S657
Adherence of Helicobacter pylori to Gastric Carcinoma Cells: Analysis by Flow Cytometry Bruce E. Dunn, Magda Altmann, and Gail P. Campbell
From the Laboratory Service, Denver ~terans Administration Medical Center; and the Department of Pathology, University of Colorado Health Sciences Center, Denver, Colorado
Infection by the gram-negative, spiral, microaerophilic bacterium Helicobacterpylori, formerly known as Campylobacter pylori [1], is a primary cause of nonautoimmune (type B) gastritis and may be a permissive factor in peptic ulcer disease [2-7]. Since its initial isolation by Warren [8] and Marshall [9] in 1983, H. pylori has been isolated from biopsied gastric tissue by investigators worldwide [2, 3]. The pathogenic mechanisms whereby H. pylori induces gastric inflammation are not well understood. H. pylori does not invade gastric tissue but is found in association with the apical surfaces of gastric mucus-secreting cells [4, 10]. Some strains of H. pylori secrete cytotoxins [11] and mucinases [12], although the significance of these factors in vivo is unknown. Indirect evidence suggests that H. pylori-specific attachment sites may exist on gastric mucus-secreting cells. For example, H. pylori is found only in association with this type of cell in humans [2, 10, 13]. The organism is often seen in proximity to sites of gastric metaplasia in both the duodenal bulb and the esophagus [6, 7]. However, it has not been observed in association with intestinal metaplasia when present in biopsied gastric tissue [2, 3]. Electron microscopy reveals that the adherence of H. pylori to human gastric mucosa involves "attachment pedestals" similar to those observed for adherent enterotoxigenic strains of Escherichia coli [4]. In experimentally infected gnotobiotic piglets, H. pylori is found only in the stomach and not in other regions of the gastrointestinal tract [14, 15].
Reprints and correspondence: Dr. Bruce E. Dunn, Laboratory Service (113LR), John L. McClellan Memorial Veterans Hospital, 4300 West 7th Street, Little Rock, Arkansas 72205. Financial support: Medical Research Service of the Veterans Administration and The Procter & Gamble Company. Reviews of Infectious Diseases 1991;13(Suppl 8):8657-64 This article is in the public domain.
Development of an in vitro model for study of the adherence properties of H. pylori is essential to an understanding of why the bacterium is found only in association with gastric mucus-secreting cells in vivo. Towardthis end, Neman-Simha and Megraud [16] have demonstrated differential adherence of H. pylori to four epithelial cell lines in vitro, using immunofluorescence and scanning electron microscopy. Evans et al. have presented evidence for receptor-mediated adherence of H. pylori to mouse Y-1 adrenal cell monolayers [17]. We have independently developed a method for quantitative assessment of the adherence of H. pylori to cultured epithelial cells, using immunofluorescence and flow cytometry; this method is based on techniques developed for study of the adherence of Chlamydia trachomatis to fibroblasts and leukocytes [18, 19]. The present study assesses the time course, saturability, reversibility, and inhibition of the adherence of H. pylori to cultured gastric carcinoma cells.
Materials and Methods Cell culture. The following cell lines were obtained from the American Type Culture Collection (Rockville, MD): Intestine407 (ATCC CCL6), KATO III (ATCC HTB 103), and Hs 746T (ATCC HTB 135); these cell lines are referred to herein as Int 407, KATO, and HTB 135, respectively. Embryonal carcinoma (EC) cells and a yolk sac-like epithelial cell line (402AX) were provided by Dr. Wendell Speers, Department of Pathology, University of Colorado Health Sciences Center, Denver. Cells were cultured in RPMI 1640 medium (Gibco, Grand Island, NY) supplemented with 10% fetal bovine serum containing penicillin, streptomycin, and amphotericin B. Stock cultures were passaged at intervals of 3-4 days. For adherence experiments 106 cells were transferred into Falcon dishes (diameter, 100 mm; Falcon Plastics, Oxnard, CA) and were incubated in 10 mL of culture medium for 24 hours be-
Downloaded from http://cid.oxfordjournals.org/ at New York University on June 7, 2015
An in vitro assay using immunofluorescence and flow cytometry for quantitative assessment of the adherence of Helicobacter pylori to cultured human gastric carcinoma (KATO III) cells was developed. Adherence was rapid, saturable, energy dependent, mannose resistant, and significantly inhibited by fetuin, a glycoprotein containing N-acetylneuraminyllactose. Pretreatment of KATO cells with neuraminidase from Clostridium perjringens, however, did not reduce adherence of H. pylori. Ultrastructurally, adherent cells of H. pyloriwere associated with indentations of KATO cell membranes. KATO cells should prove useful in the investigation of mechanisms of adherence of H. pylori to mammalian cells. Ultimately, this flow cytometric assay may be helpful in assessment of the adherence of laboratory strains of H. pylori directly to surface mucous cells dissociated from biopsied human gastric tissue.
