs (Fig. 8C). To confirm these results, we subjected WT and cPLA₂α^−/− cells to a fractionation that yielded cytoplasmic and nuclear fractions. In WT cells, FOXO1 was localized largely in the cytoplasmic fraction, whereas in the cPLA₂α^−/− cells, FOXO1 was present predominantly in the nuclear fraction (Fig. 8D).

Figure 8.

cPLA₂α regulates FOXO1 phosphorylation. A) mRNA levels of transcription factors that are known to regulate p27^Kip1 and cyclin D1 expressions were quantified by quantitative PCR in unsynchronized cPLA₂α^−/− MEFs and MCs cells. B) Immunoblotting shows the expression of FOXO1 in MEFs and MCs. ERK1 was used as a loading control. C) Immunostaining of FOXO1 in WT and cPLA₂α^−/− MEFs. D) WT and cPLA₂α^−/− MEFs were fractioned into cytoplasmic (CE) and nuclear (NE) extracts. FOXO1 levels were assessed by immunoblotting in each fraction. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and lamin A/C were used as specific cytoplasmic and nuclear markers, respectively. E) Levels of phospho-FOXO1 (Ser256) in WT and cPLA₂α^−/− MEFs. PGE₂ and PI3K inhibitor (Ly294002) were used in specified experiments. F) Levels of phospho-FOXO1 (Ser256) in control (Ctrl) or cpla₂αa/b double-morphant zebrafish embryos in the absence or presence of PGE₂. G) HEK293 cells were transfected with a FOXO1 promoter-driven luciferase reporter along with the indicated plasmids, and the promoter activity was assessed. RU, relative units. **P < 0.01.

FOXO1 contains conserved phosphorylation sites for AKT (39), and AKT-mediated phosphorylation of FOXO1 was shown to result in its exclusion from the nucleus and a subsequent inhibition of transcriptional activation of FOXO1 target genes (40). Therefore, we tested whether cPLA₂α modulates FOXO1 phosphorylation through AKT, reflected by phosphorylation on residue Ser256. Immunoblotting analysis showed reduced FOXO1 phosphorylation in cPLA₂α^−/− MEFs compared with WT (Fig. 8E). Addition of PGE₂ to cPLA₂α^−/− cells increased FOXO1 phosphorylation to a comparable level with WT cells. PGE₂-mediated increase in FOXO1 phosphorylation was prevented by addition of Ly294002, a PI3K inhibitor, indicating that this phosphorylation event was fully dependent on PI3K/AKT signaling. Likewise, in cpla₂α double-morphant embryos, FOXO1 was dephosphorylated, indicating a common regulatory mechanism in zebrafish and mammals (Fig. 8F). Finally, transfection of an FOXO1 luciferase reporter that contains a FOXO1-responsive element, in the presence of FOXO1, led to increased luciferase activity, whereas expression of cPLA₂α, but not phospholipase-deficient cPLA₂α^S228A, in the presence of FOXO1, resulted in significant decrease in luciferase activity (Fig. 8G). Altogether, these data indicate that cPLA₂α, through PI3K/AKT signaling, promotes phosphorylation and nuclear export of FOXO1.

cPLA₂α, RAF/MEK/ERK, and PI3K/AKT pathways cooperatively regulate G₁ progression

Two pathways thought to play an important role in committing quiescent cells into S phase are the rapidly accelerated fibrosarcoma (RAF)/MEK/ERK and PI3K/AKT pathways that cooperate to induce cyclin D1 and repress p27^Kip1 expression (41, 42). Our data indicated that cPLA₂α acts through PGE₂ signaling and via the PI3K/AKT/FOXO1 pathway. Additionally, cPLA₂α activation is enhanced through phosphorylation by ERK at Ser505 (43). We therefore designed a set of experiments to explore the role of each signaling pathway in mitogen-induced cell proliferation. First, we examined phosphorylation of major components of the RAF/MEK/ERK and PI3K/AKT pathways in the presence or absence of cPLA₂α. Serum-deprived WT and cPLA₂α^−/− MEFs were stimulated with PDGF and analyzed at indicated times by immunoblotting using phospho-specific antibodies. As expected, upon PDGF addition, cPLA₂α (Ser505), ERK 1/2 (Thr 204), AKT (Ser473), and FOXO1 (Ser256) were rapidly phosphorylated within 30 min, with maximum extent of phosphorylation peaked between 1 and 2 h and the phosphorylation dropping out starting at 12 h (Fig. 9A). There was no difference in phospho-ERK levels between WT and cPLA₂α^−/− MEFs; however, phosphorylation of FOXO1 and AKT was down-regulated in cPLA₂α^−/− cells, consistent with AKT and FOXO1 phosphorylation downstream of cPLA₂α signaling. Total proteins remained unchanged in all time points (Fig. 9A). Next, serum-deprived WT and cPLA₂α^−/− cells were stimulated with PDGF, and cyclin D1 and p27^Kip1 expression was evaluated after 16 h in the presence or absence of pyrrophenone or PGE₂ receptor 4 (EP₄) antagonist (L-161982). EP₄ agonists have been shown to activate the PI3K/AKT pathway (44). As expected, in cPLA₂α^−/− cells, pyrrophenone and L-161982 had no observable effect on cyclin D1 and p27^Kip1 expression (Fig. 9B). In WT cells, both pyrrophenone and L-161982 comparatively blocked PDGF-induced induction of cyclin D1 and repression of p27^Kip1. PGE₂ treatment led to an increase in cyclin D1 and a decrease in p27^Kip1 expression in both WT and cPLA₂α^−/− MEFs. These data confirm our findings in zebrafish and suggest that the cPLA₂α-dependent cell proliferation is mediated through PGE₂ signaling and via activation of the EP₄.

