A LARGE PROPORTION OF reproductive tract disease (RTD) syndromes, including urethritis, cervicitis, endometritis, and salpingitis, has no identified etiology and cannot be attributed to major sexually transmitted pathogens such as Chlamydia trachomatis or Neisseria gonorrhoeae. Many of these syndromes have serious sequelae such as pelvic inflammatory disease, scarring of fallopian tubes resulting in ectopic pregnancies and sterility, perinatal and puerperal morbidity, and increased risk of HIV transmission and acquisition. Polymerase chain reaction (PCR) tests to detect Mycoplasma genitalium inpatient specimens have allowed studies showing an association of this organism with several of these syndromes, including urethritis, 1–13 cervicitis, 14,15 and endometritis, 16 suggesting a role for M. genitalium in the etiology of these idiopathic cases.
Mycoplasma genitalium was first isolated from urethral specimens from 2 of 13 men with urethritis in 1981. 17 However, the difficulty of isolating this fastidious organism precluded epidemiologic studies assessing its association with RTD syndromes until PCR tests specific for the M. genitalium MgPa adhesin gene were developed. 18,19 Subsequent to these initial studies, other PCR assays to detect M. genitalium were devised, several targeting the MgPa adhesion gene 2–4,20 and 2 targeting the 16S ribosomal gene, 1,21 allowing several independent studies showing an association of M. genitalium with urethritis and nongonococcal urethritis (NGU) in men. 1–13 Other studies have also assessed the association of M. genitalium with RTD in women by PCR assays. Cohen et al. 16 detected this organism in the endometrium and demonstrated its association with endometritis among women with suspected pelvic inflammatory disease. In addition, 2 studies 14,15 have shown an association of M. genitalium with cervicitis, in contrast to a third study which failed to find an association. 22 These findings underscore the need for future clinical studies assessing the disease associations of, risk factors for, and appropriate treatment regimens for M. genitalium infection.
The currently used PCR assays detect M. genitalium-specific PCR products after agarose gel electrophoresis with 2,4,19 or without 3,18,20 subsequent hybridization with probes specific for the M. genitalium PCR products, by sequencing of PCR products, 12 or by real-time PCR. 21 However, these assays are not adapted for high-throughput processing and thus are not optimal for large-scale screening of patient specimens. In the current study, we describe an adaptation of an MgPa PCR assay 4 from a 48- to 96-well format thermocycler, incorporation of an internal control to detect samples inhibited for amplification, and development of a 96-microwell detection assay to facilitate throughput. We report the use of this assay with cervical and male urine specimens and compare the effect of different urine specimen processing methods on detection of M. genitalium infection by PCR.
Materials and Methods
Bacterial Strains, Plasmids, and Oligonucleotides
Bacterial strains and plasmids used in this study are listed in Table 1. Oligonucleotides used as primers or probes in this study were purchased from either Gibco Life Technologies (Rockville, MD) or Operon (Alameda, CA) and are listed in Table 2. Escherichia coli Top 10 strains containing pMgPa, pMH, or pIC plasmids were cultured on L agar plates containing 30 mg/mL kanamycin or 100 μg/mL ampicillin. Plasmids were purified with either the Wizard system (Promega, Madison, WI) or the Qiagen (Valencia, CA) plasmid preparation kit. M. genitalium strain G37 was cultured in SP4 broth with glucose. 17 Growth in this medium was indicated by fermentation of glucose (detected by a color change) and confirmed by the presence of typical clusters of M. genitalium cells viewed by DAPI (4′6-diamidino-2-phenylindole; Sigma, St. Louis, MO) staining 23 using an Epifluorescent microscope-HBO 50 (Carl Zeiss, Inc, Thornwood, NY) with a filter set 02 and F-achromat 40×/0.65 objective lens.
;)
TABLE 1: Strains and Plasmids Used in This Study
TABLE 2: Oligonucleotide Primers and Probes Used in This Study
Plasmid pMgPa was constructed by cloning the MgPa PCR product from M. genitalium strain G37 into pCR2.1-TOPO using the TA cloning kit according to the manufacturer's instructions (Invitrogen, Carlsbad, CA). The recombinant plasmid was transformed into Top 10 E. coli cells, also provided in the TA cloning kit.
