Graphical abstract graphic file with name fx1.jpg [33]Open in a new tab Highlights * • Diabetes leads to a reduction of TRX1 level in the periodontium of periodontitis * • TRX1 promotes the osteogenesis by regulating Wnt/β-catenin signaling pathway * • TRX1 shows promising therapeutic effects in periodontitis with diabetes __________________________________________________________________ Orthopedics; Molecular biology; Cell biology Introduction Periodontitis is a prevalent oral inflammatory disease that violates the periodontium including gingiva, alveolar bone and periodontal ligament.[34]^1 It may cause tooth loss and increase the risk of chronic systemic inflammatory disease such as diabetes, Alzheimer’s disease and cardiovascular disease.[35]^1^,[36]^2^,[37]^3 More than 40% of global adults have periodontitis.[38]^4 Gold standard periodontal treatments at the clinic contain scaling and root planing, designed to remove the supragingival and subgingival plaque. Antimicrobials administration is an adjuvant therapy in patients with deep periodontal pockets.[39]^5 These therapies are markedly effective for most patients. However, in patients with inflammatory comorbidities such as diabetes, obesity and metabolic syndrome (MetS), the results are more complicated.[40]^6^,[41]^7^,[42]^8 As a major risk factor for periodontitis, diabetes has attracted considerable research attention. Periodontitis patients with uncontrolled diabetes usually develop severe periodontitis and exhibit dramatic alveolar bone resorption.[43]^9 In addition, they showed a high recurrence rate of periodontitis even after standard therapy. In consequence, it is more difficult to treat periodontitis with diabetes. Oxidative stress induced by periodontal pathogens and their metabolites is one of the most important mechanisms of tissue destruction in periodontitis.[44]^10^,[45]^11 High levels of ROS cause oxidative damage of intracellular proteins, lipids and nucleic acids, leading to cell dysfunction and tissue break.[46]^12 Specifically, patients with poorly controlled diabetes (fasting glucose over 110 mg/dL and hemoglobin A1c over 7%) displayed increased ROS production and decreased antioxidant capacity, which aggravated the inflammatory responses in periodontitis and exacerbated alveolar bone absorption.[47]^13^,[48]^14 Moreover, the repair capacity of resident cells such as fibroblast cells and periodontal ligament stem cells (PDLSCs) was reduced under oxidative stress.[49]^15^,[50]^16 Hence, redox modulation of the periodontal microenvironment is critical for inflammation control and tissue repair in periodontitis patients with diabetes. For antioxidant treatment, the drug development based on endogenous antioxidants is considered as a classical strategy. The thioredoxin (TRX) system comprising NADPH, TRX reductase (TrxR), and TRX, is an indispensable endogenous antioxidant defense system in mammalian.[51]^17^,[52]^18 Specifically, TRX1, a cytosolic form of TRX, has attracted much interest. Deletion of TRX1 in mice leads to early embryonic mortality.[53]^19 Additionally, TRX1 plays a protective role in hypertension[54]^20 and atherosclerosis.[55]^21 In relation to diabetic-related diseases, TRX activity in the aorta was found to be decreased in diabetic rats, and hyperglycemia inhibited exercise-induced rise of TRX activity.[56]^22 Conversely, the development of diabetic nephropathy was inhibited in TRX1 transgenic mice.[57]^23 What’s more, TRX1 has been shown to alleviate Alzheimer’s-like pathological changes in diabetic encephalopathy model.[58]^24 In conclusion, TRX1 has an important protective effect on oxidative stress-related diseases, and can be considered a potential therapeutic target for diabetic-related injuries. However, whether TRX1 plays a protective role in periodontitis with diabetes remains unclear. Regarding periodontal tissue repair, restoring alveolar bone is crucial as it provides essential support for dental structures. Promoting the osteogenesis of periodontal tissue cells is consequently important. Interestingly, studies found that the expression of TRX1 in human mesenchymal stem cells was up-regulated during osteogenic induction.[59]^25 In addition, inhibiting TRX-interacting protein (TXNIP), an endogenous inhibitor of TRX1 has been reported to retrieve Wnt/β-catenin signaling in endothelial cells treated with high glucose.[60]^26 Wnt pathway is a major signal pathway related to osteogenic differentiation. Therefore, TRX1 may also play a role in osteogenesis. Based on the aforementioned reasons, the aim of this study was to investigate the role of TRX1 in the modulation of osteogenesis. Furthermore, the secondary objective was to explore the therapeutic effect of TRX1 on periodontitis with diabetes and evaluate whether TRX1 could be a therapeutic target for oxidative stress-associated refractory periodontitis. Results Diabetes exacerbates periodontitis and reduces the expression of TRX1 in periodontal tissues In order to evaluate the influence of diabetes on the progression of periodontitis, we constructed periodontitis mice model with or without diabetes. The modeling diagram were presented in [61]Figure S1A. The body weight and the blood glucose levels indicated that the diabetes model was developed successfully ([62]Figures S1B and S1C). Micro-CT analysis showed that compared to normal mice, diabetic mice displayed slight alveolar bone absorption, while both P mice and DP mice exhibited obvious alveolar bone loss, especially the DP mice ([63]Figures 1A and 1D). H&E staining also indicated that DP mice showed more alveolar bone loss and epithelium broken ([64]Figures 1B and 1E). Furthermore, osteoclasts activation leads to bone resorption and negatively regulate neonatal bone formation. TRAP staining illustrated that the number of osteoclasts in P mice and DP mice was significantly increased, with more osteoclasts observed in the periodontium of DP mice ([65]Figures 1C and 1F). Thus, diabetes exacerbated the periodontal destruction in mice with periodontitis. Figure 1. [66]Figure 1 [67]Open in a new tab Periodontitis mice with diabetes exhibited more serious periodontal tissues broken than those without diabetes (A) The representative 3D images of mice maxillae in four groups constructed by micro-CT. The straight line extends from CEJ to ABC, indicating alveolar bone loss. (B) H&E staining of maxillae sections showing the periodontal structure. The dotted line extents from CEJ to ABC, indicating alveolar bone absorption. First molar: left; Second molar: right. Scale bar = 200 μm. (C) Trap staining of maxillae sections showed the activated osteoclasts in periodontium. Cells stained red color were osteoclasts. Scale bar = 100 μm. (D and E) The statistical analysis of CEJ-ABC distance of 3D images (D) and H&E staining (E) in each group (n = 5 per group). (F) The number of osteoclasts per slice were quantified (n = 5 per group). All data were represented as the means ± SEM. ∗, P ＜0.5, compared with the control group; #, P ＜0.5, compared with the periodontitis group; &, P ＜0.5, compared with the diabetes group. To explore the involvement of TRX1 in periodontitis with or without diabetes, we performed IHC staining to detect the TRX1 expression in periodontal tissues. The results revealed that the expression of TRX1 was reduced in P, D, and DP mice, with DP mice showing the greatest decline of TRX1 expression ([68]Figures 2A and 2B). Western blot analysis further confirmed lower levels of TRX1 protein in the periodontal tissues of P and D mice, and the lowest TRX1 protein levels in DP mice ([69]Figures 2C and 2D). Furthermore, DHE staining showed that the generation of ROS in periodontal cells was increased in DP mice compared to P mice ([70]Figures 2E and 2F). These findings suggested that periodontitis with diabetes exhibited significant decrease in TRX1 expression and enhanced oxidative stress. Figure 2. [71]Figure 2 [72]Open in a new tab The expression of TRX1 was dramatically reduced in periodontitis mice with diabetes (A) Immunohistochemistry staining of TRX1 in four groups. Positive cells were stained brown in cytoplasm and nucleus. (B) The ratio of TRX1 positive cells in periodontium was analyzed (n = 5 per group). (C) The expression of TRX1 in periodontal tissues of mice in four groups were analyzed by western blot. (D) The statistical analysis of TRX1 expression. (E) DHE staining of periodontal cells. (F) The statistical analysis of DHE fluorescence intensity (n = 5 per group). All data were represented as the means ± SEM. ∗, P ＜0.5, compared with the control group; #, P ＜0.5, compared with the periodontitis group; &, P ＜0.5, compared with the diabetes group. The osteogenic differentiation of PDLSCs is impaired in inflammatory and diabetic conditions PDLSCs have promising potential in regenerating bone, cementum, and periodontal ligament-like tissues. It has been discovered that the damage to PDLSCs is closely linked to the development of periodontitis.[73]^27 In a miniature pig periodontitis model, Ding et al. successfully regenerated alveolar bone and cured periodontitis using allogeneic PDLSCs.[74]^28 To investigate the impact of periodontal inflammation and diabetic environment on PDLSCs, we conducted experiments where PDLSCs were stimulated with lipopolysaccharide (LPS) and advanced glycation end-products (AGEs). [75]Figures 3A–3D illustrated that treatment with LPS or AGEs reduced the protein level of TRX1, while increasing ROS production. The combination of LPS and AGEs led to the greatest decrease of TRX1 and the highest increase of ROS. Furthermore, ALP activity test and alizarin red staining showed that treatment with LPS or AGEs inhibited the mineralization of PDLSCs, particularly when stimulated with both LPS and AGEs ([76]Figures 3D and 3E). Western blot analysis further confirmed the down-regulation of osteogenic-related proteins including COL1 and OPN. Interestingly, the expression of TRX1 in osteogenic-induced PDLSCs increased over time ([77]Figures 3F and 3G). These results indicated that the osteogenesis capacity of PDLSCs was significantly suppressed under inflammatory and diabetic conditions. Moreover, TRX1 was implicated in osteogenic differentiation of PDLSCs. Figure 3. [78]Figure 3 [79]Open in a new tab Treatment with LPS and AGEs reduced the expression of TRX1 and inhibited the osteogenic differentiation of PDLSCs (A) Western bolt analysis of TRX1 in LPS and AGEs stimulated PDLSCs. (B) The statistical analysis of TRX1 expression (n = 3 per group). (C) ROS detection with DHE staining. Scale bar = 100 μm. (D) The statistical analysis of mean fluorescence intensity (MFI). (E) Alizarin red staining of osteogenic-induced PDLSCs treated with LPS and AGEs on day 14. Scale bar = 100 μm. (F) Quantitative analysis of alizarin red staining. (G) The analysis of ALP activity. (H) Western bolt analysis of TRX1 and osteogenesis-related proteins (COL1, OPN) in osteogenic-induced PDLSCs on day 7 and day 14. (I) The statistical analysis of western blot results (n = 3 per group). All data were represented as the means ± SEM. ∗, P ＜0.5, compared with the OM group; #, P ＜0.5, compared with the OM + LPS group; &, P ＜0.5, compared with the OM + AGEs group. TRX1 modulates the osteogenic differentiation of PDLSCs