cAMP inhibits CSF-1-stimulated tyrosine phosphorylation but augments CSF-1R-mediated macrophage differentiation and ERK activation

Authors

  • Nicholas J. Wilson,

    1. Arthritis and Inflammation Research Centre, Department of Medicine (RMH/WH), University of Melbourne, Royal Melbourne Hospital, Parkville, Victoria, Australia
    Search for more papers by this author
    • Present address
      DNAX Research Institute, 901 California Ave, Palo Alto, CA, USA

  • Maddalena Cross,

    1. CRC for Chronic Inflammatory Diseases, Department of Medicine (RMH/WH), University of Melbourne, Royal Melbourne Hospital, Parkville, Victoria, Australia
    Search for more papers by this author
  • Thao Nguyen,

    1. CRC for Chronic Inflammatory Diseases, Department of Medicine (RMH/WH), University of Melbourne, Royal Melbourne Hospital, Parkville, Victoria, Australia
    Search for more papers by this author
  • John A. Hamilton

    1. Arthritis and Inflammation Research Centre, Department of Medicine (RMH/WH), University of Melbourne, Royal Melbourne Hospital, Parkville, Victoria, Australia
    2. Department of Medicine (RMH/WH), University of Melbourne, Royal Melbourne Hospital, Parkville, Victoria, Australia
    3. CRC for Chronic Inflammatory Diseases, Department of Medicine (RMH/WH), University of Melbourne, Royal Melbourne Hospital, Parkville, Victoria, Australia
    Search for more papers by this author

N. J. Wilson, DNAX Research Institute, 901 California Ave., Palo Alto, CA, 94304–1104, USA
Fax: +1 650 496 1200
Tel: +1 650 496 1223
E-mail: nick.wilson@dnax.org

Abstract

Macrophage colony stimulating factor (M-CSF) or CSF-1 controls the development of the macrophage lineage through its receptor tyrosine kinase, c-Fms. cAMP has been shown to influence proliferation and differentiation in many cell types, including macrophages. In addition, modulation of cellular ERK activity often occurs when cAMP levels are raised. We have shown previously that agents that increase cellular cAMP inhibited CSF-1-dependent proliferation in murine bone marrow-derived macrophages (BMM) which was associated with an enhanced extracellular signal-regulated kinase (ERK) activity. We report here that increasing cAMP levels, by addition of either 8-bromo cAMP (8BrcAMP) or prostaglandin E1 (PGE1), can induce macrophage differentiation in M1 myeloid cells engineered to express the CSF-1 receptor (M1/WT cells) and can potentiate CSF-1-induced differentiation in the same cells. The enhanced CSF-1-dependent differentiation induced by raising cAMP levels correlated with enhanced ERK activity. Thus, elevated cAMP can promote either CSF-1-induced differentiation or inhibit CSF-1-induced proliferation depending on the cellular context. The mitogen-activated protein kinase/extracellular signal-related protein kinase kinase (MEK) inhibitor, PD98059, inhibited both the cAMP- and the CSF-1R-dependent macrophage differentiation of M1/WT cells suggesting that ERK activity might be important for differentiation in the M1/WT cells. Surprisingly, addition of 8BrcAMP or PGE1 to either CSF-1-treated M1/WT or BMM cells suppressed the CSF-1R-dependent tyrosine phosphorylation of cellular substrates, including that of the CSF-1R itself. It appears that there are at least two CSF-1-dependent pathway(s), one MEK/ERK dependent pathway and another controlling the bulk of the tyrosine phosphorylation, and that cAMP can modulate signalling through both of these pathways.

Abbreviations
8BrcAMP

8-bromo cAMP

BMM

bone marrow-derived macrophages

CSF

colony stimulating factor

CSF-1R

CSF-1 receptor

EGF

epidermal growth factor

ERK

extracellular signal-regulated kinase

FITC

fluorescein isothiocyanate

M-CSF

macrophage colony stimulating factor

MEK

mitogen-activated protein kinase/extracellular signal-related protein kinase kinase

NGF

nerve growth factor

PGE1

prostaglandin E1

PKA

protein kinase A

RTK

receptor tyrosine kinase

WT

wild-type

A key cytokine controlling macrophage lineage development from bone marrow precursors by proliferation and differentiation is macrophage-colony stimulating factor (M-CSF or CSF-1) [1]. The CSF-1 receptor (CSF-1R) is the homodimeric receptor tyrosine kinase (RTK), c-Fms [2]. The survival-promoting and proliferative actions of CSF-1 have been widely studied. Evidence supporting a possible differentiation function for CSF-1 include the observations that a significant response to CSF-1 of human bone marrow cells in vitro is differentiation into macrophages [3] and that mutations in c-fms have been associated with acute myeloid leukaemia and myelodysplastic syndromes [4]. A new model for CSF-1-induced macrophage differentiation has recently been developed by transfecting CSF-1R into the immature M1 myeloid cell line [5]; addition of CSF-1 to these cells led to a rapid appearance of macrophage-like cells.

Agents which raise intracellular cAMP have often been shown to modulate cell growth and/or differentiation most likely via protein kinase A (PKA) activation. We, and others, have shown previously that the proliferative response to CSF-1 of murine bone marrow-derived macrophages (BMM) and a subpopulation of human monocytes is dramatically suppressed by raising intracellular cAMP [6–8]. Our studies also indicated that early biochemical responses of BMM to CSF-1, namely protein synthesis, Na+/H+ exchange activity, Na+/K+ ATPase activity and c-myc mRNA expression, were not inhibited [9]; however, the CSF-1-induced mRNA expression of cyclin D1 [10], of three genes whose products are associated with the DNA synthesis machinery (the M1 and M2 subunits of ribonucleotide reductase and proliferating cell nuclear antigen) and of c-myc at later times following CSF-1 addition were reduced by cAMP elevation [9]. In a number of cell systems inhibition of extracellular signal-regulated kinase (ERK) activity by increased intracellular cAMP has often been correlated with suppression of growth factor-induced proliferation. However, in BMM we have reported that 8-bromo cAMP (8BrcAMP), despite being a dramatic G1 phase proliferation inhibitor, increased ERK activity both in the absence and presence of CSF-1 in a mitogen-activated protein kinase/extracellular signal-regulated protein kinase kinase (MEK)-dependent manner [11]. It was also found that an acute but not a sustained elevation of c-fos mRNA expression due to 8BrcAMP was also MEK dependent [11].

As human monocytes differentiate in vitro, increases in intracellular cAMP levels occur [12]; in addition, increasing the cAMP levels in the human myeloid cell lines, U937 and HL60, promoted their differentiation into macrophage-like cells [13–15]. Raising intracellular cAMP in other cellular systems can also regulate their differentiation in response to various stimuli] for example, cAMP can enhance osteoclast differentiation by receptor activator of NF-κB ligand (RANKL) [16] and neuronal differentiation by nerve growth factor (NGF) or epidermal growth factor (EGF)[17]. Although the signaling mechanism(s) underlying the differentiating promoting effect of cAMP remains unclear there are reports that for cAMP-induced neurite outgrowth from PC12 cells, ERK activation is greatly enhanced [18], while other reports suggest that activation of the EGF receptor (EGFR) can occur [17]. Others have shown that the EGF-induced tyrosine phosphorylation and kinase activity of the EGFR can be inhibited by raising intracellular cAMP which was mediated by direct phosphorylation by PKA of a particular serine residue of the EGFR [19]. These studies suggest that cAMP can modulate RTK activity and affect downstream signal transduction.

Given the above background on cAMP-dependent biology, we explored in this study the effects of enhanced cAMP levels on CSF-1-induced macrophage differentiation in M1/WT cells, i.e. M1 cells expressing the normal or ‘wild-type’ CSF-1R [5]. We report that agents that increase intracellular cAMP, namely 8BrcAMP and prostaglandin E1 (PGE1), potentiate the CSF-1-induced M1/WT differentiation. We also show that increasing intracellular cAMP in the absence of CSF-1 can induce differentiation of M1/WT cells but not in M1 cells lacking a functional CSF-1R. While not leading to tyrosine phosphorylation of the CSF-1R or to significantly increased CSF-1R degradation, 8BrcAMP and PGE1 dramatically reduced both basal and CSF-1R-dependent tyrosine phosphorylation, including that of the CSF-1R itself. In spite of this suppressed CSF-1-dependent tyrosine phosphorylation an increase in CSF-1-dependent ERK activity was noted in the presence of these agents. Similar molecular changes were made in CSF-1-treated BMM where elevated cAMP suppresses proliferation. Thus CSF-1-stimulated ERK activity is another early CSF-1-dependent biochemical response which is not suppressed by increasing intracellular cAMP concentration; it would appear that these particular responses are independent of a pathway(s) involving most of the CSF-1R-dependent tyrosine phosphorylation.

Results

cAMP augments CSF-1-induced differentiation of M1/WT cells

We have previously established a model of CSF-1-induced macrophage differentiation by transfecting the CSF-1R into a population of M1 myeloid leukemic cells lacking the CSF-1R [5]; the resultant cell population is referred to as M1/WT cells to indicate that they express the normal or wild-type receptor and to distinguish them from populations expressing mutated CSF-1R [5]. We took the precaution of removing CSF-1R positive cells from the starting population to avoid any confounding influence from the endogenous receptor. As shown before [5] and again in Fig. 1A, panel ii, the M1/WT cells rapidly (within 24-72 h) differentiate into macrophage-like cells upon treatment with CSF-1, with the cells becoming more irregular in appearance and adhering to the tissue culture surface; cells lacking the transfected receptor (M1/parental) do not respond to CSF-1 [5]. We show in Fig. 1A, panel iv, that 8BrcAMP, a stable analogue of cAMP, can enhance the macrophage-like morphological changes induced in M1/WT cells by CSF-1; more cells become much larger, acquire long projections, have a more granular appearance and adhere strongly to the tissue culture surface. Similar results (data not shown) were obtained by treating CSF-1-stimulated M1/WT cells with PGE1 (1 µm) which raises intracellular cAMP by activating adenylate cyclase. Treatment of M1/WT cells with 8BrcAMP (Fig. 1A, panel iii) or PGE1 (data not shown) alone had a slight morphologic effect at 24 h.

Figure 1.

8BrcAMP augments CSF-1-induced differentiation in M1/WT cells. (A) M1/WT cells were either (i) untreated, (ii) treated with CSF-1 (5000 U·mL−1), (iii) 8BrcAMP (1 mm), or (iv) a combination of CSF-1 (5000 U·mL−1) and 8BrcAMP (1 mm) for 24 h. Cell morphology was examined by light microscopy (20× magnification). This experiment was repeated six times with similar results; a representative experiment is shown. (B) M1/WT and M1/parental cells were either untreated (dashed lines) or treated with CSF-1 (5000 U·mL−1) (dark lines), or 8BrcAMP (1 mm) (light lines) for 72 h; cells were incubated with anti-Mac-1 IgG and FITC-conjugated anti-IgG2b secondary antibody and the median fluorescence intensity of Mac-1 staining was determined by flow cytometry. This experiment was repeated four times with similar results; representative FACS plots are shown. (C) M1/WT cells were either untreated or treated with CSF-1 (5000 U·mL−1), 8BrcAMP (1 mm) or a combination of CSF-1 (5000 U·mL−1) and 8BrcAMP (1 mm) for 72 h; cells were incubated with anti-Mac-1 IgG and FITC-conjugated anti-IgG2b secondary antibody and the percentage of Mac-1 positive staining cells was determined by flow cytometry. This experiment was repeated four times with similar results; means and standard deviation are shown.

We next examined the influence of elevated cAMP on expression of the integrin, Mac-1, another marker of CSF-1-induced macrophage differentiation in M1/WT cells [5,20]. It can be seen in Fig. 1B that 8BrcAMP can up-regulate Mac-1 expression in M1/WT cells to a similar level to that seen for CSF-1, while neither agent could up-regulate its expression in M1/parental cells. Figure 1C shows quantitative data on the increase in Mac-1 expression due to both 8BrcAMP and CSF-1 in M1/WT cells and demonstrates that 8BrcAMP can augment the enhanced expression induced by CSF-1. The effect of 8BrcAMP alone on M1/WT cell differentiation but not on that in M1/parental cells suggests that cAMP may interact in some way with CSF-1R-associated signaling to induce M1 cell differentiation (see below). It should be noted that the actual percentage of Mac-1 positive cells induced by CSF-1 alone was not increased by coaddition of 8BrcAMP (data not shown).

cAMP requires a functional CSF-1R to induce differentiation of M1 cells

In order to determine whether a fully functional CSF-1R was required for the action of 8BrcAMP on macrophage-like differentiation of M1 cells, we tested 8BrcAMP and CSF-1 addition on M1/parental cells and on M1/807 cells, i.e. M1 cells which express CSF-1R with a point mutation at tyrosine 807 thereby reducing CSF-1R function. The mutation at tyrosine 807 behaves like a kinase dead mutation in that upon activation there is minimal tyrosine phosphorylation of either the CSF-1R or downstream substrates, including overall tyrosine phosphorylation and that of specific substrates such as Shc and p42/44 MAPK [20]. Figure 2A shows that this mutation behaves in a similar manner in M1/807 cells as there is minimal tyrosine phosphorylation of the CSF-1R after CSF-1-stimulation compared with M1/WT cells. We and others have previously reported that myeloid cells expressing the tyrosine 807 mutation do not differentiate in response to CSF-1 [20,21]. Using these cells and the percentage of Mac-1 positive cells as a readout for differentiation it can be seen in Fig. 2B that 8BrcAMP induces differentiation to similar levels to that of CSF-1 in M1/WT cells but neither CSF-1 nor 8BrcAMP induce differentiation in M1/parental or M1/807 cells (Fig. 2B); the coaddition of CSF-1 and 8BrcAMP does not result in differentiation of either M1/parental or M1/807 cells (data not shown). Similar results were obtained when PGE1 (1 µm) was used to raise intracellular cAMP and when Mac-1 expression was monitored (data not shown).

Figure 2.

8BrcAMP-mediated M1 cell differentiation is dependent on a functional CSF-1R. (A) M1/WT cells and M1/807 cells were either untreated or treated with CSF-1 (5000 U·mL−1) or 8BrcAMP (1 mm) for the times indicated. CSF-1R was immunoprecipitated from protein lysates with an anti-CSF-1R IgG, proteins separated by 1D SDS/PAGE, western blotted and probed with an anti-phosphotyrosine IgG. (B) M1/parental, M1/WT and M1/807 cells were either untreated or treated with CSF-1 (5000 U·mL−1) or 8BrcAMP (1 mm) or for 72 h; cells were incubated with anti-Mac-1 IgG and FITC-conjugated anti-IgG2b secondary antibody and the percentage of Mac-1 positive staining cells was determined by flow cytometry. This experiment was repeated four times with similar results; means and standard deviation are shown. (C) M1/WT cells were either untreated or treated with CSF-1 (5000 U·mL−1) or 8BrcAMP (1 mm) in the presence (black bars) or absence (unfilled bars) of blocking CSF-1R antibody (AFS-98, 10 µg·mL−1) for 72 h; cells were incubated with anti-Mac-1 IgG and FITC-conjugated anti-IgG2b secondary antibody and the percentage of Mac-1 positive staining cells was determined by flow cytometry. This experiment was repeated four times with similar results; means and standard deviations are shown. (D) M1/WT cells were either untreated or treated with CSF-1 (5000 U·mL−1), 8BrcAMP (1 mm) or 8BrcAMP (1 mm) for 30 min prior to CSF-1 (5000 U·mL−1) for the times indicated. CSF-1R was immunoprecipitated from protein lysates with an anti-CSF-1R IgG, proteins separated by 1D SDS/PAGE, western blotted and probed with an anti-phosphotyrosine IgG (pCSF-1R) or with an anti-CSF-1R IgG (CSF-1R). A representative blot of three independent experiments is shown.

These data indicate that differentiation of M1 cells by increased cAMP concentrations is dependent upon a functional CSF-1R. To assess the possibility of autocrine production of CSF-1 leading to M1/WT cell differentiation in response to 8BrcAMP we used a blocking antibody against the CSF-1R. The blocking antibody effectively inhibited M1/WT cell differentiation induced by CSF-1, even at the high concentration used (5000 U·mL−1), but had no effect on 8BrcAMP induced differentiation (Fig. 2C). These results suggest that cAMP does not induce M1/WT cell differentiation via autocrine production of CSF-1, but that cAMP can somehow modulate CSF-1R-dependent signalling even in the absence of CSF-1.

cAMP inhibits CSF-1-induced tyrosine phosphorylation of the CSF-1R and its downstream substrates

Following CSF-1 addition, the CSF-1R-induced tyrosine phosphorylation of cellular substrates, including the CSF-1R itself, is presumed to form part of the relevant signal transduction cascades modulating cellular responses [20–24]. It might therefore be expected that cAMP elevation could enhance CSF-1R-induced tyrosine phosphorylation in M1/WT cells and maybe induce tyrosine phosphorylation of the CSF-1R itself as part of the differentiation induction program described above. Figure 2A demonstrates that, as expected, CSF-1 induces a rapid and transient tyrosine phosphorylation of the CSF-1R; however, addition of 8BrcAMP does not lead to tyrosine phosphorylation of the CSF-1R. To ensure that 8BrcAMP treatment did not simply induce tyrosine phosphorylation of the CSF-1R with different kinetics to CSF-1 we extended the time course of CSF-1 and 8BrcAMP treatment. Figure 2D shows that CSF-1-induced tyrosine phosphorylation of the CSF-1R is still maintained at 10 min but has returned to basal by 2 h post stimulation. Again we did not observe tyrosine phosphorylation of the CSF-1R induced by 8BrcAMP treatment at any time point. Surprisingly, given that the addition of 8BrcAMP and CSF-1 together results in more pronounced differentiation of M1/WT cells, we found that 8BrcAMP treatment could suppress the rapid tyrosine phosphorylation of the CSF-1R induced by CSF-1 (Fig. 2D). The effect of 8BrcAMP was still maintained even up to 2 h after CSF-1 addition. The effects on CSF-1R tyrosine phosphorylation are unlikely due to reduced CSF-1R levels because it can also be seen in Fig. 2D that the reduction in total CSF-1R levels following CSF-1 treatment, due to internalization and degradation [25–27], is only slightly more pronounced in the presence of elevated intracellular cAMP. Importantly treatment with 8BrcAMP alone for up to 30 mins, i.e. the pretreatment time before CSF-1 addition, does not reduce the levels of the CSF-1R in the M1/WT cells.

We extended these findings by examining the effects of cAMP on CSF-1-induced tyrosine phosphorylation of cellular substrates. Figure 3A confirms the effect of 8BrcAMP addition on the transient CSF-1-induced CSF-1R tyrosine phosphorylation at the 4 min time point of following CSF-1 addition, which is approximately the optimal time point for this phosphorylation and for downstream substrates [20]. It can be seen in Fig. 3D, as indicated by probing western blots of whole cell lysates with antibodies raised against phosphotyrosine, that CSF-1 treatment of M1/WT cells leads to a rapid and transient tyrosine phosphorylation of many cellular proteins. In contrast, 8BrcAMP again does not induce tyrosine phosphorylation in M1/WT cells and appears to suppress the basal tyrosine phosphorylation in these cells; 8BrcAMP pretreatment of M1/WT cells prior to CSF-1-stimulation also significantly suppressed CSF-1-mediated tyrosine phosphorylation.

Figure 3.

cAMP inhibits CSF-1-induced tyrosine phosphorylation but augments CSF-1-induced ERK activation. (A) M1/WT cells were either untreated or treated with CSF-1 (5000 U·mL−1) or 8BrcAMP (1 mm) for the times indicated or treated with 8BrcAMP (1 mm) for 30 min prior to CSF-1 stimulation for 4 min. CSF-1R was immunoprecipitated from lysates with an anti-CSF-1R IgG, proteins were separated by 1D SDS/PAGE, western blotted and probed with an anti-phosphotyrosine IgG (pCSF-1R) (upper panel) or anti-CSF-1R IgGs (middle panel). The IgG heavy chain band is shown as a loading control (lower panel). (B) M1/WT cells were either untreated or treated with CSF-1 (5000 U·mL−1) or PGE1 (10 µm) for the times indicated or treated with PGE1 (10 µm) for 30 min prior to CSF-1 stimulation for 4 min. Protein lysates were made and treated as in (A). (C) BMM cells and lysates were treated exactly as described in (A). (D) M1/WT cells were treated as in (A), whole cell lysates were separated by 1D SDS/PAGE, western blotted and probed with either anti-phosphotyrosine IgGs (α-PY) (upper panel) or anti-phospho-ERK IgGs (pERK) (middle panel). Total ERK is shown as a loading control (lower panel). (E) M1/WT cells were either untreated or treated with CSF-1 (5000 U·mL−1) or PGE1 (10 µm) for the times indicated or treated with PGE1 (10 µm) for 30 min prior to CSF-1 stimulation for 4 min. Protein lysates were made and treated as in (D). (F) BMM cells and lysates were treated exactly as described in (D). All experiments were repeated four times, with representative blots shown.

As for the effects on differentiation, similar results were found for both CSF-1R and overall tyrosine phosphorylation after addition of the more physiological stimulus, PGE1 (Fig. 3B,E, respectively).

As referred to above, CSF-1-induced proliferation of BMM is inhibited by cAMP elevation [6,9]. It can be observed that CSF-1-induced tyrosine phosphorylation of cellular proteins and, more specifically, the CSF-1R is also reduced by 8BrcAMP in BMM (Fig. 3F,C, respectively) and is therefore not restricted to M1/WT cells. The effects on CSF-1R tyrosine phosphorylation are again not due to reduced CSF-1R levels.

cAMP enhances CSF-1-induced ERK activation

The above data suggest that CSF-1-induced tyrosine phosphorylation is not critical for the CSF-1-induced macrophage differentiation in M1/WT cells as there is an inverse correlation between the effects of raised cAMP on these two parameters. We recently showed that 8BrcAMP enhances CSF-1-induced ERK activity in proliferating BMM via a MEK-dependent pathway [11]. We therefore decided to explore whether this might also be the case for a cell system which differentiates in response to CSF-1. In order to do this we probed western blots of lysates treated similarly to those shown in Fig. 3D-F with anti-phospho-ERK IgG. Figure 3D,E (lower panels) show that CSF-1 addition results in rapid ERK activation in M1/WT cells; the analogous BMM data can be seen in Fig. 3F. Pretreatment of both cell types with 8BrcAMP and M1/WT cells with PGE1 enhanced the CSF-1-stimulated ERK activity with some slight stimulation even in the absence of CSF-1. Although we found elevated cAMP led to an increase in CSF-1-induced ERK activation we did not observe altered kinetics of CSF-1-induced ERK activation by pretreatment of BMM with 8BrcAMP as an explanation for the increased activation (Fig. 4A). The increase in CSF-1-induced ERK activation by cAMP is perhaps surprising given the suppression of tyrosine phosphorylation noted.

Figure 4.

ERK activation is required for M1/WT cell differentiation in response to CSF-1 or 8BrcAMP. (A) M1/WT cells were treated with CSF-1 (5000 U·mL−1) for 4, 10 and 30 min or with 8BrcAMP (1 mm) for 30 min prior to the same CSF-1 time course. Protein lysates were collected and proteins separated by 1D SDS/PAGE, western blotted and probed with anti-phospho-ERK IgGs (pERK). Blots were stripped and reprobed with an anti-ERK IgG (ERK) as a loading control. (B) M1/WT cells were treated with (+) and without (−) PD98059 (50 µm) for 30 min prior to stimulation with either CSF-1 (5000 U·mL−1) for 4 min or 8BrcAMP (1 mm) for 30 min. Protein lysates were collected and proteins separated by 1D SDS/PAGE, western blotted and probed with anti-phospho-ERK IgGs (pERK). Blots were stripped and reprobed with an anti-ERK IgG (ERK) as a loading control. (C) M1/WT cells were treated with CSF-1 (5000 U·mL−1) in the absence or presence of PD98059 (50 µm). Cells were incubated for 72 h, washed, then incubated with anti-Mac-1 IgG and FITC-conjugated anti-IgG2b secondary antibody, and the percentage of Mac-1 positive staining cells was determined by flow cytometry. This experiment was repeated four times with similar results; means and standard deviation are shown. (D) M1/WT cells were treated with CSF-1 (5000 U·mL−1) in the absence (i) or presence (iii) of PD98059 (50 µm), or with 8BrcAMP (1 mm) in the absence (ii) or presence (iv) of PD98059 (50 µm). Cells were incubated for 72 h and cellular morphology was examined by light microscopy (20× magnification). This experiment was repeated six times with similar results.

ERK activation is required for M1/WT cell differentiation in response to CSF-1 and 8BrcAMP

Seeing that ERK activation correlated with the degree of differentiation in M1/WT cells, we determined whether the MEK inhibitor, PD98059, might suppress the macrophage-like differentiation. We have previously shown that PD98059 (50 µm) inhibited the increased ERK activity in CSF-1-treated BMM suggesting MEK dependence [11]; we confirm that this result is also true for M1/WT cells (Fig. 4B) and show that PD98059 also inhibits the weak ERK activation induced by 8BrcAMP alone in M1/WT cells. Figure 4C demonstrates that PD98059 (50 µm) could effectively inhibit CSF-1-mediated Mac-1 expression in M1/WT cells, suggesting MEK/ERK involvement. Figure 4D illustrates that PD98059 addition inhibited both CSF-1 and 8BrcAMP-mediated M1/WT cell differentiation as determined by morphology after 72 h incubation. It was also found that the differentiation induced by 8BrcAMP or PGE1 in the presence of CSF-1 was suppressed by the inclusion of PD98059 or UO126 (data not shown). Thus a MEK/ERK pathway would appear to be necessary for the differentiating actions of both CSF-1 and cAMP in M1/WT cells.

Discussion

We have provided evidence that cAMP can enhance CSF-1-induced macrophage differentiation in M1/WT cells. Of more direct significance, our findings are consistent with the report that human monocyte differentiation is accompanied by an increase in intracellular cAMP [12]. Also, as discussed previously in the context of suppression of CSF-1-driven macrophage lineage proliferation [6,9], it is likely that macrophage populations in vivo, particularly at sites of inflammation, will be exposed to prostaglandins with the potential to elevate intracellular cAMP. From our studies the effects on macrophage lineage differentiation by cAMP elevating agents should now be considered as a possibility. Even though further experiments are required to exclude formally the involvement of endogenous CSF-1, our data with neutralizing antibody (Fig. 2C) suggest that cAMP can somehow interact with CSF-1R-dependent signal transduction cascades even in the absence of CSF-1 and are consistent with the observation that the CSF-1R−/− mouse shows a more severe phenotype than that of the op/op mouse with an inactivating mutation in the CSF-1 gene [28]; in that study the authors suggested that there may be CSF-1-independent activation of the CSF-1R.

In the experiments reported above we extended our prior studies on the inhibition by elevated cAMP on the CSF-1-dependent proliferation in BMM [9] by comparing some of the responses in CSF-1-treated M1/WT cells. The overall biological response to CSF-1 in each of these two cell populations is determined by the nature of the target cell, i.e. either differentiation or proliferation. Increases in cAMP in M1/WT cells potentiated the differentiation induced by CSF-1 while inhibiting the mitogenic action of CSF-1 on BMM [10]. However, these particular biological responses to elevated cAMP [10] may not be so different as they both involve cell cycle exit; for BMM it is also possible, by analogy with human monocytes [12] and other cell types [18,29], that the cell cycle exit in response to elevated cAMP may ultimately be accompanied by differentiation. In support of this proposed analogy in the responses of the two cell types and the possible usefulness of the M1/WT cells as a model for understanding the early biochemical responses to CSF-1 in normal cell types, cAMP promoted an increase in ERK activity in both CSF-1-treated M1/WT cells and BMM [11]. In the context of our findings and the reports describing increases in cAMP associated with macrophage differentiation [12–15], it is worth noting that the cAMP phosphodiesterase, PDE4, is down-regulated during macrophage differentiation [30]. In addition, the possible relationship between the function of PDE4 splice variants during macrophage differentiation [30,31] and ERK activity [32–35] is worth exploring in the light of our findings above.

In spite of the potentiation of the CSF-1-induced differentiation and ERK activity in M1/WT cells by cAMP elevation, we found, perhaps surprisingly, that instead of there being more tyrosine phosphorylation there was in fact a reduction in the degree of tyrosine phosphorylation of CSF-1R and that was found generally in whole cell lysates, both in the absence and presence of CSF-1. Again supporting the analogy drawn above between the early responses to CSF-1 in both cell types, these changes in tyrosine phosphorylation were similar in CSF-1-treated BMM. These findings with tyrosine phosphorylation raise questions about the significance of acute CSF-1R-dependent tyrosine phosphorylation for the subsequent cellular responses to CSF-1. They suggest that the bulk of the CSF-1R-dependent tyrosine phosphorylation in M1/WT cells, including that of CSF-1R itself, is not critical for CSF-1-dependent differentiation and ERK activity. For BMM, even though it could be interpreted that the reduced CSF-1-dependent tyrosine phosphorylation due to enhanced cAMP correlates with reduced proliferation, the two responses may not be linked, particularly as CSF-1-dependent ERK activity is also enhanced by elevated cAMP in this cell type. In our previous studies with CSF-1-stimulated BMM we could find no evidence for a suppression by raised intracellular cAMP on early responses such as Na+/H+ antiport activity, Na+/K+ ATPase activity, protein synthesis, etc. [9]; however, we also reported a MEK-dependent increase in c-fos mRNA expression [11]. Therefore, like M1/WT differentiation, there are a number of CSF-1-dependent responses that are intact in BMM treated with cAMP elevating agents. These responses also inversely correlate with the majority of CSF-1-induced tyrosine phosphorylation.

Based on the analogous acute biochemical responses of CSF-1-treated M1/WT cells and BMM to elevated intracellular cAMP we propose that there are at least two independent pathways emanating from the activated CSF-1R. One pathway(s) involves extensive tyrosine phosphorylation of numerous substrates, including the CSF-1R, and can be blocked by increases in cAMP, while the other(s) is MEK/ERK-dependent. We previously reported observations in CSF-1-treated BMM consistent with this model where we found that the antiproliferative effect of 8BrcAMP correlated with reduced cyclin D1 and delayed c-myc mRNA expression, while PD98059 addition did not lower such mRNA expression [36]. However, when 8BrcAMP was combined with PD98059, dramatic apoptosis was noted in CSF-1-treated BMM [36], in the CSF-1-treated 32D myeloid cell line [37], and in CSF-1-treated M1/WT cells (NJ Wilson and JA Hamilton, unpublished observation). These data indicate that the two proposed CSF-1-dependent pathways converge at some point to promote cell survival in some cell types with this convergence being critical for this parameter.

The suppressive effects of cAMP on CSF-1-dependent tyrosine phosphorylation reported above are consistent with those found in a recent paper describing inhibition by cAMP of EGF-stimulated EGFR tyrosine phosphorylation and tyrosine phosphorylation of cellular proteins [20] but are unlike the data in other recent reports showing rapid tyrosine phosphorylation of the EGFR in response to PGE2 in colon cancer cells [38] and of the EGFR and NGFR in PC12 cells by raised intracellular cAMP [17]. Therefore, as for the work of Barbier et al. for EGF-treated cells [20], it would seem that CSF-1R-dependent tyrosine phosphorylation in M1/WT cells and BMM is an additional example where RTK-dependent tyrosine phosphorylation is inhibited by cAMP, possibly via PKA activity. Whether this inhibition is actually at the level of CSF-1R itself is unknown as it is possible that a downstream tyrosine kinase(s) may be responsible for much of the CSF-1-dependent tyrosine phosphorylation. For example, we and others have demonstrated a role for Src in CSF-1-mediated receptor tyrosine phosphorylation [5,39], differentiation [5] and proliferation [40], while others have shown that cAMP can down-regulate Src-family members via PKA directly activating the Src inhibitory kinase, CSK [41,42]. However, 8BrcAMP alone was able to differentiate M1 cells only when they expressed a CSF-1R which is capable of leading to cellular differentiation, suggesting that cAMP may in fact interact with an active CSF-1R itself in some way.

We are presently unsure as to which direct substrates cAMP may be acting upon to exert the effects discussed. Although cAMP is usually thought to activate PKA it has recently been shown that increasing cellular cAMP can activate the guanine nucleotide exchange factors, exchange protein activated cAMP (Epac)1 and Epac2, independently of PKA activation [43,44]. It has been demonstrated that Epac does not alter cAMP-mediated ERK activation but can mediate other cAMP-stimulated events, such as exocytosis and cell adhesion. Epac1 has also recently been shown to be expressed in macrophages where it is responsible for cAMP-mediated suppression of phagocytosis [45]. Whether Epac activation may also alter CSF-1R tyrosine phosphorylation is unknown. Others have shown that increases in cAMP can trans-modulate the EGFR [17,19,38] and that PKA can directly phosphorylate the EGFR on a particular serine residue [19]. Interestingly, the CSF-1R also contains a serine residue in a PKA consensus phosphorylation motif, which is located within the kinase-insert region (NJ Wilson and JA Hamilton, unpublished observation). Although we are currently determining the significance of this serine on CSF-1R tyrosine phosphorylarion we cannot rule out the possibility that cAMP, through PKA or Epac, may instead activate a tyrosine phosphatase which results in the observed decrease in both basal and CSF-1-induced tyrosine phosphorylation.

In contrast to the data for EGFR-dependent tyrosine phosphorylation reported by Pai et al. [38], the surprising result for both M1/WT cells and BMM is that there is still elevation of CSF-1-induced ERK activation by cAMP and not inhibition. Using PD98059, we provided evidence before for a partial dependence of CSF-1-induced DNA synthesis on a MEK/ERK-dependent pathway [36] while others, using a similar strategy, reported that ERK activity is essential for CSF-1-induced proliferation in Bac1.2F5 and BMM cells [46,47]; from our findings above it would appear for CSF-1-induced differentiation in M1/WT cells that such a pathway(s) is critical. However, whether other nonspecific effects of PD98059 are occurring over time need to be excluded before definitive conclusions can be drawn about the significance of the MEK/ERK pathway in CSF-1-treated cells.

The significance of the pathway(s) involving the bulk of the CSF-1R-dependent tyrosine phosphorylation and which lies downstream of cAMP action awaits clarification. However, given the increasing number of reports showing increased intracellular cAMP affecting RTKs and/or subsequent downstream signalling, elevation of cAMP may be a common mechanism by which cells alter the cellular outcome of RTK activation.

Experimental procedures

Cells and media

M1 murine myeloid cells were a gift from N Nicola (Walter and Eliza Hall Institute, Melbourne, Australia) and were maintained in DMEM (Trace Biosciences, Sydney, Australia) with 10% newborn calf serum (Trace Biosciences) at 37 °C in 5% humidified CO2. M1 cells expressing wild-type (WT)-CSF-1R (M1/WT) or Y807F-CSF-1R (M1/807) were constructed as described previously [5,20]. BMM were obtained as described previously [48] and maintained at 37 °C in 5% humidified CO2 in RPMI with 10% fetal calf serum and 30% L-cell conditioned medium, as a source of CSF-1. The BMM were deprived of CSF-1 for 18 h to growth-arrest them before the start of the experiments.

Antibodies and reagents

Mac-1-expressing hybridoma cells were from the American Tissue Culture Collection (Manassus, VA, USA). The following reagents were purchased as follows: phospho-ERK antibody (New England Biolabs, Inc., Beverly, MA, USA); ERK antibodies (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA); 4G-10 antibody raised against phosphotyrosine (Upstate Biotechnology, Charlottesville, VA, USA); anti-CSF-1R IgG (Upstate Biotechnology); HRP-conjugated rabbit anti-goat and swine anti-rabbit IgGs (DAKO, Glostrup, Denmark); PD98059 (New England Biolabs, Inc., Ipswich, MA, USA); PGE1 and PGE2 (ICN); and the sodium salt of 8BrcAMP (Sigma Chemical Co., St Louis, MO, USA). Recombinant human CSF-1 [49] was donated by Chiron Corp., Emeryville, USA. Assays to determine the levels of lipopolysaccharide (LPS) in PGE1, PGE2, 8BrcAMP and CSF-1 were routinely performed. Reagents found negative for LPS were used in all experiments.

Light microscopy

Cell morphology was examined with a Leica inverted microscope prior to image acquisition with a Sony digital Hyper HAD colour video camera (Sony Corporation, Tokyo, Japan) and Leica q500mc Windows software (Leica, Cambridge, UK). In experiments assessing the role of ERK inhibition on M1/WT cell differentiation cells were treated for 72 h with the MEK inhibitor, PD98059 (50 µm), with either CSF-1 or 8BrcAMP, and the morphology examined as above.

Mac-1 staining

Mac-1 staining was performed as described previously [20]. Briefly, 2 × 105 cells were incubated with hybridoma cell supernatant containing antibody raised against Mac-1 or 50 µL isotype-matched control (rat anti-mouse IgG2b), left on ice for 1.5 h, washed with ice cold NaCl/Pi, incubated with fluorescein isothiocyanate (FITC)-conjugated anti-rabbit IgG, left on ice for a further 30 min and finally washed with ice cold NaCl/Pi again. Fluorescence was measured using a FACS Calibur flow cytometer (Becton Dickinson, San Jose, CA, USA). Acquisition was restricted to 10 000 events for each sample and Mac-1 positive cells were calculated by subtracting the isotype-matched control value from the Mac-1 positive value. Median fluorescence intensity was determined by calculating the median of Mac-1 fluorescence for 10 000 events.

Western blot analysis

Western blot analysis was performed as described previously [20]. Briefly, cells were harvested at 1 × 107 cells in 300 µL lysis buffer (5 mm EDTA, 25 mm Hepes pH 7.4, 100 mm NaCl, 1% Triton X-100 and 10% glycerol containing 70 IU·mL−1 aprotinin, 10 µg·mL−1 leupeptin, 100 mm NaF, 0.1 mm Pefabloc and 200 µm sodium orthovanadate). Proteins were size-separated on 10% (w/v) SDS/polyacrylamide gels and then transferred to Hybond C (Amersham, Baulkham Hills, NSW, Australia). Membranes were then immunoblotted with appropriate antibody and subjected to chemiluminescence (Amersham ECL reagents and Hyperfilm).

Immunoprecipitation

Cytosolic lysates were prepared as follows: 1 × 107 cells were scraped in lysis buffer (as above) and left on ice for 10 min before centrifugation at 5000 g for 5 min. The pellets were discarded and immunoprecipitations performed by incubating lysates (100 µg) with 1 µg of antibodies raised against CSF-1R overnight at 4 °C. Twenty microlitres of a 50% (v/v) slurry of Protein A-Sepharose 4B (Pharmacia, Uppsala, Sweden) were added to the lysates and incubated for 30 min at 4 °C. An equal volume of 2× SDS/PAGE sample buffer was added and the immunoprecipitates were boiled for 5 min and separated on 10% (w/v) SDS/polyacrylamide gels before western blotting.

Ancillary