- Top of page
- Materials and methods
Objective Polymorphonuclear leukocytes (PMN) play a central role in the elimination of most extracellular pathogens, and an impairment of their functions predisposes an individual towards local and systemic bacterial and fungal infections. Here we describe a rapid and easy-to-perform cytofluorometric assay for investigation of PMN activity using Candida albicans and Staphylococcus aureus as target organisms.
Methods Phagocytes were stained with anti-CD13-RPE antibody, and microorganisms were stained with calcein-AM. Oxidative burst production was measured by oxidation of dihydroethidium. The percentage of killed target organisms after ingestion was determined by staining with ethidium-homodimer-1 after lysis of human cells. The dyes and procedures used in this method were chosen after comparison of different stains and cell preparation techniques described in previous assays.
Results Concerning phagocytosis, the percentages of active phagocytes and of ingested microorganisms were determined. Furthermore, the method allowed measurement of the resulting percentage of PMNs producing respiratory burst, and of the percentage of killed microorganisms. We minimized artifactual changes, which might have been the reason for the difficulties and conflicting results of other cytofluorometric methods.
Conclusions The described method provides a new whole blood cytofluorometric assay, which combines rapid and simple handling with high reproducibility of results obtained by investigation of PMN activity using Candida albicans and Staphylococcus aureus as target organisms.
- Top of page
- Materials and methods
In recent years increasing interest has been shown in the evaluation of PMN functions, and many investigators have developed flow cytometric approaches to assay their activation. However, there are still surprisingly conflicting results regarding neutrophil phagocytosis and oxidative burst activity. We describe a simple and rapid assay, which reduced handling difficulties, allowed measurement at multiple time intervals of incubation and minimized artefactual changes. The assay was designed to investigate the phagocytosis of C. albicans and S. aureus, the resulting percentage of PMNs producing respiratory burst and the percentage of killed microorganisms.
Many of the procedures used for isolation of PMNs for analysis [5–7, 11] themselves cause artefactual changes such as altered expression of surface antigens , morphologic changes , aggregation and activation of PMNs  and reduction of the oxidative burst, chemotaxis and increased release of lysosomal enzymes . The use of whole blood for investigation reduced these effects.
Anticoagulation was performed using 5 IE of sodium-heparin/mL blood. EDTA, citrate as used by Model et al , and oxalate form complexes with calcium ions and therefore impair PMN functions . Fluoride inhibits enzymes involved in oxidative metabolism , and higher concentrations of heparin reduce phagocytosis and the respiratory burst .
For the provision of a sufficient nutrient supply to the cells during incubation, a percentage of 0.4% glucose in the RPMI solution proved necessary. The yeasts seemed to be particularly sensitive to stressful conditions and tended to develop pseudohyphae , which is why some investigators added 0.5 µg of amphotericin B per mL to the samples. This addition led, in our investigations, to an approximately 10% higher percentage of dead yeasts (data not shown), corresponding to the findings of Schaumann and Shah . Contrary findings, however, are reported by Martin and Bhakdi  and Nugent and Couchot , who excluded an effect of low concentrations of amphotericin B. In any case, our results indicate that, microscopically, there is no development of pseudohyphae after addition of 0.4% glucose.
For the staining of living bacteria and yeasts we used calcein-AM, which appeared to be most suitable after comparison of different dyes for microorganisms. Previously used fluorescence stains such as FITC [1–4,6,20] and BCECF-AM [5–11] showed a high sensitivity for pH changes, which occur in the phagolysosomes and cause a change in the emitted wavelength of light. Others, such as Texas Red , showed toxic effects on the microorganisms (data not shown). Additionally, calcein-AM stains only the cytoplasm of living cells, causes no damage to the cell membrane and has a high fluorescence and diffusion rate.
Our investigations confirmed the results of Perticarari et al  that no preopsonization of microorganisms was necessary, since the incubation time within whole blood during the test ensures sufficient opsonization of target cells for phagocytosis in a physiologic environment.
