Attachment and Ingestion Stages of Bacterial Phagocytosis

Transcription

1 JOURNAL OF CLINICAL MICROBIOLOGY, Jan. 1988, p /88/ $02.00/0 Copyright C 1988, American Society for Microbiology Vol. 26, No. 1 Enzyme-Linked Immunosorbent Assay for Quantitation of Attachment and Ingestion Stages of Bacterial Phagocytosis ABED ATHAMNA AND ITZHAK OFEK* Department of Human Microbiology, Sackler School of Medicine, Tel Aviv University, Received 10 June 1987/Accepted 30 September 1987 Tel Aviv 69978, Israel Research on phagocytosis of bacteria is often hampered by the inability to distinguish quantitatively between bacteria that have been ingested by phagocytic cells and those which are attached to the surface of the cells. A method using the enzyme-linked immunosorbent assay technique to simply and accurately measure the rate of bacterial ingestion by phagocytic cells is described. The method is based on the ability of antibacterial antibodies to bind to bacteria attached to but not internalized by phagocytic cells. The attached bacteria were quantitated by enzyme-linked immunosorbent assay. Compared with the number of bacteria at zero time (17 bacteria attached per phagocyte) only 10 to 20% of the bacteria remained attached to phagocytic cells after incubation for 30 min at 37 C. The decrease in detected attached bacteria at 37 C was due to internalization of the bacteria by phagocytic cells, since upon disruption of the monolayer, most of the ingested bacteria were recovered, and at 4 C, most of the bacteria remained extracellularly attached. The proposed attachment and ingestion assay is easy to perform, allows the detection of specific attachment of test bacteria, and provides objective quantitation of attached and ingested bacteria. Most importantly, the assay allows testing of ingestion rates of bacteria under many variables on the same day. Phagocytosis is one of the most important host defense mechanisms against microbial infections (3). Phagocytic tests are widely used to characterize the molecular mechanisms of phagocytosis of many infectious agents by diverse types of phagocytes as well as to estimate the function of phagocytes of a patient in a clinical setting. The phagocytosis process includes three consecutive stages: attachment, ingestion, and killing (7). In several studies, attempts were made to separately measure the attachment and the ingestion stages as well as the intracellular killing of bacteria. In these studies, an antimicrobial agent was used to kill extracellular bacteria and bacteria attached to polymorphonuclear leukocytes (8, 10, 15, 17, 19). However, the antimicrobial agents used to eliminate extracellular bacteria may bind and penetrate phagocytes (20). Other studies used metabolically radiolabeled bacteria with [3H]uracil, which does not penetrate phagocytic cells and thus permits separate enumeration of the attached and ingested bacteria by pulse-labeling the bacterium-phagocyte mixture, followed by recording the decrease in uptake of radioactivity (5). This method, however, requires metabolically active bacteria, a situation that does not always concide with the expression of certain specific components which mediate attachment of the organisms to phagocytes. Furthermore, bacteria attached to but not ingested by phagocytes may have altered activity to incorporate the radioactive label. Fluorescence-labeled antibodies, which do not penetrate phagocytes, were used to selectively label attached versus ingested bacteria (2, 6). The total number of bacteria in the Giemsa-stained monolayer and the fluorescence-labeled bacteria associated with phagocytes are enumerated under light and fluorescence microscopes, respectively. The method, however, is cumbersome and subjective. In this study, we describe an enzyme-linked immunosorbent assay (ELISA) to separately quantitate bacteria attached to phagocytic cells from bacteria ingested by such cells. In this method, bacteria extracellularly attached to the * Corresponding author. 