|Home | About | Journals | Submit | Contact Us | Français|
Vaginal microbicides with activity towards organisms that cause sexually transmitted infections have been proposed as a strategy to reduce transmission. Small-molecule inhibitors of Chlamydia trachomatis serovar D belonging to the class of salicylidene acylhydrazides (INPs) have been shown to work through a mechanism that involves iron restriction. Expanding on this work, ten INPs were tested against a lymphogranuloma venereum strain of C. trachomatis serovar L2, Neisseria gonorrhoeae, and hydrogen peroxide-producing Lactobacillus crispatus and Lactobacillus jensenii. Seven INPs had minimal inhibitory concentrations (MICs) and minimal bactericidal concentrations of <50 µM towards C. trachomatis L2. Three INPs had an MIC <12.5 µM against N. gonorrhoeae. Inhibition by was reversed by iron, holo-transferrin and holo-lactoferrin but not by the iron-poor forms of these compounds. The compounds exhibited no bactericidal activity toward Lactobacillus. The INPs were not cytotoxic to HeLa 229 cells. When INP 0341 was tested in a mouse model of a Chlamydia vaginal infection there was a significant reduction in the number of mice shedding C. trachomatis up to 4 days after infection (P < 0.01). In summary, select INPs are promising vaginal microbicide candidates as they inhibit the growth of two common sexually transmitted organisms in vitro, are active in a mouse model against C. trachomatis, are not cytotoxic and do not inhibit organisms that compose the normal vaginal flora.
Prevention of sexually transmitted infections (STIs) in the era of the human immunodeficiency virus (HIV) epidemic has taken on new importance [1–6]. As such, multiple strategies in addition to antibiotics and antivirals are needed to address this problem since the context for transmission of STIs varies greatly depending on geographical, economic and social situations. Progress in vaccine development for most STIs is still at the bench level; currently, the only STI vaccine available is for high-risk types of human papillomavirus. Meanwhile, pathogens such as Chlamydia trachomatis and Neisseria gonorrhoeae not only have the potential to cause significant morbidity on their own but have been linked to the increased transmission of HIV [1,2].
One strategy to reduce the transmission of STIs is vaginal microbicides [7,8]. In part this approach has been proposed based on the overall acceptance rate of contraceptive vaginal spermicides. Several agents have been proposed for incorporation into a vaginal gel and some are currently under development [8–12]. We have recently described a salicylidene acylhydrazide, INP 0341, that when tested in vitro attenuates the infectivity of a common genital strain, serovar D, of the intracellular pathogen C. trachomatis .
In this study, we have expanded our investigation of salicylidene acylhydrazides (INPs) against two STIs, a lymphogranuloma venereum (LGV) strain of C. trachomatis (serovar L2) and N. gonorrhoeae, as well as against bacterial organisms that are members of the normal human vaginal flora. In addition, INP 0341 was tested in a mouse model to determine whether this compound could attenuate a Chlamydia genital infection. A compound able to selectively attenuate both C. trachomatis and N. gonorrhoeae while not affecting the normal bacterial flora or vaginal mucosa would be an excellent candidate to formulate and test as a vaginal microbicide.
Chlamydia trachomatis serovar L2 (434), C. trachomatis serovar D (UW-3), N. gonorrhoeae (ATCC 49226 and ATCC 43069), Lactobacillus jensenii strain 6G (ATCC 25258), Lactobacillus crispatus (ATCC 33197), Escherichia coli (ATCC 25922), Enterococcus faecalis (ATCC 29212), Neisseria mucosa (ATCC 19695), Neisseria lactamica (ATCC 23970), Neisseria flava (ATCC 13120) and HeLa 229 cells were obtained from the American Type Culture Collection (Manassas, VA). In addition, a clinical isolate of Neisseria subflava was included in the testing. Neisseria gonorrhoeae strains FA 1090, FA 7163, FA 7034 and FA 6916 were obtained from the laboratory of Dr Fred Sparling (University of North Carolina, Chapel Hill, NC) [14,15].
HeLa cells and Chlamydia stocks were raised as previously described . Neisseria spp. were maintained on chocolate agar (BD Becton Dickinson, Franklin Lakes, NJ), E. coli and E. faecalis on 5% sheep blood agar (BD Becton Dickinson) and Lactobacillus spp. on Difco™ Lactobacilli MRS agar (BD Becton Dickinson). All plates were incubated at 37 °C in 5% CO2 for 48 h. Organisms were subcultured twice prior to being used in an assay.
INPs (Creative Antibiotics, Umeå, Sweden) were dissolved in dimethyl sulphoxide (DMSO) (Fisher Scientific, Fair Lawn, NJ) at a concentration of 25 mM and stored at −20 °C. Ferrous ammonium sulphate (FeSO4) (Sigma-Aldrich, St Louis, MO) was diluted in distilled water and filter-sterilised (0.22 µM) prior to use. Stock solutions (500 nM) of human holo-transferrin and human apo-transferrin (Sigma-Aldrich) were dissolved in sterile water, and human lactoferrin saturated with iron (Sigma-Aldrich) and human apo-lactoferrin (Sigma) were dissolved in sterile phosphate-buffered saline and stored at −20 °C.
HeLa 229 cells were grown to confluency in 1-dram glass shell vials overnight at 37 °C in 5% CO2. Confluent monolayers were infected and stained as previously described . The resulting inclusion-forming units (IFU) were viewed using an Olympus BH2 microscope (Olympus, Center Valley, PA), enumerated and compared with untreated controls.
A lactate dehydrogenase (LDH) assay was used to determine the cytotoxic effects of INPs on HeLa cells. The LDH assay was performed with INP-treated and untreated HeLa cells using the Cytotox-96 kit (Promega, Madison, WI) according to the manufacturer’s instructions. The percentage of LDH leakage was calculated using the formula (LDH activity in medium/total LDH × 100) . Each sample was assayed in triplicate.
Serial dilutions of INPs (1.5–100 µM) in DMSO were added to brain–heart infusion (BHI) broth (BD Becton Dickinson) with and without supplementation with 2% IsoVitaleX (BD Becton Dickinson) in 96-well microtitre plates. A standardised suspension of bacteria from an overnight culture was added to achieve 1 × 105 colony-forming units (CFU)/mL. Plates were incubated at 37 °C in 5% CO2 for 24 h and subsequently the MIC was defined as the lowest concentration of the compound that prevented macroscopic growth.
To determine the MBC99 (the lowest concentration of INP at which the bacterial count was reduced to <99% of the inoculum), aliquots were cultured from wells with no visible growth and the highest INP concentration with visible growth. Neisseria spp. were subcultured on chocolate agar, Lactobacillus spp. on MRS agar, and E. faecalis and E. coli were subcultured on 5% sheep blood agar. All plates were incubated at 37 °C in 5% CO2 for 24–48 h.
The MIC for C. trachomatis L2 was determined by infecting HeLa cell monolayers at a multiplicity of infection of 0.2 IFU with two-fold dilutions of INPs (0.75–50 µM). Following incubation for 36 h, one set of cultures was fixed and stained to determine the MIC and the transitional MIC (MICTP) and one set was sonicated and passed to fresh monolayers to determine the MBC. The MICTP was the concentration of the compound at which abnormal inclusion size was observed for >90% of the inclusions, the MIC was the next highest concentration tested, and the MBC was the lowest concentration of the compound where <99% IFU were observed after passage to HeLa cells compared with control cultures .
Neisseria gonorrhoeae grown in BHI broth with IsoVitaleX was incubated with the INP with and without free iron (Fe2+), iron-rich human holo-transferrin or human holo-lactoferrin and their corresponding iron-poor (apo) forms. These compounds were also added to INP-treated C. trachomatis L2 cultures before centrifugation onto HeLa cell monolayers. Following incubation, cultures were fixed, stained and inclusions were counted as previously described . Neisseria cultures were incubated for 24 h before subculture to chocolate agar. The average number of CFU generated by the INP-treated N. gonorrhoeae cultures with and without an added iron source was compared with cultures incubated in media alone. All assays were performed in triplicate.
Total genomic Chlamydia DNA was isolated from cell cultures at time 0, immediately after centrifugation (early) and 36 h (late) post infection. The Promega Wizard® Genomic DNA Purification Kit (Promega) was used to extract genomic DNA according to the manufacturer’s instructions. DNA samples were precipitated, washed, reconstituted in 10 mM HEPES (pH 8.0), stored at −20 °C and/or used in real-time quantitative polymerase chain reaction (qPCR) assay. An ultraviolet spectrophotometer was used to determine the DNA concentration in each sample.
Plasmid control constructs were made from C. trachomatis L2 ompA, which was amplified using the primers 5’-ATAAGAATGCGGCCGCAATGAAAAAA CTCTTGAAA-3’ and 5’-GACTAGTTTAGAAGCGGAATTGTGCATTTACGTGA-3’. The PCR product and the pTRG vector (Stratagene, La Jolla, CA) were treated with NotI and SpeI enzymes (New England Biolabs, Beverly, MA) prior to ligation and transformation of E. coli DH5α (New England Biolabs). The resulting construct (pTRG-ompA) was propagated in E. coli and was subsequently purified using the Qiagen Miniprep Kit (QIAGEN, Valencia, CA).
A qPCR assay employing the AB Biosystems SYBR Green (PE Applied Biosystems, Warrington, UK) reaction mix and method was used to determine the amount of chlamydial DNA in INP-treated and untreated cultures. A 174-bp segment was amplified from the pTRG-ompA template. A 7-point standard curve was generated by amplifying known amounts of pTRG-ompA templates. Input DNA for all samples was normalised by DNA concentration (ng/µL). Thermal cycling reaction conditions were 95 °C for 2 min, followed by 40 cycles of 95 °C for 15 s, 54 °C for 30 s and 72 °C for 30 s. To determine relative numbers of Chlamydia genomic copies in each sample, threshold cycle (Ct) values between the standard and the test samples were compared . Assays were performed in triplicate.
A mouse model, as previously described, was used to test the ability of INP 0341 to attenuate a genital infection by C. trachomatis serovar D . Seven-to-eight-week-old female C3H/HeJ (H-2k) mice (Jackson Laboratories, Sacramento, CA) received two subcutaneous doses of medroxyprogesterone acetate (SICOR Pharmaceuticals, Inc., Irvine, CA) on Days 10 and 3 before a vaginal challenge with 1 × 103 IFU of C. trachomatis in 0.01 mL of 0.2 M sucrose/0.02 M sodium phosphate (pH 7.2) (SPG). Mice were treated intravaginally 2 days prior to challenge, 1 h before, with the challenge dose of Chlamydia, 4 h and 8 h after challenge, and daily up to 5 days after vaginal challenge with 0.02 mL of 1 mM INP 0341 dissolved in 20% DMSO and diluted in SPG. Control mice were treated the same but received SPG with DMSO without INP 0341. Vaginal swabs were collected 4 days after infection and were cultured, stained and evaluated as previously described . The experiment was repeated three times with 5 to 15 mice per experimental group.
Fisher’s exact test was used to determine differences between mouse groups regarding the number of mice infected. Mann–Whitney U-test was performed to evaluate the number of IFU recovered from two groups of mice. In vitro data were analysed by Student’s t-test and Mann–Whitney U-test using SigmaStat 3.5 software (SYSTAT Software, Inc., Richmond, CA). A P-value of <0.05 was considered significant.
The effect of ten INPs at a concentration of 25 µM against C. trachomatis L2 could be grouped by light microscopy into those that gave maximum, moderate or no appreciable inhibition (Fig. 1). INPs 0341, 0149, 0161, 0029 and 0007 were the most inhibitory, yielding either no inclusions or inclusions that were difficult to see by light microscopy owing to their abnormally small size. Cultures incubated in the presence of INPs 0400 and 0269 resulted in Chlamydia inclusions with great heterogeneity in size and appearance. INPs 0328, 0143 and 0406 did not appear to inhibit Chlamydia growth compared with control cultures.
To quantitate the inhibitory effect of the compounds, copies of C. trachomatis L2 ompA in 36 h cultures were measured by qPCR. The yield of Chlamydia DNA corresponded to the appearance of the L2 inclusions (Fig. 1). INPs giving maximal inhibition yielded from 0.6% to 1.7% (log10 3.25 to 3.67 vs. 5.45) of Chlamydia DNA compared with control cultures (P < 0.05). INPs 0400 and 0269 with moderate inhibition yielded Chlamydia DNA that was 4.3% to 5.1% (log10 4.07 to 4.15 vs. 5.45) compared with control cultures (P < 0.05). INPs 0328, 0143 and 0406 yielded Chlamydia DNA amounts similar to control cultures (P > 0.05).
The MICTP, MIC and MBC for the INPs correlated with light microscopy and qPCR performed with 25 µM INP (Fig. 1). Corroborating the qPCR data, the two compounds with the lowest MIC/MBC were INP 0341 and INP 0149, whilst INPs 0143, 0328 and 0406 exhibited no inhibitory activity.
The MICs/MBCs of the INPs were established for N. gonorrhoeae, commensal Neisseria spp., hydrogen peroxide-producing L. jensenii and L. crispatus, E. faecalis and E. coli. No inhibition of growth of N. gonorrhoeae was seen with INPs 0328, 0400 and 0406. In contrast, INPs 0007, 0029 and 0341, which also exhibited good activity against C. trachomatis, gave the lowest MIC for N. gonorrhoeae at 12.5 µM. With few exceptions, the MBC and MIC values were the same, suggesting a bactericidal effect of the active compounds towards N. gonorrhoeae. In contrast, the commensal Neisseria strains tested were less affected by the compounds, with MICs one- to two-fold higher than N. gonorrhoeae.
None of the INPs had bactericidal activity up to 100 µM against either L. jensenii or L. crispatus, two of the most common hydrogen peroxide-producing members of the vaginal normal flora. Escherichia coli was not inhibited by any of the compounds tested. Enterococcus faecalis was inhibited by INP 0007, 0341 and 0400, with MICs ranging from 12.5 µM to 50 µM. With these three compounds, the MIC and MBC values were the same or within one dilution, suggesting a bactericidal effect against E. faecalis.
HeLa cells infected with C. trachomatis L2 were incubated for 36 h with and without INP 0341 (25 µM) along with the iron-rich holo-form of lactoferrin and transferrin and their corresponding iron-poor apo-forms as well as iron supplied as Fe2+. Control cultures without INP 0341 but with added Fe2+, holo- and apo-lactoferrin or holo- and apo-transferrin appeared similar to control cultures, therefore none of the added iron sources alone had an effect on the number or appearance of Chlamydia inclusions. Normal inclusions were not seen in cultures incubated with 25 µM INP 0341 compared with the control. However, when holo-transferrin, holo-lactoferrin or Fe2+ was added, the inhibitory effect of INP 0341 was abrogated. In contrast, there was no reversal of the inhibitory effect when cultures containing INP 0341 were incubated with apo-transferrin or apo-lactoferrin.
To investigate whether iron-binding compounds were able to reverse the inhibitory effect of INP 0341 on N. gonorrhoeae, in addition to the two ATCC strains that contained the full complement of lactoferrin and transferrin genes, transferrin and lactoferrin mutants were obtained from Dr Fred Sparling. One of these strains (FA 1090) is a human isolate that was found to lack lactoferrin receptor protein expression but expressed the transferrin receptor proteins . The other three strains have been constructed in vitro to either possess the full complement of lactoferrin and transferrin genes [FA 7163 (lactoferrin+/transferrin+)] or to lack one or both of the lactoferrin and/or transferrin genes [FA 6916 (lactoferrin−/transferrin−) and FA 7034 (lactoferrin+/transferrin−)] . Unlike the ATCC strains, the laboratory-constructed strains did not grow well in BHI broth alone therefore the MICs/MBCs of all strains were compared using BHI broth supplemented with 2% IsoVitaleX (Table 1). There was no significant difference in the MIC/MBC whether or not the strains possessed genes for the lactoferrin or transferrin receptor proteins. The general pattern of activity of the ten INP compounds was the same for all N. gonorrhoeae strains, with INPs 0007, 0029 and 0341 being the most inhibitory, similar to that seen with C. trachomatis.
The same iron-carrier proteins used to test for reversal of inhibition of C. trachomatis by INP 0341 were used in a bactericidal assay with N. gonorrhoeae (Fig. 2). Here, 1 ×105 CFU/mL in BHI broth supplemented with IsoVitaleX was incubated overnight with and without INP 0341 (6.25 µM). With the three strains of N. gonorrhoeae that had the full complement of lactoferrin and transferrin genes (ATCC 49226, ATCC 43069 and FA 7163), free iron supplied as FeSO4, holo-lactoferrin and holo-transferrin, but not apo-lactoferrin and apo-transferrin, were able to reverse the inhibition by INP 0341. In contrast, these iron sources were not able to reverse the inhibition in laboratory-constructed strains lacking one or more of the lactoferrin and transferrin receptor genes.
LDH measurement of INP-treated cells incubated for up to 48 h showed no significant membrane damage as determined by the lack of leakage of LDH into the supernatant media (P > 0.05) (data not shown).
Mice were treated intravaginally with 1 mM INP 0341 prior to, at the time of and after infection with 1 × 103 IFU of C. trachomatis serovar D. Control mice were treated the same as the INP-treated group minus INP 0341. There was a significant difference 4 days after infection between the two groups in terms of the number of mice infected as judged by a vaginal culture (P = 0.013). Of the control mice, 17/20 (85%) were culture-positive, in contrast to the INP-treated mice where 12/25 (48%) were culture-positive 4 days after infection. There was also a significant difference in the number of IFU recovered 4 days after infection (Fig. 3) (P < 0.01).
STIs remain a worldwide problem especially in light of their association with increased HIV acquisition [1–6,8]. Vaginal microbicides have been proposed as an approach to reduce the spread of STIs [7–9], an attractive concept since it would afford women the ability to protect themselves . A limited number of formulations of vaginal microbicides have been investigated in clinical trials, but in general results have been disappointing [22,23]. Some of this is due to the fact that they were shown to damage the vaginal epithelium resulting in inflammation and ulceration that ultimately led to an increase in STIs. The work presented here lays the foundation for the development of a vaginal microbicide based on local iron withholding as a strategy to selectively attenuate pathogenic bacterial growth. The ability of pathogenic organisms to acquire iron is essential to their survival . However, the human host keeps iron tightly regulated and bound to limit the free supply available. We have explored the concept of introducing iron-sequestering compounds into this ‘tug-a-war’ between host and pathogen in order to favour the host and thus limit the infectivity of the pathogen.
The INPs used in this report were originally described to interfere with the type three secretion system (T3S) of several virulent Gram-negative bacilli including Yersinia and Chlamydia [13,25–27]. In an attempt to dissect the mechanism for this apparent effect on the Chlamydia T3S, we explored their role as a chelating agent. Using different metal ions to reverse the inhibition of the genital pathogen C. trachomatis serovar D, we found this effect to be specifically reversed by iron . Any effect these compounds had on host cell proliferation was also reversed by iron, leading us to hypothesise that these compounds were able to indirectly limit Chlamydia growth by restricting available iron within the host. Chlamydia, like most intracellular pathogens, is dependent upon host cell iron supplies [28,29].
With the ultimate goal of developing a vaginal microbicide, we wanted to expand our findings to determine whether other genital bacterial pathogens would be inhibited by these compounds. LGV serovars of C. trachomatis are more aggressive and invasive than the other genital serovars of C. trachomatis. LGV strains are prevalent in underdeveloped nations, many of which also have a high prevalence of HIV. However, over the last decade there have been several reports of increasing numbers of LGV from countries in Western Europe and North America [30,31]. The results obtained with the INPs for C. trachomatis serovar L2 mirrored those that we had previously reported for serovar D. We also wanted to determine whether lactoferrin could compete with INPs since it is in high concentration in the vaginal tract and a key protein in maintaining iron homeostasis. Apo-lactoferrin did not interfere or compete with the inhibitory properties of the INP; only when it was saturated with iron in its holo form did it reverse the INP effect.
Neisseria gonorrhoeae was also inhibited by the INPs and in general those compounds that were effective against C. trachomatis also had activity toward N. gonorrhoeae. This pathogen also requires iron for growth and has multiple mechanisms for acquiring iron, including high-affinity receptors for both transferrin and lactoferrin [14,32,33]. We attempted to determine whether natural and laboratory strains lacking lactoferrin or transferrin receptors were inhibited by the INPs and, if so, could the inhibition be reversed by the iron-rich form of either lactoferrin or transferrin. Our results showed that INP inhibition could be reversed by an iron source with strains possessing the full complement of lactoferrin and transferrin receptor genes, but was not reversed with strains lacking one or both set of genes coding for these receptors. The reason for this is not clear, but resolving this may add to our understanding of the complex iron acquisition systems in N. gonorrhoeae that have both unique features but also share some common pathways within the bacterial membrane. Results with N. gonorrhoeae corroborated those obtained with Chlamydia, suggesting that iron withholding is one, if not the main, mechanism by which the INPs attenuate the growth of these STI agents.
When developing a vaginal microbicide, candidate compounds should have minimal activity toward organisms that comprise the normal vaginal flora [34,35]. The hydrogen peroxide-producing lactobacilli have been shown to be part of the innate vaginal defence, being essential in maintaining homeostasis of the vaginal tract . Since lactobacilli are somewhat unique in that they do not require iron for growth but instead rely on manganese, we reasoned that INPs should not have appreciable activity against these organisms . To test this hypothesis, both L. crispatus and L. jensenii, two of the lactobacilli most frequently recovered from the normal human vaginal tract [34,35], were employed. None of the compounds tested had bactericidal activity towards L. crispatus and L. jensenii. Therefore, whilst the compounds were able to kill the genital pathogens, they had minimal to no effect on organisms considered essential for the maintenance of pH and bacterial balance in the vagina.
In the vaginal milieu many factors may compromise the activity of the INPs. Among them are the vaginal secretions, pH, lactoferrin and blood cells. In addition, a microbicide needs to be carefully formulated for ease of delivery and retention of activity. To begin to address the in vivo activity of the INP, a genital mouse model of a Chlamydia infection was employed. Here it was possible to demonstrate that the group of mice that were INP-treated had a significantly lower rate of infection compared with the control untreated group. These experiments demonstrate the proof of principle that INP 0341 can attenuate C. trachomatis infection in vivo.
In conclusion, the in vitro results presented here demonstrate that select INPs, in particular 0029 and 0341, inhibited both C. trachomatis and N. gonorrhoeae, had minimal to no effect towards organisms considered to be part of the normal vaginal flora, were not cytotoxic to HeLa cells and were active in vivo. These findings suggest that they are good candidates to consider further for formulation and continued testing as a vaginal microbicide.
The authors wish to thank Dr Fred Sparling (University of North Carolina, Chapel Hill, NC) for supplying strains of N. gonorrhoeae used in this investigation.
This work was supported by NIH/NIAID grants AI-71104 and AI-0079775 (both to EMP).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.