material, and neutrophil suspensions containing intracellular bacteria and ..... in the two fractions indicated a higher proportion of cells with hypersegmented.
Aust. J. Exp. Biol. Med. Sci., 63 (Pt. 4) 361-370 (1985)
©EVALUATION OF INTRACELLULAR KILLING OF BACTERIA BY ENRICHED POPULATIONS OF MOUSE PERITONEAL EXUDATE NEUTROPHILS by P. H. HART, L. K. SPENCER, P. J. MCDONALDAND J. J. FINLAY-JONES (From the Clinical Microbiology Unit, School of Medicine, Flinders University of South Australia, Bedford Park, South Australia 5042.) (Accepted for publication April 29, 1985.) Summary. Elicited, mouse peritoneal exudate cells were fractionated by centrifugation on discontinuous Percoll density gradients. TVo subpopulations of neutrophils, each of greater than 90% purity, were isolated at discontinuous density gradient interfaces different from the region of mononuclear ceU enrichment (i.e., 1-0694-1 0871 and 1-0872-1-1002 g/ml for neutrophils and less than 1-0694 g/ml for mononuclear cells). Peritoneal exudate cells were mixed with Proteus mirabitis in the presence of 1% normal mouse serum for 30 min. The mixtures were fractionated on gradients of Percoll diluted with a clacium-free medium. Populations of cells banding at densities greater than 1 -0693 g/ml were washed free of gradient material, and neutrophil suspensions containing intracellular bacteria and which were relatively free of extracellular bacteria were isolated. Less than 7% of the total bacteria present was extracellular. The continuing extracellular presence of a heat-labile component of normal mouse serum was essential for maximal intracellular kill of P. mirabitis by mouse peritoneal neutrophils.
INTRODUCTION The mouse is a commonly used species for the study of the pathogenesis of bacterial infections, yet studies of the interaction of bacteria with murine neutrophils, a first line of host defence, are not so numerous. The studies that exist have used peritoneal exudate cells of varying neutrophil content (Bjornson and Michael, 1971; Winkelstein, Smith and Shin, 1975; Baron and Proctor, 1982; Finlay-Jones et at., 1984) or whole blood (Jones and Dyster, 1973). Density gradient centrifugation has been used to prepare enriched neutrophil suspensions from murine peritoneal exudate cells using continuous gradients of Percoll (Watt, Burgess and Metcalf, 1979), discontinuous metrizamide (Lopez, Strath and Sanderson, 1981) or discontinuous Ficolldiatrizoate (Rabellino et al., 1978) and from blood using discontinuous gradients of Ficoll-diatrizoate (Penttila, O'Keefe and Jenkin, 1982). Discontinuous Percoll gradients have been used to prepare, from human blood, neutrophils whose chemotactic and bactericidal activities were unimpaired (Dooley, Simpson and Meryman, 1982). We have modified this technique to allow enrichment of Abbreviations used in this paper: cfu, colony forming units; HBSS, Hanks' Balanced Salt solution, Ca** and Mg** free; i.p., intraperitoneal; MPBS, Mouse osmolality phosphate-buffered saline; NMS, Normal mouse serum; RPMI, RPMI 1640 medium supplemented with 20 mM Hepes, pH 7-2.
