https://doi.org/10.1177/00220345251370995 Journal of Dental Research 1 –9 © The Author(s) 2025 Article reuse guidelines: sagepub.com/journals-permissions DOI 10.1177/00220345251370995 journals.sagepub.com/home/jdr Research Reports: Clinical Introduction Lipopolysaccharide (LPS) is a crucial virulence factor of gram- negative bacteria, constituting 10% to 15% of bacterial outer membrane molecules and accounting for up to 75% of its total surface area (Erridge et al. 2002). The primary role of LPS is to preserve the cell shape and structural integrity of the outer membrane, as the cellular wall is the first line of defense against the immune system of the host. LPS consists of 3 distinct domains, namely, O-antigen, the immunodominant, hydrophilic portion pointing out of the cell membrane; the core oligosaccharide, embedded in the cytoplasmic side of the inner membrane and influencing cell wall phenotype, poring, efflux, and responses against the environment; and lipid A, which is anchored in the outer membrane and is the primary immunos- timulatory component of LPS (Valvano 2015). LPS quantification is based on the biological activity of the lipid A portion of the molecule. The gold standard method still relies on the Limulus amebocyte lysate assay (LAL) and its ability to agglutinate upon encountering LPS. Thus, determin- ing LPS concentration is based on an indirect assay that mea- sures host innate immunity response, namely, the amebocytes enzymatic cascade triggered by LPS and eventually resulting in coagulation, immobilization, and neutralization of the invad- ing gram-negative bacteria (Munford 2016). Translocation of LPS to the bloodstream causes endotox- emia, which is associated with low-grade inflammation, 1370995 JDRXXX10.1177/00220345251370995Journal of Dental ResearchDeterminants of Saliva and Serum Lipopolysaccharide Activity research-article2025 1Department of Oral and Maxillofacial Diseases, University of Helsinki, Helsinki, Finland 2Neurology, Helsinki University Hospital and University of Helsinki, Helsinki, Finland 3Centre for Host-Microbiome Interactions, Faculty of Dentistry, Oral & Craniofacial Sciences, King’s College London, UK 4Institute of Dentistry, University of Turku, Turku, Finland 5Department of Computing, University of Turku, Turku, Finland 6School of Medicine, Institute of Dentistry, University of Eastern Finland, Kuopio, Finland A supplemental appendix to this article is available online. Corresponding Authors: M. Manzoor, Department of Oral and Maxillofacial Diseases, University of Helsinki, PO Box 63, Helsinki, 00014, Finland. Email: muhammed.allippara@helsinki.fi P.J. Pussinen, Department of Oral and Maxillofacial Diseases, University of Helsinki, PO Box 63, Helsinki, 00014, Finland and School of Medicine, Institute of Dentistry, University of Eastern Finland, Kuopio, Finland. Email: pirkko.pussinen@helsinki.fi Oral Microbial Determinants of Saliva and Serum Lipopolysaccharide Activity M. Manzoor1 , J. Putaala2, S. Zaric3 , J. Leskelä1 , A. Dong3, E. Könönen4, L. Lahti5, S. Paju1, and P. J. Pussinen1,6 Abstract Lipopolysaccharide (LPS) is a virulence factor of gram-negative bacteria, and endotoxemia or translocation of LPS in serum plays a significant role in oral and systemic pathologies. The contribution of the oral microbiome composition to saliva LPS activity and endotoxemia remains unclear. We investigated whether salivary and serum LPS levels are associated with oral microbiome diversity, taxonomic profiles, and functional characteristics. The oral microbiome was analyzed using metagenomic sequencing of saliva from 298 individuals enrolled in a multicenter case-control study, SECRETO (NCT01934725). Serum and salivary LPS activities were measured, and multiple linear regression models were fitted to identify the microbial taxa that predicted LPS levels. MaAsLin2 (Microbiome Multivariable Associations with Linear Models) was used to determine the associations of microbial functional features and LPS levels. Salivary alpha diversity was positively associated with serum LPS but negatively associated with salivary LPS, smoking, and antibiotic use in the preceding 1 to 6 mo. Community composition (beta diversity) differed between the salivary LPS tertiles (P = 0.001) but not between serum LPS tertiles. In total, 10 oral taxa associated with serum LPS tertiles and 59 with salivary LPS tertiles were identified. Prevotella, Neisseria, Leptotrichia, and Porphyromonas had significant positive associations with salivary LPS, whereas Fusobacterium had a negative association. Among these genera, Prevotella sp. E13_17, P. gingivalis, L. wadei, and F. nucleatum were the species with the strongest associations. Among the 1,016 oral microbiome metabolic features, several were linked to the biosynthesis of LPS, lipid A, and O-antigen pathways. The oral microbiome composition was strongly associated with salivary LPS activity in addition to weaker links to serum LPS. Oral microbiota–derived LPS activity in saliva was associated with microbial metabolism characterized by the predominance of proliferation and biosynthesis pathways. Our study indicates that dysbiosis of the oral microbiome is a source of increased salivary and serum LPS activity. Keywords: oral microbiome, metagenomic sequencing, gram-negative bacteria, endotoxemia, biomarkers, periodontitis 2 Journal of Dental Research 00(0) activation of the complement system and the contact activation pathway, and metabolic alterations (Pussinen et al. 2022). Accordingly, LPS may be a potent molecular mediator between periodontitis and cardiometabolic disorders (Liljestrand et al. 2017), especially since LPS may enter the circulation via inflamed periodontal tissues (Geerts et al. 2002; Lockhart et al. 2008). Severe forms of periodontitis with a high grade of pro- gression are associated with elevated serum LPS levels (Shaddox et al. 2011; Leira et al. 2021), which decrease after periodontal treatment (Pussinen et al. 2004). Although the gut microbiome has long been considered the main source of endotoxemia, the oral microbiome may also contribute to the systemic inflammatory burden (Mukherjee et al. 2025). Furthermore, the mouth is the starting point of the gastrointes- tinal tract, and the 2 microbiomes are continuously connected to each other. Nevertheless, the composition of the oral micro- biome in endotoxemia has not been investigated previously. In patients with unstable periodontitis, 4.9% of serum LPS variation was explained by saliva LPS levels (Liljestrand et al. 2017). Although saliva LPS may be used as a biomarker of gram-negative microbiome dominance and thus periodontitis, LPS has not been investigated as comprehensively as other bac- terial biomarkers (Dong et al. 2025). Saliva LPS activity is posi- tively associated with periodontal diagnosis (Zaric et al. 2022; Leskelä et al. 2024); alveolar bone loss; a cluster of subgingival oral, oropharyngeal, or nasopharyngeal species (Liljestrand et al. 2017); and cumulative risk score of periodontitis (Liukkonen et al. 2020). Functional profiling of the oral micro- biome revealed enriched LPS biosynthesis in periodontitis and its related systemic complications (Zhao et al. 2022; Oh et al. 2023). However, no studies have directly established the compo- sition of the oral microbiome regarding salivary LPS levels. We investigated the composition and functional features of the oral microbiome that are associated with salivary LPS activity and endotoxemia. Materials and Methods Study Subjects and Recruitment The present study, SECRETO Oral (Leskelä et al. 2024), is a substudy of the international, multicenter SECRETO case-con- trol study (Putaala et al. 2017). All participants provided written informed consent before participating in the study. This study was conducted in accordance with the STROBE guidelines and the Declaration of Helsinki. This study was approved by the Helsinki and Uusimaa Hospital District (362/13/03/00/2012) and the local Ethics Committees at each recruiting center. At the time of recruitment, all participants underwent a thorough structured interview, and a detailed clinical history was obtained from medical records. Clinical Oral Examination All oral examinations were conducted between April 2014 and February 2020 by a single periodontist (S. Paju). Periodontal diagnosis was defined according to the 2017 classification (Chapple et al. 2018), and the participants were divided into 3 groups: periodontal health, gingivitis, and periodontitis. Clinical moderate or extensive caries lesions (levels 3–6) were detected and classified using the International Caries Detection and Assessment System. Sample Collection and Metagenomic Sequencing The procedures for saliva sample collection, processing, DNA extraction, and sequencing have been previously described (Manzoor, Leskelä, Pietiäinen, Martinez-Majander, Könönen, et al. 2024; Manzoor, Leskelä, Pietiäinen, Martinez-Majander, Ylikotila, et al. 2024; Manzoor et al. 2025). In brief, DNA was extracted from stimulated saliva samples followed by metage- nomic sequencing using an Illumina NovaSeq 6000. LPS measurement LPS activity in serum was measured using the Limulus amebo- cyte lysate assay coupled with a chromogenic substrate (LAL) (Hycult Biotech, The Netherlands) and in saliva using the fluo- rescent recombinant Factor C assay (EndoZyme II Go Plate, bioMérieux) with dilutions of 1:5 for serum and 1:1,000 for saliva in endotoxin-free water. LPS activity within the sample type was categorized into tertiles to obtain low, medium, and high categories. The cutoff levels for serum were 0.331 EU/mL and 0.473 EU/mL and for saliva were 6,848 EU/mL and 14,453 EU/mL. Statistical Analysis All analyses were performed using R (v. 4.2.2). We computed alpha diversity metrics (Shannon and Inverse Simpson indices) and explored their relationship with LPS using Spearman cor- relations and linear regression models. Permutational multi- variate analysis of variance (PERMANOVA) tests used the Bray–Curtis dissimilarity between LPS tertiles, and the princi- pal coordinate analysis plot was used to visualize the beta diversity at the species level. We performed linear discriminant analysis effect size (LEfSe) analysis (Segata et al. 2011) at dif- ferent taxonomic levels using LPS tertiles, applying Kruskal– Wallis and Wilcoxon tests (P < 0.05) for significance between LPS tertiles within the 2 sample types. Functional profiles were generated for each sample using HUMAnN 3 using the default settings (Beghini et al. 2021). Differential abundance analysis