Local periodontal injection, systemic heart repair: A dual-functional hydrogel for non-invasive concurrent treatment of periodontitis and myocardial infarction

Abstract

Periodontitis and myocardial infarction (MI), the leading cause of mortality worldwide, represent globally prevalent inflammatory diseases with bidirectional pathophysiological links. Despite the urgent demand for non-invasive strategies capable of alleviating local periodontal destruction while mitigating associated systemic cardiovascular complications, no integrated treatment modality currently exists. To address this challenge, a protein-loaded antibacterial hydrogel, thiolated chitosan/AMP-PEG-maleimide (C1.5P4/BMP-2), with novel sustained protein release properties was developed. This hydrogel features an interconnected microporous architecture that: (1) enables high-efficiency BMP-2 protein encapsulation, (2) preserves protein bioactivity while ensuring sustained release, thereby addressing the recognized challenge of hydrogel-based protein delivery, (3) confers potent antibacterial properties, (4) facilitates remote cardiac function improvement by resolving periodontal inflammation. In murine models, locally, it attenuated alveolar bone loss (2-fold greater bone regeneration vs controls) while systemically improving post-MI cardiac function (75.6% higher ejection fraction). Mechanistically, the hydrogel's protein-protective microenvironment synergized with its antimicrobial action selectively inhibiting Gram-negative (G⁻) anaerobic pathogens (primary periodontal culprits) while enriching Gram-positive (G⁺) commensals, regulating biofilm G⁻/G⁺ ratio. Concomitantly, this dual action modulates the oral-cardiac inflammatory axis, specifically downregulating B2 cell/TNF-α signaling to mitigate systemic inflammation associated with MI. Collectively, this study presents a novel non-invasive protein-stabilizing hydrogel that addresses periodontitis-MI comorbidity through sustained osteogenic factor delivery coupled with microbiome-immune modulation.

Graphical abstract

Highlights

• •The first dual-targeting hydrogel for periodontitis-MI comorbidity via a single periodontal injection.
• •High protein loading capacity with sustained bioactivity and tunable injectability for periodontal pockets.
• •Gram-negative selective antimicrobial and osteogenic synergy of the hydrogel for periodontitis.
• •Dual-organ healing ability of the hydrogel via mediating microbiome-immune crosstalk.

1. Introduction

Periodontitis, a globally prevalent infectious disease affecting over 30% of adults, originates primarily from dysbiotic subgingival pathogens and dysregulated host proinflammatory responses [1,2]. The complex interplay between periodontal microbes and host immune system leads to the accumulation of pro-inflammatory mediators and progressive deterioration of the supporting apparatus in periodontium. Clinical manifestations include deep periodontal pocket, bleeding gums, gingival recession, alveolar bone resorption, and potentially tooth loss [3]. The unique anatomical challenges posed by periodontal pocket morphology and alveolar bone defect heterogeneity significantly complicate therapeutic interventions and regenerative outcomes [4,5]. Current treatment paradigms mainly involve conventional mechanical scaling and root planning, coupled with the administration of antimicrobial drugs such as metronidazole into the periodontal pocket [6]. However, the escalating issue of antibiotic resistance among oral pathogens, exacerbated by widespread antimicrobial use in periodontal therapy, underscores the critical need for developing alternative non-antibiotic bioactive strategies [5,6].

As an infectious disease, periodontitis not only impairs periodontium, but also exhibits bidirectionally associated with metabolic syndrome and overall systemic health [7]. Given the high prevalence of shared risk factors, such as sedentary lifestyles, smoking, and obesity, in adult populations, increasing attention has been focused on elucidating the link between periodontitis and cardiovascular diseases (CVD) [8], which encompasses a spectrum of pathological cardiac conditions including atherosclerosis and MI. Atherosclerosis arises from local ischemia due to vascular fibrosis or lipid occlusion, whereas the limited regenerative capacity of cardiomyocytes contributes to acute MI and stroke [[9], [10], [11], [12]].

MI, as a pivotal manifestation of CVD, has gained extensive attention due to its high incidence and mortality rate [13]. Current therapies for MI include pharmacotherapy, percutaneous coronary intervention, ventricular assist devices, and heart transplantation [14]. However, these strategies are constrained by donor organs scarcity, thrombosis, vascular restenosis, and immune rejection [15]. Consequently, there is an urgent need to develop novel, noninvasive, and biocompatible therapeutic alternatives. One promising direction is to explore therapeutic insights from comorbid conditions. Notably, periodontitis and MI are generally linked to each other via shared mechanisms including immune regulation, circulating signaling molecules, and bacteremia primarily caused by oral pathogens including Fusobacterium nucleatum (Fn), Porphyromonas gingivalis (Pg), Aggregatibacter actinomycetemcomitans (Aa), Micromonas, and Prevotella intermedia (Pi) which can translocate from subgingival biofilms into systemic circulation, exacerbating cardiovascular pathology [[16], [17], [18], [19]]. Recent studies indicate that approximately 42.5% of MI patients had mild to severe periodontitis [16,20], further underscoring the clinical relevance of this association. Given these findings, identification of a non-invasive bioengineering therapeutic strategy targeting common shared pathogenic pathways in both periodontitis and MI could offer a dual-therapeutic solution, addressing the clinical challenges posed by these interconnected conditions.

Hydrogels are a group of polymer materials with three-dimensional (3D) structures. Owing to their biomimetic structure resembling the native extracellular matrix, high hydration capacity, tunable degradation and injectability, they have been widely used for biomedicine and biotechnology [21]. Additionally, their controllable gelation, fluidity, and sustained-release properties [22,23] further support their use as topical drug delivery systems that reduce dosing frequency [24]. In the realm of periodontal therapy, conventional hydrogels mainly supply antibiotics or simple antimicrobials [24,25], a strategy that risks accelerating resistance and overlooks the multifactorial nature of the disease [4,26]. Therapeutic proteins (e.g., bone morphogenetic protein 2, BMP-2) and peptides present a promising alternative by enabling multi-targeted regeneration and immunomodulation at nanogram doses, a feat unachievable with small molecules. Nevertheless, the effective loading and delivery of proteins via hydrogels remain challenging, as they require mild processing conditions, preserved bioactivity, and controlled release kinetics. Addressing these challenges represents a critical step toward the development of next-generation hydrogels for synergistic treatment of periodontitis and MI.

