Sequence data

A total of 2,649,963 merged sequenced reads were obtained from all samples of both groups. The clean tags ranged between 35,446 and 71,753 reads after quality control. After removing chimera sequences, the valid tags ranged between 29,387 and 68,865 reads, with an average of 61,627 sequences for each sample. Finally, 120 to 1048 ASVs were identified. OLP and HC groups shared 981 ASVs. Moreover, the rarefaction curve tended to be flat, indicating that the 16 S rRNA gene sequences identified in this study represented the majority of the bacteria present in saliva samples.

Species richness and diversity of oral microbiota in OLP and HC groups

Alpha diversity analysis, based on Chao 1 and Shannon indexes, did not reveal significant differences in species richness and diversity, respectively, of oral microbiota between OLP and HC groups (P > 0.05; Fig. 1A, B).

Fig. 1

Alpha and Beta diversity analysis of microbiota in saliva samples of OLP and HC groups. (A) Chao 1 (B) Shannon index of diversity. P > 0.05; (C)Principal coordinates analysis (PCoA) plot constructed using the Binary Jaccard algorithm. P = 0.041

The PCoA plot revealed obvious separation between HC and OLP groups, indicating that the overall structure of the bacterial community in the two groups was significantly different (P = 0.041). Moreover, the OLP patients exhibited distinct discrete characteristics, indicating extensive heterogeneity (Fig. 1C). PC1 explained 7.36% variability, whereas PC2 explained 4.16% variability.

Phylum- and genus-level identification of saliva microbiota in OLP and HC groups

All operational taxonomic units obtained from both groups were clustered into 9714 ASVs, representing 42 phyla, 105 classes, 243 orders, 390 families, and 755 genera.

At the phylum level, 98% sequences belonged to Proteobacteria, Firmicutes, Bacteroidota, Fusobacteria, Actinobacteria, and Patescibacteria in both OLP and HC groups. Moreover, Patescibacteria exhibited higher abundance in the HC group than in the OLP group (Fig. 2A).

Fig. 2

Analysis of relative abundance of microbiota in saliva sample of HC and OLP groups. Donut chart of main communities at (A) phylum and (B) genus levels in HC (inner ring) and OLP (outer ring). *P < 0.05; (C) LDA (threshold was set at 3) using LefSe algorithm

At genus level, the relative abundance of 15 genera was > 1% in both OLP and HC groups, with Neisseria and Streptococcus accounting for approximately 20%, and Haemophilus and Prevotella accounting for approximately 10%. The abundance between two groups showed no significant difference (Fig. 2B).

LDA using LefSe revealed differences between OLP and HC groups at different taxonomic levels, including 3 phyla, 4 classes, 6 orders, 10 families. and 16 genera. Moreover, compared with those in the HC group, the abundance of Patescibacteria (phylum level), Gracilibacteria (class level), and Absconditabacteriales_SR1 (order, family, and genus levels) was significantly decreased in the OLP group, whereas that of Achromobacter and Citrobacter at the genus level was significantly increased (Fig. 2C).

Abundance of LAB in OLP and HC groups

Next, we investigated the composition of LAB at the genus level in OLP and HC groups. The abundance of Streptococcus, Selenomonas, Lactobacillus, Abiotrophia, and Enterococcus did not exhibit significant differences between OLP and HC groups, whereas the abundance of Lactococcus was significantly lower in the OLP group than in the HC group. Moreover, the abundance of Lactococcus lactis was significantly decreased in OLP patients compared with that in HC participants (Fig. 3). The real-time PCR of swab sample in cohort 2 further validated that Lactococcus lactis presented lower amount in patients with OLP rather than HC with statistically significance (Fig. 4).

Fig. 3

Relative abundance of LAB in saliva sample of HC and OLP groups. (A) lactic acid bacteria at the genus level and (B) Lactococcus lactis in HC and OLP groups. *P < 0.05

Fig. 4

Relative quantification of Lactococcus lactis in swab sample of HC and OLP groups. *P < 0.05

Potential impact of LAB on microbiome composition shift in OLP

Further, we investigated the co-occurrence of LAB and other bacterial genera in both groups. Lactococcus and Lactococcus lactis were significantly negatively related with Fusobacterium (ρ=-0.377, P = 0.013; ρ=-0.368, P = 0.015). Moreover, Aggregatibacter (ρ=-0.352, P = 0.021; ρ=-0.336, P = 0.028) showed similar relevance. Streptococcus was significantly negatively correlated with Fusobacterium (ρ=-0.594, P < 0.001), Alloprevotella (ρ=-0.510, P < 0.001), Prevotella (ρ=-0.331, P = 0.030), and Leptotrichia (ρ=-0.329, P = 0.031). However, Gemella was positively associated with Streptococcus (ρ = 0.470, P = 0.001); (Fig. 5A; Appendix 1).

Fig. 5

Relationship of certain bacteria in saliva sample (A) Heatmap showing correlation between lactic acid bacteria (LAB) and other bacterial genera. *P < 0.05, **P < 0.01, ***P < 0.001 (B) Correlation between key bacteria and REU score, which is indicative of oral lichen planus disease severity

Relationship between LAB and clinical features of OLP patients

The association between LAB and other key microbes (the differential flora between OLP and HC or LAB related flora mentioned before) with OLP disease severity was analyzed using Spearman’s rank correlation coefficient. LAB, namely Streptococcus, Lactococcus, and Lactococcus lactis, were negatively correlated with REU score. However, the non-LAB genera, Fusobacterium, Alloprevotella, Aggregatibacter, Leptotrichia, and Prevotella were positively correlated with REU score (Fig. 5B).