Increased Polymorphonuclear Leukocyte cGMP Levels Induced by the Human Lymphokine, Leukocyte Migration Inhibitory Factor (LIF) Klaus Bendtzen and Rene Klysner Abstract: The possible involvement in vitro of 3',5'-cyclic GMP (cGMP) in the mechanism of action of the lymphokine, leukocyte migration inhibitory factor (LIF), was investigated. Partially purified LIFrich supernatants, but not their control counterparts, induced a 2-fold increase in the cGMP levels of purified human polymorphonuclear (PMN) leukocytes. The effect was not influenced by heatinactivated horse serum; it was manifested within 3 rain of exposure to LIF and it subsided within 180 rain. LIF and the supematant factor responsible for the cGMP-generating effect were both rendered inactive by treatment with the serine esteraseand protease inhibitor, phenylmethylsulfonyl fluoride, indicating that these factors are closely related, if not identical. A potent phosphodiesterase inhibitor, dipyridamole (2 × 10-4 M), induced a 3- to 5-fold increase in PMN leukocyte cGMP levels, but combined treatment with purified LIF and dipyridamole did not add to this effect. This suggeststhat the cGMP-generalingfactor acts on the biochemicalpathway that degrades cGMP. Key Words: Polymorphonuclear leukocyte; Lymphokine; Leukocyte migration inhibitory factor;, cGMP
INTRODUCTION 3',5'-cyclic AMP (cAMP) and 3',5'-cyclic GMP (cGMP) are known to influence the functions of lyrnphocytes and other effector cells involved in the immune response (Bourne et al., 1974; Watson, 1975; Lornnitzer et aL, 1976). Thus, antigen- and mitogen-induced lymphocyte transformation and lymphokine production are inhibited by increased levels of cAMP; neutrophil functions such as phagocytosis, release of lysosomal enzymes, chemotactic responsiveness and spontaneous mobility are inhibited by increased levels of cAMP (Zurier et al., 1974; Rivkin et al., 1975), and chernotactic responsiveness and lysosornal enzyme release are stimulated by increased levels of cGMP (Estensen et al., 1973; Zurier et al., 1974). The possible involvement of cyclic nucleotides in the impaired cell migration induced by the lyrnphokine, leukocyte migration inhibitory factor (LIF), has been suggested recently (Lomnitzer et al., 1976; Bendtzen and Palit, 1977). Human LIF is a serine esterase and protease (Bendtzen, 1977a), and the polyrnorphonuclear (PMN) leukocyte response to this lymphokine is markedly Received November 2, 1978; accepted February 26, 1979. From the Laboratory of Clinical Immunology, Medical Department TA, RJgshospitaletUniversity Hospital and Pharmacological Institute, University of Copenhagen, Copenhagen, Denmark. Address requests for reprints to: Dr. Klaus Bendtzen, Tufts-New England Medical Center, Division of Allergy, 136 Harrison Avenue, Boston, MA 02111. © ElsevierNorthHolland,Inc.,1979 [mmunopharmacolo~/1.323-330(1979)
K. Bendtzen and R. Klysner
depressed by cAMP and by agents known to increase intraceUular cAMP concentrations (Bendtzen and Palit, 1977). Since LIF by itself has no significant effect on the leukocyte cAMP levels, a role for cGMP in LIF-induced migration inhibition seems possible (Bendtzen et al., 1977). This is further stressed by the results of recent experiments indicating that the enzymatic activity of this lymphokine is subject to specific regulatory control by cGMP (Bendtzen, 1977d; 1979a,b). This article will document that partially purified, LIF-rich supernatant fluids induce a transient but significant increase in PMN leukocyte cGMP levels, and present indirect evidence indicating that a serine esterase, possibly LIF, is responsible for this effect. MATERIALS AND METHODS
Production and Partial Purification of Leukocyte Migration Inhibitory Factor The methods previously described were used (Bendtzen, 1979a). In brief, 2.5 × 10~ mononuclear celis/ml in serum-free medium TC-199, containing 25 mM Hepes buffer, 500 IU/ml penicillin, 500 p~g/ml streptomycin, and 20 U/ml nystatin (TC-199), were incubated for 22 hr with 50 p.g/ml of concanavalin A (Con A). The supematants were harvested, desalted on a Sephadex G-50 column, and lyophilized. Pooled, 100-fold concentrated supematant fluids, pretreated with 1 mM tosyl tysine chloromethyl ketone to inhibit histidine esterases, were then chromatographed on a Sephadex G-100 column. The LIF-rich fractions were pooled and mixed in a column with agarose-conjugated rabbit immunoglobulins against protein contaminants of LlF-rich supematants. The LIF-rich eluates were lyophilized and stored at -20°C until use. Controls obtained from supernatants of unstimulated mononuc]ear cells were reconstituted with Con A and processed in parallel. The concentration of serine esterases in the purified material was approximately 10 T M M, as judged by affinity labeling with tritiated diisopropylfluorophosphate (Bendtzen, 1979a).
