Intercellular mitochondrial transfer alleviates pyroptosis in dental pulp damage

Abstract Mitochondrial transfer is emerging as a promising therapeutic strategy for tissue repair, but whether it protects against pulpitis remains unclear. Here, we show that hyperactivated nucleotide‐binding domain and leucine‐rich repeat protein3 (NLRP3) inflammasomes with pyroptotic cell death was present in pulpitis tissues, especially in the odontoblast layer, and mitochondrial oxidative stress (OS) was involved in driving this NLRP3 inflammasome‐induced pathology. Using bone marrow mesenchymal stem cells (BMSCs) as mitochondrial donor cells, we demonstrated that BMSCs could donate their mitochondria to odontoblasts via tunnelling nanotubes (TNTs) and, thus, reduce mitochondrial OS and the consequent NLRP3 inflammasome‐induced pyroptosis in odontoblasts. These protective effects of BMSCs were mostly blocked by inhibitors of the mitochondrial function or TNT formation. In terms of the mechanism of action, TNF‐α secreted from pyroptotic odontoblasts activates NF‐κB signalling in BMSCs via the paracrine pathway, thereby promoting the TNT formation in BMSCs and enhancing mitochondrial transfer efficiency. Inhibitions of NF‐κB signalling and TNF‐α secretion in BMSCs suppressed their mitochondrial donation capacity and TNT formation. Collectively, these findings demonstrated that TNT‐mediated mitochondrial transfer is a potential protective mechanism of BMSCs under stress conditions, suggesting a new therapeutic strategy of mitochondrial transfer for dental pulp repair.

in the odontoblast layer, and mitochondrial oxidative stress (OS) was involved in driving this NLRP3 inflammasome-induced pathology. Using bone marrow mesenchymal stem cells (BMSCs) as mitochondrial donor cells, we demonstrated that BMSCs could donate their mitochondria to odontoblasts via tunnelling nanotubes (TNTs) and, thus, reduce mitochondrial OS and the consequent NLRP3 inflammasome-induced pyroptosis in odontoblasts. These protective effects of BMSCs were mostly blocked by inhibitors of the mitochondrial function or TNT formation. In terms of the mechanism of action, TNF-α secreted from pyroptotic odontoblasts activates NF-κB signalling in BMSCs via the paracrine pathway, thereby promoting the TNT formation in BMSCs and enhancing mitochondrial transfer efficiency. Inhibitions of NF-κB signalling and TNF-α secretion in BMSCs suppressed their mitochondrial donation capacity and TNT formation. Collectively, these findings demonstrated that TNT-mediated mitochondrial transfer is a potential protective mechanism of BMSCs under stress conditions, suggesting a new therapeutic strategy of mitochondrial transfer for dental pulp repair.