362
P- H. HART, L. K. SPENCER, P. J. McDONALD AND J. J. FINLAY-JONES
neutrophils from murine peritoneal exudates, and, in addition, have used it for the separation of neutrophils, a proportion of which have intracellular bacteria, from extracellular bacteria in bacterium-neutrophil mixtures. This has assisted the study of intracellular killing by neutrophils as a process distinct from phagocytosis. MATERIALS AND METHODS Mice Male mice of the BALB/c strain aged 4-12 weeks were used. They were housed according to the guidelines of the National Health and Medical Research Council and the Commonwealth Scientific and Industrial Research Organisation of Australia. Bacteria The test organism, P. mirabilis, was isolated from intra-abdominal abscesses induced in mice by a complex inoculum of mouse colonic and caecal contents (Nulsen et al., 1983) and identified by the API 20E system (Montalieu-Vercieu, France). Log-phase bacteria were prepared as previously described (Finlay-Jones et ai, 1984). The viable count was assessed from predetermined standard curves by measuring absorbance at 420 nm. Enumeration of viable bacteria was performed by spreading 0-1 ml volumes of appropriately diluted bacterial samples on C.L.E.D. agar (Oxoid, Basingstoke, England) and counting of colonies after an overnight aerobic incubation at 37°. Sera Blood from the retro-orbital venous plexus of normal mice was collected under anaesthesia. The serum obtained was pooled, filtered and frozen at -70°. Heated normal mouse serum was prepared by incubating this serum at 56° for 30 min. Using an indirect immunofluorescence assay (Finlay-Jones et ai, 1984), seven preparations of normal mouse serum (NMS) were shown to contain negligible amounts of anti-/? mirabitis antibody (mean titre = 1/13, Range: not detected in undiluted serum—1/4). Opsonisation of P. mirabilis 1 X 10' f! mirabilis suspended in 1 ml RPMI 1640 medium supplemented with 20 mM Hepes, pH 7-2 (RPMI) (Flow Laboratories, McLean, VA, U.S.A.) were incubated with 1 ml NMS in a 37° reciprocating water bath for 30 min. Bacteria were washed twice with 0-9% saline before dilution and incubation with phagocytic cells. Peritoneal exudate cell collection Mice were injected intraperitoneally (i.p.) with 1 ml Brain Heart Infusion Broth (Oxoid, Basingstoke, England) prepared with tap water. After 3-5 h the mice were sacrificed by cervical dislocation and the peritoneal cavities were washed out initially with 5 ml and then 3 ml ice-cold mouse osmolality phosphatebuffered saline (MPBS) (Sheridan and Finlay-Jones, 1977). Harvested cells were washed twice with MPBS (175 g, 7 min, 4°) with 1 ml underlays of foetal calf serum (Flow Laboratories, North Ryde, Australia) before resuspension in Hanks' Balanced Salt Solution, Ca** - and Mg** - free, (HBSS—Commonwealth Serum Laboratories, Melbourne, Australia) supplemented with 10 mM Hepes, pH 7-2 at approximately 5 X 10' cells/ml. Contaminating erythrocytes were not lysed. Preparation of peritoneal exudate cells with intraceltular bacteria Peritoneal exudate cells (5 x 10' cells/ml HBSS) were incubated with non-opsonised P. mirabilis (1 X 10' bacteria/ml) in the presence of 1% (v/v) normal mouse serum for 30 min (i.e., bacterium: peritoneal cell ratio = 2:1). A 1% NMS supplement was used because phagocytosis was thereby associated with minimal intracellular killing. The 90 x 13 mm polycarbonate assay tubes (Disposable Products, South Australia) were incubated at 37° on a rotating platform (Nutator: Clay Adams, Becton, Dickinson and Co., Parsippany, NJ, USA) at 20 turns/min. At 0 and 30 min, 0-05 ml samples were taken for the determination of viable organisms (to assess whether any killing had occurred during the period of phagocytosis) and for the preparation of cytocentrifuge smears. The latter allowed a visual inspection of cell morphologies as well as a measurement of the percentage of phagocytic cells with internalized bacteria. After 30 min
ACTIVITIES OF MOUSE NEUTROPHILS
363
