Interaction between p-AMPK and RIPK3 reveals a mechanism by which AdipoAI suppresses necroptosis to attenuate periodontitis

Abstract

Background

Necroptosis is implicated in the pathogenesis of various inflammatory diseases, including periodontitis. This study aimed to investigate the molecular mechanisms underlying necroptosis in gingival fibroblasts (GFs) and to evaluate the therapeutic potential of AdipoAI, a novel adiponectin receptor agonist, administered locally, along with its regulatory effect on necroptosis.

Methods

Analysis of single-cell RNA sequencing data using the AUCell scoring system demonstrated significant activation of necroptosis in human periodontal tissues, with GFs identified as the primary cellular target. Integrated approaches, including molecular docking, co-immunoprecipitation, and immunofluorescence, revealed a physical association between phosphorylated AMP-activated protein kinase (p-AMPK) and Receptor-Interacting Protein Kinase 3 (RIPK3). Additionally, we established a mouse model of experimental periodontitis and an LPS/AZD’5582/z-VAD-fmk (LAZ)-induced necroptosis model in hGFs. Techniques including Western blotting, flow cytometry, and transmission electron microscopy were employed to evaluate the effects of AdipoAI administration.

Results

Local administration of AdipoAI significantly alleviated gingival inflammation and alveolar bone resorption in mice. In hGFs, AdipoAI effectively suppressed the activation of p-RIPK3 and phosphorylated Mixed Lineage Kinase Domain-Like protein (p-MLKL) and reduced the production of inflammatory factors. Mechanistic investigations revealed that AdipoAI binds to Adiponectin Receptor 1 (AdipoR1) - adiponectin receptor interacting protein (APPL1) to activate the AMPK signaling pathway, facilitating the association between p-AMPK and RIPK3, thereby inhibiting RIPK3/MLKL-mediated necroptosis.

Conclusions

This study is the first to unveil a mechanism by which AdipoAI suppresses GF necroptosis through p-AMPK-mediated RIPK3 regulation, offering new strategic insights and experimental evidence for targeted periodontitis therapy.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12967-026-08623-y.

Background

As a form of programmed cell death, necroptosis is a crucial defense mechanism against certain pathogenic invasions [1]. Necroptosis is primarily organized by the necrosome, which comprises receptor-interacting protein kinase 1 (RIPK1), receptor-interacting protein kinase 3 (RIPK3), and mixed lineage kinase domain-like protein (MLKL) [2]. RIPK1 primarily acts as a scaffold protein and upstream regulator whose kinase activity is often dispensable, while RIPK3 and MLKL function as the key executioners. Among them, RIPK3 serves as the core kinase whose activation represents the decisive step in initiating downstream death signaling. The pathological potential of necroptosis is further amplified by the transmissible nature of RIPK1 / RIPK3 amyloid fibrils, which propagate death signals in neurodegenerative contexts [3]. The phosphorylated MLKL triggers mitochondrial mtDNA release, leading to cGAS–STING activation and IFN-β production, and drives intestinal inflammation in murine IBD models [4]. In experimental periodontitis studies, Mlkl^-/- mice showed significantly reduced alveolar bone resorption [5]. This suggests that RIPK3/MLKL-mediated necroptosis considerably contributes to the pathogenesis of periodontitis [6–8]. Our previous studies have found that local injection of GSK’872 or sh-Mlkl can alleviate periodontitis by inhibiting necroptosis [9]. Collectively, RIPK3 serves as the master regulator and critical checkpoint, and its inhibition represents a promising upstream strategy for treating necroptosis-driven inflammatory disorders.

The precise control of RIPK3 activity is orchestrated by a intricate network that integrates upstream regulators, diverse post-translational modifications, and conformational shifts, with the AMPK-RIPK3 dialogue acting as a central processor for metabolic and inflammatory stress responses. AMP-activated protein kinase (AMPK), a central sensor of cellular energy status, has gained increasing recognition for its anti-inflammatory and cytoprotective roles. Accumulating evidence indicates a complex interplay between AMPK signaling and cell death pathways [10, 11]. Recent advances have elucidated a sophisticated bidirectional regulatory network between AMPK and RIPK3 in necroptosis control [12]. Emerging evidence identifies RIPK3 as an upstream kinase that directly binds to and phosphorylates AMPK at the critical threonine residue (T172), thereby connecting programmed cell death with metabolic stress responses [13]. Conversely, activated AMPK initiates a negative feedback loop by phosphorylating the E3 ubiquitin ligase Parkin, which subsequently ubiquitinates RIPK3 and disrupts necrosome assembly [14]. This intricate crosstalk represents a crucial self-limiting mechanism that fine-tunes necroptotic signaling. Therefore, elucidating the regulatory mechanism within the AMPK-RIPK3 axis holds promise for yielding novel therapeutic targets for necroptosis-driven inflammatory and degenerative diseases.