62 phagocytic monolayer after various periods of incubation are quantitated by adding antimicrobial antibodies, which do not penetrate phagocytes, followed by adding horseradish peroxidase-labeled anti-rabbit immunoglobulin G and suitable substrate. Under these conditions, we show that the decrease in the amount of substrate reduced as a function of time of incubation of the bacterium-phagocyte mixture reflects internalization of bacteria by phagocytes. Furthermore, by combining the ELISA technique with viable count measurements, we were able to compare the rate of killing to the rate of ingestion. (The participation of Abed Athamna in this study was in partial fulfillment of Ph.D. requirements.) MATERIALS AND METHODS Bacteria. A kanamycin- and cephalothin-resistant isolate of Klebsiella pneumoniae from urine was used. This strain possesses two phenotypes, one expressing mannose-binding activity and the other lacking such activity (9). The absence and the presence of the activity in the two phenotypes were confirmed by aggregation with mannan-containing yeast cells as described previously (12). The various phenotypes were grown in brain heart infusion broth (Difco Laboratories, Detroit, Mich.) for 48 h at 37 C under static conditions. The bacteria were harvested by centrifugation and washed in phosphate-buffered saline (PBS; 0.1 M NaCI, 0.02 M P04 [ph 7.2]). Stock solutions of the bacteria were stored at -70 C, thawed, washed once with PBS, and adjusted photometrically to the appropriate density in the required medium before use in the phagocytosis assay. The concentration of bacteria in the suspension was determined by direct counts in a Petroff-Hausser chamber (C. A. Hausser and Son, Philadelphia, Pa.) as described previously (11). Bacteria were preopsonized with specific antibodies, if required, as described elsewhere (1). Isolation of mouse peritoneal macrophages. The collection of peritoneal macrophages was performed as described previously (2). Briefly, mouse peritoneal macrophages were

2 QUANTITATION OF BACTERIAL PHAGOCYTOSIS BY ELISA VOL. 26, collected from ICR strain male mice weighing 20 to 25 g. The cells were harvested by washing the peritonea of mice with a cold solution of M-199; they were centrifuged (400 x g for 5 min), suspended in M-199 medium, and counted in a hemacytometer. The percentage of macrophages was estimated to be more than 95%. The cell suspension was then adjusted to the desired concentration before use in the phagocytosis assay. Antibodies. K. pneumoniae bacteria were suspended in saline to a concentration of 10'0/ml and killed by heating for 1 h at 60 C. The heat-killed bacteria were used as vaccines to inject rabbits. The animals were injected intravenously three times a week with increasing amounts from 0.2 to 1.0 ml of the vaccines for 4 weeks. The rabbits were sacrificed 8 days after the last injection, and blood was collected by intracardiac aspiration. All antisera were heat inactivated before being used to pretreat bacteria. Determination of attachment, ingestion, and intracellular killing. Equal volumes of 2 x 106 macrophages per ml and 5 x 108 preopsonized or nonopsonized bacteria per ml in M-199 solution supplemented or not supplemented with the inhibitory sugar methyl-a-d-mannoside were mixed in a tube. The mixture was rotated end over end at 4 C (25 rpm) for 30 min, washed four times by centrifugation at 300 x g for 5 min to remove nonattached bacteria, and suspended in M-199 medium. The number of any remaining nonadherent bacteria in the bacterium-phagocyte mixture was insignificant in samples taken from the last wash as determined by the ELISA described below. The washed phagocyte-bacterium mixture was shifted to 37 or 4 C, and timed experiments for the determination of the ingestion and killing rates of bacteria attached to macrophages were performed by withdrawing 100 pli of samples for each determination at various times. To estimate the ingestion rate, one of the timed 100-pul samples was distributed into each well of a microtiter plate (96-well, flat-bottomed microtiter plates; Linbro/Titertek; Flow Laboratories, Inc., McLean, Va.). The microtiter plate was centrifuged, and the sedimented phagocyte-bacterium suspension was fixed with methyl alcohol for 10 min. Bacteria extracellularly attached to the monolayer of phagocytes were quantitated by the ELISA technique as described previously for bacterial attachment to animal cells (13, 18) with minor modifications. Briefly, to each well was added 100 pi of specific antibacterial serum diluted 1:500 in 5% bovine serum albumin containing 10 pug of human immunoglobulin G (Sigma Chemical Co., St. Louis, Mo.) per ml to block the Fc receptor of the macrophage. After incubation for 30 min at 37 C, the monolayers were washed five times with PBS, followed by the addition of 100 pul of peroxidaselabeled horseradish anti-rabbit immunoglobulin G diluted 1:5,000 in PBS-5% bovine serum albumin for 30 min at 37 C. After five washes with PBS, 100 pul of a substrate, 2,2'-azinodi-(3-ethylbenthiazoline sulfate), was