Figure 9.

cPLA₂α, RAF/MEK/ERK, and PI3K/AKT pathways cooperate to promote G₁ progression. A) Serum-deprived WT and cPLA₂α^−/− MEFs were treated with PDGF. Phosphorylation state of the specified kinases was evaluated by immunoblotting at indicated times. B) Serum-deprived WT and cPLA₂α^−/− MEFs were treated with PDGF in the presence or absence of cPLA₂α inhibitor (pyrrophenone) or EP₄ receptor antagonist (L161982). p27^Kip1 and cyclin D1 protein levels were assessed at 0 h (control) and 16 h. Rescue experiments were performed by adding PGE₂ to the specified reactions. C, D) Serum-deprived WT and cPLA₂α^−/− MEFs were treated with PDGF in the presence or absence of MEK inhibitor (PD98059), pyrrophenone, or PI3K inhibitor (Ly294002). Phosphorylation of AKT and FOXO1 was assessed at 1 h (C), and protein (C) and mRNA (D) expression levels of p27^Kip1 and cyclin D1 transcripts were assessed at 0 h (control) and 16 h. E) Serum-deprived WT and cPLA₂α^−/− MEFs were treated with PDGF in the presence or absence of specified inhibitors. BrdU⁺ population was evaluated in WT and cPLA₂α^−/− cells at 24 or 28 h, respectively. F) WT MEFs that conditionally express EGFPΔRaf-1:AR were serum deprived and pretreated with specified inhibitors prior to addition of R1881 to activate RAF. P27^Kip1 mRNA levels were assessed at 0 h (control) and 16 h. G) Cell cycle profile of unsynchronized WT and cPLA₂α^−/− MEFs in the presence or absence of 0.1 µM of FOXO1-specific inhibitor AS1842856 (AS). H) Percentage of cells in each phase of the cell cycle in WT and cPLA₂α^−/− MEFs in the presence or absence of 0.1 µM AS (n = 3). *P < 0.05; **P < 0.01.

Next, using pharmacologic inhibitors, we sought to determine the role of each pathway in the regulation of G₁ phase downstream of the PDGF receptor. WT and cPLA₂α^−/− MEFs were serum deprived and pretreated with specific inhibitors of MEK (PD98059), cPLA₂α (pyrrophenone), and PI3K (Ly294002), separately or in combination, and then were stimulated with PDGF. Phosphorylation of AKT and FOXO1 was assessed after 1 h, and protein and mRNA expression of cyclin D1 and p27^Kip1 was evaluated after 16 h. PD98059 had no effect on AKT or FOXO1 phosphorylation, whereas pyrrophenone and Ly294002, alone or in combination, down-regulated phosphorylation of both proteins (Fig. 9C). As expected, pyrrophenone did not change the phosphorylation state of AKT or FOXO1 in cPLA₂α^−/− cells. All 3 inhibitors blocked, to a certain degree, the PDGF-induced induction of cyclin D1 and repression of p27^Kip1 (Fig. 9C, D). Combination of inhibitors led to a marked change in cyclin D1 and p27^Kip1 expression close to the level observed in serum-deprived cells. These data indicate that these 3 pathways cooperatively regulate cyclin D1 and p27^Kip1 expression during G₁ phase. To determine the role of cPLA₂α, RAF/MEK/ERK, and PI3K/AKT pathways on G₁ progression, serum-starved WT and cPLA₂α^−/− MEFs were pretreated with specified inhibitors targeting each pathway and then were stimulated with PDGF to reenter cell cycle. BrdU incorporation was assessed in WT and cPLA₂α^−/− cells at 24 and 28 h, respectively. We chose these time points based on the maximum S phase population in each cell line as was established in Fig. 6F. We observed a significant decline in BrdU uptake in WT cells upon inhibition of each pathway, whereas in cPLA₂α^−/− cells, only MEK and PI3K inhibitors decreased BrdU incorporation (Fig. 9D). Here, we also observed a cumulative effect in the case of multiple treatments. These data are in line with our findings in Fig. 9D and show that these 3 pathways cooperatively regulate cell cycle progression in response to mitogenic stimulation.