As a core energy sensor and metabolic regulatory hub, AMPK participates in the progression of periodontitis by regulating inflammatory responses, bone metabolic balance, and cellular autophagy. AMPK activation inhibits NF-κB signaling, reducing pro-inflammatory cytokine secretion [15]. ENTR1 suppresses macrophage M1 polarization via AMPK phosphorylation, alleviating bone loss—an effect reversed by AMPK inhibitor Compound C (CpC) [16]. Poly-T sequence-modified gold nanorods activate macrophage AMPK, inhibiting M1 differentiation and reducing TNF-α, IL-6, and IL-1β [17]. AMPK is crucial for bone homeostasis, upregulating osteogenic markers in rBMSCs and increasing calcified nodule formation [18]. It downregulates RANKL/OPG ratio to inhibit osteoclast differentiation, while activating mTORC1 to promote osteoblast proliferation and enhancing β-catenin stability to potentiate Wnt signaling, thereby reducing alveolar bone loss [19]. AMPK initiates autophagy by phosphorylating ULK1 and regulating PINK1/Parkin-mediated mitophagy, clearing damaged organelles to inhibit NLRP3 inflammasome activation [20]. Additionally, AMPK enhances antioxidant capacity by promoting FOXO nuclear translocation, upregulating SOD and catalase, and maintains mitochondrial homeostasis to reduce apoptosis [21]. In summary, AMPK serves as a central node connecting immune dysregulation, metabolic disorders, and bone destruction in periodontitis, representing a promising therapeutic target.

Adiponectin (APN), which is an endogenous bioactive peptide or protein, exerts multiple biological effects, such as inflammation inhibition and bone metabolism regulation [22–24]. These effects are mediated through the activation of downstream signaling pathways, such as AMPK, NF-κB, and MAPK, which bind to AdipoR1, AdipoR2, and T-cadherin [25]. However, limitations such as off-target effects, suboptimal drug delivery efficiency, and limited tissue regenerative capacity hinder their clinical translation prospects. Recently, research on adiponectin receptor agonists has gained attention because they can mimic the physiological functions of adiponectin. Studies have shown that APN and its receptor agonists can mitigate supporting tissue destruction in experimental periodontitis, making them potential candidates for periodontitis treatment [26]. Previously, we screened and reported a novel anti-inflammatory APN receptor agonist, called Adipo anti-inflammation agonist (AdipoAI). Through a series of in vivo and in vitro experiments, we observed that AdipoAI exhibited potent anti-inflammatory effects [27]. AdipoAI attenuates periodontitis in diabetic rats by inhibiting hGF-elicited macrophage migration, and it concurrently reduces alveolar bone loss and gingivitis in a diabetic peri-implantitis context [28, 29]. However, the effectiveness of the topical use of AdipoAI in the treatment of periodontitis is not yet clear, and its specific mechanism, especially in inhibiting necroptosis in GFs, remains unclear.

This study aims to elucidate the specific mechanism by which p-AMPK physically associates with RIPK3 to suppress necroptosis in GFs. Concurrently, we will investigate whether local administration of AdipoAI activates AMPK to inhibit necroptosis to alleviate gingival inflammation and alveolar bone loss in a mouse model of periodontitis. The findings are expected to provide preclinical evidence supporting the potential of AdipoAI as a candidate therapeutic for necroptosis-driven chronic inflammatory diseases.

Methods

Specimen collection, hGF isolation, culture and treatment

Human gingival tissues were collected from the interdental papillae of 10 healthy individuals and 5 samples from patients with chronic periodontitis. Among the 10 healthy samples, 5 were used for histological and biochemical detection of necroptosis markers, while the remaining 5 were used for isolation of human gingival fibroblasts (hGFs) for in vitro experiments. The sample collection and research were approved by the Ethics Committee of Southern Medical University (Approval No. NFEC2024-533). The basic information of the 15 participants is shown in Appendix Table 1. Standardized gingival collar biopsies measuring 5 mm in length and 3 mm in width were obtained from donors under local anesthesia. Gingival biopsies from healthy individuals were harvested from areas without BOP that met the criteria for good oral health and had a probing depth (PD) ≤3 mm. Biopsies from periodontitis patients were obtained from areas exhibiting severe inflammation and bone loss (BOP positive and PD > 5 mm). The other inclusion criteria were: 1) no smoking, 2) no systemic diseases, 3) no intake of antibiotics or anti-inflammatory medications in the past 3 months, 4) no periodontal therapy within the last 6 months, 5) no pregnancy or breastfeeding, 6) no acute infections or allergies, and 7) no immunosuppressant treatment in the past 3 months. Healthy gingival tissue specimens were obtained during crown lengthening procedures, while biopsies from periodontitis patients were collected prior to the extraction of teeth with no retention value.

hGFs were isolated from human gingival tissues according to an earlier study [30]. Human gingival tissues were washed three times, cut into pieces of 1 mm², and cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco, USA) supplemented with 10% fetal bovine serum (FBS; Gibco, USA) and 1% penicillin/streptomycin solution (Gibco, USA). When the cells had grown to 80–100% confluence, the hGFs were passaged with 0.01% trypsin (Gibco, USA). The hGFs were used for experiments between passages three and seven. hGFs were pretreated with 20 μM DMSO or 2.5, 5, 10, or 20 μM AdipoAI for 24 h. Then, the cells were washed with PBS. Medium containing LAZ plus 20 μg/mL Escherichia coli-derived lipopolysaccharide (LPS; Sigma‒Aldrich), 1 μM AZD’5582 (MedChemExpress, Shanghai, P.R.C.), and 50 μM z-VAD-fmk (MedChemExpress, Shanghai, P.R.C.) was added.

