Objectives Until recently, PCR had been used to detect but not quantify Treponema pallidum. To understand infection kinetics of this uncultivable organism, a real-time PCR assay was developed to quantify 47 kDa membrane lipoprotein gene DNA (tpp47).
Methods Assay specificity was determined against DNA from humans, skin organisms and sexually transmitted pathogens. tpp47 DNA (Nichols strain) was used to construct a standard curve for T pallidum quantification. Blood and ulcer samples were obtained from 99 patients being investigated or screened for syphilis and tpp47 was quantified.
Results The assay was specific, not cross-reactive with other organisms tested and sensitive, with a detection limit of a single copy of tpp47 DNA. For ulcer samples, the assay was 100% sensitive and 97.14% specific. Sensitivity fell to 34.1% for blood samples but specificity remained high (100%). tpp47 DNA was more commonly detected, and at a higher copy number, in blood of patients with secondary infection (sensitivity 57.89%) compared with primary infection. Quantity of tpp47 DNA was higher in primary infection ulcers, especially in HIV-1-positive patients, than in ulcers persisting into secondary disease.
Conclusions Quantifying T pallidum provides insight into syphilis infection kinetics: Ulcers of primary disease in HIV-1-positive patients are perhaps more infectious and the presence and load of T pallidum bacteraemia is variable, with a peak in the secondary stage. Quantitative PCR has the potential to map T pallidum infection and to highlight the impact of HIV on syphilis.
- Treponema pallidum
This is an open-access article distributed under the terms of the Creative Commons Attribution Non-commercial License, which permits use, distribution, and reproduction in any medium, provided the original work is properly cited, the use is non commercial and is otherwise in compliance with the license. See: http://creativecommons.org/licenses/by-nc/2.0/ and http://creativecommons.org/licenses/by-nc/2.0/legalcode.
Statistics from Altmetric.com
A better understanding of syphilis may be obtained by quantitatively mapping the course of infection. However, the causative organism, Treponema pallidum subspecies pallidum (T pallidum), cannot be easily cultured, so detection relies on direct visualisation, molecular detection and, rarely, rabbit infectivity testing. None of these methods is ideal for studying T pallidum pathogenesis or infection kinetics. Dark ground microscopy (DGM) to count organisms in ulcer exudate or cerebrospinal fluid is highly operator dependent and cannot be used for blood.1 Rabbit infectivity testing, often described as the gold standard for diagnosing syphilis,2 is highly sensitive but impractical. The non-specific anti-cardiolipin-based tests, for example, rapid plasma reagin (RPR), can be used as a surrogate marker of T pallidum burden to stage infection and monitor treatment response but are prone to false-positive results and do not reflect the kinetics of the infection since they take months to resolve and may never become negative. Furthermore, HIV-1 infection, which can alter the course of T pallidum infection and response to treatment, is also known to have an impact on treponemal serology.3 4
Molecular detection of T pallidum DNA by PCR is now established for the diagnosis of syphilitic ulcers.5 6 These assays have been based on the detection of various T pallidum genes, including polA,7–9 flaA,10 tmpA11 and tpp47,12–14 which are highly conserved across T pallidum subspecies. Until recently, molecular studies were qualitative and little was known about the quantity of T pallidum in clinical samples during the course of syphilis.
To study T pallidum infection kinetics in humans, a real-time PCR assay has been developed to detect and quantify T pallidum tpp47 DNA in samples from patients with syphilis.
Population and specimens
A prospective, cross-sectional study was conducted at St Mary's Hospital, London, between July 2006 and January 2008. The study had research ethics approval and samples were collected after obtaining written consent. All patients being investigated for syphilis (including routine screens) were eligible. Investigation and treatment for sexually transmitted infections (STIs) were performed in accordance with national guidelines.15
Six millilitres of blood were collected into EDTA for T pallidum DNA detection and quantification from all participants. Samples collected between July and December 2006 were processed immediately and DNA was extracted as described below. Samples collected after this period were stored at −80°C until DNA was extracted. Ulcers were cleaned and abraded with sterile water and gauze prior to sampling. After 1–2 min, the accumulated exudate was absorbed onto Sno Strips filter paper (Laboratoire Chauvin, Aubenas, France) held at the ulcer edge until dry. DGM was performed on genital ulcers according to clinic policy prior to the collection of study samples.
