Lymphogranuloma venereum (LGV) is caused by the more invasive L serovars (L1, L2, L2a, and L3) of Chlamydia trachomatis and is considered to be a rare disease outside resource-poor countries. Since the end of 2003, an ongoing outbreak of LGV proctitis has been reported in Europe and north America among men who have sex with men, which has been strongly associated with HIV infection. A total of 492 confirmed cases of rectal LGV infection were reported to April 2007 in the United Kingdom alone.1 Several molecular diagnostic methods specific for LGV detection have been reported since the occurrence of this outbreak.2^–5 These diagnostic tests are either standard PCR or real-time PCR based with a much shorter turnaround time than culture or PCR restriction fragment length polymorphism (RFLP) genotyping methods. In this study, we have modified our previously reported real-time duplex PCR assay into a quadriplex format that can detect an LGV-specific, a non-LGV-specific target sequence, the C trachomatis plasmid and the human RNase P gene. The performance of the real-time duplex and quadriplex assays were compared and evaluated using rectal specimens.

METHODS

Clinical specimens

Rectal swab specimens obtained from patients presenting with symptoms suggestive of LGV/non-LGV proctitis were referred to the Sexually Transmitted Bacterial Reference Laboratory, Health Protection Agency (HPA) of the United Kingdom during 2005–2006 for LGV testing. DNA extraction methods performed by the HPA have been described previously.3 A combination of DNA extracts (N  =  31) and swab fluids (N  =  58) stored at −20°C, without patient identifiers, were sent to the Centers for Disease Control and Prevention (CDC) for real-time multiplex PCR testing. The QIAamp DNA mini kit (Qiagen, Valencia, California, USA) was used for DNA extraction from the swab fluids performed at the CDC. This study protocol was approved by the Institutional Review Board at the CDC.

Real-time PCR testing and genotyping at the HPA

The chlamydial status of rectal specimens was confirmed at the HPA using an inhouse PCR assay as described previously.2 If specimens tested positive for C trachomatis DNA, an LGV-specific real-time simplex PCR assay reported by Morré et al3 was performed. For specimens testing negative by this LGV-specific PCR, a commercial artus C trachomatis Plus PCR Kit (Qiagen Hamburg, Germany) was used to detect the omp1 and the cryptic plasmid. All real-time PCR reactions were run on a Rotor-Gene 3000 instrument (Corbett Research, Sydney, Australia). Specimens that tested positive for LGV were genotyped by PCR–RFLP as described previously.2 The LGV-specific PCR and genotyping results from the HPA were not revealed to the CDC until the completion of real-time multiplex PCR testing.

Real-time multiplex PCR assays

The real-time duplex PCR assay, based on a unique 36-base pair (bp) deletion region of the polymorphic membrane protein H gene (pmpH) that only occurs among LGV strains, has been reported previously.2 Briefly, a set of PCR primers amplifies a 168-bp DNA fragment that encompasses the deletion region and two TaqMan probes were used: one detects only chlamydial serovars A–K and the second probe detects all chlamydial serovars (A–L3). The real-time quadriplex PCR assay was developed to include an LGV-specific, a non-LGV-specific target sequence, a C trachomatis plasmid target and the human RNase P gene as an internal control. The PCR primer set was designed to amplify a 130-bp DNA fragment encompassing the deletion region of pmpH described above. The LGV-specific TaqMan MGB probe described by Morré et al3 was incorporated in the assay. The non-LGV-specific TaqMan probe was the same as in our duplex assay but was designed to hybridise to the complementary DNA strand. The quadriplex assay also contained PCR primer and probe sets targeting an 87-bp region of the C trachomatis cryptic plasmid and a 73-bp region of the human RNase P, respectively. The DNA sequence, fluorescent tag of the probe and the optimised concentration of oligonucleotide used for the individual PCR target are described in table 1. All real-time multiplex PCR assays were performed on a Rotor-Gene 3000 real-time PCR instrument with 25 μl reaction volume and 10 μl DNA sample using the thermal cycling conditions described previously.2 A final concentration of 1 × PCR buffer (Applied Biosystems, Foster City, California, USA), 4 mmol magnesium chloride, 200 μmol each of dATP, dGTP, dCTP and dUTP, 0.5 units of uracil-N-glycosylase (Applied Biosystems) and one unit of AmpliTaq Gold DNA polymerase (Applied Biosystems) were used in the quadriplex assay. Purified genomic DNA from serovars D and L2 strain of C trachomatis were used as positive controls. The analytical sensitivity of the quadriplex PCR assay was determined using a series of 10-fold dilutions of purified C trachomatis genomic DNA from serovars D and L2. In addition, the limit of detection was also examined in spiked samples containing a high ratio of LGV to non-LGV DNA (∼10⁵ : 1 ) and vice versa (∼10⁶ : 1). The specificity of the assay was tested with a collection of C trachomatis serovars A–L3 and a panel of pathogenic and commensal genital tract microorganisms. Purified L2 genomic DNA was obtained from Advanced Biotechnologies, Inc, Columbia, Maryland, USA. The degree of agreement between the two real-time multiplex PCR assays was measured using the kappa test.