In this study, we engineered a microporous, protein-protective hydrogel by crosslinking thiolated chitosan with antimicrobial peptide (AMP)-conjugated 8-arm PEG-maleimide (PEG-Mal-AMP) and unmodified PEG-maleimide (PEG-Mal), creating a dual-functional system capable of sustained BMP-2 release while maintaining bioactivity. Unlike conventional carriers, this hydrogel's dynamic thiol-maleimide network ensures high protein-loading efficiency and prolonged osteogenic signaling, synergizing with its inherent antibacterial properties (Tet213-AMP) to address both local periodontal regeneration and systemic MI complications via a single periodontal injection. We comprehensively characterized the hydrogel's mechanochemical properties including shear-thinning behavior, tunable stiffness, and controlled degradation to optimize its in vivo performance. In vitro, the system demonstrated dual functionality: osteogenesis (>4-fold higher ALP activity vs controls), and gram negative pathogen-selective antimicrobial action (>4-log reduction in all tested periodontal pathogens). Using a ligature-induced periodontitis model and a novel periodontitis-MI comorbidity model, we validated its therapeutic efficacy, revealing mechanistic insights into its dual-tissue targeting via microbiome-immune modulation (B2 cell/TNF-α axis downregulation). This work establishes a novel protein-stabilizing hydrogel that concurrently resolves oral and cardiovascular inflammation, offering a translatable dual-targeted therapeutic strategy for periodontitis-associated systemic diseases (Scheme 1).

Scheme 1.

Fabrication and application of a dual-function hydrogel platform for periodontitis and comorbid MI.

2. Results and discussion

2.1. Formation of hydrogels

CS was thiolated using Traut's reagent to produce CS-SH, with successful modification confirmed by ¹H NMR (Fig. S1A). Quantitative analysis via Ellman's reagent (5,5′-dithiobis-(2-nitrobenzoic acid) revealed a thiol group content of 470 μmol/g on the modified CS. An AMP with a N-terminal cysteine was then conjugated to maleimide-functionalized PEG-Mal via thiol-maleimide “click” chemistry, yielding PEG-Mal-AMP. The synthesis efficiency was verified by ¹H NMR, which revealed characteristic AMP peak corresponding to a grafting efficiency of 98.8% (Figs. S1B and S5). Fourier-transform infrared (FTIR) analysis further confirmed the conjugation, as evidenced by the disappearance of the maleimide carbon-carbon double bond peak (Fig. 1A).

Fig. 1.

Synthesis and characterization of the C1.5P4 hydrogel. (A) FT-IR spectra of PEG-Mal-AMP and PEG-maleimide. (B) Compositions of C1P4, C1.5P4, and C2P4 hydrogels. (C-D) Rheological properties of the hydrogels assessed by dynamic time scan and frequency sweep. (E) Viscosity measurements of the hydrogels. (F) Degradation of the hydrogels over a 30-day period. (G) Swelling behavior of the hydrogels after 24 h of incubation in PBS. (H) Cumulative release of BMP-2 protein from the hydrogels over 12 days. (I) SEM images of the surface morphology of the hydrogels. (J) Pore size analysis of the hydrogels. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

The CS-PEG-Mal-AMP hydrogel was formed through a thiol-Michael addition reaction between the thiol groups on CS-SH and the maleimide groups on PEG-Mal. To optimize the gelation conditions, precursor solutions with varying concentrations of PEG-Mal-AMP, PEG-Mal and CS-SH were prepared (Fig. 1B). Increasing precursor concentrations enhanced crosslinking densities and significantly shortened gelation times. Dynamic time sweep analysis revealed that the G′ of the C1P4, C1.5P4 and C2P4 formulations surpassed the G″ at 82, 58 and 34 s respectively, indicating the onset of gelation (Fig. 1C). This was further supported by inverted vial tests, which confirmed successful hydrogel formation (Fig. S1C). Rheological tests by frequency sweep were also conducted to evaluate its mechanical behavior. The frequency sweep analysis showed that within the range of 0.1 to 10 Hz, the G′ of C1.5P4 was about ten times higher than the G″ (Fig. 1D), indicating a robust and stable network structure. Notably, both C1.5P4 and C2P4 exhibited higher G′ values (∼600 Pa) compared to C1P4, possibly due to increased availability of reactive thiol and maleimide groups facilitating more extensive crosslinking. Additionally, the hydrogel exhibited shear-thinning behavior, as evidenced by a decrease in viscosity with increasing shear rate, confirming its potential for injectability (Fig. 1E).

Although mature BMP-2 lacks free thiol groups and is not expected to participate in the hydrogel's thiol-maleimide crosslinking reaction, we characterized the modulus and viscosity of C1.5P4/BMP-2 to confirm whether it induces any unintended alterations to the hydrogel's properties. As shown in Figs. S10 and S11, C1.5P4/BMP-2 exhibited slight differences in modulus and viscosity to blank C1.5P4, which may be attributed to van der Waals Forces or hydrogen bonding between BMP-2 and the hydrogel.

The in vitro degradation behavior of the hydrogels was evaluated, demonstrating that degradation proceeded predominantly via surface erosion coupled with bulk degradation mechanisms. Over a 30-day incubation period, both C1.5P4 and C2P4 exhibited time-dependent degradation with residual weights of 10.6% and 21.4% at day 22, respectively. In contrast, C1P4, featuring a less densely crosslinked network, underwent 94.7% degradation by day 22 (Fig. 1F). The release kinetics of PEG-Mal-AMP were also evaluated, the results demonstrated that C1.5P4 exhibited release profiles highly similar to C2P4, confirming its controlled release potential (Fig. S9).

Hydrogel swelling behavior was assessed following 24-h incubation in PBS at 37 °C. All formulations exhibited substantial swelling, with the equilibrium swelling ratio (ESR) inversely proportional to crosslinking density. C1P4 demonstrated the highest ESR, indicative of a looser network structure, while C1.5P4 and C2P4 showed significantly lower ESRs, consistent with the rheological measurements.

BMP-2, a potent osteogenic factor essential for repairing irreversible alveolar bone resorption induced by periodontitis, was physically encapsulated within the hydrogels at 2.0 μg/mL, a concentration validated and widely adopted in well-established preclinical studies of periodontal tissue regeneration [[27], [28], [29], [30]] and its in vitro release behavior was examined (Fig. 1H). This dosage was selected to balance effective osteogenic induction for periodontal bone repair with the minimization of dose-dependent safety risks, such as ectopic bone formation, in local application scenarios. BMP-2 release occurred primarily through diffusion and hydrogel scaffold degradation. C1P4 hydrogel showed significant BMP-2 burst release, with over 62% released by day 1 and nearly 80% released by day 4. In contrast, C1.5P4 and C2P4, with higher crosslinking density, demonstrated more controlled release kinetics. Notably, C1.5P4 sustained BMP-2 release for over 12 days, achieving a cumulative release exceeding 80%, while C2P4 had a limited release less than 50%.

The morphology of lyophilized hydrogels was characterized by SEM. All the hydrogels exhibited a typical interconnected porous microstructure with uniform pore distribution (Fig. 1I and J), which facilitates efficient oxygen and nutrient transport. Quantitative analysis showed that the average pore sizes of C1P4, C1.5P4, and C2P4 were 92.9 μm, 84.7 μm, and 56.2 μm, respectively (Fig. 1I and J), indicating a clear inverse relationship between pore size and crosslinking density that C1.5P4 and C2P4 formed denser internal architectures compared to the more open structure of C1P4.