| INTRODUCTION
Odontoblasts have been demonstrated to be the main functioning cells for dental pulp tissue repair. 1 These highly specialized cells are arranged along the dentin-pulp interface and connected with each other through tight junctions, thus forming the front line of defence against external stimuli. 2 It is, therefore, crucial to preserve odontoblast functionality and survival in the reparative process after pulp injury. Mitochondria are the central coordinators of energy metabolism, and alterations in their function lead to cellular oxidative stress (OS) and pyroptosis. 3 Our previous study showed that mitochondrial damage in odontoblasts occurred in early pulpitis, leading to overproduction of mitochondrial reactive oxygen species (mtROS) and activation of nucleotide-binding domain and leucine-rich repeat pro-tein3 (NLRP3) inflammasome-induced pyroptosis. 4 These effects eventually result in the structural disruption of the dentin-pulp complex, which exacerbates pulpitis pathology. 5 Together, mitochondrial damage and subsequent pyroptosis of odontoblasts are critical early events responsible for dental pulp injury and the progression of the disease.
Recently, mitochondrial transfer has appeared as an emerging therapeutic strategy as it can restore the bioenergetic requirements of injured cells. 6,7 As active mitochondrial donor cells, BMSCs are considered to be highly promising because of their abundant sources and multipotent differentiation and anti-inflammatory properties. 8 Several studies have discovered that BMSCs can donate their mitochondria to injured cells via TNTs, actin-based intercellular channels that allow for direct communication between adjacent cells. 9 The injured cells thus acquire functional mitochondria and regain their respiratory function and capacity for oxidative metabolism. 10 Moreover, TNTs allow for bidirectional or unidirectional transfer of mitochondria in different cell types and contribute to the treatment of various inflammatory disorders, including neuroinflammation, retinal dysfunction and osteoarthritis. [11][12][13][14] However, whether BMSCs can provide mitochondria to odontoblasts and whether transferred mitochondria can rescue oxidative damage and pyroptosis in odontoblasts are still unclear.
The protective effects of BMSCs are linked to their capacity to migrate to the site of damaged tissue and to reduce oxidative damage, increase mitochondrial number, promote cell survival and inhibit pyroptosis locally. 11,15 Mechanistically, damage to tissues releases intracellular damage-associated molecular patterns (DAMPs) to further amplify the inflammation process. 16 Subsequently, the inflammatory factors released from injured cells could stimulate the mitochondrial biogenesis and TNT formation in BMSCs, thereby enhancing the mitochondrial transfer and rescue ability of BMSCs to injured cells. 17 However, the conclusive mechanistic studies responsible for mitochondrial transfer under inflammatory conditions are still lacking.
In this study, we found that mitochondrial OS promotes NLRP3 inflammasome activation and pyroptosis in odontoblasts, which is the major cause of pulpitis. Furthermore, we confirmed that odontoblasts could receive BMSC-derived mitochondria through TNTs, thereby reducing mitochondrial dysfunction and NLRP3 inflammasomeinduced pyroptosis. In further exploring the underlying mechanisms, we found that NF-κB signalling is involved in the TNT formation. TNF-α secreted by pyroptotic odontoblasts promoted the TNT formation in BMSCs via NF-κB signalling, resulting in an increase in the mitochondrial transfer. This study furthers our understanding of TNTmediated mitochondrial transfer under stress conditions and advances its potential use for new therapeutic approaches in dental pulp repair.

| Human dental pulp tissue samples
All the dental pulp tissues were obtained as discarded biological samples from the School and Hospital of Stomatology, Wuhan University, following the approval of Institutional Ethics Medical Committee of Wuhan University (2019 LUNSHENZI(A48)). Normal pulps from teeth extracted for various reasons served as a control group (n = 31).
Inflamed pulps extracted from patients diagnosed with irreversible pulpitis served as a pulpitis group (n = 65). All patients involved in this study provided a verbal informed consent and were diagnosed according to the published guidelines. In addition, patients receiving antibiotics or anti-inflammatory drugs were excluded.
Mouse bone marrow mesenchymal stem cells (mBMSCs) used in this study were isolated from 4-week-old male C57BL/6 mice. Briefly, bone marrow was collected from femora and tibiae of mice and plated in tissue culture flasks containing DMEM supplemented with 10% FBS and antibiotics. After 2 days, we removed nonadherent cells in the supernatant and cultured adherent mBMSCs and used them after 4-10 passages.

| Bioinformatic and transcriptomic analysis
Total RNA was extracted from the cells using TRIzol (#15596018, Invitrogen) according to the protocol provided by the manufacturer.
A total quantity of 2 μg of RNA per sample from the mDPC6T, LA-mDPC6T and LA-mDPC6T + mBMSC (LA-mDPC6T sorted after coculture with mBMSCs) was used for analysis. The quality of extracted RNA was assessed using the Agilent 2100 Bioanalyzer Filtering and quality controls were applied according to the standard procedure. GO

| Immunoblotting analysis
Immunoblotting analysis was performed according to a previously described protocol. 4 In brief, equal amounts (20 μg) of total protein were loaded onto 10% SDS-polyacrylamide gels and separated by electrophoresis. Then, the proteins were electrophoretic transfer onto a PVDF membrane in tris-glycine buffer. After blocking with 5% skimmed milk in tris-buffered saline supplemented with 0. (LiCor) software. β-actin and Tom20 were used as the internal reference to normalized protein expression levels. The experiment was repeated at least three times.