added to each well. The color was allowed to develop for 15 to 30 min and was then read at 405 nm with a microtiter plate reader (Dynatech Industries, Inc., McLean, Va.). Controls consisted of wells without bacterium-phagocyte suspension or wells with macrophages lacking bacteria to ensure that the macrophages did not cross-react with the first antibody and to determine the nonspecific binding of antibodies. In addition, we determined the effect of products released by the phagocytebacterium mixture on bacterial surface antigens which react with the first antibody. For this purpose, nonconditioned medium or conditioned medium derived from the phagocytebacterium mixture preincubated at 37 C for 30 min was added to wells containing dried bacteria (see below) and incubated at 37 C for 30 min, followed by determination of the number of bacteria by ELISA. A standard curve made for each test served to derive the number of bacteria per monolayer. For this purpose, bacteria of known concentrations in distilled water were allowed to dry overnight in wells of the microtiter plate. The ELISA test was performed on the immobilized bacteria as described above. The ELISA values in OD405 units (i.e., optical density at 405 nm) were plotted as a function of the number of bacteria in each well. The curve obtained was used to calculate thé number of bacteria attached to the phagocytic cell monolayer from the ELISA values obtained in the test experiment. The standard curve was adjusted for loss of dried bacteria as follows. To determine loss of dried bacteria due to washings, cells of K. pneumoniae were grown in broth containing'10,uci of ['4C]glucose (Amersham Corp., Buckinghamshire, England) per ml for 48 h at 37 C under static conditions. Thé radiolabeled bacteria were harvested by centrifugation and washed free of excess radioactivity in PBS. The bacteria were adjusted as described above to the desired concentration and counted in a, counter (1211 RackBeta; LKB Instruments, Inc., Rockville, Md.). The radiolabeled bacteria contained 4,500 cpm/107 bacteria. The bacterial suspension was diluted in distilled water and dried into two sets of flat-bottomed Immulon Removawell strips (Dynatech). One of the two sets was washed six times, and the radioactivity of individual wells of the two sets was counted. After the determination of the number of bacteria was completed, the same monolayers were washed and stained to quantitate the number of phagocytes. The determination of the number of phagocytes per well was based on selective staining of the phagocyte nuclei with methylene blue, followed by extraction of the stain as described elsewhere for tissue culture cells (4). Briefly, monolayers were stained with 100 pil of 1% methylene blue solution for 10 min, followed by washing with boric acid buffer (ph 8.6). The stain was extracted by adding 0.1 N HCI and reading at 620 nm in a microtiter plate reader (Dyn'atech). A standard curve made for each test served to derive the number of phagocytes per monolayer. For this purpose, various concentrations of macrophages in 100 pil of M-199 medium were sedimented in wells of a microtiter plate, fixed with methyl alcohol, and stained as described above. The OD620 values of the extracted stain were plotted as a function of the number of phagocytes in each well to obtain a standard curve. This curve was used to estimate the number of phagocytes in each 'experimental well after extracting and reading the methylene blue stain from each test monolayer. To estimate the intracellular killing, the second 100-pl timed sample was placed on a MacConkey agar plate containing 25,ug of kanamycin and cephalothin per ml. The number of viable bacteria was calculated from the number of CFU after overnight incubation at 37 C. Chemicals. All of the chemicals and reagents used in this study were from commercial sources of the highest purity available. RESULTS Standard curves of bacteria and macrophages. In a separate set of experiments, ['4C]glucose-radiolabeled bacteria were dried on the bottom of microtiter plate wells to see whether washings of the wells removed dried bacteria. We found that there was about a 20% loss of counts per minute after