Plasmid pIC, the internal control for the Mg PCR, consisted of the MgPa PCR product with the internal portion replaced with a larger, irrelevant piece of DNA (Fig. 1). The internal 117-bp Dde I-Eco RI fragment of the MgPa PCR product was replaced with the 458-bp Dde I-Eco RI fragment from the hhd BA gene (Hemophilus ducreyi hemolysin gene Accession no. U32175) in several steps. First, a 2000-bp fragment from the H. ducreyi hhd BA gene was amplified with primers hhdBAstart and InAR and then digested with Eco RI and ligated to the MgPa PCR product, also digested with Eco RI. Second, the correct construct containing the 5′ portion of the MgPa product and 3′ portion of the hemolysin PCR product was amplified with primers modMgPa1 and InAR. This PCR product was inserted into pCR2.1-TOPO, creating pMH, and transformed into Top 10 cells to facilitate the addition of the final fragment. The 3′ portion of the MgPa amplicon containing the sequences homologous to the 3′ primer (modMgPa3) was added by amplifying pMH with primers modMgPa1 and modPA3hhdA, a primer that contained sequence homologous to the 3′ region of the hhd BA amplicon and the modMgPa3 primer sequence. This PCR reaction was then diluted, reamplified with primers modMgPa1 and modMgPa3, confirmed by restriction digests, ligated into pCR2.1-TOPO, transformed into Top 10 cells, and sequenced. The resulting plasmid, pIC, was purified and used as the internal control template DNA. As constructed, the inhibition control is amplified with the same primers as the MgPa gene target, modMgPa1 and modMgPa3, and yields a 630-bp product (Fig. 1, lanes 1 and 3) as compared with a 258-bp product for the M. genitalium adhesin gene (Fig. 1, lanes 1 and 2).
Fig. 1: Construction and amplification of the internal control for the MgPa PCR assay. (A) Schematic of the inhibition control before cloning into pCR2.1-TOPO to create pIC. The Eco RI-Dde I fragment of the MgPa amplicon was replaced by a larger sequence of irrelevant DNA as described in Materials and Methods. Both the IC and MgPa targets amplified with primers modMgPa1 and modMgPa3. (B) The MgPa amplicon (258 bp) and internal control amplicon (630 bp) separated on a 1.2% agarose gel and stained with ethidium bromide. The PCR templates present in the reaction were: Lane 1—Both pIC and M. genitalium DNA. Lane 2—M. genitalium DNA. Lane 3—pIC.
PCR Assays
Four variations of the M. genitalium PCR assay targeting the MgPa gene were performed: 1) MgPa-S (PCR assay targeting the MgPa gene and using a 48-well thermocycler and a S outhern blot-based detection system as described), 4 2) MgPa-IS (the MgPa-S assay with the addition of an i nternal inhibition control), and 3) MgPa-IMW (MgPa PCR assay performed in the 96-well thermocycler with an internal inhibition control and a 96-well EIA m icro w ell detection system), and 4) the MgPa-MW (MgPa-IMW PCR performed without the addition of an i nternal inhibition control). Details of each assay are described subsequently in this article.
MgPa-S and MgPa-IS assays.
The method for the MgPa-S PCR assay on urine specimens in the NGU sample set has been described. 4 In this assay, no internal control was included, unlabeled primers were used, and the resulting unlabeled PCR products were detected by Southern blot hybridization using a biotinylated probe (Mg probe-S). The MgPa-IS assay is a modification of the MgPa-S assay in which 2 fg (32 copies) of pIC, the internal control, was added to the PCR reaction and allowed to amplify with the same primer set. Internal control-specific PCR products were detected as a 630-bp PCR product hybridizing to IC probe-S after detection of the 258-bp MgPa products with Mg probe-S using the same hybridization conditions and Southern blots.
MgPa-IMW and MgPa-MW assays.
For the MgPa-IMW assay, the PCR was performed in a Perkin Elmer 9600 thermocycler (Perkin Elmer, Boston, MA) using the following cycling conditions: 4 minutes at 94°C, 35 cycles of 10 seconds at 94°C, 50 seconds at 55°C, 45 seconds at 72°C, and an extension of 5 minutes at 72°C. Reaction mixtures contained a final volume of 100 μL, a final concentration of 1× amplification buffer (diluted from 10× amplification buffer; Promega, Madison, WI), 4 mmol/L MgCl2, 0.1 μmol/L each biotin-labeled primer (bModPA1 and bModPA3), 200 μmol/L each deoxynucleoside triphosphate, 2.5 U Taq polymerase (Promega), and 2 fg (32 copies) of the internal control plasmid, pIC. The thermocycling conditions were modified from previously published protocols using a 480 thermocycler 4,14,16 by shortening the thermocycling times from 3 hours (MgPa-S PCR) to 2.25 hours (MgPa-IMW PCR). Because the 9600 thermocycler has a heated lid, mineral oil can be omitted from the PCR reaction tubes, significantly facilitating subsequent pipetting of this sample for PCR product analysis.
Biotinylated PCR products generated in the MgPa-IMW PCR assay were allowed to bind to streptavidin-coated plates, denatured, and then detected with a digoxigenin labeled probe using the PCR ELISA kit (Roche Diagnostics Systems, Branchburg, NJ) according to the manufacturer's instructions-Protocol B and illustrated in Figure 2. Five μl of the PCR reaction was diluted in 160 μL of PBS, pH 7.5, 0.1% Tween 20, and incubated in streptavidin-coated wells of a 96-well plate at 42°C for 1 hour. Plates were washed and the bound biotinylated PCR products were denatured and incubated for 1 hour at 42°C with a digoxigenin-labeled probe (either the MgPa or the IC probe in separate wells) added at a concentration of 100 ng/mL in a 175-μL volume. After washing the plates to remove unbound probe, the bound digoxigenin-labeled probe was detected after incubation with antidigoxigenin peroxidase-conjugated antibody, washing to remove excess unbound antibody, followed by a color reaction with substrate ABTS (2,2′-Azino-di-3-ethylbenzthiazoline sulfonate). The resulting color production was quantitated at 405 nm in a microtiter plate reader (Titertek Multiskan MC, Flow Laboratories, McLean, VA) after 30 minutes incubation at 37°C followed by 2 hours at ambient temperature, both in the dark. A reagent blank containing all detection reagents except the PCR reactions was performed in parallel with the test samples and served to blank the plate reader.