Bioassay of Leukocyte Migration Inhibitory Factor Activity A slightly modified indirect leukocyte migration agarose technique, originally described by Clausen (1972), was employed, using PMN leukocytes from normal donors as indicator cells. These were obtained by dextran sedimentation of heparinized venous blood and subsequent centrifugation of the buffy coat cells through an Isopaque-Ficoll gradient (BOyum, 1976). No attempt was made to remove contaminating erythrocytes, since these cells did not influence cell migration. Samples (10 p.I), containing 1.5 × 10~ leukocytes, 95-98% of which were viable PMN cells, were tested in quadruplicate for migration under agarose and a migration index MI was determined by the formula: MI --
mean migration area of cells in LIF-rich supematant mean migration area of cells in control supernatant
Preparation of Human Polymorphonuclear Leukocytes Polymorphonuclear leukocytes for cGMP determination were obtained by dextran sedimentation and gradient separation as outlined above. The leukocyte-rich cell pellet consisted of 94% neutrophils, 2% eosinophils, 3% mononuclear cells, and 0 - 1 % basophils. Removal of contaminating erythrocytes was not considered necessary, since identical results were obtained in separate experiments on PMN leukocytes depleted of red cells. Abbreviations: Con A: concanavalin A; cAMP: 3',5'-cyclic AMP; cGMP: 3',5'-cyclic GMP; LIF: leukocyte migration inhibitory factor; Mh migration index; PMN: polymorphonuclear;
PMSF: phenylmethylsulfonyl fluoride.
Increased PMN Leukocyte cGMP by LIF
Effect of Leukocyte Migration Inhibitory Factor on Polymorphonuclear cGMP Levels LyophiIized, partially purified LlF-rich and control supernatant materials were dissolved at 37°C in TC-199, which, unless noted otherwise, contained 10% heat-inactivated horse serum. Then, 107 washed PMN leukocytes, containing approximately an equal number of erythrocytes, were resuspended in 90/~1 of these supernatant fluids and incubated at 37°C in a humidified 5% CO2 air atmosphere. As a further control, ceils were incubated in parallel in TC-199 alone. After various times, the cultures were terminated by the addition of 100/zl boiling Tris-EDTA buffer (50 mM Tris, 5 mM EDTA; pH 7.4), immediately followed by 3 rain in a boiling water bath. In sealed tubes, this heat treatment caused less than 5% loss of sample volume clue to evaporation. After centrifugation for 20 min at 1000 g, the supernatants were removed and stored at -20°C before cGMP assay.
