The role of reactive oxygen species (ROS) in the ... - Wiley Online Library

Vol. 43, No. 2, October ] 997

BIOCHEMISTRY and MOLECULAR BIOLOGY INTERNATIONAL Pages 399-408

THE ROLE OF REACTIVE OXYGEN SPECIES (ROS) IN T H E E F F E C T O R MECHANISMS OF HUMAN ANTIMYCOBACTERIAL IMMUNITY Nadeem Fazal

Department of Immunology, The University of Birmingham, Birmingham, U. K. Correspondence and present address:

Burn and Shock Trauma Institute, Stritch School of Medicine, Loyola University Chicago, 2160 South First Avenue, Maywood, Illinois 60153. USA E-mail: [email protected] .luc.edu Received July 22, 1997 SUMMARY The aim of this study was to determine the role of reactive oxygen species (ROS) in checking the growth of intracellular mycobacteria within human phagocytes. Peripheral blood-derived neutrophils and monocyte-derived macrophages were isolated from Chronic Granulomatous Disease (CGD) patients and normal healthy human volunteers. CGD patients are known to have a defect in the NADPH oxidase pathway, resulting in their neutrophils and monocyte-derived macrophages being unable to generate oxygen radicals which are required to kill intracellular bacteria. The cells were then infected with Bacilte Calmette Guerin (BCG) or Mycobacterium avium, and the bacterial growth in each cell type determined by Colony Forming Units (CFU) estimate. The results obtained indicate that there was no demonstrable inhibition in the intracellular mycobacterial growth within neutrophils or macrophages derived from either Chronic Granulomatous Disease (CGD:deficient in NADPH oxidase pathway) or normal healthy volunteers. Macrophage treatment with either IFN-y or TNF-c~ had no effect. KEY WORDS: Mycobacteria, Chronic Granulomatous Disease, IFN-7, TNF-c% Neutrophils, Macrophages, Fazal's Microcolony Counting Method.

INTRODUCTION Phagocytes (neutrophils, monocytes, macrophages) employ both oxygen-dependent and oxygen independent microbicidal mechanisms to kill invading micro-organisms [1, 2, 3, 4]. The relative contribution of each mechanism depends upon the host immune response and the species of bacteria phagocytized, which in turn elicits the anti-microbial capacity of the phagocytes, through cytokines released by T-cells and macrophages. The role of neutrophils in mycobacterial

1039-9712/97/020399-10505.00/0 399

Copyright 9 1997 by Academic Press Australia. All rights o/reproductionin any fiJrm reserved.

Vol. 43, No. 2, 1997

BIOCHEMISTRY and MOLECULAR BIOLOGY INTERNATIONAL

immunity however is yet to be fully determined. Neutrophils have been known to be among the first cells to arrive at the site of infection [2], and the macrophage migration that follows the neutrophil influx, is also thought to result from chemotactic factors released by the neutrophils, in response to mycobacteria. Patients with Chronic Granulomatous Disease (CGD) are known to have a defect in the NADPH oxidase pathway, resulting in their phagocytes being unable to generate oxygen radicals [5]. Earlier studies on neutrophil:mycobacteria interactions have not clearly implicated the role of reactive oxygen species [6, 7]. Experiments were then performed in this study, to find if reactive oxygen species (ROS) were involved in killing of mycobacteria, as CGD

neutrophils/macrophages would then be predicted to be less effective in killing

nonpathogenic BCG or Mycobacterium avium as compared to normal healthy controls. MATERIALS AND METHODS

Preparation of bacteria Bacillus Calmette-Guerin (BCG) and Mycobacterium avium stock solutions were grown in Middlebrook 7H9 broth (Difco, Detroit, Michigan, USA) supplemented with 10 % Middlebrook ADC enrichment media (Difco) and 0.02 % Tween 80, at 37~ in 5% CO 2. The bacterial suspensions were pelleted, resuspended in RPMI-1640, enumerated and frozen in 1.0 ml aliquots at -70~ till further use.

Isolation and infection of human neutrophils Neutrophils were isolated from 10 ml samples of venous blood drawn from CGD patients and healthy human volunteers m Li-Heparin LH/10 tubes (Sars~edt, UK), thoroughly mixed with Dextran 110 (C, P, Pharmaceuticals, Ltd, Wrexham, U.K) and left to sediment at 37~ for 15-25 minutes. Each leucocyte-rich upper layer such obtained was then loaded on to Ficoll-paque density 1.077 +/- 0.001 g/ml (Pharmacia, Uppsala, Sweden), and centrifuged at 400 g for 25 minutes. The neutrophil-enriched cell pellet was resuspended in lml of ice cold distilled water to lyse the contaminating red blood cells (RBC's). The cell suspension was then washed twice by centrifugation at 400 g for 10 minutes and the cell pellet finally resuspended in HBSS/MOPS (Gibco, Paisley, Scotland/Sigma, St. Louis, USA) to a concentration of 107 cells/ml. Purity of neutrophil cell preparation at this stage was 98 % as evidenced by trypan blue exclusion. BCG suspension at 108/ml bacteria were then added to cell suspension, transferred into a 25ml tissue culture flask (Flow, England) on a rotating shaker (G24 Environmental incubator shaker, New Brunswick Scientific Co., Edison, N. J, USA) at 20 rpm for 30 minutes at 37~ % CO 2. After phagocytosis, the infected cells were pelleted to remove non-phagocytized bacteria, resuspended in fresh HBSS, and incubated at 37~ in 5 % CO 2. Samples were taken at zero-time, 1 h and 2 h post-infection by removing 200 ~tl aliquots of the reaction mixture, to determine the viable 9

/

counts.

Preparation and infection of human macrophages Defibrinated human venous blood was obtained, centrifuged for 10 minutes at 400 g (MSE, Mistral 2000) to allow separation and removal of the serum. The buffy coat fraction was carefully removed, layered over Ficoll-paque and centrifuged at 400 g for 25 minutes. The cells 400

Vol. 43, No. 2, 1997


from the interface were removed, washed and adjusted to 2.5 xl06/ml in antibiotic-free RPMI1640 +10 % unheated autologous human serum. The cells were seeded in 96-well, flat-bottomed microtitre tissue culture plates (Flow laboratories, England) in 200 ~tl aliquots/well. Following overnight incubation at 37~ in 5 % CO 2, the wells were washed to remove nonadherent cells. The adherent monocyte-enriched cultures were then incubated for 4 days at 37~ in 5 % CO 2 to allow differentiation into macrophages [8]. Infection experiments were performed as described [9, 10, 11]. Briefly, a BCG stock aliquot was thawed, sonicated in a sonicating waterbath (Heathkit) for 3 minutes to disrupt the BCG clumps without affecting bacterial viability and adjusted to the required infection concentration in RPMI-1640. The bacteria were then incubated in 10 % unheated autologous human serum for half an hour at 37~ to allow opsonisation and then added to the cell cultures in 20 ~tl aliquots/well. The plates were then incubated for 4 h to allow phagocytosis to occur and the non-phagocytized bacteria were removed. 100gl aliquots/well of fresh medium were then added and sample wells were harvested at different time intervals post-infection by lysis of the adherent macrophage monolayer with 20 ~tl of 2 % saponin. The viable counts were determined by CFU estimate.

Fazal's Microcolony Counting Methodfor determining mycobacterial growth/viability The detailed method is described elsewhere [10]. Briefly, 50 ~tl samples of cell lysates containing liberated bacteria were added in 96 well flat-bottomed tissue culture plates containing Middlebrook 7H9 broth + 0.2 % glycerol + 10 % ADC supplement. Serial 5-fold dilutions of the samples were made. After 10-14 days of incubation in sealed plates at 37~ in 5 % CO 2, visible mycobacterial microcolonies were then counted under an inverted microscope. The method was modified for enumerating M. avium and the plate cultures were visualized after 1-3 days. The counts are represented as CFU/ml in the following experiments.

