Epidermal growth factor (EGF) triggers nuclear calcium signaling through the intranuclear phospholipase Cδ-4 (PLCδ4)

Abstract

Calcium (Ca²⁺) signaling within the cell nucleus regulates specific cellular events such as gene transcription and cell proliferation. Nuclear and cytosolic Ca²⁺ levels can be independently regulated, and nuclear translocation of receptor tyrosine kinases (RTKs) is one way to locally activate signaling cascades within the nucleus. Nuclear RTKs, including the epidermal growth factor receptor (EGFR), are important for processes such as transcriptional regulation, DNA-damage repair, and cancer therapy resistance. RTKs can hydrolyze phosphatidylinositol 4,5-bisphosphate (PI(4,5)P₂) within the nucleus, leading to Ca²⁺ release from the nucleoplasmic reticulum by inositol 1,4,5-trisphosphate receptors. PI(4,5)P₂ hydrolysis is mediated by phospholipase C (PLC). However, it is unknown which nuclear PLC isoform is triggered by EGFR. Here, using subcellular fractionation, immunoblotting and fluorescence, siRNA-based gene knockdowns, and FRET-based biosensor reporter assays, we investigated the role of PLCδ4 in epidermal growth factor (EGF)-induced nuclear Ca²⁺ signaling and downstream events. We found that EGF-induced Ca²⁺ signals are inhibited when translocation of EGFR is impaired. Nuclear Ca²⁺ signals also were reduced by selectively buffering inositol 1,4,5-trisphosphate (InsP₃) within the nucleus. EGF induced hydrolysis of nuclear PI(4,5)P₂ by the intranuclear PLCδ4, rather than by PLCγ1. Moreover, protein kinase C, a downstream target of EGF, was active in the nucleus of stimulated cells. Furthermore, PLCδ4 and InsP₃ modulated cell cycle progression by regulating the expression of cyclins A and B1. These results provide evidence that EGF-induced nuclear signaling is mediated by nuclear PLCδ4 and suggest new therapeutic targets to modulate the proliferative effects of this growth factor.

Introduction

The spatial-temporal distribution of calcium (Ca²⁺) signals contributes to the versatility of this second messenger (1). For example, increases in Ca²⁺ within the cell nucleus selectively promote cellular events such as gene transcription (2), cell proliferation, and tumor growth (3). Moreover, nuclear Ca²⁺ signals can be regulated independently of cytosolic Ca²⁺ signals (4). This is possible because the nucleus contains the machinery necessary for Ca²⁺ mobilization (5–8). Specifically, the nuclear envelope is contiguous with the endoplasmic reticulum (ER)³ (9) and has invaginations that can reach deep within the nucleoplasm (10). These invaginations express InsP₃Rs, store and release Ca²⁺ in an InsP₃-sensitive fashion (5, 11), and have been referred to as the nucleoplasmic reticulum (NR) (5, 10). The distribution of InsP₃R isoforms and also Ca²⁺ signaling in the nucleus become altered in various disease states, such as fatty liver disease (12) and cholangiocarcinoma (13). Intranuclear InsP₃ is formed from PI(4,5)P₂ hydrolysis, which is induced by growth factors such as hepatocyte growth factor (7) and insulin (8). Some RTKs undergo nuclear translocation upon activation, which appears necessary for initiation of nuclear Ca²⁺ signals (7, 8), so this could be one mechanism to regulate Ca²⁺ release locally.

Nuclear localization of EGFR is of clinical relevance, as it is correlated with cancer prognosis (14) and relates to resistance to various cancer therapies (15, 16). Nuclear EGFR function has been the subject of extensive investigation, as has been nuclear targets of EGFR. EGFR can shuttle from the cell surface to the nucleus (17, 18) where it acts as a transcriptional regulator (19, 20), transmits signals (21, 22), and is involved in processes such as cell proliferation, tumor progression, and DNA repair and replication (23). However, it is unknown whether EGFR is linked to nuclear calcium release.

PLC could be involved in primary nuclear Ca²⁺ signaling because activation of InsP₃ receptors generally is involved in the release of Ca²⁺ in the nucleus (24). In mammals, the PLC family is composed of 13 isozymes, divided into six classes as follows: β, γ, δ, ϵ, ζ, and η, according to their structures (25). All these isozymes catalyze the reaction of PI(4,5)P₂ cleavage, generating InsP₃ and diacylglycerol (DAG), but each one possesses unique physiological functions (25). Some PLCs are described within the nucleus in specific cell types such as PLCβ1, PLCγ1, PLCδ1, PLCδ4, and PLCζ (26, 27). The activity of these nuclear PLCs is related to the regulation of several cellular processes, such as proliferation and differentiation. These particular isoforms also have been linked to specific diseases, including myelodysplastic syndromes (PLCβ1), neurological diseases (PLCγ1), and infertility (PLCζ and PLCδ1–4) (28).

PLCδ4 has been identified as a nuclear protein in different cell types and may be involved in proliferative processes (27, 29–31). It was first purified from regenerating rat liver protein extracts (32), and its gene was cloned (33) from a regenerating rat liver cDNA library, indicating a possible role of PLCδ4 in cell proliferation. Despite its nuclear localization, it is unknown how human PLCδ4 is activated within the nucleus. This work investigates whether EGFR activates nuclear Ca²⁺ signaling via PLCδ4.

Results

EGFR translocates to the nucleus and induces Ca²⁺ signals

Stimulation with EGF induces its RTK EGFR to translocate to the nucleus (18), similar to what has been shown for other RTKs, including the hepatocyte growth factor (HGF) receptor, c-Met, and the insulin receptor (7, 8). EGFR translocated to the nucleus in both primary rat hepatocytes and SKHep-1 cells, a human liver cancer cell line (Fig. 1A). EGFR protein levels in the nuclear fraction were higher than in nonstimulated control cells as soon as 2.5 min after stimulation with EGF (Fig. 1A). In SKHep-1 cells, EGFR decreased in the non-nuclear fraction as it increased in the nuclear fraction. The peak of EGFR nuclear translocation in this cell line was at 10 min (Fig. 1A). Super-resolution immunofluorescence showed that EGFR accumulated in the nucleus within 10 min of exposure to EGF (Fig. 1B). EGF stimulation induced bigger EGFR clusters throughout the cells, including at the nucleus, as demonstrated previously (15). These data demonstrate that EGF induces EGFR to translocate to the nucleus.