Assay for cGMP The cGMP levels of combined cells plus media were assayed with a radioimmunoassay kit obtained from The Radiochemical Centre, Amersham, U.K. In most experiments, 0.067 pmol 3H-cGMP was added to each test tube, i.e., only one-sixth of that prescribed by the manufacturer. In order to achieve a zero-dose binding of 40-50%, the antiserum was diluted to half the concentration recommended. One hundred microliters of sample or standard were incubated with 50 Izl of the above-mentioned amount of 3H-cGMP and anUserum for 90 rain at 4°C. The reaction was stopped by adding 1 ml of 60% saturated ammonium sulfate to the tubes. After 10 rain, the tubes were centrifuged for 20 rain at 2800 g at 4°C and the supernatants were decanted. The precipitates were redissolved and counted for 10 min in a scintillation counter after mixing with 100/A of water and 1.2 ml of Insta-Gel (Packard, Downers Grove, Ill.). Using this procedure, the detection limit was 0.03 pmol (2 standard deviations at zero dose). Calibration curves were plotted each day of assay using known amounts of cGMP to inhibit the binding of 3H-cGMP. An almost linear mid-portion of the curve was always obtained, and the cGMP concentration range tested fell within this area. In one experiment, the standard curve and a dilution profile of a measured PMN leukocyte extract were superimposable. 2',3' cGMP, 5'GTP, 5'GMP, 3'GMP, 2',3'cAMP, and 5'ATP (20 nmol) produced no significant displacement of 3H-cGMP binding. A similar amount of 3',5'cAMP and 5'GDP produced an interference corresponding to that obtained with 1 pmol of 3',5'cGMP. Since the nucleotides were not purified prior to assay and the antiserum used for the interference study had been repeatedly thawn and frozen, these interferences can be regarded as maximal interference in the binding assasy. Moreover, the concentration of 3',5'cAMP and 5'GDP used in this study most probably greatly exceeded the concentrations present in the PMN leukocyte extracts (Bendtzen et al., 1977). Bovine serum albumin and human gamma globulin (50/zg) did not interfere with 3H-cGMP binding. The recovery of cGMP added to the cell suspensions was 72 - 6% (mean ± 1 SD; n = 6), indicating either trapping of a portion of the cGMP or the presence of a factor that interfered with the binding of antibody to cGMP. However, since the variation in cGMP recovery was small and the recovery was the same whether the cells were treated for 10 rain at 37°C with LIF or control fractions, individual recovery experiments were not carried out. The presence of erythrocytes did not influence the recovery of cGMP. Deviations between triplicate determinations of leukocyte cGMP were less than 7% and values reported for cGMP are not corrected for the decreased recovery. RESULTS Leukocyte migration inhibiting factor appears to be a serine esterase and protease by virtue of its susceptibility to the irreversible enzyme inhibitor, phenylmethyLsulfony] fluoride (PMSF), and by the specific ability of arginine esters and amides to protect LIF against PMSF-induced inactivation
K. Bendtzen and R. Klysner
(Bendtzen, 1977a-c). Recently, competition experiments have shown that cGMP, but not other mononucleotides, also preserves LIF activity in the presence of PMSF, suggesting a specific LIF affinity for cGMP (Bendlzen, 1977d). In view of the latter finding, initial experiments were carried out to investigate the possibility that LIF by itself possessed cGMP phosphodiesterase activity. Purified, 10-fold concentrated LIF-rich and control preparations were incubated with 3 × 10 -9 M cGMP (0.3 pmol cGMP in 100 ~I TC-199) at 37°C for various times up to 6 hr. The reactions were terminated by addition of boiling Tris-EDTA buffer and by further boiling for 3 min. Neither of the supernatants affected the cGMP concentrations, indicating the absence of any appreciable amount of cGMP phosphodiesterase activity (data not shown). Experiments were next carried out to determine the effect of LIF on the cGMP levels of human PMN leukocytes. As shown in Figure 1, LIF-rich solutions induced a transient but significant 2-fold increase in the cGMP concentrations of leukocytes, whereas control fluids and medium alone (data not shown) had no influence on the levels of the nucleotide. The presence of LIF activity was ascertained by bioassay (MI = 0.66 _+ 0.04; n = 4). Since the bioassay of LIF activity is always carried out in the presence of serum which promotes viability and migration of neutrophils in vitro, all fluids were supplemented with horse serum. The mechanism by which LIF induces inhibition of PMN leukocyte mobility is