| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Departments of
* Internal Medicine I and
Medical Microbiology and Hygiene, University of Cologne, Germany
Correspondence: Pia Hartmann, M.D., Department of Internal Medicine I, Division of Infectious Diseases, Joseph-Stelzmann-Strasse 9, 50924 Cologne, Germany.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: granulocytes • host defense • atypical mycobacteria
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Monocytes and macrophages are thought to be responsible for ingestion and killing of MAC in analogy to mouse models of MAC disease. Yet both ingestion and intracellular growth inhibition of MAC by monocytes of AIDS patients have been shown to be normal in vitro [10 ]. M. avium, like other pathogenic mycobacteria, is ingested by macrophages via a number of receptors, including the complement receptor CR3 [11 ]. Several strains of M. avium have been described to resist the oxidative bactericidal mechanisms of human macrophages [12 ].
Although neutrophils have historically been considered to be inconsequential in mycobacterial infection [13 ], and the capability of human neutrophils to kill Mycobacterium tuberculosis has been discussed controversially in the literature [14 15 16 ], several animal studies indicate that neutrophils may be involved in the defense against MAC disease. In mouse strains susceptible to MAC, neutrophil infiltrates were reduced in comparison with infiltrates in resistant mice after intraperitoneal inoculation of M. avium [17 ]. Also, the investigation of lungs and regional lymph nodes from mice infected with M. avium revealed macrophages containing not only many acid-fast bacilli but also neutrophil-derived lactoferrin [18 , 19 ]. Thus, neutrophils have been thought to increase the macrophage’s effectiveness in eliminating mycobacteria by initially phagocytizing M. avium, then causing partial degradation of the bacteria and releasing enzymatic granules that macrophages phagocytize in turn and use for the further killing process [18 ]. Other in vitro and in vivo experiments indicate that granulocyte-macrophage (GM)- and granulocyte colony-stimulating factor (G-CSF) augment the capacity of mice and human neutrophils to inhibit the growth of M. avium [20 , 21 ]. The administration of G-CSF has been shown to have a beneficial effect on survival of AIDS patients with MAC disease [22 ]. These data support the hypothesis that the first step in the defense against MAC is taken by neutrophils. Phagocytosis of MAC by neutrophils has been shown in mice [18 , 23 ] and recently observed microscopically in humans [24 ].
We have initiated a study of the functional response of neutrophils to MAC infection. Here we show by flow cytometry and by microscopy that human neutrophils suspended in autologous plasma phagocytize MAC in vitro. In addition, we observed a time-dependent decrease in colony-forming units of MAC after exposure to neutrophils, indicating that human neutrophils kill M. avium. These observations are supported by an accompanying electron microscopy (EM) study. EM sections showed intracellular MAC as early as 15 min after exposure to granulocytes. Signs of bacterial degradation could be observed at 60 and 120 min.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Staining of mycobacteria for flow cytometry and fluorescence
microscopy
M. avium was stained with Auramine O (Sigma,
Deisenhofen), a biological stain with the molecular formula
C17H21N3 • HCl,
max: 434 nm in water. Staining procedure was as follows:
approximately 50 µg (wet weight) of M. avium was diluted
in 500 µL phosphate-buffered saline (PBS) and centrifuged for 30 s at 10,000 g in a microfuge. After discarding the
supernatant the pellet was resuspended in 500 µL 10% dimethyl
sulfoxide (DMSO) in PBS. This centrifugation procedure was repeated
four times, increasing the concentration of DMSO with each resuspension
of the pellet (20, 40, 75, and 100%). Finally 500 µL of cells in
100% DMSO were mixed with 500 µL of a 1% Auramine O solution in
absolute ethanol and incubated for 30 min at room temperature. The
staining solution was centrifuged for 30 s at 10,000 g
and the pellet resuspended in PBS at approximately 1 x
107 bacteria/mL.