the assay tubes were placed on ice before dilution with cold Hepes-buffered HBSS and addition to HBSSdiluted Percoll to make a 45% solution (density = 1-0575 g/ml). Percott sotutions Percoll (density 1-130 ± 0-005 g/ml, osmolality 20 mmol/kg H^O, viscosity 10 ± 5 cP at 20° Pharmacia Fine Chemicals AB, Uppsala, Sweden) was diluted as described by Dooley et at. (1982). A stock solution labelled 100% was prepared by mixing 9 volumes of Percoll with 1 volume of 10 x HBSS. To obtain mixtures of the desired densities, the stock solution of Percoll was diluted with Hepes-buffered HBSS. Mixing ratios were calculated according to the equation of Ulmer and Flad (1979), i.e., density (g/ml) = (% of Percoll stock solution x 0-001186) -t- 1-0041. The stock solution was diluted to solutions of 81%, 70%, 55%, 50% and 45% Percoll with densities of 1-1002, 1-0871, 1-0693, 1-0634 and 1-0575 g/ml, respectively. The Percoll solutions had osmolalities of 325, 316, 312, 305 and 300 mmol/kg HjO, respectively, as determined using a freezing point depression Osmometer (Model 3W, Advanced Instruments, USA). Isosmolar solutions were prepared by dilution of the Percoll stock solution with Hepes-buffered HBSS to a concentration of 81%; further dilutions were performed with MPBS (333 mmol/kg H^O) (Sheridan and Finlay-Jones, 1977). Density gradient centrifugation Preformed gradients were prepared in polycarbonate 100 x 16 mm tubes (Disposable Products, South Australia). Three ml of the 81% Percoll solution were added to each tube. Using a pasteur pipette and without disturbance of the interface, 2 ml of the 70% Percoll mixture were added, followed by 2 ml of each of the 55% and 50% mixtures. Finally, 3 ml of a 45% Percoll mixture containing the cell suspension to be fractionated were loaded on top of the gradient (approximately 5 x 10' cells per gradient). The tubes were centrifuged at 1600 g for 30 min at 10° (IEC PR-6000, MA, USA). Cell bands (uppermost first) were harvested from each gradient with pasteur pipettes. Gradient material between cell bands was discarded if not turbid. Cell yields, recoveries and biological activities were evaluated after two washes with MPBS (175 g, 4°, 7 min). The ratio of nuclear area to nuclear shape of harvested cells was calculated using the methods of Jarvis (1981). The distribution of density marker beads (Pharmacia Fine Chemicals AB, Uppsala, Sweden) was examined in separate gradients. Bacterial suspensions were fractionated on similar gradients. After centrifugation, gradient materials of different densities were diluted and centrifuged (1500 g, 4°, 15 min). Volumes (0-1 ml) of diluted samples of resuspended pellets were spread on C.L.E.D. agar and colonies counted after overnight aerobic incubation at 37°. Assay of phagocytie kitting Percoll gradient enriched neutrophils (5 x 10') resuspended in RPMI supplemented with 2% NMS were placed into 90 x 13 mm polycarbonate tubes. Cell suspensions were prewarmed to 37° before addition of 1 X iti' P. mirabitis that had been opsonised in 50% NMS. The final volume in each tube was 1 ml. The use of 50% NMS to preopsonise bacteria and of 2% NMS in the assay mixture ensured adequate phagocytosis and killing of bacteria. The assay tubes were mixed by use of a rotating platform (Nutator) at 20 turns/min at 37°. Incubations continued for 90 min; samples for enumeration of viable P. mirabitis were withdrawn at regular intervals (every 30 or 45 min). As previously described (Finlay-Jones et at., 1984), assay samples (0-05 ml) were added to 5 ml of 0-1% Triton X-100 (Ajax Chemicals, Sydney, Australia) in 0-9% saline at room temperature for leucocyte lysis. After appropriate dilution in saline, 0-1 ml volumes were spread on C.L.E.D. agar and colonies counted after overnight aerobic incubation at 37°. Changes in the concentration of viable P. mirabitis over a 90 min period were calculated for each tube as follows: Alog|j| (cfu/ml) = [logji, (cfu/ml) at time t] - [log|^ (cfu/ml) at time 0] Assay of intracettutar kitting fottowing phagocytosis This assay was used to quantify intracellular killing in the absence of ongoing phagocytosis. 5 x IO* Percoll-fractionated, twice-washed neutrophils with intracellular bacteria were incubated with 10% (v/v) serum in a final volume of 1 ml RPMI. Incubations were performed at 37° in 90 x 13 mm polycarbonate tubes on a rotating platform (Nutator) at 20 turns/min. For enumeration of viable intracellular bacteria, samples of 0-05 ml were withdrawn at zero time, i.e., when all suspensions were removed from ice. Further samples were taken either after 45 and 90 min or after 30, 60 and 90 min, ahd, following lysis with 0-1% Triton X-100, diluted and plated on C.L.E.D. agar as described above.
364
P. H. HART, L. K. SPENCER, P. J. McDONALD AND J. J. FINLAY-JONES
Determination of extracellular bacteria Triplicate samples of 0-2 ml were placed into microtitre wells (Linbro, VA, USA) and 0-2 ^Ci 'Hthymidine (Amersham International, England) in a volume of 0-02 ml added. After 30 min incubation at 37° the cell and bacterial suspensions were harvested with 5% trichloroacetic acid on to glass fibre paper using a cell harvester (Titertek, Skatron, Norway). After washing with H^O, then methanol, the filters were dried at 60° and the radioactivity determined using ACS II scintillation fluid (Amersham) and a Mark II liquid scintillation system (Searle Analytic Inc., IL, USA). A standard curve of radioactivity incorporated vs. viable count was constructed using neutrophil-free bacterial suspensions. Expression of results Mean values are given ± one standard deviation. The numbers of observations are indicated. The Student's t test was used to determine the significance of differences between means.
RESULTS Fractionation of peritoneal exudate cells on Percoll Peritoneal exudates harvested 3-5 h after i.p. injection of 1 ml brain heart infusion broth contained a mean of 66% neutrophils (Range 54"yo-75%, n = 19). Centrifugation of these cell suspensions on preformed Percoll density gradients of varying osmolality resulted in the isolation of five well-differentiated bands of cells. One band was harvested from the top of the gradient, the others contained the cells banding at the interface between Percoll solutions of different densities. The pellets contained erythrocytes and negligible numbers of other cells. Table 1 shows a typical fractionation profile. Banding profiles of density marker beads were consistent with the densities of the Percoll solutions used, and with the maintenance of a discontinuous gradient during centrifugation. TABLE I Fractionation profile of celts harvested from elicited peritoneal exudates of mice and fractionated on a discontinuous Percolt density gradient of varying osmolality. Percofl Density* (g/ml) f-0575 f-0634 f-0693 f-0871 f-1002
Differential count (%)
Osmofality
Fraction (mosM/kg 1 300 2 305 3 312 4 3f6 5 325 Unfractionated •maximum density of cells.
up
Proportion of Recovered cells Neutro- Monocytes Macrophils lymphocytes phages Mast cells 5 19 49 32 — 3 6 58 36 — 7 8 81 ff — 58 89 fl 27 — 1 98 1 6f 26 f2 1
The viability of both mononuclear and polymorphonuclear cells after Percoll fractionation as judged by exclusion of trypan blue was greater than 95%. The recovery of cells after fractionation on Percoll gradients and subsequent washing x 2 to remove the gradient material was 62 ± 9% (n = 8). Two cell fractions enriched for neutrophils were isolated in the higher density region of the gradients, i.e., fractions 4 and 5. By examination of cytocentrifuge smears, studies of nuclear morphology of neutrophils isolated in the two fractions indicated a higher proportion of cells with hypersegmented nuclei in fraction 5. Approximately equal numbers of neutrophils were isolated in these two fractions. This was not the case with peritoneal exudate cells that
ACTIVITIES OF MOUSE NEUTROPHILS
365
had been briefly exposed to P. mirabilis before Percoll density gradient centrifugation. A greater proportion of neutrophils was found in fraction 4. For all functional assays, fractions 4 and 5 were combined to yield a viable, approximately 90% pure, neutrophil population. There was a 70 ± 16% (n = 9) recovery of neutrophils apphed to the gradients in these two bands. Functional activity of Percoll-fractionated neutrophils We were concerned that the fractionation procedure might modify the ability of neutrophils to phagoeytose and kill. Assays of phagocytie killing of P. mirabilis indicated, however, that enriched neutrophil suspensions from Percoll gradients were fully functional. A representative experiment is shown (Table 2). The unfractionated population, the Percoll gradient-derived neutrophil-enriehed subpopulations, and the remixed population (neutrophilenriched plus neutrophil-depleted subpopulations from Percoll gradients) all phagocytosed and killed the organism, with greater killing noted with higher neutrophil proportion. TABLE 2 Phagocytie kitting of opsonised P. mirabilis by cett mixtures prepared by combining non-fractionated cett suspensions with cetIs from neutrophit-enriched fractions harvested from Percott density gradients. Peritoneal exudate cells' 7-6 X 10' unfractionated 5-5 X 10'PercoU-fractionatedt 7-6 X 10' remixed^: ' t t §
Neutrophils (% of total cells)
A log|g cfu/ml§
66 88 68
-1-7 -3-1 -2-3
final concentration of neutrophils: 5-0 ± 0-2 x 10' cells/ml. cells from fractions 4 and 5. all fractions from the Percoll gradients were recombined. change in the concentration of P. mirabitis [log^^ (cfu/ml)] after 90 min.
