Culture and identification of Borrelia spirochetes in human vaginal and seminal secretions

Background: Recent reports indicate that more than 300,000 cases of Lyme disease are diagnosed yearly in the USA. Preliminary clinical, epidemiological and immunological studies suggest that infection with the Lyme disease spirochete Borrelia burgdorferi (Bb) could be transferred from person to person via intimate human contact without a tick vector. Failure to detect viable Borrelia spirochetes in vaginal and seminal secretions would argue against this hypothesis. Methods: Patients with and without a history of Lyme disease were selected for the study after informed consent was obtained. Serological testing for Bb was performed on all subjects. Semen or vaginal secretions were inoculated into BSK-H medium and cultured for four weeks. Examination of genital cultures and culture concentrates for the presence of spirochetes was performed using light and darkfield microscopy, and spirochete concentrates were subjected to Dieterle silver staining, anti-Bb immunohistochemical staining, molecular hybridization and PCR analysis for further characterization. Immunohistochemical and molecular testing was performed in three independent laboratories in a blinded fashion. Positive and negative controls were included in all experiments. Results: Control subjects who were asymptomatic and seronegative for Bb had no detectable spirochetes in genital secretions by PCR analysis. In contrast, spirochetes were observed in cultures of genital secretions from 11 of 13 subjects diagnosed with Lyme disease, and motile spirochetes were detected in genital culture concentrates from 12 of 13 Lyme disease patients using light and darkfield microscopy. Morphological features of spirochetes were confirmed by Dieterle silver staining and immunohistochemical staining of culture concentrates. Molecular hybridization and PCR testing confirmed that the spirochetes isolated from semen and vaginal secretions were strains of Borrelia, and all cultures were negative for treponemal spirochetes. PCR sequencing of cultured spirochetes from three couples having unprotected sex indicated that two couples had identical strains of Bb sensu stricto in their semen and vaginal secretions, while the third couple had identical strains of B. hermsii detected in their genital secretions. Conclusions: The culture of viable Borrelia spirochetes in genital secretions suggests that Lyme disease could be transmitted by intimate contact from person to person. Further studies are needed to evaluate this hypothesis.


Introduction
Lyme disease is the most common human tick-borne disease in the world today (Stricker & Johnson, 2014). It is transmitted by Ixodes ticks and is caused by the spirochete Borrelia burgdorferi (Bb) (Burgdorfer et al., 1982). Bb is phylogenetically related to the spirochetal agent of syphilis, Treponema pallidum (Gupta et al., 2013). T. pallidum is transmitted sexually between partners through contact of mucosal membranes, gaining access to the bloodstream through microabrasions and then disseminating systemically (Ho & Lukehart, 2011;LaFond & Lukehart, 2006). The close phylogenic relationship of Bb to T. pallidum suggests that this mode of transmission might be possible for Bb.
In addition to theoretical considerations, evidence for non-vector transmission of Bb is based on animal models. Proof of contact transmission of Bb -without involvement of an arthropod vectorwas established by two studies in mice. Burgess et al. (1986) caged uninfected deer mice with experimentally-infected deer mice and demonstrated transmission of Bb by seroconversion of contactexposed mice from negative to positive and by the isolation of Bb from the blood of one contact-exposed mouse 42 days after initial contact. A study by Wright & Nielsen (1990) demonstrated that white-footed mice were susceptible to oral infection and transmitted infection to each other through direct contact. Furthermore, sexual transmission of Bb has been proposed in a canine model. Bb was transmitted to uninfected female dogs in estrus via semen by natural breeding with male dogs infected experimentally with Bb (Gustafson, 1993). Successful transmission of infection from male dogs to female dogs was shown by seroconversion of female dogs from negative to positive as well as the detection of Bb DNA in the tissue of fetuses from resulting pregnancies. If contact transmission of Bb occurs in mice and sexual transfer occurs in dogs, it is not unreasonable to postulate similar routes of infection in humans.
We sought to determine if viable Borrelia spirochetes could be recovered from human vaginal and seminal secretions, an important first step to investigate whether sexual transmission of these spirochetes among humans is possible.
Materials and methods 1. Research subject selection Control subjects who were asymptomatic without a history of Lyme disease and patients with a history of Lyme disease were recruited for the study after written informed consent to collect and publish their data was obtained. Approval for sample collection was obtained from the Western Institutional Review Board, Olympia, WA (WIRB ® #20141439). Further approval for sample testing was obtained from the Institutional Review Board of the University of New Haven, West Haven, CT. Serological testing of all participants after coding of their blood samples was performed by IGeneX Reference Laboratories, Palo Alto, CA in a blinded fashion.
Patients were considered positive for Lyme disease if they were serologically positive by CDC criteria and/or IGeneX criteria, as previously described (Engstrom et al., 1995;Ma et al., 1992), or if they had musculoskeletal, neurocognitive and/or cardiac symptoms clinically consistent with a Lyme disease diagnosis, as described elsewhere (Donta, 2014;Smith et al., 2014). None of the patients were taking antibiotics at the time of testing.

Borrelia cultures
Borrelia spirochetes were cultured as previously described (Bankhead & Chaconas, 2007;Middelveen et al., 2013b;Middelveen et al., 2014a). The inoculum for blood culture was prepared as follows: 10 milliliters of whole blood was collected by sterile venipuncture from each patient. Samples sat at room temperature for 10 to 15 minutes allowing clotting to occur. Red blood cells (RBCs) were separated by low speed centrifugation. Barbour-Stoner-Kelly H (BSK-H) complete medium was used for cultures with the addition of 6% rabbit serum (Sigma Aldrich, #B8291) and the following antibiotics: phosphomycin (0.02 mg/ml), rifampicin (0.05 mg/ml), and amphotericin B (2.5 µg/ml) (Sigma Aldrich).
The culture medium described above was inoculated for blood culture with the spun serum containing white blood cells and some RBCs, and for genital culture with either ejaculated semen or vaginal secretions collected by intravaginal swabbing with a sterile cotton-tipped swab. Blood and genital cultures were incubated at 32°C in an Oxoid anaerobic jar (Thermo Scientific) containing an AnaeroGen sachet (Thermo Scientific) to provide an anaerobic environment. Cultures were incubated for four weeks and checked weekly by light and/or darkfield microscopy for visible motile spirochetes.
All cultures were processed for microscopic imaging and PCR by centrifuging the culture fluid at 15,000 g for 20 minutes to concentrate spirochetes. The supernatant was discarded and the pellet retained. The pellet samples were coded and processed in a blinded fashion for subsequent experiments.
3. Dieterle silver staining Dieterle silver staining was performed using two fixation methods. In the standard method, formalin-fixed, paraffin-embedded pellets

Amendments from Version 2
We have addressed all of the latest referee comments. Please see the detailed response to those comments. In particular, we have added control data for the Dieterle silver staining and anti-Bb immunostaining. We have also provided details for the molecular hybridization testing and we have justified the variation in morphology in our samples with appropriate references from the medical literature. In addition, we have added a discussion of why our PCR results are not due to contamination. We wish to emphasize that our article demonstrates culture detection of live Borrelia spirochetes in genital secretions from Lyme disease patients using a combination of light and darkfield microscopy, silver staining, immunostaining, molecular hybridization and PCR techniques. Although any of these techniques could yield faulty results, the combination of all modalities performed in three independent laboratories provides substantial corroborative support for the findings of our study. We also wish to emphasize that our study does not claim to prove sexual transmission of Lyme disease. We are simply showing that Borrelia spirochetes can be cultured from genital secretions of Lyme disease patients. We hope that the referees will reconsider their negative reviews of the article after reading our detailed responses to their comments, in the true spirit of peer review.

REVISED
were sectioned and stained with Dieterle silver stain as previously described (Aberer & Duray, 1991;Middelveen et al., 2013a). In the newer method, culture fluid was spread and dried on a SuperFrost™ Plus microscope slide (Fisher Scientific) and fixed by incubating the slide in acetone for 10 minutes at -20°C, as previously described (Sapi et al., 2013). Dieterle silver staining was performed on the acetone-fixed slide.
Positive and negative culture controls were prepared for comparison purposes with plasma from Bb-inoculated mice and uninfected mice followed by Dieterle silver staining using the standard method. Control cultures of mixed Gram-positive and mixed Gram-negative bacteria were also subjected to Dieterle staining. The control processing and staining was performed at McClain Laboratories LLC, Smithtown, NY.