Incubation of polymorphonuclear neutrophils (PMNs) with MAC
Leukocytes were obtained as a supernatant after sedimentation of
3 mL venous blood from healthy donors for 45 min at 1 g on 3
mL Histopaque 1.077 (Sigma, Deisenhofen). Twenty microliters of
supernatant plasma containing the leukocytes were incubated at 37°C
with 10 µL suspension of stained M. avium cells in a final
volume of 1 mL Hanks’ balanced salt solution (HBSS; Sigma) with 10 mM
HEPES, pH 7.4 (Serva, Heidelberg) leading to a neutrophil/MAC ratio of
approximately 1:5. A negative control was kept on ice at 0°C for the
entire incubation period. Just before being added to the cells, the
mycobacteria were washed six times in PBS (microfuge, 2 min at 10,000
g) to clear the solution of soluble Auramine O and then
pretreated in a sonicator bath at 100 W for 10 min to disperse clumps
of mycobacteria. To rule out cell surface staining of the granulocytes
by soluble Auramine O, granulocytes were also incubated with 10 µL of
the supernatant remaining after the last washing step. After incubation
the probes were stored on ice and 10 µL propidium iodide (PI; Serva,
Heidelberg) were added to each probe to identify dead granulocytes.
Routinely, neutrophils were incubated for 15 min. To evaluate the time
course of the phagocytic process the incubation period was extended for
up to 60 min and interval measurements were done every 5 min over the
entire incubation period.
Differential sedimentation assay
The influence of the Auramine staining procedure on the
granulocytes’ capacity to bind and phagocytize mycobacteria was
evaluated in a differential sedimentation assay. M. avium
was metabolically labeled by growing the cells in Middlebrook 7H12
medium containing 1.44 kBq/mL [14C]palmitic acid (Bactec
12b medium, Becton Dickinson). Bacteria were harvested during
exponential growth phase, washed in PBS and sonicated as described
above. Bacterial cells were used without any further treatment (I), or
pretreated with the increasing DMSO step gradient concentrations alone
(II) and with the final Auramine staining step (III).
Granulocyte-associated mycobacteria were quantified after differential
sedimentation by centrifugation. This was done through coincubation of
bacteria and granulocytes at 37°C for 15 min as described above,
pelleting the phagocytes by centrifugation at low relative centrifugal
force (300 g, 10 min) and counting the cell-associated
14C activity. To assess the validity of the differential
centrifugation approach an unstained bacterial suspension was
centrifuged in the absence of granulocytes under identical conditions
(IV; see Fig. 1
). The radioactivity of each probe was measured and
calculated as percentage of the total bacterial radioactivity measured
for MAC in suspension in the absence of granulocytes.
|
Fluorescent microscopy
To verify that the green fluorescence measured by flow cytometry
was due to intracellular Auramine-stained M. avium,
incubated probes were evaluated under a Zeiss fluorescent microscope at
x1000 magnification with a 50-W mercury high-pressure lamp and an
excitation filter with transmission from 355 to 425 nm and a
suppression filter set at 460 nm. For each donor four slides were
prepared and 100 cells evaluated on each slide.
Killing activity
To evaluate whether the uptake of MAC by human neutrophils is
associated with killing activity, another series of experiments was
performed. MAC suspension was appropriately diluted and pretreated by
sonication as described above. Aggregates of bacteria were removed by
low-speed centrifugation (300 g, 10 min). A solution of 1000
µL HBSS with HEPES, pH 7.4, containing PMN (2.5 x
106/mL) and MAC (2.5 x 107/mL) was
incubated at 37°C for defined intervals with a minimum of 15 min and
a maximum incubation period of 3 h. A negative control was kept on
ice for the entire incubation period. Granulocytes were separated from
non-phagocytized mycobacteria in the microfuge (10 min/300
g). The supernatant was removed and intracellular and
cell-associated MAC were released by resuspending the remaining
granulocyte pellets in 1000 µL sodium dodecyl sulfate (SDS) 0.1%.
The supernatant and the lysate were then diluted in a dilution series
of four 1:10 (v/v) steps and each dilution was plated on Middlebrook
7H10 agar. Colony-forming units (CFUs) were counted after a time period
of 7–14 days to estimate neutrophil killing activity. The multiplicity
of infection (MOI) was evaluated as CFU at 15 min/2.5 x
106 PMNs. For the various incubation periods the killing
index expressed in percent [25
, 26
] was
calculated as percent of CFU at 15 min ([CFU at 15 min - CFU at
x min] x 100/CFU at 15 min) based on the observation in
the flow cytometry experiments that maximum uptake of MAC by PMNs was
reached at 15 min and killing activity was minimal within that period
of time. Neutrophils from nine individuals were measured in
quadruplicate.