of the functional features was performed using MaAsLin 2 (Mallick et al. 2021). Multiple testing correction using the false discovery rate was applied to control for false positives. See the supplementary methods for further details. Results Characteristics of the Study Population A total of 298 participants were included in this study. The demographic and clinical characteristics of the participants are Determinants of Saliva and Serum Lipopolysaccharide Activity 3 summarized in Table 1. Sex, age, education level, smoking his- tory, alcohol consumption, obesity, diabetes, hypertension, bleeding on probing, periodontal pockets ≥4 mm, and peri- odontal health status distribution were similar across the serum and salivary LPS tertiles. History of stroke and antibiotic use during the preceding 1 to 6 mo differed between serum LPS tertiles, whereas only the prevalence of caries differed between salivary LPS tertiles. Impact of Alpha Diversity and Oral Microbiome Composition on LPS No significant correlation was observed between serum LPS levels and alpha diversity metrics (Shannon index, R2 = 0.072, P = 0.229; inverse Simpson index, R2 = 0.029, P = 0.625) (Fig. 1A). In contrast, a significant negative correlation was found between salivary LPS and Shannon (R2 = −0.124, P = 0.033) and inverse Simpson diversity (R2 = −0.162, P = 0.004, Fig. 1B). We found significant differences in Shannon (P = 0.047) and inverse Simpson (P = 0.009) diversity between the salivary LPS tertiles but not for serum (Fig. 1C, D). Both Shannon (P = 0.023) and inverse Simpson (P = 0.004) differed between salivary LPS tertiles 2 and 3. Linear regression analyses showed that serum LPS was positively associated with Shannon diversity index (P = 0.016), explaining 8.42% of the variance. Salivary LPS was negatively associated with inverse Simpson diversity (P = 0.017), while smoking and antibiotic use negatively affected both diversity indices. The model explained 9.0% of the variance in inverse Simpson diversity (Appendix Table 1). PERMANOVA showed that the microbiome composition did not significantly differ across the serum LPS tertiles (F = 0.80, R2 = 0.005, P = 0.675) (Fig. 2A). In contrast, the microbiome composition differed significantly between the salivary LPS tertiles (F = 3.60, R2 = 0.024, P = 0.001) (Fig. 2B). The differences were highly significant between all tertiles. Taxonomic Composition and Differential Abundance The oral microbiota was dominated by the phyla Bacilliota, Actinomycetota, Pseudomonadota, Bacteroidota, and Table 1. Characteristics of the Study Participants. Serum Tertiles (n = 298) Saliva Tertiles (n = 298) Characteristic Low (n = 100) Medium (n = 99) High (n = 99) P Value Low (n = 100) Medium (n = 99) High (n = 99) P Value Male sex 55 (55.0) 61 (61.6) 60 (60.6) 0.592 51 (51.0) 65 (65.7) 60 (60.6) 0.102 Mean (SD) age, y 40.1 (7.7) 40.3 (7.6) 39.2 (7.9) 0.576 39.8 (7.8) 39.0 (8.0) 40.7 (7.2) 0.315 Education 0.598 0.789 Primary or secondary 46 (46.0) 41 (41.4) 39 (39.4) 45 (45.0) 39 (39.4) 42 (42.4) or higher 52 (52.0) 58 (58.6) 58 (58.6) 55 (55.0) 58 (58.6) 55 (55.6) Smoking, ever 42 (42.0) 45 (45.5) 42 (42.4) 0.866 47 (47.0) 39 (39.4) 43 (43.4) 0.556 Diagnosed diabetes 6 (6.0) 3 (3.03) 3 (3.03) 0.609 3 (3.0) 6 (6.1) 3 (3.03) 0.538 Heavy alcohol consumption 17 (17.0) 20 (20.2) 17 (17.7) 0.840 21 (21.0) 17 (17.2) 18 (18.2) 0.773 Hypertension 23 (23.0) 19 (19.2) 20 (20.2) 0.790 18 (18.0) 20 (20.2) 24 (24.24) 0.546 Obesity 43 (43.0) 45 (45.5) 53 (53.5) 0.298 41 (41.0) 46 (46.5) 54 (54.6) 0.157 History of stroke 58 (58.0) 56 (56.6) 32 (32.3) <0.001 42 (42.0) 49 (49.5) 55 (55.6) 0.159 Antihypertensive medicationa 5 (5.0) 6 (6.1) 3 (3.03) 0.625 2 (2.0) 6 (6.1) 6 (6.1) 0.296 Statin medicationa 4 (4.0) 3 (3.03) 6 (6.1) 0.605 3 (3.0) 5 (5.1) 5 (5.1) 0.730 Antibiotic useb 10 (10.0) 11 (11.1) 23 (23.2) 0.014 11 (11.0) 13 (13.1) 20 (20.2) 0.160 Caries 36 (36.0) 41 (41.4) 41 (41.4) 0.666 35 (35.0) 49 (49.5) 34 (34.34) 0.048 Bleeding on probing (%) 0.400 0.328 <30 29 (29.0) 22 (22.2) 32 (32.3) 35 (35.0) 24 (24.24) 24 (24.24) 30–50 53 (53.0) 49 (49.5) 45 (45.5) 49 (49.0) 48 (48.5) 50 (50.5) >50 17 (17.0) 25 (25.3) 21 (21.2) 16 (16.0) 24 (24.24) 23 (23.2) 0.835 0.395 Number of all periodontal pockets (≥4 mm) 0–2 35 (35.0) 31 (31.3) 34 (34.3) 36 (36.0) 32 (32.3) 32 (32.3) 3–10 47 (47.0) 42 (42.4) 45 (45.5) 42 (42.0) 41 (41.4) 51 (51.5) ≥11 17 (17.0) 23 (23.2) 19 (19.2) 22 (22.0) 23 (23.2) 14 (14.1) Periodontal status 0.559 0.100 Healthy 23 (23.0) 16 (16.2) 25 (25.3) 29 (29.0) 18 (18.2) 17 (17.2) Gingivitis 50 (50.0) 56 (56.6) 51 (51.5) 53 (53.0) 54 (54.6) 50 (50.5) Periodontitis 26 (26.0) 24 (24.2) 21 (21.2) 17 (17.0) 24 (24.24) 30 (30.3) Data are expressed as n (%) unless otherwise indicated. Missing data: diagnosed diabetes n = 1, bleeding on probing n = 5, number of all periodontal pockets n = 5, periodontal status n = 6. Bold indicates P < 0.05. aCurrent medication registered at recruitment time or within 6 mo. bPrescribed antibiotic treatment within 6 mo before oral examination. 4 Journal of Dental Research 00(0) Figure 1. Associations between lipopolysaccharide (LPS) levels and microbial alpha diversity. Associations between (A) serum and (B) saliva LPS and alpha diversity (Shannon and inverse Simpson diversity). Comparison of Shannon and inverse Simpson diversity across (C) serum and (D) salivary LPS tertiles. Figure 2. Principal coordinates analysis (PCoA) plots illustrate beta diversity (dissimilarity between microbial communities) using the Bray–Curtis dissimilarity at the species level for (A) serum and (B) salivary lipopolysaccharide (LPS) tertiles. Ellipses represent the 95% confidence intervals for each group. Determinants of Saliva and Serum Lipopolysaccharide Activity 5 Fusobac teriota. At the genus level, Streptococcus, Prevotella, Neisseria, Veillonella, and Haemophilus were dominant. We observed differences in abundance at the phylum and genus levels for both serum and salivary LPS tertiles (Fig. 3A, B). Venn diagrams show shared and unique core genera and spe- cies across the serum and salivary LPS tertiles (Fig. 3C, D). Among 89 genera found, 59 (66.3%) genera were shared across all serum LPS tertiles, whereas 7 (7.9%), 7 (7.9%), and 4 (4.5%) were unique for tertiles 1, 2, and 3, respectively. Among the 89 genera, 60 (67.4%) were shared between salivary ter- tiles, whereas 11 (12.4%), 2 (2.2%), and 7 (7.9%) were unique for tertiles 1, 2, and 3, respectively. Cepunavirus (Streptococcus phage CP-7), Leuconostoc, and Moraxella were unique genera in both serum and saliva tertile 1, whereas Staphylococcus was unique in tertile 3 of both samples. Among the 276 identified species, 205 (74.3%) were shared across all serum LPS tertiles and 200 (74.6%) between the salivary LPS tertiles (Appendix Fig. 1). LEfSe analysis revealed that many taxa were differentially enriched between the serum and salivary LPS tertiles (Appendix Table 2). When comparing serum LPS tertiles, Prevotella histi- cola was significantly enriched in tertile 1, Porphyromonas species (including P. gingivalis and P. endodontalis) in tertile 2, and Rothia dentocariosa and Fusobacterium nucleatum in tertile 3. Salivary LPS tertiles showed even more significant differences. Genera Streptococcus (including 5 species), Haemo- philus (1), Rothia (2), Gemella (2), and Actinomyces (3) were enriched in tertile 1. In contrast, taxa enriched in the LPS tertile 3 included the phylum Bacteroidota and genera Prevotella (including 3 species), Veillonella (1 species), Schaalia (1 spe- cies), Leptotrichia (1 species), and Selenomonas (2 species). Gram-Negative Bacteria and LPS Among gram-negative species, no significant correlation was observed between alpha diversity metrics and serum or sali- vary LPS levels (Appendix Fig. 2A, B) or tertiles (Appendix Fig. 2C, D). Microbiome composition did not differ signifi- cantly between the serum LPS tertiles (P = 0.437), whereas the composition differed between salivary LPS tertiles (F = 3.72, R2 = 0.025, P = 0.001). Bacteroidota (especially genus Prevotella) increased, while Bacillota (especially Streptococcus) decreased with increasing salivary LPS tertiles (Appendix Fig. 2E, F). Figure 3. Microbial community composition in the saliva metagenome across lipopolysaccharide (LPS) tertiles. Stacked bar charts show the mean relative abundance of microbial taxa at the phylum and genus levels across different LPS tertiles for (A) serum and (B) saliva. Venn diagram depicting the shared and unique core genera across different (C) serum and (D) saliva LPS tertiles. Overlapping regions indicate genera shared between tertiles, while nonoverlapping areas represent unique core genera specific to each LPS tertile. 6 Journal of Dental Research 00(0) Oral Microbial Species as Predictors of LPS Levels Multiple linear regression analysis was conducted to examine the relationship between bacterial genera or species and serum or salivary LPS levels. The model explained little variance in serum LPS levels (F = 1.022, P = 0.429) with no statistically significant core genera. In contrast, the model for salivary LPS was significant (F = 5.673, P < 0.001) and explained 19.3% of the variance. Prevotella (P < 0.001), Neisseria (P < 0.001), Leptotrichia (P < 0.001), and Porphyromonas (P = 0.020) showed significant positive associations with salivary LPS, whereas Fusobacterium (P = 0.041) showed a negative associ- ation (Appendix Table 3). Further species-level analysis was performed based on the significant genera from the initial model, which explained 17.2% of the variance in salivary LPS levels (P < 0.001). P. gin- givalis (P = 0.005), L. wadei (P < 0.001), and Prevotella sp. E13_17 (P < 0.001) were positively associated with salivary LPS, whereas F. nucleatum (P < 0.001) showed a negative association (Appendix Table 3). Metabolic Functional Shifts Associated with High LPS Levels A total of 1,016 metabolic pathways were significantly associ- ated with salivary LPS levels but not with serum LPS levels (Appendix Table 4). Metabolites associated with high salivary LPS levels were predominantly products of pathways related to biosynthesis of nucleosides, amino acids, and cofactors and electron carriers (Fig. 4A). At the taxonomic level, these metabolites were most fre- quently linked to members of the genera Prevotella, followed by Campylobacter, Leptotrichia, and Veillonella (Fig. 4B). Figure 4. Functional pathway abundances differed significantly across saliva lipopolysaccharide (LPS) levels. (A) Number of significant metabolites grouped by metabolic super classes, (B) bacterial genera associated with the significant metabolites, and (C) specific metabolic pathways related to LPS biosynthesis, showing significant associations with high LPS levels. Key contributing taxa are indicated, with coefficients reflecting the strength of the association. Metabolic pathway profiles were generated using HUMAnN 3. Associations with saliva LPS levels were determined using MaAsLin 2. Determinants of Saliva and Serum Lipopolysaccharide Activity 7 Specific metabolic pathways involved in the biosynthesis of LPS, lipid A, and O-antigen were also significantly associated with high salivary LPS levels (P < 0.05) (Fig. 4C). Discussion The present work shows that 33% to 34% of the oral genera and 25% to 26% of the species were associated with LPS levels in both specimens. Serum LPS activity was positively associ- ated with richness and evenness of oral microbiome species, especially P. gingivalis, P. endodontalis, R. dentocariosa, and F. nucleatum. Still, the core genera explained only about 4% of the serum LPS variance, and none of them reached statistical significance. Salivary LPS activity was negatively associated with oral microbiome diversity, and the core microbiome composition differed between the LPS tertiles. A model com- posed of the core genera, Prevotella, Neisseria, Leptotrichia, Porphyromonas, and Fusobacterium, explained 19% of the salivary LPS variance. In addition, high salivary LPS levels were associated with functional pathways indicative of active microbial metabolism and proliferation within the oral micro- biome. Oral microbiota–derived LPS activity in saliva demon- strated a reduction in oral microbiome variety and enrichment of obligate or facultative anaerobic species (i.e., typical signs of dysbiosis). LPS and endotoxemia derived from the dysbiotic oral microbiome are strong stimulators of both local and systemic inflammatory and immunological responses (Pussinen et al. 2022). However, the oral microbiome composition or function- ality associated with saliva or circulating LPS activity has not been investigated. In the present study, high serum LPS activ- ity was related to high microbial diversity in saliva, but this was mainly characterized by gram-positive genera such as Staphylococcus, Streptomyces, and Lactiplantibacillus. In con- trast, the analysis performed among gram-negative genera alone did not show significant associations. Furthermore, serum LPS activity was not associated with any metabolic pro- file of the oral microbiome. Despite weak relationships observed with serum LPS activity, our study does not exclude the hypothesis that serum LPS levels are associated with the composition of the oral microbiome. In the present population, serum LPS activity was not sig- nificantly associated with periodontal status, but the serum LPS/lipoteichoic acid (LTA) ratio increased with increasing periodontitis severity (Leskelä et al. 2024). Severe periodonti- tis is characterized by reduced species richness and loss of bio- diversity (Kumar 2017), whereas the present analyses revealed increased diversity with high serum LPS activity. It is possible that the subgingival plaque microbiome is more strongly asso- ciated with serum LPS activity than the saliva microbiome, but this has been explored in only 1 earlier study using targeted microbial analyses (Liljestrand et al. 2017). In that study, sali- vary LPS activity had a significant positive association with a cluster of subgingival gram-negative species, but no associa- tion with serum LPS activity was observed (Liljestrand et al. 2017), supporting our results. The oral microbiome may be primarily associated with serum LPS via the oral–gut axis (Kitamoto and Kamada 2022). Although people swallow 1.5 × 1012 oral bacteria daily, it has been estimated that more than 99.9% of the bacteria cannot survive the acidic environment of the stomach. However, some oral species, such as Prevotella, can pass through the stomach barrier and invade the terminal ileum and colon (Mukherjee et al. 2025). Thus, it is plausible that numerous bacterial cells, metabolites, or virulence factors derived from the oral cavity reach the gut, consequently con- tributing to dysbiosis, which may lead to increased gut perme- ability and endotoxemia (Mukherjee et al. 2025). High salivary LPS was associated with decreased microbial diversity, suggesting that the microbiome is enriched by gram- negative species. Indeed, the microbiome of participants with high salivary LPS was enriched with gram-negative genera, such as Campylobacter, Leptotrichia, Selenomonas, and Veillonella, but especially Prevotella. This is aligned with our earlier results showing that salivary LPS activity is associated especially with plaque phyla Pseudomonatota, Fusobacteriota, and Bacteroidota (Kopra et al. 2021), whereas no association with the phylum Spirochaetota was observed. The outer mem- brane of Spirochaetota expresses several lipoproteins and gly- colipids, such as LPS, but several Treponema do not possess LPS or its chemical properties deviate from those described for classic gram-negative LPS (Schultz et al. 1998). The Prevotella genus, including species such as P. histicola, P. intermedia, P. jejuni, and P. melaninogenica, explained the highest proportion