Phosphoinositide 3-Kinase Activity Is Essential for all-trans-Retinoic Acid-induced Granulocytic Differentiation of HL-60 Cells¹

The phosphorylation of the d-3 position of the inositol ring of phosphoinositides is caused by a family of enzymes known as PI 3-Ks,³ of which the most studied, termed PI 3-K type I, is a heterodimer of a regulatory M_r 85,000 (or 55,000) protein and a p110 (reviewed in Ref. 1). The p85 subunit is a multidomain protein without intrinsic catalytic activity, which mediates association of the enzyme with activated tyrosine kinases (1). Five types of p85-like regulatory subunits exist: p85α, p85β, p55 (termed γ or p55^PIK), and two alternative splices of p85α (p55α and p50α; Refs. 1 and 2). All five interact with p110s, of which three distinct isoforms have been identified: α, β, and δ. p110 is responsible for the formation of phosphoinositides with a phosphate group at the d-3 position of the inositol ring (1). The 3-phosphoinositides are not substrates of any known PLC, are not components of the canonical phosphoinositide turnover pathway, and may themselves act as intracellular mediators rather than precursors of inositol phosphate second messengers (1).

Increased levels of PIP3 have been detected during diverse cellular processes, including mitogenesis, indicating that PI 3-K plays a role in cell growth and transformation (1, 3). However, PI 3-K is activated also by factors that trigger cellular differentiation, such as nerve growth factor in PC12 cells (4) and DMSO in murine erythroleukemia (MEL) cells (5), suggesting that PI 3-K and its lipid products may also play a role in cell maturation.

PI 3-K is present in the HL-60 promyelocytic cell line (6), a well-established model for the study of granulocytic differentiation when the cells are treated with retinoids, which act through a nuclear receptor belonging to the transcription factor family (7). To determine whether PI 3-K plays a role in HL-60 cell differentiation, we monitored the amount and the activity of the enzyme in whole cells during the differentiation triggered by ATRA. Here, we demonstrate that PI 3-K increases after ATRA treatment and that specific inhibition of PI 3-K activity and down-modulation of the protein prevent the differentiation of this cell line.

HL-60 cells were cultured and differentiated as described previously (8). Cell differentiation was quantitated as described previously (8). We also performed an estimate of the differentiation levels by a morphological examination of the nuclear shape on 4′-6-diamidino-2-phenylindole-stained cells. At least 200 cells were counted, and the percentage of granulocytes was evaluated. After 4 days, >90% of cells were fully differentiated, with results overlapping those obtained with CD11b expression (8).

The experiments of inhibition of PI 3-K activity were performed as follows: the cells were pretreated with 100 nm WT for 24 h and then cultured in the presence of ATRA plus WT for the indicated times. Because WT is unstable at 37°C in culture medium, it was added every 6 h to the cell culture. As controls, cells were incubated in the presence of only WT for the same times.

Membrane-depleted nuclei were isolated as described previously (8). Nuclear purity was assessed by ultrastructural analysis and marker enzyme assays, as reported previously (8, 9).

Proteins from the purified fractions (50 μg) were separated and blotted as described previously (8). Polyclonal antibody against the p85 subunit of PI 3-K was purchased from Upstate Biotechnology, Inc. (Lake Placid, NY).

The assays of PI 3-K activity were performed on 100 μg of protein of total cells in the presence of 10 mm HEPES, 1 mm EDTA (pH 7.5), and[γ-³²P]ATP (10 μCi/sample). Incorporation was for 15 min at room temperature, and the reaction was stopped by the addition of chloroform, methanol, and HCl (200:100:0.75, v/v/v), followed by two washes with chloroform, methanol, and 0.6 n HCl (3:48:47, v/v/v).

The lipid-containing organic phase was resolved on oxalate-coated TLC plates (Silica Gel 60; Merck) developed in isopropanol:acetic acid:H₂O (65:1:34, v/v/v).