unknown, however, and the possibility that LIF exerts its effect indirectly, for instance, by the activation of a serum component, which in turn causes migration inhibition, has never been fully explored. It was therefore of interest to determine whether serum was necessary for the observed effect on PMN leukocyte cGMP levels, or, if not, whether serum might potentiate this effect. As shown in Table I, such effects were not seen, even though the experiments were carried out at a low concentration of LIF to facilitate detection of a possible effect of serum. Bioassay of the 2-fold concentrated supernatant fluids confirmed the presence of LIF activity (MI: 0.77 ± 0.04; n = 4). The lymphocyte supematants did not by themselves contain measurable amounts of cGMP, since 50 p~l, 10-fold concentrated LIF and control fractions, each containing 10% serum, contained less than 0.03 pmol cGMP; the detection limit of the cGMP assay. To determine whether the active principle in partially purified LIF-rich preparations, like LIF itself, was a serine esterase, LIF-rich and control preparations were incubated at 37°C for 30 min with I mM PMSF. Unreacted PMSF and low molecular weight reaction products were then reFigure I Effectsof purified, 5-fold concentrated, LIF-rich ( I ....... 0) and control ( 0 - - - 0) supernatant.fluids on PMN leukocyte cGMP levels. The initial concentration of cGMP was 0.29 ± 0.05 pmol/lO 7leukocytes; n = 8. The number of experiments and standard deviations are s h o w n . Except for the 22-hr incubations, 4 different batches of supernatants were used. The LIF-induced increases in cellular cGMP levels were statistically significant after 3, 10, and 30 rain of incubation (p < 0.05; Mann-Whitney rank sum test).
Increased PMN Leukocyte cGMP by LIF
Lack of effect of heat-inactivated horse serum on the LIF-induced increase in PMN leukocyte cGMP levels"
Deviation from initial cGMP level
Medium alone Medium + LIF Medium + LIF + horse serum
10 10 10
1.1 _ 0.3 2.1 _+ 1.0 2.2 _+ 1.1
Medium alone Medium + LIF Medium + LIF + horse serum
30 30 30
0.9 _+ 0.3 2.0 _+ 0.7 1.9 _ 0.5
"The initial amount of cGMP was 0.30 _ 0.05 pmol/107 leukocytes; n = 8. Results are means _+ 1 SD of duplicate determinations using 4 different LIF-rich materials, each 2-fold concentrated.
moved by dialysis at 4°C for 72 hr. This treatment abolished LIF activity (MI: 0.87 -+ 0.06; n = 4), whereas LIF-rich solutions processed in parallel but in the absence of PMSF still possessed significant I.IF activity (MI: 0.71 -+ 0.05; n = 4). As shown in Figure 2, PMSF-treated LIF completely lost its ability to increase PMN leukocyte cGMP levels. This indicates not only that the factor responsible for the observed effect on leukocyte cGMP is a serine esterase and/or a serine protease, but also that the enzyme in order to exert this effect must be present in its active form. Unfortunately, attempts to specify this enzyme in more detail by using neutralizing antibodies against LIF were unsuccessful, since the purified rabbit immunoglobulins by themselves induced a 4-fold increase in PMN leukocyte cGMP concentrations. Another approach, to block the hydrolytic activity of LIF on its presumed natural substrate on the PMN leukocyte by culturing
Figure 2 Effect of PMSF on the factor(s) responsible forthe increased levels of PMN leukocyte cGMP. Purified, 5-fold concentrated, LIF-rich (left) and control supematant fluids (right) were preincubated with 1 mM PMSF, extensively dialyzed and tested for cGMP generating effect ( ll). Supernatants processed in parallel but without PMSF were also tested ( D). The initial concentration of cGMP was 0.30 +. 0.06 pmol/10' /eukocytes; n = 8. Results are means +-- 1 SD of duplicate determinations using 4 different batches of supernatants. Only nontreated, LIF-rich, supernatant fluids were able to induce a statistically significant increase in cGMP levels (p < 0.05; Mann-Whitney rank sum test). d -~
d ¢J t--
INCUBATION TIME (MIN)
K. Bendtzen and R. Klysner
cells exposed to LIF in the presence of a high affinity synthetic substrate for LIF also met with difficulties, and highly variable results were obtained. Finally, in an attempt to determine the point of action of the cGMP-inducing enzyme, the effect of the phosphodiesterase inhibitor, dipyridamole, was investigated. While 2 × 10 -4 M dipyridamole by itself induced a 3- to 5-fold increase in PMN leukocyte cGMP levels after 10 to 30 rain of incubation, combined treatment of cells with LIF and the inhibitor did not potentiate this effect (Table 2). In these experiments, two different batches of purified, 5-fold concentrated LIF and control supernatant fluids were used, and bioassay confirmed the presence of LIF activity (MI: 0.63 and 0.69, respectively). This implies an action of the cGMP-generating enzyme on the cellular reactions that degrade cGMP rather than on the biochemical pathway that produces cGMP. DISCUSSION From the results of competition experiments in which LIF was coincubated with phosphate esters