of salivary LPS variability (7.8%). These species are frequent colonizers of the oral cavity, with high proportions especially in saliva (Könönen and Gursoy 2022). Microbial communities and the host have a bilateral relationship, in which dysbiosis alters the immune responses resulting in inflammation, which further disrupts microbial homeostasis (Kilian et al. 2016). This also enhances the production of several virulence factors, such as LPS, leading to improved survival and increased pathogenic- ity (Curtis et al. 2020). LPS of Prevotella induces the release of proinflammatory cytokines from phagocytic cells, thus contrib- uting to tissue destruction and bone resorption in periodontitis. The biological activity of LPS correlates with number, length, and saturation degree of its lipid A acyl chains. Longer acyl chains lead generally to stronger immune responses due to stronger hydrophobic interactions with immune cell receptors, such as toll-like receptor 4 accessory protein MD-2: hexa- acylated lipid A (e.g., from Escherichia coli) elicit strong responses, while penta-acylated (e.g., from Prevotella) or tetra- acylated forms (e.g., from P. gingivalis) trigger weaker or altered immune activation (Park et al. 2009; Sharma et al. 2022). However, P. gingivalis lipid A, for example, has a heterogeneous mixture of acyl chains containing 4 to 5 fatty acids (Reife et al. 2006). Thus, the composition of lipid A and its specific LPS activity vary between species and even between strains. The predicted metabolic pathways associated with high salivary LPS levels were most commonly features of Prevotella species. Several enriched pathways reflected heightened bio- synthetic activity, including nucleoside, amino acid, cofactor, and cell wall biosynthesis. These pathways are integral to 8 Journal of Dental Research 00(0) microbial proliferation and consequently to the production of inflammatory molecules, potentially indicating a proinflam- matory oral microenvironment. Although several bacterial fatty acids are involved lipid storage processes, their main location is the hydrophobic portion of the bacterial membrane (Cronan and Thomas 2009). The identified fatty acid biosyn- thesis pathways may also be involved in the biosynthesis of LPS because of their deviating paths to provide specific fatty acids, which are components of LPS lipid A, such as 3-hydroxy and short-chain fatty acids (Raetz and Whitfield, 2002). Other enriched pathways in high salivary LPS included those associ- ated with the biosynthesis of O-antigen and KDO (3-deoxy-D- manno-octulosonic acid) in the inner core region of LPS, suggesting enhanced proliferation and virulence. The use of any systemic antibiotics was registered up to 6 mo prior to recruitment. Antibiotic use was negatively associ- ated with the diversity of the saliva microbiome. In earlier pub- lications describing adult and adolescence populations, the postantibiotic recovery of the oral microbiome has been reported to last from 3 to 8 mo (Raju et al. 2020; Kopra et al. 2021). Similarly, several antibiotics prescribed during the pre- ceding year were related to decreased serum and salivary LPS activity, whereas the effect was short-term and could be seen only until 1 mo (Kopra et al. 2021). Based on this information, we excluded from the present work participants who had used antibiotics within the preceding month. It is important to note that the assay used is not specific to any bacterial species or LPS type but determines the activity derived from a mixture of LPS molecules in the sample. The biological activity of LPS depends on its structure; thus, the results are correlated with LPS concentration to only a limited extent (Munford 2016). However, the biological effects of LPS are determined by its activity, that is, the host response, which is a clinically relevant measure. Importantly, LPS activity assays are highly sensitive to environmental variables, such as the composition of different sample matrices, which cannot be directly compared with each other. The results need to be inter- preted within these limitations, which emphasize the impor- tance of further methodological validation in the future. In conclusion, Prevotella, Neisseria, Leptotrichia, Porphy- romonas, and Fusobacterium genera were the strongest deter- minants of salivary LPS activity, whereas no significant associations with serum LPS activity were observed. Oral microbiota–derived LPS activity in saliva was associated with highly active microbial metabolism characterized by prolifera- tion, biosynthesis, and virulence. The observed taxonomic and functional signatures may enhance both local and systemic inflammation. Although the determination of salivary LPS activity may not be utilizable in clinical practice, it is a highly relevant measure in clinical research. Author Contributions M. Manzoor, contributed to data acquisition, analysis, and interpre- tation, drafted the manuscript; J. Putaala, S. Paju, contributed to conception and design, data acquisition, analysis, and interpretation, critically revised the manuscript; S. Zaric, J. Leskelä, contributed to data acquisition, analysis, and interpretation, critically revised the manuscript; A Dong, contributed to data acquisition, critically revised the manuscript; E. Könönen, contributed to data acquisi- tion and interpretation, critically revised the manuscript; L. Lahti, contributed to data interpretation, critically revised the manu- script; P. J. Pussinen, contributed to conception and design, data acquisition, analysis, and interpretation, drafted the manuscript. All authors gave final approval and agree to be accountable for all aspects of the work. Acknowledgments We acknowledge the Biomedicum Functional Genomics Unit (FuGU) at the University of Helsinki for their assistance with metagenomic sequencing. Declaration of Conflicting Interests The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The SECRETO Oral study was funded by the Research Council of Finland grants (316777 and 355532 for S.P. and 340750 for P.J.P.), the Finnish Dental Society Apollonia (for P.J.P.), and the Sigrid Juselius Foundation (for P.J.P.). The SECRETO study was funded by the Research Council of Finland (286246, 318075, and 322656 for J.P.), the Helsinki and Uusimaa Hospital District (TYH2014407, TYH2018318 for J.P.), and Sigrid Juselius Foundation. In addition, we acknowledge the funding from the Academy of Medical Sciences (SGL023/1035), MRC IAA King’s College London (MR/X502923/1), EPSRC IAA (EP/ X525571/1) (for S.Z.), and the King’s-China Scholarship Council PhD Scholarship Program (202108410182 to A.D.). ORCID iDs M. Manzoor https://orcid.org/0000-0002-7939-1563 S. Zaric https://orcid.org/0000-0003-4127-0253 J. Leskelä https://orcid.org/0000-0001-9670-339X P. J. Pussinen https://orcid.org/0000-0003-3563-1876 Data Availability Statement Genomic data have been deposited in the European Genome- Phenome Archive (EGA) and are accessible under study accession number EGAS00001007505. The source code for the analyses is available at https://doi.org/10.5281/zenodo.15308333. References Beghini F, McIver LJ, Blanco-Míguez A, Dubois L, Asnicar F, Maharjan S, Mailyan A, Manghi P, Scholz M, Thomas AM, et al. 2021. Integrating taxo- nomic, functional, and strain-level profiling of diverse microbial communi- ties with bioBakery 3. Elife. 10:e65088. doi:10.7554/eLife.65088 Chapple ILC, Mealey BL, Van Dyke TE, Bartold PM, Dommisch H, Eickholz P, Geisinger ML, Genco RJ, Glogauer M, Goldstein M, et al. 2018. Periodontal health and gingival diseases and conditions on an intact and a reduced periodontium: consensus report of workgroup 1 of the 2017 World Workshop on the Classification of Periodontal and Peri-Implant Diseases and Conditions. J Clin Periodontol. 45(suppl 20):S68–S77. Determinants of Saliva and Serum Lipopolysaccharide Activity 9 Cronan JE, Thomas J. 2009. Bacterial fatty acid synthesis and its relationships with polyketide synthetic pathways. Methods Enzymol. 459:395–433. Curtis MA, Diaz PI, Van Dyke TE. 2020. The role of the microbiota in peri- odontal disease. Periodontol 2000. 83(1):14–25. Dong A, Proctor G, Zaric S. 2025. Diagnostic accuracy of microbiome- derived biomarkers in periodontitis: systematic review and meta-analysis. J Periodontal Res [epub ahead of print 13 Jan 2025]. doi:10.1111/jre.13377 Erridge C, Bennett-Guerrero E, Poxton IR. 2002. Structure and function of lipo- polysaccharides. Microbes Infect. 4(8):837–851. Geerts SO, Nys M, De MP, Charpentier J, Albert A, Legrand V, Rompen EH. 2002. Systemic release of endotoxins induced by gentle mastication: asso- ciation with periodontitis severity. J Periodontol. 73(1):73–78. Kilian M, Chapple IL, Hannig M, Marsh PD, Meuric V, Pedersen AM, Tonetti MS, Wade WG, Zaura E. 2016. The oral microbiome—an update for oral healthcare professionals. Br Dent J. 221(10):657–666. Kitamoto S, Kamada N. 2022. Periodontal connection with intestinal inflamma- tion: microbiological and immunological mechanisms. Periodontol 2000. 89(1):142–153. Könönen E, Gursoy UK. 2022. Oral Prevotella species and their connection to events of clinical relevance in gastrointestinal and respiratory tracts. Front Microbiol. 12:798763. doi:10.3389/fmicb.2021.798763 Kopra E, Lahdentausta L, Pietiäinen M, Buhlin K, Mäntylä P, Hörkkö S, Persson R, Paju S, Sinisalo J, Salminen A, et al. 2021. Systemic antibi- otics influence periodontal parameters and oral microbiota, but not sero- logical markers. Front Cell Infect Microbiol. 11:774665. doi:10.3389/ fcimb.2021.774665 Kumar PS. 2017. From focal sepsis to periodontal medicine: a century of exploring the role of the oral microbiome in systemic disease. J Physiol. 595(2):465–476. Leira Y, Fragkiskos D, Orlandi M, Suvan J, Nibali L, Tonetti MS, Belibasakis GN, Bostanci N, D’Aiuto F. 2021. Severe periodontitis and biomarkers of bacterial burden. Results from a case-control and intervention clinical trial. Front Oral Health. 2:615579. doi:10.3389/froh.2021.615579 Leskelä J, Putaala J, Martinez-Majander N, Tulkki L, Manzoor M, Zaric S, Ylikotila P, Lautamäki R, Saraste A, Suihko S, et al. 2024. Periodontitis, dental procedures, and young-onset cryptogenic stroke. J Dent Res. 103(5): 494–501. Liljestrand JM, Paju S, Buhlin K, Persson GR, Sarna S, Nieminen MS, Sinisalo J, Mäntylä P, Pussinen PJ. 2017. Lipopolysaccharide, a possible molecu- lar mediator between periodontitis and coronary artery disease. J Clin Periodontol. 