| Coimmunoprecipitation (CoIP) analysis
CoIP analysis was performed using the Pierce Co-Immunoprecipitation Kit (#26149, Thermo Fisher Scientific) by following the manufacturer's instruction. In brief, antibodies of anti-FUNDC1 (#ab224722, Abcam) and anti-IgG (#ab6709, Abcam) were immobilized by amino link plus coupling resin. Cell lysates were precleared using the control agarose resin and then added to the spin column containing antibody-coupled resin and incubated overnight at 4 C. The bound proteins were eluted by elution buffer, followed by immunoblotting analysis.

| TMRM and MitoSOX assay
Mitochondrial membrane potential was measured using Tetra-

| Measurement of ATP
Cells were collected by lysis buffer provided in the ATP assay kit (#S0026, Beyotime). The sample lysates were then centrifuged, and the supernatants were harvested for the ATP content analysis. Total protein concentration in supernatants was determined using the BCA assay (#P0011, Beyotime) according to the manufacturer's protocol. Before the ATP analysis, supernatants were mixed with ATP detection solution, and then intensity readings were taken with a microplate spectrophotometer (Bio-Rad). The amount of ATP was calculated based on ATP standards and normalized to the concentration of total protein in each sample.

| Immunofluorescence
Tissue sections or cells were fixed in 4% paraformaldehyde for 10 min.

| Fluorescence imaging-based analysis of mitophagy
The cultured cells were incubated with 100 nM Lyso-Tracker Green (#C1047S, Beyotime) for 30 min to label lysosome, followed by the addition of 100 nM MitoAPC (#MT12, Dojindo) and further incubation for 30 min to label mitochondria. After wash with PBS, images were captured under a confocal microscope (LSM880, Carl Zeiss), and colocalization of lysosome and mitochondria was determined using the ZEN 3.5 (Carl Zeiss) software that shows co-staining with green and red fluorescences.
2.14 | Enzyme-linked immunosorbent assay (ELISA) and caspase1 activity assay Caspase1 activity was evaluated using the caspase1 activity assay kit (#C1101, Beyotime Biotechnology) according to the manufacturer's instructions. The substrate was Ac-YVAD-pNA (acetyl-Tyr-Val-Ala-Asp p-nitroanilide). Caspase1 could catalyse Ac-YVAD-pNA into the formazan product p-nitroaniline (pNA). In brief, samples were incubated with Ac-YVAD-pNA at 37 C for 1 h. The absorbance values of pNA at 405 nm were tested using a microplate spectrophotometer (Bio-Rad), and caspase1 activity was determined using the pNA standard solutions prepared in parallel.

| Fluorescence resonance energy transfer (FRET) analysis
Briefly, cells grown on chamber slides were treated as indicated, fixed in FRET images were further analysed using the PIX-FRET plugins for ImageJ. 19

| Mitochondria cytoplasmic fractionation
The cells were collected in the mitochondrial separation reagent provided in the cell mitochondria isolation kit (#89874, Thermo Fisher Scientific) and homogenized with the Dounce homogenizer. The homogenate was centrifuged at 3000g for 10 min. Next, the supernatant was collected and further centrifuged at 11,000g for 10 min, thereby obtaining the pellet as mitochondria-enriched fraction for the subsequent experiments.