Electron microscopy processing
Accompanying the killing study samples for electronic microscopy
were prepared as follows: at selected time points during infection (15,
60, and120 min), cells were fixed for 1 h at room temperature with
2.5% glutaraldehyde in 0.1 M cacodylate buffer, pH 7.2, containing 0.1
M sucrose, 5 mM CaCl2, and 5 mM MgCl2. Cells
were washed overnight at 4°C with sucrose-containing cacodylate
buffer and postfixed for 1 h at room temperature with 1% osmium
tetroxide (OsO4) in the same buffer. The cells were
concentrated in 2% agar in cacodylate dehydrated buffer, and treated
for 1 h at room temperature with 1% uranyl acetate in Veronal
buffer. Samples were dehydrated in a graded series of acetone and
embedded in Epon. Thin sections were stained with 2% uranyl acetate
and lead citrate.
Study population
Eleven healthy donors volunteered for the experiments: four
females and seven males. They were between 24 and 42 years old. Their
mean granulocyte count was 8135 ± 2337/µL.
Statistics
Statistical significance of the differential sedimentation assay
was assessed by the Student’s t test. P values
less than 0.05 were considered statistically significant. CFU values
and killing index, respectively, were analyzed by variance analysis
using SPSS 6.1.(SPSS, Chicago, IL). Post hoc pairwise
comparisons were done using Scheffes’ test.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Differential sedimentation assay
In three independent experiments neutrophil-associated
14C-labeled but unstained bacteria reached 19.2 ±
6.5% of the total bacterial radioactivity measured for a suspension of
radioactively labeled bacteria in the absence of granulocytes.
DMSO-treated and fully stained bacteria neutrophil associated at
slightly lower percentages, 16.1 ± 1.3 and 13.9 ± 2.5% of
bacterial radioactivity, respectively, but these differences were not
significant by statistical analysis. In contrast, centrifugation of
radioactively labeled bacteria in the absence of granulocytes showed
only low activity in the sediment (IV; Fig. 1
).
Flow cytometry
To define the population of mycobacteria in the FSC/SSC window a
suspension of 10 µL 10-fold diluted unstained mycobacteria in 1000
µL HBSS were analyzed. Mycobacteria appeared as small cells (FSC
<400) of weak refraction (SSC <400). After gating the observed
unstained bacterial population, a suspension of 10 µL 10-fold diluted
Auramine-stained M. avium in HBSS was evaluated within the
gate. FL 1 was measured with a median fluorescence of 559.65 ±
221.63. When measuring 20 µL PMN in 1000 µL HBSS without
mycobacteria an irregular population appeared within the mycobacteria
gate, probably cell fragments, and therefore this population was not
further taken into account. The thus defined mycobacterium gate was
kept as standard for all measurements.
Granulocytes in the leukocyte-rich supernatant of whole blood from 11 healthy donors were measured for their ability to phagocytize Auramine O-stained M. avium. For each donor four replicate incubations were evaluated in the same experiment. Phagocytosis of mycobacteria is demonstrated by a progressive fluorescence shift over time on a logarithmic scale comparing FL1 of granulocytes incubated at 37°C with those incubated at 0°C. FL 1 for granulocytes incubated at 37°C was significantly higher than for those kept on ice after 5 min. A twofold fluorescence shift was observed at 15 min. (Figures 2 and 3 ). Analysis by Cell Quest 3.1f showed that at 37°C only 24.6 ± 5.4% of the granulocytes within the granulocyte gate contributed to FL1. This was also expressed by a remaining fraction of MAC that was still present within the mycobacteria gate even after 60 min of incubation with a consistent FL1 of 212.34 ± 43.87. Median fluorescence FL1 for all probes of all donors is shown in Table 1 . Granulocytes incubated with the supernatant of the M. avium solution had a median fluorescence of 12.80 ± 0.66 (n = 6) showing that neutrophil surface staining by soluble Auramine O remaining after the washing procedure was low. The intraindividual variability for all measured variables was very low (Fig. 3) .
|
|
|
|
|
Fluorescence microscopy
Using the fluorescence microscope, 35 ± 12% of granulocytes
from the 11 donors showed four to six bright white/green fluorescent
mycobacteria, which appeared clearly intracellular, as could be
demonstrated by adjusting the focus at different depths. Only a few
extracellularly associated mycobacteria were observed (data not shown).
