Enhancement of macrophage superoxide anion production by ...

Vol. 35, No. 5

ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, May 1991, p. 796-800

0066-4804/91/050796-05$02.00/0 Copyright C 1991, American Society for Microbiology

Enhancement of Macrophage Superoxide Anion Production by Amphotericin B ELIZABETH WILSON,"2 LISA THORSON,"2 AND DAVID P. SPEERTl.2.3* Division of Infectious Diseases, British Columbia's Children's Hospital,' and Departments of Pediatrics2 and Microbiology,,3 University of British Columbia, Vancouver, British Columbia V5Z 4H4, Canada Received 31 October 1990/Accepted 1 February 1991

Amphotericin B (AmB) appears to have some important immunomodulatory effects, but its mechanism of action has not been explained. We investigated the effects of AmB on activation of human monocyte-derived macrophages. Macrophages cultured in the presence of AmB had an enhanced capacity to produce superoxide anion after stimulation with phorbol myristate acetate. This enhancement was dose dependent within a therapeutic range of AmB levels (0.1 to 3.0 mg/liter). Macrophages cultured in the presence of AmB had enhanced surface expression of Ia antigen; phagocytosis of unopsonized zymosan, opsonized Staphylococcus aureus, or erythrocytes opsonized with C3bi or immunoglobulin G paradoxically appeared to be reduced, but results did not achieve statistical significance. AmB appears to activate macrophages and may do so via direct effects on the plasma membrane.

gift from Genentech Corp. (South San Francisco, Calif.), was stored at -20°C. It was thawed and diluted in RPMI 1640 for use. Macrophage culture. Human monocyte-derived macrophages were prepared as described previously (19). Fresh, heparinized venous blood from healthy adult volunteers was mixed with an equal volume of isotonic saline and centrifuged on a cushion of Ficoll-Hypaque (Pharmacia Canada, Inc., Dorval, Quebec, Canada). The mononuclear cell layer was washed three times with RPMI 1640 supplemented with penicillin (100 U/ml) and streptomycin (100 jxg/ml) (RPMIPS). The cells (106/ml) were cultured with RPMI-PS and 15% autologous serum in Teflon beakers (Savillex, Minnetonka, Minn.) for 5 days at 37°C in a humidified atmosphere with 5% carbon dioxide. For some experiments, a relatively pure monocyte preparation was obtained by further centrifugation of the mononuclear cell layer on preformed Percoll gradients by a modification of the method of Wright and Silverstein (28). Percoll (Pharmacia) was first brought to isotonicity by diluting it 1:10 in Hanks balanced salt solution (HBSS; GIBCO). A total of 22 ml of this solution, 1 ml of autologous serum and 14.7 ml of HBSS were distributed to each of two 40-ml Oak Ridge-type centrifuge tubes (Nalge Co., Rochester, N.Y.) and centrifuged at 30,000 x g for 15 min at 4°C. Mononuclear cells (2 x 107 to 5 x 107/ml; maximum, 5 ml) were layered onto the Percoll gradient and centrifuged at 1,000 x g for 20 min at 4°C. The upper cell band was aspirated and washed twice in RPMI-PS. The cells were cultured in Teflon beakers as described above. Cell viability was determined by trypan blue exclusion and always exceeded 95%. This method yielded mononuclear cells, of which 90 to 95% were monocytes, as determined by morphology on cytocentrifuge preparations and by nonspecific esterase staining. AmB (in the form of Fungizone or a pure preparation), sodium deoxycholate, or gamma interferon was added to some cultures after 72 h, and the macrophages were washed free of it before each experiment. Macrophages were harvested after 5 days in culture and washed three times in ice-cold RPMI 1640 with 10 mM HEPES (N-2-hydroxyeth-

Activation of macrophages is important in host defense against intracellular pathogens (8), and a biochemical correlate of activation is the production of reactive oxygen metabolites such as superoxide anion (02-) (1). Amphotericin B (AmB), the mainstay of systemic antifungal therapy for many years (14), has known immunomodulatory effects: Medoff et al. (13) noted that some patients with systemic candidiasis improved with doses of AmB too low to result in candidacidal levels in serum. Furthermore, mice pretreated with AmB show increased survival after experimental infection with Listeria monocytogenes (23), an intracellular parasite against which AmB has no antibacterial activity. Macrophages to which AmB is added after treatment with interferon demonstrate enhanced tumor cell lysis (17), another recognized correlate of macrophage activation. AmB is also an oxidant (5); lethal effects on fungi may involve oxidative damage (18). To understand better the immunomodulatory activities of AmB, we investigated its effect on the capacity of human macrophages to mount a respiratory burst. We observed that macrophages cultured in the presence of AmB were activated; they had an enhanced capacity to produce superoxide anion and showed increased expression of Ia antigen. These observations and a description of the phagocytic capacity of AmB-treated macrophages form the basis of this report. MATERIALS AND METHODS Reagents. AmB was reconstituted from a lyophilized formulation (50 mg of AmB and 41 mg of sodium deoxycholate; Fungizone; Squibb, Montreal, Quebec, Canada) in 10 ml of sterile water. Further dilutions were made in RPMI 1640 (GIBCO Laboratories, Grand Island, N.Y.). A fresh vial of AmB was reconstituted for each experiment. Sodium deoxycholate (Sigma) was diluted in an analogous manner and was added at a concentration equivalent to that in 3 mg of diluted Fungizone per liter. Pure AmB (Sigma) is highly insoluble in water. It was therefore dissolved in dimethyl sulfoxide (22) at 5 mg/ml, diluted in 5% glucose, adjusted to pH 9.0, and filter sterilized. Recombinant human gamma interferon, a * Corresponding author. 796

VOL. 35, 1991

ylpiperazine-N'-2-ethanesulfonic acid; BDH Chemicals, Poole, England) and 0.2% bovine serum albumin (Sigma). Assessment of superoxide anion production. Superoxide dismutase-inhibitable reduction of ferricytochrome c (26) was measured. A total of 4 x 105 to 1 x 106 macrophages was suspended in HBSS with 0.1% gelatin (GHBSS). To the cells were added 0.1 mM ferricytochrome c (type IV; Sigma), 500 U of catalase (Sigma), and 100 ng of phorbol myristate acetate (Sigma), to a total volume of 1 ml in GHBSS. Negative controls also containing 0.1 mg of superoxide dismutase (Sigma) were run in parallel. Reaction mixtures were incubated for 20 min at 37°C in a shaking water bath, transferred to ice, and centrifuged at 300 x g for 10 min at 4°C. Supernatants were promptly removed to iced tubes, diluted 1 in 4 with GHBSS, and read on a Bausch and Lomb Spectronic 2000 scanning spectrophotometer (MiltonRoy, Chicago, Ill.) from 560 to 530 nm. 02 production was calculated from the change in optical density by using an extinction coefficient for cytochrome c of 21.1 mM-' (24). Assessment of phagocytosis. Particles for phagocytosis included (i) zymosan (Sigma), a 0.1% suspension prepared exactly as described previously (19); (ii) erythrocytes opsonized with immunoglobulin G (EIgG), which were prepared as described previously (28), within 48 h of use; (iii) erythrocytes opsonized with C3bi (EC3bi) as described previously (28) or by incubation sequentially with anti-sheep erythrocyte IgM (Cappel) and C5-deficient serum (Sigma) and used within 2 weeks of preparation; and (iv) Staphylococcus aureus 502A. The bacteria were grown overnight from a frozen stock in Mueller-Hinton broth (Difco, Detroit, Mich.), washed in phosphate-buffered saline (PBS; pH 7.4), and adjusted to 109 CFU/ml spectrophotometrically. The bacteria were opsonized by gentle agitation in 10% pooled normal human serum for 20 min at 37°C, washed twice in PBS, and resuspended in RPMI 1640. Macrophages were plated onto glass coverslips on the day of each experiment as described previously (19). Briefly, 100-pul aliquots of harvested macrophages at 105 cells per ml were placed onto circular glass coverslips and were incubated for 1 h at 37°C in a humidified 5% CO2 atmosphere to permit macrophage adherence. The coverslips were washed in PBS and transferred to 24-well plastic trays containing 0.9 ml of RPMI 1640 per well. Particles were added to the macrophages as described previously (20) and incubated for 60 min at 37°C. The coverslips were then washed with PBS, fixed in glutaraldehyde, and mounted onto glass slides. Phagocytosis or binding was assessed as described previously (20). To distinguish EIgG binding from ingestion, one set of EIgG-treated coverslips was rinsed with distilled water prior to fixing in order to lyse uningested erythrocytes. Assessment of la antigen expression. (i) Flow cytometry. Macrophages, which were cultured from Percoll gradientprepared monocytes (minimum, i0' per test), were washed in PBS containing 1% fetal bovine serum (Flow Laboratories, McLean, Va.), incubated in an ice bath with isothiocyanate-labeled T3/13 (Coulter, Hialeah, Fla.), washed with PBS-fetal bovine serum, and fixed in 1% paraformaldehyde. Fluorescence per cell was measured by using an Epics C Flow Cytometer (Coulter), using standard gating parameters.

(ii) PCFIA. Cells were stained as described above, and fluorescence was assessed by a solid-phase particle concentration fluorescence immunoassay (PCFIA) by using a Pandex Screen Machine (Pandex Laboratories, Baxter Healthcare Corp., Mundelein, Ill.) as described previously (2). Approximately 104 cells were added to individual wells of

x20

. .~ . ._

MACROPHAGE ACTIVATION BY AMPHOTERICIN B 25

o

0' 0 CE

15

....

797

......... ........................

_.........

ZX..............

.........

..

................

> ^...........................

20.

00.

E i

E

.......

..............

Control

Amphotericin FIG. 1. Superoxide anion production by paired samples 'of macrophages cultured for 48 h in the presence (3 mg/liter) or absence (control) of AmB. Results are from 10 separate experiments on different days and are expressed as nanomoles of superoxide anion produced in 20 min by 10' macrophages. fluoricon filter assay plates, washed by flow filtration, and read by front-surface fluorometry. All determinations were done in duplicate, and results were recorded in relative fluorescence units. RESULTS Inspection of harvested cells showed no difference in size or morphology of the cells treated with AmB for 48 h compared with the control cells. At the range of AmB doses used (0.1 to 3 mg/liter), there was no evidence of cellular toxicity manifested either by reduction in harvested cell numbers or failure to exclude trypan blue. Furthermore, the ability of macrophages to produce superoxide anion confirmed their viability. Superoxide anion production. Incubation of human monocyte-derived macrophages with 3 mg of AmB per liter for 48 h prior to harvesting and washing resulted in an enhanced capacity for 02- production following stimulation with phorbol myristate acetate (control cells, 4.36 ± 4.28 nmol of 02per 1 cells; AmB-treated cells, 13.5 +8.28 nmol of 02- e 106 cells; P < 0.05 by Student's t test). Figure 1 shows the results from 10 separate experiments. In each case, the 02produced by the AmB-treated macrophages exceeded that of the control cells. Incubation with sodium deoxycholate at a dose equivalent to that in 3 mg of Fungizone per liter did not result in a significant enhancement Of 02- production (controls, 1.2 and 2.5 nmol of 02- per 106 cells; sodium deoxycholate, 1.9 and 3.7 nmol of 02- per 106 cells), while pure AmB gave results as high as those achieved with the commercial preparation (controls, 1.1 and 15.1 nmol Of 02per 106 cells; 3 mg of pure AmB per liter, 32.9 and 27.9 nmol of 02- per 106 cells; 3 mg of commiercial AmB per liter, 5.7 and 25.3 nmol of 02- per 106 cells) (results are for two different donors on different days). On no occasion was there significant 02- production in the presence of superoxide dismutase. AmB did not directly stimulate 02- production in the absence of phorbol myristate acetate (data not shown). AmB pretreatment of macrophages enhanced 02- in a do se-dependent fashion between 0. 1 and 3.0 mg/liter (Fig.- 2) . la kntigen expression. la antigen density was assessed by both flow cvtometry and PCFIA. Figure 3 shows an example

WILSON ET AL.

798

ANTIMICROB. AGENTS CHEMOTHER.

10

TABLE 1. Effect of AmB or gamma interferon on Ia antigen expression and superoxide anion production by macrophages

.-l

la to

8

U

O3

la

Condition

a

Superoxide production (nmol Of 02- produced/ 106 macrophages)

14

Whole MNCb Monocytes' Whole MNC Monocytes

14 0

Ps d

6

-

Control

b 0 14

Ia antigen expressiona

AmB (3 mg/liter) Gamma interferon (10 U/ml)

'a

d0

607 1,122 2,423

3.4 10.0 49.9

1.7 11.8 71.8

a Relative fluorescence units. Mean of duplicate determinations. b See text for details on leukocyte preparation. MNC, Mononuclear cells.

2 1.

O _

0.0

906 2,294 2,376

0.5

2.0 2.5 1.0 1.5 Amphoteriain B (mg/i)

3.0

FIG. 2. Superoxide production by macrophages cultured for 48 h in the presence of different concentrations of AmB. Results are from two typical experiments on different days.

of results obtained by fluorescence-activated cell sorting by using Percoll gradient-separated, monocyte-derived macrophages. The AmB-treated macrophages had a greater amount of surface-expressed Ia antigen, as demonstrated by a shift of the curve to the right and increases in both mean and peak channel fluorescence (control, mean fluorescence = 139 and 129; peak channel = 82 and 80; for AmB, mean fluorescence = 171 and 161; peak channel = 255 and 255 [data are from separate experiments on different days]; P
20 6-20 1-5 0

4.6 25.5

26.1 26.3

33.6 30.0

35.7 18.1

a Values are means of results from four experiments performed on different days.

culture 48 h prior to harvest, because this is the usual dosage interval for therapeutic administration. Research into the immunomodulatory effects of AmB has been extensive, but results have been conflicting, both between and within the various types of leukocytes studied. AmB enhances murine macrophage tumor cell killing in vitro (17) and improved survival of mice infected with L. monocytogenes (23). In contrast to our findings, one group showed an improved phagocytic capacity of murine peritoneal macrophages for polystyrene beads in AmB-treated mice (11). In neutrophil studies, two groups found that AmB suppressed chemotaxis and chemiluminescence by human neutrophils in vitro (3, 12), while another group (22) showed stimulation of human and canine neutrophil chemiluminescence and oxygen uptake by high doses of commercial AmB, but they found no enhancement of superoxide anion production. We found it difficult to assess the effects of AmB on human neutrophils because the drug is toxic to these cells in the absence of serum (7), the condition under which our experiments were performed. AmB also interferes with the measurement of superoxide anion if AmB is left in the test mixture, presumably by competing with cytochrome c for reduction. All of our experiments were performed with macrophages which were washed free of AmB. Macrophages have a low constitutive capacity for superoxide anion production (10), but exposure to AmB resulted in approximately a threefold enhancement. AmB effects on superoxide anion production were biphasic (Fig. 2), with enhancement at low and high, but not at intermediate, concentrations. This phenomenon may have been due to the fact that hydrogen peroxide can inhibit superoxide production (16). Since AmB induces superoxide and hydrogen peroxide release, net superoxide measured may have been determined by both stimulatory and inhibitory influences. Enhanced superoxide production was seen whether monocytes were cultured in the presence of large numbers of lymphocytes (whole mononuclear cell fraction) or in relative purity after Percoll gradient centrifugation. This suggests that AmB acted directly on the monocyte-macrophages and not via stimulation of the lymphocytes present in culture to produce lymphokines. Nonetheless, it is possible that the small number of lymphocytes present in the Percoll-purified preparation was sufficient to influence the state of macrophage activation. This enhancement was not unique to human cells, but was also seen after overnight culture of rat peritoneal macrophages with AmB (data not shown). All experiments reported in Fig. 1 were performed with macrophages from different donors on different days. It appears that there were two populations of donors, one with low macrophage oxidative activity and the other with a higher baseline activity that increased markedly with AmB priming. We did not attempt to determine whether inherent differences exist among individuals or whether the two

799

groups of responders reflect the wide day-to-day variability of the in vitro biological assay. A study on inbred mouse strains (21) showed parallel inheritance of tissue catalase activity and response to AmB immunostimulation; those animals with high catalase activity showed high baseline macrophage H202 production which was enhanced following AmB treatment, whereas those with low catalase activity showed greater AmB toxicity, which was thought to be due to reduced H202 catabolism. Whether catalase activity variation also occurs in human macrophages was not addressed in this study, but we found no evidence of differences in AmB-mediated toxicity between the high- and low-responder groups. We did not establish the mechanism by which AmB enhances the oxidative burst of macrophages. It has been suggested previously (6) that formation of transmembrane AmB-sterol complexes induces alterations in membrane organization, with consequent changes in catalytic and receptor functions. AmB may bind to the membrane and induce conformational changes that permit greater exposure of membrane enzymes such as NADPH oxidase, thereby enhancing the respiratory burst after stimulation by phorbol myristate acetate. Alternatively, the drug may increase protein synthesis so that more respiratory burst enzymes are available. Sokol-Anderson et al. (18) suggested that AmB triggers a cascade of oxidation reactions and kills fungi by a process of auto-oxidation. This mechanism probably does not explain our observations, because the AmB was removed from the test system prior to measurement of superoxide production. Even in the case of gamma interferon, the best-established macrophage activator, the mechanism by which oxidative metabolism is enhanced is poorly understood (15). The diminished phagocytic capacity of AmB-treated macrophages suggests that it functions via an effect on the plasma membrane. By binding to the cell membrane, AmB may disrupt the phospholipid-cholesterol balance and alter the fluidity of the membrane (4), thus reducing the cell's ability to engulf particles. Gamma interferon has an analogous effect. It activates macrophages to produce enhanced quantities of superoxide but suppresses the capacity for binding via the receptors for C3b and C3bi (27). The fact that treated cells appeared to be no smaller than controls, an observation that was confirmed when flow cytometry was performed (cf. reduced cell size with gamma interferon), suggests that it is not a lack of available membrane that is responsible for the diminished binding and phagocytosis. If our observation of enhanced respiratory burst activity in macrophages exposed to AmB in vitro has an in vivo correlate, this might explain the observation of Medoff et al. (13) that some patients with systemic candidiasis respond to subcandidacidal doses of the drug. At a time when infections with fungi and other intracellular parasites are an increasing problem because of immunosuppressive diseases and therapies, elucidation of the mode of action of available drugs and the complex relationships between drugs, phagocytes, and microorganisms becomes ever more important. ACKNOWLEDGMENTS This work was supported by a grant from the Medical Research Council of Canada. D.P.S. is a scholar of the Canadian Cystic Fibrosis Foundation. REFERENCES 1. Babior, B. M., R. S. Kipnes, and J. T. Curnutte. 1973. The production by leukocytes of superoxide, a potential bactericidal agent. J. Clin. Invest. 52:741-744.

800

WILSON ET AL.

2. Bethell, D. R., M. Dawson, and M. J. LaFoe. 1985. Characterisation of monoclonal antibodies to cell surface antigens by particle concentration fluorescence immunoassay. Biotechniques 3:466-473. 3. Bjorksten, B., C. Ray, and P. G. Quie. 1976. Inhibition of human neutrophil chemotaxis and chemiluminescence by amphotericin B. Infect. Immun. 14:315-317. 4. Blanke, T. J., J. R. Little, S. F. Shirley, and R. G. Lynch. 1977. Augmentation of murine immune responses by amphotericin B. Cell. Immunol. 33:180-190. 5. Brajtburg, J., S. Elberg, D. R. Schwartz, A. Vertut-Croquin, D. Schlessinger, G. S. Kobayashi, and G. Medoff. 1985. Involvement of oxidative damage in erythrocyte lysis induced by amphotericin B. Antimicrob. Agents Chemother. 27:172-176. 6. Chapman, H. A., Jr., and J. B. Hibbs, Jr. 1978. Modulation of macrophage tumoricidal capability by polyene antibiotics: support for membrane lipid as a regulatory determinant of macrophage function. Proc. Natl. Acad. Sci. USA 75:4349-4353. 7. Chunn, C. J., P. R. Starr, and D. N. Gilbert. 1977. Neutrophil toxicity of amphotericin B. Antimicrob. Agents Chemother. 12:226-230. 8. Edelson, P. J. 1982. Intracellular parasites and phagocytic cells: cell biology and pathophysiology. Rev. Infect. Dis. 4:124-135. 9. Holz, R. W. 1974. The effects of the polyene antibiotics nystatin and amphotericin B on thin lipid membranes. Ann. N.Y. Acad. Sci. 35:469-479. 10. Johnston, R. J., Jr., C. A. Godzik, and Z. A. Cohn. 1978. Increased superoxide anion production by immunologically activated and chemically elicited macrophages. J. Exp. Med. 148:115-127. 11. Lin, H., G. Medoff, and G. S. Kobayashi. 1977. Effects of amphotericin B on macrophages and their precursor cells. Antimicrob. Agents Chemother. 11:154-160. 12. Marmer, D. J., B. T. Fields, G. L. France, and R. W. Steele. 1981. Ketoconazole, amphotericin B and amphotericin B methyl ester: comparative in vitro and in vivo toxicological effects on neutrophil function. Antimicrob. Agents Chemother. 20:660665. 13. Medoff, G., W. E. Dismukes, R. I. Meade I1I, and J. M. Moses. 1972. A new therapeutic approach to candida infections: a preliminary report. Arch. Intern. Med. 130:241-245. 14. Medoff, G., and G. S. Kobayashi. 1980. Strategies in the treatment of systemic fungal infections. N. Engl. J. Med. 302:145-155. 15. Murray, H. W. 1988. Interferon-gamma, the activated macrophage, and host defense against microbial challenge. Ann. Intern. Med. 108:595-608.

ANTIMICROB. AGENTS CHEMOTHER. 16. Oosting, R. S., L. van Bree, J. F. van Iwaarden, L. M. G. van Golde, and J. Verhoef. 1990. Impairment of phagocytic functions of alveolar macrophages by hydrogen peroxide. Am. J. Physiol. 259:L87-L94. 17. Perfect, J. R., D. L. Granger, and D. T. Durack. 1987. Effects of antifungal agents and interferon-gamma on macrophage cytotoxicity for fungi and tumour cells. J. Infect. Dis. 156:316-323. 18. Sokol-Anderson, M. L., J. Brajtberg, and G. Medoff. 1986. Amphotericin B-induced oxidative damage and killing of C. albicans. J. Infect. Dis. 154:76-83. 19. Speert, D. P., and S. C. Silverstein. 1985. Phagocytosis of unopsonized zymosan by human monocyte-derived macrophages: maturation and inhibition by mannan. J. Leukocyte Biol. 38:655-658. 20. Speert, D. P., S. D. Wright, S. C. Silverstein, and B. Mah. 1988. Functional characterization of macrophage receptors for in vitro phagocytosis of unopsonized Pseudomonas aeruginosa. J. Clin. Invest. 82:872-879. 21. Stein, S. H., J. R. Little, and K. D. Little. 1987. Parallel inheritance of tissue catalase activity and immunostimulatory action of amphotericin B in inbred mouse strains. Cell. Immunol. 105:99-109. 22. Supapidhayakul, S.-R., L. R. Kizlaitis, and B. R. Andersen. 1981. Stimulation of human and canine neutrophil metabolism by amphotericin B. Antimicrob. Agents Chemother. 19:284289. 23. Thomas, M. Z., G. Medoff, and G. S. Kobayashi. 1973. Changes in murine resistance to Listeria monocytogenes infection induced by amphotericin B. J. Infect. Dis. 127:373-377. 24. Van Gelder, B. F., and E. C. Slater. 1962. The extinction coefficient of cytochrome c. Biochim. Biophys. Acta 58:593595. 25. Vertut-Doi, A., P. Hannaert, and J. Bolard. 1988. The polyene antibiotic amphotericin B inhibits the Na+/K+ pump of human erythrocytes. Biochem. Biophys. Res. Commun. 157:692-697. 26. Weening, R. S., R. Wever, and D. Roos. 1975. Quantitative aspects of the production of superoxide radicals by phagocytizing human granulocytes. J. Lab. Clin. Med. 85:245-252. 27. Wright, S. D., P. A. Detmers, M. T. C. Jong, and C. Meyer. 1986. Interferon-gamma depresses binding of ligand by C3b and C3bi receptors on cultured human monocytes, an effect reversed by fibronectin. J. Exp. Med. 163:1245-1259. 28. Wright, S. D., and S. C. Silverstein. 1982. Tumor-promoting phorbol esters stimulate C3b and C3b' receptor-mediated phagocytosis in cultured human monocytes. J. Exp. Med. 156:1149-1164.