RESULTS

Neutrophil:mycobacteria infection assays ~These infection experiments were performed to determine the role of oxidative respiratory burst in human phagocytes in the control of nonpathogenic mycobacteria. Neutrophils isolated from two Chronic Granulomatous Disease (CGD) patients and four normal healthy human donors were infected with BCG at a bacteria to cell ratio of 5:1. Nonphagocytized bacteria were removed and the number of intracellular BCG determined at zerotime (time 0) i.e., 30 minutes post-infection by CFU estimation. This determines an index of percentage of mycobacteria phagocytized by neutrophils. All the bacteria at this time were intracellular as evidenced by Ziehl-Neelsen staining of the infected neutrophil cultures (results not shown). The ability of neutrophils to kill BCG within the first two hours after phagocytosis was then assessed by performing CFU estimates. The viability counts (CFU) performed 1 and 2 hours postphagocytosis however did not show any evidence of killing or inhibition of BCG growth in any of the donor neutrophil cultures. Likewise, nentrophils of the two CGD patients were not any

401

Vol. 43, No. 2, 1997


different than normal human controls in allowing the unabated

growth/survival

of BCG

(Table: 1). It is of interest to note that percentage of mycobacteria phagocytized, i.e., zerotime counts, differed in donor neutrophils both from CGD patients and normal human controls, but the trend was the same, regardless of initial uptake of bacterial inoculum (Table I).

Macrophage :mycobacteria infection assays

In subsequent experiments, four-day old human monocyte-derived macrophages from a CGD patient and a healthy human volunteer were treated with IFN-,/(100 U/ml) overnight, followed by infection with either BCG or ~ avium, at a bacteria to cell infection ratio of 5:1. Control macrophage cultures not pre-incubated with cytokines, but similarly infected were also included. Nonphagocytized bacteria were removed and the number of intracellular bacteria determined as zerotime (time 0) i.e., 4 hours post-infection. Intracellular growth of bacteria was then followed over a 3 days post-infection period, which has previously been established to be the optimum time point to look at the survival of intracellular mycobacteria [10, 11 ]. The results showed that BCG as well as M. avium grew intracellularly in cytokine-treated or untreated macrophage cultures, as assessed by CFU estimation by Fazal's microcolony counting method (Table: 2). Moreover, BCG and M. avium liberated from cytokine-treated macrophage cultures showed an enhanced intracellular growth in both macrophage cultures from the CGD patient (BCG + 95 %, M. avium + 107 %) and healthy control (BCG + 9 1 % , M~ avium + 67 %) as compared to

untreated controls (Table: 2). Hence, CGD cells (deficient in NADPH oxidase activity) were not any different in allowing growth of intracellular mycobacteria than normal human cells. In some experiments human cells were treated with TNF-c~ (1000 U/ml) and IFN-7 (100 U/ml) for 4 h prior to infection, or added simultaneously at the time of infection. However, such strategies failed to alter the fate of intracellular bacteria (data not shown).

DISCUSSION It is known that in tuberculosis, neutrophils are among the first cells to arrive at the site of infection [2, 3]. Moreover, the subsequent macrophage migration that follows the neutrophil influx is known to

result from chemotactic factors released by neutrophils in response to

mycobacteria [4]. What lethal effects could these short lived phagocytic ceils have, was the subject of our interest. Previously, it has been reported that human neutrophils kill M. tuberculosis. [7, 13 ], while others failed to reproduce these findings [ 12]. Chronic granulomatous

402

Vot. 43, No. 2, 1997


O

o

V r.~

O

E

II

.

E

8

0~

~b ]

~

-.

~

~-

~.