5658
Dunn, Altmann, and Campbell
of PBS and resuspended in 1 mL of 1% p-formaldehyde in PBS. Stained cells were evaluatedat a magnification of xl,OOO with an oil immersion objective on an epifluorescence microscope (Carl Zeiss, New York) that was equipped with an M35W camera attachment. For photography, Kodak Tri-X pan black and white film (ASA 400; Eastman Kodak, Rochester, NY) was used. Analysis of bacterial adherence by flow cytometry. An EPICS-C device (Coulter Electronics, Hialeah, FL) with a Coherent (Palo Alto, CA) argon-ion laser tuned to 488 DID at 500 mW was used for flow cytometry. Green fluorescence was analyzed with a 488-nm dichroic mirror, a 515-nm longpass interference filter, a 530-nm short-pass interference filter, and a log-linear amplifier. Fluorescence data were acquired in log mode (70 channels equivalent to 1 log decade) on a 256-channel scale. Cellular debris, unbound bacteria, and cell clumps were excluded from further analysis by volume gating. The mean fluorescence channel number of a peak representing cells alone or cells with adherent bacteria was determined by means of the Intgra program on a Coulter Easy 88 microcomputer system linked to the flow cytometer. The mean fluorescence channel was calculated by an analysis of more than 2,000 individual cells per determination. Mean fluorescence intensity (MFI) was calculated from the mean fluorescence channel by the function MFI = e{[ln 1,000/ 255]x}, which simplifies to MFI = e(0.027lx), where e is the base of the natural logarithm system and x is the mean fluorescence channel number generated by the Intgra program. In some cases H. pylori organisms alone were harvested at 7,000 g for 15 minutes, stained by the two-step fluorescent antibody procedure described above, and then analyzed by flow cytometry. Under these conditions bacteria clumped together. Thus it was not possible to determine the specific fluorescence of individual bacteria or the number of adherent bacteria on cells corresponding to a given MFI value. The data generated by flow cytometry therefore allowed assessment only of relative differences in bacterial adherence between various experimental treatments. Comparison ofH. pylori adherence to different eel/lines. H. pylori strain 87-263 was used in comparisons of the adherence of H. pylori to the five cultured cell lines studied. The B:C ratio in all cell lines tested was f\JI00:1. Relative cell volume was determined from the forward-angle light scatter of each line. Cell suspensions and bacteria were mixed for 1 hour before determination of MFI. Saturability of adherence. H. pylori strains 87-263, 8897, 88-101, 85-456, and 84-182 were used for assessment of the saturability of adherence to KATO cells. Each strain was harvested from 30 to 50 chocolate agar plates and was suspended at an OD 450 of 4.5 (5.0 ± 1.1 x 1010 cfu/mL). KATO cells were mixed for 30 minutes with serial 10-fold dilutions of bacteria, with an initial B:C ratio of 10,000:1. The MFI was determined for each dilution as described above. Rate ofadherence. H. pylori strain 87-263 was used for
Downloaded from http://cid.oxfordjournals.org/ at New York University on June 7, 2015
fore use. In general, KATO cells did not attach readily to dishes; most cells remained in suspension under the conditions used. In contrast, the other cell lines studied generally attached to dishes during culture. Cells were gently harvested in culture medium with a sterile rubber policeman - but without use of proteolytic enzymes-to preserve putative receptors for bacteria. Subsequently, by repeated aspiration and expulsion from a Pasteur pipette, cells were dissociated into a suspension consisting primarily of individual cells, with relatively few doublets, triplets, or cell clumps. Culture ofH. pylori. The strains of H. pylori used in this study were from the culture collection of the Campylobacter Laboratory at the Denver Veterans Administration Medical Center. All strains were cultured originally from biopsied human gastric tissue and were identified, passaged, and stored as described previously [20]. H. pylori strains used in adherence studies included 87-263, 88-97, 88-101, 85-456, and 84182. All strains were grown on chocolate agar plates (Remel, Lenexa, KS) at 37°C for 24 hours in an atmosphere containing 5% O2 , 7.5% CO 2 , and 7.5% H 2 • Cultures were harvested in 20 mM PBS (pH 7.4) with sterile swabs. Viable colony counts and turbidity measurements were performed routinely for determination of bacteria-to-cell (B:C) ratios in adherence experiments. Adherence ofH. pylori to cultured cells. Samples containing lQ6 cells were pelleted, and 1 mL of PBS containing H. pylori in the desired concentration was added (~50 of 1.0 = 5.0 ± 1.3 