3 64 ATHAMNA AND OFEK washing (six times) of the wells to which 107 bacteria were added (4,300 and 3;500 cpm before and after washing, respectively). There was a negligibie loss of bacteria due to washing in weils to which bacteria in the range of 2 x 105 to 50 x 105 were added. Insignificant differences were observed between the number of dried bacteria (106 bacteria per well) treated with conditioned medium and bacteria pretreated with nonconditioned medium (ELISA values of and 0.69 ± 0.1, respectively), suggesting that putative products released during phagocytosis do not damage bacterial surface antigens that are recognized by the antibodies. A linear relationship between the number of immobilized bacteria on the wells of the microtiter plate and ELISA values was obtained within the range of 2 x 105 to 50 x 105 bacteria per well (Fig. la). Likewise, the OD620 of the extracted methylene blue stain of sedimented macrophages was linear in the range of 105 to 106 macrophages per well (Fig. 1B). These standard curves were made in each experiment and used to determine the number of bacteria per macrophage, which was calculated by dividing the total number of bacteria by the total number of phagocytes per well. Attachment of bacteria to macrophages. Mannose-specific lectin associated with type 1 fimbriae of enterobacterial species mediates attachment of bacteria to phagocytic cells (1, 2, 14, 16). To investigate whether the ELISA technique could be used to measure such specific attachment to macrophages, we used two K. pneumoniae phenotypes, one of which expresses type 1 fimbriae associated with mannosebinding activity and one of which does not (9). The attachment of fimbriated K. pneumoniae to peritoneal macrophages increased linearly with an increased number of bacteria in the suspension (data not shown) and reached 17 bacteria attached per macrophage at a 1,000:1 ratio of bacteria added per macrophage (Fig. 2). At this ratio, only 1 to 2 bacteria of the nonfimbriated phenotype attached per macrophage. The attachment of fimbriated K. pneumoniae was inhibited by 2.5% methyl-a-d-mànnoside, indicating that the binding is mediated by mannose-specific lectin associated with type 1 fimbriae. Ingestion and killing of bacteria by macrophages. The fate of bacteria attached to peritoneal macrophages at 4 C was monitored by shifting the temperature of incubation of the macrophage-bacterium mixture to 37 C. The number of bacteria attached to macrophages decreased from 17 at zero 1.0 r (A)! 0.75 o m w a.j ci r (B) o A B C D FIG. 2. Attachment of bacteria to macrophages. Mouse peritoneal macrophages were mixed with mannose-specific K. pneumoniae in the absence (A) or the presence (B) of methyl-ex-d-mannoside and with non-mannose-specific K. pneumoniae in the absence (C) or the presence (D) of methyl-at-d-mannoside at 4 C for 30 min. Nonattached bacteria were removed by differential centrifugation, and samples were withdrawn to quantitate the number of bacteria attached per macrophage, as described in Materials and Methods. Data are means of triplicates; vertical lines show the standard deviation. time to 8 and 4 bacteria per macrophage during 15 and 30 min of incubation, respectively (Fig. 3). To see whether this decrease in the number of attached bacteria was due to the internalization of the attached bacteria by the phagocytes during'the incubation period, we 4 performed parallel experiments in which the cell suspension was incubated at 4 C to prevent ingestion or was disrupted by distilled water after incubation at 37 C to detect intracellular bacteria. Compared with the number at zero time, there was no significant decrease in the number of bacteria attached per phagocyte in timed samples at 4 C, and most of the decrease in the number of bacteria attached per phagocyte during 15 min of incubation at 37 C was recovered in disrupted phagocytes. While most of the attached bacteria were ingested during 45 min of incubation at 37 C, only 65% of the ingested J. CLIN. MICROBIOL. à à à à 'à O O BACTERIA, 10I MACROPHAG, 10 FIG. 1. Standard curves for the determination of numbers of bacteria and macrophages. (A) ELISA values (OD405) as a function of increasing number of immobilized K. pneumoniae cells. (B) Values of extracted methylene blue stain (OD620) as a function of number of sedimented peritoneal macrophages. The assays were performed as described in Materials and Methods TIME (MIN) FIG. 3. Ingestion of bacteria attached to macrophages. A macrophagl-bacterium suspension preincubated 4nC at as described in Fig. 2 was shifted 37(D ()) to or left at4nc(s), and timed samples were withdrawn to determine extracellularly attached bacteria or intracellular bacteria (M ) after disruption with distilled water as described in Materials and Methods. Data are means of triplicates; vertical fines show the standard deviation.

