All related operations involving animals follow the “Guidelines for Care and Use of Laboratory Animals” (NIH Publication No.85-23, revised 1996) published by the US National Institutes of Health (Bethesda, MD). And approved by the Animal Care and Use Committee at the Wenzhou Medical University (wydw2020-0634). All Male C57/BL6 mice (6–7 weeks old and 20–25 g weight) were purchased from the SLAC Laboratory Animal Center of Shanghai (Shanghai, China).

The myocardial I/R model was operated as our previous study (21). In detail, the mice were first anesthetized with isoflurane and intubated. Then a thoracotomy was performed to expose the heart in the third or fourth rib. The left anterior descending coronary artery was ligated with a 7-0 silk thread to cause myocardial ischemia. After 30 min of ischemia, the silk thread was untied and the blood was reperfused for 4 h. Mice in the sham group underwent the same operation but were not ligated. The operation uses an animal anesthesia machine/ventilator from RWD Life Science, the pressure is adjusted to 0.4 after isoflurane perfusion, flow rate two, stroke volum (μl) 300, strokes/min 180. The isoflurane pumping was stopped immediately after the start of reperfusion, and the air was maintained until awakening.

We selected a total of 6 mouse heart samples, of which three were treated by sham operation, and three were ischemia for 30 min and reperfused for 4 h. After the samples were collected, they were transported to Beijing on dry ice in accordance with the biological sample storage requirements, and Beijing Novogene completed high-throughput second-generation sequencing.

After all C57/BL6 mice were reared to 7 weeks, they were randomly divided into five groups: (a) Sham group-mice were subjected to sham operation and treated with vehicle (saline); (b) I/R group- mice were subjected to I/R intervention and treated with vehicle (saline); (c) I/R+CQ group-mice were subjected to I/R intervention after intraperitoneal injection of chloroquine and treated with vehicle (saline); (d) I/R+DAPA group-mice were subjected to I/R intervention and treated with dapagliflozin (40 mg/kg/day); (e) I/R+ DAPA+CQ group-mice were subjected to I/R intervention after intraperitoneal injection of chloroquine and treated with dapagliflozin (40 mg/kg/day). Treatment with dapagliflozin (MCE, China, 40 mg/kg/day) or vehicle by gavage was started 7 days before the cardiac I/R intervention. The dose of chloroquine (Sigma, USA) is 10 mg/kg, which is injected intraperitoneally 10 min before each operation, which has been mentioned in our previous article (21).

Evans Blue/Triphenyltetrazolium chloride (TTC, Sigma, USA) staining was used to detect the myocardial infarct size induced by I/R injury. The coronary artery was ligated again, and the abdominal aorta was injected with 200 μL of 2% Evans Blue, the heart was immediately taken out, placed in a mold for sectioning, and then incubated in 1% TTC staining solution at 37°C for 30 min, after finishing overnight in 4% paraformaldehyde and filming. The images were analyzed using Image Pro software (Media Cybernetics, Inc. Bethesda, MD), and infarct size was calculated using a method as previously described (21), infarct size was compared when area at risk that were similar among groups.

Blood was taken 4 h after reperfusion, and the serum was separated by centrifugation at 7,500 rpm for 15 min, and immediately detected or temporarily frozen at −80°C. Serum TnI concentration was measured using an enzyme-linked immunosorbent assay (ELISA) measurement of Troponin I kit (Elabscience, E-EL-M0086c). Serum LDH activity was measured using a Lactate dehydrogenase (LDH) assay kit (Nanjing jiancheng, China). Serum creatine kinase-MB (CK-MB) was measured with a creatine kinase MB isoenzyme Assay Kit (Nanjing jiancheng, China).

Serum concentrations and cell culture supernatant medium level of IL-1β were measured by the corresponding ELISA kits (Elabscience, E-EL-M0037) according to the manufacturer's instructions.

The cardiac activity of caspase-1 was calculated via Colorimetric Assay Kit (Solarbio, China). First, the total protein was extracted according to the instructions and formulated into a 2 mg/ml system.

