Multiple Doses of Diacylglycerol and Calcium Ionophore Are Necessary to Activate AP-1 Enhancer Activity and Induce Markers of Macrophage Differentiation*

In contrast to phorbol esters, multiple doses of diacylgycerols are needed to differentiate U937 human monoblastic leukemic cells to a macrophage-like phenotype. Although both of these agents similarly activate protein kinase C in vitro, it is not known why these agents appear to have differing biologic effects. One possibility is that they regulate gene transcription in slightly different ways. Regulation of gene transcription by phorbol esters is complex and involves the stimulation of the transactivating proteins Jun and Fos which form dimers and bind to the AP-1 enhancer elements (5'-TGAGTCA-3'). To understand whether diacylglycerols regulate gene transcription similarly to phorbol esters and to examine whether activation of AP-1 enhancer activity is correlated with differentiation, we have treated U937 human monoblastic leukemic cells with these agents and examined activation of transcription from AP-1 enhancer elements. We find that, although a single dose of diacylglycerol, like phorbol esters, is sufficient to elevate mRNA levels of both the c-jun and c-fos protooncogenes, in contrast to phorbol esters there is no increase in either Jun protein or activation of AP-1 enhancer activity. However, multiple doses of this agent given over 24 h stimulate repeated elevations in c-jun and c-fos mRNA, increases in Jun protein, and enhancer activation. Treatment of U937 cells with ionomycin, a calcium ionophore, also stimulates an increase in c-jun mRNA, but neither activates AP-1 enhancer activity nor stimulates differentiation of these cells. However ionomycin functions to enhance the effects of diacylglycerols both on transcriptional activation and U937 differentiation. These results suggest a complex regulation of AP-1 enhancer activity in U937 cells by diacylglycerols involving both transcriptional and post-transcriptional regulatory mechanisms. Maximal activation of AP-1 enhancer elements, and not changes in jun and fos mRNA, is correlated with increases in markers of U937 differentiation. These changes may be important in the early events leading to differentiation of hematopoietic cells.

In contrast to phorbol esters, multiple doses of diacylgycerols are needed to differentiate US37 human monoblastic leukemic cells to a macrophage-like phenotype.
Although both of these agents similarly activate protein kinase C in vitro, it is not known why these agents appear to have differing biologic effects. One possibility is that they regulate gene transcription in slightly different ways. Regulation of gene transcription by phorbol esters is complex and involves the stimulation of the transactivating proteins Jun and Fos which form dimers and bind to the AP-1 enhancer elements (5'-TGAGTCA-3').
To understand whether diacylglycerols regulate gene transcription similarly to phorbol esters and to examine whether activation of AP-1 enhancer activity is correlated with differentiation, we have treated U937 human monoblastic leukemic cells with these agents and examined activation of transcription from AP-1 enhancer elements. We find that, although a single dose of diacylglycerol, like phorbol esters, is sufficient to elevate mRNA levels of both the c-jun and c-fos protooncogenes, in contrast to phorbol esters there is no increase in either Jun protein or activation of AP-1 enhancer activity. However, multiple doses of this agent given over 24 h stimulate repeated elevations in c-jun and c-fos mRNA, increases in Jun protein, and enhancer activation. Treatment of U937 cells with ionomycin, a calcium ionophore, also stimulates an increase in c-jun mRNA, but neither activates AP-1 enhancer activity nor stimulates differentiation of these cells. However ionomycin functions to enhance the effects of diacylglycerols both on transcriptional activation and U937 differentiation. These results suggest a complex regulation of AP-1 enhancer activity in U937 cells by diacylglycerols involving both transcriptional and post-transcriptional regulatory mechanisms.
Maximal activation of AP-1 enhancer elements, and not changes in jun and fos mRNA, is correlated with increases in markers of U937 differentiation.
These changes may be important in the early events leading to differentiation of hematopoietic cells.
The addition of phorbol esters to both fresh samples of human leukemic cells and leukemic cell lines, HL-60 and * This work was supported by United States Public Health Service (to M. K. and A. S. K.) and the American Cancer Society and Office of Naval Research (to A. S. K.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
ll To whom correspondence and reprint requests should be addressed. Tel.: 205-934-4436. U937 cells, induces these cells to differentiate into macrophages (Kraft et al., 1989;. This activation of differentiation is mediated at least in part by phorbol ester activation of protein kinase C and translocation of this enzyme to the cell membrane (Kraft et al., 1987;Ebeling et al., 1985). Unlike phorbol esters (Morin et al., 1987), a single dose of a diacylglycerol which activates protein kinase C in uitro (Sharkey et al., 1984) has little effect on the phenotype of these cells. Although identical results have not been reported by all laboratories (Kreutter et al., 1985;Yamamoto et al., 1985), multiple doses of diacylglycerols administered to leukemic cells over 24 h appear to induce changes in surface markers and a decrease in c-myc transcription characteristic of differentiated leukemic cells (Ebeling et al., 1985;Salehi et al., 1988). Since diacylglycerols and phorbol esters activate protein kinase C in a similar fashion (Sharkey et al., 1984), it is unknown why multiple doses of diacylglycerol are needed to mimic the effects of phorbol esters in uivo. In addition, it is not clear why calcium ionophores, which have no effect on differentiation of these cells alone (Morin et al., 1987), enhance the biologic activity of diacylglycerols.
Diacylglycerols are important intracellular second messengers (Rink et al., 1983;Rittenhouse-Simmons, 1979;Rittenhouse, 1982). Increases in diacylglycerols and intracellular calcium induced by the binding of polypeptide growth factors activate protein kinase C (Nishizuka, 1984) and stimulate the transcription of specific gene products (Kaibuchi et al., 1986;Salehi et al., 1988). Like diacylglycerols, phorbol esters bind to the regulatory portion of protein kinase C; however, although both activate protein kinase C similarly in vitro (Sharkey et al., 1984), they regulate this enzyme differently in viuo. For example, prolonged incubation of cells with phorbol esters causes the eventual degradation of all cellular protein kinase C (Kraft et al., 1987); this does not occur when diacylglycerols are used (Ways et al., 1987). Also, the addition of diacylglycerols to cells labeled with [32P]orthophosphate does not stimulate protein phosphorylation identical with that induced by phorbol esters (Morin et al., 1987, Strulovici et al., 1989, and, unlike phorbol esters, the application of diacylglycerols to U937 human monocytic leukemia cells was unable to induce interleukin l@ mRNA (Strulovici et al., 1989). It has been suggested (Strulovici et al., 1989) that the differences between diacylglycerols and phorbol esters may reside in the inability of diacylglycerols to translocate protein kinase C to the membrane for prolonged periods, to activate specific isoforms of protein kinase C, or to stimulate the activation of specific mRNAs.
Phorbol ester induction of gene transcription has been shown to be mediated by consensus DNA sequences AP-l-2-3 recognized by phorbol ester-modulated transacting factors Multiple Doses of DiC8 Activate AP-1 Enhancer Activity 18167 (Chiu et al., 1988). The AP-1 enhancer sequence, 5'-TGAGTCA-3', binds the proteins encoded by the c-jun and c-fos protooncogenes (Angel et al., 1987a, 1988a, Lee et al., 1987. The addition of phorbol esters to both quiescent fibroblasts and epidermal cells (Berstein and Colburn, 1989;Lamph et al., 1985;Quantin and Breathnach, 1988) stimulates an increase in the transcription of mRNA for both of these genes. Although Jun is capable of binding to this sequence as a homodimer, the formation of a Jun/Fos heterodimer (Halazonetis et al., 1988;Kouzarides and Ziff, 1989;Neuberg et al., 1989) greatly enhances the ability of these proteins to stimulate transcription from the AP-1 enhancer sequence. Because both Jun and Fos are phosphorylated proteins (Barber and Verma, 1987;Boyle et al., 1989), it is possible that post-translational modifications of these proteins may play an important role in the regulation of AP-1 enhancer activity. Not only phorbol esters but epidermal growth factor (Kerr et al., 1989) and tumor necrosis factor (Y (Brenner et al., 1989) appear to regulate gene transcription through AP-1 enhancer elements. Recent evidence demonstrates that the addition of differentiating agents, phorbol esters, bryostatin, and vitamin DJ, to human myeloid leukemic cells stimulates increases in c-jun RNA (Sherman et al., 1990, Aureux et al., 1990) and suggests that c-jun levels may be correlated with differentiation.
To determine whether diacylglycerols activate AP-1 enhancer activity in a fashion identical with phorbol esters and whether the activation of the AP-1 enhancer activity is correlated with differentiation, we have compared the effects of diacylglycerols with or without calcium ionophore and phorbol esters on U937 human monoblastic leukemia cells. We find that although a single dose of diacylglycerol (DiC8)' induces elevations in both c-jun and c-fos RNA, unlike phorbol esters, it neither activates the AP-1 enhancer to stimulate transcription nor induces the differentiation of these cells. In contrast, multiple doses of DiC8 given at a 2-h intervals stimulate 1) repeated elevations in the messenger RNA levels for these protooncogenes, 2) increases in Jun protein, 3) activation of transcription from the AP-1 enhancer element, and 4) the induction of markers of differentiation. These data demonstrate that increases in c-jun and c-fos messenger RNA levels do not necessarily correlate with activation of AP-1 enhancer elements, suggesting a complex post-transcriptional control of AP-1 enhancer activity in these hematopoietic cells. and RNA-isolated using guanidinium thiocyanate/phenol/ chloroform (Kraft et al., 1989). Approximately 20 pg of RNA from each time point was loaded onto a 1.5% agarose/formaldehyde gel. After electrophoresis, the RNA was transferred to nitrocellulose and cross-linked with UV light, and the nitrocellulose blot was incubated with labeled probes (Kraft et al., 1987). Purified fragments of c-jun (Angel et al., 1988b), c-fos (Miller et al., 1984), and cu-tubulin (Cleveland, 1980)  Immunoprecipitation-2 x lo7 cells were incubated in methioninefree medium for % h prior to the addition of 150 &i/ml of [35S] methionine for 1 h. Immunoprecipitation was then carried out as previously described (Chiu et& 1988). Cells were lysed in 1 ml of ice-cold radioimmunoDreciDitation buffer (RIPA). and the lvsates were clarified by centrjfugaiion at 15,000 X g for 36 min at 4 'c. 2 ~1 of anti-trpE-c-jun antisera were added to the lysate and incubated at 4 "C overnight. Antibodv

RESULTS
Adding phorbol esters to U937 leukemic cells stimulates an increase in c-jun and c-fos RNA levels starting within 15 min of addition and decreasing after l-2 h (Fig. 1). Changes occur in both the 2.7-and 3.4-kb c-jun mRNA over a similar time course. Likewise, both the HL-60 and PLB-985 leukemic cell lines (Tucker et al., 1988) which are induced to differentiate into macrophages by phorbol esters yielded similar changes in both of these mRNAs (data not shown) when treated with phorbol esters. To examine the phorbol ester induction of Jun protein synthesis, we labeled the U937 cells with [35S]methionine and immunoprecipitated equivalent numbers of [35S] methionine counts at varied times after phorbol ester treatment. Immunoprecipitation of Jun protein (Fig. 2) demonstrates that it is transiently induced, peaking at 6 h, and by 24 h decreases to baseline. At all times examined, small amounts of Jun protein were present in these cells. Similar results were obtained when equivalent numbers of cells were subjected to this immunoprecipitation procedure. To evaluate the activation of the AP-1 enhancer element by Jun and Fos during differentiation of U937 cells, we electroporated chimerit plasmids containing a single copy of the phorbol ester responsive element (AP-1 site) (-73/+63) within the collagenase promoter, ligated to the chloramphenicol acetyltransferase (CAT) (Angel et al., 198713) gene. The addition of PMA (+) to these cells stimulated CAT activity from the construct containing an AP-1 site but not from one lacking an AP-1 site (-60/+63) (Fig. 3). Likewise, the addition of PMA to these cells stimulates the activity of constructs containing five copies of the AP-1 site upstream of the herpes simplex virus-tk promoter (5 X AP-l), but not the activity of tk-CAT (-enh) lacking this enhancer element. A construct containing the CMV promoter was used as a control. However, phorbol  ester treatment was found to stimulate the transcription of this promoter 2-3 fold. Since the 5 X AP-1 construct was stimulated to a different degree than the CMV promoter, the increases are not simply secondary to changes in RNA synthesis induced during differentiation.
Although the effects of PMA on this cell line involve the inhibition of cell growth and DNA synthesis, the time course of increase in c-jun RNA levels is similar to that of those cell lines stimulated to grow by specific mitogens (Ryder and Nathans, 1988;Ryseck et al., 1988).
Both calcium ionophore (May et al., 1989) and diacylglycerol (for reviews, Bell, 1986;Nishizuka, 1984) activate protein kinase C and, when added to leukemic cells either together or separately in multiple doses, induce macrophage-like differentiation (Ebeling et al., 1985;Morin et al., 1987). The addition of a single dose of either ionomycin (1 PM) or DiC8 (100 pM) to U937 cells induces increases in total cellular c-&n and c-fos RNA over a similar time course as that induced by phorbol esters (Figs. 1 and 4). Densitometry of Northern blots and the tubulin controls has allowed us to quantify the changes induced by this agent. Maximal doses of DiC8 induce a g-fold change in relative c-jun levels while ionomycin stimulated a 4-fold change. The 5.8-fold change induced by PMA was at a level similar to these other agents. The time course of elevation of c-jun and c-fos by lower doses of both iono-mycin and DiC8 was identical with that at higher concentrations.
Because the Fos and Jun proteins interact to form a more effective transcriptional activator than Jun alone operating through the AP-1 site (Chiu et al., 1988;Halazonetis et al., 1988;Rauscher et al., 1988), we determined the effect of DiC8 and ionomycin on 5 x AP-1 CAT activity. This construct was electroporated into U937 cells, and the effect of either ionomycin alone or in combination with DiC8 was tested. Neither a single dose of ionomycin given over a broad concentration range (Fig. 5A) nor a single dose of 100 FM DiC8 was capable of activating 5 x AP-1 CAT (Fig. 5B). Ionomycin alone appeared to slightly inhibit the 5 x AP-1 activity without effecting cell viability. However, the addition of a single dose of ionomycin and DiC8 together stimulated a 4.5-fold increase in CAT activity. Multiple repeat doses of ionomycin and DiC8 given at 2-h intervals caused marked increases in activity, with 12 doses stimulating enzyme activity to approximately the same level as a single dose of PMA (53.5 uersu.s 72.6) ( A, the effect of ionomycin on CAT activity. U937 cells were electroporated with 5 X AP-1 CAT and divided into 7 aliquots. One aliquot remained untreated. Each of the remaining aliquots was treated with varying doses of ionomycin, and one also with PMA. After 24 h, cells were homogenized, and CAT activity was determined. Data presented are representative of experiments performed in triplicate. The average of the fold activation of the triplicate determinations + S.D. is as follows: 5, 0.8 + 0.4, 1, 1 f 0.5, 0.5, 0.9 + 0.6; 0.05, 0.7 f 0.6; 0.005, 0.8 + 0.5. The percent acetylation of each treatment was divided by the control (C) value to obtain the fold activation. B, the effect of combined ionomycin and DiC8 on CAT activity. U937 cells were electroporated with 5 X AP-1 CAT, and cells were divided into 10 aliquots. One aliquot was not treated and remained as a control (C). One aliquot was treated with 0.1 pM PMA. The rest of the aliquots were treated with 100 pM DiC8 (D) alone or with 1 NM ionomycin. These treatments were carried out every 2 h either IX, 3X, 6X, or 12X. CAT activity was expressed first as percent acetylation of chloramphenicol. The percent acetylation of each treatment was divided by the control (C) value to obtain the fold activation. The data shown are representative of experiments performed in triplicate. The average of these triplicate determinations f SD. is as follows: I X D, 1.6 -C 0.5; I X DI, 4.9 + 1.2; 3 X D, 5.8 + 1.8; 3 x DI, 17.8 + 4.3; 6 x D, 11 f. 3.7; 6 x DI, 35 + lo; 12 x D, 15 f 6.8; 12 x DI, 55 + 12; PMA, 69 + 22. 5B). These data demonstrate that a single dose of ionomycin or DiC8 is sufficient to induce an elevation in c-jun RNA but that multiple doses of both agents together are necessary to maximally stimulate 5 x AP-1 CAT activity.
To explain why a single dose of ionomycin or DiC8 failed to stimulate 5 x AP-1 CAT activity, we labeled U937 cells with [35S]methionine and immunoprecipitated Jun protein after the addition of various combinations of DiC8 and ionomycin or PMA alone. Immunoprecipitation was done at 6 h because Jun protein levels were maximal at this time point (Fig. 3). In comparison to PMA, little change was seen in Jun protein levels with either a single dose of DiC8 or ionomycin. Also, multiple doses of ionomycin had little effect on the levels of this protein. However, when U937 cells were treated with multiple doses of DiC8 with or without ionomycin, marked increases in Jun protein occurred (Fig. 6). The prominent 43-kDa protein seen in both Figs. 6 and 3 was not competed with trp-Jun fusion protein. Control immunoprecipitations performed at time points earlier than 6 h after treatment exclude the possibility that DiC8 and ionomycin stimulated changes in Jun protein over a different time course than PMA (data not shown). These results demonstrate that activation of the 5 x AP-1 CAT construct correlates with the ability of these agents to increase immunoprecipitable Jun protein.
To understand why multiple doses of DiC8 are necessary to elevate Jun protein, we examined changes in c-jun and cfos RNA levels after multiple DiC8 treatments. We found that, when DiC8 was given at 2-h intervals, each dose of DiC8 induced a repeated increase in c-jun (Fig. 7) and c-fos messenger RNA levels (identical data not shown). With each dose of diacylglycerol, c-fos RNA levels were rapidly induced and then declined over 2 h toward baseline. Densitometry of the c-jun and tubulin Northern blots and calculation of relative changes in c-jun mRNA levels demonstrates a 4-8-fold induction of cellular levels of c-jun by a single dose of DiC8 and ionomycin which returns toward baseline by 2 h and is again induced by subsequent doses of DiC8. A third dose of DiC8 again induces an increase in c-jun mRNA.
The addition of phorbol esters to U937 cells  induces increases in 1) superoxide anion production, 2) responsiveness to chemotactic peptide, 3) expression of Leu-11 surface receptors, and 4) a marked decrease in cell surface OKT9 (May et al., 1989) with little effect on surface IgG. Examination of these surface markers of differentiation 24 h after treatment with DiC8 and/or ionomycin demonstrates that the addition of a single dose of either agent is not sufficient to induce changes in Leu-11 or OKT9. However, the addition of multiple doses induces a fall in the percentage of cells positive for OKT9 (86% to 44%) and a rise in the percentage of cells positive for Leu-11 (2.0% to 39%) (Fig. 6B). The reduction in OKT9 and the increase in Leu-11 expression induced by three doses of these agents was not as dramatic as the changes seen after PMA treatment. The surface levels of IgGi or OKM-1 which are unaffected by PMA treatment are also unchanged by these treatments. The activation of AP-1 enhancer activity and increases in Jun protein appear to correlate with the induction of differentiation markers in U937 cells.

DISCUSSION
Activators of leukemic differentiation including bryostatin, vitamin D1, and phorbol esters induce similar increases in Cjun mRNA (Sherman et al., 1990;Auwerx et al., 1990). By comparing the effects of PMA and DiC8 (with or without ionomycin) on U937 cells, we found that a single dose of A   I  I  I  I  I  I  I  I  I  I  2  I  I  A, immunoprecipitation of Jun protein in DiC8 or ionomycin-treated U-937 cells. U937 cells were treated with 1 pM ionomycin, alone or in combination with 100 pM DiC8, either once (IX) or three times (3x). Control cells were either untreated or treated with 0.1 pM PMA. After treatment, cells remained in DMEM medium plus 10% calf serum for 6 h. They were then spun down and placed in methionine-free medium for % h with 150 &i/ml of [YS]methionine. Cells were then spun down again, homogenized and immunoprecipitated as described under "Experimental Procedures." An aliquot of the radioimmune precipitation buffer homogenate of PMAtreated cells was immunoprecipitated with normal rabbit serum (NRS). Equivalent numbers of counts per min were loaded in each lane. B, the effect of DiC8 and ionomycin treatments on the induction of cell surface markers of differentiation. U937 cells were treated as described above and at 24 h were sent for FACS analysis of surface markers. The number shown is the percentage of cells expressing each marker and is the average of triplicate experiments carried out over a 3-month period. N.D., experiment not done. diacylglycerol stimulated an elevation of both c-jun and c-{OS RNA over a similar time course and to the same extent as phorbol esters, but, unlike PMA, did not induce an increase in AP-1 enhancer activity or differentiation markers. In contrast, multiple doses of DiC8 which induce the differentiation of U937 cells cause an elevation both in Jun protein levels and an activation of AP-1 enhancer activity. In addition, we have also shown that other hematopoietic differentiating agents , including interferon y, granulo- U937 cells were treated with DiC8 (100 PM) and ionomycin (1 PM) at zero time and with DiC8 (100 pM) at 2 and 4 h. At each time point, total RNA was isolated as described under "Experimental Procedures." The Northern blot was probed with c-jun and tubulin and scanned on a densitometer. The c-jun levels were divided by the oc-tubulin amounts and expressed as relative RNA levels. The baseline is given an arbitrary value of 1.0. The time at which cells were treated is shown by arrows.
correlate with the induction of differentiation. Instead, activation of transcription from the plasmid containing the 5 X AP-1 enhancer element linked to the CAT gene correlates with the induction of macrophage-like differentiation by a wide range of agents. This work suggests a complex regulation of AP-1 enhancer activity.
Like diacylglycerols, the calcium ionophore ionomycin also stimulates increases in c-jun messenger RNA levels. Our results are similar to those found in a wide range of hematopoietic cell lines using the calcium ionophore A23187 (Auwerx et al., 1990). This increase can be maximally induced at nanomolar concentrations of ionomycin suggesting that changes in intracellular calcium concentrations are sufficient to induce changes in c-jun mRNA levels. However, although the increases in mRNA occur over a similar time course and to a similar extent as after PMA treatment, there is no activation of either Jun protein synthesis or activation of AP-1 enhancer activity. The addition of ionomycin in combination with DiC8 enhances the ability of the latter agent both to induce Jun protein and AP-1 enhancer activity. The effects of calcium ionophores closely parallel the biologic activity of this compound. Alone, calcium ionophores have little effect on the differentiation of human leukemic cells, but greatly enhance the ability of DiC8 to induce differentiation markers (Morin et al., 1987), and, although diacylglycerols induce the translocation of protein kinase C to the membrane calcium ionophores, they cause little effect on this enzyme (Morin et al., 1987). Increases in intracellular calcium may be sufficient to activate c-jun mRNA synthesis, but activation of protein kinase C may be necessary to allow for both Jun protein synthesis and activation of AP-1 enhancer activity.
Transcriptional and post-translational mechanisms appear to play a role in control of Jun activity. Jun protein activates its own transcription through an AP-1 sequence in the Jun promoter (Angel et al., 1988a). Because this increase occurs rapidly in the presence of cycloheximide, existing cellular Jun protein is likely post-translationally modified to induce changes in transcription (Angel et al., 1988a). Evidence demonstrates that dephosphorylation of the Jun protein may enhance the ability of this transactivating protein to bind the AP-1 sequence (Boyle et al., 1989), suggesting that this is the mechanism for control of c-jun transcriptional activation. In U937 cells, nuclear run-off experiments (data not shown)