Experimental periodontitis model

All animal procedures were approved by the Animal Care and Use Committee of Nanfang Hospital (No. IACUC-LAC-20230207-004). The animals were maintained under temperature and humidity control (22±2 °C; 55 ± 10% humidity) on a 12-h light-12-h dark cycle. All surgeries were performed in a specific pathogen free (SPF) animal laboratory. The mice were randomly divided into five groups (n = 5 per group): untreated control, periodontitis, periodontitis + DMSO, periodontitis + AdipoAI, and periodontitis + GSK’872. To induce experimental periodontitis, a sterile 5–0 silk suture was tied in a subgingival position around the right second maxillary molars of male C57BL/6N mice, as previously described [31]. DMSO, AdipoAI, or GSK’872 suspended in PBS was injected into the buccal and palatal gingiva of the ligated tooth. On Day 14, the animals were euthanized, and maxillae were collected for analysis.

Single-cell RNA sequencing data processing and analysis

To systematically characterize the cellular heterogeneity of gingival tissues, we integrated the public scRNA-seq dataset GSE164241 (GEO) with our in-house gingival scRNA-seq data (OMIX: OMIX017754) [32].

Raw 10x Genomics expression matrices were processed using Seurat (v4.3.0). Cells were retained if they met the following criteria: nFeature_RNA > 300, nFeature_RNA < 7000, nCount_RNA > 1000, and mt_percent < 10%. Data were normalized using NormalizeData (LogNormalize), and 2,000 highly variable genes were identified using FindVariableFeatures (“vst”). After scaling with ScaleData (regressing out mt_percent), PCA was performed, and batch effects were corrected using Harmony via IntegrateLayers (method = HarmonyIntegration, orig.reduction= “pca”, new.reduction=“harmony”). UMAP visualization and graph-based clustering were performed using the first 30 Harmony dimensions, with a resolution of 0.5.

Cell types were annotated using canonical marker genes from the CellMarker and PanglaoDB databases. The principal populations included T/NK cells, endothelial cells, epithelial cells, fibroblasts, plasma cells, B cells, pericytes, macrophages, mast cells, and neutrophils. Sample information and cellular composition are provided in Appendix Table 3, and marker genes for annotation are listed in Appendix Table 4.

Programmed cell death pathway activity was quantified using AUCell, which calculates an enrichment score for a given gene set based on gene expression ranking in each cell. Scores were calculated for pyroptosis, apoptosis, necroptosis, ferroptosis, and cuproptosis using gene sets from MSigDB (v7.4) and relevant literature (Appendix Table 2).

Statistical analysis

All data in the figures are presented as the mean ± SEM of independent experiments. The data were analyzed using Prism 7 (GraphPad Software). One-way analysis of variance (ANOVA), followed by Tukey’s honestly significant difference post hoc test, was used to compare differences between multiple groups. Student’s t-test was used to compare differences between two groups. A p value of <0.05 was regarded as statistically significant.

Appendix methods

Additional methods are shown in the Appendix Material.

Results

Necroptosis is activated in gingival fibroblasts of human periodontitis tissues

To investigate the role of programmed cell death (PCD) in periodontitis, we first analyzed the scRNA-seq data from healthy and periodontitis gingival samples. The activity of PCD pathways was assessed using AUCell scoring. We found that the overall PCD signature was significantly elevated in periodontitis tissues compared to healthy controls (Fig. 1A). Notably, among various PCD marker genes, the expression of key necroptosis-specific markers was particularly prominent (Fig. 1B). Given this observation, we specifically evaluated the necroptosis signature. Both the collective gene expression level (Fig. 1C) and pathway activity (Fig. 1D) of the necroptosis feature gene set were significantly higher in periodontitis gingival tissues (Appendix Fig. 1). To identify the primary cell type undergoing necroptosis, we analyzed the cellular composition of the gingival mucosa. Among the ten major cell types identified, fibroblasts exhibited the highest necroptosis signature, based on exploratory scRNA-seq analysis. Furthermore, this necroptosis signature in fibroblasts was more pronounced in periodontitis samples than in healthy controls. Importantly, immunofluorescence staining confirmed the activation of necroptosis in gingival fibroblasts. We observed clear co-localization of the key necroptosis regulator p-RIPK3 with the fibroblast marker FSP-1 (Fig. 1E), as well as co-localization of the necroptosis executioner protein p-MLKL with FSP-1 (Fig. 1F). Quantitative analysis demonstrated that p-RIPK3 and p-MLKL signals were significantly increased in FSP-1-positive fibroblast-like stromal cells in periodontitis gingival connective tissues compared to healthy controls (Fig. 1G). Collectively, these exploratory observations suggest that necroptosis-related markers are increased in human periodontitis gingival tissues, and that this signal is particularly evident in fibroblast-like stromal cells.

Fig. 1.

Exploratory analysis of necroptosis-related markers in human gingival tissues. (A) Box plots display the area under the curve (AUC) values representing the collective gene expression levels of the programmed cell death (PCD) feature gene set in healthy and periodontitis gingival tissues. (B) Dot plots illustrate the expression of PCD marker genes. Expression values are normalized and scaled averages. (C) Box plots show the AUC values of the collective gene expression levels of the necroptosis feature gene set in human gingival tissues. (D) Box plots demonstrate the activity of the necroptosis feature gene set in healthy and periodontitis human gingival tissues. (E, F) Representative immunofluorescence (IF) staining images showing the co-localization of the necroptosis marker p-RIPK3/p-MLKL (red) with the FSP-1 (green) in healthy and periodontitis gingival tissues. (G) Quantification of p-RIPK3 and p-MLKL signals within FSP-1-positive areas. Data are presented as the ratio of p-RIPK3-positive area to FSP-1-positive area (left) and the ratio of p-MLKL-positive area to FSP-1-positive area (right)

Validation of the interaction between p-AMPK and RIPK3 proteins

Previous studies have suggested a potential functional interplay between AMPK, a central regulator of energy metabolism, and RIPK3, a key mediator of cell death, in cellular stress responses. However, whether these two proteins can interact in a cellular context remains unclear. To address this, we performed a series of validation experiments. The molecular docking analysis predicted potential binding interfaces between p-AMPK and RIPK3 (Fig. 2A), which were mainly composed of hydrogen bonds and salt bridges (Fig. 2B). Immunofluorescence staining of tissue sections further corroborated this finding by showing that p-AMPK and RIPK3 were present in the same cellular compartments, indicating that they might function together in the same signaling pathway (Fig. 2C, D). This co-localization suggests that p-AMPK and RIPK3 are spatially proximal within cells, raising the possibility of a functional association that may influence necroptosis. Co-immunoprecipitation (Co-IP) was performed according to the manufacturer’s instructions, and the precipitated complexes were analyzed by Western blotting. Most critically, Co-IP assays demonstrated a physical association between p-AMPK and RIPK3, indicating that these two proteins exist within the same protein complex (Fig. 2E). This finding provides biochemical evidence that the two proteins can form a complex.

Fig. 2.

A novel regulatory interaction between p-AMPK and RIPK3 proteins. (A) Molecular docking analysis between p-AMPK and RIPK3. left: surface pattern diagram; right: ribbon pattern diagram. (B) Molecular docking analysis of the interaction between p-AMPK and RIPK3. 2D pattern diagram. (C, D) The co-localization of p-AMPK and RIPK3 in mouse gingiva was quantified by immunofluorescent (IF) staining. n = 5 per group. (E) Lysates were prepared from hGfs and incubated with antibodies specific to p-AMPK (or isotype control for negative control) by Co-immunoprecipitation (co-IP) experiment. The immunoprecipitates were then subjected to SDS-PAGE and analyzed by Western blot using antibodies against RIPK3 to determine if the interaction occurred. No interaction was observed in the negative control (IgG), confirming the specificity of the pull-down. n = 3 per group

In summary, this study is the first to demonstrate a physical interaction between p-AMPK and RIPK3 in a cellular context, suggesting a novel pathway through which AMPK may influence cell survival and metabolism. Although whether this interaction is direct and which domains mediate the binding remain to be fully elucidated, our findings provide biochemical evidence that p-AMPK and RIPK3 can associate within cells, offering a new direction for related research.

AdipoAI binds to AdipoR1/APPL1, leading to the activation of AMPK

To perform its functions, APN exerts its functions by binding to its receptors, AdipoR1 and AdipoR2 [33]. By binding to AdipoR1, APN activates a series of downstream signaling pathways, one of which involves APPL1. Acting as an adaptor protein binding to the intracellular domain of AdipoR1, APPL1 facilitates the activation of downstream signaling molecules. AdipoR1, as a membrane receptor, tends to have relatively stable expression, whereas APPL1 is a more plastic downstream effector protein. Hence, AdipoAI may primarily exert its effects by activating the downstream signaling pathways of the AdipoR1 receptor, rather than directly influencing the expression of the receptor itself.

IHC staining revealed markedly elevated APPL1 protein levels in the LIP+AI group compared to the control group (Fig. 3A, B, Appendix Fig. 2). The AMPK signaling pathway is a classic mechanism through which APN exerts its effects. Recent studies have suggested that AMPK also plays a regulatory role in necroptosis, a process that is increasingly recognized as important for treating various diseases. Therefore, we investigated whether AdipoAI could activate AMPK to inhibit necroptosis (Appendix Fig. 3). The WB results demonstrated a considerable increase in the protein levels of APPL1 and p-AMPK after co-stimulation with AdipoAI and LAZ compared to the control group (Fig. 3C, D).

Fig. 3.

