Activation of Protein Kinase C-ζ and Phosphatidylinositol 3′-Kinase and Promotion of Macrophage Differentiation by Insulin-Like Growth Factor-I¹

The protein kinase C (PKC)³ family consists of a group of phospholipid/diacylglycerol-dependent Ser/Thr kinases that are involved in signal transduction during hemopoiesis (1). Although all of the 10 identified PKC isoforms are dependent upon phosphatidylserine, members of the PKC family are grouped further into three subfamilies: conventional PKC, novel PKC (nPKC), and atypical PKC (aPKC) (2). In contrast to conventional PKCs, neither nPKCs nor aPKCs require Ca²⁺ for activation, and aPKCs do not bind phorbol esters. Overexpression of PKC-ζ, a member of the Ca²⁺-independent aPKC subfamily, causes monocytic differentiation (3). These PKC-ζ-transfected monocytes displayed an increase in the expression of the c-jun proto-oncogene as well as enhanced AP-1-binding activity. Similarly, PKC-ζ has recently been suggested to mediate the actions of insulin and serum mitogenic factors in adipogenesis (4). Moreover, platelet-derived growth factor (PDGF) treatment of NIH-3T3 cells leads to a direct association of PKC-ζ with Ras-GTP protein, which serves to recruit PKC-ζ to plasma membrane (5). Two phosphorylation products of phosphatidylinositol 3-kinase (PI 3-kinase), phosphatidylinositol-3,4-P₂ and phosphatidylinositol-3,4,5-P₃, are now known to activate PKC-ζ, implying that this aPKC may be a downstream target of PI 3-kinase (6).

Classical growth factor receptors with intrinsic tyrosine kinase activity, such as those for insulin and insulin-like growth factor-I (IGF-I), activate PI 3-kinase in a number of cells (7, 8, 9). IGF-I acts as a survival factor to inhibit apoptosis via a PI 3-kinase-dependent pathway in both murine promyeloid cells (9, 10) and neurons (8, 11, 12). The activation of PI 3-kinase has recently been shown to be a crucial event that occurs during normal B lymphocyte differentiation (13). Indeed, inhibition of PI 3-kinase by treatment with either wortmannin or LY294002 results in a complete block of CD40 ligand-induced Ig production. Similarly, inhibition of PI 3-kinase activity blocks the differentiation of myoblasts into myotubes, thus establishing that this enzyme is critical to the development of skeletal muscle cells (14). These data establish an essential role for PI 3-kinase in the differentiation of both B lymphocytes and muscle cells, and they suggest a potential role of this signaling pathway in myeloid development.

IGF-I and its major inducer, growth hormone, augment a number of functional activities of B cells, T cells, and macrophages (15, 16), and both of these proteins are synthesized by leukocytes (17, 18, 19). IGF-I also has been reported to promote the differentiation of granulocytes (20, 21) and B lymphocytes (22, 23). For example, Landreth et al. (1992) demonstrated that stromal cell-derived IGF-I increases expression of the cytoplasmic μ heavy chain during pro-B cell maturation (22) by promoting the expansion of IL-7-dependent pro-B cells (23). Although IGF-I stimulates PI 3-kinase in murine promyeloid cells (9), and the D3-phosphorylated phosphoinositides of this enzyme activate PKC-ζ (6), the potential role of IGF-I in macrophage development is unknown. We recently reported that HL-60 cells express abundant cell surface receptors for IGF-I that are biologically active, as assessed by a 2.5-fold increase in proliferation following addition of as little as 10 ng of IGF-I (24). In this study, we establish that IGF-I also promotes the differentiation of these cells into macrophages when cultured with vitamin D₃. IGF-I, but not vitamin D₃, induces a substantial increase in the activity of PKC-ζ. More importantly, inhibition of the enzymatic activity of PI 3-kinase abrogates the enhancement of macrophage differentiation caused by IGF-I and completely blocks the IGF-I-induced activation of PKC-ζ. These data show that IGF-I activates both PI 3-kinase and PKC-ζ and promotes the vitamin D₃-induced differentiation of promyeloid cells.

