In the present study, Lactobacillus spp. collected from the buccal mucosa, tongue and plaque of 21 patients treated for cancer in the head and neck region were identified to species level and their acid-producing capacity, and anti-microbial activity was determined. The most prevalent species were L. paracasei, L casei/rhamnosus and L. fermentum of which L. paracasei was the most prominent species at 12 months post treatment. The highest proportion of isolates that showed strong acid-producing capacity using sugars was found for L. paracasei of which four showed strong acid-producing capacity using sorbitol. The proportion that showed weak acid-production capacity using xylitol was highest among L. casei/rhamnosus both at six, and 12 months post treatment. Anti-microbial activity was seen for one isolate against S. aureus and for 42 isolates (71%) against S. mutans. All but one patient had Lactobacillus with antimicrobial activity against S. mutans at least at one time-point. Antimicrobial activity was most prevalent among L. paracasei (81%).
In our previous study, using the same method for identification, the most frequently isolated species from plaque in irradiated patients 3-5 years post treatment were L. fermentum 29%, L. casei 22%, L. rhamnosus 19% and L. paracasei 7% . In the present study, a slightly lower prevalence of L. casei/rhamnosus was found, 33%. The prevalence of L. paracasei was much higher, 53%, compared with the previous study. There was also a difference in the prevalence of L. fermentum since no L. fermentum were isolated from the plaque at 12 months post RT in the present study. A possible explanation to the discrepancies might be that in the former study several isolates were collected from either anterior or posterior tooth surfaces , while in the present study only 1-2 isolates was collected from samples where plaque from four sites had been pooled.
To the best of our knowledge, there are no other previous reports on the prevalence of different Lactobacillus species in the oral cavity and their acid-producing capacity at different time-points in relation to treatment for cancer in the head and neck region.
Acid-producing capacity and time-point
The results of the present study suggest that there might be differences in acid-producing capacity among Lactobacillus isolated at different time-points. For some species a lower proportion showing strong acid-producing capacity was seen post treatment, while others showed a higher proportion. With the method used for identification it is not possible to determine whether the same strain was isolated at different time-points or if it were other strains. It is possible that the persistent acid environment due to the reduced salivary secretion rate enhanced the growth of certain Lactobacillus strains and species more than others.
Acid-producing capacity and species
In accordance with our previous study , the proportion of L. fermentum isolates that showed strong acid-producing capacity using sugars and sugar-alcohols was lower compared with L. paracasei and L. casei/rhamnosus. A high proportion of the L. paracasei isolates was able to decrease the pH to < 5.0 using sucrose, glucose and fructose, which is also in accordance with the results of our previous study . It has been shown that it takes a longer time for the pH to reach 5.5 using glucose for L. fermentum, 4.18 h, compared with L. casei/paracasei, 2.87 h, and L. rhamnosus 2.27 h [19 ]. Also, the final pH is higher for L. fermentum, pH 4.58, compared with L. casei/paracasei, pH 4.02 and L. rhamnosus pH 3.89 . In our previous study, the proportion of strains that were able to produce acids using sorbitol was highest among L. paracasei (49%), while L. fermentum rarely had that ability . These results are also in line with those in the present study where 19% of the L. paracasei isolates showed a strong acid-producing capacity using sorbitol and 66% weak; while only 20% of the L. fermentum isolates had this ability (Fig 2b).
In the present study, anti-microbial activity was determined on agar plates where growth inhibition of the test microorganism was registered. This method has been used for the determination of the anti-microbial activity for enterococcus strains at out laboratory . Lactobacillus spp. can have several activities or abilities, which can contribute to inhibition of other species such as bacteriocins, bi-products of fermentation like lactic acid and hydrogen peroxide, or their ability to reduce the pH . Due to the method used for the determination of possible inhibition in the present study, no information about which activities or abilities that were responsible for the effect can be obtained. Further studies are therefore needed to elucidate the mechanisms involved in this inhibitory effect.
All but one patient had Lactobacillus with anti-microbial activity against S. mutans at least at one time-point. A high proportion of the Lactobacillus isolates, 71%, had an anti-microbial activity against S. mutans of which 3% showed strong activity and 68% weak. High proportions of Lactobacillus with anti-microbial activity against S. mutans have previously been reported also in other studies [6, 16, 21]. Twenty-five percent of the Lactobacillus isolates collected from caries-free and caries-active children and young adults were able to completely inhibit S. mutans, while 75% showed no or low inhibition . Among Lactobacillus isolates collected from both healthy persons and periodontitis patients, the proportion with ability to inhibit S. mutans was 69%  and in another study including 67 Lactobacillus isolates, 55% could inhibit S. mutans . In the present study, most isolates showed weak anti-microbial activity, thus a possible inhibitory effect on S. mutans in vivo is questionable.
Lactobacillus spp. from patients without caries experience have been shown to have a higher inhibitory effect on S. mutans than Lactobacillus from patients with arrested or active caries . In the present study, we found no clear correlation between presence of Lactobacillus spp. with anti-microbial activity against S. mutans, and growth of S. mutans. No correlation with caries status was found. This lack of correlation may be due to the relatively low number of patients included in the present study. Other plausible explanations may be that also other bacteria are involved in caries development like for example Scardovia wiggsiae and Bifidobacterium dentium [23, 24]. In addition, other factors are involved in the caries process such as the amount and quality of saliva, intake frequency of easily fermentable carbohydrates, oral hygiene level, and fluoride exposure .
Species able to inhibit S. mutans have previously been reported among L. plantarum, L. paracasei, L. salivarius, L. rhamnosus and L. fermentum, but the proportion of the isolates having this ability were not shown . Similar results have also been shown by Simark-Mattsson et al.,  where the highest proportions of the strain able to inhibit S. mutans were found for L. paracasei (47% of the strains), L. plantarum (30%), and L. rhamnosus (23%). In the present study, the highest proportion of isolates with anti-microbial activity against S. mutans was also found among L. paracasei, 81%. However, in the present study a high proportion of L. fermentum also showed anti-microbial activity, 70%, while the one L. fermentum included in the study by Simark-Mattsson et al.,  could not inhibit S. mutans. A plausible explanation for the divergent results is that different populations were included, Simark-Mattsson et al.,  isolated Lactobacillus from children and young adults, while in the present study Lactobacillus were isolated from patients with a mean age of 59 years who had undergone treatment for cancer in the head and neck region, and therefore presented a specific oral environment.
Only one of the 59 isolates, an L. acidophilus strain collected from the tongue at 12 months post treatment, showed an anti-microbial activity against S. aureus and the patient showed no growth of S. aureus at any of the sampling sites or time-points. Among the patients, the detection frequencies of S. aureus were 57% at six months post treatment and 43% at 12 months post treatment. It is not unusual that L. acidophilus have this anti-bacterial activity against S. aureus, all of eight previously tested strains were able to inhibit S. aureus and seven of them had a bactericidal effect .
In the present study, none of the isolates showed an anti-microbial activity against C. albicans, which is in congruence with Köll et al., , who reported that none of the 67 Lactobacillus isolates from healthy persons had any anti-microbial activity against C. albicans. In another study where the anti-microbial activity of Lactobacillus isolates from saliva of healthy volunteers was tested, it was found that two out of four L. paracasei sp paracasei were able to inhibit the growth of two out of three C. albicans strains. Furthermore one of two L. rhamnosus strains inhibited two out of the three C. albicans strains . It is difficult to explain the divergent results, but it is possible that the ability to inhibit C. albicans is strain-specific. Both Lactobacillus and C. albicans are frequently found in patients after cancer treatment , which implies that the Lactobacillus that these patients harbour do not inhibit C. albicans. Another explanation might be that the acidic environment that is present in many subjects with reduced salivary secretion rate after cancer treatment is a stronger promoting factor for the growth of C. albicans. C. albicans were detected in 13 of the 21 patients (62%) at six months post treatment and in 15 of 21 (71%) at 12 months post treatment.
None of the isolates in the present study showed an anti-microbial activity against E. faecalis. Two probiotic strains, an L. plantarum and an L. rhamnosus have previously been shown to inhibit an E. faecalis strain . It is possible that the ability to inhibit or have an anti-microbial activity against E. faecalis is not a common trait among wild Lactobacillus strains, especially not among oral Lactobacillus since enterococci are not part of the resident oral microflora. Enterococci were detected in two of the 21 patients (9%) at six months post treatment and in only one patient at 12 months post treatment.
With the method used L. casei and L. rhamnosus could not be distinguished since their patterns on the gel was identical. To differentiate between these two species, they could have been subjected to sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) for whole cell protein analysis [9, 30]. However, little is known about differences between these two species and it is unlikely that such information would have added much to the results.