IL-4 induces cAMP and cGMP in human monocytic cells

Human monocytes, preincubated with IFN-γ respond to IL-4 by a cGMP increase through activation of an inducible NO synthase. Here, IL-4 was found to induce an accumulation of cGMP (1 – 3 min) and cAMP (20 – 25 min) in unstimulated monocytes. This was impaired with NOS inhibitors, but also with EGTA and calcium/calmodulin inhibitors. These results suggest that: (1) IL-4 may stimulate different NOS isoforms in resting and IFN-γ activated monocytes, and (2) cAMP accumulation may be partially dependent on the NO pathway. By RT-PCR, a type III constitutive NOS mRNA was detected in U937 monocytic cells. IL-4 also increased the [Ca2+]i in these cells. Different NOS may thus be expressed in monocytic cells depending on their differentiation and the signals they receive.


Introduction
Interleukin-4 (IL-4) is a 20kDa glycoprotein, produced by activated T lymphocytes (Th2), mast cells, basophils and other cell types, which displays a broad spectrum of biological activity on haematopoietic cells. On monocytes/macrophages, IL-4 enhances the expression of MHC class II antigens, adhesion molecules of the LFA-1 (leucocyte functional antigen) family 2 and CD23. In contrast, IL-4 down-regulates the expression of Fc3,RI (CD64) and Fc3,RII (CD32) and inhibits the induction of Fc3tRIII (CD16) by IFN-3t 4 or TGF-[3. 5 IL-4 also down-regulates the expression of CD14 on monocytes. 6 IL-4 acts as an antiinflammatory mediator by decreasing the release of IL-1, TNF-a, IL-6, IL-8 and prostaglandin E2 by human monocytes and by reducing their capacity to produce reactive oxygen species. 7-12 IL-4 also induces the secretion of Gand M-CSF by monocytes 3 and triggers their differentiation into macrophage-like cells type in vitro.
The binding component of the IL-4 receptor belongs to the family of haemapoietin receptors 14a5 and is associated, at least in some cell types, to the y chain previously reported for the 16 22 IL-2 receptor.
The importance of protein tyrosine phospholation in IL-4 signalling is now 23 28 29 well recognized and recently, the Jak-3 and 0 Lsk 3 have been identified as the major tyrosine kinases involved. Several transcription factors are selectively activated through IL-4-induced tyrosine phosphorylation. 3 '2 We recently described that IL-4 induced an accumulation of cGMP in monocytes of certain donors, which was greatly enhanced by preincubation with IFN-3t and was dependent on the t-arginine-dependent pathway of nitric oxide synthase (NOS) activation, since it was markedly reduced in the presence of NOS inhibitors. 33 We also reported that IL-4 displayed a biphasic effect on the capacity of cultured monocytes to release nitrite in their supernatants. 4 in the present work, we have further analysed the effect of It-4 on the production by monocytic cells of the cyclic nucleotides cAMP and cGMP.
Isolation of human monocytes: Blood samples were obtained from cytapheresis of normal donors, as residues of platelet preparations, thorugh the courtesy of the blood transfusion centres of Cr6teil and H6tel-Dieu. Blood donors did not show contamination by HIV and hepatitis for 15 min at 4C. The neutralized supernatants B, or suffer from manifestations of allergy. Otherwere kept frozen at -70C until cGMP and wise their clinical and immunological status was cAMP determinations. After acetylation of the not known. Peripheral blood mononuclear cells samples, the concentrations of cGMP and cAMP (PBMC) were purified as previously described. 2 were determined by specific radioimmunoassays, Briefly, blood samples were centrifuged at according to the specifications of the manu-200 x g for 10 min in order to eliminate the factumrs (kits from Amersham France, Les Ulis residual contaminating platelets. Then, samples or from NEN, Dupont, Dreieich, Germany). The were diluted in RPMI 1640 medium (Bioproduct, significance of the results was analysed by the tes Ulis, France) and PBMC were harvested after Student's t-test, adapted for small samples. a 20 min centrifugation at 800 x g on Ficoll-Hypaque (Pharmacia, Uppsala, Sweden)gradient. Identification of constitutive NO synthase in Adherent cell populations were obtained by incu-monocytic cells by reverse transcription-poly- Fluo-3/Ca complex: 535nm), as described previously) 6 The free calcium concentration [Ca + ]i in the cells is related to the fluorescence intensity (F) by the following equation: ca2 + ]i Kd X (F Fmin)/(Fmax F) where Fmax is determined by measuring fluorescence after the cells are lysed in 0.1% Triton X-100 (Sigma), whereas Fmin iS determined after addition of 10mM EGTA. The effective constant Kd of Fluo-3 in solution has been reported to be 400nM. Fluo-3 is analysed in the visible range, thus the problems associated with the use of UV light are avoided. In addition, the large increase of the fluorescence emission by binding of Ca 2 + to Fluo-3 allows a great sensitivity for measuring small variations in calcium concentration. However, upon binding of Ca 2 + there are no shifts in its excitation or emission spectrum, which makes Fluo-3 non-ratioable and therefore will compromise the accuracy of absolute [Ca 2 + ]i determinations. For these reasons, our data have been presented as fluorescence intensities.