Fig. 2: Schematic and performance of the MgPa-IMW assay. (A) Schematic. The PCR is performed using biotinylated primers and the resulting biotinylated MgPa PCR products are captured (through their biotin moieties) on streptavidin-coated microwell plates. After incubation with a digoxigenin-labeled probe at 42°C to allow hybridization to internal sequences of the PCR products, the microwells are washed to remove unbound probe. Peroxidase-labeled antibodies to digoxigenin are added and bind to the digoxigenin-labeled probe DNA, if present, and excess (unbound) antibodies are removed. M. genitalium-specific PCR products are then detected by the formation of blue color on addition of the substrate, ABTS, which forms a blue product in the presence of peroxidase. (B) Performance of the MgPa-IMW assay showing the lack of cross-reactivity between the MgPa and IC probes. Column 1 contains detection reagents alone. The following columns contain detection reagents and PCR products resulting from amplification of (2 and 5) M. genitalium DNA, (3 and 6) internal control DNA, and (4 and 7) M. genitalium and internal control DNA. The Mg probe-P (labeled with digoxigenin), added to columns 2–4, reacts only with wells containing M. genitalium, but not internal control, PCR products. Likewise, IC probe-P (labeled with digoxigenin), added to columns 5-7, reacts only in wells containing IC, but not M. genitalium, PCR products. Serial 2-fold dilutions of these PCR products, added as indicated to rows 1-7, show decreasing amounts of the blue indicator color with decreasing amounts of PCR product. The negative controls (containing no M. genitalium or pIC template in the PCR reaction) are clearly negative.
The MgPa-MW assay was performed identical to the MgPa-MW assay except without the inclusion of the internal control in the PCR assay and without the microwell assay using the IC probe.
Determination of the Limit of Detection of M. genitalium DNA
Whole-cell DNA from M. genitalium strain G37 was purified using Qia DNA mini kit as recommended by the manufacturer (Qiagen, Valencia, CA). Concentrations of purified M. genitalium whole-cell DNA and pIC were determined using a DNA Fluorometer (model TKO100; Hoefer Scientific Instruments, San Francisco, CA), calibrated human placenta DNA (Sigma) as an internal reference standard, and Hoecht dye 33258 (Sigma) according to directions (provided by Hoefer). The number of genome equivalents of M. genitalium genomic DNA was then calculated by converting micrograms of DNA to micromoles based on the size of the genome (580 kbp; 1 kb of DNA is 6.6 × 105 Da) followed by division by the number of molecules in a mole (Avogadro's number, 6.23 × 10 23). A similar analysis was performed with pIC (4531 bp).
To determine the limit of detection of M. genitalium DNA in the PCR assays, 16 replicates of each concentration (128–1 genome copy) of serial 2-fold dilutions of M. genitalium DNA were analyzed using the MgPa-MW assay and a positive cutoff of 0.11. Similarly, 16 replicates each of the MgPa-IMW assay with 64, 32, or 16 copies of pIC were analyzed. This procedure measures the probability of a positive test as a function of the mean number of genome copies detected per tube.
Patient Specimens
Specimens from 2 clinical studies were used to evaluate the MgPa-IMW assay by comparing results from either the MgPa-IS or MgPa-S assays to those of the MgPa-IMW assay. The University of Washington institutional review board approved all studies. In the “cervicitis” study, cervical exudates, collected in 1984-1985 from women seen at the Seattle STD clinic, were frozen at -20°C until analysis, processed to remove inhibitors, then used to show the association of M. genitalium with mucopurulent cervicitis using the MgPa-IS PCR assay as described, 14 although the development of the internal control and details of the MgPa-IS procedure were not described. In the current study, a subset of these cervical specimens (50 MgPa-IS-positive and a random sample of 50 MgPa-IS-negative specimens) was used to set cutoff levels of the MgPa-IMW assay and test the concordance of the MgPa-IMW and MgPa-IS PCR assays.
Urine specimens from a second clinical study designated the “NGU” study in the current paper had been previously used to show an association of M. genitalium with NGU using the MgPa-S assay as described. 4 In the current study, all 32 MgPa-S-positive and an equal number of MgPa-S-negative urine specimens (the next consecutive negative specimen after a MgPa-S-positive specimen) were analyzed by the MgPa-IMW assay.
Specimen processing, cervicitis study.
Specimens from the cervicitis study were processed using the AMPLICOR CT/NG specimen preparation kit according to the manufacturer's directions (Roche, “swab procedure”). If the specimen was inhibited for PCR as determined by the internal control in the Southern blot-based MgPa-IS assay, it was diluted 1:5 and reanalyzed by the MgPa-IS assay, or further purified by the MasterPure DNA Purification kit recommended by the manufacturer (Epicentre, Madison, WI) and as previously reported. 14 The purified uninhibited specimens were then used to evaluate the ability of the MgPa-IMW assay to detect M. genitalium in the current study.
Specimen processing, NGU study.
Three different specimen-processing procedures were used before PCR testing of the urine specimens in the NGU study: 1) DNA purification by Instagene Matrix procedure. In this procedure, used in our initial analysis of this sample set, 4 urine specimens were centrifuged the day of collection to pellet bacteria, then pellets were suspended in buffer and frozen at -70°C. After thawing, bacteria were concentrated further by centrifugation and DNA was purified using lysing medium, proteinase K, and Instagene Matrix (Bio-Rad Laboratories, Hercules, CA). Purified specimens were analyzed by the MgPa-S assay as described. 4 2) Roche urine procedure. Urine specimens (50 μL) were treated as described in the AMPLICOR CT/NG specimen preparation kit (Roche, “urine procedure”) as recommended by the manufacturer, which entailed diluting the specimen in Wash Buffer (Roche), heating, then collecting the bacteria by centrifugation. The resulting bacterial pellet was suspended in lysis reagent, mixed, incubated for 15 minutes, then diluted in Specimen Diluent (Roche). This mixture was centrifuged to remove cellular debris, if present, then the resulting supernatant (50 μL) was added to the 100-μL amplification reaction. 3) Epicentre procedure. One hundred fifty microliters of urine was processed with the MasterPure DNA Purification kit according to the manufacturer's instructions (Epicenter). Briefly, this purification included the following steps: proteinase K digestion, detergent lysis, removal of proteins by precipitation, and collection of purified nucleic acid from the resulting supernatant by precipitation. The resulting pellet was suspended in 30 μL distilled water and a 4-μL aliquot was added to the 100-μL PCR reaction.
The amount of preprocessed urine specimen used in the 3 different urine processing methods was 80 μL, 50 μL, and 16 μL for the Instagene Matrix, Roche, and Epicentre procedures, respectively. Bacteria were collected from fresh urine before processing by the Instagene Matrix procedure in our initial analysis 4 and on frozen urine with the Roche and Epicentre procedures in the current study.
Statistical Methods
The sensitivity of the assay described previously was characterized as a curve showing the probability of a positive result as a function of the mean concentration of M. genitalium DNA in the sample. The following functional form to model this curve,
Pr(positive result ‖ μ) = 1 − φexp(-θμ)
where μ is the mean concentration of M. genitalium, θ indexes the sensitivity of the test, and φ is specificity. If φ = 1, then θ could be interpreted as the probability of a positive test when exactly one DNA molecule of M. genitalium is present in the sample. Maximum likelihood methods were used to fit the model to this data. 24 Comparisons of curves were based on hypothesis tests about θ.
Simple concordance was used to measure agreement between assays. Kappa was used to measure the chance-corrected agreement between assays. 25
Results
We modified the MgPa PCR assay to facilitate high-throughput processing and to detect specimens inhibited for PCR using an inhibition control. The use of a 9600 rather than 480 model thermocycler allowed greater numbers of samples to be analyzed at a time, shortened the thermocycling times, and facilitated subsequent processing of specimens, because mineral oil could be excluded from the PCR tubes as a result of the heated lid format of the 900 thermocycler. The method used to detect PCR products was altered from a Southern blot-based assay (designated MgPa-S) to a microwell plate assay (designated MgPa-MW), illustrated in Figure 2, significantly decreasing the time for detection of PCR products from 4 days to 6 hours. The development of the internal control (IC), the determination of the limit of detection of M. genitalium DNA in the MgPa assay, both with (MgPa-IMW) and without (MgPa-MW) the internal control, the selection of positive and negative cutoff values, and the performance of the MgPa-IMW assay in cervical and male urine specimens are described subsequently.
Detection of Internal Control in the MgPa-IMW PCR
The internal control was developed to identify patient samples that were inhibited for amplification and thus might be misclassified as negative. Briefly, a portion of the internal sequence of the insert in pMgPa was replaced with a larger irrelevant sequence, thus maintaining the MgPa primer-binding sites, increasing the size from 258–630 bp (Fig. 1), and providing different target sequences that could be specifically detected by the MgPa and IC probes (Fig. 2B).
Limit of Detection of M. genitalium DNA in the MgPa-IMW
The limit of detection of M. genitalium DNA in the MgPa-MW and MgPa-IMW assays were not significantly different, and neither were the values of the MgPa-IMW assay for 16, 32, and 64 copies of pIC, which were therefore combined for statistical analysis (Fig. 3). The observed rates of detection for a mean of one genome copy per assay were 38% (6 of 16 replicates) and 33% (16 of 48 replicates) for the MgPa-MW and MgPa-IMW assays, respectively, and predicted to be 18% and 16% based on the model depicted in Figure 3. Also based on the model (see “Statistical Methods”), we estimate that, after accounting for the Poisson variability in the number of M. genitalium copies per PCR tube, the probability of detecting exactly 1 copy of M. genitalium (ie, θ) was 0.200 and 0.175 for MgPa-MW and MgPa-IMW, respectively. The 95% detection limit for M. genitalium DNA was 15 and 17 mean copies (Fig. 3) and 13.4 and 15.5 actual copies (data not shown) for the MgPa-MW and MgPa-IMW assays, respectively.
Fig. 3: Limit of detection of the MgPa-MW and MgPa-IMW assays showing the probability of a positive PCR test versus the mean number of genomic M. genitalium copies per assay. (—) = MgPa-IMW assay (with internal control, combining data for 16, 32 and 64 copies of pIC DNA; (….) = MgPa-MW assay (without internal control), Δ = observed proportion of positive tests at each concentration of M. genitalium DNA in the MgPa-IMW assay.
Determination of Positive and Negative Cutoff Values for M . genitalium in the MgPa-IMW Assay Using Cervical Specimens
Negative and positive cutoff values for the MgPa-IMW were first estimated by calculating the mean of the negative OD405 values plus 2 (OD405 = 0.05) and 3 (OD405 = 0.10) standard deviations, respectively. These values were then further refined using a simplified variation of receiver operator characteristic (ROC) analysis 26 to correlate with specimens known to be negative and positive by the MgPa-IS assay. Thus, we plotted the MgPa-IMW OD405 values of cervical samples previously analyzed by PCR and Southern blotting, 14 determined the range of negative values was from 0.00–0.10, and applied a negative cutoff value of 0.10 (Fig. 3). Because the majority of MgPa-S- or MgPa-IS-positive patients yielded OD405 values≥0.25, and very few patients had MgPa-IMW values between 0.10 and 0.25, we set a positive cutoff value of 0.25 and an equivocal zone of >0.10 and <0.25 (Fig. 3). All positive specimens were repeated to confirm that these results were not the result of spurious contamination or specimen mislabeling. Equivocal specimens were repeated in duplicate and were defined as positive with a results of 1 positive (≥0.25) or 2 equivocal reactions (>0.10 and <0.25) in these duplicate repeats.
Using the cutoffs derived from ROC analysis and criteria for positive specimens, all 50 positive and all 50 negative cervical specimens, categorized by the MgPa-IS assay in our previous study, 14 were concordant in the MgPa-IMW assay. Of the 50 positive specimens, 3 resulted in equivocal values (shown in Fig. 4), which repeated as equivocal twice (2 specimens) or positive once (one specimen) and thus were classified as positive. Five of the 100 cervical samples initially tested positive (OD405 1.04, 1.01, 0.35, 0.28, 0.26), but, when repeated, were negative, and thus were classified as negative and were plotted by the negative corrected value in Figure 4. These 5 specimens had tested negative by the MgPa-IS assay and therefore were probably true negatives.
Fig. 4: Distribution of the OD405 values for the MgPa-IMW plate assay performed on patient samples classified as negative or positive by the MgPa-IS assay (cervical specimens) or MgPa-S (urine specimens). Specimens in groups were: (1) MgPa-IS-negative cervical specimens, (2) MgPa-IS-positive specimens, (3 and 5) MgPa-S-negative urine specimens, and (4 and 6) MgPa-S-positive urine specimens. Urine specimens treated by the Roche protocol (groups 3 and 4) and those treated by the Epicentre protocol (groups 5 and 6) were comparable. All samples in the equivocal zone were repeated in duplicate and 6 of 7 resolved as positive, defined as 1 positive or 2 equivocal values on repeat in duplicate. This scattergraph was plotted using Strata version 7 (Strata Corp, College Station, TX).
Determination of Positive and Negative Cutoff Values for the Internal Control in the MgPa-IMW Assay Using Cervical Specimens
We chose a cutoff value of <0.25 for inhibited samples, reasoning that samples with amplification of the IC control to levels ≥0.25 OD405 could also amplify M. genitalium DNA, if present. Using this criteria, all 12 replicates of the IC set at 64 and 16 copies per PCR were positive and the IC was added at 32 copies per PCR in the MgPa-IMW assay. The OD405 values for the internal control were >1.0 in 97 of the 100 cervical samples and >.25 and <1.0 in the remaining 3 specimens (data not shown), confirming that the IC was set at the appropriate level to amplify in these uninhibited specimens, which had been previously treated to remove inhibitors and tested by the MgPa-IS assay. 14
Comparison of Two Different Urine Treatment Methods for the MgPa-IMW PCR
Two different treatment protocols, designated Roche and Epicentre, were 95% (61 of 64) concordant (kappa = 0.90) for the detection of M. genitalium DNA in the urethritis study sample set by the MgPa-IMW assay (Table 3). Thus, 23 specimens were positive and 38 were negative by both methods and 3 were discrepant in the 2 assays. Five specimens treated by the Roche method were initially inhibited for amplification, but after these specimens were frozen at -70°C, thawed, then reanalyzed, all were no longer inhibited.