44(8):784–792. Liukkonen J, Gürsoy UK, Könönen E, Akhi R, Salminen A, Liljestrand JM, Pradhan-Palikhe P, Pietiäinen M, Sorsa T, Persson GR, et al. 2020. Immunological and microbiological profiling of cumulative risk score for periodontitis. Diagnostics (Basel). 10(8):560. doi:10.3390/diagnos tics10080560 Lockhart PB, Brennan MT, Sasser HC, Fox PC, Paster BJ, Bahrani-Mougeot FK. 2008. Bacteremia associated with toothbrushing and dental extraction. Circulation. 117(24):3118–3125. Mallick H, Rahnavard A, McIver LJ, Ma S, Zhang Y, Nguyen LH, Tickle TL, Weingart G, Ren B, Schwager EH, et al. 2021. Multivariable association discovery in population-scale meta-omics studies. PLoS Comput Biol. 17(11):e1009442. doi:10.1371/journal.pcbi.1009442 Manzoor M, Leskelä J, Pietiäinen M, Martinez-Majander N, Könönen E, Niiranen T, Lahti L, Sinisalo J, Putaala J, Pussinen PJ, et al. 2024. Shotgun metagenomic analysis of the oral microbiome in gingivitis: a nested case- control study. J Oral Microbiol. 16(1):2330867. doi:10.1080/20002297.20 24.2330867 Manzoor M, Leskelä J, Pietiäinen M, Martinez-Majander N, Könönen E, Sinisalo J, Putaala J, Pussinen PJ, Paju S. 2025. Oral microbiome dysbiosis in cryptogenic ischemic stroke patients with high-risk patent foramen ovale. Sci Rep. 15(1):11535. doi:10.1038/s41598-025-95728-x Manzoor M, Leskelä J, Pietiäinen M, Martinez-Majander N, Ylikotila P, Könönen E, Niiranen T, Lahti L, Sinisalo J, Putaala J, et al. 2024. Multikingdom oral microbiome interactions in early-onset cryptogenic ischemic stroke. ISME Commun. 4(1):ycae088. doi:10.1093/ismeco/ ycae088 Mukherjee S, Chopra A, Karmakar S, Bhat SG. 2025. Periodontitis increases the risk of gastrointestinal dysfunction: an update on the plausible patho- genic molecular mechanisms. Crit Rev Microbiol. 51(1):187–217. Munford RS. 2016. Endotoxemia-menace, marker, or mistake? J Leukoc Biol. 100(4):687–698. Oh S, Lee H-J, Park KU. 2023. Metagenomic characterization of the microbi- omes in five different body habitats of otherwise healthy individuals with periodontal disease. Front Cell Infect Microbiol. 13:1257816. doi:10.3389/ fcimb.2023.1257816 Park BS, Song DH, Kim HM, Choi BS, Lee H, Lee JO. 2009. The structural basis of lipopolysaccharide recognition by the TLR4-MD-2 complex. Nature. 458(7242):1191–1195. Pussinen PJ, Kopra E, Pietiäinen M, Lehto M, Zaric S, Paju S, Salminen A. 2022. Periodontitis and cardiometabolic disorders: the role of lipopolysac- charide and endotoxemia. Periodontol 2000. 89(1):19–40. Pussinen PJ, Vilkuna-Rautiainen T, Alfthan G, Palosuo T, Jauhiainen M, Sundvall J, Vesanen M, Mattila K, Asikainen S. 2004. Severe periodontitis enhances macrophage activation via increased serum lipopolysaccharide. Arterioscler Thromb Vasc Biol. 24(11):2174–2180. Putaala J, Martinez-Majander N, Saeed S, Yesilot N, Jäkälä P, Nerg O, Tsivgoulis G, Numminen H, Gordin D, von Sarnowski B, et al. 2017. Searching for explanations for cryptogenic stroke in the young: revealing the triggers, causes, and outcome (SECRETO): rationale and design. Eur Stroke J. 2(2):116–125. Raetz CR, Whitfield C. 2002. Lipopolysaccharide endotoxins. Annu Rev Biochem. 71:635–700. Raju SC, Viljakainen H, Figueiredo RAO, Neuvonen PJ, Eriksson JG, Weiderpass E, Rounge TB. 2020. Antimicrobial drug use in the first decade of life influences saliva microbiota diversity and composition. Microbiome. 8(1):121. doi:10.1186/s40168-020-00893-y Reife RA, Coats SR, Al-Qutub M, Dixon DM, Braham PA, Billharz RJ, Howald WN, Darveau RP. 2006. Porphyromonas gingivalis lipopolysac- charide lipid A heterogeneity: differential activities of tetra- and penta- acylated lipid A structures on E-selectin expression and TLR4 recognition. Cell Microbiol. 8(5):857–868. Schultz CP, Wolf V, Lange R, Mertens E, Wecke J, Naumann D, Zähringer U. 1998. Evidence for a new type of outer membrane lipid in oral spirochete Treponema denticola. Functioning permeation barrier without lipopolysac- charides. J Biol Chem. 273(25):15661–15666. Segata N, Izard J, Waldron L, Gevers D, Miropolsky L, Garrett WS, Huttenhower C. 2011. Metagenomic biomarker discovery and explanation. Genome Biol. 12(6):R60. doi:10.1186/gb-2011-12-6-r60 Shaddox LM, Wiedey J, Calderon NL, Magnusson I, Bimstein E, Bidwell JA, Zapert EF, Aukhil I, Wallet SM. 2011. Local inflammatory markers and systemic endotoxin in aggressive periodontitis. J Dent Res. 90(9): 1140–1144. Sharma G, Garg N, Hasan S, Shirodkar S. 2022. Prevotella: an insight into its characteristics and associated virulence factors. Microb Pathog. 169:105673. doi:10.1016/j.micpath.2022.105673 Valvano MA. 2015. Genetics and biosynthesis of lipopolysaccharide. In: Tang Y-W, Sussman M, Liu D, Poxton I, Schwartzman J, editors. Molecular med- ical microbiology. 2nd ed. London: Academic Press. p. 55–89. doi:10.1016/ B978-0-12-397169-2.00004-4 Zaric S, Strachan A, Kurushima Y, Dong A, McIlwaine C, Harrington Z, Nibali L, Foey A, Ide M. 2022. Evaluating clinical utility of subgingival and sali- vary endotoxin activity levels as periodontal biomarkers. Front Oral Health. 3:1029806. doi:10.3389/froh.2022.1029806 Zhao Y, Feng Y, Ye Q, Hu J, Feng Y, Ouyang Z, Zhao J, Chen Y, Tan L, Chen N, et al. 2022. The oral microbiome in young women at different stages of periodontitis: Prevotella dominant in stage III periodontitis. Front Cell Infect Microbiol. 12:1047607. doi:10.3389/fcimb.2022.1047607