To clarify the role of TRX1 in the osteogenic differentiation of PDLSCs, we utilized LV-h-TRX1 shRNA vector to knock down the TRX1 gene in PDLSCs. The expression of TRX1 was found to be reduced approximately 3-fold, as detected by qRT-PCR and western blot ([80]Figures 4A–C). As anticipated, the knockdown of TRX1 led to an increase of ROS production ([81]Figures 4D and 4E), a decrease of ALP activity ([82]Figure 4F) and a reduction of mineralization capacity in PDLSCs ([83]Figures 4G and 4H). Encouragingly, administration of rhTRX1 to TRX1 knockdown PDLSCs reversed the increase of ROS and the decrease of mineralization. Correspondingly, the expression of osteogenic-related proteins (OPN and COL1), as well as TRX1 itself, was inhibited in TRX1 knockdown PDLSCs during osteoinduction. However, the inhibition was partially rescued by treatment with rhTRX1 ([84]Figures 4I and 4J). These findings strongly suggested that TRX1 plays a critical role in maintaining redox balance and regulating osteogenesis. Figure 4. [85]Figure 4 [86]Open in a new tab TRX1 knock-down promoted ROS production and impeded the osteogenic differentiation of PDLSCs (A) qRT-PCR analysis of TRX1 in PDLSCs. (B) Western bolt analysis of TRX1 in PDLSCs. (C) The statistical analysis of TRX1 expression (n = 3 per group). (D) ROS accumulation in TRX1 knockdown PDLSCs treated with or without rhTRX1 were measured by DHE probe. Scale bar = 100 μm. (E) The statistic of MFI. (F) The analysis of ALP activity. (G). Alizarin red staining of osteogenic-induced PDLSCs on day 14. Scale bar = 100 μm. (H) Quantitative analysis of alizarin red staining. (I) Western bolt analysis of osteogenic differentiation related proteins (OPN, COL1) and TRX1 in osteogenic-induced PDLSCs treated with or without rhTRX1. (J) The statistical analysis of Western blot results (n = 3 per group). All data were represented as the means ± SEM. ∗, P ＜0.5, compared with the control group; #, P ＜0.5, compared with the LV-NC group; &, P ＜0.5, compared with the LV-sh-TRX1 group. TRX1 facilitates the osteogenesis of PDLSCs by activating Wnt/β-catenin signaling To explore the mechanism underlying the regulatory function of TRX1 in osteogenic differentiation, we performed RNA-seq comparing TRX1 knockdown PDLSCs and normal PDLSCs under osteogenic induction. We found that approximately 614 genes were differentially expressed in TRX1 knockdown PDLSCs compared to normal PDLSCs ([87]Figure 5A). KEGG signal pathway enrichment analysis revealed that the differentially expressed genes (DEG) were mainly enriched in the hippo signal pathway, Wnt signal pathway, TGF-β signal pathway, as well as cell pathways related to apoptosis, adhesion and cell cycle ([88]Figure 5B). Among these pathways, Wnt signaling is known to paly a crucial role in osteogenesis. Next, we investigated the involvement of Wnt/β-catenin signal pathway in the regulation of osteogenic differentiation by TRX1 in PDLSCs. Western blot analysis indicated that TRX1 knockdown in PDLSCs significantly downregulated the expression of p-GSK3β and β-catenin during osteogenic induction. And rhTRX1 administration rescued the expression of p-GSK3β and β-catenin ([89]Figures 5C and 5D). Figure 5. [90]Figure 5 [91]Open in a new tab TRX1 regulated the osteogenic differentiation of PDLSCs through Wnt/β-catenin signal pathway (A) Heatmap of DEGs in PDLSCs transfected with LV-TRX1 shRNA (kd, n = 3) vs. PDLSCs transfected with LV-NC (con, n = 3) under osteoinduction. (B) The KEGG dot diagram showing the changed signaling pathway. (C) Western bolt analysis of p-GSK3β, GSK3β and β-catenin in osteogenic-induced PDLSCs treated with or without rhTRX1. (D) The statistical analysis of western blot results (n = 3 per group). All data were represented as the means ± SEM. ∗, P ＜0.5, compared with the control group; #, P ＜0.5, compared with the LV-NC group; &, P ＜0.5, compared with the LV-sh-TRX1 group. To further elucidate the TRX1-Wnt/β-catenin mediated osteogenesis, we used XAV939, a selective inhibitor of Wnt/β-catenin signal pathway. The results demonstrated that XAV939 diminished the pro-osteogenesis effect of TRX1. Specifically, the mineralization and the ALP levels of TRX1 knockdown PDLSCs treated with rhTRX1 were significantly decreased ([92]Figures 6A and 6B), and the expression of osteogenesis related proteins (OPN, RUNX2, and OSX) was also reduced ([93]Figures 6C and 6D). These findings revealed that TRX1 promoted the osteogenic differentiation of PDLSCs through activating Wnt/β-catenin signal pathway. Figure 6. [94]Figure 6 [95]Open in a new tab The selective inhibition of Wnt/β-catenin signal pathway abolished the pro-osteogenesis effect of rhTRX1 administration (A) The representative picture of alizarin red staining on day 14. (B) The representative picture of ALP staining on day 7. (C) Western bolt analysis of osteogenic differentiation related proteins (OPN, RUNX2, OSX) and the key proteins of Wnt/β-catenin signaling (β-catenin, p-GSK3β) (n = 3 per group). (D) The statistical analysis of western blot results. All data were represented as the means ± SEM. ∗, P ＜0.5, compared with the LV-NC group; #, P ＜0.5, compared with the LV-sh-TRX1 group; &, P ＜0.5, compared with the LV-sh-TRX1+rhTRX1 group. Administration of rhTRX1 reduces ROS accumulation and activates Wnt/β-catenin signaling in periodontal tissues Given that rhTRX1 counteracted ROS generation, activated Wnt/β-catenin signal pathway and promoted the osteogenic differentiation of TRX1 knock-down PDLSCs, we further hypothesized that rhTRX1 have therapeutic effect in periodontitis. To confirm this, we injected rhTRX1 or an equal volume of vehicle