Cell culture medium was prepared from powdered RPMI 1640 (MediaTech, Herndon, VA) supplemented with 2 g/L sodium bicarbonate, 100 U/ml penicillin, and 100 μg/ml streptomycin (Sigma Chemical Co., St. Louis, MO). FBS (HyClone Laboratories, Logan, UT), which was heat inactivated for 30 min before use, as well as all other reagents contained <25 pg endotoxin/ml (Limulus amebocyte lysate assay; Associates of Cape Cod, Woods Hole, MA). The 1α,25-(OH)₂D₃ (vitamin D₃) was kindly provided by Dr. Milan Uskokovic (Hoffmann-La Roche, Nutley, NJ), and all-trans retinoic acid (RA) was purchased from Sigma Chemical Co. Both vitamin D₃ and RA were dissolved in ethanol. Human rIGF-I was purchased from Intergen (Purchase, NY), and nonglycosylated human rIGF-binding protein-3 (rIGFBP-3) was kindly provided by Celtrix (Santa Clara, CA). The mouse anti-human IGF type I receptor mAb (αIR3, IgG1k, free of sodium azide) was purchased from Oncogene Science (Uniondale, NY), the rat anti-human CD11b mAb (Mac-1, IgG2bk) was from BioSource International (Camarillo, CA), and the irrelevant isotype-matched murine IgG1 and rat IgG2b control mAbs were purchased from Sigma Chemical Co. The F(ab′)₂ fragment of FITC-conjugated goat anti-rat Ab was obtained from Cappel (Durham, NC). The mouse anti-phosphotyrosine (PY) Ab (4G10) and rabbit anti-rat PKC-ζ Abs were purchased from Upstate Biotechnology (Lake Placid, NY).

Human HL-60 promyeloid cells were kindly supplied by American Type Culture Collection (Rockville, MD). Cells were grown in RPMI 1640 medium supplemented with 10% FBS at 37°C in 7% CO₂ and 95% humidity. Cells in the exponential phase of the cell cycle were washed three times with RPMI and then deprived of residual serum growth factors by culturing for 24 h in serum-free medium supplemented with 12.5 μg/ml human transferrin (Sigma Chemical Co.) and 30 nM sodium selenite (Sigma Chemical Co.). In differentiation assays, cells were plated in this fresh serum-free RPMI medium at a density of 3 × 10⁵ cells/ml. Cells were cultured with vitamin D₃ (1 μM), RA (1 μM), or an equivalent concentration of ethanol (0.1%, v/v) for the indicated time period before analysis by flow cytometry.

In experiments in which a mAb directed against the IGF-I receptor (αIR3) was used, cells were incubated with αIR3 (5 μg/ml) or an irrelevant isotype-matched murine IgG1 Ab for 1 h at 37 C before other treatments. Similarly, to study the effect of IGF-binding proteins on macrophage differentiation, preincubation of IGFBP-3 (250 ng/ml) or the same amount of an irrelevant serum protein (BSA) was conducted 1 h before other treatments.

HL-60 cells undergo differentiation to a macrophage or a granulocyte phenotype following addition of vitamin D₃ and RA, respectively (25). Flow-cytometric methods were used to measure initiation of macrophage differentiation by characterizing surface expression of the α-subunit of the β₂ integrin CD11b. This cell membrane marker, which is absent on human immature myeloid cells, has been widely used as an early differentiation marker in HL-60 cells (26, 27). Cells were washed once in PBS (1.5 M NaCl, 19 mM NaH₂PO₄·H₂O, and 8.4 mM Na₂HPO₄) supplemented with 0.5% FBS and 0.25% BSA, and were then incubated with an anti-CD11b (Mac-1) mAb or its isotype-matched control mAb for 30 min at 4°C. After two washes with dilution buffer, cells were then incubated with FITC-conjugated goat anti-rat secondary Ab for another 30 min at 4°C. After two final washes, cells were fixed in PBS containing 1% formaldehyde until analysis by flow cytometry (EPICS V; Coulter Corp., Hialeah, FL). Bitmaps were established to include at least 5000 cells of uniform size using immunofluorescence intensity of control cells that were treated identically and stained with an isotype-matched mAb.

The apoptotic population was quantitated by using a double-staining procedure employing both propidium iodide and Hoechst 33342 (28). The charged dye propidium iodide (Sigma Chemical Co.) is excluded by cells with intact membranes, whereas Hoechst 33342 (Sigma Chemical Co.) stains DNA in fixed cells. Those cells expressing low fluorescence intensity with both Hoechst 33342 and propidium iodide were used to define and quantitate the number of apoptotic cells. Serum-deprived HL-60 cells (1 × 10⁶) were treated as indicated and washed once in PBS and incubated in 100 μl of propidium iodide (20 μg/ml) at 4 C for 30 min. Cells were washed three times and subsequently fixed in 1.9 ml 25% ethanol. Finally, 50 μl of Hoechst 33342 solution (60 μg/ml) was added, and cells were washed once more with PBS and then analyzed by flow cytometry.