Anti-Bb immunostaining A. McClain Laboratories.
Blood and genital culture pellets from coded patient samples were processed in a blinded fashion for special staining at McClain Laboratories. Formalin-fixed, paraffin-embedded pellets were sectioned and stained with anti-Bb immunostain for spirochete detection, as previously described (Middelveen et al., 2013a;Middelveen et al., 2014a). In brief, immunostaining was performed using an unconjugated rabbit anti-Bb polyclonal antibody (Abcam ab20950), incubated with an alkaline phosphatase probe (Biocare Medical #UP536L), followed by a chromogen substrate (Biocare Medical #FR805CHC), and counterstained with hematoxylin. Positive and negative culture controls were prepared for comparison purposes with plasma from Bb-inoculated mice and uninfected mice followed by anti-Bb immunostaining. Culture pellets from fungalinfected human skin samples, mixed Gram-positive bacteria and mixed Gram-negative bacteria were also prepared for comparison purposes as negative anti-Bb immunostain controls to exclude crossreactivity with commonly encountered microorganisms. Staining was titrated to determine optimal antibody dilutions to achieve positive staining of spirochetes while minimizing background staining (Middelveen et al., 2013a;Middelveen et al., 2014a).

B. University of New Haven.
Coded samples were processed in a blinded fashion for Bb immunostaining as previously described (Sapi et al., 2013). Culture fluid was spread and dried on a SuperFrost™ Plus microscope slide (Fisher Scientific) and fixed by incubating the slide in acetone for 10 minutes at -20°C. Dried, fixed culture fluid was submerged under 100 µl of polyclonal FITC-labeled rabbit anti-Bb antibody (Thermo Scientific #PA-1-73005) diluted 1:50 in 1× PBS buffer with 1% BSA (Sigma Aldrich #A9418). For negative controls, the antibody was omitted and replaced with normal rabbit serum. The slides were then incubated for 1 hour at 37°C in a humidified chamber, washed with 1× PBS for 5 minutes at room temperature, rinsed twice in double distilled water and dried in a laminar air-flow hood for 10 minutes. The slides were mounted with Vectashield mounting medium (Vector Labs) and viewed with fluorescent microscopy at 400× magnification with a Leica DM2500 microscope (Sapi et al., 2013).

Molecular hybridization using Bb DNA probe
The Bb molecular beacon DNA probe was generously provided by Dr. Alan MacDonald. Probe FlaB (sequence of 23 mer TGGGAGTT-TCTGGTAAGATTAAT) was derived from the Bb open reading frame (ORF) BB0147 (approximately 1100 mer) of the flagellin B gene. A nucleotide Basic Local Alignment Search Tool (BLAST) search of the 23 mer sequence disclosed no matches in the human genome or in any other life form other than the Bb sequence of BB0147.
Bb detection with the molecular beacon was performed as previously described (Middelveen et al., 2014a) on coded samples in a blinded fashion using the following protocol: paraffin sections were dewaxed by baking at 60°C, then immersed in serial 100% xylene baths followed by serial immersion through baths of 100% ethanol, 90% ethanol, 80% ethanol, and finally in distilled H 2 O, and then airdried. Fixed sections were immersed in 20 µl of the working DNA beacon solution. The sectioned specimen was covered with a layer of plastic cut from a Ziploc ® freezer bag and was heated at 90°C for 10 minutes to denature DNA and RNA. The heat was first reduced to 80°C for 10 minutes, then the slides were removed from heat and allowed to gradually cool to 24°C. The slides were washed in PBS, covered with 30% glycerol and a glass coverslip, then examined under an EPI Fluor microscope. Staining of test specimens was performed alongside staining of positive and negative controls. The positive control was prepared by embedding a known Bb strain in agarose, formalin-fixing the specimen then blocking in paraffin and staining sections as described above. from cultures was followed by BLAST comparison with known Borrelia sequences, as previously described (Mayne et al., 2012).

B. University of New
Haven. DNA samples were extracted from blood, vaginal or seminal cultures by lysing cells overnight in 180 µl tissue lysis buffer (Qiagen) and 20 µl Proteinase K (Qiagen) at 56°C in a shaking water bath followed by phenol:chloroform extraction the next day. The DNA was resuspended in 50-100 µl 1×TE buffer.
A published TaqMan assay targeting a 139-bp fragment of the gene encoding the Borrelia 16S rRNA was used for the detection of Borrelia in DNA extracted from patient samples (O'Rourke et al., 2013). All reactions were carried out at a final volume of 20 µl and consisted of 900 nM of each primer, 200 nM of probe, and 10 µl of 2× TaqMan Universal PCR Master Mix (Applied Biosystems) and 1 nanogram of DNA. Amplifications were carried out on a CFX96 Real-Time System (Bio-Rad), and cycling conditions consisted of 50°C for 2 minutes, 95°C for 10 minutes, followed by 40 cycles of 95°C for 15 seconds and 60°C for 60 seconds. Fluorescent signals were recorded with CFX96 Real-Time software and Cq threshold was set automatically. The reactions were performed in triplicate with positive and negative controls.
Nested PCR primers for the genes encoding the Borrelia 16S rRNA, fla and pyrG loci were used as previously described (Clark et al., 2013;Margos et al., 2010;Sapi et al., 2013). Reactions were carried out in a final volume of 50 µl using 10 µl template DNA. Final concentrations were 2× Buffer B (Promega), 2 mM MgCl 2 , 0.4 mM dNTP mix, 2 µM of each primer, and 2.5 U Taq polymerase (Invitrogen). "Outer" primers were used in the first reaction. "Inner" primers were used for the nested reaction, in which 1 µl of PCR product from the first reaction was used as template for the second. Cycling parameters were as follows: 94°C for 5 minutes followed by 40 cycles of denaturation at 94°C for 1 minute, annealing for 1 minute (temperature based on the primer set used), and extension at 72°C for 1 minute, with a final extension step at 72°C for 5 minutes. PCR products were visualized on 1-2% agarose gels. Sanger sequencing was used for gene analysis, as previously described (Margos et al., 2010).

Patient data
All patient data are shown in Table 1. The control group included four asymptomatic patients (two males and two females). All four were seronegative for Bb.
The patient group included six male subjects and seven female subjects, including four pairs of partners (Patients 6 and 7, 8 and 9, 10 and 11, and 12 and 13, respectively). Eleven of the 13 patients selected for the study were serologically positive for Lyme disease. Patient 1 was serologically equivocal and patient 8 was seronegative, although Bb plasmid DNA was detected in whole blood and serum from this patient.

Light and darkfield microscopy
Blood cultures from 11 patients were incubated for four weeks and checked weekly for spirochete growth using light and darkfield microscopy. Motile spirochetes and/or motile spherules were  Figure 1A). See Dataset, data file 1.
Most genital cultures grew very well and contained abundant spirochetes, but some blood cultures contained few spirochetes. Therefore, to better document the presence of spirochetes in culture, the culture fluid was concentrated into pellets by centrifugation (Table 3). Spirochetes and/or spherules were detected by sectioning and special staining of paraffin blocked pellets in all the patient blood and genital cultures concentrated by centrifugation, except for blood and genital culture pellets from Patient 1 that were lost during paraffin blocking (Table 3). Control genital culture samples were sent directly for PCR testing and were not subjected to light and darkfield microscopy.

Immunohistochemistry
A. Dieterle silver staining. The culture samples of uninfected mouse plasma, mixed Gram-positive bacteria and mixed Gram-negative bacteria failed to stain with Dieterle silver stain using the standard staining method. In contrast, the culture sample of Bb-infected mouse plasma stained positive for spirochetes with Dieterle silver stain (Dataset, data file 2A).
Using standard Dieterle staining, spherules and/or spirochetal forms were visible in all patient genital cultures ( Figure 1B). Spirochetes were detected in all patient genital culture pellets except for Patient 1, whose pellet was lost during processing (  the newer fixation method, spirochetes and sperm cells were visible in semen samples and showed distinct morphology ( Figure 1C). Sperm cells are known to stain with silver stains (Pathak et al., 1979;Schmid et al., 1983). Sperm cells were seen in all semen samples except for Patients 2 and 6, who had vasectomies (data not shown). Since control genital cultures had no visible spirochetes, the control samples were sent directly for PCR testing and were not subjected to Dieterle silver staining. See Dataset, data file 2.

