Department of Pediatric Hepatology and Gastroenterology, Saiseikai Yokohamashi Tobu Hospital, Kanagawa, 230-8765, Japan.
Department of Pediatrics, Yokohama Minami Kyousai Hospital, Kanagawa, 236-0037, Japan.
Sci Rep. 2018 Apr 3;8(1):5480. doi: 10.1038/s41598-018-23870-w.
Primary sclerosing cholangitis (PSC) is a liver disease known for its frequent concurrence with inflammatory bowel disease. Dysbiosis of the gut microbiota in PSC was reported in several studies, but the microbiological features of the salivary microbiota in PSC have not been established. Here we compared the salivary microbial communities of 24 pediatric-onset PSC patients, 16 age-matched ulcerative colitis (UC) patients, and 24 healthy controls (HCs) by analyzing the bacterial 16S rRNA gene sequence data. The species-richness (α-diversity) showed no significant between-group differences, whereas the overall salivary microbiota structure (β-diversity) showed significant differences among the three groups. Taxonomic assignment revealed that the PSC salivary microbiota were characterized by significant decreases in the abundance of Rothia and Haemophilus compared to the HC group, and significantly decreased Haemophilus and increased Oribacterium compared to the UC group. By combining the genera selected by the random forest algorithm in machine learning, followed by confirmation with 10-fold cross-validation, we were able to distinguish the PSC group from the HC group with the area under the curve (AUC) of 0.7423, and from the UC group with the AUC of 0.8756. Our results indicate the potential of salivary microbiota as biomarkers for a noninvasive diagnosis of PSC.
原发性硬化性胆管炎(PSC)是一种已知常伴有炎症性肠病的肝脏疾病。几项研究报道了 PSC 肠道微生物群落的失调,但 PSC 唾液微生物群落的微生物特征尚未确定。在这里,我们通过分析细菌 16S rRNA 基因序列数据,比较了 24 名儿科起病的 PSC 患者、16 名年龄匹配的溃疡性结肠炎(UC)患者和 24 名健康对照(HC)的唾液微生物群落。物种丰富度(α多样性)在组间无显著差异,而唾液微生物群落结构(β多样性)在三组间有显著差异。分类学分配表明,与 HC 组相比,PSC 唾液微生物群落的 Rothia 和 Haemophilus 丰度显著降低,与 UC 组相比,Haemophilus 丰度降低,Oribacterium 丰度增加。通过机器学习中的随机森林算法选择的属结合 10 倍交叉验证进行验证,我们能够以 0.7423 的曲线下面积(AUC)将 PSC 组与 HC 组区分开来,以 0.8756 的 AUC 将 PSC 组与 UC 组区分开来。我们的结果表明唾液微生物群落作为 PSC 非侵入性诊断的生物标志物具有潜力。