~

~

N

'.~

~r.) O

=

..=

+dc~ r~

r)

403

e-~

Vol. 43, No. 2, 1997


TABLE: 2 Macrophage infection assay

Md~Controla

Cytokinesf

Zerotime b

Md? CGD

Day 3e

% GEd

Ae

4.2 4.6

+262 +353

11.6 10.1

+269 +202 +67

Zerotime Day 3

% GE

k

BCG x IOs CFU/ml

+

1.6 1.3

+91

1.2 1.0

3.9 4.2

+325 +420

+95

10.0 12.3

+ 129 +236

+107

MAI x 10s CFU/ml

+

4.3 5.0

7.7 5,2

Data represents mean of triplicates obtained from 5 experiments (SD < 10% of mean not shown). aMacropbage cultures obtained from a healthy human donor were used as control bIntracellularnumber of bacteria were determined 4 hours post-infectionas zerotime. e3 days post-infection dpercentage change in bacterial growth was calculated by = [CFU at day 3/CFU at zerotime] x 100 (+) Growth enhancement. eA % GE with IFN-~/(-)% GE without IFN-y fMacrophage cultures were pre-activated with IFN-y(1000 U/ml) for 18 hours prior to infection. * Statistically significant (p < 0.05). Students t test was performed for statistical analysis.

disease (CGD) patients whose phagocytes are known to have a defect in NADPH-oxidase pathway, resulting in their neutrophils unable to generate oxygen radicals [5] have also shown to have at most a modest bactericidal capacity against a virulent strain of tubercle bacilli [14, 15]. In one of the studies [14], gamma interferon was found to overcome the ROS defect in CGD cells. Antimycobacterial activity of healthy neutrophils have also been found to be unaffected by scavengers of free-oxygen-radicals [ 15]. Lack of consistent results and variability of the methods employed

to determine the viability of bacteria was considered to be responsible for this

discrepancy. Moreover, the fact that these in vitro experiments failed to explain the in vivo situations where neither neutrophil disorders, like leucocyte adhesion deficiencies (CD 18/CD 11, etc.), nor neutropenia has been associated with increased susceptibility to mycobacterial infections [16J. In an attempt to clarify the current blurred picture of the role for neutrophil in antimycobacterial immunity these experiments were undertaken. Two

404

types of human

Vol. 43, No. 2, 1997


phagocytes, i.e., neutrophils-short lived, and macrophages-long lived cells, and two types of donors cells, i.e., obtained from CGD patients that lack capacity to produce reactive oxygen species, and normal human with normal capability to produce oxygen burst, were used in the following experiments. Similarly, two strains of mycobacteria, i.e., BCG-slow growing, and M. avium-fast

growing,

were utilized to explore the role of oxygen burst in human

antimycobacterial immunity. A variety of parameters were thus available to assess the fate of intracellular mycobacteria within human phagocytes. Our results however show that both cells, including the one lacking NADPH oxidase activity, and the other possessing such an activity, allow unrestricted growth of intracellular mycobacteria. Moreover, cells obtained from CGD patients were not any different from cells obtained from normal healthy controls in phagocytizing and allowing both, fast- and slow-growing mycobacteria to survive within human phagocytes.

TNF-ct plays an important role in both protective and pathological immune response in synergism with IFN-,{ and their role in the induction of iNOS pathway has been also suggested [17]. Especially in murine models iNOS is implicated in mycobacterial killing [18]. in our studies, we have measured TNF-a production and NO in the culture supernatants of the infected monocyte-derived macrophage monolayers in these infection experiments. The presence of TNFct in the culture supernatants was considered a sign of macrophage activation, and nitrite production as a marker of iNOS activity of the infected cells. TNF-c~ was detected by a TNF-c~ 'specific ELISA, and nitrite levels were determined by chemiluminiscence as previously described [11]. Within the time frame of the assay, no nitrite was detected at day 1 or 3 postinfection in the culture supernatants of infected monolayers (data not shown). The levels of TNFc~ detected in the culture supernatants of M. avium infected samples were higher than those observed with BCG-infected cells. These results showing a differential pattern of TNF-ct production are in conformity to previous reports, where the difference in Lipoarabinomannan (LAM) structure of mycobacterial cell wall was attributed to the difference in TNF-c~ production between the rapid and slow growing mycobacterial species [27]. We have previously shown that neither enhancement nor inhibition of TNF-a production substantially altered the pattern of intracellular survival of BCG and M. avium within human phagocytic cells [11]. The levels of TNF-c~ produced in the preceding experiments were similar to what we found in our earlier

405

Vol. 43, No. 2, 1997


studies, e.g. 1000-1500 ng per 5 x

10 4

macrophages [11]. However, our results verify no killing

or inhibitory effect on the intracellular growing mycobacteria in the presence of enhanced production of TNF-c~ by the infected cells. Cytokine host response to the invading mycobacteria determines its ultimate fate. Our results then throw into question the current dogma which states that macrophages kill intracellular mycobacteria following activation via appropriate Thl-associated cytokines such as IFN-y, IL-2 & TNF-c~ [19, 20, 21, 22].

IFN-7 however has been widely reported to activate murine

macrophages to inhibit the intracellular growth of mycobacteria [23], notably inhibitor of tuberculosis in murine peritoneal macrophages [24]. Conversely, r-IFN-~, has been reported to cause only a modest inhibition, or even increased growth of intracellular M. tuberculosis in human macrophages [19, 25]. In our studies we compared the effects of exposing macrophages to IFN-3', before and after the infection, on subsequent intracellular growth of mycobacteria. The results failed to find any killing effect or inhibition of mycobacteria. Similarly, macrophage monolayers pre-treated with IFN-~, and continuously incubated with IFN-~{ also showed lack of any mycobactericidal effect. On the contrary in some of the donor macrophage cultures there was an enhanced intracellular growth as mentioned in the results section. Hence, our data suggests that IFN-? can activate human macrophages, as indicated by TNF-~ production but this activation was inadequate to sustain an inhibition of intracellular mycobacteria. Our contention is supported by published results on the effects of IFN-7 on mycobacteria [19, 23, 24, 25]. Thus far there has not been an unequivocal demonstration of killing of mycobacteria by human phagocytes. Although human neutrophils have been reported to kill M. tuberculosis [7, 13] based on determination by a radiometric assay. Such parameters of assessing growth index of mycobacteria do not establish the viability of the organisms [10]. Our studies have therefore relied on CFU counts which determine the actual viability and growth of dividing bacteria. Moreover, intracellular parasite like mycobacteria escape killing by long lived macrophage cells by finding a niche to survive in these phagocytizing cells, is unlikely to succumb to short lived cells like neutrophils. Consistent with our findings, a recent paper has elegantly described unabated growth of non-pathogenic M. smegmatis in adherent monocyte-derived human macrophage cultures [30]. As discussed in our previous papers [ 9-11], we have noticed that whenever the bacteria liberated from phagocytic cells following lysis are put to a system of enumeration of viable bacteria. The

406

Vol. 43, No. 2, 1997

BIOCHEMISTRYand MOLECULAR BIOLOGY INTERNATIONAL

presence of cell lysate itself may act as an inhibition factor to retard the growth of bacteria, or actually enhance the subsequent growth of bacteria by providing an enriched nutrient media conducive to the growth. Some researchers have used the approach of spinning down such cell debris to purify bacteria and thus minimize such errors. However such an approach is vulnerable to loss of bacteria, leading to erroneous results showing enhanced mycobactericidal activity. We have often observed this variable phenomenon in our infection experiments (unpublished results ).

There is heterogeneity of human macrophage populations in showing such a variable