The Newmarket IgM/IgG enzyme immunoassay (EIA) (Alere, Stockport, UK) was used to screen for T pallidum infection. Fujirebo T pallidum particle agglutination (TPPA) (Mast Diagnostics, Bootle, UK) and Bio-kit (RPR (Launch Diagnostics, Longfield, UK) were used to confirm and quantify the serological response, respectively.
Baseline and convalescent syphilis serology were documented in addition to STI results. Patients' final clinico-microbiological diagnoses were recorded as one of the following: herpes simplex virus (HSV), primary syphilis, secondary syphilis, other syphilis (comprising latent and symptomatic late disease), neurosyphilis and asymptomatic contacts of syphilis and non-specific genital ulceration. Case definitions are given in table 1.
Demographic data, sexual orientation and HIV-1 status, if known, were also recorded. Ninety-nine patients were recruited of whom 44 had a final clinical diagnosis of syphilis: 14 had primary syphilis, 19 secondary syphilis, 8 early latent syphilis and 3 neurosyphilis (all during the secondary stage). Of the 55 non-syphilis cases, 17 (31%) were contacts of patients with syphilis and 38 had non-syphilitic ulcers of which half were confirmed as genital herpes. The remaining 19 (35%) patients had no STI and were diagnosed with non-specific genital ulceration, which is consistent with studies examining the cause of genital ulceration.17 18
The study population is representative of the current UK syphilis epidemic.19 Ninety-seven (98%) of the patients were men, 88 (88.7%) were Caucasian with a median age of 36 years and just over half (52, 52.5%) were HIV positive. Neither of the two women were HIV positive (full data shown in online supplementary table 1).
Sample handling and DNA extraction
Aliquots of 400 μl of EDTA whole blood (WB) were either immediately treated with 60 μl of proteinase K (Qiagen, Crawley, UK) or frozen at −80°C for future processing. DNA was extracted using the Qiagen QIAamp DNA Mini kit (blood spin protocol) with the following modifications: 400 μl of sample and buffer AL were used; the first incubation was for 3 h or overnight; final elution was into 120 μl of buffer AE following 5 min incubation. Extraction was performed in a separate laboratory to the PCR and new aliquots of reagent were used for each batch.
Whole Sno Strips were placed in microcentrifuge tubes with 400 μl phosphate buffered saline and 20 μl proteinase K. DNA was extracted using the Qiagen QIAamp DNA Mini kit, buccal swab spin protocol with the above modifications. Quantitative PCR (qPCR) results for ulcer samples are presented as copies/strip.
A 178 bp segment of the 1377 bp T pallidum tpp47 gene (GenBank accession number M88769.1) was amplified using the following primers: forward 5′CGAGGAATACAAGATTACGAACG3′, reverse 5′ACGTGCAGAAAAACTATCCTCAG3′ (nucleotides 525–547 and 702–680, respectively). Primer sequences, product size and reaction conditions were optimised for sensitivity and reproducibility (data not shown). Each reaction contained 10 μl extracted DNA, 0.6 units LightCycler Uracil-DNA glycosylase (UDG) (Roche, Mannheim, Germany), 0.5 μm each of forward and reverse primers (Invitrogen, Paisley, UK), 4 μl FastStart Plus reaction mixture (Roche, Mannheim, Germany) (containing FastStart Taq DNA polymerase, reaction buffer, MgCl2, SYBR Green I dye and dNTP mix) and PCR-grade water to a final 20 μl volume.
Real-time PCR was performed using a Roche LightCycler 2, software version 4.0, with the following conditions: UDG activation at 40°C for 10 min, preincubation at 95°C for 10 min and then 46 cycles of 95°C for 10 s, 60°C for 5 s, 72°C for 8 s and 81°C for 10 s (acquisition). Two no-template controls were included in each experiment, and standard precautions to avoid contamination were employed.
Samples were run in duplicate and discordant results were repeated. Negative samples were diluted 1:10 and repeated if inhibition was suspected.
During assay development, PCR product identity was confirmed using both agarose gel electrophoresis (2% agarose gel containing Tris-acetate-EDTA buffer, 0.6 mg ethidium bromide) and DNA sequencing using the Applied Biosystems dRhodamine kit and 3100 capillary sequencer. Using NCBI BLAST software, the 178 bp amplicon showed 100% sequence homology in comparison with published T pallidum 47 kDa gene sequences.