2.2. In vitro cytocompatibility and osteogenesis promotion of the hydrogels

Given the critical requirement of biocompatibility for biomedical hydrogels, cytotoxicity was systematically evaluated using live/dead staining and CCK-8 assays. As shown in Fig. 2A, the number of cells encapsulated within the three hydrogels increased significantly from day 1 to days 3 and 5, confirming their capacity to support cell growth and proliferation. Quantitative analysis revealed only minimal number of dead cells, underscoring the excellent cytocompatibility of these hydrogels. None of the hydrogel extracts exhibited cytotoxicity, and cells maintained full viability under all conditions after 24-h incubation (Fig. 2B).

Fig. 2.

In vitro bone-forming and antibacterial abilities of hydrogels. (A) Effects of the hydrogels on hPDLFs viability and proliferation measured by live/dead staining. (B) Cytotoxicity of the hydrogels on hPDLFs measured by CCK-8 assay. (C-E) Effects of BMP-2--encapsulated hydrogels on the expression of ALP, OCN, and COL-1 in hPDLFs analyzed by qRT-PCR. (F-G) Osteogenic effects of BMP-2-encapsulated hydrogels on hPDLFs measured by ALP staining. (H-I) Calcium deposition effects of BMP-2-encapsulated hydrogels on hPDLFs measured by Alizarin red staining. (J-L) BacTiter-Glo microbial cell viability assay of Fn, Aa, and Pg. RLU, relative luminescence unit. (M) Representative fluorescence images of live Fn, Aa, and Pg after 4 h of different treatment. Scale bar: 100 μm. (N) Representative images of live Fn, Aa, and Pg grown on sheep blood agar plates. (O-Q) Quantifications of live Fn, Aa, and Pg based on the results exemplified in (N). CFU, colony-forming units. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

The osteogenic potential of BMP-2-loaded hydrogels was further assessed by quantifying osteogenesis-related gene expression via qRT-PCR. As shown in Fig. 2C, D, and 2E, following 7 days of co-culture with human periodontal ligament fibroblasts (hPDLFs), the C1P4/BMP-2, C1.5P4/BMP-2, and C2P4/BMP-2 groups all demonstrated significantly upregulated expression of ALP, OCN, and COL-1 compared to the control group. These molecular findings were corroborated by functional assay by ALP staining and Alizarin Red S staining as well. As shown in Fig. 2F and G, ALP staining intensity after 7 days of co-culture was markedly higher in the C1P4/BMP-2 and C1.5P4/BMP-2 groups than in the control and C2P4/BMP-2 groups. Similarly, Alizarin Red S staining after 14 days revealed significantly greater calcium deposition in the C1P4/BMP-2 and C1.5P4/BMP-2 groups compared to control and the C2P4/BMP-2 group (Fig. 2H and I). Based on the integrated analysis of in vitro release results and osteogenic induction capacity of the three hydrogels, C1.5P4 was identified as the optimal formulation, exhibiting a balanced profile of sustained BMP-2 release and robust osteogenesis promotion, and was thus selected for subsequent in vivo evaluations.

Among the three formulations, C1P4 and C1.5P4 exhibited the closest physicochemical profiles. Nevertheless, a modest 0.5% (w/v) increase in CS-SH concentration substantially enhanced the performance of C1.5P4, as reflected by a 3.9-fold rise in storage modulus, an approximate 22% reduction in the initial BMP-2 burst release at day 1, a slower degradation rate, an approximately 8% decrease in pore size, and a significant upregulation of osteogenic and mineralization markers (Fig. 1, Fig. 2). These characteristics align well with the periodontal pocket microenvironment, which demands sustained drug release and robust osteogenic potential, making C1.5P4 the most suitable formulation for periodontal applications.

2.3. Antibacterial ability of the hydrogel in vitro

Periodontal disease is a bacteria-induced inflammatory disorder, with pathogenic plaque biofilms dominated by subgingival bacteria Fn, Aa, and Pg serving as the primary etiological agents. Therefore, targeting bacterial colonization is a key strategy for mitigating periodontal inflammation. To evaluate the antibacterial ability of the hydrogel, overnight cultures of periodontal pathogens Fn, Aa, and Pg, each at a concentration of 10⁶ CFU/mL, were resuspended in fresh culture medium, exposed to the hydrogels, and incubated anaerobically at 37 °C in 96-well plates. As the hydrogel incorporates antimicrobial AMP and CS-both with established antibacterial properties, bacterial viability decreased to less than 1% after 4 h of incubation (Fig. 2J–L). The antimicrobial ability was further demonstrated by confocal imaging (Fig. 2M), blood agar plating and CFU enumeration (Fig. 2N–Q), all of which exhibited consistent antimicrobial trends, confirming the hydrogel's potent antibacterial effects against periodontal pathogens.

2.4. Anti-inflammatory and osteogenic effects of C1.5P4/BMP-2 hydrogel on periodontitis in vivo

Given the unique structure of the C1.5P4/BMP-2 hydrogel, we evaluated its capacity to mitigate inflammation and promote osteogenesis in a ligature-induced periodontitis (LIP) mouse model, as the accumulation of pathogenic plaque biofilms and pro-inflammatory immune responses are key contributors to alveolar bone loss. Periodontitis was induced in eight-week-old mice, followed by weekly injections of the hydrogel into the gingival sulcus until sacrifice at week 3 (Fig. 3A). Quantitative micro-CT revealed significantly exacerbated alveolar bone loss in LIP-treated mice compared to healthy controls (Fig. 3B and C), whereas C1.5P4/BMP-2 hydrogel treatment markedly attenuated bone loss, as evidenced by increased bone volume to total volume (BV/TV) and trabecular thickness (Tb.Th), along with reduced bone surface-to-volume ratio (BS/BV) (Fig. 3D–F).

Fig. 3.

C1.5P4/BMP-2 hydrogel alleviates periodontitis in mice. (A) Schematic illustration of the experimental design. PD: periodontal disease. (B) Quantification of alveolar bone loss in the periodontium after different treatment. (C) Representative Micro-CT images of alveolar bone. (D-F) Quantification of BV/TV, BS/BV, and Tb.Th based on Micro-CT analysis. (G) H&E and Masson's trichrome staining of periodontal tissues. D, dentin; A, alveolar bone. N.S., not significant. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. N = 6 per group for all experimental groups.

These protective effects were further verified by histological staining. As illustrated in Fig. 3G, a deep periodontal pocket was observed in the LIP group, whereas treatment with the C1.5P4/BMP-2 hydrogel significantly reduced this pro-inflammatory response, as evidenced by the presence of a shallow periodontal pocket in the LIP + C1.5P4/BMP-2 group. These findings collectively confirm the hydrogel's ability to simultaneously suppress inflammation and restore alveolar bone integrity in periodontitis.