and the inhibitor PMSF and from the results of direct enzyme assays using a radioactive substrate, it has recently become clear that LIF is subject to selective regulatory control by cGMP (Bendtzen, 1977d, 1979a,b). These findings and the fact that PMN leukocytes treated with cAMP-elevating agents, such as isoproterenol, papaverine, and cAMP itself, escape LIF-induced migration inhibition (Bendtzen and Palit, 1977), prompted investigations of the PMN leukocyte cyclic nucleotide levels before and during treatment with LIF. In a previous communication, LIF was reported not to influence the cAMP levels of PMN leukocytes (Bendtzen et al., 1977). In this article, a significant, 2-fold increase in the cGMP concentrations is reported. The elevated cGMP levels were observed when PMN leukocytes, predominantly neutrophils, were cultured in supernatants obtained from Con A-stimulated lymphocytes but not when cultured in supernatants of unstimulated lymphocytes or in medium alone. The active supematants were partially purified with respect to LIF. They did not by themselves contain appreciable amounts of cGMP and their effect on PMN leukocyte cGMP was not influenced by the presence or absence of serum. The finding that PMSF-inactivated supernat~ints, formerly rich in active LIF molecules, completely lost their ability to increase neutrophil cGMP levels provides suggestive evidence that the LIF molecule was in fact responsible for this effect. Table 2
Effect of the phosphodiesterase inhibitor, dipyridamole, on the amount of cGMP in PMN leukocytes treated with purified I_IF-rich and control supematant fluids" pmol cGMP/IO 7 leukocytes Exp. no 1
Exp. no 2
Incubation time (rain)
Incubation time (rain)
Medium alone Medium plus 2 x 10 .4 M dipyridamole
Control alone Control plus 2 x 10 .4 M dipyridamole
LIF alone LIF plus 2 x 10-4 M dipyridamole
a Results are means of duplicate cGMP determinations.
Increased PMN Leukocyte cGMP by LIF
The natural substrate for LIF is unknown as is the biological significance of the previously reported influence of cGMP on the molecular biology of LIF. In fact, the exact biological role of LIF is uncertain, since the lymphokine may have effects on neutrophiis other than those involved strictly in nondirectional cell migration. In view of the findings reported here, cGMP might possibly be an intracellular mediator of such a variety of effects. Thus, LIF might initiate a series of consecutive biochemical reactions leading to an increase in cellular cGMP levels, probably by reducing the rate of degradation of cGMP (Table 2). Such a model would also provide an attractive mechanism by which LIF activity might be regulated. Many similar examples of end-product regulatory systems are known in biology (Stadtman, 1970). If LIF is responsible for the elevation of PMN leukocyte cGMP levels, and if cGMP is a second mediator of LIF activity, addition of cGMP- or cGMP-generating agents to PMN leukocytes should induce migration inhibition. However, cGMP and its dibutyryl derivative, at concentrations up to 10 -s M, do not appear to affect cell migration under agarose (Bendtzen and Palit, 1977). Moreover, even agents such as papaverine and dipyridamole, which raise the cellular levels of cGMP as well as cAMP cause enhancement of PMN leukocyte migration (Bendtzen and Palit, 1977). The reason for this is entirely unknown, although the cyclic nucleotide-generating agents may have effects on neutrophils other than those involved in cyclic nucleotide metabolism. One of many possible explanations would be that an LIF-induced increase in PMN leukocyte cGMP concentration is not directly responsible for the impairment of cellular mobility. This, of course, would not rule out a possible role of cGMP in other leukocyte functions that might be regulated by LIF. Thus, definite proof that LIF influences PMN leukocyte cGMP levels and the possible biological implications of this effect will have to await further progress in purification and specific biochemical and immunochemical assay of this lymphocyte mediator. The authors wish to acknowledge the technical assistance of Ms. Lidy Broersma and the secretarial assistance of Ms. Joan Connolly. This work was supported by grants from the Danish Medical Research Council.