induced cAMP response was not significantly affected, indicating that the increase in cAMP was not secondary to a release of prostaglandins (not shown).
IL-4 stimulates cyclic GMP accumulation in human monocytes: The kinetics of IL-4 driven cAMP and cGMP accumulation were compared in unstimulated monocyte preparations from the same donors and typical results are presented in Fig. 3. It was regularly observed that the peak of cGMP accumulation (between 3 and 10 min) preceded that of late cAMP accumulation. For eleven donors tested, the range of cGMP accu-  Table 1.
Inbibitors of the nitric oxide syntbase pathway impair the IL-4-driven cyclic nucleotides accumulation in resting monocytes: The heterodimeric soluble guanylyl cyclases, which convert GTP into cGMP, are activated by NO-generating compounds through interactions of the NO radical with their haeminic prosthetic groups. In order to test whether the observed accumulation of cGMP in response to IL-4 was due to a generation of NO, two classical inhibitors of the Iarginine pathway of NO synthase activation were added to unstimulated monocytes prior to their incubation with IL-4: t-NMMA, and nitro-t-arginine which, at variance with L-NMMA, does not compete with t-arginine for its transport into the cells. As seen in Table 2, the IL-4-driven cGMP accumulation was almost totally abolished in the Table 1. IL-4-driven cyclic nucleotide accumulation in human monocytes is inhibited with a specific neutralizing anti-lL-4 antibody cAMP (fmol/106 cells) cGMP (fmol/106 cells) presence of these competitive inhibitors of t-arginine for the NOS. Interestingly, IL-4-driven cAMP accumulation was also markedly affected with t-NMMA as shown in Fig. 4, suggesting that both cAMP and cGMP production were related to the NO pathway.
Effect of calcium chelator and calmodulin inbibitors on IL-4 driven cGMP accumulation: Several NOS isoforms have been described, which differ in their sensitivity to calcium for the stimulation of their catalytic activity. Various inhibitors of the Ca/CaM complex were thus tested  and, as seen in Table 3, the IL-4-stimulated cGMP increase was greatly reduced either in the presence of the calcium chelator EGTA or of the calmodulin inhibitor W7, both reagents being ineffective alone at the concentrations tested. Similar results were obtained with calmidazolium, another inhibitor of the Ca/CaM complex (not shown). These results therefore suggested that the NOS involved may be different from the inducible type II NOS detected in IFN-7-preincubated monocytes and could be activated through variations in intracytoplasmic calcium concentration and therefore be related to the constitutive NOS isoforms.  Detection by RT-PCR of ype III eNOS mRNA in human monoytic cells: Total RNA from human monocytes and U937 cells were reverse transcribed and amplified by the polymerase chain reaction, using primers specific for the constitutive endothelial human NO synthase. As seen in Fig. 6, a fragment of 277 bp was readily detected in U937 cells, which corresponds to the size expected from the sequence of the endothelial NO synthase. Although much weaker than that of U937 cells, this band was also observed with monocytes (not shown). Treatment of U937 cells with either IL-4, or IL-4 and anti-C23 mAb, or IFN-7, did not modify the expression of cNOS.