x 108 cfu/mL). Cells and bacteria were agitated at 150 rpm at 37°C for 30 minutes unless specified otherwise. After incubation a 14-mL volume of 15% sucrose was added, with mixing; tubes were then centrifuged at 200 g for 5 minutes for harvesting of cells. In each experiment an aliquot of both supernatant and pellet was examined by modified gram staining [3]. The pellet contained cells, most or all of which (depending on the B:C ratio used) showed adherent bacteria. In general, few nonadherent bacteria were observed in the pellet. Supernatants lacked cells but contained numerous bacteria. Centrifugation of H. pylori alone at 200 g for 5 minutes through 15% sucrose did not result in the formation of a visible pellet. Under these conditions few bacteria were present in the bottom of the centrifugation tube. Fixation and immunofluorescent staining. Cells with or without adherent bacteria, obtained by centrifugation as described above, were incubated for 20 minutes at 4°C in 100 #LL of a 1:10 dilution of hyperimmune rabbit serum raised against whole cells of H. pylori strain 85-456 [21] (supplied by Drs. G. I. Perez-Perez and M. 1. Blaser, Vanderbilt University, Nashville, TN). After being washed with 15 mL of PBS, cells were incubated for an additional 20 minutes at 4°C in 100 ~L of fluorescein isothiocyanate (FITC)-conjugated goat antiserum to rabbit proteins (diluted 1:10; Jackson Immunoresearch Laboratories, West Grove, PA). Preliminary experiments demonstrated that both concentrations of antiserum used were saturating. Cells were then washed in 15 mL
RID 1991;13 (Suppl 8)
RID 1991;13 (Suppl 8)
Adherence of H. pylori to KATO Cells
1 mg/mL in PBS. Bacteria and cells were then mixed for 30 minutes at 37°C in the presence of the same concentrations of these compounds. Cells were washed and analyzed as described above. In control experiments asialofetuin (Type II; Sigma) was used at a concentration of 1 mg/mL. Electron microscopy. Cells and bacteria were incubated in PBS at 37°C for 1 hour with agitation; nonadherent bacteria were then separated by centrifugation as described above. Cells were fixed in 2 % glutaraldehyde in 0.1 M cacodylate buffer (pH 7.3) at 4°C for 1 hour; they were then postfixed in 1% cacodylate-buffered osmium tetroxide, dehydrated in a graded series of ethanol solutions, and embedded in plastic containing 34% Poly/Bed 812 embedding medium (Polysciences, Warrington, PA), 16% Araldite epoxy resin (Polysciences), and 49% dodecenylsuccinic anhydride (Polysciences). Thick sections (1.0 ~m) were cut with glass knives, stained with 1% toluidine blue, and observed for orientation and for the presence of bacteria. Thin sections (60-80 nm) were cut with diamond knives, stained with lead citrate and uranyl acetate, and observed with a Philips model CM 12 electron microscope (Philips Electronics, Skokie, IL) operated at 60 kYo Statistical analysis. In general, inhibition experiments and an equal number of control experiments were performed three to five times each at a given concentration of inhibitor. Data are expressed as the mean percentage change in MFI in the presence of inhibitor as compared with the corresponding control value. The statistical significance of differences between MFI values for various treatments was determined by means of the pooled t test.
Results Individual bacteria adherent to KATO cells were apparent by fluorescence microscopy (figure 1). Adherent bacteria were present in a diffuse pattern on KATO cells and were generally
Figure 1. Fluorescence photomicrograph of H. pylori strain 87263 adherent to three adjacent KATO cells (arrows). This photomicrograph shows only a single focal plane. Adjustment of the microscope revealed many additional adherent bacteria in other focal planes. Bar = 5 J,tm.
Downloaded from http://cid.oxfordjournals.org/ at New York University on June 7, 2015
measurement of the rate of adherence to KATO cells. Cells and bacteria were mixed for intervals of 5-120 minutes. Free bacteria were removed by centrifugation, and MFI was determined as described above. The B:C ratio used in this and subsequent experiments was rvl00:1. Stability 0/ adherence. H. pylori strain 87-263 was used for measurement of the stability of adherence to KATO cells. After cells and bacteria had been mixed at 37°C for 30 minutes, the mixture was separated into two equal portions. One portion was centrifuged immediately for the removal of nonadherent bacteria and then held unfixed at 4°C. The other portion was incubated in 10 mL of bacterium-free PBS at 37°C for 60 minutes with agitation. Cells were then harvested as described above. Both portions of the bacteria-cell mixture were stained simultaneously before analysis. Inhibition ofadherence. H. pylori