| Data analysis
Statistically significant differences were determined by Student's t test (for two groups) and one-way analysis of variance (ANOVA) (for multiple groups), as needed, using the GraphPad Prism 9.0 (GraphPad Software). Pearson's correlation coefficient was used for correlation analyses. p < 0.05 was considered to be statistically significant (*p < 0.05, **p < 0.01 and ***p < 0.001). Error bars represent the mean value ± standard deviation (mean ± SD). Tom20 and Cyto C are commonly used as mitochondrial markers to reflect mitochondrial mass. 20 Immunoblotting analysis revealed that LPS + ATP-induced NLRP3 inflammasome activation and pyroptosis were alleviated by NAC and MitoTEMPO treatment, as reflected by decreased protein expression of NLRP3, cleaved-caspase1 and IL-1β (Appendix Figure 1J,K) and decreased mRNA expression of IL-6 and CXCL10 (Appendix Figure 1L). Recent studies have revealed that the mitochondrial localization of NLRP3 is essential for activating the NLRP3 inflammasome. 21,22 Consistent with our data, immunoblotting of NLRP3 expression in isolated mitochondria was increased by LPS + ATP stimulation ( Figure 1J,K). FRET analysis was performed to (O,P) Whole-cell lysates from mDPC6T and LA-mDPC6T cells were immunoprecipitated with FUNDC1 and IgG antibodies and then immunoblotted with the indicated antibodies, and the ratio between immunoprecipitated LC3B to FUNDC1 was quantified (n = 4). Data are displayed as the mean ± SD. Statistical significance was determined using Student's t test and one-way analysis of variance (*p < 0.05; **p < 0.01; ***p < 0.001).
further assess the mitochondrial localization of NLRP3 (Appendix Figure 2A). FRET analysis revealed that the LPS + ATP treatment facilitated NLRP3 translocation to mitochondria, and this process was abolished by NAC or MitoTEMPO (Appendix Figure 2B,C). These data demonstrate that mitochondrial OS, NLRP3 inflammasome activation and pyroptosis are present in pulpitis and suggest that improving mitochondrial homeostasis and rescuing mitochondrial mass could alleviate these phenomena in pulpitis.

| Mitochondria from mBMSCs are selectively transferred to injured mDPC6T cells
BMSCs are the active mitochondrial donor cells for various cell types. 23 To explore whether BMSCs could donate their mitochondria to mDPC6T cells, mBMSCs were subjected to coculture with healthy mDPC6T cells (control) or injured mDPC6T cells (LPS + ATP stimulation). In brief, CellTrace CFSE-labelled mDPC6T cells (mDPC6T-CFSE) F I G U R E 2 Mitochondria from mBMSCs are transferred to mDPC6T cells and the transfer rate is enhanced under stressed conditions. (A) Schematic representation of the coculture experimental design for the detection of mitochondrial transfer. CellTrace CFSE-labelled mDPC6T cells (mDPC6T-CFSE) were primed with LPS (1 or 5 μg/mL) for 24 h, followed by incubation with ATP (2 mM) for different times (0, 12 or 24 h). MitoAPC-labelled mBMSCs (mBMSC-MitoAPC) were added to mDPC6T cells in the ATP treatment step. The mitochondrial transfer rate was determined by flow cytometry. (B,C) mDPC6T cells were primed with LPS (1 and 5 μg/mL) for 24 h, followed by incubation with both ATP (2 mM) and mBMSCs for 24 h. mDPC6T cells without LPS or ATP treatment were denoted as the control. The mitochondrial transfer rate from mBMSCs to mDPC6T cells was determined by flow cytometry (n = 5). (D,E) mDPC6T cells were primed with LPS (1 μg/mL) for 24 h, followed by incubation with both ATP (2 mM) and mBMSCs for different times (0, 12 or 24 h). The mitochondrial transfer rate from mBMSCs to mDPC6T cells was determined by flow cytometry (n = 5). (F) Immunofluorescence staining of mitochondrial transfer from mBMSC-MitoRFP to mDPC6T-CFSE under stressed or normal conditions. The white arrowheads indicate the mitochondria from mBMSCs (red arrowheads) within mDPC6T cells (green arrowheads). Data are displayed as the mean ± SD. Statistical significance was determined using one-way analysis of variance (**p < 0.01; ***p < 0.001).
were primed with LPS, followed by coculture with MitoAPC-labelled mBMSCs (mBMSC-MitoAPC) in the culture medium supplemented with ATP. Subsequently, the total cells were harvested and subjected to flow cytometry analysis (Figure 2A). After 24 h of coculture, the flow cytometry analysis showed a low mitochondrial transfer rate from mBMSCs to healthy mDPC6T cells, while injured mDPC6T cells received a significantly higher number of mitochondria from mBMSCs ( Figure 2B,C), indicating that mBMSCs prefer to donate their mitochondria to injured mDPC6T cells rather than healthy mDPC6T cells.
Moreover, the mitochondrial transfer rate was similar at low (1 μg/ mL) and high (5 μg/mL) LPS concentrations. We next investigated the effects of varying the coculture time on the mitochondrial transfer rate at low LPS concentration. Flow cytometry analysis showed a time-dependent increase in the mitochondrial transfer rate ( Figure 2D,E). In addition, treatment with ATP alone did not significantly increase mitochondrial transfer from mBMSCs to mDPC6T cells (Appendix Figure 3B,C). We then expressed CellLight Mitochondria-RFP in mBMSCs (mBMSC-RFP) to further confirm intercellular mitochondrial transfer. RFP-labelled mitochondria from mBMSCs were detected in several of the (LA-)mDPC6T-CFSE samples ( Figure 2F).
Collectively, these results indicate that mBMSCs have the capability of selectively donating their mitochondria to injured mDPC6T cells.