AdipoAI activates the AdipoR1/APPL1/AMPK signaling axis in GFs. (A, B) IHC staining of AdipoR1、APPL1 and p-AMPK in the mouse gingival tissue and corresponding data quantification. n = 5 per group. (C, D) The levels of AdipoR1、APPL1 and p-AMPK proteins in hGfs and corresponding data quantification, were quantified by WB. n = 3 per group

AdipoAI suppressed RIPK3/MLKL-mediated necroptosis in hGfs

Next, we determined whether AdipoAI could inhibit necroptosis in hGFs in a concentration-dependent manner. The LAZ-induced hGF necroptosis model used in this study was established based on our previous work, in which pharmacological and genetic inhibition of the RIPK3/MLKL pathway confirmed that LAZ-induced hGF death was largely dependent on RIPK3/MLKL-mediated necroptosis. Therefore, in the present study, this previously validated model was used to evaluate whether AdipoAI suppresses RIPK3/MLKL-associated necroptotic signaling in hGFs. CCK-8 experiments indicated that AdipoAI had no significant cytotoxic effect on hGFs at concentrations of 40 μM or lower (Fig. 4C). To determine the optimal concentration of AdipoAI for inhibiting necroptosis, we treated hGFs with different doses of AdipoAI before LAZ stimulation (Appendix Fig. 5). WB showed that 20 μM AdipoAI could inhibit the expression of p-RIPK3 and p-MLKL (Fig. 4A, B). AdipoAI pretreatment significantly increased cell viability compared to the DMSO+LAZ group (Fig. 4D, E). AdipoAI administration considerably reduced the percentage of Annexin V⁺/PI⁺ cells, which was similar to that in the control group (Fig. 4F, Appendix Fig. 6A). The TEM results demonstrated that hGFs treated with LAZ exhibited typical necrotic morphological features, including an enlarged cell volume, a permeable plasma membrane, a translucent cytoplasm, and swollen organelles, including mitochondria, the endoplasmic reticulum, and the Golgi complex. AdipoAI alleviated ultrastructural damage in necroptotic cells (Fig. 4G, Appendix Fig. 4). Similarly, the mRNA levels of IL-6, IL-1β, and COX2 were significantly upregulated in hGFs treated with LAZ but were considerably inhibited by AdipoAI treatment in a dose-dependent manner (Fig. 4H). These data indicated that AdipoAI inhibited RIPK3/MLKL-mediated necroptosis and the production and release of inflammatory cytokines.

Fig. 4.

AdipoAI suppressed RIPK3/MLKL-mediated necroptosis and inflammation in hGfs. (A, B) The expression levels of RIPK3、p-RIPK3、MLKL and p-MLKL in hGfs and corresponding data quantification, as determined by Western blotting (WB). Glyceraldehyde–3–phosphate delydrogenase (GAPDH) was used as a loading control. Relative protein expression was normalized to that of the internal control.control. The data are shown as the mean±SEM. *p < 0.05; **p < 0.01; ***p < 0.001 and ****p < 0.0001 by unpaired Student’s t-test. n = 3 per group. (C) Survival rate of hGfs treated with different concentrations of AdipoAI. n = 3 per group. (D) The viability of hGfs that were pretreated with AdipoAI and then stimulated with LAZ was quantified by CCK8 assays. n = 3 per group. (E) The viability of hGfs that were pretreated with AdipoAI and then stimulated with LAZ was quantified by LDH assays. n = 3 per group. (F) Flow cytometric analysis of PI- and annexin V- stained cells. n = 3 per group. (G) Representative microphotographs of transmission electron microscopy. hGfs showed normal morphology in the DMSO and AI groups. The LAZ+DMSO group exhibited typical features of necrosis, including mitochondrial swelling (green arrows), vacuole-like changes (red arrows) and endoplasmic reticulum dilation (yellow arrows). The necrotic changes in the LAZ+AI groups were significantly attenuated. n = 3 per group. (H) The proinflammatory genes IL-6, IL-1β and COX-2 in LAZ-stimulated hGfs was quantified by qRT‒PCR. n = 3 per group

Inhibitory effect of AdipoAI on necroptosis in periodontitis mice

IHC staining revealed that p-RIPK3 and p-MLKL expression was significantly increased in the gingival connective tissue of periodontitis mice compared to controls. Global p-RIPK3 and p-MLKL expression significantly declined in the LIP+AI group (Fig. 5A, D). In particular, p-RIPK3/p-MLKL and the fibroblast marker fibroblast-specific protein-1 (FSP-1) were colocalized, and the p-MLKL⁺ area within FSP-1⁺ fibroblasts was significantly larger in periodontitis mice than in the control group, indicating a higher occurrence of necroptosis in the former. AdipoAI treatment considerably inhibited necroptosis activation in gingival tissues, exhibiting similar inhibitory effects as the RIPK3-specific inhibitor GSK’872 (Fig. 5B, C, E, F). Our preliminary research results show that targeting necroptosis of GFs can effectively alleviate periodontitis. Hence, AdipoAI may treat periodontitis by inhibiting necroptosis in GFs.

Fig. 5.