2.5. Anti-inflammatory effects of C1.5P4/BMP-2 hydrogel on mice with both periodontitis and MI in vivo

To investigate the hydrogel's systemic effects, MI was induced at week 3 via left anterior descending artery ligation. Successful induction of MI was confirmed by characteristic echocardiographic findings, notably a significant decrease in the EF to below 50%. The detailed treatment strategy is illustrated in Fig. 4A. Among all groups, mice in the LIP-MI group showed the most pronounced changes in echocardiographic images, and reduction in echocardiographic parameters including EF, FS, and increased LVDd, LVIDs. In contrast, mice treated with the C1.5P4 and C1.5P4/BMP-2 hydrogel, particularly the latter, exhibited significant improved cardiac condition compared to the LIP-MI and MI-alone groups. This was evidenced by significantly higher EF (75.6%), FS (96.9%), and lower LVDd, and LVIDs (Fig. 4B–F) compared to LIP-MI group. Correspondingly, Sirius red staining confirmed the largest infarct size in the LIP-MI group, whereas hydrogel treatment significantly reduced infarction and improved cardiac performance (Fig. 4G and H). Changes in alveolar bone loss, heart weight and body weight were also assessed (Fig. 4I). The C1.5P4/BMP-2 hydrogel-treated group showed significant attenuation of alveolar bone loss (Fig. S8) and a reduced HW/BW ratio, indicating mitigated periodontitis progression and alleviated myocardial infarction. Notably, the hydrogel also exerted a preventive effect against MI as preventing periodontitis progression mitigated subsequent MI severity (Fig. S6), whereas no functional impact was observed in MI-only mice receiving hydrogel spreading on healthy periodontium (Fig. S2), confirming that therapeutic benefits require periodontal pocket penetration. Finally, in vitro pro-inflammatory effects and in vivo cytotoxicity assessment were performed to demonstrate the biocompatibility of the hydrogel (Fig. 4J–S7, S12 and S13). These analyses did not detect notable BMP-2-associated cytotoxicity, tissue damage, or aberrant inflammatory responses in the examined tissues. Additionally, no ectopic bone formation was found in the cardiac tissues of any experimental group over the course of the study, which supports the safety profile of this BMP-2 delivery strategy for periodontal applications and helps address key translational considerations surrounding BMP-2 use. Collectively, the hydrogel's potent anti-inflammatory effects and significant therapeutic benefits in treating comorbid periodontitis and MI highlight its mechanistic contributions from local periodontal repair, which further modulates systemic inflammation to exert cardiac protective effects, positioning it as a promising therapeutic strategy for managing inflammatory comorbidities with systemic manifestations.

Fig. 4.

C1.5P4/BMP-2 hydrogel attenuates LIP-exacerbated MI in mice. (A) Schematic overview of the experimental workflow. (B) Representative echocardiographic images obtained immediately before sacrifice. (C-F) Quantification of cardiac function parameters, including EF, FS, LVDd, and LVIDs, based on echocardiographic measurements. (G) Representative sirius red staining images of cross-sectional mouse heart samples post-sacrifice. Red-stained tissues indicate fibrosis. (H) Quantification of infarct size in mouse hearts based on the results exemplified in (G). (I) Heart weight-to-body weight ratio measured post-sacrifice. (J) H&E staining of liver, spleen, lung, and kidney tissues post-sacrifice. Scale bar: 100 μm. N = 12 (Control), 11 (MI), 11 (LIP-MI), 8 (LIP-MI + C1.5), 8 (LIP-MI + C1.5P4), 13 (LIP-MI + C1.5P4/BMP-2) for (B-J). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

2.6. Antibacterial ability of the hydrogel in vivo

Given that periodontal disease is initiated by subgingival bacterial biofilms, we conducted comprehensive 16S rRNA sequencing to evaluate the hydrogel's antimicrobial effects in both periodontal and systemic tissues. Full-length 16S rRNA sequencing was performed on periodontal ligature and heart tissue samples following hydrogel therapy. The specific composition tables (top 10) of microbial communities were obtained by statistically analyzing the leveled Amplicon Sequence Variant (ASV)/Operational Taxonomic Unit (OTU) tables (Fig. 5A–B and H-I). In the periodontium, Simpson's index, an index for species diversity [31], revealed a significant difference, indicating that the hydrogel application altered the local microbial community structure. Although heart tissue showed a decreasing trend in Observed species (p = 0.055), an index for species richness, this non-significant result suggests the hydrogel's primary antimicrobial effects are localized to directly contacted periodontal tissues (Fig. 5C and J). Rarefaction curves, which standardize sequencing depth for ASV/OTU richness comparison, displayed markedly lower diversity indices in hydrogel-treated groups, particularly in heart tissue, further supporting the hydrogel's antibacterial properties (Fig. 5D and K).

Fig. 5.

Microbial alterations in mice with both periodontitis and MI following C1.5P4/BMP-2 hydrogel therapy. (A-B) Genus-level (A) and species-level (B) microbial community composition in periodontal ligatures. (C) α-diversity metrics (Simpson index, Shannon index, Chao1 richness, and Observed species count) in periodontal ligatures. (D) Rarefaction curves illustrating the relationship between sequencing depth and species richness in periodontal ligatures. (E) PERMANOVA of differentially abundant microbial taxa in periodontal ligatures. (F) Principal Coordinates Analysis (PCoA) of microbial communities in periodontal ligatures based on Bray-Curtis dissimilarity. (G) OPLS-DA highlighting distinct clustering patterns of microbial communities in periodontal ligatures. (H-I) Genus-level (H) and species-level (I) microbial composition in hearts. (J) α-diversity metrics (Simpson index, Shannon index, Chao1 richness, and Observed species count) in hearts. (K) Rarefaction curves correlating sequencing depth with species richness in hearts. (L) PERMANOVA of microbial composition differences between treated and control groups in hearts. (M) PCoA analysis of microbial community structure in hearts using Bray-Curtis dissimilarity. (N) OPLS-DA demonstrating clustering patterns of microbial communities in hearts. N = 6 for all panels.

The β-diversity of bacterial communities was evaluated subsequently. Permutational Multivariate Analysis of Variance (PERMANOVA), a non-parametric method based on permutation tests [32], demonstrated significant hydrogel-induced restructuring of periodontal microbiota (pseudo-F = 1.923274, permutations = 999, p = 0.02), whereas no statistical changes were observed in heart tissue, which possibly attribute to the indirect hydrogel exposure (Fig. 5E and L). Complementary multivariate analyses-including Principal Coordinates Analysis (PCoA) and Orthogonal Partial Least Squares Discriminant Analysis (OPLS-DA) reinforced the PERMANOVA findings, revealing distinct treatment-dependent clustering in periodontal microbiota but only a mild separation trend in heart microbiota (Fig. 5F–G, M − N). These integrated results demonstrate the hydrogel's potent and localized antibacterial effects, with significant modulation of periodontal microbiota composition.