B. Anti-Bb immunostaining.
I. Culture fluid -University of New Haven Genital culture fluid from Patient 1 was fixed on a SuperFrost™ Plus microscope slide and was stained with FITC-labelled polyclonal anti-Bb antibody. Staining was strongly positive, revealing well-defined spirochetes morphologically consistent with Bb ( Figure 2A). The polyclonal antibody was not reactive to T. denticola (data not shown).

II. Culture pellets -McClain laboratories
The culture sample of uninfected mouse plasma failed to stain with anti-Bb immunostain. In contrast, the culture sample of Bb-infected mouse plasma stained positive for spirochetes with anti-Bb immunostain (Dataset, data file 3A). Control fungalinfected human skin cultures, Gram-positive bacterial cultures and Gram-negative bacterial cultures all failed to stain for spirochetes with the anti-Bb immunostain (Dataset, data file 3A).
Anti-Bb immunostaining was positive for all genital cultures except for Patient 1, whose pellet was lost during processing (Table 4). Immunostaining revealed both spiral and globular Bb forms ( Figure 2B). Since control genital cultures had no visible spirochetes, the control samples were sent directly for PCR testing and were not subjected to immunostaining. See Dataset, data file 3.

Molecular hybridization
Hybridization with the Fla B probe was positive for genital culture pellets from Patients 2-9 ( Table 4). The culture pellet from Patient 1 was lost during processing. The molecular probe showed intense staining in vaginal secretions and less intense staining in semen samples ( Figure

B. University of New
Haven. PCR testing using the TaqMan assay for Borrelia 16S rRNA sequence was positive in blood culture pellets from seven of nine patients tested (Table 5B). Patients 1 and 5 had negative results in blood culture pellets using the TaqMan assay, but both were positive by nested PCR for the pyrG gene. In addition, nested PCR targeting the fla gene was performed on blood culture pellets from Patients 2, 3 and 4, and nested PCR targeting the 16S rRNA gene was performed on the blood culture pellet from Patient 6. The samples were positive, and sequencing revealed 99-100% homology with Bb sensu stricto strain B-31 (Table 5B). See Dataset, data file 7.
PCR testing using the TaqMan assay for Borrelia 16S rRNA sequence was negative in all four control genital culture pellets, and nested PCR targeting the pyrG and fla genes was negative in all four control samples, confirming the results of the TaqMan assay (Table 5B). In contrast, eight of nine patients were positive for Taq-Man 16S rRNA sequence in the genital culture pellets. Patient 6 was negative using the TaqMan assay for 16S rRNA sequence but positive using nested PCR targeting a different portion of the 16S rRNA gene (Table 5B). Nested PCR targeting the fla gene (Patient 3) and the 16S rRNA gene (Patients 3 and 7) was also performed on genital culture pellets and was positive in those patients, confirming the results of the TaqMan assay. Patient 12 had positive PCR targeting the pyrG gene with confirmatory sequencing (see below).
6. Sequencing of Borrelia detected in blood and genital cultures PCR isolates of the vaginal culture from Patient 1 (Australian Biologics) and the seminal culture from Patient 3 (University of New Haven) were subjected to Sanger sequencing and BLAST analysis and showed 97-99% homology with Bb sensu stricto strain B-31 (Table 5A and Table 5B). See Datasets, data files 6 and 7. PCR isolates of blood cultures from Patients 2, 3, 4 and 6 were subjected to Sanger sequencing and BLAST analysis at University of New Haven and showed 99-100% homology with Bb sensu stricto strain B-31 (Table 5B). See Dataset, data file 7.
PCR isolates of genital cultures from three couples having unprotected sex (Patients 6-7, 10-11 and 12-13) were subjected to Sanger sequencing and BLAST analysis. Patients 6, 7, 10, 11 and 13 had sequencing done at Australian Biologics, while Patient 12 had sequencing done at University of New Haven. Sequencing revealed that the first and third couples had Borrelia strains that matched Bb sensu stricto strain B-31 (Table 6). In contrast, the second couple had PCR sequences that matched B. hermsii strain YOR. Thus the Borrelia strain shared by this couple differed significantly from the strains identified in the other couples. See Dataset, data file 6.