Immunofluorescence staining is used in myocardial tissues and cells. Use frozen tissue sections or 4% paraformaldehyde-fixed cells, rupture the cell membrane with 0.3% Triton for 10 min, block with 5% BSA for 1 h at room temperature, and then incubate with the corresponding primary antibody at 4°C overnight. Incubate with the corresponding secondary antibody for 1 h in the dark the next day, and finally add DAPI to stain the nucleus. After completion, add anti-fluorescence quenching solution and take pictures under the microscope (Leica, Germany).

Isolate primary cardiomyocytes from the heart of C57/BL6 mice born within 1–3 days. The specific method can refer to the study by Ehler et al. (22). The detailed steps for isolation of primary cardiomyocytes are written in the Supplementary Material.

These PCCMs were then used to simulate the H/R model (22). Hypoxic conditions was performed in a hypoxic incubator with an N₂ concentration of 94% and a O₂ concentration of 1% and a CO₂ concentration of 5% for 6 h in medium deprived of serum. Then PCCMs were replaced with normal medium and put into ordinary incubators (CO₂ concentration of 5% and O₂ concentration of 21%) for 18 h to simulate reoxygenation injury. For in vitro experiments, DAPA(10 μM) was administered for 1 h (23) and CQ was administered for 24 h at 10 μM (22) in advance. Before starting H/R, replace with a new medium.

The PCCMs were plated in 24-well plates and treated separately. After processing, use the Cell Counting Kit-8 (Beyotime, China) to measure the number of viable cells according to the manufacturer's instructions. At the same time, the upper medium of PCCMs after treatment was tested for cytotoxicity with LDH Assay Kit.

Hoechst 33342/PI (Beyotime, China) has been reported to be used to label dead or apoptotic cells. We seeded PCCMs on cell slides in 24-well plates. After completing the corresponding treatments, we added Hoechst 33342 and PI working solution in order according to the manufacturer's instructions. Finally, we added the anti-fluorescent quenching solution and took it under a fluorescent microscope for observation.

Lysotracker Red can be selectively retained in the acidic lysosomes, thereby achieving specific fluorescent labeling of lysosomes (24). In our experiments, we used Lysotracker Red (Beyotime, China) to detect the number of intact lysosomes in an H/R model simulated by PCCMs.

PCCMs lysates were centrifuged for 10 min at 14,000 g at 4°C. 500 μL of the lysate (1000 μg of total protein) was added to the monoclonal primary antibody of the target protein at 4°C overnight. The next day, agarose beads were added and incubated for 3 h. The agarose beads were washed 3 times and then mixed with SDS for western blot analysis. The above reagents were all purchased from absin (abs955, China).

First, the PCCMs are grouped and processed as before. After the H/R treatment, dilute Fluo-4AM (Beyotime, China) and SBFI AM to working concentrations according to the instructions, add them to the culture medium, incubate at 37°C for 30 min, wash with PBS and then put them back into the cell incubator for 30 min, and select the 334/505 wavelength for detection.

For overexpression plasmid transfection, 1 μg plasmid was added in 500 μl of Opti-MEM, followed by addition of 3 μl of Lipofectamine RNAiMAX in 500 μl of Opti-Mem and incubated for 5 min. The two mixtures were pooled and incubated further for 10 min at RT. The respective transfection mixture was then added to the cells and mixed by gently swirling the plate. The plate was then incubated at 37°C for 18 h in a 5% CO2 incubator. After incubation, media was added to the cells, and the cells were further incubated until harvest i.e., 48 h from the beginning of transfection.

Myocardial tissue/cell total protein preparation and concentration determination kits are from Beyotime (Shanghai, China). Proteins of different molecular weights were separated on SDS-PAGE gels, transferred to 0.22 μm PVDF membranes, and blocked with skimmed milk powder then incubated with primary antibodies of NLRP3 (A12694, ABclonal), Caspase-1 (22915-1-AP, proteintech), ASC (abs135824, absin), IL-1β (16806-1-AP, proteintech), Cleaved-IL-1β (A1112, ABclonal), Atg-5 (12994T, CST), Beclin-1 (3738S, CST), P62 (23214S, CST), LC3B (ab243506, abcam), LAMP1 (3243S, CST), LAMP2 (66301-1-Ig, proteintech),CTSD (21327-1-AP, proteintech), CTSB (12216-1-AP, proteintech), and GAPDH (10494-1-AP, proteintech) at 4°C overnight. After incubating with the corresponding secondary antibody the next day, protein bands were scanned by a ChemiDoc MP device (Bio-Rad, Hercules, CA, USA). Finally, ImageJ software (NIH, USA) was used for analysis.