Melting curve analysis was performed on all amplicons from clinical samples and compared with an in-run standard in order to confirm amplification of the correct target. Fluorescence emission in the 530 nm channel at a linear temperature transition rate of 0.1°/s from 60°C to 95°C was monitored with continuous acquisition. The melting temperature (Tm) of the correct 178 bp amplicon was 82.5°C.
In order to confirm non-cross-reactivity with commensal treponemes, the infective agents of other STIs, common skin organisms or human DNA, the assay was tested against DNA from these sources (see online supplementary table 2).
Preparation of the standard curve and limit of assay detection
T pallidum (Nichols strain) bacteria were enumerated using DGM to give an initial concentration of 6×108 organisms/ml (suspended in glycerol and phosphate buffer). DNA was extracted from 10 μl aliquots using Qiagen QIAamp DNA Mini kit as per the manufacturer's protocol and eluted into 60 μl of molecular-grade water to produce a concentration of 106 T pallidum genome equivalents/10 μl. Serial fivefold dilutions were made using molecular-grade water from 1/5 to 1/108. DNA detection was unreliable beyond 1/106 dilution (CT 38.04), thus 1 copy/10 μl was determined to be the lowest limit of reliable detection. An example of an in-run standard curve used for T pallidum DNA quantification in clinical samples is shown in the online supplementary data.
In order to be sure that the standard curve was biologically relevant, detection of T pallidum DNA in WB and water were compared. Serial 10-fold dilutions of T pallidum DNA were added to syphilis-seronegative WB, quantified and found to be similar to dilutions in molecular-grade water described above (see online supplementary table 3).
Bacteria and DNA
T pallidum (Nichols) organisms were obtained through passage in rabbit testes and donated by Nigel Appleton (Newmarket Laboratories Ltd, Suffolk, UK). DNA from T pallidum subspecies pertenue and Treponema denticola was donated by Sheila Lukehart (University of Washington, Seattle, USA). All other samples were donated by Stuart Philip (St Mary's Hospital, London, UK). Human DNA samples were obtained from blood donated for research by patients seronegative for T pallidum.
Statistical analysis was performed using PASW statistics, V.18, SPSS, Chicago, IL. Significance levels were calculated using the Mann–Whitney U test (two sided, 95% CIs).
During the validation, as previously described in a rabbit model,10 extraction and quantification of tpp47 DNA from WB gave more consistent results than from peripheral blood mononuclear cells or plasma (data not shown). Thus, only ulcer and WB results are presented.
Table 2 details all patients included in the study and gives serological, DGM and PCR results.
Amplification was observed neither with DNA extracted from the panel of STI and other organisms nor with human DNA. T pallidum subspecies pertenue was detected by the assay. This result was expected, as there are very few genetic differences between T pallidum subspecies.
Sensitivity and specificity of tpp47 detection in ulcer and WB samples
Patients' final clinico-microbiological diagnoses (as detailed in table 1) were compared with T pallidum qPCR results.
The sensitivity and specificity for T pallidum detection in syphilitic ulcers, primary or secondary, were 100% and 97·14%, respectively. The details of a potential false-positive result are given below. With the exception of secondary syphilis (57.89%), sensitivity for T pallidum in WB samples was low (34.09% for all stages overall) while specificity was 100%. The sensitivity and specificity of WB PCR compared with serology for the diagnosis of primary disease were 27.27% and 100%, respectively.
All 14 patients with primary syphilis had PCR-positive ulcers and 10 were DGM positive. All four patients with DGM-negative ulcers had positive serology at enrolment. Compared with the final diagnosis, the sensitivity and specificity of DGM for syphilis ulcers (primary and secondary combined) were 70.59% and 87.88%, respectively, with no significant difference observed for primary or secondary cases analysed individually. The sensitivity of PCR compared with DGM for the diagnosis of primary disease was 100%.
tpp47 DNA was detected in ulcer exudates of all 14 patients diagnosed with primary syphilis. Median copy number of tpp47 DNA from ulcers was 1832 copies/strip (range 251–14 244), which did not differ significantly between those processed immediately ((n=9) median 2577 copies/strip (427–14 244)) and those processed after storage ((n=5) median 1064 copies/strip (251–2115) (p=0.162)). The median number of ulcer tpp47 copies/strip in HIV-1-positive and HIV-1-negative patients were 2115 (757–14 244) and 648 (251–2742) (p=0.048), respectively. Mean CD4 count of the seven HIV-1-infected patients was 501 (range 270–920) and CD4 count was not associated with T pallidum load.
