Prevalence of Mycoplasma genitalium in humans is still not clear. We have developed a sensitive and specific serological assay for M. genitalium using lipid-associated membrane proteins (LAMPs) as antigens. Antibodies to LAMPs from M. genitalium showed little cross-reactivity to LAMPs from antigenically similar M. pneumoniae. For validity testing, urines from 104 patients were tested by PCR for M. genitalium. All 15 PCR+ patients had M. genitalium-LAMPs antibodies. Moreover, none of 64 antibody-negative patients were PCR+. Serological study of 1800 patients of various diseased groups and healthy blood donors showed M. genitalium was primarily a sexually transmitted microbe that infected patients with AIDS (44.0%), intravenous drugs users with or without HIV infection (42.5%), and also HIV− patients attending STD clinics (42.6%). Only 5.5% HIV− healthy blood donors and 1.3% HIV+ hemophiliacs tested positive. M. genitalium has been associated with acute non-gonococcal urethritis in male patients. However, many sexually active men and women appear to be chronically infected or colonized by the microbe without apparent clinical symptoms and may continue to transmit the organism through sexual contacts.
Mycoplasma genitalium was originally isolated from the urethra of two homosexual patients with non-gonococcal urethritis . Although subsequent studies suggested organisms were present in urogenital tracts of male and female patients , few further isolations of the mycoplasma have been reported. This lack of isolation is most likely because the organism is too fastidious to grow in present culture environments. The organism, along with M. pneumoniae was later identified in throat specimens from patients with respiratory diseases, suggesting that the respiratory tract could also be a primary site of infection for M. genitalium. Recently, both of these two closely related mycoplasmas were isolated in mixture from synovial fluid . Studies also showed infection of M. genitalium may be associated with development of acute non-gonococcal urethritis (NGU) [5–7]. Interest in M. genitalium increased following a report from Luc Montagnier and associates detecting the organism by PCR in blood of one patient with AIDS . The finding suggested there might be higher incidence of systemic infection by M. genitalium in HIV-infected patients with AIDS. However presently, prevalence of the microbe in various AIDS risk groups and in the healthy general population is still largely unclear.
In order to study distribution, scope of infection, mode of transmission of M. genitalium, and the possible disease processes associated with its infection, a serological test capable of detecting evidence of infection by the organism with high sensitivity is needed. Most importantly, the test must adequately differentiate with high specificity M. genitalium from other mycoplasmas, especially M. pneumoniae which shares many M. genitalium antigenic properties [9–12]. Many individuals may have significant antibody titers to M. pneumoniae due to previous exposure to this common respiratory pathogen. A good assay should measure antibodies reacting specifically with M. genitalium, but not M. pneumoniae cross-reacting antibodies.
Mycoplasmal lipid-associated membrane proteins (LAMPs) and adhesin proteins are exposed on the cell surface [13–17], highly antigenic , and are likely important immunogenic targets for hosts' responses in mycoplasmal infections . Antibodies to LAMP antigens of each individual species of Mycoplasma are apparently highly species-specific and do not cross-react with those of other species. Thus, mycoplasmal LAMPs have been used as the target antigens with a high degree of specificity in serological assays detecting M. penetrans, M. salivarium, or M. pirum-specific antibodies in serum samples from patients with a wide spectrum of clinical illness [19, 20].
In this study, we specifically compared LAMP antigens of M. genitalium and those of M. pneumoniae. Although highly prevalent in the normal subjects, antibodies to M. pneumoniae LAMPs would not cross-react with M. genitalium LAMPs. The M. genitalium-specific antibody assay using LAMPs strongly correlated with PCR evidence of M. genitalium infection in the urines in a group of IVDUs and a group of HIV-infected asymptomatic (AD) patients. M. genitalium-specific antibodies were not prevalent in the general population. In contrast, M. genitalium-specific antibodies were highly prevalent in male homosexuals and intravenous drug users (IVDUs), with or without HIV-1 infection, as well as in HIV-negative non-AIDS patients attending clinics. Most of these sexually transmitted diseases (STD) clinic patients did not have symptoms of acute NGU. Western blotting (WB) showed the antibodies' major immunoreactivity was directed against six M. genitalium LAMPs, particularly P160, P150, P123, P112, P66 and P62. Although earlier reports with PCR suggested a relationship between M. genitalium and acute NGU [6, 7], this seroepidemiological study has revealed a markedly higher level of prevalence of specific M. genitalium antibodies in many sexually active groups of patients, but not in the general healthy population. Infection and host colonization by the organism through sexual contacts are apparently much more frequent than previously recognized.
Materials and methods
The serum sample preparations, the procedures for preparation of LAMPs from both mycoplasmas, the ELISA and WB techniques of studying antibodies to LAMPs were all described previously in detail [19, 21]. Briefly, 0.5 µg/ml M. genitalium or M. pneumoniae LAMPs were coated on Nunc-Immuno F96 MaxiSorp plates with 100 µl in each well. After overcoating with 0.1% BSA, 100 µl human serum or plasma (diluted 1/250 in 10% normal goat serum, 2% BSA and 0.3% Nonidet P-40 (NP-40) in PBS) was added to each well. Subsequent steps included addition of 100 µl of 1/1000 biotin-labeled antibody of goat anti-human IgG-λ (Kirkegaard and Perry Laboratories (KPL Inc.) and 1/20 000 peroxidase-labeled streptavidin (KPL) that were prepared in 10% normal goat serum, 2% BSA, 0.1% NP-40, 1×PBS (Diluent I). The plates were developed by adding 100 µl of 2,2 β-azino-di-[3-ethyl-benzthiazoline-sulfonate] peroxidase substrate solution into each well. The plates were washed 6 times with PBS (pH 7.2) plus 0.05% NP-40 (Solution A) between each step. Positive samples were repeated in triplicate and tested in plates coated with BSA without LAMPs.
In WB analysis, LAMPs (about 30 µg) from M. genitalium or M. pneumoniae TX-114 extract were separated by SDS-polyacrylamide gel (14 cm×12 cm×0.75 cm) electrophoresis and electroblotted on a BA-85 nitrocellulose membrane (Schleicher and Schuell). The membrane was blocked with 10% fetal bovine serum and 1% BSA in PBS pH 7.2 and cut into 4 mm strips. Each strip was incubated with 2 ml of 1/250 human serum at 25°C for 15 h with shaking. The strips were washed six times with Solution A, incubated at 25°C with 1/1000 biotin-labeled antibody of goat anti-human IgG-γ for 3 h, incubated at 25°C with 1/10 000 peroxidase-labeled streptavidin in Diluent I for 90 min, and developed at 37°C for 20 min with the 4-chloro-1-naphthol peroxidase substrate system (KPL).
Urine sediments for PCR amplification were collected by centrifugation at 3000×g for 15 min at 5°C and the resulting pellets resuspended in 1/10 volume of the original urine. 500 µl of each urine sediment was digested with 200 µg/ml of proteinase K in 50 mM Tris-HCl, pH 8.3, 5 mM EDTA at 56°C for 2 h, then at 37°C for 1 h, followed by extraction with equal volume of phenol-chloroform (vol/vol=1:1) for 30 min. The urine extract was desalted and concentrated twice by diluting with 5 mM Tris-HCl, pH 8.0, 0.5 mM EDTA to 2 ml and centrifugation at 3600×g for 25 min in Centricon-30 (Amicon, Beverly, MA) each time, and finally recovered in 50–100 µl. The M. genitalium-specific primers and probe for detection of M. genitalium by PCR were as previously described . PCR was performed in an automated DNA thermal cycler (PE 9600; Perkin Elmer, Foster City, CA) with 100 µl reaction mixture containing 10 µl urine samples, 200 µM each of dATP, dCTP, dGTP, and dTTP; 0.3 µM of each primer, and 3 U of AmpliTaq® DNA polymerase (Perkin Elmer, Foster City, CA) in 10 mM Tris-HCl, pH 8.3, 50 mM KCl, 2 mM MgCl2, and 0.001% gelatin. Samples were denatured at 94°C for 25 s, and primers were annealed at 55°C for 40 s with 2 s increments for each additional cycle and extended at 72°C for 30 s for a total of 45 cycles. The positive PCR samples were first identified by dot-blot hybridization with 32P-labeled probe and further verified by size fractionation on 6% polyacrylamide gel and Southern blotting .
Distribution of reaction intensity to M. genitalium LAMPs in ELISA by serum samples from normal blood donors
M. genitalium LAMPs were isolated, purified and coated on ELISA plates to test antibody reactivity in serum samples from 384 healthy blood donors. Fig. 1 shows distribution of ELISA OD readings produced by the serum samples from these control subjects. Less than 4% of individuals in the general population had antibodies that produced an immune reaction with intensity measured at OD405 higher than 1.0. More than 75% of individuals had an OD405 reading less than 0.1. The mean of the sum OD405 readings for sera from the control subjects with reading =0.7 (n=365) plus 4 standard deviations (S.D.) was used as the cut-off for positivity in the ELISA test in this study.
