Diabetes exacerbates experimental peri-implantitis in mice with elevated IL-17A-associated inflammation and IL-17F upregulation

Abstract

This study investigated the role of interleukin-17 (IL-17) and the oral microbiome in peri-implant inflammation and bone loss under hyperglycemic and normoglycemic conditions. Wild-type (WT) and diabetic (db/db) mice with maxillary implants underwent ligature placement with or without IL-17A neutralization. Bone loss, osteoclast activity, inflammatory cytokines, Th17/Treg balance, and expression of IL-17Family members were analyzed. The oral microbiota was profiled by 16S rRNA sequencing, and its inflammatory potential was evaluated by co-culture with immune cells. Diabetic db/db mice exhibited greater peri-implant bone loss, osteoclast numbers, and RANKL/OPG ratios than WT, accompanied by elevated Il17a expression, reduced anti-inflammatory cytokines, enhanced Th17-associated inflammatory features, and altered FOXP3⁺ cell profiles. IL-17A neutralization significantly attenuated, but did not fully normalize, heightened inflammatory responses in db/db mice, whereas ligature-induced Il17f upregulation was observed only in db/db mice. Microbial alterations were partially shifted toward control profiles by IL-17A inhibition in WT mice, while diabetes-associated changes persisted regardless of ligation or anti-IL-17A. In vitro, peri-implant microbiota induced pro-inflammatory cytokine responses in splenocytes, with residual inflammatory responses remaining more evident in DB-derived microbiota after IL-17A inhibition. These findings suggest that peri-implantitis in diabetes is exacerbated by heightened IL-17-mediated inflammation and persistent microbial alterations, underscoring the need for more comprehensive therapeutic approaches to address the disease under diabetic conditions.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1038/s41598-026-58925-w.

Introduction

Peri-implantitis, characterized by inflammation and progressive bone loss, threatens the long-term success of dental implants. Diabetes mellitus (DM), a systemic condition that impairs healing, disrupts immune homeostasis, and promotes chronic inflammation^1,2, nearly doubles the risk of peri-implantitis^3,4. With the global prevalence of DM continuing to rise, elucidating the underlying immune and microbial mechanisms and establishing effective treatments for peri-implantitis in diabetic patients have become urgent priorities^5,6.

Among the six isoforms of the interleukin-17 (IL-17) cytokine family, IL-17A has been most extensively studied and is recognized as a major mediator of pro-inflammatory responses⁷. IL-17A, primarily secreted by Th17 cells, promotes neutrophil recruitment, inflammatory cytokine production, RANKL-dependent osteoclastogenesis, and soft tissue degradation, thereby contributing to inflammatory bone loss and peri-implant tissue destruction^8,9. These findings support the central role of IL-17A in local immune dysregulation during peri-implant inflammation.

DM can further amplify IL-17-associated inflammatory responses by disrupting immune homeostasis. Under diabetic conditions, Th17 expansion and IL-17 overproduction are favored, whereas regulatory T cell differentiation and suppressive function are impaired^10–12. This Th17/Treg-related imbalance may enhance inflammatory signaling, promote osteoclastogenesis, and delay resolution of tissue inflammation. In diabetic mouse models, elevated IL-17A expression accelerates periodontal bone loss and shifts the oral microbiome toward a more pathogenic profile, effects that can be reversed, at least in part, by local IL-17A neutralization¹³.

In addition to IL-17A, IL-17F shares structural and functional similarities with IL-17A and can participate in overlapping inflammatory pathways^7–9. However, compared with IL-17A, the involvement of IL-17F in oral inflammatory bone loss remains less well defined. In particular, whether IL-17F is upregulated in peri-implant tissues under diabetic conditions, and whether IL-17A blockade is sufficient to control IL-17Family-associated inflammatory responses in this setting, remain unclear.

Hyperglycemia also alters the ecological landscape of the oral cavity, favoring the overgrowth of pathogenic genera such as Porphyromonas, Tannerella, and Capnocytophaga¹⁴. This dysbiosis is positively correlated with blood glucose and HbA1c levels¹⁵. Because IL-17 responses are shaped by microbial stimuli and may also influence microbial community structure, IL-17-mediated inflammation and diabetes-associated microbial alterations may form a bidirectional feedback loop that sustains peri-implant inflammation¹⁶.

While IL-17A has been independently implicated in peri-implantitis and diabetes-associated inflammation, its direct contribution to peri-implant pathology in diabetic hosts has not been clearly defined. Moreover, the potential involvement of additional IL-17Family members, particularly IL-17F, in diabetes-associated peri-implant inflammation remains largely unexplored.

This study aimed to investigate the pathological contribution of IL-17A to peri-implant inflammation and bone loss using a ligature-induced peri-implantitis model in diabetic and normoglycemic mice. We further examined whether diabetic conditions were associated with altered IL-17Family expression, particularly IL-17F upregulation, and whether IL-17A neutralization differentially affected peri-implant inflammatory responses and microbiota-associated immune activation in diabetic and normoglycemic hosts.

Results

Diabetes exacerbates ligature-induced peri-implant bone loss and osteoclastogenic activity

Body weight showed a moderate increase within each group during the experimental period, ranging from 18.5 g to 34.2 g in WT mice and from 29.7 g to 57.8 g in DB mice (Fig. 1C). Blood glucose levels remained unchanged in WT mice at around 200 mg/dL. In contrast, DB mice exhibited a progressive increase from 416.9 mg/dL to 581.1 mg/dL, confirming the development of the diabetic phenotype (Fig. 1D).

Fig. 1.

Experimental protocol and monitoring of body weight and blood glucose. (A) Schedule of the surgical procedure. (B) Clinical observations before extraction, after tooth extraction, 5 weeks after extraction, at implant placement, 4 weeks after implant placement, at ligation, and 2 weeks after ligation in WT and DB mice. (C) Body weight was monitored during the experimental period. (D) Blood glucose levels were measured during the experimental period. Final sample sizes: n = 6–7 mice/group. WT, wild-type mice; DB, diabetic mice; ctrl, control; IgG, ligation with IgG isotype control administration; xIL-17A, ligation with anti-IL-17A antibody administration.

We next examined whether diabetic conditions affected ligature-induced peri-implant bone loss. Micro-CT showed no baseline differences between WT and DB controls. Ligation significantly increased peri-implant bone loss in both groups (p < 0.05), with greater resorption in DB than WT mice (p < 0.01). Anti-IL-17A treatment markedly reduced ligature-induced bone loss in both WT and DB mice (p < 0.05, Fig. 2A–C).

Fig. 2.

Evaluation of alveolar bone resorption, osteoclast differentiation, and RANKL/OPG ratios. (A) Representative 3D and 2D sagittal micro-CT images around implants in WT mice. (B) Representative 3D and 2D sagittal micro-CT images around implants in DB mice. (C) Quantification of bone loss volume around implants. (D) Representative TRAP-stained images around implants in WT mice. Scale bar: 150 μm. (E) Representative TRAP-stained images around implants in DB mice. Scale bar: 150 μm. (F) TRAP-positive cells around implants were counted and normalized to the WT control group. mRNA expression levels in gingival tissue were detected by qPCR: (G) Tnfsf11 (RANKL), (H) Tnfrsf11b (OPG), and (I) RANKL/OPG ratio calculated based on RANKL and OPG mRNA expression levels. Final sample sizes: micro-CT analysis, n = 6–7 mice/group; TRAP staining, n = 3 mice/group; qPCR analysis, n = 3–4 mice/group. Mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. WT, wild-type mice; DB, diabetic mice; ctrl, control; IgG, ligation with IgG isotype control administration; xIL-17A, ligation with anti-IL-17A antibody administration.

Consistent with the micro-CT findings, TRAP staining revealed no differences between WT and DB controls, but ligation significantly increased TRAP-positive cells in both groups (p < 0.0001), more prominently in DB mice (p < 0.01). Anti-IL-17A treatment markedly reduced osteoclast numbers in both WT and DB mice (p < 0.05, Fig. 2D–F).

To further examine osteoclastogenic signaling, qPCR analysis was performed for Tnfsf11 (RANKL) and Tnfrsf11b (OPG). Baseline Tnfsf11 and Tnfrsf11b did not differ between WT and DB mice. Ligation significantly upregulated Tnfsf11 in both groups, which was attenuated by anti-IL-17A (p < 0.0001, Fig. 2G). Tnfrsf11b was unchanged in WT but significantly decreased in DB mice after ligation (p < 0.01, Fig. 2H). Anti-IL-17A markedly increased Tnfrsf11b expression in DB mice (p < 0.0001). Consequently, Tnfrsf11b expression under ligation was lower in DB than WT mice (p < 0.0001), resulting in an elevated RANKL/OPG ratio in DB mice, which was reduced by anti-IL-17A (p < 0.0001, Fig. 2I).

Anti-IL-17A attenuates ligature-induced peri-implant inflammatory responses

We then analyzed peri-implant gingival tissues to determine whether IL-17A neutralization affected cytokine expression and immune cell infiltration in WT and DB mice. Baseline Il1b, Tnfa, and Il17a did not differ between WT and DB controls. After ligation, these pro-inflammatory cytokines were upregulated in both groups, and anti-IL-17A treatment attenuated their expression (Fig. 3A–C, p < 0.05), except for Tnfa in WT mice. Notably, Il17a was significantly higher in DB than WT ligature mice (p < 0.05, Fig. 3C).

Fig. 3.

Detection of gingival mRNA levels of inflammatory cytokines and immune cell infiltration. mRNA expression levels in gingival tissue were detected by qPCR: (A) Il1b (IL-1β), (B) Tnfa (TNF-α), (C) Il17a (IL-17A), (D) Il10 (IL-10), (E) Il1rn (IL-1ra), and (F) Vegfa (VEGF-A). (G) Representative H&E-stained images around implants in WT and DB mice. Scale bar: 150 μm. (H) Inflammatory cells around implants were counted and quantified using ImageJ. Final sample sizes: qPCR analysis, n = 3–4 mice/group; H&E staining, n = 3 mice/group. Mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. WT, wild-type mice; DB, diabetic mice; ctrl, control; IgG, ligation with IgG isotype control administration; xIL-17A, ligation with anti-IL-17A antibody administration.

Among anti-inflammatory cytokines, Il10, Il1rn, and Vegfa showed greater reduction in DB than in WT mice after ligature (Fig. 3D–F, p < 0.01). Il10 was unchanged in WT, but in DB mice, it was suppressed by ligation and restored by antibody treatment (p < 0.05). Il1rn returned to baseline in WT but not DB mice following antibody treatment. Vegfa increased after ligation and decreased with antibody treatment in WT (p < 0.01), but showed no differences among DB groups.

In line with these cytokine changes, HE staining showed no differences between WT and DB controls, but ligation significantly increased immune cell infiltration in both groups (p < 0.0001), with DB mice exhibiting greater infiltration (p < 0.0001). Anti-IL-17A markedly reduced infiltration in both WT and DB mice (Fig. 3G, H).