AdipoAI inhibited RIPK3/MLKL-mediated necroptosis in GFs in the gingival tissues of mice with periodontitis. (A, D) Representative immunohistochemistry (IHC) staining images of p-RIPK3、p-MLKL in the periodontium and corresponding data quantification. n = 5 per group. Scale bar, 100 μm. (B) Representative immunofluorescence (IF) staining images showing the co-localization of the necroptosis marker p-RIPK3 (red) with the gingival fibroblast marker FSP-1 (green) in mouse gingival tissues from the indicated groups. (C) Representative immunofluorescence (IF) staining images showing the co-localization of the necroptosis marker p-MLKL (red) with the gingival fibroblast marker FSP-1 (green) in mouse gingival tissues from the indicated groups. (E, F) Quantification analyses of necroptosis in mouse gingival fibroblasts. Data are presented as the ratio of p-RIPK3-positive area to FSP-1-positive area (E) and the ratio of p-MLKL-positive area to FSP-1-positive area (F). n = 5 mice per group. Scale bar, 50 μm. FSP-1 was used as a loading control. Error bars represent the mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001 and ****p < 0.0001 compared to the respective control groups by one-way analysis of variance

Involvement of AdipoR1/APPL1/AMPK in the process of AdipoAI inhibition of necroptosis in GFs

To validate the role of AdipoAI in inhibiting LAZ-induced necroptosis in hGFs via the AdipoR1/APPL1/AMPK pathway, we employed siRNA to knock down the AdipoR1 gene in the cells. We stimulated the cells with the AMPK agonist AICAR (Fig. 6A, B). The WB results showed that the inhibitory effect of AdipoAI was abolished with the addition of si-AdipoR1. The AICAR group considerably suppressed the expression of p-RIPK3 and p-MLKL, and this inhibitory effect persisted in the si-AdipoR1+AICAR group (Fig. 6C, D). Flow cytometry analysis confirmed that AdipoAI significantly reduced the number of necrotic cells in control cells, but this protective effect was abolished upon AdipoR1 knockdown. In contrast, AICAR treatment reduced necrotic cell death regardless of AdipoR1 knockdown (Fig. 6E, Appendix Fig. 6B). These data confirm that AdipoAI exerts its inhibitory effect on necroptosis through the AdipoR1/APPL1/AMPK signaling pathway.

Fig. 6.

AdipoAI activates APPL1/AMPK pathway in the process of inhibiting necroptosis in hGfs. (A, B) The levels of APPL1, AdipoR1, p-AMPK in hGfs (treated with AI/DMSO/AICAR) with or without si-AdipoR1 transfection and corresponding data quantification, were measured by WB. (C, D) The necroptosis-related proteins in hGfs (treated with AI/DMSO/AICAR) with or without si-AdipoR1 transfection and corresponding data quantification, were measured by WB. (E) Flow cytometric analysis of PI- and annexin V-stained cells. n = 3 per group

Local administration of AdipoAI alleviated gingival inflammation and bone resorption in periodontitis mice

To assess whether local application of AdipoAI could alleviate periodontitis in vivo, we used a ligature mouse model (Fig. 7A). Compared to the control group, ligature placement resulted in significant alveolar bone resorption and attachment loss, along with increased infiltration of inflammatory cells in the gingiva. AdipoAI considerably mitigated these effects (Fig. 7B). The results of BV/TV, Tb.N, and Tb.Sp indicated that periodontitis-induced alveolar bone loss could be rescued by AdipoAI (Fig. 7C, D). With H&E staining, the gingival epithelium of the LIP group was observed to have a significant elongation of the epithelial nail process and disordered periodontal fibers. However, inflammation of the gingival tissue in the LIP+AI mice was significantly inhibited, and the periodontal fibers were more regular than the LIP group (Fig. 7E). Moreover, AdipoAI treatment significantly decreased the mRNA levels of Il-6, Il-1β, and Cox-2 in the gingiva of ligated mice (Fig. 7F). These findings affirm that local application of AdipoAI effectively attenuates gingival inflammation and bone resorption in periodontitis mice.

Fig. 7.

Local administration of AdipoAI attenuates gingival inflammation and bone resorption in mice with periodontitis. (A) Procedural timeline: mice were adaptively fed for 1 week, an experimental periodontitis model was established by ligating with silk thread for 2 weeks. Simultaneously, mice in the LIP+AI group were administered AdipoAI, and those in the LIP+GSK’872 group were administered GSK’872, while those in the other groups were treated with DMSO as a control. (B) Methylene blue staining of the maxillae (red lines outline the alveolar bone crest (ABC) and the cemental-enamel junction (CEJ)). n = 5 per group. (C) Micro-CT visualization of alveolar bone (AB) loss. n = 5 per group (red lines show the CEJ-ABC distance). (D) Alveolar bone measurement by micro-CT analysis. n = 5 per group. (E) HE staining of gingivae between the first and second maxillary molars. n = 5 per group. Scale bar, 100 μm. (F) The expression of the proinflammatory genes Il-6, Il-1β and cox-2 in gingival tissues from mice was quantified by qRT‒PCR. n = 5 per group. Error bars represent the mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 compared to the respective control groups by one-way ANOVA