| mBMSCs inhibit the mitochondrial OS and pyroptosis in mDPC6T cells through mitochondrial transfer
Next, we investigated whether mBMSCs could rescue injured mDPC6T cells through the transfer of mitochondria. We generated mBMSCs without the mitochondrial function (ρ o mBMSCs) via ethidium bromide (EtBr)-mediated depletion of mtDNA. The depletion of mtDNA in ρ o mBMSCs, as confirmed by immunoblotting analysis (Appendix Figure 3D,E), resulted in reduced ND1 mRNA expression (Appendix Figure 3F) and decreased cellular ATP levels (Appendix Figure 3G). We collected LA-mDPC6T cells in mixed cultures by cell sorting using FACS ( Figure 3A). Immunoblotting analysis showed that Tom20 and Cyto C protein expression in LA-mDPC6T cells was increased after coculture with mBMSCs or ρ o mBMSCs ( Figure 3B,C), indicating the transfer of donor mitochondria to recipient cells. The transfer rate from mBMSCs to mDPC6T cells was similar to that from ρ o mBMSCs to mDPC6T cells (Appendix Figure 3H,I). The flow cytometry analysis revealed that only mBMSCs could restore the mitochondrial function in LA-mDPC6T cells, and ρ o mBMSCs showed no significant impact on the mitochondrial function ( Figure 3D-G). We then examined the effect of mBMSCs and ρ o mBMSCs on the NLRP3 inflammasome pathway and pyroptosis. NLRP3, cleaved-caspase1 and IL-1β protein expression in LA-mDPC6T cells was inhibited after coculture with mBMSCs ( Figure 3H,I). However, ρ o mBMSCs showed an inability to inhibit the expression of these proteins. Similar results were observed for the inhibition of IL-6 and CXCL10 mRNA expression ( Figure 3J), caspase1 activity ( Figure 3K) and IL-1β protein expression ( Figure 3L). These results suggest that mBMSCs can rescue injured mDPC6T cells through the transfer of healthy mitochondria, rather than non-functional mitochondria.
We performed RNA-sequencing analysis to further confirm the protective effect of mBMSCs. The GSEA analysis showed that the NLR pathway was significantly enriched in the LA-mDPC6T group compared to the mDPC6T group, but this result was reversed by coculture of LA-mDPC6T cells with mBMSCs ( Figure 3M). Moreover, by investigating the metabolic gene expression in LA-mDPC6T cells, we found increased expression of genes involved in mitochondrial apoptosis (BNIP3 and UCP2). The expression levels of antioxidase (GPX1 and ALDH1L2) and mitochondrial metabolic genes (PPARGC1A, PIF1 and HSPA1B) were significantly decreased ( Figure 3N). In contrast, coculture of LA-mDPC6T cells with mBMSCs led to an opposite change in the expression of these genes ( Figure 3N). GO enrichment analysis revealed that the ATP metabolic and metabolic processes were markedly enriched in the LA-mDPC6T + mBMSC group compared to the LA-mDPC6T group ( Figure 3O).
Together, these results demonstrate that the mitochondrial transfer of mBMSCs assisted in the rebalancing of mitochondrial redox homeostasis and inhibited NLRP3-mediated pyroptosis in recipient cells.