Killing activity
After a culture period of 7–14 days the CFUs/1000 µL of MAC
rescued after 15 min of incubation at 37°C with neutrophils from nine
independent donors were 2.72 ± 0.13 x 106/mL,
yielding an average MOI for granulocytes in the presence of MAC of
1.08 ± 0.13.
CFUs here shown in a representative experiment (Table 2 ) declined over time comparing MAC released after an incubation period of 15 min with those released after 30, 45, 60, and 120 min. Logarithmic decline of CFUs was 1.4 by 120 min. As shown by the killing index (KI) in the nine independent donors, half of the bacteria phagocytized at 15 min were killed at 45 min; KI, 53.73 ± 7.73%, and killing was virtually complete at 120 min; KI, 96.58 ± 0.71% (Fig. 6 ). No significant difference in CFUs could be detected in the supernatant for the different incubation periods (data not shown), thus confirming the flow cytometry data that no further phagocytosis occurred after 15 min of incubation. CFUs in the probe kept on ice for 180 min remained stable. All differences for CFUs and KI (incubation at 15 vs. 30 min, 30 vs. 45 min, 45 vs. 60min, and 60 vs. 120 min) were statistically significant (P < 0.05).
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Because the flow cytometry method depends on the use of fluorescently stained M. avium, it was important to assess possible deleterious effects of the Auramine staining procedure on bacterial binding and uptake by neutrophils. For this specific purpose only we used a differential sedimentation method that measures cell-associated bacteria after a low-speed centrifugation to eliminate extracellular bacteria as far as possible. This method has also been used in the past to demonstrate phagocytic uptake, but has been subject to much criticism because it is very difficult to differentiate between intracellular particles and those adherent to the cell surface [30 ]. The Auramine staining procedure did not lead to a significant difference in neutrophil-associated bacteria.
M. avium were also clearly visible within the neutrophils by fluorescence microscopy. Repeated observations showed that 23–47% of the neutrophils contained multiple bacteria, thus suggesting avid phagocytosis. However, it is clear from the flow cytometric and microscopic data that not all neutrophils are capable of MAC ingestion.
Results from nine independent experiments indicated clearly that neutrophils not only phagocytize M. avium, but are capable of killing them. Significant killing was observed within 15 min of the completion of phagocytosis and almost complete within 2 h. The time course of the killing experiments is in concordance with the time course of phagocytosis measured by flow cytometry, as there was no significant change in CFUs of extracellular bacteria in the supernatant obtained at different incubation times after 15 min. In the killing experiments only 10% of MAC were engulfed at 15 min, as judged by an initial bacteria-to-cell ratio of 10:1 and the determined effective MOI of 1. This is similar to our observation by flow cytometry that with a bacteria-to-leukocyte ratio of 5:1 the capacity to phagocytize MAC was only expressed by 20–30% of granulocytes.
Strong support of the data obtained by flow cytometry and in the killing experiments is provided by electron microscopy, which clearly showed intraphagosomal MAC even at the earliest time point of 15 min.
No increase of intraphagosomal bacteria over time was observed, as also found by flow cytometry. However, the data are not yet sufficient to allow a detailed analysis of the kinetics of the phagocytic process. The fact that phagocytosis of MAC by human neutrophils was not a frequent event is again in concordance with our flow cytometry data showing only 20–25% of neutrophils being capable of phagocytizing MAC as well as with our killing experiments showing that only 10% of the added bacteria are ingested at 15 min of infection. Although macrophages continue to phagocytize up to hundreds of mycobacteria over several hours, phagocytosis of MAC by human granulocytes is completed at 15 min. The reason for this phenomenon is not clear. Flow cytometry measurements of phagocytosis by PMNs for other bacteria such as Escherichia coli or staphylococci show a similar time course with phagocytosis being complete between 15 and 30 min. However, with these bacteria, 70–90% of granulocytes within the granulocyte gate participate in the phagocytic process and nearly all bacteria are phagocytized [31 , 32 ].