Figure 1.

EGF induces EGFR nuclear translocation and clathrin-mediated endocytosis-dependent Ca²⁺ signals. A, Western blot analysis of EGFR in non-nuclear and nuclear fractions isolated from resting (control) or EGF-stimulated hepatocyte and SKHep-1 cells. α-Tubulin and lamin B1 were used as purity controls for non-nuclear and nuclear fractions, respectively. Indicated kDa values represent the position of the referenced bands from pre-stained protein standard. Densitometric measurements of cellular fractions show that nuclear EGFR is maximal within 10 min of stimulation in SKHep-1 cells and within 2.5–5 min (p < 0.01) in hepatocytes when compared with the control (0 min). Values are scaled relative to the initial amount in the non-nuclear fraction and relative to the amount at 2.5 min in the nuclear fraction (mean ± S.E.; n = 3) (nuclear SKHep-1 fractions, one-way ANOVA, F(4,10) = 3.215; p = 0.0611; nuclear hepatocyte fractions, one-way ANOVA, F(4,10) = 8.530, p = 0.0029). B, gated STED super-resolution images of control SKHep-1 cells (right image) and 10 min after EGF stimulation (middle image). Serial optical sections were collected for three-dimensional reconstruction; y-z sections are shown at the right of each image. EGFR is represented in green; lamin B1 is in red, and the nuclear compartment marked in blue. Note that EGFR redistributes to the region of the nuclear envelope as well as within the nuclear interior (arrows). Bar = 10 μm. Scatter plot (right panel) showing the quantification of EGFR that co-localizes with the nucleus before and after EGF stimulation. n = 7–12 cells for each group. The image collection settings for fluorescence quantification was adjusted according to the cells stimulated with EGF to avoid nuclear-saturated pixels of EGFR clusters. ***, p < 0.001 (Student's t test). C, Western blotting (top) analysis of clathrin heavy chain 2 (CHC2) expression in nontreated, Lipofectamine only, control siRNAs-treated, or CHC2 siRNAs-treated SKHep-1 cells. α-Tubulin was used as a loading control. Bar graph shows the summary of the Western blottings (n = 4); ***, p < 0.001. (One-way ANOVA, F(5,18) = 34.41; p < 0.0001.) D, tracings represent Fluo-4/AM fluorescence intensity changes, normalized by the baseline, by the time of EGF stimulation. Lower panel, bar graph compiling peaks of Fluo-4/AM fluorescence intensity in EGF stimulated at 15 min, control siRNA-treated (n = 15), Lipofectamine only (n = 14), control siRNA 1 (n = 10), control siRNA 2 (n = 10), CHC2 siRNA 1 (n = 13), or CHC2 siRNA 2-treated (n = 13) SKHep-1 cells. Fluorescence changes over time from whole cells were normalized and represented as fluorescence intensity (F) by baseline fluorescence (F₀) multiplied by 100. ***, p < 0.001. (One-way ANOVA, F(5,69) = 18.78; p < 0.0001.) Experiments were performed on at least 3 different days.

Translocation of EGFR to the nucleus depends on clathrin-mediated endocytosis (17), so we used clathrin heavy chain 2 (CHC2) siRNAs to disrupt this endocytic pathway and thereby inhibit internalization of EGFR. Fig. 1C shows that siRNA treatment reduced CHC2 expression by 94 ± 3% using the CHC2 siRNA 1 and by 87 ± 8% using the CHC2 siRNA 2, relative to control (p < 0.001). Stimulation of control cells with EGF led to a gradual Ca²⁺ increase with some superimposed oscillations, similar to the Ca²⁺ signal pattern induced by other growth factors (7, 8), but CHC2 knockdown diminished the peak of EGF-induced Ca²⁺ signals by 85 ± 2% (p < 0.001) using the CHCH2 siRNA 1 and by 100 ± 2% (p < 0.001) using the CHCH2 siRNA 2, compared with control (Fig. 1D). This indicates that most of the Ca²⁺ signals induced by EGF depend on EGFR internalization.

EGF triggers intranuclear PI(4,5)P₂ hydrolysis and InsP₃ formation

EGF induces intracellular Ca²⁺ transients through InsP₃ (34), so we investigated whether EGF-induced InsP₃ formation in the nucleus was responsible for increasing intranuclear Ca²⁺. SKHep-1 cells were transfected with cytosolic or nuclear InsP₃-buffer constructs whose targeting was confirmed by subcellular localization of the monomeric red fluorescent protein (mRFP) (Fig. 2A). Fluo-4/AM was used to monitor Ca²⁺ release in response to EGF. Cytosolic InsP₃-buffer decreased the Ca²⁺ response by 50.5 ± 6.1% in the nucleus and by 53.8 ± 4.5% in the cytoplasm (Fig. 2B), but EGF-induced Ca²⁺ signals were decreased by 86 ± 2.7% in the nucleus and by 96 ± 3.9% in the cytoplasm in cells expressing the nuclear InsP₃-buffer (Fig. 2B). This suggests that EGF triggers nuclear Ca²⁺ signals by inducing the formation of InsP₃ within the nucleus, similar to what has been observed in liver cells stimulated with HGF (7) or insulin (8). To investigate whether EGF induces nuclear PI(4,5)P₂ hydrolysis to produce InsP₃, PI(4,5)P₂ was measured in nuclear fractions of hepatocytes. Nuclei from EGF-stimulated cells contained 64 ± 1.5% less PI(4,5)P₂ than nuclei from control cells (Fig. 2D). These findings provide evidence that EGF triggers nuclear PI(4,5)P₂ hydrolysis and local InsP₃ production to generate Ca²⁺ signals.

Figure 2.

