Competitive Selection of Lactic Acid Bacteria That Persist in the Human Oral Cavity

Abstract

Lactic acid bacteria (LAB) might offer opportunities as oral probiotics provided candidate strains persist in the mouth. After intake of a mixture of 69 LAB, strains of Lactobacillus fermentum and Lactobacillus salivarius were especially recovered. Coaggregation with other microbes is likely not a prerequisite for persistence since L. salivarius strongly coaggregated with typical oral cavity isolates, whereas L. fermentum failed to display this phenotype.

TEXT

Certain strains of lactic acid bacteria (LAB) are of interest as probiotics, which are defined as “live microorganisms that when administered in adequate amounts confer a health benefit on the host” (7). For oral health applications, despite broad interest of the scientific and industrial communities (2, 4, 17), functional criteria for selection of probiotics are in their infancy, and correlations between in vitro data and human intervention studies are scarce (6, 15). One potential mechanism of an oral probiotic is the inhibition of growth and maintenance of detrimental resident bacteria in specific oral sites. The screening of lactic acid bacterial species from oral cavities led to the identification of strains of Lactobacillus paracasei subsp. paracasei and Lactobacillus rhamnosus, which inhibited the growth of oral pathogens in vitro, including Streptococcus mutans and Porphyromonas gingivalis (20). Probiotic effects have also been demonstrated in vivo. The probiotic Streptococcus salivarius K12 is proposed to persist in the oral cavity, where it changes the bacterial community and improves oral malodor parameters (1). Similar observations have been reported for Weissella cibaria (11).

For strategies to decrease the activity or abundance of the detrimental bacteria, colonization, or at least temporal persistence of probiotic bacteria, is a phenotypic trait, which is highly likely to be required to achieve a functional health benefit (9, 19). The work presented here evaluates the competitive persistence of a range of LAB in the human mouth. A total of 69 food-grade LAB strains from the Lactobacillus, Lactococcus, and Streptococcus genera were evaluated for their persistence in vivo in the human oral cavity. The strains were obtained from the NIZO culture collection as well as public culture collections (Table 1). Spontaneous rifampin-resistant mutants were selected upon subculturing the wild-type strains in medium containing 10 μg/ml rifampin and subsequently in 50 μg/ml rifampin. The growth rates of the rifampin-resistant mutants were similar to that of wild-type cells in laboratory culture medium (data not shown).

Table 1.

LAB examined for persistence in the human mouth^a

Species	Source	Alternate	Originc
Lactobacillus acidophilus	NIZO867	LMG 7943, DSM 20079	N/A
	NIZO221	ATCC 4357	N/A
	NIZO222		N/A
	NIZO223		N/A
	NIZO225		N/A
	NIZO229		N/A
	NIZO267		N/A
L. brevis	NIZO2927	NCIMB 8840	Human saliva
	NIZO289		Cheese
	NIZO2019		Cheese
	NIZO1322	LMG 7944, DSM 20054	Human feces
	NIZO293		Cheese
	NIZO2491		Pork pickled sausage
Lactobacillus bulgaricus	5.2		Campina starter culture
	2.3		Campina starter culture
Lactobacillus casei subsp. casei	NIZO2928	NCIMB 8822	Human saliva
	NIZO2929	NCIMB 8823	Human saliva
	NIZO637		N/A
	NIZO889		N/A
	NIZO931		N/A
L. delbrueckii subsp. lactis	NIZO235	ATCC 7830	N/A
	NIZO2944	DSM 20073	Saliva
L. fermentum	NIZO2930	NCIMB 701751	Saliva
	NIZO2931	NCIMB 700335	Human oral strain
	NIZO2517	LMG 9846	Saliva
	NIZO2932	NCIMB 8828	Human saliva
	NIZO2933	NCIMB 8829	Human saliva
	NIZO2934	NCIMB 8830	Human saliva
	NIZO307	ATCC 9338	Human oral cavity
	NIZO1220	LMG11441	N/A
L. paracasei subsp. paracasei	NIZO2935	NCIMB 700680	Oral source
	NIZO2936	NCIMB 702713	Child saliva
	NIZO2518	DSM 20020	Child saliva
	NIZO2945	DSM 4905	Oral cavity
	NIZO1480	DSM 20244	Milk
	NIZO632		N/A
	NIZO1353	DSM 5622, ATCC25302	N/A
Lactobacillus pentosus	NIZO2514		Bamboo shoot, pickled
Lactobacillus plantarum	NIZO631		N/A
	NIZO2519	LMG 9212	Human saliva
	NIZO1315		N/A
	NIZO1699		Soakwater of soy beans
	NIZO1317	DSM 20174, LMG6907	Pickled cabbage
	NIZO2029		Raw-milk cheese
	NIZO1843		N/A
	NIZO2484		Pork pickled sour sausage
	NIZO2260	299v, DSM 9843	Human intestine
	NIZO2500		Pork pickled sour sausage
	NIZO2532		Shrimp pickled sausage
	NIZO1836	NCIMB 8826, WCFS1, LMG9211	Human saliva
L. reuterib	NIZO2691		Breast milk (BioGaia product)
L. rhamnosusb	NIZO1665	LGG	Human origin
L. salivarius	NIZO880		Human intestine
	NIZO881		Human intestine
L. salivarius subsp. salivarius	NIZO2938	NCIMB 8816	Human saliva
	NIZO2521	DSM 20555	Saliva
	NIZO2520	DSM 20554	Saliva
	NIZO2943	DSM 20492	Human saliva
L. lactis subsp. cremoris	NIZO42		N/A
	NIZO47		Starter
	NIZO57		N/A
	NIZO706		N/A
L. lactis subsp. diacetylactis	NIZO86		Starter
L. lactis subsp. lactis	NIZO2051		Raw-milk curd
	NIZO8	R5	N/A
	NIZO14		N/A
S. thermophilus	NIZO133		N/A
	NIZO2269		N/A
	NIZO122		Raw-milk cheese