| TNT mediates mitochondrial transfer from mBMSCs to mDPC6T cells
Mitochondrial transfer through TNTs is one of the mechanisms used by MSCs to repair tissue damage and to promote tissue  and IL-1β protein expression in LA-mDPC6T cells cocultured with mBMSCs or CB-mBMSCs (n = 3). Data are displayed as the mean ± SD. Statistical significance was determined using one-way analysis of variance (*p < 0.05; **p < 0.01; ***p < 0.001).
F I G U R E 5 Legend on next page.
regeneration. 14 We next tested for TNT-mediated transfer of mitochondria between mDPC6T cells and mBMSCs. mBMSC-MitoRFP were cocultured with (LA-)mDPC6T-CFSE, and phalloidin was used to stain F-actin in the TNTs. We observed that mBMSCs and (LA-) mDPC6T cells physically connected via TNTs, which enabled the consecutive passage of mitochondria ( Figure 4A,B). We also observed the colocalization of RFP-labelled mitochondria within the TNTs. In addition, RFP-labelled mitochondria were detected in mDPC6T cells, supporting mitochondrial transfer from mBMSCs to mDPC6T or LA-mDPC6T cells via TNTs ( Figure 4A Figure 4A). We found that physical contact but not microvesicle release or exocytosis was a prerequisite for mitochondrial transfer (Appendix Figure 4A). In direct coculture, pre-treatment of mBMSCs with CB significantly inhibited its protective effect on the mitochondrial function, which was reflected in decreased TMRM and increased MitoSOX levels

| TNF-α activates the NF-κB signalling in mBMSCs and promotes mitochondrial transfer
Previous results have indicated that mBMSCs prefer to donate their mitochondria to injured mDPC6T cells. We next investigated the potential molecular mechanism underlying this phenomenon. The GO and KEGG pathway enrichment analyses revealed marked enrichment of the inflammatory response and cytokine-cytokine receptor interaction between the mDPC6T group and the LA-mDPC6T group (Appendix Figure 5A). Therefore, we speculate that LA-mDPC6T cells We then aimed to identify the exact inflammatory factors responsible for activating the NF-κB pathway in mBMSCs. Using neutralizing antibodies, we blocked TNF-α, IL-6 and CXCL1 in a transwell system and examined NF-κB pathway-associated protein expression in mBMSCs. Only inhibition of TNF-α prevented upregulation of p-IKKβ, p-IκBα and p-p65 protein expression in mBMSCs cocultured with LA-mDPC6T cells (Appendix Figure 6A,B) and suppressed TNT formation  3). Data are displayed as the mean ± SD. Statistical significance was determined using one-way analysis of variance (*p < 0.05; **p < 0.01; ***p < 0.001).
(Appendix Figure 6C,D). Blocking TNF-α also reduced the mitochondrial transfer from mBMSCs to mDPC6T cells during direct coculture (Appendix Figure 6E). Conversely, recombinant TNF-α alone sufficed to induce NF-κB protein expression levels similar to those of LA-mDPC6T cells in mBMSCs (Appendix Figure 6F,G). Treatment of mBMSCs with TNF-α alone upregulated TNT formation (Appendix Figure 6H,I). Moreover, the addition of TNF-α to the culture medium during direct coculture promoted mitochondrial transfer from mBMSCs to mDPC6T cells (Appendix Figure 6J). Together, LPS + ATP induced mitochondrial dysfunction and subsequent NLRP3 activation and pyroptosis in mDPC6T cells, resulting in the production of a variety of inflammatory cytokines, such as IL-1β and IL-6. Among these inflammatory cytokines, we identified TNF-α as a major factor in promoting TNT-mediated mitochondrial transfer during the coculture of mBMSCs with mDPC6T cells and confirmed that TNT formation was NF-κB-dependent. Importantly, mitochondria from mBMSCs are selectively transferred to injured mDPC6T cells, which protects mDPC6T cells from injury.