Discussion
In this study using standard and published culture, immunohistochemical, molecular hybridization and PCR techniques, we have shown that Borrelia strains are present in semen and vaginal secretions from patients with Lyme disease. Simultaneous testing for treponemal spirochetes was negative in genital secretions of all Lyme disease patients, confirming the specificity of Borrelia detection in these patients. Furthermore we have shown that couples having unprotected sex have virtually identical strains of Borrelia in their genital secretions, suggesting that Borrelia spirochetes might be transmitted from person to person without a tick vector.
As expected, PCR sequencing of cultured Borrelia from semen and vaginal secretions yielded primarily Bb sensu stricto strains, reflecting the North American origin of our study subjects. In addition, PCR sequencing of genital secretions from one couple yielded identical strains of Bb sensu stricto strains in two different laboratories. However, we were surprised to find one couple with     et al., 1986;Burgess & Patrican, 1987;Burgess, 1989;Burgess, 1992;Wright & Neilsen, 1990). Bb has been shown to survive in stored semen from dogs, rams and bulls (Kumi-Diaka & Harris, 1995). Furthermore, seminal transmission of Bb has been noted in dogs, as described above (Gustafson, 1993). In contrast, contact transmission of Bb could not be demonstrated in Lewis rats and Syrian golden hamsters (Moody & Barthold, 1991;Woodrum & Oliver, 1999). Technical limitations in the study of these highly inbred rodents including limited contact between animals and failure to perform molecular testing may have contributed to the negative results.
While it is not possible to perform controlled sexual transmission studies of Borrelia in humans, several investigators have speculated that this mode of transmission is possible (Bach, 2001;Harvey & Salvato, 2003;Stricker et al., 2004). The suggestion that Bb could be transmitted sexually was initially proposed by Bach in 2001. He observed that sexually active patients had a marked propensity for antibiotic failure and speculated that re-infection occurred by intimate person-to-person contact. Bb DNA was detected by PCR technology in human breast milk, umbilical cord blood, semen and vaginal secretions taken from patients presenting at his practice (Bach, 2001).
The study of a group of chronically ill Bb-seropositive and PCRpositive patients in Houston, Texas -a non-endemic area -provided epidemiological evidence that Lyme disease could spread in the absence of a suitable vector (Harvey & Salvato, 2003). In the absence of infected ticks, intimate person-to-person transfer was implicated as the probable means of transmission (Harvey & Salvato, 2003 -Wierzbanowska, 1993;Metcalf, 2001;Miceli et al., 2010;Milazzo et al., 2001).
The number of spirochetes needed to infect an animal or human varies according to strain-specific biological and transmission factors. In mouse studies of experimental Borrelia infection, the 50% infectious dose was 18 spirochetes with tick salivary gland extract and 251 spirochetes with tick midgut extract (Cook, 2014). Transmission studies of syphilis using "human volunteers" found that the 50% infectious dose was approximately 57 organisms (LaFond & Lukehart, 2006). At present, the spirochetal load in genital secretions from Lyme disease patients is unknown, but it appears that genital infection could be induced by a relatively small number of organisms based on the studies outlined above. It is known that seminal plasma inhibits the immune response to Gram-negative pathogens (Brooks et al., 1981), while the female genital tract induces immune factors that may be conducive to spirochete survival (Clark & Schust, 2013;Wira et al., 2005). The role of the male and female genital tracts in tolerance and propagation of Borrelia infection merits further study.
Lyme disease diagnosis is based largely upon serological testing using CDC-sanctioned two-tier surveillance criteria supported by FDA-approved commercial test kits. While most patients in this study did have positive serological test results for Lyme borreliosis, some were considered serologically negative, and the majority of our study subjects did not meet the positive standard as defined by the CDC surveillance criteria (CDC, 2014a). We were able to detect Borrelia spirochetes in the blood and/or genital secretions of all patients who were clinically diagnosed with Lyme disease, demonstrating that the CDC surveillance protocol is inadequate diagnostically. Inadequate diagnostic methodology undoubtedly results in under-reporting of Lyme disease, and at least one group has speculated that this substandard methodology is considered The dataset contains data files 1, 2, 2a, 3, 3a, 4, 5, 5a, 6 and 7. Detailed legends for each files can be found in the text file provided.
acceptable because Borrelia is not sexually transmitted (Lange & Sayyedi, 2002). In addition, if Borrelia spirochetes were transmitted sexually, then patients with false-negative results may unknowingly spread the infection to sexual partners.
The 2011 CDC case definition for Lyme disease states that a positive Bb culture confirms the diagnosis of the disease (CDC, 2014b). Although culture of Borrelia genital isolates may be a useful diagnostic laboratory methodology in the future, detecting and characterizing cultured Borrelia isolates is not straightforward, and both false-positive and false-negative results could occur. In our experience, human clinical isolates from genital secretions frequently propagate prolifically in culture, but on occasion they do not. In such instances, the culture must be concentrated and specific staining should be conducted to ascertain the presence of spirochetes.
Once detected, spirochetes must be characterized genetically for specific identification. PCR is currently the most reliable means for correctly identifying cultured isolates, but even this methodology has drawbacks and limitations (Lange & Sayyedi, 2002;Nolte, 2012).
There are currently no standardized FDA-approved PCR protocols or kits available for Bb detection, so commercial PCR testing constitutes an array of "home brew" assays using different methodologies such as real-time PCR and nested PCR, with various primers targeting different genes, yielding wide differences in sensitivity and specificity (Nolte, 2012;Schmidt, 1997;Yang et al., 2012). False negatives can result because primers may be strain-specific and may not detect all Borrelia genotypes, and fluids such as blood, semen and vaginal secretions may contain substances inhibitory to the PCR process (Lange & Sayyedi, 2002;Nolte, 2012;Yang et al., 2012). The potential for false-positive PCR testing may also arise if there is DNA contamination in the laboratory, and appropriate positive and negative controls must be included in the assay (Lange & Sayyedi, 2002;Nolte, 2012). We experienced differences in primer specificity in our clinical isolates and also found that inhibition occurred, particularly in semen cultures.
Another complicating factor in Borrelia isolation is the morphological variation of the spirochete, which includes spherical, granular or cystic forms. Morphological variants of Bb, some of which are not culturable, are well documented in the medical literature (Barthold et al., 2010;Hodzic et al., 2014;Kurtti et al., 1987;MacDonald, 2013;Meriläinen et al., 2015;Mursic et al., 1996). These variants may play a role in infection, enabling Bb and other pathogenic spirochetes to evade the immune system (Döpfer et al., 2012;Menten-Dedoyart et al., 2012;Mursic et al., 1996). Limited Bb growth and non-spiral morphology are thought to be induced by unfavorable environmental conditions (Brorson et al., 2009), and these features appear to be consistent with our observations. We found that Borrelia growth was more vigorous with more long slender morphological variants in cultures of genital secretions compared to cultures of blood, and we speculate that the human circulatory system is a more hostile environment for Borrelia than the human reproductive system.  , 2015). A study from military treatment facilities in the USA "unexpectedly" found no association between the incidence of Lyme disease and the prevalence of infected ticks, and the rate of Lyme disease was 2.6 times higher in officers than enlisted men (Rossi et al., 2015). Second, while sexually transmitted diseases like herpes simplex virus (HSV) and gonorrhea show an urban predominance, Lyme disease has a more rural distribution (Craig, 2014). However, Lyme disease is acquired in more ways than HSV and gonorrhea, and the rate of sexual transmission is unknown at present. Thus the epidemiology of Lyme disease may differ from other sexually transmitted diseases based on these undefined variables. Third, the transmission of HIV can be traced from one sex partner to another using HIV strain typing. Based on our study, a similar transmission pattern using Borrelia strain typing may be seen once larger studies are performed among couples having unprotected sex. In summary, sexual transmission of Borrelia is plausible in light of our limited knowledge about the risk of acquiring Lyme disease.
In conclusion, we have shown that Borrelia spirochetes are present in semen and vaginal secretions of patients with Lyme disease. Furthermore, virtually identical strains of Borrelia are present in couples having unprotected sex, suggesting that transmission via intimate contact without a tick vector may occur. The epidemiology and clinical risk of Borrelia sexual transmission remain to be determined.

Consent
Written informed consent to publish clinical details and study results was obtained from each participant. The third version of the paper from Middelveen , "Culture and identification of spirochetes in et al.
Borrelia human vaginal and seminal secretions" represents a well designed and nicely executed study strongly supported by multiple results that were obtained by three independent laboratories. The impact of the presented data is significant as it highlights the alternative way of the distribution of the causative agent of Lyme disease, spirochetes, rather than the traditionally accepted Borrelia burgdorferi sensu stricto vector-borne way. The well balanced study presents the cultivation of live and replicating spirochetes from genital secretions, confirming the possibility of person-to-person transmission of Lyme disease spirochetes. Supported by multiple techniques, the obtained results are clear, consistent and confirming. The authors do not bring the conclusion or speculate that spirochetes transmitted from person-to-person might trigger Lyme disease in humans, as this postulation needs to be deeply studied. However, confirmation of the fact that live Borrelia are present in semen and vaginal secretions of patients with Lyme disease and can be cultured in active and replicating form from them is the first and significant step in analysis of person-to-person transmission of the causative agent of Lyme disease. This may be one of the reasonable explanations of the increased numbers of confirmed cases of Lyme disease worldwide, in addition to the others that are not discovered yet.

If applicable, is the statistical analysis and its interpretation appropriate? Not applicable
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
No competing interests were disclosed.

Competing Interests:
We have read this submission. We believe that we have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. Please be aware that I have invested as much time as I can in the review process for this article, and will be unable to contribute further.
No competing interests were disclosed.

Competing Interests:
I have read this submission. I believe that I have an appropriate level of expertise to state that I do not consider it to be of an acceptable scientific standard, for reasons outlined above.

Robert Smith
School of Life Sciences, University of Glasgow, Glasgow, UK This current version of the paper is one which contributes significantly to the field. I re-affirm my original approval of what I consider is a well executed and well presented study and is worthy of indexation for rational scrutiny and further investigations by the wider medical and scientific communities.
No competing interests were disclosed. Competing Interests: The design of these experiments to test the hypothesis that viable spirochetes exist in genital secretions is appropriate. However, the implementation of the experiments, presentation, and interpretation of data is questionable.

I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
Culture of from secretions: These are not sterile sources, yet there is no mention of B. burgdorferi contamination with other microbes. The authors should show (supplemental data) staining of secretions and cultures from healthy controls. One would expect some level of background staining.
Images shown in Figures 1, 2 and 3 are not specifically detecting spirochetes and are questionable without appropriate controls. The darkfield and silver stain are of very little use, given that "pseudospirochetes" of flagellar bundles can forms in body fluid cultures. No positive or negative controls are shown. The straight morphology and varying lengths of that depicted in Figure 1A (and in all of the dataset 1 darkfield samples provided) are inconsistent with morphology. Specific staining B. burgdorferi shown in Fig 2A is more consistent with spirochete morphology and a legitimate detection method (albeit a low quality image). However, the authors indicate in the Methods section that DAPI counterstain was used but no blue staining is apparent. Does this mean that there were no other sources of DNA in the culture (bacteria or cells)? Clearly, there is ample contamination with other bacteria (seen in darkfield images of dataset 1). Concerning immunostaining with the FlaB probe, the methods are not described and no positive or negative control samples are shown. With the exception of Patient 10, none of the silver stained samples resemble spirochetes.
This reviewer encourages readers to examine the images provided in supplementary datasets.
Concerning the molecular detection, the sequence for the 23mer FlaB DNA probe should be provided, as should evidence of validation with controls. If multiple genes were amplified from cultures, then those sequences should also be provided. Assuming there was a positive control, 98-100% homology cannot distinguish between a legitimate infection and PCR contamination (with introduction of errors from PCR). Sequences from patients 10 and 11 as is misleading, given that they are short, disconnected B. hermsii regions of homology. Given that 16s was used, the entire product should line up. BLAST input was not provided for Patient 12 and the region of homology for patient 13 was only 50 bp (where for patient 1, 6 and 7 it was ~120 bp).
In conclusion, this manuscript would be made much clearer if all of the control (positive and negative) data were also included. Validation of the assays employed is unapparent. The paper cited from the Australian Biologics assay indicated PCR detection of in patients at a much higher frequency Borrelia than IgG or IgM immunoblot detection of exposure. This is inconsistent with multiple reports. Furthermore, the PCR primers are listed as "proprietary" so an assessment of specificity cannot be made. Sending blinded samples to an impartial 3rd party for testing using validated assays is recommended.
No competing interests were disclosed.