4 VOL. 26, 1988 QUANTITATION OF BACTERIAL PHAGOCYTOSIS BY ELISA o 50 o TIME (MIN) FIG. 4. Ingestion and killing rates of bacteria by macrophages. A macrophage-bacterium suspension preincubated at 4 C as described in Fig. 2 was shifted to 37 C, and timed samples were withdrawn to determine killing (O) or ingestion (0) of bacteria by peritoneal macrophages as described in Materials and Methods. M, Ingestion of antibody-coated bacteria (opsonized). Data are means of triplicates; vertical lines show the standard deviation. bacteria were killed during 60 min of incubation at 37 C (Fig. 4). DISCUSSION The purpose of this study was to provide a useful method for measuring the various stages of bacterial phagocytosis in the same phagocyte-bacterium mixture. The method is based on the observation that extracellular attachment of bacteria to animal cells can be detected by the ELISA technique (13, 18). The ingestion and the attachment stages were separately measured by taking advantage of the fact that bacteria internalized by phagocytes no longer react with specific antibodies (2, 6). The facts that no ingestion of bacteria took place at 4 C and that at 37 C most of the ingested bacteria could be detected after disrupting the monolayer suggested that the ELISA method detects truly ingested bacteria. Nevertheless, the ELISA method cannot distinguish between partial ahd complete ingestion or between bacteria attached and internalized but accessible to the antibodies because of potential ingestion by phagosomes that are open to the outside; We used bacteria that attached to phagocytes via a mannose-specific lectin expressed on the bacterial surface. It has been shown that mannose-specific lectin-mediated interaction of Escherichia coli with phagocytes involves attachment to and ingestion and killing by the phagocytes (2, 14, 16). We could confirm these studies by the ELISA technique. Compared with the ingestion rate of antibody-coated bacteria, the ingestion rate of bacteria attached via mannose-specific lectin is somewhat lower. Phagocytosis measured by ELISA, as reported here, offers some advantages which make it suitable as the assay of choice in many research and clinical laboratories. First, and probably most important, many variables may be tested on the same day, e.g., type of phagocytic cells, type of bacterial strains, and various treatments or potential inhibitors. Second, the attachment is evaluated by the objective criteria of an ELISA, which avoid the use of radioactivity yet can achieve similar sensitivity; e.g., the method is sensitive enough to detect as few as 5 x 104 bacteria. Furthermore, the preparation of antibody specific for each test organism can be avoided by using antibodies against common surface antigens shared by a defined group of bacteria. In summary, the ELISA phagocytosis assay appears to be one test of choice when discrimination between the attachment and the ingestion stages in the phagocytosis process is needed. Moreover, the experiments presented here show that using the ELISA along with intracellular killing should lead to a better understanding of the relationship among the binding, ingestion, and killing stages of bacterial phagocytosis. ACKNOWLEDGMENT The excellent technical assistance of Rina Heiber is greatly appreciated. LITERATURE CITED 1. Bar-Shavit, Z., R. Goldman, I. Ofek, D. Mirelman, and N. Sharon Mannose residues on phagocytes as receptors for the attachment of Escherichia coli and Salmonella typhi. Biochem. Biophys. Res. Commun. 