strain 87-263 was used for evaluation of the effects of temperature (4°C vs. 37°C) on bacterial adherence to KATO cells. Control cells and bacteria were mixed for 30 minutes at 37°C and then processed as described above. To assess adherence at 4°C, cells and bacteria were placed on ice in separate tubes for 10 minutes and then were mixed with agitation for 30 minutes at 4 °C. Nonadherent bacteria were separated from cells by centrifugation in ice-cold 15 % sucrose. To test the effects of metabolic inhibition on adherence, H. pylori strain 87-263 was incubated for 10 minutes in 10 mM KCN in PBS at 37°C, and bacteria and cells were then mixed in PBS for 30 minutes at 37°C. Cells were stained and assayed as described above. In some cases samples of bacteria exposed to 10 mM KCN in PBS or to PBS alone for 10-30 minutes were observed directly under cover slips on glass slides with a Nikon Labophot microscope (Nikon, Garden City, NY) at a magnification of x500. As a test for possible inhibition of adherence by D-mannose, KATO cells and H. pylori strain 87-263 were incubated for 30 minutes at 37°C at a final concentration of D-mannose (Sigma Chemical, St. Louis) of 25 mg/mL, and the adherence assay was performed as described above. The potential role of cellular glycoproteins in adherence of H. pylori was assessed by pretreatment of KATO cells for 30 minutes at 37°C with neuraminidase from Clostridium perfringens (Type X; Sigma) at a final concentration of 0.125 or 0.25 U/mL in PBS (pH 6.0). Cells were then washed three times in PBS, and the flow cytometric adherence assay was performed as described above. N-Acetylneuraminyllactose and fetuin have previously been shown to inhibit hemagglutination by H. pylori [22] and to inhibit adherence of H. pylori to Y-1 adrenal cells [17]. To test the possible inhibitory effects of these compounds on the adherence of H. pylori strain 87-263 to KATO cells, both cells and bacteria were incubated for 10 minutes at 37°C either with N-acetylneuraminyllactose from bovine colostrum (Sigma) at a concentration of 0.25 mg/mL in PBS or with fetuin (prepared by gel filtration; Sigma) at a concentration of
S659
..J
...I
A
B
W
0
&L
0 a:
w m
:E ~
~
v;
z w
~
~
w
u
z
w
o (/) w
a:::
0
:::>
...J
lL
W
>
~ w
a:::
170 160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 0
KATO
HTB
INT
402AX
EC
Figure 3. Relative fluorescence intensity of five epithelial cell lines exposed in suspension to H. pylori strain 87-263 at a B:C ratio of 1"\.1100:1. In this and subsequent experiments, MFI was calculated from the mean channel number as described in the text. In this experiment MFI was normalized to that for KATO cells, which showed greater MFI than any other cell line except EC cells. Each value represents the mean of three or four determinations (± SEM).
light scatter and direct microscopic observation (data not shown). EC cell populations were generally composed of small undifferentiated cells (70 %-90 %) and large differentiated cells (10%-30%) under the culture conditions used. The EC cells exposed to H. pylori demonstrated MFI 1'\.115% greater than that of KATO cells (figure 3). Saturability ofadherence. KATO cells exposed to one bacterium per cell (log [B:C ratio] = 0) could be discriminated from an unexposed population of cells (data not shown). The MFI increased as the B:C ratio was increased to 1'\.1100:1 (log [B:C ratio] = 2). Exposure of suspended cells to bacteria at B:C ratios of up to 10,000:1 (log [B:C ratio] = 4) did not increase the MFI further (figure 4). All five H. pylori strains tested gave results similar to those depicted for KATO cells in figure 4 (data not shown).
z
w >
1000~------------------,
~
~ v; z
..J W
a:
w
~
~
w
100
•
U Z
LOG FLUORESCENCE INTENSITY
Figure 2. Comparison of log green fluorescence, measured on a 256-channel scale, of KATO cells unexposed (A) and exposed (8) to H. pylori strain 87-263. Cells and bacteria (B:C ratio, ""100:1) were mixed for 30 minutes at 37°C; then cells were pelleted through sucrose as described in the text. The midpoint between histograms of the two populations was chosen as the cutoff between negatively and positively staining cells (arrow). The mean channel numbers calculated with the Intgra program were 53.1 for peak A and 166.6 for peak B, corresponding to calculated MFI values of 4.2 and 91.4 units, respectively. Thus, in this example, peak B exhibited an MFI 21.8 times greater than that of cells in peak A.
W U VI
w
a:::
o
:::>
...J
I.L
~:E
l+-----~----_t_---~---__i
o
234 LOG [B:C RATIO]
Figure 4. Effects on MFI of increasing the B:C ratio (H. pylori strain 87-263 to KATO cells in suspension) in two experiments. In both instances MFI was maximal at B:C ratios of ~100:1 (log [B:C ratio] = 2).