Antibody cross-reactivity to LAMPs from M. genitalium and to LAMPs from the closely related species M. pneumoniae
Since previous studies clearly showed M. pneumoniae and M. genitalium share many antigenic determinants, LAMPs prepared from M. pneumoniae were tested in parallel for antibody reactivity in the identical 384 serum samples from the blood donors previously tested for M. genitalium. Contrary to the finding of low prevalence of M. genitalium antibodies, more than 60% of these normal subjects had antibodies that produced reactions with OD405 readings higher than 1.0, and more than 80% with higher than 0.8 (Fig. 1). Thus, antibodies to LAMPs of M. pneumoniae are apparently highly prevalent in the general healthy population, but these M. pneumoniae-specific antibodies evidently do not react with M. genitalium LAMPs.
Western blotting for antibodies to LAMPs of M. genitalium and M. pneumoniae
The immune reactivity of antibodies in serum from representative individuals tested positive or negative to LAMPs from M. genitalium and M. pneumoniae by ELISA was further analyzed with Western blot (WB). For M. genitalium, 160, 150, 123, 112, 66, and 62 KDal molecular mass proteins were the most prominent LAMP antigens recognized by serum from patients with AIDS or non-AIDS patients from STD clinics who had positive M. genitalium-specific antibodies (strips G–N) (Fig. 2). These serum samples produced reactions with OD405 readings higher than 1.0 in ELISA. In comparison, two other groups of serum produced little reaction in WB: serum that tested positive for antibodies to M. pneumoniae LAMPs (OD>1.0) but negative for antibodies to M. genitalium LAMPs in ELISA (strips D–F) and serum tested negative for both M. genitalium and M. pneumoniae LAMPs in ELISA (strips A–C). Antibodies to M. pneumoniae LAMPs were similarly analyzed in WB (Fig. 3). Two prominent LAMP antigens, 180 KDal and 120 KDal molecular mass proteins, were consistently recognized with high intensity by serum from patients with positive M. pneumoniae-specific antibodies (OD>1.0 in ELISA) (strips G–N). Some other LAMPs with lower molecular masses were also occasionally recognized by serum with positive antibodies to M. pneumoniae. More importantly however, two other groups of serum produced no or little reaction in WB: serum that tested positive for antibodies to M. genitalium LAMPs but negative to M. pneumoniae LAMPs in ELISA (strips D–F) and serum tested negative for both M. genitalium and M. pneumoniae LAMPs in ELISA (strips A–C).
Correlation of antibody positivity to M. genitalium LAMPs in serum and PCR positivity for M. genitalium DNA in urine
To further verify the validity of the serological assay, we collected urines from some of the patients (n=104) whose serum was tested for M. genitalium LAMPs-specific antibodies. Urines from 68 HIV-infected asymptomatic (AD) patients (24 seropositive and 44 seronegative) and 36 IVDUs (16 seropositive and 20 seronegative) were blindly examined by PCR for presence of M. genitalium DNA. Details of this PCR study is described in Section 2. In this study, 7 of 68 urines from AD patients and 8 of 36 urines from IVDUs were PCR-positive (Table 1). All the IVDUs or AD patients who had evidence of active infection by M. genitalium in the urogenital tract, i.e. the 15 patients tested positive by PCR, tested positive for M. genitalium-specific antibodies in ELISA. Equally important, none of the 64 IVDUs and AD patients tested negative for M. genitalium antibodies were found PCR-positive. Thus, the PCR study of urine samples from this group of 104 patients revealed an excellent correlation between positive evidence of M. genitalium infection or colonization by PCR and positive serum antibodies to M. genitalium LAMPs (Odds Ratio: 78.4, 95% CI: 4.5–1361; Fisher's Exact Test: two-sided P value>0.0001). Overall, there were 25 out of 104 patients who were seropositive but PCR-negative for the mycoplasma in their urines. The serological assay for antibodies to M. genitalium LAMPs is apparently highly sensitive detecting all the patients known to be infected by the mycoplasma.