Discussion

Current research indicates that AMPK can regulate necroptosis-related diseases through multiple mechanisms. In acute lung injury, calcitonin gene-related peptide (CGRP) inhibits alveolar epithelial cell necroptosis by activating the AMPK/L-OPA1 signaling pathway [34]. In spinal cord injury, exosomes derived from Schwann cells also alleviate mitochondrial dysfunction and necroptosis after spinal cord injury through the AMPK signaling pathway-mediated mitochondrial autophagy [35]. AMPK activation also prevents necroptosis by promoting Keap1-mediated PGAM5 degradation [36]. Despite these well-established roles of AMPK in regulating necroptosis in various disease contexts, whether this pathway contributes to necroptosis in periodontitis remains unclear. Necroptosis involvement in the progression of periodontitis offers a potential therapeutic target. Elevated levels of RIPK1, MLKL, p-RIPK3, and p-MLKL have been found in gingival tissues of patients with chronic periodontitis, confirming the participation of necroptosis in disease progression [37]. RIP1-mediated necroptosis plays a role in the pathological process of periodontitis in mice, and Nec-1 inhibits necroptosis, reducing periodontal tissue inflammation and bone resorption in periodontitis [38]. O-GlcNAcylation has been found to prevent RIPK3–RIPK1 hetero- and RIPK3–RIPK3 homo-interactions, thus inhibiting downstream innate immune and necroptotic signaling [39], ). Furthermore, microRNA-214 regulates necroptosis by targeting ATF4 in diabetes-associated periodontitis [40]. Nec-1 also has a protective effect by reducing cell death and promoting the regeneration of ectopic periodontal tissue-like structures by inhibiting necroptosis [41]. These findings collectively establish necroptosis as a key pathological process in periodontitis, yet the upstream metabolic regulators of this pathway—particularly AMPK—have not been explored. Building on this evidence, our findings reveal that the AMPK-RIPK3 axis critically regulates gingival fibroblast necroptosis in periodontitis, thereby bridging a significant gap between AMPK-mediated metabolic control and periodontal inflammation.

Mechanistically, the physical association between p-AMPK and RIPK3 identified in our study aligns with recent findings on the metabolic regulation of necroptosis and further extends this area of research. Previous studies have demonstrated that AMPK can suppress necroptosis through indirect mechanisms. AMPK-Parkin inhibits the formation of the RIPK1–RIPK3 complex by promoting RIPK3 ubiquitination, alleviating Parkinson’s inflammatory diseases [14]. AICAR activates p-AMPK to stabilize pro-caspase-8, enhance RIPK3 degradation, inhibit acinar cell necroptosis, and improve severe acute pancreatitis [42]. Beyond these indirect regulatory pathways, our study demonstrates that in gingival fibroblasts, p-AMPK physically associates with RIPK3, interferes with necrosome assembly, and negatively regulates necroptosis. This competitive binding mechanism suggests that p-AMPK can directly intercept necroptotic signaling by disrupting necrosome assembly. Notably, the inhibitory effect of AdipoAI on necroptosis is not exclusively dependent on this interaction. Other signaling mechanisms or pathways through which AdipoAI exerts its protective effects against necroptosis may highlight the complexity of its action. Therefore, while the p-AMPK-RIPK3 interaction reveals a key regulatory mechanism, further studies are needed to fully define the pathways through which AdipoAI inhibits necroptosis. An interesting comparison emerges when considering the target selectivity of AMPK. Metabolic stress promotes TRAIL receptor-mediated activation of RIPK1, whereas AMPK directly inhibits RIPK1 activity by phosphorylating RIPK1 at Ser415, thereby blocking cell death and inflammatory responses [12]. In contrast, the present study found that AMPK preferentially targets RIPK3 rather than RIPK1, suggesting that the selectivity of AMPK toward necroptosis regulatory molecules may depend on the tissue microenvironment or the nature of the stimulus.

RIPK1 has traditionally been regarded as an essential upstream mediator of necroptosis. However, recent studies have found that necroptosis can also occur independently of RIPK1 under certain conditions. Studies have shown that necroptosis can be executed directly through RIPK3 and MLKL in certain cell types or specific signaling environments. For example, with some viral infections, the protein ZBP1 (DAI) can directly activate RIPK3, bypassing the need for RIPK1 and thus activating MLKL and triggering necroptosis [43]. Furthermore, under specific genetic conditions where cells lack certain upstream inhibitors, such as cIAPs or FADD, RIPK3 and MLKL can be directly activated without the involvement of RIPK1 [44]. These findings establish RIPK3 and MLKL as core components of the necroptotic signaling machinery and highlight their potential as therapeutic targets for diseases associated with this form of cell death. Given the central role of RIPK3 in executing necroptosis, elucidating the regulatory mechanisms governing its activity is crucial for understanding its contribution to the pathogenesis of periodontitis. The present study identifies an interaction between AMPK and RIPK3, revealing an active negative regulatory mechanism in which p-AMPK physically binds to RIPK3—rather than functionally replacing RIPK1—thereby interfering with necrosome assembly. This mode of regulation adds a layer of complexity to the necroptotic signaling network and underscores the existence of multifaceted control mechanisms that govern cell fate decisions within inflammatory microenvironments.