EGF triggers Ca²⁺ release by nuclear InsP₃ and nuclear PI(4,5)P₂ hydrolysis. A, confocal images of SKHep-1 cells loaded with Fluo-4/AM expressing cytosolic InsP₃-buffer, nuclear InsP₃-buffer, or mRFP (control). Images show cells before EGF stimulus (baseline) and the peak (15 min) of EGF-induced Fluo-4/AM fluorescence intensity changes. Expression of constructs was checked by detection of mRFP. Bar = 10 μm. B, bar graph showing Fluo-4/AM fluorescence intensity peak in cytosolic and nuclear regions of control, cytosolic, or nuclear InsP₃-buffer expressing cells upon EGF stimulus (8–11 cells in each group: *, p < 0.05; **, p < 0.01; and ***, p < 0.001) (cytosol: one-way ANOVA, F(2,24) = 17.17, p < 0.0001; nuclear, one-way ANOVA, F(2,26) = 28.03; p < 0.0001.) C, average traces of SKHep-1 expressing cytosolic or nuclear InsP₃-buffer or control cells stimulated with EGF. Cytosolic or nuclear regions for fluorescent intensity measurements were selected using the assistance of the digital image of contrast (D.I.C) images, see squares at the nucleus and non-nuclear regions. Nuclear but not cytosolic InsP₃-buffer blocked the peak of EGF response in both compartments. D, bar graph represents the amount of PI(4,5)P₂ of nucleus isolated from control or EGF-stimulated hepatocytes (n = 6). 5 min of EGF stimulation reduces nuclear PI(4,5)P₂ by 64 ± 1.5% (p < 0.05) (Student's t test).

EGF stimulates intra-nuclear PKC activity

PI(4,5)P₂ hydrolysis generates not only InsP₃ but also DAG, which can activate PKC (24). Increases in nuclear DAG can either participate in translocation of PKCs, from the cytosol to the nucleus, or can directly activate PKCs that reside in the nucleus (35). To determine whether EGF triggers nuclear PKC activity, we used a Förster resonance energy transfer (FRET) reporter based on PKC activity and tagged to a nuclear localization signal (NucCKAR) (36). The nuclear localization of this construct was confirmed by intra-nuclear detection of cyan fluorescent protein (CFP) and yellow fluorescent protein (YFP) fluorescence by confocal microscopy (Fig. 3A). Phosphorylation of the reporter by PKC results in decreased rather than increased FRET (36), so the CFP/YFP ratio was monitored to reflect PKC activity. NucCKAR-expressing SKHep-1 cells were stimulated with EGF, phorbol 12–13-dibutyrate (PdBu) as a positive control for PKC activation, or arginine vasopressin (AVP) to reflect cytosolic activation of PKC, and then changes in FRET were followed for 30 min (Fig. 3B). PdBu increased CFP/YFP emission of NucCKAR (Fig. 3, B and C), and similar results were obtained with EGF treatment (Fig. 3, B and C), indicating that either stimulus increases nuclear PKC activity. Unlike EGF, stimulation with AVP, which activates a prototypical G-protein–coupled receptor in the plasma membrane, did not increase the CFP/YFP emission ratio of NucCKAR (Fig. 3, B and C). This result suggests that subplasmalemmal formation of DAG in response to AVP does not activate nuclear PKC. Treatment with the broad-range PKC inhibitor Gö6983 (Fig. 3D) decreased the EGF-induced CFP/YFP emission ratio, providing additional evidence that EGF triggers PKC activity within the nucleus.

Figure 3.

EGF induces nuclear PKC activity. A, confocal images of a representative SKHep-1 cell expressing NucCKAR shows CFP and YFP emission as well as a transmitted image of the cell, confirming nuclear localization of the construct. Bar = 10 μm. B, representative tracings of NucCKAR CFP/YFP emission normalized by baseline show changes in emission ratio during 30 min of perfusion with 200 nm phorbol 12–13-dibutyrate (PdBu) (n = 6), 100 ng/ml EGF (n = 10), or 500 nm AVP (n = 3). C, bar graph compiling the average results shown in B. NucCKAR FRET ratio change in EGF-stimulated cells in the presence (n = 8) or absence (n = 10) of the PKC inhibitor Gö6983 (250 nm). The PKC inhibitor significantly attenuates EGF-induced NucCKAR FRET changes: *, p < 0.05 (one-way ANOVA, F(3,23) = 5.645; p = 0.0002).

PLCγ1 is localized to the cytosol and PLCδ4 is in the nucleus

PLC mediates PI(4,5)P₂ hydrolysis (28) and formation of InsP₃ and DAG (29, 30), so the relative role played by several candidate PLC isoforms during EGF stimulation was investigated. PLCγ1 is the primary isoform typically thought to bind to and be activated by RTKs (1). PLCδ4 has been recognized as a predominantly nuclear PLC isoform, and its expression is increased in liver regeneration (33), so we compared the localization of these two isoforms in SKHep-1 cells. Confocal immunofluorescence showed that PLCγ1 is mostly localized to the cytosol, and this pattern does not change upon EGF stimulation (Fig. 4A). In contrast, PLCδ4 was detected exclusively in the nucleus of SKHep-1 cells, demonstrated by both confocal microscopy (Fig. 4B) and Western blotting (data not shown). The subcellular localization of these PLC isoforms was also assessed in primary rat hepatocytes. Similar to what was observed in SKHep-1 cells, PLCγ1 was found in non-nuclear fractions of both control and EGF-stimulated cells, whereas PLCδ4 was only detected in the nuclear fractions (Fig. 4C). The amount of PLCγ1 or PLCδ4 in each cell fraction was not altered by stimulation with EGF (Fig. 4C), indicating that there is no translocation of these PLCs between nuclear and non-nuclear compartments in response to EGF. Further examination showed that EGF stimulation for up to 10 min does not induce PLCγ1 to translocate from the cytoplasm to the nucleus (Fig. 4D).

Figure 4.