As shown in Fig. 9D, E, pretreatment with the MEK inhibitor down-regulated cyclin D1 and up-regulated p27^Kip1 expression in both WT and cPLA₂α^−/− cells. This suggests that the regulation of the 2 proteins by the MEK/ERK pathway is independent of cPLA₂α expression. To further confirm this finding, MEFs expressing a conditionally active form of RAF-1 (EGFPΔRaf-1:AR) were generated (41). Serum-deprived MEFs remained untreated or were treated with specified inhibitors and then stimulated with androgen analog R1881 to activate EGFPΔRaf-1:AR. Activation of RAF led to a suppression of p27^Kip1 mRNA (Fig. 9F). Addition of pyrrophenone had no further effect on p27^Kip1 expression. PD98059 was used as a positive control and significantly inhibited suppression of p27Kip1 mRNA. Thus, our data indicate that the RAF/MEK/ERK pathway can affect cell proliferation but does so via a cellular process that is independent of the growth regulatory process of cPLA₂α. Cumulatively, the data presented in Fig. 9 suggest that the levels of p27^Kip1 and cyclin D1 expression are under the triple control of the RAF/MEK/ERK, PI3K/AKT, and cPLA₂α/PGE₂ pathways. Next, we analyzed the cell cycle profile of WT and cPLA₂α^−/− MEFs in the absence or presence of a specific FOXO1 inhibitor, AS1842856 (45). Addition of 0.1 µM AS1842856 to unsynchronized cPLA₂α^−/− MEFs for 24 h significantly increased the S population and significantly decreased the G₀/G₁ population to a level comparable to that of WT cells (Fig. 9G, H). This indicates that the inhibition of FOXO1 alone is sufficient to rescue the phenotype observed in cPLA₂α^−/− cells, confirming that cPLA₂α modulates G₁ progression mainly through FOXO1 activity.

DISCUSSION

Regulation of cell proliferation results from a dynamic balance between stimuli that promote cell proliferation and factors that inhibit cell growth. The mechanism of G₁ delay or arrest following cPLA₂α deletion may be explained by down-regulation of cyclin D1 and up-regulation of p27^Kip1. Cyclin Dl is a critical target of proliferative signals following growth factor stimulation of quiescent cells, and inhibition of its expression can cause a G₁ arrest (46). Concomitant with increased cyclin D1 expression, suppression of p27^Kip1 is necessary for growth factor-induced G₀/G₁ progression (47). In fact, p27^Kip1 is the only Cdk inhibitor whose protein expression decreases as mitogen-induced cells enter the cell cycle. Accordingly, sustained expression of p27^Kip1 is shown to cause G₁ arrest (45). p27^Kip1 protein levels, which vary during the cell cycle, are modulated both at the level of transcription or through changes in its half-life, which is in part controlled by the ubiquitin-proteasome pathway (47–49). In the absence of cPLA₂α, it is possible that a modest increase in p27^Kip1 levels might lead to inhibition of Cdk activity followed by a decrease in p27^Kip1 phosphorylation and then an increase in the half-life of p27^Kip1 protein. This leads to pRb hypophosphorylation, which delays entry into S phase. The diminished Cdk4 activity on pRb protein in cPLA₂α^−/− MEFs supports this hypothesis. In our synchronization experiments, we did not observe a delay in S or G₂/M phase or a change in cyclin A expression or Cdk2 activity. These findings strongly suggest that the delay or arrest does not occur at S or G₂/M phases but rather at the G₁/S phase transition. Based on these findings, we propose that depleting cPLA₂α results in a prolonged or blocked G₁ phase, resulting in fewer rounds of proliferation, in turn leading to reduced cell number.

Lack of cPLA₂α in both zebrafish and murine cells resulted in a decrease in cyclin D1 and an increase in p27^Kip1 levels, indicating a high conservation of function between lower vertebrates and mammals. In zebrafish, lack of both cpla₂αa and b during embryo development resulted in a degenerative phenotype. Reduction of cyclin D1 expression was associated with impaired development of the eye and the head region in zebrafish embryos (15). It is noteworthy that cPLA₂α^−/− mice do not show any abnormality during development (5). This may be explained by redundancy provided by other PLA₂ family proteins during mouse embryogenesis. It is also possible that cPLA₂α plays a more crucial role during early development in zebrafish than in mice.