into the periodontal tissues of P and DP mice. As shown in [96]Figures 7A and 7B, rhTRX1 treatment significantly reduced ROS generation in P and DP mice, which is beneficial for tissue repair. Interestingly, although the level of ROS in the DP-vehicle group was higher than that in the P-vehicle group, there was no statistical difference between the P-rhTRX1 group and DP-rhTRX1 group. Moreover, the expression of TRX1, p-GSK3β, and β-catenin was noticeably up-regulated in periodontal tissues following rhTRX1 injection, while the vehicle-treated groups displayed low expression of these proteins, especially the DP-vehicle group ([97]Figures 7C and 7D). Additionally, IF staining analysis indicated that rhTRX1 injection enhanced the expression of osteogenic-related proteins (OPN, COL1) in periodontal tissues ([98]Figures 7E and 7F). These results demonstrated that administration of rhTRX1 alleviated oxidative stress in the periodontal microenvironment, activated Wnt/β-catenin signaling and promoted osteogenesis. Figure 7. [99]Figure 7 [100]Open in a new tab Local injection of rhTRX1 scavenged excessive ROS and enhanced Wnt/β-catenin signaling (A) ROS level of periodontal cells in four groups were examined by DHE staining. (B) The statistical analysis of DHE fluorescence intensity. (C) The protein level of TRX1, p-GSK3β, GSK3β and β-catenin in periodontal tissues of mice treated with vehicle or rhTRX1 were analyzed by western blot. (D) The statistical analysis of western blot results. (E) Immunofluorescence staining of osteogenic-related proteins (ALP, OPN). Scale bar = 100 μm. (F) MFI quantification of OPN and COL1 in the periodontal tissues. n = 5 per group. AB, alveolar bone; PDL, periodontal ligament; R, root. All data were represented as the means ± SEM. ∗, P ＜0.5, compared with the P-vehicle group; #, P ＜0.5, compared with the DP-vehicle group. Treatment with rhTRX1 promotes alveolar bone repair in periodontitis mice with diabetes The micro-CT images of 3D reconstruction showed that the CEJ-ABC distance was lower in rhTRX1-treated groups compared to vehicle-treated groups ([101]Figures 8A and 8D). Particularly, rhTRX1 injection significantly improved bone loss in DP mice. H&E staining analysis revealed better integrity of gingival epithelium, fewer infiltrating inflammatory cells, and more alveolar bone in rhTRX1-treated groups than those in vehicle-treated groups ([102]Figures 8B and 8E). Moreover, TRAP staining results suggested that the activated osteoclasts in periodontium was less in rhTRX1-treated groups than that in vehicle-treated groups. Notably, compared to PBS injection, rhTRX1 administration significantly reduced the number of osteoclasts and enhanced bone restoration in DP mice ([103]Figures 8C and 8F). Overall, these findings indicated that local injection of rhTRX1 facilitated periodontal tissues repair in both P mice and DP mice, with particularly desirable therapeutic effects observed in DP mice. Figure 8. [104]Figure 8 [105]Open in a new tab Local injection of rhTRX1 significantly restored the periodontal tissues of DP mice (A) The representative 3D images of mice maxillae in vehicle or rhTRX1-treated groups constructed by micro-CT. The straight line extends from CEJ to ABC, indicating periodontal bone loss. (B) H&E staining of maxillae sections showing the periodontal structure. The dotted line extends from CEJ to ABC, indicating periodontal bone loss. Second molar: left, Third molar: right. Scale bar = 200 μm. (C) Trap staining of maxillae sections showing the activated osteoclasts in periodontium. Cells stained red color were osteoclasts. Scale bar = 100 μm. (D and E) The statistical analysis of CEJ-ABC distance of 3D images (D) and H&E staining (E) in each group (n = 5 per group). (F) The number of osteoclasts in each group were quantified (n = 5 per group). All data were represented as the means ± SEM. ∗, P ＜0.5, compared with the P-vehicle group; #, P ＜0.5, compared with the DP-vehicle group. Discussion The current study investigated the role of TRX1 in PDLSCs osteogenesis and the therapeutic effect of TRX1 in periodontitis with diabetes. The results evidenced that TRX1 promoted the osteogenic differentiation of PDLSCs by activating Wnt/β-catenin signal pathway and rhTRX1 injection significantly improved the alveolar bone restoration in periodontitis mice with diabetes. Both diabetes and periodontitis are highly prevalent chronic diseases. In recent years, comorbidity has been demonstrated to describe their relationship for that periodontitis probably coexisting but independently with diabetes and they affect each other.[106]^13^,[107]^29^,[108]^30 Researchers found that oxidative stress is a key cause of tissue destruction and prevents tissue repair in periodontitis.[109]^13^,[110]^31 Moreover, periodontitis in patients with diabetes exhibited enhanced oxidative stress in periodontal tissues, thus persistently exacerbating inflammatory responses and alveolar destruction.[111]^32 In line with these, our study detected higher ROS levels in the periodontal cells of periodontitis mice with diabetes than in those without diabetes. More alveolar bone resorption was also observed in periodontitis mice developed with diabetes. Antioxidant therapy strategies for oxidative stress-related diseases have been extensively explored, however, few therapeutic agents have been developed in the clinic. Moderate doses of ROS are known to be biologically necessary for cell signaling. It is consequently important to understand the exact pathogenesis behind oxidative damage in disease progression and the mechanism of action of antioxidants so that we can select appropriate agents to interfere. TRX1 is a critical antioxidant protein ubiquitously expressed in various cells. In the early stages of inflammation, excess ROS induces TRX1 production to defend against oxidative stress by directly clear ROS as well as activates the endogenous antioxidant enzymes including superoxide dismutase(SOD), glutathione peroxidase(GPX), and catalase(CAT).