Although considerable evidence has been presented that neutrophils may play a role in mycobacterial infection, the mechanisms involved remain unclear. Neutrophils have been reported to be present only during the initial phase of mycobacterial infection [13 , 33 , 34 ]. However Silva et al. found a granulocytic response beyond the first 30 days and up to 3 months in mice inoculated intraperitoneally with M. avium. They also observed by light and electron microscopy the phagocytosis of mycobacteria by murine granulocytes and the ingestion of granulocytes containing mycobacteria by murine macrophages [18 ]. The same group observed that the increased susceptibility of beige mice to M. avium is reduced by transfusing neutrophils from wild-type C57BL/6 mice to C57BL/6 mice with the beige mutation [35 ]. In analogy to Silva, Pedrosa assessed neutrophil recruitment in the peritoneal cavity after intraperitoneal infection and to the spleen and liver after intravenous inoculation of mice with M. tuberculosis. The authors conclude that neutrophil host defense mechanisms against M. tuberculosis are not associated with phagocytic activity but may be of an immunomodulary nature [36 ]. Our results do not exclude an immunomodulary neutrophil defense against mycobacteria. However, the phagocytic and killing activity of human neutrophils against MAC are clearly shown by our data. Our data were acquired in a human in vitro system working with peripheral blood in the absence of tissue resident macrophages. It is interesting to note that we observed two populations of neutrophils, one that does not phagocytize MAC and a smaller active population that phagocytizes and kills MAC. One could hypothesize that phagocytosis of MAC by the active neutrophil population initiates a cascade of immunomodulary activity involving the non-phagocytizing population as well as tissue resident macrophages.
Kazanjian and co-workers have recently published three papers concerning the possible role of neutrophils in MAC infection. There are clear discrepancies between their results and ours. They reported growth inhibition of M. avium in the presence of human neutrophils that was further enhanced by the application of G-CSF in vitro [21 ] and in vivo [37 ] but not by in vivo administered GM-CSF [38 ]. In their experiments human neutrophils from HIV-negative donors inhibited the growth of M. avium at 7 days of incubation with the bacteria compared with peripheral blood mononuclear cells and macrophages. The significance of these results in terms of neutrophil function is questionable because there was no clear growth inhibition by neutrophils when compared to medium alone and the authors state that by 48 h the viability of neutrophils was less than 10%. It is particularly difficult to reconcile these results obtained by Kazanjian et al. over days with ours in which it is clear that maximum phagocytosis is achieved within 15 min and killing is already effective within 30 min.
N’Diaye et al. showed by fluorescent microscopy uptake of as many as 40% of fluorescein isothiocyanate (FITC)-labeled MAC by human neutrophils within 40 min. In their experiments serum-opsonized zymosan and MAC was engulfed even up to 90% [24 ]. However, they used a higher cell/bacteria ratio of 1:50 and FITC labeling alone may opsonize the mycobacteria as has been shown for Bordetella pertussis [27 ].
The data presented here clearly demonstrate that human neutrophils can phagocytize MAC in autologous plasma in the absence of added opsonins. Phagocytosis occurs within 15 min and the phagocytized bacteria are nearly completely killed by 120 min. The flow cytometry method to measure phagocytosis of Auramine-stained mycobacteria is rapid and semiquantitative and can be used to analyze not only peripheral blood cells but also those in bronchoalveolar lavage fluid. It should be useful in investigating the influence of cytokines and to help to illuminate the cooperation of human neutrophils and monocyte/macrophages in the defense against mycobacterial infections.
|
ACKNOWLEDGEMENTS |
|---|
Received February 15, 2000; revised October 25, 2000; accepted October 27, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. Sato, J. Schorey, V. A. Ploplis, E. Haalboom, L. Krahule, and F. J. Castellino The Fibrinolytic System in Dissemination and Matrix Protein Deposition During a Mycobacterium Infection Am. J. Pathol., August 1, 2003; 163(2): 517 - 531. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Perskvist, M. Long, O. Stendahl, and L. Zheng Mycobacterium tuberculosis Promotes Apoptosis in Human Neutrophils by Activating Caspase-3 and Altering Expression of Bax/Bcl-xL Via an Oxygen-Dependent Pathway J. Immunol., June 15, 2002; 168(12): 6358 - 6365. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
