lable at ScienceDirect

ARTICLE IN PRESS

Anaerobe xxx (2009) 1–6

Contents lists avai

Anaerobe

journal homepage: www.elsevier .com/locate/anaerobe

Clinical microbiology

In vitro inhibitory activity of human vaginal lactobacilli against pathogenicbacteria associated with bacterial vaginosis in Kenyan women

Martin N. Matu a,b,*, George O. Orinda a, Eliud N.M. Njagi a, Craig R. Cohen c, Elizabeth A. Bukusi b

a Department of Biochemistry and Biotechnology, Kenyatta University, P.O. Box 43844, 00100, Nairobi, Kenyab Kenya Medical Research Institute, Centre for Microbiology Research, P.O. Box 19464, 00202 Nairobi, Kenyac Department of Obstetrics, Gynecology and Reproductive Sciences, University of California, San Francisco, USA

a r t i c l e i n f o

Article history:Received 15 July 2008Received in revised form4 November 2009Accepted 9 November 2009Available online xxx

Key words:Vaginal lactobacilliBacterial vaginosisIndicator bacteriaAntagonistic activityHydrogen peroxide

* Correspondence to: Martin Matu, African MedicClinical and Diagnostic Programme, Kenya. Tel.: þ254810901.

E-mail address: martinmatu@gmail.com (M.N. Ma

1075-9964/$ – see front matter � 2009 Elsevier Ltd.doi:10.1016/j.anaerobe.2009.11.002

Please cite this article in press as: Matu Mdoi:10.1016/j.anaerobe.2009.11.002

a b s t r a c t

Lactobacilli have been shown to inhibit in vitro growth of many pathogens and have been used asprobiotics to treat a broad range of gastrointestinal and/or vaginal disorders. We sought to determine thein vitro inhibitory potential of lactobacilli of vaginal origin to some bacteria associated with bacterialvaginosis (BV), to characterize the inhibitory substances produced by these lactobacilli and to assess H2O2

production. Vaginal specimens were obtained by swabbing the lateral vaginal walls from 107 women twomonths following BV treatment. One hundred and fifty eight Lactobacillus spp. were isolated in 82 of the107 women. Lactobacillus jensenii was the predominant strain isolated among these women (29/158;18.4%). Among 158 culture supernatants tested for antibacterial activity against BV-associated bacteria,none inhibited the growth of Bacteroides fragilis while 23% (37/158), 28% (45/158) and 29% (46/158)inhibited the growth of Prevotella bivia, Gardnerella vaginalis and Mobiluncus spp. respectively. The lac-tobacilli produced supernatants with a pH range between 2.62 and 6.71; the highly acidic (pH 2–3.99)supernatants were more inhibitory to the indicator strains. There was significant reduction in the meanzones of inhibition following chemical and physical treatment of the supernatants (p ¼ 0.0025). Acid,bacteriocins and H2O2 demonstrated potential for antagonism of the bacterial pathogens. Thesesubstances may augment each other rather that each working independently on the pathogens.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Bacterial vaginosis (BV) is a condition characterized bydecreased hydrogen peroxide (H2O2) producing lactobacillifollowing a shift in vaginal microflora from the predominantlylactobacilli flora to an increase in gram variable coccobacilliconsistent with Gardnerella vaginalis or Bacteroides species [1]. Thiscondition has been associated with various gynaecological andobstetrical complications, including adverse pregnancy outcomessuch as amnionitis, premature rupture of membranes, post deliveryendometritis and preterm birth, pelvic inflammatory diseases (PID)and increased risk of acquisition of HIV-1 [2]. Antibiotic therapywith metronindazole or clindamycin has been shown to eliminateBV-associated bacteria. However, there is a high recurrence of BVwith up to 50% of treated cases experiencing relapses [3]. Relapses

al and Research Foundation,721374830/020 6994619/020

tu).

All rights reserved.

N, et al., In vitro inhibitory

have been associated with new sexual partners, having multiplepartners and unprotected sexual intercourse [4]. Evidence suggeststhat failure to establish normal flora following antimicrobialtherapy contributes to the recurrence [3]. Lactobacilli which arepredominant in a normal vaginal flora play a major role in protec-tion against colonization by pathogenic microorganisms such asthose that cause BV and other sexually transmitted diseasesincluding HIV [1,5,6]. Lactobacilli restrict the growth of pathogensthrough several mechanisms. First, lactobacilli produce lactic acidfrom carbohydrate metabolism which helps to maintain a lowvaginal pH i.e.<4.5. This acidity has been shown to be cidal to manyurogenital pathogens including HIV [5]. Secondly, different Lacto-bacillus spp. produce hydrogen peroxide (H2O2) which is alsoknown to be inhibitory to many pathogenic microorganisms. H2O2

suppress the endogenous pathogenic bacteria to maintain a healthyvaginal microflora [7] and has been shown to inhibit in vitro growthof G. vaginalis [8,9]. Production of bacteriocins by vaginal lactoba-cilli has also been demonstrated [10,11]. Bacteriocins are protein-aceous antibacterial compounds and exhibit bactericidal activityagainst species closely related to the producer strain with someshowing activity against unrelated species [12]. In addition,

activity of human vaginal lactobacilli against..., Anaerobe (2009),

M.N. Matu et al. / Anaerobe xxx (2009) 1–62

ARTICLE IN PRESS

lactobacilli exert a protective role against pathogens by competitionfor nutrients and exclusion of the pathogens to adhesion on hostcells [13,14]. The natural inhibition of some common pathogenicbacteria by lactobacilli may be important in understanding theinitiation of vaginal infections or recurrent BV associated with anunexplained decrease in vaginal lactobacilli [15]. The aim of thestudy was to determine the in vitro inhibition potential of lacto-bacilli of vaginal origin cultured from women two months afterreceiving treatment for BV to some bacteria associated with BV. Wealso sought to characterize the inhibitory substances produced bythese Lactobacillus isolates.

2. Materials and methods

2.1. Isolation and identification of lactobacilli from vaginalspecimen

Vaginal specimens were obtained from 107 women two monthsfollowing treatment for BV with metronindazole 400 mg twicedaily for one week. Lateral vaginal walls were swabbed with sterilecotton tipped applicators (Puritan Medical Products Company LLC,Gillford, Maine, USA). Lactobacilli were isolated by inoculating thespecimen onto three media; Rogosa agar (BBL Co inc.) containing0.2% glacial acetic acid, Columbia agar (Difco, Detroit, USA) con-taining colistin and nalidixic acid and Brucella agar (Difco, Detroit,USA) containing hemin (0.1%) and vitamin K (0.02%). The inocu-lated plates were incubated anaerobically at 37 �C for 48 h usingeither anaerobic jars (Mitsubishi Inc. USA) containing anaerobic gasgenerating kits (AnaeroPack) or anaerobic chamber system (FormaScientific). Identification of Lactobacillus spp. was performed byphenotypic method as previously described [16]. All the isolateswere initially tested for colony morphology, gram reaction, catalaseand indole production and nitrate reduction. Further character-ization was performed by carbohydrate fermentation [17,18].Lactobacillus spp. were tested for ability to ferment the followingsugars; lactose, cellobiose, rhamnose, mannitol, sucrose, melezi-tose, xylose, raffinose and arabinose as described by Jousimies-Somer et al. [17]. A modified microtiter plate method described byDavies et al. [18] was used for the carbohydrate fermentation.Briefly, 100 ml of tryptone broth containing bromothymol blue dyewas added to eight wells across a 96 wells microtiter plate. Twentyfive microliters of the appropriate carbohydrate solution was addedper well. Seventy five microliters of an overnight culture suspen-sion of the each Lactobacillus spp. in brain heart infusion brothwithout glucose was added to all eight wells containing differentcarbohydrates. The plates were sealed with microtiter adhesivecovers and incubated for 18–24 h at 37 �C anaerobically. Organismsin the wells that turned yellow after incubation were considered tobe fermenters while those that remained green were considered tobe non-fermenters. Wells containing dextrose alone were inocu-lated with Lactobacillus acidophilus ATCC 4356 probiotic strain aspositive control for fermentation while wells containing tryptonebroth and each of the respective carbohydrates without bacterialinoculation served as negative control.

2.2. Determination of hydrogen peroxide (H2O2) production

The test was carried out on Brucella agar containing 5 mg/ml ofvitamin K, 0.01 mg/ml of horseradish peroxidase 0.25 mg/ml tet-ramethylbenzidine (TMB) (Sigma LA, USA) [6]. A single colony frompure culture of Lactobacillus isolates was inoculated on the freshlyprepared agar and incubated anaerobically for 48 h at 37 �C. Theculture was then exposed to air at room temperature for up to30 min. The horseradish peroxidase hydrolyzes H2O2 to releaseoxygen that oxidizes TMB to form blue pigment that dyes the

Please cite this article in press as: Matu MN, et al., In vitro inhibitorydoi:10.1016/j.anaerobe.2009.11.002

colonies with blue colour. H2O2 production was noted by produc-tion of bluish pigmentation on colonies of H2O2 positive strains.Colonies of non-H2O2-producing strains did not turn blue on themedium.

2.3. Detection of acid production

Pure cultures of the different strains of lactobacilli were grownin brain heart infusion broth (pH 7.2) overnight at 37 �C supple-mented with 0.5% glucose. The amount of acid produced by theLactobacillus strains was measured indirectly by determining thepH of the culture supernatant by using pH meter (Corning pH meter610A).

2.4. Assay for inhibitory activity

Tests for inhibitory potential of the isolated strains of Lactoba-cillus spp. was carried out as described below.

2.4.1. Preparation of culture supernatantsPure colonies of lactobacilli were inoculated into brain heart

infusion broth and incubated overnight at 37 �C anaerobically. Thelactobacilli culture was centrifuged at 10,000 rpm for 5 min, and thesupernatant was used for the antagonism assays described below.

2.4.2. Deferred antagonism well assayAntibacterial activity of indicator bacteria by Lactobacillus

isolates was assessed by a modified deferred antagonism well assaydescribed by McLean and Rosenstein [19]. The indicator bacteria;Bacteroides fragilis, G. vaginalis, Prevotella bivia and Mobiluncus spp.were used to make a seeding suspension containing approximately1 � 107 colony forming units (cfu/ml). These organisms had beenisolated from women with BV from a different study and frozen at�80 �C. For the purpose of the study, these organisms werecultured from stock onto brucella agar containing supplements (asindicated section 2.1) for 48 h at 37 �C anaerobically. Pure culture ofeach of these organisms was used to make a suspension in sterileyeast extract broth equivalent to 0.5 McFarland standard withapproximately 107 cfu/ml of the indicator bacteria. Using sterileswabs, this suspension was inoculated onto agar surface of freshbrucella agar base containing hemin (0.1%) and vitamin K (0.02%) towhich 6 mm wells had been made using cork borer. Lactobacilliculture supernatant was filter sterilized through a 0.45 mm poly-ethersulfone membrane filter (Nalgene, Rochester, USA) and 100 mlof the supernatant was added to 6 mm wells made into the agar.The inoculated plates were incubated anaerobically at 37 �C for 48 heither in anaerobic chamber system model 1025 (Forma Scientific)or in rectangular jars with gas generating kits (AnaeroPack system,Mitsubishi Gas Chemical Co., Inc.). The zones of inhibition weremeasured from the outer edge of bacterial growth using callipers.Zones of 1 mm and above were considered as inhibition.

2.4.3. Characterization of inhibitory substanceThe culture supernatants of the Lactobacillus spp. that showed

antibacterial activity to indicator bacteria in the well antagonismassay in 2.4.2 above were tested for potential role of acids andbacteriocins using a method described by Jin et al. [20].

2.4.3.1. Assay for bacteriocins. To assay for bacteriocins aliquots ofculture supernatants of the lactobacilli cultured in 2.4.1 weretreated with three enzymes in the same tube which included;trypsin from beef extract (BDH), pepsin and proteinase K solutioneach at a final concentration of 1 mg/ml for 60 s at room temper-ature. These enzymes were selected for their ability to degradebacteriocins. 100 ml of each of the three enzyme solutions was

activity of human vaginal lactobacilli against..., Anaerobe (2009),

Table 1The total number of Lactobacillus species isolated among sampled women.

Lactobacillus species BV Intermediate Normal Prevalence

L. acidophilus 10 (20) 2 (4.7) 11 (16.9) 23 (14.6)L. brevis 1 (2) 0 3 (6.7) 4 (2.5)L. casei 2 (4) 1 (2.3) 0 3 (1.9)L. cateneformis 2 (4) 0 3 (6.7) 5 (3.2)L. crispatus 0 1 (2.3) 0 1 (0.6)L. fermentum 3 (6) 5 (11.6) 12 (18.5) 20 (12.6)L. iners 6 (12) 7 (16.3) 10 (15.4) 23 (14.6)L. jensenii 10 (20) 6 (13.9) 13 (20) 29 (18.4)L. minutus 3 (6) 0 2 (3.1) 5 (3.2)L. plantanarum 6 (12) 12 (27.9) 2 (3.1) 20 (12.6)L. paracasei 0 1 (2.3) 1 (1.5) 2 (1.3)L. paraplantanarum 4 (8) 2 (4.7) 3 (6.7) 9 (5.7)L. rhamnosus 0 1 (2.3) 4 (6.2) 5 (3.2)L. salivarius 0 2 (4.7) 1 (1.5) 3 (1.9)L. vaginalis 3 (6) 3 (6.9) 0 6 (3.8)

50 (31.6) 43 (27.2) 65 (41.1) 158

Table 2Number of women with different BV status colonized with Lactobacillus species.

Vaginal flora status Lactobacilli colonization H202 Lactobacillus spp.

Not colonized(%)

Colonized(%)

Positive (%) Negative (%)

Normal 6 (17) 30 (83) 11 (37) 19 (63)Intermediate 7 (23) 23 (77) 8 (35) 15 (65)Bacterial vaginosis 12 (29) 29 (71) 12 (41) 17 (59)

M.N. Matu et al. / Anaerobe xxx (2009) 1–6 3

ARTICLE IN PRESS

added to the same tube containing 1 ml of the culture supernatantof each strain of lactobacilli to be tested and allowed to stand atroom temperature for 1 min. The treated supernatant was thentested for antagonistic activity to indicator bacteria by agar diffu-sion assay described in section 2.4.2 above. The indicator bacteriawere seeded onto agar plates and 100 ml of the enzymes treatedculture supernatant of lactobacilli was transferred into the wellsmade prior to seeding of indicator bacteria. The plates wereobserved for inhibition after incubation for 24 h at 37 �C. Zones ofinhibition were measured as in the antagonism assay describedabove.

2.4.3.2. Effect of pH on antimicrobial activity. Culture supernatantsof an overnight culture of Lactobacillus spp. were neutralized to pH7.2 by addition of 1 N NaOH. To 2 ml of the culture supernatant, 1 NNaOH was added drop wise to bring the pH to 7.2. 100 ml of thetreated supernatant was used for antagonism assay to test inhibi-tory activity to the selected BV associated organisms by the modi-fied differed antagonism assay described above.

2.4.3.3. Effect of heat on antimicrobial activity. Two millilitrealiquots of the overnight culture supernatant of inhibitory lacto-bacilli was put into a sterile test tube and allowed to boil in a hotwater bath at 100 �C for 30 min. To test heat sensitivity, 100 ml ofheated culture supernatant was added to wells and the residualactivity was assayed by well antagonism assay using the sameindicator organisms as used in the above assays.

2.5. Statistical analysis

All data were complied and analyzed using SPSS version 15.0statistical software package (SPSS Inc. Chicago, IL.). Fisher’s exacttest was used for analysis of categorical variables and radii of zonesof inhibition were measured and differences analyzed by Mann–Whitney U test. P values less than 0.05 were considered significant.

3. Results

3.1. Distribution of Lactobacillus spp. among women with differentvaginal status

A total of 158 different Lactobacillus strains were isolated from 82(76.7%) women found colonized with Lactobacillus strains. Amongthe Lactobacillus spp. isolated were; 29 (18.4%) Lactobacillus jensenii,23 (14.6%) L. acidophilus, 23 (14.6%) Lactobacillus iners, 20 (12.6%)Lactobacillus fermentum, 20 (12.6%) Lactobacillus plantanarum, 9(5.76%) Lactobacillus paraplantanaum, 6 (3.8%) Lactobacillus vagi-nalis, 5 (3.2%) Lactobacillus rhamnosus, 5 (3.2%) Lactobacillus minu-tus, 5 (3.2%) Lactobacillus cateneformis, 4 (2.5%) Lactobacillus brevis, 3(1.9%) Lactobacillus casei, 3 (1.9%) Lactobacillus salivarius, 2 (1.3%)Lactobacillus paracasei and 1 (0.6%) Lactobacillus crispatus (Table 1).We evaluated the distribution of the Lactobacillus spp. in womenwith different vaginal status; 65 of 158 (41.1%) Lactobacillus specieswere isolated from women with normal vaginal flora, while 50(31.6%) and 43 (27.2%) Lactobacillus species were isolated from BVand intermediate vaginal flora respectively. The distribution ofLactobacillus species isolated from women with BV ranged from1 (2%) for L. brevis to 10 (20%) for L. acidophilus and L. jensenii each.Distribution in women with intermediate vaginal flora ranged from1 (2.3%) L. rhamnosus, L. crispatus, L. paracasei and L. casei each to 12(27.9%) L. plantanarum. Those isolated from women with normalvaginal flora ranged from 1(1.5%) for L. salivarius and L. paracaseieach, to 13 (20%) for L. jensenii. There was no significant difference inthe distribution of Lactobacillus spp. isolated in women with thedifferent categories of vaginal flora (p ¼ 0.32, Table 1).

Please cite this article in press as: Matu MN, et al., In vitro inhibitorydoi:10.1016/j.anaerobe.2009.11.002

3.2. Lactobacillus spp. colonization among women with differentvaginal status

Eighty two (77%) of the 107 women sampled were colonized byLactobacillus spp., 30 (83%) had normal flora, 23 (77%) had inter-mediate flora and 29 (71%) had BV. Although more women withnormal flora were colonized by lactobacilli, there was no significantassociation between the vaginal flora and colonization by lactoba-cilli (p > 0.05). Of the 30 women colonized by lactobacilli who hadnormal flora, 11 (37%) had H2O2 Lactobacillus spp. compared to 8(35%) and 12 (41%) of the women with intermediate and BV florarespectively. More women with BV had H2O2 Lactobacillus spp.(p < 0.001; Table 2).

3.3. Inhibitory activity

Culture supernatants obtained from the 158 Lactobacillus spp.were tested for antibacterial activity against BV-associated bacteria.The supernatants were assayed against one strain of G. vaginalis,P. bivia, B. fragilis and Mobiluncus spp. None of the Lactobacillus spp.tested was able to inhibit growth of B. fragilis, 37 of 158 (23%)inhibited growth of P. bivia, 45 of 158 (28%) inhibited growth of G.vaginalis, and 46 of 158 (29%) of the Lactobacillus spp. inhibitedgrowth of Mobiluncus spp. (Fig. 1).

3.4. Inhibition zones

Mobiluncus spp., G. vaginalis, Bacteroides spp. and P. bivia,organisms commonly associated with BV as reported in number ofstudies were selected to be used as indicator bacteria [9,21]. Theinhibitory Lactobacillus spp. produced zones of inhibition with thefollowing ranges; 1–8 mm, 1.5–8 mm and 1.5–7 mm for Mobiluncusspp., G. vaginalis and P. bivia, respectively. The G. vaginalis was moresusceptible to inhibition than the other two susceptible indicatorbacteria; the mean (mm � SD) zones of inhibition were5.2 mm � 2.0 for G. vaginalis as compared to 4.8 mm � 1.8 and

activity of human vaginal lactobacilli against..., Anaerobe (2009),

Fig. 1. Distribution of the number of inhibitory and non-inhibitory strains of Lacto-bacillus spp. to the different indicator bacteria.

Table 3Relationship between pH of supernatant and growth inhibition of BV-associatedbacteria.

Level of acidity n (158) Indicator bacteria inhibited

G. vaginalis(%)

P. bivia(%)

Mobiluncusspp. (%)

B. fragilis(%)

Highly acidicd 53 (33.5) 26 (49.1) 18 (33.9) 25 (47.2) 0Moderately acidicc 33 (20.9) 14 (42.4) 13 (39.4) 15 (45.4) 0Slightly acidicb 26 (16.5) 5 (19.2) 5 (19.2) 5 (19.2) 0Very slightly acidica 46 (29.1) 1 (2.3) 1 (2.3) 1 (2.3) 0

a pH range of 6.0–6.99.b pH ¼ 5.0–5.99.c pH ¼ 4.0–4.99.d pH ¼ 2.0–3.99.

Table 4Mean inhibition zone among the 16 selected inhibitory Lactobacillus spp. aftertreatment.

Treatment Mean zone of inhibition of the indicator strains (mm)

G. vaginalis Mobiluncus spp. P. bivia

No treatment 4.31 4.09 2.96Boiling 3.11 0a 0a

pH 7.2 3.2 2 0a

Proteases 3 2.1 0a

a Indicator bacteria not inhibited.

M.N. Matu et al. / Anaerobe xxx (2009) 1–64

ARTICLE IN PRESS

3.0 mm� 1.3 for Mobiluncus spp. and P. bivia, respectively. B. fragiliswas not inhibited by any of the Lactobacillus spp. tested. The meanzones sizes of inhibition produced against the different indicatorbacteria were statistically significant (p < 0.0001).

3.5. Production of acid by the lactobacilli

The pH of the supernatant was measured before use in theinhibition assay. The pH ranged between 2.62 and 6.71 with a meanpH of 4.82 � 1.1, and a mode of 4. Fifty three of the 158 (33.5%)produced highly acid pH (2–3.99), 33 of 158 (20.9%), 26 of 158(16.5%) and 46 of 158 (29.1%) produced culture with moderatelyacidic pH (4–4.99), slightly acidic pH (5–5.99) and very slightlyacidic pH (6–6.99), respectively. More of the highly acidic Lacto-bacillus spp. culture supernatants inhibited the growth of G. vagi-nalis and Mobiluncus spp. while most of those that inhibited P. biviawere moderately acidic. Generally, there was an inverse relation-ship between inhibition of G. vaginalis and Mobiluncus spp., andincrease in pH of the supernatant, supernatants that were moreacidic inhibited many of the indicator bacteria. There was a signif-icant difference between inhibition of indicator bacteria anddecreasing pH (p < 0.0001; Table 3).

3.6. Effect of treatment on zones of inhibition

Sixteen of 33 different Lactobacillus spp. that inhibited three ofthe indicator bacteria were randomly selected for further analysis.Sensitivity to physical treatment by boiling the lactobacilli super-natants at 100 �C for 30 min and treatment with chemicalsubstances using proteases and raised pH (pH 7.2) were assessed.None of culture supernatants tested inhibited P. bivia i.e. boiling for30 min, adjustment of pH to 7.2 and after proteases treatment for1 h at room temperature, but some of the supernatants continuedto inhibit G. vaginalis and Mobiluncus spp. producing smaller zonesof inhibition compared to the untreated supernatant (Table 4). Afterboiling the supernatant, raising pH to 7.2 and treating with prote-ases, the inhibitory supernatants produced mean inhibition zone of3.11 mm, 3.2 mm and 3.0 mm respectively. There was significantreduction in the mean zones of inhibition after chemical andphysical treatment (p ¼ 0.0025; Table 4).

3.7. Relationship between the Lactobacillus spp., and production ofacid, H2O2 and inhibitory capacity

Eighty eight (56%) of the 158 Lactobacillus spp. producedsupernatants with pH less than 5, while the rest 70 (44%) produced

Please cite this article in press as: Matu MN, et al., In vitro inhibitorydoi:10.1016/j.anaerobe.2009.11.002

supernatants with pH greater than 5. As shown in Table 5 morethan 50% each of the following Lactobacillus spp. produced super-natants with pH greater than 5; L. acidophilus, L. crispatus, L. fer-mentum, L. jensenii, L. plantanarum, L. salivarius. The differences inthe acid production by the different species were not significant(p¼ 0.1762; Table 5). Forty eight (30%) of the Lactobacillus spp. wereH2O2 producers, over 30% of each of the following species wereH2O2 producers; L. casei, Lactobacillus cartenaformis, L. jensenii, L.minutus, L. paracasei, Lactobacillus paraplantanarum, L. salivariusand L. vaginalis. There were no differences observed between thespecies isolated and the production of H2O2 (p > 0.05; Table 5). Weassessed the relationship between the Lactobacillus spp. and theability to inhibit the indicator bacteria. A higher percentage of the L.acidophilus isolates was more inhibitory than the other strains(p ¼ 0.0361; Table 5).

4. Discussion

Among 107 women treated for BV whose vaginal swabs weretaken for evaluation of H2O2-producing lactobacilli, a large diversityof lactobacilli was seen. L. jensenii was the most predominantspecies isolated from vaginas of these women. In general, differentstudies have reported L. crispatus, L. jensenii, L. acidophilus andL. fermentum to be the predominant species in vaginal tracts ofwomen with normal flora in majority Caucasian populations [22–24]. In other studies, Lactobacillus gasseri has been reported to bethe predominant strain of Lactobacillus colonizing the vagina. Aslimand Kilic [25] reported having isolated L. gasseri as the predominantstrain in Turkish women while Song et al. [26] (1999) reportedL. crispatus and L. gasseri as the predominant species colonizing thevaginas of Japanese women. In work done in Africa, Jin et al. [27] ina study done in Uganda found Lactobacillus reuteri, L. crispatus,L. vaginalis and L. jensenii to be the predominant species. A study inNigeria found L. iners and L. gasseri to be the predominant species[28]. These differences may be attributed to racial differencesamong the different populations’ samples and the methods ofidentification. Some investigators have used molecular methods ofidentification [23,24] while some have used phenotypic methods[25] for lactobacilli identification.

activity of human vaginal lactobacilli against..., Anaerobe (2009),

Table 5Production of H2O2 and acid by different species of lactobacilli.

Lactobacillusspp.

Acid production H2O2 production Inhibition

pH < 5 pH � 5 Positive Negative Inhibitory Not inhibitory

L. acidophilus 15 (65) 8 (35) 5 (21) 18 (79) 15 (65) 8 (35)L. brevis 2 (50) 2 (50) 1 (25) 3 (75) 2 (50) 2 (50)L. casei 1 (33) 2 (67) 1 (33) 2 (67) 1 (33) 2 (67)L. cateneformis 1 (20) 4 (80) 2 (40) 3 (60) 0 5 (100)L. crispatus 1 (100) 0 0 1 (100) 0 1 (100)L. fermentum 12 (60) 8 (40) 5 (25) 15 (75) 5 (25) 15 (75)L. iners 9 (39) 14 (61) 5 (21) 18 (79) 2 (9) 21 (91)L. jensenii 19 (66) 10 (34) 10 (34) 19 (66) 11 (38) 18 (62)L. minutus 1 (20) 4 (80) 3 (60) 2 (40) 2 (40) 3 (60)L. plantanarum 14 (70) 6 (30) 4 (20) 16 (80) 4 (20) 16 (80)L. paracasei 1 (50) 1 (50) 1 (50) 1 (50) 0 2 (100)L. paraplantanarum 4 (44) 5 (56) 3 (33) 6 (67) 3 (33) 6 (67)L. rhamnosus 1 (20) 4 (80) 2 (40) 3 (60) 1 (20) 4 (80)L. salivarius 3 (100) 0 2 (67) 1 (33) 1 (33) 2 (67)L. vaginalis 4 (67) 2 (33) 4 (66) 2 (33) 2 (33) 4 (67)

M.N. Matu et al. / Anaerobe xxx (2009) 1–6 5

ARTICLE IN PRESS

Lactic acid bacteria have been shown to inhibit the in vitro growthof many pathogens and have been used as probiotics in both humansand animals to treat a broad range of gastrointestinal and/or vaginaldisorders [15,29–31]. The inhibition has been attributed to produc-tion of organic acids such as lactic acid [32], H2O2 production andbacteriocins [15] and competition between pathogens and lactoba-cilli for adhesion and nutrients [13,14]. In this study, only 21% (33/158) of the isolated strains of lactobacilli inhibited all three indicatorspecies. The lactobacilli produced zones of inhibition of growth inthe range 1.5–8 mm with G. vaginalis being the most susceptible to invitro inhibition by lactobacilli. McLean and Rosenstein [19] tested 60vaginal isolates of lactobacilli for their ability to inhibit the growth ofG. vaginalis, Bacteroides spp., and P. bivia and found all the testedlactobacilli inhibited each of the indicator bacteria.

Lactic acid produced by vaginal lactobacilli contributes to themaintenance of a low vaginal pH and high redox potential whichcan inhibit growth of other bacteria [19,31,33]. McLean and Rose-nstein [19] observed that strains that produced a more acidicculture were the most inhibitory, therefore demonstrating the roleof organic acids in inhibition pathogen growth. Similarly, in thepresent study, lactobacilli isolates that produced highly acidicsupernatant (pH 2.62–3.99) were more inhibitory.

In this study, sixteen of 33 strains of lactobacilli that wereinhibitory to all three indicator species were selected for furthercharacterization. Their culture supernatant was evaluated for effectof treatment by physical (heat treatment at 100 �C for 30 min), andchemical substances (treatment with proteases for 1 h and pH to 7.2with 1 N NaOH). We found that treatment of the supernatantdecreased the inhibition of the three indicator bacteria. This indi-cated that acid and bacteriocins were potential contributing factorsto inhibition by these lactobacilli.

The restoration of the ecologic balance has lately been proposedto control the infectivity of pathogenic microbes in the vagina andbladder. The genus Lactobacillus has been applied in differenthuman and animal tracts as probiotic microorganisms with thisobjective [28,29,34]. In development of appropriate biotherapeuticremedy for BV such as probiotic lactobacilli, proper identificationand understanding of the properties of the dominant vaginal lac-tobacilli is vital. There has been scanty information on thepredominant vaginal lactobacilli in women in the sub-SaharanAfrica where BV is highly prevalent.

5. Limitations of the study

The population of women sampled was highly selective (twomonths post BV treated women) and therefore may not representAfrican women. The study used phenotypic methods for

Please cite this article in press as: Matu MN, et al., In vitro inhibitorydoi:10.1016/j.anaerobe.2009.11.002

identification of the Lactobacillus spp. and these methods have beenreported to have limitations in proper identification and differen-tiation of some Lactobacillus spp.

6. Conclusion

Our findings suggest a potential role of acid, bacteriocins andH2O2 in antagonism of the bacterial pathogens. G. vaginalis andMobiluncus spp. were the more susceptible to inhibition than P.bivia. L. jensenii, L. acidophilus and L.iners were the most predomi-nant strains of lactobacilli and therefore when assessing fora potential bacteriotherapy it may be worth while using thesepredominant vaginal colonists. L. acidophilus strains in addition tobeing predominant were found to be more inhibitory than otherstrains. Our findings also indicated that H2O2-producing lactobacilliare present in both women with and without BV and we suggestthat, besides H2O2 other factors play a role in inhibition of vaginalpathogens. Of the tested lactobacilli only 21% of the isolatesinhibited all the three indicator bacteria. We found that many of thelactobacilli tested did not inhibit the BV-associated bacteria testedand therefore may not have been protective to the women; thiswould partially explain why many of these women would haverecurrent infection. However since some of the tested isolated hadgood inhibitory power, we encourage further in vivo studies, suchas clinical trials designed to test the capacity of these strains toprevent and manage urogenital tract infections in females.

Acknowledgements

We thank the Director, Kenya Medical Research Institute and thestaff of Centre for Microbiology Research for their invaluablesupport. We thank John Kiiru, Musa Otieno, Ian Waweru of KenyaMedical Research Institute and Eric Ndiang’ui of Egerton Universitywho provided valuable input before and during the study.

Source of funding

This study was supported in part through the study of Dr. ElizabethBukusi, a fellow of the International AIDS Research and TrainingProgram at the University of Washington, supported by the FogartyInternational Centre (#T22TW00001) and in part by: FIRCA grantNo: R03 TW05820, the University of Washington Centre for AIDSResearch (P30-AI-27757) and a WHIN supplementary grant (HD40540-04).

activity of human vaginal lactobacilli against..., Anaerobe (2009),

M.N. Matu et al. / Anaerobe xxx (2009) 1–66

ARTICLE IN PRESS

References

[1] Hillier S, Holmes KK. In: Holmes KK, Sparling PF, Mardh PA, editors. Sexuallytransmitted diseases. 3rd ed. New York: McGraw-Hill; 1999. p. 563–86.

[2] Sewankambo N, Gray RH, Wawer MJ. HIV-1 infection associated withabnormal vaginal flora morphology and bacterial vaginosis. Lancet1997;350:546–50.

[3] Hillier SL, Krohn MA, Rabe LK, Klebanoff SJ, Eschenbach DA. The normalvaginal flora, H2O2-producing lactobacilli, and bacterial vaginosis in pregnantwomen. Clinical and Infectious Diseases 1993;16(4):273–81.

[4] Wilson J. Managing recurrent bacterial vaginosis. Journal of Sexually Trans-mitted Infection 2004;80:8–11.

[5] Hawes SE, Hillier SL, Benedetti J, Stevens CE, Koutsky LA, Wolner-Hanssen P,et al. Hydrogen peroxide-producing lactobacilli and acquisition of vaginalinfections. Journal of Infectious Diseases 1996;1996(174):1058–63.

[6] Eschenbach DA, Davick PR, Williams BL, Klebanoff SJ, Young-Smith K,Critchlow CM, et al. Prevalence of hydrogen peroxide producing species innormal women and women with bacterial vaginosis. Journal of ClinicalMicrobiology 1989;27:251–6.

[7] Aroutcheva A, Gariti D, Simon M, Shott S, Faro J, Simoes JA, et al. Defensefactors of vaginal lactobacilli. American Journal of Obstetrics and Gynaecology2001;185:375–9.

[8] Pascual LM, Daniele MB, Cristina P, Barberis T. Lactobacillus species isolatedfrom the vagina: identification, hydrogen peroxide production and non-oxynol-9 resistance. Contraception 2006;73:78–81.

[9] Klebanoff SJ, Hillier SL, Eschenbach DA, Waltersdorph AM. Control of themicrobial flora of the vagina by H2O2-generating lactobacilli. Journal ofInfectious Diseases 1991;164:94–100.

[10] Sobel JD. Is there a protective role for vaginal flora? Current InfectiousDiseases 1999;1:379–83.

[11] Alpay KS, Aydin F, Kilic, SS, Kilic AO. Antimicrobial activity and characteristicsof bacteriocins produced by vaginal lactobacilli. Turkish Journal of MedicalSciences 2003;33:7–13.

[12] Cintas LM, Herranz C, Hernandez PE, Casaus MP, Nes IF, Hernandez PE. Review:bacteriocins of lactic acid bacteria. Food Science and Technology International2001;7:281–305.

[13] Reid G, Burton J. Use of Lactobacillus to prevent infection by pathogenicbacteria. Microbes Infection 2002;4:319–24.

[14] Vesterlund S, Karp M, Salminen S, Ouwehand AC. Staphylococcus aureusadheres to human intestinal mucus but can be displaced by certain lactic acidbacteria. Microbiology 2006;152:1819–26.

[15] Karaoglu SA, Faruk A, Kilic SS, Kilic AO. Antimicrobial activity and character-istics of bacteriocins produced by vaginal lactobacilli. Turkish Journal ofMedical Science 2003;33:7–13.

[16] Tynkkynen S, Satokari R, Saarela M, Mattila-Sandho T, Saxelin M. Comparisonof ribotyping, randomly amplified polymorphic DNA analysis, and pulsed-fieldgel electrophoresis in typing of Lactobacillus rhamnosus and L. casei strains.Applied and Environmental Microbiology 1999;65:3908–14.

[17] Jousimies-Somer HR, Summanen P, Citron DM, Baron EJ, Wexler HM,Finegold SM. Wadsworth anaerobic bacteriology manual. 6th ed. Belmont,Calif: Star Publishing Co.; 2002.

Please cite this article in press as: Matu MN, et al., In vitro inhibitorydoi:10.1016/j.anaerobe.2009.11.002

[18] Davies JA, Mitzel JR, Beam WE. Carbohydrate fermentation patterns of Neis-seria meningitidis determined by a microtiter method. Applied Microbiology1971;21(6):1072–4.

[19] McLean NW, Rosenstein IJ. Characterisation and selection of a Lactobacillusspecies to re-colonise the vagina of women with recurrent bacterial vaginosis.Journal of Medical Microbiology 2000;49:543–52.

[20] Jin LZ, Ho YW, Abdullah N, Ali MA, Jalaludin S. Antagonistic effects of intestinalLactobacillus isolates on pathogens of chicken. Letter of Applied Microbiology1996;23:67–71.

[21] Verhelst R, Verstraelen H, Claeys G, Verschraegen G, Simaey LV, De Ganck C,et al. Comparison between gram stain and culture for the characterization ofvaginal microflora: definition of a distinct grade that resembles grade Imicroflora and revised categorization of grade I microflora. BMC Microbiology2005;5:61–72.

[22] Boris S, Suarez JE, Vazquez F, Barbes C. Adherence of human vaginal lacto-bacilli to vaginal epithelial cells and interaction with uropathogens. Infectionand Immunity 1998;66:1985–9.

[23] Antonio MA, Hawes SE, Hillier SL. The identification of vaginal Lactoba-cillus species and the demographic and microbiologic characteristics ofwomen colonized by these species. Journal of Infectious Diseases1999;180:1950–6.

[24] Anukam KC, Osazuwa EO, Ahonkhai I, Reid G. 16S rRNA gene sequence andphylogenetic tree of Lactobacillus species from the vagina of healthy Nigerianwomen. African Journal of Biotechnology 2005;4(11):1222–7.

[25] Aslim B, Kilic E. Some probiotic properties of vaginal lactobacilli isolated fromhealthy women. Japanese Journal of Infectious Disease 2006;59:249–53.

[26] Song Y, Kato N, Matsumiya Y, Liu C, Kato H, Watanabe K. Identification of andhydrogen peroxide production by fecal and vaginal lactobacilli isolated fromJapanese women and newborn infants. Journal of Clinical Microbiology1999;37:3062–4.

[27] Jin L, Tao L, Pavlova SI, So J-S, Kiwanuka N, Namukwaya Z, et al. Speciesdiversity and relative abundance of vaginal lactic acid bacteria from women inUganda and Korea. Journal of Applied Microbiology 2007;102:1107–15.

[28] Anukam KC, Osazuwa EO, Ahonkhai I, Reid G. Lactobacillus vaginal microbiotaof women attending a reproductive health care service in Benin city, Nigeria.Sexually Transmitted Diseases 2006;33:59–62.

[29] Falagas ME, Betsi GI, Athanasiou S. Probiotics for the treatment of women withbacterial vaginosis. Clinical Microbiology and Infection 2007;13(7):657–64.

[30] Piyawan VS, Bilasoia S, Supamala O. Antagonistic activity against pathogenicbacteria by human vaginal lactobacilli. Anaerobe 2006;12:221–6.

[31] Rolfe DR. The role of probiotic cultures in the control of gastrointestinal health.Journal of Nutrition 2000;130:396–402.

[32] Oyetayo VO. Phenotypic characterisation and assessment of the inhibitorypotential of Lactobacillus isolates from different sources. African Journal ofBiotechnology 2004;7:355–7.

[33] Holmes KK, Chen KC, Lipinski CM, Eschenbach DA. Vaginal redox potential inbacterial vaginosis (non-specific vaginitis). Journal of Infectious Diseases1985;152:379–82.

[34] Redondo-Lopez V, Cook RL, Sobel JD. Emerging role of lactobacilli in thecontrol and maintenance of the vaginal bacterial micro-flora. Review ofInfectious Diseases 1990;12:856–72.

activity of human vaginal lactobacilli against..., Anaerobe (2009),