Data were presented as mean ± SD, and were analyzed using SPSS version 21.0 (IBM, Armonk, NY, USA). All the data we used were normally distributed. A two-tailed unpaired Student's t-test was conducted to compare two experimental groups or one way analysis of variance (ANOVA) followed by Duncan's T3 multiple-range test was used for the comparison of more than two groups. P < 0.05 were considered as statistically significant.

We first evaluated the effect of DAPA on myocardial I/R injury by assessing myocardial infarct size. After DAPA treatment, the area of myocardial infarction in mice was significantly decreased for area at risk that were similar among groups (Figures 1A–C). DAPA significantly decreased cTnI, CK-MB, and LDH levels after I/R (Figures 1D–F).

Coincidentally, we also found that DAPA treatment significantly improved cell viability of PCCMs after hypoxia/re-oxygenation injury (H/R). The optimal concentration of DAPA was 10 μM (Figure 1I). Subsequent treatment with 10 μM DAPA reduced LDH in the cell supernatant (Figure 1J) and a significant decrease in myocardial cell death (Figures 1G,H).

Many factors cause myocardial I/R injury; to explore which factor is the main factor in our model, we performed high-throughput sequencing on the I/R injury area. Sequencing results suggest that inflammatory response is the central part of the biological process (Figure 2A).

We speculated that DAPA treatment would reduce the inflammatory response in I/R injury. First, we detected inflammatory proteins in the tissues and found that DAPA treatment significantly reduced the expression levels of Pro-IL-1β and Cleaved-IL-1β (Figures 2B–D). We also detected the secretion level of IL-1β in the serum of mice. It was found that DAPA treatment also reduced the secretion of IL-1β in the serum of I/R mice (Figure 2E). Coincidentally, DAPA also improved H/R injury of PCCMs in vitro. DAPA treatment decreased expression of Pro-IL-1β and Cleaved-IL-1β in vitro (Figures 2F–H) and reduced the secretion of IL-1β in the PCCMs culture medium (Figure 2I).

Through sequencing, we have found that the inflammatory response dominates in our I/R model, and DAPA treatment inhibits this inflammatory response. In order to explore how inflammation is activated here, we performed KEGG pathway analysis on the sequencing results and screened the top 10 using “count number” as the standard and found that the signaling pathways directly related to inflammation are “NF-kappa B” and “NOD-like receptor” (Figure 3A). And through screening with “fold enrichment” as the standard, it is found that “NOD-like receptor” is more important (Figure 3B).

So we selected NLRP3 inflammasome as the verification target. By detecting the level of NLRP3 inflammasome-related protein, we found that after DAPA treatment, the expression of NLRP3 and ASC were reduced (Figures 3C–F,J,K), and the production of Cleaved-caspase-1 was also significantly inhibited (Figures 3C,D,G,H,L,M). On the other hand, the enzyme activity of caspase-1 activated during I/R injury also decreased significantly after DAPA treatment (Figure 3I).

Recently, autophagy has been implicated in I/R (25) and may be related to inflammation (19). We analyzed the expression of autophagy-related proteins by Western blot analysis to determine the effect of DAPA on autophagy in myocardial I/R injury. The results showed that LC3B-II, a representative protein of autophagy, was significantly increased in I/R, accompanied by an increase of Atg-5 and Beclin-1, which are related to autophagosome production (26). P62, a ubiquitinated protein that may indicate autophagic flux (26), was also increased (Figures 4A–E,G–K). The formation of autophagosomes was increased, and the destruction of autophagosomes was reduced, which led to the accumulation of autophagosomes.

DAPA treatment reduced the expression of LC3B-II, Atg-5, beclin-1, and P62 (Figures 4A–E,G–K). We next explored whether the reduction of autophagosomes after DAPA treatment was due to inhibition of autophagosome production, increased destruction, or both. To show that this phenomenon occurs in myocardial cells, we co-stained LC3 and cTnT (cardiac-specific markers) and measured their relative concentrations by immunofluorescence (Figure 4F). The results were consistent with those of Western blot assays. When PCCMs were stained for LC3 and imaged by immunofluorescence microscopy, the results were the same: pre-treatment of DAPA reduced the accumulation of autophagosomes (Figure 4L).