We further evaluated IL-17A protein expression in peri-implant tissues. IHC revealed no baseline differences in IL-17A between WT and DB controls. Ligation significantly increased IL-17A expression in both groups, which was reduced by anti-IL-17A treatment (p < 0.05). DB mice showed higher IL-17A levels than WT under ligation (p < 0.001, Fig. 4A, B).

Fig. 4.

Identification of IL-17A expression and CD4⁺FOXP3⁺ cells in peri-implant tissues. (A) Representative IHC images of IL-17A expression around implants in WT and DB mice. Scale bar: 150 μm. (B) IL-17A-positive cells were quantified using ImageJ. (C) Representative immunofluorescence images of CD4, FOXP3, and DAPI staining. APC-conjugated anti-CD4 antibody and anti-FOXP3 antibody were used as primary antibodies, and Alexa Fluor 488-conjugated anti-rabbit IgG was used as the secondary antibody. Scale bar: 150 μm. (D) CD4⁺FOXP3⁺ cell abundance was quantified as integrated intensity using ImageJ. Final sample sizes: IL-17A IHC, n = 3 mice/group; CD4⁺FOXP3⁺ IF analysis, n = 3 mice/group. Mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. WT, wild-type mice; DB, diabetic mice; ctrl, control; IgG, ligation with IgG isotype control administration; xIL-17A, ligation with anti-IL-17A antibody administration.

Because IL-17A-associated inflammation may be linked to local immune regulatory profiles, IF analysis was performed to assess CD4⁺FOXP3⁺ cell abundance within peri-implant tissue sections. In WT mice, ligation reduced the abundance of CD4⁺FOXP3⁺ cells, which increased after anti-IL-17A treatment. In DB mice, CD4⁺FOXP3⁺ cell abundance did not significantly differ among treatment groups (Fig. 4C, D).

Diabetes is associated with enhanced Il17a expression and Il17f upregulation

Because anti-IL-17A treatment attenuated but did not fully normalize several inflammatory responses in DB mice, we next examined the expression profile of IL-17Family members in peri-implant gingival tissues. The qPCR results showed that in WT mice, only Il17a was upregulated after ligation, while other isoforms remained low. In DB mice, Il17a expression was markedly elevated compared with WT (p < 0.001), and Il17f expression was significantly increased (p < 0.0001), indicating selective upregulation of Il17f under diabetic conditions (Fig. 5A).

Fig. 5.

Gene expression levels of IL-17Family members and IL-17F protein expression in the gingival tissue after ligation. (A) mRNA expression levels of IL-17Family members in the gingival tissue were detected by qPCR and normalized to Gapdh. (B) Representative IHC images of IL-17F expression around implants in WT and DB mice. Scale bar: 150 μm. (C) IL-17F-positive cells were quantified using ImageJ. Final sample sizes: IL-17Family qPCR, n = 3 mice/group; IL-17F IHC, n = 3 mice/group. Mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. WT, wild-type mice; DB, diabetic mice; ctrl, control; IgG, ligation with IgG isotype control administration; xIL-17A, ligation with anti-IL-17A antibody administration.

Consistent with the gene expression data, IHC confirmed that the low level of IL-17F-positive cells did not differ among WT groups but was significantly increased in DB mice after ligation (p < 0.01). This increase was not reduced by anti-IL-17A antibody treatment (Fig. 5B, C).

IL-17A neutralization differentially affects peri-implant microbiota-associated responses in WT and DB mice

We next examined whether peri-implant microbiota profiles differed between WT and DB mice after ligation and IL-17A neutralization. 16S rRNA sequencing analysis showed that the microbiome of DB controls already differed from those of WT at baseline (p < 0.01, Fig. 6A). In WT mice, ligation caused a marked microbial shift, with significant reductions in richness-type α-diversity in two of the four models, observed features and Chao1 index, and clear β-diversity separation from controls (p < 0.05, Fig. 6B, C). In contrast, DB mice showed minimal response to ligation, with no significant changes in any α- or β-diversity indices (Fig. 6B, D). In WT mice, anti-IL-17A treatment did not change β-diversity compared with controls (Fig. 6C), and DB mice similarly exhibited no significant changes relative to their controls (Fig. 6D). Taxonomic profiling confirmed persistent divergence in each group after any treatments (Fig. S1).

Fig. 6.

Analysis of the composition and pro-inflammatory potential of microbiota from peri-implant sites in WT and DB mice. (A) Principal coordinate analysis (PCoA) based on Bray–Curtis distances comparing WT control and DB control microbiota, with box plots quantifying distances to DB control. (B) Comparison of α-diversity indices, including observed features, Chao1 index, Shannon index, and Simpson index, across WT and DB groups with or without ligation and anti-IL-17A treatment. (C) PCoA plots of microbial communities in WT mice under ligature with IgG or anti-IL-17A treatment, with box plots showing distances to WT control. (D) PCoA plots of microbial communities in DB mice under ligature with IgG or anti-IL-17A treatment, with box plots showing distances to DB control. (E–I) Relative mRNA expression of inflammatory cytokines in splenocytes co-cultured with microbiota from each group: (E) Il1b (IL-1β), (F) Tnfa (TNF-α), (G) Il17a (IL-17A), (H) Il17f (IL-17F), and (I) Il10 (IL-10). Control indicates splenocytes cultured without bacterial stimulation. Final sample sizes: 16S rRNA sequencing, n = 5 mice/group; splenocyte co-culture assay, n = 5 independent microbiota preparations/group. Mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. WT, wild-type mice; DB, diabetic mice; ctrl, control; IgG, ligation with IgG isotype control administration; xIL-17A, ligation with anti-IL-17A antibody administration.

To assess the functional inflammatory response in vitro, splenocytes were co-cultured with fixed bacteria expanded from peri-implant microbiota. In both WT and DB mice, Il1b, Tnfa, Il17a, and Il17f expression was elevated when stimulated with microbiota from the ligation group (p < 0.05, Fig. 6E–H). These responses were attenuated by microbiota from the IL-17A antibody-treated group in WT mice. In DB mice, however, residual inflammatory responses persisted, particularly for Tnfa expression. Il10 expression decreased following co-culture with microbiota from the ligation group but was markedly restored by microbiota from the IL-17A antibody-treated group (p < 0.0001, Fig. 6I). In contrast, Il10 expression in DB mice remained unchanged across treatment groups.

Discussion

The present study extends previous findings on diabetes-associated IL-17–mediated inflammation and oral dysbiosis by providing an integrated in vivo analysis of peri-implantitis under diabetic conditions. Within this model, diabetes was associated with enhanced peri-implant bone loss and IL-17A-associated inflammatory responses, Il17f upregulation, imbalanced Th17/Treg-associated immune profiles, and altered peri-implant microbiota. Micro-CT and TRAP staining showed that DB mice exhibited greater bone resorption and osteoclast formation than WT mice under ligature-induced peri-implantitis (Fig. 2A-F). These observations are consistent with previous reports indicating increased susceptibility to bone loss under diabetic conditions^17,18. Neutralization of IL-17A attenuated bone loss and osteoclast numbers in both genotypes, accompanied by reduced RANKL expression, supporting a contributory role of IL-17A in osteoclastogenic activity during peri-implant inflammation (Fig. 2A–F). OPG expression displayed divergent patterns, remaining unaffected in WT mice but decreasing in DB mice following ligation (Fig. 2H). Because OPG expression in periodontal ligament cells can be modulated by IL-17A in a dose-dependent manner¹⁹, the genotype-specific reduction of OPG in DB mice may have contributed to the elevated RANKL/OPG ratio and greater bone resorption (Fig. 2I).

DB mice also displayed a more pronounced peri-implant inflammatory response than WT mice (Figs. 3 and 4). Clinical and meta-analytic evidence has shown that diabetes significantly increases the risk and severity of peri-implantitis and implant loss²⁰. In the present study, pro-inflammatory cytokines were elevated in DB mice, and IL-17A antibody treatment significantly attenuated these responses in both WT and DB mice (Fig. 3A–C). However, these responses were not fully normalized under diabetic conditions, suggesting that additional diabetes-associated inflammatory mechanisms may contribute to the residual inflammatory phenotype. Although baseline Il17a expression did not differ between genotypes, ligation induced a stronger upregulation in DB mice, suggesting that diabetic conditions may amplify IL-17-associated inflammatory responses (Fig. 3C). Clinical studies have shown that salivary IL-17 levels are higher in periodontitis patients and become even higher when type 2 diabetes is present²¹. The IL-23/IL-17Axis has also been implicated in periodontal destruction and bone resorption²², and pro-inflammatory cytokines such as IL-1β, IL-17, IL-6, and TNF-α have been associated with the progression of peri-implantitis²³. Thus, the present findings support the relevance of IL-17A-associated inflammation in diabetes-exacerbated peri-implant tissue destruction.

Impaired resolution of inflammation may also contribute to sustained peri-implant inflammation in DB mice (Fig. 3D–F). Although IL-17A blockade attenuated inflammatory responses, several anti-inflammatory mediators failed to return to normal levels in DB mice, suggesting that diabetes compromises the resolution of inflammation. Previous reports demonstrated that macrophages from db/db mice exhibit defective efferocytosis, leading to prolonged inflammation characterized by reduced IL-10 production²⁴. In patients with type 2 diabetes, anti-inflammatory cytokines such as IL-10, IL-4, IL-13, and TGF-β are also significantly reduced²⁵. Consistently, DB mice displayed greater infiltration of immune cells in peri-implant tissues (Fig. 3G, H). These features may limit the capacity of diabetic tissues to resolve inflammation, even when pro-inflammatory signaling is partially suppressed.

IHC confirmed elevated IL-17A protein expression, and Th17-related inflammatory features were more prominent in DB mice but diminished after IL-17A antibody administration (Fig. 4A, B). In contrast, CD4⁺FOXP3⁺ cell abundance differed between WT and DB mice, suggesting that diabetes is associated with altered Th17/Treg-related immune profiles in peri-implant tissues (Fig. 4C, D). Disruption of the Th17/Treg balance has been widely implicated in the pathogenesis of periodontitis and diabetes-associated bone loss²⁶. However, because FOXP3⁺ cell abundance was assessed by two-dimensional IF staining, these data should not be interpreted as definitive quantification of total Treg populations or as direct evidence of impaired Treg suppressive function.

Currently, six isoforms of the IL-17Family are recognized (Supplementary Table 2). In addition to Il17a mRNA expression, we observed more abundant mRNA and protein expression of Il17f in the peri-implant gingival tissue of DB mice than WT mice following ligation (Fig. 5A–C). Recent evidence suggests that IL-17F is elevated in metabolic disorders such as obesity and metabolic syndrome, where its serum levels correlate with BMI and contribute to low-grade systemic inflammation²⁷. Functional studies have shown that IL-17A and IL-17F share overlapping inflammatory roles, meaning that inhibition of only one isoform may not sufficiently reduce inflammation²⁸. Importantly, IL-17F upregulation was observed under diabetic conditions in our model and was not reduced by IL-17A neutralization. Although this finding does not establish a causal role for IL-17F, it suggests that IL-17A alone may not fully explain IL-17Family–associated inflammatory responses in diabetic peri-implantitis. Together with insufficient anti-inflammatory regulation, this diabetes-associated inflammatory profile may contribute to heightened inflammatory and osteolytic responses at the 14-day endpoint.

Our results demonstrated that DB mice exhibited a distinct microbiome composition compared with WT at baseline (Fig. 6A), supporting the notion that type 2 diabetes mellitus is associated with pathogenic oral microbiome alteration even under clinically healthy periodontal conditions¹³ and consistent with studies showing diabetes-induced oral dysbiosis independent of local inflammatory challenge²⁹. In WT mice, ligation markedly reduced α-diversity and caused clear β-diversity separation (Fig. 6B, C), whereas DB mice showed only minor post-ligation changes (Fig. 6B, D), suggesting pre-existing alteration and inflammation-driven dysbiosis³⁰. Because hyperglycemia may limit microbial recovery³¹, the limited effect of IL-17A neutralization in DB mice suggests that host metabolic status may constrain immune-mediated modulation of the oral microbiome.

Splenocytes were selected as a readily available mixed primary immune cell population to compare microbiota-induced cytokine responses under standardized in vitro conditions. Cellular analyses suggested that diabetes does not necessarily potentiate basal inflammatory responses in splenocytes, consistent with previous reports showing comparable baseline expression of TNF-α and MCP-1 in macrophages from diabetic and non-diabetic donors³². In both WT and DB mice, anti-IL-17A treatment reduced the pro-inflammatory potential of isolated microbiota from anti-IL-17A–treated sites, which induced lower expression of Il1b, Il17a, and Il17f in splenocytes than microbiota from ligature-treated sites (Fig. 6E, G, H). However, the present study does not establish that IL-17A directly controls microbial composition in vivo. In DB mice, Tnfa induction remained relatively elevated, and Il10 was not restored (Fig. 6F, I), suggesting that microbiome-associated cytokine responses under diabetic conditions are accompanied by insufficient anti-inflammatory responses.

In gingival tissues, Il10 expression was markedly increased after IL-17A blockade, demonstrating tissue-specific induction of anti-inflammatory responses (Fig. 3D). Genetic knockout of IL-17A can increase IL-10 expression at local tissue sites³³, whereas Il10 upregulation was not observed in our splenocyte-microbiome co-culture system. This discrepancy may reflect the absence of stromal or epithelial components required for IL-10 induction and the tissue-specificity of IL-17A–IL-10 crosstalk³⁴. Under diabetic conditions, impaired Treg and B10 cell function may further restrict IL-10 production^35,36. In addition, the persistence of IL-17F under diabetic conditions highlights an additional inflammatory axis that is not targeted by IL-17A blockade. This interpretation is supported by previous studies showing that IL-17F exerts inflammatory functions similar to, and partially redundant with, IL-17A²⁸. Indeed, dual targeting of IL-17A and IL-17F results in more pronounced attenuation of inflammatory responses than selective IL-17A inhibition³⁷. Together, these findings suggest that the diabetic peri-implant environment is characterized by diabetes-associated microbial alterations, insufficient anti-inflammatory responses, and an IL-17Family–associated inflammatory profile that is not fully ameliorated by IL-17A neutralization. However, the present data do not establish a causal relationship between microbial alterations, IL-17F upregulation, and local tissue destruction.

Several limitations should be considered. First, the 14-day ligature-induced model used in this study reflects active peri-implant inflammatory and osteolytic responses at the experimental endpoint, rather than a sustained chronic peri-implantitis process. In addition, implant survival prior to ligature placement was lower in db/db mice than in WT mice, and the marked body weight difference between genotypes may represent a potential confounding factor, including obesity-associated metabolic load and altered mechanical loading on the maxilla and alveolar bone. The use of machined, non-surface-modified titanium implants should also be considered, as clinically used rough-surface implants may show different patterns of osseointegration, biofilm accumulation, and peri-implant tissue responses. Second, although IL-17F was upregulated in peri-implant tissues of diabetic mice, its functional role was not directly tested by IL-17F-specific inhibition or combined IL-17A/IL-17F blockade. Similarly, Th17/Treg-related changes were assessed based on cytokine expression and CD4⁺FOXP3⁺ cell abundance in tissue sections, without direct evaluation of Treg suppressive function or flow cytometric quantification. Third, microbiome findings provide supportive and contextual evidence of diabetes-associated changes in the peri-implant environment, but do not establish a causal relationship between microbial alterations, IL-17 signaling, and bone loss. The splenocyte co-culture assay allowed comparison of microbiota-induced cytokine responses under standardized in vitro conditions, but it does not reproduce the local epithelial, stromal, and tissue-resident immune cell architecture of peri-implant tissues. Future studies using IL-17F-specific or dual IL-17A/IL-17F blockade, functional microbiome approaches, and local oral tissue-derived models will be needed to further validate these findings.

In summary, our study suggests that peri-implantitis inflammation and bone loss are exacerbated in diabetes and are characterized by enhanced IL-17A-mediated immune responses, accompanied by additional up-regulation of IL-17F and persistent microbial alteration. Although IL-17A neutralization significantly attenuated inflammation and bone loss, these changes were not fully normalized in diabetic mice. This suggests that IL-17A is an important but not exclusive contributor to diabetes-exacerbated peri-implant inflammation, while the functional role of IL-17F remains to be determined. It supports the concept of IL-17 targeting and microbiome modulation, in combination with glycemic control, to achieve long-term implant stability in diabetic patients.

Materials and methods

Animals

Wild-type (WT, C57BL/6J) and leptin receptor-deficient (db/db, B6.BKS(D)-Leprdb/J) mice (4 weeks old; Jackson Laboratory, Bar Harbor, ME) were used. Mice were housed under standard laboratory conditions and fed a solid diet until tooth extraction, after which they were switched to a powdered diet (Bio-Serv, Flemington, NJ, USA) to minimize mechanical disturbance to the surgical sites. All animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Nova Southeastern University (approval number: 2023.08.XH2) and were conducted in accordance with institutional and national guidelines and the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines. WT and db/db mice were randomly assigned to six experimental groups: WT control, WT + ligature + IgG (WT Lig + IgG), WT + ligature + anti-IL-17A (WT Lig+anti-IL-17A), DB control, DB + ligature + IgG (DB Lig + IgG), and DB + ligature + anti-IL-17A (DB Lig+anti-IL-17A). Each group initially consisted of 10 mice to ensure sufficient biological replicates after potential exclusions. Before ligature placement, implant retention was confirmed by visual inspection and gentle mobility testing of the implant head using a dental explorer. In WT mice, 24 of 30 implants were successfully retained before ligature placement, corresponding to an implant survival rate of 80.0%. In db/db mice, 18 of 30 implants were successfully retained before ligature placement, corresponding to an implant survival rate of 60.0%. Mice with implant loss or obvious implant mobility were excluded from subsequent ligature placement and downstream analyses. Final sample sizes for each analysis are indicated in the corresponding figure legends. All statistical analyses were performed using biological replicates (individual mice) as the experimental unit. Anti-IL-17A monoclonal antibody (Bio X Cell, Lebanon, NH, USA) or isotype-matched IgG control was administered via intragingival injection according to the experimental schedule described in this study to neutralize local IL-17A activity.

Tooth extraction, implantation, peri-implantitis induction, and peri-implant treatment

The entire surgical procedure was conducted as described in previous reports¹⁸. Briefly, four-week-old mice underwent extraction of the left maxillary first and second molars under isoflurane inhalation anesthesia using a nose cone (continuous flow of 1–2% isoflurane mixed with oxygen, 1 L/min) (Dechra Veterinary Products, Overland Park, KS, USA). To prevent postoperative infection, sulfamethoxazole/trimethoprim (850 µg/170 µg per mL) was administered in drinking water for two weeks after extraction. For postoperative analgesia, extended-release buprenorphine (0.5 mg/kg; Patterson Veterinary, MN, USA) was administered subcutaneously for three days. After a five-week healing period, a screw-shaped titanium implant with a machined surface and no additional surface treatment (1.0 mm in length and 0.5 mm in diameter) was placed into the healed edentulous ridge of the maxilla. Mice were allowed to heal for an additional four weeks to permit osseointegration. Implant stability was confirmed by visual inspection and gentle mobility testing of the implant head using a dental explorer before further experimental procedures. Experimental peri-implantitis was induced by subgingival placement of a 7 − 0 silk ligature around the implant head to promote plaque accumulation and localized inflammation. For local treatment, mice received intragingival injections adjacent to the implant site with either isotype-matched control IgG (0.5 µg per site) or anti-IL-17A neutralizing antibody (0.5 µg per site; Bio X Cell, Lebanon, NH, USA), according to the experimental schedule. Two weeks after ligature placement, mice were euthanized by CO₂ inhalation, and peri-implant tissues and maxillary bone specimens were harvested for subsequent micro-CT, histological, immunohistochemical, and molecular analyses. The experimental timeline and representative images of the surgical procedures are shown in Fig. 1A and B.

Body weight measurement, blood glucose detection

Body weight was recorded monthly in wild-type (WT) mice and weekly in diabetic (db/db) mice throughout the experimental period. Blood glucose levels were measured using a portable glucometer (OneTouch, LifeScan, Milpitas, CA, USA) with approximately 10 µL of blood collected from the tail vein. Because db/db mice exhibit metabolic instability and rapid fluctuations in blood glucose levels, measurements in this group were performed weekly to monitor glycemic status more accurately, whereas WT mice with stable glucose levels were monitored monthly. This differential monitoring schedule was applied to capture disease-related metabolic changes while minimizing unnecessary handling stress in WT mice.

Micro-computed tomography scanning and bone loss volume analysis

Maxillae containing the implants were fixed in 4% paraformaldehyde and scanned using a high-resolution micro-computed tomography system (SkyScan, Bruker, Billerica, MA, USA). Scanning was performed under standard settings for small-animal bone analysis (voxel size: 9 μm; voltage: 50 kV; current: 500 µA), and images were reconstructed using the manufacturer’s software (NRecon, Bruker) according to the recommended protocols. Three-dimensional morphometric analysis was conducted using CTAn software (Bruker). A standardized volume of interest (VOI) was defined around each implant as follows: mesial–distal boundaries were set at the width of the implant head; buccal–lingual boundaries extended from the median palatal suture to the buccal alveolar ridge; and coronal–apical boundaries were defined as a 1.0-mm region from the implant platform toward the apex, excluding the implant body. Within this VOI, total volume (TV) and bone volume (BV) were calculated, and empty space volume (ESV) was determined as ESV = TV − BV. Peri-implant bone loss was quantified as ESV (mm³) within the defined VOI. Representative three-dimensional reconstructions were generated using CTVox software (Bruker) for visualization.

Histology: hematoxylin and eosin staining

Maxillae with surrounding gingival tissues were fixed and decalcified in 10% EDTA at 4 °C for three weeks, with decalcification solution replaced every other day. After decalcification, titanium implants were carefully removed because they cannot be sectioned using standard paraffin microtomy. The remaining peri-implant tissues and residual implant socket area were then embedded in paraffin and sectioned longitudinally for histological analysis. Hematoxylin and eosin (H&E) staining was performed using standard histological protocols. Stained sections were imaged using a light microscope (Carl Zeiss, Jena, Germany). Inflammatory cell infiltration was quantified in peri-implant connective tissues using four standardized regions of interest (ROIs; 400 μm × 400 μm) per section at 40× magnification. Cell numbers were quantified using ImageJ software (NIH, version 1.54d), and the mean value of the four ROIs was calculated for each specimen and used for statistical analysis.

Osteoclastogenic activity: tartrate-resistant acid phosphatase (TRAP) staining

Maxillae with gingiva were decalcified in 10% EDTA, paraffin-embedded, and sectioned mesiodistally. TRAP staining was performed using a modified standard protocol, and sections were counterstained with hematoxylin. TRAP-positive multinucleated osteoclasts along the surface of alveolar bone were counted in mesial and distal regions of osteotomy under 40× magnification (Zeiss microscope). Values were averaged per sample and presented as fold change relative to WT controls.

Immunohistochemistry (IHC) and immunofluorescence (IF) staining

Paraffin-embedded maxillary sections were deparaffinized and rehydrated through graded ethanol, followed by antigen retrieval using standard heat-induced epitope retrieval methods. For immunohistochemistry (IHC), sections were incubated with primary antibodies against IL-17A (ab79056, 1:200, Abcam, Cambridge, UK) and IL-17F (210-401-B31, 1:200, Rockland Immunochemicals, PA, USA). Signal detection was performed using an HRP-based detection system (VECTASTAIN^® ABC-HRP Kit; Vector Laboratories, CA, USA) with ImmPACT^® VIP substrate, followed by hematoxylin counterstaining.

For immunofluorescence (IF) staining, sections were incubated with APC-conjugated anti-CD4 antibody (BioLegend, CA, USA) and rabbit anti-FOXP3 antibody (Abcam), followed by Alexa Fluor 488–conjugated secondary antibody (Invitrogen, Thermo Fisher Scientific, MA, USA). Nuclei were counterstained with DAPI. Fluorescent images were captured using an EVOS M5000 fluorescence microscope (Thermo Fisher Scientific), and IHC images were obtained using a Zeiss light microscope (Carl Zeiss, Jena, Germany). Quantitative analysis was performed using ImageJ software (NIH, version 1.54d). For each specimen, three representative sections were analyzed, and within each section, four standardized regions of interest (ROIs; 400 μm × 400 μm) in peri-implant connective tissues were evaluated at 40× magnification. For IHC, positively stained cells were manually counted and expressed as cell number per unit area (cells/mm²). For IF, signal intensity was quantified as mean integrated density per unit area (integrated density/mm²). Values from all ROIs were averaged to obtain a single representative value per animal for statistical analysis.

Quantitative RT-PCR

Total RNA was extracted from peri-implant gingival tissues (n = 3–4 mice per group) using the PureLink RNA Mini Kit (Invitrogen, Thermo Fisher Scientific, Carlsbad, CA, USA) according to the manufacturer’s instructions. RNA concentration and purity were assessed spectrophotometrically prior to downstream processing. Complementary DNA (cDNA) was synthesized using the Verso cDNA Synthesis Kit (Thermo Fisher Scientific). Quantitative real-time PCR was performed using PowerUp SYBR Green Master Mix (Applied Biosystems, Thermo Fisher Scientific) with gene-specific primers on a real-time PCR system under standard cycling conditions. All reactions were performed in technical duplicates. Relative gene expression levels were calculated using the comparative ΔΔCt method with Gapdh as the internal reference gene. Primer sequences used in this study are listed in Supplementary Table 1.

Bacterial culture

To obtain peri-implant–associated microbiota for co-culture experiments, silk ligatures were carefully removed from implant sites and rinsed in sterile phosphate-buffered saline (PBS). Biofilm samples were cultured on Luria–Bertani (LB) agar and blood agar plates under both aerobic and anaerobic conditions to recover a broad spectrum of oral bacteria. To preserve microbial diversity, biofilm isolation and expansion were adapted from a previously established in vitro oral biofilm culture model³⁸, with modification by using ligature-associated microbiota from murine peri-implant sites as the inoculum.

After visible colony formation, all colonies from each sample were pooled and inoculated into LB broth, followed by incubation for 18 h to generate bacterial suspensions. Bacteria were then fixed in 4% paraformaldehyde for 30 min at room temperature and washed thoroughly with sterile PBS to remove residual fixative. Fixed bacterial preparations were used for subsequent co-culture with primary murine splenocytes.

Co-culture of bacteria with splenocytes

For each sample, bacterial suspensions obtained from aerobic and anaerobic cultures were pooled at a 1:1 ratio before fixation and co-culture with splenocytes. The pooled bacterial suspensions were washed twice with sterile phosphate-buffered saline (PBS) by centrifugation (6,000 × g, 5 min, 4 °C) and fixed in 4% paraformaldehyde for 30 min at room temperature. After fixation, bacteria were washed thoroughly with PBS to remove residual fixative. Bacterial concentrations were adjusted based on optical density measured at 600 nm (OD600) using a spectrophotometer (NanoDrop, Thermo Fisher Scientific), with OD600 = 0.1 corresponding to approximately 1 × 10⁸ CFU/mL. Splenocytes were isolated from spleens of C57BL/6J mice by mechanical dissociation, followed by red blood cell lysis (Gibco), filtration through a 70-µm cell strainer, and resuspension in Iscove’s Modified Dulbecco’s Medium (IMDM) supplemented with 10% fetal bovine serum (FBS) and 1% antibiotic–antimycotic solution. Cells were seeded in 12-well plates at a density of 1 × 10⁵ cells per well and co-cultured with fixed bacteria at a bacteria-to-cell ratio of 500:1. Co-cultures were incubated for 12 h at 37 °C in a humidified atmosphere containing 5% CO₂. After incubation, splenocytes were collected, washed with PBS, and subjected to total RNA extraction using the PureLink RNA Mini Kit (Invitrogen, Thermo Fisher Scientific) for subsequent quantitative RT-PCR analysis of inflammatory cytokine gene expression.

16S rRNA gene sequencing and sample processing

Oral microbiota samples were collected at the end of the experimental period. For ligature-induced peri-implantitis groups, ligatures were carefully removed from implant sites and rinsed in sterile phosphate-buffered saline (PBS) to recover biofilm-associated microorganisms. For control groups (implant without ligature), samples were obtained by swabbing peri-implant sites using sterile PBS-soaked swabs. All samples were immediately snap-frozen in liquid nitrogen and stored at − 80 °C until further processing. Microbial DNA was extracted using the DNeasy^® PowerSoil^® Pro Kit (QIAGEN, Hilden, Germany) according to the manufacturer’s instructions, and DNA concentration and purity were assessed using a NanoDrop spectrophotometer (Thermo Fisher Scientific, MA, USA). Amplicon libraries targeting the V3–V4 hypervariable regions of the 16S rRNA gene were prepared using standard Illumina-compatible protocols and sequenced by a commercial service provider (CD Genomics, Shirley, NY, USA) on an Illumina MiSeq platform with paired-end reads (2 × 300 bp). Raw sequencing reads were processed using the DADA2 pipeline for quality filtering, error correction, chimera removal, and inference of amplicon sequence variants (ASVs). To minimize technical batch effects, datasets were processed independently using identical parameters prior to merging into a single feature table. Downstream analyses were performed using QIIME2 (version 2025.4) as described previously³⁹. Alpha diversity was evaluated using observed features, Chao1 richness, Shannon index, and Simpson index. Beta diversity was assessed based on distance matrices and visualized by principal coordinate analysis (PCoA). Taxonomic classification was performed using a reference database implemented in QIIME2, and relative abundance profiles were visualized as heatmaps to identify group-associated microbial shifts.

For statistical analysis of microbiome data, beta diversity was assessed using Bray–Curtis dissimilarity and visualized by principal coordinate analysis (PCoA). Differences in beta diversity between groups were evaluated using PERMANOVA with 999 permutations in QIIME2. A p-value < 0.05 was considered statistically significant. Differences in the relative abundance of taxa between groups were assessed using appropriate non-parametric tests implemented in QIIME2.

Statistical analysis

All statistical analyses were performed using GraphPad Prism version 10 (GraphPad Software, San Diego, CA, USA). Data are presented as mean ± standard error of the mean (SEM), unless otherwise indicated. Individual mice were considered the experimental units for all in vivo analyses. For datasets involving genotype and treatment as two independent factors, two-way ANOVA was used to evaluate the main effects of genotype, treatment, and their interaction. When appropriate, post hoc multiple-comparison tests were performed to evaluate treatment-induced changes within each genotype. Comparisons between two groups were performed using an unpaired two-tailed Student’s t-test. Normality of data distribution was assessed prior to parametric testing. A p-value < 0.05 was considered statistically significant.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

T M ; Investigation, data curation, Interpretation of data, Writing-original draft. J C J : Investigation, reviewed and edited manuscript. T C ; sequence data analysis, interpretation of data, reviewed and edited manuscript. G C : Investigation, reviewed and edited manuscript. N K : investigated and interpreted data, reviewed and edited manuscript. E D A , M O , S S , S Y, S R, A M, and A A ; data curation, reviewed and edited manuscript. X He ; sequence data analysis, interpretation of data, reviewed and edited manuscript. S V ; Supervision. M S ; Supervision. J L ; Supervision. T K ; Supervision. X Han ; Project Administration/oversight, Conceptualization, Writing—review, editing, and revision.

Funding

This study was supported by a grant from the National Institutes of Dental and Craniofacial Research (NIDCR) DE032156 to XH, DE029709 to TK, and DE027648 to MS. The authors would like to acknowledge the use of NSU’s Vivarium Facility and the Bruker SkyScan 1276 microCT acquired with funding from the National Institutes of Health through grant number S10 OD032199-01A1.

Data availability

The datasets generated and/or analysed during the current study are available in the NCBI Sequence Read Archive (SRA) repository under BioProject accession number PRJNA1433895.

Declarations

Competing interests

The authors declare no competing interests.

Ethics statement

All procedures were approved by the IACUC of Nova Southeastern University (protocol #2023.08.XH2). The conduct of this study adhered to the ARRIVE guidelines.