TABLE 3: Comparison of Urine Specimen Treatment Methods and PCR Assays for the Detection of M. genitalium DNA in Urine Specimens
Comparison of MgPa-S and MgPa-IMW PCR Results for Urine Specimens
The results of the MgPa-IMW assay were compared with those previously obtained on these urine specimens, processed the same day of collection, and analyzed by the Southern blot-based PCR assay (MgPa-S). 4 Of the 32 urine specimens positive for M. genitalium by the MgPa-S assay, 25 (78%) and 24 (75%) were positive by the MgPa-IMW PCR after treatment by the Roche and Epicentre procedures, respectively (Table 3). All 32 MgPa-S-negative samples were also negative by the MgPa-IMW assay irrespective of the urine treatment protocol used. Of the MgPa-S-positive, MgPa-IMW-negative specimens, 6 were negative by both urine treatment methods, 2 were negative after Epicentre but not Roche treatment, and 1 was negative after Roche but not Epicentre treatment. The range of OD405 values for the MgPa-IMW assay performed in this sample set is shown in Figure 4. Of the 4 specimens in the equivocal range, 3 repeated as positive. Thus, the MgPa-S assay was 89% and 88% concordant (kappa = 0.78 and 0.75) with the Mg-IMW assay using Roche and Epicentre treatment protocols, respectively.
Analysis of Urine Specimens Discordant by the MgPa-IMW and MgPa-S Assays
We retested a subset of the frozen urine specimens (n = 19) by the MgPa-S PCR to resolve the discrepancy between the MgPa-IMW (performed on frozen urine after 2-3 y storage) and MgPa-S assays (in which bacteria were collected by centrifugation the day of collection, then frozen until analysis). In this analysis, we found that only 1 of the 8 discrepant specimens (MgPa-S-positive and MgPa-IMW-negative after the Epicentre or Roche treatments, or both), all 5 of 5 positive controls (MgPa-S- and MgPa-IMW-positive) and none of 5 negative controls (MgPa-S- and MgPa-IMW-negative) were positive by the MgPa-S assay after processing by the Epicentre technique (Table 3). Adjusting for the loss of M. genitalium DNA in the frozen specimens (indicated by a negative MgPa-S assay on repeat) and assuming that the remaining concordant-positive and -negative specimens would be consistent with their previous test results, the concordance of the MgPa-IMW and MgPa-S assays was 97% (62 of 64).
Discussion
We have developed a high-throughput M. genitalium-specific PCR assay that is adapted to a 96-well format, incorporates an internal inhibition control, and detects biotinylated PCR products using a colorimetric microwell plate assay. This MgPa-IMW PCR assay, targeting the M. genitalium MgPa adhesin gene, detects a mean of 1 and 17 genome copies of purified M. genitalium DNA with 16% and 95% probability, respectively, close to the detection limit of a single-copy gene, given Poisson distribution of target molecules in the analysis tubes. The MgPa-IMW was 100% and 87-88% concordant with cervical and urine specimens evaluated by the MgPa-IS and MgPa-S assays, respectively.
We used a conservative strategy to define positive specimens in the MgPa-IMW assay, with positive (≥0.25) and negative (≤0.10) OD405 cutoff values and an equivocal range (0.10 > equivocal <0.25) to detect samples that might amplify more robustly by repeat PCR but that we were reluctant to call positive with a single low OD405 value. This strategy might have reduced our sensitivity slightly (one MgPa-S-positive urine sample was equivocal then repeated as negative), but maintained high specificity. In addition, we confirmed all MgPa-IMW-positive results with repeat testing, thus controlling for sporadic contamination or specimen mislabeling. Such contamination, although rare, would greatly affect the results of studies evaluating disease associations of, and risk factors for, M. genitalium, particularly in low prevalence populations. Indeed, 5 cervical specimens that were initially positive in the MgPa-IMW assay were negative both by a repeat MgPa-IMW assay and by the MgPa-IS assay, consistent with their interpretation as negative.
The MgPa-S and MgPa-IMW assays were 88% and 89% concordant for urine specimens treated by the Roche and Epicentre methods, respectively. This lower concordance level is most likely the result of lysis of bacteria (which therefore would not be collected in the urine processing procedures) in additional freeze–thaw cycles and the 2-3 years between analysis by MgPa-S PCR and MgPa-IMW assays. This interpretation is supported by the MgPa-S-negative results in the discordant MgPa-S-positive/MgPa-IMW-negative specimens when retested after frozen storage. However, other reasons for the discrepancy include differences in the amount of patient samples used for PCR following the Instagene (80 μL) and Epicentre (20 μL) purification methods, loss of DNA in urine specimens using storage, or differences in bacterial load between aliquots of the urine specimen. Further studies are needed to determine the effect of long-term frozen storage on urine specimens for detection of M. genitalium by PCR.