Anti-PY-associated PI 3-kinase activity was determined in cell lysates, as described previously (9). Phosphatidylinositol lipid kinase activity was measured in anti-PY-associated immunoprecipitates, which is a more sensitive measure of PI 3-kinase activity than that found in immunoprecipitates using an Ab against the p85 subunit of this enzyme (29, 30). Serum-deprived cells (10⁸) were incubated with optimal concentrations of wortmannin (1 μM), LY294002 (25 μM), or DMSO solvent (0.1% v/v) for 1 h before other treatments. The optimal dose of both inhibitors was selected based upon our preliminary experiments to achieve a total blockage (>95% rather than 50% inhibition) of enzymatic activity (data not shown) so that we could better evaluate the role of PI 3-kinase in the activation of PKC-ζ and macrophage differentiation. Our working concentrations are consistent with other reports in which wortmannin (1 μM) (9) and LY294002 (50 μM) (31) effectively blocked PI 3-kinase activity in cells of the hemopoietic lineage. Cells were then incubated with vitamin D₃ (1 μM) for 30 min before stimulation with IGF-I (100 ng/ml) for 5 min. The cells were centrifuged at 400 × g and homogenized at 4°C in cell lysis buffer containing 1% Nonidet P-40, 50 mM Tris · HCl, 100 mM NaCl, 50 mM NaF, 10 mM sodium pyrophosphate, 2 mM orthovanadate, 2.5 mM benzamidine, 1 mM PMSF, and 1 μM DTT. Cell lysates were then incubated with a mouse anti-PY Ab 4G10 (2 μg; Upstate Biotechnology) at 4°C overnight and then with protein A-agarose beads for 2 h. The immunoprecipitated complex was washed twice with PBS/1% Nonidet P-40/1 mM DTT, twice with 0.5 M LiCl/1 mM DTT/100 mM Tris · HCl, and twice with 10 mM NaCl/1 mM DTT/10 mM Tris · HCl. Lipid kinase activity was measured directly by incubating the precipitated beads in a reaction mixture containing L-α-phosphatidylinositol (0.33 mg/ml), 20 mM Na HEPES, 0.4 mM EGTA, 0.4 mM NaPO₄, 10 mM MgCl₂, and 2 μCi/nmol [γ-³²P]ATP. After a 15-min incubation at room temperature, and following termination of the reaction by addition of 15 μl 4 N HCl, lipids were extracted once in chloroform/methanol (1:1 (v/v)) and once in 0.15 N HCl/methanol (1:1 (v/v)). Labeled phospholipid contained in the organic phase was separated by TLC in chloroform/methanol/ammonium hydroxide (75:58:17 (v/v)). Dried TLC plates were exposed to Kodak XRP5 autoradiographic film (Eastman Kodak, Rochester, NY) at −80°C for 24 h and subsequently exposed to PhosphorImager storage screens (Molecular Dynamics, Sunnyvale, CA). Individual signal intensity was measured on a series 400 PhosphorImager, using the ImageQuant 3.3 software (Molecular Dynamics) (9).

Cells (10⁸) were preincubated with wortmannin (1 μM), LY294002 (25 μM), or the DMSO diluent for 1 h. Cells were then treated with vitamin D₃ for 30 min before incubation with IGF-I (100 ng/ml) for 5 min. Subsequently, cells were extracted with cell lysis buffer containing 1% Nonidet P-40, 50 mM Tris · HCl, 100 mM NaCl, 50 mM NaF, 10 mM sodium pyrophosphate, 2 mM orthovanadate, 2.5 mM benzamidine, 1 mM PMSF, and 1 μM DTT. The anti-PKC-ζ rabbit antiserum (2 μg) or control rabbit serum was added to equal amounts of cell lysates (200 μg), followed by immunoprecipitation with protein A-agarose, as described above. The Ab immunoprecipitated equal amounts of PKC-ζ in each treatment group, as confirmed by Western analysis using the same anti-PKC-ζ Ab (data not shown). The washed immununoprecipitates were subjected to the kinase reaction for 30 min at 30°C in 50 μl of kinase buffer containing 0.5 mM EGTA, 10 mM MgCl₂, 20 mM Na HEPES (pH 7.4), 50 μM ATP, and 1 μCi [γ-³²P]ATP with 2 μg MBP as a substrate. The reaction was terminated by the addition of 5% TCA, and 25 μl of reaction mixture was transferred to p81 phosphocellulose paper. After three rinses with 1% phosphoric acid, radioactivity on the filter disc was determined in a Beckman LS6000IC scintillation counter (Beckman Instruments, Fullerton, CA).

All experiments were repeated for at least three times. Data were analyzed using the Statistical Analysis System (32), with Student’s t test being used to detect differences between treatments.

Although generally recognized as a classical growth factor acting at the G₁ to S transition phase of the cell cycle (33), IGF-I has been reported to promote differentiation of both skeletal and neuronal tissues (34, 35). However, the potential role of IGF-I in macrophage development is unknown. We recently reported that HL-60 cells express easily detectable receptors for IGF-I (24), and these cells are well known to differentiate along the macrophage lineage in the presence of vitamin D₃. As these progenitors develop into more mature myeloid cells, they express the α-subunit of the β₂ integrin CD11b, and this surface marker is used routinely to evaluate their maturation toward the macrophage phenotype (26, 27). Indeed, induction of CD11b is very well correlated with the expression of other myeloid differentiation markers, such as the CD14 LPS receptor (26) and CD67 (36). In these experiments, HL-60 cells were cultured with vitamin D₃ to induce their differentiation along the macrophage lineage, and flow-cytometric histograms depicting these results are shown in Figure 1,A. When HL-60 cells were cultured in medium only, very few expressed the integrin adhesion molecule, CD11b, amounting to only 6 ± 1% when replicated in three independent experiments (Fig. 1,B). In serum-containing medium, vitamin D₃ induced HL-60 cells to differentiate along the macrophage lineage in a dose-dependent fashion (data not shown). At an optimal dose of vitamin D₃ (1 μM), expression of the CD11b differentiation marker increased from 6 ± 1% to 70 ± 5% (Fig. 1,B; p < 0.01; n = 3). HL-60 cells maintained in defined serum-free medium were able to survive for at least 6 days, as determined by >95% of these cells excluding trypan blue. Incubation with vitamin D₃ in this serum-free medium, however, induced only a moderate increase in the proportion of cells expressing CD11b (Fig. 1,B; 5 ± 1% vs 25 ± 3%; p < 0.01). Interestingly, addition of exogenous IGF-I (100 ng/ml) to vitamin D₃-treated cells resulted in a threefold increase in the proportion of cells expressing CD11b (Fig. 1 B; 78 ± 5%; p < 0.01) in serum-free medium, suggesting that IGF-I can totally replace the serum requirement for macrophage differentiation. IGF-I or FBS alone, in the absence of vitamin D₃, failed to promote the differentiation of HL-60 cells (6 ± 1% and 7 ± 1%, respectively). Similarly, we observed that a small percentage of cells treated with vitamin D₃ alone became adherent, irregularly shaped cells with mononuclear, but not polymorphonuclear, nuclei. Cells treated with vitamin D₃ alone also expressed detectable amounts of CD14 and the nonspecific esterase, α-naphthyl acetate esterase, both of which are expressed predominantly on cells of the macrophage/monocytic series (26, 37, 38). However, addition of IGF-I (100 ng/ml) in conjunction with vitamin D₃ markedly increased the proportion of adherent mononuclear cells that expressed these markers (data not shown). These are the first data to show that exogenous IGF-I, a classical growth factor, promotes cellular differentiation along the macrophage lineage.

CSFs such as IL-3 promote differentiation of promyeloid cells by preventing their apoptotic cell death (39). We and others have demonstrated that IGF-I inhibits apoptosis of murine promyeloid cells (9), fibroblasts (40), and PC12 pheochromocytoma cells (8). We therefore asked whether vitamin D₃ induces HL-60 cells to undergo apoptosis in serum-free medium, and whether IGF-I prevents this cell death, thus allowing the completion of vitamin D₃-initiated macrophage differentiation processes. As summarized in Table I, vitamin D₃ alone induced macrophage differentiation in a small proportion of HL-60 cells when compared with those in serum-free medium only (6 ± 2% vs 24 ± 3% CD11b-positive cells; p < 0.01; n = 3). However, vitamin D₃ did not induce apoptotic cell death (6 ± 2% apoptotic cells) compared with that of cells incubated in serum-free medium (6 ± 3% apoptotic cells). Furthermore, IGF-I tripled expression of the CD11b differentiation marker (24 ± 3% vs 73 ± 4%; p < 0.01) in vitamin D₃-treated cells cultured in serum-free medium compared with those incubated with vitamin D₃ alone. IGF-I did not affect the proportion of apoptotic cells in either the absence (5 ± 2%) or presence (5 ± 2%) of vitamin D₃ in serum-free medium, and similar results were observed in the presence of FBS. In contrast, IGF-I promoted expression of the CD11b Ag in RA-treated HL-60 cells cultured in serum-free medium compared with those incubated with RA alone (Table I; 22 ± 2% vs 50 ± 4%; p < 0.01) while rescuing these cells differentiating along the granulocytic pathway from apoptosis (31 ± 3% vs 6 ± 3% apoptotic cells; p < 0.01). Similar to results presented in Figure 1, cells incubated with either differentiating agent in the presence of either IGF-I in serum-free medium or 10% FBS expressed equivalent amounts of CD11b, again suggesting that IGF-I can replace the serum requirement for differentiation along the myeloid lineage. These data establish that IGF-I promotes development of HL-60 cells along both the granulocytic and monocytic lineages, but that IGF-I does not act as a survival factor to promote macrophage differentiation.