[112]^33 At the same time, with chronic pro-inflammatory conditions, it seems different. Several studies have confirmed reduced antioxidant capacity in individuals with periodontitis as assessed by SOD, CAT, and GPX activity in plasma, serum, saliva, and gingival fluid in participants.[113]^34^,[114]^35^,[115]^36^,[116]^37 While in another report, the mRNA levels of peroxiredoxin 2 and SOD2 in the gingivae were slightly affected by periodontitis, but were significantly induced in periodontitis with uncontrolled diabetes.[117]^38 Samuel found the protein level of TRX1 in the myocardium of diabetic sham rats was evidently reduced compared to their controls.[118]^39 In agreement, we observed lower expression of TRX1 in periodontal tissues of periodontitis mice and diabetic mice. What’s more, periodontitis mice with diabetes exhibited the greatest reduction of TRX1 and the most severe alveolar bone resorption. In vitro experiments also showed a significantly decrease of TRX1 expression accompanied by an increase of ROS generation in PDLSCs stimulated with LPS plus AGEs. The osteogenic differentiation of PDLSCs was correspondingly inhibited. It is interesting to note that diabetic mice, similar to periodontitis mice, exhibited increased ROS generation and decreased TRX1 expression in the periodontal tissues. However, diabetic mice did not show significant destruction of the periodontium. We consider that it is the imbalance of the host immune system and bacterial infection ultimately induces periodontium destruction. In diabetic mice without periodontitis, the periodontal tissue was able to maintained homeostasis despite the high level of ROS and low expression of TRX1. Similarly, mice with MetS induced by high-fat diet showed increased alveolar bone loss when subjected to ligature-induced periodontitis, but mice with MetS alone did not appear noticeably bone destruction.[119]^40 Recent emerging evidence has indicated that microRNAs (miRNAs) was involved in periodontal tissue homeostasis and may contribute to increased oxidative stress probably by dysregulating the post-transcription of proteins implicated in ROS generation as well as antioxidant defense.[120]^41^,[121]^42 Specifically, miRNA 7a-5p and miRNA 21-3p were found to be significantly increased in the gingival crevicular fluid of individuals with periodontitis.[122]^42 It is possible that the decrease of TRX1 and the increase of ROS closely related to miRNAs. Further exploration is needed to fully understand the relationship between miRNAs, ROS and periodontal tissue homeostasis. A preliminary study showed that TRX1 mediated metformin-induced osteogenesis and increase of the femoral bone density in osteoporotic mice.[123]^43 Interestingly, we found that TRX1 was up-regulated during the osteoinduction of PDLSCs. Further experiments revealed that TRX1 contributed to osteogenesis regulation, as TRX1 knockdown dramatically reduced the osteogenic differentiation of PDLSCs, while adding rhTRX1 rescued the mineralization. Few researches have investigated the effect of TRX1 on osteogenesis regulation, and the regulation mechanism remains unclear. To address this, we performed RNA-seq analysis of TRX1 knockdown PDLSCs and the normal PDLSCs under osteogenic induction. KEGG analysis of the DEGs revealed changes in Wnt signaling. The Wnt/β-catenin pathway is a canonical pathway mediated osteogenic differentiation.[124]^44 Previous studies have suggested that TRX1 may interact with Wnt/β-catenin signal pathway. Shen et al. found that endothelial-specific knockdown of TXNIP significantly rescued β-catenin activity in endothelial cells treated with high glucose.[125]^26 Moreover, treatment with PX12, a potent inhibitor of TRX1, led to a decrease of β-catenin level in pulmonary artery smooth muscle cells.[126]^45 Conversely, overexpression of TRX1 in human adipose derived mesenchymal stem cells improved β-catenin/TCF promoter activities.[127]^46 A similar conclusion was drawn in our study that knockdown of TRX1 suppressed the Wnt/β-catenin pathway in PDLSCs during osteoinduction. In addition, interference of TRX1 expression increased ROS generation, while the addition of rhTRX1 restored β-catenin signaling and reduced ROS accumulation. These findings indicated that TRX1 can enhance osteogenic differentiation by activating Wnt/β-catenin pathway. As a form of exogenous TRX1, rhTRX1 has been proven to exert therapeutic effect in oxidative stress related disease such as atherosclerosis[128]^20 and hypertention[129]^21 by counteracting oxidative stress. In this research, local injection of rhTRX1 was certified to reduce ROS production in periodontal tissues of periodontitis mice, improve β-catenin activation and contribute to the repair of gingival tissue as well as alveolar bone. Importantly, rhTRX1 treatment significantly lowered the CEJ-ABC distance and restored gingival epithelial of periodontitis mice with diabetes. In addition, TRX1 may play a role in preventing the degradation of periodontal ligaments under conditions of periodontitis and diabetes. This hypothesis was supported by the observation that the expression of COL1, one of the main structural components of ligaments, was strong in the periodontal ligaments of mice treated with rhTRX1, while it was weak in the vehicle-treated mice. Besides its antioxidant property, TRX1 also possesses anti-inflammatory, anti-apoptotic and pro-angiogenesis properties. El Hadri et al. reported that rhTRX1 induced the polarization of anti-inflammatory macrophages (M2) and reduced the number of pro-inflammatory macrophages (M1) in vivo.