|Patient group||Number tested||Antibodies by ELISA||PCR positivity in urines||Percent (%) with PCR positivity|
|HIV+ AD||68||+||24||7||7/24 (29%)|
|Patient group||Number tested||Antibodies by ELISA||PCR positivity in urines||Percent (%) with PCR positivity|
|HIV+ AD||68||+||24||7||7/24 (29%)|
High prevalence of antibodies to M. genitalium LAMPs in HIV-negative patients attending STD clinics
In ELISA using M. genitalium LAMPs, 44% of patients with AIDS (n=341) and 32.3% of HIV-infected AD patients (n=127) were positive for M. genitalium-specific antibodies. We also found 42.5% intravenous drugs users (n=308) (59 of 137 HIV+ patients and 72 of 171 HIV− patients) had antibodies to M. genitalium. In contrast, only 5.5% of normal blood donors (n=384) and 3.5% of patients with various malignant diseases (n=144) had antibodies that reacted with M. genitalium LAMPs. In addition, we tested serum from 165 patients with hemophilia (79 HIV-positive and 86 HIV-negative). Only 1.3% of sera from hemophiliac patients tested positive for antibodies to M. genitalium. However, 141 of 331 (42.6%) HIV-negative patients attending STD clinics in three different geographic areas, Brooklyn, Milwaukee and southern California tested positive for antibodies to the mycoplasma (Fig. 4) The results show that M. genitalium is not a simple opportunistic infection occurring in HIV-infected immune compromised patients, but is instead, a highly prevalent sexually transmitted mycoplasma. The surprisingly high rate of infection by this mycoplasma in the STD clinic patients as well as IVDUs has not previously been appreciated . In comparison, many fewer samples from cancer patients in the hospital, hemophiliacs with or without HIV infection, or the healthy general population known to have lower rate of sexual contacts tested positive.
Identifying infections due to M. genitalium is difficult in that culturing is rarely successful and usual histopathological techniques designed to reveal bacteria do not stain the wall-free microbe. PCR has been shown to be successful in detecting M. genitalium infection including more recently in the joints of patients with rheumatoid arthritis . However, preparation of clinical samples, especially those suspected to carry PCR inhibitory factor(s) , as well as analysis of artifacts or contaminants easily introduced at any step of the procedure have to be meticulously monitored. Complications such as these have prevented most clinical microbiology laboratories from using PCR as a general tool for diagnosis of mycoplasmal infection. Measurement of incidence or frequency of infection in a large group of patients with various clinical presentations is particularly difficult. In addition, the primary site of infection or colonization must first be identified before a meaningful assessment by PCR can be accomplished.
Current serological assays detecting mycoplasma infections, like those detecting other bacterial infections, are cumbersome to perform and are generally less sensitive and less specific compared to serological assays detecting viral infections. Metabolic inhibition (MI) and growth inhibition (GI) assays measure antibodies inhibiting metabolism or hindering growth of a specific species of mycoplasma [27, 28]. These assays require careful and stringent controls of culture conditions, inoculation titers of mycoplasma as well as culture incubation times. Non-specific, cross-reacting antibodies to house-keeping proteins commonly shared by different mycoplasmas as well as other inhibitory factors such as various antibiotics in serum samples of patients often complicate the analysis. Thus, MI and GI serological assays have not been widely applied by clinical microbiology laboratories. A sensitive and specific serological assay such as ELISA or WB for rapid detection of M. genitalium infections, similar to the techniques used in viral infections, is needed. However, the antigenic complexity of mycoplasmas is much greater than that of viruses. Previously, the unique target antigens on mycoplasmas for human hosts' serological responses were not identified.
LAMPs are highly antigenic, species-specific, lipid-modified proteins anchored on the exterior surface of mycoplasma membranes [13, 14]. Thus, they are the most likely candidates to serve as major immunogenic targets of host serological responses . In this study, we further demonstrated the uniqueness or the specificity of antigenic epitopes of LAMPs, even between those species of mycoplasmas previously known to be antigenically closely related, such as M. genitalium and M. pneumoniae[9–11]. WB also clearly confirmed little cross-reactivity between M. genitalium-specific antibodies and M. pneumoniae-specific antibodies to LAMPs of each species. Four major LAMPs (P160, P150, P112 and P66) of M. genitalium appeared to be most consistently recognized by M. genitalium-specific antibodies. In addition, 104 cryopreserved urine samples collected from a group of 36 IVDUs and 68 AD patients were examined by PCR for M. genitalium. Our results showed every patient with a positive urine by PCR (n=15) tested positive for the LAMP antibodies to the mycoplasma. And none of the antibody-negative patients (n=64) had a positive PCR reaction. Thus, this test of antibodies to M. genitalium LAMPs had a high sensitivity detecting 100% of known infected persons (true positives). Epidemiologically, one would expect a test with high specificity to give negative results for samples from individuals in a truly negative population. In this study, the individuals of ‘true negatives’ who had never been infected or exposed previously to M. genitalium could not be known for sure. Although some individuals in the otherwise low-exposure general population must have been exposed to this mycoplasma before, 95% tested negative by this presumably highly sensitive test. In other populations that were also much less likely to be infected by or exposed to this organism through multiple sexual contacts such as cancer patients in the hospital and hemophiliacs with or without HIV infection, 96% and 99% tested negative, respectively. This mycoplasmal antibody test clearly had a high specificity as well.
PCR of 104 patients' urines revealed strong association between positive finding of the serum IgGs in the patients and finding M. genitalium in their urines (Fisher's exact test: P<0.0001). Evidently more patients tested positive for antibodies that identified both patients with active infection that could be detected by PCR as well as patients who were recently exposed to the organism but the immune system cleared the organism. Some patients could have also received treatment of antibiotics during this period of time. In this context, since M. genitalium was previously identified in respiratory tracts of patients , some antibody-positive patients with a negative PCR in urines might have an infection at a site other than the urogenital tract. Likewise, some of them may have also cleared the organism, but antibodies persist.
In addition to male homosexuals and IVDUs, the newly developed serological test revealed that HIV-negative patients attending STD clinics also have nearly 10-fold higher frequencies of M. genitalium infection than the normal population. The extremely high incidence of M. genitalium infection in many groups of sexually active patients was not previously recognized. Since most of the HIV-negative patients attending STD clinics were likely heterosexuals, M. genitalium is clearly transmissible through both homosexual and heterosexual contacts. Recent studies from different laboratories suggested that other than chlamydia, mycoplasmas are probably the most common cause of NGU [5–7]. As noted earlier, recent PCR study of urine samples also showed significant association between M. genitalium infection and development of acute NGU . However, the data here suggest that patients who are colonized or infected by M. genitalium remain largely free of clinical symptoms such as acute NGU. Many of these sexually active patients appear to be unknowingly infected or colonized by M. genitalium and are likely to unknowingly transmit the organism to their sexual partners.
Although patients chronically infected or colonized by M. genitalium often have no apparent symptoms of acute illness, a distinct form of pathogenesis may be associated with persistent infection and parasitic colonization by certain mycoplasmas. We reported that prolonged close interaction between mycoplasmas and mammalian cells may gradually but significantly alter eukaryotic hosts' biological and physiological properties . Potential biological significance of chronic parasitism by bacterial and mycoplasmal prokaryotic agents of seemingly low virulence could include oncogenesis [29–32]. The true significance of pathogenesis associated with chronic infection or colonization by M. genitalium in urogenital tract awaits further study.
In this study, it is interesting to note that IVDUs appeared to have very high frequency of M. genitalium infection (Table 1) suggesting the organism might also be transmitted through parenteral routes. Previously, experimental infection with M. genitalium in chimpanzees through genital tract inoculation produced systemic infection and positive blood cultures showing that parenteral transmission of the organism may be possible. However, the finding that HIV-infected hemophiliacs had very low prevalence of M. genitalium infection (even lower than the control general population) supports the primary mode of transmission of M. genitalium is sexual, not parenteral. Therefore, IVDUs could more likely obtain M. genitalium infections through frequent unprotected sexual contacts. This would favor the argument that many IVDUs could have obtained HIV infections through frequent unprotected sexual contacts. On the other hand, M. genitalium may be able to survive in the shared needles better than in processed blood products.
The authors thank Dr. W.M. McCormack and Ms. M. Cummings of SUNY Health Science Center, Brooklyn, NY, Dr. S.L. Inhorn and Mr. J.R. Pfister of State Laboratory of Hygiene, Madison, WI, as well as Drs. S.S. Appleton, Loma Linda University, Loma Linda, CA, R.C. Thorsen, Disease Control, Riverside County, CA, and R. Alexander, San Bernardino County Laboratories, CA, for providing serum samples of patients attending STD clinics; Dr. Harvey Alter of Department of Transfusion Medicine, Clinical Center, NIH, Bethesda, MD, for providing serum samples of HIV-infected asymptomatic patients. The authors also thank Dr. D.J. Wear and Ms. S. Ditty for helping in manuscript preparation. This study was supported in part by the American Registry of Pathology.