2.7. Mechanistic study in vivo

2.7.1. Microbial community regulation-probiotics enrichment and G⁻/G⁺ ratio downregulation

To explore the hydrogel's therapeutic mechanisms, as well as the core oral-cardiac microbial and immune links underlying these two comorbid conditions, we analyzed microbial taxa in periodontium and heart tissue groups. A total of 298 species were shared across all tissue samples (Fig. S3A). To identify therapeutic microbial markers responsible for the changes, two well-established analytical methods, random forest and linear discriminant analysis effect size (LEfSe) [33] were employed (Fig. 6A–E). In periodontium tissue, random forest analysis identified Pediococcus and Enterococcus (Fig. 6A and B) as the 2 most significantly upregulated genera following hydrogel therapy, with Pediococcus pentosaceus and Enterococcus faecalis identified as the top two upregulated species (Fig. S4). This finding was consistent with LEfSe that Enterococcaceae (family), Enterococcus (genus), Pediococcus (genus), Pediococcus pentosaceus (species) and Enterococcus faecalis (species) were upregulated in hydrogel therapy groups (Fig. 6E). Notably, Pediococcus pentosaceus is widely recognized as a probiotic in numerous studies due to its antibacterial and antifungal activities, contributing to the maintenance of oral microbiome homeostasis [[34], [35], [36], [37]]. Similarly, specific strains of Enterococcus faecalis have been used as probiotics to help maintain intestinal health and support the immune system, with selected isolates showing safety for human consumption and metabolic benefits [38,39]. Therefore, these microbes, recognized for their probiotic properties, might help maintain oral microbiome balance and overall health.

Fig. 6.

Impacts of C1.5P4/BMP-2 hydrogel on microbial interaction in mice with both periodontitis and MI. (A-B) Random Forest analysis of microbial communities in periodontal ligatures at the genus and species levels. (C-D) Random Forest analysis of microbial communities in hearts at the genus and species levels. (E) LEfSe-derived taxonomic cladogram illustrating differentially abundant microbial taxa in periodontal ligatures and hearts, with interconnecting lines highlighting taxa shared between compartments. (F) Heatmap of Spearman's correlation coefficients between differentially abundant genera in periodontal ligatures and hearts. “+”: bacteria significantly changed in mice with C1.5P4/BMP-2 hydrogel therapy; “-”: bacteria significantly changed in mice with no C1.5P4/BMP-2 hydrogel treatment. G⁻: Gram negative species; G⁺: Gram positive species. ∗p < 0.05, ∗∗p < 0.01. (G-H) Differentially abundant microbial taxa in periodontal ligatures at genus level. (I-J) Differentially abundant microbial taxa in hearts at genus level. N = 6 for all panels. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

In cardiac tissue, Corynebacterium (G⁺) and Absiella (G⁺) were the 2 most significantly upregulated genera following hydrogel injection (Fig. 6C and D), which was also consistent with LEfSe results that Corynebacteriaceae (family), Corynebacterium (genus), Erysipelotrichales (order), Erysipelotrichaceae (family), and Absiella (genus) were upregulated in infarcted site after hydrogel therapy. LEfSe further identified the probiotic taxa Lactobacillales (order), Enterococcaceae (family), and Enterococcus (genus) as the top shared ones upregulated in all hydrogel therapy groups, suggesting potential oral-cardiac microbial mediating potential. Conversely, LEfSe analysis demonstrated significant downregulation of pathogenic genera Prevotella (genus) and Prevotellaceae (species), Bacteroidota (phylum) and Bacteroidales (order) in hydrogel therapy group (Fig. 6E). Notably, the well characterized human periodontal pathogens Prevotella intermedia and Pg (formerly Bacteroides gingivalis) are classified within these downregulated taxa, highlighting the hydrogel's capacity to target key human periodontal pathogens.

Spearman correlation coefficients heatmaps were generated for shared genera between periodontitis and MI at genus (Fig. 6F, with p < 0.05, and absolute value of the rho (R) value > 0.6), family, and order levels (Fig. S3B and C) to compare relative abundances of taxa conserved across both conditions. Corresponding species fold changes were shown in Fig. 6G–J. Periodontal Pediococcus was associated with cardiac Absiella. Similarly, Enterococcus, which was the keystone connector in periodontal biofilm, was correlated with cardiac Corynebacterium and Absiella. Notably, all enriched microbes in hydrogel therapy group in periodontium (Pediococcus and Enterococcus at genus levels), and 4/5 enriched microbes after hydrogel therapy in the heart tissue (Enterococcus, Corynebacterium, Absiella and Mogibacterium) were G⁺, while most of the taxonomy identified decreased after hydrogel therapy (Pseudomonas, Prevotella, Zoogloea, Proteus, Chryseobacterium, Ammoniphilus, Desulfomicrobium, Hyalangium, Chitinophaga, Moraxella, and Lysobacter in periodontium (11/13), and Syntrophorhabdus and Variovorax in heart tissue) were G⁻, indicating a reduction of G⁻/G⁺ ratio in the microbial community composition, while G⁻ are generally more virulent to host. This also aligned with the earlier results that the hydrogel inhibited the growth of key G⁻ periodontal pathogens (Fig. 2M–Q), underscoring its potential to control periodontal pathogens, mitigate microbial dissemination to the systemic circulation, thus relieving periodontitis-associated cardiac dysfunction.

However, while these microbiome shifts correlate with cardiac function improvement, direct causal evidence remains elusive; and cardiac benefits may also derive from direct antimicrobial or BMP-2-mediated effects independent of microbiome changes. Future mechanistic studies, such as microbiota transplantation, will be essential to validate these relationships.

2.7.2. Adaptive immunity regulation-B2 cell/TNF-α signaling downregulation

Apart from microbial dysbiosis, immune dysregulation also critically contributes to the pathogenesis of both periodontitis and MI, with immune cells playing central roles in both diseases’ progression [[40], [41], [42]]. We next determined the immune cells that are activated and the mechanism through which the hydrogel may cure periodontitis and associated MI. To achieve this, mouse submandibular lymph nodes and cardiac tissues were collected, and single cell suspensions were prepared and subjected to flow cytometry to identify activated populations. Innate and adaptive immune subsets, including lymphocytes (T and B cell), macrophage, and myeloid cells (including neutrophils and monocyte populations) were evaluated (Fig. 7). Notably, only the number of CD45⁺CD3^-B220⁺ B cells in both submandibular lymph nodes and heart tissues was significantly decreased after hydrogel treatment (Fig. 7D). The number of CD45⁺CD11b⁺ myeloid cells demonstrated a decreasing trend only in submandibular lymph nodes (Fig. 7B). No significant changes in CD45⁺CD11b⁺CD64⁺ macrophages were detected in either tissue (Fig. 7C).

Fig. 7.

Impacts of C1.5P4/BMP-2 hydrogel on cardiac and periodontal immune cell subsets in LIP-MI mice. Representative flow cytometry plots of CD45⁺CD3⁺ T lymphocytes, CD45⁺CD11b⁺ myeloid cells, CD45⁺CD11b⁺CD64⁺ macrophages, and CD45⁺CD3^-B220⁺ B cells (A-D), and quantification of the percentage of the cells (E-H) in periodontal tissues and hearts with or without C1.5P4/BMP-2 hydrogel treatment. N = 6 for for all panels. ∗p < 0.05, ∗∗p < 0.01 as compared to LIP-MI group. N.S., not significant.

The host immune system includes two major B cell subtypes with distinct inflammatory roles. To identify the subtype responsible for these effects, we further analyzed B cell populations. As shown in Fig. 8A–B and E-F, hydrogel treatment led to a significant reduction in the number of CD45⁺CD3^-B220⁺CD5⁻CD23⁺ B2 cells in both mouse submandibular lymph nodes and heart tissues. Given that B2 cells circulate between the bloodstream and lymphoid tissues, these results indicated that the hydrogel either suppresses systemic B2 cells to mitigate inflammation or suppresses B2 cells in submandibular lymph nodes to reduce the number of B2 cells circulating to cardiac tissue, leading to a final reduction in total B2 cells in mice. This finding aligns with recent evidence that B2 cells can markedly exacerbate MI via the oral-heart axis, thereby underscoring their critical role in this axis [19].

Fig. 8.

Alterations of B-cell subset and pro-inflammatory cytokine expression in LIP-MI mice following C1.5P4/BMP-2 hydrogel therapy. (A-B) Representative flow cytometry plots of CD5^-B220⁺ cells (arrow), CD5⁻CD23⁺B220⁺ (B2) cells (in periodontal tissues and hearts. (C-D) Representative flow cytometry plots of TNF-α and IL-6 positive cells in CD5⁻CD23⁺ B2 cells. (E-H) Quantitative analysis of B-cell subsets and cytokine positive cells in periodontal tissues and hearts. (I-J) Immunofluorescence staining of CD23 (B2 cell marker) and TNF-α in periodontium and hearts. N = 6 for for all panels. ∗p < 0.05 compared to LIP-MI group. N.S., not significant.

Tumour necrosis factor-α (TNF-α) and interleukin-6 (IL-6) are critical regulators of B cell function [19,43], as well as the key mediators of bone tissue regeneration and apoptosis [44], Importantly, we noticed a significant reduction of TNF-α-positive cells after the hydrogel therapy in both lymph node and heart tissue (Fig. 8C–D, G-H). This selective modulation of TNF-α highlights the hydrogel's potential to modulate the periodontal microenvironment and alleviate MI-associated cardiac inflammation through B2 cell/TNF-α signaling, thereby establishing a mechanistic link between its local anti-inflammatory effects in the oral cavity and therapeutic benefits in periodontitis-associated MI.

2.8. Limitations and future directions

While this study demonstrates the therapeutic efficacy of the C1.5P4/BMP-2 hydrogel for periodontitis and comorbid MI, key considerations remain for its clinical translation. The association between hydrogel-mediated microbiome modulation and improved cardiac function has yet to be definitively established as a causal relationship, and additional mechanistic investigations would help to elucidate the nature of this regulatory link. Additionally, considering the BMP-2 translational concerns, long-term in vivo follow-up biocompatibility evaluation may be preferred to assess the long-term safety of BMP-2 delivery via this hydrogel, including potential risks of ectopic bone formation, systemic distribution, chronic systemic BMP-2 exposure, and associated impacts on systemic bone metabolism. Addressing these limitations will help optimize the hydrogel's translational potential as well as exploring the feasibility of combining this hydrogel with clinical periodontal therapies for personalized treatment regimens.

3. Conclusions

In summary, we developed a multifunctional hydrogel system, C1.5P4/BMP-2, as a promising therapeutic platform for the treatment of periodontitis and its associated MI. This injectable, highly biocompatible hydrogel exhibits antimicrobial, osteoinductive, and anti-inflammatory properties. Through sustained release of AMP and BMP-2, C1.5P4/BMP-2 effectively manages periodontitis and associated MI by: (1) selectively enriching probiotics, reducing the microbial G⁻/G⁺ ratio and inhibiting microbial dissemination; (2) downregulating proinflammatory B2 cell/TNF-α signaling axis in inflamed periodontium, which subsequently attenuates systemic inflammation and improves cardiac function via the periodontal-systemic axis. This integrated local-to-systemic dual-targeted strategy provides a translatable therapeutic approach for managing local infections and their systemic complications, particularly CVD.

4. Materials and methods

4.1. Synthesis of thiolated CS (CS-SH)

CS-SH was synthesized with slight modifications to previously reported methods [45]. Briefly, chitosan (100 mg) was dissolved in 1% (w/v) acetic acid solution, followed by pH adjustment to 5.5 using 5 M NaOH. The solution was then reacted with 2-iminothiazolidine hydrochloride (20 mg) for 16 h at room temperature. Following a 3-fold dilution with water, the product was dialyzed sequentially against: (a) 5 mM HCl, (b) 5 mM HCl/1% NaCl, and (c) 5 mM HCl/0.4 mM HCl. The solution was acidified to pH 4 using 10% (v/v) HCl, lyophilized and stored at −20 °C till use. Thiol group content of CS-SH was quantified using Ellman's reagent (5,5′-dithiobis-(2-nitrobenzoic acid), Sigma-Aldrich, D218200) following a standard colorimetric protocol.

4.2. Synthesis of PEG-Mal-AMP

The PEG-Mal-AMP conjugate was synthesized via thiol-maleimide coupling [46]. Specifically, PEG-Mal (25.2 mg) was dissolved in PBS (6 mL, pH 7.4). An equimolar quantity (4 mg) of antimicrobial peptide Tet213-AMP (sequence: KRWWKWWRRC; purchased from MedChemExpress, China) was added to achieve a 1:1 M ratio. The reaction proceeded at 30 °C for 6.5 h under constant agitation. The resulting conjugate was purified through dialysis against ultrapure water for 72 h, lyophilized and stored at −20 °C for further use.

4.3. Antibacterial hydrogels preparation

Hydrogels were fabricated by stoichiometrically crosslinking CS-SH with PEG-Mal and PEG-Mal-AMP. Specifically, 4% (w/v) solutions of PEG-Mal and PEG-Mal-AMP conjugates were prepared separately in PBS (10 mM, pH 8.2), while CS-SH was dissolved in the same buffer to final concentrations of 1%, 1.5%, or 2% (w/v). The precursor solutions were combined at a 1:1:1 M ratio (PEG-Mal: PEG-Mal-AMP: CS-SH) under vigorous vortex mixing, followed by incubation at 37 °C until complete gelation occurred. For BMP-2-loaded formulations, BMP-2 (2.0 μg/mL) was incorporated into PEG-Mal or PEG-Mal-AMP solutions prior to mixing with CS-SH.

4.4. Hydrogel characterization

Hydrogel viscoelastic properties were quantified using a Kinexus rheometer (Malvern) equipped with 25 mm parallel plates. Samples (700 μL) were loaded at a 0.75 mm gap and equilibrated at 37 °C. Frequency sweep analyses (0.1-10 Hz) were performed at 1% constant strain, confirmed to be within the linear viscoelastic region. Storage modulus (G′) and loss modulus (G″) were recorded in triplicate.