RESULTS

Sensitivity and specificity of the real-time quadriplex PCR assay

The limit of detection for the quadriplex PCR assay for either LGV or non-LGV was 10–100 genomic copies/reaction with a similar PCR efficiency (using the pmpH target, data not shown). In samples spiked with both LGV and non-LGV DNA, the assay had a limit of detection of 10–100 copies/reaction for non-LGV in the presence of approximately 9 × 10⁵ copies of LGV. The sensitivity of detecting LGV was 10–100 copies/reaction in the presence of approximately 4 × 10⁶ copies of non-LGV. The C trachomatis plasmid target was detected in spiked samples containing either 1–10 copies of LGV or non-LGV. The quadriplex assay correctly identified C trachomatis serovars A–K as non-LGV and L1–L3 as LGV. No crossreactivity was observed when tested against a panel of pathogenic and commensal genital tract microorganisms (data not shown).

Comparison of real-time duplex PCR versus quadriplex PCR assay

Good agreement (85 of 89 specimens, 95.5%) was found between the two real-time multiplex PCR assays for the detection of C trachomatis DNA (kappa value 0.93, 95% CI 0.86 to 0.99). Both assays identified 34 LGV, 35 non-LGV C trachomatis-positive and 16 negative specimens (table 2). Of two specimens that tested non-LGV by the duplex PCR, one was found to be a mixed infection (positive for pmpH, plasmid and RNase P targets), whereas the other was positive only for plasmid and RNase P targets but not the pmpH target by the quadriplex PCR. Two additional specimens that gave equivocal results (same specimen tested twice with differing results) for non-LGV by the duplex PCR and reported as being positive for non-LGV C trachomatis also tested positive only for plasmid target and human DNA control by the quadriplex PCR. In addition, six specimens that tested negative by the duplex assay were found to be invalid as a result of the failure to detect RNase P DNA by the quadriplex PCR. After spiking approximately 10³ genomic copies of C trachomatis DNA into these six invalid specimens, three were found to contain PCR inhibitors by the quadriplex PCR (data not shown).

Comparison of LGV-specific real-time simplex PCR and genotyping versus real-time multiplex PCR

Thirty-three of 34 LGV-positive specimens detected by both duplex and quadriplex PCR assays also tested positive by the LGV-specific simplex PCR performed at the HPA. Genotyping using PCR–RFLP analysis confirmed that all LGV-positives were serovar L2. No discrepancies were found among the 16 specimens that tested negative. The single specimen identified by the quadriplex PCR as a mixed LGV/non-LGV infection was reported as LGV positive by the HPA. One specimen that tested non-LGV positive by the duplex PCR but tested positive only for plasmid and RNase P targets by the quadriplex PCR was also reported as non-LGV by the HPA. Two additional specimens, which yielded equivocal results for non-LGV by the duplex PCR and tested positive only for the plasmid and human DNA targets by the quadriplex PCR, were also reported as equivocal for non-LGV by the HPA either with late positive signals (threshold cycles >40) in only one of their real-time PCR assays, or with discordant results from two different real-time PCR assays.

DISCUSSION

In the past, the laboratory diagnosis of LGV infection has been based either on culture or nucleic acid amplification testing followed by genotyping of positives with PCR-based RFLP analysis and/or sequencing of omp16^–8 and serological testing.9 The utility of serological methods is limited either as a result of the lack of species specificity or the broad crossreactivity of antibody responses especially in LGV. A recent study by Van der Snoek et al,10 however, suggested that high titres of species-specific IgA antibody and the age of the infected individual appeared to be of possible diagnostic value for (early) detection of LGV proctitis. Currently, there are no US Food and Drug Administration-cleared commercially available nucleic acid amplification tests for C trachomatis detection using rectal specimens in the United States. In addition, these commercial diagnostic tests do not differentiate LGV from non-LGV biovars. Genotyping or serotyping methods for C trachomatis allow specific discrimination between biovars but are time consuming, require specially trained personnel and a sophisticated laboratory setting.