Three patients had tpp47 DNA detectable in WB with a copy number 1–2 logs lower (median 127 copies/ml, 127–168) than for ulcers. It did not differ according to HIV status and was associated with a modal RPR of 4. The mean duration of symptoms for these patients was 8 days (range 7–10 days), compared with 30.5 days (1–90) for those who were tpp47 DNA PCR negative in WB.
tpp47 DNA was detected in the blood of 11/19 (58%) patients with secondary syphilis. This rate of detection was higher (86%) in samples that had been processed immediately compared with 42% when DNA extraction was performed after storage (p=0.02) (see online supplementary figures). The mean tpp47 DNA copy number in WB samples extracted immediately was 516 copies/ml and the modal RPR for these patients was 128.
T pallidum was detected in WB of 8/15 (53%) HIV-1-infected versus 3/4 (75%) uninfected (28%) patients. Mean CD4 count of HIV-1-infected patients with tpp47-positive WB was 448, not significantly different from those tpp47 negative in WB (492, p=0.594).
Three patients had muco-cutaneous oral lesions, of which two were positive for tpp47 DNA. Four patients with secondary syphilis still had anogenital ulcers of which three were sampled and all were tpp47 DNA positive. When compared, the mean tpp47 DNA copy number was lower in these anogenital ulcers persistent in secondary disease (351 copies/strip) than in primary chancres (3498 copies/strip) (p=0.032).
tpp47 DNA was detected in WB of one of the eight patients with latent disease. This HIV-1-positive untreated patient, with an initial RPR titre of 1:131 072, had been lost to follow-up for 2 years prior to the study. RPR at the time of recruitment was 1:256.
One ulcer sample from a patient with a 1-week history of multiple non-tender genital ulcers thought to be atypical genital herpes and treated with acyclovir had detectable tpp47 DNA. DGM was negative as were the contemporaneous and convalescent syphilis serology and PCR for herpes simplex virus types one and two. This may represent a false-positive PCR result, but it is not known whether this patient received coincidental treponemocidal antibiotics that could have prevented seroconversion.
Conclusions and discussion
For 20 years, PCR has been used with a variety of clinical and rabbit specimens to detect, type and quantify T pallidum.5 6 10 20 21 PCR reliably detects T pallidum in ulcer specimens and is now established, often as part of a multiplex assay, for the diagnosis of primary syphilis.5 The more modest success of T pallidum DNA PCR in other clinical samples has been attributed to their lower treponemal load.9
There are few data on the quantification of T pallidum by PCR in clinical samples, thus the timing of bacteraemia and duration of post-treatment as well as the impact of HIV-1 infection on the course of T pallidum infection are uncertain. Although cross-sectional, the data presented here provide insight into the range of T pallidum DNA copies that can be detected in samples taken at different stages of disease.
Since several groups had found the tpp47 gene to be reliable and specific for T pallidum detection,5 9 12 13 new primers were designed for a real-time PCR assay. These have been shown to be equally specific for T pallidum detection with no cross-reactivity to T denticola, any of the common commensal skin flora or sexually transmitted organisms tested. Sensitivity was maintained when used to detect T pallidum DNA extracted from human WB samples with no inhibition by the high concentration of human DNA.
Storage at −80°C prior to DNA extraction impaired the detection and quantification of T pallidum in WB but not in ulcer samples. Cruz et al reported a similar effect, with detection of T pallidum DNA in WB of patients with secondary syphilis just 30% compared with 63% when the samples were processed within a few hours of collection.22 The maximum interval between blood sampling and DNA extraction, which is likely to be short, remains to be established. Additional study is also needed to optimise T pallidum collection from ulcer samples. While Sno Strips are shown here to be efficient at absorbing exudate and bacterial DNA, the volume (which is in part dependent on the ulcer area) of ulcer exudate will influence final copy number in the same way as the concentration of T pallidum in the ulcer fluid. Quantifying a human reference gene from ulcers (eg, β-globin) and comparing it with tpp47 load may also help standardisation. It remains to be shown, however, that the amount of human DNA detected would not simply mirror that of T pallidum.