Swelling behavior was assessed by immersing hydrogels in PBS at 37 °C for 24 h to remove unreacted components, followed by lyophilization and dry weight (W0) measurement. Samples were then rehydrated in PBS (37 °C, 100 rpm orbital shaking) with weights recorded at predetermined intervals (W_t). The percentage of weight loss was calculated as:

Weight loss (%) = (W0-Wt)/W0 × 100%

where W0 and Wt represent the initial dry weight and the remaining weight at time t, respectively.

Hydrogel microstructure was analyzed by scanning electron microscopy (SEM). Briefly, samples were lyophilized, sputter-coated with 10 nm gold/palladium (Q150T ES Plus, Quorum), and imaged using a Mira 3 SEM (TESCAN) at 5 kV accelerating voltage. Pore size distribution was quantified using NIH ImageJ software.

4.5. BMP-2 and PEG-Mal-AMP release kinetics

BMP-2 release profiles were characterized using 200 μL hydrogels immersed in 2 mL PBS (pH 7.4, 37 °C) with orbital shaking (60 rpm). At predetermined intervals, 0.5 mL supernatant was collected and replaced with fresh PBS to maintain sink conditions. Released BMP-2 concentration in the supernatant was determined via a commercial BMP-2 ELISA Kit.

PEG-Mal-AMP release profiles were characterized using 1 mL hydrogels immersed in 30 mL PBS (pH 7.4, 37 °C) with orbital shaking (60 rpm) for up to 14 days. At predetermined intervals, 0.2 mL supernatant was collected and replaced with fresh PBS to maintain sink conditions. Released PEG-Mal-AMP concentrations were quantified by HPLC.

4.6. In vitro cytocompatibility and cytotoxicity

Human periodontal ligament fibroblasts (hPDLFs, Sciencell) were seeded in 24-well plates at 1 × 10⁴ cells/mL in Dulbecco's modified Eagle's medium (DMEM, Gibco) supplemented with 10% (v/v) fetal bovine serum (FBS, Gibco). Hydrogels formed in situ in Transwell inserts (0.4 μm pore, Corning) were placed above the hPDLFs. Viability was assessed at predetermined intervals using a live/dead staining kit (Beyotime, Shanghai).

Cytocompatibility was further evaluated using a Cell Counting Kit-8 (CCK-8, Beyotime). Hydrogel extracts were prepared by incubating sterile hydrogel samples (100 μL) in 1 mL serum-free DMEM for 48 h at 37 °C. The conditioned media was supplemented with 10% FBS prior to exposure to subconfluent hPDLFs. After 48 h of exposure, cell viability was quantified by adding CCK-8 reagent (10% v/v) and recording absorbance at 450 nm following the manufacturer's protocol.

4.7. Alkaline phosphatase activity

hPDLFs were seeded in the lower chambers of transwell plates at a density of 5x10⁴ cells/mL. Osteogenic induction medium (10 mM β-glycerophosphate, 50 μg/mL ascorbic acid, and 10⁻⁷ M dexamethasone) was added and incubated for 12 h. Subsequently, hydrogels containing BMP-2 were then placed in the upper chambers. After 7 days of culture, the cultures were fixed with 4% paraformaldehyde for 30 min, rinsed with PBS, and stained using BCIP/NBT solution (Beyotime). Following three washes with double-distilled water, the cells were observed under a microscope.

4.8. Alizarin Red S staining

hPDLFs were osteogenically differentiated as described above. Hydrogels encapsulating BMP-2 were placed in transwell inserts and maintained for 14 days. After washing three times with PBS, the cells were fixed with 4% paraformaldehyde for 30 min, rinsed with PBS again, and stained with freshly prepared 2% Alizarin Red S solution (Beyotime) for 30 min. Stained cells were then observed under a microscope.

4.9. Real-time qPCR

Following 7-day coculture with BMP-2-loaded hydrogels in transwells, hPDLFs were lysed and total RNA was extracted via TRIzol reagent. The isolated RNA was then reverse transcribed into cDNA using PrimeScript™ RT reagent kit (Takara), and the expression of osteogenic differentiation-related genes, including osteocalcin (OCN), alkaline phosphatase (ALP), and collagen type I (COL-1) were quantified by real-time PCR (RT‒qPCR). Gene expression was normalized to the housekeeping gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH), via the 2^−ΔΔCT method.

4.10. Antibacterial ability evaluation

The antibacterial efficacy of the hydrogels against representative key periodontal pathogens Fn (ATCC 25586), Pg (W50), and Aa (ATCC 700685) was evaluated by colony-forming unit (CFU) assay. Briefly, 100 μL of hydrogel in 96-well plates were inoculated with bacterial cultures (1 × 10⁶ CFU/mL) and incubated anaerobically for 6 h. The control group was treated with an equal volume of bacterial cultures in PBS. After incubation, bacteria were resuspended in PBS and collected from the hydrogel surface by gently pipetting. Bacteria obtained were either observed with a confocal microscope or cultured on sheep blood agar plates (Thermo Fisher, USA) and enumerated the following day.

4.11. Animal experimental set up

All experiments were approved by the Institutional Animal Care Committee of Shanghai Ninth People's Hospital (SH9H-2025-A1672-1). Male C57BL/6 mice (8-week-old) from Charles River were acclimated for 7 days in specific pathogen-free (SPF) condition prior to experimentation, and mice were randomly assigned to experimental groups with 6 mice per group for the periodontitis model. The periodontitis model was established by ligating the bilateral maxillary second molars of mice with 5-0 silk sutures to initiated LIP. Hydrogels (20 μL) were administered subgingivally on a weekly basis.

For the periodontitis-MI (LIP-MI) cohort, MI was induced two weeks post-LIP, with group sizes of N = 12 (Control), 11 (MI), 11 (LIP-MI), 8 (LIP-MI + C1.5), 8 (LIP-MI + C1.5P4) and 13 (LIP-MI + C1.5P4/BMP-2). Specifically, mice were continuously anesthetized with isoflurane respiration and positioned supine. After thoracic shaving and disinfection, a left thoracotomy exposed the heart. The left anterior descending coronary artery was permanently ligated with 7-0 polypropylene suture. Successful occlusion was confirmed by regional myocardial pallor and ST-segment elevation on intraoperative echocardiography. Hydrogels (20 μL) were administered subgingivally twice weekly, with treatments initiated two weeks after LIP establishment.

4.12. Cardiac function evaluation

Successful MI modeling was confirmed via echocardiography using the FUJIFILM VisualSonics Vevo 3100 LT system, with quantitative analysis of parameters including left ventricular ejection fraction (EF), Fractional Shortening (FS), Left Ventricular End-Diastolic Dimension (LVDd), and Left Ventricular Internal Dimension in Systole (LVIDs). The diagnostic criteria for successful MI modeling required concurrent demonstration of: (1) EF reduction ≥30% from baseline, (2) FS decrease ≥25%, and (3) characteristic ventricular remodeling evidenced by increased LVDd (>10%) and reduced anterior wall thickness during diastole.