PLCγ1 is a cytosolic and PLCδ4 is a nuclear protein. A, representative confocal images of SKHep-1 cells before and after EGF stimulation showing PLCγ1 in red, EGFR in green, and TO-PRO nuclear dye in blue. Merged images reveal that PLCγ1 is predominantly expressed outside of the nucleus in both conditions (n = 15 cells). B, representative confocal images of control and EGF-stimulated SKHep-1 cells showing PLCδ4 in red, which is exclusively intra-nuclear (n = 15 cells). C, Western blotting showing PLCγ1 and PLCδ4 expressions in non-nuclear and nuclear fractions from control or EGF-stimulated hepatocytes. α-Tubulin and lamin B1 were used as purity controls for nuclear and non-nuclear fractions. Indicated kDa values represent the position of the reference bands from pre-stained protein standard. Subcellular localizations of PLCγ1 and PLCδ4 in hepatocytes are not altered by EGF stimulation for 10 min (n = 3). D, tracings of normalized amounts of PLCγ1 in non-nuclear and nuclear fractions of SKHep-1 cells upon EGF stimulation with different time points, quantified from Western blotting experiments (n = 3).

PLCδ4 participates in EGF-induced nuclear PI(4,5)P₂ hydrolysis, InsP₃ signaling, Ca²⁺ signals, and PKC activity

An siRNA approach was employed to determine the relative roles of PLCδ4 and PLCγ1 in EGF-induced nuclear PI(4,5)P₂ hydrolysis and Ca²⁺ signaling. Specific siRNAs reduced PLCγ1 expression by 90 ± 14 and 91 ± 27% using siRNAs 1 and 2, respectively. PLCδ4 expressions were reduced by 75 ± 7 and 86 ± 5% using siRNAs 1 and 2, respectively (p < 0.01) (Fig. 5A). Nuclear fractions were isolated from these cells after EGF stimulation, and PI(4,5)P₂ was quantified. EGF significantly increased nuclear PI(4,5)P₂ hydrolysis in control siRNAs and PLCγ1 siRNA-treated cells, but not in PLCδ4 siRNA-treated cells (Fig. 5B). Furthermore, cells treated with PLCδ4 siRNAs significantly reduced nuclear InsP₃ production, nuclear PKC activity, and Ca²⁺ signals, but this was not seen in PLCγ1 siRNA-treated cells (Fig. 5, C–E). These results suggest that PLCδ4 but not PLCγ1 participates in nuclear PI(4,5)P₂ hydrolysis, InsP₃ production, PKC activation, and Ca²⁺ signaling triggered by EGF stimulation.

Figure 5.

Intra-nuclear PLCδ4 mediates EGF-induced nuclear PI(4,5)P₂ depletion and Ca²⁺ signals. A, Western blotting showing PLCγ1 and PLCδ4 expression in SKHep-1 cells under control, Lipofectamine, control siRNAs, siRNAs for PLCδ4, or PLCγ1 siRNAs-treated conditions. α-Tubulin was used as a loading control. The molecular mass of each protein analyzed is on the right side. Bar graphs show the summary of the Western blottings. Specific siRNAs reduced PLCγ1 expression by 90 ± 14 and 91 ± 27% using siRNAs 1 and 2, respectively (one-way ANOVA, F(7,16) = 26.60; p < 0.001). PLCδ4 expressions were reduced by 75 ± 7 and 86 ± 5% using siRNAs 1 and 2, respectively (**, p < 0.01) (one-way ANOVA, F(7,32) = 5.808; p = 0.0002). ns, not significant. PLCγ1 siRNAs do not alter each PLCδ4 expression level or PLCδ4 siRNAs do not alter each PLCγ1 expression levels. B, graph shows PI(4,5)P₂ quantification in nuclei isolated from SKHep-1 cells before (−EGF group) and after 10 min of stimulation with EGF (+EGF groups). EGF-induced PI(4,5)P₂ hydrolysis was blocked by PLCδ4 siRNAs, but not by PLCγ1 siRNAs, compared with cells not treated with EGF (−EGF). Observe that the −EGF group has no difference with PLCδ4 siRNA-treated cells, or PLCγ1 has no difference with EGF-treated cells. n = 3. **, p < 0.01 (one-way ANOVA F(8,18) = 88.26; p < 0.0001). C, left: representative average tracings of nuclear InsP₃ sensor (FIRE-1nuc) shows changes in YFP/CFP emission normalized by baseline in EGF-stimulated groups over 20 min. Smoothing filter was used to better discern the tracings. Right: bar graph compiling the average results shows peak intensity in control (n = 11), control siRNA 1 (n = 13), control siRNA 2 (n = 13), PLCδ4 siRNA 1 (n = 11), PLCδ4 siRNA 2 (n = 12), PLCγ1 siRNA 1 (n = 11), or PLCγ1 siRNA 2-treated (n = 11) SKHep-1 cells. InsP₃ production is significantly reduced by PLCδ4 siRNAs (100 ± 0.5% for both siRNAs, *, p < 0.05), but not by PLCγ1 siRNAs, compared with control groups (one-way ANOVA, F(7,93) = 6.717; p < 0.0001). D, left: representative average tracings of NucCKAR CFP/YFP emission normalized by baseline shows changes in emission ratio during 25 min of perfusion with EGF. Smoothing filter was used to better discern the tracings. Right: bar graph compiling the average results shows peak intensity in control (n = 12), control siRNA 1 (n = 11), control siRNA 2 (n = 9), PLCδ4 siRNA 1 (n = 7), PLCδ4 siRNA 2-treated (n = 7), PLCγ1 siRNA 1 (n = 6), or PLCγ1 siRNA 2-treated (n = 8) SKHep-1 cells. PKC activation is significantly reduced by PLCδ4 siRNAs (100.7 ± 0.4% for siRNA 1 and 100 ± 0.7% for siRNA 2, *, p < 0.05), but not by PLCγ1 siRNAs, compared with control groups (Kruskal-Wallis, H = 37.85, df = 7, p < 0,0001; Dunn's multiple comparisons test). E, left: average tracings represent Fluo-4/AM fluorescence intensity changes in response to EGF stimulation. Fluorescence changes over time from whole cells were normalized and are represented as fluorescence intensity (F) divided by baseline fluorescence (F₀), multiplied by 100%. Right: bar graph compiling the average results shows peak intensity in control (n = 20), Lipofectamine (n = 21), control siRNA 1 (n = 13), control siRNA 2 (n = 10), PLCδ4 siRNA 1 (n = 13), PLCδ4 siRNA 2-treated (n = 10), PLCγ1 siRNA 1 (n = 5), or PLCγ1 siRNA 2-treated (n = 9) SKHep-1 cells. Ca²⁺ is significantly reduced by PLCδ4 siRNAs (95.9 ± 2.5% for siRNA 1 and 96.9 ± 2.6% for siRNA 2, *, p < 0.05) but not by PLCγ1 siRNAs, compared with control groups (one-way ANOVA, F(7,92) = 15.66; p < 0.0001).