The radiolabeled products were identified by comparison with standard lipids obtained after incubation of purified PI 3-K enzyme (generous gift of S. Volinia, Department of Morphology and Embryology, University of Ferrara) with phosphatidylinositol, phosphatidylinositol 4-phosphate, and phosphatidylinositol 4,5-bisphospate, respectively, in conditions that optimize PI 3-K activity.

Cells in PBS were plated onto 0.1% poly-l-lysine-coated glass slides, fixed, and processed as previously described (10). Samples were reacted with either a polyclonal (dilution, 1:80) or monoclonal (1:40) antibody directed against the M_r 85,000 subunit of PI 3-K (Upstate Biotechnology, Inc.) with identical results. Unless otherwise stated, photographs shown were obtained with monoclonal antibodies. Confocal laser scanning microscopy and image processing analysis were performed as described previously (10).

In the transfection experiments, we used the plasmid pcDNA3-p85α (11), containing a fragment of the bovine p85α cDNA (nucleotides 35–408) in sense or antisense orientation.

Exponentially growing HL-60 cells were transfected by Lipofectin carrier (Life Technologies, Inc. Paisley, United Kingdom). Briefly, 10 μg of specific plasmid DNA were mixed with 0.5 μg of pCH110 DNA (Pharmacia, Uppsala, Sweden), a plasmid DNA that constitutively expresses the β-galactosidase gene as control of transfection efficiency, and incubated in RPMI 1640 at room temperature for 10 min in the presence of Lipofectin reagent. A total of 2 × 10⁶ cells were collected by centrifugation, washed three times in PBS, resuspended in RPMI 1640, and mixed to the Lipofectin-DNA mixture. After 8 h of incubation at 37°C, the cells were collected by centrifugation, washed three times in PBS, resuspended in RPMI 1640 plus 10% fetal bovine serum, and grown for 48 h. After this period, the number of β-galactosidase-positive cells was counted to determine the efficiency of transfection, which ranged between 15 and 20%. Cells were then treated or not with 1 μm ATRA for 4 days. Daily, 100,000 cells were collected and cytocentrifuged onto a glass slide to determine the differentiation rates.

The results were expressed as means+ SD of three or more experiments performed in duplicate. The two-tailed Student’s t test was used for statistical analysis.

PI 3-K changes in HL-60 cells along the ATRA-induced differentiation pathway was first examined by immunochemical analysis (Fig. 1 a), which showed a significant increase of immunoreactivity in whole lysates of cells treated for 96 h with the agonist.

PI 3-K activity, measured as PIP3 recovery, increased significantly after ATRA treatment (Fig. 1,b). The contemporary administration of ATRA and WT, a fungal product known to inhibit agonist stimulated PI 3-K activity, largely reduced the recovery of PIP3 in both control and treated conditions (Fig. 2,a). The in situ analysis by fluorescence microscopy demonstrated that the treatment with WT impaired the morphological changes of the nucleus typically observed in the granulocytic-like differentiation of HL-60 cells (Fig. 2,b). Viability and cell proliferation were not affected by the administration of WT both with and without ATRA treatment, whereas the percentage of differentiated cells dramatically decreased (Fig. 2, c–e).

To further assess the role of PI 3-K in granulocytic differentiation, we transiently transfected HL-60 cells with pcDNA3-p85 containing a fragment of the p85 gene, in sense and antisense orientation. The response of the transfected cells was consistent with a PI 3-K modulation. In fact, when probed by Western blot, the amount of PI 3-K in cells transfected with the antisense construct was reduced (Fig. 3,a). As expected, the percentage of cells transfected with the sense construct and differentiating after ATRA treatment overlapped that of mock-transfected cells (data not shown) and constituted the control experiment (Fig. 3,b). On the contrary, ATRA treatment of cells transfected with the antisense construct resulted in a reduced number of terminally differentiated cells at all of the examined days, with an inhibition that closely reflects the percentage of transfected cells (Fig. 3 b).