IGF-I binds to the IGF-I receptor as well as the insulin receptor, although with a 50- to 100-fold reduction in affinity (41). A murine mAb against the IGF-I receptor, αIR3, specifically recognizes the extracellular α-subunit of the IGF-I receptor (42) and inhibits a number of receptor-mediated activities (24). We therefore asked whether αIR3 would inhibit IGF-I-enhanced macrophage differentiation of vitamin D₃-induced HL-60 cells. Cells were preincubated with αIR3 (5 μg/ml) for 30 min before addition of IGF-I (100 ng/ml). As expected, IGF-I significantly increased the proportion of CD11b-expressing cells treated with vitamin D₃ from 25 ± 4% to 76 ± 4% (Fig. 2; p < 0.01; n = 3). Neutralization of the IGF-I receptor with αIR3 led to a complete inhibition of this IGF-I-promoted cell differentiation in the presence of vitamin D₃ (Fig. 2; 76 ± 4% vs 26 ± 3%; p < 0.01). However, at the same concentration, an irrelevant isotype-matched control Ab (IgG1) had no effect (Fig. 2; 73 ± 2%). These results clearly demonstrate that the IGF-I-promoted enhancement of differentiation is mediated exclusively by the IGF-I receptor.

Normal human serum contains abundant amounts of IGF-I (∼200 ng/ml) and IGF-II (∼700 ng/ml) (43), and similar amounts of IGF-I are contained in FBS (44). IGF-I was identified recently as the relevant serum factor responsible for maintaining the survival of human vascular smooth muscle cells (45). IGF-II also binds and activates the IGF-I receptor in HL-60 cells (24). Since HL-60 cells are conventionally differentiated in FBS-containing medium (as shown in Table I), we tested whether IGF-I in serum is also responsible for promoting macrophage differentiation induced by vitamin D₃. To test this hypothesis, cells were preincubated with αIR3 (5 μg/ml) for 30 min before culturing with vitamin D₃ in 10% FBS-containing medium (Fig. 3). The proportion of CD11b-positive macrophages in vitamin D₃-induced HL-60 cells increased from 24 ± 5% to 70 ± 3% (p < 0.01; n = 3) by addition of 10% FBS. Preincubation with αIR3 in cells cultured with vitamin D₃ in the presence of FBS reduced CD11b expression to 35 ± 3% (p < 0.01), which was statistically similar to that of cells treated with vitamin D₃ alone. At the same concentration, the irrelevant isotype-matched control Ab had no effect. Human rIGFBP-3 also has a high affinity for IGF-I and has been shown to inhibit a number of activities of IGF-I (24, 46). We therefore tested the possibility that IGFBP-3 would inhibit vitamin D₃-induced macrophage differentiation in serum-containing medium. Similar to the results obtained with αIR3, addition of IGFBP-3 (250 ng/ml) reduced CD11b expression to 34 ± 4% (p < 0.01) in vitamin D₃-stimulated HL-60 cells (Fig. 3), amounting to a 78% inhibition. At the same concentration (250 ng/ml), an irrelevant serum protein (BSA) had no effect, excluding the possibility of a nonspecific protein effect of IGFBP-3. We interpret these data to indicate that both an Ab to the receptor and a protein that specifically binds the ligand are capable of inhibiting the unbound form of IGF-I in serum, and therefore suppress the development of HL-60 cells into macrophages.

PI 3-kinase has been demonstrated to play a key role in normal B lymphocyte differentiation (13). Indeed, inhibition of this lipid kinase completely blocks the CD40 ligand-induced Ig production. We therefore asked whether PI 3-kinase is activated during IGF-I-promoted macrophage differentiation in promyeloid cells induced by vitamin D₃. We found that vitamin D₃ alone was unable to activate anti-PY-immunoprecipitable PI 3-kinase activity, whereas addition of IGF-I resulted in a sevenfold increase in kinase activity (Fig. 4 B; p < 0.01; n = 3). Indeed, the IGF-I-induced increase in PI 3-kinase activity was nearly the same in either the presence or the absence of vitamin D₃ (7- ± 1- vs 8- ± 2-fold, respectively; p > 0.10).

To determine whether PI 3-kinase plays a central role in IGF-I-enhanced macrophage differentiation, two different PI 3-kinase inhibitors (wortmannin and LY294002) were used to block this phospholipid-mediated signal-transduction pathway (47). Wortmannin and LY294002 differ in their mechanisms of inhibition of PI 3-kinase. Whereas wortmannin inhibits PI 3-kinase by irreversibly binding the p110 catalytic subunit (48), LY294002 is a competitive antagonist for the ATP binding site of this kinase (49). Cells were preincubated with an optimal concentration of wortmannin (1 μM) or LY294002 (25 μM), both of which inhibited by greater than 95% (Fig. 4 B; p < 0.01; n = 3) the IGF-I-induced enhancement of phosphatidylinositol phosphorylation in cells treated with vitamin D₃.

Members of PKC family, in particular aPKC-ζ, have recently been demonstrated to mediate signals from growth factors during cell growth and differentiation, and have also been suggested to be one of the downstream signaling targets of PI 3-kinase (6). Of particular relevance is the finding that overexpression of PKC-ζ in human promonocytic U937 cells induces their differentiation along the monocyte pathway (3). We therefore tested the possible involvement of endogenous PKC-ζ in IGF-I signaling during macrophage differentiation. HL-60 cells were incubated in serum-free medium with vitamin D₃ (1 μM) in the presence or absence of IGF-I (100 ng/ml) for 5 min. Cell lysates were then immunoprecipitated with an anti-PKC-ζ Ab, and the resulting kinase activity was measured in the immunocomplex. While vitamin D₃ alone failed to activate PKC-ζ, addition of IGF-I to vitamin D₃-treated cells induced a sixfold increase in protein phosphorylation in the anti-PKC-ζ-precipitated complex (Fig. 5; p < 0.05; n = 3). Indeed, the IGF-I-elicited increase in the PKC-ζ activity was not statistically different in either the presence (6 ± 1) or absence (10 ± 3) of vitamin D₃. More importantly, inhibition of PI 3-kinase was directly related to the diminished PKC-ζ activity (Fig. 5) because treatment with either wortmannin (1 μM) or LY294002 (25 μM) led to a 95 ± 3% and 82 ± 2% inhibition, respectively, of PKC-ζ activity (p < 0.05). These results established that PKC-ζ is activated during the development of HL-60 cells into macrophages, and that IGF-I rather than vitamin D₃ is responsible for induction of this enzyme. The IGF-I-induced activation of PKC-ζ is likely to be a downstream target of PI 3-kinase.

Since IGF-I induces significant PKC-ζ activity, and this activation is blocked by two different inhibitors of PI 3-kinase, we tested the important possibility that these PI 3-kinase inhibitors might also suppress IGF-I-enhanced macrophage differentiation. As expected, vitamin D₃ induced CD11b expression in only a small proportion of cells (Fig. 6; 26 ± 4%; n = 3), and addition of IGF-I led to a threefold increase in expression of this leukocyte differentiation marker (79 ± 4%; p < 0.01). Both PI 3-kinase inhibitors, LY294002 (25 μM) and wortmannin (1 μM), suppressed IGF-I-enhanced CD11b expression (Fig. 6; 26 ± 2% and 26 ± 3%, respectively; p < 0.01) to levels equivalent to cells treated with only vitamin D₃. This is consistent with the finding that vitamin D₃ alone induces the activity of neither PI 3-kinase (Fig. 4) nor PKC-ζ (Fig. 5). In the absence of IGF-I, LY294002 and wortmannin did not affect CD11b expression in cells treated with vitamin D₃ alone (Fig. 6; 22 ± 2% and 21 ± 2%, respectively), indicating that the two PI 3-kinase inhibitors were specific for IGF-I. These results, together with the earlier experiments measuring PKC-ζ and PI 3-kinase activation, strongly suggest that IGF-I-enhanced macrophage differentiation occurs concomitantly with activation of PI 3-kinase and its putative downstream PKC-ζ pathway.

These data establish that IGF-I induces the activation of PKC-ζ by a mechanism that involves PI 3-kinase, and that IGF-I plays a critical role in the development of macrophages from their progenitor cells. Immature myeloid progenitors express very little CD11b, which is a component of the heterodimeric CR3 complement receptor, and in this work we demonstrate that addition of IGF-I to vitamin D₃-treated progenitors increases the expression of CD11b by threefold (Fig. 1). Indeed, this enhancement of differentiation by exogenous IGF-I is mediated via its own receptor rather than the potential involvement of the insulin receptor (Fig. 2) since a mAb directed against the IGF-I receptor completely abrogates IGF-I-potentiated macrophage differentiation. These data further suggest that IGF-I is a critical peptide in serum that promotes macrophage maturation because both IGFBP-3 and a mAb against the IGF-I receptor profoundly inhibit vitamin D₃-induced CD11b expression in serum-containing conditions (Fig. 3). More importantly, IGF-I-enhanced macrophage differentiation occurs concomitantly with the activation of PI 3-kinase and its putative downstream target PKC-ζ (Figs. 4 and 5). Inhibition of the activity of PI 3-kinase blocks the activation of both PKC-ζ (Fig. 5) and the IGF-I-promoted enhancement of macrophage differentiation (Fig. 6). Taken together, the present results argue for a novel role for IGF-I during the development of macrophages by a mechanism that induces the activation of both PI 3-kinase and PKC-ζ.

The PKC-ζ isoform has been demonstrated recently to mediate cellular differentiation in adipocytes (4) and neuronal cells (50). Although overexpression of PKC-ζ stimulates the phenotypic expression of monocytic maturation markers (3), the possibility that this endogenous Ser/Thr kinase is activated in clonal progenitor cells induced by vitamin D₃ has not been addressed. Our experiments demonstrate that IGF-I not only promotes macrophage differentiation, but also stimulates PKC-ζ, and that both of these events are dependent upon the activation of a wortmannin- and LY294002-sensitive pathway, putatively PI 3-kinase. These findings are of importance because the stimulation of PI 3-kinase by IGF-I is required to prevent apoptosis of promyeloid progenitor cells (9) and PC 12 pheochromocytoma cells (8). Recent evidence has demonstrated that protein kinase B (c-Akt) is a downstream signaling target of PI 3-kinase since suppression of PI 3-kinase activity leads to inhibition of this enzyme in vivo (12, 51). Although certain members of both nPKC and aPKC, such as PKC-ζ, are likely to be direct downstream targets of PI 3-kinase-triggered signals (6), the possibility of a potential linkage between PI 3-kinase and PKC-ζ in IGF-I-promoted macrophage differentiation has not yet been tested. Our experiments demonstrate that IGF-I enhances macrophage differentiation as well as activation of both PI 3-kinase and PKC-ζ, and that inhibition of PI 3-kinase results in suppression of both PKC-ζ and expression of the differentiation marker CD11b. Interestingly, the jun/fos transcriptional factor acts synergistically with the vitamin D receptor to bind AP-1 DNA binding sites. Overexpression of PKC-ζ leads to the induction of this transcription factor (3), suggesting a direct interaction between this Ser/Thr kinase and the AP-1 heterodimer. Although members of the mitogen-activated protein (MAP) kinase family have been reported to transduce extracellular signals during differentiation in other cell types, including immature thymocytes (52), adipocytes (53), and myoblasts (54), preliminary data do not appear to support a major role for Erk1 and Erk2 in IGF-I-enhanced macrophage differentiation (data not shown). This conclusion is supported by a recent report that identified PKC-ζ as a novel cell growth suppressor that acts in v-raf-transformed NIH-3T3 cells via a Raf/MEK/ MAP kinase-independent mechanism (55).

Although these data clearly establish that IGF-I activates both PI 3-kinase and PKC-ζ, our experiments do not exclude the possibility that other PI 3-kinase family members may be involved in the signaling events that lead to enhanced macrophage differentiation. It is now recognized that there is a growing family of PI 3-kinase-related proteins, currently consisting of at least nine members (reviewed in 56 . All of the carboxyl-terminal regions of these proteins share homology with those found in the catalytic domain of the classical PI 3-kinases. The enzymatic activity of two mammalian homologues of this PI 3-kinase family, mTOR (57) and DNA-PKcs (58), is inhibited in vitro by both wortmannin and LY294002 at concentrations similar to those used in our in vivo experiments. However, neither mTOR nor DNA-PKcs are currently known to phosphorylate phosphatidylinositol, whereas IGF-I clearly induces the activity of PI 3-kinase. At the concentrations used in our experiments, neither wortmannin nor LY294002 directly inhibits the enzymatic activity of other non-PI 3-kinase-related signaling molecules that are ubiquitously expressed, such as Raf, MEK, PKC, protein kinase A, or Src kinase (49, 59). Although myosin light chain kinase can be inhibited by LY294002 at concentrations higher than 100 μM (40% inhibition), it is not blocked at the concentration (25 μM) used in our study (60). These data are consistent with the idea that one or more members of the PI 3-kinase family are responsible for the activation of PKC-ζ in IGF-I-treated myeloid progenitor cells.