Moreover, the expression of P62 was reduced after the addition of DAPA. P62 is a selective autophagy receptor and is degraded by autophagy; thus, increased levels of p62 reflect the inhibition of autophagy (27). Combined with our previous results, we speculated whether there is a relationship between this blocked autophagy flux and increased inflammation and whether hindering autophagosome degradation affects the inflammation change in I/R injury.

One of the normal elements of autophagy flux, the successful degradation of autophagosomes, requires the participation of normally-functioning lysosomes (28). We used chloroquine (CQ) to disrupt lysosomal function and block autophagosome degradation (22).

After CQ exposure, we first detected autophagy-related proteins by Western blotting to determine whether CQ inhibition was achieved. Compared with the DAPA group, the DAPA + CQ group showed a significant increase in LC3B-II (Figures 5A,B), indicating an increase in autophagosome accumulation. Similarly, in the DAPA + CQ group, levels of P62 were significantly higher than those in the DAPA group (Figures 5A,C). Therefore, we inferred that the degradation of autophagosomes was reduced after the addition of CQ. We concluded that the use of CQ in our model successfully inhibited the degradation of autophagosomes. We also examined whether this phenomenon occurred in myocardial cells. Using tissue immunofluorescence technology, we found that the results on both sides were consistent (Figure 5D). We also discovered that in the comparison between the I/R group and the I/R + CQ group, the expression of LC3B-II and P62 in the I/R + CQ group was further enhanced. Blocked degradation of autophagosomes when I/R occurred appeared to be incomplete.

We then asked whether I/R damage changed after using CQ. We observed that damage in the DAPA + CQ group increased following the addition of CQ. Compared with the DAPA group, the DAPA + CQ group increased the myocardial infarction area (Figures 5E–G). Similarly, we tested myocardial infarction markers in the serum, including cTnI, CK-MB, and LDH, and found that the treatment effect of the DAPA group was reversed by CQ (Figures 5H–J). Compared to the I/R group, the I/R + CQ group results were exacerbated to varying degrees (Figures 5E–J). Therefore, the elimination of autophagosomes plays a more critical role in myocardial I/R injury. Impaired clearance will increase I/R damage. Conversely, promoting the elimination of autophagosomes can reduce I/R injury, and DAPA may play an important role here.

What is the relationship between clearance of autophagosomes and I/R inflammation? When autophagosome clearance was blocked by CQ, inflammation in the DAPA + CQ group was significantly greater than that in the DAPA group. The most direct manifestation was the increased expression of pro-IL-1β and cleaved-IL-1β (Figures 5K,P,Q). Serum IL-1β secretion increased after CQ administration (Figure 5R). The expression of NLRP3 and ASC also changed similarly as inflammation (Figures 5K,L,M). With cleaved-caspase-1 (Figures 5K,N,O,S), which represents the protein that inflammasomes activate for further function, inflammation in the I/R + CQ group was further exacerbated than the I/R group. The observations suggest that clearance of autophagosomes in I/R injury is also closely related to the occurrence and progression of inflammation.

Combining these observations, we found that DAPA may play a role by restoring autophagy flux and autophagosome degradation, which was damaged by I/R injury.

Next, we explored whether these results would occur in isolated myocardial cells using CQ and monitoring marker molecules related to autophagy. As shown in Figures 6A–D, when the pathway of autophagosome degradation was inhibited, the number of autophagosomes in PCCMs increased. Coincidentally, under H/R conditions, autophagosome clearance was blocked incompletely, which helps to clarify whether the changes in vivo and in vitro were consistent.

We tested whether damage after H/R injury changed after CQ inhibited autophagosome clearance in PCCMs. The addition of CQ further aggravated the damage of H/R, which can be seen from the comparison between the H/R and H/R + CQ groups (Figures 6G,H). Conversely, when comparing the DAPA + CQ group with the DAPA group, CQ eliminated the improvement attributable to DAPA (Figures 6G,H). Hoechst/PI double-stained PCCMs also showed the same change in the percentage of PCCMs deaths (Figures 6E,F). To further explore whether H/R injury and DAPA-related changes are related to inflammation, we examined the expression of inflammation-related proteins. As expected, CQ reversed the DAPA induced reduction of inflammation and aggravated H/R-induced inflammation, manifested by upregulation of pro-IL-1β and cleaved-IL-1β protein levels (Figures 6I,N,O) and increased IL-1β secretion in the cell supernatant (Figure 6P). In addition, observing the NLRP3 inflammasome index, the results showed that after CQ suppressed autophagosome clearance, levels of NLRP3, ASC, and cleaved-caspase-1 increased (Figures 6I–M).