PLCδ4 participates in cell cycle progression

Nuclear Ca²⁺ is important for cell proliferation (3), an important downstream effect of EGF. Therefore, we investigated whether PLCδ4 is involved in EGFR-mediated cell proliferation. Cell proliferation as measured by either cell counting (Fig. S1) or BrdU incorporation (Fig. 6A) was decreased by over 67% in cells in which PLCδ4 expression was reduced. SKHep-1 proliferation was restored when PLCδ4 expression was recovered (Fig. S1). Cell death does not mediate the decreased proliferation because annexin V–FITC and propidium iodide assays did not reveal cell death between control and siRNA-treated groups (Fig. 6B). Next, the role of PLCδ4 in cell cycle progression was investigated. To understand the basis for this, the effects of PLCδ4 on the expression of various cyclins were studied (Fig. S1). Expression of cyclin B1, a G₂/M-phase checkpoint protein, and cyclin A, an S-phase checkpoint protein, was decreased in PLCδ4 knockdown cells compared with cells treated with control siRNA (Fig. 6C). A nuclear InsP₃-buffer also decreased the expression of cyclins A and B. These findings suggest that nuclear PLCδ4 affects cell cycle progression, in part by affecting cyclin expression.

Figure 6.

Knockdown of PLCδ4 diminishes cell growth and decreases cyclins A and B1 without affecting cell death. A, effects of PLCδ4 depletion on cell proliferation. BrdU measurements were performed 48 h after transfection. *, p < 0.05 (EGF−, one-way ANOVA, F(5,35) = 8.042; p < 0.0001; EGF+, one-way ANOVA, F(5.35) = 9.570; p < 0.0001). B, quantification of cells labeled with annexin V-FITC and propidium iodide. The bar graph shows the percentages of cells labeled with either annexin V-FITC or propidium iodide or double-labeled with annexin V-FITC plus propidium iodide compared with all cells from each field. Each treatment was made in triplicate, and for each one, three different images were counted using ImageJ software; then, the means were used for statistical analysis. All groups were compared with control groups (control, Lipofectamine, and control siRNAs). Doxorubicin (Dox) 10 μg/ml was used as positive control, and it differs from all experimental groups. ***, p < 0.001. Annexin V staining did not reveal differences in the level of cell death between control and siRNA-treated groups (one-way ANOVA, F(6,71) = 20.64; p < 0.0001). C, Western blotting showing cyclins A and B1 expression in SKHep-1 cells under controls or siRNA-treated conditions. α-Tubulin was used as a loading control. Representative image of three independent experiments. Middle and right graphs, quantification of the Western blottings of cyclins that were normalized with the expression of α-tubulin (n = 3). *, p < 0.05, and ***, p < 0.001 (cyclin A, one-way ANOVA, F(6,14) = 17.92; p < 0.0001; cyclin B1, one-way ANOVA, F(5,12) = 69.17; p < 0.0001). D, Western blotting showing the effects of nuclear InsP₃ buffer on the expression of the cyclins A and B1. Right graph, quantification of the Western blottings of cyclins that were normalized with the expression of α-tubulin (n = 3). **, p < 0.01 (cyclin A and B1 used a Student's t test).

Discussion

EGFR has several direct effects on signaling in the nucleus, including interaction with transcription factors (37) and phosphorylation (21). This work extends this repertoire of intranuclear actions by showing that EGFR translocates to the nucleus of hepatic cells to initiate InsP₃-mediated Ca²⁺ signals. EGFR likely translocates from the plasma membrane to the nucleus via the ER (20, 23), and a similar trafficking route has been described for both viruses and toxins as well (38–40). Trafficking of other RTKs from the plasma membrane to the nucleus has also been described. For example, c-Met that is biotinylated at the plasma membrane can be recovered from the nucleus after cells are stimulated with HGF (7). Similarly, fluorescently-labeled EGF can be tracked from the cell membrane to the cell nucleus over a 10-min period (18). Furthermore, super-resolution imaging can be used to quantify the amount of EGF/EGFR clusters that accumulate in the nucleus (18). The peak in nuclear translocation of EGFR coincides with the peak in Ca²⁺ signals (Fig. 1, A–C). This work builds on the previous observation that translocation of EGFR to the nucleus depends on dynamin and clathrin-mediated endocytosis (17) by showing that knockdown of clathrin reduces EGF-induced Ca²⁺ signals. These findings are consistent with the idea that EGFR must translocate to the nucleus in order to initiate nuclear Ca²⁺ signals, similar to what has been shown for c-Met (7).

Ca²⁺ signaling in the nucleus rather than cytosol is important for proliferation and tumor growth (3). Nuclear Ca²⁺ can also regulate gene expression (2, 4). Thus, increases in nuclear Ca²⁺ could be one mechanism for EGF to promote those cellular processes. The nucleus contains the machinery necessary for InsP₃ receptor-mediated Ca²⁺ release (5, 7, 8, 24). Although PLC-mediated PI(4,5)P₂ hydrolysis and formation of InsP₃ and DAG within the cytoplasm have been described extensively (1), there are fewer studies showing that PI(4,5)P₂ hydrolysis also occurs within the nucleus. This work provides evidence that EGF triggers PI(4,5)P₂ hydrolysis and InsP₃ formation within the nucleus (Figs. 2 and 5), similar to what has been observed in response to stimulation with HGF (7) or insulin (8).

PLC within the nucleus is involved in signal-transduction pathways that are separate from those activated by isoforms localized in other compartments (41). The most common paradigm links PLCβ isoforms with G-protein–coupled receptors, whereas RTKs typically are associated with PLCγ isoforms (42). For example, the activated nuclear metabotropic glutamate 5 (mGlu5) receptor couples to G_q/11 and a nuclear isoform, PLCβ1, to generate the InsP₃-mediated release of Ca²⁺ from Ca²⁺-release channels in the nucleus of primary striatal neurons (43). This work provides evidence that EGF induces hydrolysis of nuclear PI(4,5)P₂ by the intra-nuclear PLCδ4, rather than by PLCγ1 (Fig. 5). It remains unclear whether PLCδ4 is activated directly or indirectly, as a direct binding of PLCδ4 with nuclear EGFR could not be demonstrated by co-immunoprecipitation (data not shown).

There are a number of previous studies about the function of nuclear PLCs. Some studies suggest that PLC positively regulates the cell cycle. PLCβ1 can regulate insulin-like growth factor-1–stimulated DNA synthesis, and its overexpression increases proliferation in the absence of mitogenic stimuli (44, 45). Furthermore, PLCβ1 in the nucleus regulates transcriptional levels of genes such as cyclin D3 and c-jun activation in skeletal muscle cells (45, 46). This study demonstrates that PLCδ4 also plays a role in cellular proliferation and specifically in cell-cycle control. Knockdown of PLCδ4 leads to decreased cell proliferation without affecting apoptosis in SKHep-1 cells (Fig. 6). The cell cycle profile is accompanied by reduced expression of cyclins A and B1 (Fig. 6). Consistent with this, decreased cyclin B1 could interfere with the transition through the G₂/M phase of the cell cycle. Furthermore, it was demonstrated that PLCδ4 knockdown impaired human mesenchymal stem/stromal cells proliferation, without inducing cell death (31).

The direct target of PLC signaling is typically PKC, which is activated by the PI(4,5)P₂ hydrolysis products DAG and InsP₃ (35). Because EGF activates PKC within the nucleus (Fig. 3), and several nuclear proteins are PKC substrates, this may also be relevant for processes such as gene expression and changes in chromatin structure (47). Activation of PKC may affect cell cycle progression as well (48, 49). PKCα is necessary for PLCβ1-mediated regulation of cyclin D3 and cell proliferation in human erythroleukemia cells (50), and PKCα regulates cyclin B1 leading to effects on G₂/M transition. PKC inhibitors can decrease cyclin B1 expression consequent to PKC down-regulation (51). The increase in nuclear DAG level observed during the G₂ phase is sufficient to activate nuclear PKCα/PKCβ. PKC phosphorylates lamin B1 at sites involved in mitotic disassembly of the nuclear lamina, one of the most critical events in the G₂/M transition (52).

In light of these considerations, the current findings identify novel signaling events that may link EGFR nuclear translocation to stimulation of cell growth (Fig. 7). Activation of this signaling pathway may, in turn, be important for hepatocyte proliferation, especially because liver regeneration is absolutely dependent on the presence of either EGF or HGF (53). This is also consistent with the observation that PLCδ4 expression is increased in liver regeneration as well as in liver cancer cell lines (33, 55). Furthermore, it is consistent with data from PLCδ4 knockout mice that show increases in bromodeoxyuridine (BrdU) are delayed after partial hepatectomy, although liver regeneration is not delayed (55). Therefore, strategies to selectively inhibit PLCδ4 could be important for selective inhibition of liver cancer, which is the third leading cause of cancer death worldwide.

Figure 7.

Proposed model: EGF triggers nuclear PI(4,5)P₂ hydrolysis, local InsP₃ production, and Ca²⁺ release into the nucleoplasm through intra-nuclear PLCδ4, which mediates cell proliferation. Scheme of the proposed EGF-induced nuclear Ca²⁺ signaling pathway. Upon EGF binding EGFR is activated and initiates signal transducer cascades. EGFR is shuttled to the nucleus to induce downstream targets. PLCδ4 is a nuclear enzyme induced by EGF stimulus and leads to PI(4,5)P₂ hydrolysis that generates InsP₃ and DAG. InsP₃ binds to InsP₃R present on the inner nuclear envelope triggering Ca²⁺ release in the nucleoplasm. DAG can activate nuclear-resident PKCs. Ca²⁺ and PKCs can activate downstream targets which ultimately result in cell proliferation.

Experimental procedures

Cell culture and EGF stimulus

The SKHep-1 is a human liver cancer cell line (ATCC, Manassas, VA). It was cultured at 37 °C in 5% CO₂ in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) containing 10% fetal bovine serum, 1 mm sodium pyruvate, 100 units/ml penicillin, and 100 μg/ml streptomycin (Life Technologies, Inc.).

Hepatocytes were isolated from healthy male Sprague-Dawley rats (190–200 g; Charles River Laboratories), as described previously (56). Primary cells were cultured at 37 °C in 5% CO₂ in Williams' medium E (Life Technologies, Inc.) containing 10% fetal bovine serum, 100 units/ml penicillin, and 100 μg/ml streptomycin and plated on collagen-coated coverslips (50 μg/ml) (BD Biosciences). Hepatocytes were used within 24 h of isolation. All animal studies were approved by the Yale University Institutional Animal Care and Use Committee under Protocol no. 2018-07602. Serum-starved (14–16 h) cells were treated with 100 ng/ml human recombinant EGF (Life Technologies, Inc.) or recombinant rat EGF (R&D Systems) at 37 °C in 5% CO₂.

Subcellular fractionation and Western blotting

Cells were washed twice with ice-cold 10 mm PBS, pH 7.4 (PBS) (Sigma), harvested by scraping, and lysed in a buffer with 20 mm HEPES, pH 7.0, 10 mm KCl, 2 mm MgCl₂, 0.5% Nonidet P-40. After incubation on ice for 10 min, the cells were homogenized by vortex. The homogenate was centrifuged at 1500 × g for 5 min to sediment the nuclei. The supernatant was then centrifuged at a maximum speed of 16,100 × g for 20 min, and the resulting supernatant formed the non-nuclear fraction. The nuclear pellet was washed three times with lysis buffer to remove any contamination from cytoplasmic membranes. The isolated nuclei were resuspended in NETN buffer (150 mm NaCl, 1 mm EDTA, 20 mm Tris-HCl, pH 8.0, 0.5% Nonidet P-40), and the mixture was sonicated briefly to aid nuclear lysis. Nuclear lysates were collected after centrifugation at 16,100 × g for 20 min at 4 °C. Protease and phosphatase inhibitors (Sigma) were added to all buffers. Proteins were quantified by Bradford assay (Sigma).