Downloaded from http://cid.oxfordjournals.org/ at New York University on June 7, 2015
too numerous to count visually. After incubation for 30 minutes at a B:C ratio of 100:1 and subsequent staining with a twostage FITC-Iabeled H. pylori antibody method, KATO cells with associated bacteria stained 20-50 times more intensely than did cells without bacteria and could easily be discriminated by flow cytometry. Fluorescence histograms of KATO cells unexposed and exposed to H. pylori strain 87-263 are shown in figure 2. The cutoff between negativelyand positively stained cells was chosen as the midpoint between the histograms of the two cell populations (channel number 100). Cells unexposed to bacteria showed
Adherence of Helicobacter pylori to Gastric Carcinoma Cells: Analysis by Flow Cytometry Bruce E. Dunn, Magda Altmann, and Gail P. Campbell
From the Laboratory Service, Denver ~terans Administration Medical Center; and the Department of Pathology, University of Colorado Health Sciences Center, Denver, Colorado
Infection by the gram-negative, spiral, microaerophilic bacterium Helicobacterpylori, formerly known as Campylobacter pylori [1], is a primary cause of nonautoimmune (type B) gastritis and may be a permissive factor in peptic ulcer disease [2-7]. Since its initial isolation by Warren [8] and Marshall [9] in 1983, H. pylori has been isolated from biopsied gastric tissue by investigators worldwide [2, 3]. The pathogenic mechanisms whereby H. pylori induces gastric inflammation are not well understood. H. pylori does not invade gastric tissue but is found in association with the apical surfaces of gastric mucus-secreting cells [4, 10]. Some strains of H. pylori secrete cytotoxins [11] and mucinases [12], although the significance of these factors in vivo is unknown. Indirect evidence suggests that H. pylori-specific attachment sites may exist on gastric mucus-secreting cells. For example, H. pylori is found only in association with this type of cell in humans [2, 10, 13]. The organism is often seen in proximity to sites of gastric metaplasia in both the duodenal bulb and the esophagus [6, 7]. However, it has not been observed in association with intestinal metaplasia when present in biopsied gastric tissue [2, 3]. Electron microscopy reveals that the adherence of H. pylori to human gastric mucosa involves "attachment pedestals" similar to those observed for adherent enterotoxigenic strains of Escherichia coli [4]. In experimentally infected gnotobiotic piglets, H. pylori is found only in the stomach and not in other regions of the gastrointestinal tract [14, 15].
Reprints and correspondence: Dr. Bruce E. Dunn, Laboratory Service (113LR), John L. McClellan Memorial Veterans Hospital, 4300 West 7th Street, Little Rock, Arkansas 72205. Financial support: Medical Research Service of the Veterans Administration and The Procter & Gamble Company. Reviews of Infectious Diseases 1991;13(Suppl 8):8657-64 This article is in the public domain.
Development of an in vitro model for study of the adherence properties of H. pylori is essential to an understanding of why the bacterium is found only in association with gastric mucus-secreting cells in vivo. Towardthis end, Neman-Simha and Megraud [16] have demonstrated differential adherence of H. pylori to four epithelial cell lines in vitro, using immunofluorescence and scanning electron microscopy. Evans et al. have presented evidence for receptor-mediated adherence of H. pylori to mouse Y-1 adrenal cell monolayers [17]. We have independently developed a method for quantitative assessment of the adherence of H. pylori to cultured epithelial cells, using immunofluorescence and flow cytometry; this method is based on techniques developed for study of the adherence of Chlamydia trachomatis to fibroblasts and leukocytes [18, 19]. The present study assesses the time course, saturability, reversibility, and inhibition of the adherence of H. pylori to cultured gastric carcinoma cells.
Materials and Methods Cell culture. The following cell lines were obtained from the American Type Culture Collection (Rockville, MD): Intestine407 (ATCC CCL6), KATO III (ATCC HTB 103), and Hs 746T (ATCC HTB 135); these cell lines are referred to herein as Int 407, KATO, and HTB 135, respectively. Embryonal carcinoma (EC) cells and a yolk sac-like epithelial cell line (402AX) were provided by Dr. Wendell Speers, Department of Pathology, University of Colorado Health Sciences Center, Denver. Cells were cultured in RPMI 1640 medium (Gibco, Grand Island, NY) supplemented with 10% fetal bovine serum containing penicillin, streptomycin, and amphotericin B. Stock cultures were passaged at intervals of 3-4 days. For adherence experiments 106 cells were transferred into Falcon dishes (diameter, 100 mm; Falcon Plastics, Oxnard, CA) and were incubated in 10 mL of culture medium for 24 hours be-
Downloaded from http://cid.oxfordjournals.org/ at New York University on June 7, 2015
An in vitro assay using immunofluorescence and flow cytometry for quantitative assessment of the adherence of Helicobacter pylori to cultured human gastric carcinoma (KATO III) cells was developed. Adherence was rapid, saturable, energy dependent, mannose resistant, and significantly inhibited by fetuin, a glycoprotein containing N-acetylneuraminyllactose. Pretreatment of KATO cells with neuraminidase from Clostridium perjringens, however, did not reduce adherence of H. pylori. Ultrastructurally, adherent cells of H. pyloriwere associated with indentations of KATO cell membranes. KATO cells should prove useful in the investigation of mechanisms of adherence of H. pylori to mammalian cells. Ultimately, this flow cytometric assay may be helpful in assessment of the adherence of laboratory strains of H. pylori directly to surface mucous cells dissociated from biopsied human gastric tissue.