| DISCUSSION
With the progression of caries, odontoblasts are affected first because they are the first line of cells to come into contact with oral bacteria. 24 The bacteria-induced mitochondrial damage of odontoblasts is associated with the aberrant production of mtROS and is thought to contribute to the pathogenesis of pulpitis. 4 Given the essential role of mitochondrial homeostasis in pulpitis, we postulated that delivery of heathy mitochondria to odontoblasts might compensate for mitochondrial damage and restore a basal metabolic capacity. In this study, we provide direct experimental evidence that mBMSCs can provide their mitochondria to injured mDPC6T cells through TNTs, thus preserving mitochondrial oxidative metabolism and bioenergetic requirements. Moreover, we revealed that the paracrine effect of mDPC6T cells on mBMSCs is the major driver of mitochondrial transfer.
Mitochondria are essential for energy metabolism and normal cellular function. Any mitochondrial abnormality can cause a bioenergetic crisis, eventually leading to inflammation and cell death. 25,26 Recent findings suggest that mtROS released from damaged mitochondria are the major trigger of NLRP3 inflammasome activation. 3,21 Consistently, our results showed that increased mtROS mediated NLRP3 inflammasome activation and subsequent pyroptosis of mDPC6T cells. In addition, mitochondrial damage facilitates the recruitment of NLRP3 to the mitochondria and may enhance its activation by allowing for efficient sensing of mtROS. 27 Here, we found a similar phenomenon, in which inhibition of mtROS decreased NLRP3 mitochondrial localization and inflammasome activation. These findings provide new mechanistic links among NLRP3 inflammasome activation, mitochondrial dysfunction and pyroptosis during pulpitis.
Intercellular mitochondrial transfer has recently been described as a new mechanism to rescue mitochondrial function in injured cells. 28,29 In this study, we demonstrated that mBMSCs could donate their mitochondria to mDPC6T cells, thereby improving mitochondrial function and protecting mDPC6T cells from pyroptosis. RNAsequencing analysis further confirmed that mitochondria transferred from mBMSCs to mDPC6T cells may contribute to the mitochondrial biogenesis of mDPC6T cells. These results indicated that BMSCs could induce mitochondrial metabolic reprogramming in damaged cells and that mitochondrial donation could sustain the energy supply and redox balance.
Various routes are involved in mitochondrial transfer, such as TNT formation, microvesicle release, gap junctions and other routes of transfer. 23 In this study, we revealed that TNTs are the main route of mitochondrial transfer between mBMSCs and mDPC6T cells. More interestingly, TNT formation in mBMSCs was increased in the pyroptosis model, and the mitochondrial transfer efficiency from mBMSCs to mDPC6T cells was also enhanced. The cells displaying numerous aberrancies and injuries are more permissive to mitochondrial transfer than their healthy cell counterparts. 28 A variety of stress signals released by recipient cells are sensed by MSCs resulting in enhanced mitochondrial biogenesis and TNT formation by MSCs through retrograde signalling, thereby preparing MSCs for mitochondrial donation. 6,10 In addition, inflammatory factors released by cells suffering from OS and pyroptosis have also been postulated to trigger donation of mitochondria. 30 Our results found that pyroptotic mDPC6T cells released TNF-α that induced the TNT formation in mBMSCs via the NF-κB pathway, which enhanced mitochondrial donation. The NF-κB pathway has been shown to be involved in the TNT formation and stability by modulating actin interactions. 31 In addition to NF-κB, M-Sec-induced actin polymerization is an important initiating step of TNT formation. 32 Following initiation, the expression of the actin regulatory protein Eps8, mammalian target of rapamycin (mTOR) and CDC42 appears necessary for the elongation of TNTs. 14 Furthermore, Miro1, a Rho GTPase that helps mitochondrial movement along microtubules, is rate-limiting for mitochondrial transfer. 33 However, the exact regulatory mechanism of mitochondrial transfer and TNT formation remains unknown, and further experiments are required to investigate these cellular processes.
In conclusion, we revealed that mBMSCs promote reparative mechanisms in pulpitis through TNT-mediated mitochondrial transfer. These findings highlight a potential clinical application of mitochondrial transfer and provide an important basis for the future development of mitochondria-targeted therapy for dental pulp repair.

AUTHOR CONTRIBUTIONS
Konghuai Wang carried out the experimental work, the collection and interpretation of data and drafted the manuscript; Lu Zhou participated in the study design, the collection and analysis of data and preparation of the manuscript; Hanqing Mao and Jiayi Liu carried out the study design, the analysis and interpretation of data and critically revised the manuscript; Zhi Chen and Lu Zhang conceived and designed the study and critically revised the manuscript. All authors have approved the final manuscript and have agreed to be accountable for all the aspects of the work.