Preparation of neutrophils with intracellular bacteria P. mirabilis (1 x lO^/ml), peritoneal exudate cells (5 x 10^/ml) and Hepes-buffered HBSS containing 1% NMS were mixed for 30 min at 37°. A serum supplement of 1% ensured phagocytosis occurred without intraeellular killing. The viable bacterial count in the mixture increased by an average 0-2 log,o cfu/ml (Range 0 - 0-3 log,o cfu/ml, n = 6). At the end of the incubation period, 38 ± 12% (n = 6) of neutrophils contained membrane-associated or intracellular bacteria, each with a mean load of 2-7 ± 0-5 P. mirabilis. The neutrophil-baeteria suspensions were then loaded on to Percoll gradients. After centrifugation the percentage of neutrophils in fractions 4 and 5 with intracellular bacteria was 39 ± 14% (n = 6). Thus, intracellular bacterial loads remained approximately constant during fractionation on Percoll at 10°. The average number of intracellular or cell membrane-associated bacteria after fractionation was 2-3 + 0-4 (n = 6). The banding profiles of bacterial suspensions in Percoll gradients were determined. When bacterial suspensions (5 x 10* P. mirabilis) were fractionated on discontinuous Percoll gradients, 66 ± 3% (n = 3) of the bacteria recovered were isolated in fraction 4. Serum-opsonised and nonopsonised bacteria showed similar banding profiles.
366
P. H. HART, L. K. SPENCER, P. J. McDONALD AND J. J. FINLAY-JONES
Extracellular bacteria in Percoll-fractionated neutrophil suspensions As ^H-thymidine does not cross the neutrophil membrane (Braconier and Odeberg, 1979), we used incorporation of ^H-thymidine to estimate the extracellular bacteria in suspensions of bacteria and neutrophils with intracellular organisms. Other parameters measured were cell-associated bacteria (in cytocentrifuge smears) and total viable bacteria (by determining cfu of bacteria in lysed neutrophil suspensions). These determinations were made at various stages in the preparative procedure. In a representative experiment 56-2% of the organisms in the mixture of bacteria and peritoneal cells loaded on to a Percoll gradient were extracellular. In the combined fractions 4 and 5 after centrifugation and two washings to remove gradient material, this was reduced to 6-5% of total. In 4 experiments estimates of the extracellular bacteria, as a proportion of total, after Percoll fractionation and washing, ranged from 2-4Vo to 6-9%. These results were significantly better than with the use of differential centrifugation without prior Percoll gradient fractionation to reduce the numbers of extracellular bacteria, in which the latter was in excess of of total. TABLE 3 Intracellular killing of P. mirabiUs following phagocytosis by neutrophil-enriched leucocytes. Intracellular killing' Concentration of No. of Range Mean ± S.D. experiments extracellular serum 5 tO% normal mouse serum 5 10% heated normal mouse serum 3 No serum • decrease in the concentration of P. mirabilis [- Alog,^
19 ± 0-4 0 9 ± 0-2 01 ± 0 3
15 -- 2 4 0 6 -- 11 ( - 0 2) - 0 2
(cfu/ml)] after 90 min.
Assessment of intracellular killing of P. mirabilis after termination of phagocytosis Neutrophils (5 x 10*) harvested from Percoll density gradients and containing intracellular bacteria previously phagocytosed in the presence of 1% NMS were placed in tubes at 4° and various extracellular serum supplements added. Changes in the concentration of P. mirabilis, measured 90 min after incubation of these tubes at 37°, are shown in Tkble 3. The killing of intracellular P. mirabilis in the presence of 10% heated normal mouse serum was significantly less than that measured in the presence of an equal amount of non-heated mouse serum (p < 0-001); however, the killing was significantly greater than the kilUng measured in the absence of serum (p