The number of inhibited specimens was very low in the NGU and cervicitis sample sets, and thus the benefit of including the inhibition control could not be assessed in these studies. However, because the prevalence of inhibited specimens varies between specimen types, specimen treatment methods, patient characteristics, and geographic locale of the study 27,28 (Totten, unpublished data), we anticipate that the inclusion of the inhibition control will reduce false-negative PCR results resulting from PCR inhibition in other sample sets. Furthermore, the inhibition control will aid in the assessment of optimal specimen treatment protocols and in the determination of the maximum amount of patient specimen that can be added to the PCR assay to maximize sensitivity without introducing inhibition.
The sensitivity and specificity of the MgPa-IMW assay for the detection of “true-positives” inpatient specimens could not be evaluated as a result of the lack of an established “gold standard” for detection. Culture is time-consuming, difficult, and performed in only a few specialized laboratories. However, the results of the MgPa-S assay, which correlated well with the MgPa-IMW assay in the current study, were confirmed by culture in a subset of male urine specimens; 9 of 12 PCR-positive specimens were culture-positive (2 were contaminated and 1 failed to grow) and all 8 PCR-negative specimens were culture-negative. 4 Additionally, in other studies, we confirmed 50 of 51 MgPa-IS-positive cervical specimens either by the rDNA PCR assay 1 or by the sequence of the MgPa amplicon, 14 and confirmed all 13 MgPa-S-positive endometrial specimens by the rDNA PCR. 1,16 Although these studies support a high specificity for the MgPa-IS and MgPa-IMW M. genitalium PCR assays, further studies are needed to evaluate their relative sensitivity and specificity in larger sample sets, with different specimen types, and in comparison to other M. genitalium PCR assays targeting other sites on the the MgPa gene 2–4,18–20,29 and the 16S rDNA gene, 1,21 including a recently developed microwell plate-based M. genitalium-specific rDNA assay. 30 Differences between these PCR assays could include their detection limits for M. genitalium DNA, their detection of possible divergent target sequences between M. genitalium strains, their possible cross-reaction with other organisms present in the patient specimens, and the specimen treatment protocol used.
Advantages of the MgPa-IMW assay compared with the Southern blot-based MgPa PCR assay include the shorter turnaround time (2 rather than 4 days), the numeric positive and negative values giving clear cutoff levels, and the detection of samples inhibited for amplification. This assay performed well after treatment with both the Roche diluent lysis and the Epicentre treatment methods. We recommend performing the MgPa-IMW PCR assay on batches of 82 samples with 6 controls: 2 positive controls of M. genitalium DNA, 2 PCR-negative controls (reagents alone), a positive specimen treatment control (M. genitalium cells), a negative specimen treatment control (reagents alone), a positive development control (detection reagents with M. genitalium and internal control PCR products), in separate wells, and an 8-well strip to blank the plate assay in the microtiter plate reader (detection reagents with no PCR products). The negative controls serve to detect contamination as well as reaffirm the negative cutoff levels for the microwell assay. We also recommend repeating all positive specimens to control for PCR contamination or mislabeling of tubes. With M. genitalium becoming increasingly recognized as a possible etiologic agent in idiopathic cases of urethritis, cervicitis, and endometritis, we hope that the MgPa-IMW PCR assay will enhance studies designed to determine the association of this organism with reproductive tract disease syndromes, risk factors and risk markers for infection, possible sequelae of infection, and optimal treatment regimens.
References
1. Bjornelius E, Lidbrink P, Jensen JS.
Mycoplasma genitalium in non-gonococcal urethritis—a study in Swedish male STD patients. Int J STD AIDS 2000; 11: 292–296.
2. Mena L, Wang X, Mroczkowski TF, Martin DH.
Mycoplasma genitalium infections in asymptomatic men and men with urethritis attending the New Orleans Sexually Transmitted Disease Clinic. Clin Infect Dis 2002; 35: 1167–1173.
3. Morency P, Dubois MJ, Gresenguet G, et al. Aetiology of urethral discharge in Bangui, Central African Republic. Sex Transm Infect 2001; 77: 125–129.
4. Totten PA, Schwartz MA, Sjostrom KE, et al. Association of
Mycoplasma genitalium with nongonococcal urethritis in heterosexual men. J Infect Dis 2001; 183: 269–276.
5. Horner PJ, Gilroy CB, Thomas BJ, Naidoo RO, Taylor-Robinson D. Association of
Mycoplasma genitalium with acute non-gonococcal urethritis. Lancet 1993; 342: 582–585.
6. Janier M, Lassau F, Casin I, et al. Male urethritis with and without discharge: a clinical and microbiological study. Sex Transm Dis 1995; 22: 244–252.
7. Taylor-Robinson D, Horner PJ. The role of
Mycoplasma genitalium in non-gonococcal urethritis. Sex Transm Infect 2001; 77: 229–231.
