Comparison of Bacterial and Fungal Adherence to Vaginal Exfoliated Epithelial Cells and Human Vaginal Epithelial

Transcription

1 INFECTION AND IMMUNITY, Feb. 1982, p /82/ $02.00/0 Vol. 35, No. 2 Comparison of Bacterial and Fungal Adherence to Vaginal Exfoliated Epithelial Cells and Human Vaginal Epithelial Tissue Culture Cells JACK D. SOBEL, PAULINE MYERS, MATTHEW E. LEVISON,* AND DONALD KAYE Department of Medicine, Medical College of Pennsylvania and Philadelphia Veterans Administration Hospital, Philadelphia, Pennsylvania Received 13 April 1981/Accepted 9 October 1981 The adherence of four bacterial species and Candida albicans to a new in vitro tissue culture model of human vaginal stratified squamous epithelium was investigated and compared with in vitro adherence to vaginal exfoliated cells. Gardnerella vaginalis, group B streptococci, Lactobacillus sp., and C. albicans adhered well to both exfoliated and tissue culture cells. Similarly, a piliated fecal isolate of Escherichia coli, but not a nonpiliated vaginal isolate of E. coli, adhered well to both cell types. Adherence of the piliated E. coli was markedly inhibited by preincubation of bacteria with D-mannose. No inhibition of adherence by D-mannose of G. vaginalis, nonpiliated E. coli, and C. albicans was demonstrated. Scanning electron microscopy of tissue cultures showed nonuniform distribution of adherent microorganisms with diminished adherence in areas of active mitosis and proliferation and increased adherence to mature flat cells, often in the process of desquamation. Studies of bacterial adherence to exfoliated epithelial cells have revealed considerable cellto-cell variation in the number of adherent bacteria, suggesting a relatively nonuniform population of exfoliated cells (4, 6, 9, 16). This may in part represent differences in cell maturation or viability, resulting in changes in the character and number of available receptor sites on the cell membrane (4, 9). Attempts have therefore been made to use alternate experimental models to investigate cellular characteristics crucial to bacterial adherence. These include the use of animal-derived whole organ cultures as well as in vitro monolayer tissue cultures derived from nonhuman or fetal cells (1, 2, 5, 10, 12). We recently reported a simple but reliable method for establishing noncontinuous lines of multilayer squamous cell tissue cultures obtained as a tissue outgrowth from human vaginal biopsy explants (14). Using light microscopy as well as scanning and transmission electron microscopy, a striking morphological similarity was observed between the squamous epithelial cells of this model and fresh viable exfoliated vaginal epithelial cells (VEC). Transmission electron microscopy revealed evidence of keratinization of the superficial layers of the multilayered squamous outgrowth. However, no evidence was presented of functional similarity in the properties of cultured and exfoliated cells. In this study, data are presented indicating that bacterial and fungal adherence in both the exfoliated cells and the in vitro-grown tissue culture cells is similar. MATERIALS AND METHODS Human vaginal tissue cultures. Tissue was obtained from the vaginal cuff, 1 cm from the os of the uterine specimens removed by hysterectomy for nonmalignant diseases. The biopsy specimens were immediately placed in cold Williams medium E (GIBCO Laboratories, Grand Island, N.Y.) containing 10% heat-inactivated fetal bovine serum, penicillin (50 jig/ ml), streptomycin (50 jig/ml), and amphotericin B (1 jig/ml). Tissue samples were sliced immediately to about 1-cm3 fragments and subsequently, using apposed scalpels, to about 2-mm cubes. Cell cultures were initiated by placing the 2-mm cubes in individual slide chambers (Labtec) containing 1.5 ml of Williams culture medium supplemented with penicillin and gentamicin. The culture chambers were placed in a 5% CO2 incubator, 37 C, and left undisturbed for 14 days. Emphasis was given to placing tissue fragments in culture within 30 min of biopsy. After several days of incubation, most epithelial cell fragments had attached and produced islands of epithelial cell outgrowth. Culture media were replaced once a week. All adherence experiments were performed on multilayers obtained after 14 to 35 days of growth. Explants from three patients were tested against each microbial strain twice, and the results were pooled for each microorganism. By using the above explant system, fibrous tissue was excised and similarly prepared for fibroblast primary monolayers. Exfoliated VEC. Six healthy, nonpregnant volunteer female medical students, mean age 25 years (range, 22 to 29 years), were used. All of the volunteers were 697

2 698 SOBEL ET AL. asymptomatic and sexually active with regular menses and had a normal gynecological examination. None of the subjects used vaginal douches or medication, nor did any use oral hormonal contraceptives; instead they relied on the use of a diaphragm and spermicidal cream. Previous studies in this laboratory have noted no difference in attachment related to the menstrual cycle (J. Sobel, D. Kaye, and M. E. Levison, Clin. Res. 27:357A, 1979). Therefore, cells were obtained, without regard to the time in the menstrual cycle, from the mucosal surface of the midvaginal wall by gentle scraping with an Ayre spatula and were immediately transferred to minimal essential medium (MEM), ph 7.2 (GIBCO). Adherence experiments began within 30 min of sampling. Exfoliated cells from each of the six volunteers were tested against each microbial strain twice, and the results were pooled for each microorganism. Bacteria. The following vaginal isolates were studied: Escherichia coli, Lactobacillus species, group B streptococcus, and Gardnerella vaginalis. Equal portions of an 18-h broth culture were stored at -20 C. For adherence testing, a frozen portion of each bacteria was thawed 24 h before the experiment, transferred to sheep blood agar plates (Difco Laboratories, Detroit, Mich.) (E. coli, group B streptococcus, and Lactobacillus sp.) or Casman agar (G. vaginalis), and incubated at 37 C in 5% CO2 for 24 h. In additional experiments, a fecal isolate of E. coli was thawed and incubated in nutrient broth at 37 C for 48 h under stationary conditions to encourage the formation of type 1 pili. The presence of pili was confirmed by negative tungsten staining on transmission electron microscopy. Candida albicans. A single vaginal isolate of C. albicans was studied. The organism was grown in Sabouraud glucose broth, and equal portions were stored at -20 C. Forty-eight hours before the experiment, the frozen specimen was thawed and transferred to Sabouraud agar; 24 h later, isolated colonies were placed in Sabouraud glucose broth at 37 C for 24 h. Adherence to exfoliated cells. The method used was similar to that described by Mardh and Westrom (7). Vaginal epithelial cells were washed free of nonadherent vaginal bacteria by centrifugation three times in 20 ml of MEM at 800 rpm for 10 min. The cells were resuspended in 10 ml of MEM and subjected to filtration through a membrane filter with a pore size of 8,um (Millipore Corp., Bedford, Mass.), which retains cells but allows free bacteria to pass. The cells were resuspended in MEM and adjusted to a final concentration of 2 x 104 washed epithelial cells/ml as determined in a hemacytometer counting chamber. The bacteria were scraped from the agar, suspended in phosphate-buffered saline (PBS), ph 7.2, and washed three times in 20 ml of PBS by centrifugation at 2,500 rpm. After resuspension of bacteria in PBS, the bacterial suspensions were passed through a 25- gauge needle to obtain single organisms as determined by Gram stain. The final suspension of bacteria was adjusted to 108/ml by using a spectrophotometer after initial quantitative cultures had confirmed the bacterial density. One milliliter of bacteria suspension was mixed with 1 ml of the epithelial cell suspension (5 x 103 bacteria to one epithelial cell) and incubated aerobically on a INFECT. IMMUN. rocker at 37 C for 30 min (ph 7.2). Control tubes contained epithelial cells suspended in MEM at ph 7.2, diluted 1:2 in PBS. Epithelial cells were then washed free of nonadherent bacteria by differential centrifugation at 800 rpm and filtration through an 8-,um membrane filter. To transfer the cells to slides, the membranes were gently pressed against the glass slides which had previously been coated with a thin layer of albumin. After this procedure, less than 1% of epithelial cells remained on the filter as detected by microscopic examination of Gram-stained filters. Cells attached to the slide were air dried, fixed in 95% ethanol for 10 min, and Gram stained. The number of bacteria adherent to epithelial cells was counted under a light microscope (x1,000). In each experiment, at least 50 epithelial cells were studied, and all experiments were performed in duplicate. The C. albicans adherence assay was performed in an identical manner except that a lower inoculum of yeast cells, i.e., 5 x 105/ml, was used (25 yeast per epithelial cell), and the yeast cells were incubated with the VEC for 2 h in Williams medium. Adherence to VEC tissue culture. Tissue culture medium was aspirated, and the outgrowths were gently washed three times with sterile PBS to remove any residual culture medium containing antibiotics as well as fetal calf serum which might interfere with microbial adherence. Thereafter, 0.5 ml of PBS was added to each slide chamber and a further 0.5 ml of bacterial (or yeast) test suspension in PBS (5 x 107 bacteria or 2.5 x 105 yeast). The slide chambers were then incubated in an agitating water bath at 37 C for 30 min with the bacterial suspension or for 2 h with yeast suspensions. After the period of incubation, each slide chamber was aspirated and washed three times by gentle rinsing with sterile PBS. The slide chambers were then inverted to allow air drying of the explant outgrowths, fixed in 95% ethanol, and Gram stained. The number of microorganisms adherent to the tissue culture was counted under a light microscope (x1,000) and expressed as adherent microorganisms per visual field (xl,000). With each explant, 50 fields were counted, and in each experiment two duplicate explants were studied. For comparison of adherence, experiments on exfoliated cells obtained from volunteers were done simultaneously on the same day as experiments on the tissue culture system against each of the microbial strains. All experiments were repeated at least twice on consecutive days, using similar isolates and fresh exfoliated cells from the same volunteer and fresh explants from the same biopsy. It is difficult to quantitatively compare adherence to exfoliated cells with adherence to tissue culture cells because, although the number of bacteria associated with individual exfoliated cells can be identified, the tightly packed, multilayered tissue culture does not permit identification of bacteria with individual epithelial cells. However, the relative adherence of the different microorganisms to exfoliated cells and to tissue culture can be determined. In additional experiments, 25 mg of D(+)-mannose broth per ml (Sigma Chemical Co., St. Louis, Mo.) was added to bacterial or fungal suspensions for 60 min before the incubation as described above with exfoliated cells from three volunteers or the tissue cultures from three patients. Scanning electron microscopy was used to survey

3 VOL. 35, 1982 the epithelial surface for the distribution of bacteria. After fixation in glutaraldehyde, specimens were washed twice in phosphate and fixed in 5% osmium tetroxide in phosphate buffer for 15 min. The specimens were dehydrated in a graded ethanol series (50, 70, 90, and 100% for 5 min each). Thereafter, they were rinsed twice in amyl acetate, subjected to critical point drying for 30 min, coated with gold in a vacuum evaporator, and examined with a scanning electron microscope (Jeol Ltd., JEM model 504). RESULTS All of the bacterial species adhered well to exfoliated VEC (Fig. 1), except for a vaginal isolate of E. coli which was nonpiliated (Table 1). The adherence to explants of G. vaginalis and group B streptococci was similar and greater than that of the Lactobacillus sp., which in turn was greater than that of the piliated E. coli and least for the nonpiliated E. coli. Relative adherence to exfoliated cells was similar to that of explants (Table 1). In control experiments, no adherence of any of the above species to fibroblast monolayers was observed. The vaginal isolate of C. albicans adhered well to exfoliated cells and the tissue culture (24.8 ± 3.1 yeast/ VEC and 51.7 ± 11.8 yeast/visual field, respectively). ADHERENCE TO VAGINAL EPITHELIAL CELLS 699 Preincubation of E. coli in D-mannOSe (25 mg/ ml) for 60 min resulted in a 76 and 87% reduction in adherence of the piliated fecal E. coli strain to exfoliated cells and explants, respectively. No effect was seen when C. albicans, G. vaginalis, or nonpiliated E. coli were preincubated in mannose (Table 2). Bacterial adherence of G. vaginalis to vaginal explant outgrowth was compared at different ages of the explant, i.e., at 3, 4 and 5 weeks, to evaluate the effect of maturation on adherence. No marked differences were observed when the different-aged explants were compared. Maintaining cultures for larger periods was associated with extensive desquamation of the cultured cells and did not allow further comparison. Both light microscopic and scanning electron microscopic observation of bacteria-tissue culture interaction showed nonuniform distribution of adherent bacteria. Bacteria adhered to most of the epithelial cells, but in particular to those cells in the process of desquamating. Moreover, they did not adhere to newly dividing young cells. Least adherence was detected by light microscopy in the areas of the tissue culture with the highest mitotic activity. Figure 2 demonstrates by scanning electron microscopy the adherence of G. vaginalis to explants. Downloaded from on April 23, 2018 by guest FIG. 1. Scanning electron micrograph demonstrating adherence of G. vaginalis organisms to surface of two adjacent exfoliated VEC. (a) x3,000; (b) x10,000.

4 700 SOBEL ET AL. TABLE 1. Adherence of five bacterial isolates to exfoliated VEC and to the multilayer tissue culture Adherence Multilayer tissue Exfoliated Organism culture (mean cells (mean bacteria/visual bacteria/vec field t SE) ± SE) E. coli (vaginal) 1.4 ± ± 0.9 E. coli (faecal) 40.9 ± ± 6.1 Lactobacillus sp ± ± 4.1 G. vaginalis ± ± 5.2 Group B Streptococci ± ± 3.5 INFECT. IMMUN. DISCUSSION The capacity of bacteria to adhere to mucosal epithelial cells is a crucial factor in establishment of the indigenous bacterial flora of a mucosal surface. In studying the bacteria-epithelial cell interaction, most studies have relied on the use of exfoliated skin, buccal, vaginal, or uroepithelial cells (7, 8). Limitations of this method include loss of viability of cells together with a nonuniform population of cells obtained during cell collection. Above all, this method fails to allow for further investigation of factors such as viral infection and hormonal and nutritional factors that may act upon epithelial cell receptor sites and influence bacterial-epithelial cell membrane interaction. Although these drawbacks may be partially overcome by using monolayer tissue cultures derived from mature, nonmalignant tissue, such cultures might not correlate functionally with the parent in vivo cells. In this study, using a simple method previously described (14), the relative adherence to the explant was similar to that of exfoliated cells. Adherence was relatively poor for the strain of E. coli which failed to piliate. The marked adherence to vaginal exfoliated cells of G. vaginalis, group B streptococcus, Lactobacillus sp., and piliated E. coli in contrast to the poor adherence exhibited by the nonpiliated vaginal E. coli isolate is similar to our previous observations studying the effect of the menstrual cycle on adherence to exfoliated vaginal cells (15). These observations provide suggestive evidence of a functional similarity between the multilayered explant cells and the exfoliated VEC. Moreover, the over 75% inhibition of adherence of piliated E. coli to both tissue culture and exfoliated cells by mannose is highly suggestive of the presence of functional mannose-like receptors on both the explants and exfoliated cells. Both light and scanning electron microscopy revealed irregular distribution of adherence of microorganisms to various types of epithelial cells. B. F. King and J. J. Mais (J. Cell Biol. 79:43a, 1978), in a scanning electron microscopy study of the vaginal mucosa of rhesus monkeys, observed that the outermost surfaces of desquamating cells were totally devoid of bacteria, although the bacteria were in large numbers in the intercellular space beneath the desquamating cells. In the present study, epithelial cells in the process of desquamating from the explant had large numbers of adherent bacteria. Similarly, Ramphal et al. (11) found Pseudomonas aeruginosa to adhere preferentially to desquamating tracheal epithelial cells. Keratinization of cells in squamous epithelium increases the closer the cells are to the surface. Sklavounou and Germaine (13) recently observed that keratinization of oral epithelial cells strongly enhanced the adherence of certain oral streptococci to oral mucosal cells. These data suggest the presence of unique surface constituents on the more superficial keratinized cell or the loss of certain surface constituents. Mouse VEC tissue cultures have been shown to be responsive to hormonal stimulation (3), although the local effect of hormones on the number of bacterial receptors and adherence kinetics is unknown. The present model provides an opportunity to study the effect of these factors on bacterial adherence to cells that resemble human in vivo cells, both structurally and functionally. TABLE 2. Culture Effect of preincubation of test microorganisms in D-mannose (25 mg/ml) for 60 min at 37 C before admixture and incubation with exfoliated cells and tissue culture explant G. vaginalis C. albicans E. coli (vaginal) E. coli (faecal) Control D-Mannose Control D-Mannose Control -Mannose Control D-Mannose Exfoliated cells (mean bacteria/ VEC + SE) ± ± ± ± ± ± ± 0.9a Multilayer cells (mean bacterial visual field ± SE) ± ± ± ± ± ± ± ± 0.4a a P < 0.01 for comparison of adherence in presence and absence of D-mannose.

5 VOL. 35, 1982 ADHERENCE TO VAGINAL EPITHELIAL CELLS o ewi W, FIG. 2. Scanning electron micrography demonstrating adherence of G. vaginalis to surface of tissue culture VEC (magnification, x10,000). Note microridges and villous-like appendages of epithelial cells. The microorganisms appear to be anchored to surface by weblike filaments as shown in Fig. 1. ACKNOWLEDGMENT We express our thanks to the Philadelphia College of Osteopathic Medicine for providing their scanning electron microscope facilities. LITERATURE CITED 1. Aldridge, K. E., and B. C. Cole Immunofluorescence and electron microscopy of the attachment of Mycoplasma synoviae to chicken embryo fibroblasts. Infect. Immun. 21: Fitzgerald, T. J., R. C. Johnson, J. N. Miller, and J. A. Sykes Characterization of the attachment of Treponema pallidum to cultured mammalian cells and the potential relationship of attachment to pathogenicity. Infect. Immun. 18: Flaxman, B. A., D. P. Chopra, and D. Newman Growth of mouse vaginal epithelial cells in vitro. In Vitro 9: Fowler, J. E., Jr., and T. A. Stamey Studies on the introital colonization in women with recurrent urinary infections. VII. The role of bacterial adherence. J. Urol. 117: Gabridge, M. G., H. Gunderson, S. L. Schaeffer, and Y. D. Bardenstahl Ciliated respiratory epithelial monolayers: new model for Mycoplasma pneumoniae infection. Infect. Immun. 21: Gibbons, R. J., and J. Van Houte Selective bacterial adherence to oral epithelial surfaces and its role as an ecological determinant. Infect. Immun. 3: Mardh, P. A., and L. Westrom Adherence of bacteria to vaginal epithelial cells. Infect. Immun. 13: Ofek, I., D. Mirelman, and N. Sharon Adherence of Escherichia coli to human mucosal cells mediated by mannose receptors. Nature (London) 265: Parsons, C. L., H. Anwar, C. Stauffer, and J. D. Schmidt In vitro adherence of radioactively labeled Escherichia coli in normal and cystitis-prone females. Infect. Immun. 26: Quaroni, A., J. Wands, R. L. Trelstad, and K. J. Isselbacher Epithelioid cell cultures from rat small intestine. Characterization by morphologic and immunologic criteria. J. Cell Biol. 80: Ramphal, R., P. M. Small, J. W. Shands, Jr., W. Fischlschweiger, and P. A. Small, Jr Adherence of Pseudomonas aeruginosa to tracheal cells injured by influenza infection or by endotracheal intubation. Infect. Immun. 27: Sanford, B. A., A. Shelokov, and M. A. Ramsay Bacterial adherence to virus-infected cells: a cell culture model for bacterial superinfection. J. Infect. Dis. 137: Sklavounou, A., and G. R. Germaine Adherence of oral streptococci to keratinized and non-keratinized human oral epithelial cells. Infect. Immun. 27: Sobel, J. D., R. Tchao, J. Bozzola, M. E. Levison, and D. Kaye Human vaginal epithelial multilayer tissue culture. In Vitro 15: Sobel, J. D., J. Schneider, D. Kaje, and M. E. Levison Adherence of bacteria to vaginal epithelial cells at various times in the menstrual cycle. Infect. Immun. 32: Svanborg Eden, C., B. Eriksson, and L. A. Hanson Adhesion of Escherichia coli to human uroepithelial cells in vitro. Infect. Immun. 18: