A DPP inhibitor suppresses periodontitis via antibacterial effect targeting Porphyromonas gingivalis Authors: Yukari Aoki-Nonaka, Yukako Minato, Koushi Hidaka, Yuko Warita, Daiki Ando, Hnin Yu Lwin, Aoi Matsugishi-Nasu, Naoki Takahashi, Akihiro Nakamura, Wataru Ogasawara, Mizuki Sekiya, Yasumitsu Sakamoto, Koichi Tabeta Published: 2026 ABSTRACT Porphyromonas gingivalis is the most common periodontal pathogen. P. gingivalis dipeptidyl peptidase 7 (PgDPP7) belongs to a new class of serine peptidases, family S46. S46 peptidases are absent in mammals. Therefore, these enzymes are promising targets for novel antibacterial agents. In this study, inhibitors were designed based on the cocrystal structures of valyl-tyrosine and phenylalanyl-tyrosine, which bind to the active centers of DPP7 derived from bacteria, and dipeptide derivatives that inhibit PgDPP7 were obtained. The active compound KGDI-109, the first peptidyl inhibitor of S46 peptidases, exerted an inhibitory effect against P. gingivalis growth at a minimum inhibitory concentration of 1.56 µM. In C57BL/6 N male mice with induced periodontitis, the oral administration of KGDI-109 significantly suppressed alveolar bone resorption and reduced the amount of P. gingivalis in the oral cavity, indicating that the DPP inhibitor suppresses periodontal disease by its antibacterial activity. This dipeptide derivative did not inhibit the growth of other oral bacteria, and its antibacterial action was presumed to target bacteria possessing DPP, particularly P. gingivalis. Furthermore, KGDI-109 may be more effective than azithromycin in maintaining the gut microbial diversity and reducing adverse health effects. KGDI-109 can be a novel treatment for periodontal diseases targeting P. gingivalis.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-52648-8. Introduction Periodontitis is an inflammatory disease caused by infection with periodontal bacteria, which results in the destruction of periodontal tissues1. Among periodontal pathogens, Porphyromonas gingivalis has long been regarded as a major causative bacterium due to its virulence factors and its role in severe periodontitis. Hajishengallis et al. proposed that P. gingivalis is a keystone bacterium that induces dysbiosis in the oral microbiota, thereby initiating the onset of periodontitis2–4. However, due to the lack of appropriate tools, its causal role in humans remains unverified. To address this issue and translate the etiological concept to therapeutic strategies, antibiotics specific to P. gingivalis could provide direct evidence for its causal role. P. gingivalis generates its metabolic energy by fermenting amino acids; extracellular oligopeptides are degraded into dipeptides by dipeptidyl peptidases (DPPs)5,6. These degraded peptides are taken up into the cytoplasm and serve as nutrients essential for P. gingivalis growth7. DPP7 and DPP11 are members of the S46 exopeptidase family in P. gingivalis8,9. S46 peptidases are widely distributed in anaerobic Gram-negative species of the genera Bacteroides, Parabacteroides, and Porphyromonas, but they are not found in mammals9,10. Antibiotics that target molecules absent in mammals may be associated with a lower risk of side effects; therefore, these peptidases represent promising targets for the development of new antibacterial agents. Sakamoto et al. developed the DPP11 inhibitor SH-5, which exhibits antibacterial activity against P. gingivalis11, but no DPP7 inhibitor has been developed. P. gingivalis DPP7 (PgDPP7) is also important for metabolism; therefore, a DPP7 inhibitor could be a specific antibiotic to P. gingivalis. Therefore, this study aimed to develop a new peptide-type PgDPP7 inhibitor and investigate its effects on periodontal pathogenic bacteria and its efficacy in preventing periodontal diseases. P. gingivalis-specific therapeutics represent an indispensable tool for definitively establishing its causal role in human periodontitis. Materials and methods Reagents KGDI-109, a DPP7 inhibitor, was synthesized (Figs. S1 and S2). Details regarding reagents are provided in the Supplementary Information. Kinetic characterization and inhibition assays The full nucleotide sequence of the codon-optimized gene and the details on protein production and purification of DPP7 are provided in the Supplemental Materials and Methods and Fig.S3. Protein overproduction and purification were performed as previously described12. The kinetic parameters of purified DPP7 (10 nM) were determined by nonlinear least-squares fitting of the experimental data to the Michaelis–Menten equation using Excel Solver (Microsoft, Redmond, WA, USA), with methionyl-l-leucyl-4-methylcoumaryl-7-amide (1.56, 3.13, 6.25, 12.5, 25, 50, 100, and 200 µM; Met-Leu-MCA; Peptide Institute, Osaka, Japan) as the substrate. The kcat was calculated based on the total protein concentration and reported as an apparent value, as active-site titration was not performed. Inhibition assays were performed using a substrate concentration of 100 µM. Briefly, purified DPP7 (10 nM) was preincubated at 25 °C in a reaction buffer containing 50 mM sodium phosphate (pH 7.0), 5 mM EDTA, 0.005% Tween 20, and various concentrations (1.56–100 µM) of the inhibitor KGDI-109. Reactions were initiated by adding the Met-Leu-MCA substrate (100 µM), followed by continuous fluorescence monitoring every 1 min over 20 min at 25 °C (multipoint kinetic assay). The fluorescence intensity resulting from MCA release was measured using an Infinite 200 PRO microplate reader (Tecan, Männedorf, Switzerland) with excitation at 355 nm and emission at 460 nm. The half-maximal inhibitory concentrations (IC50 values) were determined by fitting experimental data from three independent