We further show that cPLA₂α acts via its PGE₂ product acting on the EP₄ receptor and via the PI3K/AKT pathway, which leads to FOXO1 hyperphosphorylation (Fig. 10). There is an increasing body of evidence indicating that the G-coupled EP₄ receptor activates PI3K/AKT kinase, and activation of the AKT is now recognized as a critical mechanism associated with activation of multiple downstream pathways that increase cell proliferation and cell survival (44). PI3K/AKT-mediated phosphorylation of FOXO1 in this context is a highly conserved pathway also present Caenorhabditis elegans, where the PI3K and AKT homologs (AGE-1 and AKT, respectively) regulate the activity of a forkhead transcription factor, DAF-16, in a pathway involved in the regulation of survival in response to nutrient starvation (50). In addition to FOXO1 hyperphosphorylation, we also observed a decrease in transcriptional levels of FOXO1 in WT cells. This may also be explained through reduced PI3K/AKT signaling in WT cells because AKT activation was shown to inhibit transcriptional activation of FOXO1 (44).

Figure 10.

The model proposed in this study. PDGF is known to activate PI3K/AKT and RAF/MEK/ERK pathways. ERK phosphorylates cPLA₂α, which leads to activation of cPLA₂α and production of PGE₂. Binding of PGE₂ to EP₄ receptor activates the PI3K/AKT/FOXO1 pathway, ultimately enhancing G₁ progression through FOXO1 nuclear export and modulating cyclin D1 and p27^Kip1 expression. Moreover, the RAF/MEK/ERK signaling pathway regulates cyclin D1 and p27^Kip1 expression through cPLA₂α-independent mechanisms. MEK (PD98059), cPLA₂α (pyrrophenone), PI3K (Ly294002), and FOXO1 (AS1842856) inhibitors were used in this study to elucidate the role of each pathway in the regulation of the G₁ progression. C, cytoplasm; N, nucleus.

The FOXO1 transcription factor in its active form mediates cell cycle arrest, DNA repair, and apoptosis (51). FOXO1 controls cell proliferation and survival by regulating the expression of genes involved in cell cycle progression [p27^Kip1, p130 (RB2), cyclin D1, and B-cell lymphocytic leukemia proto-oncogene 6], and apoptosis (Bim, Fas ligand, TNF-related apoptosis-inducing ligand, and Bcl-XL) (52). Cell cycle arrest observed in our cPLA₂α-deficient cells can be explained by the increase in the activity of FOXO1 because a specific inhibitor of FOXO1 completely reversed the cell cycle defect in these cells. Despite a significant increase in FOXO1 activity in cPLA₂α-deficient cells, we did not observe an increase in apoptosis in these cells compared to WT cells. The most plausible explanation for this observation is that there is cell-type–specific gene regulation of pro- and/or antiapoptotic genes. This may depend on the type of tissue and the type of damage, and also other coexisting transcription factors. Distinct cofactors must exist that are involved in the activation or repression of the various FOXO1 target genes. In fact, FOXO1 in its active state has been shown to induce apoptosis in certain tissues; however, in other cell types, activated FOXO1 induces cell cycle arrest accompanied by a low level of cell death (53).

Our study identifies an important role for cPLA₂α in induction of cell cycle progression at the G₁ phase in MCs through inhibition of FOXO1 activity. Mesangial proliferation is characteristic of a number of chronic progressive glomerular diseases, and a large number of mediators, including PDGF, have been implicated as MC mitogens (24, 25). Moreover, FOXO1 is a tumor suppressor gene, and several molecules and strategies aiming to restore activity of FOXO1 are currently being developed for anticancer therapies. These findings lead us to suggest that antagonizing cPLA₂α activity may be effective in protecting tissues and organ function from disease mechanisms that involve excessive cell proliferation, including progressive acute and/or chronic glomerulonephritis and tumorigenesis. Because potent and specific inhibitors of cPLA₂α are currently available (22), this pathway could be exploited for therapeutic interventions.

Glossary

AA

arachidonic acid

Ad-cPLA₂α

adenoviral cytosolic phospholipase A₂α

AKT

protein kinase B

APC

adenomatous polyposis coli

BrdU

5-bromo-2′-deoxyuridine

Cdk

cyclin-dependent kinase

CM

conditioned medium

cPLA₂α

cytosolic phospholipase A₂α

EGFP

enhanced green fluorescent protein

EP₄

prostaglandin E₂ receptor 4

FBS

fetal bovine serum

FOXO1

Forkhead box protein O1

HEK

human embryonic kidney

HEPES

4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

hpf

hours postfertilization

Mb

megabase

MC

mesangial cell

MEF

mouse embryonic fibroblast

MO

morpholino

PDC

phosducin

PDGF

platelet-derived growth factor

PG

prostaglandin

PGE₂

prostaglandin E₂

PLA₂

phospholipase A₂

pRb

retinoblastoma

PTGS2

prostaglandin-endoperoxide synthase 2

RAF

rapidly accelerated fibrosarcoma

TCA

trichloroacetic acid

WT

wild-type