T pallidum DNA was detected in the ulcers of all 14 patients diagnosed with primary syphilis, a similar finding to the work of Palmer et al in 200212 and unsurprising, given that the chancre is the site of bacterial inoculation. Moreover, tpp47 DNA PCR from ulcer exudates appears to more reliably include or exclude T pallidum infection than DGM. T pallidum DNA was also detected in ulcer samples in secondary syphilis with three out of four genital and two out of three oral lesions being positive. Fewer organisms/strip were detected in these lesions than in primary chancres (351 vs 3498), which likely represents ulcer healing and resolution.
T pallidum bacteraemia is thought to occur from the outset of the primary infection and to persist at all stages of disease thereafter,23 and while these stages are well defined, the rate of bacterial dissemination in humans is unknown. In rabbits inoculated intratesticularly, T pallidum flaA DNA could be detected at low load in blood after 24 h and at much higher load after 7–10 days. The increased detection after 1 week corresponded with rising RPR titres and widespread treponemal dissemination into organs.10 The detection rates observed in WB for patients with primary (3/14, 21.4%) and secondary (11/19, 58%) syphilis are lower than the detection rates for ulcers, but are comparable with those reported previously.12 22 A higher detection rate in WB in secondary syphilis corresponds to a greater number of organisms and higher RPRs, similar to the picture reported in the rabbits 7–10 days post-inoculation. T pallidum DNA detection in blood has been reported in 0–58.3% of patients with latent disease.9 14 22–25 Kouznetsov et al reported higher detection of T pallidum DNA in peripheral blood mononuclear cell samples in early latent (4/4) than in late latent (1/) cases, and in our study, quantifiable low-load (70 copies) bacteraemia was found in only one patient (1/7, 14.3%), who had early latent disease.14
Combining the qualitative and quantitative findings, a pattern of bacterial dissemination emerges. In primary disease, high numbers of organisms in ulcers but low copy number or no tpp47 DNA were detected in blood. In secondary disease, the healing ulcers have decreased load but bacteraemia is more frequently detected and at higher load than primary disease. By the latent stage, treponemal bacteraemia is less commonly detected and when present is at low load.
Syphilis and HIV-1 are known to interact in a number of ways. Both HIV-1 transmission and acquisition are facilitated by the presence of syphilitic ulcers and there is evidence that those ulcers and the clinical course of syphilis can be affected by HIV-induced immunosuppression.3 Rompalo et al studied 214 patients with genital ulcers and found HIV-1-positive men to be 11% more likely to have secondary syphilis with a concomitant primary chancre.26 Similarly, in this study, HIV-1-positive patients with secondary syphilis were more likely to present with unhealed primary ulcers. HIV-1-positive patients had a higher number of organisms in primary ulcers than patients known to be HIV-1 uninfected. Although this association did not reach a high level of statistical significance, the presence of more organisms suggests that HIV-1-positive patients with primary syphilis are more infectious and for longer, which may help explain the high rates of syphilis–HIV co-infection.
This study had small sample sizes in some groups, which limited the statistical significance of associations. It is also appreciated that longitudinal sampling of the same patients during the course of their infection would more accurately reflect infection kinetics but would not be ethical. Importantly though, we demonstrate that quantification of T pallidum has the potential to determine the kinetics of the infection and the impact of HIV-1 co-infection on those kinetics. In this study, a pattern of T pallidum load by site and stage is described with important observations related to the infectiousness of HIV-positive patients with primary syphilis.
Real-time PCR can be used to quantify Treponema pallidum in clinical samples and provide insight into the kinetics of syphilis infection.
The amounts of T pallidum found in ulcers and blood at different stages of syphilis follow a biologically plausible pattern.
Larger numbers of T pallidum in ulcers may mean HIV-1-positive patients can more easily transmit syphilis.
Funding National Institute for Health Research, London. Other funders: NIHR BRC (Imperial College London).
Competing interests The corresponding author has had full access to all the data in the study and had final responsibility for the decision to submit for publication. These data were presented, in part, at the IUSTI World Congress, Cape Town, 2009.
Ethics approval This study was conducted with the approval of the Hounslow and Hillingdon Local Research Ethics Committee. Ref: 06/Q0407/29.
Provenance and peer review Not commissioned; externally peer reviewed.