4.13. Micro-CT

After sacrifice, the alveolar bone specimens were fixed in 4% polymerized formaldehyde at 4 °C for 48 h and scanned via a high-resolution micro-CT system (Bruker, Belgium). The scanning data underwent three-dimensional reconstruction via Analyze 12.0 software (PerkinElmer), with standardized thresholds set to equate bone mineral density to a grayscale range of 3000-8000 units, ensuring data accuracy and consistency.

4.14. Histological analysis

Tissue samples of the maxillary alveolar bone, heart, livers, spleens, lungs and kidneys were fixed, embedded and sectioned at 5 μm. The livers, spleens, lungs and kidneys sections were then subjected to hematoxylin and eosin (H&E) staining to visualize their general morphological structures, maxillary alveolar bones were subjected to H&E and Masson's trichrome staining to differentiate collagen fibers from other components within the bone tissue, and heart sections were sent for Sirius red staining to analyze the distribution and content of collagen fibers in the cardiac tissue.

4.15. Flow cytometry

Cardiac tissues were minced into pieces smaller than 1 mm³ and then enzymatically digested in RPMI 1640 containing collagenase IV (1 mg/mL, Sangon), dispase II (0.5 mg/mL, Sigma), and DNase I (20 U/mL, Applichem) at 37 °C for 40 min with gentle agitation. Digests were sequentially filtered through 100 μm and 70 μm strainers (BD) to remove undigested tissue fragments and cell aggregates. Viable cells were collected and resuspended. For the submandibular cervical lymph nodes, the tissues were mechanically dissociated through 70 μm strainers in PBS supplemented with 2% FBS. Single-cell suspensions were stained with fluorochrome-conjugated antibodies and subjected to flow cytometry and analyzed using a BD LSRFortessa X-20 Cell Analyzer.

Antibodies used were: Fc-block (101320, Biolegend), Zombie NIR™ Fixable Viability Kit (423105, Biolegend), CD45-BV421 (103134, Biolegend), CD3-APC (100235, Biolegend), CD4-PE/Cy7 (100527, Biolegend), CD8-BV510 (100752, Biolegend), CD11b-PE/Cy7 (101215, Biolegend), Ly-6G-PE (127607), Ly-6C-BV510 (128033, Biolegend), CD64-FITC (139315, Biolegend), CD45R/B220-FITC (553087, BD), CD23-PE/Cy7 (101613, Biolegend), CD5-BV711 (100639, Biolegend), TNF-α-BV605 (506329, Biolegend), and IL-6-APC (504507, Biolegend).

4.16. Full-length 16S analysis

Following euthanasia, fresh periodontal ligatures and heart tissues were carefully isolated. Genomic DNA was extracted using a DNA extraction kit (QIAGEN, Germany), quantified via NanoDrop (Thermo Fisher, USA) and verified via 1% agarose electrophoresis. The universal PCR primers (F: AGAGTTTGATCMTGGCTCAG R: ACCTTGTTACGACTT) were adapted to amplify the full-length 16S rRNA (V1-V9 regions) genes. SMRT (Single Molecule Real-Time) sequencing of community DNA fragments was conducted by Personalbio (Shanghai) via the PacBio Sequel next-generation sequencing platform. Raw reads were processed in QIIME 2 2019.4 [47]. Taxonomy classification of ASVs was conducted using the classify-sklearn naïve Bayes taxonomy classifier within the feature-classifier plugin [48], referenced against the Greengenes 13_8 99% OTUs sequences [49].

Microbiome analyses included: calculation of α-diversity indices (Simpson index, Shannon index, Chao1 richness, and Observed species) and β-diversity on rarefied ASV tables; assessment of group structural differences via PERMANOVA; visualization of shared ASVs using Venn diagrams; identification of differentially abundant taxa using LEfSe (LDA score >3.0, Kruskal-Wallis α < 0.05); supervised modeling of group separation via OPLS-DA and Random Forest; model validation by ROC-AUC and permutation testing (n = 1000). Spearman correlations were computed between significant microbial features and metabolite concentrations.

4.17. Hemolysis assay

Mouse whole blood was centrifuged at 2000 rpm for 10 min to collect red blood cells (RBCs), which were washed 2-3 times with PBS and resuspended to a 5% (v/v) stock solution in PBS. A 50 μL hydrogel was mixed with 400 μL RBC stock solution, adjusted to 1 mL with PBS, and incubated at 37 °C for 3 h. RBC suspensions with 600 μL PBS (negative control) or Triton X-100 (0.1%, v/v, positive control) were set up in parallel. After incubation, 500 μL culture media was collected and the mixture was centrifuged at 2000 rpm for 10 min, and the supernatants were collected for spectrophotometric analysis at 540 nm.

4.18. In vivo degradation

CY5.5-PEG-SH (MELOPEG, China) was incorporated into the hydrogel at a dosage of 1/20 the amount of PEG-Mal-AMP prior to the experiment. Male C57BL/6 mice were anesthetized, and the left dorsal scapular region was depilated with depilatory cream before imaging session to avoid hair interference. A 100 μL aliquot of CY5.5-PEG-SH -labeled hydrogel was subcutaneously injected into the prepared area. Imaging was performed at predetermined time points using the Vieworks Visque system. Regions of interest (ROIs) were defined around the injection sites, and mean fluorescence intensity was quantified using the system software to reflect residual hydrogel amount over time.

4.19. Statistical analysis

All the data are presented as the means ± SD and were analyzed using Prism (GraphPad Software, Inc.). Two-way ANOVA was used for multigroup comparisons, and Student's t-test was used for two-group comparisons. A value of p < 0.05 was considered statistically significant.

CRediT authorship contribution statement

Qinfeng Ding: Writing – review & editing, Writing – original draft, Software, Resources, Project administration, Methodology, Investigation, Data curation. Yilong Teng: Writing – review & editing, Visualization, Validation, Methodology, Formal analysis. Hong Zhu: Methodology, Investigation. Zhitong Jia: Methodology, Investigation. Shuo Xu: Methodology. Bin Hu: Methodology. Boyan Chen: Methodology, Formal analysis. Jiewen Dai: Project administration. Jian Sun: Methodology, Formal analysis, Data curation. Ping Nie: Formal analysis. Zhixiang Liu: Validation. Qijing Chen: Methodology. Xudong Wang: Supervision, Project administration, Funding acquisition. Xiaoyang Xu: Writing – review & editing, Writing – original draft, Project administration, Conceptualization. Xueqing Zhang: Project administration, Data curation, Conceptualization. Shengzhong Duan: Writing – review & editing, Writing – original draft, Supervision, Funding acquisition, Conceptualization.

Ethics approval and consent to participate update

All procedures involving animals were approved by the Institutional Animal Care Committee of Shanghai Ninth People's Hospital (SH9H-2025-A1672-1).

All authors are in compliance with all relevant ethical regulation.

Declaration of competing interest

Xiaoyang Xu is an editorial board member for Bioactive Materials and was not involved in the editorial review or the decision to publish this article. All authors declare that there are no competing interests.

Appendix B. Supplementary data

The following is the Supplementary data to this article: