Recurrent aphthous stomatitis (RAS) is a common oral mucosal disorder for which no curative treatment is available. We previously reported that decreased Streptococcus salivarius and increased Acinetobacter johnsonii on the oral mucosa are associated with RAS risk. The purpose of this study was to identify antibiotics that selectively inhibit A. johnsonii but minimally inhibit oral mucosal commensals. S. salivarius KCTC 5512, S. salivarius KCTC 3960, A. johnsonii KCTC 12405, Rothia mucilaginosa KCTC 19862, and Veillonella dispar KCOM 1864 were subjected to antibiotic susceptibility test using amoxicillin, cefotaxime, gentamicin, clindamycin, and metronidazole in liquid culture. The minimal inhibitory concentration (MIC) was defined as the concentration that inhibits 90% of growth. Only gentamicin presented a higher MIC for A. johnsonii than MICs for S. salivarius and several oral mucosal commensals. Interestingly, the growth of S. salivarius increased 10~200% in the presence of sub-MIC concentrations of gentamicin, which was independent of development of resistance to gentamicin. In conclusion, gentamicin may be useful to restore RAS associated imbalance in oral microbiota by selectively inhibiting the growth of A. johnsonii but enhancing the growth of S. salivarius.

ReαlITent aphthous stomatitis (RAS) appears to be one of the most common oral diseases. However, the defmitive etiology of RAS is not well established, though many etiologic faαors have been suggested and examined. The present study was petformed to investigate the association between HIA and Korean recurrent aphthous stomatitis pa디ents. πle proportions of class 1 and class II HIA types expressed in 49 Korean subjects affected by recurrent aphthous stomatitis (RAS) and in 50 healthy controls were deteffi1Í11ed by microlymphocytotoxicity test. 까le sig띠ficance of the data was analyzed by chi-square test and Fisher' s exact test. πle proportions of HIA-Cw1 and -DR9 antigens were significantly higher in RAS pa디ents (p(0.05), whereas those of HIA-DR4 and -DQ2 antigens were significantly lower (p(0.05). 까le odds ratio (OR) were 2.8 for HIA-Cw1 and 2.7 for -DR9. πle etiologic fractions (EF) were 0.262 and 0.193, respectively. The results s맹gest that, in Koreans, there Í$ a sig띠ficant relation between HIA antigens and RAS. Genetic faαors ， reflected in the HIA type, may play an in1portant role in the development of RAS.