experiments to a three-parameter logistic curve (LL.3) using the drc package in R (version 4.3.0)13. To account for the technical limitation of the 100 µM maximum inhibitor concentration (imposed by the 1.0% dimethyl sulfoxide limit), the bottom plateau of the curve was constrained to zero. The inhibition constants (Ki) were calculated using the Cheng-Prusoff equation14. The competitive nature of the inhibition was validated based on the Ki and IC50 relationship. Results are reported as the mean with 95% confidence intervals (CI). Bacterial culture P. gingivalis W83, P. gingivalis ATCC 33277, Fusobacterium nucleatum ATCC 25586, and Prevotella intermedia ATCC 25611 were cultured in modified Gifu anaerobic medium broth (Nissui, Tokyo, Japan) in an anaerobic jar in the presence of AnaeroPack™ (Thermo Fisher Scientific, Waltham, MA, USA) at 37 °C. Aggregatibacter actinomycetemcomitans was cultured in tryptic soy broth supplemented with 6 mg/mL yeast extract and 0.4 mg/mL sodium bicarbonate in an anaerobic jar in the presence of AnaeroPack™ at 37 °C. Streptococcus mitis ATCC 903 was cultured in brain heart infusion broth (Thermo Fisher Scientific) at 37 °C under aerobic conditions. The bacterial suspension concentration was determined by measuring the optical density at 600 nm using a UV-visible Spectronic Genesys 10 Bio spectrophotometer (Thermo Fisher Scientific). Colony-forming units (CFU) were standardized using the growth curve of each bacterium15. Inhibition of DPP enzyme activity P. gingivalis W83 was diluted to 1 × 108 CFU/mL, and 100 µL of the cell suspension was transferred to a 96-well plate and incubated with KGDI-109 (0–50 µM) in an anaerobic jar in the presence of AnaeroPack™ at 37 °C for 30 min. Each peptidyl-MCA substrate (PEPTIDE INSTITUTE, INC., Osaka, Japan) was added at 20 µM and incubated for 1 h. The fluorescent emitter produced by hydrolysis of the peptidyl-MCA substrates was measured with excitation at 365 nm and emission at 415–445 nm using GloMax® Explorer (Promega, Madison, WI, USA). Met-Leu-MCA has an amino acid sequence recognized by DPP5 and DPP716,17, whereas Phe-Met-MCA has an amino acid sequence characteristically recognized by DPP717. The controls Gly-Pro-MCA and Leu-Asp-MCA have amino acid sequences characteristically recognized by DPP4 and DPP11, respectively9,16. Evaluation of the antibacterial activity against oral bacteria The minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) for KGDI-109 were determined using a microplate dilution assay15,18. P. gingivalis W83, F. nucleatum, P. intermedia, A. actinomycetemcomitans, and S. mitis were diluted to 1 × 106 CFU/mL, and 100 µL of the cell suspension of each bacterium was transferred to a 96-well plate and incubated in the presence of KGDI-109 (0–200 µM). P. gingivalis and F. nucleatum were incubated for 72 h, P. intermedia was incubated for 48 h, and A. actinomycetemcomitans and S. mitis were incubated for 24 h. The bacterial suspension concentration was determined by measuring the optical density at 600 nm using a UV-visible Spectronic Genesys 10 Bio spectrophotometer (Thermo Fisher Scientific). For the time-killing assay, P. gingivalis W83 was diluted to 1 × 108 CFU/mL and treated with KGDI-109 and azithromycin (AZM) at 200 µM and AZM at 1.98 µM. The killing curves were determined using the ATP assay (Promega). Biofilm quantification P. gingivalis ATCC 33277 was diluted to 1 × 108 CFU/mL. The bacterial suspension was transferred into 96-well flat-bottomed plates (Techno Plastic Products AG, Trasadingen, Switzerland) with final KGDI-109 concentrations of 0–200 µM and incubated for 48 h. Biofilm quantification was performed using a crystal violet (Chroma-Gesellschaft Co. Ltd., Münster, Germany) staining method19. Cell viability testing The human monocyte cell line THP-1 (1 × 105 cells/well) was incubated for 48 h with 10 ng/mL phorbol 12-myristate 13-acetate (Sigma-Aldrich) in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin–streptomycin to differentiate macrophage-like cells. The human oral epithelial cell line Ca9-22 (1 × 105 cells/well) was cultured in Dulbecco’s modified Eagle’s medium (Thermo Fisher Scientific) supplemented with 10% FBS and 1% penicillin–streptomycin. These cells were treated with KGDI-109 for 24 h at 37 °C and 5% CO2. Cell viability was assessed using thiazolyl blue tetrazolium bromide (Sigma-Aldrich). Absorbance was measured at 570 nm using SpectraMax® ABS Plus. Mice Nine-week-old male C57BL/6 N/Crl mice (purchased from The Jackson Laboratory Japan, Inc., Kanagawa, Japan) were maintained under specific pathogen-free conditions and fed regular chow. The details are presented in Supplementary Information. A total of 32 mice were randomly divided into the following groups for the collection of alveolar bone, ligature, and feces samples, as well as body weight data: control group ( no-treatment control : nonligation and noninfection), a periodontitis-induced group ( periodontitis control : ligation and P. gingivalis infection), a KGDI-109-administered periodontitis group (ligation, P. gingivalis infection, and KGDI-109 administration), and an AZM-administered periodontitis group (ligation, P. gingivalis infection, and AZM administration). Several samples were excluded from the analysis owing to the following reasons: injury, loss of ligature thread, and deficiencies in the sample collection process. A total of 16 mice were randomly assigned to the aforementioned groups for a histological analysis of gingival tissue sections. The thread was tied to the mouse under anesthesia, and the mouse was euthanized after the experiment. In this experiment, euthanasia was performed via cervical dislocation under anesthesia before the collection of various samples. Anesthesia was administered via intraperitoneal injection of a mixture of medetomidine (0.3 mg/kg body weight), midazolam (4 mg/kg body weight), and butorphanol tartrate (5 mg/kg body weight) at a dosage of 0.1 mL per 10 g body weight (mice body weight: 20–25 g at the start of the experiment). All animal experiments were approved by the Animal Care and Use Committee of Niigata University and conducted in accordance with the Regulations and Guidelines on Scientific and Ethical Care and Use of Laboratory Animals of the Science Council of Japan and the ARRIVE guidelines. Periodontitis induction in mice and KGDI-109 administration A 5 − 0 silk thread was gently ligated around the left maxillary second molar under anesthesia. The threads were kept in place for 7 days. During this period, P. gingivalis W83 (2 × 109 CFU), suspended in 200 µL phosphate-buffered saline and 4.5% carboxymethyl cellulose (FUJIFILM Wako Pure Chemical Corporation), was orally administered using a feeding needle once a day to the periodontitis-induced group, whereas the vehicle alone was administered to the control group. KGDI-109 was prepared at a dose of 20 mg/kg body weight in sterile water and administered orally as drinking water. Alveolar bone resorption quantification The maxillary bones were stained with 0.5% methylene blue to delineate the cementoenamel junctions and alveolar bone crest. The exposed root area showing alveolar bone resorption surrounded by the cementoenamel junctions and alveolar bone crest of the maxillary second molar was measured using a stereomicroscope (Stemi 305; ZEISS, Jena, Germany)20. All measurements were performed independently by two masked examiners (HYL and DA). Histological analysis of periodontal tissue The maxillae were fixed in 4% paraformaldehyde phosphate buffer solution for 24 h, decalcified in 10% EDTA for 2 weeks, mounted in paraffin, and sectioned at a thickness of 4 μm in the sagittal direction along the long axis of the teeth. The sections were stained with hematoxylin and eosin (H-E; FUJIFILM Wako Pure Chemical Corporation) solutions and a tartrate-resistant acid phosphatase (TRAP) staining kit (Cosmo Bio Corporation, Tokyo, Japan) and imaged using a fluorescence microscope BZ-X710 (KEYENCE, Osaka, Japan). Each image selected for analysis was representative of its group. Quantification of the bacterial amount All 5 − 0 silk threads collected were adjusted to 3 mm in length. Bacterial DNA was isolated using DNeasy Blood & Tissue Kits (QIAGEN, Hilden, Germany). DNA was mixed with PowerUp™ SYBR™ Green Master Mix (Thermo Fisher Scientific) for qPCR and analyzed using QuantStudio™ 1 Real-Time PCR Instrument (Thermo Fisher Scientific). The total bacterial value was assumed to represent the biofilm amount, and normalization was performed by dividing the value for P. gingivalis (Δ42-Ct) by the total bacterial value (Δ42-Ct). The oligonucleotides used are shown in Table S1. Intestinal microflora analysis The following procedures were performed under contract to Bioengineering Lab. Co., Ltd. (Kanagawa, Japan). The details are shown in Supplementary Information. DNA was extracted from the fecal samples. The libraries were prepared using a 2-step tailed PCR method. The primers used are shown in Table S2. The V3/V4 variable region (341f-805r) of the 16S rDNA in the samples was sequenced using the MiSeq system and MiSeq Reagent Kit v3 (Illumina, Inc., San Diego, CA, USA) at 2 × 300 bp. The representative sequences were compared with 97% operational taxonomic units using the Greengene database (version 13_8). Statistical analysis Statistical analyses were performed using GraphPad Prism version 7.0 (GraphPad Software, Inc., Boston, MA, USA). The statistical analysis details are shown in the figure legends. Results Design of KGDI-109, a DPP7 inhibitor The novel dipeptide derivative KGDI-109 is a compound that targets P. gingivalis and comprises P2 4-nitro-phenylalanine and P1 O-methyltyrosine isopropyl ester (Fig. 1a). This design is based on the X-ray cocrystal structure analysis of DPP7, which revealed numerous water molecules in the S1 and S2 pockets interacting with Val-Tyr while leaving unoccupied space21, and the synthesis of dipeptide analogs with several modifications to the side chains and C-terminus of Val-Tyr and Phe-Tyr12 targeting Stenotrophomonas maltophilia DPP7. KGDI-109 is a modified analog designed based on comparisons of substituents and their functions. Specifically, it combines, a P2-substituted phenylalanine, which exhibited strong P. gingivalis growth inhibition, with an isopropyl ester at the C-terminal structure, which exhibited low cytotoxicity and stability against human carboxylesterase (Patents JP7228828 and JP7836044). KGDI-109 was synthesized using conventional peptide condensation and deprotection reactions (Fig. S1). Starting with tert-butyloxycarbonyl (Boc)-protected O-methyltyrosine, the carboxylic acid was converted to an isopropyl ester using 2-bromopropane. After Boc deprotection, the Boc-protected 4-nitro-phenylalanine was coupled using the 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide-1-hydroxybenzotriazole method. The intermediate was purified by flash chromatography, deprotected, and recrystallized to give KGDI-109 as the hydrochloride form in high yield with > 98% purity (Fig. S2). Fig. 1Development of the DPP7 peptide inhibitor. (a) Chemical structure of KGDI-109. (b) Characterization and inhibition assay using purified PgDPP7. SDS–PAGE analysis of purified PgDPP7, showing a single band corresponding to the theoretical molecular weight of 77.8 kDa (left). Purified enzyme (2 µg) was subjected to SDS–PAGE using a 10% acrylamide gel; a representative cropped image is shown, and the original, full-length gel image is provided in Supplementary Figure S4. Michaelis–Menten plot of PgDPP7 activity against the Met-Leu-MCA substrate (center). Inhibition curves of PgDPP7 activity by KGDI-109 in the presence of 100 µM substrate (right). Error bars represent the standard deviations from three independent experiments. IC50 and Ki and values represent the 95% CI from three independent experiments. (c) Inhibition of DPP7 enzyme activity by KGDI-109 in P. gingivalis. The fluorescent emitter produced by the hydrolysis of the peptidyl-MCA substrates was measured (n = 3). P. gingivalis was preincubated with KGDI-109 (0–50 µM), and the peptidyl-MCA substrates were added at 20 µM and incubated for 1 h. The peptidyl-MCA substrates, Met-Leu-MCA, Phe-Met-MCA, Gly-Pro-MCA, and Leu-Asp-MCA, are shown in the figure as substrates for DPP5/7, DPP7, DPP4, and DPP11, respectively. Data are presented as means ± SD from three independent experiments. Significant values are indicated by ***P < 0.001 vs. 0 µg/mL. Comparisons were performed using Dunnett’s multiple comparison test. Development of the DPP7 peptide inhibitor. (a) Chemical structure of KGDI-109. (b) Characterization and inhibition assay using purified PgDPP7. SDS–PAGE analysis of purified PgDPP7, showing a single band corresponding to the theoretical molecular weight of 77.8 kDa (left). Purified enzyme (2 µg) was subjected to SDS–PAGE using a 10% acrylamide gel; a representative cropped image is shown, and the original, full-length gel image is provided in Supplementary Figure S4. Michaelis–Menten plot of PgDPP7 activity against the Met-Leu-MCA substrate (center). Inhibition curves of PgDPP7 activity by KGDI-109 in the presence of 100 µM substrate (right). Error bars represent the standard deviations from three independent experiments. IC50 and Ki and values represent the 95% CI from three independent experiments. (c) Inhibition of DPP7 enzyme activity by KGDI-109 in P. gingivalis. The fluorescent emitter produced by the hydrolysis of the peptidyl-MCA substrates was measured (n = 3). P. gingivalis was preincubated with KGDI-109 (0–50 µM), and the peptidyl-MCA substrates were added at 20 µM and incubated for 1 h. The peptidyl-MCA substrates, Met-Leu-MCA, Phe-Met-MCA, Gly-Pro-MCA, and Leu-Asp-MCA, are shown in the figure as substrates for DPP5/7, DPP7, DPP4, and DPP11, respectively. Data are presented as means ± SD from three independent experiments. Significant values are indicated by ***P < 0.001 vs. 0 µg/mL. Comparisons were performed using Dunnett’s multiple comparison test. Inhibition of P. gingivalis DPP7 by KGDI-109 The inhibitory effect of KGDI-109 on purified PgDPP7 was investigated. PgDPP7 was produced in E. coli and purified by liquid chromatography, resulting in a homogeneous enzyme preparation that showed a single band on SDS–PAGE corresponding to its theoretical molecular weight of 77.8 kDa. The kinetic parameters for Met-Leu-MCA were determined using this purified enzyme, and the kcat and Km values were 0.197 ± 0.004 s− 1 and 87.8 ± 2.6 µM, respectively. The IC50 and Ki values in inhibition assays with KGDI-109, conducted in the presence of 100 µM substrate, were determined to be 53.8 µM (95% CI: 50.2–57.4) and 25.1 µM (95% CI: 23.5–26.8), respectively (Fig. 1b). The competitive nature of this inhibition was supported by the design of KGDI-109 as a substrate-mimetic dipeptide analog. Furthermore, it was validated by the kinetic relationship where the Ki value was approximately half of the IC50 at a substrate concentration (100 µM) near the Km (87.8 µM), which is consistent with the theoretical requirement for competitive inhibition. The specific inhibition of DPP7 enzyme activity by KGDI-109 was evaluated. The fluorescence emission of the MCA substrates designed for each DPP was measured upon bacterial uptake and hydrolysis. KGDI-109 administration inhibited the degradation of DPP7 and DPP5 substrates in a concentration-dependent manner. Furthermore, it exerted a stronger inhibitory effect on the degradation of substrates that react more specifically with DPP7. On the other hand, it did not inhibit the degradation of substrates corresponding to DPP4 and DPP11 (Fig. 1c). KGDI-109 exhibited a selective antibacterial effect against P. gingivalis The MIC and MBC of KGDI-109 were determined to show its antibacterial effects (Table 1). KGDI-109 showed an inhibitory effect against P. gingivalis at an MIC of 1.56 µM but did not show an inhibitory effect against F. nucleatum, P. intermedia, A. actinomycetemcomitans, or Streptococcus mitis. Table 1MIC and MBC Values.(µM)KGDI-109AZMERYCHXMICMBCMICMBCMICMBCMICMBC Porphyromonas gingivalis 1.566.250.390.390.20.398.88.8 Fusobacterium nucleatum > 200> 2000.390.7812.5507070 Prevotella intermedia > 200> 2000.10.20.10.13535 Aggregatibacter actinomycetemcomitans > 200> 2000.20.780.391.5617.517.5 Streptococcus mitis > 200> 2000.23.130.051.567070AZM, azithromycin; ERY, erythromycin; CHX, chlorhexidine; MIC, minimum inhibitory concentration; MBC, minimum bactericidal concentration. Table 1MIC and MBC Values.(µM)KGDI-109AZMERYCHXMICMBCMICMBCMICMBCMICMBC Porphyromonas gingivalis 1.566.250.390.390.20.398.88.8 Fusobacterium nucleatum > 200> 2000.390.7812.5507070 Prevotella intermedia > 200> 2000.10.20.10.13535 Aggregatibacter actinomycetemcomitans > 200> 2000.20.780.391.5617.517.5 Streptococcus mitis > 200> 2000.23.130.051.567070AZM, azithromycin; ERY, erythromycin; CHX, chlorhexidine; MIC, minimum inhibitory concentration; MBC, minimum bactericidal concentration. MIC and MBC Values. AZM, azithromycin; ERY, erythromycin; CHX, chlorhexidine; MIC, minimum inhibitory concentration; MBC, minimum bactericidal concentration. Conversely, other antibacterial drugs, such as AZM, ERY, and CHX, showed a broad antibacterial spectrum against all of these bacteria. KGDI-109 was compared with AZM. The growth-inhibitory activity of KGDI-109 against P. gingivalis was concentration-dependent (Fig. 2a). The antibacterial kinetics study showed that KGDI-109 significantly reduced bacterial ATP activity in a relatively shorter time compared with AZM (Fig. 2b). Additionally, KGDI-109 significantly inhibited biofilm formation by P. gingivalis (Fig. 2c). The cytotoxicity study results are shown in Fig. 2d. Ca9-22 and THP-1 cells were exposed to KGDI-109 for 24 h. KGDI-109 at 200 µM did not affect cell viability in either Ca9-22 or THP-1 cells. Fig. 2KGDI-109 had an antibacterial effect against P. gingivalis. (a) Growth rate of P. gingivalis with KGDI-109. The rate was calculated for each sample using the following formula: (average OD value of sample − average OD value of medium) / (average OD value of control − average OD value of medium) × 100. P. gingivalis was diluted to 1 × 106 CFU/mL and treated with 0–100 µM KGDI-109 for 72 h (n = 3). Significant values are indicated by **P < 0.01 and ***P < 0.001, vs. 0 µg/mL. (b) The antibacterial kinetics of KGDI-109. P. gingivalis was diluted to 1 × 108 CFU/mL and treated with KGDI-109 at 200 µM and AZM at 1.98 µM for 0–24 h (n = 4). Significant values are indicated by ***P < 0.001 (control vs. KGDI-109), †††P < 0.001 (control vs. AZM), and ‡ ‡ ‡P < 0.001 (KGDI-109 vs. AZM). Comparisons were performed using Dunnett’s multiple comparison test. (c) The inhibition of biofilm formation by KGDI-109 or AZM against P. gingivalis. The remaining amount of biofilm was evaluated using crystal violet staining (n = 4). (d) Cell viability test. Ca9-22 and THP-1 were treated with KGDI-109 for 24 h and analyzed using the thiazolyl blue tetrazolium bromide assay (Ca9-22: n = 3; THP-1: n = 4). Data are presented as means ± SD. Comparisons were performed using Dunnett’s multiple comparison test. Significant values are indicated by *P < 0.05, **P < 0.01, and ***P < 0.001, vs. 0 µg/mL. n.s. = not significant. Three repeated experiments were conducted. KGDI-109 had an antibacterial effect against P. gingivalis. (a) Growth rate of P. gingivalis with KGDI-109. The rate was calculated for each sample using the following formula: (average OD value of sample − average OD value of medium) / (average OD value of control − average OD value of medium) × 100. P. gingivalis was diluted to 1 × 106 CFU/mL and treated with 0–100 µM KGDI-109 for 72 h (n = 3). Significant values are indicated by **P < 0.01 and ***P < 0.001, vs. 0 µg/mL. (b) The antibacterial kinetics of KGDI-109. P. gingivalis was diluted to 1 × 108 CFU/mL and treated with KGDI-109 at 200 µM and AZM at 1.98 µM for 0–24 h (n = 4). Significant values are indicated by ***P < 0.001 (control vs. KGDI-109), †††P < 0.001 (control vs. AZM), and ‡ ‡ ‡P < 0.001 (KGDI-109 vs. AZM). Comparisons were performed using Dunnett’s multiple comparison test. (c) The inhibition of biofilm formation by KGDI-109 or AZM against P. gingivalis. The remaining amount of biofilm was evaluated using crystal violet staining (n = 4). (d) Cell viability test. Ca9-22 and THP-1 were treated with KGDI-109 for 24 h and analyzed using the thiazolyl blue tetrazolium bromide assay (Ca9-22: n = 3; THP-1: n = 4). Data are presented as means ± SD. Comparisons were performed using Dunnett’s multiple comparison test. Significant values are indicated by *P < 0.05, **P < 0.01, and ***P < 0.001, vs. 0 µg/mL. n.s. = not significant. Three repeated experiments were conducted. KGDI-109 inhibited periodontal tissue destruction in an experimental murine model of periodontitis The effect of KGDI-109 on periodontal disease was investigated in a mouse model of periodontitis caused by oral P. gingivalis infection. No abnormalities in health status or body weight were observed following KGDI-109 administration (Fig. 3a). KGDI-109 administration significantly inhibited bone resorption in periodontitis-induced mice (Figs. 3b and c). In gingival tissue sections stained with H-E, attachment loss was observed in the periodontitis-induced group compared with the control group, and it was suppressed in the KGDI-109-treated group (Figs. 3d and e). In the same experiment, osteoclast activities in periodontal destruction were assessed by evaluating the number of TRAP-positive cells. The number of TRAP-positive cells adjacent to the bone surface in the interalveolar septa was significantly decreased in KGDI-109-treated mice compared with periodontitis-induced mice (Figs. 3f and g). Fig. 3KGDI-109 inhibited alveolar bone destruction in an experimental murine model of periodontitis. (a) Change in body weight. The change in body weight is expressed as a percentage, and the average weight is set as 100% for each group at the start of the experiment. Data are presented as means ± SD. Significant values are indicated as *P < 0.05 and ***P < 0.001 (the AZM group vs. the periodontitis group). (b) Representative stereomicroscope images of the maxillae from each group. (c) Quantification of alveolar bone destruction: measurement of the exposed root surface area. Data are presented as means ± SD (control: n = 7, periodontitis: n = 7, KGDI-109-periodontitis: n = 6, and AZM-periodontitis: n = 5). (d) Histological findings of gingival tissue sections stained with H-E. (e) Quantification of attachment loss level. The attachment loss level was measured in each section as the distance from the cementoenamel junctions to the gingival margin of the distal first and mesial second molars in the maxilla. Data are presented as means ± SD (n = 4). (f) Representative TRAP staining images of the interalveolar septal region. (g) The number of TRAP-positive osteoclasts per millimeter of the alveolar bone surface was measured using ImageJ 1.53t software (n = 4). Scale bars = 100 μm. Black arrows indicate the locations of the cementoenamel junctions, and black triangles indicate the locations of the alveolar bone crest. B = bone, D = dentin, and P = pulp. Comparisons were performed using Dunnett’s multiple comparison test. Significant values are indicated as *P < 0.05, **P < 0.01, and ***P < 0.001, vs. the periodontitis group. Three repeated experiments were conducted. KGDI-109 inhibited alveolar bone destruction in an experimental murine model of periodontitis. (a) Change in body weight. The change in body weight is expressed as a percentage, and the average weight is set as 100% for each group at the start of the experiment. Data are presented as means ± SD. Significant values are indicated as *P < 0.05 and ***P < 0.001 (the AZM group vs. the periodontitis group). (b) Representative stereomicroscope images of the maxillae from each group. (c) Quantification of alveolar bone destruction: measurement of the exposed root surface area. Data are presented as means ± SD (control: n = 7, periodontitis: n = 7, KGDI-109-periodontitis: n = 6, and AZM-periodontitis: n = 5). (d) Histological findings of gingival tissue sections stained with H-E. (e) Quantification of attachment loss level. The attachment loss level was measured in each section as the distance from the cementoenamel junctions to the gingival margin of the distal first and mesial second molars in the maxilla. Data are presented as means ± SD (n = 4). (f) Representative TRAP staining images of the interalveolar septal region. (g) The number of TRAP-positive osteoclasts per millimeter of the alveolar bone surface was measured using ImageJ 1.53t software (n = 4). Scale bars = 100 μm. Black arrows indicate the locations of the cementoenamel junctions, and black triangles indicate the locations of the alveolar bone crest. B = bone, D = dentin, and P = pulp. Comparisons were performed using Dunnett’s multiple comparison test. Significant values are indicated as *P < 0.05, **P < 0.01, and ***P < 0.001, vs. the periodontitis group. Three repeated experiments were conducted. KGDI-109 reduced the amount of P. gingivalis in the oral cavityand had an effect on the intestinal flora less than that of AZM The bacterial count in the oral cavity was measured by qPCR after 7 days of KGDI-109 administration to confirm the effectiveness of the specific antibacterial activity of KGDI-109 against P. gingivalis in the periodontitis mouse model. The amount of total bacteria in the oral cavity showed no significant change between groups, whereas the amount of P. gingivalis was significantly reduced following KGDI-109 treatment (Fig. 4a). Fig. 4The effects of KGDI-109 on bacteria in the oral cavity and intestine. (a) The amount of bacteria in the oral cavity. The Ct values of the universal 16S rDNA and P. gingivalis genes from the ligature samples were analyzed using qPCR. The relative quantity of P. gingivalis was normalized to that of total bacterial gene copies, the ratio of periodontitis group as 1. Data are presented as means ± SD (control: n = 6, periodontitis: n = 7, KGDI-109-periodontitis: n = 6, and AZM-periodontitis: n = 5). Significant values are indicated as *P < 0.05 and **P < 0.01 (Dunnett’s multiple comparison test). n.s. = not significant. Three repeated experiments were conducted. (b) The following number of mice were used for the analysis of the intestinal microbiota (control: n = 8, periodontitis: n = 5, KGDI-109-periodontitis: n = 8, and AZM-periodontitis: n = 7). Relative abundance of intestinal bacterial taxa at the genus level in each experimental group. (c) Bar graphs and dot plots of alpha diversity based on observed features. Data are presented as means ± SD. Significant values are indicated as *P < 0.05, **P < 0.01, and ***P < 0.001 (Tukey’s multiple comparison test). (d) Principal coordinate analysis score plot of the intestinal microbiota profiles of the four groups using unweighted UniFrac distance. P = 0.002, for the AZM group vs. the control group (analysis of similarity test). The effects of KGDI-109 on bacteria in the oral cavity and intestine. (a) The amount of bacteria in the oral cavity. The Ct values of the universal 16S rDNA and P. gingivalis genes from the ligature samples were analyzed using qPCR. The relative quantity of P. gingivalis was normalized to that of total bacterial gene copies, the ratio of periodontitis group as 1. Data are presented as means ± SD (control: n = 6, periodontitis: n = 7, KGDI-109-periodontitis: n = 6, and AZM-periodontitis: n = 5). Significant values are indicated as *P < 0.05 and **P < 0.01 (Dunnett’s multiple comparison test). n.s. = not significant. Three repeated experiments were conducted. (b) The following number of mice were used for the analysis of the intestinal microbiota (control: n = 8, periodontitis: n = 5, KGDI-109-periodontitis: n = 8, and AZM-periodontitis: n = 7). Relative abundance of intestinal bacterial taxa at the genus level in each experimental group. (c) Bar graphs and dot plots of alpha diversity based on observed features. Data are presented as means ± SD. Significant values are indicated as *P < 0.05, **P < 0.01, and ***P < 0.001 (Tukey’s multiple comparison test). (d) Principal coordinate analysis score plot of the intestinal microbiota profiles of the four groups using unweighted UniFrac distance. P = 0.002, for the AZM group vs. the control group (analysis of similarity test). The effects of KGDI-109 on the intestinal microbiota were further investigated. Figure 4b presents the relative abundance of intestinal bacterial taxa at the genus level. The AZM group showed a different profile from the other groups. Figure 4c presents the result of alpha diversity analysis conducted to evaluate the microbial community diversity within a sample. Although the AZM group showed a significant decrease in alpha diversity compared with the periodontitis group, the KGDI-109 group showed higher alpha diversity than the periodontitis group (Fig. 4c). Furthermore, the AZM group differed greatly from the other groups in composition, whereas the KGDI-109 group had a composition like that of the control group, indicating maintained diversity of the intestinal flora (Fig. 4d). Discussion To the best of our knowledge, this is the first study to demonstrate the periodontal disease inhibitory effect of a DPP inhibitor that targets P. gingivalis. In this study, inhibitors were designed based on the cocrystal structures of valyl-tyrosine and phenylalanyl-tyrosine, which bind to the active centers of DPP7 derived from bacteria, and the dipeptide derivative KGDI-109 that inhibited PgDPP7 was obtained. The results showed that KGDI-109 inhibited amino acid metabolism by DPP7, thereby depleting the nutrient source necessary for P. gingivalis growth. Oral administration of KGDI-109 significantly suppressed alveolar bone resorption. The amount of P. gingivalis in the oral cavity was significantly reduced, indicating that the DPP inhibitor suppresses periodontal disease by its antibacterial activity against P. gingivalis. Figure 5 shows an overview of this study. Fig. 5An overview of this study. KGDI-109, a DPP inhibitor, showed an antibacterial effect targeting P. gingivalis via inhibition of amino acid metabolism. KGDI-109 is a new antibacterial agent that suppresses periodontitis. An overview of this study. KGDI-109, a DPP inhibitor, showed an antibacterial effect targeting P. gingivalis via inhibition of amino acid metabolism. KGDI-109 is a new antibacterial agent that suppresses periodontitis. P. gingivalis is an asaccharolytic bacterium that degrades proteins via gingipains and DPPs and exclusively utilizes amino acids as an energy source5,6. Arginine- and lysine-specific proteinases, RgpA, RgpB, and Kgp22–24, are present in the cell membrane or outer membrane vesicles and among the major virulence factors of P. gingivalis25–27. Extracellular proteins are initially degraded to oligopeptides via gingipains and transported into the periplasmic space28. DPPs are localized in the periplasm and degrade oligopeptides to dipeptides or tripeptides29,30. As this bacterium mainly takes up protein nutrients as dipeptides or tripeptides, DPPs are essential for its survival and proliferation. RgpA, RgpB, and Kgp demonstrate high specificity for the basic amino acid lysine or arginine; thus, KGDI-109, whose S1 subsite structure features hydrophobic side chains, is not expected to strongly bind to gingipains. Figure 1c suggests that KGDI-109 acts as a substrate-competitive inhibitor, particularly for substrates that specifically react with DPP7. DPP7 has been reported to prefer hydrophobic residues at the P18 and P216 positions. The suppression of alveolar bone resorption by KGDI-109, which likely does not inhibit gingipain activity, is primarily due to bacterial count reduction by the DPP inhibitor, thereby suppressing proliferation. In a periodontitis model in which mice were infected with the P. gingivalis W83 strain, metronidazole, which does not directly inhibit gingipain31, has been shown to suppress alveolar bone resorption32. This indicates that reducing the amount of bacteria through antimicrobial effects weakens their pathogenicity. Additionally, even when infecting SPF mice with the W50 strain, which demonstrates strong gingipain activity like W83, no alveolar bone resorption occured2. Therefore, the mechanism of periodontal disease induction does not necessarily originate from gingipain. In previous reports, the triple knockout mutant of DPPIV, DPP-7 and PTP-A indicated a 40% decrease in growth rate in minimal medium GA, whereas the growth of single- and double-knockout mutants was like that of W8333. The growth of the mutant strain NDP212, which lost dpp4, dpp5, dpp7, and dpp11, was significantly more retarded than the wild type in an anaerobic bacterial culture medium; however, this growth inhibition was limited29. The Met-Leu-MCA hydrolysis activity increased in the quadruple-deficient strain (NDP212) compared to the wild type. Such an excessive increase in activity suggests that it is caused by the induction of acylpeptidyl oligopeptidase (PGN_1349) and other unknown peptidases that complement and compensate for the four DPP deficiencies. Furthermore, the DPP11 activity is elevated in gingipain-deficient mutant cells, suggesting that it compensates for the loss of gingipain activity29. Elevated DPP4 and PTP-A activities have also been reported in porT mutants that do not express functional gingipain34. Therefore, the limited growth inhibition observed in the DPP knockout strain29,33 suggests that the absence of a specific DPP may be compensated for by other DPP types. Furthermore, gene-deficient strains may have established alternative metabolic pathways optimized for an environment that lacks DPP during repeated proliferation. Unlike the limited growth inhibition reported in DPP knockout strains, the pharmacological inhibition by KGDI-109 in this study resulted in a potent growth-inhibitory effect. This may be because the rapid action of the inhibitor may not allow sufficient time for such a compensation or adaptation to occur. Furthermore, the potent growth-inhibitory effect of KGDI-109 may be further enhanced by its ability to simultaneously block redundant pathways, as demonstrated by its dose-dependent inhibition of Met-Leu-MCA hydrolysis—a substrate recognized by both DPP5 