Immunochemical (Fig. 4,a) and confocal (Fig. 4,b) analysis of the subcellular distribution of PI 3-K along the differentiation process, showed that the enzyme was discretely present in both cytoplasmic and nuclear compartments. In the cytosolic fractions no significant variations were detectable at the times examined, whereas in isolated nuclei, which showed a constitutive content of p85, PI 3-K increased progressively reaching the highest levels at 96 h, in correspondence with a fully differentiated phenotype (Fig. 4,a). Immunocytochemical analysis revealed the prevalence of the enzyme in the cytoplasm and confirmed a progressive increase of nuclear PI 3-K (Fig. 4 b).

PI 3-K has been shown to be directly associated (via its p85 subunit) with and activated by various normal and oncogenic tyrosine kinases, suggesting that it may regulate proliferation and/or other functions of normal and neoplastic cells (12). However, a role for PI 3-K in differentiation has also been shown. In rat PC12 cells, PI 3-K is necessary for multiple steps of neurite outgrowth during nerve growth factor-stimulated differentiation, and chronic treatment with WT inhibits the late phase of neurite extension (13). When murine erythroleukemia cells undergo erythroid differentiation by the addition of DMSO, an increase of PI 3-K activity occurs, and WT administration delayed the differentiation process (5). An involvement of PI 3-K in the adipocytic differentiation of 3T3-L1 cell was also reported. In this work, treatment with WT significantly inhibited adipogenesis of cells treated with isobutylmethylxanthine, dexamethasone, and insulin (14). More recently, it has also been shown that the inhibition of PI 3-K activity by LY294002 abrogates the insulin-like growth factor I-induced enhancement of macrophage-like differentiation in HL-60 cells treated with vitamin D₃ (15).

Here, we provide biochemical and morphological evidence for a specific involvement of PI 3-K activity in HL-60 granulocytic differentiation. PI 3-K is relevant to this process because reduction of PI 3-K activity by WT prevented the effects of ATRA. Accordingly, the decrease of protein expression obtained by antisense transfection inhibited granulocytic differentiation.

Although increased PI 3-K activity was originally associated with activation of receptor tyrosine kinases, it also occurs in response to activation of receptors lacking intrinsic tyrosine kinase activity, such as interleukin 2, interleukin 3, granulocyte macrophage colony-stimulating factor, and T-cell receptors (reviewed in Ref. 3). Our data provide the first evidence that also an agonist with a nuclear receptor induces PI 3-K activity, which promotes differentiation.

The action of PI 3-K in ATRA-driven differentiation is conceivably mediated by the downstream targets of PI 3-K products, which include Akt (PKB), p70^S6 kinase, cytoskeletal proteins (1), and PKC ζ (15). The link between PI 3-K and PKC has been recently elucidated and involves a PDK1 phosphorylation of the activation loop sites of PKC ζ, and PKC δ in a PI 3-K-dependent manner (16). It has been recently suggested that PIP3 activates PLC γ, revealing a novel mechanism for mutual regulation between these two enzymes that participate in the control of phosphoinositide metabolism (17, 18). Recent data also showed a cross-communication between nuclear PLC and PI 3-K pathways (19).

The ATRA-dependent differentiation process of HL-60 induced changes not only in the cytoplasm but also in the nuclear compartment. This observation is related to our previous findings describing the nuclear translocation of PLC β and γ and of specific PKC α and ζ during granulocytic differentiation of HL-60 cells (8, 20). On the other hand, nuclear PI 3-K has been described in cells other than HL-60, such as PC12 (21), Saos-2 (22), and rat liver (19) cells, suggesting that translocation of this enzyme to the nucleus is a widespread occurrence.

It is worthwhile to note that different inositol lipid-modifying enzymes are recruited during granulocytic differentiation, and in particular, PI 3-K may play a relevant role in this process, as demonstrated here by inhibition and down-modulation experiments. It can be informative to explore in future if an additional modulating pathway might rely on cross-talk between different enzymes.

We are grateful to Lewis Cantley for helpful discussion and suggestions. We thank Wataru Ogawa (Kobe University School of Medicine, Kobe, Japan) for the plasmids used in the transfection experiments and Stefano Volinia for a generous supply of the purified PI 3-K.