The Global Burden of Disease Study 2021 revealed that approximately 220 million people across China have severe periodontitis, with an age-standardized prevalence rate of 10.8%, and is the leading cause of irreversible alveolar bone resorption and tooth loss [45]. Although mechanical therapy is a cornerstone of periodontitis treatment, its efficacy in eliminating microbes from periodontal pocket linings remains limited. Local drug administration offers distinct advantages, including rapid onset, high bioavailability at target sites, and minimal systemic exposure. Among local delivery methods, gingival injection has been widely employed in preclinical studies and has shown efficacy with agents such as icariin, which reduces inflammation and promotes tissue regeneration in miniature pigs [46], and calcitonin, which attenuates alveolar bone resorption by modulating osteoclast activity in rats [47]. Currently, research on periodontitis treatments in animal models with adiponectin and its receptor agonists mainly involves systemic drug administration, with limited exploration of local application methods. Adiponectin and its receptor agonists have been shown to reduce alveolar bone loss and inflammation in periodontitis [48, 49]. Furthermore, we have found that AdipoAI reduces alveolar bone resorption and gingival inflammation associated with diabetes-related periodontitis [28]. For instance, local injection of APN attenuated orthodontic tooth movement and reduced relapse in rats [50]. Similarly, local intraventricular injection of APN and APR improved the volume of infarction and neurofunctional deficits induced by transient middle cerebral artery occlusion in mice [51]. These findings suggest that local drug delivery is a feasible strategy for periodontitis treatment and support the rationale for evaluating locally administered AdipoAI. Based on these precedents and the advantages of targeted delivery, we selected local injection of AdipoAI into gingival tissue as the administration method.

Several limitations of this study should be acknowledged. First, although a physical association between p-AMPK and RIPK3 was observed, whether this interaction is direct and the precise binding domains involved remain to be elucidated using advanced techniques such as surface plasmon resonance or cryo-electron microscopy. Second, the relatively short observation period in the animal model limits assessment of the long-term safety and potential drug resistance associated with sustained local AdipoAI injection. Future studies should address this limitation by extending the treatment duration in chronic periodontitis models and systematically evaluating the long-term histological and systemic effects of repeated local AdipoAI administration. Third, although FSP-1/S100A4 is commonly used as a fibroblast-associated marker, it is not entirely specific for fibroblasts under inflammatory conditions. Therefore, the FSP-1-based localization should be interpreted as indicating fibroblast-like stromal cells, and future studies using multiple fibroblast markers and lineage-specific approaches are warranted. Fourth, the human gingival tissue data should be interpreted as exploratory tissue-level validation rather than definitive clinical evidence. Larger, independently recruited cohorts with formal sample-size calculations are required to further confirm the clinical relevance of these findings.

Conclusions

This study demonstrates for the first time an interaction between p-AMPK and RIPK3, and shows that local administration of AdipoAI ameliorates gingival inflammation and alveolar bone resorption in a preclinical periodontitis model. Co-localization of p-RIPK3 and p-MLKL with FSP-1 in gingival connective tissues indicated that necroptosis was activated in fibroblast-like stromal cells in the periodontitis model, and AdipoAI treatment reduced the expression of these necroptotic markers. Mechanistically, AdipoAI binds to AdipoR1/APPL1 to activate the AMPK pathway, which facilitates the association of p-AMPK to RIPK3, thereby suppressing RIPK3/MLKL-mediated necroptosis and subsequent inflammatory responses. Together, these findings indicate that local AdipoAI treatment suppresses RIPK3/MLKL-associated necroptotic signaling through the AdipoR1/APPL1/AMPK pathway and alleviates experimental periodontitis in a preclinical model, suggesting that AdipoAI is a promising candidate for further investigation in periodontitis therapy.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

GFs

Gingival fibroblasts

AdipoAI

Adipo anti-inflammation agonist

AdipoR1/2

Adipo receptor 1/2

APPL1

Adaptor protein, phosphotyrosine interacting with PH domain and leucine zipper 1

MLKL

Mixed lineage kinase domain-like

RIPK

Receptor interacting protein kinase

DAMPs

Danger-associated molecular patterns

CEJ-ABC

Cement-to-enamel junction-alveolar bone crest

Author contributions

Wei Qiu, Dian Ding contributed to conception, design, data acquisition, analysis, interpretation, and drafted and critically revised the manuscript; Fuchun Fang and Hongle Wu contributed to conception, design, data acquisition, analysis, interpretation and critically revised the manuscript; Jun Shao, Yuxuan Huang, Zehao Chen contributed to data acquisition, analysis and interpretation and critically revised the manuscript; Ruiming Guo and Huaxuan Zhao contributed to data acquisition and analysis and critically revised the manuscript. All authors gave final approval and agreed to be accountable for all aspects of the work ensuring integrity and accuracy.

Funding

This study was supported by the National Natural Science Foundation of China grants (82270982, 82,470,975), Guangdong Provincial Natural Science Foundation (2026A1515012942), Guangzhou Key Research and Development Program (2024B03J0667), Natural Science Foundation of Guangdong Province (2024A1515010840), Basic and Applied Basic Projects of Huadu District in Guangzhou (24HDQYLH16).

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

All animal procedures were conducted in accordance with the National Regulations for the Administration of Laboratory Animals (China) and the ARRIVE 2.0 guidelines, and were approved by the Institutional Animal Care and Use Committee (IACUC) of Nanfang Hospital (No. IACUC-LAC-20230207-004). Every effort was made to minimize animal suffering. The collection and use of human gingival tissues were approved by the Ethics Committee of Southern Medical University (No. NFEC2024-533). Written informed consent was obtained from each donor, and the study was conducted in accordance with the Declaration of Helsinki (2013 revision).

Consent for publication

Not applicable.

Competing interests

The authors have declared that no competing interest exists.