^aThe LAB strains were routinely grown in preferred laboratory culture medium under anaerobic conditions (90% N₂, 5% H₂, and 5% CO₂). Streptococci and lactococci were grown in M17 medium (Oxoid, Hampshire, United Kingdom) supplemented with 1 % lactose (or glucose when mentioned) at 30°C and 42°C, respectively. Lactobacilli were grown in MRS medium (Merck, Darmstadt, Germany) at 37°C.

^bIncluded for reference purposes.

^cN/A, not known.

The rifampin-resistant LAB strains were separately cultured overnight in the presence of 50 μg/ml rifampin, washed, and mixed in a final volume of 30 ml of saline at a concentration of approximately 2 × 10⁸ CFU per strain. Ethical approval for human studies was given by the Commissie Mensgebonden Onderzoek regio Wageningen. Three subjects that were previously confirmed to lack rifampin-resistant oral bacteria held the mixture in their mouth for 1 min, gently washing the liquid around their oral cavity, after which the mixture of bacteria was spit out. Saliva, tongue scrapings, and tooth swabs were collected by the subjects 5 min, 15 min, 1 h, 4 h, 24 h, 13 days, and 28 days after administration. Tongue scrapers (DA Retail B.V., Zwolle, The Netherlands) were rinsed in 5 ml saline and swabs in 1 ml saline. The subjects did not consume any food, but were allowed to drink water, during the first 4 h after receiving the oral rinse, and subsequently no dietary or behavioral restrictions were imposed.

Enumeration of total rifampin-resistant bacteria was performed on standard medium containing 50 μg/ml rifampin (Fig. 1). The highest numbers of colonies from all three volunteers were recovered from saliva, ranging from 10⁷ CFU/ml 5 min after rinsing to 10⁵ to 10⁶ CFU/ml 4 h later. In saliva samples, the numbers of rifampin-resistant bacteria from subject 2 declined >10⁵-fold within the first 24 h, whereas the colony recovery in saliva samples from subjects 1 and 3 dropped only 10³-fold. Dental swabs consistently contained smaller amounts of LAB inoculants than the other samples, and tongue scrapings showed considerable variation among the subjects. Rifampin-resistant bacteria were still recovered 13 days after administration in the saliva from subjects 1 and 3 in concentrations of 5 × 10¹ and 7 × 10³ CFU/ml saliva and in subject 3 even after 28 days, indicating that in some individuals one or more of the administered strains display a very high level of persistence.

Fig. 1.

Total numbers of rifampin-resistant LAB recovered from the oral cavity at different times during the first 24 h after administration. The saliva (A), tongue (B), and teeth (C) of the three subjects (subject 1, diamonds; subject 2, squares; subject 3, triangles) enumerated independently. Limit of detection was 10 CFU/ml of sample.

From each subject, 30 rifampin-resistant bacterial isolates were selected on the basis of colony morphology, type of sample, and time point (mostly 24 h after administration) of the LAB strains. Six isolates were collected at the 13- and 28-day time points. Species identification was performed using V1-V3 16S rRNA gene sequencing (12) (see Table S1 in the supplemental material). Fifty-eight percent of the isolates were identified as being Lactobacillus fermentum, while only 12% of the strains in the oral rinse were L. fermentum. Also, strains of Lactobacillus salivarius and Lactobacillus (para)casei were recovered frequently among the isolates. This result is in agreement with other studies reporting that these species are commonly found in the normal oral microbiota (14). Isolates of Lactobacillus brevis, Lactobacillus delbrueckii, Lactococcus lactis, and Streptococcus thermophilus were not among the 96 isolates examined, suggesting that they are unable to form persistent populations in the mouth.

Two discriminative colony types of L. fermentum were isolated. GTG-5 PCR identification (16) showed that these represented L. fermentum NIZO1220 (flat, rough-edged colonies) and NIZO2930 (pink, large colonies) (Fig. 2). For L. salivarius, molecular typing according to GTG-5 PCR was not sufficient. RAPD4 (5′-AAGAGCCCGT-3′), M13 (5′-GAGGGTGGCGGTTCT-3′), and Box-A1R (5′-CTACGGCAAGGCGACGCTGACG-3′) PCRs assisted in the partial differentiation of the L. salivarius strains recovered from the subjects (Fig. 3). Six out of the nine L. salivarius oral isolates examined showed RAPD4 PCR patterns shared among L. salivarius strains NIZO880, NIZO881, and NIZO2938. The remaining L. salivarius isolates were likely strains NIZO2520 and/or NIZO2943. The diversity of L. salivarius in the recovered bacterial isolates suggests that L. salivarius strains commonly persist for extended periods in the oral cavity compared to the other species tested.

Fig. 2.

Dendrogram and GTG5 PCR fingerprints for comparison of L. fermentum strains included in the oral rinse and isolates from the oral cavity collected during the first persistence trial. For strains, NIZO numbers are denoted. Isolates are indicated by subject number and isolate number.

Fig. 3.

Dendrogram and PCR fingerprints for comparison of L. salivarius strains included in the oral rinse and isolates from the oral cavity collected during the first persistence trial. The comparison is based on the combined PCR fingerprints obtained by RAPD4, M13, and BOX-A1R.

In a second human study, rifampin-resistant L. fermentum NIZO1220 and L. salivarius NIZO2521 were administered in concentrations of 10⁹ CFU to the oral cavity of 5 subjects, and the persistence of these strains was followed over time, as described above. Surprisingly, rifampin-resistant colonies were recovered from the oral cavity of subject 1 prior to receiving the oral rinse. This subject was the same individual as subject 3 in the initial oral persistence trial. Identification by 16S rRNA gene sequencing and RAPD PCR methodology showed that this individual harbored at least two different strains of rifampin-resistant L. salivarius which were distinct from strain NIZO2521 (data not shown). A similar long persistence was reported for Lactobacillus rhamnosus GG that was identified in saliva from a female subject 5 months after the use of L. rhamnosus GG (21).

For at least 24 h after administration, the inoculated strains were found in amounts of 10² to 10⁵ CFU/ml saliva (Fig. 4). Thereafter, L. fermentum or L. salivarius strains ranged between 10 and 1,000 CFU/ml saliva at 2 and 5 days after administration and returned to baseline levels in each of the subjects within 15 days, although a high interindividual variation was observed.

Fig. 4.

Recovered total numbers of rifampin-resistant colonies at different time points in five subjects (subject 1, closed diamonds; subject 2, closed squares; subject 3, closed triangles; subject 4, open circles; subject 5, asterisks) and two sampling sites were enumerated independently (saliva [A] and tongue scrapings [B]). Limit of detection was 10 CFU/ml of sample.

L. fermentum NIZO1220 and L. salivarius NIZO2521 were individually counted in samples on the basis of colony morphology. Since subject 1 had rifampin-resistant bacteria in the mouth prior to taking the oral rinse, this subject was excluded from further analysis at the group level. In the majority of samples, L. fermentum NIZO1220 was recovered in 1- to 2-log-higher numbers than L. salivarius NIZO2521, although not always significant (Fig. 5). These findings confirm that L. fermentum NIZO1220 and L. salivarius NIZO2521 are LAB with relatively high persistence capacities in the human oral cavity.

Fig. 5.

Relative persistence (CFU/ml) of L. fermentum NIZO1220 (black bars) and L. salivarius NIZO2521 (striped bars) in the oral cavity of 4 healthy human subjects (subjects 2 to 5), as measured in saliva (A) and on tongue scrapings (B). No rifampin bacteria were recovered from subjects 2 to 5 before oral administration of the two candidate probiotic strains. Subject 1 harbored rifampin-resistant bacteria before administration and was therefore excluded from the analysis. Limit of detection was 10 CFU/ml of sample.

Previous studies evaluating individual strains have shown variable capacities of LAB to colonize the human mouth. Lactobacillus reuteri ATCC 55730 that was associated with an in vivo reduction of S. mutans (18) disappeared in almost 50% of subjects within 24 h (3). LGG was maintained in only 66% of the participating subjects after the first day of discontinuation of its intake (21). Our study is in line with these observations, since the same strains of the species L. rhamnosus and L. reuteri were included in our initial collection of strains. In contrast, S. salivarius K12 persisted in the human oral cavity for a period of up to 2 weeks (1).

Coaggregation is proposed as a mechanism by which oral bacteria adhere to each other and as a result may colonize persistently in biofilms in the host oral cavity (13). For example, the capacity of orally administered Weissella cibaria isolates to inhibit resident oral bacteria is proposed to be at least partially determined by the capacity of this bacterium to coaggregate with target strains, including Fusobacterium nucleatum, Treponema denticola, and Prevotella loescheii (11, 13).

To evaluate whether adherence to other oral bacteria might be a factor influencing the persistence characteristics of LAB in the mouth, the ability to adhere and coaggregate with oral bacteria was investigated for the 2 most persistent strains of L. fermentum and all 6 L. salivarius strains included in the oral rinse. Coaggregation capacity of lactic acid bacteria was performed with cultured representatives of common oral microorganisms that are implicated as causative agents of bad breath or caries (Table 2).

Table 2.

Strains of oral bacteria used in this study^a

Species	Strain identifier	Source
Porphyromonas gingivalis	HG66	ACTA, Amsterdam
Porphyromonas endodontalis	HG181	ACTA, Amsterdam
Prevotella intermedia	HG110	ACTA, Amsterdam
Prevotella melaninogenica	HG73	ACTA, Amsterdam
Peptostreptococcus anaerobius	HG578	ACTA, Amsterdam
Fusobacterium nucleatum	HG646	ACTA, Amsterdam
Tannerella forsythia	HG1245	ACTA, Amsterdam
Streptococcus mutans	UA 159	ACTA, Amsterdam
Streptococcus mutans	NIZO B1215	NIZO culture collection
Streptococcus mutans	C180-2	ACTA, Amsterdam

^aStreptococcus mutans was grown on M17 containing 1% glucose at 37°C. The other strains were grown in brain heart infusion medium (Merck, Darmstadt, Germany) at 37°C.

L. salivarius NIZO2520, NIZO2521, and NIZO2943 coaggregated with the majority of the target strains, with the exception of S. mutans and Prevotella melaninogenica (Table 3). Small aggregates indicated that L. salivarius NIZO2521 also coaggregated slightly but significantly with P. melaninogenica HG73 (see Fig. S1 in the supplemental material). In comparison, L. fermentum strains NIZO1220 and NIZO2930 and L. salivarius strains NIZO880, NIZO881, and NIZO2938 did not coaggregate with any of the oral strains. Possible explanations for the persistence of L. fermentum may be the ability to adhere to species that were not tested or directly to dental surfaces, e.g., by adhesion to salivary proteins. Indeed, in vitro assays revealed a considerable degree of variation of adherence of individual bacterial strains to salivary proteins (10), which indicates that coaggregation is not the sole mechanism by which bacteria can persist in the oral cavity.

Table 3.

Coaggregation of mixtures of Lactobacillus and oral bacteria

Strain	Scoreb for L. salivarius:
	NIZO2521	NIZO2520	NIZO2943
Controla	0	0	0
Fusobacterium nucleatum HG646	3	3	3
Peptostreptococcus anaerobius HG578	3	3	3
Porphyromonas endodontalis HG181	2	2	2
Porphyromonas gingivalis HG66	4	4	4
Prevotella intermedia HG110	3	3	3
Prevotella melaninogenica HG73	0	0	0
S. mutans B1215	0	0	0
Tannerella forsythia HG1245	4	3	3

^aL. salivarius without oral bacteria.

^bData are provided for coaggregation after 2 h of incubation. These results are consistent with the findings observed at 4 h and 24 h (data not shown). Scores are based on visual inspection, using the following scoring criteria (5): 0, no visible aggregates in the cell suspension; 1, small uniform coaggregates in suspension; 2, definite coaggregates easily seen but suspension remained turbid; 3, large coaggregates which settled rapidly, leaving some turbidity in the supernatant fluid; 4, clear supernatant fluid and large coaggregates which settled immediately.

One important caveat which might prevent the use of Lactobacillus species as oral probiotics is that members of this genus have also been associated with childhood caries because of their strong acidifying characteristics, although their presence was not sufficient to explain all cases of caries (8). Therefore, probiotic characteristics of the selected strains should be carefully monitored in vivo, e.g., for the absence of a contribution to dental decay, and not only based on in vitro characteristics.

In conclusion, the ability of a bacterial strain to persist in the oral cavity is likely to support oral probiotic efficacy. The approach we presented here can serve as an initial step in the selection of candidate probiotic strains aiming to promote oral health. L. fermentum and L. salivarius strains display the best extended oral persistence relative to other LAB. Further evaluation of these strains should examine their effects on the composition and activity of the endogenous oral microbiota and should be complemented with determination of the possible consequences for certain health parameters, including exhaled volatile sulfur compounds (VSC), reduced levels of S. mutans, or other clinically relevant characteristics.