and DPP7. The MIC may be lower than the IC50 for DPP7 based on the DPP dependence of peptide metabolism in P. gingivalis and the potential for KGDI-109 to simultaneously inhibit multiple DPPs. Furthermore, DPPs targeted by KGDI-109 are localized in the periplasmic space. KGDI-109 is highly lipophilic and readily permeates the outer membrane, whereas its inner membrane permeability is low. Consequently, the compound equilibrates across the outer membrane and is localized in the narrow periplasmic space, where the target enzymes may act as a reservoir35, the total local concentration likely exceeds the MIC, which is the external concentration. Further research is needed to elucidate the detailed mechanism of action of the inhibitor. KGDI-109 did not inhibit tumor necrosis factor-α production induced by stimulation with toll-like receptor agonists containing P. gingivalis lipopolysaccharide (Fig. S5), indicating that KGDI-109 does not have anti-inflammatory effects. Additionally, KGDI-109 administration reduced the number of TRAP-positive cells in the periodontal tissue of mice. However, it did not have a direct effect on osteoclast differentiation in vitro (Fig. S6a). These findings indicate that the inhibition of periodontal tissue destruction and decrease in TRAP-positive cells are the result of inflammatory reaction suppression by the antibacterial activity of KGDI-109. Furthermore, KGDI-109 did not affect osteoblast calcified nodule formation (Fig. S6b). These findings indicate that KGDI-109 has no anti-inflammatory or direct effects on bone metabolism and that the inhibition of alveolar bone resorption in the mouse model is due to its specific antibacterial activity against P. gingivalis. In this study, the bacterial load adhering to the ligature was measured by qPCR (Fig. 4a). Although AZM has a broad antimicrobial spectrum, the total bacterial did not change. This may be due to the persistence of a certain amount of biofilm formed by oral commensal bacteria around the ligature, even after azithromycin administration, and the potential inclusion of dead bacteria within the biofilm in qPCR results. Antibiotics such as ampicillin and vancomycin have been reported to cause dramatic changes in the abundance and diversity of the gut microbiota. However, they do not have a significant impact on the oral microbiota, indicating that the oral microbiota, characterized by frequent exposure to various environmental factors, is more resilient to antibiotic disruption36. Our findings indicate a reduction in P. gingivalis following KGDI-109 administration, suggesting that it alters the composition rather than the quantity of the oral biofilm. Future studies are needed to investigate changes in the composition and diversity of the oral microbiota to provide a more detailed understanding of the mechanism of action of KGDI-109. In this study, P. gingivalis infection altered the gut microbiota and reduced bacterial alpha diversity (Fig. 4), consistent with previous reports37. In Fig. 4b, the genus Porphyromonas was not dominant in the P. gingivalis-administered group, likely because P. gingivalis does not colonize the gut37,38. Even in the absence of intestinal colonization, P. gingivalis has been reported to induce ileal inflammation and IL9+ CD4+ lamina propria T cells, thereby impairing the gut epithelial barrier function by inducing dysbiosis of the gut microbiota37. The advantages of using a narrow-spectrum antibacterial agent are sparing antibiotic use and curtailing the emergence of resistance39. In addition, maintaining beneficial microbiota can prevent complications such as antibiotic-associated diarrhea39,40. KGDI-109 exhibited that the change in the intestinal flora was less than that of AZM and that the diversity of the resident bacterial flora was maintained via selective inhibition of P. gingivalis (Fig. 4). KGDI-109 may be more effective than AZM in maintaining the diversity of lactic acid bacteria, which is essential for maintaining health and reducing adverse health effects. Moreover, clinical evaluation of P. gingivalis-specific KGDI-109 will elucidate the etiology of periodontal disease. KGDI-109 may also be beneficial in preventing systemic diseases associated with P. gingivalis, such as diabetes and rheumatoid arthritis41,42. This study has several limitations. First, the number of bacterial species examined in this study was limited. KGDI-109 may exhibit antibacterial activity against other bacteria if they have DPPs like those of P. gingivalis. Second, further investigations into the biological stability of KGDI-109 and its clinical administration methods are needed. Heat treatment at 100 °C for 10 min did not affect the antibacterial activity of KGDI-109 (Table S3). In addition to the effects of acids and enzymes, considering their behavior in the body, such as their half-life, is necessary. Third, the model used in this experiment has some limitations. Bacterial flora changes and inflammatory bone resorption occur even with ligation alone43, therefore, investigating the effects of DPP inhibitors in settings other than combined models is necessary. Furthermore, verifying their effects on the human bacterial flora and in periodontitis remains a challenge. Further studies are needed to determine the efficacy and side effects in humans. Conclusion KGDI-109, a DPP inhibitor, showed an inhibitory effect against P. gingivalis growth and suppressed alveolar bone resorption. KGDI-109 has the potential to be a novel treatment for periodontal diseases, targeting P. gingivalis and maintaining gut microbial diversity. Supplementary Material Below is the link to the electronic supplementary material. 41598_2026_52648_MOESM1_ESM 41598_2026_52648_MOESM1_ESM