8. Jensen JS, Orsum R, Dohn B, Uldum S, Worm AM, Lind K.
Mycoplasma genitalium: a cause of male urethritis? Genitourin Med 1993; 69: 265–269.
9. Gambini D, Decleva I, Lupica L, Ghislanzoni M, Cusini M, Alessi E.
Mycoplasma genitalium in males with nongonococcal urethritis. Sex Transm Dis 2000; 27: 226–229.
10. Johannisson G, Enstrom Y, Lowhagen GB, et al. Occurrence and treatment of
Mycoplasma genitalium in patients visiting STD clinics in Sweden. Int J STD AIDS 2000; 11: 324–326.
11. Maeda S, Tamaki M, Nakano M, Uno M, Deguchi T, Kawada Y. Detection of
Mycoplasma genitalium in patients with urethritis. J Urol 1998; 159: 405–407.
12. Yoshida T, Maeda S, Deguchi T, Ishiko H. Phylogeny-based rapid identification of mycoplasmas and ureaplasmas from urethritis patients. J Clin Microbiol 2002; 40: 105–110.
13. Pépin J, Sobela F, Deslandes S, et al. Etiology of urethral discharge in West Africa: the role of
Mycoplasma genitalium and
Trichomonas vaginalis. Bull World Health Organ 2001; 79: 118–126.
14. Manhart LE, Critchlow CW, Holmes KK, et al. Mucopurulent cervicitis and
Mycoplasma genitalium. J Infect Dis 2003; 187: 650–657.
15. Uno M, Deguchi T, Komeda H, et al.
Mycoplasma genitalium in the cervices of Japanese women. Sex Transm Dis 1997; 24: 284–286.
16. Cohen CR, Manhart LE, Bukusi EA, et al. Association between
Mycoplasma genitalium and acute endometritis. Lancet 2002; 359: 765–766.
17. Tully JG, Taylor-Robinson D, Cole RM, Rose DL. A newly discovered mycoplasma in the human urogenital tract. Lancet 1981; 1: 1288–1291.
18. Palmer HM, Gilroy CB, Furr PM, Taylor-Robinson D. Development and evaluation of the polymerase chain reaction to detect
Mycoplasma genitalium. FEMS Microbiol Lett 1991; 61: 199–203.
19. Jensen JS, Uldum SA, Sondergard-Andersen J, Vuust J, Lind K. Polymerase chain reaction for detection of
Mycoplasma genitalium in clinical samples. J Clin Microbiol 1991; 29: 46–50.
20. Deguchi T, Gilroy CB, Taylor-Robinson D. Comparison of two PCR-based assays for detecting
Mycoplasma genitalium in clinical specimens. Eur J Clin Microbiol Infect Dis 1995; 14: 629–631.
21. Yoshida T, Deguchi T, Ito M, Maeda S, Tamaki M, Ishiko H. Quantitative detection of
Mycoplasma genitalium from first-pass urine of men with urethritis and asymptomatic men by real-time PCR. J Clin Microbiol 2002; 40: 1451–1455.
22. Casin I, Vexiau-Robert D, De La Salmoniere P, Eche A, Grandry B, Janier M. High prevalence of
Mycoplasma genitalium in the lower genitourinary tract of women attending a sexually transmitted disease clinic in Paris, France. Sex Transm Dis 2002; 29: 353–359.
23. Russell W, Newman C, Williamson D. A simple cytochemical method for demonstration of DNA in cells infected with mycoplasmas and viruses. Nature 1975; 253: 461–462.
24. Hughes JP, Totten PA. Estimating the accuracy of polymerase chain reaction-based tests using endpoint dilution. Biometrics 2003; 59: 505–511.
25. Fisher L, vanBelle G. Biostatistics. A Methodology for the Health Sciences. New York: Wiley and Sons, 1993.
26. Metz CE. Basic principles of ROC analysis. Semin Nucl Med 1978; 8: 283–298.
27. Mahony J, Chong S, Jang D, et al. Urine specimens from pregnant and nonpregnant women inhibitory to amplification of
Chlamydia trachomatis nucleic acid by PCR, ligase chain reaction, and transcription-mediated amplification: identification of urinary substances associated with inhibition and removal of inhibitory activity. J Clin Microbiol 1998; 36: 3122–3126.
28. Toye B, Woods W, Bobrowska M, Ramotar K. Inhibition of PCR in genital and urine specimens submitted for
Chlamydia trachomatis testing. J Clin Microbiol 1998; 36: 2356–2358.
29. Cadieux N, Lebel P, Brousseau R. Use of a triplex polymerase chain reaction for the detection and differentiation of
Mycoplasma pneumoniae and
Mycoplasma genitalium in the presence of human DNA. J Gen Microbiol 1993; 139: 2431–2437.
30. Jensen JS, Borre MB, Dohn B. Detection of
Mycoplasma genitalium by PCR amplification of the 16S rRNA gene. J Clin Microbiol 2003; 41: 261–266.