78: Bar-Shavit, Z., R. Goldman, I. Ofek, N. Sharon, and D. Mirelman Mannose-binding activity of Escherichia coli: a determinant of attachment and ingestion of the bacteria by macrophages. Infect. Immun. 29: Bellanti, J. A Mechanisms of immunity to bacterial disease, p In J. A. Bellanti (ed.), Immunology III. The W. B. Saunders Co., Philadelphia. 4. Bracha, R., and D. Mirelman Virulence of Entamoeba histolytica trophosoites. Effect of bacteria, microaerobic conditions and metronidazole. J. Exp. Med. 160: Cohen, P. S., S. Renate, J. Syter, and K. Vosbeck Quantitation of the rate of the ingestion of Escherichia coli by human polymorphonuclear neutrophils using 3H-uracil. J. Microbiol. Methods 3: Dilworth, J. A., J. O. Hendly, and G. L. Mandell Attachment and ingestion of gonococci by human neutrophils. Infect. Immun. 11: Jonston, R. B., Jr The host response to invasion by Streptococcus pneumoniae: protection and the pathogenesis of tissue damage. Rev. Infect. Dis. 3: Maaloe, O On the dependence of the phagocytosis stimulating action of immune serum on complement. Acta Pathol. Microbiol. Scand. 24: Maayan, M. C., I. Ofek, O. Medalia, and M. Aronson Population shift in mannose-specific fimbriated phase of Klebsiella pneumoniae during experimental urinary tract infection in mice. Infect. Immun. 49: MacKaners, J. B The phagocytosis and inactivation of staphylococci by macrophages of normal rabbits. J. Exp. Med. 112: Ofek, I., and E. H. Beachey Mannose binding and epithelial cell adherence of Escherichia coli. Infect. Immun. 22: Ofek, I., and E. H. Beachey Bacterial adherence. Adv. Intern. Med. 25: Ofek, I., H. S. Courtney, D. M. Schifferli, and E. H. Beachey Enzyme-linked immunosorbent assay for adherence of bacteria to animal cells. J. Clin. Microbiol. 24: Ohman, L., J. Hed, and O. Stendahi Interaction between human polymorphonuclear leukocytes and two different strains of type 1 fimbria bearing Escherichia coli. J. Infect. Dis. 146: Ran, J. A., C. Watanakunakorn, and J. P. Phair A modified assay of neutrophil function: use of lysostaphin to differentiate defective phagocytosis from impaired intracellular killing. J. Lab. Clin. Med. 78: Silverblatt, F. J., J. S. Dreyer, and S. Schauer Effect of pili on suspectibility of Escherichia coli to phagocytosis. Infect. Immun. 24:

5 66 ATHAMNA AND OFEK 17. Solberg, C. O Protection of phagocytized bacteria against antibiotics. Acta Med. Scand. 191: Stanislawski, L., W. A. Simpson, D. Hasty, N. Sharon, E. H. Beachey, and I. Ofek Role of fibronectin in attachment of Streptococcus pyogenes and Escherichia coli to human cell lines and isolated oral epithelial cells. Infect. Immun. 48: Van der Broek, P. I., F. A. M. Dehn, P. C. Leijh, M. P. Van der J. CLIN. MICROBIOL. Barsehaar, and R. van Furth The use of lysostaphin in in-vitro assays of phagocyte function: adherence to and penetration into granulocytes. Scand. J. Immunol. 15: van Furth, R., L. van Zwet, and P. C. Leijh In vitro determination of phagocytosis and intracellular killing by polymorphonuclear and mononuclear phagocytes, p In D. M. Wier (ed.), Handbook of experimental immunology, vol. 2. Blackwell Scientific Publications, Ltd., Oxford.