These observations suggest that H/R-impaired autophagosome clearance cause damage, which DAPA can partially offset by promoting autophagosome clearance.

How does DAPA reduce I/R-induced inflammation and improve I/R injury by promoting autophagy? Autophagy can remove toxic waste, including some compounds that affect excessive inflammation (29). In addition, inflammasome components, including NLRP3, can be cleared by autophagy (19). Based on these reports, we believe that autophagosomes in our model system clear NLRP3 and that DAPA may play an important role.

We used co-immunoprecipitation and fluorescence co-localization to investigate this phenomenon further. Through Western blot analysis, we observed an interaction between P62 and NLRP3, possibly promoted by DAPA, in the H/R and H/R + DAPA groups (Figures 7A–D). Fluorescence results identified co-localization of autophagosome marker LC3 and inflammasome marker NLRP3 (Figure 7E). These findings suggest that eliminating autophagosomes may remove the inflammasome component NLRP3, and the application of DAPA may accelerate the phagocytosis of NLRP3 by autophagosomes.

Autophagosomes phagocytose the inflammasome component NLRP3, and although DAPA may strengthen this process, it is insufficient to degrade the autophagosomes adequately. We have shown that DAPA can promote the clearance of autophagosomes in I/R, but the successful clearance of autophagosomes requires normal lysosomes for proper degradation. Disruption of lysosomal function or impaired lysosomal degradation capacity will affect autophagy flux (28).

To determine if DAPA improved the function of lysosomes in I/R, we used Lysotracker red probe for acid staining to mark normal lysosomes. As shown in Figure 7J, the number of lysosomes with an acidic environment decreased in the H/R group but increased in the H/R + DAPA group. In addition, we detected some indicators of lysosomes at the protein level, such as LAMP1, which closely tracks lysosome numbers, and LAMP2, a critical protein of autophagosome-lysosome fusion (18), that showed reduced expression in the H/R group but were upregulated in the H/R + DAPA group (Figures 7F–H). CTSD (Cathepsin D) and CTSB (Cathepsin B), typical lysosomal labeled proteases, are involved in the degradation of autophagosomes (30). We found that CTSD expression was downregulated in the H/R group but increased after DAPA treatment (Figures 7F,I). We used LAMP2 to co-stain with CTSB and observed a decrease in fluorescence in H/R; the overlap between LAMP2 and CTSB was also reduced (Figure 7K). These results suggest that (a) the lysosomal membrane permeability changed in H/R, resulting in a decrease in the number of lysosomes with a typical acidic environment, and (b) some cathepsins in the lysosome were released into the cytoplasm. Although these factors may be responsible for reducing autophagosome degradation, DAPA appears to attenuate this damage.

Studies have suggested that SGLT-2i can act on NHE1 (sodium hydrogen ion transporter 1) on cardiomyocytes and inhibit its activity (12), thereby producing physiological effects; others believe that it is myocardial protection produced by NCX (sodium-calcium ion transporter) effect (31). We used Auto Dock Tools to perform molecular simulations on dapagliflozin and NHE1 (PDBID: 3v5s) and NCX (I-TASSER) and found that dapagliflozin and NHE1, and NCX have binding ability (Figures 8A–E). Further, we observed that after DAPA treatment, Ca²⁺ and Na⁺ in PCCMs have been effectively alleviated.

We used plasmids to upregulate the activities of NHE1 Figures 8F,G and NCX in PCCMs to prove whether DAPA's intervention is related to these two receptors. The results showed that as the activity of NHE1 increased during H/R injury, the assembly and activation of NLRP3 inflammasomes also increased (Figures 8H–N), indicating that the increased activity of NHE1 eliminated the treatment brought by DAPA. Similarly, the upregulation of NCX activity also showed the above results (Figures 8O–U).