Bendtzen K (1977a) Human leukocyte migration inhibitory factor (LIF). I. Effect of synthetic and naturally occurring esterase and protease inhibitors. Scand J Immunol 6:125. Bendtzen K (1977b) Human leukocyte migration inhibitory factor (LIF). II. Partial biochemical characterization of the substrate specificities for this lymphokine. Scand J lmmunol 6:133. Bendtzen K (1977c) Human leukocyte migration inhibitory factor (LIF). M. Further investigations on the serine protease nature of this lymphokine and its preference for arginine amides. Scand J Immunol 6:1055. Bendtzen K (1977d) Human leukocyte migration inhibitory factor (LIF). IV. 3',5'-cGMP protects LIF against inactivation by the esterase inhibitor phenylmethylsulfony] fluoride. Scand J lmmunol 6:1357. Bendtzen K (1979a) Determination of the human lymphokine leukocyte migration inhibitory factor (LIF) by a sensitive radioenzymatic assay. Inhibitory effect of cGMP on the esterolytic activity of highly purified LIF. J Clin Lab lmmunol 2:37. Bendtzen K (1979b) Leukocyte migration inhibitory factor. A serine esterase released by stimulated human lymphocytes. Kinetic analysis and inhibition by cyclic GMP. Biochim Biophys Acta 566:183. Bendtzen K, Palit J (1977) Modulation of human leucocyte migration inhibitory factor (LIF) by 3',5'-cyclic AMP, 3',5'-cyclic GMP and agents known to influence intracellular cyclic nucleotide metabolism. Acta Pathol Microbiol Scand [C]85:317. Bendtzen K, Thode J, Nistrup Madsen S (1977) Effect of human leucocyte migration inhibitory factor (LIF) on 3',5'-cyclic AMP levels of peripheral blood leucocytes. Acta Patho! Microbiol Scand [C]85:473.
K. Bendtzen and R. Klysner
Bourne, HR, Lichtenstein LM, Melmon KL, Henney CS, Weinstein Y, Shearer GM (1974) Modulation of inflammation and immunity by cyclic AMP. Science 184:19. BCyum A (1976) Isolation of lymphocytes, granulocytes and macrophages. Scand J Immunol Suppl 5:9. Clausen JE (1972) Migration inhibitory effect of cell-free supernatants from mixed human lymphocyte cultures. J Immunol 108:453. Estensen RD, Hill HR, Quie PG, Hogan N, Goldberg ND (1973) Cyclic GMP and cell movement. Nature 245:458. Lomnitzer R, Rabson AR, Koornhof HJ (1976) The effects of cyclic AMP on leucocyte inhibitory factor (LIF) production and on the inhibition of leucocyte migration. Clin Exp lm muno124:42. Rivkin I, Rosenblatt J, Becker EL (1975) The role of cyclic AMP in the chemotactic responsiveness and spontaneous motility of rabbit peritoneal neutrophils. J lmmunol 115:1126. Stadtman ER (1970) Mechanisms of enzyme regulation in metabolism. Enzymes 1:397. Watson J (1975) Cyclic nucleotides as intracellular mediators of B cell activation. Transplant Rev 23:223. Zurier RB, Weissmann G, Hoffstein S, Kammerman S, Tai, HH (1974) Mechanisms of lysosomal enzyme release from human leucocytes. II. Effects of cAMP and cGMP, autonomic agonists, and agents which affect microtubule function. J Clin Invest 53:297.