enhanced or inhibited effect on the growth of mycobacteria, This leads us to speculate if there is a factor present in the cell lysate responsible for causing such an effect in an in vitro system. There are two evident approaches to understand such a behavior of cultured injected human cells. Infection of phagocytic cells by mycobacteria may induce either a Thl or Th2 response, depending upon various cytokines produced. Thl response leads to protective effect which in tissue culture environment is shown as reduction of bacterial growth, or a Th2 response leading to cytokine production responsible for favoring intracellular growth of bacteria causing pathogenic effects. These include cytokines such as IL-4, 1L-6, IL-10 & Transforming Growth Factor-beta (TGF-[3) [28, 29]. In these experiments both a rapid growing mycobacteria-M, avium, and a sl0w growing-BCG are utilized to address such a possibility. Further studies are however required to answer this question. We surmise that in an in vitro environment of the laboratory, mycobacteria replicate 9intracellularly within neutrophils and human monocyte-derived macrophages. In the battlefield of a granuloma-hallmark of tuberculous infections, a variety of phagocytic cells arrive and discharge their armamentarium to get rid of invading mycobacteria. In the current setting to identify the role of individual cytokine or a cell type may not represent the true in vivo scenario, because both an intracellular as well as a intracellular environment is required resulting in an effective immune response. Experiments are thus being designed whereby a natural environment will be provided to address the mechanisms of effector functions of human antimycobacterial immunity. REFERENCES 1,

2. 3.

Hasset JD, Cohen MS. FASED J 1989;3:2574-2582. Bloch H. Am Rev Tuberculosis 1948;58;667-70 Montgomery L J, Lemon W. J Thorac Cardiovasc Surgery 1933; 2;429-38.

407

Vol. 43, No. 2, 1997

4.

5. 6. 7. 8.

9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30.

BIOCHEMISTRYond MOLECULAR BIOLOGY INTERNATIONAL

Antony VB, Sahn SA, Harada RN. Chest 1983;83 (suppl):95-96. Cumutte JT, Babior BM. Adv Human Genet 1987;16;229-298. Jones GS, Amirault HJ, Anderson BR. J Infect Dis 1990; 162;700-704. Brown AE, Holzer TJ, Anderson BR. (1987), J Infect Dis 1987;156; 985-989. Fazal N, Lammas DA, Pithie AS, Rahelu M, Gaston JSH, Kumararatne DS. Clinical & Experimental Immunology 1995;99;82-89. Pithie AD, Laminas DA, Fazal N, Rahelu M, Bartlett R, Gaston JSH, Kumararatne DS. FEMS Immunology & Medical Microbiology 1995; 11, 145-154. Fazal N, Bartlett R, Lammas DA, Kumararatne DS. FEMS Immunology & Microbiology, 1992;105:355- 362. Fazal N, Bartlett R, Lammas DA, Raykundalia C, Kumararatne DS. FEMS Immunology & Microbiology, 1992;105:337-346. Denis M. J Infec Dis 1991;163:919. Anderson AB. J Inject Dis 1991;156:985-989. Sechler JMG, Malech HL, White CJ. Proc Natl AcadSci 1977;85:4874-8 Denaver MC. Interferon and CGD: Current opinion in Immunology 1991 ;3:61-64 Johnston RB, NEJM, 1984;310:1237-43 Marshall BG, Chambers MA, Wangoo A, Shaw RJ, Young DB, Infect. lmmun 1997, 65, No. 5: 1931-1935. Chan J, Tanaka D, Carrol J, et al, Infect. Immun 1995, 63:736-40 Rook GAW, Steele J, Ainsworth M, Champion BR. Clin. Exp. Immunol. 1986; 59:333338. Crowle AJ. Res Microbiology 1990;141 ;231-236 Flesch IEA, Kaufmann SHE. Infect lmmun 1990;58;2675-77 Walker L, Lowrie DB. Nature 1981; 293;69-70. Kaufmann SHE. Ann Rev Immunol 1993;11;129-63. Rook GAW, Steele J, Ainsworth M, Champion BR. Immunology 1986;59: 333-338. Douvas GS, Looker DL, Vatter AE, Crowle AJ. Infect lmmun 1985;50;1-8. Rook GAW, AI Attiyah RJ. Tubercle 1990;72;768-770. Havell EA. J Infec Dis 1986; 153:690 Appleburg R, Orme IM, Desouza MIP, Silva MT. Immunology 1992;76: 553-559. Bermudez LE. J Immunol 1993;150:1838-1845 Barker K, Hongxia F, Clare C, Kaplan G, Barker J, Wilhelmine JH, Cohn Z A. Infect Immun 1996; 64:2, 428-433.

408