Robust Vaginal Colonization of Macaques with a Novel Vaginally Disintegrating Tablet Containing a Live Biotherapeutic Product to Prevent HIV Infection in Women

MucoCept is a biotherapeutic for prevention of HIV-1 infection in women and contains a human, vaginal Lactobacillus jensenii that has been genetically enhanced to express the HIV-1 entry inhibitor, modified cyanovirin-N (mCV-N). The objective of this study was to develop a solid vaginal dosage form that supports sustained vaginal colonization of the MucoCept Lactobacillus at levels previously shown, with freshly prepared cultures, to protect macaques from SHIV infection and to test this formulation in a macaque vaginal colonization model. Vaginally disintegrating tablets were prepared by lyophilizing the formulated bacteria in tablet-shaped molds, then packaging in foil pouches with desiccant. Disintegration time, potency and stability of the tablets were assessed. For colonization, non-synchronized macaques were dosed vaginally with either one tablet or five tablets delivered over five days. Vaginal samples were obtained at three, 14, and 21 days post-dosing and cultured to determine Lactobacillus colonization levels. To confirm identity of the MucoCept Lactobacillus strain, genomic DNA was extracted from samples on days 14 and 21 and a strain-specific PCR was performed. Supernatants from bacteria were tested for the presence of the mCV-N protein by Western blot. The tablets were easy to handle, disintegrated within two minutes, potent (5.7x1011 CFU/g), and stable at 4°C and 25°C. Vaginal administration of the tablets to macaques resulted in colonization of the MucoCept Lactobacillus in 66% of macaques at 14 days post-dosing and 83% after 21 days. There was no significant difference in colonization levels for the one or five tablet dosing regimens (p=0.88 Day 14, p=0.99 Day 21). Strain-specific PCR confirmed the presence of the bacteria even in culture-negative macaques. Finally, the presence of mCV-N protein was confirmed by Western blot analysis using a specific anti-mCV-N antibody.


Introduction
Human mucosal surfaces such as the non-keratinized epithelium of the vagina are colonized by commensal flora [1,11,21,50]. The vagina, together with its microflora, constitutes a dynamic ecosystem with important host defense mechanisms that promote reproductive health in women. In healthy women of childbearing age, the vaginal flora is dominated by lactobacilli (10 7 -10 9 CFU per gram of fluid) [41]. These facultative anaerobes metabolize glucose to lactic acid, contributing to the maintenance of a low vaginal pH (3.6-4.5) that accounts for a major part of the nonspecific defense to reduce the risk of acquiring sexually transmitted diseases [5]. Depletion of vaginal lactobacilli is associated with establishment of opportunistic bacterial infections [14] and an increased risk of acquiring human immunodeficiency virus (HIV) and herpes simplex virus type 2 (HSV-2) [9,10,30,40,43]. Consequently, there has been considerable interest in exploring an ecologic approach to re-populate the vagina with lactobacilli, termed Lactobacillus replacement therapy. One unique biologic approach being pursued is the use of a 'live' recombinant human vaginal Lactobacillus as a self-renewable topical microbicide against vaginal HIV transmission [8,27,28].
Some smaller animals, including mice, rats, dogs, hamsters, and guinea pigs are reported to harbor low numbers of vaginal lactobacilli [34,38]. Apparently, lactobacilli are not the dominant species in the vaginal flora as in the case of reproductive-age women. In addition, these smaller animals have an estrous cycle, so their reproductive physiology as well as anatomy is also very different from that of humans [34,38]. Thus, selection of a relevant animal model that is closely related to humans and sensitive to mucosal HIV infection is crucial. We sought to establish a nonhuman primate model that could more accurately mimic the human female lower reproductive tract. This model would help to test colonization, persistence, efficacy, and safety of a 'live' topical microbicide against mucosal transmission of pathogens in the vagina.
Among nonhuman primates, the rhesus macaque (Macaca mulatta) reproductive system shares many similarities to the human reproductive system. They have been widely used in simian immunodeficiency virus and simian-human immunodeficiency virus vaginal challenge models [19,29] to test various candidate microbicides [26,45,[47][48][49]. In order to develop a rhesus macaque model for evaluating Lactobacillus-based microbicides, we analyzed the cultivable vaginal flora over one menstrual cycle of 13 Chinese rhesus macaques that were housed at two separate facilities. We recovered abundant endogenous vaginal Lactobacillus species in 12 of 13 animals and identified Lactobacillus johnsonii as the predominant species in these macaques. Moreover, vaginal colonization of a human vaginal Lactobacillus strain was established in these animals. This is the first report of persistent vaginal colonization of human Lactobacillus in the Chinese rhesus macaque. These findings demonstrate that the female Chinese rhesus macaque is a potentially useful model to test vaginal Lactobacillus replacement therapy for prevention of urogenital infections and to support the pre-clinical evaluation of Lactobacillus-based topical microbicides.

Experimental animals
Thirteen sexually mature, non-hormonal-treated female Chinese-origin rhesus macaques, ranging in age from 4 to 10 years, were individually housed at two separate AAALAC International accredited facilities in Maryland. Specifically, five monkeys were housed at Southern Research Institute and eight monkeys were housed

Sampling
Prior to sample collections or intravaginal administration of vaginal suppository of azithromycin or lactobacilli, the animals were sedated with an injectable anesthetic, ketamine hydrochloride, at a dose of 10 mg/kg. Once sedated, the animals were removed from their cage and taken to a central procedure room. Vaginal and rectal cultures were obtained with polyester-tipped swabs (Becton-Dickinson, Cockeysville, MD, USA) and transported overnight to the Microbiology Core Laboratory at Osel Inc., CA in Port-A-Cul anaerobic transport tubes (Becton-Dickinson), which preserve the viability of anaerobic and aerobic bacterial specimens [4]. Approximately 130-150 mg of vaginal fluid (containing some sloughed-off vaginal epithelial cells) was obtained with the swabs. Vaginal fluid was smeared immediately onto pH-indicator strips (EMD Chemicals Inc., Darmstadt, Germany) to estimate vaginal pH values. They typically varied from pH 4 to 8 in Chinese rhesus macaques undergoing normal menstrual cycles.

Microbiology
The microflora was analyzed by standard culturedependent assays, with modifications of previously described methods [35]. Briefly, the vaginal and rectal swabs were removed from the transport tubes and serially diluted in saline buffer in an anaerobic chamber, Bactron IV (Sheldon Manufacturing, Cornelius, OR, USA). Serial dilutions of the bacterial samples were then plated on different agar plates and incubated anaerobically at 37°C with 5% hydrogen, 5% CO 2 , and 90% nitrogen for 4-5 days, or aerobically at 37°C with 5% CO 2 for 24-48 hours. The media used for anaerobic culture were Brucella blood agar, phenylethyl alcohol (PEA) agar with 5% sheep blood, laked blood with kanamycin and vancomycin agar, bacteroides bile esculin agar (Anaerobe Systems, Morgan Hill, CA, USA), and Mann Rogosa Sharpe agar (MRS) (EMD Chemicals Inc., Gibbstown, NJ, USA). The media used for aerobic culture were tryptic soy agar with 5% sheep blood, PEA with 5% sheep blood, MacConkey agar, Candida isolation agar (Hardy Diagnostics, Santa Maria, CA, USA), MRS agar and Rogosa agar (Becton, Dickinson and Co., Franklin Lakes, NJ, USA). Cultures on the primary plates were screened, semi-quantitated, and sub-cultured for isolation of pure cultures. They were then examined for pigment production and hemolysis of blood agar. Additional Gram stain and biochemical tests (e.g. catalase, oxidase, urease and indole production) were also performed for identification of the bacterial isolates.
Bacterial identification by 16S ribosomal RNA (rRNA) gene sequencing Bacterial identification to the species level was performed using 16S rRNA gene sequencing. A single colony of bacteria was lysed in 100 ll of 'PrepMan Ultra' Sample Preparation Reagent (Applied Biosystems, Foster City, CA, USA) to prepare a DNA template for polymerase chain reaction (PCR) amplification. 16S rDNA fragments (900 base pairs) were amplified with universal bacterial primers 8f (5¢-AGA GTT TGA TCC TGG CTC AG-3¢) and 926r (5¢-CCG TCA ATT CCT TTR AGT TT-3¢, R = A/G) [11], purified with ExoSAP-IT (USB, Cleveland, OH, USA), and sequenced with a primer 519r (5¢-GWA TTA CCG CGG CKG CTG-3¢, W = A/T, K = G/T) [32]. Resulting sequences were subjected to nucleotidenucleotide BLAST (blastn) in comparison to known 16S rRNA genes in the public databases to identify the species of Lactobacillus. A species was assigned to an isolate when it shared 98% or higher identity to known genes. For construction of a phylogenetic tree, full-length 16S rRNA genes of Lactobacillus were PCR amplified using primers 8f and 1492r (5¢-TAC GGY TAC CTT GTT ACG ACT T-3¢, Y = C/T) [21]. The resulting PCR products were cloned into pGEM-T Easy Vector System (Promega, Madison, WI, USA) and sequenced with universal primers SP6 and T7.
Assay for detection of hydrogen peroxide production by Lactobacillus isolates Hydrogen peroxide (H 2 O 2 ) production by Lactobacillus strains was tested on MRS agar supplemented with tetramethylbenzidine and horseradish peroxidase (Sigma-Aldrich, St. Louis, MO, USA) [42]. Plates were incubated anaerobically at 37°C for 24 hours and then exposed to ambient air at room temperature. Colonies were observed for color development (from white to blue, indicating H 2 O 2 production) for 30 min. H 2 O 2 production was ranked as +++, ++, and +, if blue color developed in less than 10, 20, and 30 min, respectively. H 2 O 2 production was ranked as negative if colonies remained white within 30 min.

Carbohydrate fermentation by Lactobacillus isolates
Phenotypic diversity among Lactobacillus isolates to ferment carbohydrates was evaluated using API 50 CH carbohydrate fermentation strips (bioMerieux, Inc., Marcy-I'Etoile, France) [6]. Pure Lactobacillus colonies that were grown anaerobically on MRS agar plates for 24 hours were harvested with a sterile swab, and resuspended in an API 50 CHL medium to turbidity equal to or greater than 0.5 McFarland. Strips were incubated anaerobically for 24-48 hours and results were recorded.

Repetitive extragenic palindromic (Rep)-PCR
Chromosomal DNA of Lactobacillus culture at early stationary phase was isolated by using the DNeasy Blood and Tissue kit (Qiagen, Valencia, CA, USA). An aliquot of chromosomal DNA was used for Rep-PCR using primers Rep-R (5¢-III NCG NCG NCA TCN GGC-3¢, I = inosine, N = C/G/A/T) and Rep-L (5¢-NCG NCT TAT CNG GCC TAC-3¢, N = C/G/ A/T) in the PCR mixtures [2]. PCR was performed with a PTC-200 model thermal cycler (MJ Research, Waltham, MA, USA). The PCR reaction was initiated by incubating the reaction mixture at 95°C for 7 min to activate the Taq DNA polymerase (Invitrogen, Carlsbad, CA, USA), followed by 35 cycles of 90°C for 30 s, 40°C for 1 min, and 65°C for 8 min. The reaction was terminated with an extension step of 65°C for 16 min. An aliquot of each reaction mixture was resolved on a 1% agarose gel at 30 V for 18 hours.

In vitro testing of antibiotic susceptibility of vaginal Lactobacillus isolates
A broth microdilution method [33] was used to test the minimum inhibitory concentrations (MICs) of antibiotics for vaginal Lactobacillus isolates.

Clearance of vaginal Lactobacillus by vaginal suppository of azithromycin
Fatty acid-based suppositories containing 200 mg of azithromycin were compounded by Foer's Pharmacy (Bethesda, MD, USA) and administered intravaginally once a day for five consecutive days by digital manipulation to sedated animals.
Intravaginal administration of L. johnsonii and human vaginal isolate of L. jensenii to Chinese rhesus macaques Endogenous vaginal isolates of L. johnsonii isolated from two Chinese rhesus macaques and a human vaginal isolate of L. jensenii 1153 [28] were grown overnight to late log/early stationary phase in MRS broth. The bacterial cells were centrifuged and washed with PBS (pH 7.0). Approximately 10 9 CFU of Lactobacillus cells were mixed in 1 ml of fresh MRS medium and 2 ml of 2.7% hydroxyethyl cellulose (HEC) solution [44], and inoculated intravaginally into the sedated macaques for 7 consecutive days for L. jensenii 1153 [28] or every other day for three times for L. johnsonii recovered from rhesus macaques. Sedated animals were closely monitored by veterinary staff.

Lactobacillus was isolated among endogenous vaginal microflora of Chinese rhesus macaques
The vaginal microflora of nine naı¨ve Chinese rhesus macaques that have not been treated with antibiotics was analyzed from samples collected at multiple time points over one menstrual cycle (36-46 days). Cultivable bacterial species isolated from these animals over the course of the studies are listed in Table 1. The most common bacterial species isolated were Enterococcus and Staphylococcus species (100%). Surprisingly, Lactobacillus was the next most common bacterial species isolated from these animals (89%). Six out of nine animals harbored H 2 O 2 -producing Lactobacillus species, while four out of nine animals harbored non-H 2 O 2 -producing Lactobacillus species. Among the anaerobes, Gram-positive cocci Peptoniphilus sp., Peptostreptococcus sp., and Anaerococcus sp., and Gram-negative rods Bacteroides sp. and Porphyromonas sp. were the most frequently and the most prevalent ones isolated. Yeast was not isolated from any of the animals.

Identification and characterization of endogenous vaginal Lactobacillus in Chinese rhesus macaques
Lactobacillus isolated from a total of 13 Chinese rhesus macaques housed at two separate locations was analyzed. Analysis of 16S rRNA gene sequences determined that the predominant Lactobacillus species recovered from the vagina of 12 of 13 macaques is L. johnsonii (Table 2). Endogenous vaginal Lactobacillus was not recovered from one of the animals (ABL-5) over the course of the study. A majority of the macaques only harbored one vaginal Lactobacillus species, but a few of them harbored more than one. The level of cultivable lactobacilli ranged from 10 2 to 10 8 CFU per vaginal swab. Two isolates of L. johnsonii that had different colony morphology on the MRS agar plates were recovered from one macaque, SRI-3. When carbohydrate fermentation profiles were compared between these two vaginal L. johnsonii isolates, differences were noted in the fermentation of d-lactose, d-raffinose and d-tagatose, demonstrating that they are different strains (Table 2). Other vaginal Lactobacillus species were also recovered from three macaques which harbored L. johnsonii. Lactobacillus murinus was isolated from ABL-6, L. amylovorous was isolated from ABL-8, and L. acidophilus and L. amylovorous were recovered from ABL-7 (Table 2). However, these other Lactobacillus species constituted the minority, as they were recovered less consistently (at one time point only) than L. johnsonii.
Most of the endogenous lactobacilli recovered from Chinese rhesus macaques are strong H 2 O 2 producers ( Rectal swabs were also collected from the above-mentioned nine naı¨ve macaques (ABL 5-8 and SRI 1-5), and endogenous lactobacilli were recovered from all samples analyzed, including ABL-5 that did not harbor vaginal lactobacilli at any sampling times. Lactobacillus johnsonii was found in seven of the nine animals, but it is not the predominant Lactobacillus species in the rectum as in the case of the vagina. In the rectum, many different Lactobacillus species were recovered, including Lactobacillus reuteri, Lactobacillus animalis, Lactobacillus ingluviei, Lactobacillus salivarius, Lactobacillus sobrius and Lactobacillus mucosae. In most instances, more than one species of rectal lactobacilli was recovered from each individual macaque. Seven of the nine macaques harbored the same Lactobacillus species in both the vagina and the rectum ( Table 2), but it was subsequently shown that they have different biochemical and genetic profiles. Additional biochemical tests and chromosomal DNA fingerprinting conclusively determined that the rectal and vaginal Lactobacillus isolates from the same animal are uniquely different from each other ( Table 2 and Fig. 1). All L. johnsonii, of both vaginal and rectal origins, could ferment d-glucose, d-fructose, d-maltose, and sucrose. However, a few rectal L. johnsonii isolates (ABL-R6A, ABL-R6B and SRI-R3) could not utilize several sugars (d-galactose, N-acetylglucosamine, amygdalin, arbutin, esculin ferric citrate, salicin, d-cellobiose, d-trehalose, and gentiobiose) that were fermented by all of the vaginal isolates. d-ribose, dulcitol, d-mannitol, d-sorbitol, and methyl-ad-glucopyranoside were used by some of the rectal but none of the vaginal L. johnsonii isolates. l-arabinose was fermented only by L. murinus isolated from the rectum of animal ABL-6 (ABL-R6C), and d-turanose was utilized only by the L. acidophilus isolates, of both vaginal and rectal origins, from ABL-7 (ABL-V7B and ABL-R7B) ( Table 2).  Rep-PCR was also performed to determine the genomic fingerprint patterns of the Lactobacillus isolates. The fingerprint patterns were consistent in three independent rep-PCR runs, and a representative figure is shown in Fig. 1. As shown in Fig. 1, fingerprints of the vaginal isolates of L. johnsonii from animals ABL 1-4 are each different from one another, indicating that they are different isolates (Fig. 1A, lanes 2-5). When fingerprints of L. johnsonii (Fig. 1A, lanes 6-8) isolated from the same animal ABL-6 were analyzed, the band patterns were similar, but not identical, indicating again that the vaginal and rectal Lactobacillus isolates from the same animal are uniquely different from each other. Similar results were observed for the vaginal and rectal lactobacilli isolates from other animals: L. johnsonii isolates from ABL-7 (Fig. 1A, lanes 11 and 12), L. acidophilus isolates from ABL-7 (Fig. 1A, lanes 13 and 14), L. johnsonii from SRI-1 (Fig. 1B, lanes 2 and 3), L. johnsonii from SRI-3 (Fig. 1B, lanes 5-7), L. johnsonii from SRI-4 ( Fig. 1B, lanes 8 and 9), and L. johnsonii from SRI-5 (Fig. 1B, lanes 10 and 11). Although the fingerprints of the vaginal and rectal L. johnsonii isolates from animal ABL-8 seemed almost identical (Fig. 1A, lanes 16 and  17), carbohydrate fermentation profiles revealed differ-ences in the fermentation of methyl-ad-glucopyranoside and d-lactose, and they also differed in terms of H 2 O 2 production (Table 2). Similarly, vaginal and rectal L. murinus (Fig. 1A, lanes 9 and 10) from ABL-6 also have very similar fingerprints, but they differed in the fermentation of multiple carbohydrates (Table 2). Together, carbohydrate fermentation profiles, H 2 O 2 production, and genomic fingerprints revealed that the Lactobacillus isolates of the same species from the same animals are indeed different (Table 2 and Fig. 1).
Lactobacillus johnsonii isolated from Chinese rhesus macaques are closely related to Lactobacillus gasseri Sequence data of the 1537-nucleotide region of the 16S rRNA gene of the vaginal L. johnsonii isolates from Chinese rhesus macaques were aligned using Clustalw and a phylogenetic tree was constructed showing the relationship of L. johnsonii and a few representative human vaginal isolates. As shown in Fig. 2, all vaginal L. johnsonii isolates from the 12 macaques clustered together. They also have a very close phylogenetic relationship to a human vaginal L. johnsonii isolate from our collection (L. johnsonii 135-1). The 16S rRNA gene-based phylogenetic tree also showed that L. johnsonii has a closer relationship to human vaginal strains of L. gasseri than to Lactobacillus crispatus or L. jensenii.

Dynamic nature of endogenous vaginal L. johnsonii in Chinese rhesus macaques
In order to understand the dynamic nature of vaginal Lactobacillus flora over menstrual cycles, the level of endogenous vaginal lactobacilli was monitored at four time points over a period of 36 days in five Chinese rhesus macaques housed at Southern Research Institute. The macaques did not receive hormonal treatment, so their menstrual cycles were not synchronized. On average, between 10 5 and 10 7 CFU per vaginal swab of L. johnsonii was recovered in all macaques. For animal SRI-5, lactobacilli were not recovered at two time points (days 1 and 22). This may be because the amount of lactobacilli present at those time points was below the detection level of our microbiologic techniques. It was also noted that the level of endogenous lactobacilli dropped precipitously (to 10 4 ) during or immediately after menses (Fig. 3). For other non-lactobacilli endogenous bacteria, a similar pattern of changes was also observed. The average level of other endogenous microflora recovered was also about 10 6 to 10 7 CFU per vaginal swab, and less variety and lower quantity of each bacterial species was recovered during or immediately after menses (data not shown).

Colonization of a human vaginal isolate of L. jensenii 1153 in Chinese rhesus macaques
To test the potential use of Chinese rhesus macaques as a model for studying colonization of human vaginal Lactobacillus, the macaques were inoculated with a human vaginal isolate of L. jensenii 1153 for seven consecutive days. Bacterial concentrations of 10 6 and 10 5 CFU per vaginal swab were recovered 1 and 5 days after final bacterial inoculation, respectively (Fig. 4A). Lactobacillus jensenii 1153 forms small smooth colonies on the MRS agar plates, which can be differentiated from the large rough colonies of the endogenous L. johnsonii. L. jensenii isolates were further confirmed by 16S rRNA gene sequencing. In one of the macaques, 9 · 10 6 CFU per vaginal swab of L. jensenii was recovered 26 days after the final bacterial inoculation (data not shown). Apparently, the introduction of the human vaginal Lactobacillus isolate has no significant impact on the diversity or quantity of the endogenous vaginal flora of the macaques (data not shown).

Clearance of vaginal Lactobacillus and restoration of L. johnsonii in Chinese rhesus macaques
A human vaginal isolate of L. jensenii 1153 strain and the Lactobacillus isolates recovered from Chinese rhesus macaques were sensitive to erythromycin and azithromycin, an erythromycin derivative (data not shown). We chose to use azithromycin as a vaginal suppository to clear vaginal Lactobacillus as it is acid stable (which is important in a healthy vaginal setting) and is better tolerated by animals and humans [17,24]. By administering a vaginal suppository of azithromycin for 5 consecutive days, the exogenous human vaginal lactobacilli previously administered to the animals were readily cleared from the vaginal flora of Chinese rhesus macaques (Fig. 4B). In addition, no lactobacilli were recovered up to at least 36 days after the final azithromycin administration (Fig. 4B). Furthermore, endogenous vaginal L. johnsonii could be restored in these macaques. Respective isolates of L. johnsonii from individual macaques were reintroduced intravaginally into each of the two macaques. Approximately 10 7 CFU per vaginal swab of L. johnsonii were recovered 12 days after the final bacterial inoculation (Fig. 4B), demonstrating that endogenous L. johnsonii was restored in the Chinese rhesus macaques. On the contrary, no human vaginal Lactobacillus isolates were recovered from the 'reconstructed' microflora after azithromycin treatment, showing effective clearance of the pre-existing lactobacilli by the antibiotic.

Discussion
Macaques are commonly used for evaluation of the safety and efficacy of topical vaginal microbicides for preventing sexually transmitted diseases in humans [26,45,[47][48][49]. The similarities between the reproductive anatomy and physiology of human and rhesus macaques [36] prompted us to establish a nonhuman primate model suitable to evaluate the colonization of our Lactobacillus-based topical microbicide candidates. This study describes the use of the Chinese rhesus macaque as a potential model for persistent vaginal colonization of human Lactobacillus, and for pre-clinical evaluation of Lactobacillus-based topical microbicides. Our data demonstrated that Chinese rhesus macaques naturally harbor endogenous L. johnsonii among their vaginal flora.
Although several small animal models (such as rabbits, mice and rats) could possibly support transient vaginal colonization of Lactobacillus [7,31,34], they are not suitable or relevant for our purpose. All of these small animals require treatment with estrogen, and the profiles of their indigenous microflora show no resemblance to the microflora of humans [22,34]. Rabbits in particular are not suitable, as it has been reported that laboratory rabbits are usually in precoital status. They do not undergo estrous cycle stages and have little mucous secretion before mating. Vaginal environments with poor mucus secretion are inadequate for bacterial proliferation. Thus, the vaginal environment in precoital rabbits was suggested to be comparable to that during diestrus or anestrus in other animal species [34]. We found that diestrus does not support Lactobacillus colonization in a mouse model [28]. It also appears that the predominance of lactobacilli and low vaginal pH are distinct characteristics of the human vaginal microflora [5,20]. However, small animals have shown minimal numbers of vaginal lactobacilli and neutral vaginal pH [22,31,34]. In addition, anaerobic bacteria were the predominant flora in vagina of humans, and the total number of bacteria was higher than that in small laboratory animal models [18,34]. The lack of similarity prompted us to look for a primate model that better mimics the human vaginal environment for the testing of live topical vaginal microbicides. The normal vaginal flora we identified from the Chinese rhesus macaques in this study was generally similar to the findings of Doyle et al. [12] and Scorpio et al. [39] (Table 1). However, there are several major differences in the findings of these studies. First, we isolated and identified vaginal Lactobacillus in 12 of 13 of the animals we studied, and some of these macaques harbored more than one species of Lactobacillus over the course of the study. Lactobacillus johnsonii was the predominant endogenous vaginal Lactobacillus species we recovered, and it was isolated from animals housed at both ABL and Southern Research Institute. On the contrary, Doyle et al. did not find any lactobacilli in 37 rhesus macaques, and instead, reported Gardnerella-like or Gardnerella-probable organisms in about one-third of the animals, and Mobiluncus curtisii in half of the animals they studied. Although sampling time and bacterial identification methods were different in these two studies, it is difficult to determine whether these reasons alone contributed to the discrepancies. It was also unclear whether naı¨ve animals were used in the studies of Doyle et al. For instance, endogenous vaginal Lactobacillus is sensitive to certain antibiotic treatments, as evidenced in our studies. Although Scorpio et al. reported Lactobacillus spp. as one of the most commonly cultured bacteria from the 11 Chinese rhesus macaques in their study, it is unclear what species these lactobacilli are and how many animals harbor them [39].
In healthy women of childbearing age, the vaginal In this study, we demonstrated the presence of abundant endogenous vaginal Lactobacillus in rhesus macaques that have been widely used to test various candidate microbicides. In this animal model, endogenous vaginal Lactobacillus microflora could serve as a complimentary marker for preclinical toxicity evaluation of potential vaginal microbicide candidates. In addition, this is a potentially useful model for vaginal Lactobacillus colonization to support the pre-clinical evaluation of Lactobacillus-based topical microbicides. Indeed, we have successfully colonized a genetically modified L. jensenii strain that produces a heterologous protein for at least 87 days in this rhesus macaque model (unpublished data). We also determined that it is unnecessary to clear the endogenous flora for the colonization of exogenous human Lactobacillus in Chinese rhesus macaques (unpublished data). The decrease in the quantity of vaginal L. johnsonii recovered during or immediately after menses in macaques was similar to what has been reported with humans [13]. Therefore, immediately after menses should be the best time to intravaginally administer ecologically appropriate Lactobacillus.