Although IGF-I has been viewed as a progression factor that is required for promoting cells to advance through the cell cycle (33, 61), our findings that this peptide also enhances vitamin D₃-induced development of macrophages are consistent with those of IGF-I augmenting the differentiation of fat (62), muscle (34), and nerve cells (35). By using a defined serum-free system, we demonstrate that IGF-I increases the expression of a surface differentiation marker, CD11b, as promyeloid progenitors develop into mature macrophages in the presence of vitamin D₃. IGF-I acts to promote rather than to initiate this process because it is ineffective in the absence of vitamin D₃. These results support accumulating evidence that this classical growth factor significantly promotes cell differentiation. For example, both IGF-I and IGF-II, acting through the IGF-I receptor, stimulate myogenesis in the absence of other inducers of differentiation (63, 64). Indeed, autocrine secretion of IGF-II causes spontaneous differentiation (65), whereas expression of antisense IGF RNA blocks skeletal muscle differentiation in vitro (66). Recent data show that IGF-I also promotes initial progression through the cell cycle and subsequently induces differentiation in myoblast cells (34), suggesting a biphasic effect of IGF-I in promoting clonal expansion, followed by cellular differentiation.

Similar to the development of adipocytes and neuroblastoma cells, both lymphocytes and myeloid cells require two signals for differentiation. For example, IGF-I alone does not induce the differentiation of human SH-SY5Y neuroblastoma cells, but it potently promotes their differentiation in the presence of PMA as an induction factor (35). More significantly, IGF-I alone is able to replace the serum requirement for differentiation of 3T3-L1 preadipocytes induced by dexamethasone and 1-methyl-3-isobutylxanthine (62), suggesting IGF-I acts as a critical factor to support fat cell development. In the absence of these inducers of differentiation, however, IGF-I has no effect. During hemopoiesis, stromal cell-derived IGF-I induces the formation of pre-B cells from murine bone marrow progenitor cells, as assessed by expression of cytoplasmic μ heavy chains (22). This B cell differentiation process is abrogated by either an anti-IGF-I Ab or pretreatment of stromal cells with an antisense oligonucleotide to IGF-I mRNA. However, IGF-I does not induce pre-B cell development in the absence of IL-7 (23), which is similar to our findings that IGF-I does not induce macrophage differentiation in the absence of vitamin D₃. Instead, IGF-I, as well as c-kit ligand, potentiates IL-7-induced B cell maturation. Similarly, during myeloid cell development from primary human bone marrow cells, IGF-I also promotes the formation of granulocytic colonies induced by granulocyte CSF, granulocyte-macrophage CSF, or IL-3, but as in our experiments, has little effect in the absence of these initiation signals (21). Therefore, it appears that at least two signals are needed for the optimal development of myeloid progenitor cells into macrophages: 1) a specific inducer such as vitamin D₃ is required for initiation of the differentiation process, and 2) IGF-I activates both PI 3-kinase and PKC-ζ and enhances, but does not induce, macrophage differentiation. This is similar to the classical requirements for macrophage activation in which both priming (e.g., IFN-γ) and inducing (e.g., LPS) signals are clearly defined (67).

The mechanism by which IGF-I synergizes with vitamin D₃ to promote the differentiation of hemopoietic cells remains unknown. Indeed, since vitamin D₃ has only been used to induce hemopoietic precursor cells to differentiate toward the monocyte/macrophage lineage in serum-containing medium (26, 27), it has been difficult to separate the role of serum from that of vitamin D₃ in this process. Our data clearly show that the use of a defined serum-free system for myeloid cell differentiation is necessary to separate the signals derived from vitamin D₃ and IGF-I. During macrophage differentiation in vitamin D₃-stimulated HL-60 cells in serum, expression of both c-jun and c-fos has been identified as a lineage-specific marker of this process (68). The steroid/nuclear vitamin D receptor acts synergistically with a number of nonreceptor transcriptional factors, such as jun/fos, at the level of cooperative DNA binding at the AP-1 site (69). Since IGF-I activates the AP-1 transcriptional factor (70), which is also regulated by PKC-ζ (3, 55), perhaps a mechanism by which IGF-I synergizes with vitamin D₃ to promote macrophage differentiation is at the level of activation of the AP-1 transcriptional factor.

In conclusion, we have established a new and important role for IGF-I in promoting vitamin D₃-induced macrophage differentiation. By using a defined serum-free differentiation system, we demonstrate that although vitamin D₃ initiates the differentiation program, IGF-I, acting through its own receptor, dramatically promotes this process, as indicated by a threefold increase in expression of the CD11b mature leukocyte Ag. More importantly, IGF-I enhances macrophage differentiation via a pathway that includes the concomitant activation of both PI 3-kinase and PKC-ζ. Inhibition of IGF-I-inducible PI 3-kinase abrogates PKC-ζ activity and suppresses vitamin D₃-induced macrophage differentiation. Taken together, these results suggest a novel role for the classical growth factor IGF-I in promoting macrophage differentiation and activating PKC-ζ, both of which occur concomitantly with the activation of PI 3-kinase.