Western blotting was performed and detected as described previously (7). In brief, the primary antibodies used were as follows: rabbit polyclonal anti-EGFR (1:500, Santa Cruz Biotechnology, catalog no. SC3); mouse monoclonal anti-EGFR clone 8G6.2 (1:500, Millipore/Sigma, catalog no. 05-1047); mouse monoclonal anti-α-tubulin (1:2000, Sigma, catalog no. T6199); rabbit polyclonal anti-lamin B1 (1:2000, Abcam, catalog no. ab16048); monoclonal anti-clathrin (1:500, Abcam, catalog no. b21679); mouse monoclonal anti-PLCγ1 (1:1000, Cell Signaling Technology, catalog no. I2822); rabbit polyclonal anti-PLCδ4 (1:500, Santa Cruz Biotechnology, catalog no. H-250); rabbit polyclonal Erk1/2 (1:1000; Cell Signaling, catalog no. 9101); rabbit monoclonal anti-phospho-Erk1/2 Thr-202/Tyr-204 (1:1000; Cell Signaling, catalog no. 4370); rabbit monoclonal anti-AKT (1:1000, Cell Signaling, catalog no. 4691); rabbit monoclonal anti-phospho-AKT Thr-308 (1:1000; Cell Signaling, catalog no. 13038), cyclins A, B1, D1, D2, D3, E, E2 and H (1:500 to 1000, Cell Signaling Technology, Kit catalog no. 9869). The membranes were developed using ECL Plus (GE Healthcare). Subsequently, the films were scanned and analyzed using ImageJ software. All antibodies used were found to have the correct molecular weight by Western blotting. The antibodies for clathrin, PLCδ4, and PLCγ1 were validated with at least two specific siRNAs.

Immunofluorescence

Confocal and super-resolution immunofluorescence imagings were performed as described previously (18). In brief, cells were labeled with rabbit polyclonal anti-EGFR (1:250, Santa Cruz Biotechnology, catalog no. SC3) or mouse monoclonal anti-EGFR clone 8G6.2 (1:200, Millipore/Sigma, catalog no. 05-1047) and rabbit polyclonal anti-PLCγ1 (1:200, Cell Signaling, catalog no. I2822) and polyclonal anti-PLCδ4 (Santa Cruz Biotechnology, catalog no. H-250) and then incubated with secondary antibodies conjugated to Alexa Fluor 488 or 546 (Life Technologies, Inc.). Hoechst 33342 (Life Technologies, Inc.) was used as a marker for the nuclear compartment. Images were collected using a Leica SP8 Gated STED super-resolution microscope with a ×100, 1.4 NA objective lens.

siRNA transfection

Cells were transfected using Lipofectamine RNAiMAX reagent (Life Technologies, Inc.) according to the manufacturer's instructions. 50 nm siRNAs were used for clathrin heavy chain knockdown (CHC2 siRNA 1, 5′-GCAAUUAAUUAAGGCUGACCGUAtt-3′, Life Technologies, Inc., catalog no. s223263; CHC2 siRNA 2, 5′-AGGUGGCUUCUAAAUAUCAUGAACA-3′, IDT, catalog no. hs.Ri.CLTC.13.1). 25 nm siRNAs were used for PLCγ1 knockdown (PLCγ1 siRNA 1, 5′-GAAUCGUGAGGAUCGUAUAtt-3′, Life Technologies, Inc., catalog no. s10632; and PLCγ1 siRNA 2, 5′-GACCUCAUCAGCUACUAUGAGAAAC-3′, IDT, catalog no. hs.Ri.PLCG1.13.1). For PLCδ4 knockdown, 50 nm of the following siRNAs were used: ON-TARGETplus SMARTpool (PLCδ4 siRNA 1, 5′-CAAGAAGUUCAGCGGUUAU-3′, 5′-GCUCAAUCCCAUACCGACA-3′, 5′-GACCAAUGGCUGAGCGAUU-3′, and CAACAAGGUUACCGCCACA); Dharmacon, catalog no. L-005065-01 (PLCδ4 siRNA 2, 5′-AAGGCAGGUUGCAACUAGAAAUUCA-3′) (IDT, catalog no. hs.Ri.PLCD4.13.3). 50 nm of the following nontargeting siRNAs were used: Silencer^TM Select Negative Control No. 2 siRNA (control siRNA 1, Life Technologies, Inc., catalog no. 4390846) and IDT Scrambled negative control DsiRNA (control siRNA 2, 5′- CUUCCUCUCUUUCUCUCCCUUGUGA-3′, catalog no. 51-01-19-08). Further analysis was performed 48 h after transfection.

InsP₃-buffer constructs and calcium imaging

Generation of InsP₃-buffer constructs was as described previously (7). Briefly, the InsP₃-binding domain (residues 224–605) of the human type I InsP₃ receptor was tagged with an mRFP and with a nuclear localization signal or nuclear exclusion signal to generate the nuclear and cytosolic constructs, respectively. The adenoviruses were built by ViraQuest Inc. Cells were infected with 10 multiplicities of infection, and analysis was performed in 48 h. For calcium imaging, cells were incubated with 5 μm Fluo-4/AM (Life Technologies, Inc.) and placed in a custom-made perfusion chamber for EGF stimulation. Images were collected with a Zeiss LSM 710 DUO NLO. Nuclear and cytosolic regions were selected with the assistance of the brightfield images.

Nuclear PI(4,5)P₂ extraction and detection

PI(4,5)P₂ from samples was detected by a mass ELISA kit (Echelon Biosciences, UT). 5 min after EGF stimulation, cell nuclei were isolated as described previously (7, 8), washed three times with PBS, and followed by PI(4,5)P₂ extraction as recommended by the kit manufacturer. In summary, the protein was precipitated with 0.5 m TCA and followed by neutral lipid extraction using MeOH/CHCl₃ (2:1) and acidic lipid extracted with MeOH, CHCl₃, 12 m HCl (80:40:1). Lipids collected from split organic phase were dried before analysis.

FRET-based PKC activity reporter

The NucCKAR construct was developed by Gallegos et al. (36) and obtained from Addgene (plasmid no. 14869). Cells were transfected using FuGENE 6 (Promega), and microscope studies were performed in 48 h. 200 nm PdBu (Sigma) was used as a positive control. 250 nm Gö6983 (Tocris Bioscience, Bristol, UK) was used as a PKC inhibitor to confirm the specificity of EGF response. Pre-stimulation baseline was performed with only DMSO before perfusion with PdBu and Gö6983 as a vehicle control. Cells were placed in a custom-made perfusion chamber for agonist stimulation. CFP and YFP emission intensities were collected using a Zeiss LSM 710 DUO NLO with a ×63, 1.4 NA objective lens. With this construct, increased PKC activity results in decreased rather than increased FRET (36). Therefore, CFP/YFP rather than YFP/CFP emission ratio was calculated to reflect PKC activity (36) and was normalized using Microsoft Excel software.

FRET-based InsP₃ reporter

The FRET-based InsP₃-biosensor with nuclear localization signal (FIRE-1nuc) was developed by Kapoor et al. (54). Cells were transfected using FuGENE 6 (Promega), and microscopic studies were performed in 48 h. Cells were placed in a custom-made perfusion chamber for agonist stimulation. CFP and YFP emission intensities were collected using a Zeiss LSM 710 DUO NLO with a ×63, 1.4 NA objective lens. YFP/CFP emission ratio was calculated and normalized using ZEN (Zeiss), Microsoft Excel, and GraphPad Prism software.

Proliferation assays

SKHEP-1 cells were seeded into 96-well plates (Corning, NY) at a density of 1 × 10⁴ and transfected with Lipofectamine RNAiMAX with control siRNAs, PLCδ4 siRNAs, and PLCγ1 siRNAs. The cells were transfected with 50 nm of each siRNA. Cells without transfection reagent and with Lipofectamine were used as a control group. The Cell Proliferation ELISA kit (Roche Applies Science, offered by Sigma) was used according to the manufacturer's instructions. Briefly, after 48 h, cells were cultivated with BrdU (0.2 μm) for 2 h in serum-free media, fixated for 30 min, and incubated with anti-BrdU–POD for 90 min at room temperature. After three washing steps, 3,3′,5,5′-Tetramethylbenzidine was added, and the plate was read at 370 and 492 nm (Synergy 2, Biotek).

For proliferation curve assay, SKHep-1 cells were transfected with control siRNA 1 or PLCδ4 siRNA 1. After 48 h, cells were collected by trypsinization and counted in a Neubauer chamber using trypan blue for viability exclusion in triplicate.

For the proliferation recovery experiment, the Tet-pLKO-neo plasmid (Addgene ID. 21916) was used to express shRNAs under the control of a doxycycline promoter. The PLCD4 shRNAs target sequences were obtained from “The RNAi Consortium Collection” (https://portals.broadinstitute.org/gpp/public/),⁴ and the shRNAs were designed and cloned according to Addgene instructions. In summary, shRNA sequences were subcloned into Tet-pLKO-neo after digestion with AgeI and EcoRI. We used the following target sequences for the shRNAs: PLCδ4 shRNA 1 (5′-AGAGCAGCGTCGAGGGATATA-3′); PLCδ4 shRNA 2 (5′-ACTACCACTTCTACGAGATAT-3′); scramble shRNA 1 (5′-CCTAAGGTTAAGTCGCCCTCG-3′); and scramble shRNA 2 (5′-CUUCCUCUCUUUCUCUCCCUUGUGA-3′). SKHep-1 cells were transfected and selected with 500 ng/ml G418. Stable cells were treated with 1 μg/ml doxycycline for 48 h and then were kept without doxycycline for another 48 h for PLCδ4 expression recovery. BrdU assay was performed as described above.

Cell death assay

For this experiment, the cells were incubated with 10 μg/ml doxorubicin for 24 h, as a positive control for cell death. The cells were incubated with annexin V–FITC (Life Technologies, Inc.) and propidium iodide (Life Technologies, Inc.) for 1 h. Images were collected with a ×10 and ×20 objective lens on an Olympus IX70 inverted fluorescence microscope (Olympus America, Melville, NY), and the data obtained were quantified by counting cell labeled for annexin V–FITC, propidium iodide, double-labeled, and crossed with bright-field images used to obtain total cell count. Each treatment was made in triplicate, and three different images were counted, and the means were used for statistical analysis.

Statistical analysis

The significance of changes in treatment groups was determined by Student's t test or one-way analysis of variance with Tukey's multiple comparison tests if not otherwise stated, using GraphPad Prism software. Data are represented as mean ± S.E.

Author contributions

M. C. d. M., M. A. R., A. C. d. A. C., J. A. Q. A. F., D. S., A. M. G., M. H. N., and D. A. G. conceptualization; M. C. d. M., M. A. R., A. C. d. A. C., J. A. Q. A. F., M. K.-L., and D. A. G. formal analysis; M. C. d. M., A. C. d. A. C., J. A. Q. A. F., M. K.-L., M. H. N., and D. A. G. investigation; M. C. d. M., A. C. d. A. C., G. A. M., M. H. N., and D. A. G. methodology; M. C. d. M., M. A. R., J. A. Q. A. F., G. A. M., D. S., A. M. G., M. H. N., and D. A. G. writing-review and editing; M. A. R., A. C. d. A. C., J. A. Q. A. F., D. S., A. M. G., M. H. N., and D. A. G. writing-original draft; M. A. R. and D. A. G. project administration; A. C. d. A. C., J. A. Q. A. F., and M. K.-L. data curation; J. A. Q. A. F. and D. A. G. supervision; J. A. Q. A. F. and M. H. N. visualization; G. A. M. and M. H. N. resources; M. H. N. and D. A. G. funding acquisition.