Competing Interests:
I have read this submission. I believe that I have an appropriate level of expertise to state that I do not consider it to be of an acceptable scientific standard, for reasons outlined above.

The design of these experiments to test the hypothesis that viable spirochetes exist in genital secretions is appropriate. However, the implementation of the experiments, presentation, and interpretation of data is questionable.
Culture of B. burgdorferi from secretions: These are not sterile sources, yet there is no mention of contamination with other microbes. The authors should show (supplemental data) staining of secretions and cultures from healthy controls. One would expect some level of background staining.
We acknowledge that these are not sterile sources. Three antibiotics were used to limit the growth of normal flora. Although cultures may contain other flora, obtaining pure Borrelia isolates was not the purpose of the study. The negative controls from healthy subjects did not show PCR evidence of spirochetal infection, as seen in Table 5 and Dataset 7. Real-time PCR failed to show spirochetal infection in genital cultures from negative controls (Table 5A), and both TaqMan PCR and nested PCR using a variety of targets failed to show infection in genital cultures from negative controls (Table 5B). To address the Borrelia referee's concern about control silver staining and immunostaining, however, we have performed additional control silver stains and immunostains using a culture from uninfected mouse plasma, which showed no staining for with either technique. Cultures of Borrelia fungal-infected human skin, mixed gram positive strep and staph (from dairy mastitis cultures) and mixed gram negative fecal coliforms (from canine feces cultured on McConkey Agar) also failed to show significant immunostaining for . In contrast, Borrelia plasma culture from a mouse infected with spirochetes processed at McClain Borrelia Laboratory showed positive silver staining and immunostaining. We have included this information in Datasets 2A and 3A and added a description in the Results section on Pages 8 and 9. Figures 1, 2 and 3 are not specifically detecting spirochetes and are questionable without appropriate controls. The darkfield and silver stain are of very little use, given that "pseudospirochetes" of flagellar bundles can forms in body fluid cultures. No positive or negative controls are shown. Figure 1 shows darkfield imaging and silver staining indicating that cultures are living and viable. The Figure shows typical spirochetal morphology, and these viable spirochetes were confirmed to be by PCR, as shown in Table 5 and Dataset 7. Figure 2 shows Borrelia immunostaining that is specific for Borrelia and correlates with positive PCR detection in these samples. Figure 3 shows molecular hybridization with the FlaB probe that is highly specific for . The positive and negative controls for this experiment are shown in Borrelia Supplemental Figure 1. Although each figure by itself might not indicate the presence of , the combination of darkfield imaging, silver staining, immunostaining, PCR testing Borrelia using a range of targets and molecular hybridization provides strong evidence for viable in these genital cultures, as stated repeatedly in the article. Borrelia

Images shown in
The straight morphology and varying lengths of that depicted in Figure 1A (and in all of the dataset 1 darkfield samples provided) are inconsistent with B. burgdorferi morphology.
We respectfully disagree with the referee, and we have included a discussion of morphological variation and pleomorphism in cultures in the Discussion section on Borrelia morphological variation and pleomorphism in cultures in the Discussion section on Borrelia Page 16 with multiple references. Furthermore, to our knowledge nobody has ever studied in seminal or vaginal cultures, so nobody knows what morphological variants would Borrelia be present. Since morphological variation and pleomorphism is acknowledged in Borrelia the medical literature, visual identification alone is not reliable and that is why we used silver staining, immunostaining and molecular detection to characterize the isolates as . Borrelia Specific staining shown in Fig 2A is more consistent with spirochete morphology and a  legitimate detection method (albeit a low quality image). However, the authors indicate in the Methods section that DAPI counterstain was used but no blue staining is apparent. Does this mean that there were no other sources of DNA in the culture (bacteria or cells)? Clearly, there is ample contamination with other bacteria (seen in darkfield images of dataset 1).
We agree with the referee that immunostaining in Figure 2A is consistent with spirochete morphology. The DAPI counterstain was only used in the FlaB probe validation study, as shown in Supplemental Figure 1. We have modified the Methods description to reflect this fact.
Concerning immunostaining with the FlaB probe, the methods are not described and no positive or negative control samples are shown. Wth the exception of Patient 10, none of the silver stained samples resemble spirochetes.
The FlaB probe is a DNA probe, not an immunostain. This DNA hybridization technique is described in detail in Methods section 5 (Page 4) with appropriate references, and positive and negative controls for the technique are shown in Supplemental Figure 1. Again, we have included references to studies showing morphological variation and pleomorphism in cultures in the Discussion section on Page 16. As there are problems with Borrelia specificity of visual identification using silver staining, we used this method to provide corroborative evidence only. It is the collective evidence --including darkfield imaging, immunostaining and molecular detection of borrelial DNA that is important. We used a broad approach with various methods so that there was no reliance on any single methodology, as stated repeatedly in the article.

This reviewer encourages readers to examine the images provided in supplementary datasets.
We encourage readers to do the same. The Dataset images support the findings described in the article.

Concerning the molecular detection, the sequence for the 23mer FlaB DNA probe should be provided, as should evidence of validation with controls.
We have provided the FlaB probe sequence in Methods section 5 (Page 4), courtesy of Dr. Alan MacDonald. Again, this is a molecular technique intended to provide corroborative evidence --along with all the other techniques. A detailed vaildation study was provided in Supplemental Figure 1 clearly showing that the molecular probe was specific for Bb sensu . stricto If multiple genes were amplified from cultures, then those sequences should also be

If multiple genes were amplified from cultures, then those sequences should also be provided. Assuming there was a positive control, 98-100% homology cannot distinguish between a legitimate infection and PCR contamination (with introduction of errors from PCR).
The gene sequences are described in Table 6, and BLAST analyses are shown in Dataset 6. The argument against PCR contamination is presented in the Discussion section on Page 17.
Sequences from patients 10 and 11 as B. hermsii is misleading, given that they are short, disconnected regions of homology. Given that 16s was used, the entire product should line up. BLAST input was not provided for Patient 12 and the region of homology for patient 13 was only 50 bp (where for patient 1, 6 and 7 it was ~120 bp).
The sequences are short, but they were the closest match in those samples. The B. hermsii referee ignores the fact that this couple was the only one with a distinct sequence Borrelia that was detected in both sexual partners and not in any other individual or couple in our small study. The BLAST analysis for Patient 12 was performed using a different program, and labeling for the sequence input has been provided. For Patient 13, the region of homology was short, but it was virtually identical to the sequence in her sexual partner (Patient 12), as shown in Table 6.

In conclusion, this manuscript would be made much clearer if all of the control (positive and negative) data were also included. Validation of the assays employed is unapparent. The paper cited from the Australian Biologics assay indicated PCR detection of Borrelia in patients at a much higher frequency than IgG or IgM immunoblot detection of exposure.
This is inconsistent with multiple reports. Furthermore, the PCR primers are listed as "proprietary" so an assessment of specificity cannot be made. Sending blinded samples to an impartial 3rd party for testing using validated assays is recommended.
As outlined above, we have included control data for the PCR and molecular hybridization testing, which is the "acid test" for the presence of in genital cultures. The assays Borrelia described in our study are internally consistent since the darkfield imaging, silver staining and immunostaining have been validated by the PCR and molecular hybridization procedures. In contrast to the referee's comment, there is ample evidence for "seronegative" Lyme disease, and we have included a discussion of this entity on Page 16 with appropriate references. We used two independent laboratories for PCR testing and molecular hybridization studies, and the results using a range of primers are consistent between the two laboratories. We intend to send blinded samples to independent laboratories for confirmation of our results. In light of our detailed responses to the referee comments, we believe that this preliminary study merits reconsideration for approval by the referee. have, produced a well written and informative account of their findings on the "Culture et al. and identification of Borrelia spirochetes in human vaginal and seminal secretions" from subjects diagnosed as Lyme Disease patients by serological and clinical parameters compared to control individuals who were asymptomatic and serologically negative for the Borrelia burgodorfei. Of most interest is their report of the PCR sequencing of cultured spirochetes from three couples having unprotected sex which indicated that two couples had identical strains of in their semen Bb sensu stricto and vaginal secretions, while the third couple had identical strains of . B. hermsii I believe the study is worthy of acceptance and approve it for the following reasons:

None
The rationale and hypothesis behind the investigation is clearly stated.
The work has been carefully planned, employing an array of pertinent experimental approaches: routine light microscopy, histological silver staining and immunohistological staining methodology, DNA probe analysis and PCR to demonstrate the presence of spirochetes in cultures Borrelia prepared from semen or vaginal secretions from Lyme disease subjects serologically tested for Bb, whilst samples from Bb negative control subjects did not yield spirochetes.
The study is well executed, with samples analysed following blind coding, and techniques performed in a number of independent laboratories.
The data sets, tables and figures are presented in a detailed and informative manner.
The authors demonstrate an authoritative awareness of previous work in the area, citing relevant animal studies which had investigated the possible sexual transmission of Borrelia, and give a plausible explanation for the discrepancies between the findings by Guftafson in dogs and those of Moody and Woodrun in inbred rodent models.
In this second version of the article, the authors have taken pains to address the main concerns of the reviewer Donta, and also answered the comments raised by Baker and others, to version one of the manuscript.
In their replies, Middelveen and Stricker have, in my opinion, more than allayed the criticisms with regard to technical details, and also those concerning aspects such as CDC surveillance criteria and of the unknown spirochetal load.
The conclusions drawn, that the culture of viable spirochetes in human genital secretions Borrelia suggests that Lyme disease could be transmitted by intimate contact from person to person, are justified and of significance. Their findings to date warrant a larger study and further attention by the scientific and medical community.
No competing interests were disclosed.

Competing Interests:
I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. There are a number of issues that mitigate against the authors' conclusion that Lyme disease can be transmitted sexually.
While there are conflicting reports from animal studies that there can be transmission by contact between animals and other studies that appear to better controlled that do not provide such evidence, there is no obvious epidemiological evidence in humans that this is a likely possibility. Apart from the sociologic implications of claiming that intimate or even casual transmission is possible, there needs to be more compelling evidence that this might be the case than is offered in this report.
Specific issues in this report: No evidence that samples were blinded.
The numbers of patients were too small from which to draw meaningful conclusions.
The actual serologic data on patients should be presented in order to be more properly assessed.
PCR-DNA analyses should have been done on the original specimens; again in a blinded fashion, and weekly during the culture period. And, as PCR-DNA testing is much more sensitive than cultures, without this information, the validity of the presented information remains in question.
Concurrent samples of other body fluids, i.e. blood, saliva, should have been included.
Even if presuming the presence of the Lyme in vaginal secretions or semen, the numbers borrelia present would have not been sufficient to cause any transmission of infection, as, with any infectious process, there needs to be a critical inoculum to establish infection. This is the case with transmission by ticks.
Borrelia may be spirochetes as are and , but the transmission of the latter Treponemes Leptospira are obviously through body fluids, and not by tick or other vectors. It is not clear that infections transmitted by ticks are also transmitted by intimate or casual contact.
If it was true that 8/11 samples were positive, one would expect a much more obvious clinical picture of transmission by intimate or casual contact, which is not the case.
PCR-DNA analyses and long-term cultures can be subject to contamination, making the data here more difficult to interpret.
No competing interests were disclosed.

Competing Interests:
I have read this submission. I believe that I have an appropriate level of expertise to state that I I have read this submission. I believe that I have an appropriate level of expertise to state that I do not consider it to be of an acceptable scientific standard, for reasons outlined above.

Co-written with Marianne J. Middelveen
There are a number of issues that mitigate against the authors' conclusion that Lyme disease can be transmitted sexually.
We did not conclude that . Based on preliminary "Lyme disease can be transmitted sexually" editorial comments, we were careful to state that our microscopy, immunochemistry, molecular hybridization and PCR analysis showing live, culturable in semen and Borrelia vaginal secretions suggests that these spirochetes could be transmitted in that manner. Our study does not prove this form of transmission, and we do not make this claim anywhere in the text.
While there are conflicting reports from animal studies that there can be transmission by contact between animals and other studies that appear to better controlled that do not provide such evidence, there is no obvious epidemiological evidence in humans that this is a likely possibility. Apart from the sociologic implications of claiming that intimate or even casual transmission is possible, there needs to be more compelling evidence that this might be the case than is offered in this report.
The statement that "better controlled" studies do not provide support for contact transmission of in animals is contrary to the examples in mice and dogs described Borrelia in the Introduction and Discussion sections of our article. We have pointed out that the two studies on highly inbred rodents that allegedly showed lack of such transmission did not use PCR techniques and therefore may have missed this transmission. The fact that the CDC now admits to more than 300,000 new cases of Lyme disease each year in the USA (and perhaps as many as one million new cases, as outlined in Stricker & Johnson, 2014) is suggestive that other forms of transmission occur, as noted in the Discussion section on page 10.

No evidence that samples were blinded.
In response to the referee's comment, we have noted that the laboratory testing was performed on coded samples in a blinded fashion. This has been noted in the Abstract and reiterated throughout the Methods section.

The numbers of patients were too small from which to draw meaningful conclusions.
The " " are that microscopy, immunochemistry, molecular meaningful conclusions hybridization and PCR analysis demonstrates live, culturable in semen and vaginal Borrelia secretions from Lyme disease patients. Although the numbers are relatively small, we feel that our detailed study supports this conclusion.

3.The actual serologic data on patients should be presented in order to be more properly assessed.
We can include the actual serologic data as an original Dataset, but we don't see how that would alter the experimental findings in our study. The serologic data is presented in Table  1 and the Results section. In response to the referee's comment, we have added more detail about the serologic and diagnostic criteria in the Methods section on page 4, with supporting references.
4. PCR-DNA analyses should have been done on the original specimens; again in a blinded fashion, and weekly during the culture period. And, as PCR-DNA testing is much more sensitive than cultures, without this information, the validity of the presented information remains in question.
PCR-DNA analysis was done on the " in a blinded fashion in conjunction original specimens" with the microscopy, immunochemistry and molecular hybridization analysis. Repeated testing at weekly or other intervals is beyond the scope of this pilot study, and this type of testing should certainly be explored in future Lyme disease studies.

Concurrent samples of other body fluids, i.e. blood, saliva, should have been included.
As shown in Table 5, we did do PCR testing on concurrent blood samples in some patients. Saliva testing for would have been of interest, but this form of testing Borrelia Borrelia requires further investigation and is beyond the scope of our study.

Even if presuming the presence of the Lyme borrelia in vaginal secretions or semen, the numbers present would have not been sufficient to cause any transmission of infection, as, with any infectious process, there needs to be a critical inoculum to establish infection. This is the case with transmission by ticks.
In response to the referee's comment, we have researched the number of spirochetes necessary for transmission of infection in mice and infection in B. burgdorferi T. pallidum humans. The results show that a very small number of spirochetes (as little as 18 organisms) are required for transmission of infection, and we have included this information in the Discussion on page 11. We have also noted that seminal plasma and the female genital tract may provide a relatively permissive environment for spirochetes compared to blood, skin and other immune sites, making transmission theoretically easier via the genital route.

Borrelia may be spirochetes as are Treponemes and Leptospira, but the transmission of the latter are obviously through body fluids, and not by tick or other vectors. It is not clear that infections transmitted by ticks are also transmitted by intimate or casual contact.
In response to the referee's query, we have provided examples of other agents ( , Babesia , ) that are proven or postulated to be transmitted by both tickbite and Chlamydia Coxiella intimate contact. This information is included in the Discussion on page 11 with supporting references.
If it was true that 8/11 samples were positive, one would expect a much more obvious clinical picture of transmission by intimate or casual contact, which is not the case.
As noted above, the substantial numbers of new Lyme disease cases each year suggests As noted above, the substantial numbers of new Lyme disease cases each year suggests additional forms of transmission beyond a tickbite. At this time, the true epidemiology of Lyme disease is unknown because the CDC surveillance system only captures less than 10% of Lyme disease cases, as noted in the Discussion on page 11. Other epidemiological studies have suggested that some infected patients may be relatively asymptomatic ( ), so transmission via intimate contact resulting in less obvious Harvey & Salvato, 2003 infection is plausible. The risk of this form of transmission and correlation with symptoms merits further study. Our report simply raises the possibility, and rejecting the report will shove this issue under the rug to the detriment of Lyme patients.
9. PCR-DNA analyses and long-term cultures can be subject to contamination, making the data here more difficult to interpret.
PCR-DNA analysis is subject to contamination. That is why we did blinded testing that always included negative controls in three different laboratories using microscopy, immunochemistry and molecular hybridization to confirm the PCR findings. Although PCR testing alone might be "difficult to interpret", the combination of experimental techniques done in different laboratories makes interpretation much more reliable. We hope that the referee will change his opinion after reading the revised manuscript and our responses to his comments.
No competing interests were disclosed. Competing Interests: It is probably not surprising that this work, reporting isolation of viable in human vaginal and Borrelia seminal fluid has incited such vigorous commentary; the epidemiological implication of sexual transmission is significant. However, the vigor of the commentary highlights the importance of evaluating the work based on its content rather than concern about the work's implications.

Discuss this Article
Phylogenetic analysis of spirochaetes places as the basal spirochaetal clade, suggesting Leptospiraceae that the ancestral mode of transmission for spirochaetes was through host urine (Radhey et al. 2013). While has evolved mechanisms that allow efficient tick-vectored transmission, there is no reason Borrelia to expect that ancestral mode(s) of transmission would disappear. Incidences of contact or fluid-based transmission may simply be eclipsed by the more efficient arthropod-vectored mode of transmission in most circumstances. Published evidence for non arthropod-vectored contact transmission is extensive. These reports of non-arthropod vectored contact transmission indicate urine/fecal -oral infection routes.
There are multiple reports of viable shed into the urine, in a variety of species. Bosler Borrelia burgdorferi and Schultze (1986) report that bacteria can be cultured from the urine of the wild reservoir host, , although Wright and Neilsen (1990) were not able to find microscopic evidence of Peromyscus leucopus spirochaetes in mouse urine. Similarly, live spirochaetes have been reported to have been cultured from spirochaetes in mouse urine. Similarly, live spirochaetes have been reported to have been cultured from urine from cows (Burgess 1988). Culture from canine urine has been found by some researchers (Grauer et al. 1988) but not others (Appel et al 1993).
The shedding of viable bacteria is not, in itself, sufficient for transmission. However, documentation of infection arising from oral exposure is strong. Ducks can acquire infection orally (Burgess 1989) and, in turn, the bacteria are shed in feces, providing the biological pre-requisites for non-arthopod vectored transmission. Cats also were able to be infected orally (Burgess 1983), detected by seroconversion, direct histology of tissue and bacterial culture from an orally-infected cat. There are multiple reports of seroconversion of uninfected dogs housed in arthropod-free clinical settings with infected dogs (Burgess 1986, Greene et al. 1988, Cerri et al. 1994. The mode of transmission was inferred to be via exposure to viable bacteria in their kennel-mate's urine. Appropriately, the most work on non-arthropod vectored Borrelia transmission has been with the wild reservoir species and Peromyscus maniculatus Peromyscus Burgess and Patrican (1987) provide thorough documentation of oral acquisition of infection, leucopus. assessed by seroconversion and xenotransmission to naive ticks in .

Peromyscus maniculatus
Direct evidence for contact transmission is provided by carefylly controlled experiments in which infected and individuals were able to induce seroconversion in contact-exposed mice P. leucopus P. maniculatus (Burgess et al. 1986). In both species, uninfected mice caged with infected mice became infected. Infection in the absence of an arthropod vector, i.e. contact transmission of infection, was detected by seroconvertion and by culture of viable spirocheates (identified as by immunofluorescence) B. burgdoreferi from the blood of a contact-infected mouse. This work was extended by Wright and Nielsen (1990) who sought to determine whether infection was transmitted through sexual or body fluid contact. Wright and Nielsen (1990) confirmed that could be orally infected. Importantly, they showed that both P. leucopus contact with body fluids/waste and sexual contact were able to transmit infection to co-habiting previously uninfected mice. Infection was detected by seroconversion, histological analysis and IFA. Similarly, Lord et al. (1994) also conclude that contact transmission is taking place in because infection Peromyscus spp. was identified in wild mice in the winter when no ticks were active. These studies provide multiple examples of non-arthropod vectored infection in a variety of species. However, there is certainly evidence that contact-mediated infection may differ biologically from arthropod-vectored infection; Burgess et al.(1986), Wright andNielson (1990) and Appels et al (1993) all note differences in the serological profile of animals infected by ticks vs. other methods, with non-tick infected animals showing less seroreactivity. Clinical manifestations may also differ, which might reflect the role of other pathogens transmitted by the tick vectors versus pure cultures of .

Borrelia burgdorferi
Given the number of published reports of contact transmission and the variety of animals in which this has been documented, this begs the question of why contact transmission was not found in the Moody and Barthold (1991) study using Lewis rats and in the Woodrum and Oliver (1999) study using Syrian hamsters. Although contact transmission was assessed with 30 days co-habitation, sexual transmission was assessed after the rats were in contact for one night. This was sufficient for mating but might not have been sufficient for transfer of if sexual transmission is inefficient. In the work on hamsters, the Borrelia authors caution against extrapolating the results from hamsters to the wild rodents that are the natural reservoir species for . Presumably extrapolation of result to humans is similarly problematic. Borrelia If the conflicting results on contact and sexual transmission reflect species-specific differences in the immune response to , and consequently the duration and location of infection, this raises the issue Borrelia of whether the results of any animal model can be extrapolated to humans. Any animal model in which most individuals are asymptomatic or subclinical and/or spontaneously clear the infection are probably not going to be a good model for transmission studies in humans as both the duration of infection and tissue/organ distribution of the bacteria will be different from that in humans. This makes the Middelveen et indicated in my previous comments, there is ample evidence that this is not likely to occur. My earlier reference to Koch's postulates is pertinent and quite appropriate in that context. Furthermore, since these authors and their colleagues have a history of publishing results suggesting extensive laboratory contamination by species , these findings ought not to be accepted for Borrelia indexation without independent confirmation by a third party using blinded specimens, even if they might be inconvenienced by such a reasonable request under such circumstances. We are sorry that Phillip Baker is upset by the inconvenient truth presented in our study showing that live, culturable Borrelia spirochetes are present in semen and vaginal secretions from patients with Lyme disease. As pointed out in a previous comment, his reference to Koch's postulates is inappropriate because according to Wikipedia these criteria for infection are largely obsolete in the age of molecular biology . As for the epidemiology of Lyme disease, readers should judge for themselves whether the number of infected patients makes no difference in how we view the disease. The trivial illness that Baker and other "Lyme denialists" cling to has disappeared, and we need to accept that Lyme disease is a major epidemic no matter how inconvenient that truth may be .

References
As for Baker's suggestion that Borrelia cannot survive the vaginal pH of 3.8-4.5, there are several major flaws in that argument: First, according to Wikipedia the vaginal pH in pre-menopausal women is in that range, but in post-menopausal women the pH increases to 6.5-7.0, which would be more permissive for spirochete survival . Second, seminal plasma and vaginal mucus contain an alkaline medium that protects sperm and may also protect the spirochetal agents of syphilis and Lyme disease in the vagina . Third, Borrelia is known to form cysts that can survive in extreme conditions, including low pH, and then revert to the spiral form in more favorable circumstances . Fourth, as stated in our article the vaginal mucosa produces immune factors that would allow Borrelia to survive once it reaches that location. Thus Borrelia has ample means to survive in the "hostile" vaginal environment, and this important topic merits further study. Baker's commentary on our article uses selective information that supports his point of view while ignoring contradictory evidence. He starts with two animal models that appear to refute contact or sexual transmission of Lyme disease, and he concludes with selective epidemiological evidence from the CDC that obliquely supports his view. Sandwiched in between is a flawed discussion of maternal-fetal transmission of Lyme disease that has no bearing on sexual transmission of the spirochetal infection.
Although we have addressed the animal model and epidemiological issues in our article, we will address these issues in greater detail below. We will leave it to readers to find the flaws in Baker's denial of maternal-fetal transmission of Lyme disease, since this form of transmission has been documented in canine, bovine, murine and human reports from the peer-reviewed medical literature .
Baker cites two animal studies as proof that Lyme disease cannot be sexually transmitted . The study by Moody and Barthold was performed using Lewis rats, while the study by Woodrum and Oliver was performed using Syrian hamsters. In limiting his analysis to these two studies, Baker ignores the compelling observations in mice, birds, cats and dogs that support contact or sexual transmission of Lyme disease without a tick vector, as outlined in our report . The Lewis rat and Syrian hamster studies had significant flaws. First, these rodents are highly inbred with all Lewis rats descended from a single breeding colony, and all Syrian hamsters descended from a single female breeder . The effect of inbreeding on contact transmission of is unknown. Second, these rodents are poor models for human Lyme Borrelia disease. The Lewis rat develops a transient arthritis that fades without treatment following intraperitoneal injection of , while the Syrian hamster may be colonized by for long periods without Borrelia Borrelia showing any signs of disease . Since the disease pattern in these animals differs significantly from the human pattern, it is not surprising that transmission of Borrelia would differ as well. In fact, the immunocompetent Syrian hamster has been abandoned as a model for human Lyme disease because of these differences. Third and perhaps most important, both studies used limited culture and immunological methods to determine sexual transmission of , and more sensitive detection of spirochetes by Borrelia molecular techniques might have yielded positive results, as shown in other rodent experiments using PCR . Thus the selective animal models cited by Baker fail to refute the possibility of sexual transmission of Lyme disease in humans.
Baker attempts to use CDC surveillance statistics to further disprove sexual transmission of Lyme disease. He points to the fact that 95% of reported CDC surveillance cases occur in 12 states in the USA, and that these reported cases are seasonal and have a male and child predominance. As pointed out in our article, however, CDC surveillance cases account for less than 10% of total Lyme cases (about 30,000 cases out of more than 300,000 annual cases in the USA). Thus the epidemiology of the vast majority of Lyme cases is not covered by CDC surveillance reporting and may differ significantly from the epidemiology reported by the CDC. To support this view, as noted in our study, a recent report from Australia found that the prevalence of Lyme disease cases aligned with the sexually active population on that continent . Baker also questions whether our Lyme disease testing was accurate since it differs from the serological parameters used for CDC surveillance criteria. However the CDC surveillance criteria rely on testing that misses more than half of Lyme disease cases , and these criteria are not recommended for diagnosis of Lyme disease . The reason why Baker thinks that insensitive and incomplete epidemiology statistics trump microscopy, immunochemistry, molecular hybridization and PCR testing that demonstrates live, culturable spirochetes in genital secretions from our Lyme disease patients is unclear. Borrelia Baker bemoans the fact that because of the inconvenient evidence presented in our study, Lyme disease patients may suddenly be concerned about sexual transmission of their disease. Because Baker does not represent or treat patients, he may be unaware that this question is being asked by patients in clinicians' represent or treat patients, he may be unaware that this question is being asked by patients in clinicians' offices every day as spouses and sexual partners become symptomatic. Should Lyme disease patients be concerned about sexual transmission? Absolutely. We currently treat Lyme disease with 60-year-old antibiotics that are marginally effective against a chronic infectious disease . Baker and his research associates should focus on the risk of sexual transmission of and strive to develop new treatments Borrelia for this alarming epidemic infection along the lines of the global "Manhattan Project" that has successfully dealt with the HIV/AIDS epidemic . Our suffering Lyme disease patients deserve nothing less. ehrlichiosis in an Australian cohort. The concept of sexual transmission of borreliosis, which has been resurrected recently by Middelveen et , was refuted years ago by the well-designed and controlled studies of Moody and Barthold , as well al. as Woodrum and Oliver , internationally known experts on Lyme disease. These investigators used well-characterized animal models of borreliosis in which infection is much more disseminated and profound than it ever is in humans. It should be noted that, in the United States, Lyme borreliosis has historically been defined as a tick borne infection caused by .

Borrelia burgdorferi sensu lato
To determine if borreliosis can be transmitted by direct contact, Moody and Barthold housed three-day-old -or three-week-old-Lewis rats, deliberately infected with , with normal, B. burgdorferi uninfected rats for 30 days. As expected, all deliberately infected rats continued to be actively infected, 30 days later; however, none of the uninfected rats acquired infection after 30 days of intimate direct contact with their infected house mates.
In other experiments, Moody and Barthold were unable to demonstrate venereal transmission of borreliosis from seven infected females-or six infected males -to uninfected rats of the opposite sex.
In the work of Woodrum and Oliver , six female Syrian hamsters infected with were mated B.burgdorferi with six uninfected males; conversely, three infected males were mated with six uninfected females. None of the uninfected hamsters became infected after mating with an infected partner of the opposite sex, indicating that borreliosis is not sexually transmitted. These investigators also failed to demonstrate contact transmission of between infected female -or male-hamsters and uninfected B. burgdorferi hamsters of the opposite sex. Also, it was not possible to transmit borreliosis to uninfected hamsters with urine or feces from infected hamsters.
Sadly, preliminary oral reports of the observations of Middelveen have already generated an et al. inordinate amount of fear and anxiety within the lay community due to sensationalized reports of their unconfirmed findings by an uncritical -and often naïve -press. This has already caused much harm. To date, I have received numerous inquiries from distraught individuals, wondering if they now should even consider marrying their spouse-to-be for fear of contracting Lyme disease that some mistakenly believe to be incurable. Some fear the possibility of giving birth to an infected or congenitally deformed child, because their spouse or spouse-to-be had been diagnosed and treated for Lyme disease in the past. To examine the issue of transmission of infection, Moody and Barthold inoculated pregnant in utero female Lewis rats with viable , at four days of gestation. All of the inoculated pregnant B. burgdorferi females became seropositive as expected, and could be cultured from their spleens at 20 B. burgdorferi days of gestation; however, their placentas and fetuses were culture negative, indicting the lack of in utero transmission.
Moody and Barthold used two different experimental protocols to determine if transplacental transmission of occurs. One protocol involved six non-pregnant infected females that were B. burgdorferi subsequently mated and became pregnant. Three of the females were allowed to carry to full term, whereas the remaining three were sacrificed just prior to parturition. All offspring and offspring-to-be were found to be culture negative for , as well as seronegative for antibody specific for B. B. burgdorferi , indicating that transplacental transmission of infection does not occur. burgdorferi In the second protocol, six females were infected via tick bite after becoming pregnant, and were allowed to carry their fetuses to birth; all were negative for infection. The results of these studies likewise failed to provide evidence for the transplacental transmission of naturally acquired borreliosis.
Other investigators examined the possibility of congenital birth defects in humans with Lyme disease by doing a rather large comparative study involving 5,000 infants, half from an area in which Lyme disease was endemic and half as controls from an area without Lyme disease . They found no significant differences in the overall incidence of congenital malformations between the two groups.
In another study, involving 1,500 subjects including controls, no increased risk of giving birth to a child with a congenital heart defect was noted in women who had either been bitten by a tick or had been treated for Lyme disease during or before pregnancy .
Finally, an extensive analysis of the world literature revealed "that an adverse outcome due to maternal infection with at any point during pregnancy in humans is at most extremely rare" .

B. burgdorferi
In view of the above considerations, the work by Middelveen is not credible and must be view with et al. much skepticism. There is no evidence to indicate that vaginal and seminal secretions provide a suitable environment for the growth of Borrelia and that the mere presence of Borrelia in vaginal and seminal secretions -if such an observation is indeed valid and can be confirmed independently -results in disease. In the absence of such proof, no conclusions can be drawn as to the sexual transmission of Lyme disease by of .

: Addendum added in proof
It should be noted that the basic epidemiology of vector-borne Lyme disease is strikingly different from that for sexually transmitted diseases (STDs). For example, annual data provided by the CDC indicate that about 95% of all reported cases of Lyme disease occur in 12 States, and the incidence of Lyme disease is seasonal and peaks with increased tick activity. Also, in the case of STDs, there are more husband-wife paired cases rather than the preponderance of cases of males and children with Lyme disease. Most important, Middelveen state that "… the majority of our study subjects meet the positive et al.
did not standard as defined by the CDC surveillance criteria." One must therefore wonder if their study subjects ever had Lyme disease in the first place.

References
Middleveen, MJ, Burke, J., Sapi, E., Culture and identification of Borrelia spirochetes in human et al.