[130]^47 Ma et al. found that rhTRX1 treatment inhibited neuronal cell apoptosis in mice with ischemic cerebral infarction.[131]^48 Additionally, rhTRX mediated the repair of myocardial damage in diabetic mice by promoting angiogenesis.[132]^49 Bone tissue repair is a complex process involving multiple cells including osteoblasts, osteoclasts, endothelial cells and leukocyte. The proportion of MI/M2 in periodontal tissues of periodontitis mice with diabetes was increased.[133]^50 Moreover, Isola G et al. reported that the level of CD33^+KDR^+ endothelial progenitor cells was negatively correlated with the extent of periodontitis.[134]^51 Besides, the in vivo experiment of this study showed that rhTRX1 administration reduced the number of osteoclasts. Thus, in addition to promoting osteogenesis, rhTRX1 may restore periodontal tissue by acting on macrophages to reduce proinflammatory factors, acting on endothelial cells to promote angiogenesis, and alleviating osteoclastogenesis. Further studies are needed to elucidate the speculation. In summary, the present study highlighted that TRX1 has the ability to promote osteogenesis by regulating Wnt/β-catenin pathway. Furthermore, rhTRX1 exert a promising therapeutic effect in periodontitis with diabetes by counteracting ROS generation and activating β-catenin signaling. These findings suggested that TRX1 could be a potential therapeutic target for refractory periodontitis associated with oxidative stress. Limitations of the study The mechanisms by which TRX1 regulate Wnt/β-catenin signaling and whether ROS are involved in the process are still unknown. Further investigations are needed to elucidate the underlying mechanisms. In addition, as rhTRX1 was directly injected into periodontal tissues, multiple doses were necessary to achieve a good therapeutic effect. To reduce the frequency and dose of rhTRX1 administration, a sustained-release delivery system should be considered. Another limitation is that the inflammation in periodontal tissue is not thoroughly analyzed in the in vivo study. It would be beneficial to assess the expression of inflammatory factors as well as the infiltration of inflammatory cells, which would provide a more comprehensive understanding of the effects of TRX1 on periodontal tissue inflammation. STAR★Methods Key resources table REAGENT or RESOURCE SOURCE IDENTIFIER Antibodies __________________________________________________________________ anti-osteopontin/OPN rabbit primary antibody Zen-Bioscience (Chengdu, China) Cat#340690 anti-collagen1/COL1 rabbit primary antibody Zen-Bioscience (Chengdu, China) Cat#501352 Alexa Fluor 488, Goat Anti-Rabbit IgG EMAR(Beijing, China) EM35140-02 Anti-Collagen1 antibody abcam ab233080 anti-Thioredoxin-1/TRX1 antibody abcam ab86255; RRID: [135]AB_1925501 beta Catenin Rabbit mAb Zen-Bioscience (Chengdu, China) [136]R22820 GSK3β Rabbit pAb Zen-Bioscience (Chengdu, China) Cat#340449 Phospho-GSK3β (Ser9) Rabbit pAb Zen-Bioscience (Chengdu, China) Cat#310010 GAPDH Mouse mAb Cell Signaling Technology Cat#97166; RRID: [137]AB_2756824 RUNX2 Polyclonal antibody Proteintech Cat#20700-1-AP; RRID: [138]AB_2722783 SP7 Rabbit Polyclonal Antibody OriGene Technologies Cat#TA381910 __________________________________________________________________ Bacterial and virus strains __________________________________________________________________ Human lentivirus-sh-TRX1 vectors Hanbio Biotechnology Co., Ltd. (Shanghai, China) N/A __________________________________________________________________ Chemicals, peptides, and recombinant proteins __________________________________________________________________ streptozotocin（STZ） Sigma-Aldrich S0130 recombinant human TRX1（rhTRX1） R&D Cat#1970-TX-500 Lipopolysaccharides (Escherichia coli O111:B4) Sigma-Aldrich L5293 advanced glycation end products(AGEs) Bioss(Beijing, China) bs-1158P β-glycerophosphate Sigma-Aldrich G9422 ascorbic acid Sigma-Aldrich AX1775 dexamethasone Sigma-Aldrich D4902 __________________________________________________________________ Critical commercial assays __________________________________________________________________ Tartrate-resistant acid phosphatase Stain Kit (TRAP) kit Jiancheng Technology (Nanjing, China) D023-1-1 Alkaline phosphatase assay (ALP) kit Jiancheng Technology (Nanjing, China) A059-2-2 Dihydroethidium (DHE) Applygen(Beijing, China) C1300-2 Hoechst 33342 Thermo Scientific Cat#62249 __________________________________________________________________ Deposited data __________________________________________________________________ Raw and analyzed data used for RNA sequencing analysis this paper [139]GSE235930 Experimental Models: cells Periodontal ligament stem cells This paper N/A __________________________________________________________________ Experimental models: Organisms/strains __________________________________________________________________ Mouse: C57BL/6 GemPharmatech Co., Ltd (Nanjing, China) N000013 __________________________________________________________________ Oligonucleotides __________________________________________________________________ Primers for TRX1: Forward: GGGATGTTGGCGATGCA Reverse: CCAGCTACTTGAGGTCCATCTTC This paper N/A Primers for GAPDH: Forward: GGGAAACTGTGGCGTGAT Reverse: GAGTGGGTGTCGCTGTTGA This paper N/A __________________________________________________________________ Software and algorithms __________________________________________________________________ VGStudio MAX 1.2.1 Volume Graphics N/A ImageJ Fiji National Institutes of Health [140]https://imagej.net/software/fiji/ Prism 7.0 GraphPad N/A "edgeR" packages Bioconductor [141]http://bioconductor.org/packages/release/bioc/html/edgeR.html "heatmap" package The Comprehensive R Archive Network (CRAN) [142]https://cran.r-project.org/ "clusterprofile" package Bioconductor [143]http://bioconductor.org/packages/release/bioc/html/clusterProfiler .html __________________________________________________________________ Other __________________________________________________________________ High fat diet containing 60% lard Research Diets, Inc D12492 [144]Open in a new tab Resource availability Lead contact Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Zhi Song ([145]songzh@mail.sysu.edu.cn). Materials availability This study did not generate new unique reagents. Experimental model and study participant details Animal models Forty eight five-week-old male C57BL/6 mice were purchased from GemPharmatech Co., Ltd (Nanjing, China). All experiments were completed in Laboratory Animal Center of Sun Yat-sen University and all procedures met the guidelines of the Animal Care and Use Committee. Diabetic mice model was generated with high-fat diet and streptozotocin (STZ) injection. Briefly, a high-fat diet containing 60% lard (Research Diets, New Brunswick, NJ, USA) was given for 12 weeks, and followed by daily intraperitoneal injection with small doses of STZ (30 mg/kg; Sigma-Aldrich, St. Louis, MO, USA) for 5 consecutive days. Mice with random blood glucose levels above 16.7 mmol/L were identified as diabetic mice. Experimental periodontitis model were established by ligature method. Briefly, A 5-0 silk ligature was tied around the maxillary second molar and kept for 10 days. Primary cell culture Impacted wisdom teeth without carious and inflammation were collected from aldult patients informed consent at Guanghua School of Stomatology, Sun Yat-sen University, Guangzhou, China. PDLSCs were isolated by tissue explant-collagenase digestion method and cultured in Dulbecco’smodified Eagle’s medium (DMEM, Gibco, USA) supplemented with 10% fetalbovine serum (FBS, Gibco, USA) and 1% penicillin-streptomycin (Gibco, USA). Method details Mice treatment To investigate the influence of diabetes on periodontitis, the mice were randomly divided into four groups: (1) normal mice (C, n = 5); (2) periodontitis model without diabetes (P, n = 5); (3) diabetic mice without periodontitis (D, n = 5); (4) periodontitis model with diabetes (DP, n = 5). To investigate the effectiveness of rhTRX1 on periodontitis, the mice were randomly assigned into four groups: (1) periodontitis mice treated by vehicle (P + vehicle, n = 5); (2) periodontitis mice treated by 40 ug rhTRX1 (P + rhTRX1, n = 5); (3) periodontitis mice with diabetes treated by vehicle (DP + vehicle, n = 5); (4) periodontitis mice with diabetes treated by 40ug rhTRX1 (DP + rhTRX1, n = 5). Total 5 μL of rhTRX1(8 μg/μL, R&D, MN, USA) or vehicle was injected into the periodontal lesions in the palatal side of the second molar teeth once a week after the ligature was removed. Four weeks later, the maxillae of mice were harvested. Micro-CT analysis Formalin-fixed maxillae were scanned with a high-resolution micro-CT scanner (Scanco Medical AG, Bassersdorf, Switzerland). The scan parameters were set as 70 kV, 114 mA and 10 μm increments. Three-dimensional (3D) images were reconstructed using VGStudio MAX 1.2.1 (Volume Graphics, Heidelberg, Germany). The distance between the cementoenamel junction (CEJ) and the alveolar bone crest (ABC) was measured at six predetermined sites on both the buccal and palatal sides to assess alveolar bone loss according to previous studies.[146]^48 Histological staining and immunohistochemistry (IHC) Formalin-fixed maxillae were decalcified in 10% EDTA for four weeks, dehydrated in ascending grade ethanol and embedded in paraffin. Sections with 4-μm thickness were stained using H&E solution and TRAP kit (Jiancheng Technology, Nanjing, China). Number of osteoclasts were counted and CEJ-ABC distance were measured to assess alveolar bone absorption. IHC staining was conducted to detect the expression of TRX1 in periodontal tissues. The anti-TRX1 primary antibody was diluted in 1:200. Immunofluorescence (IF) staining Paraffin maxilla sections with 4-μm thickness were treated for antigen retrieval and blocked with goat serum (Bosterbio, Wuhan, China), subsequently incubated with diluted anti-osteopontin/OPN rabbit primary antibody and anti-collagen1/COL1 primary antibody (both from Zen-Bioscience, Chengdu, China) for 12 h at 4°C. After wash with PBS, sections were incubated with diluted Alexa Fluor 488 goat anti-rabbit antibody (1:500, EMAR, Beijing, China) for 1 h and 4′,6-diamidino-2-phenylindole (DAPI) for 5 min in dark at room temperature. IF signals were recorded by confocal laser scanning fluorescence microscopy (Zeiss, Oberkochen, Germany) and analyzed using ImageJ Fiji software. Cell treatment AGEs are crucial molecules involved in the development of various diabetic related diseases. Their accumulation contributes to diabetes-related injury by binding to receptor for advanced glycationend products (RAGE), which further activate the production of ROS and the regeneration of inflammatory cytokines.[147]^52^,[148]^53 In the in vitro experiment related to diabetes, AGEs can be used to simulate diabetic conditions.[149]^54 We used 1 μg/ml LPS (Escherichia coli O111:B4; L5293, Sigma) or 50 μg/ml AGEs (Bioss, Beijing, China) to stimulate PDLSCs. PDLSCs were divided into four groups: control group (Control), LPS group (LPS), AGEs group (AGEs) as well as LPS plus AGEs group (LPS + AGEs). After 72 h’ incubation, cells were collected to test the expression of TRX1 and ROS accumulation. For osteogenic induction, cells were cultured in osteogenic-inductive medium (OM) consisting Dulbecco’smodified Eagle’s medium (DMEM, Gibco, USA),10% fetalbovine serum (FBS, Gibco，USA)，1% penicillin-streptomycin (Gibco, USA), 10 mM β-glycerophosphate (Sigma-Aldrich, MO, USA), 50 μM ascorbic acid (Sigma-Aldrich, MO, USA), and 100 nM dexamethasone(Sigma-Aldrich, MO, USA). 7 days later, the alkaline phosphatase (ALP) activity was detected with an ALP assy kit (Jiancheng, Nanjing, China). Briefly, the cells were lysed using 0.5% Triton. The cell lysate was then incubated with the detection reagent for 15 min at 37°C. The optical density of the cell lysate-detection reagent mixture was measured at 520 nm. And the ALP staining was conducted using an ALP staining kit (Solarbio, Beijing, China). 14 days later, alizarin red staining was performed to access the influence of LPS or AGEs on cells mineralization capacity. And Western blot analysis was conducted to test the expression of osteogenic related proteins. Lentivirus transfection of PDLSCs Human lentivirus-sh-TRX1 vectors (LV-sh-TRX1) and negative control vectors (LV-NC) expressing the ZsGreen protein were purchased from Hanbio Biotechnology Co., Ltd. (Shanghai, China). PDLSCs at passage 3 were seeded into six-well culture dishes at a density of 4×10^5 cells per well. When the cells were about 40% confluence, the lentivirus vectors were added into the medium. Polybrene (4 μg/ml; Hanbio, Shanghai, China) was used to improve tansfection efficiency. After 48 h’ incubation, the infected cells were selected with 2 μg/ml puromycin (Hanbio, Shanghai, China). We performed qRT-PCR and Western blot to analyze the expression of TRX1. Quantitative real-time PCR (qRT-PCR) Total RNAs were extracted from cells using RNA quick purification kit (Esunbio, Shanghai, China) and synthesized into complementary DNA (cDNA) using PrimeScript RT Master Mix (TaKaRa, Kyoto, Japan). The expression of TRX1 was tested by qPCR SYBR Green Master Mix (Yeasen, Shanghai, China) on a LightCycler 96 system (Roche, Sweden). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as the housekeeping gene. Western blot Total proteins were extracted from cells and tissues lysates. A total of 20 μg protein were separated by 4–20% Bis-Tris SDS-PAGE (GenScript, Nanjing, China) and transfered onto a polyvinylidene fluoride membrane (PVDF; Millipore, MA, USA). The membrane was blocked with 5% (M/V) skimmed milk for 1 h, incubated overnight with diluted primary antibody at 4°C and then incubated with diluted secondary antibodies (1:2000, Cell Signaling Technology, Danvers, MA, USA) at room temperature for 1 h. The target bands were developed using a chemiluminescence HRP reagent (Millipore, MA, USA) and signals were analyzed using ImageJ software. The primary antibodies used in this study were listed as follows, anti-Thioredoxin-1/TRX1 antibody (1:1000, Abcam), anti-Collagen1/COL1 antibody (1:1000,Abcam), anti-Osteopontin/OPN antibody (1:1000, Zen-Bioscience), anti-β-catenin antibody (1:1000, Zen-Bioscience), anti-p-GSK3β antibody (1:1000, Zen-Bioscience), anti-GSK3β antibody (1:1000, Zen-Bioscience), anti-SP7/OSX antibody (1:1000, OriGene), anti-RUNX2 antibody (1:1000, Proteintech) and anti-GAPDH antiybody (1:1000, Cell Signaling Technology). RNA sequencing analysis PDLSCs transfected with LV-sh-TRX1 and LV-NC were cultured in osteogenic medium (OM) for 7 days. Total RNA in PDLSCs was collected for poly-A transcriptome sequencing analyze (completed by Guangzhou Ruibo Biotechnology Co., Ltd). R (V 3.6.1) and "edgeR" packages were used to standardize the cleaned data. T-test analysis was performed and p < 0.05 as well as fold change (FC) | FC | >2 were used as criteria to screen differentially expressed mRNAs. Cluster analysis was carried out on differential mRNAs, and the "heatmap" package was used to draw the heatmap; The "clusterprofile" package was used for KEGG pathway enrichment analysis. ROS detection We utilize dihydroethidium (DHE; Applygen, Beijing, China) probe to detect ROS production in cells. Periodontal cells were prepared as follows. Maxillary periodontal tissues were cut into small fragments and digested in 3 mg/mL collagenase type I plus 4 mg/mL dispase solutions (both from Sigma-Aldrich, MO, USA) at 37°C for 30 min and further filtered with a nylon filter (mesh size 70 μm) to acquire single cell suspensions. The cells were then incubated with medium containing 10 nM DHE reagent for 30 min at 37°C. After wash with PBS, the fluorescence signal was detected and record by flow cytometry (BD LSRFortessa; Becton Dickinson, San Diego, CA, USA). For ROS measurement in PDLSCs, cells treated with LPS or AGEs were incubated with 10 nM DHE reagent for 30 min at 37°C. Hoechst 33342 (Thermo Scientific) was used to stain nucleus. The images were obtained by a confocal microscope and IF signals were analyzed using ImageJ software. Quantification and statistical analysis All data were analyzed using Prism 7.0 software (GraphPad, San Diego, CA). One-way analysis of variance (ANOVA) with Tukey’s post hoc test was carried out to determine statistical significance between groups and P ＜ 0.05 was considered significant. Error bars represented the standard error of the mean. In vitro experiments were independently performed at least three times. Acknowledgments