Discussion
We have previously observed that, for a fraction of the donors tested, IL-4 induced a cGMP increase in unstimulated monocytes; a much higher response was detected when monocytes were preincubated with IFN-7; this was inhibited in the presence of L-NMMA and was not affected by calcium chelation, suggesting the activation of IL-4 triggers P and cGMP in monocytes an iNOS. In addition, IL-4 induced a slight production of nitrite by monocytes, which was potentiated by IFN-y and inhibited by t-NMMA.
These results suggested that the sequential exposure of monocytes to IFN-y and IL-4 elicited the release of NO from t-arginine, which in turn was able to stimulate soluble guanylate cyclase. We have also reported that IL-4 induced an heterogeneous and late accumulation of nitrite in monocyte supernatants. 4 In the present work, we show that IL-4 induces cAMP and cGMP generation in unstimulated human monocytes. A rather large heterogeneity in the response of the different individuals was observed, which may be explained by the fact that, except for the absence of HIV or hepatitis infection and of allergy, the immunological status of these donors was largely unknown. One cannot exclude the possibility that some of these volunteers were primed otherwise which may explain the variation in the response observed and the biphasic cAMP response which was detected for some of them. We have no direct explanation for this early cAMP increase; tentatively, one may speculate that it could be achieved through a cross stimulation of 2 receptors, inasmuch as we have previously demonstrated a functional interplay between the response of monocytes to IL-4 and to 2 agonists.
Hart et al. did not detect significant cAMP accumulation in elutriation-enriched monocytes stimulated by IL-4 or IL-4 plus LPS; however, their experiments were performed with low concentration of It-4 (1 ng/ml), that, in our hands too, only triggered a slight cAMP increase, whereas higher IL-4 concentrations stimulated significant increases in both cyclic nucleotides. The IL-4 receptor binding unit and its y chainassociated component do not belong to the family of receptors with seven spanning domains, capable of interacting with Gst and leading to the activation of adenylate cyclase (AC). Further. more, whether the associated y chain has been detected in lymphoid cells, it appears to be absent in monocytic cells, as revealed by RT-PCR. 38'39 The late accumulation of c_AMP suggested therefore an indirect mechanism of action of IL-4. Activation of AC via the induction of prostaglandin by IL-4 was ruled out, since the It-4-driven cAMP accumulation was not significantly affected in the presence of indomethacin, an inhibitor of cyclooxygenase. Alternatively,-the observed IL-4-induced cAMP accumulation could result from AC stimulation through the activation of PKC or of the calcium/calmodulin complex. 4 The role of protein kinase C (PKC) and cyclic nucleotides in IL-4 signalling is controversial. Arruda and HO 4 presented evidence that, in human monocytes, IL-4 signalling involved PKC translocation to a nuclear fraction, without parallel increase in [Ca 2+ ]i, suggesting a possible role for phosphatidylcholine hydrolysis in this process; the same group has shown recently that IL-4 signalling in monocytes and U937 cells involves the activation of a phosphatidylcholine-specific phospholipase C (PC-PLC), but not of phosphatidylinositol 4,5-bisphosphate phospholipase C (PIP2-PEg), PEA2, PLD nor sphingomyelinase. 42 In contrast, our data clearly indicate that 1L-4 stimulates an increase in [Ca 2+ ]i in human promonocytic U937 cells, despite the usual considerable variation which was observed when the response of individual cells was measured. The fluorescent probe we used, Fluo-3 is slightly more convenient for detecting small variations in calcium concentration than Fura-2, used by Ho et al., but this is not sufficient to explain the reasons for this discrepancy. Alternatively there may exist differences between various sub-clones of U937 cells.
Our kinetic studies indicate that the late peak of c_AMP generation is preceded by an increase in cGMP. At variance with the results observed with rodent monocytes, where a 30-100 fold difference between cAMP and cGMP levels has been reported, 4 our data indicate that the basal concentration of cGMP is roughly of the same order of magnitude as that of cGMP, which is more in agreement with the results obtained for human monocytes by Li et a/. 44 The differential kinetics of the cGMP and cAMP increases suggested a possible link between the accumulation of these two cyclic nucleotides. Indeed, both cGMP and cAMP accumulation were suppressed by I-NMMA and nitro-t-arginine, two inhibitors of the NO synthase (NOS). Nitric oxide (NO), which has been identified in recent years as a pleiotropic mediator, 45-48 is a short-lived radical generated by oxidation of the guanidino group of I-arginine by constitutive or inducible NO synthases. [49][50][51][52] Soluble guanylate cyclase is activated by NO through interaction with its haeminic group, which leads to increased cGMP concentration. The subsequent cAMP increase we observed following addition of IL-4 to monocytes could then result from an interplay in the cyclic nucleotide phosphodiesterases (CN-PDE) network.
Noteworthy, Li et al. 44 54 The cGMP increase might activate the cGMPinhibitable cAMP-specific phosphodiesterase, (type III CN-PDE), that would result in subsequent cAMP accumulation, independent of the activation of adenylyl cyclase and might explain the delayed phase of cAMP increase. Experiments are in progress to study this mechanism and preliminary results suggest that both chemical NO donators and permeant cGMP analogues, such as Sp-cGMPS, are able to trigger cAMP accumulation in monocytic cells.
Both cGMP and cAMP increases were affected in the presence of EGTA, a calcium chelator, and of W-7 or calmidazolium, two inhibitors of the calcium/calmodulin complex. In contrast with the calcium/calmodulin-dependent activation of constitutive NOS, the activation of inducible forms of NOS (iNOS) is regulated by transcription; once translated, the protein tightly associates with calmodulin, explaining the relative insensitivity of iNOS to modifications of the intra-48 cellular calcium concentration. However, the iNOS from human hepatocytes, or Hep-NOS, is partially inhibited by calcium chelators and by antagonists of calmodulin; furthermore, natural and recombinant Hep-NOS might differ in their sensitivity to calcium and calmodulin. 52 The NO synthase activated by IL-4 in unstimulated monocytes may therefore be of the constitutive type, inasmuch as an endothelial type cNOS mRNA can be detected in these cells by RT-PCR. The presence of both inducible iNOS (type II) and constitutive endothelial NOS (type III) mRNA has been recently reported in human monocytes and monocytic cell lines, such as U937; 55 interestingly, stimulation with IFN-, and TNF-cz led to an induction of iNOS mRNA expression, whereas cNOS mRNA was reduced in the same time. A reciprocal regulation of cNOS and iNOS mRNA has already been observed in other cell types, suggesting the control of their expression by coordinated mechanisms. In our investigation, however, we did not detect a significan reduction of cNOS mRNA following the addition of IFN-2 alone. In addition, treatment with either IL-4, or IL-4 and anti-C23, did not modify the expression of cNOS mRNA in U937 cells. In contrast, these treatments were found to induce the expression of the inducible iNOS, both at the mRNA level, as detected by RT-PCR, and at the protein level, as revealed by immunochemistry (Dugas et al.; manuscript in preparation and Reference 56).
Preliminary immunofluorescence experiments, using a mouse monoclonal antibody against antihuman endothelial NOS (Affinity Research Products, Nottingham, UK) confirm the existence of an endothelial type cNOS protein in monocytic cells such as U937. Experiments are in progress to determine if this protein is enzymatically active and whether it is also present in human blood monocytes. Our results also indicate that a calcium increase, by itself, is not sufficient to trigger the activation of cNOS in monocytic cells. They suggest that IL-4 provides other signals or co-factors that are required to obtain a full catalytic activity. One such co-signal could be related to tyrosine phosphorylation, inasmuch as the catalytic activity of another type of constitutive NOS, the neuronal enzyme, is impaired in the presence of tyrosine kinase inhibitors. 57 Indeed, preliminary data suggest that IL-4-induced cGMP accumulation may be impaired by tyrosine kinase inhibitors.
These results indicate that different isoforms of NO synthase, constitutive and inducible, may be expressed in human monocytes, depending on their stage of differentiation and the combination of signals provided.