5658
Dunn, Altmann, and Campbell
of PBS and resuspended in 1 mL of 1% p-formaldehyde in PBS. Stained cells were evaluatedat a magnification of xl,OOO with an oil immersion objective on an epifluorescence microscope (Carl Zeiss, New York) that was equipped with an M35W camera attachment. For photography, Kodak Tri-X pan black and white film (ASA 400; Eastman Kodak, Rochester, NY) was used. Analysis of bacterial adherence by flow cytometry. An EPICS-C device (Coulter Electronics, Hialeah, FL) with a Coherent (Palo Alto, CA) argon-ion laser tuned to 488 DID at 500 mW was used for flow cytometry. Green fluorescence was analyzed with a 488-nm dichroic mirror, a 515-nm longpass interference filter, a 530-nm short-pass interference filter, and a log-linear amplifier. Fluorescence data were acquired in log mode (70 channels equivalent to 1 log decade) on a 256-channel scale. Cellular debris, unbound bacteria, and cell clumps were excluded from further analysis by volume gating. The mean fluorescence channel number of a peak representing cells alone or cells with adherent bacteria was determined by means of the Intgra program on a Coulter Easy 88 microcomputer system linked to the flow cytometer. The mean fluorescence channel was calculated by an analysis of more than 2,000 individual cells per determination. Mean fluorescence intensity (MFI) was calculated from the mean fluorescence channel by the function MFI = e{[ln 1,000/ 255]x}, which simplifies to MFI = e(0.027lx), where e is the base of the natural logarithm system and x is the mean fluorescence channel number generated by the Intgra program. In some cases H. pylori organisms alone were harvested at 7,000 g for 15 minutes, stained by the two-step fluorescent antibody procedure described above, and then analyzed by flow cytometry. Under these conditions bacteria clumped together. Thus it was not possible to determine the specific fluorescence of individual bacteria or the number of adherent bacteria on cells corresponding to a given MFI value. The data generated by flow cytometry therefore allowed assessment only of relative differences in bacterial adherence between various experimental treatments. Comparison ofH. pylori adherence to different eel/lines. H. pylori strain 87-263 was used in comparisons of the adherence of H. pylori to the five cultured cell lines studied. The B:C ratio in all cell lines tested was f\JI00:1. Relative cell volume was determined from the forward-angle light scatter of each line. Cell suspensions and bacteria were mixed for 1 hour before determination of MFI. Saturability of adherence. H. pylori strains 87-263, 8897, 88-101, 85-456, and 84-182 were used for assessment of the saturability of adherence to KATO cells. Each strain was harvested from 30 to 50 chocolate agar plates and was suspended at an OD 450 of 4.5 (5.0 ± 1.1 x 1010 cfu/mL). KATO cells were mixed for 30 minutes with serial 10-fold dilutions of bacteria, with an initial B:C ratio of 10,000:1. The MFI was determined for each dilution as described above. Rate ofadherence. H. pylori strain 87-263 was used for
Downloaded from http://cid.oxfordjournals.org/ at New York University on June 7, 2015
fore use. In general, KATO cells did not attach readily to dishes; most cells remained in suspension under the conditions used. In contrast, the other cell lines studied generally attached to dishes during culture. Cells were gently harvested in culture medium with a sterile rubber policeman - but without use of proteolytic enzymes-to preserve putative receptors for bacteria. Subsequently, by repeated aspiration and expulsion from a Pasteur pipette, cells were dissociated into a suspension consisting primarily of individual cells, with relatively few doublets, triplets, or cell clumps. Culture ofH. pylori. The strains of H. pylori used in this study were from the culture collection of the Campylobacter Laboratory at the Denver Veterans Administration Medical Center. All strains were cultured originally from biopsied human gastric tissue and were identified, passaged, and stored as described previously [20]. H. pylori strains used in adherence studies included 87-263, 88-97, 88-101, 85-456, and 84182. All strains were grown on chocolate agar plates (Remel, Lenexa, KS) at 37°C for 24 hours in an atmosphere containing 5% O2 , 7.5% CO 2 , and 7.5% H 2 • Cultures were harvested in 20 mM PBS (pH 7.4) with sterile swabs. Viable colony counts and turbidity measurements were performed routinely for determination of bacteria-to-cell (B:C) ratios in adherence experiments. Adherence ofH. pylori to cultured cells. Samples containing lQ6 cells were pelleted, and 1 mL of PBS containing H. pylori in the desired concentration was added (~50 of 1.0 = 5.0 ± 1.3 x 108 cfu/mL). Cells and bacteria were agitated at 150 rpm at 37°C for 30 minutes unless specified otherwise. After incubation a 14-mL volume of 15% sucrose was added, with mixing; tubes were then centrifuged at 200 g for 5 minutes for harvesting of cells. In each experiment an aliquot of both supernatant and pellet was examined by modified gram staining [3]. The pellet contained cells, most or all of which (depending on the B:C ratio used) showed adherent bacteria. In general, few nonadherent bacteria were observed in the pellet. Supernatants lacked cells but contained numerous bacteria. Centrifugation of H. pylori alone at 200 g for 5 minutes through 15% sucrose did not result in the formation of a visible pellet. Under these conditions few bacteria were present in the bottom of the centrifugation tube. Fixation and immunofluorescent staining. Cells with or without adherent bacteria, obtained by centrifugation as described above, were incubated for 20 minutes at 4°C in 100 #LL of a 1:10 dilution of hyperimmune rabbit serum raised against whole cells of H. pylori strain 85-456 [21] (supplied by Drs. G. I. Perez-Perez and M. 1. Blaser, Vanderbilt University, Nashville, TN). After being washed with 15 mL of PBS, cells were incubated for an additional 20 minutes at 4°C in 100 ~L of fluorescein isothiocyanate (FITC)-conjugated goat antiserum to rabbit proteins (diluted 1:10; Jackson Immunoresearch Laboratories, West Grove, PA). Preliminary experiments demonstrated that both concentrations of antiserum used were saturating. Cells were then washed in 15 mL
RID 1991;13 (Suppl 8)
RID 1991;13 (Suppl 8)
Adherence of H. pylori to KATO Cells
1 mg/mL in PBS. Bacteria and cells were then mixed for 30 minutes at 37°C in the presence of the same concentrations of these compounds. Cells were washed and analyzed as described above. In control experiments asialofetuin (Type II; Sigma) was used at a concentration of 1 mg/mL. Electron microscopy. Cells and bacteria were incubated in PBS at 37°C for 1 hour with agitation; nonadherent bacteria were then separated by centrifugation as described above. Cells were fixed in 2 % glutaraldehyde in 0.1 M cacodylate buffer (pH 7.3) at 4°C for 1 hour; they were then postfixed in 1% cacodylate-buffered osmium tetroxide, dehydrated in a graded series of ethanol solutions, and embedded in plastic containing 34% Poly/Bed 812 embedding medium (Polysciences, Warrington, PA), 16% Araldite epoxy resin (Polysciences), and 49% dodecenylsuccinic anhydride (Polysciences). Thick sections (1.0 ~m) were cut with glass knives, stained with 1% toluidine blue, and observed for orientation and for the presence of bacteria. Thin sections (60-80 nm) were cut with diamond knives, stained with lead citrate and uranyl acetate, and observed with a Philips model CM 12 electron microscope (Philips Electronics, Skokie, IL) operated at 60 kYo Statistical analysis. In general, inhibition experiments and an equal number of control experiments were performed three to five times each at a given concentration of inhibitor. Data are expressed as the mean percentage change in MFI in the presence of inhibitor as compared with the corresponding control value. The statistical significance of differences between MFI values for various treatments was determined by means of the pooled t test.
Results Individual bacteria adherent to KATO cells were apparent by fluorescence microscopy (figure 1). Adherent bacteria were present in a diffuse pattern on KATO cells and were generally
Figure 1. Fluorescence photomicrograph of H. pylori strain 87263 adherent to three adjacent KATO cells (arrows). This photomicrograph shows only a single focal plane. Adjustment of the microscope revealed many additional adherent bacteria in other focal planes. Bar = 5 J,tm.
Downloaded from http://cid.oxfordjournals.org/ at New York University on June 7, 2015
measurement of the rate of adherence to KATO cells. Cells and bacteria were mixed for intervals of 5-120 minutes. Free bacteria were removed by centrifugation, and MFI was determined as described above. The B:C ratio used in this and subsequent experiments was rvl00:1. Stability 0/ adherence. H. pylori strain 87-263 was used for measurement of the stability of adherence to KATO cells. After cells and bacteria had been mixed at 37°C for 30 minutes, the mixture was separated into two equal portions. One portion was centrifuged immediately for the removal of nonadherent bacteria and then held unfixed at 4°C. The other portion was incubated in 10 mL of bacterium-free PBS at 37°C for 60 minutes with agitation. Cells were then harvested as described above. Both portions of the bacteria-cell mixture were stained simultaneously before analysis. Inhibition ofadherence. H. pylori strain 87-263 was used for evaluation of the effects of temperature (4°C vs. 37°C) on bacterial adherence to KATO cells. Control cells and bacteria were mixed for 30 minutes at 37°C and then processed as described above. To assess adherence at 4°C, cells and bacteria were placed on ice in separate tubes for 10 minutes and then were mixed with agitation for 30 minutes at 4 °C. Nonadherent bacteria were separated from cells by centrifugation in ice-cold 15 % sucrose. To test the effects of metabolic inhibition on adherence, H. pylori strain 87-263 was incubated for 10 minutes in 10 mM KCN in PBS at 37°C, and bacteria and cells were then mixed in PBS for 30 minutes at 37°C. Cells were stained and assayed as described above. In some cases samples of bacteria exposed to 10 mM KCN in PBS or to PBS alone for 10-30 minutes were observed directly under cover slips on glass slides with a Nikon Labophot microscope (Nikon, Garden City, NY) at a magnification of x500. As a test for possible inhibition of adherence by D-mannose, KATO cells and H. pylori strain 87-263 were incubated for 30 minutes at 37°C at a final concentration of D-mannose (Sigma Chemical, St. Louis) of 25 mg/mL, and the adherence assay was performed as described above. The potential role of cellular glycoproteins in adherence of H. pylori was assessed by pretreatment of KATO cells for 30 minutes at 37°C with neuraminidase from Clostridium perfringens (Type X; Sigma) at a final concentration of 0.125 or 0.25 U/mL in PBS (pH 6.0). Cells were then washed three times in PBS, and the flow cytometric adherence assay was performed as described above. N-Acetylneuraminyllactose and fetuin have previously been shown to inhibit hemagglutination by H. pylori [22] and to inhibit adherence of H. pylori to Y-1 adrenal cells [17]. To test the possible inhibitory effects of these compounds on the adherence of H. pylori strain 87-263 to KATO cells, both cells and bacteria were incubated for 10 minutes at 37°C either with N-acetylneuraminyllactose from bovine colostrum (Sigma) at a concentration of 0.25 mg/mL in PBS or with fetuin (prepared by gel filtration; Sigma) at a concentration of
S659
..J
...I
A
B
W
0
&L
0 a:
w m
:E ~
~
v;
z w
~
~
w
u
z
w
o (/) w
a:::
0
:::>
...J
lL
W
>
~ w
a:::
170 160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 0
KATO
HTB
INT
402AX
EC
Figure 3. Relative fluorescence intensity of five epithelial cell lines exposed in suspension to H. pylori strain 87-263 at a B:C ratio of 1"\.1100:1. In this and subsequent experiments, MFI was calculated from the mean channel number as described in the text. In this experiment MFI was normalized to that for KATO cells, which showed greater MFI than any other cell line except EC cells. Each value represents the mean of three or four determinations (± SEM).
light scatter and direct microscopic observation (data not shown). EC cell populations were generally composed of small undifferentiated cells (70 %-90 %) and large differentiated cells (10%-30%) under the culture conditions used. The EC cells exposed to H. pylori demonstrated MFI 1'\.115% greater than that of KATO cells (figure 3). Saturability ofadherence. KATO cells exposed to one bacterium per cell (log [B:C ratio] = 0) could be discriminated from an unexposed population of cells (data not shown). The MFI increased as the B:C ratio was increased to 1'\.1100:1 (log [B:C ratio] = 2). Exposure of suspended cells to bacteria at B:C ratios of up to 10,000:1 (log [B:C ratio] = 4) did not increase the MFI further (figure 4). All five H. pylori strains tested gave results similar to those depicted for KATO cells in figure 4 (data not shown).
z
w >
1000~------------------,
~
~ v; z
..J W
a:
w
~
~
w
100
•
U Z
LOG FLUORESCENCE INTENSITY
Figure 2. Comparison of log green fluorescence, measured on a 256-channel scale, of KATO cells unexposed (A) and exposed (8) to H. pylori strain 87-263. Cells and bacteria (B:C ratio, ""100:1) were mixed for 30 minutes at 37°C; then cells were pelleted through sucrose as described in the text. The midpoint between histograms of the two populations was chosen as the cutoff between negatively and positively staining cells (arrow). The mean channel numbers calculated with the Intgra program were 53.1 for peak A and 166.6 for peak B, corresponding to calculated MFI values of 4.2 and 91.4 units, respectively. Thus, in this example, peak B exhibited an MFI 21.8 times greater than that of cells in peak A.
W U VI
w
a:::
o
:::>
...J
I.L
~:E
l+-----~----_t_---~---__i
o
234 LOG [B:C RATIO]
Figure 4. Effects on MFI of increasing the B:C ratio (H. pylori strain 87-263 to KATO cells in suspension) in two experiments. In both instances MFI was maximal at B:C ratios of ~100:1 (log [B:C ratio] = 2).
Downloaded from http://cid.oxfordjournals.org/ at New York University on June 7, 2015
too numerous to count visually. After incubation for 30 minutes at a B:C ratio of 100:1 and subsequent staining with a twostage FITC-Iabeled H. pylori antibody method, KATO cells with associated bacteria stained 20-50 times more intensely than did cells without bacteria and could easily be discriminated by flow cytometry. Fluorescence histograms of KATO cells unexposed and exposed to H. pylori strain 87-263 are shown in figure 2. The cutoff between negativelyand positively stained cells was chosen as the midpoint between the histograms of the two cell populations (channel number 100). Cells unexposed to bacteria showed
