We thank David Kelly for his assistance in obtaining pediatric sera. REFERENCES. 1. Amsel, R., P. A. Totten, C. A. Spiegel, K. C. S. Chen, D. Eschenbach, and.
CLINICAL AND DIAGNOSTIC LABORATORY IMMUNOLOGY, Sept. 1996, p. 567–569 1071-412X/96/$04.0010 Copyright q 1996, American Society for Microbiology
Vol. 3, No. 5
Humoral Antibody to Mobiluncus curtisii, a Potential Serological Marker for Bacterial Vaginosis JANE R. SCHWEBKE,1* SHELLIE C. MORGAN,1
AND
SHARON L. HILLIER2
Division of Infectious Diseases, University of Alabama at Birmingham, Birmingham, Alabama 35294-0006,1 and Department of Obstetrics and Gynecology, University of Pittsburgh/Magee Women’s Hospital, Pittsburgh, Pennsylvania 152132 Received 22 March 1996/Returned for modification 15 May 1996/Accepted 3 July 1996
While bacterial vaginosis (BV) is a polymicrobial syndrome, Mobiluncus spp. are the organisms most highly associated with this condition. It is possible that serum antibody to Mobiluncus spp. could be used as a serological marker for BV. Using immunofluorescence techniques, we studied the prevalence of antibody to M. curtisii among three cohorts—pregnant women, pediatric patients, and sexually inexperienced women. The prevalence of antibody in each of these three groups was 75, 6, and 0%, respectively. Of the three pediatric patients with antibody to Mobiluncus curtisii, two were neonates, and the only class of antibody detected was immunoglobulin G. Among the cohort of pregnant women, the presence of antibody could not be correlated with a clinical history of BV. Serum antibody to M. curtisii could be a useful serological marker for BV. The lack of correlation of antibody positivity to historical information regarding BV suggests that unrecognized or undiagnosed episodes of BV may be common. Bacterial vaginosis (BV) is the most frequent cause of vaginal discharge in the United States (16). Additionally, BV has been associated with complications of pregnancy and postsurgical gynecological infections and is epidemiologically related to pelvic inflammatory disease (7, 12). BV is a polymicrobial syndrome characterized by large concentrations of anaerobic and facultatively anaerobic organisms within the vaginal fluid (20). Although no single organism is known to be the etiologic agent of BV, Mobiluncus spp. are the bacteria which are most highly associated with this syndrome. Mobiluncus spp. are found in up to 85% of women with BV and are rarely isolated from normal controls (8, 17). They therefore represent an important marker of this condition. A serological assay for BV is not currently available. Such an assay would be a valuable tool for studying the role of BV in infections such as pelvic inflammatory disease and would help elucidate the pathogenesis of BV itself. Preliminary work in our laboratory has shown that systemic antibody is produced against Mobiluncus spp. Thus, we decided to investigate the utility of antibody to Mobiluncus spp. as a serological marker of BV in a cohort of pregnant women for whom sexual histories and current vaginal microbiological data were available.
IFA. All sera were screened for antibody to Mobiluncus spp. by an indirect immunofluorescence assay (IFA). Organisms were harvested from Columbia sheep blood agar plates after 4 to 5 days of anaerobic incubation. The organisms were washed three times in phosphate-buffered saline (PBS) and then adjusted to a concentration of approximately 1.5 3 108 organisms per ml. The organisms were immediately fixed in 3% formalin for 15 min, centrifuged, and then suspended in M199 medium (Sigma Chemical Co., St. Louis, Mo.). Approximately 3 3 105 organisms were spotted onto each well of multiple-well slides (Cel-line, Newfield, N.J.) and allowed to air dry for 1 h. Slides were fixed in 70% ethanol for 15 min. Diluted serum was applied to each well. The slides were incubated in a humidity chamber at 378C for 30 min and then washed three times for 2 min each time with PBS and once in deionized H2O. The slides were then incubated for 30 min at 378C with fluorescein-conjugated antiserum (either goat antihuman immunoglobulin G [IgG] [whole molecule], goat anti-human IgM [Cappel Research Products, Durham, N.C.], or goat anti-human IgA [Zymed Laboratories, San Francisco, Calif.]) at appropriate working dilutions made in PBS–2% Tween 80. The slides were washed as before and air dried. Coverslips were applied with Tris-glycerol mounting medium (pH 8.5) and examined with an epifluorescence microscope at a magnification of 3100. Initially, sera were screened against M. curtisii curtisii and M. mulieris as well as SLH-29 and SLH-2, previously described antigenic subgroups of M. curtisii and M. mulieris respectively (19). However, screening of the first 10 specimens confirmed the previous experience that M. mulieris (and SLH-2) consistently yielded negative results. Therefore, screening was continued using only M. curtisii curtisii and SLH-29. Sera were initially screened at a 1:8 dilution prepared with PBS–1% bovine serum albumin, and if samples tested positive they were diluted in doubling dilutions to 1:256. Adsorption studies. Adsorption studies were performed with two of the positive serum samples. The organisms (M. curtisii curtisii and M. mulieris) were each harvested from brucella agar plates and washed twice in PBS. Patient serum was adsorbed three times with the bacteria (5:1 [vol/vol] at 48C on ice for 1 h). Adsorbed sera were run against M. curtisii curtisii antigen in the IFA described above. Statistical analysis. Statistical comparisons were made by using the EpiInfo 6 software program (5). Fisher’s exact test or the chi-square test was used to compare categorical variables, and Wilcoxon’s test was used to compare continuous variables.
MATERIALS AND METHODS Serological specimens. Sera from 59 pregnant women who participated in a study of prematurity and infection were available. The sera were collected during the enrollment visit and stored at 2208C. Available historical data included basic demographics, weeks of gestation at the time of delivery, and sexual history (the number of lifetime partners and a history of sexually transmitted diseases and BV). Available clinical data included the presence or absence of BV as determined by vaginal Gram stain (15) at the time of enrollment as well as the presence or absence of Mobiluncus spp. as documented by culture performed according to methods described previously (17). Sera from four sexually inexperienced women were available to be used as negative controls. Additionally, sera from 47 pediatric patients undergoing testing at the Children’s Hospital in Birmingham, Ala., were available. For the latter group, each sample of serum was labeled with the patient’s date of birth and stored at 2208C.
RESULTS The adult female patients ranged in age from 17 to 38 years, with a median age of 25 years. The racial composition was 84% white–16% African-American. The median number of lifetime sexual partners was 4.5, and 20 of 59 (34%) of the patients admitted to a history of sexually transmitted diseases. The self-reported history of BV was low (7%); however, 25% of the patients were found to have BV at enrollment, and the presence of Mobiluncus spp. in 73% of these subjects was docu-
* Corresponding author. Mailing address: 1900 University Blvd., 229 Tinsley Harrison Tower, Birmingham, AL 35294-0006. Phone: (205) 975-5665. Fax: (205) 934-5155. Electronic mail address: jschwebke @uabid.dom.uab.edu. 567
568
SCHWEBKE ET AL.
CLIN. DIAGN. LAB. IMMUNOL. TABLE 2. Prevalence of serum IgG antibody to M. curtisii by cohort
TABLE 1. Characteristics of patients with and without serum IgG antibody to M. curtisii Resulta for patients with: Characteristic
Antibody present (n 5 44)
Antibody absent (n 5 15)
Median age (yr) Race (no. [%] African-American) Median no. of lifetime sexual partners Median gestation at delivery (wk) Previous infections BV (no. [%]) STDb (no. [%]) Infections at enrollment BV (no. [%]) M. curtisii (no. [%]) Serum IgM to M. curtisii (no. [%]) Presence of BV (historical or current) or Mobiluncus spp. (no. [%])
25 5 (11) 4.5 37.5
22 4 (27) 4.5 39
4 (9) 13 (29)
0 (0) 7 (47)
12 (27) 4 (9) 18 (41) 22 (50)c
3 (20) 1 (7) 9 (60) 1 (7)
a For age, P 5 0.04; for presence of BV (historical or current) or Mobiluncus spp., P 5 0.02. P values for all other comparisons were not significant. b STD, sexually transmitted diseases. c IgG or IgM seropositive.
Cohort
No. IgG positive/no. IgG negative (%)
Sexually inexperienced women Pediatric patients Pregnant women
0/4 (0) 3/47 (6)b 44/59 (75)
a b
Pa
0.006 ,0.001
P value from comparison with cohort of pregnant women. Two of the three IgG-positive patients were newborns.
fluorescence, suggesting that the assays were specific for M. curtisii. The pediatric control patients ranged in age from 18 days to 10 years. IgG antibody to M. curtisii was detected in 3 of 47 (6%) of these patients. Two were neonates with titers of 1:32 and 1:128, while a 3-year-old girl had a titer of 1:16. IgM antibody to M. curtisii was not detectable in these three patients (Table 2). DISCUSSION
mented. Mobiluncus spp. were isolated from an additional five patients without vaginal Gram stain evidence of BV. IgG antibodies to M. curtisii (titer . 1:8) were present in 44 of 59 (75%) of the women. Of these, 35 of 44 (80%) had titers of $1:32. The presence of antibody to SLH-29, an antigenic variant of M. curtisii, paralleled that of antibodies against typical M. curtisii in all cases, although titers against SLH-29 were generally one to two dilutions lower. IgM antibody was detected in 27 patients, including 8 patients in whom IgG was not detected. Titers for IgM ranged from 1:8 to 1:64, the majority being between 1:16 and 1:32. Circulatory IgA was not detected in any of the patients. Antibody was not detected in any of the four sexually inexperienced controls (see Table 2). A comparison of patients with and without antibody to M. curtisii is shown in Table 1. There were no significant differences between the groups with respect to race or median number of weeks of gestation at delivery. Although their number was not statistically significant, women who recalled a prior episode of BV or who had evidence of BV at the time of enrollment more often had IgG antibody to Mobiluncus curtisii. All four of the women who recalled a prior episode of BV had this antibody. Of the 15 women who had BV at enrollment, 12 (80%) were IgG positive. This antibody was also present in 30 of 42 (71%) of the women without BV at enrollment (vaginal cultures were unavailable for two patients). The patients with IgG antibody were significantly older than the women without this antibody (P 5 0.04). Of the 20 women who recalled a history of sexually transmitted diseases, 13 (65%) were antibody positive. There was no apparent association between the presence of IgM antibody and the presence of BV and/or Mobiluncus spp. at enrollment. However, if serum antibody positivity to Mobiluncus curtisii was defined by the presence of IgG or IgM, the women with either a history of BV or current evidence of BV and/or Mobiluncus spp. were significantly more likely to be antibody positive than the women without these clinical findings (P 5 0.02). The results of the IFA using M. curtisii curtisii as the antigen and sera adsorbed with M. mulieris were identical to those of the unabsorbed sera (positive titers). Sera from the same patients whose sera had been adsorbed with M. curtisii showed no
The immunological factors which are involved in disease initiation and modulation are poorly understood for human vaginal infections, especially BV. In contrast, the immunology of bovine vaginal infections, such as bovine trichomoniasis and genital vibriosis (caused by Campylobacter fetus), has been studied in detail and vaccines are widely used to control both of these conditions (2, 3, 18). The immunology of periodontal disease, another mixed anaerobic infection within a mucosal site, has also been described (22, 23). A better understanding of the immunology of vaginal infections could provide helpful insight into understanding the pathogenesis of these diseases. Vibrio-like organisms have long been demonstrated in patients with BV (4, 9, 15, 16). Currently, Mobiluncus spp. are recognized as the predominant “vibrio” in this syndrome, although others such as Wolinella spp. and C. fetus have been identified (8, 21). Mobiluncus spp. are highly associated with BV, being found in up to 85% of patients with this condition, and are rarely isolated from women with Lactobacillus-predominant flora (8, 17). Thus, of all the organisms associated with BV, Mobiluncus spp. are the most specific markers. Antibodies to Mobiluncus spp. have been previously demonstrated by others. Larsson et al. reported a case of postoperative infection due to Mobiluncus spp. (species not determined) and anaerobic cocci in a patient who had undergone abdominal hysterectomy. By indirect immunofluorescence, they found antibody to M. curtisii at an initial titer of 1:320. Interestingly, the serum did not show reactivity against M. mulieris (11). Larsson et al. further reported significant increases in antibody titer to Mobiluncus spp. among three patients monitored for abnormal uterine bleeding (10). Moi et al. studied antibody against Mobiluncus spp. among women with and without BV as well as a group of male blood donors. They found IgG antibody to Mobiluncus spp. to be present among most of the subjects, although the actual numbers were not provided. The subset of women with BV had significantly higher titers than did the other groups. Titers of antibody to M. mulieris were significantly lower than those to M. curtisii. IgM antibodies were most often detected among the women with BV and were not seen at all among the males. Serum IgA was demonstrated in a few of the women with BV but in none of the other patients (14). The presence of IgG antibody against M. curtisii in 75% of
VOL. 3, 1996
SEROLOGICAL MARKER FOR BACTERIAL VAGINOSIS
the women in our cohort was higher than we anticipated. The two possible explanations are (i) that the antibody response detected was nonspecific or (ii) that BV occurs more frequently than was previously appreciated. Several pieces of evidence argue against the former. No antibody response to M. mulieris was detected among a subset of 10 of the patients studied, yet antibody responses for SLH-29, an antigenic subgroup of M. curtisii (19), paralleled those for M. curtisii in all cases. Secondly, the majority of titers were $1:32. The results of the adsorption study strongly suggest that the antibodies responsible for the observed fluorescence in the IFA were specific, since they were completely removed by adsorption with M. curtisii and not affected by adsorption with M. mulieris. Among the control patients, antibody was not detected in the four sexually inexperienced women. Finally, antibody to Mobiluncus curtisii was detected in only 3 of 47 pediatric control patients, 2 of whom were newborns. The absence of IgM in these patients suggests passive transfer of antibody to the neonates from the mothers. The prevalence of diagnosed BV varies with the population studied, from 25 to 60% (1, 6, 13). In our cohort, the prevalence of BV at enrollment was 25%; however, only four patients admitted to a prior history of BV. It is known that 50% of women with BV are asymptomatic (1), and it is highly likely that many women have had prior episodes of BV which were either asymptomatic or underdiagnosed. The actual lifetime prevalence of BV among women could be quite high, as our antibody assay suggests. Therefore, any comparisons between the presence of antibody and the history of disease are compromised by the lack of accurate information on the incidence of BV. In summary, we have demonstrated the potential usefulness of the presence of antibody to Mobiluncus curtisii as a serological marker for BV. The seroprevalence of IgG antibody to M. curtisii in a cohort of pregnant females was found to be 75%, suggesting that BV may be more common than has been previously recognized. Further investigation into the timing and effect of systemic and local antibody to Mobiluncus spp. as a serological marker for BV could help to elucidate the pathogenesis of this complex syndrome. ACKNOWLEDGMENTS This work was supported by a grant from the NIH (PHSUO1AI31494). We thank David Kelly for his assistance in obtaining pediatric sera. REFERENCES 1. Amsel, R., P. A. Totten, C. A. Spiegel, K. C. S. Chen, D. Eschenbach, and K. K. Holmes. 1983. Non-specific vaginitis. Diagnostic criteria and microbial and epidemiologic associations. Am. J. Med. 74:14–22. 2. BonDurant, R. H., R. R. Corbeil, and L. B. Corbeil. 1993. Immunization of virgin cows with surface antigen TF 1.17 of Tritrichomonas foetus. Infect. Immun. 61:1385–1394.
569
3. Corbeil, L. B., J. R. Duncan, G. D. Schurig, C. E. Hall, and A. J. Winter. 1974. Bovine venereal vibriosis: variations in immunoglobulin class of antibodies in genital secretions and serum. Infect. Immun. 10:1084–1090. 4. Curtis, A. H. 1913. A motile curved anaerobic bacillus in uterine discharges. J. Infect. Dis. 12:165–169. 5. Dean, A. G., J. A. Dean, and D. Columbeer. 1994. EpiInfo, version 6: a word processing database, and statistics program for epidemiology on microcomputers. Centers for Disease Control and Prevention, Atlanta. 6. Embree, J., J. J. Caliando, and W. M. McCormack. 1984. Non-specific vaginitis among women attending a sexually transmitted diseases clinic. Sex. Transm. Dis. 11:81–84. 7. Eschenbach, D. A. 1993. Bacterial vaginosis and anaerobes in obstetricgynecologic infection. Clin. Infect. Dis. 16:S282–S287. 8. Holst, E., B. Wathne, B. Hovelius, and P.-A. Mårdh. 1987. Bacterial vaginosis: microbiological and clinical findings. Eur. J. Clin. Microbiol. 6:536–541. ¨ ber die Natur der Scheidenkeime, speziell u 9. Kro¨nig, I. 1895. U ¨ber das Vorkommen anaerober Streptokokken im Scheidensekret Schwangerer. Zentralbl. Gynaekol. 19:409–412. 10. Larsson, P. G., B. Bergman, U. Forsum, and C. Påhlson. 1990. Treatment of bacterial vaginosis in women with vaginal bleeding complications or discharge and harboring Mobiluncus. Gynecol. Obstet. Invest. 29:296–300. 11. Larsson, P. G., B. Bergman, C. Påhlson, and L. Gotthardsson. 1986. Mobiluncus-specific antibodies in a post operative infection. Am. J. Obstet. Gynecol. 154:1167–1168. 12. Martius, J., and D. A. Eschenbach. 1990. The role of bacterial vaginosis as a cause of amniotic fluid infection, chorioamnionitis and prematurity—a review. Arch. Gynecol. Obstet. 247:1–13. 13. Moi, H. 1990. Prevalence of bacterial vaginosis and its association with genital infections, inflammation, and contraceptive methods in women attending sexually transmitted disease and primary health clinics. Int. J. Sex. Transm. Dis. AIDS 1:86–94. 14. Moi, H., G. Fredlund, E. To¨rnqvist, and D. Danielsson. 1991. Mobiluncus species in bacterial vaginosis: aspects of pathogenesis. APMIS 99:1049–1054. 15. Nugent, R. P., M. A. Krohn, and S. L. Hillier. 1991. Reliability of diagnosing bacterial vaginosis is improved by a standardized method of Gram stain interpretation. J. Clin. Microbiol. 29:297–301. 16. Rein, M. F., and K. K. Holmes. 1983. “Non-specific vaginitis,” vulvovaginal candidiasis, and trichomoniasis: clinical features, diagnosis and management. Curr. Clin. Top. Infect. Dis. 4:281–315. 17. Roberts, M. C., S. L. Hillier, F. C. Schoenknecht, and K. K. Holmes. 1985. Comparison of Gram stain DNA probe and culture for the identification of species of Mobiluncus in female genital specimens. J. Infect. Dis. 152:74–77. 18. Schurig, G. D., J. R. Duncan, and A. J. Winter. 1978. Elimination of genital vibriosis in female cattle by systemic immunization with killed cells or cellfree extracts of Campylobacter fetus. J. Infect. Dis. 138:463–472. 19. Schwebke, J. R., S. A. Lukehart, M. C. Roberts, and S. L. Hillier. 1991. Identification of two new antigenic subgroups within the genus Mobiluncus. J. Clin. Microbiol. 29:2204–2208. 20. Spiegel, C. A., R. Amsel, D. Eschenbach, F. Schoenknecht, and K. K. Holmes. 1980. Anaerobic bacteria in non-specific vaginitis. N. Engl. J. Med. 303:601–607. 21. Spiegel, C. A., and M. Roberts. 1984. Mobiluncus gen. nov., Mobiluncus curtisii subsp. curtisii sp. nov., Mobiluncus curtisii subsp. holmesii subsp. nov., and Mobiluncus mulieris sp. nov., curved rods from the human vagina. Int. J. Syst. Bacteriol. 34:177–184. 22. Underwood, K., K. Sjo ¨stro¨m, and R. Darveau. 1993. Serum antibody opsonic activity against Actinobacillus actinomycetemcomitans in human periodontal diseases. J. Infect. Dis. 168:1436–1443. 23. Whitney, C., J. Ant, B. Moncla, B. Johnson, R. C. Page, and D. Engel. 1992. Serum immunoglobulin G antibody to Porphyromonas gingivalis in rapidly progressive periodontitis: titer, avidity, and subclass distribution. Infect. Immun. 60:2194–2200.