The staining of phagocytes was performed using a CD13 antibody linked to R-phycoerythrine (RPE), which is selective for granulocytes and monocytes. In contrast to other studies, we marked the phagocytes after the incubation with the microorganisms, because the staining of PMNs involves washing and centrifugation procedures, which affect the phagocytes [12–14]. Furthermore, the linking of antibodies to PMNs might itself cause an alteration of their functions. Fluorescence microscopic investigations with the chosen concentration of antibodies showed no antibody-capping, which is probably due to the low concentration of antibodies used and the presence of human plasma in the samples. Therefore, there was no necessity to remove surplus antibodies or to add bovine serum albumin.
The determination of the percentage of ingested microorganisms gives important additional information about the efficiency of PMNs. This applies even more for bacteria than for yeasts, because each phagocyte can ingest widely differing numbers of staphylococci during the incubation. To facilitate this measurement, it was necessary to omit the quenching procedure usually performed in cytofluorometric experiments.
Fluorescence microscopic investigations showed that, after incubation, an average of 6% of the phagocytes had microorganisms attached to their surface. We centrifuged samples after incubation with yeasts, as well as with bacteria at different accelerations (60–300g). The pellets were resuspended in PBS and again investigated microscopically. After centrifugation at at least 200g and storage of the samples on ice (which suppresses the mobility of cells) for 45 min, no attached microorganisms were detectable . Centrifugation at 250g is performed in any case in the course of the PMN antibody staining. Accordingly, no quenching procedure seemed necessary. We could nevertheless demonstrate that the fluorescence of calcein-AM could be suppressed by using 2 mg of azure A in PBS per sample for quenching after the incubation.
For the determination of the oxidative burst activity a similar principle is used in nearly all previously described assays: during the oxidative burst, a colorless dye is oxidized and is then detectable by the flow cytometer [1–4,9]. Dihydrorhodamine 123 (DHR), used by Rothe and Valet , Vowells et al  and Hirt et al , is a very sensitive intracellular dye, which, however, emits a green fluorescence and therefore interferes with the detection of the microorganisms in our assay. As it was an important aim to measure all parameters in one assay, a different dye had to be found. Model et al  used 4-carboxydihydrotetramethylrosamine succinimidyl ester (RS-SE) in a method for measuring only the oxidative burst. RS-SE showed the lowest non-specific oxidation and very sensitive detection of the PMN oxidative burst compared with DHR, DHE and dihydrotetramethylrosamine (RS) . However, due to its labeling of the plasma membrane and a consequent possible influence on PMN function, it could not be used. DHE was used by Perticarari et al  and its fluorescence is detectable in both the orange and red channels. Perticarari et al detected its fluorescence in FL2, which would interfere with the emission of the RPE dye of the phagocytes. In our method, therefore, the fluorescence was detected in FL3. We used a concentration 150 times lower than that of Rothe and Valet , which may account for the fact that, in contrast to their findings, no non-specific oxidation of DHE was observed .
For the determination of the percentage of dead microorganisms, it was necessary to lyse the phagocytes. In the recent literature different techniques are described, but most have proved to have major or minor disadvantages . Use of distilled water [22,23] led to complete lysis of granulocytes, while the monocytes were not affected. Gelatine , sodium deoxycholate (DOC) , lysolecithin  and digitonin  resulted in satisfactory lysis of phagocytes, but the concentrations required caused severe damage to the microorganisms. Especially for DOC, the findings stand in contrast to the results of Martin and Bhakdi , who did not, however, use counterstaining with a dye, especially for detection of dead microorganisms. The use of ultrasonics  gave satisfactory results but was too time-consuming for routine measurement. With the chosen combination of Triton X-100 and Tween-20, we obtained results corresponding in quality to those of ultrasonic treatment, with a minimum of handling difficulties and time consumption.
Martin and Bhakdi  determined the percentage of dead target cells by measuring the decrease in calcein-AM-fluorescence compared to living microorganisms. Our own experiments, however, showed that the kinetics of the killing were faster than the diffusion of the dye out of the damaged microorganisms. It was not possible to use propidium iodide , because this dye stained even living cells after the necessary duration of the killing assay. Finally, with adapted cytometric parameters, satisfactory results were obtained by using EthD-1, in contrast to the findings of Haugland , who stated that only mammalian cells could be stained with that dye.
The described assay, in contrast to previous investigations [1–7,31], focuses only on the presence of the different fluorescent signals in the specific tests rather than on the quantity of the detected fluorescence. Therefore, no statement could be made about a possible correlation between the degree of particle uptake and the degree of DHE oxidation. The reason for this is the use of living microorganisms, which, in contrast to, for example, latex particles, makes a calibration impossible. This is due, on the one hand, to the fact that the staining capacity of living microorganisms is a priori variable, and on the other hand, to the fact that the signal may vary as a consequence of diffusion of the dye or morphologic changes of the cells during the assay.
A typical feature of cytofluorometric investigations is the high statistical significance of the results obtained. For every single evaluated parameter we measured about 5000 PMNs. The intra-assay variation of the results is around 1–2%, obtained by investigation of six samples of the same donor, which were independently processed in parallel. The inter-assay variability of a single donor amounts to 5–10% between 10 blood samples investigated over a period of 4 weeks and is therefore normally significantly larger than the intra-assay variability. Concerning the phagocytosis, burst activity and killing, the inter-assay variations of 10 different donors were about 7–14% . It seems that individual variability influences the kinetics of the reaction as well as the end results.
For investigations with bacteria, an optimal target cell/PMN ratio of 50 : 1 had to be used. Furthermore, it proved necessary to perform incubation with the bacteria at 30 °C instead of 37 °C, because at 37 °C the immune reaction with staphylococci is too fast to obtain optimal kinetics. Despite the lower incubation temperature the assay with S. aureus still showed a quicker increase and a nearly two-fold higher final percentage of phagocytically active PMNs. The percentage of ingested microorganisms also showed a quicker increase with bacteria with similar final results for both microorganisms. These results might be due to the higher target-to-effector cell ratio and the resulting higher frequency of collisions between bacteria and PMNs. Furthermore, phagocytes can also ingest more than one yeast, and therefore the maximum achievable proportion of phagocytosing PMNs is already, on a statistical basis, lower when using a PMN/yeast ratio of 1 : 1. In addition, the smaller cell size of bacteria may accelerate the phagocytosis, and with that the percentage of phagocytosing PMNs as well as that of ingested staphylococci.
In contrast to previous assays [5–11], which determined the decrease of the non-phagocytosed target cells, we performed a direct measurement of the ingested microorganisms. The advantage is that perturbation of the measurement by debris is excluded. Investigation of the percentage of PMNs displaying oxidative burst activity after phagocytosis revealed a quicker increase and much higher final results using yeasts. A possible explanation might be the fact that C. albicans is exclusively killed by the oxidative burst , and other toxic substances tend to play no role . However, the lower incubation temperature used with bacteria might possibly affect the burst production, but, on the other hand, the fast kinetics of phagocytosis at the same temperature indicate that this may not be the only reason. Using Candida, the kinetics for the oxidative burst-producing PMN correspond to those of ingested yeasts ( Figure 3A). This implies that, after phagocytosis of yeasts, nearly all PMNs start the production of oxygen-dependent toxic substances, while after ingestion of bacteria, even if the oxidative burst takes place, the more important part of the destruction of target cells is mediated by non-oxygen-dependent mechanisms. This view is supported by the finding that the percentage of dead microorganisms showed a quicker increase and slightly higher final results for staphylococci. An additional explanation for these results lies in the observation of other authors [34,35] that, in contrast to bacteria, yeasts are not damaged in plasma, but only within phagocytes.
Our findings demonstrate the necessity of the simultaneous determination of all the investigated parameters and the importance of different target organisms to obtain objective results.