PCR-based diagnostic tests are preferred over culture or genotyping methods because of their higher sensitivity and rapid turnaround time. A number of conventional PCR assays have been described for the detection of C trachomatis DNA,4 7 9 including a duplex PCR that targets both the plasmid and omp1.11 Relatively few have, however, been developed for the differential diagnosis of LGV.5 12 Morré et al3 recently described a real-time simplex PCR assay that is specific for the detection of LGV but the assay still needs a secondary test to identify non-LGV C trachomatis infection and mixed infections. Subsequently, Halse et al4 described a real-time multiplex PCR that is capable of detecting all serovars of C trachomatis by targeting the plasmid and simultaneously identifying the serovar L2 specifically. That assay includes a spiked internal plasmid control to monitor PCR inhibition and provides rapid screening for serovar L2; however, L1 and L3 serovars, and a mixed infection with a non-LGV strain would remain undetected if further molecular testing were not performed. Our previously described duplex PCR assay is not capable of differentiating a co-infection caused by LGV and non-LGV strains. In addition, both simplex and duplex PCR assays do not include an internal control to monitor PCR inhibition and specimen adequacy. The strength of our quadriplex PCR assay lies in its ability specifically to detect individual LGV or non-LGV infection as well as mixed infections. Furthermore, the assay was able to identify specimens that either contained PCR inhibitors or were not adequately collected for PCR. This assay also requires additional molecular testing in order to differentiate among LGV serovars.

The reported analytical sensitivity of the real-time simplex and multiplex PCR assays for LGV detection, depending on the PCR target, ranges from 0.01 inclusion-forming units to 50 genomic copies; and approximately one to 25 genomic copies for non-LGV C trachomatis using the plasmid target.2^–4 In this study, the limit of detection of our quadriplex PCR assay for either LGV or non-LGV was approximately 10–100 genomic copies/reaction using the pmpH target. It is not surprising that the analytical sensitivity of a multiplexed assay is slightly reduced when compared with a simplex PCR. The potential causes include, but are not limited to, competition from the different targets and primers, the preferential amplification of one target sequence over another.13^–15

Both real-time multiplex PCR assays performed at the CDC had 100% concordance for the specific detection of rectal LGV and also had a high agreement (33 of 34, 97.1%) with the results obtained from the LGV-specific simplex PCR assay and PCR–RFLP genotyping performed by the HPA. These results suggest that the optimised quadriplex PCR conditions used had minimal probability of providing false negatives or positives for LGV diagnosis. Of the four specimens that gave discordant results between the duplex and the quadriplex PCR assay, one was caused by co-infection of LGV and non-LGV strains that inevitably would be identified as non-LGV by the duplex assay as a result of the inherent limitation of the assay design. On the other hand, the LGV-specific simplex PCR by the HPA identified this specimen as LGV and further supports the likelihood of a mixed infection. Of the three rectal specimens identified as non-LGV, two gave split results by the duplex assay and all tested positive only for plasmid target and human DNA by the quadriplex PCR; the most likely explanation for these discrepancies could be the presence of low copy numbers of targets. The only specimen that tested positive for non-LGV by the duplex PCR but tested positive only for plasmid target and human DNA by the quadriplex PCR was also reported as non-LGV by the HPA. It is not clear whether low copy numbers of targets play a role or some unknown factors within this particular specimen competed or interfered with target detection in the quadriplex assay. In the case of inconclusive test results arising from the quadriplex assay, we suggest that individual simplex LGV-specific and non-LGV-specific PCR should be repeated for confirmation.

The recently reported C trachomatis strains with a 377-bp deletion in the cryptic plasmid in some regions of Sweden16 has no impact on our assay performance because the deletion area is located within the open reading frame 1 of cryptic plasmid, whereas our real-time PCR primers target open reading frame 5. Urogenital isolates of C trachomatis that lack the cryptic plasmid have also been documented;17^–19 however, in our study we did not observe any specimens that tested positive for either LGV or non-LGV but negative for the plasmid target. The frequency of mixed chlamydial infections found in rectal specimens is not well documented. Morré et al3 recently reported two rectal specimens that were initially typed as non-LGV serovars E and D by PCR–RFLP and sequencing but subsequently tested positive for LGV by the simplex PCR. Therefore, in that study population, a co-infection rate of LGV and non-LGV infection in the rectum was estimated to be 6.7% (2/30). In the current study, we identified only one rectal specimen (1/89, 1.1%) with a mixed LGV and non-LGV infection. The capability of the quadriplex assay to identify mixed infections makes it a useful tool to study the prevalence of co-infection by LGV and non-LGV C trachomatis strains. Alternatively, sequential rounds of simplex PCR testing can be performed to identify co-infections and to reduce further the probability of failing to detect a target that had very low copy numbers in a mixed infection.

In the past few years, rectal LGV has emerged as a significant problem among men who have sex with men in Europe and north America. A rapid and accurate diagnosis of the disease is crucial for effective patient management because the duration of treatment required for LGV is longer than that for non-LGV chlamydial infections. Real-time multiplex PCR assays have a potential application at point-of-care facilities and can be cost-effective as a result of the capability of detecting up to six target organisms in one assay (eg, Rotor-Gene 6000). The real-time multiplex PCR not only enables the early and easy detection of chlamydial infections, but can also be an efficient screening tool for asymptomatic or mildly symptomatic cases during routine testing for chlamydial infection as well as in an outbreak situation.